We want to provide articles for the assignment. There are 21 articles. can you add 5 more
sustainability
Article
Revisiting the Relation between Renewable
Electricity and Economic Growth:
A Renewable–Growth Hypothesis
Minyoung Yang and Jinsoo Kim *
Department of Earth Resources and Environmental Engineering, Hanyang University, 222 Wangsimni-ro,
Seongdong-gu, Seoul 04763, Korea; yminy94@hanyang.ac.kr
* Correspondence: jinsookim@hanyang.ac.kr; Tel.: +82-2-2220-2241
Received: 13 March 2020; Accepted: 10 April 2020; Published: 13 April 2020
����������
�������
Abstract: Global concern about the climate crisis has incited movements for switching to renewable
electricity. Renewable electricity can contribute to economic growth as an input factor (electricity
generation) and also as an industry (renewable manufacturing). We introduce a new hypothesis,
the renewable–growth hypothesis, to investigate the role of the renewable manufacturing industry in
the energy–growth nexus study. To test the hypothesis, we select a target country group using the
market share of the renewable manufacturing industry and conduct the Granger causality test for solar
photovoltaic and wind power. The autoregressive distributed lag bounds testing approach is applied
for the causality test. The results show that renewable electricity Granger causes economic growth in
target countries, which supports the renewable–growth hypothesis. However, the hypothesis did not
hold in countries that export renewable power facilities more than they install them for domestic
demand. We believe that the renewable–growth hypothesis would be secured soon if renewable
electricity expands broadly over the world.
Keywords: renewable–growth hypothesis; renewable electricity; economic growth; renewable
manufacturing; energy–growth nexus
1. Introduction
International attention to global warming comprises the global effort for carbon reduction.
However, if we keep our pledges at the current level, the world’s temperature will increase to almost
twice the limit referred to in the Paris Agreement by the end of the century. Climate change is
now referred to as the “climate crisis” [1]. This phenomenon also draws attention to the topic of
energy, particularly in the electricity industry. As part of that, many countries have displayed their
transition to renewable energy from fossil fuels, which are considered the main source of carbon
emissions. However, many governments, especially those in Asia, are still using coal-fired power
stations. Even worse, the demand growth for gas as an alternative to coal is emerging [2]. Transitioning
to renewable energy is a means for solving problems caused by the climate crisis.
Indeed, renewable energy accounted for an estimated 18.1% of total final energy consumption
(TFEC) in 2017. Modern renewables composed 10.6% of TFEC, with an estimated 4.4% growth in
demand compared to 2016. Particularly in the power sector, renewable energy has grown to account
for more than 33% of the world’s total installed power-generating capacity in 2018. Solar photovoltaics
(PV) comprises 55% of renewable capacity, after the additional installation of around 100 gigawatts in
2018, and is followed by wind power (28%) [3].
Economic development and sustainability are important not only for carbon reduction but also for
the global trend promoting renewable energies [4–6]. Developed economies promote renewable sources
Sustainability 2020, 12, 3121; doi:10.3390/su12083121 www.mdpi.com/journal/sustainability
http://www.mdpi.com/journal/sustainability
http://www.mdpi.com
https://orcid.org/0000-0001-6496-1714
http://dx.doi.org/10.3390/su12083121
http://www.mdpi.com/journal/sustainability
https://www.mdpi.com/2071-1050/12/8/3121?type=check_update&version=
2
Sustainability 2020, 12, 3121 2 of 22
to strengthen the energy security of supply and control their greenhouse gas emissions [7]. Similarly,
understanding the causal relationship between renewable energy and economic growth is significant
for a country’s economic development and as the basis for policymakers. Therefore, extensive research
has been conducted on the relationship between them. In the 1980s, many studies investigated the
relationship between energy consumption and economic growth [8]. “Energy–growth nexus” is a term
referring to the link between energy consumption and economic growth. The energy–growth nexus
has reached a more disaggregated level over time [9].
An analysis of renewable energy growth is needed considering the described circumstances.
The facility cost of renewable electricity accounts for a higher percentage in power generation costs
compared to non-renewable energy [10]. This means that, if we increase solar facilities, the demand
for all sectors related to them also increases. This might lead to positive effects on economic growth.
Furthermore, renewable electricity is capital intensive and has a value-added effect. According
to Ernst & Young [11], the solar industry in the European Union 28 represented more than 81,000
full-time equivalents and more than EUR 4600 million gross value-added (GVA). They mentioned the
installed capacity has a significant impact on job and GVA creation because there is a direct impact
on manufacturing and services needed. Thus, the increasing demand for renewable electricity has a
positive effect on such industries and is related to economic growth.
However, if a country generally imports equipment and produces its own electricity, renewable
electricity demand has less or no positive impact on economic growth. Given the cost structure of
the renewable electricity industry, growth is driven by demand growth in a related facility. It has a
different growth mechanism when a country simply purchases equipment. If so, it is likely to have a
positive impact on the economy of the country producing that facility. This means that if demand for
renewable electricity increases, there is no induced effect like great value-added or additional demand
in other industries or there is only an impact on the country’s imports concerning the overall facility.
Therefore, the relationship between renewable electricity demand and economic growth may not be
evident in these countries.
There are four hypotheses on the energy–growth nexus considering the number of cases that
could be the result of the analysis [12]. First, the “neutrality hypothesis” implies the absence of
a causal relationship between energy consumption and GDP. The second one, the “conservation
hypothesis,” suggests that energy conservation policies may be implemented with little or no adverse
effect on economic growth. In the third one, the “growth hypothesis” means that energy consumption is
important for economic growth. The last one that implies bidirectional causality between the two factors
is the “feedback hypothesis.” It is possible to derive energy policy through an analysis of the above
hypothesis. The energy conservation policy means reducing energy consumption for economic growth.
If a specific country’s energy–growth nexus follows the growth hypothesis, energy conservation
policies may have an adverse impact on economic growth. However, if under the neutrality hypothesis,
energy conservation policies may not have much impact. The main purpose of the energy–growth
nexus study is to examine which hypothesis is investigated in a specific country or group of countries.
The results of the energy–growth nexus study have yielded mixed results within hypotheses. Existing
hypotheses are focused on energy itself and do not consider specific industries. Therefore, the existing
hypothesis is not enough to cover the impact of renewable energy manufacturing.
Due to the feature originating from the renewable manufacturing industry, renewable electricity
and economic growth will show a positive relationship. Thus, this study tries to fill the gap with a
new hypothesis, which is the “renewable–growth hypothesis.” We analyze the time-series data of
countries to confirm the renewable–growth relationship. The main contributions of this study are as
follows. First, this study investigates if the development of the renewable energy generation sector has
a positive effect on economic growth considering the features of the renewable power sector. Second,
a new perspective of renewable–growth hypothesis is proposed, and the national group supporting
the analysis is established. This makes it possible to draw implications for the policy direction of
Sustainability 2020, 12, 3121 3 of 22
fostering the renewable electricity industry. Furthermore, this study presents the issues to promote
further studies.
The paper is organized as follows. Section 2 provides a review of the literature. Section 3
describes the countries to be analyzed and the data to be used for analysis and explains the model
and methodology. Section 4 provides empirical results. Section 5 concludes the paper and provides
remarks and policy implications.
2. Literature Review
2.1. Overview of Energy–Growth Nexus
The energy–growth nexus has been studied in order to confirm the direction of the causality and
its hypothesis has been well developed via various studies (see Table 1). They can be divided into
country-specific studies [13–16] and multi-country studies [17–19]. Some of the studies attempted
to examine the causal relationship in the energy–growth nexus in both developed and developing
countries or by using different period data [8–10,18,20–22]. A study can be conducted for a specific
research purpose. Pao and Fu [23] tried to investigate the relationship between economic growth
and energy consumption in Brazil. Contrary to the previous study, this study covered various types
of energy. They found mixed results: A conservation hypothesis between non-renewable energy
consumption and economic growth, a growth hypothesis between non-hydroelectric renewable energy
consumption and economic growth, and a feedback hypothesis between economic growth and total
renewable energy consumption. Others focused more on the causal relationship between economic
growth and energy consumption [24,25]. Apergis and Danuletiu [24] examined the relationship
between economic growth and renewable energy consumption for 80 countries using the long-run
causality test. They concluded that there is a bidirectional causality between renewable energy
consumption and economic growth in the long run. Kazar and Kazar [25] investigated the relationship
between development and renewable electricity net generation values for 154 countries with a panel
analysis. They found the presence of bidirectional causality in the short run and that the causal
relationship differs both in the short run and long run depending on the human development level.
2.2. Electricity and Economic Growth
The energy–growth nexus has been studied in various countries and on a more disaggregated
level [9]. In addition to the interest in climate change, many countries and policymakers have tried
to implement an electricity conservation policy. The confirmed results of research have been used
as a basis for implementing such a policy and to establish the right policy direction for the country.
The electricity–growth nexus for a single country has been studied and developed for that reason.
Ramcharran [26] studied the electricity–growth nexus in Jamaica and found that the country is energy
dependent. Ghosh [27] and Narayan and Smyth [28] investigated energy consumption and economic
growth in India and Australia, respectively. They found Granger causality from economic growth
to electricity consumption in both countries. In addition, research is being conducted in various
countries such as Korea [29], Bangladesh [30], Cyprus [31], China [32], Turkey [33], Malaysia [34],
Lebanon [35], and Nigeria [36]. However, due to the omitted variable bias, the studies that use only
energy consumption and economic growth as variables should be interpreted with caution. In that
context, some studies have attempted to make econometric transformations by adding employment
variables to the bivariate model [37] or setting up the model reflecting the structural beaks [38].
Lorde et al. [39] constructed a multivariate model using the neoclassical production model to examine
the economic theory. In various studies, bivariate and multivariate models were analyzed, in order or
simultaneously. The impact on economic growth has been studied for the implementation of national
power-related policies until recently [40–43].
An electricity–growth nexus may be established by setting up a group of countries depending
on the research purpose. Yoo [44] analyzed the relationship between electricity consumption and
Sustainability 2020, 12, 3121 4 of 22
economic growth in ASEAN countries for similar characteristics in the electricity sector. This study
tried to confirm the relationship between electricity consumption and economic growth within a
similar-featured group but with some differences in terms of investment in the power sector. Each of
the four countries showed similar results with the two countries, respectively. The Organization of the
Petroleum Exporting Countries (OPEC) is expected to be greatly influenced by energy conservation
due to the characteristics of oil-producing countries. However, despite the obvious similarities,
the results vary depending on the model and the time period [45,46]. This “no consensus” feature of the
electricity–growth nexus was observed among the countries in the group even when the econometric
method was modified. Acaravci and Ozturk [47] emphasized that they have results that conflict with
the existing literature through the identification of the Granger causality of panel data. At the same
time, they stated and emphasized that this issue deserves more attention and needs further research.
Some studies approach the characteristics of a country from an economic perspective. Apergis
and Payne [48] analyzed a total of 88 countries using the panel vector error correction models based on
the level of economic income. The study showed different results depending on whether a country
was assessed in terms of the short run or the long run and also depending on its development level.
In the long run, bidirectional causality exists in both high-income and upper-middle-income panels
and lower-middle-income country panels. The growth hypothesis is satisfied in lower-middle-income
country panels and low-income country panels in the short run. Even in the same group, mixed results
were observed according to the short-run and long-run views [49], but some studies show a consensus
between studies analyzing the same country [50].
Over time, research has been further disaggregated in various aspects, with national concerns
shifting to energy transitions rather than just energy conservation. As a result, studies distinguishing
renewable and non-renewable electricity from total electricity have begun [51–57]. Ibrahiem [51]
mentions the limitations of the structure of the Egyptian electricity market, such as the crude-oil
shortage, the need for renewable growth, or the transition on power mix. Apergis and Payne [52]
conducted the panel causality test for the emerging market, including Egypt. Apergis and Payne [53]
extend their work by examining the causal dynamics between renewable and non-renewable electricity
consumption and economic growth in Central America. They show the negative bidirectional causality
between renewable and non-renewable electricity consumption and conclude that the cause of this is
that imported petroleum products raise concerns about the security of the region’s energy supply. One of
their contributions is their investigation of the negative bidirectional causality between renewable and
non-renewable electricity consumption. They believe that this is due to imported petroleum products
raising concerns about the security of the region’s energy supply.
The analysis of renewable electricity is relevant to climate change and policies for global
warming mitigation. Based on the results of the study, each country is recommended to increase its
investment in renewable energy projects and vice versa (Table 1). According to the confirmed results
of Al-mulali et al. [54], they recommended that Latin American countries should encourage not only
the investment for renewable energy projects but also reduction in the role of non-renewable sources
in electricity consumption. In recent years, these recommendations have been implemented in a way
analysis of non-standard Granger causality [55–57]. In this way, it is possible to confirm the effect of
specific energy sources [55] or conduct the analysis not only on aggregate electricity models but also
those disaggregated into renewable and non-renewable models [57]. Despite the development of the
research scope by the economic and econometric approach, the field of energy–growth nexus must still
be investigated not only in terms of methodology but also in economics or policy.
Sustainability 2020, 12, 3121 5 of 22
Table 1. The previous study-related electricity growth nexus.
Author Country Period Methodology Finding
Ramcharran (1990) [26] Jamaica 1970–1986 Granger causality
EC → Y
Ghosh (2002) [27] India 1950–1997 Granger causality
Y → EC
Narayan and Smyth (2005) [28] Australia 1966–1999 Multivariate Granger causality Y → EC
Yoo (2005) [29] Korea 1970–2002
Cointegration
VECM (vector error correction model)
Brown parameter stability test
Y ⇔ EC
Yoo (2006) [44] 4 countries 1971–2002
Hsiao’s version of Granger causality
Standard Granger causality test
Y → EC
(Thailand, Indonesia)
EC → Y
(Singapore, Malaysia)
Mozumder and Marathe (2007) [30] Bangladesh 1971–1999
Cointegration
VECM
Y → EC
Zachariadis and Pashourtidou (2007) [31] Cyprus 1960–2004
Cointegration
Granger causality
VECM
Y ⇔ EC
Yuan et al., (2007) [32] China 1978–2004
Cointegration
Hodrick–Prescott (HP) filter
Granger causality
EC → Y
Squalli (2007) [45] All OPEC members 1980–2003
Cointegration
ARDL (autoregressive distributed lag) Bounds Test
Toda and Yamamoto causality test
Y → EC
(Indonesia, Libya, Iraq, Algeria,)
EC → Y
(Kuwait, Venezuela)
Y ⇔ EC
(Iran, Venezuela, Qatar)
Mixed outcomes with different models
(Nigeria, Saudi Arabia, Indonesia, Kuwait,
and UAE)
Halicioglu (2007) [33] Turkey 1968–2005
ARDL bounds test
Granger causality
Y → EC
Tang (2008) [34] Malaysia 1972–2003
ARDL bounds test
Toda and Yamamoto causality test
Brown parameter stability test
EC → Y
Abosedra et al., (2009) [35] Lebanon 1995–2005 Granger causality EC → Y
Akinlo (2009) [36] Nigeria 1980–2006
Cointegration
Granger causality
VECM
EC → Y
Ghosh (2009) [37] India 1970–2006
ARDL bounds test
Cointegration
VECM
Y→Electricity supply (short-run)
Sustainability 2020, 12, 3121 6 of 22
Table 1. Cont.
Author Country Period Methodology Finding
Acaravci (2010) [38] Turkey 1968–2005
Cointegration
Granger causality
VECM
EC → Y
Lorde et al., (2010) [39] Barbados 1980–2006
Granger causality
VECM
VAR (vector auto regressive)
EC → Y (short-run)
EC ←→ Y (long-run)
Yoo and Kwak (2010) [46]
Argentina
Brazil
Chile
Columbia
Ecuador
Peru
Venezuela
1975–2006
Johansen cointegration
Hsiao’s Granger causality
EC → Y (Argentina, Brazil, Chile,
Columbia, Ecuador)
Y ⇔ EC (Venezuela)
Y == EC (Peru)
Chandran et al. (2010) [40] Malaysia 1971–2003
ARDL bounds test
Granger causality
EC → Y (short-run)
EC → Y (long-run)
Acaravci and Ozturk (2010) [47] 15 transition countries 1990–2006
Pedroni cointegration
Granger causality
Y == EC
Apergis and Payne (2011a) [48] 88 countries 1990–2006 Panel cointegration test
Y ⇔ EC (high income and
upper-middle-income country panels)
EC → Y (short-run, lower-middle-income
country panel)
Y ⇔ EC (long-run, lower-middle-income
country panel)
EC → Y (short-run, low income-country
panel)
Ozturk and Acravci (2011) [49]
11 Middle East and North
Africa (MENA) countries
1971–2006
ARDL bounds test
Granger causality
Y → EC (short-run, Israel, Oman)
EC → Y (long-run, Egypt and Saudi Arabia)
Y == EC (Iran, Jordan, Morocco, Syria)
Bekhet and Othman (2011) [41] Malaysia 1970–2009
Cointegration
Granger causality
EC → Y (long-run)
Shahbaz et al. (2011) [42] Portugal 1971–2009
ARDL bounds test
Granger causality
VECM
EC → Y (short-run)
Y ⇔ EC (long-run)
Apergis and Payne (2011b) [52] 16 emerging market economies 1990–2007
Panel cointegration
Panel Granger causality
Y → REC (short-run)
Y ⇔ REC (long-run)
Y ⇔ NREC
Apergis and Payne (2012) [53] 6 Central American countries 1990–2007 Panel cointegration test
REC → Y (short-run)
Y ⇔ REC (long-run)
Y ⇔ NREC
Al-mulali et al. (2014) [54] 18 Latin American countries 1980–2010 Panel cointegration test
Y ⇔ REC (long-run)
NREC → Y (short-run)
Y ⇔ NREC (long-run)
Sustainability 2020, 12, 3121 7 of 22
Table 1. Cont.
Author Country Period Methodology Finding
Halkos and Tzeremes (2014) [55] 36 countries 1990–2011 Nonparametric analysis Based on Growth hypothesis
Ibrahiem (2015) [51] Egypt 1980–2011 ARDL bounds testing approach Y ⇔ REC (long-run)
Kumari and Sharma (2016) [43] India 1974–2014
Cointegration
Granger Causality
Y → EC
Atems and Hotaling (2018) [56] 174 countries 1980–2012
The system generalized method of moments
(GMM) approach
Positive relationship between Y, renewable
electricity generation and non-renewable
electricity generation
Al-Mulali et al. (2018) [50]
Gulf Cooperation Council
(GCC) member
1980–2014
Panel cointegration test
Panel Granger causality test
Y → EC (short-run)
Y ⇔ EC (long-run)
Aydin (2019) [57] 26 OECD countries 1980–2015
Cross-sectional dependence test
Panel unit root test
Dumitrescu-Hurlin (DH) panel causality test
Panel frequency domain causality
Y == EC (DH)
Y ⇔ REC, NREC (panel frequency)
Note: Y, EC, REC, NREC mean economic growth, electricity consumption, renewable electricity consumption, and non-renewable electricity consumption, respectively. EC → Y, growth
hypothesis; Y → EC, conservation hypothesis; Y ⇔ EC, feedback hypothesis; Y == EC, neutrality hypothesis.
Sustainability 2020, 12, 3121 8 of 22
2.3. Revisited Nexus Study
This study tries to investigate the causal relationship between renewable electricity and economic
growth. Given what previous studies have referred to as further studies [9,47,56], we “revisited”
the nexus study. As mentioned in the previous “Literature Review” section, many studies refer
to the existence of a consensus in the nexus field. Nowadays, to try to clarify the related points,
several revisited studies are being conducted. Andrew and Bothwell [58] mentioned the limitation of
using a bivariate or trivariate model within South Africa. They try to revisit the electricity–growth
nexus by using a multivariate model considering the economic aspects of a country like export,
employment, and consumer price index. Zortuk and Karacan [59] revisited the energy–growth nexus
by selecting countries included in panel data. The ex-Soviet countries located in Central and Eastern
Europe and the Caucasus are now in transition into free market economics. Therefore, analysis panel
data containing the related countries can clarify the direction of the causal relationship. We also try to
revisit the electricity–growth nexus in the same country as the previous study, but we will be using a
different methodology and have a different purpose to them.
In this study, we have three distinctions from the previous study. First, the study’s purpose is
to confirm the renewable electricity–growth nexus by considering the characteristics of renewable
energy. This feature is discussed in detail in the next section. Second, to reflect the characteristics
that vary depending on the source, we built a model at the renewable source level. The differences
in the characteristics of energy sources can affect the investigated findings [56]. There are cases in
which previous studies have considered disaggregated level like sources [26,55]; however, only specific
sources (steam, hydro, diesel, and gas turbine) were used for data availability and analytical
purposes [26]. Halkos and Tzeremes [55] conducted their analysis using wind power, biomass,
solar power, and geothermal data. However, the availability of the data used differed from country to
country, and a non-parametric technique was applied. The last difference is that generation data is
used for the application of the first and second implications mentioned above. The various differences
between generation and consumption could affect interpretation in this study, which aimed at reflecting
renewable characteristics rather than electricity [40,56]. This study is meaningful in that it analyzes
through a revisiting flow in which the implication of the study is made by both the setting of the
country and the variables.
3. Data and Methodology
3.1. Renewable–Growth Hypothesis
In this study, we analyze the renewable electricity and economic growth nexus considering
renewable manufacturing industries. As mentioned above, the cost structure of renewable electricity
is quite different from the conventional thermal power. In the renewable generation sector, much of
the cost goes to equipment or facilities, which means a capital expenditure is much higher than an
operational expense. On the other hand, the thermal power plants that use fossil fuels have substantial
operational costs, although their capital cost is also considerable. Promoting renewable electricity could
thus have a positive effect on economic growth through the production of renewable manufacturing
industries such as wind turbine and solar panel manufacturing. In other words, renewable electricity
can contribute to economic growth through industrial outputs as well as an energy input. We want to
call this mechanism the ‘renewable–growth’ hypothesis in the context of the energy–growth nexus.
To test the renewable–growth hypothesis, we choose a country in which renewable electricity
might lead to the development of related industries. We confirm that the renewable–growth nexus
in a country includes companies that have developed in the renewable electricity sector. This study
attempts to examine the positive relationship between economic growth and renewable electricity
demand in such countries based on the market share of renewable manufacturing companies. At the
moment, we are analyzing solar PV and wind power, which are expected to grow quickly [3]. We might
expect to see a causal relationship by identifying the more disaggregated sectors.
Sustainability 2020, 12, 3121 9 of 22
Based on sales volume or revenue, we selected a global renewable energy company in solar [60] and
wind [61] power. According to the collected data, we identified the country the company belongs to for
analytical purposes. The analysis is conducted on solar and wind power, separately. Tables A1 and A2
show the companies and corresponding countries with the highest sales volume. For the solar PV,
we have included the component in the plant and the market share of module companies. Four countries
are analyzed for the solar PV: China (including Hong Kong), Canada, the USA, and Korea. A key
component of the wind manufacturing industry is the turbine, which accounts for 60% of the total
capital expenditure [62]. Thus, the standard for selecting the country included in the wind model
is a global wind-turbine company [Table A2]. Wind models include six countries: China, the USA,
Denmark, Germany, India, and Spain.
3.2. Data and Estimation Procedure
Consistent with the previous literature, we use real GDP as the dependent variable. Independent
variables are electricity generation, gross fixed capital, and labor force. Due to the features of renewables,
to transmission, to distribution losses, or even to theft, bias can exist in the investigated result [56].
This feature was also affected by country development level [40,63]. Therefore, we consider electricity
generation data. Considering the connectivity to the power grid and intermittent characteristics of
renewable energy, the detailed consumption data of each source is not available. To include this,
the previous study used aggregate level consumption data. However, this study’s purpose is to
identify the effects of the nexus depending on the presence of manufacturing companies reflecting the
manufacturing feature of each source. Reflecting on the characteristics of the energy source means it
must be analyzed at a disaggregated level. Therefore, generation data is used for analysis, and solar
PV and wind are configured separately.
We adopt the annual time series data of each country from 1980 to 2017. The data are obtained
from the World Bank Development Indicators, World Energy Balances, and US Energy Information
Administration and defined as follows: Real GDP (Y) in billion 2010 USD using purchasing power
parity, fossil fuel electricity generation (NRE) defined in kilowatt hours, electricity generation from
solar PV (RES) defined in kilowatt hours, electricity generation from wind (REW) defined in kilowatt
hours, real gross fixed capital formation in current US dollars (K), and total labor force (L) in millions.
We convert the unit of real gross fixed capital formation into constant 2010 US dollars using the GDP
deflator from WDI. Additionally, all variables are converted to per capita and the natural logarithms
are transformed.
3.3. Estimation Strategy
We assume that, due to the characteristics of renewables, renewable electricity generation
and economic growth have a relationship. To confirm the causal relationship between them,
we must check the stationarity of time series data. If data are not stationary, the estimated
model might be a spurious regression and results may lack robustness. Common methods
applied to unit root tests are the augmented Dickey–Fuller (ADF) [64], Phillips–Perron (PP) [65],
and Kwiatkowski–Phillips–Schmidt–Shin (KPSS) [66]. If the null hypothesis of the ADF and PP tests
are rejected, we conclude that time series data has a unit root. The KPSS test is applied to investigate
the stationarity, and, if the null hypothesis is rejected, the data may be nonstationary time series data.
The KPSS test is more suitable for testing small samples due to the lower lag truncation parameter, and it
might complement the limitation of the ADF and PP tests [67]. Therefore, as a tool for cross-checking
and for the robustness of results, we conduct ADF, PP, and KPSS tests together.
Before the Granger causality analysis, it is necessary to test for cointegration. There are several
methodologies related to cointegration, and the Johansen test [68] is the most common method as it is
more generally applicable than the Engle–Granger test [69]. Since this test is common and well known,
Sustainability 2020, 12, 3121 10 of 22
we provide just a brief overview of this method. Johansen [68] modeled time series as a reduced rank
regression, and we can trace test and maximum eigenvalue. The model is given as the following:
∆Zt = ω +
q−1∑
i=1
∆Zt−i + Π∆Zt−1 + εt (1)
The Zt is a vector of p variables consisting of an (n × 1) column, and ω is an (n × 1) vector of
constant terms. Γ means coefficient matrices, and ∆ is a difference operator. The εt follows distribution
as N(0, Σ). The Π means the coefficient is known as the impact matrix and contains information about
the long-run relationships. By employing this method, it is possible to determine the number of the
cointegrating vectors of the model.
We can also employ the cointegration test with the autoregressive distributed lag (ARDL) bounds
test approach [70]. By using ARDL bounds testing, we can expect to take advantage of the following.
First, it is possible to apply such a method irrespective of whether the regressors are I (0) or I (1).
Furthermore, compared to the Johansen cointegration techniques, estimating with a smaller sample
size is possible [71]. According to Narayan [72], the ARDL bounds test gives a reasonable critical value
if the number of samples is between 30 and 80. Furthermore, the ARDL model can be derived in the
form of an error correction model. Thi means it is possible to confirm that the long-run and short-run
causality is the same as in VECM. In this study, we can establish the ARDL model as the following:
∆GDPt = α0 +
q∑
i=1
α1∆GDPt−i +
q∑
i=1
α2∆NREt−i +
q∑
i=1
α3∆REt−i +
q∑
i=1
α4∆Kt−i+
λ1GDPt−1 + λ2NREt−1 + λ3REt−1 + λ4Kt−1 + µt
(2)
∆REt = β0 +
q∑
i=1
β1∆REt−i +
q∑
i=1
β2∆NREt−i +
q∑
i=1
β3∆GDPt−i +
q∑
i=1
β4∆Kt−i+
ζ1REt−1 + ζ2NREt−1 + ζ3GDPt−1 + ζ4Kt−1 + µ2t
(3)
where GDP, NRE, RE, and K denote the logarithm form of real GDP, fossil fuel electricity
generation, solar PV or wind electricity generation, and gross domestic capital formation, respectively.
The parameters α, β are the short-run dynamic coefficient and λ, ζ are the corresponding long-run
multipliers of each ARDL model. The µ represents the white noise error term.
To determine the existence of cointegration, we test the lagged value jointly by using the F
test. However, as the F statistics from [70] are for several samples, we usually use the critical
value from Narayan in the small sample size analysis [72]. The null hypotheses of each model are
Ho : λ1 = λ2 = λ3 = 0 and Ho : ζ1 = ζ2 = ζ3 = 0. If the computed F statistics exceed the upper
bound value, the null hypothesis is rejected. However, if the F statistics fall below the lower bound,
we can conclude that we cannot reject the no cointegration hypothesis. However, if the value exists
between the upper bound and lower bound, the results are inconclusive. Through the cointegration
test, if we find the evidence for a long-run relationship between variables, we can conduct the test to
check the existence of Granger causality.
According to the results from the cointegration test, there are two cases we must consider. If our
variables are not cointegrated, we perform the test as a vector autoregressive (VAR) model in first
differenced variable form. If we can confirm the existence of a long-run relationship, we can conduct
the model with the error correction term. Thus, the model has a cointegration, and the Granger
causality relationships are written as the vector error correction models (VECM) given below:
∆GDPt = α0 +
q∑
i=1
α1i∆GDPt−i +
q∑
i=1
α2i∆NREt−i +
q∑
i=1
α3i∆REt−i +
q∑
i=1
α4i∆Kt−i+
λECTt−1 + µ1t
(4)
Sustainability 2020, 12, 3121 11 of 22
∆REt = β0 +
q∑
i=1
β1i∆REt−i +
q∑
i=1
β2i∆NREt−i +
q∑
i=1
β3i∆GDPt−i +
q∑
i=1
β4i∆Kt−i+
ζECTt−1 + µ2t
(5)
The null hypothesis of the Granger causality from renewable electricity generation to GDP is
Ho : α1 = α2 = . . .αq = 0. To test the short-run causality of renewable electricity to GDP, we impose
restrictions on all the lagged renewable electricity generation data using the F test. This is equivalent
to the test of lagged first differences of the Granger causality from renewable electricity generation
to GDP.
Whether or not they are integrated, the economic variables can be integrated into different orders.
In that case, the Wald test will not have an asymptotic chi-square distribution, and VECM cannot be
applied for the Granger causality test. Toda and Yamamoto [73] suggest the procedure for solving this
problem. According to the standard stationary test, we can determine the order of the variables. Let m
be the maximum order of integration, and l be the appropriate maximum lag length in VAR. Then take
the preferred VAR model and add the m additional lags of each of the variables. We can test the
Granger causality using that model. However, we must test the hypothesis with only the coefficients
of the first l lagged values. Rejection of the null hypothesis implies the existence of Granger causality.
4. Empirical Results
4.1. Unit Root Test Results
Tables A3 and A4 present the result of a stationarity test for the solar PV case and the wind power
case from ADF, PP, and KPSS test. We use Stata 14.0 for the analysis. The definition of GDP is real GDP,
FOG is fossil-fuel electricity generation, RE is renewable electricity generation, and FXC is fixed capital
formation. According to the model, RE is divided into solar PV (RES) and wind power (REW). In the
process of examining stationarity, the maximum lag length was set to 4 considering the characteristics
of the economic variables.
As mentioned above, ADF, PP, and KPSS tests were conducted simultaneously for robustness.
If the differenced form does not reject the null hypothesis in more than two tests, the second differenced
variable is tested. The first differenced variable is then analyzed based on the statistics considering
the trend. Additionally, if we get the result that the first differenced is stationary in at least two tests,
we do not write the statistics of the second differenced form in the table. Two or more tests indicate the
same order of integration, and we set that value as the integration order of that variable. The results
show that most variables in both models are in the form of I (1). Only the GDP variable of Spain in
wind power has I (2). Since the order of integration of the variables is mixed, the analysis proceeds
with Toda and Yamamoto’s procedure.
Therefore, we constructed the VAR model with the variables to find the optimal lag length of the
model setting the maximum lag as 4. Lag-order selection is based on Bayesian information criterion
(BIC). In each model, the VAR model was constructed according to the optimal time difference obtained
in each country, autoregression was checked, and the time difference was adjusted. The order of
integration of variables and the optimal lag of the VAR model are summarized in Tables 2 and 3.
Table 2 is for the solar PV, and Table 3 is for wind power. To investigate the Granger causality of each
country and model, we reconstructed the VAR model with the sum of these two values as the lag
length of the new model.
Sustainability 2020, 12, 3121 12 of 22
Table 2. Integration order and optimal lag results for solar photovoltaics (PV).
Country Variables Order of Integration (m) Optimal Lag (l)
Canada
GDP I(1)
1
FOG I(1)
RES I(1)
FXC I(1)
China
GDP I(1)
3
FOG I(1)
RES I(1)
FXC I(1)
USA
GDP I(1)
4
FOG I(1)
RES I(1)
FXC I(1)
Korea
GDP I(1)
3
FOG I(1)
RES I(1)
FXC I(1)
Table 3. Integration order and optimal lag results for wind power.
Country Variables Order of Integration (m) Optimal Lag (p)
China
GDP I(1)
3
FOG I(1)
REW I(1)
FXC I(1)
USA
GDP I(1)
5
FOG I(1)
REW I(1)
FXC I(1)
Denmark
GDP I(1)
2
FOG I(1)
REW I(1)
FXC I(1)
Germany
GDP I(1)
2
FOG I(1)
REW I(1)
FXC I(1)
India
GDP I(1)
1
FOG I(1)
REW I(1)
FXC I(1)
Spain
GDP I(2)
2
FOG I(1)
REW I(1)
FXC I(1)
4.2. Results of the Granger Causality
Tables 4 and 5 summarize the results of the solar PV and wind power, respectively. The country
shows that the growth hypothesis presented in the previous nexus study are Canada for the solar PV,
and Germany, India, and Spain for wind power. China and the USA were included in both models,
satisfying the conservative and feedback hypothesis, respectively. Denmark was the only country with
a neutral hypothesis. The renewable–growth relationship is investigated in Canada, USA, and Korea
Sustainability 2020, 12, 3121 13 of 22
for the solar PV and the USA, Germany, India, and Spain for wind power. China in both models and
Denmark in the wind power show different results based on our hypothesis.
Table 4. The results of the Granger causality in solar PV.
Country Granger Causality Test Statistics
Hypothesis
Conventional Renewable–Growth
Canada
RES → GDP 3.27 *
Growth Yes
GDP → RES 0.20
China
RES → GDP 3.61
Conservative No
GDP → RES 8.70 **
USA
RES → GDP 20.25 ***
Feedback Yes
GDP → RES 25.74 ***
Korea
RES → GDP 18.14 ***
Feedback Yes
GDP → RES 16.48 ***
Note: *, **, and *** denote the rejection of the null hypothesis of no relationship at the 5% and 1% significance
level, respectively.
Table 5. The results of the Granger causality wind power.
Country Granger Causality Test Statistics
Hypothesis
Conventional Renewable–Growth
China
REW → GDP 0.79
Conservative No
GDP → REW 23.84 ***
USA
REW → GDP 18.48 ***
Feedback Yes
GDP → REW 17.12 ***
Denmark
REW → GDP 0.09
Neutral No
GDP → REW 0.28
Germany
REW → GDP 14.00 ***
Growth Yes
GDP → REW 4.25
India
REW → GDP 2.75 *
Growth Yes
GDP → REW 0.04
Spain
REW → GDP 6.26 **
Growth Yes
GDP → REW 0.71
Note: *, **, and *** denote the rejection of the null hypothesis of no relationship at the 5% and 1% significance
level, respectively.
In this study, the standard Granger causality test was also conducted to confirm the reliability of
the results. The difference of the standard Granger causality is after the stationarity test, which means
identifying the cointegration of the model. The Johansen method was used to confirm the cointegration,
and the bounds test was also performed considering the small sample size. If cointegration exists,
the VECM model is constructed to confirm the causality of the long run and short run. If the results
indicate no cointegration in both tests, a VAR model with a differenced variable must be built, and the
short run causality is checked using that model. The lag length of each model is the same as that of
Toda and Yamamoto. Tables A5 and A6 summarize the standard Granger causality results. Only China,
for the solar PV, shows a conflicting result in the cointegration test. The variables in this model are
all I (1), so we follow the bounds test result due to the small sample size. Furthermore, because
the lag lengths of Canada for solar PV and India for wind power are both 1, the short-run term
disappears when constructing the VECM model. Therefore, both cases were analyzed only in the
long-run relationship. However, due to the integration order of the variables, we used and analyzed
the results using the Toda and Yamamoto procedure.
Sustainability 2020, 12, 3121 14 of 22
5. Discussion
The study aims to investigate the renewable–growth hypothesis of the countries with a renewable
manufacturing industry. According to the results, the long-run relationship between renewable
electricity generation and economic growth is present in all countries. However, in some countries,
the results differ from our hypothesis. The results can be attributed to diverse factors like economic or
socio-cultural points existing in different countries [45,47,52,74].
In China, the opposite of the renewable–growth hypothesis was found in both solar PV and wind
power. China is the largest producer of both solar and wind components. The total installed solar PV
was 175 GW in 2018, while the total for wind power was 185 GW in 2018 [75]. However, China’s exports
of manufactured goods to other countries are much higher than its domestic use [60,61]. Therefore,
the causality test with renewable electricity generation in China may not reflect the positive effect of
the renewable electricity industry. China’s renewable electricity is also a reason for the ambiguous
factor of the renewable–growth relationship. The renewable electricity industry induces economic
growth in the indirect sectors, such as the development of the manufacturing sector and employment
for facilities and the direct impact of the generation itself. Electricity curtailment due to grid issues is a
chronic problem in China [76,77]. The resulting loss of generation has a negative effect on economic
growth [60].
Denmark also has a similarity with China. In Denmark, renewable energy accounts for more than
50% of total electricity production [78]. However, the increase in installed wind energy capacity is
the smallest among countries with wind power [75]. In the absence of increased domestic facilities,
it is difficult to identify the growth hypothesis with given variables. Furthermore, Siemens Gamers
has manufacturing facilities worldwide [79]. Vestas’ manufacturing facilities are also located in
other countries [80]. Thus, even if they satisfy the renewable–growth hypothesis, they may not
be revealed under the given variables. In the rest of the target countries, we can examine the
renewable–growth hypothesis.
The renewable electricity industry of countries showing a growth hypothesis has been more
established in recent years. Germany, which shows the renewable–growth relationship between
wind power and economic growth, had the largest wind power capacity in 2017. Germany leads the
European Union (EU) in terms of power capacity with 56.1 GW, followed by Spain with 23.2 GW,
which is also included in our analysis model. The manufacturing facility of Gamesa, which merged
with Danish renewable company Siemens in 2016, still exists in Spain, and this could have a positive
effect on the country. According to GWEC [81], one of the leading turbine suppliers for the EU is
Siemens Gamesa. While installation numbers have slowed, the total capacity of India is 32.8 GW,
taking the fourth position in the world’s largest wind markets. Because the share of wind-generated
electricity in India’s total electricity consumption is still 4.35%, there is still growth potential in India’s
renewable electricity market [81]. The USA has the highest installed capacity after China among our
target countries, both for solar PV and wind power. In 2018, the total installed capacity was 51 GW
for solar PV and 94 GW for wind power [75]. In Korea, the installed solar PV capacity is 8 GW, but
the number of related companies and employment is increasing due to recent aggressive renewable
electricity policies. Canada has also installed a solar PV capacity of 3 GW, and while it is not growing
much, it is considered to have high potential as the capacity continues to rise.
Under our hypothesis, the driving force of economic growth is the electricity industry itself,
and other industries have induced effects from the increase of renewable generation. We used solar
PV generation data and wind power generation data as variables to cover these points. However,
these variables do not reflect the growth effect of the renewable manufacturing industry effectively in
some countries. Development of relevant industries can lead to the growth of domestic products usage.
So we set renewable electricity data as a variable in our model. This is a limitation of the study in that
it does not consider the characteristics of trade. In the same line, the diverse factors that are inherent
in countries need to be considered. The various political, economic, and socio-cultural factors could
cause a bias in the nexus study. Therefore, further research to secure the renewable–growth hypothesis
Sustainability 2020, 12, 3121 15 of 22
is necessary. Developing variables to consider the export-focused characteristic and renewable energy
policy of each country could also be the next step of the renewable–growth hypothesis. Furthermore,
the existence of renewable growth may not be revealed in the results due to bias arising from the
small sample size. It is necessary to overcome this limitation in the future with continuous research
on specific energy sources. A country’s characteristics may vary and be embodied in the results,
so it also becomes the motivation for additional research. Lastly, our study is the first to present the
renewable–growth hypothesis. However, it must be noted that while we use an analysis strategy by
using Granger causality, it is unknown whether the relationship is positive or negative [82]. A more
diversified approach could make it clear whether the hypothesis is correct or not.
6. Conclusions
The increasing concerns about the climate crisis cause trends such as energy transition from fossil
fuels to renewable energy. According to the change in the power mix, the renewable electricity industry
is affected both directly and indirectly. The difference between conventional electricity and renewable
electricity sources lies in the weight of manufacturing-related parts. This is the case with solar PV and
wind power, both comprising the world’s highest growth in renewable electricity installation. Due
to such characteristics, the increase in renewable energy generation has the additional effect of not
only leading to advantages in the generation industry but also the development of related industries
and increase in employment. As this will have a positive effect on economic growth, we have thus
formulated the renewable–growth hypothesis, which is one step further from the growth hypothesis of
the existing nexus studies.
In this study, we tried to investigate the relationship between renewable electricity and economic
growth. We constructed and analyzed more disaggregated models like solar PV and wind power.
Additionally, we established the target group based on global companies in manufacturing. It was
expected that the renewable–growth relationship would be represented more accurately in those
countries. The variables within the model are real GDP, fossil fuel electricity generation, renewable
electricity generation, and fixed capital formation. Renewable electricity generation means the solar
PV data for solar PV and wind power data in wind power. Because our purpose is to confirm the
renewable–growth hypothesis, we used electricity generation as a variable, considering the difference
between generation and consumption data.
The analysis showed that the renewable–growth hypothesis was satisfied except only China for
solar PV and China and Denmark for wind power. These results are due to the characteristics of the
country’s power and power-related manufacturing industries. We can thus conclude that national
and industry characteristics can influence the results of the analysis. Furthermore, we can provide the
direction for further studies. First, renewable electricity time series data has a relatively small sample
size. Therefore, if data is acquired over time, it is necessary to analyze a larger sample size. Doing this
would determine the renewable–growth relationship more accurately. Second, considering additional
variables to reflect the characteristics of industry or country would be necessary. However, as we can
see from the results of other countries, the parameters were appropriate for verifying the hypothesis.
Thus, while considering the general case, further research is necessary to reflect the characteristics of
each country. The study’s main contributions are its presentation of a new perspective in the form of
the renewable–growth hypothesis and the establishment and analysis of a target group that reflects the
characteristics of the renewable electricity industry.
Author Contributions: Conceptualization, J.K. and M.Y.; Methodology, J.K. and M.Y.; Validation, J.K.; Formal
Analysis, M.Y.; Investigation, J.K. and M.Y.; Data Curation, M.Y.; Writing—Original Draft Preparation, M.Y.;
Writing—Review & Editing, J.K.; Supervision, J.K.; Funding Acquisition, J.K. All authors have read and agreed to
the published version of the manuscript.
Funding: This work was supported by the Human Resources Development program (No. 20194010201860) of the
Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Korea government
Ministry of Trade, Industry and Energy.
Sustainability 2020, 12, 3121 16 of 22
Conflicts of Interest: The authors declare no conflict of interest.
Appendix A
Table A1. The highest production company in solar PV (module).
Rank Company Country Production [MW]
1 Jinko Solar China 9000
2 JA Solar China 8500
3 Canadian Solar Canada 8310
4 Hanwha Q CELLS Korea 8000
5 Trina Solar Ltd. China 8000
6 Risen China 6600
7 GCL System China 5400
8 Talesun China 4500
9 Suntech/Shunfeng China 3300
10 Znshine Solar China 3200
11 Seraphim China 3000
12 Chint/Astronergy China 2500
13 First Solar USA 2200
14 Eging China 2000
15 BYD China 1700
Table A2. The market share of global wind power companies.
Rank Company Country Market Share (%, 2014)
1 Vestas Denmark 16%
2 Siemens Gamesa Denmark 15%
3 Goldwind China 12%
4 GE Wind USA 10%
5 Enercon Germany 7%
6 Nordex Germany 6%
7 Envision China 6%
8 Senvion
Germany
(India)
3%
9 Suzlon India 3%
10 Guodian UP China 3%
11 Ming Yang China 2%
Appendix B
Table A3. Unit root test results for solar PV.
Country Variables
Test Statistics
ADF PP KPSS
Canada
GDP
Level −2.144 −6.105 0.225 ***
First difference −4.556 *** −26.242 *** 0.116
FOG
Level −1.233 −3.816 0.416 ***
First difference −6.374 *** −40.340 *** 0.0413
RES
Level −2.427 −10.865 0.113
First difference −3.013 ** −15.118 ** 0.0943
FXC
Level −2.565 −7.935 0.25 ***
First difference −5.095 *** −31.560 *** 0.0992
Sustainability 2020, 12, 3121 17 of 22
Table A3. Cont.
Country Variables
Test Statistics
ADF PP KPSS
China
GDP
Level −2.371 −10.118 0.183 **
First difference −3.094 ** −15.071 ** 0.0984
FOG
Level −1.804 −8.698 0.2 **
First difference −3.546 ** −17.539 ** 0.187 **
RES
Level −0.372 −0.427 0.7 ***
First difference
1 −4.122 ** −24.522 *** 0.116
FXC
Level −1.750 −6.024 0.413 ***
First difference −3.875 *** −17.560 ** 0.128 ’
USA
GDP
Level −0.485 −0.967 0.301 ***
First difference −5.216 *** −30.371 *** 0.244 ***2
FOG
Level 1.072 2.428 0.343 ***
First difference −6.544 *** −41.387 *** 0.158 *2
RES
Level −2.923 −7.352 0.266 ***
First difference −5.637 *** −35.112 *** 0.0927
FXC
Level −2.716 −11.397 0.142 ’
First difference −5.800 *** −34.972 *** 0.0615
Korea
GDP
Level −3.063 −8.906 0.126 ’
First difference −2.654 * −13.941 ** 0.293 ***2
FOG
Level −1.362 −5.247 0.113
First difference −2.902 * −52.064 *** 0.0957
RES
Level −2.557 −11.407 0.111
First difference −2.963 * −46.013 *** 0.112
FXC
Level −2.222 −8.004 0.284 ***
First difference −5.671 *** −34.438 *** 0.06
Note: *, **, and *** indicate the level of significance at 10%, 5%, and 1% for ADF, PP. ’, *, **, and *** indicate the level
of significance at 10%, 5%, 2.5%, and 1% for KPSS. 1 first difference form using trend when testing ADF and PP. 2 the
result of second difference form is stationary.
Table A4. Unit root test results for wind power.
Country Variables
Test Statistics
ADF PP KPSS
China
GDP
Level −2.371 −10.118 0.183 **
First difference −3.094 ** −15.071 ** 0.0865
FOG
Level −1.804 −8.698 0.164 *
First difference −3.546 ** −17.539 ** 0.158 *2
REW
Level −2.863 −10.116 0.0846
First difference −4.715 *** −28.629 *** 0.0773
FXC
Level −1.750 −6.024 0.413 ***
First difference −3.875 *** −17.560 ** 0.128 ’2
USA
GDP
Level −0.485 −0.967 0.301 ***
First difference −5.216 *** −30.371 *** 0.244 ***2
FOG
Level 1.072 2.428 0.343 ***
First difference −6.544 *** −41.387 *** 0.158 *2
REW
Level −2.515 −10.241 0.305 ***
First difference
1 −7.100 *** −43.660 *** 0.033
FXC
Level −2.716 −11.397 0.142 ’
First difference −5.800 *** −34.972 *** 0.0615
Sustainability 2020, 12, 3121 18 of 22
Table A4. Cont.
Country Variables
Test Statistics
ADF PP KPSS
Denmark
GDP
Level −1.684 −3.320 0.414 ***
First difference −4.619 *** −27.620 *** 0.0744
FOG
Level −0.391 −1.660 0.426 ***
First difference −8.083 *** −48.551 *** 0.0307
REW
Level −2.446 −1.864 0.448 ***
First difference −4.422 *** −24.953 *** 0.0876
FXC
Level −2.696 −13.698 0.175 *
First difference −4.836 *** −28.317 *** 0.102
Germany
GDP
Level −1.291 −2.180 0.444 ***
First difference −4.301 *** −25.521 *** 0.114
FOG
Level −1.094 −6.726 0.225 ***
First difference −6.980 *** −44.514 *** 0.0939
REW
Level −0.777 −0.912 0.459 ***
First difference −2.665 * −12.125 * 0.331 ***2
FXC
Level −3.843 ** −11.353 0.128 ’
First difference −2.662 * −30.106 *** 0.0918
India
GDP
Level −1.422 −2.319 0.454 ***
First difference −3.879 *** −22.200 *** 0.0682
FOG
Level −2.775 −3.588 0.411 ***
First difference −4.405 *** −26.979 *** 0.167 *2
REW
Level −1.771 −5.858 0.318 ***
First difference
1 −5.555 *** −34.473 *** 0.162 *
2
FXC
Level −1.843 −6.219 0.342 ***
First difference −6.157 *** −38.722 *** 0.0848
Spain
GDP
Level −2.854 −5.346 0.239 ***
First difference −2.704 1 −15.080 ** 0.225 ***
Second
difference
−8.087 *** −45.537 *** 0.0283
FOG
Level −1.146 −7.526 0.173 *
First difference −5.891 *** −44.595 *** 0.0633
REW
Level −0.946 −3.469 0.394 ***
First difference
1 −4.827 *** −29.612 *** 0.328 ***
2
FXC
Level −2.425 −7.233 0.166 *
First difference −3.404 ** −22.619 *** 0.117
Note: *, **, and *** indicate the level of significance at 10%, 5%, and 1% for ADF, PP. ’, *, **, and *** indicate the level
of significance at 10%, 5%, 2.5%, and 1% for KPSS. 1 first difference form using trend when testing ADF and PP. 2:
the result of second difference form is stationary.
Sustainability 2020, 12, 3121 19 of 22
Appendix C
Table A5. The results from the standard Granger causality test in solar PV.
Country
Cointegration Granger Causality
Johansen Bounds Test Short-Run Long-Run
Canada Rank 1 6.388 *** – 1 RES → Y ***
China Rank 1 1.935 RES → Y ** – 2
USA Rank 1 4.867 * No causality RES → Y *
Korea Rank 1 10.649 *** Y → RES ***
RES → Y ***
Y → RES ***
Note: ’, *, **, and *** indicate the level of significance at 10%, 5%, 2.5%, and 1% for bound-testing. *, **, and ***
indicate the level of significance at 10%, 5%, and 1% for Granger causality. 1 the lag of this model is 1, so we can
investigate the long-run relationship result only. 2 according to the bounds test result, we investigate the short-run
relationship using only VAR.
Table A6. The results from the standard Granger causality test in wind power.
Country
Cointegration Granger Causality
Johansen Bounds Test Short-Run Long-Run
China Rank 1 5.951 *** No causality No causality
USA Rank 1 4.737 *** Y → REW ** No causality
Denmark Rank 2 4.041 ’ No causality
REW → Y ***
Germany Rank 1 8.590 ***
REW → Y *
Y → REW *
REW → Y ***
India Rank 2 5.888 *** – 1 REW → Y **
Spain Rank 3 16.836 ***3 REW → Y ** Y → REW *
Note: ’, *, **, and *** indicate the level of significance at 10%, 5%, 2.5%, and 1% for Bound-testing. *, **, and ***
indicate the level of significance at 10%, 5%, and 1% for Granger causality. 1 the lag of this model is 1, so we can
investigate the long-run relationship result only. 2 this model does not have cointegration, so we can investigate the
short-run relationship only. 3 while this model includes the I (2) variable, we applied the bounds test for reference.
1. Damian, C. Climate Crisis: 11,000 Scientists Warn of ‘Untold Suffering’. The Guardian. 5 November 2019.
Available online: https://www.theguardian.com/environment/2019/nov/05/climate-crisis-11000-scientists-
warn-of-untold-suffering (accessed on 2 January 2020).
2. Pierre, F.; Matthew, W.J.; Michael, O.; Robbie, M.A.; Judith, H.; Glen, P.P.; Wouter, P.; Julia, P.; Stephen, S.;
Corinne, L.Q.; et al. Global Carbon Budget 2019. Earth Syst. Sci. Data 2019, 11, 1783–1838. [CrossRef]
3. REN21 (Renewable Energy Policy Network for the 21st Century). Renewables 2018: Global Status Report;
REN21 Publications: Paris, France, 2019.
4. Kaygusuz, K. Energy for sustainable development: Key issues and challenges. Energy Sources Part B Econ.
Plan. Policy 2007, 11, 73–83. [CrossRef]
5. Kaygusuz, K.; Yuksek, O.; Sari, A. Renewable energy sources in the European union: Markets and capacity.
Energy Sources Part B Econ. Plan. Policy 2007, 2, 19–29. [CrossRef]
6. Park, S.; Kim, J. The effect of interest in renewable energy on US household electricity consumption: An
analysis using Google Trends data. Renew. Energy 2018, 127, 1004–1010. [CrossRef]
7. Moselle, B. Why Support Renewables? In Proceedings of the EPRG Spring Research Seminar, University of
Cambridge, Cambridge, UK, 13 May 2011.
8. Kraft, J.; Kraft, A. Interfuel substitution and energy consumption in the industrial sector. Appl. Energy 1980,
38, 348–349. [CrossRef]
9. Ozturk, I. A literature survey on energy–growth nexus. Energy Policy 2010, 38, 340–349. [CrossRef]
https://www.theguardian.com/environment/2019/nov/05/climate-crisis-11000-scientists-warn-of-untold-suffering
https://www.theguardian.com/environment/2019/nov/05/climate-crisis-11000-scientists-warn-of-untold-suffering
http://dx.doi.org/10.5194/essd-11-1783-2019
http://dx.doi.org/10.1080/15567240500402560
http://dx.doi.org/10.1080/15567240500400887
http://dx.doi.org/10.1016/j.renene.2018.05.044
http://dx.doi.org/10.1016/0306-2619(80)90018-5
http://dx.doi.org/10.1016/j.enpol.2009.09.024
Sustainability 2020, 12, 3121 20 of 22
10. Harold, A.; Pablo, R.; Michael, T. Renewable Power Generation Costs in 2018; International Renewable Energy
Agency: Abu Dhabi, UAE, 2019.
11. EY (Ernst & Young). Available online: https://www.solarpowereurope.org/wp-content/uploads/2018/08/
Solar-PV-Jobs-Value-Added-in-Europe-November-2017 (accessed on 2 January 2020).
12. Payne, J.E. Survey of the international evidence on the causal relationship on the causal relationship between
energy consumption and growth. J. Econ. Stud. 2010, 37, 53–95. [CrossRef]
13. Yu, E.S.H.; Hwang, B.K. The relationship between energy and GNP: Further results. Energy Econ. 1984, 6,
186–190. [CrossRef]
14. Payne, J.E. On the dynamics of energy consumption and output in the US. Appl. Energy 2009, 86, 575–577.
[CrossRef]
15. Soytas, U.; Sari, R.; Ozdemir, O. Energy Consumption and GDP Relations in Turkey: A Co-integration and
Vector Error Correction Analysis. Glob. Bus. Technol. Assoc. 2001, 1, 838–844.
16. Soytas, U.; Sari, R. Energy consumption, economic growth, and carbon emissions: Challenges faced by an
EU candidate member. Ecol. Econ. 2009, 68, 1667–1675. [CrossRef]
17. Masih, A.; Masih, R. Energy consumption and real income temporal causality, results for a multi-country
study based on cointegration and error-correction techniques. Energy Econ. 1996, 18, 165–183. [CrossRef]
18. Soytas, U.; Sari, R. Energy consumption and GDP: Causality relationship in G-7 countries and emerging
markets. Energy Econ. 2003, 25, 33–37. [CrossRef]
19. Narayan, P.K.; Smyth, R. Energy consumption and real GDP in G7 countries: New evidence from panel
cointegration with structural breaks. Energy Econ. 2008, 30, 2331–2341. [CrossRef]
20. Apergis, N.; Tang, C.F. Is the energy-led growth hypothesis valid? New evidence from a sample of 85
countries. Energy Econ. 2013, 38, 24–31. [CrossRef]
21. Cheng, B.S. Energy Consumptions and economic growth in Brazil, Mexico and Venezuela: A time series
analysis. Appl. Energy 1997, 4, 671–674. [CrossRef]
22. Sari, R.; Soytas, U. The growth of income and energy consumption in six developing countries. Energy Policy
2007, 35, 889–898. [CrossRef]
23. Pao, H.T.; Fu, H.C. Renewable energy, non-renewable energy and economic growth in Brazil. Renew. Sustain.
Energy Rev. 2013, 25, 381–392. [CrossRef]
24. Apergis, N.; Danuletiu, D.C. Renewable energy and economic growth: Evidence from the sign of Panel
long-run causality. Int. J. Energy Econ. Policy 2014, 4, 578–587.
25. Kazar, G.; Kazar, A. The Renewable energy production-economic development nexus. Int. J. Energy
Econ. Policy 2014, 4, 312–319.
26. Ramcharran, H. Electricity consumption and economic growth in Jamaica. Energy Econ. 1990, 12, 65–70.
[CrossRef]
27. Ghosh, S. Electricity consumption and economic growth in India. Energy Policy 2002, 30, 125–129. [CrossRef]
28. Narayan, P.K.; Smyth, R. Electricity consumption, employment and real income in Australia evidence from
multivariate Granger causality tests. Energy Policy 2005, 33, 1109–1116. [CrossRef]
29. Yoo, S.H. Electricity consumption and economic growth: Evidence from Korea. Energy Policy 2005, 33,
1627–1632. [CrossRef]
30. Mozumder, P.; Marathe, A. Causality relationship between electricity consumption and GDP in Bangladesh.
Energy Policy 2007, 35, 395–402. [CrossRef]
31. Zachariadis, T.; Pashourtidou, N. An empirical analysis of electricity consumption in Cyprus. Energy Econ.
2007, 29, 183–198. [CrossRef]
32. Yuan, J.; Zhao, C.; Yu, S.; Hu, Z. Electricity consumption and economic growth in China: Cointegration and
co-feature analysis. Energy Econ. 2007, 29, 1179–1191. [CrossRef]
33. Halicioglu, F. Residential electricity demand dynamics in Turkey. Energy Econ. 2007, 29, 199–210. [CrossRef]
34. Tang, C.F. A re-examination of the relationship between electricity consumption and economic growth in
Malaysia. Energy Policy 2008, 36, 3077–3085. [CrossRef]
35. Abosedra, S.; Dah, A.; Ghosh, S. Electricity consumption and economic growth, the case of Lebanon.
Appl. Energy 2009, 86, 429–432. [CrossRef]
36. Akinlo, A.E. Electricity consumption and economic growth in Nigeria: Evidence from cointegration and
co-feature analysis. J. Policy Modeling 2009, 31, 681–693. [CrossRef]
https://www.solarpowereurope.org/wp-content/uploads/2018/08/Solar-PV-Jobs-Value-Added-in-Europe-November-2017
https://www.solarpowereurope.org/wp-content/uploads/2018/08/Solar-PV-Jobs-Value-Added-in-Europe-November-2017
http://dx.doi.org/10.1108/01443581011012261
http://dx.doi.org/10.1016/0140-9883(84)90015-X
http://dx.doi.org/10.1016/j.apenergy.2008.07.003
http://dx.doi.org/10.1016/j.ecolecon.2007.06.014
http://dx.doi.org/10.1016/0140-9883(96)00009-6
http://dx.doi.org/10.1016/S0140-9883(02)00009-9
http://dx.doi.org/10.1016/j.eneco.2007.10.006
http://dx.doi.org/10.1016/j.eneco.2013.02.007
http://dx.doi.org/10.1080/758530646
http://dx.doi.org/10.1016/j.enpol.2006.01.021
http://dx.doi.org/10.1016/j.rser.2013.05.004
http://dx.doi.org/10.1016/0140-9883(90)90009-5
http://dx.doi.org/10.1016/S0301-4215(01)00078-7
http://dx.doi.org/10.1016/j.enpol.2003.11.010
http://dx.doi.org/10.1016/j.enpol.2004.02.002
http://dx.doi.org/10.1016/j.enpol.2005.11.033
http://dx.doi.org/10.1016/j.eneco.2006.05.002
http://dx.doi.org/10.1016/j.eneco.2006.09.005
http://dx.doi.org/10.1016/j.eneco.2006.11.007
http://dx.doi.org/10.1016/j.enpol.2008.04.026
http://dx.doi.org/10.1016/j.apenergy.2008.06.011
http://dx.doi.org/10.1016/j.jpolmod.2009.03.004
Sustainability 2020, 12, 3121 21 of 22
37. Ghosh, S. Electricity supply, employment and real GDP in India: Evidence from cointegration and
Granger-causality tests. Energy Policy 2009, 37, 2926–2929. [CrossRef]
38. Acaravci, A. Structural breaks, electricity consumption and economic growth: Evidence from Turkey. J. Econ.
Forecast. 2010, 2, 140–154.
39. Lorde, T.; Waithe, K.; Francis, B. The importance of electrical energy for economic growth in Barbados. Energy
Econ. 2010, 32, 1411–1420. [CrossRef]
40. Chandran, V.G.R.; Sharma, S.; Madhavan, K. Electricity consumption–growth nexus: The case of Malaysia.
Energy Policy 2010, 38, 606–612. [CrossRef]
41. Bekhet, H.A.; Othman, N.S. Causality analysis among electricity consumption, consumer expenditure, gross
domestic product (GDP) and foreign direct investment (FDI): Case study of Malaysia. J. Int. Financ. Econ.
2011, 3, 228–235.
42. Shahbaz, M.; Tang, C.F.; Shabbir, M.S. Electricity consumption and economic growth nexus in Portugal using
cointegration and causality approaches. Energy Policy 2011, 39, 3529–3536. [CrossRef]
43. Kumari, A.; Sharma, A.K. Analyzing the causal relations between electric power consumption and economic
growth in India. Electr. J. 2016, 29, 28–35. [CrossRef]
44. Yoo, S.H.; Kim, Y. Electricity generation and economic growth in Indonesia. Energy 2006, 31, 2890–2899.
[CrossRef]
45. Squalli, J. Electricity consumption and economic growth: Bounds and causality analyses of OPEC members.
Energy Econ. 2007, 29, 1192–1205. [CrossRef]
46. Yoo, S.H.; Kwak, S.Y. Electricity consumption and economic growth in seven South American countries.
Energy Policy 2010, 38, 181–188. [CrossRef]
47. Acaravci, A.; Özturk, I. Electricity consumption–growth nexus: Evidence from panel data for transition
countries. Energy Econ. 2010, 32, 604–608. [CrossRef]
48. Apergis, N.; Payne, J.E. A dynamic panel study of economic development and the electricity
consumption–growth nexus. Energy Econ. 2011, 33, 770–781. [CrossRef]
49. Ozturk, I.; Acaravci, A. Electricity consumption and real GDP causality nexus: Evidence from ARDL bounds
testing approach for 11 MENA countries. Appl. Energy 2011, 88, 2885–2892. [CrossRef]
50. Al-Mulali, U.; Tang, C.F.; Tan, B.W.; Ozturk, I. The nexus of electricity consumption and economic growth in
Gulf cooperation council economies: Evidence from non-stationary panel data methods. Geosyst. Eng. 2019,
22, 40–47. [CrossRef]
51. Ibrahiem, D.M. Renewable electricity consumption, foreign direct investment and economic growth in Egypt:
An ARDL approach. Procedia Econ. Financ. 2015, 30, 313–323. [CrossRef]
52. Apergis, N.; Payne, J.E. Renewable and non-renewable electricity consumption–growth nexus: Evidence
from emerging market economies. Appl. Energy 2011, 88, 5226–5230. [CrossRef]
53. Apergis, N.; Payne, J.E. The electricity consumption–growth nexus: Renewable versus non-renewable
electricity in Central America. Energy Sources Part B Econ. Plan. Policy 2012, 7, 423–431. [CrossRef]
54. Al-mulali, U.; Fereidouni, H.G.; Lee, J.Y. Electricity consumption from renewable and non-renewable sources
and economic growth: Evidence from Latin American countries. Renew. Sustain. Energy Rev. 2014, 30,
290–298. [CrossRef]
55. Halkos, G.E.; Tzeremes, N.G. The effect of electricity consumption from renewable sources on countries’
economic growth levels: Evidence from advanced, emerging and developing economies. Renew. Sustain.
Energy Rev. 2014, 39, 166–173. [CrossRef]
56. Atems, B.; Hotaling, C. The effect of renewable and nonrenewable electricity generation on economic growth.
Energy Policy 2018, 112, 111–118. [CrossRef]
57. Aydin, M. Renewable and non-renewable electricity consumption-economic growth nexus: Evidence from
OECD countries. Renew. Energy 2019, 136, 599–606. [CrossRef]
58. Phiri, A.; Nyoni, B. Re-visiting the electricity–growth nexus in South Africa. Stud. Bus. Econ. 2016, 11, 97–111.
[CrossRef]
59. Zortuk, M.; Karacan, S. Energy–growth nexus revisited: An empirical application on transition countries.
Environ. Dev. Sustain. 2018, 20, 605–623. [CrossRef]
60. BNEF (Bloomberg New Energy Finance). 3Q 2018 Global PV Market Outlook; BNEF Publications: New York,
NY, USA, 2018.
http://dx.doi.org/10.1016/j.enpol.2009.03.022
http://dx.doi.org/10.1016/j.eneco.2010.05.011
http://dx.doi.org/10.1016/j.enpol.2009.10.013
http://dx.doi.org/10.1016/j.enpol.2011.03.052
http://dx.doi.org/10.1016/j.tej.2016.04.008
http://dx.doi.org/10.1016/j.energy.2005.11.018
http://dx.doi.org/10.1016/j.eneco.2006.10.001
http://dx.doi.org/10.1016/j.enpol.2009.09.003
http://dx.doi.org/10.1016/j.eneco.2009.10.016
http://dx.doi.org/10.1016/j.eneco.2010.12.018
http://dx.doi.org/10.1016/j.apenergy.2011.01.065
http://dx.doi.org/10.1080/12269328.2018.1521753
http://dx.doi.org/10.1016/S2212-5671(15)01299-X
http://dx.doi.org/10.1016/j.apenergy.2011.06.041
http://dx.doi.org/10.1080/15567249.2011.639336
http://dx.doi.org/10.1016/j.rser.2013.10.006
http://dx.doi.org/10.1016/j.rser.2014.07.082
http://dx.doi.org/10.1016/j.enpol.2017.10.015
http://dx.doi.org/10.1016/j.renene.2019.01.008
http://dx.doi.org/10.1515/sbe-2016-0009
http://dx.doi.org/10.1007/s10668-016-9901-9
Sustainability 2020, 12, 3121 22 of 22
61. BNEF (Bloomberg New Energy Finance). Technical and Financial Service in the Wind Sector; BNEF Publications:
New York, NY, USA, 2018.
62. BNEF (Bloomberg New Energy Finance). Onshore Wind the Experience Curve Revisited; BNEF Publications:
New York, NY, USA, 2018.
63. Karki, S.K.; Michael, D.M.; Hossein, S. Energy and environment in the ASEAN: Challenges and opportunities.
Energy Policy 2005, 33, 499–509. [CrossRef]
64. Dickey, D.A.; Fuller, W.A. Distribution of the Estimators for Autoregressive Time Series with a Unit Root.
J. Am. Stat. Assoc. 1979, 74, 427–431.
65. Phillips, P.C.B.; Perron, P. Testing for a unit root in time series regression. Biometrika 1988, 75, 335–346.
[CrossRef]
66. Kwiatkowski, D.; Phillips, P.C.B.; Schmidt, P.; Shin, Y. Testing the null hypothesis of stationarity against the
alternative of a unit root. J. Econom. 1992, 54, 159–178. [CrossRef]
67. Lin, B.; Moubarak, M. Renewable energy consumption—Economic growth nexus for China. Renew. Sustain.
Energy Rev. 2014, 40, 111–117. [CrossRef]
68. Johansen, S. Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector Autoregressive
Models. Econometrica 1991, 59, 1551–1580. [CrossRef]
69. Engle, R.F.; Granger, C.W.J. Co-integration and error correction: Representation, estimation, and testing.
Econometrica 1987, 55, 251–276. [CrossRef]
70. Pesaran, M.H.; Shin, Y.; Smith, R. Bounds testing approaches to the analysis of level relationships. J. Appl.
Econom. 2001, 16, 289–326. [CrossRef]
71. Fosu, O.E.; Magnus, F.J. Bound Testing Approach to Cointegration: An Examination of Foreign Direct
Investment Trade and Growth Relationships. Am. J. Appl. Sci. 2006, 3, 2079–2085.
72. Narayan, P.K. The Saving and investment nexus for China: Evidence from cointegration tests. Appl. Econ.
2005, 37, 1979–1990. [CrossRef]
73. Toda, H.Y.; Yamamoto, T. Statistical inferences in vector auto regressions with possibly integrated processes.
J. Econom. 1995, 66, 225–250. [CrossRef]
74. Jin, T.; Kim, J. Relationship between coal consumption and economic growth for OECD and non-OECD
countries. Geosyst. Eng. 2016, 19, 48–56. [CrossRef]
75. BP (British Petroleum). Available online: https://www.bp.com/content/dam/bp/business-sites/en/global/
corporate/pdfs/energy-conomics/statistical-review/bp-stats-review-2019-full-report (accessed on 17
February 2020).
76. NDRC (National Development and Reform Commission). Available online: https://www.ndrc.gov.cn/xxgk/
zcfb/tz/201803/t20180323_962694.html (accessed on 25 February 2020).
77. Lin, B.; Wang, Y. Inconsistency of economic growth and electricity consumption in China: A panel VAR
approach. J. Clean. Prod. 2019, 229, 144–156. [CrossRef]
78. REN21 (Renewable Energy Policy Network for the 21st Century). Advancing the Global Renewable Energy
Transition; REN21 Publications: Paris, France, 2018.
79. Simens Gamesa. Available online: https://www.siemensgamesa.com/about-us/location-finder (accessed on
27 February 2020).
80. Vestas. Available online: http://us.vestas.com/manufacturing (accessed on 27 February 2020).
81. GWEC (Global Wind Energy Council). Available online: https://gwec.net/wp-content/uploads/vip/GWEC_
PRstats2017_EN-003_FINAL (accessed on 6 February 2020).
82. Taeyoung, J.; Jinsoo, K. Coal Consumption and Economic Growth: Panel Cointegration and Causality
Evidence from OECD and Non-OECD Countries. Sustainability 2018, 10, 660. [CrossRef]
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
http://dx.doi.org/10.1016/j.enpol.2003.08.014
http://dx.doi.org/10.1093/biomet/75.2.335
http://dx.doi.org/10.1016/0304-4076(92)90104-Y
http://dx.doi.org/10.1016/j.rser.2014.07.128
http://dx.doi.org/10.2307/2938278
http://dx.doi.org/10.2307/1913236
http://dx.doi.org/10.1002/jae.616
http://dx.doi.org/10.1080/00036840500278103
http://dx.doi.org/10.1016/0304-4076(94)01616-8
http://dx.doi.org/10.1080/12269328.2015.1087348
https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/energy-conomics/statistical-review/bp-stats-review-2019-full-report
https://www.bp.com/content/dam/bp/business-sites/en/global/corporate/pdfs/energy-conomics/statistical-review/bp-stats-review-2019-full-report
https://www.ndrc.gov.cn/xxgk/zcfb/tz/201803/t20180323_962694.html
https://www.ndrc.gov.cn/xxgk/zcfb/tz/201803/t20180323_962694.html
http://dx.doi.org/10.1016/j.jclepro.2019.04.396
https://www.siemensgamesa.com/about-us/location-finder
http://us.vestas.com/manufacturing
https://gwec.net/wp-content/uploads/vip/GWEC_PRstats2017_EN-003_FINAL
https://gwec.net/wp-content/uploads/vip/GWEC_PRstats2017_EN-003_FINAL
http://dx.doi.org/10.3390/su10030660
http://creativecommons.org/
http://creativecommons.org/licenses/by/4.0/.
Overview of Energy–Growth Nexus
Electricity and Economic Growth
Revisited Nexus Study
Renewable–Growth Hypothesis
Data and Estimation Procedure
Estimation Strategy
Unit Root Test Results
Results of the Granger Causality
References
resources
Article
Spatial Evolution of the Energy and Economic Centers
of Gravity
Géza Tóth and Tekla Sebestyén Szép *
Faculty of Economics, University of Miskolc, 3515 Miskolc-Egyetemváros, Hungary; Geza.Toth@ksh.hu
* Correspondence: regtekla@uni-miskolc.hu
Received: 11 April 2019; Accepted: 21 May 2019; Published: 24 May 2019
����������
�������
Abstract: Increasing energy demand and economic performance can be observed in emerging markets
and, in parallel, their share in world energy use and in global GDP is growing as well. It causes
significant spatial shifts and calls attention for a new geography of energy demand. The main purpose
of this study is to reveal the spatial distribution of energy use and economic growth focusing on
the link between them. Developing gravity models, we identify the economic and energy centres
of gravity in the world and on different continents and reveal their movements between 1990 and
2015, in particular, the directions of the shifts. Bi-dimensional regression analysis and the method
of standard distance are applied to compare these movements. The study utilizes cartograms to
visualize how the space is changed and distorted by the field of force. It can be stated that the
economic and energy centre of gravity can be found in the Mediterranean Basin, but a slow and
gradual shift to the east can be observed. Currently it reflects the dominance of the north, but it marks
the position loss of the northern hemisphere and the greater importance of
developing economies
(in the southern hemisphere).
Keywords: bi-dimensional regression; economic growth; energy consumption; gravity model; spatial
models; structural change
1. Introduction
National and global energy consumption is influenced by many factors: e.g., the economic growth,
population changes, urbanization processes, structural changes (alteration in the share of primary,
secondary, and tertiary sectors), sectoral energy intensity, energy efficiency improvements, the energy
mix and, finally, national energy policy.
1.1. Global Trends in Energy Use and Economic Growth—A Dynamic Shift
There are significant differences between developed nations and developing countries (emerging
market and developing economies, see the description of the category in International Monetary
Fund (IMF) [1]) regarding the energy intensity of their economy. As Figure 1 shows, while developed
countries had produced more than four-fifths of the global gross domestic product (GDP) in 1990, this
figure dropped below 65% by 2015. At the same time these nations have contributed to the world’s
energy use to a lesser extent (compared with other country groups): in 1990 half of the energy demand
derived from developed countries, while in 2015 it was slightly more than 40% (according to BP [2] the
energy use in non-OECD countries overtook Organisation for Economic Co-operation and Development
(OECD) countries in 2008). Based on these ratios, many conclusions can be drawn: the economy of
developed countries uses energy more efficiently, which can be explained by the development stage of
the service sector, the applied technologies, and environmental and energy policies.
Modifications of these proportions cannot be considered as a final and closed process. According
to the World Energy Council [3], scenarios report that primary energy demand during the period
Resources 2019, 8, 100; doi:10.3390/resources8020100 www.mdpi.com/journal/resources
http://www.mdpi.com/journal/resources
http://www.mdpi.com
https://orcid.org/0000-0002-9233-1899
https://orcid.org/0000-0001-9985-1881
http://dx.doi.org/10.3390/resources802010
0
http://www.mdpi.com/journal/resources
https://www.mdpi.com/2079-9276/8/2/100?type=check_update&version=
2
Resources 2019, 8, 100 2 of 19
2014–2060 is expected to grow at a slower rate (between 2014 and 2030 the growth rate will be probably
1.4–1.7%, while between 2030 and 2060 it is likely to range from 0.5–1%). However, the demand
for energy, especially for crude oil and other fossil fuels, will stem from the emerging countries
(not from the developed and industrialized nations), especially China and India. The modernization,
industrialization, and urbanization processes contribute strongly to the growing and insatiable hunger
for energy sources [4]. This tendency calls attention to the shift in economic and energy centres of gravity.
Resources 2019, 8, x FOR PEER REVIEW 2 of 19
Modifications of these proportions cannot be considered as a final and closed process. According
to the World Energy Council [3], scenarios report that primary energy demand during the period
2014–2060 is expected to grow at a slower rate (between 2014 and 2030 the growth rate will be
probably 1.4–1.7%, while between 2030 and 2060 it is likely to range from 0.5–1%). However, the
demand for energy, especially for crude oil and other fossil fuels, will stem from the emerging
countries (not from the developed and industrialized nations), especially China and India. The
modernization, industrialization, and urbanization processes contribute strongly to the growing and
insatiable hunger for energy sources [4]. This tendency calls attention to the shift in economic and
energy centres of gravity.
Figure 1. Share of global GDP and total final energy consumption (TFEC) among different country
groups (1990, 2015, %). Source: own compilation based on World Bank databases [5,6].
In this study—as Table 1 shows—the energy consumption and economic growth are examined
in five-year periods. In the group of developed countries a slight increase in energy use can be
observed between 1990 and 2000 (it is 0.9% during the time period of 1990–1995 and 1.7% between
1995 and 2000 based on World Bank databases [5,6]). After that the growth stops; moreover, from the
mid-2000s it shows a declining tendency. The causes of this are probably related to a number of events
and trends: the economic and financial crisis of 2008–2010, economic development focusing on
sustainability, structural change (outsourcing of energy-intensive industrial sub-sectors to
developing countries), and energy efficiency measures (these countries take an active part in the
climate change mitigation). However, the fuel-exporting and the developing countries show an
intensive growth of energy consumption until 2010. As a consequence, the centre of gravity shifts
from the northern hemisphere to the southern hemisphere. In our view the main reason for this is
that, at the turn of the millennium, the world economy moved towards an unprecedented boom and
global economic prosperity. After 2010 the total final energy consumption shows declining growth,
and the main causes are in China. Looking at the Chinese economy the marks of possible overheating
can be seen for years, meaning that the balancing mechanisms operate only poorly (increasing
imbalances related to the Chinese yuan renminbi/US dollar (CNY/USD) exchange rate, the massive
trade surplus with the United States of America (USA), decreasing savings, extremely high
investment rate, and the danger of the shadow bank system). Although China’s economy continues
83.6
64.9
52.3
41.
5
0.6
0.4
4.0
1.2
6.9
8.9
19.5
17.7
1.1
1.2 2.5
3.2
4.3
20.4 17.9
32.2
3.5 4.2 3.7 4.3
0
%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
GDP (1990) GDP (2015) TFEC (1990) TFEC (2015)
%
GDP (current US$) and total final energy consumption (TFEC) (TJ)
Developing countries – Latin
America and the Caribbean
Developing countries – Asia
Developing countries – Africa
Fuel-exporting countries
Economies in transition
Developed economies, other
OECD members, other European
Union member states
Figure 1. Share of global GDP and total final energy consumption (TFEC) among different country
groups (1990, 2015, %). Source: own compilation based on World Bank databases [5,6].
In this study—as Table 1 shows—the energy consumption and economic growth are examined
in five-year periods. In the group of developed countries a slight increase in energy use can be
observed between 1990 and 2000 (it is 0.9% during the time period of 1990–1995 and 1.7% between
1995 and 2000 based on World Bank databases [5,6]). After that the growth stops; moreover, from
the mid-2000s it shows a declining tendency. The causes of this are probably related to a number of
events and trends: the economic and financial crisis of 2008–2010, economic development focusing on
sustainability, structural change (outsourcing of energy-intensive industrial sub-sectors to developing
countries), and energy efficiency measures (these countries take an active part in the climate change
mitigation). However, the fuel-exporting and the developing countries show an intensive growth
of energy consumption until 2010. As a consequence, the centre of gravity shifts from the northern
hemisphere to the southern hemisphere. In our view the main reason for this is that, at the turn of
the millennium, the world economy moved towards an unprecedented boom and global economic
prosperity. After 2010 the total final energy consumption shows declining growth, and the main causes
are in China. Looking at the Chinese economy the marks of possible overheating can be seen for
years, meaning that the balancing mechanisms operate only poorly (increasing imbalances related to
the Chinese yuan renminbi/US dollar (CNY/USD) exchange rate, the massive trade surplus with the
United States of America (USA), decreasing savings, extremely high investment rate, and the danger
of the shadow bank system). Although China’s economy continues to grow, its rate of growth has
steadily declined since 2010, which suits the main goals of the Chinese leadership [7]. Naturally this
affects energy consumption as well. China is responsible for more than one fifth of the global energy
demand (in 2017 it had become the world’s largest net importer), so it has a significant role in not only
the energy demand of developing countries but in influencing energy prices.
Resources 2019, 8, 100 3 of 19
Table 1. Changes of GDP (current USD) and total final energy consumption (TFEC, [TJ]) in different country groups (%). Source: own calculation based on World Bank
databases [5,6].
Title
1995/1990 2000/1995 2005/2000 2010/2005 2015/20
10
Change in
GDP (%)
Change in
TFEC (%)
Change in
GDP (%)
Change in
TFEC (%)
Change in
GDP (%)
Change in
TFEC (%)
Change in
GDP (%)
Change in
TFEC (%)
Change in
GDP (%)
Change in
TFEC (%)
Developed economies, other
OECD members, other European
Union member countries
6.5 0.9 1.5 1.7 6.5 0.6 3.9 –0.2 1.0 −0.3
Economies in transition −4.4 −10.5 −4.6 −3.1 21.0 1.7 12.4 −0.4 −1.6 −3.4
Fuel-exporting countries −0.8 −2.0 2.3 0.5 14.5 2.3 17.0 3.2 1.6 1.6
Developing economies 10.4 3.4 3.3 1.1 9.8 5.7 17.9 5.0 7.8 3.3
from this:
Africa 10.8 3.7 6.8 0.6 12.0 6.7 18.9 5.5 10.6 3.5
Asia 4.2 2.0 −0.1 2.1 12.3 3.3 10.0 2.4 1.6 2.8
Latin America and the Caribbean 12.7 2.7 −0.5 3.3 5.1 2.0 18.0 3.6 0.2 2.0
World 6.3 0.6 1.7 1.3 7.3 2.4 7.0 2.1 2.5 1.3
Resources 2019, 8, 100 4 of 19
The economies in transition present an interesting group. In these countries of the former Soviet
Union as a result of regime change and the radical structural change (relate to the decline of heavy
industry) the energy use sharply declined until 2000 (related to the decline of heavy industry), and
apart from a short time period it continuously decreases.
The visualization and map view of the processes described above are limited; just a few examples
can be found in the literature focusing on centres of gravity (such as Zhang et al. [8] and Wang et al. [9]).
Analysing the geographical differences in global energy use and in the world economy and explaining
spatial processes are essential to realize and solve global problems. These provide information to the
field of energy security, help tackle global environmental challenges, and contribute to carrying out
deeper energy studies. Spatial modelling serves as a tool for examining the shift in the world’s energy
and economic centres of gravity and it helps to visualize this spatial evolution.
Our main research questions are the following. How did the energy and economic centres of
gravity move during the period 1990–2015? Can there be a strong link between energy use and
economic growth identified using the centre of gravity method and the model of two-dimensional
regression? Is there significant difference between the global level and the level of continents?
1.2. Energy Consumption and Economic Growth
Opinions about the role of energy in economic growth range themselves along a very wide scale.
Both the representatives of the ecological economics and energy economics argue that the main reason
for economic growth is energy and energy use (as a production factor). Moreover, there is a radical
viewpoint according to which energy consumption is the only indicator of development (such as the
Olduva theory of Duncan [10]). This is explained by the important and unquestionable role of energy
in the production process. Proponents of this stance argue that there is no economic activity for which
there is no need for energy [11,12], based on the simple fact that production is a workflow and work
needs energy investment [13].
Kovács [14,15] emphasizes in his research the strong relationship between energy consumption
and economic growth; moreover, in his view “the economic welfare and well-being of a state is
characterized by the level of energy use” [14] (p. 63), and “the mineral production and meeting
the energy demand is one most important ways to enhance economic welfare and the standard of
living” [15] (p. 47). Lakatos and Lakatosné Szabó [16] have a similar opinion and examine the
connection between global GDP and oil demand.
Researchers from the field of ecological and energy economics tend to agree with the primary
role of energy in economic growth. However, there is a significant difference in how they see this
determinant role. Stern and Cleveland [11] emphasize energy availability, and according to Murphy
and Hall [13] the increase in the amount of energy used is important. Berndt and Wood [17] and
Schurr [18] were pioneers in recognizing the economic importance of energy quality: in their view
the best available energy source at that time was electrical energy and furthermore—primarily—the
improvements in energy quality contributed the most to the decreasing energy intensity. Stern [19]
argues for the magnitude of energy quality in economic growth as well.
Ayres et al. [20,21] and Akizu et al. [22] explain the unprecedented economic growth during the
first and second industrial revolution by the continuously declining real price of energy and with
the increasing availability of energy. These changes “made human workers vastly more productive
than they would otherwise have been” and most of the social classes gained access to energy (such as
internal combustion engines) [21] (p. 185). In human history the first, second, and third industrial
revolutions resulted in significant development, and it was “a result of a combination of the availability
of a particular resource and the technological development of processing the available source” [22]
(p. 18046). Most of the new technological improvements have been connected to previously unknown
or not efficiently used energy sources (coal to the steam engine; crude oil to internal combustion
engines; nuclear energy to cheap electrical energy).
Resources 2019, 8, 100 5 of 19
A strong relationship can be observed not only between energy use and economic growth
but between energy consumption and human development index (HDI) as well [23]. The basic
assumption of the decoupling theory is also that there is a strong and causal relationship between these
former two indicators. The term decoupling refers to breaking the link between the environmental
pressure and economic driving forces (economic growth, GDP) and it looks at a possible way towards
sustainable development.
This topic (spatial movements of global energy use, main causes of consumption, and reorganization
of global energy markets) can be interpreted as the border of energy economics, energy geographics
(Munkácsy [24] and Calvert [25] describe this in detail), and energy geopolitics (related to global
economic growth and the increasing share of emerging nations). However, this study is closer to
energy economics, because our main goal is to reveal the connection between economic growth and
energy consumption and identify how these factors affect each other in space and time.
The analysis of the correlation and causal relationship between economic growth and energy
consumption has been a central topic in energy economics for several decades. In this field of economic
research the study of Kraft and Kraft [26] was the pioneering work, analysing the relationship between
energy consumption and gross national product (GNP) in the USA between 1947 and 1974. In the
more than four decades since then many publications have been issued but their results are often
contradictory. These studies demonstrate the positive and strong correlation between the economic
growth and energy consumption (e.g., [27–30]). However, it is not clear which factor is the dependent
or independent one. According to Cleveland et al. [31] “the correlation between aggregate energy
consumption and GDP in both industrialised and developing countries is undeniable. While industrial
countries appear to have partly decoupled growth in energy consumption from growth in GDP in
recent years, the close correlation reappears when energy consumption is weighted by the quality of
different fuel types” (cited by Sorrell and Dimitropoulos [27] (p. 8)).
The starting point for our study is the assumption that there is a positive and strong correlation
between energy use and economic growth in the world economy. That is why we find it expedient to
examine the co-movement of these two indicators. Our hypotheses are as follows:
• The energy and economic centres of gravity move closely together not only globally, but also on a
continental level.
•
can be drawn from the analysis of centres of gravity (from the gravity models) related
to the causality directions of economic growth and energy use.
1.3. Fields of Force in the World, or the Centre of Gravity Method
Using methods based on gravity models is not new in the economic literature; Nagy [32] gives a
comprehensive study on its application. Kincses and Tóth [33] found that the gravitational models
based on physical analogy are used for spatial flow and to delimit catchment areas.
In the pioneering work of Quah [34]—which relies heavily on the experiences of Klein [35] and
Grether and Mathys [36]—the dynamics of the global economy’s centre of gravity are described
between 1980 and 2007 based on 693 identifiable locations on Earth. He concluded that the world’s
economic centre of gravity is in the North Coast of Africa.
Based on the literature it can be stated that the research topic of energy centres of gravity is
underrepresented; only a few papers can be found focusing on it (i.e., [8,9,37]). Fesharaki [37] first uses
the term energy centre of gravity. In his analysis he argues that the centre of gravity of the world’s
energy market is shifting to the Asia-Pacific region, but his conclusion is mainly based on simple
statistical analysis; no gravity model calculations were carried out.
The gap between Chinese energy supply and demand by energy sources (coal, oil, gas, electricity)
is examined in Zhang et al. [8] using gravity models, for the time period of 1997–2009. The study
reveals the spatial variations between energy production and consumption, making suggestions for
infrastructure development.
Resources 2019, 8, 100 6 of 19
The main purpose of Wang et al. [9] is to study the spatial distribution and centres of gravity for
global energy supply and demand (oil, gas and coal) using the centre of gravity theory. Not only the
movement of the centres of gravity is calculated, revealing movement towards the south-southeast,
but the speed of this shift as well. However, we observe a lack of explanations and detection of
main causes.
In the current study, we employ the gravity model to examine global and also continental
energy and economic centres of gravity. In our view this study goes beyond the research objective
of Quah [34], as our main goal is not to determine the economic centre of gravity and to follow its
movements but to compare these spatial changes with the shift in energy centres of gravity. In this way
the co-movements (convergence or divergence processes) become identifiable, which should serve
to identify the relationship between energy use and economic growth and to analyse its spatiality.
We strive to reveal the main causes and give comprehensive explanations of the shifts and their trends.
Our main goal is to identify the economic and energy centres of gravity in the world and on
different continents and to reveal their movements, in particular the directions of the shifts. The rest of
this paper is organized as follows: The
section introduces the centre of gravity
method and bi-dimensional regression, and shows the database and the country classification system
applied in the study. The results connecting to the shift in energy and economic centres of gravity are
highlighted in maps and further conclusions are drawn based on two-dimensional regression, the
Chow test, and the calculation of standard distance. Finally, the last chapter concludes this study.
2. Materials and Methods
Annual data for 204 data points (i.e., countries, territorial units, and provinces), as listed below,
are applied in the calculations collected from World Bank databases [5,6]:
• GDP (current USD); and
• Total final energy consumption (in terajoule (TJ); abbreviated as TFEC).
The following country groups are made based on UN [38] and IMF [1] country classification
methodology:
• developed economies are the advanced economies based on IMF [1], other OECD members, and
other European Union member states;
• economies in transition based on IMF [1]: economies in transition in South-Eastern Europe and the
Commonwealth of Independent States and Georgia (we note here that there is significant overlap
between the category of economies in transition and the category of fuel-exporting countries; if a
country is listed in the latter category, it is not listed in the former);
• fuel-exporting countries based on IMF [1] same category;
• developing countries: every other country located in Asia, in Latin America and the Caribbean,
and in Africa; and
• The sample period is from 1990–2015.
2.1. The Gravity Model with Special Regard to the Centre of Gravity Method
Based on Nemes Nagy [39] the application of this model should be achieved as follows:
“the co-ordinates of the gravity centres in a planar system consisting of n elements can be calculated
as the weighted arithmetical means of the co-ordinates of the points in condition that the location of
the points in the system of co-ordinates (map) is fixed and all the points are associated with ‘weights’.
The centre of gravity represents an optimal point: the weighted sum of the distances between gravity
centre and the basic points is minimal” [39] (pp. 75–76). The equations are given as:
x =
∑n
i=1 fixi∑n
i=1 f
i
(1)
Resources 2019, 8, 100 7 of 19
y =
∑n
i=1 fiyi∑n
i=1 fi
(2)
where x and y are the points of gravity (x, y); xi and yi are co-ordinates of basic points and fi shows
their weights.
The calculation can be carried out on any spatial unit, we can find examples for global analyses
(such as Grether and Mathys [36] and Quah [34]) or regional analysis (such as Zhang et al. [8]).
As Nemes Nagy [39] confirms, not only the population can be used as the role of weight, but any other
social or economic variable (such as income, number of employees, energy use or emissions, etc.).
Analysing the spatial transformation and the shift of the centre of gravity can reveal the distance and
direction of changes in spatial structure. A disadvantage of the model is that “substantial changes in
spatial structure can occur without a slight movement of the gravity centre, when the changes (growth
or increase) take place symmetrically around the gravity centre” [39] (p. 76). Here we note that in this
study global processes are examined. The complexity of world economic system and the diversity of
the world’s countries exclude the possibility of changes in energy use and economic growth taking
place symmetrically around the gravity centre, which eliminates this problem.
Limitations of Gravity Model
Similar to an earlier study of ours [40] the geometric centre of the area for all countries is weighted
by nominal GDP and TFEC obtained from the World Bank World Development Indicators and the
Sustainable Energy 4 All databases. This method—especially for geographically large countries—raises
many problems, because the geometric centre is far from any significant city or industrial centre.
However, from our point of view, it is not the spatial location of the centres of gravity that is important
but the changes over time and the co-movements of the two sets of points. Although the literature
contains a method that eliminates the negative outcomes of the projection [41,42], we have chosen not
to apply it. In our opinion, the original version of the centre of gravity method is best suited for the
main objectives of this study. Naturally, the specific location of the centres of gravity is not considered
relevant because the economic activity and the energy use are often confined to only a part of the
country (such as China), while we weighted the geometrical centre of the entire national territory
in this analysis. The examination of tendencies and spatial co-movements is much more important,
as appropriate conclusions may be drawn from the results with respect to the direction of the processes
taking place in the world economy.
The new information resulted by this method provides additional evidence for identifying
world-wide processes and economic restructuring. ETRS89 (European Terrestrial Reference System) is
applied to locate the points and create the maps, which is the current version of the standard spatial
reference system operated by Eurostat. Our aim is to examine all the countries of the world’s economic
and energy centre of gravity together, while examining the continents (Africa, the Americas (North
and Latin America), Asia, Australia together with Oceania, and Europe) separately. Here we note that
in our model, Russia is considered as a part of Europe.
2.2. Application of Two-Dimensional Regression
The point set obtained by the gravitational calculation using GDP is worth comparing with the
calculation using the total final energy consumption, examining how the space is changed and distorted
by the field of force. The comparison, of course, can be done by a simple cartographic representation,
but with such a large number of points, that does not really promise good results. It is much better to
use a two-dimensional regression.
Two-dimensional regression is one the methods of comparing partial shapes. The comparison
is possible only if one of the point coordinates in the coordinate systems differing from each other
is transformed to another coordinate system by an appropriate rate of displacement, rotation, and
scaling. Thus, it is possible to determine the degree of local and global similarities of shapes, as well
as their differences, which are based on the unique and aggregated differences between the points
Resources 2019, 8, 100 8 of 19
of the shapes transformed into a common coordinate system. The method was developed by W. R.
Tobler, who published a study describing this procedure in 1994 after the precedents of the 1960s and
1970s [43–47]. As for the equation relating to the calculation of the Euclidean version, see [48–50].
In Table 2 x and y are the coordinates of independent shapes; a and b are the coordinates of
dependent shapes; ai and bi represent the coordinates of dependent shapes in the system of independent
shapes; α1 determines the measure of horizontal shift, while α2 determines the measure of vertical
shift; β1 and β2 are to derive the scale and angle values. Hence, α1, α2 represents a translation relative
to the original location: up or down; and β1 and β2 indicates by how much the points have been
rotated and expanded or contracted. Φ and Θ determine the angle of shifting. SST is the total square
sum of difference, SSR is the square sum of difference explained by regression and SSE is the square
sum of difference not explained by the regression (residual). Further details about the background of
the two-dimensional regression can be seen in [50] (pp. 14–15).
Table 2. The equations of the bi-dimensional Euclidean regression. Source: [47,49] as cited by [50]
(p. 14).
Step Equation
1. Equation of the regression
(
A
′
B′
)
=
(
α1
α2
)
+
(
β1
β2
−
β2
β1
)
∗
(
X
Y
)
2. Scale difference Φ =
√
β21 + β
2
2
3. Rotation Θ = tan−1
(
β2
β1
)
4.Calculation of β1 β1 =
∑
(ai−a)∗(xi−x)+
∑
(bi−b)∗(yi−y)∑
(xi−x)
2
+
∑
(yi−y)
2
5.Calculation of β2 β2 =
∑
(bi−b)∗(xi−x)−
∑
(ai−a)∗(yi−y)∑
(xi−x)
2
+
∑
(y−y)2
6. Horizontal shift α1 = a −β1 ∗ x + β2 ∗ y
7. Vertical shift α2 = b −β2∗x −β1∗y
8. Correlation based on error terms r =
√√√
1 −
∑[
(ai−a′i )
2
+(bi−b′i )
2
]
∑[
(ai−a)
2
+(bi−b)
2
]
9. Breakdown of the square sum of the
difference
∑[
(ai − a)
2
+
(
bi − b
)2]
=∑[(
a′i − a
)2
+
(
b′i − b
)2]
+
∑[(
ai − a
′
i
)2
+
(
bi − b
′
i
)2]
SST = SSR + SSE
10. Calculation of A′ A′ = α1 + β1(X)−β2(Y)
11. Calculation of B′ B′ = α2 + β2(X) + β1(Y)
2.3. Examination of Shift in Centres of Gravity (Standard Distance)
To describe the movement of centres of gravity, the standard of distance is measured. It is an
indicator for spatial deviation of points. It measures the degree to which features are concentrated
or dispersed around the gravity centre (in an input feature class) [51]. The larger the value, the less
characteristic it is that the given social or economic variable concentrates around the gravity centre [39].
The first step in measuring the standard distance is to calculate the mean centre of the centre of gravity,
which is actually the average mean of the coordinates of the centre of gravity. The average movement
of coordinates is compared with the mean both in the whole time period and in its two phases, as well.
Basically the standard distance is the average and weighted distance of centres of gravity; it is the
radius of a circle drawn around an average centre. The formula of standard distance is:
SD =
√√√√
n∑
i=1
(
Xi − X
)2
n
+
n∑
i=1
(
Yi − Y
)2
n
(3)
where SD is the standard distance (in this case geographical coordinates), Xi and Yi are coordinates of
gravity centres, X and Y are the coordinates of the average centres and n is the number of years.
Resources 2019, 8, 100 9 of 19
2.4. Chow Test for Structural Changes
In a situation when a significant shift occurs in the direction of a centre of gravity, we obtain a
structural break. This can be tested with the Chow test. Actually the Chow test contributes to verifying
whether one or two separate regression lines best fit a split set of data. To do that, the time series data
are divided into two subsamples (A and B) based on a date. For calculating the Chow F statistic the
formula is [52] (p. 334):
Fc =
(RSSR − RSS1 − RSS2)/k
(RSS1 + RSS2)/(n − 2k)
(4)
where RSSR is the residual sum of squares of the model using the entire pooled sample, RSS1 is the
residual sum of squares of the model using only subsample A, RSS2 is the residual sum of squares of
the model using only subsample B. A and B are two distinct subsamples. The null hypothesis for the
test is that there is no break point. Under H0, the F-statistic follows an F-distribution with k and n − 2k
degrees of freedom (k is the total number of parameters).
3. Results
Before presenting our calculation results and the spatial evolution of the energy and economic
centres of gravity, we highlight the changes in the position of world nations regarding changes in the
total final energy consumption and GDP between 1990 and 2015 based on World Bank databases [5,6].
It reveals the energy and economic trends in a structured way. In the visualization process the
formerly-presented country classifications were followed. The four corners of Figure 2 show four types
of development pathway. In the countries, located in the upper left corner of the Figure 2, while the
average annual growth rate of total final energy consumption is above the world average, the economic
growth rate does not follow it (it was under world average). In the bottom right corner, the trends are
the opposite of that. In the upper right corner, the average annual growth rate of both indicators is
above the world average, in the bottom left corner it is below average.
Resources 2019, 8, x FOR PEER REVIEW 9 of 19
F = (RSS − RSS − RSS )/k(RSS + RSS )/(n − 2k) (4)
where RSSR is the residual sum of squares of the model using the entire pooled sample, RSS1 is the
residual sum of squares of the model using only subsample A, RSS2 is the residual sum of squares of
the model using only subsample B. A and B are two distinct subsamples. The null hypothesis for the
test is that there is no break point. Under H0, the F-statistic follows an F-distribution with k and n −
2k degrees of freedom (k is the total number of parameters).
3.
Before presenting our calculation results and the spatial evolution of the energy and economic
centres of gravity, we highlight the changes in the position of world nations regarding changes in the
total final energy consumption and GDP between 1990 and 2015 based on World Bank databases
[5,6]. It reveals the energy and economic trends in a structured way. In the visualization process the
formerly-presented country classifications were followed. The four corners of Figure 2 show four
types of development pathway. In the countries, located in the upper left corner of the Figure 2, while
the average annual growth rate of total final energy consumption is above the world average, the
economic growth rate does not follow it (it was under world average). In the bottom right corner, the
trends are the opposite of that. In the upper right corner, the average annual growth rate of both
indicators is above the world average, in the bottom left corner it is below average.
Figure 2. Changes in the position of world nations regarding changes in total final energy
consumption and GDP (1990–2015, average annual growth rate, %). Source: own compilation based
on World Bank databases [4,5].
Furthermore, Figure 2 gives additional information. It specifies the top 10 energy consumer
countries based on their share of global final energy consumption [6]. In this list there are four
developed countries (in order of energy use the rank is: the USA, Japan, Germany, and Canada) and
these nations can be found in the bottom left corner. This placement means that the average annual
growth rate of GDP and TFEC are under the world average. This can be explained by the significant
share of the tertiary sector (which is much less energy intensive), energy efficiency improvements,
stagnation of population numbers (Japan, Germany) and small but balanced growth (Canada, the
USA). These countries have already performed absolute or relative decoupling, so they could succeed
Brazil
Canada
China
Germany
India
Indonesia
Iran
Japan
Russia
USA
developed economies
economies in transition
developing economies
fuel-exporting countries
-5
0
5
10
15
-5 0 5 10 15 20 25
C
ha
ng
es
o
f T
FE
C
(
19
90
-2
01
5,
%
)
Changes of GDP (1990-2015, %)
world average
world average
Figure 2. Changes in the position of world nations regarding changes in total final energy consumption
and GDP (1990–2015, average annual growth rate, %). Source: own compilation based on World Bank
databases [4,5].
Resources 2019, 8, 100 10 of 19
Furthermore, Figure 2 gives additional information. It specifies the top 10 energy consumer
countries based on their share of global final energy consumption [6]. In this list there are four
developed countries (in order of energy use the rank is: the USA, Japan, Germany, and Canada) and
these nations can be found in the bottom left corner. This placement means that the average annual
growth rate of GDP and TFEC are under the world average. This can be explained by the significant
share of the tertiary sector (which is much less energy intensive), energy efficiency improvements,
stagnation of population numbers (Japan, Germany) and small but balanced growth (Canada, the
USA). These countries have already performed absolute or relative decoupling, so they could succeed
in breaking the link between environmental pressure and economic driving forces (for more details see
Szlávik and Sebestyén Szép [53]).
Considering the group of transition economies, only Russia (located in the bottom right corner
of Figure 2) belongs to this top 10 list. While the average annual growth rate of its TFEC is below
the world average, its GDP is above average. In Russia the negative consequences of the decline of
heavy industry, the deindustrialization and restructuring, can still be felt. According to the World Bank
databases [4,5] and Weiner [54] the Russian economy managed to reach the production level of 1991
(the year of the dissolution of the Soviet Union) only in 2006.
In the upper right corner of Figure 2 (where the average annual growth rate of both indicators is
above average) are located China, India, Brazil, Indonesia, and Iran. The last two belong to the group
of fuel-exporting countries (we note here that, in Indonesia, the export of fuel exceeds import by a few
percentage points), and Iran was the 10th largest net exporter in 2015 as a result of the lifted sanctions
(connected to the Iran nuclear deal). However, some duality characterizes Iran. The energy efficiency of
its economy (calculated as units of GDP per unit of energy) based on World Bank databases [5,6] is far
from the world average. In 2014 it was 5.6 USD (2011, purchasing power parity (PPP)), while the global
average at that time was 7.9 USD. However, at 3023.5 kilogram of oil equivalent (koe) the energy use per
capita is extremely high (the world average is 1919.4 koe), the same as in the European Union. The main
reasons are in connection with the high share of energy-intensive industrial subsectors (oil production
and refining) in the economy, the less-developed transport sector (road transport dominancy), and the
substantial subsidizing of energy prices (in the latter case the government recently embarked on an
aggressive energy price reform—see Moshiri [55] for more details). As for China, India and Brazil,
they clearly appear in the top 10 as a result of their expansive economic policies, increasing population
and the industrialization process, producing these high rankings as the largest energy consumers in
the world.
Hereinafter the shift in energy and economic centres of gravity are presented (Figures 3 and 4).
In Africa both these centres of gravity nearly overlap with the geometric centre of the continent, located
in the Central African Republic and in the Democratic Republic of the Congo. The shifts are small,
but in many cases these move hectically: while the energy centre of gravity clearly moved towards the
north (primarily as a result of rapid population and economic growth in North Africa), the economic
centre of gravity followed that route until 2001/2002, after which it turned to the south until 2010, from
which it moved again towards the north.
Probably the continent of the Americas shows the clearest picture about the vulnerability of
developing countries lacking export diversification. These nations gained extra benefit from export-led
growth, but, in parallel, the exposure of such economies to economic crisis and price volatility of raw
materials is very high. As a result of reform processes after the debt crisis in 1980s, South America
showed intensive economic growth in the first half of the 1990s. During this time period (1990–1995)
the economic centre of gravity moves towards the southeast. However, the 1997–1998 Asian financial
crisis strongly influenced their economies and this led to steps backwards in the development that had
only just begun. As a consequence, the centre of gravity shifts towards the northwest again.
In the early 2000s the bursting of the dot-com bubble resulted in an economic slowdown in North
America, but the economic performance of Argentina and Brazil developed, and as a consequence
the economic centre of gravity shifted again towards the southeast. For a long time, it seemed
Resources 2019, 8, 100 11 of 19
that the 2008–2009 financial crisis would avoid South America, but, finally, the negative effects of
the price decline of raw materials (and later its stagnation) and the declining growth of the world
economy (especially in China) reached the continent. The investors lost their appetite for emerging
(and developing) markets, so these negative tendencies turned the shift of centre of gravity towards
the northwest.
By contrast the energy centre of gravity moved clearly towards the southeast during the examined
time period and it is not influenced by the various economic crises, either. One factor is in connection
with improvements in energy efficiency in North America, while another is the decrease in the TFEC
due to the development of the service sector and structural changes. As North America becomes more
and more efficient, the energy use per capita is shrinking (between 1990 and 2015 it reduced by more
than 10%) and the energy intensity improves (less energy is required to generate one unit of GDP).
Energy intensity increases by 55% during the examined time period in the USA, based on World Bank
databases [4,5]. In contrast, in the global South the energy use per capita continuously increases thanks
to the rising standard of living and economic development, while the energy intensity of the economy
does not significantly change.
Resources 2019, 8, x FOR PEER REVIEW 11 of 19
reduced by more than 10%) and the energy intensity improves (less energy is required to generate
one unit of GDP). Energy intensity increases by 55% during the examined time period in the USA,
based on World Bank databases [4,5]. In contrast, in the global South the energy use per capita
continuously increases thanks to the rising standard of living and economic development, while the
energy intensity of the economy does not significantly change.
In the case of Europe, remember that Russia is considered as a part of Europe in this study. Our
decision to include it is primarily justified by the fact that that most of the Russian economic
performance and population is tied to Europe, geographically. In this study we ignore the
geopolitical theories according to which Europe is only the western peninsula of the giant
“supercontinent” of Eurasia and that find Eurasia (as a political force field) to be one of the most
important geopolitical concepts (see Brzezinski [56]). The geometric centre of Russia is located in the
eastern part of Russia. Figure 3 highlights that Europe is the only continent where the economic and
energy centres of gravity move together, but a five-year delay can be observed. The direction of the
shift was west-southwest, but while the economic centre had turned to the east with the economic
growth in 2000, in the case of TFEC this change happened later (only in 2005). This can be related to
the fact that after the regime change in East-Central Europe the industrial output sharply declined
(over a few years) but as a result of the economic liberalism and the fast adoption of privatization
policies, the volume of inward FDI flows increased, contributing to significant economic growth. The
banking system started to develop and the share of the tertiary sector increased. The region gradually
regained its economic position in the continent. After 2005 the energy use tendency stopped falling,
and stagnation of energy intensity could be observed. The re-industrialization process became more
significant after the 2008–2009 financial crisis and the national share of the industrial sector increased
again. As a consequence, the energy centre of gravity shifted again towards the east.
Figure 3. Shift in economic and energy centres of gravity for the continents of Africa, Europe, and the
Americas (1990–2015). Source: own compilation.
Figure 3. Shift in economic and energy centres of gravity for the continents of Africa, Europe, and the
Americas (1990–2015). Source: own compilation.
In the case of Europe, remember that Russia is considered as a part of Europe in this study.
Our decision to include it is primarily justified by the fact that that most of the Russian economic
performance and population is tied to Europe, geographically. In this study we ignore the geopolitical
theories according to which Europe is only the western peninsula of the giant “supercontinent” of
Eurasia and that find Eurasia (as a political force field) to be one of the most important geopolitical
concepts (see Brzezinski [56]). The geometric centre of Russia is located in the eastern part of Russia.
Figure 3 highlights that Europe is the only continent where the economic and energy centres of gravity
Resources 2019, 8, 100 12 of 19
move together, but a five-year delay can be observed. The direction of the shift was west-southwest, but
while the economic centre had turned to the east with the economic growth in 2000, in the case of TFEC
this change happened later (only in 2005). This can be related to the fact that after the regime change in
East-Central Europe the industrial output sharply declined (over a few years) but as a result of the
economic liberalism and the fast adoption of privatization policies, the volume of inward FDI flows
increased, contributing to significant economic growth. The banking system started to develop and
the share of the tertiary sector increased. The region gradually regained its economic position in the
continent. After 2005 the energy use tendency stopped falling, and stagnation of energy intensity could
be observed. The re-industrialization process became more significant after the 2008–2009 financial
crisis and the national share of the industrial sector increased again. As a consequence, the energy
centre of gravity shifted again towards the east.
Resources 2019, 8, x FOR PEER REVIEW 12 of 19
In Asia the economic centre of gravity was located in China between 1990 and 2015; however,
the energy centre of gravity moved to the eastern part of Bangladesh after 2007. Until 1995 the
economic centre of gravity moved towards the east as a result of the increasing economic
performance of the Tiger Cub Economies. After that, with its successful reform and opening-up
policy, China became an increasingly important player on this continent. In parallel with that, after
the turn of the millennium the economic centre of gravity shifted west. From 2014–2015 a sharp
movement can be observed towards the south-southeast, caused by declining Chinese economic
growth.
Figure 4. Shift in economic and energy centres of gravity in the continent of Asia and of Australia
together with Oceania (1990–2015). Source: own compilation.
The movement of the energy centre of gravity is less dynamic in Asia than the economic centre.
Up to 2000 it moves towards the southeast, after that clearly towards the west—this is in connection
with the growing energy hunger of India. According to IEA [57], in 2016 India was the 2nd largest
net coal importer (after China) and 3rd largest net oil importer in the world. However, probably this
shift is smaller—compared with the GDP—because the structural change of the economy on the
continent is delayed; though the share of the industry is still significant, serious restructuring cannot
be observed.
In the case of Australia and Oceania the economic and energy centres of gravity move within
Australia. The size of the shift is not significant; Australia dominates the geographical region.
We conclude that globally the examined time period can be divided into three main parts (Figure
5). Between 1990 and 1995 the shift of the energy and economy centres of gravity passes in the
opposite direction. The economic centre of gravity moves to the southeast, indicating the rise of China
and the success of the Chinese economic reform (for more details see Simon [58]). In spite of that, the
energy centre of gravity moves towards the south-southwest direction. This can be explained by the
dramatic structural change, with special regard to the decline in the industrial output and to the post-
Soviet heavy industrial meltdown, which was one of the most significant energy-intensive industrial
sectors earlier. From 1995–2000 no significant shift can be identified; this time period can be labelled
Figure 4. Shift in economic and energy centres of gravity in the continent of Asia and of Australia
together with Oceania (1990–2015). Source: own compilation.
In Asia the economic centre of gravity was located in China between 1990 and 2015; however, the
energy centre of gravity moved to the eastern part of Bangladesh after 2007. Until 1995 the economic
centre of gravity moved towards the east as a result of the increasing economic performance of the
Tiger Cub Economies. After that, with its successful reform and opening-up policy, China became an
increasingly important player on this continent. In parallel with that, after the turn of the millennium
the economic centre of gravity shifted west. From 2014–2015 a sharp movement can be observed
towards the south-southeast, caused by declining Chinese economic growth.
The movement of the energy centre of gravity is less dynamic in Asia than the economic centre.
Up to 2000 it moves towards the southeast, after that clearly towards the west—this is in connection
with the growing energy hunger of India. According to IEA [57], in 2016 India was the 2nd largest net
coal importer (after China) and 3rd largest net oil importer in the world. However, probably this shift
Resources 2019, 8, 100 13 of 19
is smaller—compared with the GDP—because the structural change of the economy on the continent is
delayed; though the share of the industry is still significant, serious restructuring cannot be observed.
In the case of Australia and Oceania the economic and energy centres of gravity move within
Australia. The size of the shift is not significant; Australia dominates the geographical region.
We conclude that globally the examined time period can be divided into three main parts (Figure 5).
Between 1990 and 1995 the shift of the energy and economy centres of gravity passes in the opposite
direction. The economic centre of gravity moves to the southeast, indicating the rise of China and the
success of the Chinese economic reform (for more details see Simon [58]). In spite of that, the energy
centre of gravity moves towards the south-southwest direction. This can be explained by the dramatic
structural change, with special regard to the decline in the industrial output and to the post-Soviet
heavy industrial meltdown, which was one of the most significant energy-intensive industrial sectors
earlier. From 1995–2000 no significant shift can be identified; this time period can be labelled as path
searching. One year the centre of gravity moves towards one direction, the next year towards the
opposite direction. This tendency is caused by the fact that after collapse of the Soviet Union in 1991 the
world became unipolar, but it was a very short-lived situation; new challengers appeared (emerging
markets) on the world stage. Naturally, these processes pull the centres of gravity in different directions.
Around 2001–2002 a kind of uncertainty ceased, and both the energy and the economic centres of
gravity moved to the east-southeast. After 2001–2002 both centres of gravity moved together; the size
and direction of the shift were nearly exactly the same.
Resources 2019, 8, x FOR PEER REVIEW 13 of 19
as path searching. One year the centre of gravity moves towards one direction, the next year towards
the opposite direction. This tendency is caused by the fact that after collapse of the Soviet Union in
1991 the world became unipolar, but it was a very short-lived situation; new challengers appeared
(emerging markets) on the world stage. Naturally, these processes pull the centres of gravity in
different directions. Around 2001–2002 a kind of uncertainty ceased, and both the energy and the
economic centres of gravity moved to the east-southeast. After 2001–2002 both centres of gravity
moved together; the size and direction of the shift were nearly exactly the same.
Between 1990 and 2015 the energy and economic centres of gravity could be found in the
Mediterranean Basin (but here we note it moves from Spain—the Iberian Peninsula—towards Tunis
and the eastern coast of the Mediterranean Sea), which refers to the dominance of North. However,
this movement towards the southeast indicates the position loss of the Northern Hemisphere and the
fact that the developing (emerging) countries (global south) are becoming more and more important.
Figure 5. Shift in global economic and energy centres of gravity (1990–2015). Source: own compilation.
3.1. Movement of Gravity Centres
To describe the movement of gravity centres the standard of distance is measured. Table 3 shows
the results. It can be stated that the movement of economic centres of gravity is much larger (on every
continent and every time period) than the movement of energy centres of gravity. Generally, the
movement of gravity centres was smaller between 1990 and 2000 than in the second time period
(between 2001 and 2015).
Figure 5. Shift in global economic and energy centres of gravity (1990–2015). Source: own compilation.
Between 1990 and 2015 the energy and economic centres of gravity could be found in the
Mediterranean Basin (but here we note it moves from Spain—the Iberian Peninsula—towards Tunis
and the eastern coast of the Mediterranean Sea), which refers to the dominance of North. However,
Resources 2019, 8, 100 14 of 19
this movement towards the southeast indicates the position loss of the Northern Hemisphere and the
fact that the developing (emerging) countries (global south) are becoming more and more important.
3.1. Movement of Gravity Centres
To describe the movement of gravity centres the standard of distance is measured. Table 3 shows
the results. It can be stated that the movement of economic centres of gravity is much larger (on every
continent and every time period) than the movement of energy centres of gravity. Generally, the
movement of gravity centres was smaller between 1990 and 2000 than in the second time period
(between 2001 and 2015).
Table 3. Weighted standard distance of economic and energy centres of gravity (degree) Source:
own calculation.
Region Time GDP TFEC
World
1990–2015 5.90 5.28
1990–2000 4.22 1.68
2001–2015 6.08 3.91
Africa
1990–2015 1.30 0.77
1990–2000 1.16 0.36
2001–2015 0.97 0.43
Americas
1990–2015 2.65 1.54
1990–2000 1.49 0.53
2001–2015 2.65 1.13
Asia
1990–2015 6.60 1.15
1990–2000 1.80 0.58
2001–2015 3.38 0.80
Australia
1990–2015 0.46 0.27
1990–2000 0.31 0.13
2001–2015 0.42 0.22
Europe
1990–2015 2.36 2.86
1990–2000 1.91 0.43
2001–2015 1.96 0.83
3.2. Results of Bi-Dimensional Regression
One of the main objectives of the bi-dimensional regression is to reveal how geometric displacement
can be used to get from the point set obtained by the gravitational calculation using GDP to the point
set obtained by the gravitational calculation using TFEC. The results presented in Table 4 show a strong
correlation between the point set obtained by the gravitational calculation using GDP and TFEC in Asia
(the value of it is 0.746), but globally, in America, and in Australia, it is only moderate (0.532, 0.615, and
0.621, respectively) between 1990 and 2015. In Africa and in Europe only a weak positive correlation
can be observed (the correlation value for the former is 0.378 and for the latter 0.182). The value of Φ is
under 1 (in every case), which means basically a zoom out, but here we compare not spatial forms but
only points. Based on the methodology of bi-dimensional regression, Θ is the rotation angle. If Θ = 0,
then the XY coordinate system should not be rotated; if it is negative it indicates a clockwise rotation.
This latter can be seen for the Americas and Australia, while for the other continents (Asia, Africa,
Europe, and the world) the relationship shows an anti-clockwise rotation. The SSR (sum of squares
due to regression) is the highest in Asia at over 80% and it has moderate explanation power for the
American and Australian continent (similar to the results of the correlation analysis). However, the
SSR (the square sum of difference explained by regression) is only 7% in Europe, so the explanatory
power of the model is very small and the SSE (sum of squares of errors/residuals that is not explained
by the regression) is extremely high (its value is 93.45%). For Africa the SSR is also low (26.52%), while
globally it is moderate (48.67%).
Resources 2019, 8, 100 15 of 19
Table 4. Two-dimensional regression between the gravitational centres calculated under GDP and total
final energy consumption (1990–2015, 1990–2000, 2001–2015). Source: own calculation.
Continent Time r α1 α2 β1 β2 Φ Θ
SST
(%)
SSR
(%)
SSE
(%)
World
1990–2015 0.532 24.78 9.40 0.63 0.11 0.64 0.17 100 48.67 51.33
1990–2000 0.519 20.87 24.93 0.26 0.09 0.27 0.36 100 46.61 53.39
2001–2015 0.894 24.43 5.06 0.73 0.03 0.73 0.05 100 95.94 4.06
Africa
1990–2015 0.378 14.36 1.36 0.29 0.00 0.29 0.01 100 26.52 73.48
1990–2000 0.428 17.70 −0.42 0.14 0.11 0.18 0.75 100 33.32 66.68
2001–2015 0.144 18.20 3.05 0.08 0.00 0.08 −0.01 100 4.09 95.91
Americas
1990–2015 0.615 −51.93 6.01 0.43 −0.09 0.44 −0.21 100 61.41 38.59
1990–2000 0.522 −68.59 22.80 0.25 −0.01 0.25 −0.02 100 47.00 53.00
2001–2015 0.852 −56.14 9.87 0.38 −0.07 0.38 −0.18 100 92.47 7.53
Asia
1990–2015 0.746 81.07 21.94 0.15 0.02 0.15 0.11 100 80.28 19.72
1990–2000 0.366 93.71 10.03 0.08 0.14 0.15 0.11 100 25.06 74.94
2001–2015 0.841 75.48 26.70 0.19 −0.04 0.19 −0.21 100 91.47 8.53
Australia
1990–2015 0.621 75.67 −8.53 0.47 −0.04 0.47 −0.09 100 62.32 37.68
1990–2000 0.130 143.15 −16.98 −0.01 −0.08 0.08 0.05 100 3.36 96.64
2001–2015 0.782 66.24 −10.54 0.53 −0.02 0.53 −0.03 100 84.90 15.10
Europe
1990–2015 0.182 31.11 36.36 0.32 0.04 0.32 0.12 100 6.55 93.45
1990–2000 0.741 23.12 −26.90 1.60 0.09 1.60 0.06 100 79.66 20.34
2001–2015 0.531 28.62 41.69 0.21 0.00 0.21 −0.01 100 48.44 51.56
The investigation using these maps allows us to conclude that around 2000 a change of direction
took place in the world, which refers to a structural break: until then the economic centre of gravity
moved towards the west and the energy centre of gravity shifted towards the southwest. After
2000 both of them turned towards the east. To test the related hypothesis, the Chow Test is applied.
The dataset is split into two parts. The H1 hypothesis for the test is that the year 2000 serves as a
break point globally, which can be explained by rapid growth in the price of raw materials and by
economic prosperity. The examined time period is split into two main parts: the first is from 1990–2000,
the second contains the years of 2001–2015. To do the Chow test the pooled-OLS model is applied and
we do not suppose the causality direction a priori, so we make our calculations into both directions.
The results are shown in Table 5. Based on the results, we can reject the null hypothesis and we accept
the presence of a structural break in the year 2000.
Table 5. Chow test for structural break. Source: own calculation.
Case I: Dependent Variable:
logGDP
Case II: Dependent Variable:
logTFEC
Test statistic F(2, 4977) = 27.2525 F(2, 4977) = 14.5053
p-value P(F(2, 4977) > 27.2525) = 0.000 P(F(2, 4977) > 14.5053) = 0.000
Note: Null hypothesis: no structural break.
Based on the results of the Chow test we carried out a two-dimensional regression analysis for
both time periods. The results (Table 5) verify that a sharp structural break related to the energy and
economic centres of gravity is found at the turn of the millennium. Globally and in the case of the
Americas, Asia and Australia the correlation between the two centres of gravity became stronger.
In contrast, in the case of Africa and Europe a weakening relationship can be observed (the value
of SSR is 4.09% in Africa and 48.44% in Europe); moreover, the relationship between the point sets
obtained by the gravitational calculations using GDP and TFEC not only becomes weaker but clearly
takes another direction. This is the primary reason for the weak correlation during the whole time
period. But the trend of the angle of shifting (Θ) shows no regularity, so—for the time being—in our
Resources 2019, 8, 100 16 of 19
view the method cannot be applied for the analysis of causality directions between economic growth
and energy use.
4. Discussion
It is predicted that global energy use will continue to grow by around a third or so by 2040
and the industrial energy consumption will account for around half of the increase in it [1]. From a
geographical view, all of the growth is in emerging economies (mainly in China and India), which
account for half of the growth in global energy demand. Significant changes in global energy demand
can occur within a short time that affect not only the international trade of fuels and global cash flows
but also energy security issues of the developed world. This creates new challenges for global energy
systems and governance, with a deepening gap between energy importers and exporters.
Energy security is at the top of the list related to the most urgent energy and foreign policy
priorities (in most countries). We believe that understanding the geographies of energy demand, as well
as the spatial variation in energy consumption, is critical not only in determining the substitutability
of fossil fuel resources but also in energy security issues. This means the uninterrupted physical
availability of the energy sources at affordable prices, and having regard for the environment and the
principles of sustainability. In the long-term, energy security means all of the investments which, if
carried out in time, ensure that supply can meet the demands arising during economic development
(having regard to environmental sustainability). In the short-term, this means the ability of energy
systems to immediately react to any sudden changes, keeping the balance of demand and supply. The
degree of energy security will not be satisfactory if the energy is physically not available or is sold at a
price which most role-players cannot afford. The unpredictability of the future and the dynamics of
international relations make the planning of energy systems more difficult: potential threats must be
identified on the basis of the available (and never complete) information. However, possible responses
and assets can be improved through continuous assessments, and the flexibility of the system can be
enhanced. Political decision-makers play a key role in energy security; as they are the ones who have
information about the whole system, and they are responsible for the synchronization of the activities
of people working in the sector.
5. Conclusions
In our study the centre of gravity method (a member of the family of gravity models) and
bi-dimensional regression were applied to examine the shift in economic and energy centres of gravity
during the period of 1990–2015 in the world and on each of the continents. The basic hypotheses were
partly confirmed and the following statements can be made:
• The economic and energy centre of gravity can be found in the Mediterranean Basin, but a slow
and gradual shift to the east can be observed. Currently it reflects the dominance of the north, but
it marks the decreasing role of the northern hemisphere (position loss) and the greater importance
of the developing (emerging) economies (in the southern hemisphere).
• A strong correlation between the point set obtained by the gravitational calculation using GDP
and the total final energy consumption is observed between 1990 and 2015 only in Asia. Globally,
in the Americas and in Australia and Oceania the strength of the relationship is moderate, while
in Africa and in Europe the relationship is much weaker. These results are confirmed by the SSE
(the square sum of difference not explained by the regression).
• The year 2000 can be interpreted as a turning point (or as a structural change). After that point
the co-movement of the two examined indicators becomes conspicuous. This can be increasingly
perceived in the Americas, in Asia, in Australia, and globally. However, in the case of Africa and
Europe an opposite and weakening correlation can be observed.
Resources 2019, 8, 100 17 of 19
• The application of the gravity model and bi-dimensional regression analysis can be limited to
identify the causality directions among these examined variables; further investigation of this
topic should be conducted.
This study contributes to apply kindly new approaches and methods (such as centre of gravity
method, two-dimensional regression, standard distance) to examine the spatial movements of energy
demand and to give additional information about the new geography of energy. In the future this
investigation should be developed further. Beyond the national level, the consideration of the regional
level is necessary. It would enable us to measure the shifts more accurately and to obtain deeper
knowledge about the causality relationship between energy consumption and economic growth.
Author Contributions: Conceptualization: G.T. and T.S.S.; methodology: G.T. and T.S.S.; software: G.T. and T.S.S.;
validation: G.T. and T.S.S.; formal analysis: G.T. and T.S.S.; investigation: G.T. and T.S.S.; resources: G.T. and
T.S.S.; data curation: G.T. and T.S.S.; writing—original draft preparation: G.T. and T.S.S.; writing—review and
editing: G.T. and T.S.S.; visualization: G.T.; supervision: G.T. and T.S.S.; project administration: T.S.S.; funding
acquisition: T.S.S.
Funding: This research was supported by project no. EFOP-3.6.2-16-2017-00007, titled “Aspects on the development
of intelligent, sustainable and inclusive society: social, technological, innovation networks in employment and
digital economy”. The project has been supported by the European Union, and co-financed by the European
Social Fund and the budget of Hungary.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to
publish the results.
1. IMF. World Economic and Financial Surveys. World Economic Outlook. Database—WEO Groups and
Aggregates Information. 2017. Available online: https://www.imf.org/external/pubs/ft/weo/2017/01/weodata
/groups.htm (accessed on 15 September 2018).
2. BP. Energy Outlook 2018. 2018. Available online: https://www.bp.com/content/dam/bp/en/corporate/pdf/ene
rgy-economics/energy-outlook/bp-energy-outlook-2018 (accessed on 15 September 2018).
3. World Energy Council. World Energy Scenarios 2017. 2017. Available online: https://www.worldenergy.org/
wp-content/uploads/2017/05/LAC-Scenarios_Full-Report_FINAL (accessed on 15 September 2018).
4. Zhang, N.; Lior, N.; Jin, H. The energy situation and its sustainable development strategy in China. Energy
2011, 36, 3639–3649. [CrossRef]
5. World Bank. World Development Indicators. Database. 2018. Available online: http://databank.worldbank.o
rg/data/reports.aspx?source=world-development-indicators (accessed on 25 September 2018).
6. World Bank. Sustainable Energy 4 All. Database. 2018. Available online: https://datacatalog.worldbank.org/
dataset/sustainable-energy-all (accessed on 25 September 2018).
7. Szabó, Z.S. A válság Hatása Kínára, és Peking Gazdaságpolitikai Válaszai. Magyar Tudomány 2010, 4.
Available online: http://www.matud.iif.hu/2010/04/08.html (accessed on 6 May 2018).
8. Zhang, Y.; Zhang, J.; Yang, Z.; Li, J. Analysis of the distribution and evolution of energy supply and demand
centres of gravity in China. Energy Policy 2012, 49, 695–706. [CrossRef]
9. Wang, W.; Zhang, M.; Li, P. Exploring temporal and spatial evolution of global energy production and
consumption. Renew. Sustain. Energy Rev. 2014, 30, 943–949. [CrossRef]
10. Duncan, R. The Olduvai Theory—Energy, Population and Industrial Civilization. Soc. Contract 2005, 16, 6.
11. Stern, D.I.; Cleveland, C.J. Energy and Economic Growth; Rensselaer Working Papers in Economics,
Encyclopedia of Energy; Academic Press: San Diego, CA, USA, 2004; pp. 35–51.
12. Stern, D.I. The role of energy in economic growth. Ann. N. Y. Acad. Sci. 2011, 1219, 26–51. [CrossRef]
13. Murphy, D.J.; Hall, C.A.S. Adjusting the economy to the new energy realities of the second half of the age of
oil. Ecol. Model. 2011, 223, 67–71. [CrossRef]
14. Kovács, F. Energiaigények és a világ szénkészletei. A Miskolci Egyetem Közleménye A sorozat, Bányászat 2007,
71, 63–74. (In Hungarian)
15. Kovács, F. A megújuló energiafajták várható arányai az energiaigények kielégítésében. A Miskolci Egyetem
Közleménye A sorozat, Bányászat 2007, 71, 47–62. (In Hungarian)
https://www.imf.org/external/pubs/ft/weo/2017/01/weodata/groups.htm
https://www.imf.org/external/pubs/ft/weo/2017/01/weodata/groups.htm
https://www.bp.com/content/dam/bp/en/corporate/pdf/energy-economics/energy-outlook/bp-energy-outlook-2018
https://www.bp.com/content/dam/bp/en/corporate/pdf/energy-economics/energy-outlook/bp-energy-outlook-2018
https://www.worldenergy.org/wp-content/uploads/2017/05/LAC-Scenarios_Full-Report_FINAL
https://www.worldenergy.org/wp-content/uploads/2017/05/LAC-Scenarios_Full-Report_FINAL
http://dx.doi.org/10.1016/j.energy.2011.01.035
http://databank.worldbank.org/data/reports.aspx?source=world-development-indicators
http://databank.worldbank.org/data/reports.aspx?source=world-development-indicators
https://datacatalog.worldbank.org/dataset/sustainable-energy-all
https://datacatalog.worldbank.org/dataset/sustainable-energy-all
http://www.matud.iif.hu/2010/04/08.html
http://dx.doi.org/10.1016/j.enpol.2012.07.012
http://dx.doi.org/10.1016/j.rser.2013.11.027
http://dx.doi.org/10.1111/j.1749-6632.2010.05921.x
http://dx.doi.org/10.1016/j.ecolmodel.2011.06.022
Resources 2019, 8, 100 18 of 19
16. Lakatos, I.; Lakatosné Szabó, J. A nem Konvencionális Szénhidrogének Jelentősége a XXI. Században.
2008. Available online: http://www.mernokkapu.hu/fileok/2/Koolaj-Foldgaz (accessed on 8 May 2017).
(In Hungarian).
17. Berndt, E.R.; Wood, D.O. Technology, prices and the derived demand for energy. Rev. Econ. Stat. 1975, 3,
259–268. [CrossRef]
18. Schurr, S.H. Energy Efficiency and Productive Efficiency: Some Thoughts Based on American Experience.
Energy J. 1982, 3, 3–14. [CrossRef]
19. Stern, D.I. Energy Quality. 2009. Available online: http://mpra.ub.uni-muenchen.de/16857/1/MPRA_paper_1
6857 (accessed on 4 December 2018).
20. Ayres, R.U.; Ayres, L.W.; Warr, B. Exergy, power and work in the US economy, 1900–1998. Energy 2003, 28,
219–273. [CrossRef]
21. Ayres, R.U.; Warr, B. Accounting for growth: The role of physical work. Struct. Chang. Econ. Dyn. 2005, 16,
181–209. [CrossRef]
22. Akizu, O.; Urkidi, L.; Bueno, G.; Lago, R.; Barcena, I.; Mantxo, M.; Basurko, I.; Lopez-Guede, J.M. Tracing
the emerging energy transitions in the Global North and the Global South. Int. J. Hydrogen Energy 2017, 42,
18045–18063. [CrossRef]
23. Arto, I.; Capellán-Pérez, I.; Lago, R.; Bueno, G.; Bermejo, R. The energy requirement of a developed world.
Energy Sustain. Dev. 2016, 33, 1–13. [CrossRef]
24. Munkácsy, B. Energiaföldrajz és Energiatervezés; Eötvös Loránd Tudományegyetem, Természettudományi
Kar, Földrajztudományi Központ, Környezet- és Tájföldrajzi Tanszék: Budapest, Hungary, 2018; 135p,
Available online: http://munkacsy.web.elte.hu/energiafoldrajz%20tankonyv (accessed on 15 May 2018).
(In Hungarian)
25. Calvert, K. From “energy geography” to “energy geographies”: Perspectives on a fertile academic borderland.
Prog. Hum. Geogr. 2016, 40, 105–125. [CrossRef]
26. Kraft, J.; Kraft, A. On the relationship between energy and GNP. J. Energy Dev. 1978, 3, 401–403.
27. Sorrell, S.; Dimitropoulos, J. UKERC Review of Evidence for the Rebound Effect; Technical Report 5: Energy,
productivity and economic growth studies; UKERC: London, UK, 2007; 192p. Available online:
http://ukerc.rl.ac.uk/UCAT/PUBLICATIONS/UKERC_Review_of_Evidence_for_the_Rebound_Effect_-_T
echnical_Report_5_Energy,_productivity_and_economic_growth_studies (accessed on 10 April 2019).
28. Jin, S. The effectiveness of energy efficiency improvement in a developing country: Rebound effect of
residential electricity use in South Korea. Energy Policy 2007, 35, 5622–5629. [CrossRef]
29. Tsani, S.Z. Energy consumption and economic growth: A causality analysis for Greece. Energy Econ. 2010,
32, 582–590. [CrossRef]
30. Warr, B.; Ayres, R.; Eisenmenger, N.; Krausmann, F.; Schandl, H. Energy use and economic development: A
comparative analysis of useful work supply in Austria, Japan, the United Kingdom and the US during 100
years of economic growth. Ecol. Econ. 2010, 69, 1904–1917. [CrossRef]
31. Cleveland, C.J.; Kaufmann, R.K.; Stern, D.I. Aggregation and the role of energy in the economy. Ecol. Econ.
2000, 32, 301–317. [CrossRef]
32. Nagy, G. A gravitációs modell felhasználásának lehetőségei a várostérségek lehatárolásában. Területi Stat.
2011, 6, 656–673. (In Hungarian)
33. Kincses, Á.; Tóth, G. Gravitációs modell alkalmazása a térszerkezet vizsgálatára. Területi Statisztika 2012,
52, 479–491. Területi Statisztika 2012, 52, 479–491. (In Hungarian)
34. Quah, D. The global economy’s shifting centre of gravity. Glob. Policy 2011, 2, 3–9. [CrossRef]
35. Klein, L.R. Measurement of a shift in the world’s centre of economic gravity. J. Policy Model. 2009, 31, 489–492.
[CrossRef]
36. Grether, J.; Mathys, N.A. Is the world’s economic centre of gravity already in Asia? Area 2010, 42, 47–50.
[CrossRef]
37. Fesharaki, F. Asia as the centre of gravity of the world energy system. Energy 1996, 21, 999–1003. [CrossRef]
38. UN. World Economic Situation and Prospects; UNCTAD: New York, NY, USA, 2017; Available online: https://ww
w.un.org/development/desa/dpad/wp-content/uploads/sites/45/publication/2017wesp_full_en (accessed
on 15 September 2018).
39. Nemes Nagy, J. Spatial gravity centres of the dynamics and the crisis in Hungary. Hung. Stat. Rev. 2002, 80,
75–85.
http://www.mernokkapu.hu/fileok/2/Koolaj-Foldgaz
http://dx.doi.org/10.2307/1923910
http://dx.doi.org/10.5547/ISSN0195-6574-EJ-Vol3-No3-1
http://mpra.ub.uni-muenchen.de/16857/1/MPRA_paper_16857
http://mpra.ub.uni-muenchen.de/16857/1/MPRA_paper_16857
http://dx.doi.org/10.1016/S0360-5442(02)00089-0
http://dx.doi.org/10.1016/j.strueco.2003.10.003
http://dx.doi.org/10.1016/j.ijhydene.2017.04.297
http://dx.doi.org/10.1016/j.esd.2016.04.001
http://munkacsy.web.elte.hu/energiafoldrajz%20tankonyv
http://dx.doi.org/10.1177/0309132514566343
http://ukerc.rl.ac.uk/UCAT/PUBLICATIONS/UKERC_Review_of_Evidence_for_the_Rebound_Effect_-_Technical_Report_5_Energy,_productivity_and_economic_growth_studies
http://ukerc.rl.ac.uk/UCAT/PUBLICATIONS/UKERC_Review_of_Evidence_for_the_Rebound_Effect_-_Technical_Report_5_Energy,_productivity_and_economic_growth_studies
http://dx.doi.org/10.1016/j.enpol.2007.05.028
http://dx.doi.org/10.1016/j.eneco.2009.09.007
http://dx.doi.org/10.1016/j.ecolecon.2010.03.021
http://dx.doi.org/10.1016/S0921-8009(99)00113-5
http://dx.doi.org/10.1111/j.1758-5899.2010.00066.x
http://dx.doi.org/10.1016/j.jpolmod.2009.05.005
http://dx.doi.org/10.1111/j.1475-4762.2009.00895.x
http://dx.doi.org/10.1016/0360-5442(96)00083-7
https://www.un.org/development/desa/dpad/wp-content/uploads/sites/45/publication/2017wesp_full_en
https://www.un.org/development/desa/dpad/wp-content/uploads/sites/45/publication/2017wesp_full_en
Resources 2019, 8, 100 19 of 19
40. Tóth, G.; Nagy, Z. The world’s economic centre of gravity. Reg. Stat. 2016, 6, 177–180. [CrossRef]
41. Jenness, J. Weighted Mean of Points v. 1.2c. 2004. Available online: http://www.jennessent.com/arcview/wei
ghted_mean.html (accessed on 23 November 2018).
42. Durocher, S.; Leblanc, A.; Skala, M. The projection median as a weighted average. J. Comput. Geom. 2017, 8,
78–104. [CrossRef]
43. Tobler, W.R. Map Transformations of Geographic Space. Ph.D. Thesis, University of Washington, Seattle,
WA, USA, 1961.
44. Tobler, W.R. Computation of the Correspondence of Geographical Patterns. Pap. Reg. Sci. Assoc. 1965, 15,
131–139. [CrossRef]
45. Tobler, W.R. A Computer Model Simulating Urban Growth in the Detroit Region. Econ. Geogr. 1970, 46,
234–240. [CrossRef]
46. Tobler, W.R. Comparisons of Plane Forms. Geogr. Anal. 1978, 10, 154–162. [CrossRef]
47. Tobler, W.R. Bidimensional Regression. Geogr. Anal. 1994, 26, 187–212. [CrossRef]
48. Tobler, W.R. On the First Law of Geography: A Reply. Ann. Assoc. Am. Geogr. 2004, 94, 304–310. [CrossRef]
49. Friedman, A.; Kohler, B. Bidimensional Regression: Assessing the Configural Similarity and Accuracy of
Cognitive Maps and Other Two-Dimensional Data Sets. Psychol. Methods 2003, 8, 468–491. [CrossRef]
50. Dusek, T. Kétdimenziós regresszió a területi kutatásokban. Területi Stat. 2011, 51, 11–22. (In Hungarian)
51. ESRI Developer Network. Standard Distance (Spatial Statistics). 2018. Available online: http://edndoc.esri.com/ar
cobjects/9.2/net/shared/geoprocessing/spatial_statistics_tools/standard_distance_spatial_statistics_.htm (accessed
on 2 April 2018).
52. Ramanathan, R. Bevezetés az Ökonometriába Alkalmazásokkal; Panem Kiadó: Budapest, Hungary, 2003.
(In Hungarian)
53. Szlávik, J.; Sebestyén Szép, T. Delinking of energy consumption and economic growth in the Visegrad group.
Geogr. Tech. 2017, 12, 139–149. [CrossRef]
54. Weiner, C. Az Állam Szerepe az Orosz Gazdaságban; MTA Közgazdaság- és Regionális Tudományi Kutatóközpont
Világgazdasági Intézet: Budapest, Hungary, 2016; Available online: http://real.mtak.hu/44950/1/MT_119_We
iner (accessed on 5 May 2018).
55. Moshiri, S. The effects of the energy price reform on households consumption in Iran. Energy Policy 2015, 79,
177–188. [CrossRef]
56. Brzezinski, Z. The Grand Chessboard: American Primacy and Its Geostrategic Imperatives, 2nd ed.; Basic Books:
New York, NY, USA, 2016; 316p.
57. IEA. Key World Energy Statistics 2018. 2018. Available online: https://www.iea.org/publications/freepublica
tions/publication/KeyWorld2017 (accessed on 5 May 2018).
58. Simon, G., Jr. Reform és növekedés Kínában. Közgazdasági Szle 2001, XLVIII, 673–692. (In Hungarian)
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
http://dx.doi.org/10.15196/RS06210
http://www.jennessent.com/arcview/weighted_mean.html
http://www.jennessent.com/arcview/weighted_mean.html
http://dx.doi.org/10.20382/jocg.v8i1a5
http://dx.doi.org/10.1007/BF01947869
http://dx.doi.org/10.2307/143141
http://dx.doi.org/10.1111/j.1538-4632.1978.tb00004.x
http://dx.doi.org/10.1111/j.1538-4632.1994.tb00320.x
http://dx.doi.org/10.1111/j.1467-8306.2004.09402009.x
http://dx.doi.org/10.1037/1082-989X.8.4.468
http://edndoc.esri.com/arcobjects/9.2/net/shared/geoprocessing/spatial_statistics_tools/standard_distance_spatial_statistics_.htm
http://edndoc.esri.com/arcobjects/9.2/net/shared/geoprocessing/spatial_statistics_tools/standard_distance_spatial_statistics_.htm
http://dx.doi.org/10.21163/GT_2017.122.12
http://real.mtak.hu/44950/1/MT_119_Weiner
http://real.mtak.hu/44950/1/MT_119_Weiner
http://dx.doi.org/10.1016/j.enpol.2015.01.012
https://www.iea.org/publications/freepublications/publication/KeyWorld2017
https://www.iea.org/publications/freepublications/publication/KeyWorld2017
http://creativecommons.org/
http://creativecommons.org/licenses/by/4.0/.
Global Trends in Energy Use and Economic Growth—A Dynamic Shift
Energy Consumption and Economic Growth
Fields of Force in the World, or the Centre of Gravity Method
Materials and Methods
The Gravity Model with Special Regard to the Centre of Gravity Method
Application of Two-Dimensional Regression
Examination of Shift in Centres of Gravity (Standard Distance)
Chow Test for Structural Changes
Results
Movement of Gravity Centres
Results of Bi-Dimensional Regression
Conclusions
References
Journalof Business & Economic Studies, Vol. 17, No. 2, Fall 2011
20
Financial Development and Economic Growth:
Evidence From Jordan Economy
Zakia A. Mishal, Yarmouk University, Irbid, Jordan
Abstract
Whether financial market development causes, or is caused by, economic growth is an unsettled
issue. This subject has greater relevance for Jordan, given the tremendous improvements in
banking systems and stock market activities in the last decade. Causality tests were conducted by
co-integration testing. The results provide evidence of a stable long-run equilibrium relationship
between financial markets’ development and economic development. The causality test results
showed a bi-directional causality between economic growth and banking system developments.
Moreover, the results demonstrate that economic growth leads to the growth of stock market not
vice versa.
Keywords: Economic Growth, Banking Sector Development, Stock Market Development
JEL Classifications: E 44, G 10, O 16, O 50
Introduction
Several attempts have been made to investigate the role of the financial sector in
economic growth. The issue has been of great interest, generating a considerable amount of
debate among economists for many years. In addition, the growing importance of stock markets
recently opened a new avenue of research into the relationship between financial development
and economic growth, which focuses on the effects of stock market development over and above
the effects of the banking system (Levine & Zervos, 1998). Undoubtedly, stock markets are
expected to increase economic growth by increasing the liquidity of financial assets, making
global and domestic risk diversification possible, promoting wiser investment decisions, and
solving institutional financial problems by increasing shareholders’ interest/value. Thus, the
stock market is important for investors and policy makers because of the benefits it provides to
the economy, and because it is often cited as a barometer of business direction.
Given the controversy that surrounds this issue, it seemed relevant to further research in
order to identify the mechanism through which financial markets influence economic growth.
Various channels have been suggested: mobilizing domestic savings, allocating capital
proficiency, and diversifying risks (Caporale, Howells, & Soliman, 2005).
Stock markets and banks clearly are substitute sources for corporate finance, because,
when a firm issues new equity, its borrowing needs from the banking system decline; it is then
possible that stock market development may hamper economic growth if it happens at the
expense of banking system development (Fry, 1997; Mayer, 1988;). On the other hand, increased
stock market capitalization may be accompanied by an increase in the volume of bank business.
If not providing an increase in new lending, financial intermediaries may provide complementary
services to issuers of new equity such as underwriting. Thus, at the aggregate level, the
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
21
development of the stock market is likely to go hand in hand with the development of the
banking system (Arestis, Demetriades, & Luintel, 2001).
The underlying objective of the present study was to explore the long-run relationship
between financial markets’ development and the real sector growth in Jordan. If there has been
a
long-run relationship, what is the exact direction of the relationship? This issue in the Middle
East has not been analyzed thoroughly despite tremendous growth in recent years. The Middle
East countries have gone through the process of upgrading their stock markets and developing
their regulations. Jordan is one of the most open to foreign investors. The study conducted by
Sedik and Petri (2006) indicated that the Amman Stock Exchange (ASE) compares favorably
with many other markets in the region in terms of investment restrictions, regulatory
environment, and transparency. These characteristics make the ASE a good representative of
other emerging stock markets. In this regard, the results reported in this study may be relevant
for other emerging markets of similar characteristics and stage of development. In addition, there
is an advantage in looking at a specific country when including the stock market development in
the empirical model, since using a cross-country study might face the problems of a huge
reduced sample size.
Overview of the Literature
An extensive volume of literature and research work has emerged attempting to highlight
the role of the financial market in growth. Levine (1997, 2002) provided a comprehensive survey
about the subject. In his theoretical study, Singh (1997) examined the importance of stock market
development for the economic growth of developing countries. Some studies, such as King and
Levine (1993a, b); Levine and Zervos (1996, 1998); and Liang and Reichert (2007) indicated
that in most cases stock market improvements promote growth dramatically, especially in
developed countries. Levine and Zervos (1996) also made a distinction between the financial
services offered by credit and equity markets and suggested that they may complement each
other. (Other studies found the same results using the Granger Causality approach and applying
data for a single country (Dep & Mukherjee, 2008; Guru-Gharan, Rahman, & Parayita, 2009;
Shahbaz, Ahmed, & Ali, 2008;; Somoye, Akintoye, & Oseni, 2009). Alkhathlan (2009) for
Saudi Arabia, and Mazur and Alexander (2001) for New Zealand found that the banking system
had only a strong, positive effect on economic growth. Agrawalla and Tureja (2007, 2008) for
India found that stock market development caused long-term growth in output. For Zimbabwe,
Oyama (1997) found that money supply (M2) and market interest rate explained changes in stock
prices.
Other studies used data from several countries and found that equity markets play a
significant role in the banking sector’s economic growth (Arestis, et al, 2001; Kassimatis &
Spyrou, 2001; Liang & Reichert, 2007). The studies by Caporale et al. (2005), and Adjasi and
Biekpe (2009) indicated that investment productivity is the channel through which stock markets
enhance growth. Thus, any theoretical indication of a link between improvements in financial
markets and faster economic growth remains ambiguous.
The current study attempted to examine whether a relationship exists between financial
market development and output growth in the case of developing a small open economy such as
Jordan.
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
22
Research Methodology
The empirical analysis in the present study was based on multivariate Granger causality
tests within an error correction framework. The first step in the analysis was to test the
macroeconomic series for stationarity through unit root tests. The stationarity of the series was
investigated by employing the unit root tests developed by Dickey and Fuller (1979, 1981), and
Phillips and Perron (1988). The joint use of both tests attempted to overcome the common
criticism that unit root tests have limited power in finite samples to reject the null hypothesis of
nonstationarity. Thus, the Augmented Dickey-Fuller (ADF) t-tests and Phillips and Perron
(1988) Z(tα) tests are used to test for the presence of a unit root for the individual time series and
their first differences.
The next step for co-integration. The advantage of carrying out co-integration testing was
to provide was testing evidence of the existence of a stable long-run equilibrium relationship
between the macroeconomic variables and financial market variables, which was interesting
from a theoretical perspective. In addition, the advantage of carrying out co-integration testing
was that the causality tests were preceded by co-integration testing because the existence of co-
integration has implications on the way in which causality testing is carried out (Granger, 1988;
Toda & Phillips, 1993). The present study used Johansen Cointegration Tests and Vector Error
Correction Model (VECM) to avoid potential misspecification biases that might result from the
use of a more conventional VAR modeling technique; if the variables used in the VAR model
were cointegrated, then the model may have been misspecified because it excluded an additional
channel of influence resulting from a long-term equilibrium relationship among these variables
(Engle & Granger, 1987). Finally, the causality tests among the co-integrated variables were
undertaken.
Data and Measurement
The current study focused on Jordan’s economy spanning a period of 31 years. The
model was estimated using yearly data for the period 1978 to 2009. The reason for choosing an
annual data was the unavailability of quarterly or monthly data for GDP for a sequent long time
frame. The variables that were used are as follows:
1. Economic development: measured by growth rate of nominal GDP (Y).
2. Stock market development: measured by three proxies: The first proxy was market-
capitalization ratio (size proxy) defined by the ratio of market capitalization to GDP
(MCY). The second proxy was value-traded ratio (activity proxy) defined by the ratio of
trading volume to GDP (VTY), (this ratio did not directly measure the cost of buying and
selling shares; it averaged the cost of equity transactions as a share of national output
over a long time frame. There would be a less trading if it is costly to buy and sell
equities. However, this ratio was particularly suitable to capture the stock market
liquidity). The third proxy was stock prices return defined by the growth rate of share
price index weighted by market capitalization (SR). Stock prices reflect the marginal
productivity of capital; thus increases in stock prices would result in an increase in the
marginal productivity of capital, which would be linked directly to an increase in
investment activities. Fama (1981) and Barro (1990) also explained changes in stock
prices and considered these changes as an important component of variation in the market
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
23
value of capital. Thus, changes in stock prices would cause changes in the market value
of capital, in addition to changes in output.
3. Banking system development was measured by the ratio of domestic bank credit to
nominal GDP (BY). In addition, the market lending interest rate (I) was used, since it
affects the borrowing needs from the banking system. These credit-based indicators were
more likely to exhibit a stable long-run relationship with output than deposit-based ones
(Arestis, 2001).
The data on market capitalization, trading value, and share price index were collected
from the Amman Stock Exchange (ASE), while that of GDP, bank credit, market lending
interest rate are collected from several issues of the central bank of Jordan (CBJ).
Table 1 shows the variables definitions and abbreviations, and Figure 1) represents the
variables in the study during the study period.
Table 1. Variable Definitions
Definition Calculation Abbreviation
Annual growth rate of
nominal GDP log (Yt) – log (Yt-1) Y
Market capitalization ratio1
(first difference) mcyt – mcyt-1 MCY
Value traded ratio2 (first
difference) vtyt – vtyt-1 VTY
Annual growth rate of share
price index3 log (spi)4t – log(spi)t-1 SR
Bank credit ratio5 (first
difference) byt – byt-1 BY
Market lending interest rate
(first difference) 6mlit – mlit-1 I
Notes: 1Defined by the ratio of market capitalization to GDP. 2Defined by the ratio of trading
volume to GDP. 3Stock prices return defined by the growth rate of share price index weighted by
market capitalization. 4spi is the stock price index weighted by market capitalization. 5Defined by
the ratio of domestic bank credit to nominal GDP.
Empirical Results
The first step of the empirical part of this study was subjecting the data to diagnostic
tests; such as unit root tests and Johansen co-integration test. The main purpose of these tests was
to search for the model that best fits the data set. The empirical results reported for Jordan were
based, as mentioned before, on annual observations for the period 1978 to 2009. All the data
were expressed in logarithms.
Response Analysis
For further evidence on the relationships between the growth rate of GDP and the
financial markets, an impulse response function was employed on the multivariate VECM to
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
24
trace the effect of a one-time shock to one of the innovations on current and future values of the
endogenous variables. The impulse function is designed to identify the dynamic effects of an
exogenous and temporary shock in one variable, say bank credit ratio, on another variable, say
the GDP growth. The short-term response is obtained from the first step of the impulse response
analysis. It measures the immediate impact on the growth of GDP when an exogenous shock in
bank credit ratio occurs. On the other hand, the long-run response is obtained by allowing all
variables to respond over time to a shock in bank credits in the first step. To obtain the long-run
effects, the number of steps for the impulse response analysis was set at 12 periods. Figure 2
shows responses of GDP growth, market capitalization ratio, value traded ratio, and stock returns
to Cholesky One S.D. change in bank credit ratio innovation. As expected, following an
expansionary bank credit, the GDP growth decreased as shown in Figure 2. However, this effect
was not significant quantitatively, since the output growth converged gradually back toward its
long-run equilibrium level.
Table 2 reports the augmented Dickey-Fuller and Phillips-Perron test statistics for the
levels and first difference of all nominal variables in this study. According to the results shown
in Table 2, the ADF and PP tests for unit roots suggested that the variables considered in this
study are all non-stationary in their levels but stationary in first difference.
The second step is testing for co-integration among the variables using the Johansen’s
(1988) methodology, that is, the trace (trace) and the maximum Eigen value (max) statistics. In
general, if two series are found to be co-integrated, then the inference of a long-run equilibrium
relation between them is sufficiently robust, except for a stationary disturbance with finite
variance. Moreover, in the presence of co-integration, the long-run elasticity of GDP, with
respect to the other variables (or vice versa), can be estimated without specifying any dynamics
and without an a priori determination of causality, since both variables are endogenous and can
be treated symmetrically (Ahmad, 2001).
Since the results derived from these tests were sensitive to the selection of the lag length,
two criteria for lag order selection were used: AIC (Akaik Information Criterion), and SC
(Schwarz -Information Criterion). The test results suggested using a lag length (which has white
noise residuals) of two lags. Subsequent analysis, therefore, proceeded with the use of VAR with
lag lengths k = 2. Given that there were six variables in the model (n = 6), there could be a
maximum of five co-integrating vectors; thus, r would be equal to 0,1, 2, 3, 4, or 5.
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
25
.00
.05
.10
.15
.20
.25
1980 1985 1990 1995 2000 2005
annual growth of (GDP)
–
1
.0
–
0.5
0.0
0.5
1.0
1.5
1980 1985 1990 1995 2000 2005
annual change of (MCY)
–
0.8
–
0.4
0.0
0.4
0.8
1
.2
1.6
1980 1985 1990 1995 2000 2005
annual change of (VTY)
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1980 1985 1990 1995 2000 2005
annual growth of (stock price index)
–
.12
–
.08
–
.04
.00
.04
.08
.12
1980 1985 1990 1995 2000 2005
annual change of (BY)
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
1980 1985 1990 1995 2000 2005
annual change of interest rate (I)
Figure 1. The Variables Used in the Analysis During the Period 1978 to 2009
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
26
Table 2. The ADF and PP Tests for Unit Roots
Test
statistic GDP BY MCY VTY I Spi Comment
Levela. Variable t = 1978-2009
ADF 3.55(1) -1.53(1) -1.22(1) -1.52(1) -2.39(1) -0.76(1) Not I(0)
PP 5.10(3) -1.84(3) -1.69(3) -1.99(3) -1.65(3) -092(3) Not I(0)
First differencea
ADF 2.54(1)b –
4.03(1)*
–
2.61(1)***
–
4.44(1)*
–
3.17(1)**
–
3.76(1)*
I(1)
PP –
3.28(3)**
– 4.24(3)
*
– 6.71(3)* – 6.10(3)
*
–
2.95(3)***
–
5.57(3)*
I(1)
a with intercept and no trend.
b non-stationary at the first difference.
***,**,* denote significance at the 1%, 5%, and 10% levels respectively.
Note. ADF stands for Augmented Dickey-Fuller; PP for Phillips-Perrone. Numbers in brackets are
number of lags used in the ADF test in order to remove serial correlation in the residuals, these lag
lengths are chosen based on Akaike’s Information Criterion (AIC) and Schwartz Bayesian Criterion
(BIC). The truncated lag for PP tests was obtained based on a Newey-West adjustment with lag three for
the sample period 1978-2009.
Table 3. Tests for Co-integration Using the Johansen Procedure
Model: Y = f(BC, SMC, REM, SPI) period sample 1978-2009
p=2a
Test statistics
hypothesis
r=0b
r1
r 2
r 3 r 4
Trace test 209.4** 121.5** 57.5**
max test 87.9** 64.1** 28.5*
a The lag length p was chosen based on AIC (Akaik information criterion), and SC (Schwarz information
criterion).
b r is the number of co-integrating vectors. The Trace test indicate 3 co-integration equations at both 1%
and 5% , while Max eigen value test indicate 3 and 2 co-integration equations at 5% and 1% levels
respectively.
*, ** Indicates statistical significance at 1% and 5% critical levels.
Critical values for choosing the number of co-integrating vectors are taken under the assumption that
there is no deterministic trend in the data.
Results of co-integration rank tests for the model are presented in Table 3. The value of
the trace test (trace ) indicated that the null hypothesis of three (r 3 ) co-integrating vectors can
be rejected at the 1% and 5% levels. That is, it suggested the presence of three co-integrating
vectors between annual growth of GDP (Y), annual change of market capitalization ratio (MCY),
annual change of value traded ratio (VTY), annual growth rate of stock returns (SR), annual
change of bank credit ratio (BY), and annual change in market lending interest rate (I) (see Table
3). At the same time, the Max-eigen value test indicated the existence of 3 co-integrating
equations at both 1% and 5% levels, and 2 co-integrating equations at 1% level. Consequently,
Jordanian growth rate of output, market capitalization ratio, bank credit ratio, stock returns,
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
27
value-traded ratio, and market lending interest rates were co-integrated. The estimated
normalized coefficients of this co-integrating relationship (the s) were significantly different
from zero for all the variables.
These results suggested the existence of a long-run relationship between Y and the other
variables in the study over the entire period 1978 to 2009.
Estimated Error Correction Models
The short-run dynamics, or the direction of causality between the variables in the co-
integration equation, were examined by estimating an error correction model. Given the presence
of co-integrating relationship, the Engle and Granger (1987) error correction specification could
be used to test for Granger causality. The error correction representation for our model for six
variable cases is written as follows (we ignore the other five equations to save space):
0tY
117
65432
1
1 ][
tk
ktkktkktkttt
m
EC
SRIBYVTYMCYY
Where Y, the annual growth rate of GDP; MCT, market capitalization ratio; VTY, value traded
ratio; BY, bank credit ratio; I, market lending interest rate; and SR, the annual growth rate of
share price index; ECt-1 is the vector error correction for the model, and Ɛ1 is uncorrelated
disturbances, and m is the lag length.
In order to make valid inferences on causality, all the variables must be stationary. Thus,
the annual growth of GDP (Y) and stock return (SR), and the first differences of the variables
(MCYt, VTYt, BYt, It), and the residuals (ECs) obtained from the co-integrating vector were
included in the Granger causality test structure. The above structure focused on the short-run
dynamics among Y, MCY, VTY, BY, I, and SR, and at the same time the long-run information,
which was contained in the error correction vector (EC). For each variable in the system, at least
one channel of Granger causality was active: either in the short-run through the joint tests of
lagged-differences or in a statistically significant EC.
The coefficient of EC contained information about whether the past values of variables
affected the current values of the variable under study. The size and statistical significance of the
coefficient on the error correction term in each error correction model measured the tendencies
of each variable to return to equilibrium. A significant coefficient implies that past equilibrium
errors play a role in determining the correct outcomes. For example, if α7 in the above equation
were significant, it could be concluded that Y responded to disequilibria in its relationship with
the independent variables. The short-run dynamics were captured through the individual
coefficients of the difference terms.
Co-integration tests suggested a long-run relationship, but they do not indicate the
direction of this relationship. The significance of the error correction coefficient was determined,
by the t-ratio given below the coefficient for ECt-1. In each specification, the magnitude of the
error correction coefficient indicated the speed of adjustment of any disequilibrium toward a
long-run equilibrium state. A statistically significant EC coefficient implied that past equilibrium
errors played a role in determining current outcomes. The short-run dynamics were captured by
the individual coefficients of the differenced terms. Even if the coefficients of the lagged
changes in the dependent variables were not statistically significant, Granger causality could still
exist as long as the coefficient of EC were statistically different from zero (Choudry, 1995).
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
28
Table 4. Regression Results Based on Johansen Error Correction Procedure
Independent
variables
GDP
Development
Equation
Eq.1
Stock Market
Development Equations
Eq.2 Eq.3 Eq.4
Bank Market
Development
Equations
Eq.5 Eq.6
ΔY ΔMCY ΔVTY ΔSR ΔBY Δi
aECt-1 -0.014
(-0.68)
-0.162
(-3.4 )*
-0.05
(-1.24 )
-0.54
(-2.66 )**
-0.037
(-2.96)*
-0.041
(-0.17)
ΔYt-1 -0.56
(-1.73)***
-3.40
(-4.6 )*
-2.00
(-3.37 )*
-6.85
(-2.19)**
-0.026
(-0.13)
-0.762
(-0.20)
ΔYt-2 0.525
(1.35)
-3.45
(-3.88 )*
-1.82
(-2.55 )**
-6.46
(-1.72)***
-0.733
(-3.13)*
-2.24
(-0.49)
ΔMCYt-1 -0.127
(-0.33 )
-3.10
(-3.53 )*
-0.844
( -1.20)
-5.89
(-1.59)
-0.612
(-2.65)**
-3.30
(-0.73)
ΔMCYt-2 -0.134
(-0.60 )
-0.87
(-1.71)***
-0.534
(-1.32 )
-0.96
(-0.45)
-0.23
(-1.72)***
-3.48
(-1.34)
ΔVTYt-1 0.125
(1.10 )
-0.704
(-2.7)**
-1.14
(-5.40 )*
-0.54
(-0.48)
0.024
(0.34)
0.64
(0.47)
ΔVTYt-2 0.125
(1.12 )
-0.25
(-0.96)
-0.365
(-1.78)***
-1.19
(-1.10)
-0.07
(-1.00)
1.15
(0.87)
ΔSRt-1 0.088
(0.81 )
1.25
(5.10)*
0.78
(3.92)*
-1.72
(-1.65)***
0.19
(2.86)**
1.10
(0.85)
ΔSRt-2 -0.026
(-0.28 )
0.80
(3.73)*
0.623
(3.65)*
0.37
(0.41)
0.145
(2.59)**
1.41
(1.284)
ΔBYt-1 -0.212
(-0.30 )
0.62
(0.39)
-0.130
(-0.10)
2.86
(0.42)
1.032
(2.45)**
7.21
(0.88)
ΔBYt-2 0.94
(2.31 )**
-3.47
(-3.71)*
-0.372
(-0.50)
-7.45
(-1.90)***
-0.55
(2.23)**
1.11
(0.23)
ΔIt-1 -0.01
(-0.39 )
0.094
(1.60)
-0.055
(-1.17 )
0.32
(1.27)
0.016
(1.04)
0.17
(0.56)
ΔIt-2 0.046
(2.12 )**
-0.105
(-2.10)**
-0.040
( -0.94)
-0.037
(-0.18)
-0.013
(-0.98)
-0.17
(-0.65)
Constant -0.005
(-0.41 )
-0.08
(-2.70)**
-0.030
( -1.40)
-0.170
(-1.50)
-0.0134
(-1.92)***
-0.025
(-0.18)
Adjusted R2 0.56 0.97 0.97 0.56 0.68 0.18
D.W Test 1.89 1.75 2.12 1.86 1.75 2.17
F-Testb 3.85 63.58 64.98 3.59 5.34 1.45
aECt-1 is the one period lagged error correction term from the co-integrating equation. Numbers in
parentheses below ECt-1 are t-statistics for H0: αi = 0.
b The F-statistics tests the joint significance of lagged values of the independent variables. The values in
parenthesis are the t-test. * , **, and *** denotes significance at 1% , 5% , and 10% significance level
respectively.
The results in Table 4 clearly show significant error correction terms for the variables,
namely market capitalization ratio (MCY), stock price return (SP), and bank credit ratio (BY)
only. That is, the variables market capitalization ratio; stock price return, and bank credit ratio
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
29
adjusted to disequilibrium from the long-run relationship, while GDP growth, value of shares
traded ratio, and market lending interest rate did not significantly respond to deviations from the
long-run relationship. That is because the lagged coefficients on the independent variables in the
error correction model represented short-run Granger causality, while the coefficient on the error
correction (EC) term in the error correction model reflected long-run Granger causality.
Causality Tests Results
The results of the estimated multivariate VECM (vector error correction model) are
presented in Table 5. The Causality test results could be summarized between output growth,
banking system development, and stock market development in a more candid manner as
presented in Table 6. It is clear from Table 6 that the banking system development is a vehicle of
economic growth, through the bi-directional causality, which runs both sides between bank credit
ratio (BY) and GDP growth (Y), and a uni-directional causality, which runs from lending interest
rate to GDP growth. Beside that, the banking system development affects the stock market
development. A bi-directional causality runs from BY to MCY (market capitalization ratio), and
a uni-directional causality runs from lending interest rate to MCY. On the other hand, the study
strongly contradicted the assumption that Jordanian’s stock market is a vehicle of economic
growth. The results showed that the only effect of the stock market on the economy was through
bank credits. The results showed the existence of a bi-directional causality runs between bank
credit ratio (BY) and market capitalization ratio (MCY). As mentioned before, just the variables
market capitalization ratio (MCY), stock price return (SR), and bank credit ratio (BY) adjusted to
disequilibrium from the long-run relationship, which means the causality runs both ways for
long-run and short-run relationships between stock market development (MCY and SR), and
banking system development (through BY and I). In general, the results showed that stock
market did not Granger cause the GDP growth in Jordan, while the banking system development
did Granger cause the GDP growth. This result implies that the revival of stock market could not
be taken as a leading indicator of the revival of the economy in Jordan. That is, stock market
development (proxied by market capitalization ratio [size proxy], value traded ratio [activity
proxy], and stock returns) cannot be considered as a barometer of business direction, or it cannot
be relied upon to measure changes in the general economic activities in Jordan. Alkhathlan
(2009) reported similar results for Saudi Arabian economy. The study results were also
consistent with Agrawalla and Tuteja (2007, 2008) for India. Kassimatis and Spyrou, (2001), in
their studies for five emerging countries, indicated that equity markets have a role to play only in
liberalized economies.
Response Analysis
For further evidence on the relationships between the growth rate of GDP and the
financial markets, an impulse response function was employed on the multivariate VECM to
trace the effect of a one-time shock to one of the innovations on current and future values of the
endogenous variables. The impulse function is designed to identify the dynamic effects of an
exogenous and temporary shock in one variable, say bank credit ratio, on another variable, say
the GDP growth. The short-term response is obtained from the first step of the impulse response
analysis. It measures the immediate impact on the growth of GDP when an exogenous shock in
bank credit ratio occurs. On the other hand, the long-run response is obtained by allowing all
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
30
variables to respond over time to a shock in bank credits in the first step. To obtain the long-run
effects, the number of steps for the impulse response analysis was set at 12 periods. Figure 2
shows responses of GDP growth, market capitalization ratio, value traded ratio, and stock returns
to Cholesky One S.D. change in bank credit ratio innovation. As expected, following an
expansionary bank credit, the GDP growth decreased as shown in Figure 2. However, this effect
was not significant quantitatively, since the output growth converged gradually back toward its
long-run equilibrium level.
Table 5. VEC Pairwise Granger Causalitya/Block Exogeneity Wald Test
Dependent variable: Δ GDP growth
Variable Chi-sq df Probability
Δ Bank credit ratio (BY) 5.53 2 0.062
Δ Lending interest rate (I) 4.66 2 0.097
Dependent variable: Δ MCY (Market Capitalization Ratio)
Δ GDP growth 23.67 2 0.000
Δ value traded ratio (VTY) 7.90 2 0.019
Δ Stocks return (SR) 26.35 2 0.000
Δ bank credit ratio (BY) 14.5 2 0.000
Δ lending interest rate ( I) 7.10 2 0.029
Dependent variable: Δ VTY (Value Traded Ratio)
Δ GDP growth 12.05 2 0.002
Δ Stocks return (SR) 15.82 2 0.000
Dependent variable: Stocks Return (SR)
Δ GDP growth 5.18 2 0.074
Δ market capitalization
ratio (MCY)
8.45 2 0.014
Dependent variable: Bank Credit ratio (BY)
Δ GDP growth 13.40 2 0.001
Δ market capitalization
ratio (MCY)
9.85 2 0.007
Δ Stocks return (SR) 8.31 2 0.015
Note. aThe table shows just the relations which has a significant statistics for the independent variables in
the model.
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
31
Table 6. Summary of the Causality Tests between Output Growth, Banking System Growth and
Stock Market Growth
Output development
Banking system
development Stock market development
Y ↔ BYa BY ↔ Ya SR ↔ MCYa
Y → MCYb BY ↔ MCYa SR → VTY
Y → VTY I → MCY VTY → MCY
Y → SR I → Y SR → BY
MCY ↔ BY
a
a ↔ represent bi-directional causality.
b → represent uni- directional causality
(Regression Results Based on Johansen Error Correction Procedure)
Figure 2. Response of (1) GDP Growth, (2) Market Capitalization, (3) Value Traded and
Stock Returns to Cholsky One S. D. Change in Bank Credits Innovations.
Figure 3 shows responses of bank credit ratio, market capitalization ratio, value traded
ratio, and stock market returns to Cholesky One S.D. change in GDP growth innovation.
-.05
-.04
-.03
-.02
-.01
.00
1 2 3 4 5 6 7 8 9 10 11 12
R e sp ons e of GDP g r owth
-.4
-.3
-.2
–
.1
.0
.1
1 2 3 4 5 6 7 8 9 10 11 12
Response of Market capitaliza tion r at io ( MC Y)
-.25
-.20
-.15
-.10
-.05
.00
.05
1 2 3 4 5 6 7 8 9 10 11 12
Respons e of Va lue tr aded r a tio (VTY)
-.6
-.5
-.4
-.3
-.2
-.1
.0
.1
.2
1 2 3 4 5 6 7 8 9 10 11 12
Response of Stock ret ur ns ( SR)
FI GU R E 2: R esp o nse o f : ( 1) GDP growth, (2) Market capitalization, (3) Value t r ad ed , a n
( 4) S to ck ret u rn s t o Cholesky one S.D change in Bank credits innovat io n
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
32
Following a shock in GDP growth, bank credit jumped in period 1, as shown in Figure 3(1), and
converged gradually back at period 12 towards its long-run equilibrium level. The stock returns
growth (SR) also jumped and reached levels higher than its long-run norm, then converged
gradually back towards its long-run equilibrium level.
-.024
-.020
-.016
-.012
-.008
–
.004
.000
.004
1 2 3 4 5 6 7 8 9 10 11 12
(1) Response of D(BY) to Cholesky
One S.D. GDP growth Innovation
-.3
-.2
-.1
.0
.1
.2
1 2 3 4 5 6 7 8 9 10 11 12
(2) Response of D(MCY) to Cholesky
One S.D. GDP growth Innovation
-.25
-.20
-.15
-.10
-.05
.00
.05
.10
1 2 3 4 5 6 7 8 9 10 11 12
(3) Response of D(VTY) to Cholesky
One S.D. GDP growth Innovation
-.25
-.20
-.15
-.10
-.05
.00
.05
.10
.15
1 2 3 4 5 6 7 8 9 10 11 12
(4) Response of SR to Cholesky
One S.D. GDP growth Innovation
(Regression Results Based on Johansen Error Correction Procedure)
Figure 3: Responses of: (1) Bank Credit Ratio, (2) Market Capitalization Ratio, (3) Value Traded
Ratio, and (4) Stock Market Returns to Cholesky One S.D. GDP Growth Innovation
Conclusions and Policy Implications
The main contribution of this study was to provide empirical evidence for the relationship
between the economic development and financial sector developments. We examined the causal
relationships between the GDP growth, banking sector development, and stock market
development in a multivariate vector error correction model. The findings provided strong
evidence of a stable long-run relationship between the banking sector and economic growth, and
between the banking sector and the stock market. The study reported a bi-directional causality
between banking sector development and economic growth in the long run, and a bi-directional
causality between the banking sector and stock market. The causality runs from GDP growth to
the stock market and not in the opposite direction, implied that the health of the stock market was
not reflective of an improvement in the health of the economy in Jordan. This finding had many
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
33
implications for the kind of transactions in the Jordanian stock market in the recent years. The
findings of this study suggested that the policies relating to the stock market should be directed
toward the creation of transparent and guiding investors to take a long-term view rather than
serving as a caterer to satisfy the needs of speculators.
References
Adjasi, C. K. D., & Biekpe, N. (2009). Do stock markets matter in investment growth in Africa?
Journal of Developing Areas, 43(1), 109-120.
Agrawalla, R. K., & Tuteja, S. K. (2007). Causality between stock market development and
economic growth. Journal of Management Research, 7(3), 158-168.
Agrawalla, R. K., & Tuteja, S. K. (2008). Share prices and macroeconomic variables in India.
Journal of Management Research, 8(3), 136-146.
Ahmad, J. (2001). Causality between exports and economic growth: What do the econometric
studies tell us? Pacific Economic Review, 6(1), 147-167.
Alkhathlan, K. A. (2009). The role of the financial system in the economic growth: Some
evidence from Saudi Arabian economy. Journal of International Finance and Economics,
9(4), 1-13.
Amman Stock Exchange. [on line] http://www.ase.com.jo
Arestis, P., Demetriades, P. O., & Luintel, K. B. (2001), Financial development and economic
growth: The role of stock markets. Journal of Money, Credit and Banking, 33(1), 16-41.
Barro, R. J. (1990). The stock market and investment. Review of Financial Studies, 3(1), 115-
131.
Caporale, G. M., Howells, P. & Soliman, A. M. (2005). Endogenous growth models and stock
market development: Evidence from four countries. Review of Development Economics,
9(2), 166-176.
Central bank of Jordan [on line] http://www.cbj.gov.jo
Choudry, T. (1995). Long-run money demand function in Argentina during 1935-1962: Evidence
from cointegration and error correction models. Applied Economics, 27,66-167.
Dep, S. G., & Mukherjee, J. (2008). Does stock market development cause economic growth? A
time series analysis for Indian economy. International Research Journal of Finance and
Economics, 21, 142-149.
Dickey, D. A., & Fuller, W. A. (1981). Likelihood ratio statistics for autoregressive time series
with a unit root. Econometrica, 49, 1057-1072.
Dickey, D. A., & Fuller, W. A. (1979), Distribution of the estimation for autoregressive time
series with a unit root. Journal of American Statistical Association, 79, 355-367.
Engle, R .F. & Granger, C.W. J. (1987). Co integration and error correction: Representation,
estimation and testing. Econometrica, 55, 251-276.
Fama, E. F. (1981). Stock returns, real activity, inflation and money. American Economic
Review, 71, 545-565.
Fry, M. J. (1997). In favour of financial liberalization. The Economic Journal, 107, 754-770.
Granger, C. (1988). Some recent development in a concept of causality. Journal of
Econometrics, 39, 199-211.
Guru-Gharan, K. K., Rahman, M., & Parayitam, S. (2009). Influences of selected
macroeconomic variables on U.S. stock market returns and their predictability over
varying time horizons. Academy of Accounting and Financial Studies Journal, 13(1), 13-
31.
Journal of Business & Economic Studies, Vol. 17, No. 2, Fall 2011
34
Johansen, S. (1988). Statistical analysis of cointegration vectors. Journal of Economic Dynamics
and Control, 12, 231-254.
Kassimatis, K., & Spyrou, S. I. (2001). Stock and credit market expansion and economic
development in emerging markets: Further evidence utilizing cointegration analysis.
Applied Economics, 33, 1057-1064.
Khedhiri, S., & Muhammad, N. (2008). Empirical analysis of the UAE stock market volatility.
International Journal of Finance and Economics, 15, 249-260.
King, R. G., & Levine, R. (1993a). Finance and growth: Schumpeter might be right, Quarterly
Journal of Monetary Economics, 108, 717-738.
King, R. G., & Levine, R. (1993b). Finance, entrepreneurship and growth, Journal of Monetary
Economics, 32, 513-542.
Levine, R. (1997). Financial development and economic growth: Views and agenda. Journal of
Economic Literature, 35, 688-726.
Levine, R. (2002). Bank-based or market-based financial systems: Which is better? Journal of
Intermediation, 11, 1-30.
Levine, R., & Zervos, S. (1996). Stock market development and long-run growth. World Bank
Economic Review, 10, 323-339.
Levine, R., & Zervos, S. (1998). Stock markets, banks and economic growth. American
Economic Review, 88, 537-558.
Liang, H-Y., & Reichert, A. (2007). Economic growth and financial sector development. The
international Journal of Business and Finance Research, 1(1), 68-78.
Mayer, C. (1988). New issues in corporate finance. European Economic Review, 32, 1167-1188.
Mazur, E. A., & Alexander, R. J. (2001). Financial sector development and economic growth in
New Zealand. Applied Economics Letters, 8, 545-549.
Oyama, T. (1997). Determinants of stock prices: The case of Zimbabwe. International Monetary
Fund (WP/97/117), 1-44.
Phillips, P., & Perron, P. (1988). Testing for a unit root in time series regression. Biometrica, 75,
333-346.
Sedik, T. S., & Petri, M. (2006). The Jordanian stock market: Should you invest in it for risk
diversification or performance? International Monetary Fund (WP/06/187), 1-36.
Shahbaz, M., Ahmed, N., & Ali, L. (2008). Stock market development and economic growth:
Ardl causality in Pakistan. International Research Journal of Finance and Economics,
14, 182-195.
Singh, A. (1997). Financial liberalization stock markets and economic development. The
Economic Journal, 107, 771-782.
Somoye, R. O. C., Akintoye, I. R., & Oseni, J. E. (2009). Determinants of equity prices in the
stock markets. International Research Journal of Finance and Economics, 30, 177-189.
Toda, H. Y., & Phillips, P. C. B. (1993), Vector Autoregressions and Causality. Econometrica,
61, 1367-1393.
Woertz, E. (2006). GCC stock markets at risk. Gulf Research Center Working Papers, Dubai.
Copyright of Journal of Business & Economic Studies is the property of Dowling College and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder’s express written
permission. However, users may print, download, or email articles for individual use.
RJOAS,11(107), November 2020
160
DOI https://doi.org/10.18551/rjoas.2020-11.19
THE RELATIONSHIP BETWEEN ECONOMIC GROWTH AND AGRICULTURAL
LAND-USE INTENSITY
Wei Zhu
ORCID: 0000-0001-7822-0095
Ruimei Wang*
ORCID: 0000-0001-9964-9421
College of Economics and Management, China Agricultural University, Beijing, China
*E-mail: wangruimei@cau.edu.cn
ABSTRACT
This study aims to investigate the relationship between economic growth and agricultural
land-use intensity for 86 countries over the period 1990-2017. Specifically, this study analyzes
the relationship between economic growth and fertilizer use intensity, pesticide use intensity,
cropping frequency, production value of crops per area, GHG emissions per area. The results
suggest the existence of an inverted-U-shaped relationship between economic growth and
each of the agricultural land-use intensity indicators, which is consistent with the
Environmental Kuznets Curve (EKC) hypothesis. This implies that in the early stages of
economic growth, agricultural land-use intensity rises steadily as per capita income grows,
and then begins to decline as per capita income continues to grow after reaching the turning
point. A mediation analysis is conducted to test the potential causes behind the relationship.
KEY WORDS
Environmental Kuznets Curve, inverted-U-shaped relationship, land-use, economic growth.
The world population has grown steadily, people with higher purchasing power comes a
greater demand for processed food, meat, dairy, and fish, all of which add pressure to the
food supply system (Godfray et al., 2010). At the same time, more than 820 million people in
the world are still hungry today, about 2 billion people in the world experience moderate or
severe food insecurity (FAO et al., 2019). Feeding a growing world population may require an
additional 2.7-4.9 Mha of cropland per year on average (Lambin and Meyfroidt, 2011), but
there is substantially less potential additional cropland than is generally assumed once
constraints and trade-offs are taken into account (Lambin et al., 2013). Intensification of
agricultural land use may have the potential to satisfy the increasing demand for food
(Godfray et al., 2010). However, intensification of agricultural land use can lead to ecological
damage, soil degradation, environmental pollution, and create health risks to livestock and
humans. For example, irrigation practices can limit the amount of water available for other
uses; runoff from pesticides and fertilizers can degrade water quality; the use of agricultural
chemicals can enhance the growth of invasive plants, which can contribute to decreases in
biodiversity; carbon emission from agricultural land use accounts for about 9% of total carbon
emissions (Cassou, 2018; EPA, 2020; FAOSTAT, 2017; Reay et al., 2012). To address these
negative impacts, it is now widely recognized that agricultural land use must be sustainable,
that is, the use of agricultural land at rates that do not exceed the capacity of Earth to replace
them (Godfray et al., 2010; Pellegrini and Fernandez, 2018; Tilman et al., 2011).
The Environmental Kuznets Curve (EKC) is often used to analyze the relationship
between economic growth and various indicators of environmental degradation. The common
point of most studies is that environmental degradation increases faster than per capita
income in the early stages of economic growth and slows down relative to GDP growth in
higher per capita income levels (Dinda, 2004; Grossman and Krueger, 1995). This implies that
there is an inverted-U-shaped relationship between economic growth and the environmental
indicator (Stern, 2004). Environmental indicators used in most studies include atmospheric
indicators, freshwater indicators, land indicators, oceans, seas coasts and biodiversity
indicators (Sarkodie and Strezov, 2019).
mailto:wangruimei@cau.edu.cn
RJOAS, 11(107), November 2020
161
With respect to the relationship between economic growth and agricultural land-use
intensity, there are some studies have separately tested the EKC hypothesis on pesticide use,
fertilizer use, and cultivated area, etc. (Singh and Narayanan, 2015) observed a U-shaped
relationship between fertilizer consumption per hectare and per capita income, an
inverted-U-shaped relationship between pesticide consumption per hectare and per capita
income, using data of the Indian states in 1990–2008. (Zhang et al., 2016) found that
agricultural inputs (consumption of fertilizer, pesticide, and agricultural film) and economic
growth in the Three Gorges Reservoir Area in China between 2003 and 2012 fulfill the EKC
hypothesis. (RodrÍGuez-Meza et al., 2004) analyzed agricultural land use at the household
level, obtained the evidence of an EKC relating farmed area to per capita income. (Managi,
2006) demonstrated that pesticide use followed the inverted-U-shaped curve for US
agriculture in 48 states in 1970-1997. (Ghimire and Woodward, 2013) found a N-shaped
relationship between the pesticide use and per capita GDP using data of 94 countries in
1990-2000. Using data of 119 countries in 1990-2009, (Schreinemachers and Tipraqsa, 2012)
found that as per capita GDP increased, growth in the intensity of pesticide use diminished
and then became negative, suggesting the existence of an inverted-U-shaped curve.
(Hedlund et al., 2020) demonstrate a positive relationship between economic development
and pesticide consumption over time, with no decline in use at higher levels of economic
development. (Ali et al., 2017)’s results do not support the hypothesis of an
inverted-U-shaped relationship between agricultural growth and CO2 emissions in the long
run or the short run in case of Pakistan during 1960 to 1990. However, the agricultural
land-use intensity can be described from three dimensions including input intensity, output
intensity, and human-induced but unintended outcomes (Erb et al., 2013). To the best of our
knowledge, no previous research has investigated the relationship between economic growth
and various agricultural land-use intensity indicators using the same data source. The current
study attempts to fill this gap.
In this study, we examine data compiled by the Food and Agriculture Organization (FAO)
on fertilizer use, pesticide use, cropping area, production value, and net carbon emissions for
86 countries over the period 1990-2017. This study aims to investigate the relationship
between economic growth and each of the agricultural land-use intensity indicators.
Furthermore, a mediation analysis is conducted to test the potential causes behind the EKC
for agricultural land-use intensity.
MATERIALS AND METHODS OF RESEARCH
Data and variable definitions. The data for our analysis were extracted from FAO
database (FAOSTAT, 2017). Following (Erb et al., 2013), the agricultural land-use intensity,
which is the dependent variable in this study, integrates three dimensions: (a) input intensity
measured by fertilizer use intensity, pesticide use intensity, and cropping frequency; (b) output
intensity measured by production value of crops per area; (c) the associated system-level
impacts of agricultural land-based production measured by greenhouse gas (GHG) emissions
per area. The fertilizer use intensity is measured by the use of chemical and mineral fertilizers
per area of cropland, which corresponds to the sum of arable land and permanent crops. The
pesticide use intensity is measured by the use of pesticides per area of cropland. The
cropping frequency is defined as the ratio of area harvested to area of cropland. The
production value of crops per area is defined as the ratio of gross production value of crops
(constant 2014-2016 million US$) to area of cropland. The GHG emissions per area is defined
as the ratio of GHG emissions to area of cropland. The GHG emissions includes emissions
from crop residues, cultivation of organic soils, manure applied to soils, manure left on
pasture, and synthetic fertilizers, expressed as Gg CO2 and CO2eq (from CH4 and N2O).
Due to the availability of data, a total of 86 countries are included in this study. Although the
averaging and simplifications in the data mean that the data are imperfect, we have the
advantage of being able to include 86 countries with all the agricultural land-use intensity
indicators for each country. As data for all years are available for all the 86 countries, the
panel data are strongly balanced. The fertilizer use data are in a time series from 2002 to
RJOAS, 11(107), November 2020
162
2017, the production value data are from 1990 to 2016, and the data of other indicators are
from 1990 to 2017. The rationale for using panel data is to minimize unobserved
heterogeneity bias and some dynamics of land use that are difficult to investigate with
cross-sectional data only (Ghimire and Woodward, 2013).
The major explanatory variable of interest in this study is economic growth which is
measured by per capita GDP (expressed in constant 2015 US$). Changes in the proportions
of different categories of crops will affect the agricultural land-use intensity, thus the
proportions of cereals, vegetables, fruits, roots and tubers, coarse grain, and oil crops are
used as control variables. The descriptive statistics are provided in Table 1. The correlation
coefficients between various agricultural land-use intensity indicators are shown in Table 2.
Table 1 – Descriptive statistics
Variable Obs. Mean Std. Dev. Min. Max.
Fertilizer use intensity (kg/ha) 1376 117.726 99.032 0.745 579.800
Pesticide use intensity (kg/ha) 2408 2.940 3.453 0.005 19.450
Cropping frequency 2408 0.909 0.344 0.0729 2.661
Production value of crops ($/ha) 2322 1053.302 825.645 54.441 5046.765
GHG emissions (kg CO2eq/ha) 2408 1650.516 1248.244 110.979 9229.642
GDP per capita ($) 2408 13026.980 16039.740 172.498 83056.550
Share of cereals 2408 0.451 0.163 0.003 0.887
Share of vegetables 2408 0.044 0.077 0.001 0.709
Share of fruits 2408 0.077 0.091 0.002 0.684
Share of roots and tubers 2408 0.041 0.055 0.001 0.462
Share of coarse grain 2408 0.206 0.119 0.000 0.460
Share of oil crops 2408 0.111 0.121 0.000 0.772
Table 2 – Correlation coefficients between various agricultural land-use intensity indicators
Fertilizer use
intensity
Pesticide use
intensity
Cropping
frequency
Production value
of crops
GHG emissions
Fertilizer use intensity 1.000
Pesticide use intensity 0.616 1.000
Cropping frequency -0.003 -0.178 1.000
Production value of crops 0.570 0.543 0.082 1.000
GHG emissions 0.587 0.417 0.066 0.500 1.000
According to the method adopted in the previous literature (e.g.,(Bimonte and Stabile,
2017; Kuznets, 1955; Pontarollo and Serpieri, 2020; Sarkodie and Strezov, 2019)), this study
defined the empirical model as follows:
ln𝑦𝑖𝑡 = 𝛽0 + 𝛽1ln𝐺𝐷𝑃𝑖𝑡 + 𝛽2 ln𝐺𝐷𝑃𝑖𝑡
2 + 𝜃𝑖 + 𝛾𝑡 + 𝜀𝑖𝑡 （1）
ln𝑦𝑖𝑡 = 𝛽0 + 𝛽1ln𝐺𝐷𝑃𝑖𝑡 + 𝛽2 ln𝐺𝐷𝑃𝑖𝑡
2 + 𝛽𝑗 𝑍𝑖𝑗𝑡 + 𝜃𝑖 + 𝛾𝑡 + 𝜀𝑖𝑡 （2）
Where: i and t represent country i and year t, respectively; y represents agricultural land-use
intensity indicators; GDP denotes per capita GDP; z represents other control variables; θ
indicates the country’s fixed effect; γ indicates the year fixed effect; β’s are the coefficient
estimates of the regressors; ε represents the random error term.
In a typical method for estimating the equations (1) and (2), the EKC hypothesis is valid
if β1>0, β2<0. This implies that there is an inverted-U-shaped relationship between economic
growth and agricultural land-use intensity. The ‘‘turning point’’ income, where the agricultural
land-use intensity is at a maximum, is given by exp(-β1/(2β2)).
RESULTS OF STUDY
In the empirical estimation this study proceeds as follows. First, the reduced form model
shown in equation (1) is estimated separately for fertilizer use intensity, pesticide use intensity,
cropping frequency, production value of crops per area, and GHG emissions per area.
Second, the proportions of different categories of crops are controlled in the model shown in
equation (2). This enables us to check for the robustness of the coefficients associated with
per capita GDP and its squared term to the inclusion of additional control variables. The
RJOAS, 11(107), November 2020
163
equations were estimated using random effect estimator (RE), fixed effects estimator (FE),
and panel-correction standard error (PCSE) models respectively. The joint significance test of
the year dummy variable indicates that the year dummy variable should be included in the
model, thus a two-way fixed effects model should be adopted. Hausmann test indicates that
FE model rather than RE model should be chosen. Modified Wald test for groupwise
heteroskedasticity rejects the null hypothesis of homoscedasticity. Pesaran’s test of
cross-sectional independence cannot reject the null hypothesis that there is no
cross-sectional correlation. Therefore, PCSE model rather than FE model should be chosen.
The estimation results for fertilizer use intensity are presented in Table 3. The
coefficients estimated by FE and PCSE are the same, thus only the results of RE and PCSE
are shown in the table. The coefficients of per capita GDP and its squared term obtained by
different estimation methods have little difference.
Table 3 – Estimation results for fertilizer use intensity
RE PCSE RE PCSE RE PCSE RE PCSE
lnGDP 3.469*** 3.693*** 3.453*** 3.832*** 3.473*** 3.714*** 3.485*** 3.839***
(1.099) (0.407) (1.137) (0.461) (1.105) (0.452) (1.132) (0.478)
(lnGDP)
2
-0.168*** -0.177*** -0.172*** -0.194*** -0.167*** -0.179*** -0.171*** -0.190***
(0.0617) (0.0227) (0.0633) (0.0285) (0.0618) (0.0259) (0.0628) (0.0292)
Country’s fixed effect
Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed effect
No No Yes Yes No No Yes Yes
Share of cereals -2.068 -2.083** -1.389 -1.328*
(1.441) (0.816) (1.320) (0.797)
Share of vegetables -0.222 -0.846 0.0702 -0.501
(1.563) (1.066) (1.477) (0.972)
Share of fruits -4.386** -4.671*** -3.672* -3.891***
(2.235) (1.105) (2.159) (1.120)
Share of roots and
tubers
-1.420 -0.757 -0.772 -0.0388
(1.611) (1.656) (1.402) (1.589)
Share of coarse grain -2.609** -2.311*** -2.136* -1.902***
(1.316) (0.741) (1.235) (0.722)
Share of oil crops -1.858 -1.169 -1.359 -0.569
(1.352) (0.857) (1.254) (0.810)
Constant -12.90*** -13.92*** -12.49** -13.99*** -10.95** -11.71*** -11.31** -12.64***
(4.810) (1.855) (5.017) (1.906) (4.900) (2.208) (5.065) (2.170)
N 1376 1376 1376 1376 1376 1376 1376 1376
R-sq 0.923 0.927 0.926 0.930
Note: Robust standard errors appear in parentheses; asterisks *, **, and *** indicate significance at the 10%, 5%,
and 1% levels, respectively.
Table 4 – Estimation results for pesticide use intensity
RE PCSE RE PCSE RE PCSE RE PCSE
lnGDP 2.405* 2.137*** 2.627** 2.598*** 2.450** 2.322*** 2.545** 2.583***
(1.241) (0.210) (1.272) (0.215) (1.168) (0.221) (1.200) (0.220)
(lnGDP)
2
-0.101 -0.0826*** -0.128* -0.134*** -0.110* -0.102*** -0.123* -0.133***
(0.0707) (0.0123) (0.0748) (0.0131) (0.0660) (0.0136) (0.0702) (0.0135)
Country’s fixed effect Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed effect No No Yes Yes No No Yes Yes
Share of cereals -1.861 -1.803*** -1.880 -1.951***
(1.873) (0.641) (1.886) (0.641)
Share of vegetables 0.466 0.725 0.145 0.0420
(2.316) (0.968) (2.196) (0.952)
Share of fruits 1.013 1.373* 0.956 1.115
(2.418) (0.779) (2.411) (0.772)
Share of roots and
tubers
0.360 0.995 0.200 0.375
(2.272) (0.787) (2.119) (0.767)
Share of coarse grain -1.944 -1.744** -1.829 -1.867***
(1.918) (0.694) (1.923) (0.706)
Share of oil crops 0.905 1.279* 0.622 0.787
(1.770) (0.660) (1.726) (0.661)
Constant -12.75** -13.13*** -12.68** -13.84*** -11.44** -12.48*** -11.27** -12.58***
(5.366) (0.903) (5.357) (0.919) (5.376) (1.114) (5.386) (1.114)
N 2408 2408 2408 2408 2408 2408 2408 2408
R-sq 0.886 0.888 0.891 0.893
Note: Robust standard errors appear in parentheses; asterisks *, **, and *** indicate significance at the 10%, 5%,
and 1% levels, respectively.
RJOAS, 11(107), November 2020
164
Table 5 – Estimation results for cropping frequency
RE PCSE RE PCSE RE PCSE RE PCSE
lnGDP 0.349* 0.374*** 0.363* 0.375*** 0.395* 0.416*** 0.433** 0.455***
(0.204) (0.0523) (0.201) (0.0563) (0.215) (0.0562) (0.208) (0.0627)
(lnGDP)
2
-0.0162 -0.0171*** -0.0182 -0.0175*** -0.0180 -0.0183*** -0.0233* -0.0233***
(0.0125) (0.00339) (0.0123) (0.00389) (0.0132) (0.00379) (0.0125) (0.00451)
Country’s fixed
effect
Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed effect No No Yes Yes No No Yes Yes
Share of cereals 0.218 0.292*** 0.177 0.243**
(0.364) (0.113) (0.364) (0.118)
Share of
vegetables
-1.238** -1.178*** -1.399** -1.320***
(0.570) (0.284) (0.575) (0.287)
Share of fruits -0.699 -0.676*** -0.759 -0.739***
(0.615) (0.200) (0.642) (0.199)
Share of roots and
tubers
-1.031* -0.958*** -1.180** -1.108***
(0.598) (0.315) (0.569) (0.310)
Share of coarse
grain
0.945** 1.053*** 0.917** 1.003***
(0.465) (0.126) (0.467) (0.126)
Share of oil crops 0.309 0.369*** 0.175 0.253**
(0.415) (0.122) (0.439) (0.125)
Constant -1.953** -2.177*** -1.927** -2.159*** -2.396** -2.535*** -2.286** -2.496***
(0.829) (0.205) (0.836) (0.212) (0.974) (0.257) (0.968) (0.270)
N 2408 2408 2408 2408 2408 2408 2408 2408
R-sq 0.934 0.935 0.942 0.943
Note: Robust standard errors appear in parentheses; asterisks *, **, and *** indicate significance at the 10%, 5%,
and 1% levels, respectively.
Table 6 – Estimation results for production value of crops per area
RE PCSE RE PCSE RE PCSE RE PCSE
lnGDP 1.044*** 1.011*** 1.355*** 1.353*** 1.254*** 1.224*** 1.458*** 1.475***
(0.349) (0.0906) (0.350) (0.0890) (0.326) (0.0866) (0.319) (0.0897)
(lnGDP)
2
-0.0320 -0.0283*** -0.0677*** -0.0666*** -0.0490** -0.0445*** -0.0752*** -0.0749***
(0.0208) (0.00595) (0.0215) (0.00619) (0.0197) (0.00579) (0.0196) (0.00626)
Country’s
fixed effect
Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed
effect
No No Yes Yes No No Yes Yes
Share of
cereals
0.511 0.616*** 0.440 0.492***
(0.451) (0.182) (0.413) (0.185)
Share of
vegetables
1.725*** 1.709*** 1.012** 0.935***
(0.527) (0.302) (0.441) (0.247)
Share of fruits 1.497** 1.544*** 1.304* 1.286***
(0.735) (0.265) (0.784) (0.256)
Share of roots
and tubers
0.865 1.075*** 0.348 0.420
(0.695) (0.240) (0.640) (0.262)
Share of
coarse grain
-0.600 -0.290 -0.540 -0.365*
(0.578) (0.226) (0.558) (0.206)
Share of oil
crops
0.927** 1.041*** 0.418 0.509***
(0.448) (0.181) (0.443) (0.177)
Constant 0.123 0.978*** 0.0423 0.457 -0.824 -0.355 -0.549 -0.415
(1.469) (0.352) (1.435) (0.332) (1.426) (0.390) (1.352) (0.407)
N 2322 2322 2322 2322 2322 2322 2322 2322
R-sq 0.953 0.959 0.955 0.960
Note: Robust standard errors appear in parentheses; asterisks *, **, and *** indicate significance at the 10%, 5%,
and 1% levels, respectively.
The estimated coefficients are positive for the linear term and negative for squared term,
which suggests the existence of an inverted-U-shaped relationship between per capita GDP
and fertilizer use intensity. This is consistent with the EKC hypothesis. The estimated turning
point is 24407.154 US$ (2015 prices). This means that as per capita income increases at the
initial stages, fertilizer use intensity follows the same path but begins to decline after reaching
the turning point.
The estimation results for pesticide use intensity are provided in Table 4. The estimated
coefficients are positive for the linear term and negative for squared term, which suggests the
existence of an inverted-U-shaped relationship between per capita GDP and pesticide use
intensity. The estimated turning point is 16490.279 US$ (2015 prices). This means that in the
RJOAS, 11(107), November 2020
165
early stages of economic development, pesticide use intensity rises steadily as per capita
income grows, and then begins to decrease as per capita income continues to grow once the
turning point is reached.
The estimation results for cropping frequency are shown in Table 5. The results suggest
the existence of an inverted-U-shaped relationship between per capita GDP and cropping
frequency and the validity of the EKC. The estimated turning point is 17395.181 US$ (2015
prices). In other words, as per capita income increases, cropping frequency also increases at
first and then starts declining after the turning point.
The estimation results for production value of crops per area are presented in Table 6.
The results indicate that there is an inverted-U-shaped relationship between per capita GDP
and production value of crops per area, and confirm the EKC hypothesis. The estimated
turning point is 18891.398 US$ (2015 prices).
The estimation results for GHG emissions per area are provided in Table 7. The results
confirm the EKC hypothesis. The estimated turning point is 17387.489 US$ (2015 prices).
This means that GHG emissions per area increases faster than per capita GDP in the early
stages of economic growth and slows down relative to per capita GDP growth in higher per
capita income levels.
Table 7 – Estimation results for GHG emissions per area
RE PCSE RE PCSE RE PCSE RE PCSE
lnGDP 0.779** 0.797*** 0.874*** 0.915*** 0.747** 0.767*** 0.846*** 0.900***
(0.339) (0.0516) (0.337) (0.0527) (0.322) (0.0514) (0.319) (0.0580)
(lnGDP)
2
-0.0319 -0.0327*** -0.0436** -0.0466*** -0.0295 -0.0302*** -0.0419** -0.0460***
(0.0202) (0.00312) (0.0204) (0.00340) (0.0195) (0.00313) (0.0196) (0.00388)
Country’s fixed effect Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed effect No No Yes Yes No No Yes Yes
Share of cereals 0.301 0.349*** 0.359 0.388***
(0.518) (0.135) (0.512) (0.137)
Share of vegetables 0.687 0.639*** 0.469 0.367**
(0.708) (0.197) (0.659) (0.183)
Share of fruits 0.425 0.427** 0.440 0.413**
(0.608) (0.194) (0.609) (0.191)
Share of roots and tubers 0.500 0.467** 0.280 0.
166
(0.647) (0.238) (0.601) (0.246)
Share of coarse grain 0.00615 0.0846 -0.00530 0.0184
(0.559) (0.177) (0.553) (0.186)
Share of oil crops -0.132 -0.107 -0.297 -0.302**
(0.482) (0.128) (0.471) (0.133)
Constant 2.910** 3.036*** 2.975** 2.931*** 2.801** 2.794*** 2.892** 2.708***
(1.401) (0.223) (1.382) (0.233) (1.385) (0.247) (1.356) (0.292)
N 2408 2408 2408 2408 2408 2408 2408 2408
R-sq 0.950 0.952 0.951 0.953
Note: Robust standard errors appear in parentheses; asterisks *, **, and *** indicate significance at the 10%, 5%,
and 1% levels, respectively.
DISCUSSION OF RESULTS
To better understand the impact of economic growth on agricultural land-use intensity,
we continue to discuss the potential causes behind the EKC for agricultural land-use intensity.
According to (Dinda, 2004), several factors are responsible to shape the EKC, including
income elasticity of environmental quality demand, international trade, market mechanism,
regulation, etc. For the agricultural land-use intensity indicators, fertilizer use intensity,
pesticide use intensity, and GHG emissions per area are closely related to environmental
degradation, thus, as income grows, people achieve a higher standard of living and care more
for the quality of environment they live in and demand for better environment induces
structural changes in agriculture that tends to reduce agricultural chemical use and GHG
emissions. Increasing food demand requires more input including farmland, fertilizer,
pesticide, labor, energy, etc. Therefore, agricultural land-use intensity will be affected by
population and income level that affect food demand. On the other hand, technological
progress occurs with economic growth and the obsolete and dirty technologies are replaced
by upgraded new and cleaner technology, which improves environmental quality. This is the
RJOAS, 11(107), November 2020
166
technique effect of economic growth (Dinda, 2004). Sustainable intensification has been
gaining attention in policy discussions as an appropriate means to use land in order to
increase food supplies while protecting environmental security (Petersen and Snapp, 2015).
Meanwhile, because the adoption of technological innovations and modern management
practices involves high fixed costs, the adoption rate of small-scale farms is relatively low
(Feder et al., 1985). Some empirical studies highlighted the importance of farm size in
understanding changes in land use intensity (van der Sluis et al., 2016; Wu et al., 2018; Zhu
and Wang, 2020). Because land resources are limited, the expansion of farm size is mainly
due to the transfer of agricultural labor to other industries. (Liu et al., 2016) found there is an
inverted-N-shaped relationship between rural out-migration and arable land use intensity.
While (Gray and Bilsborrow, 2014) found that migrant departure has a positive effect on
agricultural activities. (Caulfield et al., 2019) demonstrated that remittances received from
rural out-migrations were associated with an increase in the use of pesticides and chemical
fertilizers. (Zhang et al., 2020) found that off-farm employment has a significant influence on
chemical fertilizer use. (Jiang et al., 2013) revealed that urban expansion driven by population
urbanization decreased the agricultural land use intensity. While (You et al., 2018) found that
urban population proportion has a positive effect on input intensity.
Table 8 – Estimation results for mediation analysis
Total pop. Urban
pop.
Farm
size
Fertilizer Pesticide Currency Value Emission
lnGDP 0.680*** 0.570*** -0.264*** 2.523*** 2.668*** 0.410*** 1.037*** 0.515***
(0.0385) (0.0244) (0.0533) (0.430) (0.265) (0.0661) (0.0844) (0.0628)
(lnGDP)
2
-0.0556*** -0.0360*** 0.0249*** -0.0988*** -0.133*** -0.0187*** -0.0460*** -0.0149***
(0.00356) (0.00152) (0.00393) (0.0252) (0.0162) (0.00450) (0.00582) (0.00433)
ln(total pop.) 4.702*** 2.571*** 0.0820 -1.125*** 0.240
(1.180) (0.422) (0.111) (0.108) (0.151)
(ln(total pop.))
2
-0.235*** -0.147*** -0.00925* 0.0673*** 0.000776
(0.0694) (0.0219) (0.00558) (0.00508) (0.00875)
ln(urban pop.) 0.730 0.0929 0.317*** 1.153*** 1.031***
(0.456) (0.231) (0.0820) (0.101) (0.0994)
(ln(urban pop.))
2
-0.504*** -0.180** 0.0544*** 0.244*** 0.210***
(0.0777) (0.0761) (0.0195) (0.0285) (0.0264)
ln(farm size) -0.646*** 0.0685 -0.122*** -0.262*** -0.550***
(0.119) (0.0511) (0.0158) (0.0224) (0.0146)
(ln(farm size))
2
0.0919*** 0.114*** 0.0690*** 0.0347*** -0.00307
(0.0333) (0.0277) (0.00532) (0.00678) (0.00786)
Control of categories
of crops
No No No Yes Yes Yes Yes Yes
Country’s fixed effect Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed effect Yes Yes Yes Yes Yes Yes Yes Yes
Constant 5.856*** -2.774*** -0.131 -30.91*** -24.49*** -2.437*** 6.675*** 2.119**
(0.118) (0.107) (0.196) (6.382) (2.575) (0.659) (0.791) (0.930)
N 2408 2408 2408 1376 2408 2408 2322 2408
R-sq 0.998 0.991 0.980 0.938 0.895 0.947 0.965 0.968
Note: Robust standard errors appear in parentheses; asterisks *, **, and *** indicate significance at the 10%, 5%,
and 1% levels, respectively.
The literature shows that various economic indicators, such as total population, urban
population proportion, farm size, maybe potential mediators of the effects of economic growth.
In order to test the effect of these potential mediators, we perform a mediation analysis
(Mackinnon and Fairchild, 2009). Due to the availability of data, we detect the mediated
effects of total population, urban population proportion, and farm size. The cropland area per
rural population is used as the proxy variable of farm size. The data for our analysis were
extracted from FAO database (FAOSTAT, 2017). The brief estimation results for mediation
analysis are shown in Table 8, and the detailed estimation results are provided in Tables
S1-S6 in Appendix. According to (Preacher and Hayes, 2008), the estimation results of
separate models and one model including all mediators indicate that total population, urban
population proportion, and farm size are mediators of the effects of economic growth on the
agricultural land-use intensity.
RJOAS, 11(107), November 2020
167
CONCLUSION
This study investigates the relationship between economic growth and agricultural
land-use intensity indicators, including fertilizer use intensity, pesticide use intensity, cropping
frequency, production value of crops per area, and GHG emissions per area, using data of 86
countries in 1990-2017. The results demonstrate the existence of an inverted-U-shaped
relationship between economic growth and each of the agricultural land-use intensity
indicators, which is consistent with the EKC hypothesis. This implies that in the early stages of
economic growth, agricultural land-use intensity rises steadily as per capita income grows,
and then begins to decline as per capita income continues to grow after reaching the turning
point. Furthermore, the mediation analysis shows that total population, urban population
proportion, and farm size are mediators of the effects of economic growth on the agricultural
land-use intensity.
It should be noted that this study has several limitations. First, the indicators of
agricultural land-use intensity other than the five indicators analyzed above have not been
discussed in this study. Second, some of the variables, such as cropping frequency and farm
size, are proxy variables, so that there may be measurement errors, that may lead to
inaccurate estimates of these variables. Third, the potential mediated effects of technological
progress, government expenditure on agriculture, foreign direct investment, etc. have not
been tested in this study. In future research, more accurate and detailed data can be used to
address the above limitations.
ACKNOWLEDGEMENTS
This work was supported by the National Social Science Fund of China [16AJY013].
REFERENCES
1. Ali G, Ashraf A, Bashir MK, Cui S. Exploring environmental Kuznets curve in relation to
green revolution: A case study of Pakistan. Env. Science & Policy 2017; 77: 166-171.
2. Bimonte S, Stabile A. Land consumption and income in Italy: a case of inverted EKC.
Ecological Economics 2017; 131: 36-43.
3. Cassou E. Pesticides. Agricultural Pollution. The Challenge of Agricultural Pollution:
Evidence from China, Vietnam, and the Philippines. World Bank, Washington, DC., 2018.
4. Caulfield M, Bouniol J, Fonte SJ, Kessler A. How rural out-migrations drive changes to
farm and land management: A case study from the rural Andes. Land Use Policy 2019; 81:
594-603.
5. Dinda S. Environmental Kuznets Curve Hypothesis: A Survey. Ecological Economics 2004;
49: 431-455.
6. EPA. Report on the Environment. https://www.epa.gov/report-environment/
land-use#note4 [accessed 10 July 2020], 2020.
7. Erb KH, Haberl H, Jepsen MR, Kuemmerle T, Lindner M, Muller D, et al. A conceptual
framework for analysing and measuring land-use intensity. Current Opinion in
Environmental Sustainability 2013; 5: 464-470.
8. FAO, IFAD, UNICEF, WFP, WHO. The State of Food Security and Nutrition in the World
2019. FAO, Rome, 2019.
9. FAO Statistical Databases. http://www.fao.org/faostat/en/ (accessed 17 March 2020).
10. Feder G, Just RE, Zilberman D. Adoption of agricultural innovations in developing
countries: A survey. Economic development and cultural change 1985; 33: 255-298.
11. Ghimire N, Woodward RT. Under- and over-use of pesticides: An international analysis.
Ecological Economics 2013; 89: 73-81.
12. Godfray HC, Beddington JR, Crute IR, Haddad L, Lawrence D, Muir JF, et al. Food
security: the challenge of feeding 9 billion people. Science 2010; 327: 812-8.
13. Gray CL, Bilsborrow RE. Consequences of Out-Migration for Land Use in Rural Ecuador.
Land use policy 2014; 36.
14. Grossman GM, Krueger AB. Economic growth and the environment. Quarterly Journal of
RJOAS, 11(107), November 2020
168
Economics 1995; 110: 353-377.
15. Hedlund J, Longo SB, York R. Agriculture, Pesticide Use, and Economic Development: A
Global Examination (1990-2014) Rural Sociology 2020; 85: 519-544.
16. Jiang L, Deng X, Seto KC. The impact of urban expansion on agricultural land use
intensity in China. Land Use Policy 2013; 35: 33-39.
17. Kuznets S. Economic growth and income inequality. Amer. Economic Review 1955: 1-28.
18. Lambin EF, Gibbs HK, Ferreira L, Grau R, Mayaux P, Meyfroidt P, et al. Estimating the
world’s potentially available cropland using a bottom-up approach. Global Environmental
Change 2013; 23: 892-901.
19. Lambin EF, Meyfroidt P. Global land use change, economic globalization, and the looming
land scarcity. Proc Natl Acad Sci USA 2011; 108: 3465-72.
20. Liu G, Wang H, Cheng Y, Zheng B, Lu Z. The impact of rural out-migration on arable land
use intensity: Evidence from mountain areas in Guangdong, China. Land Use Policy 2016;
59: 569-579.
21. Mackinnon DP, Fairchild AJ. Current Directions in Mediation Analysis. Curr Dir Psychol Sci
2009; 18: 16.
22. Managi S. Are there increasing returns to pollution abatement? Empirical analytics of the
Environmental Kuznets Curve in pesticides. Ecological Economics 2006; 58: 617-636.
23. Pellegrini P, Fernandez RJ. Crop intensification, land use, and on-farm energy-use
efficiency during the worldwide spread of the green revolution. Proc Natl Acad Sci U S A
2018; 115: 2335-2340.
24. Petersen B, Snapp S. What is sustainable intensification? Views from experts. Land Use
Policy 2015; 46: 1-10.
25. Pontarollo N, Serpieri C. Testing the Environmental Kuznets Curve hypothesis on land use:
The case of Romania. Land Use Policy 2020; 97.
26. Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behav Res Methods 2008; 40:
879-91.
27. Reay DS, Davidson EA, Smith KA, Smith P, Melillo JM, Dentener F, et al. Global
agriculture and nitrous oxide emissions. Nature Climate Change 2012; 2: 410-416.
28. RodrÍGuez-Meza J, Southgate D, GonzÁLez-Vega C. Rural poverty, household responses
to shocks, and agricultural land use: panel results for El Salvador. Environment and
Development Economics 2004; 9: 225-239.
29. Sarkodie SA, Strezov V. A review on Environmental Kuznets Curve hypothesis using
bibliometric and meta-analysis. Science of the Total Environment 2019; 649: 128-145.
30. Schreinemachers P, Tipraqsa P. Agricultural pesticides and land use intensification in high,
middle and low income countries. Food Policy 2012; 37: 616-626.
31. Singh AP, Narayanan K. Impact of economic growth and population on agrochemical use:
evidence from post-liberalization India. Environment, Development and Sustainability
2015; 17: 1509-1525.
32. Stern DI. The Rise and Fall of the Environmental Kuznets Curve. World Development
2004; 32: 1419-1439.
33. Tilman D, Balzer C, Hill J, Befort BL. Global food demand and the sustainable
intensification of agriculture. Proc Natl Acad Sci U S A 2011; 108: 20260-4.
34. van der Sluis T, Pedroli B, Kristensen SBP, Lavinia Cosor G, Pavlis E. Changing land use
intensity in Europe – Recent processes in selected case studies. Land Use Policy 2016;
57: 777-785.
35. Wu Y, Xi X, Tang X, Luo D, Gu B, Lam SK, et al. Policy distortions, farm size, and the
overuse of agricultural chemicals in China. Proc Natl Acad Sci USA 2018; 115: 7010-7015.
36. You H, Hu X, Wu Y. Farmland use intensity changes in response to rural transition in
Zhejiang province, China. Land Use Policy 2018; 79: 350-361.
37. Zhang T, Ni J, Xie D. Assessment of the relationship between rural non-point source
pollution and economic development in the Three Gorges Reservoir Area. Environ Sci
Pollut Res Int 2016; 23: 8125-32.
38. Zhang Y, Long H, Li Y, Ge D, Tu S. How does off-farm work affect chemical fertilizer
application? Evid. from China’s mountainous and plain areas. Land Use Policy 2020; 99.
39. Zhu W, Wang R. Impact of farm size on intensity of pesticide use: Evidence from China.
Science of The Total Environment 2020.
The manufacturing sector plays an important role
in the development of modern economy all over the
world. Manufacturing – as a sub – sector of the indus-
try-refers to the production of raw materials and other
factors of production, such as labor, land and capital,
or goods and services through the production pro-
cess. With the multiplier effect, the share of the manu-
facturing industry sectors in the world value added
is around 20%. Similarly, the sector also covers about
1 in 9 of the total employment in the world (UNIDO,
2017). As to put it, the share of the sector in GDP is
also very high and crucial. For example, in Turkey, the
average manufacturing industry share in total GDP is
40% but it varies yearly. While the shares of agriculture
and service sectors in GDP were declining in years re-
cently, the share of the industrial sector was increas-
ing (BAT, 2017).
To fully realize the importance of the industry
in evaluating the above figures, the interaction be-
tween the other sectors of the industrial sector and
other economic activities on the economy scale has
been examined for many years. In the 1920s, British
economist Allyn Young suggested that network-type
connections between sectors of this type were the
Nuri Hacievliyagil, Ibrahim Halil Eksi
Abstract
This study examines the relationship between bank credits and performance and growth of manufactur-
ing sub-sectors. Industrial Production Index was used for a different approach as a dependent variable.
Indications of the autoregressive distributed lag (ARDL) bound co-integration test support the theory that
bank credits are more effective than loan rates on industrial production of sub-sectors. Moreover, the increase
in bank credit leads to the rise of industrial production in all the sub-sectors, except Machinery. According to
the Toda Yomamato causality test results, there are different degrees of causalities in means of the impor-
tance of bank loans for industrial production. On the other hand, in all sub-sectors except machinery and
chemical sub-sectors, causality relations were observed at different grades beginning from loan interest rates
to industrial production. As a result, this study concludes with the evidence of supply leading hypothesis via
the financial sector leads and causes economic growth.
Keywords: Economic Growth; Bank Credit; Manufacturing Sector
Jel classification: E59, L69, O47
1. InTRODuCTIOn
Nuri Hacievliyagil, PhD
Inonu University, Turkey
E-mail: nuri.hacievliyagil@inonu.edu.tr
Ibrahim Halil Eksi, PhD, Assoc. Prof.
Gaziantep University, Turkey
E-mail: ieksi@gantep.edu.tr
South East European Journal of Economics and Business
Volume 14 (1) 2019, 72-91
DOI: 10.2478/jeb-2019-0006
Copyright © 2019 by the School of Economics and Business Sarajevo
A Micro BAsed study oN BANk credit ANd ecoNoMic
Growth: MANufActuriNG suB-sectors ANAlysis
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
73South East European Journal of Economics and Business, Volume 14 (1) 2019
main source of increasing returns on the scale of the
economy. Industrial activities could have a synergistic
effect on economic growth, even if it creates a scale
effect in the sense that Allyn Young has mentioned, or
even if these returns cannot be observed or qualita-
tive parsed (Arisoy, 2008). Another well-known British
economist Nicholas Kaldor (1968), who was inspired
by Allyn Young’s views, confirmed the existence of
such an influence and concluded the industrial sector
as the propulsive power of economic growth and laid
the first law known as its own name.
According to first law of Kaldor, the issue is exist-
ence of a positive relationship between the growth
rate of the manufacturing industry and the rate of
economic growth. According to this law, productivity
in the manufacturing industry increases faster than
in the other sectors because of the increasing returns
in the manufacturing industry, and consequently, the
economy grows rapidly. The second law of Kaldor, also
known as the Verdoorn law, poses a positive relation-
ship between labor productivity and manufacturing
industry production in the manufacturing industry
sector, while the third law argues that the increase in
production in the manufacturing industry increases
productivity in other sectors and ultimately increases
the efficiency in the whole economy. (Mamgain, 1999;
Pons-Novell and Viladecans-Marsal, 1998).
The importance of the manufacturing industry in
terms of the country’s economy raises the financing
status of companies operating in the sector. In this
sense, developing countries undertake the most im-
portant roles of this play. Apart from bank loans, it is
seen that the manufacturing sector has extremely
scarce resources in the financial market for develop-
ing countries, plus the use of other financial resources
is very limited.
The development of any sector depends largely
on the funding of the country’s financial system.
Currently, the Turkish financial systems provide sec-
tor companies with funding in three main ways: first,
capital financing from the capital markets; second,
debt financing by issuing commercial bonds; third,
debt financing from the bank loans. (Duan, 2008) In
fact, as on Turkey extend, because commercial bond
issuance and capital financing from the capital mar-
ket are limited to some super-large enterprises, com-
panies cannot benefit enough from the capital mar-
kets. Therefore, manufacturing sector companies are
forced to use bank loans. As a result, the dependence
degree of bank credits on Turkish firms is becoming
too much.
As in many developing countries, the level of bank
loan usage is quite high, as a result, companies in
Turkey are not able to benefit from the capital markets
sufficiently. In a conducted survey, 48% of compa-
nies consult the use of bank loans in Turkey (Demirci,
2017). According to the numbers taken from the of-
ficial authorities, 20% of all credits used by manu-
facturing industry sector in 2011. This figure was at a
17% level in 2016 with a small depreciation. The men-
tioned amount used in the manufacturing sector and
the indication increased to ₺378,680,695 out of the
total amount loan of ₺2,102,789,474 in 2017. As can
be seen from the figures, approximately around 20%
of total cash loans are usually given to firms in Turkish
manufacturing industry (BRSA, 2017). The downtrend
in capacity utilization rates seen since end-2017 has
had a dampening effect on the demand for new in-
vestments in the Turkish manufacturing industry
(CBRT, 2018). On 2018, the growth of the Turkish man-
ufacturing sector, which is the most strategic one and
very important for the country’s economy, using ap-
proximately 24% of the cash loans where construction
sector using 12%, wholesale and retail trade sector
using 15%, general service sector using 13%, energy
sector using 10% (CBRT, 2018).
In Turkey, various reasons have been motivated
to conduct a survey on the impact of bank loans on
industrial production. Turkey is a popular country in
Europe and Asia region. The country has great poten-
tial in many sectors of the economy, including manu-
facturing, agriculture and tourism. The GDP numeral
of Turkey in 2016 was 857 billion US Dollars. Thanks
to this figure, Turkey is the 17th largest country in the
world (T24, 2017). As a result of the growth in eco-
nomic activity, the ratio of the total financial debt
of firms to GDP remained at 60 percent during 2017
(CBRT, 2018). Even though aggregate corporate finan-
cial leverage has increased to around 65 percent since
the beginning of 2018 due to exchange rate develop-
ments, the ratio of corporate loans to GDP remains
below global, G20 and EME (emerging economies) av-
erages. (BIS, 2018). Graphic 1 shows a comparison of
the ratio of corporate loans to GDP in some emerging
economies with Turkey.
The contribution of bank credits to economic
growth is highly related to the sectors where credits
are available and the added value created by these
sectors. Lending to high value-added and high-pro-
ductivity sectors contributes to the growth of GDP,
and enables the use of resources in the financial sec-
tor effectively, allowing for more new resources to be
created. Gross value added can be used to analyze the
sector’s contribution to GDP by analyzing the efficien-
cy of the sector’s credit usage on a production basis.
In this context, the manufacturing industry is the sec-
tor that contributed most to GDP in Turkey is also the
highest share in total loans. Despite the fact that the
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
74 South East European Journal of Economics and Business, Volume 14 (1) 2019
Graphic 1: Bank Credit to the Private Non-Financial Sector As a Percentage of GDP
Source: https://stats.bis.org/statx/srs/table/f2.4
13
4,
1
46
6
46
,6
6
4,
2
56
,7
65
,4
53
,9
2013
14
0,
4
54
,5
66
55
7
2014
BANK CRED
China R
15
2,
7
55
,7
65
,8
56
,7
4
IT TO THE P
AS A PER
Russia Bras
15
2,
7
55
,3
66
,8
56
,2
66
,6
59
,3
2015
PRIVATE NO
RCENTAGE O
sil India
15
7,
1
51
,8
62
,3
59
,3
20
N‐FINANCIA
OF GDP
South Afric
53
,1
64
,8
61
,5
16
AL SECTOR
ca Turkey
15
6,
1
50
,6
59
,7
52
,9
63
,5
2017
62
,5
Graphic 2: Share of Corporate Sector’s Financial Debt in GDP (%)
Source: CBRT, 2018, Financial Stability Report-II. http://www.tcmb.gov.tr/
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
75South East European Journal of Economics and Business, Volume 14 (1) 2019
Turkish manufacturing industry has largely used cred-
it, the share of total credits has fallen over time (CBRT,
2018).
According to Turkish foreign trade data, the share
of high-tech manufacturing industry products in im-
ports corresponds to five times the share of exports
and makes a structural contribution to the current
account deficit (CBRT, 2018). Therefore, the financial
support of sub-sectors with high technology will have
positive reflections on both macroeconomic terms
such as GDP and current account balance in terms of
productivity and micro-level production and employ-
ment decisions. Increasing competitiveness of the real
sector in the global technology products market will
increase the level of development of the country in
the medium-long term. From here, as an indicator of
the growth of the country and the sector, the relation-
ship between industrial production and bank loans
need to be analyzed.
Bank loans are indispensable for the realization of
economic growth and for the completion of develop-
ment. In the case of the transfer of this principle to the
sectors which will be produced through financial in-
stitutions, especially in the manufacturing sector and
agriculture sector, efficiency is realized and leads to
development. In fact, although there are publications
on this transmission mechanism applied to agricul-
ture in many places, such studies are not being carried
out for manufacturing sub-sectors.
In this study, it is aimed to investigate the produc-
tivity of the manufacturing industry sector, and see
how much the sector companies are related to the use
of bank loans. In this respect, the effects of foreign fi-
nancing sources in the increase of production can be
determined on the sub-sector basis and the indus-
trial policies to be implemented can be highlighted.
Usage of sub-sector Industrial Production Index data
in countries, such as Turkey, those are not published
in the industrial production index, is bringing a differ-
ent perspective. Thus, this study is different from other
studies and it contributes to the literature while it is
addressing the sub-sectors. In particular, the lack of
work done for sub-sectors in Turkey for the production
of goods and services is remarkable. Our new point of
view will also help to cover this gap in the sub-sectors.
The rest of the work will be divided into four:
Part one: the review of literature, part two: materials
and methods, part three: empirical results, and finally
part four is the conclusion and policy implications.
2. THEORY
Banking constitutes a bank credit channel, and the
credit channel2 mechanisms. Bank credit channel; as
a result of monetary policy practices, affects the over-
all country’s output by changing the amount of credit
given to the firm. For example, a restrictive monetary
policy reduces the amount of credits that banks give
to firms by reducing bank reserves and deposits. The
decrease in credits affects the investment expendi-
tures in a negative direction and causes decrescent
national income. (Mishkin, 1995).
On the other hand, it is assumed that banks’ lend-
ing requests will increase with the expansion of mon-
etary policy of increased bank deposits and reserves.
Therefore, increasing the investment demand of firms
increases the consumption level and ultimately the
total production level. This mechanism, in which the
full functionality of the credit channel is provided,
can be broken in some cases. Companies that have a
less expensive bond, securities or more return to the
sale of goods will negatively affect the processing of
the bank credit channel. In this case, the bank credit
channel will try to reduce the impact of monetary
policy on the real economy (Romer, 1990; Oliner and
Rudebusch, 1996; Meltzer, 1995; Kashyap et al., 1996;
Çamoğlu and Akıncı, 2012; Ümit, 2016).
Another failure in the mechanism of credit channel
will occur in case of an imbalance between the return
of the lender banks and the cost of the borrowers that
will have to endure. Bernanke and Gertler (1995) said
that monetary policy does not affect only the real in-
terest rates, but also the foreign financing premium,
Table1: Value Added and Productivity with Credit Share of Manufacturing Sector in GDP
Sector
Gross Value Added
(Contribution to GDP, %)
Value Added Per
Hour Worked
(at prices of 2005 TL)
Share in Commercial Loans (%)
2007 2012 2017 Ave. 2009-15 2007 2012 2017
Change
2007-2017
Manufacturing
Sector
19,5 18,6 20,6 12,4 37,7 32,4 26,6 -11,1
Source: CBRT, 2018, Financial Stability Report-II. http://www.tcmb.gov.tr/
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
76 South East European Journal of Economics and Business, Volume 14 (1) 2019
which expresses the difference between the firms›
own internal resources and external resources. In ad-
dition to the changes in monetary policy, Bernanke
and Gertler (1995) emphasized that both short-term
interest rates and foreign financing premiums need to
be managed properly, additionally real expense and
activities will run smoothly through a credit channel
that operates properly in a suitable monetary policy.
Investigating the factors that affect the efficiency
and proper functioning of the bank credit channel,
the relationship between bank credit and economic
growth has been an important and extensive sub-
ject. Schumpeter was the pioneer of the subject and
he stated the importance of technological innova-
tion in long-run economic growth by bank credits
(Gurgul and Lach, 2012). Schumpeter mentioned the
importance of banking system in facilitating capi-
tal outlays in productive investment. Thereafter, the
Schumpeterian aspect has really improved and be-
came a supply-leading hypothesis. Today, advocates
of supply-side hypothesis believe that the produc-
tivity of financial institutions, financial activities and
especially the economies that bring these two com-
ponents together will increase. Moreover, the supply-
side hypothesis advocates of countries with more
developed financial systems say they grow faster and
develop their economies faster. On the other hand,
the demand-side hypothesis argues that as the econ-
omy grows, the financial system will strengthen and
financial development will be possible. Advocates of
demand-side hypothesis says that economic growth is
an important factor for financial development. In this
respect, economic growth stimulates growth in the
real sector, and it stimulates the financial sector.
3. LITERATuRE
It is observed that studies on bank credits and eco-
nomic growth are concentrated in two groups. In the
first group, macroeconomic studies of bank credits
and economic growth issues were observed. A sig-
nificant correlation was observed between bank loans
and economic growth in some of the related stud-
ies that began with Schumpeter in 1912 (Gurgul and
Lach, 2012). Patrick (1966), Greenwood and Jovanovic
(1990) , King and Levine (1993), Demetriades and
Hussein (1996), De Gregorio and Guidotti (1995), Rajan
and Zingales (1998), Das and Maiti (1998), Levine et al.
(2000), Beck et al., (2000), Christopoulos and Tsionas
(2004), Demirguc-Kunt and Levine (2008), Mishra et.
al. (2009a) and (2009b), Pradhan (2010a), Hassan et.
al. (2011), Banerjee (2012), Iyoboyi (2013), Ebi and
Emmanuel (2014), Sehrawat and Giri (2015), Mohanty
et al. (2016) concluded that increase in bank credit
(or financial development) leads to higher economic
growth. All these studies found significant evidence
and consequently supported the supply leading hy-
pothesis. However, Boyreau-Debray (2003), Guariglia
and Poncet (2008), Leitao (2012), Sehrawat and Giri
(2016) and Chow et al. (2018) found that bank cred-
its showed a negative impact on economic growth.
Some, in addition, found significant evidence that
economic growth will create demand for the various
financial services that the financial system will pro-
vide. Supporters of demand leading hypothesis were
Robinson (1952), Kar and Pentecost (2000), Ansari
(2002), Favara (2006), Kandir et. al. (2007), Chakraborty
(2008), Ceylan and Durkaya (2010), Pradhan (2010b),
Oluitan (2012) and Ak et. al. (2016). There are also
a considerable number of studies in the literature
that cannot observe the relationship between bank
credits and economic growth: Odedokun (1998), Wa
(2002), Aziz and Duenwald (2002), Bloch and Tang
(2003) Demetriades and Andrianova (2003), Chen
(2006) Shan and Jianhong (2006), Loayza and Ranciere
(2006), Estrada et al. (2010), Onder and Ozyıldırım
(2010), Kumar (2011), Demetriades and James (2011),
and Lu and Shen (2012).
The second group was micro-studies; it observed
the examining bank credits and sector or firm per-
formances. Like macro-studies, micro-studies have
reached various results. Some studies have claimed
that bank loans are positive or vice versa to industri-
al sector growth, meanwhile some other researchers
have found a weak or no relationship between bank
lending and sector growth.
Studies examining the effect of bank credit to the
public sector in the past literature have found weak in-
teraction. King and Levine, (1993), Odedokun, (1998),
Levine, (2002), and Beck et al., (2005) are examples of
these studies. Researchers have found that bank loans
are delivered with political motivation and idle areas.
Some other past studies Gurley and Shaw (1955) and
Adve (1980), favored positive effect of bank credits on
industrial growth. In other respect, Ajayi (2000), Wa
(2002), Bloch and Tang (2003) argued that the inex-
pressive positive and negative relation between bank
loans and the growth of the sector was widespread in
time.
After examining past literature on micro-studies
-bank credits and sector or firm performances- it came
to an end of concentrating on more details of recent
studies. It has been taken a closer look at the sector-
based manufacturing and credit relationships that
constitute the basis for this study.
Kelly and Everett (2004) examined Irish private
sector bank credits. Although they warned about
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
77South East European Journal of Economics and Business, Volume 14 (1) 2019
inflationary economic policies and financial crises, the
growth of Manufacturing, Building and Construction,
Hotels and Catering and Education positively affected
from bank loans. As a result, they helped in order to re-
sume stable and reliable real growth in the economy.
If the bank loans are similar to a useful drug for com-
panies, it can be said that Larrain (2006) used this drug
in three main indicators. Larrain (2006) stated that the
first drug in the industrial production was bank loans.
Later, this drug, which caused idiosyncratic volatil-
ity reduction, also caused the reduction of volatility.
Therefore, the links between sectors and GDP are in-
creasing with the bank loan drug. Larrain (2006) em-
phasized that short-term debt is more negatively as-
sociated with firm activity when you increase the dose
of the drug. Izhar and Tariq (2009), which examined
the efficiency of the Indian agricultural sector be-
tween 1992-2005, found that bank loans did not grow
at the rate in which the aggregate agricultural output
grew. Moreover, according to the research, credits did
not determine agricultural output in India.
Tawose (2012), who tries to measure the perfor-
mance of the industrial sector, compared the loans
and advances of commercial banks to the industrial
sector and total savings, interest rate and inflation
rates. Finding a long-term relationship between these
variables, Tawose (2012) found that the variables con-
sidered in the short term had a positive effect on the
performance of the industrial sector. On the contrary,
it was emphasized that loans and advances of long-
term banks extended to the industry sector had an
insignificant negative impact on sector performance.
One of the studies conducted between the growth of
the agricultural sector and commercial bank loans is
Toby and Peterside (2014). Although there was a sig-
nificant weak relationship between commercial bank
loans and the contribution of agriculture to GDP, there
was a significant positive relationship between bank
loans and agricultural contribution to GDP. In the
same study, Toby and Peterside (2014) looked at rela-
tions with commercial bank loans on the manufactur-
ing sector, and found that the manufacturing contri-
bution to GDP was in a significant inverse relationship
with bank loans. According to Buono and Formai
(2014), bank credit was positively and significantly
correlated with export incomes. Higher initial levels of
productivity, collateral and credit rating is associated
with a higher export growth were examined in order
to estimate the structural OLS within firm leveled con-
trols such as size and productivity. The analysis strong-
ly suggested that there is a positive and causal link
between access to bank credit and total revenues and
consequent firm growth.
Ebi and Emmanuel (2014) showed that bank
credits impacted the manufacturing sub-sector posi-
tively and significantly. Sub-sector borrowers’ (mining
and quarry firms) the output was positively correlated
and determined by bank credits. Moreover, contrary
to expectations, they could not determine an impor-
tant relationship between interest rate and industy
and its sub-sector outputs. Chisasa (2014), who stud-
ies in South Africa, found that the bank’s loan reduced
agricultural output in the short term and emphasized
that the uncertainties in the country’s corporate loan
should be eliminated. Chisasa (2014) reached differ-
ent results at macro and micro level. He stated that
supply-side findings are predominant on agricultural
sector basis, but that there is a causal relationship be-
tween economic growth and finance at macro level
and therefore there is a demand-side relationship.
In another study examining the level of agricultural
output, Nnamocha and Eke (2015) discovered that al-
though long-term bank credit and industrial output
had a positive effect, only industrial output influenced
the level of agricultural output in the short term.
According to the research results, low agricultural in-
vestment will lead to low agricultural output, which
is due to low credit opportunity or high lending rate.
Adeola and Ikpesu (2016) showed that agricultural
output, trade credit and money supply were not co-
integrated. After sharing that the money supply and
the loans used for agriculture increased the agricultur-
al sector and the level of agricultural output in Nigeria,
they showed that the effects of the loans at this point
were very weak by the method of decomposition of
the variance they applied. Sogules and Nkoro (2016)
claimed that a long run relationship exists between
bank credits and manufacturing sector output.
In their study on the manufacturing sector in
Nigeria, John and Therhemba (2016) found that bank
loans, advances and large money supply had an im-
pact on the output of the manufacturing sector.
However, they stated that the output of the manufac-
turing sector is declining by being affected by high
inflation and high interest rates. Using quarterly data
of syndicated loans given by Italian banks, Dörr et. al.
(2017) stated that increases in the borrowing costs
and default rates of firms have led to severe decreases
in the assets side of the banks ‘ balance sheets. Strong
international banks have hesitated to lend money to
Italian firms with distressed balance sheets. The de-
cline in the productivity of the country led to a reduc-
tion in the country’s productivity, resulting from the
credit restrictions experienced as a result of this hesita-
tion. Companies that could not provide enough credit
from banks faced a decline in productivity as well as
investment and employment. Dörr et. al. (2017), who
revealed this situation with their findings, found that
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
78 South East European Journal of Economics and Business, Volume 14 (1) 2019
the reason why Italian firms could not achieve the
desired level of productivity was troubled banks and
credit supply shocks.
Ume et. al. (2017) in their study on the manufac-
turing sector identified each explanatory variable
(commercial bank loans, total savings, interest rate
and inflation rate) and subsequent lags as important
functions at 5%, excluding foreign exchange rates and
lags in the output of the sector. They argues that the
output of the manufacturing sector has a set rate in
order to reach equilibrium the behavior of the short-
term shocks in the long-term, has reached the conclu-
sion that this period is about 3 years. Vovchak (2017)
examined the relationships of the firms with the bank
during the crisis and the credit situation, stated that
it was very difficult for the firms to work with health-
ier banks by changing their banks during the crisis.
It has determined that banks applying core deposit
financing provided firms with a lower credit rating
than other loans during the crisis. Ramcharran (2017)
indicated increasing productivity (output elasticity)
of bank credit from 0.76 to 1.23 to small and medium
size industries sector. The sector’s efficiency improved
from returns to scale of −0.89 to 0.607. The main rea-
son for the increase in efficiency was the increase in
the productivity of the bank loan Dimelis et. al. (2017)
claims that the growth of firms over 2075 firms in the
euro zone before the crisis (2008 before the mortgage
crisis) is directly related to bank loans and moves in
the same direction. For the post-crisis period, it has
revealed that this relationship did not last long, and
that bank loans contributed only to the slow-growing
companies. However, for firms with a high growth
rate, this relationship has disappeared and no contri-
bution has been observed. Diallo (2018) analyzes and
shows that the bank’s productivity loosens credit re-
strictions and increases the growth rate of financially
hooked industries during the 2008 crisis.
As can be seen in the literature studies, it is found
that the industrial manufacturing sector is generally
taken into consideration in comparison with other
sectors. Besides, the only micro-based research in
which the manufacturing industry and bank loans
are handled in Turkey is the study of Demirci (2017).
However in this study, the manufacturing industry was
considered as a whole. In the econometric analysis of
Demirci (2017), which determined that the production
in the manufacturing industry sector acted together
with bank loans, monthly manufacturing industry
production index data were used between 1999-2015.
The monthly cash loan volume of the manufacturing
industry sector, which was given by domestic banks,
was included in the analysis and causality test appli-
cation was made. In the long term, the causality from
the production to the bank loan was determined and
it was emphasized that the financial sector followed
the real economy. The macro findings of the study
showed that the demand leading hypothesis supports
the results.
The existing studies show that bank credit has an
important and significant role in increasing econom-
ic growth and increasing importance of bank credit.
Therefore, finding the determinants of bank credit is
an issue that attracts attention all over the world. The
increasing use rate of bank loans by companies today
is an important issue that should be discussed in the
interaction with world economies. However, what
kind of a policy is to be applied for the world econo-
mies and the effect of bank loans on the manufactur-
ing sector is unfortunately a questioned answer with
an insufficient number of literatures. Although Kaldor
(1968) has already confirmed the existence of such an
influence and saw the industrial sector as the driving
force of economic growth, there are few studies on
the manufacturing sub-sectors. (Ajayi, 2000; Tawose,
2012; Sogules and Nkoro, 2016) This study will help to
cover this gap in the existing literature especially for
emerging markets.
4. MATERIALS AnD METHODS
4.1. Model Specification and Data Estimation
Procedure
In studies examining the effect of bank credit with
economic growth, Johansen Method was used by
Chisasa (2014), Chisasa and Makina (2015), Olokoyo
et al. (2016), Sogules and Nkoro (2016); OLS Method
was used by Chisasa and Makina (2013) VAR analy-
sis was used by Adeola and Ikpesu, (2016), and The
Generalised Method of Moments (GMM) estima-
tion was used by Nkurunziza (2010), Petkovski and
Kjosevski (2014). Following the studies; Ang, (2007);
Ma and Jalil, (2008); Hasanov and Huseynov, (2013);
Iyoboyi, (2013); Tripathi and Kumar, (2015); Abubakar
and Kassim, (2016); Ume at al., (2017) we prefer ARDL
approach due to its specific advantages over other
techniques.
In the application section of the study, two fre-
quently used methods were exercised in time series
analysis in literature. Firstly, to examine the long and
short-run impact of bank credits on manufacturing
sub-sectors, ARDL Bounds Testing Approach (ARDL)
applied. Although the equation based ARDL method,
which has different properties from the other cointe-
gration methods presented by Pesaran et. al. (2001)
in literature, does not take into account the number
of co-integrating relationships between the basic
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
79South East European Journal of Economics and Business, Volume 14 (1) 2019
variables and the weak exogeneity problem, it comes
with many advantages. It is an important advantage
that it allows estimation of short and long-term coef-
ficients simultaneously while making estimation with
least squares method. The most important advantage
of this is that it is independent of whether the regres-
sors are I (0) or I (1) or both, with the bouncing endog-
eneity problems and preference for studying variables
with small time series data samples. (Hasanov and
Huseynov, 2013).
The choice of the ARDL test was based on the ad-
vantages listed above. Due to the counted advantag-
es of the ARDL method, residual-based technique by
Engle and Granger, and the Full-Maximum Likelihood
(FML) test based on Johansen and on Johansen
and Juselius could not be preferred. (Nyasha and
Odhiambo 2017). Especially, for this study, less data
and small sample size are the main reasons for choos-
ing ARDL method according to other cointegration
methods.
After implementing ARDL method, diagnostic tests
are also checked out. Moreover, after the estimation,
the overall stability of the empirical model is checked
by using cumulative sum (CUSUM) and cumulative
sum of squares (CUSUMSQ) methods by Brown et al.
(1975). The CUSUM and CUSUMSQ tests do not re-
quire the determination of break points as it is in the
Chow test. Therefore, there is no need to determine
the breaking dates in advance.
Co-integration does not imply the direction of cau-
sality between two variables. Hence, after ensuring
the co-integration between bank credits and manu-
facturing industry index, their causal relationship is
examined onward. In this sense, Toda-Yamamoto (TY)
causality test was applied to investigate the causal re-
lationship between these two variables.
The reason for choosing Toda-Yamamoto cau-
sality method is that the lag lengths do not change,
the maximum lag with VAR technique is taken and
therefore the results are more reliable. The tradition-
al Granger causality method causes data loss in long
lags by changing the lag length. In this manner, Toda-
Yamamoto causality method provides superiority
over traditional Granger causality method. (Leshoro,
2017). In addition, TY technique removes the prob-
lem of power and size property in estimating unit root
test for long-run relationship (Rahman et al., 2017).
The main difference of this method from the others
is; there is not a requirement for the variables to be
stationary.
For utilization of the bounds test procedure, the fol-
lowing regressions are estimated for each sub-sector:
LIPI = f (LCRD, LLR)
LIPI refers to the industrial production index of each
sub-sector. In previous studies (e.g., Demetriades and
Hussein, 1996; Levine, 2002; Aslan and Kucukaksoy,
2006; Ang, 2008; Jalil et al., 2010; Hasanov and
Huseynov, 2013) it was stated that commercial bank
loans to the private sector were a method to accel-
erate economic growth compared to other types of
funding. Consequently, the principle independent
variable is LCRD, the total commercial bank loans
opened by the banking sector to the manufacturing
industry sub-sectors (Kelly and Everett, 2004; Larrain,
2006; Izhar and Tariq, 2009; Tawose, 2012; Buono and
Formai, 2014; Ebi and Emmanuel, 2014; Nnamocha
and Eke, 2015; Adeola and Ikpesu, 2016; Vovchak,
2017; Ramcharran, 2017). In the literature, other in-
dependent variable, lending rate (LLR), which is used
commonly in related studies (Ma and Jalil, 2008; Oni
et. al., 2014; Bayar and Tokpunar, 2014; Nnamocha and
Eke, 2015; Ume et al., 2017) expresses interest rates
that banks apply to commercial loans.
It is observed that many of the micro-studies
on the effects of bank loans have used sector GDP
as a dependent variable (Rajan and Zingales, 1998;
Nnamocha and Eke, 2015; Toby and Peterside, 2014;
Diallo, 2018). These measures can be seen as sum of
firm credit/GDP, private credit/GDP or stock market
capitalization/GDP. In addition, to show relationship
between credits and sector GDP growth, few empirical
studies use either sales growth or employment growth
or both (Bottazzi et al., 2001; Audretsch et al., 2004;
Covin et al., 2006; Coad and Rao, 2008; Giotopoulos
and Fotopoulos, 2010; Dimelis et al., 2016; Dimelis et
al., 2017). Moreover, Pham (2014) used the change in
R&D spending to measure the sector GDP growth in-
stead by scaling it over total assets and net sales.
Almost all micro studies on the effects of bank
credits in Turkey can be observed to have used GDP
as dependent variable (Kar and Pentecost, 2000;
Aslan and Kucukaksoy, 2006; Onder and Ozyıldırım,
2010; Kaya et al., 2013; Bayar and Tokpunar, 2014). It
is clear that GDP is a very general concept and is un-
der the influence of all other sectors. In this regard, it
is thought that in countries that do not publish their
sectoral GDP like Turkey, by bringing a different view-
point, using Industrial Production Index as Sub-sector
Industrial Production Index data in micro-based stud-
ies is more appropriate.
The selection of the Industrial Production Index
as a dependent variable instead of GDP is also in
line with Kaldor’s first law that shows that the indus-
trial sector is the “engine of growth”. According to
Kaldor (1968), the growth of the industrial sector in-
creases the level of efficiency not only in its own but
also in other sectors with its wide division of labor. It
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
80 South East European Journal of Economics and Business, Volume 14 (1) 2019
is more appropriate to measure the development of
the industry with the added contribution made to
the industry sector. For this purpose, the Industrial
Production Index is used to represent industrial de-
velopment in Turkey. This index gives the production
from various sub-sectors (Arisoy, 2008). This produc-
tion is the trigger of economic growth according to
Kaldor. According to his law, there is a positive rela-
tionship between the growth of the industrial sector
and economic growth from the first to the second. In
other words, the faster the industrial sector grows, the
faster the economy will grow. Therefore, the course of
the Industrial Production Index is very important and
chosen as the dependent variable. Very few studies
were found which take the same approach, Bahmani
and Saha (2016) and Demirci (2017), while the review
was taking place in the literature. However, Demirci
(2017) is only used Industrial Production Index just for
manufacturing industry sector in general. In this re-
spect, this research is different from other studies and
it is a contribution to the literature.
Moreover, Costantini (2013) explains the impor-
tance of the Industrial Production Index in his study.
According to Costantini (2013), industrial production
is the most important indicator used to explain total
business cycle fluctuations. In this point of view, cy-
clical indicators of the manufacturing sector -such
as sector-GDP- can be obtained from the Industrial
Production Index Series. In fact, industrial production
is an index that has been researched and estimated by
many previous studies.
Independent variable, LCRD (the total commer-
cial bank loans opened by the banking sector), is the
main source of funding for the country’s economy,
manufacturing industry and the private sector, espe-
cially in developing countries. For instance, the credit
volume allocated to the sub-sectors of the manufac-
turing industry is an important variable that can dem-
onstrate the sector’s financial intermediation services
and the transfer of such activities to other productive
sectors through investment expenditure. According
to Nkurunziza (2010), the use of bank loans can af-
fect firm growth in two ways, positive and negative.
Adverse economic conditions, instability and macro-
economic fluctuations have a negative effect on the
firm’s ability to repay its loans. However, if the loans al-
low the firm to solve its liquidity problem, increase its
profitability and expand its expansion, the use of bank
loans will positively affect its growth.
Using bank loans as explanatory variable, Chisasa
(2014) explains that liquidity-enhancing loans will
allow for relaxation, particularly in input payments.
Development of new technologies can be helpful with
the use of credit, increasing the technical equipment
and production technology can be achieved. Finally,
by increasing the density of fixed inputs of loans, more
efficient use of resources will increase production out-
put and increase profitability. All these reasons are
considered as the reason why bank loans are explana-
tory variable by Chisasa (2014).
In terms of showing financial costs, Lending Rate
(LLR) that banks apply to commercial loans can be
considered an important variable. Since the high lend-
ing rate will increase financial costs, it also affects the
savings to be transferred directly to the investment.
Low rate, low cost, high savings, high investment
and consequently, it leads to an increase in econom-
ic growth. Otherwise, a negative relationship among
growth and interest rates could occur in manufactur-
ing sub-sectors. (John and Therhemba, 2016; Ume et
al., 2017)
Banks’ lending rate (LLR) is an important variable
that affects the decision of firms. Will the company
borrow from the banking sector or explore other
sources of finance? LLR is the most important argu-
ment for this decision.
King and Levine (1993) suggest a positive relation-
ship among growth, interest rates and financial depth.
Inspired by McKinnon and Shaw, Ma and Jalil (2008)
highlighted the importance of savings. They empha-
sized that investments will increase with the effective
distribution of the resources caused by the savings. Ma
and Jalil (2008) pointed out that governments should
pursue policies to implement high interest rates to
increase savings incentive. They emphasized that the
acceleration of economic growth will be through the
transfer of savings to be collected by the high inter-
est rate. They argued that financial depth would come
with a well-managed interest rate policy and positive
interest rate, and consequently increase economic
growth.
The model is estimated using monthly time-series
of Turkish data between 01/2010 and 09/2017. Data
are taken from Banking Regulation and Supervision
Agency (BRSA) and the Central Bank of the Republic
of Turkey (CBRT) websites. In this study, sub-sectors
of industrial production indices, which can be found
as well as credit numbers, were investigated. The
sub-sectors studied in this matching are Mining and
Quarrying (MQ), Food and Beverage (FB), Textile and
Clothing (TC), Wood and Furniture (WF), Paper (PP),
Chemistry (CH) and Machinery (MC).
As the study uses monthly time series, it optimized
to a maximum lag length of 12 periods. The optimal
lag length of each variable that enters the model is
chosen based on the Schwarz-Bayesian selection
criterion.
Turkish industrial production index data are
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
81South East European Journal of Economics and Business, Volume 14 (1) 2019
provided in the form of 2010 = 100. Accordingly, other
two variables also were brought to the 2010 = 100 lev-
els. Before analysis, the data were brought to logarith-
mic form and seasonally adjusted by the Census X12
procedure.
5. EMPIRICAL RESuLTS
Following the standard procedures for variables
with time series properties, it is considered that first
step should be the statistical features of the series
starting with the descriptive statistics (see Table 2).
ARDL bound test is based on one of the F-statistic
tests, Wald Test. Pesaran et al., (2001) has claimed two
critical values, called lower and upper critical values,
for testing co-integrating relationship among the vari-
ables examined in researches. According to this argu-
ment, the lower bound critical values assume no co-
integrating relationship. Thus all variables included in
the analysis are I(0).On the other hand upper bound
critical values reject null of no co-integration and
therefore all variables are I(1). Null of no co-integration
is rejected, in case the calculated Wald test (F-statistic)
is greater than upper bound critical value .Null of no
co-integration is not rejected if calculated F-statistic
is less than lower critical bound value. Pesaran et al.,
(2001) concludes that results are inconclusive if calcu-
lated F-statistic falls between upper and lower bound
critical values. It’s been used the Schwarz-Bayesian
selection criterion for selecting optimal lag length be-
cause it is useful for small sample size. The F-statistic
bound test results are shown in Table 3.
Schwarz Bayesian Criteria used in order to deter-
mine appropriate lag structure of ARDL procedure.
Appropriate ARDL Model for Machinery is (2, 0, 0) and
F-statistic is 5.649 which exceed upper bound critical
value at %5 level. Estimated ARDL model are (1,1,0),
(1,4,2), (2,0,1), (5,0,7), (2,2,2) and (2,0,0) with F statistics
are 18.719, 34.107, 6.764, 6.263, 9.406 and 7.505 for
Table 2: Descriptive Statistic and Correlations
Sectors Descriptive Statistic Correlation Matrix
Mean Std.Dev. Min. Max. LLR LIPI LCRD
MQ
LCRD 232.6776 108.557 91.29 422.29 LLR 1.000 .413 .706
LIPI 112.3896 14.708 73.4 142.94 LIPI 1.000 .470
LLR 143.229 28.045 94.75 189.65 LCRD 1.000
FB
LCRD 216.681 80.815 81.63 383.69 LLR 1.000 .496 .718
LIPI 114.0635 14.568 81.69 138.26 LIPI 1.000 .584
LLR 143.229 28.045 94.75 189.65 LCRD 1.000
TC
LCRD 245.743 104.247 85.79 436.66 LLR 1.000 .418 .709
LIPI 106.761 7.483 90.78 121.75 LIPI 1.000 .477
LLR 143.229 28.045 94.75 189.65 LCRD 1.000
WF
LCRD 261.0549 105.188 82.64 462.83 LLR 1.000 .480 .708
LIPI 118.5874 15.004 85.38 150.82 LIPI 1.000 .673
LLR 143.229 28.045 94.75 189.65 LCRD 1.000
PP
LCRD 169.4937 50.425 88.35 289.3 LLR 1.000 .686 .711
LIPI 123.6153 16.650 88.34 156.24 LIPI 1.000 .859
LLR 143.229 28.045 94.75 189.65 LCRD 1.000
CH
LCRD 232.8666 99.590 83.31 421.36 LLR 1.000 .554 .715
LIPI 115.1482 12.693 83.95 147.31 LIPI 1.000 .702
LLR 143.229 28.045 94.75 189.65 LCRD 1.000
MC
LCRD 208.9306 81.604 87.8 380.41 LLR 1.000 .716 .526
LIPI 133.6957 19.847 74.37 174.83 LIPI 1.000 .608
LLR 143.229 28.045 94.75 189.65 LCRD 1.000
Mining and Quarrying (MQ), Food and Beverage (FB), Textile and Clothing (TC), Wood and Furniture (WF), Paper (PP),
Chemistry (CH), Machinery (MC), the industrial production index of sub-sectors (LIPI), the total commercial bank loans
(LCRD), interest rates (LLR)
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
82 South East European Journal of Economics and Business, Volume 14 (1) 2019
MQ, FB, TC, WF, PP and CH. Consequently, the null hy-
pothesis is rejected thus it is concluded that long-run
relationship exists among variables.
To understand the short-term adjustment process,
it’s needed to look at the sign and the magnitude of
the coefficient of the error correction term (ECT). If the
value of this coefficient is between 0 and –1, the cor-
rection to the model in the period is a fraction of the
error in period t–1. If the value is between –1 and –2,
then the ECT will produce damped oscillations in this
model about its equilibrium path (Alam and Quazi,
2003). The study’s short run adjustment period ac-
cording to the error correction term that is negative
and statistically significant. The negative value shows
that there exists an adjustment speed from short-run
disequilibrium towards the long-run equilibrium. In
this study, ECM term coefficients are negative for all
sub-sectors.
The long run coefficients of variables are displayed
in Table 4. It observed at first glance that coefficient
of bank credit is positively significant for all sub-sec-
tors except MQ. The results suggest that an increase
in bank credit leads to raise industrial production in-
dex in FB, TC, WF, PP, CH and MC. On the other hand,
the lending rate is positively correlated with industrial
production index in only MQ and WF. It means that,
bank credit is more effective than lending rate on in-
dustrial production index of sub-sector.
In the long run, relationship was observed between
the bank credits and the industrial production index
in FB, TC, WF, PP, CH and MC sub-sectors. However, in
the short run there is no significant relationship was
observed in such sectors. The lending rate is positively
correlated with industrial production index in only
MQ and WF sub-sectors in the long-run. Despite, in
the short-run, only FB sub-sector’s industrial produc-
tion index negatively correlated with interest rates.
The fact that none of the sectors affected in the short-
and long-term parameters with each other, indicates
that each sector has its own structural characteristics.
In the long run, the absence of a relationship be-
tween bank lending rates and industrial production in
most sub-sectors may be associated with the structur-
al characteristics of the sector. The capital structure of
the manufacturing industry sector firms is managed
by more fixed assets. This may be the reason of fail-
ure to find a meaningful long-term relationship with
credit interest rates. Moreover, long term interest rates
of credits always higher than short term interest rates.
Highness of long-run interest rates may be one of the
reasons why the relationship is positive in two sub-
sectors and not in others.
The short run parameters in Table 4 indicate that
bank credit and lending rate are related to industrial
production index. According to findings, industrial
production index is negatively affected by bank credit
only on MQ and lagged values of bank credit on FB
sub-sectors. In other sub-sectors, no significant rela-
tionship was observed between the bank credits and
the industrial production index. On the other hand,
industrial production index is significantly affected
in a negative way by bank credit lending rates only
in FB sub-sector. In addition, industrial production
is significantly affected positively by lagged value of
Table 3: Estimated ARDL models and Bound F-test.
Sectors Model F-Stat. ECM(-1) Critical Value %1 Critical Value %5
MQ (1, 1, 0) 18.719 -0.858 I (0) = 4.13 I (0) = 3.1
[0.000] I (1) = 5 I (1) = 3.87
FB (1, 4, 2) 34.107 -1.189 I (0) = 4.13 I (0) = 3.1
[0.000] I (1) = 5 I (1) = 3.87
TC (2, 0, 1) 6.764 -0.788 I (0) = 4.13 I (0) = 3.1
[0.000] I (1) = 5 I (1) = 3.87
WF (5, 0, 7) 6.236 -0.913 I (0) = 4.13 I (0) = 3.1
[0.000] I (1) = 5 I (1) = 3.87
PP (2, 2, 2) 9.406 -0.872 I (0) = 4.13 I (0) = 3.1
[0.000] I (1) = 5 I (1) = 3.87
CH (2, 0, 0) 7.505 -0.815 I (0) = 4.13 I (0) = 3.1
[0.000] I (1) = 5 I (1) = 3.87
MC (2, 0, 0) 5.649 -0.321 I (0) = 4.13 I (0) = 3.1
[-0.815] I (1) = 5 I (1) = 3.87
Critical values are obtained from (Pesaran et al., 2001). Numbers in brackets are p-values.
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
83South East European Journal of Economics and Business, Volume 14 (1) 2019
bank credit lending rates in WF sub-sector. Although
some of the test values are not significant, their nega-
tive signs give us a clue. Additionally, any significant
relation between the lending rates and the industrial
production index of other sub-sectors could not be
found.
The absence of a relationship between bank lend-
ing rates and industrial production could be because
of various reasons. In some emerging economies,
although interest rates are favourable, there are a
number of other factors limiting the credit supply of
enterprises, especially through banks. These include
many factors such as the high level of collateral and
insurance costs, the legal structure of enterprises, the
age of the enterprises, the attitude of the banks, and
the excess of formalities of providing credit. (Arıcay
and Kok, 2009) According to the findings of Bhaird
and Lucey (2006), long-term borrowing has a nega-
tive relationship with the enterprise age. This result
indicates that the funds formed in the enterprise over
time reduce the borrowing necessity. In addition,
the study conducted by the Government of Scotland
(2008) examined the difficulties faced by businesses
in accessing bank financing. The results of the study
found that almost half of the firms seeking bank loans
were close to using alternative financing (grants, soft
loans and equity financing). The consultancy problem
and high accounting costs experienced in applying to
Table 4: The results of long and short run
The results of long and short run
LIPI MQ FB TC WF PP CH MC
Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient
Panel A. Long Run Estimates
Cons. 3.967* *3.641 4.117* 4.379* 2.768* 3.806* 4.189*
LCRD 0.102 *0.18 0.063* 0.24* 0.427* 0.165* 0.219*
LLR 0.041* 0.03 0.042 0.189** -0.024 0.0106 -0.086
Panel B. Short Run Estimates
LIPI (-1) – – *-0.313 -0.286 *-0.348 **-0.234 *-0.478
LIPI (-2) – – – 0.181 – – –
LIPI (-3) – – – 0.423** – – –
LIPI (-4) – – – 0.250** – – –
LCRD ***-0.25 -0.106 0.052 0.125 0.027 0.13 -0.021
LCRD (-1) – *-0.392 – – 0.21 – –
LCRD (-2) – -0.198 – – – – –
LCRD (-3) – *-0.455 – – – – –
LLR -0.032 **-0.154 -0.131 -0.107 -0.089 0.007 -0.117
LLR (-1) – *-0.213 0.154 -0.123 – –
LLR (-2) – – – 0.355** – – –
LLR (-3) – – – 0.124 – – –
LLR (-4) – – – 0.016 – – –
LLR (-5) – – – 0.084 – – –
LLR (-6) – – – 0.392* – – –
Panel C. Diagnostic Statistics
Serial 1.078 [0.39] 1.013 [0.447] 0.996 [0.46] 0.684 [0.759] 0.754 [0.693] 0.407 [0.956] 0.349 [0.976]
Arch 0.462 [0.929] 0.324 [0.982] 0.343 [0.977] 0.627 [0.81] 1.322 [0.227] 0.42 [0.95] 0.28 [0.99]
Ramsey 0.196 [0.658] 1.698 [0.196] 0.90 [0.345] 0.088 [0.766] 0.125 [0.724] 1.764 [0.187] 1.634 [0.204]
CUSUM Stabil Stabil Stabil Stabil Stabil Stabil Stabil
CUSUMQ Stabil Stabil Stabil Stabil Stabil Stabil Stabil
*, ** and *** indicate statistical significance at 10, 5 and 1% level respectively. Diagnostic tests results based on F-statistic,
numbers in brackets are p-values.
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
84 South East European Journal of Economics and Business, Volume 14 (1) 2019
banks were among the main complaints of the firms.
It has been determined that the lack of collateral and/
or the lack of trading record has prevented the bank
from receiving loans. The study determined that the
firms affected by these reasons accounted for about
one third of the non-creditors.
The current value of the industrial production in-
dex is not related to the sub-sectors growth. The one-
period lagged value of the industrial production index
is adversely affected by the growth of the PP, TC, CH
and MC sub-sectors. When the causes of this finding
are investigated, it is thought that delay may be a loss
of time. Considering the return on investment period,
it will take time for goods and services to emerge. The
current volume of goods and services can only be ex-
plained by past financial resources used in production.
Same situation occurs on one-period lagged value of
lending rates. It is negatively related to PP sub-sector
growth while current value is not.
Checking stability properties of parameters is also
a very important issue. This study used CUSUM and
CUSUMQ tests for examining the stability properties.
The CUSUM test is based on cumulative error terms
associated with the observation set and is drawn be-
tween two critical points showing 5 percent signifi-
cance. In the tests, it was observed that the CUSUM
and CUSUMQ test statistics remained within critical
limits at 5% significance level. So, the null hypothesis
is accepted and the model is stable. This means that
the estimated parameters were stable during the pe-
riod of examination.
Diagnostic tests results are shown in Table 4. The
result of Breusch-Godfrey LM test rejects serial cor-
relation for the equations. ARCH test results support
that residuals are homoscedastic for all sub-sectors.
Finally, Ramsey-Reset test results confirm the correct
functional form.
While positive correlation was observed between
bank credits and industrial production index in some
of the studies (Kelly and Everett, 2004; Tawose, 2012;
Manova, 2013; Toby and Peterside, 2014; Buono and
Formai, 2014; Ebi and Emmanuel, 2014; Nnamocha
and Eke, 2015; Sogules and Nkoro, 2016; John and
Therhemba, 2016; Dörr et al., 2017; Ume et al., 2017;
Vovchak, 2017; Ramcharran, 2017; Dimelis et al., 2017)
negative correlation was observed in others (Larrain,
2006; Ma and Jalil, 2008; Izhar and Tariq, 2009; Toby
and Peterside, 2014; Chisasa, 2014). In this research
we have reached to a conclusion which supports both
findings. There was a positive relationship between
bank loans and industrial production index in the WF
sub-sector, while a negative relationship was found
in the MQ and FB sub-sectors. The situation is chang-
ing for the lending rate and the industrial production
index. For the FB, TC, WF and MC sub-sectors, a nega-
tive relationship between interest rates and GDP vari-
ables for (Ma and Jalil, 2008; Ebi and Emmanuel, 2014;
Nnamocha and Eke, 2015; John and Therhemba, 2016;
Ume et al., 2017) were also identified. The study could
not confirm the positive relationship between interest
rates and growth variables. But for CH sub-sector, hav-
ing a positive sign countable through studies are only
Ma and Jalil (2008) and Tawose (2012).
After the ARDL test, Toda-Yomomato (TY) test was
performed. TY technique is performed through two
steps; First, the maximum lag length (k) determined by
using either the Schwarz Information Criteria (SIC) and
the maximum order of integration (d). Secondly, cau-
sality through the Wald test is determined. The modi-
fied Wald (MWALD) test is adopted in this technique
in order to restrain the parameters of the VAR model
along with the asymptotic chi-square distribution.
According to the TY test results Table 5), in all the
sectors except Machinery (MC), there is a degree of
significance towards the industrial production in-
dex from bank credits. In addition to the results, in
all other sectors except MC and CH sectors, causality
relation was observed at different rates from credit
Table 5: Toda-Yomamato Causality Results
Sectors LCRD » LIPI LLR » LIPI
Wald Stat Asym P-Value Wald Stat Asym P-Value
MQ 31.239*** 0.002 39.75*** 0
FB 20.987* 0.051 25.655*** 0.007
TC 59.206*** 0 36.627*** 0
WF 80.204*** 0 23.683** 0.022
PP 46.607*** 0 19.433* 0.079
CH 25.273** 0.014 15.47 0.162
MC 9.278 0.596 0.009 0.925
*, ** and *** indicate statistical significance at 10, 5 and 1% level respectively. Max lag length criteria are taken 12.
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
85South East European Journal of Economics and Business, Volume 14 (1) 2019
interest rates to industrial production index. There are
also some studies in the literature in which the causal-
ity relation from bank credits to industrial production
index is partially or completely observed (Toby and
Peterside, 2014; Chisasa, 2014; Rahman et al., 2015;
Stolbov, 2017; Qamruzzaman and Jianguo, 2017)
However, there are also studies in which the causality
relationship from bank loans to industrial production
index cannot be found (Dal Colle, 2011; Marques et
al., 2013; Tripathi and Kumar, 2015; Adeola and Ikpesu,
2016).
The casual relationship between the industrial pro-
duction index and the bank credit volume and credit
interest rates give clues that production is financed
by bank loans. If this relationship is positive, the low
credit interest rates will increase the use of credits and
the increase in the banks’ lending volume will increase
the production. The process will be reversed if the re-
lationship is negative.
Evidence of supply leading hypothesis has been
found whereby financial sector is leading and causing
economic growth. There was positive relationship be-
tween loans to the private sector and industrial pro-
duction index growth, in the case of Turkey. It reveals
that it was the financial system which would create
various types of economic growth to which the secto-
ral and sub-sectoral would respond.
As to continue on macro base findings; an impor-
tant idea states that interest rate is positively corre-
lated with savings under financially repressing econ-
omies especially in developing countries. In these
economies the higher interest rates encourage sav-
ings and decrease consumption which is called sub-
stitution effect. Consequently, higher interest rates
increase income for those people with high levels of
savings, which is called income effect (Ma and Jalil,
2008). While this study’s short-run findings support
substitution effect in one sector significantly, it does
not support substitution and income effects in certain
sub-sectors. According to findings, on MQ and FB sub-
sectors’ bank credits have negative effect on growth
which means substation effect is not an issue. On the
meantime, on WF sub-sector’s bank credits have posi-
tive effect on growth which means substation effect
could occur. On the other hand, FB, TC, WF and MC
sub-sectors’ interest rates have negative effect on
growth which does not support income effect.
6. COnCLuSIOnS AnD POLICY IMPLICATIOnS
This paper examines the relationship between
manufacturing industry sub-sector growth and
financial sectors’ bank credits of Turkey, an advanced
emerging market of the world, over the period be-
tween 2010 and 2017. The data has been analyzed by
using Auto Regressive Distributed Lag (ARDL) model
and Toda Yomamato Causality test to capture the na-
ture of relationship between seven manufacturing
industry sub-sectors including Mining and Quarrying
(MQ), Food and Beverage (FB), Textile and Clothing
(TC), Wood and Furniture (WF), Paper (PP), Chemistry
(CH), and Machinery (MC) in Turkey context.
In this study, industrial production index (LIPI) was
used instead of the GDP data of the manufacturing
industry sub-sectors. It clearly brought a contribution
to literature as a dependent variable. Independent
variables of research are total commercial bank loans
(LCRD) opened by the banking sector to the manufac-
turing industry sub-sectors, and interest rates (LLR)
applied by commercial banks to commercial credits.
It would not be appropriate to compare the find-
ings of this study with the findings of other previous
studies. Because this study was carried out on the ba-
sis of manufacturing industry sub-sectors and indus-
trial production index which is used as an indicator of
economic development and sector GDP. The use of in-
dustrial production index of sub-sectors could not be
found in other studies. In this respect, this research is
unique and different from other studies and it is a con-
tribution to the literature.
Findings of this study support that bank credits are
more effective than loan rates on industrial produc-
tion index of sub-sectors in the long-run. Moreover,
an increase in bank credit leads to the rise of indus-
trial production index. On the long-run parameters,
bank credit is positively correlated with industrial
production index except for Mining and Quarrying
sub-sectors. In addition, on short-run findings, indus-
trial production index is negatively affected by bank
credit only on Mining and Quarrying and lagged val-
ues of bank credit on Foods and Beverages sub-sector.
According to the Toda Yomamato causality test results,
in all the sub-sectors except Machinery, there are dif-
ferent degrees of causalities in the level of significance
from bank loans to industrial production index. On
the other hand, in all sub-sectors except Machinery
and Chemical sub-sectors, causality relation was ob-
served at different grades from loan interest rates to
industrial production index.
One of the macro-based results of this study is find-
ing evidence of supply leading hypothesis whereby
financial sector is leading and causing economic
growth. Another result of this study supports substi-
tution effect in one sector while it does not support
substitution and income effects in certain sectors.
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
86 South East European Journal of Economics and Business, Volume 14 (1) 2019
Following implications can be deduced from the
study:
1. The causality between the industrial production in-
dex and the bank credit volume and credit interest
rates emphasizes the significant linkage of produc-
tion with bank lending. This connection indicates
that production can be increased by improving
credit conditions.
2. Results reveal that industrial sub-sector managers
should be very careful about their financing deci-
sions in Turkey. Any consequences of adequate re-
sources, sub-sector firms with higher outstanding
total debt and higher capital intensity may more
adversely affected. Hence, credit supply shocks
could significantly affect the firms’ investment deci-
sions and productivity. (Dörr et al., 2017) External
financing is another option that managers could
follow in case of high rates and shortage of loans in
Turkey.
3. Another implication could be about banking prac-
tices. The fact that existence of causal relationship
towards growth from loans, especially the situation
that comes up after the Asian financial crisis which
can be summarized as credit rationing, does not ex-
ist in Turkey.
4. Macro-based implications; firstly, performing
policies and future institutional reforms in Turkey
must enable more efficient allocation of resources.
Accordingly, it is suitable for the Turkish govern-
ment to design policies that will encourage banks
to create an enabling environment to distribute
credits by making more funds available for the
manufacturing industrial sub-sector as so this will
increase the level of industrial output in the coun-
try and contribute to increased economic growth.
5. Having found banks as actual financier of manu-
facturing industrial sub-sectors, there is a need to
promote the banking sector in Turkey. Appropriate
policies should be implemented by the Central
Bank of the Republic of Turkey and other monetary
authorities alike and should be pursued by the
strengthening of the banking sector.
6. Recommendation could be given to Turkish gov-
ernment and Central Bank authorities on determi-
nation of interest rates. Policies that lower interest
rates (cost of capital) should be pursued by govern-
mental agencies.
7. Future researches should focus on other sub-sec-
tors of the economy. Another point that should not
be forgotten is that the findings of this study may
be unique. It should be emphasized that the sub-
sector findings cannot be generalized, that each
sector should be evaluated within its own dynam-
ics, and that the relationship between bank loans
and industrial production index growth may only
belong to that sub-sector.
8. It is important to stress that the findings of this
study may be specific to Turkish manufacturing in-
dustry sub-sectors. Results cannot be generalized
across countries. More detailed analysis is required
to explain the uncovered country-level causal
patterns.
9. It may also be possible to test for causality in a
time-series framework using alternative economet-
ric techniques.
Endnote: __________________________________
1 The credit channel operates through two mechanisms:
the balance sheet and the bank credit channel (Bernanke
and Gertler, 1995; Kashyap and Stein, 2000).
2 The monetary policy implemented by the monetary
authorities is called monetary transmission mechanisms
in the process of influencing the real economy. Monetary
transmission mechanisms are grouped under five headings:
interest rate channel, exchange rate channel, asset prices
channel, expectations channel and credit channel (Mishkin,
1995; 1996; and 2001).
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
87South East European Journal of Economics and Business, Volume 14 (1) 2019
REFEREnCES
Abubakar, A. and Kassim, S., 2016. Sectoral Impact of Bank
Credit in Malaysia: ARDL Modelling Approach Pertanika
Social Sciences & Humanities. 24 (S): 205 – 211.
Adeola, O. and Ikpesu, F., 2016. An empirical investigation
of the impact of bank lending on agricultural output in
Nigeria: A vector autoregressive (VAR) approach. The
Journal of Developing Areas, 50(6): 89-103.
Adve, S., 1980. Financial Practices in Indian Corporate Sector,
Inter-Group and Inter-Size Differences. Economic and
Political Weekly, 15(8): M2-M10.
Ajayi, D. D., 2000. The Determinants of the Volume of
Production Subcontracting in Nigeria. Nigeria Journal
of Economic and Social Studies (NJESS), 42 (1): 95-111.
http://www.africabib.org/rec.php?RID=260155500
Ak, M., Altıntaş, N. and Şimşek, A., 2016. Türkiye’de finansal
gelişme ve ekonomik büyüme ilişkisinin nedensellik an-
alizi. Doğuş Üniversitesi Dergisi, 17(2): 151-160.
Alam, I. and Quazi, R., 2003. Determinants of capital flight:
An econometric case study of Bangladesh. International
Review of Applied Economics, 17(1): 85-103.
Ang, J. B., 2007. Financial deepening and economic devel-
opment in Malaysia. Economic Papers, 26(3): 249-260.
Ang, B. J., 2008. What are the mechanisms linking finan-
cial development and economic growth in Malaysia?
Economic Modeling, 25 (1): 38 –53
Ansari, M. I., 2002. Impact of Financial Development, Money,
and Public Sending on Malaysian National Income: An
Econometric Study. Journal of Asian Economics, 13(1):
72-93.
Ayrıçay, Y., and Kök, D., 2009. Factors Affecting the
Level of Bank Loans in SMEs Finance: A Study in
Kahramanmaraş. Mustafa Kemal University Journal of
Social Sciences Institute, 6 (11): 109-131
Arisoy, I., 2008. Türkiye›de Sanayi Sektörü-İktisadi Büyüme
İlişkisinin Kaldor Hipotezi Çerçevesinde Test Edilmesi.
Turkish Economic Association Discussion Paper. 2008
(1): 1-31.
Aslan, Ö. and Küçükaksoy, İ., 2006. Finansal gelişme ve
ekonomik büyüme ilişkisi: Türkiye ekonomisi üzerine
ekonometrik bir uygulama. İ.Ü. Ekonometri ve İstatistik
Dergisi, (4): 25-38.
Audretsch, D. B., Klomp, L., Santarelli, E. and Thurik, A. R.,
2004. Gibrat›s law: are the services different?. Review of
Industrial Organization, 24(3): 301-324.
Aziz, J. and Duenwald, C., 2002. Growth-financial interme-
diation nexus in China. International Monetary Fund
Working Paper, No. 2002-2/194: 3-14.
Bahmani-Oskooee, M. and Saha, S., 2016. Asymmetry
Cointegration Between The Value Of The Dollar and
Sectoral Stock Indıces in The US. International Review of
Economics & Finance, 46: 78-86.
Banerjee, K., 2012. Credit and growth cycles in India: An
empirical assessment of the lead and lag behavior. RBI
Working Paper Series, (DEPR: 22/2011) (No. id: 4699).
BAT, 2017. The Bank Association of Turkey (BAT) Web Site,
https://www.tbb.org.tr/en/home
Bayar, G. and Tokpunar, S., 2014. Determinants of Turkish
Manufacturing Industry Production in Sub-Sectors: A
Panel Data Analysis. Business and Economics Research
Journal, 5(1): 67-85.
Beck T., Cull R., and Jerome A. T., 2005. Bank Privatization and
Performance: Empirical Evidence from Nigeria. World
Bank Policy Research Working Paper No: 3511.
Beck, T., Levine, R. and Loayza, N. , 2000. Finance and the
Sources of Growth. Journal of financial economics, 58(1):
261-300.
Bernanke, B. S. & Gertler, M. 1995. Inside The Black Box: The
Credit Channel of Monetary Policy Transmission. Journal
of Economic Perspectives, 9, 4: 27-48.
Bhaird C. ve Lucey B., 2006. Capital Structure and the
Financing of SMEs: Empirical Evidencefrom an Irish
Survey. Entrepreneurship: Occupational Choice and
Financing Conferance, 6-7 June, Copenhagen: Centre for
Economic and Business Research. (6)
BIS, 2018. The Bank for International Settlements (BIS) Web
Site, https://stats.bis.org/statx/srs/table/f2.4
Bloch, H. and Tang, S. H.K., 2003. The role of Financial
Development in Economic Growth. Progress in
Development Studies, 3(3): 243-251.
Bottazzi, G., Dosi, G., Lippi, M., Pammolli, F. and Riccaboni, M.,
2001. Innovation and corporate growth in the evolution
of the drug industry. International Journal of Industrial
Organization, 19(7): 1161-1187.
Boyreau-Debray, G., 2003. Financial Intermediation and
Growth: Chinese Style. World Bank Policy Research
Working Paper, No. 3027: 1-23.
Brown, R. L., Durbin, J. and Evans, J. M., 1975. Techniques for
testing the constancy of regression relationships over
time. Journal of the Royal Statistical Society, Series B
(Methodological): 149-192.
BRSA, 2017. Banking Regulation and Supervision Agency
(BRSA) Web Site https://www.bddk.org.tr/WebSitesi/
English.aspx
Buono, I., and Formai, S., 2014. Bank Credit and Firm Export:
Is there really a link. FREIT Working Paper, 1-20. http://
www.etsg.org/ETSG2014/Papers/041
Çamoğlu, S. M. & Akinci, M. 2012. Türkiye’de Sektörel Banka
Kredilerinin Gelişimi: Bir Zaman Serisi Analizi. Yönetim ve
Ekonomi: Celal Bayar Üniversitesi İktisadi ve İdari Bilimler
Fakültesi Dergisi, 19, 1: 193-210.
CBRT. 2018. Financial Stability Report-II. The Central Bank of
the Republic of Turkey (CBRT) Web Site. Available at
http://www.tcmb.gov.tr/wps/wcm/connect/7be5ffea-
3a41-4533-933a-db63c8943bdc/Fir_TamMetin27 ?M
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
88 South East European Journal of Economics and Business, Volume 14 (1) 2019
OD=AJPERES&CVID=&CACHE=NONE&CONTENTCACHE
=NONE
CBRT. 2017.The Central Bank of the Republic of Turkey
(CBRT) Web Site. Available at http://www.tcmb.gov.tr/
wps/wcm/connect/en/tcmb+en
Ceylan, S. and Durkaya, M., 2010. Türkiye’de Kredi Kullanımı-
Ekonomik Büyüme İlişkisi. Atatürk Üniversitesi İktisadi ve
İdari Bilimler Dergisi, 24(2):21-35
Chakraborty, I., 2008. Does financial development cause eco-
nomic growth? The case of India. South Asia Economic
Journal, 9(1): 109-139.
Chen, H., 2006. Development of financial intermediation
and economic growth: The Chinese experience. China
Economic Review, 17(4): 347-362.
Chisasa, J., 2014. An empirical study of the impact of bank
credit on agricultural output in South Africa (Doctoral
dissertation).University of South Africa (Unisa), http://
uir.unisa.ac.za/bitstream/handle/10500/18511/thesis_
chisasa_j ?sequence=1
Chisasa, J. and Makina, D., 2013. Bank Credit and Agricultural
Output in South Africa: A Cobb-Douglas Empirical
Analysis. International Business & Economics Research
Journal. 12 (4): 387-398.
Chisasa, J. and Makina, D., 2015. Bank Credit And Agricultural
Output In South Africa: Cointegration, Short Run
Dynamics And Causality. The Journal of Applied Business
Research. 31 (2): 489-500.
Chow, W. W., Fung, M. K. & Man-Kwong Leung. 2018.
Finance–growth nexus in China from an endogenous
switching perspective, The Journal of International Trade
& Economic Development, 27, 4: 443
Christopoulos, D. and Tsionas E., 2004. Financial
Development and Economic Growth: Evidence from
Panel Unit Root and Cointegration Tests. Journal of
Development Economics, 73: 55-74.
Coad, A. and Rao, R., 2008. Innovation and firm growth
in high-tech sectors: A quantile regression ap-
proach. Research Policy, 37(4): 633-648.
Costantini, M., 2013. Forecasting the industrial production
using alternative factor models and business survey
data. Journal of Applied Statistics. 40 (10): 2275-2289.
Covin, J. G., Green, K. M. and Slevin, D. P., 2006. Strategic
process effects on the entrepreneurial orientation–sales
growth rate relationship. Entrepreneurship Theory and
Practice, 30(1): 57-81.
Dal Colle, A., 2011. Finance–growth nexus: does causality
withstand financial liberalization? Evidence from cointe-
grated VAR. Empirical Economics, 41(1): 127-154.
Das, P. K. & Maiti, P. 1998. Bank credit, output and bank de-
posits in West Bengal and selected states. Economic and
political Weekly,33, 47/48: 3081-3088.
De Gregorio, J. and Guidotti, P. E., 1995. Financial Development
and Economic Growth. World Development, 23(3):
433-448.
Demetriades, P. O. and Andrianova, S., 2003. Finance
and growth: what we know and what we need to
know. Financial Development and Economic Growth,
Palgrave Macmillan UK: 38-65 https://lra.le.ac.uk/bit-
stream/2381/4424/1/dp03-15 (25-01-2018)
Demetriades, P. O. and Hussein A. K., 1996. Does Financial
Development Cause Economic Growth? Time Series
Evidence from 16 Countries. Journal of Development
Economics, 51(2): 387-411.
Demetriades, P. O. and James, G. A., 2011. Finance and growth
in Africa: The broken link. Economics Letters, 113(3):
263-265.
Demirci, N. S., 2017. Relationship between Production
and Bank Credits in Manufacturing Industry Sector:
Cointegration and Causality Analysis for Turkey. Dokuz
Eylül Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 19(1):
35-61.
Demirguc-Kunt, A., and Levine R., 2008. Finance, Financial
Sector Policies and Long Run Growth. The World Bank
Development Research Group, Policy Research Working
Paper No: 4469.
Diallo, B., 2018. Bank efficiency and industry growth during
financial crises. Economic Modelling, 68: 11-22.
Dimelis, S., Giotopoulos, I. and Louri, H., 2016. What de-
termines the growth of micro firms in the euro
area?. Economics and Business Letters, 5(4): 145-151.
Dimelis, S., Giotopoulos, I., and Louri, H., 2017. Can Firms
Grow Without Credit? A Quantile Panel Analysis in
the Euro Area. Journal of Industry, Competition and
Trade, 17(2): 153-183.
Dörr, S., Raissi, M. M. and Weber, A., 2017. Credit-Supply
Shocks and Firm Productivity in Italy. International
Monetary Fund, No. 17/67.
Duan, Y., 2008. Analysis of the Effect of the Financial Support
on China’s Equipment Manufacturing. International
Conference on Information Management, Innovation
Management and Industrial Engineering: 74-77.
Ebi, B. O. and Emmanuel, N., 2014. Commercial Bank Credits
and Industrial Subsector’s Growth in Nigeria. Journal of
Economics and Sustainable Development, 5(9): 1-10.
Estrada, G., Park, D., and Ramayandi, A., 2010. Financial
Development and Economic Growth in Asia. ADB
Working Paper No 233. https://www.adb.org/sites/de-
fault/files/publication/28277/economics-wp233
Favara G., 2006). An Empirical Reassessment of the
Relationship between Finance and Growth. IMF Working
Paper, 1-27. http://www.afi.es/eo/an%20empirical%20
reassessment
Giotopoulos, I. and Fotopoulos, G., 2010. Intra-industry
growth dynamics in the Greek services sector: firm-level
estimates for ICT-producing, ICT-using, and non-ICT in-
dustries. Review of Industrial Organization, 36(1): 59-74.
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
89South East European Journal of Economics and Business, Volume 14 (1) 2019
Greenwood, J., and Jovanovic, B., 1990. Financial
Development, Growth and the Distribution of Income.
Journal of Political Economy, 98: 1076-1107.
Guariglia, A. and Poncet, S., 2008. Could financial dis-
tortions be no impediment to economic growth af-
ter all? Evidence from China. Journal of Comparative
Economics, 36(4): 633-657.
Gurgul, H. & Lach, L. 2012. Financial development and eco-
nomic growth in Poland in transition: causality analy-
sis. Czech Journal of Economics and Finance (Finance a
Uver), 62, 4: 347-367.
Gurley, J.G. and Shaw E.S., 1955. Financial Aspects of
Economic Development. The American Economic
Review, 45(4): 515-538.
Hasanov, F., and Huseynov, F., 2013. Bank credits and
non-oil economic growth: Evidence from Azerbaijan.
International Review of Economics & Finance, 27:597-610.
Hassan, M. K., Sanchez, B. and Yu, J. S., 2011. Financial de-
velopment and economic growth: New evidence from
panel data. The Quarterly Review of economics and fi-
nance, 51(1): 88-104.
Iyoboyi, M., 2013. Bank and Non‐Bank Financial Deepening
and Economic Growth: The Nigerian Experience (1981–
2010). Economic Notes, 42(3): 247-272.
Izhar, A. and Tariq, M., 2009. Impact of institutional credit on
aggregate agricultural production in India during post
reform period. MPRA paper, 17075.
Jalil, A., Feridun, M., and Ma, Y., 2010. Finance–growth nexus
in China revisited: New evidence from principal compo-
nents and ARDL bounds tests. International Review of
Economics and Finance, 19(2): 189–195
John, E. E. and Terhemba, I. P. , 2016. Commercial Bank Credit
and Manufacturing Sector Output in Nigeria. Journal
of Economics and Sustainable Development, 7(16):
189-196.
Kaldor, N., 1968. “Productivity and Growth in Manufacturing
Industry: a reply”. Economica. 35: 385-391.
Kandir, S. Y., İskenderoğlu, Ö. and Önal, Y.B., 2007. Finansal
Gelişme ve Ekonomik Büyüme Arasındaki İlişkinin
Araştırılması. Ç.Ü. Sosyal Bilimler Enstitüsü Dergisi, 16(2):
311-326
Kar, M. and Pentecost, E. J., 2000. Financial Development
and Economic Growth in Turkey: Further Evidence on
the Causality Issue. Loughborough University Economic
Research Paper, 27.
Kashyap, A. K., Stein, J. C. & Wilcox, D. W. 1996. Monetary
policy and credit conditions: Evidence from the compo-
sition of external finance: Reply. The American Economic
Review, 86, 1: 310-314.
Kashyap, A. K. & Stein, J. C. 2000. What do a million observa-
tions on banks say about the transmission of monetary
policy?. American Economic Review, 90, 3: 407-428.
Kaya, A., Gülhan, Ü. and Güngör, B., 2013. Türkiye ekonomis-
inde finans sektörü ve reel sektör etkileşimi. Akademik
Araştırmalar ve Çalışmalar Dergisi, 5(8): 2-15
Kelly, J. and Everett, M., 2004. Financial liberalisation and
economic growth in Ireland. Central Bank Quarterly
Bulletin, 4(3): 91-112.
King, R. G., and Levine, R., 1993. Finance and Growth:
Schumpeter Might Be Right. Quarterly Journal of
Economics, 108: 717-738.
Kumar, R. R., 2011. Do Remottances, Exports and Financial
Development Matter for Economic Growth? A Case
Study of Pakistan Using Bounds Approach. Journal of
International Academic Research, 11(1): 18-27.
Larrain, B., 2006. Do banks affect the level and composition
of industrial volatility?. The Journal of Finance, 61(4):
1897-1925.
Leitão, N. C., 2012. Bank credit and economic growth: a dy-
namic panel data analysis. Economic Research Guardian:
256-267.
Leshoro, T. L. 2017. Trade union inflation expectations and
the second-round effect in South Africa: Toda-Yamamoto
causality approach. University of South Africa (Unisa)
Economic Research Working Papers, 13
Levine, R., 2002. Bank-Based or Market-Based Financial
Systems: Which is Better? Journal of Financial
Intermediation, 11(4): 398-428.
Levine, R., Loayza, N., and Beck, T., 2000. Financial intermedi-
ation and growth: Causality and causes. Journal of mon-
etary Economics, 46(1): 31-77.
Loayza, N.V. and Ranciere, R., 2006. Financial Development,
Financial Fragility, and Growth. Journal of Money, Credit
and Banking, 38(4): 1051-1076.
Lu, C. H., and Shen, C. H., 2012. Bank Lending and Economic
Growth of Chinese Cities. China & World Economy, 20(2):
59-79.
Lucas, R., 1988. On the Mechanics of Economic Development.
Journal of Monetary Economics, 22: 3-42.
Ma, Y. and Jalil, A., 2008. Financial development, economic
growth and adaptive efficiency: A comparison between
China and Pakistan. China & World Economy, 16(6):
97-111.
Mamgain, V., 1999. Are the Kaldor–Verdoorn Laws Applicable
In The Newly Industrializing Countries? Review of
Development Economics. 3(3): 295-309.
Manova, K., 2013. Credit constraints, heterogeneous
firms, and international trade. Review of Economic
Studies, 80(2): 711-744.
Marques, L. M., Fuinhas, J. A. and Marques, A. C., 2013. Does
the stock market cause economic growth? Portuguese
evidence of economic regime change. Economic
Modelling, 32: 316-324.
Meltzer, A. H. 1995. Monetary, credit and (other) transmis-
sion processes: A monetarist perspective. Journal of
Economic Perspectives, 9, 4: 49-72.
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
90 South East European Journal of Economics and Business, Volume 14 (1) 2019
Mishkin, F. S. 1995. Symposium on the Monetary Transmission
Mechanism. Journal of Economic Perspectives, 9, 4: 3-10.
Mishkin, F. S. 1996. The Channels of Monetary Transmission:
Lesson For Monetary Policy. National Bureau Of
Economic Research (NBER) Working Series, Working
Paper 5464: 1-27.
Mishkin, F. S. 2001. The Transmission Mechanism and the
Role of Asset Prices in Monetary Policy. National Bureau
Of Economic Research (NBER) Working Series, Working
Paper 8617: 1-21.
Mishra, P. K., Das, K. B. and Pradhan, B. B., 2009a. Empirical
Evidence on Indian Stock Market Efficiency in Context of
the Global Financial Crisis, Global Journal of Finance, 1(2):
149-157.
Mishra, P. K., Das, K. B. and Pradhan, B. B., 2009b. Credit mar-
ket development and economic growth in India. Middle
Eastern Finance&Economics, 5(3):92-106
Mohanty, A. R., Kumar S.and Patra, S. K., 2016. Bank credit
and economic growth: An empirical evidence from
Indian states. International Journal of Current Research,
8, (09): 39145-39151.
Nkurunziza, J.D., 2010. The effect of credit on growth and
convergence of firm size in Kenyan manufacturing. The
Journal of International Trade & Economic Development.
19 (3): 465-494.
Nnamocha, P. N. and Eke, C. N., 2015. Bank Credit and
Agricultural Output in Nigeria (1970 – 2013. An Error
Correction Model (ECM) Approach. British Journal of
Economics, Management & Trade 10(2): 1-12.
Nyasha S. and Odhiambo N.M., 2017. Bank Versus Stock
Market Development in Brazil: An ARDL Bounds Testing
Approach. South East European Journal of Economics
and Business 12 (1): 7-21
Odedokun, M. O., 1998. Financial intermediation and eco-
nomic growth in developing countries. Journal of
Economic studies, 25(3): 203-224.
Oliner, S. D. & Rudebusch, G. D. 1996. Monetary policy and
credit conditions: evidence from the composition of
external finance: comment. The American Economic
Review, 86, 1: 300-309.
Olokoyo, F. O., Adetiloye, K. A., & Ikpefan, O. A., 2016. Bank›s
Intermediation Role and Industrial Output in Developing
Economies. The Social Sciences, 11(24): 5838-5844.
Oluitan, R. O., 2012. Bank credit and economic growth:
Evidence from Nigeria. International Business and
Management, 5(2): 102-110.
Onder, Z. and Ozyildirim, S., 2010. Banka kredilerinin bölgesel
büyümeye etkileri. METU Studies in Development, 37(3):
267-292.
Oni, I. O., Akinlo, A. E. and Oladepo, E. D., 2014. Impact of
Bank Credit on the Real Sector: Evidence from Nigeria.
Global Journal of Business Research,8(3):39-47.
Patrick, H., 1966. Financial Development and Economic
Growth in Underdeveloped Countries. Economic
Development and Cultural Change, 14: 174-189.
Pesaran, M. H., Shin, Y. and Smith, R. J., 2001. Bounds testing
approaches to the analysis of level relationships. Journal
of Applied Econometrics, 16(3): 289-326.
Petkovski, M. and Kjosevski, J., 2014. Does banking sec-
tor development promote economic growth? An em-
pirical analysis for selected countries in Central and
South Eastern Europe. Economic Research-Ekonomska
Istraživanja. 27 (1): 55-66.
Pham, V. D. H., 2014. Credit market disruption and cor-
porate innovation: an empirical analysis. Saint Mary’s
University Research Report:1-67. Available at http://
s p r i n g e r l i n k . co m . l i b ra r y 2 . s m u. c a / b i t s t re a m / h a n –
d l e / 0 1 / 2 6 1 2 9 / P h a m _ V u D i e m H a n g _ M R P _ 2 0 1 4 .
pdf?sequence=1&isAllowed=y
Pons-Novell, J. and Viladecans-Marsal, E., 1998. Kaldor’s laws
and spatial dependence: evidence for the European re-
gions. Regional Studies. 33 (5): 443-451.
Pradhan, R. P., 2010a. Financial deepening, foreign direct
investment and economic growth: are they cointegrat-
ed. International Journal of Financial Research, 1(1): 37.
Pradhan, R. P. P., 2010b. The Nexus between Finance, Growth
and Poverty in India: The Cointegration & Causality
Approach1.Asian Social Science, 6(9):114-122.
Qamruzzaman, M. and Jianguo, W., 2017. Financial inno-
vation and economic growth in Bangladesh. Financial
Innovation, 3(1): 1-24.
Rahman, M ., Noman, A. H. M. & Shahari, F. 2017. Does eco-
nomic growth in Malaysia depend on disaggregate en-
ergy?. Renewable and Sustainable Energy Reviews, 78:
640-647.
Rahman, M., Shahbaz, M. and Farooq A., 2015. Financial
Development, International Trade and Economic Growth
in Australia: New Evidence from Multivariate Framework
Analysis. Journal of Asia-Pacific Business, 16(1): 21-43
Rajan, R. G. and Zingales, L., 1998. Financial Dependence and
Growth. American Economic Review, 88(3): 559–586.
Ramcharran, H., 2017. Bank lending to small business
in India: Analyzing productivity and efficiency. The
Quarterly Review of Economics and Finance, 65: 16–24.
Robinson, J., 1952. The model of an expanding economy.
The Economic Journal, 62(245): 42-53.
Romer, P. M. 1990. Endogenous technological change. Journal
of Political Economy, 98, 5, Part 2: S71-S102.
Scottish Governement, 2008. Access to Bank Finance for
Scottish SMEs. Scottish Government Publications. https://
www2.gov.scot/Publications/2008/09/08110445/2
Sehrawat, M. & Giri, A. K. 2015. The role of financial devel-
opment in economic growth: Empirical evidence from
Indian states. International Journal of Emerging Markets,
Vol. 10 Issue: 4, pp.765-780, https://doi.org/10.1108/
IJoEM-05-2014-0064
A MiCro BASEd Study oN BANk CrEdit ANd ECoNoMiC Growth: MANufACturiNG SuB-SECtorS ANAlySiS
91South East European Journal of Economics and Business, Volume 14 (1) 2019
Sehrawat, M. & Giri, A. K. 2016. Panel data analysis of financial
development, economic growth and rural-urban income
inequality: Evidence from SAARC countries. International
Journal of Social Economics, Vol. 43 Issue: 10, pp.998-
1015, https://doi.org/10.1108/IJSE-10-2014-0211
Shan J., and Jianhong Q., 2006. Does Financial Development
lead Economic Growth? The case of China. Annals of
Economics and Finance, 7(1): 197-213.
Sogules, I. W. and Nkoro, E., 2016. Bank Credits and Performance
of Manufacturing Sector inNigeria,1970-2013.The
Journal of Social SciencesResearch,2(7):129-132.
Stolbov, M., 2017. Causality between credit depth and
economic growth: evidence from 24 OECD coun-
tries. Empirical Economics, 53(2): 493-524.
T24. T24 Online News Paper. 2017. Available
at http://t24.com.tr/haber/turkiye-ekonomisi-dunyada
-kacinci-sirada,430739.
Tawose, J. O. B., 2012. Effects of bank credit on industrial
performance in Nigeria. International Business and
Management, 4(2): 158-168.
Toby, A. J. and Peterside, D. B., 2014. Analysis of the role of
banks in financing the agriculture and manufacturing
sectors in Nigeria. International Journal of Research in
Business Management, 2(2): 9-22.
Tripathi, V., & Kumar, A. 2015. Linkages between Real &
Financial Sector: Empirical Evidence from India Using
ARDL Approach. Financial Markets And Economic
Development. Available at https://www.researchgate.
net/profile/Arnav_Kumar4/publication/281118050_
Linkages_between_Real_Financial_Sector_Empirical_
E v i d e n c e _ f r o m _ I n d i a _ u s i n g _ A R D L _ a p p r o a c h /
links/55d72e7408aec156b9a9ff6f/Linkages-between-
Real-Financial-Sector-Empirical-Evidence-from-India-
using-ARDL-approach
Ume, K. E., Obasikene, A., Oleka, C., Nwadike, A. and
Okoyeuzu, C., 2017. The Relative Impact of Bank Credit on
Manufacturing Sector in Nigeria. International Journal of
Economics and Financial Issues, 7(2): 196-201.
Umit, A. Ö. 2016. Türkiye’de Banka Kredi Kanalının İşleyişini
2001 Krizi Sonrasında Yeniden Değerlendirme: Ampirik
Analiz. Abant İzzet Baysal Üniversitesi Sosyal Bilimler
Enstitüsü Dergisi. 16 (1): 97-123
UNIDO, 2017. Industrial Development Report 2018. Demand
for Manufacturing: Driving Inclusive and Sustainable
Industrial Development. United Nations Industrial
Development Organization, Vienna. https://www.unido.
org/
Vovchak, T., 2017. Bank Credit, Liquidity Shocks and Firm
Performance: Evidence from the Financial Crisis of 2007-
2009. CERGE-EI Working Paper Series, 584: 1-19. https://
www.cerge-ei.cz/pdf/wp/Wp584
Wa, H.N., 2002. Financial Development and Economic
Growth in Macau. MonetaryAuthority ofMacau,https://
p d fs. s e m a n t i c s c h o l a r. o rg / 3 5 3 0 / d 7 c b c 0 e 6 f 9 8 5 e f –
f37e2535d24d0cd1bf8fc9
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/521638
4
The Pattern and Causes of Economic Growth in
India
Article in Oxford Review of Economic Policy · February 200
7
DOI: 10.1093/oxrep/grm012 · Source: RePEc
CITATIONS
10
8
READS
39,13
0
2 authors:
Some of the authors of this publication are also working on these related projects:
Participatory research on irrigated farms – Tanzania View project
Tara Akshar Research Project View project
Kaushik Basu
Cornell University
249 PUBLICATIONS 9,759 CITATIONS
SEE PROFILE
Annemie Maertens
University of Sussex
34 PUBLICATIONS 535 CITATIONS
SEE PROFILE
All content following this page was uploaded by Annemie Maertens on 18 June 2014.
The user has requested enhancement of the downloaded file.
https://www.researchgate.net/publication/5216384_The_Pattern_and_Causes_of_Economic_Growth_in_India?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_2&_esc=publicationCoverPdf
https://www.researchgate.net/publication/5216384_The_Pattern_and_Causes_of_Economic_Growth_in_India?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_3&_esc=publicationCoverPdf
https://www.researchgate.net/project/Participatory-research-on-irrigated-farms-Tanzania?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_9&_esc=publicationCoverPdf
https://www.researchgate.net/project/Tara-Akshar-Research-Project?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_9&_esc=publicationCoverPdf
https://www.researchgate.net/?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_1&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Kaushik-Basu-4?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_4&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Kaushik-Basu-4?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_5&_esc=publicationCoverPdf
https://www.researchgate.net/institution/Cornell_University?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_6&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Kaushik-Basu-4?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_7&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Annemie-Maertens?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_4&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Annemie-Maertens?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_5&_esc=publicationCoverPdf
https://www.researchgate.net/institution/University_of_Sussex?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_6&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Annemie-Maertens?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_7&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Annemie-Maertens?enrichId=rgreq-f02a8994f438538be4d88959482ce9f5-XXX&enrichSource=Y292ZXJQYWdlOzUyMTYzODQ7QVM6MTA5NDIyNjA0OTg0MzIwQDE0MDMwOTk3OTAyODE%3D&el=1_x_10&_esc=publicationCoverPdf
Oxford Review of Economic Policy, Volume 23, Number 2, 2007, pp.143 – 167
The pattern and causes of economic growth
in India
Kaushik Basu∗ and Annemie Maertens∗∗
Abstract This paper presents the broad macro parameters of the growth of the Indian economy since
the nation’s independence and a cross-country evaluation of where India stands, drawing out the patterns
discernible in these aggregative statistics. The paper gives an overview of the on-going debate on the
components of the Indian growth and the relative importance of the different policies in the 1980s and 1990s.
It contributes to this debate by identifying the landmark years, and analysing the politics behind some of the
economics. The paper also analyses the factors behind the changes in India’s savings rate and the relation
between growth and development, on the one hand, and the nature of labour market regulation, on the other.
Key words: India, growth, labour market, savings
JEL classification: O10, O5
3
I. Introduction
The mainspring of an economy’s growth and take-off continues to puzzle economists. Even
though, thanks to years of sustained research, many of the pieces of the jigsaw puzzle are in
place, it remains very difficult to predict when an economy that has floundered for decades
might suddenly take off. The economy, embedded as it is in politics, culture, and institutions,
is a sufficiently complex organism for this not to be surprising. However, growth tends to
beget growth, though, of course, missteps can bring it to a halt. Hence, our understanding of
an economy’s rapid growth has to focus largely on what causes the first stirrings.
∗Cornell University and Indian Statistical Institute, New Delhi, kb40@cornell.edu
∗∗Cornell University, am445@cornell.edu
The paper has benefited greatly from the comments and suggestions received from Alaka Basu, V. Bhaskar,
Annelies Deuss, Bishnupriya Gupta, Abhijit Patnaik, and two anonymous referees of this journal. We are also
grateful to Charan Singh of the Reserve Bank of India for help and advice with data at several stages, and to
Ashokankur Datta and Namrata Gulati of the Indian Statistical Institute, New Delhi, for research assistance.
doi: 10.1093/icb/grm0
1
2
The Authors 2007. Published by Oxford University Press.
For permissions please e-mail: journals.permissions@oxfordjournals.org
144 Kaushik Basu and Annemie Maertens
What this paper attempts is to analyse and understand the constellation of forces that
has determined the growth performance of the Indian economy, including its long period of
hibernation and sudden, recent show of dynamism. The first task in such an undertaking is to
get the facts right. Over the last 4 or 5 years India has been getting a better press than ever
before, since its independence in 1947. Is this good press really justified? If the economy is
growing faster, when did the take-off occur? This is important to investigate not just to satisfy
idle curiosity but to understand the various forces that may have triggered the dynamism; and
that in turn is important for the crafting of policies to sustain the growth and spread its spoils
more evenly across the population.
Section II of the paper presents the broad macro parameters of growth of the Indian
economy since independence and also a cross-country evaluation of where India stands.
It then goes on to discuss the broad patterns that can be discerned in these aggregative
statistics pertaining to India, including the sectoral statistics, and gives a brief overview
of the on-going debate on the components of Indian growth and the relative importance
of the different policies in the 1980s and 1990s. Section III contributes to this debate by
identifying the landmark years, and analysing the politics that occurred behind the scenes,
and the extent to which they helped or hindered economic progress. Section IV looks at a
critical microeconomic component of the overall growth — labour-market behaviour, and is
followed by some concluding remarks in section V.
II. Growth: trends and patterns
(i) Backdrop
Thanks to a long history of data collection, the basic numbers of the Indian economy are, for a
poor country, well documented. At the time of its independence in 1947, India had: a literacy
rate of 18 per cent; an investment rate of around 9 per cent of its GDP; life expectancy at birth
of 32 years; an annual population growth rate of 1.25 per cent; and an average annual growth
rate of GDP of around 3 per cent. In 2005/6, India had: a literacy rate of around 60 per cent;
an investment rate of around 30 per cent of its GDP; life expectancy at birth of 63 years; an
annual population growth rate of 1.5 per cent; and an annual growth rate of GDP of around
8.4 per cent.
1
Given that the focus of the paper is on growth, let us take a look at GDP growth and
growth rate as displayed in Figure 1 and Table 1. We have graphed the natural log of the
GDP, rather than the actual GDP, so that one can read the growth rate directly from the slope
of the graph. A straight line represents a constant rate of growth. Table 1 presents annual
averages of growth rates and averages over plan periods.
Without going into any detailed analysis as yet, but just glancing at the data in Table 1, it
seems that the rate of growth in the 1950s, 1960s, and 1970s has been fluctuating around 3.
5
1 Sources, respectively: Selected Education Statistics (age 15 and above), Ministry of Human Resource
Development, from Indiastat; our Table 1; Age-group Wise Expectation of Life in India, Ministry of Health
and Family Welfare, from Indiastat;Dyson et al. (2004, Table 2.1, p. 20); our Table 1; estimated literacy rate age
15 and above of UNESCO; our Table 1; World Development Indicators 2006, World Bank; Dyson et al. (2004,
Table 2.1, p. 20); our Table 1.
The pattern and causes of economic growth in India 145
Table 1: Annual growth rate of real GDP and gross capital formation, 1950 – 200
6
Year Annual
growth rate
of GDP at
factor costs
Gross
domestic
capital
formation (%
of GDP at
factor cost)
Year Annual
growth rate
of GDP at
factor costs
Gross
domestic
capital
formation (%
of GDP at
factor cost)
1950/1 9.07 1980/1 7.2 22.45
1951/2 2.3 11.59 1981/2 6 22.34
1952/3 2.8 8.32 1982/3 3.1 21.7
9
1953/4 6.1 8.08 1983/4 7.7 20.69
1954/5 4.2 10.04 1984/5 4.3 22.16
1955/6 2.6 13.64 Average 5.6
Average 3.6 1985/6 4.5 24.2
1956/7 5.7 15.76 1986/7 4.3 23.47
19 578 −1.2 14.73 1987/8 3.8 25.23
1958/9 7.6 12.64 1988/9 10.5 26.48
1959/60 2.2 13.36 1989/90 6.7 27.23
Average 3.5 Average 5.9
1960/1 7.1 15.23 1990/1 5.6 29.27
1961/2 3.1 14.18 1991/2 1.3
25
1962/3 2.1 15.95 1992/3 5.1 26.25
1963/4 5.1 15.31 1993/4 5.9 25.39
1964/5 7.6 15.26 1994/5 7.3 28.72
Average 5.0 Average 5.0
1965/6 −3.7 17.47 1995/6 7.3 29.77
1966/7 1 18.19 1996/7 7.8 26.94
1967/8 8.1 15.17 1997/8 4.8 26.94
1968/9 2.6 14.23 1998/9 6.5 24.59
1969/70 6.5 15.99 1999/2000 6.1 28.43
Average 2.8 Average 6.5
1970/1 5 16.68 2000/1 4.4 26.37
1971/2 1 17.46 2001/2 5.8 24.97
1972/3 −0.3 16.53 2002/3 3.8 27.51
1973/4 4.6 18.81 2003/4 8.5 29.57
1974/5 1.2 18.28 2004/5 7.5 33.04
Average 2.3 Average 6.0
1975/6 9 18.79 2005/6 9.0a
1976/7 1.2 19.78 2006/7 9.2a
1977/8 7.5 20.
11
1978/9 5.5 23.85
1979/80 −5.2 22.85
Average 3.5
Per 5-year
plan periods
I. 1951 – 56 3.6 VI. 1980 – 85 5.6
II. 1956 – 61 4.2 VII. 1985 – 90 6.0
III. 1961 – 66 2.8 VIII. 1992 – 97 6.7
IV. 1969 – 74 3.3 IX. 1997 – 02 5.5
V. 1974 – 79 4.8
Notes: Up to 1999/2000, old series (base: 1993/4). From 2000/1 onwards, new series (base: 1999/2000).
Averages: authors’ own calculations.a Latest estimates, released by the Ministry of Finance, Economic Survey
2006/7.
Source: Reserve Bank of India, Handbook of Statistics on the Indian Economy 2006 (Table 1 and 237).
146 Kaushik Basu and Annemie Maertens
Figure 1: Ln GDP (at constant prices), 1950 – 2005
11
11.5
12
12.5
13
13.5
14
14.5
5
0
-5
1
5
5
-5
6
6
0
-6
1
6
5
-6
6
7
0
-7
1
7
5
-7
6
8
0
-8
1
8
5
-8
6
9
0
-9
1
9
5
-9
6
0
0
-0
1
ln
G
D
P
[
a
t
F
a
c
to
r
C
o
s
t
]
Source: Reserve Bank of India, Handbook of Statistics on the Indian
Economy 2006 (Table 2 — old series; base: 1993/4).
per cent per annum, the so-called ‘Hindu rate of growth’.2 With an average annual rate of
population growth of 1.9 per cent, this results in an average annual growth in per-capita GDP
of around 1.6 per cent. From the late 1970s, the rate of growth exhibits an upward trend,
averaging around 6 per cent for the period 1980 – 2005.
To get a basic idea of the absolute numbers involved, Table 2 gives the size of the Indian
population, the real GDP (at market prices), and the real GDP per capita (at market prices).
The key difference between the GDP at factor costs and the GDP at market prices is that the
latter includes indirect taxes net of subsidies. As the latter is considered a better measure of
the standard of living, we have opted to report the absolute figures of the GDP and GDP per
capita at market prices.
From Table 2 it is clear that while the Indian population has more than doubled since the
1960s, GDP has increased more than eightfold since then. As the population figures for India
are based on projections from Census of India data,3 we have opted not to show the entire
time series for population or GDP per capita.
To conclude this introduction, let us ask one more factual question: how has India done
vis-à-vis other nations, especially other developing countries? Has it really done better or is
it simply the fact of a large country beginning to grow that has caught the media’s attention
and imagination?
In order to answer this question we assembled the purchasing power parity (PPP)-corrected
national income and per-capita national income data of the World Bank for 109 countries.4
There was a trade-off involved. As one goes further back, data, especially when we want
2 ‘Hindu rate of growth’ is the tongue-in-cheek expression, coined by the Indian economist, the late Raj Krishna,
to capture the frustrations India’s planners faced with growth. No matter what they did, growth seemed, invariably,
to revert back to 3.5 per cent per annum, almost as if this magic figure was written in the land’s scriptures. The
possibility of Hinduism having something to do with economic growth was earlier suggested by B. P. R. Vithal.
3 The first census of India was carried out throughout the 1860s and completed in 1871. Since then there have
been 13 more censuses, one per decade, the latest one being the 2001 census.
4 The PPP-corrected GDP takes into account the difference in prices of goods and services between countries.
As the exchange rate only takes into account the differences in tradable goods and services and several
countries
have non-market-based exchange-rate determination, it is arguable that the PPP allows us to make more meaningful
comparisons of standards of living across countries.
The pattern and causes of economic growth in India 147
Table 2: Population, GDP, and GDP per capita at market prices, selected years
Year Population (in
millions
GDP (in millions
constant 2000 US$)
GDP per capita
(constant 2000 US$)
1960 435 76,283 175
1965 487 91,054 187
1970 548 113,606 207
1975 613 130,913 213
1980 687 152,621 222
1985 765 198,167 259
1990 850 268,023 316
1995 932 345,394 371
2000 1,016 457,377 450
2005 1,095 641,926 586
Source: World Development Indicators 2006, World Bank.
them PPP-corrected, get sparse and more and more countries have to be left out. We chose
to go back to 1975, when the PPP-corrected data became available for the first time. There
are 109 countries for which data are available without break from that year to current times.
For each year since 1975, we ranked these 109 countries based on PPP-corrected GDP and
per-capita GDP.
From this assembled data set (not shown here) it is clear that not only has India done better
over time vis-à-vis itself, but even in comparison to others. In terms of GDP per capita, India
ranked 90th among these nations in 1975. The rank fluctuated a little between 1975 and 1982,
falling to 93rd and rising again to 90th. From 1982 onwards there has been a steady and
monotonic improvement, with India’s per-capita GDP (PPP-corrected), rising to 75th rank in
2004. There are two or three countries that did better than India over this period, the most
notable being China, which was 108th among the 109 countries in 1975 but had risen to the
rank of 58th by 2004. But, as follows from the fact of rank improvement, India overtook
numerous nations during the last three decades.
In terms of GDP ranking the improvement has also been marked, though this is tempered
by the fact that some of the poorer economies have had a faster growth of population,
especially over the last two decades. In 1985, India’s PPP-corrected GDP was the 8th largest
in the world, and by 2004 it was the fourth largest, with only the USA, China, and Japan
ahead.5
Despite this rank improvement, India and South Asia in general are still among the
poorest regions in the world (see Table 3). Indeed, a quick look at one of the most important
development indicators, the population below the poverty line, shows us that despite the
decline in poverty headcount ratio, from 55 per cent in 1973/4 to 29 per cent in 1998/9, India
still accounts for a large absolute number of poor people, close to 30 m in 2000.6
This change in India’s growth rate and improved economic performance vis-à-vis other
nations triggered off a change in global perception not just in academic writing, but in the
5 If India’s rank is measured in GDP (constant 2000 international dollars), India ranks 13th.
6 Sources: Reserve Bank of India, Handbook of Statistics of the Indian Economy 2005 – 2006 (Table 172) and
National Sample Survey (NSS) 55th Round Official Estimates. Note that the measurement of poverty has been a
hotly debated subject especially since the 55th round of the NSS tried to change the reference period of household
consumption (see Deaton and Drèze, 2002; Himanshu and Sen, 2005; Lancaster and Ray, 2005; Subramanian, 2006,
ch. 10).
148 Kaushik Basu and Annemie Maertens
Table 3: GDP per capita (constant 2000 US$), selected comparisons
1965 1970 1980 1990 1995 2000 2004
India 187 207 222 316 371 450 538
Sub-Saharan Africa 485 536 577 520 484 504 537
South Asia 197 220 234 326 377 446 522
East Asia and Pacific 145 176 273 481 735 952 1,254
Latin America and Caribbean 2,276 2,616 3,568 3,262 3,555 3,854 3,906
World 2,843 3,316 3,974 4,555 4,748 5,237 5,516
Note: The East Asia and Pacific and the Latin America and Caribbean aggregates do not contain
the high-income countries.
Source: World Development Indicators 2006, World Bank.
media and popular business publications, that India was a newly emerging and dynamic
economy and, in very recent years, it has been repeatedly compared to China.7 This was
unthinkable to most India watchers even a decade ago. This popular celebration of the
economy gives rise to a host of questions. While the strengthening of the growth rate is
beyond doubt, the recorded overall growth rate is not the only indicator one uses to judge an
economy and so the question arises as to whether the fundamentals are as strong as the media
make them out to be. What are the strengths and weaknesses of the economy? Is there reason
to expect that the growth will be sustained? What are the right policies for sustaining the high
growth and spreading its spoils better among the population? Even though growth is higher,
can we pin down when exactly the breaks occurred?
(ii) Growth patterns and hypotheses
Turning to details of the growth performance, let us take another look at the growth rate of
the GDP in Figure 2 and Table 1. Observe that the spikes in annual growth rates have not
changed very much over the years; it is the downturns that have become less severe and
frequent. Before 1980, there were 4 years when the GDP recorded negative growth rates:
1957/8, 1965/6, 1972/3, and 1979/80. Since 1980 never has the economy, as measured by
GDP, shrunk in any year, though per-capita GDP fell once — during 1990/1 — owing to the
First Gulf War and a sharp decline in remittances and exports. GDP grew slowly that year
but by less than the population growth.
Owing to the huge amount of noise, the trends are not too evident to the naked eye. But
once we smooth out these annual fluctuations and look, instead, at averages of several years
of growth (see Table 1), a pattern emerges. The average growth holds steady till about the
mid-1970s and then, somewhere after that, it begins to move up, and that upward incline has
persisted till current times. This is corroborated by the average, annual growth-rate figures
for each of the 5-year plan periods. Average annual growth nearly touched the 5 per cent
mark during the Fifth Plan period, 1974 – 9, and has never dropped below that since. The
sharp spike occurred during the Eighth Plan period, 1992 – 7, when annual growth averaged
6.7 per cent. All the portents are that, during the Tenth Plan period, the economy will grow
at close to 8 per cent per annum. Given that India’s population growth rate is much slower
than it used to be three or four decades ago (1.5 per cent in 2004/5 as against 2.22 per cent in
7 This changing perception is cited and discussed in Basu (2006b).
The pattern and causes of economic growth in India 149
Figure 2: Growth rate of GDP in India, 1950 – 2006
-6
-4
-2
0
2
4
6
8
10
12
5
5
-5
6
6
0
-6
1
6
5
-6
6
7
0
-7
1
7
5
-7
6
8
0
-8
1
8
5
-8
6
9
0
-9
1
9
5
-9
6
0
0
-0
1
0
5
-0
6
G
ro
w
th
R
a
te
o
f
G
D
P
[
a
t
fa
c
to
r
c
o
s
ts
]
Source: Table 1.
1971/28) this means that the rise in per-capita income growth rate from the 1960s and 1970s
to current times has been even more marked.
More formal evidence that the GDP growth series exhibits a structural break at the end of
the 1970s/beginning of the 1980s can be found in Virmani (1997, 2004a), Wallack (2003),
Rodrik and Subramanian (2004b), and Balakrishnan and Parameswaran (2006). The last, for
instance, use a regression-based least-squares approach that does not arbitrarily partition the
data according to pre-selected break points and identify 1978/9 as a structural break year for
the GDP growth series. These authors challenged the standard view held in the 1990s by the
public and a large majority of economists that the policy reforms of the early 1990s had caused
or played a major role in the growth acceleration (views held by, for instance, Ahluwalia
(2002) and Srinivasan and Tendulkar (2003)). A new view emerged, led by Williamson and
Zagha (2002), De Long (2003), Rodrik and Subramanian (2004a,b), Panagariya (2004), and
Virmani (2004a), that the surge in growth rate in India happened around 1980 and could
therefore not be attributed entirely to the new economic policies of the early 1990s.
While it is difficult to dispute that a rise in growth rate took place before the 1990s, it is
possible to argue that there was further acceleration after the reforms of the 1990s, which can
be attributed to those reforms. Further, we are inclined to argue — as, indeed, some others
have done — that the growth in the 1980s was not of a sustainable nature, since it relied too
much on deficit financing and excessive foreign borrowing (Basu, 2004; Panagariya, 2004;
Srinivasan, 2005).
To understand this debate and the components of the post-1980s growth further, let us
take a look at the results of the growth accounting exercise of Bosworth et al. (2007). The
objective of growth accounting is to decompose the economic growth rate of a country
into contributions of different factors. Assuming a certain aggregate production function and
competitive markets, the method identifies the contribution of the different factors (such as
labour and physical capital) and a residual, called total factor productivity (TFP).9 Changes
8 Source: Dyson et al. (2004, Table 2.1, p. 20).
9 Thereby hinting at the main critique of this approach: that TFP is a residual and, as such, incorporates also all
kinds of shocks, such as political turmoil, external shifts, and measurement errors.
150 Kaushik Basu and Annemie Maertens
Table 4: Contributions to growth (in annual percentage rate of change)
Contribution of:
Selected Output Employment Output Physical Land Education Factor
periods per worker capital productivity
1960 – 73 3.3 2.0 1.3 1.1 −0.2 0.1 0.2
1973 – 83 4.2 2.4 1.8 0.9 −0.2 0.3 0.6
1983 – 93 5.0 2.1 2.9 0.9 −0.1 0.3 1.7
1993 – 9 7.0 1.2 5.8 2.4 −0.1 0.4 2.8
1999 – 2004 6.0 2.4 3.6 1.2 0.1 0.4 2.0
1960 – 2004 4.7 2.0 2.6 1.2 −0.1 0.3 1.2
1960 – 80 3.4 2.2 1.3 1.0 −0.2 0.2 0.2
1980 – 2004 5.8 1.9 3.8 1.4 0.0 0.4 2.0
Source: Bosworth et al. (2007, Table 3).
Table 5: Percentage of GDP (at factor costs) by industry of origin
Year Agriculture,
forestry,
fishing, mining,
and quarrying
Manufacturing,
construction,
and electricity,
gas, and water
supply
Trade, hotel,
transport, and
commu
nication
Financing,
insurance, real
estate, and
business
services
Public
administration
and defence
and
other
services
1950/1 59 13 12 7 9
1960/1 55 17 14 6 9
1970/1 48 20 16 6 11
1980/1 40 22 18 7 12
1990/1 35 24 19 10 12
2000/1 27 24 22 13
15
2005/6 23 24 25 13 14
Source: Ministry of Finance, Economic Survey 2005 – 2006, Table 1.3.
in the TFP represent changes in efficiency and/or changes in production technology. Table 4
shows the results of this exercise.
Table 4 shows that the pre-1980 growth is mainly associated with an increase in factors
while the post-1980 growth is associated with some increase in factors, but more importantly
an increase in TFP.10 Looking at the entire time series, they conclude that the TFP growth
took off around the early 1980s, and has shown an increasing trend since then. This finding
is consistent with other studies on TFP growth (Rodrik and Subramanian, 2004b; Virmani,
2004b). Despite the large structural change in the economy (see Table 5), this increase in
TFP, according to these authors, mainly reflects an improvement of the performance of
the individual sectors rather than a re-allocation of resources from low-productivity sectors
(agriculture) to higher productivity sectors (manufacturing and services).11
10 Other studies confirm this general trend (see, for instance, Dholakia, 2002). A more detailed discussion can be
found in Virmani (2004b).
11 According to Rodrik and Subramanian (2004b), a structural shift can only explain 10 per cent of the TFP
growth.
The pattern and causes of economic growth in India 151
But how did this sudden surge in TFP come about? Rodrik and Subramanian (2004b)
suggest that even though the reforms of the 1980s, which consisted of some industrial
liberalization measures, lowering of tax rates, and limited import liberalization, were not
substantial, this small trigger could have elicited a large response in TFP because India
was below its production possibility frontier. The increases in TFP would, in that case, just
be a reflection of the move towards the frontier rather than a shift of the frontier itself.12
As a whole, they see an attitudinal shift towards ‘pro-business’ policies (in contrast to the
‘pro-market’ policies of the 1990s) as crucial in explaining the surge in the aggregate growth
rate and TFP.13 They share this view with other authors, such as Panagariya (2004), who see
the switch from a ‘positive’ list approach, where restrictions are the rule and few exceptions
are allowed, in the 1980s towards a ‘negative’ list approach in the 1990s as crucial.
Let us now turn to the disaggregate figures of growth and TFP. Beginning with the
primary sector, from Table 6 (column 1) it is clear that the growth rate of this sector has
been extremely volatile. The (arithmetic) average annual growth rate of the entire series
is 5.5 per cent and the standard deviation is 3.8 per cent. Despite the irregular nature of
this time series, Balakrishnan and Parameswaran (2006) were able to discern a structural
break, namely a positive break around 1964/5. Notably, this break is situated slightly before
the onset of the Green Revolution (around 1967/8), which in India mainly consisted of the
spread of high-yield rice and wheat varieties. According to them, this structural break ‘may
owe something to the steady expansion in irrigated area in the decade and a half preceding
the mid-sixties’.14 Indeed, given the significance of agriculture as a share of GDP in the
1950s – 1980s, it seems that a good or a bad year in terms of rains could have a large impact
on the overall growth rate. Consider, for example, the GDP growth rates of the good years
1958/9, 1967/8, and 1988/9 in Table 6. Virmani (2004a), however, contests this conventional
wisdom and argues that there has been no change in the impact of rainfall fluctuations on the
Indian economy during the last 50 years. As far as TFP is concerned, Bosworth et al. (2007)
show that, taking 1960 as the index year (1960 = 1), the growth of TFP in the agriculture
sector fluctuates around the index 1 up to the mid- to late 1980s, after which an increasing
trend can be discerned. In figures, they find that the TFP growth changes from −0.2 per cent
per year during the period 1960 – 73, to 0.9 per cent per year during 1973 – 83, and to 1.2 per
cent during 1983 – 99.
Despite the fact that the growth figures of the last few years do not seem structurally
different from the growth figures of the previous decades, there is talk of an agrarian crisis
in India. This is caused by the declining public investments in agriculture (a trend which
started in the early 1980s), the decline in agriculture as a share of the GDP associated with
relatively little reallocation of employment (the primary sector contributes 20 per cent of the
GDP but has a share of 60 per cent of the employment), the fact that poverty in India is
a predominantly rural phenomenon, and the rise in farmer suicides, mainly in the states of
Karnataka, Andhra Pradesh, and Maharashtra (Vaidyanathan, 2006).
12 The reforms of the 1980s are extensively discussed by, among others, Kohli (2006a), Virmani (2004b), and
Panagariya (2004).
13 Rodrik and Subramanian (2004b) provide evidence for this attitudinal shift by the government in the early
1980s that favoured the interests of existing businesses rather than new entrants or consumers. This evidence has
been contested by Srinivasan (2005).
14 The current gross irrigated area is 40 per cent of the cultivated area. This area has increased greatly over the last
40 years (see Ministry of Agriculture, 2004, Table 14.2).
152 Kaushik Basu and Annemie Maertens
Table 6: Annual growth rate of real GDP at factor cost by industry of origin
Agriculture,
forestry,
fishing,
mining, and
quarrying
Manufacturing,
construction,
and
electricity,
gas, and
water supply
Trade,
hotels,
transport,
and commu-
nication
Financing,
insurance,
real estate,
and business
services
Public
administra-
tion,
defence, and
other
services
1951/2 1.8 4.5 2.7 2.3 3.0
1952/3 3.1 0.1 3.2 4.2 2.1
1953/4 7.5 6.4 3.7 1.4 3.1
1954/5 3.0 8.6 6.4 3.7 3.6
1955/6 −0.8 11.2 7.3 4.0 3.1
Average 2.9 6.1 4.6 3.1 3.0
1956/7 5.4 8.8 7.4 1.6 3.8
1957/8 −4.2 −1.0 3.3 3.8 4.5
1958/9 9.9 7.1 5.1 2.8 4.1
1959/60 −0.8 7.1 6.3 3.8 4.3
Average 2.5 5.6 5.3 3.0 3.9
1960/1 7.0 10.5 8.5 2.1 4.9
1961/2 0.3 7.2 6.5 4.3 4.7
1962/3 −1.5 6.4 6.0 3.4 7.1
1963/4 2.4 10.6 7.1 3.1 6.6
1964/5 8.9 7.3 6.7 2.7 6.6
Average 3.3 8.4 7.0 3.1 6.0
1965/6 −10.2 3.0 2.0 3.0 4.0
1966/7 −1.3 3.4 2.6 1.8 4.6
1967/8 14.3 3.1 4.4 2.7 3.9
1968/9 0.0 5.2 4.6 4.9 4.5
1969/70 6.4 8.1 5.4 4.2 5.5
Average 1.5 4.5 3.8 3.3 4.5
1970/1 6.5 1.8 4.8 4.2 5.5
1971/2 −1.7 2.7 2.3 5.2 4.5
1972/3 −4.6 3.5 2.5 3.9 3.3
1973/4 6.9 1.1 4.2 2.4 2.6
1974/5 −1.3 1.3 6.2 −0.3 4.7
Average 1.1 2.1 4.0 3.1 4.1
1975/6 12.9 6.1 9.0 6.9 3.5
1976/7 −5.4 9.3 4.6 7.9 2.8
1977/8 9.7 7.2 6.4 4.9 2.7
1978/9 2.3 8.0 8.1 7.1 4.3
1979/80 −12.2 −3.4 −0.4 1.0 7.3
Average 1.0 5.3 5.5 5.5 4.1
1980/1 12.9 4.0 5.7 1.9 4.1
1981/2 5.7 7.4 6.2 8.3 2.6
1982/3 0.0 2.9 4.6 10.4 8.0
1983/4 9.1 8.7 4.9 10.0 3.9
1984/5 1.5 6.2 5.1 8.5 6.8
Average 5.7 5.8 5.3 7.8 5.1
1985/6 1.0 4.7 7.9 10.2 6.5
1986/7 0.2 6.2 5.9 11.3 7.0
1987/8 −1.0 7.0 5.2 8.4 7.2
1988/9 15.4 8.6 6.0 11.4 6.4
1989/90 1.9 10.7 7.4 12.6 8.3
Average 3.3 7.4 6.5 10.8 7.1
(continued overleaf )
The pattern and causes of economic growth in India 153
Table 6: (Continued)
Agriculture,
forestry,
fishing,
mining, and
quarrying
Manufacturing,
construction,
and
electricity,
gas, and
water supply
Trade,
hotels,
transport,
and commu-
nication
Financing,
insurance,
real estate,
and business
services
Public
administra-
tion,
defence, and
other
services
1990/1 4.6 7.4 4.9 7.7 4.1
1991/2 −1.1 −1.0 2.5 12.0 2.6
1992/3 5.4 4.3 5.6 5.9 4.6
1993/4 3.9 5.6 7.1 13.4 3.5
1994/5 5.3 10.3 10.4 5.6 3.2
Average 3.6 5.3 6.1 8.9 3.6
1995/6 −0.3 12.3 13.3 8.2 7.9
1996/7 8.8 7.7 7.8 7.0 6.3
1997/8 −1.5 3.8 7.8 11.6 11.7
1998/9 5.9 3.8 7.7 7.4 10.4
1999/2000 0.6 4.9 8.5 10.6 12.2
Average 2.6 6.5 9.0 8.9 9.7
2000/1 0.2 6.7 7.1 4.1 4.7
2001/2 5.8 2.8 9.2 7.3 3.9
2002/3 −5.6 6.8 9.1 8.0 3.8
2003/4 9.6 7.9 12.0 4.5 5.4
2004/5 1.2 8.9 10.6 9.2 9.2
Average 2.1 6.6 9.6 6.6 5.4
Source: Ministry of Finance, Economic Survey 2005 – 2006, Table 1.6. From 2001/2 new series at 1999/2000
prices. Before 2001/2 at 1993/94 prices. Averages: authors’ own calculations.
Turning to industry, let us first look at the industry figures in column 2 of Tables 5 and
6. The share of the industrial sector has increased over the last 50 years from 13 per cent at
the time of independence to 24 per cent. As such, the structure of the economy of India is
nowadays of a very different nature from that of China, where industry represents nearly 50
per cent of the economy. The time series in column 2, Table 6, which seems at first sight quite
volatile, has an (arithmetic) average of approximately 6 per cent and a standard deviation of
2.9 per cent. According to Balakrishnan and Parameswaran (2006), the manufacturing series
exhibits three structural breaks. The first negative structural break is in the mid-1960s. The
second positive structural break is in 1982/3, and the third negative structural break is in
1994/5. They interpret these figures as evidence against the hypothesis that manufacturing
led the acceleration in the GDP growth rate at the beginning of the 1980s. Virmani (2004a),
however, comes to the exact opposite conclusion: ‘this [analysis] shows that the growth rate
of manufacturing accelerated after 1980 – 81. This contributed to the acceleration of the rate
of growth of GDP from 1981.’ And the debate does not end there. According to Bosworth
et al. (2007), the TFP growth of industry has been slowing down, not accelerating, during
the post-reform period. They conclude that ‘these results are disappointing in light of the
attention that has been devoted to the on-going liberalization of the trade and regulatory
regimes for goods production’. Their conclusions related to manufacturing are very similar to
those for industry as a whole. They, among others, thereby provide counter-evidence to the
studies of Ahluwalia (1995) and Unel (2003), who concluded that manufacturing experienced
a surge of productivity in the 1980s. Goldar and Mitra (2002) take the more sceptical line
that these differences in findings can be attributed to a variety of measurement issues.
154 Kaushik Basu and Annemie Maertens
Finally, let us take a look at the service-sector figures in the remaining columns of Tables 5
and 6. Table 6 shows that, since the 1980s, the services have shown a more consistent
higher annual growth rate than the industrial and the agricultural sectors of the economy.
Bosworth et al. (2007) show that, according to their growth-accounting analysis, this increase
in growth rate is mainly due to an increase in TFP. This is rather puzzling as services are
normally considered as an area of limited productivity growth. They suggest a number of
explanations for this phenomenon, such as incorrect measurement of prices in the service
industry. Srinivasan (2005) even suggests that the higher wages in the public sector might be
driving a spurious increasing TFP. This is clearly an area which needs further research.
Given the predominance of the services in the Indian economy (see Table 5), namely 52
per cent of the GDP in 2005/6, several authors have concluded that this sector is driving
the current growth that India witnesses. According to the analysis of Balakrishnan and
Parameswaran (2006), services led the acceleration in the growth of GDP in India in the
late 1970s to early 1980s (see, also, Babu, 2005). In this regard also, the growth in India
is of a very different nature from the growth in China, where services contribute only 33.2
per cent to the GDP (Panagariya, 2004). As only 23 per cent of the population is employed
in India’s service sector and the growth in employment in this sector has been low, many
scholars have concluded that India is caught in the groove of ‘jobless growth’. While the
growth in employment need not match the growth in each sector, the discrepancy between
the contribution of the primary sector to the GDP (20 per cent) and its share in the workforce
(60 per cent of the population) is indeed worrying.
As illustrated in the previous section, despite its limitations, growth accounting can provide
some insight into the proximate causes of growth. In addition, other techniques, such as
growth regressions, are often employed to analyse the impacts of the immediate causes, such
as quantity and quality of labour and capital, as well as the ‘fundamental determinants’ of
growth. It is, however, important to note that both methods have been heavily criticized in
the literature.15
Without going into the technical details of these debates, let us conclude this section
by taking a look at these ‘fundamental determinants of growth’ for the Indian case. Most
scholars seem to agree on the fundamental determinants of growth: physical and human
capital investments, quality of institutions or governance, and investment climate. While the
gross domestic capital formation ratio (33 per cent — see Table 1) is rather high, India’s low
literacy rate (61 per cent) could potentially become a constraint on India’s long-term growth
prospects. And what about India’s institutions and investment climate?
In order to answer this question, we took a look at the World Bank’s ‘ease-of-business’
indicator, which can be viewed as a measure of investment climate. This indicator ranks
economies in terms of their ‘ease of business’, from 1 to 175, with 1 referring to the best.
It averages the country’s percentile rankings in 10 categories, giving equal weight to each
category. Each category in its turn averages the country’s percentile rankings on different
15 See, for instance, Durlauf et al. (2004). See also Bosworth and Collins (2003) for additional references on these
critiques.
The pattern and causes of economic growth in India 155
Table 7: Cross-country comparison of ease-of-business indicators (2006)
Average rank
Category Low-income
countries
Lower-
middle-
income
countries
Upper-middle-
income
countries
High-income
countries
India
Starting a business 116 98 71 45 88
Dealing with licences 115 94 75 49 155
Employing workers 107 85 81 70 112
Registering property 113 96 76 48 110
Getting credit 121 80 69 32 65
Protecting investors 94 96 67 50 33
Paying taxes 105 101 75 52 158
Trading across bordersa 130 96 71 30 139
Enforcing contracts 111 95 91 39 173
Closing a business 117 94 87 26 133
Overall rank 133 97 67 27 134
GNI per capita ($) 436 2,037 6,431 30, 763 7
20
Notes: a The ‘trading across borders’ measure does not include tariffs or trade taxes.
Source: World Bank, Doing Business 2007.
sub-categories.16 According to the latest indicators, India ranks 134 out of 175 and is
situated around the average rank of low-income countries, above the majority of the Sub-
Saharan African countries and below most South- and South-East-Asian and Latin-American
developing countries. From Table 7, one can see that India ranks relatively high in the ‘getting
credit’ and ‘protecting investors’ spheres, but particularly low in the areas of ‘dealing with
licences’, ‘paying taxes’, and ‘enforcing contracts’. In these areas, as well as in the areas of
‘employing workers’ and ‘closing a business’, most analysts would agree that further reforms
are needed.
The ‘dealing with licences’ indicator is of particular interest in the Indian case as it is often
viewed as a residue of the Licence Raj that characterized India before the 1980s (Aghion
et al., 2006).17 According to the World Bank figures, India has not improved much over
the last 2 years. In India one needs on average 270 days to complete all the procedures
required to build a standardized warehouse in the construction industry; this is considerably
higher than the average of the low-income countries (231 days). India’s cost measure of this
indicator gives a slightly more optimistic picture. Obtaining the necessary licences to build
the warehouse costs 606 per cent of the GNI per capita, which is in between the averages
of the low-income countries (996 per cent) and the lower-middle-income countries (558 per
cent).
16 Note that the World Bank figures consider only the official costs and times involved for a standardized firm,
assuming perfect knowledge about the procedures, and these measures most likely underestimate the real costs
involved. In addition, the rankings do not take into account that the opportunity cost of time differs across countries;
one day waiting in India is not the same as one day waiting in the USA. On a similar note, as the gross national
income (GNI) is much higher in the high-income countries, a low cost as a percentage of the GNI is in a way
‘easier’ to achieve; also, as this cost is not calculated as a percentage of the PPP GNI, the actual perceived costs in
developing countries might be lower than is suggested by the World Bank figures. It is unclear how these data issues
affect the relative ranking of the countries.
17 Some first-hand descriptive accounts of India’s burgeoning bureaucracy occur in Basu (2007b).
156 Kaushik Basu and Annemie Maertens
Many scholars have argued that India’s performance is surprisingly low when compared
to the quality of its institutions.18 Rodrik and Subramanian (2004b), for instance, using
geography, openness, and economic and political institutions as fundamental determinants of
growth, conclude that:
India’s level of income was about a quarter of what it should be given the strength of its
economic institutions. On the other hand, if political institutions are the true long-run
determinant of income, India’s income is about 15 per cent of what it should be. India
has thus been a significant under-achiever in the sense that it has not exploited the
potential created by having done the really hard work of building institutions.
In their analysis they used settler mortality, fraction of the population speaking one of the
major languages of Western Europe, or fraction of the population speaking English as an
instrument for institutions, as described in Rodrik et al. (2002).
III. The political economy of growth
The previous section gave an overview of India’s growth performance over the last 50 years
and briefly outlined the elements of the current debate in the literature on India’s growth.
This section tells the story behind the numbers.
The first real big growth year for India, 1975/6, was also one that stood out politically as one
of the most salient, if not notorious, years for the nation. That year the country’s GDP grew
by 9 per cent, a figure that has been surpassed only twice since then.19 It was also the year in
which the then Prime Minister, Indira Gandhi, declared a state of emergency and established
dictatorial control over the nation. This would last for 2 years. In 1977 Indira Gandhi called
an election. There is no way of knowing if this was prompted by an exaggerated sense of
popularity on her part or because of a genuine fatigue she felt with totalitarian control. But
the fact of the matter is that she was routed at the polls, and she would return to power
(re-elected) only in 1980.
Some of the growth spurt of the early Emergency period would be undone in 1979/80, the
worst-performing year in the history of independent India, but 1974 – 9, as already noted, was
nevertheless to be the cross-over plan period when average annual growth rate closed in on
the 5 per cent mark.
Thanks to the totalitarian embarrassment of 1975, most Indian commentators are loathe
to identify that year as a break in the trend for the economy.20 Yet there is no denying
that it was, even though the sustainability of that growth impulse may be questionable. The
next improvement would come in the early 1980s, when not only did the growth rate pick
up further but, as pointed out above, the country broke ranks with other nations. The next
change, which in our opinion was the large and sustainable change and is often thought of as
a structural break, occurred in 1991, when, pushed by a macroeconomic crisis, itself caused
18 See, for instance, Srinivasan (2005).
19 One of these is the current year, 2006/7, the growth rate for which is estimated to be 9.2 per cent.
20 To the extent that the value of democracy is not purely instrumental but as an end in itself, 1975 – 7 must overall
go down as dark years in India’s history. For a discussion of India’s democracy and development, see Sen (2004).
The pattern and causes of economic growth in India 157
by the First Gulf War and the drying up of foreign reserves, Indian undertook the most major
reforms since independence.21
If these are the three landmark years — and clearly this matches reasonably well with the
statistical analysis — what were the factors that led to them? There were important policy
changes, it is true, but what does not always get adequate credit in the case of India are the
two closely related variables that are identified in standard growth theory as among the most
significant factors — the savings and investment rates.22 These rates, which were traditionally
very low in India, around 15 per cent till the late 1960s, began climbing all through the
1970s and crossed the 20 per cent mark in 1978/9. This must have contributed to the greater
growth momentum of the mid-1970s, and, in fact, the persistent growth that we have seen
since the early 1980s. The extra spike in the year 1975/6 did probably owe something to the
Emergency. Trains do run on time in the first flush of dictatorship and there is anecdotal
evidence that this and other work-related efficiencies were adopted in that year. But that first
flush soon vanished and, thanks to severe shortages in infrastructural resources, the economy,
too, slowed down severely by the end of the 1970s.
A natural question that arises is: what caused the rise in the savings rate? Unfortunately,
this question has not received sufficient attention and no clear answer is available. It seems
to us, however, that it had something to do with the nationalization of banks that Indira
Gandhi announced in 1969.23 After nationalization, the banks were forced to open branches
in remote, ‘unprofitable’ areas. This, coupled with the impetus that came with the formation
of the state-owned Unit Trust of India in 1964, may have prompted greater savings by making
savings easier and safe (Shetty, 2007). Table 8 shows that there was, indeed, a phenomenal
increase in the number of bank branches in India, following the nationalization; so some
prima-facie evidence for this hypothesis is, indeed, there.
The opening of branches and making savings and borrowing outlets available to poorer
citizens were explicit objectives of the nationalization.24 And while its impact on savings
awaits formal investigation, there are other kinds of related studies that do suggest the
nationalization of banks had a large impact on ordinary citizens.25 Burgess and Pande
(2005), for instance, test whether this large state-led bank-branch expansion programme was
associated with poverty reduction in India, given that an integral element of the programme
was branch expansion into rural locations without banks. The paper’s main finding is that
branch expansion into such rural locations in India significantly reduced rural poverty. It
21 Why the Indian reforms came so late, and only when the nation was up against the wall, is itself an interesting
question. It may have something to do with India’s democracy, which is quite unique. All the currently developed
democratic nations adopted democracy with universal suffrage, after the process of industrialization was firmly in
place. India adopted universal suffrage at independence, at a level of poverty with few parallels, and so it has had
to contend with the opinion of the poor in ways that are quite alien to the industrialized nations of today (Varshney,
2007). This may also have something to do with the tenacity of India’s labour laws, discussed below.
22 See Majumdar (1997) for discussion of growth theory in the context of the Indian economy.
23 The Supreme Court of India initially declared the nationalization to be invalid. But Indira Gandhi amended the
law and passed the nationalization decision by an ordinance.
24 ‘The banking system touches the lives of millions and has to be inspired by larger social purpose and has to
subserve national priorities and objectives such as rapid growth of agriculture, small industries and exports, raising
of employment levels, encouragement of new entrepreneurs and development of backward areas. For this purpose it
is necessary for the government to take direct responsibility for the extension and diversification of banking services
and for the working of a substantial part of the banking system.’ (Bank Company Acquisition Act, 1969).
25 Economic analysis of banking reform is a relatively scarce discipline in India. For recent work, see Banerjee
et al. (2003, 2004).
158 Kaushik Basu and Annemie Maertens
Table 8: Number of bank branches in India (scheduled and
non-scheduled commercial banks)
Year Number
1941 2,074a
1951 4,119
1961 5,113
1969 9,051
1971 12,985
1976 23,656
1981 38,047
1986 53,397
1991 62,740
1996 64,937
2001 67,856
2005 70,324
Notes: a India and Burma (officially now Myanmar). Note the figures
from 1991 onwards are from 31 March, while those before that date
are from 31 December.
Source: Reserve Bank of India, Statistical Tables Related to Banks in
India, 2005 – 2006 and previous issues.
Table 9: Nominal and real lending rates, 1970 – 2004
Year 1970 – 4 1975 – 9 1980 – 4 1985 – 9 1990 – 4 1995 – 9 2000 – 4
Nominal lending rate 9.0 10.4 13.3 13.9 16.7 15.2 11.4
Inflation rate 15.3 4.7 9.3 6.7 11 5.3 5.2
Real interest rate −4.9 5.9 3.9 6.8 5.2 9.4 5.9
Note: The nominal lending rate is an average of the rates for four major lending institutions. The inflation rate
is measured by the annual rate of change in the wholesale price index for all commodities.
Source: Bosworth et al. (2007, Table 12)
seems natural to expect that such a large banking initiative did cause a boost in savings,
especially since it coincided with the rise in India’s savings rate.
Another reason for the increasing savings rate could be the increasing real interest rates
(Table 9). These are positive and show an increasing trend since 1974. Yet, most detailed
studies of savings find a rather weak connection between interest rates and savings (Rao,
2007; Shome, 2007), suggesting that consumers are more interested in long-run prospects
and the facilities for saving than the immediate lure of interest. As Shome (2007, p. 464)
remarks, ‘But it is clearly the growth in financial intermediation that stands out most as the
main driver of savings.’
Let us look at some statistics on the growth of savings in India. Bosworth et al. (2007)
report that not only have national savings risen considerably since the 1980s, but that, in
particular, household savings have risen from 10 to 25 per cent of GDP during the last
30 years. Half of this is in the form of financial savings, which can be channelled back into
other sectors as investment. Public-sector saving, however, has not performed as well. From
a high of around 4 per cent in the 1970s, it became negative in the late 1990s, recovering
only recently. These trends are noted in Figure 3. Note that savings are reported here as
percentage of GNP at factor costs. It is also important to note that savings have risen
The pattern and causes of economic growth in India 159
Figure 3: Domestic savings per sector
-5
0
5
10
15
20
25
30
35
1
9
7
0
-7
1
1
9
7
2
-7
3
1
9
7
4
-7
5
1
9
7
6
-7
7
1
9
7
8
-7
9
1
9
8
0
-8
1
1
9
8
2
-8
3
1
9
8
4
-8
5
1
9
8
6
-8
7
1
9
8
8
-8
9
1
9
9
0
-9
1
1
9
9
2
-9
3
1
9
9
4
-9
5
1
9
9
6
-9
7
1
9
9
8
-9
9
2
0
0
0
-0
1
2
0
0
2
-0
3
%
o
f
G
N
P
a
t
fa
c
to
r
c
o
s
ts
Public Sector Savings Household Physical Savings
Corporate Sector Savings Household Financial Savings
Note: Old series (base: 1993/4) used for calculation.
Source: Reserve Bank of India, Handbook of Statistics on the Indian Economy 2006 (Tables 1
and 10).
particularly in the post-liberalization years. Whether there is a link between higher savings
and post-liberalization policies needs to be further researched.
To conclude, the rise in India’s savings rate deserves more detailed investigation than has
occurred thus far. Given that India is currently in the midst of another sharp rise in savings
(the first since the late 1970s), this is a subject of contemporary relevance.
The second acceleration — in the early 1980s — probably owes something to altered
policies, and, of course, it came with a background of higher investment and savings rates.
The change in policy regime makes for interesting political sleuthing. It seems to have much
to do with Indira Gandhi’s altered perceptions. The reading of descriptive accounts of her
regime (for instance, Dhar, 2000; Frank 2002) suggests that unlike her father, Nehru, India’s
first Prime Minister, Indira Gandhi never had strong convictions about economic policy. Her
early commitment to ‘socialism’ was arguably prompted by an instinctive following of her
father’s convictions and policies, without any deep convictions of her own. She nationalized
banks and established control over grain trade as homage to her father without a coherent
plan for the whole economy as such.
By the mid-1970s she was under the influence of another man — her younger son, Sanjay
Gandhi. Sanjay was not committed to any well-thought-out ideology, but was wary of
Congress socialism. What was notable about him was his vaulting ambition and, along with
his entrepreneurial friends, he pushed India towards crony capitalism. Much has been written
about his disproportionate influence on his mother. In an interview that he gave in July 1976,
he openly criticized the Communist Party of India (an ally of Indira Gandhi) and disparaged
earlier policies of the Congress. Mrs Gandhi was upset by the interview and summoned
P. N. Dhar, who headed her secretariat, and told him, ‘Sanjay has done something terrible
and I am upset’ (Dhar, 2000, p. 325). The conversation that followed, where she asked Dhar
to do the damage control, suggested to him that she was ‘afraid’ of Sanjay’s ‘displeasure’. As
time progressed and she felt more and more isolated from her own party and other politicians,
160 Kaushik Basu and Annemie Maertens
she turned increasingly to her son, who had visions, without wisdom, of an entrepreneurial
revolution, mainly under the ownership of him and his friends.26
By the early 1980s India had started out on a path of openly capitalist development. Even
though this was done with no systematic vision and with favours doled out to those close
to the government, it boosted growth, as the statistical analysis above shows. The economy
had for so long been shackled by bureaucratic rules and red tape that the release from these,
however small, caused a rise in growth. Moreover, by now India had higher investment and
savings rates to support this.
By the late 1980s, even though the country was growing fast, it was beginning to borrow
heavily from its future, which makes us believe that the growth impulse of the 1980s
would not have been sustainable without sharp changes in policy. The fiscal deficit was
growing, international debt was reaching record levels, and the debt – service ratio had
become untenable. The meltdown happened in 1990/1. The First Gulf War was the proximate
cause, but the bubble was anyway ready to burst. Huge amounts have been written on this
crisis27 and we will not go into that here. But the crisis became an impetus for economic
reform. By 1991 government had changed hands. Narasimha Rao was Prime Minister and
Manmohan Singh was his Finance Minister. Under their stewardship a reform started, more
far-reaching than any since the early days of Nehru’s government. Industrial licensing was
discarded and the astronomical import tariff rates were set on a sharp downward course. The
first 2 years of the reform were a difficult time for the economy. But in terms of overall
growth rate and performance in the international sector, the Indian economy has not looked
back since then. From 1994 to 1997 the economy grew at a rate of above 7 per cent for 3
successive years, slowing down a little after that as a result of the general East-Asian crisis;
but over the last few years the growth rate has picked up again. It has not dropped below
7.5 per cent per annum since 2003 and has thrice crossed the 8 per cent mark. What has
been powering this new growth and was the big success of the reforms was the international
sector. India’s foreign-exchange balance started rising from a precarious low in 1991, when
the country was on the verge of default, to a very comfortable level. As Table 10 shows, the
reserves had fluctuated but, on balance, remained low for several decades, up to 1990. Since
then, with the reforms marking an excellent dummy variable, it has grown sharply. Currently,
India is among the world’s five largest foreign reserve holders.
Even exports have risen, especially when one includes software and information-
technology-related invisibles within exports.28 According to World Bank figures (see
Figure 4), exports, as a percentage of GDP, crossed 10 per cent for the first time in
1992 and are currently over 19 per cent. With the rise in foreign-exchange balance and the
confidence of success in the software and pharmaceuticals sector, Indian corporations have
gone on a spree of buying international companies, an activity unheard of 10 years ago. It
is this international presence and visibility that has given India somewhat disproportionate
global media attention. But that in itself is an advantage, since it has boosted confidence in
the country, pouring money into the Indian bourses, and become partly self-fulfilling.
While the economic reforms of 1991 – 3 lie behind the international success of the country,
there is more to the story, especially over the last 4 or 5 years and this, once again, intertwines
with politics, this time global politics.
26 Sanjay Gandhi died when a plane that he was flying crashed on 23 June 1980.
27 See, for instance, Desai (1994), Srinivasan (2000), Ahluwalia (2002), Bardhan (2004), Basu (2004), and
Chidambaram (2007).
28 For an analysis of India’s success in software and information technology, see Kapur (2002) and Murthy (2004).
The pattern and causes of economic growth in India 161
Table 10: Foreign-exchange reserves in India, selected years
Year Foreign-
exchange
reserves ($
millions)
Aggregate
export of goods
and services ($
millions)
Short-term debt,
(as % of foreign
reserves)
Debt – service
ratio
1977 5,824 6,354
1990 5,834 18,477 129 35
1994 25,186 26,855 14 26
1998 32,490 34,298 16 18
2002 75,428 52,512 10 14
2005 130,000 68,000 5.7 6.2
Sources: Government of India, Ministry of Finance, Economic Survey (various years); press
releases of the Ministry of Commerce.
Figure 4: Export and imports as a percentage of GDP in
India, 1965 – 2004
0
5
10
15
20
25
1
9
6
5
1
9
7
0
1
9
7
5
1
9
8
0
1
9
8
5
1
9
9
0
1
9
9
5
2
0
0
0
%
o
f
G
D
P
Exports of goods and
services (% of GDP)
Imports of goods and
services (% of GDP)
Source: World Bank, World Development Indicators 2006.
There are two factors to which it is worth drawing attention.29 First, there was one
unintended positive spillover of the last US presidential election. During the election
campaign, outsourcing back-office work to developing countries came under heavy criticism,
with some television shows, such as that of Lou Dobbs, attacking US entrepreneurs for
profiteering by outsourcing work. This had a huge advertisement effect for the advantages of
outsourcing. Most poor countries would not be able to afford such advertising on American
television; they suddenly got it for free. Small US entrepreneurs, who were unaware of
29 A more detailed discussion of this occurs in Basu (2006b).
162 Kaushik Basu and Annemie Maertens
this profit opportunity, learned about it and began outsourcing, and India’s already large
outsourcing business received another boost.
Second, having an alignment of interest with the world’s most powerful nation, the USA,
can have large benefits for an economy. We have seen this happen in the case of South Korea
since the 1950s. Shifts in the global geo-political balance of power have suddenly brought
India into the ambit of US interest. As long as the USA’s main foreign policy concern was
Communism and the USSR, it had little use for India. But now, with terrorism being the
main global concern of the USA, it has shared interests with India, which are deeper than the
current tactical ties with Pakistan. Also, with the rise of China, the USA has apprehensions
of a new unipolar world with China at the centre or, what is only marginally better, a direct
face-off with China in a bipolar world. In the event of a showdown between these two
countries, the risks for the USA are huge. If China were to cut off or monitor trade flows
through the Strait of Malacca, the biggest sea route for trade in today’s world, this could have
very large consequences for the US economy. China also has a disproportionate leverage on
the value of the dollar, thanks to its large foreign-exchange holdings. India is viewed by the
United States as a counterpoise for all these risks, and India’s economy has benefited from
this emerging geo-political advantage.30
But we want to turn now from these macroeconomic aggregates and broad global issues
to the microeconomic foundations of what is happening in India. The advantages of global
politics can easily be dissipated, as we saw in the case of the former Soviet Union, if a
country’s economic ‘nuts and bolts’ do not function well. There are now increasing data on
the microeconomic institutions that permit businesses to thrive and grow and play a crucial
role in an economy’s long-run trajectory. The next section examines some of these issues.
IV. Microeconomic foundations
India’s initial focus on the international sector has paid off handsomely. But to sustain
this growth, microeconomic issues — better distribution of income, improved labour-market
functioning, the control of corruption, and more efficient institutions for business and
enterprise — need greater attention. These are often referred to as second-generation reforms.
There is no effort here to cover all these microfoundational issues, each of which could be
the subject of a full-length paper, but we comment on one, namely, labour-market regulation,
where we have some insights to offer which are not common knowledge.
While the Indian economy is booming, there is evidence that workers are not partaking
in the boom adequately. Employment is not growing as fast as working-age population, nor
are wages rising as rapidly as per-capita income. There are many reasons for this — some to
do with forces of globalization that are beyond the Indian government’s policy reach. But
much of it has to do with the ‘culture’ that pervades India’s labour markets, which in turn is
a consequence of the complicated and ill-conceived laws that govern the labour market.
In India there are 45 laws at the national level and close to four times as many at the
level of state governments that monitor the functioning of labour markets. This complexity is
reflected in the World Bank’s ease-of-business indicators, where India ranked 112 out of 175
countries in the category ‘employing workers’ in 2006 (with 1 being the best). Even though
30 Interestingly, Indo – Chinese relations have also improved steadily since Rajiv Gandhi’s visit to China in 1989;
and trade between India and China has grown exponentially over the last 4 years (see Ramesh, 2005).
The pattern and causes of economic growth in India 163
recent changes in the regulation of several Indian states have resulted in a lower ‘difficulty
of hiring index’, ‘rigidity of hours index’, ‘rigidity of employment index’, and ‘firing costs’,
India still scores high on the ‘difficulty of firing’ index (70/100), which is considerably higher
than the average of the low-income countries (44/100).
Some of these labour laws date back to the nineteenth century. They were meant to control
conflict and keep the labour market efficient. Unfortunately, the experience has been to the
contrary. According to recent World Bank estimates, in 2004 there were 482 cases of major
work stoppages, resulting in 15 m human days of work loss (World Bank, 2006). Between
1995 and 2001, around 9 per cent of factory workers were involved in these stoppages. The
figure for China is close to zero. On the other hand, the wages of Chinese workers are rising
much faster than those of India’s workers. These facts, we would argue, are not unrelated.
Most of India’s labour laws were crafted with scant respect for ‘market response’. If X
seemed bad, the presumption was that you had simply to enact a law banning X. But the fact
that each law leads entrepreneurs and labourers to respond strategically, often in complicated
ways, was paid no heed. In a poor country no one with any sensitivity wants workers to
lose their jobs. So what does one do? The instinct is to make it difficult for firms to lay off
workers. That is exactly what India’s Industrial Disputes Act, 1947, did, especially through
the amendments of 1976 and 1982, for firms in the formal sector and employing more than
100 workers.
But in today’s globalized world, with volatile and shifting demand, firms have responded
to this by keeping their labour force as small as possible. It is little wonder that in a country
as large as India fewer than 10m workers are employed in the formal private sector. Some
commentators have argued that India’s labour laws could not have had much of a consequence
since most of them apply only to the formal sector. What they fail to realize is that one reason
the formal sector has remained minuscule is because of these laws and also the culture that
these laws have spawned (Basu, 2006a).
Several recent studies have analysed the impacts of labour regulations on firm productivity,
patterns of specialization, and technological progress. According to Besley and Burgess
(2004), increasing pro-worker regulation has a negative impact on investment and productivity
in the registered manufacturing sectors. What is also interesting about their findings is the lack
of evidence that such policies improve labor interests. Aghion and Burgess (2003) confirm
these results and in addition show that the negative impact of having stricter labour regulations
on productivity has increased in the post-liberalization period. Kochar et al. (2006), based
on their analysis of the patterns of specialization of Indian firms, suggest that not only is the
level of productivity of existing firms affected by stringent regulations, but new firms are also
kept from entering as a result.
What is needed in India is not a law that allows employers to fire workers at will, but one
that allows for different kinds of contracts. Some workers may sign a contract for a high
wage, but one that requires them to quit at short notice; others may seek the opposite. This
would allow firms to employ different kinds of labour depending on the volatility of the
market they operate in.
Much of the debate on labour laws has been misconstrued. What is needed is not change
in labour laws and policy to elicit sacrifice from organized labour, as some economists
have suggested. Indian workers, whether they be in the organized sector or the unorganized
sector, are too poor for that. The need is for changes in order to create greater private-sector
demand for labour, which would boost wages and employment. We believe that India’s
poorly construed labour laws have been so persistent because of an intellectual failure — to
164 Kaushik Basu and Annemie Maertens
wit, the inability to grasp that, in some contexts, it may be in the worker’s own interest to be
able to waive some rights that have been granted to him/her.
It should be clarified that we are not making the case for all workers to be given the right to
give up their right either to call strikes or to continue to work, but simply arguing that there
are contexts where it is worthwhile giving them the meta-right to give up some right. One
has to weigh lots of pros and cons before making a general recommendation. We feel that the
right to strike has other advantages so that, barring some very special cases, workers should
always have this right. On the other hand, we believe that in India workers should be given
the right to sign contracts with different kinds of firing or retrenchment rules, and that doing
so is likely to cause such a rise in aggregate demand for labour that all workers will be better
off. While much of the current debate in India on labour laws is conducted as though worker
interests are pitted against business interests, in reality it is between thinking clearly and not
thinking clearly.
It should also be added that flexibility in hiring and firing is not the only problem. India’s
complex web of legislation has led to a system of dispute resolution that is incredibly slow.
Data from the Ministry of Labour reveal that in the year 2000 there were 533,038 disputes
pending in India’s labour courts; of these, 28,864 had been pending for over 10 years. If India
is to be a vibrant global economy, this has to change.
In brief, the need is to move to a system that (i) makes room for more flexible contracts
in the labour market, (ii) has a minimal welfare net for workers who are out of work, and
(iii) resolves labour-market disputes more quickly.
V. Conclusion
To conclude, if India wants to sustain and raise even higher its current growth, the main
bottlenecks in the Indian economy will need to be addressed. These are infrastructure (roads,
expensive freight rates, power supply, ports, and airports), labour and bankruptcy regulations,
and the high level of corruption in the government bureaucracy. In addition, the current
erratic and low growth pattern of the agricultural sector, and the rising inequality — between
states, between rural and urban areas, and within urban and within rural areas mainly since
the 1990s — are a concern.
Of these numerous factors, we have addressed only a few in this paper. Each of these
factors deserves inquiry, research, and policy initiative, but in concluding we remark briefly
on just one of them — the subject of inequality.
Comparing the ratio of the income shares of the richest 10 per cent over the poorest 10
per cent in India with other countries, one may be tempted to conclude that inequality in
India is not abnormally high. According to the World Bank’s World Development Indicators
2006, this ratio was 7.2 in India (in 2000), compared to 18.36 in China (in 2001), 48 in
Guatemala (in 2002), and 15.9 in the USA (in 2000). As such, India’s current inequality
seems to be low and comparable to some of the Western European nations. But one has
to remember that a poor country will have a natural tendency for greater equality, since
people cannot survive below a certain level of income. To take an extreme case, a country
that has a per-capita income equal to the subsistence income will, by definition, have no
income inequality. Hence, despite the seemingly encouraging inequality ratio mentioned
above, inequality — especially when it results in higher poverty — is a serious problem for
India. This could lead to political tensions and could destabilize the otherwise optimistic
growth scenario. But, even if it does not dampen the country’s growth prospects, it seems to
The pattern and causes of economic growth in India 165
us that greater equity and the reduction of poverty are valuable ends in themselves. Indeed, it
is arguable that growth is valuable precisely because it enables a country to banish poverty
and achieve greater equality. India’s trajectory over the last 15 years has been remarkable,
but there will truly be reason to celebrate this when the overall gains filter down to the poorest
and the most deprived sections of India’s vast population.
References
Acharya, S. (2004), ‘India’s Growth Prospects Revisited’, Economic and Political Weekly, 9 October,
4537 – 42.
Aghion, P., and Burgess, R. (2003), ‘Liberalization and Industrial Performance: Evidence from India and the
UK’, conference paper presented at the Future of Globalization conference at Yale University in 2003.
— — Redding, S., and Zilibotti, F. (2006), ‘The Unequal Effects of Liberalization: Evidence from
Dismantling the License Raj in India’, Centre for Economic Performance, London School of Economics,
CEP Discussion Papers.
Ahluwalia, I. J. (1995), ‘India: Industrial Development Review’, The Economist Intelligence Unit and
UNIDO.
Ahluwalia, M. S. (2002), ‘Economic Reforms in India since 1991: Has Gradualism Worked?’, Journal of
Economic Perspectives, 16(3), 67 – 88.
Babu, S. M. (2005), ‘India’s Recent Economic Growth, Some Limits and Limitations’, Economic and Political
Weekly, 23 July, 3249 – 52.
Balakrishnan, P., and Parameswaran, M. (2006), ‘Identifying Growth Regimes in India’, paper presented at
the South and South East Asian Econometric Society, 19 December, Chennai, India.
Banerjee, A. V., Cole, S., and Duflo, E. (2003), ‘Bank Financing in India’, MIT Department of Economics
Working Paper.
— — — (2004), ‘Banking Reform in India’, MIT Department of Economics Working Paper.
Bardhan, P. (2004), ‘Disjunctures in the India Reform Process: Some Reflections’, in K. Basu (ed.), India’s
Emerging Economy: Performance and Prospects in the 1990s and Beyond, Cambridge, MA, MIT Press,
and New Delhi, Oxford University Press.
Basu, K. (2004), ‘The Indian Economy: Up to 1991 and Since’, in K. Basu (ed.), India’s Emerging Economy:
Performance and Prospects in the 1990s and Beyond, Cambridge, MA, MIT Press, and New Delhi,
Oxford University Press.
— (2006a), ‘Labor Laws and Labor Welfare in the Context of the Indian Experience’, in A. de Janvry and
R. Kanbur (eds), Poverty, Inequality and Development, New York, Springer.
— (2006b), ‘India Globalizing’, in D. A. Kelly, R. S. Rajan, and G. H. L. Goh (eds), Managing Globalization:
Lessons from China and India, Singapore, World Scientific Publishing.
— (ed.) (2007a), Oxford Companion to Economics in India, New Delhi, Oxford University Press.
— (2007b), The Retreat of Democracy and other Itinerant Essays on Globalization, Economics and India,
New Delhi, Permanent Black.
Besley, T., and Burgess, R. (2004), ‘Can Labor Regulation Hinder Economic Performance? Evidence from
India’, Quarterly Journal of Economics, 119(1), 91 – 134.
Bhagwati, J. N. (2004), ‘International Flows of Humanity’, ch. 14, in In Defense of Globalization, Oxford
University Press, 208 – 18.
Bosworth, B. P., and Collins, S. M. (2003), ‘The Empirics of Growth: An Update’, Working Paper,
Washington, DC, Brookings Institution.
— — Virmani, A. (2007), ‘Sources of Growth in the Indian Economy’, forthcoming in India Policy
Forum.
Burgess, R., and Pande, R. (2003), ‘Do Rural Banks Matter? Evidence from the Indian Social Banking
Experiment’, BREAD Working Paper No. 037, Harvard University, Bureau for Research and Economic
Analysis of Development.
— — (2005), ‘Do Rural Banks Matter? Evidence from the Indian Social Banking Experiment’, American
Economic Review, 95(3), 780 – 95.
166 Kaushik Basu and Annemie Maertens
Chidambaram, P. (2007), ‘Fiscal Policy Reforms since 1991’, in K. Basu (ed.), Oxford Companion to
Economics in India, New Delhi, Oxford University Press.
Deaton, A., and Drèze, J. (2002), ‘Poverty and Inequality in India: A Reexamination’, Economic and Political
Weekly, 7 September, 729 – 48.
De Long, B. J. (2003), ‘India since Independence: An Analytic Growth Narrative’, in D. Rodrik (ed.), In
Search of Prosperity: Analytic Narratives on Economic Growth, Princeton, NJ, Princeton University
Press.
Desai, A. V. (1994), My Economic Affair, New Delhi, Wiley Eastern.
Dhar, P. N. (2000), Indira Gandhi, the ‘Emergency’ and Indian Democracy, New Delhi, Oxford University
Press.
Dholokia, B. H. (2002), ‘Sources of India’s Accelerated Growth and the Vision of the Indian Economy in
2020’, The Indian Economic Journal, 29(4), 27 – 46.
Drèze, J., and Sen, A. (2002), India, Economic Development and Social Opportunity, Oxford University
Press.
Durlauf, S., Johnson, P. A., and Temple, J. R. W. (2004), ‘Growth Econometrics’, University of
Wisconsin – Madison, Department of Economics Research Paper.
Dyson, T., Cassen, R., and Visaria, L. (2004), Twenty-first Century India, Population, Economy, Human
Development and the Environment, Oxford University Press.
Forbes, N. (2002), ‘Doing Business in India: What has Liberalization Changed?’, in A. O. Krueger (ed.),
Economic Policy Reforms and the Indian Economy, Chicago, IL, and London, University of Chicago
Press, 129 – 67.
Frank, K. (2002), Indira: The Life and Times of Indira Nehru Gandhi, London, Harper Collins.
Goldar, B., and Mitra, A. (2002), ‘Total Factor Productivity Growth in Indian Industry: A Review of Studies’,
in B. S. Minhas (ed.), National Income Accounts and Data Systems, New Delhi, Oxford University Press,
218 – 37.
Himanshu and Sen, A. (2005), ‘Poverty and Inequality in India: Getting Closer to the Truth’, in A. Deaton
and V. Kozel (eds), The Great Indian Poverty Debate, New Delhi, Macmillan.
Kapur, D. (2002), ‘The Causes and Consequences of India’s IT Boom’, India Review, 1, 91 – 110.
Khilnani, S. (1998), The Idea of India, New York, Farrar, Straus, & Giroux.
Kochar, K., Kumar, U., Rajan, R., Subramanian, A., and Tokatlidis, I. (2006), ‘India’s Pattern of Development:
What Happened, What Follows’, IMF Working Paper No. WP/06/22.
Kohli, A. (2006a), ‘Politics of Economic Growth in India, 1980 – 2005, Part I: The 1980s’, Economic and
Political Weekly, 1 April, 1251 – 9.
— (2006b), ‘Politics of Economic Growth in India, 1980 – 2005: Part II: The 1990s and Beyond’, Economic
and Political Weekly, 8 April, 1361 – 70.
Lancaster, G., and Ray, R. (2005), ‘On Setting the Poverty Line, Based on Nutrient Prices’, Economic and
Political Weekly, 1 January, 46 – 56.
Majumdar, M. (1997), ‘The East Asian Miracle and India’, The Asiatic Society, Monograph, Calcutta, 1 – 51.
Mathur, K. B. L. (2004), ‘The Growth Rate Mystery’, Economic and Political Weekly, 17 July, 3228 – 31.
Ministry of Agriculture (2004), Agricultural Statistics at a Glance, Department of Agriculture and
Cooperation, India.
Murthy, N. R. N. (2004), ‘The Impact of Economic Reforms on Industry in India: A Case Study of the
Software Industry’, in K. Basu (ed.), India’s Emerging Economy: Performance and Prospects in the
1990s and Beyond, Cambridge, MA, MIT Press, and New Delhi, Oxford University Press.
Panagariya, A. (2004), ‘Growth and Reforms during the 1980s and 1990s’, Economic and Political Weekly,
19 June, 2581 – 94.
Rakshit, M. (2007), ‘Oil Price Shocks’, in K. Basu (ed.), Oxford Companion to Economics in India, New
Delhi, Oxford University Press.
Ramesh, J. (2005), Making Sense of Chindia: Reflections on China and India, New Delhi, India Research
Press.
Rao, D. T. (2007), ‘Do Savings, Investment and Interest Rates Respond to Each Other?’, in G. S. Ganesh
Prasad (ed.), Transformation, Transition or Stagnation?, Chennai, Madras Institute of Development
Studies.
Rodrik, D., and Subramanian, A. (2004a), ‘Why India Can Grow at 7 Per Cent a Year and More: Projections
and Reflections’, Economic and Political Weekly, 17 April, 1591 – 6.
The pattern and causes of economic growth in India 167
Rodrik, D., and Subramanian, A. (2004b), ‘From ‘‘Hindu Growth’’ to Productivity Surge: The Mystery of the
Indian Growth Transition’, IMF Working Paper No. WP/04/77.
— — Trebbi, F. (2002), ‘Institutions Rule: The Primacy of Institutions over Geography and Integration
in Economic Development’, NBER Working Paper 9305, Washington, National Bureau of Economic
Research.
Sen, A. (2004), ‘Democracy and Secularism in India’, in K. Basu (ed.), India’s Emerging Economy:
Performance and Prospects in the 1990s and Beyond, Cambridge, MA, MIT Press, and New Delhi,
Oxford University Press.
Shetty, S. L. (2007), ‘India’s Savings Performance since the Advent of Planning’, in K. L. Krishna and
A. Vaidyanathan (eds), Institutions and Markets in India’s Development, New Delhi, Oxford University
Press.
Shome, P. (2007), ‘Savings and Investment’, in K. Basu (ed.), Oxford Companion to Economics in India,
New Delhi, Oxford University Press.
Srinivasan, T. N. (2000), Eight Lectures on India’s Economic Reforms, New Delhi, Oxford University Press.
— (2005), ‘Comments on ‘‘From ‘Hindu Growth’ to Productivity Surge: The Mystery of the Indian Growth
Transition’’’, IMF Staff Papers, 52(2).
— Tendulkar, S. (2003), Reintegrating India with the World Economy, Washington, DC, Institute for
International Economics.
Subramanian, A. (2006), ‘Growth Experience’, in K. Basu (ed.), Oxford Companion to Economics in India,
New Delhi, Oxford University Press.
Subramanian, S. (2006), Rights, Deprivation and Poverty: Essays in Concepts and Measurement, New Delhi,
Oxford University Press.
Unel, B. (2003), ‘Productivity Trends in India’s Manufacturing Sectors in the Last Two Decades’, IMF
Working Paper No. WP/03/22, Washington, DC, International Monetary Fund.
Vaidyanathan, A. (2006), ‘Farmers’ Suicides and the Agrarian Crisis’, Economic and Political Weekly, 23
September, 4009 – 13.
Varshney, A. (2007), ‘India’s Democratic Challenge’, Foreign Affairs, 86, 93 – 106.
Virmani, A. (1997), ‘India: Crises, Reform and Growth’, Economic and Political Weekly, 9 August.
— (2004a), ‘India’s Economic Growth: From Socialist Rate of Growth to Bharatiya Rate of Growth’,
Working Paper No. 122, New Delhi, Indian Council for Research on International Economic Relations.
— (2004b), ‘Sources of India’s Economic Growth: Trends in Total Factor Productivity’, Working Paper
No. 131, New Delhi, Indian Council for Research on International Economic Relations.
Wallack, J. S. (2003), ‘Structural Breaks in Indian Macroeconomic Data’, Economic and Political Weekly,
11 October, 4312 – 15.
Williamson, J., and Zagha, R. (2002), ‘From Slow Growth to Slow Reform’, Washington, DC, Institute for
International Economics.
World Bank (2006), ‘India’s Employment Challenge’, Work in Progress, Washington, DC, World Bank.
View publication statsView publication stats
https://www.researchgate.net/publication/5216384
CONTENTS PAGE
I. Introduction …………………………………………………………………………………………………………..
4
II. Export Sophistication and Growth: Measurement and Stylized Facts ……………………………
5
III. Sharp vs. Blunt Instrument …………………………………………………………………………………….
8
IV. Empirical Results ………………………………………………………………………………………………….
9
V. Robustness ………………………………………………………………………………………………………….
12
VI. Conclusion ……………………………………………………………………………………………………….. .19
TABLES
Table 1. Growth and Export Sophistication: OLS and Fixed Effects (5-Year Panel) …………
10
Table 2. Growth and Export Sophistication: IV Estimation (5-Year Panel) ……………………..
11
Table 3. Other Measures of Export Sophistication: IV Estimation (5-Year Panel)……………..13
Table 4. The Spillover Effect: IV Estimation (5-Year Panel) …………………………………………
14
Table 5. Robustness: IV Estimation (5-Year Panel) ………………………………………………………
15
Table 6. GMM-System Estimation (5-Year Panel) ……………………………………………………….
16
Table 7. Dissecting GMM (5-Year Panel) ……………………………………………………………………
17
Table 8. IV Estimation (10-Year Panel) ………………………………………………………………………
18
Table 9. IV Estimation, Instrument Set: Median (5-Year Panel) ……………………………………. 18
Table 10. IV Estimation, Instrument Set: Weighted Mean (5-Year Panel) ……………………….
19
FIGURES
Figure 1. Log Initial Export Sophistication vs. 5-Year Ahead Log Real GDP per Capita Growth
Conditional on Initial Real GDP per Capita …………………………………………………………………..
6
Figure 2. Log Real Manufacturing Exports per Capita vs. 5-Year Ahead Log Real GDP per
Capita Growth Conditional on Initial Real GDP per Capita…………………………………………….. 6
References ………………………………………………………………………………………………………………. 21
Appendix ………………………………………………………………………………………………………………… 24
©International Monetary Fund. Not for Redistribution
Underline
Underline
Underline
4
I. INTRODUCTION
Determining the causes of economic growth is the grail sought by a large empirical growth
literature. To understand the causes of economic growth, what is sought is not an exhaustive list
of growth drivers which is probably unfathomable but rather a brief list of key determinants
of growth. However, uncovering key factors of growth is hard in practice. Data have
measurement errors; there are too many growth drivers to consider compared to data available;
and causation is hard to distinguish from correlation. In this paper, we attempt to tackle one of
the key problems in the empirical growth literature, namely, the causation vs. correlation or
endogeneity problem, in determining the key causes of growth in a cross-country setting.
The standard technique used in the literature to tackle the endogeneity problem is to use an
instrumental variable (IV) estimation. This technique relies on finding an instrumental variable
outside the model that is both relatively strongly correlated with the endogenous variable of
interest ( strength) and at the same time uncorrelated with the residual or innovation
term of the growth regression ( validity). Many growth studies seem to suffer from a
violation of one or both of these two conditions. Bazzi and Clemens (2013) showed that some
prominent growth studies focus on different determinants of growth while using the same
instrumental variables. As a result, they collectively problem.
In this case, at least one of the instrumental variable estimations must be invalid and possibly all
could be invalid.
Over the last two to three decades, the number of determinants of growth explored in the
literature grew much faster than the stock of instrumental variables available, which makes
tackling the blunt instrument issue crucial. For example, population size was used as an
instrument in a myriad of growth regressions to instrument for different endogenous variables
such as trade (Spolaore and Wacziarg 2005), international aid (Rajan and Subramanian 2008) or
export sophistication (Hausmann, Hwang, and Rodrik 2007) without necessarily controlling for
other studies explanatory variables.2 Moreover, even when the instrument does not suffer from
the problem, it could be weak, producing estimates that could misinform the reader
about the true effects on growth.3 It is not an exaggeration to say that thanks to Bazzi and
Clemens (2013) we know that we may not know much about key growth determinants from
instrumental variable growth regressions. We suggest a way to address their criticism.
In this paper we revisit the study of the main determinants of growth while avoiding the blunt
and weak instrument problems. Our instrumentation technique consists in using, as an instrument
for each endogenous variable, the average of the same variable in the neighboring marine and
land countries. The instruments we propose have the advantage of being variable-specific and
time-varying and the method produces strong instruments. The relatively
strong correlations of growth determinants between a country and its neighbors suggest that
geographic proximity can lead to imitation in trade openness, quality of institutions, education,
2 Bazzi and Clemens (2013) also argue that even when multiple instruments are used for the same endogenous
variable, in many cases population contains all the relevant information and other instruments are weak.
3 Kraay (2015), following the approach suggested in Bazzi and Clemens (2013), finds a similar weak instrument
problem in several studies of growth and inequality. A recent study by Berg et. al. (forthcoming) on growth and
inequality uses a variety of robustness checks to address this problem.
©International Monetary Fund. Not for Redistribution
5
financial development, and other policies. Moreover, we show evidence that the spillover effect
from neighbors or time-invariant country features do not affect our main conclusions.
We find that export sophistication is the only robustly significant determinant of growth among
the standard determinants such as human capital, trade, financial development, and institutions.
Moreover, in the presence of export sophistication, other standard growth determinants mostly
become statistically insignificant. One potential implication of our result is that improvements in
human capital, trade, financial development or institutions would raise economic growth to the
extent that they contribute to increasing export sophistication. We also show evidence that export
orientation of domestic production, as opposed to domestic production per se or specialization in
manufacturing, is critical.
In their seminal paper, Hausmann, Hwang, and Rodrik (2007) proposed a measure of export
sophistication and argued that it was a key causal factor of growth. The measure is based on the
weighted average of the degree of sophistication of the goods exported, which is measured by the
average GDP per capita of all the countries exporting such a good. Replicating Hausmann,
Hwang, and Rodrik (2007), Bazzi and Clemens (2013), show that in addition to the blunt
instrument problem of using the population variable as an instrument for export sophistication,
the problem of weak instruments could not be readily dismissed in the estimation. In this paper,
we first recalculate the export sophistication variable, extending it to 2014. Then we not only
resurrect the result of Hausmann, Hwang, and Rodrik (2007), but also show that export
sophistication is the only robust variable when the standard factors of growth are included in the
regression and the averages of variables in the neighboring countries are used
as instruments.
Moreover, we propose other proxies for export sophistication such as real manufacturing exports
per capita or the share of manufacturing exports in total exports of goods and find broadly
similar results.
II. EXPORT SOPHISTICATION AND GROWTH: MEASUREMENT AND STYLIZED FACTS
As the experience of many oil-exporting countries shows, in the absence of improvement in
export sophistication, economic growth is fleeting (Cherif and Hasanov 2016). Although many
oil exporters have grown for periods of time on the back of large oil income flows, sustained
growth has not materialized as productivity growth has been stagnant or even negative. The
authors argue that the main source of productivity gains stems from the production of
sophisticated tradable goods, which in turn could be proxied by the degree of sophistication of
exports. This type of production and exports have been lacking in many oil exporters.
Export sophistication has a strong positive association with the 5-year ahead real GDP per capita
growth controlling for the level of initial GDP per capita (Figure 1).4 The level of sophistication
of each good is measured as the weighted average of real GDP per capita of all countries that
export that good a proxy for the level of sophistication. If a good is typically exported by rich
countries (poor countries), it will have a high (low) sophistication level. Export sophistication
4 The plot represents the residuals of the pooled OLS regression of growth on the initial logarithm of real GDP per
capita vs. the residuals of the pooled regression of export sophistication on the initial logarithm of real GDP per
capita. The slope of the fitted line should be equal to the coefficient of export sophistication in the pooled growth
regression controlling for initial income. The plot excludes a few outliers with 5-year growth rates over 20 percent.
©International Monetary Fund. Not for Redistribution
6
(EXPY in Hausmann, Hwang, and Rodrik 2007) is defined as the export-share weighted average
of sophistication levels of the count
Figure 1. Log Initial Export Sophistication vs. 5-Year Ahead Log Real GDP per Capita Growth
Conditional on Initial Real GDP per Capita
We also use alternative proxies of export sophistication such as the share of manufacturing in
goods exports and real manufacturing exports per capita. Both measures have high coefficients
of correlation with EXPY, about 60 and 75 percent, respectively. These measures have also a
strong positive correlation with the 5-year ahead real GDP per capita growth controlling for the
level of initial real GDP per capita (Figure 2).
Figure 2. Log Real Manufacturing Exports per Capita vs. 5-Year Ahead Log Real GDP per
Capita Growth Conditional on Initial Real GDP per Capita
AGO
AGO
AGO
AGO
AGO
AGO
AGO
AGO
AGOALB
ALB
ALB
ALB ALB
ALB
ALBALB
ALB
ARE
ARE
ARE
ARE
ARE
ARE
ARE
ARE
AREARG
ARG
ARGARG
ARG
ARG
ARG
ARGARG
ARG
ARG
ARM
ARM
ARMARM
ATG
ATG
ATG
ATG
ATG
ATGATG
ATG
ATG
AUS
AUS
AUSAUSAUSAUS
AUS
AUSAUS
AUSAUS
AUT
AUT
AUTAUT
AUT
AUT
AUT
AUT
AUTAUTAUT
AZE
AZE
AZE
BDI
BDI
BDI
BDI BDI
BDI
BDI
BDI
BDI
BDI
BDI
BEL
BEL
BELBEL
BEL
BEL
BEL
BEL
BELBELBEL
BEN
BEN
BEN
BEN
BEN
BEN
BEN
BEN
BEN BEN
BEN
BFA
BFA
BFA
BFA
BFA
BFA
BFA
BFA BFA
BFA
BFA
BGD
BGD
BGD BGD
BGD
BGD
BGDBGD
BGR
BGR
BGR
BGR
BGR
BGR
BGR
BGR
BGR
BHR
BHR
BHR
BHR
BHR
BHR
BHR
BHR
BHR
BHS
BHS
BHS
BHS
BHS
BHS
BHS
BHS
BHS
BIH
BIH
BIH
BLR
BLR
BLR
BLRBLZ
BLZ
BLZ
BLZ
BLZ
BLZBLZ
BLZ BLZ
BOL
BOL
BOL
BOL
BOL
BOL
BOL
BOLBOL
BOL
BOL
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRB
BRB
BRB
BRB
BRB
BRB
BRB
BRB
BRBBRB BRB
BRN
BRN
BRN
BRN
BRN
BRN
BRN
BRNBRN
BTN
BTN
BTN
BTN
BTNBTNBTN
BTN
BTN
BWABWA
BWA
CAF
CAF
CAF
CAF
CAFCAF
CAFCAF
CAF
CAF
CAF
CAN
CANCANCAN
CAN
CAN
CAN
CAN
CAN
CAN
CAN
CHE
CHE
CHE
CHE
CHE
CHE
CHE
CHE
CHECHECHE
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHN
CHNCHN
CHN
CIV
CIV
CIV
CIV
CIV
CIV
CIV
CIV
CIV
CIV
CIV
CMR
CMR
CMRCMR CMR
CMR
CMR
CMR
CMR
CMR
CMR
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COG
COG
COG COG
COG
COG
COG
COG
COG
COGCOG
COL
COL COLCOL
COL
COL COL
COL
COL
COL
COL
COM
COM
COM
COM COM
COM COM
COM
COM
COMCOM
CPV
CPV
CPV
CPV
CPV CPV
CPV
CPV
CPV
CPV
CRI CRI
CRI
CRI
CRI
CRI
CRICRI CRI
CRICRI
CYP
CYP
CYP
CYP
CYP
CYP
CYPCYPCYP
CYP
CYP
CZE
CZE
CZE
CZE
DEU
DEU
DEU
DEU
DEU
DEU
DEUDEU
DEU
DEUDEU
DJI
DJI
DJI
DJI
DJI
DJI
DJI
DJI
DJI
DMA
DMA
DMA
DMA
DMA
DMA
DMA
DMA
DMA
DNKDNK
DNK
DNKDNK
DNK
DNK
DNK
DNK
DNKDNK
DOM
DOM
DOM
DOM
DOM
DOM
DOM
DOM
DOM
DOM
DOM
DZA
DZA
DZA
DZA
DZA
DZADZA
DZA
DZA
DZADZAECU
ECU
ECU
ECU
ECU
ECU
ECU
ECU
ECU
ECU
ECU
EGY
EGY
EGY
EGY
EGY
EGYEGY
EGY
EGY
EGY
EGY
ESP
ESP
ESP
ESPESP
ESP
ESP
ESP
ESP
ESP
ESP
EST
EST
EST
EST
ETH
ETH
ETH
ETH
ETH
ETH
ETH
ETH
ETH
ETH ETH
FIN
FIN
FIN
FINFIN
FIN
FIN
FIN
FIN
FIN
FIN
FJI
FJI
FJI
FJI
FJI
FJI
FJIFJI
FJI
FJI
FJI
FRA
FRA
FRAFRA
FRA
FRA
FRA
FRA
FRA
FRAFRA
GAB
GAB
GAB
GAB
GAB
GAB
GAB
GAB GAB
GAB
GBR
GBRGBRGBRGBR
GBR
GBR
GBRGBR
GBR
GBR
GEO
GEO
GEOGEO
GHA
GHA
GHA
GHA
GHA
GHA GHA
GHA
GHA
GHA
GHA
GIN
GIN
GIN
GIN
GIN
GIN
GIN
GIN GIN
GIN GIN
GMB
GMB
GMB
GMB
GMB
GMB GMB
GMB
GMB
GMB
GMB
GNB
GNB
GNB
GNB
GNB
GNB
GNB
GNB
GNB
GNB
GNB
GNQ
GNQ
GNQ
GNQ
GNQ
GNQ
GNQ
GNQ
GNQ
GRC
GRC
GRC
GRC
GRC
GRCGRC
GRC
GRC
GRC
GRC
GRD
GRD
GRD
GRD
GRD
GRD
GRD
GRD
GRD
GTM GTM GTM
GTM
GTM
GTM
GTMGTM
GTM
GTMGTM
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HND
HND HND
HND
HND
HND
HNDHND
HND
HNDHND
HRVHRV
HRV
HRV
HTI
HTI
HTI
HTI
HTI
HTI
HTI
HTI
HTI
HTI
HTI
HUN
HUN
HUN
HUN
HUN
HUN
HUN
HUN
HUN
IDN
IDN
IDN IDN
IDN
IDN
IDN
IDN
IDN
IDNIDN
IND
IND
INDIND
IND
IND
IND
IND
IND
IND
IND
IRL
IRL
IRL
IRL
IRL
IRLIRL
IRL
IRL
IRL
IRL
IRN
IRN
IRN
IRN
IRN
IRN
IRN IRN
IRN IRN
IRN
IRQ
IRQ
IRQ
IRQ
IRQ
IRQ
IRQ
IRQ
ISL
ISL
ISLISL
ISL
ISL
ISL
ISL
ISL
ISL
ISL
ISR
ISR
ISR
ISR
ISR
ISR
ISRISR
ISR
ISR ISR
ITA
ITA
ITA
ITA
ITA
ITA
ITA
ITA
ITA
ITAITA
JAM
JAM
JAM
JAM
JAM
JAM
JAM
JAM
JAM
JAM
JAM
JOR
JOR
JOR
JOR
JOR
JOR
JORJOR
JOR
JOR
JOR
JPN
JPN
JPNJPNJPN
JPN
JPNJPNJPNJPNJPN
KAZ
KAZ
KAZ
KAZ
KEN
KEN KEN
KEN
KEN
KEN
KEN
KEN
KEN
KENKEN
KGZ
KGZKGZKGZ
KHM
KHM
KHM
KHM
KHM
KHM
KHM
KHMKHM
KNA
KNA
KNA
KNA
KNA
KNA
KNA
KNA
KNA
KOR
KOR
KOR
KOR
KOR
KOR
KOR
KOR
KOR
KOR
KOR
KWT
KWT
KWT
KWT
KWT
KWT
KWT
KWT
KWT
LAO
LAO
LAO
LAO
LAO
LAO
LAO
LAOLAO
LBN
LBN
LBN
LBN
LBN
LBNLBN
LBN
LBN
LBR
LBR
LBR
LBR
LBR
LBR
LBR
LBR
LBR
LCA
LCA
LCA
LCA
LCA
LCA
LCALCA
LCA
LKA
LKA
LKA
LKA LKA
LKA
LKA
LKA
LKA
LKA
LKA
LSO
LSO
LSO
LTU
LTU
LTU
LTU
LUX
LUXLUX
LVA
LVA
LVA
LVA
MAR
MAR
MARMAR
MAR
MAR
MAR
MAR
MARMAR
MAR
MDA
MDA
MDA
MDA
MDG
MDG
MDG
MDG
MDG
MDG
MDG
MDG
MDG
MDG MDG
MDV
MDV
MDV
MDV
MDV
MDV
MDV
MDV
MDV
MEX
MEX MEX
MEX
MEXMEX MEX
MEX
MEXMEX
MEX
MKD
MKD
MKD
MKD
MLI
MLI
MLI
MLI
MLI
MLI
MLI
MLI
MLI
MLI
MLI
MLT
MLT
MLT
MLT
MLT
MLT
MLTMLT
MLTMLT
MLT
MMR
MMR
MMR
MMR
MMR
MMR
MMR
MMR
MMR
MMR
MNE
MNG
MNG
MNG
MNG
MNG
MNG
MNGMNG
MNG
MOZ
MOZ
MOZ
MOZ
MOZ
MOZ
MOZ
MOZ
MOZ
MOZMOZ
MRT
MRT
MRT
MRT
MRT
MRTMRT
MRT
MRT
MRT
MUS
MUS
MUS
MUS
MUS
MUS
MUS
MUS
MUS
MUS
MUS
MWI
MWI
MWI
MWIMWI
MWI
MWIMWI
MWI
MWI
MYS
MYS
MYS
MYS
MYS
MYS
MYSMYSMYS
MYS
NAMNAMNAM
NER
NER
NER
NER
NER
NER
NER
NER
NERNER
NER
NGA
NGA
NGA
NGA
NGA
NGA
NGA
NGA
NGA
NGA
NGA
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NLD
NLD
NLD
NLD
NLD
NLD
NLD
NLD
NLDNLD
NLD
NOR
NOR
NOR
NOR
NOR
NOR
NORNOR
NOR
NOR
NOR
NPL
NPL
NPLNPL
NPLNPLNPLNPL NPL
NPLNPL
NZL
NZL
NZL
NZL
NZL
NZL
NZLNZL
NZL
NZL
NZL
OMN
OMN
OMN
OMN
OMN
OMN
OMN
OMN
OMN
PAK
PAK
PAK
PAK
PAKPAK
PAK
PAK
PAK
PAKPAK
PAN
PAN
PAN
PAN
PAN
PAN
PAN
PANPAN
PAN
PAN
PER
PER
PER
PER
PER
PER
PER
PER
PER
PER
PER
PHL
PHL
PHLPHL
PHL
PHL
PHL
PHL
PHL
PHL
PHL
POL
POL
POLPOL
POL
POL
POL
POL
POL
PRT
PRT
PRT
PRT
PRT
PRT
PRT
PRT
PRTPRT
PRT
PRY
PRY
PRY
PRY
PRY
PRY PRY
PRY
PRYPRY PRY
QAT
QAT
QAT
QAT
QAT
QAT
QAT
QAT
QAT
ROM
ROM
ROM
ROM
ROM
ROM ROM
ROM
ROM
ROM
ROM
RUS
RUS
RUS
RUS
RWA
RWA
RWA
RWA
RWA
RWA
RWA
RWARWA
RWA
SAU
SAU
SAU
SAU
SAU
SAU
SAU
SAU
SAU
SDN
SDN
SDN
SDNSDN
SDN
SDN
SDN
SDN
SEN
SEN
SEN
SEN
SENSEN
SEN
SEN
SEN
SENSEN
SGP
SGP
SGP SGP
SGP
SGP
SGP
SGP
SGP
SGP
SGP
SLE
SLE
SLE SLESLE
SLE
SLE
SLE
SLE
SLE
SLE
SLV
SLV
SLV
SLV
SLV
SLV
SLV
SLVSLV
SLV
SLV
SRB
SRB
STP
STP
STP STP
STP STP
STP
STP
STP
SUR
SURSUR
SURSUR
SUR
SUR
SUR
SUR
SVK
SVKSVK
SVK
SVN
SVN
SVN
SVN
SWE
SWE
SWE
SWE
SWESWE
SWE
SWE
SWE
SWESWE
SWZ
SWZSWZ
SYC
SYC
SYC
SYC
SYC
SYC
SYC
SYC
SYC
SYR
SYR
SYR
SYR
SYR
SYR
SYR
SYR
SYR
SYR
SYR
TCD
TCD
TCD
TCD
TCDTCD
TCD TCD
TCD
TCD
TCD
TGO
TGO
TGO
TGO
TGO
TGO
TGO
TGO
TGO
TGO
TGO
THA
THA
THA
THA
THA
THA
THA
THA
THA
THA
THA
TJK
TJK
TJK
TJK
TKM
TKM
TKM
TKM
TTO
TTO
TTO
TTO
TTO
TTO
TTO
TTOTTO
TTO
TTO
TUN
TUN TUN
TUN
TUN
TUN
TUN
TUN
TUNTUN
TUN
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TWN
TWN
TWN
TWN
TWN
TWN
TWN
TWN
TWNTWN
TWN
TZA
TZA
TZA
TZA
TZA
TZA
TZA
TZA
TZA
TZA
UGA
UGA
UGA
UGA
UGA
UGA
UGA
UGA
UGA
UGA
UGA
UKR
UKR
UKR
UKR
URY
URYURY
URY
URY
URYURY
URY
URY
URY
URY
USA
USA
USA
USAUSAUSA
USA
USA
USA
USA
USA
UZB
UZB
UZB
UZB
VCT
VCT
VCT
VCTVCT
VCT
VCT
VCT VCT
VEN
VEN
VEN
VEN
VEN
VEN
VEN
VEN
VEN
VEN
VEN
VNM
VNM
VNM
VNM
VNM
VNM
VNM
VNM
VNM
YEM
YEM
YEM
YEM
YEM
ZAF
ZAF
ZAFZAF
ZAF
ZAF
ZAF
ZAF
ZAFZAF
ZAF
ZMB
ZMB
ZMB
ZMB
ZMB
ZMB
ZMB
ZMB
ZMB
ZMB
ZWE
ZWE
ZWE
ZWE
ZWE
ZWEZWE
ZWE
ZWE
ZWE
-.
2
-.
1
0
.1
.2
5
-Y
e
a
r
re
a
l
G
D
P
p
e
r
c
a
p
it
a
g
ro
w
th
|
i
n
it
ia
l
in
c
o
m
e
-2 -1 0 1 2
Log initial export sophistication | initial income
AGO
AGO
AGO
ALBALB
ALB
ARE
ARE
ARGARG
ARG ARG
ARG
ARG
ARG
ARGARG
ARG
ARG
ARM
ARMARM
ATG
ATG
ATG
ATG
AUS
AUSAUSAUSAUS
AUS
AUSAUS
AUSAUS
AUT
AUTAUT
AUT
AUT
AUT
AUT
AUTAUTAUT
AZE
AZE
BDI
BDI
BDI
BDI
BDI
BDI
BEL
BEL
BELBEL
BEL
BEL
BEL
BEL
BELBELBEL
BENBEN
BEN
BEN BEN
BENBEN
BEN
BFA
BFA
BFA
BFA
BFA
BFABFA
BFA
BFA
BGD
BGDBGD
BGD
BGD
BGDBGD
BGR
BGR
BGR
BHR
BHR
BHR
BHR
BHR
BHR
BHR
BHR
BHR
BHS
BHS
BHS
BHS
BHS
BHS
BIH
BIH
BLR
BLR
BLR BLZ
BLZ
BLZ
BLZBLZBLZ
BLZBLZ
BOL
BOL
BOL
BOL
BOL
BOL
BOL
BOLBOL
BOL
BOL
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRA
BRB
BRB
BRB
BRB
BRB
BRB
BRBBRBBRB
BRN
BTN
BTN
BWABWA
BWA
CAF
CAF
CAF
CAF CAF
CAF
CAF
CAF
CAF
CAN
CANCANCAN
CANCAN
CAN
CAN
CAN
CAN
CAN
CHECHE
CHE
CHE
CHE
CHE
CHE
CHE
CHECHECHECHL
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHL
CHN
CHN
CHN
CHNCHN
CHN
CIV
CIV
CIV
CIVCIV
CIV
CIV
CIV
CIV
CMR
CMR
CMRCMRCMR
CMR
CMR
CMR
CMR
CMR
COG
COG
COGCOG
COG
COG
COG
COG
COL
COLCOL COL
COL
COL COL
COL
COL
COL
COL
COM
COM
COM
COM
CPV
CPV
CPV
CPV
CPV
CRICRI
CRI
CRI
CRI
CRICRI CRI
CRICRI
CYP
CYP
CYP
CYP
CYPCYPCYP
CYP
CYP
CZE
CZE
CZE
CZE
DEU
DEU
DEU
DEU
DEU
DEU
DEUDEU
DEU
DEUDEU
DJI
DMA
DMA
DMA
DMA
DMA
DMA
DMA
DNKDNK
DNK
DNKDNK
DNK
DNK
DNK
DNK
DNKDNK
DOM
DOM
DOM
DOMDOM
DOM
DZA
DZA
DZA
DZADZA
DZA
DZA
DZADZAECUECU
ECU
ECU
ECUECU
ECU
ECU
ECU
ECU
ECU EGY
EGY
EGY
EGY
EGY EGY
EGY
EGY
EGY
EGY
ESP
ESP
ESP
ESP ESP
ESP
ESP
ESP
ESP
ESP
ESP
EST
EST
EST
EST
ETH
ETH
ETH ETH
FIN
FIN
FINFIN
FIN
FIN
FIN
FIN
FIN
FIN
FJI
FJI
FJI
FJI
FJI
FJI
FJI
FJI
FRAFRA
FRAFRA
FRA
FRA
FRA
FRA
FRA
FRAFRA
GAB
GAB
GAB
GAB GAB
GBR
GBRGBRGBRGBR
GBR
GBR
GBRGBR
GBR
GBR
GEO
GEOGEO
GHA
GHA
GHA
GHA
GHA
GHA
GHA
GHA
GINGIN
GIN
GMB
GMB
GMB
GMB
GMB
GMB
GNB
GNB
GNB
GNQ
GRC
GRC
GRC
GRC
GRC
GRCGRC
GRC
GRC
GRC
GRC
GRD
GRD
GRD
GRD
GRD
GRD
GTM
GTMGTM
GTM
GTM
GTMGTM
GTM
GTMGTM
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HKG
HND HND
HND
HND
HND
HNDHND
HND
HNDHND
HRVHRV
HRV
HRV
HTI
HTI
HTI
HTI
HUN
HUN
HUN
HUN
HUN
HUN
HUN
HUN
HUN
IDN
IDN IDN
IDN
IDN
IDN
IDN
IDN
IDNIDN
IND
IND
INDIND
IND
IND
IND
IND
IND
IND
IND
IRL
IRL
IRL
IRL
IRLIRL
IRL
IRL
IRL
IRL
IRN
IRN
IRN
IRNIRN
IRN
IRQ
IRQ
IRQ
ISL
ISL
ISLISL
ISL
ISL
ISL
ISL
ISL
ISL
ISL
ISR
ISR
ISR
ISR
ISR
ISR
ISRISR
ISR
ISRISR
ITA
ITA
ITA
ITA
ITA
ITA
ITA
ITA
ITA
ITAITA
JAM
JAM
JAM
JAM
JAM
JAM
JAM
JAM
JAM
JOR
JOR
JOR
JOR
JOR
JORJOR
JOR
JOR
JOR
JPN
JPN
JPNJPNJPN
JPN
JPNJPNJPNJPNJPN
KAZ
KAZ
KAZ
KAZ
KEN
KEN
KEN
KEN
KEN
KENKEN
KGZ
KGZKGZ KGZ
KHM
KHM
KHMKHM
KNA
KNA
KNA
KNA
KOR
KOR KOR
KOR
KOR
KOR
KOR
KOR
KOR
KOR
KOR
KWT
KWT
KWT
KWT
KWT
KWT
LAO
LBN
LBN
LBN
LBN
LBR
LBR
LBR
LCA
LCA
LCA
LCA
LCA
LCALCA
LKA
LKA
LKA
LKA LKA
LKA
LKA
LKA
LKA
LKA
LSO
LSO
LTU
LTU
LTU
LTU
LUX
LUXLUX
LVA
LVA
LVA
LVA
MAR
MAR
MAR MAR
MAR
MAR
MAR
MAR
MARMAR
MAR
MDA
MDA
MDA
MDA
MDG
MDG
MDG
MDG
MDG
MDG
MDG
MDG
MDG
MDGMDG
MDV
MDV
MDV
MEX
MEXMEX
MEX
MEX MEX MEX
MEX
MEXMEX
MEX
MKD
MKD
MKD
MKD
MLI
MLI MLI
MLI
MLI
MLI
MLI
MLI
MLI
MLT
MLT
MLT
MLT
MLT
MLT
MLTMLT
MLTMMR
MMR
MMR
MMR
MNGMNG
MOZ
MOZ
MOZMOZ
MRT
MRT
MRT
MUS
MUS MUS
MUS
MUS
MUS
MUS
MWI
MWI
MWIMWI
MWI
MWIMWI
MWI
MWI
MYS
MYS
MYS
MYS
MYS
MYS
MYSMYSMYS
MYS
NAMNAMNAMNER
NER
NER
NER
NER
NER NER
NER
NGANGA
NGA
NGA
NGA
NGA
NGA
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NIC
NLDNLD
NLDNLD
NLD
NLD
NLD
NLD
NLDNLD
NLD
NOR
NOR
NORNOR
NOR
NOR
NORNOR
NOR
NOR
NOR
NPL
NPL NPLNPLNPLNPL
NPL
NZL
NZL
NZL
NZL
NZL
NZLNZL
NZL
NZL
NZL
OMN
OMN
OMN
OMN
OMN
OMN
OMN
PAKPAK
PAK
PAKPAKPAK
PAK
PAK
PAK
PAKPAK
PAN
PAN
PAN
PAN
PAN
PAN
PAN
PANPAN
PAN
PAN
PERPER
PER
PER
PER
PER
PER
PER
PER
PER
PER
PHLPHL
PHL PHL
PHL
PHL
PHL
PHL
PHL
PHL
PHL
POL
POL
POL
POL
POL
PRT
PRT
PRT
PRT
PRT
PRT
PRT
PRT
PRTPRT
PRT
PRYPRY
PRY
PRY
PRY
PRY PRY
PRY
PRYPRYPRY
QAT
QAT
QAT
QAT
QAT
QAT
RUS
RUS
RUS
RWA
RWA
SAU
SAU
SAU
SAU
SAU
SAU
SAU
SAU
SDN
SDN
SDN
SDN
SDN
SDN
SDN
SDN SENSEN
SEN
SEN
SEN
SEN
SEN
SENSEN
SGP
SGP
SGP SGP
SGP
SGP
SGP
SGP
SGP
SGP
SGP
SLE
SLE
SLV
SLV
SLV
SLV
SLV
SLV
SLVSLV
SLV
SLVSTP
STP
STP
SUR
SUR
SUR
SUR
SUR
SVK
SVKSVK
SVK
SVN
SVN
SVN
SVN
SWE
SWE
SWE
SWE
SWESWE
SWE
SWE
SWE
SWESWE
SWZ
SWZ
SYC
SYC
SYC
SYC
SYC
SYC
SYCSYR
SYR
SYR
SYR
SYR
SYR
SYR
SYR
TCD
TCD
TCD
TCD
TGO
TGO
TGO
TGO
TGO
TGO
TGO
TGO
TGO
TGO
THA
THA
THA
THA
THA
THA
THA
THA
THA
THA
THA
TJK
TKM
TTO
TTO
TTO
TTO
TTO
TTOTTO
TTO
TTO
TUN
TUNTUN
TUN
TUN
TUN
TUN
TUN
TUNTUN
TUN
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TUR
TZA
TZA
TZA
TZAUGA
UGA
UGA
UGA
UKR
UKR
UKR
URY
URY
URY
URYURY
URY
URY
URY
URYUSA
USA
USA
USAUSAUSA
USA
USA
USA
USA
USA
VCT
VCT
VCT
VCTVCT
VEN
VENVEN
VEN
VEN
VEN
VEN
VEN
VEN
VEN
VEN
VNM
VNM
VNM
YEM
YEM
YEM
YEM
ZAF
ZAF
ZAF
ZAFZAF
ZAF
ZMB
ZMB
ZMB
ZMB
ZMB
ZMB
ZWE
ZWEZWE
ZWE
ZWE
ZWE
-.
2
-.
1
0
.1
.2
5
-Y
e
a
r
re
a
l
G
D
P
p
e
r
c
a
p
it
a
g
ro
w
th
|
i
n
it
ia
l
in
c
o
m
e
-6 -4 -2 0 2 4
Log initial real manufacturing exports per capita | initial income
©International Monetary Fund. Not for Redistribution
7
In a macro setting, finding valid and strong instruments is not a straightforward task and the
GMM estimation, using own lagged variables, should bypass this problem (and in
theory could help avoid the blunt instrument issue). Unfortunately, many of these instruments
turn out to be weak, especially when using a smaller number of instruments. Bazzi and Clemens
(2013) show that several of the seminal papers they examine do not survive the opening of the
, use their term. In particular, the authors analyze the excludability of the
country size (population and area) instrument and examine the instrument validity (e.g. using the
test of Hahn, Ham, and Moon 2011), perform various tests of underidentification of the
instrument set and tests of weak instruments (e.g. using Kleibergen-Paap and Cragg-Donald test
statistics), and use estimation methods robust to weak instruments (e.g. a testing procedure using
Kleibergen 2002). Interestingly, the only determinant of growth, which seemed to broadly
survive their comprehensive analysis, is the export sophistication variable of Hausmann, Hwang,
and Rodrik (2007). However, even in this case, the regressions do not pass all the tests. Weak
instrument tests fail in the case of smaller or collapsed number of instruments, and the validity
of the population variable is questioned. However, the weak-instrument robust confidence set
estimation indicates that the export sophistication variable has a positive and statistically
significant effect.
In the following section, we study the standard determinants of growth put forth by the literature
in addition to export sophistication human capital, quality of institutions, trade openness,
financial development, foreign direct investment (FDI), saving rates, investment, government
size, and the Gini coefficient. We use real GDP per capita for 1960-2014 from the Penn World
Tables 9.0 (Feenstra, Inklaar, and Timmer 2015). Years of schooling come from Barro and Lee
(2013).
Export sophistication data (EXPY) of Hausmann, Hwang, and Rodrik (2007) are computed using
the World Trade Flows data (Feenstra and Romalis 2014) for 1962-2014. We also compute a
structural EXPY measure S-EXPY which discounts the high share of commodity exports of
high-income commodity exporters correcting the artificially high EXPY of commodity exporters
(see Appendix for details).
trade (exports plus imports, percent of GDP) as a measure of trade openness, domestic credit to
the private sector (percent of GDP) as a measure of financial development, manufacturing
exports (in constant USD and percent of GDP), manufacturing production (real value added in
local currency and percent of GDP), FDI (percent of GDP), government consumption (percent of
GDP) as a measure of government size, and gross fixed capital formation (percent of GDP). The
Economic Outlook database, and the Gini
coefficient is from SWIID v4. The law and order indicator, measuring the strength and
impartiality of the legal system and the assessment of popular observance of the law, is used as a
proxy of the quality of institutions and
database (the data start in 1984). We also use a corruption indicator from the same source.
©International Monetary Fund. Not for Redistribution
8
III. SHARP VS. BLUNT INSTRUMENT
In this section, we describe our instrumental variable methodology. Finding valid and strong
instruments in the cross-country setting is challenging. As argued by Bazzi and Clemens (2013),
many papers use the same instruments such as population and area for different variables. In
addition, these instruments suffer from validity and possibly weak instrument problems. To
illustrate the blunt instrument problem, suppose that growth could be (potentially) explained by
two factors and such that:
Let us assume that two studies use growth regressions (A) and (B) which have the following
forms:
Let us suppose that one study uses (A) and instrumental variable , and a second study uses (B)
while relying on the same instrumental variable . Assuming that is a valid instrument in both
(A) and (B) and that and/or are significant determinants of growth, is problematic. Indeed,
(A) could be re-written as , while (B) could be re-written as
. If is correlated with , then the latter is also correlated with the error term of
(A), and the same logic applies to (B). In other words, at least one of the studies relies on an
invalid instrumental variable (and it could be the case for both).
To remedy the blunt instrument problem, we propose the sharp instrument solution. Our
method instruments for variables of a country with the average values of these variables in its
neighboring countries. The advantage of this IV method is that it generates variable-specific
instruments and can be applied to a wide range of explanatory variables thus bypassing the
problem of blunt instruments described above. We also test for the strength of our instruments
(correlation with the variables for which they are instruments).
We argue that using the average of a variable in neighboring countries as an instrument is likely
to satisfy the exclusion restriction from the growth equation (validity of instruments) while at the
same time, it should be variable. The exclusion
restriction requires that the innovation or error term in the growth regression be uncorrelated
with the instruments for explanatory variables the average values of those variables in
neighboring countries. If the validity assumptio
In contrast, some
the growth regressions. Chua (1993) and Ades and Chua (1993) show that various practices and
traits that are unfavorable to growth could spill over from neighboring countries and add simple
regression. Easterly and Levine (1998) control
for the neighbor th regression and instrument it with the
Growth and its determinants in neighboring countries could be related in several ways. We
©International Monetary Fund. Not for Redistribution
9
propose different methods and controls to verify that our instruments are not invalid due to some
unaccounted correlation with the residuals of the growth regression. Governments, firms and
households in neighboring countries could imitate each other because of regional competition,
common languages, or cultural proximity.5 In particular, Bahar, Hausmann, and Hidalgo (2014)
show that a country is more likely to export a product if its neighbor is exporting it. This type of
effect would explain the strength of our instrument without invalidating it.
A country could be affected by spillovers from its neighbors mostly, but not exclusively, through
trade and finance. Being close to a country that is growing fast could encourage FDI and
technological transfers as was the case in East Tigers. 6 We offer several types of
robustness checks to verify that our instrument remains valid (see the next section). First, we
control for the average real GDP per capita or real GDP in neighboring countries as a proxy for
the spillover effect. Second, we modify the weighting of the instrument to mitigate a potential
violation of the exclusion restriction of instruments based on simple averages. We use the
median of variables of neighboring countries and the weighted average of variables of neighbors,
in which weights are inversely proportional to real GDP. The median neighbor is less likely to be
the main trading partner of a country, while the weights based on the inverse of real GDP
mitigate the impact of large neighbors on the construction of instruments. This weighting scheme
is inversely related to the size, a key predictor of trade links in the gravity model, assigning a
smaller weight to bigger neighbors.
Neighboring countries in general share and climate,
which are likely to affect growth. This could invalidate our instruments if not accounted for. We
use latitude, ethnic fractionalization, and a dummy for Sub-Saharan Africa to control for some of
these features. We also run a fixed effect IV regression. If we properly control for spillovers,
common traits
iction.
IV. EMPIRICAL RESULTS
Running ordinary least squares (OLS) and fixed effects (FE) regressions (Table1), we find that
many standard growth determinants are correlated with the growth rate. Regressing growth on
initial log real GDP per capita and export sophistication and controlling for each standard
determinant of growth separately (columns 1-5) yields mostly highly statistically significant
coefficients (law and order is, however, not statistically significant) with the expected signs
except for private sector credit. The coefficient on private sector credit is negative but this could
be due to potential nonlinearities in the finance-growth nexus found in the literature (e.g. Arcand,
Berkes, and Panizza 2015 and Demetriadis and Rousseau 2016). Increasing private credit could
5 Riva
in a sporting contest.
6 Typically, emerging and low-income countries have strong trade and financial links with advanced countries or
large emerging markets, which are remote. Meanwhile, a developing economy is usually surrounded by other
developing economies with little trade and financial links. In the absence of such links, it is plausible that there is no
spillover effect from neighboring countries.
©International Monetary Fund. Not for Redistribution
10
be correlated with higher vulnerabilities, financial instability and lower growth (e.g. Popov 2014
and Levine, Lin, and Xie 2016).
The coefficient estimates for export sophistication we find when including more controls,
are
consistent with the relationship shown in Figure 1. A 10 percent increase in export sophistication
is associated with about 0.2-0.3 percent increase in the annual growth rate. The regression with
all variables (columns 6-7) also yields statistically significant coefficients with the expected
signs for all the variables. The same regressions with fixed effects result in a similar statistically
significant estimate on export sophistication as in OLS regressions (columns 8-11). However,
adding law and order reduces the sample size substantially and makes the export sophistication
parameter statistically insignificant (columns 12-14). These regressions, however, show us
correlations between growth and standard growth determinants and to infer causality, we turn to
instrumental variable estimations.
Table 1. Growth and Export Sophistication: OLS and Fixed Effects (5-Year Panel)
Using an instrumental variable estimation, based on the average of variables of neighboring
countries as an instrument for each explanatory variable, we find that export sophistication is the
key determinant of growth (Table 2). A 10 percent increase in export sophistication, measured by
EXPY, increases the average annual growth rate in the next 5 years by about 0.6-0.7 percent.
This result, namely the statistical significance and the magnitude of the coefficient on export
sophistication, is robust across most estimations and is about as robust as the initial real GDP per
capita.7 The coefficient obtained is two to three times larger than the coefficient in OLS or FE
regressions suggesting a large downward bias. In column 1, real GDP per capita is assumed to be
exogenous, while it is not assumed to be exogenous in column 2. In each case, our instruments
are
value of the instrument for export sophistication for Mexico is the average of export
7 All estimations include a constant and time dummies (not shown). In this specification, fixed effects are subsumed
variables. It is a less
-LEV
Dependent variable: (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)
5-year ave. annual growth rate OLS OLS OLS OLS OLS OLS OLS FE FE FE FE FE FE FE
Log export sophistication 0.027*** 0.024*** 0.027*** 0.029*** 0.029*** 0.024*** 0.022*** 0.020*** 0.026*** 0.024*** 0.012 0.008
(0.004) (0.004) (0.004) (0.004) (0.006) (0.005) (0.006) (0.005) (0.006) (0.006) (0.010) (0.010)
Log real GDP per capita -0.010*** -0.015*** -0.012*** -0.010*** -0.012*** -0.010*** -0.015*** -0.044*** -0.050*** -0.052*** -0.049*** -0.097*** -0.072*** -0.070***
(0.002) (0.002) (0.002) (0.002) (0.003) (0.002) (0.002) (0.006) (0.008) (0.006) (0.005) (0.023) (0.012) (0.012)
Years of schooling 0.003*** 0.004*** 0.004*** -0.003 -0.003 -0.002
(0.001) (0.001) (0.001) (0.002) (0.002) (0.002)
Trade (% of GDP) 0.017*** 0.005** 0.004** 0.037** 0.010 0.014**
(0.005) (0.002) (0.002) (0.015) (0.008)
(0.007)
Credit to private sector (% of GDP) -0.007** -0.010*** -0.013*** -0.010* -0.028*** -0.028***
(0.003) (0.003) (0.003) (0.005) (0.005) (0.005)
Law and order 0.001 0.003*** 0.002** -0.000 0.004** 0.004**
(0.001) (0.001) (0.001) (0.002) (0.002) (0.002)
Observations 1,592 1,226 1,376 1,333 748 609 601 1,592 1,226 1,376 1,333 748 609 601
Adjusted R-squared 0.082 0.136 0.118 0.092 0.088 0.119 0.159 0.156 0.216 0.207 0.211 0.284 0.301 0.300
\# of countries 171 137 168 167 134 117 117
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
©International Monetary Fund. Not for Redistribution
11
i.e. Belize, Cuba, Guatemala, Honduras, and the U.S. In
both specifications, tests for weak instruments suggest that the instruments are strong. The
conditional likelihood ratio (CLR) confidence set of Moreira (2003), which is robust to the
weak-instrumentation of one endogenous variable (column 1), indicates that the coefficient on
export sophistication is in the range of 0.06 to 0.09. We find that the coefficient estimates stay
within this range as we add more variables to the regression (columns 3-5) and increase to about
0.1-0.15 in other specifications (columns 6-9).
Table 2. Growth and Export Sophistication: IV Estimation (5-Year Panel)
Controlling for other determinants of growth (one at a time) such as human capital, law and
order, trade, and financial development do not much affect the coefficient or significance of
export sophistication or initial income (columns 3-6). All variables are considered endogenous
and are instrumented using the average values of those variables in neighboring countries. Years
of schooling and trade are not statistically significant (columns 3-4). The effects of credit and
law and order are negative (columns 4-6), albeit at lower significance levels than those for export
sophistication or initial income. The negative IV coefficient on credit is also obtained in OLS
and FE regressions. The negative coefficient on law and order is more surprising, especially
since the coefficient in OLS and FE regressions, is positive. One potential explanation is that it
could also exhibit some nonlinearities similar to the private credit variable. In addition, this
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Dependent var.: 5-year ave. annual growth rate IV IV IV IV IV IV IV IV IV-FE
Log export sophistication 0.072*** 0.063*** 0.065*** 0.074*** 0.077*** 0.105*** 0.148*** 0.154**
(0.009) (0.009) (0.013) (0.013) (0.012) (0.028) (0.055) (0.070)
Log real GDP per capita -0.024*** -0.020*** -0.024*** -0.024*** -0.023*** -0.021*** -0.007* -0.026*** -0.116***
(0.003) (0.003) (0.004) (0.005) (0.003) (0.006) (0.004) (0.008) (0.043)
Years of schooling 0.001 0.004** -0.003 0.057
(0.001) (0.002) (0.004) (0.046)
Trade (% of GDP) -0.017 -0.006 -0.047
(0.013) (0.011) (0.032)
Credit to private sector (% of GDP) -0.023* -0.001 -0.003
(0.013) (0.013) (0.023)
Law and order -0.016** 0.001 -0.013*
(0.006) (0.003) (0.007)
Observations 1,590 1,590 1,216 1,369 1,319 748 606 598 983
\# of endogenous variables 1 2 3 3 3 3 5 6 3
\# of instruments 13 13 13 14 14 9 11 12 14
\# of excluded instruments 1 2 3 3 3 3 5 6 6
Cragg-Donald F stat 347.0 136.2 40.7 12.9 37.9 9.6 4.4 0.8 1.4
Kleibergen-Paap F stat 267.5 104.8 38.4 10.8 24.8 7.7 2.5 1.0 1.3
Kleibergen-Paap LM test p-value 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.11
H_0: t-test size>10% (p-value) | KP 0.00 0.00 0.00 0.17 0.00 0.47 1.00 1.00 1.00
H_0: t-test size>25% (p-value) | KP 0.00 0.00 0.00 0.00 0.00 0.01 0.79 1.00 0.99
H_0: t-test size>10% (p-value) | CD 0.00 0.00 0.00 0.07 0.00 0.26 1.00 1.00 1.00
H_0: t-test size>25% (p-value) | CD 0.00 0.00 0.00 0.00 0.00 0.00 0.33 1.00 0.99
H_0: t-test rel-bias>10% (p-value) | KP 0.00 0.00 0.00 0.00 0.00 0.04 0.95 1.00 1.00
H_0: t-test rel-bias>30% (p-value) | KP 0.00 0.00 0.00 0.00 0.00 0.00 0.43 0.94 0.88
H_0: t-test rel-bias>10% (p-value) | CD 0.00 0.00 0.00 0.00 0.00 0.01 0.68 1.00 1.00
H_0: t-test rel-bias>30% (p-value) | CD 0.00 0.00 0.00 0.00 0.00 0.00 0.08 0.97 0.85
Hansen J-test p-value 0.17
Lower CLR bound 0.06
Upper CLR bound 0.09
H0: Beta_EXPY=0 | CLR p-value 0.00
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
©International Monetary Fund. Not for Redistribution
12
regression exhibits a potential weak instrument problem. Kleibergen-Paap (KP) Wald statistic is
much smaller, and the null hypothesis that the actual size of the t-test of the coefficients equal
zero at the 5 percent level is greater than 10 percent cannot be rejected (but is rejected if the size
is greater than 25 percent). The tests for weak instruments in other estimations suggest that the
instruments are not weak.
In the IV regression with all the determinants of growth while excluding export sophistication
(column 7), only years of schooling and initial income remain significant although the coefficient
on initial income drops sharply. With export sophistication (column 8), years of schooling
becomes insignificant; law and order has a negative coefficient at 10 percent significance level;
and export sophistication and initial income are strongly significant with a larger coefficient of
0.15 on export sophistication. In these two regressions, we no longer obtain favorable statistics
for strong instruments.8 It is likely that including several endogenous explanatory variables in the
growth regression results in the weak instrument problem as endogenous variables and
instruments are probably all correlated with each other resulting in weak identification.
Nonetheless, the coefficient on export sophistication is statistically significant at 1 percent level
in this specification as well. Since the effect of years of schooling in the regression without
export sophistication is statistically significant and has a meaningful sign, we specify our
baseline regression with initial income, years of schooling and export sophistication. Lastly, we
add fixed effects to this specification and confirm our previous finding that export sophistication
remains statistically significant and robust with a positive and relatively large effect on growth.
However, the fixed effects IV regression, in which the equation is differenced and the dependent
variable is the change in the growth rate, has the weak instrument problem as well. This suggests
that it is harder to predict endogenous variables that are growth rates rather than levels using
as instruments.
V. ROBUSTNESS
We experiment with alternative proxies of export sophistication manufacturing exports as a
regression, we find that both the share of manufacturing exports and real manufacturing exports
per capita have all significant and positive coefficients (Table 3, columns 2 and 4). Including
both EXPY and another measure of manufacturing exports results in quasi-multicollinearity and
insignificant coefficients (columns 3 and 5). The weak instrument tests show that the regressions
with both measures are plagued with the weak instrument problem.
Adding a control for manufacturing production, we find that export sophistication seems to be
more important than manufacturing production in affecting growth. Manufacturing value added
as a share of GDP is statistically significant in the regression with EXPY (columns 6-7) but real
manufacturing value added per capita (in logs) is not statistically significant (columns 8-9).
However, with other proxies for export sophistication, manufacturing value added as a share of
GDP is no longer statistically significant while the coefficient of log real manufacturing value
added per capita is negative, which seems to pick up the effect of the initial income variable
8 Since the reported test statistics are based on 2-3 endogenous variables from Stock and Yogo (2005), and we have
a total of 6 endogenous variables, the thresholds used are relatively conservative.
©International Monetary Fund. Not for Redistribution
13
(columns 10-11). Export sophistication proxies have positive and statistically significant
estimates in all regressions. This suggests that export orientation is important in the growth
process and that producing manufacturing, and potentially sophisticated, goods without
exporting them may not be sufficient to increase long-run growth.
Table 3. Other Measures of Export Sophistication: IV Estimation (5-Year Panel)
We control for the average logarithm of real GDP (or real GDP per capita) in the neighboring
countries to capture directly spillover effects. Doing so should also mitigate a potential violation
e average GDP or GDP per capita in
initial income.9
statistically significant at the 10 percent level (Table 4, column 1). Excluding schooling, it
becomes statistically insignificant (column 2). The coefficient on EXPY is statistically
significant and similar to other estimates. In the baseline regression (Table 4, column 2), the
coefficient of the spillover effect as measure by average real GDP of neighbors is positive and
strongly significant while EXPY is no longer significant (column 3). However, when we exclude
schooling, which is not significant in most of our regressions (see Tables 2-10), including
column 3 regression when EXPY is not included, we obtain a positive and statistically
significant coefficient on EXPY (column 4).
9 We exclude the duplicate countries and the country in question for which the instrument is calculated when
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
Dependent var.: 5-year ave. annual growth rate
Log export sophistication 0.065*** 0.053 0.209 0.063*** 0.083***
(0.013) (0.051) (0.830) (0.021) (0.020)
Log real GDP per capita -0.024*** -0.012*** -0.026* -0.027*** -0.025*** -0.003 -0.023*** -0.002 -0.029** -0.013*** -0.005
(0.004) (0.003) (0.014) (0.005) (0.009) (0.004) (0.008) (0.009) (0.013) (0.003) (0.009)
Years of schooling 0.001 0.004*** 0.003** 0.001 0.008 -0.003 -0.003 0.003 0.002 0.003* 0.004*
(0.001) (0.001) (0.001) (0.002) (0.028) (0.003) (0.003) (0.002) (0.003) (0.002)
(0.003)
Manufactures exports (% of merchandise exports) 0.000*** -0.000 0.000***
(0.000) (0.000)
(0.000)
Log real manufacturing exports per capita 0.012*** -0.034 0.012***
(0.003) (0.180)
(0.004)
Manufacturing value added (% of GDP) 0.004*** 0.003** 0.001
(0.002) (0.001)
(0.001)
Log real manufacturing value added per capita -0.004 -0.003 -0.022**
(0.008) (0.009) (0.011)
Observations 1,216 947 947 946 946 828 799 799 770 671 651
\# of endogenous variables 3 3 4 3 4 3 4 3 4 4 4
\# of instruments 13 13 14 13 14 13 14 13 14 14 14
\# of excluded instruments 3 3 4 3 4 3 4 3 4 4 4
Cragg-Donald F stat 40.7 109.7 2.6 13.0 0.0 6.0 5.1 13.1 6.6 5.9 5.5
Kleibergen-Paap F stat 38.4 100.0 1.7 11.6 0.0 4.9 4.0 15.1 5.7 4.0 5.9
Kleibergen-Paap LM test p-value 0.000 0.000 0.013 0.000 0.800 0.000 0.000 0.000 0.000 0.000 0.000
H_0: t-test size>10% (p-value) | KP 0.000 0.000 1.000 0.117 1.000 0.822 0.996 0.021 0.968 0.996 0.961
H_0: t-test size>25% (p-value) | KP 0.000 0.000 0.822 0.000 1.000 0.077 0.286 0.000 0.080 0.289 0.067
H_0: t-test size>10% (p-value) | CD 0.000 0.000 1.000 0.062 1.000 0.689 0.982 0.060 0.931 0.962 0.974
H_0: t-test size>25% (p-value) | CD 0.000 0.000 0.607 0.000 1.000 0.031 0.125 0.000 0.038 0.069 0.093
H_0: t-test rel-bias>10% (p-value) | KP 0.000 0.000 0.960 0.002 1.000 0.247 0.608 0.000 0.288 0.611 0.256
H_0: t-test rel-bias>30% (p-value) | KP 0.000 0.000 0.634 0.000 1.000 0.054 0.122 0.000 0.022 0.123 0.017
H_0: t-test rel-bias>10% (p-value) | CD 0.000 0.000 0.865 0.001 1.000 0.132 0.381 0.001 0.176 0.262 0.317
H_0: t-test rel-bias>30% (p-value) | CD 0.000 0.000 0.374 0.000 1.000 0.021 0.040 0.000 0.008 0.018 0.027
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
©International Monetary Fund. Not for Redistribution
14
The coefficient for the spillover effect as measured by real GDP per capita of neighbors is
statistically insignificant using median or weighted averages of neighbors for the instruments
(columns 5-6 and 9-10). However, the estimates when using real GDP of neighbors as the
spillover effect are similar irrespective of the weighting schemes for the instruments used
(columns 7-8 and 11-12). The tests show that instruments are not weak when using real GDP as a
measure of the spillover effect. Overall, we find that the spillover effect, even if present, does not
invalidate our initial finding that export sophistication is a key growth determinant.10
Table 4. The Spillover Effect: IV Estimation (5-Year Panel)
Further, we explore other potential explanatory variables in the baseline regression and examine
the robustness of our results (Table 5). We add such variables as investment to GDP ratio, the
national saving rate, FDI, government consumption, the Gini coefficient, and corruption
(columns 3-9). The coefficient on export sophistication varies in the range of 0.05-0.08 and is
statistically significant in line with the previous results. However, some of these regressions
suffer from the weak instrument problem. We also include a measure of export sophistication
Ding and Hadzi-Vaskov (2017) that results in a robust and positive estimate
which is even larger than in previous regressions (column 2).11 Another measure of EXPY we
use structural EXPY, or S-EXPY which corrects for the share of commodities in exports also
produces a strong and positive coefficient although it is two to three times smaller in magnitude
(column 1).
10 The results with other proxies for export sophistication are broadly the same. We also control for the growth rates
of neighbors in the previous 5-year period, but obtain statistically insignificant results and weak instruments.
Regression results are not included in Table 4 and are available upon request.
11 The same study also computes a standardized EXPY but it is highly correlated with the original EXPY and
produces similar results (with a coefficient closer to our estimates).
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Dependent var.: 5-year ave. annual growth rate
Log export sophistication 0.084*** 0.065*** 0.002 0.043** 0.109*** 0.066** 0.010 0.057** 0.045*** 0.066*** -0.001 0.033*
(0.019) (0.015) (0.016) (0.020) (0.032) (0.027) (0.016) (0.024) (0.012) (0.015) (0.016) (0.019)
Log real GDP per capita -0.057*** -0.005 -0.016*** -0.018*** -0.080* 0.007 -0.018*** -0.023*** -0.023 -0.029* -0.019*** -0.017***
(0.018) (0.029) (0.004) (0.004) (0.041) (0.047) (0.004) (0.006) (0.015) (0.015) (0.004) (0.004)
Years of schooling 0.001 0.002* 0.001 0.001 0.003** 0.005***
(0.001) (0.001) (0.002) (0.001) (0.001) (0.001)
Log real GDP per capita of neighbors 0.025* -0.015 0.038 -0.027 0.001 0.005
(0.014) (0.025) (0.030) (0.036) (0.010) (0.009)
Log real GDP of neighbors 0.012*** 0.005 0.013*** 0.005 0.011*** 0.007**
(0.002) (0.003) (0.002) (0.003) (0.002) (0.003)
Observations 1,216 1,489 1,216 1,489 1,216 1,489 1,216 1,489 1,216 1,489 1,216 1,489
\# of endogenous variables 4 3 4 3 4 3 4 3 4 3 4 3
\# of instruments 14 13 14 13 14 13 14 13 14 13 14 13
\# of excluded instruments 4 3 4 3 4 3 4 3 4 3 4 3
Cragg-Donald F stat 4.1 3.2 18.7 17.8 1.0 1.1 18.8 15.6 2.8 7.8 20.3 22.6
Kleibergen-Paap F stat 3.8 2.4 16.7 15.3 0.9 0.8 17.1 12.9 2.6 7.0 17.2 18.9
Kleibergen-Paap LM test p-value 0.00 0.01 0.00 0.00 0.06 0.11 0.00 0.00 0.00 0.00 0.00 0.00
H_0: t-test size>10% (p-value) | KP 1.00 0.98 0.05 0.02 1.00 1.00 0.04 0.06 1.00 0.56 0.04 0.00
H_0: t-test size>25% (p-value) | KP 0.32 0.42 0.00 0.00 0.96 0.85 0.00 0.00 0.59 0.01 0.00 0.00
H_0: t-test size>10% (p-value) | CD 1.00 0.96 0.02 0.00 1.00 1.00 0.02 0.02 1.00 0.46 0.01 0.00
H_0: t-test size>25% (p-value) | CD 0.27 0.27 0.00 0.00 0.94 0.78 0.00 0.00 0.55 0.01 0.00 0.00
H_0: t-test rel-bias>10% (p-value) | KP 0.64 0.70 0.00 0.00 0.99 0.96 0.00 0.00 0.85 0.07 0.00 0.00
H_0: t-test rel-bias>30% (p-value) | KP 0.14 0.36 0.00 0.00 0.88 0.81 0.00 0.00 0.36 0.01 0.00 0.00
H_0: t-test rel-bias>10% (p-value) | CD 0.59 0.54 0.00 0.00 0.99 0.93 0.00 0.00 0.83 0.04 0.00 0.00
H_0: t-test rel-bias>30% (p-value) | CD 0.11 0.21 0.00 0.00 0.84 0.73 0.00 0.00 0.32 0.00 0.00 0.00
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Instruments: Average Instruments: Median Instruments: Weighted Average
©International Monetary Fund. Not for Redistribution
15
Table 5. Robustness: IV Estimation (5-Year Panel)
We also use the GMM methodology using our sharp instrument as an additional instrument in
the regression. By doing so, we are revisiting the main specification of Hausmann, Hwang, and
Rodrik (2007). We present our results in Table 6 (constant and time dummies are not shown).
The standard GMM instruments are two lags of the explanatory variables (levels in the
difference equation, DIF) and one lagged difference of the variables in the level equation (LEV).
We add the neighbor of the same explanatory variables assumed exogenous (that is,
IV instruments) in the GMM setting (implying the difference of the variables in the difference
equation and the levels of variables in the level equation). We also experiment with using sharp
instruments as GMM instruments and excluding standard GMM instruments from the first stage.
Export sophistication has a highly significant and positive coefficient in all specifications.
Moreover, the magnitude of the coefficient remains relatively stable around 0.05-0.1 and is
similar to that found in IV regressions.
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Dependent var.: 5-year ave. annual growth rate
Log export sophistication 0.061** 0.067*** 0.078*** 0.053*** 0.098*** 0.079*** 0.055***
(0.024) (0.018) (0.020) (0.012) (0.022) (0.020) (0.017)
Log real GDP per capita -0.010*** -0.024*** -0.026*** -0.027*** -0.027*** -0.021*** -0.042*** -0.020*** -0.029***
(0.002) (0.006) (0.005) (0.004) (0.006) (0.005) (0.008) (0.005) (0.008)
Years of schooling 0.000 0.008*** 0.002 0.001 0.002 0.002 0.004** 0.001 0.001
(0.001) (0.002) (0.002) (0.002) (0.002) (0.001) (0.001) (0.002) (0.002)
Log structural export sophistication, S-EXPY 0.022***
(0.004)
Log export sophistication (Ding and Hadzi-Vaskov) 0.443***
(0.147)
Gross fixed capital formation (% of GDP) 0.000
(0.001)
National saving rate (% of GDP) 0.000
(0.000)
Foreign direct investment (% of GDP) -0.487*
(0.291)
Government consumption (% of GDP) -0.001
(0.001)
Gini coefficient (net) 0.001**
(0.000)
Corruption -0.010***
(0.003)
Ethnic fractionalization 0.000*
(0.000)
Latitude 0.000
(0.000)
Sub-Saharan Africa dummy -0.036***
(0.007)
Observations 1,216 1,200 1,004 1,047 936 1,052 716 675 980
\# of endogenous variables 3 3 4 4 4 4 4 4 5
\# of instruments 13 13 14 14 13 14 14 10 16
\# of excluded instruments 3 3 4 4 4 4 4 4 5
Cragg-Donald F stat 88.0 23.2 3.2 11.0 0.2 22.8 11.3 16.5 15.2
Kleibergen-Paap F stat 88.4 17.2 2.6 5.1 0.8 19.4 6.9 16.5 13.5
Kleibergen-Paap LM test p-value 0.00 0.00 0.00 0.00 0.06 0.00 0.00 0.00 0.00
H_0: t-test size>10% (p-value) | KP 0.00 0.01 1.00 0.98 1.00 0.01 0.92 0.06 0.51
H_0: t-test size>25% (p-value) | KP 0.00 0.00 0.61 0.13 0.97 0.00 0.03 0.00 0.00
H_0: t-test size>10% (p-value) | CD 0.00 0.00 1.00 0.49 1.00 0.00 0.45 0.06 0.31
H_0: t-test size>25% (p-value) | CD 0.00 0.00 0.45 0.00 1.00 0.00 0.00 0.00 0.00
H_0: t-test rel-bias>10% (p-value) | KP 0.00 0.00 0.87 0.39 1.00 0.00 0.15 0.00 0.00
H_0: t-test rel-bias>30% (p-value) | KP 0.00 0.00 0.38 0.04 0.91 0.00 0.01 0.00 0.00
H_0: t-test rel-bias>10% (p-value) | CD 0.00 0.00 0.76 0.01 1.00 0.00 0.01 0.00 0.00
H_0: t-test rel-bias>30% (p-value) | CD 0.00 0.00 0.24 0.00 0.99 0.00 0.00 0.00 0.00
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
©International Monetary Fund. Not for Redistribution
16
Table 6. GMM-System Estimation (5-Year Panel)
Comparing the GMM estimation of the difference equation and the level equation, we find
favorable test statistics using the level equation specification. The GMM-LEV estimator that
uses smaller or collapsed number of instruments (Table 7, column 7) than the usual GMM
estimator (Table 7, column 6) does not suffer from the weak instrument problem and produces a
positive and statistically significant coefficient on EXPY, similar to the IV estimates. Using own
explanatory variables and sharp instruments as GMM instruments results in the weak instrument
problem (columns 8-9). However, it seems strong identification comes from using sharp
instruments as IV instruments (column 10). The J-test (column 7) suggests that overidentifying
restrictions are not valid, but using Hahn, Ham, and Moon (2011) test for instrument validity, we
find that the instruments based on the averages are valid (p-value of about 0.7). In the
difference equation specification (columns 1-5), the tests for weak instruments suggest that the
instruments are weak. The coefficient on the initial log real GDP per capita is statistically
significant and has the usual negative sign. However, the coefficient on schooling becomes
negative and is statistically significant in a few estimations (columns 1-3). The identification
seems to come from sharp instruments used as IV instruments (column 5) in which the
coefficient on EXPY is statistically significant at 10 percent. However, the weak instrument
problem is present in this specification as well. These results indicate that it is the cross-country
variation stemming from the level equation estimation rather than the time series variation in the
difference equation estimation that produces parameter identification and favorable test
statistics.12
12 If we assume fixed effects are correlat
not produce robust estimates of the parameters and the differenced variables result in the weak instrument problem.
Table 2, column 9, and Table 4, column 12, show estimations considering fixed effects. The coefficient on export
sophistication is similar to that in other estimations.
(1) (2) (3) (4) (5)
Dependent var.: 5-year ave. annual growth rate
Own variable 2-
4
period lags,
GMM
instr
Neighbor
averages, IV
instr
Neighbor
averages and 1-
period lags, IV
instr
Neighbor
averages and 1-
2 period lags,
GMM instr
Neighbor
averages and 2-
4 period lags,
GMM instr
Log export sophistication 0.052*** 0.059*** 0.084*** 0.111*** 0.098**
(0.018) (0.022) (0.026) (0.033) (0.042)
Log real GDP per capita -0.027** -0.022*** -0.029*** -0.030*** -0.017**
(0.011) (0.006) (0.007) (0.011) (0.008)
Years of schooling 0.008*** 0.001 0.000 -0.003 -0.005
(0.003) (0.002) (0.002) (0.004) (0.005)
\# of observations 1226 1216 1119 1226 1226
\# of countries 137 136 136 137 137
\# of instruments 22 13 15 19 22
\# of overidentifying restrictions 9 0 1 6 9
Hansen J-test p-value 0.146 . 0.005 0.025 0.012
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1 ©International Monetary Fund. Not for Redistribution 17
Table 7. Dissecting GMM (5-Year Panel)
averages are used, and
the 10-year panel regressions, we find that export sophistication is statistically significant in most
estimations (Table 8). The coefficients are similar in magnitude and the tests for weak
instruments suggest that in several estimations, the instruments are as strong as in the 5-year
panel case.
erform the same estimations IV using 5-
year panels with the instrument set that uses the median of and the weighted mean (with
weights equal to the . The results are broadly
unchanged compared to the instrument set based on simple averages (Tables 9-10). This shows
that our results, tests and coefficients, are mostly robust to changing the aggregation method for
the instrument calculation. Using the weights of the inverse of real GDP for our instruments
interest. Even though the weak instrument tests are not favorable in a few estimations, the overall
conclusion remains the same (Table 10). Export sophistication is still a key determinant in
growth regressions.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
DIF DIF-Collapse DIF-Collapse DIF-Collapse DIF-Collapse LEV LEV-Collapse LEV-Collapse LEV-Collapse LEV-Collapse
Dependent var.: 5-year ave. annual growth rate
Own vars,
GMM instr
+ Neighbor
averages,
IV instr
Own vars,
GMM instr +
Neighbor
averages, IV
instr
Own vars,
GMM instr
Neighbor
averages 1-2
period lags,
GMM instr
Neighbor
averages 1-2
period lags, IV
instr
Own vars,
GMM instr
+ Neighbor
averages,
IV instr
Own vars,
GMM instr +
Neighbor
averages, IV
instr
Own vars,
GMM instr
Neighbor
averages,
GMM instr
Neighbor
averages, IV
instr
Log export sophistication 0.036*** 0.038*** 0.068 0.228*** 0.065***
(0.011) (0.011) (0.068) (0.077) (0.013)
Log real GDP per capita -0.018*** -0.018*** -0.126 -0.052*** -0.024***
(0.004) (0.004) (0.121) (0.018) (0.004)
Years of schooling 0.003*** 0.003*** 0.044 -0.018 0.001
(0.001) (0.001) (0.039) (0.014) (0.001)
Differenced log export sophistication 0.021 0.057* 0.039 -0.067 0.173*
(0.019) (0.030) (0.057) (0.103) (0.093)
Differenced log real GDP per capita -0.142*** -0.110*** -0.193*** -0.060* -0.126**
(0.025) (0.033) (0.048) (0.034) (0.052)
Differenced years of schooling -0.037*** -0.056** -0.111*** -0.055 0.060
(0.013) (0.029) (0.034) (0.043) (0.060)
Observations 1,080 1,080 1,089 944 944 1,216 1,216 1,226 1,080 1,216
\# of endogenous variables 3 3 3 3 3 3 3 3 3 3
\# of instruments 75 21 18 14 14 37 16 13 12 13
\# of excluded instruments 66 12 9 6 6 27 6 3 3 3
Cragg-Donald F stat 1.9 1.8 1.5 1.2 0.9 9.7 39.0 0.5 2.3 40.7
Kleibergen-Paap F stat 2.5 1.7 1.6 1.5 0.9 10.4 36.8 0.3 2.5 38.4
Kleibergen-Paap LM test p-value 0.01 0.03 0.06 0.09 0.23 0.00 0.00 0.32 0.01 0.00
C-stat (p-value) 0.00 0.03 . . . 0.00 0.00 . . .
Hansen J-test p-value 0.00 0.03 0.47 0.31 0.52 0.00 0.00 . . .
H_0: t-test size>10% (p-value) | KP 1.00 1.00 1.00 1.00 1.00 1.00 0.00 1.00 0.98 0.00
H_0: t-test size>25% (p-value) | KP 1.00 1.00 1.00 0.99 1.00 1.00 0.00 0.97 0.40 0.00
H_0: t-test size>10% (p-value) | CD 1.00 1.00 1.00 1.00 1.00 1.00 0.00 1.00 0.99 0.00
H_0: t-test size>25% (p-value) | CD 1.00 1.00 1.00 1.00 1.00 1.00 0.00 0.94 0.46 0.00
H_0: t-test rel-bias>10% (p-value) | KP 1.00 1.00 1.00 1.00 1.00 0.10 0.00 0.99 0.68 0.00
H_0: t-test rel-bias>30% (p-value) | KP 0.97 0.93 0.89 0.82 0.96 0.00 0.00 0.95 0.33 0.00
H_0: t-test rel-bias>10% (p-value) | CD 1.00 1.00 1.00 1.00 1.00 0.25 0.00 0.99 0.73 0.00
H_0: t-test rel-bias>30% (p-value) | CD 1.00 0.91 0.93 0.91 0.95 0.00 0.00 0.92 0.39 0.00
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
©International Monetary Fund. Not for Redistribution
18
Table 8. IV Estimation (10-Year Panel)
Table 9. IV Estimation, Instrument Set: Median (5-Year Panel)
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Dependent var.: 10-year ave. annual growth rate IV IV IV IV IV IV IV IV IV-FE
Log export sophistication 0.085*** 0.074*** 0.061*** 0.091*** 0.081*** 0.142** 0.225 0.020
(0.013) (0.013) (0.016) (0.024) (0.020) (0.062) (0.180) (0.052)
Log real GDP per capita -0.030*** -0.026*** -0.027*** -0.030*** -0.024*** -0.032*** -0.007 -0.035 -0.062**
(0.005) (0.005) (0.005) (0.009) (0.005) (0.012) (0.005) (0.022) (0.026)
Years of schooling 0.002 0.003 -0.009 -0.000
(0.001) (0.002) (0.014) (0.034)
Trade (% of GDP) -0.042** -0.015 -0.096
(0.021) (0.019) (0.112)
Credit to private sector (% of GDP) -0.033 0.017 0.014
(0.025) (0.018) (0.044)
Law and order -0.021 -0.002 -0.026
(0.013) (0.004) (0.021)
Observations 724 724 596 627 618 365 306 301 363
\# of endogenous variables 1 2 3 3 3 3 5 6 3
\# of instruments 7 7 8 8 8 6 8 9 9
\# of excluded instruments 1 2 3 3 3 3 5 6 6
Cragg-Donald F stat 147.6 53.3 20.8 6.3 19.3 2.5 1.5 0.2 0.4
Kleibergen-Paap F stat 103.7 38.9 17.9 4.8 11.3 2.2 0.9 0.2 0.4
Kleibergen-Paap LM test p-value 0.00 0.00 0.00 0.00 0.00 0.01 0.03 0.28 0.66
H_0: t-test size>10% (p-value) | KP 0.00 0.00 0.00 0.83 0.13 0.99 1.00 1.00 1.00
H_0: t-test size>25% (p-value) | KP 0.00 0.00 0.00 0.08 0.00 0.48 0.99 1.00 1.00
H_0: t-test size>10% (p-value) | CD 0.00 0.00 0.00 0.66 0.00 0.98 1.00 1.00 1.00
H_0: t-test size>25% (p-value) | CD 0.00 0.00 0.00 0.03 0.00 0.40 0.95 1.00 1.00
H_0: t-test rel-bias>10% (p-value) | KP 0.00 0.00 0.00 0.26 0.00 0.75 1.00 1.00 1.00
H_0: t-test rel-bias>30% (p-value) | KP 0.00 0.00 0.00 0.06 0.00 0.41 0.93 1.00 1.00
H_0: t-test rel-bias>10% (p-value) | CD 0.00 0.00 0.00 0.11 0.00 0.68 0.99 1.00 1.00
H_0: t-test rel-bias>30% (p-value) | CD 0.00 0.00 0.00 0.02 0.00 0.33 0.76 1.00 1.00
Hansen J-test p-value . . . . . . . . 0.11
Lower CLR bound 0.06
Upper CLR bound 0.12
H0: Beta_EXPY=0 | CLR p-value 0.00
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Dependent var.: 5-year ave. annual growth rate IV IV IV IV IV IV IV IV IV-FE
Log export sophistication 0.076*** 0.078*** 0.076*** 0.082*** 0.082*** 0.122*** 0.116*** 0.108
(0.010) (0.011) (0.013) (0.014) (0.013) (0.032) (0.030) (0.067)
Log real GDP per capita -0.025*** -0.026*** -0.028*** -0.029*** -0.025*** -0.025*** -0.005 -0.025*** -0.109**
(0.004) (0.004) (0.004) (0.005) (0.004) (0.007) (0.005) (0.006) (0.043)
Years of schooling 0.001 0.003 -0.001 -0.007
(0.001) (0.002) (0.003) (0.037)
Trade (% of GDP) -0.003 -0.016 -0.022
(0.011) (0.012) (0.016)
Credit to private sector (% of GDP) -0.021* 0.004 -0.006
(0.013) (0.014) (0.016)
Law and order -0.018** 0.001 -0.009*
(0.007) (0.003) (0.005)
Observations 1,590 1,590 1,216 1,369 1,319 748 606 598 983
\# of endogenous variables 1 2 3 3 3 3 5 6 3
\# of instruments 13 13 13 14 14 9 11 12 14
\# of excluded instruments 1 2 3 3 3 3 5 6 6
Cragg-Donald F stat 313.1 100.7 43.3 17.7 40.1 8.3 4.4 2.3 1.4
Kleibergen-Paap F stat 232.7 71.0 39.3 14.5 23.9 6.9 2.4 2.3 1.3
Kleibergen-Paap LM test p-value 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.09
H_0: t-test size>10% (p-value) | KP 0.00 0.00 0.00 0.03 0.00 0.57 1.00 1.00 1.00
H_0: t-test size>25% (p-value) | KP 0.00 0.00 0.00 0.00 0.00 0.01 0.82 0.92 0.99
H_0: t-test size>10% (p-value) | CD 0.00 0.00 0.00 0.00 0.00 0.39 1.00 1.00 1.00
H_0: t-test size>25% (p-value) | CD 0.00 0.00 0.00 0.00 0.00 0.00 0.34 0.93 0.99
H_0: t-test rel-bias>10% (p-value) | KP 0.00 0.00 0.00 0.00 0.00 0.08 0.96 0.99 1.00
H_0: t-test rel-bias>30% (p-value) | KP 0.00 0.00 0.00 0.00 0.00 0.01 0.48 0.53 0.87
H_0: t-test rel-bias>10% (p-value) | CD 0.00 0.00 0.00 0.00 0.00 0.03 0.69 0.99 1.00
H_0: t-test rel-bias>30% (p-value) | CD 0.00 0.00 0.00 0.00 0.00 0.00 0.08 0.55 0.85
Hansen J-test p-value . . . . . . . . 0.10
Lower CLR bound 0.06
Upper CLR bound 0.09
H0: Beta_EXPY=0 | CLR p-value 0.00
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
©International Monetary Fund. Not for Redistribution
19
Table 10. IV Estimation, Instrument Set: Weighted Mean (5-Year Panel)
VI. CONCLUSION
This paper explores the determinants of growth based on an instrumental variable technique that
factor of growth is instrumented by its average in the neighboring countries. We show that export
sophistication whether proxied by EXPY, the share of manufacturing exports
exports, or real manufacturing exports per capita stands out as an important and robust
determinant of growth. This is further confirmed by verifying the strength of the
Bazzi and Clemens (2013).
Although standard growth determinants are not robust in the regressions, they may be important
to the extent they help improve export sophistication.
The technique we propose could be applied to other empirical studies suffering from the blunt
instrument problem. It offers a variable-specific, dynamic and plausibly valid instrument for as
many variables as needed. The striking result in our study is that overall, the instruments passed
the instrument strength tests. Correlations among neighborin
mimetic forces could be at play, where economic agents learn from across the borders in formal
and informal ways. It suggests that competition with immediate neighbors is a potent factor in
the diffusion of technologies and policies. Perhaps a , far away from advanced
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Dependent var.: 5-year ave. annual growth rate IV IV IV IV IV IV IV IV IV-FE
Log export sophistication 0.062*** 0.055*** 0.031** 0.060*** 0.072*** 0.077** 0.269 0.157**
(0.010) (0.011) (0.013) (0.020) (0.016) (0.036) (0.444) (0.072)
Log real GDP per capita -0.021*** -0.018*** -0.015*** -0.015** -0.020*** -0.013** 0.004 -0.009 -0.135**
(0.004) (0.004) (0.004) (0.007) (0.005) (0.006) (0.011) (0.026) (0.059)
Years of schooling 0.003** 0.000 -0.024 0.082
(0.001) (0.004) (0.052) (0.085)
Trade (% of GDP) -0.059 -0.047 -0.215
(0.046) (0.050) (0.443)
Credit to private sector (% of GDP) -0.027** -0.008 0.111
(0.014) (0.025) (0.246)
Law and order -0.013* 0.004 -0.033
(0.008) (0.006) (0.061)
Observations 1,590 1,590 1,216 1,369 1,319 748 606 598 983
\# of endogenous variables 1 2 3 3 3 3 5 6 3
\# of instruments 13 13 13 14 14 9 11 12 14
\# of excluded instruments 1 2 3 3 3 3 5 6 6
Cragg-Donald F stat 260.4 95.7 38.0 1.6 29.6 4.4 0.6 0.0 1.0
Kleibergen-Paap F stat 192.8 72.8 33.9 1.5 20.5 3.3 0.4 0.0 1.1
Kleibergen-Paap LM test p-value 0.00 0.00 0.00 0.03 0.00 0.00 0.18 0.60 0.13
H_0: t-test size>10% (p-value) | KP 0.00 0.00 0.00 1.00 0.00 0.95 1.00 1.00 1.00
H_0: t-test size>25% (p-value) | KP 0.00 0.00 0.00 0.67 0.00 0.25 1.00 1.00 1.00
H_0: t-test size>10% (p-value) | CD 0.00 0.00 0.00 1.00 0.00 0.87 1.00 1.00 1.00
H_0: t-test size>25% (p-value) | CD 0.00 0.00 0.00 0.62 0.00 0.12 1.00 1.00 1.00
H_0: t-test rel-bias>10% (p-value) | KP 0.00 0.00 0.00 0.88 0.00 0.52 1.00 1.00 1.00
H_0: t-test rel-bias>30% (p-value) | KP 0.00 0.00 0.00 0.61 0.00 0.20 0.99 1.00 0.92
H_0: t-test rel-bias>10% (p-value) | CD 0.00 0.00 0.00 0.85 0.00 0.32 1.00 1.00 1.00
H_0: t-test rel-bias>30% (p-value) | CD 0.00 0.00 0.00 0.56 0.00 0.08 0.98 1.00 0.94
Hansen J-test p-value . . . . . . . . 0.02
Lower CLR bound 0.04
Upper CLR bound 0.08
H0: Beta_EXPY=0 | CLR p-value 0.00
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
©International Monetary Fund. Not for Redistribution
A HISTORICAL PATTERN OF ECONOMIC GROWTH
IN DEVELOPlNG COUNTRIES
KANAME AKAMATSU
I. THE HISTORICAL STAGES OF ECONOMIC GROWTH
IN DEVELOPING COUNTRIES
T is impossible to study the economic growth of the developing coun-I
tries in modern times without considering the mutual interactions
between these economies and those of the advanced countries. When
Western European Capitalism began to expand its production and trade
on a world-wide scale, it awakened the less-developed areas of the
world to modern economic development. This article will take up the
Asian afea, including Japan, as object of examination. Economic growth
in the Asian area was brought about by the eastward advance of Western
European capitalism. In this intermingling of Western European and
Asian economies the following historical stages can be observed.
The first stage is the period when native Asian industry developed
as a result of the exchange of native Asian products for Western European
industrial products.
The second stage is the period when the native handicraft industry
crumbled because manufactured consumer goods flowed into the Asian
area after the Industrial Revolution in Western Europe.
The third stage is the period when Western European capital and
techniques infiltrated the Asian area for the large-scale production of
primary goods, such as raw materials and provisions necessary for the
Western European economy, as well as for the construction of railroads
and highways. During this period the exchange of Western European
consumer goods for native primary produ.cts came to be established.
The fourth stage is the period when Western European capital. came
into the developing countries to develop modern industries, including the
industries processing raw materials produced ip those areas.
The fifth stage is the period when native capital began to run the
industries processing native raw materials. In this period a conflicting
relationship was generated between consumer gdods imported from the
advanced countries and those of the native processing industries. However,
in this period, capital goods came to be imported from the advanced
countries for the consumer-goods indus.tries in the developing countries and,
r~ ~
4 THE DEVELOPING ECONOMIES
in consequence, there was a conspicuous change from consumer goods to
capital goods in the import structure.
The sixth stage is the period when manufactured goods in general
began to be produced by native industries, whether the raw materials
were domestically available or not. The capital goods required by these
industries were imported at the expense of the induction of foreign capital_.
and of the export of primary products,
The seventh stage is the period when the industrialization of the
developing countries became so advanced as to make possible the export
of manufactured consumer goods, and when the domestic production of
some capital goods gradually came to the fore.
These stages, however, overlap each other and cannot be clearly
classified nor applied to every developing country in Asia. For instance,
J~pan has attained the position of an advanced country in comparison
with other Asian countries. She has attained a stage higher than the
seventh stage. India and China can be said to have partially reached
the seventh stage. The other countries, however, have not yet developed
to the stage of capital goods production.
1. Development in Connection with Heterogeneous Economic Interrelation.
For the classification of the above-mentioned stages with respect to
the interrelation between an advanced economy and a less-advanced
economy, we can call the stages from the frst to the third the period
of differentiation of economic structures into advanced and less-advanced.
Primarily, the Western European economy and the Asian economy have
heterogeneous characters deriving from different natural environments, ways
of life, and cultures. Out of this heterogeneous economic relationship,
as a Inatter of course, something of an international division (if labour is
formed. When heterogeneous specialities are produced in different en-
vironments, international relations are formed therein and trade is com-
menced. Thus the specialities of Asian countries came to be exchanged
for Western European products. However, at this stage, it is premature
to call rt an “mternatronal drvlslon of labour”. The international division
of labour comes into being only after the specialities trade, which has
been initiated by heterogeneous interrelationship, has stimulated the pro-
duction of specialities for export. This is none other than the hetero-
geneization of the international economies. At this juncture, Western
European capital set out to exploit farms and mines in the Asian area
and to develop the Asian economy as a complementary economy to that
of Western Europe.
Heterogerieous economies have a possibility of creating complementary
A HISTORICAL PATTERN OF ECONOMIC GROWTH 5
relationship provided that the products of the one area can become the
object of wants by the inhabitants of the other area. Western Europeans
have manufactured in the Asian area the products which they themselves
wanted and have turned Asia into a complementary area of Europe by
so developing them. In this way, modern development of the Asian
economy has been greatly accelerated by its heterogeneization with the
Western European economy.
This heterogeneization of the international economy, i,e. the interna-
tional division of labour, necessarily accompanies structural changes in the
Asian economy and partially created a situation to be called “structural
contradiction”.
One of these changes is that, as indicated in the above-mentioned
second stage, many handicraft industries which had existed among the
natives were destroyed by manufactured consumer goods imported in ex-
change for the export of native specialities. Of course, the import of
such items as matches and glasswares from Western Europe was instru-
mental in the elevation of living standard of the native population in
view of the fact that those articles had not been produced in the locality.
But, for instance, in various parts of Asia clothing had been produced
by the handicraft industry, which was destroyed as a result of the import
of cotton cloth from abroad, particularly from England. The handicrafts-
men were thrown out of work and ivere degraded to being farm labourers.
If the handicraftsmen were driven out of work and their standard of
living downgraded to that of labourers, this situation could be called a
“structural contradiction.”I
When Western European industrial products which have the same
l “Examples are easy to find of underdeveloped countries whose entire culture has
been impoverished as trading contacts with the outside world have developed. In
13a:gdad, for example, of the old handicrafts for which the city was famous th~,re
survive only a few silversmiths who themselves have adopted patterns from abroad
requiring less craftsmanship.” (G. Myrdal, Economic Theory and Underdeveloped
Regions. London : Gerald Duckworth & Co., 1957, p. 52)
In the early years of the Meiji Era in Japan, “Handicraftsmen were threatened by
the newly imported plant industries, and the invasion of cheap merchandise produced
by European and American capitalism wrought havoc upon them.” (Tetsuji Kada,
Melji Shoki Shakai Keizai Shis5shi ((History of Social Economic Thought in the
Early Years of the Meiji Era)), Tokyo : Iwanami Shoten, 1937, p. 535)
K. Marx quoted in his Discours sur la Question du Libre Echange (1848) the report
that after the Industrial Revolution in England, English woven cloth was imported
into India at a time when exports of Indian cotton goods came to a stop, resulting
in the starvation of Indian fabric handicraftsmen.
“The process of the decay of the Indian industries was rendered complete by the
competition of cheap machine-made goods from England.” (V. Sundara Rajan. An
Economic History of India, 1757-1947, Baroda : East & West Book House, 1955, p. 118).
6 THE DEVELOPING ECONOMIES
uses as those of native handicraft industry products are imported, homo-
geneization of the international economy through the contact of these
two products can be observed. Heterogeneity can be complementary
and co-accelerative, while homogeneity can be substitutive and competitive.
This substitutive and competitive relationship caused the native handicraft
industry to submit to the pressure of the factory industries of the advan-
ced countries. However, the structural contradiction caused by this homo-
geneization was the cost of creating an international division of labour
which constituted the essential tendency of the internatiqnal economy at
that time; and this contradiction was gradually dissolved with the growth
of modern industries in the less-advanced countries.
The second structural contradiction as a result of international economic
heterogeneization is that the advanced Western European countries came
to politically govern the Asian countries as their colonies. This contra-
diction is what the theory of imperialism explains. The problem here is
that the Western countries wanted to monopolize their colonies as com-
plementary areas of their home economies and to maintain everlasting
heterogeneous interrelations between their home economies and colonial
economies. This is realized in their policy of heterogeneizing colonial
economies.l
The heterogeneization policy was enforced, for instance, by England
on its American colonies. On the one hand they encouraged in America
those of their industries which could not have developed in England in order
to make America a complementary economic area, while on the other
hand they oppressed the growth, in the colonies, of the same kind of
industries as in England. In other words, they discouraged the homo-
geneization of the American colonial economy with the English home
economy. However, the industrialization of the colony and its homogenei-
zation with the home economy eventually became an essential trend of the
age until the heterogeneization policy of the home country could no longer
l The following testimonial of a person called Martin is quoted in Rajan’s An Economic
History of India, 1757-1947 : “Montgomery Martin, another witness, said : ‘We
have during the period of a quarter of a century compelled the Indian territories to
receive our manufactures, our woolens duty free, our cottons at 2lh per cent and
other articles in proportion ; while we have continued during that period to levy
almost prohibitory duties or duties varying from 10 to 20, 30, 50, 100, 500 and 1000
per cent upon articles, the produce from our territories . . .”‘ (ibid., p. 1 16)
“The policy of the government aimed at making India, ‘the agricultural farm of
England.”‘ (ibid., p. 1 19)
In the Dutch Indies, ” measures to replace goods imported from the Netherlands
by native J:)roducts were unlikely to meet with approval in the Second Chamber”.
(J. S. Furnivall. Netherland India, A Study of Plural Economy, New York : Macmillan
Co., 1944, p. 332)
A HISTORICAL PATTERN OF ECONOMIC GROWTH 7
check the homogeneization trend and resulted in the Declaration of In-
dependence of the American colonies. A similar process can be seen in
the case of India.
2. Development in Connection with Homogeneous Economic Interrelation
The Asian economy made remarkable progress in the course of its
heterogeneization with the Western European economy, but it soon switched
over to development toward homogeneization. Although this is partly
due to the fact that the less-advanced Asian area could not pose as a
primary products producer forever, the chief impetus came from Western
European capitalism.
Western European capital first went into the exploitation of farnis
and mineral resources to produce primary products on a large scale, thus
effectuating heterogeneization with the economy of the home country, but
it eventually turned to establish in the colonies processing industries similar
to those in the home country. Here started the above-mentioned fourth
stage.l
England had inhibited the rise of industries in the American colonies
which were the same as those in the home country, but with the indepen-
dence of America and the French Revolution, Western Europe entered
upon the age of Liberalism. Accordingly, it became impossible to restrain
Western European capitalism from transplanting industries to the colonies
in pursuit of larger profits. It was natural that in those lands, in
particular India and China, which had all the three beneficial conditions,
i.e. Iow wages peculiar to Asia, cheap raw materials obtainable from
native resources, and a selling market with a big population, Western
capitalism came to build modern industries. HoWever, industries and
trade thus built up by the foreign capital could not help giving birth to
modern industries by national capital, ‘as mentioned in the fifth stage.
Thus, national capital set out to build modern industries to cope with
the imported products as well as with the industrial goods manufactured
by imported capital. The degree of industr.ialization by means of national
capital varies according to country. In Japan, modernization of industries
was carried out rapidly almost exclusively through her national capital, and
this is partly because she had been almost free from Western European coloni-
zation. Also, India and China saw remarkable progress in the development
of their modern industries through national capital. But, in other colonial
countries no noticeable advance of national capital was marked until after
l “The factory industry in India has been founded and built up by British businessmen.”
(Rajan, op. cit., p. 127)
r~
8 THE DEVELOPlNG ECONOMIES
World War II. The rise of modern industry on the basis of national
capital in a less-developed area means homogeneization with Western
European industry. Here arises, first of all, a conflicting relationship between
imported consumer goods and native-produced consumer goods, which gives
birth to economic nationalism in the less-developed countries. This
economic nationalism movement first takes shape in the raising of import
tariffs on imported consumer goods or in the direct limitation of imports.
If the protective policy is effective and imports are checked while produc-
tion by national capital increases, the native industry might be said to
have attained the take-off stage.1 However, the policy of protecting
domestic industry by means of an import check should be adopted only
when ample development of the protected industry can be foreseen. Should
this development fail to occur, nationalism in the less-developed countries
may, on the contrary, impoverish the national economy.
The economic nationalism of the developing countries at first esta-
blishes consumer goods production on the basis of national capital, and
then it proceeds to the national capitalization of industries so far operated
by foreign capital and, further, to the production of capital goods by
national capital. This series of developments signifies that the developing
countries advance through the stages of homogeneization with the industries
of the advanced countries.
3. Development Stages of High-Degree Heterogeneization and High-Degree
Homogeneization
The establishment of consumer goods production with national capital,
within its limitation, means homogeneization with the advanced country’s
economy and development through mutual conflict. However, national
capital, in most cases, has to import capital goods in order to produce
consumer goods by itself. Here arises a heterogeneous complementary
relationship of capital goods importation and consumer goods production.
In this case, the premise is that the export industry of the advanced
countries has naturally developed from the production of consumer goods
to that of capital goods. Consequently, as the consumer goods industries
* In India, “tariff protection was thus an effective instrument for assisting industries
and s.timulating industrialization . . . The change in the content of her external trade
refiects the partial industrialization which took place between 1914 and 1939. Imports
bf consumer goods declined from 37 per cent in 1926-27 to 20 per cent in 1938-39 ;
while the imports of raw materials increased from 16 per cent in 1922-23 to 24 per
cent in 1938-39. The imports of machinery and other capital goods which formed
19 per cent of the total imports in 1926-27, constituted 25 per cent in 1938-39″.
(Rajan. op. cit., p. 125-26)
A HISTORICAL PATTERN OF ECONOMIC GROWTH 9
rise in the less-advanced countries, they form heterogeneous relations with
the capital goods industries of the advanced countries. This is called
high-degree heterogeneization, in which a high-degree conrplementary
relation is formed which differs from the frst stage heterogeneous relation
between the native specialities industry and the Western European con-
sumer goods industry. That is, the above-mentioned fifth stage sets in.
At this stage native nationalism levies lower tariffs on capital goods
imports in contradiction to the check policy on consumer goods imports,
with the import of the former being promoted. However, national capital
alone may not always be sufficient to import machinery and plants, nor
is the export of special products enough to cover the import of capital
goods. Either of these cases, or a situation in which both existed, could
cause difficulties in the international balance of payments. This makes
the induction of foreign capital necessary. Foreign capital had already
been induced in the above-mentioned third or fifth stage, mainly in the
stage of heterogeneization of the less-developed country’s economy to the
advanced country’s economy, which led to the appearance of imperialism
or colonialism by way of the capital exports of advanced countries.
In the course of imperialism, the capital exports of advanced coun-
tries are positive and active, while the capital import of developing coun-
tries are rather negative and passive. In the course of the homogeneization
of the less-advanced country’s economy with the advanced country’s econo-
my, the economic nationalism of the less-advanced countries becomes positive
in its own economic development and induces foreign capital voluntarily
for its own need. Economic nationalism tries to utilize fo.reign capital
for its homogeneization with the advanced country’s economy while exclu-
ding submission to colonialism. The capital export of advanced countries
or the capital import of less-advanced countries is, namely, the exchange
of capital goods of advanced countries for primary products and consumer
products of less-advanced countries through the international division of
labour. It is a heterogeneous relationship higher than the international
trade of previous stages between manufactured consumer goods and primary
products, which co.nstituted the stage of high-degree heterogeneization.
However, this stage also involves the possibility of developing into
the stage of high-degree homogeneization. Examples are the development
of the steel industry in India and China, where steel is already being
produced as a capital good, and moreover the production of secondary
goods with steel as a material is now beginning, the domestic production
of machinery and plants which ‘are necessary for the consumer goods
industries is also starting. If the first stage of the homogeneization pro-
cess is the process of the consumer goods industries of a less-advanced
10 THE DEVELOPlNG ECONOMIES
country catching up with those of an advanced country, this development
of capital goods production in a less-advanced country shold be called
the stage of high-degree homogeneization.
4. Homogeneization of Synthetic Materials and Natural Materials
What should be added lastly in the process of the heterogeneization
as well as the homogeneizatipn of an advanced couhtry’s economy and
a less-advanced country’s economy is the development of synthetic
material industries in advanced countries. This has led to the production
of artificial materials taking the place of natural materials in many fields
of industrial materials such as dyestuffs, rubber, textiles, fertilizers, fat,
and oil. There is a possibility that these fields will be enlarged. Here
arises the homogeneization of natural materials industry with synthetic
materials industry. While the primary industry depends largely upon
nature, the synthetic industty depends upon a high level of scientific tech-
nique. Although conspicuous heterogeneity exist~ between the two, they
are homogeneous in their uses, and present a conflicting and substitutive
relationship. This homogeneization process has been caused by the
elevation of scientific techniques in advanced countries. As against the
previous process, which occurred as a developmental process in less-
advanced countries, in this process new motive factors have been created
on the part of advanced countries. This means, in one view, that as the
old imperialism, which had penetrated into less-advanced countries seeking
for materials and provisions as primary products, came to be gradually
eliminated by nationalism, advanced countries began to produce artificial
materials by means of chemical syntheses and adopted re-agriculturization
policies in an effort to lessen their dependency upon less-advanced coun-
tries for those materials and provisions. From another angle, it might
be said that the advanced countries have attempted to homogeneize their
industries with the primary industries of less-advanced countries by synthetic
industries and re-agriculturization in order to stave off the pursuit by
less-advanced countries through their industrialization.
At any rate, the homogeneization of the synthetic industry and the
primary industry is a matter of concern to the less-advanced countries and
casts a dark shadow on the prospect of expansion of their exports of
primary products. The only way for less-advanced countries to overcome
this contradiction will be to weaken their vertical dependency upon the
advanced industrial countries by pushing forward their own industrializa-
tion. In the meantime, the elevation of the national income level of
less-advanced countries through industrialization will create a new horizontal
A HISTORICAL PATTERN OF ECONOMIC GROWTH 1 1
international division of labour irrelevant to the homogeneization of the
industrial structure.
II. THE WILD-GEESE-FLYING PATTERN (GANKO KEITAI) OF
INDUSTRIAL DEVELOPMENT IN DEVELOPlNG COUNTRIES
The development process of heterogeneization and homogeneization
of an advanced country’s economy and a less-advanced country’s economy,
as well as that of high-degree heterogeneization and homogeneization, can
be generally formulated into a historical theory called the “wild-geese-
flying pattern” of the industrial development of less-advanced countries.
Th~ wild-geese-flying pattern of industrial development denotes the develop-
ment after the less-advanced country’s economy enters into an international
economic relationship with the advanced countries. This theory leaves
out of consideration the period during which less-advanced countries are
in the stage of a closed self-sufflcient economy or during which there is
no international trade of any significance with a neighbouring country,
since their economic structures are homogeneous with each other. A sort
of formula for the industrial development of less-advanced countries after
they have Qpened trade ports and entered into large-scale trade relations
with the advanced Western European countries is the hereby termed wild-
geese-flying pattern of industrial development.
Wild geese fly in orderly ranks forming an inverse V, just as air-
planes fly in formation. This flying pattern of wild geese is metaphori-
cally applied to the below figured three time-series curves each denoting
import, domestic production, and export of the manufactured goods in
less-advanced countries. Figure I is a model of the time-series curves of
Japan’s import, domestic production, and export in respect of cotton yarn,
cotton cloth, spinning and weaving machinery as well as general machines
and tools from about 1 870 to World War II. The term wild-geese-
flying pattern was derived from this model figure.
Explanation of the meaning of the wild-geese-flying pattern will be
given in a later paragraph. I will only point out in this context the
following facts with regard to the economic development of less-advanced
countries. First, for all industrial goods there exists a sequential order,
from import to domestic production and further to export. Secondly, the
time for the curves of domestic production and export to go beyond that
of import will come earher in crude goods and later in refined goods,
and similarly, earlier in consumer goods, and later in capital goods.
Thirdly, the import curve falls in proportion to the rise of the domestic
production curve, and it is probable that the export curve will sooner or
12 THE DEVELOPING ECONOM肥S
Figure1。WILD－GEESE－FLYING PATTERN
ノ《
、
．・爪、＼
！／ ＼＿
Cotton yarn
’【＼
7 ’
1
！〆一＼
ノ ＼
！！／ ＼
Spinning＆Weaving
Machinery
NoteSl
Cotton cbth Machines＆Tools
L These curves cover中e period from about1870to World War II
2。 一 一一一一 Import
－
Production
田一ロー
Export
3。Vertical line denotes value
1ater begin to fall with respect to crude goods or consumer goods and
the domestic production curve of these goods will also decline in the future．1
1．画n4α膨脚1曜4一σεεSε一F1ア’ngP副εrn
First stage： The wild－geese－fiying pattem is classified into several
forms，the丘rst of which is that manufactured consumer goods are import－
ed from advanced countries and，in exchange for them，special products
of less－advanced countries are exported． As mentioned above，some
■ For the“wild－geese－fiying pattem”of industrial development please refer to my trea－
tises listed below；
Kaname Akamatsu，“Waga Kuni Y6mO KOgy6hin no BOeki SOsei”（The Trend of
Foreign Trade in Manu£actured Woolen Goods in Japan），in5hδgアδKε’z躍Roπ3δ
（Joumal of Nagoya Higher Commercial School），1935，pp．1295gg．（in Japanese）；
“Waga Kuni Keizai Hatten no SOg6Bensh6hO”（Synthetic Dialeotics of Industrial
Development in Japan），∫巌．，July，1937，pp．1793gg．（in Japanese）；“Waga Kuni
Sangy6Hatten no Gank6Keitai，tokuni Kikai Kigu K6gyO ni tsuite”（Types in the
Development of Our Imported Industries，With Special Re‘erence to the Machine
and Tool Industry），in Thε』隣’o∫躍加sh∫Roπ3δ，Tokyo l Nov．1956，pp．68s‘7g。（i皿
Japanese）1“A Theory of Unbalanced Growth㎞the World Eco塾omy”，in昭81師ル
schψ1’chεsイ4κh’v，，Hamburg：1961シBd。86，Heft2、pp。1963四。伽English）．
A HISTORICAL PATTERN OF ECONOMIC GROWTH 1 3
manufactured consumer goods thus imported have a destructive effect on
the native handicraft industry of the less-advanqed countries, with the re-
sult that most labourers discharged from the domestic handicraft industry
move into the export industry for special products. Thus, the process of
the above-mentioned heterogeneization or the international division of
labour gets under way.
Second stage : Domestic production of previously imported consumer
goods comes ihto existence, entailing the import of capital goods for the
consumer goods industry, as mentioned above. Homogeneization in this
field with advanced countries spontaneously effects a high-degree hetero-
geneization between the capital goods industry of advanced countries and
the consumer goods industry or primary industry of less-advanced coun-
tries. The fundamental problem in this case is that there should be foster-
ed a domestic consumer goods industry powerful enough to win in the
competition with imported consumer goods and to recover the home
market from the hands of foreign industries. The primary condition for
this consumer goods industry to be established is the existence of a do-
mestic market for these goods. Such a market has already been developed
by the imported goods, which fact has been overlooked in Nurkse’s theory.
Nurkse holds that in less-advanced countries no demand for industrial goods
exists because of their poverty, and in consequence, there is no incentive
to investment. However, the reverse being the case, manufactured con-
sumer goods are being imported in exchange for special products and
form a big export market for the advanced industrial countries.l
In this process of recovering the domestic market, there will arise a
struggle of economic nationalism in less-advanced countries. This presup-
poses the accumulation of capital and the technological adaptability of
the people in those countri~s. Further, it calls for the government’s pro-
tective policy to encourage and promote the consumer goods industries.
Moreover, for the establishment of such domestic industries, there must
be an abundant supply of raw materials, which may be found in the
native resources as primary products, or at times as in the case of Japan,
must for the most part be imported from abroad. In the latter case, not
only capital goods such as machinery but also raw materials must be im-
ported from abroad as production materials.
In Japan, with the rise of the cotton spinning industry since the
1 880’s, her domestic market was gradually recovered from imported pro-
ducts. Prior to this, the Ansei Treaty of 1 8 58 had deprived Japan of
* Albert O. Hirschman. The Strategy of Economic Development, (New Haven : Yale
University Press, 1958) p. 120. In this book the author stresses the important role
of imports in causing the industrialization of developing countries.
/
14 THE DEVELOPlNG J3CONOMIES
her tariff autonomy making it impossible for her to apply a protective
tariff policy. It was probably due to the exellent technological adaptabi-
lity of the Japanese ,in addition to low wages that her domestic products
could overcome this ‘ handicap and win out over the competition with im-
ported goods.
Such requirements are not always met in less-advanced countries.
Therefore, it is a big problem whether these countries may be able to go
through the same development process as did Japan.
Furthermore, establishment of’ a consumer goods industry presupposes
the presence of energy resources, granted that machinery is imported. In
Japan, there are some resources of coal and fairly rich resources of water
power, which started to be exploited parallel with the rise of consumer
goods industries. As a result, coal mining and electric power enterprises
rapidly developed, and so did railroad and shipping enterprises. Such
industries as those of energy, traffic, etc. form the external economy or
the environmental economy. These should start with the birth of the
consumer goods industry. The establishment of these envirofimental
industries is made possible by the existence of a demand market for
energy through the consumer goods industry. The development of the
energy industry as well as of the traffic industry will in turn accelerate
industrialization. Thus, in the second stage of the wild-geese-flying
development, the consumer goods industry springs up accompanied by the
energy industry and the trafiic industry as environmental industries to
support it. Leaving the environmental industries out of consideration, we
can see conspicuously, in the second stage, the development of the do-
mestic production of hitherto imported consumer goods and increased im-
ports of capital goods as well as gradually decreasing importation of con-
sumer goods. At what time the expanded domestic productioh turns to
decrease imports should be empirically studied for each indi･vidual com-
modity. In case the increase in domestic demand surpasses that of dom-
estic production, the increase of imports will continue despite the increase
in domestic production. However, it is clear that at a certain stage of
the development of domestic production, the imports of commodities of
the same kind as domestic products will show a decreasin~ trend.
Third stage : This is the stage when the domestic consumer goods
industry develops into the export industry. By this time most of the
domestic markets have turned into markets for domestic industrial goods.
As production is put on a larger scale for mass production, the products
are exported in increasing numbers to overseas markets. Simultaneously,
the domestic production of hitherto imported machinery comes to the
fore, while the import of capital goods which are substitutive for domes-
A*’ HISTORICAL PATTERN OF ECONOMIC GROWTH 15
tic machinery begins to decline in turn. In Japan, spinning machinery
began to be domestically manufactured around 1 900, and today India has
started the manufacture of spinning and weaving machines.
As stated above, a wild-geese-fiying pattern in regard to the consumer
goods industry has been seen in the developments starting from the im-
port period of the first stage, through the period of domestic production
of the second stage, and up to the implementation of the export industry
of the third stage. In addition, capital goods (excluding raw materials)
to be invested in the consumer goods industry ‘also start being domesti-
cally produced. If imported goods such as machinery ar~ not merely
used as they are, but are improved and adapted to the various conditions
of the country, natural, racial, and technical, this series of consumer and
capital goods manufacturing industries, such as cotton yarn spinning,
cotton cloth, and spinning machinery industries, which are originally the
industries introduced from abroad, will become domestic industries rooted
and assimilated in the country during the third stage. This is the do-
mestic industrialization of the import industry. For instance, Japan’s own
automatic looms have been exported to England, the origin of the modern
cotton industry.
Fourth stage : In the third stage the consumer goods industry was
already homogeneized with that of the advanced countries, attaining the
same standard as that of the advanced countries ; therefore, those count-
ries are no longer less-advanced countries as far as this industry is con-
cerned but have joined the ranks of advanced countries as an exporter
of t.hese goods. In the fourth stage. this advanced status is further ele-
vated. A characteristic phenomenon of this stage is that the export of
consumer ,goods begins to decline. This is attributable to the fact that
consumer goods are put -into production in other less-advanced countries
and development in a wild-gees~-flying pattern is under way. Another
feature is that in this stage, capital goods domestically produced in the
third stage begin to be exported. In other words, in place of the decreas-
ing export of consumer goods, capital goods are expofted and reach the
sta~e of high-degree heterogeneization in regard to other less-advanced
countries, as mentioned above. However, the domestic production of
machi,nery as well as its export to less-developed countries forms a stage
of high-degree homogeneization towards advanced countries, causing a
conflicting relationship with the advanced countries with respect to capital
goods export.
This fourth stage of wild-geese,-flying development is what Japan has
now arrived at. Although Japan’s export of consumer goods to less-
advanced countries is showing a downward trend, there is an indication
1l~
1 6 THE DEVELOPlNG ECONOMIES
that the export of consumer goods to advanced countries is on the in-
crease. The reason is that in advanced countries the capital goods indu-
stry has developed tremendously, while the consumer goods industry is
becoming stagnant, and in addition high costs due to high wages make
the import of consumer goods from less-advanced countries more profit-
able. Thereupon, what had been imported from advanced countries in
the early development stages of less-advanced countries are now, converse-
ly, exported to advanced countries from the less-advanced countries. Thus,
there arises a tendency in which the most advanced industrial countries
import consumer goods from medium-advanced industrial countries, result-
ing in an international division of labour through high-degree heterogene-
ization. The wild-geese-flying pattern sees its completion in the fourth
stage, with respect to capital goods such as machinery, by going through
the importation beginning from the second stage, the initiation of domes-
tic production in the third stage, and the switch-over to export in the
fourth stage. Here, domestic industrialization is also achieved for the
capital goods industry. However, there is a possibility that another new
stage will be developed in regard to the capital goods industry. It is the
stage during which the export of spinning machinery, for instance, is
eventually decreased, and this may inevitably happen when other less-
advanced countries develop the domestic production of the machinery.
2. Wild-Geese-Flying Pattern Development from Crude Goods to Elaborate
Goods
In the foregoing paragraphs, I have referred to the fundamental wild-
geese-flying pattern of development describing how consumer goods and
capital goods form a: wild-geese-flying order in three or four different
stages. It must be noted in this connection that these consumer goods,
and capital goods as well, vary both in kind and quality. Take cotton
cloth as an example: there are numerous differences in quality from the
crude to the elaborate according to the diff:erence of denier and process-
ing grade. Also, the kinds vary between low-class and high-class goods.
Moreover, as to capital goods, there are a number of differences in
kind and quality constituting a big discrepancy in the elaborateness of the
machinery make-up. Consequently, taking the existence of international
trade for granted, the industrial development of less-advanced countries
will, as a matter of course, take the form of the development of a wild-
geese-flying pattern from crude goods towards elaborate goods. First, im-
port of crude industrial goods is initiated, and in time it dev. elops into
their domestic production and, further, into their export. Meanwhile, the
A HISTORICAL PATTERN OF ECONOMIC GROWTH 17
import of crude goods decreases, but conversely the import of elaborate
goods increases in proportion to the increase in the national income.
However, the grade of domestic production is raised from crude goods to
elaborate goods, and that of export goods is also elevated proportionate-
ly. When the development of less-advanced countries reaches the third
stage with the consumer goods industry developing into the export indus-
try, crude goods are exported first as a matter of course, in which case
the importing countries are naturally less advanced than the exporting
countries. Take Japan as an example: Before the industrialization of Japan,
other countries in Asia, including China, had industrial structures homo-
geneous with Japan, chiefly producing such agricultural products as rice,
tea, silk, etc. However, the industrialization of Japan made the industrial
structure of Japan heterogeneous with those of other Asian countries,
with the result that crude textile goods was first exported to those coun-
tries. As similar textile industrie~ came into being in the less-developed
areas of Asia, Japan turned to produce and export higher quality textile
goods. The elevation of quality was in compliance with the rise in the
income standard of the less-developed area, and at the same time, it
resulted in expanding its export market to the advanced industrial coun-
tries of the west, which are bringing their production and export to bear
upon heavy and chemical industry. The development of a wild-geese-
fiying pattern from crude goods to elaborate goods was also the develop-
ment of the export market from a low income area to a high income area.
As for capital goods, those which were exported in the fourth stage
of the wild-geese-flying development were of a comparatively simple type,
and their destinations were less-developed areas. In Japan, however, as
more and more elaborate and complicated machines were imported from
the advanced industrial countries, there took place the domestic produc-
tion of more elaborate machines which were increasingly exported. Thus,
as in the case of textile goods the export market expands from the less-
developed area to the advanced area to the degree that products become
more elaborate.
3. Development of Advanced and Less-Advanced Countries in a Wild-
Geese-Flying Pattern
The third type of a wild-geese-flying pattern is clear from what is stated
above. The countries of the world form a wild-geese-flying order from
the advanced countries which have reached the stage of high-degree heavy
and chemical industries to the less-advanced countries which are still in
the stage of primary industries. The less-advanced ‘wild geese’ are chas-
1 8 THE DEVELOPlNG ECONOMIES
ing those ahead of them, some gradually and others rapidly, following
the course of industrial development in a wild-geese-flying pattern, as
described above.
The advanced ‘wild geese’ which are in the lead are flying onward,
incessantly achieving technological innovations and trying to maintain a
certain distance of heterogeneous difference from the less-advanced ‘wild
geese’. It can be said in this connection, however, that this technological
innovation, combined with the tendency on the part of less-advanced
countries that their economies become homogeneous with those of advanc-
ed countries, has the function of producing a contradiction in the world
economy in that it creates synthetic materials who.se products become
homogeneous with natural raw materials and foodstuffs of primary indus-
tries. But, at the same time, the contradiction due to economic homo-
geneization is continually being overcome by the other function of tech-
nological innovation, that is, to heterogeneize the economies of advanced
countries with those of less-advanced – countries. If the development of
advanced countries and less-advanced countries in the wild-geese-flying
pattern is to proceed at the same speed, the homogeneization by the
progress of the less-advanced countries means the heterogeneization by a
similar progress in the advanced countries, and substitutive conflicts which
arise every moment will be offset by complementary co-accelerations which
take place every moment. However, these countries, advanced and less-
advanced, do not necessarily go forward at the same speed in their
development of a wild-geese-flying pattern, nor do they always make
gradual progress, but they are at times dormant and at other times make
leaping advances. That the economies of advanced countries are some-
times stagnant and sometimes make leaping advances causes the econo-
mies of less-advanced countries to make similar movements. Some of
the less-advanced countries always remain in a stagnant state falling more
and more behind in the wild-geese-flying order, while others, like Japan,
joined the ranks of advanced countries by making rapid advances and
are strengthening a high degree of homogeneization. However, the wild
geese order of industrial development from the advanced countries to the
less-advanced countries is not a one-series row, but is divided into several
wild-geese-flying rows, one following another. There is a wild-geese-fly-
ing group with America taking the lead, and a Western European group
with England and Germany taking the lead, as well as a comparatively
small group with Japan taking the lead. It is needless to say that those
groups consisting of advanced and less-advanced countries complexly in-
termingle with each other.
A RISTORICAL PATTERN OF ECONOMIC GROWTH 19
皿．SoME ILLUSTRATIONS OF THE WILD・GEESE－FLYING
DEVELOPMENT
The above－mentioned theory of the wild－geese・Hying development of
less－advanced countries is based upon an empirical study of the economic
development of J哉pan over the period since the early Meiji Era，when
she entered the realm of intemational trade．A model of the series of
actual statistical figures is already shown in Figure 1． Let us show in
the fo1正owing血gure several cases of wild－geese－Hying development base（1
0n actua1血gures． The bottleneck in such studies is that it is dif五cult t
o
make a retrospective study on the series of・a certain commodity going
back to its original kind because 霊he commodity classi五cations both in
trade statistics and in production statistics were often changed an（1diver－
si五ed during the per孟od of observation． I have herein graphed out in
Figure2Japan’s import，domestic production，an（1export of cotton yarn
to begin with，despite its duplication with the model in Figure1，as t1丘s
denotes the typical wild－geese－Hying Pattem of manufact皿e（1consumer
goods，
Figure2． IMPORT，PRODUCTION，AND EXPORT
OF COTTON YARN IN JAPANF
Sources：
1000
100
ユ0
thousand bales
卜㌔ 卜㌧ 卜 卜㌧ 卜 卜㌧ 卜 o 卜㌧ OD （ハ O r－l c団 oつ LO
～ ～ ～ 9 ～ ～ ～ ～
器 巽 器 ～ 零 曽 舘 躍
OD OD σ》
け の ぱ
T6y6Keizai Shimp6・sha，1Whoπβδθ耐Sθかαπ（CloseReview of JapaneseForeig取
Trade〉，1935．Ministry ofFinance，θ磁oんμBδ襯漉遅ρアδ（ChronologicalTable
of Japanese Foreign Trade），Ministry of Commerce an41ndustry，1◎δ
Tδたθ∫助δ（Plant Statistics Table），Ministry of Intemational Trade and
Industry，κ6gyσ rδたε∫πyδ（lndustry S亡atistics Table），
Production
z
．窪．
ノト、
＼、
、墨 、
￥
イ￥
■
∠、
＼ ノ
ソlmport、
＼ ／
！
、∠
V
20 THE DEVELOPlNG ECONOMIES
Figure 2 shows that the import of cotton yarn kept on increasing
after Japan opened her ports under the Ansei Treaty (1858) until 1880.
However, the domestic production of cotton yarn rapidly increased from
1870 to 1880, while the import of cotton yarn began to decline. The
export of cotton yarn began to increase from the 1 890’s. At the turn of
the century the growing export showed a speedy rise, surpassing the de.-
clining import. One of the interesting points about the wild-geese-flying
pattern of cotton yarn in Figure 2 is that imports decreased down to a
negligible figure around 1 9 1 O and then again increased up to World War
II. Other points are that exports began to decline from their peak of
1908-1917 and that production declined in the period after World War
II to a point lower than that of prewar days. A11 this was due to the
rise of cotton yarn spinning in the countries less advanced than Japan,
which caused the downward trend of Japan’s cotton yarn export. Japa-
nese spinning companies advanced to the Shanghai area of China. Mean-
while, the cotton spinning industry built by Chinese national capital flou-
rished there also, at least as far as coarse thread cotton yarn was con-
cerned, having reached the third stage of wild-geese-flying development.
Consequently, with respect to coarse yarn it became advantageous for
Japan to import it rather than to produce it at home, and the Japanese
cotton spinning industry was levelled up to the production of fine yarn.
Japan began to place an importance on the fostering of heavy, chemical
and machinery industry after the outbreak of the Manchurian Incident
in 1 931 and so it became even necessary for her to depend partly upon
the import of cotton yarn. This resembles the above-mentioned circum-
stances under which the Western European countries, England in particular,
which reached the higher stage of industrialization centred on heavy,
chemical, and machinery industries, turned to import cotton yarn and
cotton cloth from the less-developed areas.
Shown in Figure 3 is a wild-geese-flying pattern development of the
Indian cotton cloth industry in comparison with the cotton industry of
Japan. As long-range statistics are unavailable regarding the Indian cot-
ton industry, only the period covering years before and after World War
II has been taken up. As the statistical materials do not cover long-range
development, we cannot clarify in this figure the process in which imports
of cotton cloth first increased and when the rise of domestic production
followed. However, the importation of Indian cottolL cloth took a down-
ward trend after the 1 936 peak, while the exportation took an upward
trend at about that time until it attained its peak in the period after
World War II, particularly during the war in Korea. It is presumed that
had it not been for Korean War, India’s export of cotton cloth would
A HISTORICAL PATTERN OF ECONOMIC GROWTH
Figure 3. IMPORT, PRODUCTION, AND EXPORT
OF COTTON CLOTH IN INDIA
l0,000 millions
lOOO millions
lOO millions
21
10 million yards
~r to c¥1 Lo ~ o co cY) co Lo Lo ~t lo Lo ol ~
Note : The unit of cotton cloth production is metric.
Sources : For export and import, Statistical Abstractfor India (prewar figures) and Statis-
tical Abstract for the Commonwealth (postwar figures) ; for production, U.N.,
Statistical Yearbook, 1956.
have drawn a curve of grad.ual increase. The wild-geese-flying pattern
shown in Figure 3 is the third stage, denoting the period of Indian cotton
industry established as an export industry.
As for the wild-geese-fiying pattern development of the machine and
tool industry in Japan, Figure I shows its model. Figure 4 shows the
import, domestic production, and export of Japan’s machines and tools
based on statistical data. The lower part of Figure 4 shows the trends
after the early years of Meiji, i.e. from 1868 to World War II, while
the upper part shows trends from the terminatidn of the war in 1 945 to
1 955. The production statistics by money value in Figure 4 show the
results of research conducted every five years between 1909 and 1919,
and every year after that period. Trade statistics can be traced back to
1 868. The import of machinery shown in the lower part of Figure 4
mounted to the one million yen bracket in the 1 870’s. Considering that
22 THE DEVELOPING ECONOMIES
Figure 4. IMPORT, PRODUCTION AND EXPORT, OF
MACHlNES AND TOOLS IN JAPAN
Domar
rooo ~iui.~*
800
400
loo ~i lion=
80
8~~~
lo ~iui.~*
l ooo *iui.~,
800
400
200
1 mj =.~
80
roo th.*==~d=
10 ~ ‘i.ns
Yen
+ I mj ‘on I ~jj[to~
1 oo th.**.~d= Ioo th.~=.~d* ~
R
”
Source : Tame as Figure 2.
Japan’s import of cotton cloth was then in the five million yen bracket,
needless to say the importS at that time were mainly consumer goods.
Industrialization in Japan made a long step forward in the ,1 890’s, and
from that time on the import of machinery rapidly increased. Although
it is not possible to determine in this figure the time when the produc-
tion of machinery surpassed the minimum one-hundred thousand yen
bracket, it is clear that it was sometime before 1 890 when export statis-
tics started. Though the export of machinery started comparatively early,
among the items exported in the early period were Jinrikishas listed as an
important item. Original invention in Japan as it is, it is out of consi-
deration in the study of a wild-geese-flying development. Before and
after 1 900 such consumer goods as table-clocks and wall-clocks became
important items, and thereafter the export of machinery for production
means appeared. On the occasion of World War I, Japan showed a rapid
increase in the production as well as import of machinery, while its im-
ports showed a slow decline. A setback was seen after that war, but after
A HISTORICAL PATTERN OF ECONOMIC GROWTH 23
the Manchurian Incident in 1 931 the same trends as in World War I
appeared as a result of the heavy industrialization of Japan, with the
export curve in 1939 having surpassed the import line. As shown in the
upper part of Figure 4, after World War 11 the export curve took a
position above the import turve, but the imports rose rapidly out of the
necessity for the modernization of Japan’s industry. This should be re-
garded as proof of the industrial revolution in restored Japan. As mo-
dernization is achieved in almost all areas of industry, the import curve
will be stagnant while the export curve will continue to rise rapidly.
Next, Iet us pick up from among the products of Japan’s machine
and tool industry two or three items which had completed the course of
wild-geese-flying development before World War II. As such, spinning
machinery, bicycles (including parts) , and electric machinery are shown
in Figure 5 . Of these three industrial goods, it is shown that the in~-
port of spinning machinery tose rapidly after the early years of the Meiji
Era ; this reflects the establishment of Japan’s cotton spinning industry
in the 1 890’s. At about the time of World War I, Japan’s spinning in-
dustry advanced conspicuously, followed by the increased import of spin-
ning and weaving machinery. This peak was reached about 1918-1922
and thereafter the imports show a downward trend. This was of course
due to the rapid advance of Japan’s spinning and weaving machinery
industry before and after World War I. The amount of production sur-
paSsed that of import in the 1920’s. The exports increased accordingly,
surpassing the import line in the 1930’s. As Japan’s economy rushed
into a semi-war structure with the outbreak of the Sino-Japanese war in
1 938, both production and export declined. Though they rose again after
the war, a future decline in export and consequently in production is
foreseen if the development of the developing countries in a wild-geese-
flying pattern is taken into account. As for the import of electric
machinery, it showed a gradual increase, though not as rapid as that of
spinning and weaving machinery, until it reached its peak in 1 923-1927,
later than spinning and weaving machinery, and then began to decline.
When production statistics were started in 1 909, the amount of produc-
tion was less than that of import, but soon after that the production
curve exceeded the import curve. The exports surpassed the imports at
about the same time as in the case of spinning and weaving machinery.
Lastly, the importation of bicycles was recorded from the 1 890’s, a
little later than the two cases above. It increased gradually until it rea-
ched its peak in 1923-1927, as was the case with electric machinery, and
then it declined suddenly. The pro.duction exceeded the import during
World War I, while exports advanced very rapidly, surpassing imports
24 THE DEVELOPING ECONOMIES
100mmlons
80
40
20
10mi”ions
4
2
1mimon
400
200
ユ00tbousands
800
400．
200
100mi聴ions
80
40
20
10mi聡ions
4
2
1mi”ion
400
200
100thousands
工OO mil『ions
80
40
20
10mi闘ions
4
2
1miUion
400
200
ユOO thousands
Figure5． A WILD－GEESE・FLYING PATTERN OF
MACHINERY（PREWAR）IN JAPAN
Sp血ing＆Weaving Machinery
Production
Export
ノ 、一
！1 、 、
！
！
1 ノ
lmport
￥
、
ノ
ノ ’
、
Electric Machinery
Production
Export
／ ＼
’1
、 、
、 ＼ 、 ノ 、 、
－
！ ノ lmport
！
1
！ ノ
ノ
ノ ．（
Bicycles
Production
’
噛
’ Export
￥
1 ￥
γ ノ
／
■
’ ＼Import
1
ノ
￥
！
！ ノ
卜
OD
～
oO
．OD
oo
qo
匙
OD
oo
卜
o
～
oo
o
doo
聖
oo
o
卜 N 卜 d 卜
O H H d
q
～ ～ ～ ～ ～
α》 oo “り op oo
O o 【‘ H “
①
H
Source＝ Same as Figure2。
q
の
～
oo
q
卜
oり
～
α》．
ρD
OD σ9
0り OD
σ、 σ、
r4 r6
A HISTORICAL PATTERN OF ECONOMIC GROWTH 25
earlier than the above two commodities. The production of bicycles is
comparatively simple from technical point of view. There fore, although
their import was started later than the other items, the domestic produc-
tion rose earlier and the export increase was more rapid, with the result
that imports quickly dwindled almost to a negligible point. This is different
from the case of electric machinery, whose import has shown downward
trends and yet is lingering around a fairly high level because it involves
complicated techniques as well as a variety of types.
So far we have graphed out wild-geese-flying patterns of several
commodities based on statistical data to illustrate the historical stages of
the economic development of less-advanced countries. For the furtherance
of this sort of study, statistical data should be prepared in connection
with many industrial goods, for example, the various kinds of electric
machinery. Also, statistics of the importation, production, and exporta-
tion of those commodities in less-advanced countries must be collected
over the longest possible period of time. Provided this is accomplished,
the following would be possible. Firstly, it could be determined in what
stage of wild-geese-flying development a certain less-adyanced country
stands in respect of various commodities of consumer goods and capital
goods. Secondly, how the production and trade of the country is being
transformed according to each development stage could be made clear,
and at the same time how they would change in the future could be
foreseen. Thirdly, other countries trading with the above country may
be able to predict a falling trend of the export of a certain item of
consumer goods or a rising trend of the export of a certain item of
capital goods. Of course, to make such predictions we must make an
inquiry into how far it is possible for a certain country to advance the
stages of a wild-geese-flying development. Some countries may advance
quickly and some slowly, while others may stand still at a certain stage.
There is such a case as the growth of the Japanese transistor radio in-
dustry after World War 11 in which the wild-geese-flying development
completed its course from import to domestic production and, further, to
export in the short period of several years. In Japan, there are also such
cases as electronic instruments which are still in the import stage, and
other cases where domestic production has reached thesecond stage but it
is uncertain whether it can develop into an export industry of the third stage.
These various patterns of wild-geese-fiying development should be
explained by studies on the economic circumstances of the individual
countries. At the same time it should be noted that the industrial policy
of a country has a great influence on the wild-geese-flying pattern, as,
for instance, import restrictions cause a sharp decline in the import curve.
Accelerat ing t he world’s research.
Food consumption patterns and
economic growth. Increasing
affluence and the use of natural
resources
P. Gerbens-leenes
Appetite
Cite this paper
Get the citation in MLA, APA, or Chicago styles
Downloaded f rom Academia.edu
Related papers
Consumpt ion pat t erns and t heir ef f ect s on land required f or f ood
P. Gerbens- leenes
Short comings of a growt h- driven f ood syst em
Michalis Hadjikakou
A met hod t o det ermine land requirement s relat ing t o f ood consumpt ion pat t erns
S. Nonhebel, P. Gerbens- leenes, W. Ivens
Download a PDF Pack of t he best relat ed papers
https://www.academia.edu/18139385/Food_consumption_patterns_and_economic_growth_Increasing_affluence_and_the_use_of_natural_resources?auto=citations&from=cover_page
https://www.academia.edu/18139385/Food_consumption_patterns_and_economic_growth_Increasing_affluence_and_the_use_of_natural_resources?from=cover_page
https://www.academia.edu/18139387/Consumption_patterns_and_their_effects_on_land_required_for_food?from=cover_page
https://www.academia.edu/38469862/Shortcomings_of_a_growth_driven_food_system?from=cover_page
https://www.academia.edu/17821841/A_method_to_determine_land_requirements_relating_to_food_consumption_patterns?from=cover_page
https://www.academia.edu/18139385/Food_consumption_patterns_and_economic_growth_Increasing_affluence_and_the_use_of_natural_resources?bulkDownload=thisPaper-topRelated-sameAuthor-citingThis-citedByThis-secondOrderCitations&from=cover_page
Research report
Food consumption patterns and economic growth. Increasing affluence
and the use of natural resources
P.W. Gerbens-Leenes a,*, S. Nonhebel b, M.S. Krol a
a Faculty of Engineering Technology, Water Engineering and Management, University of Twente, P.O. Box 217, 7500 AE, The Netherlands
b Center for Energy and Environmental Studies (IVEM), University of Groningen, Nijenborg 4, 9747 AG Groningen, The Netherlands
Introduction
At present, the world faces enormous challenges over food
security (Millenium Ecosystems Assesment, 2005), which threaten
the availability and quality of natural resources such as arable
lands, freshwater and natural areas (FAO, 2003; Hoekstra &
Chapagain, 2008; WWF, 2007). The potential impacts of climate
change are likely to worsen this situation (Fischer, Van Velthuizen,
Shah, & Nachtergaele, 2002). Globally, food consumption gives rise
to the greatest use of land (FAO, 2003; Penning de Vries et al., 1995)
and freshwater (Falkenmark, 1989; FAO, 2003; Hoekstra &
Chapagain, 2008; Rockstrom, 1999; Rosegrant & Ringler, 1998)
and is an important cause of greenhouse gas emissions (Carlsson-
Kanyama, Engström, & Kok, 2005; Kramer, 2000). The current
growth in the world population requires the production of more
food. As well as population growth, most areas of the world
have shown economic development that resulted in increased
purchasing power, causing not only a demand for more food
(Latham, 2000) but also for different food. Studies on human
nutrition have shown that worldwide a nutrition transition is
taking place, in which people shift towards more affluent food
consumption patterns (FAO, 2003; Grigg, 1995; Popkin, 2002).
Globalization of nutrition includes shifts from local markets
towards global trade in commodities and processes in which
people and ideas spread throughout the world (Lang, 2002) and
thereby change consumption. Since the beginning of the eigh-
teenth century, the nutrition transition that accompanied eco-
nomic development has caused large shifts in food consumption
patterns in Europe and the United States (Fogel & Helmchen, 2002).
When economic development occurs in developing countries as
well, as is the case in China today (IMF, 2010), nutritional changes
put additional pressure on limited natural
resources.
The use of natural resources for food is the combined effect of a
specific consumption pattern and production system. Several
scientists describe the complex links between sustainable
consumption and the limited availability of natural resources
(e.g. Hertwich, 2005). Duchin (2005) provides an overview of
studies on energy and land required for food and suggests that a
shift from affluent consumption patterns towards a Mediterra-
nean-type pattern, characteristic of Greece in the 1960s, has
favorable impacts on the environment.
Appetite 55 (2010) 597–608
A R T I C L E I N F O
Article history:
Received 11 March 201
0
Received in revised form 1 September 20
10
Accepted 14 September 2010
Keywords:
Dietary change
Economic development
Natural resource use
Nutrition transition
Food consumption patterns
A B S T R A C T
This study analyzes relationships between food supply, consumption and income, taking supply, meat
and dairy, and consumption composition (in macronutrients) as indicators, with annual per capita GDP
as indicator for income. It compares food consumption patterns for 57 countries (2001) and gives time
trends for western and southern Europe. Cross-sectional and time series relationships show similar
patterns of change. For low income countries, GDP increase is accompanied by changes towards food
consumption patterns with large gaps between supply and actual consumption. Total supply differs by a
factor of two between low and high income countries. People in low income countries derive nutritional
energy mainly from carbohydrates; the contribution of fats is small, that of protein the same as for high
income countries and that of meat and dairy negligible. People in high income countries derive
nutritional energy mainly from carbohydrates and fat, with substantial contribution of meat and dairy.
Whenever and wherever economic growth occurs, food consumption shows similar change in direction.
The European nutrition transition happened gradually, enabling agriculture and trade to keep pace with
demand growth. Continuation of present economic trends might cause significant pressure on natural
resources, because changes in food demand occur much faster than in the past, especially in Asia.
� 2010 Elsevier Ltd. All rights reserved.
Abbreviations: A%, average supply of nutritional energy from animal sources (%); E%,
energy percentage; FAO, Food and Agriculture Organization of the United Nations;
GDP, gross domestic product; GE, grain equivalents; G-K dollars, 1990 International
Geary-Khamis dollars; PPP, purchasing power parity; WHO, World Health
Organization.
* Corresponding author.
E-mail address: p.w.gerbens-leenes@utwente.nl (P.W. Gerbens-Leenes).
Contents lists available at ScienceDirect
Appetite
journal homepage: www.elsevier.com/locate/appet
0195-6663/$ – see front matter � 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.appet.2010.09.013
http://dx.doi.org/10.1016/j.appet.2010.09.013
mailto:p.w.gerbens-leenes@utwente.nl
http://www.sciencedirect.com/science/journal/01956663
http://dx.doi.org/10.1016/j.appet.2010.09.013
Agriculture is the basis for the production of food, providing
commodities such as wheat and raw milk. The number of
agricultural commodities that are important for the global supply
of food is limited to about 21 when expressed in terms of weight of
annual global production (FAO, 2010). The 15 main categories of
crop commodities, expressed as annual global production (kg), are
sugar cane, root crops, vegetables, maize, paddy rice, wheat, fruits,
potato, sugar beet, cassava, soybean, barley, pulses, oil seed rape
and sorghum; the six main animal commodities are raw milk, pork,
poultry, beef, mutton and goat’s meat (FAO, 2010). These
commodities provide the ingredients for a large number of
different food items, such as pizza or cheese. The mass of food
items mainly consists of only four different components: water
and the three macronutrients carbohydrates, fats and proteins
(FAO, 2010; Voedingscentrum, 1998a, 1998b; Whitney & Rolfes,
1999). Food items form the basis of food consumption patterns,
defined as the consumption of specific food items and their
combination in dishes and meals. These patterns show large
temporal and spatial differences, mainly caused by the availability
of commodities, cultural aspects and economic factors (Whitney &
Rolfes, 1999). The requirements for specific natural resources for
each food item are determined by the production system. In
general, food items show large differences in the requirement for
land (Gerbens-Leenes & Nonhebel, 2005), energy (Kramer, 2000)
and freshwater (Hoekstra & Chapagain, 2007), resulting in
substantial variations in requirements for natural resources
between food consumption patterns. As a rule, affluent western-
style food consumption patterns need more natural resources than
those of poor developing countries (e.g. Duchin, 2005; Gerbens-
Leenes & Nonhebel, 2002; Hoekstra & Chapagain, 2007). Food
items that are typical of affluent patterns are fats, drinks and foods
derived from animal sources, such as milk, cheese and meat. These
items have a substantial impact on natural resources, either
through heavy consumption (for example of beer), or through large
specific resource requirements per unit of food. For example, in
western countries the contribution of fats to land requirements is
about 25% and that of meat about 30% (Gerbens-Leenes &
Nonhebel, 2002), while the contribution of meat to energy and
freshwater requirements is also about 30% (Gerbens-Leenes &
Hoekstra, 2007; Kramer, 2000). It is therefore important to identify
the relationship between economic growth and more affluent food
consumption patterns.
Hundreds of detailed studies from the nutritional, social and
agricultural sciences, as well as food security studies, are
available. Nutritional and social studies express consumption in
terms of specific food items (e.g. Mennell et al., 1992; Receveur,
Boulay, & Kuhnlein, 1997; Whitney & Rolfes, 1999), agricultural
studies on global food security simplify consumption to basic
and affluent diets and show them in grain equivalents (GE) (e.g.
Penning de Vries et al., 1995), while other studies address food
security as the average per capita availability of commodities
(FAO, 2003). The agricultural and food security studies often
show time trends, emphasizing the need to increase agricultural
production. The effect of income on food consumption patterns
is recognized as one of the factors that determine food choice
(e.g. Van der Boom-Binkhorst et al., 1997; Von Braun, 1988; Von
Braun & Paulino, 1990; Ivens et al., 1992; Musaiger, 1989;
Vringer & Blok, 1995; Wandel, 1988; Whitney & Rolfes, 1999; De
Wijn & Weits, 1971). Among the poorest people, be they
individuals or nations, diets tend to be composed principally of
cheap starchy staple foods: wheat, rice, potatoes, cassava and
the like (Jobse-van Putten, 1995; Poleman & Thomas, 1995).
Existing research often focuses on health issues and changes in
time. General relationships between economic change and the
rate of change in food consumption patterns are also important
to be explored.
The nutrition transition began in developed countries 300
years ago. It coincided with great economic growth (Maddison,
2003). If developing countries follow the same route, it would
mean a major shift in the balance between global food demand
and supply, with considerable consequences for natural
resources. It is therefore important to investigate whether there
are general relationships between economic growth and food
consumption patterns. This is also important when growth is
negative. The FAO, for example, estimated that in 2007 75 million
people were pushed into undernourishment as a result of higher
food prices mainly caused by an increase in commodities used for
bio-energy, bringing the total number of hungry people in the
world to 923 million (FAO, 2008). It is possible that the current
financial crisis will diminish purchasing power and so increase the
risk of a drop in food intake.
When food consumption patterns are expressed in terms of
food items, differences among the patterns are large and studying
them requires a great amount of data and time. This paper
presents an analysis of nutritional changes due to economic
growth, so being situated between detailed consumption pattern
analyses in terms of specific foods items, which are only valid for a
limited group of consumers, and the coarse agricultural analyses
based on simplified diets in terms of GE. The paper expresses
changes in terms of macronutrients and related nutritional
energy. The specific aims of this study are to quantify shifts in
food consumption patterns that accompany economic develop-
ment. The research questions are: (i) what are the trends in
national per capita food supply, measured in terms of nutritional
energy and macronutrients that follow economic changes? (ii)
what are the trends in individual per capita food consumption, i.e.
the food actually eaten, measured in terms of nutritional energy
and macronutrients, that accompany economic changes? (iii) in
which regions will large changes in food supply and consumption
occur in the next 10 years? To analyze the impact of economic
changes on food consumption patterns this study addresses
similarities among patterns in terms of composition, rather than
differences in terms of food items. This approach makes it possible
to identify general trends. By differentiating between national per
capita supply and individual consumption, it also shows trends for
the gap between supply and consumption. The study analyzes
cross-sectional and time series relationships, revealing general
trends. These trends provide a better understanding of the
connection between food consumption and environment and can
contribute to environmental studies that aim to indicate
transition pathways towards a more sustainable use of natural
resources.
[()TD$FIG]
National
agricultural
production
Import Export
Gap
Consumer
supply
Consumption
(food eaten)
Fig. 1. Simplified food system overview. Consumer supply is defined as per capita
national food availability and is a function of national production + import � export.
Consumption is food actually eaten. The difference between supply and consumption
is the gap caused by food chain losses.
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608598
Food systems
Food systems include production and related supply of
commodities and foods as well as consumption, defined here as
the food actually eaten. Figure 1 shows a simplified overview of the
system. Agricultural production provides crop and animal com-
modities. On a national level, supply is a function of national
production plus imports minus exports. Consumer supply is defined
here as the per capita national availability of food, consumption as
actual consumed food. The difference between supply and
consumption is the gap caused by losses in the food chain.
Production
Starting in 1961, FAO food balance sheets (FAO, 2010) provide
information for almost all countries in the world on the annual
market supply of agricultural commodities. The food industry
selects and processes agricultural commodities to manufacture
food items (Catsberg & Kempen-van Dommelen, 1997). It often
divides commodities into several fractions based on composition
characteristics, such as the fat, protein or carbohydrate content
(the macronutrients). The fractions form the basic ingredients for
food items, when industry joins in and processes ingredients.
Soybeans, for example, are split into an oil and an oil cake fraction
(Kramer & Moll, 1995). Oil is a basic ingredient in margarines, oil
cake in livestock feed. In the western world, technological
developments in agriculture, transportation and food conservation
at the end of the 19th century prompted the expansion of the food
industry and food preparation shifted from households to industry
(Jobse-van Putten, 1995), so stimulating the nutrition transition.
Figure 1 shows that agriculture, inside or outside the national
borders, determines consumer supply available for
consumption.
Sometimes wastesarereused,forexamplemanureor waste fromthe
food industry for livestock feed (Nonhebel, 2004). In every link of the
system losses take place. For example, the food industry processes
1.4 kg of wheat to manufacture 1 kg of flour (Kramer & Moll, 1995). A
study on household consumption in Sweden has estimated that
during meals 10% of the food remains behind on the plate and is
wasted (Karlsson,2001).Theselosses cause a gap betweenconsumer
supply and consumption, i.e. the food actually eaten.
Consumption
Consumer supply concerns the food available for consumption at
a national level. Consumers buy it in shops or sometimes produce it
in
their gardens (Fernandes & Nair, 1986; Pallot & Nefedova, 2003).
The repeated arrangements of consumption, characterized by types
and quantities of food items and their combination in dishes and
meals, are termed food consumption patterns (Gerbens-Leenes &
Nonhebel, 2002). Factors such as preferences, habits, availability,
tradition, culture and income influence these patterns (Van der
Boom-Binkhorst et al., 1997; Von Braun, 1988; Von Braun & Paulino,
1990; Ivens et al., 1992; Musaiger, 1989; Vringer & Blok, 1995;
Wandel, 1988; Whitney & Rolfes, 1999; De Wijn & Weits, 1971). For
instance, when income increases, peoplespend more money on food
(Pindyck & Rubinfeld, 2005; Vringer & Blok, 1995).
Consumption, or human nutrition, concerns the food actually
eaten. For nutrition the composition of food in terms of
macronutrients, its fats, carbohydrates and proteins, is important,
because they provide energy and are essential for the functions of
the human body. Food surveys ask respondents what they have
eaten and provide detailed information on the composition of
consumed food (see also Appendix B). Humans can derive energy
from different combinations of macronutrients. This flexibility
contributes to variations in the macronutrient composition of
nutrition and to
differences in food consumption patterns.
The composition of food
Four components, water, carbohydrates, fats and proteins,
dominate the composition of every commodity (FAO, 2010;
Voedingscentrum, 1998a, 1998b; Whitney & Rolfes, 1999). The
macronutrient content of commodities, such as wheat, soybean or
pork, is genetically determined, so that all crop and animal
commodities show a specific composition (kg macronutrient per
kg dry matter) (Penning de Vries et al., 1989; Schmidt-Nielsen,
1988). Based on composition, commodities form four categories:
(i) starchy staples, crops that mainly provide carbohydrates with
few proteins; (ii) protein-rich crops, which provide proteins as well
as carbohydrates; (iii) oil crops, providing plant-based fats for the
production of oil, and carbohydrates and proteins for feed (Penning
de Vries et al., 1989); and (iv) animal commodities, which provide
high quality proteins and fats (Whitney & Rolfes, 1999). The
composition of a commodity determines its suitability for a food
item (Whitney & Rolfes, 1999). For example, starchy staples, such
as wheat, can be used for bread or pasta, while oil crops, such as oil
seed rape, provide oil (Voedingscentrum, 1998a, 1998b) for
margarines. Consumption changes can cause shifts in the
macronutrient composition of food consumption patterns includ-
ing a demand for different commodities.
The food system and natural resources
Several studies have shown that natural resource use varies
greatly between food items and food consumption patterns (e.g.
Engelenburg van, Rossum van, Blok, & Vringer, 1994; Hoekstra &
Chapagain, 2008; Kok, Biesiot, & Wilting, 1993; Kramer & Moll,
1995; Tukker & Jansen, 2006). Meat, fats, and drinks especially have
relatively large requirements for energy, land and freshwater
(MJ kg�1, m2 kg�1, m3 kg�1). One kilogram of pork providing
2000 kilocalories (Voedingscentrum, 1998a, 1998b), for example,
requires 86 MJ of energy (Kramer & Moll, 1995), 9 m2 of land
(Gerbens-Leenes & Nonhebel, 2005) and 4850 L of freshwater
(Hoekstra & Chapagain, 2008) for its production. In comparison, 1 kg
of paddy rice providing even more nutritional energy (3500 kilo-
calories) requires 22 MJ energy (Kramer & Moll, 1995), 3 m2 land
(Gerbens-Leenes, 2006) and 2300 L of water (Hoekstra & Chapagain,
2008). The example shows that changes in food consumption
patterns can have a considerable impact on natural resource use.
Methods and data
To analyze the relationship between food supply, food
consumption, and the contribution of animal foods to supply on
the one hand and income on the other, this study assesses cross-
sectional and time series relationships.
Units of calculation
The study does not express food supply and consumption in
terms of foods, but simplifies supply and consumption and uses
macronutrient composition (fats, carbohydrates and proteins) as
units of calculation. It indicates per capita food supply in the
fraction of nutritional energy provided by the macronutrients, the
macronutrient energy percentage (E%), as is common in nutrition
research (Whitney and Rolfes, 1999). It shows the contribution of
animal foods to supply in the fraction of nutritional energy derived
from animal sources (A%), and total food supply as availability of
nutritional energy (kilocalories per capita per day). E% and A% are
calculated by
protein E% ¼
P � kcal: p
E
� 100% (1)
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608 599
fat E% ¼
F � kcal: f
E
� 100% (2)
carbohydrate E% ¼
E � ððP � kcal: pÞ þ ðF � kcal: fÞÞ
E
� 100% (3)
A% ¼
A
E
� 100% (4)
where P is the average daily supply of protein (grams), kcal.p the
nutritional energy supply of protein (4 kilocalories per gram), E the
average daily per capita supply of nutritional energy (kilocalories),
F the average daily supply of fat (grams); kcal.f the nutritional
energy supply of fat (9 kilocalories per gram) and A the average
daily per capita supply of nutritional energy from animal sources
(kilocalories). The study derives data on per capita supply from
FAO food balance sheets (FAO, 2010) and values on nutritional
energy for protein and fat from the Dutch Nutrition Council
(Voedingscentrum, 1998a, 1998b).
Average per capita income depends, among other things, on the
development status of an economy, the size of households and
income distribution in a country. The World Bank (2005) calculates
economic output as gross domestic product (GDP). Information on
average per capita GDP is available for most countries from various
sources. The only database, however, that covers recent informa-
tion on GDP as well as historical economic developments is that of
Maddison (2003) which provides information on the economic
development status of almost all countries in the world on a
national and per capita basis from the Middle Ages onwards. It
expresses average GDP in 1990 International Geary-Khamis dollars
(G-K dollars). The Geary-Khamis method is an aggregation method
in which international prices and a countries purchasing power
parity, depicting relative country price levels, are estimated
simultaneously from a system of linear equations and expressed
in G-K dollars (United Nations Statistics Division, 2006). We derive
data on per capita GDP from Maddison (2003) as an indicator for
income (expressed as dollars per capita per year) for the cross-
sectional and time series relationships.
Cross-sectional relationships
The study assesses cross-sectional relationships between
income and food supply, the contribution of animal foods and
the composition of supply for 57 countries in 2001.
Appendix A
gives an overview of these countries. Countries from Africa, Asia,
Eastern Europe, Latin America, the Middle East and the OECD, in
different stages of development and with more than five million
inhabitants, have been selected. These countries form two clusters
of developed and developing countries, with relatively high and
relatively low GDPs, however. To also cover countries with average
incomes, three small transition countries with GDPs in between
the two extremes, the United Arab Emirates, Estonia and Slovenia,
are added, clustered into a small country group.
Time series relationships
Over the last millennium Europe has shown continuous
economic growth (Maddison, 2003). Between 1700 and 2000,
for example, per capita GDP in France increased from 900 to 21,000
dollars and in Great Britain from 1250 to 20,000 dollars. Between
1961 and 2001, Italy, Greece, Spain and Portugal showed a three- to
fourfold increase of per capita GDP. These periods were accompa-
nied by large changes in food consumption patterns (Jobse-van
Putten, 1995; FAO, 2010; Fogel & Helmchen, 2002). Most studies of
historical food consumption describe changes in a qualitative way
(e.g. Jobse-van Putten, 1995; Mennell et al., 1992) and do not
provide quantitative data. An exception is the analysis by Fogel and
Helmchen (2002), which has quantified nutritional energy supply
for France and Great Britain between 1700 and 2000 (kilocalories
per capita per day). To evaluate per capita food supply over time,
this study first assesses a time series relationship between supply
and income in France and Great Britain over a period of three
centuries. It combines data from Fogel and Helmchen (2002) with
GDP data from Maddison (2003).
Secondly, the study evaluates a four-decade time series
relationship in southern Europe. For Italy, Spain, Portugal and
Greece, it assesses the relationship between the increase of per
capita supply, changes in the composition of food consumption
and changes in the contribution of animal foods on the one hand
and income on the other over the period 1961–2001. It applies
Eqs. (1)–(4) and derives data from the FAO (2010) and combines
this with data on GDP from Maddison (2003).
Confirmation of trends and the gap between supply and consumption
This study firstly assesses the relationship between GDP and
national per capita food supply. Secondly, to evaluate whether
information from food surveys can confirm trends found, the study
assesses the relationship between consumption, expressed as
nutritional energy intake (kilocalories per capita per day), and
annual per capita GDP. A number of surveys have been done in
developing countries (FAO, 2005) and two time series are available
for developed countries, the Netherlands (Voedingscentrum/TNO,
1998) and the United States (USDA, 2005). The study combines
data from 31 food surveys from 26 countries (see Appendix B) with
information on GDP from Maddison (2003). Thirdly, it evaluates
the size of the gap between national per capita supply and
consumption. This provides information on food losses in the
supply chain. To confirm trends, the study also compares the fat
E%
of urban and rural consumption. Data for this were derived from 11
surveys in developing countries that made a distinction between
urban and rural patterns (see Appendix B).
Results and discussion
Per capita income and food supply
The cross-sectional analysis indicates a relationship between
per capita food supply and income (GDP) following a power-law
dependency E = 850 � GDP0.14, thus showing an income elasticity
of about 0.14. Figure 2 shows that supply varies between 1600
kilocalories per capita per day for low GDPs and 3800 kilocalories
for high GDPs, a difference of a factor of almost two and a half. The[()TD$FIG]
Fig. 2. Relationship between annual per capita GDP (dollars) and nutritional energy
supply (kilocalories per capita per day) based on data from 57 countries in 2001. The
solid line shows the power-law regression (income elasticity 0.14, R2 = 0.71), the
shaded zone is the 90% confidence band.
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608600
figure also shows that for low GDPs increases in food supply per
unit of GDP are large, while for high GDPs increases are much
smaller and income does not seem to affect supply any longer.
The power-law relationship results in an R2 of 0.71, the
elasticity 0.14 inhibits a standard error of 0.013 (t-stat 11, p value
3.8 � 10�15), while the logarithm of the constant is 2.93 with
standard error 0.048 (t-stat 60.5, p value 3.7 � 10�49). Data
analysis suggests that income elasticity decreases with per capita
income, but this does not result in a more significant relationship.
The figure also shows an appreciable scatter around the identified
relationship; the grey zone indicates the 90% confidence band. This
scatter may be due to socio-economic factors, such as the share of
income spendable on food, income distribution, or cultural
differences in food consumption patterns.
Figure 3 shows that in France and Great Britain increasing per
capita GDP parallels greater food supply, which doubles over the
three centuries considered from 1700 kilocalories per capita per
day in 1700–3500 kilocalories in 2000. The largest increase per
unit of GDP occurs for incomes below 5000 dollars, while above
this level the increase gradually slows down. The figure also shows
the relationship identified in the cross-sectional analysis and the
confidence band around it. Results for food supply in France and
Great Britain are within the confidence band and qualitatively
similar to trends in that analysis and show an income elasticity of
0.23: E = 395GDP0.23. The R2 is 0.88, the elasticity 0.23 inhibits a
standard error of 0.024 (t-stat 9.6, p value 5.6 � 10�7), while the
logarithm of the constant is 2.6 with standard error 0.087 (t-stat
30, p value 1.3 � 10�12). Quantitative differences may be due to,
among other factors, limitations of the spatio-temporal analogue
of comparing spatial differences between countries to temporal
changes within countries.
Figure 4a–c depicts the results for southern Europe in the period
1961–2001. Figure 4a shows that nutritional energy supply
increased from 2500 kilocalories per capita per day for a GDP of
3000 dollars (Portugal, 1961) to 3700 for a GDP of 12,500 dollars
(Greece, 2001). Again, results are qualitatively consistent with the
cross-sectional results (power-law relationship and confidence
band); the income elasticity found here is 0.21. The relation reads
E = 475GDP0.21. The R2 is 0.78, the elasticity 0.21 inhibits a standard
error of 0.026 (t-stat 8, p value 2.6 � 10�7), while the logarithm of
the constant is 2.7 with standard error 0.10 (t-stat 26, p value
9.3 � 10�16). Moreover, Fig. 4c shows that the fraction of food
supply from animal sources is explained well by per capita income
following a power-law with income elasticity of 0.43:
A% = 0.41 � GDP0.43, R2 is 0.88, the elasticity 0.43 inhibits a
standard error of 0.037 (t-stat 11.5, p value 9.9 � 10�10), while
the logarithm of the constant is �0.39 with standard error 0.15 (t-
stat �2.6, p value 0.016).
Per capita income and composition of food consumption
Figure 5a and b shows the relationship between the macronu-
trient composition of consumption and annual per capita GDP for
the cross-sectional analysis. The fraction of nutritional energy
provided by proteins does not change with income and is between
9 and 18 E%. The carbohydrate and fat E%, however, show a
connection with GDP. It can be estimated from Fig. 5a that in
countries with low GDPs, below 5000 dollars, people derive
[()TD$FIG]
Fig. 3. Relationship between annual per capita GDP (dollars) and nutritional energy
supply (kilocalories per capita per day) for France and Great Britain between 1700
and 2000 also showing the relationship of the cross-sectional analysis of Fig. 2. The
solid line and shaded zone denote the relation identified in the cross-sectional
analyses and the confidence band around it.
[()TD$FIG]
Fig. 4. (a) shows the relationship between annual per capita GDP (dollars) and
nutritional energy supply (kilocalories per capita per day) for southern Europe
between 1961 and 2001. The solid line and shaded zone denote the relationship
identified in the cross-sectional analyses and the confidence band around it. (b)
shows
the relationship between annual per capita GDP and the composition of food
consumption patterns in terms of the fraction of nutritional energy derived from fat
(fat E%), protein (protein E%), and carbohydrate (carbohydrate E%) for southern
Europe between 1961 and 2001. (c) shows the relationship between annual per
capita GDP and the composition of food consumption patterns in terms of the
fraction of nutritional energy from animal sources (%) for southern Europe between
1961 and 2001, and the relationship based on
the cross-sectional analysis.
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608 601
nutritional energy mainly from carbohydrates, with a small
fraction from fats. In Bangladesh, for example, the country with
the lowest GDP in the analysis, people derive 80% of nutritional
energy from carbohydrates and 11% from fats. In consumption in
countries with high GDPs, carbohydrates are less important and
more energy is provided by fats. The average consumer in the US,
France and Denmark, for instance, derives 45–50 E% from
carbohydrates and 40 E% from fats. The figure also shows that
for countries with low GDPs, small changes in income cause large
changes in the composition of consumption, while for countries
with high GDPs, small income changes do not affect composition
because saturation has already occurred.
Although over the period 1961–2001 there were no average
annual per capita GDPs of southern Europe countries below 3500
and above 19,000 dollars, when comparing Figs. 4b–5a, results
show similar trends. Figs. 4c and 5b show the relationship between
the fraction of nutritional energy derived from animal sources (A%)
and GDP. For countries with low GDPs, A% is almost negligible; for
countries with high GDPs, the fraction is about 25–40%. In
Bangladesh, for example, A% is only 3%, while in Denmark A% is
40%. The figures show that for low GDPs, differences per unit of
GDP are large; for high GDPs, differences per unit GDP are much
smaller. It should be stressed, however, that A% indicates the
fraction of energy derived from animal foods and not the amount
consumed. Some countries with a high GDP, for example Canada
and the US, show relatively small consumption of animal foods, 27
and 30 E% respectively. In absolute numbers, however, for the
average Canadian the annual supply was 101 kg of meat and
204 kg of milk, while for a US citizen it was 121 kg of meat and
262 kg of milk. In an OECD country with a lower GDP than the US
and Canada, the Netherlands, the relative consumption of animal
foods is larger than in the US and Canada, 34 E%, but actual meat
supply is less with 90 kg per capita per year and milk supply larger
with 336 kg (FAO, 2010). The example also shows that consump-
tion of animal foods does not increase indefinitely. Jobse-van
Putten (1995), for instance, has also shown that in the western
world high income groups consume less meat that low income
groups. This trend has also been observed in Portugal (Rodrigues,
Caraher, Trichopoulou, & De Almeida, 2008). Meat and milk require
relatively large natural resource use (Wirsenius, 2003). Therefore
this result is important from an environmental perspective.
In general, animal protein is of better quality than plant-based
protein (Whitney & Rolfes, 1999). For most countries the protein E%
does not show great differences. An increase in the fraction of
nutritional energy derived from animal foods, therefore, does not
imply an increase in the fraction of protein, but rather an
improvement in protein quality. Especially for developing countries
this trend is important, because it improves the quality of the
consumption pattern.
Figure 5a and b also show that in some countries consumers
deviate from trends. In Japan, for example, a high GDP is combined
with a relatively small per capita food supply, while the
composition of consumption also resembles a pattern related to
a lower GDP. This indicates that factors other than GDP, such as
culture, also affect consumption.
Figure 6 shows the fat E% of per capita consumption derived
from 11 surveys in developing countries that make a distinction
between urban and rural patterns. Apart from Egypt, urban
consumption has a greater fat E% than rural consumption. The
surveys were done in countries with relatively low GDPs, i.e.
within the range where large differences in composition of food
consumption occur per unit of GDP. It is likely that per capita GDP
was higher for urban populations, which would explain the
difference in fat E%. The result confirms the other trends.
Trends
Results of the cross-sectional and time series relationships all
show that large changes in food supply and composition of
consumption occur for relatively low annual per capita GDPs,
below 5000 dollars, while for a GDP between 5000 and 12,5000
dollars changes are relatively small and above a GDP of 12,500
dollars food supply and the composition of consumption become
quite stable. This is in accordance with many detailed studies of
specific consumer groups, which have shown that increasing
societal affluence causes shifts in the consumption of specific foods
[()TD$FIG]
Fig. 5. (a)showsthe relationshipbetween annual percapita GDP andthe composition
of food consumption patterns in terms of the fraction of nutritional energy derived
from fat (fat E%), protein (protein E%) and carbohydrate (carbohydrate E%). (b) shows
the relationship between annual per capita GDP and the composition of food
consumption patterns in terms of the fraction of nutritional energy from animal
sources (A%); the solid line denotes the power-law function with income elasticity
0.52, R2 = 0.73. The relationships were based on data from 57 countries in 2001.
[()TD$FIG]
Fig. 6. Fraction of nutritional energy derived from fat (fat E%) for urban and rural
populations in nine countries based on data from 11 food surveys in developing
countries (see Appendix B).
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608602
and commodities. It is also in accordance with Sen (1981) the Nobel
prize winner who found that food shortages do not result from a lack
of food but from a lack of access to food. By simplifying per capita
consumption further than existing studies, we identify strong
similar trends and shows general patterns in nutritional changes.
WhenFigs.2 and 5a are compared,animportantresult isthat for low
income countries the increase in supply happens faster than the
change in composition. This is relevant for environmental studies.
An increase in low incomes means initially that people buy more of
the same foods. When this is the case the use of natural resources is
linear to supply in terms of food calories. Next, people shift towards
consumingmore fatsandanimal foods and this changecontinues for
longer. This requires the production of commodities that entail a
different and possibly increased use of natural resources.
Per capita income and nutritional energy intake
Most of the available food surveys used in this study (see
Appendix B) were done in countries with large differences in per
capita income. Sixteen surveys indicate nutritional energy intakes
between 2000 and 2500 kilocalories per capita per day, which is in
the range of actual physical requirements (Whitney & Rolfes,
1999). Eight surveys show intakes between 1700 and 2000
kilocalories, while five studies report intakes of over 2500
kilocalories per capita per day. We find no relationship between
nutritional energy intake and annual individual per capita GDP,
however. For the Netherlands and the US, for example, two
countries with a high GDP, energy intakes are between 2000 and
2500 kilocalories per capita per day. This is similar to intakes in
countries with low GDPs. Twelve food surveys make a distinction
between rural and urban consumption, but do not find substantial
differences between energy intakes. The general impression is that
in western countries per capita food consumption, expressed as
nutritional energy intake, has increased over the last few decades.
In science and anecdote it is often assumed that obesity is partly
driven by eating specific foods or by eating in general (De Graaf,
2006; Mela, 2006; Van den Bos & De Ridder, 2006), while most
people consider dieting to be a solution to the problem of obesity
(Polivy & Herman, 2006). We show that nutritional energy intake is
far more constant than food supply or the composition of
consumption.
The gap between supply and consumption
Results show that when income increases the availability of
food also increases (expressed as per capita supply in kilocalories),
but actual consumption remains the same. This means that with
increasing income, the size of the gap between actual per capita
consumption and supply grows. For low GDPs the study finds a
ratio of supply to actual consumption of about 1.0. For high GDPs,
however, this ratio is higher, about 1.8, which indicates that about
half of supply is not eaten at all. This gap is larger than estimates
showing that in 1995 total losses in the food chain in the US were
27% of total supply (Scott Kantor, Lipton, Manchester, & Oliveira,
1997). That study, however, excludes weight reductions that occur
when commodities are processed into final food products, so
producing estimates lower than our results. A possible explanation
is that affluent countries with a high GDP also have more
industrialized food industries with longer food chains that are
probably less efficient, than countries with low GDPs. Another
explanation might be that in countries with a low GDP, where food
is more scarce, people prepare and consume food more efficiently
and so generate fewer and smaller waste streams. The FAO food
balance sheets do not provide information to confirm this
hypothesis. The evaluation of the increasing gap between supply
and per capita consumption that coincides with increasing GDP
requires further research. The result is important for environmen-
tal sciences, because it means that rising incomes are accompanied
by a less efficient use of natural resources.
Uncertainty and inaccuracy of results
Four factors cause uncertainty and inaccuracy of results. These
are: (i) data quality; (ii) the use of average data; (iii) the use of
supply data and (iv) the use of inhomogeneous data. A fifth factor
adds to uncertainty in interpreting the results: (v) uncertainty in
spatio-temporal analogues.
Data quality
The first factor that contributes to uncertainty and inaccuracy is
data quality. This study derives data on food supply from FAO food
balance sheets (FAO, 2010), for which the FAO obtains data from
national datasets. However, data from different countries probably
are not of equal quality, as this depends on the degree of
development of national statistical organizations. Within coun-
tries, data quality varies between years. Major events, such as
political instability, or improvements in the methods of statistical
organizations affect data quality. Even in countries with high-
standard statistical organizations, different sources provide
different data. For the Netherlands in 2000, for example, per
capita butter consumption varies by a factor of three among
datasets. According to the FAO the Dutch consume 2.1 kg of butter
per capita per year (FAO, 2010), according to the Statistics
Netherlands (CBS) 3.3 kg (LEI-DLO/CBS, 2002) and Eurostat
estimates 6.8 kg (LEI-DLO/CBS, 2002). The FAO adjusts basic data
and estimation/imputation of the missing data is necessary in
order to maintain a certain degree of consistency, completeness
and reliability in the food balance sheets (FAO Statistics Division,
2008). Although for some countries data quality might be poor, the
FAO food balance sheets are the only source of information
available to perform an analysis of the type presented here.
The study derives data on GDP from Maddison (2003), who has
expressed GDP in 1990 International Geary-Khamis dollars. The
Geary-Khamis system is an aggregation method in which
international prices and a country’s Purchasing Power Parity,
depicting relative country price levels, are estimated from a system
of linear equations and expressed in G-K dollars (United Nations
Statistics Division, 2006). Individual country GDP values can be
substantially different depending on the PPP methodology used,
however. To compare cross-sectional and time series relationships
this study prefers to apply only one source of GDP data and
therefore uses the database of Maddison (2003), since it covers
both historical and recent global information.
Data on historical food supply are obtained from the historical
analysis of Fogel and Helmchen (2002). That study reported
nutritional energy intakes below physiological requirements. In
general, nutritional energy requirements are constant per unit of
body mass (Whitney & Rolfes, 1999). Average energy intakes below
the physiological requirement might be possible, though, if the fact
that three centuries ago people were smaller and had more
children, with less body mass, is taken into account.
The use of average data
The second factor that contributes to uncertainty and inaccu-
racy is the use of average data. Per capita data are derived from
information on a national level and are therefore average numbers.
In some countries, disparity in income distribution is large and
differences in food consumption occur among population groups.
These differences are not reflected in national data, which means
that the use of average data underestimates trends found here.
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608 603
The use of supply data
The third factor that contributes to uncertainty and inaccuracy
is the use of FAO supply data that exclude non-market production.
Sometimes people produce food outside the market, for example in
their gardens (Fernandes & Nair, 1986; Pallot & Nefedova, 2003).
The FAO food balance sheets give supply data on a national level
and do not take non-market production into account (FAO, 2010).
By thus excluding non-market production, we probably underes-
timate supply.
The composition of supply and consumption
Before food is available for consumption, commodities and
foods go through complete food chains, from farm to fork, in which
processes are used to produce the final foods. In all chain links and
in transportation between links losses occur. Neglecting losses and
excluding non-market production could have an effect on the
assessments. To analyze the use of supply rather than consumption
data in the analysis of trends in the composition of national food
supply, this study compares differences between the composition
of per capita consumption and related supply. Every commodity
has a specific composition in terms of macronutrients, such as the
fat E%. Results have already shown that the contribution of protein
to nutritional energy of consumption, the protein E%, is stable, but
that fat and carbohydrate E% vary, reflecting a difference in the
composition of consumption. We assume that a difference in fat E%
between per capita consumption and related supply reflects a
difference in composition. For the comparison the study uses
information on the fat E% obtained from 18 food surveys, marked
with an asterisk (*) in Appendix B. The study calculates the fat E% of
related supply using Eq. (2), deriving data from the FAO (2010).
Figure 7 shows that the fat E% of consumption and related supply
are similar, indicating that the macronutrient compositions are the
same. Eventual losses or non-market production do not cause shifts
in the composition. This justifies the use of FAO data for the analysis
of trends in the composition of national food supply.
The use of food survey data
The fourth factor for uncertainty and inaccuracy is that the
study derives information from food surveys that were probably all
performed in different ways, generating different types of
inaccuracies and uncertainties. For example, people tend to
underreport consumption (CBS, 1994; Kok et al., 1993). The way
food surveys have addressed this problem has probably varied
among surveys, causing diverse inaccuracies. The lowest value is
for the Philippines at 1800 kilocalories per capita per day and the
highest for Jordan, 3200 kilocalories per capita per day. These
examples show that methods used for surveys have probably
differed, generating under- and overestimations, an unquantified
inhomogeneity that the present analysis cannot compensate for.
The use of spatio-temporal analogues
An additional fifth factor of uncertainty in interpreting the
results concerns the use of results from cross-sectional analyses for
drawing temporal inferences. Interrelations between countries
and globalization impact on developments in the economy,
agricultural technology and cultural preferences in, especially,
developing countries in a way that may significantly deviate from
historical and present trends in developed countries. Illustrations
of this are the quantitative differences between the cross-sectional
results and the results from the long and medium-term
longitudinal analyses of European countries. Deviations found,
however, are restricted to quantitative results, qualitative findings
were found to be robust. Projections based on the identified
relations are therefore quantitatively uncertain.
Future changes
Although there are many uncertainties and despite the use of
rough estimates, differences among countries, developments in
time and differences between urban and rural populations, all
results show similar changes in direction. It is stressed, however,
that results obtained here cannot be taken at face value. They give
an indication of the direction of changes in food supply, the
composition of consumption and the contribution of animal foods
and of their magnitude. Combined with estimates of increases in
GDP, the study provides a tool to quantify these changes and
indicate where and when they will probably take place.
The most important finding of this paper is that the main
changes occur for per capita annual incomes below 12,500 dollars.
If trends found here are also valid for the future, this has important
consequences in the coming decade not only for food security, but
also for the use of natural resources such as arable land and
freshwater. Currently, about 85% of the world population lives in
six regions: (i) the OECD countries, (ii) Latin America, (iii) Africa,
(iv) China, (v) India and (vi) the rest of Asia. Table 1 shows the
nutrition, GDP and population characteristics of these regions.
In four regions per capita income levels are below 5000 dollars
per year, i.e. within the range where the largest changes occur.
China, India and the rest of Asia combine low GDPs with large
growth rates. This means that in the next 10 years considerable
changes are likely to occur in Asia. If the Asian countries maintain
economic growth along existing lines, the next decade might show
a substantial increase in per capita food supply, while the
composition of consumption might shift towards the affluent
patterns of western countries, characterized by substantial
consumption of fats and animal foods and limited consumption
of starchy staples. Latin America and Africa will probably see little
economic growth. There, population growth will be the main
driver in increasing total food demand. For the OECD no substantial
changes are likely, because food consumption in these countries
has already reached saturation level and population size is more or
less stable.
To estimate food demand for the period 2003–2030, the FAO
(2003) has indicated that developing regions will show a shift
[()TD$FIG]
0
10
20
30
40
0 10 403020
Fat E% food surveys
F
a
t
E
%
f
o
o
d
b
a
la
n
c
e
s
h
e
e
ts
Fig. 7. Comparison between the fraction of nutritional energy derived from fat (fat
E%) of actual consumption and of related supply. Data on fat E% of consumption
were derived from food surveys, the fat E% of related supply was calculated from
FAO food balance sheets.
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608604
towards increased food supply, as well as greater consumption of
specific commodities such as cereals, sugar, oils and animal foods,
while consumption of pulses, roots and tubers will decrease. When
information from our study is combined with estimates of GDP
growth, results can contribute to the scenario analysis of the FAO
and provide additional information on when and where changes
are likely to occur.
The impact on natural resources
Increased supply or a shift in consumption towards foods with
greater requirements for natural resources will have impacts on
the production system and the pressure on these natural resources.
The European transition towards affluent food consumption
patterns was a gradual process, taking place over centuries. The
economies developed step by step and agriculture could keep pace
with the growth in demand. Today, economic growth occurs at a
much faster rate (Maddison, 2003). Especially for developing
countries with a low GDP, this process contributes to pressure on
agriculture to produce sufficient food of a required quality in the
coming decade.
At present 38% of the global land area is in use for food
production (FAO, 2003) and sustainable options to increase this are
few. Moreover, in the last 10 years a global deceleration of yield
growth occurred (FAO, 2003). The pressure on freshwater is also
great. Humans already use 86% of freshwater, mainly for
agriculture. This water can for a large part be attributed to
the consumption of animal foods (Hoekstra & Chapagain, 2007).
The expected global increase in consumption of these foods,
therefore, will put additional pressure on the availability of
freshwater. While land and freshwater are mainly needed in
agriculture, energy requirements occur in all links of a food chain.
In the past decades energy requirements for food increased not
only due to consumption of different foods, but also due to the use
of other production and transportation methods (Gerbens-Leenes,
2006). It can therefore be expected that the general trends of
consumption found in this study will also have a considerable
impact on energy use. Knowledge of the impact of different food
items and categories on the use of resources provides a tool to
indicate pathways towards more sustainable consumption, for
example by increased efficiency, prevention or substitution.
Conclusions
The study confirms that throughout the world a nutrition
transition is taking place, in which people shift towards more
affluent food consumption patterns. This transition is taking place
at different stages and at different paces. The cross-sectional and
time series relationships show similar patterns of change. For low
income countries, an increase in per capita GDP is accompanied by
changes towards the food consumption patterns of western
countries, characterized by a large gap between supply and actual
consumption. Whereas actual consumption remains stable, total
supply (kilocalories per capita per day) differs by a factor of two
between low and high income countries. In this way economic
growth also causes a shift towards a more inefficient food system,
with greater use of natural resources. A second characteristic of
changes in consumption is the switch in the fraction of nutritional
energy from carbohydrates to fats and to animal foods, while the
protein fraction remains stable. People with low incomes derive
nutritional energy mainly from carbohydrates; the contribution of
fats to nutritional energy is small, that of protein the same as for
high incomes and that of animal sources negligible. People with
high incomes derive nutritional energy mainly from carbohydrates
and fats and the contribution of animal sources is substantial. In
general, whenever and wherever economic growth occurs, per
capita food supply and the composition of supply and consumption
show the same change of direction. The results of the study
are based on a simplified food system using rough estimates. In
reality, the food system is far more complex and there are
many factors, such as culture, that influence food consumption
patterns. By simplifying the system, the study shows general
trends that would not have been found in a more detailed analysis.
For specific situations, however, results might deviate from trends
found in here.
The importance for environmental studies is that results show
that the largest changes in food consumption patterns, and thus
the largest increase in the use of natural resources, occurs in the
range of incomes below 5000 dollars per year, i.e. in developing
countries. With an income of above 12,500 dollars saturation has
occurred and per capita use of natural resources for food does not
necessarily increase any further. In the coming 10 years large
changes in food consumption patterns are likely to occur in Asia
and especially in China and India, two countries that combine great
economic growth, low income levels and poor food consumption
patterns. The European transition occurred gradually, enabling
agriculture and trade to keep pace with the growth in demand.
Changes in economic circumstances change the demand for food.
A continuation of present economic trends might cause a
considerable pressure on the food system, because changes are
occurring much faster than they did in Europe and causing
additional pressure on finite natural resources.
Table 1
Per capita nutrition characteristics in 2001, GDP characteristics (dollars), expected national GDP growth and population characteristics for six regions (85% of the global
population).
Region Nutrition characteristics 2001a GDP characteristicsb Population characteristics
Energy
supplyc
Fat
E%
Protein
E%
Energy from
animal sources (%)
Annual per
capita GDP 2001
National
GDP growthd
Estimated annual
per capita GDP 2015
Size 2001
(billion)e
Annual
growthf
Size 2015
(billion)
China 2953 26 11 20 3800 8% 11,600 1.29 0.7% 1.42
India 2385 19 9 8 1926 6% 4,300 1.03 1.4% 1.25
OECD 3493 36 12 27 21538 2% 29,500 0.89 0.4% 0.94
Asiag 2540 18 9 9 2760 7% 7,000 0.76 1.3% 1.20
Africa 2519 18 10 7 1615 4% 2,900 0.52 2.6% 0.74
Latin America 2905 26 11 20 6174 2% 8,400 0.45 1.3% 0.54
a Source: FAO (2010).
b Source: Maddison (2003).
c Kilocalories per capita per day.
d Based on data from the International Monetary Fund, 2010IMF (2010) for 2001–2005.
e Source: FAO (2005).
f Source: FAO (2003).
g Without China and India.
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608 605
References
Boom-Binkhorst van der, F. H., Winkelman, M. L. J., Lith van, A., von Lossonczy von
Losoncz, T. O., Amesz, M. F., & Schure-Remijn, P. J. M. (1997). Voedingsmiddelen en
dranken in het Nederlandse voedingspatroon [Food items and beverages in the
Dutch food package]. In A. Stafleu, J. M. Veen, & W. H. Vredebregt-Lagas (Eds.), Mens
en Voeding (pp. 233–234). People and foodBaarn, the Netherlands: Nijgh en van
Ditmar.
Braun von, J. (1988). Effects of technological change in agriculture on food consump-
tion and nutrition. Rice in a West African Setting. World Development, 9, 1083–
1098.
Braun von, J., & Paulino, L. (1990). Food in sub-Saharan Africa, Trends and policy
challenges for the 1990s. Food Policy, 505–517.
Carlsson-Kanyama, A., Engström, R., & Kok, R. (2005). Indirect and direct energy
requirements of city households in Sweden. Options for reduction, lessons for
modeling. Journal of Industrial Ecology, 9(1–2), 221–235.
Catsberg, C. M. E., & Kempen-van Dommelen, 1997. The knowledge of Foods. Intro,
Baarn, the Netherland (in Dutch).
Centraal bureau voor de statistiek (CBS), hoofdafdeling statistieken van inkomen en
consumptie. (1994) [Central Bureau of statistics on income and consumption.
(1994)]. Gebruikershandboek budgetonderzoek 1992 [User manual expenditure sur-
vey 1992]. Voorburg, the Netherlands: CBS.
De Graaf, C. (2006). Effects of snacks on energy intake. An evolutionary perspective.
Appetite, 47, 18–23.
Duchin, F. (2005). Sustainable consumption of food. A framework for analyzing
scenarios about changes in diets. Journal of Industrial Ecology, 9(1–2), 99–
114.
Engelenburg van, B. C. W., Rossum van, T. F. M., Blok, K., & Vringer, K. (1994).
Calculating the energy requirements of household purchases, a practical step
by step method. Energy Policy, 22, 648–656.
Falkenmark, M. (1989). Water scarcity and food production. In D. Pimentel & C. W. Hall
(Eds.), Food and natural resources (pp. 164–191). San Diego, CA: Academic Press.
FAO. (2003). In J. Bruinsma (Ed.), World Agriculture. Towards 2015/2030. An FAO
Perspective. London: Earthscan Publications.
FAO. (2010). Food Balance Sheets. http://www.fao.org
FAO. (2008). The state of food and agriculture 2008. Biofuels. Prospects, risks and
opportunities. Rome.
Fernandes, E. C. M., & Nair, P. K. R. (1986). An evaluation of the structure and function of
tropical homegardens. Agricultural systems, 21, 279–310.
Fischer, G., Van Velthuizen, H., Shah, M., & Nachtergaele, F. O., 2002. Global Agro-
ecological Assessment for Agriculture in the 21st century: Methodology and Results.
International Institute for Applied Systems Analysis, Laxenburg, Austria/Food and
Agriculture Organization of the United Nations, Rome, Italy.
Fogel, R. W., & Helmchen, L. A. (2002). Economic and technological development and
their relationships to body size and productivity. In B. Caballero and B. M. Popkin
(Eds.), The nutrition transition, diet and disease in the developing world (pp. 9–24).
Food Science and Technology International Series. London, San Diego: Academic
Press.
Gerbens-Leenes, P. W. (2006). Natural resource use for food. Land, water and energy in
production and consumption systems. Thesis University of Groningen, Groningen,
the Netherlands.
Gerbens-Leenes, P. W., & Hoekstra, A. Y. (2007). The water footprint of the Netherlands.
Analysis of international impacts and options for change. CE&M research report
2007R-008/WEM-005. ISSN 1568-4652. Enschede, the Netherlands: University
of Twente.
Gerbens-Leenes, P. W., & Nonhebel, S. (2002). Consumption patterns and their effects
on land required for food. Ecological Economics, 42, 185–199.
Gerbens-Leenes, P. W., & Nonhebel, S. (2005). Food and land use. The influence of
consumption patterns on the use of agricultural resources. Appetite, 45, 21–31.
Grigg, D. (1995). The nutrition transition in Western Europe. Journal of Historical
Geography, 22(1), 247–261.
Hertwich, E. G. (2005). Consumption and industrial ecology. Journal of Industrial
Ecology, 9(1–2), 1–6.
Hoekstra, A. Y., & Chapagain, A. K. (2007). Water footprints of nations. Water use by
people as a function of their consumption pattern. Water Resource Management, 21,
35–48.
Hoekstra, A. Y., & Chapagain, A. K. (2008). Globalization of water. Sharing the planet’s
freshwater resources. London: Blackwell Publishing, London.
International Monetary Fund (IMF). (2010). World Economic Database, September
2004. http://www.imf.org/external/pubs/ft/weo/2004/02/data/
Ivens, W. P. M. F. , Dankert, G., Eng van der, P. J., Faber, D. C., van Keulen, H., Klaver, W.,
et al. (1992). World food production. Heerlen, the Netherlands: Open universiteit.
Jobse-van Putten, J. (1995). Eenvoudig Maar Voedzaam [Simple but nutritious]. Nijme-
gen/Amsterdam, the Netherlands: SUN/P.J. Meertens-Instituut.
Karlsson, R. (2001). Svinn i storhushåll. mängder, sammansättning och möjligheter att
minska [Losses in food service institutions. Options for reductions]. Master’s Thesis.
Stockholm, Sweden: Department of Systems Ecology, Stockholm University.
Kok, R., Biesiot, W., & Wilting, H. C. (1993). Energie-intensiteiten van voedingsmiddelen
[Energy intensities of food items]. IVEM-report 59. Groningen, the Netherlands:
Center for Energy and Environmental Studies (IVEM).
Kramer, K. J. (2000). Food matters. Thesis University of Groningen, the Netherlands.
Kramer, K. J., & Moll, H. C. (1995). Energie voedt, nadere analyses van het indirecte
energieverbruik van voedingsmiddelen [Energy feeds, analysis of indirect energy use of
food items]. IVEM-report 77. Groningen, the Netherlands: Center for Energy and
Environmental Studies (IVEM).
Landbouw-Economisch Instituut (LEI-DLO), Centraal Bureau voor de Statistiek (CBS)
[Agricultural Economic Institute (LEI-DLO), Central Bureau of Statistics (CBS)].
(2002). Landbouwcijfers 2002 [Agricultural data 2002]. ‘s-Gravenhage, the Nether-
lands: LEI/CBS.
Lang, T., 2002. Can the Challenges of poverty, sustainable consumption and good health
governance be addressed in an era of globalization? In B. Caballero and B. M. Popkin
(Eds.), The nutrition transition. diet and disease in the developing world (pp. 51–71).
Food Science and Technology International Series. London: Academic Press.
Latham, J. R. (2000). There’s enough food for everyone, but the poor can’t afford to buy
it. Nature, 404, 222.
Maddison, S. (2003). The world economy. Historical statistics. Paris, France: OECD.
Mela, P. J. (2006). Eating for pleasure or just wanting to eat. Reconsidering sensory
hedonic responses as a driver for obesity. Appetite, 47, 3–9.
Mennell, S., Murcott, A., & Otterloo van, H. A. (1992). Sociology of food, eating diet and
culture. London: International Sociological Association SAGE Publications.
Millenium Ecosystems Assesment, 2005. Millennium Ecosystem Assessment Synthesis
Report. Pre-publication Final Draft Approved by MA Board on March 23, 2005.
Coordinated by the United Nations Environment Programme.
Musaiger, A. O. (1989). Changes in food consumption patterns in Bahrein. Nutrition and
Health, 6, 183–188.
Nonhebel, S. (2004). On resource use in food production systems. The value of
livestock as ‘rest-stream upgrading system’. Ecological Economics, 48(2), 221–230.
Pallot, J., & Nefedova, T. (2003). Trajectories in people’s farming in Moscow oblast
during the post-socialist transformation. Journal of Rural Studies, 19, 345–362.
Penning de Vries, F. W. T., Jansen, D. M., Ten Berge, H. F. M., & Bakema, A. L. (1989).
Simulation of ecophysiological processes of growth in several annual crops (pp. 63–
64). Wageningen, the Netherlands: Centre for Agricultural Publishing and Docu-
mentation (Pudoc).
Penning de Vries, F. W. T., Van Keulen, H., & Rabbinge, R. (1995). Natural resources and
limits of food production in 2040. In J. Bouma (Ed.), Eco-regional approaches for
sustainable land use and food production. The Netherlands: Kluwer Academic
Publishers.
Pindyck, R. S., & Rubinfeld, D. L. (2005). Microeconomics (sixth edition). Upper Saddle
River, New Yersey, 07458, USA: Pearson Education, Inc.
Poleman, Th. T. , & Thomas, L. T. (1995). Income and dietary change. Food Policy, 20(2),
149–159.
Polivy, J., & Herman, C. P. (2006). An evolutionary perspective on dieting. Appetite, 47,
30–35.
Popkin, B. M. (2002). The dynamics of the dietary transition in the developing world. In
B. Caballero and B. M. Popkin (Eds.), The nutrition transition. diet and disease in the
developing world (pp. 111–129). Food Science and Technology International Series.
London: Academic Press.
Receveur, O., Boulay, M., & Kuhnlein, H. V. (1997). Decreasing traditional food use
affects diet quality for adult dene/métis in 16 communities of the Canadian
Northwest territories. Journal of Nutrition, 127, 2179–2186.
Rockstrom, J. (1999). On-farm green water estimates as a tool for increased food
production in water scarce regions. Physics and Chemistry of the Earth, Part B, 24(4),
375–383.
Rodrigues, S. S. P., Caraher, M., Trichopoulou, A., & De Almeida, M. D. V. (2008). Portuguese
households’ diet quality (adherence to Mediterranean food pattern and compliance
with WHO population dietary goals). Trends, regional disparities and socioeconomic
determinants. European Journal of Clinical Nutrition, 62, 1263–1272.
Rosegrant, M. W., & Ringler, C. (1998). Impact of food security and rural development of
transferring water out of agriculture. Water Policy, 1, 567–586.
Schmidt-Nielsen, K. (1988). Animal physiology. Adaptation and environment (third ed.).
New York: Press Syndicate of the University of Cambridge.
Scott Kantor, L., Lipton, K., Manchester, A., & Oliveira, V. (1997). Estimating and
addressing America’s Food Losses. Food review, 20(1), 2–12.
Sen, A. (1981). Poverty and famines. An essay on entitlement and deprivation. Oxford
Clarendon Press.
Tukker, A., & Jansen, B. (2006). Environmental impacts of products. A detailed review of
studies. Journal of Industrial Ecology, 10(3), 159–182.
United Nations Statistics Division. (2006). http://unstats.un.org/unsd/methods/icp/
gdp/gdp06_htm.htm
United States Department of Agriculture. (USDA) Agricultural Research Service 2005.
Food and Nutrient Database for Dietary Studies, 1.0. http://www.ars.usda.gov/
Services/docs.htm?docid=7637
Van den Bos, R., & De Ridder, D. (2006). Evolved to satisfy our immediate needs. Self
control and the rewarding properties of food. Appetite, 47, 24–29.
Voedingscentrum. (1998a). In B. C. Breedveld, J. Hammink, & J. H. M. van Oosten (Eds.),
Nederlandse Voedingsmiddelentabel. Dutch food itemsDen Haag, the Netherlands:
Voedingscentrum[[nl]]Food center.
Voedingscentrum, & TNO. (1998). Zo eet Nederland 1998 [This is how the Netherlands eats
1998]. Den Haag, the Netherlands: Voedingscentrum [Food Centre].
Vringer, K., & Blok, K. (1995). The direct and indirect energy requirements of house-
holds in the Netherlands. Energy Policy, 23(10), 893–910.
Wandel, M. (1988). Household food consumption and seasonal variations in food
availability in Sri Lanka. Ecology of Food and Nutrition, 22, 169–182.
Whitney, E. N., & Rolfes, S. R. (1999). Understanding nutrition (8th ed., pp. 3–40).
Belmont, USA: Wadsworth Publishing Company.
Wijn de, J. F., & Weits, J. (1971). Steensma’s Voedingsleer. Steensma’s Food Science].
Amsterdam/Haarlem, the Netherlands: Scheltema en Holkema.
Wirsenius, S. (2003). The biomass metabolism of the food system. A model-based
survey of the global and regional turnover of food biomass. Journal of Industrial
Ecology, 7(1), 47–80.
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608606
http://www.fao.org/
http://www.imf.org/external/pubs/ft/weo/2004/02/data/
http://unstats.un.org/unsd/methods/icp/gdp/gdp06_htm.htm
http://unstats.un.org/unsd/methods/icp/gdp/gdp06_htm.htm
http://www.ars.usda.gov/Services/docs.htm%3Fdocid=7637
http://www.ars.usda.gov/Services/docs.htm%3Fdocid=7637
World Bank. (2005). East Asia’s Dollar Influx-Signal for Change. Special focus. Gender
Equality in East Asia. https://publications.worldbank.org/subscriptions/WDI
WWF, 2007. Allocating scarce water. A WWF primer on water allocation, water rights and
watr markets. WWF-UK, Godalming, UK.
Appendix A
Overview of the 57 countries for which this study performed
the cross-sectional analysis.
Africa: Algeria, the Democratic Republic of Congo, Côte d’ Ivoire,
Egypt, Ethiopia, Ghana, Kenya, Morocco, Nigeria, Sudan, Tanzania,
and South Africa
Asia: Bangladesh, China, India, Indonesia, Malaysia, Pakistan,
the Philippines, Sri Lanka, Thailand, and Vietnam
Eastern Europe: Poland
Latin America: Argentina, Brazil, Chile, Colombia, Ecuador,
Guatemala, Mexico, Peru, and Venezuela
Middle East: Israel, and Syria
OECD: Austria, Belgium, Canada, Denmark, Finland, France,
Germany, Greece, Iceland, Ireland, Italy, Japan, the Netherlands,
Portugal, Spain, Sweden, Turkey, United Kingdom, United States
Additional, small countries: The United Arab Emirates, Estonia,
Slovenia.
Appendix B
Overview of countries and national food surveys used in this
study. The 18 food surveys that provide data on fat E% are marked
with an asterisk *.
Argentina
Britos, S., & Scacchia, S. (1998). Disponibilidad y consumo de
alimentos en Argentina. Escuela de Nutrición [Food availability and
consumption in Argentina. School of Nutrition]. Argentina: Uni-
versidad Nacional de Buenos Aires [National University of Buenos
Aires].
Bangladesh*
Jahan, & Hossein. (1998). Malnutrition in Bangladesh: Bangladesh
National Nutrition Survey, 1995–96. Bangladesh: Institute of
Nutrition and Food Science, Dhaka University.
Brazil
Galeazzi, M. A. M., & Falchoni Jr., P. (1998). Inquérito de Consumo
Alimentar da Area Metropolitana de Brası́la-Relatório [Nutrition survey
in the area of Brasilia-Relatorio]. Brası́lia: Técnico-Secretaria de Saúde
de Brası́lia.
Cambodja
National Institute of Statistics (NIS/MOP). (1996). Socio-
Economic Survey of Cambodja. Data from the Multi-Indicator Cluster
Survey (MICS) of the Socio-Economic Survey of Cambodja (SESC)
sponsored by the Asian Development Bank in collaboration with the
UNICEF/UNDP/CARERE and ILO. Cambodja: Royal Government of
Cambodja.
China*
Ge, K., Zhai, F., & Yan, H. (1996). Institute of Nutrition and Food
Hygiene (INFH) 1985. Summary Report of the 2nd National Nutrition
Survey in 1982. Beijing, China: Institute of Nutrition and Food
Hygiene.
Ge, K., Zhai, F., & Yan, H. (1996). The dietary and nutritional
status of Chinese population. 3rd National Nutrition Survey, 1992.
Beijing, China: People’s Med. Pub. House.
Colombia
Ministerio de Agricultura DANE-DRI-PAN. (1984). Encuesta
Nacional de Alimentatión, Nutrición y Vivienda DANE-PAN-DRI 1981
[Ministry of Agriculture DANE-DRI-PAN 1984. National Feeding,
Nutrition and Housing Survey DANE-PAN-DRI 1981]. Bogotá:
Franza Pardo T-Bogotá (Mimeógrafo).
Costa Rica
Ministerio de Salud. (1996). Ministerio de Salud 1996. Encuesta
Nacional de Nutrición. Fasciculo No 1: Consumo Aparente [Ministry of
Health 1996. National Nutrition Survey. Fascicle No 1: Apparent
Consumption]. San José, Costa Rica.
Egypt*
Hassanyn, A. S. (2000). Food Consumption Pattern and Nutrients
Intake Among Different Population Groups in Egypt. Final Report
(Part 1). Egypt: Nutrition Institute, WHO/EMRO.
El Salvador
Asociación Demográfica Salvadorĕna (ADS), Ministerio de Salud
Pública y Asistencia Social (MSPAS), & Instituto de Nutrición de
Centro América y Panamá (INCAP) [Salvadoran Demographic
Association, Ministry of Public Health and Social Assistance, &
Institute of Nutrition of Central America and Panama (INCAP)].
(1990). Evaluación de la Situación Alimentaria Nutricional en El
Salvador [Evaluation of the nutritional situation in El Salvador]. El
Salvador: ESA NES-88.
Equador
Freire, W. (1988). Diagnóstico de la situación alimentaria y
nutricional y de salud de la población ecuatoriana menor de cinco años
– DANS -1986 [Diagnosis of the alimentary, nutritional and health
state of the Ecuadorian population less than five years – DANS -1986].
Quito, Equador: CONADE, MSP.
Iran*
Djazayery, A., & Samimi, B. (1996). (Surveys for 1983 and 1992)
Food consumption and energy intake patterns in the rural and
urban areas of Iran, 1983–1992. Agricultural Economics and
Development, 4, 218–248.
Jamaica
Simeon, D. T., & Patterson, A. W. (1994). Energy and protein
accessibility at the household level in Jamaica: Results from a national
survey 1989. Jamaica: CFNI.
Jordan*
Department of Statistics (DOS). (1997). Household Income and
Expenditure Survey. Amman, Jordan.
Madagaskar
FAO. (2004). L’état de l’insécurité alimentaire 2001 dans le monde
[The state of food insecurity in the world]. Rome, Italy: Organisation
des Nations Unies pour l’alimentation et l’agriculture [Food and
Agriculture Organisation of the United Nations], http://
www.fao.org.
Mali*
FAO. (2005). Profiles nutritionnels par pays [Nutritional profiles
per country]. Mali: Departement Economique et Social, Alimenta-
tion et nutrition [Department of Economic and social affairs, food
and nutrition]. http://www.fao.org/es/nutrition/mli-f.stm.
Mexico*
INNSZ. (1990). Encuesta Nacional de Alimentación en el Medio
Rural ENAL 1989 [National Feeding Survey in Rural Areas ENAL 1989].
México: INCMNSZ.
Avila, A., Shamah, T., & Chavez, A. (1997). Enquestas de
Alimentación y Nutrición en el Medio Rural, 1996. Resultados
por entidad [Feeding and Nutrition Surveys in Rural Areas,
1996. Results by organization]. INNSZ, DEDESOL, DIF, SSA, Golernos
de los Estados [Governments of the States]. Mexico: IMSS, INI, Unicef.
The Netherlands*
Voedingscentrum [Food Center], & TNO. (1998). Zo eet Neder-
land 1998 [This is how the Netherlands eats 1998]. Den Haag, the
Netherlands: Voedingscentrum [Food Center].
Panama
Ministerio de Salud. (1992). Ministerio de Salud 1992. Encuesta
Nacional de Consumo de Alimentos. Panamá: Departamento de
Nutrición y Dietética Panamá [Ministry of Health 1992. National
Food Consumption Survey. Panamá: Dietetic and Nutrition Depart-
ment Panamá].
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608 607
https://publications.worldbank.org/subscriptions/WDI
http://www.fao.org/
http://www.fao.org/
http://www.fao.org/es/nutrition/mli-f.stm
Instituto de Nutrición de Centro América y Panamá (INCAP),
Oficina de Investigaciones Internationales de Salud, & Ministerio
de Salud Pública y Asistencia Social (MSPAS) [Institute of Nutrition
of Central America and Panama (INCAP), International Health
Research Office, & Ministry of Public Health and Social Assistance
(MSPAS)]. (2000). Evaluación nutricional de El Salvador 1969
[Nutritional Assessment of El Salvador 1969].
Peru
Amat, C., & Curonisy, P. (1981). La alimentación en el Perú
[Feeding in Peru]. Lima, Perú: Centro de Investigación [Research
Center University of the Pacific].
Philippines*
Food and Nutrition Research Institute of the Department of
Science and Technology (FNRI-DOST) of the Philippines. (2000).
National Survey of 1993: Final Results.
Sri Lanka*
Department of Census and Statistics. (1993). Household Income
and Expenditure Survey 1990/91, Final Report. Department of Census
and Statistics. Sri Lanka: Ministry of Policy Planning and
Implementation.
Turkey*
Hundd, & Moh. (1997). Food consumption survey in 7 provinces,
Project Report. Ankara, Turkey: Hacettepe University, Department
of Nutrition and Dietetics, Ministry of Health.
United States*
United States Department of Agriculture (USDA) Agricultural
Research Service. (2005). Food and Nutrient Database for Dietary
Studies, 1.0. http://www.ars.usda.gov/Services/docs.htm?docid=
7637.
Venezuela*
Luna Bazó, P., & Bracho, M. (1987). Encuesta Nacional de
Nutrición. Area Socio Alimentaria ‘‘Encuesta de Consumo’’. Mimeo-
grafiado [National Nutrition Survey. Socio Alimentary Field ‘‘Con-
sumption Survey’’. Mimeografiado]. Caracas, Venezuela: Instituto
Nacional de Nutrición, Direccion Téchnica [National Nutrition
Institute, Technical Direction].
Vietnam*
Tu Giay, & Chu Quoc Lap. (1990). Final report on the subject
64D.01.01 of the National Research Programme National General
Survey 1989. Hanoi, Vietnam: The Governmental Science and
Technology Committee, NIN.
National Institute of Nutrition (NIN). (1995). Sentinel food and
nutrition surveillance system data. Hanoi, Vietnam: NIN.
Zimbabwe
Bursztijn, P. G. (1985). A diet survey in Zimbabwe. Human Nutr.
Appl. Nutr. 39 (5), 376–388.
P.W. Gerbens-Leenes et al. / Appetite 55 (2010) 597–608608
http://www.ars.usda.gov/Services/docs.htm%3Fdocid=7637
http://www.ars.usda.gov/Services/docs.htm%3Fdocid=7637
Administrator
20005418coverv05b
A Macroeconomics Reader
A Macroeconomics Reader brings together in a single volume a collection of
key articles on developments and current debates in modern
macroeconomics. The articles have been selected to provide the reader with
an accessible and informative overview of a number of the most significant
issues and controversies in macroeconomics.
Areas covered include:
Keynesian Economics and the Keynesian Revolution
The Monetarist Counter-Revolution
The Challenge of Rational Expectations and New Classical Macroeconomics
The Real Business Cycle Approach to Economic Fluctuations
New Keynesian Economics
The Renaissance of Economic Growth Analysis
Brian Snowdon is Principal Lecturer in Economics at the University of
Northumbria. His previous books include Markets, Intervention and
Planning (1987) and (with Howard Vane and P.Wynarczyk) A Modern Guide
to Macroeconomics (1994).
Howard R.Vane is Reader in Economics at Liverpool John Moores
University. His previous books (with J.Thompson) include Current
Controversies in Macroeconomics (1992) and An Introduction to
Macroeconomic Policy (1993).
A Macroeconomics Reader
Edited by
Brian Snowdon and
Howard R.Vane
London and New York
First published 1997
by Routledge
11 New Fetter Lane, London EC4P 4EE
Simultaneously published in the USA and Canada
by Routledge
29 West 35th Street, New York, NY 10001
Routledge is an imprint of the Taylor & Francis Group
This edition published in the Taylor & Francis e-Library, 2003.
© 1997 BRIAN SNOWDON and HOWARD R.VANE
All rights reserved. No part of this book may be reprinted or
reproduced or utilized in any form or by any electronic,
mechanical, or other means, now known or hereafter
invented, including photocopying and recording, or in any
information storage or retrieval system, without permission in
writing from the publishers.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British
Library
Library of Congress Cataloging in Publication Data
A catalog record for this book has been requested
ISBN 0-203-44396-9 Master e-book ISBN
ISBN 0-203-27587-X (OEB Format)
ISBN 0-415-15715-3 (hbk)
0-415-15716-1 (pbk)
To Jean and Christine
Contents
Preface xi
1 The development of modern macroeconomics: a rough guide
Brian Snowdon and Howard R.Vane 1
Part I Keynesian economics and the Keynesian revolution 27
Introduction 29
2 Keynesian economics: the search for first principles
Alan Coddington
Journal of Economic Literature (1976) 14, December,
pp. 1258–73 36
3 On different interpretations of the General Theory
Don Patinkin
Journal of Monetary Economics (1990) 26, October, pp. 205–43 55
4 Keynes’s General Theory: interpreting the interpretations
Bill Gerrard
Economic Journal (1991) 101, March, pp. 276–87 95
5 The fall and rise of Keynesian economics
Alan S.Blinder
Economic Record (1988) December, pp. 278–94 109
6 Price flexibility and output stability: an old Keynesian view
James Tobin
Journal of Economic Perspectives (1993) 7, Winter, pp. 45–65 135
Part II The monetarist counter-revolution 157
Introduction 159
7 The role of monetary policy
Milton Friedman
American Economic Review (1968) 58, March, pp. 1–17 164
viii Contents
8 The structure of monetarism
Thomas Mayer
Kredit und Kapital (1975) 8, pp. 191–215, 292–313 180
9 Monetarism: an interpretation and an assessment
David Laidler
Economic Journal (1981) 91, March, pp. 1–28 216
10 The monetarist controversy revisited
Franco Modigliani
Contemporary Policy Issues (1988) 6, October, pp. 3–18 247
Part III The challenge of rational expectations and new classical
macroeconomics 263
Introduction 265
11 After Keynesian macroeconomics
Robert E.Lucas and Thomas J.Sargent
After the Phillips Curve: Persistence of High Inflation
and High Unemployment (1978) Boston, MA: Federal
Reserve Bank of Boston, pp. 49–72 270
12 A child’s guide to rational expectations
Rodney Maddock and Michael Carter
Journal of Economic Literature (1982) 20, March, pp. 39–51 295
13 The Ricardian approach to budget deficits
Robert J.Barro
Journal of Economic Perspectives (1989) 3, Spring, pp. 37–54 314
14 The new-classical contribution to macroeconomics
David Laidler
Banca Nazionale Del Lavoro Quarterly Review (1986)
March, p. 27–55 334
Part IV The real business cycle approach to economic fluctuations 359
Introduction 361
15 Theory ahead of business cycle measurement
Edward C.Prescott
Federal Reserve Bank of Minneapolis Quarterly Review
(1986) Fall, pp. 9–22 366
Contents ix
16 Some skeptical observations on real business cycle theory
Lawrence H.Summers
Federal Reserve Bank of Minneapolis Quarterly Review
(1986) Fall, pp. 23–7 389
17 Understanding real business cycles
Charles I.Plosser
Journal of Economic Perspectives (1989) 3, Summer,
pp. 51–77 396
18 Real business cycles: a new Keynesian perspective
N.Gregory Mankiw
Journal of Economic Perspectives (1989) 3, Summer, pp. 79–90 425
Part V New Keynesian economics 437
Introduction 439
19 The reincarnation of Keynesian economics
N.Gregory Mankiw
European Economic Review (1992) 36, April, pp. 559–65 445
20 New-Keynesian economics today: the empire strikes back
Brian Snowdon and Howard Vane
American Economist (1995) 39, Spring, pp. 48–65 452
21 What is new-Keynesian economics?
Robert J.Gordon
Journal of Economic Literature (1990) 28, September,
pp. 1115–71 478
22 New and old Keynesians
Bruce Greenwald and Joseph Stiglitz
Journal of Economic Perspectives (1993) 7, Winter, pp. 23–44 552
Part VI The renaissance of economic growth analysis 575
Introduction 577
23 Catching up, forging ahead, and falling behind
Moses Abramovitz
Journal of Economic History (1986) 46, June, pp. 385–406 582
24 Technological catch up and diverging incomes:
patterns of economic growth 1960–88
Steve Dowrick
Economic Journal (1992) 102, May, pp. 600–10 604
x Contents
25 Policy implications of endogenous growth theory
G.K.Shaw
Economic Journal (1992) 102, May, pp. 611–21 616
26 The origins of endogenous growth
Paul M.Romer
Journal of Economic Perspectives (1994) 8, Winter, pp. 3–22 628
27 Perspectives on growth theory
Robert M.Solow
Journal of Economic Perspectives (1994) 8, Winter, pp. 45–54 649
Index 660
Preface
Anyone who teaches ‘intermediate undergraduate’ macroeconomics can
vouch that there now exist a large number of excellent textbooks they can
choose from and direct their students to buy. This was certainly not the case
when we were undergraduate and graduate students in the late 1960s and
early 1970s. However, when it comes to recommending journal articles for
students to read, two main problems arise. First, many important and/or
seminal papers display a high degree of technical virtuosity and
mathematical sophistication which most students find extremely demanding,
to say the least. Second, students often experience difficulty in gaining access
to popular articles which are inevitably in heavy demand.
The present volume seeks to alleviate these two problems. The main aim
of the book is to bring together in a single volume a collection of insightful
articles, for intermediate undergraduates, which shed light on the
development of, and selected important controversies within, modern
macroeconomics. As such the book will serve as a supplementary text to be
read alongside a main macroeconomics textbook. The articles which make
up this book of readings have been chosen to provide the reader with
accessible and predominantly non-technical, reflective papers which
critically assess and/or survey important areas/issues in the development of,
and selected controversies within, modern macroeconomics. Choosing the
final collection of twenty-six articles which make up this volume proved to
be a more difficult task than we first anticipated (every economist would no
doubt select their own favourite collection). One important criterion for
selection has been the need to keep an ever watchful eye on production costs
so that the book could be published at a price which is affordable to students.
Our original proposal, which contained a collection of thirty-eight articles,
would have been prohibitively expensive to produce, especially given the
disproportionately high copyright permission fees levied by certain journals.
The final selection of twenty-six articles chosen for this volume we believe
are both thought provoking and insightful.
The book follows a structured direction, tracing the origins and
development of modern macroeconomics in historical perspective around the
main schools of macroeconomic thought through to the focus of much
xii Preface
current debate and research, namely the issue of economic growth. In
following this structured pattern we have consciously not included readings
on such areas as, for example, consumption, investment, the demand for and
supply of money, and open economy macroeconomics; topics which have
previously occupied a significant proportion of macroeconomics readers.
After an introductory chapter in which we present for the student reader a
selective review of some of the most important developments in
macroeconomics since the mid-1930s, the collection of articles is divided into
six main parts covering
• Keynesian Economics and the Keynesian Revolution
• The Monetarist Counter-Revolution
• The Challenge of Rational Expectations and New Classical
Macroeconomics
• The Real Business Cycle Approach to Economic Fluctuations
• New Keynesian Economics
• The Renaissance of Economic Growth Analysis
Each part starts with a brief introduction in which we place the articles
included in an overall context. Our intention is not to comment in detail
upon the central points raised in the articles concerned, merely to provide a
frame of reference for the reader. In the introduction to each part we have
included references which are particularly recommended for additional
reading, together with some essay-style questions for review. Having
completed such additional reading the reader will be in a much better
position to make an informed answer to these and other questions in
macroeconomics.
By bringing together in a single volume a collection of key supplementary
readings we hope the book will attract a wide readership among
intermediate undergraduates, as well as postgraduates and teachers, in the
field of macroeconomics and the history of economic thought.
We wish to express our gratitude to all the authors, and journals, who
have given their permission to reprint the collection of articles in this
volume. Our final thanks to Katherine Eade, Kerry Douglas, Angie Dell and
Sue Barlow for their patience, humour and co-operation in typing various
parts of the final manuscript. Any remaining errors (few in number, it is
hoped) are our responsibility.
Brian Snowdon
Howard R.Vane
1 The development of modern
macroeconomics
A rough guide
Brian Snowdon and Howard R.Vane
Any economics student who graduated from university in the late 1960s and
early 1970s, as we both did, would have found macroeconomics a much
easier and far less controversial subject to study then than it is today. Since
the breakdown of the Keynesian consensus in the early 1970s,
macroeconomics has been in a ‘state of disarray’ (Brunner 1989) having
witnessed the appearance of a number of conflicting and competing
approaches. As a result modern macroeconomics is a rapidly changing
diverse subject with a built-in tendency to generate deep divisions. These
divisions have led to the formation of schools of thought consisting of
economists who share a broad vision of how macroeconomic phenomena are
generated. In order to better understand current controversies it is necessary,
in our view, to know how macroeconomic thought has developed since
Keynes’s General Theory (1936) was published. In this opening chapter we
briefly survey some of the important developments in the evolution of
macroeconomics since the mid-1930s. Our purpose is not to critically assess
in detail the central tenets underlying and policy implications of the main
macroeconomic schools of thought, rather it is to provide a background
discussion in order to help place the readings that follow in context (for a
more detailed survey of competing schools of thought see Phelps 1990;
Chrystal and Price 1994; Snowdon et al. 1994).
Although there are significant differences between the various schools of
thought, the work of Keynes remains a central point of reference because, as
Vercelli (1991) argues, all the schools define themselves in relation to the
ideas originally put forward by Keynes in his General Theory, either as a
development of some version of his thought or as a restoration of some
version of pre-Keynesian classical thought. A unifying theme in the evolution
of modern macroeconomics has been an ‘ever-evolving classical Keynesian
debate’ (see Gerrard 1996). Although elsewhere (Snowdon et al., 1994) we
have identified seven schools of thought which have been influential in the
development of macroeconomic analysis since the mid-1930s, each of these
schools can be viewed as adhering to one of two basic positions in terms of
broad vision. Gregory Mankiw (1989)—reprinted in Part IV—describes these
2 The development of modern macroeconomics
two positions by distinguishing between the classical school and the
Keynesian school with respect to their faith in the ‘invisible hand’.
The classical school emphasizes the optimization of private economic
actors, the adjustment of relative prices to equate supply and demand, and
the efficiency of unfettered markets. The Keynesian school believes that
understanding economic fluctuations requires not just studying the
intricacies of general equilibrium, but also appreciating the possibility of
market failure on a grand scale.
Hence, although it is possible to distinguish between orthodox Keynesians,
new Keynesians and post-Keynesians, all three groups are united in the
belief that aggregate economic instability represents ‘some sort of market
failure on a grand scale’ (Mankiw 1990). In contrast the majority of
economists who have been prominent in the monetarist, new classical, real
business cycle and Austrian schools have tended to place their faith in
market forces as an equilibrating mechanism and question the capacity and
desirability of government intervention as a means of achieving the major
macroeconomic objectives. Following Gerrard (1996) the seven schools of
thought identified above can also be differentiated and classified as
o r t h o d o x , n e w o r r a d i c a l . T h e t w o ‘ o r t h o d o x ’ s c h o o l s , ‘ I S – L M
Keynesianism’ and ‘neoclassical monetarism’, dominated macroeconomic
theory in the period up to the mid-1970s. Since then three ‘new’ schools
have been highly influential. The new classical, real business cycle and
new Keynesian schools place emphasis on issues relating to aggregate
supply in contrast to the orthodox schools which focused their research
primarily on the factors determining aggregate demand and the
consequences of demand-management policies. In particular the new
schools share the view that macroeconomic models should be based on
solid microeconomic foundations. The ‘radical’ post-Keynesian and
Austrian schools are both critical of mainstream analysis whether it be
orthodox or new. Since modern macroeconomics has been most influenced
by the orthodox and new schools we will confine our discussion here to
their contributions (see Davidson 1994 and Chick 1995 for a discussion of
post-Keynesian macroeconomics; Garrison 1994 presents the case for the
Austrian approach).
In the final section of this opening chapter we briefly review the
renaissance of economic growth analysis which since the mid-1980s has
moved into the centre stage of macroeconomic research after twenty years of
relative neglect. The consequences of economic growth for economic welfare
are so important that many prominent macroeconomists who previously
concentrated their research efforts on the analysis of business cycles have
now turned their attention to theoretical and empirical issues arising out of
the burgeoning endogenous growth literature (see Barro and Sala-i-Martin
1995; Mankiw 1995).
The development of modern macroeconomics 3
KEYNESIAN ECONOMICS AND THE KEYNESIAN REVOLUTION
The birth of modern macroeconomics can be traced back to the 1930s, and
in particular the publication of John Maynard Keynes’s (1936) General
Theory of Employment, Interest and Money. Prior to the 1930s the dominant
view, in what we now call macroeconomics, was the classical approach that
within capitalist market economies which are subject to periodic shocks the
market mechanism would operate quickly and efficiently to restore full
employment equilibrium. In such circumstances government intervention to
stabilize the economy was believed to be neither necessary nor desirable.
However, the experience of the 1920s and 1930s in Britain, and that of all
major capitalist market economies during the 1930s, appeared to shatter the
classical assumption that full employment was the normal state of affairs. In
Britain the rate of unemployment never fell below 10 per cent between 1921
and 1938, and actually exceeded 20 per cent in 1931 and 1932. In the United
States unemployment reached a peak of 25 per cent in 1933 and was still
almost 10 per cent in 1941. Writing against this background Keynes (1936)
put forward a new and revolutionary theory to explain, and provide a
remedy for, the then-prevailing persistent and severe unemployment. In doing
so Keynes was responding to what undoubtedly was the most significant
macroeconomic event of the twentieth century; the Great Depression gave
birth to modern macroeconomics.
The causes of the Great Depression are still the subject of considerable
dispute and finding a plausible explanation for the global economic
c o l l a p s e d u r i n g t h e e a r l y 1 9 3 0 s r e m a i n s t h e ‘ H o l y G r a i l ’ o f
macroeconomics (seeC. D.Romer 1993; Bernanke 1995). In an extensive
study of the US experience Gordon and Wilcox (1981) concluded that the
cause of the Great Depression can be traced to a series of domestic
spending shocks, both monetary and non-monetary. The initial decline in
output during the 1929–31 period can be traced to a decline in
consumption and residential investment expenditures. After September 1931
the recession was turned into the Great Depression by the perverse actions
of the Federal Reserve in letting the money supply decline drastically (see
Friedman and Schwartz 1963). Hence the Great Depression resulted from a
shift of the aggregate demand curve to the left and the impact of the
monetary contraction was transmitted worldwide via the operation of the
gold standard (see Eichengreen 1992). The self-equilibrating tendencies of
the market failed to come into play and as a result we have the best known
example of massive monetary non-neutrality. In the face of enormous
unemployment, nominal wages failed to adjust sufficiently to shift the
aggregate supply curve so as to restore full employment (Bernanke and
Carey 1996). The worldwide decline in aggregate demand as an
explanation for the Great Depression has its origin in the work of Keynes,
for it was his analysis in the General Theory which turned economists’
attention away from the classical emphasis on supply-side factors.
A l t h o u g h e c o n o m i s t s c e r t a i n l y e x a m i n e d ( w h a t w e n o w c a l l )
4 The development of modern macroeconomics
macroeconomic issues prior to the publication of the General Theory, even
Pigou argued that Keynes was the first economist to bring together real and
monetary factors ‘in a single formal scheme, through which their interplay
could be coherently investigated’ (quoted in Solow 1986). The dominance
of Keynes in the field of macroeconomics prior to his death in 1946 is
clearly illustrated by looking at data on citations for the period 1920–44
Table 1.1 Most cited macroeconomists 1920–30
Table 1.2 Most cited macroeconomists 1931–5
Table 1.3 Most cited macroeconomists 1936–9
The development of modern macroeconomics 5
(see Tables 1.1–1.4). The outstanding feature of these tables is the extent to
which Keynes had come to dominate macroeconomics by the mid-1930s.
In Solow’s (1986) view, the General Theory ‘has certainly been the most
influential work of economics of the 20th century, and Keynes the most
important economist’.
A central theme of Keynes’s analysis is the contention that capitalist
market economies are inherently unstable and are capable of coming to rest
‘in a chronic condition of sub-normal activity for a considerable period
without any marked tendency, either towards recovery or towards complete
collapse’ (Keynes 1936:249). This instability was in Keynes’s view
predominantly the result of fluctuations in aggregate demand and the Great
Depression resulted from a sharp fall in investment expenditure ‘occasioned
by a cyclical change in the marginal efficiency of capital’. The resulting
unemployment was involuntary and reflected a state of deficient aggregate
demand. Given the weak equilibrating powers of the market mechanism in
these circumstances the implication of Keynes’s analysis was that fiscal and
monetary policy could correct the aggregate instability exhibited by market
economies and help stabilize the economy at full employment. Once full
employment is restored Keynes accepted that ‘the classical theory comes into
its own again’ and Keynes was optimistic that limited government
intervention could remedy the shortcomings of the invisible hand (see Keynes
1936:379). Managed capitalism, with a commitment to full employment,
was the kind of system Keynes had in mind when in the concluding section of
his famous essay The End of Laissez-Faire (1924) he argued that:
For my part I think that capitalism, wisely managed, can probably be
made more efficient for attaining economic ends than any alternative yet
in sight, but that in itself it is in many ways objectionable. Our problem is
Table 1.4 Most cited macroeconomists 1940–4
Source: Deutscher (1990)
6 The development of modern macroeconomics
to work out a social organisation which shall be as efficient as possible
without offending our notions of a satisfactory way or life.
(reprinted in Keynes 1972:294)
Keynes objected to mass unemployment which his analysis defined as largely
involuntary. In such a situation the economy can be said to be operating in
what Tobin (1992) refers to as ‘the Keynesian regime’. Here aggregate
economic activity is demand constrained and additional ‘effective’ demand
creates its own supply since the economy has the necessary spare capacity
during a recession. However, once full employment is restored, the economy
operates in the supply constrained classical regime. Here supply creates its
own demand. Whereas the classical model recognizes only the supply
constrained regime, Keynes and Keynesians believe that the economy is
capable, at different times, of being in either regime.
As early as the mid-1950s the consensus which was beginning to emerge
in macroeconomics particularly in the USA was labelled the neoclassical
synthesis by Samuelson:
In recent years 90 per cent of American economists have stopped being
‘Keynesian economists’ or ‘anti-Keynesian economists’. Instead they have
worked towards a synthesis of whatever is valuable in older economics
and in modern theories of income determination. The result might be
called neoclassical economics and is accepted in its broad outlines by all
but about 5 per cent of extreme left-wing and right-wing writers.
(Samuelson 1955:212, emphasis added)
The initial synthesis proceeded along two lines of inquiry. The first studied
the long-run movement of output by identifying the determinants of the trend
and ignoring fluctuations around the trend. Significant contributions were
made during this period to the development of growth theory (see Hahn and
Matthews 1964). The second line of inquiry concentrated on the analysis of
short-run fluctuations around the trend. At the centre of this analysis lay the
Hicks-Hansen IS-LM framework (Hicks 1937; Hansen 1953; Young 1987;
Darity and Young 1995). During this period a great deal of macroeconomic
research was devoted to refining the four basic building blocks of the IS-LM
model, namely the consumption function, the investment function and the
demand for, and supply of, money.
During the late 1950s and early 1960s a consensus emerged with respect
to the ‘Keynes v. Classics’ debate in which it was generally accepted that at
the theoretical level, once the Pigou or wealth effect of falling prices on
consumption expenditure is taken into account, then unemployment
equilibrium is possible in the Keynesian IS-LM model only where downward
money wage rigidity prevents the classical automatic adjustment to full
employment. Nevertheless at the practical policy level it was conceded that
the process of adjustment via the Pigou effect might be so weak and slow
The development of modern macroeconomics 7
that interventionist policies (notably expansionary fiscal policy) would be
required in order to achieve the primary stated objective of full employment
(see Snowdon et al. 1994: Chapter 3). With a relatively inelastic IS curve
and a relatively elastic LM curve Keynesianism became synonymous with
‘fiscalism’ and policies to fine tune the macroeconomy.
The publication of the results of Bill Phillips’s (1958) statistical
investigation into the relationship between the level of unemployment and
wage inflation, and Richard Lipsey’s (1960) subsequent theoretical rationale
for the curve, proved to be another important development during this period
(see Santomero and Seater 1978; Wulwick 1987). The Phillips curve was
quickly adopted by orthodox Keynesian economists for three main reasons.
First, it provided an explanation of price determination, and inflation, which
was missing in the then-prevailing macroeconomic model. Within the IS-LM
model the price level is assumed to be fixed at less than full employment
with the result that changes in aggregate demand affect only the level of
output and employment. Up to full employment money wages are assumed to
be fixed and unresponsive to changes in aggregate demand. The Phillips
curve allowed the orthodox theory of output and employment determination
to be linked to a theory of wage and price inflation (see Lipsey 1978).
Second, the Phillips curve appeared to provide rare evidence of a stable
relationship between unemployment and inflation that had existed for almost
a century. Third, the curve provided an insight into the problem that policy-
makers face of simultaneously achieving high levels of employment with
price stability given the trade-off between wage inflation and unemployment.
As such the Phillips curve was interpreted by many orthodox Keynesians as
implying a stable long-run trade-off which offered the authorities a menu of
possible inflation-unemployment combinations for policy choice (see for
example Samuelson and Solow 1960). Up to at least the mid-to-late 1960s
the prevailing Keynesian consensus in macroeconomics was one in which the
IS-LM model was used to explain the determination of output and
employment, while the Phillips curve enabled policy-makers to predict the
rate of inflation which would result from different target levels of
unemployment being attained by activist demand-management policies. This
consensus position was first seriously challenged by Milton Friedman, who
launched the monetarist attack against orthodox Keynesian analysis and
policy-activism during the 1950s and 1960s.
THE MONETARIST COUNTER-REVOLUTION
Friedman’s starting-point was one in which he sought to re-establish the
quantity theory of money approach to macroeconomic analysis which had
been usurped by the Keynesian revolution. In the mid-to-late 1940s and the
1950s the then-prevailing Keynesian orthodoxy emphasized real demand
disturbances (notably fluctuations in investment and autonomous
consumption) as the main cause of fluctuations in money or nominal
8 The development of modern macroeconomics
income, predominantly in the form of changes in real income. In contrast,
within the quantity theory approach, changes in the money stock are
regarded as the predominant, though not the only, factor explaining
changes in money income. Nevertheless Friedman (1956) initially presented
his now famous restatement of the quantity theory of money as a theory of
the demand for money rather than a theory of the general price level or
money income. In his paper he asserted that the demand for money
function was stable, an assertion which lies at the heart of the modern
quantity theory approach to macroeconomic analysis. If the demand for
money function is stable then velocity will also be stable, changing in a
predictable manner if any of the limited number of variables in the demand
for money function should change (see Laidler 1993). Over the period of
the mid-to-late 1950s to the mid-1960s various empirical evidence was put
forward in support of the belief that most of the observed instability in the
economy could be attributed to factors which affected the supply of money
independently of any change in the demand for money. In the latter case it
was claimed that changes in the demand for money tend to take place
gradually or result from events set in motion by prior changes in the supply
of money.
The most persuasive evidence to support the belief that changes in the
stock of money play a largely independent role in cyclical fluctuations was
presented by Milton Friedman and Anna Schwartz (1963) in their influential
study of the Monetary History of the United States (see Lucas 1994; Miron
1994). Friedman and Schwartz found that the only times when there was an
absolute fall in the money stock corresponded with the six periods of major
economic contraction over the period 1867–1960. Furthermore from studying
the historical circumstances underlying the changes that occurred in the
money stock during the six major recessions, they argued that the factors
producing monetary contraction were mainly independent of contemporary
or prior changes in money income and prices. In such circumstances
Friedman and Schwartz interpreted monetary changes as the cause, rather
than the consequence, of major recessions. For example, re-examining the
monetary history of the period of the Great Depression, they argued that the
depression became ‘Great’ only as a consequence of the failure of the Federal
Reserve to prevent a dramatic decline in the money stock; between October
1929 and June 1933 the money stock fell by about a third. According to
Friedman and his associates, the Great Depression demonstrated the potency,
rather than the ineffectiveness, of monetary change and monetary policy (see
Hammond 1996). However, because of the length and variability of the time
lag involved between the implementation of monetary policy and its effects
on money income (see Friedman 1958; 1961) it was suggested that
discretionary monetary policy could turn out to be destabilizing. In
consequence Friedman argued that the money supply should be allowed to
grow at a fixed rate in line with the underlying growth of output to ensure
long-term price stability.
The development of modern macroeconomics 9
The direction of the monetarist attack against the Keynesian demand-
management policies and policy-activism changed at the end of the 1960s
when Friedman (1968) augmented the basic Phillips curve with the expected
rate of inflation as an additional variable determining the rate of change of
money wages (Phelps 1967 provided a similar analysis from a non-
monetarist perspective). While A Monetary History has undoubtedly been
Friedman’s most influential book in the macroeconomics sphere, his 1967
presidential address to the American Economic Association published as The
Role of Monetary Policy’ (1968) has certainly been his most influential
article. In 1981 Robert Gordon described this paper as probably the most
influential article written in macroeconomics in the previous twenty years.
More recently James Tobin (1995), one of Friedman’s most eloquent,
effective and long-standing critics, has gone even further, describing the 1968
paper as ‘very likely the most influential article ever published in an
economics journal’ (emphasis added). Friedman’s utilization of Wicksell’s
concept of the ‘natural rate’ in the context of unemployment was in
rhetorical terms a ‘masterpiece of marketing’ (see Dixon 1995) just as the
application of the term ‘rational’ to the expectations hypothesis turned out to
be in the rise of new classical economics during the 1970s. The impact of
Professor Friedman’s work forced Keynesians to restate and remake their
case for policy-activism even before that case was further undermined by the
penetrating theoretical critiques of Robert Lucas and other leading new
classical economists. In line with orthodox neoclassical microeconomic
analysis, Friedman suggested that the demand for and supply of labour
should be specified in real not money terms. Friedman denied the existence
of a permanent (long-run) trade-off between inflation and unemployment,
and put forward the ‘natural rate of unemployment’ hypothesis. Five main
implications for the role and conduct of stabilization policy derive from the
view that the long-run Phillips curve is vertical at the natural rate of
unemployment. First, the authorities can reduce unemployment below the
natural rate only in the short run and then only because inflation is not fully
anticipated. The assumption underlying orthodox monetarist analysis is that
expected inflation adjusts to actual inflation only gradually in line with the
so-called ‘adaptive’ expectations hypothesis. Second, any attempt to
maintain unemployment permanently below the natural rate will result in
accelerating inflation. Third, if governments wish to reduce the natural rate
of unemployment in order to achieve higher output and employment levels
they should pursue supply-management policies which are designed to
improve the structure and functioning of the labour market and industry
rather than demand-management policies. Fourth, the natural rate is
compatible with any rate of inflation which in turn is determined by the rate
of monetary expansion in line with the quantity theory tradition. Given the
belief that inflation is essentially a monetary phenomenon (see Friedman
1970) propagated by excessive monetary growth monetarists argue that
inflation can be reduced only by slowing down the rate of growth of the
10 The development of modern macroeconomics
money supply. In monetarist analysis the output/employment cost of
disinflation depends on three main factors namely: (1) whether the
authorities follow a path of rapid or gradual monetary contraction (cold
turkey v. gradualism); (2) the extent of institutional adaptations such as
indexation (see Friedman 1974); and (3) the speed economic agents adjust
their inflation expectations downwards which in large part depends on the
credibility of any anti-inflation strategy. Finally, in a world of fixed
exchange rates inflation is viewed as an international monetary phenomenon
explained by an excess-demand expectations model. Monetarists attribute the
acceleration of inflation that occurred in western economies in the late 1960s
primarily to an increase in the rate of monetary expansion in the United
States in order to finance increased spending on the Vietnam War (see for
example Johnson 1972; Laidler 1976). In practice the US determined
monetary conditions for the rest of the world, a situation which eventually
proved unacceptable to other countries and helped lead to the breakdown of
the Bretton Woods system in the early 1970s.
During the early 1970s the subject of the possible existence of a long-run
vertical Phillips curve became a controversial issue in the Keynesianmonetarist
debate and numerous empirical studies of the expectations—augmented
Phillips curve were undertaken. However by the mid-to-late 1970s at least as
far as the United States was concerned, the majority of mainstream Keynesians
had come to accept that the Phillips curve was vertical in the long run (see
Blinder 1988—reprinted on pp. 109–34). While the controversy over the slope
of the long-run Phillips curve was largely laid to rest, the associated
controversy over the role for short-run interventionist stabilization policy
continued unabated. Even if the long-run Phillips curve is vertical, Keynesian
arguments justifying intervention to stabilize the economy in the short run can
be made on the grounds of the length of time required for the economy to
return to the natural rate of unemployment and the potential to identify and
respond to economic disturbances.
The failure of inflation to slow down both in the US and UK economies in
1970–1, despite rising unemployment, and the subsequent simultaneous rise
of unemployment and inflation during the 1970s destroyed the idea that there
might be a permanent long-run trade-off between inflation and
unemployment. These events also verified the predictions of Friedman’s
model and contradicted the then-prevailing Keynesian views. As a result of
these developments there is little doubt that Milton Friedman became ‘the
most influential macroeconomist’ from the late 1960s to the mid-1970s (see
Snowdon and Vane 1997). With hindsight 1976, the year when he was
awarded the Nobel Prize for Economics, probably marked the pinnacle of
Friedman’s influence in academia even if monetarism had yet to rise (and
fall) in the policy-making arena following the initiation of the Volcker and
Thatcher disinflations (see Friedman 1977; Blinder 1987).
Although within academia monetarism is no longer the influential force it
was in the late 1960s and early 1970s, a large part of the reason for this
The development of modern macroeconomics 11
apparent decline can be attributed to the fact that mainstream
macroeconomics has absorbed the insights of monetarism with a small ‘m’.
The expectations-augmented Phillips curve is now a standard part of the
Keynesian-monetarist synthesis, although modern hysteresis theories of
unemployment challenge Friedman’s natural rate hypothesis which denies the
importance of aggregate demand factors in influencing the equilibrium rate
of unemployment (see Cross 1995; 1996). Of enormous importance has been
Friedman’s numerous contributions which have succeeded in reminding
economists that their knowledge of how the economy functions is limited.
Friedman’s view is that by claiming more than can be delivered economists
have on too many occasions encouraged the general public ‘to expect
standards of performance that as economists we do not know how to
achieve’ (Friedman 1972). This was a lesson that monetarists themselves
were to learn during the ‘great velocity decline’ during the 1979–82 period
when the stable demand for money function began to suffer the same fate as
had befallen the stable Phillips curve (see Laidler 1985; Modigliani 1988—
reprinted on pp. 247–61). As a result ‘hard core’ monetarism with a capital
‘M’—devoted to Friedman’s advocacy of a rigid monetary growth rate rule—
has few remaining supporters. However, the majority of modern Keynesians
recognizing the political, economic and informational constraints facing
policy-makers also now accept that in practice the opportunity for frequently
exploiting fiscal policy for stabilization purposes is extremely limited. In
addition whatever controversies remain over aggregate fluctuations
Friedman has undoubtedly won one important debate, that relating to the
determinants of sustained inflation. A clear majority of economists and
central banks emphasize the rate of growth of the money supply when it
comes to explaining and combating inflation over the long run. This allows
mainstream macroeconomists to attribute temporary bouts of inflation to
non-monetary causes such as supply shocks.
During the 1970s theoretical developments in macroeconomics were
dominated by the new classical school. The contributions of Lucas and his
associates cast further doubt on the mature synthesis model even when
modified to incorporate the expectations-augmented Phillips curve.
THE CHALLENGE OF RATIONAL EXPECTATIONS AND NEW
CLASSICAL MACROECONOMICS
During the 1970s the new classical approach to macroeconomics replaced
monetarism as the main rival to Keynesianism. Underlying the approach,
which is often taken to be synonymous with the work most notably of
Robert Lucas, Thomas Sargent, Robert Barro, Edward Prescott and Neil
Wallace, is the joint acceptance of three main sub-hypotheses. First, the
rational expectations hypothesis which is associated with the work of John
Muth in the context of microeconomics. In his seminal article Muth (1961)
suggested ‘that expectations since they are informed predictions of future
12 The development of modern macroeconomics
events are essentially the same as the predictions of the relevant economic
theory’. In the Muthian version of the rational expectations hypothesis
which has been incorporated into new classical models, economic agents’
subjective expectations of economic variables will coincide with the true or
objective mathematical conditional expectations of those variables, with
the crucial implication that economic agents will not form expectations
which are systematically wrong over time. Second, new classical models
are Walrasian in that all observed outcomes are viewed as ‘market-
clearing’ at each point of time, given the assumption that markets
continuously clear all possible gains from trade have been exploited and
utility has been maximized. Third, new classical models incorporate an
aggregate supply hypothesis based on two orthodox microeconomic
assumptions, namely that (1) rational decisions taken by workers and firms
reflect optimizing behaviour on their part and (2) the supply of labour by
workers, and output by firms, depends upon relative prices. The new
classical approach to aggregate supply derives from the highly influential
work of the 1995 Nobel Laureate Robert Lucas (1972, 1973). This work
has given rise to the so-called Lucas ‘surprise’ supply function (in effect a
restatement of the expectations-augmented version of the Phillips curve)
where output deviates from its natural level only in response to errors in
price (inflation) expectations (see Blanchard 1990).
By combining the Friedman-Phelps natural rate hypothesis with the
assumption of continuous market clearing and the rational expectations
hypothesis, Lucas was able to demonstrate rigorously how a short-run
Phillips curve would result if inflation was unanticipated due to incomplete
information. Since a short-run trade-off between a real variable
(unemployment) and a nominal variable (inflation) breaks the classical
dichotomy, the work of Lucas was crucial in that it demonstrated that the
classical model is compatible with the Phillips curve phenomena providing
the assumption of perfect information is abandoned. By invoking the Lucas
aggregate supply hypothesis monetary shocks can have a temporary
influence on real variables, that is unanticipated money is non-neutral.
Within this framework Lucas (1975, 1977) was able to develop an
equilibrium monetary explanation of the business cycle.
During the 1970s it would be no exaggeration to say that there was a
‘rational expectations revolution’ (Taylor 1989). The combination of the
rational expectations, continuous market clearing and aggregate supply
hypotheses within new classical models produces six highly controversial
policy implications. The first of these, the policy ineffectiveness
proposition, was initially presented in two influential papers by Thomas
Sargent and Neil Wallace (1975, 1976). As ‘forward-looking’ rational
economic agents will take into account any ‘known’ monetary rule in
forming their expectations, new classical models predict that the authorities
will be unable to influence output and employment even in the short run by
pursuing a systematic monetary policy. Furthermore any attempt to affect
The development of modern macroeconomics 13
output and employment by random or non-systematic monetary policy will,
new classicists argue, only increase the variation of output and
employment around their natural levels. Second, in contrast to both
Keynesianism and monetarism the new classical approach implies that as
long as announced monetary contraction is believed to be credible, rational
economic agents will immediately revise downwards their inflation
expectations enabling the authorities to engineer painless disinflation (see
Blackburn and Christensen 1989). Only where policy announcements lack
credibility will inflation expectations fail to fall sufficiently to prevent the
economy from experiencing output/employment costs. Third, closely
related to the importance of policy credibility is the problem of dynamic
time-inconsistency first highlighted by Finn Kydland and Edward Prescott
(1977) in support of monetary policy being conducted by rules rather than
discretion. The problem can be illustrated as follows (see Fischer 1990).
Suppose the authorities announce a policy of monetary contraction to
reduce inflation. If the policy is believed and economic agents revise
downwards their inflation expectations then authorities who are not bound
by a fixed monetary growth rate rule will have an incentive to cheat or
renege on their announced policy in order to reduce unemployment. In
circumstances where the authorities have such discretionary powers, and
have in consequence an incentive to cheat, the credibility of announced
policies will be significantly weakened. Since the difficulty of gaining
credibility derives from the authorities having discretionary powers with
respect to monetary policy the problem could be overcome by transferring
the responsibility of anti-inflation policy to an independent central bank.
Fourth, associated with the work of Robert Barro (1974; 1989—reprinted
on pp. 314–33) is the highly controversial Ricardian debt equivalence
theorem limiting the usefulness of tax changes as a stabilization
instrument. According to this theorem a bond-financed tax cut will leave
consumption unchanged as the private sector will fully anticipate the future
tax liability required to meet interest-payments on, and repayment of, the
debt. The fifth main policy implication of the new classical approach
concerns what policies the authorities should pursue if they wish to increase
output/reduce unemployment permanently. Given that changes in output
and employment are held to reflect the equilibrium supply decisions of
firms and workers, given their perceptions of relative prices, it follows
from this view that the appropriate policy measures to increase output and
reduce unemployment are those that increase the microeconomic incentives
for firms and workers to supply more output and labour. The final
implications of the new classical approach for the formulation of
macroeconomic policy concerns what is known as the ‘Lucas critique’ of
econometric policy evaluation after the title of Robert Lucas’s (1976)
seminal paper in which the proposition first appeared. Lucas concluded
that macroeconometric models should not be used to predict the
consequences of alternative policies since the parameters of such models
14 The development of modern macroeconomics
may change as economic agents adjust their expectations and behaviour to
the new policy environment.
Despite the enormous influence of these and other developments, by about
1980 the Barro-Lucas-Sargent-Wallace monetary surprise explanation of the
business cycle had reached both a theoretical and empirical impasse. On the
theoretical front the implausibility of the assumption relating to
informational confusion was widely recognized (Tobin 1980). New classical
theorists argue that nominal rigidities are implausible in a world of rational
agents who will always exhaust the opportunities for mutually beneficial
trade (see Barro 1979). With sticky prices ruled out on methodological
grounds new classical models were left without an acceptable explanation of
the business cycle involving money to output causality. On the empirical
front while early work, in particular the seminal papers by Robert Barro
(1977, 1978) seemed to support the policy ineffectiveness proposition,
subsequent studies most notably by Frederic Mishkin (1982) and Robert
Gordon (1982) found evidence that suggested that both unanticipated and
anticipated monetary policy affects output and employment. The depth of the
recessions in both the USA and UK in the 1980–2 period following the
Reagan and Thatcher deflations provided further ammunition to the critics.
In addition opponents of the new classical approach drew attention to
aggregate price and money supply data which are readily available to
economic agents at a relatively low cost and questioned how this could be
reconciled with the magnitude and length of actual business cycles
supposedly caused by incomplete information. These criticisms led a number
of economists who were sympathetic to the new classical approach to
develop a mark II version which, while maintaining the assumptions of
rational expectations and continuous market clearing, has reverted to a full
information assumption relating to monetary developments and views
business cycles as being predominantly caused by persistent real (supply-side)
shocks rather than monetary (demand-side shocks) to the economy. Leading
exponents and/or contributors to the so-called real business cycle approach
include John Long, Charles Plosser, Robert Barro, Robert King, Finn
Kydland and Edward Prescott.
THE REAL BUSINESS CYCLE APPROACH TO ECONOMIC
FLUCTUATIONS
During the 1970s, with the rebirth of interest in business cycle research,
economists became more involved with research into the statistical properties
of economic time series. One of the main problems in this work is to
separate trend from cycle. The conventional approach has been to imagine
that the economy evolves along a path reflecting an underlying trend rate of
growth described by Solow’s neoclassical growth model (Solow 1956). This
approach assumes that the supply determined long-run trend component of
GDP is smooth with short-run fluctuations about trend being primarily
The development of modern macroeconomics 15
determined by demand shocks. This conventional wisdom was accepted by
Keynesian, monetarist and new classical economists alike until the early
1980s. The demand-shock models of all three groups interpret output
deviations from trend as temporary. Whereas Keynesians such as James
Tobin (1987) feel that such deviations could be severe and prolonged and
therefore justify the need for corrective action, monetarists, and especially
new classical economists, reject the need for activist stabilization policy
having greater faith in the equilibrating power of market forces. Nelson and
Plosser’s (1982) important paper challenged this conventional wisdom. The
research of Nelson and Plosser into macroeconomic time series led them to
conclude that
macroeconomic models that focus on monetary disturbances as a source of
purely transitory fluctuations may never be successful in explaining a large
fraction of output variation and that stochastic variation due to real factors
is an essential element of any model of macroeconomic fluctuations.
Nelson and Plosser reached this important conclusion because in their
research into US data they were unable to reject the hypothesis that GDP
follows a random walk. Nelson and Plosser argue that most of the changes
in GDP that we observe are permanent in that there is no tendency for output
to revert to its former trend following a shock.
These findings of Nelson and Plosser have radical implications for
business cycle theory. If shocks to productivity growth due to technological
change are frequent and random then the path of output following a random
walk will exhibit features which resemble a business cycle. In this case
however the observed fluctuations in GDP are fluctuations in the natural
(trend) rate of output not deviations of output from a smooth deterministic
trend. What looks like output fluctuating around a smooth trend is in fact
fluctuations in the trend itself due to a series of permanent shocks with each
permanent productivity shock determining a new growth path. Whereas
following Solow’s seminal work economists have traditionally separated the
analysis of growth from the analysis of fluctuations, the work of Nelson and
Plosser suggests that the economic forces determining the trend are not
different from those causing fluctuations. Since permanent changes in GDP
cannot result from monetary shocks in a new classical world because of the
neutrality proposition embedded in the natural rate hypothesis the main
forces causing instability must be real shocks. Nelson and Plosser interpret
their findings as placing limits on the importance of monetary theories of the
business cycle and that real disturbances are likely to be a much more
important source of output fluctuations. If there are important interactions
between the process of growth and business cycles then the conventional
practice of separating growth theory from the analysis of fluctuations is
illegitimate. By ending the distinction between trend and cycle, real business
cycle theorists have begun to integrate the theory of growth and fluctuations
16 The development of modern macroeconomics
(see Kydland and Prescott 1982). Hence the basic real business cycle model
is a stochastic dynamic equilibrium growth model.
The starting-point of the real business cycle approach (see Plosser 1989—
reprinted on pp. 396–424; Stadler 1994) is the assumption that the economy
is subjected to random supply-side shocks, most notably large random
fluctuations in the rate of technological progress. These large and random
shocks to the production function result in fluctuations in relative prices to
which rational economic agents respond. According to this approach
observed fluctuations in output and employment are equilibrium phenomena
and are the outcome of rational economic agents responding optimally to
unavoidable changes in the economic environment. For example the
approach controversially assumes that fluctuations in employment reflect
voluntary changes in the amount of labour people wish to supply. As
fluctuations in output and employment are seen as Pareto efficient responses
to shocks to the production function the approach implies that monetary
policy is irrelevant in explaining such fluctuations (see Van Els 1995).
However, the main policy implication of the approach is that because the
existence of fluctuations in GNP do not imply the failure of markets to clear
and are instead regarded as fluctuations in the natural (trend) rate of output,
the government should not attempt to reduce these fluctuations through
stabilization policy not only because such attempts are unlikely to achieve
their desired objective, but also because reducing instability would reduce
welfare (see Prescott 1986—reprinted on pp. 366–88). While monetary policy
has no real effects, the government could do a great deal of harm if its
taxation and spending policies distorted output and employment from the
optimal amounts chosen by firms and workers. Needless to say the approach
is highly controversial and has been subjected to a number of criticisms (see
Summers 1986—reprinted on pp. 389–95; Mankiw 1989—reprinted on pp.
425–36). One final feature of the approach worth noting is that concerning
the development of the calibration method. Rather than attempting to
provide models capable of conventional econometric testing, real business
cycle theorists have instead developed the calibration method in which the
simulated results of their specific models (when hit by random shocks) in
terms of key macroeconomic variables are compared with the actual
behaviour of the economy. Real business cycle theory therefore represents a
specific approach to macroeconomic modelling (see Lucas 1980; Danthine
and Donaldson 1993).
NEW KEYNESIAN ECONOMICS
The third approach which has dominated the more recent development of
macroeconomics is new Keynesian economics. Since the mid-1980s the new
Keynesian school has emerged as the main rival to the new classical
approach. Leading exponents and/or contributors to the approach include
Gregory Mankiw, Olivier Blanchard, George Akerlof, Janet Yellen, David
The development of modern macroeconomics 17
Romer, Joseph Stiglitz and Bruce Greenwald. While most new Keynesian
analysis incorporates the rational expectations and natural rate hypotheses,
it does not incorporate the new classical assumption of continuous market
clearing. Indeed the central focus of one important strand of the burgeoning
new Keynesian literature has been to explore a variety of reasons for wage
and price stickiness that prevent market clearing. This has involved research
into the causes of (1) nominal wage stickiness (e.g. via overlapping long-
term wage contracts); (2) nominal price stickiness (e.g. arising from menu or
adjustment costs faced by monopolistically competitive firms); (3) real
rigidities in both the labour market (e.g. via efficiency wage, insider-outsider
and implicit contract models) and product market (e.g. via customer
markets); and (4) co-ordination failures (see for example Gordon 1990—
reprinted on pp. 478–551; Mankiw and Romer 1991; D.Romer 1993).
One problem with the new Keynesian developments is that there is no
single new Keynesian model, rather the research programme has led to a
multiplicity of explanations of wage and price rigidities, and their
macroeconomic consequences. Although the numerous explanations to be
found in the literature are not necessarily mutually exclusive and often
complement each other, it is the case that different economists within the new
Keynesian research programme emphasize various aspects and causes of
market imperfections and their macroeconomic consequences (see Stiglitz
1992). Bearing this in mind we draw attention to three policy implications
that derive from new Keynesian analysis. First, in new Keynesian models
which emphasize wage and price stickiness money is no longer neutral and
policy effectiveness is re-established. For example, Stanley Fischer (1977)
and Edmund Phelps and John Taylor (1977) have shown that nominal
demand disturbances are capable of producing real effects in models
incorporating rational expectations providing the new classical assumption
of continuous market clearing is abandoned. In such models systematic
monetary policy can help stabilize the economy. Second, the gradual
adjustment of prices and wages in new Keynesian models implies that any
policy of monetary disinflation, even if credible and anticipated by rational
economic agents, will lead to a substantial recession (sacrifice ratio) in terms
of output and employment. Furthermore in some new Keynesian analysis the
equilibrium rate of unemployment is affected by the path taken by the actual
rate of unemployment so that the natural rate is affected by the aggregate
demand (see Cross 1996). In circumstances where unemployment remains
above the natural rate for a prolonged period the natural rate itself will tend
to increase due to so-called hysteresis effects. Not only will those who are
unemployed suffer a deterioration of their human capital (skills)
exacerbating the problem of structural unemployment, but the number of
long-term unemployed is also likely to increase. In the latter case it is
claimed that such outsiders exert little influence on wages and are unable to
price their way back into jobs, and as a result the natural rate of
unemployment rises. Such hysteresis effects provide new Keynesians with a
18 The development of modern macroeconomics
strong case to boost aggregate demand during a protracted recession. While
monetarism and new classical macroeconomics undermined the case for
‘fine’ tuning, new Keynesians have championed the case for ‘rough’ or
‘course’ tuning where policies are designed to offset or avoid the more
serious macro-level problems. Lastly, contrary to the new classical approach,
new Keynesian analysis has provided a rationale for the existence of
involuntary unemployment as an equilibrium phenomenon. For example, in
efficiency wage models (see Yellen 1984) firms are reluctant to cut wages
even in the face of an excess supply of labour (persistent unemployment)
since such a policy would be counter-productive as lower efficiency/
productivity would result. In summary the bulk of new Keynesian research
has sought to develop models with coherent microfoundations, in order to
explain why prices and wages adjust only gradually, and in doing so have
sought to re-establish a case for policy effectiveness and justify
interventionist policies (both supply and demand—management policies) to
stabilize the economy. However, it should be noted that a second strand of
new Keynesian work demonstrates how money is not neutral even if prices
and wages are perfectly flexible (see Greenwald and Stiglitz 1993—reprinted
on pp. 552–74). Greenwald and Stiglitz follow Keynes (1936) and argue that
increasing price flexibility could well be destabilizing.
THE RENAISSANCE OF ECONOMIC GROWTH ANALYSIS
The causes of the enormous differences in living standards across time and
space have long been of interest to economists. Since rising living standards
depend in the long run on economic growth it is extremely important for
economists to understand and quantify this process. The consequences of
even small differences in growth rates when compounded over time, are
striking. The example provided by Barro and Sala-i-Martin (1995) clearly
brings this point out. The 1.75 per cent growth rate of real per capita income
in the USA between 1870 and 1990 enabled real GDP per head to rise from
$2,244 to $18,258. If the growth rate had been 0.75 per cent over the same
period, income per capita would have risen to $5,519. However, if the
growth rate had been 2.75 per cent the real income per person in 1990 would
have been an astonishing $60,841! While short-run fluctuations in output
have important welfare consequences (in the opinion of the majority of
mainstream economists) it is evident that ‘the welfare implications of long-
run growth swamp any possible effects of the short-run fluctuations that
macroeconomics traditionally focuses on’ (D.Romer 1996). As we have seen
Keynesian, monetarist and new classical analysis was primarily concerned
with trying to understand short-run instability of output, employment and the
price level. Some prominent macroeconomists now seem to regard these
previous efforts as perhaps misguided given that the economic instability
since the second world war has been a relatively ‘minor problem’ (Lucas,
The development of modern macroeconomics 19
1987 p.30). Barro and Sala-i-Martin (1995) are more forceful and are worth
quoting:
if we can learn about government policy options that have even small
effects on the long-term growth rate, then we can contribute much more to
improvements in standards of living than has been provided by the entire
history of macroeconomic analysis of countercyclical policy and fine-
tuning. Economic growth…is the part of macroeconomics that really
matters.
Given the importance of economic growth it is surprising that economists’
interest in the theoretical and empirical issues relating to the causes and
consequences of growth has itself been cyclical. Dynamic issues were of
major concern to the classical economists who sought to understand the
nature and causes of the ‘Wealth of Nations’. But following the ‘marginalist
revolution’ in the 1870s neoclassical economists turned the focus of their
attention towards problems associated with static microeconomic issues
relating to the efficient allocation of resources. With the onset of the Great
Depression and the subsequent Keynesian revolution, economists quite
understandably switched their interest towards the causes of short-run
aggregate instability. However, some notable Keynesian economists (e.g.
Hansen) feared that the Great Depression represented more than a severe
example of a business cycle downswing. Rather it was feared that there may
be a long-run tendency for capitalist economies to produce an actual rate of
growth less than the underlying growth of productive potential. Hansen’s
stagnation thesis was an important contributing factor leading to a
reawakening of interest in long-run issues in the post-war period (for a
discussion of the stagnation thesis see Ackley 1966: Chapter 18).
Between 1956 and 1970 economists refined and developed the Solow-
Swan neoclassical growth model (see Solow 1956; 1957; Swan 1956). But
thereafter, until the mid-1980s macroeconomic research was predominantly
concerned with business cycle issues in the wake of the 1973 oil shocks and
theoretical developments which absorbed the rational expectations
hypothesis into macroeconomic analysis. Following the contributions of Paul
Romer (1986) and Robert Lucas (1988) the study of economic growth has
once again become a vibrant research area. In 1996 the first issue of a new
Journal of Economic Growth was launched and many well-known
macroeconomics textbooks now have their discussion of economic growth at
the beginning rather than at the end of the text (see Mankiw 1994; D.Romer
1996). Economic growth has returned as an active research area and is
central to contemporary macroeconomics.
The essential starting-point to any discussion of economic growth is the
Solow growth model which has as its centrepiece the standard neoclassical
production function (see Solow 1994—reprinted on pp. 649–59). This
wellknown framework illuminates how growth of the capital stock and
20 The development of modern macroeconomics
labour force interact with technological progress to produce more output.
Given a Cobb-Douglas production function with diminishing returns to
factors and constant returns to scale, economic growth is the result of
changes in both the quality and quantity of factor inputs. Household savings
are converted via investment into a higher capital stock which generates
growth of output. However because of diminishing returns the marginal
product of capital declines as the capital-labour ratio rises. The
accumulation of reproducible inputs contributes less and less to growth and
will approach zero unless population growth and technological change allow
the quality and quantity of non-reproducible factors to rise. In Solow’s model
the long-run rate of growth is driven by the exogenous factors of labour force
growth and technological progress and is independent of the rate of
investment. A change in the savings rate can have only a temporary effect on
growth although the savings rate does influence the level of output per head.
The growth of long-run income per head depends on total factor productivity
which is driven by exogenous improvements in technology. Hence long-run
growth appears to be beyond the influence of government policy.
A striking implication of the Solow model is that if technical progress is
regarded as a public good and freely available to all countries, rich and
poor, there should be no cross-country divergence of growth rates of income
per capita. Differences in the level of per capita income result from
variations in the capital-labour ratio. Hence countries with low initial
incomes per capita due to low capital-labour ratios have the potential for
rapid growth which will allow them to catch up the high income countries.
This convergence hypothesis is conditional on the underlying determinants of
the steady state (see Abramovitz 1986—reprinted on pp. 582–603; Baumol
1986). The empirical evidence on the extent of convergence suggests that this
process has been present among some industrialized countries but apart from
a limited number of ‘star’ performers (particularly in East Asia) the
convergence between developed and developing countries has been limited or
absent. As a result there is growing evidence of widening disparities of
income across the world’s economies (see Dowrick 1992—reprinted on pp.
604–15).
Given the absence of convergence when a broad sample of countries are
considered a number of theorists, notably P.Romer (1986) and Lucas (1988),
have developed growth models where two essential assumptions of the Solow
model are abandoned, namely that technological change is exogenous and
that all countries have the same access to technological opportunities (see
P.Romer 1994—reprinted on pp. 628–48). These endogenous theories of
growth have dominated the new literature since the mid-1980s (see Van de
Klundert and Smulders 1992; Pack 1994). In some models of endogenous
growth there are constant returns to broad capital accumulation and
investment in physical and human capital can permanently raise the growth
of output per head. Other models stress endogenous innovation and reject the
idea of a universally available technology. In this case poor countries may
The development of modern macroeconomics 21
fail to catch up the leading countries because of ‘idea gaps’ rather than
‘object gaps’ (see P.Romer 1993; Crafts 1996). Lucas (1993) has argued that
‘the main engine of growth is the accumulation of human capital—of
knowledge—and the main source of differences in living standards among
nations is differences in human capital’. Endogenous growth theory also
gives rise to important policy implications which imply that governments
can influence the long-run growth rate (see Shaw 1992—reprinted on pp.
616–27).
CONCLUSION
Since the mid-1930s there has been considerable progress in our
understanding of macroeconomic phenomena. Following the monetarist
contributions a consensus of economists now accept that sustained inflation is
a monetary phenomenon. New classical theory has produced numerous
insights with respect to the role and conduct of stabilization policy. Real
business cycle theorists have caused everyone to rethink basic issues
associated with economic fluctuations. The insights provided by new
Keynesian economists have transformed the microfoundations of the supply
side of models which stress the importance of aggregate demand disturbances
in explaining aggregate instability. Finally economists are once more
producing valuable research into the causes of economic growth. In the
chapters which follow we hope the reader can capture some of the
excitement which has been a constant feature of the controversies which have
characterized macroeconomics since Keynes first stimulated its development.
REFERENCES
*Titles marked with an asterisk are particularly recommended for additional reading.
*Abramovitz, M. (1986) ‘Catching Up, Forging Ahead, and Falling Behind’, Journal of
Economic History 46, June, pp. 385–406.
Ackley, G. (1966) Macroeconomic Theory, New York: Collier-Macmillan.
Barro, R.J. (1974) ‘Are Government Bonds Net Wealth?’, Journal of Political Economy
82, November/December, pp. 1095–117.
Barro, R.J. (1977) ‘Unanticipated Money Growth and Unemployment in the United
States’, American Economic Review 67, March, pp. 101–15.
Barro, R.J. (1978) ‘Unanticipated Money, Output, and the Price Level in the United
States’, Journal of Political Economy 86, August, pp. 549–80.
Barro, R.J. (1979) ‘Second Thoughts on Keynesian Economies’, American Economic
Review 69, May, pp. 54–9.
*Barro, R.J. (1989) ‘The Ricardian Approach to Budget Deficits’, Journal of Economic
Perspectives 3, Spring, pp. 37–54.
Barro, R.J. and X.Sala-i-Martin (1995) Economic Growth, New York: McGrawHill.
*Baumol, W. (1986) ‘Productivity Growth, Convergence and Welfare’, American
Economic Review 76, December, pp. 1072–85.
Bernanke, B.S. (1995) ‘The Macroeconomics of the Great Depression: A Comparative
Approach’, Journal of Money, Credit and Banking 27, pp. 1–28.
22 The development of modern macroeconomics
Bernanke, B.S. and K.Carey (1996) ‘Nominal Wage Stickiness and Aggregate Supply in
the Great Depression’, Quarterly Journal of Economics 111, August, pp. 853–83.
*Blackburn, K. and M.Christensen (1989) ‘Monetary Policy and Policy Credibility:
Theories and Evidence’, Journal of Economic Literature 27, March, pp. 1–45.
*Blanchard, O.J. (1990) ‘Why Does Money Affect Output? A Survey’, in B.M. Friedman
and F.H.Hahn (eds) Handbook of Monetary Economics, Amsterdam: North Holland.
Blinder, A.S. (1987) Hard Heads, Soft Hearts: Tough-Minded Economics for a Just
Society, New York: Addison-Wesley.
*Blinder, A.S. (1988) ‘The Fall and Rise of Keynesian Economies’, Economic Record
December, pp. 278–94.
*Brunner, K. (1989) ‘The Disarray in Macroeconomics’, in F.Capie and G.Wood (eds)
Monetary Economics in the 1980s, London: Macmillan.
Chick, V. (1995) ‘Is There a Case for Post-Keynesian Economics?’ Scottish Journal of
Political Economy 42, February, pp. 20–36.
Chrystal, K.A. and S.Price (1994) Controversies in Macroeconomics, 3rd edn, New
York: Harvester Wheatsheaf.
*Crafts, N. (1996) ‘Post-Neoclassical Endogenous Growth Theory: What are its Policy
Implications?’, Oxford Review of Economic Policy 12, pp. 30–47.
Cross, R. (ed.) (1995) The Natural Rate of Unemployment: Reflections on 25 Years of
the Hypothesis, Cambridge: Cambridge University Press.
Cross, R. (1996) ‘The Natural Rate: An Attractor for Actual Unemployment or an
Attractee?’, Scottish Journal of Political Economy 43, August, pp. 349–64.
Danthine, J.P. and J.B.Donaldson (1993) ‘Methodological and Empirical Issues in Real
Business Cycle Theory’, European Economic Review 37, January, pp. 1–35.
Darity, W. and W.Young (1995) ‘IS-LM: An Inquest’, History of Political Economy 27,
Spring, pp. 1–41.
Davidson, P. (1994) Post Keynesian Macroeconomic Theory: A Foundation for
Successful Economic Policies for the Twenty-First Century, Aldershot: Edward Elgar.
Deutscher, P. (1990) R.G.Hawtrey and the Development of Macroeconomics, London:
Macmillan.
Dixon, H. (1995) ‘Of Coconuts, Decomposition and a Jackass: The Genealogy of the
Natural Rate’, in R.Cross (ed.) The Natural Rate of Unemployment: Reflections on
25 Years of the Hypothesis, Cambridge: Cambridge University Press.
*Dowrick, S. (1992) ‘Technological Catch Up and Diverging Incomes: Patterns of
Economic Growth 1960–88’, Economic Journal 102, May, pp. 600–10.
Eichengreen, B. (1992) Golden Fetters: The Gold Standard and the Great Depression,
1919–1939, New York: Oxford University Press.
Fischer, S. (1977) ‘Long-Term Contracts, Rational Expectations and the Optimal Money
Supply Rule’, Journal of Political Economy 85, February, pp. 191–205.
Fischer, S. (1990) ‘Rules Versus Discretion in Monetary Policy’, in B.M.Friedman and
F.H.Hahn (eds) Handbook of Monetary Economics, Amsterdam: North Holland.
Friedman, M. (1956) ‘The Quantity Theory of Money: A Restatement’, in M.Friedman
(ed.) Studies in the Quantity Theory of Money, Chicago: University of Chicago
Press.
Friedman, M. (1958) ‘The Supply of Money and Changes in Prices and Output’, in
M.Friedman (1969) The Optimum Quantity of Money and Other Essays, Chicago:
Aldine.
Friedman, M. (1961) ‘The Lag in the Effect of Monetary Policy’, in M.Friedman (1969)
The Optimum Quantity of Money and Other Essays, Chicago: Aldine.
*Friedman, M. (1968) ‘The Role of Monetary Policy’, American Economic Review 58,
March, pp. 1–17.
*Friedman, M. (1970) The Counter-Revolution in Monetary Theory, IEA Occasional
Paper no. 33, London: Institute of Economic Affairs.
The development of modern macroeconomics 23
Friedman, M. (1972) ‘Have Monetary Policies Failed?’, American Economic Review
62, May, pp. 11–18.
Friedman, M. (1974) Monetary Correction, IEA Occasional Paper no. 41, London:
Institute of Economic Affairs.
* Friedman, M. (1977) ‘Nobel Lecture: Inflation and Unemployment’, Journal of Political
Economy 85, June, pp. 451–72.
Friedman, M. and A.J.Schwartz (1963) A Monetary History of the United States,
1867–1960, Princeton, NJ: Princeton University Press.
Garrison, R. (1994) ‘Interview with Roger Garrison’, in B.Snowdon, H.R.Vane and
P.Wynarczyk, A Modern Guide to Macroeconomics: An Introduction to Competing
Schools of Thought, Aldershot: Edward Elgar.
*Gerrard, B. (1996) ‘Review Article: Competing Schools of Thought in
Macroeconomics—An Ever-Emerging Consensus?’, Journal of Economic Studies
23, pp. 53–69.
Gordon, R.J. (1981) ‘Output Fluctuations and Gradual Price Adjustment’, Journal of
Economic Literature 19, June, pp. 493–530.
Gordon, R.J. (1982) ‘Price Inertia and Policy Ineffectiveness in the United States, 1890–
1980’, Journal of Political Economy 90, December, pp. 1087–117.
*Gordon, R.J. (1990) ‘What is New Keynesian Economics?’, Journal of Economic
Literature 28, September, pp. 1115–71.
*Gordon, R.J. (1993) Macroeconomics, 6th edn, New York: HarperCollins.
Gordon, R.J. and J.A.Wilcox (1981) ‘Monetarist Interpretations of the Great
Depression: An Evaluation and Critique’, in K.Brunner (ed.) The Great Depression
Revisited, London: Martinus Nijhoff.
*Greenwald, B.C. and J.E.Stiglitz (1993) ‘New and Old Keynesians’, Journal of Economic
Perspectives 7, Winter, pp. 23–44.
Hahn, F.H. and R.C.O.Matthews (1964) ‘The Theory of Economic Growth: A Survey’
, Economic Journal 74, December, pp. 779–902.
*Hammond, J.D. (1996) Theory and Measurement: Causality Issues in Milton
Friedman’s Monetary Economics, Cambridge: Cambridge University Press.
Hansen, A.H. (1953) A Guide to Keynes, New York: McGraw-Hill.
Hicks, J.R. (1937) ‘Mr. Keynes and the “Classics”: A Suggested Interpretation’,
Econometrica 5, April, pp. 147–59.
Johnson, H.G. (1972) ‘Inflation: A Monetarist View’, in H.G.Johnson (ed.) Further
Essays in Monetary Economics, London: Macmillan.
Keynes, J.M. (1936) The General Theory of Employment, Interest and Money, London:
Macmillan.
Keynes, J.M. (1972) Essays in Persuasion, vol. IX, London: Macmillan.
Kydland, F.E. and E.C.Prescott (1977) ‘Rules Rather than Discretion: The Inconsistency
of Optimal Plans’, Journal of Political Economy 85, June, pp. 473–91.
Kydland, F.E. and E.C.Prescott (1982) ‘Time to Build and Aggregate Fluctuations’,
Econometrica 50, November, pp. 1345–70.
*Laidler, D.E.W. (1976) ‘Inflation in Britain: A Monetarist Perspective’, American
Economic Review 66, September, pp. 485–500.
Laidler, D.E.W. (1985) ‘Monetary Policy in Britain: Successes and Shortcomings’, Oxford
Review of Economic Policy 1, pp. 35–43.
Laidler, D.E.W. (1993) The Demand for Money: Theories, Evidence and Problems, 4th
edn, New York: HarperCollins.
Lipsey, R.G. (1960) ‘The Relationship between Unemployment and the Rate of Change
of Money Wage Rates in the UK 1862–1957: A Further Analysis’, Economica 27,
February, pp. 1–31.
*Lipsey, R.G. (1978) ‘The Place of the Phillips Curve in Macroeconomic Models’, in
A.R.Bergstrom (ed.) Stability and Inflation, Chichester: John Wiley.
24 The development of modern macroeconomics
Lucas, R.E. Jr (1972) ‘Expectations and the Neutrality of Money’, Journal of Economic
Theory 4, April, pp. 103–24.
Lucas, R.E. Jr (1973) ‘Some International Evidence on Output-Inflation Tradeoffs’,
American Economic Review 63, June, pp. 326–34.
Lucas, R.E. Jr (1975) ‘An Equilibrium Model of the Business Cycle’, Journal of Political
Economy 83, December, pp. 1113–44.
Lucas, R.E. Jr (1976) ‘Econometric Policy Evaluation: A Critique’, in K.Brunner and
A.H.Meltzer (eds) The Phillips Curve and Labor Markets, Amsterdam: North
Holland.
*Lucas, R.E. Jr (1977) ‘Understanding Business Cycles’, in K.Brunner and A.H. Meltzer
(eds) Stabilization of the Domestic and International Economy, Amsterdam: North
Holland.
*Lucas, R.E. Jr (1980) ‘Methods and Problems in Business Cycle Theory’, Journal of
Money, Credit and Banking 12, November, pp. 696–715.
Lucas, R.E. Jr (1987) Models of Business Cycles, Oxford: Basil Blackwell.
Lucas, R.E. Jr (1988) ‘On the Mechanics of Economic Development’, Journal of
Monetary Economics 22, July, pp. 3–42.
Lucas, R.E. Jr (1993) ‘Making a Miracle’, Econometrica 61, March, pp. 251–72.
Lucas, R.E. Jr (1994) ‘Review of Milton Friedman and Anna J.Schwartz’s A Monetary
History of the United States, 1867–1960’, Journal of Monetary Economics 34, pp.
5–16.
*Mankiw, N.G. (1989) ‘Real Business Cycles: A New Keynesian Perspective’, Journal
of Economic Perspectives 3, Summer, pp. 79–90.
*Mankiw, N.G. (1990) ‘A Quick Refresher Course in Macroeconomics’, Journal of
Economic Literature 28, December, pp. 1645–60.
*Mankiw, N.G. (1994) Macroeconomics, 2nd edn, New York: Worth.
*Mankiw, N.G. (1995) ‘The Growth of Nations’, Brookings Papers on Economic
Activity, pp. 275–326.
Mankiw, N.G. and D.Romer (eds) (1991) New Keynesian Economics, Cambridge:
MIT Press.
Miron, J.A. (1994) ‘Empirical Methodology in Macroeconomics: Explaining the Success
of Friedman and Schwartz’s A Monetary History of the United States, 1867–1960’,
Journal of Monetary Economics 34, pp. 17–25.
Mishkin, F.S. (1982) ‘Does Anticipated Monetary Policy Matter? An Econometric
Investigation’, Journal of Political Economy 90, February, pp. 22–51.
*Modigliani, F. (1988) ‘The Monetarist Controversy Revisited’, Contemporary Policy
Issues 6, October, pp. 3–18.
Muth, J.F. (1961) ‘Rational Expectations and the Theory of Price Movements’,
Econometrica 29, July, pp. 315–35.
Nelson, C. and C.I.Plosser (1982) ‘Trends and Random Walks in Macroeconomic
Time Series: Some Evidence and Implications’, Journal of Monetary Economics 10,
September, pp. 139–62.
*Pack, H. (1994) ‘Endogenous Growth Theory: Intellectual Appeal and Empirical
Shortcomings’, Journal of Economic Perspectives 8, Winter, pp. 55–72.
Phelps, E.S. (1967) ‘Phillips Curves, Expectations of Inflation and Optimal
Unemployment over Time’, Economica 34, August, pp. 254–81.
Phelps, E.S. (1990) Seven Schools of Macroeconomic Thought, Oxford: Oxford
University Press.
Phelps, E.S. and J.B.Taylor (1977) ‘Stabilizing Powers of Monetary Policy under Rational
Expectations’, Journal of Political Economy 85, February, pp. 163–90.
Phillips, A.W. (1958) ‘The Relation between Unemployment and the Rate of Change of
Money Wage Rates in the United Kingdom, 1861–1957’, Economica 25, November,
pp. 283–99.
The development of modern macroeconomics 25
*Plosser, C.I. (1989) ‘Understanding Real Business Cycles’, Journal of Economic
Perspectives 3, Summer, pp. 51–77.
*Prescott, E.C. (1986) ‘Theory Ahead of Business Cycle Measurement’, Federal Reserve
Bank of Minneapolis Quarterly Review Fall, pp. 9–22.
*Romer, C.D. (1993) ‘The Nation in Depression’, Journal of Economic Perspectives 7,
Spring, pp. 19–39.
*Romer, D. (1993) ‘The New Keynesian Synthesis’, Journal of Economic Perspectives
7, Winter, pp. 5–22.
Romer, D. (1996) Advanced Macroeconomics, New York: McGraw-Hill.
Romer, P.M. (1986) ‘Increasing Returns and Long-Run Growth’, Journal of Political
Economy 94, October, pp. 1002–37.
* Romer, P.M. (1993) ‘Idea Gaps and Object Gaps in Economic Development’, Journal
of Monetary Economics 32, pp. 543–73.
*Romer, P.M. (1994) ‘The Origins of Endogenous Growth’, Journal of Economic
Perspectives 8, Winter, pp. 3–22.
Samuelson, P.A. (1955) Economics, 3rd edn, New York: McGraw-Hill.
Samuelson, P.A. and R.M.Solow (1960) ‘Analytical Aspects of Anti-Inflationary Policy’,
American Economic Review 50, May, pp. 177–94.
*Santomero, A.M. and J.J.Seater (1978) ‘The Inflation-Unemployment Trade-Off: A
Critique of the Literature’, Journal of Economic Literature 16, June, pp. 499–544.
Sargent, T.J. and N.Wallace (1975) ‘Rational Expectations, the Optimal Monetary
Instrument and the Optimal Money Supply Rule’, Journal of Political Economy 83,
April, pp. 241–54.
Sargent, T.J. and N.Wallace (1976) ‘Rational Expectations and the Theory of Economic
Policy’, Journal of Monetary Economics 2, April, pp. 169–83.
*Shaw, G.K. (1992) ‘Policy Implications of Endogenous Growth Theory’, Economic
Journal 102, May, pp. 611–21.
*Snowdon, B. and H.R.Vane (1997) ‘Modern Macroeconomics and its Evolution from
a Monetarist Perspective: An Interview with Professor Milton Friedman’, Journal of
Economic Studies 24.
*Snowdon, B., H.R.Vane and P.Wynarczyk (1994) A Modem Guide to Macroeconomics:
An Introduction to Competing Schools of Thought, Aldershot: Edward Elgar.
Solow, R.M. (1956) ‘A Contribution to the Theory of Economic Growth’, Quarterly
Journal of Economics 70, February, pp. 65–94.
Solow, R.M. (1957) ‘Technical Change and the Aggregate Production Function’, Review
of Economics and Statistics 39, August, pp. 312–20.
Solow, R.M. (1986) ‘What is a Nice Girl Like You Doing in a Place Like This?
Macroeconomics after Fifty Years’, Eastern Economic Journal 12, July/September,
pp. 191–8.
*Solow, R.M. (1994) ‘Perspectives on Growth Theory’, Journal of Economic
Perspectives 8, Winter, pp. 45–54.
*Stadler, G.W. (1994) ‘Real Business Cycles’, Journal of Economic Literature 32,
December, pp. 1750–83.
Stiglitz, J.E. (1992) ‘Methodological Issues and the New Keynesian Economies’, in
A.Vercelli and N.Dimitri (eds) Macroeconomics: A Survey of Research Strategies,
Oxford: Oxford University Press.
*Summers, L.H. (1986) ‘Some Sceptical Observations on Real Business Cycle Theory’,
Federal Reserve Bank of Minneapolis Quarterly Review Fall, pp. 23–7.
Swan, T.W. (1956) ‘Economic Growth and Capital Accumulation’, Economic Record
32, November, pp. 334–61.
Taylor, J.B. (1989) ‘The Evolution of Ideas in Macroeconomics’, Economic Record 65,
June, pp. 185–9.
*Tobin, J. (1980) ‘Are New Classical Models Plausible Enough to Guide Policy?’,
Journal of Money, Credit and Banking 12, November, pp. 788–99.
26 The development of modern macroeconomics
Tobin, J. (1987) Policies For Prosperity: Essays in a Keynesian Mode, Brighton:
Wheatsheaf.
*Tobin, J. (1992) ‘An Old Keynesian Counterattacks’, Eastern Economic Journal 18,
Fall, pp. 387–400.
Tobin, J. (1995) ‘The Natural Rate as New Classical Economies’, in R.Cross (ed.) The
Natural Rate of Unemployment: Reflections on 25 Years of the Hypothesis,
Cambridge: Cambridge University Press.
Van de Klundert, T. and S.Smulders (1992) ‘Reconstructing Growth Theory: A Survey’,
De Economist 140, pp. 177–203.
Van Els, P.J.A. (1995) ‘Real Business Cycle Models and Money: A Survey of Theories
and Evidence’, Weltwirtshaftliches Archiv 131, pp. 223–63.
Vercelli, A. (1991) Methodological Foundations of Macroeconomics: Keynes and Lucas,
Cambridge: Cambridge University Press.
Wulwick, N. (1987) ‘The Phillips Curve: Which? Whose? To Do What? How?’, Southern
Economic Journal 53, pp. 834–57.
Yellen, J.L. (1984) ‘Efficiency Wage Models of Unemployment’, American Economic
Review 74, May, pp. 200–5.
Young, W. (1987) Interpreting Keynes: The IS-LM Enigma, Boulder, CO: Westview.
Part I
Keynesian economics and the
Keynesian revolution
Introduction
Keynesian economics, in all its many forms, challenges the legacy of Adam
Smith’s basic theorem that competitive markets are capable of converting the
self-interested behaviour of millions of individuals into a desirable
macroeconomic outcome. To Keynes the ‘invisible hand’ mechanism is
fundamentally flawed in that capitalist market systems seem incapable of
generating the full utilization of societies scarce labour and capital resources
except for limited periods of time. Given that the economic system is subject
to periodic aggregate demand and supply shocks, a key question for
macroeconomic theorists is: ‘Will the economy, once displaced from its full
employment equilibrium, return to that desirable state in a reasonable period
of time via the normal functioning of the price mechanism operating without
assistance from the “visible hand” of government intervention?’ Keynesians
of all persuasions share the view that some degree of selective government
intervention, via fiscal and monetary policies, can improve upon the
‘invisible hand’ inspired non-interventionist stance of the classical economists
and their modern day disciples (see Snowdon et al. 1994; Tobin 1996; Shaw
1997).
Although few, if any, economists have had an impact on the development
of modern macroeconomics to compare with that of John Maynard Keynes,
the essential message of his magnum opus, The General Theory of
Employment, Interest and Money (1936), remains the subject of continuing
controversy some sixty years following its publication. This is hardly
surprising given the impact the General Theory has had in three important
areas. First, the resultant theoretical revolution in the then newly created
field of macroeconomics has inspired an ongoing theoretical debate
concerning the equilibrating properties of the price mechanism. Second, in
parallel to the theoretical revolution, it led to a policy revolution which
represented a major shift in thinking which subsequently encountered
political and intellectual resistance from those who remain wedded to the
classical laissez-faire philosophy associated with the invisible hand doctrine
(see Moggridge 1993; Skidelsky 1996). Third, as Colander and Landreth
(1996) note, it led to a pedagogical (textbook) revolution which reflected the
real need for economists to devise ways of distilling Keynes’s ideas and
30 Keynesian economics and the Keynesian revolution
presenting his theoretical innovations in a format which made them more
accessible to a wider audience. In the latter case the rapid acceptance of
Hicks’s (1937) IS-LM model as the mainstream presentation of Keynes’s
ideas (e.g. Hansen 1953) and Samuelson’s (1948) best-selling textbook were
crucial in the ‘Keynesianization’ of academia. However, one unfortunate
consequence of this pedagogical revolution was that it gave further impetus
to the already growing separation of microeconomics from macroeconomics.
Hence much of the modern debate between new Keynesians and new
classical/real business cycle theorists has arisen from conflicting views about
how to reconcile these two main branches of economics. The burgeoning
literature on the microfoundations of macroeconomics since the early 1970s
is testimony to the legacy of the Keynesian revolution (see Part V). The
ongoing debate relating to the interpretation and relevance of Keynes’s ideas
remains an important source of controversy in macroeconomics and forms a
major theme in the chapters which follow.
Alan Coddington’s 1976 Journal of Economic Literature article,
‘Keynesian Economics: The Search for First Principles’ (reprinted on pp. 36–
54) inquires into the various ways in which economic analysis has been used
in an attempt to come to terms with Keynes’s attack on classical economics
and methods of analysis. Coddington identifies the classical method as
‘reductionist’ because the central idea is to analyse market phenomena on
the basis of rational choices made by individual economic agents. Those
Keynesians who see Keynes’s work as a ‘frontal assault’ on the reductionist
programme are classified as ‘fundamentalists’ by Coddington.
Fundamentalist Keynesians place a great deal of emphasis on Keynes’s 1937
Quarterly Journal of Economics article with its extensive discussion and
focus on the influence of uncertainty, expectations and ignorance.
Fundamentalist Keynesians regard Keynes as having liberated economic
analysis from the strait-jacket of equilibrium theorizing. Coddington’s second
school of Keynesian thought is labelled ‘hydraulic Keynesianism’ and refers
to the mainstream textbook Keynesianism associated with the neoclassical
synthesis. The crucial feature of hydraulic (IS-LM) Keynesianism is the belief
that stable relationships exist between variables at the aggregate level which
provide governments with the potential leverage to influence real income
and employment via fiscal and monetary policy. Coddington shows how this
hydraulic approach to theorizing, by suppressing the role of price
adjustment, is in conflict with reductionist market theory. Finally,
Coddington discusses the contributions of Clower and Leijonhufvud, who
argue that Keynesian economics is best understood within a disequilibrium
trading framework. Their work is classified as ‘reconstituted reductionism’
and is important in that it challenges the neoclassical synthesis view which
had downplayed Keynes’s contribution as a pure economic theorist.
The second article reprinted in this part (pp. 55–94) is Don Patinkin’s
1990 Journal of Monetary Economics article, ‘On Different Interpretations of
the General Theory’. Patinkin argues that it was not until the 1960s that
Introduction 31
major differences among the various interpretations of the General Theory
began to appear. An important question is raised: Why are there so many
‘vastly different’ interpretations of the General Theory? Patinkin attributes
this phenomenon to (1) the non-mathematical style of the General Theory,
(2) its presentation of new and controversial ideas which contained ‘some
obscurities and even inconsistencies’, (3) the failure of Keynes to pull
together an explicit and complete model, this being left to numerous
subsequent interpreters, and (4) the political implications and message of the
General Theory which implied an extension of government intervention in
the economy. Patinkin defends the IS-LM interpretation of Keynes which
became the hallmark of mainstream Keynesianism in the 1950s and 1960s
(see also Patinkin 1990; Darity and Young 1995). In doing so Patinkin
provides a critique of the Post-Keynesian (fundamentalist) interpretation of
the General Theory as well as the more recent reappraisal of Keynes by
Allan Meltzer (1988). Patinkin rejects Meltzer’s contention that Keynes
favoured rules rather than discretion in the conduct of economic policy.
The chapters by Coddington and Patinkin show clearly that Keynes’s
General Theory has given rise to a number of different interpretations and
generated a variety of research programmes. Bill Gerrard, in his article
‘Keynes’s General Theory: Interpreting the Interpretations’ taken from the
March 1991 issue of the Economic Journal (reprinted on pp. 95–108), puts
forward the view that the reason why there is still so much controversy
surrounding Keynes’s General Theory is in part due to ‘different
presuppositions made about the nature of interpretation’. Gerrard argues that
the study of hermeneutics (the study of interpretation) can help pierce the
‘doctrinal fog’ surrounding Keynesian economics. He presents two variants
of the ‘atomistic’ approach to interpretation, namely the objectivist and
relativist approaches. The former seeks to recover the author’s original
message which is hidden due to author-and reader-generated confusion, as
well as disagreement over the stock of relevant textual evidence. The
relativist approach treats interpretation as being determined by the
worldview of the reader. Hence there is no single essential meaning. Gerrard
argues that both varieties of the atomistic view of interpretation are flawed
and suggests a second alternative, the ‘organicist’ approach. Of special
interest, given our earlier discussion of the chapters by Coddington and
Patinkin, is the proposition that the existence of multiple interpretations is
not a problem. Multiple interpretation is taken as evidence of the high
reference power and fertility of the text under examination. Indeed the main
achievement of Keynes’s General Theory ‘was its ability to generate a
diversity of research programmes’. Gerrard concludes that we should ‘stop
worrying about multiple interpretations of Keynes’s General Theory’ as it
represents a strength rather than a weakness of his contribution.
During the 1970s Keynesian economics was generally regarded as
exhibiting strong symptoms of crisis suggesting terminal decline. It was
during this period that new classical economics (see Part III) began to
32 Keynesian economics and the Keynesian revolution
provide penetrating critiques of both the theoretical and empirical
foundations of Keynesian orthodoxy. In a highly entertaining and insightful
article (reprinted on pp. 109–34) published in the December 1988 issue of the
Economic Record, Alan Blinder surveys macroeconomic developments
following the demise of the original stable Phillips curve analysis circa 1972.
Blinder traces out the reasons for the declining influence of the Keynesian
approach to macroeconomic analysis during the 1970s as well as the
resurgence of Keynesian theorizing within academia during the 1980s (see
also Blinder 1987). Blinder begins by discussing the essential characteristics
which in his view identify what it means to be a Keynesian. Of particular
importance is the emphasis placed on aggregate demand disturbances as a
major source of aggregate fluctuations, the significance of wage and price
stickiness, the need to reduce excessive levels of involuntary unemployment
and the potential role for some degree of stabilization policy via ‘coarse
tuning’. Blinder goes on to argue that the fall of Keynesian economics had
more to do with perceived theoretical shortcomings than empirical failure
once the lessons of the expectations-augmented Phillips curve, and supply
shocks of the 1970s, had been absorbed into mainstream macroeconomic
analysis. In addition he identifies certain internal forces within academia
which worked against the prevailing orthodoxy including the attraction that
new classical economics generated for the new generation of technically
sophisticated graduate students and faculty members seeking tenure (see
Snowdon and Vane 1996). Finally, Blinder suggests that a fourth important
factor contributing to the fall of Keynesianism was the resurgence of
Conservative ideology in the United States and elsewhere. When it comes to
explaining the resurgence of Keynesian economics in the 1980s Blinder
draws attention to the growing body of empirical evidence mounted against
the new classical model as well as theoretical deficiencies. At the same time
new theoretical developments within the Keynesian paradigm have resulted
in the reemergence of Keynesian models with stronger microfoundations (see
Part V). Important developments here include the use of imperfect
competition as the basic microfoundation for Keynesian models, the use of
efficiency wage theories to explain involuntary unemployment, the
incorporation of costly price adjustment as an influence on optimal decision
making and hysteresis models of unemployment which question Friedman’s
(1968) natural rate hypothesis (reprinted in Part II, pp. 164–79). As a result
of these and other developments Blinder believes that once more Keynesian
economics is on the ‘ascendancy in academia’.
The final paper reprinted in this part (pp. 135–55) is by one of the world’s
foremost Keynesian economists, James Tobin. In an article entitled ‘Price
Flexibility and Output Stability: An Old Keynesian View’, Tobin presents an
articulate and spirited defence of what he considers to be the essence of
Keynesianism. In proudly declaring himself to be an ‘old Keynesian’ Tobin is
keen to distance himself from the new Keynesian views of economists such as
Gregory Mankiw (see Part V and Tobin 1994). Tobin’s article was originally
Introduction 33
published as part of a symposium on new Keynesian economics in the Winter
issue of the Journal of Economic Perspectives, which also contains articles by
D.Romer (1993), Greenwald and Stiglitz (1993– reprinted on pp. 552–74)
and King (1993). According to Tobin the central Keynesian proposition is not
nominal price rigidity (as stressed by many new Keynesians) but the
principle of ‘effective’ demand. Tobin argues that Keynesian economics in
the spirit of Keynes’s General Theory ‘neither asserts nor requires nominal
wage and/or price rigidity’. Tobin also differs from much new Keynesian
analysis in stressing the importance of real shocks to aggregate demand
rather than nominal disturbances. Given such disturbances Tobin shows how
Keynes provided an explanation of aggregate fluctuations where greater
wage and price flexibility would more than likely turn recession into
depression due to the adverse expectational effects generated during a regime
of deflation. Tobin dismisses the practical relevance of the Pigouvian real
balance (or wealth) effect as an equilibrating mechanism relying as it does
on a highly risky regime of deflation. Tobin concludes by restating the ‘old’
Keynesian policy position; ‘that in the absence of activist “feedback”
policies, monetary and fiscal, flexibility may well be destabilizing, both to
prices and to real macro variables’.
REFERENCES
*Titles marked with an asterisk are particularly recommended for additional reading.
*Abel, A.B. and B.S.Bernanke (1995) Macroeconomics, 2nd edn, Chapter 12, New
York: Addison Wesley.
*Backhouse, R. (1995) Interpreting Macroeconomics: Explorations in the History of
Economic Thought, London: Routledge.
*Backhouse, R. (1997) ‘The Rhetoric and Methodology of Modern Macroeconomics’,
in B.Snowdon and H.R.Vane (eds) Reflections on the Development of Modern
Macroeconomics, Aldershot: Edward Elgar.
Barro, R.J. and V.Grilli (1994) European Macroeconomics, Chapter 21, London:
Macmillan.
Blaug, M. (1990) John Maynard Keynes: Life, Ideas, Legacy, London: Institute of
Economic Affairs.
*Blaug, M. (1994) ‘Recent Biographies of Keynes’, Journal of Economic Literature 32,
September, pp. 1204–15.
*Bleaney, M. (1985) The Rise and Fall of Keynesian Economics, London: Macmillan.
Blinder, A.S. (1987) Hard Heads Soft Hearts: Tough Minded Economics for a Just
Society, New York: Addison Wesley.
*Colander, D.C. and H.Landreth (eds) (1996) The Coming of Keynesianism to America:
Conversations with the Founders of Keynesian Economics, Aldershot: Edward
Elgar.
Darity, W. and W.Young (1995) ‘IS-LM: An Inquest’, History of Political Economy 27,
Spring, pp. 1–41.
Davidson, P. (1994) Post Keynesian Macroeconomic Theory: A Foundation for
Successful Economic Policies for the Twenty-First Century, Aldershot: Edward Elgar.
Fletcher, G.A. (1987) The Keynesian Revolution and its Critics: Issues of Theory and
Policy for the Monetary Production Economy, London: Macmillan.
34 Keynesian economics and the Keynesian revolution
*Friedman, M. (1968) ‘The Role of Monetary Policy’, American Economic Review 58,
March, pp. 1–17.
*Froyen, R.T. (1996) Macroeconomics: Theories and Policies, 5th edn, Chapters 6, 7
and 8, London: Prentice-Hall.
Gerrard, B. and J.Hillard (eds) (1992) The Philosophy and Economics of J.M. Keynes,
Aldershot: Edward Elgar.
*Greenwald, B.C. and J.E.Stiglitz (1993) ‘New and Old Keynesians’, Journal of Economic
Perspectives 7, Winter, pp. 23–44.
Hansen, A.H. (1953) A Guide to Keynes, New York: McGraw-Hill.
Hicks, J.R. (1937) ‘Mr. Keynes and the “Classics”: A Suggested Interpretation’,
Econometrica 5, April, pp. 147–59.
*Hillard, J. (ed.) (1988) J.M.Keynes in Retrospect: The Legacy of the Keynesian
Revolution, Aldershot: Edward Elgar.
*Jansen, D.W., C.D.Delorme and R.B.Ekelund, Jr (1994) Intermediate Macroeconomics,
Chapters 3, 4, 5 and 6, New York: West.
Keynes, J.M. (1936) The General Theory of Employment, Interest and Money, London:
Macmillan.
Keynes, J.M. (1937) ‘The General Theory of Employment’, Quarterly Journal of
Economics 51, February, pp. 209–23.
King, R.G. (1993) ‘Will the New Keynesian Macroeconomics Resurrect the IS-LM
Model?’ Journal of Economic Perspectives 7, Winter, pp. 67–82.
*Mankiw, N.G. (1994) Macroeconomics, 2nd edn, Chapters 9, 10 and 11, New York:
Worth.
Meltzer, A.H. (1988) Keynes’s Monetary Theory: A Different Interpretation, Cambridge:
Cambridge University Press.
*Moggridge, D.E. (1993) Keynes, 3rd edn, London: Macmillan.
Patinkin, D. (1990) ‘In Defense of IS-LM’, Banca Nazionale Del Lavoro Quarterly
Review March, pp. 119–34.
Romer, D. (1993) ‘The New Keynesian Synthesis’, Journal of Economic Perspectives
7,Winter, pp. 5–22.
Samuelson, P.A. (1948) Economics, New York: McGraw-Hill.
*Shaw, G.K. (1988) Keynesian Economics: The Permanent Revolution, Aldershot:
Edward Elgar.
*Shaw, G.K. (1997) ‘How Relevant is Keynesian Economics Today?’, in B. Snowdon
and H.R.Vane (eds) Reflections on the Development of Modern Macroeconomics,
Aldershot: Edward Elgar.
*Skidelsky, R. (1996) Keynes, Oxford: Oxford University Press.
Snowdon, B. and H.R.Vane (1996) ‘The Development of Modern Macroeconomics:
Reflections in the Light of Johnson’s Analysis after Twenty-Five Years’, Journal of
Macroeconomics 18, Summer, pp. 381–401.
*Snowdon, B., H.R.Vane and P.Wynarczyk (1994) A Modern Guide to Macroeconomics:
An Introduction to Competing Schools of Thought, Chapters 1, 2 and 3, Aldershot:
Edward Elgar.
Tobin, J. (1994) ‘Interview with James Tobin’, in B.Snowdon, H.R.Vane and P.
Wynarczyk, A Modern Guide to Macroeconomics: An Introduction to Competing
Schools of Thought, Aldershot: Edward Elgar.
Tobin, J. (1996) Full Employment and Growth: Further Keynesian Essays on Policy,
Aldershot: Edward Elgar.
QUESTIONS
1 Why are there so many different interpretations of Keynes’s General
Theory?
Introduction 35
2 To what extent did Keynes make a distinct break from the classical
economists and their methods of analysis?
3 ‘In the old Keynesian view there are two regimes: the Keynesian regime
where demand creates its own supply and the classical regime where
economic activity is supply constrained’. Explain and discuss.
4 Examine the importance of price flexibility in the Keynes v. Classics debate.
Is price flexibility stabilizing?
5 What were the main reasons for the demise of Keynesianism during the
1970s? To what extent have Keynesian models been rehabilitated in recent
years?
6 How relevant is Keynesian economics today?
7 The era of discretionary demand management is over because the Keynesian
view of governments as benign social welfare maximizers is discredited
beyond repair’. Critically appraise this view.
8 How and why do the ‘fundamentalist’ and ‘hydraulic’ interpretations of
Keynes differ?
9 To what extent did the neoclassical synthesis interpretation of Keynes
introduce an artificial separation of microeconomics from macroeconomics?
10 ‘Creating visions is much more difficult than refining them’. What are the
main features of Keynes’s vision created in the General Theory?
2 Keynesian economics
The search for first principles
Alan Coddington
Journal of Economic Literature (1976) 14, December, pp. 1258–73
INTRODUCTION
In The General Theory of Employment, Interest and Money [13, 1936] and
elsewhere, Keynes attacked a body of theory that he designated ‘Classical’.
In posing a threat to the ‘Classical’ system—or at least to a recognizable
caricature of it—Keynes also called into question the method of analysis by
which this system was constructed. The purpose of this article is to inquire
into the various ways in which methods of economic analysis have come to
terms with this threat, either by responding to it or by reinforcing it with
further threats; it seeks to ask the question: ‘What has to be changed or
sacrificed in order to accommodate Keynesian ideas within standard methods
of analysis?’ Its theme will be the variety of ways in which this may be done:
three broad types will be presented and the contrasts between them explored.
The first task accordingly is to characterize the method of analysis of that
body of theory in opposition to which Keynes presented his own. Very
broadly, this method consisted of analyzing markets on the basis of the
choices made by individual traders. Thus, the resulting theory operates at
two distinct levels—that of individual choice, and that of market
phenomena—even though the connection between the two levels may be
provided only by an analysis of the choices of a ‘representative’ trader.
Moreover, in order to provide a basis for a manageable analysis of market
phenomena, the analysis of individual choice has to be of a particularly
stereotyped and artificial kind. This method of analysis, using market theory
based on choice theory of a type that allows the two levels to be connected,
I will refer to as ‘reductionism’, on the grounds that the central idea is the
reduction of market phenomena to (stylized) individual choices.
Considerations of tractability impose restrictions on the kind of choice
theory on which the market theory can be based: the theory cannot deal with
choices in all the idiosyncratic detail in which actors conceive of them, nor
in terms of the elusive and wayward manner in which actors make up their
minds; stable objectives and well-defined constraints are needed to provide a
Keynesian economics: first principles 37
firm enough foundation for market theory. And just as the choice theory has
to be restricted in the interests of building up to market theory, so the market
theory has to be restricted in the interests of working back to choice theory.
Overwhelmingly, reductionist theorizing has confined its attention to
situations of market equilibrium; for these situations a choice theory basis is
relatively straightforward. There may be, in accordance with the standard
schedules, a gap between market demand and market supply, but the choice
theory from which each of these schedules is derived supposes that all
choices are realizable; accordingly, the standard schedules can tell us
nothing about what will happen when the traders attempt to do what, in the
aggregate, is impossible; nor are the schedules likely to persist as the traders
become aware of the difficulty of doing what they had regarded, in making
their (intended) choices, as straightforward. Overwhelmingly, therefore,
reductionist theory has been concerned with the connection between
equilibrium states of market phenomena and the choice logic from which
these states could be generated. It should be noted, however, that this concern
with market equilibrium is not a defining characteristic of reductionism: it is
rather a way in which reductionist theorizing has been rendered manageable.
FUNDAMENTALIST KEYNESIANISM
If Keynes’s ideas are to be seen as a threat to the reductionist program, the
question naturally arises of how serious a threat they are: of how
fundamental the aspects are that are threatened. Those who have seen
Keynes’s work as a frontal assault on the whole reductionist program, I will
refer to as ‘fundamentalist Keynesians’. It is the purpose of this section to
expound such an interpretation, to consider some of the difficulties in
sustaining it, and briefly to discuss its significance for economic theorizing.
Like the interpretation of the work of any active mind, the interpretation
of Keynes’s writings requires the use of selection and emphasis: it requires a
view as to what is central and what merely peripheral, what is essential and
what merely incidental, in his writings; in this way apparent inconsistencies
and obscurities may readily be resolved, at least to the satisfaction of those
adhering to that interpretation. For fundamentalists, what is central and
essential in Keynes’s writing is to be found primarily in his article The
General Theory of Employment’ [14, 1937] in the Quarterly Journal of
Economics of 1937, an article concisely restating the argument of the
General Theory in response to various critics; in the General Theory itself,
the essence is said to lie in Chapter 12, ‘The State of Long-Term
Expectations’, and, to a lesser extent, in Chapter 17, The Essential Properties
of Interest and Money’. The kind of considerations to be found in these
places can be traced back at least to the work of ten years earlier in The End
of Laissez-Faire [12, 1926] and, with hindsight, still further back.
An early statement of the fundamentalist position was provided by Hugh
Townsend [28, 1937]. He argued that the kind of considerations raised by
38 Alan Coddington
Keynes in his theory of liquidity preference have quite devastating
consequences for reductionist price theory if they are allowed to apply to all
assets. Once all prices are seen as money prices, and all assets as bearing a
liquidity premium, price theory becomes enmeshed in the same tangle of
expectational and conventional elements that characterize Keynes’s theory of
the rate of interest. On this view, the hope of extracting ‘real’ (relative)
prices from their monetary context looks bleak; although we should not, as
Hicks [9, 1946: ch. 12] has pointed out, confuse Keynes’s innovation in
analytical procedure (in dealing with the rate of interest in association with
money-holding decisions rather than with borrowing and lending) with his
substantive contributions. Nevertheless, the threat to the reductionist
program does, on this view, indeed appear to be a fundamental one.
Perhaps the most uncompromising, and certainly the most tirelessly
eloquent, exponent of fundamentalist Keynesianism is G.L.S.Shackle [25,
1967; 26, 1972; 27, 1974]. His own work has centered on the irreducibly
creative element in human choice: its basis in constructs of the choosing
mind. His appreciation of Keynes’s contributions to economic theory has,
accordingly, centered around this same concern, and naturally he sees these
matters of expectation, uncertainty, and ignorance—matters of the provision
of knowledge-surrogates in the face of knowledge deficiency—as of the
essence. A most succinct distillation of Shackle’s reading of Keynes has been
provided by B.J.Loasby [17, 1976].
One further line of thought must be mentioned in the present context: this
is one that has attempted to use the fundamentalist aspect of Keynesianism as
a way of clearing the ground to permit a return to a certain cluster of
doctrines and concerns that are variously referred to as ‘classical’ (as distinct
from ‘neoclassical’) or ‘Neo-Ricardian’. The objective of this school, whose
most distinguished practitioner is Joan Robinson, is to produce a hybrid of
Keynesianism with those aspects of Ricardo’s work that were appropriated
by Marx: Ricardo minus Say’s law and the quantity theory of money.
Keynes’s QJE paper of 1937, to which fundamentalists attach such great
importance, is, first and foremost, an attack on the kind of choice theory that
is required for the reductionist program. As against the clearly specified and
stable objectives and constraints required by reductionist theorizing, Keynes
emphasizes that the basis of choice lies in vague, uncertain, and shifting
expectations of future events and circumstances: expectations that have no
firm foundation in circumstances, but that take their cues from the beliefs of
others, and that will be sustained by hopes, undermined by fears and
continually buffeted by ‘the news’. He was drawing attention to both the
importance and the elusiveness of the state of business confidence, and the
way it unfolds. Keynes focused on the conventional element in valuation: the
way in which valuations may persist to the extent that they are shared, but
are thereby rendered sustainable in the face both of minor events and of
changes in circumstances, but vulnerable to anything that threatens this
conventional basis. In the course of a riot, for example, the moods and
Keynesian economics: first principles 39
feelings of the rioters may be widely shared until, at a later stage when the
riot has lost its force, the moods and feeling may generally and rapidly
revert to normal. The coordination of such crowd behavior and its
characteristic dynamics arise from the fact that the participants are taking
their cues directly from one another. Reductionist choice theory as it has
been developed does not shed any light on decisions involving such
immediate and strong interdependence as this.
Once its choice-theoretic foundations are threatened, the whole
reductionist program is called into question; for without them the market
theory would have nothing on which to stand, nothing to which it could
be reduced. The concept of market equilibrium is in this way left exposed
to attack. For without a clearly-specified and stable basis in choice logic,
the idea of market equilibrium is no longer connected to the realizability
of individuals’ intentions in the aggregate. This does not mean that
market equilibrium cannot be rehabilitated; what it means is that the
sustainability of equilibrium must depend on conditions that are confined
to the level of the market. For the fundamentalist, however, Keynes’s
ideas require the rethinking and reconstruction of the whole body of
reductionist theory: its choice-theoretic basis and the equilibrium theory
of markets that rests on it.
The objections to equilibrium theorizing have been elaborated by
fundamentalist Keynesians. Joan Robinson has shown that if the idea of
equilibrium is pursued relentlessly, then as the concept becomes all-embracing it
becomes paralyzed by its own logic: equilibrium becomes a state of affairs that
is, strictly, unapproachable: unless it already exists, there is no way of attaining
it [20, 1953]. Similarly, in the work of G.L.S.Shackle, the idea of general
equilibrium is shown to require the pre-reconciliation, one with another, of all
present and future choices of all economic actors [26, 1972]. On either ground it
would follow that the standard use of the method of comparative statics (or,
better, ‘comparative equilibria’) to analyze the effects of changes in
circumstances, is strictly unwarranted and illegitimate.1 Of course, this line of
thought would have nihilistic consequences for the entire corpus of economic
theory and in particular for its applicability; in this respect, the line of thought
reaches a purist and impractical conclusion that is in marked contrast to
Keynes’s own highly eclectic approach to economic theory.
The concept of equilibrium is accordingly seen by fundamentalists not as
a useful simplification for economic theorists, but as a distraction.2 The
essence of Keynes’s thought is seen as the liberation from equilibrium
theorizing, as an escape from the restrictions that it imposes on our thinking.
This, however, is not so much a matter of what Keynes said, as of what we
are led to if we follow his line of thought, taking the QJE article as the
definitive guide to its direction.
Where we are led by a line of thought depends a great deal, of course, on
where we are disposed to go. Fundamentalists have, correspondingly,
contributed freely of their own preoccupations in arriving at interpretations
40 Alan Coddington
of Keynes’s thought. At their most uninhibited, fundamentalist Keynesians
have presented Keynes’s ideas as an escape from the essential ‘timelessness’
of the modes of thought he attacked. More concretely, they have presented
his central message regarding employment as concerning the existence of a
liquid asset in a world of uncertainty, thus providing a retreat from the
holding of real assets and the associated commitment to
(employmentgenerating) production of a particular output. This theme has
been much elaborated by Shackle and is concisely expounded by Loasby; in
Joan Robinson’s work, however, we find its place taken by a preoccupation
with the heterogeneity of capital goods: the fact that individual items of the
capital stock that history bequeaths to us cannot be costlessly transformed
into one another, but exist in particular forms, embodying particular
techniques, reflecting the superseded expectations of the past. The problems
raised by the existence of liquid assets and durable, functionally specific
capital assets, are not, however, unrelated; the nature of capital goods means
that holding them involves a kind of commitment, while the nature of
liquidity allows an escape from that particular commitment.
Fundamentalist Keynesianism, in seeing Keynes’s ideas as a wholesale
onslaught on the reductionist program, does not see those ideas as providing
a substitute for that program. Rather, it sees Keynes’s own ideas as a first
step in a thorough-going revision of economic theory. Accordingly, it sees
what Keynes did constructively as merely a makeshift, an improvisation, a
stop-gap. To take the constructive part of Keynes’s work (in developing the
consumption function, the marginal efficiency of capital schedule, etc.) as
being the substance or result of ‘the Keynesian Revolution’ would therefore
betoken a failure of nerve, a betrayal of fundamentalist principles.3
In order to sustain the fundamentalist interpretation, it is necessary to
postulate that Keynes himself had occasional lapses. Thus, Joan Robinson
[23, 1973:3] writes:
there were moments when we had some trouble in getting Maynard to see
what the point of his revolution really was, but when he came to sum it
up after the book was published he got it into focus.
Here she refers, of course, to the QJE article of 1937.
Again, she writes [21, 1964:75]:
The General Theory broke through the unnatural barrier and brought
history and theory together again. But for theorists the descent into time
has not been easy. After twenty years the awakened Princess is still dazed
and groggy.
Keynes himself was not quite steady on his feet.
She then goes on to refer [21, 1964:75] to Keynes’s (‘highly suspicious’)
remark about the timeless multiplier [13, 1936:122].
Keynesian economics: first principles 41
A major embarrassment for fundamentalists is to be found in the final
chapter of the General Theory. Here we find Keynes arguing as follows [13,
1936:378–9]:
if our central controls succeed in establishing an aggregate volume of
output corresponding to full employment as nearly as is practicable, the
classical theory comes into its own from this point onwards. If we suppose
the volume of output to be given, i.e. to be determined by forces outside
the classical scheme of thought, then there is no objection to be raised
against the classical analysis of the manner in which private self-interest
will determine what in particular is produced, in what proportions the
factors of production will be combined to produce it, and how the value of
the final product will be distributed between them.
This is abundantly clear, and in obvious conflict with the fundamentalist
view of Keynes’s thought being subversive of the whole classical
(‘reductionist’) scheme. Accordingly, we find Joan Robinson writing [21,
1964:92], in connection with this passage, of the ‘fallacy’ that Keynes fell
into, and remarking sadly that, ‘He was himself partly to blame for the
perversion of his ideas’ and ‘Keynes himself began the reconstruction of the
orthodox scheme that he had shattered’ [22, 1971: ix].
A further embarrassment for fundamentalists is that Keynes indicated
quite clearly that he found nothing to object to in Hicks’s distillation [8,
1937] of the General Theory into the IS-LM framework, or what has come to
be known as ‘the income-expenditure model’, quite devoid of any
fundamentalist characteristics.4 This again must be seen as some kind of
momentary lapse on Keynes’s part if the fundamentalist interpretation is to
be sustained, at any rate if Keynes himself is to be allowed to be a
fundamentalist Keynesian.
What, then, does fundamentalism add up to? It does not provide any sort
of determinate theory or model of how the economy functions at the
aggregate level; it does not enable one to make any definite predictions
about the likely effects of alternative policies or circumstances. On the
contrary, it is a viewing point from which such constructions would appear
as rather desperate makeshifts of transient applicability. Fundamentalist
Keynesianism is concerned with the texture rather than the direction, as it
were, of the economic process.
To stress the basis of all economic activity in more or less uncertain
expectations is precisely to emphasize the openness and incompleteness of
economic theorizing and explanation. It does not itself provide any kind of
fixed mechanism according to which the unfolding of events takes place; but
it does show how one would set about constructing a narrative of events. It is
a view about where the gaps are in the causal chains that can be identified in
the economy: the points at which the economic process is susceptible to
influence. We can accordingly begin to appreciate the deep ambivalence of
42 Alan Coddington
this standpoint towards economic policy. On the one hand, it sees
potentiality for enormous leverage, the whole economic process moving in
response to changing states of mind and consciousness; on the other hand,
the very precariousness of this vision leads very naturally to thoroughgoing
scepticism about the predictability of the effects of deliberate attempts to
apply leverage in pursuit of political objectives. The point of view in itself
provides no guidance on whether the precariousness is so pervasive as to
undermine the potential for political leverage. That is to say: the wayward
and unruly character of individual choices—and in particular investment
decisions—is seen as an impediment to economic functioning; but the
question that must be faced from a policy point of view is whether it is a
greater impediment to the self-regulation of the economy than it is to the
workings of discretionary fiscal and monetary policy. This matter would
involve not just the consideration of an impediment to economic functioning,
but a comparison between its inhibiting effects on alternative modes of
economic regulation. More broadly, the comparison also arises among the
alternative effects of investment decisions taken within alternative
institutional frameworks (various powers and responsibilities having been
given to agencies of the State), whose regulative capacities then also become
a part of the appropriate comparison.
In summary, we can say that fundamentalist Keynesians are united in
seeing Keynesian ideas as posing a threat to the whole reductionist program;
and that their primary concern has been to reinforce this threat with further
threats. When it comes to providing an alternative to the reductionist
program, however, matters are less unified. There is a marked contrast, for
example, between the prospectus offered by Joan Robinson for the
completion of the Keynesian revolution and the insight offered by Shackle
into its integrity and essence. And when we move from the critical to the
constructive aspects of fundamentalism, not only are matters less unified,
they are also less definite. In Loasby’s work, this indefiniteness is
transformed into a methodological principle [17, 1976:167]:
If one can summarise in one sentence the theory of employment set forth
by Keynes in his [QJE] article of 1937, it is this: unemployment in a
market economy is the result of ignorance too great to be borne. The
fully-specified macroeconomic models miss the point—which is precisely
that no model of this situation can be fully specified.
HYDRAULIC KEYNESIANISM
During the 1940s and 1950s, there appeared a number of expositions of
‘Keynesian economies’, attempting to make the ideas accessible to students,
and even to intelligent laymen. What these works had in common, quite
apart from matters of substance, was an unmistakable enthusiasm for (what
were taken to be) Keynes’s ideas. This enthusiasm was at times unrestrained
Keynesian economics: first principles 43
to the point of excitement; it was the authors of these works who spoke
without reservation of a ‘Keynesian Revolution’, one of the books in fact
having this title [15, Lawrence R.Klein, (1949) 1966]. It is some indication of
the level of enthusiasm reached by these expositors and popularizers that
one, Jan Pen, wrote a book setting out and discussing a particular
specification of a static ‘Keynesian’ model of relationships between a small
number of macroeconomic aggregates and gave it the title Modern
Economics [19, 1965]. It is not my intention here, however, to attempt to
chart the process of the diffusion and popularization of Keynes’s ideas.5
The period of Keynesian enthusiasm was really the post-war period: the
ideas went cantering briskly through the 1950s and early 1960s; faltered
sometime in the mid-1960s and stumbled into the 1970s.6 This, at any rate,
is the picture as its emerges at the level of popular influence, at the level of
widely and influentially held views on macroeconomic policy; at the level,
that is, of Keynesianism as a doctrine about how a largely decentralized
economy may be subject to broad (as opposed to detailed) central control or
influence through the instrument of the budget. It is tempting to adopt the
practice of referring to this doctrine as ‘fiscalism’ to show that it is a
particular variant (and perhaps a corruption or vulgarization) of Keynes’s
ideas. At any rate, it is important to keep distinct the ups and downs of
Keynesianism as a policy doctrine from those of Keynesianism as an
academically respectable theory of the functioning of a capitalist economy at
the aggregate level.7 Indeed, the esteem in which the two aspects have been
held has tended to move in opposite directions, the period when ‘fiscalist’
policy enthusiasm was at its height being a time at which the intellectual
interest in the underlying theory had become moribund. Again, the demise of
‘fiscalism’ in the late 1960s and early 1970s was accompanied by a
reawakening of interest in the underlying theoretical conceptions. (We shall
have more to say about this revival in the next section.)
All this should not be allowed to give the impression, which would be
quite mistaken, that the fiscal enthusiasm stemming from Keynes’s ideas did
not include, or could not provide, a theory in support of its policy doctrine.
It could and it did. What, then, we are led to ask, is the theoretical basis for
fiscalist enthusiasm? How is it to be characterized as one of the strands in
the development of Keynesian thought? It is to these questions we now turn.
The theoretical content of the body of ideas that has been propagated
through the educational system in the West since World War II as ‘Keynesian
Economies’ (by, for example, Paul Samuelson’s pedagogically authoritative
textbook [24, 1973]) I shall proceed to refer to as ‘hydraulic Keynesianism’.
This designation reflects the view that the natural and obvious way to regard
elementary textbook Keynesianism is as conceiving of the economy at the
aggregate level in terms of disembodied and homogeneous flows. Of course,
conceiving of the macro-economy in this way will be fruitful only to the
extent that there exist stable relationships between these overall flows. And it
is my contention that the central characteristic of ‘hydraulic Keynesianism’ is
44 Alan Coddington
the belief that such stable relationships do exist at the aggregate level. It is
this belief that gives some point to the hydraulic conception; without such a
belief the conception would simply be a matter of national income
accounting, not of economic theory.
It should be noted that the flows involved in this conception are flows of
expenditure, income or output. That is to say, neither prices nor quantities
per period make a separate appearance: they appear inextricably in the
contribution each makes to the overall flows of spending and receipts. It
should now be apparent why the belief in the existence of, and the attempt to
establish, stable relationships between the overall flows is radically
inconsistent with reductionism. For any reductionist program must give a
crucial role in its theorizing to prices as such (not to the contribution they
make to overall spending flows). The grounds for this view are that it is
prices as such that provide the incentives that individuals face in making the
choices on which the whole scheme is to rest. This does not mean that
hydraulic Keynesianism can allow no part at all to be played by prices;
when we come to think of such prices as embracing wage rates and interest
rates, we can see that this cannot be so. Correspondingly, it does not mean
that reductionism is incapable of allowing overall flows to play any part in
its scheme. Since these are alternative programs for theorizing, rather than
alternative theories, they revolve around matters of emphasis. They do not
concern what can or cannot play a part in a theory, but what can or cannot
play a central part.
In fact, contrary to the standpoint associated with reductionism, hydraulic
Keynesianism is a scheme in which there is only one agency making
deliberate acts of choice; that one agency is ‘the government’. And it is the
belief that there are indeed stable relations among the various overall flows
in the economy that provides a basis for ‘the government’ to pursue its policy
goals regarding the overall level of economic activity and hence, relatedly,
of the level of employment. It is the stability of these aggregate relationships
that provides ‘the government’ with the leverage it needs to influence those
flows that are not under its direct control. By making deliberate choices for
the flows it does control (via the budget), and bearing in mind the (allegedly)
stable relationships between this and the other flows that are objects of
concern for economic policy, ‘the government’ can, in principle, exercise an
indirect control on the overall level (although not the composition) of the
flows that are not the objects of anyone’s deliberate choice. That is the story.
On the face of it, it may appear a major triumph in the march of human
reason: a dramatic and irreversible extension of the boundaries of political
responsibility. Instead of unemployment and depression being seen and
accepted passively, like the weather, they are to be seen as matters for
human will and design, something that human agency, through the
instrument of central government, could actually resist and remedy.8 As an
idea it looked both simple and good; accordingly, it was, at the end of the
Keynesian economics: first principles 45
war, rapidly assimilated to both the policy statements and rhetoric of all
major political parties.9
In summary, it can be seen that the hydraulic approach is in conflict with
reductionist market theory. The hydraulic approach shows how things would
work when market prices (and wages) will not, or will not quickly enough,
or will not be allowed to, perform their allocative role; it analyzes a
situation in which prices are failing, both as disseminators of information
about relative scarcities and in the provision of incentives to act on the basis
of that information.
If the central message of the General Theory is that overall employment is
more a matter of the demand for output than of real wages, except when ‘full
employment’ already obtains, then that message is certainly embodied in the
hydraulic approach. As such, it is an audacious simplification, which is, on
the face of it, in conflict with the corpus of reductionist theorizing.
Furthermore, as a way of thinking about macroeconomic policy, it seems to
work to some extent, sometimes. The intellectual problem that it raises,
however, is that of its own scope. What we need to know are the
circumstances in which, and the extent to which, the operation of an
economy may be conceived of in hydraulic terms. There are various
approaches to this question. A familiar one is provided by the IS-LM
apparatus, within which it can be readily shown that the economy exhibits
the characteristics of the hydraulic model to the extent that the interest
elasticity of expenditure is low and of the demand for money is high; with a
zero interest elasticity of expenditure and an indefinitely large interest
elasticity of demand for money, the operation of the economy would be
exactly in accordance with the hydraulic model: changes in expenditure
flows would lead to changes in output flows without any repercussions on the
rate of interest. In sum, it follows that the economy may exhibit the
characteristics of the hydraulic model to the extent that the interest rate is
impeded, for whatever reason, in its attempts to respond to changes in
expenditure.
Since the IS-LM apparatus was put forward by Hicks, however, we have
had something like 30 years’ experience of demand management policies
based on the assumption that the economy exhibits marked hydraulic
characteristics in the short-run; and the question of why these policies have
been less effective at some times than others naturally raises in a practical
way the question of the scope of the hydraulic conception. It is therefore of
considerable interest that Hicks, in a revision of Keynesian economics in the
light of this experience, does not adopt his own IS-LM apparatus for the
purpose [11, 1974: ch. 1]. Rather, he provides an alternative framework in
which the possibility of an expansion in demand being translated into an
expansion of output depends crucially on the structure of inventories at the
outset of the process. In particular, it depends crucially on there being
plentiful stocks of materials to sustain investment projects until decisions to
increase the output of these materials are taken; or what amounts to the
46 Alan Coddington
same thing, it depends on there being ample foreign exchange reserves
representing command over foreign inventories of materials. Unless this
condition is met, the attempted expansion can easily run up against
bottlenecks and dissipate itself in the diversion of resources from other uses
and, notoriously, in creating balance of payments problems. Of course, if the
increase in investment expenditure is translated into a net increase in real
investment, the multiplier process can set in, and adequate stocks of
consumer goods will then be required to avoid bottlenecks at this stage and
sustain the expansion.10
There are, of course, other approaches to the question of the scope of
hydraulic theorizing. Indeed, the Monetarist arguments against Keynesian
conclusions may be seen as one possible answer to this question: namely,
that the scope of hydraulic theorizing is practically nonexistent. In these
arguments the Keynesian conclusions are undermined by the reintroduction
of a choice-theoretic basis of the standard reductionist type. As we shall see
in the next section, the work of Clower and Leijonhufvud may also be seen
as contributing to this question of scope, although this is not how either of
them presented his work.
RECONSTITUTED REDUCTIONISM
During the 1960s there emerged a school of thought, associated primarily
with the names of Robert W.Clower [2, 1969] and Axel Leijonhufvud [16,
1968], concerned with reappraising Keynes’s contribution to economics.
These writers presented their work as concerned with reestablishing and
reasserting the discontinuity between Keynesian economics and its
alternatives, a discontinuity that they saw as having been blurred and finally
lost to view by the various activities of interpretation, condensation, and
reconstruction that came in the wake of the General Theory. It is within this
perspective accordingly that the contribution of Clower and Leijonhufvud to
our understanding of Keynes has been discussed and appraised. My purpose
here, however, will be to present the dispute between Clower and
Leijonhufvud, on the one hand, and those whose views they were combating,
on the other, as a family quarrel within the reductionist program. Most
fundamentally, the family quarrel is about the expendability of the concept
of equilibrium: the Clower-Leijonhufvud position being that the concept of
equilibrium should be abandoned in the interests of a more thorough-going
reduction of Keynesian ideas to choice logic. The thesis is that once
equilibrium has been abandoned and one focuses on a process of trading at
disequilibrium prices, then one has a framework that is entirely congenial to
Keynesian ideas, unlike the framework of equilibrium theorizing which, on
this view, leaves room for them in only the most attenuated and ad hoc
form. The problem then becomes one of providing a more sophisticated
specification of the constraints on individual choices, opening up the
possibilities for theoretically novel and challenging forms of market inter-
Keynesian economics: first principles 47
dependence arising from a schematization of the process of disequilibrium
trading.
In order to lead up to my characterization of the work of Clower and
Leijonhufvud, it is appropriate to begin by discussing each writer’s own
characterization of his work: how each of them conceived of the task he had
set himself. I will argue that their own characterizations are in various
respects unsatisfactory, and that my alternative is not therefore gratuitous. I
shall not, however, attempt to substantiate the designation of the work of
these writers as reductionist. I shall take it that once the idea of what is
involved in the reductionist program is appreciated, it should be clear that
this work falls within the program.
Let us take Clower first. Having advanced the ‘dual decision hypothesis’
as a basis for expecting consumer spending to depend on current income,
Clower goes on to speak unguardedly of Keynes having had this theory of
household behavior ‘at the back of his mind when he wrote the General
Theory’ [2, 1969:290]. Clower goes on immediately to admit that ‘I can find
no direct evidence in any of his writings to show that he ever thought
explicitly in these terms.’ After advancing what he takes to be ‘indirect
evidence’ for this, he concludes that ‘Keynes either had a dual-decision
hypothesis at the back of his mind, or most of the General Theory is
theoretical nonsense.’ The picture here seems to be one of Keynes with a
mind full of ideas, some of which he got onto the pages of the General
Theory, the task being to work out what the remainder must have been. This
is a problem of reading not so much between the lines as off the edge of the
page. In his conclusion, however, Clower maintains, rather more soberly,
that his purpose has been ‘simply to clarify the formal basis of the Keynesian
revolution and its relation to orthodox thought’ (emphasis added) [2, 1969;
295]. This then leaves the task quite up in the air, for it is not explained to
the reader how this relates to the previous concern with what Keynes had ‘at
the back of his mind’.
Turning to Leijonhufvud, we find that he is at some pains to try to make
clear the task he has set himself. First, he makes it plain that the doctrine-
historical question of ‘what Keynes really said’ is a strictly secondary matter
for his purposes [16, 1968:9]. ‘The primary purpose,’ he explains, ’remains
…to provide a fresh perspective from which the income-expenditure theory
may be reconsidered’ [16, 1968:9–10]. (The ‘income-expenditure theory’ is
Leijonhufvud’s label for the ‘conceptual framework which has crystallized
out of the debate triggered by the General Theory’ [16, 1968:6].) This seems
straightforward enough. The difficulty arises because what was presented
was not just ‘Leijonhufvud’s fresh perspective’, but rather the fresh
perspective that Leijonhufvud claimed to have distilled from the General
Theory itself. On the face of it, the task appears to be to get a perspective on
the whole debate by going back to the origins of it. But the question arises of
how the responsibility for this new perspective is to be apportioned between
Keynes and Leijonhufvud. Keynes may well have provided the inspiration for
48 Alan Coddington
the task, but if the product of the distillation is to be presented as a (purified)
‘Economics of Keynes’ to be contrasted with the (corrupted) ‘Keynesian
Economies’, then we are back in the realms of mind-reading, especially as
this ‘Economics of Keynes’ can be read into the General Theory only with
what seems to me to be a great deal of ingenuity and determination. So
although Leijonhufvud at first seems to be concerned with the rather modest
task of finding a fresh perspective from which the development of Keynesian
Economics can be surveyed or appraised, it turns out that he is in search of
the one perspective from which the Keynesianness of these developments can
be judged. What looks at first like a search for new angles turns out to be a
search for authenticity.
But it is not just a matter of authenticity, for the fundamental presumption
that underlies the work of Clower and Leijonhufvud is that Keynes said
something important, not only for economic policy, but for economic theory.
They are saying: ‘Let us read the General Theory in a search for theoretical
innovation.’ In other words, far from being engaged in disinterested exegesis
(as the concern for authenticity might suggest), they were concerned with
reworking with a view to rejuvenating (by which standards they must be
judged to have had some success).
How, then, is the task that Clower and Leijonhufvud set themselves to be
expressed and understood? The view I want to advance is that they were
setting themselves the task of constructing a framework that would provide
room or scope for Keynesian ideas. This quite rightly takes it for granted
that we already have a good rough idea what Keynesian ideas are: of what
the General Theory was driving at. What was wanted was a theoretical
niche in which what were taken to be Keynes’s insights could take root and
thrive. The motive for this search was evidently the recognition that the
framework of general equilibrium theory that had been widely adopted for
attempts at precise expression of Keynesian ideas leaves practically no room
or scope for them: they may appear in only the most attenuated and ad hoc
form.
On its own terms, then, the essence of the Clower-Leijonhufvud position is
this: that in order to accommodate Keynesian ideas, we have to abandon
equilibrium theorizing and address ourselves to an understanding of the
process of disequilibrium trading. In my terms, however, it is not just
equilibrium theorizing that has been shown to be uncongenial to Keynesian
ideas, but rather equilibrium theorizing within the reductionist program. And
one can see why this should be so without even taking any detailed view
about the workings of the economy. For within reductionism everything boils
down to acts of choice within a well-specified system of objectives,
constraints and forms of interdependence; and in equilibrium theorizing we
confine our attention to situations in which all the independently arrived at
choices can be simultaneously realized. It then follows rather naturally,
irrespective of any details of market forms or institutional arrangements, that
such a system leaves no room for the ‘unintended’ and ‘involuntary’: for
Keynesian economics: first principles 49
malfunctioning and disorder. It follows, however, from my characterization
of such theorizing that there are two distinct possibilities for the
accommodation of Keynesian ideas: (i) the abandonment of equilibrium and
(ii) the abandonment of reductionism. Clower and Leijonhufvud consider
only the former possibility. We can see, however, that the claim that
equilibrium theorizing must be abandoned in order to accommodate
Keynesian ideas postulates that theorizing must be carried out in accordance
with the reductionist program; but this is something that Clower and
Leijonhufvud simply take for granted.
The whole question of whether Keynesian ideas should be accommodated
by abandoning equilibrium theorizing rather naturally raises the question of
what use Keynes himself made of the concept of equilibrium.11 It is certainly
true that Keynes made use of the term ‘equilibrium’. But before we conclude
that if Keynes could express his ideas in these terms then they must be
perfectly compatible with equilibrium theorizing, we must pause to consider
the meaning of equilibrium and the uses to which an equilibrium concept
might be put. We must bear in mind that it is entirely in keeping with
Keynes’s eclecticism that his use of the term equilibrium could have been a
rather desperate improvisation at one stage in the ‘long struggle of escape’.
An equilibrium is a configuration which, once attained, will be
maintained provided the underlying circumstances (formally, the parameters
and exogenous variables) remain unchanged. Accordingly, the interest and
usefulness of an equilibrium construction, as an end in itself, depends on a
question which is, in principle, an empirical one, namely: what is the range
of variability of the underlying circumstances over the order of magnitude of
the time involved in adjusting (near enough) to its equilibrium
configuration?12 That is to say, if the underlying circumstances are fairly
stable relative to the speed of adjustment of the endogenous variables, the
equilibrium configuration of the system becomes a matter of some interest in
itself and may provide a reasonably useful substitute for becoming involved
in the complexities of the adjustment process. It is something to know where
we are heading, provided we have some grounds for believing that we will
get most of the way there before we start heading somewhere else.
It is in the light of these considerations that we can say why Keynes’s use
of equilibrium constructions was a peculiar one: He was concerned with
discussing, among other things, the instability of the underlying
circumstances of his construction. That is, one of his focuses of interest was
precisely the failure of his equilibrium construction to satisfy the conditions
for the routine usefulness of an equilibrium construction. Therefore, in
arriving at an appreciation of Keynes’s method, it is not enough to ask the
nature of his construction; we must enquire also into its mode of animation.
When we have reason to expect relatively stable underlying circumstances,
the construction may be animated according to the method of comparative
statics. When the animation is endemic, when one is concerned, as it were,
with the restlessness of the underlying circumstances, the use of the
50 Alan Coddington
construction becomes less straightforward, and certainly less mechanical.
Whether, in this case there is anything much left of the concept of
equilibrium is a matter of no particular importance. What is important is to
see that, just as one does not expect to quell a riot by taking a photograph of
it, neither did Keynes’s makeshift use of the equilibrium concept involve the
expectation that he could freeze the economy in a particular state. Shackle
has expressed this idea with characteristic elegance [25, 1967:182]:
At each curtain rise the General Theory shows us, not the dramatic
moment of inevitable action, but a tableau of posed figures. It is only after
the curtain has descended again that we hear the clatter of violent scene-
shifting.
We have seen that Clower and Leijonhufvud’s version of Keynesianism is a
reconstituted reductionism: it addresses itself not to the state of equilibrium,
but to the problem of attaining it.13 It asks the question how a decentralized
market economy might, with some degree of effectiveness, perform the task
that the Walrasian auctioneer would perform smoothly. To ask this question,
one needs a construction in which prices adjust less than instantaneously to
economic circumstances, so that at any point in time the prices may be
effectively providing incentives to act, but the information they reflect will
not be appropriate for the equilibrium that is being approached.
Now it may well be that formulating this question raises some of the most
profound questions in macro-and monetary economics; but we are still in
need, for the practical deployment of Keynesian ideas, of a usable
simplification such as the hydraulic approach provides. And the use of such a
simplification will require an awareness of the circumstances under which it
may be expected to work tolerably well: an awareness of its scope. This is
where a reconstituted reductionism may play a part. For in order to examine
the scope of a theory in which prices fail altogether to play their (ideal)
allocative role, one needs a theory in which there is a partial failure in this
respect. This latter theory could then be used to interpret the practical
successes and failures of the hydraulic approach: as a way of trying to
distinguish the circumstances conducive to its being an adequate
simplification. Accordingly, we should see Leijonhufvud’s book as not so
much about the economics of Keynes as about the scope of the economics of
Keynes. Clower and Leijonhufvud claim to have shown that Keynes was
trying to adapt the reductionist method to the expression of his own ideas by
refocusing it on situations of market disequilibrium. But in displaying the
analytical unmanageability of such a program, they make it clear that,
insofar as Keynes was able to come to any definite conclusions about
economic functioning, he must have short-circuited such problems.
Within the hydraulic approach, employment problems are quite distinct
from allocation problems; they arise at the aggregate level, and they are
independent of relative prices and the composition of demand or output. The
Keynesian economics: first principles 51
thrust of the reconstituted reductionist approach, however, is to present
unemployment as a by-product or even a species of allocation problem. But
if this formulation does not set any definite limits on the scope of the
hydraulic simplification, all it can suggest is a general scepticism regarding
the appropriateness of aggregate tools to deal with problems that are seen as
involving the internal composition of those aggregates; this, however, adds
nothing to what we already know, namely that the hydraulic approach is a
simplification and abstracts from allocation problems. The question that still
remains is essentially a question of decomposability. It is the question of the
separability of employment problems from the allocation problems on which
they are, in practice, superimposed. To what extent may we disregard the
allocative structure of macroeconomic aggregates? Just how blunt an
instrument is demand management? If the reconstituted reductionist
approach could be made tractable without collapsing into the Monetarist
simplification, it could be expected to shed some light on these matters (as
indeed the Monetarist simplification itself has done).
CONCLUSION
In this paper we have considered three varieties of Keynesianism: the
fundamentalist, the hydraulic, and the reconstituted reductionist approaches.
Each one has been located in relation to the reductionist program: the
fundamentalist approach by its rejection of the choice theory that is essential
to and the (equilibrium) market theory that is typical of reductionist
theorizing; the hydraulic approach by its short-circuiting of reductionist
market theory and its eschewal of formal choice theory foundations; and the
reconstituted reductionist approach by its attempt to make room for
Keynesian ideas within the reductionist program by refocusing the market
theory on disequilibrium states whilst retaining the standard choice-theoretic
foundations.
It remains only to make some comments on the relationship of the
approaches to one another; the thrust of these comments will be that the
various approaches are, in their contribution to understanding, largely
complementary.
The fundamentalist approach provides a very general critique of the
methods of reductionism with regard to both its style of choice theory and
the equilibrium theory of markets with which it is typically associated. As
such it clears the ground for the introduction of Keynesian ideas; at the same
time it forms a kind of backdrop against which hydraulic thinking can
thrive, and, as it turns out, reductionism can reappear in a modified form.
Hydraulic thinking can thrive because, in the absence of standard
reductionist results, one needs some drastic simplification in order to say
anything at all definite regarding forecasting or policy. (The alternative
candidate is the drastic simplification provided by the quantity theory of
money and its modern variants.) Reductionism can then reappear because, in
52 Alan Coddington
making use of a drastic simplification, one is led to ask questions about its
scope and limits; these questions will concern why the economy may not
work in the way that standard reductionist theory indicates and are questions
that could be formulated in a modified and expanded reductionist
framework.
Thus, the fundamentalist approach clears the ground for Keynesian ideas,
the hydraulic approach provides the dangerous simplification that makes
them at all definite and manageable, and a loosened reductionism provides
the reservations and qualifications that provide guidance on the scope of this
simplification. The matter may be expressed cryptically in terms of Keynes’s
‘long struggle of escape’. We may say that what he escaped from was
(unreconstituted) reductionism; what he escaped to was the hydraulic
approach; and what he went through in the process of struggle has been
preserved in the fundamentalist approach. For a generation brought up on
Keynesian ideas, however, a sense of intellectual liberation is far more likely
in the struggle of escape from hydraulic thinking into a reconstituted form of
reductionism. In treading this particular path, Clower and Leijonhufvud were
quite right to identify their work with that of Keynes; they differ from him
only in their direction of travel.
ACKNOWLEDGEMENTS
I would like to acknowledge various forms of indebtedness in connection
with this chapter: to, among others, Thanos Skouras, Alan Peacock, David
Currie, and an anonymous referee of this journal for helpful comments on an
earlier draft; to Michael Kennedy and Colette Bowe for indispensable
guidance at various stages; and to the University of Manchester for the
Hallsworth Fellowship in Political Economy, which made this work possible.
NOTES
1 This argument is elaborated in Coddington [3, 1975] and Loasby [17, 1976: ch. 3].
2 Thus: The argument stops when…the equilibrium lullaby hushes further inquiry’
[Robinson, 21, 1964:80]. But this soporific effect is never reconciled with the
concurrently held view that ‘The concept of equilibrium, of course, is an
indispensable tool of analysis’ [21, 1964:78].
3 An immediate difficulty for fundamentalists is the fact that the QJE article of
1937, after having advanced the arguments already discussed, goes on to stress
the importance of the consumption function, which is then deployed (anticipating
terminology I will introduce at a later stage) in a thoroughly hydraulic fashion.
4 See Keynes’s letter of 31 March 1937, to Hicks in Hicks [10, 1973:9–10].
5 But see John Kenneth Galbraith’s ‘How Keynes came to America’ for some
interesting insights into the way Keynesian ideas made their entry into the US
academic economics establishment [5, 1971].
6 For an attempt at intellectual stock-taking at that time, see my ‘Rethinking Economic
Policy’ [4, 1974].
Keynesian economics: first principles 53
7 Reflecting on the fragmentation of Keynesian thought, Axel Leijonhufvud makes
the following observation: ‘For some time now, contentment with this state of the
arts has rested on the motto “The Theoretically Trivial is the Practically Important
and the Practically Important is the Theoretically Trivial.” It is a disturbing formula
which can hardly be a permanent basis for the further development of the field’
[16, 1968].
8 This changed attitude did not come easily or quickly, and fundamental attitudes
had been undergoing a process of erosion for some decades by the time Keynes
came on the scene. For a painstaking documentation of this process in Britain, see
José Harris [7, 1972].
9 The major bridge in Britain between Keynesian doctrines and political platforms
was William Henry Beveridge [1, 1944]. The ideas were given official recognition in
the White Paper Employment Policy [6, 1944].
10 This analysis can readily be transformed from a ‘fixprice’ basis to a ‘flexprice’ one;
in which case the precondition for a successful expansion becomes that the prices
of the various goods needed to sustain the expansion while changes in expenditure
are being translated into changes in output are significantly below normal, so that
as the expansion proceeds traders will release their stocks onto the market as
prices rise. If this condition is not satisfied, the expansionary impetus will be
wholly dissipated in price increases [11, Hicks, 1974:23–30].
11 For a detailed exegesis of this point, see Don Patinkin [18, 1976:113–19].
12 We are here avoiding the large question of whether the system may approach an
equilibrium configuration without shifting the equilibrium that is being approached.
13 In order to do this, Clower and Leijonhufvud avoid Joan Robinson’s ultra-strict
logic of equilibrium according to which the equilibrium state is unapproachable
and hence the problem of attaining it insoluble.
REFERENCES
1 Beveridge, William Henry [Sir], Full Employment in a Free Society, London: Allen
and Unwin, 1944.
2 Clower, Robert W., ‘The Keynesian Counter-Revolution: A Theoretical Appraisal’,
in Monetary Theory, edited by Robert W.Clower, Harmondsworth: Penguin, 1969.
3 Coddington, Alan, ‘Creaking Semaphore and Beyond: A Consideration of Shackle’s
“Epistemics and Economies”’, British Journal for the Philosophy of Science 1975,
26(2), pp. 151–63.
4 ——‘Re-thinking Economic Policy’, Political Quarterly Oct.-Dec. 1974, 45(4), pp.
426–38.
5 Galbraith, John Kenneth, A Contemporary Guide to Economics, Peace and
Laughter, Boston, MA: Houghton Mifflin; London: André Deutsch, 1971.
6 Great Britain Parliament, Employment Policy, Cmd 6527, London: HMSO, May
1944.
7 Harris, José, Unemployment and Politics: A Study in English Social Policy, 1886–
1914, Oxford: Clarendon Press, 1972.
8 Hicks, John R., ‘Mr. Keynes and the “Classics”: A Suggested Interpretation’,
Econometrica April 1937, 5, pp. 147–59.
9 ——Value and Capital, 2nd edn, Oxford: Clarendon Press, [1939] 1946.
10 ——‘Recollections and Documents’, Economica Feb. 1973, 40(1), pp. 2–11.
11 ——The Crisis in Keynesian Economics, New York: Basic Books, Oxford: Basil
Blackwell, 1974.
54 Alan Coddington
12 Keynes, John Maynard, The End of Laissez-faire, London: Woulf, Hogarth Press,
1926.
13 ——The General Theory of Employment, Interest and Money, London: Macmillan,
1936.
14 ——‘The General Theory of Employment’, Quarterly Journal of Economics Feb.
1937, 51(2), pp. 209–23.
15 Klein, Lawrence R., The Keynesian Revolution, London: Macmillan, 1949; 2nd
edn, 1966.
16 Leijonhufvud, Axel, On Keynesian Economics and the Economics of Keynes, New
York: Oxford University Press, 1968.
17 Loasby, Brian J., Choice, Complexity and Ignorance, Cambridge: Cambridge
University Press, 1976.
18 Patinkin, Don, ‘Keynes’ Monetary Thought: A Study of its Development’, History
of Political Economy, Spring 1976 8(1), pp. 1–150.
19 Pen, Jan, Modern Economics. Translated from the Dutch by Trevor S.Preston,
Harmondsworth: Penguin, 1965.
20 Robinson, Joan, ‘The Production Function and the Theory of Capital’, Review of
Economic Studies 1953–54, 21(2), pp. 81–106.
21 ——Economic Philosophy, Chicago: Aldine, 1962; Harmondsworth: Penguin,
1964.
22 ——Economic Heresies, New York: Basic Books; London: Macmillan, 1971.
23 ——‘What has become of the Keynesian Revolution?’, in After Keynes, edited by
Joan Robinson, Oxford: Basil Blackwell, 1973.
24 Samuelson, Paul A., Economics, 9th edn of Economics: An Introductory Analysis,
New York: McGraw-Hill, [1948] 1973.
25 Shackle, G.L.S., The Years of High Theory, Cambridge: Cambridge University
Press, 1967.
26 ——Epistemics and Economics, Cambridge: Cambridge University Press, 1972.
27 ——Keynesian Kaleidics, Edinburgh: Edinburgh University Press, 1974.
28 Townshend, Hugh, ‘Liquidity-Premium and the Theory of Value’, Economic
Journal, March 1937, 47(1), pp. 157–69.
55
3 On different interpretations of the
General Theory
Don Patinkin
Journal of Monetary Economics (1990) 26, October, pp. 205–43
During the first quarter-century after the publication of the General Theory,
there were no significant differences among the various interpretations of
this book. Such differences began to appear only in the 1960s. These
interpretations are critically examined and an explanation given of their
emergence.
I
To paraphrase Ecclesiastes, of making many interpretations of the General
Theory there is no end, and that is what intrigues me. Why should there be
different interpretations of this book? More to the point, what does it mean
to provide a different interpretation of the General Theory thirty, forty, and
even fifty years after it was published? What new information became
available at those respective times to provide a basis for such new
interpretations? And let me immediately say that one cannot answer that
question by pointing to the hitherto unpublished or obscurely published
materials in the Royal Economic Society’s monumental thirty-volume edition
of Keynes’s Collected Writings. For the volumes of that edition which contain
materials that relate to the General Theory (XIII, XIV, and XXIX) were not
published until the 1970s, several years after different interpretations of the
book had already been advanced. Furthermore, there is little reliance on
these materials even in interpretations which appeared after the publication
of these volumes. I might also add that the new classical macroeconomics
has presented what it regards as fundamental criticisms of the General
Theory, not different interpretations of it.
I am, of course, fully aware of the fact that much more than fifty years
later we continue to get different interpretations of, for example, Smith,
Ricardo, and especially Marx. But in these cases the difference in time is
itself a partial explanation: for no matter how many and how detailed the
studies we have of these writers and their respective periods, we still cannot
have a feeling for the full social, political, and economic context in which
they wrote. We still will not be aware of some of the events and/or
discussions to which they alluded. We still will not fully know what mind-set
56 Don Patinkin
on the part of their readers they took for granted and what details they
accordingly did not bother specifying. And we also might be misled by
words which today have a different meaning, a different connotation, from
what they had at the time they were written.
The situation with respect to the General Theory is quite different. Here
we know much more about the man and his times. Indeed, those of my
generation and older may even have personal recollections of those times,
those dark depression years during which the General Theory was written—
even if (as in my case) they are recollections of adolescent days. And though
there may be some exceptions (I shall return to this point later), there has
been little if any change in the meaning of words since the time of the
General Theory, and its historical allusions are ones that we are aware of
and can (though sometimes only after close contextual study) identify.
So why are there such vastly different interpretations of the General
Theory? Or to ask the question from another viewpoint: why are there not
different interpretations of other classic works of the period? Why are there
not different interpretations of John Hicks’s Value and Capital, published
three years after the General Theory,1 or of Paul Samuelson’s Foundations of
Economic Analysis, published in 1947?
There are in part some obvious answers to this question, and I will
ultimately come to them. At the moment, I would like to say that though the
question of different interpretations of the General Theory has long intrigued
me, it is with hesitancy that I undertake to discuss it—and this for two
reasons. First, the question that I have raised is one in the field of literary
criticism and hermeneutics, and hence a question which rapidly involves us
in deep philosophical issues. And these are fields in which, to say the least, I
have no expertise. Second, I do not undertake this discussion as a
dispassionate observer, but as one who has over the past years himself
presented an interpretation of the General Theory, and in this context even
criticized other interpretations. So it is for you the reader to make whatever
allowances you feel are called for by these facts.
I have, of course, tried to obtain a minimal outsider’s understanding of the
hermeneutical issues inherent in my inquiry. But as I have attempted to
understand the successive waxing and waning of different theories of literary
criticism and hermeneutics since World War II—of the New Criticism, of
Reader-Response Criticism, of Deconstruction and so forth—I have again
thought of Ecclesiastes and his conclusion that ‘much study is a weariness of
the flesh’. At the same time it has been most comforting for me to learn from
my limited venture into this field that the half-life of a theory in literary
criticism is even shorter than one in postwar macroeconomics, and that the
debates between the protagonists of the various theories of interpretation are
equally intense, protracted—and inconclusive.
There is one such debated issue which is of paramount importance for my
purpose: namely, the question of the significance that we should attach to the
intention of the author when we interpret a text. The views of the various
On different interpretations of the General Theory 57
schools of interpretation range here from the deconstructionists, who on
philosophical grounds maintain that the meaning of a text is indeterminate
and may always be construed by different readers in different ways—to
which they add that the original intention of an author (even when he
explicitly declares it) is again subject to different interpretations; to the
approach of Stanley Fish (1980), who maintains that though in theory there
are an unlimited number of possible interpretations of a text, in practice
there are only a finite number of different ‘interpretive communities’, each
with its own rules and conventions, and each accordingly with a ‘legitimate’
interpretation; to what I understand is today considered the conservative
view of E.D.Hirsch (1967: ch. 5), who believes that though we cannot
achieve absolute certainty, we can—by applying the inductive and
probabilistic methods of all sciences—present a construal of the author’s
intention and explain why it is preferable to all others,2 and who accordingly
maintains (1976:90–1) that ‘unless there is a powerful overriding value in
disregarding an author’s intention (i.e., original meaning), we who interpret
as a vocation should not disregard it…. To treat an author’s words merely as
grist for one’s own mill is ethically analogous to using another man merely
for one’s own purposes’ (italics in original). As an example of such a
‘powerful overriding value’ Hirsch cites the possibility that ‘one might fudge
on original meaning for the sake of young, impressionable children’ (ibid.:
90). I think that we here can safely ignore that danger.
I am obviously not qualified to express an opinion with respect to the
validity of these and other theories of interpretation. I shall, however, take
advantage of the degrees of freedom afforded by this lack of consensus
among specialists in the field to choose among them and invoke the
authority of Hirsch for the emphasis that I have always given in my
interpretation of the General Theory on Keynes’ intentions in writing it—its
original meaning. I also invoke Hirsch’s authority (1967:209ff.; 1976:124ff.)
to justify my use of any evidence that throws light on these intentions:
evidence not only from the text itself (pace the New Criticism, which
contends that this should constitute the sole basis for interpretation), but also
from the historical context in which the book was written, from the context
of Keynes’ other writings and activities, from the reactions of his
contemporaries to the book, and so forth. I am also encouraged by the fact
that this approach to the history of economic ideas accords with the one that
Quentin Skinner (1969) has followed in his influential studies in the history of
political ideas. I must admit that both Hirsch (1967: passim) and Skinner
(1976:68) emphasize that one cannot say that the original meaning of a text is
the only possible reading. At the same time, I do think that this meaning can
be used to justify the rejection of interpretations that differ greatly from it.
I would also like to say that whatever may be the proper hermeneutical
principle to follow with respect to literary works, Hirsch’s view is to my
mind the correct one with respect to scientific writings—and for present
purposes let me include economics as a science. As Robert Merton (1957) has
58 Don Patinkin
emphasized, priorities play a major role in the reward system of science, so
for this purpose alone the scientist writes with the intention of conveying to
his profession a precise and definite message. And this is a fortiori so in the
case of an economist whose message implicitly or explicitly includes the
advocacy of certain policy measures. So in interpreting scientific writings,
and especially writings in the social sciences, we should make use of all
available evidence as to the author’s intention.
II
Keynes’ intention in writing the General Theory—its original meaning—is
already indicated in its title, which lists ‘Employment’ as the first of the
subjects to be dealt with. The General Theory (GT) is divided into six
Books. In Book I—entitled ‘Introduction’—Keynes presents ‘a brief summary
of the theory of employment to be worked out in the course of the following
chapters’ (GT: 27). This is his ‘theory of effective demand’, which in this
introductory Book is presented under the explicit simplifying assumptions of
a constant level of investment (which presupposes a constant rate of interest)
and a constant money wage rate, an assumption with which Keynes tells us
he will ‘dispense later’ (GT: 27–9). The central message of this theory and its
analytical novelty (as I have shown on earlier occasions: Patinkin 1976: chs
8–9; 1982: ch. 1) is that changes in output themselves act as an equilibrating
force to bring aggregate demand and supply—or, equivalently, planned
investment and saving—into equality at a level that need not be one of full
employment. In Keynes’ words: ‘The novelty in my treatment of saving and
investment consists, not in my maintaining their necessary aggregate
equality, but in the proposition that it is, not the rate of interest, but the level
of incomes which (in conjunction with certain other factors) ensures this
equality’ (Keynes 1937a: 211; see also GT: 31, lines 16–23; 179: lines 2–6).
And this was his explanation of the ‘paradox of poverty in the midst of
plenty’ (GT: 30): his explanation of the seemingly endless depression in the
Western world that was creating misery for millions of unemployed and even
endangering the existence of its democratic institutions.
In this Book, Keynes also tells us that once his theory of effective demand
is set out, ‘we shall find that the theory of prices falls into its proper place as
a matter which is subsidiary to our general theory’ (GT: 31–2). I shall later
indicate the specific sense in which it is subsidiary. Now, however, I would
like to suggest that with this statement Keynes also intended to highlight the
difference between the central message of the General Theory and that of his
Treatise on Money, whose subject was indeed the price level as analyzed by
what he termed ‘The Fundamental Equations for the Value of Money’ (so the
title of Chapter 10 of that book).
After a ‘digression’ from the ‘main theme’ in Book II for the purpose of
clarifying various concepts, Keynes devotes ‘Book III: The Propensity to
Consume’ and ‘Book IV: The Inducement to Invest’ to those two components
On different interpretations of the General Theory 59
of effective demand. In the latter Book, Keynes drops the assumption of a
constant level of investment and explains how this level is determined by the
marginal-efficiency-of-capital schedule in conjunction with the rate of
interest, which rate is in turn determined by the liquidity-preference schedule
in conjunction with the quantity of money.
Finally, in ‘Book V: Money-Wages and Prices’, Keynes drops (as he had in
Book I said he would) the assumption of a constant money wage and devotes
the first chapter of this Book (Chapter 19, entitled ‘Changes in Money-
Wages’) to an analysis of the effects of such changes—explaining at the
beginning of this chapter that ‘it was not possible…to discuss this matter
fully until our own theory had been developed’ (GT: 257). And he goes on in
this chapter (which I have accordingly always regarded as the apex of the
General Theory) to tie together all the analytical elements of the preceding
chapters in order to argue that a decline in money wages in the face of
unemployment might create such perverse expectations and such a wave of
bankruptcies that the level of aggregate demand, hence effective demand,
and hence employment, would remain unchanged. From this followed the
main policy conclusion of the General Theory, namely, that ‘there is,
therefore, no ground for the belief that a flexible wage policy is capable of
maintaining a state of continuous full employment;—any more than for the
belief that an open-market monetary policy is capable, unaided, of achieving
this result. The economic system cannot be made self-adjusting along these
lines’ (GT: 267). Consequently government must take ‘an ever greater
responsibility for directly organising investment’ (GT: 164; see also p. 378)
in order to assure that total expenditures on investment in the economy—
augmented by the multiplier effect—will supplement expenditures on
consumption to the extent necessary to bring aggregate demand to its full
employment level.
Book V also contains ‘Chapter 20: The Employment Function’ and
‘Chapter 21: The Theory of Prices’, and I shall have something to say about
them later. I am also deferring until later some observations about Book VI,
entitled ‘Short Notes Suggested by the General Theory’, which is the final
one of the General Theory.
We can obtain a deeper understanding of Keynes’ intentions in writing
the General Theory by reading it in the context of his earlier writings on
the problem of unemployment. Thus in his 1925 Economic Consequences of
Mr. Churchill, he attributed the increase in the level of unemployment that
then took place to England’s return to the gold standard at prewar parity,
thus overvaluing the pound. The analysis of his 1930 Treatise was of a
general nature and attributed unemployment to entrepreneurial losses
associated with too high a real wage. In both cases Keynes contended that
if only nominal wages could be simultaneously and equiproportionately
reduced, the problem would be solved: in the Economic Consequences of
Mr. Churchill this reduction was to be accompanied by one in domestic
prices as well, so that the real wage was to be reduced only in terms of
60 Don Patinkin
international prices, thus offsetting the overvalued pound (Keynes,
Collected Writings, hereafter JMK IX: 211, 228–9); in the Treatise it was to
be a reduction of real wages in terms of domestic prices as well, thus
eliminating business losses (JMK: 141, 151, 244–5, 265, 281). In both
cases, however, the illocutionary force of Keynes’ discussion was not to
actually advocate a policy of dealing with unemployment by reducing
nominal wages, but to highlight the fact that in practice the resistance of
workers would make this impossible, so that the alternative policies that he
was advocating were called for. (In the passages from the Treatise just
cited, Keynes repeatedly expresses the view that only in a totalitarian
state—‘in Bolshevist Russia or in Fascist Italy’ (ibid.: 244)—could such a
wage reduction be carried out.)
The unemployment of the 1930s, however, constituted (according to
Thomas Kuhn 1970: chs 6–8) an ‘anomaly’ for Keynes’ earlier analysis of
the problem, and this for two reasons. First, unemployment had become a
worldwide phenomenon, and so could not be explained in terms of the
specific circumstances of Britain. Second, from 1929 to 1933—that is, from
the time of writing the Treatise to that of writing the General Theory—
money wages had fallen in the United States by over a quarter, but to no
avail insofar as unemployment was concerned. True, the price level had
fallen even more, thus resulting in an increase in the real wage; but this too
was part of the anomaly.
So in addition to the basic theoretical criticisms to which the Treatise was
subjected (see Patinkin 1976:54–8), this empirical experience necessitated
two important revisions in Keynes’ earlier views: it showed that in practice
money wages could fall drastically (even if not simultaneously), but it also
showed that this would not help to solve the problem of unemployment. It
was to this experience that Keynes alluded when at the beginning of the
General Theory he wrote that ‘it is not very plausible to assert that
unemployment in the United States in 1932 was due either to labour
obstinately refusing to accept a reduction of money-wages or to its
obstinately demanding a real wage beyond what the productivity of the
economic machine was capable of furnishing’ (ibid.: 9).
Keynes’ new theory of effective demand, together with his acceptance of
the ‘first classical postulate’ that ‘the [real] wage is equal to the marginal
product of labour’ (GT: 5, italics deleted), provided the answer to both these
anomalies. For if as a result of the adverse reactions described in Chapter 19,
a reduction in money wages in the face of unemployment would actually be
followed by a decrease in effective demand and hence in the level of output
and corresponding input of labor, this would result in an increase in the
marginal product of labor and hence in its real wage. And this is the
meaning of Keynes’ cryptic statement in ‘Book I: Introduction’ of the General
Theory that ‘there may exist no expedient by which labour as a whole can
reduce its real wage to a given figure by making revised money bargains
with the entrepreneurs’ (GT: 13, italics in original). This was a specific
On different interpretations of the General Theory 61
instance of Keynes’ general principle that ‘the propensity to consume and the
rate of new investment determine between them the volume of employment,
and the volume of employment is uniquely related to a given level of real
wages—not the other way round’ (GT: 30).
III
I began this lecture by saying that I was intrigued by the existence today of
widely different interpretations of the General Theory. In this connection
there are two most significant facts: first, there were no such wide differences
in the interpretations that were presented in the years immediately following
its publication; and second, almost a quarter of a century was to elapse
before such differences did appear.
The contemporary interpretations that I have in mind began with the 1935
‘missionary’ talk in which Robert Bryce (who had attended Keynes’ lectures
in the successive years 1932–4) brought the gospel according to Keynes to
what he many years later described as ‘the nearest concentration of heathen
available from Cambridge’—namely, Hayek’s seminar at the London School
of Economics (Bryce 1977:40, 129–45). It continued with Joan Robinson’s
1937 book Introduction to the Theory of Employment’, with the respective
review articles that appeared in the years 1936 to 1938 by Champernowne
(1936), Hansen (1936), Harrod (1937), Hicks (1936, 1937), Lange (1938),
Lerner (1936), Meade (1937) and Reddaway (1936); and to the celebrated
1936 Quarterly Journal of Economics symposium with the participation of
Leontief, Robertson, Taussig and Viner. This symposium concluded with a
reply by Keynes (1937b) which has played a major role in other
interpretations of the General Theory and which will be discussed in the next
section.
In a letter that he wrote in August 1936 to Harrod commenting on a draft
of the latter’s review, Keynes identified the three major components of his
book as being the theory of effective demand (with its ‘psychological law’ of
a less-than-unity marginal propensity to consume), the marginal efficiency of
capital, and the theory of liquidity preference (JMK XIV: 84–6). And almost
a decade later, about a year before his death, Keynes (according to an entry
from March 1945 in James Meade’s diary) gave a ‘lecture’ to a government
committee in which he once again designated these components as the main
ones of the General Theory.3 These are also the components—in varying
ways and degrees of emphasis, and in some cases with explicit attention also
being paid to Keynes’ corollary discussions of the multiplier, of the inefficacy
of a reduction in the money wage rate as a means of reducing
unemployment, and of the determination of price given this rate—that we
find in the foregoing contemporary interpretations.
What is even more important for my present purposes is what we do not
find in them. Thus (with the exception of Robertson 1936: §1) we do not find
discussions of the aggregate supply function: or of wage-induced cost
62 Don Patinkin
inflation; or of the distribution of income; or of the ‘animal spirits’, the
irrational aspects of economic behavior which influence economic decisions
and thus allegedly make it impossible to speak—even in the short-run
context which is the major concern of the General Theory—of a stable
investment function and a corresponding equilibrium; or of the material
contained in the chapter on social philosophy in Book VI. Nor do we find
any discussions which interpret the General Theory as denying the
possibility that proper government fiscal policy can assure full employment
in a capitalist economy.
And what is equally significant is that though Keynes wrote letters to
most of his reviewers, commenting on their respective interpretations, he
generally approved of them, and in any event did not criticize them for
omitting any of the foregoing points.4 On the contrary, the only omission
for which he criticized Harrod’s review in the aforementioned August 1936
letter was precisely that it did not ‘mention effective demand or, more
precisely, the demand schedule for output as a whole, except in so far as it
is implicit in the multiplier’ (JMK XIV: 85, italics in original). And in
essence this was also Keynes’ major criticism of Lerner’s review (which he
otherwise termed ‘splendid’) in his letter to him two months before (JMK
XXIX: 214–16).
Let me in particular emphasize that in his correspondence with Harrod,
Reddaway, Meade, and Hicks on their respective review articles, Keynes
did not express any objection to the fact that each in his own way had
presented the analysis of the General Theory in terms of a general-
equilibrium system of simultaneous equations. When in August 1935
Harrod had tried to convince Keynes that he should view his analysis in
that way, Keynes vehemently rejected his suggestion (JMK XIII: 526–65,
especially pp. 531–2, 545–6, 548, 553–4 and 557). But in the course of the
year, Keynes had apparently come around to accepting it. Thus in his
aforementioned 1936 letter to Harrod, Keynes wrote: ‘I like your paper
(may I keep the copy you have sent me?) more than I can say. I have found
it instructive and illuminating, and I really have no criticisms. I think you
have reorientated the argument beautifully’ (JMK XIV, p. 84). Similarly, in
his March 1937 letter to Hicks on the latter’s IS-LM interpretation, Keynes
wrote that he ‘found it very interesting and really have next to nothing to
say by way of criticism’. And his main criticism was that Hicks’
investment function depended only on current income, whereas Keynes felt
that ‘whilst it may be true that entrepreneurs are over-influenced by present
income’, nevertheless ‘expected income for the period of investment is the
relevant variable’ (JMK XIV: 79–81). Again, he ended an August 1936
letter to Reddaway with ‘I enjoyed your review of my book in the
Economic Record, and thought it very well done’ (JMK XIV: 70). And in a
September 1936 postcard to Meade, Keynes wrote: ‘Thanks for the copy of
your paper. It’s excellent. I have no criticisms to suggest’ (cited by Young
1987:34).
On different interpretations of the General Theory 63
Furthermore, on at least one occasion, Keynes expressed his approval of a
general-equilibrium interpretation of his book, not only in correspondence,
but also in print. Specifically, in the course of an exchange with Robertson in
the pages of the 1938 Economic Journal, Keynes described Lange’s 1938
review article—which like those of Reddaway, Hicks and Harrod (as Lange
himself pointed out in its opening footnote) presented such an
interpretation—as one ‘which follows very closely and accurately my line of
thought’ (JMK XIV: 232, n. 1).5
And now let me anticipate an issue that I will later discuss and emphasize
that all of the above simultaneous-equation interpretations of the General
Theory can essentially be regarded as variations of IS-LM: the distinctive
feature of Hicks’ version was that it also provided a diagrammatic
presentation. Thus Keynes’ approval of all these reviews also constituted his
consistent approval of the IS-LM interpretation of the General Theory’.6
The IS-LM interpretation was elaborated upon by Modigliani in his
influential 1944 article; it appeared in Klein’s classic 1947 work on The
Keynesian Revolution (pp. 87–8); and it also played an important role in
Hansen’s Guide to Keynes (1953:107, 143–8). And as we all know, the IS-
LM interpretation became the standard representation of the Keynesian
system in the macroeconomic textbooks that began to appear in the 1960s,
and thereby became the hallmark of ‘mainstream Keynesianism’.
I might also add that the interpretation presented in Chapters XIII: 4 and
XIV: 1 and 3 of the 1956 and subsequent (1965 and 1989) editions of my
Money, Interest, and Prices is essentially that of IS-LM—with the difference
that I regarded its equilibrium position as being of a Marshallian short-run
nature which, if disturbed by a decline in the money wage rate in the face of
unemployment, might (in accordance with Keynes’ argument in Chapter 19)
bring the economy to a new short-run position of unemployment equilibrium,
but would not restore full employment. In Keynes’ words, in his chapter on
‘The General Theory of Employment Re-stated’,
it is an outstanding characteristic of the economic system in which we live
that, whilst it is subject to severe fluctuations in respect of output and
employment, it is not violently unstable. Indeed it seems capable of
remaining in a chronic condition of sub-normal activity for a considerable
period without any marked tendency either towards recovery or towards
complete collapse. Moreover, the evidence indicates that full, or even
approximately full, employment is of rare and short-lived occurrence.
(GT: 249–50)
This disequilibrium approach enabled me to dispense with the Hicks-
Modigliani assumptions of a rigid money-wage rate (which, as emphasized
in section II above, Keynes had dropped in Chapter 19 of the General
Theory) and/or ‘liquidity trap’ [about which Keynes had said that ‘whilst this
limiting case might become practically important in future, I know of no
64 Don Patinkin
example of it hitherto’ (GT: 207)],7 assumptions which are required if one
interprets the General Theory as describing a position of long-run
unemployment equilibrium, i.e. one which remains unchanged. Similarly,
when in my Keynes’ Monetary Thought I said that ‘in most cases, I do not
think that my views on these issues differ basically from the traditional ones’
(Patinkin 1976:10; see also p. 100), it was the IS-LM interpretation, with this
difference, that I had in mind.
IV
When did significantly different interpretations of the General Theory begin
to appear? There is no mention of them either in Schlesinger’s (1956) survey
article ‘After Twenty Years: The General Theory’ or in Harry Johnson’s
(1961) corresponding article on The General Theory after Twenty-Five
Years’. And to the best of my knowledge, it is just about that time that such
interpretations do begin to appear. I am referring in particular to Sidney
Weintraub’s (1961) Classical Keynesianism, Monetary Theory, and the Price
Level and to George Shackle’s survey article of the same year on ‘Recent
Theories Concerning the Nature and Role of Interest’. In the latter, Shackle
(1961:228) referred to Keynes’ 1937 article in the aforementioned Quarterly
Journal of Economics symposium and said that ‘no reader of Keynes’s
article…will be in doubt that Keynes looking back saw as the main theme of
his book the commanding importance of uncertainty and of the conventions
by which the insoluble problems it poses, and the nonsense it makes of pure
“rational calculation”, can be shelved in order to make life possible at all’.
And a few years later, in his book Years of High Theory, Shackle (1967)
devoted a whole chapter to Chapter 12 of the General Theory (which bears
the title ‘The State of Long-Term Expectation’) and to Keynes’ 1937 article;
subtitled his (Shackle’s) chapter ‘Keynes’s Ultimate Meaning’; claimed that
this theme is ‘the message of the General Theory, and…the only part of it
which Keynes troubled to reproduce’ in his 1937 article (Shackle 1967:130);
and termed this article the ‘third edition’ of the General Theory, after having
designated the Treatise on Money as the first one (Shackle 1967:136; see also
Shackle 1973). And in a still later article, Shackle (1982:438) explicitly
rejected Hicks’ IS-LM interpretation on the grounds that ‘the elemental core
of Keynes’ conception of economic society is uncertain expectation, and
uncertain expectation is wholly incompatible and in conflict with the notion
of equilibrium’.8
Weintraub’s interpretation of the General Theory emphasized instead the
analysis of the determination of the price level, particularly in the case of
wage-induced cost inflation. He stated that he will ‘make no effort to refer to
Keynes even though I think the tenor and the text will sustain me’ (1961:3).9
Weintraub also entered a plea for the abandonment of the IS-LM
interpretation, as well as that of the 45°-cross diagram, on the alleged
ground that these deal only with real quantities (ibid.: 5–10, 18–22). And he
On different interpretations of the General Theory 65
advocated instead the use of the aggregate demand and supply curves of
Chapter 3 of the General Theory, which reflect changes in the price level by
virtue of their being expressed in nominal money terms.
Let me begin with Shackle and say that there is no question that
expectations and uncertainty play an essential role in the General Theory.
But Shackle takes Keynes’ discussion of uncertainty in the Quarterly Journal
article somewhat out of the context in which it appears. Specifically, in the
first part of this article, Keynes briefly discusses the criticisms of Leontief,
Robertson, Taussig and Viner—and then states that Viner’s is ‘the most
important of the four comments’ (JMK XIV: 110). Now, Viner’s major
criticism was directed at Keynes’ theory of liquidity preference. And this was
the reason that Keynes went on to devote most of the second part of his reply
to explicating the nature of the uncertainty that generates this preference.
There is, however, little if anything in this exposition which differs from that
of Chapter 12 of the General Theory. In particular, Keynes’ well-known
statement in the Quarterly Journal article that the uncertainty which
characterizes so much of economic life is one for which ‘there is no scientific
basis on which to form any calculable probability whatever’ (JMK XIV: 114)
is equivalent to the at least equally well-known statement in Chapter 12 that
‘our decisions to do something positive, the full consequences of which will
be drawn out over many days to come, can only be taken as a result of
animal spirits—of a spontaneous urge to action rather than inaction, and not
as the outcome of a weighted average of quantitative benefits multiplied by
quantitative probabilities’ (GT: 161). Furthermore, contrary to Shackle’s
aforementioned statement in Years of High Theory (1967:130), the discussion
of uncertainty is not ‘the only part’ of the General Theory that Keynes
reproduces in his 1937 article: for Keynes goes on in it to discuss both the
theory of effective demand and the marginal efficiency of capital (JMK XIV:
119–23). In brief, Keynes’ 1937 Quarterly Journal article emphasizes the
same three basic components of the General Theory that he had emphasized
in his 1936 letter to Harrod (see p. 61).10
I must also point out that Keynes concludes his discussion of ‘animal
spirits’ in the General Theory with the statement that
We should not conclude from this that everything depends on waves of
irrational psychology. On the contrary, the state of long-term expectation
is often steady, and, even when it is not, the other factors exert their
compensating effects.
(GT: 162)
Similarly, Keynes writes:
There are not two separate factors affecting the rate of investment,
namely, the schedule of the marginal efficiency of capital and the state of
confidence. The state of confidence is relevant because it is one of the
66 Don Patinkin
major factors determining the former, which is the same thing as the
investment demand-schedule.
(GT: 149)
Thus even after taking account of the major influence of ‘the state of
confidence’ on expectations, Keynes still speaks of a determinate investment
demand schedule in the short-run context which is the concern of the central
message of the General Theory.11 Needless to say, in a longer-run context
Keynes emphasized that significant fluctuations ‘in the market estimation of
the marginal efficiency of different types of capital’ would be ‘likely’ (GT:
164). Indeed, Keynes’ explanation of the business cycle in Chapter 22 of the
General Theory that he devoted to this subject was in terms of ‘a cyclical
change in the marginal efficiency of capital’ (GT: 313).12
Let me finally note that the presentation of the nonprobabilistic nature of
economic uncertainty can hardly be considered to be a contribution of the
General Theory: it had been emphasized long before by Frank Knight (1921)
in his classic Risk, Uncertainty, and Profit. Furthermore, as Samuelson
(1946:320) observed many years ago, the General Theory ‘paves the way for
a theory of expectations, but it hardly provides one’ (see also Hart 1947).
I turn now to Sidney Weintraub’s interpretation and first of all note that
in contrast with his view that the aggregate supply curve is a basic
component of the General Theory, this curve was not among the three such
components of his book that Keynes set out in his August 1936 letter to
Harrod, described in the preceding section. Similarly, as indicated in that
section, this curve was not even referred to by any of the contemporary
reviewers of the book, with the exception of Robertson; nor did Keynes
complain about this in his correspondence with them. And to this I add
that, as I have elsewhere shown (Patinkin 1982:142–53), Keynes himself
did not have a clear notion of the nature of this curve, and particularly
about its mathematical properties (viz., its slope and convexity). Thus in
Chapter 3 of the General Theory on ‘The Principle of Effective Demand’,
the properties of this curve—in contrast with those of the aggregate demand
curve—are not specified. Furthermore, in the only place in the book where
they are specified (GT: 55, n. 2), Keynes does so incorrectly. And though
Keynes devotes Chapter 20 to a discussion of what he calls the employment
function, and which he defines as the ‘inverse function’ of the aggregate
supply function (GT: 280), the various elasticity formulas which he
presents do not explicitly describe the properties of the latter, beyond
implying the obvious one that its slope is positive.
Insofar as Weintraub’s emphasis on the price level is concerned, I have
already noted Keynes’ statement in Book I that ‘the theory of prices falls
into its proper place as a matter which is subsidiary to our general theory’
(GT: 32). For if the level of employment and hence the marginal product of
labor is determined by the level of effective demand, and if (as Keynes
assumed) price is equal to marginal cost,13 then for any given money-wage
On different interpretations of the General Theory 67
rate the level of effective demand also determines the price. In the words of
Chapter 21 of the General Theory on ‘The Theory of Prices’: ‘The general
price-level (taking equipment and technique as given) depends partly on the
wage-unit [i.e. on the money-wage rate] and partly on the volume of
employment’ (GT: 295). And I might also point out that, contrary to
Weintraub’s criticism, Hicks’ IS-LM article deals not only with real
quantities, but also with the respective price levels of consumption and
i n v e s t m e n t g o o d s a s d e t e r m i n e d i n t h i s w a y ( H i c k s 1 9 3 7 : 1 0 3 ) .
Furthermore, the standard presentation of IS-LM includes a description of
how changes in the price level cause shifts in the LM curve and hence
affects the equilibrium position.
Note the parallel treatment in the General Theory of the determination of
the real wage and the determination of the price level: in both cases this is
achieved as a by-product of the theory that constitutes its central message–
namely, the theory of effective demand that explains the level of output in
the economy, hence its level of employment, and hence the marginal product
of labor.
V
The quarter-century and more since the writings of Shackle and Weintraub in
the early 1960s has seen the presentation of innumerable interpretations of
the General Theory, far more than I could possibly discuss. My impression,
however, is that most of them are variations on either the IS-LM
interpretation of ‘mainstream Keynesianism’, or the interpretation of so-
called ‘Post-Keynesianism’, to which I shall in a moment turn. I shall not,
however, discuss such well-known works as those of Leijonhufvud (1968) and
Chick (1983), for their avowed main purpose is to study the General Theory
as the point of departure for their respective contributions to (in the words of
the title of Chick’s book) ‘macroeconomics after Keynes’. Similarly, the
subtitle of Leijonhufvud’s book is ‘A Study in Monetary Theory’ (see also
ibid., p. 9). Accordingly, I consider these works to be outside my terms of
reference, strictly construed.14
Let me then briefly discuss Post-Keynesian economics. This is actually
only in part concerned with the interpretation of the General Theory. In this
context it is largely a combination of George Shackle’s interpretation of the
book in terms of the overriding impact of uncertainty on economic behavior,
and Sidney Weintraub’s interpretation in terms of the major importance in
the book of the aggregate supply curve and of the analysis of the
determination of the price level.15 And I have nothing to add to my criticisms
of these interpretations in the preceding section. In addition, Post-Keynesian
economics—particularly as expounded by the so-called ‘Modern Cambridge
School’ of Joan Robinson (in her postwar period), Piero Sraffa, and their
followers—has a major concern with the development and elaboration of the
theories of Marx, Kalecki, and Sraffa (see Harcourt 1987) and is in this
68 Don Patinkin
respect also outside my terms of reference.16 The same is true of the work of
Nicholas Kaldor in taking the General Theory as his point of departure for
developing theories of income distribution and growth, respectively—issues
which were not among the concerns of the General Theory (see Patinkin
1976:19–20; see also p. 80 of this chapter).
A Post-Keynesian cum Modern-Cambridge-School work that is however in
large part devoted to an interpretation of the General Theory is Murray
Milgate’s 1982 book, Capital and Employment: A Study of Keynes’s
Economics. The Modern Cambridge School rejects both marginal analysis
and the notion of capital as a factor of production, and then implicitly makes
use of ‘productivity ethics’17 to deny the moral justification of profits in a
capitalist economy. And a major purpose of Milgate’s book is to attribute
this rejection to the General Theory as well. Indeed, Milgate presents the
transition from the Treatise to the General Theory as a transition from the
marginal viewpoint to the nonmarginal one.
At first sight, this would seem to be ‘mission impossible’. For whereas the
theory of value qua marginal analysis—and even the term ‘marginal’—is
completely absent from the Treatise (see Patinkin 1976:13, 47, 94), it is
repeatedly and consistently applied in the General Theory, the marginal
productivity of labor as determining its real wage, marginal cost as
determining price, and the marginal efficiency of capital as determining
investment decisions. And surely to speak of ‘the marginal efficiency of
capital’ is to regard capital as a factor of production.
Milgate (1982:166), however, explains that there is no reference to the
theory of value in the Treatise because it is too ‘obvious’ to be mentioned.
On the other hand, the notion of the marginal efficiency of capital—to which
Keynes devotes a whole chapter in the General Theory, which he describes
there as equivalent to Irving Fisher’s (1930:155, 159, 168) marginal rate of
return over cost, and which in his aforementioned August 1936 letter to Roy
Harrod (see p. 61 above) he listed as one of the three major analytical
components of his book—that notion [tells us Milgate (1982:9Iff.)] is among
the ‘inconsistent’ ‘remnants of the orthodox position’ that are to be found in
the General Theory. Finally, Milgate interprets the General Theory as
presenting a theory of long-run unemployment that necessarily exists in a
capitalist economy, and (using a singular definition of the term) then
interprets this view as constituting a rejection of ‘orthodox marginalist
theory’ (1982:96–7).18
In this connection I should also refer to Luigi Pasinetti—a leading member
of the Modern Cambridge School—who contends that ‘the marginal-
efficiency-of-capital schedule, which might, at a first superficial look, appear
as belonging to marginal economic analysis, when examined more deeply
turns out to have a rather different origin’—namely, classical nonmarginal
analysis (Pasinetti 1974:43). Pasinetti also rejects Hicks’ IS-LM general-
equilibrium interpretation of the General Theory on the grounds that the
correct interpretation of this book is in terms of a system of equations which
On different interpretations of the General Theory 69
is ‘decomposed’ into a causal chain by which the rate of interest is
determined by the liquidity-preference equation alone (i.e. independently of
the level of income), that the rate of interest so determined is then inserted
into the savings=investment equation to determine the level of income, and
that it was Hicks who has ‘broken up Keynes’ basic chain of arguments’ by
‘introducing income’ into the liquidity-preference equation, thus yielding a
system of equations which cannot be so ‘decomposed’ (Pasinetti 1974:74).
This contention stands in direct contradiction to the fact that the liquidity-
preference function which Keynes presented and analyzed in Chapter 15 of
the General Theory depends on both income and the rate of interest, and has
the specific form M=L1(Y)+L2(r) (see GT: 199).
The rejection of Hicks’ IS-LM interpretation (though generally not on the
grounds advanced by Pasinetti, but on those of Shackle—namely, on the
grounds that it attributes to the General Theory the analysis of the
determination of an equilibrium situation and of the equilibrium level of
income in particular: see previous section) is another hallmark of Post-
Keynesianism. Indeed, its rejection of this interpretation is so vehement as to
bring it to follow Joan Robinson (1962a: 27–9; 1962b: 100–2; see also
1962c: 76; 1979: xi) in denouncing the kind of analysis which it represents as
‘bastard Keynesianism’.19 Some adherents of the Modern Cambridge School
(see e.g. Harcourt 1980:151) have even claimed that Keynes himself
‘explicitly’ rejected Hicks’ interpretation in his 1937 letter to him. The
discussion of this letter in section III above, as well as of Keynes’ approval of
what are essentially other ‘IS-LM interpretations’ of his book, shows how
unfounded this claim is. And though that was probably not his intention in
making it, Richard Kahn’s (1984:160) complaint that ‘it is tragic that Keynes
made no public protest when they [i.e. IS-LM interpretations of the General
Theory] began to appear’, is to my mind further evidence that Keynes
basically accepted them.
Let me also note that the one diagram that we do find in the General
Theory (p. 180) is logically equivalent to the IS curve. For though drawn
with different axes, this diagram shows different combinations of the rate of
interest and the level of income in which the commodity market is in
equilibrium. Furthermore, Keynes goes on to say that this diagram alone
cannot determine the equilibrium levels of these variables; ‘if, however, we
introduce the state of liquidity-preference and the quantity of money and
these between them tell us that the rate of interest is r2, then the whole
position becomes determinate’ (GT: 181). Here, then, is the spirit of IS-LM—
the determination of the equilibrium level of income by the interaction
between the markets for commodities and money—even if not its precise
geometrical form.
In addition to making use of the aforementioned arguments of Shackle
and Weintraub, Post-Keynesians have also attempted to support their
rejection of IS-LM with the claim that Hicks himself has since then—to
quote Joan Robinson (1978: xiv)—‘repented’ (note the choice of language
70 Don Patinkin
associated more with religious disputes than with scholarly disagreements:
Hicks has not only been wrong, but like Jeroboam son of Nebat in the
Book of Kings, he has sinned and made others to sin).20 Robinson’s claim
was based on Hicks’ statement in a 1976 article that the IS-LM diagram ‘is
now much less popular with me than I think it still is with many other
people’ (pp. 289–90). On this I have two observations: First, neither this
article, nor Hicks’ later article on ‘IS-LM—An Explanation’ (1981)—which
Post-Keynesians also cite in this connection—say that IS-LM is not a proper
interpretation of the General Theory: on the contrary, it seems to me that
in both articles Hicks is actually very careful not to say this. Indeed, in the
sentence immediately preceding the one from the 1976 article that
Robinson quoted, Hicks referred to Keynes’ March 1937 letter to him (see
p. 62 above) and said: ‘I think I am justified in concluding from that letter
that Keynes did not wholly disapprove of what I had made of him’.
Similarly, in his 1974 Crisis in Keynesian Economics, Hicks referred to his
IS-LM interpretation and said:
To many students I fear it is the Keynes theory. But it was never intended
as more than a representation of what appeared to be a central part of the
Keynes theory. As such I think it is still defensible.*
*It would appear that Keynes himself accepted it as such. See his letter of
March 1937…in JMK XIV, pp. 79–81.
(Hicks 1974:6, original italics and footnote)
(See also part 1 of Hicks’ 1981 article, as well as Hicks 1973:10.)
My second, and far more important point, however, is that whatever
Hicks might have said or thought forty years after his 1937 article does not
change the fact that at that time he presented IS-LM as an interpretation of
the General Theory, and that Keynes accepted it as such. And let me
immediately add that this last statement is not inconsistent with my emphasis
on the importance of the author’s intent in interpreting a text: for there is a
fundamental difference between intent at the time of writing and ‘retroactive
intent’ many years later.21
VI
In the opposite direction from Milgate’s book on the political spectrum is
that of the well-known conservative economist, Allan Meltzer, whose book is
entitled Keynes’s Monetary Theory: A Different Interpretation (1988). Almost
one-third of the book is devoted to a chapter entitled ‘The General Theory. A
Different Perspective’, which to a large extent repeats the contents of a 1981
article of his with an essentially identical title. In this chapter Meltzer
contends that some of the major differences between his interpretation and
others are that his does not depend on either money illusion in the labor
market, or wage rigidity, or the ‘liquidity trap’, and that (on the other hand)
On different interpretations of the General Theory 71
it emphasizes the role of expectations. In a criticism of this article (Patinkin
1983), I showed (inter alia) that in point of fact Meltzer’s interpretation does
not differ in these respects from some earlier interpretations, and even from
some macroeconomic textbooks—and I see no point in repeating that
demonstration here. My present concern is instead with two of Meltzer’s
contentions that are indeed different from interpretations of Keynes hitherto.
The first is his (Meltzer’s) contention that a major objective of the General
Theory was to stress the importance of ‘a higher level of investment until the
capital stock reached a social optimum’ (1988:300; see also pp. 7, 118, 185–
6). The second is his attempt to create the impression that Keynes shared the
conservative view in favor of rules in the present-day debate on ‘rules vs.
discretion’, and that indeed this represents a consistent strand in Keynes’
policy views from the 1920s on (Meltzer 1988:8–9, 182–92, 200, 293–7).
Meltzer’s first contention is based on Keynes’ discussion in Chapter 24 of
the General Theory, entitled ‘Concluding Notes on the Social Philosophy
towards which the General Theory Might Lead’, in which Keynes says that if
as a result of what he terms ‘a somewhat comprehensive socialisation of
investment’ (GT: 378) a continuous state of full employment could be
assured, then investment would proceed at such a rapid rate that the
economy would achieve his ‘aim of depriving capital of its scarcity-value
within one or two generations’ (GT: 377), with salutary consequences for the
distribution of income.22
That here and in other writings Keynes considered it desirable to increase
the stock of capital is a point of Meltzer’s (1988:186) that is well taken. But
to contend that this was a major theme of the General Theory is to tear this
book out of the historical context in which it was written, and to tear
Chapter 24 out of its context in the book. For the General Theory was
written at a time not only of idle men, but of idle factories; so it is surely
farfetched to think that a major concern of Keynes’ at that time was to
increase capital investment in order to further increase the stock of idle plant
and equipment. Instead the crucial role of investment expenditures in the
General Theory was to supplement consumption expenditures in order (with
the help of the multiplier) to raise aggregate demand to its full employment
level. In brief, the passage in Chapter 24 referred to in the preceding
paragraph has to do, not with the major role of investment in the General
Theory, but with its role (incidental to the central message of the book) in
ultimately bringing about Keynes’s version and vision of the classical
stationary state.
Insofar as the context of Chapter 24 within the General Theory is
concerned, let us remember that it appears in Book VI, the last one of the
General Theory, and one that could have been omitted without affecting its
central message. Of this there are several indications, of which the casual
title of this Book—namely, ‘Short Notes Suggested by the General Theory’
—is itself one. Further indications are provided by the nature of the three
chapters which constitute this Book. Thus Keynes claims no novelty for the
72 Don Patinkin
explanation of the cycle which he presents in Chapter 22, entitled ‘Notes
on the Trade Cycle’ (GT: 314–15). Chapter 23 (entitled ‘Notes on
Mercantilism, the Usury Laws, Stamped Money and Theories of Under-
Consumption’) is, as its title indicates, a miscellanea. It is also significant
that, in contrast with the careful procedure that he followed with other
chapters of the General Theory (see JMK XIII and XXIX), Keynes
apparently made little effort to subject the drafts and/or proofs of this
chapter, as well as of Chapter 24, to the criticism of Kahn, Joan Robinson,
Hawtrey, and Robertson. Indeed, it would seem that the proofs of these two
chapters were sent only to Harrod, who sufficed with a brief comment on
Chapter 23 (to which Keynes did not react), and none on Chapter 24 (see
JMK XIII: 526, 542, 555; JMK XIV: 351).
In brief, Chapter 24 is part of a Book that is essentially an appendage to
the General Theory, and one should not let the appendage to a text wag its
body.
Let me turn now to Meltzer’s contention that in his policy prescriptions
throughout his career, Keynes was an advocate of rules as against discretion.
Strictly speaking, that contention is outside my terms of reference: for as I
have emphasized on earlier occasions (1976:12–13, 135–6; 1982:14, 212–
13)—and as Meltzer (1988:115) recognizes—the General Theory is (as
indicated by its title) a book that is devoted almost entirely to theory. I wish
nevertheless to deal with Meltzer’s discussion of this point because it
illustrates some additional pitfalls of interpretation.
To begin with, there is no evidence that Keynes ever gave specific thought
to the issue of rules vs. discretion, and I share Quentin Skinner’s (1969:32–4)
skepticism about attributing to earlier writers specific views about issues
with which they were never concretely confronted. Furthermore, even if
Keynes had thought of this issue, its meaning in his days was different from
what it is today. Specifically, when in 1936 Henry Simons published his
famous essay on ‘Rules versus Authorities in Monetary Policy’, the rule that
he advocated as a contracyclical policy was that of stabilizing the price level
by means of variations in the quantity of money—with no indication of how
the extent of this variation should in each case be determined. Clearly, this
policy would today be described not as one of rules, but of discretionary
‘fine-tuning’.
Let me for the moment leave this fundamental point aside and consider
the two pieces of evidence that Meltzer cites in support of his contention. The
first is the fact that from his earliest writings on through Bretton Woods,
Keynes advocated ‘rules for international monetary arrangements’ that
called for ‘fixed, but adjustable, exchange rates’ (Meltzer 1988:242–3; see
also pp. 209–11)]. Without committing myself one way or the other on the
validity of Meltzer’s treatment of this subject—Moggridge’s (1980: esp. 58–
60) treatment of it provides somewhat different emphases—let me simply say
that the policy issue of fixed vs. flexible exchange rates is one frequently
considered to be in a category of its own, unrelated to the issue of rules vs.
On different interpretations of the General Theory 73
discretion. Suffice it to note the views of Milton Friedman (1953:157ff.), who
with equal forcefulness has advocated both a monetary rule and flexible
exchange rates.
The second piece of evidence on which Meltzer attempts to base his
contention that Keynes was an advocate of rules as against discretion
consists of excerpts from Keynes’ contributions to the wartime discussions
that took place in the Treasury on the problem of maintaining full
employment in the postwar period. In this connection, Meltzer (1988:187)
cites Keynes’ statement in a 1943 letter to Meade (who played a major
role in these discussions) that ‘if the bulk of investment is under public or
semi-public control and we go in for a stable long-term programme,
serious fluctuations are enormously less likely to occur’ (JMK XXVII:
326), and on the basis of this statement contends that ‘Keynes’s
stabilization proposal did not depend on prompt changes in the amount of
public works’. Similarly, at a later point in his book, Meltzer (1988:294)
refers to his discussion in Chapter 4 and writes: ‘Keynes’s postwar fiscal
policy proposals…aim at stabilizing the rate of investment at a
permanently higher level’.
Now, the passage in Keynes’ letter to Meade which Meltzer cites is
actually a loose quotation from a memorandum that Keynes had just written
on ‘The Long-Term Problem of Full Employment’ and which he enclosed
with his letter. In particular, this memorandum contains the following
passage:
If two-thirds of three-quarters of total investment is carried out or can be
influenced by public or semi-public bodies, a long-term programme of a
stable character should be capable of reducing the potential range of
fluctuation [in the level of employment] to much narrower limits than
formerly.
(JMK XXVII: 322)
And what is significant about this passage for my present purpose is that it
does not specify any rule for the government to use in determining its actual
level of investment in the ‘two-thirds to three-quarters [range] of total
investment’ or in the ‘less than 71/2 per cent or more than 20 per cent
[range] of the net national income’ indicated in the passage in the
memorandum which follows. In brief, what Keynes regards as ‘stable’ is not
the level of ‘public or semi-public’ investment, but the need to carry it out.
And in this he was reflecting the policy conclusion of the General Theory
that ‘a somewhat comprehensive socialisation of investment will prove the
only means of securing an approximation to full employment’ (GT: 378; see
also p. 164).
That Keynes was not thinking in terms of a fixed rule that would
determine the level of government investment is also clear from the fact that
he goes on in the aforementioned memorandum to say that:
74 Don Patinkin
The main task should be to prevent large fluctuations by a stable long-
term programme. If this is successful it should not be too difficult to offset
small fluctuations by expediting or retarding some items in this long-term
programme.
(JMK XXVII: 322, original italics)
Similarly, in a related 1943 memorandum, Keynes wrote:
if we can find ways of retarding or accelerating the long-term programme to
offset unforeseen short-term fluctuations, all the better. No reason, surely, why
the Treasury should not be fairly constructive and optimistic on this heading.
(JMK XXVII: 356; see also the similar statement on p. 323; see also
the first paragraph of his letter to Meade reproduced on p. 319)
Thus, unlike many of today’s advocates of, say, the constant-money-growth
rule, Keynes’ advocacy of a ‘long-term programme of a stable character’ did
not exclude the desirability of also carrying out short-term discretionary
policies when necessary.23
All this is brought out more clearly and systematically in the historic May
1944 White Paper on ‘Employment Policy’ which (I think it fair to say) largely
reflected Keynes’ view on how to carry out the full-employment policy that he
advocated—or which (at the very least) advocated policies to which he had no
basic objections. Indeed, not only did Keynes write the foregoing memoranda
in the context of the Treasury discussions which culminated in this White
Paper, but he even prepared notes for the Chancellor of the Exchequer to use in
presenting it to the House of Commons (see JMK XXVII: 374–9). Now,
Chapter V of the White Paper is entitled ‘Methods for Maintaining Total
Expenditure’ at its full-employment level and inter alia outlines ‘the measures
by which the Government propose, as part of their long-term policy, to
influence the volume of capital expenditure, private and public’ (ibid., §57,
italics added). The chapter then goes on to say that
for the purpose of maintaining general employment it is desirable that
public investment should actually expand when private investment is
declining and should contract in periods of boom…. The procedure which
the Government have in mind is as follows. All local authorities will
submit annually to the appropriate Department their programme of
capital expenditure for the next five years…. These programmes will then
be assembled by an appropriate co-ordinating body under Ministers and
will be adjusted, upward or downward, in the light of the latest
information on the prospective employment situation…. The machinery
envisaged in this paragraph will enable the Government to set each year a
target for the whole volume of public works in the succeeding year.
In order that public investment may be more quickly mobilized to
redress the balance of private investment the Government also intend to
On different interpretations of the General Theory 75
seek means of reducing the time-lag which ordinarily intervenes between
a decision to undertake public capital expenditure and the actual start of
the work. Speed here is crucial, for if a decline in demand can be caught
quickly enough and corrected, a comparatively modest amount of
compensating expenditure will be sufficient to restore the balance…. The
Government believe that in the past the power of public expenditure,
skilfully applied, to check the onset of a depression has been
underestimated.
(ibid. §§ 62–6)
And, significantly enough, this passage echoes one of the comments that
Keynes made in February 1944 on what was essentially a draft of the White
Paper.24, 25
I have quoted these passages in extenso in order to clarify the nature of
Keynes’ policy proposals, and in particular the meaning of the phrase ‘long-
term policy’. The picture that thus emerges from Keynes’ wartime writings is
quite at variance with Meltzer’s contention (1988:293) that they present
Keynes as a ‘proponent of discretionary action constrained by well-defined
policy rules’—and my emphasis is of course on ‘well-defined’. Thus, to say
the least, Keynes’ wartime memoranda and correspondence provide little if
any support for Meltzer’s repeated contention that they show that Keynes
‘did not favor the discretionary policies to manage short-term changes in
aggregate demand that are called Keynesian’ (ibid.: 308; see also pp. 4–5,
122, 295)—whose nature Meltzer does not specify, but by which he
presumably means policies of the kind advocated in the United States by
(say) Robert Solow and James Tobin.
Meltzer (1988:294) also tries to create the impression that Keynes was not
in favor of ‘deficit financing’. Here, however, Meltzer fails to take account
of the fact that there is a fundamental difference between his usage of this
term and Keynes’. For as part of his postwar proposals, Keynes
recommended dividing the overall government budget into a current (or
ordinary) budget (which ideally would be financed entirely by taxes) and a
capital budget (within which framework the government would carry out its
long-term program of investment, which would be financed primarily by
borrowing). And he used the term ‘deficit financing’ only with respect to
borrowing to finance a deficit in the ordinary budget.
This basic distinction is spelled out in the following passage from a 1945
memorandum in which Keynes advocated the establishment of a separate
capital budget:
It is important to emphasise that it is no part of the purpose of the
Exchequer or the Public Capital Budget to facilitate deficit financing, as I
understand this term. On the contrary, the purpose is to present a sharp
distinction between the policy of collecting in taxes less than the current
non-capital expenditure of the state as a means of stimulating
76 Don Patinkin
consumption, and the policy of the Treasury’s influencing public capital
expenditure as a means of stimulating investment.
(JMK XXVII: 406, first set of italics added)
Keynes then goes on in this memorandum (ibid.: 406–8) to list the various
kinds of receipts on capital account to finance the capital expenditures,
including, of course, loans from the public (‘net receipts…of public debt held
by the private sector’).26
And the distinctly different roles that Keynes assigned to these two budgets
in his contracyclical policy is brought out most clearly in the following
comment that he made in 1942 on proposed postwar budgetary policy: ‘I
should not aim at attempting to compensate cyclical fluctuations by means
of the ordinary Budget. I should leave this duty to the capital budget’ (JMK
XXVII: 278; see also pp. 352–3). Thus it was only in his specific sense of the
term that Keynes objected to ‘deficit financing’. In, however, the sense in
which it has been used in (say) the United States and the United Kingdom
(where the budget is an overall one, which includes both current and capital
expenditures), the contracyclical policy that Keynes advocated can only be
described as one of deficit financing.
Let me now return to Meltzer’s interpretation of Keynes’ views on the
issue of ‘rules vs. discretion’ and note that, over and above my foregoing
criticisms, is the fundamental question that I raised above as to the
meaningfulness of attributing to writers views about issues with which they
were never concretely confronted. And in the case of Keynes, this question is
particularly relevant. For Keynes was—and is—well known for the fact that
he changed his policy recommendations in accordance with changing
circumstances. Thus in 1929–30 he advocated public works, as against a
lowering of the rate of interest, as the means of dealing with Britain’s
problem of unemployment; but after Britain abandoned the gold standard in
1931, he advocated doing so by lowering this rate (the ‘cheapmoney policy’);
and when by 1933 this failed to produce the desired results, he returned to
the advocacy of public works.27
Now, the postwar world which Keynes generally envisaged in the various
wartime documents referred to above was one—in the words of the General
Theory (p. 249)—‘capable of remaining in a chronic condition of subnormal
activity for a considerable period without any marked tendency either towards
recovery or towards complete collapse’. So even if we accept the validity of
Meltzer’s interpretation of Keynes’ views in these documents on the issue of
‘rules vs. discretion’ (and I have expressed serious doubts about that), there is
little if any validity in his attempt to infer from it how Keynes would today
regard this issue for the postwar world of vastly different circumstances that
has actually emerged: a world which, far from remaining in any ‘chronic
condition’, has experienced prolonged investment booms and unprecedented
rapid growth in some periods, and unemployment and recessions in others; a
world which in different periods has experienced different degrees of both
On different interpretations of the General Theory 77
demand-inflation and cost-inflation, accompanied sometimes by full
employment and sometimes by different rates of unemployment; a world with
an international monetary system which for the past two decades has been in a
constant state of flux, and thus vastly different from the relatively stable
system that Keynes envisaged at Bretton Woods.
Might I finally say that if one nevertheless insists upon making inferences
from Keynes’ writings about what his view would today be on the issue of
‘rules vs. discretion’ (and needless to say, I do not), then a more appropriate
basis for such inferences is his 1930 Treatise on Money, which deals with a
world that experiences significantly changing circumstances, a world subject
to both booms and depressions, to both inflation and deflation. The
contracyclical policy that Keynes advocated for this world was one that he
termed ‘The Management of Money’ (title of Book VII of the Treatise) and
described in the first chapter (31) of this Book as follows:
Thus the art of the management of money consists partly in devising
technical methods by which the central authority can be put in a position
to exercise a sensitive control over the rate of investment, which will
operate effectively and quickly, and partly in possessing enough
knowledge and prognosticating power to enable the technical methods to
be applied at the right time and in the right degree to produce the effects
on prices and earnings which are desirable in the interests of whatever
may be the prescribed ultimate objective of the monetary system which is
being managed.
(Treatise II: 189–90)
And the ‘technical method’ which Keynes went on to specify in Chapter 37 of
the Book—after having taken due account of the fact that the monetary
authorities could not produce effects ‘instantaneously’, and could not be
‘expected always to foresee the operation of non-monetary factors in time to
take measures in advance to counteract their influence on prices’—was to
stabilize the price level by central-bank monetary policy in the form of
variations of the short-term rate of interest, with the purpose of influencing
the long-term rate and hence the level of investment (see ibid., especially pp.
304–5, 309–10, 315–16, 325–35): in brief, a policy that (like that of Henry
Simons’ described above) would today clearly be regarded as an example of
discretionary ‘fine-tuning’.28
VII
I conclude this chapter with an attempt to answer the question with which I
began: why are there different interpretations of the General Theory? Why
are there not different interpretations of Value and Capital and Foundations
of Economic Analysis? All of these are canonical texts, and canonical texts
attract different interpretations. A partial answer to this question is to be
78 Don Patinkin
found in the fact that there is a fundamental difference between these texts.
For without minimizing their basic contribution to our discipline, Value and
Capital and Foundations were books that elaborated, rigorized and extended
economic theory within the existing paradigm. Furthermore, they were books
that presented their analysis in mathematical terms, thus leaving little if any
ambiguity as to their intended meaning.
Not so the General Theory: here was a book which presented a new and
at-the-time strange paradigm. It was a pioneering work that introduced new
concepts (e.g. the very notion of an ‘aggregate demand function’) and new
ways of thinking (see Patinkin 1976:83, 98–9). And so it is not surprising
that it left some obscurities and even inconsistencies, as well as some loose
ends. Indeed, it never pulled together its various analytical components into
an explicit and complete model: this task was left for its contemporary
interpreters.
An equally if not more important difference is the fact that the works of
Hicks and Samuelson had no political implications. This is not true from the
Marxist viewpoint, which presumably regards them as rationalizing and hence
justifying the functioning of a capitalist market economy. It is however true
from the viewpoint of the Western democratic society to which these works
were addressed: for neither of them expressed any specific view as to the extent
and manner in which government should intervene in such an economy. In
contrast, the General Theory had a clear political message: government
intervention in the form of contracyclical public and semi-public investment was
necessary to assure the existence of full employment in a capitalist economy;
and only after full employment was thus assured could the market mechanism
be relied upon to function without further intervention (GT: 378–9).
And these are the differentiae of the General Theory which have created
fertile ground for different interpretations.
Now, with two of the contemporary interpreters of the General Theory,
Joan Robinson at Cambridge and Roy Harrod at Oxford, Keynes had carried
out intensive discussions of earlier drafts of the book (see JMK XIII and
XXIX, passim). Three others (Bryce, Champernowne, Reddaway) had been
students who had attended the lectures that Keynes had given in the process
of writing it (see Rymes 1989: ix; JMK XIV: 59; Austin Robinson 1977:33).
Others (Hicks, Lerner, Meade) had in one way or another been aware of the
work in progress.29 And all of the contemporary interpreters were directly
experiencing the seemingly endless years of the Great Depression which
formed the background of the General Theory. So despite its ambiguities,
there was little difference between them as to both its analytical and political
message.
A quarter of a century later, however, and a fortiori half a century later,
some of the interpreters and most of their readers were of a generation who
knew not the Great Depression. Furthermore, the memories of those years
had in general dimmed in a postwar world facing economic problems that
were vastly different from the prewar one. And—no less important– Keynes
On different interpretations of the General Theory 79
was no longer around to defend himself against various would-be interpreters
(as he had on at least one occasion, when in April 1938 he concluded a long
correspondence with E.S.Shaw on a note that the latter had sent him on the
‘finance motive’, with the comment that ‘there is a good deal in it which I
cannot accept as anything like an accurate version of what I am driving
at…. I am really driving at something extremely plain and simple which
cannot possibly deserve all this exegesis’ (JMK XXIX: 281–2). In these
circumstances the obscurities and loose ends of the General Theory provided
ample opportunities for different interpreters with different—and quite
familiar—motivations: in particular, some interpreters wanted to invoke the
authority of a canonical text in support of their prior theoretical views; and
others wanted to invoke the authority of a canonical figure in support of
their prior political views.30
Milgate, Shackle and Weintraub were clearly motivated by the first of
these considerations. In section V above, I have said enough about Milgate’s
attempt to recreate Keynes in the image of the Modern Cambridge School.
Insofar as Shackle is concerned, already in his 1938 book on Expectations,
Investment and Income he had emphasized the importance of expectations
and uncertainty in economic life; at the same time, while acknowledging his
indebtedness to Chapter 12 of the General Theory, he did not claim that the
‘animal spirits’ of this chapter constituted the central message of that book.
Quite the contrary: he went on to describe the analysis of the book in a way
that is completely consistent with Hicks’ IS-LM interpretation. In Shackle’s
words at that time:
[Keynes’] General Theory of Employment, Interest and Money really
deals with the formal interdependencies of economic variables at a
moment of time. Expectations, the quantity of money, and the schedules
of propensity to consume and of liquidity preference being all given, there
is a certain level of the investment-flow, which, in view of the aggregate
income corresponding to it, and the given quantity of money, will both
evoke and be evoked by a certain rate of interest. In other words, from the
knowledge specified we can determine values of certain main economic
variables which are mutually consistent and can hold simultaneously.
(Shackle 1938:2, original italics)
Similarly, Shackle’s 1949 book on Expectation in Economics (p. 60, footnote)
contains only one passing reference to the General Theory and, what is even
more significant, does not contain any reference to Keynes’ 1937 Quarterly
Journal of Economics article. To the best of my knowledge, it was only in a
1953 article on ‘A Chart of Economic Theory’ (pp. 217, 222–4) that we find
the first indications of Shackle’s interpretation of the General Theory cum QJE
1937 article that has been described in section IV above.
A similar story, albeit over a much shorter time span, and somewhat more
complicated, holds for Sidney Weintraub and his wage-cost-markup theory of
80 Don Patinkin
price. In order to present this story in its proper perspective, let me first note
that there is little concern with the theory of price per se—and a fortiori little
concern with this specific theory—in Weintraub’s 1958 book on An Approach
to the Theory of Income Distribution, whose main concern is precisely what
its title says (but see p. 57, n. 16, and pp. 100–1). And though this book
contains many references to the General Theory, I think it fair to say that
Weintraub did not contend that the concern of his book was a major one of
the General Theory. In any event, it is a fact that the latter contains no
significant discussion of the distribution of income; and that, indeed, at the
beginning of the General Theory (p. 4), Keynes distinguishes sharply
between his concern with the level of total output and the ‘Ricardian
tradition…[which] expressly repudiated any interest in the amount of the
national dividend, as distinct from its distribution’ (ibid., n. 1, original
italics).
There is, however, one respect in which Weintraub’s 1958 book
foreshadows the interpretation of the General Theory that he was to present
in his 1961 book on Classical Keynesianism, Monetary Theory, and the Price
Level, for in Chapter 2 of the 1958 book, Weintraub develops the properties
of what he denotes as aggregate demand and aggregate supply curves and
states (ibid., p. 24) that ‘in a way, then, this chapter constitutes a restatement
of the main part of macroeconomic theory [and here a footnote provides
references to various articles and notes on the aggregate supply curve]. Like
all such analysis, its origin is to be found in Keynes’s General Theory’.31 On
the other hand, and in sharp contrast with his 1961 book, Weintraub’s 1958
book (ibid.: 153–6) presents Hicks’ IS-LM analysis without in any way
criticizing it.
We come now to Weintraub’s 1959 book on the General Theory of the
Price Level, Output, Income Distribution, and Economic Growth, whose
central message is precisely his theory of price determination by wage-cost-
markup. And in the Preface to the book, he describes this theory as ‘a
glimmer of an idea that had been stirring in [his] mind for some time…a
single idea…able to unify important parts of the theory of the price level,
output, income distribution and growth theory’. It is, however, most
significant that Weintraub does not yet ascribe this theory to Keynes.
It is only in his article a year later on ‘The Keynesian Theory of Inflation:
The Two Faces of Janus?’ (1960) that Weintraub (after referring to his 1959
book) does make such an ascription. And here he went on to describe the
relation of his wage-cost-markup theory of price to Keynes’ in the following
words: ‘I think I “know” what Keynes would have emphasized, but I refrain
from any attempt at documentation, for it is equally possible for those who
hold alternative views to find appropriate supporting passages in Keynes’
General Theory or his later discussion in How to Pay for the War to sanction
their doctrinal interpretation’ (1960:144). A year later this article was
reprinted as Chapter 2 of his Classical Keynesianism, Monetary Theory, and
the Price Level (1961), in which the foregoing quotation appears on p. 27. A
On different interpretations of the General Theory 81
similar ascription to Keynes—which has already been cited on p. 64 and in
note 9—also appears in Chapter 1 of the book (1961:3). And in presenting
his interpretation of the General Theory in both the article (1960: § § 1–2)
and the book (1961:5–10, 18–22, 35–8), Weintraub emphasizes the role of
the aggregate supply curve in Keynes’ analysis, and now firmly rejects both
the 45°-cross diagram and the IS-LM one.
Needless to say, the distinction that I have made between theoretical and
political motivations is not an absolute one. Thus Milgate’s attempt to
identify Keynes’ theory with that of the ‘Modern Cambridge School’ is also
an attempt to invoke Keynes’ authority for the ‘much more interventionist
stance’ of this School (Milgate 1982: preface). Similarly, Weintraub’s 1961
interpretation of the General Theory as having a primary concern with
wage-cost-push inflation was clearly related to his policy proposals—in
opposition to those that he attributed to the ‘Keynesians’—for dealing with
the inflation of the 1950s (1961:5–9).
Meltzer’s book too reflects a combination of both motivations. Thus at
many points in it (see e.g. Meltzer 9, 15, 177, 184, 207) he invokes Keynes’
authority for his (Meltzer’s) specific theoretical view of uncertainty as
imposing on the economy an ‘excess burden’ that manifests itself in a lower
rate of investment, and hence lower stock of capital, than would otherwise
exist. But what stands out most in Meltzer’s interpretation is his political
motivation, and particularly his desire to claim the political mantle of
Keynes for the conservative rules-as-against-discretion type of contracyclical
policy that he advocates (see e.g. Meltzer 1983, 1987; Brunner and Meltzer
1983: esp. 97–100), and for the related conservative opposition to ‘deficit
financing’. And no less important is his desire to deny that mantle to the
‘Keynesians’: to drive a wedge between ‘Keynes and the Keynesians’ by
denying that the latter are the legitimate heirs of the Master’s views on the
nature of the contracyclical policy that should be adopted.
Let me conclude this discussion of politically motivated interpretations by
noting that the nature of such motivations is well illustrated by the following
passage:
On questions of policy the differences can never be resolved. Even such an
apparently simple problem as, for instance, the extension of public works
as a remedy for unemployment, is found to give rise to violent conflicts of
interest…. Revolutionaries who regard unemployment as only one of the
evils of a system of private enterprise are not anxious for capitalist
governments to learn the trick of reducing fluctuations in trade, and so
deprive them of the most obvious, though not the most fundamental, of
their objections to the system. The adherents of laissez-faire, on the other
hand, fear that, if it once became clear to the public that state interference
can reduce unemployment, the public might begin to think that state
interference could do much else besides.
(Robinson 1937:126–7)
82 Don Patinkin
This is a passage from the Introduction to the Theory of Employment by
Joan Robinson—in her prewar period.
VIII
At the beginning of this chapter I expressed the view that though the original
meaning which an author intended might not be the only legitimate one, at
the same time this meaning can be used to justify the rejection of
interpretations that differ greatly from it. And now I leave it for you the
reader to decide which if any of the interpretations of the General Theory
that I have discussed—including my own version of the IS-LM
interpretation—should be rejected on those grounds.
ACKNOWLEDGEMENTS
This chapter originated in my Keynes Lecture delivered at the British
Academy in November 1989. An augmented version of the lecture appears in
the Proceedings of the Academy for 1989. The present chapter contains some
minor changes and additions to that version, and is published here with the
kind permission of the Academy.
Work on this lecture was begun while serving in September 1989 as the
James S.McDonnell Scholar at the World Institute for Development
Economics Research (WIDER) in Helsinki. During this visit I benefited
greatly from discussions with my fellow McDonnell Scholars, Frank Hahn
and Robert Solow. I am indebted to both the James S.McDonnell Foundation
and WIDER for making possible this most fruitful visit.
Without in any way burdening them with responsibility for the end
product, I wish to express my deepest appreciation to Menahem Brinker,
Geoffrey Hartman and Ilana Pardes for guiding my reading in the field of
hermeneutics, and for patiently discussing with me many of the problems
that I there encountered. Again, without burdening them with any
responsibility for the views expressed, I am also indebted to Brinker and
Pardes, as well as to Chaim Barkai, Stanley Fischer, David Laidler, Donald
Moggridge, Luigi Pasinetti, Robert Skidelsky, and Roy Weintraub, for
valuable criticisms of earlier drafts of this chapter. Similarly valuable
comments were received on the draft presented at the Department of
Economics Seminar of the Hebrew University of Jerusalem. Finally, I wish to
thank Vivian Nadir for her most pleasant and efficient technical assistance.
Reprinted articles are referred to by the date of original publication; for
convenience, however, the page references to such articles are to the reprint. All
references to the writings of Keynes are to the relevant volumes of the Royal
Economic Society’s edition of his Collected Writings. These volumes are,
respectively, referred to as JMK XIII, JMK XXVII, and so forth. The Treatise on
Money is frequently referred to as the Treatise; and the General Theory of
Employment, Interest and Money is referred to as the General Theory or GT.
On different interpretations of the General Theory 83
This research was supported by the Basic Research Foundation
administered by the Israel Academy of Sciences and Humanities, to which I
express my thanks. I also wish to thank the Israel Academy itself for the
pleasant and helpful atmosphere in which most of the work on this chapter
was carried out.
NOTES
1 Note that the 1979 exchange between Coddington and Hicks deals—as indicated
by the title of Coddington’s article—with ‘Hicks’s Contribution to Keynesian
Economies’, and not with Value and Capital proper.
2 For an example of the use of probabilistic notions to draw a hypothetical regression
line in order to determine what I have called the central message of a text, see
Patinkin (1982, pp. 16–18).
3 The relevant passage in this entry reads: The meeting was devoted to a long academic
lecture by Keynes on the whole of the General Theory—the equality between
savings and investment; the inducement to invest; the propensity to consume; and
liquidity preference’ (Meade 1990:48). See also Keynes’ notes for this ‘lecture’ as
reprinted in JMK XXVII: 388ff.
4 See the letters (written at various dates during the period 1935 to 1938) reproduced
in JMK XIV to Joan Robinson (pp. 148–50), Harrod (pp. 84–6), Reddaway (p.
70) and Hicks (pp. 70–83); as well as the letters in JMK XXIX to Bryce (p. 150) and
Lerner (pp. 214–16). See also Hicks (1973).
Though in the correspondence with Robertson on his review Keynes did refer
to the former’s discussion of the aggregate supply curve (JMK XIV: 89), he did not
do so in his published reply.
5 In this footnote, Keynes goes on to say: ‘The analysis which I gave in my General
Theory of Employment is the same as the ‘general theory’ explained by Dr. Lange
on p. 18 of this article [corresponding to p. 176 of the reprint], except that my
analysis is not based (as I think his is in that passage) on the assumption that the
quantity of money is constant’.
6 Harrod, Hicks and Meade had all presented their reviews as papers at the September
1936 European meetings of the Econometric Society at Oxford: see the report by
Phelps Brown (1937). For a fascinating behind-the-scenes account of these meetings,
see Chapter 1 of Young (1987), who inter alia reproduces a letter from Hicks to
Meade (ibid: 33 and 35) which shows that Hicks had read both Meade’s and Harrod’s
papers before completing his own IS-LM paper, which is the one he presented at the
meetings. Young (ibid.: 33 and 98 ff.), however, also emphasizes that Hicks never
claimed originality for the IS-LM equations, but only for the diagram.
Having mentioned Young’s book, I must add that at some points it reflects the
‘Modern Cambridge School’ cum Post-Keynesian interpretation of the General
Theory which is criticized in sections V and VI of this chapter.
I might also note that in a March 1937 letter to Joan Robinson (reproduced in
JMK XIV: 149), Keynes wrote that he did not ‘feel any objection’ to her publishing
her Introduction to the Theory of Employment (1937), Chapter 2 of which presents
an interpretation of the General Theory which is essentially a verbal rendition of
the subsequent elementary-textbook 45°-cross diagrammatic explanation of the
determination of the equilibrium level of income.
84 Don Patinkin
7 But see pp. 111–13 of my Keynes’ Monetary Thought for a discussion of what
may possibly be some ambivalence on this point in the General Theory.
8 It is this conclusion which led Coddington (1983:93ff.) to designate the
interpretation of Keynes by Shackle and those who have been influenced by him
(see section V of this chapter) as ‘fundamentalist Keynesianism’, in the sense that it
interprets the General Theory as being fundamentally opposed to traditional
equilibrium theory. Coddington himself criticized this view (ibid.: 88, 97–100); see
also note 10 below.
Coddington (1983:102 ff., which reproduces the corresponding material of his
1976 article) contrasted this approach with that of ‘hydraulic Keynesianism’, which
is essentially his term for mainstream Keynesianism as represented by (say) IS-LM.
His explanation of his choice of this term is that it reflects the fact that this approach
describes the macroeconomy ‘in terms of disembodied and homogeneous flows’
between which ‘there exist stable relationships’. Readers of Coddington’s
posthumously published book have sometimes wondered whether his choice of
this term was also influenced by the hydraulic mechanism that A.W. Phillips—with
the assistance of W.T.Newlyn—constructed to illustrate the workings of what is
essentially the IS-LM model of the General Theory (see Phillips 1950). In reply to
a query of mine on this point, Mark Blaug has informed me that he once raised this
question with Coddington and was told that ‘the idea came from his memory of
the Phillips machine, which he had never seen but which was legendary at L.S.E.’
(cited from Blaug’s letter with his kind permission). Susan Howson, however, has
pointed out to me that Coddington might also have been influenced by Shackle’s
reference in his Years of High Theory (1967:189) to the ‘misleading’ ‘reduction of
economics to hydraulics’, and that Shackle himself, according to Arthur Brown
(1988), was ‘no doubt thinking of the Phillips-Newlyn model, which arrived in
Leeds shortly before he did’ (ibid.: 38). I am indebted to Howson for this additional
information, as well as for referring me to Nicholas Barr’s illuminating article on
‘The Phillips Machine’ (1988). Barr (ibid.: 330–4) reports that the machine fell
into disuse in the late 1950s, but that a process of renovation was undertaken in
1987. Tony Atkinson has kindly informed me that the renovated machine was
unveiled in 1989 and is on display once again at LSE (see postscript p. 94).
I might note that a hydraulic mechanism to illustrate the working of economic
principles was first constructed by Irving Fisher in 1893 in connection with his
1892 doctoral dissertation on Mathematical Investigations in the Theory of Value
and Prices. In Fisher’s case, however, the principles illustrated were the utility-
maximizing conditions of general-equilibrium analysis. Pictures of the 1893
mechanism, as well as a ‘somewhat improved and simplified’ one constructed in
1926, appear as frontispieces to the 1925 reprint of Fisher’s dissertation. (The
quotation is from Fisher’s Preface to this reprint.)
9 In all fairness, I should add that this passage goes on to say:
I am aware that other Keynesians can select other passages to corroborate their
doctrinal position so that such controversies on ‘what Keynes really meant’, like
those on ‘what Marx—or Marshall—really meant’, are likely to be peculiarly
barren and futile. It is with Keynesianism then, not with Keynes, that I am
concerned. For myself I acknowledge full indebtedness to his tremendous work,
rather than to deny or camouflage it as is becoming fashionable. My ideas, like
those of all modern Keynesians, emanate from it.
On different interpretations of the General Theory 85
10 See also Coddington’s (1983:59–60) implicit criticism of Shackle’s interpretation
of Keynes’ 1937 article.
11 For further evidence that in using the phrase ‘animal spirits’ Keynes did not intend
to imply that ‘the determination of investment is entirely arbitrary’, see Robin
Matthews (1984:209–12), who also cites supporting evidence from Joan Robinson.
Matthews (ibid.) also reproduces the following interesting information on this
phrase that he received from Donald Moggridge:
The origins of ‘animal spirits’ seem to go a long way back in Keynes. The earliest
reference comes in a set of lecture notes which are in the Marshall Library
Collection, entitled ‘Notes on Modern Philosophy I—Descartes, Leibnitz,
McTaggart’s Lectures, Ertemann’s [sic] History—[includes Spinoza’s Ethics]’. In
the part concerning Descartes as regards life and biology the text runs: ‘The body
is moved by animal spirits—the fiery particles of the blood distilled by the heat of
the heart. They move the body by penetrating and moving the nerves and
muscles…. But does not this increase the amount of motion? No. for the animal
spirits are always in motion—the will only directs them’.
Keynes then adds a comment that reads ‘unconscious mental action’.
In reply to a query from me, Moggridge has kindly supplemented this
information with the following:
The documents referred to are a set of handwritten reading/lecture notes taken
by JMK…The notes are in JMK’s papers, then in the Marshall [Library] and now
in King’s [College Library], JMK, according to Harrod (1951, p. 61) went to
J.E.McTaggart’s lectures on general philosophy during the 1902–3 academic
year. Erdmann’s philosophy is presumably J.E.Erdmann’s A History of
Philosophy, an English translation of which appeared in 1890.
12 It is in this longer-run context that I also interpret Keynes’ statement in a May 1936
letter to Hubert Henderson (dealing with the latter’s criticism of the General Theory):
‘I should. I think, be prepared to argue that in a world ruled by uncertainty with an
uncertain future linked to an actual present, a final position of equilibrium, such as one
deals with in static economics, does not properly exist’ (JMK XXIX: 222, italics added).
13 Here is another problematic aspect of Weintraub’s interpretation: for this attributes to
Keynes a ‘theory of the price level’ in terms of a ‘wage-cost markup’ (Weintraub
1978:64). This may be the reason Weintraub prefaced his attribution with the comment
that in doing so he was ‘taking only mild liberties with Keynes’. Surely, however, there
is a fundamental difference between the cost-markup theory of price and the marginal-
cost theory of price, and it is the latter that Keynes consistently uses throughout the
General Theory (see e.g. p. 55 (bottom), p. 283 (lines 4–5) and pp. 294–295).
For my present purposes, however, this is a secondary issue, my major point
being that in any event the theory of price determination is a subsidiary concern of
the General Theory.
14 I must admit to some inconsistency here: for the subtitle of Chick’s book is ‘A
Reconsideration of the General Theory’. On the other hand, in the preface to her
book (p. vii) she states:
This is not a book in the history of economic doctrine as such, which is concerned
with illuminating the author’s point of view as brightly as possible on his own
terms. I hope at several points to have done that, though I do not claim that this
86 Don Patinkin
book reveals ‘what Keynes really meant’…it is a philosophical impossibility to
know what someone else ‘really meant’; what matters is to make coherent sense
for oneself of what an author says and to evaluate its relevance to the problem at
hand. (italics in original)
This last sentence clearly reflects an implicit acceptance of the deconstructionist
approach to hermeneutics: see section I of this chapter.
I should also note that in what may in part have been Leijonhufvud’s reaction to
the criticisms of his interpretation of the General Theory presented in the review
articles of his book by Grossman (1972) and Jackman (1974), Leijonhufvud (on
pp. i-ii of his English foreword to the 1978 Japanese edition of his book) emphasized
that his book was about ‘theoretical problems that were current problems in the
early or mid-sixties…. What Keynes might have meant, etc., was not one of the
problems. Doctrine history was not what the book was about’.
15 See the references to these economists in the writings of such Post-Keynesians as Davidson
(1972; 1981:154–5) and Eichner (1979). Shackle’s interpretation is also implicitly accepted
by Joan Robinson in her paper on ‘What Has Become of the Keynesian Revolution’
(Robinson 1973:3, n. 1, and text to which it is attached) and in her Foreword to Eichner
(1979: xi). Though he does not refer explicitly to Shackle, Minsky (1975:38) rejects the
IS-LM interpretation of the General Theory on the same grounds that Shackle did. On
Weintraub, see also the survey article by Eichner and Kregel (1975).
16 As is also its contention that Kalecki independently discovered the General Theory—
a contention that I have examined in detail and rejected in Chapter 3 of my
Anticipations of the General Theory? (Patinkin 1982).
17 On which, see Frank Knight’s (1923) classic critique.
18 Another example of Milgate’s approach to the interpretation of the General Theory
is provided by his discussion of Keynes’ criticism of the classical theory of interest
in Chapter 14 of the book. Here Milgate (1982:111) contends that this criticism
‘has been progressively obscured by conventional interpretations’. And he then
presents his alternative interpretation, which is based on his objection to the
‘unqualified acceptance of the final text of Chapter 14’ (ibid.: 122) and on his
resort instead (ibid: 111–23) to his interpretation of the earlier draft of this chapter
and the related Keynes-Harrod correspondence reproduced in JMK XIII.
19 On the same Shacklian grounds, Robinson could also have applied this epithet to
the 45°-cross interpretation of the General Theory presented in Chapter 2 of her
1937 Introduction to the Theory of Employment (see note 6 above). In particular,
Robinson concluded the exposition of this chapter with the following words:
To sum up:…It is through changes in income that the equality of saving and
investment is preserved. Thus the level of income is determined by the rate of
investment and the desire to save; given the desire to save, the level of income that
will rule is governed by the rate of investment. And given the rate of investment
the level of income is determined by the desire to save.
(1937:16)
In a 1970 note on ‘Quantity Theories Old and New’, Robinson also objects to IS-LM
on the grounds that it provided ‘a mollifying version of his [Keynes’] system of ideas
which turned it back once more into a variant of the quantity theory’ (ibid.: 507)!
20 I Kings 14:16. See Joan Robinson’s (1977:10) statement about ‘the IS/LM model
with which generations of students have been taught to misinterpret the General
On different interpretations of the General Theory 87
Theory’. Robinson’s (1978) statement about Hicks’ having ‘repented’ is quoted
approvingly by Richard Kahn (1984:160), who adds that Keynes’ 1937 letter to
Hicks left Hicks ‘unrepentant’. This religious overtone is not unusual in Post-
Keynesian writings. Thus Sidney Weintraub (1976) entitled his criticism of IS-LM
‘Revision and Recantation in Hicksian Economies’. Similarly, Paul Davidson
(1989:23, n. 3) claimed that Axel Leijonhufvud ‘recanted’ on a certain point that
he (Davidson) considered to be at variance with Post-Keynesian teachings.
21 For another instance of disregarding ‘retroactive intent’—even when it is that of
God Himself—see the wonderful Talmudic story cited at the beginning of my ‘In
Defense of IS-LM’ (1990).
22 In a 1943 letter to Josiah Wedgwood, Keynes also said that ‘it would be in the
interests of the standards of life in the long run if we increased our capital quite
materially’ (JMK XXVII: 350).
23 This would also seem to be the conclusion that Dimsdale (1987:224) draws from
his study of Keynes’ wartime memoranda: ‘His [Keynes’] views on the design of
counter-cyclical policy were a development of the ideas which he had put to the
Committee on Economic Information from 1935 onwards. He favoured the use of
fiscal measures, based on variations in public and semi-public investment, to offset
the trade cycle’.
24 Namely, the comment in which he advocated preparing
a regular survey and analysis of the relationship between sources of savings and
different types of investment and a balance sheet showing how they have been
brought into equality for the past year, and a forecast of the same for the year to
come. If aggregate demand gave signs of being deficient, the analysis would indicate
a deflationary gap exactly corresponding to the inflationary gap which we have
so often discussed during the war. This survey and balance sheet…would give an
annual opportunity for examining whether the state of demand during the ensuing
year looked like being adequate to maintain employment and national income at
the desirable level and for the Government to explain to Parliament what steps it
had in view to remedy a prospective disequilibrium in either direction.
(JMK XXVII: 368–9)
25 In the sentence before the one that Meltzer quotes from Keynes’ aforementioned
letter to Meade, Keynes said that ‘it is quite true that a fluctuating volume of public
works at short notice is a clumsy form of cure and not likely to be completely
successful’ (JMK XXVII: 326). And I think that Keynes was here making a distinction
between public works carried out ‘at short notice’ and those that would be carried
out in accordance with the ‘long-term programme’ that he advocated in the
following sentence of the letter.
26 The financing by borrowing of the public-works, expenditures that he advocated
for dealing with depressions was a consistent feature of Keynes’ policy thinking.
Thus see his statement in a 1933 letter to The Times that ‘I contemplate that public
works would be paid for out of loans’ (JMK XXI: 200), as well as his description
of the financing of the government ‘programme of domestic investment’ that (under
certain conditions) he advocated in the Treatise (II: 337–8). Indeed, in his prewar
writings Keynes frequently referred to such public-works expenditures as ‘loan
expenditures’: see e.g. the 1929 Can Lloyd George Do It? (JMK IX: esp. 115–21;
written together with Hubert Henderson), the 1933 Means to Prosperity (JMK IX:
346–50, 354–5, 364–6), the General Theory (p. 128, n. 1), and the 1937 ‘How to
88 Don Patinkin
Avoid a Slump’ (JMK XXI: 390). The discussion in Means to Prosperity (JMK IX:
bottom of p. 347) also makes it clear that when Keynes speaks of ‘balancing the
budget’, he is not including ‘loan expenditures’ in the budget.
27 This is an obvious oversimplification of what was actually a complex sequence of
events. For details, see Patinkin (1982: ch. 8) and references there cited. In early
1931, Keynes also recommended the imposition of tariffs, and dropped this
recommendation after the devaluation a few months later.
28 Note that this fact in itself constitutes a refutation of Meltzer’s contention (1988:8)
that ‘from the [1923] Tract [on Monetary Reform] through his subsequent major
works’, Keynes ‘favored principal reliance on rules, with strict limits on discretionary
action and policy surprises’.
29 Joan Robinson and others had discussed the General Theory with Lerner before
the latter wrote his review of the book (JMK XIV: 148, esp. n. 1). Robinson (1978:
xiv-xv) and Kahn (1984:182–3) have also described a weekend meeting in August
1933 with Meade (who, together with them, had been a member of the 1931
‘Cambridge Circus’) and Lerner which was devoted to clarifying various aspects of
Keynes’ thinking up to that point. Meade was also in close contact in other contexts
with his Oxford colleague, Roy Harrod, as well as with Kahn (see Young 1989:52–
6, 70–3, et passim), and it is reasonable to assume that in this way he continued to
receive some information about the development of the General Theory. There was
also some minimal contact between Hicks and Keynes (Hicks 1973:7–8).
30 For different approaches to The Politics of Interpretation, see the 1983 collection
of papers and related discussions under that title, edited by WJT.Mitchell.
31 In the opening footnote of this chapter, Weintraub notes that ‘with some minor
changes’, it is reproduced from his 1957 Economic Journal article on ‘The Micro-
Foundations of Aggregate Demand and Supply’. Though it is most unlikely that
this was one of the ‘minor changes’ that he had in mind, it does seem to me that the
connection he makes in this chapter between his aggregate supply curve and that
of the General Theory is somewhat stronger than the connection he made in his
1957 article. Specifically, the opening paragraph of that article consists of the
following declaration:
This article attempts to construct an aggregate supply function without reference
to what Keynes ‘really’ meant. Unlike recent contributions to the Economic Journal
(and here a footnote provides references to the same articles and notes that were
later listed in the footnote at the beginning of Chapter 2 of his 1958 book, just
referred to in the text), it foregoes the attempt to link the concept, in detail and
with supporting references, to Keynes. While discretion replaces valour, it remains
beyond dispute that the very discussion must acknowledge the General Theory
as the source and inspiration.
REFERENCES
Reprinted works are cited in the text by year of original publication; the page references
to such works in the text are, however, to the pages of the reprint in question.
Barr, Nicholas (1988) ‘The Phillips machine’, LSE Quarterly 2, 305–37.
Brown, Arthur (1988) ‘A worm’s eye view of the Keynesian revolution’, in John Hillard
(ed.) J.M.Keynes in Retrospect: The Legacy of the Keynesian Revolution (Edward
Elgar, Hants, England) 18–44.
On different interpretations of the General Theory 89
Brunner, Karl and Allan H.Meltzer (1983) ‘Strategies and tactics for monetary control’,
Carnegie-Rochester Conference Series on Public Policy 18, 59–103.
Bryce, Robert B. (1935) ‘An introduction to a monetary theory of employment’, in
Patinkin and Leith (1977:128–45). Also reproduced in JMK XXIX, 132–50.
Champernowne, D.G. (1936) ‘Unemployment, basic and monetary: The classical analysis
and the Keynesian’, Review of Economic Studies 3, 201–16. As reprinted in
Lekachman (1964:153–73).
Chick, Victoria (1983) Macroeconomics after Keynes: A Reconsideration of the General
Theory (Philip Allan, Oxford).
Coddington. Alan (1976) ‘Keynesian economics: The search for first principles’, Journal
of Economic Literature 14, 1258–73.
Coddington, Alan (1979) ‘Hicks’s contribution to Keynesian economies’, Journal of
Economic Literature 17, 970–88.
Coddington, Alan (1983) Keynesian Economics: The Search for First Principles (George
Allen & Unwin, London).
Davidson, Paul (1972) Money and the Real World (Macmillan, London).
Davidson, Paul (1981) ‘Post Keynesian economics: solving the crisis in economic theory’,
in Daniel Bell and Irving Kristol (eds) The Crisis in Economic Theory (Basic Books,
New York) 151–73.
Davidson, Paul (1989) ‘Keynes and money’, in Roger Hill (ed.) Keynes, Money and
Monetarism (Macmillan, London) 2–26.
Dimsdale, N.H. (1987) ‘Keynes on British budgetary policy 1914–1946’, in Michael
J.Boskin, John S.Flemming, and Stefano Gorini (eds) Private Saving and Public Debt
(Basil Blackwell, Oxford) 208–33.
Eichner, Alfred S. (ed.) (1979) A Guide to Post-Keynesian Economics (M.E.Sharpe,
White Plains, NY).
Eichner, Alfred S. and J.A.Kregel (1975) ‘An essay on post-Keynesian theory: a new
paradigm in economies’, Journal of Economic Literature 13, 1293–314.
Fish, Stanley (1980) Is There a Text in this Class? The Authority of Interpretive
Communities (Harvard University Press, Cambridge, MA).
Fisher, Irving (1892) ‘Mathematical investigations in the theory of value and prices’,
Transactions of the Connecticut Academy 9, July. Reprinted with a new preface
(Yale University Press, New Haven, CT, 1925). Reprinted (Kelley, New York, 1901).
Fisher, Irving (1930) The Theory of Interest (Macmillan, New York). Reprinted (Kelley
and Millman, New York, 1954).
Friedman, Milton (1953) Essays in Positive Economics (University of Chicago Press,
Chicago, IL).
Grossman, Herschel I. (1972) ‘Was Keynes a “Keynesian”? A review article’, Journal of
Economic Literature 10, 26–30.
Hansen, Alvin H. (1936) ‘Mr. Keynes on underemployment equilibrium’, Journal of
Political Economy, 44, 667–86.
Hansen, Alvin H. (1953) A Guide to Keynes (McGraw-Hill, New York).
Harcourt, G.C. (1980) ‘A post-Keynesian development of the “Keynesian” model’, in
Edward J.Nell (ed.) Growth, Profits, and Property: Essays in the Revival of Political
Economy (Cambridge University Press, Cambridge) 151–64.
Harcourt, G.C. (1987) ‘Post-Keynesian economies’, in John Eatwell, Murray Milgate,
and Peter Newman (eds) The New Palgrave: A Dictionary of Economics, vol. 3
(Macmillan, London) 924–8.
Harrod, R.F. (1937) ‘Mr. Keynes and traditional theory’, Econometrica 5, 74–86. As
reprinted in Lekachman (1964:124–52).
Harrod, R.F. (1951) The Life of John Maynard Keynes (Macmillan, London).
Hart, A.G. (1947) ‘Keynes’ analysis of expectations and uncertainty’, in S.E.Harris
(ed.) The New Economics: Keynes’ Influence on Theory and Economic Policy (Knopf,
New York) 415–24.
90 Don Patinkin
Hicks, J.R. (1936) ‘Mr. Keynes’ theory of employment’, Economic Journal 46, 238–
53. As reprinted in Hicks (1982:83–99).
Hicks, J.R. (1937) ‘Mr. Keynes and the “Classics”: a suggested interpretation’,
Econometrica 5, 147–59. As reprinted in Hicks (1982:100–15).
Hicks, J.R. (1939) Value and Capital (Clarendon Press, Oxford).
Hicks, J.R. (1973) ‘Recollections and documents’, Economica 40, 2–11.
Hicks, J.R. (1974) The Crisis in Keynesian Economics (Basil Blackwell, Oxford). Hicks,
J.R. (1976) ‘Time in economies’, in A.M.Tang et al (eds) Evolution, Welfare and
Time in Economics: Essays in Honor of Nicholas Georgescu-Roegen. As reprinted
in Hicks (1982:282–300).
Hicks, J.R. (1979) ‘On Coddington’s interpretation: a reply’, Journal of Economic
Literature 17, 989–95.
Hicks, J.R. (1981) ‘IS-LM: an explanation’, Journal of Post Keynesian Economics 3,
139–54. As reprinted in Hicks (1982:318–31).
Hicks, J.R. (1982) Money, Interest and Wages: Collected Essays on Economic Theory,
vol. II (Basil Blackwell, Oxford).
Hirsch, E.D.Jr (1967) Validity in Interpretation (Yale University Press, New Haven,
CT).
Hirsch, E.D.Jr (1976) The Aims of Interpretation (University of Chicago Press, Chicago,
IL).
Jackman, Richard (1974) Keynes and Leijonhufvud, Oxford Economic Papers 26,
259–72.
Johnson, Harry G. (1961) The General Theory after Twenty-Five Years, American
Economic Association Papers and Proceedings 51, 1–17.
Kahn, Richard F. (1984) The Making of Keynes’ General Theory (Cambridge University
Press, Cambridge).
Keynes, John Maynard (1923) A Tract on Monetary Reform. As reprinted in Keynes,
Collected Writings, Vol. IV.
Keynes, John Maynard (1925) The Economic Consequences of Mr. Churchill. As
reprinted in Keynes, Collected Writings, Vol. IX, 207–30.
Keynes, John Maynard (1930) A Treatise on Money, Vol. I: The Pure Theory of Money.
As reprinted in Keynes, Collected Writings, Vol. V.
Keynes, John Maynard (1930) A Treatise on Money, Vol. II: The Applied Theory of
Money. As reprinted in Keynes, Collected Writings, Vol. VI.
Keynes, John Maynard (1933) The Means to Prosperity. As reprinted in Keynes,
Collected Writings, Vol. IX, 335–66.
Keynes, John Maynard (1936) The General Theory of Employment Interest and Money.
As reprinted in Keynes, Collected Writings, Vol. VII.
Keynes, John Maynard (1937a) ‘Alternative theories of the rate of interest’, Economic
Journal 47, 241–52. As reprinted in Keynes, Collected Writings, Vol. XIV, 201–
215.
Keynes, John Maynard (1937b) ‘The general theory of employment’, Quarterly Journal
of Economics 51, 209–23. As reprinted in Keynes, Collected Writings, Vol. XIV,
109–23.
Keynes, John Maynard (1937c) How to Avoid a Slump. As reprinted in Keynes, Collected
Writings, Vol. XXI, 384–95.
Keynes, John Maynard (1938) ‘Comment on D.H.Robertson, “Mr. Keynes and
‘Finance’”’, Economic Journal 48, 318–22. As reprinted in Keynes, Collected
Writings, Vol. XIV, 229–233.
Keynes, John Maynard (1940) How to Pay for the War. As reprinted in Keynes, Collected
Writings, Vol. IX, 367–439.
Keynes, John Maynard, The General Theory and After: Part I, Preparation, edited by
Donald Moggridge, Vol. XIII of Keynes, Collected Writings.
On different interpretations of the General Theory 91
Keynes, John Maynard, The General Theory and After: Part II, Defense and
Development, edited by Donald Moggridge, Vol. XIV of Keynes, Collected Writings.
Keynes, John Maynard, Activities 1931–1939: World Crises and Policies in Britain
and America, edited by Donald Moggridge, Vol. XXI of Keynes, Collected Writings.
Keynes, John Maynard, Activities 1940–1946: Shaping the Post-war World:
Employment and Commodities, edited by Donald Moggridge, Vol. XXVII of Keynes,
Collected Writings.
Keynes, John Maynard, The General Theory and After: A Supplement, edited by Donald
Moggridge, Vol. XXIX of Keynes, Collected Writings.
Keynes, John Maynard, Collected Writings, Vols I-VI (1971), Vols VII-VIII (1973),
Vols IX–X (1972), Vols XI–XII (1983), Vols XIII–XIV (1973), Vols XV–XVI (1971),
Vols XVII–XVIII (1978), Vols XIX–XX (1981), Vol. XXI (1982), Vol. XXII (1978),
Vols XXIII–XXIV (1979), Vols XXV–XXVII (1980), Vol. XXVIII (1982), Vol. XXIX
(1979), Vol. XXX (bibliography and index, 1989) (Macmillan, for the Royal
Economic Society, London).
Keynes, John Maynard and Hubert Henderson (1929) Can Lloyd George do it? An
Examination of the Liberal Pledge. As reprinted in Keynes, Collected Writings, Vol.
IX, 86–125.
Klein, Lawrence (1947) The Keynesian Revolution (Macmillan, New York).
Knight, Frank H. (1921) Risk, Uncertainty and Profit (Houghton Mifflin, New York).
Knight, Frank H. (1923) ‘The ethics of competition’, Quarterly Journal of Economics
37, 579–624. As reprinted in Knight, The Ethics of Competition and Other Essays
(Harper & Bros., New York, 1935) 41–75.
Kuhn, Thomas S. (1970) The Structure of Scientific Revolutions, 2nd enlarged edn
(University of Chicago Press, Chicago, IL).
Lange, Oskar (1938) ‘The rate of interest and the optimum propensity to consume’,
Economica 5, 12–32. As reprinted in Readings in Business Cycle Theory, Selected
by a Committee of the American Economic Association (Blakiston, for the American
Economic Association, Philadelphia, PA, 1944) 169–92.
Leijonhufvud, Axel (1968) On Keynesian Economics and the Economics of Keynes: A
Study in Monetary Theory (Oxford University Press, New York).
Leijonhufvud, Axel (1978) Foreword to the Japanese edition of Leijonhufvud (1968)
(Toyo Keizai Shinposha, Tokyo).
Lekachman, Robert (ed.) (1964) Keynes’ General Theory: Reports of Three Decades
(St Martin’s Press, New York).
Leontief, Wassily (1936) ‘The fundamental assumption of Mr. Keynes’ monetary theory
of unemployment’, Quarterly Journal of Economics 51, 192–7.
Lerner, Abba P. (1936) ‘The general theory’, International Labour Review 34, 435–54.
As reprinted in Lekachman (1964:203–22).
Matthews, R.C.O. (1984) ‘Animal spirits’, Proceedings of the British Academy 70,
209–29.
Meade, J.E. (1937) ‘A simplified model of Mr. Keynes’ system’, Review of Economic
Studies 4, 98–107.
Meade, J.E. (1990) The Cabinet Office Diary: 1944–46, edited by Susan Howson and
Donald Moggridge, Vol. IV of The Collected Papers of James Meade (Unwin Hyman,
London).
Meltzer, Allan H. (1981) ‘Keynes’s General Theory: a different perspective’, Journal of
Economic Literature 29, 34–64.
Meltzer, Allan H. (1983) ‘Monetary reform in an uncertain environment’, Cato Journal
3, 93–112.
Meltzer, Allan H. (1987) ‘Limits of short-run stabilization policy’, Economic Inquiry
25, 1–13.
Meltzer, Allan H. (1988) Keynes’s Monetary Theory: A Different Interpretation
(Cambridge University Press, Cambridge).
92 Don Patinkin
Merton, Robert K. (1957) ‘Priorities in scientific discovery’, American Sociological
Review 22, 635–59. As reprinted in Merton, The Sociology of Science: Theoretical
and Empirical Investigations, edited by Norman W.Storer (University of Chicago
Press, Chicago, IL, 1973) 286–324.
Milgate, Murray (1982) Capital and Employment: A Study of Keynes’s Economics
(Academic Press, London).
Minsky, Hyman P. (1975) John Maynard Keynes (Columbia University Press, New
York).
Mitchell, W.J.T. (ed.) (1983) The Politics of Interpretation (University of Chicago Press,
Chicago, IL).
Moggridge, D.E. (1986) ‘Keynes and the international monetary system 1909–46’, in
Jon S.Cohen and G.C.Harcourt (eds) International Monetary Problems and Supply-
side Economics: Essays in Honour of Lorie Tarshis (Macmillan, London) 56–83.
Modigliani, Franco (1944) ‘Liquidity preference and the theory of interest and money’,
Econometrica 12, 45–88. As reprinted in Readings in Monetary Theory, Selected by
a Committee of the American Economic Association (Blakiston, for the American
Economic Association, Philadelphia, PA, 1951) 186–240.
Pasinetti, Luigi L. (1974) Growth and Income Distribution: Essays in Economic Theory
(Cambridge University Press, Cambridge).
Patinkin, Don (1956) Money, Interest and Prices: An Integration of Monetary and
Value Theory (Row, Peterson, Evanston, IL).
Patinkin, Don (1965) Money, Interest and Prices: An Integration of Monetary and
Value Theory, 2nd edn (Harper and Row, New York).
Patinkin, Don (1976) Keynes’ Monetary Thought: A Study of its Development (Duke
University Press, Durham, NC).
Patinkin, Don (1982) Anticipations of the General Theory? And Other Essays on
Keynes (University of Chicago Press, Chicago, IL).
Patinkin, Don (1983) ‘New perspectives or old pitfalls? Some comments on Allan
Meltzer’s interpretation of the General Theory’, Journal of Economic Literature 21,
47–51.
Patinkin, Don (1989) Money, Interest, and Prices: An Integration of Monetary and
Value Theory, 2nd abridged edn with a new introduction (MIT Press, Cambridge,
MA).
Patinkin, Don (1990) ‘In defense of IS-LM’, Banca Nazionale del Lavoro: Quarterly
Review 172, 119–34.
Patinkin, Don, and J.Clark Leith (eds) (1977) Keynes, Cambridge and the General
Theory: The Process of Criticism and Discussion Connected with the Development
of the General Theory (Macmillan, London).
Phelps Brown, E.H. (1937) ‘Report of the Oxford meeting, September 25–29, 1936’,
Econometrica 5, 361–83.
Phillips, A.W. (1950) ‘Mechanical models in economic dynamics’, Economica NS 17,
283–305.
Reddaway, W.B. (1936) ‘The general theory of employment, interest and money’,
Economic Record 12, June. As reprinted in Lekachman (1964, 99–107).
Robertson, D.H. (1936) ‘Some notes on Mr. Keynes’ general theory of employment’,
Quarterly Journal of Economics 51, 168–91.
Robinson, Austin (1977) ‘Keynes and his Cambridge colleagues’, in Patinkin and Leith
(1977:25–38).
Robinson, Joan (1937) Introduction to the Theory of Employment (Macmillan,
London).
Robinson, Joan (1962a) Essays in the Theory of Economic Growth (Macmillan,
London).
Robinson, Joan (1962b) Review of ‘Money, trade and economic growth’ by Harry
On different interpretations of the General Theory 93
G.Johnson, Economic Journal 72, 690–2. As reprinted in part in Robinson, Collected
Economic Papers, Vol. III (Basil Blackwell, Oxford, 1975) 100–2.
Robinson, Joan (1962c) Economic Philosophy (Aldine, Chicago, IL).
Robinson, Joan (1970) ‘Quantity theories old and new’, Journal of Money, Credit and
Banking 2, 504–12.
Robinson, Joan (1973) ‘What has become of the Keynesian revolution?’, in Joan
Robinson (ed.) After Keynes (Basil Blackwell, Oxford).
Robinson, Joan (1977) ‘What are the questions?’, Journal of Economic Literature 15,
1318–39. As reprinted in Robinson, Collected Economic Papers, Vol. V (Basil
Blackwell, Oxford, 1980) 1–31.
Robinson, Joan (1978) Contributions to Modern Economics (Basil Blackwell, Oxford).
Robinson, Joan (1979) Foreword to Eichner (1979, xi-xxi).
Rymes, Thomas K. (ed.) (1989) Keynes’s Lectures, 1932–35: Notes of a Representative
Student (Macmillan, London).
Samuelson, Paul A. (1946) ‘Lord Keynes and the General Theory’, Econometrica 14,
187–200. As reprinted in Lekachman (1964:315–31).
Samuelson, Paul A. (1947) Foundations of Economic Analysis (Harvard University
Press, Cambridge, MA).
Schlesinger, James R. (1956) ‘After twenty years: The General Theory’, Quarterly Journal
of Economics 70, 581–602.
Shackle, G.L.S. (1938) Expectations, Investment and Income (Oxford University Press,
Oxford).
Shackle, G.L.S. (1949) Expectation in Economics (Cambridge University Press,
Cambridge).
Shackle, G.L.S. (1953) ‘A chart of economic theory’, Metroeconomica 5, 1–10. As
reprinted in Shackle, Uncertainty in Economics: And Other Reflections (Cambridge
University Press, Cambridge, 1955) 217–26.
Shackle, G.L.S. (1961) ‘Recent theories concerning the nature and role of interest’,
Economic Journal 71, 209–254. As reprinted in Shackle, The Nature of Economic
Thought: Selected Papers 1955–1964 (Cambridge University Press, Cambridge,
1966)225–81.
Shackle, G.L.S. (1967) The Years of High Theory: Invention and Tradition in Economic
Thought, 1926–1939 (Cambridge University Press, Cambridge).
Shackle, G.L.S. (1973) ‘Keynes and today’s establishment in economic theory: a view’,
Journal of Economic Literature 11, 516–19.
Shackle, G.L.S. (1982) ‘Sir John Hicks’ “IS-LM: An explanation”: a comment’, Journal
of Post Keynesian Economics 4, 435–8.
Simons, Henry (1936) ‘Rules versus authorities in monetary policy’, Journal of Political
Economy 44, 1–30. As reprinted in Simons, Economic Policy for a Free Society
(University of Chicago Press, Chicago, IL, 1948) 160–83.
Skinner, Quentin (1969) ‘Meaning and understanding in the history of ideas’, History
and Theory 8, 3–53. As reprinted in Tully (1988:29–78).
Skinner, Quentin (1976) ‘Motives, intentions and the interpretation of texts’, in D.
Newton de Molina (ed.) On Literary Intention (Edinburgh University Press,
Edinburgh) 210–21. As reprinted in Tully (1988:68–78).
Taussig, F.W. (1936) ‘Employment and the national dividend’, Quarterly Journal of
Economics 51, 198–203.
Tully, James (ed.) (1988) Meaning and Context: Quentin Skinner and his Critics (Polity
Press, Cambridge).
Viner, Jacob (1936) ‘Mr. Keynes on the causes of unemployment’, Quarterly Journal
of Economics 51, 147–67. As reprinted in Lekachman (1964:235–53).
94 Don Patinkin
Weintraub, Sidney (1957) ‘The micro-foundations of aggregate demand and supply’,
Economic Journal 67, 455–70.
Weintraub, Sidney (1958) An Approach to the Theory of Income Distribution (Chilton,
Philadelphia, PA).
Weintraub, Sidney (1959) A General Theory of the Price Level, Output, Income
Distribution, and Economic Growth (Chilton, Philadelphia, PA).
Weintraub, Sidney (1960) ‘The Keynesian theory of inflation: the two faces of Janus?’,
International Economic Review 1, 143–55.
Weintraub, Sidney (1961) Classical Keynesianism, Monetary Theory, and the Price
Level (Chilton, Philadelphia, PA).
Weintraub, Sidney (1976) ‘Revision and recantation in Hicksian economics: a review
article’, Journal of Economic Issues 10, 618–27. As reprinted in Weintraub et al.
(1978, 113–23).
Weintraub, S. et al. (1978) Keynes, Keynesians and Monetarists (University of
Pennsylvania Press, Philadelphia, PA).
Young, Warren (1987) Interpreting Mr. Keynes: The IS-LM Enigma (Polity Press,
Cambridge).
Young, Warren (1989) Harrod and his Trade Cycle Group: The Origins and
Development of the Growth Research Programme (Macmillan, London).
Government publications
Great Britain (1944) Employment Policy, Parliamentary papers, Cmd 6527 (HMSO,
London).
Postscript to note 8
In reply to my query about the conjectures of Howson and Brown, Shackle
wrote me: ‘I can say that my use of the term “hydraulic” was not influenced
by the Phillips machine, unless quite unconsciously. I cannot say whether
Coddington may have had this term suggested to him by The Years of High
Theory.’
4 Keynes’s General Theory
Interpreting the interpretations
Bill Gerrard
Economic Journal (1991) 101, March, pp. 276–87
Keynes’s General Theory has given rise to a variety of Keynesian research
programmes. The development of these different Keynesian research
programmes is well documented (see e.g. Coddington 1976; Gerrard 1988;
Hamouda and Harcourt 1988 for surveys of Keynesian and post-Keynesian
economics). However, less attention has been paid to explaining a striking
feature of this Keynesian diversity, namely, the stress placed on discovering
the real meaning of Keynes’s General Theory. The legitimacy of any
particular Keynesian research programme has been judged with regard to the
authenticity of its implied interpretation of Keynes. Inevitably this concern
for authenticity has generated much controversy, enveloping Keynesian
economics in a ‘doctrinal fog’ (Blaug 1980:221). This chapter attempts to
pierce that Keynesian doctrinal fog. The central thesis is that the causes of
the controversy surrounding Keynes’s General Theory lie, in part, in the
different presuppositions made about the nature of interpretation. It is argued
that much light can be shed on the Keynesian debate by drawing on the
study of hermeneutics.
The structure of the chapter is as follows. The first section discusses the
atomistic view of interpretation which is implicitly presupposed by most
contributors to the Keynesian debate. Two variants of the atomistic view are
considered: the objectivist/essentialist approach and the relativist approach.
The second section provides an alternative presupposition, the organicist
view of interpretation, as exemplified by Ricoeur’s dialectical approach in
hermeneutics. The chapter concludes with a re-examination of the Keynesian
debate in the light of the organicist view of interpretation.
THE ATOMISTIC VIEW OF INTERPRETATION
A principal aim of Keynesian economics has been to give a definitive answer
to the question ‘What does Keynes’s General Theory really mean?’ Much of
the resulting controversy arises from the nature of the question itself. In
asking the question an atomistic view of interpretation is presupposed. The
author, the text and the reader are treated as individual atomistic entities
which are interrelated in a purely external manner: the author produces the
96 Bill Gerrard
text which the reader interprets. There are two variants of the atomistic view
of interpretation: the objectivist/essentialist approach and the relativist
approach.
The objectivist/essentialist approach
From the perspective of the objectivist/essentialist approach, the aim of
interpretation is the rational reconstruction of the text in order to recover the
author’s original meaning. Interpretation is seen to be problematic because
the author’s meaning is hidden. The latency of the original meaning creates
confusion and generates the possibility of multiple interpretations. The task
of the interpreter is to resolve this confusion by discovering the ‘true’
meaning of a text. This presupposes that the true meaning is knowable.
Within the objectivist/essentialist approach, Keynes’s General Theory is
viewed as containing a single essential meaning which is hidden as the result
of the confusion created either by Keynes himself or by the economics
profession in its reading of Keynes. The belief that Keynes’s General Theory
contains a single essential meaning is shared by most interpreters of Keynes.
Leijonhufvud (1968) claims to have found the ‘economics of Keynes’ as
opposed to ‘Keynesian economies’, while Shackle (1967: ch. 12) seeks
Keynes’s ‘ultimate meaning’. Fender (1981:1–2) sets out to find the ‘exact
nature of the theoretical contribution of Keynes’; similarly, Chick (1983: v)
attempts to remedy the fact that the ‘macroeconomics that has been
developed after Keynes, though claiming inspirations from the General
Theory, in my view has not, with some outstanding exceptions, been
macroeconomics after the manner of Keynes—with the method and
perspective and insight of Keynes’.
From the objectivist/essentialist perspective, it is necessary to explain why
the essential meaning of the General Theory is hidden. There are three broad
types of explanations.
The confusion is author-generated
A number of writers have suggested that Keynes himself is the cause of the
confusion. There are a number of variants of this ‘author-generated
confusion’ thesis.
(i) Technical incompetence
It is often argued that Keynes had limited analytical abilities. For example,
Hahn (1982: x, xi) writes that ‘I consider that Keynes had no real grasp of
formal economic theorising (and also disliked it), and that he consequently
left many gaping holes in his theory.’ This follows a famous remark by
Shove that ‘Maynard had never spent the twenty minutes necessary to
understand the theory of value’ (quoted in Robinson 1964:79).
Keynes’s General Theory: Interpretations 97
(ii) The ‘vision’ thesis
Confusion arises as the inevitable consequences of the difficulties which
Keynes faced in trying to formulate his underlying vision in a precise
analytical manner. It is a line of argument originating with Schumpeter
(1946:501) who distinguishes between Keynes’s vision, that is, his ‘view
about the basic features of society, about what is and what is not
important’ and Keynes’s technique, that is, the ‘apparatus by which he
conceptualises his vision and which turns the latter into concrete
propositions or theories’. According to Schumpeter the General Theory is
the final result of a long struggle by Keynes to make his vision analytically
operative. Leijonhufvud (1968:10, 11) adopts Schumpeter’s distinction,
arguing that Keynes was not entirely successful in translating his vision
into a logically watertight model.
(iii) Stylistic difficulties
Keynes is often accused of a lack of clarity. This is suggested by O’Donnell
(1989a: 6) as a reason for the difficulties in the interpretation of Keynes.
O’Donnell approvingly quotes Wittgenstein’s maxim that if ‘anything can be
said, it can be said clearly’. Leijonhufvud (1968:10, 11) goes as far as to say
that the General Theory is ‘a badly written book’ and that the need for
‘repairs’ has led to confusion because different writers have corrected
Keynes’s model in inappropriate ways. It is also argued that Keynes’s style is
too loose and vague. This echoes Whitehead’s criticism of Keynes’s
dissertation on probability as using the style of literature, not the style of
logic and philosophy.
(iv) Inconsistencies
Some writers have suggested that Keynes did not have a coherent and
consistent vision. It is an argument used by Leijonhufvud but also by
Robinson (1973:3) in her well-known comment that ‘there were moments
when we had some trouble in getting Maynard to see what the point of his
revolution really was’.
(v) The ‘grand mistake’ thesis
An extreme example of the ‘author-generated confusion’ thesis is the neo-
Ricardian argument that Keynes made a mistake in retaining the
neoclassical concept of the marginal efficiency of capital (Milgate 1982).
This created the possibility that, if the rate of interest is sufficiently low,
there will be sufficient investment to maintain full employment. According to
Milgate this undermined the principle of effective demand, the essence of the
General Theory. Furthermore it forced Keynes to develop explanations of
98 Bill Gerrard
interest rate maladjustment; hence Keynes’s misguided emphasis on
expectations and liquidity preference.
The confusion is reader-generated
An alternative explanation of the confusion surrounding the General Theory
is to focus on the actions of the audience. Again there are a number of
variants of this explanation.
(i) Inappropriate framing
Readers have interpreted Keynes relative to an inappropriate frame of
reference. This results in the development of a variety of ‘subjective’
interpretations based on personal beliefs and ideological and normative
biases. In the process the ‘objective’ meaning of Keynes becomes lost. This
line of argument is epitomised by Leijonhufvud’s distinction between
‘Keynesian economies’ and the ‘economics of Keynes’. It is an argument
repeated by Fitzgibbons (1988:1–5) when he points towards the problem of
‘systematically biased interpretation’.
(ii) Selective reading
A closely related variant to the inappropriate framing argument is the
problem created by readers considering only parts of the text. Thus
O’Donnell (1989a: 4) sees the main reason for the multiple interpretations of
Keynes as the ‘tendency to base interpretations on selected parts, rather than
the whole of his relevant writings’.
(iii) Reliance on secondary sources
The tendency towards multiple interpretations of Keynes has been
exacerbated by the tendency to read about Keynes rather than to read Keynes
himself.
The confusion is generated by differences in the composition of the stock of
relevant textual evidence
Different interpretations may arise because of differences between
interpreters with regard to the definition of the text to be analysed. There are
two variants of this explanation.
(i) Which text?
With an author as productive as Keynes, an inevitable problem is whether to
interpret the target text in isolation or in the context of the author’s other
Keynes’s General Theory: Interpretations 99
related writings. This can lead to multiple interpretations if readers find
inconsistencies between different texts. The definition of ‘other relevant
writings’ may vary between interpreters, particularly with regard to the
relative weights to be attached to earlier and later writings as well as to
formal writings and more informal sources such as speeches, unpublished
papers and private correspondence. Thus, for example, the recent emergence
of the ‘new’ Keynesian fundamentalism associated with Carabelli (1988),
Fitzgibbons (1988) and O’Donnell (1989a) amongst others, represents a shift
of weight in favour of Keynes’s early philosophical papers, a source largely
ignored by previous interpreters.
(ii) Changes in availability
Not only can there be ‘subjective’ differences about the definition of the text,
there may also be ‘objective’ differences over time as the stock of documents
available for interpretation changes. Such changes have been particularly
important in the interpretation of Keynes’s General Theory. The publication
of the Collected Writings, especially volumes XIII and XIV, shed
considerable light on the development of Keynes’s thought immediately
before and after the publication of the General Theory. The stock of
documents expanded subsequently with the publication of volume XXIX as
the result of the discovery of a laundry basket of previously unknown papers
by Keynes. These included early drafts of the General Theory focusing on
the concept of a monetary production economy.
These various sources of confusion create the possibility of multiple
interpretations. Convergence towards the correct interpretation is usually
presumed to be ensured by the use of consistency with the textual evidence
as the criterion of choice between competing interpretations (for example,
Leijonhufvud 1968:8). Interpretation is viewed, therefore, as a scientific
problem. The scientific nature of interpretation has been highlighted by
Stigler (1965) with particular reference to the problem of multiple
interpretations of Ricardo. Stigler argues that hand-picked quotations are
insufficient to validate any particular interpretation. He proposes instead
the use of two different principles of interpretation: the principle of
scientific exegesis and the principle of personal exegesis. Scientific exegesis
is interpretation which aims to maximise the value of a text to the science.
In this case the text of an interpretation is its consistency with the main
analytical conclusions of the author. This type of interpretation is
concerned with the ‘strong’ form of the text, that is, an amended form of
the text which removes ‘blemishes’ such as logical errors and tautologies.
Personal exegesis, on the other hand, aims to discover what the author
really believed and thus the test of an interpretation becomes consistency
with the author’s style, that is, what the author actually wrote. According
to Stigler this latter form of interpretation is of no direct relevance to
scientific progress.
100 Bill Gerrard
Stigler’s separation of scientific and personal exegesis has been the subject
of criticism by Aksoy (1989) and Hollander (1989). Both question whether
scientific exegesis can really be considered as interpretation when it denies any
significance to what the author really believed. Hollander criticises scientific
exegesis for leading to the distortion of an author’s writings since it justifies
disregarding those parts of an author’s writings deemed by the interpreter to be
‘blemishes’. Hollander describes this as a ‘lazy man’s procedure’ and argues
for the need to explain the ‘residuals’, that is, the differences between the
‘strong’ form of the text and what the author actually wrote.
The controversy surrounding Stigler’s scientific approach to interpretation
is suggestive of an important, but often overlooked, characteristic of
interpretation. ‘Consistency with the textual evidence’ is inadequate as an
objective criterion for assessing competing interpretations. The definition of
both ‘consistency’ and ‘textual evidence’ is open to debate. The textual
evidence may be defined as the single target text only or it may include other
relevant texts. Likewise ‘consistency’ is open to multiple interpretation.
Stigler suggests two alternative definitions: consistency with the author’s
beliefs (i.e. personal exegesis). O’Donnell (1989b) has proposed a three stage
consistency test for any interpretation of Keynes. According to O’Donnell,
interpretations should be: (i) internally consistent; (ii) consistent with
quotations taken in context; and (iii) consistent with all of an author’s
writings. This definition of consistency appears to be unnecessarily restrictive
in at least two ways. First, it seems to impose on an author’s writings a
degree of integration and continuity over time which may be unwarranted.
Second, it seems to exclude evidence drawn from sources other than the
author’s own writings.
Thus the choice between competing interpretations cannot be purely
objective. Such choices must always be made on the basis of certain
presuppositions. Any particular interpretation has a conventional foundation,
a set of presuppositions which are treated as beyond doubt. This
conventional foundation includes presuppositions about the aims of
interpretation, the consistency criterion and the relevant evidence. Different
writers may base their interpretations on different conventional foundations.
Furthermore the very presupposition that the author’s meaning is hidden
provides the means by which any particular interpretation can be rendered
consistent with the text. Any apparent inconsistency between an
interpretation and the text can be explained away as the result of author-
generated confusion. The objectivist/essentialist approach has a built-in
immunising stratagem with which to protect the validity of any particular
interpretation. The parallel with recent developments in the philosophy of
science is clear. The Duhem-Quine thesis on the underdeterminacy of
empirical testing has led to the recognition that science consists of theoretical
structures with conventional foundations as exemplified by Kuhn’s theory of
paradigms and Lakatos’s notion of scientific research programmes with hard
cores.
Keynes’s General Theory: Interpretations 101
The relativist approach
The contradictions within the objectivist/essentialist approach to
interpretation have led some to adopt the relativist approach. Rather than
viewing interpretation as the discovery of the author’s meaning hidden
within the text, the relativist approach treats interpretation as the product of
the reader imposed on the text. Thus the analysis of interpretation moves
from being text-centred to being reader-centred. There is no single essential
meaning within a text, only a variety of reader-determined meanings.
Interpretations are always relative to the frame of reference within which the
reader is operating. This implies that there is no objective standard for a
rational evaluation of alternative interpretations. The choice of
interpretation is made by the individual reader on the basis of a subjectively
determined frame of reference. Any interpretation can only ever be treated as
consistent with its own underlying frame of reference. In this sense, one
interpretation is as good as another.
From the relativist perspective, to ask the question ‘What does Keynes’s
General Theory really mean?’ is really to ask ‘What does Keynes’s
General Theory mean to me, given my frame of reference?’ Readers
choose that interpretation which makes sense within their own world-
view. For example, since the neo-Keynesians adopt the choice-theoretic
and market-theoretic perspective of mainstream economic theory, it
follows that they interpret Keynes’s General Theory to be dealing with
the effects of imperfections such as non-atomistic market structures or
imperfect information which prevent the price mechanism from operating
effectively.
The relativist approach, however, also suffers from self-contradictions. To
adopt the relativist approach is to accept that the interpretive process is
indeterminate. Relativism implies that ‘anything goes’; consistency with the
text is always relative to a subjectively determined frame of reference. This
brings the validity of the process of interpretation itself into question. The
relativist approach denies that interpretation can be the pursuit of
understanding beyond that which is relative to the individual’s own frame of
reference. But this runs counter to the explicitly stated aims of those who
engage in interpretation. Interpretations are advanced on the basis of
intellectual and cognitive properties which are deemed to transcend
subjectivist concerns. In order to justify the public presentation of an
interpretation, interpreters adopt an objectivist/essentialist approach when
advocating their own particular interpretations. Relativism cannot be
followed consistently since it would deny the intellectual properties which
interpreters always claim for their own interpretations. Relativism is a
position that can be entertained but never occupied. The relativist approach,
through the force of its own logic, becomes self-contradictory. Relativism
always reverts to a form of ‘back-door’ objectivism in which writers use the
relativist approach to criticise the interpretations of others but claim
102 Bill Gerrard
authenticity for their own interpretations. Practised in this form, relativism is
but a variant of the ‘reader-generated confusion’ thesis.
To summarise: much of the confusion surrounding the interpretation of
Keynes’s General Theory arises from the presupposed atomistic view of
interpretation. The atomistic view, in both its objectivist/essentialist and
relativist forms, is self-contradictory. These contradictions have been carried
over into the interpretation of Keynes. Thus the task of piercing the
Keynesian doctrinal fog requires, as a first step, moving beyond the atomistic
view of interpretation.
THE HERMENEUTIC APPROACH TO INTERPRETATION
An alternative view of interpretation is the organicist view which stresses the
importance of context. The writing and reading of texts are processes which
are socially and historically contingent. The author, the text and the reader
are not atomistic entities but form a dynamic whole in which the nature of
any part is defined by its interrelationship with the other parts. It is a general
vision of the process of interpretation which finds its expression primarily in
the study of hermeneutics.
Hermeneutics is the study of interpretation. It originated as the study of
the principles of biblical exegesis, deriving its name from Hermes, the
messenger of the Gods. Gradually hermeneutics extended its scope beyond
the confines of biblical exegesis, becoming the study of textual interpretation
in general. It was systematised in the nineteenth century in the writings of
Schleiermacher who set out the basic principles of hermeneutics.
Schleiermacher stressed the importance of context, arguing that the historical
conditions of the author must be taken into account in recovering the
original meaning of the text. In particular, it was crucial to understand the
original audience to whom the text was directed. Following Schleiermacher,
hermeneutics became closely associated with the German historicist school,
especially Dilthey, who came to regard the hermeneutic approach as a
general methodology for the human sciences. During the twentieth century
hermeneutics has undergone at least two marked changes in orientation. In
the writings of Heidegger and Gadamer, hermeneutics emerged as a general
philosophical position, becoming an integral part of the work of the
Frankfurt School. Later hermeneutics took a more linguistic turn, becoming
synonymous with the structuralist approach to the study of language and
symbols.
The ever-changing nature of hermeneutics means that there is no single
coherent set of hermeneutic principles. Instead there are different conflicting
traditions. Hirsch (1976) identifies three such traditions:
1 The biblical/intuitionist tradition, in which the concern is for the ‘spirit’
of the text. Meaning is seen as not fully expressible in words, implying
that the understanding of a text requires getting behind the text.
Keynes’s General Theory: Interpretations 103
Schumpeter’s notion of vision is suggestive of the intuitionist tradition.
Interpretation is more than logical deduction. It requires intuition in
order to achieve ‘communion’ with the spirit of the text.
2 The legal/positivist tradition, in which the concern is for the ‘letter of the
law’, that is, what is actually written. Meaning is seen as wholly
contained within the text and recoverable by means of logical deduction.
It is a tradition which has its origins in part in Jewish Talmudic
scholarship. The latter is explicitly invoked by Patinkin (1978) in
support of his contention that the students of Keynes’s thought should
concern themselves only with what Keynes actually said.
3 The metaphysical tradition, in which the original meaning of the text is
seen as unrecoverable, and, hence, unknowable. Given the seemingly
nihilistic implications of the metaphysical tradition for the interpretive
process, it is a tradition that has little parallel in the Keynesian debate.
Given these very different traditions in hermeneutics it is important from the
outset that hermeneutics should not be seen as a set of unquestionable
principles which can settle the debates over Keynes’s General Theory once
and for all. Hermeneutics can be a double-edged sword.
A key concept in hermeneutics is the hermeneutic circle which arises from
a general paradox within any organicist approach. To know the whole, one
needs to know the parts. But the parts can be known only in the context of
their interdependencies within the whole. Thus to know the whole requires
that the whole be pre-known. This circularity has a ‘narrow’ and ‘wider’
sense. In the narrow sense the text itself is treated as a whole while in the
wider sense the text is treated as part of the historical context.
Within hermeneutics, the hermeneutic circle has been subject to two very
different interpretations (Hirsch 1976). The ‘old’ or ‘romanticist’
hermeneutics considers the aim of interpretation to be the recovery of the
author’s original meaning. The hermeneutic circle is the recognition of the
need to understand the historical context of the text. In contrast, the ‘new’
hermeneutics considers the hermeneutic circle as the rejection of objectivity.
Interpretation is always relative to the historical context of the reader. Thus,
just as with the atomistic approach to interpretation, the hermeneutic
approach is caught in an objectivist-versus-relativist controversy.
An attempt to transcend the objectivist-relativist duality in hermeneutics is
to be found in the writings of Ricoeur (1976, 1981). Following Hirsch,
Ricoeur views interpretation as a process of guess and validation in which
interpreters marshall evidence to show a particular interpretation to be more
or less probably true. Interpretation involves the logic of qualitative
probability and uncertainty. However, unlike Hirsch, he does not view the
aim of interpretation as the recovery of the author’s original meaning which
he considers to be ‘a lost psychical event’. But the denial of this objective
standard does not entail that anything goes in interpretation. The text
contains potential meaning to be actualised by readers but it constrains the
104 Bill Gerrard
possibilities of that actualisation. For Ricoeur interpretation is the
appropriation of the reference power of a text, that is, the actualisation of a
text’s ability to disclose possible ways of looking at the world. Interpretation
involves what Gadamer termed the ‘fusion of horizons’. The world horizon
of the reader becomes fused with the world horizon of the author with the
text acting as the mediating link. The resultant interpretation is not the
function of the text or the reader alone. Rather it is the outcome of a
dialectical process.
From the writings of Ricoeur and others it is possible to develop a
hermeneutic framework for the analysis of interpretation using the following
definitions:
• understanding: the construction of a text’s meaning in its own terms
• explanation: the presentation of the understood meaning in terms
accessible to a particular audience
• interpretation: the process of dynamic interaction between understanding
and explanation
• meaning: the outcome of the process of interpretation
• judgement: the construction of a relation between the text and something
external to it
• criticism: the explanation and evaluation of the judgement
• application: the process of dynamic interaction between judgement and
criticism
• significance: the outcome of the process of application.
For Ricoeur these concepts are abstract poles within a concrete whole.
Interpretation is a dialectic of understanding and explanation in which the
understanding of a text affects its explanation and vice versa. Likewise
application is a dialectic of judgement and criticism. At a higher level there
is also a dialectical interaction between interpretation and application.
The dialectical approach of Ricoeur and other related organicist analyses
of interpretation are able to provide an alternative set of presuppositions
about the nature of interpretation to that associated with the atomistic view.
This alternative set of presuppositions can be formulated in terms of four
propositions.
Proposition 1 Interpretation is a multi-dimensional process with multiple
objectives
Interpretation involves understanding and explanation as well as judgement
and criticism. It aims not only to provide meaning to a text but also to
evaluate the significance of the text. The complexity of interpretation needs
to be recognised by writers in order to overcome the confusion. This
conclusion is drawn by Aksoy (1989:744) in his application of the
hermeneutic perspective to the debates surrounding the interpretation of
Keynes’s General Theory: Interpretations 105
Ricardo: ‘there is a great deal of confusion about the meaning and the
purpose of the act of interpretation. In most cases, interpreters are found to
be careless and inconsistent about the differences between interpretation,
judgement, evaluation and application.’
Proposition 2 Interpretation is not an objective process but this does not imply
that anything goes
There is no objective criterion of consistency with the textual evidence on
which to base the rational evaluation of alternative interpretations. But
anything does not go. The interpretive process is historically and socially
contingent. Fish (1980) argues that interpreters are social beings, constrained
by the set of characteristic beliefs of the particular interpretive community to
which the interpreters belong. The interpretive community acts as the
intellectual arbiter, setting the cognitive standards by which new
interpretations are judged. The acceptance of a new interpretation involves
both demonstration and persuasion. The interpretive community must be
persuaded to change its paradigm, that is, the conventional foundations
underlying the currently-accepted interpretation. It is only after the
interpretive community has been persuaded to change its paradigm that the
explanatory power of the new interpretation can be demonstrated.
Proposition 3 The significance of an interpretation is not determined by
correspondence with the author’s original meaning
The significance of an interpretation is something external to the text. It
depends on the successful application of the insights gained from the text. In
the case of an economics text, its significance can be seen as its ability to
explain economic behaviour. The question of whether or not this is what the
author really meant is largely irrelevant in this context. From this
perspective, Stigler’s distinction between personal and scientific exegesis can
be justified to the extent that it represents the hermeneutic distinction
between meaning and significance.
Proposition 4 The existence of multiple interpretations is not a problem
Multiple interpretations of a single text arise partly because of the
inescapable vagueness of the written word. It is this vagueness which gives
the text its reference power, its ability to disclose different possible ways of
looking at the world. The text constrains the range of potential meanings but
cannot determine the actualisation of any specific meaning by the reader.
The continued existence of multiple interpretations need not, therefore, be
treated as a problem to the extent that it is indicative of the high reference
power of the text and the heterogeneity of the readership. It only becomes a
problem if interpreted as such from an objectivist/essentialist perspective.
106 Bill Gerrard
CONCLUSION: HOW TO PIERCE THE KEYNESIAN DOCTRINAL FOG
Much of the doctrinal fog surrounding the interpretation of Keynes has been
generated by the confusion of interpretation and application. Interpreting
Keynes’s General Theory has involved two questions, the question of
interpretation, ‘What does Keynes’s General Theory mean?’, and the
question of application ‘What is the significance of Keynes’s General Theory
for explaining how the economy actually works?’ The two questions are
interdependent but they have tended to become interwoven in a very
confusing manner. For example, Leijonhufvud (1968:1) differentiates between
what he calls the ‘doctrinal-historical’ question of understanding Keynes’s
General Theory and the task of discovering fresh perspectives with which to
understand the economy and with which to assess economic theory.
Leijonhufvud considers the latter to be the primary objective, the doctrinal-
historical question being ‘strictly secondary’. Yet the validity of the fresh
perspective discovered by Leijonhufvud is assessed on the basis of it having
‘a much firmer foundation in Keynes’s writing than can be claimed for the
[income-expenditure] interpretation’ (p. 11). Thus Leijonhufvud considers his
interpretation as the ‘economics of Keynes’. Leijonhufvud does not assess his
interpretation with regard to its relevancy to the understanding of the
economy and economic theory but in terms of its authenticity as an exegesis
of the real meaning of Keynes’s General Theory.
Piercing the Keynesian doctrinal fog requires disentangling the question of
interpretation and the question of application. The significance of Keynesian
economics depends on its ability to provide an understanding of how the
economy actually works. The significance of Keynesian economics does not
depend on being the economics of Keynes. What Keynes himself believed is a
question for the historians of economic thought, not for macroeconomists.
This is not to say that interpretation is unimportant; quite the opposite.
Rather the point is that the usefulness of an interpretation depends on its
ability to generate a better understanding of economic behaviour. The
interpretation of Keynes’s General Theory has relevance for macroeconomics
if and only if it can provide access to new understandings of the macro
economy. Distinguishing between meaning as the aim of interpretation and
significance as the aim of application will do much to overcome the
confusion surrounding Keynes’s General Theory.
The continuing achievement of Keynes’s General Theory is its ability to
generate a diversity of research programmes. This diversity has resulted
from the fusion of Keynes’s horizon as expressed in the General Theory with
a series of very different horizons due to his interpreters and is evidence of
the high reference power of Keynes’s General Theory. It is a text which
continues to disclose a number of different possible ways of looking at the
macro economy. The recognition of this will clear the Keynesian doctrinal
fog. The multiple interpretations of Keynes’s General Theory need no longer
be interpreted as a problem yet to be solved.
Keynes’s General Theory: Interpretations 107
This conclusion that the economics profession should stop worrying about
the multiple interpretations of Keynes’s General Theory is an example of the
therapeutic role that hermeneutics can have in economics. Hermeneutics can
help economists become more self-aware about some of the causes of
controversies in the subject-field, thereby helping to resolve these
controversies. This is the pragmatic justification for introducing
hermeneutics into economics. But hermeneutics can be a double-edged
sword. Hermeneutics provides an attitude of mind, not a set of ready-made
answers. Hermeneutic understanding can only emerge through a dialectical
process of application to specific problems, the interpretation of Keynes’s
General Theory being one such application. If hermeneutics is introduced
into economics as an abstract methodology, it will generate more heat than
light. Hermeneutics is a study of controversy but it is also a study in
controversy. The emergence of the rhetorical approach in economics
exemplifies the possible dangers of replaying the old controversies of other
subject-fields. Hermeneutics can be an effective medicine but should be
marked ‘handle with care’.
ACKNOWLEDGEMENTS
This chapter represents a development of themes contained in Gerrard
(1989). I should like to thank Paul Anand, Roger Backhouse, John Brothwell,
Meghnad Desai, Athol Fitzgibbons, John Hillard, Brian Hillier and the
participants at the RES conference as well as the Economic Journal editors
and an anonymous referee for much in the way of helpful criticism. The
usual disclaimer applies.
REFERENCES
Aksoy, E.G. (1989) ‘Problems of interpretation: hermeneutic circle, objectivity and
validity in the interpretations of Ricardo’, Proceedings of the 16th Annual Meeting
of the History of Economics Society, Richmond, VA, 2, 725–47.
Blaug, M. (1980) The Methodology of Economics: Or How Economists Explain,
Cambridge: Cambridge University Press.
Carabelli, A. (1988) On Keynes’s Method, London: Macmillan.
Chick, V. (1983) Macroeconomics after Keynes: A Reconsideration of the General
Theory, Oxford: Philip Allan.
Coddington, A. (1976) ‘Keynesian economics: the search for first principles’, Journal
of Economic Literature 14, 1258–73.
Fender, J. (1981) Understanding Keynes, Brighton: Wheatsheaf.
Fish, S. (1980) Is There a Text in This Class? The Authority of Interpretive Communities,
Cambridge, MA: Harvard University Press.
Fitzgibbons, A. (1988) Keynes’s Vision, Oxford: Clarendon Press.
Gerrard, B. (1988) ‘Keynesian economics: the road to nowhere?’, in J.V.Hillard (ed.)
J.M.Keynes in Retrospect: The Legacy of the Keynesian Revolution, Aldershot:
Edward Elgar.
Gerrard, B. (1989) ‘Some notes on interpreting Keynes’ General Theory’, IRISS,
University of York Discussion Paper 140.
108 Bill Gerrard
Hahn, F.H. (1982) Money and Inflation, Oxford: Basil Blackwell.
Hamouda, O.F. and Harcourt, G.C. (1988) ‘Post-Keynesianism: from criticism to
coherence?’, Bulletin of Economic Research 40, 1–33.
Hirsch, E.D. (1976) The Aims of Interpretation, Chicago: Chicago University Press.
Hollander, S. (1989) ‘Principles of textual interpretation: illustrated by Ricardian growth
theory’ , Proceedings of the 16th Annual Meeting of the History of Economics
Society, Richmond, VA, 2, 763–74.
Leijonhufvud, A. (1968) On Keynesian Economics and the Economics of Keynes,
Oxford: Oxford University Press.
Milgate, M. (1982) Capital and Employment, London: Academic Press.
O’Donnell, R. (1989a) Keynes: Philosophy, Economics and Politics, London: Macmillan.
——(1989b) ‘Keynes on probability, expectations and uncertainty’, paper presented at
the 9th Keynes Seminar, University of Kent, Canterbury.
Patinkin, D. (1978) ‘Keynes’ aggregate supply function: a plea for common sense’,
History of Political Economy 10, 577–96.
Ricoeur, P. (1976) Interpretation Theory: Discourse and the Surplus of Meaning, Forth
Worth, TX: Texas Christian University Press.
——(1981) Hermeneutics and the Human Sciences, Cambridge: Cambridge University
Press.
Robinson, J. (1964) Economic Philosophy, Harmondsworth: Penguin.
——(1973) ‘What has become of the Keynesian revolution?’ In After Keynes, Oxford:
Basil Blackwell.
Schumpeter, J.A. (1946) ‘John Maynard Keynes 1883–1946’, American Economic
Review 36, 495–518.
Shackle, G.L.S. (1967) The Years of High Theory, Cambridge: Cambridge University
Press.
Stigler, G. (1965) ‘Textual exegesis as a scientific problem,’ Economica 32, 447–50.
5 The fall and rise of Keynesian
economics
Alan S.Blinder
Economic Record (1988) December, pp. 278–94
Keynesian economics came under much criticism in the 1970s. This chapter
argues that the decline in Keynesian economics and the rise in, notably, new
classical economics in this period related to their respective theoretical
appeal rather than their ability to explain developments in the
macroeconomy. As this has become increasingly recognized, and with the
development of sound microeconomic foundations, Keynesian economics has
again been on the rise.
1 INTRODUCTION
According to T.S.Kuhn’s The Structure of Scientific Revolutions (1962),
progress in ‘normal science’ requires an agreed-upon theoretical framework
or ‘paradigm’ within which researchers work to solve puzzles. The stage is
set for a paradigm change when anomalies are discovered and documented.
After a period of turmoil and extensive questioning of old assumptions, a
new theory may emerge which explains not only the anomalies but also the
phenomena encompassed by the old theory. If so, the scientific revolution
succeeds; although the new theory may itself be subsequently supplanted by
a still newer one. Implicitly, a progressive science rarely, if ever, goes back
to a previously discarded theory, for that theory was rejected for good
reasons.
For a period of roughly 35 years, Keynesian theory provided a central
paradigm for macroeconomists, and considerable progress was made on
several empirical fronts. It was widely recognized that some of the
ingredients of Keynesian economics (e.g. money illusion and/or nominal
wage rigidity) rested on slender to non-existent microtheoretic foundations;
and there were always dissenters. But, thought of as a collection of empirical
regularities that fit together into a coherent whole, the theory worked
tolerably well.
In the 1970s, however, the Keynesian paradigm was rejected by a great
many academic economists, especially in the United States, in favour of
what we now call new classical economics. By about 1980, it was hard to
find an American academic macroeconomist under the age of 40 who
110 Alan S.Blinder
professed to be a Keynesian. That was an astonishing intellectual turnabout
in less than a decade, an intellectual revolution for sure.
Scientists from another discipline might naturally surmise that the data of
the 1970s had delivered a stunning and unequivocal rejection of the
Keynesian paradigm. They would look for some decisive observation or
experiment that did to Keynes what the orbit of Mercury did to Newton. But
they would look in vain.1 I argue in Section 3 that there was no anomaly,
that the ascendancy of new classicism in academia was instead a triumph of
a priori theorizing over empiricism, of intellectual aesthetics over
observation and, in some measure, of conservative ideology over liberalism.
It was not, in a word, a Kuhnian scientific revolution.
If this is so, it helps explain a phenomenon that a Kuhnian would find
puzzling: macroeconomics is already in the midst of another revolution
which amounts to a return to Keynesianism—but with a much more rigorous
theoretical flavour. The first stages of the Keynesian counter-revolution—
which is still in progress—are summarized and evaluated in Section 4. But
before doing this, I must define precisely what I mean by Keynesianism. This
I do in Section 2.
2 WHAT IT MEANS TO BE A KEYNESIAN
The word ‘Keynesian’ means many things to many people. Decades ago, it
was a carelessly applied label for economic liberals and interventionists in
general. For a while in the late 1970s and early 1980s it became a pejorative
term more or less synonymous with old-fashioned. No two people have
precisely the same definition of Keynesian economics. But, as one of the few
American economists of my generation who never shunned the label, I feel
entitled to my own definition. To me, the heart of Keynesianism consists of
six principal tenets.
First and foremost, Keynesian economics is a theory of aggregate demand
and of the effects of aggregate demand on real output and inflation. The first
three tenets follow from this.
1 A Keynesian believes that aggregate demand is influenced by a host of
economic decisions, both private and public, and sometimes behaves
erratically. Some decades ago, there were active, impassioned debates over
the propositions that (a) monetary policy is powerless because money
demand is infinitely elastic, or (b) fiscal policy is powerless because money
demand is totally inelastic. But both of these are dead issues now. Essentially
all Keynesians and most monetarists now believe that both fiscal and
monetary policy affect aggregate demand.2 Many new classicals, however,
believe in debt neutrality—the doctrine that substitutions of debt for taxes
have no effects on total demand.
2 According to Keynesian theory, changes in aggregate demand, whether
anticipated or unanticipated, have their greatest short-run impact on real
output and employment, not on prices, and the short run lasts long enough to
The fall and rise of Keynesian economics 111
worry about.3 In textbook expositions, this idea is conveyed by a short-run
aggregate supply curve that is upward sloping, and probably quite flat
except at high levels of capacity utilization, so that changes in aggregate
demand are normally not dissipated in higher prices. In macroeconometric
models, the same idea is captured by treating output and employment as
demand-determined in the short run and letting an inertial Phillips curve
determine inflation.
For a theoretical model to produce real effects from anticipated monetary
policy, it is usually necessary to have some sort of nominal rigidity in the
model; otherwise, an injection of money is like a currency reform which
changes all prices equiproportionately.4 So Keynesian models generally
either assume or try to rationalize nominal rigidities. Because supply and
demand curves derived from standard neoclassical maximizing principles are
always homogeneous of degree zero in nominal quantities, this is not an easy
task. Real effects of government purchases, however, are readily explained
on strictly neoclassical grounds.5
Since prices do not absorb all shocks to demand, fluctuations in any
component of spending will cause sympathetic movements in output. In most
Keynesian models, the latter are larger than the former because of the
multiplier; but a multiplier greater than one is not central to Keynesian
analysis. A positive real multiplier is.
Although real effects from demand fluctuations are often called
‘Keynesian effects’, most monetarists accept the idea as well—at least as it
pertains to monetary policy. So this tenet does not really divide those two
schools of thought. However, at least some new classicals insist that changes
in money affect real output only if they are unanticipated.
3 Keynesians believe that goods markets and, especially, labour markets
respond only sluggishly to shocks, i.e. that prices and wages do not move
quickly to clear markets. This issue, once again, divides Keynesians more
from new classicals than from monetarists—although monetarists probably
place more faith in the economy’s natural servomechanism than Keynesians
do. Milton Friedman (1968:13), for example, has written that ‘Under any
conceivable institutional arrangements, and certainly those that now prevail
in the United States, there is only a limited amount of flexibility in prices
and wages.’ In current parlance, that would certainly be called a ‘Keynesian’
position.
The next three tenets have to do directly with policy; and here Friedman
and other monetarists part company with most Keynesians.
4 To a Keynesian, the actual levels of employment and unemployment
have no special claim to optimality—partly because unemployment is
subject to the caprice of aggregate demand, and partly because they believe
that markets clear only gradually. In fact, Keynesians typically see
unemployment as both too high on average and too variable, although they
know that rigorous theoretical justification for these positions is hard to
come by. Keynesians also feel certain that periods of recession or depression
112 Alan S.Blinder
are economic maladies, not Pareto-optimal responses to unattractive
technological opportunities. All this is summarized in the term ‘involuntary
unemployment’, which Keynesians deplore even though it has proved
notoriously difficult to define.6 On this tenet, new classicals differ sharply
from Keynesians, with monetarists somewhere in between.
5 Many, but not all, Keynesians advocate activist stabilization policy to
reduce the amplitude of business cycles, which they rank among the most
important of all economic problems. Here monetarists generally join new
classicals, as well as some conservative Keynesians, in doubting both the
efficacy of stabilization policy and the wisdom of attempting it. Some new
classicals go even further and question whether business cycles are a serious
problem at all.7
The argument that economic knowledge is not secure enough to support
what used to be called fine tuning is by now widely accepted, even by most
Keynesians. Yet many Keynesians believe that more modest goals for
stabilization policy—coarse tuning, if you will—are not only defensible but
sensible. For example, an economist need not have detailed quantitative
knowledge of lag structures to prescribe a dose of expansionary monetary
policy when the unemployment rate is 10 per cent or more—as it has been in
many countries in the 1980s. Furthermore, and this may be the most
important point, the nature of government seems to abhor a vacuum of
economic advice. If economists with admittedly limited knowledge refuse to
offer their expert (if uncertain) counsel, assorted quacks with no knowledge
at all will surely rush in to fill the void.
6 Finally, and even less unanimously, many Keynesians are more
concerned about combating unemployment than about conquering inflation.8
However, there are plenty of anti-inflation Keynesians; most of the world’s
current and past central bankers, for example, merit this title whether they
like it or not. Needless to say, relative attitudes toward unemployment and
inflation heavily influence the policy advice that economists give and that
policy-makers accept. As a broad generalization, I think it safe to say that
Keynesians are typically more aggressive about expanding aggregate
demand than are non-Keynesians.
My six tenets divide naturally into two equal groups: the first three are
clearly assertions about positive economics while the last three are mostly
normative. The division of Keynesian economics into positive and normative
components is central to understanding both the academic debate and its
relevance to policy.
Positive Keynesianism is a matter of scientific judgement. A positive
Keynesian believes that both monetary and fiscal policy can change
aggregate demand, that fluctuations in aggregate demand have real effects,
and that prices and wages do not move rapidly to clear markets. No policy
prescriptions follow from these beliefs alone. And, as I have indicated, many
economists who do not call themselves Keynesian would nevertheless accept
the entire list.
The fall and rise of Keynesian economics 113
Normative Keynesians add both value judgements and political
judgements to the preceding list. A normative Keynesian believes that
government should use its leverage over aggregate demand to reduce the
amplitude of business cycles. He or she is probably also far more interested
in filling in cyclical troughs than in shaving off peaks. These normative
propositions are based on judgements that (a) macroeconomic fluctuations
significantly reduce social welfare, (b) the government is knowledgeable and
capable enough to improve upon free-market outcomes, and (c)
unemployment is a more important problem than inflation.9
The long, and to some extent continuing, battle between Keynesians and
monetarists, you will note, has been primarily fought over the normative
issues—particularly (b) and (c).10 Thus, by my definitions, most monetarists
are positive Keynesians but not normative Keynesians. So are other
conservatives who shun the label Keynesian. Protagonists in these debates
agree on most positive issues but make different value judgements and seat-
of-the-pants political judgements, and so reach different conclusions about
policy. Their disagreements in many ways mirror disagreements among
policy-makers.
The briefer, but more intense, debate between Keynesians and new
classicals had, by contrast, been fought primarily over the tenets of positive
Keynesianism. New classicals argue that anticipated changes in money do
not affect real output; that markets, including the labour market, clear
quickly by price;11 and that business cycles may be Pareto optimal. Here
‘objective’ scientific evidence can be brought to bear and, in my judgement,
the evidence on all three issues points strongly in the Keynesian direction.12
But rather than try to summarize that evidence now, I want to make only one
point: that arguments over the positive aspects of Keynesian economics are
potentially resolvable by the accumulation of scientific evidence in a way
that disputes over normative issues are not.
Before leaving the realm of definition, let me underscore several glaring
and intentional omissions.
First, I have said nothing about rational expectations. Many Keynesians
are doubtful about the validity of rational expectations as a behavioral
hypothesis, as was Keynes himself.13 Others are willing to accept it. But,
when it comes to the large issues with which I have concerned myself so far,
nothing much rides on whether or not expectations are rational. In
particular, rational expectations models with sticky prices—like those of
Fischer (1977) and Taylor (1980), for example—are thoroughly Keynesian by
my definition. Details of model construction and quantitative answers to
specific questions do, of course, depend on how expectations are modelled.
And, for some issues, the expectational mechanism is crucial.14 But, for the
most part, these are not central to the debate between new classical and
Keynesian economists.15
The second omission is the natural rate hypothesis. Pre-1970
Keynesianism included a Phillips curve that was negatively sloped even in
114 Alan S.Blinder
the long run. This idea was rejected theoretically by Milton Friedman
(1968), a monetarist, and Edmund Phelps (1968), a Keynesian, and shortly
thereafter was also rejected in econometric studies by Keynesians like Robert
Gordon (1972). Since about 1972, a Phillips curve that is vertical in the long
run has been an integral part of Keynesian economics. So the natural rate
hypothesis played essentially no role in the intellectual ferment of the 1972–
85 period. Ironically, however, questions about its validity are now playing a
role in the Keynesian renaissance. Specifically, models with hysteresis have
reopened the theoretical and empirical debate over the natural rate
hypothesis, especially for Europe. (More on this below.)
Third, I have ignored the choice between monetary and fiscal policy as
the preferred instrument of stabilization policy. People differ along this
dimension and occasionally change sides. By my definition, however, it is
perfectly possible to be a Keynesian and still believe either that responsibility
for stabilization policy should in principle be ceded to the monetary
authority or that it is in practice so ceded.
3 THE FALL OF KEYNESIAN ECONOMICS
To start with, let me first dispose of the view, promoted in some quarters,
that the demise of Keynesian economics was due to the doctrine’s poor
empirical predictions. Robert Lucas (1981:559) wrote that ‘Keynesian
orthodoxy is in deep trouble, the deepest kind of trouble in which an applied
body of theory can find itself: It appears to be giving seriously wrong
answers to the most basic questions of macroeconomic policy’. He was
talking about the collapse of the Phillips curve in the US during the 1970s,
which he and Thomas Sargent had characterized as ‘econometric failure on
a grand scale’ (Lucas and Sargent 1978:57).
It is, of course, true that pre-1972 Phillips curves were ill-equipped to
handle the food and energy shocks that dominated the period from 1972 to
1981 and, in consequence, badly underestimated inflation. But it is also true
that Keynesians quickly added supply-side variables (like oil or import
prices) to what had up to then been an entirely demand-oriented theory.16
Soon thereafter supply shocks were also appended to empirical Phillips
curves.17 By the early 1980s, numerous studies had documented the fact that
a conventional Phillips curve equation with a supply-shock variable (any one
of several will do) fits the US data of the 1970s and 1980s extremely well.18
The charge that empirical Keynesian models were, in Lucas and Sargent’s
(1978) words, ‘wildly incorrect’ is, well, wildly incorrect.
One objection frequently raised by supporters of new classical economics
is that saving the Phillips curve after the fact by adding supply variables is
like saving Ptolemaic astronomy by adding a new epicycle. I disagree. Any
economic model is fundamentally a set of statements about the behaviour
underlying supply and demand and the nature of the shocks impinging on
each. For example, an empirical model of the market for wheat consists of a
The fall and rise of Keynesian economics 115
negatively sloped demand curve, a positively sloped supply curve, and some
assumptions about the shocks hitting each. Analogously, a macroeconomic
model must specify not only aggregate demand and supply behaviour but
also the nature of the shocks that buffet the economy.
The empirical correlations implied by either sort of model depend on both
the model’s structure and the shocks that predominate during a particular
historic period. For example, the same structural model of the wheat market
will predict that price and quantity are negatively correlated if most of the
shocks emanate from the supply side but positively correlated if most of the
shocks come from the demand side. Analogously, pre-1973 Keynesian theory
produced a negatively sloped statistical Phillips curve because of an unstated
assumption that macroeconomic shocks come solely from the demand side—
an assumption proved wrong by the events of the 1970s and 1980s. The very
same model generates a positively sloped Phillips curve if the shocks come
from the supply side, which is just what the econometric evidence says
happened in the 1973–81 period.
If you don’t trust econometrics, the following back-of-the envelope
calculation should help drive home the point. Keynesian economists in the
US in the 1960s and early 1970s developed what I used to call the Brookings
Rule of Thumb: that each point-year of unemployment above the natural
rate reduces the rate of inflation by 0.4–0.5 of a percentage point. Using a
5.6 per cent natural rate, the US experienced about 15 point-years of extra
unemployment between 1980 and 1985 and, during those years, the inflation
rate declined about 6–7 percentage points. Once you see how well the rule of
thumb worked, you understand why conventional Phillips curves fit data
from the 1980s so well.
Why, then, was the alleged demise of the Phillips curve trumpeted so
loudly and so widely? I think the reason was the conjunction of two events–
one historical, the other intellectual.
First, when supply shocks came to dominate the data in the 1970s, the
familiar negative correlation between inflation and unemployment—which is
clearly visible on a scatter diagram of data for the 1950s and 1960s–
disappeared. The Phillips curve could no longer be depicted in two
dimensions. To those too unsophisticated to distinguish between a simple
correlation and a multivariate relationship, that seemed equivalent to the
death of the Phillips curve.
Second, Lucas’s (1976) insightful critiques of econometric policy
evaluation provided an elegant a priori argument for why an empirical
Phillips curve might collapse under the weight of a more inflationary
policy.19 Briefly, the argument went like this. A prototypical empirical
Phillips curve explains inflation by lagged inflation and unemployment:
(1)
116 Alan S.Blinder
but is meant to signify a theory in which inflation really depends on expected
inflation and unemployment:
(2)
It thus embodies an auxiliary hypothesis that the distributed lag a(L) t-1 is a
good statistical proxy for expected inflation. Lucas pointed out, correctly,
that (1) will continue to fit the data well only as long as a(L) t-1 remains the
best predictor (i.e. the rational expectation) of inflation. If policy changes,
the best forecasts of future inflation might also change, making (1) break
down even if (2) is stable.
Academic readers of Lucas put two and two together and jumped like
lemmings to the wrong conclusion. The facts were (a) that inflation rose and
(b) that the correlation between inflation and unemployment changed. The
(untested) assertion was that the Lucas critique explained why (b) followed
from (a): the government had adopted a more inflationary policy, which in
turn had changed a(L).
It was remarkable how uncritically the Lucas critique was accepted. Had
governments really decided to ‘ride up’ the Phillips curve toward higher
inflation, as Lucas claimed, or had they simply encountered bad luck from the
supply side? The former was assumed even though the latter seems clearly to
have been the dominant factor quantitatively.20 Did the more inflationary
environment shift the distributed lag a(L)? Rather than seek evidence on this
point, partisans of the Lucas critique became econometric nihilists. Theory, not
data, was supposed to answer such questions; and theory allegedly said yes.
But, in fact, a rise in inflation need not mean that the univariate
autoregressive representation of inflation must change (other than its
constant). Whether or not the lag coefficients a(L) actually shifted in the
early 1970s is an empirical question. To investigate whether or not such a
shift took place, I estimated simple autoregressions for US inflation over the
period 1955:2 to 1987:4 subperiods. As a way of guarding against the
danger of choosing among competing regressions on the basis of prior
beliefs, the lag length (four quarters) and the price index (the GNP deflator)
were specified a priori and never changed. I tested for statistically significant
breaks in the autoregression at the ends of 1970, 1971, 1972 and 1973. The
resulting F statistics were as follows:
break period F statistic
1970:4/1971:1 0.92
1971:4/1972:1 0.85
1972:4/1973:1 0.77
1973:4/1974:1 0.21
None of these F statistics is remotely close to conventional significance levels.
Thus, there is no evidence for a shift in the lag coefficients a(L). And that, in
turn, suggests that the breakdown of the old-fashioned Phillips curve cannot be
The fall and rise of Keynesian economics 117
attributed to the reason emphasized by Lucas. The strongest evidence for a
break emerges if the sample is split 1955:2–1970:4 vs. 1971:1–1987:4. In that
case, the a(L) coefficients sum to 0.73 in the first period and 0.88 in the
second, which is an increase, though not a dramatic one.
I have already noted that, once supply variables are added, contemporary
Phillips curves look much like their ancestors of 1973. Supply shocks not
only provide a more parsimonious explanation for both the rise of inflation
and the fall of the Phillips curve, but one that can be substantiated
empirically. Yet academic economists, at least American academic
economists, opted en masse for Lucas’s explanation, deserting Keynesianism
in the process. Why? The rest of this section gives my personal answers.
They are rooted in the sociology of science, in attachment to theory, and in
ideology—not in empiricism. I take up the three factors in turn.
The sociology of economics
Many people have observed that economics has become a highly technical
subject in recent decades, more so in the US than elsewhere. And technicians,
of whatever discipline, prize technique; it’s how the young cut their teeth.
The rational expectations revolution was a godsend for aspiring young
technicians. It not only pushed macroeconomic theory in more abstract and
mathematical directions, but brought in its wake a new style of econometrics
that was far more technically demanding than the old methods it sought to
replace.21
The tools needed to carry out the new brands of theory and econometrics
could not be found in the kit bags of the older economists, which gave the
young a heavy competitive edge. Not only were they better trained
mathematically and, being younger, more flexible of mind, but also they
were less distracted by other pursuits and hence more willing and able to
absorb the new techniques. As an extra bonanza, the Lucas critique provided
a reason to shun the previously accumulated stock of econometric results as
unreliable. Thus freed of any need to absorb the knowledge of the past,
newly minted Ph.D. economists could concentrate on developing what they
saw as the wave of the future.
It was a recipe for generational conflict within the discipline and, sure
enough, the young were recruited disproportionately into the new classical
ranks while few older economists converted.22 Traditional Keynesian tools
like IS-LM and large-scale macroeconometric models came to be viewed as
relics of the past and, in a strange kind of guilt by association, Keynesian
ideas like those discussed in Section 2 also came to be seen as outmoded. By
1980 or so, the adage ‘there are no Keynesians under the age of 40’ was part
of the folklore of the (American) economics profession.
The saying, of course, was meant to encompass only academic economists
and, indeed, only those in the elite institutions. In fact, virtually no non-
academic economists converted to new classicism. Why the sharp bifurcation
118 Alan S.Blinder
between professors on the one hand and business and government economists
on the other? Part of the answer is that scholars are naturally the producers
of new ideas while practitioners are the consumers. Fundamental debates
over theory and statistical method belong in the academy, where the
protagonists are better equipped to deal with them and have the luxury of a
long time horizon. That, I suppose, is what ivory towers are for.
But another part of the explanation lies in the different market tests the
two groups must meet. In academia, as in fashion, it is more important to be
fresh and creative than to be correct. Cute models, after all, make snappy
papers; the real world can be left to less original minds. I have heard it said
that the surest route to academic success is to devise a clever proof of an
absurd proposition. And dazzling displays of technical fireworks, perhaps
accompanied by some impenetrable prose, regularly impress referees and
editors of scholarly journals.
Incentives are quite different in business or government, where the
important thing is to produce the right answer—or, rather, to appear to
produce the right answer. Methodological innovation and purity count for
little, cuteness for nothing, and technical virtuosity is unappreciated. A
professional forecaster seeks accuracy, not scholarly kudos. A policy analyst
wants to communicate with policy makers, not to dazzle them with
technique.
That new classical ideas failed to migrate from the academy to the worlds
of business and government—as Keynesian ideas had done 40 years earlier—
suggests that they failed to meet the non-academic market test: they did not
produce useful results. But that is getting ahead of my story.
The nature of economic theory
The triumph of new classical ideas in academia was also rooted in the
nature of economic theory and in economists’ fierce loyalty to it. We
economists proudly distinguish ourselves from the lower social sciences by
pointing to our illustrious theoretical heritage. In the economist’s world,
rational and self-interested people optimize subject to constraints. The
resulting decision rules equating ‘marginal this’ to ‘marginal that’ lead to
supplies and demands, which interact in markets to determine prices. These
prices, in turn, guide the allocation of resources and the distribution of
income. If not interfered with, markets tend to be highly competitive and
have a strong tendency to clear by price. (Here the consensus begins to fray
a bit.)
These are the canons of our faith. They are what gives economics the
unity and cohesion that, say, sociology lacks. Rightly or wrongly, they also
imbue economists with an imperialistic attitude toward the other social
sciences—rather like Kipling’s attitude toward India. We have a tight theory;
they don’t. We should treat the heathen kindly, if condescendingly, while we
firmly propagate the faith.
The fall and rise of Keynesian economics 119
Notice, however, that the central economic paradigm is entirely
microeconomic. Keynesian macroeconomics coexists with it uneasily at best.
In at least some Keynesian models, workers are less than rational. (For
example, they may harbour money illusions.) Relative wages and notions of
fairness probably matter in labour markets. Decision-makers frequently
bump into corners, so that optimal decisions are no longer described by
neoclassical marginal conditions. Markets may not clear, and in fact may
display surpluses for long periods of time; so trading takes place at ‘false
prices’. In all these respects and others, Keynesians have long been infidels in
the neoclassical temple.
The strength of neoclassical fundamentalism has ebbed and flowed over
the decades. The worldwide depression of the 1920s and 1930s undermined it
severely, thus paving the way for the Keynesian revolution. The prosperity of
the 1960s and early 1970s probably helped restore it. New classical
economics was quite explicitly a revival of neoclassical orthodoxy, a return
to what Lucas (1987), echoing Marshall, called ‘the only engine for the
discovery of truth’.
Keynesians had long felt an agonizing tension between the
macroeconomics they taught on Mondays and Wednesdays and the
microeconomics they taught on Tuesdays and Thursdays. New classicals
explicitly sought to end this tension by making macroeconomics more like
microeconomics. All supply and demand decisions were to be derived
rigorously from neoclassical ‘first principles’. Aggregate demand and supply
schedules were to be viewed as blow-ups of interior solutions to individual
optimization problems. Markets were to be viewed as perfectly competitive
and clearing. If necessary, bothersome empirical phenomena like involuntary
unemployment were to be ignored, defined out of existence, or ingeniously
rationalized by convoluted theoretical arguments.
Methodological purity has a seductive attraction to mathematically
minded technicians—which helps explain why rational expectations came to
be so intimately tied up in the debate. Modelling expectations as rational—
that is, as optimal subject to informational constraints—is the analogue of
modelling consumers as maximizing utility and producers as maximizing
profits. Rational expectations was therefore a natural accompaniment– and,
indeed, a major impetus—to the ‘back to basics’ movement. It was no
accident, then, that those who favoured frictionless, optimizing, market-
clearing models were immediately attracted to rational expectations as a
behavioural hypothesis without bothering to look for evidence. Linking
rational expectations to new classicism (thus leaving ‘irrational’ expectations
to the Keynesians) helped the new theory win converts in the same way that
celebrity endorsements help sell products. Theoretically minded economists
were predisposed to believe in rational expectations and, at first, took the
new classical baggage along with it.
120 Alan S.Blinder
The role of conservative ideology
There were also ideological overtones in the neoclassical revival which I
have yet to mention, but which played an important role.
The basic neoclassical paradigm is profoundly conservative, as other
social scientists—and, sometimes, our own students—remind us. Those who
take it seriously as a description of the economy tend toward the Panglossian
view of private economic transactions and look askance at government
intervention. When this world view is transported from microeconomics to
macroeconomics, it leads to theoretical models in which business cycles are
benign, unimportant, or inevitable—perhaps all three. And it leads, as usual,
to laissez-faire policy recommendations. For example, Edward Prescott
(1986) asserts that ‘costly efforts at stabilization are likely to be
counterproductive’ because the free-market business cycle is Pareto optimal.
Keynesians, as I indicated in Section 2, do not buy any of this. They argue
that the very existence of macroeconomics as a subdiscipline owes to the
massive market failures that we observe during recessions but which the
neoclassical paradigm rules out. They believe that recessions are important,
malign, and ameliorable, and so are ready to support government
interventions designed to stabilize aggregate economic activity. As James
Tobin once remarked, they worry more about Okun gaps than Harberger
triangles.
The relative strengths of conservative and liberal ideology obviously vary
both over time and through space. My argument is that new classical theory
could have attracted a large following only in a country and at a time when
right-wing ideology was on the ascendancy, as was true in the United States
in the 1970s and 1980s.23 Though we academics live in ivory towers, the
social winds blow there, too.
Many observers have noticed that the new classical revolution was mainly
restricted to the United States; it never really caught on in Europe. That was
no coincidence, I think, for right-wing ideology has long found more
adherents in the US than in Europe. The timing was also no accident; new
classicism took root just when the political balance in the US was shifting
toward the right. I don’t believe such ideas would have sold in American
academia during the 1960s.
What I have just said about the theoretical and ideological roots of new
classical economics could equally well have been said about old classical
economics. But the 1970s did not witness a revival of Pigou, or even of
Friedman. It saw, instead, a movement towards the high-tech economic
theory of Lucas and the high-tech econometrics of Sargent. The secret to the
success of the new classical economics is that it managed to be at once
ideologically backward looking and technologically forward looking. Given
the temper of the times, that was a winning formula.
Or, rather, it was a winning formula in academia. Outside the academy,
the emphasis on theoretical purity (at the possible expense of empirical
The fall and rise of Keynesian economics 121
validity) and technical wizardry were liabilities, not assets. In addition, the
leaders of the new school, particularly Lucas and Sargent, were disinclined
to press their views on policy-makers because they deemed macroeconomic
science insufficiently developed to support such advice. Finally, as we shall
see in the next section, the empirical implications of new classical theory
were wide of the mark. For all these reasons, the theory that swept academia
made hardly a ripple in the world of policy.
4 THE RISE OF KEYNESIAN ECONOMICS
I have argued that empirical evidence played little or no role in the fall of
Keynesian economics in academia, which I have attributed instead to the
theory’s weak microeconomic underpinnings, to the curious sociology of our
discipline, and to the rise of right-wing ideology.24 The story behind the
recent resurgence of Keynesianism is quite different, for here the empirical
failures of new classical economics are central. In addition, however, new
strains of theory are beginning to resolve the tension between
microeconomics and macroeconomics in a fascinating way. Whereas new
classical economists sought to remake macroeconomics in the image of
neoclassical microeconomics, recent developments in economic theory may
eventually lead to a reformulation of micro theory that resembles Keynesian
economics. I will discuss each of these in turn, beginning with empirics.
Empirical evidence against the new classical paradigm
In view of the normally strong interplay between events and ideas, it is
somewhat astounding that new classical economics caught on during the
second half of the 1970s—a time when most of the world’s industrial
economies were struggling to emerge, often unsuccessfully, from deep and
long recessions.
True to its classical roots, new classical theory emphasized the ability of a
competitive price-auction economy to cure recessions by wage-price
deflation. Its early forms attributed downturns to misperceptions about
relative prices (such as real wages) that arise when people do not know the
current price level, and implied that unemployment should vibrate randomly
around its natural rate. But such misperceptions surely cannot be large in
societies in which price indexes are published monthly and the typical
monthly inflation rate is under 1 per cent; and they cannot be persistent if
expectations are rational. Yet economic fluctuations in the late 1970s and
1980s were both large and persistent.
Later versions of new classical theory replaced monetary misperceptions
with changes in perceived intertemporal terms of trade and added several
features which produced persistent movements in employment and output.25
But empirical research has never been able to find large intertemporal
substitution effects. And theories that generate employment fluctuations from
122 Alan S.Blinder
the supply side of the labour market stumble over the facts that labour
supply looks to be quite inelastic (at least in the US) and real wages are
nearly constant over the business cycle. They also have a hard time making
the jump from persistent changes in employment to persistent—not to
mention involuntary—unemployment. In stark contrast, the Keynesian model
may be theoretically untidy; but it is certainly a model of persistent,
involuntary unemployment.
So the events of the late 1970s seemed to support the incumbent theory
and undermine the challenger. Yet the challenger prevailed. Curious.
Next came the 1980s, which were ushered in by another oil shock but
were dominated by the Reagan-Volcker fiscal and monetary policy shocks
and the European depression. I think it fair to say that new classical
economics shed little light on any of these events. The events, however, cast
deep shadows across the theory.
First according to new classical theory, a correctly perceived deceleration of
money growth affects real output only via its effects on anticipated inflation and
real interest rates. Virtually no one thinks real interest rate effects are very large,
which is why simple models often ignore them. Yet when the Federal Reserve
and the Bank of England announced that monetary policy would be tightened to
fight inflation, and then made good on their promises, severe recessions followed
in each country.26 Could it have been that the tightening was unanticipated?
Perhaps in part. The Fed did seem to get carried away, and perhaps both central
banks lacked credibility at first. But surely the broad contours of the restrictive
policies were anticipated, or at least correctly perceived as they unfolded.27
Old-fashioned Keynesian theory, which says that any monetary restriction
is contractionary because firms and individuals are locked into nominal
contracts, seems more consistent with actual events, even though it doesn’t
explain why nominal contracts exist. Strike one against the new theory.
Second, an offshoot of new classical theory due to Barro (1974) argued
that debt-financed tax reductions should have neither real nor nominal effects
because rational agents, correctly perceiving their future tax liabilities or
those of their heirs, would act to offset them. The only observable
consequence of such a policy, on this view, should be a rise in private saving
to offset the government dissaving.
Naive Keynesian analysis, by contrast, sees the same event as an outward
shift of the IS curve. If the LM curve is unchanged, real interest rates, real
output, and the price level should all rise. If, as happened in the US, the
stimulus to demand is snuffed out by contractionary monetary policy, real
interest rates should rise even more. There is no reason to expect the private
saving rate to rise.
Econometric studies of the Barro hypothesis have yielded highly
inconclusive results. The answer seems to depend on who asks the question.28
Observation of the real world seems to deliver a stronger verdict, however.
Taxes were cut massively in the US between 1981 and 1984. Given the thin
economic rationale for the policy, the Reagan tax cuts come as close to a
The fall and rise of Keynesian economics 123
truly exogenous fiscal experiment as we are ever likely to get—just the sort
of thing that helps scholars discriminate among competing theories. What
happened? The private saving rate did not rise. Real interest rates soared,
even though a surprisingly large part of the shock was absorbed in exchange
rates rather than in interest rates (so that net exports were crowded out rather
than domestic investment). Real GNP growth seems not to have been
affected; it grew at about the same rate as it had in the recent past.
It would be unfair to say that neoclassical theory offers no explanation for
these events. A sudden rise in the productivity of capital in the US would be
expected to raise domestic interest rates (and rates of return), draw in capital
from abroad (thus causing a current account deficit), and appreciate the
currency. The only trouble with this explanation is that the alleged jump in
the productivity of capital is unobservable and unexplained. Why, for
example, did it not also happen in other countries?29 Why did measured
productivity growth not accelerate? Furthermore, neither private saving nor
investment really rose much as a share of US GNP. The neoclassical
explanation does successfully explain the puzzling rise in the US stock
market. But, if the productivity of capital soared only in the US, why did
stock markets boom all over the world? And if the rise in capital’s
productivity was global, why did capital come pouring into the United
States? Strike two against new classical theory.
Third, we then have the nasty matter of the European depression which,
in some countries, has been as long and as deep as the depression of the
1920s and 1930s and which, at this writing, is still in progress. The
Keynesian explanation is straightforward. Governments, led by the British
and German central banks, decided to fight inflation by highly restrictive
monetary and fiscal policies. The anti-inflationary crusade was facilitated by
the European Monetary System which, in effect, spread the stern German
monetary policy all over Europe. If Keynesian theory has any trouble
explaining these events, it is because modern versions which incorporate the
natural rate hypothesis are not Keynesian enough. (More on this below.)
The new classical explanation of the European depression is…well,
frankly, I am not sure there is one. Proponents of new classicism, and
conservative economists in general, point to microeconomic interferences in
labour markets. But most of these policies (like generous unemployment
insurance) were in place in 1973 when unemployment was extremely low. In
my country, three strikes and you are out. It is therefore not surprising that
new classical economics began to lose supporters.
Even this recent history might not have been decisive, given the insular
attitudes of academic economists. But there was more scholarly evidence as well.
First, new classical economists had made the Phillips curve a test case and
interpreted it in their favour. But, as I have already related, a succession of
econometric studies in the 1980s all concluded that the empirical Phillips
curve was alive and well once you allowed for supply shocks, at least in the
US. Gordon (1987) argues much the same for Europe.
124 Alan S.Blinder
Second, the newly developed technology for estimating models with
rational expectations began to be applied; and the results were almost
uniformly unfavourable to the new classical view.30 Normally, the trio of
hypotheses that (a) expectations are rational, (b) decision rules are first-order
conditions to well-defined optimization problems, and (c) markets clear had to
be tested jointly. And almost always the joint hypothesis was resoundingly
rejected. Which was the weak leg of the tripod? Most economists, instinctively
attracted by rational expectations, thought it was market clearing. But it really
didn’t matter, for the new classical edifice required the entire lot.
Finally, the validity of the rational expectations hypothesis itself was
called into question. Directly observed expectational data were used to test
for rationality. Mostly, these were tests of the weak forms of rationality:
unbiasedness and/or efficiency. They did not, and could not, test for the
much stronger form of rational expectations required by new classical
theory: that people’s subjective expectations match the mathematical
expectations implied by the model. None the less, most of these tests rejected
rational expectations.31
So by 1983 or 1984, academic macroeconomics was in the following
somewhat embarrassing position. Keynesian economics had been maligned on
the grounds that its theoretical foundations were prosaic at best, non-existent
at worst, and certainly inelegant. Its heir apparent, new classical economics,
boasted an elegant and technically sweet theory which passed internal
consistency checks with flying colours, but which failed miserably when it
came to consistency with observation. In the shorthand that was used both then
and now, Keynesian economics was ‘bad theory’ which none the less seemed
consistent with the facts while new classical economics was ‘good theory’
which, unfortunately, did not describe the way the world works.
This is, it seems to me, a curious usage of the terms ‘good’ and ‘bad’– one
which reflects the academic economist’s preoccupation with elegance and
mathematical structure over relevance and empirical accuracy. By these
criteria, the ‘good theory’ is not the one that explains the data best, but
rather the one that is truest to neoclassical orthodoxy—which sees people as
self-interested and maximizing individuals, who calculate well, have no
money illusion, and don’t leave unexploited profit opportunities. That is the
attitude of a mathematician who deals in logical constructs, not of a scientist
who deals in facts. If real people are social beings who care not just about
their own well-being but also about their relative position in society, who are
not very good at doing calculations or deflating by a price index, and who
have other things to do besides maximizing all the time, then Keynesian
theory may be the ‘good’ theory after all—even if it is contaminated by ideas
from other social sciences.
Of course, a theory can be judged good or bad only relative to some
competitor. There are several senses in which Keynesian theory was and is
not good enough. One is that empirical problems continue to beset
macroeconometric models built in the Keynesian tradition. The collapse of
The fall and rise of Keynesian economics 125
the LM curve is just the most obvious of these empirical failures, not the
only one. Another problem is that the Keynesian model has such a weak
microtheoretic structure that it is hard—some would say impossible—to do
welfare economics with it. While most Keynesians believe that successful
stabilization policies improve social welfare, the theory itself does not really
justify that belief.
In any case, the view in academia was then (and in some circles still is)
that economists had to choose between a tight theory with severe empirical
problems and a sloppy theory that none the less worked better empirically.
There were two ways to proceed. Either efforts could be made to make
Keynesian economics more theoretically respectable, or energy could be
devoted to bringing new classical economics into closer contact with reality.
Research is proceeding in both directions. In my judgement, the work that is
being done along the first route is much the more interesting and promising,
so I will dwell on that.
New theoretical foundations for Keynesian economics
Four new developments in economic theory, all of them still in progress,
seem to me not only to shore up the theoretical foundations of Keynesianism,
but actually to push micro theory in a Keynesian direction. None of them
puts sticky nominal wages at centre stage.
Monopolistic competition
The first idea is to build a macro structure on the foundations of
monopolistic, rather than perfect, competition. This helps produce a
Keynesian environment in two respects. First, it leads to theoretical models
in which firms always want to sell more at current prices because price
exceeds marginal cost. Second, output levels in monopolistic equilibria are
generally below the social optima, which echoes the Keynesian idea that
employment is typically too low. The knotty intellectual problem was
always that monopolistic competition theory pertains strictly to relative
prices while nominal magnitudes matter in Keynesian macroeconomics.
Mankiw (1985) and Akerlof and Yellen (1985) solved this problem at
more or less the same time by adding fixed costs of changing nominal
prices to the model.32 Suppose the money supply (M) falls by a small
amount. Fixed costs of changing prices will deter some firms from cutting
their nominal prices even though their first-bet nominal prices in a
frictionless world would be proportional to M. In consequence, the price
level will fall less than proportionately to M and real balances, and hence
aggregate demand, will decline. More than likely, so will social welfare.
On the up side, a small enough rise in M will induce only some firms to
pay the fixed costs of raising their prices. So real balances, aggregate
demand, and social welfare will all rise.
126 Alan S.Blinder
Mankiw and Akerlof and Yellen pointed to a kind of externality later
made more precise by Blanchard and Kiyotaki (1987). This idea is important
because economists like to rest the case for government intervention on
externalities.
The argument goes as follows. Since each firm sits at the top of its profit
hill, a small deviation from its first-best relative price has only a small effect
on its profits. But because the pre-existing (monopoly) distortion causes
output to be below the socially optimal level, the loss in social welfare is
greater than the drop in profits. Although the firm loses little from its
deviation from optimality, society loses much.
The so-called aggregate demand externality arises in the following way–
which should sound familiar to Keynesians. In equilibrium, individual firms
do not find it profitable to reduce their prices. Yet, if all firms would cut
their prices simultaneously, real balances would rise, aggregate demand
would expand, and all firms’ profits (and social welfare) would rise. In a
decentralized economy, there is no way to achieve such coordinated price
cutting. But a sufficiently large rise in the money stock can accomplish the
same thing—just as Keynes suggested more than 50 years ago.
This new strain of theorizing is appealing because it relies on just three
seemingly realistic assumptions: (a) that demand curves for individual firms
slope down: (b) that firms maximize profits; and (c) that lumpy transactions
costs are incurred whenever a nominal price is changed. However, it does
not provide a complete theoretical justification for Keynesian economics for
several reasons.
The first is a technical point to which I will return. In a dynamic
economy, fixed costs of price adjustment should lead firms to allow their
relative prices to drift away from their first-best profit-maximizing levels
most of the time. In that case, firms are not atop their profit hills, so a small
change in a relative price (caused, say, by inflation) may have a large effect
on profits—not the small effect envisioned by the theory.
Furthermore, while the theoretical results of monopolistic competition
models are consistent with Keynesian insights, they lack certain important
Keynesian features. For one thing, the main finding is that output is normally
too low, not that it is too variable. Hence the obvious policy intervention is
an output subsidy, not macro stabilization policy. Using this class of model
to justify the Keynesian belief that output is too variable turns out to be quite
tricky.33 For another, the models do not produce any natural notion of
involuntary unemployment which, as I noted earlier, plays a central role in
the Keynesian tradition.
Efficiency wages
The next group of theories I will consider addresses itself directly to the
involuntary unemployment question. Several microeconomic theories of the
labour market based on imperfect, and usually asymmetric, information
The fall and rise of Keynesian economics 127
show that the market can be in equilibrium—in the sense that there are no
unexploited profit opportunities—with supply unequal to demand. The
simplest, and to me the most appealing, of these is the efficiency wage
model. It also seems to accord best with common sense.
Here is a simple example that makes the point starkly.34 Suppose output,
f(eL), depends on labour input in efficiency units, where L is physical labour
input and e indicates effort. Suppose further, and this is the efficiency wage
hypothesis, that e rises when the real wage, w, rises. Then profits:
f(e(w)L)—wL,
will be maximized when two conditions hold. First, the marginal product of
labour must equal the wage per efficiency unit, w/e(w). Second, the wage must
be set at the point (if there is one!) at which the function e(w) has unit elasticity.35
The second condition fixes the equilibrium wage, call it w*, on purely
technological grounds. Given w*, the first condition then determines optimal
employment, L*, as long as labour supply at w* is at least as great as L*.
An equilibrium with unemployment arises if L* happens to fall below
labour supply at w*. It is a true equilibrium, not just a long-lasting
disequilibrium, because profit-maximizing firms have no interest in reducing
wages. It has involuntary unemployment that persists for an indefinite period
of time because, at wage w*, labour supply exceeds labour demand. All this
sounds very Keynesian. But there is a hitch. Like the monopolistic
competition models, efficiency wage theories are fundamentally models of
relative prices and real wages. They have nothing to say about nominal
magnitudes, and hence allow no role for nominal money, until they are
altered to include fixed costs of changing nominal wages or prices.36 Nor, in
their current state of development, do they have much to say about
fluctuations in employment.
Efficiency wage models do, however, have at least one more Keynesian
aspect that I think important: they focus attention on relative wages.37 Ask
yourself why higher wages enhance productivity. Theorists have provided
many possible answers, but the most plausible for advanced, industrial
nations (where malnutrition is not the issue) is that workers who are paid
well are inclined to perform better for their employers. Such behaviour can
be rationalized if workers care about relative wages, as Keynes believed.
Fixed costs and inertia
A third important recent development in micro theory is the revision of the
standard theory of optimization to include fixed costs of changing a decision
variable. This idea was mentioned earlier in the context of pricing decisions,
where it helps impart inertia to the price level. But it also has obvious
applications to inventory behaviour (where the idea originated), to the
demands for both consumer and producer durables, to the demand for
128 Alan S.Blinder
money, and to portfolio choice more generally.38 Though the mathematics
can get complicated quite quickly, the basic idea is completely intuitive and
easy to grasp. I illustrate it in the case of a consumer durable, but the same
idea applies in other contexts.
Suppose a consumer must pay a fixed transactions cost whenever she
purchases a durable good, such as a car. Then, each time one of the basic
determinants of demand for automobiles (such as income, interest rates, or
relative price) changes, she is faced with the following choice. If she switches
to the new first-best optimal car, she must pay a fixed cost. If she does not,
she suffers an implicit utility cost from having a suboptimal car. Obviously,
it does not pay her to adjust her car purchases continuously, for that would
entail exorbitant transactions costs. Rather, optimal behaviour leads to a
decision rule something like the so-called (S, s) rule of inventory theory:
when the quality of the car deteriorates to some lower bound, s, purchase a
new car of quality S; otherwise, do nothing. The parameters S and s are
chosen optimally in view of transactions costs, income, and other pertinent
information.
Once you think about continuous reoptimization and (S, s) as alternative
models of behaviour, even for a little while, two things become clear. First,
(S, s) is almost certainly more descriptive of the way people and businesses
actually behave; they do nothing for long periods of time and then make
large changes in their behaviour. Second, (S, s) behaviour is almost certainly
a more sensible theoretical model on a priori grounds than continuous
reoptimization.
Obviously, (S, s)-type reasoning provides different microfoundations for
the traditional aggregative behavioural equations (consumption, investment,
money demand, etc.) that constitute macroeconomic models. But what does
it have to do with Keynesian versus new classical economics? Three things,
principally. First, in a world with important fixed costs, optimizing agents
are typically not at neoclassical tangencies where marginal this equals
marginal that. Instead, behaviour displays a substantial amount of inertia. In
the case of price setting, that is a characteristic Keynesian position, as I
noted earlier.39 Second, it leads us to expect to see occasional large
adjustments to what appear to be small economy-wide shocks as a
substantial number of decision-makers trip their (S, s) boundaries
simultaneously. Third, the (S, s) view of the world suggests that a more
volatile economic environment imposes real costs on individuals and
businesses by making them trip their (S, s) barriers more frequently.40
Though I have not seen the argument worked out formally, this would seem
to support the traditional Keynesian advocacy of stabilization policy.
Hysteresis
Finally, it is important to mention the development of modern models of
‘hysteresis’, that is, models in which the economy’s equilibrium state depends
The fall and rise of Keynesian economics 129
on the path we follow to get there, for these bring Keynesian economics back
with a vengeance.
Old-fashioned Keynesian models assumed such hysteresis, without
thinking much about it, and without using the fancy name. For example, the
simplest Keynesian cross model, if taken literally, asserts that equilibrium
can occur at any level of output, if aggregate demand is high enough.
Similarly, the original Phillips curve implied that the economy could achieve
equilibrium at a wide variety of (permanent) unemployment rates, each with
its own unique (permanent) inflation rate. Both of these ideas were swept
away by the natural rate revolution in the late 1960s and early 1970s and
came to be thought of as muddled thinking. Economists, Keynesian or
otherwise, came to believe that the long-run Phillips curve was vertical; no
matter what happened to the economy in the interim, it could come to rest
only at a unique ‘natural’ rate of unemployment determined by
microeconomic factors.
Spurred on, I think, by observing what has happened in Europe, modern
theorists are now constructing models that do not have the natural rate
property. In these super-Keynesian models, expansionary demand
management policies can raise employment permanently. In a neat reversal
of Say’s Law, demand here creates its own supply. Why might this be so?
One simple and obvious mechanism is based on human capital. Suppose
workers who are more experienced are also more productive, perhaps due to
learning by doing on the job, and that, conversely, human capital
deteriorates when not in use. Then a demand-induced boom will build
human capital and hence raise potential GNP for the future; so output can be
permanently raised. Conversely, a recession which idles workers will deplete
the human capital stock and hence lead to lower potential GNP in the future.
There is, then, no ‘natural’ level of employment. The equilibrium level
depends on what came before.
The most popular and best developed hysteresis models nowadays are
based not on human capital, but rather on the conflict between insiders and
outsiders.41 Suppose unions decide on wages, mindful of the fact that higher
wages lead to lower employment but caring only about its members (the
insiders). Suppose further that only employed workers are union’s members;
outsiders have no voice in the union’s decisions. Now let a recession lead to
lay-offs. The union’s membership shrinks and, hence, the union begins to
give more weight to higher wages and less weight to higher employment.
That is, its optimal wage rises and its optimal employment level falls. The
outsiders who lack jobs object to this decision, but have no way to change it;
they are disenfranchized. The lower employment level brought about by the
recession therefore becomes permanent. In the other direction, of course, this
vicious circle becomes a virtuous circle. If a demand-induced boom leads to
new hiring, some outsiders are transformed into insiders and the protected
level of employment rises.42
130 Alan S.Blinder
5 IN CONCLUSION
The empirical evidence does not yet dictate that we adopt these four
theoretical innovations. Most industries seem monopolistically rather than
perfectly competitive; but no one has yet established that the major costs of
price adjustments (and other adjustments) are fixed rather than variable. Nor
is the evidence on the efficiency wage hypothesis overwhelming. And
hysteresis seems to characterize some economies some of the time, not all
economies all the time. But at least these hypotheses have not been refuted
by the data.
If we put all four of these theoretical features together—an act of extreme
chutzpah, to be sure—a thoroughly Keynesian world emerges. Decision
variables, including nominal prices and wages, are inertial. Markets often
equilibrate with excess supply. So, in particular, involuntary unemployment
is common and firms have chronic excess capacity. At least within some
range, the economy’s equilibrium can be changed by demand management
policies because there is no natural rate. Again within some range, welfare
can be improved by expanding aggregate demand and by reducing the
amplitude of cyclical fluctuations.
This world is different in every particular from the world envisioned by
the new classical economics. But its theoretical foundations are no less
strong, and perhaps stronger, which is why Keynesian economics now seems
to be on the ascendancy in academia. More importantly, it sounds more like
the world we live in, which is why some of us find new Keynesian theorizing
so hopeful.
ACKNOWLEDGEMENTS
I am grateful to Ben Bernanke, John Campbell, Stephen Goldfeld, John
Seater, Steven Sheffrin and Robert Trevor for helpful comments on an earlier
draft.
NOTES
1 See Blinder (1979; 1987a: Ch. 3).
2 That does not, however, preclude the possibility that, for example, monetary policy
might cancel the macro effects of fiscal policy by controlling nominal GNP.
3 That belief need not preclude the possibility that, for example, the division of any
given change in nominal GNP into real effects and price effects might depend on
whether or not the change is anticipated.
4 By monetary policy I mean, for example, an increase in base money paid out in
lump-sum transfers. Open-market swaps of money for bonds are often (not always)
non-neutral because they change interest rates. Monetary non-neutrality can also
be rationalized by distribution effects; but these are typically considered unimportant
and are rarely the focus of the debate.
5 See, for example, Barro (1981b).
The fall and rise of Keynesian economics 131
6 For this reason, I have proposed that we rename involuntary unemployment as
‘pornographic unemployment’ (Blinder 1988).
7 See Lucas (1987) for an example. For a rebuttal, see Blinder (1987b).
8 I include myself here (see Blinder 1987a: Ch. 2; 1988).
9 I would not want to place the last of these beyond the realm of positive economics.
There is a huge literature on the social costs of unemployment and inflation, and
many Keynesians like myself have concluded from this evidence that the costs of
low inflation are both small and readily avoidable. None the less, value judgements
are still involved in the trade-off between unemployment and inflation.
10 The one prominent exception was mentioned above: the old debate over whether
or not the LM curve is vertical. This has long been a dead issue.
11 The ‘price’ may be multifaceted. Complicated contractual agreements are allowed
within the new classical approach.
12 I do not intend to join the philosophical debate over whether there is such a thing
as objective scientific evidence.
13 See, for example, Blinder (1987b) and Lovell (1986).
14 One example is the effects of anticipated future changes in policy.
15 It should be noted that some new classicals disagree and see rational expectations
as much more fundamental to the debate.
16 For the rudimentary theory, see Phelps (1978) and Gordon (1975). Phelps’ ideas
on the subject were first presented at an American Enterprise Institute meeting in
April 1974; Gordon’s were first offered at a meeting of the Savings and Loan
Association that same month. Already in January 1974, Princeton graduate
students were being asked (by me!) to analyse supply shocks in Keynesian models
in examinations.
17 See Gordon (1977).
18 Some examples are Ando and Kennickell (1985), B.Friedman (1983), Gordon
(1985) and Perry (1983). There are others.
19 Though Lucas’s paper was published only in 1976, it had been given at a Carnegie-
Rochester conference in April 1973 and was well known in academic circles years
before it was published.
20 Blinder (1979; 1982) traces the relevant history for the US and supports the
statement, which holds even though the world-wide boom of 1972–3 was surely
demand-induced.
21 Thomas Sargent and Lars Hansen led in developing the new econometric methods.
Sargent always referred to it as ‘a technology’.
22 Regretfully, I have no data to support this quantitative assertion. Lucas, Sargent,
Barro and Wallace are, of course, ‘older economists’ in this context. But they were
the founders of the new school of thought.
23 Symmetrically, a conservative might argue that Keynesian ideas could only have
caught on in a milieu (like that of the Great Depression) in which left-wing ideology
was ascendant. Neither statement says anything about the validity of either doctrine.
24 That does not mean that Keynesianism encountered no empirical problems. The
most prominent one was probably the collapse of the money-demand equation in
the US, Canada, and many other countries. While this is commonly, and correctly,
considered a disaster for monetarism, it also poses a serious problem for the
Keynesian LM curve.
25 For a review, see Barro (1981a).
132 Alan S.Blinder
26 The story is even worse than this because money growth did not actually decelerate,
except fleetingly, in either country. But that has to do with financial innovation and
the collapse of the money-demand equation, which is as much a problem for
Keynesian theory as for new classical theory.
27 In the case of the 1973–5 recession, Blinder (1981) points out that ‘unanticipated
money’, as defined empirically by Barro and Rush (1980), does not come close to
explaining the recession. I know of no similar calculation for the 1980s, but it
would also not come close since it was declining velocity growth, not declining
money growth, that made money tight in 1981–2.
28 For a comprehensive review of the evidence, see Brunner (1986).
29 The US corporate tax cuts enacted in 1981 have been suggested as an explanation.
But there is controversy about this. See Blanchard and Summers (1984), Niskanen
(1988) and Bosworth (1985).
30 See, for example, Rotemberg (1984).
31 Lovell (1986) offers a convenient summary of many studies, one of them by Muth
(1985)!
32 Actually, Akerlof and Yellen (1985) appealed to ‘near rationality’ rather than to
fixed costs. But the two amount to essentially the same thing. Also, the costs of
changing prices do not have to be exclusively fixed to make the theory work.
33 See Ball and Romer (1987). However, DeLong and Summers (1988) argue that
Keynesians ought to be more concerned with output being too low and less
concerned with it being too variable.
34 For much greater detail, including applications to markets other than that for
labour, see Greenwald and Stiglitz (1987a; 1987b).
35 The proof is straightforward. Maximizing profits with respect to w gives the first-
order condition f’(eL)e’(w)=1, while maximizing with respect to L gives f’(eL)=w/
e(w). Putting the two together implies we’(w)/e=1.
36 That is what Akerlof and Yellen (1985) do.
37 I have elaborated on this theme at greater length in Blinder (1988).
38 On inventories, see Blinder (1981a). On consumer durables, see Bar-Ilan and Blinder
(1988). On investment, see Dixit (1988). On the demand for money, see Bar-Ilan
(1987). On portfolio choice, see Grossman and Laroque (1987).
39 This need not always be true. Caplin and Spulber (1987) show that price-level
inertia may be absent in certain cases. But these are very special steady state cases.
40 However, the (S, s) range can (and presumably will) be widened if volatility increases.
41 See Lindbeck and Snower (1986), Blanchard and Summers (1986) and Drazen
and Gottfries (1987).
42 Acceptance of this model does not necessarily lead to advocacy of expansionary
demand-management policies. It might lead, instead, to policies that weaken the
power of insiders.
REFERENCES
Akerlof, George and Yellen, Janet L. (1985) ‘Can Small Deviations from Rationality
Make Significant Differences to Economic Equilibria?’, American Economic Review
75, 4, September, 708–20.
Ando, Albert F. and Kennickell, Arthur (1985)‘ “Failure” of Keynesian Economics and
“Direct” Effects of Money Supply: A Fact or a Fiction’, University of Pennsylvania,
mimeo.
The fall and rise of Keynesian economics 133
Ball, Laurence and Romer, David (1987) ‘Are Prices Too Sticky?’ National Bureau of
Economic Research working paper 2171, February.
Bar-Ilan, Avner (1987) ‘Stochastic Analysis of Money Demand Using Impulse Control’,
Dartmouth College, mimeo.
Bar-Ilan, Avner and Blinder, Alan (1988) ‘Consumer Durables and the Optimality of
Usually Doing Nothing’, January, mimeo.
Barro, Robert J. (1981a) ‘The Equilibrium Approach to Business Cycles’, in his Money,
Expectations, and Business Cycles, Academic Press, New York, 41–78.
——(1981b) ‘Output Effects of Government Purchases’, Journal of Political Economy
89, 6, December, 1086–121.
——(1974) ‘Are Government Bonds Net Wealth?’, Journal of Political Economy 82,
November/December, 1095–117.
Barro, Robert J. and Mark Rush (1980) ‘Unanticipated Money and Economic Activity’,
in Stanley Fischer (ed.) Rational Expectations and Economic Policy, University of
Chicago Press, Chicago, 23–48.
Blanchard, Olivier and Kiyotaki, Nobuhiro (1987) ‘Monopolistic Competition and
the Effects of Aggregate Demand’, American Economic Review 77, 4, September,
647–66.
Blanchard, Olivier and Summers, Lawrence (1984) ‘Perspectives on High World Real
Interest Rates’, Brookings Papers on Economic Activity 2, 273–324.
——(1986) ‘Hysteresis and the European Unemployment Problem’, NBER
Macroeconomics Annual 1986, 15–78.
Blinder, Alan S. (1988) ‘The Challenge of High Unemployment’, American Economic
Review 78, May, 1–15.
——(1987a) Hard Heads, Soft Hearts: Tough-Minded Economics for a Just Society,
Addison-Wesley.
——(1987b) ‘Keynes, Lucas and Scientific Progress’, American Economic Review 78,
May, 130–6.
——(1982) ‘The Anatomy of Double Digit Inflation in the 1970s’, in R.E.Hall (ed.)
Inflation: Causes and Effects, University of Chicago Press, Chicago, 261–82.
——(1981a) ‘Retail Inventory Behaviour and Business Fluctuation’, Brookings Papers
on Economic Activity 2, 443–505.
——(1981b) ‘Supply-Shock Inflation: Money, Expectations, and Accommodation’, in
M.J.Flanders and A.Razin (eds) Development in an Inflationary World, Academic
Press, 61–101.
——(1979) Economic Policy and the Great Stagflation, Academic Press.
Bosworth, Barry P. (1985) ‘Taxes and the Investment Recovery’, Brookings Papers on
Economic Activity 1, 1–45.
Brunner, Karl (1986) ‘Fiscal Policy in Macro Theory: A Survey and Evaluation’, in
R.W.Hafer (ed.) The Monetary versus Fiscal Policy Debate, Rowman & Allanheld,
33–116.
Caplin, Andrew and Spulber, D. (1987) ‘Menu Costs and the Neutrality of Money’,
Quarterly Journal of Economics, November.
DeLong, Bradford and Summers, Lawrence (1988) paper in preparation for Brookings
Panel, September.
Dixit, Avinash (1988) ‘Entry and Exit Decisions Under Uncertainty’, Princeton University,
May, mimeo.
Drazen, Allan and Gottfries, Nils (1987) ‘Seniority Rules and the Persistence of
Unemployment in a Dynamic Optimizing Model’, mimeo.
Fischer, Stanley (1977) ‘Long-Term Contracts, Rational Expectations, and the Optimal
Money Supply Rule’, Journal of Political Economy 85, 191–205.
Friedman, Benjamin (1983) ‘Recent Perspectives In and On Macroeconomics’, National
Bureau of Economic Research, working paper 1208.
134 Alan S.Blinder
Friedman, Milton (1968) ‘The Role of Monetary Policy’, American Economic Review
58, 1–17.
Gordon, Robert J. (1987) ‘Wage Gaps vs. Output Gaps: Is There A Common Story for
All of Europe?’, National Bureau of Economic Research working paper 2454,
December.
——(1985) ‘Understanding Inflation in the 1980s’, Brookings Papers on Economic
Activity, 263–99.
Gordon, Robert J. (1977) ‘Can the Inflation of the 1970s Be Explained?’, Brookings
Papers on Economic Activity, 253–77.
——(1975) ‘Alternative Responses of Policy to External Supply Shocks’, Brookings
Papers on Economic Activity 1, 183–206.
——(1972) ‘Wage-Price Controls and the Shifting Phillips Curve’, Brookings Papers
on Economic Activity, 385–421.
Greenwald, Bruce and Joseph E.Stiglitz (1987a) ‘Keynesian, New Keynesian and New
Classical Economies’, Oxford Economic Papers 39, 119–32.
——(1987b) ‘Imperfect Information, Credit Markets and Unemployment’, European
Economic Review 31, 444–56.
Grossman, Sanford and Laroque, Guy (1987) ‘The Demand for Durables and Portfolio
Choice under Uncertainty’, Princeton University, June, mimeo.
Kuhn, Thomas S. (1962) The Structure of Scientific Revolutions, University of Chicago
Press, Chicago.
Lindbeck, Assar and Snower, Dennis (1986) ‘Wage Setting, Unemployment, and Insider-
Outsider Relations’, American Economic Review 76, May, 235–39.
Lovell, Michael E. (1986) ‘Tests of the Rational Expectations Hypothesis’, American
Economic Review 76, 1, March, 110–24.
Lucas, Robert E. Jr (1987) Models of Business Cycles, Basil Blackwell.
——(1981) ‘Tobin and Monetarism: A Review Article’, Journal of Economic Literature
XIX, June, 558–67.
——(1976) ‘Econometric Policy Evaluation: A Critique’, in Karl Brunner and Allan
H.Meltzer (eds) The Phillips Curve and Labor Markets, Carnegie-Rochester
Conferences on Public Policy, vol. 1, North-Holland Publishing Company,
Amsterdam.
Lucas, Robert E. Jr and Sargent, Thomas (1978) ‘After Keynesian Macroeconomics’,
in After the Phillips Curve; Persistence of High Inflation and High Unemployment,
Conference Series 19, Federal Reserve Bank of Boston.
Mankiw, N.Gregory (1985) ‘Small Menu Costs and Large Business Cycles: A
Macroeconomic Model of Monopoly’, Quarterly Journal of Economics 100, May,
529–39.
Muth, John F. (1985) ‘Short Run Forecasts of Business Activity’, Indiana University,
March, mimeo.
Niskanen, William A. (1988) Reaganomics: An Insider’s Account of the Policies and
the People, Oxford University Press, Oxford.
Perry, George L. (1983) ‘What Have We Learned About Disinflation?’, Brookings
Papers on Economic Activity, 587–602.
Phelps, Edmund S. (1978) ‘Commodity-Supply Shock and Full-Employment Monetary
Policy’, Journal of Money, Credit and Banking X, 2, May, 206–21.
——(1968) ‘Money-Wage Dynamics and Labor Market Equilibrium’, Journal of Political
Economy 78, 678–711.
Prescott, Edward (1986) ‘Theory Ahead of Business Cycle Measurement’, Federal
Reserve Bank of Minneapolis Research Department Staff Report 102, February.
Rotemberg, Julio J. (1984) ‘Interpreting Some Statistical Failures of Some Rational
Expectations Macroeconomic Models’, American Economic Review 74, 188–93.
Taylor, John B. (1980) ‘Aggregate Dynamics and Staggered Contracts’, Journal of
Political Economy 88, 1–23.
6 Price flexibility and output stability
An old Keynesian view
James Tobin
Journal of Economic Perspectives (1993) 7, Winter, pp. 45–65
In this symposium I shall play the role in which I was cast, the
unreconstructed old Keynesian. Time was when I resisted labels and schools,
naively hoping that our fledgling science was outgrowing them. I had, to be
sure, been drawn into economics when the General Theory was an exciting
revelation for students hungry for explanation and remedy of the Great
Depression. At the same time, I was uncomfortable with several aspects of
Keynes’ theory, and I sought to improve what would now be called the
microfoundations of his macroeconomic relations.
The synthesis of neoclassical and Keynesian analysis achieved in the
1950s and 1960s promised a reconciliation of the two traditions, or at least
an understanding of the different contexts to which each applies. The hope
and the promise were premature, to say the least. Since 1973 the dominant
trend in macroeconomics has dismissed Keynesian theory. Nevertheless,
Keynesian models continue to prove useful in empirical applications,
forecasting and policy analysis. Macro-econometric models are mostly built
on Keynesian frameworks. The gulfs between doctrine and observation,
between theory and practice, are chronic sources of malaise in our discipline.
I have benefitted from Gregory Mankiw’s ‘refresher course’ in modern
macroeconomics (1990). He writes that recent developments—
methodological, new classical, and new Keynesian—are to old
macroeconomics as Copernicus was to Ptolemy. It just takes time before
Copernican truths can outdo Ptolemaic approximations in practical
applications.
Considering the alternatives, I do not mind being billed as a Keynesian,
an old Keynesian at that. But old Keynesians come in several varieties, and
I speak for no one but myself. Nor do I defend the literal text of the General
Theory. Several generations of economists have criticized, amended, and
elaborated that seminal work. I shall argue for the validity of the major
propositions that distinguish Keynesian macroeconomics from old or new
classical macroeconomics.
SUMMARY OF THE KEYNESIAN CASE
The central proposition of Keynesian economics is commonly described as
follows: ‘According to the Keynesian view, fluctuations in output arise
136 James Tobin
largely from fluctuations in nominal aggregate demand. These fluctuations
have real effects because nominal wages and prices are rigid’ (Ball et al.
1988:1). On the contrary, I shall argue that Keynesian macroeconomics
neither asserts nor requires nominal wage and/or price rigidity. It does assert
and require that markets not be instantaneously and continuously cleared by
prices. That is a much less restrictive assumption, and much less
controversial. It leaves plenty of room for flexibility in any commonsense
meaning of the word.
Keynesian models were said to be vulnerable to the charge that ‘the
crucial nominal rigidities were assumed rather than explained’, although ‘it
was clearly in the interests of agents to eliminate the rigidities they were
assumed to create…. Thus the 1970s and 1980s saw many economists turn
away from Keynesian theories and toward new classical models with flexible
wages and prices’ (Ball et al. 1988:2). Those market-clearing models have
not just flexible prices but perfectly and instantaneously flexible prices, an
assumption that is surely more extreme, more arbitrary, and more devoid of
foundations in individual rational behavior than the imperfect flexibility of
Keynesian models.
The central Keynesian proposition is not nominal price rigidity but the
principle of effective demand (Keynes 1936: ch. 3). In the absence of
instantaneous and complete market clearing, output and employment are
frequently constrained by aggregate demand. In these excess-supply regimes,
agents’ demands are limited by their inability to sell as much as they would
like at prevailing prices. Any failure of price adjustments to keep markets
cleared opens the door for quantities to determine quantities, for example
real national income to determine consumption demand, as described in
Keynes’ multiplier calculus.
For this reason, Keynesian macroeconomics alleges that capitalist
societies are vulnerable to very costly economy-wide market failures.
Individuals would be willing to supply more labor and other resources in
return for the goods and services the employment of those resources would
enable them to consume now or in the future, but they cannot implement this
willingness in market transactions. As the quotation from Ball et al. (1988)
suggests, many contemporary theorists cannot believe any theory that
implies socially irrational market failures. They suspect that individual
irrationalities are lurking somewhere in the theory. In continuously price-
cleared competitive markets, they know, individually rational behaviour
implies collectively rational outcomes. But this theorem does not apply if
markets and price-setting institutions do not produce perfectly flexible
competitive prices. Individual rationality does not necessarily create the
institutions that would guarantee ‘invisible hand’ results. Keynes was not
questioning the rationality of individual economic agents; he was arguing
that their behavior would yield optimal results if and only if they as citizens
organized the necessary collective institutions and government policies. In
the same spirit though in different contexts, some modern theoretical
Price flexibility and output stability 137
research has shown that welfare-improving policies may be designed even
when asymmetries of information and incompleteness of markets prevent the
achievement of global optima.
Ball, Mankiw, Romer and others style themselves as New Keynesians.
Their program is to develop improved microeconomic foundations for
imperfectly flexible prices. In the process, they hope to illuminate the
paradox that individually rational or near-rational behavior can result in
significant collective market failures. These are certainly laudable
objectives. In the end, I suspect, the program will not change the essential
substance of Keynesian macroeconomics. But it will make Keynes more
palatable to theorists.
In Keynesian business cycle theory, the shocks generating fluctuations are
generally shifts in real aggregate demand for goods and services, notably in
capital investment. Keynes would be appalled to see his cycle model
described as one in which ‘fluctuations in output arise largely from
fluctuations in nominal aggregate demand’ (Ball et al. 1988:2). The
difference is important. The impact on real purchases of a one-time 1 percent
shock to aggregate nominal spending will be eroded if and as nominal prices
increase in response, and eliminated once prices have risen by the same 1
percent as nominal spending did. But suppose it is real demand that initially
rises 1 per cent. At the prevailing prices nominal spending will rise 1 percent
too. But if and as prices rise in response the 1 percent real demand shock
becomes an ever larger amount of nominal spending. Its impact is not
mechanically eroded by the price response; if it is absorbed, the process is
subtle and indirect.
The big issue between Keynes and his ‘old classical’ opponents was the
efficacy of the economy’s natural market adjustment mechanisms in restoring
full employment equilibrium, once a negative real demand shock had pushed
the economy off that equilibrium. Keynes and Keynesians said those
mechanisms were weak, possibly nonexistent or perverse, and needed help
from government policy. That is still the major question of macroeconomic
theory and policy, even though new classical economists finesse it by
assuming that the economy can never be pushed out of equilibrium even for
a moment. Keynes’ classical contemporaries and predecessors would never
have drawn real-world lessons from theories based on such an assumption.
Their successors strain credulity when their models imply that markets are
cleared and joblessness is voluntary when measured unemployment is 10
percent as truly as when it is 5 percent.
Keynesian theory of nominal wage stickiness does not deserve the disdain
with which it is commonly regarded. It is not dependent on ‘money illusion’.
But Keynes certainly would have done better to assume imperfect or
monopolistic competition throughout the economy, in both product and labor
markets. In markets of these kinds, nominal prices are decision variables for
sellers or buyers or are determined by negotiations between them. They
therefore move only at discrete intervals. Despite considerable effort over the
138 James Tobin
years to give macroeconomics improved microfoundations along these lines,
there is plenty of scope for the ‘New Keynesian’ program of theoretical and
empirical research on this topic.
In the absence of perfect flexibility, does greater flexibility of nominal
prices strengthen the equilibrating mechanisms, or does it weaken them?
Keynes doubted that the problems of involuntary unemployment and
underutilized capacity would be mitigated by greater flexibility of nominal
wages and prices. On the whole, he favored stable nominal wages. Critics of
Keynesian macroeconomics forget this strand of the argument when they
assume that without absolute ‘rigidity’ aggregate demand could never be
deficient. Fortunately, this issue has been receiving greater attention in the
last few years, with considerable support for Keynes’ position.
MACROECONOMICS WITH EFFECTIVE DEMAND CONSTRAINED
The empirical relevance of Keynesian economics is based on its assertion
that situations of pervasive excess supply often occur. An advanced capitalist
industrial economy is frequently in a state in which most labor and product
markets are not clearing at prevailing prices. As a result, workers are
involuntarily unemployed and capital capacity is underutilized. The effective
constraint on output is the aggregate demand for goods and services;
likewise the effective constraint on employment is the amount of labor
required to produce that output.
Keynesian unemployment must be differentiated from both frictional and
classical unemployment. Frictional unemployment occurs because of
microeconomic flux. Demands and supplies are continually shifting, bringing
unemployment and excess capacity in some sectors and contemporaneous
labor shortages and capacity bottlenecks elsewhere. The gross aggregates of
these frictional excess supplies and excess demands vary together positively
over time. In contrast, cyclical excess supplies and demands are negatively
correlated; in economy-wide recessions and depressions, excess-supply
markets and sectors predominate, while the reverse is true in inflationary
booms. The amount of frictional unemployment depends on the strength of
intersectoral shocks and on the mobility of factors of production in
responding to them. Large and protracted shocks, for example in technology
or in supplies and prices of key commodities like energy, convert frictional
unemployment to structural unemployment. Neither is remediable by
demand expansion alone.
A common species of classical unemployment occurs when jobs are
limited because of excessive real wage rates imposed by governmental or
trade union regulations. For individuals who would like to work at or below
the wage floor, such unemployment is involuntary. For the workers
collectively whose bargaining strength or political clout established the
regulations, the unemployment could be regarded as the voluntary
consequence of their exercise of monopoly power.
Price flexibility and output stability 139
Identification of observed unemployment as classical or Keynesian is
sometimes difficult. In either case unemployment might be observed to be
associated with real wages above their full employment equilibrium values.
In the Keynesian case, this could result from perfect competition among
producing firms; they would be paying workers the high marginal products
associated with low employment. The big difference between the two cases is
that in the Keynesian case, but not in the classical case, real wages would
decline on their own and output and employment would increase in response
to expanded demand. In the classical case removal of the regulations would
be essential.
There are several variations on the classical unemployment theme. One
case is queuing for a high-wage job. An artificially high wage in a particular
sector could draw workers from employment elsewhere to wait and hope.
This model was originally designed to explain the heavy unemployment in
the urban centers of developing countries, where the queuing requires living
near the scarce jobs, far from alternative means of subsistence in traditional
agriculture. It fits less well in advanced economies, where workers can
search and apply for better jobs while employed. Another source of
voluntary unemployment may be unemployment insurance benefits and other
transfers that increase the reservation prices of persons without jobs.
However, in the United States, where unemployment is measured by large
household surveys conducted monthly by the Census, persons without jobs
will be counted not as unemployed but as ‘not in labor force’ unless they
report they have been actively searching. Although some misreporting
doubtless occurs, it is small, not always in the same direction, and cannot
begin to account for the cyclical variability of unemployment rates.
Agents who are unable to sell as much as they would like at prevailing
prices restrict demands in other markets. Unemployed workers cut their
consumption. Demand-constrained firms restrict their hiring of labor and
their purchases of other inputs. Keynes’ insight that quantities actually sold,
if smaller than sales desired at existing prices, will keep demands in other
markets below equilibrium values, was rediscovered and elaborated by self-
styled ‘disequilibrium theorists’ 30 years later (Barro and Grossman 1971).
In old Keynesian economics, multiplier theory formalized the determination
of quantities by quantities. It did not and does not, however, preclude the
relevance of other determinants of demand, notably prices and interest rates.
In this respect it is more general than most of its latter-day extensions in
‘disequilibrium theory’. In demand-constrained regimes, any agent’s increase
in demand—for example, more investment spending by a business firm—has
positive externalities. It will increase the attainable consumption of third
parties. In some modern literature, this idea of Keynes is revived and
elaborated under the label ‘strategic complementarity’ (Cooper and John
1988).
Liquidity constraints are an important but extreme form of effective
demand constraint. Some wage earners, no doubt, depend on each week’s
140 James Tobin
wages to buy the goods for that week’s consumption. But Keynes’ principle
does not depend on such short horizons for consumption-smoothing.
Expectations of future spells of unemployment, enhanced by present and
recent experience, can limit the current consumption and durables purchases
even of long-horizon households. Liquidity constraints and prospective
effective demand constraints can also limit business investment. Common
observation suggests that households and businesses, and governments too,
differ widely in their horizons, i.e. the length of the future period over which
expected resources are regarded as potentially available for spending today.
These horizons, moreover, doubtless change over time with circumstances
and behavior.
The multipliers relating change in aggregate demand to demand shocks,
from policies or other events, are not as large as they were thought to be
when the concept was first introduced and estimated in the 1930s. One
reason is a substantial structural change in democratic capitalist economies.
Governments are much larger relative to private sectors than before World
War II, and their fiscal institutions are ‘built-in stabilizers’. Their
expenditures are quite unresponsive to current business conditions, while
their revenues (net of transfers to the private sector) are cyclically sensitive
and thus moderate swings in private incomes. A second reason is that
economists have come to recognize that, thanks to accommodating capital
markets as well as to their own foresight, most economic agents have
horizons longer than one year.
While this consideration implies that multipliers of transient shocks are
lower than for permanent changes, it by no means implies that they are zero.
Both consumption and investment appear to be sensitive to contemporaneous
and recent incomes. For most agents capital markets are far from perfect; in
particular future and current labor incomes are not fungible. Moreover,
expectations of economic futures, individual, national, and global, are
influenced by current events, perhaps to an irrational extent.
As Keynes explicitly observed, his theory refers to economies with
incomplete markets. In his day futures markets were rare, and contingent
futures markets even rarer. They are still scarce. As Keynes explained,
decisions not to spend now are not coupled with any definite orders for future
or contingent deliveries. Typically they result in accumulations of assets that
can be spent on anything at any future time. The multiplier effects of lower
current spending propensities are not offset by specific and firm expectations
of higher future demands.
BUSINESS CYCLES AS DEMAND FLUCTUATIONS
According to Keynesian macroeconomics, business cycles are fluctuations in
aggregate effective demand, carrying output and employment in their wake.
They do not reflect movements in market-clearing supply-equals-demand
equilibria.
Price flexibility and output stability 141
Supplies of labor and other factors of production move fairly smoothly
from year to year and from cycle to cycle. So does economy-wide factor
productivity, largely reflecting technological progress. Equilibrium output
and employment cannot be as variable as actual cyclical observations. In the
neoclassical neo-Keynesian synthesis, trend growth is supply-determined;
markets are cleared; supply truly creates its own demand. In cyclical
departures from trend, demand evokes its own supply. Keynesian short-run
macroeconomics does not pretend to apply to problems of longrun growth
and development.
Equilibrium cycle theories (Plosser 1989) are unconvincing. They rely on
incredible volatility in technology, retrogressive as well as progressive. They
rely on extreme intertemporal substitutions among work, leisure, and
consumption. Or they contrive informational asymmetries and misperceptions
that seem easy to correct. For example, a few years ago a popular theory
attributed business cycles to confusions by suppliers of products and labor
between increases in their own real prices, on the one hand, and economy-
wide inflation, on the other. Evidently businesses and households were
assumed to ignore the flood of current statistics on prices and money supplies.
I am using the word equilibrium to mean Walrasian market-clearing by
prices, as is the current usage of both new classical macroeconomists and
disequilibrium theorists. Keynes used it otherwise, to refer to a position of
rest. That is why he referred to outcomes with involuntary unemployment as
equilibria on a par with full employment, and why he termed his theory
‘general’ in the title of his book. The basic issue is not semantic. It is
whether situations of general excess supply can and do exist for significant
periods of time, whether or not they are called equilibria.
Some passages of the General Theory can be read to assert that
involuntary unemployment is much more than a temporary cyclical
phenomenon, that it is in the absence of remedial policies a chronic defect of
capitalism. This was a natural enough view in the 1930s. In Alvin Hansen’s
American Keynesianism (e.g. Hansen 1938) secular stagnation was a central
proposition. Formally, however, the analysis of the General Theory is limited
to a time period short enough that the changes in capital stock resulting from
non-zero investment can be ignored.
Postwar Keynesians, for the most part, have not regarded protracted
depression as a likely outcome.1 Chronic inflationary gaps could also occur,
and alternations between excess-supply and excess-demand regimes were
highly probable. Keynesian macroeconomics is two-sided. Deviations on
both sides of Walrasian market-clearing can occur, though not necessarily
with symmetrical symptoms. Excess demand in aggregate is mainly an
‘inflationary gap’, generating unfilled orders and repressed or open inflation,
rather than significant extra output and employment. Macroeconomic
stabilization requires two-sided countercyclical demand management.
In any case, habitual application of Keynesian remedies reinforces
whatever natural mechanisms tend to return the economy to its full
142 James Tobin
employment growth path. Expectations that those remedies will be used
contribute to the stability of that equilibrium path.
THE EFFICACY OF CLASSICAL ADJUSTMENT MECHANISMS:
INTEREST RATES
Suppose that shocks to current real demands for goods and services create, at
existing prices and wages, excess supplies of labor and capital services.
What are the variables whose changes would avert or eliminate
macroeconomic disequilibrium? The leading candidates are current prices,
which include both wages of labor as well as prices of products, and interest
rates, which involve future as well as current prices. In what follows, I shall
set forth Keynesian skepticism regarding the efficacy of these classical
adjustment mechanisms.
If these mechanisms respond instantaneously to shocks, no actual
discrepancy between demand and supply will occur or be observed. The
shocks will be wholly absorbed in the market-clearing variables. This is the
assumption of equilibrium business cycle theory and of the ‘real business
cycles’ approach. It is this assumption that, among other things, enables new
classical macroeconomists to dismiss out of hand real aggregate demand
shocks and to react with incredulity when Keynesians mention them.
However, if these adjustments do not occur instantaneously but take real
time, then Keynesian situations of excess supply do occur. They occur even if
prices and interest rates are falling at the same time. The consequence is that
the quantity adjustments of the multiplier process start working counter to
the possible equilibrating effects of interest rate and price reductions.
In standard Walrasian/Arrow-Debreu theory, perfect flexibility of all
wages and prices, present and future, would maintain full employment
equilibrium. Short of that, an old question of macroeconomic theory is
whether, given current nominal wages and prices, changes in future
money wages and prices—that is, in nominal interest rates—could do
the job.
In old classical macroeconomics, interest rates are the equilibrators of
both capital markets and goods markets. Their adjustment is crucial to the
Say’s Law story, which dismisses as vulgar superficiality notions that an
economy could suffer from shortfalls in demand for commodities in
aggregate. Market interest rates keep investment equal to saving at their full-
employment levels—and therefore keep aggregate demand equal to full
employment output—even if nominal product prices and wages stay put.
Indeed classical doctrine is that the real equilibrium of the economy is
independent of nominal prices, as if it were the outcome of moneyless
frictionless multilateral Walrasian barter.2
Can interest rates do the job? The Keynesian insight is that the
institutionally fixed nominal interest rate on currency, generally zero, limits
the adjustment of nominal interest rates on non-money assets and imparts to
Price flexibility and output stability 143
them some stickiness even when they are above zero. As a result, after an
aggregate demand shock they may not fall automatically to levels low
enough to induce sufficient investment to absorb full employment saving. As
a result, aggregate demand—consumption plus investment—will fall short of
full employment supply.
The case for significant non-zero interest elasticity of money demand is
simply that the opportunity costs of holding money fall as the interest rates
available on non-money substitutes decline. As those rates approach the
interest paid on money itself, zero at the lowest, the opportunity costs vanish.
The interest rate on money sets the floor for other nominal market interest
rates. The familiar specific money demand models—transactions costs, risk
aversion, regressive interest rate expectations—all depend on the fixed
nominal interest floor.
The interest-elasticity of money demand is a key parameter in
macroeconomic theory. Three cases can be distinguished. One is a classical
extreme, often associated with the quantity theory of money: the elasticity is
zero. At the other extreme is the Keynesian liquidity trap: market interest
rates are so close to the floor that people are on the margin indifferent
between money and other assets. In between is the vast middle ground,
where the interest-elasticity of money demand is somewhere between zero
and negative infinity. Undergraduate students of macroeconomics know, or
used to know, that in standard models monetary policy can effectively alter
spending in the classical and intermediate cases but not at the liquidity trap
extreme. They also know, or used to know, that fiscal policy is effective in
the liquidity trap and intermediate cases but not at the classical, monetarist
extreme.
My focus here is somewhat different. The question is the efficacy of
market interest rates as automatic stabilizers in the face of real demand
shocks, when monetary quantities, fiscal parameters, and other policy
instruments are fixed. The answer is not in dispute for the two extremes: they
work in the classical case and not in the liquidity trap. Who owns the middle
ground? Quantity theorists used to contend that classical propositions obtain
everywhere outside the liquidity trap. But the middle ground belongs to the
Keynesians. Real demand shocks will move aggregate income despite their
effects on interest rates, for the same reason that fiscal policies will do so.
Unless the real supply of money is increased by monetary policy or by price
reduction, the interest rate will not fall enough after a negative aggregate
demand shock (the same thing as a negative investment-minus-saving shock)
to maintain investment-equals-saving equality at full employment. The
interest rate that would do that job would also require additional money
supply—unless money demand is perfectly inelastic with respect to market
interest rates.
Recent structural changes have made the monetary system more
monetarist, more like what the quantity theorists said it always was. Bank
deposit interest rates, even on the checkable deposits used for most
144 James Tobin
transactions, now are market-determined and move up and down along with
rates on non-money assets. The differential between them, the opportunity
cost important in cash management, is less systematically related to the
general level of interest rates than it used to be. This development has
undoubtedly made the demand for deposits less elastic with respect to the
interest rates that matter for demands for goods and services (on these
developments see Tobin 1983).
However, the zero floor on nominal interest rates is still there. The
monetary base, currency held outside banks plus bank reserves, remains
interest-free. The money market in which the demand for and supply of bank
reserves are equated is the fulcrum of the banking system and of the entire
structure of interest rates. States of nature in which equilibrium would
require negative real interest rates still have positive probability. Since
nominal rates cannot be negative, full employment would not be possible in
those contingencies unless expected inflation made real rates negative. The
possibility of these states will influence the portfolio and investment decisions
of rational agents.
Money demand is not the whole story. Keynes also stressed liquidity
preference in a different form, sticky long-term interest rates. Because
traditional expectations of future long rates persist in slumps, current long
rates do not automatically follow short rates down far enough to induce the
spurts in investment needed for recovery.
Classical and new classical theories assert that capital markets generate
equilibrium real rates independently of what is happening to nominal interest
rates and commodity prices. But the evidence is that nominal interest rates
do matter. Changes in them are usually changes in real rates. Likewise
changes in inflation expectations are not fully offset by changes in nominal
rates. The ‘Fisher equation’ asserts that real interest rates are independent of
nominal rates and inflation expectations, but Irving Fisher himself concluded
from his empirical investigations that the proposition held if at all only in
very long runs. Modern research has confirmed his findings.
THE EFFICACY OF CLASSICAL ADJUSTMENT MECHANISMS:
NOMINAL WAGES AND PRICES
If interest rate adjustments cannot suffice, no matter how rapidly asset
markets clear, the job falls to nominal prices. If it is a crime not to accept
the instantaneous clearing by prices of product and labor markets as the
foundation of macroeconomics, then Keynes and Keynesians are certainly
guilty. But it is a caricature of Keynesian economics, no less false because it
is widely believed, to attribute to Keynesians the assumption that nominal
prices are perfectly rigid, for the entire time period over which the analysis is
intended to apply. In fact Keynes himself did not contend that nominal prices
and product prices are fixed independently of amounts of excess supply or
demand, and neither do most Keynesians today.
Price flexibility and output stability 145
The ‘fixprice’ method used in many textbooks was a convenient device
for expounding the Keynesian calculus of adjustments of quantities to
quantities and to interest rates. It was carried to extreme in modern
formal ‘disequilibrium theory’. The method is misleading when it conveys
the impression that Keynesian economics assumes price rigidities and
indeed is defined by that assumption. It is especially misleading if it gives
the idea that such an assumption is necessary. This impression of
Keynesian theory, whether the result of caricatures by its enemies or
careless expositions by its friends, appears to be the source of the
defection of many economists.
Consider a spectrum of the degree of nominal price flexibility from
complete flexibility at one extreme to complete rigidity at the other.
Complete flexibility means instantaneous adjustment, so that prices are
always clearing markets, jumping sufficiently to absorb all demand or
supply shocks. Complete rigidity means that nominal prices do not change
at all during the period of analysis. In between are various speeds of price
adjustment, various lengths of time during which markets are not clearing.
Here again, as in the case of interest rate effects and despite common
beliefs to the contrary, Keynesians own the middle ground. It is not true
that only the arbitrary and gratuitous assumption of complete rigidity
converts nominal demand shocks into real demand shocks and brings
multipliers and IS-LM processes into play. Any degree of stickiness that
prevents complete price adjustment at once has the same qualitative
implications, and can even be treated by the fixprice method on an ‘as if
basis.
Keynes argued that nominal wages would not fall rapidly in response to
excess supplies of labor. At the same time, he asserted that real wages could
fall if product prices rose as necessary to induce firms to expand
employment. This asymmetry led many critics to suppose that Keynes was
attributing ‘money illusion’ to workers and to dismiss Keynesian theory out
of hand. Why would workers accept a cut in real wages achieved by an
increase in the price of wage goods but resist cuts in money wages? Keynes’
reason for this asymmetry is both empirically realistic and theoretically
impeccable. Workers are concerned primarily with relative wages, with how
their pay compares with the pay of those to whom they regard themselves at
least equal in merit. Those concerns do not depend on money illusion, they
are certainly not irrational, and there is a great deal of empirical evidence of
their importance.
Labor markets are disaggregated and desynchronized. To any single
worker or local group, a nominal wage cut appears to be a loss in relative
wages; there is no assurance that others will also take cuts. On the other
hand, an increase in the cost of living is the same for everybody. Workers
may be perfectly prepared to receive lower real wages with unchanged
relative wages, but labor market institutions give them no way to
communicate this willingness.
146 James Tobin
The hole in this story is that it does not explain how the relative-wage
concerns of employed workers prevail when there are unemployed workers
willing to work for less pay—real, nominal, and relative. The power of
insiders vis-à-vis employers and outsiders evidently derives from the costs of
turnover among members of an interdependent working team. Insider power
has lately been the subject of considerable theoretical and empirical inquiry,
notably by Assar Lindbeck and his colleagues (Lindbeck and Snower 1990).
Labor economists have long observed that queues of job-seekers outside the
factory gate have little effect on the wages paid to employees inside. Hard
times do bring wage cuts, but usually by so damaging the financial and
competitive positions of employers that they can credibly threaten layoffs of
senior workers and even plant closings and bankruptcies.
All Keynesian macroeconomics really requires is that product prices and
money wages are not perfectly flexible, whatever may be the rationale for
their behavior. After all, the Walrasian auctioneer of classical
macroeconomics is itself not an implication of optimizing behavior. It is a
fictitious institution with no presumptive priority over alternative
institutional assumptions.
Seeking to win the game on his opponents’ home field, Keynes pretended
to be assuming pure competition in all markets. But his insights regarding
labor markets implicitly recognized that wages are administered or
negotiated prices, and for that reason alone are not perfectly flexible, not
prices set in impersonal auction markets. His product markets, however,
remained Marshallian. Given money wages and given the overall aggregate
demand constraint, competition equated product prices to marginal cost.
Thus real wages were equal to marginal productivity. But, as the existence of
excess supply would imply, those wages exceeded the wages necessary to
induce workers to supply the actual volume of employment.
Marginal productivity theory implies that real wages and employment or
output would be negatively correlated in business cycles. But this implication
has been repeatedly refuted by empirical observations. This is not a blow to
Keynesian policy recommendations, quite the contrary. If it is possible to
expand demand and increase output and employment without lowering real
wages, so much the better—there is less reason to worry that observed
unemployment may be classical.
Clearly product markets, as well as labor markets, should be modeled as
imperfectly competitive. There too prices are decision variables, a fact that
at the very least suggests that they don’t change every hour. When the
economy is in a Keynesian excess-supply regime, dynamics of adjustment
determine the paths of wages, markups, and product prices. The path of real
wages lies between the classical labor demand the supply curves, and could
be either pro-cyclical or counter-cyclical. Likewise, the paths of output and
employment typically diverge from production functions. In the past 50 years
a great deal of empirical work has been done on these relationships. Phillips
curves and Okun’s law are among the best known examples.
Price flexibility and output stability 147
In addition, more formal models of nominal price inertia have been
developed. Arthur Okun (1981) provided a theory of ‘invisible handshakes’,
in which price adjustments are moderated in the interest of maintaining long-
run customer-supplier relationships. Stanley Fischer (1977) and John Taylor
(1980) formalized wage stickiness in models of overlapping staggered
contracts. These models can apply even to non-union shops where wages are
administered rather than negotiated; employers with large work forces
change announced wage scales periodically. In a monograph that has
attracted too little attention, Katsuhito Iwai (1981) gave Keynesian
macroeconomics rigorous microfoundations in a model of monopolistic
competition. A microeconomic world of imperfect competition is a
Keynesian macroeconomic world, where nominal prices are imperfectly
flexible.
Keynes’ explanation of money-wage stickiness is the usual focus of
discussion and criticism. It is to the second strand of his argument,
commonly ignored, that I wish to direct major attention. Even if money
wages and prices were more flexible, even if excess supplies of labor were to
lead more rapidly to cuts in money wages, this greater flexibility would not
prevent or cure unemployment. Given a contractionary shock in aggregate
demand, deflation of money wages and prices would not restore real demand
to its full employment value. This classical market-clearing adjustment
mechanism was, in Keynes’ view, much too frail to bear the weight of
macroeconomic stabilization. In fact, Keynes recommended stability rather
than flexibility in money wages.
Keynes did not challenge the efficacy of price adjustment mechanisms in
clearing particular markets in the Marshallian partial equilibrium theory on
which he had been reared. He did challenge the mindless application of
those mechanisms to economy-wide markets. Founding what came to be
known as macroeconomics, he was modeling a whole economy as a closed
system. He knew he could not use the Marshallian assumption that the
clearing of one market could be safely described on the assumption that the
rest of the economy was unaffected.
Consider the difference between a local market for a particular kind of
worker and the national market for all labor. Excess supply in the local
printing trades, for example, would in a competitive market cause printers’
wages to fall. Declining nominal wages would be declining real wages; both
would be falling relative to the rest of the economy. The adjustments
themselves would not have any noticeable effects on local printing firms’
schedules of demand for printers or on workers’ supply schedules. But
suppose there is an economy-wide excess supply of labor. How is the
conventional adjustment apparatus to be deployed?
The orthodox instinct is to think of the price in this market as the real
wage. It is in terms of the real wage that the employers’ downward sloping
demand schedule, following the law of diminishing marginal productivity, is
expressed. In the same terms are expressed workers’ marginal choices
148 James Tobin
between the consumption rewards of paid employment and the utilities of
other uses of time. The orthodox expectation and prescription is that real
wages fall to eliminate unemployment.
But, Keynes asks, how do workers and employers engineer an economy-
wide reduction in real wages? The unemployment is nation-wide, but the
markets where wages are set are decentralized. In every local market it is
the money wage, not the real wage, that is determined. If money wage
rates fall in all these excess-supply local labor markets, will real wages in
fact fall?
It is certainly far from obvious. The relevant labor demand curves are the
nominal values of marginal products. These values will fall, the demand
curves shift down, if and as product prices fall. Product prices will fall
because nominal labor incomes decline along with wage rates; as a result,
workers’ money demands for the products they produce will decline too.
Here, then, is a case in which demand and supply schedules do not stay put
while the price adjustment to excess supply takes place. It is illegitimate to
appeal to the intuition that seems so credible for single markets. Instead, the
question is whether proportionate deflation of all nominal prices will or will
not increase aggregate effective real demand.3
Two issues in this debate need to be distinguished. The first concerns the
relation of real aggregate demand to the price level. The second concerns its
relation to the expected rate of change of prices. In discussing them, I shall
not distinguish between money wages and nominal product prices or between
their rates of change, but rather follow the assumption, conventional in this
debate, that they move together. I remind you that the theoretical argument
refers to a closed economy—maybe the United States in years gone by, or
post-1992 Europe, or the whole OECD area.
Keynes in Book I of the General Theory denied that real aggregate
demand was related at all to the price and money wage level. In effect, he
turned the classical neutrality proposition against the classicals. If all money
wages and prices are lowered in the same proportion, how can real
quantities demanded be any different? Thus, if a real shock makes real
demand deficient, how can a purely nominal price adjustment undo the
damage?
Actually Keynes himself provided an answer in Chapter 19. If the nominal
quantity of money remains the same, its real quantity increases, interest rates
fall, and real demand increases. This scenario is often called the ‘Keynes
effect’. This mechanism would fail if demand for money became perfectly
elastic with respect to interest rates—as in the liquidity trap discussed
above—or if demand for goods for consumption and investment were
perfectly inelastic.
Pigou (1943, 1947), Patinkin (1948, 1956 [1965]), and other authors
provided another scenario, the ‘Pigou effect’ or ‘real balance effect’, which
alleges a direct effect of increased wealth, in the case at hand taking the
form of the increased real value of base money, on real consumption demand
Price flexibility and output stability 149
(possibly also on investment demand as wealth-owners seek to maintain
portfolio balance between real and nominal assets). This effect does not
depend on reduction of interest rates.
To an astonishing degree, the theoretical fraternity has taken the real
balance effect to be a conclusive refutation of Keynes. Perhaps it does refute
his claim to have found underemployment equilibria. If involuntary
unemployment and excess capacity are pushing nominal wages and prices
down, the economy is not in equilibrium in any sense. It is not in a position
of rest, markets are not clearing, and expectations are not being realized.
Equilibrium requires wages and prices so low that the purchasing power of
net monetary wealth is so great that aggregate real demand creates jobs for
all willing workers. In principle, as Leontief observed, prices could be low
enough to enable you to buy the whole GNP for one thin dime.
Nevertheless the real balance effect is of dubious strength, and even of
uncertain sign. Most nominal assets in a modern economy are ‘inside’ assets,
that is the debts of private agents to other private agents. They wash out in
accounting aggregation, leaving only the government’s nominal debt to the
private sector as net wealth. Some, though probably not all, of that debt is
internalized by taxpayers. The base of the real balance effect is therefore
quite small relative to the economy—in the United States the monetary base
is currently only 6 percent of GNP. A 10 percent increase in the value of
money would increase net wealth by 0.6 percent of GNP and, if the marginal
propensity to spend from wealth were generously estimated at 0.10, would
increase spending by 0.06 percent of GNP.
While Don Patinkin (1948) stressed the theoretical importance of the real
balance effect, he disclaimed belief in its practical significance. In the Great
Depression, he pointed out, the real value of net private balances rose 46
percent from 1929 to 1932, but real national income fell 40 percent.
That inside assets and debts wash out in accounting aggregation does
not mean that the consequences of price changes on their real values wash
out. Price declines make creditors better off and debtors poorer. Their
marginal propensities to spend from wealth need not be the same. Common
sense suggests that debtors have the higher spending propensities—that is
why they are in debt! Even a small differential could easily swamp the
Pigou effect—gross inside dollar-denominated assets are 200 percent of
United States GNP.
Irving Fisher (1933) emphasized the increased burden of debt resulting
from unanticipated deflation as a major factor in depressions in general and
in the Great Depression in particular. Therefore, I like to call the reverse
Pigou-Patinkin effect the Fisher wealth redistribution effect (not to be
confused with other Fisher effects). It is quite possible that this Fisher effect is
stronger than the Pigou and Keynes effects combined, particularly when
output and employment are low relative to capacity.4
150 James Tobin
AGGREGATE DEMAND AND THE RATE OF CHANGE OF PRICES
The previous argument refers to levels of nominal wages and prices. An even
more important argument refers to rates of change. The Keynes and Pigou
effects compare high prices and low as if they were timeless alternatives,
without worrying about the process of change from high to low in real time.
Economists of their day argued in this way quite consciously, as dictated by
the rules of the comparative statics games they were playing.
The process of change works on aggregate demand in just the wrong
direction. Greater expected deflation, or expected disinflation, is an increase
in the real rate of interest, necessarily so when nominal interest rates are
constrained by the zero floor of the interest on money. Here is another Fisher
effect, another factor Fisher stressed in his explanation of the Great
Depression. Keynes stressed it too, as a pragmatic dynamic reinforcement of
the lesson of his static general theory.
The problematic stability of price adjustment is evident in Figure 6.1.
Here the horizontal axis represents expected price deflation or inflation, x.
The vertical axis represents p the log of the price level. An upward sloping
curve like plots combinations (x, p) of expected price change and price level
that generate the same aggregate real demand E. The slope reflects the
assumptions that demand is related negatively to the price level and
positively to its expected rate of change. In given circumstances, a higher
curve refers to a lower demand E and a lower curve to higher demand. The
curvature of the E* loci reflects the assumption that the ‘Keynes effect’ of
increases in real money balances in lowering interest rates declines as those
balances increase and interest rates fall.
Figure 6.1 The problematic stability of price adjustment
Price flexibility and output stability 151
Suppose that initially the ‘isoquant’ E*1 makes demand equal to full
employment equilibrium output Y*, here taken to be constant. Points above
or left of that isoquant are positions where E is lower than Y*,
characterized by Keynesian unemployment. Points below or right of E*1 are
positions of macroeconomic excess demand. In Figure 6.1, the equilibrium
inflation rate (expected and actual) and price are (0, p1). Suppose now that
a discrete one-time negative shock to real demand shifts the isoquant for
E=Y* down to E*1 so that the new equilibrium inflation rate and price are
(0, p2). The old isoquant E
*
1 now implies an E lower than Y*. To restore
equilibrium the price level must fall from p1 to p2. How is the price decline
t o b e a c c o m p l i s h e d ? O n e s c e n a r i o i s t h e Wa l r a s i a n m i r a c l e , a n
instantaneous precipitous vertical descent, so that there is no time interval
during which actual or expected price changes are other than zero. If jumps
of that kind in p are excluded, there is no path of actual price changes and
rationally expected prices that avoids departure from E=Y* during the
transition. It would take a burst of positive inflation, actual and expected,
to offset the negative demand shock, as at point A. But this would move the
price level in the wrong direction.
The likely scenario is a path like B or C in Figure 6.1: the excess supply
that now characterizes the initial equilibrium point (0, p1) and the first
isoquant starts prices declining, and the anticipation of their decline is bad
for aggregate demand. Along B the real balance effect is strong enough to
overcome the negative effects of the deflation; aggregate demand E is
increasing as the path hits lower isoquants. The new equilibrium may be
attained, though probably by a damped cyclical process. Along C, however,
the price level effect is too weak to win out, and the gap of E and Y below
Y* is increasing.
Fisher and Keynes were right. In Tobin (1975), I exhibited a simple formal
macroeconomic system, classical in the sense that it has only one
equilibrium, which is characterized by full employment, indeed by a
‘natural’ rate of unemployment. Given a zero natural real growth rate and a
constant nominal monetary base, the price level is constant in that
equilibrium.
Several specifications of the short-run dynamics of this model are
possible. One is a Keynesian specification, as follows: (1) Production
increases when desired purchases exceed actual current output, but not by
the full amount of the gap. This adjustment can be thought of as response
to undesired changes in inventories or unfilled orders. (2) Nominal prices
follow expectations plus or minus a ‘Phillips curve’ adjustment to the
difference between actual and full employment output. (3) Price change
expectations adapt to the difference between actual and expected inflation
or deflation.
Alternatively, the price change expectations could be regarded as rational
expectations of the Phillips curve price adjustment mechanisms. That is, the
impossibility of instantaneous jumps to the new equilibrium would be as
152 James Tobin
intrinsic to the structure of the system as the system’s static equations
themselves.
The stability of this system requires, first, that the dynamics of output at
constant prices, involving marginal propensities to spend and adjustments to
excess or deficient inventories and other manifestations of demand/ output
gaps, is stable. Assuming this condition is met, stability depends on the
relative strengths of the price level effects on demand—both ‘Keynes’ and
‘Pigou’ as modified by ‘Fisher wealth redistribution’—and the real interest
effect—another ‘Fisher’—of expected deflation (or disinflation). The latter is
the product of two coefficients, the response of price change expectations to
actual change (equal to one if expectations are rational) and the response of
real demand to expected price change. The real interest effect may well
dominate if the real balance effect is weak, especially if the Fisher wealth
redistribution effect overshadows it, and if the demand for money is highly
sensitive to interest rates. The equilibrium is then unstable. Moreover,
because of the curvature of the E* loci, the system could be stable locally but
unstable for large displacements.
I have experimented with simulations of a discrete-time approximation to
this model, subjecting it to stochastic shocks to real aggregate demand. One
extreme case is ‘Walrasian’: prices vary from period to period as necessary
to keep goods markets always cleared, prices are always anticipated to equal
their expected value corresponding to zero shock, and both output and
aggregate demand always equal equilibrium full employment output. An
opposite extreme is ‘rigid-price Keynesian’: prices are constant at their
expected equilibrium value and expectations of price change are constant at
zero. In between the extremes, nominal prices adjust with some inertia to
excess real demand or supply, and expectations of price change adapt, more
or less speedily, to observed changes.
In these simulations the underlying ‘fixprice’ dynamics are stable, and its
parameters are the same in all cases. ‘Greater price flexibility’ can mean two
things: (1) a larger Phillips curve coefficient relating price change to excess
real demand or supply; (2) if expectations are taken to be adaptive, a larger
coefficient of adaptation of price change expectations to actual price changes.
The issue is whether greater price flexibility increases or decreases the
ratio between the standard deviation of the actual output gap and the
standard deviation of the stochastic real demand shock. That ratio is zero in
the Walrasian case, where the shock is always wholly absorbed in prices. It
is of course positive for the rigid-price case. What happens in the
intermediate cases? Not surprisingly, the results depend mainly on the same
condition that determines stability or instability with respect to a single
unrepeated shock. Greater flexibility in sense (1), a faster ‘Phillips’
adjustment, diminishes the test ratio when the stability condition is met—that
is, the price level effect on demand is negative and bigger than the price
change effect—and raises it otherwise. Greater flexibility in sense (2), faster
adjustment of price expectations, always raises the test ratio.5
Price flexibility and output stability 153
POLICIES, EXPECTATIONS AND STABILITY
Keynes stressed the central role of long-term expectations. He had in mind in
particular expectations of real variables—effective demands and real returns
on investments. They might be either stabilizing or destabilizing. If business
managers believe that recessions will be quickly reversed, their actions will
help to bring about recoveries. If they expect business activity to continue to
be subnormal or to fall further, their pessimism may turn recession into
depression. That is why policies and policy expectations are very important.
After World War II, widespread perception that government fiscal and
monetary policies would keep recessions short and shallow helped to keep
them short and shallow. In these circumstances, the economy would work
well if, as Keynes advocated, employers and workers kept average money
wage rates stable, so that actual and expected price and wage changes were
not a source of instability.
In the 1930s both Fisher and Keynes saw deflation as a cause of
depression in production and employment, and advocated monetary and gold
policies of reflation for recovery. Today, however, unexpectedly high prices
are regarded as bearish economic news, and unexpectedly low prices as
bullish. Is this a paradox? Does it mean that price flexibility is stabilizing
after all? Again, policies and policy expectations are crucial. Today the
public understands the high priorities central banks attach to inflation
control. If prices are above the path to which the central bank is committed,
it will take measures to contract demand. The faster private agents respond
by lowering prices and wages, the sooner the monetary authorities will
reflate. In this sense, price flexibility is stabilizing.
In contrast, extrapolative expectations are destabilizing. Policies—policy
rules if you like—that create and sustain regressive expectations of output
and price departures from equilibrium are stabilizing. Those facts are wholly
consistent with the contentions of Fisher and Keynes, and of this paper, that
in the absence of activist ‘feedback’ policies, monetary and fiscal, flexibility
may well be destabilizing, both to prices and to real macro variables.
Governments and central banks should not expect disinflation or deflation
alone to maintain or restore full employment.
ACKNOWLEDGEMENT
I would like to express my gratitude for the faithful and valuable research
assistance of Mitchell Tobin, Yale College 1992 (no relation).
NOTES
1 In Tobin (1955), stagnation is one possibility, the stable solution of a non-linear
model whose unstable solution is a repetitive cycle.
2 Dudley Dillard (1988) calls this the ‘barter illusion’ of classical economics.
154 James Tobin
3 In formal general equilibrium theory the stability of markets determining relative
prices cannot be guaranteed without special assumptions. This is a fortiori true if
money is introduced and markets determine nominal prices. See the survey by
Franklin Fisher (1987).
4 I have exhibited a dominant Fisher effect and examined its macroeconomic
consequences in an IS-LM model that also has a Keynes effect, in Tobin (1980: ch.
1). See also Caskey and Fazzari (1987).
5 At long last the question whether price flexibility (in any sense short of the Walrasian
auctioneer fairy tale) is stabilizing has begun to receive considerable attention. De
Long and Summers (1986) have investigated this question using the Fischer-Taylor
staggered-contract model (Fischer 1977; Taylor 1980), amended to allow both
price-level and price-change effects on demand. Their most interesting simulation
has the intuitively desirable property that close to the limit of perfect price flexibility
greater price flexibility means greater real stability, while farther away from it the
reverse is true. Similar results are obtained by Caskey and Fazzari (1988) and
Chadha (1989).
REFERENCES
Ball, Lawrence, N.Gregory Mankiw, and David Romer, ‘The New Keynesian Economics
and the Output-Inflation Tradeoff’, Brookings Papers on Economic Activity 1988
1, 1–65.
Barro, Robert, and Herschel Grossman, ‘A General Disequilibrium Model of Income
and Employment’, American Economic Review March 1971, 61, 82–93.
Caskey, John, and Steve Fazzari, ‘Aggregate Demand Contractions with Nominal Debt
Commitments’, Economic Inquiry October 1987, 25, 583–97.
Caskey, John, and Steven Fazzari, ‘Price Flexibility and Macroeconomic Stability: An
Empirical Simulation Analysis’, Washington University Department of Economics,
working paper 118, January 1988.
Chadha, Binky, ‘Is Increased Price Inflexibility Stabilizing?’, Journal of Money, Credit
and Banking November 1989, 21, 481–97.
Cooper, Russell, and Andrew John, ‘Coordinating Coordination Failures in Keynesian
Models’, Quarterly Journal of Economics August 1988, 100, 441–63.
De Long, J.Bradford, and Lawrence H.Summers, ‘Is Increasing Price Flexibility
Stabilizing?’ American Economic Review December 1986, 76, 1031–44.
Dillard, Dudley, ‘The Barter Illusion in Classical and Neoclassical Economies’, Eastern
Economic Journal October-December 1988, 14, 299–318.
Fischer, Stanley, ‘Long-term Contracts, Rational Expectations, and the Optimal Money
Supply Rule’, Journal of Political Economy February 1977, 85:1, 191–205.
Fisher, Franklin, M., ‘Adjustment Processes and Stability’, in John Eatwell, Murray
Milgate, and Peter Newman (eds) The New Palgrave: A Dictionary of Economics,
London: Macmillan, 1987, 26–9.
Fisher, Irving, ‘The Debt-Deflation Theory of Great Depressions’, Econometrica October
1933, 1, 337–57.
Hansen, Alvin H., Full Recovery or Stagnation, New York: W.W.Norton, 1938.
Iwai, Katsuhito, Disequilibrium Dynamics (Cowles Foundation Monograph 27), New
Haven, CT: Yale University Press, 1981.
Keynes, John Maynard, The General Theory of Employment, Interest, and Money,
New York: Harcourt Brace, 1936.
Lindbeck, Assar, and Dennis J.Snower, The Insider-Outsider Theory of Employment
and Unemployment, Cambridge, MA: MIT Press, 1990.
Price flexibility and output stability 155
Mankiw, N.Gregory, ‘A Quick Refresher Course in Macroeconomics’, Journal of
Economic Literature December 1990, 28, 1645–60.
Okun, Arthur M., Prices and Quantities: A Macroeconomic Analysis, Washington,
DC: Brookings Institution, 1981.
Patinkin, Don, ‘Price Flexibility and Full Employment’, American Economic Review
September 1948, 38, 543–64.
Patinkin, Don, Money, Interest, and Prices, New York: Harper and Row, 1956, 2nd
edn, 1965.
Pigou, Arthur Cecil, ‘The Classical Stationary State’, Economic Journal December 1943,
53, 313–51.
Pigou, Arthur Cecil, ‘Economic Progress in a Stable Environment’, Economica August
1947, 14, 180–90.
Plosser, Charles I., ‘Understanding Real Business Cycles’, Journal of Economic
Perspectives Summer 1989, 3:3, 51–77.
Taylor, John, ‘Aggregate Dynamics and Staggered Contracts’, Journal of Political
Economy February 1980, 88, 1–23.
Tobin, James, ‘A Dynamic Aggregative Model’, Journal of Political Economy April
1955, 63, 103–15.
Tobin, James, ‘Keynesian Models of Recession and Depression’, American Economic
Review (Papers and Proceedings), May 1975, 55, 195–202.
Tobin, James, Asset Accumulation and Economic Activity, Oxford: Basil Blackwell,
1980.
Tobin, James, ‘Financial Structure and Monetary Rules’, Kredit und Kapital 1983, 16,
155–71.
Part II
The monetarist counter-
revolution
Introduction
‘Monetarism’, a term first introduced by Karl Brunner (1968), refers to a
school of economic thought which initially evolved in the United States to
attack the orthodox Keynesian analysis and associated policy-activism of the
1950s and 1960s. The historical development of the orthodox monetarist
school can be traced in three main stages (see Snowdon et al. 1994). The
first main stage, from the mid-1950s to the mid-1960s, involved an attempt,
by Professor Milton Friedman and his associates, to re-establish the quantity
theory of money approach to macroeconomic analysis. In this approach
changes in the money supply are regarded as the predominant, though not
the only, factor explaining changes in money income. Friedman’s (1956)
seminal article on ‘The Quantity Theory of Money: A Restatement’ in which
he asserted that the demand for money (and by implication velocity) is a
stable function of a limited number of variables lies at the heart of the
modern quantity theory approach. However, it was the breadth and depth of
empirical evidence cited in Friedman and Schwartz’s (1963) book on A
Monetary History of the United States, 1867–1960 which was particularly
influential in reviving interest in the potency of money in generating cyclical
fluctuations. Their analysis of the US historical record provided strong
support for the view that independent movements of the stock of money
played a significant role in causing macroeconomic instability. According to
Lucas (1994) this book played ‘an important—perhaps even decisive—role in
the 1960s’ debates over stabilization policy between Keynesians and
monetarists’.
The second main stage in the development of orthodox monetarism
involved the expectations-augmented Phillips curve analysis which was
absorbed into monetarist analysis after the mid-to-late 1960s. Central to this
phase is Friedman’s 1967 Presidential Address to the American Economic
Association, subsequently published in 1968 in the American Economic
Review as ‘The Role of Monetary Policy’. In this article (reprinted on pp.
164–79) Friedman denies the existence of a permanent/long-run trade-off
between inflation and unemployment and introduces the natural rate of
unemployment hypothesis. The essence of this hypothesis is a reaffirmation
of the classical view that in the long run nominal magnitudes cannot
determine real magnitudes such as employment and output. According to
160 The monetarist counter-revolution
Friedman monetary policy cannot, other than for very limited periods,
achieve some target unemployment rate and any attempt to maintain
unemployment below the natural rate will produce accelerating inflation; a
prediction subsequently borne out by the experience of many western
economies during the 1970s (see Snowdon and Vane 1997). In the conduct of
monetary policy he prescribes that the authorities pursue a ‘stable’ rate of
monetary growth in line with the trend/long-run growth rate of the economy
to ensure long-run price stability. Friedman argues that the natural rate of
unemployment can be reduced only by appropriate supply-side
microorientated policies which improve the operation of the labour market. In
1981 Robert Gordon described Friedman’s 1968 paper as probably the most
influential article written in macroeconomics in the previous twenty years.
More recently James Tobin (1995), one of Friedman’s most eloquent, effective
and long-standing critics, has described the paper as ‘very likely the most
influential article ever published in an economics journal’ (emphasis added).
The third main stage in the development of orthodox monetarism came in
the 1970s with the incorporation of the monetary approach to balance of
payments theory and exchange rate determination into monetarist analysis.
This made monetarist analysis which, up to that time, had been implicitly
developed in the relatively closed economy context of the US economy under
the Bretton Woods system, relevant to open economies such as the UK. It also
provided an explanation of the international transmission of inflation in the
late 1960s (see Frenkel and Johnson 1976).
While it is possible to outline the historical development of orthodox
monetarism in these three main stages, it is far more problematic to specify a
set of characteristics to which all monetarists would subscribe. Within the
monetarist school there have been, and continue to be, differences of opinion
and emphasis and it is impossible to produce a definitive list of characteristics
of the monetarist viewpoint which would be universally accepted. Over the
years there have been a number of papers which have sought to summarize the
central distinguishing beliefs within the orthodox monetarist school of thought
(see for example Brunner 1970; Friedman 1970; Purvis 1980). In his 1975
Kredit und Kapital article (reprinted on pp. 180–215) on ‘The Structure of
Monetarism’, Thomas Mayer discusses twelve beliefs of varying significance
which at that time were in his view generally held by monetarists in the United
States. In Part I of his article Mayer discusses six propositions relating to
theory and techniques of analysis, and in Part II six policy-orientated
propositions. While he demonstrates the connections between the complete set
of monetarist propositions identified, Mayer argues that economists can accept
some of these propositions while rejecting others. Another attempt to specify
monetarism’s key characteristics is David Laidler’s 1981 Economic Journal
article entitled ‘Monetarism: An Interpretation and an Assessment’. This
article (reprinted on pp. 216–46) was initially prepared for presentation at a
Conference on ‘Monetarism—An Appraisal’ organized by the Royal Economic
Society (RES) in London in July 1980. The March 1981 issue of the Economic
Introduction 161
Journal also includes an article by James Tobin appraising the monetarist
counter-revolution, together with comments on both Laidler’s and Tobin’s
articles by R.C.O.Matthews and James E.Meade, presented at the RES
Conference. As the reader will discover, Laidler’s article provides both an
exposition and appraisal of monetarism in terms of four key characteristics,
namely a quantity theory approach to macroeconomic analysis, an analysis of
the division of money income fluctuations between the price level and real
income, a monetary approach to balance-of-payments and exchange-rate
theory, and an antipathy to activist stabilization policy and support for long-
run monetary policy ‘rules’.
However monetarism is defined, one characteristic which all monetarists
agree upon is the emphasis given to controlling the money stock and prior to
the 1980s ‘velocity puzzle’ a majority of monetarists supported a monetary
growth rule or target, in order to both attain and maintain long-run price
stability. The case for Friedman’s ‘stable’ monetary growth rule is
inextricably linked to empirical evidence of a stable demand for money
function (and associated trend in velocity) for the post-war period up to the
1970s, both in the US and other economies. Evidence of money demand
instability (and a shift in the trend of velocity, with velocity becoming more
erratic) especially since 1981 in the United States and elsewhere has
undermined the case for a fixed monetary growth rule. Franco Modigliani’s
1988 Contemporary Policy Issues article (reprinted on pp. 247–61) entitled
‘The Monetarist Controversy Revisited’ examines two proposed ‘feedback’
monetary growth rules, and the case for stabilization policy, in the light of
US experience during the 1980s.
Although within academia monetarism is no longer the influential force it
was in the late 1960s and early 1970s, its apparent demise can in large part
be attributed to the fact that a number of its insights have been absorbed into
mainstream macroeconomics (see Mayer 1997). Most notably the
expectations-augmented Phillips curve analysis, the view that the long-run
Phillips curve is vertical and that money is neutral in the long run are all
now widely accepted and form part of mainstream macroeconomics.
Furthermore a majority of economists and central banks emphasize the rate
of growth of the money supply when it comes to explaining and combating
inflation over the long run. But perhaps the most important and lasting
contribution of monetarism has been to persuade many economists to accept
the idea that the potential of activist discretionary fiscal and monetary policy
is much more limited than conceived prior to the monetarist counter-
revolution.
REFERENCES
* Titles marked with an asterisk are particularly recommended for additional reading.
Brunner, K. (1968) ‘The Role of Money and Monetary Policy’, Federal Reserve Bank of
St Louis Review 50, July, pp. 9–24.
162 The monetarist counter-revolution
*Brunner, K. (1970) ‘The Monetarist Revolution in Monetary Theory’,
Weltwirtschaftliches Archiv 105, March, pp. 1–30.
Frenkel, J.A. and H.G.Johnson (eds) (1976) The Monetary Approach to the Balance of
Payments, London: Allen and Unwin.
Friedman, M. (1956) ‘The Quantity Theory of Money: A Restatement’, in M. Friedman
(ed.) Studies in the Quantity Theory of Money, Chicago: University of Chicago
Press.
* Friedman, M. (1970) The Counter-Revolution in Monetary Theory, IEA Occasional
Paper no. 33, London: Institute of Economic Affairs.
*Friedman, M. (1975) Unemployment versus Inflation? An Evaluation of the Phillips
Curve, IEA Occasional Paper no. 44, London: Institute of Economic Affairs.
*Friedman, M. (1977) ‘Nobel Lecture: Inflation and Unemployment’, Journal of Political
Economy 85, June, pp. 451–72.
Friedman, M. and A.J.Schwartz (1963) A Monetary History of the United States,
1867–1960, Princeton, NJ: Princeton University Press.
*Froyen, R.T. (1996) Macroeconomics: Theories and Policies, 5th edn, Chapters 9 and
10, London: Prentice-Hall.
Gordon, R.J. (1981) ‘Output Fluctuations and Gradual Price Adjustment’, Journal of
Economic Literature 19, June, pp. 493–530.
*Hoover, K.D. (1984) ‘Two Types of Monetarism’, Journal of Economic Literature 22,
March, pp. 58–76.
Hoover, K.D. and S.M.Sheffrin (eds) (1995) Monetarism and the Methodology of
Economics: Essays in Honour of Thomas Mayer, Aldershot: Edward Elgar.
*Jansen, D.W., C.D.Delorme and R.B.Ekelund, Jr (1994) Intermediate Macroeconomics,
Chapter 8, New York: West.
* Johnson, H.G. (1971) ‘The Keynesian Revolution and the Monetarist Counter-
Revolution’, American Economic Review 61, May, pp. 1–14.
Lucas, R.E. Jr (1994) ‘Review of Milton Friedman and Anna J.Schwartz’s “A Monetary
History of the United States, 1867–1960”’, Journal of Monetary Economics 34,
August, pp. 5–16.
*Mayer, T. (1990) Monetarism and Macroeconomic Policy, Aldershot: Edward Elgar.
*Mayer, T. (1997) ‘What Remains of the Monetarist Counter-Revolution’, in B.
Snowdon and H.R.Vane (eds) Reflections on the Development of Modern
Macroeconomics, Aldershot: Edward Elgar.
Purvis, D.D. (1980) ‘Monetarism: A Review’, Canadian Journal of Economics 1,
February, pp. 96–121.
Snowdon, B. and H.R.Vane (1997) ‘Modern Macroeconomics and its Evolution from
a Monetarist Perspective: An Interview with Professor Milton Friedman’, Journal of
Economic Studies 24.
*Snowdon, B., H.R.Vane and P.Wynarczyk (1994) A Modern Guide to Macroeconomics:
An Introduction to Competing Schools of Thought, Chapter 4, Aldershot: Edward
Elgar.
Tobin, J. (1981) ‘The Monetarist Counter-Revolution Today: An Appraisal’, Economic
Journal 91, March, pp. 29–42.
Tobin, J. (1995) ‘The Natural Rate as New Classical Economies’, in R.Cross (ed.) The
Natural Rate of Unemployment: Reflections on 25 Years of the Hypothesis,
Cambridge: Cambridge University Press.
QUESTIONS
1 Compare and contrast the traditional quantity theory of money with its
modern restatement.
Introduction 163
2 What do you understand by the term ‘monetarism’? In what ways does the
monetarist approach to macroeconomics differ from the orthodox Keynesian
approach?
3 ‘Monetarist analysis seeks to explain the short-run non-neutrality of money
whilst preserving the classical view of long-run neutrality’. Explain and
discuss.
4 How do fluctuations in the quantity of money influence real variables
according to monetarist analysis?
5 What are the policy implications of Friedman’s analysis of the expectations-
augmented Phillips curve?
6 ‘The case for or against monetarism is fundamentally empirical, not
theoretical’. Discuss.
7 What were the major factors which contributed to the rise of monetarism in
academia and policy circles during the 1970s?
8 Critically examine the monetarist arguments against policy activism and
assess the case for a monetary rule.
9 What remains of the ‘monetarist counter-revolution’?
10 To what extent has Milton Friedman proved to be the most influential
macroeconomist since Keynes?
7 The role of monetary policy
Milton Friedman
American Economic Review (1968) 58, March, pp. 1–17
There is wide agreement about the major goals of economic policy: high
employment, stable prices, and rapid growth. There is less agreement that
these goals are mutually compatible or, among those who regard them as
incompatible, about the terms at which they can and should be substituted
for one another. There is least agreement about the role that various
instruments of policy can and should play in achieving the several goals.
My topic for tonight is the role of one such instrument—monetary policy.
What can it contribute? And how should it be conducted to contribute the
most? Opinion on these questions has fluctuated widely. In the first flush of
enthusiasm about the newly created Federal Reserve System, many observers
attributed the relative stability of the 1920s to the System’s capacity for fine
tuning—to apply an apt modern term. It came to be widely believed that a
new era had arrived in which business cycles had been rendered obsolete by
advances in monetary technology. This opinion was shared by economist
and layman alike, though, of course, there were some dissonant voices. The
Great Contraction destroyed this naive attitude. Opinion swung to the other
extreme. Monetary policy was a string. You could pull on it to stop inflation
but you could not push on it to halt recession. You could lead a horse to
water but you could not make him drink. Such theory by aphorism was soon
replaced by Keynes’ rigorous and sophisticated analysis.
Keynes offered simultaneously an explanation for the presumed impotence
of monetary policy to stem the depression, a nonmonetary interpretation of
the depression, and an alternative to monetary policy for meeting the
depression and his offering was avidly accepted. If liquidity preference is
absolute or nearly so—as Keynes believed likely in times of heavy
unemployment—interest rates cannot be lowered by monetary measures. If
investment and consumption are little affected by interest rates—as Hansen
and many of Keynes’ other American disciples came to believe—lower
interest rates, even if they could be achieved, would do little good. Monetary
policy is twice damned. The contraction, set in train, on this view, by a
collapse of investment or by a shortage of investment opportunities or by
stubborn thriftiness, could not, it was argued, have been stopped by
monetary measures. But there was available an alternative—fiscal policy.
The role of monetary policy 165
Government spending could make up for insufficient private investment. Tax
reductions could undermine stubborn thriftiness.
The wide acceptance of these views in the economics profession meant
that for some two decades monetary policy was believed by all but a few
reactionary souls to have been rendered obsolete by new economic
knowledge. Money did not matter. Its only role was the minor one of
keeping interest rates low, in order to hold down interest payments in the
government budget, contribute to the ‘euthanasia of the rentier’, and maybe,
stimulate investment a bit to assist government spending in maintaining a
high level of aggregate demand.
These views produced a widespread adoption of cheap money policies
after the war. And they received a rude shock when these policies failed in
country after country, when central bank after central bank was forced to
give up the pretense that it could indefinitely keep ‘the’ rate of interest at a
low level. In the United States, the public denouement came with the Federal
Reserve-Treasury Accord in 1951, although the policy of pegging
government bond prices was not formally abandoned until 1953. Inflation,
stimulated by cheap money policies, not the widely heralded postwar
depression, turned out to be the order of the day. The result was the
beginning of a revival of belief in the potency of monetary policy.
This revival was strongly fostered among economists by the theoretical
developments initiated by Haberler but named for Pigou that pointed out a
channel—namely, changes in wealth—whereby changes in the real quantity
of money can affect aggregate demand even if they do not alter interest
rates. These theoretical developments did not undermine Keynes’ argument
against the potency of orthodox monetary measures when liquidity
preference is absolute since under such circumstances the usual monetary
operations involve simply substituting money for other assets without
changing total wealth. But they did show how changes in the quantity of
money produced in other ways could affect total spending even under such
circumstances. And, more fundamentally, they did undermine Keynes’ key
theoretical proposition, namely, that even in a world of flexible prices, a
position of equilibrium at full employment might not exist. Henceforth,
unemployment had again to be explained by rigidities or imperfections, not
as the natural outcome of a fully operative market process.
The revival of belief in the potency of monetary policy was fostered also
by a re-evaluation of the role money played from 1929 to 1933. Keynes and
most other economists of the time believed that the Great Contraction in the
United States occurred despite aggressive expansionary policies by the
monetary authorities—that they did their best but their best was not good
enough.1 Recent studies have demonstrated that the facts are precisely the
reverse: the US monetary authorities followed highly deflationary policies.
The quantity of money in the United States fell by one-third in the course of
the contraction. And it fell not because there were no willing borrowers– not
because the horse would not drink. It fell because the Federal Reserve System
166 Milton Friedman
forced or permitted a sharp reduction in the monetary base, because it failed
to exercise the responsibilities assigned to it in the Federal Reserve Act to
provide liquidity to the banking system. The Great Contraction is tragic
testimony to the power of monetary policy—not, as Keynes and so many of
his contemporaries believed, evidence of its impotence.
In the United States the revival of belief in the potency of monetary policy
was strengthened also by increasing disillusionment with fiscal policy, not so
much with its potential to affect aggregate demand as with the practical and
political feasibility of so using it. Expenditures turned out to respond
sluggishly and with long lags to attempts to adjust them to the course of
economic activity, so emphasis shifted to taxes. But here political factors
entered with a vengeance to prevent prompt adjustment to presumed need, as
has been so graphically illustrated in the months since I wrote the first draft
of this talk. ‘Fine tuning’ is a marvelously evocative phrase in this electronic
age, but it has little resemblance to what is possible in practice–not, I might
add, an unmixed evil.
It is hard to realize how radical has been the change in professional
opinion on the role of money. Hardly an economist today accepts views that
were the common coin some two decades ago. Let me cite a few examples.
In a talk published in 1945, E.A.Goldenweiser, then Director of the
Research Division of the Federal Reserve Board, described the primary
objective of monetary policy as being to ‘maintain the value of Government
bonds…. This country’ he wrote, ‘will have to adjust to a 2 1/2 per cent
interest rate as the return on safe, long-time money, because the time has
come when returns on pioneering capital can no longer be unlimited as they
were in the past’ [4, p. 117].
In a book on Financing American Prosperity, edited by Paul Homan and
Fritz Machlup and published in 1945, Alvin Hansen devotes nine pages of
text to the ‘savings-investment problem’ without finding any need to use the
words ‘interest rate’ or any close facsimile thereto [5, pp. 218–27]. In his
contribution to this volume, Fritz Machlup wrote, ‘Questions regarding the
rate of interest, in particular regarding its variation or its stability, may not
be among the most vital problems of the postwar economy, but they are
certainly among the perplexing ones’ [5, p. 466]. In his contribution, John
H.Williams—not only professor at Harvard but also a long-time adviser to
the New York Federal Reserve Bank—wrote, ‘I can see no prospect of revival
of a general monetary control in the postwar period’ [5, p. 383].
Another of the volumes dealing with postwar policy that appeared at this
time, Planning and Paying for Full Employment, was edited by Abba P.
Lerner and Frank D.Graham [6] and had contributors of all shades of
professional opinion—from Henry Simons and Frank Graham to Abba
Lerner and Hans Neisser. Yet Albert Halasi, in his excellent summary of the
papers, was able to say, ‘Our contributors do not discuss the question of
money supply…. The contributors make no special mention of credit policy
to remedy actual depressions…. Inflation…might be fought more effectively
The role of monetary policy 167
by raising interest rates…. But…other anti-inflationary measures…are
preferable’ [6, pp. 23–4]. A Survey of Contemporary Economics, edited by
Howard Ellis and published in 1948, was an ‘official’ attempt to codify the
state of economic thought of the time. In his contribution, Arthur Smithies
wrote, ‘In the field of compensatory action, I believe fiscal policy must
shoulder most of the load. Its chief rival, monetary policy, seems to be
disqualified on institutional grounds. This country appears to be committed
to something like the present low level of interest rates on a long-term basis’
[1, p. 208].
These quotations suggest the flavor of professional thought some two
decades ago. If you wish to go further in this humbling inquiry, I recommend
that you compare the sections on money—when you can find them–in the
Principles texts of the early postwar years with the lengthy sections in the
current crop even, or especially, when the early and recent Principles are
different editions of the same work.
The pendulum has swung far since then, if not all the way to the position
of the late 1920s, at least much closer to that position than to the position of
1945. There are of course many differences between then and now, less in the
potency attributed to monetary policy than in the roles assigned to it and the
criteria by which the profession believes monetary policy should be guided.
Then, the chief roles assigned monetary policy were to promote price
stability and to preserve the gold standard; the chief criteria of monetary
policy were the state of the ‘money market’, the extent of ‘speculation’ and
the movement of gold. Today, primacy is assigned to the promotion of full
employment, with the prevention of inflation a continuing but definitely
secondary objective. And there is major disagreement about criteria of
policy, varying from emphasis on money market conditions, interest rates,
and the quantity of money to the belief that the state of employment itself
should be the proximate criterion of policy.
I stress nonetheless the similarity between the views that prevailed in the
late 1920s and those that prevail today because I fear that, now as then, the
pendulum may well have swung too far, that, now as then, we are in danger
of assigning to monetary policy a larger role than it can perform, in danger
of asking it to accomplish tasks that it cannot achieve, and, as a result, in
danger of preventing it from making the contribution that it is capable of
making.
Unaccustomed as I am to denigrating the importance of money, I therefore
shall, as my first task, stress what monetary policy cannot do. I shall then try
to outline what it can do and how it can best make its contribution, in the
present state of our knowledge—or ignorance.
WHAT MONETARY POLICY CANNOT DO
From the infinite world of negation, I have selected two limitations of
monetary policy to discuss: (1) it cannot peg interest rates for more than very
168 Milton Friedman
limited periods; (2) it cannot peg the rate of unemployment for more than
very limited periods. I select these because the contrary has been or is widely
believed, because they correspond to the two main unattainable tasks that
are at all likely to be assigned to monetary policy, and because essentially
the same theoretical analysis covers both.
Pegging of interest rates
History has already persuaded many of you about the first limitation. As
noted earlier, the failure of cheap money policies was a major source of the
reaction against simple-minded Keynesianism. In the United States, this
reaction involved widespread recognition that the wartime and postwar
pegging of bond prices was a mistake, that the abandonment of this policy
was a desirable and inevitable step, and that it had none of the disturbing
and disastrous consequences that were so freely predicted at the time.
The limitation derives from a much misunderstood feature of the relation
between money and interest rates. Let the Fed set out to keep interest rates
down. How will it try to do so? By buying securities. This raises their prices
and lowers their yields. In the process, it also increases the quantity of
reserves available to banks, hence the amount of bank credit, and, ultimately
the total quantity of money. That is why central bankers in particular, and
the financial community more broadly, generally believe that an increase in
the quantity of money tends to lower interest rates. Academic economists
accept the same conclusion, but for different reasons. They see, in their
mind’s eye, a negatively sloping liquidity preference schedule. How can
people be induced to hold a larger quantity of money? Only by bidding
down interest rates.
Both are right, up to a point. The initial impact of increasing the quantity
of money at a faster rate than it has been increasing is to make interest rates
lower for a time than they would otherwise have been. But this is only the
beginning of the process not the end. The more rapid rate of monetary
growth will stimulate spending, both through the impact on investment of
lower market interest rates and through the impact on other spending and
thereby relative prices of higher cash balances than are desired. But one
man’s spending is another man’s income. Rising income will raise the
liquidity preference schedule and the demand for loans; it may also raise
prices, which would reduce the real quantity of money. These three effects
will reverse the initial downward pressure on interest rates fairly promptly,
say, in something less than a year. Together they will tend, after a somewhat
longer interval, say, a year or two, to return interest rates to the level they
would otherwise have had. Indeed, given the tendency for the economy to
overreact, they are highly likely to raise interest rates temporarily beyond
that level, setting in motion a cyclical adjustment process.
A fourth effect, when and if it becomes operative, will go even farther,
and definitely mean that a higher rate of monetary expansion will
The role of monetary policy 169
corres-pond to a higher, not lower, level of interest rates than would
otherwise have prevailed. Let the higher rate of monetary growth produce
rising prices, and let the public come to expect that prices will continue to
rise. Borrowers will then be willing to pay and lenders will then demand
higher interest rates—as Irving Fisher pointed out decades ago. This price
expectation effect is slow to develop and also slow to disappear. Fisher
estimated that it took several decades for a full adjustment and more recent
work is consistent with his estimates.
These subsequent effects explain why every attempt to keep interest rates
at a low level has forced the monetary authority to engage in successively
larger and larger open market purchases. They explain why, historically,
high and rising nominal interest rates have been associated with rapid
growth in the quantity of money, as in Brazil or Chile or in the United States
in recent years, and why low and falling interest rates have been associated
with slow growth in the quantity of money, as in Switzerland now or in the
United States from 1929 to 1933. As an empirical matter, low interest rates
are a sign that monetary policy has been tight—in the sense that the quantity
of money has grown slowly; high interest rates are a sign that monetary
policy has been easy—in the sense that the quantity of money has grown
rapidly. The broadest facts of experience run in precisely the opposite
direction from that which the financial community and academic economists
have all generally taken for granted.
Paradoxically, the monetary authority could assure low nominal rates of
interest—but to do so it would have to start out in what seems like the
opposite direction, by engaging in a deflationary monetary policy. Similarly,
it could assure high nominal interest rates by engaging in an inflationary
policy and accepting a temporary movement in interest rates in the opposite
direction.
These considerations not only explain why monetary policy cannot peg
interest rates; they also explain why interest rates are such a misleading
indicator of whether monetary policy is ‘tight’ or ‘easy’. For that, it is far
better to look at the rate of change of the quantity of money.2
Employment as a criterion of policy
The second limitation I wish to discuss goes more against the grain of
current thinking. Monetary growth, it is widely held, will tend to stimulate
employment; monetary contraction, to retard employment. Why, then,
cannot the monetary authority adopt a target for employment or
unemployment—say, 3 percent unemployment; be tight when unemployment
is less than the target; be easy when unemployment is higher than the target;
and in this way peg unemployment at, say, 3 percent? The reason it cannot
is precisely the same as for interest rates—the difference between the
immediate and the delayed consequences of such a policy.
170 Milton Friedman
Thanks to Wicksell, we are all acquainted with the concept of a ‘natural’
rate of interest and the possibility of a discrepancy between the ‘natural’ and
the ‘market’ rate. The preceding analysis of interest rates can be translated
fairly directly into Wicksellian terms. The monetary authority can make the
market rate less than the natural rate only by inflation. It can make the
market rate higher than the natural rate only by deflation. We have added
only one wrinkle to Wicksell—the Irving Fisher distinction between the
nominal and the real rate of interest. Let the monetary authority keep the
nominal market rate for a time below the natural rate by inflation. That in
turn will raise the nominal natural rate itself, once anticipations of inflation
become widespread, thus requiring still more rapid inflation to hold down
the market rate. Similarly, because of the Fisher effect, it will require not
merely deflation but more and more rapid deflation to hold the market rate
above the initial ‘natural’ rate.
This analysis has its close counterpart in the employment market. At any
moment of time, there is some level of unemployment which has the property
that it is consistent with equilibrium in the structure of real wage rates. At
that level of unemployment, real wage rates are tending on the average to
rise at a ‘normal’ secular rate, i.e. at a rate that can be indefinitely
maintained so long as capital formation, technological improvements, etc.,
remain on their long-run trends. A lower level of unemployment is an
indication that there is an excess demand for labor that will produce upward
pressure on real wage rates. A higher level of unemployment is an indication
that there is an excess supply of labor that will produce downward pressure
on real wage rates. The ‘natural rate of unemployment’, in other words, is
the level that would be ground out by the Walrasian system of general
equilibrium equations, provided there is imbedded in them the actual
structural characteristics of the labor and commodity markets, including
market imperfections, stochastic variability in demands and supplies, the
cost of gathering information about job vacancies and labor availabilities,
the costs of mobility, and so on.3
You will recognize the close similarity between this statement and the
celebrated Phillips Curve. The similarity is not coincidental. Phillips’
analysis of the relation between unemployment and wage change is
deservedly celebrated as an important and original contribution. But,
unfortunately, it contains a basic defect—the failure to distinguish between
nominal wages and real wages—just as Wicksell’s analysis failed to
distinguish between nominal interest rates and real interest rates. Implicitly,
Phillips wrote his article for a world in which everyone anticipated that
nominal prices would be stable and in which that anticipation remained
unshaken and immutable whatever happened to actual prices and wages.
Suppose, by contrast, that everyone anticipates that prices will rise at a rate
of more than 75 per cent a year—as, for example, Brazilians did a few years
ago. Then wages must rise at that rate simply to keep real wages unchanged.
An excess supply of labor will be reflected in a less rapid rise in nominal
The role of monetary policy 171
wages than in anticipated prices,4 not in an absolute decline in wages. When
Brazil embarked on a policy to bring down the rate of price rise, and
succeeded in bringing the price rise down to about 45 percent a year, there
was a sharp initial rise in unemployment because under the influence of
earlier anticipations, wages kept rising at a pace that was higher than the
new rate of price rise, though lower than earlier. This is the result
experienced, and to be expected, of all attempts to reduce the rate of
inflation below that widely anticipated.5
To avoid misunderstanding, let me emphasize that by using the term
‘natural’ rate of unemployment, I do not mean to suggest that it is
immutable and unchangeable. On the contrary, many of the market
characteristics that determine its level are man-made and policy-made. In the
United States, for example, legal minimum wage rates, the Walsh-Healy and
Davis-Bacon Acts, and the strength of labor unions all make the natural rate
of unemployment higher than it would otherwise be. Improvements in
employment exchanges, in availability of information about job vacancies
and labor supply, and so on, would tend to lower the natural rate of
unemployment. I use the term ‘natural’ for the same reason Wicksell did—to
try to separate the real forces from monetary forces.
Let us assume that the monetary authority tries to peg the ‘market’ rate of
unemployment at a level below the ‘natural’ rate. For definiteness, suppose
that it takes 3 percent as the target rate and that the ‘natural’ rate is higher
than 3 percent. Suppose also that we start out at a time when prices have
been stable and when unemployment is higher than 3 percent. Accordingly,
the authority increases the rate of monetary growth. This will be
expansionary. By making nominal cash balances higher than people desire,
it will tend initially to lower interest rates and in this and other ways to
stimulate spending. Income and spending will start to rise.
To begin with, much or most of the rise in income will take the form of an
increase in output and employment rather than in prices. People have been
expecting prices to be stable, and prices and wages have been set for some
time in the future on that basis. It takes time for people to adjust to a new
state of demand. Producers will tend to react to the initial expansion in
aggregate demand by increasing output, employees by working longer hours,
and the unemployed, by taking jobs now offered at former nominal wages.
This much is pretty standard doctrine.
But it describes only the initial effects. Because selling prices of products
typically respond to an unanticipated rise in nominal demand faster than
prices of factors of production, real wages received have gone down—though
real wages anticipated by employees went up, since employees implicitly
evaluated the wages offered at the earlier price level. Indeed, the
simultaneous fall ex post in real wages to employers and rise ex ante in real
wages to employees is what enabled employment to increase. But the decline
ex post in real wages will soon come to affect anticipations. Employees will
start to reckon on rising prices of the things they buy and to demand higher
172 Milton Friedman
nominal wages for the future. ‘Market’ unemployment is below the ‘natural’
level. There is an excess demand for labor so real wages will tend to rise
toward their initial level.
Even though the higher rate of monetary growth continues, the rise in real
wages will reverse the decline in unemployment, and then lead to a rise,
which will tend to return unemployment to its former level. In order to keep
unemployment at its target level of 3 per cent, the monetary authority would
have to raise monetary growth still more. As in the interest rate case, the
‘market’ rate can be kept below the ‘natural’ rate only by inflation. And, as
in the interest rate case, too, only by accelerating inflation. Conversely, let
the monetary authority choose a target rate of unemployment that is above
the natural rate, and they will be led to produce a deflation, and an
accelerating deflation at that.
What if the monetary authority chose the ‘natural’ rate—either of interest
or unemployment—as its target? One problem is that it cannot know what
the ‘natural’ rate is. Unfortunately, we have as yet devised no method to
estimate accurately and readily the natural rate of either interest or
unemployment. And the ‘natural’ rate will itself change from time to time.
But the basic problem is that even if the monetary authority knew the
‘natural’ rate, and attempted to peg the market rate at that level, it would
not be led to a determinate policy. The ‘market’ rate will vary from the
natural rate for all sorts of reasons other than monetary policy. If the
monetary authority responds to these variations, it will set in train longer-
term effects that will make any monetary growth path it follows ultimately
consistent with the rule of policy. The actual course of monetary growth will
be analogous to a random walk, buffeted this way and that by the forces that
produce temporary departures of the market rate from the natural rate.
To state this conclusion differently, there is always a temporary trade-off
between inflation and unemployment; there is no permanent trade-off. The
temporary trade-off comes not from inflation per se, but from unanticipated
inflation, which generally means, from a rising rate of inflation. The
widespread belief that there is a permanent trade-off is a sophisticated
version of the confusion between ‘high’ and ‘rising’ that we all recognize in
simpler forms. A rising rate of inflation may reduce unemployment, a high
rate will not.
But how long, you will say, is ‘temporary’? For interest rates, we have
some systematic evidence on how long each of the several effects takes to
work itself out. For unemployment, we do not. I can at most venture a
personal judgement, based on some examination of the historical evidence,
that the initial effects of a higher and unanticipated rate of inflation last for
something like two to five years; that this initial effect then begins to be
reversed; and that a full adjustment to the new rate of inflation takes about
as long for employment as for interest rates, say, a couple of decades. For
both interest rates and employment, let me add a qualification. These
estimates are for changes in the rate of inflation of the order of magnitude
The role of monetary policy 173
that has been experienced in the United States. For much more sizeable
changes, such as those experienced in South American countries, the whole
adjustment process is greatly speeded up.
To state the general conclusion still differently, the monetary authority
controls nominal quantities—directly, the quantity of its own liabilities. In
principle, it can use this control to peg a nominal quantity—an exchange
rate, the price level, the nominal level of national income, the quantity of
money by one or another definition—or to peg the rate of change in a
nominal quantity—the rate of inflation or deflation, the rate of growth or
decline in nominal national income, the rate of growth of the quantity of
money. It cannot use its control over nominal quantities to peg a real
quantity—the real rate of interest, the rate of unemployment, the level of real
national income, the real quantity of money, the rate of growth of real
national income, or the rate of growth of the real quantity of money.
WHAT MONETARY POLICY CAN DO
Monetary policy cannot peg these real magnitudes at predetermined levels.
But monetary policy can and does have important effects on these real
magnitudes. The one is in no way inconsistent with the other.
My own studies of monetary history have made me extremely sympathetic
to the oft-quoted, much reviled, and as widely misunderstood, comment by
John Stuart Mill. ‘There cannot…,’ he wrote, ‘be intrinsically a more
insignificant thing, in the economy of society, than money; except in the
character of a contrivance for sparing time and labour. It is a machine for
doing quickly and commodiously, what would be done, though less quickly
and commodiously, without it: and like many other kinds of machinery, it
only exerts a distinct and independent influence of its own when it gets out of
order’ [7, p. 488].
True, money is only a machine, but it is an extraordinarily efficient
machine. Without it, we could not have begun to attain the astounding
growth in output and level of living we have experienced in the past two
centuries—any more than we could have done so without those other
marvelous machines that dot our countryside and enable us, for the most
part, simply to do more efficiently what could be done without them at much
greater cost in labor.
But money has one feature that these other machines do not share.
Because it is so pervasive, when it gets out of order, it throws a monkey
wrench into the operation of all the other machines. The Great Contraction
is the most dramatic example but not the only one. Every other major
contraction in the United States has been either produced by monetary
disorder or greatly exacerbated by monetary disorder. Every major inflation
has been produced by monetary expansion—mostly to meet the overriding
demands of war which have forced the creation of money to supplement
explicit taxation.
174 Milton Friedman
The first and most important lesson that history teaches about what
monetary policy can do—and it is a lesson of the most profound
importance—is that monetary policy can prevent money itself from being a
major source of economic disturbance. This sounds like a negative
proposition: avoid major mistakes. In part it is. The Great Contraction might
not have occurred at all, and if it had, it would have been far less severe, if
the monetary authority had avoided mistakes, or if the monetary
arrangements had been those of an earlier time when there was no central
authority with the power to make the kinds of mistakes that the Federal
Reserve System made. The past few years, to come closer to home, would
have been steadier and more productive of economic well-being if the
Federal Reserve had avoided drastic and erratic changes of direction, first
expanding the money supply at an unduly rapid pace, then, in early 1966,
stepping on the brake too hard, then, at the end of 1966, reversing itself and
resuming expansion until at least November, 1967, at a more rapid pace
than can long be maintained without appreciable inflation.
Even if the proposition that monetary policy can prevent money itself
from being a major source of economic disturbance were a wholly negative
proposition, it would be none the less important for that. As it happens,
however, it is not a wholly negative proposition. The monetary machine has
gotten out of order even when there has been no central authority with
anything like the power now possessed by the Fed. In the United States, the
1907 episode and earlier banking panics are examples of how the monetary
machine can get out of order largely on its own. There is therefore a positive
and important task for the monetary authority—to suggest improvements in
the machine that will reduce the chances that it will get out of order, and to
use its own powers so as to keep the machine in good working order.
A second thing monetary policy can do is provide a stable background for
the economy—keep the machine well oiled, to continue Mill’s analogy.
Accomplishing the first task will contribute to this objective, but there is
more to it than that. Our economic system will work best when producers
and consumers, employers and employees, can proceed with full confidence
that the average level of prices will behave in a known way in the future–
preferably that it will be highly stable. Under any conceivable institutional
arrangements, and certainly under those that now prevail in the United
States, there is only a limited amount of flexibility in prices and wages. We
need to conserve this flexibility to achieve changes in relative prices and
wages that are required to adjust to dynamic changes in tastes and
technology. We should not dissipate it simply to achieve changes in the
absolute level of prices that serve no economic function.
In an earlier era, the gold standard was relied on to provide confidence in
future monetary stability. In its heyday it served that function reasonably
well. It clearly no longer does, since there is scarce a country in the world
that is prepared to let the gold standard reign unchecked—and there are
persuasive reasons why countries should not do so. The monetary authority
The role of monetary policy 175
could operate as a surrogate for the gold standard, if it pegged exchange
rates and did so exclusively by altering the quantity of money in response to
balance of payment flows without ‘sterilizing’ surpluses or deficits and
without resorting to open or concealed exchange control or to changes in
tariffs and quotas. But again, though many central bankers talk this way,
few are in fact willing to follow this course—and again there are persuasive
reasons why they should not do so. Such a policy would submit each country
to the vagaries not of an impersonal and automatic gold standard but of the
policies—deliberate or accidental—of other monetary authorities.
In today’s world, if monetary policy is to provide a stable background for
the economy it must do so by deliberately employing its powers to that end.
I shall come later to how it can do so.
Finally, monetary policy can contribute to offsetting major disturbances in
the economic system arising from other sources. If there is an independent
secular exhilaration—as the postwar expansion was described by the
proponents of secular stagnation—monetary policy can in principle help to
hold it in check by a slower rate of monetary growth than would otherwise
be desirable. If, as now, an explosive federal budget threatens unprecedented
deficits, monetary policy can hold any inflationary dangers in check by a
slower rate of monetary growth than would otherwise be desirable. This will
temporarily mean higher interest rates than would otherwise prevail—to
enable the government to borrow the sums needed to finance the deficit—but
by preventing the speeding up of inflation, it may well mean both lower
prices and lower nominal interest rates for the long pull. If the end of a
substantial war offers the country an opportunity to shift resources from
wartime to peacetime production, monetary policy can ease the transition by
a higher rate of monetary growth than would otherwise be desirable—though
experience is not very encouraging that it can do so without going too far.
I have put this point last, and stated it in qualified terms—as referring to
major disturbances—because I believe that the potentiality of monetary
policy in offsetting other forces making for instability is far more limited
than is commonly believed. We simply do not know enough to be able to
recognize minor disturbances when they occur or to be able to predict either
what their effects will be with any precision or what monetary policy is
required to offset their effects. We do not know enough to be able to achieve
stated objectives by delicate, or even fairly coarse, changes in the mix of
monetary and fiscal policy. In this area particularly the best is likely to be
the enemy of the good. Experience suggests that the path of wisdom is to use
monetary policy explicitly to offset other disturbances only when they offer a
‘clear and present danger’.
HOW SHOULD MONETARY POLICY BE CONDUCTED?
How should monetary policy be conducted to make the contribution to our
goals that it is capable of making? This is clearly not the occasion for
176 Milton Friedman
presenting a detailed ‘Program for Monetary Stability’—to use the title of a
book in which I tried to do so [3]. I shall restrict myself here to two major
requirements for monetary policy that follow fairly directly from the
preceding discussion.
The first requirement is that the monetary authority should guide itself by
magnitudes that it can control, not by ones that it cannot control. If, as the
authority has often done, it takes interest rates or the current unemployment
percentage as the immediate criterion of policy, it will be like a space
vehicle that has taken a fix on the wrong star. No matter how sensitive and
sophisticated its guiding apparatus, the space vehicle will go astray. And so
will the monetary authority. Of the various alternative magnitudes that it
can control, the most appealing guides for policy are exchange rates, the
price level as defined by some index, and the quantity of a monetary total–
currency plus adjusted demand deposits, or this total plus commercial bank
time deposits, or a still broader total.
For the United States in particular, exchange rates are an undesirable
guide. It might be worth requiring the bulk of the economy to adjust to the
tiny percentage consisting of foreign trade if that would guarantee freedom
from monetary irresponsibility—as it might under a real gold standard. But
it is hardly worth doing so simply to adapt to the average of whatever
policies monetary authorities in the rest of the world adopt. Far better to let
the market, through floating exchange rates, adjust to world conditions the 5
per cent or so of our resources devoted to international trade while reserving
monetary policy to promote the effective use of the 95 per cent.
Of the three guides listed, the price level is clearly the most important in its
own right. Other things the same, it would be much the best of the
alternatives—as so many distinguished economists have urged in the past. But
other things are not the same. The link between the policy actions of the
monetary authority and the price level, while unquestionably present, is more
indirect than the link between the policy actions of the authority and any of
the several monetary totals. Moreover, monetary action takes a longer time to
affect the price level than to affect the monetary totals and both the time lag
and the magnitude of effect vary with circumstances. As a result, we cannot
predict at all accurately just what effect a particular monetary action will
have on the price level and, equally important, just when it will have that
effect. Attempting to control directly the price level is therefore likely to make
monetary policy itself a source of economic disturbance because of false stops
and starts. Perhaps, as our understanding of monetary phenomena advances,
the situation will change. But at the present stage of our understanding, the
long way around seems the surer way to our objective. Accordingly, I believe
that a monetary total is the best currently available immediate guide or
criterion for monetary policy—and I believe that it matters much less which
particular total is chosen than that one be chosen.
A second requirement for monetary policy is that the monetary authority
avoid sharp swings in policy. In the past, monetary authorities have on
The role of monetary policy 177
occasion moved in the wrong direction—as in the episode of the Great
Contraction that I have stressed. More frequently, they have moved in the
right direction, albeit often too late, but have erred by moving too far. Too
late and too much has been the general practice. For example, in early 1966,
it was the right policy for the Federal Reserve to move in a less expansionary
direction—though it should have done so at least a year earlier. But when it
moved, it went too far, producing the sharpest change in the rate of
monetary growth of the postwar era. Again, having gone too far, it was the
right policy for the Fed to reverse course at the end of 1966. But again it
went too far, not only restoring but exceeding the earlier excessive rate of
monetary growth. And this episode is no exception. Time and again this has
been the course followed—as in 1919 and 1920, in 1937 and 1938, in 1953
and 1954, in 1959 and 1960.
The reason for the propensity to overreact seems clear: the failure of
monetary authorities to allow for the delay between their actions and the
subsequent effects on the economy. They tend to determine their actions by
today’s conditions—but their actions will affect the economy only six or nine
or twelve or fifteen months later. Hence they feel impelled to step on the
brake, or the accelerator, as the case may be, too hard.
My own prescription is still that the monetary authority go all the way in
avoiding such swings by adopting publicly the policy of achieving a steady
rate of growth in a specified monetary total. The precise rate of growth, like
the precise monetary total, is less important than the adoption of some stated
and known rate. I myself have argued for a rate that would on the average
achieve rough stability in the level of prices of final products, which I have
estimated would call for something like a 3 to 5 percent per year rate of
growth in currency plus all commercial bank deposits or a slightly lower
rate of growth in currency plus demand deposits only.6 But it would be better
to have a fixed rate that would on the average produce moderate inflation or
moderate deflation, provided it was steady, than to suffer the wide and
erratic perturbations we have experienced.
Short of the adoption of such a publicly stated policy of a steady rate of
monetary growth, it would constitute a major improvement if the monetary
authority followed the self-denying ordinance of avoiding wide swings. It is a
matter of record that periods of relative stability in the rate of monetary
growth have also been periods of relative stability in economic activity, both
in the United States and other countries. Periods of wide swings in the rate of
monetary growth have also been periods of wide swings in economic
activity.
By setting itself a steady course and keeping to it, the monetary authority
could make a major contribution to promoting economic stability. By
making that course one of steady but moderate growth in the quantity of
money, it would make a major contribution to avoidance of either inflation
or deflation of prices. Other forces would still affect the economy, require
change and adjustment, and disturb the even tenor of our ways. But steady
178 Milton Friedman
monetary growth would provide a monetary climate favorable to the
effective operation of those basic forces of enterprise, ingenuity, invention,
hard work, and thrift that are the true springs of economic growth. That is
the most that we can ask from monetary policy at our present stage of
knowledge. But that much—and it is a great deal—is clearly within our
reach.
ACKNOWLEDGEMENTS
This chapter was the presidential address delivered at the Eightieth Annual
Meeting of the American Economic Association, Washington, DC, 29
December 1967.
I am indebted for helpful criticisms of earlier drafts to Armen Alchian,
Gary Becker, Martin Bronfenbrenner, Arthur F.Burns, Phillip Cagan, David
D.Friedman, Lawrence Harris, Harry G.Johnson, Homer Jones, Jerry
Jordan, David Meiselman, Allan H.Meltzer, Theodore W.Schultz, Anna J.
Schwartz, Herbert Stein, George J.Stigler and James Tobin.
NOTES
1 In [2], I have argued that Henry Simons shared this view with Keynes, and that it
accounts for the policy changes that he recommended.
2 This is partly an empirical not theoretical judgment. In principle, ‘tightness’ or
‘ease’ depends on the rate of change of the quantity of money supplied compared
to the rate of change of the quantity demanded excluding effects on demand from
monetary policy itself. However, empirically demand is highly stable, if we exclude
the effect of monetary policy, so it is generally sufficient to look at supply alone.
3 It is perhaps worth noting that this ‘natural’ rate need not correspond to equality
between the number unemployed and the number of job vacancies. For any given
structure of the labor market, there will be some equilibrium relation between
these two magnitudes, but there is no reason why it should be one of equality.
4 Strictly speaking, the rise in nominal wages will be less rapid than the rise in
anticipated nominal wages to make allowance for any secular changes in real
wages.
5 Stated in terms of the rate of change of nominal wages, the Phillips Curve can be
expected to be reasonably stable and well defined for any period for which the
average rate of change of prices, and hence the anticipated rate, has been relatively
stable. For such periods, nominal wages and ‘real’ wages move together. Curves
computed for different periods or different countries for each of which this condition
has been satisfied will differ in level, the level of the curve depending on what the
average rate of price change was. The higher the average rate of price change, the
higher will tend to be the level of the curve. For periods or countries for which the
rate of change of prices varies considerably, the Phillips Curve will not be well
defined. My impression is that these statements accord reasonably well with the
experience of the economists who have explored empirical Phillips Curves.
Restate Phillips’ analysis in terms of the rate of change of real wages—and even
more precisely, anticipated real wages—and it all falls into place. That is why
The role of monetary policy 179
students of empirical Phillips Curves have found that it helps to include the rate of
change of the price level as an independent variable.
6 In an article on The Optimum Quantity of Money’, I conclude that a still lower rate
of growth, something like 2 percent for the broader definition, might be better yet
in order to eliminate or reduce the difference between private and total costs of
adding to real balances. [Article was published in Friedman (1969) The Optimum
Quantity of Money and Other Essays, Chicago: Aldine.]
REFERENCES
1 H.S.Ellis (ed.) (1948) A Survey of Contemporary Economics, Philadelphia, PA.
2 Milton Friedman (1967) ‘The Monetary Theory and Policy of Henry Simons’,
Journal of Law and Economics Oct., 10:1–13.
3 ——(1959) A Program for Monetary Stability, New York.
4 E.A.Goldenweiser (1945) ‘Postwar Problems and Policies’, Federal Reserve Bulletin
Feb., 31:112–21.
5 P.T.Homan and Fritz Machlup (eds) (1945) Financing American Prosperity, New
York.
6 A.P.Lerner and F.D.Graham (eds) (1946) Planning and Paying for Full Employment,
Princeton, NJ.
7 J.S.Mill (1929) Principles of Political Economy, Bk III, Ashley ed., New York.
8 The structure of monetarism
Thomas Mayer
Kredit und Kapital (1975) 8, pp. 191–215, 292–313
In recent years the term ‘monetarism’ has come into vogue.1 Defined in a
very narrow sense it is the view that changes in the money stock are the
predominate factor explaining changes in money income, and hence is
merely a new term for ‘quantity theory’. But used in a broader sense the term
‘monetarism’ encompasses a number of other propositions apart from the
quantity theory of money. Unfortunately, this whole set of views is
commonly judged as a single unit. This contributes to an unfortunate
division of economists into monetarists and Keynesian schools with a
resulting polarization. It is my impression that the Keynesians have a
predisposition to reject all monetarist propositions on the basis of their ‘guilt
by association’ with other monetarist propositions, while monetarists have
the opposite tendency. I will therefore try to do two things in this chapter.
One is to show the interrelations between the various monetarist
propositions, and to illustrate that they do indeed form a coherent whole.
The other is to show that despite this, the connection between various
monetarist propositions is loose enough so that one can judge each one on its
own merits rather than having to accept or reject monetarist doctrine as a
whole. However, I will not try to judge the validity of monetarism.
To do this it is necessary as a first step to define the set of propositions
which characterize monetarists and distinguish them from Keynesians.
Unfortunately there is no single place where one can find a listing of all
monetarist propositions, and I have therefore had to construct my own list.2
In doing so I have tried to err on the side of inclusiveness rather than
exclusiveness, and I am dealing therefore with monetarism in the broad sense
of a Weltanschauung. Any such listing is, of course, quite arbitrary and the
reader may want to add or to delete items from the following list.3
1 The quantity theory of money, in the sense of the predominance of the
impact of monetary factors on nominal income.
2 The monetarist model of the transmission process.
3 Belief in the inherent stability of the private sector.
4 Irrelevance of allocative detail for the explanation of short-run changes
in money income, and belief in a fluid capital market.
The structure of monetarism 181
5 Focus on the price level as a whole rather than on individual prices.
6 Reliance on small rather than large econometric models.
7 Use of the reserve base or similar measure as the indicator of monetary
policy.
8 Use of the money stock as the proper target of monetary policy.
9 Acceptance of a monetary growth rule.
10 Rejection of an unemployment-inflation trade-off in favor of a real
Phillips-curve.
11 A relatively greater concern about inflation than about unemployment
compared to other economists.
12 Dislike of government intervention.
The first four of these items are ones listed by Karl Brunner in his description
of monetarism,4 while items 2 and 7–9 can be found in David Fand’s survey
of monetarism.5 On the other side of the debate James Tobin has
characterized it by items 1, 7, 8, 9 and 10 in the above list.6 Item 5, the
focus on the price level as a whole, while usually not explicit, is implicit in
typical monetarist discussion of inflation, particularly in their rejection of
cost-push inflation. Item 6, the preference for small models, while certainly
not a basic part of monetarist doctrine, is something which most monetarists
seem to have in common. Item 10, the real Phillips curve, is listed by
Leonall Andersen.7 Item 11, concern about inflation, is admittedly a rather
questionable item, based only on my general impression of monetarist
writings and verbal tradition.8 The final item, dislike of government
regulation, is a view that seems to be generally shared by monetarists, at
least in the United States.
These twelve items are, of course, not all equally significant. The first
four are the basic ones and can be used to define monetarism. A monetarist
need not accept any of the other eight.9 But monetarists do tend to accept
these other eight propositions too. And my purpose here is to describe a set of
beliefs which are shared by economists who call themselves monetarists (and
to a much lesser extent by other economists) rather than to set out a set of
beliefs which are the sufficient and necessary conditions for an economist to
be called a monetarist.10
The way I will now proceed is to start with the quantity theory, and then
take up each of the other components in the order listed, and see to what
extent they are dependent or independent of the previously discussed
components. In the first part of this chapter I will deal with the first six
propositions. In the second part, I will deal with propositions 7–12 and will
summarize the results obtained in both parts.
I THE QUANTITY THEORY
The quantity theory is the most basic component of monetarism. By the
quantity theory I mean the proposition that changes in the money stock are
182 Thomas Mayer
the dominant determinant of changes in money income.11 This is a very
general version which does not commit one to a specific theory of the
transmission process, a process treated separately in the next section. A very
important aspect of the quantity theory-Keynesian dispute involves the speed
of adaption of the economy.12 Keynesians would not—or at least should
not—deny that in the long run changes in nominal income are dominated by
changes in the money stock.
The above definition of the quantity theory clearly fits both the
Friedmanian and the Brunner-Meltzer versions. It is not at all clear, however,
that it also fits the Patinkin version. This is due to two characteristics of
Patinkin’s model. First, while changes in the stock of money ultimately bring
about equivalent changes in the price level we are not told how long this
process takes. Hence someone may completely accept the Patinkin model,
and yet, in forecasting next year’s money income might not pay very much
attention to recent changes in the money stock because these changes will
have their effects only in some far-off equilibrium.13 Second, while Patinkin’s
model tells us that changes in the money supply have proportional effects on
money income, this does not necessarily deny that changes in other variables
also affect money income. And if changes in these other variables have
important effects on income, then the essential monetarist proposition that
variations in money income are explained mainly by changes in the money
stock need no longer hold. Thus a Patinkian might well use a Keynesian
model for ordinary forecasting purposes instead of a quantity theory model.
This is not to deny that Patinkin’s model is a quantity theory model, but it
is a quantity theory model in a different sense from the way I am defining
the quantity theory here. His model has, to a large extent, a quantity
theorist’s ‘engine of analysis’,14 but the conclusions he reaches are not
necessarily those of the quantity theory in the short run as distinct from the
long run.15 Since monetarism is a policy-oriented doctrine, concerned very
much with the short run, Patinkin’s version of the quantity theory can be
excluded from it.
II THE TRANSMISSION PROCESS
The monetarist’s version of the transmission process by which changes in the
money stock affect income follows naturally from his research strategy
which is to focus on the supply and demand for real money balances.16 If the
public finds itself with excess balances it will reduce them by increasing
expenditures, presumably on both goods and bonds. By contrast, the
Keynesian focuses on relative yields, and therefore phrases the story
differently. If the public has excess money balances this must mean that the
yield on its money balances is less than the yield it can obtain on other
assets, and hence it buys other assets. Such a portfolio realignment to bring
yields (adjusted for risk, etc.) into equality is likely to involve primarily
assets which are similar to money, that is securities rather than goods.
The structure of monetarism 183
Hence, monetarists and Keynesians typically have a different range of assets
in mind when they think of the transmission process. This difference is
illustrated by the Keynesian calling the price of money the interest rate since
he thinks of money as a fund which can be either held as money or lent,
while the monetarist thinks of the price of money as the inverse of the price
level, since money is used to buy goods.17
Unfortunately, this genuine dispute, as well as disputes relating to the
measurement problem discussed below, are often obscured by a spurious
dispute about whether money affects income ‘directly’ or only ‘indirectly’.
This difference is terminological; one can reformulate the monetarist story in
terms of the interest rate and the Keynesian story in a way that omits the
interest rate. An increase in the real stock of money lowers the imputed real
interest rate on money balances. Hence, a monetarist, instead of saying that
the public has more money than it wants to hold, and thus increases
expenditures, can say that the public’s imputed interest rate on money
holdings has fallen while the yield on other assets is constant. Thus, directly
to equalize marginal yields, and indirectly because of the increase in the
money stock, the public increases expenditures. Conversely, the Keynesian
can use his liquidity preference diagram to show that an increase in the
public’s money stock means that the public is now holding more than its
optimal stock of money, and hence, to equalize rates of return on the margin,
buys securities. Essentially the point here is the following: given a demand
curve, whether for a commodity such as apples, or for the holding of money,
we can described any change either in terms of price (the interest rate) or in
terms of quantities (the stock of money). As long as we have a given demand
curve it does not matter; we must get the same answer regardless of which
axis of the diagram we look at. Hence, on a level of formal theory where
one can ignore measurement problems, it is unimportant whether one
formulates the analysis in terms of the money stock or in terms of the interest
rate.18 This dispute is spurious. It is therefore not surprising that Y.C.Park in
his careful survey of the transmission process concluded that ‘at the level of
general description there appear to be no significant differences in the
transmission process of monetary influences among a variety of monetary
economists’.19
A genuine aspect of the dispute, however, relates to the stability of the
demand for money which is part of the previously discussed hypothesis that
nominal income changes are dominated by changes in the money stock. If
the demand for money is unstable (in a numerical sense), perhaps because of
shifts in the marginal efficiency of investment, then knowledge that the
supply of money has increased no longer allows us to predict with any
degree of confidence that expenditures will actually increase. This is so
regardless of whether one phrases the process in Keynesian or monetarist
terms. The real difference between the two schools is that the Keynesian
tends to take the possibility of an unstable demand for money much more
seriously than does the monetarist, in part because he has a different theory
184 Thomas Mayer
of the interest rate.20 Hence, in predicting expenditures the Keynesian prefers
to look at what is happening to the rate of interest, thus taking account of
changes in both the demand for, and the supply of, money. The monetarist,
on the other hand, though he would agree in principle that changes in the
money supply may give a misleading answer because of changes in money
demand, does not treat this danger as seriously as does the Keynesian.
However, one must beware of exaggerating this difference. Although in
the General Theory Keynes did give the impression that the demand for
money is highly unstable, modern Keynesians no longer seem to believe this,
and instead treat the demand for money as fairly stable. On the other hand,
Friedman has stated that the quantity theorist looks upon the demand for
money as being a stable function of other variables, rather than as
necessarily being stable in a numerical sense.21
Since changes in the interest rate register demand as well as supply
shifts, they clearly have more information content than changes in the
money supply. One might, therefore, ask why anyone would look at the
money supply rather than at the rate of interest. This question brings us to
the second substantive issue, the measurement problem. The above
discussion has assumed implicity that both ‘the’ interest rate and ‘the’
money stock can be measured without error, or that they are measured with
equivalent errors. But this is questionable. Monetarists prefer to use the
money stock rather than the rate of interest because they believe that the
money stock can be measured much better. The term ‘the rate of interest’ as
used in formal theory is a theoretical term, and for any empirical work
with it it is necessary to find an accurately measurable counterpart. The
monetarist typically believes that this creates insuperable difficulties. One
difficulty is that ‘the’ rate of interest is an amalgam of a vast number of
specific long-term and short-term rates, and that there is no clear way in
which these rates can all be combined into a single measure. Term
structure theory is not a completely reliable guide. The second difficulty is
that by no means all the rates which should be combined into ‘the’ interest
rate can be observed in the market. Imputed rates used internally by
households and firms should be included, and due allowance should also be
made for borrowing costs other than the measured interest rate, for
example the cost of deteriorating balance sheet ratios. Third, what is
relevant for economic decisions is the expected real rate of interest, which
cannot be observed in the market, and cannot be approximated reliably by
econometric techniques. Since changes in the inflation rate are frequently
large relative to changes in the real interest rate, changes in the nominal
rate may be a very poor guide to changes in the expected real rate. Hence,
monetarists argue, in practice the money stock is a much better measuring
rod than is the interest rate.
It is, of course, open for Keynesians to reply that the money stock is also
measured badly. Again the problem is that the theoretical term, ‘money’, as
used in the quantity theory does not have a clear-cut empirical counterpart.
The structure of monetarism 185
Should it be approximated by M1 or M2? This is an issue on which
monetarists disagree among themselves. 22 Presumably, the proper
counterpart is some weighted mean, but there exists no reliable way of
estimating it.23 Furthermore, as in the case of the interest rate one should
make some adjustment for the anticipated inflation rate. Surely, it does
affect how the public feels about the adequacy of its cash balances. Hence,
it is open to the Keynesian to argue that despite the difficulties of
measuring ‘the’ rate of interest, it can be measured more accurately than
‘the’ money stock.
Problems of measuring the money stock are likely to seem more serious to
a Keynesian than to a quantity theorist because someone who believes that
the money stock cannot be measured accurately is likely to be skeptical of
the empirical evidence claiming to show that changes in the money stock
explain changes in money income. But it does not necessarily follow from
this that a Keynesian need be more worried about the difficulty of measuring
the money supply than about measuring the interest rate. He may well take
the position that, while neither variable can be measured accurately, the
interest rate is measured with a greater error than is the money stock. There
is certainly nothing in Keynesian theory to deny this. The exposition of the
argument in terms of the interest rate rather than the money stock, both in
the General Theory and the subsequent Keynesian literature, can often be
explained by the argument being on a high level of abstraction where
measurement problems can be ignored. Thus, while it is hard to see why a
quantity theorist would prefer to use the interest rate in his description of the
transmission process, it is not hard to see why a Keynesian may agree with a
quantity theorist in looking at the money stock rather than the rate of
interest.
A third substantive difference between the Keynesian and monetarist
transmission processes relates to the range of assets considered. The
monetarist looks at an increase in the money supply as having raised the
public’s money holdings relative to its holdings of securities and all types of
real assets. Hence, to bring marginal yields into equilibrium the public now
spends these excess balances to acquire securities, capital goods and
consumer goods. The Keynesian, however, typically treats the increase in the
money stock as affecting only investment, and not consumption.24 There are
two reasons for this. First, by looking at the interest rate the Keynesian
adopts a borrowing-cost interpretation; an increase in the money stock
lowers interest rates, and this lower cost of borrowing stimulates demand for
goods which are bought with credit; that is, it stimulates business investment,
residential construction, and perhaps investment in consumer durables.25
Demand for nondurables is not directly affected because they are usually not
bought on credit. A second reason is that the Keynesian often makes the
simplifying assumption that the propensity to consume is not directly affected
by the interest rate, so that an increase in the money stock affects only
investment.26
186 Thomas Mayer
How does this difference in the range of assets relate to the magnitude of
the impact of changes in the money stock, and hence to the question whether
changes in money income are dominated by changes in the money stock? On
a level of rather causal empiricism there is a direct relationship. If monetary
changes affect consumption as well as investment then money probably has a
much great effect on income than is the case if it can affect only ‘investment’
including perhaps consumer durables.27 But this reasoning while suggestive is
hardly conclusive. Someone might accept the Keynesian transmission
process, believing that changes in the money stock operate only via
investment, and yet he might think that, due to a high interest elasticity of
investment, this effect is very powerful. On the other hand, someone might
believe that changes in the stock of money affect both consumption and
investment, but that this total effect is quite weak.
Another substantive difference is newer. Karl Brunner and Allan Meltzer
have developed a new version of the monetarist transmission process.28 They
argue that the Friedmanian version, which is really what was discussed
above, is essentially Keynesian in its underlying theory, and they have set
out a theoretical critique of this Keynesian transmission process. It focuses on
a relative price process and stock effects which tend to bring the system
towards a classical rather than a Keynesian equilibrium.29
Thus there are four links between the hypothesis of the primacy of
changes in the quantity of money and the monetarist—as opposed to the
Keynesian—version of the transmission process. One is the stability of the
demand for money, the second is the relative measurability of money and
interest rates, the third is the range of assets considered, and the fourth
concerns the relative price effects and stock effects discussed by Brunner and
Meltzer.
Are these links compelling in the sense that someone who accepts the
monetarist story on one must also accept it on the other? The answer is, no.
Clearly, one can accept the Keynesian version of the transmission process
and yet believe that monetary factors dominate money income. All one has
to do is to believe that the interest elasticity is high for investment and low
for the liquidity preference function. Conversely, one can accept the
monetarist transmission process, and yet reject the quantity theory as an
explanation of most observed changes in income. Thus, while the demand
for money may be relatively stable (compared to the seriousness of the errors
introduced by the measurement problem), the stock of money may be even
more stable. And while someone who believes in the primacy of the
monetary impulse is likely to believe that money can be measured fairly
well, he could also believe that the interest rate can be measured just as well
or better. Moreover, changes in the quantity of money could exert all their
(strong) effects on income through investment. Finally, someone may
consider the Brunner-Meltzer analysis of the relative price and stock effects
to be valid, but might believe that in the short run and intermediate run these
effects are relatively minor. In other words, one cannot logically infer how
The structure of monetarism 187
money affects income from the strength of the monetary impulse and vice
versa.
III STABILITY OF THE PRIVATE SECTOR
Monetarists generally believe that the private sector is inherently stable if
left to its own devices and not disturbed by an erratic monetary growth rate.
Many, probably most, Keynesians deny this. The nature of this dispute is
complex. Keynesians typically do not deny that the private sector is stable in
the sense that it is damped rather than explosive. As Lawrence Klein has
pointed out, some leading Keynesian econometric models show the economy
to be stable in its response to stochastic shocks.30 However, Keynesians look
upon the private sector as being unstable in another sense. This is that it is
inherently subject to erratic shocks, primarily due to changes in the marginal
efficiency of investment. To a Keynesian many factors can, and do, cause
substantial changes in aggregate demand, changes which may then lead to
damped oscillations.
By contrast, monetarists treat aggregate demand as the resultant of a
stable demand for money and an unstable supply of money. They look upon
the private sector as stable because its demand for money is stable, and
attribute most, though certainly not all, the actually observed instability to
fluctuations in the money supply induced by the monetary authorities.31
Thus, this dispute about the stability of the private sector is tied directly into
the basic dispute about the quantity theory, the extent to which changes in
aggregate demand are explained primarily by changes in the money supply
rather than by changes in the marginal efficiency of investment, etc.32
But even so, the tie between the quantity theory and the stability of the
private sector is not complete; someone can reject the quantity theory, and
yet believe in the inherent stability of the private sector. For example, a
Keynesian who believes that fiscal policy is so badly timed that it is
destabilizing, and that monetary policy has also not been a net stabilizer,
would have to believe that the private sector is stabler than is indicated by
the actually observed fluctuations in GNP. Yet there is nothing about such a
view which is contrary to Keynesian theory, or which requires the quantity
theory as its foundation. Thus one can be a Keynesian in one’s basic theory,
and, at the same time, accept the monetarist proposition that the private
sector is inherently stable or at least stabler than the private and government
sectors combined. Admittedly, it is much harder to see how a quantity
theorist could believe in the instability of the private sector.
IV IRRELEVANCE OF ALLOCATIVE DETAIL AND BELIEF IN THE
FLUIDITY OF CAPITAL MARKETS
One of the points of distinction between the monetarists and the Keynesians
is that in trying to determine short-run changes in income the Keynesian,
188 Thomas Mayer
unlike the monetarist, typically focuses on what happens in particular sectors
of the economy. With unstable private sectors (in the sense defined above)
fluctuations can start in various sectors, or be conditioned by the particular
characteristics of a sector. For example, a rise in the interest rate may have
different effects on residential construction, and hence on total output, at a
time when mortgage lending institutions are already short of liquidity than
at a time when they have a large liquidity buffer. More fundamentally,
Keynesians predict, or explain, income by looking at expenditure motives in
each sector. Hence, they have to analyze each sector.
The monetarists, by contrast, look upon expenditures as determined by the
excess supply of, or demand for, real balances. They therefore have to look
at the behavior of only a single market, the market for real balances.33
The Keynesian’s concern with allocative detail, that is the behavior of
different sectors, is reinforced by a frequent tendency among Keynesians to
treat the capital market as imperfect so that capital rationing can occur.
Hence, in estimating aggregate demand Keynesians are not satisfied with
knowing the total amount of liquidity in the economy. They also want to
know the liquidity of specific sectors, such as financial institutions serving
the mortgage market.34 This emphasis on imperfect capital markets and
credit rationing is also connected with the common Keynesian emphasis on
borrowing conditions as the only channel through which monetary policy
operates.35 Hence, they want to know a great deal about various interest
rates and financial markets in assessing the influence of monetary factors on
money income. And their belief that capital markets are imperfect explains
why Keynesians seem much more interested in flow of funds analysis than
are most monetarists, despite the fact that the flow of funds deals with the
monetarist’s item of central concern, money.
Another reason for the Keynesian emphasis on sectorial detail is probably
the tendency of many Keynesians to favor government intervention. Efficient
government intervention obviously requires detailed knowledge of many
sectors since the intervention is likely to focus on specific ‘troubles’ in
particular sectors. Finally, as will be discussed in the next section, many
Keynesians look upon inflation as sometimes being due, at least in part, to
developments in particular sectors rather than as due to the monetarist’s
single pervasive factor.
By contrast, in explaining short-run changes in income, monetarists
usually express little interest in allocative detail.36 They make a sharp
distinction between relative prices which are affected by the fortunes of
various sectors, and the general price level which is affected by the quantity
of money. They do not build up their estimate of national income by adding
up incomes in various sectors as Keynesians do, but rather, they work ‘from
the top down’. Using changes in the money stock they estimate total
expenditures, and then, if they happen to be interested in it, they might
investigate the allocation of this fixed expenditure total among various
sectors. Their assumption of a fluid capital market fortifies monetarists in
The structure of monetarism 189
their belief that a given increase in the money stock will have more or less
the same effect on aggregate incomes, though not of course, on the relative
incomes of various sectors, regardless of where it is injected.37 And their
belief in the stability of the private sector and in the absence of a need for
government intervention gives monetarists little incentive to focus their
attention on developments in various sectors.38 This is reinforced by the fact
that the monetarist, unlike the Keynesian, does not typically try to specify
the channels through which monetary factors operate, and hence does not try
to gauge the impact of monetary factors by looking at their impact on
different sectors.
Hence, monetarists’ disregard of allocative detail in explaining short-run
income changes is a natural outgrowth of their basic position. It results from
their belief in the quantity theory, i.e. in the primacy of money supply
changes in explaining income. It is also connected with their view of the
transmission process, in which expenditure motives and the peculiarities of
individual sectors are unimportant and the borrowing cost approach to
gauging the influence of monetary factors is rejected. But this does not mean
that monetarists must necessarily de-emphasize allocative detail in their
prediction of income fluctuations. Someone might accept all the other basic
and characteristic monetarist positions, and yet believe that the capital
market is highly imperfect, that capital rationing is important, and that the
flow of funds between various sectors therefore plays some role in
determining income.39 Similarly, monetarists might favor government
intervention either because they are skeptical of the stability of the private
sector, or because they favor government intervention for some other reason;
in principle one could certainly be a monetarist and also a socialist.
At the same time, a Keynesian need not believe in the imperfection of the
capital market and the importance of capital rationing. Neither of these
ideas plays a role in the General Theory. More significantly, one can accept
the general framework of Keynesian analysis without believing in the
instability of the private sector, and in the advisability of government
intervention, and hence not be concerned with allocative problems on these
grounds. It is only the Keynesian focus on expenditure motives that provided
a basic reason for the Keynesian’s interest in allocative detail.
V THE PRICE LEVEL VERSUS INDIVIDUAL PRICES
One major distinction between monetarists and most Keynesians is the way
of looking at the price level.40 This is a subtle distinction that is seldom, if
ever, made explicit. Basically there are two ways of approaching the price
level. One is to treat it as an aggregate phenomenon, determined by the
interaction of only two factors, aggregate demand and aggregate output.
This view draws a sharp distinction between the price level as a whole and
relative prices. Specific events in particular industries, such as an increase in
the degree of monopoly, union pressure, or bad harvests obviously affect
190 Thomas Mayer
relative prices. But they affect the price level only to the extent that they also
affect either aggregate demand or output. Thus if prices rise in industry A
without raising aggregate demand, this rise in the price of A has to be
matched either by a reduction of output, or by a decline in the average of all
other prices.
The alternative way of treating the price level is to approach it as the
weighted sum of individual prices. These prices are then explained by the
interaction of supply and demand in individual industries with the pricing
policies of various industries. Changes in aggregate demand are certainly not
ignored in this framework since they affect the demand curve faced by each
industry, but there is considerable emphasis on the particular behavior of
individual industries.
Both of these ways of looking at the price level are formally correct.
While, they must therefore yield the same answers to someone who possesses
all the required information, they do lead to different research strategies, and
are therefore likely in practice to provide different answers.
Monetarists clearly use the aggregative approach to the price level. They
look at changes in the quantity of money to determine changes in aggregate
demand, and then allocate changes in aggregate demand between changes in
prices and changes in output.41 In this approach, at least in its simple
version, the pricing decisions made by any particular industry have no effect
on the overall price level, but affect only relative prices.42 Hence, the
monetarist typically rejects cost-push explanations of inflation.
It might be worth noting in passing that this rejection of all cost-push
phenomena may well be unwarrented even within the monetarist framework.
If industry A (with an inelastic demand) raises its prices, and thus reduces the
aggregate demand that is available for other industries, these industries may
respond, at least in part, not by cutting prices, but by cutting output. Insofar
as this occurs, the general price level is raised by the behavior of industry A,
and not just the relative price of commodity A. The extent to which this
happens is an empirical question, and is likely to depend upon the degree of
inflation in the economy. If prices in general are rising then, as industry A
raises its prices other industries can adjust their prices for this merely by not
raising them by as much as they otherwise would. On the other hand, at a
time when prices are generally stable they would have to lower their prices
absolutely in order to offset the rise in the price of commodity A, and there is
considerable evidence that prices are sticky downward.
The monetarist’s macroeconomic, rather than microeconomic, approach
to the price level fits in well with two of the previously discussed
characteristics of monetarism. First, insofar as the rise in the price of one
particular industry results in a price decline in other industries the economic
system is inherently stable, at least as far as cost-push inflation is concerned.
Second, if the price behavior of individual industries has no effect on the
general price level, then this is one more reason for ignoring allocative
detail. However, it should be noted, that while monetarists’ approach to the
The structure of monetarism 191
price level therefore goes along well with their belief in the stability of the
private sector and the irrelevance of allocative detail, in neither case is the
relationship one of logical entailment. One can accept the monetarist
hypotheses about the irrelevance of allocative detail, and the stability of the
private sector, and yet, at the same time, accept the Keynesian approach to
the price level.43
The typical Keynesian’s view of the price level is quite different from the
monetarist view. To be sure, in the Keynesian model the price level is also
determined by aggregate demand and supply, but to Keynesians this
formulation is not useful because they cannot take aggregate demand as
given.44 The monetarist, by contrast, can do this; if industry A raises its
price, this does not change aggregate demand which depends upon the
money stock.45 But to the Keynesian the money stock is only one of several
factors determining aggregate demand. Thus while the rise in the price of
commodity A lowers the real money stock, it may also raise the marginal
efficiency of investment, particularly in industry A. In other words, while to
the monetarist aggregate demand, as determined by the quantity of money,
functions as a budget constraint, in the Keynesian system it is a variable.
Hence, to the Keynesian it is at least possible that a rise in the price of
commodity A raises aggregate demand enough so that other prices (and
outputs) will not have to fall, and might even rise.
Since the aggregate demand effects of a rise in the price of commodity A
are uncertain, the Keynesian is tempted to ignore them. And this temptation
is frequently not resisted. A typical example is a study by Otto Eckstein
and Gary Fromm in which they investigated the effect on the wholesale
price index of the rise in the price of steel. They considered both the direct
effect as well as the indirect effect of the steel price increase being passed
forward by steel users, and concluded that ‘if steel prices had behaved like
other industrial prices, the total wholesale price index would have risen by
40 percent less over the last decade’.46 To a monetarist such a statement
gives us only an arithmetic relationship which has no economic meaning
because it ignores aggregate demand, and hence other prices.47 And,
indeed, it is hard to see how a Keynesian can really justify ignoring the
indirect repercussions.
But the roots of this oversimplification can already be found in the
General Theory since Keynes looked upon prices as determined by the wage
rate and the marginal physical product of labor. Indeed Keynes specifically
tried to bring the theory of the price level into contact with microeconomic
factors such as marginal cost, and to eliminate the dichotomy between the
determination of individual prices by marginal cost etc., and of the price
level by macroeconomic factors such as the quantity of money and its
velocity. Thus he wrote in Chapter 21 of the General Theory: ‘One of the
objects of the foregoing chapters has been to escape from this double life and
bring the theory of price as a whole back to close contact with the theory of
value.’48
192 Thomas Mayer
This Keynesian tendency to look at the price level as determined by costs
in various industries has been furthered in recent years by an extensive
empirical literature which estimates prices more on the basis of shifts in costs
than on the basis of shifts in demand.49 (However, this evidence is not always
easy to interpret because changes in costs may be the result of changes in
demand.)50 In addition, it has probably gained in acceptability from the use,
as a first approximation or as an elementary teaching tool, of the Keynesian
supply curve dichotomized at full employment. If changes in aggregate
demand affect only output and not prices until full employment is reached,
then if one is trying to explain the price level under conditions of less than
full employment, the fact that a price rise in industry A changes the demand
experienced in other industries can be ignored.
But while many—perhaps most—Keynesians treat the price level in the
way just described, this way of looking at the price level is far from being a
necessary implication of the Keynesian model. A Keynesian could focus on
the overall price level rather than on its individual component prices to the
same extent as a monetarist does without abandoning any basic part of
Keynesian theory. As pointed out above, the only way a Keynesian can
ignore the effects of the rise in the price of commodity A on the demand left
over for other commodities is to assume that this rise in the price of
commodity A generates an exactly offsetting increase in demand. But there is
nothing in Keynesian theory that requires this to occur. The increase in the
price of commodity A reduces real balances thus lowering demand. To be
sure, this may be offset by an increase in the marginal efficiency of capital,
but this need not happen. The effect of the increase in the price of commodity
A on the marginal efficiency of investment may even be negative, or if it is
positive, it need not be great enough to offset all the effect of the decline in
real balances. Keynesian theory is silent on this. Strange as it may seem,
there appears to be virtually no Keynesian literature on the effect of a rise in
a particular price on income.51 It is, of course, true that a change in demand
for other commodities could affect the output of other commodities rather
than their prices, but whether this happens or not depends upon where we are
along the aggregate supply curve.52
Thus, this dispute about the determinants of the price level is not so much
a dispute between monetarism and Keynesianism as it is a dispute between
monetarism and a particular specification of Keynesianism. And while this
specification is a popular one, and is perhaps accepted by most Keynesians,
it represents only one line of development of the basic Keynesian model.
Moreover, monetarists too need not accept the typically monetarist
position discussed above. They may argue that while a rise in the price of
commodity A will eventually lower the prices of other commodities, in the
short run it will lower their outputs rather than their prices. Hence, a
Keynesian can accept the typically monetarist view on this issue, and a
monetarist can adopt the typically Keynesian view, without either one
abandoning the fundamental Keynesian or monetarist position.53
The structure of monetarism 193
VI LARGE VERSUS SMALL MODELS
While Keynesians usually prefer large-scale structural models, monetarists
prefer small reduced-form models.54 This dispute on model size involves
many issues which are extraneous to the monetarist debate. To a large extent
it is an issue in theoretical econometrics concerned with the validity of the
single equation approach, rather than an issue in monetary economics.
Moreover, as Friedman has pointed out, it involves also the question of
whether we know enough to be able to represent complex reality by the
greatly simplified systems used even by large models.55 Hence, Friedman
considers the debate about large versus small models to be ‘almost entirely
independent of the monetarist versus Keynesian point of view’.56
But even so, there are several ways in which the use of a reduced form
model goes along well with monetarist hypotheses. One way relates to the
transmission process. If changes in the money stock affect income through a
limited number of channels then it is tempting to cover each of these
channels, and thus to use a structural model. But if monetary changes affect
the economy in a very large number of ways, as the monetarist claims, then
even a large structural model is not likely to pick up all of them. Hence a
reduced-form approach is likely to be more reliable.
Second, one of the great advantages of large structural models is that they
provide detailed information on various economic sectors. This makes large
structural models attractive to Keynesians, who are interested in allocative
detail, but does little to recommend them to monetarists who are not interested
in allocative detail. Furthermore, by focusing on expenditure motives, and
looking upon people as being consumers, investors in inventories, etc., the
Keynesian is naturally concerned with many sectors. The monetarist, on the
other hand, is concerned with people only as money holders, and hence is
interested in only one sector, the supply of and demand for money. Third,
someone who is concerned about the instability of the private sector in the
sense that erratic shifts in expenditure incentives cause serious fluctuations, is
likely to believe that to predict income one needs a large model which allows
for the impact of these erratic factors on various sectors.
The relationship between the quantity theory per se and the choice of
structural models versus reduced-form models is much less clear. Ex ante,
there is little, if any, reason why someone who believes in the strength of the
monetary impulse, should necessarily believe in the desirability of reduced-
form models. But there is an ex post relationship due to the fact that the most
famous of all reduced form models, the Andersen-Jordan model, yields
monetarist conclusions while structural models generally yield Keynesian
conclusions. But the relationship between model size and the results obtained
from the model are far from firm. Edward Gramlich has shown that
Andersen-Jordan type models can generate not only monetarist results, but
also Keynesian, or inbetween, results depending on the monetary variables
used.57
194 Thomas Mayer
Thus there are many links between various monetarist propositions and a
preference for reduced form models. But as indicated above this linkage is
not strong. A monetarist might well reject the use of reduced form models,
while a Keynesian might prefer such models since the dispute is largely a
matter of choice of estimation technique.
This concludes the discussion of the six monetarist propositions which
relate to theory and techniques of analysis. In the second part I will discuss
the remaining six policy-oriented propositions.
In the first part I selected twelve propositions characterizing the monetarist
outlook, and discussed six of them. I will now discuss the remainder, and
then summarize both parts. Three of these propositions relate to monetary
policy. They are the choice of an indicator, the choice of a target, and the
use of a monetary growth rule.58
VII MONETARY INDICATORS
A monetary policy indicator is a variable that measures the thrust, that is the
direction and magnitude, of monetary policy. It should therefore be a
variable which is closely controlled by the central bank rather than being
endogenous to the economy. Accurate data on it should be available without
delay, and they should have a very high correlation with the target, or goal,
variables. These requirements rule out both the money stock and the long-
term interest rate as monetary indicators. To be sure, to a monetarist the
stock of money is the ultimate indicator of monetary policy in a different
sense, because changes in the money stock foretell changes in income. But, at
least in the United States, accurate data on the money stock are not available
quickly, and besides, the money stock is partly endogenous, being some
distance removed from central bank actions. Hence, it cannot be used as an
indicator as the term is defined in this context. Similarly, for the Keynesian
the long-term interest rate is not an adequate indicator because it is not
under the close control of the central bank. Thus, neither monetarists nor
Keynesians can use as their indicators those variables which would fit best
into their models. Both of them have to select other indicators which are
closer to the tools used by the central bank.
Monetarists favor some measure of total reserves such as the reserve base
adjusted for changes in reserve requirements or else unborrowed reserves.
These are clearly under the control of the central bank, they are measured
accurately without delay, and they have a powerful effect on the money stock,
the monetarists’ target variable. Keynesians, on the other hand, probably use
the short-term interest rate as their favored indicator.59 The short-term rate can
then be related to one of their target variables, the long-term interest rate, via
term structure theory. And in addition, the short-term rate is a target in its own
right to Keynesians since it affects flows into depository institutions, and hence
residential construction. But this does not make it an indicator.
The structure of monetarism 195
But it is important to note that the choice of a monetary policy indicator
is to a considerable extent isolated from the rest of the Keynesian-monetarist
dispute. Monetarists choose a monetary base measure for two reasons. One
is that their analysis of the money supply process tells them that this is the
variable which best reflects monetary policy actions. The second is that they
believe the monetary base (adjusted for reserve requirement changes) to be
the best indicator of future changes in the money stock. As far as the first of
these reasons is concerned this involves little dispute with Keynesians if only
because few Keynesians have bothered to formulate a money supply
hypothesis.
Turning to the second reason, the predictive power of a base measure, it is
certainly true that one can predict the money stock fairly well in this way.
But suppose that it were shown that changes in the short-term interest rate
are an even better indicator of changes in the money stock. In this case,
monetarists should use the short-term interest rate as their indicator to
predict the money stock. And the possibility that the short-term interest rate
is a better predictor of the money stock than are various reserve measures is
by no means farfetched.60 Furthermore, if it were somehow shown that
monetary policy changes are reflected better by the Federal Funds rate than
by a reserve base measure, monetarists could abandon their money supply
hypothesis without thereby weakening their belief in any of the other
monetarist propositions.
Conversely, Keynesians could select total reserves as their policy
indicator, and use this variable, rather than the short-term rate, to predict
long-term interest rates. The unsettled state of term structure theory hardly
provides us with much confidence in trying to predict the long-term rate on
the basis of the short-term rate. Empirically, David Fand has shown that
while there is a fairly high correlation between long-term and short-term
rates, ‘in a cyclical context, the long rate is relatively independent of the
short-run movements in the short rates’.61
In addition to its use in gauging policy, a monetary indicator can also be
used to measure the thrust of the monetary impulse regardless of whether this
arises in the private or public sector. For this a monetarist may want to use
the money stock, while a Keynesian may want to use a short-term interest
rate. Thus, if the money stock is growing at, say, a 10 percent rate, while the
Federal Funds rate is 12 percent, a monetarist would call this a situation of
monetary ease, while a Keynesian would call it tight money.
This distinction has some superficial relation to the dispute about the
transmission mechanism because the Keynesian is looking at an interest rate
while the monetarist is using the money stock. But, as discussed above, this
dispute is, in part, a matter of terminology rather than a genuine dispute.
(And, as will be shown below, in part it is the result of many Keynesians not
being faithful to their Wicksellian tradition.)
Another connection is that to the Keynesian the short-term interest rate is
a valid partial indicator because it affects the flow of funds into financial
196 Thomas Mayer
intermediaries, and hence residential mortgage lending and construction.
(This is a channel stressed strongly in the FMP-model.)
Thus, here we have a component of monetarism which has only a limited
relationship to the other components. The dispute about the proper indicator
is to a considerable extent an isolated technical issue. Its intrusion into the
monetarist-Keynesian debate can perhaps be explained as an historical
accident. In the past the Federal Reserve has used short-term interest rates
and money market conditions as its indicator in a different sense from the
way the indicator concept is defined here. Instead of treating short-term rates
and money market conditions as an intermediate step on the way to long-
term interest rates or to the money stock, it looked at short-term rates and
money market conditions as an immediate guide to how its policy is
affecting income. In this way—which does not allow the money stock to be a
recognized part of the process—the use of short-term rates and money market
conditions is, of course, contrary to monetarism. But as a result of the
insights which monetarists have brought to this debate indicators are no
longer thought of in this way.
VIII MONETARY POLICY TARGETS
Obviously monetarists want to use the money stock as the target of monetary
policy. Keynesians, on the other hand, prefer to use long-term interest rates
or, in some cases, bank credit or total credit. The extent to which each of
these targets fits into the underlying theories of both schools can be seen best
by considering the arguments for each of these targets.
To start with a comparison of the interest rate target and the money stock
target there is again the measurement problem previously discussed in
connection with the transmission process.
But with respect to the problem of chosing a target, the Keynesian is less
worried about the difficulties of measuring the interest rate. This is so
because one important Keynesian channel for the impact of monetary policy
operates through the flows of funds into depository institutions. And since
such flows depend upon a comparison of interest rates of depository
institutions with open market rates, the problem of infering the expected real
rate from the nominal interest rate does not arise. (And the problem of
combining various observed and imputed rates into ‘the interest rate’ is also
less serious.) Furthermore, another channel is the effect of interest rates on
the market value of the households’ stock of securities, and hence on
consumption. Here too, the problem of measuring the interest rate is not
serious. However, for the traditional ‘cost of capital’ effect of interest rates
on investment, the measurement problem still exists.
Apart from the measurement problem the choice of a target involves
another issue which arises from our inability to predict precisely changes in
the liquidity preference schedule and in expenditure incentives.62 If we would
know very accurately the liquidity preference curve as well as expenditure
The structure of monetarism 197
incentives, then the central bank could easily select the interest rate which
would optimize its objectives. Since with a known liquidity preference curve
we can infer a particular quantity of money for each rate of interest and vice
versa, leaving aside the above discussed measurement problem, it would be
a matter of complete indifference whether the central bank picks a particular
interest rate target or a money stock target.
But in actuality the central bank does not know the liquidity preference
schedule and the strength of the expenditure incentives accurately. Suppose
that the liquidity preference curve shifts outward unexpectedly. All the
central bank observes is a rise in the interest rate. If it uses an interest rate
target it responds to this rise in the interest rate by increasing the quantity of
money sufficiently to lower it back to its previous level.63 What it does is to
satisfy the increased demand for money, or in terms of the cash balance
equation, it offsets the rise in the Cambridge ‘k’ by raising ‘M’, thus keeping
‘PT’ constant. If it had used a money stock target instead of its interest rate
target, it would have kept the money stock constant and allowed the interest
rate to rise. This increase in the interest rate would then have reduced
income below its previous (presumably optimal) level.
On the other hand, suppose that the liquidity preference curve is
predictable, but that expenditure incentives increase unexpectedly.64 This too
raises the rate of interest. If the central bank has an interest rate target and
counteracts this rise in the interest rate, it allows income to rise in an
unintended way. In other words, if expenditure incentives increase the
interest rate should also increase, thus acting as an automatic stabilizer.
Hence, if it is expenditure motives rather than the liquidity preference
function which changes in an unpredicted way, then an interest rate target
does harm, and a money stock target is preferable. But if it is the liquidity
preference function which is the unpredictable one, then an interest rate
target is superior.65
On both of these issues a monetarist prefers a money stock target.
Regarding the measurement problem, someone who accepts the monetarist
transmission process believes that the money stock can be measured more
accurately than can the interest rate. On the relative predictability of the
liquidity preference function and the expenditure functions a quantity theorist
considers the liquidity preference function (i.e. the demand for money) to be
the stabler of the two.66 Hence, the monetarist’s preference for a money stock
target over an interest rate target can be seen as an implication of the
quantity theory and its transmission mechanism.
Apart from the money stock and the long-term interest rate there is a third
major potential target for monetary policy. This is a credit measure, such as
bank credit or total credit. Here too, the quantity theory and the monetarist’s
version of the transmission process decide the issue for the monetarist. As a
quantity theorist, the monetarist believes that the effect of changes in the
money stock on income is more important than the effect of changes in bank
credit, for otherwise he or she would hold a quantity theory of bank credit
198 Thomas Mayer
rather than a quantity theory of money. Moreover, in the analysis of the
transmission process the monetarist rejects a credit and borrowing cost
interpretation.67
The matter is more complex for Keynesians. As indicated above, the
problem of measuring the interest rate is of serious concern to them only
with respect to the cost of capital channel. Since different Keynesians attach
different weights to this channel, it is hard to say how significant the
measurement problem is for the Keynesian’s choice of a target. Furthermore,
a Keynesian may—or may not—be concerned about the difficulties of
measuring the money stock.
With regard to the second issue, the relative predictability of the liquidity
preference and expenditure functions, Keynes originally considered both the
liquidity preference function and the investment function to be erratic
without indicating which was the more unstable. Modern Keynesians, on the
other hand, have de-emphasized the speculative motive for liquidity
preference which for Keynes was the source of its instability, and appear to
believe that the liquidity preference function is fairly stable and predictable.
On the other hand, Keynesians also believe that investment and
consumption, while unstable, are predictable. It is therefore not really clear
whether Keynesians typically consider the liquidity preference function or the
expenditure functions to be the more predictable. Perhaps there is a
presumption that, on the whole, they consider the demand for money to be
the more predictable variable which should make them prefer a money stock
target.
Moreover, insofar as they are the intellectual heirs of the Wicksellian
tradition, Keynesians should prefer a money stock target to an interest rate
target. It was Wicksell who taught us the dangers of keeping the money rate
of interest fixed (as happens with an interest rate target) when the natural
rate of interest changes. All in all, Keynesian theory is more or less neutral
on the issue of the money stock versus the interest rate as the target.
The third potential target is the volume of bank credit, or total credit.
Some Keynesians have accepted such targets and they, of course, differ
sharply from the monetarists. But one can be a good Keynesian while
rejecting the reasoning of the Radcliffe Report.
IX THE MONETARY GROWTH RULE
The next component of monetarism is the constant money growth rule. Such
a rule fits well into the monetarist framework on several counts. First, it is
closely related to the quantity theory. If the demand for money is indeed
constant when adjusted for trend, then a constant growth rate of the supply
of money would result in income too growing at a constant rate.68 Hence,
someone who accepts the quantity theory of money is much more likely to
favor constant money growth than is someone who believes either that the
demand for money is unstable, or that fluctuations in income are largely due
The structure of monetarism 199
to nonmonetary factors, factors which the central bank can offset.69 Second,
a belief in constant money growth also fits in with the monetarist’s belief that
the private sector is inherently stable. If this is the case there is at best a
limited amount of good that could be accomplished by variations in the
money growth rate. Third, belief in a constant money growth rate requires
acceptance of a money stock target, for the monetary growth rule is really
only a special version of the use of a monetary target; it merely sets a
specific, unvarying target.
In addition, the constant money growth rule also has some connection,
albeit a looser connection, with two other components of monetarism, the
disinterest in allocative detail, and the monetarist view of the price level.
Someone who is interested in allocative detail is likely to be concerned, from
time to time, with the impact of financial stringency on a particular sector,
such as residential construction. He is therefore likely to feel, at least
occasionally, that the monetary growth rate should be changed to protect a
particular sector. A monetarist who believes that allocative detail is outside
the purview of macroeconomic stabilization policy is much less likely to feel
this way. The monetarist view of the price level reinforces the case for a
monetary rule by implying that one of the factors which might cause
someone to favor variations in the monetary growth rate, cost-push inflation,
does not occur.
Having seen how the monetary growth rule fits into the rest of
monetarism let us see to what extent it conflicts with Keynesian theory. It
does conflict in one way because Keynesians look upon velocity as being
variable; a belief connected with their view that the private sector is
unstable, and with their emphasis on the interest elasticity of the demand for
money. Hence, to the Keynesians a constant rate of monetary growth would
not result in an acceptable degree of income stability. However, Keynesians
may well accept some of the other arguments mentioned above which cause
monetarists to favor a constant growth rate. Thus, Keynesians need not
consider it desirable to change the money growth rate to accommodate
particular sectors of the economy. And similarly, they need not accept the
likelihood of cost-push inflation, or they may feel that while cost-push
inflation is a serious possibility it should be resisted by not creating the
additional money stock demand at higher prices. Moreover, as pointed out
above, Keynesians may well accept the use of a money stock target.
Despite the fact that the monetary growth rule fits in so well with a large
number of monetarist propositions, it is in a very important way a separate
issue, independent of the validity of all other monetarist propositions. This is
so because the main arguments for a constant monetary rule are essentially
quite different from what has been discussed so far. They are that monetary
policy affects the economy with long and unpredictable lags, or that the
central bank is likely to be inefficient and follow goals other than income
stabilization.70 These hypotheses are not derivable from other monetarist
propositions, nor do they conflict in any important way with Keynesian
200 Thomas Mayer
propositions. Yet while, strictly speaking, these two hypotheses are neither
necessary nor sufficient conditions for the desirability of a monetary rule,
they are close to it.71 Thus, if it were shown conclusively that the lags of
monetary policy are so long and variable that discretionary monetary policy
is likely to be destabilizing, or that the central bank is too inefficient to
operate a successful stabilization policy, then many—probably most–
Keynesians would support a monetary rule. And concern that a discretionary
stabilization policy may be destabilizing is far from being a monetarist
monopoly. In fact, a classic article warning of this danger was written by a
Keynesian, A.W.Phillips.72
Conversely, if it were shown conclusively that discretionary policy can
stabilize the economy, then probably most monetarists would reject the
monetary growth rule. To be sure, a monetarist with beliefs in a stable
demand for money, and in the inherent stability of the private sector, is likely
to expect that even a successful stabilization policy will do relatively little
good, but it could still do some good. Hence, it is not surprising that belief in
a stable monetary growth rule is not a component of Friedman’s definition of
monetarism.73 Thus, the debate about a monetary growth rule transcends the
issue of monetarism versus Keynesianism.74
X ABSENCE OF AN INFLATION-UNEMPLOYMENT TRADE-OFF
Having looked at the basic theory of the monetarists, their choice of
estimation procedures, and their views on monetary policy there remain
three monetarist propositions having to do with economic policy in general.
One of these is the monetarists’ belief that, except in the short run, the
Phillips curve is in real terms, so that, at most, there exists a very limited
trade-off between inflation and unemployment.
The real Phillips curve is related to three of the previously discussed
monetarist propositions, the quantity theory, the stability of the private
sector, and the stable monetary growth rate. If the Phillips curve (over the
time span relevant for analysis) is in real terms, then an increase in the
quantity of money does not affect real income, but affects only prices since it
merely changes the wage unit. Moreover changes in Keynesian variables
such as fiscal policy then have no lasting effect on real income.75
But a Keynesian could accept the real Phillips curve and still claim that
changes in the marginal efficiency of investment are more important than
changes in the monetary growth rate in explaining short run fluctuations in real
income. This is so because, with a belief in the instability of the private sector, a
Keynesian believes that much of the time the economy is in a situation where the
marginal efficiency of investment has changed, and the nominal wage has not
yet adapted to this change, so that real income is affected.
The stable money growth rate rule too has a connection with the real
Phillips curve. One objection to it is that it would not allow the central bank
to intervene when unemployment becomes too high. But if there exists only a
The structure of monetarism 201
very short-run trade-off between unemployment and inflation such
intervention would do little good, and hence a monetary growth rate rule
becomes more acceptable.76
Having seen that the real Phillips curve fits into the monetarist
framework, to what extent is it inconsistent with the Keynesian framework?
One obvious inconsistency arises in an historical context. In the General
Theory Keynes sharply rejected Pigou’s assumption that workers bargain for
a real wage (which is what the real Phillips curve says), and argued instead
that workers bargain for a certain money wage.
A second inconsistency relates to the current Keynesian, or neo-Keynesian,
model. In this model the Phillips curve fixed in nominal terms is used to
determine the price level. If a real Phillips curve is substituted for the
nominal Phillips curve a Keynesian has no way of determining the
equilibrium price level.77 In this way, the acceptance of the real Phillips
curve would weaken Keynesian theory.
But despite this, the debate about the real or nominal nature of the
Phillips curve is to a considerable extent independent of the Keynesian-
monetarist debate. It is essentially an empirical issue which has to be
resolved by detailed studies of the labor market, rather than by settling the
monetarist-Keynesian debate in some other way, and then deducing the
nature of the Phillips curve from the result reached in the monetarist-
Keynesian debate. If empirical studies were to show conclusively that the
Phillips curve is in real terms a Keynesian could surely accept this result
without abandoning Keynesian theory in favor of monetarism. Conversely, if
the empirical evidence were to show that the Phillips curve is fixed in
nominal terms, a monetarist could easily live with this conclusion.
XI CONCERN ABOUT INFLATION
Monetarists appear to be more concerned than are Keynesians about the
disadvantages of unanticipated inflation, and to be relatively less concerned
about the disadvantages of unemployment.78 This choice between these two
evils can be related to several of the foregoing characteristics of monetarists.
One is that the quantity theorist pays much more attention to the likelihood
of price changes than does the Keynesian. Indeed, one of the standard
criticisms which monetarists make of Keynesians is to accuse them of
assuming that the price level is constant.79 And someone who considers price
level changes to be a serious possibility will obviously be concerned much
more about potential inflation than someone who more frequently takes the
price level as constant.
Second, there is the belief in the inherent stability of the private sector at
an acceptable rate of unemployment. While modern Keynesians may readily
concede that underemployment cannot be an equilibrium, they still stress
that serious underemployment may occur frequently, and continue for a very
long time. The monetarist, by contrast, has a stronger belief in the corrective
202 Thomas Mayer
forces that bring the private sector close to full employment if it is left
undisturbed by government policy. Hence, the monetarist worries less about
unemployment than the Keynesian does.
Third, there is the monetary growth rule. A stable monetary growth rule
would limit the potential inflation rate by denying the economy the
additional liquidity needed during an inflation. Hence, someone who is very
concerned about inflation, and the inflationary bias of the political process,
might be led by this to favor the monetary growth rule.80 On the other hand,
if velocity falls or productivity increases to an extent unanticipated when the
monetary rule is instituted, substantial unemployment might result. Hence a
Keynesian who is very concerned about unemployment may, for this reason,
reject a stable money growth rule.
A fourth, rather tenuous, connection is that, by accepting a real Phillips
curve monetarists abandon any hope of being able, except in the short run,
to lower unemployment at the cost of inflation. And while this may not make
monetarists more concerned about inflation, it causes them to oppose as
essentially useless inflationary policies which aim at raising employment.
But again, the issue under discussion is far removed from the main area of
monetarist-Keynesian contention. For example, if we had conclusive
evidence on the validity of the quantity theory and the monetarist
transmission process, it would probably do little to change our relative
degree of concern about inflation and unemployment. This depends much
more on other issues, such as the effects of inflation on income distribution,
and on fundamentally ethical judgments.
XII DISLIKE OF GOVERNMENT INTERVENTION
The final characteristic of monetarists, at least in the United States, is a
dislike of government intervention. This is not limited to macroeconomics; in
general monetarists appear to be much more satisfied with the outcome of
market processes than most Keynesians are. There is, of course, no way of
proving that this attitude should be considered a component of monetarism,
rather than a characteristic which those economists who are monetarists
happen to have for extraneous reasons. However, a dislike of government
regulations fits very well with most of the previously discussed components
of monetarism. Thus, a belief in the quantity theory implies that there should
be no countercyclical fiscal policy. Moreover, a countercyclical fiscal policy
might result in the government sector expanding in a recession more than it
shrinks in the expansion, so that it grows secularly.81 In any case, if the
private sector is inherently stable no countercyclical policy may be needed or
be desirable. Someone who objects to government regulations is less likely to
be interested in allocative detail than someone who has to have information
about various sectors to plan government policy. And conversely, if the
behavior of various sectors does not matter for macroeconomic policy, some
government regulations should be abolished. Furthermore, if the behavior of
The structure of monetarism 203
the price level is essentially independent of the pricing policies and wage
policies followed in ‘strategic industries’ then this is another reason why
some government regulations are unnecessary.
Using the money stock rather than interest rates or bank credit as the
target of monetary policy means that the government can leave the
determination of interest rates and bank credit to free market, and can
confine its attention to the stock of money, something just about always
considered outside the domain of the private market. A monetary growth
rule obviously reduces the need for discretionary policy. And if the Phillips
curve is such that one cannot successfully trade off unemployment and
inflation then here is another task the government should not attempt.
Finally, there are several links between a concern about inflation and concern
about the growth of government. One is that inflation can easily lead to political
pressures for the imposition of wage and price controls. A second is that, given
a progressive tax system, inflation raises the share of the government sector with
the resulting temptation to increase government expenditures. A third link is that
since one way that government expenditures have risen is through inflationary
finance, prevention of inflation may indirectly limit government expenditures.
Fourth, deficit expenditures when financed by newly created money, as is so
often the case, tend to be inflationary.
A critic of monetarism might therefore be tempted to claim that
monetarism is basically an ‘ideological’ doctrine; that it consists of finding
seemingly technical reasons to hide a basic commitment in favor of
unfettered capitalism. But this temptation to play amateur psychoanalyst
should be firmly resisted. A monetarist can reply to it very easily by
reversing the argument, and claiming that the ideological element in the
debate rests with the Keynesians; that it is their ideological commitment to
government regulations and the growth of bureaucracy that makes them
reject the monetarist’s sound arguments on various technical issues of
monetary economics.
On a more worthwhile level than such name-calling it should be noted
that while opposition to government regulations fits in well with
monetarism, it is still a very loose connection in one important sense.82 One
can be a radical and yet accept all the other monetarist propositions
discussed above. Thus, radicals might even accept the constant monetary
growth rule on the basis that this is the best one can do under capitalism.83 In
fact, a planner in an almost totally controlled economy, such as China,
should find the quantity theory more useful than the Keynesian theory.84
Conversely, one can be a right-wing extremist without being a monetarist.
XIII SOME OTHER DIFFERENCES
If one wants to look for a common thread connecting various monetarist
propositions one need not confine oneself to an ideological consideration
since there is a methodological element available.
204 Thomas Mayer
We live in a world too complex for our intellectual apparatus. We must
therefore do either of two things. One is to take account of a great many
factors at the cost of being able to see their interrelations only in a vague,
clouded way. The other is to simplify drastically, and to look at only a few
factors. Along these lines one can classify economists into ‘cloud makers’
and into ‘oversimplifies’, to use two derogatory terms. Using this
dichotomy the Keynesian is a cloud maker while the monetarist is an
oversimplifer.85 Thus the quantity theory is simpler than the Keynesian
theory in the sense of taking account of fewer variables.86 The picture is
less obvious as far as the monetarist transmission process is concerned. The
monetarist view of this process is certainly more cloudy and less clear than
the Keynesian one, since monetarists believe that it works through a large
number of channels, some of which they cannot specify. However, a vague
transmission process, when combined with Friedman’s methodological
views results in a simple, rather than a complex, view of the world.
Friedman finds a close relationship between changes in money and in
nominal income, and presumably does not feel greatly worried by the fact
that it is difficult to specify the transmission process.87 He stresses
predictive power rather than descriptive realism.88
The monetarist’s hypothesis that the private sector is inherently stable also
helps to simplify the analysis, since, if true, this means that we do not have
to concern ourselves in macroeconomics with fluctuations in expenditure
motives. Hence, one can dispense with the detailed Keynesian analyses of
consumption and investment, as well as many complex business cycle
theories. Monetarists’ disinterest in allocative detail obviously also simplifies
macroeconomics. The same is true for their use of small, rather than large,
econometric models, and for their focus on the overall price level rather than
on the prices charged in individual industries.
Using total reserves rather than a combination of short-term interest rates
and money market conditions as an indicator of monetary policy helps to
simplify the analysis of monetary policy. Indeed, monetarists have criticized
the use of money market conditions because of the complexity and vagueness
it introduces.89 The use of a stable money growth rate also obviously
simplifies the conduct of monetary policy. Indeed, one of the leading
monetarist arguments for it is that we do not have the required information,
such as knowledge of lags, to do better with discretionary policy than a
simple growth rate rule does. And a Phillips curve that does not allow for
any unemployment-inflation trade-off simplifies macroeconomics by
removing one very difficult question, selection of the optimal trade-off. Only
two components of monetarism, the use of a money stock target, and the
concern about inflation do not fit the picture of monetarism as simplification.
There exists also another element that links six monetarist propositions.
This is the monetarist’s skepticism about how much we really know about
the short-run workings of the economy. Monetarists generally seem to be less
optimistic about this than are Keynesians. If we really do know little about
The structure of monetarism 205
the short-run behavior of the economy, then the monetarist transmission
process is less subject to the criticism that it does not try to spell out the
channels of monetary influence in any detail. Any attempt to do this could
then be considered presumptuous. Second, if our knowledge of short-run
economic behavior is limited, then we may not have an adequate framework
for using information about allocative detail. Third, we then do not know
enough to build useful large-scale econometric models. Fourth, the less our
knowledge, the weaker is the case for ‘fine tuning’, and the stronger is the
case for a monetary rule,90 and hence the use of a money stock target.
Finally, the less we know about the economy the less likely are government
regulations to improve it.
XIV CONCLUSION
This chapter has dealt with various propositions that make up monetarism,
broadly defined, and showed that they form a coherent whole. With one
exception (the use of a total reserve measure as the indicator of monetary
policy) they fit together in the sense that definitive proof of the validity of
one of the more basic propositions would increase the plausibility of some of
the other propositions. Figure 8.1 shows relations which have been traced
here between the various propositions.
But this does not mean that monetarism is a paradigm which must be accepted
or rejected as a whole. As pointed out above, with the exception of the quantity
theory itself, and perhaps its transmission process, every single proposition of
monetarism is one which Keynesians could accept while rejecting others, and still
maintain their adherence to basic Keynesian theory. In particular, the policy
propositions are readily detachable from the theoretical propositions of
monetarism, and can be accepted without qualms by a Keynesian. Conversely,
someone who accepts some of the monetarist propositions, including the two most
basic ones (the quantity theory and the monetarist version of the transmission
process) need not therefore accept all the others.
Hence, a good case can be made for abolishing the term ‘monetarism’
altogether, and for treating each proposition independently. This would reduce
the unfortunate polarization of economists into monetarists and anti-monetarists,
with the accompanying tendency to accept or reject various propositions on a
basis other than the empirical evidence bearing on them.91 Admittedly, this may
well be the counsel of perfection since the term ‘monetarism’ is now so well
established and convenient. But eclecticism is fully justified.92
ACKNOWLEDGEMENTS
I am indebted for helpful comments to Karl Brunner, Thomas Cargill, H.
Cheng, Benjamin Friedman, Milton Friedman, Michael Hamburger, Michael
Keran, Allan Meltzer, Franco Modigliani, Manfred J.M.Neumann, Roger
Spencer, Edward Shaw, Daniel Vencill, and to members of workshops and
F
ig
u
re
8
.1
I
n
te
rr
el
at
io
n
o
f
m
o
n
et
ar
is
t
p
ro
p
o
si
ti
o
n
s
N
o
te
s:
a
m
ea
su
re
m
en
t
p
ro
b
le
m
s
b
ra
n
ge
o
f
as
se
ts
c
o
n
si
d
er
ed
c
st
ab
il
it
y
o
f
d
em
an
d
f
o
r
m
o
n
ey
d
re
la
ti
ve
p
ri
ce
a
n
d
s
to
ck
e
ff
ec
ts
e
d
is
in
te
re
st
i
n
e
x
p
en
d
it
u
re
m
o
ti
ve
s
f
u
n
im
p
o
rt
an
ce
o
f
ex
p
en
d
it
u
re
m
o
ti
ve
s
an
d
o
f
p
ec
u
lia
ri
ti
es
o
f
se
ct
o
rs
g
li
tt
le
c
o
n
ce
rn
a
b
o
u
t
in
st
ab
il
it
y
h
ag
gr
eg
at
e
d
em
an
d
d
et
er
m
in
ed
b
y
q
u
an
ti
ty
o
f
m
o
n
ey
i
im
m
u
n
it
y
o
f
co
st
-p
u
sh
i
n
fl
at
io
n
j
ir
re
le
va
n
ce
o
f
p
ri
ce
b
eh
av
io
r
in
i
n
d
iv
id
u
al
i
n
d
u
st
ri
es
k
d
is
in
te
re
st
i
n
p
ar
ti
cu
la
r
se
ct
o
rs
l
re
su
lt
r
ea
ch
ed
b
y
m
o
d
el
s
m
gr
o
w
th
r
u
le
a
s
sp
ec
ia
l
ty
p
e
o
f
m
o
n
et
ar
y
ta
rg
et
n
n
o
n
ee
d
t
o
h
el
p
p
ar
ti
cu
la
r
se
ct
o
rs
o
li
tt
le
n
ee
d
o
f
o
ff
se
t
fl
u
ct
u
at
io
n
s
p
p
re
ve
n
t
u
n
an
ti
ci
p
at
ed
i
n
fl
at
io
n
q
fo
cu
s
o
n
p
ri
ce
f
le
x
ib
il
it
y
r
n
at
u
ra
l
ra
te
o
f
u
n
em
p
lo
ym
en
t
s
ab
se
n
ce
o
f
u
n
em
p
lo
ym
en
t-
in
fl
at
io
n
t
ra
d
e-
o
ff
t
a
d
a
p
ti
v
e
ex
p
ec
ta
ti
o
n
s
a
n
d
em
p
h
a
si
s
o
n
d
is
ti
n
ct
io
n
b
et
w
ee
n
re
a
l
a
n
d
n
o
m
in
a
l
m
a
g
n
it
u
d
es
u
re
al
v
ar
ia
b
le
s
u
n
af
fe
ct
ed
b
y
in
fl
at
io
n
v
ab
se
n
ce
o
f
tr
ad
e-
o
ff
a
m
el
io
ra
te
s
p
o
te
n
ti
al
l
o
ss
f
ro
m
m
o
n
et
ar
y
ru
le
w
in
ef
fe
ct
iv
n
es
s
o
f
fi
sc
al
p
o
li
cy
x
in
fo
rm
at
io
n
o
n
s
ec
to
rs
n
o
t
n
ee
d
ed
y
in
te
rv
en
ti
o
n
i
n
s
ec
to
rs
n
o
t
n
ee
d
ed
f
o
r
m
ac
ro
ec
o
n
o
m
ic
p
o
li
cy
z
n
o
i
n
te
rf
er
en
ce
w
it
h
i
n
te
re
st
r
at
es
o
r
cr
ed
it
v
o
lu
m
e
A
n
o
d
is
cr
et
io
n
ar
y
m
o
n
et
ar
y
p
o
li
cy
B
le
ss
d
an
ge
r
o
f
w
ag
e
an
d
p
ri
ce
c
o
n
tr
o
ls
C
in
fl
at
io
n
ar
y
im
p
ac
t
o
f
go
ve
rn
m
en
t
ex
p
en
d
it
u
re
f
in
an
ce
d
b
y
n
ew
m
o
n
ey
c
re
at
io
n
D
in
fl
at
io
n
r
ai
se
s
go
ve
rn
m
en
t
re
ce
ip
ts
E
in
fl
at
io
n
ar
y
fi
n
an
ce
f
ac
il
it
at
es
a
d
d
it
io
n
al
e
x
p
en
d
it
u
re
s
F
fo
cu
s
o
n
t
h
e
m
o
n
ey
s
to
ck
r
a
th
er
t
h
a
n
o
n
i
n
te
re
st
r
a
te
a
n
d
d
is
re
g
a
rd
o
f
th
e
fi
n
a
n
ci
a
l
in
te
rm
ed
ia
ti
o
n
-r
es
id
en
ti
al
c
o
n
st
ru
ct
io
n
c
h
an
n
el
The structure of monetarism 207
seminars at the Board of Governors, Federal Reserve System, the San
Francisco Federal Reserve Bank, and MIT, none of whom are responsible for
any remaining errors.
NOTES
1 The term ‘monetarism’ was originated by Karl Brunner (The Role of Money and
Monetary Policy’, Federal Reserve Bank of St Louis Review 50, July 1968, pp. 8–
24), and was popularized by David Fand; see for instance his ‘Monetarism and
Fiscalism’, Banca Nazionale del Lavoro Quarterly Review 94, September 1970,
pp. 3–34 and ‘Ein monetaristisches Modell des Geldwirkungsprozesses’, Kredit
und Kapital 3 (1970), pp. 361–85.
2 Similarly, there is no authoritative listing of Keynesian propositions. I have
interpreted Keynesian theory as a theory represented by the views of such economists
as James Duesenberry, Franco Modigliani, Paul Samuelson and James Tobin rather
than by the more extreme views which can be found in the writings of economists
such as Alvin Hansen. Hence, what I am calling ‘Keynesian theory’ is to some
extent a synthetic theory which has probably been influenced by monetarism.
3 Throughout my discussion deals only with monetarist school as it exists in the
United States.
4 Brunner, The “Monetarist Revolution” in Monetary Theory’, Weltwirtschaftliches
Archiv 105, no. 1, 1970, pp. 1–30.
5 Fand also mentions another item, the monetarist’s belief in long and variable lags.
But at present many Keynesians also believe that monetary policy has long lags.
David Fand, ‘Monetarism and Fiscalism’, loc. cit.
6 James Tobin, The New Economics One Decade Older (Princeton, NJ, pp. 58–9).
Actually, as far as item 7 is concerned Tobin refers to the money stock rather than
total reserves, but this is a minor difference. Also, item 8 is implicit, rather than
explicit in Tobins’s list. Paul Samuelson (‘Reflections on the Merits and Demerits of
Monetarism’ in James Diamond (ed.) Issues in Fiscal and Monetary Policy, Chicago,
IL, 1971, pp. 7–21) lists the quantity theory and the monetary growth rate rule as
the two basic propositions of monetarism. To these he adds the belief in wage and
price flexibility, and in the response of the interest rate to inflation (two propositions
which can be treated as part of the quantity theory) and a belief in the real nature of
the Phillips curve with the associated belief in a natural rate of unemployment. He
then stated (p. 20) that ‘there is no reason why monetarists should believe this except
that all of these notions happen to be believed by one man, Professor Friedman’.
(For a similar statement see James Tobin, op. cit., p. 62.) This not only ignores the
work of Brunner and Meltzer, but also ignores the various linkages discussed below.
7 ‘The State of the Monetarist Debate’, Federal Reserve Bank of St Louis Review 56,
September 1973, pp. 5–6. Although Andersen states that it is also accepted by
‘many other economists’ it is frequently rejected by Keynesians.
8 For a typical example see James Tobin’s criticisms of the policy recommendations
made by the, mainly monetarist, ‘Shadow Open Market Committee’ (James Tobin,
‘Monetary Policy in 1974 and Beyond’, Brookings Papers on Economic Activity
1974:1, pp. 219–32.
9 Thus Allan Meltzer wrote (private communication) ‘I do not accept any but points
1 to 4 as part of monetarism. The other points are, for me, propositions that I
208 Thomas Mayer
accept to varying degrees. Many are unrelated to monetarism. For example, your
point 5 is a Hicksian proposition about composite goods. It should be accepted by
all economists.’ It is certainly true that item 5 can be considered as a theorem about
composite goods, but there is still a decision to be made as a matter of research
strategy, rather than as a matter of formal theory, whether one analyzes the general
price level as a single unit or by looking at individual prices.
10 However, I have omitted the international aspect of monetarism, the proposition
that with fixed exchange rates a country’s money stock and price level depend not on
its own monetary policy, but on the whole world’s monetary policy. At least in the
United States, this proposition has not played much of a role in monetarist discussions.
But adding it to my list would not change my conclusions because, as Harry Johnson
has pointed out, ‘a properly understood Keynesian approach to the system as a
whole would produce the same conclusion’ (H.G.Johnson and A.R.Nobay (eds)
Issues in Monetary Economics, London, Oxford University Press, 1974, p. 50.)
Besides, it is essentially part of the first proposition, the quantity theory. I have also
omitted an item mentioned by Brunner (‘The Role of Money and Monetary Policy’,
op. cit., p. 9), the belief that the monetary authorities can control the stock of money.
This is now accepted by many Keynesians as well, though admittedly, Keynesians
tend to qualify it more than monetarists do.
The various hypotheses I describe as monetarist do, of course, predate the
development of the term ‘monetarism’. The quantity theory, together with its
transmission process, has an ancient history, as do, though perhaps to a lesser
extent, the next three items. Items 6, 7 and 10 are newer because the problems they
present are newer. A hundred years ago nobody was worried about the proper size
of an econometric model, or about the correct monetary indicator. Debate about
items 8 and 9 can, to some extent, be traced back to the banking school-currency
school debate. The final two items again, do have a long history. What is new about
monetarism is therefore primarily its combination of hypotheses into a single doctrine.
11 The modern Keynesian theory differs from the quantity theory in denying that
changes in the money stock dominate changes in income, but it does not claim that
changes in the money stock are unimportant. According to J.R.Hicks (The Crisis
in Keynesian Economics, Oxford, 1974, pp. 31–2) Keynes himself ‘must surely in
some sense, perhaps a very weak sense, have been a monetarist. He has nevertheless
been read to imply that there is nothing to be done with money.’
12 Thus the growing literature on search costs is relevant to the monetarist-Keynesian
debate and monetarists attach more importance to search costs than do Keynesians.
13 To be sure, if there are long lags in the effects of money on income then one might
predict next years’s income by changes in the money stock in previous years, but if
the lags are highly variable, even this would not work.
14 Patinkin’s model uses the quantity theory’s analytic procedures insofar as it focuses
on the gap between desired and actual real balances. However, it is Keynesian in its
use of capital theory since, as Patinkin has argued, the Cambridge school did not
use capital theory in its monetary analysis to any significant extent. (See Don
Patinkin, ‘Keynesian Monetary Theory and the Cambridge School’, in H.G.Johnson
and A.R.Nobay, op. cit., pp. 3–30.)
15 This difference between looking at the quantity theory as an engine of analysis and
looking at it as the conclusion that money matters a great deal is at the heart of a
dispute between Friedman and Patinkin. Patinkin, focusing on the fact that
Friedman—like Keynes, but unlike pre-Keynesian quantity theorists—uses capital
The structure of monetarism 209
theory in his monetary analysis, has argued that Friedman’s theory is more a
Keynesian than a quantity theory. (Don Patinkin, ‘The Chicago Tradition, the
Quantity Theory and Friedman’, Journal of Money, Credit and Banking I, Feb.
1969, pp. 46–70, and ‘Friedman on the Quantity Theory and Keynesian
Economies’, Journal of Political Economy 80, September/October 1972, pp. 883–
905.) Friedman’s reply was to object to Patinkin’s ‘propensity to take the “quantity
theory” to mean one thing, and one thing only, namely the long-run proposition
that money is neutral, even though he fully recognizes, indeed insists, that quantity
theorists (myself included) were concerned mostly with short-run fluctuations’.
(‘Comments on the Critics’, Journal of Political Economy 80, September/October
1972, p. 932.) Perhaps the point should be stated differently by saying that
Friedman classifies theories on the basis of the conclusions they reach, while Patinkin
classifies them on the basis of the analytic method used.
16 I will not describe the monetarist transmission processes here in any detail.
Friedman’s variant stresses substitution effects, and the influence of changes in the
money stock on the nominal interest rate while the Brunner-Meltzer variant stresses
relative price and stock effects. Both variants attach much importance to the
distinction between nominal and real rates of interest, and more generally, pay
greater attention to price changes than Keynesians typically do.
I am discussing only the transmission process for changes in the quantity of
money, and not for fiscal policy, etc. The monetarist argument that fiscal policy
changes result in counteracting changes, such as ‘crowding out’, which offset them
after some time, is really part of the previously discussed monetarist proposition,
that changes in money income are explained largely by changes in the money stock.
17 Thus in commenting on a draft of this chapter Milton Friedman wrote (private
communication): ‘I believe an important distinction between Keynesian and
monetarist views is one that I have not myself stressed sufficiently but that comes out
in the course of some of your comments. This is the distinction between money and
credit and most particularly in what one regards as the price of money. The Keynesian
approach invariably regards the interest rate as the price of money whereas the
quantity theory approach regards the interest rate as the price of credit and the
inverse of the price level as the price of money. This is exremely important in connection
with the way in which the demand curve for money is used.’
18 Cf. Milton Friedman, ‘A Theoretical Framework for Monetary Analysis’, National
Bureau of Economic Research, Occasional Paper 112 (New York, 1971), p. 28.
19 Y.C.Park, ‘Some Current Issues on the Transmission Process of Monetary Policy’,
International Monetary Fund, Staff Papers, March 1972, p. 38.
20 Many monetarists believe that if the quantity of money is increased the nominal
interest rate declines only very temporarily. It soon rises back to its previous level,
and, due to the Fisher effect, even exceeds it. The monetarist therefore looks upon
the expected real interest rate as fairly stable. Hence, one of the factors which can
cause fluctuations in the quantity of money demanded, changes in the expected
real interest rate, seems much less important to the monetarist than to the Keynesian.
Another important reason why monetarists take the demand for money as stable
is that, as discussed below, the monetarist treats expenditure incentives as much
more stable than the Keynesian does, and hence considers the expected real rate of
interest, and therefore the demand for money, to be stabler than Keynesians do.
21 ‘The Quantity Theory of Money: A Restatement’, reprinted in Milton Friedman,
The Optimum Quantity of Money (Chicago, 1969) ch. 2.
210 Thomas Mayer
22 In the United States the growth rates of M1 and M2 have diverged widely, presumably in
large part due to restrictions on interest payments on deposits. For example, between
December 1972 and December 1973 M1 grew at a 6.1 percent rate while M2 (excluding
large certificates of deposit) grew at an 8.9 percent rate, that is at a 69 percent greater rate.
23 Some attempts have been made to settle this issue by seeing whether M1 or M2 have
a closer correlation with income. But these attempts founder on the fact that the
‘reverse causation’ bias may be greater for one measure than for the other.
24 To be sure, in the mainly Keynesian Federal Reserve–MIT–Penn model the interest
rate has a strong effect on consumption. But this is not true for the more typically
Keynesian models.
25 According to Karl Brunner (‘The Monetarist Revolution in Monetary Theory’,
op.cit., p. 3) the borrowing-cost interpretation is post-Keynesian rather than part
of Keynes’ own thought.
26 Keynes’ evidence for the interest inelasticity of consumption is extremely casual
(The General Theory, London, 1936, pp. 93–4), but this rather arbitrary judgment
allowed him to make a great simplification. This is to dichotomize his model into
decisions made about the disposition of income (to save it or consume it) and
decisions made about asset composition (to hold money or bonds). He did not
have to consider the feedback effect of asset decisions on consumption through
changes in the propensity to consume as the interest rate changes.
27 Although this is no more than a surmise I suspect that the debate about the
channels of monetary influence received some of its impetus from the fact that at
one time empirical studies of business investment behavior showed the interest
rate as playing, at best, a very small role. Hence, monetarists had a strong reason
to argue that changes in the money stock do not operate just through business
investment, while Keynesians had an incentive to treat business investment as the
only link between changes in the money stock and income.
28 Karl Brunner and Allan Meltzer, ‘Money, Debt and Economic Activity’, Journal of
Political Economy 80, September/October 1972, pp. 951–77; Karl Brunner, ‘A
Survey of Selected Issues in Monetary Theory’, Schweizerische Zeitschrift für
Volkswirtschaft und Statistik 107, 1971, pp. 1–146.
29 Y.C.Park (op. cit., p. 31) has argued that ‘Brunner and Meltzer—contrary to their
claim—accept the Keynesian view of the nature of the transmission process; what they
seem to reject is the heuristic simplification of reality with regard to the range of assets
considered in the Keynesian income/expenditure theory.’ This statement is very much
open to question if one treats as ‘Keynesian’ not every single factor mentioned in the
General Theory and post-Keynesian writings, but only those which are stressed.
30 ‘The State of the Monetarist Debate: Comment’, Federal Reserve Bank of St Louis
Monthly Review 55, September 1973, p. 11.
31 An approach which looks at expenditure incentives is likely to come up with
different results than one which focuses on the demand for money. The latter–on
an intuitive level at least—seems stable, while—again on an intuitive level–
expenditure incentives seem highly variable. Obviously, these two intuitions are in
conflict due to Walras’ Law. Perhaps the resolution of this conflict is that while the
incentives for particular expenditures looked at one at a time seem unstable, much
of this instability averages out in the sense that one sector may be depressed while
another is in a boom.
32 Leonall Andersen (‘The State of the Monetarist Debate’, Federal Reserve Bank of
St Louis Monthly Review 55, September 1973, pp. 2–8) has pointed to another
The structure of monetarism 211
factor as the difference between Keynesian and monetarist views on the stability of
the private sector, the length of time it takes to return to the neighborhood of
equilibrium when the economy is subjected to a shock.
33 This does not mean that monetarists can ignore all institutional detail. They have to
consider numerous institutional factors (which differ among countries) in their analysis
of the money supply process. But this is different from concern with allocative detail.
34 Thus in the, mainly Keynesian, Federal Reserve-MIT-Penn model, one of the major
channels by which monetary changes affect income is credit rationing.
35 Obviously, a large sophisticated model, like the above mentioned one, can have
several channels, and is not confined to borrowing costs. But for most Keynesian
expositions borrowing costs are the channel.
36 This does not mean that monetarists are uninterested in allocative detail per se.
They are often strongly interested in it because they look upon government
interference with financial markets as creating very serious problems. Thus they
oppose the suppression of financial deepening. In the United States monetarists
are much more critical of Regulation Q (the limitation of interest payments on
bank deposits) than are Keynesians. It is only with respect to the use of allocative
detail as a predictor of short-run changes in income that monetarists have shown
less interest in it than Keynesians.
37 But to the extent that the velocity of money differs in various sectors the monetarist
has an incentive to analyze the distribution of money between various sectors. For
a notable example see Richard Selden, The Postwar Rise in the Velocity of Money’,
Journal of Finance 16, December 1961, pp. 483–545.
38 This statement is subject to one qualification. The monetarist is likely to pay a great
deal of attention to the efficiency of one sector, the financial sector, and to point out
the distortions created in this sector by government regulations.
39 Admittedly, capital rationing tends to make the demand for money less stable.
40 One way of determining whether someone is a Keynesian or a monetarist is to ask
for a quick and intuitive answer to the following question: ‘Suppose the price of
petroleum rises. What will this do to the average of other prices?’
41 See, for example, Keith Carson, ‘A Monetarist Model for Economic Stabilization’,
Federal Reserve Bank of St Louis Review 52, April 1970, pp. 7–25.
42 This is subject to the caveat that the central bank might raise the money stock to
maintain output when some prominent industries raise their wages and prices, or
when unemployment develops.
43 The private sector may be stable even in the sense of being immune to cost-push
inflation even if individual price increases do not result in corresponding price
decreases in other sectors. This is so if, and only if, the forces making for cost-push
are weak. Similarly, erratic shifts in expenditure motives could destabilize the private
sector even if the monetarist’s approach to the price level is correct. And allocative
detail would then be important.
44 Cf. Sidney Weintraub, Keynes and the Monetarists (New Brunswick, NJ, 1973) ch. 7.
45 Admittedly, this reasoning is only a first approximation, for it ignores the fact that
an increase in the price level, by raising the interest rate, raises velocity. However,
monetarists may feel justified in ignoring this effect as minor because they may
believe that the interest elasticity of the demand for money is low.
46 Otto Eckstein and Gary Fromm, ‘Steel and the Postwar Inflation’, Study Paper 2,
US Congress, Joint Economic Committee, 86th Congress, 1st Session, Washington,
DC, 1959, p. 34.
212 Thomas Mayer
47 See Denis Karnosky, ‘A Primer on the Consumer Price Index’, Federal Reserve
Bank of St Louis Review 56, July 1974, p. 7.
48 General Theory, op. cit., p. 293.
49 See William Nordhaus, ‘Recent Developments in Price Dynamics’, in Board of
Governors, Federal Reserve System, The Econometrics of Price Determination,
Conference (Washington, DC 1972). See also W.Godley and W.Nordhaus, ‘Pricing
in the Trade Cycle’, Economic Journal 82, September 1972, pp. 853–82. Perhaps
this tendency of Keynesians to treat prices as cost determined represents a partial
fusion of the Keynesian and institutionalist schools.
50 A leading monetarist, Phillip Cagan, has suggested that the dependence of price
changes on changes in costs can be explained as a short run phenomenon resulting
from the difficulties which firms have in coordinating their price changes (Phillip
Cagan, Inflation: The Hydra-Headed Monster, Washington, DC, 1974, pp. 21–4.)
51 The only serious Keynesian discussion of this issue I know of is Abraham Bergson’s
‘Price Flexibility and the Level of Income’, Review of Economics and Statistics
XXV, February 1943, pp. 2–5.
52 It is not clear whether a Keynesian is more likely than a monetarist to believe that
the change will be in output rather than in prices. On the one hand, a Keynesian is
more likely to stress price inflexibility and situations of underemployment. On the
other hand, many monetarists stress expectational effects, and anticipatory pricing
in inflation. Insofar as prices are set in anticipation of inflation, a decline in demands
is likely to affect output rather than prices even during an inflation when downward
price flexibility is not a problem.
53 And while monetarists frequently consider prices to be fairly flexible, one can be a
monetarist without this belief.
54 However, a number of fairly small Keynesian models do exist. It may be worth
noting that if one is trying to evaluate the Keynesian-monetarist debate by comparing
the predictive powers of monetarist and Keynesian models one should compare
the monetarist model (i.e. the Andersen-Jordan model), not with large Keynesian
models such as the Wharton model, as is sometimes done, but with small Keynesian
models. Thus, the finding that the Andersen-Jordan model does well compared to
the Wharton and OBE models (cf. Yoel Haitovsky and George Treyz, ‘Forecasts
with Quarterly Macroeconomic Models, Equation Adjustment and Benchmark
Predictions: The U.S. Experience’, Review of Economics and Statistics LIV, August
1972, pp. 317–25) is not as important for the Keynesian-monetarist dispute as is
the finding that the Andersen-Jordan model’s performance is not outstanding
when compared to that of small Keynesian models. (See S.K.McKnees, ‘A
Comparison of the GNP Forecasting Accuracy of the Fair and St. Louis Econometric
Models’, in Federal Reserve Bank of Boston, New England Economic Review,
September/October 1973, pp. 29–34, and J.W.Elliot, ‘A Direct Comparison of
Short-Run GNP Forecasting Models’, Journal of Business 46, January 1973, pp.
33–60). The trouble with the Wharton or OBE model may be its structural, rather
than its Keynesian, characteristics.
55 See Milton Friedman, ‘Comment’, in Universities-National Bureau Committee for
Economic Research, Conference on Business Cycles (New York, 1951), pp. 112–14.
56 Private communication.
57 ‘The Usefulness of Monetary and Fiscal Policy as Discretionary Stabilization Tools’,
Journal of Money, Credit and Banking III, May 1971, Part 2, pp. 506–32.
58 The indicators-targets dichotomy has been challenged by Benjamin Friedman
The structure of monetarism 213
(‘Targets, Instruments and Indicators of Monetary Policy’, Journal of Monetary
Economics 1, October 1975). However, since I am dealing here with the dispute
between monetarists and Keynesians both of whom generally use this dichotomy,
I am accepting it without questioning its validity.
59 I have phrased this statement in such a tentative way because I am far from certain
that most Keynesians really prefer the short-term rate as their indicator. Unlike the
monetarists, Keynesians have not written much on this topic.
60 See Richard Davis and Frederick Schadrack, ‘Forecasting the Monetary Aggregates
with Reduced Form Equations’, in Federal Reserve Bank of New York, Monetary
Aggregates and Monetary Policy (New York, 1974), pp. 60–71. See also Fred
J.Levine, ‘Examination of the Money-Stock Control Approach of Burger, Kalish,
and Babb’, Journal of Money, Credit and Banking V, November 1973, pp. 924–
38; and James Pierce, and Thomas Thomson, ‘Some Issues in Controlling the
Stock of Money’, in Federal Reserve Bank of Boston, Controlling Monetary
Aggregates II: The Implementation (Boston n.d.), pp. 115–36.
61 David Fand, ‘A Time Series Analysis of the “Bills-Only” Theory of Interest Rates’,
Review of Economics and Statistics XLVIII, November 1966, p. 369.
62 For a detailed exposition of this argument see William Poole, ‘Optimal Choice of
Monetary Policy Instruments in a Simple Stochastic Macro Model’, Quarterly Journal
of Economics 84, May 1970, pp. 197–216; and ‘Rules of Thumb for Guiding
Monetary Policy’ in Board of Governors, Federal Reserve System, Open Market
Policies and Operating Procedures, Staff Studies, Washington, DC, pp. 135–89.
63 The assumption that the money growth rate and the interest rate are negatively
correlated is justified by the analysis being only very short run.
64 It is worth noting that what is relevant is not the stability of either the IS or LM
curve, but its predictability since the central bank can readily offset predictable
fluctuations.
65 A third aspect of the choice between a money stock target and an interest rate
target relates to the problem of lags in the effects of monetary policy. Since many
types of expenditures respond only slowly to a change in the interest rate the
effects of monetary policy tend to be delayed. But this delay can be offset if interest
rates initially overshoot their new level. (See Donald Tucker, ‘Dynamic Income
Adjustments to Money Supply Changes’, American Economic Review LVI, June
1966, pp. 433–49.) Insofar as the central bank follows a money stock target such
an overshoot occurs automatically. But with an interest rate target, the central
bank may fail to allow for the required overshoot. And even if it aims for an
overshoot, it does not know how large it should be.
66 The monetarists look upon expenditure motives as stable too, unless disturbed by
variations in the money growth rate, since they treat the private sector as stable,
but even so, they take the demand for money as the stabler one.
67 A fourth potential target, and money market conditions, is hardly taken seriously
anymore, at least in the United States.
68 A monetary growth rule is supposed to provide a growth rate of money income
which is stable, though this may be a stable rate of inflation or deflation.
69 This conclusion is subject to the caveat that in their formal theory monetarists
consider the demand for money to be stable only in a functional sense. Hence, if
many of the variables in the money demand functions fluctuate, the demand for
money, and therefore income, would also fluctuate under a constant money supply
rule. But according to Friedman, and perhaps to most monetarists, this distinction
214 Thomas Mayer
between the functional stability and the constancy of the demand for money does
not create a serious problem. Insofar as the demand for money is a function of
permanent income or wealth it is likely to grow at a steady rate. To be sure, it is also
a function of the nominal rate of interest. But fluctuations in the nominal interest
rate are largely the result of previous fluctuations in the money growth rate and
prices. Hence, given a constant money growth rule, velocity would tend to be fairly
stable in a numerical, as well as a functional, sense.
70 Another reason sometimes given for a monetary growth rule is that it reduces arbitrary
government interference, substituting as it does the rule of law for the rule of humans.
71 They are not really necessary conditions, because someone might advocate the
monetary growth rule solely on the basis that it curbs arbitrary government power.
They are also not really sufficient conditions because some might reject the rule,
even though it would stabilize income, because they believe that monetary policy
should be used to stabilize particular sectors of the economy, to help government
finance, or to obtain balance of payments equilibrium, etc.
The belief that stabilization policies are actually destabilizing may appear to
conflict with one Keynesian proposition, the instability of the private sector. If the
government sector has been a net contributor to instability it would seem that the
private sector must be relatively stable. But this reasoning is questionable. At least
in the United States, discretionary fiscal policy has frequently not behaved
countercyclical; government expenditures have frequently risen at times of high
activity. Similarly, if one accepts a money stock measure of monetary policy it also
has usually not been countercyclical in the post-war period.
72 ‘Some Notes on the Estimation of Time-Forms of Reactions in Interdependent
Dynamic Systems’, Economica 23, May 1956, pp. 99–113.
73 The Counter-Revolution in Monetary Theory (London 1970), p. 26.
74 In any case, the debate about stable money growth versus discretionary policy is in
the process of becoming technologically obsolete. Recent work suggests that an
intermediate position, a stable central bank reaction function to changes in income,
may well be superior to both a fixed money growth rule and to ad hoc discretionary
policy. (See J.Phillip Cooper, Development of the Monetary Sector, Prediction and
Policy Analysis in the FRB-MIT-Penn Model, Lexington, MA, 1974.)
75 Cf. Jerome Stein, ‘Unemployment, Inflation and Monetarism’, American Economic
Review LXIV, December 1974, pp. 867–87. Two other ways in which the real
Phillips curve fits in well with the quantity theory are the quantity theory’s emphasis
on the distinction between real and nominal magnitudes, and the use of adaptive
expectations in both the modern quantity theory and the real Phillips curve analysis.
76 The direction of the connection between the real Phillips curve and the monetary
growth rule is from the real Phillips curve to the growth rate rule rather than vice versa.
77 Insofar as prices are changing, Keynesians could use an expectational adjustment
model to derive a modified Phillips curve which would then allow them to determine
the price level. But if the inflation rate stays constant long enough for expectations
to have fully adapted, Keynesians could predict neither the price level nor the
unemployment rate unless they have independent information on what the natural
rate of unemployment is. However, the same is true for monetarists. They also
need a specialist in labor markets to tell them the natural rate of unemployment.
78 And there are monetarist objections even to fully anticipated inflation. As Friedman
has pointed out (The Optimum Quantity of Money, op. cit., ch. 1) the price level
should be falling to induce the public to hold the optimum quantity of money.
The structure of monetarism 215
79 See, for example, Milton Friedman, ‘Comments on the Critics’, op. cit., pp. 917–18.
80 Admittedly, a constant monetary growth rate, if set at too high a level, might result
in inflation. But this would be a fully anticipated inflation.
81 See James Tobin, op. cit., p. 63. However, Tobin also points out a negative
relationship; insofar as fiscal policy has little, or no, effect on income, inflation
cannot be used as an excuse for cutting the budget.
82 See ibid., p. 63.
83 Radicals, unless they are Marxist, need not reject the monetarist’s belief in the
inherent stability of the private sector since their objection to capitalism could be
founded on grounds other than instability.
84 The Keynesian’s marginal efficiency of investment and the multiplier play little, or
no, role in determining income in a controlled economy. On the other hand, since
the public has freedom to adjust its money holdings the quantity theory is relevant.
85 This does not imply that the quantity theorist thinks we live in a simple world. One
may want to use simple models precisely because the world is so complex that no
complex, but still manageable, model can do it justice. This can be seen readily on
an empirical level. If we try to forecast a variable which has determinants of only
moderate complexity we tend to use a standard ‘explanatory’ regression. But if we
try to forecast a variable with extremely complex determinants we are more likely
to use a naive model or some other autoregressive scheme.
86 The Brunner-Meltzer version of the quantity theory gives the impression of being
more complex than the Keynesian theory since it criticizes Keynesian theory for
ignoring some important effects. But this appearance is due, in part, to the fact that
when Brunner and Meltzer criticize the Keynesian model they focus on the greatly
oversimplified IS-LM diagram which does not give the full Keynesian story. Although
they introduce some additional variables, they omit some of the Keynesian variables.
87 To be sure, Friedman believes that the mere correlation of money and income is not
enough to establish the quantity theory, that a plausible transmission process is needed.
(See Milton Friedman and Anna Schwartz, ‘Money and Business Cycles’, Review of
Economics and Statistics XLV, February 1963, Supplement, p. 59.) However, a vague,
generalized sketch of the transmission process may suffice for this.
88 Karl Brunner, too, has rejected the type of descriptive realism that tests theories by
evaluating the validity of their assumptions. (See his ‘Assumptions and the Cognitive
Quality of Theories’, Synthese 20, 1969, pp. 501–25.
89 See Karl Brunner and Allan Meltzer, Some General Features of the Federal Reserve’s
Approach to Policy, US Congress, House, Committee on Banking and Currency,
Subcommittee on Domestic Finance, 88th Congress, 2nd Session (Washington, DC 1964).
90 This statement is subject to the objection that a great deal of knowledge is required
to decide on the correct long run growth rate rule. But monetarists believe that the
economy can adapt itself to any monetary growth rate as long as this rate is stable.
91 As Cyrus Gordon (Riddles in History, New York, 1974, p. 156) has put it, ‘all
schools of thought are in reality “schools of un-thought” to the extent that they
prevent us from going to where the facts should lead us’.
92 Thus Karl Brunner has argued that: ‘the four major issues [in the monetarist
debate] allow a variety of combinations…. The evolution of such a spectrum with
a “middle ground” should enrich our future research activities. Such activities
should yield substantive results over the years to the extent that economists
successfully avoid the “media propensity” of equating all issues with ideological
positions’, “Commentary on “The State of the Monetarist Debate”’, Federal Reserve
Bank of St Louis Review 55, September 1973, p. 14.
9 Monetarism
An interpretation and an assessment
David Laidler
Economic Journal (1981) 91, March, pp. 1–28
Like beauty, ‘monetarism’ tends to lie in the eye of the beholder, and before
it can be assessed it must be defined. Though there have been several
valuable attempts over the years to specify monetarism’s key characteristics,1
I shall not rely upon them in this chapter. Each of them has been heavily
conditioned by its time and place of writing, and monetarism has evolved
over the years in response to changing circumstances, and in different ways
in different places, as new hypotheses have either been developed or
absorbed. Thus, I will begin this chapter with my own characterisation of
monetarism. In my view, the key characteristics of monetarism are as
follows:
1 A ‘quantity theory’ approach to macroeconomic analysis in two distinct
senses: (a) that used by Milton Friedman (1956) to describe a theory of
the demand for money, and (b) the more traditional sense of a view that
fluctuations in the quantity of money are the dominant cause of
fluctuations in money income.
2 The analysis of the division of money income fluctuations between the
price level and real income in terms of an expectations-augmented
Phillips curve whose structure rules out an economically significant long-
run inverse trade-off between the variables.
3 A monetary approach to balance-of-payments and exchange-rate theory.
4 (a) Antipathy to activist stabilisation policy, either monetary or fiscal,
and to wage and price controls, and (b) support for long-run monetary
policy ‘rules’ or at least prestated ‘targets’, cast in terms of the
behaviour of some monetary aggregate rather than of the level of
interest rates.
The first characteristic categorises the theoretical core of monetarism as it
developed in the 1950s and 1960s, the second and third represent theory
developed or absorbed by monetarists since the mid-1960s, while the fourth
summarises a view of macroeconomic policy issues which, even though it is
neither logically implicit in their positive analysis, nor their exclusive
property, has remained reasonably constant among monetarists since the
mid-1950s.
Monetarism: An interpretation and assessment 217
Before discussing these characteristics of monetarism in detail, let me
deal briefly with two propositions that some might feel should be included
in the above list. First, on the one hand, monetarists have frequently been
accused of failing to give any account of the transmission mechanism of
monetary policy, and have had attributed to them a belief in some
mysterious ‘direct’ influence of money on expenditure; on the other hand
they have themselves sometimes referred to a characteristically ‘monetarist
model’ of that same transmission mechanism cast in terms of portfolio
substitution among a wide variety of assets including reproducible capital,
and even perhaps non-durable consumption goods. I believe that this is and
always has been a non-issue. The claim that monetarists have failed to
specify their transmission mechanism has been untrue from the very outset
(see, for example, Brunner 1961; Friedman and David Meiselman 1963;
Friedman and Anna Schwartz 1963b), and although the mechanism
propounded in those papers is a good deal more sophisticated and better
grounded in relative price theory than that embodied in the textbook
macroeconomic models of the 1950s, or in the econometric models of that
vintage, there is no essential difference between it and that analysed for
example by James Tobin and his associates.2
Second, monetarists are often said to prefer ‘small’ to ‘big’ econometric
models, and their views about the importance of the quantity of money for
the determination of the general price level have undoubtedly led them to
take highly aggregated systems seriously. Moreover, early large-scale
econometric models were not constructed so as to highlight any strong effects
of money on economic activity. Monetarists criticised them, as much for
being Keynesian, as for being ‘big’. Even so, subsequent developments have
clearly shown that ‘big’ models can easily take on some very monetarist
characteristics, while the Albert Ando and Franco Modigliani (1965) and
Michael De Prano and Mayer (1965) papers demonstrate that single
equation reduced form techniques can as well produce ‘Keynesian’ as
‘monetarist’ results.3 Empirical analysis of all sorts has been used by both
sides in the monetarist controversy, and if there is a method of empirical
research more frequently associated with monetarist work than Keynesian, it
is not small model or single equation econometrics, but National Bureau
techniques of business cycle analysis.4 Thus though empirical techniques
have, in specific instances, provided something to argue about, there seems
to me to be no clear dividing line between the statistical methodology of
monetarists and their opponents about which one can usefully generalise.
I THE QUANTITY THEORY OF MONEY
It has often been said that Friedman’s celebrated essay on the Quantity
Theory could just as well have been called ‘The Theory of Liquidity
Preference—a Restatement’. Harry Johnson (1962) argued that Friedman’s
work on the demand for money should be viewed as a development of a
218 David Laidler
fundamentally Keynesian capital theoretic approach to monetary theory and
Don Patinkin (1969) later documented that it was indeed just that. However,
I would stress the word development here, for ‘Keynesian’ though Friedman’s
model is, it is no more Keynes’ model than Keynes’ ‘Marshallian’ theory of
income determination is Marshall’s theory; and it differed from other
developments of Keynes’ theory of liquidity preference that appeared at
about the same time in a number of ways.
First, it abstracted from any specific characteristics that money might
have because it is a financial asset; Friedman treated money instead ‘as if’ a
service-yielding consumer durable to which the permanent income hypothesis
of consumption could be applied, just as Margaret Reid (1962) applied it to
housing, or the contributors to Arnold Harberger (1960) did to a variety of
other durable goods. In this respect Friedman’s approach stands in sharp
contrast to the analyses of William Baumol (1952) and Tobin (1956; 1958)
as it does in its claim to be a theory of the total demand for money in the
macroeconomy rather than of some component of that demand. Second,
Friedman explicitly recognised inflation as an own rate of return on money
and postulated a well-determined functional relationship between the
expected inflation rate and the demand for money, a relationship whose
existence Maynard Keynes (and some of his disciples) explicitly denied (see
Roy Harrod 1971).
Finally, and so obviously that the matter is usually overlooked,
Friedman asserted that the demand for money was, as an empirical matter,
a stable function of a few measurable arguments. Keynes did not believe
that—his empirically stable relationship was the consumption function—
and nor did (or perhaps do) many of his British followers.5 Moreover, pre-
Keynesian monetary theorists did not believe in an empirically stable
demand for money function either. Though they often enough assumed a
constant velocity of circulation that is by no means the same thing, and in
any event, they typically did so in order to make their analytic points with
the maximum of clarity, and not with the intention of stating a belief about
the nature of the real world. It is only with the publication of Friedman’s
essay that statements to the effect that the velocity of circulation is, as a
practical matter, a stable function of a few arguments become central to
debates about monetary economics. Its stress on this hypothesis makes
monetarism a very different doctrine from classical and neoclassical
economics, no matter what other similarities there may be, though it
should be noted explicitly that the econometricians among American
Keynesians have not found it necessary to adopt a monetarist label as a
result of contemplating the possibility of the empirical stability of the
relationship.6
In 1971 it was possible to argue that this characteristic monetarist belief
in a stable demand for money function was well supported by empirical
evidence as I did in Laidler (1971). However, the 1970s have produced a
good deal of evidence to suggest that the relationship has shifted in an
Monetarism: An interpretation and assessment 219
unpredicted way in a number of countries. There is not space to go into
details here, but I would be willing to defend the following assertions.7
First, the instability in question is often presented, particularly in the
United States, as a matter of a cumulative deterioration in the ability of the
function to track data. This cumulative deterioration is largely an illusion
stemming from the use of dynamic simulations of relationships containing a
lagged dependent variable. A one-time shift of such a function will, as a
matter of arithmetic, lead to a cumulative deterioration of its dynamic
simulation goodness of fit that should not be read as implying a continuous
tendency of the relationship to shift. On the other hand, I do not believe we
can safely conclude that such one-time shifts in the demand for money
function have not occurred, despite the fact, again particularly in the United
States, that some formulations of the relationship turn out to deteriorate
significantly less than others during the 1970s. When important issues like
the stability of the demand for money function begin to depend, for example,
on just which interest rate or rates one uses to proxy the opportunity cost of
holding money, I believe that the correct conclusion is not that the variable
which provides the best fit this time around is the ‘right’ one, but that our
knowledge of the details of the relationship is more fragile than we thought.
Finally, arguments to the effect that the demand for money function has not
‘really’ shifted, that we can restore its stability by taking note of institutional
change and redefining ‘money’ so as to take account of its effects, need to be
handled carefully. They are relevant to the interpretation of economic
history, but the successful conduct of policy requires that specific actions be
taken vis-à-vis precisely defined aggregates in order to achieve particular
policy goals. To say, after the event, that our policy did not work because
new assets evolved whose existence affected the outcome of our policies in a
way that we could have forecast had we only been able to foresee their
invention, may be true, but it is not very helpful in enabling us to do better
next time, unless the evolution in question was, as it sometimes can be, the
predictable outcome of some policy action or other.
Shifts in the demand measured for money function are not a new
phenomenon. Evidence drawn from more than one country shows that the
demand for money function shifted as the institutional framework evolved
long before 1974. To cite but four examples: the income elasticities of
demand for money seem to have fallen significantly in both the United States
and Britain in the twentieth century (see Laidler 1971), the abolition of
interest payments on demand deposits in the United States in 1933 was
associated with a change in the nature of the demand function for narrow
money (see Charles Lieberman 1980), as was the growth of Savings and
Loan Associations in the 1940s (see Cagan and Schwartz 1975), or in
Britain, the introduction of ‘Competition and Credit Control’ in 1971. Such
shifts in the demand for money function are not new, then, but they are
important. Though two of the above examples were the result of policy
changes and might have been predicted ex ante, two were not. In any event
220 David Laidler
these effects of institutional change on the demand for money function have
important implications for our views on the proper conduct of monetary
policy, as I shall argue in Section 4.
In the traditional vocabulary of economics, the phrase ‘quantity theory of
money’ referred to a theory of (or better an approach to the analysis of) the
relationship between the supply of money and the general price level. The
characteristic monetarist belief that variations in the supply of money are the
‘dominant impulse’ (to borrow Brunner’s phrase) causing fluctuations in
money income is clearly related to this traditional version of the quantity
theory, but modern monetarists are more clearcut in their attribution of a
dominant causative role to the money supply than were quantity theorists of
earlier vintages.8 The difference here is surely attributable to monetarists’
belief in a stable demand for money function, because earlier quantity
theorists spent much of their time contemplating the empirical possibility of
autonomous shifts in velocity. However, it takes more than a belief in a
stable demand for money function to yield the monetarist view of these
matters.
Setting aside the important complications that arise in the open economy,
there are two ways in which a conventional analytic model of the IS-LM
variety can be made to produce ‘monetarist’ results. First in its
underemployment form, if, relative to expenditure, the demand for money is
insensitive to interest rates then the quantity of money comes to dominate the
determination of the level of real income. Now obviously a monetarist must
deny that the interest elasticity of demand for money is infinite, and this has
been done often and explicitly, but it is mainly in Britain that such a denial
has been thought to amount to a distinctively monetarist statement. A
number of textbook writers (including myself) have gone to the other extreme
and used the assumption of a zero interest elasticity of demand for money to
generate monetarist propositions from an under-employment IS-LM model.
However, Friedman’s (1959) study of the United States function is a notable
exception to the general tendency of demand for money studies—including
those of such monetarists as Brunner and Meltzer (e.g. 1963)—to find a
significant interest elasticity of demand for money, and his inability to find a
relationship turned out to be the result of faulty statistical method (see
Laidler 1966; Friedman 1966). Thus, the existence or non-existence of a
statistically significant interest elasticity of demand for money has not been a
serious issue between monetarists and their opponents for at least fifteen
years. If it had been, it is hard to see how monetarists, not least Friedman
could have contributed to the analysis of the welfare costs of inflation, or
how Friedman and Meiselman could have accepted their own evidence of the
importance of autonomous expenditure as an influence on money income
during the depression years with such equanimity.9
If we rule out the vertical LM curve, we can still get an IS-LM model to
produce monetarist results if we assume full employment, and then postulate
that the major source of disturbance is variations in the level—or rate of
Monetarism: An interpretation and assessment 221
change of—the nominal money supply. With the determinants of velocity,
except the expected rate of inflation, thus pinned down at full employment,
and with fluctuations in money income thus reduced to fluctuations in the
price level, the characteristics of the demand for money function—other than
its stability and homogeneity in the general price level and its sensitivity to
fluctuations in the expected inflation rate– become quite irrelevant to the
relationship between the quantity of money and money income. A Keynesian
of course would agree, as an analytic matter, with this proposition, but
would probably deny what the monetarist would claim: namely that, if the
IS-LM model is to be used as a framework for discussion at all—and there
are some monetarists, notably Brunner and Meltzer, who would not want to
use it at any price—then this full employment version of it is frequently the
empirically relevant one.
To put matters this way is, in effect, to say that monetarists’ belief in the
quantity theory as a theory of money income boils down to the view that
sustained inflation is caused by an expanding money supply. This is not too
far from the mark, and much of the spread of monetarism since the mid-
1960s stems from its ability to provide a readily comprehensible
explanation of inflation along these lines. However, to cast the monetarist
approach to the analysis of inflation in terms of a ‘full employment’ IS-LM
model is difficult to justify except as a very first approximation. Though
monetarists are among those who have written at considerable length about
the interaction of the quantity of money and the price level in models
where ‘full employment’ is the rule, the models in question have been long-
run equilibrium growth models, not versions of short-run IS-LM analysis;
in any event the ‘money and growth’ literature and, to a lesser extent, that
dealing with ‘money and welfare’, even though it builds on Friedman’s
formulation of the relationship between the demand for real balances and
the expected rate of inflation as a well-defined inverse function, is properly
viewed, not as an offshoot of monetarism, but as an extension of Patinkin’s
(1956) theoretical analysis of the classical dichotomy and the neutrality of
money to deal with the long-run properties of a growing economy, in the
presence of variations in the rate of change of the nominal money supply.10
In dealing with the interaction of the quantity of money, money income,
and prices, the essential monetarist contribution has been to postulate the
existence of stable relationships among these variables as an empirical
matter, and to draw practical conclusions about the proper conduct of
short-run stabilisation policy from studying their nature, and the ‘money,
growth and welfare’ literature has next to nothing to say about these
matters.
When it comes to empirical propositions about the relationship between
money and money income, what was once monetarist heresy is now close to
being received orthodoxy. In this respect monetarism has made an important
positive contribution to macroeconomics. In the United States it seems now
to be widely accepted that the correlation between the quantity of money and
222 David Laidler
money income that long runs of time series data display is not just the result
of coincidence, but does in fact constitute evidence for the existence of a
causative relationship that has run primarily from money to money income
rather than vice versa. The weight of the evidence produced by Friedman
and his various collaborators (not to mention predecessors) and the
persuasiveness of their arguments, has changed enough minds to warrant the
conclusion that, in an important sense, ‘we are all monetarists’ now.
Elsewhere in the world, not least in Britain, there has been a similar
movement of opinion. Certainly one no longer hears much about velocity
being variable ‘almost without limit’. However, one does hear more about
‘reverse causation’ in Britain as an explanation of the correlation between
money and money income than one does in the United States (I shall take
this matter up below).
Even so, monetarist doctrine asserts not just that variations in the
quantity of money lead to systematic variations in money income, but also,
that those variations are primarily in prices rather than real income.
Although, as I have already noted, much of monetarism’s popular appeal
stems from its claim to provide an easily comprehensible theory of
inflation, that theory of inflation is by no means universally accepted. The
view that the influence of money on money income falls on its real income
component and not on prices has constituted a ‘Keynesian’ alternative to
the monetarist position on these matters and the ‘expectations-augmented
Phillips Curve’ has provided a focus for debate about them.11 That is why a
particular set of beliefs about its nature is a vital ingredient of monetarist
doctrine.
II THE EXPECTATIONS AUGMENTED PHILLIPS CURVE
The notion of a trade-off between inflation and unemployment was widely
prevalent in Keynesian literature even before Arthur Brown (1955), William
Phillips (1958) and Richard Lipsey (1960) formalised it in terms of what
seemed to be an empirically stable functional relationship. Monetarists have
long doubted its existence, instead asserting a belief in the ‘inherent stability’
of the private sector in the absence of policy-induced monetary disturbances,
by which they have usually meant nothing more complex than that the
system tends in and of itself to operate at or near ‘full employment’,
regardless of the inflation rate, if policy-makers do not upset matters. The
papers of Edmund Phelps (1967) and Friedman (1968) provided a framework
in terms of which differences of opinion about these matters could be stated
sharply enough to be confronted with empirical evidence. Although some
commentators (e.g. Helmut Frisch 1977) treat the Phillips curve as providing
an alternative theory of inflation to the monetarist approach, this is surely a
mistake. In its expectations-augmented form, it emerged at the turn of the
decade to provide what Friedman (1970) called ‘the missing equation’ in the
monetarist model of inflation.
Monetarism: An interpretation and assessment 223
It is possible to derive this ‘missing equation’ from two very different
theoretical bases, and disagreements here are of quite fundamental
importance for macroeconomics, but the first round in the debate about the
expectations augmented Phillips curve, and the one that was crucially
relevant to monetarism, paid little attention to these matters. It was almost
entirely empirical because the relationship in question enabled alternative
viewpoints about important and pressing policy issues to be formulated and
investigated in an easily manageable way. With ∆p the inflation rate, ∆pe
the expected inflation rate, and y some measure, either direct or indirect, of
the deviation of output from its ‘full employment’ level, and v a ‘catchall’
vector of other influences, systematic as well as random, the general form of
the relationship may be written as follows:
∆p=gy+b∆pe+v. (1)
A whole spectrum of beliefs about the nature of inflation may be expressed
in terms of this simple equation, depending upon the values assigned to its
parameters. Thus, the extreme ‘sociological’ view of the determination of
the price level, that was widely prevalent in Britain in the early 1970s,
would predict that the parameters g and b were essentially equal to zero,
implying that monetary policies, if they had any effect on money income,
would influence real income alone.12 The behaviour of prices, in this view,
was determined by exogenous factors that would all go into the catchall
vector v. At the other extreme, typical monetarists of the early 1970s would
argue that g was positive, so that inflation would, relative to expectations,
be low in a depressed economy, and high in an over-expanded one. They
would also argue that the coefficient b on expected inflation would be
equal to unity, and would supplement equation (1) with some formula for
the formation of expectations, typically based on the error learning
hypothesis, that ensured that, eventually, any constant actual inflation rate
would come to be fully anticipated. For them, therefore, any trade-off
between inflation and deviations of output from full employment was a
temporary one which vanished in the long run. The typical ‘American
Keynesians’ of the same vintage would agree with the monetarists about
the parameter g, and about the reasonableness of assuming that
expectations would eventually catch up with experience, but would assign
a value of less than unity to the parameter b, thus ensuring that though the
price in terms of inflation of increasing output was higher in the long run
than in the short run, it did not, as the monetarists asserted, ever become
infinitely high.13 They might also argue that equation (1) omitted to
mention explicitly many factors that in particular times and places might
have an important influence on the inflation rate, and which it will suffice
here to think of as being captured in v.
There is not space here to survey the extensive empirical literature that
these issues generated, but its upshot may be summarised easily enough. The
evidence that, other things equal, inflation varies with the level of aggregate
224 David Laidler
demand is overwhelming. To the extent that differences of opinion here ever
set monetarism apart from other points of view—and I think they probably did
in Britain, though not in North America—then surely we have here another
case of ‘we are all monetarists now’.14 There has also been a swing towards
the typically monetarist belief that in the long run there is no economically
significant inflation-output trade-off. The more rapid inflation of the 1970s and
the more sophisticated methods of modelling expectations developed over the
same period have provided empirical evidence of a type that we did not have
in 1971 to support this belief. There is still substantial disagreement though on
the question of how fast the economy converges on the long-run solution.
Finally there is more of a consensus about the importance of the influence of
‘other’ factors on the inflation rate than there was. Monetarists are willing to
agree that factors such as the activities of OPEC, unexpected real shocks, or
sudden changes in the level of indirect taxes, can affect the behaviour of the
price level ‘temporarily’ against the background of long-run trends determined
by monetary factors; Keynesians, particularly American ones, in their turn are
now willing to agree that the long-run trend of inflation may well be
determined by monetary factors while continuing to stress the importance of
special factors for the short run. However, as we shall now see, there is much
less of a consensus about the theoretical basis of the Phillips curve than there is
about its empirical properties.
As originally analysed by Lipsey, the Phillips curve dealt with the reaction
of the money wage to the existence of a general condition of excess demand
for labour in the economy, and therefore of the general price level to the
excess demand for goods. Excess demand was conceived of, not as a purely
ex ante notion such as we meet in theoretical analyses of Walrasian
tatonnement, but as a realised quantity such as appears in models of
economies made up of markets characterised by sticky prices. In their
original critiques of the Phillips curve, Friedman (1968) and Phelps (1967)
both concentrated on the point that disequilibrium in the labour market
might be expected to bring pressure to bear on real wages rather than on
money wages per se, and that what would happen to the latter would
therefore be critically influenced by what was thought to be happening to the
general price level. Each of them, though Phelps more explicitly so than
Friedman, treated unemployment as a quantity signal that conveyed to
economic agents the desirability of varying prices, and hence seemed to be
providing a crucial correction to what remained a fundamentally Keynesian
approach to the analysis of wage and price stickiness.15
On the other hand, most of the contributors to the well-known Phelps
(1969) volume started from a very different theoretical basis to provide
an explanation of the interaction of output and prices, though the
similarity of their conclusions to those stated by Phelps and Friedman at
first distracted attention from what in retrospect was the much more
important theoretical matter of different premises.16 According to this
alternative approach, which was anticipated by Irving Fisher (1911), the
Monetarism: An interpretation and assessment 225
expectations-augmented Phillips curve is in fact an aggregate supply
curve. Equation (1) is derived from
y=(l/g)(p-pe) (2)
combined with the following definition of the expected rate of inflation
∆pe=pe-p-1 (3)
Brunner and Meltzer were quick to adopt this interpretation of the
expectations augmented Phillips curve. They had already developed a view
of the transmission of monetary impulses in asset markets that stressed the
role of relative prices as signalling devices, and found it easy enough to
extend that line of reasoning to the markets for output and labour services as
well (see Meltzer 1969).17 By now there can be no doubt that this aggregate
supply curve interpretation of inflation employment interaction is the
dominant one among monetarists. However, not all monetarists have
accepted it (see, for example, Cagan 1979), and as I shall now argue, it
raises issues that go well beyond the traditional subject matter of the
monetarist debate.
To say that the Phillips curve is an aggregate supply curve is to say that
fluctuations in output and employment in response to price level variations
represent the voluntary choices of individuals operating in markets which are
continually clearing. Since voluntary choices made on the basis of erroneous
expectations are by no means the same thing as choices that lead to the
outcome which agents would have desired, this is not to deny that deviations
of output and unemployment from the ‘natural’ levels they would attain if
expectations were fulfilled represent a serious problem.18 However, it is to
locate the cause of unemployment, not in the failure of markets to bring
together all willing buyers and sellers in ex ante mutually satisfactory
trades, but rather in a failure of markets (and other social institutions as well
perhaps) to convey sufficient information to enable the expectations upon
which those trades are based to be formed accurately in an economy
subjected to stochastic shocks.
If fluctuations in output and employment about their natural rates are the
result of the failure of expectations to be realised, the manner in which
expectations are formed must play a vital role in their analysis. That is why
the ‘rational expectations’ hypothesis is a natural supplement to the
aggregate supply curve interpretation of the Phillips curve. If agents suffer
losses in utility as a result of making expectational errors, they have an
incentive to use all available information in forming their expectations up to
the point at which the marginal benefit from improving their accuracy
equals the marginal cost of doing so. The rational expectations hypothesis
does not say that every agent’s expectations are always as accurate (i.e. have
as small a variance) as they would be if agents were equipped with a ‘true’
econometric model of the economy in which they operate (though it is
226 David Laidler
sometimes convenient to formulate it that way in analytic and empirical
exercises), but it does say that their expectations will not be wrong
systematically over time and to that extent will resemble those generated by
such a ‘true’ model in being unbiased and serially uncorrelated. Agents who
form expectations in a manner that leads to systematic error will find
themselves persistently making the wrong choices; hence in the very course
of their market activities, they will be provided gratis with the information
necessary to eliminate that systematic error.
If each individual makes only random errors in forming expectations, two
questions naturally arise: how does it happen that at a particular moment the
expectations of a predominant number of agents in the economy should be in
error in one particular direction so that aggregate output and employment
come to deviate from their ‘natural rates’, and how does it happen that the
fluctuations in output and employment which are observed in any actual
economy come to display that pattern of serial correlation summarised in the
term ‘business cycle’? The answer to the first question given by Lucas (1972) is
by now well known. If individuals have more up-to-date information about the
money prices that rule in the markets in which they operate as sellers than
about others, then in order to assess the pattern of relative prices upon which
their quantity decisions rest, they must form expectations about the behaviour
of other money prices. An unforeseen shock affecting the whole economy
which leads to a change in the general price level will influence individual
money prices, and will have its consequences everywhere misread as reflecting
relative price changes. Hence quantities supplied will everywhere change.
If that was all there was to it, output and employment fluctuations would
be random over time. However, if there are time delays in getting
information to agents, if there are costs of adjusting output decisions once
taken, or if some of the goods over-produced in error in the face of a positive
unexpected shock to the price level are durable, then the effects of that shock
will persist over time.19 By the time its effects on output have petered out
there will be too many durable goods in the economy—capital will be ‘too
deep’—and the marginal productivity of labour in terms of consumption in
industries producing durable goods will fall. If workers prefer to take extra
leisure when their marginal productivity is low, and if the price system
operates so as to inform them of when that is the case, there will be a
voluntary fall in the level of employment that will persist until the structure
of the economy’s capital stock is restored. The objection to this explanation
of the cycle, that it predicts more wage variability than we observe in the
real world, can be countered by arguments to the effect that firms and
households find it mutually beneficial to enter into wage contracts under
which wages do not instantaneously fluctuate in tune with the marginal
productivity of labour, but under which firms are permitted to lay off
workers in such a way that the latter still take more leisure at times when
their marginal productivity is unusually low, even though the behaviour of
wages no longer signals the fact.
Monetarism: An interpretation and assessment 227
Readers will find the last paragraph reminiscent of the Austrian business
cycle theory of the 1920s and 1930s, and that is no accident. It is the
Austrians, and not, as Solow (1980) has suggested, Pigou, who are the
predecessors of Lucas, Sargent and their associates. Like Ludwig von Mises
and Friedrich von Hayek, they have set themselves the task of producing a
theory of the business cycle that is firmly based on the notion that all market
phenomena represent the harmonious outcome of the voluntary choices of
maximising individuals. However these neo-Austrians have gone beyond
their predecessors to produce a theory in which output and employment as
well as prices fluctuate as a result of such voluntary choices. Whatever we
may think of the empirical relevance of that theory, and its proponents show
an admirable, and un-Austrian, willingness to submit their ideas to empirical
tests,20 we must surely agree that its very construction represents an
intellectual achievement of the highest order.
One can admire a theory without agreeing with it, and there are many
including myself who would challenge the basic assumption upon which the
analysis just discussed is based, namely that it is legitimate to model the
economy ‘as if ’ markets always clear. It is one thing to agree that
commodity and asset markets dominated by specialist traders ought, and
indeed do, display the characteristics associated with continuous clearing
and rational expectations, and quite another to attribute similar
characteristics to the markets for many components of final output, and
above all to the labour market. One may follow Hicks (1974) in
distinguishing between ‘flexprice’ and ‘fixprice’ markets, assign the labour
market to the latter category, and argue that the interaction of inflation and
unemployment is best analysed on the premise that the Phillips curve
represents the disequilibrium response of prices to a mismatching of supply
and demand.
Of course the ‘neo-Austrians’ are well aware that there is no Walrasian
auctioneer to set prices, and no recontracting to ensure that trade takes place
only at market clearing prices; but they do assert that individual agents—or
their representatives—are acute enough in their bargaining to ensure that
money wages and prices universally behave ‘as if markets operated along
such Walrasian lines, that they perceive the possibility of realising mutual
gains by adjusting wages downward when excess supply turns up in the
market in which they are operating, and act upon that perception.21 However
one can have no difficulty accepting the proposition that, even in labour
markets, if it is mutually beneficial to lower money wages (or their rate of
change), agents will discover this and will agree to do so, but still find it
hard to understand how the relevant information is conveyed to the agents in
question without the intervention of quantity signals. In a Walrasian market,
the auctioneer can discover that the price is too high by adding up notional
supplies and finding out that they exceed notional demands, to use Robert
Clower’s (1965) terms, but how can participants in any actual labour market
find out that money wages there are too high without some of them
228 David Laidler
discovering first that they are unable to sell all the services that they would
like at the going rate?
If adjustments in the level (or rate of change) of money wages and prices
to aggregate demand shocks are anything other than instantaneous, then
markets fail to clear, trade takes place at false prices, and quantity signals,
perhaps amplified by multiplier effects, become an integral part of the
mechanism whereby monetary changes are transmitted to the behaviour of
the price level. This line of analysis is as ‘Keynesian’ in spirit as the clearing
market approach is ‘Austrian’, and its existence permits one to subscribe to
the expectations-augmented Phillips curve without also being committed to a
clearing market rational expectations approach to the analysis of economic
fluctuations. Moreover, the approach in question does not differ from the
clearing market view in denying that individuals perceive and then engage
in all available mutually beneficial trades. It simply denies that they do so
infinitely rapidly. I do not see why, as for example Barro (1979) has
suggested, to postulate an infinite speed of price adjustment in the face of
excess demand or supply is to conform to sound microeconomic principles,
and to postulate anything significantly slower is to propose an ‘ad hoc non-
theory’.
The non-clearing market approach to analysing inflation employment
interaction is not obviously incompatible with the notion of rational
expectations. If output fluctuations convey information about the appropriate
behaviour concerning price setting as this approach suggests, they can be
regarded as constituting one of the ingredients of the expectations upon
which such behaviour is based. In that case the term ∆pe in equation (1) can
be thought of as summarising influences upon expectations other than
quantity signals.22 To say this begs the question of what those ‘other
influences’ on expectations might be, but leaves open the possibility that the
same type of information to which the rational expectations hypothesis
draws our attention could be incorporated without difficulty into models
based on the non-clearing market approach. Observations on the past
behaviour of the money supply, for example, might well provide agents with
information about the appropriate way to set prices, and might be included
among those ‘other’ influences, as might, in an open economy, variations in
prices ruling elsewhere in the world economy, variations in exchange rates,
and so on.23
The non clearing-market interpretation of the Phillips curve needs to be
reconciled with the basic facts of the business cycle. Once given, why do
output signals not result in an immediate adjustment of prices to a market
clearing level? The answer here is straightforward—a quantity signal will
lead to a response in price behaviour only to the extent that agents believe
that the shock which gave rise to it will persist into the future. Inability to
disentangle short-term from persistent shocks will lead to a tendency to
underreact to quantity signals, and hence to cause them to be drawn out over
time. I would conjecture that the Austrian-style arguments about the role of
Monetarism: An interpretation and assessment 229
errors made with respect to the production of durable goods in the business
cycle can be superimposed upon this fundamentally Keynesian explanation
of the persistence of output fluctuations should anyone wish to do so.24
Although theoretical analysis of the interaction of output, employment
and prices in terms of an expectations-augmented Phillips curve can thus
proceed along two very different lines, it is a mistake to treat debate about
these issues as simply the latest round in the monetarist controversy. Though
monetarists and Keynesians are in much closer agreement than they were
about the empirical stability of the demand for money function, and about
the empirical nature of output-inflation interaction, they still take the same
diametrically opposed views on the proper conduct of macroeconomic policy
that they did in the mid-1950s, and divisions of opinion here do not, as I
shall argue below, depend upon differences of views about the theoretical
basis of price-output interaction. Since the policy debate is undoubtedly a
continuation of the monetarist controversy, and since disputes about the
theoretical basis of the Phillips curve clearly deal with a new set of issues, it
seems to me to be misleading to treat what I have here termed the neo-
Austrian view as synonymous with monetarism, as for example Frank Hahn
(1980) does. I shall discuss the policy aspects of the monetarist controversy in
Section 4, but before I do so, it will be convenient to discuss the place of the
monetary approach to balance of payments and exchange rate theory in
monetarist doctrine.
III THE MONETARY APPROACH TO BALANCE OF PAYMENTS
AND EXCHANGE RATE ANALYSIS
The monetary approach to balance of payments and exchange rate analysis
represents in some respects a revival of the English classical approach to
these problem areas. However, the monetary approach differs in important
ways from classical analysis, and the very characteristics that thus
distinguish it are borrowed from closed economy monetarism.25 Most
important, advocates of the monetary approach postulate the existence of a
stable demand for money function, not just as a working simplification, but
as an empirical hypothesis; it is this hypothesis that transforms the approach
from an accounting framework into a body of substantive theory.
Furthermore, in early statements of the doctrine, its proponents tied down the
real income argument of that function by assuming full employment but they
soon learned how to replace this assumption with an expectations-augmented
Phillips curve approach to price-output interaction.26 In effect the monetary
approach to balance of payments and exchange rate analysis provided the
means whereby these characteristically monetarist hypotheses were made
relevant to economies other than the United States which, under the Bretton
Woods system, was about as close an approximation to a closed economy
that was also a separate political entity as the world has ever seen.
Monetarism thus came to be important only outside the United States, not
230 David Laidler
least in Britain, in alliance with the monetary approach to balance of
payments and exchange rate analysis.
Until 1971 the world was on a system of fixed exchange rates against the
United States dollar. Under such a system the existence of a stable demand
for money function, whose arguments are beyond the direct control of the
domestic authorities, implies that the money supply is an endogenous
variable that must adjust to demand. Given this insight, evidence that
suggests, for example in the United Kingdom in the 1950s and 1960s, that
causation seems to have run predominantly from money income to money,
rather than vice versa, is no embarrassment to monetarists provided that
they are also willing to attribute most of the variation in money income to
causative factors originating abroad. Moreover, although the expectations-
augmented Phillips curve tells us that in general we should expect to find no
stable inverse trade-off between inflation and unemployment, post-war
United Kingdom data do display just such a well determined relationship
down to 1967, and this fact needs explaining. The monetary approach to
balance of payments analysis suggests two complementary reasons why this
should be the case. First it notes that, so long as a fixed exchange rate is to
be maintained, the prices of tradable goods sold domestically are going to be
determined in the long run, not domestically, but on world markets, and
from this it follows that the domestic price level’s long-run behaviour is
going to be constrained by the behaviour of prices in the world at large.
Economic agents do not have to be more than merely sensible to perceive
this fact and to incorporate it into their expectations. If world prices are
relatively stable, and they were until the late 1960s, then so are inflation
expectations, and our expectations-augmented Phillips curve, equation (1),
no matter how we interpret its microeconomic origins, will predict that the
data will generate a stable inflation-unemployment trade-off.
This explanation of the existence of a stable inflation-unemployment
trade-off in post-war Britain is an important component of what may fairly
be called monetarist hypotheses, about the nature of the stop-go cycle in the
1950s and 1960s and about the degeneration of that economy’s performance
in the 1970s, which contrast strongly with conventional ‘Keynesian’ accounts
of the same phenomena. The latter begin from the proposition that Britain
has a peculiarly high marginal propensity to import, so that, under the
Bretton Woods system, attempts to run the economy at a high degree of
capacity utilisation, though they produced only a small and on the whole
acceptable amount of inflation, were frustrated by balance of payments
pressure which forced a reversal of policy. The monetarist hypothesis about
stop-go, on the other hand, has it that high levels of demand were associated
with high rates of domestic credit expansion which, under fixed exchange
rates, generated balance of payments problems in large measure as an
alternative to domestic inflationary pressure. The conventional view seemed
to imply that Britain’s economic performance could be improved by adopting
exchange rate flexibility and allowing a depreciating currency to offset the
Monetarism: An interpretation and assessment 231
balance of payments effects of a high propensity to import. With a flexible
exchange rate, the economy could be run at a higher level of capacity
utilisation and could grow more rapidly without interference from a balance
of payments ‘constraint’. According to this view a series of exogenous shocks
and the autonomous activities of trade unions undermined a basically well-
founded strategy when it was adopted in the 1970s. The monetarist view, on
the other hand, argues that the adoption of exchange rate flexibility replaced
a balance of payments problem with a domestic inflation problem when
expansionary policies were pursued, and did nothing to influence the
economy’s ability to sustain either a higher level or rate of growth of real
income. For the monetarist, therefore, the deterioration of British economic
performance after 1972 was the predictable (and predicted) consequence of a
policy of expanding aggregate demand against a background of exchange
rate flexibility.27
Now the monetary approach to balance of payments analysis does far
more than make monetarist analysis relevant to Britain. It also permits the
explanation of the international spread of inflation in the late 1960s in terms
of the repercussions in the world economy of United States monetary
expansion, and it treats the breakdown of the Bretton Woods system as the
culmination of this process. However, it is only fair to note that such
analysis performs less well in the face of the behaviour displayed by the
international monetary system since exchange rates began to float in the
early 1970s. The prediction that the behaviour of exchange rates can be
analysed fruitfully as if determined in efficient asset markets does seem to be
supported by the data. However, a basic postulate of the monetary approach
is that the equilibrium value of the exchange rate between any two
currencies reflects purchasing power parity. Just as data generated under
fixed rates show that the price levels of particular economies can display
considerable autonomy for substantial periods of time, so under flexible
exchange rates systematic and persistent deviations of exchange rates from
purchasing power parity do seem to be possible. Though purchasing power
parity considerations underlie the behaviour of long period averages of data,
implying that, ultimately the terms of trade between countries are
independent of monetary factors, there seems to be ample room for short-run
deviations from the long-run pattern. Just why this should be the case, and
what explains the patterns of such deviations as we observe, are important
and, at the moment, open questions.28
Be all that as it may, the present regime of flexible exchange rates came
into being because the authorities in various countries learned that they
could not control such politically important variables as domestic inflation
and unemployment while continuing to adhere to the Bretton Woods
arrangements. The diversity of inflation rates among countries since 1971
supports the view that the adoption of flexible rates allows such variables to
have their behaviour predominantly determined at home; and long before the
1970s, monetarists, not least of course Friedman, argued that the adoption of
232 David Laidler
exchange rate flexibility was a necessary prerequisite to the pursuit of
monetarist policies in individual countries. In the 1970s we have seen the
emergence of conditions under which individual countries could implement
independent monetary policies, and as I have suggested above, it is mainly
on the matter of policy prescriptions that sharp differences between
monetarists and their opponents persist. I shall therefore devote the
penultimate section of this chapter to a discussion of these matters.
IV POLICY ISSUES
As we have seen, when it comes to propositions about the demand for money
function, the relationship between money and money income, and output
inflation interaction, there is a real sense in which ‘we are all monetarists
now’. The issues that nowadays distinguish monetarists from their opponents
concern the conduct of economic policy. As he did in the 1950s the
monetarist still wants fiscal policy to stick mainly to its traditional tasks of
influencing resource allocation and the distribution of income and wealth,
and monetary policy to adhere to some simple rule under which the
monetary aggregates do not react to short-run fluctuations either in real
output or prices; the Keynesian on the other hand is still a proponent of
activist stabilisation policy.
These policy issues are not independent of the theoretical questions that
we have discussed earlier, and indeed, much of the current popularity among
monetarists of the neo-Austrian approach to the analysis of price-output
interaction stems from the erroneous belief that it provides the only sound
basis for scepticism about the effectiveness of activist stabilisation policies.
Many Keynesians focus their attacks on that same piece of analysis in the
belief, just as erroneous, that if they succeed in refuting it, they also succeed
in restoring the case for activist stabilisation policy. Now the approach in
question does indeed imply that output and employment can be influenced by
policy only to the extent that it causes prices to vary in a way that agents in
the private sector do not foresee, while the rational expectations hypothesis
tells us that if such effects were systematic, the private sector would discover
the fact, adapt to it, and thereby render policy ineffective. It follows at once
that the only macroeconomic policy that can influence income and
employment is a purely random one, and no supporter of ‘fine tuning’ could
possibly recommend that.
The argument just sketched out is logically watertight. So is this counter-
argument: if inflation-output interaction reflects the role of quantity signals
in the mechanism whereby various shocks, including those imparted by
policy, have their effects transmitted to prices, the way is opened for
monetary and fiscal policy to exert a systematic influence upon output and
employment. However, there is much more than this to be said about the
feasibility and desirability of activist policies. If there was not, how could it
be that Friedman (1960) was able systematically to state his views on policy
Monetarism: An interpretation and assessment 233
more than a decade before Lucas (1976) and Sargent and Neil Wallace
(1975) developed the theoretical arguments that are now so widely regarded
as the only logical underpinning of those views? The Lucas-Sargent-Wallace
analysis certainly provides a sufficient basis for monetarist policy
prescriptions, but it is not a necessary basis for them: it is one thing to say
that the world is so structured that policy can systematically influence output
and employment in the short run, and another thing altogether to say that
policy makers have enough knowledge to use that ability in a way that will
be beneficial.
If it is agreed that in the long run the Phillips curve is essentially
vertical—or perhaps even positively sloped if allowance is made for super-
non-neutralities—then that certainly does not rule out the possibility of the
economy slipping below its natural rate of output in a short run that may be
of considerable duration, or the possibility that there exists an appropriate
menu of monetary and fiscal policies that might hasten its return to that
natural rate without generating any serious costs during the transition. As a
first step to exploiting this possibility though, those in charge of policy would
need to know what the natural rates of output and employment actually are.
As a second step, they would need accurate information upon where the
economy actually is, and where it would move in the absence of a policy
change, not to mention at what pace. Armed with this not inconsiderable
amount of information, policy makers would know that they were in a
position where it might be useful to deploy some policy measure or other. To
design the policy would of course require them to know about the size and
time path of the economy’s response to the measures they might take, factors
which even the loosest application of the rational expectations idea tells us
are likely to be influenced by the policy measures themselves.
Now I will readily agree that we have the mathematical and statistical
tools available for tackling the design of stabilisation policy along the
foregoing lines, and I also agree that our econometric models contain
answers to all the quantitative questions that I have just raised. However the
conclusion that I draw from all this is that we are probably rather good at
fine tuning econometric models.29 One can rest the monetarist case against
activist policy on the proposition that markets always clear and that
expectations are rational, but one can also rest it on the much more down-to-
earth proposition that we are too ignorant of the structure of the economies
we live in and of the manner in which that structure is changing to be able
safely to implement activist stabilisation policy in the present environment,
or in the foreseeable future.
Among the penalties for making errors in fine tuning that concern
monetarists are those that come in the form of uncomfortably high and
perhaps accelerating inflation that would result from setting over-optimistic
targets for employment and output. Thus, if there is something in the policy
environment that weakens the ability of the inflation rate to accelerate, the
penalties for such errors are milder, and the case against fine tuning
234 David Laidler
developed above can be softened a little. In the 1950s and 1960s, there can
be little doubt that the British authorities did succeed in fine tuning income
and employment variables within the rather narrow bounds laid down by
what then appeared to be balance of payments constraints. The monetarist
interpretation of that period implies that the background of monetary
stability implicit in the commitment to a fixed exchange rate was the real
constraint on how far fine-tuning policy could be pushed and also that it
provided the necessary conditions for its limited success. However the fact
remains that the experience in question does show that a limited degree of
fine tuning is feasible if only a background of long-run price stability is
assured, and is seen to be assured.
It is hard for a monetarist to see how one could avoid assigning to
monetary policy the role of providing that necessary assurance.30 A fixed
exchange rate regime is one way of tying down monetary policy, and the
adoption of some sort of a money supply growth rule would be an
alternative. But this means that fine tuning would have to be by fiscal policy.
Such a conclusion will be of little consolation to American Keynesians who
are forced by the inability of American institutions to deliver rapid changes
reliably in fiscal variables to assign to monetary variables a far more
important role in stabilisation policy than their British counterparts ever did.
However it may do a little to cheer up the British, for whom a return to the
days of ‘never had it so good’ might be a welcome relief from the
consequences of ‘going for growth’.
As should be apparent from the last few paragraphs, I regard the question
of whether governments should or should not indulge in a limited amount of
fiscal fine tuning as a secondary issue for monetarists.31 Related questions
concerning public sector borrowing and the share of the public sector in
National Income are even more peripheral to the monetarist debate. No
matter what the public perception of these matters might be, I insist that
monetarist doctrine tells one that there are severe limits to the extent to
which public sector borrowing can be financed by money creation, and
beyond that has nothing to say about whether a ‘high’ or ‘low’ level of such
borrowing is in and of itself desirable. Similarly monetarism offers no
guidance as to how big the public sector of any economy ought to be. It is a
macroeconomic doctrine and the issues at stake in debates about the size of
the public sector, the welfare state, and so on are fundamentally
microeconomic in nature.
Monetarism however has had a good deal to say about wage and price
control policies. It has opposed them, not just for ideological reasons, but for
the much more down to earth reason that they have not been expected to
work.32 This position has been mainly and justifiably defended on the basis
of empirical evidence: in the post-Korean war period it is hard indeed to find
any wage-price control scheme that has not produced disappointing results
over any period longer than a few months. However monetarists have also
sometimes opposed controls on theoretical grounds, particularly in the
Monetarism: An interpretation and assessment 235
context of open economies. They have noted that under fixed exchange rates
the behaviour of world prices and hence the domestic prices of traded goods
cannot be controlled by domestic regulations, any more than can the money
supply. They have also pointed out that under flexible rates, though the
money supply is under control, neither the exchange rate nor world prices
can be regulated separately. In either case in an open economy wage and
price controls inevitably impinge upon ‘the domestic component’ of the price
level and are hence policies towards relative prices. For that reason, they
cannot for long influence the behaviour of the general price level, unless they
are accompanied by a battery of quantitative restrictions, not least on
foreign trade, that very few of their advocates have been willing to
contemplate.
In the 1960s wage and price controls came to be regarded as an
alternative to monetary policy in the control of inflation, and in the early
1970s serious attempts were made in both Britain and the United States to
use them as such. In both cases the attempts failed sufficiently dramatically
that the proponents of controls now regard them at best as supplementary
devices to be deployed in harmony with more traditional demand side
policies rather than as a serious alternative to such measures. Though such a
viewpoint stops short of the blanket opposition to controls that, along with
other monetarists, I would still be willing to defend, it does represent a
substantial move in a monetarist direction from positions taken in the early
1970s. Here, as in other instances, much of the heat has gone out of the
monetarist controversy.33
There is more to practical monetarism than scepticism about fiscal fine
tuning and opposition to wage and price controls. Its key positive tenet is
that monetary weapons should be assigned to the attainment and
maintenance of long-run price stability, and hence that those same monetary
weapons not be used for fine-tuning purposes. In this respect, as with the
other components of the doctrine which we considered earlier, there has been
a considerable growth in the acceptance of monetarism. Propositions about
the desirability of setting rules and targets for the growth of monetary
aggregates are now commonplace in the statements of central banks. If
monetarists complain—and they do—about the failure of Keynesian policies
since the mid-1960s, then simple fairness requires them to say something
about the lessons that they have learned about the viability of their own
policy proposals from what many observers believe to have been widespread
and sustained efforts to apply them during the 1970s.
The first thing to be said on this score is that the case for monetary
growth-rate rules, as initially stated by Friedman (and Edward Shaw) was
put in terms of the capacity of such a policy to maintain stability in an
already stable economy—it was a policy prescription for staying out of
trouble. However, it has been only since our economies have found
themselves deeply in trouble that monetarist policy proposals have attracted
the attention of policy makers. There is much less unanimity among
236 David Laidler
monetarists about how to tackle the problem of restoring stability than there
is about how to maintain it. Though all monetarists would agree that a
return to a modest growth rate of some monetary aggregate or other is the
long-run goal, the neo-Austrians would favour a rapid return to such a rule,
while those of us who take a more traditional view of the nature of the
Phillips trade-off have advocated ‘gradualism’.
Unless we take the cynical view that the rhetoric of central bankers
bears no relationship to their intentions, we must conclude that in a
number of places attempts have been made to implement gradualist
policies. There are two questions to be asked about those attempts: first,
is it the case that those attempts have resulted in a systematic and
gradual reduction in the rate of growth of any monetary aggregate?
Second, if such attempts have anywhere been successful, did that success
lead to a reduction in the inflation rate? As is well known, policy has in
the main failed on the first count. Only in Canada, to the best of my
knowledge, have the authorities set, and on the whole succeeded in
achieving, pre-stated monetary growth targets over an extended period. It
is equally well known that the single most important reason for this
f a i l u r e , a t l e a s t i n t h e U n i t e d S t a t e s a n d B r i t a i n , h a s b e e n t h e
unwillingness of those in charge of monetary policy to give up setting
interest rate targets when they adopted targets for the money supply,
combined with a proclivity to stick with the interest rate target when the
two came into conflict, as they inevitably had to sooner or later. This has
not been universally the case, however. Germany and Switzerland have
had difficulty sticking to money supply targets because of concern with
the behaviour of the exchange rates rather than interest rates, as Sumner
(1980) has noted, while political concern over the exchange rate and
interest rates during the winter of 1979–80 posed a serious threat to the
continuation of the Canadian experiment.
It would be easy enough to argue in the light of all this that recent
experience offers essentially no test of monetarist gradualism, but that seems
to be going too far. Monetarists have usually treated questions of income
distribution and resource allocation as separate and distinct from those of
monetary policy. This dichotomy is a useful one when the problem for
monetary policy is to maintain already existing stability, but can all too
easily lead one to neglect the way in which monetary policy interacts with
allocation and distribution when its implementation requires sharp (albeit
temporary) increases in interest rates. A key factor here is of course the
political importance of the housing market, and of the behaviour of
mortgage interest rates. In retrospect, it is clear that monetarists did not do a
very good job of educating policy makers—both elected and otherwise—
about the problems that adopting monetarist policies would generate in this
area. Some of us did raise these matters, but apparently not loudly enough.34
High interest rates have turned out to be more difficult for politicians to face
up to than high unemployment rates, and that was not foreseen.
Monetarism: An interpretation and assessment 237
There are also technical problems with implementing monetarist policies.
The manipulation of interest rates as the centrepiece of monetary policy long
antedates the Keynesian revolution, and was quite appropriate in economies
whose monetary rule was to maintain convertibility into gold or some other
currency at a fixed price. However the day-to-day operating procedures of
central banks, the very organisation of their decision-making processes, not
to mention the structure of the private markets in which they operate are all
geared by force of tradition to making and implementing decisions about
interest rates. Although monetarists have done a great deal of work on the
basic economics of the money supply process under different policy regimes,
and though some of them, notably Brunner and Meltzer, have frequently
scolded their colleagues for neglect of these issues, hindsight suggests that
they did not recognise the extent to which the problem of implementing a
different monetary policy might require a basic overhaul of institutions if it
was to be solved, an overhaul that might involve a considerable break with
traditional practices, and hence be hard to implement, or that, if they did,
they were unable to convince policy makers to undertake that overhaul at
the same time as they adopted monetarist rhetoric.
If central banks, apart from the Bank of Canada, have not in fact
succeeded in smoothly slowing down monetary expansion rates in a
sustained way, a number of them have nevertheless managed to create
contractions in monetary growth rates that have been sharp and persistent
enough to bite. Associated with these contractions have been the ‘shifts’ of
the demand for money function that I discussed earlier in this chapter. As the
reader will recall, I argued that these shifts were, in all probability, real
phenomena, and not statistical artefacts, that such shifts were nothing new,
and that they were probably to be explained, at least in part, by institutional
changes which themselves might plausibly be interpreted as a response to
monetary policy. I believe that these shifts of the demand for money function,
relatively small though they have been, force us to reassess a fundamental
tenet of practical monetarism, namely the injunction to fix ex ante a growth
rate rule for the money supply, and then ensure adherence to it by taking
away from the monetary authorities the discretion to do otherwise.
Objections to such a proposal have frequently been cast in terms of the
question ‘How are you going to define the money supply for purposes of
implementing this policy?’ The answer typically given has been that it does
not much matter, because if the rate of growth of one monetary aggregate is
pinned down, all the others will end up behaving consistently, at least on
average over the kind of time periods for which stability in monetary policy
is really important. That answer is surely valid if one is dealing with an
economy in which there is no institutional change in the private sector, but
that does not make it as adequate a response to the question as I once
thought it did. Suppose we agreed to set a rule for the growth rate of Ml and
that initially we could agree on what assets to include in that aggregate.
What if after the rule had been implemented some new asset, for example a
238 David Laidler
new kind of chequing account, evolved? Perhaps the demand function for Ml
as initially defined would then shift, but if ex post we included the new asset
in our definition of M1 we might still be able to show that the demand for
narrow money had not ‘really’ shifted, after all.
Such problems would not arise if we were not too specific in laying down
the precise definition of money that was to bind policy makers in the future.
However, to do that would leave it open to the discretion of someone at
some time in the future to decide just how to define the monetary aggregate
whose rate of growth was tied down with a rule, and that amounts to giving
them the discretion to ignore the rule in question. It is hard to resist the
implication that it does not seem to be possible, let alone desirable, to
eliminate all scope for discretionary policy in a world in which the monetary
system is in a state of evolution. I hasten to add that this does not imply that
attempts to implement short-run fine tuning of the economy by way of
manipulating interest rates are all of a sudden alright, or that it is fruitless to
require central banks to announce target ranges for monetary expansion
over, say, one or two year time horizons. However it does imply that it is as
a practical matter impossible to prevent policy makers doing the wrong
things if they so wish by tying them down to a monetary growth-rate rule.
Unless we can accurately foresee the path that innovations in the financial
sector are going to take, someone somewhere is going to have to be granted
the discretion to deal with them when they arise. The monetarist injunction
not to use monetary policy for fine tuning is not affected by these
considerations, but the proposal that the once and for all enactment of a
simple rule can lead to that injunction being implemented is undermined.
That seems to me to be a rather severe criticism of monetarist policy
doctrine.
V CONCLUDING COMMENTS
As the reader will by now have seen, it is my view that the core of monetarism
has consisted of a series of empirical propositions and policy prescriptions, all
of which are quite consistent with mainstream economic theory. One can
approach the analysis of social questions in terms of the maximising behaviour
of individual agents without believing in a stable demand for money function,
or a vertical long-run Phillips curve, but evidence that such relationships exist
need in no way disturb one’s theoretical preconceptions. Although there have
been episodes in the monetarist debate where the relevance of mainstream
economics to the analysis of such social questions as inflation and
unemployment has been vigorously questioned, particularly in Britain, it has
mainly been about questions amenable to being settled with reference to
empirical evidence, as Mayer (1978) has also argued.
Viewed in this light, I would suggest that, in all but one aspect, the
monetarist debate is as close to being over as an economic controversy
ever is. The demand for money function does seem to be more stable over
Monetarism: An interpretation and assessment 239
time than the early critics of monetarism suggested, while shifts in it have
been neither new phenomena, nor of sufficient magnitude seriously to
undermine long-run relationships between money and money income.
Puzzles about ‘reverse causation’ in the data for countries such as Britain
cease to be puzzles when the openness of the economy and the nature of
the exchange-rate regime are taken account of. There is now much less
disagreement about the empirical nature of the interaction of real income
and inflation: there is a short-run trade-off between inflation and
unemployment and it does seem to vanish in the long run. Though we
should not underrate the importance of the consensus that has been
achieved on the foregoing issues—or neglect to mention explicitly that the
consensus in question is not universal—this does not mean that there is
now no controversy in macroeconomics. As we have seen two areas
remain contentious.
First, one aspect of the monetarist debate remains alive, and that
concerns the proper conduct of monetary policy. I doubt that my own view,
that the case for governing monetary policy by rules is impossible to
sustain in the face of careful consideration of the influence of institutional
change on the behaviour over time of the demand for money function, will
find a great deal of support among monetarists at present, while I would be
surprised to find it regarded as sufficient of a concession to ‘fine tuning’,
and it really is no such thing, to satisfy the Keynesians. Thus, I would
expect debates about this matter to keep the monetarist controversy alive
for a while yet.
The other, and in my view far more important, issue has to do with the
market-theoretic foundations of macroeconomics. The issues raised by Lucas
and his collaborators are not the issues that have traditionally concerned
participants in the monetarist debate and it is misleading to approach them
as if they were. The debate about the assumptions of clearing markets and
rational expectations as a basis for macroeconomics is a new one, and as
Brian Kantor (1979) has suggested is really about whether Keynes’ General
Theory carried economics forward or took it on a fruitless detour. Though it
has very little to do with monetarism, it nevertheless concerns issues of
fundamental theoretical importance for macroeconomics. Let us hope that
this new controversy proves to be as fruitful as the monetarist controversy
has been.
ACKNOWLEDGEMENTS
I have benefited greatly from the extensive comments of John Foster,
Milton Friedman, John Helliwell, Geoffrey Kingston, Clark Leith, Thomas
Mayer, Ronald Shearer and George Zis, none of whom is to be held
responsible for the views that I espouse. The financial support of the Social
Science and Humanities Research Council of Canada is gratefully
acknowledged.
240 David Laidler
NOTES
1 See, for example, James Boughton (1977), Karl Brunner (1970), Nicholas Kaldor
(1970), Harry Johnson (1972), Thomas Mayer (1978), Franco Modigliani (1977),
Douglas Purvis (1980). This list is far from exhaustive.
2 This is the judgement of Johnson (1962) and Brunner (1970), among others.
3 Consider, for example, the London Business School model of the UK economy (see
Jim Ball and Terry Burns 1976). The Canadian RDX2 model also seems to me to
fall into this category.
4 See, for example, Friedman and Schwartz (1963a) and Philip Cagan (1979). Note
that such monetarists as Brunner and Meltzer, however, do not use National Bureau
techniques. They are mainly associated with the Chicago branch of monetarism.
5 The Radcliffe Report (1959) is based on the proposition that the demand for
money function is essentially nonexistent as a stable relationship. For a later
statement of the same point of view see Kaldor (1970) or Joan Robinson (1970).
6 Note in particular that the Keynesian James Tobin was the author of a pioneering
econometric study of the demand for money function (see Tobin 1947). See also
his review of Friedman and Schwartz (1963a), where further econometric estimates
of the demand for money function are presented (Tobin 1965).
7 I have dealt with the matters taken up here in much greater length in Laidler
(1980).
8 But as with all such blanket judgements as this there are important exceptions.
Irving Fisher’s empirical work on the relationship between money and prices
presented in The Purchasing Power of Money (1911) is not so far removed from
modern monetarism.
9 But this of course is not to say that Friedman has always paid as much attention to
the interest elasticity of the demand for money as his critics might have wished. See
for example the various reviews of the monetary history of the United States, but
note that the monetarist Allan Meltzer (1965) was as critical on this score as any
other reviewer. What we are here dealing with is a characteristic of some of the
work of one, albeit the most important, monetarist rather than of monetarism in
general.
10 Mayer (1978) argues, correctly I believe, that Patinkin should not be regarded as
a monetarist. This of course is not to deny the important influence that Patinkin’s
work had on subsequent monetary analysis. See, for example, Jonson (1976b).
11 There seems to have been a systematic shift in British opinion from the Radcliffe
view that money does not matter at all, to the view that money matters for real
income but not for prices. To trace this development is beyond the scope of this
paper. However, the work of Richard Kahn shows clearly that it has taken place.
Compare his evidence to the Radcliffe Committee with, for example, Kahn (1976).
12 See Peter Wiles (1973) for a particularly extreme version of the sociological approach
to inflation.
13 I have in mind here, in particular, the work of Robert Solow (1968) and James
Tobin (1972).
14 See Anthony Santomero and John Seater (1978) for a well-balanced survey of the
evidence on these matters.
15 Notice that in some of his subsequent writings on inflation-unemployment
interaction Friedman adopts an aggregate supply curve interpretation of the Phillips
curve (see, for example, Friedman 1975).
Monetarism: An interpretation and assessment 241
16 See the papers by Armen Alchian, Robert E.Lucas and Leonard Rapping, Donald
Gordon and Alan Hynes, and Dale Mortensen, all in the Phelps (1969) volume.
17 In later work carried out by Brunner and Meltzer and their associates, a version
appears of the aggregate supply curve in which the rate of change of output rather
than the level of output affects the rate of inflation. This form of the relationship
appears to stem from their tendency to treat the expected inflation rate as
synonymous with the rate of change of the expected price level. See Brunner and
Meltzer (1978) and particularly the comments there by Bennett McCallum.
18 Thus, though I agree with much of what Willem Buiter (1980) has to say about
this theory of employment, I cannot accept his characterisation of it as ‘The
Macroeconomics of Dr. Pangloss’. It might be noted that in the aggregate supply
curve interpretation of the Phillips curve, the natural unemployment rate becomes
a long-run equilibrium concept. In the price reaction function interpretation of the
relationship it seems to me to be synonymous with the Keynesian concept of the
minimum feasible unemployment rate. For a perceptive discussion of some of the
issues involved here see Thomas Wilson (1976).
19 I base the following arguments on the papers of Lucas (1975), Thomas Sargent
(1976) and Lucas (1977). The first two of these papers are extremely technical and
I am by no means sure that I am doing justice to them in the discussion that
follows. Milton Friedman has pointed out to me that one can say that errors are
random or systematic only if one is also specific about both the time at which
expectations are formed, and the period for which they are formed. If one is now
planning for, say a five-year horizon, then the rational expectations hypothesis
permits the actual value of any variable to deviate systematically from its ex ante
expected value over any interval of less than five years. This matter is clearly related
to questions raised by adjustment lags, the durability of certain goods, and so on,
since the horizon over which a decision taken now is likely to be binding is also
presumably the horizon over which a rational agent would seek to form expectations
about relevant variables. To the best of my knowledge, the published literature has
not recognised this point explicitly, and it deserves much more attention than I
have space to give it here.
20 To comment on the empirical work in question, notably that of Robert J.Barro
(1978), would take us beyond the scope of this chapter.
21 Robert J.Barro (1979) presents a particularly forceful and clearcut statement of
what I am calling the ‘neo-Austrian’ view on these matters. Robert Solow (1979)
might be regarded as providing a traditional Keynesian rebuttal of this line of
argument. Note that questions of the relevant time horizon, raised in note 19, are
again relevant here.
22 Alternatively, as Michael Wickens has suggested to me, we may think of ∆pe as
being a rational expectation of inflation conditional upon information available at
an earlier time than that at which the quantity signal is received.
23 I base the foregoing discussion on conversations and correspondence that I have
had with Marcus Miller and Peter Jonson on various occasions. See also Clements
and Jonson (1979).
24 A more extensive account of these matters is given in Laidler (1975: ch. I). Note
that Brunner, Alex Cuckeirman and Meltzer (1979) provide an analysis of persistent
shocks within an aggregate supply curve framework. Note also that Peter Howitt
(1979) argues that, once explicit attention is paid to the role of inventories in the
price setting process, the contrast between clearing-market and non-clearing-market
242 David Laidler
approaches to economic modelling becomes blurred, and to some extent semantic
rather than substantive in nature.
25 The locus classicus for pioneering work on the monetary approach to balance of
payments analysis is, of course, Frenkel and Johnson (1976).
26 See, for example, Laidler (1975: ch. 9) and Jonson (1976 a).
27 It is worth pointing out that I set out much of the foregoing argument in my 1972
Lister Lecture. See Laidler (1975: ch. 10), where the lecture is reprinted. The
argument is developed in further detail in Laidler (1976 a).
28 Frenkel (1980) provides a useful and accessible overview of the issues involved
here and the evidence on them.
29 John Helliwell has suggested to me that the application of policy optimisation
techniques to such models is better regarded as a test of their validity than as a
preliminary to actual policy making.
30 It is worth noting that the Radcliffe Committee (1959) regarded the task of
monetary policy to be the achievement of background stability for the economy.
Their view differed from the monetarist approach to the same issue in putting
interest rates at the centre of the policy making process rather than any monetary
aggregate. In the kind of sociological theorising about inflation that was particularly
popular in Britain in the early 1970s incomes policy was to be assigned the task of
stabilising prices and expectations.
31 I would emphasise that this is not a new position on my part. It is one that I have
consistently taken. Of course the questions about the effectiveness of fiscal policy
are important ones for macroeconomists, and the Brown University Conference
on Monetarism (see Jerome Stein 1976) dealt almost exclusively with such issues.
I accept Purvis’ (1980) judgement that the outcome of that conference was to show
beyond a reasonable doubt that ‘fiscal policy matters’ but also his judgement that
in retrospect the debate about the effectiveness of fiscal policy has not been the
most important one in the monetarist debate, however important an issue it might
be in its own right for macroeconomics.
Finally note that the foregoing discussion ignores the question as to whether,
even if we had enough knowledge to ensure that fine tuning could be used beneficially,
the political process would permit it to be used in that way. This question, as
Milton Friedman has pointed out to me, is a vital one in any practical debate about
activist policies.
32 Of course there has been a considerable ideological content to the monetarist
debate and I would not deny that for a moment. Nor would I take the position that
there is anything reprehensible about ideological debates per se. I play these issues
down in this chapter not because, from a broader perspective I would regard them
as unimportant, but because my expertise as an economist does not put me in a
position to say anything very useful about them.
33 Michael Parkin, Michael Sumner and Robert Jones (1972) is still an admirable
source of information about wage and price controls in the British economy.
Michael Walker (1976) contains much useful information on other countries.
Note that the views that I state here about the importance of using wage and price
controls, if they are to be used, in conjunction with monetary and fiscal policy,
rather than instead of such policies, are those of the McCracken Committee. See
McCracken et al (1977).
34 See Laidler (1976b), particularly Chapters 7 and 9, for an earlier statement of my
own views on the role of the housing market and its interaction with monetary
policy and inflation. I readily acknowledge that the source is an obscure one.
Monetarism: An interpretation and assessment 243
REFERENCES
Ando, A. and Modigliani, F. (1965) ‘The relative stability of monetary velocity and the
investment multiplier’, American Economic Review 55 (September), pp. 693–728.
Ball, R.J. and Burns, T. (1976) ‘The inflationary mechanism in the U.K. economy’,
American Economic Review 66 (September), pp. 478–84.
Barro, R.J. (1978) ‘Unanticipated money, output and the price level in the United
States’, Journal of Political Economy 86 (August), pp. 549–81.
——(1979) ‘Second thoughts on Keynesian economies’, American Economic Review
69 (May), papers and proceedings, pp. 54–9.
Baumol, W.J. (1952) ‘The transactions demand for cash: an inventory theoretic approach’,
Quarterly Journal of Economics 66 (November), pp. 545–56.
Boughton, James M. (1977) ‘Does Monetarism Matter’, in Elmus Wicker (ed.) Lilley
Conference on Recent Developments in Economics April 21–23. Bloomington, IN,
Indiana University.
Brown, Arthur J. (1955) The Great Inflation 1939–1951, London: Oxford University
Press.
Brunner, K. (1961) ‘The report of the Commission on money and credit’, Journal of
Political Economy 69 (December), pp. 605–20.
——(1970) ‘The monetarist revolution in monetary theory’, Weltwirtschaftliches Archiv
105, pp. 1–30.
——(1978) ‘Issues of post-Keynesian monetary analysis’, in Thomas Mayer (1978).
Brunner, K., and Meltzer, A.H. (1963) ‘Predicting velocity: implications for theory and
policy’, Journal of Finance 18 (May), pp. 319–54.
——(eds) (1978) The Inflation Problem. Carnegie-Rochester Public Policy Conference
Series, Amsterdam: North-Holland.
Brunner, K., Cukierman, A. and Meltzer, A.H. (1979) ‘Stagflation, persistent
unemployment and the permanence of economic shocks’, Carnegie-Mellon University
(mimeo).
Buiter, W.H. (1980) ‘The macroeconomics of Dr. Pangloss: a critical survey of the New
Classical Macroeconomics’, Economic Journal 90 (March), pp. 34–50.
Cagan, P. (1978) ‘Monetarism in historical perspective’. In Thomas Mayer (1978).
——(1979) Persistent Inflation, New York: Columbia University Press.
Cagan, P. and Schwartz, A.J. (1975) ‘Has the growth of money substitutes hindered
monetary policy?’ Journal of Money, Credit and Banking 7 (May), pp. 137–60.
Clements, K.W. and Jonson, P.D. (1979) ‘Unanticipated money, disequilibrium modelling
and rational expectations’, Economic Letters 2, pp. 303–8.
Clower, R.W. (1965) ‘The Keynesian counterrevolution: a theoretical appraisal’, The
Theory of Interest Rates (ed. F.H.Hahn and F.P.R.Brechling), London: Macmillan.
De Prano, M. and Mayer, T. (1965) ‘Tests of the relative importance of autonomous
expenditure and money’, American Economic Review 55 (September), pp. 729–52.
Fisher, I. (1911) The Purchasing Power of Money, New York.
Frenkel, J. (1980) ‘Flexible Exchange Rates in the 1970s’, in Laurence H.Meyer (ed.)
Stabilisation Policies Lessons from the 70s and Implications for the 80s, St Louis,
Washington University and Federal Reserve Bank of St Louis.
Frenkel, J. and Johnson, H.G. (1976) The Monetary Approach to Balance of Payments
Theory, London: George Allen and Unwin.
Friedman, M. (1956) ‘The quantity theory of money: a restatement’, in Studies in the
Quantity Theory of Money (ed. M.Friedman), Chicago: University of Chicago Press.
——(1959) ‘The demand for money: some theoretical and empirical results’, Journal of
Political Economy 67 (June), pp. 327–51.
——(1960) A Program for Monetary Stability, New York: Fordham University Press.
——(1966) ‘Interest rates and the demand for money’, Journal of Law and Economics
9 (October).
244 David Laidler
——(1968) ‘The role of monetary policy’, American Economic Review 58 (March), pp.
1–17.
Friedman, M. (1970) ‘A theoretical framework for monetary analysis’, Journal of
Political Economy 78 (March-April), pp. 193–238.
——(1975) Unemployment Versus Inflation, London: Institute of Economic Affairs.
Friedman, M. and Meiselman, D. (1963) ‘The relative stability of monetary velocity and
the investment multiplier in the United States, 1898–1958. Commission on Money
and Credit’, Stabilization Policies, Englewood Cliffs, NJ: Prentice-Hall.
Friedman, M. and Schwartz, A.J. (1963a) A Monetary History of the United States,
1867–1960, Princeton, NJ: Princeton University Press for the NBER.
——(1963b) ‘Money and business cycles’, Review of Economics and Statistics 45
(February), supplement, pp. 32–64.
Frisch, H. (1977) ‘Inflation theory 1963–1975: a second generation survey’, Journal
of Economic Literature 15 (December), pp. 1289–317.
Hahn, F.H. (1980) ‘Monetarism and Economic Theory’, Economica NS 47 (February),
pp. 1–18.
Harberger, A.C. (ed.) (1960) The Demand for Durable Goods, Chicago: University of
Chicago Press.
Harrod, R.F. (1971) ‘Discussion paper’, in Monetary Theory and Monetary Policy in
the 1970’s (ed. G.Clayton, J.C.Gilbert and R.Sedgewick), London: Oxford University
Press.
Hicks, J.R. (1974) The Crisis in Keynesian Economics, Oxford: Basil Blackwell.
Howitt, P.W. (1979) ‘Evaluating the non-market-clearing approach’, American Economic
Review 69 (May), Papers and Proceedings, pp. 60–4.
Johnson, H.G. (1962) ‘Monetary theory and policy’, American Economic Review 52
(June), pp. 335–84.
——(1972) Inflation and the Monetarist Controversy, De Vries Lectures, 1971.
Amsterdam: North Holland.
Jonson, P.D. (1976a) ‘Money and economic activity in the open economy, the United
Kingdom 1880–1970’, Journal of Political Economy 84 (October), pp. 979–1012.
——(1976b) ‘Money, prices and output: an integrative essay’, Kredit und Kapital 9,
pp. 499–518.
Kahn, R.F. (1976) ‘Inflation: a Keynesian view’, Scottish Journal of Political Economy
23 (February), pp. 11–15.
Kaldor, N. (1970) ‘The new monetarism’, Lloyd’s Bank Review (July), pp. 1–18.
Kantor, B. (1979) ‘Rational expectations and economic thought’, Journal of Economic
Literature 17 (December), pp. 1422–75.
Laidler, D. (1966) ‘The rate of interest and the demand for money: some empirical
evidence’, Journal of Political Economy 74 (December), pp. 545–55.
——(1971) ‘The influence of money on economic activity: a survey of some current
problems’, in Monetary Theory and Monetary Policy in the 1970’s (ed. G.Clayton,
J.C.Gilbert and R.Sedgwick), London: Oxford University Press.
——(1975) Essays on Money and Inflation, Manchester: University of Manchester
Press; Chicago: University of Chicago Press.
——(1976a) ‘Inflation in Britain: a monetarist analysis’, American Economic Review
76 (September), pp. 485–500.
——(et al.) (1976b) Study on the Possible Part Played by Certain Primary Non-
employment Incomes in the Inflationary Process in the United Kingdom, Commission
of the European Communities, Brussels.
——(1980) ‘The demand for money in the United States: yet again’ (ed. K.Brunner and
A.H.Meltzer), Carnegie-Rochester Conference Series, The State of Macroeconomics,
Amsterdam: North Holland.
Lieberman, C. (1980) ‘The long-run and short-run demand for money revisited’, Journal
of Money, Credit and Banking.
Monetarism: An interpretation and assessment 245
Lipsey, R.G. (1960) ‘The relationship between unemployment and the rate of change of
money wage rates in the United Kingdom, 1862–1957’, Economica 27, pp. 1–31.
Lucas, R.E., Jr (1972) ‘Expectations and the neutrality of money’, Journal of Economic
Theory 4, pp. 103–24.
——(1975) ‘An equilibrium model of the business cycle’, Journal of Political Economy
83 (November-December), pp. 1113–44.
——(1976) ‘Econometric policy evaluation: a critique’, The Phillips Curve and Labour
Markets (ed. K.Brunner and A.H.Meltzer), Carnegie-Rochester Conference Series,
Amsterdam: North Holland.
——(1977) ‘Understanding business cycles’, in Stabilization of the Domestic and
International Economy (ed. K.Brunner and A.H.Meltzer), Carnegie-Rochester
Conference Series, Amsterdam: North Holland.
McCracken, P. et al. (1977) Towards Full Employment and Price Stability (McCracken
Report), Paris: OECD.
Mayer, T. (1978) The Structure of Monetarism, New York: W.W.Norton.
Meltzer, A.H. (1965) ‘Monetary theory and monetary history’, Schweizerische Zeitschrift
Volkswirtschaft und Statistik (Spring), pp. 409–22.
——(1969) ‘Money intermediation and growth’, Journal of Economic Literature 7
(March), pp. 27–56.
Modigliani, F. (1977) ‘The monetarist controversy, or should we forsake stabilization
policies?’, American Economic Review 67 (March), pp. 1–19.
Parkin, J.M., Sumner, M.T. and Jones, R.A. (1972) ‘A survey of the econometric evidence
on the effects of incomes policy on the rate of inflation’, in Incomes Policy and
Inflation (ed. J.M.Parkin and M.T.Sumner), Manchester: University of Manchester
Press.
Patinkin, D. (1965) Money, Interest and Prices, 2nd edn, originally published 1956,
New York: Harper Rowe.
——(1969) ‘The Chicago tradition the quantity theory and Friedman’, Journal of
Money, Credit and Banking 1 (February), pp. 46–70.
Phelps, E. (1967) ‘Phillips curves expectations of inflation and optimal unemployment
over time’, Economica NS 34 (August), pp. 254–81.
——(et al.) (1969) Microeconomic Foundations of Employment and Inflation Theory,
New York: W.W.Norton.
Phillips, A.W. (1958) ‘The relation between unemployment and the rate of change of
money wage rates in the United Kingdom, 1861–1957’, Economica NS 25
(November), pp. 283–99.
Purvis, D.D. (1980) ‘Monetarism—a review’, Canadian Journal of Economics 1
(February), pp. 96–121.
Radcliffe Committee (Committee on the Working of the Monetary System) (1959)
Report, London: HMSO.
Reid, M. (1962) Housing and Income, Chicago: University of Chicago Press.
Robinson, J. (1970) ‘Quantity theories old and new’, Journal of Money, Credit and
Banking 2 (November), pp. 504–12.
Santomero, A.M. and Seater, J.J. (1978) ‘The inflation-unemployment trade-off: a
critique of the literature’, Journal of Economic Literature 16 (June), pp. 499–544.
Sargent, T.J. (1976) ‘A classical macroeconomic model for the United States’, Journal
of Political Economy 84 (April), pp. 207–38.
Sargent, T.J. and Wallace, N. (1975) ‘Rational expectations, the optimal monetary
instrument and the optimal money supply rule’, Journal of Political Economy 83
(April), pp. 241–54.
Solow, R.M. (1968) ‘Recent controversies on the theory of inflation: an eclectic view’, in
Symposium on Inflation—Its Causes, Consequences and Control (ed. S.Rousseas),
New York: Calvin K.Kazanjian Economics Foundation.
246 David Laidler
Solow, R.M. (1979) ‘Alternative approaches to macroeconomic theory: a partial view’,
Canadian Journal of Economics 12 (August), pp. 339–54.
——(1980) ‘On theories of unemployment’, American Economic Review 70 (March),
PP. 1–11.
Stein, J. (1976) Monetarism, Amsterdam: North-Holland.
Sumner, M.T. (1980) ‘The operation of monetary targets’, in K.Brunner and A.H.
Meltzer (eds) Monetary Institutions and the Policy Process, Carnegie-Rochester
Public Policy Conference Series, Amsterdam: North-Holland.
Tobin, J. (1947) ‘Liquidity preference and monetary policy’, Review of Economics and
Statistics 29 (May), pp. 124–31.
——(1956) ‘The interest elasticity of transactions demand for cash’, Review of
Economics and Statistics 38 (August), pp. 241–7.
——(1958) ‘Liquidity preference as behaviour towards risk’, Review of Economic
Studies 25 (February), pp. 65–86.
——(1965) ‘The monetary interpretation of history’, American Economic Review 55
(June), pp. 464–85.
——(1972) ‘Inflation and unemployment’, American Economic Review 62 (March),
pp. 1–18.
Walker, M. (ed.) (1976) The Illusion of Wage and Price Control, Vancouver, BC: Fraser
Institute.
Wiles, P. (1973) ‘Cost inflation and the state of economic theory’, Economic Journal
83, pp. 377–98.
Wilson, T. (1976) ‘The natural rate of unemployment’, Scottish Journal of Political
Economy 23 (February), pp. 99–107.
10 The monetarist controversy revisited
Franco Modigliani
Contemporary Policy Issues (1988) 6, October, pp. 3–18
I am delighted with the success of this conference honoring Michael
Hamburger, and I am impressed by both the attendance and the outpouring
of love, respect, and admiration for his work. I wish I could claim some
credit for this success since, after all, he was a student of mine.
Unfortunately, I cannot do so in good conscience. One reason for his coming
to the Carnegie Institute of Technology for his graduate work was to study
with me. But because I left soon after he arrived, he was my student for only
a short time–a year or so. Since leaving Carnegie, I often have felt a bit
guilty for abandoning Mike just a year after he arrived.
My topic for today actually was inspired by the last letter that Michael
wrote to me—a letter of congratulations on the Nobel award, dated Oct. 15,
1985. It was a very warm letter in which he first gave me undeserved credit
by asserting that ‘you probably had more influence on my career than
anyone else.’ He proceeded to say that ‘to keep you up-to-date on my
activities, I am enclosing two of my latest reports. I believe that you will
approve of the evolution of my thinking.’
I must confess that because we were buried under a deluge of wonderful
moving messages, I was unable to read the enclosures at that time. It was only
when I heard the heartbreaking news of his death that I suddenly remembered
I had not quite finished reading his message. I finally managed to do so.
CHARACTERISTICS OF MONETARISM
The enclosures were two issues of his monthly publication, Economic
Insights. I found their content—especially that of August 1985, entitled
‘Money Growth Is Not Excessive’—quite extraordinary and gratifying. That
report showed, without a shadow of a doubt, that Michael had abandoned
‘Monetarism’ with a capital ‘M’, as defined in my American Economic
Association presidential address (1977). He had rejoined the ranks of
monetarists with a small ‘m’—those like me and countless others concerned
with the role of money in our economy.
As I suggested in that paper, the difference between a true Monetarist and
the rest of us has little to do with views regarding the workings of the
248 Franco Modigliani
economy. By now, no significant disagreements—at least between Keynesians
and Monetarists—exist over the proposition that money counts. To be sure,
Monetarists sometimes attribute to Keynesians the view that Keynesian
monetary theory rests on the liquidity trap. That mechanism may be of
notable logical significance, but hardly anyone would claim that it is of
practical significance under present circumstances. Nor can there be
differences of principle regarding fiscal policy. Monetarists, including Milton
Friedman, presumably now accept the view that the demand for money
depends on interest rates. Therefore, they must also agree with others that
fiscal policy—at least government expenditure—has some real effects. Nor is
there any difference in accepting the view that money is neutral in the long
run. The vertical Phillips curve is generally accepted now, and I would say
that it was widely adopted within just a few years of the time the hypothesis
was formulated.
Instead, to determine what really distinguishes the M’s from the rest, we
must search along very different lines. First, how stable is the economy in the
absence of stabilization policies? For the Monetarists, the economy is stable
enough so that stabilization policies are not needed. For the non-Monetarists,
the economy suffers from enough instability so as to benefit, in principle,
from active stabilization. Second, how effective are stabilization policies?
According to non-Monetarists, stabilization policies can contribute to
stability not only in principle but also in practice. For Monetarists,
discretionary stabilization destabilizes the economy as a result of long and
variable lags, and because those charged with carrying out the policy are
ineffective. And finally, true Monetarists will argue that even if stabilization
were both needed and feasible, we still should not trust the government with
the necessary power since the bureaucracy has the propensity to act in its
own self-interest. Non-Monetarists like myself recognize that public servants
are not infallible saints and that, under some circumstances, they may make
mistakes and/or be influenced by self-interest. By no means, however, do
these dangers justify depriving ourselves of all discretion, at least where it
matters—to wit, in well-administered countries where government actions
are open to public scrutiny.
Given these fundamental beliefs, Monetarists conclude that one must
forsake stabilization policies. This implies, in particular, forsaking the use of
monetary policy as a stabilization device and instead requiring money to
grow mechanically at a rate fixed once and for all. I have come to
appreciate how Monetarists view the holiness of this principle by watching
Friedman advising on the appropriate monetary policy in diverse complex
situations and each time coming up, unfailingly, with the same practical
answer: 3 percent.
Accordingly, I conclude that the essential characteristics distinguishing
Monetarists from others is their insistence on a steady growth of money—
fixed for a long time, if not forever, and in any event not responsive to
developing circumstances. The only exception that they have made to their
The monetarist controversy revisited 249
principles of constant growth has occurred when the economy, for whatever
reason, has reached a high level of inflation. Then they have recommended a
money growth path beginning with a rate close to that of inflation and
declining gradually until the growth coincides with the real growth trend of
the economy and, they hope, inflation has been squeezed out of the economy.
But even in this case, the path is fixed in advance.
Michael’s article in Economic Insights (August 1985) to which he called
my attention showed that he had both the honesty and the flexibility to give
up the Monetarist position he had long held by expressing his approval of
the course that the Fed pursued in 1985. This course involved abandoning
the growth target, set at the beginning of the year as part of a program of
gradual re-entry from inflation, and accepting a much higher growth of
money. The purpose of the higher growth was to track the initial GNP target
in the face of a rise in money demand or, equivalently, of a large decline in
velocity, amounting to 5 percent, for the first half of the year. Not only was
he abandoning Monetarism, he was quite open about it: ‘I believe you will
approve of the evolution of my thinking.’ By mid-1986, he not only was
approving of the unprecedented high growth rate of Ml to offset the
continuing velocity decline, but even was complaining because the Fed failed
to take into account the prospective velocity decline by announcing an
insufficiently high target. This caused a large discrepancy between
realization and target which, in his view, disturbed the financial markets (see
Economic Insights June 1986). In short, Michael had concluded that it was
wrong to enforce, at all costs, a constant or transiently declining growth of
money. Instead, the time had come to target income and to adjust money
growth and money targets as necessary so as to achieve the income target in
light of changing velocity.
One must recognize that Hamburger was not the first to defect from
Monetarism. Already in 1983, the Council of Economic Advisers gave up its
earlier prescription of ‘a gradual reduction in the rate of growth of the
money stock until the rate is consistent with price stability’ (Economic
Report of the President, 1983, p. 23). Influenced by the 1982 experience, and
especially by the sharp fall in velocity, the Council acknowledged the need
‘to balance the principle of stable money growth with the need to take
account of changing asset preferences that may alter the velocity of
money…One possible way [to do so] is to use the observed behavior of
nominal GNP to guide a gradual recalibration of the monetary growth
targets…Basing the recalibration of monetary targets on nominal GNP is
consistent with the basic principle of pursuing a stable monetary policy’
(Economic Report of the President, 1983, pp. 23–4).
But probably the first to abandon the Monetarist bandwagon was the
Federal Reserve itself. It did so as early as 1982 by responding to the large
decline in velocity with an unusually high infusion of money. Actually, some
question exists as to whether the Fed abandoned Monetarism in 1982 or
instead had never really embraced it. We know that around 1979, in the face
250 Franco Modigliani
of continuing and rising inflation, the Fed had adopted a policy of renewed
emphasis on monetary growth targets announced in advance—in place of
interest rate targets. However, favoring money growth targets is a necessary
but by no means sufficient condition to being admitted to the ranks of true
Monetarists.
MONETARISM AND THE FED
The main reason to question the seriousness of the Fed’s commitment to
Monetarism is that despite its announcement of targets during these early
years, the growth of money supply appeared extremely jagged—periods of
very rapid growth followed by periods of stagnation or even decline—and
hence was a far cry from a constant growth rate. However, economists
generally agree that so long as the variability of money is merely short run,
it should not be of much consequence. On the other hand, the behavior over
longer periods—such as four-quarter changes—does not appear inconsistent
with Monetarists’ prescriptions.
One may begin by looking at the target’s behavior for the four-quarter
change. Table 10.1 reports target and actual four-quarter changes for all
monetary aggregates. The table also reports the Fed’s so-called ‘economic
projections’ and actual four-quarter changes for its monetary policy
objectives—namely GNP, real GNP, the GNP deflator, and unemployment.
One can see from the first row, which relates to M1, that up until 1982
the M1 target declines year after year. This is true whether we rely on the
midpoint or on the upper limit of the target range. (The target rises from
1979 to 1980, but one may attribute this to the fact that the 1979 target
related to a somewhat different aggregate, M1A.) The decline is not large,
but the path traced certainly is broadly consistent with the Monetarists’
prescription. Similarly, the actual growth of M1 tended to decline through
1981 up until the third quarter of 1982, even if the behavior of the actual
growth was less steady than that of the target.
But several other aspects of the Fed’s behavior appear quite inconsistent
with Monetarism. One conspicuous example is the rather broad target range
that the Fed announced for M1 and the other aggregates: for the period
1979–82, typically 3 percentage points and not less than 2 1/2. Monetarists
holding that 3 percent growth is the only acceptable target must regard this
much room as a devious way to reintroduce discretion.
Another highly suspicious practice is that throughout this period, the Fed
announced targets for a substantial array of aggregates—generally four,
though not always the same four. These targets generally could be expected
to prove inconsistent with each other ex post. Thus, the only effect of this
practice was to give the Fed extra leeway regarding any one target, in
particular M1, by giving the Fed the option regarding the target to which it
would adhere.
The monetarist controversy revisited 251
One more highly revealing aspect of the Fed’s behavior is that throughout
the years 1979–82, the actual growth of M1 exceeded the targeted growth
except in 1981. Nor could one attribute this excess to the pursuit of a
different target or targets since during this period, the actual growth of M2
and M3 also was on the high side of the corresponding targets without
exception. To be sure, the excess for M1 and the other targets was not very
large if compared with the upper limit of the target range. But when one
considers the very large range announced and the extra potential leeway
resulting from the multiple targets, one must conclude that—contrary to
Monetarist tenets—the Fed felt no strong obligation to stick to the targets
whenever doing so seemed to interfere with the pursuit of other non-
Monetarist goals.
This evidence suggests rather convincingly that the Federal Open Market
Committee (FOMC) never really embraced Monetarism. The reason that the
FOMC shifted to an emphasis on aggregates was not a Monetarist
conversion but rather a matter of convenience. First, it had difficulty
selecting appropriate interest rate targets in the presence of significant and
variable inflation. With inflation, the usual difficulty of determining
appropriate interest rates is compounded by that of assessing price
expectations and the pervasive effects of taxation. Second, the Fed was very
conscious of the need for a tough anti-inflationary policy at the cost of very
high interest rates, but it was concerned about the antagonism that such a
policy surely would engender and that, in the end, might even interfere with
the Fed’s ability to carry out the task. In addition, the FOMC might also tend
to procrastinate when circumstances call for enforcing unusually high
interest rates. Adoption of Monetarist-like growth targets could reduce this
potential hostility and procrastination, since doing so would make
unnecessary the explicit targeting of unpopular sky-high interest rates and
since part of the blame could be shifted to the Monetarists, who were then
riding high.
The variable that the Fed seemed especially concerned with and
endeavored to track was not the growth of M1 (or any other monetary
aggregate) but that of nominal GNP—specifically, its real implications. The
main evidence supporting this proposition is that throughout the years 1979
to 1981, GNP was uniformly within the projected range and was generally
close to the middle while the monetary aggregates were missed and
systematically exceeded. Note also that the GNP path—both projected and
achieved—was characterized by a slow but steady decline in the growth rate.
This suggests a courageous choice of final targets and a very skilled
execution. The Fed had no illusion that a sudden burst of ‘rational
expectations’ would make it possible to slow down money and nominal GNP
growth without adverse consequences for output and employment. Quite the
contrary, the chosen path of nominal GNP was seen as leading to painful
consequences. Thus, in 1980, the targeted real growth range fell almost
entirely on the negative side and the actual growth was negative. In
T
ab
le
1
0
.1
F
ed
er
al
R
es
er
ve
‘
ta
rg
et
s’
a
n
d
a
ct
u
al
c
h
an
ge
s
in
m
o
n
et
ar
y
ag
gr
eg
at
es
a
n
d
‘
p
ro
je
ct
io
n
s’
a
n
d
ac
tu
al
c
h
an
ge
s
in
o
th
er
v
ar
ia
b
le
s
N
o
te
s:
a
F
o
r
1
9
7
9
,
th
e
va
lu
es
a
re
f
o
r
M
lA
.
F
o
r
1
9
8
1
,
th
e
va
lu
es
a
re
a
d
ju
st
ed
f
o
r
th
e
N
O
W
a
cc
o
u
n
t
sh
if
t.
b
D
ef
in
it
io
n
c
h
an
ge
s
af
te
r
1
9
8
2
(
an
d
a
ls
o
s
ea
so
n
al
a
n
d
b
en
ch
m
ar
k
a
d
ju
st
m
en
ts
).
S
ee
F
ed
er
al
R
es
er
ve
B
u
ll
et
in
,
F
eb
ru
ar
y
1
9
8
4
,
p
.
8
4
.
c
F
o
r
th
e
ye
ar
s
1
9
7
9
t
o
1
9
8
2
,
th
e
se
ri
es
s
h
o
w
n
r
ef
er
t
o
b
an
k
c
re
d
it
.
N
o
te
:
D
at
a
ar
e
ra
te
s
o
f
ch
an
ge
,
fo
u
rt
h
q
u
ar
te
r
to
f
o
u
rt
h
q
u
ar
te
r.
F
o
r
u
n
em
p
lo
ym
en
t,
d
at
a
ar
e
th
e
ra
te
s
in
e
ac
h
f
o
u
rt
h
q
u
ar
te
r.
So
u
rc
es
:
F
ed
er
al
R
es
er
ve
B
u
ll
et
in
,
va
ri
o
u
s
is
su
es
.
T
ar
ge
ts
a
re
f
ro
m
‘
M
o
n
et
ar
y
P
o
li
cy
R
ep
o
rt
t
o
t
h
e
C
o
n
gr
es
s’
(
M
P
R
C
),
p
u
b
li
sh
ed
i
n
F
eb
ru
ar
y
o
f
th
e
gi
ve
n
y
ea
r
o
r
af
te
rw
ar
d
.
R
ev
is
io
n
s
ar
e
fr
o
m
M
P
R
C
,
p
u
b
li
sh
ed
i
n
J
u
ly
o
r
af
te
rw
ar
d
.
A
ct
u
al
v
al
u
es
o
f
m
o
n
et
ar
y
an
d
c
re
d
it
a
gg
re
ga
te
s
ar
e
fr
o
m
M
o
n
et
ar
y
P
o
li
cy
O
b
je
ct
iv
es
f
o
r
1
9
8
7
,
Su
m
m
ar
y
R
ep
o
rt
o
f
th
e
F
ed
er
al
R
es
er
ve
B
o
ar
d
,
1
0
,
F
eb
.
1
9
8
7
,
p
.
1
1
.
A
ct
u
al
v
al
u
es
o
f
o
th
er
v
ar
ia
b
le
s
fo
r
va
ri
o
u
s
ye
ar
s
ar
e
fr
o
m
t
h
e
fo
ll
o
w
in
g
so
u
rc
es
:
1
9
7
9
–1
9
8
2
,
M
P
R
C
,
p
u
b
li
sh
ed
i
n
F
eb
ru
ar
y
fo
ll
o
w
in
g
th
e
gi
ve
n
y
ea
r;
1
9
7
9
–1
9
8
5
,
St
at
is
ti
ca
l
A
n
n
ex
o
f
th
e
F
ed
er
al
R
es
er
ve
B
u
ll
et
in
,
p
u
b
li
sh
ed
i
n
A
p
ri
l
fo
ll
o
w
in
g
th
e
gi
ve
n
y
ea
r;
1
9
8
6
,
E
co
n
o
m
ic
I
n
d
ic
at
o
rs
,
F
eb
ru
ar
y
1
9
8
7
,
p
.
3
.
T
ab
le
1
0
.1
(
co
n
ti
n
u
ed
)
254 Franco Modigliani
1981. the projection centered on zero and the outcome also was close to
zero. The observed behavior of target setting and realizations during the
period prior to 1983 indicates that the Fed was concerned primarily with
tracking final targets, essentially GNP. Even though monetary targets were
announced regularly, presumably they represented not a true commitment
but rather the best guess as to how much monetary growth would be required
so as to achieve the final target, perhaps shaved down to placate the
Monetarists. If the announced target ever proved to be inconsistent with
achieving the final target, then the monetary targets would be set aside—as
happened repeatedly between 1979 and 1982. But up until 1981, the
inconsistencies were reasonably small.
As a result, during these early years, the actual behavior of M1 had a
Monetarist quality. One must explain this by the circumstance that up until
1982. no serious perceived discrepancy existed between Monetarist
prescriptions and a desirable course for GNP. However, beginning with the
sharp decline in velocity in 1982, the inconsistencies became glaring. The
Fed finally was forced to choose between either staying close to the money
growth target or tracking GNP and giving up the pretense of Monetarism.
Unfortunately, this realization had occurred too late to prevent a serious
contraction which, following years of no growth or even declines, played
havoc with the world economy.
Since 1982, the relationship between the choice of targets and the
realizations of monetary and final targets is similar to that prevailing during
the earlier period, but it is even less coherent with Monetarist behavior. In
1983. the monetary targets mostly were increased but nonetheless the
realizations exceeded the targets—some by large margins—while GNP
growth again fell within the target ranges. In 1984, the monetary
aggregates—including M1—were mostly on track, but then so was GNP.
Hence, one cannot tell whether the Fed truly was concerned with tracking
GNP or with tracking M1. During 1985 and 1986, on the other hand, the
Table 10.1 data leave little doubt that the Fed again endeavored to track
GNP and, to this end, was prepared to let M1 grow at a rate much higher
than targets as well as Monetarist standards. Unfortunately, the M1 growth
proved neither large enough nor timely enough to achieve the GNP targets,
particularly the real ones.
On the whole, the evidence suggests that during this second period also,
the Fed was continuing to behave in a non-Monetarist fashion by focusing on
tracking the GNP targets while disregarding the monetary growth targets.
What can one learn from this rich experience of the years since 1979
about the controversy surrounding stabilization policies? I find this
experience quite instructive in providing strong support for the proposition
stated in my earlier paper (Modigliani 1977): ‘[A] private enterprise
economy using an intangible money needs to be stabilized, can be stabilized
and therefore should be stabilized.’
The monetarist controversy revisited 255
CASE FOR STABILIZATION POLICY
The proposition that the economy must be stabilized is amply supported by
the dramatic variability of M1 velocity, especially since 1981. The
implications of such variability are as follows. Suppose that the Fed had
followed a strict Monetarist policy of increasing M1 at a constant rate. For
this particular period, suppose that the Fed chose a rate around 5 percent—a
figure mentioned frequently in Policy Statements of the Shadow Open
Market Committee. Then the growth of Ml would have been lower than the
actual by 2.3 percentage points in 1980, 3.6 percentage points in 1982, 5.2
percentage points in 1983, 7.1 percentage points in 1985, and 10.2
percentage points in 1986. The discrepancies in the money stock are even
more enlightening. In 1980, it would have fallen short of the actual by 2.1
percent of the actual and the gap would have risen to 9.9 percent by 1983, to
nearly 16.0 percent by 1985, and to 23.4 percent by 1986. This enormous
shortfall of money clearly would have played havoc with the economy. How
much would nominal GNP have declined? If the velocity were little affected
by the money supply—a viewpoint that Monetarists occasionally embrace–
then one would have to conclude that nominal GNP would have been one-
fourth, or roughly $1 trillion, lower than it actually was by the end of 1986.
At the same time, the GNP contraction presumably would have led to a price
level somewhat below the current one. Hence, the contraction in real output
and employment would have been somewhat smaller than one-fourth.
In addition, the assumption that velocity would remain constant in the
face of a large contraction in the money supply is not tenable. First, much
evidence—including the Hamburger demand for money equation—indicates
that the demand for money adjusts slowly to its final determinant. According
to the Hamburger estimate, the adjustment speed could be placed at about
11 percent a quarter, or roughly 40 percent a year. This slow adjustment
implies that in the face of a contraction in the money supply, the velocity
would decline initially or, equivalently, the GNP would tend to contract by a
percentage appreciably larger than the change in money supply. At the same
time, the rise in interest accompanying the contraction of M1 could be
expected to raise the velocity. By relying on the standard Hicksian IS-LM
paradigm, one can readily establish that the elasticity of GNP with respect to
a change in money depends on three elasticities: that of saving with respect
to income and that of investment and money demand with respect to the
interest rate. If, as a rough approximation, we put the elasticity of saving at
unity, the short-run elasticity of investment at 0.5, and use Hamburger’s
parameter estimates (speed of adjustment, 40 percent a year; elasticity of
money demand with respect to interest rates in first year of 20 percent), we
find that the short-run elasticity of GNP with respect to a change in the
money supply is somewhere around 0.7, though one might expect it to
decline substantially as the system gradually adjusts to the change. This is
256 Franco Modigliani
true if wages are rigid, though one might expect them to adjust in the longer
run.
But even with these qualifications, the impact of the money shortfall under
a Monetarist policy of constant and relatively low money growth would
have been devastating. This experience demonstrates clearly that at least
some circumstances exist in which Monetarists’ constant money growth
policy with no discretion would prove disastrous. In other words, the system
needs a stabilizing monetary policy.
My next proposition is that an effective stabilization policy is feasible.
The experiment since the late 1970s illustrates this point well. During the
period 1979 to 1986, actual GNP fell squarely within the projected range in
six of these years. During the remaining two years, 1982 and 1985, GNP fell
short of target and the discrepancies were substantial. But one point to note
about these two years is that the actual growth of money was well above the
M1 growth target set for the year which, in turn, was above the Monetarists’
target. This means that the Fed used discretion in deviating from the money
targets so as to approach the GNP target without overshooting it.
I conclude, therefore, that using discretion to track GNP paid off
handsomely during this period in terms of a smooth re-entry from inflation.
Of course, as all neo-Keynesians knew, we could not achieve that re-entry
without substantial unemployment—especially following the second oil
shock. But the Fed had the courage to stick to the path, though it may have
tried to disinflate too quickly in 1982. And since that year, the Fed has
produced a path with both declining inflation (at least until 1987) and rising
output and employment.
POLICY RULES
The recent experience with unstable velocity has led to a string of defections
from Monetarism besides those of the Fed, the Council of Economic Advisers,
and Michael Hamburger. Among those who have given up Monetarism (i.e. a
money growth rate either constant or fixed long in advance) are previously
hard-line Monetarists such as Allan Meltzer, Bennett McCallum, William
Poole, and Thomas Mayer. All have been forced to recognize that the original
formulation of Monetarism cannot be viable after the experience since the late
1970s. Thus, they have endeavored to find a solution that, in keeping with the
spirit of Monetarism, avoids giving the monetary authority any discretionary
powers. This has led them to propose replacing the original constant growth
rule with some other kind of rule–one that will adjust money growth so as to
allow it to reflect variation in velocity. Ideally, one would want to adjust the
change in money growth for the current-period velocity change. But, of course,
the current velocity is not known. One could draw some inferences about it
from a model of the economy, but a Monetarist could never recommend this
since it would violate the principle of no discretion. These considerations have
led both Meltzer (1984) and Mayer (1987) to propose a ‘rule’ according to
The monetarist controversy revisited 257
which one year’s growth of some stated aggregate will equal a constant that is
related to the desirable long-run growth trend of nominal income, less the
percentage increase in velocity over some recent past year. The only difference
between these two authors’ proposals is that Mayer suggests applying the rule
to Ml whereas Meltzer recommends applying it to the base. McCallum (1984)
also has proposed a rule that would add to the constant trend a correction
‘increasing the rate [of money growth] if nominal GNP is below its target and
vice versa.’ But one can show that this rule reduces to the Meltzer-Mayer rule
so long as the ‘correction’ of money growth during each period equals the
previous year’s deviation of GNP from target.
The rule has intuitive appeal, at least if one supposes that velocity can be
described by a random walk with moderate variance. Under these
conditions, the proposed rule could well represent a substantial improvement
over the Monetarist rule. But then, as we have just seen, the latter rule would
have performed so miserably that outperforming it would be no great
achievement. The relevant question is how it would have performed in terms
of income variability, both absolutely and relative to the actual performance
of the Fed, especially over the recent difficult period of erratic velocity
behavior.
One can endeavor to answer this question by simulating the results of the
Meltzer et al. rule.1 To this end, recall that the proposed money supply rule
can be written as
(1)
Here, (t) is the current-period money growth called for by the rule, a is the
long-term growth trend, and v(t-1) is the change in velocity over the recent
past, which we shall identify operationally with the previous year. In
addition, we have the relationship
(2)
where (t) is the change in velocity that actually would occur during period
t if m(t) were equal to (t). Unfortunately, (t) is not directly observable. To
get around this difficulty, let us first assume, that (t) can be approximated
with v(t)—the change actually observed during the period—even though
generally the change in money supply under the rule, (t), would differ from
the actual, m(t). With this assumption, we can substitute for (t) from
equation (1) into equation (2):
(3)
This can readily be used to compute . Column (1) of Table 10.2 shows the
resulting value of for the years 1980 to 1986, assuming for a the value of
5 percent.2 Column (2) shows the corresponding money growth called for by
rule (1). For comparison, columns (3) and (4) show the actual growth in GNP
and money supply.
258 Franco Modigliani
The table indicates that letting the money supply be determined according
to the latest rule the new Monetarism favors would have produced rather
devastating and destabilizing results, at least under the assumption of
(t)=v(t). From column (1), we observe that despite the correction for growth,
the rise of money supply would have been on average rather inadequate. The
actual money growth was a hefty 9.1 percent which, judging from the ex post
result, could not be deemed significantly excessive. By contrast, the money
growth pursuant to the rule hardly matched the long-run trend of 5 percent.
This may seem surprising given the velocity fall during much of this period,
but the lagged velocity actually rose slightly. To be sure, the average growth
rate may not be too relevant since it depends on the somewhat arbitrary choice
of a in equation (1). More revealing is that the money supply growth
according to the rule is even more variable than is the actual growth—even if
by a small margin, or a range of about 11 versus 10 percent.
However, what matters most is not the growth of money but the behavior of
income. Given the unstable growth of the money supply, one should not be
Table 10.2 Simulations of Mayer et al. Money Supply Rule
Notes:
Column (1): From equation (3) and data in columns (2), (3), and
(4) of this table.
Column (2): From equation (1) and the relevant data in Table
10.1. For the purpose of computation, a=5.0 and
the growth rate of velocity, v(t-1), is approximated
by the percentage change in nominal GNP less the
percentage change in M1, in the previous year, as
given in columns (3) and (4) of this table.
Column (3): From Table 10.1, row indicating actual nominal
GNP growth.
Column (4): From Table 10.1, row indicating actual M1 growth.
Column (5): From equation (4) and columns (2), (3), and (4) of
this table.
The monetarist controversy revisited 259
surprised that the behavior of nominal income growth under the rule is also
quite unsatisfactory. The average growth of nominal income turns out to be
only 2.9 percent as compared with the actual growth of 7.5 percent. Clearly,
growth under the rule is inadequate by a wide margin. But again, the most
damaging characteristic of the rule’s performance is the great instability in the
path of income. The variations in actual income growth stay within a range of
7 percentage points, which reduces to 5 if one omits 1982. But for the path
generated by the rule, the range becomes 16.7 percentage points (from -6.0
percent in 1985 to 10.7 percent in 1983) and does not decline by omitting
1982. The greater instability of income under the rule reflects the rule’s calling
for money supply growth substantially lower than the actual precisely during
those years when the income growth already was inadequate—such as 1982,
1985, and 1986—or when the actual growth did not seem excessive, such as
1980. Similarly, the only year in which the rule calls for money growth larger
than the actual is 1983, when the historical growth already was the largest.
During two of these years, 1982 and 1985, the rule would have resulted in
actual contraction in nominal income whereas the actual path shows none.
During the two remaining years, 1980 and 1986, the nominal growth is so
small as to imply a decline in real income.
Altogether, in no single year does the rule call for a growth rate clearly
more appropriate than the historical one. In fact, one can verify that with the
single exception of 1984, such performance is distinctly inferior to even a
mechanical 5 percent growth rate.
However, these results must be qualified in that many may question our
assumption that the velocity under the rule can be equated with that
observed historically. We should expect instead that a faster (or a slower)
money growth should affect velocity in two ways: (i) through the gradual
adjustment of money demand and (ii) through the effect of money on interest
rates and thereby on aggregate demand. Therefore, we have attempted to
estimate the extent to which velocity, under the rule, would differ from the
observed one. To this end, we recall that in the short run, an increase in
money above that observed should tend to be accompanied by a fall in
velocity. In fact, we suggested that this fall could be placed at 0.2, implying
a short-run elasticity of income with respect to money of about 0.7. Allowing
also for the slow adjustment of money demand on which Hamburger
reported, we estimated the change in income resulting from the Mayer et al.
rule but with adjusted velocity, y*, as
(4)
where the last term reflects the slow adjustment of money demand. The growth
of income, y*, estimated with this alternative procedure is reported in column
(5) of Table 10.2, while is unchanged and is given in column (2).
Column (5) indicates that allowing for a variable velocity may, on the
whole, improve somewhat the performance of the money rule—see e.g. the
smaller contraction during 1982 and 1985 and the larger growth during 1980
260 Franco Modigliani
and 1986. The outcome of the rule still remains quite poor compared with the
historical behavior: y* still is appreciably smaller than y when y already was
too small (1982 and 1985) and appreciably larger when y already was quite
large (1981 and 1983). As a result, the variability of y* remains quite large.
The range of more than 18 percentage points far exceeds that of y or even —
or even a 5 percent rule. Thus, whether or not we allow for a velocity
response, the tests seemingly lead to the same negative conclusion and suggest
that the main source of instability from applying the rule arises from the sharp
acceleration and deceleration of velocity such as those occurring during 1982,
1983, and 1985. The discretionary policy of the Fed weathered the storm much
better than did the proposed rule, as evaluated in column (1) or column (5).
The conclusion that one should draw from this experiment seems clear
enough: the new Monetarist rule of Mayer et al. is entirely unacceptable
since in some circumstances, such as the recent episode, it would lead to
instability of major proportions.3 To be sure, our simulations are not
exhaustive. In particular, one could experiment with somewhat different lags
on which to base the velocity correction. But such experiments could not
radically change our conclusions. Indeed, the attraction of Monetarists’ rules
has been that they were mechanical and did not require fine tuning: so long
as money grew at a constant rate, the rate itself was a mere detail. But rules
that depend critically on the specification of the lags would only reinforce
mistrust of monetary management by rule.
This, in turn, leads me to formulate two suggestions. In the first place,
those proposing the new rule should go back to the drawing board and
experiment with less naive rules by taking into account the earlier control
literature.4 Second, if this exercise leads to a truly promising rule, then room
for a compromise could exist between Monetarists who insist on a rule and
everybody else who is unwilling to trust major decisions to any mechanical
preprogrammed rule. I suggest a setup under which the rule normally would
be used to determine the growth of money. However, the monetary authority
would have the option of making a discretionary departure from the rule if it
provided an explicit justification for the departure, under appropriate
procedure. Reasons for such a departure would have to include impending
macro policy changes, such as major changes in expenditures or taxes and in
net exports. Under the proposed eclectic arrangement, the rule would serve
as a check on discretion while discretion would serve as a check on the rule.
This check is vital since, as Table 10.2 indicates, the risks of instability and
disastrous outcomes are much too great to justify placing ourselves entirely
at the mercy of any mechanical rule.
NOTES
1 The test that follows has some similarity to that presented in my 1964 article. Of
course, the money supply rule tested is a different one. Furthermore, at least in the
The monetarist controversy revisited 261
variant described below, the simulation allows for the feedback of the money path
on velocity, including the effect of lags.
2 The percentage change in velocity is approximated as the difference between the
percentage change in GNP and that in Ml, which can be calculated from columns
(3) and (4) of Table 10.2.
3 Poole (1986) has proposed a different rule in which the change in money supply is
given by a constant (or a declining trend) plus a fraction of the change in interest
rates. This clearly is an improvement if the velocity change comes mostly from
interest rates, though not if they come from other sources. Poole provides a graph
comparing the money supply according to his rule with the actual for the period
1915–85. Unfortunately, this graph is not very enlightening for several reasons: (i)
the scale is very small; (ii) the graph does not show the rates of change; and (iii) in
the simulation, he uses the simultaneous interest rate which clearly would be
unavailable in practice.
I have tried to replicate Poole’s simulation for the period 1980–6 by using
annual data and a one-quarter lag for interest rates. The results indicate that
Poole’s rule is a definite improvement over the Mayer et al. rule, notably in 1982,
but still must be rated as unreliable. For example, during two years, the Ml growth
that his rule calls for is only about half as large as the actual while real GNP is
below target.
4 The literature is quite voluminous and extends over several decades. The following
selection is purely indicative: Phillips (1954, 1957), Cooper and Fischer (1974),
Chow (1975), Craine et al. (1976), Tinsley and von zur Muehlen (1981).
REFERENCES
Chow, G.C. (1975) Analysis and Control of Dynamic Economic Systems, New York:
Wiley.
Cooper, J.P. and S.Fischer (1974) ‘A Method for Stochastic Control of Non-linear
Econometric Models and an Application’, Annals of Economic and Social
Measurement January.
Craine, R., A.Havenner and P.Tinsley (1976) ‘Optimal Macro Economic Control
Policies’, Annals of Economic and Social Measurement May.
McCallum, B.T. (1984) ‘Monetarist Rules in the Light of Recent Experience’, American
Economic Review, Papers and Proceedings May.
Mayer, T. (1987) ‘Replacing the FOMC by a PC’, Contemporary Policy Issues April,
pp. 31–43.
Meltzer, A. (1984) ‘Overview’, in Price Stability and Public Policy, Federal Reserve Bank
of Kansas City.
Modigliani, F. (1964) ‘Some Empirical Tests of Monetary Management and of Rules vs.
Discretion’, Journal of Political Economy June.
——(1977) ‘The Monetarist Controversy or, Should We Forsake Stabilization Policies?’
American Economic Review March.
Phillips, A.W. (1954) ‘Stabilization Policy in a Closed Economy’, Economic Journal
June.
——(1957) ‘Stabilization Policy and the Time-Forms of Lagged Responses’, Economic
Journal June.
Poole, W. (1986) ‘Is Monetarism Dead?’, Business Economics October.
Tinsley, P. and P.von zur Muehlen (1981) ‘A Maximum Probability Approach to Short-
Run Policy’, Journal of Econometrics January.
Young, W. 6, 31, 62
Zeldes, S. 322
Zervos, S. 654
Part III
The challenge of rational
expectations and new classical
macroeconomics
Introduction
During the 1970s (at least as far as the United States was concerned) the new
classical research programme replaced monetarism as the main counter-
revolutionary theory to Keynesianism. The primary objective of early new
classical theorists was to build macroeconomic models based on firm
microeconomic foundations involving continuous market clearing and
optimization by economic agents. The leading US exponents of, and/or
contributors to, the first phase of this research programme include Robert
Lucas Jr, Thomas Sargent, Robert Barro, Edward Prescott and Neil Wallace.
In their 1978 paper entitled ‘After Keynesian Macroeconomics’ (reprinted on
pp. 270–94), Robert Lucas and Thomas Sargent argue that Keynesian models
in the 1960s were fatally flawed as they lacked sound microfoundations and
incorporated an adaptive expectations hypothesis which is inconsistent with
maximizing behaviour. In addition to these theoretical failings they also
argue that in the 1970s Keynesian macroeconometric models experienced
‘econometric failure on a grand scale’ and that being subject to the Lucas
critique could not be used to guide policy. In their place Lucas and Sargent
advocate ‘equilibrium’ new classical models (in which markets always clear
and ‘rational’ agents optimize) suggesting that equilibrium models can
account for the main features of the business cycle with economic
fluctuations being triggered by unanticipated (monetary) shocks. In the final
section of their paper they respond to four lines of criticism which have been
raised against equilibrium new classical models relating to the assumption of
continuous market clearing, the observed persistence of cyclical movements
in output and employment, linearity, and the neglect of learning. For a
discussion of the rhetoric surrounding this article see Backhouse (1997).
New classical models are based on the (highly controversial) assumption
of continuous market clearing, together with the rational expectations and
aggregate supply hypotheses. In these models fluctuations in output and
employment reflect the voluntary response of rational economic agents who
misperceive money price changes for relative price changes due to
incomplete information. Rodney Maddock and Michael Carter’s 1982
Journal of Economic Literature article (reprinted on pp. 295–313) entitled ‘A
Child’s Guide to Rational Expectations’ examines a number of key issues
266 Challenge of new classical macroeconomics
surrounding the structure and policy implications of new classical models.
The article is entertainingly couched in the form of a play involving a
conversation between two students, one Keynesian in orientation and the
other more disposed to monetarism. In a series of scenes set in a student’s
union, first over coffee and later over a beer, the two students discuss the
basic ideas underlying rational expectations; the implications for
stabilization policy of combining the rational expectations hypothesis with a
new classical aggregate supply hypothesis (most notably surrounding the
policy impotence or ineffectiveness proposition); criticisms of the new
classical approach; and finally evidence on, and the significance of, ‘new
classical’ rational expectations models.
The structure of new classical models leads to a number of controversial
policy implications which, in addition to the policy ineffectiveness
proposition, feed into the debate over the role and conduct of stabilization
policy. One of the most debated policy implications that result from the new
classical approach concerns the ‘Ricardian equivalence theorem’, a term first
introduced by James Buchanan (1976). The theorem is also sometimes known
as the ‘Ricardo-Barro equivalence theorem’ as it was developed in its
modern form by Robert Barro (1974) but was first articulated by the famous
nineteenth-century English economist David Ricardo, who ironically
expressed doubts about its validity. In essence this theorem states that the
effects of an increase in government expenditure on the economy will be the
same whether it is financed by increased taxation or borrowing from the
private sector. In the latter case it is claimed that the private sector will
merely react to a bond-financed increase in government expenditure by
saving more in the present period in order to meet future tax liabilities. In
this approach changes in the government budget deficit caused by tax
changes have no effect on aggregate demand. In his 1989 Journal of
Economic Perspectives article (reprinted on pp. 314–33) on ‘The Ricardian
Approach to Budget Deficits’, Robert Barro discusses five major theoretical
objections which have been put forward in the literature to Ricardian
equivalence, and the empirical evidence on the effects of budget deficits on
interest rates, saving and the current account balance.
The final contribution is David Laidler’s 1986 Banca Nazionale Del
Lavoro Quarterly Review article entitled ‘The New-Classical Contribution
to Macroeconomics’ (reprinted on pp. 334–58). Laidler critically assesses the
basic premises underlying new classical macroeconomics and considers what
contributions of lasting importance the new classical school has made to
macroeconomics since the early 1970s. After examining the relationship
between monetarism and new classical macroeconomics, Laidler considers
the relative explanatory power of these two doctrines with respect to the
empirical evidence and the new classical claim that their approach
incorporates a superior analytical method. He concludes by suggesting that
the application of certain basic ideas which underlie the new classical
approach to macroeconomics has been unnecessarily restrictive. In
Introduction 267
particular, he questions that the maximizing behaviour of agents should
always be analysed in the context of continuously clearing markets, and that
agents’ rational expectations accord with those that would be generated with
knowledge of the ‘true’ model of the economy. However, Laidler also accepts
that the insistence on equilibrium modelling of individuals as the basis for
macroeconomic modelling has been both valuable and beneficial.
Although the early 1980s witnessed the demise of the mark I (monetary
surprise) version of the new classical approach (not least due to the
implausibility of supposed information gaps relating to aggregate price level
and money supply data, and the failure of empirical tests to provide strong
support for the policy ineffectiveness proposition) it set down the seeds for the
development of a mark II version, commonly referred to as the real business
cycle approach, in which real supply-side factors rather than monetary
impulses are emphasized in explaining economic fluctuations. Before turning
to four articles which consider various aspects of the real business cycle
approach in Part IV, it would be useful to summarize the contribution that
the first phase of this new classical revolution has made to macroeconomics.
That contribution can be seen in four main directions. First, it has led to the
widespread practice of applying equilibrium modelling to macroeconomic
analysis. Second, it has accelerated the movement towards the now much
more widely accepted view that any satisfactory macroeconomic analysis
needs to be based on firm microfoundations. Third, it has led to the
widespread adoption of the rational expectations hypothesis in
macroeconomics. Finally, the insights provided by the incorporation of the
rational expectations hypothesis into macroeconomic models, together with
the literature on time-inconsistency and the Lucas critique, have led
economists to critically reappraise the traditional approaches to policy-
making and evaluation.
REFERENCES
*Titles marked with an asterisk are particularly recommended for additional
reading.
*Abel, A.B. and B.S.Bernanke (1995) Macroeconomics, 2nd edn, Chapter 11, New
York: Addison Wesley.
*Backhouse, R.E. (1997) ‘The Rhetoric and Methodology of Modern
Macroeconomics’, in B.Snowdon and H.R.Vane (eds) Reflections on the
Development of Modern Macroeconomics, Aldershot: Edward Elgar.
Barro, R.J. (1974) ‘Are Government Bonds Net Wealth?’, Journal of Political Economy
82, November/December, pp. 1095–117.
Barro, R.J. (1984) ‘What Survives of the Rational Expectations Revolution? Rational
Expectations and Macroeconomics in 1984’, American Economic Review 74, May,
pp. 179–82.
*Barro, R.J. (1989) ‘New Classical and Keynesians, or the Good Guys and Bad Guys’,
Schwiez Zeitschrift für Volkswirtschaft und Statistik 3, pp. 263–73.
*Barro, R.J. and V.Grilli (1994) European Macroeconomics, Chapter 20, London:
Macmillan.
268 Challenge of new classical macroeconomics
Buchanan, J.M. (1976) ‘Barro on the Ricardian Equivalence Theorem’, Journal of
Political Economy 84, April, pp. 337–42.
*Carter, M. and R.Maddock (1984) Rational Expectations: Macroeconomics for the
1980s?, London: Macmillan.
*Dornbusch, R. and S.Fischer (1994) Macroeconomics, 6th edn, Chapter 9, New York:
McGraw-Hill.
*Froyen, R.T. (1996) Macroeconomics: Theories and Policies, 5th edn, Chapter 11,
London: Prentice-Hall.
*Gordon, R.J. (1993) Macroeconomics, 6th edn, Chapter 7, New York: HarperCollins.
*Hall, R.E. and J.B.Taylor (1993) Macroeconomics, 4th edn, Chapter 15, New York:
W.W.Norton.
* Hoover, K.D. (1992) ‘The Rational Expectations Revolution: An Assessment’, Cato
Journal Spring/Summer, pp. 81–96.
*Jansen, D.W., C.D.Delorme and R.B.Ekelund, Jr (1994) Intermediate Macroeconomics,
Chapter 9, New York: West.
Kantor, B. (1979) ‘Rational Expectations and Economic Thought’, Journal of Economic
Literature 17, December, pp. 1422–41.
Klamer, A. (1984) ‘Levels of Discourse in New Classical Economies’, History of Political
Economy 16, Summer, pp. 263–90.
*Klamer, A. (1984) The New Classical Macroeconomics: Conversations with New
Classical Economists and their Opponents, Brighton: Wheatsheaf.
*Lucas, R.E. Jr (1977) ‘Understanding Business Cycles’, in K.Brunner and A.H. Meltzer
(eds) Stabilization of the Domestic and International Economy, Amsterdam: North
Holland.
Maddock, R. (1984) ‘Rational Expectations Macrotheory: A Lakatosian Case Study
in Programme Adjustment’, History of Political Economy 16, Summer, pp. 291–
310.
*Mankiw, N.G. (1994) Macroeconomics, 2nd edn, Chapters 11 and 16, New York:
Worth.
Minford, P. (1997) ‘Macroeconomics: Before and After Rational Expectations’, in
B.Snowdon and H.R.Vane (eds) Reflections on the Development of Modern
Macroeconomics, Aldershot: Edward Elgar.
*Shaw, G.K. (1984) Rational Expectations: An Elementary Exposition, Brighton:
Wheatsheaf.
*Sheffrin, S.M. (1996) Rational Expectations, 2nd edn, Cambridge: Cambridge
University Press.
*Snowdon, B. and H.R.Vane (1996) ‘The Development of Modern Macroeconomics:
Reflections in the Light of Johnson’s Analysis after Twenty-Five Years’, Journal of
Macroeconomics 18, Summer, pp. 381–401.
*Snowdon, B., H.R.Vane and P.Wynarczyk (1994) A Modern Guide to Macroeconomics:
An Introduction to Competing Schools of Thought, Chapter 5, Aldershot: Edward
Elgar.
*Tobin, J. (1980) ‘Are New Classical Models Plausible Enough to Guide Policy?’,
Journal of Money, Credit and Banking 12, November, pp. 788–99.
QUESTIONS
1 ‘Orthodox monetarism and early new classical models represent two types
of monetarism’. Critically examine this view.
2 Explain why in the new classical market clearing model the Phillips curve
emerges as a result of imperfect information about the aggregate price
level.
Introduction 269
3 Critically assess the importance of the rational expectations hypothesis within
the new classical approach to macroeconomics. What are the implications
for stabilization policy?
4 ‘The crucial assumption in new classical analysis is market clearing not
rational expectations’. Do you agree?
5 In new classical models only unanticipated monetary shocks can affect real
output and employment. How plausible is this theory as an explanation of
economic fluctuations?
6 Assess the significance of the Lucas critique with respect to the construction
of macroeconomic models.
7 Compare and contrast the traditional and Ricardian approaches to the impact
of budget deficits.
8 Outline the dynamic time inconsistency argument in favour of a monetary
rule. Does this analysis point towards the desirability of having an
independent central bank?
9 To what extent do the real output and employment effects of disinflation
depend on the credibility and reputation of the policy-maker?
10 Evaluate the contribution made to macroeconomic analysis by new classical
theorists in the period 1972–82. What remains of the first phase of the new
classical revolution?
11 After Keynesian
macroeconomics
Robert E.Lucas and Thomas J.Sargent
After the Phillips Curve: Persistence of High Inflation and High
Unemployment (1978) Boston, MA: Federal Reserve Bank of Boston,
pp. 49–72
INTRODUCTION
For the applied economist, the confident and apparently successful
application of Keynesian principles to economic policy which occurred in the
United States in the 1960s was an event of incomparable significance and
satisfaction. These principles led to a set of simple, quantitative relationships
between fiscal policy and economic activity generally, the basic logic of
which could be (and was) explained to the general public, and which could
be applied to yield improvements in economic performance benefiting
everyone. It seemed an economics as free of ideological difficulties as, say,
applied chemistry or physics, promising a straightforward expansion in
economic possibilities. One might argue about how this windfall should be
distributed, but it seemed a simple lapse of logic to oppose the windfall itself.
Understandably and correctly, this promise was met at first with skepticism
by non-economists; the smoothly growing prosperity of the Kennedy-Johnson
years did much to diminish these doubts.
We dwell on these halcyon days of Keynesian economics because, without
conscious effort, they are difficult to recall today. In the 1970s, the US
economy has undergone its first major depression since the 1930s, to the
accompaniment of inflation rates in excess of 10 percent per annum. These
events have been transmitted (by consent of the governments involved) to
other advanced countries and in many cases have been amplified. These
events did not arise from a reactionary reversion to outmoded, ‘classical’
principles of tight money and balanced budgets. On the contrary, they were
accompanied by massive governmental budget deficits and high rates of
monetary expansion: policies which, although bearing an admitted risk of
inflation, promised according to modern Keynesian doctrine rapid real
growth and low rates of unemployment.
That these predictions were wildly incorrect, and that the doctrine on
which they were based is fundamentally flawed, are now simple matters of
fact, involving no novelties in economic theory. The task which faces
contemporary students of the business cycle is that of sorting through the
wreckage, determining which features of that remarkable intellectual event
After Keynesian macroeconomics 271
called the Keynesian Revolution can be salvaged and put to good use, and
which others must be discarded. Though it is far from clear what the
outcome of this process will be, it is already evident that it will necessarily
involve the reopening of basic issues in monetary economics which have
been viewed since the 1930s as ‘closed’, and the re-evaluation of every
aspect of the institutional framework within which monetary and fiscal
policy is formulated in the advanced countries.
This chapter is in the nature of an early progress report on this process of
re-evaluation and reconstruction. We begin by reviewing the econometric
framework by means of which Keynesian theory evolved from disconnected,
qualitative ‘talk’ about economic activity into a system of equations which
could be compared to data in a systematic way, and provide an operational
guide in the necessarily quantitative task of formulating monetary and fiscal
policy. Next, we identify those aspects of this framework which were central
to its failure in the 1970s. In so doing, our intent will be to establish that the
difficulties are fatal: that modern macroeconomic models are of no value in
guiding policy, and that this condition will not be remedied by modifications
along any line which is currently being pursued.
This diagnosis, if successful, will suggest certain principles which a useful
theory of business cycles must possess. In the latter part of this chapter we
shall review some recent research which is consistent with these principles.
MACROECONOMETRIC MODELS
The Keynesian Revolution was, in the form in which it succeeded in the
United States, a revolution in method. This was not Keynes’s [13] intent, nor
is it the view of all of his most eminent followers. Yet if one does not view
the revolution in this way, it is impossible to account for some of its most
important features: the evolution of macroeconomics into a quantitative,
scientific discipline, the development of explicit statistical descriptions of
economic behavior, the increasing reliance of government officials on
technical economic expertise, and the introduction of the use of
mathematical control theory to manage an economy. It is the fact that
Keynesian theory lent itself so readily to the formulation of explicit
econometric models which accounts for the dominant scientific position it
attained by the 1960s.
As a consequence of this, there is no hope of understanding either the
success of the Keynesian Revolution or its eventual failure at the purely
verbal level at which Keynes himself wrote. It will be necessary to know
something of the way macroeconometric models are constructed and the
features they must have in order to ‘work’ as aids in forecasting and policy
evaluation. To discuss these issues, we introduce some notation.
An econometric model is a system of equations involving a number of
endogenous variables (variables that are determined by the model),
exogenous variables (variables which affect the system but are not affected
272 Robert E.Lucas and Thomas J.Sargent
by it), and stochastic or random shocks. The idea is to use historical data to
estimate the model, and then to utilize the estimated version to obtain
estimates of the consequences of alternative policies. For practical reasons, it
is usual to use a standard linear model, taking the structural form.1
Here yt is an (Lx1) vector of endogenous variables, xt is a (Kx1) vector of
exogenous variables, and εt and ut are each (Lx1) vectors of random
disturbances. The matrices Aj are each (LxL); the Bj’s are (LxK), and the Rj’s
are each (LxL). The (Lxl) disturbance process ut is assumed to be a serially
uncorrelated process with Eut=0 and with contemporaneous covariance
matrix =∑ and =0 for all t≠s. The defining characteristic of the
exogenous variables xt is that they are uncorrelated with the ε’s at all lags so
that is an (LxK) matrix of zeroes for all t and s.
Equations (1) are L equations in the L current values yt of the endogenous
variables. Each of these structural equations is a behavioral relationship,
identity, or market clearing condition, and each in principle can involve a
number of endogenous variables. The structural equations are usually not
‘regression equations’,2 because the εt’s are in general, by the logic of the
model, supposed to be correlated with more than one component of the
vector yt and very possibly one or more components of the vectors yt-1,…yt-m.
The structural model (1) and (2) can be solved for yt in terms of past y’s
and x’s and past shocks. This ‘reduced form’ system is
The reduced form equations are ‘regression equations’, that is, the
disturbance vector is orthogonal to yt-1,…,yt-r-m,xt,…,xt-n-r. This follows
from the assumptions that the x’s are exogenous and that the u’s are serially
uncorrelated. Therefore, under general conditions the reduced form can be
estimated consistently by the method of least squares. The population
parameters of the reduced form (3) together with the parameters of a vector
autoregression for xt,
(1)
(2)
(3)
(4)
After Keynesian macroeconomics 273
where Eat=0 and Eat·xt-j=0 for j�1 completely describe all of the first and
second moments of the (yt, xt) process. Given long enough time series, good
estimates of the reduced form parameters—the Pj’s and Qj’s– can be obtained
by the method of least squares. Reliable estimates of those parameters is all
that examination of the data by themselves can deliver.
It is not in general possible to work backwards from estimates of the P’s
and Q’s alone to derive unique estimates of the structural parameters, the
Aj’s, Bj’s, and Rj’s. In general, infinite numbers of A, B, and R’s are
compatible with a single set of P’s and Q’s. This is the ‘identification
problem’ of econometrics. In order to derive a set of estimated structural
parameters, it is necessary to know a great deal about them in advance. If
enough prior information is imposed, it is possible to extract estimates of the
(Aj, Bj, Rj)’s implied by the data in combination with the prior information.
For purposes of ex ante forecasting, or the unconditional prediction of the
vector yt+1,yt+2,…given observation of ys and xs, s�t, the estimated reduced
form (3), together with (4), is sufficient. This is simply an exercise in a
sophisticated kind of extrapolation, requiring no understanding of the
structural parameters or, that is to say, of the economics of the model.
For purposes of conditional forecasting, or the prediction of the future
behavior of some components of yt and xt conditional on particular values of
other components, selected by policy, one needs to know the structural
parameters. This is so because a change in policy necessarily alters some of
the structural parameters (for example, those describing the past behavior of
the policy variables themselves) and therefore affects the reduced form
parameters in highly complex fashion (see the equations defining Ps and Qs,
below (3)). Without knowledge as to which structural parameters remain
invariant as policy changes, and which change (and how), an econometric
model is of no value in assessing alternative policies. It should be clear that
this is true regardless of how well (3) and (4) fit historical data, or how well
they perform in unconditional forecasting.
Our discussion to this point has been at a high level of generality, and the
formal considerations we have reviewed are not in any way specific to
Keynesian models. The problem of identifying a structural model from a
collection of economic time series is one that must be solved by anyone who
claims the ability to give quantitative economic advice. The simplest
Keynesian models are attempted solutions to this problem, as are the large-
scale versions currently in use. So, too, are the monetarist models which
imply the desirability of fixed monetary growth rules. So, for that matter, is
the armchair advice given by economists who claim to be outside the
econometric tradition, though in this case the implicit, underlying structure is
not exposed to professional criticism. Any procedure which leads from the
study of observed economic behavior to the quantitative assessment of
alternative economic policies involves the steps, executed poorly or well,
explicitly or implicitly, which we have outlined above.
274 Robert E.Lucas and Thomas J.Sargent
KEYNESIAN MACROECONOMETRICS
In Keynesian macroeconometric models structural parameters are identified
by the imposition of several types of a priori restrictions on the A
j
’s, B
j
’s, and
Rj’s. These restrictions usually fall into one of the following categories:4
(a) A priori setting of many of the elements of the A
j
’s and B
j
’s to zero.
(b) Restrictions on the orders of serial correlation and the extent of the cross
serial correlation of the disturbance vector ε
t
, restrictions which amount
to a priori setting many elements of the Rj’s to zero.
(c) A priori categorization of variables into ‘exogenous’ and ‘endogenous’.
A relative abundance of exogenous variables aids identification.
Existing large Keynesian macroeconometric models are open to serious
challenge for the way they have introduced each category of restriction.
Keynes’s General Theory was rich in suggestions for restrictions of type
(a). It proposed a theory of national income determination built up from
several simple relationships, each involving a few variables only. One of
these, for example, was the ‘fundamental law’ relating consumption
expenditures to income. This suggested one ‘row’ in equations (1) involving
current consumption, current income, and no other variables, thereby
imposing many zero-restrictions on the Ai and Bj. Similarly, the liquidity
preference relation expressed the demand for money as a function of income
and an interest rate only. By translating the building blocks of the Keynesian
theoretical system into explicit equations, models of the form (1) and (2)
were constructed with many theoretical restrictions of type (a).
Restrictions on the coefficients Ri governing the behavior of the ‘error
terms’ in (1) are harder to motivate theoretically, the ‘errors’ being by
definition movements in the variables which the economic theory cannot
account for. The early econometricians took ‘standard’ assumptions from
statistical textbooks, restrictions which had proved useful in the agricultural
experimenting which provided the main impetus to the development of
modern statistics. Again, these restrictions, well-motivated or not, involve
setting many elements in the Ri’s equal to zero, aiding identification of the
model’s structure.
The classification of variables into ‘exogenous’ and ‘endogenous’ was also
done on the basis of prior considerations. In general, variables were classed
as ‘endogenous’ which were, as a matter of institutional fact, determined
largely by the actions of private agents (like consumption or private
investment expenditures). Exogenous variables were those under
governmental control (like tax rates, or the supply of money). This division
was intended to reflect the ordinary meaning of the word ‘endogenous’ to
mean ‘determined by the [economic] system’ and ‘exogenous’ to mean
‘affecting the [economic] system but not affected by it’.
By the mid-1950s, econometric models had been constructed which fit
time series data well, in the sense that their reduced forms (3) tracked past
After Keynesian macroeconomics 275
data closely and proved useful in short-term forecasting. Moreover, by
means of restrictions of the three types reviewed above, it was possible to
identify their structural parameters Ai, Bj, Rk. Using this estimated structure,
it was possible to simulate the models to obtain estimates of the
consequences of different government economic policies, such as tax rates,
expenditures or monetary policy.
This Keynesian solution to the problem of identifying a structural model
has become increasingly suspect as a result of developments of both a
theoretical and statistical nature. Many of these developments are due to
efforts to researchers sympathetic to the Keynesian tradition, and many were
well advanced well before the spectacular failure of the Keynesian models in
the 1970s.5
Since its inception, macroeconomics has been criticized for its lack of
‘foundations in microeconomic and general equilibrium theory’. As astute
commentators like Leontief [14] (disapprovingly) and Tobin [37]
(approvingly) recognized early on, the creation of a distinct branch of theory
with its own distinct postulates was Keynes’s conscious aim. Yet a main
theme of theoretical work since the General Theory has been the attempt to
use microeconomic theory based on the classical postulate that agents act in
their own interests to suggest a list of variables that belong on the right side
of a given behavioral schedule, say, a demand schedule for a factor of
production or a consumption schedule.6 But from the point of view of
identification of a given structural equation by means of restrictions of type
(a), one needs reliable prior information that certain variables should be
excluded from the right-hand side. Modern probabilistic microeconomic
theory almost never implies either the exclusion restrictions that were
suggested by Keynes or those that are imposed by macroeconometric models.
To take one example that has extremely dire implications for the
identification of existing macro models, expectations about the future
prices, tax rates, and income levels play a critical role in many demand
and supply schedules in those models. For example, in the best models,
investment demand typically is supposed to respond to businessmen’s
expectations of future tax credits, tax rates, and factor costs. The supply of
labor typically is supposed to depend on the rate of inflation that workers
expect in the future. Such structural equations are usually identified by the
assumption that, for example, the expectation about the factor price or rate
of inflation attributed to agents is a function only of a few lagged values of
the variable itself which the agent is supposed to be forecasting. However,
the macro models themselves contain complicated dynamic interactions
among endogenous variables, including factor prices and the rate of
inflation, and generally imply that a wise agent would use current and
many lagged values of many and usually most endogenous and exogenous
variables in the model in order to form expectations about any one
variable. Thus, virtually any version of the hypothesis that agents behave
in their own interests will contradict the identification restrictions imposed
276 Robert E.Lucas and Thomas J.Sargent
on expectations formation. Further, the restrictions on expectations that
have been used to achieve identification are entirely arbitrary and have not
been derived from any deeper assumption reflecting first principles about
economic behavior. No general first principle has ever been set down
which would imply that, say, the expected rate of inflation should be
modeled as a linear function of lagged rates of inflation alone with weights
that add up to unity, yet this hypothesis is used as an identifying restriction
in almost all existing models. The casual treatment of expectations is not a
peripheral problem in these models, for the role of expectations is
pervasive in the models and exerts a massive influence on their dynamic
properties (a point Keynes himself insisted on). The failure of existing
models to derive restrictions on expectations from any first principles
grounded in economic theory is a symptom of a somewhat deeper and more
general failure to derive behavioral relationships from any consistently
posed dynamic optimization problems.
As for the second category, restrictions of type (b), existing Keynesian
macro models make severe a priori restrictions on the Rj’s. Typically, the Rj’s
are supposed to be diagonal so that cross equation lagged serial correlation
is ignored and also the order of the εt process is assumed to be short so that
only low-order serial correlation is allowed. There are at present no
theoretical grounds for introducing these restrictions, and for good reasons
there is little prospect that economic theory will soon provide any such
grounds. In principle, identification can be achieved without imposing any
such restrictions. Foregoing the use of category (b) restrictions would
increase the category (a) and (c) restrictions needed. In any event, existing
macro models do heavily restrict the R’s.
Turning to the third category, all existing large models adopt an a priori
classification of variables into the categories of strictly endogenous
variables, the yt’s, and strictly exogenous variables, the xt’s. Increasingly, it
is being recognized that the classification of a variable as ‘exogenous’ on
the basis of the observation that it could be set without reference to the
current and past values of other variables has nothing to do with the
econometrically relevant question of how this variable has in fact been
related to others over a given historical period. Moreover, in light of recent
developments in time series econometrics, we know that this arbitrary
classification procedure is not necessary. Christopher Sims [34] has shown
that in a time series context the hypothesis of econometric exogeneity can
be tested. That is, Sims showed that the hypothesis that xt is strictly
econometrically exogenous in (1) necessarily implies certain restrictions
that can be tested given time series on the y’s and x’s. Tests along the lines
of Sims’s ought to be used as a matter of course in checking out
categorizations into exogenous and endogenous sets of variables. To date
they have not been. Prominent builders of large econometric models have
even denied the usefulness of such tests.7
After Keynesian macroeconomics 277
FAILURE OF KEYNESIAN MACROECONOMETRICS
Our discussion in the preceding section raised a number of theoretical
reasons for believing that the parameters identified as structural by the
methods which are in current use in macroeconomics are not structural in
fact. That is, there is no reason, in our opinion, to believe that these models
have isolated structures which will remain invariant across the class of
interventions that figure in contemporary discussions of economic policy. Yet
the question of whether a particular model is structural is an empirical, not a
theoretical, one. If the macroeconometric models had compiled a record of
parameter stability, particularly in the face of breaks in the stochastic
behavior of the exogenous variables and disturbances, one would be
skeptical as to the importance of prior theoretical objections of the sort we
have raised.
In fact, however, the track record of the major econometric models is, on
any dimension other than very short-term unconditional forecasting, very
poor. Formal statistical tests for parameter instability, conducted by sub-
dividing past series into periods and checking for parameter stability across
time, invariably reveal major shifts (for one example, see [23]). Moreover,
this difficulty is implicitly acknowledged by model-builders themselves, who
routinely employ an elaborate system of add-factors in forecasting, in an
attempt to offset the continuing ‘drift’ of the model away from the actual
series.
T h o u g h n o t , o f c o u r s e , d e s i g n e d a s s u c h b y a n y o n e ,
macroeconometric models were subjected in the 1970s to a decisive test.
A key element in all Keynesian models is a ‘trade-off between inflation
and real output: the higher is the inflation rate; the higher is output (or
equivalently, the lower is the rate of unemployment). For example, the
models of the late 1960s predicted a sustained unemployment rate in the
United States of 4 percent as consistent with a 4 percent annual rate of
inflation. Many economists at that time urged a deliberate policy of
inflation on the basis of this prediction. Certainly the erratic ‘fits and
starts’ character of actual US policy in the 1970s cannot be attributed to
recommendations based on Keynesian models, but the inflationary bias
on average of monetary and fiscal policy in this period should,
according to all of these models, have produced the lowest average
unemployment rates for any decade since the 1940s. In fact, as we
know, they produced the highest unemployment since the 1930s. This
was econometric failure on a grand scale.
This failure has not led to widespread conversions of Keynesian
economists to other faiths, nor should it have been expected to. In economics,
as in other sciences, a theoretical framework is always broader and more
flexible than any particular set of equations, and there is always the hope
that, if a particular specific model fails, one can find a more successful one
based on ‘roughly’ the same ideas. It has, however, already had some
278 Robert E.Lucas and Thomas J.Sargent
important consequences, with serious implications both for economic policy-
making and for the practice of economic science.
For policy, the central fact is that Keynesian policy recommendations
have no sounder basis, in a scientific sense, than recommendations of non-
Keynesian economists or, for that matter, non-economists. To note one
consequence of the wide recognition of this, the current wave of
protectionist sentiment directed at ‘saving jobs’ would have been answered,
in the late 1960s, with the Keynesian counter-argument that fiscal policy
can achieve the same end, but more efficiently. In the late 1970s, of course,
no one would take this response seriously, so it is not offered. Indeed,
economists who in the late 1960s championed Keynesian fiscal policy as an
alternative to inefficient direct controls increasingly favor the latter as
‘supplements’ to Keynesian policy. The idea seems to be that if people
refuse to obey the equations we have fit to their past behavior, we can pass
laws to make them do so.
Scientifically, the Keynesian failure of the 1970s has resulted in a new
openness. Fewer and fewer economists are involved in monitoring and
refining the major econometric models; more and more are developing
alternative theories of the business cycle, based on different theoretical
principles. In addition, increased attention and respect are accorded to the
theoretical casualties of the Keynesian Revolution, to the ideas of Keynes’s
contemporaries and of earlier economists whose thinking has been regarded
for years as outmoded.
At the present time, it is impossible to foresee where these developments
will lead. Some, of course, continue to believe that the problems of existing
Keynesian models can be resolved within the existing framework, that these
models can be adequately refined by changing a few structural equations, by
adding or subtracting a few variables here and there, or perhaps by
disaggregating various blocks of equations. We have couched our preceding
criticisms in such general terms precisely to emphasize their generic
character and hence the futility of pursuing minor variations within this
general framework.
A second response to the failure of Keynesian analytical methods is to
renounce analytical methods entirely, returning to ‘judgmental’ methods.
The first of these responses identifies the quantitative, scientific goals of the
Keynesian Revolution with the details of the particular models so far
developed. The second renounces both these models and the objectives they
were designed to attain. There is, we believe, an intermediate course, to
which we now turn.
EQUILIBRIUM BUSINESS CYCLE THEORY
Economists prior to the 1930s did not recognize a need for a special branch
of economics, with its own special postulates, designed to explain the
business cycle. Keynes founded that subdiscipline, called macroeconomics,
After Keynesian macroeconomics 279
because he thought that it was impossible to explain the characteristics of
business cycles within the discipline imposed by classical economic theory, a
discipline imposed by its insistence on adherence to the two postulates (a)
that markets be assumed to clear, and (b) that agents be assumed to act in
their own self-interest. The outstanding fact that seemed impossible to
reconcile with these two postulates was the length and severity of business
depressions and the large-scale unemployment which they entailed. A related
observation is that measures of aggregate demand and prices are positively
correlated with measures of real output and employment, in apparent
contradiction to the classical result that changes in a purely nominal
magnitude like the general price level were pure ‘unit changes’ which should
not alter real behavior. After freeing himself of the strait-jacket (or
discipline) imposed by the classical postulates, Keynes described a model in
which rules of thumb, such as the consumption function and liquidity
preference schedule, took the place of decision functions that a classical
economist would insist be derived from the theory of choice. And rather than
require that wages and prices be determined by the postulate that markets
clear—which for the labor market seemed patently contradicted by the
severity of business depressions—Keynes took as an unexamined postulate
that money wages are ‘sticky’, meaning that they are set at a level or by a
process that could be taken as uninfluenced by the macroeconomic forces he
proposed to analyze.
When Keynes wrote, the terms ‘equilibrium’ and ‘classical’ carried certain
positive and normative connotations which seemed to rule out either
modifier being applied to business cycle theory. The term ‘equilibrium’ was
thought to refer to a system ‘at rest’, and both ‘equilibrium’ and ‘classical’
were used interchangeably, by some, with ‘ideal’. Thus an economy in
classical equilibrium would be both unchanging and unimprovable by policy
interventions. Using terms in this way, it is no wonder that few economists
regarded equilibrium theory as a promising starting point for the
understanding of business cycles, and for the design of policies to mitigate or
eliminate them.
In recent years, the meaning of the term ‘equilibrium’ has undergone such
dramatic development that a theorist of the 1930s would not recognize it. It
is now routine to describe an economy following a multivariate stochastic
process as being ‘in equilibrium’, by which is meant nothing more than that
at each point in time, postulates (a) and (b) above are satisfied. This
development, which stemmed mainly from work by K.J.Arrow [2] and
G.Debreu [6], implies that simply to look at any economic time series and
conclude that it is a ‘disequilibrium phenomenon’ is a meaningless
observation. Indeed, a more likely conjecture, on the basis of work by Hugo
Sonnenschein [36], is that the general hypothesis that a collection of time
series describes an economy in competitive equilibrium is without content.8
The research line being pursued by a number of us involves the attempt to
discover a particular, econometrically testable equilibrium theory of the
280 Robert E.Lucas and Thomas J.Sargent
business cycle, one that can serve as the foundation for quantitative analysis
of macroeconomic policy. There is no denying that this approach is ‘counter-
revolutionary’, for it presupposes that Keynes and his followers were wrong
to give up on the possibility that an equilibrium theory could account for the
business cycle. As of now, no successful equilibrium macroeconometric
model at the level of detail of, say, the FMP model, has been constructed.
But small theoretical equilibrium models have been constructed that show
potential for explaining some key features of the business cycle long thought
to be inexplicable within the confines of classical postulates. The equilibrium
models also provide reasons for understanding why estimated Keynesian
models fail to hold up outside of the sample over which they have been
estimated. We now turn to describing some of the key facts about business
cycles and the way the new classical models confront them.
For a long time most of the economics profession has, with some reason,
followed Keynes in rejecting classical macroeconomic models because they
seemed incapable of explaining some important characteristics of time series
measuring important economic aggregates. Perhaps the most important
failure of the classical model seemed to be its inability to explain the positive
correlation in the time series between prices and/or wages, on the one hand,
and measures of aggregate output or employment, on the other hand. A
second and related failure was its inability to explain the positive
correlations between measures of aggregate demand, like the money stock,
and aggregate output or employment. Static analysis of classical
macroeconomic models typically implied that the levels of output and
employment were determined independently of both the absolute level of
prices and of aggregate demand. The pervasive presence of the above
mentioned positive correlations in the time series seems consistent with
causal connections flowing from aggregate demand and inflation to output
and employment, contrary to the classical ‘neutrality’ propositions.
Keynesian macroeconometric models do imply such causal connections.
We now have rigorous theoretical models which illustrate how these
correlations can emerge while retaining the classical postulates that markets
clear and agents optimize.9 The key step in obtaining such models has been
to relax the ancillary postulate used in much classical economic analysis that
agents have perfect information. The new classical models continue to
assume that markets always clear and that agents optimize. The postulate
that agents optimize means that their supply and demand decisions must be
functions of real variables, including perceived relative prices. Each agent is
assumed to have limited information and to receive information about some
prices more often than other prices. On the basis of their limited
information—the lists that they have of current and past absolute prices of
various goods—agents are assumed to make the best possible estimate of all
of the relative prices that influence their supply and demand decisions.
Because they do not have all of the information that would enable them to
compute perfectly the relative prices they care about, agents make errors in
After Keynesian macroeconomics 281
estimating the pertinent relative prices, errors that are unavoidable given
their limited information. In particular, under certain conditions, agents will
tend temporarily to mistake a general increase in all absolute prices as an
increase in the relative price of the good that they are selling, leading them
to increase their supply of that good over what they had previously planned.
Since everyone is, on average, making the same mistake, aggregate output
will rise above what it would have been. This increase of output will rise
above what it would have been. This increase of output above what it would
have been will occur whenever this period’s average economy-wide price
level is above what agents had expected this period’s average economy-wide
price level to be on the basis of previous information. Symmetrically,
average output will be decreased whenever the aggregate price turns out to
be lower than agents had expected. The hypothesis of ‘rational expectations’
is being imposed here because agents are supposed to make the best possible
use of the limited information they have and are assumed to know the
pertinent objective probability distributions. This hypothesis is imposed by
way of adhering to the tenets of equilibrium theory.
In the preceding theory, disturbances to aggregate demand lead to a
positive correlation between unexpected changes in the aggregate price level
and revisions in aggregate output from its previously planned level. Further,
it is an easy step to show that the theory implies correlations between
revisions to aggregate output and unexpected changes in any variables that
help determine aggregate demand. In most macroeconomic models, the
money supply is one determinant of aggregate demand. The preceding
theory easily can account for positive correlations between revisions to
aggregate output and unexpected increases in the money supply.
While such a theory predicts positive correlations between the inflation
rate or money supply, on the one hand, and the level of output on the other,
it also asserts that those correlations do not depict ‘trade-offs’ that can be
exploited by a policy authority. That is, the theory predicts that there is no
way that the monetary authority can follow a systematic activist policy and
achieve a rate of output that is on average higher over the business cycle
than what would occur if it simply adopted a no-feedback, X-percent rule of
the kind Friedman [8] and Simons [32] recommended. For the theory predicts
that aggregate output is a function of current and past unexpected changes in
the money supply. Output will be high only when the money supply is and
has been higher than it had been expected to be, i.e. higher than average.
There is simply no way that on average over the whole business cycle the
money supply can be higher than average. Thus, while the preceding theory
is capable of explaining some of the correlations long thought to invalidate
classical macroeconomic theory, the theory is classical both in its adherence
to the classical theoretical postulates and in the ‘nonactivist’ flavor of its
implications for monetary policy.
Small-scale econometric models in the sense of the second section of this
chapter have been constructed which capture some of the main features of
282 Robert E.Lucas and Thomas J.Sargent
the equilibrium models described above.10 In particular, these models
incorporate the hypothesis that expectations are rational, or that all
available information is utilized by agents. To a degree, these models
achieve econometric identification by invoking restrictions in each of the
three categories (a), (b) and (c). However, a distinguishing feature of these
‘classical’ models is that they also heavily rely on an important fourth
category of identifying restrictions. This category (d) consists of a set of
restrictions that are derived from probabilistic economic theory, but play no
role in the Keynesian framework. These restrictions in general do not take
the form of zero restrictions of the type (a). Instead, the restrictions from
theory typically take the form of cross-equation restrictions among the Aj, Bj,
Cj parameters. The source of these restrictions is the implication from
economic theory that current decisions depend on agents’ forecasts of future
variables, combined with the implication that these forecasts are formed
optimally, given the behavior of past variables. These restrictions do not
have as simple a mathematical expression as simply setting a number of
parameters equal to zero, but their economic motivation is easy to
understand. Ways of utilizing these restrictions in econometric estimation
and testing are being rapidly developed.
Another key characteristic of recent work on equilibrium
macroeconometric models is that the reliance on entirely a priori
categorizations (c) of variables as strictly exogenous and endogenous has
been markedly reduced, although not entirely eliminated. This development
stems jointly from the fact that the models assign important roles to agents’
optimal forecasts of future variables, and from Christopher Sims’s
demonstration that there is a close connection between the concept of strict
econometric exogeneity and the forms of the optimal predictors for a vector
of time series. Building a model with rational expectations necessarily forces
one to consider which set of other variables helps forecast a given variable,
say income or the inflation rate. If variable y helps predict variable x, then
Sims’s theorems imply that x cannot be regarded as exogenous with respect
to y. The result of this connection between predictability and exogeneity has
been that in equilibrium macroeconometric models the distinction between
endogenous and exogenous variables has not been drawn on an entirely a
priori basis. Furthermore, special cases of the theoretical models, which
often involve side restrictions on the Rj’s not themselves drawn from
economic theory, have strong testable predictions as to exogeneity relations
among variables.
A key characteristic of equilibrium macroeconometric models is that as a
result of the restrictions across the Aj, Bj, and Cj’s, the models predict that in
general the parameters in many of the equations will change if there is a
policy intervention that takes the form of a change in one equation that
describes how some policy variable is being set. Since they ignore these
cross-equation restrictions, Keynesian models in general assume that all
other equations remain unchanged when an equation describing a policy
After Keynesian macroeconomics 283
variable is changed. Our view is that this is one important reason that
Keynesian models have broken down when there have occurred important
changes in the equations governing policy variables or exogenous variables.
Our hope is that the methods we have described will give us the capability to
predict the consequences for all of the equations of changes in the rules
governing policy variables. Having that capability is necessary before we
can claim to have a scientific basis for making quantitative statements about
macroeconomic policy.
At the present time, these new theoretical and econometric developments
have not been fully integrated, although it is clear they are very close, both
conceptually and operationally. Our preference would be to regard the best
currently existing equilibrium models as prototypes of better, future models
which will, we hope, prove of practical use in the formulation of policy. But
we should not understate the econometric success already attained by
equilibrium models. Early versions of these models have been estimated and
subjected to some stringent econometric tests by McCallum [20], Barro [3],
[4], and Sargent [27], with the result that they do seem capable of explaining
some broad features of the business cycle. New and more sophisticated
models involving more complicated cross-equation restrictions are in the
works (Sargent [29]). Work to date has already shown that equilibrium
models are capable of attaining within-sample fits about as good as those
obtained by Keynesian models, thereby making concrete the point that the
good fits of the Keynesian models provide no good reason for trusting policy
recommendations derived from them.
CRITICISM OF EQUILIBRIUM THEORY
The central idea of the equilibrium explanations of business cycles as
sketched above is that economic fluctuations arise as agents react to
unanticipated changes in variables which impinge on their decisions. It is
clear that any explanation of this general type must carry with it severe
limitations on the ability of governmental policy to offset these initiating
changes. First, governments must somehow have the ability to foresee
shocks which are invisible to private agents but at the same time lack the
ability to reveal this advance information (hence defusing the shocks).
Though it is not difficult to write down theoretical models in which these
two conditions are assumed to hold, it is difficult to imagine actual
situations in which such models would apply. Second, the governmental
countercyclical policy must itself be unanticipatable by private agents
(certainly a frequently realized condition historically) while at the same
time be systematically related to the state of the economy. Effectiveness
then rests on the inability of private agents to recognize systematic patterns
in monetary and fiscal policy.
To a large extent, criticism of equilibrium models is simply a reaction to
these implications for policy. So wide is (or was) the consensus that the task
284 Robert E.Lucas and Thomas J.Sargent
of macroeconomics is the discovery of the particular monetary and fiscal
policies which can eliminate fluctuations by reacting to private sector
instability that the assertion that this task either should not, or cannot be
performed is regarded as frivolous independently of whatever reasoning and
evidence may support it. Certainly one must have some sympathy with this
reaction: an unfounded faith in the curability of a particular ill has served
often enough as a stimulus to the finding of genuine cures. Yet to confuse a
possibly functional faith in the existence of efficacious, re-active monetary
and fiscal policies with scientific evidence that such policies are known is
clearly dangerous, and to use such faith as a criterion for judging the extent
to which particular theories ‘fit the facts’ is worse still.
There are, of course, legitimate issues involving the ability of equilibrium
theories to fit the facts of the business cycle. Indeed, this is the reason for our
insistence on the preliminary and tentative character of the particular models
we now have. Yet these tentative models share certain features which can be
regarded as essential, so it is not unreasonable to speculate as to the
likelihood that any model of this type can be successful, or to ask: what will
equilibrium business cycle theorists have in ten years if we get lucky?
Four general reasons for pessimism which have been prominently
advanced are (a) the fact that equilibrium models postulate cleared markets,
(b) the assertion that these models cannot account for ‘persistence’ (serial
correlation) of cyclical movements, (c) the fact that econometrically
implemented models are linear (in logarithms), and (d) the fact that learning
behavior has not been incorporated. We discuss each in turn in distinct
subsections.
Cleared markets
One essential feature of equilibrium models is that all markets clear, or that
all observed prices and quantities be explicable as outcomes of decisions
taken by individual firms and households. In practice, this has meant a
conventional, competitive supply-equals-demand assumption, though other
kinds of equilibrium can easily be imagined (if not so easily analyzed). If,
therefore, one takes as a basic ‘fact’ that labor markets do not clear one
arrives immediately at a contradiction between theory and fact. The facts we
actually have, however, are simply the available time series on employment
and wage rates, plus the responses to our unemployment surveys. Cleared
markets is simply a principle, not verifiable by direct observation, which
may or may not be useful in constructing successful hypotheses about the
behavior of these series. Alternative principles, such as the postulate of the
existence of a third-party auctioneer inducing wage ‘rigidity’ and non-
cleared markets, are similarly ‘unrealistic’, in the not especially important
sense of not offering a good description of observed labor market institutions.
A refinement of the unexplained postulate of an uncleared labor market
has been suggested by the indisputable fact that there exist long-term labor
After Keynesian macroeconomics 285
contracts with horizons of two or three years. Yet the length per se over
which contracts run does not bear on the issue, for we know from Arrow and
Debreu that if infinitely long-term contracts are determined so that prices
and wages are contingent on the same information that is available under
the assumption of period-by-period market clearing, then precisely the same
price-quantity process will result with the long-term contract as would occur
under period-by-period market clearing. Thus equilibrium theorizing
provides a way, probably the only way we have, to construct a model of a
long-term contract. The fact that long-term contracts exist, then, has no
implications about the applicability of equilibrium theorizing. Rather, the
real issue here is whether actual contracts can be adequately accounted for
within an equilibrium model, that is, a model in which agents are
proceeding in their own best interests. Stanley Fischer [7], Edmund Phelps
and John Taylor [26], and Robert Hall [12] have shown that some of the
‘non-activist’ conclusions of the equilibrium models are modified if one
substitutes for period-by-period market clearing the imposition of long-term
contracts drawn contingent on restricted information sets that are
exogenously imposed and that are assumed to be independent of monetary
and fiscal regimes. Economic theory leads us to predict that costs of
collecting and processing information will make it optimal for contracts to
be made contingent on a small subset of the information that could possibly
be collected at any date. But theory also suggests that the particular set of
information upon which contracts will be made contingent is not immutable
but depends on the structure of costs and benefits to collecting various kinds
of information. This structure of costs and benefits will change with every
change in the exogenous stochastic processes facing agents. This theoretical
presumption is supported by an examination of the way labor contracts differ
across high-inflation and low-inflation countries and the way they have
evolved in the United States since the mid-1950s.
So the issue here is really the same fundamental one involved in the
dispute between Keynes and the classical economists: is it adequate to regard
certain superficial characteristics of existing wage contracts as given when
analyzing the consequences of alternative monetary and fiscal regimes?
Classical economic theory denies that those characteristics can be taken as
given. To understand the implications of long-term contracts for monetary
policy, one needs a model of the way those contracts are likely to respond to
alternative monetary policy regimes. An extension of existing equilibrium
models in this direction might well lead to interesting variations, but it seems
to us unlikely that major modifications of the implications of these models
for monetary and fiscal policy will follow from this.
Persistence
A second line of criticism stems from the correct observation that if agents’
expectations are rational and if their information sets include lagged values
286 Robert E.Lucas and Thomas J.Sargent
of the variable being forecast, then agents’ forecast errors must be a serially
uncorrelated random process. That is, on average there must be no
detectable relationships between this period’s forecast error and any previous
period’s forecast error. This feature has led several critics to conclude that
equilibrium models are incapable of accounting for more than an
insignificant part of the highly serially correlated movements we observe in
real output, employment, unemployment and other series. Tobin has put the
argument succinctly in [38]:
One currently popular explanation of variations in employment is
temporary confusion of relative and absolute prices. Employers and
workers are fooled into too many jobs by unexpected inflation, but only
until they learn it affects other prices, not just the prices of what they sell.
The reverse happens temporarily when inflation falls short of expectation.
This model can scarcely explain more than transient disequilibrium in
labor markets.
So how can the faithful explain the slow cycles of unemployment we
actually observe? Only by arguing that the natural rate itself fluctuates,
that variations in unemployment rates are substantially changes in
voluntary, frictional, or structural unemployment rather than in
involuntary joblessness due to generally deficient demand.
The critics typically conclude that the theory attributes only a very minor
role to aggregate demand fluctuations and necessarily depends on
disturbances to aggregate supply to account for most of the fluctuations in
real output over the business cycle. As Modigliani [21] characterized the
implications of the theory: ‘In other words, what happened to the United
States in the 1930s was a severe attack of contagious laziness.’
This criticism is fallacious because it fails to distinguish properly between
‘sources of impulses’ and ‘propagation mechanisms’, a distinction stressed by
Ragnar Frisch in a classic 1933 paper [9] that provided many of the
technical foundations for Keynesian macroeconometric models. Even though
the new classical theory implies that the forecast errors which are the
aggregate demand ‘impulses’ are serially uncorrelated, it is certainly
logically possible that ‘propagation mechanisms’ are at work that convert
these impulses into serially correlated movements in real variables like
output and employment. Indeed, two concrete propagation mechanisms have
already been shown in detailed theoretical work to be capable of performing
precisely that function. One mechanism stems from the presence of costs to
firms of adjusting their stocks of capital and labor rapidly. The presence of
these costs is known to make it optimal for firms to spread out over time
their response to the relative price signals that they receive. In the present
context, such a mechanism causes a firm to convert the serially uncorrelated
forecast errors in predicting relative prices into serially correlated
movements in factor demands and in output.
After Keynesian macroeconomics 287
A second propagation mechanism is already present in the most classical
of economic growth models. It is known that households’ optimal
accumulation plans for claims on physical capital and other assets will
convert serially uncorrelated impulses into serially correlated demands for
the accumulation of real assets. This happens because agents typically will
want to divide any unexpected changes in the prices or income facing agents.
Thus, the demand for assets next period depends on initial stocks and on
unexpected changes in the prices or income facing agents. This dependence
makes serially uncorrelated surprises lead to serially correlated movements
in demands for physical assets. Lucas [16] showed how this propagation
mechanism readily accepts errors in forecasting aggregate demand as an
‘impulse’ source.
A third likely propagation mechanism is identified by recent work in
search theory.11 Search theory provides an explanation for why workers who
for some reason find themselves without jobs will find it rational not
necessarily to take the first job offer that comes along but instead to remain
unemployed for some period until a better offer materializes. Similarly, the
theory provides reasons that a firm may find it optimal to wait until a more
suitable job applicant appears so that vacancies will persist for some time.
Unlike the first two propagation mechanisms mentioned, consistent
theoretical models that permit that mechanism to accept errors in forecasting
aggregate demand as an impulse have not yet been worked out for mainly
technical reasons, but it seems likely that this mechanism will eventually
play an important role in a successful model of the time series behavior of
the unemployment rate.
In models where agents have imperfect information, either of the first two
and most probably the third mechanism is capable of making serially
correlated movements in real variables stem from the introduction of a
serially uncorrelated sequence of forecasting errors. Thus, theoretical and
econometric models have been constructed in which in principle the serially
uncorrelated process of forecasting errors is capable of accounting for any
proportion between zero and one of the steady-state variance of real output
or employment. The argument that such models must necessarily attribute
most of the variance in real output and employment to variations in
aggregate supply is simply wrong logically.
Linearity
Most of the econometric work implementing equilibrium models has
involved fitting statistical models that are linear in the variables (but often
highly non-linear in the parameters). This feature is subject to criticism on
the basis of the indisputable principle that there generally exist non-linear
models that provide better approximations than linear models. More
specifically, models that are linear in the variables provide no method of
detecting and analyzing systematic effects of higher than first-order moments
288 Robert E.Lucas and Thomas J.Sargent
of the shocks and the exogenous variables on the first moments of the
endogenous variables. Such systematic effects are generally present where
the endogenous variables are set by risk-averse agents.
There is no theoretical reason that most applied work has used linear
models, only compelling technical reasons given today’s computer
technology. The predominant technical requirement of econometric work
which imposes rational expectations is the ability to write down
analytical expressions giving agents’ decision rules as functions of the
parameters of their objective functions and as functions of the parameters
governing the exogenous random processes that they face. Dynamic
stochastic maximum problems with quadratic objectives, which give rise
to linear decision rules, do meet this essential requirement, which is their
virtue. Only a few other functional forms for agents’ objective functions
in dynamic stochastic optimum problems have this same necessary
analytical tractability. Computer technology in the foreseeable future
seems to require working with such a class of functions, and the class of
linear decision rules has just seemed most convenient for most purposes.
No issue of principle is involved in selecting one out of the very restricted
class of functions available to us. Theoretically, we know how to
calculate via expensive recursive methods the non-linear decision rules
that would stem from a very wide class of objective functions; no new
econometric principles would be involved in estimating their parameters,
only a much higher computer bill. Further, as Frisch and Slutsky
emphasized, linear stochastic difference equations seem a very flexible
device for studying business cycles. It is an open question whether for
explaining the central features of the business cycle there will be a big
reward to fitting non-linear models.
Stationary models and the neglect of learning
Benjamin Friedman and others have criticized rational expectations models
apparently on the grounds that much theoretical and almost all empirical
work has assumed that agents have been operating for a long time in a
stochastically stationary environment. As a consequence, typically agents are
assumed to have discovered the probability laws of the variables that they
want to forecast. As Modigliani made the argument in [21]:
At the logical level, Benjamin Friedman has called attention to the
omission from [equilibrium macroeconomic models] of an explicit
learning mechanism, and has suggested that, as a result it can only be
interpreted as a description not of short-run but of long-run equilibrium in
which no agent would wish to recontract. But then the implications of
[equilibrium macroeconomic models] are clearly far from startling, and
their policy relevance is almost nil (p. 6)
After Keynesian macroeconomics 289
But it has been only a matter of analytical convenience and not of necessity
that equilibrium models have used the assumption of stochastically
stationary ‘shocks’ and the assumption that agents have already learned the
probability distributions that they face. Both of these assumptions can be
abandoned, albeit at a cost in terms of the simplicity of the model.12 In fact,
within the framework of quadratic objective functions, in which the
‘separation principle’ applies, one can apply the ‘Kalman filtering formula’
to derive optimum linear decision with time dependent coefficients. In this
framework, the ‘Kalman filter’ permits a neat application of Bayesian
learning to updating optimal forecasting rules from period to period as new
information becomes available. The Kalman filter also permits the
derivation of optimum decision rules for an interesting class of nonstationary
exogenous processes assumed to face agents. Equilibrium theorizing in this
context thus readily leads to a model of how process non-stationarity and
Bayesian learning applied by agents to the exogenous variables leads to
time-dependent coefficients in agents’ decision rules.
While models incorporating Bayesian learning and stochastic non-
stationarity are both technically feasible and consistent with the equilibrium
modeling strategy, almost no successful applied work along these lines has
come to light. One reason is probably that non-stationary time series models
are cumbersome and come in so many varieties. Another is that the
hypothesis of Bayesian learning is vacuous until one either arbitrarily
imputes a prior distribution to agents or develops a method of estimating
parameters of the prior from time series data. Determining a prior
distribution from the data would involve estimating a number of initial
conditions and would proliferate nuisance parameters in a very unpleasant
way. It is an empirical matter whether these techniques will pay off in terms
of explaining macroeconomic time series; it is not a matter distinguishing
equilibrium from Keynesian macroeconometric models. In fact, no existing
Keynesian macroeconometric model incorporates either an economic model
of learning or an economic model in any way restricting the pattern of
coefficient nonstationarities across equations.
The macroeconometric models criticized by Friedman and Modigliani,
which assume agents have ‘caught on’ to the stationary random processes
they face, give rise to systems of linear stochastic difference equations of the
form (1), (2), and (4). As has been known for a long time, such stochastic
difference equations generate series that ‘look like’ economic time series.
Further, if viewed as structural (i.e. invariant with respect to policy
interventions) the models have some of the implications for countercyclical
policy that we have described above. Whether or not these policy
implications are correct depends on whether or not the models are structural
and not at all on whether the models can successfully be caricatured by
terms such as ‘long run’ or ‘short run’.
It is worth re-emphasizing that we do not wish our responses to these
criticisms to be mistaken for a claim that existing equilibrium models can
290 Robert E.Lucas and Thomas J.Sargent
satisfactorily account for all the main features of the observed business cycle.
Rather, we have argued that no sound reasons have yet been advanced which
even suggest that these models are, as a class, incapable of providing a
satisfactory business cycle theory.
SUMMARY AND CONCLUSIONS
Let us attempt to set out in compact form the main arguments advanced in
this chapter. We will then comment briefly on the main implications of these
arguments for the way we can usefully think about economic policy.
First, and most important, existing Keynesian macroeconometric models
are incapable of providing reliable guidance in formulating monetary, fiscal
and other types of policy. This conclusion is based in part on the spectacular
recent failures of these models, and in part on their lack of a sound
theoretical or econometric basis. Second, on the latter ground, there is no
hope that minor or even major modification of these models will lead to
significant improvement in their reliability.
Third, equilibrium models can be formulated which are free of these
difficulties and which offer a different set of principles which can be used to
identify structural econometric models. The key elements of these models are
that agents are rational, reacting to policy changes in a manner which is in
their best interests privately, and that the impulses which trigger business
fluctuations are mainly unanticipated shocks.
Fourth, equilibrium models already developed account for the main
qualitative features of the business cycle. These models are being subjected
to continued criticism, especially by those engaged in developing them, but
arguments to the effect that equilibrium theories are, in principle, incapable
of accounting for a substantial part of observed fluctuations appear due
mainly to simple misunderstandings.
The policy implications of equilibrium theories are sometimes
caricatured, by friendly as well as unfriendly commentators, as the assertion
that ‘economic policy does not matter’ or ‘has no effect’.13 This implication
would certainly startle neoclassical economists who have successfully
applied equilibrium theory to the study of innumerable problems involving
important effects of fiscal policies on resource allocation and income
distribution. Our intent is not to reject these accomplishments, but rather to
try to imitate them, or to extend the equilibrium methods which have been
applied to many economic problems to cover a phenomenon which has so far
resisted their application: the business cycle.
Should this intellectual arbitrage prove successful, it will suggest
important changes in the way we think about policy. Most fundamentally, it
directs attention to the necessity of thinking of policy as the choice of stable
‘rules of the game’, well understood by economic agents. Only in such a
setting will economic theory help us to predict the actions agents will choose
to take. Second, this approach suggests that policies which affect behavior
After Keynesian macroeconomics 291
mainly because their consequences cannot be correctly diagnosed, such as
monetary instability and deficit financing, have the capacity only to disrupt.
The deliberate provision of misinformation cannot be used in a systematic
way to improve the economic environment.
The objectives of equilibrium business cycle theory are taken, without
modification, from the goal which motivated the construction of the
Keynesian macroeconometric models: to provide a scientifically based means
of assessing, quantitatively, the likely effects of alternative economic
policies. Without the econometric successes achieved by the Keynesian
models, this goal would be simply inconceivable. Unless the now evident
limits of these models are also frankly acknowledged, and radically different
new directions taken, the real accomplishments of the Keynesian Revolution
will be lost as surely as those we now know to be illusory.
ACKNOWLEDGEMENTS
We wish to acknowledge the benefit of criticism of an earlier draft by
William Poole and Benjamin Friedman.
NOTES
1 Linearity is a matter of convenience, not of principle. See pp. 287–88.
2 A ‘regression equation’ is an equation to which the application of ordinary least
squares will yield consistent estimates.
3 In these expressions for P
s
and Q
s
, take matrices not previously defined (for example,
any with negative subscripts) to be zero.
4 These three categories certainly do not exhaust the set of possible identifying
restrictions, but in Keynesian macroeconometric models most identifying restrictions
fall into one of these three categories. Other possible sorts of identifying restrictions
include, for example, a priori knowledge about components of Σ, and cross-
equation restrictions across elements of the A
j
,B
j
, and C
j
’s. Neither of these latter
kinds of restrictions is extensively used in Keynesian macroeconometrics.
5 Criticisms of the Keynesian solutions of the identification problem along much the
following lines have been made in Lucas [17], Sims [33], and Sargent and Sims [31].
6 [This note was added in revision, in part in response to Benjamin Friedman’s
comments.] Much of this work was done by economists operating well within the
Keynesian tradition, often within the context of some Keynesian macroeconometric
model. Sometimes a theory with optimizing agents was resorted to in order to
resolve empirical paradoxes by finding variables that had been omitted from some
of the earlier Keynesian econometric formulations. The works of Modigliani and
Friedman on consumption are good examples of this line of work, a line whose
econometric implications have been extended in important work by Robert Merton.
The works of Tobin and Baumol on portfolio balance and of Jorgenson on
investment are also in the tradition of applying optimizing microeconomic theories
for generating macroeconomic behavior relations. Since the late 1940s, Keynesian
econometric models have to a large extent developed along the line of trying to
model agents’ behavior as stemming from more and more sophisticated optimum
292 Robert E.Lucas and Thomas J.Sargent
problems. Our point here is certainly not to assert that Keynesian economists have
completely foregone any use of optimizing microeconomic theory as a guide. Rather,
it is that, especially when explicitly stochastic and dynamic problems have been
studied, it has become increasingly apparent that microeconomic theory has very
damaging implications for the restrictions conventionally used to identify Keynesian
macroeconometric models. Furthermore, as Tobin [37] emphasized long ago, there
is a point beyond which Keynesian models must suspend the hypothesis either of
cleared markets or of optimizing agents if they are to possess the operating
characteristics and policy implications that are the hallmarks of Keynesian
economics.
7 For example, see the comment by Albert Ando [35, esp. pp. 209–10], and the
remarks of L.R.Klein [24].
8 For an example that illustrates the emptiness at a general level of the statement that
‘employers are always operating along dynamic stochastic demands for factors’,
see the remarks on econometric identification in Sargent [29]. In applied problems
that involve modeling agents’ optimum decision rules, one is impressed at how
generalizing the specification of agents’ objective functions in plausible ways quickly
leads to econometric under-identification. A somewhat different class of examples
is seen in the difficulties in using time series observations to refute the view that
‘agents only respond to unexpected changes in the money supply’. A distinguishing
feature of the equilibrium macroeconometric models described below is that
predictable changes in the money supply do not affect real GNP or total employment.
In Keynesian models, predictable changes in the money supply do cause real GNP
and employment to move. At a general level, it is impossible to discriminate between
these two views by observing time series drawn from an economy described by a
stationary vector random process (Sargent [28]).
9 See Edmund S.Phelps et al [25] and Lucas [15], [16].
10 For example, Sargent [27]. Dissatisfaction with the Keynesian methods of achieving
identification has also led to other lines of macroeconometric work. One line is the
‘index models’ described by Sargent and Sims [31] and Geweke [10]. These models
amount to a statistically precise way of implementing Wesley Mitchell’s notion that
there is a small number of common influences that explain the covariation of a large
number of economic aggregates over the business cycle. This ‘low dimensionality’
hypothesis is a potential device for restricting the number of parameters to be
estimated in vector time series models. This line of work is not entirely a-theoretical
(but see the comments of Ando and Klein in Sims [35]), though it is distinctly
unKeynesian. As it happens, certain equilibrium models of the business cycle do
seem to lead to low dimensional index models with an interesting pattern of variables’
loadings on indexes. In general, modern Keynesian models do not so easily assume
a low-index form. See the discussion in Sargent and Sims [31].
11 For example [19], [22] and [18].
12 For example, see Crawford [5] and Grossman [11].
13 A main source of this belief is probably Sargent and Wallace [30], in which it was
shown that in the context of a fairly standard macroeconomic model, but with
agents’ expectations assumed rational, the choice of a reactive monetary rule is of
no consequence for the behavior of real variables. The point of this example was to
show that within precisely that model used to rationalize reactive monetary policies,
such policies could be shown to be of no value. It hardly follows that all policy is
ineffective in all contexts.
After Keynesian macroeconomics 293
REFERENCES
1 Ando, Albert, ‘A Comment’, in [35].
2 Arrow, Kenneth J., ‘The Role of Securities in the Optimal Allocation of Risk Bearing’,
Review of Economic Studies 31 (1964): 91–6.
3 Barro, Robert J., ‘Unanticipated Money Growth and Unemployment in the United
States’, American Economic Review 67 (1977): 101–15.
4 ——‘Unanticipated Money, Output and the Price Level in the United States’, Journal
of Political Economy 86 (1978): 549–80.
5 Crawford, Robert, ‘Implications of Learning for Economic Models of Uncertainty’,
Manuscript. Pittsburgh: Carnegie-Mellon University, 1971.
6 Debreu, Gerard, The Theory of Value, New York: Wiley, 1959.
7 Fischer, Stanley, ‘Long-term Contracts, Rational Expectations, and the Optimal
Money Supply Rule’, Journal of Political Economy 85 (1977): 191–206.
8 Friedman, Milton, ‘A Monetary and Fiscal Framework for Economic Stability’,
American Economic Review 38 (1948): 245–64.
9 Frisch, Ragnar, ‘Propagation Problems and Impulse Problems in Dynamic
Economies’, reprinted in AEA Readings in Business Cycles, edited by R.A. Gordon
and L.R.Klein, vol. X, 1965.
10 Geweke, John, ‘The Dynamic Factor Analysis of Economic Time Series’, in Latent
Variables in Socio-Economic Models, edited by D.Aigner and A.Goldberger, pp.
365–83. Amsterdam: North Holland, 1977.
11 Grossman, Sanford, ‘Rational Expectations and the Econometric Modeling of
Markets Subject to Uncertainty: A Bayesian Approach’, Journal of Econometrics
3 (1975): 255–72.
12 Hall, Robert E., ‘The Macroeconomic Impact of Changes in Income Taxes in the
Short and Medium Runs’, Journal of Political Economy 86 (1978): S71-S86.
13 Keynes, J.M., The General Theory of Employment, Interest, and Money, London:
Macmillan, 1936.
14 Leontief, W., ‘Postulates: Keynes’ General Theory and the Classicists’, in The New
Economics, Keynes’ Influences on Theory and Public Policy, edited by S. Harris,
Clifton, NJ: Augustus Kelley, 1965.
15 Lucas, R.E. Jr, ‘Expectations and the Neutrality of Money’, Journal of Economic
Theory 4, no. 2 (April 1972): 102–23.
16 ——‘An Equilibrium Model of the Business Cycle’, Journal of Political Economy
83, no. 6 (December 1975): 1113–4.
17 ——‘Econometric Policy Evaluation: A Critique’, in The Phillips Curve and Labor
Markets, edited by K.Brunner and A.H.Meltzer, Carnegie-Rochester Conference
Series on Public Policy, 1:19–46, Amsterdam: North Holland, 1976.
18 Lucas, R.E. Jr., and Prescott, Edward C. ‘Equilibrium Search and Unemployment’,
Journal of Economic Theory 7 (1974): 188–209.
19 McCall, John, ‘The Economics of Information and Optimal Stopping Rules’,
Journal of Business 38 (1965): 300–17.
20 McCallum, Bennett, ‘Rational Expectations and the Natural Rate Hypothesis:
Some Consistent Estimates’, Econometrica 44 (1976): 43–52.
21 Modigliani, Franco, ‘The Monetarist Controversy, or Should We Forsake
Stabilization Policies?’, American Economic Review (March 1977): 1–19.
22 Mortensen, Dale T., ‘A Theory of Wage and Employment Dynamics’, in [25].
23 Muench, T., Rolnick, A., Wallace, N., and Weiler, W. ‘Tests for Structural Change
and Prediction Intervals for the Reduced Forms of Two Structural Models of the
U.S.: The FRB-MIT and Michigan Quarterly Models’, Annals of Economic and
Social Measurement 313 (1974).
24 Okun, Arthur, and Perry, George L. (eds) Brookings Papers on Economic Activity,
1973, vol. 3, remarks attributed to Lawrence Klein, p. 644.
294 Robert E.Lucas and Thomas J.Sargent
25 Phelps, E.S. et al.,Microeconomic Foundations of Employment and Inflation Theory,
New York: Norton, 1970.
26 Phelps, E.S. and Taylor, John, ‘Stabilizing Powers of Monetary Policy under Rational
Expectations’, Journal of Political Economy 85 (1977): 163–90.
27 Sargent, T.J. ‘A Classical Macroeconometric Model for the United States’, Journal
of Political Economy (1976).
28 ——‘The Observational Equivalence of Natural and Unnatural Rate Theories of
Macroeconomics’, Journal of Political Economy (June 1976).
29 ——‘Estimation of Dynamic Labor Demand Schedules Under Rational
Expectations’, Journal of Political Economy (December 1978).
30 Sargent, J.J. and Wallace, Neil, ‘“Rational” Expectations, the Optimal Monetary
Instrument, and the Optimal Money Supply Rule’, Journal of Political Economy
83 (1975): 241–54.
31 Sargent, T.J., and Sims, C.A. ‘Business Cycle Modeling Without Pretending to
Have Too Much A Priori Economic Theory’, in New Methods in Business Cycle
Research, edited by C.Sims, Federal Reserve Bank of Minneapolis, 1977.
32 Simons, Henry C., ‘Rules Versus Authorities in Monetary Policy’, Journal of Political
Economy 44 (1936): 1–30.
33 Sims, C.A., ‘Macroeconomics and Reality’, Econometrica 48 (1980): 1–48.
34 ——‘Money, Income, and Causality’, American Economic Review (September 1972).
35 ——(ed.) New Methods in Business Cycle Research: Proceedings from a
Conference, Federal Reserve Bank of Minneapolis, October, 1977.
36 Sonnenschein, Hugo, ‘Do Walras’ Identity and Continuity Characterize the Class
of Community Excess Demand Functions?’ , Journal of Economic Theory 6 (1973):
345–54.
37 Tobin, James, ‘Money Wage Rates and Employment’, in The New Economies,
edited by S.Harris, Clifton, NJ: Augustus Kelley, 1965.
38 ——‘How Dead is Keynes?’, Economic Inquiry (October 1977) 15:459–68.
12 A child’s guide to rational expectations
Rodn Maddock and Michael Carter
Journal of Economic Literature (1982) 20, March, pp. 39–51
DRAMATIS PERSONAE
(In order of speaking)
Ernie, first student, is something of a Keynesian.
Bert, second student, is more inclined to monetarism.
Scene I Prologue
Scene II The idea of rational expectations
Deriving the impotence results
Criticisms
Scene III Testing
Significance
Conclusion
Appendix A Aggregate supply
Appendix B Algebra of the model
References
SCENE I
Prologue
(Two students sharing coffee in the union of an Australian university.)
Ernie: Did you read that ridiculous article in Challenge the other day?
Bert: Which?
Ernie: Somebody named Bennett McCallum was saying that rational
expectations proved that the government could not stabilize the economy.
Hang on, I’ve got it here: ‘An accurate understanding of how expectations
are formed leads to the conclusion that short-run stabilization policies are
untenable’ (McCallum 1980:37). I don’t know how they could develop
theories like that. It’s pretty obvious that government policy does affect the
economy in the short run.
Bert: He didn’t say they could not affect the economy in the short run or
even in the long run. The key word is stabilize.1 Just think about what’s
happened in the last few years—record inflation and record unemployment.
You don’t call that stabilization, do you?
296 Rodney Maddock and Michael Carter
Ernie: Well, maybe they’ve been stable at high levels but I take your point.
There does seem to have been some breakdown of the ways in which the
government can influence the macroeconomy. Do these rational expectations
blokes think they have a model to explain stagflation?
Bert: Yes, they do. It’s caused by misguided governments following Keynesian
policies that haven’t worked, don’t work, and won’t work in the future.2
Ernie: I suppose they advocate doing nothing and letting the ‘free market’ do
its worst. Great! They sound just like Friedman and all those old-fashioned
monetarists. They have always said inflation was just a monetary
phenomenon and macro policy couldn’t shift the economy to higher levels of
employment.
Bert: Yes, that’s right. Most economists now agree that the long-run Phillips
curve is vertical. 3 That means that there exists a natural rate of
unemployment.4 Government policy can bring about a departure from that
only in the short run and then only by fooling people. But you can’t fool all
the people all the time. Therefore, systematic policy is ineffective.5
Ernie: I’m not at all sure that the long-run Phillips curve is vertical.6 We
used to have about 1 per cent unemployment; now we seem to be stuck at
about 8 per cent. How can you explain that with a vertical Phillips curve.
‘The Phillips curve is vertical but moves around a lot’—hardly seems much
of a theory.7 Even if it is vertical and we can’t get away from it except by
fooling people, clearly the government can fool people. Every time it
changes policy the people don’t know about the new policy for a while so it
takes time before they catch up.8
Bert: But that’s just what rational expectations is all about! It suggests that
people anticipate the effects of the new policy. If that’s true, then the policies
won’t cause any increase in employment!
Ernie: How on earth are people supposed to anticipate the effects of policy?
I just can’t see it. Have they all got econometric models under the sink?9
Bert: (Angrily) Now you’re just being silly. Have you read any of the basic
literature—Lucas, Sargent, Wallace, and so on?
Ernie: I’ve looked at some but it just seems unreal—too many equations. They
never define exactly how they think anybody forms these ‘rational expectations’.
Bert: Look, I’ve got to go to my macro lecture. How about we meet again
tomorrow, and I’ll introduce you to the magnificent world of rational
expectations. Same time?
SCENE II
The idea of rational expectations
(In the same place, next day.)
Bert: Well, are you ready to try to understand what rational expectations is
about?
A child’s guide to rational expectations 297
Ernie: Yes. Have you got it figured out yet?
Bert: I’ve been thinking about it. Let’s go through it systematically. First, we can
talk about just what rational expectations are. Then we can look at the way the
policy impotence result is derived. By then we should have a pretty clear idea of
what this line of research is all about so we can try to figure out how it relates
to the Phillips curve, monetarism, econometric models, and all that. OK?
Ernie: Alright. What’s the definition of rational expectations? What on earth
might irrational expectations be?
Bert: First things first. Let’s start with familiar ground. What would you say
is the basic behavioral assumption of economic behavior?
Ernie: Utility maximization, I suppose.
Bert: More or less. I would say that the basic assumption about individual
behavior is that economic agents do the best they can with what they have.
This principle forms the basis of consumption theory, production theory,
human capital theory and so on.
Ernie: So it’s the basis of microeconomics.10 But what’s that got to do with
expectations?
Bert: Everything. At its most fundamental, rational expectations theorists
argue that the same principle should be applied to the formation of
expectations. If you want a definition, how about: rational expectations is
the application of the principle of rational behavior to the acquisition and
processing of information and to the formation of expectations.11
Ernie: Am I to infer that my utility function and I sit down together and
rationally decide how much information I should acquire in order to form
the expectations that will help me maximize my utility? Incredible!
Bert: Yes, you can attack it that way if you like, but that’s a more general
criticism of utility theory which we can argue about some other time. All I’m
saying here is that if one considers economic agents to be rational
maximizers, then it’s consistent12 to consider information gathering and
expectation formation as determined by the same procedure.
Ernie: OK. So you’d insist upon a rational expectations postulate that
private economic agents gather and use information efficiently. That means
you believe the marginal costs of gathering and using information are
equated to their marginal benefits. McCallum doesn’t agree with you. He
says: ‘Individual agents use all available and relevant information’13 and it
seems to me that Sargent and Lucas say the same. It almost seems as if they
think information is a free good.14
Bert: That’s a good point. Many theorists have ignored the costs of
information used in the formation of expectations. That is one of my
criticisms of the literature. But I think it is useful to distinguish between
rational expectations as a principle of informational efficiency and rational
expectations as it appears in some of the macroeconomic literature.15
Ernie: The term ‘rational’ is quite confusing in the context and you are right
that the distinction between the two things is important. But what difference
does rational expectations make to individuals? Can you give examples?
298 Rodney Maddock and Michael Carter
Bert: The example most often used in the literature involves the allocation of
time between labor and leisure.16 In deciding how many hours to work this
period, an individual must take account of expected future wages and not
just the present wage. For example, if you expect the real wage to be $10 per
hour this week, and $1 next week, then it makes sense to work as much as
possible this week, and have some time off next week. Therefore the number
of hours worked in any period, that is, the labor supply, will depend not only
on the current real wage but on expected future real wages. A rational
expectation of real wages will take into account all available information,
including the effects of government policy.
Ernie: But my old man works 40 hours every week—he doesn’t have much
choice.
Bert: But your old man’s boss does. When he is deciding whether to hire
more people or lay them off, he needs to take into account future prices and
wages. His expectations should be based on all the available information.
This includes, among other things, the impact of future government policy.
Ernie: OK. I see how the level of employment might depend upon
expectations and how ‘good’ expectations are better than ‘bad’ ones. But I
can’t see why that means that there is no room for government policy.
Deriving the impotence result
Bert: Without realizing it you’ve just made a very important distinction. The
relationship between the level of employment and expectations is logically
quite separate from beliefs about how expectations are formed. The
conclusion that there is no scope for government policy—the impotence
result—depends crucially upon imposing a special assumption about
expectations—rational expectations—upon a special type of macroeconomic
model.
Ernie: Well I think I understand the meaning of rational expectations. What
types of macro models do rational expectations theorists use?
Bert: (Drawing a diagram—Figure 12.1.) Most of them work with the idea
that the levels of output and prices are determined by the intersection of an
aggregate demand and aggregate supply function. The aggregate supply
curve is taken to be vertical, so that output cannot deviate from Y
n
as a
direct result of any change in the level of demand. Thus government policies
designed to change the level of aggregate demand are not likely to be
effective. The level Y
n
is the output associated with equilibrium in the labor
market at the natural rate of unemployment so we can call Y
n
the natural
rate of output or income for the economy.17
Consider the possibility that the government takes action that may, at first
blush, be supposed to increase output. For example, let it act to increase
nominal income and aggregate money demand. Money wage rates will tend
to rise, and if workers regard this as equivalent to an increase in real wages,
employment will increase and output will temporarily rise to a level higher
A child’s guide to rational expectations 299
than Yn. But if production is carried on subject to diminishing returns to
labor, prices will rise relative to nominal wages, and real wages will fall.
When workers realize this, employment will fall back to its original position,
and output will return to Yn. At this point, nominal wage rates and prices are
higher (the nominal demand curve crosses the vertical supply curve at a
higher level), but output and employment are back where they started. Since
the aggregate supply curve had not shifted, the possibility of increasing
employment and output arises only as long as people confuse nominal
changes in wages (for example) with real changes. This means that
government policy will increase the level of income in real terms only if it is
able to fool people into confusing nominal changes with real ones.18
Ernie: That’s ridiculous! The ‘natural’ rate of unemployment depends
intimately upon all sorts of government policies—for example tax laws,
minimum wage laws, immigration policy, school-leaving age, etc., etc. Do
you really mean that the government can’t change aggregate supply by
increasing the investment allowance? Or by going to war, for that matter?
Bert: You’re right, you’re right! I should have been more careful. Clearly
government policy can alter the natural rate of unemployment or, if you like,
the position of the aggregate curve. What I should have said is that the only
way in which government policy can bring about deviations from the natural
rate of unemployment is by inducing private agents to have mistaken
expectations. Let’s write down a simple model.
(Bert’s scribbling is attached as Appendix B. For those who like
mathematical descriptions it should make the discussion clearer but is not a
necessary adjunct.)
Ernie: That makes your position clearer. The actual aggregate supply
function implies that deviations of actual output from the natural rate are
directly proportional to deviations of actual prices from expected prices.
Figure 12.1
300 Rodney Maddock and Michael Carter
Since people with rational expectations never make mistakes about policy
rules, policy will never fool them, and output will never deviate from its
natural rate as a result of any policy rule.
Bert: That’s the idea but you’ve put it too strongly. If government policies
are random, they will be effective although not necessarily desirable. It’s the
systematic component of policy that the theory suggests will be ineffective.
Ernie: I’m not too sure about the neutrality-of-money proposition19 generally
but will let it ride for now. You explain the rest of the argument—then I’ll
put my objections one by one.
Bert: Now the point of rational expectations is that people won’t be
surprised by any systematic policy. Any government that relies upon a policy
rule—one that has a fixed growth of the money supply, or one systematically
related to income or unemployment—will never cause any deviation from
the natural rate.20 A random policy will affect real output. But any policy
rule that is systematically related to economic conditions, for example one
designed with stabilization in mind, will be perfectly anticipated, and
therefore have no effect on output or employment. In other words, to have
real effects, monetary policy must be completely unpredictable. Any
systematic policy will be impotent.21
Ernie: Can we put it this way? Rational expectations are based on all
available information. The available information set includes the
government policy rule. Therefore, a rational expectation of inflation, for
example, will include the anticipated effects of government policy so that
the policy will have no effect on output.22
Bert: Yes, that pretty well sums it up.
Criticisms
Ernie: Now that I think I understand what you’re on about, can I tell you
what I think is wrong with the model?
Bert: OK.
Ernie: First, I don’t like your supply curve. There are lots of criticisms one
could make, but the most important in the context of the model is that you
assume an extreme form of the neutrality of money.23 Perfectly anticipated
inflation has no real effects in your model. That’s clearly wrong. Buiter
(1980) put the standard argument in terms of the portfolio readjustments
required because inflation changes the real rate of return on those financial
assets which have a zero nominal return. Personally, I think the distortions
introduced by the progressive tax structure in an inflationary situation are
far more important empirically.
Bert: Yes, but all models are approximations.
Ernie: True, but not all approximations are good approximations! Here’s
another problem. If expectations are rational, then expectation errors should
be randomly distributed over time. A straightforward implication of that for
this model is that the level of output (or unemployment) is uncorrelated over
A child’s guide to rational expectations 301
time. Yet everybody knows that the GNP and unemployment series have a
high degree of serial correlation.24 We tend to go through a series of years in
which unemployment is below the ‘natural rate’, and then a series of years in
which it is above the ‘natural rate’. It doesn’t seem to be distributed very
randomly. Compare the 1960s and 1970s in Australia—it’s the old story of
business cycle expansion and contraction.
Bert: I can’t deny the serial correlation in the unemployment or
income series. Most rational expectations models include lagged
income or lagged unemployment as explanatory variables in the
supply function.25 This does make the models fit the data better, but
there is no good theoretical justification for it. Lucas is the only one I
know who really addresses the issue.26 He relies on the well-known
‘fact’ that all people live on islands. At the end of each trading period
people choose a new island at random. Since they don’t know the
history of their new island, they can’t distinguish immediately between
real and nominal effects.27
Ernie: These island models seem appropriate to a society in which the fastest
form of communication is a floating coconut. Hasn’t Lucas ever heard of
radio and the telephone?28
Bert: I have to agree with you. I said that the explanations for persistence
weren’t very convincing, especially when the government regularly publishes
lots of statistics. The newspapers carry stock exchange prices every day. The
information seems to be essentially free.
There is another line of argument though. If prices change suddenly, firms
can increase their production less than their sales by using up some of their
inventories. If there is no price shock in the next period production would
then be raised to build inventories back to their original level. Thus there
would be an increase in production to meet the original stock and for as long
thereafter as the restocking took.
Ernie: But that implies there should be a strong relationship between
inventory cycles and output cycles and that’s not really true, is it?
Bert: Well, the relation is far from perfect. I was really just suggesting that
in an economy characterized by durable goods it shouldn’t be too difficult to
accept that adjustments of various sorts will have effects that persist.29 We
really don’t have a good explanation for persistence (serial correlation). I
willingly concede that point. What’s next?
Ernie: OK. Even if all that information is freely available, you assume that
all the agents know the correct model of the economy. How…
Bert: No, I don’t. Well, not me really. I mean that rational-expectations
people don’t necessarily say that everybody knows the correct model of the
economy. They suggest that some arbitrage process takes place whereby the
people who have the correct model dominate the outcome.30 If there are
misapprehensions, then well-informed agents can make profits at the expense
of the ill-informed. This will inevitably lead the system to converge to the
302 Rodney Maddock and Michael Carter
rational expectations equilibrium. As your old mate, John Maynard Keynes,
said:
actions based on inaccurate anticipations will not long survive
experiences of a contrary character, so that the facts will soon override
anticipation except where they agree.
(Keynes 1930:160)
Ernie: Granted there is a role for arbitrage. But how do we know that
expectations of the experts will converge on the true value? Give me any
rational expectations model and I think I can show you a reasonable
adjustment process that will not converge to the rational expectations
equilibrium. Bert: And I can probably show you one that can. Unless the
theorists specify an adjustment mechanism, we can’t really argue about this
point. Rational expectations theorists haven’t addressed this problem.31
Ernie: That’s a big gap in your theory. But let me read to you what Robert
Shiller says:
Even if a model does eventually converge on a rational expectations
equilibrium, it may take such a long time to do so that, since the structure
of the economy changes occasionally, the economy is never close to a
rational expectations equilibrium.
(Shiller 1978:39)
To recalculate a quarterly econometric model after a change in policy rule
might take 20 quarters. To estimate the effect of policy based on the new
estimates might take another 20 quarters. Thus, even if the process
converges, each stage in the convergence to the new equilibrium could take
five years—by which time we may all be dead!
Bert: But if the Government’s objective is to stabilize the economy, then it
wants to make sure that private expectations are rational, and hence it will
inform the public of any new policy rule.32
Ernie: That doesn’t get you out of hot water. First, the learning problem
doesn’t concern the policy rule alone. Agents also have to learn the structure
of economy, which is subject to change. Econometric modellers don’t have
an outstanding record of success, do they? Second, why do you assume that
the government’s objective is to stabilize the economy? It seems to me that
the government’s real objective is to remain in power—you know, the
political business cycle idea.33 And if that is their objective it may be in their
interest to hide information and fool the public. If that is the case the voters
can hardly be expected to believe the signals the government is sending out
and the whole macroeconomic process degenerates into a guessing game.
Bert: But you must agree that most macroeconomics does assume the objective
of stabilization. This literature falls into that tradition which is concerned with
government policy rules designed to achieve macroeconomic stabilization.34
Ernie: Let’s go and get a beer.35
A child’s guide to rational expectations 303
SCENE III
(In the union bar)
Testing
Bert: Now the testing is a bit tricky. It’s a pretty young research program
and there are no well-accepted testing procedures as yet. The principal
difficulty is that we are really testing a joint hypothesis—the economic
model and the expectations mechanism. That makes it difficult to decide just
where the responsibility for failures of tests really lie.
Ernie: But why can’t you just test the expectations mechanism directly? Ask
people what they expect, and see if they are right?
Bert: Stephen Turnovsky (1970) and James Pesando (1975) and a couple of
other people have done that, but the results have been inconclusive.
Economists traditionally don’t like surveys, anyway.
Ernie: Well, how have rational expectations protagonists tried to test the
theory?
Bert: Basically they have taken two different approaches. Have a look at the
supply function again. The natural rate hypothesis will allow non-random
deviations only if there are expectations errors of some sort. Under the
rational expectations hypothesis there are no expectations errors or at most,
only random ones. Thus, deviations from the natural rate of output must be
random.36 In particular, deviations cannot be systematically related to any
other explanatory variable, for example the (lagged) money supply or the
wage rate. And so the first type of test is essentially to see whether deviations
from the natural rate are systematically related to any other variables. This
test has been applied in a number of different ways especially by Sargent
(1973; 1976). In those papers the joint hypothesis—natural rate and rational
expectations—was rejected in a number of cases, jointly rejected because
they were jointly tested. With slightly different specifications, they weren’t.
Sargent concludes that rational expectations is ‘not obscenely at variance
with the data.’37
Ernie: Remarkable resilience, eh!
Bert: Yes, what’s more, his next paper was entitled The Observational
E q u i v a l e n c e o f N a t u r a l a n d U n n a t u r a l R a t e T h e o r i e s o f
Macroeconomics.’38 This initiated the second approach to testing rational
expectations. It was based on the idea that what rational expectations
models add, compared with other models, is that price expectations take
into account the policy rule the government is using. If that rule does not
change, ordinary models and rational expectations models will fit any data
set equally well, though probably with different parameters. What it means
is that you can only distinguish between rational expectations macro
models and ordinary ones if the policy rule has changed.39 It seems a
reasonable idea to me.
304 Rodney Maddock and Michael Carter
Ernie: Maybe.
Bert: Well, anyway, it gave Sargent and Salih Neftci (1978) an idea for a
new type of test. First, they estimated the government’s policy rule by
regressing the money supply on past levels of income. They then analysed
the results to see if there had been any significant changes in the relationship,
that is, to see if the policy rule had changed.40 They found changes in 1929
and 1964. They then looked at some ordinary macroeconomic models to see
whether the parameters had changed at about the same time that the policy
rule changed. In each case they found some evidence that it had.41
Ernie: Well, that seems like a reasonable sort of approach. Really it depends
on the rational expectations idea—that people change their behavior as
policy changes—rather than on the natural rate proposition. I’m inclined to
agree that people take into account what the government is trying to do
when they plan for the future but the extreme form of the natural rate of
unemployment is an over-simplification of a complex world. Of course, the
test is a bit tricky. It’s not really clear to me that that’s the way to estimate
the government policy rule and then, even if it is, there are probably other
non-rational expectations mechanisms which would suggest a change in the
econometric structure as a result of a change in policy rule. There’s no real
alternative hypothesis involved in the test!
Bert: Everyone agrees it’s not a very strong test; it’s titled ‘A Little Bit of
Evidence…’, but it is suggestive and does focus upon the relation between
expectations and the policy rule and not upon the natural rate. It gets away
from the dogmatic form of the natural rate, impotence of policy area, and
focuses upon the positive contribution of the rational expectations idea.
Significance
Ernie: OK, we’ve covered the model and the evidence, such as it is. What’s
all the fuss about?
Bert: Well, it really nails the Phillips curve. Much post-war stabilization
policy has been based on the idea that there is a trade-off between
unemployment and inflation that the government can exploit by
influencing aggregate demand—the so-called Phillips curve. Friedman
largely undermined that with the natural rate idea. He said that policy
worked only by fooling people and that in the long-run they could not
be fooled. This still left the way open for effective short-run policy. If
people have rational expectations they won’t be fooled. If people have
rational expectations they won’t be fooled by systematic policy even in
the short run so there is no scope for short-run policy either. This
explains why the Phillips curve became unstable the moment policy-
makers tried to exploit it.
Ernie: But the explanation depends really heavily, as you’ve already agreed,
upon the particular macroeconomic model you have set up.42 The
A child’s guide to rational expectations 305
monetarists have grabbed this idea43 to support their traditional position that
active government policy is not desirable.44 It’s largely been their baby until
now. The reason is not hard to see. The impotence result strongly supports
their ideological position that the government should keep its hands off the
economy.45
Bert: That’s rather harsh. Monetarists are no more ideological than other
economists. What would be your ‘un-ideological’ view of rational
expectations?
Ernie: I’ve been wondering how a future historian of thought might assess
it. I think rational expectations theory will be seen as a very important
development in economics, but not because of the impotence result.
Rational expectations are important in any situation in which market
behavior is influenced by expectations. Take the case of an aggregate
demand deficiency in a Keynesian model. The usual argument is that
monetary expansion will work through affecting interest rates so that
gradually the economy shifts to a higher output level. With rational
expectations the shift to the new level would be extremely rapid. If business
people understand the economic implications of expansionary government
policy, they can expand their output in anticipation of those effects rather
than waiting for the rise in demand to be obvious in the market. In that
case, far from policy being impotent, rational expectations may make
policy more effective.
This example, by the way, illustrates the fact that most of the rational
expectations literature has a particular economic model built in, one in
which all markets clear instantaneously; unemployment is, therefore,
voluntary, hence ‘natural’; and money is necessarily neutral. But if that
model is not applicable, policy need not be impotent, and, as said, rational
expectations may make it more rapidly effective.
What’s more rational expectations can be applied in microeconomic
situations.46 Cobweb models of dynamic behavior in commodity markets
depend upon ‘irrationality of expectations’—the idea that next year’s prices
will be the same as this year’s or even a simple extrapolation of it. Clearly,
producers can do better than that, and if they have rational expectations the
market is likely to approach its equilibrium quite quickly.
Bert: You’re right! The idea can be applied in a wide variety of models but
what about your general conclusions about policy formation?
Ernie: Yes, I agree that there are important implications for policy design.
Private economic agents are intelligent decision-makers and can be expected
to take the effects of government policy changes into account in deciding
their behavior. This means that the policy-maker must anticipate the effect of
policy on private expectations and the consequent changes in behavior. In
practical terms it means that we need to know a lot more about the
availability and use of information by private decision-makers.47 Thus, the
focus of the theory of policy should be on expectations and information and
on their role in determining behavior.
306 Rodney Maddock and Michael Carter
Bert: I think you can go further. People used to think that the only reason
stabilization policy didn’t work well was that policy-makers didn’t have
enough knowledge about the structure of the economy. Rational expectations
has taught us that the problem may not be just one of the absolute
knowledge of the authorities but rather of how much more or less they know
than the public does—a problem of relative knowledge. If this is true, the
problem will always be with us.
Conclusion
Ernie: Well, I started off inclined towards Robert A.Gordon’s view that
rational expectations is an example of a recent development in economics ‘in
which theory proceeds with impeccable logic from unrealistic assumptions to
conclusions that contradict the historical record’ (1976:5). But now I see that
that’s a bit too harsh.
Most of the research on rational expectations has exhibited great technical
competence, ‘impeccable logic’, and considerable ingenuity. This has
contributed in no small measure to its apparent success, and to the confusion
and uncertainty which rational expectations have aroused in the rest of the
economics profession. The fundamental simplicity of the ideas involved has
become obscured by overly rigorous development, and especially by the
unconvincing resort to extraneous constructions, such as the ‘islands’
mentioned above.
Undoubtedly, it is the impotence of policy results that has aroused most
attention. Yet these results depend very heavily on a particular type of
macroeconomic model usually embodying a strong form of the natural rate
hypothesis. If you start with ‘classical’ models in which policy can have no
real effects, it is hardly surprising that you get results in which policy is
impotent. Because of this the novelty of rational expectations has become
bundled up with tired and worn notions of the way in which the world
works. It is vitally important to unbundle these ideas.
The rational expectations hypothesis, in itself, should not be provocative
to economists. It merely brings expectations within the scope of individual
maximizing behavior. Expectations used to be handled within economic
models on an ad hoc basis. Rational expectations provides a way of
incorporating expectations which is consistent with the orthodox economic
theorizing.
The development of rational expectations theory will make a more
significant contribution to economics in the impetus it gives to research on
the vital areas of learning and expectations formation. It brings to the fore
questions about the availability and use of information. Instead of being the
finale of the monetarist’s case against policy intervention, it should be seen
as the prologue for a revitalized theory of expectations, information and
policy.
Bert: I guess you’re right. Let’s go and get another beer.
A child’s guide to rational expectations 307
APPENDIX A: AGGREGATE SUPPLY
The underlying inspiration for rational expectations macro models is derived
from the notion of general equilibrium. With price flexibility, for given
endowments and skills, a condition of general equilibrium requires
equilibrium in the labor markets. In such a world all unemployment is
voluntary, everybody who wants a job has one. Every individual has labor
hours and assets allocated according to some personal optimum. The
remaining unemployment can be termed the ‘natural’ rate of unemployment
and the level of output termed the ‘natural’ level of output.
Abstracting from inter-industry shifts in production, the only way output
can change is through a change in employment. To increase or decrease the
level of output government policy must alter the equilibrium in the labor
markets. But if the natural rate of unemployment represents an optimal
position for private actors, how can government policy affect it?
The models rational expectations theorists usually work with suggest that
this is possible only if the government is able to fool people. If people
confuse nominal wage changes for real ones, they might reallocate their
portfolios and their hours of work, and thus increase output. While allowing
for this possibility the models suggest that such a change would not be
desirable for the worker (representing a suboptimal decision) and would be
avoided if they had rational expectations. They suggest that the labor supply
decision is made in real terms so that labor market equilibrium is
independent of prices which, in turn, is taken to imply that output is
independent of prices. This result is presented in a vertical aggregate supply
curve.
Alternative macro-economic theories suggest that the optimal allocation
decisions of private actors will be affected by changes in prices, but not just
because people are fooled. If this were true, increases in aggregate demand
could increase output and employment even with rational expectations. One
argument for the proposition suggests that people don’t hold money in their
asset portfolios simply for transaction purposes. If prices go up, the
desirability of holding such money goes down, changing people’s private
allocation decisions, and perhaps the rate of capital formation or number of
hours worked. Thus it might be said that the rational expectations models
assume that the only motive for holding money is the transactions motive.
APPENDIX B: ALGEBRA OF THE MODEL
(1)
(2)
(3)
308 Rodney Maddock and Michael Carter
where y
t
= income
y¯ = income level corresponding to the natural rate of
unemployment
p
t
= prices
= price expectations
x
t
= government policy instrument, e.g. money supply
I
t-1
= all information available at time t-1
u
t
= random error term; Eu
t
=0
E = expectations operator.
Equating demand and supply, we obtain the following reduced form
equation:
Now, by the rational expectations assumption (3)
That is, the deviation of output from the ‘natural’ level y¯ depends only on
the unsystematic component of government policy (x–Ext). To see this,
assume that the government uses the following policy rule:
(4)
(5)
Subtracting (5) from (4)
Substituting (6) in equation (1)
(7)
(6)
where v
t
is a random variable, Ev
t
=0.
Then,
(8)
(9)
A child’s guide to rational expectations 309
Deviations of yt from y¯ are thus entirely random. This implies that
systematic government policy is impotent (in this model), since the
systematic component of any policy will be incorporated in Ext, and
therefore be cancelled out in forming xt-Ext.
ACKNOWLEDGEMENTS
Our thanks to Neville Cain for his inspiration for this chapter and to our
colleagues at ANU, notably Malcolm Gray, Adrian Pagan and Jim
Trevithick, for their comments. We are also grateful to Fred Gruen and to an
anonymous referee and the editor of Journal of Economic Literature for their
assistance. All faults, of course, remain ours.
NOTES
1 Usually defined as minimization of the variance around some fixed macroeconomic
objectives (Gregory Chow 1970).
2 There is a clear ideological component to much rational-expectations work and
opponents will be tempted to dismiss the theory on ideological grounds. Later we
suggest that there are merits in the theory quite separate from its use to support
particular propositions about the role of government.
3 See M.Friedman (1968) and Robert J.Gordon (1976) for views on this. Appendix
A deals with the issue in some more detail.
4 ‘Natural’ in the sense that everybody who wants a job at the going wage has one.
This definition denies the possibility of unemployment arising from a failure of
effective demand and hence from the ‘Keynesian’ problem (Edmond Malinvaud
1977). There is no necessary connection between vertical Phillips curves and a
natural rate of employment.
5 This is Friedman’s proposition that in the long run anticipated and actual economic
values must be equal so that policies that work through illusions, or systematically
wrong anticipations, will be ineffective in that long run.
6 Gordon (1976) considers a number of possibilities mainly relying on different
forms of sluggish price adjustment.
7 Robert Hall (1975) makes this criticism. As he interpreted the evidence, most of
the variation in output came from changes in the natural rate, provoking questions
about the importance of a theory which only explained deviations from the natural
rate. It would be a useful theory if it explained the movements in the rate itself.
Subtracting (9) from (8)
Putting this in (7)
310 Rodney Maddock and Michael Carter
8 John Taylor (1975) explores the possibilities for policy while people learn the new
rule. Benjamin Friedman (1979) addresses the same question.
9 John Muth (1961:317) in outlining what he meant by rational expectations
anticipated this criticism.
10 There is clearly some tension in macroeconomics between its empirical behavioral
aspects (e.g. the consumption function) and its derivation of insights from a
microeconomic basis (e.g. permanent income hypothesis). The micro foundations
of macroeconomics literature, for example Geoffrey Harcourt (1977) attempts to
resolve this conflict but, so far, not very successfully.
11 This is not the approach usually adopted by rational expectations theorists (n.
15). It is, however, closer to the usual economic methodology and seems preferable.
12 That is, consistent with the methodological approach of explaining all behavior in
terms of utility maximization.
13 McCallum (1980:38). In fact, Ernie has quoted McCallum out of context. He goes
on to admit that information costs are neglected for simplicity.
14 Edgar Feige and Douglas Pierce (1976) consider the implications of costly
information for rational expectations.
15 The distinction seems important for clarifying ideas within macroeconomics. The
all-information approach adopted by Sargent et al. should ideally be given another
name, for example ‘Muth expectations’, since the adjective ‘rational’ is normally
reserved in economics to describe the outcome of a utility maximization process.
J.J.Sijben writes: ‘Muth’s view implies that economic agents build up their
expectations as if they are fully informed of the process which ultimately generates
the real outcome of the variable concerned’ (1980:66). Pushed further, McCallum
follows the line that all models are ‘unrealistic’, which seems to lead him to the
position that theories stand or fall on their predictions.
16 Rational expectations in labor supply decisions have fairly obvious corollaries on
capital investment decisions (Robert Lucas, 1975, for example).
17 Appendix A deals with the problems of the vertical aggregate supply curve in more
detail. It should be noted that the models are usually expressed in logarithms so
that the real debate concerns rates of change rather than levels. The distinction is
neglected here.
18 As Thomas Sargent and Neil Wallace suggest ‘it must somehow trick the public’
(1976:177). The argument is more complex with capital in the model as may be
clear from Appendix A.
19 The idea that changes in money supply do not influence people’s preferred hours
of work, portfolio holdings, etc. Again, this is considered in Appendix A.
20 Sargent and Wallace (1976:177–8), put this argument in almost the same form.
Expectations can be wrong but not systematically wrong (i.e. biased), hence there
is no scope for systematic policy.
21 Gordon (1976) makes this clear. See especially pp. 200–1.
22 This follows the usual solution method followed by rational expectations models.
See Lucas (1973) for an example.
23 Some criticisms are discussed in Appendix A.
24 This was Hall’s criticism (1975), and is also put by Gordon to Sargent (1973:478).
25 Lucas (1973) introduced the lagged term with a footnote explaining that not all
deviation from the natural rate of unemployment could be accounted for by the
error in expectations terms.
26 Lucas (1975) attempts a systematic explanation for the serial correlation in terms
A child’s guide to rational expectations 311
of information lags. See Rodney Maddock (1979) for a discussion of the importance
of persistence for the rational expectation program.
27 ‘The idea behind this island abstraction is not, of course, to gain insight into
maritime affairs, or to comment on the aimlessness of life. It is intended simply to
capture in a tractable way the fact that economic activity offers agents a succession
of ambiguous, unanticipated opportunities which cannot be expected to stay fixed
while more information is collected. It seems safe and, for my purposes, sensible to
abstract here from the fact that in reality this situation can be slightly mitigated by
the purchase of additional information’ (Lucas 1975:1120).
28 Lucas does actually mention the problem in the quotation in note 27, but makes
nothing of it.
29 These issues are raised in a penetrating discussion of the problem of persistence by
Gordon (1981).
30 For example, Muth (1961) argued that economists could sell the information
profitably if expectations were not rational. Since he wrote, many have done so.
This suggests that market forces would tend to drive decisions to those rationally
based.
31 Shiller (1978:38) focuses upon the issue of convergence. There seem to be two
separate issues involved. Since rational expectations for this period depend upon
estimates about the future while the future depends in part upon present
expectations, there need be no unique rational expectation for the current period.
In many models, methods of adjusting expectations (i.e. forecasts) of the future
will either converge on a rational-expectations solution or explode. The implicit
argument of protagonists seems to be that since we do not observe prices exploding
off to infinity we need only consider converging cases. This type of counter-factual
reasoning is somewhat dubious. The dynamics of expectation formation might
still be explosive but some other fact or—e.g. policy action—act to constrain the
explosive tendency.
32 Sargent and Wallace (1976:181–3) argue this point.
33 The nature of the problem when the government’s objectives vary over time does
not seem to have been well explored. Clearly, rational expectations forces
economists to think more about the precise nature of learning.
34 Following the tradition of Dutch economists Jan Tinbergen (1952) and Henri
Theil (1958).
35 Following a sound Australian tradition.
36 Actually, some allowance in the tests is made for persistence by the inclusion of
lagged values of the dependent variable.
37 The theory suggests that no extra information would significantly contribute to
the prediction. The evidence would thus appear to falsify the theory. Sargent
(1976a), instead, went on to try an alternative type of test. See especially p. 233.
38 Sargent (1976b) and Sargent and Wallace (1975) start to develop this idea.
39 Since expectations are usually unobservable they are eliminated from econometric
models to be estimated by introducing an equation about their relation to known
variables. The parameters of this equation then become embedded in the actual
reduced form equations which are estimated. An ‘equation’ for rational expectations
incorporates the parameters of the policy rule being used by the government.
Under the assumption of rational expectations these parameters of the policy rule
become embedded in the reduced form (i.e. estimating) equations of the economy.
312 Rodney Maddock and Michael Carter
Thus they suggest that structure of the economy, as measured by usual econometric
models, will appear to change whenever the policy rule changes.
40 Using Ml there was a policy change, but with M2 none appeared to have taken
place.
41 The tests were non-parametric.
42 Alan Preston and Adrian Pagan (1982: ch. 10) explore the way in which the
impotence result depends on particular specifications of the macroeconomic model
being considered. They develop a more general model that includes the usual
rational expectations models as particular cases.
43 Lucas and Leonard A.Rapping used adaptive expectations (1969a), then rational
distributed lag expectations (1969b) before Lucas first introduced rational
expectations (1972). Where adaptive expectations assumed that people just simply
adapted to past errors, rational distributed lag expectations were the very best
econometrically predicted estimates of prices derived from analysis of all past price
information.
44 Any stable understandable rule would have no effect if people were exactly able to
predict it. By that test a ‘discretionary’ rule and a fixed money growth rate rule
would be equally impotent. If stochastic effects of discretionary rules are allowed
for however, these cannot be predicted and would introduce fluctations into the
system. See Sargent and Wallace (1976) for a discussion of the issues.
45 McCallum’s (1980) popularization of their position carries the inference that the
results are quite robust, but that does not appear to be the case. See Preston and
Pagan (1982).
46 The seminal article by Muth (1961) deals with microeconomic market situations.
47 Sargent and Wallace (1975:251) provide a start in this direction by modelling a
case where government has an information advantage over private actors.
REFERENCES
Buiter, Willem, ‘The Macroeconomics of Dr. Pangloss: A Critical Survey of the New
Classical Macroeconomics’, Economic Journal March 1980, 90(1), pp. 34–50.
Chow, Gregory, ‘Optimal Stochastic Control of Linear Economic Systems’, Journal of
Money, Credit and Banking Aug. 1970, 2(3), pp. 291–302.
Decanio, Stephen, ‘Rational Expectations and Learning from Experience’, Quarterly
Journal of Economics 1979, 93(1), pp. 47–57.
Feige, Edgar, and Pearce, Douglas, ‘Economically Rational Price Expectations’, Journal
of Political Economy 1976, 84(3), pp. 499–522.
Friedman, Benjamin, ‘Optimal Expectations and the Extreme Information Assumptions
of “Rational Expectations” Macromodels’, Journal of Monetary Economics Jan.
1979, 5(1), pp. 23–41.
Friedman, Milton, ‘The Role of Monetary Policy’, American Economic Revolution
March 1968, 58(1), pp. 1–17.
Gordon, Robert A., ‘Rigor and Relevance in a Changing Institutional Setting’, American
Economic Review 1976, 66(1), pp. 1–14.
Gordon, Robert J., ‘Recent Developments in the Theory of Inflation and Unemployment’,
Journal of Monetary Economics April 1976, 2(2), pp. 185–220.
——‘Output Fluctuations and Gradual Price Adjustment’, Journal of Economic
Literature June 1981, pp. 493–530.
Hall, Robert E., ‘The Rigidity of Wages and the Persistence of Unemployment’, Brookings
Papers on Economic Activity 1975, 2, pp. 301–49.
A child’s guide to rational expectations 313
Harcourt, Geoffrey (ed.) The Microeconomic Foundations of Macroeconomics, London:
Macmillan, 1977.
Keynes, John, A Treatise on Money, vols I and II, New York: Harcourt Brace, 1930.
Lucas, Robert E. Jr, ‘Testing the Natural Rate Hypothesis’, in The Econometrics of
Price Determination, Conference, edited by Otto Eckstein, Washington, DC: Federal
Reserve Board, 1972, pp. 50–9.
——‘Some International Evidence on Output-Inflation Tradeoffs’, American Economic
Review June 1973, 63(3), pp. 326–34.
——‘An Equilibrium Model of the Business Cycle’, Journal of Political Economy Dec.
1975, 83(6), pp. 1113–44.
——‘Econometric Policy Evaluation: A Critique’, Journal of Monetary Economics
Supplementary Series 1976, 1, pp.19–46.
Lucas, Robert E.Jr and Rapping, Leonard A., ‘Price Expectations and the Phillips
Curve’, American Economic Review June 1969a, 59(3), pp. 342–50.
——‘Real Wages, Employment and Inflation’, Journal Political Economy Sept./Oct.
1969b, 77(5), pp. 721–54.
Maddock, Rodney, ‘Rational Expectations, Political Business Cycles and the Course of
Macroeconomic Theory’, Unpublished Ph.D thesis, Duke University, 1979.
Malinvaud, Edmond, The Theory of Unemployment Reconsidered, New York: Wiley;
Oxford: Blackwell, 1977.
McCallum, Bennett T., ‘The Significance of Rational Expectations Theory’, Challenge
Jan./Feb. 1980, pp. 37–43.
Muth, John, ‘Rational Expectations and the Theory of Price Movements’, Econometrica
July 1961, 29(3), pp. 315–35.
Neftci, Salih and Sargent, Thomas, ‘A Little Bit of Evidence on the Natural Rate
Hypothesis from the U.S.’, Journal of Monetary Economics 1978, 4(1), pp. 315–
19.
Pesando, James, ‘A Note on the Rationality of the Livingston Price Expectations’,
Journal of Political Economy Aug. 1975, 83(4), pp. 849–58.
Preston, Alan and Pagan, Adrian, The Theory of Economic Policy: Statics and Dynamics,
New York: Cambridge University Press, 1982.
Sargent, Thomas, ‘Rational Expectations, the Real Rate of Interest, and the Natural
Rate of Unemployment’, Brookings Papers on Economic Act, 1973, 2, pp. 429–72.
——‘A Classical Macroeconomic Model for the United States’, Journal of Political
Economy April 1976a, 84(2), pp. 207–37.
——‘The Observational Equivalence of Natural and Unnatural Rate Theories of
Macroeconomics’, Journal of Political Economy June 1976b, 84(3), pp. 631–40.
Sargent, Thomas, and Wallace, Neil, ‘“Rational” Expectations, the Optimal Monetary
Instrument and the Optimal Money Supply Rule’, Journal of Political Economy
1975, 83(2), pp. 241–55.
——‘Rational Expectations and the Theory of Economic Policy’, Journal of Monetary
Economics April 1976, 2(2), pp. 169–83.
Shiller, Robert J., ‘Rational Expectations and the Dynamic Structure of Macroeconomic
Models: A Critical Review’, Journal of Monetary Economics Jan. 1978, 4(1), pp. 1–44.
Sijben, J.J., Rational Expectations and Monetary Policy, Alphen aan den Rijn, The
Netherlands: Sijthoff and Noordhoff, 1980.
Taylor, John, ‘Monetary Policy during the Transition to Rational Expectations’, Journal
of Political Economy 1975, 83(5), pp. 1009–21.
Theil, Henri, Economic Forecasts and Policy, Amsterdam: North-Holland, 1958.
Tinbergen, Jan, On the Theory of Economic Policy, Amsterdam: North-Holland, 1952.
Turnovsky, Stephen J., ‘Empirical Evidence on the Formation of Price Expectations’,
Journal American Statistical Association Dec. 1970, 65(332), pp. 1441–54.
13 The Ricardian approach to
budget deficits
Robert J.Barro
Journal of Economic Perspectives (1989) 3, Spring, pp. 37–54
In recent years there has been a lot of discussion about US budget deficits.
Many economists and other observers have viewed these deficits as harmful
to the US and world economies. The supposed harmful effects include high
real interest rates, low saving, low rates of economic growth, large current-
account deficits in the United States and other countries with large budget
deficits, and either a high or low dollar (depending apparently on the time
period). This crisis scenario has been hard to maintain along with the robust
performance of the US economy since late 1982. This performance features
high average growth rates of real GNP, declining unemployment, much
lower inflation, a sharp decrease in nominal interest rates and some decline
in expected real interest rates, high values of real investment expenditures,
and (until October 1987) a dramatic boom in the stock market.
Persistent budget deficits have increased economists’ interest in theories
and evidence about fiscal policy. At the same time, the conflict between
standard predictions and actual outcomes in the US economy has, I think,
increased economists’ willingness to consider approaches that depart from
the standard paradigm. In this paper I will focus on the alternative theory
that is associated with the name of David Ricardo.
THE STANDARD MODEL OF BUDGET DEFICITS
Before developing the Ricardian approach, I will sketch the standard model.
The starting point is the assumption that the substitution of a budget deficit
for current taxation leads to an expansion of aggregate consumer demand. In
other words, desired private saving rises by less than the tax cut, so that
desired national saving declines. It follows for a closed economy that the
expected real interest rate would have to rise to restore equality between
desired national saving and investment demand. The higher real interest rate
crowds out investment, which shows up in the long run as a smaller stock of
productive capital. Therefore, in the language of Franco Modigliani (1961),
the public debt is an intergenerational burden in that it leads to a smaller
stock of capital for future generations. Similar reasoning applies to pay-as-
you-go social security programs, as has been stressed by Martin Feldstein
The Ricardian approach to budget deficits 315
(1974). An increase in the scope of these programs raises the aggregate
demand for goods, and thereby leads to a higher real interest rate and a
smaller stock of productive capital.
In an open economy, a small country’s budget deficits or social security
programs would have negligible effects on the real interest rate in
international capital markets. Therefore, in the standard analysis, the home
country’s decision to substitute a budget deficit for current taxes leads mainly
to increased borrowing from abroad, rather than to a higher real interest
rate. That is, budget deficits lead to current-account deficits. Expected real
interest rates rise for the home country only if it is large enough to influence
world markets, or if the increased national debt induces foreign lenders to
demand higher expected returns on this country’s obligations. In any event,
there is a weaker tendency for a country’s budget deficits to crowd out its
domestic investment in the short run and its stock of capital in the long run.
However, the current-account deficits show up in the long run as a lower
stock of national wealth—and correspondingly higher claims by foreigners.
If the whole world runs budget deficits or expands the scale of its social
insurance programs, real interest rates rise on international capital markets,
and crowding-out of investment occurs in each country. Correspondingly, the
world’s stock of capital is lower in the long run. These effects for the world
parallel those for a single closed economy, as discussed before.
THE RICARDIAN ALTERNATIVE
The Ricardian modification to the standard analysis begins with the
observation that, for a given path of government spending, a deficit-financed
cut in current taxes leads to higher future taxes that have the same present
value as the initial cut. This result follows from the government’s budget
constraint, which equates total expenditures for each period (including
interest payments) to revenues from taxation or other sources and the net
issue of interest-bearing public debt. Abstracting from chain-letter cases
where the public debt can grow forever at the rate of interest or higher, the
present value of taxes (and other revenues) cannot change unless the
government changes the present value of its expenditures. This point
amounts to economists’ standard notion of the absence of a free lunch—
government spending must be paid for now or later, with the total present
value of receipts fixed by the total present value of spending. Hence, holding
fixed the path of government expenditures and non-tax revenues, a cut in
today’s taxes must be matched by a corresponding increase in the present
value of future taxes.1
Suppose now that households’ demands for goods depend on the expected
present value of taxes—that is, each household subtracts its share of this
present value from the expected present value of income to determine a net
wealth position. Then fiscal policy would affect aggregate consumer demand
only if it altered the expected present value of taxes. But the preceding
316 Robert J.Barro
argument was that the present value of taxes would not change as long as
the present value of spending did not change. Therefore, the substitution of a
budget deficit for current taxes (or any other rearrangement of the timing of
taxes) has no impact on the aggregate demand for goods. In this sense,
budget deficits and taxation have equivalent effects on the economy—hence
the term, ‘Ricardian equivalence theorem’.2 To put the equivalence result
another way, a decrease in the government’s saving (that is, a current budget
deficit) leads to an offsetting increase in desired private saving, and hence to
no change in desired national saving.
Since desired national saving does not change, the real interest rate does
not have to rise in a closed economy to maintain balance between desired
national saving and investment demand. Hence, there is no effect on
investment, and no burden of the public debt or social security in the sense of
Modigliani (1961) and Feldstein (1974). In a setting of an open economy
there would also be no effect on the current-account balance because desired
private saving rises by enough to avoid having to borrow from abroad.
Therefore, budget deficits would not cause current-account deficits.
THEORETICAL OBJECTIONS TO RICARDIAN EQUIVALENCE
I shall discuss five major theoretical objections that have been raised against
the Ricardian conclusions. The first is that people do not live forever, and
hence do not care about taxes that are levied after their death. The second is
that private capital markets are ‘imperfect’, with the typical person’s real
discount rate exceeding that of the government. The third is that future taxes
and incomes are uncertain. The fourth is that taxes are not lump sum, since
they depend typically on income, spending, wealth, and so on. The fifth is that
the Ricardian result hinges on full employment. I assume throughout that the
path of government spending is given. The Ricardian analysis applies to shifts
in budget deficits and taxes for a given pattern of government expenditures; in
particular, the approach is consistent with real effects from changes in the level
or timing of government purchases and public services.
In many cases it turns out that budget deficits matter, and are in that sense
non-Ricardian. It is important, however, to consider not only whether the
Ricardian view remains intact, but also what alternative conclusions emerge.
Many economists raise points that invalidate strict Ricardian equivalence, and
then simply assume that the points support a specific alternative; usually the
standard view that a budget deficit lowers desired national saving and thereby
drives up real interest rates or leads to a current-account deficit. Many criticisms
of the Ricardian position are also inconsistent with this standard view.
Finite horizons and related issues
The idea of finite horizons, motivated by the finiteness of life, is central to
life-cycle models—see, for example, Franco Modigliani and Richard
The Ricardian approach to budget deficits 317
Brumberg (1954) and Albert Ando and Franco Modigliani (1963). In these
models individuals capitalize only the taxes that they expect to face before
dying. Consider a deficit-financed tax cut, and assume that the higher future
taxes occur partly during the typical person’s expected lifetime and partly
thereafter. Then the present value of the first portion must fall short of the
initial tax cut, since a full balance results only if the second portion is
included. Hence the net wealth of persons currently alive rises, and
households react by increasing consumption demand. Thus, as in the
standard approach sketched above, desired private saving does not rise by
enough to offset fully the decline in government saving.
A finite horizon seems to generate the standard result that a budget deficit
reduces desired national saving. The argument works, however, only if the
typical person feels better off when the government shifts a tax burden to his
or her dependants. The argument fails if the typical person is already giving
to his or her children out of altruism. In this case people react to the
government’s imposed intergenerational transfers, which are implied by
budget deficits or social security, with a compensating increase in voluntary
transfers (Barro 1974). For example, parents adjust their bequests or the
amounts given to children while the parents are still living. Alternatively, if
children provide support to aged parents, the amounts given can respond
(negatively) to budget deficits or social security.
The main idea is that a network of intergenerational transfers makes the
typical person a part of an extended family that goes on indefinitely. In this
setting, households capitalize the entire array of expected future taxes, and
thereby plan effectively with an infinite horizon. In other words, the
Ricardian results, which seemed to depend on infinite horizons, can remain
valid in a model with finite lifetimes.
Two important points should be stressed. First, intergenerational transfers
do not have to be ‘large’; what is necessary is that transfers based on
altruism be operative at the margin for most people.3 Specifically, most
people must be away from the corner solution of zero transfers, where they
would, if permitted, opt for negative payments to their children. (The results
also go through, however, if children typically support their aged parents.)
Second, the transfers do not have to show up as bequests at death. Other
forms of intergenerational transfers, such as inter vivos gifts to children,
support of children’s education, and so on, can work in a similar manner.
Therefore, the Ricardian results can hold even if many persons leave little in
the way of formal bequests.
One objection to Ricardian equivalence is that some persons, such as
those without children, are not connected to future generations (see James
Tobin and Willem Buiter 1980:86 ff.). Persons in this situation tend to be
made wealthier when the government substitutes a budget deficit for taxes.
At least this conclusion obtains to the extent that the interest and principal
payments on the extra public debt are not financed by higher taxes during
the remaining lifetimes of people currently alive. However, the quantitative
318 Robert J.Barro
effects on consumption tend to be small. For example, if the typical person
has 30 years of remaining life and consumes at a constant rate, a one-time
budget deficit of $100 per person would increase each person’s real
consumption demand by $1.50 per year if the annual real interest rate is 5
percent, and by $2.10 per year if the real interest rate is 3 percent.4
The aggregate effect from the existence of childless persons is even
smaller because people with more than the average number of descendants
experience a decrease in wealth when taxes are replaced by budget deficits.
(In effect, although some people have no children, all children must have
parents.) In a world of different family sizes, the presumption for a net effect
of budget deficits on aggregate consumer demand depends on different
propensities to consume out of wealth for people with and without children.
Since the propensity for those without children tends to be larger (because of
the shorter horizon), a positive net effect on aggregate consumer demand
would be predicted. However, the quantitative effect is likely to be trivial.
Making the same assumptions as in the previous example, a budget deficit of
$100 per capita would raise real consumption demand per capita by 30 cents
per year if the real interest rate is 5 percent, and by 90 cents if the real
interest rate is 3 percent.
A variety of evidence supports the proposition that intergenerational
transfers—defined broadly to go beyond formal bequests—are operative for
most people. Michael Darby (1979: ch. 3) and Laurence Kotlikoff and
Lawrence Summers (1981) calculate that the accumulation of households’
assets in the United States for the purpose of intergenerational transfers is far
more important than that associated with the life cycle. This observation
suggests that most people give or receive intergenerational transfers—a
conclusion that supports the Ricardian position. Franco Modigliani (1988)
contests this conclusion, but Laurence Kotlikoff (1988) shows that
Modigliani’s findings derive from an extremely nar row view of
intergenerational transfers. Modigliani focuses on bequests at death, and he
also does not treat interest earnings on prior bequests as income attributable
to intergenerational transfers.
Some authors accept the idea that intergenerational transfers are
important, but argue that the motivation for the transfers matters for the
results. Douglas Bernheim, Andrei Shleifer and Lawrence Summers (1985)
consider the possibility that bequests, instead of being driven by altruism, are
a strategic device whereby parents induce their children to behave properly.
Some imaginative evidence is presented (involving how often children visit
and communicate with their parents) to document the importance of strategic
bequests. In this strategic model, if the government redistributes income from
young to old (by running a deficit or raising social security benefits), the old
have no reason to raise transfers to offset fully the government’s actions.
Instead, the old end up better off at the expense of the young, and aggregate
consumer demand rises. Then, as in the standard approach, real interest
rates increase or domestic residents borrow more from abroad.
The Ricardian approach to budget deficits 319
One shortcoming of this approach is that it treats the interaction between
parents and children as equivalent to the purchases of services on markets. In
this setting parents would tend to pay wages to children, rather than using
bequests or other forms of intergenerational transfers. These features—as
well as the observation that most parents seem to care about their children’s
welfare—can be better explained by introducing altruism along with a desire
to influence children’s behavior. In this case Ricardian equivalence may or
may not obtain. Consider the utility that a parent would allocate to his or
her child if there were no difficulty in motivating the child to perform
properly. Suppose that the parent can design a credible threat involving
bequests that entails the loss of some part of this utility for the child. (Note
that if no threats are credible, the whole basis for strategic bequests
disappears.) If the threat is already large enough to induce the behavior that
the parent desires, Ricardian equivalence still holds. For example, if the
government runs a budget deficit, the parent provides offsetting transfers to
the child, and thereby preserves the child’s level of utility, as well as the
behavior sought by the parent. On the other hand, the parent may have to
allow excess utility to the child to secure a sufficient threat against bad
performance. Then a budget deficit enables the parent to reduce the child’s
utility (as desired), while maintaining or even enhancing the threat that
influences behavior. In this case Ricardian equivalence would not hold.
Other economists argue that the uncertainty of the time of death makes
many bequests unintended, and that such bequests would not respond very
much to budget deficits. The imperfection of private annuity markets is
usually mentioned to explain why unintended bequests are significant. But
this reasoning is backwards, since annuities do not entail greater adverse
selection problems than many other types of insurance. The small amount of
private annuities outstanding, other than the substantial amount in the form
of pensions, reflects primarily a lack of demand, which itself is an indication
that people desire to make the most of the bequests that occur. In any event,
since the Ricardian results involve a broad concept of intergenerational
transfers, rather than especially bequests at death, a focus on formal bequests
is misplaced.
Imperfect loan markets
Many economists argue that the imperfection of private credit markets is
central to an analysis of the public debt; see, for example, Robert Mundell
(1971). To consider this argument, assume that a closed economy consists of
two types of infinite-lived economic agents; those of group A who have the
same discount rate, r, as the government (and are therefore willing to hold
the government’s debt), and those of group B who have the higher discount
rate, The constituents of group A would include large businesses,
pension funds, and some individuals. The members of group B, such as small
businesses and many households, possess poor collateral; therefore, loans to
320 Robert J.Barro
these people imply large costs of evaluation and enforcement. It follows that
the members of group B face higher borrowing rates (even after an
allowance for default risk) than the government. Whether or not they are
actually borrowing, the high discount rate for group B corresponds to a
high rate of time preference for consumption and a high marginal return on
investment.
Suppose that the government cuts current taxes and runs a budget deficit.
Further, assume that the division of the tax cut between groups A and B—say
fifty-fifty—is the same as the division of the higher future taxes needed to
service the extra debt. Since those from group A experience no net change in
wealth, they willingly hold their share of the extra public debt. For group B,
where the discount rate exceeds r, the present value of the extra future
taxes falls short of the tax cut. The members of this group are better off
because the tax cut effectively enables them to borrow at the lower interest
rate, r. This cut in the effective borrowing rate motivates the members of
group B to raise current consumption and investment.
In the aggregate a budget deficit now raises aggregate demand, or
equivalently, the aggregate of desired private saving increases by less than
one-to-one with the government’s deficit. It follows that the real interest rate
r, which applies to group A and the government, must rise to induce people
to hold the extra public debt. Hence there is crowding out of consumption
and investment by members of group A. For group B, the opportunity to
raise current consumption and investment means that the rate of time
preference for consumption and the marginal return to investment would
decline. That is, the discount rate r falls. Thus, the main effects are a
narrowing of the spread between the two discount rates, r and , and a
diversion of current expenditures from group A to group B. In the aggregate
investment may either rise or fall, and the long-term effect on the capital
stock is uncertain. The major change, however, is a better channeling of
resources to their ultimate uses. Namely the persons from group B—who
have relatively high values for rates of time preference and for marginal
returns to investment—command a greater share of current output. In any
event the outcomes are non-neutral, and in that sense non-Ricardian.
The important finding from the inclusion of imperfect loan markets is that
the government’s issue of public debt can amount to a useful form of
financial intermediation. The government induces people with good access to
credit markets (group A) to hold more than their share of the extra public
debt. Those with poor access (group B) hold less than their share, and
thereby effectively receive loans from the first group. This process works
because the government implicitly guarantees the repayment of loans
through its tax collections and debt payments. Thus loans between A and B
take place even though such loans were not viable (because of ‘transaction
costs’) on the imperfect private credit market.
The Ricardian approach to budget deficits 321
This much of the argument may be valid, although it credits the
government with a lot of skill in the collection of taxes from people with
poor collateral (which is the underlying source of the problem for private
lenders). Even if the government possesses this skill, the conclusions do not
resemble those from the standard analysis. As discussed before, budget
deficits can amount to more financial intermediation, and are in that sense
equivalent to a technological advance that improves the functioning of loan
markets. From this perspective it is reasonable to find a reduced spread
between various discount rates and an improvement in the allocation of
r e s o u r c e s . I f t h e g o v e r n m e n t r e a l l y i s b e t t e r a t t h e p r o c e s s o f
intermediating, more of this activity—that is, more public debt—raises
perceived wealth because it actually improves the workings of the
economy.
In the preceding analysis, the imperfection of credit markets reflected
costs of enforcing the collection of loans. A different approach, followed by
Toshiki Yotsuzuka (1987) in his extension of the models of Mervyn King
(1986) and Fumio Hayashi (1987), allows for adverse selection among
borrowers with different risk characteristics. Individuals know their
probabilities of default, but the lenders’ only possibility for learning these
probabilities comes from observing the chosen levels of borrowing at going
interest rates. In this setting the government’s borrowing amounts to a loan
to a group that pools the various risk classes. Such borrowing matters if the
private equilibrium does not involve similar pooling. However, by
considering the incentives of lenders to exchange or not exchange
information about their customers, Yotsuzuka argues that the private
equilibrium typically involves a pooled loan of limited quantity at a
relatively low interest rate. Then the high-risk types may borrow
additional amounts at a high interest rate. (The assumption is that this
additional borrowing is not observable by other lenders.) In this case the
government’s borrowing replaces the private pooled lending, and leads to
no real effects. That is, Ricardian equivalence holds despite the imperfect
private loan market where high-risk people face high marginal borrowing
rates. The general lesson again is that Ricardian equivalence fails because
of imperfect credit markets only if the government does things in the loan
market that are different from, and perhaps better than, those carried out
privately.
Uncertainty about future taxes and incomes
Some economists argue that the uncertainty about individuals’ future taxes—
or the complexity in estimating them—implies a high rate of discount in
capitalizing these future liabilities (Martin Bailey 1971:157–8; James
Buchanan and Richard Wagner 1977:17, 101, 130; Martin Feldstein
1976:335). In this case, a substitution of a budget deficit for current taxes
raises net wealth because the present value of the higher expected future
322 Robert J.Barro
taxes falls short of the current tax cut. It then follows that budget deficits
raise aggregate consumer demand and reduce desired national saving.
A proper treatment of uncertainty leads to different conclusions. Louis
Chan (1983) first considers the case of lump-sum taxes that have a known
distribution across households. However, the aggregate of future taxes
and the real value of future payments on public debt are subject to
uncertainty. In this case a deficit-financed tax cut has no real effects.
Individuals hold their share of the extra debt because the debt is a perfect
hedge against the uncertainty of the future taxes. (This analysis assumes
that private credit markets have no ‘imperfections’ of the sort discussed
earlier.)
Suppose now that future taxes are still lump sum but have an uncertain
incidence across individuals. Furthermore, assume that there are no
insurance markets for relative tax risks. Then a budget deficit tends to
increase the uncertainty about each individual’s future disposable income.
Chan (1983:363) shows for the ‘usual case’ (of non-increasing absolute risk
aversion) that people react by reducing current consumption and hence, by
raising current private saving by more than the tax cut. Consequently, the
effects on real interest rates, investment, the current account, and so on are
the opposites of the standard ones.
The results are different for an income tax (Chan 1983:364–6; Robert
Barsky, Gregory Mankiw and Stephen Zeldes 1986). Suppose that each
person pays the tax τyi, where yi is the person’s uncertain future income.
Suppose that there are no insurance markets for individual income risks,
and that τ is known. (The analysis thus abstracts from uncertainties in
relative tax rates across individuals.) In this case a budget deficit raises the
future value of τ and thereby reduces the uncertainty about each
individual’s future disposable income. In effect, the government shares the
risks about individual disposable income to a greater extent. It follows that
the results are opposite to those found before; namely, a budget deficit
tends to raise current consumption and hence, to raise private saving by
less than the tax cut.
Overall, the conclusions depend on the net effect of higher mean future
tax collections on the uncertainty associated with individuals’ future
disposable incomes. Desired national saving tends to rise with a budget
deficit if this uncertainty increases, and vice versa.
The timing of taxes
Departures from Ricardian equivalence arise also if taxes are not lump
sum; for example, with an income tax. In this situation, budget deficits
change the timing of income taxes, and thereby affect people’s incentives to
work and produce in different periods. It follows that variations in deficits
are non-neutral, although the results tend also to be inconsistent with the
standard view.
The Ricardian approach to budget deficits 323
Suppose, for example, that the current tax rate on labor income, t1,
declines, and the expected rate for the next period, τ2, rises. To simplify
matters, assume that today’s budget deficit is matched by enough of a surplus
next period so that the public debt does not change in later periods. Because
the tax rate applies to labor income, households are motivated to work more
than usual in period 1 and less than usual in period 2. Since the tax rate does
not apply to expenditures (and since wealth effects are negligible here),
desired national saving rises in period 1 and falls in period 2. Therefore, in
a closed economy, after-tax real interest rates tend to be relatively low in
period 1—along with the budget deficit—and relatively high in period 2—
along with the surplus. In an open economy, a current-account surplus
accompanies the budget deficit, and vice versa. Hence the results are non-
Ricardian, but also counter to the standard view. (Temporary variations in
consumption taxes tend to generate the standard pattern where real interest
rates, current-account deficits, and budget deficits are positively correlated.)
Unlike in the Ricardian case where debt and deficits do not matter, it is
possible in a world of distorting taxes to determine the optimal path of the
budget deficit, which corresponds to the optimal time pattern of taxes. In
effect, the theory of debt management becomes a branch of public finance;
specifically, an application of the theory of optimal taxation.
One result is that budget deficits can be used to smooth tax rates over
time, despite fluctuations in government expenditures and the tax base.5 For
example, if time periods are identical except for the quantity of government
purchases—which are assumed not to interact directly with labor supply
decisions—optimality dictates uniform taxation of labor income over time.
This constancy of tax rates requires budget deficits when government
spending is unusually high, such as in wartime, and surpluses when spending
is unusually low.
Constant tax rates over time will not be optimal in general; for example,
optimal tax rates on labor income may vary over the business cycle. To the
extent that some smoothing is called for, budget deficits would occur in
recessions, and surpluses in booms. If optimal tax rates are lower than
normal in recessions and higher than normal in booms, the countercyclical
pattern of budget deficits is even more vigorous. The well-known concept of
the full-employment deficit, as discussed in E.Gary Brown (1956) and
Council of Economic Advisers (1962:78–82), adjusts for this cyclical
behavior of budget deficits.
The tax-smoothing view has implications for the interaction between
inflation and budget deficits if the public debt is denominated in nominal
terms. Basically, the fiscal authority’s objective involves the path of tax rates
and other real variables. Therefore, other things equal, a higher rate of
expected inflation (presumably reflecting a higher rate of monetary growth)
motivates a correspondingly higher growth rate of the nominal, interest-
bearing debt. This response keeps the planned path of the real public debt
invariant with expected inflation. This behavior means that differences in
324 Robert J.Barro
expected rates of inflation can account for substantial variations in budget
deficits if deficits are measured in the conventional way to correspond to the
change in the government’s nominal liabilities. This element is, however, less
important for an inflation-adjusted budget deficit, which corresponds to the
change in the government’s real obligations (Jeremy Siegel 1979; Robert
Eisner and Paul Pieper 1984).
With perfect foresight, the strict tax-smoothing model implies constant tax
rates. More realistically, new information about the path of government
spending, national income, and so on, would lead to revisions of tax rates.
However, the sign of these revisions would not be predictable. Thus, in the
presence of uncertainty, tax smoothing implies that tax rates would behave
roughly like random walks.
It is possible to use the tax-smoothing approach as a positive theory of
how the government operates, rather than as a normative model of how it
should act.6 Barro (1979; 1986) shows that this framework explains much of
the behavior of US federal deficits from 1916 to 1983, although the deficits
since 1984 turn out to be substantially higher than predicted. Over the full
sample, the major departures from the theory are an excessive reaction of
budget deficits to the business cycle (so that tax rates fall below ‘normal’
during recessions) and an insufficient reaction to temporary military
spending (so that tax rates rise above normal during wars). These departures
are found also by Chaipat Sahasakul (1986), who looks directly at the
behavior of average marginal tax rates. Barro (1987: section 3) finds for the
British data from the early 1700s through 1918 that temporary military
spending is the major determinant of budget deficits. Also, unlike the US
case, the results indicate a one-to-one response of budget deficits to
temporary spending.
Full employment and Keynesian models
A common argument is that the Ricardian results depend on ‘full
employment’, and surely do not hold in Keynesian models. In standard
Keynesian analysis (which still appears in many textbooks), if everyone
thinks that a budget deficit makes them wealthier, the resulting expansion of
aggregate demand raises output and employment, and thereby actually
makes people wealthier. (This result holds if the economy begins in a state of
‘involuntary unemployment’.) There may even be multiple, rational
expectations equilibria, where the change in actual wealth coincides with the
change in perceived wealth.
This result does not mean that budget deficits increase aggregate demand
and wealth in Keynesian models. If we had conjectured that budget deficits
made people feel poorer, the resulting contractions in output and
employment would have made them poorer. Similarly, if we had started with
the Ricardian notion that budget deficits did not affect wealth, the Keynesian
results would have verified that conjecture. The odd feature of the standard
The Ricardian approach to budget deficits 325
Keynesian model is that anything that makes people feel wealthier actually
makes them wealthier (although the perception and actuality need not
correspond quantitatively). This observation raises doubts about the
formulation of Keynesian models, but says little about the effect of budget
deficits. Moreover, in equilibrium models that include unemployment (such
as models with incomplete information and search), there is no clear
interplay between the presence of unemployment and the validity of the
Ricardian approach.
EMPIRICAL EVIDENCE ON THE ECONOMIC EFFECTS OF
BUDGET DEFICITS
It is easy on theoretical grounds to raise points that invalidate strict
Ricardian equivalence. Nevertheless, it may still be that the Ricardian view
provides a useful framework for assessing the first-order effects of fiscal
policy. Furthermore, it is unclear that the standard analysis offers a more
accurate guide. For these reasons it is especially important to examine
empirical evidence.
The Ricardian and standard views have different predictions about the
effects of fiscal policy on a number of economic variables. The next three
sections summarize the empirical evidence on interest rates, saving, and the
current-account balance.
Interest rates
The Ricardian view predicts no effect of budget deficits on real interest rates,
whereas the standard view predicts a positive effect, at least in the context of
a closed economy. Many economists have tested these propositions
empirically (for a summary, see US Treasury Department 1984). Typical
results show little relationship between budget deficits and interest rates. For
example, Charles Plosser (1982:339) finds for quarterly US data from 1954
to 1978 that unexpected movements in privately held federal debt do not
raise the nominal yield on government securities of various maturities. In
fact, there is a weak tendency for yields to decline with innovations in
federal debt. Plosser’s (1987: Tables VIII and XI) later study, which includes
data through 1985, reaches similar conclusions for nominal and expected
real yields. Paul Evans (1987b) obtains similar results for nominal yields
with quarterly data from 1974 to 1985 for Canada, France, Germany, Japan,
the United Kingdom, and the United States.
Evans (1987a: Tables 4–6) finds for annual US data from 1931 to 1979
that current and past real federal deficits have no significant association with
nominal interest rates on commercial paper or corporate bonds, or with
realized real interest rates on commercial paper. Over the longer period from
1908 to 1984, using monthly data, there is some indication of a negative
relation between deficits and nominal or real interest rates (Evans 1987a:
326 Robert J.Barro
Tables 1–3). Evans also explores the effects of expected future budget deficits
or surpluses. He assumes that people would have expected future deficits in
advance of tax cuts, such as in 1981, and future surpluses in advance of tax
hikes. But interest rates turn out typically not to rise in advance of tax cuts
and not to fall in advance of tax hikes.
Overall, the empirical results on interest rates support the Ricardian view.
Given these findings it is remarkable that most macroeconomists remain
confident that budget deficits raise interest rates.
Consumption and saving
Many empirical studies have searched for effects of budget deficits or
social security on consumption and saving. Most of these studies—
exemplified by Levis Kochin (1974) and the papers surveyed in Louis
Esposito (1978)—rely on estimates of coefficients in consumption functions.
Basically, the results are all over the map, with some favoring Ricardian
equivalence, and others not.
The inconclusive nature of these results probably reflects well-known
identification problems. The analysis does not deal satisfactorily with the
simultaneity between consumption and income, and also has problems with
the endogeneity of budget deficits. For example, deficits and saving (or
investment) have strong cyclical elements, and it is difficult to sort out the
causation in these patterns. Because of these problems, I regard as more
reliable some results that exploit situations that look more like natural
experiments.
One such study, a comparison of saving in Canada and the United
States was carried out by Chris Carroll and Lawrence Summers (1987).
They note that the private saving rates in the two countries were similar
until the early 1970s, but have since diverged; for 1983–5 the Canadian
rate was higher by about six percentage points. After holding fixed some
macroeconomic variables and aspects of the tax systems that influence
saving, the authors isolate a roughly one-to-one, positive effect of
government budget deficits on private saving. That is, the rise in the
private saving rate in Canada, relative to that in the United States,
reflected the greater increase in the Canadian budget deficit as a ratio to
GNP. Thus, as implied by the Ricardian view, the relative values of the
net national saving rates in the two countries appeared to be invariant
with the relative values of the budget deficits. These results are
particularly interesting because the focus on relative performance in
Canada and the United States holds constant the many forces that have
common influences on the two countries. It may be that this procedure
lessens the problems of identification that hamper most studies of
consumption functions.
Recent fiscal policy in Israel comes close to a natural experiment for
studying the interplay between budget deficits and saving.7 In 1983 the gross
The Ricardian approach to budget deficits 327
national saving rate of 13 percent corresponded to a private saving rate of
17 percent and a public saving rate of–4 percent. In 1984 the dramatic rise
in the budget deficit led to a public saving rate of–11 percent. (A principal
reason for the deficit was the adverse effect of the increase in the inflation
rate on the collection of real tax revenues.) For present purposes, the
interesting observation is that the private saving rate rose from 17 percent to
26 percent, so that the national saving rate changed little; actually rising
from 13 percent to 15 percent. Then the stabilization program in 1985
eliminated the budget deficit, along with most of the inflation, so that the
public saving rate increased from –11 percent in 1984 to 0 in 1985–6 and –
2 percent in 1987. The private saving rate decreased dramatically at the
same time—from 26 percent in 1984 to 19 percent in 1985 and 14 percent in
1986–7. Therefore, the national saving rates were relatively stable, going
from 15 percent in 1984 to 18 percent in 1985, 14 percent in 1986, and 12
percent in 1987. The main point is that this evidence reveals the roughly
one-to-one offset between public and private saving that the Ricardian view
predicts.
Finally, I should note the ‘Reagan experiment’, which featured large US
budget deficits from 1984 to 1987 during a peacetime boom. (While an
interesting experiment—applauded on scientific grounds even by opponents
of Reagan—the magnitudes are much less dramatic than those in Israel.)
Unfortunately, the effects of recent US budget deficits on US investment and
saving are controversial, especially because it is unclear whether recent
investment and saving rates are high or low.
National accounts measures of rates of net investment and net national
saving are low, and have often been cited. But the ratio of real gross
investment (broadly defined to include purchases of consumer durables) to
real GNP averaged 27.9 percent from 1984 to 1987, as compared to an
average of 23.8 percent from 1947 to 1987. In fact, the recent investment
ratios represent a post-World War II high. If saving is measured (as I would
argue is appropriate) by the change in the real market value of assets, recent
saving rates have not been low. For example, the change in real household
net worth as a ratio to real GNP averaged 11.2 percent from 1984 to 1987,
as compared to a mean of 10.1 percent from 1949 to 1987.8 Thus, while a
good portion of recent US budget deficits may qualify as exogenous, it is not
yet clear how these deficits affected US investment and saving.
Current-account deficits
Popular opinion attributes the large current-account deficits in the United
States since 1983 to the effects of budget deficits. Figure 13.1 shows the
values since 1948 of the ratio of the total government budget surplus
(national accounts’ version) to GNP (solid line) and the ratio of net foreign
investment to GNP (dotted line).9 Through 1982 there is no association
between these two variables (correlation=-0.02). However, including the
328 Robert J.Barro
data since 1983 raises the correlation to 0.37. In effect, the US data since
World War II reveal a single incident—the period since 1983—when budget
and current-account deficits have been high at the same time. While this co-
movement is interesting, it does not by itself provide strong support for the
view that budget deficits cause current account deficits.
Evans (1988: Tables 1–5) carried out a cross-country empirical
investigation of the relation between budget and current account deficits. He
looked first at annual, post-World War II data for Canada, with the United
States used as a proxy for the rest of the world. (In a model where budget
deficits matter, the current account deficit responds to the home country’s
budget deficit relative to the budget deficit in the rest of the world.) Then he
looked at quarterly data since 1973 on the United States, Canada, France,
Germany, and the United Kingdom, with the aggregate of the major
industrialized countries (other than the country under study) used to represent
the rest of the world. Evans’s overall finding is that the results are consistent
with the Ricardian hypothesis that current-account balances are independent
of budget deficits. (Only the estimates for Germany suggest a positive
relation between budget and current account deficits, but the results in this
case are not statistically significant.) Evans reacts to his findings with the
question, ‘If large U.S. budget deficits did not produce the large U.S. current-
account deficits of the 1980s, what did?’ and concludes that this question is
an interesting topic for future research (Evans 1988:31).
Figure 13.1 US budget and current-account surpluses, 1948–87
Note: Data are seasonally adjusted, quarterly values from Citibase
The Ricardian approach to budget deficits 329
CONCLUDING OBSERVATIONS
The Ricardian approach to budget deficits amounts to the statement that the
government’s fiscal impact is summarized by the present value of its
expenditures. Given this present value, rearrangements of the timing of
taxes—as implied by budget deficits—have no first-order effect on the
economy. Second-order effects arise for various reasons, which include the
distorting effects of taxes, the uncertainties about individual incomes and tax
obligations, the imperfections of credit markets, and the finiteness of life. To
say that these effects are second order is not to say that they are
uninteresting; in fact, the analysis of differential taxation in the theory of
public finance is second order in the same sense. However, careful analysis
of these effects tends to deliver predictions about budget deficits that differ
from those of standard macroeconomic models.
I have argued that empirical findings on interest rates, consumption and
saving, and the current-account balance tend mainly to support the
Ricardian viewpoint. However, this empirical analysis involves substantial
problems about data and identification, and the results are sometimes
inconclusive. It would be useful to assemble additional evidence, especially
in an international context.
Although the majority of economists still lean toward standard
macroeconomic models of fiscal policy, it is remarkable how respectable the
Ricardian approach has become in the last decade. Most macroeconomists
now feel obligated to state the Ricardian position, even if they then go on to
argue that it is either theoretically or empirically in error. I predict that this
trend will continue and that the Ricardian approach will become the
benchmark model for assessing fiscal policy.
There is a parallel between the Ricardian equivalence theorem on
intertemporal government finance and the Modigliani-Miller (1958) theorem
on corporate finance. Everyone knows that the Modigliani-Miller theorem is
literally incorrect in saying that the structure of corporate finance does not
matter. But the theorem rules out numerous sloppy reasons for why this
structure might have mattered, and thereby forces theoretical and empirical
analyses into a disciplined, productive mode. Similarly, I would not predict
that most analysts will embrace Ricardian equivalence in the sense of
concluding that fiscal policy is irrelevant. But satisfactory analyses will
feature explicit modeling of elements that lead to departures from Ricardian
equivalence, and the predicted consequences of fiscal policies will flow
directly from these elements.
ACKNOWLEDGEMENTS
I am grateful for support of research from the National Science Foundation.
Also, I appreciate the high quality comments provided by the editors of
Journal of Economic Perspectives.
330 Robert J.Barro
NOTES
1 The calculations use the government’s interest rate in each period to calculate
present values, and assume perfect foresight with respect to future government
expenditures and taxes. For further discussion see Ben McCallum (1984) and
Robert Barro (1989).
2 The term, Ricardian equivalence theorem, was introduced to macroeconomists by
James Buchanan (1976). After Gerald O’Driscoll (1977) documented Ricardo’s
reservations about this result, some economists have referred to the equivalence
finding as being non-Ricardian. But, as far as I have been able to discover, David
Ricardo (1951) was the first to articulate this theory. Therefore, the attribution of
the equivalence theorem to Ricardo is appropriate even if he had doubts about
some of the theorem’s assumptions. As to whether the presence of this idea in
Ricardo’s writings is important for scientific progress, I would refer to Nathan
Rosenberg’s (1976:79) general views on innovations in the social sciences: ‘what
often happens in economics is that, as concern mounts over a particular problem
…an increasing number of professionals commit their time and energies to it. We
then eventually realize that there were all sorts of treatments of the subject in the
earlier literature…. We then proceed to read much of our more sophisticated
present-day understanding back into the work of earlier writers whose analysis
was inevitably more fragmentary and incomplete than the later achievement. It was
this retrospective view which doubtless inspired Whitehead to say somewhere that
everything of importance has been said before—but by someone who did not
discover it.’ (This last point relates to ‘Stigler’s Law’, which states that nothing is
named after the person who discovered it.)
3 Philippe Weil (1987) and Miles Kimball (1987) analyze conditions that ensure an
interior solution for intergenerational transfers. Douglas Bernheim and Kyle Bagwell
(1988) argue that difficulties arise if altruistic transfers are pervasive. See Barro
(1989) for a discussion of their analysis.
4 The assumption is the real debt remains permanently higher by the amount of the
initial deficit. For some related calculations, see Merton Miller and Charles Upton
(1974: ch. 8) and James Poterba and Lawrence Summers (1987: section I).
5 For discussions of the tax-smoothing model of budget deficits, see A.C.Pigou
(1928: ch. 6) and Robert Barro (1979; 1986).
6 A colleague of mine argues that a ‘normative’ model should be defined as a model
that fits the data badly.
7 I am grateful to Ed Offenbacher for calling my attention to the Israeli experience.
The data, all expressed in US dollars, are from Bank of Israel (1987).
8 Household net worth comes from Board of Governors of the Federal Reserve
System (1988). The nominal year-end figures were divided by the fourth-quarter
GNP deflator. The Federal Reserve numbers include stocks, housing, and consumer
durables at estimated market value, but bonds at par value. I made no adjustments
for households’ liabilities for future taxes associated with the government’s debt
net of assets. There is a conceptual problem here because some of this liability is
already reflected in the market values of households’ stocks, housing, and so on.
Also, the Federal Reserve’s measures of government liabilities and assets are not
well developed.
9 The data are quarterly, seasonally adjusted values from Citibase. The results are
similar if the federal surplus is used instead of the total government surplus.
The Ricardian approach to budget deficits 331
REFERENCES
Ando, Albert, and Franco Modigliani, ‘The “Life Cycle” Hypothesis of Saving: Aggregate
Implications and Tests’, American Economic Review March 1963, 53, 55–84.
Bailey, Martin J., National Income and the Price Level, 2nd edn, New York: McGrawHill,
1971.
Bank of Israel, Annual Report, Jerusalem, 1987.
Barro, Robert J., ‘Are Government Bonds Net Wealth?’, Journal of Political Economy
November/December 1974, 82, 1095–117.
Barro, Robert J., ‘On the Determination of the Public Debt’, Journal of Political Economy
October 1979, 87, 940–71.
Barro, Robert J., ‘U.S. Deficits since World War I’, Scandinavian Journal of Economics
1986, 88, no. 1, 195–222.
Barro, Robert J., ‘Government Spending, Interest Rates, Prices, and Budget Deficits in
the United Kingdom, 1701–1918’, Journal of Monetary Economics September 1987,
20, 221–47.
Barro, Robert J., ‘The Neoclassical Approach to Fiscal Policy’, in Robert J.Barro (ed.)
Modern Business Cycle Theory, Cambridge, MA: Harvard University Press, 1989.
Barsky, Robert B., N.Gregory Mankiw, and Stephen P.Zeldes, ‘Ricardian Consumers
with Keynesian Propensities’, American Economic Review September 1986, 76,
676–91.
Bernheim, B.Douglas, Andrei Shleifer, and Lawrence H.Summers, ‘The Strategic Bequest
Motive’, Journal of Political Economy December 1985, 93, 1045–76.
Bernheim, B.Douglas and Kyle Bagwell, ‘Is Everything Neutral?’, Journal of Political
Economy April 1988, 96, 308–38.
Board of Governors of the Federal Reserve System, Balance Sheets for the U.S. Economy,
1948–87, Washington, DC, October 1988.
Brown, E.Cary, ‘Fiscal Policy in the’ Thirties: a Reappraisal’, Journal of Political Economy
December 1956, 46, 857–79.
Buchanan, James M., ‘Barro on the Ricardian Equivalence Theorem’, Journal of Political
Economy April 1976, 84, 337–42.
Buchanan, James M. and Richard E.Wagner, Democracy in Deficit, New York: Academic
Press, 1977.
Carroll, Chris and Lawrence H.Summers, ‘Why Have Private Savings Rates in the
United States and Canada Diverged?’, Journal of Monetary Economics September
1987, 20, 249–79.
Chan, Louis K. C, ‘Uncertainty and the Neutrality of Government Financing Policy’,
Journal of Monetary Economics May 1983, 11, 351–72.
Council of Economic Advisers, Annual Report, Washington, DC: US Government
Printing Office, 1962.
Darby, Michael, R., The Effects of Social Security on Income and the Capital Stock,
Washington, DC: American Enterprise Institute, 1979.
Eisner, Robert and Paul Pieper, ‘A New View of the Federal Debt and Budget Deficits’,
American Economic Review March 1984, 74, 11–29.
Esposito, Louis, ‘Effect of Social Security on Saving: Review of Studies Using U.S. Time-
Series Data’, Social Security Bulletin May 1978, 41, 9–17.
Evans, Paul, ‘Interest Rates and Expected Future Budget Deficits in the United States’,
Journal of Political Economy February 1987a, 95, 34–58.
Evans, Paul, ‘Do Budget Deficits Raise Nominal Interest Rates? Evidence from Six
Industrial Countries’, Journal of Monetary Economics September 1987b, 20, 281–
300.
Evans, Paul, ‘Do Budget Deficits Affect the Current Account?’, unpublished, Ohio State
University, August 1988.
332 Robert J.Barro
Feldstein, Martin S., ‘Social Security, Induced Retirement, and Aggregate Capital
Accumulation’, Journal of Political Economy September/October 1974, 82, 905–
926.
Feldstein, Martin S., ‘Perceived Wealth in Bonds and Social Security: a Comment’,
Journal of Political Economy April 1976, 84, 331–6.
Hayashi, Fumio, ‘Tests for Liquidity Constraints: a Critical Survey and some New
Observations’, in Truman F.Bewley (ed.) Advances in Econometrics, Fifth World
Congress, Cambridge: Cambridge University Press, 1987.
Kimball, Miles S., ‘Making Sense of Two-Sided Altruism’, Journal of Monetary
Economics September 1987, 20, 301–26.
King, Mervyn A., ‘Tax Policy and Consumption Smoothing’, unpublished, London
School of Economics, April 1986.
Kochin, Levis A., ‘Are Future Taxes Anticipated by Consumers?’, Journal of Money,
Credit and Banking August 1974, 6, 385–394.
Kotlikoff, Laurence J., ‘Intergenerational Transfers and Savings’, Journal of Economic
Perspectives Spring 1988, 2, 41–58.
Kotlikoff, Laurence J. and Lawrence H.Summers, ‘The Role of Intergenerational
Transfers in Aggregate Capital Accumulation’, Journal of Political Economy August
1981, 89, 706–732.
McCallum, Ben T., ‘Are Bond-financed Deficits Inflationary? A Ricardian Analysis’,
Journal of Political Economy February 1984, 92, 123–35.
Miller, Merton H. and Charles W.Upton, Macroeconomics, a Neoclassical Introduction,
Homewood, IL: Irwin, 1974.
Modigliani, Franco, ‘Long-run Implications of Alternative Fiscal Policies and the Burden
of the National Debt’, Economic Journal December 1961, 71, 730–55.
Modigliani, Franco, ‘The Role of Intergenerational Transfers and Life Cycle Saving in
the Accumulation of Wealth’, Journal of Economic Perspectives Spring 1988, 2, 15–
40.
Modigliani, Franco and Richard Brumberg, ‘Utility Analysis and the Consumption
Function: an Interpretation of Cross-Section Data’, in K.K.Kurihara (ed.) Post-
Keynesian Economics, New Brunswick, NJ: Rutgers University Press, 1954.
Modigliani, Franco and Merton H.Miller, ‘The Cost of Capital, Corporation Finance
and the Theory of Investment’, American Economic Review June 1958, 48, 261–97.
Mundell, Robert A., ‘Money, Debt, and the Rate of Interest’, in R.A.Mundell, Monetary
Theory, Pacific Palisades, CA: Goodyear, 1971.
O’Driscoll, Gerald P., ‘The Ricardian Nonequivalence Theorem’, Journal of Political
Economy February 1977, 85, 207–10.
Pigou, A.C., A Study in Public Finance, London: Macmillan, 1928.
Plosser, Charles I., ‘Government Financing Decisions and Asset Returns’, Journal of
Monetary Economics May 1982, 9, 325–52.
Plosser, Charles I., ‘Further Evidence on the Relation between Fiscal Policy and the
Term Structure’, Journal of Monetary Economics September 1987, 20, 343–67.
Poterba, James M. and Lawrence H.Summers, ‘Finite Lifetimes and the Savings Effects
of Budget Deficits’, Journal of Monetary Economics September 1987, 20, 369–91.
Ricardo, David, ‘Funding System’, in Piero Sraffa (ed.) The Works and Correspondence
of David Ricardo, volume IV, Pamphlets and Papers, 1815–1823, Cambridge:
Cambridge University Press, 1951.
Rosenberg, Nathan, Perspectives on Technology, Cambridge: Cambridge University
Press, 1976.
Sahasakul, Chaipat, ‘The U.S. Evidence on Optimal Taxation over Time’, Journal of
Monetary Economics November 1986, 18, 251–75.
Siegel, Jeremy J., ‘Inflation-Induced Distortions in Government and Private Saving
Statistics’, Review of Economics and Statistics April 1979, 61, 83–90.
The Ricardian approach to budget deficits 333
Tobin, James and Willem Buiter, ‘Fiscal and Monetary Policies, Capital Formation, and
Economic Activity’, in George M. von Furstenberg (ed.) The Government and Capital
Formation, Cambridge: Ballinger, 1980.
US Treasury Department, The Effect of Deficits on Prices of Financial Assets: Theory
and Evidence, Washington, DC: US Government Printing Office, 1984.
Weil, Philippe, ‘Love Thy Children: Reflections on the Barro Debt Neutrality Theorem’,
Journal of Monetary Economics May 1987, 19, 377–91.
Yotsuzuka, Toshiki, ‘Ricardian Equivalence in the Presence of Capital Market
Imperfections’, Journal of Monetary Economics September 1987, 20, 411–36.
14 The new-classical contribution
to macroeconomics
David Laidler
Banca Nazionale Del Lavoro Quarterly Review (1986) March, pp.
27–55
INTRODUCTION
Macroeconomics is prone to ‘revolutions’—intellectual upheaval in which
some new idea or ideas claiming to establish fresh and valid insights into the
workings of the economic system sweep away a prevailing orthodoxy. Since
the mid-1930s the ‘Keynesian revolution’ has overwhelmed ‘classical
economies’ so-called, to be succeeded in turn by a ‘monetarist revolution’
which seemed to overthrow ‘Keynesian’ economics. Since the early 1970s
‘monetarism’ has in turn yielded to a ‘new-classical revolution’ which self-
consciously, and much more thoroughly than monetarism, has sought to
reestablish macroeconomics on foundations that bear a close resemblance to
those of certain strands in pre-Keynesian economics.1 In every case, the
superiority of the ‘new’ approach has undoubtedly been oversold by its
adherents, but, at the same time, insights and tools of lasting value have also
been added to the corpus of economic knowledge.
This chapter is devoted to assessing new-classical ideas, and to asking
what of lasting importance this school of macroeconomics has contributed
since the early 1970s. It deals in turn with the relationship between new-
classical economics and monetarism, the relative explanatory power of these
two bodies of doctrine over empirical evidence, and the claims of new-
classical economics to embody a superior analytic method. It argues that,
although the particular ways in which new-classical macroeconomics has
applied its basic ideas, notably in its insistence that the interaction of the
maximizing behaviour of individuals be analysed in the context of
continuously clearing markets, and that agents’ expectations be represented
by the predictions of the true model of the economy in which they operate,
are unnecessarily restrictive, its stress on equilibrium behaviour conditioned
by the state of individual agents’ expectations as a basis for macro modelling
is nevertheless valuable, and has been salutary for the discipline.
MONETARISM AND NEW-CLASSICAL MACROECONOMICS
New-classical macroeconomics was initially a response to the inflation of the
1960s and 1970s, and to monetarist analysis of that inflation. Indeed, in its
New-classical contribution to macroeconomics 335
earliest manifestations, it appeared to be nothing more than an attempt to
restate monetarist analysis with greater rigour than its pioneers—notably
Milton Friedman—had achieved.2 In order to put matters in perspective it
will be helpful to recall the nature of the intellectual problem which that
inflation created for most macroeconomists. Quite simply the empirical
evidence it generated proved to be utterly inconsistent with then prevailing
Keynesian views about how the economy worked, and about how policy
could be used to improve its performance. Expansionary demand side
policies, predominantly fiscal, could, according to that orthodoxy, generate
lasting reductions in unemployment at the cost of somewhat higher, but
nevertheless stable, inflation. When the Keynesian experiment occurred, it
failed.3 Gains in output and employment, where they materialized at all,
proved to be temporary, and inflation, instead of shifting once to a new
higher level, rose continuously.
Monetarist macroeconomics (whose components were available before the
event, be it noted) explained these facts by arguing: first that Keynesian
orthodoxy had underestimated the role of the quantity of money as an
influence on aggregate demand in general and the behaviour of prices in
particular; and second that the idea of a stable inflation-unemployment
trade-off—the Phillips curve—was based on an implicit assumption that the
private sector of the economy suffered from perpetual money illusion. To the
pressure of aggregate demand as a proximate influence on the inflation rate,
Friedman (1968)—not to mention Phelps (1967)—added the expected rate of
inflation. Furthermore, because Friedman viewed inflation expectations as
deriving from past experience, and as being formed in such a way that
expectations would in fact come to catch up with experience eventually, he
argued that any attempt to reduce the unemployment rate below that
determined by the normal frictions inherent in the labour market would lead,
in the long run, not to higher, but to rising, inflation.
From the point of view of policy prescriptions and empirical judgements
about the reliability of particular functional relationships in the economy,
monetarism presented a clear alternative to Keynesian orthodoxy, but
constituted no radical theoretical challenge to it. Keynesian models already
contained a demand for money function, and if monetarism was correct in
arguing that this relationship was more stable than had in the past been
believed, such a modification could easily enough be accommodated.4 If
expected inflation belonged as an extra variable in the Phillips curve, and
depended upon the past behaviour of inflation, that would alter one’s view
of what demand management policy could accomplish, but it did not
require any fundamental change in economists’ vision of how the economy
worked. There is no stronger evidence in favour of the latter judgement
than the fact that the first explicit monetarist analytic models were
recognizable extensions of the IS-LM model.5 Moreover large-scale
Keynesian econometric systems proved easily able to absorb monetarist
ideas as well.
336 David Laidler
The difficulty here was that the new version of the Phillips curve was
hardly more satisfactory than the old one from an analytic point of view.
Though the proposition that money wages and prices tend, given
expectations, to rise faster the higher the level of aggregate demand in the
economy, might be a plausible enough empirical generalization, it does not
constitute an explanation of the phenomenon which relates it to the
purposeful maximizing behaviour of individual economic agents. The
monetarist ‘expectations-augmented Phillips curve’ was an empirical
observation in need of an explanation, not a well grounded structural
relationship in its own right. In attempting to provide an explanation of it,
new-classical economists, and in particular Robert E.Lucas Jr (1972), set in
motion the ‘New-Classical Revolution’, based upon two analytic devices,
namely the aggregate supply curve and the rational expectations hypothesis.6
Though the rational expectations idea has probably attracted more attention,
it is its use of a particular version of the aggregate supply curve which
constitutes the most fundamental innovation of new-classical economics.
Keynesian macroeconomics (including its monetarist variation) can
accommodate rational expectations, but it cannot be reconciled with the
universal existence of the continuously clearing flexible price competitive
markets which are a sine qua non of the ‘aggregate supply curve’
explanation of the Phillips curve.
Sticky prices lie at the very heart of Keynesian macroeconomics, and it
explains quantity fluctuations in goods and labour markets as equilibrating
movements arising because prices do not immediately change when
aggregate demand shifts. The postulate of price flexibility lies at the centre
of new-classical economics. It has it that prices always move to equilibrate
markets when demand shifts, but that individual agents, who are not fully
informed about the behaviour of all money prices in the economy, mistake
money price changes in the markets for the goods they sell for relative price
changes. Hence they respond by changing the quantities of goods they
supply. In the aggregate, an unperceived demand increase which raises the
general price level therefore causes an expansion of output along an
aggregate supply curve, and a fall of demand causes a contraction. Output
and employment fluctuation such as we observe in the real world are,
according to new-classical economics, voluntary responses to misperceived
price signals. They occur because prices change. Keynesian economics
(including its monetarist variant) explains quantity changes as occurring
because prices do not change fast enough to keep markets cleared. In this
vital matter the contrast between the two approaches could not be more
stark.7
Now the clearing markets hypothesis of new-classical economics is
logically compatible with the idea that expectations are naively extrapolated
from past experience, but the use of the two ideas in conjunction certainly
strains credulity. If agents are in no way tied down by sticky prices, and
make costly errors in quantity decisions because of faulty expectations about
New-classical contribution to macroeconomics 337
the behaviour of prices in markets other than those in which they are
currently active as sellers, they have every incentive to make their
expectations as accurate as possible, and to use all available information in
order to do so. Maximizing agents should be presumed to form expectations,
as Sargent and Wallace (1973:328) put it, so that they ‘depend, in a proper
way, on the same things that economic theory says actually determine that
variable’. Hence, though the literature of the 1960s and early 1970s does
contain examples of models which combine clearing markets with adaptive
expectations, such hybrids soon vanished to be replaced by a substantial
body of new-classical theory, based upon the twin hypotheses of clearing
markets and rational expectations.8
THE CASE FOR NEW-CLASSICAL MACROECONOMICS
Economists have no clearly agreed criteria for deciding among competing
bodies of theory, but certain factors are widely accepted as being relevant.
The ability to explain past events, or (even better) to forecast future ones, is
highly valued, as is the closely related capacity to yield insights into the
nature of policy options available, and into the likely outcome of whichever
option is chosen. Also important are matters of logical coherence, and
intellectual compatibility with other available and accepted doctrines.
Proponents of new-classical macroeconomics have, at various times, claimed
it to be superior to Keynesian and monetarist alternatives on all three
criteria.9
As will already be apparent, I quite agree that the western world’s
experience with inflation and unemployment of the 1970s constitutes a
massive refutation of ‘Keynesian economies’ as the term was understood in
the mid-1960s. Nor would I deny that the new-classical macroeconomics of
the late 1970s, emphasizing as it did the role of the quantity of money in
generating inflation, and the crucial role played by expectations in the
inflationary process, provided a superior explanation of that experience. If
we were forced to make a choice between these two alternatives alone, we
would have to accept the claims of Lucas and Sargent (1978) that their brand
of macroeconomics is the only respectable one available. However, we are
not forced to make this choice.10
Before the inflation of the 1970s was dreamed of, monetarists, such as
Friedman (e.g. 1959) and Brunner and Meltzer (e.g. 1963) had been
attacking Keynesian orthodoxy for underestimating the importance of the
quantity of money. Furthermore, Friedman and Phelps (surely no monetarist)
had, as we have seen, criticized the idea of a permanent inflation-
unemployment trade-off in the mid-1960s on the grounds that the behaviour
of inflation expectations, themselves endogenous to the structure of the
economy, would render any such trade-off temporary. As we have also noted,
however, these ideas could be, and eventually were, easily incorporated into
otherwise orthodox Keynesian models, but Keynesian models so modified do
338 David Laidler
very well indeed in explaining the 1970s.11 A system in which prices are
sticky (though not rigid), in which quantities change to absorb demand side
shocks in the short run, and in which inflation expectations though mainly
backward looking, are endogenous, can account for the 1970s experience at
least as well as any new-classical system based on price flexibility, clearing
markets and rational expectations. To put it in terms of labels, the empirical
experience of the 1970s does not force one to reject the ‘monetarist’ variation
on the ‘Keynesian’ model and embrace ‘new-classical macroeconomics’.
The methodological criteria proposed by new-classical economists in
defence of their work have much in common with those sketched above, and
implicitly or explicitly adopted by economists in general. If they did not, it
would be hard to explain why their arguments have proved so widely
persuasive. However, though claims to superior predictive power, and to
deeper insights into the nature of economic policy processes, have certainly
been made from time to time on behalf of new-classical macroeconomics, it
has also, from the very outset, been presented as the product of a major
advance in the application of analytical methods; and, with the passage of
time, its proponents have come to place increasing emphasis on this last
factor, claiming that their macroeconomics is more logically coherent and
more closely related to micro theory than anything which went before it. It is
certainly true, as we shall now see, that these are the strongest arguments in
favour of new-classical economics.
To begin with, and uncontroversially, new-classical economists tell us that
an important purpose of macroeconomic models is to deduce predictions
about the behaviour of an economy when subjected to various shocks.
Equally uncontroversially, they argue that key components of such a model
should be logically coherent and well tested propositions about the behaviour
of individual agents. That these propositions about individual agents should
in turn be derived from analysis of rationally purposeful utility maximizing
behaviour might be less universally accepted, but I do not wish to quarrel
about this particular principle.12 Reasons for controversy begin to arise only
when we seek an institutional framework in terms of which it is possible to
derive coherent predictions about the behaviour of the economy as a whole
from knowledge of individual behaviour, and I shall argue in due course that
the particular choice made at this point by the new-classicals is not the only
respectable one available to us.
Be that as it may, new-classical economists propose that we model agents
as operating in an environment of perfect competition, in which markets
costlessly adjust to maintain the supply and demand for every good and
service, not least labour, in constant equilibrium. Their competitive model
differs from traditional treatments of perfectly competitive economies
inasmuch as agents in it do not have full information about the structure of
relative prices when they engage in trade. The demand and supply schedules
which determine the equilibrium structure of market prices in a new-classical
model are conditional, not upon full and accurate information about that
New-classical contribution to macroeconomics 339
same structure of market prices, but upon agents’ perceptions (expectations is
the more commonly used word) of that structure. Because agents are
supposed to be purposeful rational maximizers, they form their expectations
so that they differ from the actual values of the variables in question only to
the extent of a serially uncorrelated random error. For agents to operate on
the basis of any other kind of expectations would result in them encountering
unnecessary losses, and hence in violating the purposeful utility
maximization assumption.
The ‘rational’ approach to modelling expectations formation has been
translated by new-classical economists into the postulate that agents form
expectations ‘as if they were fully informed about the structure of the economy
in which they operate, and make mistakes only to the extent that the economy
is subjected to random exogenous shocks, either in the form of ‘policy
surprise’—any systematic component of policy behaviour being, and being
perceived to be, part of the economy’s structure,—or in more recent literature,
random fluctuations in technology, ‘real shocks’ as they are called. In such a
framework, given currently (and only rather recently) available analytic
techniques, it is possible to derive predictions about the aggregate behaviour of
the economy directly from premises concerning individual behaviour. More to
the point, these predictions in certain important ways mimic the behaviour of
real world economies, specifically in the matter of co-movements of money
wages and prices and quantities of employment and output over the course of
the business cycle, and indeed the very fact that new-classical macroeconomics
involves the exploitation of these new analytic techniques is sometimes
advanced as an argument in its favour.13
The really critical point, however, as far as the proponents of new-
classical economics are concerned, is that the above-mentioned analytic
techniques, in their current state of development, can be used to derive
macro-predictions with empirical content from nothing but well specified
micro-premises only on the assumptions of representative agents operating in
competitive markets cleared by flexible prices. A model which postulates
some form of wage or price stickiness inevitably involves the use of some
(allegedly) ad hoc element in forming the link between micro-postulates and
macro-predictions. This is not because there do not exist models of
individual maximizing behaviour that explain price stickiness, because there
obviously do, but because our current analytic capacity does not permit us
except in exceptionally simple examples (e.g. Howitt 1981) to embed such
behaviour in a model of the economy as a whole, to allow for the way in
which such behaviour might influence expectations, and then explicitly to
derive macro predictions.
As a result, those who wish both to postulate phenomena such as price
stickiness and to build models with empirical content, are led to introduce
qualitative empirical ‘laws’ into them and to permit the data to find
quantitative values for the parameters which characterize these ‘laws’. One
way of looking at the issues at stake here is in terms of alternative strategies
340 David Laidler
for evading that perennial barrier to truly rigorous macroeconomics, the
aggregation problem. The new-classical assumptions of representative agents
plus perfect competition certainly permit clearly defined links to be
established between individual and market experiments without recourse to
empirical laws, but those links are only as defensible as the assumptions that
permit them to be forged.
Even so, if we regard the presence of ‘free parameters’, as Lucas (1980)
calls them, in a model to be a fatal drawback, then new-classical
macroeconomics, with its assumptions of universal competition among
representative agents, perfect price flexibility, and rational expectations, has
no rivals. If it is objected that perhaps empirical evidence might nevertheless
have a role to play in such a judgement, the answer offered by the
proponents of new-classical economics, notably Lucas (1980), is that, since
their basic model uses no ‘free parameters’, a model which fits the facts
better, or at least as well, can always be constructed by adding one (or more)
such parameter to a basic new-classical system. Economic models are not
supposed to be descriptions of all elements of reality (whatever that might
be); and to show that greater descriptive accuracy may be achieved by the
addition of free parameters is said to be neither surprising nor compelling as
an argument against new-classical economics. I shall now turn to an
examination of this argument.
EMPIRICAL EVIDENCE AND ‘FREE PARAMETERS’
I remarked earlier that there is no completely agreed set of methodological
criteria for judging economic models. As a matter of simple logic, it cannot
be denied that, if rigorous connections between maximizing premises and
ultimate conclusions is regarded as the be all and end all of economic
analysis, then new-classical macroeconomics is indeed the only game worth
playing. The most that individuals who deny this viewpoint can do is
explain why they think that other criteria are relevant, show how the criteria
support their position, and hope that their reasoning will be taken seriously.
Such is my purpose here.
My starting point is that the ultimate aim of economic theory is to explain
observations, in the sense of deducing statements which describe such
observations from more general premises. Moreover, and quite crucially,
such premises should also yield other statements whose truth is not
contradicted by the facts. The more general the predictive power of a set of
premises (and the more propositions about purposeful maximizing behaviour,
and the fewer theoretically unsupported generalizations relying upon ‘free
parameters’ there are among them) the better. An economics which can
deduce true predictions about all the phenomena that might interest us from
nothing but premises about maximizing behaviour is presumably the ideal
towards which we are all striving. That we are unlikely to achieve this ideal
is not the point, though. Rather it is that, even if we did stumble upon it, we
New-classical contribution to macroeconomics 341
could never know this. The most we can ever be sure of about our models is
that they have not been contradicted by evidence gathered to date. In the
very nature of things we can never know that they are true in the sense that
they never will be contradicted.
As a practical matter we must always be more concerned with criteria for
choosing among less than ideal theories than with laying down unattainable
and non-operational standards of theoretical perfection. For this rather
humdrum task, primacy must be accorded to empirical evidence, because it
is surely uncontroversial that a theory which makes systematically false
predictions about some phenomenon is itself false, and in need of
modification, no matter how closely it satisfies other criteria.14 Even so we
must be careful when we advance this last proposition not also to demand
that a theory’s predictions be ‘descriptively accurate’. A theory may abstract
from all manner of phenomena, have nothing to say about them, and hence
be ‘descriptively inaccurate’ (or incomplete), but that does not make it false.
The question of falsity only arises when a theory yields definite predictions
about some phenomenon which turn out to be untrue. Descriptive inaccuracy
is an inherent quality of any abstract model; but falsity is not. To use a
standard platitude of the elementary logic class as an illustration, the reason
why the proposition ‘all swans are white’ is false is not that this statement
fails to mention feathers, and into the bargain has nothing to say about
ducks; rather it is that some black swans do exist.
My reason for denying the inherent superiority of new-classical
macroeconomics is not, therefore, that there might be interesting facts from
which it abstracts and about which it has nothing to say; rather it is that it
makes false predictions about the very phenomena with which it purports to
deal, and that if it is to be rescued, parameters every bit as ‘free’ as those
utilized in the Keynesian (or monetarist) alternative seem to be required. The
original task which new-classical economics set itself was to provide a
foundation in qualitative microeconomic reasoning for Friedman’s
propositions about the temporary nature of the inflation-unemployment
trade-off. The fact that it succeeded in doing so is, however, not an empirical
argument in its favour. That statements describing a set of already known
facts may be deduced from a model is evidence, not of its truth, but of the
logical skills of the person who constructed it. An empirical test arises only
when conclusions yielded by the same model about facts not used to
discipline its construction, and better still, initially unsuspected, are
compared with those facts.15 Here, new-classical economics finds itself in
trouble.
To begin with, it gets rid of the free parameter linking money wage and
price changes to ‘excess demand’ by postulating that the Phillips trade-off
reflects, among other parameters of the system, the elasticity of the supply of
labour with respect to real wages. In doing so it yields a testable prediction
about the quantitative relationship between inflation and employment
fluctuations. Empirical evidence shows that the relative amplitudes of those
342 David Laidler
fluctuations do not square up with what we think we know from micro-
studies about this supply elasticity. Aggregate employment fluctuations seem
to be systematically much too large relative to inflation fluctuations to be
treated as movements along a supply curve of labour when the labour force
misperceives nominal wage changes as reflecting real wage changes, and
hence to be accounted for along new-classical lines. Closely related, the
nature of the interaction of employment and real wages over the business
cycle is hard to reconcile with the new-classical postulate that the real wage
is always equal to the marginal product of labour and that employment
fluctuations involve movements along a downward sloping marginal product
schedule.16
In a new-classical world, quantities change because prices fluctuate.
Output and employment should therefore vary at least simultaneously with
(or perhaps lag behind) the price level; but it is a stylized fact of real world
business cycles that quantity changes seem to precede associated price level
changes. Moreover, if the price level is free to move to keep the supply and
demand for money in equilibrium, the economy should always be on its
long-run demand for money function; but empirical observations suggest that
the economy is often and systematically ‘off this relationship for extensive
periods of time.17 In the early 1980s, predictions about all of these
phenomena were put to the test in one real world experiment which was
surely just as damaging to the new-classical economics of the 1970s as the
experience of the 1970s was to the Keynesian orthodoxy of the 1960s. Then,
in a number of countries, sudden, but nevertheless well publicized, monetary
contractions were followed by unusually low real balances (relative to the
values of the variables determining their demand), rapid and severe output
and employment contractions, and only later by price and money wage
responses; according to new-classical economics they should have generated
price changes on the spot, and, being well publicized, only a rather mild
quantity response.18
The new-classical economist does of course have answers to all of these
questions. To begin with, monetary contraction will only have its major
effect on prices if it is expected that the authorities will persist with such a
policy. The policy must, that is to say, be credible if it is to influence
behaviour by way of its effects on expectations. In a new-classical model the
less credible is a policy, the more will the price level changes it generates be
misread for relative price changes, and the larger will be the quantity
responses. Perhaps policy was not, despite the publicity, credible in the early
1980s. As to the arrival of quantity changes before price level responses this
could have been the result either of our observations of the price level being
unreliable, because they are based upon posted prices rather than those at
which trade ‘really’ took place, or because the downturn in question did not
stem from monetary contraction after all, but from some exogenous
contractionary shift on the supply side of the economy. Why were economies
apparently ‘off’ their demand for money functions? Perhaps these functions
New-classical contribution to macroeconomics 343
were estimated using data that only imperfectly measure the true variables
upon which the demand for money depends. In this case, an apparent
departure of the economy from its demand for money function might be an
illusion created by measurement error.19
It may, of course, be that all of these propositions have some truth to
them, but it is also the case that they offer to the new-classical economist a
rich array of free parameters with which to rescue his model from empirical
evidence. How fast, and by what mechanisms does any policy become
credible? How can we test propositions about measurement error when they
result from our inability to observe the true variables? How are we to
allocate responsibility for a particular cyclical turning point between
demand side and unobservable supply side factors without referring to the
timing and amplitudes of price and quantity fluctuations? The point of all
this is not to suggest that new-classical macroeconomics is unique in relying
upon ex post ad hoc postulates about the values of free parameters to
reconcile it with empirical evidence. The criticisms which its adherents
advanced of alternative approaches for using free parameters were not
without merit. The point is rather that new-classical economics appears to be
in the same trouble as these alternative approaches, because it can avoid
recourse to free parameters for just so long as it avoids confrontation with
empirical evidence, and no longer. That can hardly be comfortable for
proponents of an approach whose major claim to superiority lies in a claim
that it avoids such problems.
Perhaps the new-classical economist would answer the foregoing argument
with a ‘so what?’ After all, Lucas (1980) did tell us that the addition of free
parameters to a new-classical model would indeed improve its predictive
performance. This answer will not quite do, however. A Keynesian (or
monetarist) model, to the extent that it relies on expectations, must also face
up to problems concerning the credibility of policy and hence is no
improvement upon a new-classical system in this respect. However, it can
dispense with conjectures about unobservable supply side shocks,
measurement error, and such, when confronted with the data. If we add the
postulate of price stickiness to an ordinary full information Walrasian
general equilibrium framework, we may model the occurrence of quantity
movements in advance of price changes in the face of demand side shocks to
the economy as an equilibrating mechanism, and we have no difficulties in
generating persistence over time in fluctuations in real variables, including
real balances. Nor do we have to puzzle over the relative magnitudes of
price-quantity fluctuations. The empirical puzzles which require new-
classical economics to add free parameters do not, that is to say, arise in the
Keynesian framework it seeks to supplant, once a free parameter
characterizing price stickiness is allowed to do its work.20
The choice here is between two models, one of which (the new-classical
model) happens to yield predictions about output fluctuations without resort
to free parameters, and one of which (the monetarist version of the
344 David Laidler
Keynesian alternative) does not; and it would be an easy one to make if
other predictions yielded by the new-classical model were empirically
supported, but, as we have seen, they are not. The choice between new-
classical and Keynesian economics is thus a choice about which free
parameters to use and at what stage in the analysis to deploy them when
modifying a standard full information Walrasian model. It is not about
whether to do without them or not.
THE PRICE STICKINESS POSTULATE
In the light of the preceding discussion, the monetarist variant of the
traditional Keynesian model begins to look attractive. Moreover as I shall
now argue, its attractiveness is further enhanced by the fact that the free
parameters it utilizes are rather harmless, linking as they do rates of change
of money wages and prices to the levels of ‘excess’ demand and supply in
particular markets.21 To begin with, though Keynesian theory does not tie
down the parameters in question to any particular quantitive value, they are
nevertheless not left to take on whatever value might be needed to reconcile
a model with any data it might encounter. These parameters are at least
required to take a non-negative sign, thus ruling out a rather wide variety of
logically possible observations whose real world occurrence would therefore
refute the Keynesian model.
More important, the price stickiness postulate amounts to a good deal
more than an unfounded ex post and ad hoc rationalization of otherwise
inexplicable observations about the interaction of quantities and prices over
time. It is, at the very least, a descriptively accurate empirical generalization
whose truth is quite independent of any macreoconomic observations. In the
real world, pricing in many branches of the labour market is characterized
by contracts which set terms for money wages and endure for rather long
time periods; similar long term contracts, also negotiated in terms of money,
do characterize many final output markets as well; the contracts in question
are not all negotiated at the same time, and they do overlap; it does follow
from these facts that, in the aggregate, money wage and price levels will
display just the kind of stickiness with respect to demand changes that
Keynesian macroeconomics postulates; and it also follows that quantities
will indeed fluctuate, as Keynesian economics says they will, instead of
prices. That is what the work of Fischer (1977), Phelps and Taylor (1977)
and Okun (1981), among others, is all about.
Moreover, the micro-economics literature does enable us to explain wage
and price stickiness in terms of maximizing behaviour. Barro (1972) and
Kawasaki et al. (1983), among others, invoke costs of changing prices as a
reason for the phenomenon. There exists a literature, surveyed by Hall
(1980), which explains wage stickiness as the outcome of contracts designed
to share the risks inherent in demand fluctuations between firms and their
employees. Mancur Olson (1984) has argued that the existence of
New-classical contribution to macroeconomics 345
rent-seeking coalitions in the market sector of the economy is likely to be
associated with wage and price stickiness, for the simple reason that such
coalitions find it easier to monitor the pricing behaviour of their members
than to enforce agreements about quantities.
What then is the difficulty about accepting wage and price stickiness? The
problem is that, though it is easy enough to explain the existence of sticky
wages and prices at the level of the individual experiment, it has not, thus
far, proved possible to explain why the stickiness in question should
characterize money wages and prices as opposed to relative wages and
prices. Thus Barro (1977b) purported to show that optimal contracts should
be concerned with relative prices, and argued that models dealing with them
cannot therefore be used to explain money wage and money price stickiness.
Since contracts set in money terms do exist in the real world, the correct
inference to draw here is that there must be something missing from the
particular maximizing models that deny their occurrence. Incredibly, new-
classical economists seem to have concluded that the maximizing models
must be correct, that the facts about contracts cannot be what they patently
are, and that they therefore must not be used as a basis for an empirical
generalization which, when inserted into macro-economic model, helps it to
yield useful predictions about the world.22
Now, if the claim of new-classical economics to be able to deduce
everything with which it deals from nothing other than fundamental premises
about tastes and technology were true, the reluctance of its proponents to use
an unexplained empirical generalization about contracts being set in terms of
money would be understandable. However, quite apart from its need for
‘free’ parameters already discussed, new-classical economics also requires us
to accept important unsupported assertions about institutional arrangements.
Consider: in every new-classical model agents trade, but the existence of
trade presupposes a system of property rights and legal arrangements
permitting their exchange; and new-classical models are frequently used to
analyse policy problems of one sort or another, but the existence of policy
presupposes both that a government of some description exists, and that this
institution has a capacity for purposeful behaviour.
We might prefer it if we could explain the existence of these social
institutions as the outcome of the maximizing behaviour of the individuals
who inhabit the economy. However, we do have to start somewhere, and our
inability to explain social institutions as the consequences of individual tastes
and technology should not prevent us from getting on with our economics.23
Precisely: but what is monetary exchange, including the practice of
contracting in money terms, if not a social institution on the same level as
property rights, markets, and government? And why should our inability to
explain it prevent us assuming it as a starting point for certain pieces of
economic analysis? However, if we do treat monetary exchange as such a
starting point, we can of course explain money wage and price stickiness in
terms of the analysis invoked above.
346 David Laidler
To sum up, the assumption of price stickiness used in conventional
Keynesian macroeconomics does permit a degree of freedom in the
determination of certain parameter values that is larger than ideal.
Moreover, we do not, in the current state of knowledge, have a full
understanding of the phenomenon. However, given the institution of
monetary exchange, money wage and price stickiness can be explained as
the result of maximizing behaviour; they do exist at the micro level, and
they do have certain implications for macroeconomic phenomena that
appear to conform to the facts. Given the choice, therefore, between a
macroeconomics which recognizes the existence of price stickiness and one
which refuses to do so, there does not seem to be very much harm done if we
opt for the former, particularly since the alternative approach also seems to
rely on a good share of free parameters and unexplained institutional
assumptions to get results with non-falsified predictive content.
RATIONAL EXPECTATIONS
The notion that the world may, and indeed ought, to be modelled as if the
activities of individual agents were co-ordinated in continuously clearing
flexible price competitive markets is one foundation of new-classical
economics. The other is the rational expectations hypothesis. The idea that
expectations about the future behaviour of prices must be important
determinants of current market behaviour is an old one, as is the closely
related proposition that, only if such expectations are fulfilled, can the
economy be said to be in full equilibrium.24 In extending these notions by
arguing, first, that we should think of expectations as being the output of an
economic model, knowledge of whose structure is attributed to agents, and
second, that for full equilibrium to rule, the model in question must be the
‘true’ one of the economy under analysis, new-classical economics has made
a contribution of immense importance to our understanding of these matters.
Economic theory has been permanently changed by these insights, and for
the better.25 That being said, I am not enthusiastic about the way in which
new-classical economists have applied these insights. Two issues in
particular are worth considering, the first having to do with the choice of the
‘model’ of the economy which one attributes to agents in analysing their
behaviour, and the second having to do with interaction between policy
authorities and the private sector, and specifically the way in which the
question of ‘credibility’ is handled.
For analytic exercises designed to reveal the long run equilibrium
properties of economic models, it is of course quite appropriate to attribute
to agents within the model knowledge of that same model. Any other basis
for expectations formation would, under some condition or other, lead agents
into systematic error, causing them to revise their method of forming
expectations. Hence, it could not be a component of a full equilibrium
structure. To say this, however, is not to say that this same procedure is
New-classical contribution to macroeconomics 347
appropriate as a foundation for applied work on any particular historical
episode.26 If it is true that expectations should ‘depend, in a proper way, on
the same things that economic theory says actually determine that variable’,
then surely, when trying to understand the behaviour of a particular
economy at a particular time in its history, we should attribute to agents
expectations based, not on what we now believe is the proper model of that
economy, but rather on what the economic theory available and believed at
that time and place said was a proper model.
We may illustrate this proposition with a concrete example. Among the
seminal papers of new-classical economics are empirical studies, by Robert
J.Barro (1977a; 1978), of the influence of money on unemployment, output
and prices in the United States since the Second World War. It is the essential
claim of these papers that only ‘unanticipated’ changes in the quantity of
money affected employment and output (relative to trend) over this period;
agents inhabiting the economy at that time are treated by Barro as believing
in the equilibrium competitive model of new-classical economics,
supplemented by a primitive version of the quantity theory of money, and as
using this model for forming their expectations.27 However, if, in the 1945–
76 period agents really had held new-classical beliefs, there would have been
no need for a new-classical revolution. As it is, we know very well that until
the mid-1970s, firm beliefs in a certain kind of Keynesian economics, whose
centrepiece was a permanent inflation unemployment trade-off, were the
common property of US policy-makers and key private sector agents alike.
Indeed the primary claim made by Lucas and Sargent (1978) to support the
scientific importance of their work was that it had undermined just this
Keynesian consensus. That being the case, logical consistency requires new-
classical economics to model the economic history of the period in question
by postulating that agents operating within the US economy used an
erroneous Keynesian model to form their expectations. To do otherwise
would be to wind up in a hopeless logical tangle.
The point illustrated here is of course quite general. New-classical
economics argues, with great persuasiveness, that the nature of agents’
information about the structure of the economy is itself an important
component of that structure. If that information changes, then so does the
economy’s behaviour. If it is right so to argue, then the state of economic
knowledge itself becomes a key ingredient of any economic model, and
economic history cannot be studied without recourse to the history of
economic thought. This latter insight is not new, of course. It is central to the
kind of Austrian economics associated particularly with the later work of
von Hayek, but he was led to this position from a starting point very similar
to the stance of contemporary new-classical economists.28 The fact that the
latter insist that agents, living at any time or place, should be thought of as
believing that the economy which they inhabit behaves ‘as if’ it was driven
by the mechanisms highlighted by a theory first advocated by a particular
group of American economists in the 1970s, certainly sets them apart from
348 David Laidler
the later Austrians. The comparison here is hardly in favour of new-classical
economists, however.
A similar type of unhistoric naïveté is to be found in the way in which
new-classical economics approaches the problem of ‘policy credibility’. It is
undoubtedly true that, in a new-classical world, a well-publicized change in,
say, monetary policy, will have its effects concentrated on prices only if the
publicity is believed. Just as traditional Keynesians—though their ancestors
here are Meade and Tinbergen, not Keynes—viewed the policy-maker’s task
as the maximization of a social utility function subject to a constraint given
by the structure of the economy, so new-classical economists regard the
typical private sector agent as maximizing a private utility function subject
to a structure determined both by the activities of other private sector agents,
and by the activities of policy-makers. Suppose that both policy-makers and
private sector agents are aware of this: how do they interact? The answer,
we are told, will be found by the application of ‘differential game theory’ in
which policy-makers and private sector agents communicate and establish
credibility with one another solely through observable behaviour.29
Ultimately in such games an ‘equilibrium’ emerges in which each agent’s
maximizing behaviour imposes a constraint on the other which leads to that
behaviour being sustained. Analysis of this type is intellectually challenging,
but a little scepticism about its empirical relevance is surely in order.
‘Policy-makers’ in the real world are not entities who exist outside of their
society and economy. They are endogenous self-interested maximizing
agents. Moreover, they interact with the private sector in many more ways
than by giving and receiving market signals to establish their credibility. In
particular, they achieve the positions that they do, and maintain them, as the
result of political processes in which private sector agents participate. A
whole literature in the area of ‘public choice’ analysis is devoted to all of
this, and I am not saying anything novel in drawing attention to these
matters. 30 I am however suggesting that to rest one’s analysis of
macroeconomic policy making on ‘differential game theory’ is simply to
ignore this critical dimension of the policy-making process. Perhaps political
institutions have nothing to do with the way in which policy is made and
changed; perhaps ideology has no influence here either; but I doubt it.
Rather, I suspect that the new-classical approach to the analysis of policy-
making, in ignoring these factors, threatens to lead us down a blind alley.
CONCLUDING COMMENTS
The bulk of this chapter has been critical of new-classical macroeconomics.
This does not mean that such analysis has nothing of importance to say to
us; quite the contrary. Though the ‘new-classical revolution’ has had
exaggerated claims made on its behalf, and it is these exaggerated claims
which I have been concerned to criticize in this chapter, it is also the case, as
noted at the very outset of this chapter, that ‘revolutions’ in macroeconomics
New-classical contribution to macroeconomics 349
usually leave behind them contributions of lasting importance to be absorbed
into the mainstream of the discipline. The new-classical revolution has
certainly done this, as I shall now argue.
Consider first the new-classical insistence on equilibrium modelling. If it is
desired to construct an economics with predictive content, then the postulate
that agents formulate purposeful and consistent plans and that they are able
to execute those plans is surely a useful starting point; but at the level of the
individual, the execution of such plans is precisely what we mean when we
speak of equilibrium behaviour. If assumptions about the nature of plans do
not permit us to say anything about actions, as they can not if we entertain
the possibility of ‘disequilibrium’ at the level of the individual agent, then an
economics based on the analysis of the individual can have no predictive
content. This idea is an old one, to be sure, having been a constant theme in
Austrian economics from Menger (1871) onwards, but a glance at the
macroeconomics literature of the 1960s will soon confirm that we had lost
sight of it, and needed to be reminded of its importance. New-classical
economics did just that.
The difficulty with new-classical economics lies not in the equilibrium
postulate per se, but in its insistence that we model the economy as a whole
as if the equilibrium strategies of individuals were formulated and executed
in an institutional framework characterized by continuously clearing
competitive markets. The fact that such a framework is the only one which,
in the current state of analytic techniques, permits a seamless connection
between the analysis of the microeconomic equilibrium of the individual and
macro behaviour, is no reason for insisting that macro-predictions obtained
by other less pristine methods are unworthy of consideration. That, though,
is what new-classical macroeconomics have, quite unjustifiably, been doing.
However, we ought not to let dissatisfaction with a particular application of
a methodological precept lead us to underestimate its general importance.
Equilibrium modelling of individuals surely ought to be the basis of
macroeconomic reasoning, and the fewer empirical generalizations about
behaviour we need to make to get from such a basis to empirically robust
predictions about the economy as a whole, the better.
Exactly parallel arguments to these may be advanced about the rational
expectations idea. This is hardly surprising, since there is a real sense in
which this hypothesis is simply a particular consequence of the purposeful
maximizing postulate. The idea that the state of agents’ knowledge, and the
nature of their expectations about future events, form a key part of the
economy’s current structure, and help to determine the outcome of current
maximizing behaviour, is hardly new. It was, as has been pointed out, a
prominent ingredient of Austrian economics, but once more, a glance at the
macroeconomic literature of the 1960s (replete as it is with exercises in
which the consequences of alternative policy measures are derived from the
same, allegedly structural, representation of the private sector of the
economy) will show how badly we needed to be reminded of this insight.
350 David Laidler
As with the equilibrium idea, criticisms of the rational expectations notion
advanced above have been of the particular and very special ways in which
it has been applied, rather than of the basic idea itself. It is at best logically
dubious to analyse historical episodes ‘as if’ agents involved in them
possessed a vision of the economy which has been created only since the
mid-1970s. When the very purpose of the analysis in question is to expose
flaws in the economics which was commonly believed during the episode
under analysis, perhaps stronger epithets are called for. Nevertheless it is
important to formulate hypotheses about the way in which the state of
knowledge influences the structure of the economy at particular times and
places, and to investigate the way in which that structure changes in the light
of actual experience and of changes in economic doctrines. That is the key
implication of the rational expectations idea for empirical work.
Problems posed by the credibility of policy for the predictive content of
macroeconomics are also real. To argue, as I have, that new-classical
economists do not seem to be following the most fruitful path in investigating
such matters (which probably lie in an analysis of the way in which private
and public sector agents interact through political processes) does not alter
the fact that it has been the new-classicals’ initial insights which have
compelled macroeconomists in general to recognize the importance of these
questions. They have stressed that a positive theory of government behaviour
must be an important factor conditioning private sector behaviour, and I
have criticized them, not for advancing this view, but for failing even to
attempt to incorporate currently available positive theories of government
into their work.
It is worth pointing out explicitly that the problems with new-classical
economics discussed in this chapter are, in a fundamental sense, different
aspects of a single issue. At least since the first publication of Smith’s Wealth
of Nations (1776) economists have been arguing about the extent to which a
society that organizes its economic activity on the basis of voluntary
exchange of private property rights can be expected to achieve a coherent
solution to problems of resource utilization and allocation (not to mention
distribution). From their arguments has emerged an increasingly clear
understanding that analysis of the institutional framework within which, and
the processes whereby, the decisions of agents are co-ordinated, and the
information upon which those decisions are based is disseminated, must lie
at the heart of any attempt to come to grips with these issues.
New-classical economists insist that we assume agents to possess, as
common knowledge, almost all systematic information about the structure of
the economy relevant to their welfare before we model their decision-
making. They also insist that, in analysing the interaction of agents, we must
assume that their behaviour is co-ordinated by a price mechanism that never
permits their plans to be incompatible for long enough to have observable
consequences. In short, new-classical economics requires that we treat certain
(and extreme) propositions about a market economy’s capacity for solving
New-classical contribution to macroeconomics 351
problems of disseminating information and co-ordinating decisions, not as
hypotheses to be questioned and investigated, but as axiomatic assumptions.
To adhere to the ‘first principles’ of analysis upon which new-classical
economics is based requires that we give up questioning the coherence of
economic activity co-ordinated by markets and confine our activities to
describing the nature of a coherence that is presumed to exist. If the
popularity of Keynesian economics in the years following the Depression
was, as Lucas is said to have told Newsweek (14 February 1985:60), ‘based
on political needs, not economic truth’ then, so, surely, as Howitt (1986) has
remarked, does the current popularity of new-classical economics reflect its
compatibility with the ideology of the New Right.
And yet the pioneers of new-classical economics are no more ideologues
than was Keynes. Disinterested seriousness about following the logic of an
argument wherever it might lead is surely the hallmark of the writings of
Lucas and Sargent, and let it be said explicitly, that, in this chapter I intend
to accuse them of no worse an offence than permitting this very seriousness
of purpose to lead them into carrying good ideas too far and sometimes in
the wrong direction. If this characterization of the ‘new-classical revolution’
is accepted, it has not, of course, in this respect been different from other
periods of advance in economic knowledge. The Keynesian revolution and
the monetarist revolution were both in their own ways equally open to
criticism on such grounds in their respective days. More to the point, in
rejecting the extremes to which new-classical economics has taken them, we
should not lose sight of the fact that the ideas in question are, after all, good
ones. When, as I hope it will, the main thrust of macroeconomics research
returns to addressing problems of Information and Co-ordination, to borrow
yet another phrase from Leijonhufvud (1982) it surely will do so with a much
clearer understanding of the role of purposeful maximizing individual
behaviour in the solution of these problems than could have been possible
had the new-classical revolution never occurred.
ACKNOWLEDGEMENTS
I am grateful to Dieter Helm, Peter Howitt, Jurek Konieczny, Thomas Mayer
and George Stadler for helpful conversations and correspondence about
aspects of this chapter.
NOTES
1 For an account of the nature of ‘revolutions’ in economics, illustrated with reference
to the Keynesianism and monetarism, see Johnson (1971). A number of
commentators (e.g. Tobin 1981; Howitt 1986) treat new-classical economics as a
‘Mark 2’ version of monetarism. For a contrary view, see Laidler (1982: ch. 1) where
I argue that whereas, from the point of view of the analytic structure of the models it
utilized, monetarism was a development of Keynesian theory, new-classical economics
352 David Laidler
in important respects is a throwback to the Austrian economics of the 1920s and
early 1930s. This theme also runs through much of this chapter but, because the
adjective ‘neo-Austrian’ seems to upset some people, I have not used it here.
2 Lucas has made this point on a number of occasions (see e.g. 1980).
3 The choice of the word ‘occurred’ is not accidental. Though in some places (e.g.
Britain) fiscal expansion in the mid-1960s and again in the early 1970s was
deliberately used in an attempt to generate real ‘growth’, the experiment in the
United States was less wholeheartedly and self-consciously ‘Keynesian’, but had a
great deal to do with the politics of financing the Vietnam War. Note that, in this
chapter, I am using the adjective ‘Keynesian’ to refer to the economics of what
Lucas (e.g. 1980) referred to as the ‘Neo-Classical Synthesis’. I am not talking
about the ‘Economics of Keynes’, to borrow Leijonhufvud’s (1968) phrase.
4 Harry Johnson as long ago as 1970 noted the effects of ‘conditioned Keynesian
reflexes’ in preventing the idea of a stable demand for money function being
incorporated into British Keynesian thought. American Keynesians, such as James
Tobin (e.g. 1981) and Franco Modigliani (e.g. 1977) were much more open minded
in this respect.
5 Thus Milton Friedman’s early 1970s ‘Monetary Framework’ (1974) was explicitly
cast in IS-LM terms, while Brunner and Meltzer (e.g. 1976a) used an extended IS-
LM model to expound their important insights about the role of credit markets in
the process of money creation. So strong were the IS-LM roots of Brunner and
Meltzer’s work at that time that at least one commentator, Dornbusch (1976), was
misled into believing that their essential contribution could be grasped without
any extension at all to the IS-LM framework. Dornbusch’s misconception did at
least have the productive consequence of provoking an exceptionally clear statement
from Brunner and Meltzer (1976b) of where they saw their contribution as lying.
This author too, in analysing inflation unemployment dynamics, used a vertical
LM curve IS-LM model as a starting point (Laidler 1973).
6 The aggregate supply curve interpretation of the Phillips curve was not a component
of early Monetarism though Friedman did accept it on at least one later occasion. In
1968 he said ‘Phillips’ analysis…contains a basic defect—the failure to distinguish
between nominal wages and real wages’ (Friedman’s italics). In 1975, (pp. 12–14), in
a pamphlet explicitly dealing with the role of rational expectations and such in
monetarist analysis, while continuing to point up this nominal-real confusion, he
characterized taking ‘the rate of change of prices as the independent variable’ as ‘the
truth’ and taking ‘the level of employment to be the independent variable’ as ‘error’
(Friedman’s italics). Friedman’s acceptance of the aggregate supply curve interpretation
of the Phillips curve, quite clearcut in this 1975 pamphlet, was never thoroughgoing,
however. Thus, the ‘framework’ of 1974 is used, without apology, as the theoretical
starting point for the analysis contained in Friedman and Schwartz (1983) and is
quite incompatible with new-classical style equilibrium macroeconomics.
7 I have discussed these issues in some detail in Laidler (1982) particularly chs 1, 3
and 4. It is this fundamental theoretical difference which leads me to treat new-
classical macroeconomics as a distinct body of analysis, rather than as a simple
extension of ‘monetarism’.
8 Thus the paper by Lucas and Rapping, among others contained in the famous
Phelps (1970) volume, was based on just such a hybrid.
9 There has been a considerable change of emphasis over time here. The claim to have
a superior method of analysis looms much larger in more recent defences of their
New-classical contribution to macroeconomics 353
work by new-classicals, than in earlier ones. Compare for example Lucas and Sargent
(1978) to their recorded comments in Klamer (1984) which led Kramer himself to
argue that the new-classical revolution was a matter of method and rhetoric, rather
than substance. Howitt (1986) rightly criticizes Klamer for this judgement.
10 As Lucas (1980) himself acknowledged. Not so Sargent. See Klamer (1984:66–7).
11 Which is not to say that no differences remained between Keynesians and
monetarists. For example, though there is little difference between this author’s
views on the way in which the economy works, and those expressed by say
Modigliani (1977) or Lipsey (1981), I am much less optimistic than are they about
the scope for stabilization policy. In general, there is considerable continuity between
the policy views of monetarists and new-classicals, and it is this continuity that
persuades Tobin (1981) and Howitt (1986) to take the work of the latter as an
extension of that of the former. I regard theoretical differences as decisive in this
matter of classification. See Laidler (1982) particularly chs 1 and 3.
12 The reader’s attention is drawn to the use of the word ‘purposeful’ here. To
adherents of revealed preference analysis, for whom consistent behaviour is logically
equivalent to utility maximization, as opposed to being a consequence of it, the
methodological case for new-classical macroeconomics, particularly as it relates to
the rational expectations idea will not perhaps be as strong as I here present it. It
might be noted, that in stressing the individualistic maximizing foundations of
their model, as opposed to its empirical content, New-classicals are reverting to a
weighting of methodological criteria used by Austrians in the 1920s and 1930s.
See especially Robbins (1935).
13 Lucas (1980) comes close to arguing along such lines. I do not find this style of
argument persuasive.
14 In his contributions to Klamer (1984) Sargent at one point appears to accept this
view of the ultimate primacy of empirical evidence. See Klamer (1984:68). However,
the general thrust of his work, and that of other new-classicals seems to be to stress
the importance of deriving results from what they take to be ‘first principles’. For
this indulgence in the ‘Cartesian fallacy’ they are, rightly in my view, taken to task
by Brunner in his contribution to Klamer (1984:191–5). The reader who is familiar
with Brunner’s methodological views on these issues will recognize the common
debt that we both owe to Karl Popper.
15 The danger here is one that Sargent is aware of. See Klamer (1984:75–6). We may
illustrate it from an earlier episode in the development of macroeconomics. Their
ability to deal with the conflict between time series and cross section evidence on the
consumption income relationship did not constitute an empirical argument in favour
of the Friedman (1957) and Modigliani and Brumberg (1954) theories of the
consumption function, but only confirmed the logical powers of their authors. They
knew about the evidence in question before they constructed their models and
calibrated them to it. The important empirical content of their theories which rendered
them testable lay in their ability to tell us about other empirical regularities, which
either had not been observed, or were not regarded as being related to the theoretical
foundations of consumer theory, until new theoretical insights were put to work.
16 On the matter of real wages and employment, see Geary and Kennan (1982).
17 See Laidler (1982: ch. 2) and Lane (1983) for discussion of this matter.
18 This is not to say that the 1980s experience was any more the outcome of a
conscious attempt to implement new-classical policies than was the 1970s the
result of a conscious Keynesian experiment. Nevertheless, before the event,
354 David Laidler
new-classical economists did make confident predictions about the outcome of
pre-announced monetary contraction. Thus, Lucas is quoted by Time magazine,
27 Aug. 1979 p. 29 as having said ‘Ideally we should announce a monetary
expansion policy of 4% annually for the next seven years and then stick to it.
People would respond, and inflation would be cured with a minimal risk of a deep
recession’. The basis of such prediction as this was the Sargent and Wallace (1976)
analysis of the effects of rational expectations on the ability of the monetary
authorities to influence real variables. Nowadays it is claimed that this paper was
taken more seriously and literally by its readers than by its authors. (See Sargent in
Klamer 1984:70–1.) Certainly, the opening of the paper in question suggests that
the analysis which it contains is to be treated as a counterexample to a prevailing
Keynesian view of policy, rather than as serious alternative, but its last two or three
pages mount a strong case for treating it as just such an alternative.
19 A New-Classical economist should treat the ‘credibility’ alibi with care. Before the
event, Sargent and Wallace (1976:181) developed what they characterized as a
‘telling argument’ against its empirical relevance. On the matter of the demand for
money, see Goodfriend (1985) and Kohn and Manchester (1985).
20 Laidler (1985) demonstrates that price stickiness is an alternative to new-classical
assumptions in generating such results rather than a supplement to them, and
does so in terms of a model in which agents are all ‘in equilibrium’ in the sense of
being able to execute their ex ante plans, albeit not with the expected results ex
post. The model in question is not of course an equilibrium framework in the sense
that markets are cleared by flexible prices. Instead quantity fluctuations play an
equilibrating role in some markets. Note that some new-classical economists (e.g.
Barro 1977b) have argued that even with sticky wage and price contracts it is
possible to model quantity fluctuations as taking place ‘as if’ they reflected
appropriate market clearing responses to variations in agents’ perceptions of the
marginal product and marginal disutility of labour. This is true enough, but hard
to take seriously as an empirical proposition, because it implies that agents have
enough information to take market clearing decisions in the absence of price signals.
For a discussion of this, see Laidler (1982:90–2).
21 One must be careful here, because the phrase ‘excess demand’ has strong overtones
of ‘disequilibrium’ about it and there is much semantic confusion in the
macroeconomics literature caused by contributors referring to any non-flexible-
price-Walrasian system as a ‘disequilibrium’ one. Excess demand is here used to
refer to the difference between the level of output at which markets currently clear,
and that at which they would clear if prices were perfectly flexible and all expectations
were completely fulfilled. The literature with which I am dealing here treats the
latter as a unique level of output, determined by tastes, technology, and market
institutions, but work by Diamond (1984) and Howitt (1985) on search equilibria
suggests that we ought not to take such uniqueness for granted once we get away
from an economy presided over by a Walrasian auctioneer. The work to which I
refer here provides a complementary analysis of potentially great importance to
short-run sticky price macroeconomic models.
22 Montgomery and Shaw (1985) have investigated the role of money wage stickiness in
an otherwise new-classical framework, have conceded it to be a pervasive phenomenon,
but have argued that it has little explanatory power over quantity fluctuations. The
basis for this last conclusion appears to be the assumption that, wage contracts
notwithstanding, money prices are perfectly flexible, and hence it misses the point of
New-classical contribution to macroeconomics 355
Keynesian analysis which models quantity fluctuations as an alternative equilibrating
mechanism to price fluctuations, and not as a response to them.
23 However, see Rowe (1985) for a pioneering attempt to come to grips with problems
of this sort.
24 The argument was well developed by Hayek (1928), and according to Hansson (1983)
a slightly later version of it, developed initially by Gunnar Myrdal, was seminal to
much Swedish dynamic economics in the 1930s. See also McCloughry (1984).
25 I have developed this argument at greater length in Laidler (1984). The interaction
of expectations and the structure of the economy is most fully developed by Lucas
(1976) in what I suspect will turn out to be the most durably important paper of
the new-classical revolution.
26 Thus, I stand by the judgement offered by Laidler and Parkin (1975:771) that
‘The Rational Expectations hypothesis…is probably better suited to a
characterization of expectations formation in the very long run.’ I do not wish to
imply that my co-author would still subscribe to this view, though.
27 The argument here abstracts from more down to earth issues such as whether the
proposition that only ‘unanticipated money’ affects output is uniquely a prediction
of new-classical macroeconomics (it isn’t—see Laidler 1985), or whether the data
actually do support Barro’s analysis (they don’t appear to—see Mishkin 1982).
28 Hayek paid increasing attention to problems of knowledge as a determinant of
economic behaviour and became less and less inclined to ascribe empirical content
to what we would now call a full rational expectations equilibrium, such as he
described (1928), from the mid-1930s onwards. The turning point in his thought
is perhaps to be found in Hayek (1937). On this matter see also McCloughry
(1984).
29 Both Lucas and Sargent recommend differential game theory in their contributions
to Klamer (1984:55, 73). It is instructive to compare their discussion of this issue
with Karl Brunner’s contribution to the same volume (Klamer 1984:185–6).
Brunner has, of course, long been acutely aware of the role of political processes in
forming policies and conditioning the private sector’s responses to them.
30 The contribution of Mancur Olson (1982; 1984), and of James Buchanan and his
associates (see e.g. Buchanan, Tollison and Tulloch 1980) to this literature are well
known. It is surely no accident that two prominent monetarists who have refused
to join the ‘New-classical revolution’, Karl Brunner and Allan Meltzer, have also
worked in the public choice area. It should also be noted that Harry Johnson drew
similar conclusions to those developed here about the interaction of expectations,
policy, and political processes as long ago as 1972.
REFERENCES
Barro, R.J. (1972) ‘A Theory of Monopolistic Price Adjustment’, Review of Economic
Studies 34, January, 17–26.
Barro, R.J. (1977a) ‘Unanticipated Money Growth and Unemployment in the United
States’, American Economic Review 67, March, 101–15.
Barro, R.J. (1977b) ‘Long-Term Contracting, Sticky Prices, and Monetary Policy’,
Journal of Monetary Economics 3, July, 305–16.
Barro, R.J. (1978) ‘Unanticipated Money, Output and the Price Level in the United
States’, Journal of Political Economy 86, August, 549–81.
356 David Laidler
Brunner, K. and Meltzer, A.H. (1963) ‘Predicting Velocity: Implications for Theory and
Policy’, Journal of Finance 18, May, 319–54.
Brunner, K. and Meltzer, A.H. (1976a) ‘An Aggregative Theory for a Closed Economy’,
in J.Stein (ed.) Monetarism, Amsterdam: North-Holland.
Brunner, K. and Meltzer, A.H. (1976b) ‘Reply’, in J.Stein (ed.) Monetarism, Amsterdam:
North-Holland.
Buchanan, J., Tollison, R.D. and Tulloch, G. (eds) (1980) Towards a Theory of the
Rent Seeking Society, College Station: Texas A&M Press.
Diamond, P.A. (1984) A Search Equilibrium Approach to the Microfoundations of
Macroeconomics, Cambridge, MA: MIT Press.
Dornbusch, R. (1976) ‘Comments on Brunner and Meltzer’, in J.Stein (ed.) Monetarism,
Amsterdam: North-Holland.
Fischer, S. (1977) ‘Long-Term Contracts, Rational Expectations and the Optimal Money
Supply Rule’, Journal of Political Economy 85, February, 191–206.
Friedman, M. (1957) A Theory of the Consumption Function, Princeton, NJ: Princeton
University Press for the NBER.
Friedman, M. (1959) ‘The Demand for Money—Some Theoretical and Empirical
Results’, Journal of Political Economy 67, June, 327–51.
Friedman, M. (1968) ‘The Role of Monetary Policy’, American Economic Review 58,
March, 1–17.
Friedman, M. (1974) Milton Friedman’s Monetary Framework, R.J.Gordon (ed.) ch.
1, Chicago: University of Chicago Press.
Friedman, M. (1975) Unemployment Versus Inflation, London: Institute of Economic
Affairs.
Friedman, M. and Schwartz, A.J. (1983) Monetary Trends in the United States and the
United Kingdom, Chicago: University of Chicago Press for the NBER.
Geary, P. and Kennan, J. (1982) ‘The Employment Real Wage Relationship: An
International Study’, Journal of Political Economy 90, August, 854–71.
Goodfriend, M. (1985) ‘Reinterpreting Money Demand Regressions’, in K.Brunner
and A.H.Meltzer (eds) Understanding Monetary Regimes Carnegie-Rochester
Conference Series, vol. 22, Amsterdam: North-Holland.
Hall, R. (1980) ‘Employment Fluctuations and Wage Rigidity’, Brookings Papers on
Economic Activity 1, 91–123.
Hansson, B.A. (1983) The Stockholm School and the Development of Dynamic Method,
London: Croom Helm.
Hayek, F.A.von (1928) ‘Intertemporal Price Equilibrium and Movements in the Value
of Money’, translated and reprinted (1984) in F.A.Hayek, Money Capital and
Fluctuations: Early Essays, R.McCloughry (ed.), Chicago: University of Chicago
Press.
Hayek, F.A.von (1937) ‘Economics and Knowledge’, Economica n.s. 4, February, 33–
54.
Howitt, P.W. (1981) ‘Activist Monetary Policy under Rational Expectations’, Journal
of Political Economy 89, April, 249–69.
Howitt, P.W. (1985) ‘Transactions Costs in the Theory of Unemployment’, American
Economic Review 75, March, 88–100.
Howitt, P.W. (1986) ‘Conversations with Economists—A Review Essay’, Journal of
Monetary Economics 18, July.
Johnson, H.G. (1970) ‘Recent Developments in Monetary Theory—A Survey’, in
D.Croome and H.G.Johnson (eds) Money in Britain, London: Oxford University
Press.
Johnson, H.G. (1971) ‘The Keynesian Revolution and the Monetarist Counter-
Revolution’, American Economic Review 61, May, 9–22.
Johnson, H.G. (1972) Inflation and the Monetarist Controversy, Amsterdam: North-
Holland.
New-classical contribution to macroeconomics 357
Kawasaki, S., McMillan J. and Zimmerman, K.F. (1983) ‘Inventories and Price
Flexibility’, Econometrica 51, May, 599–610.
Klamer, A. (1984) Conversations with Economists, Totowa, NJ: Roman and Allanheld.
Kohn, M. and Manchester, J. (1985) ‘International Evidence on Misspecification of the
Standard Money Demand Equation’, Journal of Monetary Economics 16, March,
87–94.
Laidler, D. (1973) ‘The Influence of Money on Real Income and Inflation—A Simple
Model with Some Empirical Tests for the United States 1953–72’, Manchester School
61, December.
Laidler, D. (1982) Monetarist Perspectives, Deddington: Philip Allan; Cambridge, MA:
Harvard University Press.
Laidler, D. (1984) ‘Did Macroeconomics Need the Rational Expectations Revolution?’,
in G.Mason (ed.) Macroeconomics: Theory, Policy and Evidence, Winnipeg, Manitoba:
Institute for Social and Economic Research.
Laidler, D. (1985) ‘Price Stickiness, Buffer-Stock Money, and Persistent Fluctuations in
Output and Real Balances’, UWO Department of Economics Research Report 8509
(mimeo).
Laidler, D. and Parkin, J.M. (1975) ‘Inflation—A Survey’, Economic Journal 75,
December, 741–809.
Lane, T. (1983) ‘Essays in monetary control’, unpublished Ph.D dissertation, University
of Western Ontario.
Leijonhufvud, A. (1982) On Keynesian Economics and the Economics of Keynes,
London: Oxford University Press.
Leijonhufvud, A. (1968) Information and Coordination, London: Oxford University
Press.
Lipsey, R.G. (1981) ‘Presidential Address: The Understanding and Control of Inflation—
Is There a Crisis in Macroeconomics?’, Canadian Journal of Economics 14,
November, 545–76.
Lucas, R.E. Jr (1972) ‘Expectations and the Neutrality of Money’, Journal of Economic
Theory 4, 103–24.
Lucas, R.E. Jr (1976) ‘Econometric Policy Evaluation’, in K.Brunner and A.H. Meltzer
(eds) The Phillips Curve and the Labour Market, Carnegie-Rochester Conference
Series, vol. 1, Amsterdam: North-Holland.
Lucas, R.E. Jr (1980) ‘Methods and Problems in Business Cycle Theory’, Journal of
Money, Credit and Banking 12, November, part II, 696–715.
Lucas, R.E. Jr and Rapping, L.A. (1970) ‘Real Wages, Employment and Inflation’, in
E.Phelps et al., The Microeconomic Foundations of Employment and Inflation
Theory, New York: Norton.
Lucas, R.E. Jr and Sargent, T.J. (1978) ‘After Keynesian Economies’, in After the Phillips
Curve: Persistence of High Inflation and High Unemployment, Boston, MA: Federal
Reserve Bank of Boston.
McCloughry, R. (1984) ‘Editor’s Introduction’, to F.A.Hayek, Money Capital and
Fluctuations: Early Essays, R.McCloughry (ed.), Chicago: University of Chicago
Press.
Menger, C. (1871) Principles of Economics, tr. and ed. J.Dingwall and B.Hoselitz (1950),
Glencoe, IL: The Free Press.
Mishkin, F. (1982) ‘Does Anticipated Monetary Policy Matter? An Econometric
Investigation’, Journal of Political Economy 90, February, 22–51.
Modigliani, F. (1977) ‘The Monetarist Controversy, or Should We Forsake Stabilisation
Policy’, American Economic Review 67, March, 1–19.
Modigliani, F. and Brumberg, R. (1954) ‘Utility Analysis and the Consumption Function:
An Interpretation of the Cross-Section Data’, in K.K.Kurihara (ed.) Post Keynesian
Economics, New Brunswick, NJ: Rutgers University Press.
358 David Laidler
Montgomery, E. and Shaw, K. (1985) ‘Long Term Contracts, Expectations and Wage
Inertia’, Journal of Monetary Economics 16, September, 209–26.
Okun, A. (1981) Prices and Quantities: A Macroeconomic Analysis, Washington, DC:
Brookings Institution.
Olson, M. (1982) The Rise and Decline of Nations, New Haven, CT: Yale University
Press.
Olson, M. (1984) ‘Microeconomic Incentives and Macroeconomic Decline’,
Weltwirtschaftliches Archiv 120, 631–45.
Phelps, E. (1967) ‘Phillips Curves, Expectations of Inflation and Optimal Unemployment
Over Time’, Economica NS 34, August, 254–81.
Phelps, E. et al. (1970) Microeconomic Foundations of Employment and Inflation
Theory, New York: Norton.
Phelps, E. and Taylor, J. (1977) ‘Stabilizing Powers of Monetary Policy under Rational
Expectations’, Journal of Political Economy 85, February, 163–90.
Robbins, L.C. (1935) An Essay on the Nature and Significance of Economic Science,
London: Macmillan.
Rowe, N. (1985) ‘Rules and institutions’, unpublished Ph.D dissertation, University of
Western Ontario.
Sargent, T. and Wallace, N. (1973) ‘Rational Expectations and the Dynamics of
Hyperinflation’, International Economic Review 14, 328–50.
Sargent, T. and Wallace, N. (1976) ‘Rational Expectations and the Theory of Economic
Policy’, Journal of Monetary Economics 2, May, 169–83.
Smith, A. (1776) An Inquiry into the Nature and Causes of the Wealth of Nations,
London.
Tobin, J. (1981) ‘The Monetarist Counter-Revolution Today—An Appraisal’, Economic
Journal 91, March, 29–42.
Part IV
The real business cycle
approach to economic
fluctuations
Introduction
The real business cycle school, a predominantly US school of thought,
evolved out of the new classical school in the 1980s. The research
programme associated with this second phase of equilibrium theorizing was
initiated by Kydland and Prescott’s (1982) seminal article and following
Long and Plosser (1983) has come to be referred to as the real business cycle
approach. In stark contrast to the earlier new classical models, proponents of
real business theory reject the view that unanticipated monetary shocks
generate fluctuations in output and employment. Instead the real business
cycle school views business cycles as being predominantly caused by
persistent real (or supply-side) shocks, rather than monetary (or demand-side)
shocks to the economy (some models emphasize shocks to government
expenditure: see Barro and Grilli 1994). These real shocks, the focus of
which involve large random fluctuations in the rate of technological
progress, cause the aggregate production function to shift. Rational economic
agents are held to optimally respond to the resultant fluctuations in relative
prices by altering their supply of labour and consumption. As such,
fluctuations in output and employment are regarded as Pareto efficient
responses to real technological shocks to the aggregate production function.
Furthermore observed fluctuations in output are viewed as fluctuations in the
natural (trend) rate of output, rather than deviations of output from a smooth
deterministic trend, thereby resulting in the abandonment of the established
convention of distinguishing between short-run cycles and long-run trend.
In his 1986 Federal Reserve Bank of Minneapolis Quarterly Review
article (reprinted on pp. 366–88) on Theory Ahead of Business Cycle
Measurement’, Edward Prescott, a leading exponent of the real business
cycle approach, maintains that the large fluctuations in output and
employment displayed in the US economy are in accord with the predictions
of standard economic theory, given the persistent nature of technological
shocks and significant intertemporal substitution in labour supply and
employment decisions. Prescott uses a model based on the orthodox
neoclassical growth model and argues that the artificial economy he
constructs displays fluctuations in aggregate time series with statistical
properties which are close to those experienced by the US economy since the
362 Real business cycle approach to economic fluctuations
Korean War. In conclusion he controversially suggests that, as fluctuations in
output and employment are the optimal responses to irregular technological
shocks, ‘costly efforts at stabilization are likely to be counterproductive’.
Lawrence Summer’s 1986 Federal Reserve Bank of Minneapolis Quarterly
Review article (reprinted on pp. 389–95) entitled ‘Some Skeptical
Observations on Real Business Cycle Theory’ examines four main criticisms
of the type of real business cycle model put forward by Prescott. These
criticisms, which concern the parameters used by Prescott in his model,
evidence on technological shocks and intertemporal substitution in
employment, price data, and breakdowns in the exchange mechanism, lead
Summers to reject this radical approach as an explanation of aggregate
instability.
Another leading exponent of the real business cycle approach is Charles
Plosser. In his 1989 Journal of Economic Perspectives article (reprinted on
pp. 396–424) entitled ‘Understanding Real Business Cycles’, Plosser provides
a very accessible introduction to the real business cycle approach to business
fluctuations, in which dynamic general equilibrium models are constructed
to help understand the behaviour of aggregate economic variables following
changes in the economic environment. Plosser demonstrates how the
predicted annual growth rates of real output, consumption, investment, wage
rate and hours worked, derived from the basic neoclassical model of capital
accumulation, appear to mirror a significant portion of the actual behaviour
of these five variables for the US economy between 1955 and 1985. While
real business cycle models have typically focused on real technological
shocks as the main source of economic fluctuations, Plosser also considers
some potential areas for research to develop and extend the real business
cycle approach, including shocks arising from changes in preferences and the
quantity of money. In the final contribution, also reprinted from the 1989
Summer issue of the Journal of Economic Perspectives, Gregory Mankiw in
his article entitled ‘Real Business Cycles: A New Keynesian Perspective’
acknowledges that ‘while real business cycle theory has served the important
function of stimulating and provoking scientific debate’, he predicts that it
will ‘ultimately be discarded as an explanation of observed fluctuations’.
Mankiw highlights what he regards as two fundamental weaknesses of the
approach, namely the reliance given to large technological disturbances as
the main source of economic fluctuations and the intertemporal substitution
of leisure in explaining changes in employment.
The real business cycle approach, like the new classical approach from
which it evolved, has proved to be highly controversial. Critics of the real
business cycle research programme remain convinced that it has a number of
serious deficiencies. Contrary to the ‘strong’ version (see McCallum 1989) in
which monetary disturbances have negligible, if any, consequences for real
variables, most economists believe that in the short run aggregate demand
disturbances arising from changes in monetary policy can have significant
real effects because of the nominal price and wage rigidities which
Introduction 363
characterize actual economies (see Part V). Furthermore many economists
reject the view that stabilization policy has no role to play. Nevertheless real
business cycle theorists have made two important and lasting contributions
to macroeconomics. First, real business cycle theory has challenged the pre-
1980 consensus that growth and fluctuations are distinct phenomena to be
studied separately, requiring different analytical tools. By integrating the
theory of growth and fluctuations, the direction of modern business cycle
research has been irreversibly changed. Second, rather than attempting to
provide models capable of conventional econometric testing, real business
cycle theorists, inspired by the work of Kydland and Prescott (1982) have
developed the calibration method. While calibration does not, to date,
provide a method that allows one to judge between the performance of real
and other business cycle models, it has provided an important new
contribution to the methodology of macroeconomics research (see Hoover
1995; Wickens 1995). This new research methodology for macroeconomics,
with its emphasis on the stylized facts of the business cycle (see Ryan and
Mullineux 1997) to be explained and the construction of general equilibrium
dynamic models which can replicate such stylized facts, has focused
attention on, and stimulated renewed interest into, empirical knowledge of
business cycle phenomena., In short the major contribution of the real
business cycle approach has been to raise fundamental questions relating to
the meaning, significance and characteristics of economic fluctuations.
REFERENCES
* Titles marked with an asterisk are particularly recommended for additional reading.
*Abel, A.B. and B.S.Bernanke (1995) Macroeconomics, 2nd edn, Chapters 9 and 11,
New York: Addison Wesley.
*Barro, R.J. and V.Grilli (1994) European Macroeconomics, Chapter 12, London:
Macmillan.
Danthine, J.P. and J.B.Donaldson (1993) ‘Methodological and Empirical Issues in Real
Business Cycle Theory’, European Economic Review 37, January, pp. 1–35.
*Froyen, R.T. (1996) Macroeconomics: Theories and Policies, 5th edn, Chapter 12,
London: Prentice-Hall.
*Gordon, R.J. (1993) Macroeconomics, 6th edn, Chapter 7, New York: HarperCollins.
*Hall, R.E. and J.B.Taylor (1993) Macroeconomics, 4th edn, Chapters 4 and 15, New
York: W.W.Norton.
Hoover, K.D. (1995) ‘Facts and Artifacts: Calibration and the Empirical Assessment of
Real Business-Cycle Models’, Oxford Economic Papers 47, pp. 24–44.
*Jansen, D.W., C.D.Delorme and R.B.Ekelund, Jr (1994) Intermediate Macroeconomics,
Chapter 11, New York: West.
Kydland, F.E. and E.C.Prescott (1982) ‘Time to Build and Aggregate Fluctuations’,
Econometrica 50, November, pp. 1345–70.
*Kydland, F.E. and E.C.Prescott (1990) ‘Business Cycles: Real Facts and a Monetary
Myth’, Federal Reserve Bank of Minneapolis Quarterly Review 14, Spring, pp. 3–
18.
Long, J.B. and C.I.Plosser (1983) ‘Real Business Cycles’, Journal of Political Economy
91, February, pp. 39–69.
364 Real business cycle approach to economic fluctuations
*Lucas, R.E. Jr (1977) ‘Understanding Business Cycles’, in K.Brunner and A.H. Meltzer
(eds) Stabilization of the Domestic and International Economy, Amsterdam: North
Holland.
McCallum, B.T. (1989) ‘Real Business Cycle Models’, in R.J.Barro (ed.) Modern Business
Cycle Theory, Cambridge, MA: Harvard University Press.
*Mankiw, N.G. (1994) Macroeconomics, 2nd edn, Chapter 14, New York: Worth.
*Mullineux, A.W. and D.G.Dickinson (1992) ‘Equilibrium Business Cycles: Theory
and Evidence’, Journal of Economic Surveys 6, pp. 321–58.
*Plosser, C.I. (1994) ‘Interview with Charles Plosser’, in B.Snowdon, H.R.Vane and
P.Wynarczyk, A Modern Guide to Macroeconomics: An Introduction to Competing
Schools of Thought, Aldershot: Edward Elgar.
*Ryan, C. and A.W.Mullineux (1997) ‘The Ups and Downs of Modern Business Cycle
Theory’, in B.Snowdon and H.R.Vane (eds) Reflections on the Development of
Modern Macroeconomics, Aldershot: Edward Elgar.
*Snowdon, B. and H.R.Vane (1995) ‘New Classical Macroeconomics Mark II: The
Real Business Cycle Model’, Economics and Business Education 3, Winter, pp.
153–6.
*Snowdon, B., H.R.Vane and P.Wynarczyk (1994) A Modern Guide to Macroeconomics:
An Introduction to Competing Schools of Thought, Chapter 6, Aldershot: Edward
Elgar.
*Stadler, G.W. (1994) ‘Real Business Cycles’, Journal of Economic Literature 32,
December, pp. 1750–83.
Wickens, M. (1995) ‘Real Business Cycle Analysis: A Needed Revolution in
Macroeconometrics’, Economic Journal 105, November, pp. 1637–48.
QUESTIONS
1 What are the main differences between the mark I new classical and real
business cycle models?
2 How are fluctuations in employment explained in the real business cycle
model?
3 What are the ‘stylized facts’ of the business cycle? Have business cycle
theorists successfully explained these facts?
4 To what extent have economists succeeded in developing a satisfactory
equilibrium model of the business cycle since the early 1970s?
5 Is there any role for stabilization policy in a real business model?
6 Compare and contrast the conventional approach to econometric testing
with the calibration method advocated by real business cycle theorists.
7 ‘If these theories are correct, they imply that the macroeconomics developed
in the wake of the Keynesian Revolution is well confined to the ashbin of
history’ (Summers 1986). Critically examine this view of real business cycle
theories.
8 How do real business cycle theorists reconcile their view that variations in
the quantity of money are unimportant for explaining economic fluctuations
with the empirical evidence which appears to show a money-output link?
9 Is the calculation of the Solow residual a satisfactory and reliable method
of identifying the variance of technology shocks?
Introduction 365
10 ‘Real business cycle theory is not only a competitor to Keynesian
macroeconomics but also represents a serious challenge to all monetarist
models as well as early new classical explanations of aggregate instability’.
Explain and discuss.
15 Theory ahead of business cycle
measurement
Edward C.Prescott
Federal Reserve Bank of Minneapolis Quarterly Review (1986) Fall,
pp. 9–22
Economists have long been puzzled by the observations that during
peacetime industrial market economies display recurrent, large fluctuations
in output and employment over relatively short time periods. Not uncommon
are changes as large as 10 percent within only a couple of years. These
observations are considered puzzling because the associated movements in
labor’s marginal product are small.
These observations should not be puzzling, for they are what standard
economic theory predicts. For the United States, in fact, given people’s
ability and willingness to intertemporally and intratemporally substitute
consumption and leisure and given the nature of the changing production
possibility set, it would be puzzling if the economy did not display these
large fluctuations in output and employment with little associated
fluctuations in the marginal product of labor. Moreover, standard theory also
correctly predicts the amplitude of these fluctuations, their serial correlation
properties, and the fact that the investment component of output is about six
times as volatile as the consumption component.
This perhaps surprising conclusion is the principal finding of a research
program initiated by Kydland and me (1982) and extended by Kydland and
me (1984), Hansen (1985a) and Bain (1985). We have computed the
competitive equilibrium stochastic process for variants of the constant
elasticity, stochastic growth model. The elasticities of substitution and the
share parameters of the production and utility functions are restricted to
those that generate the growth observations. The process governing the
technology parameter is selected to be consistent with the measured
technology changes for the American economy since the Korean War. We ask
whether these artificial economies display fluctuations with statistical
properties similar to those which the American economy has displayed in
that period. They do.1
I view the growth model as a paradigm for macro analysis—analogous to
the supply and demand construct of price theory. The elasticities of
substitution and the share parameters of the growth model are analogous to
the price and income elasticities of price theory. Whether or not this
paradigm dominates, as I expect it will, is still an open question. But the
Theory ahead of business cycle measurement 367
early results indicate its power to organize our knowledge. The finding that
when uncertainty in the rate of technological change is incorporated into the
growth model it displays the business cycle phenomena was both dramatic
and unanticipated. I was sure that the model could not do this without some
features of the payment and credit technologies.
The models constructed within this theoretical framework are necessarily
highly abstract. Consequently, they are necessarily false, and statistical
hypothesis testing will reject them. This does not imply, however, that
nothing can be learned from such quantitative theoretical exercises. I think
much has already been learned and confidently predict that much more will
be learned as other features of the environment are introduced. Prime
candidates for study are the effects of public finance elements, a foreign
sector, and, of course, monetary factors. The research I review here is best
viewed as a very promising beginning of a much larger research program.
THE BUSINESS CYCLE PHENOMENA
The use of the expression business cycle is unfortunate for two reasons. One
is that it leads people to think in terms of a time series’ business cycle
component which is to be explained independently of a growth component;
our research has, instead, one unifying theory of both of these. The other
reason I do not like to use the expression is that it is not accurate; some
systems of low-order linear stochastic difference equations with a
nonoscillatory deterministic part, and therefore no cycle, display key
business cycle features (see Slutzky 1927). I thus do not refer to business
cycles, but rather to business cycle phenomena, which are nothing more nor
less than a certain set of statistical properties of a certain set of important
aggregate time series. The question that I and others have considered is, Do
the stochastic difference equations that are the equilibrium laws of motion
for the stochastic growth display the business cycle phenomena?
More specifically, we follow Lucas (1977:9) in defining the business cycle
phenomena as the recurrent fluctuations of output about trend and the co-
movements among other aggregate time series. Fluctuations are by definition
deviations from some slowly varying path. Since this slowly varying path
increases monotonically over time, we adopt the common practice of
labeling it trend. This trend is neither a measure nor an estimate of the
unconditional mean of some stochastic process. It is, rather, defined by the
computational procedure used to fit the smooth curve through the data.
If the business cycle facts were sensitive to the detrending procedure
employed, there would be a problem. But the key facts are not sensitive to
the procedure if the trend curve is smooth. Our curve-fitting method is to
take the logarithms of variables and then select the trend path {tt} which
minimizes the sum of the squared deviations from a given series {Yt} subject
to the constraint that the sum of the squared second differences not be too
large. This is
368 Edward C.Prescott
The smaller is µ, the smoother is the trend path. If µ=0, the least squares
linear time trend results. For all series, µ is picked so that the Lagrange
multiplier of the constraint is 1600. This produces the right degree of
smoothness in the fitted trend when the observation period is a quarter of a
year. Thus, the sequence {�
t
} minimizes
The first-order conditions of this minimization problem are linear in Y
t
and
�t, so for every series, �=AY, where A is the same T×T matrix. The deviations
from trend, also by definition, are
Unless otherwise stated, these are the variables used in the computation of
the statistics reported here for both the United States and the growth
economies.
An alternative interpretation of the procedure is that it is a high pass
linear filter. The facts reported here are essentially the same if, rather than
defining the deviations by Yd=(I-A) Y, we filtered the Y using a high pass
band filter, eliminating all frequencies of 32 quarters or greater. An
advantage of our procedure is that it deals better with the ends of the sample
problem and does not require a stationary time series.
To compare the behaviors of a stochastic growth economy and an actual
economy, only identical statistics for the two economies are used. By definition,
a statistic is a real valued function of the raw time series. Consequently, if a
comparison is made, say, between the standard deviations of the deviations, the
date t deviation for the growth economy must be the same function of the data
generated by that model as the date t deviation for the US economy is of that
economy’s data. Our definitions of the deviations satisfy this criterion.
Figure 15.1 plots the logs of actual and trend output for the US economy
during 1947–82, and Figure 15.2 the corresponding percentage deviations
from trend of output and hours of market employment. Output and hours
clearly move up and down together with nearly the same amplitudes.
Table 15.1 contains the standard deviations and cross serial correlations
of output and other aggregate time series for the US economy during 1954–
82. Consumption appears less variable and investment more variable than
subject to
Theory ahead of business cycle measurement 369
output. Further, the average product of labor is procyclical but does not vary
as much as output or hours.
THE GROWTH MODEL
This theory and its variants build on the neoclassical growth economy of
Solow (1956) and Swan (1956). In the language of Lucas (1980:696), the
model is a ‘fully articulated, artificial economic system’ that can be used to
Figure 15.1 Actual and trend logs of US gross national product (Quarterly, 1947–82)
Source of basic data: Citicorp’s Citibase data bank
Figure 15.2 Deviations from trend of gross national product and nonfarm employee
hours in the United States (Quarterly, 1947–82)
Source of basic data: Citicorp’s Citibase data bank
370 Edward C.Prescott
generate economic time series of a set of important economic aggregates.
The model assumes an aggregate production function with constant returns
to scale, inputs labor n and capital k, and an output which can be allocated
either to current consumption c or to investment x. If t denotes the date, f:
R2 R the production function, and zt, a technology parameter, then the
production constraint is
x
t
+c
t
≤ z
t
f(k
t
,n
t
)
where xt,ct,kt,nt≥0. The model further assumes that the services provided by
a unit of capital decrease geometrically at a rate 0<δ<1:
kt+1=(1-δ)kt+xt.
Solow completes the specification of his economy by hypothesizing that some
fraction 0<σ<1 of output is invested and the remaining fraction 1-σ consumed
and that nt is a constant—say, —for all t. For this economy, the law of
motion of capital condition on zt is
k
t+1
=(1-δ)k
t
+σz
t
f(kt, ).
Once the {z
t
} stochastic process is specified, the stochastic process governing
capital and the other economic aggregates are determined and realizations of
the stochastic process can be generated by a computer.
This structure is far from adequate for the study of the business cycle
because in it neither employment nor the savings rate varies, when in fact
they do. Being explicit about the economy, however, naturally leads to the
question of what determines these variables, which are central to the cycle.
Table 15.1 Cyclical behavior of the US economy
(Deviations from trend of key variables, 1954:1–1982:4)
Source of basic data: Citicorp’s Citibase data bank
Theory ahead of business cycle measurement 371
That leads to the introduction of a stand-in household with some explicit
preferences. If we abstract from the labor supply decision and uncertainty
(that is, zt= and nt= ), the standard form of the utility function is
where ß is the subjective time discount factor. The function u: R
+
R is twice
differentiable and concave. The commodity space for the deterministic
version of this model is l∞, infinite sequences of uniformly bounded
consumptions .
The theorems of Bewley (1972) could be applied to establish existence of a
competitive equilibrium for this l∞ commodity-space economy. That
existence argument, however, does not provide an algorithm for computing
the equilibria. An alternative approach is to use the competitive welfare
theorems of Debreu (1954). Given local nonsaturation and no externalities,
competitive equilibria are Pareto optima and, with some additional
conditions that are satisfied for this economy, any Pareto optimum can be
supported as a competitive equilibrium. Given a single agent and the
convexity, there is a unique optimum and that optimum is the unique
competitive equilibrium allocation. The advantage of this approach is that
algorithms for computing solutions to concave programming problems can
be used to find the competitive equilibrium allocation for this economy.
Even with the savings decision endogenous, this economy has no
fluctuations. As shown by Cass (1965) and Koopmans (1965), the competitive
equilibrium path converges monotonically to a unique rest point or, if zt is
growing exponentially, to a balanced growth path. There are multisector
variants of this model in which the equilibrium path oscillates (see Benhabib
and Nishimura 1985; Marimon 1984). But I know of no multisector model
which has been restricted to match observed factor shares by sector, which
has a value for ß consistent with observed interest rates, and which displays
oscillations.
When uncertainty is introduced, the household’s objective is its expected
discounted utility:
The commodity vector is now indexed by the history of shocks; that is,
is the commodity point. As Brock and Mirman (1972)
show, if the {z
t
} are identically distributed random variables, an optimum to
the social planner’s problem exists and the optimum is a stationary stochastic
process with k
t+1
=g(k
t
,z
t
) and c
t
=c(k
t
, z
t
). As Lucas and Prescott (1971) show,
for a class of economies that include this one, the social optimum is the
unique competitive equilibrium allocation. They also show that for these
homogeneous agent economies, the social optimum is also the unique
372 Edward C.Prescott
sequence-of-markets equilibrium allocation. Consequently, there are
equilibrium time-invariant functions for the wage w
t
=w(k
t
, z
t
) and the rental
price of capital r
t
=r(k
t
,z
t
), where these prices are relative to the date t
consumption good. Given these prices, the firm’s period t problem is
max k
t
,n
t
≥0 {y
t
-r
t
k
t
-w
t
n
t
}
subject to the output constraint
y
t
≤z
t
f(k
t
, n
t
).
The household’s problem is more complicated, for it must form expectations
of future prices. If at is its capital stock, its problem is
and given a0-k0. In forming expectations, a household knows the relation
between the economy’s state (kt, zt) and prices, wt=w(kt, zt) and rt=r(kt, zt).
Further, it knows the process governing the evolution of the per capita
capital stock, a variable which, like prices, is taken as given.
The elements needed to define a sequence-of-markets equilibrium are
the firm’s policy functions y(kt, zt), n(kt, zt), and k(kt, zt); the household’s
policy functions x(at, kt, zt) and c(at, kt, zt); a law of motion of per capita
capital kt+1=g(kt, zt); and pricing functions w(kt, zt) and r(kt, zt). For
equilibrium, then,
• The firm’s policy functions must be optimal given the pricing functions.
• The household’s policy functions must be optimal given the pricing
functions and the law of motion of per capita capital.
• Spot markets clear; that is, for all k
t
and z
t
=n(k
t
, z
t
)
k
t
=k(k
t
, z
t
)
x(k
t
, k
t
, z
t
)+c(k
t
, k
t
, z
t
)=y(k
t
, z
t
).
(Note that the goods market must clear only when the representative
household is truly representative, that is, when a
t
=k
t
).
• Expectations are rational; that is,
g(k
t
, z
t
)=(1-δ)k
t
+x(k
t
, k
t
, z
t
).
This definition still holds if the household values productive time that is
allocated to nonmarket activities. Such time will be called leisure and
subject to
Theory ahead of business cycle measurement 373
denoted lt. The productive time endowment is normalized to 1, and the
household faces the constraints
nt+lt ≤1
for all t. In addition, leisure is introduced as an argument of the utility
function, so the household’s objective becomes the maximization of
Now leisure—and therefore employment—varies in equilibrium.
The model needs one more modification: a relaxation of the assumption
that the technology shocks zt are identically and independently distributed
random variables. As will be documented, they are not so distributed. Rather,
they display considerable serial correlation, with their first differences nearly
serially uncorrelated. To introduce high persistence, we assume
z
t+1
=ρz
t
+
t+1
where the { t+1} are identically and independently distributed and ρ is near
1. With this modification, the recursive sequence-of-markets equilibrium
definition continues to apply.
USING DATA TO RESTRICT THE GROWTH MODEL
Without additional restrictions on preferences and technology, a wide variety
of equilibrium processes are consistent with the growth model. The beauty of
this model is that both growth and micro observations can be used to
determine its production and utility functions. When they are so used, there are
not many free parameters that are specific to explaining the business cycle
phenomena and that cannot be measured independently of those phenomena.
The key parameters of the growth model are the intertemporal and
intratemporal elasticities of substitution. As Lucas (1980:712) emphasizes, ‘On
these parameters, we have a wealth of inexpensively available data from
census cohort information, from panel data describing the reactions of
individual households to a variety of changing market conditions, and so
forth.’ To this list we add the secular growth observations which have the
advantage of being experiments run by nature with large changes in relative
prices and quantities and with idiosyncratic factors averaged out.2 A
fundamental thesis of this line of inquiry is that the measures obtained from
aggregate series and those from individual panel data must be consistent. After
all, the former are just the aggregates of the latter.
Secularly in the United States, capital and labor shares of output have
been approximately constant, as has r, the rental price of capital. However,
the nation’s real wage has increased greatly—more than 100 percent since
the Korean War. For these results to hold, the model’s production function
must be approximately Cobb-Douglas:
374 Edward C.Prescott
The share parameter θ is equal to labor’s share, which has been about 64
percent in the postwar period, so θ=0.64. This number is smaller than that
usually obtained because we include services of consumer durables as part of
output. This alternative accounting both reduces labor’s share and makes it
more nearly constant over the postwar period.
The artificial economy has but one type of capital, and it depreciates at
rate δ. In fact, different types of capital depreciate at different rates, and the
pattern of depreciation over the life of any physical asset is not constant.
Kydland and I (1982, 1984) simply pick δ=0.10. With this value and an
annual real interest rate of 4 percent, the steady-state capital-annual output
ratio is about 2.6. That matches the ratio for the US economy and also
implies a steady-state investment share of output near the historically
observed average. Except for parameters determining the process on the
technology shock, this completely specifies the technology of the simple
growth model.
A key growth observation which restricts the utility function is that leisure
per capita lt has shown virtually no secular trend while, again, the real wage
has increased steadily. This implies an elasticity of substitution between
consumption ct and leisure lt near 1. Thus, the utility function restricted to
display both constant intertemporal and unit intratemporal elasticities of
substitution is
where 1/γ >0 is the elasticity of substituting between different date composite
commodities . This leaves γ and the subjective time discount factor ß
[or, equivalently, the subjective time discount rate (1/ß)-1] to be determined.
The steady-state interest rate is
As stated previously, the average annual real interest rate is about 4
percent, and the growth rate of per capita consumption /c has averaged
nearly 2 percent. The following studies help restrict γ. Tobin and Dolde
(1971) find that a γ near 1.5 is needed to match the life cycle consumption
patterns of individuals. Using individual portfolio observations, Friend and
Blume (1975) estimate γ to be near 2. Using aggregate stock market and
consumption data, Hansen and Singleton (1983) estimate γ to be near 1.
Using international data, Kehoe (1984) also finds a modest curvature
parameter γ. All these observations make a strong case that γ is not too far
from 1 . Since the nature of fluctuations of the artificial economy is not
very sensitive to γ, we simply set γ equal to 1. Taking the limit as γ 1
yields
Theory ahead of business cycle measurement 375
This leaves ß and φ still to be determined.
Hansen (1985b) has found that growing economies—that is, those with zt
having a multiplicative, geometrically growing factor (1+�)t with �>0—
fluctuate in essentially the same way as economies for which �=0. This
justifies considering only the case �=0. If �=0, then the average interest rate
approximately equals the subjective time discount rate.3 Therefore, we set ß
equal to 0.96 per year or 0.99 per quarter.
The parameter φ is the leisure share parameter. Ghez and Becker (1975)
find that the household allocates approximately one-third of its productive
time to market activities and two-thirds to nonmarket activities. To be
consistent with that, the model’s parameter φ must be near two-thirds. This is
the value assumed in our business cycle studies.
Eichenbaum et al. (1984) use aggregate data to estimate this share
parameter φ, and they obtain a value near five-sixths. The difference
between two-thirds and five-sixths is large in the business cycle context. With
φ=2/3, the elasticity of labor supply with respect to a temporary change in
the real wage is 2, while if φ=5/6, it is 5. This is because a 1 percent change
in leisure implies a φ/(φ-1) percent change in hours of employment.
We do not follow the Eichenbaum-Hansen-Singleton approach and treat φ
as a free parameter because it would violate the principle that parameters
cannot be specific to the phenomena being studied. What sort of science
would economics be if micro studies used one share parameter and aggregate
studies another?
The nature of the technological change
One method of measuring technological change is to follow Solow (1956)
and define it as the changes in output less the sum of the changes in labor’s
input times labor share and the changes in capital’s input times capital share.
Measuring variables in logs, this is the percentage change in the technology
parameter of the Cobb-Douglas production function. For the US economy
between the third quarter of 1955 and the first quarter of 1984, the standard
deviation of this change is 1.2 percent.4 The serial autocorrelations of these
changes are ρ1=-0.21, ρ2=-0.06, ρ3=0.04, ρ4=0.01, and ρ5=-0.05. To a first
approximation, the process on the percentage change in the technology
process is a random walk with drift plus some serially uncorrelated
measurement error. This error produces the negative first-order serial
correlation of the differences.
Further evidence that the random walk model is not a bad approximation
is based on yearly changes. For the quarterly random walk model, the
standard deviation of this change is 6.63 times the standard deviation of the
quarterly change. For the US data, the annual change is only 5.64 times as
376 Edward C.Prescott
large as the quarterly change. This, along with the negative first-order serial
correlation, suggests that the standard deviation of the persistent part of the
quarterly change is closer to 5.64/6.63=0.85 than to 1.2 percent. Some
further evidence is the change over four-quarter periods—that is, the change
from a given quarter of one year to the same quarter of the next year. For the
random walk model, the standard deviation of these changes is 2 times the
standard deviation of the quarterly change. A reason that the standard
deviation of change might be better measured this way is that the
measurement noise introduced by seasonal factors is minimized. The
estimate obtained in this way is 0.95 percent. To summarize, Solow growth
accounting finds that the process on the technology parameter is highly
persistent with the standard deviation of change being about 0.90.5
The Solow estimate of the standard deviation of technological change is
surely an overstatement of the variability of that parameter. There
undoubtedly are non-negligible errors in measuring the inputs. Since the
capital input varies slowly and its share is small, the most serious
measurement problem is with the labor input. Fortunately there are two
independent measures of the aggregate labor input, one constructed from a
survey of employers and the other from a survey of households. Under the
assumption of orthogonality of their measurement errors, a reasonable
estimate of the variance of the change in hours is the covariance between the
changes in the two series. Since the household survey is not used to estimate
aggregate output, I use the covariance between the changes in household
hours and output as an estimate of the covariance between aggregate hours
and output. Still using a share parameter of θ=0.75, my estimate of the
standard deviation of the percentage change in zt is the square root of var
(∆ )-2θcov(∆ 1,∆ )+θ
2cov(∆ 1,∆ 2), where the caret (^) denotes a measured
value. For the sample period my estimate is 0.763 percent. This is probably
a better estimate than the one which ignores measurement error.
Still, my estimate might under- or overstate the variance of technological
change. For example, the measurement of output might include significant
errors. Perhaps measurement procedures result in some smoothing of the
series. This would reduce the variability of the change in output and might
reduce the covariance between measured hours and output.
Another possibility is that changes in hours are associated with
corresponding changes in capital’s utilization rate. If so, the Solow approach
is inappropriate for measuring the technology shocks. To check whether this
is a problem, I varied θ and found that θ=0.85 yields the smallest estimate,
0.759, as opposed to 0.763 for θ=0.75. This suggests that my estimate is not
at all sensitive to variations in capital utilization rates.
To summarize, there is overwhelming evidence that technological shocks
are highly persistent. But tying down the standard deviation of the
technology change shocks is difficult. I estimate it as 0.763. It could very
well be larger or smaller, though, given the accuracy of the measurements.
Theory ahead of business cycle measurement 377
THE STATISTICAL BEHAVIOR OF THE GROWTH MODELS
Theory provides an equilibrium stochastic process for the growth economy
studied. Our approach has been to document the similarities and differences
between the statistical properties of data generated by this stochastic process
and the statistical properties of American time series data. An alternative
approach is to compare the paths of the growth model if the technological
parameters {z
t
} were those experienced by the US economy. We did not
attempt this because theory’s predictions of paths, unlike its predictions of the
statistical properties, are sensitive to what Learner (1983:43) calls
‘whimsical’ modeling assumptions. Another nontrivial problem is that the
errors in measuring the innovations in the z
t
process are as large as the
innovations themselves.
The basic growth model
With the standard deviation of the technology shock equal to 0.763, theory
implies that the standard deviation of output will be 1.48 percent. In fact, it
is 1.76 percent for the post-Korean War American economy. For the output of
the artificial economy to be as variable as that, the variance of the shock
must be 1.0, significantly larger than the estimate. The most important
deviation from theory is the relative volatility of hours and output. Figure
15.3 plots a realization of the output and employment deviations from trend
for the basic growth economy. A comparison of Figures 15.2 and 15.3
demonstrates clearly that, for the US economy, hours in fact vary much more
than the basic growth model predicts. For the artificial economy, hours
fluctuate 52 percent as much as output, whereas for the US economy, the
Figure 15.3 Deviations from trend of GNP and hours worked in the basic growth
economy
378 Edward C.Prescott
ratio is 0.95. This difference appears too large to be a result of errors in
measuring aggregate hours and output.
The Kydland-Prescott economy
Kydland and I (1982, 1984) have modified the growth model in two
important respects. First, we assume that a distributed lag of leisure and the
market-produced good combine to produce the composite commodity good
valued by the household. In particular,
Kydland (1983) provides justification for this preference ordering based on
an unmeasured, household-specific capital stock that, like c
t
and l
t
, is an
input in the production of the composite commodity. The economy studied
has α
0
=0.5 and η=0.1. This increases the variability of hours.
The second modification is to permit the workweek of capital to vary
proportionally to the workweek of the household. For this economy,
increases in hours do not reduce the marginal product of labor as much, so
hours fluctuate more in response to technology shocks of a given size.
The statistical properties of the fluctuations for this economy are reported
in Table 15.2. As is clear there, hours are now about 70 percent as variable
as output. This eliminates much of the discrepancy between theory and
measurement. If the standard deviation of the technology shock is 0.72
percent, then fluctuations in the output of this artificial economy are as large
as those experienced in the US economy.
A comparison of Tables 15.1 and 15.2 shows that the Kydland-Prescott
economy displays the business cycle phenomena. It does not quite demonstrate,
however, that there would be a puzzle if the economy did not display the
business cycle phenomena. That is because the parameters α0 and η have not
been well tied down by micro observations.6 Better measures of these
parameters could either increase or decrease significantly the amount of the
fluctuations accounted for by the uncertainty in the technological change.
The Hansen indivisible labor economy
Labor economists have estimated labor supply elasticities and found them to
be small for full-time prime-age males (see, for example, Ashenfelter 1984).
Heckman (1984), however, finds that when movements between employment
and nonemployment are considered and secondary workers are included,
Theory ahead of business cycle measurement 379
elasticities of labor supply are much larger. He also finds that most of the
variation in aggregate hours arises from variation in the number employed
rather than in the hours worked per employed person.
These are the observations that led Hansen (1985a) to explore the
implication of introducing labor indivisibilities into the growth model. As
shown by Rogerson (1984), if the household’s consumption possibility set has
nonconvexities associated with the mapping from hours of market production
activities to units of labor services, there will be variations in the number
employed rather than in the hours of work per employed person. In addition,
the aggregate elasticity of labor supply will be much larger than the
elasticity of those whose behavior is being aggregated. In this case
aggregation matters, and matters greatly.
There certainly are important nonconvexities in the mapping from hours
of market activities to units of labor services provided. Probably the most
important nonconvexity arises from the considerable amount of time
required for commuting. Other features of the environment that would make
full-time workers more than twice as productive as otherwise similar half-
time workers are not hard to imagine. The fact that part-time workers
typically are paid less per hour than full-time workers with similar human
capital endowments is consistent with the existence of important
nonconvexities.
Hansen (1985a) restricts each identical household to either work h̄ hours
or be unemployed. His relation is as depicted by the horizontal lines in
Note: * These are the means of 20 simulations, each of which was 116 periods
long; the numbers in parentheses are standard errors
Source: Kydland and Prescott (1984)
Table 15.2 Cyclical behavior of the Kydland-Prescott economy*
380 Edward C.Prescott
Figure 15.4. This assumption is not as extreme as it appears. If the relation
were as depicted by the curved line, the behavior of the economy would be
the same. The key property is an initial convex region followed by a
concave region in the mapping from hours of market activity to units of
labor service.
With this modification, lotteries that specify the probability of
employment are traded along with market-produced goods and capital
services. As before, the utility function of each individual is
u(c,l)=(1/3)log c+(2/3)log l.
If an individual works, l=1- ; otherwise, l=1. Consequently, if π is the
probability of employment, an individual’s expected utility is
E{u(c, l)}=(1/3)log c+(2/3)πlog (1- ).
Given that per capita consumption is and per capita hours of employment
, average utility over the population is maximized by setting c= for all
individuals. If , which equals 1-π , denotes per capita leisure, then
maximum per capita utility is
This is the utility function which rationalizes the per capita consumption and
leisure choices if each person’s leisure is constrained to be either 1-h̄ or 1.
The aggregate intertemporal elasticity of substitution between different date
Figure 15.4 Relation between time allocated to market activity and labor service
Theory ahead of business cycle measurement 381
leisures is infinity independent of the value of the elasticity for the individual
(in the range where not all are employed).
Hansen (1985a) finds that if the technology shock standard deviation is
0.71, then fluctuations in output for his economy are as large as those for the
US economy. Further, variability in hours is 77 percent as large as
variability in output. Figure 15.5 shows that aggregate hours and output for
his economy fluctuate together with nearly the same amplitude. These
theoretical findings are the basis for my statement in the introduction that
there would be a puzzle if the economy did not display the business cycle
phenomena.
Empirical labor elasticity
One important empirical implication of a shock-to-technology theory of
fluctuations is that the empirical labor elasticity of output is significantly
larger than the true elasticity, which for the Cobb-Douglas production
function is the labor share parameter. To see why, note that the capital stock
varies little cyclically and is nearly uncorrelated with output. Consequently,
the deviations almost satisfy
y
t
=θh
t
+z
t
where y
t
is output, h
t
hours, and z
t
the technology shock. The empirical
elasticity is
η=cov(h
t
, y
t
)/var(h
t
)
Figure 15.5 Deviations from trend GNP and hours worked in Hansen’s indivisible
labor economy
Source: Gary D.Hansen, Department of Economics, University of California, Santa
Barbara
382 Edward C.Prescott
which, because of the positive correlation between ht and zt, is considerably
larger than the model’s θ, which is 0.64. For the basic, Kydland-Prescott, and
Hansen growth economies, the values of η are 1.9, 1.4, and 1.3, respectively.
Because of measurement errors, the empirical elasticity for the US
economy is not well-estimated by simply computing the ratio of the
covariance between hours and output and dividing by the variance of hours.
The procedure I use is based on the following probability model:
where the cart (ˆ) denotes a measured value. The �
it
are measurement errors.
Here, the
1t
measure of hours uses the employer survey data while the
2t
measure uses the household survey data. Since these are independent
measures, a maintained hypothesis is that �
2t
and �
3t
are orthogonal. With
this assumption, a reasonable estimate of var (h
t
) is the sample covariance
between
1t
and
2t
. Insofar as the measurement of output has small
variance or �
1t
is uncorrelated with the hours measurement errors or both,
the covariance between measured output and either measured hours series is
a reasonable estimate of the covariance between output and hours. These
two covariances are 2.231×10-4 and 2.244×10-4 for the sample period, and I
take the average as my estimate of cov (h
t
, y
t
) for the US economy. My
estimate of the empirical labor elasticity of output is
This number is considerably greater than labor’s share, which is about 0.70
when services of consumer durables are not included as part of output. This
number strongly supports the importance of technological shocks in
accounting for business cycle fluctuations. Nevertheless, the number is
smaller than those for the Kydland-Prescott and Hansen growth economies.
One possible reason for the difference between the US economy and the
growth model empirical labor elasticities of output is cyclical measurement
errors in output. A sizeable part of the investment component of output is
hard to measure and therefore not included in the US National Product
Accounts measure of output, the gross national product (GNP). In particular,
a firm’s major maintenance expenditures, research and development
expenditures, and investments in human capital are not included in GNP. In
good times—namely, when output is above trend—firms may be more likely
to undertake major repairs of a not fully depreciated asset, such as replacing
the roof of a 30-year-old building which has a tax life of 35 years. Such an
Theory ahead of business cycle measurement 383
expenditure is counted as maintenance and therefore not included in GNP
even though the new roof will provide productive services for many years.
The incentive for firms to do this is tax savings: by expensing an investment
rather than capitalizing it, current tax liabilities are reduced. Before 1984,
when a railroad replaced its 90-pound rails, the expenditure was treated as a
maintenance expense rather than an investment expenditure. If these and
other types of unmeasured investment fluctuate in percentage terms more
than output, as do all the measured investment components, the volatility of
GNP is larger than measured. We do know that investment in rails was
highly procyclical and volatile in the postwar period. A careful study is
needed to determine whether the correction for currently unmeasured
investment is small or large.
Another reason to expect the American economy’s labor elasticity to be
less than the model’s is that the model shocks are perfectly neutral with
respect to the consumption and investment good transformation. Persistent
shocks which alter the product transformation frontier between these goods
would cause variation in output and employment but not in the productivity
parameters. For fluctuations so induced, the empirical labor elasticity of
output would be the true elasticity. Similarly, relatively permanent changes
in the taxing of capital—such as altering depreciation rates, the corporate
income tax rate, or the investment tax credit rate—would all result in
fluctuations in output and employment but not in the productivity
parameters.
A final reason for actual labor elasticity to be less than the model’s is the
way imports are measured. An increase in the price of imported oil, that is,
an increase in the quantity of output that must be sacrificed for a given unit
of that input, has no effect on measured productivity. From the point of view
of the growth model, however, an oil price increase is a negative technology
shock because it results in less output, net of the exports used to pay for the
imported oil, available for domestic consumption and investment. Theory
predicts that such shocks will induce variations in employment and output,
even though they have no effect on the aggregate production function.
Therefore, insofar as they are important, they reduce the empirical labor
elasticity of output.
EXTENSIONS
The growth model has been extended to provide a better representation of
the technology. Kydland and I (1982) have introduced a technology with
more than one construction period for new production capacity.7 We have
also introduced inventory as a factor of production. This improves the match
between the model’s serial correlation properties and the US postwar data,
but has little effect on the other statistics.
Kydland (1984) has introduced heterogeneity of labor and found that if
there are transfers from high human capital people to low human capital
384 Edward C.Prescott
people, theory implies that hours of the low fluctuate more than hours of the
high. It also implies a lower empirical labor elasticity of output than the
homogeneous household model.
Bain (1985) has studied an economy that is richer in sectoral detail. His
model has manufacturing, retailing, and service-producing sectors. A key
feature of the technology is that production and distribution occur
sequentially. Thus there ar e two types of inventories—those of
manufacturers’ finished goods and those of final goods available for sale.
With this richer detail, theory implies that different components of aggregate
inventories behave in different ways, as seen in the data. It also implies that
production is more volatile than final sales, an observation considered
anomalous since inventories can be used to smooth production (see, for
example, Blinder 1984).
Much has been done. But much more remains to be explored. For
example, public finance considerations could be introduced and theory used
to predict their implications. As mentioned above, factors which affect the
rental price of capital affect employment and output, and the nature of the
tax system affects the rental price of capital. Theory could be used to predict
the effect of temporary increases in government expenditures such as those in
the early 1950s when defense expenditures increased from less than 5 to
more than 13 percent of GNP. Theory of this type could also be used to
predict the effect of terms-of-trade shocks. An implication of such an exercise
most likely will be that economies with persistent terms-of-trade shocks
fluctuate differently than economies with transitory shocks. If so, this
prediction can be tested against the observations.
Another interesting extension would be to explicitly model household
production. This production often involves two people, with one specializing
in market production and the other specializing in household production
while having intermittent or part-time market employment. The fact that,
cyclically, the employment of secondary wage earners is much more volatile
than that of primary wage earners might be explained.
A final example of an interesting and not yet answered question is,
how would the behavior of the Hansen indivisible labor economy change
if agents did not have access to a technology to insure against random
unemployment and instead had to self-insure against unemployment by
holding liquid assets? In such an economy, unlike Hansen’s, people would
not be happy when unemployed. Their gain of more leisure would be
more than offset by their loss as an insurer. Answering this question is
not straightforward, because new tools for computing equilibria are
needed.
SUMMARY AND POLICY IMPLICATIONS
Economic theory implies that, given the nature of the shocks to technology
and people’s willingness and ability to intertemporally and intratemporally
Theory ahead of business cycle measurement 385
substitute, the economy will display fluctuations like those the US economy
displays. Theory predicts fluctuations in output of 5 percent and more from
trend, with most of the fluctuation accounted for by variations in
employment and virtually all the rest by the stochastic technology
parameter. Theory predicts investment will be three or more times as volatile
as output and consumption half as volatile. Theory predicts that deviations
will display high serial correlation. In other words, theory predicts what is
observed. Indeed, if the economy did not display the business cycle
phenomena, there would be a puzzle.
The match between theory and observation is excellent, but far from
perfect. The key deviation is that the empirical labor elasticity of output is
less than predicted by theory. An important part of this deviation could very
well disappear if the economic variables were measured more in conformity
with theory. That is why I argue that theory is now ahead of business cycle
measurement and theory should be used to obtain better measures of the key
economic time series. Even with better measurement, there will likely be
significant deviations from theory which can direct subsequent theoretical
research. This feedback between theory and measurement is the way mature,
quantitative sciences advance.
The policy implication of this research is that costly efforts at stabilization
are likely to be counterproductive. Economic fluctuations are optimal
responses to uncertainty in the rate of technological change. However, this
does not imply that the amount of technological change is optimal or
invariant to policy. The average rate of technological change varies much
both over time within a country and across national economies. What is
needed is an understanding of the factors that determine the average rate at
which technology advances. Such a theory surely will depend on the
institutional arrangements societies adopt. If policies adopted to stabilize the
economy reduce the average rate of technological change, then stabilization
policy is costly. To summarize, attention should be focused not on
fluctuations in output but rather on determinants of the average rate of
technological advance.
ACKNOWLEDGEMENTS
This paper was presented at a Carnegie-Rochester Conference on Public
Policy and will appear in a volume of the conference proceedings. It
appears here with the kind permission of Allan H.Meltzer, editor of that
volume. The author thanks Finn E.Kydland for helpful discussions of the
issues reviewed here, Gary D.Hansen for data series and some additional
results for his growth economy, Lars G.M.Ljungqvist for expert research
assistance, Bruce D.Smith and Allan H.Meltzer for comments on a
preliminary draft, and the National Science Foundation and the
Minneapolis Federal Reserve Bank for financial support. The views
expressed herein are those of the author alone.
386 Edward C.Prescott
NOTES
1 Others (e.g. Barro 1981; Long and Plosser 1983) have argued that these fluctuations
are not inconsistent with competitive theory that abstracts from monetary factors.
Our finding is much stronger, standard theory predicts that the economy will
display the business cycle phenomena.
2 See Solow (1970) for a nice summary of the growth observations.
3 Actually, the average interest rate is slightly lower because of risk premia. Given the
value of γ and the amount of uncertainty, the average premium is only a fraction of
a percent. See Mehra and Prescott (1985) for further details.
4 I use Hansen’s (1984) human capital-weighted, household hour series. The capital
stock and GNP series are from Citicorp’s Citibase data bank.
5 The process z
t+1
=.9z
t
+�
t+1
is, like the random walk process, highly persistent. Kydland
and I find that it and the random walk result in essentially the same fluctuations.
6 Hotz, et al. (1985) use annual panel data to estimate α
0
and η and obtain estimates
near the Kydland-Prescott assumed values.
7 Altug (1983) has introduced two types of capital with different gestation periods.
Using formal econometric methods, she finds evidence that the model’s fit is
improved if plant and equipment investment are not aggregated.
REFERENCES
Altug, S. (1983) ‘Gestation lags and the business cycle’, Working paper, CarnegieMellon
University.
Ashenfelter, O. (1984) ‘Macroeconomic analyses and microeconomic analyses of labor
supply’, in Essays on Macroeconomic Implications of Financial and Labor Markets
and Political Processes, K.Brunner and A.H.Meltzer (eds), Carnegie-Rochester
Conference Series on Public Policy 21:117–55, Amsterdam: North-Holland.
Bain, I.R.M. (1985) ‘A theory of the cyclical movements of inventory stocks’, Ph.D.
dissertation, University of Minnesota.
Barro, R.J. (1981) ‘Intertemporal substitution and the business cycle’, in Supply Shocks,
Incentives and National Wealth, K.Brunner and A.H.Meltzer (eds), Carnegie-
Rochester Conference Series on Public Policy 14:237–68, Amsterdam: North-
Holland.
Benhabib, J. and Nishimura, K. (1985) ‘Competitive equilibrium cycles’, Journal of
Economic Theory 35:284–306.
Bewley, T.F. (1972) ‘Existence of equilibria in economies with infinitely many
commodities’, Journal of Economic Theory 4:514–40.
Blinder, A.S. (1984) ‘Can the production smoothing model of inventory behavior be
saved’, Working paper, Princeton University.
Brock, W.A. and Mirman, L.J. (1972) ‘Optimal economic growth and uncertainty: the
discounted case’, Journal of Economic Theory 4:479–513.
Cass, D. (1965) ‘Optimum growth in an aggregative model of capital accumulation’,
Review of Economic Studies 32:233–40.
Debreu, G. (1954) ‘Valuation equilibrium and Pareto optimum’, Proceedings of the
National Academy of Science 70:558–92.
Eichenbaum, M.S., Hansen, L.P., and Singleton, K.S. (1984) ‘A time series analysis of
representative agent models of consumption and leisure choice under uncertainty’,
Working paper, Carnegie-Mellon University.
Friend, I. and Blume, M.E. (1975) ‘The demand for risky assets’, American Economic
Review 65:900–22.
Theory ahead of business cycle measurement 387
Ghez, G.R. and Becker, G.S. (1975) ‘The allocation of time and goods over the life cycle’,
New York: National Bureau of Economic Research.
Hansen, G.D. (1984) ‘Fluctuations in total hours worked: a study using efficiency
units’, Working paper, University of Minnesota.
——(1985a) ‘Indivisible labor and the business cycle’, Journal of Monetary Economics
16:309–27.
——(1985b) ‘Growth and fluctuations’, Working paper, University of California, Santa
Barbara.
Hansen, L.P. and Singleton, K.J. (1983) ‘Stochastic consumption, risk aversion, and the
temporal behavior of asset returns’, Journal of Political Economy 91:249–65.
Heckman, J. (1984) ‘Comments on the Ashenfelter and Kydland papers’, in Essays on
Macroeconomic Implications of Financial and Labor Markets and Political Processes,
K.Brunner and A.H.Meltzer (eds), Carnegie-Rochester Conference Series on Public
Policy 21:209–24, Amsterdam: North-Holland.
Hotz, V.S., Kydland, F.E. and Sedlacek, G.L. (1985) ‘Intertemporal preferences and
labor supply’, Working paper, Carnegie-Mellon University.
Kehoe, P.J. (1984) ‘Dynamics of the current account: theoretical and empirical analysis’,
Working paper, Harvard University.
Koopmans, T.C. (1965) ‘On the concept of optimal economic growth’, in The
Econometric Approach to Development Planning, Chicago: Rand-McNally.
Kydland, F.E. (1983) ‘Nonseparable utility and labor supply’, Working paper, Hoover
Institution.
——(1984) ‘Labor-force heterogeneity and the business cycle’, in Essays on
Macroeconomic Implications of Financial and Labor Markets and Political Processes,
K. Brunner and A.H.Meltzer (eds), Carnegie-Rochester Conference Series on Public
Policy 21:173–208, Amsterdam: North-Holland.
Kydland, F.E. and Prescott, E.C. (1982) ‘Time to build and aggregate fluctuations’,
Econometrica 50–70.
——(1984) ‘The workweek of capital and labor’, Research Department Working Paper
267, Federal Reserve Bank of Minneapolis.
Learner, E.E. (1983) ‘Let’s take the con out of econometrics’, American Economic
Review 73:31–43.
Long, J.B. and Plosser, C.I. (1983) ‘Real business cycles’, Journal of Political Economy
91:39–69.
Lucas, R.E. Jr (1977) ‘Understanding business cycles’, in Stabilization of the Domestic
and International Economy, K.Brunner and A.H.Meltzer (eds), Carnegie-Rochester
Conference Series on Public Policy 5:7–29, Amsterdam: North-Holland.
——(1980) ‘Methods and problems in business cycle theory’ Journal of Money, Credit
and Banking 12:696–715. Reprinted in Studies in Business-Cycle Theory, pp. 271–
96, Cambridge, MA: MIT Press, 1981.
Lucas, R.E. Jr and Prescott, E.C. (1971) ‘Investment under uncertainty’, Econometrica
39:659–81.
Marimon, R. (1984) ‘General equilibrium and growth under uncertainty: the turnpike
property’, Discussion Paper 624, Northwestern University, Center for Mathematical
Studies in Economics and Management Science.
Mehra, R. and Prescott, E.C. (1985) ‘The equity premium: a puzzle’, Journal of Monetary
Economics 15:145–61.
Rogerson, R.D. (1984) ‘Indivisible labor, lotteries and equilibrium’, in Topics in the
Theory of Labor Markets, ch. 1, Ph.D. dissertation, University of Minnesota.
Slutzky, E. (1927) ‘The summation of random causes as the source of cyclic processes’,
in Problems of Economic Conditions, ed. Conjuncture Institute. Moskva (Moskow),
vol. 3, no. 1, revised English version, 1937, in Econometrica 5:105–46.
Solow, R.M. (1956) ‘A contribution to the theory of economic growth’, Quarterly
Journal of Economics 70:65–94.
388 Edward C.Prescott
——(1970) Growth Theory, Oxford: Oxford University Press.
Swan, T.W. (1956) ‘Economic growth and capital accumulation’, Economic Record
32:334–61.
Tobin, J. and Dolde W. (1971) ‘Wealth, liquidity and consumption’, in Consumer
Spending and Monetary Policy: The Linkages, Monetary Conference Series 5:99–
146, Boston, MA: Federal Reserve Bank of Boston.
16 Some skeptical observations on
real business cycle theory
Lawrence H.Summers
Federal Reserve Bank of Minneapolis Quarterly Review (1986) Fall,
pp. 23–7
The increasing ascendancy of real business cycle theories of various stripes, with
their common view that the economy is best modeled as a floating Walrasian
equilibrium, buffeted by productivity shocks, is indicative of the depths of the
divisions separating academic macroeconomists. These theories deny
propositions thought self-evident by many academic macroeconomists and all of
those involved in forecasting and controlling the economy on a day-to-day basis.
They assert that monetary policies have no effect on real activity, that fiscal
policies influence the economy only through their incentive effects, and that
economic fluctuations are caused entirely by supply rather than demand shocks.
If these theories are correct, they imply that the macroeconomics
developed in the wake of the Keynesian Revolution is well confined to the
ashbin of history. And they suggest that most of the work of contemporary
macroeconomists is worth little more than that of those pursuing astrological
science. According to the views espoused by enthusiastic proponents of real
business cycle theories, astrology and Keynesian economics are in many
ways similar: both lack scientific support, both are premised on the
relevance of variables that are in fact irrelevant, both are built on a
superstructure of nonoperational and ill-defined concepts, and both are
harmless only when they are ineffectual.
The appearance of Ed Prescott’s stimulating paper, Theory Ahead of
Business Cycle Measurement’, affords an opportunity to assess the current state
of real business cycle theory and to consider its prospects as a foundation for
macroeconomic analysis. Prescott’s paper is brilliant in highlighting the appeal
of real business cycle theories and making clear the assumptions they require.
But he does not make much effort at caution in judging the potential of the
real business cycle paradigm. He writes that ‘if the economy did not display
the business cycle phenomena, there would be a puzzle’, characterizes without
qualification economic fluctuations as ‘optimal responses to uncertainty in the
rate of technological change’, and offers the policy advice that ‘costly efforts
at stabilization are likely to be counter productive’.
Prescott’s interpretation of his title is revealing of his commitment to his
theory. He does not interpret the phrase theory ahead of measurement to
390 Lawrence H.Summers
mean that we lack the data or measurements necessary to test his theory.
Rather, he means that measurement techniques have not yet progressed to the
point where they fully corroborate his theory. Thus, Prescott speaks of the
key deviation of observation from theory as follows: ‘An important part of
this deviation could very well disappear if the economic variables were
measured more in conformity with theory. That is why I argue that theory is
now ahead of business cycle measurement’.
The claims of real business cycle theorists deserve serious assessment,
especially given their source and their increasing influence within the
economics profession. Let me follow Prescott in being blunt. My view is that
real business cycle models of the type urged on us by Prescott have nothing
to do with the business cycle phenomena observed in the United States or
other capitalist economies. Nothing in Prescott’s papers or those he
references is convincing evidence to the contrary.
Before turning to the argument Prescott presents, let me offer one lesson
from the history of science. Extremely bad theories can predict remarkably
well. Ptolemaic astronomy guided ships and scheduled harvests for two
centuries. It provided extremely accurate predictions regarding a host of
celestial phenomena. And to those who developed it, the idea that the earth
was at the center seemed an absolutely natural starting place for a theory.
So, too, Lamarckian biology, with its emphasis on the inheritance of
acquired characteristics, successfully predicted much of what was observed
in studies of animals and plants. Many theories can approximately mimic
any given set of facts; that one theory can does not mean that it is even close
to right.
Prescott’s argument takes the form of the construction of an artificial
economy which mimics many of the properties of actual economies. The
close coincidence of his model economy and the actual economy leads him
to conclude that the model economy is a reasonable if abstract representation
of the actual economy. This claim is bolstered by the argument that the
model economy is not constructed to fit cyclical facts but is parameterized on
the basis of microeconomic information and the economy’s long-run
properties. Prescott’s argument is unpersuasive at four levels.
ARE THE PARAMETERS RIGHT?
First, Prescott’s claim to have parameterized the model on the basis of
wellestablished microeconomic and long-run information is not sustainable.
As one example, consider a parameter which Prescott identifies as being
important in determining the properties of the model, the share of household
time devoted to market activities. He claims that is one-third. Data on its
average value over the last century indicate, as Martin Eichenbaum, Lars
Hansen, and Kenneth Singleton (1986) have noted, an average value of one-
sixth since 1956. This seems right—a little more than half the adult
population works, and those who work work about a quarter of the time. I
Skeptical observations: real business cycle 391
am unable to find evidence supporting Prescott’s one-third figure in the cited
book by Gilbert Ghez and Gary Becker (1975). To take another example,
Prescott takes the average real interest rate to be 4 percent. Over the 30-year
period he studies, it in fact averaged only about 1 percent. This list of model
parameters chosen somewhat arbitrarily could be easily extended.
A more fundamental problem lies in Prescott’s assumption about the
intertemporal elasticity of substitution in labor supply. He cites no direct
microeconomic evidence on this parameter, which is central to his model of
cyclical fluctuations. Nor does he refer to any aggregate evidence on it.
Rather, he relies on a rather selective reading of the evidence on the
intertemporal elasticity of substitution in consumption in evaluating the
labor supply elasticity. My own reading is that essentially all the available
evidence suggests only a minimal response of labor to transitory wage
changes. Many studies (including Altonji 1982; Mankiw et al. 1985;
Eichenbaum et al. 1986) suggest that the intertemporal substitution model
cannot account at either the micro or the macro level for fluctuations in
labor supply.
Prescott is fond of parameterizing models based on long-run information.
Japan has since the mid-1950s enjoyed real wage growth at a rate four times
the US rate, close to 8 percent. His utility function would predict that such
rapid real wage growth would lead to a much lower level of labor supply by
the representative consumer. I am not aware that this pattern is observed in
the data. Nor am I aware of data suggesting that age/hours profiles are
steeper in professions like medicine or law, where salaries rise rapidly with
age.
Prescott’s growth model is not an inconceivable representation of reality.
But to claim that its parameters are securely tied down by growth and micro
observations seems to me a gross overstatement. The image of a big loose
tent flapping in the wind comes to mind.
WHERE ARE THE SHOCKS?
My second fundamental objection to Prescott’s model is the absence of any
independent corroborating evidence for the existence of what he calls
technological shocks. This point is obviously crucial since Prescott treats
technological shocks as the only driving force behind cyclical fluctuations.
Prescott interprets all movements in measured total factor productivity as
being the result of technology shocks or to a small extent measurement error.
He provides no discussion of the source or nature of these shocks, nor does
he cite any microeconomic evidence for their importance. I suspect that the
vast majority of what Prescott labels technology shocks are in fact the
observable concomitants of labor hoarding and other behaviour which
Prescott does not allow in his model.
Two observations support this judgment. First, it’s hard to find direct
evidence of the existence of large technological shocks. Consider the oil
392 Lawrence H.Summers
shocks, certainly the most widely noted and commented on shocks of the
postwar period. How much might they have been expected to reduce total
factor productivity? In one of the most careful studies of this issue, Ernst
Berndt (1980:85) concludes that ‘energy price or quantity variations since
1973 do not appear to have had a significant direct role in the slowdown of
aggregate labor productivity in U.S. manufacturing, 1973–77.’ This is not to
deny that energy shocks have important effects. But they have not accounted
for large movements in measured total factor productivity.
Prescott assumes that technological changes are irregular, but is unable to
suggest any specific technological shocks which presage the downturns that
have actually taken place. A reasonable challenge to his model is to ask how
it accounts for the 1982 recession, the most serious downturn of the postwar
period. More generally, it seems to me that the finding that measured
productivity frequently declines is difficult to account for technologically.
What are the sources of technical regress? Between 1973 and 1977, for
example, both mining and construction displayed negative rates of
productivity growth. For smaller sectors of the economy, negative
productivity growth is commonly observed.
A second observation casting doubt on Prescott’s assumed driving force is
that while technological shocks leading to changes in total factor
productivity are hard to find, other explanations are easy to support. Jon Fay
and James Medoff (1985) surveyed some 170 firms on their response to
downturns in the demand for their output. The questions asked were phrased
to make clear that it was exogenous downturns in their output that were
being inquired about. Fay and Medoff (1985:653) summarize their results by
stating that ‘the evidence indicates that a sizeable portion of the swings in
productivity over the business cycle is, in fact, the result of firms’ decisions
to hold labor in excess of regular production requirements and to hoard
labor.’ According to their data, the typical plant in the US manufacturing
sector paid for 8 percent more bluecollar hours than were needed for regular
production work during the trough quarter of its most recent downturn. After
taking account of the amount of other worthwhile work that was completed
by blue-collar employees during the trough quarter, 4 percent of the blue-
collar hours paid for were hoarded. Similar conclusions have been reached
in every other examination of microeconomic data on productivity that I am
aware of.
In Prescott’s model, the central driving force behind cyclical fluctuations is
technological shocks. The propagation mechanism is intertemporal
substitution in employment. As I have argued so far, there is no independent
evidence from any source for either of these phenomena.
WHAT ABOUT PRICES?…
My third fundamental objection to Prescott’s argument is that he does price-
free economic analysis. Imagine an analyst confronting the market for
Skeptical observations: real business cycle 393
ketchup. Suppose she or he decided to ignore data on the price of ketchup.
This would considerably increase the analyst’s freedom in accounting for
fluctuations in the quantity of ketchup purchased. Indeed, without looking at
the price of ketchup, it would be impossible to distinguish supply shocks
from demand shocks. It is difficult to believe that any explanation of
fluctuations in ketchup sales that did not confront price data would be taken
seriously, at least by hard-headed economists.
Yet Prescott offers us an exercise in price-free economics. While real
wages, interest rates, and returns to capital are central variables in his
model, he never looks at any data on them except for his misconstrual of the
average real interest rate over the postwar period. Others have confronted
models like Prescott’s to data on prices with what I think can fairly be
labeled dismal results. There is simply no evidence to support any of the
price effects predicted by the model. Prescott’s work does not resolve—or
even mention—the empirical reality emphasized by Robert Barro and Robert
King (1982) that consumption and leisure move in opposite directions over
the business cycle with no apparent procyclicality of real wages. It is finessed
by ignoring wage data. Prescott’s own work with Rajnish Mehra (1985)
indicates that the asset pricing implications of models like the one he
considers here are decisively rejected by nearly 100 years of historical
experience. I simply do not understand how an economic model can be said
to have been tested without price data.
I believe that the preceding arguments demonstrate that real business
cycle models of the type surveyed by Prescott do not provide a convincing
account of cyclical fluctuations. Even if this strong proposition is not
accepted, they suggest that there is room for factors other than productivity
shocks as causal elements in cyclical fluctuations.
…AND EXCHANGE FAILURES?
A fourth fundamental objection to Prescott’s work is that it ignores the
fact that partial breakdowns in the exchange mechanism are almost
surely dominant factors in cyclical fluctuations. Consider two examples.
Between 1929 and 1933, the gross national product in the United States
declined 50 percent, as employment fell sharply. In Europe in the mid-
1980s, employment has not risen since 1970 and unemployment has risen
more than fivefold in many countries. I submit that it defies credulity to
account for movements on this scale by pointing to intertemporal
substitution and productivity shocks. All the more given that total factor
productivity has increased more than twice as rapidly in Europe as in the
United States.
If some other force is responsible for the largest fluctuations that we
observe, it seems quixotic methodologically to assume that it plays no role at
all in other smaller fluctuations. Whatever mechanisms may have had
something to do with the depression of the 1930s in the United States or the
394 Lawrence H.Summers
depression today in Europe presumably have at least some role in recent
American cyclical fluctuations.
What are those mechanisms? We do not yet know. But it seems clear that
a central aspect of depressions, and probably economic fluctuations more
generally, is a breakdown of the exchange mechanism. Read any account of
life during the Great Depression in the United States. Firms had output they
wanted to sell. Workers wanted to exchange their labor for it. But the
exchanges did not take place. To say the situation was constrained Pareto
optimal given the technological decline that took place between 1929 and
1933 is simply absurd, even though total factor productivity did fall. What
happened was a failure of the exchange mechanism. This is something that
no model, no matter how elaborate, of a long-lived Robinson Crusoe dealing
with his changing world is going to confront. A model that embodies
exchange is a minimum prerequisite for a serious theory of economic
downturns.
The traditional Keynesian approach is to postulate that the exchange
mechanism fails because prices are in some sense rigid, so they do not
attain market-clearing levels and thereby frustrate exchange. This is far
from being a satisfactory story. Most plausible reasons why prices might
not change also imply that agents should not continue to act along static
demand and supply curves. But it hardly follows that ignoring exchange
failures because we do not yet fully understand them is a plausible
strategy.
Where should one look for failures of the exchange process? Convincing
evidence of the types of mechanisms that can lead to breakdowns of the
exchange mechanism comes from analyses of breakdowns in credit markets.
These seem to have played a crucial role in each of the postwar recessions.
Indeed, while it is hard to account for postwar business cycle history by
pointing to technological shocks, the account offered by, for example, Otto
Eckstein and Allen Sinai (1986) of how each of the major recessions was
caused by a credit crunch in an effort to control inflation seems compelling
to me.
CONCLUSION
Even at this late date, economists are much better at analyzing the optimal
response of a single economic agent to changing conditions than they are at
analyzing the equilibria that will result when diverse agents interact. This
unfortunate truth helps to explain why macroeconomics has found the task of
controlling, predicting, or even explaining economic fluctuations so difficult.
Improvement in the track record of macroeconomics will require the
development of theories that can explain why exchange sometimes works
well and other times breaks down. Nothing could be more counterproductive
in this regard than a lengthy professional detour into the analysis of
stochastic Robinson Crusoes.
Skeptical observations: real business cycle 395
NOTE
An earlier version of these remarks was presented at the 25 July 1986
meeting of the National Bureau of Economic Research Economic
Fluctuations Group.
REFERENCES
Altonji, Joseph G. (1982) ‘The intertemporal substitution model of labour market
fluctuations: an empirical analysis’, Review of Economic Studies 49 (special issue):
783–824.
Barro, Robert J. and King, Robert G. (1982) ‘Time-separable preferences and
intertemporal-substitution models of business cycles’, working paper 888, National
Bureau of Economic Research.
Berndt, Ernst R. (1980) ‘Energy price increases and the productivity slowdown in
United States manufacturing’, in The Decline in Productivity Growth, pp. 60–89,
Conference Series 22, Boston, MA: Federal Reserve Bank of Boston.
Eckstein, Otto, and Sinai, Allen (1986) ‘The mechanisms of the business cycle in the
postwar era’, in The American Business Cycle: Continuity and Change, ed. Robert
J.Gordon, pp. 39–105. National Bureau of Economic Research Studies in Business
Cycles, vol. 25, Chicago: University of Chicago Press.
Eichenbaum, Martin S., Hansen, Lars P., and Singleton, Kenneth J. (1986) ‘A time series
analysis of representative agent models of consumption and leisure choice under
uncertainty’, working paper 1981, National Bureau of Economic Research.
Fay, Jon A. and Medoff, James L. (1985) ‘Labor and output over the business cycle:
some direct evidence’, American Economic Review 75 (September): 638–55.
Ghez, Gilbert R. and Becker, Gary S. (1975) The Allocation of Time and Goods over the
Life Cycle, New York: National Bureau of Economic Research.
Mankiw, N.Gregory, Rotemberg, Julio J., and Summers, Lawrence H. (1985)
‘Intertemporal substitution in macroeconomics’, Quarterly Journal of Economics
100 (February): 225–51.
Mehra, Rajnish, and Prescott, Edward C. (1985) ‘The equity premium: a puzzle’, Journal
of Monetary Economics 15 (March): 145–61.
17 Understanding real business cycles
Charles I.Plosser
Journal of Economic Perspectives (1989) 3, Summer, pp. 51–77
The 1960s were a time of great optimism for macroeconomists. Many
economists viewed the business cycle as dead. The Keynesian model was the
reigning paradigm and it provided all the necessary instructions for
manipulating the levers of monetary and fiscal policy to control aggregate
demand. Inflation occurred if aggregate demand was stimulated ‘excessively’
and unemployment arose if demand was ‘insufficient’. The only dilemma
faced by policymakers was determining the most desirable location along
this inflation-unemployment tradeoff or Phillips curve. The remaining
intellectual challenge was to establish coherent microeconomic foundations
for the aggregate behavioral relations posited by the Keynesian framework,
but this was broadly regarded as a detail that should not deter policymakers
in their efforts to ‘stabilize’ the economy.
The return of the business cycle in the 1970s after almost a decade of
economic expansion, and the accompanying high rates of inflation, came as
a rude awakening for many economists. It became increasingly apparent that
the basic Keynesian framework was not the appropriate vehicle for
understanding what happens during a business cycle nor did it seem capable
of providing the empirically correct answers to questions involving changes
in the economic environment or changes in monetary or fiscal policy. The
view that Keynesian economics was an empirical success even if it lacked
sound theoretical foundations could no longer be taken seriously.
The essential flaw in the Keynesian interpretation of macroeconomic
phenomenon was the absence of a consistent foundation based on the choice
theoretic framework of microeconomics. Two important papers, one by
Milton Friedman (1968) and the other by Robert Lucas (1976), forcefully
demonstrated examples of this flaw in critical aspects of the Keynesian
reasoning and set the stage for modern macroeconomics.
A central feature of the Keynesian system of the 1960s was the tradeoff
between inflation and some measure of real output or unemployment.
Friedman argued that basic microeconomic principles demanded that this
long-run Phillips curve must be vertical. That is, general microeconomic
principles implied that individuals (firms) maximizing their utility (profit)
Understanding real business cycles 397
resulted in real demand (and supply) curves that are homogeneous of degree
zero in nominal prices and money income. Thus sustained inflation was
compatible with any level of real demand (or supply) of goods. A central
Keynesian tenet was therefore in stark conflict with microeconomic
principles.1
Lucas reinforced this point by arguing that microeconomic foundations
frequently implied that the sorts of behavioral relations exploited by the
Keynesian model builders were incapable of correctly evaluating changes in
economic policy. Lucas’s specific examples stressed that expectations about
future policy will systematically influence current decisions and thus alter the
behavioral relations exploited by empirical implementations of the
Keynesian analysis. Moreover, Lucas argued, expectations could not be
formulated or specified in an arbitrary manner and be consistent with
individual maximization, but should be viewed as rational in the sense of
Muth (1961).
The absence of an underlying choice theoretic framework also plagued the
dynamic elements of Keynesian models. Business cycles have long been
characterized in terms of how they evolve over time. In particular,
discussions regarding how shocks to the economic system were propagated
across time and across sectors in the economy were a central theme of
Mitchell (1927) and other early students of the business cycle such as von
Hayek (1932). The foundations of the Keynesian model, however, were static
and focused on determining output at a point in time, while treating the
capital stock as given. Dynamic elements were introduced through
accelerator mechanisms (investment and inventories) and later in the form of
price or wage adjustment equations and partial adjustment models of one
form or another. These dynamic specifications, however, did not arise from
any choice theoretic framework of maximization, but were simply
behavioral rules that characterized either agents or, more frequently, markets
in general. One economist’s behavioral formulation for dynamic adjustment
was as good as any other, and it was simply an empirical question which
one seemed to fit the data best.
These problems are fundamental. They suggest that the underpinnings of
our understanding of economic fluctuations are likely to be found somewhere
other than a suitably modified version of the Keynesian model. Indeed, there
is a growing body of research in macroeconomics that begins with the idea
that in order to understand business cycles, it is important and necessary to
understand the characteristics of a perfectly working dynamic economic
system.2 Hicks (1933:32) makes this point quite clearly, arguing that the
‘idealized state of dynamic equilibrium…give(s) us a way of assessing the
extent or degree of disequilibrium.’ In 1939, Hicks set out the basic elements
and tools for determining the character of the ‘idealized state’ in more detail
in Value and Capital. Progress towards understanding this idealized state is
essential because it is logically impossible to attribute an important portion
of fluctuations to market failure without an understanding of the sorts of
398 Charles I.Plosser
fluctuations that would be observed in the absence of the hypothesized
market failure. Keynesian models started out asserting market failures (like
unexplained and unexploited gains from trade) and thus could offer no such
understanding. Fortunately, since the late 1970s, economists have developed
the analytical tools to follow through with the Hicks program.3 The basic
approach is to build on the earlier work in growth theory to construct small-
scale dynamic general equilibrium models and attempt to understand how
aggregate economic variables behave in response to changes in the economic
environment, like changes in technology, tastes, or government policies.
Real business cycle models take the first necessary steps in evaluating and
understanding Hicks’ ‘idealized state of dynamic equilibrium’. Consequently,
these models must be at the core of any understanding economists will
provide of business cycles. This brief chapter is intended to provide readers
with an introduction to the real business cycle approach to business
fluctuations.
THE BASIC REAL BUSINESS CYCLE FRAMEWORK
Real business cycle models view aggregate economic variables as the
outcomes of the decisions made by many individual agents acting to
maximize their utility subject to production possibilities and resource
constraints. As such, the models have an explicit and firm foundation in
microeconomics. More explicitly, real business cycle models ask the
question: how do rational maximizing individuals respond over time to
changes in the economic environment and what implications do those
responses have for the equilibrium outcomes of aggregate variables?
To address these questions, it is necessary to specify the economic
environment and how it evolves through time. It also requires specifying the
criteria that economic agents use in choosing appropriate patterns of such
variables as consumption, investment and work effort. It is important in
developing a model of this sort to recognize that business cycles are
fundamentally phenomena that are characterized by their behavior through
time. For example, when we think of business cycles, we frequently think
about notions of persistence or serial correlation in economic aggregates;
comovement among economic activities; leading or lagging variables
relative to output; and different amplitudes or volatilities of various series.
The objective of any model of the business cycle is to generate a coherent
understanding of how and why these characteristics arise. Thus a model of
fluctuations must be dynamic at its most basic level and not a collection of
anecdotal behavioral rules attached to an otherwise static framework.
The neoclassical model of capital accumulation
The most basic model of economic dynamics is the neoclassical model of
capital accumulation. While many readers may be familiar with some
Understanding real business cycles 399
versions of this framework as a model of optimal economic growth—
following the work of Cass (1965), Koopmans (1965) and Solow (1956)—it is
better viewed as framework for economic dynamics (see Hicks 1965:4). As
such it is natural to consider it as the benchmark model for our
understanding of economic fluctuations as well as growth.4 What is
somewhat remarkable is that the implications for fluctuations of this
neoclassical approach have not been seriously explored until recently.5
A simple economic environment to consider is an economy populated by
many identical agents (households) that live forever. The utility of each
agent is some function of the consumption and leisure they expect to enjoy
over their (infinite) lifetimes. Each agent is also treated as having access to a
constant returns to scale production technology for the single commodity in
this economy. The production function requires both capital, which
depreciates over time, and work effort. In addition, the production
technology is assumed to be subject to temporary productivity shifts or
technological changes which provide the underlying source of variation in
the economic environment to which agents must respond. For simplicity,
assume that these shifts, past and future, are known with certainty to all
agents and thus agents have perfect foresight. The choices each consumer
must make are how to allocate their hours between work and leisure, and
how to allocate their supply of the single good between investment in future
capital and current consumption. Of course, the model imposes resource
constraints such that the sum of consumption and investment is less than or
equal to output and the sum of time spent working and at leisure is less than
or equal to some fixed amount of time in the period. Consumption, labor,
leisure, capital and investment must also be nonnegative. The Appendix
presents a mathematical summary of such a model.
This model is clearly simple and unrealistic, but for present purposes that
is an advantage. After all, the model is not intended to capture a complex
reality, but, at this point, only to provide a benchmark of the features of a
dynamic market equilibrium. It is a purely real model, driven by technology
or productivity disturbances and hence, following Long and Plosser (1983), it
has been labeled a real business cycle model. But despite this model’s
simplicity, the equilibrium behavior of the model exhibits many important
characteristics that are generally associated with business cycles.
Equilibrium outcomes
How does one think about the competitive equilibrium prices and quantities
that are implied by this framework? The first step is to recognize that all
individuals are alike, thus it is easy to imagine a representative agent,
Robinson Crusoe, and ask how his optimal choices of consumption, work
effort and investment evolve over time. Do these optimally chosen quantities
correspond to the per capita quantities that would be produced by a
competitive equilibrium involving many agents interacting in the markets for
400 Charles I.Plosser
current and future goods and labor? The answer to this question is provided
to us by Debreu (1954) and Prescott and Lucas (1972) in the affirmative. In
other words, we can interpret the utility maximizing choices of consumption,
investment and work effort by Robinson Crusoe as the per capita outcomes
of a competitive market economy.
Robinson Crusoe’s choice problem is to maximize his lifetime utility
subject to the production technology and a sequence of resource constraints,
a problem that can be viewed in the familiar framework of constrained
optimization. (See the Appendix for more detail.) Given specific functional
forms for the utility function and the production function, some initial
conditions and the sequence of productivity disturbances, one could, in
principle, derive a set of decision rules that describe Robinson’s optimal
consumption, work, and investment decisions in terms of the current
(predetermined) capital stock and the past and future productivity
disturbances. These decisions, in turn, imply an amount of total output via
the production function. The optimal quantities also imply market prices for
labor (a real wage) and one-period loans (a real interest rate). Another
important characteristic of these models is that in the absence of productivity
disturbances, Robinson Crusoe’s optimal choice of consumption, work effort,
investment, and thus output will, under a broad set of conditions, converge
to constant or steady state values.
Under most specifications of preferences and production functions, it is
impossible to solve analytically this maximization problem for the optimal
decision rules of Robinson Crusoe. Consequently, real business cycle
researchers find it necessary to compute approximate solutions to Robinson
Crusoe’s choice problem in the neighborhood of the steady state. The
approximate decision rules are linear functions of the predetermined capital
stock and all productivity disturbances. The details of the procedures
available to compute the approximately optimal quantities and competitive
prices from this framework are beyond the scope of this chapter but the
economic intuition underlying the resulting optimal decisions is relatively
straightforward.6
Responses to productivity disturbances
Imagine Crusoe observes a temporarily high value of productivity. How will
he respond? One option would be for him to consume the above normal
output holding investment and work effort fixed. This is clearly a feasible
outcome, but one that says shocks are totally absorbed within a period and
thus have no implications for future decisions or outcomes. A moment’s
reflection, however, suggests that Crusoe values future consumption and
leisure in addition to current consumption and, opportunities/technology
permitting, would prefer to consume more output in the future as well as
today. This intertemporal transfer can be accomplished in this setting
because the production function permits Crusoe to invest in capital that will
Understanding real business cycles 401
help produce output in subsequent periods. Thus investment should respond
positively to the temporary shock. The effect on work effort is ambiguous.
Current productivity is temporarily high which encourages intertemporal
substitution of current for future work and intertemporal substitution of
current consumption for leisure. Wealth, on the other hand, is higher and
that acts to reduce current and future work effort. For plausible
parameterizations of the model the substitution effect dominates so that
current work effort rises. Thus the temporary shock is propagated forward
and the effects of the shock show up in higher output, consumption and
leisure in the future. This simple intuition illustrates why variables like
output and consumption are likely to be serially correlated even when shocks
to the environment are uncorrelated and purely temporary.
If the productivity shock observed by Robinson Crusoe is more long-lived or
persistent, then his responses would be different. For example, a more persistent
increase in productivity would tend to raise wealth more significantly by raising
future output. Robinson Crusoe’s incentive to increase investment would
plausibly be reduced and his incentive to increase current consumption would be
increased. There would also be less incentive to work harder today because the
wealth effect is stronger and the intertemporal substitution effect is reduced.
Quantitative results require a more specific formulation.7
Thus, a productivity disturbance results in a dynamic response by
Robinson Crusoe that involves variations in output, work effort, consumption
and investment over many periods. It is important to stress that there are no
market failures in this economy, so Robinson Crusoe’s response to the
productivity shifts are optimal and the economy is Pareto efficient at all
points in time. Put another way, any attempt by a social planner to force
Crusoe to choose any allocation other than the ones indicated, such as
working more than he currently chooses, or saving more than he currently
chooses, are likely to be welfare reducing. Therefore, business cycle
characteristics exhibited by this economy are chosen in preference to
outcomes that exhibit no business cycles.
The decision rules summarize the solution to Robinson Crusoe’s dynamic
optimization problem. As stated above they depend explicitly on the current
and future productivity disturbances. In richer models that include
government (see below), these decision rules would also depend on current
and future actions of the government. Consequently, these rules provide the
basis for evaluating policy in a manner that is not subject to the criticism
Lucas (1976) levied on models that possess simple behavioral relations
among current and past economic variables that are assumed to be invariant
with respect to changes in the actions of government.
Supply or demand
It is common to refer to these real business cycle models as models that are
driven by aggregate ‘supply shocks’.8 While such a description seems
402 Charles I.Plosser
approximately accurate for the model driven by productivity shifts, and thus
innocuous enough, it is potentially misleading. In the first place trying to
think about these dynamic general equilibrium models in terms of supply
and demand is slippery. In these models shocks occur to either preferences,
technologies/opportunities, or resources and endowments. Unfortunately,
these shocks do not easily translate into either supply or demand
disturbances. Each type of shock will generally affect both the supply and
demand schedules in a particular market. For example, shifts in technology
influence both the supply of goods for a given level of inputs (work effort in
particular), and the demand for goods through its effect on wealth and the
labor/ leisure decision.
Second, while most of the analyses to date have focused on the version of
the model where variations in technology are the source of changes in the
environment, one could just as easily specify the changes as arising from
variations in preferences or tastes. This would lead to a real business cycle
model driven by what some would label as ‘demand shocks’. In addition, the
model can be expanded to include a government sector (discussed further
below) that could also be considered a source of ‘demand shocks’.9 Thus
there is nothing inherent in the real business model that limits it to the
analysis of variations in technology or supply.
Stochastic models and uncertainty
The discussion, at several points, has noted the explicit dependence of
Robinson Crusoe’s decisions on the future path of productivity. It is natural to
ask if the framework can be adapted to handle uncertainty, where the
productivity disturbance is a random variable whose future values are
uncertain. The answer to this question is yes and is based on the seminal
work of Brock and Mirman (1972). As in the certainty case discussed above,
however, analytical solutions for the decision rules under uncertainty are
rare.10 It has been common practice to rely on what is called certainty
equivalence. This procedure takes the linear decision rules obtained as the
approximate solution to the certainty model and replaces the future
productivity disturbances with their conditional expected value given
information available at time t.11 The resulting set of time paths for
consumption, work effort and capital are then a linear rational expectations
equilibrium rather than a perfect foresight equilibrium.
ECONOMIC GROWTH AND BUSINESS CYCLES
The neoclassical model of capital accumulation outlined in the previous
section predicts that per capita values of output, capital and consumption
will, in the absence of disturbances to productivity, converge to constants or
steady state values. The evidence, however, is that per capita values in the
United States and most other industrialized countries grow continually over
Understanding real business cycles 403
time. For example, from 1954 to 1985, per capita real GNP grew at an
average annual rate of about 1.5 percent. The basic neoclassical model does
not offer an explanation of this sustained growth in per capita values.
In a classic paper, Robert Solow (1957) argued that technical change, in
addition to the capital per worker, was an important source of variation in
output per capita.12 Solow constructed estimates of US technological change
using data from 1909 to 1949. He concluded that productivity grew at an
average rate of 1.5 percent per year during the period. Output per capita, on the
other hand, grew at an average annual rate of 1.7 percent. Solow then argued
that these productivity changes were empirically uncorrelated with changes in
capital per worker. He concluded that about 85 percent of the real per capita
growth during this period was accounted for by technological change or
productivity and only about 15 percent by increases in capital per worker. Thus,
based on Solow’s evidence, one would conclude that changes in productivity and
technology are the major factors determining economic growth.
While technological progress has been recognized as an important factor
determining economic growth, at least since Solow’s seminal work, it has
been common to think of economic growth as something that can be studied
independently of economic fluctuations. Or to put the point another way, it is
often presumed that the factors that influence growth have only second order
implications for economic fluctuations. In fact the use of the phrase ‘growth
theory’ was an intentional attempt to distinguish it from a theory of the
business cycle. As stressed by Hicks (1965:4), however, there is no
compelling economic rationale underlying this view.
The distinction between trend and fluctuation is a statistical distinction; it
is an unquestionably useful device for statistical summarizing. Since
economic theory is to be applied to statistics, which are arranged in this
manner, a corresponding arrangement of theory will (no doubt) often be
convenient. But this gives us no reason to suppose that there is anything
corresponding to it on the economic side which is at all fundamental. We
have no right to conclude, from the mere existence of the statistical
device, that the economic forces making for trend and for fluctuation are
any different, so that they have to be analyzed in different ways. It is
inadvisable to start our economics from the statistical distinction, though
it will have to come in at an appropriate point, as an instrument of
application.
Nevertheless, it has been common to think of business cycle models as
separate from models of economic growth and to characterize business cycles
as the deviations from some smooth, usually deterministic, trend that proxies
for growth. Theories of the business cycle are then constructed to explain
these deviations. Thus, while rarely explicitly recognized, tests of these
business cycle theories are actually joint tests of the model for growth (the
trend) and the model for the cycle.
404 Charles I.Plosser
Nelson and Plosser (1982) argue that real per capita output, as well as
many other economic time series, behave as if they have random walk
components (much like the log of stock prices). Random walks have the
important property that there is no tendency for the process to return to any
particular level or trend line once displaced. Thus, unpredicted shocks to
productivity permanently alter the level of productivity. Nelson and Plosser
also argue that Solow’s technology series also behaves like a random walk.
The observation that the log of productivity follows a random walk with
drift (drift meaning the changes have a non-zero mean) has some important
implications. First, a random walk is a nonstationary stochastic process,
which means that it possesses no affinity for any particular mean. Random
walks are also referred to as stochastic trends because while they may
exhibit growth, they do not fluctuate about any particular deterministic path.
If shocks to productivity are permanent, each one determines a new growth
path. Therefore, detrending economic time series with a deterministic time
trend and then assuming that the deviations from the trend will exhibit some
tendency to return to the trend line would be econometrically incorrect and
may be quite misleading.13
Second, the fact that productivity grows over time raises some additional
complications for the neoclassical model described in the previous section. In
particular, if productivity is growing then output, consumption and capital
per capita will also tend to grow over time. If, for example, output and
consumption grew at different rates, in the long run, then the consumption/
output ratio would be driven to zero or one. To prevent this, it is usually
required that these per capita values converge to constant, but equal growth
rates, so that the model possesses steady state growth. In addition, work
effort cannot grow in the steady state because available hours are bounded
from above and below. For these restrictions to be satisfied additional
requirements must be placed on the form of the production process and
utility function. Of particular importance is the requirement that permanent
technological progress must be expressible as labor augmenting or Harrod-
neutral.14
Third, King et al. (1988b) and King et al. (1987) show that the
neoclassical model with random walk technological progress implies that
output, consumption and investment per capita will all contain a common
random walk component or stochastic trend. This structure is consistent with
the empirical observations of Nelson and Plosser discussed above. In
addition, King et al. (1987) investigate the common stochastic trend
implication for output, consumption and investment and conclude that it
provides a reasonable representation of the data. As noted above, hours
worked per capita will not contain a stochastic growth component since the
number of available hours per time period is in fixed supply.
If these labor augmenting productivity shifts can be characterized as
the engine of economic growth, what does the simple neoclassical model
of optimal capital accumulation predict about the response of output,
Understanding real business cycles 405
consumption, investment, work effort and wages to these technological
shifts? The permanent change in productivity sets in motion a series of
dynamic responses that move Robinson Crusoe and the economy towards
a new growth path. For example, 1 percent permanent (once and for all)
change in labor productivity in the long run leads to a one percent
permanent increase in the level of capital stock, consumption, output and
investment once the transitory dynamics have been dissipated. These
transitory dynamics are important for understanding fluctuations. They
are initiated by the requirement that the economy must move to a
permanently higher capital stock. To get there requires substantial
increases in investment in the near term that taper off to a new higher
steady state level as the economy converges to the higher capital stock.
There will also be gradual increases in consumption and output towards
their respective higher steady state levels. Work effort will also be
temporarily high along the transition path. While wealth has increased,
which discourages current work effort, productivity is also higher which
encourages work effort. Productivity is higher because the desired or
steady capital stock has risen. Thus in the near term real interest rates
rise, which induces intertemporal substitution of current for future work
effort. The responses, and thus the fluctuations that are present in the
model, are the result of the same factors that generate economic growth.
The real business cycle model, therefore, provides an integrated approach
to the theory of growth and fluctuations.
REAL BUSINESS CYCLES AND THE 1954–85 US ECONOMY
The simple neoclassical model described earlier is clearly an incomplete
model of the US economy. Nevertheless, useful insights into the properties
of the model can be obtained by providing a more quantitative assessment
of the model’s explanatory power. The strategy is to choose explicit
functional forms for Robinson Crusoe’s utility function and production
function and then to compute the approximate equilibrium behavior of
output, consumption, investment, work effort and wages implied by the
model when the technology shifts are computed following Solow. These
predicted series can then be compared to the actual performance of the US
economy.
The first step is to specify explicit functions for the production technology
and preferences. A natural choice for the production function that also
satisfies the restrictions necessary for steady state growth is the Cobb-
Douglas formulation. There is some latitude in the choice of Robinson
Crusoe’s preferences. King et al. (1988a) derive the class of admissible
preference functions if the economy is to possess steady state growth. One
admissible utility function is logarithmic preferences. Based on these
specifications of preferences and technology and the random walk properties
of the technology shifts, approximate optimal decisions of Robinson Crusoe
406 Charles I.Plosser
can be obtained and used to calculate how he will respond to the Solow
technology shifts.15
Summary statistics for the US economy
Table 17.1 highlights some of the statistical properties of postwar business
fluctuations. The period begins in 1954 in order to avoid potential
complications raised by the very high levels of government spending during
the Korean War. Output Y is real nonfarm business product per capita and
hours N is the average fraction of the week spent working by nongovernment
employees per capita. The remaining empirical counterparts to the variables
in the model are: consumption, C, the sum of real consumption of
nondurables and services per capita; investment, I, the sum of real
nonresidential fixed investment and real consumption of consumer durables;
and the real wage rate, w, the real average hourly earnings of all production
workers.
There are, of course, a number of ways of summarizing these types of
data. I have chosen the typical practice of using sample moments to describe
the central characteristics. Growth rates are chosen because the model
predicts that log levels will possess stochastic trends (or random walk
components) so that population moments do not exist. While virtually all
empirical investigations of business cycles start by detrending the data, the
real business cycle model I have described here integrates growth and
fluctuations and provides the detrending instructions to obtain variables that
possess well-defined distributions.
The moments presented in Table 17.1 are the sample means, standard
deviations, serial correlation (autocorrelation) coefficients and correlations
with output. The mean growth rate of output and consumption is about
Table 17.1 Summary statistics 1954–85
Note: a The approximate standard error of the estimated autocorrelations is 0.18
Understanding real business cycles 407
1.5 percent per year. Wage growth is less and investment growth is
somewhat more. Hours, on the other hand, exhibit no growth at all and
actually fall by about 0.1 percent per year. Standard deviations provide
information on the relative volatility of the different series. Investment
growth is the most volatile followed by output, hours, wages and
consumption respectively. Autocorrelations measure the amount of
persistence of the series from one year to the next. For example, the
correlation coefficient between the growth in consumption in one year is
about 0.4 with the previous year’s growth in consumption. Only real wages
and consumption show much evidence of persistence in growth rates. Finally,
all the series are highly correlated with output and thus are procyclical. The
lowest correlation with output is exhibited by real wage growth with a
correlation coefficient of 0.59 and the highest is investment with a
correlation coefficient of 0.92.
Productivity shifts
In order to see more quantitatively the sorts of real economic fluctuations
generated by the simple model economy it is necessary to obtain some
measure of the productivity shocks. A crude but straightforward method is to
follow the example provided by Solow to construct a measure of the state of
productivity. Using the data described above and the gross stock of real
nonresidential fixed private capital, a Solow technology series is readily
constructed.16 The annual percentage rate of change in technology is plotted
in Figure 17.1. The picture corresponds to most observers’ impressions that
productivity growth was on average higher in the 1960s than the 1970s and
1980s. The growth rate of this 32 year period averages 0.8 percent per year
and has a standard deviation of about 1.9 percent. The maximum growth
rate is about 4.0 percent and the minimum is about -3.5 percent.
Figure 17.1 Annual growth rate of technology
408 Charles I.Plosser
There is only slight evidence of serial correlation in these growth rates so to
a first approximation it seems acceptable to view the level of productivity as
a random walk.
These computed productivity disturbances may or may not be very good
estimates of the true changes in productivity. However, the real business
cycle model delivers explicit and tight restrictions on the behavior of
consumption, hours worked, investment, and thus output, conditioned on the
disturbances to the model being of a technological source. If the measured
technological shocks are poor estimates (that is, if they are confounded by
other factors such as ‘demand’ shocks, preference shocks or change in
government policies, and so on) then feeding these values into our real
business cycle model should result in poor predictions for the behavior of
consumption, investment, hours worked, wages and output.
Real business cycles
Given the form of preferences and technology, the model is used to obtain the
responses of the simple neoclassical model to observed productivity shifts.17
These results are summarized in Panel B of Table 17.1. The model produces
sample means that are very close to the data for output, consumption and
hours, but is too low for investment and too high for wages. The model
generates the same volatility rankings for Y, C, I, but the absolute standard
deviation of investment is slightly lower and that for consumption is slightly
higher than in the actual data. The major discrepancy appears to be that in
the model the growth rate of hours has a standard deviation that is less than
one-half of that in the data. In terms of serial correlation properties, the
model generates slightly, but perhaps not significantly, more positive
autocorrelation than seems present in the data.
Figure 17.2 Annual growth rate of real output
Understanding real business cycles 409
Perhaps the numbers of most interest in the table are those in the last column
of Panel B. These are the correlation coefficients of the predicted outcomes
with the actual series and range from 0.52 for wages to 0.87 for output. To
many economists, the whole idea that such a simple model with no
gover nment, no money, no market failures of any kind, rational
expectations, no adjustment costs and identical agents could replicate actual
experience this well is very surprising. This is especially true given that most
macroeconomic research since the late 1930s stressed the importance of one
or more of the above factors in explaining business fluctuations.
Figures 17.2 to 17.6 provide a visual impression of these correlations by
plotting both the actual and predicted growth rates of each of the five
Figure 17.3 Annual growth rate of real consumption
Figure 17.4 Annual growth rate of real investment
410 Charles I.Plosser
variables. As expected from the evidence presented in Table 17.1, the growth
rate of hours worked exhibits the biggest discrepancy between actual and
predicted. Nevertheless, the simple model appears to replicate a significant
portion of the behavior of the economy during recessions as well as other
periods.
GOVERNMENT POLICIES AND SUBOPTIMAL EQUILIBRIUM
Two key features of the real business cycle model discussed so far are that
business cycles are initiated by shocks to technology and that fluctuations are
Figure 17.5 Annual growth rate of hours worked
Figure 17.6 Annual growth rate of real wage rate
Understanding real business cycles 411
Pareto optimal. Neither of these conditions, however, are necessary features
of the real business cycle approach. Many economists, for example, argue
that government tax and spending policies are an important source of real
disturbances to the economic system. The incorporation of government into
the real business cycle models makes it possible to address important
questions regarding changes in fiscal policies in the presence of distortionary
taxes. Of particular interest is the case where the tax and spending policies
are functions of the state of the economy. Variation in government spending
introduces a potential source of demand disturbances to the model. The
presence of distortionary taxes generally breaks the link between Robinson
Crusoe’s optimal decisions and Pareto efficiency, since removing the
distortions will usually raise welfare. Nevertheless, competitive equilibria
can be computed and analyzed that are not Pareto optimal but suboptimal
equilibria.
The theoretical underpinning of this line of research draws from earlier
work by Arrow (1962), Hall (1971) and Brock (1975) and a series of papers
by Romer (1983, 1986, 1987). The basic line of reasoning is that in an
economy with many agents, each can take the government’s spending and
taxing policies as given in their choice problem. The only additional
restriction is that aggregate behavior satisfy the government’s budget
constraint. These models provide an artificial laboratory for answering
questions regarding policy changes that is not subject to the criticism of
Lucas (1976).
The intuition underlying the effects of unproductive (as assumed in most
Keynesian analyses) government purchases in the neoclassical model is
basically found in Barro (1981) and Hall (1980).18 These authors emphasize
two sorts of influences. First, raising government purchases induces a
negative wealth effect that acts to reduce consumption and raise work effort
and output. Second, raising government purchases also induces intertemporal
substitution when the increase is temporary. This results in lower
consumption, lower investment, higher work effort and higher output. The
relative importance of the wealth and intertemporal substitution channels
remains unresolved. Barro and Hall assume that the intertemporal
substitution channel is quantitatively more important so that temporary
changes in government purchases are more important than the wealth
channel. Baxter and King (1988) have investigated these effects within a real
business cycle model and have concluded that for plausible values of the
parameters more persistent changes in government purchases have larger
output ‘multipliers’ than more temporary changes in purchases. Temporary
purchases on the other hand, have a more negative impact on investment
than more persistent purchases.19
The implications of distortionary taxation within the neoclassical model
have been a topic in public finance for some time. What distinguishes the
more recent work from the earlier efforts, including Hall (1971) and analyses
by Abel and Blanchard (1983) and Judd (1985) is that tax rates are assumed
412 Charles I.Plosser
to be functions of the state of the economy. King et al. (1988b) summarize
the implications of a real business cycle model under a period-by-period
balanced budget, where tax revenue is based on an output tax and
government spending is rebated as lump-sum transfers. In this case a positive
productivity shift requires a decline in the tax rate in order to maintain
budget balance. This reduction in tax rates reinforces the efforts of the
productivity shock on after-tax labor productivity and further increases work
effort in response to technology shocks. Thus work effort (and investment, for
analogous reasons) are more volatile in this economy.
THE REAL BUSINESS CYCLE RESEARCH AGENDA
The results in the previous sections indicate that the basic neoclassical model
of capital accumulation can provide an important framework for developing
our understanding of economic fluctuations. The models investigated to date,
however, are not entirely satisfactory. Indeed, it would be extraordinary if
they were. The real business cycle research program is to pursue this class of
models to determine how far the approach can take us. In this section, I
highlight some of the issues that are likely to be important for developing
and evaluating this important class of models.
Multi-sector extensions
The basic neoclassical model has been explored along various dimensions in
an effort to expand the scope of the method of analysis. Long and Plosser
(1983) explore a model with multiple sectors in order to understand the
comovement across sectors in response to shocks that are potentially sector
specific. Their interest in multiple sectors is motivated by the observation
that many sectors of the economy tend to move together but some sectors
lead while other sectors lag the general state of business activity. Multi-
sector models are the only way to address this phenomenon and understand
it since one-sector models proceed by assuming that the answer is the
existence of aggregate or common shocks.20
Black (1987) argues that multi-sector models are important, particularly if
unemployment is to be explained. Black bases his argument on the notion
that both human and physical capital is highly specialized. Shocks to either
preferences or technologies will generally require resources in the form of
labor and capital to move between sectors. Since these inputs are specialized,
it will be costly to make this adjustment. As a result, unemployment can be
expected to rise above its long-run level.
Labor markets
A major thrust of much research in the real business cycle area is to expand
and extend the basic model in ways that would result in a better match of the
Understanding real business cycles 413
model’s predictions for hours and actual hours worked. The source of conflict
is that the logarithmic preferences adopted for the purpose of the earlier
estimates imply a labor supply elasticity that is much higher than the
estimates obtained by labor economists using panel data on prime age
males. Thus, the model appears to be incapable of generating sufficient
volatility in hours without being in conflict with evidence from detailed
microeconomic investigations. This view, however, is unduly pessimistic.
Numerous approaches have been pursued (though none has been completely
satisfactory to date) that attempt to modify the model in ways that make it
compatible with the microeconomic evidence.
One approach pursued by Kydland and Prescott (1982) stresses the
importance of preference structures that are not time separable. In their
formulation, the current utility of leisure depends on past leisure in an explicit
way. This has the effect of permitting an increase in the intertemporal
substitutability of leisure which in turn makes hours worked more volatile.
Rogerson (1988) and Hansen (1985) explore the consequences of
indivisibilities in the labor supply decision that require agents to work either
full-time or not at all. This is in contrast to the simple model where agents are
permitted to vary hours worked continuously. The result is that the volatility
of hours worked in response to productivity shifts is significantly increased
while estimated labor supply elasticities would remain low for working males.
Another approach to enhancing the response hours worked in the model is
to allow for heterogeneity across agents in the economy. Examples of this
approach are found in Cho and Rogerson (1988), Kydland (1984), King et al.
(1988b) and Rebelo (1987). All of these papers suggest that there can be
important downward biases in estimates of aggregate labor supply elasticity
when there are agents with different skill levels.
Endogenous growth
Another important area for research focuses on the role played by the
technology shocks. Solow’s view of technological change included anything
that shifted the production function other than measurable capital or labor.
As an empirical proposition, Solow’s results indicate that such shifts, if
viewed as exogenous, account for a substantial portion of economic growth.
The real business cycle model stresses that these shifts play an important role
in economic fluctuations as well. This is not entirely satisfactory. It would be
useful if we had a better understanding of the economics of growth that did
not rely as heavily on such an exogenous unobservable process.
Work by Uzawa (1965), Romer (1986) and Lucas (1988) modifies the basic
neoclassical model to permit growth to be an endogenous outcome of the
technology. The key to obtaining such a result is eliminating the diminishing
returns in the production process. King and Rebelo (1986, 1988) provide
examples of this strategy and explore its implications for economic
fluctuations and certain types of fiscal policies. The idea is to permit human
414 Charles I.Plosser
capital (labor-augmenting technical change) to be produced using physical
capital and human capital as inputs to a constant returns technology. The
results are interesting and potentially important. For example, purely
temporary productivity shifts can have permanent effects on the level of
economic activity. The reason is that a change in productivity that results in
more output will generally result in some increased resources being allocated
to the production of additional human capital. Thus allocation decisions affect
the level of technology and the growth in the economy. These models have the
additional implication that such variables as output, consumption and
investment are integrated or possess a stochastic trend. This result is appealing
because as noted above, Nelson and Plosser (1982) have argued that many
economic time series appear to possess stochastic trends or random walk
components. Finally, productivity shocks in these models can initiate complex
patterns of adjustment to a new growth path. These transition paths generally
include complex changes in work effort, investment and consumption. An
understanding of these models is likely to be an important part of
understanding economic fluctuations, as well as economic development, while
reinforcing the concept that growth and fluctuations are intimately connected.
Money
Real business cycle research has focused almost exclusively on models with
no role for money. For some economists, this not only doesn’t represent
progress, but also borders on blasphemy. My view, and that of many other
real business cycle researchers, is that the role of money in an equilibrium
theory of growth and fluctuations is not well understood and thus remains an
open issue. Some researchers, including King and Plosser (1984), Kydland
(1987), Eichenbaum and Singleton (1986) and Cooley and Hansen (1988),
have explored methods of incorporating money and investigating its
implications in a real business cycle model. Unfortunately there is little
agreement on what constitutes the most fruitful approach at this time.
Nevertheless, without an understanding of the real fluctuations inherent in
the basic neoclassical model without money it will be difficult if not
impossible to measure the quantitative importance of money in actual
business fluctuations. The nature and magnitude, however, of the fluctuations
and responses in the real neoclassical model means that real business
research poses a challenge to conventional views regarding the relative
importance of money. This is particularly true given the difficulties
economists have faced in developing a convincing and coherent explanation
of the monetary transmission mechanism.
Strategies for estimation and hypothesis testing
A final part of the research agenda relates to empirical assessments of the
real business cycle approach. The approach adopted to date takes as given
Understanding real business cycles 415
the technological shocks and asks how other variables—such as
consumption, work effort and investment—respond over time to these
impulses. It would be useful to obtain an independent measure of these
shocks or identify observable variables that could proxy for them. Expanding
the models to include government, endogenous growth or international trade
are important steps in this process.
A closely related topic, and one of intense debate among researchers, is
the strategy used to investigate the implications of the model. The traditional
way of estimating and testing an economic model is to write down a set of
structural equations, estimate the parameters and test any restrictions not
necessary to identify parameters. In the context of the real business cycle
models described earlier this strategy corresponds to obtaining Robinson
Crusoe’s optimal decision rules for consumption, work effort and investment,
jointly estimating the parameters of technology and preferences and then
testing the overidentifying restrictions imposed on these decision rules.21
An alternative strategy has been pursued in much of the real business
cycle literature. The technique, made popular in this literature by Kydland
and Prescott (1982) (but more widely employed in the applied general
equilibrium literature like Ballard et al. 1985), is called ‘calibration’. The
strategy is to choose values for certain key parameters of the underlying
preferences and technologies using evidence from other empirical studies.
This restricts the number of free parameters in the model. Using these
parameter values, the stochastic properties (means, variances,
autocorrelations and cross-correlations) of certain key variables are
constructed. The remaining free parameters are chosen to yield, as closely as
possible, a correspondence between the moments predicted by the model and
those in the sample data. A formal definition of what constitutes a good fit or
the metric along which fit should be judged is not explicitly offered by
Kydland and Prescott.22
The appropriate empirical strategy for investigating the class of models
discussed in this chapter remains an open area of research. Ultimately, this
issue must be addressed and real business cycle models will have to face and
pass more stringent empirical tests than they have to date.
CONCLUSIONS
The basic framework of real business cycle analysis is the neoclassical model
of capital accumulation. This is the natural starting point to begin the study
of dynamic fluctuations. While frequently interpreted as a model of
economic growth, the neoclassical model generates fluctuations in response
to external disturbances that resemble business cycles. While real technology
shocks have occupied the central focus in the literature, other shocks arising
from preferences, government, terms of trade and eventually money can be
included. Thus real business cycle models do not have to be confined to
analyzing only technological or productivity shocks. Nevertheless, these real
416 Charles I.Plosser
technological disturbances generate rich and neglected dynamics in the basic
neoclassical model that appear to account for a substantial portion of
observed fluctuations.
Real business cycle theory is still in its infancy and thus remains an
incomplete theory of the business cycle. Yet the progress to date has had a
significant impact on research in macroeconomics. In particular, simple real
business cycle models have demonstrated that equilibrium models are not
necessarily inconsistent with many characteristics attributed to the business
cycle. In so doing these models have changed the standard by which
macroeconomic theories are judged and provided the foundations for an
understanding of business cycles that is based on the powerful choice
theoretic analysis that is at the core of economic reasoning. The appeal of
this line of research is the apparent power of some very simple economic
principles to generate dynamic behavior that was heretofore thought to be
incompatible with any notion of equilibrium. While the promise is great,
much work remains before economists have a real understanding of business
cycles.
APPENDIX: SPECIFYING A MODEL OF REAL BUSINESS CYCLES
This appendix presents a more analytical summary of the basic neoclassical
model discussed in the second section of the chapter. The first step is to
specify the economic environment by describing the preferences, technology
and endowments of the model economy.
The neoclassical model
Preferences
The economy is assumed to be populated by many identical agents (house-
holds) that live forever.23 The agents’s utility at time t is assumed to be of the
form , where C
t
is the level of consumption of the
single produced good and L
t
is the amount of leisure consumed. The utility
discount factor is assumed to be constant. Leisure is included because
variation in work effort is an important feature of short-run fluctuations and
yet is frequently absent from otherwise similar models encountered in the
growth literature. The momentary utility function u(·) is assumed to be
concave and twice continuously differentiable.
Production
The single final good, Y
t
, is produced by a constant returns to scale
production technology given by Y
t
=Θ
t
F(K
t
, N
t
), where Kt is the
predetermined capital stock (chosen at t–1) and N
t
is labor input in period t.
Θ
t
is a temporary shift factor that alters total factor productivity. The
Understanding real business cycles 417
produced commodity Y can either be consumed or invested. The production
function is also assumed to be concave and twice continuously differentiable.
Capital accumulation
The invested commodity becomes part of the capital stock that is available
on input to production next period. This capital stock evolves as Kt+1=(1–
δ)Kt+It, where It is gross investment and d is the depreciation rate of capital.
Resource constraints
Agents also face resource constraints in each period on the use of the
commodity and time. These constraints are Lt+Nt≤1 and Ct+It≤Yt, where the
time endowment is normalized to unity. These are non-negativity constraints
Lt, Ct, Nt and Kt as well.
The computation of the competitive equilibrium prices and quantities that
are implied by this framework is simplified by recognizing that all
individuals are alike. Thus, it is easy to imagine a representative agent,
Robinson Crusoe, and determine how his optimal choice of consumption,
work effort and investment evolve over time. Debreu (1954) and Prescott and
Lucas (1972) have shown that we can interpret the utility maximizing
choices of consumption, investment and work effort by Robinson Crusoe as
the per capita outcomes of a competitive market economy.
Robinson Crusoe’s choice problem is to maximize his lifetime (infinite)
utility subject to a sequence of resource constraints. The Lagrangian
associated with the maximization problem is
where 1–Nt is substituted for Lt, Kt+1–(1–δ)Kt is substituted for It,ΘtF(Kt, Nt) is
substituted for Yt, and �t, is the Lagrange multiplier associated with the
period t resource constraint Yt-Ct-It=0.
The first-order efficiency conditions for this problem are obtained by
differentiating with respect to the variables of choice at each time t, Ct,
Nt, Kt+1 and the multiplier, which yields
which must hold for all t=1, 2,…∞. Fi(·) and ui(·) denote the partial derivatives of
F and u with respect to the ith argument. In addition, it is common to assume
that the transversality condition, limt 8ß
t
�tKt+1=0, is satisfied.
418 Charles I.Plosser
Given specific functional forms for u(·) and F(·) the solution to this
maximum problem is the time paths of the four unknown choice variables,
C, N, K and � that satisfy these efficiency conditions for some initial
condition K0 and a sequence of productivity disturbances
24 These
time paths can be expressed in the form of time invariant decision rules that
take the form
The competitive market prices implied by these optimal quantities are a real
interest rate between t and t+1, rt, and a real wage rate, w. These are readily
determined to be (1+rt)=�t/(�t+1ß) and wt=ΘtF2(Kt, Nt) and are the ones that
would prevail in the spot market for labor services and a one-period
sequential loan market.25 Another important feature of this economy is that
in the absence of changes in technology (i.e. Θt=0, for all t), and given some
initial capital stock, per capita values of consumption, hours, capital and
output, converge to constants, referred to as the steady state.
Approximate solutions
Under most specifications of preferences and production functions, the four
first-order conditions given earlier constitute a set of nonlinear difference
equations. Thus it is usually impossible to solve this maximum problem
analytically for the optimal decisions rules of Robinson Crusoe.
Consequently, real business cycle researchers find it necessary to compute
approximate solutions to Robinson Crusoe’s choice problem. These
approximation procedures typically result in decision rules that are linear K
t
and the Θ’s. The details of various procedures available to compute these
approximately optimal quantities and competitive prices are beyond the
scope of this chapter. Nevertheless, the basic idea pursued in King et al.
(1988a) is intuitively straightforward and is the method employed in the
text.26
The first step in the approximation procedure is to choose a point to
approximate around. The natural choice is the stationary point or steady
state, denoted [CS, KS, NS, �S]. The second step is to express the four first-
order conditions in terms of the percentage deviations from the stationary
values (defined as etc.) and then take a linear approximation to each
condition. This results in a set of linear difference equations in percentage
deviations from the steady state.
Solving this linear system produces the approximately optimal decision
rules that correspond to the three time-invariant decision rules.27 These
decision rules are linear functions of the predetermined capital stock and the
Understanding real business cycles 419
sequence of productivity shifts. For example, efficient capital accumulation
can be written as
where µ1, µ2, �1 and �2 are complicated functions of the underlying
parameters of tastes and technology. Thus next period’s capital stock depends
on the current capital stock, the current level of productivity t and the
entire future path of shifts discounted by µ2. The conditions on the problem
pretty much guarantee that µ1<1 and µ2>1. The (approximately) optimal
decision rules for t and t take similar forms.
28
An example economy
In the text a specific example economy is used to quantitatively measure the
responses of a real business cycle model to estimated productivity shifts. As
indicated in the text preferences are taken to be logarithmic such that u(Ct,
Lt)=log(Ct)+ηLlog(Lt). The production technology is taken to be Cobb-
Douglas , or, expressing the technology shift Θt as labor-
augmenting . The technology shifts are computed
following Solow and are assumed to follow a logarithmic random walk for
purposes of computing the approximate optimal decisions.
The remaining parameters are chosen assuming the time interval is one
year and correspond to those used in King et al. (1988a, 1988b). Labor’s
share (α=0.58) is computed as the average ratio of total employee
compensation to GNP for the period 1948–85. Depreciation (δ=0.10) is
simply assumed to be 10 percent per annum. The utility discount factor
(ß=0.95) is chosen to yield a return to capital of 6.5 percent per annum,
which is the average real return to equity from 1948 to 1981. Finally, the
utility parameter ηL is chosen indirectly by specifying that steady state hours
work is 0.20 which is based on the average fraction of hours devoted to
market work during the 1948–85 period.
ACKNOWLEDGEMENTS
The author has benefited from the comments and suggestions of Marianne
Baxter, Fischer Black, Karl Brunner, Thomas Cooley, Robert King, Sergio
Rebelo, Carl Shapiro, Joseph Stiglitz and Timothy Taylor. The Bradley
Policy Research Center at the W.E.Simon Graduate School of Business
Administration provided financial support.
NOTES
1 This interpretation of Friedman’s discussion follows Lucas (1977).
420 Charles I.Plosser
2 This view is explicit in the research program initiated by Long and Plosser (1983:68).
3 Lucas (1980) presents an elegant and clear statement of the importance of our
analytical tools in improving our understanding of economic phenomena.
4 Some readers will notice that I have substituted the phrase ‘economic fluctuations’
for ‘business cycle’. I will use these terms interchangeably. My own preference is to
use the term ‘fluctuations’ since ‘business cycle’ frequently carries the connotation
that there is true periodicity present in economic activity. Virtually all of modern
macroeconomics dismisses the view that there are actual periodic cycles in economic
activity. Instead it follows the important work of Slutsky (1937) and interprets the
ups and downs in economic activity as the accumulation of random events or a
stochastic process.
5 While the growth theory literature of the 1960s is replete with discussions of
dynamic behavior of the models studied, little effort was made to relate this behavior
to the characteristics of economies associated with the business cycle. For example,
labor supply did not play a particularly important role in the growth theory
literature yet it is central to any theory attempting to address the phenomenon of
business cycles.
6 See King et al. (1988a) or Kydland and Prescott (1982) for further discussion and
examples of different methods. The Appendix summarizes the approach followed
by the former authors.
7 The economic intuition of how Robinson Crusoe responds to productivity shifts
is also discussed in Long and Plosser (1983) and can now be found in intermediate
macroeconomic textbooks (e.g. Barro 1987).
8 It is sometimes suggested that evidence of important shifts in ‘aggregate demand’ is
prima facie evidence against real business cycle models. As will be argued, this is
incorrect and takes an extraordinarily narrow view of this class of models.
9 Abel and Blanchard (1983) illustrate that under certain conditions that government
spending shocks can be modelled as negative technology shocks. This further
illustrates the potential difficulties of labelling technology shocks as supply or
demand.
10 Long and Plosser (1983) provide an example. Unfortunately, their example
possesses some special features that limit its usefulness for business cycle research.
In particular, they require 100 percent depreciation to obtain the analytical solution.
This results in hours worked being invariant to variations in productivity. As
suggested by Long and Plosser and demonstrated by King et al. (1988a), this
result does not hold when the assumption of 100 percent depreciation is relaxed.
11 Increases in computing power are making it possible to move beyond certainty
equivalence methods and linear decision rules by computing the equilibrium
numerically. For an example see Greenwood et al. (1988).
12 Solow suggested a simple way of measuring technological change. Consider any
constant returns to scale production function with neutral technological change
such as given by Y
t
=Θ
t
F(K
t
, N
t
), where Y
t
is output at time t, K is the capital input,
N is the labor input and Θ
t
measures productivity shifts over time. Solow shows
that if labor is paid its marginal product then the percentage change in productivity
or technology can be computed as ∆θ
t
=∆y
t
-∆k
t
–ω
l
(∆n
t
-∆k
t
) where lower case letters
denote logarithms, ∆ denotes first differencing (i.e. ∆θ
t
=θ
t
-θ
t-1
) and ω
l
is the relative
share of the total output going to labor (i.e. ω
l
=wN/Y where w is the real wage rate).
Thus, using observable data on y, k, n and an estimate of ω
l
estimates of technical
change can be computed.
Understanding real business cycles 421
13 There is a large literature on this issue. In addition to Nelson and Plosser (1982),
see Nelson and Kang (1981), Campbell and Mankiw (1987) and Stock and Watson
(1988).
14 See, for example, Uzawa (1961), Swan (1963) or Phelps (1966).
15 The actual functional forms and parameter values employed in this exercise are
given in the Appendix.
16 All data are taken from the CITIBASE data service except the capital stock, which
is taken from the August 1986 issue of Survey of Current Business. An estimate of
labor’s share of output is also required (see note 12).
17 The responses are computed using the stochastic version of the model that assumes
productivity shifts are known to follow a random walk. Future value of the shifts
are not known to the agents in the economy but they form rational expectations of
these shifts based on their known stochastic structure.
18 In this discussion government purchases are assumed to be financed by lumpsum
taxes or reductions in transfer payments. In this case increase in government
purchases can be viewed as negative shocks to production that enter additively. See
Abel and Blanchard (1983).
19 Another quantitative example can be found in Wynne (1988), who uses a real
business cycle model that includes government purchases to account for the behavior
of the US economy during World War II.
20 Baxter (1988) presents a quantitative analysis of a two sector model in the context
of an international real trade model.
21 See, for example, Altug (1985) and Christiano (1988).
22 The traditional econometric approach and calibration are not mutually exclusive,
however. Singleton (1988) discusses how the calibration approach of Kydland
and Prescott might be formulated in the context of the generalized method of
moments procedure proposed by Hansen (1982).
23 The use of an infinitely lived agent can also be interpreted as an finite-lived agent
with an operative bequest motive that links the current generation’s utility with
future generations’. See Barro (1974) or Miller and Upton (1974).
24 If these disturbances are known, the equilibrium prices and quantities are a perfect
foresight equilibrium. If {Θ
s
} is a stochastic process, Robinson Crusoe forms
expectations about the future values using all currently available information. In
this case the equilibrium is a rational expectations equilibrium.
25 This is but one of the market structures that would support the optimal allocations
as a competitive equilibrium. An alternative market structure in the labor market
might be that agents are paid a wage rate that corresponds to the annuitized rate
based on the present value of their entire future stream of marginal products.
26 For an alternative strategy, see Kydland and Prescott (1982).
27 Solving this system also requires imposing the transversality condition. See King et
al. (1988a) for more details of this solution technique. Several authors including
Christiano (1989) and Rebelo and Rouwenhorst (1989) have studied the accuracy
of these linear approximations.
28 Generalizing this approach to handle the case of stochastic variation in productivity
is not difficult. The method of certainty equivalence amounts to positing a specific
stochastic structure for the s and substituting their conditional expectations for
the future values.
422 Charles I.Plosser
REFERENCES
Abel, Andrew B. and Olivier J.Blanchard, ‘An Intertemporal Model of Savings and
Investment’, Econometrica May 1983, 51, 675–92.
Altug, S., ‘Gestation Lags and the Business Cycle: An Empirical Investigation’,
unpublished manuscript, University of Minnesota, 1985.
Arrow, Kenneth, ‘The Economic Implications of Learning by Doing’, Review of Economic
Studies June 1962, 29, 155–73.
Ballard, Charles L., John B.Shoven, and John Whalley, ‘General Equilibrium
Computations on the Marginal Welfare Costs of Taxes in the United States’, American
Economic Review March 1985, 75, 128–38.
Barro, Robert J., ‘Are Government Bonds Net Wealth?’, Journal of Political Economy
November-December 1974, 82, 1095–117.
Barro, Robert J., ‘Output Effects of Government Purchases’, Journal of Political
Economy December 1981, 89, 1086–1125.
Barro, Robert J., Macroeconomics, New York: John Wiley, 1987.
Baxter, Marianne, ‘Dynamic Real Trade Model: New Directions for Open Economy
Macroeconomics’, Rochester Center for Economic Research, University of Rochester,
working paper 167, 1988.
Baxter, Marianne, and Robert G.King, ‘Multipliers in Equilibrium Business Cycle
Models’, unpublished manuscript, University of Rochester, 1988.
Black, F., Business Cycles and Equilibrium, New York: Basil Blackwell, 1987.
Brock, William A., ‘A Simple Perfect Foresight Monetary Model’, Journal of Monetary
Economics April 1975, 1, 133–150.
Brock, William A. and Leonard J.Mirman, ‘Optimal Economic Growth and Uncertainty:
The Discounted Case’, Journal of Economic Theory June 1972, 4, 479–513.
Campbell, John, and N.Gregory Mankiw, ‘Are Output Fluctuations Transitory?’,
Quarterly Journal of Economics 1987, 102, 857–80.
Cass, David, ‘Optimum Growth in an Aggregative Model of Capital Accumulation’,
Econometrica July 1965, 32, 223–40.
Cho, F. and Richard D.Rogerson, ‘Family Labor Supply and Aggregate Fluctuations’,
Journal of Monetary Economics 1988, 21, 233–45.
Christiano, L., ‘Why Does Inventory Investment Fluctuate So Much?’, Journal of
Monetary Economics 1988, 21, 247–80.
Christiano, L., ‘Solving a Particular Growth Model by Linear Approximation and by
Value Function Iteration’, Federal Reserve Bank of Minnesota, Discussion Paper 9,
1989.
Cooley, Thomas F. and G.Hansen, ‘The Inflation Tax and the Business Cycle’, Rochester
Center for Economic Research, University of Rochester, working paper 155, 1988.
Debreu, Gerard, ‘Valuation Equilibrium and Pareto Optimum’, Proceedings of the
National Academy of Sciences of the U.S., 1954, 38, 886–93.
Eichenbaum, Martin, and Kenneth J.Singleton, ‘Do Equilibrium Business Cycle Theories
Explain Post-war Business Cycles?’, in S.Fischer (ed.) NBER Macroeconomics Annual,
1986, 91–134.
Friedman, Milton, ‘The Role of Monetary Policy’, American Economic Review March
1968, 58, 1–17.
Greenwood, J., Z.Hercowitz, and Gregory W.Huffman, ‘Investment, Capacity Utilization
and the Real Business Cycle’, American Economic Review June 1988, 78, 402–17.
Hall, Robert E., ‘The Dynamic Effects of Fiscal Policy in an Economy with Foresight’,
Review of Economic Studies April 1971, 38, 229–44.
Hall, Robert E., ‘Labor Supply and Aggregate Fluctuations’, in Carnegie-Rochester
Series on Public Policy, 12, Amsterdam: North Holland, 1980.
Hansen, G., ‘Indivisible Labor and the Business Cycle’, Journal of Monetary Economics
1985, 16, 309–27.
Understanding real business cycles 423
Hansen, Lars, ‘Large Sample Properties of Generalized Method of Moment Estimators’,
Econometrica July 1982, 50, 1029–54.
Hicks, John, ‘Equilibrium and the Cycle’, reprinted in John Hicks, Collected Essay on
Economic Theory, Vol. II, Money, Interest and Wages , Oxford: Basil Blackwell,
1933.
Hicks, John, Value and Capital, London: Oxford University Press, 1939.
Hicks, John, Capital and Growth, New York: Oxford University Press, 1965.
Judd, Kenneth, ‘Short Run Analysis of Fiscal Policy in a Simple Perfect Foresight
Model’, Journal of Political Economy April 1985, 93, 298–319.
King, Robert G. and Charles Plosser, ‘Money, Credit and Prices in a Real Business
Cycle,’ American Economic Review June 1984, 74, 363–380.
King, Robert G. and Sergio Rebelo, ‘Business Cycles with Endogenous Growth’,
unpublished manuscript, University of Rochester, 1986.
King, Robert G. and Sergio Rebelo, ‘Public Policy and Economic Growth’, unpublished
working paper, University of Rochester, 1988.
King, Robert G., Charles Plosser, James Stock, and Mark Watson, ‘Stochastic Trends
and Economic Fluctuations’, Rochester Center for Economic Research, working
paper 79, 1987.
King, Robert G., Charles Plosser, and Sergio Rebelo, ‘Production, Growth and Business
Cycles I. The Basic Neoclassical Model’, Journal of Monetary Economics 1988a, 21,
195–232.
King, Robert G., Charles Plosser, and Sergio Rebelo, ‘Production, Growth and Business
Cycles II. New Directions’, Journal of Monetary Economics 1988b, 21, 309–341.
Koopmans, T., ‘On the Concept of Optimal Growth,’ in The Econometric Approach to
Development Planning, Chicago: Rand McNally, 1965.
Kydland, Finn E., ‘Labor Force Heterogeneity and the Business Cycle’, in Carnegie-
Rochester Conference Series on Public Policy 1984, 21, 173–208.
Kydland, Finn E., ‘The Role of Money in a Real Business Cycle Model,’ unpublished
manuscript, Carnegie-Mellon University, 1987.
Kydland, Finn E., and Edward C.Prescott, ‘Time to Build and Aggregate Fluctuations’,
Econometrica November 1982, 50, 1345–70.
Long, John B. and Charles Plosser, ‘Real Business Cycles’, Journal of Political Economy
February 1983, 91, 1345–1370.
Long, John B. and Charles Plosser, ‘Sectoral vs. Aggregate Shocks in the Business
Cycle’, American Economic Review 1987, 77, 333–337.
Lucas, Robert E., ‘Econometric Policy Evaluation: A Critique’, in Carnegie-Rochester
Conference Series on Public Policy 1976, 1, 19–46.
Lucas, Robert E., ‘Understanding Business Cycles’, in Carnegie-Rochester Series on
Public Policy 1977, 5, 7–29.
Lucas, Robert E., ‘Methods and Problems in Business Cycle Theory’, Journal of Money,
Credit and Banking November 1980, 12, 696–715.
Lucas, Robert E., ‘On the Mechanics of Economic Development’, Journal of Monetary
Economics 1988, 22, 3–42.
Miller, Merton H. and Charles W.Upton, Macroeconomics: A Neoclassical Introduction,
Homewood, IL: Irwin, 1974.
Mitchell, W., Business Cycles: The Problem and its Setting, New York: National Bureau
of Economic Research, 1927.
Muth, Richard F., ‘Rational Expectations and the Theory of Price Movements’,
Econometrica 1961, 29, 315–35.
Nelson, Charles R. and H.Kang, ‘Spurious Periodicity in Inappropriately Detrended
Time Series’, Econometrica 1981, 49, 741–51.
Nelson, Charles R. and Charles Plosser, ‘Trends and Random Walks in Macroeconomic
Time Series: Some Evidence and Implications’, Journal of Monetary Economics
September 1982, 10, 139–67.
424 Charles I.Plosser
Phelps, Edmund S., Golden Rules of Economic Growth, New York: Norton, 1966.
Prescott, Edward C., ‘Theory Ahead of Business Cycle Measurement’, in Carnegie-
Rochester Series on Public Policy 1986, 25, 11–66.
Prescott, Edward C. and Robert E.Lucas, ‘A Note on Price Systems in Infinite
Dimensional Space’, International Economics Review June 1972, 13, 416–22.
Rebelo, Sergio, ‘Tractable Heterogeneity and Near Steady State Dynamics’, unpublished
manuscript, University of Rochester, 1987.
Rebelo, Sergio, and G.Rouwenhorst, ‘Linear Quadratic Approximations Versus Discrete
State Space Methods: A Numerical Evaluation’, unpublished working paper,
University of Rochester, 1989.
Rogerson, Richard D., ‘Indivisible Labor, Lotteries and Equilibrium’, Journal of
Monetary Economics 1988, 21, 3–16.
Romer, Paul, ‘Dynamic Competitive Equilibria with Externalities, Increasing Returns
and Unbounded Growth’, doctoral dissertation, University of Chicago, 1983.
Romer, Paul, ‘Increasing Returns and Long Run Growth’, Journal of Political Economy
1986, 94, 1002–73.
Romer, Paul, ‘Increasing Returns, Specialization, and External Economies: Growth as
described by Allyn Young’, Rochester Center for Economic Research, working paper
19, 1987.
Singleton, Kenneth J., ‘Econometric Issues in the Analysis of Equilibrium Business
Cycle Models’, Journal of Monetary Economics 1988, 21, 361–86.
Slutsky, E., ‘The Summation of Random Causes as the Source of Cyclical Processes’,
Econometrica 1937, 5, 105–46.
Solow, Robert, ‘A Contribution to the Theory of Economic Growth’, Quarterly Journal
of Economics February 1956, 70, 65–94.
Solow, Robert, ‘Technical Change and the Aggregate Production Function’, Review of
Economic Studies August 1957, 39, 312–20.
Stock, James H. and Mark H.Watson, ‘Variable Trends in Economic Time Series’,
Journal of Economic Perspectives Summer 1988, 2, 147–74.
Survey of Current Business, United States Department of Commerce/Bureau of Economic
Analysis, August 1986.
Swan, T., ‘On Golden Ages and Production Functions’, in D.Berril (ed.), Economic
Development with Special References to Southeast Asia, London: Macmillan, 1963.
Uzawa, Hirofumi, ‘Neutral Inventions and the Stability of Growth Equilibrium’, Review
of Economic Studies February 1961, 28, 118–24.
Uzawa, Hirofumi, ‘Optimal Technical Change in an Aggregative Model of Economic
Growth’, International Economic Review January 1965, 6, 18–31.
von Hayek, Friederich August, Prices and Production, London: George Routledge &
Sons, 1932.
Wynne, Mark A., ‘The Aggregate Effects of Temporary Government Purchases’,
unpublished manuscript, University of Rochester, 1988.
18 Real business cycles
A new Keynesian perspective
N.Gregory Mankiw
Journal of Economic Perspectives (1989) 3, Summer, pp. 79–90
The debate over the source and propagation of economic fluctuations rages
as fiercely in the late 1980s as it did in the late 1930s in the aftermath of
Keynes’s The General Theory and in the midst of the Great Depression.
Today, as then, there are two schools of thought. The classical school
emphasizes the optimization of private economic actors, the adjustment of
relative prices to equate supply and demand, and the efficiency of unfettered
markets. The Keynesian school believes that understanding economic
fluctuations requires not just studying the intricacies of general equilibrium,
but also appreciating the possibility of market failure on a grand scale.
Real business cycle theory is the latest incarnation of the classical view of
economic fluctuations. It assumes that there are large random fluctuations in
the rate of technological change. In response to these fluctuations,
individuals rationally alter their levels of labor supply and consumption. The
business cycle is, according to this theory, the natural and efficient response
of the economy to changes in the available production technology.
My goal in this chapter is to appraise this newly revived approach to the
business cycle. I should admit in advance that I am not an advocate. In my
view, real business cycle theory does not provide an empirically plausible
explanation of economic fluctuations. Both its reliance on large
technological disturbances as the primary source of economic fluctuations
and its reliance on the intertemporal substitution of leisure to explain
changes in employment are fundamental weaknesses. Moreover, to the extent
that it trivializes the social cost of observed fluctuations, real business cycle
theory is potentially dangerous. The danger is that those who advise policy-
makers might attempt to use it to evaluate the effects of alternative
macroeconomic policies or to conclude that macroeconomic policies are
unnecessary.
WALRASIAN EQUILIBRIUM AND THE CLASSICAL DICHOTOMY
The typical undergraduate course in microeconomics begins with partial
equilibrium analysis of individual markets. A market for a good is
characterized by a downward sloping demand curve and an upward sloping
426 N.Gregory Mankiw
supply curve. The price of the good is assumed to adjust until the quantity
supplied equals the quantity demanded.
The course then builds up to Walrasian general equilibrium. In this
Walrasian equilibrium, prices adjust to equate supply and demand in every
market simultaneously. The general equilibrium system determines the
quantities of all goods and services sold and their relative prices. The most
important theoretical result, after the existence of such a Walrasian
equilibrium, is the ‘invisible hand’ theorem: the equilibrium is Pareto
efficient.
Courses in microeconomics thus show how employment, production, and
relative prices are determined without any mention of the existence of
money, the medium of exchange. The simplest way to append money to the
model is to specify a money demand function and an exogenous money
supply. Money demand depends on the level of output and the price level.
The level of output is already determined in the Walrasian system. The price
level, however, can adjust to equate supply and demand in the money
market.
Introducing money in this way leads to the classical dichotomy (Patinkin
1956). Real variables, such as employment, output, and relative prices,
including the real interest rate, are determined by the Walrasian system.
Nominal variables, such as the price level, the nominal wage, and the
nominal interest rate, are then determined by the equilibrium in the money
market. Of course, since nominal variables do not affect real variables, the
money market is not very important. This classical view of the economy
suggests that, for most policy discussions, the money market can be ignored.
The professor of macroeconomics must in some way deal with the
classical dichotomy. Given the assumptions of Walrasian equilibrium, money
is largely irrelevant. The macroeconomist must either destroy this classical
dichotomy or learn to live with it.
Keynesian macroeconomics destroys the classical dichotomy by
abandoning the assumption that wages and prices adjust instantly to clear
markets. This approach is motivated by the observation that many nominal
wages are fixed by long-term labor contracts and many product prices
remain unchanged for long periods of time. Once the inflexibility of wages
and prices is admitted into a macroeconomic model, the classical dichotomy
and the irrelevance of money quickly disappear.
Much of the early work in the new classical revolution of the 1970s
attempted to destroy the classical dichotomy without abandoning the
fundamental axiom of continuous market clearing (Lucas 1972; 1973). These
models were based on the assumption that individuals have imperfect
information regarding prices. These individuals therefore confuse movements
in the overall price level (which under the classical dichotomy should not
matter) with movements in relative prices (which should matter). An
unanticipated decrease in the money supply leads individuals to infer that the
relative prices of the goods they produce are temporarily low, which induces
Real business cycles 427
them to reduce the quantity supplied. While the fascination with this sort of
story was substantial in the 1970s, it has attracted relatively few adherents in
the 1980s. It is hard to believe that confusion about the price level is
sufficiently great to generate the large changes in quantities observed over
the business cycle.
In contrast to both the Keynesian and the early new classical approaches
to the business cycle, real business cycle theory embraces the classical
dichotomy. It accepts the complete irrelevance of monetary policy, thereby
denying a tenet accepted by almost all macroeconomists in the late 1970s.
Nominal variables, such as the money supply and the price level, are
assumed to have no role in explaining fluctuations in real variables, such as
output and employment.
Real business cycle theory thus pushes the Walrasian model farther than it
has been pushed before. In evaluating whether it provides a successful
explanation of recessions and booms, two questions naturally arise. First,
why are there such large fluctuations in output and employment? Second,
why do movements in nominal variables, such as the money supply, appear
related to movements in real variables, such as output?
CLASSICAL AND KEYNESIAN VIEWS OF ECONOMIC
FLUCTUATIONS
The only forces that can cause economic fluctuations, according to real
business cycle theory, are those forces that change the Walrasian
equilibrium. The Walrasian equilibrium is simply the set of quantities and
relative prices that simultaneously equate supply and demand in all markets
in the economy. To understand how real business cycle theory explains the
business cycle, it is necessary to look into the fundamental forces that change
the supplies and demands for various goods and services.
Many sorts of macroeconomic disturbances can in principle generate
fluctuations in real business cycle models. For example, changes in the level
of government purchases or in the investment tax credit alter the demand for
goods and therefore affect the Walrasian equilibrium. Changes in the relative
price of oil alter the equilibrium allocation of labor among alternative uses.
Many of the macroeconomic disturbances that receive much attention among
Keynesian macroeconomists will also have important effects in real business
cycle models. There is, however, substantial disagreement between the two
schools regarding the mechanisms through which these disturbances work.
Consider the case of a temporary increase in government purchases.
Almost all macroeconomists agree that such a change causes an increase in
output and employment, and the evidence, mainly from wartime experience,
supports this prediction. Yet the explanations of this effect of government
purchases differ greatly.
428 N.Gregory Mankiw
Real business cycle theory emphasizes the intertemporal substitution of
goods and leisure (Barro 1987). It begins by pointing out that an increase in
government purchases increases the demand for goods. To achieve
equilibrium in the goods market, the real interest rate must rise, which
reduces consumption and investment. The increase in the real interest rate
also causes individuals to reallocate leisure across time. In particular, at a
higher real interest rate, working today becomes relatively more attractive
than working in the future; today’s labor supply therefore increases. This
increase in labor supply causes equilibrium employment and output to rise.
While Keynesian theory also predicts an increase in the real interest rate
in response to a temporary increase in government purchases, the effect of
the real interest rate on labor supply does not play a crucial role. Instead, the
increase in employment and output is due to a reduction in the amount of
labor unemployed or underutilized. In most Keynesian theory, the labor
market is characterized as often in a state of excess supply. In contrast, the
Walrasian approach of real business cycle theory does not allow for the
possibility of involuntary unemployment.
Both real business cycle theory and Keynesian theory thus conclude that
increases in government purchases increase output and employment. This
example shows that some of the prominent implications of Keynesian models
also come out of intertemporal Walrasian models. Macroeconomists face a
problem of approximate observational equivalence: many observed
phenomena are consistent with both the classical and Keynesian paradigms.
THE CENTRAL ROLE OF TECHNOLOGICAL DISTURBANCES
While many sorts of macroeconomic disturbances can in principle cause
economic fluctuations in real business cycle models, most attention has
focused on technological disturbances. The reason is that other sorts of
disturbances are unlikely to generate fluctuations in real business cycle
models that resemble actual economic fluctuations.
An obvious but important fact is that over the typical business cycle,
consumption and leisure move in opposite directions. When the economy goes
into a recession, consumption falls and leisure rises. When the economy goes
into a boom, consumption rises and leisure falls. Explaining this phenomenon
is potentially problematic for real business cycle theory: consumption and
leisure would often be expected to move together, since both are normal goods.
In the example of a temporary increase in government purchases, both
consumption and leisure should fall. Many other changes in the demand for
goods, such as a change due to a temporary investment tax credit, also should
cause consumption and leisure to move together.
Real business cycle theory must explain why individuals in a recession
find it rational to increase the quantity of leisure they demand at the same
time they decrease the quantity of goods they demand. The answer must be
Real business cycles 429
that the price of leisure relative to goods, the real wage, falls in a recession.
Hence, a crucial implication of real business cycle theory is that the real
wage is procyclical.1
If the production function were unchanging and demand shocks were the
source of fluctuations, real business cycle theory would have trouble
generating a procyclical real wage. Since labor input is low in a recession,
one would expect that the marginal product of labor and thus the real wage
should be high. With an unchanging production function, diminishing
marginal returns to labor would produce a countercyclical real wage, not the
procyclical real wage necessary to explain the fluctuations in consumption
and leisure.
Real business cycle theorists therefore assume that there are substantial
fluctuations in the rate of technological change. In a recession, the available
production technology is relatively unfavorable. The marginal product of
labor and thus the real wage are low. In response to the low return to
working, individuals reduce consumption and increase leisure.
Since real business cycle theory describes economic fluctuations as a
changing Walrasian equilibrium, it implies that these fluctuations are
efficient. Given the tastes of individuals and the technological possibilities
facing society, the levels of employment, output, and consumption cannot be
improved. Attempts by the government to alter the allocations of the private
market, such as policies to stabilize employment, at best are ineffective and
at worst can do harm by impeding the ‘invisible hand’.
Of all the implications of real business cycle theory, the optimality of
economic fluctuations is perhaps the most shocking. It seems undeniable that
the level of welfare is lower in a recession than in the boom that preceded it.
Keynesian theory explains the reduction in welfare by a failure in economic
coordination: because wages and prices do not adjust instantaneously to
equate supply and demand in all markets, some gains from trade go
unrealized in a recession. In contrast, real business cycle theory allows no
unrealized gains from trade. The reason welfare is lower in a recession is,
according to these theories, that the technological capabilities of society have
declined.
THE EVIDENCE ON TECHNOLOGICAL DISTURBANCES
Advocates of real business cycle theories have trouble convincing skeptics
that the economy is subject to such large and sudden changes in technology.
It is a more standard presumption that the accumulation of knowledge and
the concurrent increase in the economy’s technological opportunities take
place gradually over time. Yet to mimic observed fluctuations, real business
cycle theorists must maintain that there are substantial short-run fluctuations
in the production function.
Edward Prescott (1986) has offered some direct evidence on the
importance of technological disturbances. He examines changes in total
430 N.Gregory Mankiw
factor productivity for the United States economy—the percent change in
output less the percent change in inputs, where the different inputs are
weighted by their factor shares. This ‘Solow residual’ should measure the
rate of technological progress. Prescott points out that there are substantial
fluctuations in the Solow residual, a finding which suggests a potentially
important role for technological disturbances as a source of business cycle
fluctuations.
Figure 18.1 presents my calculation of the Solow residual and the percent
change in output yearly since 1948. (Both variables are for the private
economy less agriculture and housing services.) Like Prescott, I find
substantial fluctuations in measured total factor productivity. For example,
in 1982 total factor productivity fell by 3.5 percent, while in 1984 it rose by
3.4 percent. One might interpret these numbers as showing that the
economy’s ability to convert inputs into outputs—the aggregate production
function—varies substantially from year to year.
Figure 18.1 also shows that measured productivity is highly cyclical. In
every year in which output fell, total factor productivity also fell. If the
Solow residual is a valid measure of the change in the available production
technology, then recessions are periods of technological regress.
The Solow residual need not be interpreted as evidence regarding
exogenous technological disturbances, however. The standard explanation of
cyclical productivity is that it reflects labor hoarding and other ‘off the
production function’ behavior. Productivity appears to fall in a recession
because firms keep unnecessary and underutilized labor. In a boom the
hoarded laborers begin to put out greater effort; output increases without a
large increase in measured labor input.2
Figure 18.1 Solow residuals and output growth
Real business cycles 431
An examination of the data from the early 1940s appears to support this
standard explanation of the cyclical behavior of productivity. The increase in
output associated with the World War II build-up is most plausibly a
demand-driven phenomenon. Yet from 1939 to 1944 measured total factor
productivity grew an average of 7.6 percent per year. (By contrast, the most
productivity has grown in any year since then is 5.2 percent in 1950.) One
might interpret this finding as showing that the economic boom of the 1940s
was in fact driven by supply shocks rather than demand shocks. A more
appealing interpretation is that the Solow residual is not a good measure
over short horizons of changes in the economy’s technological abilities.
Once the Solow residual is rejected as a measure of year-to-year changes
in the available production technology, there is no longer any direct evidence
for substantial technological disturbances. Yet to generate fluctuations that
mimic observed fluctuations, real business cycle models require such
disturbances. The existence of large fluctuations in the available technology
is a crucial but unjustified assumption of real business cycle theory.
An advocate of real business cycle theory might respond that economic
models often rely on assumptions for which there is no formal evidence. Yet
more casual evidence also does not give plausibility to the assumption of
substantial technological disturbances. Recessions are important events; they
receive widespread attention from policy-makers and the media. There is,
however, no discussion of declines in the available technology. If society
suffered some important adverse technological shock, we would be aware of
it. My own reading of the newspaper, however, does not lead me to associate
most recessions with some exogenous deterioration in the economy’s
productive capabilities.
The OPEC energy price changes of the 1970s illustrate that when the
economy experiences large real shocks, these shocks are easily identifiable
and much discussed. Figure 18.1 indeed shows that the economy experienced
large negative Solow residuals in 1974 and 1979, as one might have
expected.3 Yet the five other recessions in the postwar period also exhibit
large negative Solow residuals. To explain these Solow residuals as adverse
changes in the aggregate production function, one would need to find events
with the economic significance of the OPEC price increases. The apparent
absence of such events is evidence that these recessions cannot be easily
attributed to exogenous real shocks.4
LABOR SUPPLY AND INTERTEMPORAL SUBSTITUTION
Real business cycle theorists assume that fluctuations in employment are
fully voluntary. In other words, they assume the economy always finds itself
on the labor supply curve. Yet over the typical business cycle, employment
varies substantially while the determinants of labor supply—the real wage
and the real interest rate—vary only slightly. To mimic this observed
pattern, real business cycle models require that individuals be very willing to
432 N.Gregory Mankiw
reallocate leisure over time. Individuals must significantly reduce the
quantity of labor they supply in response to small temporary reductions in
the real wage or in response to small decreases in the real interest rate.
It is unlikely, however, that individuals are so responsive to intertemporal
relative prices. Econometric evidence on labor supply typically finds that the
willingness of individuals to substitute leisure over time is slight. If leisure
were highly intertemporally substitutable, as real business cycle theorists
assume, then individuals facing expected increases in their real wage should
work little today and much in the future. Individuals facing expected
decreases in their real wage should work hard today and enjoy leisure in the
future. Yet studies of individual labor supply over time find that expected
changes in the real wage lead to only small changes in hours worked (Altonji
1986; Ball 1985). Individuals do not respond to expected real wage changes
by substantially reallocating leisure over time.
Personal experience and introspection provide another way to judge the
behavioral responses on which real business cycle models rely. One key
behavioral response is that quantity of labor supplied reacts substantially to the
real interest rate. Without such intertemporal substitution, real business cycle
models are unable to explain how a temporary increase in government
purchases increases output and employment. Yet such a behavioral response
does not seem plausible. The real interest rate is simply not a significant
consideration when individuals decide to leave their jobs or to accept new
employment. While economists can easily convince non-experts and students
that the quantity of apples demanded depends on the price of apples, it is much
harder to convince them that labor supply depends on the real interest rate. The
implication I draw from this observation is that the intertemporal substitutability
of leisure is very likely far too weak to get real business cycle models to work.
REAL BUSINESS CYCLE THEORIES WITH MULTIPLE SECTORS
The real business cycle theories I have been discussing so far treat production
as if it takes place in a single industry. This abstraction, however, is not
characteristic of all real business cycle theories.
Some real business cycle theories emphasize changes in the technologies of
different sectors, rather than economy-wide changes in technology (Long and
Plosser 1983). These models highlight the interactions among the sectors. Even
if the shocks to the different sectors are independent, the outputs of the different
sectors move together. For example, an adverse shock to one sector reduces the
wealth of the individuals in the economy; these individuals respond by
reducing their demand for all goods. An observer would see an aggregate
business cycle, even without a single aggregate shock.
To get these real business cycle models to work, however, the number of
independent sectoral shocks cannot be too great. If there were many inde-
pendent sectoral shocks and labor were mobile between sectors, then the law
of large numbers would guarantee that these shocks and their effect on the
Real business cycles 433
aggregate economy would average out to zero. To get an aggregate business
cycle, these models therefore require that there be only a few sectors and that
these sectors be subject to large technological disturbances. These models are
therefore similar to the single-sector theories and suffer from the same
weaknesses: the absence of any direct evidence for such large technological
disturbances and the implausibility of strong intertemporal substitutability of
leisure.
A second type of sectoral shock theory emphasizes the costly adjustment
of labor among sectors (Lilien 1982). These models, which depart more from
the Walrasian paradigm, assume that when a worker moves from one sector
to another, a period of unemployment is required, perhaps for job search. In
this case, independent shocks across many sectors do not offset each other.
Recessions are, according to these theories, periods of more sectoral shocks
and thus greater intersectoral adjustment.
This type of real business cycle theory may appear more plausible than
those relying on substantial aggregate productivity shocks and intertemporal
substitution. It is perhaps easier to imagine that recessions are characterized by
an unusually great need for intersectoral reallocation than by some sort of
major technological regress that makes leisure unusually attractive. Yet the
available evidence appears not to support this intersectoral story. If workers
were unemployed voluntarily in recessions because they were moving to new
jobs in other sectors, we would expect to find high unemployment coinciding
with high job vacancy. Yet observed fluctuations have just the opposite
pattern: high unemployment rates coincide with low levels of help wanted
advertising (Abraham and Katz 1986). Moreover, in contrast to the prediction
of this theory, the measured mobility of workers between sectors is strongly
procyclical (Murphy and Topel 1987). This real business cycle theory is also
unable to be plausibly reconciled with observed economic fluctuations.
MONEY AND PRICES OVER THE BUSINESS CYCLE
Before real business cycle theory entered the macroeconomic debate in the
early 1980s, almost all macroeconomists seemed to agree on one conclusion:
money matters. Both historical discussions of business cycles (Friedman and
Schwartz 1963) and more formal econometric work (Barro 1977) pointed to
the Federal Reserve as an important source of macroeconomic disturbances.
While there was controversy as to whether systematic monetary policy could
stabilize the economy, it was universally accepted that bad monetary policy
could be destabilizing.
It is ironic that real business cycle theory arose in the wake of Paul
Volcker’s disinflation. Many economists view this experience as clear
confirmation of the potency of monetary policy. Volcker announced he was
going to slow the rate of money growth to achieve a lower rate of inflation;
the rate of money growth in fact slowed down; and one of the deepest
postwar recessions followed, as did an eventual reduction in the rate of
434 N.Gregory Mankiw
inflation. This set of events is easy to explain within the context of Keynesian
theory with its emphasis on the gradual adjustment of wages and prices. It is
less easy to explain within the context of real business cycle theory.5
Robert King and Charles Plosser (1984) explain the historical association
between money and output by arguing that the money supply endogenously
responds to fluctuations in output. Standard measures of the money supply
such as Ml are mostly inside money, that is, money created by the banking
system. King and Plosser suggest that the transactions services of inside
money should be viewed as simply the ‘output’ of one sector of the economy,
the banking sector. Just as one should expect the outputs of different sectors
to move together within a multi-sector real business cycle model, one should
expect the output of the banking sector to move with the outputs of other
sectors. An increase in productivity in any sector will tend to increase the
demand for transactions services; the banking system responds by creating
more inside money. Hence, the procyclical behavior of standard monetary
aggregates cannot necessarily be interpreted as evidence that changes in
outside money caused by the monetary authority have real effects.
While the story of King and Plosser can explain the procyclical behavior
of money, it cannot explain the procyclical behavior of prices. It is a well-
documented fact that, in the absence of identifiable real shocks such as the
OPEC oil price changes, inflation tends to rise in booms and fall in
recessions. This famous Phillips curve correlation played a central role in the
macroeconomic debate of the 1960s, and it was the primary empirical
motivation for the early new classical theories in the 1970s (Friedman 1968;
Lucas 1972). Yet since the model of King and Plosser generates procyclical
money through the demand for transactions services, these fluctuations in
money will be associated with fluctuations in real balances not with
fluctuations in prices. The short-run Phillips curve has thus been left without
an explanation by real business cycle theorists.6
THE TRADEOFF BETWEEN INTERNAL AND EXTERNAL
CONSISTENCY
A good theory has two characteristics: internal consistency and external
consistency. An internally consistent theory is one that is parsimonious; it
invokes no ad hoc or peculiar axioms. An externally consistent theory is one
that fits the facts; it makes empirically refutable predictions that are not
refuted. All scientists, including economists, strive for theories that are both
internally and externally consistent. Yet like all optimizing agents, scientists
face tradeoffs. One theory may be more ‘beautiful’, while another may be
easier to reconcile with observation.
The choice between alternative theories of the business cycle—in
particular, between real business cycle theory and new Keynesian theory—is
partly a choice between internal and external consistency. Real business
cycle theory extends the Walrasian paradigm, the most widely understood
Real business cycles 435
and taught model in economics, and provides a unified explanation for
economic growth and economic fluctuations. New Keynesian theory, in its
attempt to mimic the world more accurately, relies on nominal rigidities that
are observed but only little understood. Indeed, new Keynesians sometimes
suggest that to understand the business cycle, it may be necessary to reject
the axiom of rational, optimizing individuals, an act that for economists
would be the ultimate abandonment of internal consistency.
The tension between these two goals of science will undoubtedly continue.
Each school of macroeconomic thought will highlight its strengths while
trying to improve on its weaknesses. My own forecast is that real business
cycle advocates will not manage to produce convincing evidence that there
are substantial shocks to technology and that leisure is highly substitutable
over time. Without such evidence, their theories will be judged as not
persuasive. New Keynesians, however, have made substantial progress in
recent years toward providing rigorous microeconomic foundations, the
absence of which was the fatal flaw of the Keynesian consensus of the 1960s.
While real business cycle theory has served the important function of
stimulating and provoking the scientific debate, it will, I predict, ultimately
be discarded as an explanation of observed fluctuations.
ACKNOWLEDGEMENTS
I am grateful to Lawrence Ball, Susanto Basu, Marianne Baxter, Mark Bils,
Lawrence Katz, Deborah Mankiw, David Romer, Joseph Stiglitz, Lawrence
Summers, Timothy Taylor, David Weil, and Michael Woodford for helpful
discussions and comments, and to the National Science Foundation for
financial support.
NOTES
1 Alternatively, one could explain the observed pattern without a procyclical real
wage by positing that tastes for consumption relative to leisure vary over time.
Recessions are then periods of ‘chronic laziness’. As far as I know, no one has
seriously proposed this explanation of the business cycle.
2 A related explanation of the procyclical behavior of the Solow residual has been
proposed by Hall (1987). Hall points out that if price exceeds marginal cost because
of imperfect competition, then the measured Solow residual will appear procyclical
even if the true production technology is unchanging. Alternatively, the Solow
residual could reflect endogenous changes in technology due to demand shocks:
such endogeneity might arise if, for example, learning-by-doing is important.
3 Whether changes in energy prices affect the Solow residual computed from GNP
depends on a variety of issues involving the construction of index numbers like
GNP. See Bruno and Sachs (1985:43) for a discussion.
4 Hamilton (1983) finds oil price changes are also associated with the pre-OPEC
recessions. Yet these price changes are much too small to explain plausibly such
large declines in productivity.
436 N.Gregory Mankiw
5 The disinflation is not unusual. Romer and Romer (1989) show that output
typically falls after the Fed makes an explicit decision to reduce inflation, which
they interpret as evidence against real business cycle theory.
6 Indeed, as King and Plosser (1984) point out, their model makes the counterfactual
prediction that the price level should be countercyclical: since the demand for real
outside money probably rises in a boom, and it is the outside money stock that
pins down the price level, equilibrium in the market for outside money requires
that the price level fall in a boom.
REFERENCES
Abraham, Katharine G. and Lawrence F.Katz, ‘Cyclical Unemployment: Sectoral Shifts
or Aggregate Disturbances?’, Journal of Political Economy June 1986, 94, 507–22.
Altonji, Joseph G. ‘Intertemporal Substitution in Labor Supply: Evidence from Micro
Data’, Journal of Political Economy June 1986, Part 2, 94, S176–S215.
Ball, Laurence, ‘Intertemporal Substitution and Constraints on Labor Supply: Evidence
from Panel Data’, manuscript, MIT, 1985.
Barro, Robert J., ‘Unanticipated Money Growth and Unemployment in the United
States’, American Economic Review 1977, 67, 101–15.
Barro, Robert J., Macroeconomics, New York: Wiley, 1987.
Bruno, Michael, and Jeffrey Sachs, Economics of Worldwide Stagflation, Cambridge,
MA: Harvard University Press, 1985.
Friedman, Milton, ‘The Role of Monetary Policy’, American Economic Review 1968,
58, 1–17.
Friedman, Milton, and Anna Schwartz, A Monetary History of the United States,
Princeton, NJ: Princeton University Press, 1963.
Hall, Robert E., ‘Market Structure and Macroeconomic Fluctuations’, Brookings Papers
on Economic Activity 1987:1, 285–322.
Hamilton, James D., ‘Oil and the Macroeconomy since World War II’, Journal of
Political Economy April 1983, 91, 228–48.
King, Robert G. and Charles I.Plosser, ‘Money, Credit, and Prices in a Real Business
Cycle’, American Economic Review June 1984, 74, 363–80.
Lilien, David M., ‘Sectoral Shifts and Cyclical Unemployment’, Journal of Political
Economy August 1982, 90, 777–793.
Long, John B. Jr and Charles I.Plosser, ‘Real Business Cycles’, Journal of Political
Economy February 1983, 91, 39–69.
Lucas, Robert E. Jr, ‘Expectations and the Neutrality of Money’, Journal of Economic
Theory 1972, 4, 103–124.
Lucas, Robert E. Jr, ‘International Evidence on Output-Inflation Tradeoffs’, American
Economic Review 1973, 63, 326–334.
Murphy, Kevin M. and Robert H.Topel, ‘The Evolution of Unemployment in the United
States: 1968–1985’, NBER Macroeconomics Annual 1987.
Patinkin, Don, Money, Interest, and Prices: An Integration of Monetary and Value
Theory, Evanston, IL: Row, Peterson, 1956.
Prescott, Edward, ‘Theory Ahead of Business Cycle Measurement’, Carnegie-Rochester
Conference on Public Policy Autumn 1986, 25, 11–44.
Romer, Christina, and David Romer, ‘Does Monetary Policy Matter: A New Test in the
Spirit of Friedman and Schwartz’, NBER Macroeconomics Annual, 1989.
Part V
New Keynesian economics
Introduction
The Keynesian propositions of involuntary unemployment, non-neutrality of
money and extensive market failures do not rest easily alongside the
Walrasian theory of general competitive equilibrium characterized by
continuously cleared markets. Although these two strands of economics were
brought together during the era of the neoclassical synthesis, the marriage
was always potentially explosive. By the early 1970s the Keynesian model
was under attack from both monetarists and the newly emerging new
classical theorists, the latter being committed to Walrasian formalism with
respect to their methodology. This crisis in macroeconomic theory could be
reconciled in two ways. Either macroeconomic models could be adapted so as
to be consistent with choice-theoretic neoclassical microeconomic theory
within a general equilibrium framework, or, alternatively, microeconomic
theory could be adapted so as to be consistent with Keynesian propositions.
New classical economists chose the former route, new Keynesian economists
have followed the latter route (see Greenwald and Stiglitz 1987; Snowdon et
al. 1994). Hence new Keynesian economists inhabit a brave new world where
co-ordination failures and macroeconomic externalities result from market
imperfections such as those arising out of imperfect competition, incomplete
markets, heterogeneous labour and asymmetric information (see Romer 1996;
Dixon 1997).
New Keynesians and new classical/real business cycle theorists differ
substantially over the ability and desirability of stabilizing the economy via
the use of monetary and fiscal policy. However, new Keynesians fully
endorse the call, initiated by Lucas, that macroeconomic models should be
based on coherent microfoundations. Beyond this, however, there is little
agreement on what constitutes an acceptable model of how markets work.
Whereas new classical and real business cycle theorists employ the
traditional neoclassical perfectly competitive framework based on a rational
representative agent, new Keynesians recognize that whenever information
is imperfect or markets are imperfectly competitive and incomplete the
economy will fail to be Pareto efficient. Modern Keynesians regard these
real world imperfections as the heart of the problem facing business cycle
theorists. By incorporating the insights provided by modern market theory
440 New Keynesian economics
where consumers, producers and labour market participants make decisions
with incomplete information, new Keynesians argue that they have
overcome, to a large extent, the microeconomic deficiencies of orthodox
Keynesianism. In addition, by disposing of the fictional Walrasian
auctioneer, new Keynesians claim that their models are capable of
incorporating features of real economies where co-ordination problems
abound (see Phelps 1985: chs 15–20 for an accessible introduction to
modern market theory).
Gregory Mankiw, a leading new Keynesian economist, believes that
Robert Lucas’s wide-ranging criticisms of Keynesian models relating to their
inadequate microfoundations have now been met. However, in responding
to the new classical critique, the new generation of Keynesian models differ
substantially from those which formed the basis of the neoclassical synthesis
during the 1960s. In his 1992 European Economic Review article ‘The
Reincarnation of Keynesian Economies’ (reprinted on pp. 445–51), Mankiw
highlights what he considers to be the essential differences between old and
new Keynesian economics by challenging six dubious Keynesian
propositions. In doing so Mankiw illustrates how the traditional analysis of
the classical economists, modified by the contributions of Friedman and
Lucas, has been highly influential in shaping the development of new
Keynesian thinking. In particular Mankiw questions the traditional
Keynesian emphasis on short-run rather than long-run issues, the down-
grading of monetary policy and overemphasis on discretionary policies.
Nevertheless Mankiw argues that the reincarnated Keynesianism now has
‘firm microeconomic muscle’ and is in a far stronger position in explaining
economy-wide market failure than the older Keynesian models associated
with the neoclassical synthesis.
The views of Gregory Mankiw are further highlighted in the article
(reprinted on pp. 452–77) by Brian Snowdon and Howard Vane taken from
the Spring 1995 issue of the American Economist. Following a brief survey
of some of the main features and reasons for the breakdown of the
neoclassical synthesis model, Snowdon and Vane present the text of a
detailed interview conducted with Mankiw at Harvard University in
February 1993. Mankiw’s response to a wide range of questions relating to
the development and current state of macroeconomics demonstrate that
Keynesian economics has been revived. Finally the authors present an
assessment of the current state of macroeconomics where they put forward
the view that there appears to have taken place a new Keynesian-monetarist
synthesis. This now forms the mainstream position in macroeconomics
which remains the target of attack from real business cycle theorists who
continue to follow the classical tradition of rejecting the idea of economy-
wide market failure.
An important problem with new Keynesian economics is that it
encompasses an extremely heterogeneous collection of economists and ideas.
Numerous elegant theories have been developed which are often unrelated.
Introduction 441
In a masterful survey entitled ‘What is New-Keynesian Economics?’, taken
from the September 1990 issue of the Journal of Economic Literature
(reprinted on pp. 478–551), Robert Gordon pulls together the various ideas
which make up the main themes identifiable in the modern Keynesian
literature. In Gordon’s view ‘the task of new Keynesian economics is to
explain why changes in the aggregate price level are sticky’ (see also
Gordon 1981). After surveying the evidence on price inertia Gordon goes on
to critically examine theories of nominal, and real, wage and price rigidity
including those relating to costly price adjustment (menu costs) and
efficiency wage theory. Gordon’s preference is for what he calls the input-
output approach which places the emphasis on co-ordination failure as the
main cause of macroeconomic inefficiency. With thousands of firms each
buying thousands of components, made up of ingredients supplied by
numerous other firms, the typical firm does not know the identity of their
complete set of suppliers because ‘the input-output table is so broad and
deep’. Under such conditions firms are unwilling to index their prices to
nominal aggregate demand shocks because marginal cost and marginal
revenue are imperfectly correlated with aggregate demand. This is due to
the diversity of shocks which can influence a firm’s demand and cost
conditions. Thus Gordon extends Lucas’s two-way distinction between local
and aggregate demand shocks to a four-way classification which also
includes local and aggregate cost shocks. In effect this magnifies the ‘signal
extraction’ problem raised by Lucas. In such a world ‘no individual firm
has an incentive to take the risk posed by nominal GNP indexation’. Each
firm prefers to keep control of its own price—cost relationship ‘because they
will go bankrupt if costs rise sufficiently in relation to price’.
Bruce Greenwald and Joseph Stiglitz show how there are two broad
approaches within new Keynesian economics. The first strand, emphasized
by Mankiw and Romer, concentrates on nominal price rigidities as the
essential way that actual market economies differ from that depicted in the
idealized Walrasian model (see Mankiw 1990; Romer 1993). The second
strand explores the path suggested by Keynes (1936) in Chapter 19 of the
General Theory, namely that increased wage and price flexibility might
exacerbate the economy’s downturn. In their Journal of Economic
Perspectives article (reprinted on pp. 552–74) taken from the Winter 1993
issue, Greenwald and Stiglitz explore this second strand of Keynesian
analysis and put forward a new Keynesian model where monetary policy
has real effects even when wages and prices are flexible. By blending three
ingredients, namely, risk averse firms, credit rationing and new labour
market theories, Greenwald and Stiglitz demonstrate how in a world of
imperfect information, shocks to the macroeconomy have persistent real
effects and how greater price flexibility ‘might exacerbate the problem of
economic fluctuations’ (see also Sheffrin 1989; Tobin 1993—reprinted on
pp. 135–55). In the Greenwald-Stiglitz model, because capital markets are
imperfect, firms cannot divest themselves of risk via ready access to equity
442 New Keynesian economics
finance. By resorting to debt finance, firms increase the probability that they
will face bankruptcy if the economic environment deteriorates. Hence in the
face of declining demand, firms prefer, in the first instance, to cut back
production rather than reduce the price of their product. Their theory of the
risk averse firm thus provides an explanation of why the aggregate supply
curve shifts adversely during a recession. The analysis of Greenwald and
Stiglitz also suggests that the emphasis of macroeconomic research should
shift from the product market to the capital and labour markets (see Stiglitz
1992b).
The work of new Keynesian economists has demonstrated the remarkable
resilience of Keynesian economics in the face of the powerful theoretical
counter-revolutions launched against its central doctrines. By emphasizing
numerous imperfections in labour, product and capital markets new
Keynesian analysis has developed robust microfoundations for macro-models
which reject continuous market clearing. Since the mid-1970s a ‘rich harvest’
of non-Walrasian microeconomics has been incorporated into the analysis of
aggregate instability (see Benassi et al. 1994). Although the new Keynesian
research programme is far from complete, the contribution from this ‘school’
of thought towards the reconstruction of the supply side of macroeconomic
models has been highly influential (see Mankiw and Romer 1991).
REFERENCES
*Titles marked with an asterisk are particularly recommended for additional reading.
*Abel, A.B. and B.S.Bernanke (1995) Macroeconomics, 2nd edn, Chapter 12, New
York: Addison Wesley.
Adnett, N. (1994) ‘New Keynesian Economics: Macroeconomics for the 1990s?’,
Economics and Business Education 2, Autumn, pp. 118–20.
Akerlof, G.A. and J.Yellen (eds) (1986) Efficiency Wage Models and the Labour Market,
Cambridge: Cambridge University Press.
Ball, L., N.G.Mankiw and D.Romer (1988) ‘The New Keynesian Economics and the
Output Inflation Trade-Off’, Brookings Papers on Economic Activity 1, pp. 1–65.
Benassi, C., A.Chirco and C.Colombo (1994) The New Keynesian Economics, Oxford:
Basil Blackwell.
*Colander, D. (1988) ‘The Evolution of Keynesian Economics: From Keynesian to New
Classical to New Keynesian’, in O.F.Hamouda and J.N.Smithin (eds) Keynes and
Public Policy after Fifty Years Vol. 1: Economics and Policy, Aldershot: Edward Elgar.
*Dixon, H.D. (1997) ‘The Role of Imperfect Competition in New Keynesian Economies’,
in B.Snowdon and H.R.Vane (eds) Reflections on the Development of Modern
Macroeconomics, Aldershot: Edward Elgar.
Dore, M.H.I. (1993) The Macrodynamics of Business Cycles: A Comparative Evaluation,
Chapter 7, Oxford: Basil Blackwell.
Dornbusch, R. and S.Fischer (1994) Macroeconomics, 6th edn, Chapter 9, New York:
McGraw-Hill. Fischer, S. (1988) ‘Recent Developments in Macroeconomics’, Economic
Journal 98, June, pp. 294–339.
*Froyen, R.T. (1996) Macroeconomics: Theories and Policies, 5th edn, Chapter 12,
London: Prentice-Hall.
*Gerrard, B. (1996) ‘Competing Schools of Thought in Macroeconomics: An Ever
Emerging Consensus?’ Journal of Economic Studies 23, pp. 53–69.
Introduction 443
*Gordon, R.J. (1981) ‘Output Fluctuations and Gradual Price Adjustment’, Journal of
Economic Literature 19, June, pp. 493–530.
Gordon, R.J. (1982) ‘Price Inertia and Policy Ineffectiveness in the United States, 1890–
1990’, Journal of Political Economy 90, December, pp. 1087–117.
* Gordon, R.J. (1993) Macroeconomics, 6th edn, Chapter 8, New York:
HarperCollins.
*Greenwald, B.C. and J.E.Stiglitz (1987) ‘Keynesian, New Keynesian and New Classical
Economies’, Oxford Economic Papers 39, March, pp. 119–32.
*Hall, R.E. and J.B.Taylor (1993) Macroeconomics, 4th edn, Chapter 16, New York:
W.W.Norton.
Hargreaves Heap, S.P. (1992) The New Keynesian Macroeconomics: Time, Belief and
Social Interdependence, Aldershot: Edward Elgar.
*Jansen, D.W., C.D.Delorme and R.B.Ekelund, Jr (1994) Intermediate Macroeconomics,
Chapter 10, New York: West.
Kirman, A.P. (1992) ‘Whom or What Does the Representative Individual Represent?’,
Journal of Economic Perspectives 6, Spring, pp. 117–29.
Leslie, D. (1993) Advanced Macroeconomics: Beyond IS/LM, Chapter 8, London:
McGraw-Hill.
*Mankiw, N.G. (1990) ‘A Quick Refresher Course in Macroeconomics’, Journal of
Economic Literature 28, December, pp. 1645–60.
*Mankiw, N.G. (1994) Macroeconomics, 2nd edn, Chapter 11, New York: Worth.
Mankiw, N.G. and D.Romer (eds) (1991) New Keynesian Economics, 2 vols, Cambridge:
MIT Press.
Phelps, E.S. (1985) Political Economy: An Introductory Text, New York: W.W.
Norton.
Romer, D. (1993) ‘The New Keynesian Synthesis’, Journal of Economic Perspectives 7,
Winter, pp. 5–22.
Romer, D. (1996) Advanced Macroeconomics, Chapter 6, London: McGraw-Hill.
Rotenberg, J. (1987) ‘The New Keynesian Microfoundations’, National Bureau of
Economic Research Macroeconomics Annual.
Sheffrin, S. (1989) The Making of Economic Policy, Chapter 5, Oxford: Basil Blackwell.
*Snowdon, B. and H.R.Vane (1992) ‘New Classical and New Keynesian
Macroeconomics’, Economics 28, Summer, pp. 54–62.
*Snowdon, B., H.R.Vane and P.Wynarczyk (1994) A Modern Guide To Macroeconomics:
An Introduction to Competing Schools of Thought, Chapter 7, Aldershot: Edward
Elgar.
Stiglitz, J.E. (1992a) ‘Methodological Issues and the New Keynesian Economies’, in
A.Vercelli and N.Dimitri (eds) Macroeconomics: A Survey of Research Strategies,
Oxford: Oxford University Press.
Stiglitz, J.E. (1992b) ‘Capital Markets and Economic Fluctuations in Capitalist
Economies’, European Economic Review 36, pp. 269–306.
Tobin, J. (1993) ‘Price Flexibility and Output Stability: An Old Keynesian View’, Journal
of Economic Perspectives 7, Winter, pp. 45–65.
QUESTIONS
1 To what extent have new Keynesian theorists provided more coherent
microfoundations for Keynesian models and policies?
2 According to the 1996 Nobel citation, the work of Robert E.Lucas Jr
‘transformed macroeconomic analysis and deepened our understanding of
economic policy’. In what ways has the influence of Lucas changed Keynesian
economics?
444 New Keynesian economics
3 What is new in new Keynesian economics?
4 ‘Any satisfactory rehabilitation of Keynesian models requires that the
assumption of nominal price stickiness be supported by sound microeconomic
reasoning’. How do new Keynesian theorists explain nominal price stickiness
and why is this an important issue in macroeconomic analysis?
5 How does the work of new Keynesians differ from that of both Keynes and
orthodox Keynesianism?
6 To what extent can efficiency wage and insider-outsider theories provide a
coherent explanation of involuntary unemployment as an equilibrium
phenomenon?
7 Has the natural rate hypothesis, first put forward by Friedman in 1968, been
undermined by hysteresis models of unemployment?
8 ‘In the Greenwald-Stiglitz (1993) model increased price flexibility is not a
solution to demand induced recessions. Indeed such flexibility could make
the situation worse’. Why is this the case and to what extent does this conform
more closely to the view of Keynes in the General Theory?
9 ‘The real business cycle approach has performed a useful function in raising
profound questions relating to the significance and characteristics of economic
fluctuations’. How successful have new Keynesian economists been in
preserving the view of Keynes that aggregate instability is symptomatic of
significant market failure?
10 Critically examine the contribution to the development of modern
macroeconomics of any one of the following new Keynesian contributions:
(a) Menu costs
(b) Co-ordination failure
(c) Aggregate demand externality
(d) Staggered wage/price setting
(e) Hysteresis
(f) Near rationality.
19 The reincarnation of Keynesian
economics
N.Gregory Mankiw
European Economic Review (1992) 36, April, pp. 559–65
The title of this session, ‘Keynesian Economics Today’, says much about what
has happened in macroeconomics since the early 1980s. When I began
graduate school at MIT in 1980, it was not at all clear that Keynesian
economics would still be around in the 1990s. Keynes looked as if he were
leaving macroeconomics and entering the history of thought. The leading
intellectual figure of the day was Robert Lucas, and he had this to say about
the state of Keynesian economics:
One cannot find good, under-forty economists who identify themselves or
their work as ‘Keynesian’. Indeed, people even take offense if referred to as
‘Keynesians’. At research seminars, people don’t take Keynesian theorizing
seriously anymore; the audience starts to whisper and giggle to one
another.
Lucas called his article ‘The Death of Keynesian Economies’.
From our current perspective, it is clear that this obituary was premature.
Today, Keynesian theorizing does not inspire whispers and giggles from the
audience. There are many economists under the age of forty who do not take
offense when their work is called ‘Keynesian’, and I count myself as one of them.
If Keynesian economics was dead in 1980, then today it has been reincarnated.
David Romer and I have assembled some of the articles that have been
central to this reincarnation in a two-volume collection called New Keynesian
Economics (Mankiw and Romer 1991a, 1991b). The topics covered in this
collection—such as imperfect competition, menu costs, coordination failure,
and efficiency wages—are those that have dominated discussions among
Keynesians since the early 1980s. It is too early to say there is a consensus
about how all these topics fit together. Yet one can say that the new classical
challenge has been met: Keynesian economics has been reincarnated into a
body with firm microeconomic muscle.
I am careful to call this re-emergence of Keynesian economics a
‘reincarnation’ rather than a ‘resurrection’. My dictionary defines
‘reincarnation’ as ‘the rebirth into another body’, and that describes well
Keynesian economics today. The Keynesian economics of the 1990s shares
446 N.Gregory Mankiw
the spirit of the Keynesian economics of earlier decades. Like their
predecessors, new Keynesians question the relevance of the Walrasian
paradigm in explaining economy-wide booms and busts. Old and new
Keynesians share a skepticism in the invisible hand’s ability to maintain full
employment. They both see the business cycle as a type of economy-wide
market failure.
Beyond these broad principles, however, old and new Keynesians differ
substantially. In many ways, the Keynesian economics of the 1990s does not
look like the Keynesian economics of the 1930s, or even that of the 1960s. To
some old Keynesians, new Keynesian economics may be hard to recognize as
Keynesian at all. Indeed, new Keynesian economics may appear more similar
to the classical economics of David Hume, or even to the monetarist
economics of Milton Friedman.
My goal today is to highlight some of the differences between old and new
Keynesian economics. In particular, I would like to discuss six dubious
Keynesian propositions. These are propositions that various economists in the
past have viewed as basic tenets of Keynesian economics and that, I believe,
economists today should discard.
My list of dubious Keynesian propositions is highly personal. I surely
would not claim that all new Keynesians view all of these propositions as
dubious. Yet my list is not entirely idiosyncratic. I present these dubious
Keynesian propositions in part to show the profound impact that
monetarism and new classicism has had on the thinking of my generation of
Keynesians.
DUBIOUS KEYNESIAN PROPOSITION NO. 1
Learning how the economy works is best achieved by a careful reading of
Keynes’s General Theory
Since Keynesian economics is derived, by definition, from the work of John
Maynard Keynes, one might suppose that reading Keynes is an important part
of Keynesian theorizing. In fact, quite the opposite is the case. Few young
economists—Keynesian or otherwise—concern themselves with the question of
what ‘Keynes really meant’. New Keynesians view their work as following in
the broad tradition that evolved from Keynes, but their goal is to explain the
world, not to clarify the views of one particular man. If new Keynesian
economics is not a true representation of Keynes’s views, then so much the
worse for Keynes.
The reason for this attitude is clear. Despite its remarkable contribution,
the General Theory is an obscure book: I am not sure that even Keynes
himself knew completely what he really meant. Moreover, after fifty years of
additional progress in economic science, the General Theory is an out-dated
book. The rigor with which we develop economic theories and the data and
statistical techniques with which we test our theories were unknown half a
The reincarnation of Keynesian economics 447
century ago. We are in a much better position than Keynes was to figure out
how the economy works.
DUBIOUS KEYNESIAN PROPOSITION NO. 2
The lessons of classical economics are not helpful in understanding how the
world works
Perhaps for dramatic effect, or perhaps because he was writing in the midst of
the Great Depression, Keynes minimized the lessons of classical economics.
He called classical economics a ‘special case’ of his general theory, and he
wrote that
the characteristics of the special case assumed by classical theory happen
not to be those of the economic society in which we actually live, with the
result that its teaching is misleading and disastrous if we attempt to apply
it to the facts of experience.
If the General Theory were to have had a subtitle, it might have been ‘The
Death of Classical Economies’.
Today, few macroeconomists take such a dim view of classical economics.
Most accept the natural-rate hypothesis, which interpreted broadly states that
classical economics is right in the long run. Moreover, economists today are
more interested in the long-run equilibrium. The long run is not so far away that
one can cavalierly claim, as Keynes did, that ‘in the long run we’re all dead’.
The widespread acceptance of classical economics is evidenced by the
reemergence of economic growth as an active area of research. The starting
point of all modern growth theory is the Solow growth model. The Solow
growth model is eminently classical: it begins by simply assuming that the
economy reaches full employment. Although there is continuing debate about
whether the Solow model provides an adequate description of economic
growth, the model is rarely criticized as being too classical.
DUBIOUS KEYNESIAN PROPOSITION NO. 3
Capitalist economies are threatened by the possibility of excessive saving,
which could lead to secular stagnation; deficit spending is, therefore, good
for the economy
In 1981 Martin Feldstein wrote an article called ‘The Retreat from Keynesian
Economies’, in which he wrote that
the most direct effect of Keynesian thinking has been to retard the process
of capital formation. Keynes’s own writing displayed not only a lack of
interest in the potential benefits of capital accumulation but also an
outright fear of excessive saving.
448 N.Gregory Mankiw
More recently, Feldstein (1988) has suggested that Keynes was the ‘academic
scribbler’ who unwittingly inspired Ronald Reagan’s decade of budget
deficits.
Once again, we see a stark contrast between old and new Keynesians.
Feldstein is correct that some early Keynesians feared that the economy might
suffer from ‘secular stagnation’ if the propensity to save were too great. By
contrast, few economists today believe that excessive saving threatens the
economy. Instead, almost all economists now believe that additional saving
will, in the long run, lead to additional investment rather than inadequate
aggregate demand. Indeed, rather than being concerned with excessive
saving, most American economists fear that the US saving rate is inadequate
to maintain the United States’ high standard of living.
DUBIOUS KEYNESIAN PROPOSITION NO. 4
Fiscal policy is a powerful tool for economic stabilization, and monetary
policy is not very important
The first course I took in macroeconomics used Samuelson’s famous textbook.
Like many students in my generation, my class was introduced to
macroeconomic modelling with the Keynesian cross. Our first lesson was that
fiscal policy is a powerful tool that policymakers could (and indeed should)
use to control national income. After we learned about the magic of the
multiplier, we did study several ancillary topics, such as the role of monetary
policy, but we always kept the Keynesian cross as the benchmark model of
the economy.
From a modern perspective, it seems most peculiar to begin the study of
macroeconomics with the Keynesian cross. The most striking features of this
model are what it lacks: any connection to microeconomics, any self-
correcting forces returning the economy to the natural rate, and any role for
the central bank. Moreover, as an empirical matter, the message of the
model is more wrong than right. The numerical examples regularly given to
students suggest that the multiplier is indeed quite magical. But, in the
world, fiscal policy is not so potent. For example, the quintessentially
Keynesian DRI model estimates the government-purchases multiplier
(holding the money supply constant) to be only about 0.6 (Eckstein
1983:169).
For the purpose of analyzing economic policy, a student would be better
equipped with the quantity theory of money (together with the expectations-
augmented Phillips curve) than the Keynesian cross. In the United States
today, fiscal policymakers have completely abdicated responsibility for
economic stabilization. Their inability to cope with persistently large
government deficits has left them unable even to imagine trying to reach
consensus on countercyclical fiscal policy in a timely fashion. All attempts at
stabilization are left to monetary policy. When a recession ensues, as it did
The reincarnation of Keynesian economics 449
recently in the United States, fiscal policymakers merely begin discussions
about what the Federal Reserve did wrong.
DUBIOUS KEYNESIAN PROPOSITION NO. 5
Policymakers should learn to live with inflation, because it is the cost of low
unemployment
In his Presidential Address to the American Economic Association in 1971,
James Tobin argued that the ‘zero-inflation unemployment rate’ is not
optimal. What is noteworthy about this fact is not that Tobin reached this
conclusion but that he even chose to ask the question. The very phrase ‘zero-
inflation unemployment rate’ presumes the existence of a long-run tradeoff
between inflation and unemployment. Most economists today doubt that such
a tradeoff exists. On this issue, Milton Friedman (1968) has won the hearts
and minds of my generation: in most new Keynesian models, the long-run
Phillips curve is vertical.
Perhaps surprisingly, the intellectual victory of the natural-rate hypothesis
has not led to consensus among macroeconomists on the relationship between
inflation and unemployment. Instead, the debate has shifted focus to the short-
run relationship. New Keynesians are the keepers of the faith that
policymakers face a short-run tradeoff between inflation and unemployment.
New classicals, who have devoted their energy since the early 1980s to real-
business-cycle theory, deny the existence of any tradeoff over any time
horizon.
Here we can see how misleading the labels have become. Old classical
economists, such as David Hume, asserted that money was neutral in the long
run but not in the short run. This is exactly the position held by new
Keynesians. By contrast, new classical economists claim that money is neutral
even in the short run. In advocating this position, they take the classical
dichotomy more seriously than did the classical economists themselves.
DUBIOUS KEYNESIAN PROPOSITION NO. 6
Policymakers should be free to exercise their discretion in responding to
changing economic conditions and avoid adherence to a rigid policy rule
One striking feature of the economic history since the early 1940s has been
high and persistent rates of inflation. Between 1940 and 1990, US inflation
averaged 4.6 percent per year. By contrast, between 1870 and 1940, the
inflation rate averaged only 0.4 percent per year. In other words, since
Keynes wrote the General Theory, the US economy has become far more
prone to inflation.
This is, I suspect, not merely a coincidence. At its broadest level, the
General Theory is a call for monetary and fiscal policymakers to control the
450 N.Gregory Mankiw
economy through the management of aggregate demand. In the aftermath of
the Great Depression, policymakers were ready to hear this message. The US
government accepted the challenge with the Employment Act of 1946, which
stated that ‘it is the continuing policy and responsibility of the Federal
government to…promote full employment and production’. In essence, Keynes
ushered in an era of discretionary demand policy.
It was not until the 1980s, however, that economists developed a good
understanding of why discretionary policy is intrinsically inflationary. The
literature on time inconsistency contains an important warning: at any point
in time, policymakers with discretion are tempted to inflate in order to reduce
unemployment. Economic actors, however, come to understand this
temptation and adjust their expectations of inflation accordingly. Higher
expected inflation in turn causes the short-run tradeoff between inflation and
unemployment to deteriorate. In the end, discretionary policy yields higher
inflation without lower unemployment.
The literature on time inconsistency has, in my view, provided a
persuasive case for some sort of commitment to a rule for monetary policy.
My own preference would be something like a target for nominal GNP or the
nominal wage. Without such a commitment to a policy rule, there is little
hope that the modern central bank will achieve the often stated goal of price
stability.
DOES MACROECONOMICS MAKE PROGRESS?
In some ways, the history of macroeconomic thought seems like a pendulum
swinging between two views of the economy. On the right is the classical
view of a well-functioning economy; on the left is the Keynesian view of an
economy fraught with market failure. The Great Depression of the 1930s
swung the pendulum decisively from the right to the left, and Keynes could
plausibly call classical economics ‘misleading and disastrous’. The new
classical economics of the 1970s swung the pendulum back to the right, and
Robert Lucas could plausibly proclaim the ‘death of Keynesian economies’.
The new Keynesian economics of the 1980s swung the pendulum back toward
the left (at least somewhat), and today one can plausibly say that Keynesian
economics has been reincarnated.
One might be tempted to conclude from these developments that
macroeconomics does not make progress, that it is destined to oscillate
between two irreconcilable extremes. Yet one can also take a more sanguine
view. The new classical revolution of the 1970s left an indelible mark on the
way macroeconomists of all stripes think about the economy, just as the
Keynesian revolution of the 1930s did before it. New Keynesian economics is
far different from old Keynesian economics—so different, in fact, that today
the label ‘Keynesian’ may generate more confusion than understanding. With
new Keynesians looking so much like old classicals, perhaps we should
conclude that the term ‘Keynesian’ has out-lived its usefulness. Perhaps we
The reincarnation of Keynesian economics 451
need a new label to describe the school of macroeconomics that accepts the
existence of involuntary unemployment, monetary non-neutrality, and sticky
wages and prices. Until a new label is found, however, we can safely say that
Keynesian economics is alive and well.
REFERENCES
Eckstein, Otto (1983) The DRI Model of the U.S. Economy, New York: McGrawHill.
Feldstein, Martin (1981) ‘The retreat from Keynesian economies’, The Public Interest
Summer, 92–105.
Feldstein, Martin (1988) ‘Counter-revolution in progress’, Challenge July/August,
42–6.
Friedman, Milton (1968) ‘The role of monetary policy’, American Economic Review 58
(March), 1–17.
Keynes, John Maynard (1936) The General Theory of Employment, Interest, and Money,
London: Macmillan.
Lucas, Robert E. Jr (1980) ‘The death of Keynesian economies’, Issues and Ideas Winter
(University of Chicago, Chicago, IL) 18–19.
Mankiw, N.Gregory and David Romer (eds) (1991a) New Keynesian Economics, Vol. 1:
Imperfect Competition and Sticky Prices, Cambridge, MA: MIT Press.
Mankiw, N.Gregory and David Romer (eds) (1991b) New Keynesian Economics, Vol. 2:
Coordination Failures and Real Rigidities, Cambridge, MA: MIT Press.
Tobin, James (1972) ‘Inflation and unemployment’, American Economic Review 62
(March), 1–18.
20 New-Keynesian economics today
The empire strikes back
Brian Snowdon and Howard Vane
American Economist (1995) 39, Spring, pp. 48–65
there is no single doctrine taken to be a scientific truth without the diametrically
opposed view being similarly upheld by authors of high repute…in other
fields of science these conflicts usually come to an end…It is only in the field
of economics that the state of war seems to persist and remain permanent.1
(Knut Wicksell)
INTRODUCTION
In a ‘Symposium on Keynesian Economics Today’ (Journal of Economic
Perspectives, Winter 1993) David Romer, James Tobin, Robert King, Bruce
Greenwald and Joseph Stiglitz offered a variety of perspectives with respect to
the current resurgence of Keynesian ideas which has characterized the
macroeconomics literature since the mid-1980s. In his introduction to the
symposium Gregory Mankiw noted that the ‘literature that bears the name
Keynesian is broad and it does not offer a single vision of how the economy
behaves’. However, as a leading new Keynesian he did not present his own
views in the symposium. In February 1993 we interviewed Gregory Mankiw
at Harvard University and here we present his perspective of the current state
of macroeconomics in general and what he has called the ‘Reincarnation of
Keynesian Economies’ (Mankiw 1992a).2
We first briefly review the background to the current debate before
presenting Mankiw’s assessment of some of the important issues in modern
macroeconomics. In conclusion we compare the varieties of Keynesian vision
presented by some contributors to this debate.
BREAKDOWN OF THE CONSENSUS
In 1977 James Tobin, the United States’ most distinguished ‘old’ Keynesian
economist, asked the question ‘How dead is Keynes?’ (see Tobin 1977). That
Tobin was even asking this question highlights the turmoil which had begun
to plague macroeconomics in the early 1970s and has continued ever since.
Following the publication of Keynes’s General Theory macroeconomists have
New-Keynesian economics today 453
been broadly split between those who believe that the price mechanism,
unaided by the visible hand of government, is capable of stabilizing a
capitalist market economy which is subject to periodic shocks and those, like
Tobin, who doubt the capacity of the system to self-equilibrate at a
satisfactory level of employment. The synthesis of Keynesian and neoclassical
analysis which formed the basis of a consensus in the 1950s and 1960s
appeared to have achieved an uneasy reconciliation between these two
competing views. The theoretical debate relating to the consistency of
macroeconomic equilibrium with an excess supply of labour appeared to have
been won by supporters of the invisible hand view, but as a practical matter it
was accepted that the self-righting properties of the market were too weak
and needed the helping hand of fiscal and monetary policies in order to
achieve and maintain the primary stated objective of full employment.
Keynesians of all persuasions accepted the possibility of widespread and
frequent ‘effective’ demand failures together with prolonged involuntary
unemployment. Nevertheless, apart from a small but highly vocal anti-
neoclassical group of heretics centered at Cambridge University, the majority
of Keynesians were also adherents, and seminal contributors, to the
neoclassical paradigm (Paul Samuelson and Robert Solow are the most
obvious examples). This schizophrenia could not last.
During the 1960s the synthesis became increasingly associated with an
acceptance of a stable long-run trade-off between inflation and
unemployment. With the breakdown of the Phillips curve in the late 1960s
and early 1970s it became apparent that the microeconomic underpinnings of
the supply side of Keynesian models were fundamentally flawed. The impact
of the first OPEC oil shock in 1973 made this even more apparent. As a result
Keynesianism was rejected by a growing number of academic economists
during the 1970s, especially in the USA, who were increasingly attracted to
the work of the emerging new classical school led and inspired by Robert
Lucas who for many is ‘the leading macro mountaineer of our generation’
(Parkin 1992). Lucas’s incorporation of John Muth’s rational expectations
hypothesis into a market clearing setting acted like a siren song to the
younger generation of graduate economists (Lucas 1972, 1973). By 1978
Lucas and Sargent were contemplating life ‘After Keynesian
Macroeconomics’. Soon after Lucas went so far as to claim that ‘people even
take offense if referred to as “Keynesians”. At research seminars, people don’t
take Keynesian theorizing seriously anymore; the audience starts to whisper
and giggle to one another’.3 In a similar vein, a leading ‘younger generation‘
Keynesian, Alan Blinder, confirmed that by 1980 ‘it was hard to find an
American academic macroeconomist under the age of 40 who professed to be
a Keynesian’ (Blinder 1988). Lucas’s obituary of Keynesian economics can
now be seen to have been premature. However, his critiques highlighted the
tensions which existed within economics between a flexiprice neoclassical
micro world dominated by the fundamental theories of Adam Smith and Leon
Walras, and a Keynesian superstructure where arbitrary assumptions relating
454 Brian Snowdon and Howard Vane
to nominal price and wage rigidities were the norm. This conflict was in need
of resolution as the conventional practice of separating micro from macro
analysis was no longer tenable. The new classical solution to this ‘crisis’ was
to adapt macroeconomic theory to neoclassical microeconomics. As Kevin
Hoover (1992) has noted, the new classical research programme ‘seeks not
only to revivify classical modes of equilibrium analysis, but also to secure the
euthanasia of macroeconomics’. In contrast the new Keynesian approach has
been to set about building new microfoundations for Keynesian
macroeconomics which nevertheless remain faithful to the axioms of utility
and profit maximization by individual agents. These ground rules relating to
optimizing behaviour have been set by new classical economists who insist
that no self-respecting model should contain agents who fail to ‘exhaust trades
that are to the perceived mutual advantage of exchanging parties’ (Barro
1979). In the language of Robert Lucas, any acceptable theory must not allow
$100 bills to be left lying on the pavement. Incorporating acceptable
microfoundations into macro models has been and remains the principal task
facing new Keynesian economists.
THE ‘REINCARNATION’ OF KEYNESIAN ECONOMICS
New Keynesian economics, conceived in the late 1970s, sprang to life in the
1980s. Since the essential feature of Keynesian macroeconomics is the absence
of continuous market clearing, the new Keynesian developments since the
mid-1980s have been primarily concerned with the ‘search for rigorous and
convincing models of wage and/or price stickiness based on maximizing
behaviour and rational expectations’ (Gordon 1990). In contrast to the new
classical monetary surprise and real business cycle models where price taking
rational individuals make voluntary choices with respect to quantities, new
Keynesian models contain price making, demand taking, risk-averse firms
who operate in an imperfectly competitive, uncertain world riddled with
imperfect information, transaction costs and asymmetric information
(Mankiw and Romer 1991). New Keynesian economics seeks to understand
and explain the causes of the imperfections in product, labour and capital
markets and to show how these imperfections have macroeconomic
consequences. In short ‘New Keynesianism throws bucket fulls of grit into the
smooth-running neoclassical paradigms’ (Leslie 1993). This agenda has led to
research into the causes and consequences of:
1 Nominal wage stickiness (see Fischer 1977; Taylor 1979; Laing 1993);
2 Nominal price stickiness (see Mankiw 1985; Akerlof and Yellen 1985;
Romer 1993);
3 Real rigidities (see Yellen 1984; Shapiro and Stiglitz 1984; Lindbeck and
Snower 1986; Phelps 1994);
4 Co-ordination failures (see Diamond 1982; Cooper and John 1988; Ball
and Romer 1991).
New-Keynesian economics today 455
Although new Keynesian theory is still at a rudimentary stage and there are
various strands to this diverse school, one of the leading advocates, Gregory
Mankiw (1992a) has claimed that ‘the new classical challenge has been met.
Keynesian economics has been reincarnated into a body with firm
microeconomic muscle.’ Mankiw also argues that this reincarnation was
necessary because ‘the new classical revolution seriously wounded the once
prevailing Keynesian consensus.’ In the development of Mankiw’s brand of
new Keynesianism it is evident that dissatisfaction with older style Keynesian
models emphasizing nominal wage rigidity played a crucial role. A
combination of price-taking firms, neoclassical production technology and
sticky nominal wages imply that aggregate demand contractions will be
associated with a rise in real wages during a recession i.e. real wages will
move countercyclically. By 1980 Mankiw had concluded that such models
made little sense even if modified to allow for rational expectations (e.g.
Fischer 1977) since they imply
that recessions must be quite popular. Sure, a few people get laid off. But
most people get to enjoy the higher real wages that result when prices fall
and their nominal wages do not…. If high real wages accompanied low
employment, as the General Theory and my professors had taught me,
then most households should welcome economic downturns.
(Mankiw 1991:129–30)
Since the weight of evidence suggests that real wages do not move counter-
cyclically over the business cycle and the assumption of nominal wage
rigidity seems to imply substantial departures from rationality, many
economists sympathetic to the old neoclassical synthesis view shifted their
attention from the labour market to the goods market in their search for
nominal rigidities. As Mankiw notes ‘In fact, it was thinking about the real-
wage puzzle that originally got me interested in thinking about imperfections
in goods markets and, eventually, about monopolistically competitive firms
facing menu costs’ (Mankiw 1991:132). When aggregate supply is derived
from inflexible goods prices, rather than from inflexible nominal wages, then
real wages can move procyclically or acyclically. Furthermore, new
Keynesians argue that price rigidities do not imply gross departures from
rationality given the existence of ‘near rational behaviour’ and ‘menu costs’
(see Akerlof and Yellen 1985; Mankiw 1985). In new Keynesian models the
reason why firms lay off workers during a recession is not because labour
costs are too high but for the intuitively appealing reason that sales are too
low (Mankiw 1991:106). Accordingly the new breed of Keynesian models
share the spirit of the old Keynesian economics in viewing the business cycle
as evidence of economy-wide market failure. This also implies accepting the
existence of involuntary unemployment, the non-neutrality of money, sticky
prices and wages, and non-clearing markets. However, it is important not to
assume that new Keynesians are protagonists in the monetarist-Keynesian
456 Brian Snowdon and Howard Vane
debate because new Keynesians do not hold a unified view with respect to the
relative potency of fiscal and monetary policy nor do they ‘necessarily believe
that active government policy is desirable’ (Mankiw and Romer 1991). The
work of Edmund Phelps has also inspired the emergence of a ‘structuralist’
branch to the new Keynesian school where non-monetary models are given
emphasis (see Phelps 1991: vol. III; Phelps 1994).
What progress has been made? What are the likely directions of further
research? How does new Keynesian analysis differ from old Keynesian and
new classical varieties? We sought answers to these and other questions
from Gregory Mankiw. In what follows, where appropriate, we have
provided references where the substance of Gregory Mankiw’s answers are
developed more fully or where some of the ideas discussed have been
applied.
GENERAL ISSUES
Why do you think we have so much controversy in macroeconomics
compared to microeconomics?
That is a hard question. It is certainly true that there is more agreement
among microeconomists as to how they approach things. That is, most
microeconomists start off with utility and profit maximization as the
underlying motives and go from there. Macroeconomics is in some ways
harder since you are dealing with the whole economy; the field therefore
requires more simplifying assumptions to make anything manageable, to
make the problem simpler than it really is in the world. I think there is
disagreement as to which simplifying assumptions are the most natural or the
most useful.
How important do you think it is for macroeconomics to have neoclassical
choice theoretic foundations?
Well it is certainly true that all macro phenomena are the aggregate of many
micro phenomena; in that sense macroeconomics is inevitably founded on
microeconomics. Yet I am not sure that all macroeconomics necessarily has to
start off with microeconomic building blocks. To give an analogy, all of
biology is in some sense the aggregate of particle physics, because all
biological creatures are made up of particles. That doesn’t mean that the
natural place to start in building biology is to start with particle physics and
aggregate up. Instead I would probably start with theory at the level of the
organism or the cell, not the level of the sub-atomic particle. We have a lot of
models like the IS-LM model in macroeconomics that are very useful for
studying the macroeconomy, even though those models don’t start off with the
individual unit and build up from there.
Which papers or books do you feel have had the biggest impact on the
development of macroeconomics over the last 25 years?
New-Keynesian economics today 457
The biggest impact has undoubtedly come from Lucas. He put the cracks into
the Keynesian consensus that existed in the 1960s. He really pulled
macroeconomics apart by proposing new and intriguing ideas. The
disagreements today among macro economists have largely arisen from the
critiques of Lucas and of his followers. As you know, I don’t agree with
Lucas’s solutions, but I take the problems that he pointed out very seriously. A
lot of the work that I and other new Keynesians have done are a response to
the problems that he pointed out in the old Keynesian ideas.
(See Mankiw 1990.)
To some extent you’ve answered our next question. Where did you draw
inspiration for your own work?
It’s been a combination of influences. Part comes from the older generation of
macroeconomists. I view a lot of the work I do as building on the work of
Tobin, Modigliani and Friedman. I see a lot of truth in the views they were
pushing. I also take the problems that Lucas pointed out very seriously. A lot
of new Keynesian work is trying to reformulate the 1960s Friedman-Tobin
view of the world. What is now called the neoclassical synthesis had a large
element of truth in it. On the other hand, it had problems, and Lucas pointed
those problems out very forcefully. We need to fix those problems and address
the concerns that Lucas had while still maintaining the element of truth in the
neoclassical synthesis.
(See Mankiw 1992a.)
ON KEYNES AND THE GENERAL THEORY
One interpretation of the neoclassical synthesis which emerged at the end
of the 1950s suggested that the General Theory was a special case of a
more general classical model. Would you agree with that interpretation?
I would say that the classical model and the Keynesian model make different
assumptions about adjustment of prices. I think of the classical model as
being the model that assumes complete price flexibility, and therefore
describes a horizon over which it is plausible to make such an assumption.
Probably a period of years, rather than a period of months. The Keynesian
model applies over a horizon where wages and prices are relatively inflexible
or sluggish. Both models are special cases of a more general model which
allows a varying degree of flexibility and sluggishness in prices depending on
the horizon we want to study. When we study the effect of policies over a
quarter or a decade, we want to make a different assumption about the degree
of flexibility of prices.
Why do you think there are so many conflicting interpretations of the
General Theory?
There are a lot of conflicting interpretations because Keynes had a lot of
different ideas. The ideas don’t necessarily have to be packaged all together,
458 Brian Snowdon and Howard Vane
so some people grab on to one set of ideas and say that this is really what is
central to what Keynes was saying and other people grab onto other sets of
ideas. The question is, when we look at the market imperfection that we call
the business cycle, which set of general ideas from the General Theory are the
most important? There is so much in the General Theory that it is hard to
comprehend it all at once. Some is very important, but some is not
particularly important. Disagreements come by choosing different pieces of
Keynes’s world view and emphasizing those.
(See Mankiw 1992a.)
Do you think that if Keynes had still been living in 1969 he would have
received the first Nobel Prize in Economics?
Oh undoubtedly. I think there are a few very very important economists of the
century, and there is no question that Keynes has got to be on anybody’s
shortlist.
NEW CLASSICAL MACROECONOMICS
Do you regard new classical macroeconomics as a separate school of
thought from monetarism?
I think so. My impression is that monetarism is a school of thought that says
fluctuations in the money supply are the primary cause of fluctuations in
aggregate demand and income, whereas new classicism is a particular theory
as to why fluctuations in aggregate demand might matter through an
unanticipated price surprise. This price surprise view proposed by Lucas is, I
think, the next step after monetarism. More recently, new classical economists
have turned their attention to real business cycle theory, which is the
antithesis of monetarism.
Do you think that overall the new classical contributions have had a
beneficial effect on the development of macroeconomics?
Debate is healthy, and the new Keynesian school arose largely in response to
the new classical school. In that sense it is a debate leading to greater truths,
and it has been helpful. A lot of the specific contributions, especially real
business cycle theory, are probably not going to survive the test of time. The
literature on the time inconsistency of policy is a contribution that will
survive and has probably been one of the most important contributions to
policy analysis in the past two decades.
(See Mankiw 1988a: 441–3; 1992a: 563–4.)
How important is the rational expectations hypothesis?
It is important in the sense that it has now become the working hypothesis of
all practicing macroeconomists. Economists routinely assume that people are
rational when they make decisions: they maximize utility, they ration-ally
New-Keynesian economics today 459
maximize profits, and so on. It would be peculiar for us to assume that people
are rational except when they come to form expectations and then they act
irrationally. I don’t think the rational expectations hypothesis is important in
the sense of having all the sweeping implications as was at first believed. At
first people thought that it had all sorts of properties about policy being
ineffective.
(See Mankiw et al 1987; Mankiw 1990.)
Isn’t that more to do with the market clearing assumption?
Exactly. People have come to realize that it is other assumptions, like the
market clearing assumption, that are really important and that rational
expectations in itself doesn’t have implications as sweeping as was once
thought.
You have questioned the argument that the disinflation experience of the
early 1980s both here and in Britain has provided decisive evidence
against the new classical claim of painless disinflation. Is this because the
deflation was unanticipated?
There are two new classical views. The first is the price surprise theory of
Lucas. The second is real business cycle theory. This second view says that
money anticipated or unanticipated doesn’t matter. My view of that is that it
is completely at variance with the evidence. Larry Ball has a paper that
shows systematically for a large number of countries that whenever you have
a major disinflation it is associated with a period of low output and high
unemployment (see Ball 1994). So I think that the evidence is completely
clear on that. The evidence is more favourable to early new classical theory.
You’re right that to a large extent the disinflation was unanticipated even in
the United States where Volcker said he was going to disinflate. I don’t think
people believed he was going to disinflate as fast as he did. Most measures of
expectations of inflation did not come down until after the recession was well
under way. I am sympathetic to the view that credibility is one determinant of
how costly a disinflation will be.
(See Mankiw 1986a: 218–20.)
ON KEYNESIANISM AND THE NEW KEYNESIANS
Do you regard yourself as a Keynesian?
I do but I’m always nervous about the term because the term Keynesian
can mean different things to different people, just as different people will
read the General Theory and pull out different elements as being
important. People use the word Keynesian in so many different ways that
recently I have actually tried to avoid using the term at all on the grounds
that it is more confusing than illuminating. I think of myself as a
Keynesian in the sense of believing that the business cycle represents some
sort of market imperfection on a grand scale. In that sense I think of
460 Brian Snowdon and Howard Vane
myself as a Keynesian. Milton Friedman was also a Keynesian in that
sense. My own views emerged as much from Milton Friedman as they
have from John Maynard Keynes. Some people take the word Keynesian
as meaning a belief in fine tuning the economy so that the government
controls every wiggle of ups and downs. Other people take it as a belief
that deficit spending is not a bad thing. I don’t subscribe to either of those
views. I think that the broad theme of the General Theory is that the
business cycle is something that we really need to worry about because it
is a sign of a market imperfection. In that way I am a Keynesian, but as I
said so is Milton Friedman.
(See Mankiw 1987; 1992a.)
Was the breakdown of the Phillips curve fatal for Orthodox Keynesianism?
It highlighted the absence of a good theory of aggregate supply. What
orthodox Keynesians had was a pretty good theory of aggregate demand.
The IS-LM model has held up pretty well as a general structure for thinking
about how aggregate demand is determined. The problem is once you’ve
got aggregate demand—a downward sloping curve in P-Y space—you still
need a good story for the aggregate supply curve. The Phillips curve came
out of nowhere. It is really just an empirical description of what was true in
the data without any particularly good theories as to why it should look
that way, how it would change in response to policy, and what might make
it unstable. So we never had a good theory of that, and the breakdown of
the Phillips curve made that very apparent and provided room for the more
general critique that Lucas put forward. The deficiency on the supply side
was always a weakness, but it wasn’t given attention until the Phillips curve
broke down.
(See Mankiw 1990:1647–8.)
What would you summarize as being the central propositions of new
Keynesian macroeconomics?
The central propositions are largely theoretical rather than policy oriented.
New Keynesians accept the view of the world summarized by the neoclassical
synthesis: the economy can deviate in the short term from its equilibrium
level, and monetary and fiscal policy have important influences on real
economic activity. New Keynesians are saying that the neoclassical synthesis
is not as flawed as Lucas and others have argued. The purpose of the new
Keynesian school has been largely to try to fix those theoretical problems
raised by Lucas and also accept Lucas’s argument that we need models
supported by better microeconomic foundations.
(See Ball et al 1988:149–61; Mankiw and Romer 1991, vol. 1:1–26.)
So you wouldn’t subscribe to arguments in favour of incomes policies
advocated by Post-Keynesians?
New-Keynesian economics today 461
No, not at all. When the government gets in the business of setting wages and
prices it is not very good at it. The setting of wages and prices should be left
to free markets.
So you are no Galbraithian?
Absolutely not (laughter).
How important is the theory of imperfect competition to new Keynesian
macroeconomics?
A large part of new Keynesian economics is trying to explain why firms set
and adjust prices over time in the way they do. Firms in a perfectly
competitive environment don’t have any choice over what their prices are
going to be. Competitive firms are price takers. If you want to even talk
about firms setting prices you have to talk about firms that have some ability
to do so, and those are firms that have some market power: they are
imperfectly competitive. So I think imperfect competition is central to
thinking about price setting and therefore central to new Keynesian
economics.
(See Mankiw 1985; 1988b; Ball et al 1988:156–8.)
This is strange, because if you think of the 1930s, you had Keynes and
Joan Robinson at Cambridge. Joan Robinson developed the theory of
imperfect competition and Keynes developed his General Theory. Why did
it take so long to bring these two ideas together?
I don’t think that Keynes was as worried about building his model based on
microfoundations as we are today. Joan Robinson was building the
microeconomics that would later prove to be very useful for addressing the
macroeconomics of Keynes, but Keynes, not having read Robert Lucas yet,
wasn’t worried about building the microeconomics of aggregate supply
(laughter).
In a sense haven’t the Post-Keynesians been ahead of you here? People like
Paul Davidson have for years taken imperfect competition as their
microfoundation. So are the new Keynesians simply catching up on what
the Post-Keynesians did quite a while ago?
They have a broad theme of imperfect competition, but the details are not
very similar. My impression is that the new Keynesian economics is much
more in line with the neoclassical synthesis than with the Post-Keynesians.
You will obviously be very familiar with Alan Blinder’s recent surveys. Are
they supporting the new Keynesian views? (See Blinder 1991.)
Alan is providing a way of judging a variety of different new Keynesian
views. There are a lot of new theories about wage and price rigidities. He is
trying to sort out which is right and wrong using a fairly novel perspective of
asking firms how they set wages and prices. This is terrific work, but what we
462 Brian Snowdon and Howard Vane
are going to learn in the end is still unclear. He is still producing the papers
and we haven’t seen all the results yet. The goal is to provide one way of
deciding which theories we like and which we don’t. It’s a very exciting
project.
An important distinction seems to be made by new Keynesians between
real rigidities and nominal rigidities. Why is it important to make this
distinction?
The reason is that a real rigidity, which is a rigidity in a relative price, is not
a reason for monetary non-neutrality. Unions, for example, could set rigid
real wages away from equilibrium. A rigid real wage is not going to provide
any reason to believe that money is not neutral, since it does not create any
nominal lever for money to work on. It would cause unemployment but not
monetary non-neutrality. To get monetary non-neutrality, which is a central
challenge for macro theorists, you need some nominal rigidity such as sticky
prices. Having said that, there do seem to be a variety of real rigidities in the
world; unions setting wages way above equilibrium levels for example. The
question is whether nominal and real rigidities interact. One of the big themes
of this literature, mainly due to Larry Ball and David Romer, is that real and
nominal rigidities seem to reinforce each other. The real rigidity is actually
going to make the nominal rigidity a lot more important than it would be
otherwise.
(See Ball et al 1988:153–6; Ball and Romer 1990; Mankiw and Romer
1991, vol. 2:2).
Critics of the menu cost literature, Robert Barro for example, have
suggested that this is a small peg on which to hang an explanation of the
business cycle. How can small menu costs have such large real effects on
the macro economy? (See Barro 1989.)
It is clear that menu costs are quite small. Firms don’t bear huge costs
when they change their prices. Yet it also is clear that recessions are very
costly events. The question is whether these relatively small menu costs
can be a key part of understanding this relatively costly business cycle.
This literature shows that price adjustment by firms has external effects.
When a firm decides to keep prices sticky, this could well be costly for the
economy in a way that is not costly for the firm who is making the
decision.
(See Mankiw 1985; Ball et al. 1988.)
How do efficiency wage and insider/outsider theories fit into new
Keynesian thinking?
Both of those theories provide a particular explanation for real rigidities,
such as why real wages don’t move to the equilibrium level in labour
markets. As I said before, real rigidities and nominal rigidities can
complement each other. That is, the insider/outsider and efficiency wage
New-Keynesian economics today 463
explanations for rigid real wages in some senses complement the menu cost
story of rigid prices.
(See Mankiw 1990:1658.)
Is the idea of hysteresis crucial to new Keynesian macroeconomics?
Actually I don’t think of it as being crucial. It is an interesting idea, that a
recession can have long-lived effects on the economy and leave permanent
scars after the initial cause of the recession has gone. For example, the high
unemployment in Europe in the 1980s persisted far longer than anyone could
explain with standard models. But if this idea turned out to be wrong it would
not bring down the rest of our theories. This has been an interesting, but
relatively separate question.
Do you see the concept of NAIRU, and Friedman’s natural rate, as being
the same idea or are they different?
I have always thought of them as being basically the same. Most new
Keynesian models involve some sort of natural rate; in that sense Milton
Friedman has won the debate. Most new Keynesians believe in the natural
rate hypothesis except for a small group of people working with hysteresis.
The natural rate hypothesis is pretty well entrenched.
(See Mankiw 1992a: 563; 1992b: 483–4.)
What about the concept of full employment? It was difficult to think of
doing macroeconomics 15–20 years ago without the concept of full
employment being central. What do we do about issues like involuntary
unemployment? Lucas suggests that we should abandon this concept, what
are your views on this? (See Lucas 1978.)
I think there is involuntary unemployment. Part of the new Keynesian
literature has come up with models of the labour market to explain why
involuntary unemployment exists, why real wages don’t adjust to equilibrate
labour markets. There is a lot of truth to the efficiency wage theories and the
insider/outsider theories, for example.
Do new Keynesians think of full employment as the natural rate?
I avoid the term full employment because it suggests that the natural rate is in
some sense desirable. I think there is some natural rate which is the long-run
unemployment rate that the economy tends to, that can’t be influenced by
monetary policy in the long run. That doesn’t mean that it is immutable in
response to any policy intervention. There are things that have been done to
the labour market that either increase or decrease the natural rate, things like
the minimum wage, unemployment insurance laws, labour training policies.
There are all sorts of things that the government can do to change the natural
rate. I don’t like calling it full employment because good labour market
policies might well raise employment beyond that level.
(See Mankiw 1992b: 118–39.)
464 Brian Snowdon and Howard Vane
How important do you think it is to take into account fairness when
looking at the labour market? We are thinking here of the work of George
Akerlof, Janet Yellen and Robert Solow who have stressed the idea of
fairness. Doesn’t this work suggest that perhaps new Keynesians should
start looking more closely at the psychology and sociology literature? (See
Akerlof and Yellen 1990; Solow 1990.)
Some of the papers that they have written have been extremely interesting.
I don’t think there is a lot of compelling evidence yet that we need to
abandon neoclassical assumptions. I’m not doing so yet in my work, but
I’m certainly happy to read the work of others who are doing so
(laughter).
In your edited volumes of collected papers on new Keynesian economics
you say that ‘new Keynesian macroeconomics could just as easily be
labelled new monetarist economies’. What exactly did you mean? (See
Mankiw and Romer 1991.)
The challenge raised by the real business cycle school is the question of
whether money is neutral and, if not, why not? Twenty years ago, when
Friedman and Tobin were debating, there were some things they agreed on.
They agreed on the proposition that the Federal Reserve was an important
player in the economy, that what it did really mattered. The real business
cycle school has challenged that by writing down models without any real
effects of monetary policy. What the new Keynesian models have tried to do
is establish why money is not neutral, what microeconomic imperfections are
necessary to explain monetary non-neutrality at the macro level. In this sense,
these models are trying to support both traditional Keynesian and monetarist
views.
Would you agree with Stanley Fischer that the views of Friedman, Brunner
and Meltzer are closer to those of Keynesians than they are to equilibrium
business cycle theorists? (See Fischer 1988.)
Oh yes absolutely. The essence of real business cycle models is the absence of
any role for the Federal Reserve, whereas I think Brunner, Meltzer and
Friedman would agree with Tobin that the Fed is very important. None of
them would ever argue that money is neutral in the way that real business
cycle theorists have.
(See Mankiw 1986b.)
James Tobin has suggested that good papers in economics contain
surprises. What surprises have new Keynesian papers uncovered? (See
Tobin 1988.)
One of the big surprises is that one can go a lot further with menu cost
models than people once thought. A lot of people used to see these models
as a silly way of thinking about price rigidity. What the new literature is
New-Keynesian economics today 465
trying to do is to say no, maybe we should take menu cost models seriously.
I think the complementarity between real and nominal rigidities is a
surprise. As I mentioned earlier one of the disappointing features so far of
the new Keynesian literature is that it hasn’t been as empirical as I would
have liked. That is a problem being remedied right now in some research.
Ultimately that is where the literature should go. More empirical work is
needed.
(See Mankiw 1987; Ball et al 1988:161–201; Ball and Mankiw 1992a.)
Peter Howitt has talked about a Keynesian recovery, Alan Blinder about a
Keynesian restoration, you seem to prefer the term reincarnation. Is there
something important in the different terms used? (See Howitt 1990; Blinder
1992a; Mankiw 1992a.)
I chose the term reincarnation because it means rebirth into another body.
While there are many similarities between new and old Keynesian economics,
there are also a lot of differences as well, and I wanted to emphasize that. In
some senses the spirit of Keynes has been brought back, but it doesn’t look
like the old Keynes. In fact Keynes might not recognize the new Keynesians as
Keynesians at all. In general, people might not recognize themselves after
they have been reincarnated. So that is why I used the term reincarnation
(laughter).
Would you say that your work is, with respect to Keynes, faithful in spirit,
but critical in detail?
I think that is fair. It tries to go beyond Keynes in a sense of taking
microfoundations more seriously. Alan Blinder wrote a paper ‘Keynes after
Lucas’ and I think that title pretty much describes new Keynesians. It takes
some of Keynes’s ideas seriously, and it also takes some of the critiques of
Lucas seriously as well.
(See Blinder 1986.)
Do you think Keynes would have been a new Keynesian?
I don’t know, I think Keynes was a very unpredictable fellow. I guess he
would see some things in it he would like, and some things in it he wouldn’t.
REAL BUSINESS CYCLE THEORY
You’ve argued that real business cycle theory has served an important
function in stimulating and provoking scientific debate, but you predict
that the approach will eventually be discarded. What are your main
objections to real business cycle theory? What are the weaknesses,
theoretical, empirical or both?
My objections are mainly empirical. Theoretically they are very elegant
models and that is a large part of their appeal. They are very parsimonious
466 Brian Snowdon and Howard Vane
models. But when I look at the real world I see the same things that Milton
Friedman and James Tobin do, which is a very powerful Federal Reserve
board in the United States or the Bank of England in the UK. There is a lot
of evidence across countries that periods of disinflation are periods of low
output and high unemployment. Those effects are completely absent in real
business cycle models. I think the central driving forces for the business
cycle that those models highlight—technology shocks—aren’t very
important.
(See Mankiw 1989; 1994; Campbell and Mankiw 1989).
Isn’t the pro-cyclical behaviour of the real wage a strong feature of these
theories? How do new Keynesians explain the movement of real wages
over the business cycle?
The theories do predict pro-cyclical wages. Although I’ve not looked at the
models carefully on this question, my understanding is that they predict
very pro-cyclical, real wages. While it is true that real wages are pro-
cyclical, my reading of the evidence is that they are only mildly
procyclical. Therefore, the fact that these theories predict very pro-cyclical
real wages, and the data show that they are only mildly pro-cyclical,
makes it hard to reconcile this model with the evidence. I think the real
wage evidence is not that hard to explain. If you believe in a world where
wages and prices are sluggish over time, the cyclical behaviour of the real
wage is really a question of whether wages or prices are more sluggish.
The fact that real wages are roughly a-cyclical, maybe slightly pro-
cyclical, is some indication to me that wages and prices are simply
equally sticky. This is consistent with Alan Blinder’s evidence which says
that prices change on average once a year, and we know a lot of wages
change on average once a year. So I think that explanation is consistent
with a lot of the evidence.
(See Blinder and Mankiw 1984; Mankiw et al. 1985; Mankiw 1991).
How do we explain pro-cyclical productivity? Some Keynesians seem to
suggest that it is due to labour hoarding.
The pro-cyclical behaviour of productivity is a puzzle for people who
don’t believe in technology shocks. The traditional explanation for why
productivity is pro-cyclical is labour hoarding. In recessions firms keep on
workers they don’t really need so that they can have the workers still
available when the next boom comes, and that tends to give the
appearance of pro-cyclical productivity. These theories make a lot of sense
to me. I know I work my secretary harder when I have more work to be
done; therefore her productivity is pro-cyclical. I know I work harder
when there is more work to be done (laughter). I think there is a lot of
causal evidence that labour hoarding and pro-cyclical effort are
important.
(See Mankiw 1989:83–5.)
New-Keynesian economics today 467
ON MACROECONOMIC POLICY
One of the central ideas of Keynesian economics is that an increase in
aggregate demand will stimulate the economy. Under what circumstances
do you think a government should actually stimulate demand?
There are a couple of questions. First, when should it act? Second, how
should it act? That is, should it use monetary or fiscal policy? On the first
question, one should stimulate aggregate demand when it is too low to
maintain full employment—that is when you observe very high
unemployment or when there is reason to believe that unemployment is
going to rise. The policy implications of a lot of new Keynesian theories
really go back to a lot of the policy implications of the neoclassical
synthesis of the 1960s. Some of the limitations on policy that were then
debated are still relevant today. Even if you accept everything that new
Keynesians say about prices being sluggish and so on, there is still the
question of how good the government is at responding in a timely fashion to
the shocks? In that debate, I side to a large extent with Milton Friedman.
The government is very bad at recognizing shocks in a timely fashion, and
when they do respond to shocks they often do so quite late and often
counter-productively. So while I see the business cycle as a sign of market
failure I also think it is a kind of market failure that a government is very
limited in its ability to fix. If we have a very deep persistent recession,
certainly something on the lines of the Great Depression, there is room for
the government to do something. For the relatively minor wiggles that we
have experienced in the post-war economy, it is not clear that the
government can do a lot better than it has.
(See Mankiw 1992b: 322–41.)
Do you think Keynes was politically naive in thinking that politicians
would be advised by technocrats and take the correct action? We are
thinking here of the public choice literature and the political business cycle
literature. Can we actually trust politicians once they have their hands on
the fiscal and monetary levers to use them in the right way?
I think that is a serious concern but there are a lot of ways of fixing that
problem. For example, there is a large literature showing that countries with
more independent central banks have a lower inflation on average. With less
independence in the central bank, there is more political pressure and
therefore a greater possibility of following a policy of inflating too much.
There are ways around the political problem, like making independent
central banks, which to some extent are staffed by technocrats. For that
reason an independent central bank would be better at fine-tuning the
economy, to the extent we fine tune it at all, compared to fiscal policy which
is always run by politicians.
(See Mankiw 1992b: 331–3; 1994.)
468 Brian Snowdon and Howard Vane
You’ve said that the literature on time inconsistency has provided a
persuasive case for a commitment to some sort of rule for monetary policy,
do you also support fiscal rules?
Fiscal rules have to be well crafted. A balanced budget amendment that is too
strict could be a disaster. At certain times, like recessions and wars, it is
appropriate to run budget deficits. So any fiscal rule has to take into account
those special situations where budget deficits are the appropriate policy
response. A fiscal rule by itself wouldn’t be a bad idea, but it has to be well
crafted and so far I haven’t seen one that is.
Isn’t one of the problems with devising rules that if the economy is hit by
an unforeseen shock then the government really has to renege on that rule
and take some discretionary action? It is difficult to think of a rule which
really would be binding.
There are two parts to the question. First, how might you make the rule
binding? Second, do you want to make the rule binding? One way to make
the rule binding is reputational. Many rules are rules just because long
tradition has established them as rules and people don’t want to break
tradition. Another more legalistic way of imposing rules is by writing them
into the constitution. I think the harder question you raise is do you want to
make rules binding? The question is whether you can write a rule that works
well even in response to unforeseen events. If it becomes too costly to be tied
by the rule people will stop abiding by it. What we want to do is write down
a rule that will be good in response to normal kinds of shocks. That is, you
don’t know what the shocks are going to be, but you know what kind of
shocks are possible. You’ve got oil shocks, monetary demand shocks and so
on. You write down a rule that is good in response to the kinds of shocks you
expect the economy to experience, based on the shocks experienced in the
past. Therefore, unless something completely unforeseeable happens, you
stick by the rule.
Leijonhufvud once argued that the economy can be thought of as travelling
along a corridor, as long as it stays in the corridor leave it alone, but if it
gets out of this corridor into a severe recession that is the time for
intervention. Is that what you are saying? (See Leijonhufvud 1981.)
Well no, because recessions are reasonably foreseeable. Although you don’t
necessarily know when a recession is going to occur, you know that one will
occur eventually. A recession is one of the contingencies that you want your
rule to deal with. So I don’t think a recession per se is one of those
extraordinary events that make you want to break the rule. A recession is
something you can plan for in advance. I’m talking about an event that not
only can you not predict when it is going to happen, but you have never even
thought that it might happen. For example, before 1973 people never
imagined an OPEC supply shock. The whole idea of OPEC never even
New-Keynesian economics today 469
crossed anybody’s mind. That is the type of situation where you might want
to rethink the rule. Now that we know what OPEC is capable of, we can
write down a rule that takes oil shocks into account.
What is the role of fiscal policy in new Keynesian macroeconomics?
To a large extent new Keynesian economics has been about the theory of
aggregate supply and why it is that prices adjust slowly. It has been relatively
neutral on the question of what determines aggregate demand, in particular
whether monetary or fiscal policy levers are most useful. As I mentioned a
moment ago, I am skeptical personally about the usefulness of fiscal policy in
fine tuning the economy because, at least in the United States, the Congress
acts very slowly. Even as we are doing this interview (18 February 1993) the
Congress is debating a fiscal stimulus, even though the recovery has been
going on for about a year now. By the time this fiscal stimulus actually has
an effect on the economy, my guess is that we will be pretty close to the
natural rate again. This is the perfect example of how the lags can be very
long in fiscal policy. Monetary policy is a more useful tool for stabilizing
aggregate demand.
(See Mankiw and Summers 1986.)
Do budget deficits matter?
I think they matter a lot. The main way they matter is not for short-run
macroeconomic reasons but for long-run reasons—reasons that are best
described not by Keynesian models but by growth models. The evidence as
I see it is that large budget deficits reduce national saving. And the lesson
from growth theory and growth experience across countries is that low
saving leads to low growth. This is a big problem for the United States
today.
(See Mankiw 1992b: 423–35.)
If you were advising President Clinton about macroeconomic policy for the
next three or four years what would be the kinds of policies you feel are
necessary?
My reaction to President Clinton’s speech (17 February 1993) is that I don’t
think we need the fiscal stimulus that he is proposing. Recovery is already on
its way. It wasn’t a very deep recession to start off with, so I’m not terribly
shocked that there is a mild recovery. It will take the fiscal stimulus a while
to get people employed. I am happy that he is worried about the budget
deficit, as low national saving is an important macro problem in the long
term in the United States. Yet I am disappointed that he is putting so much
emphasis on tax increases rather than spending cuts. That is really a view not
so much about macroeconomics as about the size of government. I am also
disappointed that he is giving no attention to the low rate of private saving in
the United States. I would recommend tax reforms to remove the present
disincentives toward saving. So I give him a mixed review.
470 Brian Snowdon and Howard Vane
CURRENT AND FUTURE PROGRESS IN MACROECONOMICS
Much research in the 1980s, your own included, was directed at providing
more rigorous microeconomic foundations for the central elements of
Keynesian economics. Taking an overview of the last decade how
successful do you think that research has been in providing a more
substantial micro foundation for Keynesian economics?
It has been successful at the theoretical level in the sense that one can now
say that Keynesian economics, the economics of wage and price rigidities, is
well founded on microeconomic models. There are now several
microeconomic models that people can pull off the shelf. The theoretical
challenge of Lucas and his followers has been met. It is less clear whether
this line of research is going to be successful as an empirical matter. That is,
to what extent does it yield new insights to help us understand actual
economic fluctuations? Does it give us new ways to look at data and
policies? The jury is still out on that one. There is a small empirical
literature, but I can probably count the number of empirical papers on the
fingers of two hands. I hope it is a growth area, but so far the literature has
not been as empirically oriented as I would like.
(See Ball et al 1988; Ball and Mankiw 1992a, 1992b; Mankiw 1985.)
Do you think there is some truth to the view that at the moment we have
too many theories?
Yes, I have a lot of sympathy with that view. There is too big a premium for
coming up with clever new theories in the profession. Yet I don’t know of any
way to solve this problem. Obviously I believe the things I believe, and I
can’t tell people that they should believe what I believe, just because there are
too many theories (laughter). It would be nice if macroeconomists reached a
consensus and they could do more work on details and less work on creating
brand new theories of the business cycle. Until we do naturally reach a
consensus, there is no way to enforce that by fiat.
Do you see any signs of an emerging consensus in macroeconomics?
That is a good question. I change my mind on that a lot depending on what
conference I go to (laughter). I think there are certainly groups within the
profession that are agreeing with each other. There is much agreement among
new Keynesian people like Olivier Blanchard, Larry Ball, David Romer,
George Akerlof, Alan Blinder and so on. Whether we as a group are coming
to agreement with some of the real business cycle group is hard to say. I’m
delighted that some of the people who previously worked closely with the real
business cycle models are now trying to incorporate monetary effects into
those models. That provides a hope that somewhere down the line the new
Keynesian models and the real business cycle models are going to merge to
some grand synthesis that incorporates the strengths of both approaches. That
hasn’t happened yet; that is just a hope.
(See Mankiw 1989:88–9.)
New-Keynesian economics today 471
MANKIW’S REINCARNATED KEYNESIANISM
Gregory Mankiw provides a relatively optimistic vision of the future of
Keynesian macroeconomics which in his view has been ‘re-incarnated’ since
the mid-1980s rather than ‘resurrected’ in its old form. His Keynesian vision
shares the spirit of Keynes in seeing economy-wide market failures caused by
the inability of the invisible hand to maintain full employment. However,
following the new classical critiques Mankiw, like many other Keynesians,
accepts that the fatal defect of the neo-Keynesian synthesis model was the lack
of an adequate theory of aggregate supply. To Mankiw and other new
Keynesians the wage and price rigidities characteristic of Keynesian models
could no longer remain as an assumption but required theoretically rigorous
foundations. The vitality of the new classical revolution in the 1970s was
attributable as much, if not more, to theoretical flaws in the supply side of the
Keynesian model as it was to empirical dissatisfaction (see Mankiw 1988a).
To remedy those theoretical flaws by building a Keynesian theory of
aggregate supply which can rationally account for wage and price rigidities
and hence the non-neutrality of money is for Mankiw the paramount job
facing Keynesian theorists today. Indeed Mankiw goes further and argues that
the reconstruction appears to be well on the way and that Lucas’s criticisms
relating to the microfoundations of Keynesian models have been met.
OLD V NEW KEYNESIANS
Although they share the spirit of Keynes, and Mankiw’s work has in part been
inspired by older Keynesian views, the difference between James Tobin’s
‘unreconstructed old Keynesian’ views and Mankiw’s reincarnation are
striking. For Mankiw (and Romer) nominal aggregate demand disturbances
have real effects because wages and prices are rigid (see Romer 1993). In
sharp contrast Tobin argues that ‘Keynesian macroeconomics neither asserts
nor requires nominal wage and/or price rigidity’ (Tobin 1993a). Indeed that
Keynesian economics is defined by price rigidities is especially misleading if
it suggests that such an assumption is necessary to generate Keynesian results.
For Tobin the empirical fact that markets do not clear instantaneously leaves
room, for ‘flexibility in any common sense meaning of the word’ and the
resulting excess supply regimes allow quantities to determine quantities with
output and employment constrained by deficient effective demand. Tobin also
argues that the classical equilibrating mechanisms are ‘weak’ or ‘possibly non
existent or perverse’ and certainly require help from activist government
fiscal and monetary policy. According to Tobin (1993b) the suggestion that
‘Keynesian economics is doomed without new theories to explain price and
wage rigidities is to misunderstand Keynes himself and old Keynesian
economies’. It is evident from Mankiw’s answers to our questions that his
brand of new Keynesianism has been heavily influenced by the theoretical
and empirical contributions of Friedman (1968), Lucas (1972) and Kydland
472 Brian Snowdon and Howard Vane
and Prescott (1977). In particular he questions the desirability of activist
discretionary fiscal policy as a stabilizing weapon and following the new
classical work on the dynamic inconsistency of monetary policy he has also
been persuaded by the arguments in favour of a monetary rule, something old
(and some new) Keynesians would never subscribe to.
Bruce Greenwald and Joseph Stiglitz, both leading new Keynesians, appear
to occupy an intermediate position somewhere between that of Tobin and
Mankiw. Greenwald and Stiglitz support Tobin’s old Keynesian position that
increasing wage and price flexibility might well exacerbate a recession. This
alternative new Keynesian view suggests ‘that natural economic forces can
magnify economic shocks that may seem small and that existing price
rigidities may reduce the magnitude of the fluctuations as Keynes argued’ (see
Greenwald and Stiglitz 1993a). Hence for Greenwald and Stiglitz the single-
minded focus by Mankiw and others on wage and price rigidities would
appear to be somewhat misguided. With respect to the policy implications of
new Keynesian economics Stiglitz supports a more interventionist stance than
Mankiw. Most recently Stiglitz has noted that new Keynesians ‘disagree with
virtually every one of the presumptions underlying non-interventionist
theories’ (see Stiglitz 1993: ch. 39). He argues that the government on
balance has done more to stabilize than destabilize the economy and should
certainly not bind itself to fixed rules of the kind advocated by Milton
Friedman (1968), Robert Barro (1986), Finn Kydland and Edward Prescott
(1977). Like Tobin, Stiglitz favours discretionary policies because ‘changing
economic circumstances require changes in economic policy, and it is
impossible to prescribe ahead of time what policies would be appropriate’.
Indeed Stiglitz questions whether it would ever be possible for a government
to stick by a rule because ‘if the unemployment rate becomes high,
government must and will do something regardless of what is said’ (Stiglitz
1993). However Stiglitz, like Mankiw, is not as optimistic as old Keynesians
on the ability of government to fine tune the economy. Here the monetarist
and new classical arguments have modified all new Keynesian views (and no
doubt some old ones also). Stiglitz, like Mankiw, accepts that ‘by attempting
to do too much the government may do worse than it would if it were less
ambitious’ (Stiglitz 1993).
Old and new Keynesians alike are united in their view that the traditional
IS-LM model remains the best way to think about the demand side of the
macro models although Tobin gives emphasis to real rather than nominal
demand shocks (see Tobin 1993a). Their more unified position here differs
considerably from that of equilibrium business cycle theorists. For example in
Robert King’s contribution to the Symposium on Keynesian Economics Today,
he criticizes new Keynesians like Mankiw for maintaining their faith in the
textbook IS-LM model (see King 1993; Mankiw 1990). This is because ‘of its
treatment of expectations the IS-LM model, as traditionally constructed and
currently used, is a hazardous base on which to build positive theories of
business fluctuations and to undertake policy analysis’ (King 1993). However,
New-Keynesian economics today 473
even if Keynesians old and new agree on the IS-LM interpretation of
aggregate demand a further complication arises in connection with the recent
work of Greenwald and Stiglitz on financial market imperfections and
business cycles (see Greenwald and Stiglitz 1993a; 1993b; Stiglitz 1993). The
Greenwald and Stiglitz model shows how a negative aggregate demand shock
could translate itself into a leftward shift of the aggregate supply schedule due
to firms’ increased perception of risk during an economic downturn. Due to
financial market imperfections generated by asymmetric information equity
rationed firms can only partially diversify out the risks they face. Their
resultant dependence on debt rather than new equity issues to finance
investment makes firms more vulnerable to bankruptcy the higher the level of
their output. Hence any changes in a firm’s net worth position or in their
perception of the risks they face will have a negative effect on their
willingness to produce. Risk-averse firms will be less willing to supply at
every price when the environment becomes less favourable and increasingly
uncertain. When in an economic downturn firms observe a shift in their
demand curve they must either reduce their output or their price. Risk-averse
firms prefer to adjust their output because the ‘uncertainties associated with
changing prices may be much greater’ (Stiglitz 1993). In such a world the
Greenwald-Stiglitz model suggests that wage and price flexibility may well be
destabilizing and exacerbate any economic downturn. The important
implication is that the resultant risk-based aggregate supply curve will shift
leftwards following an economic downturn initiated by an aggregate demand
shock. This results in the non-neutrality of money even if prices are perfectly
flexible.
A NEW KEYNESIAN-MONETARIST SYNTHESIS?
Not all Keynesian economists are as convinced as Gregory Mankiw and other
new Keynesians that real progress has been made since the mid-1970s. Alan
Blinder has questioned whether the ‘prodigious amounts of labour and capital
devoted to macroeconomic research since 1972 have been allocated correctly’
and Olivier Blanchard has criticized the readiness of macroeconomists to
adopt the new classical ‘quasi-religious insistence on microfoundations’ which
has led to the construction of ‘too many monsters with few interesting results’
(see Blinder 1986; Blanchard 1992). Certainly the new Keynesian
developments have been criticized for their lack of attention to empirical
research, a criticism Mankiw accepts. However, given that the new classical
critique was launched mainly from a theoretical rather than an empirical
base it is perhaps understandable that younger Keynesians, at least initially,
have concentrated their efforts on providing ‘fort Keynes’ with more solid
microfoundations rather than giving continuing emphasis to empirical work
(see Snowdon et al. 1994: ch. 7).
So where does macroeconomics go from here? David Laidler, Olivier
Blanchard and Alan Blinder have made it abundantly clear that in their view
474 Brian Snowdon and Howard Vane
a monetarist augmented mainstream macroeconomics, circa 1972, although
not perfect, ‘had solid foundations and was basically right’ (see Blanchard
1992; Laidler 1992; Blinder 1992b). By the mid-1970s the impact of supply
shocks had been successfully incorporated into the mainstream model and as
a result it has, in their view, proved capable of withstanding the new classical
challenge. Thus the current debate is now mainly between a small but very
influential group of equilibrium business cycle theorists (Robert Barro’s ‘good
guys’) and a larger group of mainstream macroeconomists who adhere to
what could justifiably be called a new Keynesian—monetarist synthesis
(Robert Barro’s ‘bad guys’: see Barro 1989). However, as we have noted
above, an important implication of the work of Greenwald and Stiglitz is that
the traditional distinction between aggregate demand and aggregate supply
disturbances, although useful as an organizing principle, may be misleading.
In a similar vein Benjamin Friedman has recently argued that ‘Many
occurrences that initially seem to represent disturbances to aggregate supply
likewise cause disturbances to aggregate demand and vice versa’ (Friedman
1992). Indeed Greenwald and Stiglitz even suggest that their theory of risk-
averse firms if combined with market clearing flexible wages and prices can
be viewed as a special case of real business cycle theory. This requires that
the financial disorganization and risk associated with recessions can be
thought of as representing a form of negative shock to technology and capital.
Perhaps here there is some hope that this line of research could lead to some
future collaboration between the ‘good guys’ and the ‘bad guys’.
CONCLUDING REMARKS
The remarkable versatility of Keynesian economics guarantees that it will
continue to serve as a relevant research programme which will influence both
theoretical developments and policy proposals. It remains to be seen if
Gregory Mankiw’s optimism with respect to the usefulness of the burgeoning
new Keynesian microfoundations literature significantly improves our
understanding of macroeconomic phenomena. We share Mankiw’s optimism.
Keynesian economics, in resurrected or reincarnated form, is alive and well.
Recent controversies surely confirm the observation made by Sir Denis
Robertson (1954) many years ago when he noted that:
Highbrow opinion is like a hunted hare; if you stand in the same place, or
nearly the same place, it can be relied upon to come round to you in a circle.
So it’s back to the future!
NOTES
1 Taken from ‘Ends and Means in Economics,’ in Selected Papers on Economic Theory
(ed. E.Lindahl, London: Allen and Unwin, 1958).
New-Keynesian economics today 475
2 This interview was one in a series held in connection with the preparation of a book
published by Edward Elgar (see Snowdon, Vane and Wynarczyk 1994).
3 Cited in Mankiw (1992a).
REFERENCES
Akerlof, G.A. and Yellen, J.L. (1985) ‘A Near-Rational Model of the Business Cycle, with
Wage and Price Inertia’, Quarterly Journal of Economics, Supplement.
Akerlof, G.A. and Yellen, J.L. (1990) ‘The Fair Wage-Effort Hypothesis and
Unemployment’, Quarterly Journal of Economics May.
Ball, L. (1994) ‘What Determines the Sacrifice Ratio?’, In Monetary Policy (ed. N.G.
Mankiw), Chicago: University of Chicago Press.
Ball, L. and Mankiw, N.G. (1992a) ‘Asymmetric Price Adjustment and Economic
Fluctuations’, Harvard Institute of Economic Research, Discussion Paper no.
1602, July.
Ball, L. and Mankiw, N.G. (1992b) ‘Relative-Price Change As Aggregate Supply Shocks’,
National Bureau of Economic Research, Working Paper no. 4168, September.
Ball, L. and Romer, D. (1990) ‘Real Rigidities and the Non-Neutrality of Money’, Review
of Economic Studies April.
Ball, L. and Romer, D. (1991) ‘Sticky Prices as Coordination Failure’, American Economic
Review June.
Ball, L., Mankiw, N.G. and Romer, D. (1988) ‘The New Keynesian Economics and the
Output-Inflation Trade-Off’, Brookings Papers on Economic Activity.
Barro, R.J. (1979) ‘Second Thoughts on Keynesian Economies’, American Economic
Review May.
Barro, R.J. (1986) ‘Recent Developments in the Theory of Rules Versus Discretion’,
Economic Journal Supplement.
Barro, R.J. (1989) ‘New Classicals and Keynesians, or the Good Guys and the Bad Guys’
, Schwiez Zeitschrift für Volkswirtschaft und Statistik.
Blanchard, O.J. (1992) ‘For a Return to Pragmatism’, in The Business Cycle: Theories
and Evidence (eds M.Belongia and M.Garflnkel), London: Kluwer Academic.
Blinder, A.S. (1986) ‘Keynes After Lucas’, Eastern Economic Journal July/September.
Blinder, A.S. (1988) ‘The Fall and Rise of Keynesian Economies’, Economic Record
December.
Blinder, A.S. (1991) ‘Why Are Prices Sticky? Preliminary Results from an Interview
Study’, American Economic Review May.
Blinder, A.S. (1992a) ‘A Keynesian Restoration is Here’, Challenge, September/ October.
Blinder, A.S. (1992b) ‘Déjà Vu All Over Again’, in The Business Cycle: Theories and
Evidence (eds M.Belongia and M.Garflnkel), London: Kluwer Academic.
Blinder, A.S. and Mankiw, N.G. (1984) ‘Aggregation and Stabilization Policies in a Multi-
Contract Economy’, Journal of Monetary Economics January.
Campbell, J.Y. and Mankiw, N.G. (1989) ‘International Evidence on the Persistence of
Economic Fluctuations’, Journal of Monetary Economics March.
Cooper, R. and John, A. (1988) ‘Coordinating Coordination Failures in Keynesian
Models’, Quarterly Journal of Economics August.
Diamond, P.A. (1982) ‘Aggregate Demand Management in Search Equilibrium’, Journal
of Political Economy October.
Fischer, S. (1977) ‘Long-Term Contracts, Rational Expectations, and the Optimal Money
Supply Rule’, Journal of Political Economy February.
Fischer, S. (1988) ‘Recent Developments in Macroeconomics’, Economic Journal June.
Friedman, B. (1992) ‘How Does It Matter?’, in The Business Cycle: Theories and Evidence
(eds M.Belongia and M.Garfinkel), London: Kluwer Academic.
Friedman, M. (1968) ‘The Role of Monetary Policy’, American Economic Review March.
476 Brian Snowdon and Howard Vane
Gordon, R.J. (1990) ‘What Is New-Keynesian Economics?’, Journal of Economic
Literature September.
Greenwald, B. and Stiglitz, J. (1993a) ‘New and Old Keynesians’, Journal of Economic
Perspectives Winter.
Greenwald, B. and Stiglitz, J. (1993b) ‘Financial Markets, Imperfections and Business
Cycles’, Quarterly Journal of Economics February.
Hoover, K.D. (ed.) (1992) The New Classical Macroeconomics, Aldershot: Edward Elgar.
Howitt, P.W. (1990) The Keynesian Recovery, Oxford: Philip Allan. King, R.G. (1993)
‘Will the New Keynesian Macroeconomics Resurrect the IS-LM Model?’, Journal of
Economic Perspectives Winter.
Kydland, F.E. and Prescott, E.C. (1977) ‘Rules Rather than Discretion: The Inconsistency
of Optimal Plans’, Journal of Political Economy June.
Laidler, D.E.W. (1992) ‘The Cycle Before New Classical Economies’, in The Business
Cycles: Theories and Evidence (eds M.Belongia and M.Garfinkel), London: Kluwer
Academic.
Laing, D. (1993) ‘A Signalling Theory of Nominal Wage Inflexibility’, Economic Journal
November.
Leijonhufvud, A. (1981) Information and Coordination: Essays in Macroeconomic
Theory, Oxford: Oxford University Press.
Leslie, D. (1993) Advanced Macroeconomics Beyond IS-LM, Maidenhead: McGrawHill.
Lindbeck, A. and Snower, D. (1986) ‘Wage Setting, Unemployment, and Insider-Outsider
Relations’, American Economic Review May.
Lucas, R.E. Jr (1972) ‘Expectations and the Neutrality of Money’, Journal of Economic
Theory April.
Lucas, R.E. Jr (1973) ‘Some International Evidence on Output-Inflation Tradeoffs’,
American Economic Review June.
Lucas, R.E. Jr (1978) ‘Unemployment Policy’, American Economic Review May.
Lucas, R.E. Jr and Sargent, T.J. (1978) ‘After Keynesian Macroeconomics’, in After the
Phillips Curve: Persistence of High Inflation and High Unemployment, Boston; Federal
Reserve Bank of Boston.
Mankiw, N.G. (1985) ‘Small Menu Costs and Large Business Cycles: A Macroeconomic
Model of Monopoly’, Quarterly Journal of Economics May.
Mankiw, N.G. (1986a) ‘Issues in Keynesian Macroeconomics: A Review Essay’, Journal
of Monetary Economics September.
Mankiw, N.G. (1986b) ‘The Allocation of Credit and Financial Collapse’, Quarterly
Journal of Economics August.
Mankiw, N.G. (1987) ‘Comment’ on J.J.Rotemberg, ‘The New Keynesian
Microfoundations’, NBER Macroeconomics Annual.
Mankiw, N.G. (1988a) ‘Recent Developments in Macroeconomics: A Very Quick Refresher
Course’, Journal of Money, Credit and Banking August.
Mankiw, N.G. (1988b) ‘Imperfect Competition and the Keynesian Cross’, Economic
Letters.
Mankiw, N.G. (1989) ‘Real Business Cycles: A New Keynesian Perspective’, Journal of
Economic Perspectives Summer.
Mankiw, N.G. (1990) ‘A Quick Refresher Course in Macroeconomics’, Journal of
Economic Literature December.
Mankiw, N.G. (1991) ‘Comment on J.J.Rotemberg and M.Woodford “Markups and
the Business Cycle”’, NBER Macroeconomics Annual.
Mankiw, N.G. (1992a) ‘The Reincarnation of Keynesian Economies’, European Economic
Review April.
Mankiw, N.G. (1992b) Macroeconomics, New York: Worth.
Mankiw, N.G. (ed.) (1994) Monetary Policy, Chicago: University of Chicago Press.
Mankiw, N.G. and Romer, D. (eds) (1991) New Keynesian Economics, 2 vols, Cambridge,
MA: MIT Press.
New-Keynesian economics today 477
Mankiw, N.G. and Summers, L.H. (1986) ‘Money Demand and the Effect of Fiscal
Policies’, Journal of Money, Credit and Banking November.
Mankiw, N.G., Rotemberg, J.J. and Summers, L.H. (1985) ‘Intertemporal Substitution
in Macroeconomics’, Quarterly Journal of Economics February.
Mankiw, N.G., Miron, J.A. and Weil, D.N. (1987) ‘The Adjustment of Expectations to a
Change in Regime: A Study of the Founding of the Federal Reserve’, American
Economic Review June.
Parkin, M. (1992) ‘Where Do We Stand?’, in The Business Cycle: Theories and Evidence
(eds M.Belongia and M.Garflnkel), London: Kluwer Academic.
Phelps, E.S. (ed.) (1991) Recent Developments in Macroeconomics, Aldershot: Edward
Elgar.
Phelps, E.S. (1994) Structural Slumps: The Modern Equilibrium Theory of
Unemployment, Interest and Assets, Cambridge, MA: Harvard University Press.
Robertson, D.H. (1954) ‘Thoughts on Meeting Some Important Persons’, Quarterly
Journal of Economics May.
Romer, D. (1993) ‘The New Keynesian Synthesis’, Journal of Economic Perspectives
Winter.
Shapiro, C. and Stiglitz, J. (1984) ‘Equilibrium Unemployment as a Discipline Device’,
American Economic Review June.
Snowdon, B., Vane, H.R. and Wynarczyk, P. (1994) A Modern Guide to Macroeconomics:
An Introduction to Competing Schools of Thought, Aldershot and Vermont: Edward
Elgar.
Solow, R.M. (1990) The Labour Market as a Social Institution, Oxford: Basil Blackwell.
Stiglitz, J. (1993) Economics, New York: W.W.Norton.
Taylor, J.B. (1979) ‘Staggered Wage Setting in a Macro Model’, American Economic
Review May.
Tobin, J. (1977) ‘How Dead is Keynes?’, Economic Inquiry October.
Tobin, J. (1988) ‘Comment’ on David Romer’s paper on ‘What Are the Costs of Excessive
Deficits?’, NBER Macroeconomics Annual.
Tobin, J. (1993a) ‘Price Flexibility and Output Stability: An Old Keynesian View’, Journal
of Economic Perspectives Winter.
Tobin, J. (1993b) ‘Price Flexibility and the Stability of Full Employment Equilibrium’, in
Monetary Theory and Thought (eds H.Barkai, S.Fischer and N. Livitian), London:
Macmillan.
Yellen, J.L. (1984) ‘Efficiency Wage Models of Unemployment’, American Economic
Review May.
21 What is new-Keynesian economics?
Robert J.Gordon
Journal of Economic Literature (1990) 28, September, pp. 1115–71
1 INTRODUCTION
Background
In the late 1970s it appeared that the US macroeconomic landscape was being
swept by a new-classical tide, and that Keynesian economics had become an
isolated backwater. In fact there is still a widespread impression that the best
and brightest young macroeconomists almost uniformly marched under the
new-classical banner as the decade of the 1980s began.1 Yet it is now
apparent that the rumours of the death of Keynesian economics were greatly
exaggerated. Building on foundations laid in the late 1970s by Stanley Fischer
(1977a) and Edmund Phelps and John Taylor (1977), a large number of
authors, young and middle-aged alike, since the late 1970s have produced an
outpouring of research within the Keynesian tradition that attempts to build
the microeconomic foundations of wage and price stickiness. The adjective
new-Keynesian nicely juxtaposes this body of research with its arch-opposite,
the new-classical approach.2
This chapter extracts the essential elements of new-Keynesian economics
for an audience of professional economists who are not specialists in the
microeconomic foundations of macroeconomics. There is no intention to
survey comprehensively every notable paper in the field, but rather to sift the
literature for the most important ideas and themes. One commentator has
asserted that the new-Keynesian literature has provided too many
explanations of wage and price stickiness, and so we apply tough standards to
the major contributions, asking whether they make an essential contribution
to an understanding of the adjustment of wages and prices. In short, our intent
is to ask what is new and what is convincing in the large literature that
collectively has become known as the new Keynesian economics.
Main themes
Like its precursor a decade previously (R.Gordon 1981), this chapter differs
from conventional surveys not just in its intent to sift and criticize rather
What is new-Keynesian economics? 479
than to provide a broad and evenhanded overview. It also contains a
substantial empirical prologue before reaching the core material on new-
Keynesian theory. The prologue (sections 1 and 2) argues that there are
three different dimensions of price stickiness (which we will label the
inertia, rate-of-change, and level effects). A brief survey of the emerging
literature in the new empirical industrial organization, together with a new
empirical time-series investigation of price adjustment across time and
countries, reveals the essential fact that any satisfactory theory of price
adjustment must explain the variability of price adjustment parameters
across industries, across countries, and across historical intervals. We
ultimately reach the verdict that much of new-Keynesian theory does not
succeed in explaining these facts.
The prologue (Sections 2 and 3) is followed by the core of the chapter,
the critical review of theoretical contributions in the new-Keynesian
literature. The review is organized by recognizing two central distinctions,
the first between price setting in product markets and wage setting in labor
markets, and the second between nominal rigidity and real rigidity. The
theoretical analysis in the chapter is organized into a treatment of main
themes and issues (Section 4), and discussions of nominal rigidity in the
product market (Section 5), real rigidity in the product market (Section 6),
and models of labor market rigidity (Section 7), followed by a conclusion
(Section 8).
The task of new-Keynesian economics is to explain why changes in the
aggregate price level are sticky, that is, why price changes do not mimic
changes in nominal GNP. Sticky prices imply that real GNP is not an object
of choice by individual workers and firms but rather is cast adrift as a
residual. Thus new-Keynesian economics is about the choices of
monopolistically competitive firms that set their individual prices and accept
the level of real sales as a constraint, in contrast to new-classical economics
in which competitive price-taking firms make choices about output.
Why do changes in the aggregate price level fail to mimic changes in
nominal GNP? Two main themes emerge from the theoretical review, (1) the
reasons for the absence of nominal GNP indexation of individual prices,
and (2) the reasons why, in the absence of such indexation, individual prices
fail fully to reflect changes in nominal GNP. Underlying the first theme is an
essential element of any industrial economy—the role of idiosyncratic
elements of cost and demand. Firms care about the relation of their own
price to their own marginal cost. But because idiosyncratic shocks cause
their own costs and demand to evolve differently than nominal aggregate
demand, firms have no reason to accept the risk involved in indexing their
price to nominal aggregate demand. The absence of nominal GNP
indexation opens the way for theories of real rigidity to explain the sources
of nominal price stickiness.
The second theme is that, in the absence of nominal GNP indexation,
changes in individual prices will respond to changes in individual marginal
480 Robert J.Gordon
costs, not changes in nominal GNP. Thus the aggregate price level will be
sticky unless firms expect changes in their own marginal costs to mimic
changes in nominal GNP. Yet they have no such expectation. In the
framework that I label the input-output approach, each of thousands of
heterogeneous firms is enmeshed in a web of intricate supplier-demander
relationships. The input-output element helps to explain why firms do not
simply assume that marginal costs will move in parallel with aggregate
nominal demand: most firms do not know the identity of all of their suppliers,
their suppliers’ suppliers, and so on. The input-output approach places equal
emphasis on the purchase-material and labor-cost components of marginal
cost and points to models of real rigidities in the labor market, including the
efficiency wage and insider-outsider models, to help explain why prices are
less flexible in some industries than in others.
An important empirical finding in Section 3 is that prices were sticky not
just in the Great Depression and the postwar era, but long before World
War I. This fact casts doubt on institutional sources of price and wage
rigidity, for example, labor unions, and reinforces our emphasis on
universal features of microeconomic structure. In our treatment, price and
wage stickiness emerges from a core set of microeconomic elements that are
timeless and placeless: a technology of transactions, heterogeneity of goods
and factor inputs, imperfect competition, imperfect information, and
imperfect capital markets. Because these core elements remove any
incentive for individual agents to focus on nominal demand in making their
own price-setting decisions, their presence supports the traditional view that
Keynesian economics is fundamentally about the macroeconomic
externalities of individual decisions and the coordination failure inherent in
a free-market economy.
The dichotomy between supply and demand
With much ground to cover, there are many interesting topics in
macroeconomics that cannot be treated here. The coverage is limited to the
determinants of aggregate supply behavior, roughly, the division of a change
in nominal GNP growth between changes in prices and output, and the role of
wage stickiness (if any) in contributing to price stickiness. The entire demand
side of the economy is omitted as beyond the scope of the chapter. In
particular, we pay no attention to the reasons why aggregate demand
fluctuations exhibit positive serial correlation, nor to the respective role of
monetary and nonmonetary demand disturbances in causing these
fluctuations, nor to the significance of changes in the behavior of money
demand and velocity that have occurred in the 1980s, nor to the merits of
monetary rules, nor to the relative merits of monetary rules versus nominal
GNP rules. These topics on the demand side can be omitted, simply because
they are not at the heart of the conflict between new-Keynesian and new-
classical macroeconomics.
What is new-Keynesian economics? 481
Omission of the demand side from the scope of the chapter leads us to skip
over those contributions, sometimes classified as new-Keynesian, which
emphasize credit rationing as a source of fluctuations in commodity demand
and as a channel through which the influence of monetary policy is transmitted
(see Blanchard and Fischer 1989:478–88).3 We also omit any treatment of
feedbacks from changes in the parameters of aggregate price stickiness to the
variance of aggregate nominal demand (see the debate between Taylor 1986,
and DeLong and Summers 1986).4 We take as a precedent for imposing a
dichotomy between supply and demand, and for assuming nominal GNP to be
exogenous, Robert Lucas’ famous paper on the international output-inflation
trade-off (1973), which assumed that nominal GNP was an exogenous random
walk. In short, we are interested here in the price times output side of the
quantity equation (MV≡PQ), to the exclusion of the money times velocity side.
However, our focus here on nominal GNP rather than money helps to
clarify one source of frequent misunderstanding in this area. New-Keynesian
macroeconomics is not limited to the question ‘Why does money affect
output?’5 If prices are sticky, then any change in nominal GNP will affect real
output, no matter whether its source is a change in the nominal money supply
or some autonomous movement of spending on consumption, investment,
government purchases, or net exports. Further, nominal price stickiness opens
the way for supply shocks, for example, a change in the relative price of oil,
to create macroeconomic externalities that supplement the initial impact on
output of the shock by induced demand feedbacks. The microeconomic
theories surveyed in this chapter apply equally to the broad question as to
why demand disturbances in money and autonomous spending, as well as
supply shocks, cause changes in real output.6
2 THE THREE DIMENSIONS OF WAGE AND PRICE STICKINESS
Price stickiness in the presence of policy feedback
A prerequisite for any theory purporting to explain wage and/or price
stickiness is a demonstration that the phenomenon of stickiness exists in
realworld data. In this section we begin by defining three different dimensions
of price stickiness and distinguish between the essential role of price stickiness
and the peripheral role of wage stickiness. The exposition is carried out
insofar as possible with a set of identities, which clarify issues without
imposing any theory at all.
By definition, the log of nominal GNP (X) must be divided between the log
of the GNP deflator (P) and the log of real GNP (Q):
X≡P+Q. (1)
Reserving uppercase letters for logs of levels and lowercase letters for
percentage changes per unit of time, we take the time derivative of (1) and
obtain:
482 Robert J.Gordon
x≡p+q, (2)
which states that any change in nominal GNP must be divided between a
change in the aggregate price level and a change in real GNP. Next, we
subtract from both sides of (2) the long-run equilibrium or natural growth rate
of real GNP (q*), and use a ‘hat’ (ˆ) to designate variables defined net of that
trend growth rate of real output:
x-q*≡p+(q-q*);
x̂ ≡p+q̂. (3)
This states that an excess of nominal GNP growth over the long-run growth
rate of real output ( ) must be accompanied by some combination of inflation
(p) and a deviation of real output from that same long-run growth rate ( ).
In many recessions and depressions over the course of the industrial era
the economy has experienced a decline in output and employment that
appears to have constrained employees to work fewer hours than they
wished at the current real wage, and firms to produce less output than they
wished at the current price. These episodes admit the possibility that actual
output and long-run equilibrium output are two distinct concepts, implying
in turn that the way is open to consider the meaning of price stickiness. For
instance, if the rate of change of prices over the business cycle is always
equal to some constant fraction (α) of the excess nominal GNP movement,
then business-cycle movements in real output ( ) must soak up the
remaining fraction (1-α):
p=ax̂,
q̂=x̂-p=(1-α)x̂. (4)
One concludes from (4) that an economy with relatively sticky prices (a small
a) must exhibit correspondingly large fluctuations in real output, as long as
fluctuations in nominal demand ( ) are independent of the price stickiness
parameter α.
It is tempting to estimate a regression equation like either line of (4) to
determine the degree of price stickiness (α). But four crucial features of the
economy—level effects, inertia effects, policy feedback, and supply shocks—
are ignored in (4) and may invalidate any interpretation of an estimated
value of α as representing a structural price-stickiness coefficient. The first
problem is that (4) ignores level or Phillips-curve effects. It is possible for
actual output to be growing at its long-run equilibrium growth rate (i.e. =0)
while being off its equilibrium growth path, that is, when there is a gap
between the levels of actual and equilibrium output. The second problem is
the possible presence of price inertia, as occurs when lagged variables
(especially lagged inflation) enter into the determination of current inflation.
We defer the introduction of level and inertia effects until the next section in
What is new-Keynesian economics? 483
order to concentrate on the other two basic problems with (4), which concern
policy feedback and supply shocks.
The third problem is the possible presence of policy feedback from
inflation to excess nominal GNP growth, as would occur with a policy of
monetary accommodation to price changes. Such feedback would be implied
when the central bank attempts to peg or stabilize interest rates, or with a
real bills doctrine in which bank loans automatically expand to meet the
needs of trade. The fourth problem arises in the presence of autonomous
supply shocks which shift the rate of price change up and down relative to
that predicted by (4). We now consider a model in which the interaction
between policy feedback and supply shocks becomes crucial in estimating the
coefficient of price adjustment (α) in an equation like (4). The subsequent
results on coefficient bias apply to literally every empirical study that has
attempted to relate price or output change to such endogenous variables as
nominal or real GNP, the money supply, or unemployment.
Consider the two-equation model:
p=a +z
=θp+e, (5)
where z is the supply-shock term and e is the demand shock. The coefficient
of policy feedback (θ) would be positive if growth in the money supply
responds positively to a contemporaneous change in the inflation rate.
It is easy to see that in a world with no supply shocks (z=0), policy
accommodation makes no difference. Here we relegate the algebra to the
source note in Table 21.1 and consider a numerical example with a 10
percent positive realization of e, a price-adjustment parameter a=0.5, and a
policy accommodation parameter θ=1.0. Then (5) is satisfied with the values
p=10, =20, and =10. A regression of p on for a sample period with no
supply shocks will recover the correct value of α, 0.5=10/20. Despite policy
feedback, we would correctly infer that the smaller the price adjustment
coefficient, the larger the amplitude of output fluctuations in . Intuitively,
because in the absence of supply shocks price change depends only on
nominal demand ( ), and any policy feedback simply ‘blows up’ price and
nominal demand change by the same proportion.
We cannot, however, recover the correct value of α in the presence of
supply shocks. With a supply shock z=10 but no demand shock (e=0), and
with the same values of α and θ, (5) is satisfied for p= =20 and =0. If no ‘z’
variable is included to capture the supply-shock effect, a simple regression of
p on will recover an incorrect value of α=1. In general, as shown in the
notes to Table 21.1, a regression of p on in a sample containing both
demand and supply shocks will yield an upward biased estimate of the price-
adjustment parameter a, the larger is the accommodation parameter (θ) and
the larger is the variance of supply shocks relative to demand shocks (σ2z/
σ2e). The problem cannot be avoided by replacing nominal GNP change ( )
484 Robert J.Gordon
by real GNP change ( ) in the first equation in (5), because this would
introduce a negative bias that works in reverse and is larger, the smaller the
extent of policy accommodation.
Table 21.1 provides examples of the bias that will result in estimating the
price-stickiness coefficient (a), when excess nominal GNP growth ( ) or
excess real GNP growth ( ) are used as the alternative explanatory demand
growth variables. Columns 3 and 4 show that there is no bias in using with
any degree of policy feedback or any importance of supply shocks, as long as
both do not occur together. Using as the explanatory variable introduces a
downward bias when there are supply shocks, even if there is no policy
feedback. In intermediate situations, as on lines 4 and 5, estimates using
alternatively and bracket the true value. Using retrieves the correct
coefficient only when policy feedback is complete, that is, when policy fully
accommodates the supply shock, as in lines 6 and 7.
We reach five important conclusions from this analysis. First, to the extent
that demand shocks have been substantially more important than supply
shocks (at least prior to the oil-shock decade of the 1970s), the degree of price
stickiness can be measured by the coefficient on excess nominal GNP change
( ) in a regression equation explaining price change, even if policy has
partially or wholly accommodated price changes. Second, in view of the
Table 21.1 Range of estimated price stickiness coefficients when the true
coefficient is α=0.25
Sources by column: With reference to the model in equation (5) in the text,
the estimated coefficient E(α) in column (3) is the true coefficient (α) plus the
ratio of the covariance of with z divided by the variance of :
where r=σ2z/σ2e. When (= -p) is used as an alternative explanatory
variable in the first equation of (5), instead of , the equation estimated is
p=ß , and the estimated coefficient E(ß) is
The coefficient shown in column (4) is the value of a that would be calculated
on the assumption that ß is true, α=ß/(1+ß).
What is new-Keynesian economics? 485
widespread view of Milton Friedman and Anna Schwartz (1963) and most
commentators that monetary policy has been accommodative (i.e. procyclical
rather than countercyclical), the price-adjustment coefficient is more likely to
be upward than downward biased when the demand variable is , thus
overstating the extent of price flexibility and tilting the conclusions against
the new-Keynesian view that prices are sticky and toward the new-classical
view that prices are flexible. Third, in the presence of partial policy
accommodation, equations with nominal and real GNP changes ( and ) as
alternative demand variables will bracket the true coefficient of price
stickiness. Fourth, any empirical study of price adjustment should attempt to
find proxies for the supply shocks themselves, rather than allowing such
shocks to remain hidden in the error term, in order to minimize these biases
that occur in the presence of policy accommodation. Fifth, any study that
does not control for supply shocks is likely to reach unreliable conclusions
regarding the extent and/or secular change in price stickiness. For instance, a
conclusion that prices had become more sticky since World War II could be
subject to the criticism that prewar price adjustment coefficients are upward
biased because of some combination of (a) greater prewar policy
accommodation and (b) a higher prewar variance of unmeasured supply
shocks.
Where the Phillips curve and price inertia fit in
One reason different authors disagree on historical changes in the extent of
price stickiness is that authors have focused on different dimensions of
stickiness. Thus far we have characterized price stickiness by a single
parameter (α), which denotes the marginal response of the rate of price
change to a change in the excess growth rate of nominal GNP. Yet this
relation between the change in prices and the change in demand stands in
contrast to the relation between the change in prices and the level of demand,
that is, the Phillips curve, that may come first to mind in connection with
price adjustment. While the Phillips curve was originally developed (Phillips
1958) as an association between the change in nominal wage rates and the
level of unemployment, it has become common to use the Phillips-curve
terminology to label any relation between the rate of change of nominal
prices or wages and the level of a utilization variable like the unemployment
rate or detrended output. Here we focus on detrended output rather than
unemployment and, because our interest is primarily in price rather than
wage stickiness, we write a Phillips-curve relation for price change:7
(6)
where is the log ratio of actual to natural output, and we indicate
explicitly the time subscript that previously has been suppressed. The
supply-shock term from (5) is included here in each subsequent price
adjustment equation, in view of our previous conclusion that adjustment
486 Robert J.Gordon
coefficients will be biased unless a careful attempt is made to control for
supply shocks.
A third dimension of price stickiness is serial correlation, sometimes simply
called inertia. A frequent specification of the postwar US inflation process
combines the Phillips curve and inertia.8
(7)
When the lagged inflation term is interpreted as a proxy for the expected rate
of inflation then (7) is called an expectational Phillips curve. Friedman’s
(1968) natural rate hypothesis (NRH) states that the coefficient on in an
expectational Phillips curve is unity,
(8)
This expression is compatible with steady fully anticipated inflation when
actual and natural output are equal and implies that inflation
steadily accelerates whenever the log output ratio is positive.
But, as originally pointed out by Sargent (1971), the NRH does not
imply that the coefficient λ in (7) must be unity. The coefficient on in (8)
could be unity, while at the same time rational agents could form their
expectations of inflation by applying a coefficient λ below unity to lagged
inflation, if this provided the best possible predictor. For instance, if
inflation were a random walk the optimal predictor would be λ=1, but if
inflation were white noise, the optimal predictor would be λ=0. By
expressing the Phillips curve in form (7) rather than (8), we recognize that
the coefficient λ may vary in different times and places, depending on the
nature of the inflation process. Further, (7) recognizes, as (8) does not, that
there may be many reasons for serial dependence in the inflation rate, of
which expectation formation is only one, and overlapping wage and price
contracts may be among the others.
Blanchard (1987b) has stressed that there are two dimensions of price
adjustment, corresponding to the two parameters λ and γ in (7).9 An
equation like (7) implies that shocks to nominal aggregate demand cause
the economy to travel through loops on a diagram plotting inflation (p)
against the output ratio , and an economy with low values of λ and γ has
‘fat loops’; that is, it exhibits relatively large output fluctuations and only a
slow incorporation of the change in nominal demand growth into the rate of
inflation.
However, in addition to the two adjustment parameters in (7), we have
already introduced a third parameter (a) in (4) and (5), which measures the
fraction of current excess nominal GNP change ( t) taking the form of price
change. How are these parameters related? The connection when we add the
explanatory variable contained in (4) to those already present in (7):
(9)
What is new-Keynesian economics? 487
While the α t term may appear to drop from the sky, in fact equation (9)
can be interpreted simply as loosening the artificial restriction in (7) that
allows only the current value of the log output ratio to enter. The more
general form (9) allows both the current and one lagged value of the output
ratio to enter as explanatory variables, as becomes transparent when we use
the identity that to rewrite (9) in either of two
equivalent forms:10
(10a)
(10b)
Note that either (10a) or (10b) reduces to (7) when the a parameter is set
equal to zero.11 If both the current and one lagged output term matter for the
rate of price change, as in (10a), this implies in (10b) that the rate of change
of prices is related to both the rate of change ( t) and the level of output.
The generalization of the Phillips-curve hypothesis contained in (9) and (10)
illustrates that the same hypothesis of price adjustment can be expressed in
several alternative forms, and that the extent of price change in response to a
change in nominal demand depends not on a single parameter, but on the
three parameters λ, α, and γ.12
Where wages fit in
Keynesian economics has traditionally been more concerned with wage
rigidity than price rigidity. Yet our discussion to this point has made no
mention of wages. This is fitting, because only price stickiness, not wage
stickiness, is a necessary condition for business cycles in real output, given a
particular path of nominal aggregate demand. There are no arithmetically
necessary implications of nominal wage rigidity for the cyclical behavior of
output or employment, because sufficient flexibility in profits could allow
prices to be flexible (so that p mimics ), even if the nominal wage rate
were absolutely fixed. Yet a world of highly flexible profits with completely
rigid wages would have economic, if not arithmetic, implications. High
profit volatility for any given firm would shift the firm’s securities out along
the mean-variance schedule and raise the average cost of capital, thus
creating pressure in two directions, toward an increase in the flexibility of
wages and toward a decrease in the flexibility of prices, both of which
would reduce the volatility of profits. In new-Keynesian economics there is
no primacy to wage rigidity as contrasted with price rigidity, and thus no
presumption that wages are less cyclically sensitive than prices. In fact,
much of the research since 1985 has been directed toward the
microfoundations of price rigidity.
or,
488 Robert J.Gordon
The nature of cyclical flexibility in real wages has always played a role in
discussions of Keynesian economics, dating back to the debate involving John
Dunlop (1938), Lorie Tarshis (1938), and John Maynard Keynes (1939). Even
though these authors are known for the criticism of the countercyclical real
wage assumption implicit in the General Theory, resulting from its
assumption of price flexibility combined with nominal wage rigidity, it is less
well known that Tarshis in 1939 soon recanted and provided evidence of a
relatively strong negative correlation between average hourly earnings and
total hours worked.13 Subsequently we shall examine new evidence on the
cyclically of real wages.
Rate of change or hysteresis effects
Equations (9) and (10) imply that there may be three quite different types
of price stickiness, indicated respectively by a relatively high value of the
λ parameter, and by relatively low values of the α and γ adjustment
parameters. The role of the inertia parameter λ is straightforward, with a
higher value of λ prolonging the duration of adjustment to changes in
nominal demand, for any given values of the α and γ parameters, and
increasing the importance of overshooting and dynamic adjustment loops.
The distinction between rate-of-change adjustment (α) and level or
Phillips-curve adjustment (γ) is clarified by examining extreme cases in
which one or the other is absent. When there is no rate-of-change effect
(α=0) we are back in the simple Phillips-curve framework in which only
the level of output matters. For any given values of the λ and γ
parameters, the acceleration of inflation implied by an output ratio of +5
percent is the same, regardless of whether the output ratio is rapidly rising
or rapidly falling.
The opposite extreme is of more interest, because it has been the focus of
so much attention in the context of high European unemployment in the
1980s. An economy lacking a level effect (γ=0) is said to be characterized by
hysteresis. Considerable theoretical work has emerged to explain hysteresis
phenomena, particularly in the context of the insider-outsider model of
employment reviewed in Section 7 (pp. 524–5). Whatever the theoretical
explanation, the presence of hysteresis would have profound implications for
both economic doctrine and policy.14 Friedman’s NRH posits a self-correction
or level effect that automatically stabilizes output at its equilibrium value in
the presence of steady nominal demand growth. With no level effect, the
economy could settle down at any arbitrary distance from its equilibrium
output path (with t=0) and experience a constant rate of inflation, with no
tendency for self-correction. And, if the NRH were abandoned, it would cast
stabilization policy adrift from its previous mooring, the task of steering the
economy toward a fixed natural rate , and open to the central
implication of hysteresis that any level of detrended output or rate of
unemployment, no matter how low or high, would be consistent with steady
What is new-Keynesian economics? 489
inflation (at a rate that depends on the history of both inflation and
unemployment).
As we see below, the pattern of price adjustment described by hysteresis is
not a novel phenomenon isolated to Europe in the 1980s, for the Phillips
curve or ‘level effect’ also vanished in the United States, the United Kingdom,
and Germany during the interwar period.15 A key implication of (9) is that
with hysteresis (γ=0) and with λ=1-α, the acceleration or deceleration of
inflation, as well as the change in detrended output, depends only on the
difference between t and pt-1, that is, whether or not excess nominal GNP
growth ratifies the inherited inflation rate:
p
t
-p
t-1
=α(
t
-p
t-1
)+z
t
, and (11)
(12)
In short, hysteresis implies that changes in both inflation and output are
completely independent of the level of detrended output, and that an
economy in the depths of a great depression can experience an acceleration
of inflation, no matter how high the level of unemployment or low the level
of detrended output, if excess nominal GNP growth exceeds last period’s
inflation rate.
Empirical estimates of the general price-adjustment model in (9) and (10)
can reveal the size of the three adjustment parameters (λ, α, and γ) in
different countries and historical eras. There remains the issue of which
alternative specification in (9) and (10) is preferable for estimation. As
argued earlier, in the presence of policy feedback and unmeasured supply
shocks, the a adjustment parameter is likely to be overstated when nominal
GNP change ( ) is included as in (9) and understated when real GNP
change ( ) is used instead as in (10b). This suggests that estimates based
alternatively on both forms are preferable, because they will ‘bracket’ the
true parameter.
We conclude from this discussion that three parameters are required to
measure the degree of price stickiness: λ measuring the extent of inertia, a
measuring the rate-of-change or hysteresis effect, and γ measuring the level or
Phillips-curve effect. Any attempt to measure changes in the degree of
stickiness over time, or differences among countries, may be flawed if it omits
any of these three parameters from empirical testing. Further, we have seen
that policy accommodation of supply shocks can bias coefficients of price
adjustment, and thus any adequate empirical investigation must make a
careful attempt to control for supply shocks as well.
Section 2 began by stressing the most obvious implication of the identity
linking changes in nominal demand, real output, and the price level. Changes
in the price level must exactly mimic changes in nominal demand if business
cycles in real output are to be avoided. Thus the requirements for perfect
price flexibility are highly restrictive: in the context of equation (10) price
changes can mimic changes in nominal demand only if α=1, λ=0, and γ=0.
490 Robert J.Gordon
Thus any combination of a rate-of-change coefficient below unity, the
presence of Phillips-curve level effects (γ>0), or the presence of inertia effects
is sufficient to generate business cycles. However, Lucas (1973) showed that
Phillips-curve level effects could be derived in a business-cycle model in
which markets clear; thus the absence of perfect price flexibility is not
sufficient to distinguish between new-classical market-clearing models and
new-Keynesian sticky-price models. Instead, the presence of price inertia (θ>0)
is crucial for rejecting the new-classical interpretation and demonstrating the
existence of price stickiness.16
3 THE VARIETY OF HISTORICAL EXPERIENCE
Diversity of response across industries
Since well before the publication of Keynes’ General Theory, for example,
Mills (1927), industrial economists have been aware that the responsiveness
of prices to changes in demand differs sharply across industries. The contrast
between the flexibility of the prices of agricultural products, and the
inflexibility of the prices of complex manufactured goods, was the point of
departure of Gardiner Means’ (1935) administered price hypothesis. In the
Great Depression every farmer knew what Table 21.2 shows.
In an economic downturn the farmer was the victim of a highly unfavorable
twist in relative prices, because the prices of agricultural products fell much
more than those of many manufactured goods, especially the agricultural
implements listed on the first line that represent one of the main purchased
inputs in the farm sector. Within the spectrum of manufactured goods, crude
products like textiles tended to exhibit more price flexibility than more
finished products like tractors and automobiles.
Unfortunately, there are few empirical studies that document these
differences systematically. Stigler and Kindahl (1970) collected prices from
buyers for a large number of products, and these data were analyzed by
Carlton (1986) to determine if there were any structural relations between
Table 21.2 Decline in Price and Production, various US industries,
1929–33
Source: Gardiner Means (1935:8)
What is new-Keynesian economics? 491
seller and buyer characteristics and the degree of price rigidity. Carlton
concludes that there is a significant degree of price rigidity: ‘It is not unusual
in some industries for prices to individual buyers to remain unchanged for
several years’ (1986:638). Unfortunately, however, neither Stigler and Kindahl
nor Carlton show that, as Means suggested, the degree of ‘complexity’ of a
product is related to price rigidity. Although Carlton did try to measure
complexity as well as other structural variables, he was able to find a
significant positive correlation only between the concentration ratio for a
product and the duration of its price rigidity (i.e. the number of months a
price remains unchanged).17
But it is important to stress another of Carlton’s findings that may be of
substantial importance in assessing the theories reviewed below. By no
means are all prices rigid or do they remain unchanged for substantial
periods of time: ‘The fixed costs of changing price at least to some buyers
may be small. There are plenty of instances where small price changes
occur’ (Carlton 1986:638). Specifically, ‘there are a significant number of
price changes that one would consider small (i.e., less than 1 percent) for
most commodities and transaction types.’ Industries where frequent price
changes are common include plywood and nonferrous metals, and
commodities with relatively long spells of rigid prices include steel, paper,
chemicals, cement, and glass. Carlton’s evidence that spells of price rigidity
can be both short and long calls into question the generality of the oft-cited
study by Cecchetti (1986) which provides evidence that newsstand prices of
magazines can remain unchanged for years (see also Kashyap 1990).
Carlton’s finding that spells are sometimes short and price changes
sometimes small would appear to call into question the theories of new-
Keynesian economists based on ‘menu costs’ of price changes, reviewed in
Section 5 (pp. 511–14). However, this apparent implication is subject to the
caveat that if demand and/or supply shift permanently, then small price
adjustments can produce large benefits and will be observed even if fixed
costs are large (Carlton 1989a: 932).
There has been remarkably little interaction between new-Keynesian
theory and the evidence provided in the emerging literature of the new
empirical industrial organization (NEIO) surveyed by Bresnahan (1989). The
overall conclusion is that there is ‘a great deal of market power, in the sense
of price-cost margins, in some concentrated industries’ (Bresnahan
1989:1052). One could emphasize the words a great deal as supporting the
emphasis by new-Keynesian theorists on models of monopolistic rather than
perfect competition. Or one could emphasize the word ‘some’ to point out
that the world is made up of both monopolistic and competitive industries.
But the matter is even more complex: one important theme of recent NEIO
work is that pricing behavior can alternate between collusive monopolistic
behavior and price wars in which a cartel temporarily collapses, implying
that a given industry is characterized neither by exclusively monopolistic nor
competitive behavior.
492 Robert J.Gordon
The theme of heterogeneity extends along other dimensions. Product
differentiation is so pervasive that ‘there is almost no industry for which
the position that there are more than 100 products is untenable: without
putting more structure on the problem, the analyst could need to estimate
literally thousands of elasticities’ (Bresnahan 1989:1045).18 Heterogeneity
extends to pricing behavior across firms in a single industry. For instance,
a study of airline competition found not only that concentration affects
price in airline city-pair markets, but also that the identity of the
competitors matters. Carlton stresses that a given seller can charge
different prices to different customers and change them at different times,
based on ‘a seller’s knowledge of his customers and on the optimality of
non-price rationing’ (Carlton 1989a). In a cross-section of industries,
numerous dimensions of structure appear to vary together, including mass
pr oduction, large-scale facilities, unionization, capital intensity,
concentration, and cyclical price rigidity, all of which are more
pronounced in the cyclically sensitive sectors of the economy, particularly
durable goods.19
As we shall see below, new-Keynesian theory has contributed relatively
little to understanding these differences across industries, and as yet there has
been virtually no research that attempts to test theories on a diversity of
industrial data. We emphasize the numerous aspects of heterogeneity across
and within industries to support several themes that emerge below, including
the importance of idiosyncratic elements of product cost and demand that
prevent firms from assuming, as in so many simple models, that their costs
and product demand will mimic the behavior of nominal aggregate demand.
Even so basic a distinction as Arthur Okun’s (1975, 1981) dichotomy between
auction and customer markets rarely surfaces in new-Keynesian writing,
much less in new-classical contributions. And the seminal work in
understanding the coexistence of auction and customer markets has been
contributed by microeconomists, especially Carlton (e.g. 1989b), who stresses
that, because of the high costs of establishing auction markets, ‘there is no
incentive for the efficient creation of markets’ (P. 7).20
Diversity of response across time and space
Just as challenging for theorists as the diversity of responses across
industries at a particular time in a particular country is the diversity of
responses across time and countries. Much of the empirical work in this
area has been within the context of a debate over whether prices, wages, or
both have become less flexible in the postwar USA as contrasted with
various periods before the Great Depression (among these studies are Allen
1989; Mitchell 1985; Sachs 1980; Schultze 1981, 1986; Taylor 1986;
R.Gordon 1980, 1982b). In related work Charles Schultze and others have
examined differences in response coefficients over both time and space for
the USA and several other major industrialized nations (see Alogoskoufis
What is new-Keynesian economics? 493
and Smith 1989; Backus and Kehoe 1988; Coe 1989; Schultze 1981, 1986;
R.Gordon 1982a, 1983).
It is beyond the scope of this chapter to track all the differences in data
and specification that contribute to the variety of conclusions that these
studies have reached; that would require a separate survey on this issue
alone. Some of the disagreements, particularly about changes in cyclical
behavior for the USA, arise because authors often do not recognize that
there are three dimensions to price and wage rigidity, as demonstrated in
Section 2. These are the degree of inertia or serial correlation (λ), the rate-
of-change or hysteresis coefficient (α), and the level or Phillips-curve
coefficient (γ). Here we provide a link between that classification scheme
and historical data by presenting estimates of the three parameters based
on price-adjustment equations (9) and (10) developed above. We address
two issues, differences in the responsiveness of prices and wages over US
history, and differences in the responsiveness of prices over the period
since 1870 for five major industrial nations (USA, UK, France, Germany
and Japan).
The empirical equations summarized in this chapter are estimated only
for nominal and real output data corresponding to the t and variables
in the theoretical price-adjustment equation (9). There is no attempt to
estimate alternative versions for other possible nominal and real demand
variables, for example, the money supply or unemployment. Annual output
data extend back much further than unemployment data—to 1855 for the
UK, 1870 for France, Germany, and the USA, and 1885 for Japan. Wage-
adjustment equations are illustrated only for the USA, pending a careful
study to determine whether wage data for other countries are consistent over
time.21
Numerous decisions must be made in the development of tests covering
such a long span of history for these nations. These include the method of
detrending and the development of proxy variables for the major supply
shocks, a critical issue in view of the likely bias in coefficient estimates
when supply shocks are left unmeasured. Another issue is the estimation of
parameter shifts over subintervals of a long historical sample period.
Details on the methodology and the regression estimates are provided in
Appendix A, which shows that it is desirable to conduct the estimation with
slightly transformed versions of (9) and (10b). This allows us to proceed
directly to Tables 21.3 and 21.4, where the underlying parameters are
unscrambled from the transformed equations and presented for different
countries and historical eras.
The estimated parameters are provided for changes in prices, nominal
wages, and real wages for the USA in Table 21.3. We are interested in the
nature of changes in the three price- and wage-adjustment parameters over
time, and also evidence on the hotly debated issue of the cyclical sensitivity of
real wages. Following our analysis in Section 2, two estimates of each
parameter are provided. The left-hand element in each column is based on an
494 Robert J.Gordon
adjustment equation in which excess nominal GNP growth is included, and
this is likely to yield an upward biased value of the rate-of-change parameter
(α) in the presence of supply shocks and policy feedback. The right-hand
element in each column replaces excess nominal GNP growth with excess real
GNP growth, and this will tend to yield a downward-biased estimate of α.
The two estimates should bracket the true value. The parameters listed in
Tables 21.3 and 21.4 are allowed to change across time periods and are
recorded when parameter-shift coefficients are statistically significant, and
identical parameters across time periods indicate that such shift coefficients
are insignificant (for details and the significance of the shift coefficients
themselves, see Tables 21.5 and 21.6 in Appendix A).
The single most striking finding in Table 21.3 is that neither prices nor
wages were more sticky in 1954–87 than 1873–1914, as measured by the
rate-of-change (α) and level (γ) coefficients.22 The sole change between pre-
World War I and post-World War II was an increase in the inertia (λ)
coefficient, and this increase was much greater for prices than wages.
Between 1915 and 1953, however, there were substantial changes. The a
parameter rose substantially during World War I, while the α parameter
virtually disappeared during 1930–53. When the estimated price-change
parameters are subtracted from the wage-change parameters, the results before
1930 and after 1953 suggest that real wages have a negligible rate-of-change
Table 21.3 Estimated price and wage adjustment parameters for the USA,
1873–1987
Notes: Equation specifications and details are provided in Appendix A. The left parameter
in each column comes from unscrambling the coefficients of equation (9�) in Appendix A,
the version containing excess nominal GNP growth (
t
) as the rate-of-change variable; the
right parameter in parentheses () comes from unscrambling the coefficients of equation (9�),
the version containing excess real GNP growth (
t
) as the rate-of-change variable.
What is new-Keynesian economics? 495
effect but a substantial procyclical level effect. That is, a persistent economic
boom causes steady upward pressure on the real wage, and a persistent
recession does the reverse. However, this finding is subject to the qualification
that the manufacturing wage data used here exaggerate the cyclical
sensitivity of economy-wide rates. When the equations are reestimated for the
postwar 1954–87 period alone with the fixed-weight nonfarm wage index
replacing the manufacturing wage index, the cyclical sensitivity of real wages
drops to zero.23
Table 21.4 compares the results for US prices with similar price equations
for the other countries. Again the most striking finding is that the α and γ
parameters were the same before World War I and after World War II in the
UK, France, and Japan, with a decline in the a coefficient only in Germany
(and one may question the linking of German data over periods when its
borders were so different). In every country but Japan the inertia effect was
much higher after World War II than before World War I. The UK and Japan
Notes: Sample period for the UK begins in 1958, for Japan begins in 1888, and for
the USA ends in 1987. Equation specifications and details are provided in Appendix
A. The left parameter in each column comes from unscrambling the coefficients of
equation (9�) in the appendix, the version containing excess nominal GNP growth
(
t
) as the rate-of-change variable; the right parameter in parentheses () comes from
unscrambling the coefficients of equation (9�), the version containing excess real GNP
growth (
t
) as the rate-of-change variable.
Table 21.4 Estimated price adjustment parameters for five countries, 1873–1986
496 Robert J.Gordon
duplicate the jump in the a coefficient already observed for the USA during
the 1915–22 interval, and both the UK and Germany exhibit a substantial
decline in the γ coefficient during the interwar period.
Implications
Is the aggregate price level highly flexible, mimicking changes in excess
nominal GNP growth? Or does the aggregate price level live a life of its own,
bearing little relation to excess nominal GNP growth and thus allowing those
nominal changes to create business cycles in real output? The conclusion from
Tables 21.3 and 21.4 is that both these statements are true. And many in-
between responses have been observed as well.
At one extreme is the very high rate-of-change coefficient for Japan
throughout, and for the USA and the UK during World War I and its
aftermath. Figure 21.1 plots the 1886–1914 data for Japan and shows how
closely price changes track excess nominal GNP changes. The figure also
exhibits cycles in the log output ratio that are small relative to the large
amplitude of nominal GNP changes. We have argued above that the best
Figure 21.1 Inflation, Adjusted Nominal GNP Growth, and the Output Ratio, Japan,
1886–1914
What is new-Keynesian economics? 497
estimates of the adjustment parameters are given by the average of the
two estimates shown in each column in Tables 21.3 and 21.4. On the basis
of these averages, it is quite apparent in Table 21.4 that the USA has the
smallest rate-of-change parameter (a) and Japan the largest, both before
World War I and after World War II, with the other countries arrayed in
between. The postwar USA also contrasts starkly with Japan in its strong
inertia effect. The top frame of Figure 21.2 shows how loose is the relation
between inflation and excess nominal GNP growth in the postwar USA,
and how large is the amplitude of output cycles relative to nominal GNP
growth cycles. Again basing conclusions on the average of the two figures
in each column, inertia effects in all countries were negligible before
World War II.
These results demonstrate the strong diversity of aggregate price-
adjustment behavior that has occurred across time and across countries. The
variety of historical responses of price changes to nominal demand changes
raises questions that new-Keynesian theorists have barely begun to address.
Perhaps the most widely noted empirical test thus far devised by
Figure 21.2 Inflation, Adjusted Nominal GNP Growth, and the Output Ratio, United
States, 1950–89
498 Robert J.Gordon
new-Keynesian economists (Ball et al. 1988) takes as its point of departure
Lucas’ (1973) demonstration that the Phillips curve becomes steeper with a
higher variance of the growth rate of nominal demand. Ball et al. show that
menu-cost theory supports the Lucas correlation but also makes the
additional prediction that increases in the mean growth rate of nominal
demand should steepen the Phillips curve, because with staggered price
setting an increase in the mean inflation rate increases the frequency of
price changes. Thus far, their empirical work in support of this theoretical
prediction has been subject to substantial criticism.24 In relation to our
empirical results of Tables 21.3 and 21.4, either the Lucas or the Ball et al.
approach can help to explain why prices became more flexible during
World War I but contribute little or nothing to an understanding of the other
main findings: the similar level and rate-of-change effects before World War
I and after World War II, despite the higher variance of nominal demand in
the earlier period; the disappearance of the level effect in the Depression
years; the emergence of inertia after World War II; and the differences in
price flexibility among the five countries.
Empirical research and the revival of Keynesian economics
Theories are often judged on their ability to explain time-series data on
aggregate variables. This is clearly evident in the interaction of events and
ideas since 1970. Theories have risen and fallen in acceptance in accord with
the correspondence of their predictions with the evolution of actual events in
the macroeconomy. To gain perspective on the development of new-Keynesian
economics, we need to understand what went wrong with the old-Keynesian
economics. Our emphasis here is the empirical failure of the Keynesian
paradigm of the 1960s, and the elements that contributed to the empirical
revival of the Keynesian approach in the 1980s. We concentrate on empirical
aspects of the contest between new-classical and new-Keynesian economics,
and we limit the scope of the chapter by omitting any theoretical critique of
either the Lucas imperfect-information (Mark I) approach or the real-business-
cycle (Mark II) variant of new-classical macroeconomics.25
In the 1960s Keynesian economists incorporated into their theoretical and
econometric models an exploitable negative long-run Phillips-curve trade-off
between inflation and unemployment. The acceleration of inflation after
1965, together with the positive correlation between inflation and
unemployment observed during much of the 1970s, caused the mid-1960s
Keynesian orthodoxy to unravel. In flowery language that amounted to a
simultaneous declaration of war and announcement of victory, Lucas and
Sargent (1978:49–50) described ‘the task which faces contemporary students
of the business cycle [as] that of sorting through the wreckage…of that
remarkable intellectual event called the Keynesian Revolution’.
It is not widely recognized that the empirical reconstruction of Keynesian
economics occurred prior to the wave of theoretical work that is now most
What is new-Keynesian economics? 499
commonly associated with the term new-Keynesian economics. Lucas and
Sargent were only partly right. Yes, the predictions of the late 1960s were
incorrect, but incorrect forecasts do not provide de facto proof that a
doctrine’s theoretical underpinnings are fundamentally flawed. The essential
element of Keynesian doctrine is non-market-clearing, which in turn
requires the gradual adjustment of prices. The 1960s version of the Phillips
relation combined three elements, (1) gradual price adjustment, (2) a long-
run trade-off, and (3) a closed-economy, demand-only approach with no role
for import prices or supply shocks. Yet only (1) is necessary to maintain the
essence of the Keynesian paradigm, non-market-clearing. The other two
elements, (2) and (3), were ephemeral empirical results, based mainly on the
15 or 20 years of US postwar data, that revealed more of the short time
horizon and closed-economy mentality of the first generation of econometric
model builders than any fundamental weakness of the non-market-clearing
approach.
The long-run trade-off result was abandoned within five years of
Friedman’s presidential address.26 This allowed the gradual-adjustment
property of the 1960s-style wage and price equations to be combined with
the longrun neutrality property advocated by Friedman. The effects of
supply shocks, including the relative prices of oil and imports, were
absorbed into the US Phillips-curve framework in my work of the mid-
1970s, which was developed alongside the work by David Laidler, Michael
Parkin, and their collaborators in the open-economy setting of the UK.27 The
result was an econometric analogy to the dynamic aggregate demand and
supply model that was introduced with the 1978–9 publication of a new
generation of economic principles and intermediate macro textbooks.28 Now
a single reduced-form econometric equation for price change, like those
summarized in Tables 21.3 and 21.4 could incorporate the effects of gradual
adjustment, of demand shocks that created a temporary positive correlation
between inflation and output, and of supply shocks that created a temporary
negative correlation. By the end of the 1970s the supply side of the economy
had been opened up to outside influences, and the list of relevant supply
shocks for the USA had grown to include not only price controls and oil
shocks, but also changes in non-oil import prices, exchange rates, tax rates,
and the minimum wage.29
This so-called gradual-adjustment price-change equation is completely
non-structural and as such is in principle highly vulnerable to the Lucas
(1976) critique. We have seen in Tables 21.3 and 21.4 that coefficients of
price adjustment are subject to substantial change when there are major
changes in the economic environment, as in World War I or the Great
Depression. The sharp US disinflation of the 1980s posed a formidable
challenge which the empirical price-adjustment equation could have failed
but did not. A central implication of the resuscitated 1980-vintage empirical
Phillips curves, the value of the sacrifice ratio of lost output required to
achieve a permanent deceleration of inflation, turned out to be surprisingly
500 Robert J.Gordon
close to predictions made in advance. This suggests that, at least in the USA,
the substantial changes in price-adjustment parameters observed in Tables
21.3 and 21.4 to have occurred in previous historical eras have been largely
absent in the postwar US setting.30
The empirical stability and predictive success of the resuscitated US
Phillips curve is highly ironic in view of the inflammatory language used by
Lucas and Sargent. If anything lay smoldering in ‘wreckage’ in the mid-
1980s, it was the few abortive attempts to estimate price equations within
the framework of Mark I new-classical macroeconomics, particularly those
by Barro (1977a, 1978; Barro and Rush 1980). So strongly was price inertia
embedded in the US data that Barro could explain price movements only by
entering a distributed lag of between four and six years of monetary
surprises that themselves lasted only a single quarter. Why agents should be
reacting with a four-year lag to a one-quarter monetary surprise was never
explained. The attraction for the economics profession of the empirical
versions of Mark I new-classical macro, like the theoretical versions, was
undermined by the discrepancy between the time lags involved in data
dissemination, measured in days or weeks, as contrasted to the lags of price
changes in response to nominal demand shocks, measured in years or half-
decades.
4 NEW-KEYNESIAN THEORY: COMMON FEATURES
Essential features of Keynesian economics
The essential feature of Keynesian macroeconomics is the absence of
continuous market clearing. Thus a Keynesian model is by definition a non-
market-clearing model, one in which prices fail to adjust rapidly enough to
clear markets within some relatively short period of time. Common to almost
all Keynesian models is the prediction that in response to a decline in nominal
demand, the aggregate price level will decline less than proportionately over
a substantial time period, during which the actual price level is above the
equilibrium price level consistent with the maintenance of the initial
equilibrium level of real output. The fact that the price level is too high
means that the subequilibrium level of output actually produced is not chosen
voluntarily by firms and workers, but rather is imposed on them as a
constraint. It is the decline in nominal demand together with the absence of
full price adjustment that causes the economic system itself to impose the
constraint on each agent; nominal demand is insufficient to generate adequate
real sales at the actual price level. Each agent faces a constraint that is
indirectly a result of its own failure to reduce sufficiently its price and this
points to a coordination failure as a central ingredient in the description of
Keynesian price stickiness.
So many people now refer to new-classical models as equilibrium business-
cycle models that the word equilibrium has been co-opted as meaning the
What is new-Keynesian economics? 501
opposite of the term Keynesian. This leads some commentators to label an
approach that is the opposite of equilibrium economics as disequilibrium
economics. In one sense this is mere semantics; it does not matter whether we
describe the USA in 1932 or Europe in the mid-1980s as being in a state of
disequilibrium or low-employment equilibrium. However, the adjective
disequilibrium is not helpful, as it conveys ‘a failure of agents to realize
perceived gains from trade’ (to use Barro’s provocative 1979 phrase). Rather,
it is best to regard the core feature of Keynesian economics as the gradual
adjustment of prices and its corollary, that output and employment are not
choice variables.
In contrast to new-classical equilibrium models, with their price-taking
firms (‘yeoman barbers’) making voluntary choices of the output level,
Keynesian non-market-clearing models turn the role of prices and output on
their head, with demand-taking firms making voluntary choices of the price
level. Thus price-setting behavior is the essence of Keynesian economics. Any
attempt to embed it in microeconomic foundations must begin from
monopolistic or imperfect competition, not perfect competition, because
Keynesian agents are inherently price setters, not price takers.
A central theme of both new-classical and new-Keynesian
macroeconomics is that accurate empirical predictions are necessary but not
sufficient conditions of an acceptable theory. In addition, a theory must
have microeconomic foundations in the behavior of utility-maximizing and
profit-maximizing individual agents. The search for tractable analytic
models to form the micro foundations often leads analysts astray, causing
them to lose sight of the forest as they construct their single exquisitely
proportioned tree. Almost all new-classical theory is conducted in the
analytically convenient setting of ‘representative agent models’, where one
can move back and forth between the individual agent and the aggregate
economy simply by adding or removing i subscripts, without having to
consider such analytically inconvenient issues as coordination failures or the
speed of price adjustment. Professional microeconomists, as distinguished
from macroeconomists who dabble in microeconomic modeling, find the
failure to confront aggregation seriously to be the most critical flaw of
representative agent modeling.31 A surprising number of new-Keynesian
models share in common the neglect of aggregation; the aggregate economy
is simply the representative agent multiplied by n. Accordingly, we shall
find unsatisfactory those new-Keynesian models that neglect aggregation
issues, and we shall emphasize the central role of interactions among
agents, including coordination failures, macroeconomic externalities, and
producer-supplier relations.
Micro agents, macro spillovers, and coordination failures
The development of new-Keynesian economics since 1980 has primarily
involved the search for rigorous and convincing models of wage and/or price
502 Robert J.Gordon
stickiness based on maximizing behavior and rational expectations. The
ground rules of this search are commonly accepted. The key ingredient in the
now-abandoned Mark I new-classical approach was not rational expectations,
but rather the assumption of continuous market clearing, as is evident in the
labels new-classical macroeconomics or equilibrium macroeconomics. Most
new-Keynesian models combine rational expectations with maximizing
behavior at the level of the individual agent. Any attempt to build a model
based on irrational behavior or submaximizing behavior is viewed as
cheating. No new-Keynesian wants to build a model with agents that Barro
(1979) could criticize as failing ‘to realize perceived gains from trade.’ So the
game is to tease a failure of macro markets to clear from a starting point of
rational expectations and the maximization of profits and individual welfare
at the micro level. In short, effects of changes in nominal aggregate demand
on real output and employment are derived in models characterized by
equilibria in which all individual agents take only those actions that make
them better off and in which no agent foregoes an opportunity to take
advantage of a ‘gain from trade’.
The development of microfoundations for wage and price stickiness does
not, of course, represent the first attempt to develop micro underpinnings for
Keynesian economics. The work of Friedman and Franco Modigliani on
consumption, Dale Jorgenson on investment, and William Baumol and
James Tobin on the demand for money were all based on profit-maximizing
behavior at the micro level. But all this work was carried out within a
partial equilibrium framework, assuming in particular that both real
income and the price level were given. A useful distinction can be made
between micro theorizing at the level of individual demand and supply
functions, and micro analysis of the market mechanisms (especially the
price system) whereby the actions of maximizing agents are coordinated.32
Even before the advent of new-classical economics, the work of Clower
(1965) and Leijonhufvud (1968) stressed interactions and spillovers among
markets and argued that the nexus of research should shift from a partial to
a general equilibrium setting.
An interesting aspect of US new-Keynesian research is the near-total
lack of interest in the general equilibrium properties of non-market-
clearing models. That effort is viewed as having reached a quick dead end
after the insights yielded in the pioneering work of Barro and Grossman
(1971, 1976), building on the earlier contributions of Patinkin (1965),
Clower, and Leijonhufvud. Explaining sticky wages and/or prices is
viewed as a tough task, and no one is prepared to anticipate its
achievement by examining broader theoretical implications.33 The disdain
shown by new-Keynesian theorists for the work of Barro and Grossman,
and the latter evolution of that line of research in the hands of Malinvaud,
Muellbauer and Portes, Benassy, Grandmont, and others—notably all
Europeans—is understandable in light of the primacy of microfoundations
models as the prerequisite for macro discourse.34 Nevertheless I find that
What is new-Keynesian economics? 503
even the most perceptive new-Keynesian commentators tend to forget the
central message of these models.
This message is that spillovers between markets imply that the failure of
one market to clear imposes constraints on agents in other markets. Most
notably, when firms in a recession experience a decline in sales at the going
price, this excess supply of commodities ‘spills over’ into a decline in labor
demand at the going real wage. In this light, I am sometimes surprised to
read otherwise sensible commentators refer to the inconsistency of one or
another new-Keynesian explanation with microeconomic evidence on the
elasticity of labor supply. Such evidence is simply irrelevant for Keynesian
macroeconomics. In a genuinely Keynesian model, agents are not in a
position to choose the amount they work or produce as output varies over the
cycle, and so the constrained amount that they do work or produce cannot be
interpreted as tracing movements along a choice-theoretic labor supply curve
or production function.
Much existing new-Keynesian theorizing is riddled with inconsistencies as
a result of its neglect of constraints and spillovers, and its focus on single
markets, one at a time, in a partial equilibrium framework. For instance,
several of the most prominent models of price determination in the presence
of adjustment costs limit the source of price stickiness to the product market;
they often assume a perfectly competitive labor market in which workers slide
up and down their labor supply curves, indifferent between economic states
that offer relatively large and small amounts of leisure. Such models stand
Keynesian economics on its head, because any satisfactory explanation of
business cycles that warrants the label Keynesian must incorporate not just
price stickiness, but in addition some element that explains the evident
unhappiness of the employed in recessions and depressions. Further, such
models fail to explain why the adjustment costs that lead to price stickiness
do not in parallel imply wage stickiness.
One important exception to this neglect of macroeconomic constraints
and spillovers is the seminal work of Cooper and John (1988) on
macroeconomic coordination failures. In several new-classical models in
which agents set output, they show that spillovers and strategic
complementarities can arise at the levels of preferences and technology or in
the organization of transactions. They reach the same conclusion as Barro
and Grossman (without making the connection) that macroeconomic
quantities belong in microeconomic choice functions. Almost alone among
recent American authors in the new-Keynesian tradition, Cooper and John
cite Jean-Pascal Benassy’s fixed-price models (1975, 1982) and conclude for
such models that ‘strategic complementarity is a distinguishing element of
models with Keynesian features’ (1988:461).
The contribution of Cooper and John reaffirms the traditional view (see
particularly Leijonhufvud 1981) that coordination failures represent the core
problem in macroeconomics. In response to a nominal demand change, no
single private agent has an incentive to move its price exactly in proportion
504 Robert J.Gordon
unless it believes that all other agents will do likewise, and will do so without
delay. In Tobin’s example,
No one can see the spectacle in the theater or stadium if everyone stands,
but who has the incentive to obey a general admonition to sit down. When
the teacher tells her grade school class there will be no picnic unless all
gum-chewing ceases, would any rational child who shares the general
liking of gum stop? Threats against everybody in general addressed to
nobody in particular rarely work.
(Tobin 1989:15)
The same point can be put differently: rational microeconomic agents care
about the relation of their own price to their own costs, not to aggregate
nominal demand. Unless a single agent believes that the actions of all other
agents will make its marginal costs mimic the behavior of nominal demand
with minimal lags, the aggregate price level cannot mimic nominal demand,
and Keynesian output fluctuations result.
A notable limitation of most formal models related to coordination
failures, including Cooper and John (1988) and Durlauf (1989: esp. p. 110), is
a classical setting of competitive output setters, rather than a Keynesian world
of monopolistic price setters. In Durlauf’s words,
the hallmark of this class of theories is the compatibility of different levels
of real activity with the same microeconomic specification of individual
firms and consumers. The key source of the multiplicity of long-run
equilibriums is the positive effect that high production by some set of
agents has on the decision of others to produce.
This approach, based in part on seminal research by Diamond (1982, 1984),
essentially concerns the cyclical behavior of productivity, the positive
response of which is claimed to reflect ‘thick markets’ as a result of ‘positive
complementarities’. However, this has little to do with the essential Keynesian
coordination failure, the absence of incentives for price-setting agents to move
their individual prices in tandem with aggregate nominal demand rather than
individual marginal cost.35
Real rigidities, nominal rigidities, and the indexation puzzle
Two central distinctions are required as a preliminary to any summary of
new-Keynesian work. The first is between price setting in product markets
and wage setting in labor markets. The second distinction is between nominal
rigidity and real rigidity.
The necessary condition for non-market-clearing is a barrier to the full
adjustment of nominal prices, that is, something that prevents movements in
nominal prices that are equiproportionate to movements in nominal demand.
What is new-Keynesian economics? 505
However, some of the new-Keynesian theories explain real rigidities as the
stickiness of a wage relative to another wage, of a wage relative to a price, or
of a price relative to another price. Explanations of real rigidities in product
markets include customer markets, inventory models, and theories of markups
under imperfect competition, while those of labor markets include implicit
contracts, efficiency wages, and insider-outsider models. But theories of real
rigidities are subject to the criticism that they do not explain nominal rigidity,
because nothing prevents each individual agent from indexing its nominal
price to nominal aggregate demand.
There is surprisingly little discussion in new-Keynesian papers of optimal
indexation nor of the relation between the absence of full indexation and the
sources of nominal rigidities. Gray (1976), Fischer (1977b), and others showed
in the mid-1970s that full CPI indexation is not optimal in the presence of
supply shocks. Intuitively, no agent can afford an indexed contract that
rigidities real wages and relative prices if supply disturbances continuously shift
the optimal relative price. However, Gray’s argument supports only indexation
to a mix of price indexes preferred by firms and workers, not zero indexation.
Failing to index is tantamount to linking the prices and wages of individual
agents to a price index whose value is constant, and this becomes increasingly
irrational as the inflation rate increases.36
Further, full indexation of the wage rate to nominal GNP escapes most of
the theoretical objections to CPI indexation, because nominal GNP indexation
leaves the price level free to move to equate the real wage to the marginal
product of labor. Adopting our previous notation with lowercase letters
representing growth rates, the condition necessary for labor’s share in
national income to remain constant is that the growth rate of the real wage
(w–p) equals the growth rate of labor’s average product (q-n):
w-p=q-n, which occurs if
w=p+q-n=x-n.
Thus the growth rate of the nominal wage rate should be indexed to the
growth rate of nominal GNP per unit of labor input (x–n). If an adverse
supply shock reduces labor’s average product, then such indexation allows
the needed reduction in the real wage, whether nominal GNP growth
remains constant and the rate of inflation increases, or whether the
inflation rate remains constant and the growth rate of nominal GNP
decelerates.
Fischer (1986:152–3, 263–9) has pondered why indexation to the price
level is so often incomplete, and why we so rarely observe indexed contracts
contingent on other variables (whether economy-wide like nominal GNP or
idiosyncratic variables like firm sales or profits). The primary barrier to
indexation may be the costs of making contracts more complicated,
particularly when it is recognized that there are conflicts along at least two
dimensions. First is the Gray-Fischer point that the presence of aggregate
supply shocks makes incomplete indexation optimal, and second is the
506 Robert J.Gordon
presence of firm-specific shocks that create a conflict between the general
market basket to which workers would prefer to index, and the firm-specific
variables to which firms would prefer to index. Parties to a contract may
differ not only in their objective functions, but also in their perceptions of the
relative importance of aggregate demand shocks, aggregate supply shocks,
and firm-specific shocks, and these perceptions may change continuously,
requiring that the form of indexation in each new contract be negotiated from
scratch.
As we review sources of rigidities in Keynesian models, we shall return to
the issue of nominal GNP indexation. Are the nominal rigidities adequate to
explain the real-world absence of such indexation? How are the two
distinctions, product versus labor market and real versus nominal rigidities,
related to each other? We begin our inquiry by reviewing models of nominal
price rigidity in product markets, beginning with the elementary example of a
textbook monopolist. This example implies that the response of price to a
shift in demand is conditional not just on the elasticity of demand and the
shape of the marginal cost curve, but crucially on the shift (if any) of
marginal cost in response to demand. Thus product and labor market
rigidities are complementary and may be of equal importance.
5 THE SEARCH FOR STRUCTURE: NOMINAL PRICE RIGIDITY IN
THE PRODUCT MARKET
The textbook monopolist and the behavior of marginal cost
The behavior of a textbook monopolist is part of relative price theory, and
therefore would appear to belong in our subsequent discussion of real
rigidities. However, the monopolist model has been used to derive the
conditions under which costs of adjustment create a barrier to changes in
nominal prices. This explains the connection between theories of nominal
price stickiness and the traditional partial equilibrium analysis of a price-
setting monopolist illustrated in Figure 21.3. Note that two special
assumptions are made in drawing Figure 21.3, that the demand curves are
linear and that the marginal cost curve is horizontal. Implications of dropping
both of these assumptions are discussed shortly.
In response to a shift in the demand curve from D0 to D1, quantity will
change unless there is an equiproportionate shift in nominal marginal
revenue and nominal marginal cost at the original level of real output. The
implied marginal cost curve that maintains a constant level of output (Q0) is
labeled ‘Required’ MC1. If, following the decline in demand, marginal cost
drops instantly to the ‘Required’ MC1 line, then the intersection of MR and
MC will drop from E to G, and the price will fall by exactly the vertical
displacement of the demand curve, from P0 to P2. Any source of incomplete
adjustment in marginal cost can then explain an incomplete adjustment of
price. For instance, if the marginal cost schedule remains fixed at MC0, the
What is new-Keynesian economics? 507
intersection of MR and MC occurs at point F, the new price is P1, and the
new quantity is Q1.
When we loosen the two special assumptions incorporated into Figure
21.3, we alter the path of the price level at output levels away from the initial
level Q0 but not the basic conclusion about the required drop in marginal cost
for the economy to remain at Q0. For instance, replacing the special
assumption of a linear demand curve with a demand curve of constant
elasticity, while retaining the assumption of a horizontal MC schedule, point
C would lie directly to the left of point A, and the price level would remain
fixed in the presence of any leftward movement of the demand curve. Second,
replacing the horizontal MC schedule with a positively sloped MC schedule
going through point D would move points C and F down and to the right,
thus increasing the response of the price level to the decline in demand and
correspondingly reducing the output response.
This analysis suggests that the primary reason for sticky price adjustment
is the sticky adjustment of marginal cost. This would appear to place the
analysis of cost stickiness at the top of the new-Keynesian research agenda.
From the standpoint of the aggregate economy, the most important cost
component is labor cost, suggesting the familiar idea from the old Keynesian
economics that wage inflexibility is the key element in price stickiness.
However, from the standpoint of the individual firm, labor cost may be less
important than purchased materials as a component of cost, and this
recognition elevates to the top of the research agenda, along with wage
determination, the formation of expectations by individual producers about
the prices of purchased materials.
While Figure 21.3 identifies the rigidity of marginal cost as the key
ingredient in price stickiness, it also leaves open a role for direct barriers to
Figure 21.3
508 Robert J.Gordon
the adjustment of price to the profit-maximizing level for the monopolist,
that is, to point C in the case of a fixed MC schedule or to point D in the
case of a fully flexible MC schedule. Point B represents a price above the
profit-maximizing levels C or D, and could be explained by costs of
adjustment of the price level emphasized by new-Keynesian theorists under
the general heading of menu costs or by old-Keynesian economists who
emphasized rules of thumb like fixed markups of price over long-run
average cost. If the price level is predetermined at point B, while marginal
costs are predetermined along the schedule MC0, output and employment
may vary up and down in response to variation in product demand without
a change in the real product wage.
This analysis of Figure 21.3 helps to organize our treatment of new-
Keynesian research on the sources of price stickiness. First we examine the
studies of direct barriers to price adjustment, independent of the behavior of
marginal cost, which cause the price to deviate from the price that would be
set by a profit-maximizing monopolist who has no costs of adjustment to
consider. These direct barriers may be subdivided into two categories, (1)
state-dependent rules, which call for price changes if the optimal price strays
outside of boundaries determined by menu costs of price adjustment, and (2)
time-dependent rules, which call for price changes at fixed and predetermined
intervals written into contracts and are in turn presumably based on the costs
of negotiating new contracts at more frequent intervals. This branch of new-
Keynesian economics reinterprets these rules as profit-maximizing when
menu-type or negotiation-type costs of adjustment are taken into account.
Then we turn to sources of stickiness in marginal cost, both in prices of
purchased materials and in wages.
It should be clear from this analysis that the labor market and product
market may be equally important in contributing sources of price rigidity.
There has been some tendency to stress product markets relatively more in
recent research and to search for some source of nominal rigidity for prices in
the form of state-dependent or time-dependent rules. Yet it is clear from the
monopolist example that any source of nominal rigidity will do: a menu cost
for wage adjustment will make marginal cost sticky and indirectly create a
source of nominal price stickiness, even if costs of adjusting prices are
completely absent.
The representative-agent model of monopolistic competition
It is clear from Figure 21.3 that the mere presence of monopolistic
competition does not create a presumption of price stickiness. Some
ingredient must be introduced either as a direct barrier to instantaneous
price adjustment or as a source of sticky marginal cost. In new-Keynesian
literature this point is most often made in the context of a simple model of
a representative-agent monopolist developed by Blanchard and Kiyotaki
(1987) and described as the canonical model of monopolistic competition by
What is new-Keynesian economics? 509
Fischer (1988).37 There are n identical producer-consumers producing goods
that are imperfect substitutes, and there are no purchased materials.
Nominal aggregate demand depends only on the nominal money supply.
Marginal cost consists of the marginal disutility of production for each
producer-consumer. The canonical model describes the determinants of
output and of the desired relative price (Pi /P). With constant returns in
production and a constant marginal disutility of work, the model is
equivalent to Figure 21.3 with a flat MC schedule and a constant-elasticity
demand curve. The producer reacts to changes in demand by changing
output while leaving the relative price constant.
Only with an upward sloping MC schedule (due either to decreasing
returns to labor in production or to an increasing marginal disutility of work)
does the producer desire to change the relative price in response to a shift in
demand. However, because there is complete symmetry across producers,
relative prices must all be equal to unity. An attempt to decrease relative
price in response to a decline in demand leads to a decrease in all nominal
prices and in the aggregate price level, and this adjustment of the aggregate
price level continues until all relative prices are back to unity. Money is
completely neutral. The only element introduced by monopolistic competition
is a declining marginal revenue schedule, which means that in equilibrium
(with Pi /P=1) price is above marginal cost rather than equal to marginal cost,
and output is lower than in competitive equilibrium.
There is no role for sticky marginal cost in the Blanchard and Kiyotaki
‘pure’ model of monopolistic competition, because the imposition of
symmetry across identical representative-agent producers has the effect of
implicitly indexing both the relative price (Pi) and marginal cost to the
aggregate price level, which in turn depends only on the nominal money
supply. Thus the new-Keynesian theorists recognize that they must go
beyond the mere introduction of monopolistic competition in order to
locate the sources of price stickiness. One route is to study direct barriers
to nominal price adjustment in the form of state-dependent or time-
dependent rules. The other direction is to study the sources of sticky
marginal cost.
S,s state-dependent pricing rules
The new menu-cost literature owes its origins to a paper by Barro (1972) on
the S,s approach to price adjustment by a profit-maximizing monopolist who
faces a lump-sum cost of adjusting prices. The common theme linking the
older S,s literature and the newer menu-cost literature is that price setters do
not change price every time the desired price level changes, but only when the
desired level deviates by more than a particular percentage from the current
price. In the S,s literature the width of the percent band is arbitrarily given,
while in the menu-cost literature the width, while also given, is presumed to
be ‘small’ and ultimately capable of being explained by particular adjustment
510 Robert J.Gordon
costs. For expository purposes these contributions may be discussed together,
because they both concern barriers to the adjustment of nominal prices and
share the common theme of a percentage band within which the price remains
fixed.
The basic S,s result is derived for a monopolist facing a stochastic additive
shift in its demand curve taking the form of a random walk without drift. The
optimal strategy for the monopolist is shown to be the selection of ‘floor’ and
‘ceiling’ bands, with the price remaining constant when the shift is within the
bands and changing fully to the new desired level when the shift is outside the
bands. The width of the band, expressed as a percentage of the current price,
depends positively on the cost of a price change and inversely on the
opportunity cost of not changing, which in turn depends on the slopes of the
demand and cost functions.38
This result is specific to a demand disturbance that is modeled as a
random walk, so that changes in the disturbance are serially independent,
and as yet optimal rules have not been derived for more general processes
in which the changes in the disturbance are serially correlated (as surely
they must be in view of serial correlation in changes in nominal demand
evident in Figures 21.2 and 21.3). Instead, most of the extensions of the S,s
approach concern inflation which is at a sufficiently rapid rate that the price
level cannot decrease, so the choice problem is simplified to choosing the
timing of price increases. Sheshinski and Weiss (1977, 1983) show that the
S,s approach carries over to inflation; now the price is increased at any
point when it sinks below the optimal price by an amount exceeding a lower
s band.
Caplin and Spulber (1987) have investigated the implications of
aggregating S,s behavior from the level of the individual to the aggregate
economy. Their striking result is that one-sided S,s rules (as are appropriate in
an economy with an inflationary bias) do not lead to price stickiness or the
non-neutrality of money. If firms face both local and aggregate shocks, their
price changes will be independent and staggered across time. But when they
do increase their individual price, they will raise it sufficiently to boost the
aggregate price level by the full amount of the aggregate shock. For example,
if demand increases in a series of one-unit steps, and adjustment costs limit
individual firms to a price increase every fourth step, then that individual
price increase will be four units and will increase the aggregate price level by
one unit.
The Caplin-Spulber result is contingent on an unrealistic assumption,
that the desired price follows a continuous and monotone path. A more
general model, which reverses their main conclusion, has been developed
by Caplin and Leahy (1989). Their main point is that when the monetary
shocks are two-sided, that is, when money can go both up and down,
without any monotonic tendency in a single direction, there can be long
periods in which the aggregate price level does not change in response to
monetary disturbances. Intuitively, money is neutral only when the
What is new-Keynesian economics? 511
economy continuously hits an upper S or lower s band, but a more general
stochastic process for money may leave it inside both bands for substantial
periods during which there is no incentive for any agent to change its
nominal price.
A difficulty in the S,s literature is that for analytical tractability all firms
are identical, and thus have price increases of equal size that differ only in
timing. This is belied by virtually all evidence on cross-section pricing
behavior, including the differing cyclical responsiveness of prices across
industries in the Great Depression (shown in Table 21.2). This evidence
suggests that elements beyond simple state-dependent pricing rules must lie
behind observed price behavior at the micro level.
The menu-cost insight and its limitations
Taking the S,s literature as a point of departure, what new insights have been
contributed by the menu-cost literature developed in the mid-1980s? The
menu-cost approach began defensively in response to those critics who argued
that costs of changing nominal prices are much too small to justify output
fluctuations of the size observed in the USA. Its key insight is that second-
order adjustment costs may have first-order social consequences, simply
because profit functions are flat on top.39 Following a change in demand there
may be little difference in the firm’s profit if it does or does not adjust its
price, and thus even small menu costs may potentially dissuade the firm from
price adjustment. Yet the social consequence of such a failure to adjust price
may be large swings in output.
The proponents of the menu-cost approach are quick to admit that this
widely used label is misleading. Included among the nominal costs of price
adjustment are not just the literal application of the label to changing prices
on menus, lists, catalogs, and other printed material, but more generally the
entire range of costs that managers must incur whenever nominal prices are
changed. Meetings, phone calls, and trips to renegotiate with suppliers all fall
under the rubric of menu costs. Included in this more general definition of
menu costs is Okun’s (1981) analysis of the product market. Okun explains
the reluctance of firms to shift from FIFO (first in first out) pricing policies to
the more economically rational behavior of replacement cost pricing as a
consequence of the perceived costs of delegating pricing authority to lower
levels of management, in contrast to general FIFO-type rules of thumb that
save these costs of delegation even if they lead to pricing decisions that may
be otherwise suboptimal. All these physical costs of printing, negotiating, and
delegating are doubtless present in the real world of business, although one
can quibble with their importance. Costs of negotiating are also a key
ingredient that motivates staggered contracting, a time-dependent rule
considered in the next section.
Whatever the nature of the menu costs, the analysis may be presented in
terms of Figure 21.3 which already provides the ingredients necessary to
512 Robert J.Gordon
illustrate the point initially made by Akerlof and Yellen (1985) and Mankiw
(1985). Following Mankiw, we examine the situation in which demand has
declined in Figure 21.3 from D0 to D1 and marginal cost has declined from
MC0 to ‘Required MC1’. The optimal price-output point is at D, and we ask
what difference is made if the firm leaves its price unchanged at P0. Figure
21.4 copies the new demand curve and shows the same points B and D as in
Figure 21.3. The difference between profits at points D and B is shown by the
rectangles T–R. However, at point B total surplus is smaller by the area S+T
than at point D. But the firm will reduce price only if the extra profit T–R
exceeds the menu cost. Mankiw shows that as the price elasticity of demand
varies from ten to two, the ratio of the social cost to the profit increment
varies from 23 to 200. His results, as do Figures 21.3 and 21.4, assume that
the marginal cost schedule is flat. In general, the flatter is the marginal cost
schedule, the smaller are the menu costs needed to make the firm’s fixed-price
decision optimal and hence to create an output response from a change in
nominal aggregate demand.
At least four important criticisms of the menu-cost approach may be
offered. Taken together, they make a strong case against this core
contribution of the new-Keynesian macroeconomics.
First, a consideration of symmetry brings the basic conclusion into
question. If the failure to reduce price in response to a demand reduction
makes output too low, then the failure to raise price in response to a demand
increase makes output too high. Yet, starting from an initial profit-
maximizing equilibrium level of output like Q0 under monopolistic
competition, society gains from additional output because price is above
marginal cost. Hence the menu-cost model fails to prove its point: social costs
in recessions are balanced by social gains in booms. Any cost from price
rigidity must involve increasing the variance of output, not changing its
Figure 21.4
What is new-Keynesian economics? 513
mean, and hence are likely to be second-order, just as the costs of changing
price are second-order. One cannot conclude that sticky prices necessarily
reduce welfare, for the comparison of two second-order effects turns on
model-dependent comparisons of parameter values.40 This argument of Ball
and Romer (1989a) greatly weakens the appeal of the menu-cost approach,
although their own model implies that the second-order social costs can be
much larger than the costs of changing price.
Second, but independent of the Ball-Romer symmetry argument, the menu-
cost approach seems flawed from the start, because it considers only costs of
price adjustment and totally ignores costs of output adjustment. This places
its assumptions in diametric opposition to other important branches of
macroeconomics, such as Tobin’s q theory of investment based on time-
dependent physical costs of changing the capital stock, or the production-
smoothing theory of inventory behavior based on the assumption that a
smooth rather than variable production minimizes cost. Costs of output
adjustment raise the cost of not changing price and tilt the firm’s decision
toward price flexibility; whether costs of output adjustment raise the social
costs of aggregate output fluctuations depends on the relative size of the
private and social costs.
Third, the two-period comparison of Figure 21.4 neglects the calculus of
costs and benefits in future periods. The proper setting is dynamic, as with the
analogous question of the ‘sacrifice ratio’ in the form of a temporary
aggregate output loss required to achieve a permanent reduction of the
aggregate inflation rate. The proper comparison is between the one-shot menu
cost and the present value of the infinite stream of losses by maintaining the
price (and output) levels at different values than the social optimum. The
Ball-Romer symmetry argument vitiates the force of this criticism, because
the infinite stream of losses when the price is set too high is balanced by a
similar stream of gains when the price is set too low.
Fourth, like the S,s approach, the menu-cost approach fails to explain why
prices of some products are more flexible than others over the business cycle.
The failing of the Mankiw version illustrated in Figure 21.4 is in this regard
similar to that of the canonical Blanchard-Kiyotaki model described earlier;
marginal cost is simply assumed to move in proportion with demand. Once
we consider the many layers of heterogeneity of products and industries
studied in the industrial organization literature of Section 3 (pp. 490–92), we
recognize that no individual firm can assume its marginal cost will be
perfectly correlated with aggregate demand. Subsequently this will lead us to
the input-output table as an essential component in the description of price
stickiness and will reinforce our previous point that the failure to consider
heterogeneity and aggregation issues is a central flaw in representative agent
modeling.
Thus we return to the original objection to models of nominal rigidity
based on adjustment costs. Any satisfactory model of price rigidity must be
514 Robert J.Gordon
able to cope with the Great Depression, yet the magnitude of demand shifts
between 1929 and 1933 would seem to swamp any reasonable guess as to the
magnitude of S,s bands or menu costs. And one does not have to dip into
history to doubt the relevance of such adjustment costs. Everyone has
witnessed the fast-changing price tags in the produce section of the
neighborhood supermarket: There seems to be nothing to prevent the price of
a pint of strawberries from moving from $1.89 to $0.59 to $1.89 in successive
weeks. Carlton’s evidence shows that prices can jump not just by large
amounts in successive weeks, but by small amounts in successive months.
Roberts et al. (1989) have found in time-series data for 20 two-digit US
industries over 1958–83 that the adjustment of nominal prices to nominal
labor and materials costs takes place extremely rapidly. For four industries,
90 percent of price adjustment occurs within the month, and for no industry is
the first-month adjustment lower than 45 percent. This provides strong
evidence that the menu-cost approach is on the wrong track, and that the key
issues concern the stickiness of both wages and materials costs, not final
goods prices. The mere fact of imperfect competition and price tags appears
to be quite compatible with nominal flexibility.
Time-dependent rules and staggered contracts
The last element to be considered in new-Keynesian explanations of
nominal price rigidity is the staggered contract model. As noted above,
new-Keynesian economics can be said to begin with the Fischer (1977a)
and Phelps-Taylor (1977) models of staggered contracts, which emphasized
wage contracts. More recently, models of staggered price contracts have
been developed by Blanchard (1983, 1986) and Ball and Romer (1989b),
among others. These models investigate the implications of staggered
overlapping price-setting intervals of constant length, and in the case of
Ball-Romer investigate the conditions necessary for firms to engage in
staggering.
The staggering and overlapping of intervals in which individual prices or
wages are fixed introduces a critical element of realism into new-Keynesian
economics. As shown by Blanchard (1983), a change in nominal demand
can affect output for a period that exceeds the length of the interval during
which prices are predetermined, which we will call the contract interval
even though there is no necessity that explicit or implicit contracts be
involved. Consider contracts of n months, with a fraction 1/n of agents
resetting contracts each month. The firms that reset their price in the first
month following a demand shift move their price not to the optimum given
the level of nominal demand, but to the optimum contingent on the fact that
the other outstanding contracts cause a fraction (n-1)/n of the aggregate
price level to be preset. Any firm adjusting all the way would cause a
suboptimal divergence of its relative price from the optimum level, given
the stickiness of other prices.
What is new-Keynesian economics? 515
The study of staggered prices takes as its point of departure that the
length of predetermination of prices reflects a balancing of the costs of
adjusting prices and the opportunity costs of nonadjustment, just as in the
S,s model. Because this decision need be made only by the profit-
maximizing price-setting monopolist, there is no need for an actual contract
with another agent. There has been little attention to the nature of the
adjustment costs or in particular their variation across industries or over
time, which is unfortunate as this might provide the element that is missing
in so many new-Keynesian models, the ability to explain cross-time and
cross-industry differences in price behavior. In particular, there is no element
in the theory that would explain why the rate-of-change (α) coefficient of
price adjustment increased in most countries during World War I, or why
the inertia (λ) effect increased in most countries except Japan after World
War II. As a separate criticism, there has been no attempt to introduce
explicit indexation into these staggered contract models, which contain no
element to explain why firms do not predetermine real prices for a time
interval (presumably to save on management decision costs) and then index
the nominal price to nominal GNP.
Instead, attention has been concentrated on the question of why there is
staggering rather than complete synchronization. Ball and Romer (1989b)
show that staggering is a stable equilibrium if there are firm-specific shocks
that arrive at different times for different firms. However, they show that
synchronization can also be an equilibrium: ‘Multiple equilibria are possible
because there is an incentive for synchronized price setters to remain bunched,
but not for staggered price setters to move toward synchronization’ (1989b:
193). There seems to be a debate as to whether firm-specific shocks are
sufficient to guarantee a staggering equilibrium, but only in the context of
simple models in which firms can choose only to change price at odd or even
dates. A more general setting in which some prices are changed weekly and
others yearly, and in which the yearly price changers have 365 possible dates
on which to change, destroys the argument that each individual price setter
still has an incentive to ‘bunch’ price changes at the same time, because there
is no such thing as the ‘same time’.41
6 SOURCES OF REAL RIGIDITY IN THE PRODUCT MARKET
Customer markets
The analysis of nominal price rigidity in Section 5 treats only the first
quadrant of our two-by-two matrix defined along the dimensions of product
versus labor market and nominal versus real rigidity. We turn now to models
of real rigidity, that is, models that explain why real wages or prices are
unresponsive to changes in economic activity. In the product market, a model
of real rigidity explains why a firm would choose to hold its relative price or
price-cost margin constant. In the context of the canonical Blanchard-Kiyotaki
516 Robert J.Gordon
monopoly model, this occurs with a constant marginal disutility of work and
constant returns to labor. In the textbook monopoly model of Figure 21.3, the
price-cost margin is fixed if the marginal cost schedule is horizontal and the
elasticity of demand is constant.
Recall that our discussion of indexation in Section 4 (pp. 504–6) introduced
a basic objection to all models of real rigidity. No matter how rigid is the
real wage or real price, what prevents the nominal price and wage from
being indexed to nominal GNP? At that point we asked whether nominal
rigidities were sufficiently important to be able to explain the absence of
nominal GNP indexation. Thus far we have concluded that nominal rigidities
based on S,s or menu-cost models are not convincing, while time-dependent
rules in the form of staggered price-setting intervals are completely
compatible with any form of indexation. Thus a critical test for the theories of
real price and wage rigidity is whether they stand up to the indexation
criticism.
Perhaps the earliest prominent model of real rigidity in product markets
explains why customers do not respond instantaneously to changes in real
prices, that is, in the price charged by one firm relative to others. Okun
(1975), building on the work of Alchian (1969) as well as Phelps and Winter
(1970), popularized the distinction between auction and customer markets.
The former are perfectly competitive. But, in the latter, costly search makes
customers willing to pay a premium to do business with customary suppliers,
and intertemporal comparison shopping discourages firms from changing
prices in response to short-run changes in demand in order to avoid giving
customers an incentive to begin exploring. Okun argues that his customer-
search model explains markup pricing practices based on full costs.
Customers appear willing to accept as fair an increase in price based on a
permanent increase in cost, whereas transitory events, whether an increase in
demand or a reduction in productivity, are not generally expected to last long
enough to justify price increases.
Okun’s approach has several critical defects. He argues that price
increases based on cost are perceived as fair, while cost increases based on
demand are not so perceived. The case for customer dissatisfaction is
difficult to argue, because any loss of goodwill created by a price increase
in a boom would be balanced by a gain of goodwill created by a price
decrease in a recession. Okun seems to be thinking of an inflationary world
in which price changes are one-sided, as in the trend-inflation S,s literature.
Further, the fairness explanation leaves open the determination of fair
behavior, and in fact what is perceived to be fair may just reflect whatever
behavior may be normal, for whatever reason. Thus Okun’s approach has
an element of circularity.
Okun’s approach also seems vulnerable to the same criticism often
directed at Lucas-type new-classical models: why should a firm be afraid
to lose customers when raising prices in response to higher nominal
demand if information on higher nominal demand is instantly available?
What is new-Keynesian economics? 517
What prevents all firms from indexing to nominal demand and advertising
price specials on items priced lower than would be warranted by the
indexation formula? As we have seen already, the elementary theory of
monopoly pricing behavior by itself suggests little for price flexibility.
Everything depends on the response of marginal cost to aggregate demand
shocks.
The independence of costs and demand
Nevertheless, there is a deeper insight in Okun’s distinction between cost and
demand. Firms raise price in response to an upward shift in the marginal cost
schedule not just because it is optimal in a textbook model, but because they
will go bankrupt if cost rises sufficiently in relation to price. There is no such
economic necessity of raising price in response to an increase in demand
when cost is fixed, and for a monopolist such a price increase is not even
optimal with a constant-elasticity demand curve and a flat MC schedule.
When OPEC raises the price of oil sharply in relation to the level of nominal
aggregate demand, everybody understands why the local service station raises
the price of gasoline at the pump, but they do not understand why an increase
in aggregate demand requires any such response of the gasoline price if the
costs of service station inputs are fixed.
This distinction hinges on the possibility that shifts in marginal cost can be
independent of shifts in aggregate demand. Our historical study of price
adjustment in Sections 2 and 3 stressed the theoretical and empirical
importance of supply shifts. Ball and Romer’s (1989b) study of staggering
emphasizes the critical role of idiosyncratic firm-specific shocks. Bertola and
Caballero (1990) emphasize the role of idiosyncratic uncertainty in explaining
infrequent price adjustment at the micro level. New-classical Mark I
macroeconomics was built on Lucas’ distinction between local and aggregate
shocks. Okun’s emphasis on cost-based pricing leads us to broaden Lucas’
two-way distinction between local and aggregate demand shocks and suggest
a four-way distinction between local and aggregate demand shocks, and local
and aggregate cost shocks.42
This four-way distinction creates two complementary sets of reasons why
firms may rationally expect marginal cost to move differently from marginal
revenue. First, marginal revenue may move with aggregate nominal demand
but marginal costs may not. This would occur if a firm believes that its costs
depend not just on nominal demand but on local supply factors (e.g. harvests,
strikes, price changes for imported materials). Second, in a situation with
nominal aggregate demand fixed, a firm might face a local shift in demand
(e.g. a decline in beer drinking in response to drunk-driving laws) that reduces
marginal revenue, while marginal cost is fixed, tied to aggregate nominal
demand. More generally, any set of covariances among the four shock
concepts is possible.
518 Robert J.Gordon
The role of the input-output table
To be a credible explanation of real price rigidity, the distinction among local
and aggregate cost and demand shocks must be embedded in a world with
many heterogeneous firms interacting within a complex input-output table.
With only two firms, each supplying the other, information would be cheap
enough to permit both firms to disentangle the local versus aggregate
component of their costs. But with thousands of firms buying thousands of
components, containing ingredients from many other firms, the typical firm
has no idea of the identity of its full set of suppliers when all the indirect links
within the input-output table are considered. Because the informational
problem of trying to anticipate the effect of a currently perceived nominal
demand change on the weighted-average cost of all these suppliers is difficult
to formulate and probably impossible to solve, since (as Bresnahan
emphasizes) thousands of elasticities are involved, the sensible firm just waits
by the mailbox for news of cost increases and then, Okun-like, passes them on
as price increases.
The input-output table approach provides a critical contribution not just to
understanding real price rigidity, but also nominal rigidity. The standard
accusation against all theories of real rigidity, made often above, is that they
are consistent with nominal flexibility achieved through indexation to
nominal demand. Yet the input-output table approach emphasizes the high
fraction of a firm’s costs that is attributable to suppliers of unknown identity,
with some unknown fraction producing in foreign countries under differing
aggregate demand conditions. This environment would give pause to any
firm considering nominal demand indexation of the product price, because the
failure of all suppliers to adopt similar indexation could lead to bankruptcy
when nominal demand declines. Thus the input-output approach borrows one
element that is basic to Keynesian economics, the coordination failure that
arises from the lack of private incentives to solve a social problem, with
another element inherited from Lucas, the distinction between aggregate and
idiosyncratic shocks.
One criticism of the input-output approach claims that with perfect
information about aggregate variables, the only equilibrium of the economy
would be for immediate adjustment of all prices to nominal shocks. Yet this
ignores the fundamental assumption that marginal cost and marginal revenue
are imperfectly correlated with aggregate demand. Under these conditions
each firm would be unwilling to index price to nominal GNP both because
marginal cost may not move with nominal GNP even if marginal revenue
were to do so, and vice versa.
A good reason for every domestic firm to refuse to index its product price
to domestic nominal demand would occur to any economist from, say,
Belgium or Chile. Because we know that purchasing power parity (PPP) fails
and that real exchange rates are volatile, why would any firm adopt
indexation of its price to domestic Belgian or Chilean nominal GDP, which
What is new-Keynesian economics? 519
would disconnect its price from the large share of its costs that are imported?
The input-output approach, by stressing the independence of marginal cost
and aggregate demand, provides an understanding of the lack of indexation
to domestic nominal GNP and thus the critical link that converts a theory of
real rigidity into a theory of nominal rigidity.43
A firm’s viability depends on the relation of price to cost, not price to
nominal GNP. Aggregate macroeconomic stability is a public good subject to
a free-rider problem. No individual firm has an incentive to take the risk
posed by nominal GNP indexation, which would take away from the firm the
essential control required of the relation of price to cost. In this sense, the
input-output explanation of nominal rigidity requires capital markets that are
imperfect enough to penalize the profit volatility that would result if a firm
tried to index its prices to nominal demand without being sure in advance
that its suppliers would do likewise.
There is another sense in which the input-output table explains nominal
rigidity. It creates a technological environment for staggered price setting,
similar to but more complex than Taylor’s staggered wage setting. Today’s
product price is based on costs set at many different dates in the past as
product components weave their way through the input-output table. This
may appear to violate the maxim that prices should be based on replacement
cost. But there are too many links in the input-output table for the producer
even to guess what the replacement cost may be. The automobile firm may
receive a notice from the headlight maker of a price increase, but no warning
of a price-increase notice that is already in the mail from the filament maker
to the headlight maker or from the copper maker to the filament maker.
Blanchard (1987b) uses the term cumulation hypotheses to describe the role of
the input-output table in translating prompt price adjustment at the individual
level to gradual price adjustment at the aggregate level. He provides
suggestive supporting evidence that in disaggregated data prices adjust faster
than in aggregate data. The automobile, headlight, filament, and copper
maker may all respond to cost increases within a day, but months can
separate the effects of a change in the price of copper from the ultimate
change in the price of automobiles.
The input-output table approach dominates menu costs in explaining why
the price of strawberries is more volatile than the price of automobiles
(because strawberries are not physically transformed from farm to market). It
explains the different rate-of-change adjustment coefficient (α) across
industries by two auxiliary assumptions. First, auction markets are distinct
from customer markets and are limited mainly to crude and intermediate
goods. Thus products like strawberries and plastics that appear relatively
early in the input-output chain have relatively flexible auction-like prices. But
what is it that creates more price rigidity for more complex products later in
the chain? Partly it is the law of large numbers that cancels out idiosyncratic
supply shocks for final products incorporating large numbers of different
purchased materials. But there also may be a role for wage rigidity, as the
520 Robert J.Gordon
prices of products embodying relatively large amounts of embodied labor,
like automobiles, tend to be more rigid than that of products embodying large
amounts of embodied land, like wheat or strawberries. Thus the input-output
approach is complementary to theories of rigidity in the labor market.
A safe compromise place to end the discussion of product-market rigidities
is to admit that the input-output approach is complementary as well to at
least one of the new-Keynesian approaches based on nominal rigidities. The
input-output approach needs some additional element to explain why we do
not observe in the real world extremely frequent small price changes every
day as firms react to each tiny cost change as it arrives in the mail through
the input-output table (while Carlton documents some such small changes,
long intervals of complete rigidity are common as well).
The core element that needs to be added to the input-output approach is a
cost to making price changes every day that causes rational managers to
concentrate price-setting decisions at discrete intervals. There is no need to
force this sort of nominal rigidity into a single semantic category; the core
factor for some firms may best be described as staggered time-dependent rules
and for others as state-dependent rules based on menu-type costs. Undoubtedly
these categories overlap because many firms face both time-dependent and
state-dependent costs. Some firms that routinely hold price-setting meetings
once a week or month to save on managerial time costs may decide at those
periodic meetings to leave some or all prices unchanged when the difference
between the current and optimal price does not yet exceed the perceived cost
of printing new menus and catalogs.
7 THE SEARCH FOR STRUCTURE: LABOR MARKET BEHAVIOR
The relation of wage and price behavior
The dominant new-Keynesian view is that nominal rigidities originate in the
product market, not the labor market. The path of wages, to use the words of
Mankiw, is ‘completely indeterminant and completely irrelevant’ (1988:446).
Yet surely this goes too far. Mankiw follows earlier writing, notably by Barro
(1977b) and Hall (1980), who have argued that wage rigidity is irrelevant for
employment determination. In the context of a long-term or even lifetime job,
there is no reason for the wage in a given time period to be equal to the
marginal product of labor. The wage can be an installment payment on a
lifetime contract.
However, the claim that sticky wages are irrelevant to allocation calls for
prices to be perfectly flexible, as is required for perfect market clearing, while
wages are sticky. We have already observed in Section 2 (pp. 487–88) that
capital markets are likely to impose a tax on the resulting profit variability.
Further, the monopolist example shows that prices will not be perfectly
flexible unless all elements of marginal cost are perfectly flexible. This brings
What is new-Keynesian economics? 521
us back to indexation: The sticky wages that are installment payments for
lifetime jobs must be fully indexed to nominal demand for Barro, Hall, and
Mankiw to be correct that sticky wages are irrelevant to allocation.
Just as it is implausible for wages to be sticky while prices are perfectly
flexible, so is the reverse, for wages to be perfectly flexible while prices are
sticky. Yet this is just what is assumed in much of the menu-cost literature
reviewed in Section 5. When menu costs lead rational firms to avoid price
changes and meet demand through changes in output, corresponding
fluctuations in labor input are required. In menu-cost models the real wage
adjusts to make workers willing to change the amount they work; that is, the
nominal wage rate is perfectly flexible. In the Blanchard-Kiyotaki canonical
model of monopolistic competition, the representative agents set their relative
price to minimize the marginal disutility of work; that is, they slide along
their voluntary labor supply curve. As Rotemberg has noted, ‘both of these
approaches have the very un-Keynesian implication that in recessions workers
are close to indifferent between working and not working’.
For new-Keynesian models to avoid inconsistency, their distinction between
small menu costs of price changes and large social costs of output changes
must apply equally in the labor and product markets. The same costs of
adjustment that inhibit price changes must apply equally to wages, which are
just another price. Sticky prices cause changes in nominal aggregate demand
to be transmitted directly to shifts in the demand curves facing not just
individual firms, but also individual workers. The Barro-Grossman spillover
model discussed in Section 4 (pp. 501–4) achieves the desired symmetric
treatment, in which sticky wages and prices cause both firms and workers to
face constraints on the amount they can sell.
Blanchard and Fischer (1989:427) state that the key issue in new-Keynesian
economics is explaining why ‘labor and output supply functions [are]
relatively flat.’ They intend this phrase to mean real rigidities.44 Yet their
choice of words is unfortunate, because it ignores the distinction between
aggregate and individual supply curves, as well as between notional and
effective supply functions. The labor supply function of an individual head of
household may be vertical, but any mechanism that rigidifies the real wage
will cause the individual to be pushed off this notional supply function. Actual
behavior traces shifts in the effective labor demand schedule and tells us
nothing about the shapes of notional functions. In our interpretation the key
issue is the explanation of wage and price rigidity, not the explanation of why
labor and output supply curves are flat.45 And at a deeper level, as argued in
Section 4 (pp. 501–4), the really central element is the coordination failure
that underlies wage and price rigidity.
Early theories of real rigidity: search models and implicit contracts
Just as theories of price stickiness can usefully be divided between theories of
nominal versus real rigidity, so can theories of wage stickiness. Reflecting the
522 Robert J.Gordon
chronological development of the field, we begin with models of real wage
rigidity and then turn to models of nominal wage rigidity.
Widely recognized as the first attempt to build structural models of labor
market behavior as the outcome of maximizing behavior was the new
microeconomics of the famous volume edited by Phelps (1970). Most of the
papers in the Phelps volume, including Phelps’ own desert island parable,
yielded market-clearing conclusions and as such should be regarded as part of
the development of new-classical rather than new-Keynesian ideas. In the
parable, workers are on isolated islands and react to a wage cut by boarding
rafts to sample wage offers on other islands. Variations in employment during
business cycles are due solely to the voluntary response of workers to changes
in the expected real wage. The parable ignores the prompt availability of
aggregate information, fails to explain layoffs and “no help wanted” signs,
and yields the counterfactual implication that voluntary quits vary
countercyclically (see Okun 1981: ch. 2).
The main contribution of the new microeconomics volume was not to
business cycle theory but rather to explain why the natural unemployment
rate is greater than zero, due chiefly to the work of Mortensen (1970a,
1970b). In a world of costly information and heterogeneous jobs and
workers, workers sample from an array of job offers and firms sample from
an array of workers. Unemployment is a voluntary activity, but all
voluntary unemployment is not socially beneficial, and government
unemployment benefits tend to stretch out the interval between searches,
imposing a social cost through the taxes levied on some to support the
extended search interval of others. The new microeconomics volume also
contained the important Phelps-Winter (1970) theory of customer markets,
based also on the assumption of imperfect information. We have seen that
this was later picked up and developed by Okun in a new-Keynesian rather
than new-classical setting.
While the new microeconomics was explicitly classical in approach, the
next wave of contributions under the heading of implicit contract theory was
the first to develop what some initially thought was a microeconomic
explanation for Keynesian wage stickiness. In the simultaneously written and
independent contributions by Azariadis (1975), Baily (1974), and Donald
F.Gordon (1974), employees were assumed to be relatively more risk averse
than their employers, mainly because of self-selection of individuals to
become entrepreneurs. Firms maximized profits by minimizing the variability
of income to their workers, who disliked variability, in effect providing a
compensation package that consisted partly of pecuniary wage payments and
partly of insurance services.
It was soon recognized that this approach provides no satisfactory
explanation of Keynesian unemployment; it justifies only a fixed-income
contract (i.e. tenure) rather than the fixed-wage variable-employment
contracts actually observed. Variable employment is explained only by a
gratuitous element patched onto the theory, government side-payments during
What is new-Keynesian economics? 523
periods of unemployment. Even when these variable employment fixed-wage
contracts are generated by the theory, they have the unKeynesian implication
that workers are equally happy when employed and unemployed.46 Further,
workers are shown to care about stability in real income, not nominal
income, so implicit contract theory has no explanation for the failure of
workers to insist on full indexation of wage contracts.
Labor unions
The effects of labor unions have been extensively analyzed by labor
economists. Bargaining models have been developed in which firms and
unions, which in turn act on behalf of their member workers, bargain over
wages and employment. Some models characterize the employment decision
as a unilateral decision of management, as it is in many contracts. These
models that are concerned only with wage setting are sometimes called the
right-to-manage model and fall between two extremes. At one extreme firms
are all powerful and are able to pay the minimum wage possible, that is, the
competitive wage. Because firms have complete control of both employment
and the wage, this subclass of models does not warrant the label bargaining
model at all; the efficiency wage models discussed in Section 7 (pp. 525–7)
fall into this class. At the other extreme is the union-monopoly model dating
back to Dunlop (1944); here too there is no bargaining, because firms set
employment and unions set the wage.
A more general model is developed by McDonald and Solow (1981), who
show that a bilateral monopoly between a firm and a union can lead to
relatively large employment fluctuations and relatively small real-wage
fluctuations, thus contributing a source of real-wage rigidity. In an
extension, McDonald and Solow (1985) examine the impact of business-
cycle fluctuations on a labor market segmented into a union primary sector
and a competitive secondary sector. Reflecting the small real-wage
fluctuations in the union sector, they show that either permanent or
temporary changes in real aggregate demand widen sector wage
differentials in recession and cause greater fluctuations in primary sector
than secondary sector employment.
Yet the formal theory of unions does not provide a general explanation of
Keynesian wage rigidity. If union members care about stability of
employment, it is difficult to understand why they are willing to tolerate a
wage rate that is set for a substantial interval, while the decision on the
amount of employment is left to the firm.47 Obviously if the wage rate is
predetermined as part of a union contract, this rigidifies marginal cost and
hence prices, and nominal demand fluctuations are transmitted to output
and employment. But, overall, the union literature leaves open the question
why the wage rather than the level of employment is set by contracts, and
why the wage rate is not indexed to nominal demand so as to stabilize
employment.
524 Robert J.Gordon
Another problem is raised by the empirical evidence of Table 21.3 (p. 494).
Unions became important in the USA only after the mid-1930s, yet the
estimated rate-of-change (α) and level (γ) effects for the USA are the same
during 1873–1914 and 1954–87. Some factor other than unions must account
for the price stickiness evident in US data for the nineteenth century. The
main contribution of unionization may have been the particular US
phenomenon of the three-year staggered wage contract, which has doubtless
contributed to the higher inertia parameter for both wages and prices. Yet
here too there are problems, because inertia in price change seems to have
increased more than for wage change (Table 21.3), and, further, inertia
increased substantially for price change in the postwar period in the UK,
France, and Germany (Table 21.4), nations in which the three-year contract is
not prevalent.
Insider-outsider theory
Another body of work that deals with the existence and persistence of
unemployment is the insider-outsider theory. The insiders are experienced
incumbent employees whose jobs are protected by a variety of labor turnover
costs which make it costly for firms to replace them. The outsiders, who are
either unemployed or work in the casual or secondary labor market, have no
such protection. Lindbeck and Snower (1986, 1988) argue that, owing to these
turnover costs, the insiders gain market power, which they use to their own
advantage, without necessarily taking fully into account the interests of the
outsiders. Further, the insiders often can influence the turnover costs
themselves by agreeing to cooperate among themselves but not with outsiders
should the latter attempt to gain employment by underbidding the insider
wage. This structure causes unemployment for the outsiders, who cannot find
jobs even though they would be willing to work for less than the prevailing
insider wage.
Although the insider-outsider theory contributes to our understanding of
union behavior, it is not primarily a contribution to the union literature. A
wide variety of labor turnover costs may well be significant even in the
absence of unions, for example, hiring, training, negotiation, litigation, and
firing costs, as well as costs that can be directly imposed by the insiders when
they shirk or fail to cooperate in the presence of underbidding outsider
entrants. Nevertheless, the insider-outsider theory does suggest a rationale for
unionization by showing how unions can organize and coordinate insiders’
rent-seeking activities.
The insider-outsider theory sheds light on a variety of labor market
phenomena, such as the persistence of unemployment, differences in
variability of employment across industries and countries, labor market
segmentation, the duration and composition of unemployment, and the
interindustry wage structure. The theory has been applied to the puzzle of
persistently high unemployment in Europe in the 1980s by Blanchard and
What is new-Keynesian economics? 525
Summers (1986) and Lindbeck and Snower (1988) and has become one of
several explanations of the hysteresis hypothesis (see Section 2, pp. 488–90),
in which the rate of unemployment depends on the history of actual
unemployment rather than, as in Friedman’s original version, being ‘ground
out’ by the microeconomic structure of the economy. The insider-outsider
approach explains the emergence of high unemployment in the 1980s as an
indirect consequence of the oil shocks of the 1970s, which created a
temporary adverse reduction in labor demand and caused the insider work
force to contract. When labor demand recovered the remaining insiders set
wages to maximize their own welfare, thereby discouraging employment and
making the high unemployment persist.48 The best evidence in support of this
approach is the work of Layard and Nickell (1987) which shows that the
demand pressure variable entering the Phillips-curve wage equation is not
total unemployment, but rather total unemployment minus the long-term
unemployed. However, to the extent that it explains the persistence of high
European unemployment by high insider real wages, it is subject to the
criticism (R.Gordon 1988) that high unemployment was immune to the
moderation of real wage growth and the disappearance of the European wage
gap in the 1980s.
Efficiency wage theory
If any development in the microeconomics of labor markets could be called
the ‘rage of the 1980s’, it is efficiency wage theory, based on the hypothesis
that worker productivity depends on the level of the real wage. When there is
such a link between the wage rate and worker efficiency, firms may rationally
pay a real wage rate that exceeds the market-clearing level. Firms may refuse
to reduce the wage to hire members of a pool of unemployed workers who
may be available at a lower wage, fearful that a reduction in real wages for
existing workers may reduce productivity by more than the gain in lower
wages. The appearance of an excess supply of labor in such a setting can be
shown to be consistent with maximizing behavior of both firms and workers.
There is substantial overlap between the insider-outsider and efficiency wage
models, as they both focus on barriers to underbidding by unemployed
outsiders. While the insider-outsider approach emphasizes the market power
of incumbent workers, the efficiency wage approach stresses the choice
problem of firms that have imperfect information about the productivity of
their employees.49
The reasons for the response of productivity to the real wage vary across
models and include effort, reduced shirking, lower turnover and training
costs, the ability of high-wage firms to screen and obtain a higher-quality
labor force, and improved morale and loyalty.50 Virtually all the literature
with implications for macroeconomics dates from the 1980s. Although most
surveys trace the germ of the idea back three decades to early work on less-
developed countries that posited a linkage among wages, nutrition, and
526 Robert J.Gordon
health (e.g. Leibenstein 1957), the terms efficiency wages and efficiency
earnings appear in Alfred Marshall’s Principles (1920:456–69). Another
precursor of the idea is the negative relationship between wages and quit
rates embedded in Phelps’ (1970) desert island parable and other early
models in the new microeconomic literature. Efficiency wage theory
provides a rare common meeting ground for mainstream and radical
economists, because the far left in US economics has taken the lead in
developing theories of dual labor markets and for setting out policy
proposals for higher minimum wages based on the assumed validity of the
efficiency wage approach.51
The basic efficiency wage result is obtained in a simple model with
identical, perfectly competitive firms and a production function in which
labor input is multiplied by an efficiency factor e that depends on the real
wage. Because the elasticity of e with respect to the real wage declines as the
real wage increases, the first-order conditions require the firm to choose an
optimal real wage rate (w*) at which this elasticity is unity. Workers are
hired up to the point where their marginal product equals the optimal wage
(w*). The intuition of the unit elasticity result is that firms forgo efficiency
gains that yield more than they cost when they pay below w*, while a wage
above w* would cost more than it yields in efficiency gains. Stated another
way, effective labor cost is minimized at w*.
Because w* is completely fixed by whatever factors of taste and technology
that determine the e function, the firm’s reaction to any change in its relative
price (i.e. a demand shock) is to cut employment while maintaining the wage
rate at w*. Firms have no incentive to cut the actual wage, because this
would actually increase their wage bill per unit of output. The extreme result
of a fixed real wage in this model stems from the assumption that a worker’s
efficiency depends on the absolute level of the real wage rather than on the
real wage relative to something else, whether some measure of economy-wide
real earnings or real wages in a perceived peer group or comparison group. A
variant of this approach, in which effort depends on the relative real wage
and on the unemployment rate (a high value of which raises effort by
increasing the cost of job loss), allows the real wage to regain some flexibility
and to depend inversely on the unemployment rate (Shapiro and Stiglitz 1984;
Summers 1988).
Several criticisms of the efficiency wage approach have been offered.52
One line is to propose that job applicants should ‘buy’ high-wage jobs
either by offering lump-sum payments or performance bonds to employers,
or by offering to work in low-wage apprentice status for an initial period.
Efficiency wage proponents point out, however, that unemployed workers
lack sufficient wealth and are risk averse, and that the same monitoring
problems that generate the efficiency wage result also make it unlikely
that banks or other financing sources will come forth to provide finance
for the initial lump-sum payments or performance bonds. This defense
does not rule out low-wage apprenticeships, which are in fact observed,
What is new-Keynesian economics? 527
but these can be interpreted alternatively as a means of sharing the cost of
training rather than as the ‘sale’ of a job by a firm. A second criticism is
that the efficiency wage model is dominated by direct payments to
workers in proportion to their efficiency, whether through piece-rate
contracts or through ‘tournaments’ that pay workers according to their
ranking by performance. The defense against this criticism is similar to
the first. Piece rates and tournaments are subject to information problems.
Workers involved in joint production do not often have a unique claim to
a ‘piece’, while payments to a team invite shirking by some members of
the team. Tournaments are also difficult to implement; there are rarely
many workers in a firm doing exactly the same tasks and no way to rank
across tasks.
Overall, the efficiency wage approach seems to be an essential ingredient
in explaining numerous aspects of microeconomic labor market behavior,
including segmented labor markets, persistent wage differentials for similar
workers that are not equalizing differences, queues for high-paid jobs, and
procyclical fluctuations of the quit rate.53 Variations on the model can explain
why firms sometimes dismiss workers instead of cutting their wage. However,
as a theoretical underpinning of the new-Keynesian paradigm, it suffers from
the same defect as all models of real rigidities. If workers gear their effort to
the real wage, there appears to be no barrier to full wage indexation that
allows firms simultaneously to maintain worker effort through maintenance
of the optimal real wage w*, while changing the nominal wage in tandem
with the nominal price in order to achieve macroeconomic self-correction.
Further, the efficiency wage theory has little to say about the sources of
variations in wage and price responsiveness over time and across countries
that were identified in Section 3.
This negative verdict applies only if a new-Keynesian explanation of
nominal wage and price rigidity is erected on the sole base of the efficiency
wage theory. However, once the input-output approach and the independence
of local and aggregate costs and demand are accepted as the underlying
reason why actual economies do not index to nominal demand, the way is
open to accept the efficiency wage approach as another source of cost rigidity
within the input-output table, of potentially equal importance with the
uncertain evolution of the prices of purchased materials. Once again, we find
that the new-Keynesian approach is most convincing when sources of real and
nominal rigidity are combined rather than when either one or the other is
proposed as the sole explanation.54
Nominal rigidities: wage contract models
Theories of wage stickiness can be based on real rigidities, as in the
approaches outlined above, or on nominal rigidities. The most influential
work that rationalizes nominal rigidities in new-Keynesian labor market
analysis is the staggered-contract approach of Fischer (1977a) and Taylor
528 Robert J.Gordon
(1980). The classification of contract rigidities as nominal is subject to the
preceding criticism—that the negotiation costs that rationalize the existence of
contracts do not rule out fully indexed contracts. The costly negotiations set
the real wage, while the nominal wage is costlessly indexed, preferably to
nominal GNP.
The Fischer-Taylor contract literature is set up entirely in nominal terms
and does not discuss the option of full nominal demand indexation, so we will
discuss it on those terms, as a source of nominal rigidity. In Fischer’s version
the wage for half the workers is set for two periods at the beginning of period
t and for the other half at t+1. The wage set for the first group can respond to
any change in the money supply in the first period but not in the second. The
greater flexibility of nominal money than of nominal wages is an assumption
rather than a result and leads to real effects of perceived monetary
disturbances that cannot occur within the new-classical framework. Fischer’s
version assumes no barriers to price flexibility and market clearing in product
markets. The unemployment his model generates during a period when
money has declined but wages have not declined is classical (because of an
excessive real wage), not Keynesian.
In the setting of an n-period rather than two-period contract model,
Taylor (1980) makes the nominal wage fixed over the life of the contract (at
a level that depends on the expected price and expected output) and setting
the price as a simple markup over the average wage rate. A monetary
disturbance falls fully on output during the period until the next contract
renegotiation. Then wages can adjust quite rapidly, because the dependence
of the negotiated wage on expected future output creates a strong feedback
loop between unemployment and wage behavior. Nevertheless, because
prices depend on wages set in any previous contract still in force, the
duration of the real output response to a nominal monetary shock can last
for much longer than the length of the contract, the same result as was
subsequently derived by Blanchard (1983) for the product market (see
Section 5, pp. 514–15).
Taylor’s approach is sufficiently plausible and important to take
seriously. Yet it is subject to at least two criticisms. First, the assumptions of
staggering and of fixed contract length are arbitrary. In some places
(especially Japan) contract expiration dates among firms in the union sector
are nearly simultaneous. If contract length depends on a balancing of
negotiation costs and the allocative costs of infrequent adjustment, one
would expect changes in contract length in response to the variability of
either local or aggregate shocks. A more general statement of this first
criticism is that the existence of nominal wage contracts is not explained
from the first principles of microeconomics. Models of optimal contracting
do not produce the nominal stickiness generated by the Taylor-type
contracting models. The Taylor approach needs to be supplemented by an
extension to wage setting of work on staggered price setting by Ball and
Romer (1989b) and others, as reviewed above.
What is new-Keynesian economics? 529
The second problem is that, once the contract expires, the adjustment of
the wage in response to expected future output is not complete but is
bounded in Taylor’s model by an arbitrary Phillips-type adjustment
coefficient. As argued by Blanchard (1987a), Taylor’s results require this
adjustment coefficient to be relatively ‘small’ and, if a cyclical response of
the markup of prices over wages is allowed, that must be ‘small’ as well.
While Blanchard’s point suggests that Taylor’s wage adjustment may be too
slow, I would argue the opposite. In particular, Taylor (1983) has claimed
that it is possible for the monetary authorities to engineer a disinflation with
no output loss. However, this result depends heavily on Taylor’s assumption
that the effect of real demand on wage-setting decisions works through
expected future real demand rather than past and current real demand. For
wage setters to use a model to calculate the implications of their current
wage-contract decisions on future real demand requires not only a universal
belief that the announced disinflationary path of nominal demand will be
maintained on target, but also a universal ability to forecast the response of
actual prices to the path of nominal demand, as is required for future real
demand to be predicted.
Where then do staggered wage contracts fit in? We have seen that full price
flexibility for a monopolist requires full flexibility of marginal cost, and
staggered contracts eliminate that full flexibility in the absence of
instantaneous nominal GNP indexation. Barro, Hall, and Mankiw have
argued that it is possible for firms to adjust their prices in proportion to a
change in nominal aggregate demand if wages do not adjust. But this is not
profit maximizing. In almost any model of monopolistic price setting, an
incomplete adjustment of wages implies less than full adjustment of profit-
maximizing prices. In this sense, new-Keynesian theorists have gone over-
board in shifting the emphasis from the labor to the product market.
CONCLUSION
We have stressed throughout the need for new-Keynesian theory to address
the most important elements of variability in the adjustment of prices along
three dimensions, the inertia effect (λ), the rate-of-change effect (α), and the
level effect (γ). The industrial organization literature contributes ample
evidence of differences in rate-of-change effects across industries. It also
shows that intervals of fixed prices lasting months or years in some
industries can coexist with frequent small price changes in other industries.
It stresses that some industries are competitive, some are monopolistic, and
some industries combine monopolistic cartels with competitive price wars.
Within industries, all firms do not exhibit the same price behavior, and
given firms do not even charge the same prices to all customers.
Heterogeneity is rampant, with hundreds of products common in many
industries, and many products combine labor with hundreds or thousands of
purchased components.
530 Robert J.Gordon
The time-series evidence shows a wide variety of price adjustment
patterns across time and countries. The inertia (λ) effect has become more
prevalent since World War II in every country but Japan. The rate-of-change
(α) and level (γ) effects were remarkably similar before World War I and
after World War II in most countries, but exhibited sharp divergences in
between. The rate-of-change effect increased sharply during and after World
War I, while the level effect virtually disappeared during the interwar
period in the US, UK and Germany. The inertia-prone postwar USA is at one
extreme, and Japan throughout most of its history is at the other extreme of
relative flexibility.
A convenient image for understanding the desirable direction of new-
Keynesian theory is a small 1×1 box set next to a gigantic n×n matrix, where
n is measured in the thousands, if not the millions. The small box represents
the identical representative agents of both new-classical models and the
canonical monopolistic competition model, with their i subscripts, and the
practice of treating the macro economy as identical to the representative
agent with the subscripts removed. The gigantic matrix represents the real
world, full of heterogeneous firms enmeshed in a web of intricate supplier-
demander relationships. This n×n matrix suggests two main themes of the
theoretical review in this chapter.
First, the key to introducing theories of real rigidity as a source of nominal
price stickiness is to find a good reason why we do not observe nominal GNP
indexation. That reason is simple, and is at the heart of all good
microeconomics. Individual firms maximize profit by setting their own
marginal cost equal to their own marginal revenue. They have no reason
whatsoever to care about nominal GNP unless it provides useful information
to supplement what they can learn from observing their ‘local’ cost and
demand. There are many reasons for firms to expect their nominal marginal
costs and local demand to contain idiosyncratic elements that cause them to
evolve independently from nominal demand. The most straightforward
argument, which is enough to make the case, is that firms in a small open
economy know that their costs are determined outside the national boundaries
within which domestic nominal demand applies. This principle generalizes to
firms in large open economies, because we know that even under flexible
exchange rates purchasing power parity does not hold over long periods, so
costs of imports and domestically produced import substitutes can evolve
independently of domestic aggregate demand.
The independence of cost and demand, and the input-output table
approach, represent two separate components in the required (but as yet
missing) new-Keynesian analysis that can come to grips with the industry,
cross-time, and cross-country facts summarized here. The idea of
independent cost and demand shocks seems crucial to come to grips with the
time-series evidence. Just as Lucas (1973) argued that Argentina had a more
vertical Phillips curve because agents knew that aggregate demand shocks
dominated local shocks, so we can argue in parallel that the increase in the
What is new-Keynesian economics? 531
rate-of-change coefficient (α) in Table 21.4 for the USA, UK and Japan
during 1915–22 reflected a recognition by price setters that the increasing
importance of aggregate disturbances created a greater than usual
correlation between changes in marginal costs and changes in aggregate
demand. Similarly, the increase in persistence (the λ parameter) observed
almost everywhere after World War II reflects a widespread belief that
government full-employment policies and the end of the gold standard
created an upward drift in prices, leading to the expectation that marginal
costs would have an upward drift and would no longer be a stationary
process.
The input-output component is complementary to the independent shocks
idea, and helps to explain why firms do not simply assume that marginal
costs will move in parallel with aggregate nominal demand: Most firms do
not know the identity of their suppliers, their suppliers’ suppliers, and so on,
because the input-output table is so broad and so deep. The input-output
component of the proposed explanation is required to grapple with the
industry evidence. Prices of corn and wheat on auction markets exhibit sharp
daily swings, subject to administered limits. Prices of strawberries exhibit
frequent sharp weekly swings. Prices of many crude materials exhibit frequent
changes, both small and large. Yet prices of newspapers and many finished
goods can remain unchanged for more than a year. A unified explanation that
explains the degree of volatility of fixity for every product may be impossible
to achieve, but the basic idea that crude materials are relatively volatile and
finished goods relatively fixed seems compatible with the input-output
approach which stresses the number of steps and number of purchased
components that are mixed together with labor input in each final good. The
input-output approach also leaves open a role for a theory of real wage
rigidity, once it is admitted that nominal GNP indexation is unlikely. The
input-output approach emphasizes the time lags in transmitting news of cost
and demand changes back and forth within the input-output table. However,
to explain why prices do not change by small amounts every day, this
approach needs to be supplemented with a plausible mixture of time-
dependent and state-dependent costs of daily price changes.
Once the independence of local costs, local demand, and aggregate
demand is admitted as the fundamental explanation for the lack of nominal
demand indexation, the way is open to take seriously new-Keynesian research
on real rigidities in the labor market. Work on union behavior and on
nominal contracting in the labor market does not appear promising, in light
of the similarity of the α and γ coefficients in most countries before World
War I and after World War II. However, the efficiency wage model has strong
persuasive power as to why firms resist real wage cuts, and the independence
of shocks and input-output table explanations contribute the needed
supplementary explanation of why real wage rigidity becomes translated into
nominal wage rigidity. The other most promising development in the labor
market literature is the insider-outsider approach, if only because the
532 Robert J.Gordon
disenfranchisement of outsiders holds up the best available ray of hope that
we have for understanding why the Phillips-curve level (λ) effect disappeared
in the USA, UK and German interwar periods, and perhaps in some European
countries in the 1980s.
Our perspective that emphasizes independent shocks and the input-output
approach reinforces the view that coordination failures are the essence of
macroeconomic inefficiency in new Keynesian models. Should the
government attempt to intervene to provide the missing coordination of
microeconomic wage and price decisions, or should its activities be limited
to the traditional Keynesian use of monetary and fiscal policy to manipulate
aggregate demand directly? Clearly, traditional forms of internalization
through tax and subsidy policy are infeasible in light of pervasive
heterogeneity among products and decision makings in the millions; to go
in this direction would mean slipping into the quagmire from which Eastern
Europe is trying to emerge. Even mandatory indexation to domestic
nominal demand may be suboptimal in many countries where an important
component of nominal marginal cost is set in foreign currencies and
responds more to foreign than to domestic aggregate demand. This shifts the
ultimate weapon for fighting business cycles back to the traditional
instrument, aggregate demand policy, but not in the form of any old-
fashioned Keynesian bias in favor of fiscal policy. If prices respond slowly
to fluctuations in nominal GNP growth, then the optimal objective of
stabilization policy should be to stabilize the growth rate of nominal GNP
growth. Whether and how this can be achieved is beyond the scope of this
chapter.55
Some commentators (particularly Blanchard 1987a) have lamented that,
far from being a set of facts looking for a theory, the new-Keynesian
paradigm suffers from too many unrelated theoretical explanations. Yet the
essential features emphasized here, the independence of shocks, and the input-
output table, embody a core set of realistic microeconomic elements: a
technology of transactions, heterogeneity of goods and factor inputs,
imperfect competition, imperfect information, and imperfect capital markets.
Unlike time-dependent or place-dependent factors like unions, these essential
features are timeless and placeless. They lead us to expect that the degree of
price flexibility in the early nineteenth century would not be much greater
than today, except insofar as the n×n matrix was smaller, with fewer steps
from primary producer to final consumer, and indeed we find a basic
similarity within each country in the α and γ parameters before World War I
and after World War II.
Recognition of the universality of these imperfections in economic life is
overdue—perhaps a campaign can be started to change economic language so
that these features will be considered the norm, rather than some aberrant or
exotic flower. Rather than thinking of basic aspects of transaction and capital-
market technology as imperfections, perhaps we could all start recognizing
that these features are part of the way that markets function.
What is new-Keynesian economics? 533
But these suggestions represent only the beginning of a needed research
program. At the truly micro-micro level of relations between individual firms
and customers, imperfections go far beyond anything that the independence of
shocks, input-output, or efficiency wage approaches can explain by themselves.
The evidence presented by Carlton that firms charge different prices to different
customers for the same product, and apply nonprice allocation rules differently
across customers, opens up a whole new dimension of heterogeneity that future
theorists will need to consider. The ultimate merger of the new empirical
industrial organization and the new Keynesian macroeconomics (it is hoped not
by leveraged buyout) seems a long way off, but it is a worthy goal to support.
APPENDIX A: THE VARIETY OF HISTORICAL EXPERIENCE:
REGRESSION METHODOLOGY AND ESTIMATES
Specification of regression equations
The aim is to estimate the three parameters λ, α, and γ in equation (9) in the
text, which is repeated here for convenience:
(9)
There may be some concern regarding the close resemblance of (9) to an
identity obtained by rearranging (3):
(a)
Comparing (9) and (a), the former includes pt-1 and excludes . Because
inertia may be absent in some historical eras (λ=0), the difference between (9)
and (a) boils down to the exclusion of . Thus if (9) is a true structural
equation, the identity (a) provides the value of the missing variable, .
This argument is more transparent when (9) is transformed to include
but to exclude :
(9′)
If (9) is a structural relation, so is (9�). Given the values of the right-hand
variables in (9�), two of which are predetermined (pt-1 and ) and one of
which is endogenous ( t), the role of the identity (a) is to determine output as
. In short, the identity shows how output must change,
given the structural price equation (9�). This just restates the basic point about
Keynesian economics: if the current price is predetermined by an equation
like (9�), then the current output level is determined as a residual.
The main estimation problem is not the fact that there is an identity
linking some of the variables in (99), but rather the endogeneity of t, which
we have discussed above in the context of policy feedback. Because the
essence of the problem is policy feedback, there can be no escape by replacing
nominal GNP by the money supply, or by using money as an instrument for
534 Robert J.Gordon
nominal GNP. And alternative versons with real GNP or unemployment are
also subject to bias if policy feedback is not complete, as illustrated in Table
21.1. Our solution, which is to bracket the a parameter by estimating
alternative versions of (9�) with t and t as alternative explanatory
variables, seems to be the best alternative.
To do this, we provide a pair of estimates for each dependent variable
(price change, nominal wage change, and real wage change for the USA, and
price change for the UK, France, Germany, and Japan). The specification for
the first member of each pair is (9�). The specification for the second member
of each pair is the transformation of (9�) that results when identity (a) is used
to replace t by t:
(9�)
The values of the three parameters �, a, and � can be easily unscrambled.
If in (9�) a1 is the estimated coefficient on t, a2 is the estimated coefficient on
pt-1, and a3 is the estimated coefficient on , then the parameters
resulting from the estimation of (9�) are γ=[a3/(1-a3)], α=a1- γ(1-a1), and
λ=a2(1+γ). If in (9�) b1 is the estimated coefficient on t, b2 is the estimated
coefficient on pt-1, and b3 is the estimated coefficient on , then the
parameters resulting from the estimation of (9�) are α=(b1-b3)/(1+b1-b3),
γ=b3(1-α), and λ=b2(1-α).
Data, detrending, and parameter shifts
Postwar data are taken from standard US and OECD sources, and data prior
to World War II are based on Balke and R.Gordon (1989) for the USA,
Feinstein (1972) for the UK, Ohkawa and Shinohara (1979) for Japan, and
national sources as summarized by Maddison (1982) for France and
Germany. Data for Japan, the UK and the US measure nominal GNP, real
GNP, and the GNP deflator. Data for Germany and France prior to World
War II measure real GNP, the CPI, and a hybrid concept of nominal GNP
equal to the CPI times real GNP. The nominal wage equations for the USA
are based on Rees’ data on average hourly earnings in manufacturing linked
in 1960 to the BLS index of average hourly earnings in manufacturing. The
real wage is this nominal wage series divided by the GNP deflator. Data
sources are given in Appendix B.
The use of output data in estimating (9�) and (9�) requires a detrending
procedure to define the , , and variables. Significant variations in
population and productivity growth over the past century prevent the use of a
single trend and require the choice of benchmark years to separate multiple
piece wise log-linear output trends. The choice of the wrong benchmark years
would introduce measurement error into all three of these variables. To avoid
the possible criticism that benchmark years might have been selected to
support or refute a particular hypothesis, all are copied from previous
What is new-Keynesian economics? 535
research directed at other issues.56 The main control for supply shocks is a set
of dummy variables to proxy the effects of government intervention both in
the form of price controls (as during World War II) and intervention to raise
prices and wages, as during the National Recovery Act period in the US
Great Depression. Also for the USA we include a variable to measure the
effect on aggregate inflation of changes in the relative prices of food and
energy.57 The specific values of the supply-shock dummy variables are given
in Appendix B.
The key issue of changing cyclical responsiveness can be addressed by two
alternative methods. One obvious way of providing information on parameter
shifts would be to estimate separate versions of (9�) and (9�) for each major
subperiod within the available data set. An alternative method, carried out
previously in R.Gordon (1983), involves estimating a single equation for the
entire period for which data are available, and then searching for parameter
shifts. If additional variables are defined as the product of the three economic
variables of interest (pt-1, t or t, and ) and ‘0, 1’ dummy variables for
each subperiod, then the t ratios on the additional variables provide estimates
of the statistical significance of parameter shifts. In developing the results
displayed in Tables 21.5 and 21.6, a search procedure was followed in an
attempt to locate parameter shifts during the following subperiods: first year
through 1914, 1915–22, 1923–38 (1930–53 for the US), and 1960–86 (1954–
87 for the US). All of the statistically significant parameter shifts are listed
separately in Tables 21.5 and 21.6. Because of severe declines in output
during wartime and postwar recovery periods, the following years are
omitted from the regression equations: 1914–24 for France and Germany, and
1939–59 for the UK, France, Germany and Japan. No years were omitted for
the USA.
Regression Results
Table 21.5 addresses the issue of changing cyclical responsiveness of prices,
nominal wage rates, and real wage rates in the USA.58 Six columns of
results are shown for the entire 1873–1987 sample period, with equations
for price, nominal wage, and real wage changes presented in pairs. The first
member of each pair uses specification (9�) in which nominal GNP change
( t) appears and the second member uses (9�) in which real GNP change ( t)
appears as an alternative. The separate lines within each group of
explanatory variables report several extra effects, that is, the coefficients on
the product of the variable concerned and a 0, 1 dummy variable for the
period shown. A parallel presentation of results for price-change equations
only is provided in Table 21.6 for the other four countries. The parameters
are unscrambled in Tables 21.3 and 21.4 of the text, and the results are
interpreted in Section 3.
536 Robert J.Gordon
Table 21.5 Equations explaining annual changes in the GNP deflator, the nominal wage
rate, and the real wage rate in the USA, 1873–1987
Notes: Supply-shock variables are defined in Appendix B
* indicates statistically significant at 1 percent level, ** at 5 percent level
Table 21.6 Equations explaining the annual inflation rate, five countries, 1873–1986
What is new-Keynesian economics? 537
APPENDIX B: DATA APPENDIX
United States
GNP, deflator, and food-energy effect
1929–87
Output and prices from National Income and Product Accounts, Tables 1.1
and 7.4., US Department of Commerce. Food-energy effect (1959–87 only) is
the difference between the growth rates of the fixed-weight consumption
deflator and the fixed-weight deflator for consumption expenditures net of
food and energy, from National Income and Product Accounts, Table 7.1.
1869–1928
Balke and Gordon (1989: Table 10).
Notes: Sample period for UK begins in 1958, and for Japan begins in 1888
Supply-shock variables are defined in Appendix B
* Indicates statistically significant at 1 percent level, ** at 5 percent level
538 Robert J.Gordon
Nominal wage rate
1960–87
BLS average hourly earnings in manufacturing, Economic Report of the
President (1989: Table B-44).
1888–1959
Rees’ series on real CPI-deflated average hourly earnings in manufacturing,
series B-70 in Long-term Economic Growth, 1860–1970, US Department of
Commerce, 1973, divided by CPI, series B-69.
Output trend
The output trend is calculated as a log-linear trend between the benchmark
years 1869, 1873, 1884, 1891, 1900, 1910, 1924, and the quarterly data for the
quarters 1949:Q1, 1954:Q1, 1957:Q3, 1963:Q3, 1970:Q2, 1974:Q2, 1979:Q3,
and 1987:Q3. For further details, see R.Gordon (1990: Appendix C).
Dummy variables
World War I: 1918=1.0, 1919–20=0.5. NRA: 1933–34=0.5, 1935–36= -0.5.
World War II: 1943–44=0.5, 1946–47=-0.5. Nixon: 1972–73=0.5, 1974=-0.3,
1975=-0.7.
France and Germany
GNP and prices
1960–86
Real GNP and deflator from OECD Statistics Paris (1988).
1870–1959
Real GNP and CPI from Maddison (1982: Appendices A and E).
Output trend
The output trend is calculated as a log-linear trend between the following
benchmark years. For France: 1870, 1875, 1882, 1892, 1899, 1904, 1912, 1924,
1939, 1951, 1964, 1972, and 1979. For Germany: 1870, 1874, 1884, 1890, 1900,
1907, 1913, 1925, 1928, 1938, 1952, 1961, 1972, and 1979. For both countries,
growth in 1979–86 is calculated by applying the 1972–79 growth in the capital-
output ratio to the observed growth of the capital stock, as in Schultze (1987).
What is new-Keynesian economics? 539
Dummy variables
France Poincare: 1926=1.0.
France Popular Front: 1936–38=0.33.
Hitler controls: 1937–38=0.5.
United Kingdom
GNP and deflator
1960–86 OECD Statistics Paris (1988).
1870–1959 Feinstein (1972).
Output trend
The output trend is calculated as a log-linear trend between the following
benchmark years: 1856, 1865, 1873, 1882, 1889, 1907, 1913, 1920, 1940,
1951, 1961, 1972, 1979, and 1987.
Dummy variables
UK World War I: 1915–18=0.25, 1919–20=-0.5. UK 1972–3 Controls: 1972–
3=0.5, 1974–75=-0.5. UK 1976–7 Social Contract: 1976=1.0, 1980=-1.0.
Japan
GNP and deflator
1960–86 OECD Statistics Paris (1988).
1870–1940 Ohkawa and Shinohara (1979: Tables A9 and A50).
Output trend
The output trend is calculated as a log-linear trend between the following
benchmark years: 1885, 1890, 1903, 1914, 1919, 1929, 1938, 1953, 1961,
1972, 1979, and 1987.
Dummy variable
Japan Oil Shock: 1974=1.0.
540 Robert J.Gordon
ACKNOWLEDGEMENTS
This research was supported by the National Science Foundation. I am
grateful to George Williams for help with the data and to Steven Allen,
Laurence Ball, Olivier J.Blanchard, Timothy Bresnahan, Charles Calomiris,
Dennis Carlton, Robert Chirinko, Russell Cooper, Stanley Fischer, Herschel
Grossman, R.Glenn Hubbard, David Laidler, John Leahy, Assar Lindbeck,
N.Gregory Mankiw, David Romer, Julio Rotemberg, Dennis Snower, John
Taylor, Andrew Weiss, and two anonymous referees for comments on one or
more earlier drafts. I am also indebted to Michael Parkin and Edmund
S.Phelps for helping to establish the etymology of the phrase ‘new-Keynesian’.
NOTES
1 The strongest written statement of the dominance of new-classical macroeconomics
among the younger generation is by Alan Blinder: ‘By about 1980, it was hard to
find an American academic macroeconomist under the age of 40 who professed to
be a Keynesian. That was an astonishing intellectual turnabout in less than a decade—
an intellectual revolution for sure…the young were recruited disproportionately
into the new classical ranks…By 1980 or so, the adage “there are no Keynesians
under the age of 40” was part of the folklore of the (American) economics profession’
(1988:278).
2 The label new-Keynesian should be attributed to Michael Parkin (1982), who has
offered me the opinion that he originated the term new-Keynesian theory, not new-
Keynesian macroeconomics. The term new-Keynesian theory was incorporated into
a chapter subsection in Phelps (1985:562) and ‘new-Keynesian model’ in a chapter
title in the fourth edition of my textbook (Gordon 1990), written in 1986. One of
the first uses of the label new-Keynesian economics in a scholarly article is by Laurence
Ball, N.Gregory Mankiw, and David Romer (1988). The word new rather than neo
to describe the recent work in the classical tradition distinguishes it from what Paul
Samuelson in the early postwar period called the neoclassical synthesis of old-
Keynesian macroeconomics and classical microeconomics. In turn, the word new
rather than neo is used for the recent work in the Keynesian tradition, so that it can
be properly juxtaposed to the new-classical approach.
3 I accept David Laidler’s objection in correspondence that Keynesian economics is
about more than wage and price stickiness and includes a treatment of ‘how the
monetary system interferes with the coordination of inter-temporal choices’. The
new-Keynesian analysis of credit rationing and other failings of the monetary system
is recognized as a legitimate research activity but falls outside the scope of this
chapter, which is delimited by the supply-demand dichotomy.
4 Nevertheless, we recognize the importance of feedback from price behavior to nominal
GNP for the econometric estimation of price adjustment coefficients and devote
considerable emphasis in Section 2 to the treatment of econometric bias that results
from such feedback.
5 This is the title of the survey by Blanchard (1987a).
6 For convenience, this introduction concludes with some references to the many available
surveys that overlap with this chapter, or that treat particular issues in more detail.
Fischer (1988) provides a broadbrush survey of macroeconomics, including demand,
supply, and policy; while Michael Bruno (1988) assesses the classical Keynesian debate
What is new-Keynesian economics? 541
from the perspective of high-inflation countries designing stabilization policies. Olivier
Blanchard (1987a) provides an extended treatment of some of the supply-side issues
that concern us here, whereas Assar Lindbeck (1988) provides a briefer treatment
from a European perspective. Blanchard and Fischer (1989: chs 8–9) provide a relatively
technical exposition of several new-Keynesian models. At the level of specific topics
within the general new-Keynesian rubric, surveys are available on labor market
developments in general (Stiglitz 1986; Katz 1988), implicit contract theory (Rosen
1985), efficiency wage theory (Katz 1986; Weiss 1990), new-Keynesian product-
market theory (Rotemberg 1987), and the interrelations between industrial
organization theory and macroeconomic price stickiness (Carlton 1989a).
7 The Okun’s law relation between detrended output and the unemployment rate
holds very closely in the postwar USA, ensuring that any conclusions developed
here for the relationship between inflation and detrended output carry over to the
relation between inflation and the unemployment rate. For plots of output, trend
output, and the unemployment rate, see R.Gordon (1990:14) for the twentieth
century and p. 324 for 1964–88.
8 In practice, the first-order autoregression on p
t-1
in (7) is too simple to capture the
dynamics in quarterly data, and higher-order autoregressive terms must be included
in regression estimation. In annual data one or two lagged inflation terms are sufficient.
9 Blanchard (1987b) presents an equation like (7) in which the rate of wage change
also appears, because he is interested in the speed of transmission of cost changes
into price changes. But the same point applies to (7), where we are interested in the
division of nominal demand changes between price changes and output changes.
10 The identity in the text, +
t
–p
t
, is identical to the identity written as equation
(3) above, in view of the fact that
t
(the rate of change of detrended output) is the
same as (the change in the log ratio of actual to trend output).
11 Early precursors of (10a) and (10b), developed and originally published in 1972–3,
are reprinted in David Laidler (1975:127, 140) and differ only in assuming that θ=1
and that z
t
=0.
12 The inclusion of both level and rate-of-change effects dates back to Richard Lipsey
(1960), who aggregated a model with heterogeneous micro labor markets
characterized by limited labor mobility between markets and showed that the rate of
change of wages would depend on both the level and rate of change of the aggregate
unemployment rate. In Lipsey’s model the economy exhibits counter-clockwise loops
in a diagram plotting wage change against the level of unemployment, while an
alternative model emphasizing the inertia (λ=1) effect generates clockwise loops.
Barro and Grossman (1976: ch. 5) derive both types of loops as special cases, as well
as the condition for one or the other type of loop to dominate.
13 I am grateful to Robert Chirinko for providing me with a copy of the Tarshis (1939)
note.
14 A valuable compendium of papers on hysteresis, including a fascinating introduction
that traces the history of the term hysteresis in both economics and science, is Cross
(1988). The first use of hysteresis-based models of inflation was by Phelps (1972).
15 I have emphasized the disappearance of the level effect in the US Great Depression in
several of my papers, especially R.Gordon and Wilcox (1981:86–92) and R.Gordon
(1983:93–6).
16 R.Gordon (1982b) shows that the Lucas (1973) model can be nested in a general
model of price adjustment like (9) and can be rejected in the presence of price
inertia.
542 Robert J.Gordon
17 I am grateful to Dennis Carlton for suggesting the wording of the last two
sentences.
18 Further evidence on the extent of product differentiation comes in detailed studies of
international trade, showing the countries at the same stage of development both import
and export goods within the same industrial categories. See Blomstrom et al. (1989).
19 This point was suggested in a letter from Bresnahan, who describes these common
features of cyclically sensitive industries as ‘some famous coincidences about industry
structure.’
20 Particularly striking is Carlton’s example of the costs of running the futures markets
in Chicago, consisting of large office buildings, expensive real estate, elaborate record
keeping, and the large time cost of the many people involved. ‘A significant fraction
of the economy of the city of Chicago is devoted to the making of markets. If a magic
spell could be cast to make transactions costless, the Chicago economy would be
devastated, at least in the short run. This emphasizes how far from costless the
making of markets really is’ (1989b: 6).
21 As we have been reminded by Allen (1989), the standard prewar series on US wages
are for production workers in manufacturing and must be linked with a postwar
series on manufacturing wages, not the wage index for the nonfarm private economy
that is most often used in studies limited to the postwar period. Allen concludes after
an exhaustive study that differences in measurement methods in either wage or
output series do not change his conclusion that the cyclical sensitivity of wages was
the same in the prewar and postwar periods.
22 This finding is consistent with that of Allen’s careful (1989) study, which examines
only wage behavior, not price behavior. Allen’s specification is similar to mine and
uses both unemployment and output gap data, but no nominal GNP data or supply
shock proxies, and is thus subject to a bias in the unemployment or output coefficients
toward zero. Allen’s conclusion claims that his study finds similar behavior prewar
and postwar, but his text reveals that he finds the same increase in the inertia effect
(coefficients on lagged inflation) as is shown in Table 21.3.
23 The respective parameter estimates for 1954–87 in an equation for the change in the
real wage are, with manufacturing wage data, λ=-0.33, α=0.12, γ=-0.12. With the
fixed-weight nonfarm wage index (spliced to the employment cost index in 1975),
the parameters are λ=0.01, α=-0.07, γ=-0.02.
24 See the numerous criticisms of the paper by Ball et al. (1988) contained in the
discussant comments by George Akerlof, Andrew Rose, and Janet Yellen, as well as
by Christopher Sims.
25 The Lucas (1972, 1973) imperfect information approach (Mark I) is now widely
viewed as unconvincing, because it is undermined by the availability of information
on the aggregate price level and money supply over a much shorter time period than
the duration of the average business cycle. Major contributions to real-business-
cycle theory (Mark II) include Kydland and Prescott (1982) and Prescott (1986). A
generally supportive survey is provided by Plosser (1989), and critical surveys include
Mankiw (1989) and McCallum (1989).
26 Simultaneous work by me (R.Gordon 1972) and by Eckstein and Brinner (1972)
showed how postwar wage and price data could be made consistent with longrun
neutrality.
27 My two papers were R.Gordon (1975, 1977). See Laidler and Parkin (1975: esp.
pp. 759–74) for a comprehensive survey of research on wage and price equations in
the late 1960s and early 1970s. Nordhaus (1972) presents a survey of US work on
econometric price markup equations.
28 As author Alan Blinder described the aggregate demand and supply model as
developed in his own textbook, ‘now the Marshallian scissors come in a giant
economy size.’
What is new-Keynesian economics? 543
29 These additional supply-shock factors are omitted in Tables 21.3 and 21.4, as their
effects are hard to discern in century-long annual data samples and instead require
the finer discrimination possible with postwar quarterly data and with the improved
fixed-weight price and wage indexes available only in the postwar period.
30 Depending on the exact price index used and the criterion of what constitutes a
permanent slowdown in the inflation rate, the US sacrifice ratio observed during the
disinflation of the 1980s was between 5 and 7. An estimate of 6.2 was calculated on
the basis of data through 1980 in R.Gordon and King (1982: Table 5, line 3);
reasons for preferring this version were given in that paper (pp. 236–7). Blanchard
(1984) also provides evidence from a quite different specification that the Phillips
curve remained relatively stable during the Volcker disinflation.
31 John Pencavel suggests to me that this critical view by microeconomists is widespread.
32 I am grateful to David Laidler for suggesting this distinction.
33 In defense of the new-Keynesian approach, Andrew Weiss has suggested to me that ‘we
have to solve the partial equilibrium problems first; these also are the most interesting.’
34 Research on general disequilibrium or fixed-price models appears to have become a
specialized European activity in macroeconomics, with near-total invisibility in a
recent survey I conducted of first-year graduate macro reading lists at the top ten
American economics departments.
35 In related work Howitt (1986) calls this effect a ‘thin market externality.’
36 McCallum (1986:409) argues that linking to a constant price index instead of to the
CPI would be chosen only by those agents whose most preferred index is negatively
correlated to the CPI.
37 The model is presented in slightly simplified form in Blanchard and Fischer
(1989:376–81) and Fischer (1988:321–3). An even simpler version with constant
marginal cost is presented by Rotemberg (1987:78–80).
38 The original result was derived by Barro (1972) and is restated by Blanchard and
Fischer (1989:402–5).
39 Laurence Ball disputes this interpretation and claims that ‘the central point [of
recent work] is that nominal rigidity has negative externalities because it exacerbates
fluctuations in real aggregate demand.’ But this is clear as a matter of definition (see
equation (4) on p. 482), is common to any theory of price stickiness, and has
nothing to do with the particular contributions of recent work.
40 This important point credited to Ball and Romer, whose paper was written in 1986,
is summarized and endorsed by Rotemberg (1987:83–5).
41 For this reason I find unconvincing the skepticism of Blanchard and Fischer
(1989:401) that it is possible to derive stable staggered contracting, as when they
write ‘the introduction of stochastic idiosyncratic shocks does not make staggering
more likely.’ Their argument is carried out entirely within an either-or choice between
even and odd dates of price changing, and they show that a 50–50 equilibrium with
the same number of firms choosing each date is unstable, because the slightest tilt in
either direction gives all the other firms an incentive to shift. But with uneven frequency
of shocks and a large number of possible dates of changing, the incentive to shift
disappears. If my optimal frequency of price change is weekly, the fact that there is
bunching with more price changes on 1 January than any of the other 364 days of
the year does not lead me to limit my price changes to once a year on 1 January.
42 I have previously (1981:520ff.) suggested a distinction between aggregate and local
components of both cost and demand with explicit reference to Lucas’ original two-
way classification.
544 Robert J.Gordon
43 Cooper (1989) provides an analysis of the interdependence between wage and price
indexation; the likelihood of wage indexation depends on whether prices are indexed,
and vice versa.
44 Blanchard has written to me that what he means by flat labor supply is ‘the set of real
wages and employment traced out as the marginal product of labor shifts’ and not
‘the competitive labor supply curve.’ The issue here is the possibly misleading choice
of words, not any substantive difference between my interpretation and that of
Blanchard and Fischer.
45 Thus I concur with Barsky and Solon (1989:29–30), who find that procyclical real
wage behavior at the individual level in micro data is consistent with noncyclical
behavior in aggregate data. This pattern reflects a cyclical variation in the ‘employment
opportunities’ (read ‘constraints’) that face both ‘stayers’ at firms who face changing
opportunities to work overtime, and ‘switchers’ who face cyclically in opportunities
for across-firm career advancement.
46 More technically, as pointed out to me by Blanchard, it is the marginal utility of
consumption that is equalized between the employed and unemployed. The ranking
of utility depends on the form of the utility function.
47 This point is made by Blanchard and Fischer (1989:453).
48 To this point the discussion of the insider-outsider model is largely based on several
paragraphs of text kindly contributed by Assar Lindbeck and Dennis Snower.
49 An excellent comparison of the two approaches is provided by Lindbeck and Snower
(1988: ch. 3).
50 Two surveys of the literature that identify those authors and papers who have studied
particular channels of efficiency wage effects are Katz (1986) and Weiss (1990).
51 On dual labor markets, see especially Doeringer and Piore (1971) and David Gordon
et al. (1982). More evidence by mainstream economists is provided by Dickens and
Lang (1985). For policy proposals based on efficiency wage assumptions, see Bowles
et al. (1983). Their policy proposal to raise the minimum wage assumes implicitly
that the current wage is below the optimum efficiency wage, whereas all the work in
the new-Keynesian tradition examines the implications of assuming that the actual
wage is already at the optimum efficiency wage level.
52 This paragraph summarizes Weiss (1990:6–10).
53 The most controversial item on this list is persistent wage differentials, as argued by
Katz and Summers in a series of papers, including Katz (1986) and Katz and Summers
(1989). For a sample of a dissenting view, see Robert Topel’s comment which appears
after the latter paper.
54 In this important conclusion we fully endorse the basic message of Akerlof and
Yellen (1985) and Ball and Romer (1987).
55 Advantages, problems, and techniques relevant to the targeting of nominal GNP growth
are discussed in Tobin (1983), Hall (1984), McCallum (1988) and R.Gordon (1985).
56 For the postwar USA, benchmark years are taken from my macroeconomics textbook
(1990) and for the other four countries from R.Gordon (1988); for the pre-World War
II period, US benchmarks are taken from Romer (1989), for France, Germany, and the
UK from Solomou (1987), and for Japan from R. Gordon (1983). Inconsistency may
result from the use of benchmark years originally selected by varying criteria—peak
output in some cases, average output in others, and the level of output consistent with a
particular unemployment rate in still others. For the European countries, where the
benchmark years before World War I are all peaks (thus eliminating any positive values
of ), the resulting series is adjusted by subtracting its (negative) mean and converting
the mean to zero. This results in a mix of positive and negative values. No such adjustments
are carried out in the interwar or postwar periods.
What is new-Keynesian economics? 545
57 The larger number of such supply-shock variables for the USA than for other
countries may indicate that supply shocks have been more important in the USA,
or they may simply indicate that I am more familiar with the history of the USA
than of the other countries. However, the extra attention given to the USA is
largely due to the inclusion of wartime data for the USA but not for the other
countries, where the years of World War II and its aftermath are excluded for all
four of the other countries, while World War I and its aftermath are excluded for
France and Germany.
58 Here the wage data are adjusted for the trend in productivity growth (using piecewise
linear trends between benchmarks), so that the dependent variable in the columns
labeled Nominal Wage is actually the change in trend unit labor cost, and in the
columns labeled Real Wage is actually the change in labor’s income share adjusted
for cyclical fluctuations in productivity.
REFERENCES
Akerlof, George and Yellen, Janet, ‘A Near-rational Model of the Business Cycle with
Wage and Price Inertia’, Quarterly Journal of Economics 1985 suppl., 100(5), pp.
823–38.
Alchian, Armen, ‘Information Costs, Pricing, and Resource Unemployment’, Western
Economic Journal June 1969, 7(2), pp. 109–28.
Allen, Steven G., ‘Changes in the Cyclical Sensitivity of Wages in the United States, 1891–
1987’, working paper without number, North Carolina State University, Aug. 1989.
Alogoskoufis, George and Smith, Ron, ‘The Phillips Curve and the Lucas Critique: Some
Historical Evidence’, working paper without number, Birkbeck College, London,
Mar. 1989.
Azariadis, Costas, ‘Implicit Contracts and Underemployment Equilibria’, Journal of
Political Economy Dec. 1975, 83(6), pp. 1183–1202.
Backus, David K. and Kehoe, Patrick J., ‘International Evidence on the Historical Properties
of Business Cycles’, working paper without number or place, May 1988.
Baily, Martin Neil, ‘Wages and Unemployment under Uncertain Demand’, Review of
Economic Studies. Jan. 1974, 41(1), pp. 37–50.
Balke, Nathan and Gordon, Robert J., ‘The Estimation of Prewar Gross National Product:
Methodology and New Evidence’, Journal of Political Economy Feb. 1989, 97(1), pp.
38–92.
Ball, Laurence, Mankiw, N.Gregory and Romer, David, ‘The New Keynesian Economics
and the Output-Inflation Trade-off’, Brookings Papers on Economic Activity 1988
(1), pp. 1–65.
Ball, Laurence and Romer, David, ‘Real Rigidities and the Non-Neutrality of Money’,
NBER Working Paper no. 2476, Oct. 1987.
——‘Are Prices Too Sticky?’, Quarterly Journal of Economics Aug. 1989a, 104(3), pp.
507–24.
——‘The Equilibrium and Optimal Timing of Price Changes’, Review of Economic Studies
1989b, 56(2), pp. 179–98.
Barro, Robert J. ‘A Theory of Monopolistic Price Adjustment’, Review of Economic
Studies Jan. 1972, 59(1), pp. 17–26.
——‘Unanticipated Money Growth and Unemployment in the United States’, American
Economic Review Mar. 1977a, 67(2), pp. 101–15.
——‘Long-term Contracting, Sticky Prices, and Monetary Policy’, Journal of Monetary
Economics July 1977b, 3(3), pp. 305–16.
——‘Unanticipated Money, Output, and the Price Level in the United States’, Journal of
Political Economy Aug. 1978, 86(4), pp. 549–80.
546 Robert J.Gordon
——‘Second Thoughts on Keynesian Economies’, American Economic Review May 1979,
69(2), pp. 54–9
Barro, Robert J. and Grossman, Herschel, ‘A General Disequilibrium Model of Income
and Employment’, American Economic Review Mar. 1971, 61(1), pp. 82–93.
——Money, Employment, and Inflation, Cambridge: Cambridge University Press, 1976.
Barro, Robert J. and Rush, Mark, ‘Unanticipated Money and Economic Activity’, in
Rational Expectations and Economic Policy, Stanley Fischer (ed.), Chicago: University
of Chicago Press for NBER, 1980, pp. 23–48, 72–3.
Barsky, Robert and Solon, Gary, ‘Real Wages over the Business Cycle’, NBER Working
Paper no. 2888, Mar. 1989.
Benassy, Jean-Pascal, ‘Neo-Keynesian Disequilibrium Theory in a Monetary Economy’,
Review of Economic Studies Oct. 1975, 42(4), pp. 503–23.
——The Economics of Market Disequilibrium New York: Academic Press, 1982.
Berndt, Ernst R. and Wood, David O., ‘Engineering and Econometric Interpretations of
Energy-Capital Complementarity’, American Economic Review June 1979, 69(3),
pp. 342–54.
Bertola, Giuseppe and Caballero, Ricardo J., ‘Kinked Adjustment Costs and Aggregate
Dynamics’, paper presented at Fifth Annual NBER Conference on Macroeconomics,
9–10, Mar. 1990.
Blanchard, Olivier J., ‘Price Asynchronization and Price Level Inertia’, in Inflation, Debt,
and Indexation, Rudiger Dornbusch and Mario Henrique Simonsen (eds), Cambridge:
MIT Press, 1983, pp. 3–24.
——‘The Lucas Critique and the Volcker Deflation’, American Economic Review May
1984, 74(2), pp. 211–15.
——‘The Wage-Price Spiral’, Quarterly Journal of Economics Aug. 1986, 101(3), pp.
543–65.
——‘Why Does Money Affect Output? A Survey’, NBER Working Paper no. 2285, June
1987a, in Handbook of Monetary Economics, B.Friedman and F.Hahn (eds),
Amsterdam: North-Holland, June 1990.)
——‘Aggregate and Individual Price Adjustment’, Brookings Papers on Economic Activity
1987b, (1), pp. 57–122.
Blanchard, Olivier J. and Fischer, Stanley, Lectures on Macroeconomics, Cambridge:
MIT Press, 1989.
Blanchard, Olivier J. and Kiyotaki, Nobuhiro, ‘Monopolistic Competition and the Effects
of Aggregate Demand’, American Economic Review Sept. 1987, 77(4), pp. 647–66.
Blanchard, Olivier J. and Summers, Lawrence H., ‘Hysteresis and the European
Unemployment Problem’, NBER Macroeconomics Annual, 1986, Cambridge: MIT
Press, pp. 15–78.
Blinder, Alan S., ‘The Fall and Rise of Keynesian Economies’, Economic Record. Dec.
1988, 64, pp. 278–94.
Blomstrom, Magnus, Lipsey, Robert E. and Ohlson, Lennart, ‘What Do Rich Countries
Trade with Each Other? R & D and the Composition of U.S. and Swedish Trade’,
NBER Working Paper no. 3140, Oct. 1989.
Bowles, Samuel, Gordon, David M. and Weisskopf, Thomas, Beyond the Waste Land,
New York: Doubleday, 1983.
Bresnahan, Timothy F., ‘Industries with Market Power’, in Handbook of Industrial
Organization, Richard Schmalensee and Robert Willig (eds), Amsterdam: North-
Holland, 1989, ch. 17, pp. 1011–57.
Bruno, Michael, ‘Theoretical Developments in the Light of Macroeconomic Policy and
Empirical Research’, NBER Working Paper no. 2757, Nov. 1988.
Caplin, Andrew S. and Leahy, John, ‘State-dependent Pricing and the Dynamics of Money
and Output’, working paper without number or place, Oct. 1989.
Caplin, Andrew S. and Spulber, Daniel F., ‘Menu Costs and the Neutrality of Money’,
Quarterly Journal of Economics Nov. 1987, 102(4), pp. 703–25.
What is new-Keynesian economics? 547
Carlton, Dennis W. ‘The Rigidity of Prices’, American Economic Review Sept. 1986,
76(4), pp. 637–58.
——‘The Theory and the Facts of How Markets Clear: Is Industrial Organization Valuable
for Understanding Macroeconomics?’ in Handbook of Industrial Organization, vol.
I, R.Schmalensee and R.D.Willig (eds), Amsterdam: North-Holland 1989a, pp. 909–
46.
——‘The Theory of Allocation and Its Implications for Marketing and Industrial
Structure’, unnumbered working paper, University of Chicago Graduate School of
Business, May 1989b.
Cecchetti, Stephen, ‘The Frequency of Price Adjustment: A Study of the Newsstand
Prices of Magazines, 1953 to 1979’, Journal of Econometrics 1986, 31(3), pp. 255–
74.
Clower, Robert W. ‘The Keynesian Counterrevolution: A Theoretical Appraisal’, in The
Theory of Interest Rates, F.H.Hahn and F.Brechling (eds), London: Macmillan, 1965,
pp. 103–25.
Coe, David T., ‘Hysteresis and Insider-Outsider Influences on Industry Wages’, working
paper without number or place, Mar. 1989.
Cooper, Russell, ‘Predetermined Wages and Prices and the Impact of Expansionary
Government Policy’, working paper without number or place, Sept. 1989.
Cooper, Russell and John, Andrew, ‘Coordinating Coordination Failures in Keynesian
Models’, Quarterly Journal of Economics Aug. 1988, 100(3), pp. 441–63.
Cross, Rod (ed.) Unemployment, Hysteresis, and the Natural Rate Hypothesis, Oxford
and New York: Basil Blackwell, 1988.
DeLong, J.Bradford and Summers, Lawrence H., ‘The Changing Cyclical Variability of
Economic Activity in the United States’, in The American Business Cycle: Continuity
and Change, Robert J.Gordon, (ed.) Chicago: University of Chicago Press for NBER,
1986, pp. 679–734.
Diamond, Peter A., ‘Aggregate Demand Management in Search Equilibrium’, Journal of
Political Economy Oct. 1982, 90(5), pp. 881–94.
——A Search Equilibrium Approach to the Micro Foundations of Macroeconomics,
Cambridge: MIT Press, 1984.
Dickens, William T. and Lang, Kevin, ‘A Test of Dual Labor Market Theory’, American
Economic Review Sept. 1985, 75(4), pp. 792–805.
Doeringer, Peter B. and Piore, Michael J., Internal Labor Markets and Manpower Analysis,
Lexington, MA: D.C. Heath, 1971.
Dunlop, John T., ‘The Movements of Real and Money Wage Rates’, Economic Journal
Sept. 1938, 48(191), pp. 413–34.
——Wage Determination under Trade Unions, New York: Macmillan, 1944.
Durlauf, Steven N., ‘Output Persistence, Economic Structure, and the Choice of
Stabilization Policy’, Brookings Papers on Economic Activity 1989, 20(2), pp. 69–
116.
Eckstein, Otto and Brinner, Roger, The Inflation Process in the United States, A Study
Prepared for the use of the Joint Economic Committee, 92 Cong. 2 session, 1972.
——Economic Report of the President 1989, Washington, DC: US GPO, 1989.
Feinstein, C.H. National Income Expenditure and Output of the United Kingdom,
1855–1965, Cambridge: Cambridge University Press, 1972.
Fischer, Stanley. ‘Long-term Contracts, Rational Expectations, and the Optimal Money
Supply Rule’, Journal of Political Economy. Feb. 1977a, 85(1), pp. 191–205.
——‘Wage Indexation and Macroeconomic Stability’, in Stabilization of the Domestic
and International Economy, Karl Brunner and Allan Meltzer (eds), Carnegie-Rochester
Conference Series on Public Policy, vol. 5. Amsterdam: North-Holland 1977b, pp.
107–47.
——Indexing, Inflation, and Economic Policy, Cambridge: MIT Press, 1986.
548 Robert J.Gordon
——‘Recent Developments in Macroeconomics’, Economic Journal. June 1988, 95(391),
pp. 294–339.
Friedman, Milton, ‘The Role of Monetary Policy’, American Economic Review Mar.
1968, 58(1), pp. 1–17.
Friedman, Milton and Schwartz, Anna, A Monetary History of the United States, 1867–
1960, Princeton, NJ: Princeton University Press for NBER, 1963.
Gordon, David M., Edwards, Richard and Reich, Michael, Segmented Work, Divided
Workers: The Historical Transformation of Labor in the United States, New York:
Cambridge University Press, 1982.
Gordon, Donald F., ‘A Neo-Classical Theory of Keynesian Unemployment,’ Economic
Inquiry Dec. 1974, 12(4), pp. 431–59.
Gordon, Robert J., ‘Wage-Price Controls and the Shifting Phillips Curve’, Brookings
Papers on Economic Activity 1972 (2), pp. 385–421.
——‘The Impact of Aggregate Demand on Prices’, Brookings Papers on Economic
Activity 1975 (3), pp. 613–62.
——‘Can the Inflation of the 1970s Be Explained?’, Brookings Papers on Economic
Activity 1977 8(1), pp. 253–77.
——‘A Consistent Characterization of a Near-Century of Price Behavior’, American
Economic Review May 1980, 70(2), pp. 243–49.
——‘Output Fluctuations and Gradual Price Adjustment’, Journal of Economic Literature
June 1981, 19(2), pp. 493–530.
——‘Why U.S. Wage and Employment Behavior Differs from that in Britain and Japan’,
Economic Journal Mar. 1982a, 92(365), pp. 13–44
——‘Price Inertia and Policy Ineffectiveness in the United States, 1890–1980’, Journal of
Political Economy Dec. 1982b, 90(6), pp. 1087–1117.
——‘A Century of Evidence on Wage and Price Stickiness in the United States, the United
Kingdom, and Japan’, in Macroeconomics, Prices, and Quantities, James Tobin (ed.),
Washington, DC: Brookings Institution, 1983, pp. 85–121.
——‘The Conduct of Domestic Monetary Policy’, in Monetary Policy in Our Times,
Albert Ando et al. (eds), Cambridge: MIT Press, 1985, pp. 45–81.
——‘Wage Gaps vs. Output Gaps: Is There a Common Story for All of Europe?’, in
Macro and Micro Policies for More Growth and Employment: Keil Symposium,
Herbert Giersch (ed.), Tubingen: J.C.B.Mohr [Paul Siebeck] 1988, pp. 97–151.
——Macroeconomics, 5th edn Glenview, IL: Scott, Foresman, 1990. Gordon, Robert J.
and King, Stephen R., ‘The Output Cost of Disinflation in Traditional and Vector
Autoregressive Models’, Brookings Papers on Economic Activity 1982 13(1), pp.
205–42.
Gordon, Robert J. and Wilcox, James, ‘Monetarist Interpretations of the Great Depression:
An Evaluation and Critique’, in The Great Depression Revisited, Karl Brunner, (ed.)
Boston, MA: Martinus Nijhoff, 1981, pp. 49–107.
Gray, Jo Anna, ‘Wage Indexation: A Macroeconomic Approach’, Journal of Monetary
Economics Apr. 1976, 2(2), pp. 221–35.
Hall, Robert E., ‘Employment Fluctuations and Wage Rigidity’, Brookings Papers on
Economic Activity 1980 (1), pp. 91–123.
——‘Monetary Policy with an Elastic Price Standard’, in Price Stability and Public Policy,
Federal Reserve Bank of Kansas City, 1984, pp. 137–59.
Howitt, Peter, ‘The Keynesian Recovery’, Canadian Journal of Economics Nov. 1986,
19(4), pp. 626–41.
Kashyap, Anil. ‘Sticky Prices: New Evidence from Retail Catalogs’, Finance and Economics
Discussion Series Working Paper no. 112, Federal Reserve Board, Washington, DC,
Mar. 1990.
Katz, Lawrence F., ‘Efficiency Wage Theories: A Partial Evaluation’, NBER Macroeconomics
Annual, 1986, 1, Cambridge: MIT Press, pp. 235–75.
What is new-Keynesian economics? 549
——‘Some Recent Developments in Labor Economics and Their Implications for
Macroeconomics’, Journal of Money, Credit, Banking Aug. 1988, 20(3), pp. 507–
22.
Katz, Lawrence F. and Summers, Lawrence H., ‘Industry Rents: Evidence and Implications’,
Brookings Papers on Economic Activity.: Microeconomics 1989, 1989, 51, pp. 209–
90.
Keynes, John M., ‘Relative Movements of Real Wages and Output’, Economic Journal.
Mar. 1939, 49(193), pp. 34–51.
Kydland, Finn E. and Prescott, Edward C. ‘Time to Build and Aggregate Fluctuations’,
Econometrica Nov. 1982, 50(6), pp. 1345–70.
Laidler, D.E.W. Essays on Money and Inflation, Manchester: Manchester University
Press, 1975.
Laidler, D.E.W. and Parkin, J.Michael, ‘Inflation—A Survey’, Economic Journal Dec.
1975, 85(340), pp. 741–809.
Layard, Richard and Nickell, Stephen, ‘The Labour Market’, in The Performance of the
British Economy, Rudiger Dornbusch and Richard Layard (eds), Oxford: Clarendon
Press, 1987, pp. 131–79.
Leibenstein, Harvey, ‘The Theory of Underemployment in Backward Economies’, Journal
of Political Economy. Apr. 1957, 65(2), pp. 91–103.
Leijonhufvud, Axel, On Keynesian Economics and the Economics of Keynes, New York:
Oxford University Press, 1968.
——Information and Coordination: Essays in Macroeconomic Theory, New York:
Oxford University Press, 1981.
Lindbeck, Assar, ‘Remaining Puzzles and Neglected Issues in Macroeconomics’, Institute
for International Economic Studies (University of Stockholm), Seminar Paper no.
424, Dec. 1988.
Lindbeck, Assar and Snower, Dennis, ‘Wage Setting, Unemployment, and Insider-Outsider
Relations’, American Economic Review., May 1986, 76(2), pp. 235–39.
——The Insider-Outsider Theory of Employment and Unemployment, Cambridge: MIT
Press, 1988.
——‘Remaining Puzzles and Neglected Issues in Macroeconomics’, Scandinavian Journal
of Economics 1989, 91(2), pp. 495–516.
Lipsey, Richard G., ‘The Relation between Unemployment and the Rate of Change of
Money Wage Rates in the United Kingdom, 1862–1957: A Further Analysis’,
Economica n.s., Feb. 1960, 27, pp. 1–31.
Long-term Economic Growth, 1860–1970, Washington, DC: US Dept. of Commerce,
1973.
Lucas, Robert E. Jr. ‘Expectations and the Neutrality of Money’, Journal of Economic
Theory Apr. 1972, 4(2), pp. 103–24.
——‘Some International Evidence on Output-Inflation Tradeoffs’, American Economic
Review June 1973, 63(3), pp. 326–34.
——‘Econometric Policy Evaluation: A Critique’, in The Phillips Curve and Labor Markets
, Karl Brunner and Allan Meltzer (eds), Amsterdam: North-Holland, Carnegie-
Rochester Series on Public Policy, a supplementary series to Journal of Monetary
Economics Jan. 1976, 1, pp. 19–46.
Lucas, Robert E. Jr and Sargent, Thomas J. ‘After Keynesian Macroeconomics’, in After
the Phillips Curve: Persistence of High Inflation and High Unemployment, Boston,
MA: Federal Reserve Bank of Boston, 1978, pp. 49–72.
Maddison, Angus, Phases of Capitalist Development, Oxford: Oxford University Press,
1982.
Mankiw, N.Gregory, ‘Small Menu Costs and Large Business Cycles: A Macroeconomic
Model’, Quarterly Journal of Economics May 1985, 100(2), pp. 529–38.
——‘Recent Developments in Macroeconomics: A Very Quick Refresher Course’, Journal
of Money, Credit and Banking Aug. 1988, 20(3, part 2), pp. 436–49.
550 Robert J.Gordon
——‘Real Business Cycles: A New Keynesian Perspective’, Journal of Economic
Perspectives, Summer 1989, 3(3), pp. 79–90.
Marshall, Alfred, Principles of Economics, London: Macmillan, 1920.
McCallum, Bennett T., ‘On “Real” and “Sticky-Price” Theories of the Business Cycle’ ,
Journal of Money, Credit and Banking Nov. 1986, 18(4), pp. 398–414.
——‘Postwar Developments in Business Cycle Theory: A Moderately Classical
Perspective’, Journal of Money, Credit, and Banking Aug. 1988, 20(3, Part 2), pp.
459–71.
——‘Real Business Cycle Theory’, in Modern Business Cycle Theory, Robert J. Barro
(ed.), Cambridge: Harvard University Press, 1989, pp. 16–50.
McDonald, Ian and Solow, Robert M., ‘Wage Bargaining and Employment’, American
Economic Review Dec. 1981, 71(5), pp. 896–908.
——‘Wages and Employment in a Segmented Labor Market’, Quarterly Journal of
Economics Nov. 1985, 100(4), pp. 1115–41
Means, Gardiner C., ‘Industrial Prices and their Relative Inflexibility’, US Senate Document
13, 74th Congress, 1st session, Washington, DC, 1935.
Mills, Frederick C. The Behavior of Prices NBER General Series, no. 11. New York: Arno
Press, 1927.
Mitchell, Daniel J.B. ‘Wage Flexibility: Then and Now’, Industrial Relations Spring 1985,
24, pp. 266–79.
Mortensen, Dale T, ‘A Theory of Wage and Employment Dynamics’, in Phelps, 1970, pp.
167–211.
——‘Job Search, the Duration of Unemployment, and the Phillips Curve’, American
Economic Review Dec. 1970b, 60(5), pp. 847–62.
National Income and Product Accounts of the United States, 1929–82. Washington,
DC: US Dept. of Commerce, 1986. (Updated beyond 1982 from Survey of Current
Business, various issues)
Nordhaus, William D., ‘Recent Developments in Price Dynamics’, in The Econometrics
of Price Determination, Otto Eckstein (ed.), Washington, DC: Board of Governors of
the Federal Reserve System, 1972, pp. 16–49.
OECD Statistics. Paris, 1988. (Data obtained from data diskette.)
Ohkawa, Kazushi, and Shinohara, Mihohei (eds) Patterns of Japanese Economic
Development: A Quantitative Appraisal. New Haven: Yale University Press, 1979.
Okun, Arthur M., ‘Inflation: Its Mechanics and Welfare Cost’, Brookings Papers on
Economic Activity, 1975 (2), pp. 351–401.
——Prices and Quantities: A Macroeconomic Analysis. Washington, DC: Brookings
Institution, 1981.
Parkin, Michael, Macroeconomics. Scarborough, Ontario: Prentice-Hall, 1984. Patinkin,
Don, Money, Interest, and Prices, 2nd edn, New York: Harper & Row, 1965. Phelps,
Edmund S., Microeconomic Foundations of Employment and Inflation Theory, New
York: Norton, 1970.
——Inflation Policy and Unemployment Theory: The Cost-benefit Approach to Monetary
Planning, New York: Norton, 1972.
—— Political Economy: An Introductory Text, New York: W.W.Norton, 1985.
Phelps, Edmund S. and Taylor, John B., ‘Stabilizing Powers of Monetary Policy Under
Rational Expectations’, Journal of Political Economy Feb. 1977, 85(1), pp. 163–90.
Phelps, Edmund S. and Winter, Sidney G.Jr, ‘Optimal Price Policy Under Atomistic
Competition’, in Phelps, 1970, pp. 309–37.
Phillips, A.W., ‘The Relation between Unemployment and the Rate of Change of Money
Wage Rates in the United Kingdom, 1861–1957’, Economica Nov. 1958, 25, pp.
283–99.
Plosser, Charles I., ‘Understanding Real Business Cycles’, Journal of Economic
Perspectives, Summer 1989, 3(3), pp. 51–77.
What is new-Keynesian economics? 551
Prescott, Edward C., ‘Theory Ahead of Business Cycle Measurement’, Carnegie-Rochester
Conference Series on Public Policy, Autumn 1986, 25, pp. 11–44.
Roberts, John M., Stockton, David J. and Struckmeyer, Charles S., ‘An Evaluation of the
Sources of Aggregate Price Rigidity’, Working Paper 99, Economic Activity Section,
Division of Research and Statistics, Board of Governors of the Federal Reserve System,
May 1989.
Romer, Christina, ‘The Prewar Business Cycle Reconsidered: New Estimates of Gross National
Product, 1869–1908’, Journal of Political Economy Feb. 1989, 97(1), pp. 1–37.
Rosen, Sherwin, ‘Implicit Contracts: A Survey’, Journal of Economic Literature Sept.
1985, 23(3), pp. 1144–75.
Rotemberg, Julio, ‘The New Keynesian Microfoundations’, NBER Macroeconomics
Annual 1987, 2, pp. 69–104.
Sachs, Jeffrey D., ‘The Changing Cyclical Behavior of Wages and Prices: 1890–1976’,
American Economic Review., Mar. 1980, 70(1), pp. 78–90.
Sargent, Thomas J., ‘A Note on the “Accelerationist” Controversy,’ Journal of Money,
Credit and Banking Aug. 1971, 3(3), pp. 721–25.
Schultze, Charles, ‘Some Macro Foundations for Micro Theory,’ Brookings Papers on
Economic Activity 1981, (2), pp. 521–76.
——Other Times Other Places: Macroeconomic Lessons from U.S. and European History
Washington, DC: Brookings Institution, 1986.
——‘Real Wages, Real Wage Aspirations, and Unemployment in Europe’, in Barriers to
European Growth: A Transatlantic View, Robert Z.Lawrence and Charles L.Schultze
(eds), Washington, DC: Brookings Institution, 1987, pp. 230–91.
Shapiro, Carl and Stiglitz, Joseph, ‘Equilibrium Unemployment as a Discipline Device’,
American Economic Review June 1984, 64(3), pp. 433–44.
Sheshinski, Eytan and Weiss, Yoram, ‘Inflation and Costs of Price Adjustment’, Review
of Economic Studies June 1977, 44(2), pp. 287–303.
——‘Optimum Pricing Policy under Stochastic Inflation’, Review of Economic Studies.
July 1983, 50(3), pp. 513–29.
Solomou, Solomos, Phases of Economic Growth, 1850–1973, Cambridge: Cambridge
University Press, 1987.
Stigler, George and Kindahl, James K., The Behavior of Industrial Prices, New York:
Columbia University Press for the NBER, 1970.
Stiglitz, Joseph, ‘Theories of Wage Rigidity’, in Keynes’ Economic Legacy: Contemporary
Economic Theories, James L.Butkiewicz, Kenneth J.Koford, and Jeffrey B.Miller (eds),
New York: Praeger, 1986, pp. 153–206.
Summers, Lawrence H., ‘Relative Wages, Efficiency Wages, and Keynesian Unemployment’,
American Economic Review May 1988, 78(2), pp. 383–8.
Tarshis, Lorie, ‘Real Wages in the United States and Great Britain’, Canadian Journal of
Economics Aug. 1938, 4(3), pp. 362–76.
——‘Changes in Real and Money Wages’, Economic Journal Mar. 1939, 49(193), pp.
150–54.
Taylor, John, ‘Aggregate Dynamics and Staggered Contracts’, Journal of Political Economy
Feb. 1980, 88(1), pp. 1–24.
——‘Union Wage Settlements’, American Economic Review Dec. 1983, 73(5), pp. 981–93.
——‘Improvements in Macroeconomic Stability: The Role of Wages and Prices’, in The
American Business Cycle: Continuity and Change, Robert J.Gordon (ed.), Chicago:
University of Chicago Press for NBER, 1986, pp. 639–77.
Tobin, James, ‘Monetary Policy: Rules, Targets, and Shocks’, Journal of Money, Credit
and Banking , Nov. 1983, 15(4), pp. 506–18.
——‘On the Theory of Macroeconomic Policy’, Cowles Foundation Discussion Paper
no. 931, Dec. 1989.
Weiss, Andrew, Efficiency Wages: Models of Unemployment, Layoffs, and Wage Dispersion,
Princeton, NJ: Princeton University Press, 1990.
22 New and old Keynesians
Bruce Greenwald and Joseph Stiglitz
Journal of Economic Perspectives (1993) 7, Winter, pp. 23–44
All Keynesians, whether new or old, would agree on three propositions. First,
during some periods—often extended—an excess supply of labor exists at the
prevailing level of real wages (and expectations concerning future wages and
prices).
Second, the aggregate level of economic activity fluctuates markedly,
whether measured by capacity utilization, GDP, or unemployment. These
fluctuations are greater in magnitude and different in pattern from any that
might be accounted for by short-run changes in technology, tastes, or
demography.
Third, money matters, at least most of the time, although monetary policy
may be ineffective in some periods (like the Great Depression).
From these three propositions follow certain important policy conclusions;
while old and new Keynesians may disagree upon the exact form of their
policy recommendations, they would agree generally that government
intervention is at least sometimes (many would argue frequently) desirable to
stabilize the level of economic activity.
Agreement upon these three propositions, and the associated policy
perspective, sets old and new Keynesians apart from advocates of other major
schools of macroeconomic thought, including new classical and real business
cycle theorists. Both of these, for instance, believe that the labor market and
other markets essentially always clear, with wages and prices adjusting
quickly to any disturbances; that shifts in the demand or supply curves for
labor can explain fluctuations in observed levels of employment; and that the
economy’s (presumably efficient) responses to shocks can explain these
fluctuations in output. In the case of real business cycles, the focus is on
shocks to technology; for many new classical theories, the focus is shocks to
the money supply.
Despite the fundamental differences in views between these different
schools, they have agreed upon two methodological premises: that
macroeconomics should be grounded in microeconomic principles, and
that understanding macroeconomic behavior requires the construction of a
(simple) general equilibrium model. The real difference arises here: real
business cycles and (to a lesser extent) new classical economists base their
New and old Keynesians 553
theories on simple (we would say simplistic) models of markets that
employ perfect information, perfect competition, the absence of
transactions costs, and the presence of a complete set of markets. They
also often employ a representative agent model.1 These assumptions often
interact: the absence of risk markets is of no import in a world in which
a l l i n d i v i d u a l s a r e i d e n t i c a l — s i n c e t h e r e i s n o o n e t o w h o m a
representative agent can transfer risk. Problems of asymmetric information
cannot arise if all individuals are identical. Moreover, the strong
assumptions allow market results to be Pareto efficient, despite the fact
that economies with imperfect information and incomplete markets are
generally not constrained Pareto efficient (Greenwald and Stiglitz 1986,
1988a).2 In contrast, modern Keynesians have identified these real world
‘imperfections’ as the source of the problem: leaving them out of the
model is like leaving Hamlet out of the play.
The insistence on micro-foundations enhances the ability of economists to
distinguish among alternative theories, and helps to set the research agenda.
Statistical analyses based on variances and covariances of the principle
aggregate time series simply do not have enough power to distinguish among
many of the alternative theories. Good macro-theories should do more. A host
of other facts clamor to be explained; for instance, good macro-theories must
explain why variations in the number of hours worked should take the form
of layoffs rather than work-sharing; why layoffs tend to be concentrated
among certain parts of the labor force; why investment in general, and
inventories and construction in particular, should be so volatile; and more.
Beyond that, the micro-foundations from which the aggregate behavior is
derived can often be tested directly. A rejection of the underlying micro-
hypotheses should suffice to cast doubt on the validity of the derived macro-
theory.
Incorporating the newer micro-foundations is the principal task ahead of
new Keynesians.3 The challenge is to choose between the myriad of ways in
which markets can be imperfect, and to decide on the central questions and
puzzles to be explained.
Different strands of research within new Keynesian economics have taken
two broadly different approaches.4 The first argues that nominal price
rigidities are the essential way in which market economies differ from the
Walrasian Arrow-Debreu model. Without such rigidities, the argument goes,
flexible prices would allow the economy to adjust quickly to whatever shocks
it experiences, maintaining all the while full employment and economic
efficiency. Early work in this area focused on constructing general
equilibrium models with price rigidities.5 More recent work has been
concerned with explaining the sources of those price rigidities, as discussed by
Romer (1993).
The second strand of new Keynesian literature explores another path
suggested by Keynes: that increased flexibility of wages and prices might
exacerbate the economy’s downturn. This insight implies that wage and price
554 Bruce Greenwald and Joseph Stiglitz
rigidity are not the only problem, and perhaps not even the central problem.
This view holds that even if wages and prices were perfectly flexible, output
and employment would be highly volatile. It sees the economy as amplifying
the shocks that it experiences, and making their effects persist. It identifies
incomplete contracts, and, in particular, imperfect indexing, as central market
failures, and it attempts both to explain the causes and consequences of these
market failures.
Clearly, these two new Keynesian approaches have different implications
for how the economy works. The first holds that the classical dichotomy
breaks down, allowing monetary policy to have effects other than on the price
level, because nominal prices are at least somewhat rigid throughout the
economy. The second approach, however, holds that monetary policy has real
effects even when wages and prices are flexible.
In addition, the nominal price rigidity theories describe how the economy
will recover from a recession as wages and prices eventually fall enough that
consumption recovers, or as capital goods wear out to the point where gross
investment is required to replace even the small amount of capital required
for the low level of output. However, neither the sources of the shocks, nor the
mechanisms by which falling prices and wages would restore the economy to
equilibrium, have received extensive attention; implicitly, in most of the
models, it appears as a hidden real balance effect—as wages and prices fall,
the real value of individuals’ holdings of money increases, and this induces
them to consume more.
The new Keynesian view that emphasizes price flexibility suggests an
alternate and more complex perspective: first, that natural economic forces
can magnify economic shocks that may seem small, and second, that existing
price rigidities may reduce the magnitude of the fluctuations, as Keynes
argued.6 Since even with perfectly flexible wages and prices, the economy
could experience substantial variations in employment, they believe the
single-minded focus on price and wage rigidities is misguided. And since
small disturbances can give rise to large effects, there is less concern about
identifying the source of the disturbance: in one case, it may be a supply
shock (the oil price shocks of 1973 and 1979), in another case it may be a
monetary shock (the Volcker recession).
BASIC INGREDIENTS
The purpose of this chapter is to describe the second strand of new Keynesian
literature and to contrast it both with the alternative strand of new Keynesian
literature based on price rigidities as well as with other points of view. The
models described here contain three basic ingredients, each playing a different
role in explaining aspects of the underlying macroeconomic quandaries, but
all based on problems which arise in economies with imperfect information
and incomplete contracts. The ingredients are: risk averse firms; a credit
allocation mechanism in which credit-rationing, risk averse banks play a
New and old Keynesians 555
central role; and new labor market theories, including efficiency wages and
insider-outsider models. These building blocks should help to explain how
price flexibility contributes to macroeconomic fluctuations and to
unemployment. In particular, the first two building blocks will explain why
small shocks to the economy can give rise to large changes in output, while
the new labor market theories will explain why those changes in output (with
their associated changes in the demand curve for labor) result in
unemployment.
RISK AVERSE FIRMS
Much of the macroeconomic behavior of firms can be explained by the fact
that firms are risk averse. Let us first explore several alternative theories as to
why firms are risk averse, and then examine the consequences of that
finding.7
A first explanation for risk averse firms has to do with imperfections in the
equity market. In traditional Keynesian theory, whether finance came from
equity or debt was not important. In our view, it is central. With equity, the
firm shares risk with those who provide finance, and the firm has no fixed
obligation to repay. With debt, the firm has a fixed obligation, and if it fails
to meet those obligations, it can be forced into bankruptcy. Thus, firms will
tend to be risk averse if they do not have ready access to equity finance, and
are therefore pushed to debt finance.
In fact, despite the seeming advantages of equity, firms finance a relatively
small fraction of their investment with new equity issues. One obvious
explanation is when firms do issue new equities, their market values tend to
decline markedly, because the market interprets issuing new shares as a
negative signal. Think of it this way: assume the owner of a firm knew the
value of the company. Then auctioning off shares in the firm is no different
from auctioning off dollar bills. If I know the number of dollar bills in my
back pocket, and auction off 1 percent shares, what is the equilibrium price?
Zero! And for an obvious reason. If there are $100,000 in my back pocket,
and you offer me less than $1000 for a 1 percent share, then I will not accept
the offer; if you offer me more, I will. The only price at which you will not
lose is a price of zero.
So how can markets for issuing new equity exist at all, in the presence of
asymmetric information? Owners of firms are risk averse, and do not have
perfect information about the value of their firm. Provided it is not too costly,
they would like to sell some of their shares and diversify their risk. But the
adverse selection effect still works with a vengeance. Those who know that
the market overvalues their shares are most anxious to sell additional shares.
Accordingly, in a rational expectations equilibrium, the ‘worst firms’ (most
overvalued, or least undervalued) are most willing to issue equities; and,
given that, issuing equity will be treated as a negative signal and the equity
market will be thin.
556 Bruce Greenwald and Joseph Stiglitz
Investors may also be generally leery of equity because of its effect on
incentives. An early version of this argument, using principal-agent theory,
pointed out that equity means that management must share the returns of its
efforts with others (Ross 1973; Stiglitz 1974). A more recent, probably more
important effect is what Robert Hall refers to as the ‘backs to the wall theory’
of corporate finance, or what Jensen (1986) refers to as the ‘free cash flow’
hypothesis. In these theories, the fixed obligations entailed by high debt
obligations can provide strong managerial incentives.
The literature offers a number of other reasons why firms may be risk
averse; the discussion here is not meant as exhaustive.8 For example, one
major strand of literature emphasizes that modern corporations are
controlled by managers who act in a risk averse manner. While managerial
incentive schemes may attempt to reduce this behavior, they do so only
imperfectly.
At this juncture, many a macroeconomist may ask: while all of this is
interesting micro-theory, what does it have to do with macroeconomics? To
answer this challenge, we have to describe a bit more how risk aversion
affects firm behavior.
A risk averse firm will be sensitive to the risk associated with any action
(including inaction). Production itself is risky; it takes time and there are no
future markets for the sale of goods. Firms are often uncertain about the
consequences of their actions (so-called ‘instrument uncertainty’) and their
uncertainty grows with the size of the change. In general, firms know more
about the status quo than about what things might be like if they changed
their actions.
The risk averse nature of firms under these conditions of uncertainty is the
basis of the ‘portfolio theory’ of the firm, in which firms simultaneously
choose all of their actions—prices, wages, employment, production, and so
on—taking into account the risk (covariances as well as variances) and
expected returns with each ‘portfolio’ of decisions. In assessing the
consequences of various actions, firms look at the effects that those actions
will have on the firm’s assets, which include cash, a set of machines, a group
of employees, a set of customers, and so on. Changes in economic
circumstances—either the firm’s willingness to bear risks, or its perceptions
concerning the riskiness or value of various assets, will lead it to want to
change that portfolio; for instance, increased uncertainty about the value of
inventories will lead it to want to hold smaller inventories.
Changes in the economic environment will in general necessitate changes
in some actions of the firm. Thus, if the demand curve for the firm’s product
shifts to the left, it must either change the price it charges, the quantity it
sells, or the inventories it holds. If it holds price constant, the quantity sold
must adjust, and conversely. Evaluating what it should do entails an
evaluation of the risks associated with each of these changes and the costs of
adjustment.
New and old Keynesians 557
The actions of firms are affected by their perceptions of risks, both
through instrument uncertainty (the uncertainty concerning the consequences
of any actions), and the uncertainty associated with the value of various
assets. At least three factors influence the risks firms face and their
willingness to bear those risks. One key factor is the overall state of the
economy. When the economy goes into a recession, and firms talk about
their pessimism or uncertainty, these perceptions have real consequences. A
second factor is the firm’s cash (or liquid asset) position. Changes in a firm’s
cash position affect how much it must borrow to maintain its production
activities. A firm’s cash position is affected by profits, and since profits are
a residual, small changes in prices may have large effects on profits, and
thus on firm liquidity, particularly for highly leveraged firms. Of course, the
lower profits also adversely affect the firm’s net worth. A third important
factor is changes in the price level. Since almost all debt is denominated in
nominal terms, such changes have large effects on firm real liquidity and
real wealth.9
The theory of the risk averse firm can thus provide an explanation of why
each firm’s supply curve, and hence the aggregate supply curve—the amount
that they are willing to produce at each level of prices (given wages)—should
shift markedly as the economy goes into a recession. The riskiness of
production has increased, and firm’s willingness and ability to bear that risk
has decreased.
To maintain the same level of economic activity, with the reduced cash
flow from lower profits, firms must borrow more. But increased debt creates a
higher probability that future returns will not be sufficient to meet these fixed
obligations. As the firm expands its production, it must borrow more,
increasing its fixed obligations; there is an increased chance of not being able
to meet those increased fixed obligations. The expected extra costs associated
with bankruptcy are what is meant by the ‘marginal bankruptcy cost’.
Normally, the necessity to borrow more resulting from lower cash flow (lower
profits) not only increases the probability of bankruptcy (at any fixed level of
economic activity) but also the marginal bankruptcy costs. Once bankruptcy
costs are taken into account, we need to modify the standard theory of the
firm, where, as a firm expands, it compares price (marginal revenue) with
marginal cost.
Thus, the aggregate supply curve shifts to the left. The shift in the firm
(and aggregate) supply curve means that the amount firms are willing to
produce, at each level of prices and wages, is reduced; conversely, it also
means that at each level of output the firm’s mark-up of price over marginal
costs (largely determined by the wage) is increased. Moreover, the same
reasoning provides an explanation of why the aggregate demand curve should
shift to the left in this situation: the firm’s demand for investment may shift
down markedly.
The theory also explains why large redistributions, like those stemming
from large price changes (like the oil price shocks of the 1970s) should have
558 Bruce Greenwald and Joseph Stiglitz
a negative effect on the economy. While increases in wealth lead to
increasing production and investment in the sectors which benefit from the
price change, there are diminishing returns; the increases of, say, production
from those who benefit are more than offset by the reductions from those
who lose.
The theory of the risk averse firm explains a number of other aspects of the
cyclical behavior of the economy. For instance, imperfections in equity
markets and the extent of leveraging on equity differ across sectors.
Construction, for instance, is an industry dominated by small firms, most of
whom do not have access to the equity market; and construction firms
typically borrow heavily to finance their construction activities. Such sectors
one would expect to be particularly volatile.
To illustrate how the risk averse theory of the firm can explain why shocks
to the economy, whether real or monetary, can have real, large, and
persistent effects, let’s trace through an example. Say that a decrease in export
prices (to lower than expected levels) reduces exporters’ net worth, leading
them to reduce their supply, and their demand for inputs from other
producers. This unexpected change in the demand curve for others’ products
leads to lower prices than expected in other sectors, with adverse effects on
their asset and liquidity position and on what they want to produce, and their
demand for inputs (including investment).
Inventory adjustments exacerbate the process: with greater perceived risk
and lower wealth, and hence reduced willingness to bear risk, firms cut back
on their desired level of inventories; this translates into a further reduction in
production. Note that the theory of risk averse firm thus offers an answer to
one of the long-standing puzzles of macroeconomics: why inventories do not
seem to perform the production smoothing role they should, with concave
production functions: if anything, inventories seem to exacerbate economic
fluctuations (Blinder and Maccini 1991).
We thus have a mechanism for the transmission, amplification, and
persistence of the effects of shocks, even with complete flexibility of wages
and prices. Such a model can explain volatility, and also provide answers to
the two other questions posed in the beginning of this chapter. If one adopts a
standard model of money demand, say with constant velocity, then
unanticipated changes in the money supply lead to unanticipated changes in
the price level, which will set off the process described above. Remember,
changes in the price level affect the value of firm debt, since that debt is
usually denominated in nominal terms.
Moreover, hiring workers is an investment. As the economy goes into a
recession, the optimal portfolio of assets for a firm includes less ‘human
capital’. Beyond that, the shadow cost of capital—taking into account, for
instance, the increased risks of bankruptcy that follow from the increased
borrowing required to finance the hiring and training costs of new
employees—is high in a recession, and thus, even if firms eventually wanted
to increase their stock of employees, the depths of a recession is not the time
New and old Keynesians 559
to make that investment. Thus, new hires are reduced.10 This gives rise to
unemployment, which results when the rate of separations exceeds the rate of
new hires. As this theory would predict, the rate of new hires shows greater
cyclical volatility than the rate of separations.
CREDIT MARKETS AND RISK AVERSE BANKS
The theory of risk averse firms takes us a considerable distance, but effects
that operate through the banking system and credit markets provide yet
another process by which shocks to the economy are amplified and their
effects propagated, and another set of reasons why monetary policy will
work, even in a world with flexible prices and wages.11
Recent economic work has emphasized that credit is not allocated in an
auction process, with whoever is willing to pay the highest interest rate
receiving the loan. Instead, lenders must face the risk that a loan will not be
repaid, and institutions, like banks, have arisen for screening loan applicants
and monitoring loans. Banks are highly leveraged; with fixed obligations (the
deposits they hold) and risky assets, banks must worry about the risk of
bankruptcy. It is now well-known that increasing interest rates may have
adverse effects both on the mix of loan applicants and on the incentives of
borrowers to undertake risky activities, and that these adverse incentive and
selection effects can be so strong that lenders’ expected returns may actually
decrease as the interest rate charged increases. This can lead to credit
rationing, with the interest rate charged being that which maximizes the
expected return to lenders, and at that interest rate, there is an excess demand
for credit.
Greenwald and Stiglitz (1990a) have extended that analysis to embrace
risk averse lenders. Like the equity-constrained firms described earlier,
banks, who must worry about the risk of bankruptcy, act in a risk averse
manner. There will still be credit rationing, with interest rates chosen to
maximize the ‘expected utility’ of the lender, or the expected returns minus
the costs of bankruptcy. But with risk averse banks, the same kinds of
factors which affect firm behavior—changes in risk perceptions and
changes in net worth, affecting the willingness to bear risk—affect bank
behavior, too.
This risk averse behavior of banks will magnify an initial negative
economic shock, and make recessions deeper and longer. The banks’
portfolio of activities can usefully be divided into recruiting and processing
new customers; making (and monitoring) loans to existing customers; and
buying a safe asset, like Treasury bills. When economic conditions worsen,
banks’ perceptions of the relative risk of loans increases; and since bad
economic conditions are often accompanied by high default rates, banks’
net worth decreases, along with their willingness to bear risks. On both
accounts, banks respond to bad conditions by shifting their portfolio
towards the safer activity: investing in Treasury bills. Equilibrium in the
560 Bruce Greenwald and Joseph Stiglitz
loan markets would be attained only at a higher real interest rate, which
would also discourage investment activity. And banks will often be
unwilling to raise interest rates, because of a fear that higher rates will have
the adverse selection effect of chasing away credit-worthy borrowers and
adverse incentive effects, inducing them to undertake greater risks (Stiglitz
and Weiss 1981).
Monetary policy still works (at times) in this situation, but not in the
accustomed way. The conventional monetary policy story has the Federal
Reserve driving down interest rates, which stimulates investment. In this
situation, though, while monetary policy may succeed in lowering the rate of
interest on Treasury bills, the change in interest rates charged by banks may
well be minimal. It may also result in little change, if any, in the supply of
loans: while there is a substitution effect associated with loans being
relatively more attractive, there is an income effect which goes the other way
(if banks have decreasing absolute risk aversion). And in the credit rationing
regime, it is the supply of loans which is critical; firms are limited in their
investment activities, and possibly even in their production activities (if they
rely on bank credit for working capital) by the lack of credit.12
However, monetary policy also works through another set of mechanisms.
Reserve requirements (when reserves are kept in accounts that bear little or no
interest) act as a tax on deposits. Higher reserve requirements raise that tax,
and reduce the wealth of banks; lower reserve requirements have the reverse
effect. Lowering the discount rate has the effect of reducing one cost facing
the bank—the cost of obtaining funds from the central bank. This change
increases the real wealth of banks, making them more willing to bear risks
and make loans. Since the ratio of loans to net worth for banks is typically
very large, relatively small changes in bank net worth can give rise to large
changes in credit availability.13
Although monetary policy can have potent effects through these channels,
it will also be relatively impotent at times. If the economy is very weak, so
that expected returns on bank loans are very low, relative to the risks
associated with them, then raising the wealth of banks may still not make
lending money look profitable.14
LABOR MARKETS
One peculiar aspect of old Keynesian analysis was that while its main
concern was unemployment, it offered little discussion of the labor market.
However, a consensus is growing that an understanding of the labor
market must be at the center of any macroeconomic theory (Lindbeck
1992).
The basic empirical puzzle in the labor market is that employment levels
change markedly, with little change in real wages. One explanation is that
the supply curve for labor is horizontal, but that would run counter to all the
microeconomic evidence, as well as introspection. Another explanation is
New and old Keynesians 561
that, by some miracle of coincidence, shifts in the demand and supply curves
have been perfectly offsetting. A recession, for example, is marked by a
leftward shift in the labor supply schedule, just as the demand schedule
moved left. But why should labor supply fall so fortuitously? Changes in real
interest rates and expectations concerning future wages could, of course,
through intertemporal substitution, induce shifts in the labor supply schedule;
but micro-evidence suggests that these intertemporal substitution effects are
far too small to obtain the desired effects. A further problem is presented by
the contradictory movements in real interest rates: in the Great Depression
they rose markedly; during the recessions of the 1950s, 1960s and early
1970s, they changed hardly at all.
New Keynesians offer an alternate interpretation. They have explored
reasons why real wages are not likely to move. As a result, shifts in demand
for labor can create a situation where people are willing to work at the going
wage, but cannot find jobs; in other words, there is involuntary
unemployment. Some of the possible reasons for sticky real wages include
efficiency wages, insider-outsider theory, imperfect competition, and implicit
contracts. Let us say a few words about each; the reader interested in a
thorough evaluation of these theories might begin with Stiglitz (1992a) or
Newbery and Stiglitz (1987).
Efficiency wage theories argue that productivity often increases with real
wages; as a result, it does not pay firms to cut wages. High wages may raise
productivity either because they attract higher quality labor; or because they
result in increased effort; or because they reduce labor turnover and save on
hiring and training costs.15 Efficiency wage theories can be used to explain
why firms do not lower wages even in the presence of an excess supply of
workers, and also why they avoid two-tier wage systems, under which new
workers are hired at lower wages than existing workers.
Insider-outsider theories and bargaining theories begin with the presence of
turnover costs, and then argue that trained ‘inside’ workers are not a perfect
substitute for untrained ‘outside’ workers. This situation gives rise to a
bargaining problem. Since ‘inside workers’ control the training process, they
would react negatively to hiring workers at lower wages who could
potentially replace them. Moreover, the fact that new workers cannot commit
themselves not to demand higher wages once trained provides a further
reason that firms do not hire ‘cheap’ new workers.
When imperfect competition exists in labor and product markets, firms set
wages, prices, and employment. Given the risk averse nature of the firm, as
described earlier, and efficiency wage and insider-outsider effects just
mentioned, a firm that is considering lower wages must face considerable
uncertainty about the possible effects on the effort, quality, and turnover of its
labor force.
To this point, the discussion has focused on the ‘supply side’ of the labor
market. But the demand side offers a puzzle as well. The demand for labor at
any real product wage can be derived in a straightforward way from the
562 Bruce Greenwald and Joseph Stiglitz
production function. The fact that employment varies considerably with small
variations in real product wages presents a puzzle.
With given technology and capital stock, if firms operate along their
supply function (with concave production functions), then a reduction in
output should be associated with an increase in real product wages, contrary
to what is observed. There are several possible explanations. One is that,
somehow, there has been a large negative change in technology. The
implausibility of this hypothesis, and the empirical evidence against it, are
matters taken up elsewhere. Second, there could be a change in the degree of
competition, and hence in the mark-up over marginal costs. Third, firms
could simply be off their supply curve. (For a critique of these alternative
explanations, see Stiglitz 1992b.)
We prefer a fourth theory, provided by the theory of the risk averse firm.
Earlier, we explained why the firm and aggregate supply curve of output
shifts as the economy goes into a recession. One can easily translate this into
a shift in the firm and aggregate demand curves for labor.
The new Keynesian research program in labor economics followed
traditional macroeconomics in seeking to explain the observed patterns of
real wages and employment. But it has also tested those explanations against
a number of other key aspects of the labor market, like why reductions in the
demand for labor take the form of layoffs rather than reduced hours for
everyone and why unemployment seems to be so concentrated in certain
groups in the population. Focusing on these characteristics of unemployment
is important, because if the reduction in the demand for labor took the form
of an equi-proportionate reduction in the hours worked by each individual,
the social and economic consequences of unemployment would be much less
than they in fact are. The labor market theories described above are able to
explain these phenomena.16
PERSPECTIVES ON ALTERNATIVE THEORIES
Our main objective in this chapter is to describe this emerging strand of new
Keynesian literature in broad terms. To this point, we have described how
theories based on informational imperfections can explain the main puzzles
mentioned at the start of the chapter: the presence and persistence of
unemployment, the variability of output, and why money matters. In fact, the
theories described here go farther, and offer an explanation of why certain
sectors of the economy exhibit greater volatility than others; why the
variability in hours worked takes the form of layoffs; and the logic behind the
cyclical patterns of inventories, hours worked and employment.
In this section, we describe the kinds of arguments that persuade us that
alternative theories are at best incomplete, at worse wrong. None of the
theories discussed in this chapter, including our own, have been fully
embodied in a large macro-econometric model. We believe that constructing
such models, together with conducting the kind of simulation exercises that
New and old Keynesians 563
have provided much of the support for real business cycles, should be on the
agenda for future research. But before subjecting a model to that sort of
extensive testing, we believe it must be shown that it can at least display the
critical basic observed facts about the economy. Thus, our discussion will
seek to identify key observations which, in our judgment, cast serious doubt
on the major competing theories.17
NEW KEYNESIAN PRICE RIGIDITIES
As mentioned earlier, one strand of new Keynesian economics has emphasized
nominal price rigidity, and used explanations that go under the name of
‘menu costs’ to explain that rigidity.
A number of facts imply that price rigidities are, at a minimum, not the
only source of economic problems like volatility and unemployment. For
example, Keynesian-like unemployment problems seem to arise even in
economies which are experiencing inflationary pressures, and thus where the
nominal wages do not need to fall, but only to rise more slowly. Moreover,
nominal wages and prices did fall in the Great Depression, as well as in other
economic downturns. We agree with Keynes that had prices fallen even faster,
the economy would have degenerated farther, rather than improving more
quickly.
Indeed, in most new Keynesian models the mechanism by which wage and
price flexibility would eventually restore the economy to full employment is
the old real balance effect. The enormous attention that the real balance effect
has received over the years hardly speaks well for the profession.
Quantitatively, it is surely an nth order effect; one calculation put it that,
even at the fastest rate at which prices fell in the Great Depression, it would
take more than two centuries to restore the economy to full employment. And
in the short run even its sign is ambiguous, as intertemporal substitution
effects may (depending on expectations) more than offset the wealth effects
(Neary and Stiglitz 1982; Grandmont 1983).
But while price rigidities may not be at the center of phenomena like
fluctuations and unemployment, and one does not have to assume price
rigidities to establish that monetary policy has real effects, the relative rigidity
of wages and prices remains a phenomenon which needs to be explained.
The menu cost literature has attempted to argue that the costs of
adjustment, like the costs of printing new menus, results in firms only
adjusting prices periodically, which is another way of saying that price
stickiness exists. From a tactical point, the advocates of menu costs beat their
critics to the punch by choosing a name—‘menu costs’—which would seem to
belittle the importance of the subject. Indeed, these costs are small, and have
become smaller as computer programs allow the printing of menus on a daily
basis at a marginal cost of pennies.
Two arguments were necessary to give these seemingly small effects any
plausible relevance (Akerlof and Yellen 1985). First, if firms are already
564 Bruce Greenwald and Joseph Stiglitz
choosing their prices optimally, then the cost of not adjusting was of second
order. Thus, while the costs of adjusting may be small, so were the benefits of
adjusting. Second, in spite of the small (second order) losses to the firm, the
losses to society could be first order.18 While both of these propositions are
correct, they are not sufficient to justify paying much attention to the menu
cost literature. Both propositions apply to any decision of the firm: they offer
no reason to single out pricing decisions.
By contrast, we have emphasized that firms must view all their decisions
together; that the costs of adjusting prices must be put in juxtaposition with
the costs of adjusting (or not adjusting) quantities. Since there is a strong
presumption that costs of adjusting outputs and inputs will be much greater
than those associated with simply adjusting prices, this would seem to argue
for quantity rigidities, and against price stickiness. But when focusing on risk,
as we have done, the conclusion changes. When a firm considers the various
ways it might react, it will perceive greater uncertainty about the
consequences of price and wage adjustments—because those consequences
depend on the uncertain responses of rival firms, customers, and workers—
than about the consequences of output adjustments. In fact, for those goods
which can be put into inventory, the only risk associated with producing too
little is the risk associated with higher production costs next period, when any
inventory deficiency must be made up. (Of course, boom times may create a
risk of running out of stock, but that risk is not important in recessionary
periods.) This portfolio theory of firm adjustment does provide an explanation
of price and wage rigidity, at least in the short run; though in the long run,
the theory suggests that prices and wages eventually do adjust.19
To be sure, if agents in the economy perfectly anticipated changes in the
money supply and if it was common knowledge that all agents in the
economy responded to changes in the money supply by changing all prices
proportionately, then money might be neutral. But since the money supply is
not perfectly observed by all agents, not all agents change prices
proportionately, and so there is no reason that they should all believe that
price changes will perfectly offset changes in the money supply. Given the
uncertainty about whether other agents will increase prices proportionately to
observed changes, it will not generally be optimal for any firm to increase its
price proportionately; thus, the beliefs about non-proportional responses to
price changes are consistent.
Thus, there is a presumption that as long as risk markets are incomplete
and firms and individuals are risk averse, and debt is imperfectly indexed,
then an expansion of the money (credit) supply will have real effects. Also,
there are distributional consequences of the manner in which the money
(credit) supply is increased. A credit expansion affects some individuals,
firms, and industries more than others. In short, money (credit) matters, but
not just because of nominal rigidities.
In fact, our theory can be seen as a particular kind of menu cost theory—
a theory which emphasizes the riskiness of adjusting prices, rather than the
New and old Keynesians 565
actual adjustment costs. But while our theory does provide a theory of price
stickiness, it argues that price stickiness is only one element, and not the most
important one, in understanding macroeconomic phenomena. And nothing
that we have said would be substantially altered if, in addition to the risk
costs which we have emphasized, fixed costs of price adjustment were
significant.20
Another major distinction between the two strands of new Keynesian
literature is whether nominal or real price rigidities are emphasized. One
strand uses nominal rigidities as an important step in explaining why
money matters. But in the alternative theory, based on the risk averse theory
of the firm with incomplete contracting and indexing, money matters more
as prices become more flexible. By contrast, to explain unemployment, it
focuses on real rigidities in the labor market (such as associated with the
efficiency wage theory). It argues that whatever happens to the product
market, unless one has a theory of real wage rigidity, one cannot explain
unemployment. For even if there were large shifts in the demand curve for
labor, if the real wage were flexible, demand and supply for labor would
equilibrate.21
There is, however, an important difference between the two approaches for
policy purposes. A menu cost theorist would focus efforts at structural
macroeconomic reform on reducing the costs and speeding the
implementation of price changes. Anti-inflation measures like those
considered in the 1970s, which penalized price changers, would have
potentially destructive consequences for overall economic welfare. A menu
cost theorist would to the contrary advocate measures which would provide
incentives for rapid nominal price adjustments. In contrast, in our model
rapid price adjustment is a two-edged sword. On the one hand, it reduces the
reliance of firms on quantity adjustment and hence might stabilize aggregate
levels of employment and output. On the other hand, greater overall price
changes would mean greater wealth transfers to and from firms, exacerbating
the financing imbalances which act to amplify the original macroeconomic
disturbances. On balance, therefore, we would regard price and wage rigidity
more as a symptom of underlying financial and labor market failures and not
as a fundamental cause of business cycles. We, therefore, would focus
structural reform on those fundamental areas rather than directly on price and
wage setting by firms.
OTHER KEYNESIAN THEORIES
Of course, there are other strands of Keynesian and new Keynesian thought
besides those focused on price rigidities. One strand which enjoyed
considerable popularity in the 1970s and 1980s was that of Tobin, which, like
our theory, emphasized the importance of risk. It used a portfolio theory to
explain the demand and supply of assets; and related firm investment to the
price of (existing) capital goods, as reflected in the price of equity, which
566 Bruce Greenwald and Joseph Stiglitz
emerged in the market equilibrium. Monetary policy affected this price, and
hence the level of investment.
The theory has had limited empirical success. One possible reason is
that firms raise little of their funds for investment through equity. What
success it has had may be due to a spurious correlation: when a firm’s
future prospects are good, firm managers invest more, and the firm’s stock
is high. There is not (necessarily) the causal connection suggested by that
theory.
That theory, as well as most other Keynesian theories, explain the effect of
monetary policy by looking at the demand for money by households. Our
theory focuses more on the effects on the banking system, and on the
implications through the credit mechanism, both as a result of credit rationing
and the behavior of the risk averse firm.22
REAL BUSINESS CYCLE THEORIES
Real business cycle theory addresses two of the three puzzles with which we
began this chapter by denying their existence: proponents of this school deny
either that (involuntary) unemployment exists or that money matters. (The
fact that monetary policy is ineffective is of little moment, since in any case
the economy is, in this view, efficient, with resources being fully used.) This
school of thought focuses on the second problem, that of economic volatility,
and proposes exogenous technology shocks as the source of that volatility.
The most telling criticisms of this view is the difficulty it has explaining the
large negative shocks that mark recession: was there a loss in technological
competence?23
Of course, if one includes economic organization in ‘technology’, and in
the information embodied in the various firms within the economy in
‘capital’, then the financial disorganization and risk associated with
recessions discussed in this chapter represents both a negative technology and
capital shock. With this expanded vocabulary, the basic model of risk averse
firms and banks, together with flexible wages and prices, and market clearing
in the labor market, can be viewed as a version of real business cycle
theory—but one with fundamentally different predictions and policy
presumptions than the standard version of the theory.
NEW CLASSICAL THEORIES
The branch of new Keynesian theory emphasized here shares a
methodological premise with at least some versions of new classical
theories: the importance of imperfect information in explaining observed
deviations from the predictions of neoclassical theory. But new classical
theories have tended to focus on the consequences of imperfect information
for the inferences firms make—say, about the desirability of changing price
or quantity. We think the difficulties firms have in inferring whether a shift
New and old Keynesians 567
in the demand curves which they face is due to a real or nominal shock may
play a role in explaining ‘why money matters’, but surely it is not the only
reason, nor even perhaps the most important one. While accepting the
importance of looking at these issues, we also emphasize the implications of
imperfect information for how markets function—the causes and
consequences, for instance, of credit rationing, limited equity markets, and
efficiency wages.
Another ingredient in new classical models attempts to explain why
unanticipated increases in prices (presumably following from an
unanticipated increase in the money supply) might elicit a larger than normal
output. Our theory provides an alternative explanation: larger-than-
anticipated increases in prices increase firms’ net worth, and this increases the
amount they are willing to supply. Our theory is not based on misperceptions:
at the time the loan contract was made, it was anticipated that, with some
probability, prices would be high.
New classical economists have also emphasized the importance of
expectations (as does King 1993), and particularly rational expectations.
Thinking about expectations is hardly new. Keynes invoked a variety of
assumptions concerning expectations, and in this, he was only reflecting the
common practice of the time.24 Today, most Keynesians believe that whether
expectations are ‘rational’ is an empirical question—one which, in important
instances, will surely be answered in the negative. For example, the stock
market crashes of 1929 or 1987 seem very difficult to reconcile with ‘rational’
expectations.
At the same time, many new Keynesians are not adverse to using the
rational expectation assumption when it is convenient to do so (for example,
Greenwald and Stiglitz 1986). One especially interesting result is that the
basic results of the models which lead to the conclusion that government
policy is ineffective do not depend on the assumptions of rational
expectations, but rather on even less realistic assumptions concerning
instantaneous market clearing. For instance, Nearly and Stiglitz (1983)
supply a model with price and wage rigidities where rational expectations
actually increased the multipliers from government action. The multipliers
were larger for an obvious reason: consumers with rational expectations
recognized that the ‘leakage’ of increased income into savings would be
translated into higher consumption in future periods; and the expectation of
this higher future income ‘spilled over’ into higher current consumption.
To be sure, rational expectations of policy changes may sometimes lead
individuals to act in a manner which undoes those policy changes, but this is
surely not the case when the government imposes taxes or subsidies which
change intertemporal prices, nor when the government engages in
redistributions which have aggregate effects. Obviously, in models with a
representative agent, redistributions make no sense, and cannot have any
effects. But this just illustrates how such models may be of little use in
addressing fundamental issues of macroeconomics.
568 Bruce Greenwald and Joseph Stiglitz
SUMMARY
The economy is a complex organization, requiring coordination of
decisions of the millions of households and firms. Unemployment and
other macroeconomic problems can be viewed as a failure of society to
solve the necessary coordination problem efficiently. The focus of our
research program has been to understand why markets and other social
institutions sometimes do not work as well as we would like. Given the
complexity of the economy, no one should expect to find a single
explanation of any of the macroeconomic phenomena under study. There
is no Holy Grail. But new Keynesian economists, whether of the first or
the second type as described in this chapter agree on two broad
propositions.
First, they agree that the Walrasian auctioneer does not really exist,
and that ‘as if’ stories about the auctioneer are a fiction that has too long
misled the profession. Instead, firms set prices and wages in an
uncoordinated fashion, facing considerable uncertainties about the
consequences of their actions. As a result, it will often be true that wages,
prices, and interest rates are not at market clearing levels (and will not
adjust rapidly to those levels), so that large parts of the economy will not
be in equilibrium.
Second, they agree that problems of coordinating prices and wages simply
cannot be studied in the context of a macroeconomy consisting only of an
aggregated representative agent, like Robinson Crusoe. It is not even clear
that an island with Robinson Crusoe and Friday provides a fertile basis for
studying macroeconomic problems, though at least this opens the possibility
of problems such as those associated with asymmetric information. Indeed, at
the core of the models discussed here is the notion that redistributions of
wealth across firms and between households and firms matter, and they
matter because there is a corporate veil created by imperfect information.
Aggregate approaches using representative agent models are not of much use
in studying these macroeconomic phenomena.
The strand of new Keynesian literature discussed and advocated here
attempts to shift the focus of the research program in two ways. It argues for
shifting the analysis of these issues from the product market to the capital and
labor markets. In addition, it argues for shifting away from a single-minded
pursuit of the consequences and causes of price rigidities; in fact, the analysis
here suggests that greater price flexibility might exacerbate the problem of
economic fluctuations.25,26 Instead, we believe that the focus should be on how
imperfections in information limit, and sometimes even eliminate, the
markets which distribute risk in modern economies; how these market
imperfections serve to amplify the shocks facing the economy and make their
effects persist; and how, when translated to the labor market and combined
with information and other problems there, they can give rise to high levels of
unemployment.
New and old Keynesians 569
NOTES
1 For a devastating attack on the underlying methodological premises of the
representative agent approach, see Kirman (1992).
2 The term ‘constrained’ in the concept of ‘constrained Pareto efficiency’ is simply
inserted to remind readers that the constraints—absence of a complete set of markets,
the imperfections of information, and so on—were indeed taken into account. Even
when the government faces these constraints, when the economy is not constrained
Pareto efficient, there exist interventions in the market which can make all individuals
better off. There are, to be sure, innumerable papers in the literature showing that
with incomplete markets and imperfect information, the economy may be constrained
Pareto efficient. The point of the Greenwald-Stiglitz (1986) paper was to show that
these papers all entail special assumptions; and that in general, the market economy
is not constrained Pareto efficient.
3 Some new Keynesians are wont to claim that this insistence on micro-foundations is
what distinguishes them from Keynes and the older Keynesians. Though much
macroeconomic analysis in the Keynesian tradition in the 1950s and 1960s did stray
from a solid grounding in micro-foundations, Keynes himself clearly argued each of
his macroeconomic relations on the basis of microeconomic analysis. In fact, we
would argue that Keynes did the best he could with the micro-foundations which were
available at the time. Macroeconomists of the 1950s and 1960s faced a dilemma: the
microeconomics that was fashionable at that time—assuming perfect information,
complete markets, and so on—was obviously inconsistent with the spirit of the
Keynesian model. It made sense for them to ignore that kind of microeconomics.
4 There are still other strands emphasizing, for example, imperfect competition or
coordination failures.
5 For example, Hansen (1951), Solow and Stiglitz (1968), Barro and Grossman (1971),
and the large subsequent literature surveyed in Benassy (1982).
6 In taking this approach, this second strand of new Keynesian thought addresses
one of the major criticisms of real business cycle theory—that real shocks to the
economy are simply not large enough to account for the magnitude of the observed
fluctuations. Standard neoclassical models have strong forces working to stabilize
the economy: price adjustments act like shock absorbers; savings and inventories
act as buffers; lags mean that even a major new innovation will take years to be
absorbed into the economy; and many shocks have offsetting effects in different
sectors, implying limited aggregative impacts.
7 For a more detailed discussion of the arguments presented in this section, and the
empirical evidence in its support, as well as a more complete list of references, see
Greenwald and Stiglitz (1987, 1988b, 1988c, 1989, 1990a, 1990b, 1991a, 1993),
Stiglitz (1992a), and Greenwald et al. (1984).
8 One explanation for why firms do not issue equities upon which we do not put
much credence is the costly state verification model, which notes that using equity
requires verifying the state (the firm’s profits), so the costs of implementing equity
contracts thus exceed that of debt contracts. While this argument has some relevance
for small businesses, firms that have already issued equity have little or no marginal
cost of verifying their state when they seek to issue additional equity.
9 A major lacuna in this theory is the failure to explain why debt contracts are
denominated in nominal terms. However, there are models, such as Cooper (1990),
which show that there may be Nash equilibrium with imperfect indexing; that is,
given that all other contracts are not indexed, firms would not want to just index
debt contracts.
570 Bruce Greenwald and Joseph Stiglitz
10 The story, as presented thus far, is not quite complete: Why don’t workers cut the
wages at which they are willing to work, and thus make it worthwhile for the firm
to hire them even though costs of capital are high? There are several answers: the
required reductions in wages are so large that workers prefer to wait (what they
expect to be the short time) until the costs of hiring are lowered; workers are not
willing to put up the cost of being hired for a whole variety of reasons, from lack
of capital, worker risk aversion, and firm moral hazard; workers cannot reduce
the firm’s risk of hiring by making wages contingent upon the performance of the
firm, for that would entail, in effect, workers taking an equity share in the firm,
and all the arguments for why equity markets fail apply with equal force here; and
workers cannot even commit themselves to charging only a low wage, once they
are trained, as insider-outsider theories have emphasized.
11 For a more extensive development of these ideas, see Stiglitz (1988, 1992a) and
Stiglitz and Weiss (1992a, 1992b).
12 Of course, this story raises the question of why firms facing credit constraints from
their banks do not turn to other sources of funds. We have explained why equity is
not a viable alternative. Other sources of funds are even less informed about
creditworthiness; they are likely to make credit available only under much less favorable
terms, or not at all. Adverse selection works to exacerbate other sources of credit, too;
the firms that avail themselves of these alternative supplies are those in dire straits.
13 Open market operations will have a similar net wealth effect on banks. However,
this will occur only to the extent that rates of interest paid on demand deposit are
held below their competitive levels, by either direct legal fiat or limitations on interbank
competition. With zero interest on demand deposits, increases in deposits (if believed
to be permanent) represent equal increases in effective bank equity; thus money
supply expansions represent a particularly powerful wealth transfer from households
to banking firms. (Letting W=bank wealth derived from deposits, r=interest rate,
D=deposits, W=r(D/r)=D). If the monetary policy is believed to be temporary, then
there may be no significant wealth effect (AW=∆M·r+∆r·M; when r is near zero, this
is near zero). However, if rates paid on demand deposits are competitive either
because, as recently, they are deregulated or because, as in the Depression, a zero
nominal rate is close to the competitive rate of interest, increases in demand deposits
through open market operations will have no significant effect on bank wealth.
14 At one level of analysis, the insights of this model can be viewed as a mild modification
of standard IS-LM theory. The LM curve is now derived not as the equilibrium in the
money market—the locus of interest rates and income levels at which the demand
for money equals the supply of money; but rather as the equilibrium in the capital
market—the locus of interest rates and income levels at which the capital market is
in equilibrium; for the capital market to be in equilibrium the demand for reserves
must be equal to the supply (otherwise banks would change their behavior) and the
demand for Treasury bills held by the public must be equal to the supply.
Operationally, the standard IS-LM curves differ in two fundamental ways
from the ones implicit in our analysis. First, we have identified a set of variables—
balance sheet variables of firms and banks, and the dispersion in those variables—
which affect both the IS and LM curves, and can cause them to shift markedly.
Second, monetary policy may shift the IS curve: firm investment depends on the
interest rates charged by banks and the credit they make available. The interest
rate charged by banks is not just the government interest rate. There may be
marked changes in the spread (for instance, they increased in 1991, so that bank
loan rates fell much less than did government interest rates). Monetary policy may
New and old Keynesians 571
affect not only the Treasury bill rate, but also the spread, so that monetary policy,
in effect, shifts the IS curve as well.
15 Workers’ efforts may be reduced if they receive less than what they perceive to be a
fair wage; while they may respond to higher wages with higher effort as part of a ‘gift
exchange’ (Akerlof 1982).
16 The precise mechanism differs among the different theories. For instance, in the
Shapiro and Stiglitz (1984) efficiency wage theory, it is the risk of being fired, and
with it loss of total rents which provides workers incentives not to shirk. But part-
time workers, with the same surplus per hour, have a lower total level of surplus.
Certain changes in the economic environment which necessitate an increase in the
wages to induce workers not to shirk will necessitate a larger increase in the wages of
part-time workers, and thus, these workers become less attractive—their costs,
adjusted for quality, increase.
17 One group of theories not discussed here, which should be mentioned briefly, are
those focusing on imperfect competition. For many of the central issues with which
macroeconomics is concerned, we do not believe that imperfect competition is central.
For example, imperfect competition can hardly explain the cyclical movements in
output and employment. While prices might be different from what they would be in
perfect competition, imperfect competition in the product market cannot explain
why the labor market does not clear. While the classical dichotomy has traditionally
been couched in terms of models with perfect competition, one can prove analogous
results from general equilibrium models with imperfect competition. However, we
do view imperfect competition as important to the extent that it allows firms to set
prices and wages. As explained earlier, the price and wage setting behavior of risk
averse firms has important macroeconomic consequences.
18 This result can also be seen as a direct corollary of the Greenwald and Stiglitz (1986)
analysis of the welfare economics of economies with imperfect information and incomplete
markets. We show there that under those conditions the economy is not (constrained)
Pareto efficient, and that whenever this is true, pecuniary externalities matter.
19 In some important cases, however, the economy exhibits nominal rigidities even in
the long run. In effect, these are cases of multiple equilibria. If each firm believes
other firms are going to keep their nominal wages rigid, it pays each firm to keep its
own nominal wages unchanged. There are thus equilibrium exhibiting nominal
wage rigidities (Stiglitz 1985). Similarly, if each firm believes other firms are going to
keep their nominal prices unchanged, it pays each firm to keep his nominal prices
unchanged. There are thus Nash equilibrium nominal price rigidities (Stiglitz 1987).
20 One empirical objection to standard menu cost theory (which is addressed by our
theory) is that while the theory would seem to explain rigidities in the adjustments in
the level of prices; it has a hard time explaining inflation inertia—that is, rigidities in
adjustments of the rate of change of prices.
21 Of course, nominal rigidities in wages and prices give rise to real wage rigidities. See
Solow and Stiglitz (1968) for a model incorporating explicitly stickiness in both. Of
course, if the costs of adjusting wages and prices differ, one would not expect the
same degree of stickiness in both markets, and thus, one would expect systematic
changes in real wages in response to particular economic disturbances.
22 For a more extended critique of the standard theory of the household’s demand for
money, see Green wald and Stiglitz (1991b).
23 For an introduction to the claims and difficulties of real business cycle theory, see the
exchange between Plosser and Mankiw in the Summer 1989 issue of Journal of
572 Bruce Greenwald and Joseph Stiglitz
Economic Perspectives. Other criticisms, besides those mentioned in the text, include
the lack of correlation across countries of the implied shocks to different industries
(which one would expect if the shocks were really technology shocks), compared to
the correlation of industries within a country. Also, this school has failed to identify
large positive shocks of the required magnitude. (Remember that once one takes
into account the shock absorbers, buffers, and lags, and that much of technology is
‘embodied’, then the implied shocks to technology must indeed be large.)
Furthermore, negative technology shocks move the factor price frontier inward;
that would imply that if real product wages remain unchanged, real interest rates
would have to fall markedly. But in fact, while real interest rates vary little in many
recessions, in other recessions, like the Great Depression and the 1982 recession,
real interest rates rose.
24 At times, Keynes seems inconsistent in his discussion of expectations. For example,
in discussions of the liquidity trap, it was argued that the value of long-term bonds—
consols—was inversely proportional to the short-term interest rate, a result which
can be justified on the basis of static expectations concerning interest rates (that is,
the expectation that future interest rates will, on average, be equal to current interest
rates) and risk neutrality. It was then argued that, when interest rates were very low,
investors were worried that the interest rate would rise, giving rise to a fall in the
price of consols. But if investors expect interest rates to fall, then the price of consols
will not be inversely proportional to the short-term interest rate, and changes in the
short-term interest rate will have negligible effects on the price of consols.
25 Thus, the work described in this chapter can be thought of as providing the theoretical
underpinnings of one of the standard interpretations of the Great Depression and
other major economic downturns, the debt-deflation theories. See for example,
Calomiris (1993) and the papers cited there. (Of course, the effects we describe do
not require actual deflation, only a slowdown in the rate of inflation relative to that
anticipated.)
26 Keynes seemed to be of that view when he concluded, ‘In the light of these
considerations, I am now of the opinion that the maintenance of a stable general
level of money wages is, on balance of considerations, the most advisable policy for
a closed system.’ Clearly, Keynes did not consider the central problem one of lack of
wage (and price) flexibility.
REFERENCES
Akerlof, G.A., ‘Labor Contracts as Partial Gift Exchange’, Quarterly Journal of
Economics, 97:4, November 1982, 543–69.
Akerlof, G.A. and J.Yellen, ‘A Near-Rational Model of the Business Cycle with Wage and
Price Inertia’ , Quarterly Journal of Economics 1985, 100, 823–38.
Barro, R.J. and H.I.Grossman, ‘A General Disequilibrium Model of Income and
Employment’, American Economic Review 1971, 61, 82–93.
Benassy, Jean-Pascal, The Economics of Market Disequilibrium, New York: Academic
Press, 1982.
Blinder, A.S. and L.J.Maccini, ‘Taking Stock: A Critical Assessment of Recent Research
on Inventories’, Journal of Economic Perspectives Winter 1991, 5:1, 73–96.
Calomiris, Charles W., ‘Financial Factors in the Great Depression’, Journal of Economic
Perspectives Spring 1993, 7:2.
Cooper, Russell, ‘Predetermined Wages and Prices and the Impact of Expansionary
Government Policy’, Review of Economic Studies April 1990, 57, 205–14.
New and old Keynesians 573
Grandmont, Jean-Michel, ‘Money and Value: A Reconsideration of Classical and
Neoclassical Monetary Theories’, Cambridge: Cambridge University Press, 1983.
Greenwald, B. and J.E.Stiglitz, ‘Externalities in Economies with Imperfect Information and
Incomplete Markets’, Quarterly Journal of Economics 101:2, May 1986 , 229–64.
Greenwald, B. and J.E.Stiglitz, ‘Imperfect Information, Credit Markets and Unemployment’,
European Economic Review February/March 1987, 31:2, 444–56.
Greenwald, B. and J.E.Stiglitz, ‘Pareto Inefficiency of Market Economies: Search and
Efficiency Wage Models’, American Economic Review 78:2, May 1988a, 351–55.
Greenwald, B. and J.E.Stiglitz, ‘Examining Alternative Macroeconomic Theories’,
Brookings Papers on Economic Activity 1988b, no. 1, 207–70.
Greenwald, B. and J.E.Stiglitz, ‘Imperfect Information, Finance Constraints and Business
Fluctuations’, in Meir Kohn and S.C.Tsiang (eds) Finance Constraints, Expectations
and Macroeconomics, Oxford: Oxford University Press, 1988c, 103–40.
Greenwald, B. and J.E.Stiglitz, ‘Toward a Theory of Rigidities’, American Economic
Review May 1989, 79:2, 364–69.
Greenwald, B. and J.E.Stiglitz, ‘Macroeconomic Models with Equity and Credit Rationing’,
in R.Glenn Hubbard (ed.) Information, Capital Markets and Investments, Chicago:
University of Chicago Press, 1990a, 15–42.
Greenwald, B. and J.E.Stiglitz, ‘Asymmetric Information and the New Theory of the
Firm: Financial Constraints and Risk Behavior’, American Economic Review May
1990b, 80:2, 160–65.
Greenwald, B. and J.E.Stiglitz, ‘Capital Market Imperfections and Labor Market
Adjustments’, paper presented at NBER/CEPR Conference on Labor Market
Dynamics, Cambridge, Massachusetts, October 1991a, in Conference Proceedings.
Greenwald, B. and J.E.Stiglitz, ‘Towards a Reformulation of Monetary Economies’, Caffee
Lectures presented to the University of Rome and the Bank of Italy, April 1991b,
Cambridge: Cambridge University Press, 1992.
Greenwald, B. and J.E.Stiglitz, ‘Financial Market Imperfections and Business Cycles’,
Quarterly Journal of Economics 108:1, 1993, 77–114.
Greenwald, B. and J.E.Stiglitz and A.Weiss, ‘Informational Imperfections in the Capital
Markets and Macroeconomic Fluctuations’, American Economic Review May 1984,
74:1, 194–99.
Hansen, Bent, A Study on the Theory of Inflation, London: Allen Press, 1951.
Jensen, Michael C., ‘Agency Costs of Free Cash Flow, Corporate Finance and Takeovers’,
American Economic Review May 1986, 76, 323–29.
King, R.G., ‘Will the New Keynesian Macroeconomics Resurrect the IS-LM Model?’,
Journal of Economic Perspectives Winter 1993, 7:1, 67–82.
Kirman, A., ‘Whom or What does the Representative Consumer Represent?’, Journal of
Economic Perspectives Spring 1992, 6:2, 117–36.
Lindbeck, Assar, ‘Macroeconomic Theory and Labor Markets’, European Economic
Review 1992, 36:2/3, 209–36.
Mankiw, N.Gregory, ‘Real Business Cycles: A New Keynesian Perspective’, Journal of
Economic Perspectives Summer 1989, 3:3, 79–90.
Neary, P. and J.E.Stiglitz, ‘Expectations, Asset Accumulation and the Real Balance Effect’,
Presented at the Dublin Meetings of the Econometric Society, September 1982.
Neary, P. and J.E.Stiglitz, ‘Toward a Reconstruction of Keynesian Economics: Expectations
and Constrained Equilibria’, Quarterly Journal of Economics Supplement 1983,
98:3, 199–228.
Newbery, D. and J.E.Stiglitz, ‘Wage Rigidity, Implicit Contracts, Unemployment and
Economic Efficiency’, Economic Journal June 1987, 97:386, 416–30.
Plosser, C.I., ‘Understanding Real Business Cycles’, Journal of Economic Perspectives
Summer 1989, 3:3, 51–77.
Romer, D., ‘The New Keynesian Synthesis’, Journal of Economic Perspectives Winter
1993, 7:1, 5–22.
574 Bruce Greenwald and Joseph Stiglitz
Ross, S.A., ‘The Economic Theory of Agency: The Principal’s Problem’, American
Economic Review May 1973, 63:2, 134–39.
Shapiro, C. and J.E.Stiglitz, ‘Equilibrium Unemployment as a Worker Discipline Device’,
American Economic Review June 1984, 74:3, 433–44. Reprinted in N.G. Mankiw
and D.Romer (eds) New Keynesian Economics, 2, Cambridge: MIT Press, 1991,
123–42.
Solow, R. and J.E.Stiglitz, ‘Output, Employment and Wages in the Short Run’, Quarterly
Journal of Economics XLXXXII, November 1968, 537–60.
Stiglitz, J.E., ‘Incentives and Risk Sharing in Sharecropping’, Review of Economic Studies
April 1974, 41:2 219–55.
Stiglitz, J.E., ‘Equilibrium Wage Distribution’, Economic Journal September 1985, 95:379,
595–618.
Stiglitz, J.E., ‘Competition and the Number of Firms in a Market: Are Duopolies More
Competitive than Atomistic Markets?’, Journal of Political Economy October 1987,
95:5, 1041–61.
Stiglitz, J.E., ‘Money, Credit and Business Fluctuations’, Economic Record December
1988, 307–22.
Stiglitz, J.E., ‘Methodological Issues and the New Keynesian Economies’, in A. Vercelli
and N.Dimitri (eds) Macroeconomics: A Survey of Research Strategies Oxford: Oxford
University Press, 1992a.
Stiglitz, J.E., ‘Capital Markets and Economic Fluctuations in Capitalist Economies’,
European Economic Review North-Holland, 36, 1992b, 269–306. (Marshall Lecture
prepared for the European Economic Association Annual Meeting, Cambridge, UK,
August 1991.)
Stiglitz, J.E. and A.Weiss, ‘Credit Rationing in Markets with Imperfect Information’,
American Economic Review June 1981, 71:3, 393–410.
Stiglitz, J.E. and A.Weiss, ‘Banks as Social Accountants and Screening Devices and the
General Theory of Credit Rationing’ in Essays in Monetary Economics in Honor of
Sir John Hicks , Oxford: Oxford University Press, 1992a.
Stiglitz, J.E. and A.Weiss, ‘Asymmetric Information in Credit Markets and Its Implications
for Macroeconomics’, Oxford Economic Papers, October 1992b (Special Issue),
44.4. Reprinted in A.S.Courakis (ed.) Financial Markets, Institutions, and Policy,
Oxford: Oxford University Press, 1993.
Part VI
The renaissance of economic
growth analysis
Introduction
In 1987 the Nobel Prize for Economics was awarded to Robert Solow for his
contributions to the theory of economic growth, the most important of which
were made in the 1950s (see Solow 1956; 1957). Solow’s research proved to
be highly influential and led to the development of the neoclassical growth
model which following the demise of the Harrod-Domar approach, has
remained the most popular basic framework for studying economic growth
(for an introduction to neoclassical growth theory see Gordon 1993; Barro
and Grilli 1994; Jansen et al. 1994; Mankiw 1994; Abel and Bernanke 1995;
Froyen 1995).
As we noted in Chapter 1 of this volume, there has been a reawakening of
economists’ interest into the long-run issue of growth since about the mid-
1980s. This work, known as endogenous growth theory, has sought to provide
a better understanding of the growth process and in so doing holds out the
prospect of providing insights which may be invaluable in helping to design
policies which could make a significant difference to the long-term growth
rate (see Crafts 1996). Solow’s neoclassical model shows how capital
accumulation could raise an economy’s growth rate in the medium term but
generates the prediction that the long-run (steady state) rate of growth is
constrained by the rate of growth of the labour force if the production
function exhibits diminishing returns to the variable factor, constant returns to
scale and zero technical progress. Hence technical progress which is assumed
exogenous in the Solow model came to be seen as the main driving force of
long-run growth. Following Solow’s (1957) article, where he showed how the
aggregate production function could be decomposed so that the contribution
of the various factor inputs towards the growth of output could be calculated,
it was discovered that a large part of measured growth was left unexplained.
This ‘Solow residual’, representing the growth of total factor productivity,
was viewed as a measure of technical progress. But since technological
change was exogenous, it meant that a large part of economic growth
remained unexplained in the neoclassical model. The Solow residual,
extracted by growth accounting procedures, was in Abramovitz’s words a
‘measure of our ignorance about the causes of growth’ (see Abramovitz 1956;
Fagerberg 1994). Endogenous growth theory has attempted to get round some
578 The renaissance of economic growth analysis
of these difficulties by reasserting the importance of capital accumulation,
proposing various ways of endogenizing technological change, emphasizing
the role of human capital and stressing the importance of research and
development. As a result endogenous growth theory raises the prospect that
sustained differences in the levels and growth rates of national income are
possible. This is in contrast to the Solow model where (conditional)
convergence across countries is implied with respect to either growth rates or
income levels (see Mankiw et al. 1992; Mankiw 1995). The large observed
differences in income between countries which persist is a problem for the
Solow model, where a country’s growth rate of per capita income should be
inversely related to its initial or starting level of income per head. In the
absence of shocks poor countries should tend to converge with rich countries
in terms of levels of per capita income.
The first article reprinted in this final part is the classic 1986 article by
Moses Abramovitz first published in the Journal of Economic History. In it
Abramovitz examines the catch-up hypothesis which asserts that ‘being
backward in the level of productivity carries a potential for rapid advance’.
Hence productivity growth rates across countries should be inversely related
to initial levels of productivity. Given that the United States is the world
leader in terms of productivity levels, other countries have the opportunity to
catch up providing they utilize unexploited technology. Providing the low
productivity countries have the ‘social capability’ to exploit the opportunities
available to them converging incomes and productivity should be observed.
Abramovitz examines the experience of a group of industrialized countries
over the past 100 years and finds support for the convergence of productivity
levels implied by the catch-up hypothesis particularly in the quarter century
following the Second World War.
Steve Dowrick’s article from the May 1992 issue of the Economic
Journal (reprinted on pp. 604–15) draws attention to the diverging growth
paths of the world’s economies which have led to a widening of income
disparities. Dowrick’s growth accounting exercise isolates the main
features and patterns of world economic growth. In particular they show
that there is supporting evidence for the technological spillover hypothesis
and that the poorest countries have tended to experience faster growth in
total factor productivity. However, this has not prevented income
divergence due to lower rates of investment and labour force participation
in the poorest countries compared to richer countries (see also Baumol
1986; Baumol et al. 1994).
As noted above, the neoclassical growth model, relying as it does on
exogenous population growth and technological change to explain long-run
growth, leaves little room for government intervention to enhance growth
performance. Keith Shaw’s 1992 Economic Journal paper (reprinted on pp.
616–27) examines the policy implications implied by the new endogenous
theories of growth. After first reviewing traditional and new growth theories,
Shaw goes on to discuss how differences in public policy can influence the
Introduction 579
‘incentives to acquire capital in both physical and human forms’. This is
particularly important in less advanced economies where protectionist
activity could lead to the transfer of skilled labour resources from the
knowledge-creating sector to the manufacturing sector leading to a decline in
growth-enhancing innovation. Also important is tax policy which by affecting
decisions to invest can have ‘real and permanent effects on the level and
growth rate of income’ (see also Stern 1991).
Paul Romer is widely acknowledged to be the leading figure in the new
growth economics. In his 1994 Journal of Economic Perspectives article
(reprinted on pp. 628–48) Romer traces the origins of the new ideas which
have dominated this branch of macroeconomic analysis during recent years.
Romer recounts two explanations which have been put forward to account
for the origin of these new theories. The first explanation, primarily
empirical in character, relates to the convergence controversy (discussed
earlier). The second is linked to the various theoretical attempts to replace
the model of perfect competition at the aggregate level. With respect to the
former issue of convergence Romer argues that this controversy ‘captures
only part of what endogenous growth has been all about’ and in many
respects ‘represents a digression from the main story behind endogenous
growth theory’. Romer argues that a significant factor driving the new
literature has been the desire to accommodate the fact that technological
change comes from what people actually do and that many individuals and
firms have market power and do not operate as price takers as in perfect
competition. In order to understand the important growth enhancing
processes of discovery, diffusion and technological advance, Romer argues
that economists must incorporate models of imperfect competition into the
analysis of economic growth. Such models are more consistent with the
empirical evidence and also call into question traditional perfectly
competitive neoclassical models where market incentives and government
policies have little impact on the crucial growth enhancing activities (see
Nelson and Romer 1996).
The final article is by Robert Solow and is also taken from the 1994
Journal of Economic Perspectives symposium on the analysis of economic
growth (reprinted on pp. 649–59). Solow seeks to place the new ideas on
endogenous growth in historical perspective. To do this Solow divides the
history of modern growth analysis into three waves, namely the Harrod-
Domar impulse, the neoclassical response and newer alternatives, i.e.
endogenous growth theories. Solow endorses the attempts made by
endogenous growth theorists to incorporate imperfect competition into their
models and also their emphasis on modelling the endogenous component of
technological change. However, he is also critical of some of the assumptions
arbitrarily introduced by new growth theorists. Solow concludes with a plea
for more research into developing workable hypotheses which may throw
light on ‘good ways to model the flow of productivity-increasing innovations
and improvements’ (see Grossman and Helpman 1994).
580 The renaissance of economic growth analysis
REFERENCES
*Titles marked with an asterisk are particularly recommended for additional reading.
*Abel, A.B. and B.S.Bernanke (1995) Macroeconomics, 2nd edn, Chapter 6, New York:
Addison Wesley.
Abramovitz, M. (1956) ‘Resource and Output Trends in the United States since 1870’,
American Economic Review, 46, May, pp. 5–23.
Abramovitz, M. (1986) ‘Catching Up, Forging Ahead, and Falling Behind’, Journal of
Economic History 46, June, pp. 385–406.
Barro, R.J. and V.Grilli (1994) European Macroeconomics, Chapter 14, London:
Macmillan.
Barro, R.J. and X.Sala-i-Martin (1995) Economic Growth, New York: McGrawHill.
*Baumol, W.J. (1986) ‘Productivity Growth, Convergence and Welfare: What the Long
Run Data Show’, American Economic Review 76, December, pp. 1072–85.
Baumol, W.J., R.R.Nelson and E.N.Wolff (eds) (1994) Convergence of Productivity:
Cross-National Studies and Historical Evidence, New York: Oxford University Press.
Boltho, A. and G.Holtham (1992) ‘New Approaches to Economic Growth’, Oxford
Review of Economic Policy 8, Winter, pp. 1–14.
*Crafts, N. (1996) ‘Post-Neoclassical Endogenous Growth Theory: What are its Policy
Implications?’, Oxford Review of Economic Policy 12, pp. 30–47.
Crafts, N. and G.Toniolo (eds) (1996) Economic Growth in Europe since 1945, Cambridge:
Cambridge University Press.
*Fagerberg, J. (1994) ‘Technology and International Differences in Growth Rates’, Journal
of Economic Literature 32, September, pp. 1147–75.
*Froyen, R.T. (1995) Macroeconomics: Theories and Policies, 5th edn, Chapter 17,
London: Prentice Hall.
*Gordon, R.J. (1993) Macroeconomics, 6th edn, Chapter 12, New York: HarperCollins.
Grossman, G. and E.Helpman (1994) ‘Endogenous Innovation in the Theory of Growth’,
Journal of Economic Perspectives 8, Winter, pp. 23–44.
*Jansen, D.W., C.D.Delorme and R.B.Ekelund, Jr (1994) Intermediate Macroeconomics,
Chapter 18, New York: West.
Lucas, R.E. Jr (1993) ‘Making a Miracle’, Econometrica 61, March, pp. 251–72.
*Mankiw, N.G. (1994) Macroeconomics, 2nd edn, Chapter 4, New York: Worth.
*Mankiw, N.G. (1995) ‘The Growth of Nations’, Brookings Papers on Economic Activity
pp. 275–326.
Mankiw, N.G., D.Romer and D.N.Weil (1992) ‘A Contribution to the Empirics of
Economic Growth’, Quarterly Journal of Economics 107, May, pp. 407–37.
*Nelson, R.R. and P.M.Romer (1996) ‘Science, Economic Growth and Public Policy’,
Challenge March-April, pp. 9–21.
Rebelo, S. (1991) ‘Long Run Policy Analysis and Long Run Growth’, Journal of Political
Economy 99, June, pp. 500–21.
Solow, R. (1956) ‘A Contribution to the Theory of Economic Growth’, Quarterly Journal
of Economics 70, February, pp. 65–94.
Solow, R. (1957) ‘Technical Change and the Aggregate Production Function’, Review of
Economics and Statistics 39, August, pp. 312–20.
Stern, N. (1991) ‘The Determinants of Growth’, Economic Journal 101, January, pp.
122–33.
Introduction 581
QUESTIONS
1 ‘Differences among countries in productivity levels create a strong potentiality
for subsequent convergence of levels, provided that countries have a social
capability adequate to absorb more advanced technologies’ (Abramovitz
1986). Explain and discuss.
2 Does convergence of productivity levels extend beyond the convergence club
of free market industrialized countries?
3 ‘One strength of Solow’s version of the neoclassical growth model is that,
despite its simplicity, it has many predictions’ (Mankiw 1995). What are the
predictions of the neoclassical growth model?
4 What are the main sources of growth in the Solow neoclassical growth model?
5 To what extent have endogenous growth theories been successful in accounting
for the long-run growth of per capita incomes without the need to invoke
exogenous technological progress?
6 ‘Better supply-side policies tend to have their pay-off in the far distant future’
(Crafts 1996). Does this mean that the short-termism of politicians will always
provide an obstacle to faster long-run growth?
7 What are the policy implications of endogenous growth theory?
8 ‘The main engine of growth is the accumulation of human capital—of
knowledge—and the main source of differences in living standards among
nations is differences in human capital. Physical capital accumulation plays
an essential but decidedly subsidiary role’ (Lucas 1993). Is this the essential
message of recent research into the growth process?
9 What factors led to the reawakening of economists’ interest into the long-run
issue of growth in the mid-1980s? What progress has been made?
10 ‘Economic growth is the part of macroeconomics that really matters’ (Barro
and Sala-i-Martin 1995). Critically examine this view.
23 Catching up, forging ahead, and
falling behind
Moses Abramovitz
Journal of Economic History (1986) 46, June, pp. 385–406
A widely entertained hypothesis holds that, in comparisons among countries,
productivity growth rates tend to vary inversely with productivity levels. A
century of experience in a group of presently industrialized countries supports
this hypothesis and the convergence of productivity levels it implies. The rate
of convergence, however, varied from period to period and showed marked
strength only during the first quarter-century following World War II. The
general process of convergence was also accompanied by dramatic shifts in
countries’ productivity rankings. This chapter extends the simple catch-up
hypothesis to rationalize the fluctuating strength of the process and explores
the connections between convergence itself and the relative success of early
leaders and latecomers.
Among the many explanations of the surge of productivity growth during
the quarter-century following World War II, the most prominent is the
hypothesis that the countries of the industrialized ‘West’ were able to bring
into production a large backlog of unexploited technology. The principal part
of this backlog is deemed to have consisted of methods of production and of
industrial and commercial organization already in use in the United States at
the end of the war, but not yet employed in the other countries of the West. In
this hypothesis, the United States is viewed as the ‘leader’, the other countries
as ‘followers’ who had the opportunity to ‘catch up’. In conformity with this
view, a waning of the opportunity for catching up is frequently advanced as
an explanation of the retardation in productivity growth suffered by the same
group of followers since 1973. Needless to say, the size of the initial backlog
and its subsequent reduction are rarely offered as sole explanations of the
speedup and slowdown, but they stand as important parts of the story.
These views about postwar following and catching up suggest a more
general hypothesis that the productivity levels of countries tend to converge.
And this in turn brings to mind old questions about the emergence of new
leaders and the historical and theoretical puzzles that shifts in leadership and
relative standing present—matters that in some respects fit only awkwardly
with the convergence hypothesis.
The pertinence of all these questions to an understanding of modern
economic growth obviously demands their continued study. The immediate
Catching up, forging ahead, and falling behind 583
occasion for this chapter, however, is the appearance of Angus Maddison’s
compilation of historical time series of the levels and growth of labor
productivity covering sixteen industrialized countries from 1870 to 1979.1
These data enable us to observe the catch-up process in quantitative terms
over a much longer span of time than was possible hitherto. At the same time,
the evidence of Maddison’s tables raises again the historical puzzles posed by
productivity leadership and its shifts.
THE CATCH-UP HYPOTHESIS
The hypothesis asserts that being backward in level of productivity carries a
potential for rapid advance. Stated more definitely the proposition is that in
comparisons across countries the growth rates of productivity in any long
period tend to be inversely related to the initial levels of productivity.
The central idea is simple enough. It has to do with the level of
technology embodied in a country’s capital stock. Imagine that the level of
labor productivity were governed entirely by the level of technology
embodied in capital stock. In a ‘leading country’, to state things sharply,
one may suppose that the technology embodied in each vintage of its stock
was at the very frontier of technology at the time of investment. The
technological age of the stock is, so to speak, the same as its chronological
age. In an otherwise similar follower whose productivity level is lower, the
technological age of the stock is high relative to its chronological age. The
stock is obsolete even for its age. When a leader discards old stock and
replaces it, the accompanying productivity increase is governed and limited
by the advance of knowledge between the time when the old capital was
installed and the time it is replaced. Those who are behind, however, have
the potential to make a larger leap. New capital can embody the frontier of
knowledge, but the capital it replaces was technologically superannuated.
So—the larger the technological and, therefore, the productivity gap
between leader and follower, the stronger the follower’s potential for growth
in productivity; and, other things being equal, the faster one expects the
follower’s growth rate to be. Followers tend to catch up faster if they are
initially more backward.
Viewed in the same simple way, the catch-up process would be self-limiting
because as a follower catches up, the possibility of making large leaps by
replacing superannuated with best-practice technology becomes smaller and
smaller. A follower’s potential for growth weakens as its productivity level
converges towards that of the leader.
This is the simple central idea. It needs extension and qualification. There
are at least four extensions:
1 The same technological opportunity that permits rapid progress by
modernization encourages rapid growth of the capital stock partly
because of the returns to modernization itself, and partly because
584 Moses Abramovitz
technological progress reduces the price of capital goods relative to the
price of labor. So—besides a reduction of technological age towards
chronological age, the rate of rise of the capital-labor ratio tends to be
higher. Productivity growth benefits on both counts. And if circumstances
make for an acceleration in the growth of the capital stock its
chronological age also falls.2
2 Growth of productivity also makes for increase in aggregate output. A
broader horizon of scale-dependent technological progress then comes
into view.
3 Backwardness carries an opportunity for modernization in disembodied,
as well as in embodied, technology.
4 If countries at relatively low levels of industrialization contain large
numbers of redundant workers in farming and petty trade, as is normally
the case, there is also an opportunity for productivity growth by
improving the allocation of labor.
Besides extension, the simple hypothesis also needs qualification.
First, technological backwardness is not usually a mere accident.
Tenacious societal characteristics normally account for a portion, perhaps a
substantial portion, of a country’s past failure to achieve as high a level of
productivity as economically more advanced countries. The same
deficiencies, perhaps in attenuated form, normally remain to keep a
backward country from making the full technological leap envisaged by the
simple hypothesis. I have a name for these characteristics. Following Kazushi
Ohkawa and Henry Rosovsky, I call them ‘social capability’.3 One can
summarize the matter in this way. Having regard to technological
backwardness alone leads to the simple hypothesis about catch-up and
convergence already advanced. Having regard to social capability, however,
we expect that the developments anticipated by that hypothesis will be clearly
displayed in cross-country comparisons only if countries’ social capabilities
are about the same. One should say, therefore, that a country’s potential for
rapid growth is strong not when it is backward without qualification, but
rather when it is technologically backward but socially advanced.
The trouble with absorbing social capability into the catch-up hypothesis is
that no one knows just what it means or how to measure it. In past work I
identified a country’s social capability with technical competence, for
which—at least among Western countries—years of education may be a
rough proxy, and with its political, commercial, industrial, and financial
institutions, which I characterized in more qualitative ways.4 I had in mind
mainly experience with the organization and management of large-scale
enterprise and with financial institutions and markets capable of mobilizing
capital for individual firms on a similarly large scale. On some occasions the
situation for a selection of countries may be sufficiently clear. In explaining
postwar growth in Europe and Japan, for example, one may be able to say
with some confidence that these countries were competent to absorb and
Catching up, forging ahead, and falling behind 585
exploit then existing best-practice technology. More generally, however,
judgments about social capability remain highly problematic. A few
comments may serve to suggest some of the considerations involved as well
as the speculative nature of the subject.
One concerns the familiar notion of a trade-off between specialization and
adaptability. The content of education in a country and the character of its
industrial, commercial, and financial organizations may be well designed to
exploit fully the power of an existing technology; they may be less well fitted
to adapt to the requirements of change. Presumably, some capacity to adapt is
present everywhere, but countries may differ from one another in this respect,
and their capacities to adapt may change over time.
Next, the notion of adaptability suggests that there is an interaction
between social capability and technological opportunity. The state of
education embodied in a nation’s population and its existing institutional
arrangements constrains it in its choice of technology. But technological
opportunity presses for change. So countries learn to modify their institutional
arrangements and then to improve them as they gain experience. The
constraints imposed by social capability on the successful adoption of a more
advanced technology gradually weaken and permit its fuller exploitation.
Thorstein Veblen said it this way:
There are two lines of agency visibly at work shaping the habits of thought
of [a] people in the complex movements of readjustment and rehabilitation
[required by industrialization]. These are the received scheme of use and
wont and the new state of the industrial arts; and it is not difficult to see
that it is the latter that makes for readjustment; nor should it be any more
difficult to see that the readjustment is necessarily made under the
surveillance of the received scheme of use and wont.5
Social capability, finally, depends on more than the content of education and
the organization of firms. Other aspects of economic systems count as well—
their openness to competition, to the establishment and operation of new firms,
and to the sale and purchase of new goods and services. Viewed from the other
side, it is a question of the obstacles to change raised by vested interests,
established positions, and customary relations among firms and between
employers and employees. The view from this side is what led Mancur Olson
to identify defeat in war and accompanying political convulsion as a radical
ground-clearing experience opening the way for new men, new organizations,
and new modes of operation and trade better fitted to technological potential.6
These considerations have a bearing on the notion that a follower’s
potential for rapid growth weakens as its technological level converges on the
leader’s. This is not necessarily the case if social capability is itself
endogenous, becoming stronger—or perhaps weaker—as technological gaps
close. In the one case, the evolution of social capability connected with
catching up itself raises the possibility that followers may forge ahead of even
586 Moses Abramovitz
progressive leaders. In the other, a leader may fall back or a follower’s
pursuit may be slowed.
There is a somewhat technical point that has a similar bearing. This is the
fact, noticed by Kravis and Denison, that as followers’ levels of per capita
income converge on the leader’s, so do their structures of consumption and
prices.7 R.C.O.Matthews then observed that the convergence of consumption
and production patterns should make it easier, rather than more difficult, for
followers to borrow technology with advantage as productivity gaps close.8
This, therefore, stands as still another qualification to the idea that the catch-
up process is steadily self-limiting.
The combination of technological gap and social capability defines a country’s
potentiality for productivity advance by way of catch-up. This, however, should
be regarded as a potentiality in the long run. The pace at which the potentiality
is realized depends on still another set of causes that are largely independent of
those governing the potentiality itself. There is a long story to tell about the
factors controlling the rate of realization of potential.9 Its general plot, however,
can be suggested by noting three principal chapter headings:
1 The facilities for the diffusion of knowledge—for example, channels of
international technical communication, multinational corporations, the
state of international trade and of direct capital investment.
2 Conditions facilitating or hindering structural change in the composition
of output, in the occupational and industrial distribution of the
workforce, and in the geographical location of industry and population.
Among other factors, this is where conditions of labor supply, the
existence of labor reserves in agriculture, and the factors controlling
internal and international migration come in.
3 Macroeconomic and monetary conditions encouraging and sustaining
capital investment and the level and growth of effective demand.
Having considered the technological catch-up idea, with its several extensions
and qualifications, I can summarize by proposing a restatement of the
hypothesis as follows.
Countries that are technologically backward have a potentiality for
generating growth more rapid than that of more advanced countries, provided
their social capabilities are sufficiently developed to permit successful
exploitation of technologies already employed by the technological leaders.
The pace at which potential for catch-up is actually realized in a particular
period depends on factors limiting the diffusion of knowledge, the rate of
structural change, the accumulation of capital, and the expansion of demand.
The process of catching up tends to be self-limiting, but the strength of the
tendency may be weakened or overcome, at least for limited periods, by
advantages connected with the convergence of production patterns as
followers advance towards leaders or by an endogenous enlargement of social
capabilities.
Catching up, forging ahead, and falling behind 587
HISTORICAL EXPERIENCE WITH CATCHING UP
I go on now to review some evidence bearing on the catch-up process. The
survey I make is limited to the sixteen countries covered by the Maddison
estimates of product per worker hour for nine key years from 1870 to 1979.10
The estimates are consistently derived as regards gross domestic product and
worker hours and are adjusted as regards levels of product per worker hour
Table 23.1 Comparative levels of productivity, 1870–1979
means and relative variance of the relatives of fifteen countries
compared with the United States
(US GDP per manhour=100)a
Notes: a 1870 and 1890. Figures in parentheses are based on
relatives with the United Kingdom=100
b Standard deviation divided by mean
Source: Calculated from Angus Maddison, Phases of Capitalist
Development (New York, 1982), Tables 5.2 and C.10
Table 23.2 The association (rank correlation) between initial levels and
subsequent growth rates of labor productivity
(GDP per manhour in sixteen countries, 1870–1979)
Source of underlying data: Maddison, Phases, Tables 5.1, 5.2, and C. 10
588 Moses Abramovitz
by the Kravis estimates of purchasing power parities for postwar years. I have
compressed the message of these data into three measures (see Tables 23.1
and 23.2).
1 Averages of the productivity levels of the various countries relative to
that of the United States, which was the leading country for most of the
period. (For 1870 and 1890, I have also calculated averages of relatives
based on the United Kingdom.) I calculate these averages for each of the
nine key years and use them to indicate whether productivity levels of
followers, as a group, were tending to converge on that of the leader.11
2 Measures of relative variance around the mean levels of relative
productivity. These provide one sort of answer to the question of whether
the countries that started at relatively low levels of productivity tended
to advance faster than those with initially higher levels.
3 Rank correlations between initial levels of productivity and subsequent
growth rates. If the potential supposedly inherent in technological
backwardness is being realized, there is likely to be some inverse
correlation; and if it works with enough strength to dominate other
forces the coefficients will be high.
The data I use and the measures I make have a number of drawbacks. The
data, of course, have the weaknesses that are inherent in any set of estimates
of GDP and manhours, however ably contrived, that stretch back far into the
nineteenth century. Beyond that, however, simple calculations such as I have
made fail, in a number of respects, to isolate the influence of the catch-up
hypothesis proper.
To begin with, my measures do not allow for variation in the richness of
countries’ natural resources in relation to their populations. Labor productivity
levels, therefore, are not pure reflections of levels of technology. In the same
way, these levels will also reflect past accumulations of reproducible capital,
both physical and human, and these may also be independent of technological
levels in one degree or another. Further, the measured growth rates of labor
productivity will be influenced by the pace of capital accumulation. As already
said, differences in rates of accumulation may reflect countries’ opportunities to
make advances in technology, but rates of capital formation may also be
independent, to some degree, of countries’ potentials for technological advance.
Finally, my measures make no allowance for countries’ variant abilities to
employ current best-practice technology for reasons other than the differences in
social capability already discussed. Their access to economies of scale is
perhaps the most important matter. If advanced technology at any time is
heavily scale-dependent and if obstacles to trade across national frontiers,
political or otherwise, are important, large countries will have a stronger
potential for growth than smaller ones.
There are many reasons, therefore, why one cannot suppose that the
expectations implied by the catch-up hypothesis will display themselves
Catching up, forging ahead, and falling behind 589
clearly in the measures I present. It will be something if the data show some
systematic evidence of development consistent with the hypothesis. And it will
be useful if this provides a chance to speculate about the reasons why the
connections between productivity levels and growth rates appear to have been
strong in some periods and weak in others.
Other countries, on the average, made no net gain on the United States in
a period longer than a century (Table 23.1, col. 1). The indication of very
limited, or even zero, convergence is really stronger than the figures suggest.
This is because the productivity measures reflect more than gaps in
technology and in reproducible capital intensity, with respect to which catch-
up is presumably possible. As already said, they also reflect differences in
natural resource availabilities which, of course, are generally favorable to the
United States and were far more important to the United States and to all the
other countries in 1870 than they are today. In 1870, the agricultural share of
United States employment was 50 percent; in 1979, percent. For the other
fifteen countries, the corresponding figures are 48 and 8 percent on the
average. The declines were large in all the countries.12 So the US advantage
in 1870 depended much more on our favorable land-person ratio than it did
in 1979. Putting it the other way, other countries on the average must have
fallen back over the century in respect to the productivity determinants in
respect to which catch-up is possible.
In other respects, however, one can see the influence of the potential for
catching up clearly. The variance among the productivity levels of the fifteen
‘follower’ countries declines drastically over the century—from a coefficient
of variation of 0.5 in 1870 to 0.15 in 1979. Not only that: the decline in
variance was continuous from one key year to the next, with only one
reversal—in the period across World War II. In the same way, the inverse
rank correlation between the initial productivity levels in 1870 and
subsequent growth rates over increasingly long periods becomes stronger and
stronger, until we reach the correlation coefficient of -0.97 across the entire
109 years.13 (Again there was the single reversal across World War II when
the association was actually—and presumably accidentally—positive.)
I believe the steadily declining variance measures and the steadily rising
correlation coefficients should be interpreted to mean that initial productivity
gaps did indeed constitute a potentiality for fast growth that had its effect
later if not sooner. The effect of the potentiality became visible in a very
limited degree very early. But if a country was incapable of, or prevented
from, exploiting that opportunity promptly, the technological growth
potential became strong, and the country’s later rate of advance was all the
faster. Though it may have taken a century for obstacles or inhibitions to be
fully overcome, the net outcome was that levels of productivity tended
steadily to even out—at least within the group of presently advanced
countries in my sample.
This last phrase is important. Mine is a biased sample in that its
members consist of countries all of whom have successfully entered into the
590 Moses Abramovitz
process of modern economic growth. This implies that they have acquired
the educational and institutional characteristics needed to make use of
modern technologies to some advanced degree. It is by no means assured—
indeed, it is unlikely—that a more comprehensive sample of countries would
show the same tendency for levels of productivity to even out over the same
period of time.14
This is the big picture. How do things look if we consider shorter periods?
There are two matters to keep in mind: the tendency to convergence within
the group of followers; and the convergence—or lack of it—of the group of
followers vis-à-vis the United States. I take up the second matter in the next
section. As to the convergence within the follower group, the figures suggest
that the process varied in strength markedly from period to period. The main
difference was that before World War II it operated weakly or at best with
moderate strength. For almost a quarter-century following the war it
apparently worked with very great strength. Why?
Before World War II, it is useful to consider two periods, roughly the
decades before 1913, and those that followed. In the years of relative peace
before 1913 I suggest that the process left a weak mark on the record for two
reasons, both connected with the still early state of industrialization in many
of the countries. First, the impress of the process was masked because farming
was still so very important; measured levels of productivity, therefore,
depended heavily on the amount and quality of farmland in relation to
population. Productivity levels, in consequence, were erratic indicators of
gaps between existing and best-practice technology. Second, social
competence for exploiting the then most advanced methods was still limited,
particularly in the earlier years and in the more recent latecomers. As the pre-
World War I decades wore on, however, both these qualifying circumstances
became less important. One might therefore have expected a much stronger
tendency to convergence after 1913. But this was frustrated by the irregular
effects of the Great War and of the years of disturbed political and financial
conditions that followed, by the uneven impacts of the Great Depression itself
and of the restrictions on international trade.
The unfulfilled potential of the years 1913–38 was then enormously
enlarged by the effects of World War II. The average productivity gap behind
the United States increased by 39 percent between 1938 and 1950; the poorer
countries were hit harder than the richer. These were years of dispersion, not
convergence.
The post-World War II decades then proved to be the period when—
exceptionally—the three elements required for rapid growth by catching
up came together.15 The elements were large technological gaps; enlarged
social competence, reflecting higher levels of education and greater
experience with large-scale production, distribution, and finance; and
conditions favoring rapid realization of potential. This last element refers
to several matters. There was on this occasion (it was otherwise after
World War I) a strong reaction to the experience of defeat in war, and a
Catching up, forging ahead, and falling behind 591
chance for political reconstruction. The postwar political and economic
reorganization and reform weakened the power of monopolistic groupings,
brought new men to the fore, and focused the attention of governments on
the tasks of recovery and growth, as Mancur Olson has argued.16 The
facilities for the diffusion of technology improved. International markets
were opened. Large labor reserves in home agriculture and immigration
from Southern and Eastern Europe provided a flexible and mobile labor
s u p p l y. G o v e r n m e n t s u p p o r t , t e c h n o l o g i c a l o p p o r t u n i t y, a n d a n
environment of stable international money favored heavy and sustained
capital investment. The outcome was the great speed and strength of the
postwar catch-up process.17
Looking back now on the record of more than a century, we can see that
catching up was a powerful continuing element in the growth experience of
the presently advanced industrial countries. The strength of the process varied
from period to period. For decades it operated only erratically and with
weakened force. The trouble at first lay in deficient social capability, a
sluggish adaptation of education and of industrial and financial organization
to the requirements of modern large-scale technology. Later, the process was
checked and made irregular by the effects of the two world wars and the
ensuing political and financial troubles and by the impact of the Great
Depression. It was at last released after World War II. The results were the
rapid growth rates of the postwar period, the close cross-country association
between initial productivity levels and growth rates, and a marked reduction
of differences in productivity levels, among the follower countries, and
between them and the United States.
Looking to the future, it seems likely that this very success will have
weakened the potentiality for growth by catching up among the group of
presently advanced countries. The great opportunities carried by that
potential now pass to the less developed countries of Latin America and Asia.
FORGING AHEAD AND FALLING BEHIND
The catch-up hypothesis in its simple form does not anticipate changes in
leadership nor, indeed, any changes in the ranks of countries in their relative
levels of productivity. It contemplates only a reduction among countries in
productivity differentials. Yet there have been many changes in ranks since
1870 and, of course, the notable shift of leadership from Britain to the United
States towards the end of the last century.18 This was followed by the
continuing decline of Britain’s standing in the productivity scale. Today there
is a widely held opinion that the United States is about to fall behind a new
candidate for leadership, Japan, and that both Europe and the United States
must contemplate serious injury from the rise of both Japan and a group of
still newer industrializing countries.
Needless to say, this chapter cannot deal with the variety of reasons—all
still speculative—for the comparative success of the countries that advanced
592 Moses Abramovitz
in rank and the comparative failure of those that fell back.19 I focus instead on
a few matters that help illustrate the ramifications of the catch-up process and
reveal the limitations of the simple hypothesis considered in earlier sections.
The congruity of technology and resources: United States as leader
Why did the gap between the United States and the average of other countries
resist reduction so long? Indeed, why did it even appear to become larger
between 1870 and 1929—before the impact of World War II made it larger
still? I offer three reasons:
1 The path of technological change which in those years offered the greatest
opportunities for advance was at once heavily scale-dependent and biased
in a labor-saving but capital-and resource-using direction. In both respects
the United States enjoyed great advantages compared with Europe or
Japan. Large-scale production was favored by a large, rapidly growing,
and increasingly prosperous population. It was supported also by a
striking homogeneity of tastes. This reflected the country’s comparative
youth, its rapid settlement by migration from a common base on the
Atlantic, and the weakness and fluidity of its class divisions. Further,
insofar as the population grew by immigration, the new Americans and
their children quickly accepted the consumption patterns of their adopted
country because the prevailing ethos favored assimilation to the dominant
native white culture. At the same time, American industry was encouraged
to explore the rich possibilities of a labor-saving but capital-and resource-
using path of advance. The country’s resources of land, forest, and
minerals were particularly rich and abundant, and supplies of capital
grew rapidly in response to high returns.20
2 By comparison with the United States and Britain, many, though not all,
of the ‘followers’ were also latecomers in respect to social capability. In
the decades following 1870, they lacked experience with large-scale
production and commerce, and in one degree or another they needed to
advance in levels of general and technical education.
3 World War I was a serious setback for many countries but a stimulus to
growth in the United States. European recovery and growth in the
following years were delayed and slowed by financial disturbances and
by the impact of territorial and political change. Protection, not
unification, was the response to the new political map. The rise of social
democratic electoral strength in Europe favored the expansion of union
power, but failed to curb the development and activities of industrial
cartels. Britain’s ability to support and enforce stable monetary
conditions had been weakened, but the United States was not yet able or,
indeed, willing to assume the role of leadership that Britain was losing.
In all these ways, the response to the challenge of war losses and defeat
after World War I stands in contrast to that after World War II.
Catching up, forging ahead, and falling behind 593
Points (2) and (3) were anticipated in earlier argument, but Point (1)
constitutes a qualification to the simple catch-up hypothesis. In that view,
different countries, subject only to their social capability, are equally
competent to exploit a leader’s path of technological progress. That is not
so, however, if that path is biased in resource intensity or if it is scale-
dependent. Resource-rich countries will be favored in the first instance,
large countries in the second. If the historical argument of this section is
correct, the United States was favored on both counts for a long time; it
may not be so favored in the future. Whether or not this interpretation of
American experience is correct, the general proposition remains: countries
have unequal abilities to pursue paths of progress that are resource-biased
or scale-dependent.
Interaction between followers and leaders
The catch-up hypothesis in its simple form is concerned with only one aspect
of the economic relations among countries: technological borrowing by
followers. In this view, a one-way stream of benefits flows from leaders to
followers. A moment’s reflection, however, exposes the inadequacy of that
idea. The rise of British factory-made cotton textiles in the first industrial
revolution ruined the Irish linen industry. The attractions of British and
American jobs denuded the Irish population of its young men. The beginnings
of modern growth in Ireland suffered a protracted delay. This is an example
of the negative effects of leadership on the economies of those who are
behind. Besides technological borrowing, there are interactions by way of
trade and its rivalries, capital flows, and population movements. Moreover,
the knowledge flows are not solely from leader to followers. A satisfactory
account of the catch-up process must take account of these multiple forms of
interaction. Again, there is space only for brief comment.
Trade and its rivalries
I have referred to the sometimes negative effects of leading-country exports on
the economies of less developed countries. Countries in the course of catching
up, however, exploit the possibilities of advanced scale-dependent
technologies by import substitution and expansion of exports. When they are
successful there are possible negative effects on the economies of leaders. This
is an old historical theme. The successful competition of Germany, the United
States and other European countries is supposed to have retarded British
growth from 1870 to 1913 and perhaps longer.21 Analogous questions arise
today. The expansion of exports from Japan and the newer industrializing
countries has had a serious impact on the older industries of the United States
and Europe, as well as some of the newer industries.
Is there a generalized effect on the productivity growth of the leaders?
The effect is less than it may seem to be because some of the trade shifts are
594 Moses Abramovitz
a reflection of overall productivity growth in the leader countries
themselves. As the average level of productivity rises, so does the level of
wages across industries generally. There are then relative increases in the
product prices of those industries—usually older industries—in which
productivity growth is lagging and relative declines in the product prices of
those industries enjoying rapid productivity growth. The former must suffer
a loss of comparative advantage, the latter a gain. One must keep an eye
on both.
Other causes of trade shifts that are connected with the catch-up process
itself may, however, carry real generalized productivity effects. There are
changes that stem from the evolution of ‘product cycles’, such as Raymond
Vernon has made familiar. And perhaps most important, there is the
achievement of higher levels of social capability. This permits followers to
extend their borrowing and adaptation of more advanced methods, and
enables them to compete in markets they could not contest earlier.
What difference does it make to the general prospects for the productivity
growth of the leading industrial countries if they are losing markets to
followers who are catching up?
There is an employment effect. Demand for the products of export-and
import-competing industries is depressed. Failing a high degree of
flexibility in exchange rates and wages and of occupational and
geographical mobility, aggregate demand tends to be reduced. Unless
macroeconomic policy is successful, there is general unemployment and
underutilization of resources. Profits and the inducements to invest and
innovate are reduced. And if this condition causes economies to succumb
to protectionism, particularly to competitive protectionism, the difficulty
is aggravated.
International trade theory assures us that these effects are transitory.
Autonomous capital movements aside, trade must, in the end, balance. But
the macroeconomic effects of the balancing process may be long drawn out,
and while it is in progress, countries can suffer the repressive effects of
restricted demand on investment and innovation.
There is also a Verdoorn effect. It is harder for an industry to push the
technological frontier forward, or even to keep up with it, if its own rate of
expansion slows down—and still harder if it is contracting. This is
unavoidable but tolerable when the growth of old industries is restricted by
the rise of newer, more progressive home industries. But when retardation of
older home industries is due to the rise of competing industries abroad, a
tendency to generalized slowdown may be present.
Interactions via population movements
Nineteenth-century migration ran in good part from the farms of Western and
Southern Europe to the farms and cities of the New World and Australia. In
the early twentieth century, Eastern Europe joined in. These migrations
Catching up, forging ahead, and falling behind 595
responded in part to the impact on world markets of the cheap grains and
animal products produced by the regions of recent settlement. Insofar they
represent an additional but special effect of development in some members of
the Atlantic community of industrializing countries on the economies of other
members.
Productivity growth in the countries of destination was aided by
migration in two respects. It helped them exploit scale economies; and by
making labor supply more responsive to increase in demand, it helped
sustain periods of rapid growth. Countries of origin were relieved of the
presence of partly redundant and desperately poor people. On the other
hand, the loss of population brought such scale disadvantages as accompany
slower population growth, and it made labor supply less responsive to
industrial demand.
Migration in the postwar growth boom presents a picture of largely similar
design and significance. In this period the movement was from the poorer,
more slowly growing countries of Southern Europe and North Africa to the
richer and more rapidly growing countries of Western and Northern Europe.22
There is, however, this difference: The movement in more recent decades was
induced by actual and expected income differences that were largely
independent of the market connections of countries of origin and destination.
There is no evidence that the growth boom of the West itself contributed to
the low incomes of the South.
Needless to say, migrations are influenced by considerations other than
relative levels of income and changing comparative advantage. I stress
these matters, however, because they help us understand the complexities of
the process of catch-up and convergence within a group of connected
countries.
Interaction via capital flows
A familiar generalization is that capital tends to flow from countries of high
income and slow growth to those with opposite characteristics or, roughly
speaking, from leaders to followers. One remembers, however, that that
description applies to gross new investments. There are also reverse flows
that reflect the maturing of past investments. So in the early stages of a
great wave of investment, followers’ rates of investment and productivity
growth are supported by capital movement while those of leaders are
retarded. Later, however, this effect may become smaller or be reversed, as
we see today in relations between Western leaders and Latin American
followers.
Once more, I add that the true picture is far more complicated than this
idealized summary. It will hardly accommodate such extraordinary
developments as the huge American capital import of recent years, to say
nothing of the Arabian-European flows of the 1970s and their reversal now
underway.
596 Moses Abramovitz
Interactions via flows of applied knowledge
The flow of knowledge from leader to followers is, of course, the very essence
of the catch-up hypothesis. As the technological gaps narrow, however, the
direction changes. Countries that are still a distance behind the leader in
average productivity may move into the lead in particular branches and
become sources of new knowledge for older leaders. As they are surpassed in
particular fields, old leaders can make gains by borrowing as well as by
generating new knowledge. In this respect the growth potential of old leaders
is enhanced as the pursuit draws closer. Moreover, competitive pressure can
be a stimulus to research and innovation as well as an excuse for protection.
It remains to be seen whether the newly rising economies will seek to guard a
working knowledge of their operations more closely than American
companies have done, and still more whether American and European firms
will be as quick to discover, acquire, and adapt foreign methods as Japanese
firms have been in the past.
Development as a constraint on change: tangible capital
The rise of followers in the course of catching up brings old leaders a mixed
bag of injuries and potential benefits. Old leaders, however, or followers who
have enjoyed a period of successful development, may come to suffer
disabilities other than those caused by the burgeoning competitive power of
new rivals. When Britain suffered its growth climacteric a century ago,
observers thought that its slowdown was itself due in part to its early lead.
Thorstein Veblen was a pioneer proponent of this suggestion, and Charles
Kindleberger and others have picked it up again.23 One basis for this view is
the idea that the capital stock of a country consists of an intricate web of
interlocking elements. They are built to fit together, and it is difficult to
replace one part of the complex with more modern and efficient elements
without a costly rebuilding of other components. This may be handled
efficiently if all the costs and benefits are internal to a firm. When they are
divided among different firms and industries and between the private and
public sectors, the adaptation of old capital structures to new technologies
may be a difficult and halting process.
What this may have meant for Britain’s climacteric is still unsettled.
Whatever that may be, however, the problem needs study on a wider scale
as it arises both historically and in a contemporaneous setting. After World
War II, France undertook a great extension and modernization of its public
transportation and power systems to provide a basis for later development
of private industry and agriculture. Were the technological advances
embodied in that investment program easier for France to carry out because
its infrastructure was technically older, battered, and badly maintained? Or
was it simply a heavy burden more in need of being borne? There is a
widespread complaint today that the public capital structure of the United
Catching up, forging ahead, and falling behind 597
States stands in need of modernization and extension. Is this true, and, if it
is, does it militate seriously against the installation of improved capital by
private industry? One cannot now assume that such problems are the
exclusive concern of a topmost productivity leader. All advanced industrial
countries have large accumulations of capital, interdependent in use but
divided in ownership among many firms and between private and public
authorities. One may assume, however, that the problem so raised differs in
its impact over time and among countries and, depending on its importance,
might have some influence on the changes that occur in the productivity
rankings of countries.
Development as a constraint on change: intangible capital and political
institutions
Attention now returns to matters akin to social capability. In the simple catch-
up hypothesis, that capability is viewed as either exogenously determined or
else as adjusting steadily to the requirements of technological opportunity.
The educational and institutional commitments induced by past development
may, however, stand as an obstacle. That is a question that calls for study.
The comments that follow are no more than brief indications of prominent
possibilities.
The United States was the pioneer of mass production as embodied in
the huge plant, the complex and rigid assembly line, the standardized
product, and the long production run. It is also the pioneer and developer
of the mammoth diversified conglomerate corporation. The vision of
business carried on within such organizations, their highly indirect,
statistical, and bureaucratic methods of consultation, planning and
decision, the inevitable distractions of trading in assets rather than
production of goods—these mental biases have sunk deep into the
American business outlook and into the doctrine and training of young
American managers. The necessary decentralization of operations into
multiple profit centers directs the attention of managers and their superiors
to the quarterly profit report and draws their energies away from the
development of improved products and processes that require years of
attention.24 One may well ask how well this older vision of management
and enterprise and the organizational scheme in which it is embodied will
accommodate the problems and potentialities of the emerging computer
and communications revolution. Or will that occur more easily in
countries where educational systems, forms of corporate organization, and
managerial outlook can better make a fresh start?
The long period of leadership and development enjoyed by the United
States and the entire North Atlantic community meant, of course, a great
increase of incomes. The rise of incomes, in turn, afforded a chance to satisfy
latent desires for all sorts of non-market goods ranging from maintenance in
old age to a safeguarded natural environment. Satisfying these demands,
598 Moses Abramovitz
largely by public action, has also afforded an ample opportunity for special
interest groups to obtain privileges and protection in a process that Mancur
Olson and others have generalized.
The outcome of this conjuncture of circumstances and forces is the mixed
economy of the West, the complex system of transfers, taxes, regulations, and
public activity, as well as organizations of union and business power, that had
its roots long before World War II, that expanded rapidly during the growth
boom of the 1950s and 1960s, and that reached very high levels in the 1970s.
This trend is very broadly consistent with the suggestion that the elaboration
of the mixed economy is a function of economic growth itself. To this one has
to add the widely held idea advanced by Olson and many others that the
system operates to reduce enterprise, work, saving, investment, and mobility
and, therefore, to constrict the processes of innovation and change that
productivity growth involves.
How much is there in all this? The answer turns only partly on a
calculation of the direct effects of the system on economic incentives. These
have proved difficult to pin down, and attempts to measure them have
generally not yielded large numbers, at least for the United States.25 The
answer requires an equally difficult evaluation of the positive roles of
government activity. These include not only the government’s support of
education, research, and information, and its provision of physical overhead
capital and of the host of local functions required for urban life. We must
remember also that the occupational and geographical adjustments needed
to absorb new technology impose heavy costs on individuals. The
accompanying changes alter the positions, prospects, and power of
established groups, and they transform the structure of families and their
roles in caring for children, the sick, and the old. Technical advance,
therefore, engenders conflict and resistance; and the Welfare State with its
transfers and regulations constitutes a mode of conflict resolution and a
means of mitigating the costs of change that would otherwise induce
resistance to growth. The existing empirical studies that bear on the
economic responses to government intervention are, therefore, far from
meeting the problem fully.
If the growth-inhibiting forces embodied in the Welfare State and in
private expressions of market power were straightforward, positive
functions of income levels, uniform across countries, that would be another
reason for supposing that the catch-up process was self-limiting. The
productivity levels of followers would, on this account, converge towards
but not exceed the leader’s. But these forces are clearly not simple, uniform
functions of income. The institutions of the Welfare State have reached a
higher degree of elaboration in Europe than in the United States. The
objects of expenditure, the structures of transfers and taxes, and people’s
responses to both differ from country to country. These institutional
developments, therefore, besides having some influence on growth rates
Catching up, forging ahead, and falling behind 599
generally, may constitute a wild card in the deck of growth forces. They
will tend to produce changes in the ranks of countries in the productivity
scale and these may include the top rank itself.
A sense that forces of institutional change are now acting to limit the
growth of Western countries pervades the writings of many economists—and,
of course, of other observers. Olson, Fellner, Scitovsky, Kindleberger,
Lindbeck, and Giersch are only a partial list of those who see these economies
as afflicted by institutional arthritis or sclerosis or other metaphorical malady
associated with age and wealth.
These are the suggestions of serious scholars, and they need to be taken
seriously. One may ask, however, whether these views take account of still
other, rejuvenating forces which, though they act slowly, may yet work
effectively to limit and counter those of decay—at least for the calculable
future. In the United States, interregional competition, supported by free
movement of goods, people, and capital, is such a force. It limits the power of
unions and checks the expansion of taxation, transfers, and regulation.26
International competition, so long as it is permitted to operate, works in a
similar direction for the United States and other countries as well, and it is
strengthened by the development in recent years of a more highly integrated
world capital market and by more vigorous international movements of
corporate enterprise.
In the ranking of countries within the group of presently advanced
industrial economies, their variant responsiveness to competition may be still
another influence making for change in rank and relative level of
productivity. As this group competes with the newly industrializing countries
of the East and South, however, the pressures of competition on their
institutional development, as distinct from their impact on particular
industries, should help the older group maintain a lead. There are, however,
still more solid grounds for a renewal of productivity advance in both Europe
and the United States and for the maintenance of a substantial lead over
virtually all newcomers. These are their high levels of general and technical
education, the broad bases of their science, and the well-established
connections of their science, technology, and industry. These elements of
social capability are slow to develop but also, it seems very likely, slow to
decay.
Finally, it is widely recognized that the process of institutional aging,
whatever its significance, is not one without limits. Powerful forces continue
to push that way, and they are surely strong in resisting reversal. Yet it is also
apparent that there is a drift of public opinion that works for modification
both in Europe and North America. There is a fine balance to be struck
between productivity growth and the material incomes it brings and the other
dimensions of social welfare. Countries are now in the course of readjusting
that balance in favor of productivity growth. How far they can go and,
indeed, how far they should go are both still in question.
600 Moses Abramovitz
CONCLUDING REMARKS
This chapter points in two directions. It shows that differences among
countries in productivity levels create a strong potentiality for subsequent
convergence of levels, provided that countries have a ‘social capability’
adequate to absorb more advance technologies. It reminds us, however, that
the institutional and human capital components of social capability develop
only slowly as education and organization respond to the requirements of
technological opportunity and to experience in exploiting it. Their degree of
development acts to limit the strength of technological potentiality proper.
Further, the pace of realization of a potential for catch-up depends on a
number of other conditions that govern the diffusion of knowledge, the
mobility of resources and the rate of investment.
The long-term convergence to which these considerations point, however,
is only a tendency that emerges in the average experience of a group of
countries. The growth records of countries on their surface do not exhibit
the uniformly self-limiting character that a simple statement of the catch-up
hypothesis might suggest. Dramatic changes in productivity rankings mark
the performance of a group’s individual members. Some causes of these
shifts in rank are exogenous to the convergence process. The state of a
country’s capability to exploit emerging technological opportunity depends
on a social history that is particular to itself and that may not be closely
bound to its existing level of productivity. And there are changes in the
character of technological advance that make it more congruent with the
resources and institutional outfits of some countries but less congruent with
those of others. Some shifts, however, are influenced by the catch-up process
itself—for example, when the trade rivalry of advancing latecomers makes
successful inroads on important industries of older leaders. There are also
the social and political concomitants of rising wealth itself that may weaken
the social capability for technological advance. There is the desire to avoid
or mitigate the costs of growth, and there are the attractions of goals other
than growth as wealth increases. A reasonably complete view of the catch-
up process, therefore, does not lend itself to simple formulation. Its
implications ramify and are hard to separate from the more general process
of growth at large.
ACKNOWLEDGEMENTS
The author acknowledges with thanks critical comments and suggestions by
Paul David and Knick Harley. This chapter is the revision of a draft read to
the Economic History Association at its New York meeting in September
1985. This, in turn, was a greatly abbreviated version of a longer paper since
published. See ‘Catching Up and Falling Behind’, Fackföreningsrörelsens
Institut für Ekonomisk Forskning (Trade Union Institute for Economic
Research), Economic Research Report no. 1 (Stockholm, 1986).
Catching up, forging ahead, and falling behind 601
NOTES
1 Angus Maddison, Phases of Capitalist Development (New York, 1982). Maddison’s
estimates of productivity levels are themselves extrapolations of base levels established
for most, but not all, countries by Irving B.Kravis, Alan Heston, and Robert Summers
in their International Comparisons of Real Product and Purchasing Power
(Baltimore, 1978) and in other publications by Kravis and his associates.
2 W.E.G.Salter, Productivity and Technical Change (Cambridge, 1960) provides a
rigorous theoretical exposition of the factors determining rates of turnover and
those governing the relation between productivity with capital embodying best
practice and average (economically efficient) technology.
3 K.Ohkawa and H.Rosovsky, Japanese Economic Growth: Trend Acceleration in
the Twentieth Century (Stanford, 1973), especially ch. 9.
4 Moses Abramovitz, ‘Rapid Growth Potential and its Realization: The Experience of
the Capitalist Economies in the Postwar Period’, in Edmond Malinvaud (ed.)
Economic Growth and Resources, Proceedings of the Fifth World Congress of the
International Economic Association, vol. 1 (London, 1979), pp. 1–30.
5 Thorstein Veblen, Imperial Germany and the Industrial Revolution (New York,
1915), p. 70.
6 Mancur Olson, The Rise and Fall of Nations: Economic Growth, Stagflation and
Social Rigidities (New Haven, 1982).
7 Kravis et al., International Comparisons; Edward F.Denison, assisted by Jean-Pierre
Poullier, Why Growth Rates Differ, Postwar Experience of Nine Western Countries
(Washington, DC, 1967). pp. 239–45.
8 R.C.O.Matthews, Review of Denison (1967), Economic Journal (June 1969), pp.
261–8.
9 My paper cited in note 4 describes the operation of these factors in the 1950s and
1960s and tries to show how they worked to permit productivity growth to rise in
so many countries rapidly, in concert and for such an extended period (‘Rapid
Growth Potential and Its Realization’, pp. 18–30).
10 The countries are Australia, Austria, Belgium, Canada, Denmark, Finland, France,
Germany, Italy, Japan, Netherlands, Norway, Sweden, Switzerland, United Kingdom
and United States.
11 In these calculations I have treated either the United States or the United Kingdom as
the productivity leader from 1870 to 1913. Literal acceptance of Maddison’s
estimates, however, make Australia the leader from 1870–1913. Moreover, Belgium
and the Netherlands stand slightly higher than the United States in 1870. Here are
Maddison’s relatives for those years (from Phases, Table 5.2):
Since Australia’s high standing in this period mainly reflected an outstandingly
favorable situation of natural resources relative to population, it would be misleading
to regard that country as the technological leader or to treat the productivity changes
in other countries relative to Australia’s as indicators of the catch-up process. Similarly,
602 Moses Abramovitz
the small size and specialized character of the Belgian and Dutch economies make
them inappropriate benchmarks.
12 Maddison, Phases, Table C5.
13 Since growth rates are calculated as rates of change between standings at the terminal
dates of periods, errors in the estimates of such standings will generate errors in the
derived growth rates. If errors at both terminal dates were random, and if those at
the end-year were independent of those at the initial year, there would be a tendency
on that account for growth rates to be inversely correlated with initial-year standings.
The inverse correlation coefficients would be biased upwards. Note, however, that if
errors at terminal years were random and independent and of equal magnitude,
there would be no tendency on that account for the variance of standings about the
mean to decline between initial and end-year dates. The error bias would run against
the marked decline in variance that we observe. Errors in late-year data, however, are
unlikely to be so large, so an error bias is present.
14 See also William J.Baumol, ‘Productivity Growth, Convergence and Welfare: What
the Long-run Data Show’, C.V.Starr Center for Applied Economics, New York
University, Research Report no. 85–27, August 1985.
15 See Abramovitz, ‘Rapid Growth Potential and its Realization’.
16 Olson, Rise and Fall.
17 Some comments on the catch-up process after 1973 may be found in Abramovitz,
‘Catching Up and Falling Behind’ (Stockholm, 1986), pp. 33–9.
18 If one follows Maddison’s estimates (Phases, Table C.I9), the long period from 1870
to 1979 saw Australia fall by 8 places in the ranking of his 16 countries, Italy by
Switzerland by 8, and the United Kingdom by 10. Meanwhile the United States rose
by 4, Germany by Norway by 5, Sweden by 7, and France by 8.
19 The possibility of overtaking and surpassing, however, was considered theoretically
by Edward Ames and Nathan Rosenberg in a closely reasoned and persuasive
article, ‘Changing Technological Leadership and Industrial Growth’, Economic
Journal, 72 (1963), pp. 13–31. They conclude that the troubles connected with
leadership and industrial “aging” that doom early leaders to decline in the
productivity scale are not persuasive. They hold that outcomes turn on a variety of
empirical conditions, the presence of which is uncertain and not foreordained.
20 These arguments are anticipated and elaborated in Nathan Rosenberg’s fertile and
original chapter, ‘Why in America?’, in Otto Mayr and Robert Post (eds) Yankee
Enterprise: The Rise of the American System of Manufactures (Washington, DC, 1981).
21 See also R.C.O.Matthews, Charles Feinstein and John Odling-Smee, British Economic
Growth, 1856–1973 (Stanford, 1983), chs. 14, 15, 17. Their analysis does not find
a large effect on British productivity growth from 1870 to 1913.
22 The migration from East to West Germany in the 1950s was a special case. It
brought to West Germany educated and skilled people strongly motivated to rebuild
their lives and restore their fortunes.
23 Charles P.Kindleberger, ‘Obsolescence and Technical Change’, Oxford Institute of
Statistics Bulletin (Aug. 1961), pp. 281–97.
24 These and similar questions are raised by experienced observers of American business.
They are well summarized by Edward Denison, Trends in American Economic
Growth, 1929–1982, (Washington, DC, 1985), ch. 3.
25 Representative arguments supporting the idea that social capability has suffered,
together with some quantitative evidence, may be found in Olson, Rise and Fall,
William Fellner, ‘The Declining Growth of American Productivity: An Introductory
Note’, in W.Fellner (ed.) Contemporary Economic Problems, 1979 (Washington,
Catching up, forging ahead, and falling behind 603
DC, 1979); and Assar Lindbeck, ‘Limits to the Welfare State’, Challenge (Dec. 1985).
For argument and evidence on the other side, see Sheldon Danzigar, Robert Haveman,
and Robert Plotnick, ‘How Income Transfers Affect Work, Savings and Income
Distribution’, Journal of Economic Literature, 19 (Sept. 1982), pp. 975–1028; and
E.F.Denison, Accounting for Slower Economic Growth (Washington, DC, 1979),
pp. 127–38.
26 See R.D.Norton, ‘Regional Life Cycles and US Industrial Rejuvenation’, in Herbert
Giersch (ed.) Towards an Explanation of Economic Growth (Tübingen, 1981), pp.
253–80; and R.D.Norton, ‘Industrial Policy and American Renewal’, Journal of
Economic Literature, 24 (March 1986).
24 Technological catch up and
diverging incomes
Patterns of economic growth 1960–88
Steve Dowrick
Economic Journal (1992) 102, May, pp. 600–10
INTRODUCTION
The pattern of worldwide economic growth since the early 1960s displays
diverging growth paths. Most economies shared the experience of high
growth rates in the 1950s and 1960s, reverting in the 1970s and 1980s to
rates which are more normal by historical standards. At the same time,
however, income disparities across the national economies of the world have
been widening. The richer economies have, in per capita terms, been growing
faster than the middle-income economies, which in turn have out-paced the
poorest economies. Moreover, within each of these broadly defined groups,
income levels have been diverging.
The divergence of growth paths of GDP per capita is perhaps surprising.
The post-war period has witnessed an explosion in world trade,
communications and the dissemination of information—all factors which
might be supposed to both encourage and enable the technologically
backward economies to learn from and adopt the production techniques of the
more advanced. At the same time, the integration of capital markets, the
emerging dominance of transnational corporations and the development of
both transport and communications technology might be expected to lead to
growth-enhancing investment in the poorer, low-wage economies.
The first of these conjectures is supported by an analysis of the sources of
economic growth. There is indeed evidence to support the technological
spillover hypothesis: the less advanced economies have tended to experience
faster growth in multi-factor productivity (although not necessarily with
respect to manufacturing technology in the poorest economies). It appears,
therefore, that income divergence has occurred in spite of technological
catching up. The proximate causes are lower rates of investment in the poorer
countries allied to declining rates of labour force participation in the poorer
countries and rising participation in the richer countries.
There are several explanations for the divergence of the growth paths of
capital and labour inputs. Employment growth, relative to total population,
is strongly enhanced in the medium term by the demographic transition
from high to low rates of population growth. The poorest group of countries
Technological catch up and diverging incomes 605
have tended to experience, however, either stable or even increasing rates of
population growth since the early 1960s. Population has tended to
decelerate only in those countries which already had higher income levels
by 1960.
There is some weak statistical evidence that aggregate rates of return on
capital investment across poorer countries may be higher than in richer
countries, in which case their low investment rates might be attributable to
capital barriers and an inability to generate substantial domestic savings out
of near-subsistence incomes. Moreover, it seems likely that complementarities
between private capital investment on the one hand and human capital and
public infrastructure on the other, lower the private returns to investment in
the poorer economies.
THE PATTERN OF GROWTH 196–4 TO 1984–8 FOR 113
COUNTRIES
Summers and Heston (1991) continue to feed the ongoing analysis of
worldwide economic growth with improved estimates of real GDP and its
principal components for an ever increasing number of countries. The sample
of countries used here consists of 113 out of their total of 138. The 25
exclusions were made either because of lack of data on key variables before
1965, or because a country belongs to OPEC. The sample breaks down into
42 African countries, 25 American, 20 Asian, 22 European and 4 Southern
Pacific.
I take as the base for measurement of economic growth the annual
average value of real GDP (1985 US$) over the period 1960–4. The end-
point is the average for the period 1984–8. The purpose of taking these five-
year averages is to remove cyclical variation from the cross-country
comparisons. Real output is deflated either by population or by the
workforce to give approximate indicators of per capita incomes and labour
productivity. Examination of the ranked productivity measure reveals
natural breaks which divide the sample into three (in a similar way to
Baumol’s 1986 divisions). The high productivity group, with 1960–4 output
per worker above $6,500 contains most of the European economies and the
richer American and Asian economies. The low productivity group, with
output per worker below $2,800 contains most of the African economies and
the poorer Asian economies. The middle group contains mostly Asian and
Latin American countries. Since output per head of population and output
per worker are very highly correlated (r=0.98), I also refer to these groups
as the rich, poor and middle-income economies—although such descriptions
ignore variations in average incomes which may be due to foreign
ownership and the terms of trade.
Table 24.1 displays some of the key statistics for the sample and for each
of the three groups. Rows 4 and 5 reveal that while the average growth rates
of GDP per worker are very similar for the three groups (ranging from 1.90
606 Steve Dowrick
to 2.06 per cent per annum for the poor and rich groups respectively), growth
of GDP per capita is substantially lower in the poor group (1.36 per cent per
annum) than in the middle group (2.16 per cent) and the rich group (2.49 per
cent). In other words, although average labour productivity across the three
groups displays only a weak tendency towards divergence, this tendency is
amplified strongly when output is measured per head of population. The
proximate reason is given in rows 11 and 12 of Table 24.1: the workforce
has, on average, grown slower than population in the poorer countries, and
vice versa in the rich countries.
The dispersion of income and productivity levels within each group can be
read from rows 6–9 of Table 24.1. In each case within group dispersion
increased between 1960–4 and 1984–8. Divergence appears to be somewhat
stronger in output per capita than in output per worker. The overall picture is
one of increasing dispersion in incomes and productivity both within and
between the three income groups. Since the early 1960s it appears that the
world’s economies have been on divergent growth paths, leading to increasing
inequality, especially in per capita output.
This divergence in world income levels can be seen in Figure 24.1 which
plots the growth of per capita GDP against the logarithm of 1960–4 real GDP
for the 113 countries. First, it is evident that growth rates have varied
tremendously, within a range of -2 to +7 per cent per annum for all but the
richest 20 countries (where the range has been between 1 and 4 per cent).
Second, a weak upward drift in the per cent scatter points is just about
discernible, at least with the aid of a least squares regression line as
displayed. Average incomes in the richer countries are tending to pull away
Table 24.1 Average values of principal variables
Notes:
a 1962 is the average for the period 1960–4 and 1986 is the average for
1984–8
b Growth rates are annual averages of logarithmic growth rates 1962–86
c Dispersion is the standard deviation of the logarithm
Technological catch up and diverging incomes 607
from the levels in the poorer countries. Closer examination also reveals a
tendency for income convergence amongst the very richest group of
countries—at the extreme right of Figure 24.1—roughly speaking the OECD
group.
Evidence such as this has been taken by some commentators, for example
DeLong (1988) to refute arguments that there has been a tendency for poorer
economies to catch up on the richer. Nevertheless, Dowrick and Nguyen
(1989) have shown that at least within the OECD group of countries there has
been a strong and consistent tendency since 1950 for technological catch up to
occur, even though income levels have not been converging since 1973. In
other words, after taking account of the rate of growth of labour and capital,
the poorer OECD countries have tended to experience faster growth in the
residual (multi-factor productivity).
If we repeat this growth accounting exercise on the 113 economies in our
sample, we reach a similar conclusion. There is a strong tendency, at least
within the upper and lower ends of the world income distribution, for multi-
factor productivity growth to be inversely related to the starting level of
productivity. Simple regression of GDP growth on initial labour
productivity (In Y/L0), the growth of the workforce and average
investment rates (INV) gives estimates of the growth accounting parameters
within each of the three income groupings. The first parameter captures the
extent of technological catch up (if negative); the second parameter measures
the elasticity of output with respect to employment; the third is an estimate of
the marginal productivity of gross capital investment. The regression results
are summarized below, omitting the constant terms, with heteroscedasticity-
consistent t-statistics in parentheses.
Figure 24.1 Growth of real GDP per capita 1960–4 to 1984–8 for 113 countries
608 Steve Dowrick
Poor economies
=-0.013 lnY/L
0
+0.89 +0.15INV
(-2.2) (2.3) (4.0) (n=42, R2=0.33, S.E.=1.87)
Middle income economies
=0.019 lnY/L
0
+0.48 +0.05 INV
(1.4) (0.9) (0.9) (n=42, R2=0.13, S.E.=1.88)
Rich economies
=-0.014 lnY/L
0
+0.95 +0.13 INV
(-3.1) (5.9) (5.6) (n=39, R2=0.63, S.E.=0.94)
It is only within the middle-income group that we fail to find evidence of
technological catch up. Within this group, however, the standard errors are
large, due to collinearity amongst the explanatory variables. The coefficients
estimated on the other two groups are well defined and remarkably similar.
The hypothesis that the parameters are in fact the same across all three
samples is not rejected at even the 20 per cent level (F8,101=1.38)—so it is
legitimate on statistical criteria to pool the samples. Pooling the three samples
gives weight to the inter-group variation in the data, as well as the intra-
group variation, so these are the preferred estimates.
All economies
=-0.006 lnY/L
0
+0.88 +0.11 INV
(-3.1) (4.6) (4.2) (n=113, R2=0.28, S.E.=1.63)
Technological catch-up is strongly significant across the whole sample. So too
are the growth of the workforce and the rate of gross investment although
nearly three-quarters of the variance in growth rates remains unexplained.1
(1)
(2)
(3)
(4)
Figure 24.2 Growth of multi-factor productivity, 1960–88
Technological catch up and diverging incomes 609
Figure 24.2 illustrates the technological catch-up tendency by plotting growth
in multi factor productivity against productivity levels.2 We can discern the
tendency for a negative correlation, with dispersion around the regression line
most pronounced for the middle-income countries.
Using these parameter estimates it is possible to decompose observed
rates of growth of per capita GDP into four elements: the amount
attributable to technological catch up; the amount due to the growth of the
workforce relative to the total population; the amount due to investment;
and the unexplained residual. The decomposition method is explained in
Dowrick and Nguyen (1989:1025). Results for each of 113 countries are
available from the author on request. Table 24.2 gives the average values
for each of the three income groups. The decomposition is also given for a
group of five rapidly growing Asia-Pacific economies consisting of Japan,
Korea, Taiwan, Singapore and Hong Kong. Their exceptional growth
record makes study of this group of newly industrializing countries (NICs)
of particular interest.
Output per capita in the poorest group of countries grew at 0.6 percentage
points per annum below the sample average. This performance occurred
despite the ‘advantage of backwardness’, or catching up effect, which
afforded them 0.6 points in above average productivity growth. Overall,
then, these economies underperformed, relative to the world average, by 1.2
points. This underperformance is, on average, attributable to two factors: the
decline in employment relative to population and the poor rate of growth of
the capital stock, which contributed -0.5 and -0.7 points respectively. The
Table 24.2 Decomposition of relative growth performance for four groups of
countries
Note: The contribution of each factor is derived from the regression parameters in (4)
and the deviation of variables for each country from the overall mean. See Appendix
for details.
610 Steve Dowrick
employment shallowing effect reflects the fact that the workforce failed to
grow at the same rate as population. The capital shallowing effect reflects the
fact that investment rates were only just over one-half of average investment
rates in the rich economies, despite much higher population growth in the
poor countries.
When this growth accounting exercise is repeated for the rich group of
countries it presents an almost exact mirror image. Above average per capita
growth of 0.5 points, on top of the slower productivity growth due to less
opportunity for technological catch up, implies that the richer countries’
growth performance was some 1.1 percentage points per annum above what
might have been expected. This performance is partly due to employment
deepening, which contributed 0.4 points, but is mainly due to capital
deepening which contributed 0.6 points.
The five rapidly growing Asian economies have each outperformed the
world economy by over four percentage points per annum, with the exception
of Japan whose growth rate has been only 3.1 points above average. These
exceptional growth rates imply that per capita output in each of these
countries is now more than double what it would have been if they had grown
at the world average rate of 2 per cent per annum. The growth decomposition
in Table 24.2 suggests that nearly half of this exceptional performance can be
attributed to faster than average growth in factor inputs relative to
population. It is particularly interesting to note that employment deepening,
i.e. raising the ratio of the workforce to the population, has been relatively
more important than capital deepening in three of these five countries. It is
only in Japan that capital deepening has substantially outweighed the
employment deepening effect.
Increasing the ratio of workers to population has also made significant
contributions to the growth record of a number of other countries, notably
the USA and Portugal (in each case contributing an average of 0.7
percentage points per annum) and Iceland and Malta (1.1 points). Capital
deepening has been particularly important in contributing to above average
annual growth rates in Norway (1.9 points), West Germany, Denmark,
France, Israel, Italy, Spain, Cyprus, Malaysia, Malta and Yugoslavia
(above 1 point in each case).
The relationship between divergence of world incomes on the one hand
and both employment deepening and capital deepening on the other hand is
illustrated in Figures 24.3 and 24.4. Figure 24.3 plots the growth in the
workforce to population ratio against the logarithm of real GDP per capita.
The simple regression line highlights the moderately strong positive
correlation in the data (r=+0.53). Since the early 1960s, the richer countries
have tended to experience a rise in the ratio of workers to population, whilst
in the poorer countries the ratio has tended to decline. Note also that some of
the middle-income countries (including Singapore, Hong Kong and South
Korea) have experienced particularly large rises in this aggregate
participation ratio.
Technological catch up and diverging incomes 611
Figure 24.4 presents a corresponding plot of investment rates against
1960–4 levels of per capita GDP. There is again a clear positive correlation
(r=+0.58). Richer countries tend to devote a greater proportion of their output
to physical investment than do the poorer countries. Allied with employment
deepening, it is this relative capital deepening in the more advanced
economies which provides a proximate explanation for the divergence of
world per capita GDP despite technological catch up.
EXPLANATIONS AND POLICIES
The growth-accounting exercises of the previous section are important in
isolating the immediate features and patterns of world economic growth.
They do not necessarily provide much depth of explanation, nor much guide
for constructive policies by either national governments or development
Figure 24.3 Growth of ratio of workers to population, 1962–86
Figure 24.4 Investment rates, 1960–88
612 Steve Dowrick
agencies, except to add weight to ‘motherhood’ prescriptions such as that
more investment and employment will promote economic growth! I will
discuss in turn some of the mechanisms that may underlie the three sources of
growth that have been identified: catch up, employment deepening and
capital deepening.
The most obvious explanation of observed productivity catch up is
technological spillover, or the ability of less advanced economies to imitate
and copy the techniques of production used by more advanced economies.
Abramovitz (1986) warns, however, that such spillovers may not be
effective if a country lacks the technical and social capability to absorb and
implement new ideas. This hypothesis is supported by several studies which
suggest reasons why catch up may be limited in the poorest economies of
the world. Dowrick and Gemmell (1991) report evidence that although the
poorest countries of sub-Saharan Africa have managed some catch-up in
agricultural productivity, they have tended to fall behind in industrial
pr oductivity, where complementarities with human capital and
infrastructural development are likely to be particularly strong. Barro
(1991) finds that low levels of educational enrolment are a substantial
impediment to growth.3
We have seen in the previous section that the contribution of employment
deepening to the faster economic growth of the richer economies is
substantial. This does not reflect, however, any systematic increase in the
rate of participation by adults in the workforce. Although rising
participation is important in a number of individual countries, on average it
is offset by increasing aged dependency. The average ratio of workers to
adults (over 15 years of age) in the richer economies has barely increased at
all since the early 1960s. The major demographic change affecting the
aggregate participation ratio in the richer countries has been the decline in
youth dependency. This in turn reflects declining birth rates in most
advanced economies. Brander and Dowrick (1990) argue that even though
economic growth is independent of steady-state birth rates, declining birth
rates provide a temporary but substantial stimulus through this labour
supply effect.
There is evidence to support this view in Figure 24.5 where the growth of
the worker to population ratio is plotted against the change in rates of
population growth between the beginning and end of the sample period. A
strong negative correlation (r=-0.69) supports the idea that declining fertility
raises, albeit temporarily, the ratio of adults to total population which in turn
raises the ratio of workers to population for a given adult participation rate.
Moreover, to the extent that female participation in the labour force is a
substitute for child-rearing, declining fertility may have an additional positive
participation effect.
The importance of demographic change for the five fastest growing Asian
economies is illustrated in Table 24.3. With the exception of Japan, all
countries experienced a sharp rise in the ratio of workers to population. In
Technological catch up and diverging incomes 613
each of these four cases the main impetus was a rise in the ratio of adults to
population, i.e. a decline in youth dependency rates. In the case of Japan, the
fall in youth dependency more than offset a fall in the adult participation
rate. In all five cases, the rising ratio of adults to population was associated
with a sharp fall in the rate of growth of population.
We can infer from these figures that the halving of population growth over
three decades, and the consequent maturing of the population age structure,
has increased substantially the supply of labour relative to total population.4
The labour supply effect of demographic change has contributed substantially
to the rapid increase in per capita output in these countries, and it is probable
that the rapid increase in per capita incomes has also contributed to the
demographic change.
Turning to investment, and referring back to Figure 24.4, the pattern to be
explained here is the strong tendency for rates of investment to rise with
income levels. In the absence of international capital movements, where
investment has to be funded entirely out of domestic savings, this pattern
might simply reflect sharply diminishing marginal utility of consumption, i.e.
if the cost of forgoing present consumption is higher, the nearer a person is to
a subsistence income. This would imply, however, that real rates of return to
investment should be substantially higher in the poorer countries and that
internationally mobile capital should flow to these economies. The simple
regression estimates (l)-(3) of gross returns to investment are difficult to
interpret because of the large standard error on the estimate for the middle-
income countries. There is some weak evidence that rates of return are
slightly higher in the low income countries, although the differences are not
statistically significant. It is possible then that it is barriers to capital mobility
which are holding back investment rates in the poorer countries.
Figure 24.5 Growth in the ratio of workers to population
614 Steve Dowrick
A more likely general explanation of the pattern of investment rates would
seem to be concerned with complementarities between physical investment
and human and social capital. Simple neo-classical growth models based on
diminishing returns to capital imply that real rates of return should be highest
in those countries with the lowest capital stock. It seems to be the case,
however, that backward infrastructure and low levels of education and
training substantially reduce rates of return to physical investment. Whilst
investment in the physical and social infrastructure might generate a higher
return, these returns would not generally be captured by private investors.
The problem then becomes one of directing international investment and aid
into infrastructural development and education and training.
NOTES
1 The magnitude of the factor input parameters might be taken to imply that there are
increasing returns to scale since the output-labour elasticity is almost 0.9 and the
output-capital elasticity is approximated by multiplying the coefficient 0.11 by the
capital-output ratio which is typically between two and three. Some of the unexplained
variance in the regression is due to differences in openness to world trade, as in
Romer (1990), and variance in industrial structure, as in Dowrick and Gemmell
(1991) as well as a multiplicity of country-specific policies and natural events.
Government consumption is not a significant explanatory variable; however there is
evidence that government consumption crowds out investment, which would explain
Table 24.3 Demographic change in five NICs
Technological catch up and diverging incomes 615
Barro’s (1991) result. The addition of dummy variables for Asia, Africa, Europe
and Latin America raises the R2 to 0.35, statistically significant at the 5% level but
not at the 1% level. The parameter estimates are not affected significantly by these
additions, except that the coefficient on investment falls from 0.11 to 0.09. A further
test of the regression model has been carried out by dividing the observations for
each country into two periods, 1960–75 and 1975–88. The hypothesis that the
regression coefficients are stable over the two periods is accepted (F
3,218
=0.4) and the
coefficient estimates for the pooled samples are very similar to those reported here.
2 Multi-factor productivity is defined here as the growth in GDP minus the contribution
of the growth of the workforce and of investment (relative to the sample averages)
calculated using the coefficients of equation (4).
3 Note however that measurement of technological catch-up in these broad cross-
section studies is fraught with difficulties. First, we have to rely on gross investment
rates to capture changes in capital stocks and gross employment to capture aggregate
labour input. Second, there are substantial gaps in standard national accounting
techniques which may lead to systematic mismeasurement of GDP growth due to the
omission of domestic production and the omission of production externalities such
as pollution and the maintenance of environmental standards.
4 This is the immediate consequence of lowering birth rates. There will in future,
however, be a rebound effect due to increases in the aged dependency rate.
REFERENCES
Abramovitz, M. (1986) ‘Catching up, forging ahead and falling behind’, Journal of
Economic History 46, pp. 385–406.
Barro, Robert J. (1991) ‘Economic growth in a cross-section of countries’, Quarterly
Journal of Economics 106, pp. 407–43.
Baumol, William J. (1986) ‘Productivity growth, convergence and welfare: what the
long-run data show’, American Economic Review 76, pp. 1072–85.
Brander, James A. and Dowrick, Steve (1990) ‘The role of fertility and population in
economic growth: new results from aggregate cross-national data’, Centre for
Economic Policy Research Discussion Paper no. 232, Australian National University,
Canberra.
DeLong, J.Bradford (1988) ‘Productivity growth, convergence and welfare: comment’,
American Economic Review 78, pp. 1138–54.
Dowrick, Steve and Gemmell, Norman (1991) ‘Industrialisation, catching up and economic
growth: a comparative study across the world’s capitalist economies’, Economic
Journal 101, pp. 263–75.
Dowrick, Steve and Nguyen, Duc-Tho (1989) ‘OECD comparative economic growth
1950–85: catch up and convergence’, American Economic Review 79, pp. 1010–30.
Romer, Paul (1990) ‘Endogenous technological change’, Journal of Political Economy
98, pp. S71–S102.
Summers, Robert and Heston, Alan (1991) ‘The Penn World Table (Mark 5): an expanded
set of international comparisons, 1950–88’, Quarterly Journal of Economics 106,
pp. 327–68.
25 Policy implications of endogenous
growth theory
G.K.Shaw
Economic Journal (1992) 102, May, pp. 611–21
INTRODUCTION
Harrod-Domar Growth Theory sanctioned the overriding importance of
capital accumulation in the quest for enhanced growth. Since budgetary
surpluses could substitute for domestic savings, fiscal policy became identified
as the primary growth instrument. Government had a role to play.
The influence of Harrod-Domar economics was far reaching. Development
agencies gave great prominence to the need to raise savings ratios. In doing
so, they were reflecting the spirit of contemporary development doctrine.
Lewis (1954, 1958), Rostow (1960) and Fei and Ranis (1964) had pinpointed
the raising of the savings ratio as the key to understanding the process of
development and the ‘take-off’ into sustained growth. In advanced market
economies concerned with the apparent success of Soviet economic growth,
similar sentiments were expressed and the ‘strict fiscal—easy money’
prescription for economic growth was frequently advocated.
The introduction of the neo-classical growth model, especially in the
contributions of Solow (1956) and Swan (1956) provided the necessary
antidote to the excessive claims made for capital accumulation. In
neoclassical one sector models, the ultimate determinant of the growth rate is
shown to be the autonomously determined rate of population expansion.
Fiscal policy is thus rendered an irrelevancy in the pursuit of higher growth
per se, although it may still have a part to play in the more esoteric pursuit of
Golden Rules to achieve Golden Ages.
What the early Harrod-Domar and neo-classical formulations of growth
theory possessed in common was the belief that the third ingredient in growth,
namely technical progress, was an exogenously determined, fortuitous and
costless occurrence—descending like manna from the heavens. Even though it
was recognized that technical progress could be the dominant element in the
growth equation, especially following the publication of Solow (1957), there
was no satisfactory account of the determinants of technical change. Indeed,
in the neo-classical growth model with exogenous population expansion and
exogenous technical change there was virtually no role for government to
play. Discussion turned to the rather sterile issue of whether government
Policy implications of endogenous growth theory 617
intervention could speed up the process of adjustment in the event of some
temporary disturbance from the steady state path. But there was no growth
policy as such.
This essentially unsatisfactory position has continued until very recent
times. The present paper offers an attempt to detail recent contributions which
seek to rectify this situation and to draw out the more important implications
for economic policy.
THE SOLOW GROWTH FORMULA
Solow (1957) devised a framework for distinguishing the contributions of
labour, capital and technical change to economic growth. This pioneering
contribution formed the basis of growth accounting exercises pursued by
growth specialists such as Denison (1962, 1967) in order to derive important
implications for policy as well as to explain international differences in actual
recorded growth rates. Whilst critics have questioned the validity of the
aggregate production function approach (Hicks 1960) and others have pointed
to inconsistencies (Hall 1991) it remains none the less a useful conceptual
starting point.
Given a production function of the form
Y
t
=A
t
F(N
t
K
t
), (1)
where A is an index of overall productivity, N and K inputs of labour and
capital respectively and the subscript t denotes the time period, the growth
rate of output, ∆Y/Y can be approximated as the sum of the growth rate of
technology, ∆A/A and the growth rate of factor inputs ∆F(N,K)/F(N,K). Since
the contribution stemming from the augmentation of factor inputs is indicated
by the marginal product of the factor multiplied by the finite change in factor
employment, we obtain
∆Y/Y=∆A/A+∂F∂N·dN/Y+∂F∂K·dK/Y. (2)
Assuming factors are paid according to their marginal product so that
∂F∂N=W/P and ∂F∂K=R/P where W is the nominal wage, R the nominal
rental cost and P the general price index, we have
∆Y/Y=∆A/A+W/P·∆N/Y+R/P·∆K/Y. (3)
Rearranging gives
∆Y/Y=∆A/A+WN/PY·∆N/N+RK/PY·∆K/K. (4)
But WN/PY is simply the share of income paid to wage labour and RK/PY is the
share accruing to the owners of employed capital. Empirically, in the context of
the United States, these shares are in the order of 0.7 and 0.3 respectively.
Consequently, we obtain as a general growth formula the conclusion
∆Y/Y=∆A/A+0.7∆N/N+0.3∆K/K. (5)
618 G.K.Shaw
The initial implications are as startling as perhaps they are misleading.
Stated simply and without qualification, they suggest that a 1 per cent
increase in output growth could be achieved by either a 1 per cent increase in
productivity growth, or a 1.4 per cent increase in employment, or a 3.3 per
cent increase in the capital stock. Even more dramatically, Solow concluded
from US time series data over the period 1909 to 1949 that gross output per
man had doubled with 87.5 per cent of the increase being attributable to
technical change and the remaining 12.5 per cent stemming from the
increased use of capital.
Abramovitz (1956) and later Kendrick (1973) were to confirm these
general findings; it now became clear that at least 50 per cent of United
States growth stemmed from the increased efficiency of productive inputs
rather than the mere augmentation of those inputs. And yet there existed no
adequate theory to account for such efficiency gains. Indeed, technical
progress measures were derived as a residual after determining the
contribution obtained from augmenting factor inputs. To all intents and
purposes the determinants of technical progress lay outside the scope and
concern of mainstream economics. This was clearly an unsatisfactory and
untenable position for the economics profession and research agendas were
modified accordingly.
Finally, it will be noted, that although the Solow growth formula had
indicated a comparatively minor role to the capital input in overall growth, it
must be recalled that productivity growth may not be independent of the rate
of investment activity. That is to say that technical progress may be embodied
in the act of investment.
The Solow growth formula, regardless of its limitations, focused attention
firmly upon the role of technical progress. If the sources of technical progress
could be identified, then the implications for government policy could be
profound. Advocates of technical progress pointed to the possible role of
education and training where government policy might be expected to exert a
major impact. The role of multi-national enterprise as a vehicle for the
transmission of technical progress to the less advanced economies also took
on new meaning. The distinction between embodied and disembodied
technical progress was also brought into sharp focus.1
KENNETH ARROW AND LEARNING BY DOING MODELS
One of the first attempts to render technical progress endogenous in growth
models was the seminal paper by Arrow (1962) incorporating the concept of
learning by doing. Arrow’s approach has been generalized and extended by
Levhari (1966) and Sheshinski (1967) without departing from Arrow’s
general well-known conclusion that socially too little is invested and
produced. This derives from the spillover effects of increased knowledge
benefiting the economy in general over and above those benefits internal to
the firm.
Policy implications of endogenous growth theory 619
In this framework, the level of knowledge is itself a productive factor
which depends upon past levels of investment. Moreover, each firm learns
from the investment activity of other firms as well as from its own
investment behavior. The productivity of a given firm is thus assumed to be
an increasing function of cumulative aggregate investment for the industry.
More broadly, knowledge acquired by labour is accordingly a function of
the total capital stock; learning at any date reflects the integral of capital
output to that date.
The essence of these early learning models can be summarized quite
simply. Each firm within the economy is assumed to operate with constant
returns to scale. A doubling of labour and capital inputs with a given state of
knowledge (assumed constant by the firm) will double output. However, the
very act of increasing the capital stock through investment by the firm raises
the level of knowledge elsewhere. The economy as a whole, therefore, is
operating subject to increasing returns. This, of course, is perfectly consistent
with decreasing marginal productivity of the intangible capital good,
knowledge.
The concept of knowledge being a factor in the production function renders
increasing returns inevitable. This follows because a doubling of all tangible
factor inputs and productive processes should double output in an
environment with a constant level of knowledge. When knowledge is
permitted to vary as well in consequence of enlarging the capital input,
increasing returns follows automatically. The notion of increasing returns
being essentially external to individual firms stems from such knowledge
being public knowledge. It is this feature, that of increasing returns being
external to the firm in the tradition of Marshall (1920) and Young (1928)
which allows competitive equilibrium to exist. In doing so, it reconciles
increasing returns with the marginal productivity theory of distribution which
would otherwise imply total factor payment being in excess of total output (if
factors are paid their marginal product). In essence, equilibrium is possible
because only labour and capital are actually compensated financially—
knowledge being treated as a public good.
Although the Arrow-Levhari-Sheshinski models rendered technical progress
endogenous and explained economic growth in the context of aggregate
increasing returns being consistent with competitive equilibrium, the steady
state solution remains, growth of the economy being equated with the
autonomously determined rate of growth of the labour force. Whilst the
Arrow model had pointed to sub-optimal levels of investment, the ultimate
determinant of economic growth remained non-amenable to policy action. A
permanent increase in the share of national income devoted to investment,
whether attained through a raising of the private sector savings ratio or by
fiscal intervention, cannot influence the ultimate long run growth path. In
other words, within these models endogenous technological change is
reflected in a level effect (via an upward raising of the production function) as
opposed to a growth effect—to adopt the terminology of Lucas (1988).
620 G.K.Shaw
It is for this reason that a recent model in this tradition, linking
productivity growth directly to investment, takes on special interest. King and
Robson (1989) invoke a technical progress function which emphasizes
‘learning by watching’. This is the idea that new investment projects in one
sector of the economy have a demonstration effect or contagion effect upon
the efficiency of other sectors, permitting a raising of their output from the
employment of existing factors. The significance of the King and Robson
analysis, however, lies in its conclusion that multiple steady state growth
paths exist, even for economies which have similar initial endowments, and
that tax policy can influence the ultimate growth path attained by the
economy. Tax policy can accordingly have real and permanent effects upon
the level and growth rate of income.
THE ROMER VERSION
A striking variant on these learning by doing models was provided by the
work of Romer (1986, 1989). Again, knowledge is taken as an input in the
production function and competitive equilibrium is rendered consistent with
increasing aggregate returns owing to externalities. However, the essential
feature of Romer’s analysis is that knowledge displays increasing marginal
productivity. Traditional growth models had postulated diminishing returns;
the rate of return upon investment and the rate of growth of income per capita
was shown to be a decreasing function of the level of capital per capita. In
these models, in the absence of technological change, per capita income
attains a constant value in steady state equilibria with no per capita income
growth. Moreover, in two sector models or international trade models, rates
of return upon capital investment and wage compensation to labour will
ultimately converge over time.
In contrast, Romer’s model offers a totally different prospect. Even whilst
retaining a fully specified competitive equilibrium, per capita income can
grow without limit and the rate of return to capital may increase. It is
important to note that in Romer’s model new knowledge, the ultimate
determinant of long-run growth, is produced by investment in research
technology which exhibits diminishing returns. That is to say, a doubling of
investment in research technology will not double knowledge. Moreover, the
increase in knowledge will not be appropriated solely by the firm under-
taking the investment. The creation of new knowledge by one firm raises
production possibilities of other firms owing to the inadequacy of patent
protection. But—and here is the crucial point of departure—the production of
goods from increased knowledge demonstrates increasing returns. In other
words, knowledge displays increasing marginal productivity.
Romer demonstrates that the three key elements of his model namely
externalities, increasing returns in the production of output and decreasing
returns in the production of new knowledge are consistent with competitive
equilibrium. Moreover, the assumption of diminishing returns to research
Policy implications of endogenous growth theory 621
technology imposes an optimal (from the viewpoint of the private investor)
upper limit to the amount of knowledge creating investment activity. Thus,
endogenous technical change is explained in terms of the acquisition of
knowledge by rational profit maximizing economic agents. It follows that
they should respond to appropriate tax and fiscal incentives.
The striking implications of Romer’s analysis however, are far more
dramatic than the immediate policy implications. The assumption of
increasing marginal productivity questions the entire conclusions of
traditional growth models. There is no longer any presupposition to
converging incomes per capita in two sector or international trade models.
Less advanced economies may experience slower rates of growth than
advanced economies thus widening the gap between rich and poor countries.
Indeed, capital and investment might well flow from poor to rich economies
given the increasing marginal productivity of capital in the latter—an
analysis reminiscent of Myrdal’s (1970) thesis of cumulative causation. Small
economies would appear to be placed at a disadvantage in the growth
process. Returns to knowledge would appear to carry positive implications for
economic integration, economic unions and common markets.
The relevance of Romer’s model turns upon the empirical question of
whether increasing marginal productivity does or does not apply to the
intangible asset knowledge. Romer claims that the conclusion of the model
analysis is consistent with the stylized facts as insisted upon by Kaldor
(1961) especially the problematic wide dispersion of growth rates across
countries.
R & D MODELS
An alternative attempt to explain technological change endogenously is
provided in the approach of Uzawa (1965), Lucas (1988) and Romer (1990).
The essential feature of these models, as Stern (1991) has highlighted, lies in
identifying a sector specializing in the production of ideas. The research
sector invokes human capital together with the existing stock of knowledge to
produce new knowledge. New knowledge enhances productivity and is
available to other sectors at virtually zero marginal cost.
What these models possess in common is an emphasis upon the importance
of human capital as being the crucial determinant in the growth process.
Production of human capital is more important than the production of
physical capital. Learning by doing, or on the job training is, of course a part
of human capital formation and may be as important as formal education.
Lucas (1988) stresses the distinction between the internal effects of human
capital where the return accrues to the individual undergoing training and the
external effects which spill over into output changes. Romer (1990) has a
similar emphasis. His formal model separates the rival component of
knowledge from the non-rival technological component. Technology is a non-
rival input; its use by one firm does not preclude its use by another. Treating
622 G.K.Shaw
knowledge as a non-rival good explains knowledge spillovers and denies a
constant returns to scale production function since it is not necessary to
replicate the non-rival inputs.
Specifically, the research sector produces ideas—or improved designs for
the production of producer durables available for final goods production. In
the Romer model, knowledge enters into production in two distinct ways. A
new design (idea) allows the production of a new intermediate input. But
also, a new design increases the total stock of knowledge and accordingly
increases the productivity of human capital employed in the research sector.
The owner of a new idea has certain property rights over its use in the
production of a new producer durable but not over its use in research. To
quote Romer directly: ‘If an inventor has a patented design for widgets, no
one can make or sell widgets without the agreement of the inventor. On the
other hand, other inventors are free to spend time studying the patent
application for the widget and learn knowledge that helps in the design of a
wodget’.
The clear implication of this distinction is that the benefits of new design
inputs are partially excludable and retainable to the party initiating the new
design. This consideration points to an important difference between the
Romer and Lucas papers. In the latter, it is assumed that the production of
human capital generates a non-rival non-excludable good. As in the case of
Arrow (1962), the production of a non-rival non-excludable good is shown to
be the side effect of the production of a conventional good. In effect,
knowledge emerges as a public good whose production cannot be explained
in terms of firm investment in research and development. By introducing
partial excludability, investment in research and development is accounted for
in terms of rational profit maximizing behaviour upon the part of firms able
to enjoy quasi-rents.
All these models conclude that research produced by the research sector
will be suboptimal because of those benefits arising which are freely available
to all. This would suggest a possible role for government in subsidizing
research. The difficulty here would lie in the adequate identification of the
relevant R & D sectors. A second best option might lie in the subsidization of
the acquisition of human capital generally.
Romer’s formal model goes further in terms of its policy implications.
Since research is explained in terms of profit maximizing behaviour, and
involves making current outlays in the anticipation of future returns, then a
clear role emerges for the rate of interest. More importantly, it suggests that
countries with greater stocks of human capital will enjoy a faster rate of
economic growth. Low levels of human capital would help to explain the
comparative lack of growth in certain underdeveloped economies. Finally, it
points to the advantages to be gained from greater involvement in
international trade and economic integration.
Brief mention may be made of the paper by King and Rebelo (1990) which
like the contribution of King and Robson (1989) stresses the role of tax policy.
Policy implications of endogenous growth theory 623
Essentially King and Rebelo build upon approaches of Uzawa (1965) and
Lucas (1988). Their concern is with the observed disparity in international
growth rates and they suggest in the context of a two sector endogenous
growth model, built upon explicit microeconomic foundations, that
differences in public policy can affect the incentives to acquire capital in both
physical and human forms. In consequence taxation policy can have adverse
effects reinforced in open economies having access to international capital
markets. Comparatively small changes in tax rates can generate development
traps (zero-growth steady states) as well as growth miracles. This emphasis
upon the incentive effects of policy on investment in human capital is
reminiscent of Schultz (1961).
SOME IMPLICATIONS FOR LESS ADVANCED ECONOMIES
From the policy perspective the most important and arguably the most
urgent considerations turn upon the implications for the developing
economies. Recent growth accounting exercises have suggested that the
percentage of growth accounted for by the ‘unexplained residual’, is much
smaller for the less advanced economy. Estimates by Chenery (1983) suggest
ratios exceeding half and less than a quarter for advanced and less
advanced economies respectively. This may of course simply reflect the
greater stock of human capital in the former. Alternatively, it may reflect
other considerations generally excluded from growth theory but which
possess particular relevance for the developing economies. Stern (1991) for
example, has stressed the importance of management, organization,
infrastructure, and sectorial transfer as key elements in the growth process
of third world economies.
The implications for Third World economies is probably of greatest
moment in terms of international trade and trade policy. This is an area of
research explored in considerable detail by Grossman and Helpman (1989,
1990, 1991a, 1991b) and Helpman (1984). Potentially, it is the less advanced
economy which stands to gain the most from the freeing of international trade
since by doing so it can draw upon the stock of world knowledge. But
technological flows from rich to poor economies are by no means automatic
(see for example Lucas 1990) which raises the issue of the role of
multinational corporations and how they respond to incentives for
technological transfer.
This leads naturally into the question of policy. The essence of modern
statements of endogenous growth is that the technical progress residual is
accounted for by endogenous human capital formation. But if the latter can be
influenced by government policy world growth may be changed accordingly.
For example, if a country possessed of a comparative advantage in R & D
activity were to subsidize research, world growth would increase. Equally, a
similar subsidy introduced by an economy relatively more efficient in
manufacturing as opposed to innovating may cause world growth to decline.
624 G.K.Shaw
With regard to trade policy, clearly protectionist activity can influence
the allocation of resources into the knowledge creating sector. Trade policies
which afford protection to the manufacturing sector may promote the
transfer of skilled labour from research activity into manufacturing which
will retard innovation. Ceteris paribus, trade policy will effect a shift of
resources from research to manufacturing in policy active countries and in
the opposite direction in policy inactive countries. However, the issue is by
no means clear cut. If the profit motive governs investment in research
activities, returns to research will rise when the sector incorporating the
resulting technology is promoted by trade policy. In a three sector model
(research, manufacturing and services) research and manufacturing may
advance together when the latter is protected at the expense of contraction
within the service sector.
Other complications emerge. Different economies will possess different
ratios of skilled human capital to unskilled labour. The opening up of trade
will then be expected to change relative prices of human capital and labour.
The sudden appearance of cheap labour available to high ratio countries may
reduce incentives to produce non-rival inputs and thus slow down the growth
process for the high ratio country—the example afforded here being the
United States vis-à-vis Mexico. Again, on reasonable assumptions concerning
the required ratio of human capital per unskilled worker for R & D activity
as compared to that for the production of industrial goods, these models are
able to predict the emergence of international licensing and multinational
investment and also to account for the growth of world trade as a percentage
of world GNP over time.
Implications also emerge for the international product cycle. Traditionally,
invention and new products occur in the advanced economy where R & D
activity is well developed. Later, either by imitation or technology transfer
they will be produced in the less advanced country and ultimately production
of these goods will migrate to the low wage economy. Accordingly, trade in
manufactured products takes place on the basis of exchange between the latest
innovative goods produced only in the advanced economy and the more
traditional goods now produced predominantly by the less advanced. The
product cycle accounts for an ever evolving pattern of international trade
with the advanced economy importing the very same goods that initially it
exported.
In the context of the product cycle model, international trade always
emerges as a contributor to faster economic growth in both advanced and less
advanced economies. In the former, the migration of production from the
advanced to the less advanced economy frees resources for use in growth
enhancing product development activity. At the same time, growth occurs
faster in the less advanced economy since the resources needed for learning
and adapting the techniques imported from the advanced economy are far
fewer than those needed for autonomous new product development. In both
cases, the subsidization of learning activities (innovation in the advanced
Policy implications of endogenous growth theory 625
economy, imitation in the less advanced) may be expected to enhance long
run growth rates.
In conclusion, it would appear clear that trade policy has the potential for
influencing long run growth paths for the world economy. However, it would
be unwise to suggest that the future is one of unlimited optimism. Numerous
difficulties present themselves. The identification of growth influencing
knowledge sectors is itself a major difficulty ex ante if not ex post. Second,
the fact that conclusions deriving from the model analysis can be so easily
overturned by the alteration of the conditions or assumptions underlying the
analysis—which for the most part are unlikely to be resolved empirically—
weakens one’s confidence in growth prescription. Moreover, in the context of
international trade and the world economy, the outcome and effects of policy
measures are themselves interdependent with the policy actions of others. This
would point to the need for the coordination of national policies or at least
the consideration of second best outcomes. None the less, and despite these
caveats one might reasonably endorse the policy proposal of Lucas (1990),
namely that economic aid programmes to developing economies might well
be tied to the recipients’ willingness and openness to accept foreign
investment upon competitive terms.
NOTE
1 Disembodied technical progress is that which is able to be exploited by the existing
stock of capital employing the same kind of labour. In contrast, embodied technical
change does not benefit older machinery; rather it is embodied in the very act of new
investment.
REFERENCES
Abramovitz, M. (1956) ‘Resource and output trends in the United States since 1870’,
American Economic Review 46, Papers and Proceedings, May, pp. 5–23.
——(1986) ‘Catching up, forging ahead, and falling behind’, Journal of Economic History
46, pp. 385–406.
Arrow, K.J. (1962) ‘The economic implications of learning by doing’, Review of Economic
Studies 29, pp. 155–73.
Arthar, W.B. (1989) ‘Competing technologies: increasing returns and lock-in by historical
events’, Economic Journal 99, March, pp. 116–31.
Baumol, W.J. (1986) ‘Productivity growth, convergence and welfare: what the longrun
data show’, American Economic Review 76, pp. 1072–85.
Blitch, Charles P. (1983) ‘Allyn Young on increasing returns’, Journal of Post-Keynesian
Economics 5, pp. 359–72.
Chenery, H.B., Robinson, S. and Syrquin, M. (1986) Industrialization and Growth: A
Comparative Study, Washington DC: World Bank.
Cohen, W.M. and Levinthal, D.A. (1989) ‘Innovation and learning: the two faces of R &
D’, Economic Journal 99, pp. 569–96.
Denison, Edward F. (1961) The Sources of Economic Growth in the United States,
Committee for Economic Development, New York.
——(1962) ‘How to raise the high-employment growth rate by one percentage point’,
American Economic Review, Papers and Proceedings, May, pp. 67–75.
626 G.K.Shaw
Denison, Edward F. (Assisted by J.Poullier) (1967) Why Growth Rates Differ: Postwar
Experience in Nine Western Countries, Washington, DC: Brookings.
Domar, E. (1947) ‘Expansion and employment’, American Economic Review 37, pp.
34–55.
Dowrick, Steve and Nguyen, Duc-Tho (1989) ‘OECD comparative economic growth
1950–85: catch-up and convergence’, American Economic Review 79, pp. 1010–30.
Feder, Gershon (1983) ‘On exports and economic growth’, Journal of Development
Economics 12, pp. 59–73.
Fei, J.C.H. and Ranis, G. (1964) Development of the Labor Surplus Economy: Theory
and Policy, Homewood, IL: Irwin.
Grossman, G.M. and Helpman, E. (1989) ‘Product development and international trade’,
Journal of Political Economy 97, pp. 1261–83.
——(1990) ‘Trade innovation and growth’, American Economic Review 80, Papers and
Proceedings, pp. 86–91.
——(1991 a) ‘Endogenous product cycles’, Economic Journal 101, pp. 1214–29.
——(1991 b) Innovation and Growth: In The Global Economy, pp. 359 Cambridge,
MA: MIT Press.
Hall, Robert (1991) ‘Invariance properties of Solow’s productivity residual’, Growth/
Productivity/Unemployment: Essays in Honor of Robert Solow’s 65th Birthday (ed.
P.Diamond), Cambridge, MA: MIT Press.
Harrod, R.F. (1939) ‘An essay in dynamic theory’, Economic Journal 49, pp. 14–33.
Helpman, E. (1984) ‘A simple theory of international trade with multinational
corporations’, Journal of Political Economy 92, pp. 451–71.
Hicks, J.R. (1960) ‘Thoughts on the theory of capital: The Corfu Conference’, Oxford
Economic Papers 12, pp. 123–32.
Kaldor, N. (1957) ‘A model of economic growth’, Economic Journal 67, pp. 591–624.
——(1961) ‘Capital accumulation and growth’, In The Theory of Capital (ed. F.A. Lutz
and D.C.Hague), London: Macmillan.
Kendrick, J. (1973) Postwar Productivity Trends in the United States 1948–1969, New
York: National Bureau of Economic Research.
King, Mervyn (1989) ‘Economic growth and the life-cycle of firms’, European Economic
Review 33, pp. 325–34.
King, Mervyn and Robson, Mark (1989) Endogenous Growth and the Role of History,
London: LSE Financial Markets Group, Discussion Paper no. 63.
King, Robert G. and Rebelo, Sergio (1990) ‘Public policy and economic growth:
developing neo-classical implications’, Journal of Political Economy 98, pp. S126–
50.
Levhari, David (1966 a) ‘Further implications of “learning by doing”’, Review of
Economic Studies 33, pp. 31–9.
——(1966 b) ‘Extensions of “Arrow’s learning by doing”’, Review of Economic Studies
33, pp. 117–32.
Levin, R.C., Cohen, W.M. and Mowery, D.C. (1985) ‘R & D appropriability, opportunity
and market structure: new evidence on some Schumpeterian hypotheses’, American
Economic Review 75, pp. 20–4.
Lewis, W.A. (1954) ‘Economic development with unlimited supplies of labour’, Manchester
School 22, pp. 139–91.
Lucas, Robert E. Jr (1988) ‘On the mechanics of economic development’, Journal of
Monetary Economics 22, pp. 3–22.
——(1990) ‘Why doesn’t capital flow from rich to poor countries?’, American Economic
Review, Papers and Proceedings, vol. 80, pp. 92–6.
Marshall, A. (1920) Principles of Economics, 8th edn, London: Macmillan.
Mokyr, Joel (1991) ‘Evolutionary biology, technological change and economic history’,
Bulletin of Economic Research 43, pp. 127–49.
Myrdal, Gunnar (1970) The Challenge of World Poverty, New York: Pantheon.
Policy implications of endogenous growth theory 627
Norsworthy, J.R. (1984) ‘Growth accounting and productivity measurement’, Review
of Income and Wealth 30, pp. 309–29.
Romer, Paul M. (1986) ‘Increasing returns and long-run growth’, Journal of Political
Economy 94, pp. 1002–37.
——(1987) ‘Growth based on increasing returns due to specialization’, American
Economic Review 77, Papers and Proceedings, pp. 56–62.
——(1990) ‘Endogenous technical change’, Journal of Political Economy 98, pp. S71–
102.
——(1990) ‘Are non-convexities important for understanding growth’, American
Economic Review, Papers and Proceedings, vol. 80, pp. 97–103.
Rostow, W.W. (1960) The Stages of Economic Growth: A Non-Communist Manifesto,
Cambridge: Cambridge University Press.
Schultz, Theodore W. (1961) ‘Investment in human capital’, American Economic Review
51, pp. 1–17.
Scott, Morris (1989) A New View of Economic Growth, Oxford: Oxford University
Press.
Segerstrom, Paul S. (1991) ‘Innovation, imitation and economic growth’, Journal of
Political Economy 99, pp. 807–27.
Sheshinski, E. (1967) ‘Optimal accumulation with learning by doing’, in Essays on the
Theory of Optimal Economic Growth (ed. K.Shell), Cambridge, MA: MIT Press .
Solow, R.M. (1956) ‘A contribution to the theory of economic growth’, Quarterly Journal
of Economics 70, pp. 65–94.
——(1957) ‘Technical change and the aggregate production function’, Review of
Economics and Statistics 31, pp. 312–20.
——(1970) Growth Theory: An Exposition, Oxford: Oxford University Press.
Stern, N.H. (1989) ‘The economics of development: a survey’ Economic Journal 99, pp.
597–685.
——(1991) ‘The determinants of growth’, Economic Journal 101, pp. 122–33.
Stokey, Nancy L. (1988) ‘Learning by doing and the introduction of new goods’, Journal
of Political Economy 96, pp. 701–17.
Swan, T.W. (1956) ‘Economic growth and capital accumulation’, Economic Record 32,
pp. 334–61.
——(1964) ‘Growth models of golden ages and production functions’, in Economic
Development with Special Reference to East Asia (ed. K.Berrill), London: Macmillan.
Uzawa, H. (1965) ‘Optimum technical change in an aggregative model of economic
growth’, International Economic Review 6, pp. 18–31.
Young, Allyn A. (1928) ‘Increasing returns and economic progress’, Economic Journal
38, pp. 527–42.
26 The origins of endogenous growth
Paul M.Romer
Journal of Economic Perspectives (1994) 8, Winter, pp. 3–22
The phrase ‘endogenous growth’ embraces a diverse body of theoretical and
empirical work that emerged in the 1980s. This work distinguishes itself from
neoclassical growth by emphasizing that economic growth is an endogenous
outcome of an economic system, not the result of forces that impinge from
outside. For this reason, the theoretical work does not invoke exogenous
technological change to explain why income per capita has increased by an
order of magnitude since the industrial revolution. The empirical work does
not settle for measuring a growth accounting residual that grows at different
rates in different countries. It tries instead to uncover the private and public
sector choices that cause the rate of growth of the residual to vary across
countries. As in neoclassical growth theory, the focus in endogenous growth is
on the behavior of the economy as a whole. As a result, this work is
complementary to, but different from, the study of research and development
or productivity at the level of the industry or firm.
This chapter recounts two versions that are told of the origins of work on
endogenous growth. The first concerns what has been called the
convergence controversy. The second concerns the struggle to construct a
viable alternative to perfect competition in aggregate-level theory. These
accounts are not surveys. They are descriptions of the scholarly equivalent
to creation myths, simple stories that economists tell themselves and each
other to give meaning and structure to their current research efforts.
Understanding the differences between these two stories matters because
they teach different lessons about the relative importance of theoretical
work and empirical work in economic analysis and they suggest different
directions for future work on growth.
VERSION 1: THE CONVERGENCE CONTROVERSY
The question that has attracted the most attention in recent work on growth is
whether per capita income in different countries is converging. A crucial
stimulus to work on this question was the creation of new data sets with
information on income per capita for many countries and long periods of time
(Maddison 1982; Heston and Summers 1991).
The origins of endogenous growth 629
In his analysis of the Maddison data, William Baumol (1986) found that
poorer countries like Japan and Italy substantially closed the per capita
income gap with richer countries like the United States and Canada in the
years from 1870 to 1979. Two objections to his analysis soon became
apparent. First, in the Maddison data set, convergence takes place only in the
years since World War II. Between 1870 and 1950, income per capita tended
to diverge (Abramovitz 1986). Second, the Maddison data set included only
those economies that had successfully industrialized by the end of the sample
period. This induces a sample selection bias that apparently accounts for most
of the evidence in favor of convergence (De Long 1988).
As a result, attention then shifted to the broad sample of countries in the
Heston-Summers data set. As Figure 26.1 shows, convergence clearly fails in
this broad sample of countries. Income per capita in 1960 is plotted on the
horizontal axis. The average annual rate of growth of income per capita from
1960 to 1985 is plotted on the vertical axis.1 On average, poor countries in
this sample grow no faster than the rich countries.
Figure 26.1 poses one of the central questions in development. Why is it
that the poor countries as a group are not catching up with the rich
countries in the same way that, for example, the low income states in the
United States have been catching up with the high income states? Both
Robert Lucas (1988) and I (Romer 1986) cited the failure of cross-country
convergence to motivate models of growth that drop the two central
assumptions of the neoclassical model: that technological change is
exogenous and that the same technological opportunities are available in all
countries of the world.
To see why Figure 26.1 poses a problem for the conventional analysis,
consider a very simple version of the neoclassical model. Let output take the
Figure 26.1 Testing for convergence
630 Paul M.Romer
simple Cobb-Douglas form Y=A(t)K1-ßLß. In this expression, Y denotes net
national product, K denotes the stock of capital, L denotes the stock of labor,
and A denotes the level of technology. The notation indicating that A is a
function of time signals the standard assumption in neoclassical or exogenous
growth models: the technology improves for reasons that are outside the
model. Assume that a constant fraction of net output, s, is saved by consumers
each year. Because the model assumes a closed economy, s is also the ratio of
net investment to net national product. Because we are working with net
(rather than gross) national product and investment, sY is the rate of growth
of the capital stock. Let y=Y/L denote output per worker and let k=K/L denote
capital per worker. Let n denote the rate of growth of the labor force. Finally,
let a ‘ˆ’ over a variable denote its exponential rate of growth. Then the
behavior of the economy can be summarized by the following equation:
ŷ=(1-ß) +
=(1-ß)[sA(t)1/(1-ß)y(-ß)/(1-ß)-n]+
The first line in this equation follows by dividing total output by the stock
of labor and then calculating rates of growth. This expression specifies the
procedure from growth accounting for calculating the technology residual.
Calculate the growth in output per worker, then subtract the rate of growth
of the capital-labor ratio times the share of capital income in total income
from the rate of growth of output per worker. The second line follows by
substituting in an expression for the rate of growth of the stock of capital
per worker, as a function of the savings rate s, the growth rate of the labor
force n, the level of the technology A(t), and the level of output per
worker, y.
Outside of the steady state, the second line of the equation shows how
variation in the investment rate and in the level of output per worker should
translate into variation in the rate of growth. The key parameter is the
exponent ß on labor in the Cobb-Douglas expression for output. Under the
neoclassical assumption that the economy is characterized by perfect
competition, ß is equal to the share of total income that is paid as
compensation to labor, a number that can be calculated directly from the
national income accounts. In the sample as a whole, a reasonable benchmark
for ß is 0.6. (In industrialized economies, it tends to be somewhat larger.) This
means that in the second line of the equation, the exponent (-ß)/(1-ß) on the
level of output per worker y should be on the order of about -1.5.
We can now perform the following calculation. Pick a country like the
Philippines that had output per worker in 1960 that was equal to about 10
percent of output per worker in the United States. Because 0.1-1.5 is equal to
about 30, the equation suggests that the United States would have required a
savings rate that is about 30 times larger than the savings rate in the
Philippines for these two countries to have grown at the same rate. If we use
2/3 instead of 0.6 as the estimate of ß, the required savings rate in the United
(1)
The origins of endogenous growth 631
States would be 100 times larger than the savings rate in the Philippines. The
evidence shows that these predicted saving rates for the United States are
orders of magnitude too large.
A key assumption in this calculation is that the level of the technology A(t)
is the same in the Philippines and the United States. (The possibility that A(t)
might differ is considered below.) If they have the same technology, the only
way to explain why workers in the Philippines were only 10 percent as
productive as workers in the United States is to assume that they work with
about 0.11/(1-ß) or between 0.3 percent and 0.1 percent as much capital per
worker. Because the marginal product of capital depends on the capital stock
raised to the power -ß, the marginal product of an additional unit of capital is
0.1-ß/(1-ß) times larger in the Philippines than it is in the United States, so a
correspondingly higher rate of investment is needed in the United States to get
the same effect on output.
Figure 26.2 plots the level of per capita income against the ratio of gross
investment to gross domestic product for the Heston-Summers sample of
countries. The correlation in this figure at least has the correct sign to explain
why poor countries on average are not growing faster than the rich
countries—that is, a higher level of income is associated with a higher
investment rate. But if ß is between 0.6 and 0.7, the variation in investment
between rich and poor countries is at least an order of magnitude too small to
explain why the rich and poor countries seem to grow at about the same rate.
In concrete terms, the share of investment in the United States is not 30 or 100
times the share in the Philippines. At most, it is twice as large.
Of course, the data in Figures 26.1 and 26.2 are not exactly what the
theory calls for, but the differences are not likely to help resolve the problem
here. For example, the display equation depends on the net investment rate
Figure 26.2 Per capita income and investment
632 Paul M.Romer
instead of the gross investment rate. Because we do not have reliable data on
depreciation for this sample of countries, it is not possible to construct a net
investment ratio. A reasonable conjecture, however, is that depreciation
accounts for a larger share of GDP in rich countries than it does in poor
countries, so the difference between the net investment rate in rich and poor
countries will be even smaller than the difference between the gross
investment rates illustrated in the figure. The display equation also calls for
output per worker rather than output per capita, but for a back-of-the-
envelope calculation, variation in income per capita should be close enough
to variation in output per worker to show that a simple version of the
neoclassical model will have trouble fitting the facts.
The way to reconcile the data with the theory is to reduce ß so that labor
is relatively less important in production and diminishing returns to capital
accumulation set in more slowly. The theoretical challenge in constructing a
formal model with a smaller value for ß lies in justifying why labor is paid
more than its marginal product and capital is paid less. To explain these
divergences between private and social returns, I proposed a model in which
A was determined locally by knowledge spillovers (Romer 1987a). I followed
Arrow’s (1962) treatment of knowledge spillovers from capital investment and
assumed that each unit of capital investment not only increases the stock of
physical capital but also increases the level of the technology for all firms in
the economy through knowledge spillovers. I also assumed that an increase in
the total supply of labor causes negative spillover effects because it reduces
the incentives for firms to discover and implement labor-saving innovations
that also have positive spillover effects on production throughout the
economy.
This leads to a functional relationship between the technology in a country
and the other variables that can be written as A (K, L). Then output for firm
j can be written as , where variables with subscripts are
ones that firm j can control, and variables without subscripts represent
economy-wide totals. Because the effect that a change in a firm’s choice of K
or L has on A is an external effect that any individual firm can ignore, the
exponent α measures the private effect of an increase in employment on
output. A 1 percent increase in the labor used by a firm leads to an α percent
increase in its output. As a result, α will be equal to the fraction of output that
is paid as compensation to labor. Suppose, purely for simplicity, that the
expression linking the stock of A to K and L takes the form A(K,L)=KγL-γ for
some γ greater than zero. Then the reduced form expression for aggregate
output as a function of K and L would be Y=K1-ßLß where ß is equal to α–γ.
This exponent ß represents the aggregate effect of an increase in employment.
It captures both the private effect a and the external effect -γ. In the
calculation leading up to the equation displayed above, it is this aggregate or
social effect that matters. According to this model, ß can now be smaller than
labor’s share in national income.
The origins of endogenous growth 633
Using a simple cross-country regression based on an equation like the
display equation, I found that the effect of the investment rate on growth was
positive and the effect of initial income on growth was negative. Many other
investigators have found this kind of negative coefficient on initial income in
a growth regression. This result has received special attention, particularly in
light of the failure of overall convergence exhibited in Figure 26.1. It suggests
that convergence or regression to the mean would have taken place if all
other variables had been held constant.
After imposing the constraint implied by the equation, I estimated the
value of ß to be in the vicinity of 0.25 (Romer 1987a: Table 4). With this
value, it would take only a doubling of the investment rate—rather than a 30-
or 100-fold increase—to offset the negative effect that a ten-fold increase in
the level of output per worker would have on the rate of growth. These
figures are roughly consistent with the numbers for the United States and the
Philippines. For the sample as a whole, the small negative effect on growth
implied by higher levels of output per worker are offset by higher investment
rates in richer countries.
Robert Barro and Xavier Sala i Martin (1992) subsequently showed that
the conclusions about the size of what I am calling ß (they use different
notation) were the same whether one looked across countries or between
states in the United States. They find that a value for ß of the order of 0.2 is
required to reconcile the convergence dynamics of the states with the equation
presented above. Convergence takes place, but at a very slow rate. They also
observe that this slow rate of convergence would be even harder to explain if
one introduced capital mobility into the model.
As a possible explanation of the slow rate of convergence, Barro and Sala
i Martin (1992) propose an alternative to the neoclassical model that is
somewhat less radical than the spillover model that I proposed. As in the
endogenous growth models, they suggest that the level of the technology A(t)
can be different in different states or countries and try to model its dynamics.
They take the initial distribution of differences in A(t) as given by history and
suggest that knowledge about A diffuses slowly from high A to low A regions.
This would mean that across the states, there is underlying variation in A(t)
that causes variation in both k and y. As a result, differences in output per
worker do not necessarily signal large differences in the marginal product of
capital. In fact, free mobility of capital can be allowed in this model and the
rate of return on capital can be equalized between the different regions.
Because the flow of knowledge from the technology leader makes the
technology grow faster in the follower country, income per capita will grow
faster in the follower as diffusion closes what has been called a technology
gap.2 The speed of convergence will be determined primarily by the rate of
diffusion of knowledge, so the convergence dynamics tell us nothing about the
exponents on capital and labor.
The assumption that the level of technology can be different in different
regions is particularly attractive in the context of an analysis of the state
634 Paul M.Romer
data, because it removes the prediction of the closed-economy, identical-
technology neoclassical model that the marginal productivity of capital can
be many times larger in poorer regions than in rich regions.3 According to
the data reported by Barro and Sala i Martin (1992), in 1880, income per
capita in states such as North Carolina, South Carolina, Virginia, and
Georgia was about one-third of income per capita in industrial states such
as New York, Massachusetts, and Rhode Island. If ß is equal to 0.6, -ß/(1-ß)
is equal to –1.5 and -1.5 is equal to about 5. This means that the marginal
product of capital should have been about five times higher in the South
than it was in New England. It is difficult to imagine barriers to flows of
capital between the states that could have kept these differences from
rapidly being arbitraged away. In particular, it would be difficult to
understand why any capital investment at all took place in New England
after 1880. But if there were important differences in the technology in use
in the two regions, the South may not have offered higher returns to capital
investment.
In a third approach to the analysis of cross-country data, Greg Mankiw,
David Romer, and David Weil (1992) took the most conservative path,
showing that it is possible to justify a low value for ß even in a pure version
of the closed economy, neoclassical model which assumes that the level of
technology is the same in each country in the world. The only change they
make is to extend the usual two-factor neoclassical model by allowing for
human capital H as well as physical capital K. They use the fraction of the
working age population that attends secondary school as a measure of the
rate of investment in human capital that is analogous to the share of physical
capital investment in total GDP.
They conclude from their cross-country growth regressions that
is a reasonable specification for aggregate output. In this
model, the exponent ß on the fixed factor of production L has been
reduced from 0.6 to 0.33. This lower value of ß is consistent with the data
on income shares because total wage payments consist of payments to both
human capital and unskilled labor. If K and H vary together across
countries, this specification implies that it takes about a three-fold increase
in investment (an increase by the factor 0.1-0.5 to be precise) to offset a ten-
fold increase in output per worker in a comparison across nations. Once
one takes account of variation in investment in schooling as well as in
investment in physical capital, a factor of three is roughly consistent with
the variation in total investment rates observed in the Summer-Heston
sample of countries.
Although Mankiw et al. (1992) do not examine the state data, it is clear
what their style of explanation would suggest. They would assume that the
same technology was available in the North and the South. Suppose that
Northern states had levels of both human capital and physical capital that
were higher than those in the Southern states in the same ratio. A value of ß
equal to 1/3, together with the fact that output per capita was about one-third
The origins of endogenous growth 635
as large in the South in 1880, would imply that rate of return on physical
capital and the wage for human capital were both about -0.5 (or about 1.7)
times higher in the Southern states than they were in the New England states.
Compared to the factor of 5 implied by the model without human capital,
these parameters would imply much smaller incentives to shift all capital
investment to the South. (They would imply, however, that human capital
would tend to migrate from the North to the South.)
The implication from this work is that if you are committed to the
neoclassical model, the kind of data exhibited in Figures 26.1 and 26.2
cannot be used to make you recant. They do not compel you to give up the
convenience of a model in which markets are perfect. They cannot force you
to address the complicated issues that arise in the economic analysis of the
production and diffusion of technology, knowledge, and information.
AN EVALUATION OF THE CONVERGENCE CONTROVERSY
The version of the development of endogenous growth theory outlined above
skips lots of detail and smooths over many complications that made this seem
like a real controversy at the time. In retrospect, what is striking is how little
disagreement there is about the basic facts. Everyone agrees that a
conventional neoclassical model with an exponent of about one-third on
capital and about two-thirds on labor cannot fit the cross-country or cross-
state data. Everyone agrees that the marginal product of investment cannot be
orders of magnitudes smaller in rich countries than in poor countries. The
differences between the different researchers concern the inferences about
models that we should draw from these facts. As is usually the case in
macroeconomics, many different inferences are consistent with the same
regression statistics.
This history has many elements in common with other stories about the
development of economics. The story starts with the emergence of new data.
These present anomalies that lead to new theoretical models, some of which
differ markedly from previous, well-accepted models. Then a more
conservative interpretation emerges that accommodates the new evidence and
preserves much of the structure of the old body of theory. In the end, we have
refined the set of alternatives somewhat, but seem to be left in about the same
position where we started, with too many theories that are consistent with the
same small number of facts.
But economists who accept this interpretation come to the conclusion that
we do not have enough data only because they restrict attention to the kind of
statistical evidence illustrated in Figures 26.1 and 26.2. They fail to take
account of all the other kinds of evidence that are available. My original
work on growth (Romer 1983, 1986) was motivated primarily by the
observation that in the broad sweep of history, classical economists like
Malthus and Ricardo came to conclusions that were completely wrong about
prospects for growth. Over time, growth rates have been increasing, not
636 Paul M.Romer
decreasing.4 Lucas (1988) emphasized the fact that international patterns of
migration and wage differentials are very difficult to reconcile with a
neoclassical model. If the same technology were available in all countries,
human capital would not move from places where it is scarce to places where
it is abundant and the same worker would not earn a higher wage after
moving from the Philippines to the United States.
The main message of this chapter is that the convergence controversy
captures only part of what endogenous growth has been all about. It may
encompass a large fraction of the recently published papers, but it
nevertheless represents a digression from the main story behind endogenous
growth theory. The story told about the convergence controversy also tends to
reinforce a message that I think is seriously misleading—that data are the
only scarce resource in economic analysis.
VERSION 2: THE PASSING OF PERFECT COMPETITION
The second version of the origins of endogenous growth starts from the
observation that we had enough evidence to reject all the available growth
models throughout the 1950s, 1960s and 1970s. What we lacked were
good aggregate-level models. This version of the origins of endogenous
growth is therefore concerned with the painfully slow progress we have
made in constructing formal economic models at the aggregate level. It
suggests that progress in economics does not come merely from the
mechanical application of hypothesis tests to data sets. There is a creative
act associated with the construction of new models that is also crucial to
the process.
The evidence about growth that economists have long taken for granted
and that poses a challenge for growth theorists can be distilled to five basic
facts.
Fact 1: There are many firms in a market economy
The fact is so obvious that we often do not bother to state it, but it clearly
will not do to have a model in which there are overwhelming forces that
tend to concentrate all output in the hands of a single, economy-wide
monopolist.
Fact 2: Discoveries differ from other inputs in the sense that many people can
use them at the same time
The idea behind the transistor, the principles behind internal combustion,
the organizational structure of a modern corporation, the concepts of
double entry bookkeeping—all these pieces of information and many more
like them have the property that it is technologically possible for
everybody and every firm to make use of them at the same time. In the
The origins of endogenous growth 637
language of public finance, ordinary goods are rival goods, but
information is nonrival.
Fact 3: It is possible to replicate physical activities
Replication implies that the aggregate production function representing a
competitive market should be characterized by homogeneity of degree one in
all of its conventional (that is, rival) inputs. If we represent output in the form
Y=AF(K, H, L), then doubling all three of K, H, and L should allow a
doubling of output. There is no need to double the nonrival inputs represented
by A because the existing pieces of information can be used in both instances
of the productive activity at the same time. (The assumption that the market
is competitive means that the existing activity already operates at the
minimum efficient scale, so there are no economies of scale from building a
single plant that is twice as large as the existing one.)
If farming were the relevant activity instead of manufacturing, we would
clearly need to include land as an input in production, and in the economy
as a whole, it is not possible to double the stock of land. This does not
change the fundamental implication of the replication argument. If
aggregate output is homogeneous of degree 1 in the rival inputs and firms
are price takers, Euler’s theorem implies that the compensation paid to the
rival inputs must exactly equal the value of output produced. This fact is
part of what makes the neoclassical model so simple and makes growth
accounting work. The only problem is that this leaves nothing to
compensate any inputs that were used to produce the discoveries that lead to
increases in A.
Fact 4: Technological advance comes from things that people do
No economist, so far as I know, has ever been willing to make a serious
defense of the proposition that technological change is literally a function of
elapsed calendar time. Being explicit about the issues here is important
nevertheless, because it can help untangle a link that is sometimes made
between exogeneity and randomness. If I am prospecting for gold or looking
for a change in the DNA of a bacterium that will let it eat the oil from an oil
spill, success for me will be dominated by chance. Discovery will seem to be
an exogenous event in the sense that forces outside of my control seem to
determine whether I succeed. But the aggregate rate of discovery is
endogenous. When more people start prospecting for gold or experimenting
with bacteria, more valuable discoveries will be found. This will be true even
if discoveries are accidental side effects of some other activity (finding gold as
a side effect of ditch-digging) or if market incentives play no role in
encouraging the activity (as when discoveries about basic molecular biology
were induced by government research grants). The aggregate rate of
discovery is still determined by things that people do.
638 Paul M.Romer
Fact 5: Many individuals and firms have market power and earn monopoly
rents on discoveries
Even though the information from discoveries is nonrival (as noted in fact
2), economically important discoveries usually do not meet the other
criterion for a public good; they typically are partially excludable, or
excludable for at least some period of time. Because people and firms have
some control over the information produced by most discoveries, it cannot
be treated as a pure public good. This information is not like a short-wave
radio broadcast that everyone can access without the permission of the
sender. But if a firm can control access to a discovery, it can charge a price
that is higher than zero. It therefore earns monopoly profits because
information has no opportunity cost.
The neoclassical model that was developed and applied by Robert Solow
(1956, 1957) and others constituted a giant first step forward in the process
of constructing a formal model of growth. The discussion of the
convergence controversy, framed as it was almost entirely in terms of the
neoclassical model, illustrates the model’s power and durability. Like any
model, the neoclassical model is a compromise between what we would like
from a model and what is feasible given the state of our modeling skills.
The neoclassical model captured facts 1, 2, and 3, but postponed
consideration of facts 4 and 5. From a theoretical point of view, a key
advantage of the model is its treatment of technology as a pure public good.
This makes it possible to accommodate fact 2—that knowledge is a nonrival
good—in a model that retains the simplicity of perfect competition. The
public good assumption also implies that knowledge is nonexcludable, and
this is clearly inconsistent with the evidence summarized in fact 5—that
individuals and firms earn profits from their discoveries. This assumption
was useful, nevertheless, as part of an interim modeling strategy that was
adopted until models with nonrivalry and excludability could be
formulated.
Endogenous growth models try to take the next step and accommodate fact
4. Work in this direction started in the 1960s. For example, Karl Shell (1966)
made the point about replication noted above, showing that it left no
resources to pay for increases in A. He proposed a model in which A is
financed from tax revenue collected by the government. Endogenous growth
models have tended to follow Arrow (1962) and emphasize the private sector
activities that contribute to technological advance rather than public sector
funding for research. A subset of these models has tried to incorporate both
fact 4 (that technological advance comes from things people do) and fact 5
(the existence of monopoly rents). These are sometimes referred to as neo-
Schumpeterian models because of Schumpeter’s emphasis of the importance of
temporary monopoly power as a motivating force in the innovative process.5
In addition, there are two other distinct kinds of endogenous growth models.
The origins of endogenous growth 639
Spillover models have already been mentioned. Linear models will be
described below.6
With the benefit of hindsight, it is obvious that growth theorists would
eventually have to do what economists working at the industry and firm
level have done: abandon the assumption of price-taking competition.
Otherwise, there is no hope of capturing fact 5. Even at the time, the point
received at least some attention. In his 1956 paper, Solow remarked in a
footnote on the desirability of extending the model to allow for
monopolistic competition. One of his students, William Nordhaus (1969),
subsequently outlined a growth model that did have patents, monopoly
power, and many firms. For technical reasons, this model still invoked
exogenous technological change, so it is not strictly speaking a model of
endogenous growth—but it could have been extended to become one.
Because a general formal treatment of monopolistic competition was not
available at the time, little progress in this direction took place for the
next 20 years.
Even though it is obvious in retrospect that endogenous growth theory
would have to introduce imperfect competition, this was not the direction that
the first models of the 1980s pursued. Both my model (Romer 1986) and
Robert Lucas’s model (1988) included fact 4 without taking the final step and
including step 5. In both of these models, the technology is endogenously
provided as a side effect of private investment decisions. From the point of
view of the users of technology, it is still treated as a pure public good, just as
it is in the neoclassical model. As a result, firms can be treated as price takers
and an equilibrium with many firms can exist.
This technique for introducing a form of aggregate increasing returns into
a model with many firms was first proposed by Alfred Marshall (1890). To
overturn the pessimistic predictions of Malthus and Ricardo, he wanted to
introduce some form of aggregate increasing returns. To derive his downward
sloping supply curve from an industry with many firms, Marshall introduced
the new notion of increasing returns that were external to any individual firm.
External effects therefore entered into economics to preserve the analytical
machinery of supply and demand curves and price taking in the presence of
increasing returns. The analysis of other kinds of external effects—smoke,
bees, and so on—came later.7
As noted in the previous discussion of spillover models, Arrow (1962)
constructed a model along these lines. In a simplified form, output for firm
j in his model can be written as Yj=A(K)F(Kj, Lj), where (as before) K
without a subscript denotes the aggregate stock of capital. For technical
reasons, Arrow, like Nordhaus, did not emphasize the fact that his model
could lead to sustained, endogenous growth. For the parameter values that
he studies, if the size of the population is held constant, growth eventually
comes to a halt.
Lucas’s model has a very similar underlying structure. There, it is
investments in human capital rather than physical capital that have spillover
640 Paul M.Romer
effects that increase the level of the technology. It is as if output for firm j
takes the form Yj=A(H)F(Kj,Hj). Both of these models accommodated facts 1–
4 but not fact 5.8
In my first paper on growth (Romer 1986), I assumed in effect that
aggregate output could be written as Y=A(R)F(Rj, Kj, Lj) where Rj stands for
the stock of results from expenditure on research and development by firm
j.9 I assumed that it is spillovers from private research efforts that lead to
improvements in the public stock of knowledge A. This seemed appealing
because it recognized that firms did research and development on purpose
and that the relevant spillovers or incomplete property rights were
associated with the results from research and development. (In the
microeconomic analysis of research and development at the industry level,
Zvi Griliches (1979) used this same kind of formulation.) But to make this
model fit within the framework of price-taking with no monopoly power, I
assumed that the function F was homogeneous of degree one in all of its
inputs, including R. This, unfortunately, violates fact 2, that research is a
nonrival good and fact 3, that only rival goods need to be replicated to
double output. If I had admitted that Rj was nonrival, the replication
argument would have implied that the firm faced increasing returns in the
inputs Rj, Kj, and Lj that it controlled, because output would double merely
by replicating Kj and Lj.
My sleight of hand in treating Rj as a rival good and making F
homogeneous of degree 1 in all three of K, L, and R may seem like a
trifling matter in an area of theory that depends on so many other short
cuts. After all, if one is going to do violence to the complexity of economic
activity by assuming that there is an aggregate production function, how
much more harm can it do to be sloppy about the difference between rival
and nonrival goods? Unfortunately, quite a bit. The distinctions between
rival and nonrival inputs, and the distinction between excludable and
nonexcludable goods, are of absolutely fundamental importance in
modeling and in policy formulation.
For years, the economic analysis of science and technology policy
consisted of little more than a syllogism. The major premise was that the
government should provide public goods and the private sector should
provide private goods. The minor premise was that basic research is a
public good and applied research is a private good. Once you think
carefully about nonrivalry and excludability, it is clear that the major
premise is misleading because it understates the possible role for collective
action. Governments can usefully provide goods that are nonrival but are
not true public goods, because they are potentially excludable. The minor
premise is simply wrong. Applied research is not an ordinary private
good. Discussion in policy circles is now taking place using new terms—
critical technologies, generic research, and pre-competitive research—that
are only vaguely defined but that take the discussion outside of the simple
dichotomy between public goods and private goods. This is probably
The origins of endogenous growth 641
useful, but it would lend needed structure to this discussion if participants
paid more attention to the distinction between the two different aspects of
publicness (nonrivalry and nonexcludability) and looked more formally at
t h e d i f f e r e n t k i n d s o f p o l i c y c h a l l e n g e s t h a t n o n r i v a l r y a n d
nonexcludability present.
The linear model branch of endogenous growth theory pursued even more
aggressively the strategy I used.10 If I could treat the part of knowledge that
firms control as an ordinary input in production—that is, as an input that is
rival and hence is not associated with increasing returns—why bother to
allow for any nonrival inputs at all? In effect, these models assumed that
output could be written as Y=F(R, K, H) for a homogenous of degree 1
production function F. These models assumed that research R, physical capital
K, and human capital H were like ordinary inputs. If there are no nonrival
goods, there are no increasing returns. It is then a relatively simple matter to
build a perfectly competitive model of growth. To simplify still further, these
models often aggregate R, K, and H into a single broad measure of capital.
Suppose we call it X. Then we can write F(X) as a linear function:
Y=F(X)=aX, hence the name linear models. If we assume that a constant
fraction of output Y is saved and used to produce more X, the model
generates persistent, endogenous growth. Relative to the neoclassical model,
these models capture fact 4—that technological change comes from
investments that people make—at the cost of abandoning fact 2, that
technology or knowledge is a nonrival good.
Proponents of the linear model and the neoclassical model have sometimes
been drawn into pointless arguments about which model is worse. Proponents
of the linear growth models point out that the neoclassical model fails to
capture fact 4. Proponents of the neoclassical model observe that the linear
model cannot capture fact 2. This dispute is partly an outgrowth of the
convergence controversy. Both sides specify that output takes the form Y=K1-
ßLß and then argue about whether ß is bigger than zero (as the proponents of
the neoclassical model claim) or close to zero (as some versions of the linear
growth model suggest).
This is not a very useful debate. There are circumstances in which each
model can be a useful expositional device for highlighting different aspects of
the growth process, but presumably the agenda for the profession ought to be
to capture both facts 2 and 4 and pick up fact 5 to boot.
NEO-SCHUMPETERIAN GROWTH
Two steps were required for the neo-Schumpeterian models of growth to
emerge. The first was that after struggling for years to preserve perfect
competition, or at least price-taking in the presence of external effects, growth
theorists had to decide to let go. It helped that economists working on
industrial organization had given them something else to hang onto. By the
late 1970s, there were aggregate models with many firms (fact 1), each of
642 Paul M.Romer
which could have market power (fact 5). The most convenient such model
was developed by Dixit and Stiglitz (1977). Ethier (1982) subsequently
showed how their model of preferences over many goods could be interpreted
as a production function that depended on a large number of inputs in
production.
Once people who were interested in growth recognized that this approach
offered the alternative to a competitive market structure, there was only one
technical detail that remained to be resolved, the detail that had kept both
Nordhaus and Arrow from producing models of endogenous growth. All
models of growth need at least one equation which describes the evolution of
something like A (t).11 This equation usually takes the form
.
A=-Aφ, (2)
where A with a dot denotes the time derivative of A. Models that produce
steady state growth fill in the blank with a constant and set the exponent φ
equal to 1. For example, if we set φ equal to 1 and insert a constant g in the
blank, we have the driving equation behind the neoclassical model with
exogenous technological change.
Mathematically, this kind of formulation is not robust. If φ turns out to be
even slightly greater than 1, the equation implies that the stock of technology
will go to infinity in finite time. When we use this same kind of model to
study population growth, this lack of robustness does not raise any particular
difficulties. We understand that functional forms are always approximations,
and that a linear differential equation leading to exponential growth is a
particularly convenient approximation. But Nordhaus and Arrow both
worked at a time when there was real concern about the knife-edge character
of the assumptions about φ.12 If it was less than one, growth eventually
stopped. If it was even slightly greater than one, everything blows up. As a
result, economists stayed well away from the edge and assumed that φ had to
be strictly less than 1. In a model like Nordhaus’s, growth can be kept going
only by adding a second kind of knowledge A2 that grows exogenously.
(Formally, bringing in exogenous technological change amounts to bringing
in a new equation in which the exponent corresponding to φ has already been
set to 1, and it only takes one equation with this property to keep things
going.)
I devoted a great deal of attention to this robustness problem in my
analysis of the spillover models. I modified other functional forms elsewhere
in the model to construct robust models of endogenous growth in which the
level of output and its rate of growth stayed finite for all time for a range of
values of φ that were strictly bigger than 1 (Romer 1983, 1986). For values
slightly less than 1, growth eventually stopped but could persist, nevertheless,
for a very long time. The mathematical analysis in this more complicated
robust model was much harder than the analysis that is possible when φ is
equal to 1. The difference between the two models is the difference between
studying the phase plane of a nonlinear differential equation system and
The origins of endogenous growth 643
solving a simple linear differential equation. Once it is clear that we could
build a complicated model that is robust, there is every reason to work with
the simple special case whenever possible.
By the late 1980s, economists like Judd (1985) and Grossman and Helpman
(1989) were working out models of growth with monopolistic competition. Like
Nordhaus and Arrow, they stayed well away from the case where φ was equal
to 1. Judd invoked exogenous technological change to keep his economy
growing. Grossman and Helpman were investigating the connection between
trade and growth, and settled for an analysis of transitional dynamics of the
model as it converged to a steady state level of income where growth stopped.
In each model, monopoly profits motivate discovery.
I took what I had learned about generating sustained growth from my
analysis of spillover models and applied it to the monopolistic competition
model. I constructed two very simple models of sustained growth that
accommodated all five of the facts cited above. One of these did not invoke
any spillover effects at all (Romer 1987b). The other combined both
monopoly power and spillovers—that is, incomplete intellectual property
rights (Romer 1990). In each of these models I set the analog of φ equal to 1.
I knew that by repeating my analysis of the spillover model, it would be
possible to construct more complicated robust models with the same
qualitative implications.
Research on endogenous growth models in which monopoly profits
motivate innovation has progressed rapidly since then and has uncovered a
number of unexpected connections between market size, international trade,
and growth, as Grossman and Helpman (1994) explain.
CONCLUSIONS
The economics profession is undergoing a substantial change in how we think
about international trade, development, economic growth and economic
geography.13 In each of these areas, we have gone through a progression that
starts with models based on perfect competition, moves to price-taking with
external increasing returns, and finishes with explicit models of imperfect
competition. It is likely that this pattern will repeat itself in other areas like
the theory of macroeconomic fluctuations.
The effects of this general trend may be far-reaching. Ultimately, it may
force economists to reconsider some of the most basic propositions in
economics. For example, I am convinced that both markets and free trade are
good, but the traditional answer that we give to students to explain why they
are good, the one based on perfect competition and Pareto optimality, is
becoming untenable. Something more interesting and more complicated is
going on here.14
In each of the areas where our understanding has changed, evidence that
challenged the models of perfect competition and supported the models with
imperfect competition had been apparent all along. Everyone knew that there
644 Paul M.Romer
was lots of intra-industry trade between developed nations and little trade
between the North and the South. Everyone knew that some developing
countries grew spectacularly while others languished. Everyone knew that
people do the things that lead to technological change. Everyone knew that
the number of locally available goods was limited by the extent of the market
in the city where someone lives and works.
In evaluating different models of growth, I have found that Lucas’s (1988)
observation, that people with human capital migrate from places where it is
scarce to place where it is abundant, is as powerful a piece of evidence as
all the cross-country growth regressions combined. But this kind of fact, like
the fact about intra-industry trade or the fact that people make discoveries,
does not come with an attached t-statistic. As a result, these kinds of facts
tend to be neglected in discussions that focus too narrowly on testing and
rejecting models.
Economists often complain that we do not have enough data to
differentiate between the available theories, but what constitutes relevant
data is itself endogenous. If we set our standards for what constitutes
relevant evidence too high and pose our tests too narrowly, we will indeed
end up with too little data. We can thereby enshrine the economic
orthodoxy and make it invulnerable to challenge.15 If we do not have any
models that can fit the data, the temptation will be to set very high
standards for admissible evidence, because we would prefer not to reject the
only models that we have.
When I look back on my work on growth, my greatest satisfaction comes
from having rejected the first round of external effects models that I tried. I
am glad that I was able to learn something about robustness and nonrivalry
from struggling with these models, but was still able to let go when a better
alternative became apparent. My greatest regret is the shift I made while
working on these external effects models, a shift that took me away from the
emphasis on research and knowledge that characterized my 1986 paper and
toward the emphasis on physical capital that characterized the empirical
work in the paper cited in the discussion of convergence (1987a). This paper
contributed to the convergence controversy and to an emphasis on the
exponents on capital and labor in aggregate production. I am now critical of
this work, and I accept part of the blame. Looking back, I suspect that I made
this shift toward capital and away from knowledge partly in an attempt to
conform to the norms of what constituted convincing empirical work in
macroeconomics. No international agency publishes data series on the local
production of knowledge and inward flows of knowledge. If you want to run
regressions, investment in physical capital is a variable that you can use, so
use it I did. I wish I had stuck to my guns about the importance of evidence
like that contained in facts 1 through 5.
If macroeconomists look only at the cross-country regressions deployed in
the convergence controversy, it will be easy to be satisfied with neoclassical
models in which market incentives and government policies have no effect on
The origins of endogenous growth 645
discovery, diffusion, and technological advance. But if we make use of all of
the available evidence, economists can move beyond these models and begin
once again to make progress toward a complete understanding of the
determinants of long-run economic success. Ultimately, this will put us in
position to offer policy-makers something more insightful than the standard
neoclassical prescription—more saving and more schooling. We will be able
to rejoin the ongoing policy debates about tax subsidies for private research,
antitrust exemptions for research joint ventures, the activities of multinational
firms, the effects of government procurement, the feedback between trade
policy and innovation, the scope of protection for intellectual property rights,
the links between private firms and universities, the mechanisms for selecting
the research areas that receive public support, and the costs and benefits of an
explicit government-led technology policy. We will be able to address the
most important policy questions about growth: In a developing country like
the Philippines, what are the best institutional arrangements for gaining
access to the knowledge that already exists in the rest of the world? In a
country like the United States, what are the best institutional arrangements for
encouraging the production and use of new knowledge?
ACKNOWLEDGEMENTS
I have benefited from comments by Jeffrey Frankel, Alan Krueger, David
Romer, Carl Shapiro, and Timothy Taylor on early drafts of this chapter.
This work was supported by NSF Grant SES 9023469 and by the Canadian
Institute for Advanced Research.
NOTES
1 The data here are taken from version IV of the Penn World Table. The income
measure is RGDP2. See Summers and Heston (1988) for details.
2 Nelson and Phelps (1966) give a theoretical model that allows for diffusion of the
technology between countries. Fagerberg (1987) interprets cross-country growth
regressions in the context of a technology gap model instead of a neoclassical model
or a spillover model. For further discussion of diffusion, see also Barro and Sala i
Martin (1994) and Jovanovic and Lach (1993).
3 See King and Rebelo (1993) for a fuller discussion of both the price and quantity
implications of the neoclassical model.
4 See Kremer (1993) for a stimulating look at this question from a very long-run
point of view.
5 Of course, Stigler’s law applies in this case: the person that any result is named after
was not the first person to derive or state the result. It just helps to have a label so
that you can keep track of the players without a scorecard.
6 Nelson and Winter (1982) developed an alternative evolutionary model of growth.
Their verbal, descriptive style of theory, which they label appreciative theory, was
flexible enough to accommodate facts 1–5. This style of work can be thought of as
a complement to formal theory, not a substitute for it. It leaves open the problem of
constructing a formal theory that could accommodate these facts.
646 Paul M.Romer
7 For an explicit treatment showing that Marshallian external increasing returns is
ultimately an untenable way to model any process involving learning or knowledge,
see Dasgupta and Stiglitz (1988).
8 Lucas actually makes A depend on per capita H rather than total H. The difference
between these two formulations is not relevant for the discussion here, but is important
for some of the other implications of the model.
9 For consistency with the rest of the discussion, I distinguish here between R and K.
In the paper, I actually dropped physical capital from consideration so that I have
only one state variable to deal with. This leads to a potential confusion because I also
used the symbol K for knowledge instead of R.
10 One of the early linear models was Uzawa (1965). Important papers in this line of
work include Becker et al. (1990), Jones and Manuelli (1990) and Rebelo (1991).
11 Sometimes other variables like H or K are used in place of A, but the basic issues are
the same.
12 See Stiglitz (1990) for a discussion of how people working on growth at the time
perceived this problem.
13 Paul Krugman has made influential contributions in all of these areas. See Krugman
(1990, 1991, 1993) for a discussion of the changes in these fields.
14 Romer (1994) offers a demonstration that, for example, the costs of trade restrictions
in a developing country can be far greater in the context of a model with imperfect
competition than they are in a model with perfect competition.
15 In their discussion of real business cycle theories and the kind of evidence used to
test them, Greg Mankiw (1989) and Robert Solow (1988) have both made a similar
point about explicit statistical versus broader kinds of evidence.
REFERENCES
Abramovitz, Moses, ‘Catching Up, Forging Ahead, and Falling Behind’, Journal of
Economic History June 1986, 46:2, 385–406.
Arrow, Kenneth J., ‘The Economic Implications of Learning by Doing’, Review of
Economic Studies June 1962, 29, 155–73.
Barro, Robert J. and Xavier Sala i Martin, ‘Convergence’, Journal of Political Economy
April 1992, 100:2, 223–51.
Barro, Robert J. and Xavier Sala i Martin, ‘Chapter 8: Diffusion of Technology’. In
R.J.Barro and X.Sala-i-Martin (eds) Economic Growth, New York: McGraw Hill,
1994.
Baumol, William J., ‘Productivity Growth, Convergence, and Welfare: What the Long-
run Data Show’, American Economic Review December 1986, 76:5, 1072–85.
Becker, G., K.Murphy and R.Tamura, ‘Economic Growth, Human Capital, and Population
Growth’, Journal of Political Economy October 1990, 98:5 Part 2, S12– S137.
Dasgupta, P. and J.Stiglitz, ‘Learning-by-Doing, Market Structure, and Industrial and
Trade Policies’, Oxford Economic Papers June 1988, 40:2, 246–68.
De Long, J.Bradford, ‘Productivity Growth, Convergence and Welfare: Comment’,
American Economic Review December 1988, 75:5, 1138–54.
Dixit, A. and J.Stiglitz, ‘Monopolistic Competition and Optimum Product Diversity’,
American Economic Review June 1977, 67:3, 297–308.
Ethier, W.J., ‘National and International Returns to Scale in the Modern Theory of
International Trade’, American Economic Review June 1982, 72:3, 389–405.
Fagerberg, Jan, ‘A Technology Gap Approach to Why Growth Rates Differ’, Research
Policy 1987, 16, 87–99.
The origins of endogenous growth 647
Griliches, Zvi, ‘Issues in Assessing the Contribution of Research and Development to
Productivity Growth’, Bell Journal of Economics Spring 1979, 10, 92–116.
Grossman, Gene, and Elhanan Helpman, ‘Product Development and International Trade’,
Journal of Political Economy December 1989, 97:6, 1261–83.
Grossman, Gene, and Elhanan Helpman, ‘Endogenous Innovation in the Theory of
Growth’, Journal of Economic Perspectives Winter 1994, 8:1, 23–44.
Heston, Alan, and Robert Summers, ‘The Penn World Trade (Mark 5): An Expanded Set
of International Comparisons, 1950–1988’, Quarterly Journal of Economics May
1991, 106, 327–68.
Jones, Lawrence, and Rodolfo Manuelli, ‘A Convex Model of Equilibrium Growth:
Theory and Policy Implications’, Journal of Political Economy October 1990, 98:5
Part 1, 1008–38.
Jovanovic, Boyan, and Saul Lach, ‘Diffusion Lags and Aggregate Fluctuations’, mimeo,
New York University, August 1993.
Judd, K.L., ‘On the Performance of Patents’, Econometrica May 1985, 53:3, 567–85.
King, Robert G. and Sergio Rebelo, ‘Transitional Dynamics and Economic Growth in the
Neoclassical Model’, American Economic Review September 1993, 83:4, 908–31.
Kremer, Michael, ‘Population Growth and Technological Change: One Million B.C. to
1990’, Quarterly Journal of Economics August 1993, 108:3, 681–716.
Krugman, Paul, Rethinking International Trade, Cambridge: MIT Press, 1990.
Krugman, Paul, Geography and Trade, Cambridge: MIT Press, 1991.
Krugman, Paul, ‘Towards a Counter-Counter Revolution in Development Theory’,
Proceedings of the Annual World Bank Conference on Development 1992,
Supplement, Washington, DC, World Bank Economic Review 1993, 15–38.
Lucas, Robert E., Jr, ‘On the Mechanics of Economic Development’, Journal of Monetary
Economics July 1988, 22:1, 3–42.
Maddison, A., Phases of Capitalist Development, Oxford: Oxford University Press,
1982.
Mankiw, N.Gregory, ‘Real Business Cycles: A New Keynesian Perspective’, Journal of
Economic Perspectives Summer 1989, 3:3, 79–90.
Mankiw, N.Gregory, David Romer, and David N.Weil, ‘A Contribution to the Empirics
of Economic Growth’, Quarterly Journal of Economics May 1992, 107, 407–37.
Marshall, Alfred, Principles of Economics, London: Macmillan, 1890.
Nelson, Richard R. and Edmund S.Phelps, ‘Investment in Humans, Technological
Diffusion, and Economic Growth’, American Economic Review May 1966, 56, 69–
75.
Nelson, Richard R. and Sidney G.Winter, An Evolutionary Theory of Economic Change,
Cambridge: Belnap Press of Harvard University Press, 1982.
Nordhaus, William D., ‘An Economic Theory of Technological Change’, American
Economic Review May 1969, 59:2, 18–28.
Rebelo, Sergio, ‘Long Run Policy Analysis and Long Run Growth’, Journal of Political
Economy June 1991, 99:3, 500–21.
Romer, Paul M., ‘Dynamic Competitive Equilibria with Externalities, Increasing Returns
and Unbounded Growth’, Ph.D. dissertation, University of Chicago, 1983.
Romer, Paul M., ‘Increasing Returns and Long-Run Growth’, Journal of Political Economy
October 1986, 94:5, 1002–37.
Romer, Paul M., ‘Crazy Explanations for the Productivity Slowdown’, in S.Fischer (ed.)
NBER Macroeconomics Annual, Cambridge: MIT Press, 1987a, 163–202.
Romer, Paul M., ‘Growth Based on Increasing Returns Due to Specialization’, American
Economic Review May 1987b, 77:2, 56–62.
Romer, Paul M., ‘Endogenous Technological Change’, Journal of Political Economy
1990, 98, S71–102.
648 Paul M.Romer
Romer, Paul M., ‘Two Strategies for Economic Development: Using Ideas and Producing
Ideas’, Proceedings of the Annual World Bank Conference on Development 1992,
Supplement, Washington, DC, World Bank Economic Review, 1993.
Romer, Paul M., ‘New Goods, Old Theory and the Welfare Costs of Trade Restrictions’,
Journal of Development Economics February 1994, 43:1.
Shell, Karl, ‘Toward a Theory of Inventive Activity and Capital Accumulation’, American
Economic Review May 1966, 56, 62–68.
Solow, Robert, ‘A Contribution to the Theory of Economic Growth’, Quarterly Journal
of Economics February 1956, 70, 65–94.
Solow, Robert, ‘Technical Change and the Aggregate Production Function’, Review of
Economics and Statistics August 1957, 39, 312–20.
Solow, Robert, ‘Growth Theory and After’, American Economic Review June 1988,
78:3, 307–17.
Stiglitz, Joseph, ‘Comments: Some Retrospective Views on Growth Theory’, in Peter
Diamond (ed.) Growth, Productivity, and Unemployment, Cambridge: MIT Press,
1990, 50–68.
Summers, Robert, and Alan Heston, ‘A New Set of International Comparisons of Real
Product and Price Levels Estimates for 130 Countries, 1950–1985’, Review of Income
and Wealth March 1988, 34:1, 1–25.
Uzawa, Hirofumi, ‘Optimum Technical Change in an Aggregative Model of Economic
Growth’, International Economic Review January 1965, 6, 18–31.
27 Perspectives on growth theory
Robert M.Solow
Journal of Economic Perspectives (1994) 8, Winter, pp. 45–54
The current wildfire revival of interest in growth theory was touched off by
articles from Romer (1986, from his 1983 thesis) and Lucas (1988, from his
1985 Marshall Lectures). This boom shows no signs of petering out. The time
is not yet ripe for stock-taking and evaluation. My goal is not nearly so
ambitious. All I want to do is to place the new thinking in some sort of
historical perspective, and perhaps sprinkle a few idiosyncratic judgments
along the way.
There have been three waves of interest in growth theory during the past
50 years or so. The first was associated with the work of Harrod (1948) and
Domar (1947); Harrod’s greater obscurity attracted more attention at the
time (and earlier, in 1939), although Domar’s way of looking at things is
more relevant to some of the current ideas.1 The second wave was the
development of the neoclassical model. I think—probably inevitably—that
some misconceptions remain about what that was all about, and why. The
third wave began as a reaction to omissions and deficiencies in the
neoclassical model, but now generates its own alternation of questions and
answers.
THE HARROD-DOMAR IMPULSE
Suppose aggregate output is for some reason—technological or any other—
proportional to the stock of (physical) capital. There is a warrant for this in
the almost-trendlessness of the observed ratio. Suppose that realized saving
and investment (net, for simplicity) is proportional to output and income.
There is similar warrant for this assumption. It follows that investment is
proportional to the stock of capital, and this fixes the trend rate of growth of
both capital and output, unless the rate of capacity utilization is allowed to
go wild. That rate of growth is the product of the investment-output ratio and
the output-capital ratio. If we think entirely in ex post terms, the saving-
income ratio and the investment-output ratio are the same thing. One of the
defining characteristics of growth theory as a branch of macroeconomics is
that it tends to ignore all the difficult economics that is papered over by that
sentence.
650 Robert M.Solow
Now suppose that the required labor input per unit of output is falling at
the rate m~ (which is to say that labor productivity is rising at the rate m~),
again for whatever reason. If the labor force is increasing at the rate n~ , a sort
of impasse arises. Total output must grow at the rate m+n on average, or else
the unemployment rate will rise indefinitely (if output growth is too slow) or
the economy will run out of labor (if growth is too fast). But we have just seen
that the growth rate must satisfy a quite independent condition: it must be
equal to the product of the saving-investment quota (s~) and the output-capital
ratio (a~). The two conditions can be reconciled only if sa=m+n. But there is no
reason why this should ever happen, because the four parameters come from
four wholly unrelated sources.
This construction seemed to have two unpalatable consequences. The first
is that observed economies should spend most of their time experiencing
either prolonged episodes of increasing or falling unemployment rates and/or
prolonged periods of rising or falling capacity utilization. There is no reason
to expect these movements to be confined to minor business-cycle dimensions
or to be quickly reversed. But that is not what the record of the main
capitalist economies looks like.
The second apparent consequence is this. Suppose the first problem can be
evaded. This might happen, for instance, in a developing country with a large
pool of rural labor. It could then have an industrial labor force growing at
whatever the required rate, sa-m, happens to be; the consequences of a
mismatch would be seen only in the waxing or waning of the rural
population. Such an economy could jack up its long-term rate of industrial
growth merely by increasing its investment quota. Under the influence of this
model, that policy was sometimes prescribed. It makes general sense. But if
economic development were that easy, it would be hard to understand why
more poor countries did not follow that route to rapid growth. Even rich
countries would surely want to take advantage of this possibility sometimes.
Something seems to be wrong with this way of looking at long-run economic
growth.
The straightforward way to avoid the first of these awkward conclusions is
to recognize that at least one of the four underlying parameters is likely to be
endogenous. Then the condition sa=m+n may have a solution most or all of
the time; and there may be a plausible adjustment process that will realize the
solution and allow uninterrupted growth to take place. Obviously the
investment-income ratio quota s~ and the output-capital ratio a~ are the natural
candidates for endogeneity.2 Kaldor (1956) and others tried to use s~ in this
way, usually by emphasizing its interpretation as a saving rate, introducing
different saving rates applying to different categories of income, especially
wages and profits, and then focussing on changes in the functional
distribution of income as the mechanism causing the aggregate saving rate to
vary endogenously. (Bertola 1992 is an interesting modern treatment of this
line of thought.) It is fair to say that this way of resolving the problem did not
catch on, partly for empirical reasons and partly because the mechanism
Perspectives on growth theory 651
seemed to require that factor prices be completely divorced from productivity
considerations.
THE NEOCLASSICAL RESPONSE
The standard neoclassical model, of course, resolves the problem by making
the output-capital ratio a~ the endogenous variable. Then labor productivity
growth m will have an endogenous component too, as capital-intensity
changes; but there may remain an exogenous component, loosely identified as
technological progress. This has several related advantages. It fits in well
with the rest of economics; the possibility of increasing the output-capital
ratio by substituting labor for capital is a comfortable and sensible device,
especially on a longish time scale. The implied adjustment mechanism is
plausible and familiar. If sa–m>n, so that labor is getting scarce relative to
capital, one might naturally expect the wage-rental ratio to rise; cost-
minimizing firms would naturally substitute capital for labor. The output-
capital ratio would fall and the economy would move closer to satisfying the
consistency condition. Similarly in reverse. (There the habit of ignoring
aggregate-demand considerations might grate a little. In periods of high
unemployment firms face weak product markets; lower wages could make
things worse.) The assumptions about diminishing returns that are required to
make this mechanism work come easily to most economists. Substitution
along isoquants is routine stuff. That does not count as evidence in favor of
the traditional assumptions, but it explains why the model feels comfortable
to economists. Besides, there is quite a bit of evidence to support the
traditional assumptions, considerably more than there is in the opposite
direction.
Notice that I have not mentioned constant returns to scale. That is because
the model can get along perfectly well without constant returns to scale. The
occasional expression of belief to the contrary is just a misconception. The
assumption of constant returns to scale is a considerable simplification, both
because it saves a dimension by allowing the whole analysis to be conducted
in terms of ratios and because it permits the further simplification that the
basic market-form is competitive. But it is not essential to the working of the
model nor even overwhelmingly useful in an age of cheap computer
simulation.
Everybody knows that fixing up the first awkward implication in this way
(the implication that economies should be experiencing prolonged swings in
unemployment and capacity utilization) also takes care of the second
awkward implication (that growth by raising an investment quota seems
somehow too easy an approach). Diminishing returns to capital implies that
the long-run rate of growth is completely independent of the saving-
investment quota. A closed economy that manages to raise or lower the
fraction of output invested, and sticks to the program, will experience a rise
or fall in its aggregate rate of growth, but only temporarily. Eventually the
652 Robert M.Solow
rate of growth relapses back to its long-run value. This underlying rate of
growth is the sum of n and the technological-progress component of m. The
only permanent effect of the maintained change in investment will be an
upward or downward shift in the level of the trend path, but not in its slope.
Increasing the rate of per capita growth is not only not easy in this model, but
also impossible unless the rate of technological progress can be altered
deliberately.
This reversal of conclusions has led to a criticism of the neoclassical
model: it is a theory of growth that leaves the main factor in economic
growth unexplained. There is some truth in that observation, but also some
residual misconception. First of all, to say that the rate of technological
progress is exogenous is not to say that it is either constant, or utterly erratic,
or always mysterious. One could expect the rate of technological progress to
increase or decrease from time to time. Such an event has no explanation
within the model, and may have no apparent explanation at all. Or else it
might be entirely understandable in some reasonable but after-the-fact way,
only not as a systematic part of the model itself.
Second, no one could ever have intended to deny that technological
progress is at least partially endogenous to the economy. Valuable resources
are used up in pursuit of innovation, presumably with some rational hope of
financial success. The patent system is intended to solidify that hope, and thus
attract more resources into the search for new products and processes. It
would be very odd indeed if all that activity had nothing to do with the actual
achievement of technological progress. The question is whether one has
anything useful to say about the process, in a form that can be made part of
an aggregative growth model. I will suggest later on that this is probably the
most promising aspect of the current third wave of growth theory, even if
much that has been written on the subject so far seems simplistic and
unconvincing.
NEWER ALTERNATIVES
The direction taken at first by the newer growth-theoretic models was not
toward a direct approach to the economics of technological progress. It was
something much simpler: a straightforward abandonment of the idea of
diminishing returns to ‘capital’ (now interpreted as the whole collection of
accumulatable factors of production, one of which might be labelled human
capital or even the stock of knowledge). This stage of the revival could be
described as a return to generalized Domar, but with sophisticated bells and
whistles. Among the bells and whistles were allowance for substitutability
between capital and labor and between various forms of capital, allowance
for only asymptotic absence of diminishing returns, the adoption of a
representative-agent set-up with infinite-horizon intertemporal optimization to
determine investment (in everything), and the introduction of monopolistic
competition as the underlying market form.
Perspectives on growth theory 653
Here I would like to interject two comments. The modelling of
imperfect competition was made necessary by the appearance of
increasing returns to scale. I have already mentioned that the presence of
increasing returns to scale is not the essence of these newer approaches. It
is perfectly possible to have increasing returns to scale and preserve all the
standard neoclassical results. What is essential is the assumption of
constant returns to capital. The presence of increasing returns to scale is
then inevitable, because otherwise the assumption of constant returns to
capital would imply negative marginal productivity for non-capital
factors. Anyway, I register the opinion that the incorporation of
monopolistic competition into growth theory is an unambiguously good
thing, for which the new growth theory can take a bow (along with a
derived curtsey to Dixit and Stiglitz).
I cannot say the same about the use made of the intertemporally
optimizing representative agent. Maybe I reveal myself merely as old-
fashioned, but I see no redeeming social value in using this construction,
which Ramsey intended as a representation of the decision-making of an
idealized policy-maker, as if it were a descriptive model of an industrial
capitalist economy. It adds little or nothing to the story anyway, while
encumbering it with unnecessary implausibilities and complexities.
Now I return to the question of constant returns to capital. It may not be
generally recognized how restrictive this assumption is. There is no tolerance
for deviation. Lucas emphasized in his 1988 article that a touch of
diminishing returns to capital (human capital in his case) would change the
character of the model drastically, making it incapable of generating
permanent growth. He did not notice that a touch of increasing returns to
capital would do the same, but in a quite different way. Since I have not seen
this acknowledged in the literature I will spell it out here.
Suppose that the production function is f(K,L), with non-decreasing returns
to capital. Treat L as constant for the moment, so we can think of this as just
f(K). Let net investment be the fraction s of output so that the time path of K
is determined by dK/dt=sf(K). It is obvious on the face that there is potential
for fairly explosive behavior if f(K) increases more and more rapidly with K.
For instance, if f(K)/K increases with K, the rate of growth of K gets faster as
K gets larger. Then the time path for this growth model has the property that
the stock of capital becomes infinite in finite time. (It is one thing to say that
a quantity will eventually exceed any bound. It is quite another to say that it
will exceed any stated bound before Christmas.) It takes a little calculus to
show that ‘fairly explosive’ puts it mildly.3
The fragility of the constant-elasticity case is worth pursuing further. I will
choose h=0.05 to represent a fairly small degree of increasing returns to
capital. If Y=K1.05, increasing K by 20 percent will increase Y by a bit more
than 21 percent. This is already a fairly weak dose of increasing returns, and
might even be empirically undetectable. Anything more would have even
more drastic consequences. The capital-output ratio is of order of magnitude
654 Robert M.Solow
about one, to be conservative. A straightforward calculation shows that
output will be infinite in about (1/sh) years.4 If s is about 0.1 and h is as small
as 0.05, a country like Germany or France will achieve infinite output in
about 200 years, or even a shorter time from ‘now’. They should live so long,
one is inclined to say.
Of course this kind of calculation should never be taken literally, but it
teaches an important lesson. The knife-edge character of the constant-returns
model can not be evaded by the obvious dodge: oh, well, so it blows up in
finite time—that time could be a million years from now, by which time we
will have evolved into God knows what. For the Land of Cockaigne to be a
million years away, 1+h would have to be so close to 1 that we would never
be able to discern the difference. The conclusion has to be that this version of
the endogenous-growth model is very un-robust. It can not survive without
exactly constant returns to capital. But you would have to believe in the tooth
fairy to expect that kind of luck.
This branch of the new growth theory seems unpromising to me on straight
theoretical grounds. If it found strong support in empirical material, one
would have to reconsider and perhaps try to find some convincing reason why
Nature has no choice but to present us with constant returns to capital. On
the whole, however, the empirical evidence appears to be less than not strong;
if anything, it goes the other way.
A particular style of empirical work seems to have sprung from the
conjunction of growth theory and the immensely valuable body of
comparative national-accounts data compiled by Summers and Heston
(1991). It rests on international cross-section regressions with the average
growth-rates of different countries as the dependent variable and various
politico-economic factors on the right-hand side that might easily affect the
growth rate if the growth rate were easily affected. I had better admit that I
do not find this a confidence-inspiring project. It seems altogether too
vulnerable to bias from omitted variables, to reverse causation, and above all
to the recurrent suspicion that the experiences of very different national
economies are not to be explained as if they represented different ‘points’ on
some well-defined surface. These weaknesses are confirmed by Levine and
Reinelt (1992) and Levine and Zervos (1992), who find that these cross-
section regressions are not robust to the choice of explanatory variables and
are otherwise statistically unprepossessing. More strictly focussed studies—I
am thinking especially of Mankiw et al. (1992) and Islam (1992)—seem to
favor some extended version of the neoclassical model.
The temptation of wishful thinking hovers over the interpretation of
these cross-section studies. It should be countered by cheerful skepticism.
The introduction of a wide range of explanatory variables has the
advantage of offering partial shelter from the bias due to omitted
variables. But this protection is paid for. As the range of explanation
broadens, it becomes harder and harder to believe in an underlying
structural, reversible relation that amounts to more than a sly way of
Perspectives on growth theory 655
saying that Japan grew rapidly and the United Kingdom grew slowly
during this or that period.
I think that the real value of endogenous growth theory will emerge from
its attempt to model the endogenous component of technological progress as
an integral part of the theory of economic growth. Here too the pioneer was
Romer (1990). Many others have followed his lead: my short list includes
Grossman and Helpman (1991), Aghion and Howitt (1992), Stokey (1991)
and Young (1991, 1993), but there are others.
This is a very hard problem for a number of reasons. For one thing, there
is probably an irreducibly exogenous element in the research and
development process, at least exogenous to the economy. Fields of research
open up and close down unpredictably, in economics as well as in science and
technology. This is reflected, for instance, in the frequency with which
research projects end up by finding something that was not even contemplated
when the initial decisions were made. There is an internal logic—or
sometimes non-logic—to the advance of knowledge that may be orthogonal to
the economic logic. This is not at all to deny the partially endogenous
character of innovation but only to suggest that the ‘production’ of new
technology may not be a simple matter of inputs and outputs. I do not doubt
that high financial returns to successful innovation will divert resources into
R&D. The hard part is to model what happens then.
A second difficulty, no doubt related to the first, is the large uncertainty
surrounding many research projects. It is possible that some of this
uncertainty is not probabilistic: if ‘Knightian uncertainty’ shows up anywhere,
it could be here. If so, then appropriate analytical techniques are lacking.
Third, it is not clear how you would know if you had a promising model.
Surface plausibility is one criterion, but hardly a sufficient one. The best
source of empirical material may be historical case studies, but then the test
of truth is bound to be fuzzy.
There are, of course, historians and sociologists, as well as economists,
who study the R&D process in contextual detail. Their insights and
conclusions are usually not in a form that can be used by a macroeconomic
model-builder, and they may even regard the necessary abstraction and
codification as a kind of violation. Even so, there is no excuse for ignoring
the generalizations that emerge from other styles of work. Models of
innovation can be constructed out of thin air, but it is surely better to use
more durable materials if they are available. The best bet, no doubt, would
be collaboration between model-builders and those who use informal
methods, to compromise between one side’s need for definiteness and the other
side’s sense of complexity.
All the difficulties notwithstanding, it seems to me that the body of work I
have just cited has an air of promise and excitement about it. Aghion and
Howitt (1992) manage to give some precision to Schumpeter’s vague notions
about ‘creative destruction’. They make a formal model in which each
innovation kills off its predecessors. It is obvious that some innovations
656 Robert M.Solow
reduce or wipe out the rents that might otherwise have accrued to previous
innovations, and this fact of life has to be taken into account in any
understanding of the process. But sometimes—who knows, maybe just as
often—innovations are complementary with predecessors and add to their
rents. This possibility matters too. Is there any non-mechanical way to take
both contingencies into account? (Schumpeter is a sort of patron saint in this
field. I may be alone in thinking that he should be treated like a patron saint:
paraded around one day each year and more or less ignored the rest of the
time.)
It seems to me that there is great merit in Alwyn Young’s (1993) project of
treating learning-by-doing as one mode of productivity increase, but not the
only one. It is an important fact of life that many instances of product
improvement and cost reduction have little to do with the R&D activity, but
originate in some other way, for instance from the cumulation of small
suggestions coming from production workers, process engineers, and even
customers. Categorical R&D spending may be an inadequate measure of the
resources devoted to increasing productivity. How to understand and model
that other way is an important question. Growth theorists might profit from
picking the brains of informed observers of industry.
This is a good place for me to insert a few more idiosyncratic criticisms of
the new wave. Much of the advanced literature uses the ‘new product’ as a
universal metaphor for innovation. Even cost reduction is often supposed to
come about via the invention of new intermediate goods. The development of
new products is certainly a prominent feature of the technological landscape,
but one is permitted to wonder if that is the only way to go, or even the best
way. Any particular metaphor can impose a bias on subsequent trains of
thought.
The idea of endogenous growth so captures the imagination that growth
theorists often just insert favorable assumptions in an unearned way; and then
when they put in their thumb and pull out the very plum they have inserted,
there is a tendency to think that something has been proved. Suppose that the
production function is Af(K, L) where A carried (Hicks-neutral) technological
progress. (The neutrality is just for clarity; it is inessential.) Successful
innovations make A larger. But how much larger?
For this purpose, take it for granted that there is something meaningful
called ‘an innovation’ and a stream of these innovations occurs as a result of
decisions made by firms. It is easy to agree that the flow of innovations per
unit time depends on the amount of resources devoted to creating them. If an
innovation generates a proportionate increase in A, then we have a theory of
easy endogenous growth. Spend more resources on R&D, there will be more
innovations per year, and the growth rate of A will be higher. But suppose
that an innovation generates only an absolute increase in A: then greater
allocation of resources to R&D buys a one-time jump in productivity, but not
a faster rate of productivity growth. I do not know which is the better
assumption, and these are only two of many possibilities. But merely to adopt
Perspectives on growth theory 657
the more powerful assumption is no more than to assume the more powerful
conclusion.
Ideally, such modelling decisions should be made in the light of facts.
Unfortunately there are not a lot of usable facts to be digested. One could
hope for some enlightment from case studies of industries, technologies, and
R&D decisions. Even that is not easy: it takes two to tango and the authors of
case studies do not like to see their insights reduced to terms in a highly-
simplified equation. Nevertheless I think the best candidate for a research
agenda right now would be an attempt to extract a few workable hypotheses
from the variegated mass of case studies, business histories, interviews, expert
testimony, anything that might throw light on good ways to model the flow of
productivity-increasing innovations and improvements. Finally I would like to
call attention to an interesting paper by Caballero and Jaffe (1993) who made
an ingenious start on exploiting whatever data there are. I am not necessarily
endorsing all their conclusions, but rather their willingness to sift through a
lot of data looking for reasonable generalizations.
NOTES
1 Harrod’s exposition tended to rest on incompletely specified behavioral and
expectational hypotheses. Domar focussed more straightforwardly on the
requirements for equilibrium of demand and supply in steady growth.
2 In principle there is no reason to exclude the endogeneity of m and n. But induced
changes in population growth, although an important matter in economic
development, seemed not to figure essentially in the rich countries for which these
models were devised. The idea of endogenous technological progress was never far
below the surface. In those days it would have seemed rash to conjure up some
simple connection between the allocation of resources and the rate of growth of
productivity. Kaldor and Mirrlees’ ‘technical progress function’ (1962) was an attempt
that apparently did not seem plausible. I would recommend Karl Shell’s papers
(1966, 1967, 1973) as an indication of how far a technically-sophisticated and well-
read economist of the time would have been willing to go. There has been some
progress since those papers, but not a whole lot.
3 The solution of this differential equation is given by . Now
suppose that the improper integral converges to a number J (which will
depend on K(t
0
) though this is not significant). Indeed the capital stock approaches
infinity as t gets closer and closer to t
0
+(J/s). If the production function will generate
infinite output from infinite capital (as with Cobb and Douglas or a better-than-unit
elasticity of substitution between labor and capital) then aggregate output and income
become infinite at that time too. Allowing employment to increase can only hasten the
date of the Big Bang. If output is finite even with infinite capital, the economy will
achieve its maximal output in finite time. That is what I meant by saying that the model
changes its character in a different way. What will make that improper integral converge?
Clearly it is more likely to do so if f(K) increases very rapidly with K. It can not do so
if f(K) is concave or linear. There are convex functions f(K) for which the integral
diverges. But increasing returns to capital helps a lot. It is easy to see that the integral
converges if f(K)=K1+h for any positive h, no matter how small.
658 Robert M.Solow
4 When f(K)=K1+h, the number J is, , which is K(t
0
)-h/h. Since
Y(t
0
)=K(t
0
)1+h,K(t
0
)-h=K(t
0
)/Y(t
0
). Thus the date of the Big Bang satisfies s(t-t
0
)=h-
1K(T
0
)/Y(t
0
). Solving for t shows that the date of the Big Bang (the end of scarcity as
we know it) occurs at t
0
+(K(t
0
)/Y(t
0
))(sh)-1.
REFERENCES
Aghion, P. and P.Howitt, ‘A Model of Growth through Creative Destruction’,
Econometrica March 1992, 60:2, 322–52.
Bertola, G., ‘Wages, Profits and Theories of Growth’, International Economic Association
Conference Paper, Varenna, Italy, 1992. In L.Pasinetti and R. Solow (eds) Economic
Growth and the Structure of Long-Term Development, New York: St Martin’s Press,
1994.
Caballero, R. and A.Jaffe, ‘How High are the Giant’s Shoulders: An Empirical Assessment
of Knowledge Spillovers and Creative Destruction in a Model of Economic Growth’,
NBER Macroeconomics Annual, 1993, Cambridge and London: MIT Press, 1993,
15–74.
Dixit, Avinash, and Joseph E.Stiglitz, ‘Monopolistic Competition and Optimum Product
Diversity’, American Economic Review June 1977, 67:3, 297–308.
Domar, E., ‘Expansion and Employment’, American Economic Review, March 1947,
37:1, 343–55.
Grossman, G. and E.Helpman, Innovation and Growth in the World Economy,
Cambridge: MIT Press, 1991.
Harrod, R.F., Towards a Dynamic Economics, London: MacMillan, 1948.
Islam, N., ‘Growth Empirics: A Panel Data Approach’, unpublished paper, Harvard
University, 1992.
Kaldor, Nicholas, ‘Alternative Theories of Distribution’, Review of Economic Studies
1956, 23:2, 83–100.
Kaldor, Nicholas and J.Mirrlees, ‘A New Model of Economic Growth’, Review of
Economic Studies June 1962, 29:3, 174–92.
Levine, R. and D.Reinelt, ‘A Sensitivity Analysis of Cross-Country Growth Regressions’,
American Economic Review September 1992, 82:4, 942–63.
Levine, R. and S.Zervos, ‘Looking at the Facts: What We Know about Policy and Growth
from Cross-Country Analysis’, Unpublished paper prepared for International
Economic Association Conference on ‘Economic Growth and the Structure of Long-
Term Development’, Varenna, Italy, 1–3, October 1992.
Lucas, R., ‘On the Mechanics of Economic Development’, Journal of Monetary Economics
July 1988, 22:1, 3–42.
Mankiw, N.G., D.Romer, and D.Weil, ‘A Contribution to the Empirics of Economic
Growth’, Quarterly Journal of Economics May 1992, 107:2, 407–37.
Romer, P., ‘Increasing Returns and Long-Run Growth’, Journal of Political Economy
October 1986, 94:5, 1002–37.
Romer, P., ‘Endogenous Technological Change’, Journal of Political Economy October
1990, 985:2, S71–102.
Shell, Karl, ‘Toward a Theory of Inventive Activity and Capital Accumulation’, American
Economic Review May 1966, 56:2, 62–68.
Shell, Karl, ‘A Model of Inventive Activity and Capital Accumulation’, in K.Shell (ed.)
Essays on the Theory of Optimal Economic Growth, Cambridge: MIT Press, 1967,
67–85.
Shell, Karl, ‘Inventive Activity, Industrial Organization and Economic Growth’. In
J.A.Mirrlees and N.Stern (eds) Models of Economic Growth, Macmillan: London,
1973, 77–100.
Perspectives on growth theory 659
Stokey, N., ‘Human Capital, Product Quality, and Growth’, Quarterly Journal of
Economics May 1991, 106, 587–616.
Summers, Robert, and Alan Heston, ‘The Penn World Table (Mark 5: An Expanded Set
of International Comparisons, 1950–1988)’, Quarterly Journal of Economics May
1991, 106:2, 327–68.
Young, A., ‘Learning by Doing and the Dynamic Effects of International Trade’, Quarterly
Journal of Economics May 1991, 106:2 369–406.
Young, A., ‘Invention and Bounded Learning by Doing’, Journal of Political Economy
June 1993, 101:3, 443–72.
Abel, A.B. 411, 577
Abraham, K.G. 433
Abramovitz, M. 20, 577, 578, 612, 618,
629
Ackley, G. 19
Adams, E. 602 (n19)
administered price hypothesis 490
agents: fluctuations 361; General
Theory 275; information 287;
maximizing 266–7, 281, 297, 337,
338–9, 349–50; optimizing 266–7,
280–1, 282, 291 (n6); price
mechanisms 350–1; and
principals 556; rational 12–13, 290,
297, 337, 361; see also representative
agents
aggregate demand: disturbances 32,
281; employment 71, 110–11, 279;
fiscal policy 110, 114; fluctuations 5,
136; government intervention 113;
inflation 223–4, 396; market
imperfections 5, 188, 305; monetarist
view 187, 188–9; money supply 281;
policy 110, 114, 467; price levels 148,
150–2, 191, 298–300, 336, 496–7;
unemployment 17
aggregate supply 12, 65, 67, 111, 298–
300, 307–9, 336, 557
Aghion, P. 655–6
Akerlof, G.A. 125–6, 454, 455, 464,
512, 563
Aksoy, E.G. 100, 104–5
Alchian, E.S. 516
Allen, S.G. 492, 542 (n21, n22)
Alogoskoufis, G. 492
Altonji, J.G. 391, 432
altruism 318–19
Andersen, L. 181, 210 (n32)
Andersen-Jordan model 193
Ando, A. 217, 317
animal spirits 65–6, 79, 84–5 (n11)
arbitrage 301–2
Arrow, K.J.: endogenous growth 639,
642; equilibrium studies 142, 279,
285, 411; knowledge spillovers 632,
639; learning by doing 618–20, 622
Ashenfelter, O. 378
Austrian school 2, 227, 229, 232, 352
(n1)
author: intention 56–8, 70; original
meaning 103, 105, 106
Azariadis, C. 522
Backhouse, R.E. 265
Backus, D.K. 493
Bailey, M.J. 321
Baily, M.N. 522
Bain, I.R. M. 366, 384
balance of payments 46, 216, 229–32
Balke, N. 537
Ball, L.: coordination failure 136;
disinflation 459; labour supply 432;
market failure 136; menu cost theory
498, 513; new Keynesianism 460;
rigidities 462; staggered prices 514,
515, 517, 528
Ballard, C.L. 415
banks, central 194, 197, 200, 236, 237,
467, 559–60
bargaining models 523, 561
Barro, R.J.: budget deficits 324; business
cycles 14, 283, 361, 393, 433, 474;
disequilibrium 139, 500; economic
growth theory 2, 18, 19, 577, 612,
615 (n1), 633, 634; equilibrium
competitive model 347; gains from
trade 454, 501, 502; government
intervention 411, 472; market
Index
Note: page numbers in italics denote figures or tables where these are separated from
their textual references.
Index 661
clearing 228, 502, 503; menu cost
462, 509; real business cycles 428;
Ricardian debt equivalence 13, 266;
savings 317; tax reduction 122–3;
wage/price 344–5, 520
Barro-Grossman spillover model 521
Barsky, R. 322
Baumol, W. 20, 218, 502, 578, 629
Baxter, M. 411
Becker, G.S. 375, 391
Benassi, C. 442
Benassy, J.-P. 503
Benhabib, J. 371
bequests 317–19
Bernanke, B.S. 3, 577
Berndt, E. 392
Bernheim, D. 318
Bertola, G. 517, 650
Bewley, T.F. 371
birth rates 612, 615 (n4)
Black, F. 412
Blackburn, K. 13
Blanchard, O.J.: credit rationing 481;
externality 126; new Keynesianism
473–4, 532; price adjustment 486,
514, 519; represenative agent 508,
509; surprise supply function 12;
taxation 411; unemployment 524,
528, 529
Blanchard-Kiyotaki model 508, 509,
513, 516, 521
Blaug, M. 95
Blinder, A.: efficiency wages 126–7;
Keynesianism 32, 453, 540 (n1); new
Keynesianism 461–2, 465, 473–4;
Phillips curve 10; production and
inventories 384, 558; rational
expectations 113
Blume, M.E. 374
Brander, J.A. 612
Brazil, wages/prices 170–1
Bresnahan, T.F. 491, 492, 518
Bretton Woods 10, 160, 229, 231
Britain: capital stock 596; exchange rates
230–1; GNP and prices 539; inflation
rates 536–7; inflation-unemployment
trade-off 230; money demand shifts
219–20; price adjustment 495;
productivity 591, 601 (n11)
Brock, W.A. 371, 402, 411
Brown, A. 222
Brown, E.C. 323
Brumberg, R. 317
Brunner, K. 1; monetarism 159, 160,
181, 220, 237; political factors 81;
quantity theory 182, 337;
transmission process 186, 217, 225
Bryce, R. 61
Buchanan, J. 266, 321, 330 (n2)
budget, government 44–5, 75–6
budget deficit: benefits 468–9; empirical
evidence of effects 325–8; GNP
327–8; inflation 323–4; interest rates
325–6; national saving 317–18;
Ricardian alternative 315–16, 325–9;
standard model 314–15; taxation
315–16, 320, 323
Buiter, W. 300, 317
business cycle 367–9; consumption/
leisure 393; economic growth 402–4;
fluctuations 140–2, 286, 420 (n4);
Keynes 66, 72, 137; labour market
523; market failure 455; money/
output 14, 433–4; stabilization
policies 112–14; supply-side shocks
14, 19; Tobin 14, 374; unemployment
226; see also equilibrium business
cycle theory; real business cycle
theories
Caballero, R. 517, 657
Cagan, P. 219, 225
calibration 16, 363, 415
Calomiris, C.W. 572 (n25)
Campbell, J.Y. 466
Canada: current-account deficits 328;
monetary growth 236, 237; savings
326
capital 68, 595–6, 649, 651
capital accumulation 398–9, 402–3, 412,
417, 577–8, 616
capital goods, heterogeneity 40
capitalism 5–6, 136–7
Caplin, A.S. 510
Carabelli, A. 99
Carey, K. 3
Carlton, D.W. 490–1, 492, 533
Carroll, C. 326
Carter, M. 265
Cass, D. 371, 399
catch-up hypothesis: developing
countries 593–6, 629; economic
growth theory 578, 582, 583–6;
historical examples 587–91; leaders/
followers 593–6; OECD countries
607–9
662 Index
Cecchetti, S. 491
Champernowne, D.G. 61
Chan, L. 322
Chenery, H.B. 623
Chick, V. 2, 96
Cho, F. 413
choice theory 36–7, 38–9, 396, 397, 456
Christensen, M. 13
Chrystal, K.A. 1
classical economics 2, 3, 5, 36, 41, 285,
427–8, 447; dichotomy 426–7;
unemployment 138, 139
Clower, R.W. 46, 47–8, 227, 502
Coddington, A. 30, 84 (n8), 95
Coe, D.T. 493
Colander, D.C. 29
competition: endogenous growth 636;
imperfect 461, 561, 571 (n17), 639,
653; monopolistic 125–6, 491–2,
508–9, 521, 652; new Keynesian
economics 491–2; perfect 491–2; and
productivity 599
complementarity, strategic 139
conservative ideology 110, 120–1
consistency, textual 100
construction industry 558
consumer price index 505
consumption 409; Friedman 502;
government 614–15 (n1); household
379; and investment 71, 366, 368–9;
leisure 366, 378, 380–1, 393, 399,
413, 416, 428–9; prices 6; and saving
326–7
contracts: fixed-income 522–3; implicit
522–3; incomplete 554; long-term
285, 344; staggered 514–15, 527–8;
wages 147, 514, 527–9
Cooley, T.F. 414
Cooper, R. 139, 454, 503, 569 (n9)
coordination failures 17, 136, 454,
493–40, 503–4, 518
Council of Economic Advisers 323
Crafts, N. 21, 577
credit 197, 209 (n17), 319–21
credit control 219
credit markets 559–60
credit rationing 481, 559
Cross, R. 11, 17
Crusoe metaphor 399–402, 405
cumulation hypotheses, input-output
table 519
current-accounts 327–8
customer markets 515–17, 522
cycle, and trend 14–15
Danthine, J.P. 16
Darby, M. 318
Darity, W. 6, 31
Davidson, P. 2, 461
De Prano, M. 217
Debreu, G. 142, 279, 285, 371, 400, 417
debt 149, 314, 319–20, 557, 569 (n9)
debt finance 442, 555
decentralization 126
deconstructionism 57
deficit financing 75, 76
deflation 14, 121, 153, 169, 172
DeLong, J.B. 481, 607, 629
demand for money 218–20, 222, 239,
248; see also aggregate demand;
effective demand
demographic change 612, 614
Denison, E.F. 586, 617
depression, European 123
Depression, Great: causes 3, 5, 173;
declining prices/production 490;
Federal Reserve 3, 8, 164–6, 174;
Keynesians 19, 164, 450; real balance
effect 149; wages/prices 563
desert island parable (Phelps) 522
destabilization 200
Deutscher, P. 4–5
developing countries: catch-up
hypothesis 593–6, 629; economic
growth 20–1, 623–5; trade 644
Diamond, P.A. 454
differential game theory 348
Dilthey, W. 102
discoveries, as inputs 636–8
discretion policies 449–50, 468
disequilibrium 46–8, 139, 145, 224, 349,
500–1
disinflation 13, 17, 436 (n5), 459
Dixit, A. 642, 653
Dixon, H. 9
Dixon, H.D. 439
Dolde, W. 374
Domar, E. 649
Donaldson, J.B. 16
Dornbusch, R. 352 (n5)
Dowrick, S. 20, 607, 609, 612
Dunlop, J.T. 488, 523
Eckstein, O. 191, 394, 448
econometric models 217, 271–2, 274,
282, 285–90, 291 (n1); rational
expectations 117, 119, 124, 282
econometric policy evaluation,
Lucas 12–14
Index 663
economic growth: analysis, Solow 579;
Barro 2, 18, 19, 577, 612, 615 (n1),
633, 634; business cycles 402–5;
catch-up 578, 582, 583–6;
convergence hypothesis 20; developed/
developing countries 20–1;
endogenous 20–1, 413–14, 577–8,
623–5, 628, 636, 639, 641–2, 655,
656; evolutionary 645 (n6);
fluctuations 15–16; GNP/hours
worked 377; governments 21, 578–9,
616; investment/output 633, 649;
Lucas 20, 21, 369, 371, 373, 413,
629, 636, 639–40, 644, 649, 653;
Mankiw 19, 577, 578, 634, 654;
national accounting data 327, 654;
neoclassical model 19–20, 369, 577,
614, 616–17, 628, 638, 649, 651–2;
neo-Schumpeterian 641–3; policies
611–14; renaissance 2, 18–21;
replication 638–9; Romer, P. 19, 20–1,
579; sectors 384; sources 612;
technological progress 403;
worldwide 604, 605–11
economic growth models 366–7, 369–73,
377–8; data restrictions 373–6;
extensions 383–4; Romer, D. 19, 634,
639, 640, 642, 644, 649, 655; Solow
14, 15, 19–20, 399, 447, 577, 616–18,
638; statistical behaviour 377–83;
Stiglitz 642, 653
Economic Journal 161, 578–9
Economic Report of the President 249
economics: academia/business 118,
120–1; aims 3, 40–1; generational
conflict 117–18; micro vs macro 30;
polarization 180; schools of 1–2, 439,
440, 473; sociology of 117–19
economies, less advanced: see developing
countries
education 584–5, 634
effective demand: failures 97, 453;
Keynes 6, 58–9, 60–1, 65, 66–7;
Tobin 33, 136, 138–40
efficiency wages: Blinder 126–7; Gordon
525–7; Mankiw 462–3; models 18,
531–2; theories 561, 571 (n16)
Eichenbaum, M.S. 375, 390, 391, 414
Eichengreen, B. 3
Eisner, R. 324
Ellis, H. 167
employees: hiring and training 558–9;
risk averse 522
employment: aggregate demand 71,
110–11, 279; deepening 612;
equilibrium 3, 137; exports/ imports
594; fluctuations 16, 342, 366, 431–2,
433; full 324–5, 463; fundamentalist
Keynesianism 40; General Theory 58,
59; government intervention 7, 59,
73–5; growing 604–5, 606, 610, 611,
612–13; hydraulic approach 50–1;
investment 71, 73–5; monetary policy
169–73; new Keynesian economics
463; output 45, 336; prices 286;
rational expectations 298; real
business cycles 431–2, 433; Snowdon
7, 160; Tobin 286, 449; wages 45,
560–1; see also labour market
Employment Act (1946) 450
endogenous growth: see economic
growth, endogenous
equilibrium: Arrow 142, 279, 285, 411;
business cycle 12, 278–83; competitive
347, 371, 399–400, 619, 620;
dynamic 397–8; employment 3, 137,
286; fundamentalist view 39–40;
general 39, 62–3, 502, 552, 553;
Keynes 49–50, 279; labour market
307; market-clearing 141, 284–5,
349; markets 37, 39, 425–6; models
288; new classical 349; new Keynesian
500–1; Phelps 285; Prescott 361–2,
366, 371, 400; prices 50; rational
expectations 346–7, 348; reconsituted
reductionism 46–9; theories 39, 289,
361–3; underemployment 149;
unemployment 6–7, 17; Walrasian
141, 170, 343, 425–7, 429
equilibrium business cycle theory
278–80, 474; criticised 283–90;
market-clearing 280; models 500; new
classical economics 501–2; technology
141, 142
equity market 555–6
error terms, econometrics 274, 287
Esposito, L. 326
Europe: depression 123;
unemployment 525
European Monetary System 123
Evans, P. 325–6, 328
exchange mechanism 393–4
664 Index
exchange rates: fixed 234;
flexibility 231; Keynes 72–3;
monetarist 229–32; theory 216;
US 176
exegesis, scientific/personal 99–100
expectations: adaptive 312 (n43); faulty
336–7; General Theory 65, 275–6;
inflation 13, 300; Keynes 572 (n24);
Lucas 226, 287, 297, 301, 336, 337;
stabilization 153; uncertain 41–2; see
also rational expectations
expenditure 41, 45
exports 593, 594
externalities 126
Fagerberg, J. 577
Fand, D. 181, 195
Fay, J. 392
Federal Reserve Bank of Minneapolis
Quarterly Review 361, 362
Federal Reserve System: destabilizing
433; Great Depression 3, 8, 164–6,
174; interest rates 168, 196;
Monetarism 249–54; monetary policy
164–6; over-reacting 177
Fei, J.C.H. 616
Feldstein, M. 315, 316, 321, 447–8
Fender, J. 96
fine-tuning 77, 112, 233–4, 467
firm theory 555–9, 565–6, 636
fiscal policy: aggregate demand 110, 114;
first order effects 325–8; Keynesian
view 7, 232–8, 448–9; limitations
161; monetarism 232–8, 248; neo-
Austrians 232, 236; New Keynesian
economics 469, 471–2
fiscalism 7, 43
Fischer, S.: Blanchard-Kiyotaki model
508; CPI 505; credit rationing 481;
long-term contracts 285, 344; new
Keynesianism 521; nominal demand
distribution 17; rational expectations
113, 455; rules/ discretion 13;
staggered wage contracts 147, 514,
527–8; wage stickiness 454, 478
Fish, S. 57, 105
Fisher, I. 68, 84 (n8), 144, 149, 153,
169, 225
Fisher effect 149, 150–1, 152
Fitzgibbons, A. 98, 99
fixed-price models 503
fixprice method 145
flexibility, destabilizing 33
flexprice/fixprice markets 227
fluctuations: agents’ response 361;
aggregate demand 5, 136; business
cycle 140–2, 420 (n4); classical
approach 427–8; economic growth
15–16; employment 16, 342, 366,
431–2, 433; exchange mechanism
393–4; Keynesian approach 427–8;
output 16, 366; prices 342; real
business cycle theory 14–16, 393–4,
425, 427–8; Romer, D. 18; and trend
367, 403
forecasting 273, 287
France: GNP and price data 538–9;
inflation rates 536–7; price
adjustment 495; productivity 596
Frenkel, J.A. 160
Friedman, B. 288, 289, 291 (n6), 474
Friedman, Milton: business cycles 433;
consumption 502; demand for money
218, 220, 222; Great Depression 3;
inflation 335, 337; as influence 471;
and Keynesianism 396, 460; market-
clearing 111; Monetary History of the
United States 8, 159; monetarism
7–11; monetary policy 281, 485;
natural rate hypothesis 113–14, 304,
486, 488; Phillips curve 434, 449;
predictive power 204; quantity theory
of money 7–8, 159–60, 182, 204,
216; ‘Role of Monetary Policy’ 9, 159;
Tobin on 9, 160; transmission
mechanism 217; unemployment 9–10,
12, 224–5; wages/inflation 9
Friend, I. 374
Frisch, H. 222
Frisch, R. 286, 288
Fromm, G. 191
Froyen, R.T. 577
fundamentalist Keynesianism 30, 37–42
Gadamer, H.-G. 104
Garrison, R. 2
GDP per capita 14–15, 604, 605–6, 607,
609
Gemmell, N. 612
gender, labour force participation 612
General Theory of Employment, Interest
and Money (Keynes): agents’ own
interests 275; atomistic view 95–102;
Index 665
effective demand 58–9; employment
58, 59; expectations 65, 275–6;
general-equilibrium interpretations
62–3; impact 3–5, 29–31, 446–7;
innovation 77–8; interpretations
55–6, 61–7, 78–9, 457–8; investment
65–6; Leijonhufvud on 47–8; national
income 274; Patinkin on 30–1, 58,
60, 63–4, 68, 78; political message 78,
81; prices 58, 67, 191; Shackle on 50,
65, 67, 79, 96; uncertainty 65, 66;
unemployment 68; wages 67
Germany: GNP and price data 538–9;
inflation rates 536–7; price
adjustment 495
Gerrard, B. 1, 2, 31, 95
Ghez, G.R. 375, 391
GNP: budget deficits 327–8; economic
growth 377; elasticity 255–6; hours
worked 377; indexation of prices
479–80; indexation of wages 505;
limitations of measurement 382–3;
monetary growth 250, 251, 255–6;
and prices 481–2, 538–9;
unemployment 301
gold standard 3, 174–5
Gordon, D.F. 522
Gordon, R.J.: coordination failure 17;
efficiency wages 525–7; on Friedman
9, 160; monetary policy 14;
neoclassical growth 577; new
Keynesian economics 441, 478–80,
500–6; Phillips curve 114, 123;
rational expectations 306, 454;
unemployment 525; US, pre-/post-war
492, 493, 535, 537
government: consumption 614–15 (n1);
economic growth 21, 616; impotence
result 298–300, 309, 311 (n42);
money supply 304; policy goals 44–5,
302; public goods 640; stabilization
295–6, 302, 304
government intervention: aggregate
demand 113; countercyclical policy
283–4; economic growth 578–9, 616;
employment 7, 59, 73–5; externalities
126; Keynesians 5, 29; monetarists
202–3, 216, 305; new Keynesian
economy 456; real business cycle
theory 410–12; unemployment 9
gradualism 236
Graham, F.D. 166
Gramlich, E. 193
Grandmont, J.-M. 563
Gray, J.A. 505
Greenwald, B.C. 18, 439, 441–2, 472,
473, 474, 553, 559
Griliches, Z. 640
Grilli, V. 361, 577
Grossman, G.M. 579, 623, 643, 655
Grossman, H. 139, 502, 503
growth: see economic growth
Hahn, F.H. 6, 96, 229
Halasi, A. 166–7
Hall, R.E. 285, 344, 411, 435 (n2), 520,
556, 617
Hamburger, M. 247, 249, 255, 259
Hammond, J.D. 8
Hamouda, O.F. 95
Hansen, A.H. 6, 19, 30, 61, 63, 141, 166
Hansen, G.D. 366, 375, 378–81, 414
Hansen, L.P. 374, 390, 413
Harberger, A. 218
Harcourt, G.C. 68, 69, 95
Harrod, R.F. 61, 62, 65, 68, 218, 649
Harrod-Domar growth theory 616,
649–51
Hart, A.G. 66
Hayashi, F. 321
Hayek, F.A. von 397
Heckman, J. 378
Helpman, E. 579, 623, 643, 655
hermeneutics 31, 56, 102–5
Heston, A. 605, 628, 654
Hicks, J.R.: capital accumulation 399;
flexprice/fixprice 227; IS-LM 6, 30,
45, 62, 63, 69; on Keynes 38, 61; on
Solow growth model 617; trends/
fluctuation 403; Value and Capital
397–8
Hirsch, E.D. 57, 102–3
Hollander, S. 100
Homan, P. 166
Hoover, K.D. 363, 454
hours worked: annual growth 410; GNP
377; layoffs/work-sharing 553, 562
household: composite commodity good
378; consumption 379; market
activity hours 380; price expectations
371, 372; production 372–3, 375,
384, 390–1
housing market 236
Howitt, P. 655–6
Howitt, P.W. 339, 351
666 Index
human capital: endogenous growth
413–14, 578; formation 623–4; as
investment 558–9, 623, 634; levels
383–4; research 621, 623;
unemployment 17–18, 129, 412
hydraulic Keynesianism 30, 42–6, 84
(n8), 94
hysteresis effect 17–18, 114, 128–9, 463,
488–90, 525
ideology: conservative 110, 120–1;
economics 270; monetarism/
Keynesianism 203; new classical
macroeconomics 351
imports 383, 593
income: cross-country data 628–36, 644;
distribution 71, 236, 604, 606–7,
610; and expenditure 41; investment
631–2; money stock 182–4; stability
199, 259; taxation 322, 620
income policies 460–1
indexation, prices/wages 505, 518–19,
521, 554
inertia, prices 127–8, 147, 479, 485–7,
494, 500, 524
inflation: aggregate demand 223–4, 396;
budget deficit 323–4; credibility 13;
cross-country comparisons 536–7;
expectations 13, 300; international
160; lags 115–17, 141; Lucas 12;
monetarists 201–2; monetary policy
175; money supply 9–10, 11, 221,
222; new classical economics 334–40,
341–2; new Keynesian economics
256, 449; Phelps 222, 224, 335, 337;
Phillips curve 200–1, 204, 335;
political pressures 203; quantity of
money 337; Romer, D. 436 (n5);
unemployment 9–10, 112, 113, 115–16,
160, 172, 200–1, 222, 230, 277,
304–5, 335, 337, 341, 396–7, 449;
US 165, 497; wage/price controls 9,
235
information: agents 287; availability/ use
306; economic behaviour 347–8;
hidden 302; imperfect 562–3, 566–7;
nonrival 637; replication 637
information costs 297, 301
innovation 638–9, 656–7
input, rival/non-rival 640
input-output 480, 518, 531
insider-outsider theory 129, 146, 462–3,
524–5, 561
institutions, maximizing behaviour 345
intention, authorial 70
interest rates: budget deficits 325–6;
classical adjustment mechanisms
142–9; Federal Reserve 168, 196;
fixed/natural 198; investment 196;
long-/short-term 194–6; manipulation
237; measurement 184, 196, 198;
monetary policy 166, 168–9, 560;
natural/real 170; Ricardo 325–6; rises
559–60
international product cycle 624
international trade theory 594
interpretation: atomistic 31, 95–102,
104; hermeneutics 56, 102–5;
multidimensional 104–5; objectivist
96–100, 103–4, 105; organicist view
31, 102–5; reader’s world-view
98–102; relativist approach 101–2,
103–4, 105; textual consistency 100
investment 409, 611; confidence 65–6;
and consumption 71, 366, 368–9;
economic growth 633, 649;
employment 71, 73–5; equity 556;
General Theory 65–6; human capital
558–9, 623, 634; income per capita
631–2; interest rates 196; Keynes 59;
neoclassical growth 614; and output
649; productivity growth 620
IS-LM model: elaborations of 63–4,
570–1 (n14); equlibrium level of
income 69; GNP 255; Hicks 6, 30, 45,
62, 63, 69; hydraulic economics 45;
Keynes 41, 62, 63; Keynesian school
2, 6–7, 30, 460, 472; monetarism
220–1, 335–6, 352 (n5); rejected 64–5,
67, 68–70; Shackle 79
Islam, N. 654
Israel, fiscal policy 326–7
Iwai, K. 147
Jaffe, A. 657
Jansen, D.W. 577
Japan: economic growth rate 610; GNP
and price data 539; inflation 496,
536–7; price adjustment 495;
productivity 591; youth dependency
613
Jensen, M.C. 556
John, A. 139, 454, 503
Johnson, H.G. 10, 64, 160, 217–18, 351
(n1)
Index 667
Jorgenson, D. 502
Journal of Economic Growth 19
Journal of Economic History 578
Journal of Economic Literature 441
Journal of Economic Perspectives 33,
362, 452, 579
Judd, K. 411
Judd, K.L. 643
Kahn, Richard 69
Kaldor, N. 68, 621, 650, 657 (n2)
Kalecki 68
Kalman filtering formula 289
Kantor, B. 239
Kashyap, A. 491
Katz, L.F. 433
Kawasaki, S. 344
Kehoe, P.J. 374, 493
Kendrick, J. 618
Keynes, John Maynard 3, 488; animal
spirits 65–6, 79, 84–5 (n11); business
cycle 66, 72, 137; classical economics
36, 41, 447; Clower on 47; coarse
tuning 32; Collected Writings 55;
deficit financing 75–6; Economic
Consequences of Mr. Churchill
59–60; effective demand 6, 58–9, 60–1,
65, 66–7; End of Laissez-Faire 5–6;
equilibrium 49–50, 279; exchange
rates 72–3; expectations 572 (n24);
fiscalism 7, 43; General Theory of
Employment, Interest and Money 3–5,
29–31, 37, 38–9, 47–8, 50, 58–61,
65, 457–8; instability of market 5;
investment 59; ‘invisible hand’ 2, 5,
29; IS-LM model 41, 62, 63; labour
market 146, 147–8; Leijonhufvud on
47–8; liquidity preference 198; ‘Long-
Term Problem of Full Employment’
73–5; managed capitalism 5–6;
marginal efficiency of capital 97–8;
Meltzer on 31, 70–1, 75–6; ‘Notes on
Mercantilism, the Usury Laws,
Stamped Money and Theories of
Under-Consumption’ 72; ‘Notes on
the Trade Cycle’ 72; Pigou 201; prices
58, 191; product market 146–7;
Quarterly Journal of Economics
(1937 article) 30, 37, 38–9, 65;
reductionist choice theory 38–9; rules/
discretion 72, 73, 76–7, 81–2; Tobin
on 6, 32–3, 275; Treatise on Money
58, 59–60, 77; unemployment 5–6,
42, 139; vision 97, 103; wages 147–8,
487–8, 572 (n26)
Keynes effect 148, 150–1, 152
Keynesian economics 3–7, 109, 110–14,
135–8, 204, 500–1; aggregate demand
188; in ascendancy 32, 121–9,
495–500; and classical economics
285; in decline 114–21, 445, 450,
452–4; econometrics 265, 274–6,
277–8; employment 324–5; fiscalism
7, 232–8, 448–9; flaws 396;
fluctuations 427–8; Friedman on 396,
460; fundamentalist 30, 37–42;
government intervention 5, 29;
hydraulic 30, 42–6, 84 (n8), 94;
inflation and recession 270; interest
rates 194, 195–6; IS-LM 2, 6–7, 30,
460, 472; liquidity preference 183;
Lucas on 9, 159, 265, 347, 351,
396–7, 445, 450, 498–500; model
size 193–4, 212 (n54); monetary
growth rule 199; monetary policy
448–9; money stock increases 185;
and new classical economics 109–10;
and new Keynesian economics 459–65,
471–3; new theoretical foundations
125–9; prices 191, 192, 394; private
sector 187; quantity theory of money
181–2; recession 111–12; relative
yields 182–3; Sargent on 114, 283,
347, 453, 498–500
Kindahl, J.K. 490, 491
Kindleberger, C. 596
King, M. 321, 620, 622
King, R.G.: consumption/leisure 393; IS-
LM 472–3; labour supply 413, 419;
money/output 434; rational
expectations 567; real business cycles
411, 414, 418; taxation 622–3;
technological progress 404, 405
Kirman, A. 569 (n1)
Kiyotaki, N. 126, 508, 509, 513, 516
Klein, L.R. 43, 63, 187
Knight, Frank 66
knowledge: applied 596; diffusion 586,
633; as input 641; marginal
productivity 620–1; new 620, 621;
nonrival good 622, 638; as
productive factor 619, 620; as public
good 619, 622; spillover 618, 622,
632, 635, 639
Kochin, L. 326
Koopmans, T.C. 371, 399
Kotlikoff, L. 318
668 Index
Kravis, I.B. 586, 588
Kuhn, T.S. 60, 109
Kydland, F.E.: calibration 363, 415;
consumption/leisure 366, 378, 413;
equilibrium 361; real business cycle
16, 414; rules/discretion 13, 471–2;
technology 383
Kydland-Prescott economy 378, 379
labour: female participation 612;
heterogeneity 383–4; hours worked
410, 553, 562; and leisure 298;
marginal productivity 68; turnover
costs 561
labour elasticity, empirical 381–3
labour hoarding 391–2, 466
labour market 520–9; business-cycle
fluctuations 523; competitive 503;
disequilibrium 224; equilibrium 307;
fairness 464; local/national 147–8;
new Keynesians 527–8, 531, 560–2;
non-clearing 284–5; price rigidity 508;
real business cycle theory 412–13;
turnover costs 524; wages 145–6,
227–8, 479
labour productivity 141, 583, 587, 588,
650, 651
labour supply 275; elasticities 378–9,
381–3, 391, 413; in excess 552;
substitution 431–2
labour unions 480, 523–4
Laidler, D.E. W.: Economic Journal article
160–1; inflation 10; monetarism
216–17; money demand 8, 11, 218,
219, 220; new classical economics
266–7, 474; price stickiness 354
(n20); supply shocks 499
Laing, D. 454
laissez-faire policy 29, 120
Landreth, H. 29
Lange, O. 61, 63
Layard, R. 525
Leahy, J. 510
Learner, E.E. 377
learning by doing 618–20, 622, 656
Leibenstein, H. 526
Leijonhufvud, A. 351, 468; coordination
failure 503–4; general equilibrium
502; General Theory 47–8; Keynes
interpreted 96, 97, 98, 99, 106;
reductionism 46, 47–8
leisure: consumption 366, 378, 380–1,
393, 399, 413, 416, 428–9; and
labour 298; lag 378; non-market
activities 372–3, 375
Leontief, W. 65, 149, 275
Lerner, A.P. 61, 62, 166
Leslie, D. 454
Levhari, D. 618
Levine, R. 654
Lewis, W.A. 616
Lieberman, C. 219
life-cycle models, finite horizons 316–19
Lilien, D.M. 433
Lindbeck, A. 146, 454, 524, 525, 560
linearity 287–8, 291 (n1)
Lipsey, R. 7, 222, 224
liquidity constraints 139–40
liquidity preference 65, 183, 196, 197,
198, 218, 274
liquidity trap 143, 248
living standards, differences 18
loan markets, imperfect 319–21
Loasby, B.J. 38, 40, 42
Long, J.B. 361, 399, 412, 432
long-term policy 75
Lucas, R.E. Jr.: econometric policy
evaluation 12–14, 115–17; economic
aid programmes 625; economic
growth 20, 21, 369, 371, 373, 413,
629, 636, 639–40, 644, 649, 653;
employment 463; expectations 226,
287, 297, 301, 336, 337; fluctuations
8, 367; government policy 411;
inflation 12; on Keynesian economics
9, 159, 265, 347, 351, 396–7, 445,
450, 498–500; local/aggregate shocks
517, 530; Mankiw on 440, 471;
neoclassical orthodoxy 119; new
classical 336–7, 340, 343, 426, 434,
453–4, 457; output-inflation trade-off
481; Phillips curve 11, 114, 498; price
indexing 441; rational expectations
336, 397, 453; real business cycles 16,
19; surprise supply functions 12, 14;
technological change 619, 621, 623;
utility maximizing 400, 401, 417
Lucas-Sargent-Wallace analysis 233
McCallum, B.T. 257, 283, 295, 297, 362
Maccini, L.J. 558
McDonald, I. 523
Machlup, F. 166
macroeconomics 1–2, 470; citations
data 4–5; microeconomic
Index 669
foundations 456, 552–3, 569 (n3);
models 271–3, 274–6; policy 467–9;
textbooks 29–30, 43–4
Maddison, A. 583, 587, 628
Maddock, R. 265
Malinvaud, E. 309 (n4)
Mankiw, N.G. 454–70 (interview):
classical/Keynesian 1–2; coordination
failure 17; economic growth 19, 577,
578, 634, 654; efficiency wages 462–3;
labour supply 391; menu cost model
512, 513; monopolistic competition
125–6; new classical economics
458–9; new Keynesian economics
440, 441, 445–6, 452, 454– 6;
rational expectations 458–9; real
business cycles 16, 362, 465–6;
recession 468; taxation 322; wages
520
marginal cost behaviour 506–8, 509
marginal productivity theory 146, 147–8,
620–1
Marimon, R. 371
market-clearing: equilibrium business
cycle 280; equilibrium models 141,
284–5, 349; Keynesians 111; Lucas
12; neo-Austrians 227–8; new
classical economics 113, 118–19, 136,
265–6, 280–1, 336–7, 339, 349, 426,
459, 502, 503; see also markets, non-
clearing
markets 2, 3; aggregate demand 5, 188,
305; choice theory 36–7; competitive
29, 338–40; discoveries 638;
equilibrium 37, 39, 425–6; failure 2,
5, 120, 136–7, 225, 398, 455; fluidity
187–9; non-clearing 228–9, 471, 500,
502–3; spillovers 502–3
Marshall, A. 119, 526, 619, 639
Marx, Karl 68
mass production 597
Matthews, R.C.O. 6, 85 (n11), 161, 586
maximizing behaviour 297, 337, 338–9,
344–5, 398, 400
Mayer, T. 160, 161, 217, 238, 257
Meade, J.E. 61, 62, 73, 74, 161
Means, G. 490, 491
Medoff, J. 392
Mehra, R. 393
Meiselman, D. 217
Meltzer, A.: deficit financing 75–6; on
Keynes 31, 70–1, 75–6; monetary
growth 257–8, 259–60; monetary
policy 237; Patinkin on 71–2, 73, 75,
76; Phillips curve 225; quantity theory
182, 220, 337; rules/ discretion 72,
73, 76–7, 81–2; transmission
mechanism 186, 210 (n29)
Meltzer-Mayer rule 257–8, 259–60
Menger, C. 349
menu cost theory 511–14; input-output
519; literature 462; models 464–5,
512, 513, 521; price rigidity 563, 571
(n20); price stickiness 508–9
Merton, R. 57–8
migration 594–5
Milgate, M. 68, 79, 81, 97
Mill, John Stuart 173
Mills, F.C. 490
Mirman, L.J. 371, 402
Miron, J.A. 8
Mirrlees, J. 657 (n2)
Mishkin, F. 14
Mitchell, D.J. B. 492
Mitchell, W. 397
models: size 193–4, 212 (n54), 217;
testing 303–4
Modern Cambridge School 67–8, 69
Modigliani, F.: business cycle 286;
econometrics 217, 289; equilibrium
models 288; IS-LM model 63; life-
cycle 317, 318; monetarism
characteristics 247–50; monetary
growth 161, 254; profit-maximizing
502; public debt 314, 316
Modigliani-Miller theorem 329
Moggridge, D.E. 29, 72–3, 85 (n11)
monetarism 7–11, 159–61, 180, 204,
206; aggregate demand 187, 188–9;
allocative detail 188–9; balance of
payments 229–32; Brunner 159, 160,
181, 220, 237; exchange rates 229–32;
Federal Reserve 249–54; fiscal policy
232–8, 248; government intervention
202–3, 216, 305; inflation 201–2;
international 208 (n10); IS-LM model
220–1, 335–6, 352 (n5); Laidler
216–17; model size 193–4, 217;
Modigliani 247–50; Monetarism
249–54, 256–60; monetary growth
rule 198–200, 202, 204; and new
classical macroeconomics 334–7; and
New Keynesian economics 473–4;
670 Index
policy issues 232–8; price level 190–1,
199; quantity theory of money 181–2,
194, 335; shocks 224; transmission
process 182–7, 193
monetary authorities 175–8
monetary contractions: see recession
monetary growth 248–9; feedback 161;
GNP 250, 251, 255–6; new classical
economics 122; rules 198–200, 202,
204, 235–7, 256–60; targets 250–1,
252–3, 254
monetary indicators 194–6
monetary policy 130 (n4), 175–8, 232–8;
aggregate demand 110, 114, 467;
employment 169–73; Federal Reserve
System 164–6; Friedman 281, 485;
Gordon 14; inflation 175; interest
rates 166, 168–9, 560; Keynesian
view 448–9; lags 208 (n13), 213
(n65); limitations 161, 167–73, 427;
potential 173–5; procyclical 485;
rules/discretion 13; stability 174–5,
469; targets 196–8; US 164–7, 194
money: as consumer durable 218;
demand function 143, 183–4, 239,
248, 426, 558; importance 552; as
machine 173–4; money income 182–6,
197, 221–2; narrow 219; neutrality
300, 449, 510–11; non-neutral 462;
and output 14, 433–4; real business
cycle theory 414, 433–4; see also
demand for money
money supply 237–8; aggregate demand
281; inflation 9–10, 11, 221, 222;
price stability 8; shocks 552
money supply rule 257–8, 261 (n3), 304
monopolists 125–6, 506–9, 521, 523,
639
Montgomery, E. 354 (n22)
Mortensen, D.T. 522
mortgage interest rates 236
Mullineux, A.W. 363
multiplier theory 61, 139, 140
Mundell, R. 319
Murphy, K.M. 433
Muth, J. 11–12, 310 (n15), 312 (n46),
397, 453
Myrdal, G. 621
NAIRU 463
national accounts data 327, 654
national income 274
natural rate hypothesis 113–14, 129,
303–4, 447, 449, 486, 488; see also
unemployment
Neary, P. 563, 567
Neftci, S. 304
NEIO (new empirical industrial
organization) 491
Nelson, C.R. 15, 404, 414
Nelson, R.R. 579, 645 (n6)
neo-Austrians 227–8, 232, 236
neoclassical economic growth: and
endogenous growth 628; investment
614; model 19–20, 616–17; Solow
369; theory 577, 649, 651–2
neoclassical model 2, 6, 416–19; capital
accumulation 398–9, 402–3, 412;
choice theory 456; Lucas 119;
taxation 411–12
neo-Schumpterian growth 641–3
new classical economics 2, 265–7;
empirical evidence against 121–5;
equilibrium 349; equilibrium business
cycles 501–2; free parameters 340–4;
fundamentalism 119; inflation/
unemployment 334–40, 341–2;
information, imperfect 566–7; and
Keynesians 109–10; Lucas on 336–7,
340, 343, 426, 434, 453–4, 457;
Mankiw on 458–9; market-clearing
113, 118–19, 136, 265–6, 280–1,
336–7, 339, 349, 426, 502, 503; and
monetarism 334–7; money changes/
output 111, 122; money supply
shocks 552; policy implications 266;
rational expectations 11–14, 346–8,
349–50; recession 122, 342–3;
representative agents 553; tax
reductions 122–3; technological
sophistication 32
new Keynesian economics 2, 16–18, 460,
500–6, 540 (n2), 554–5, 568;
competition 491–2; coordination
failure 439–40; credit markets
559–60; cross-industry differences
492; employment 463; equilibrium
500–1; fiscal policy 469, 471–2;
Gordon 441, 478–80, 500–6;
government intervention 456;
inflation 256, 449; insiders/
outsiders 462–3; and
Keynesianism 450–1, 459–65, 471–3,
498; labour market 527–8, 531,
Index 671
560–2; Mankiw on 440, 441, 445–6,
452, 454–6; markets, non-clearing
502–3; and monetarism 473–4;
monopolistic competition 508–9;
NEIO 491; price stickiness 508;
rational expectations 567; real
business cycle 425, 466; rigidities, real/
nominal 462, 520–1, 563–5; risk
aversion 555–60; Romer, D. 445,
454, 456, 460; stabilization 439–40;
Stiglitz 17, 18, 439, 441–2, 472;
Taylor 17, 478; Tobin 441, 452–3,
464; unemployment 449; wages 466
Newbery, D. 561
newly industrialising countries 591, 609,
610, 614
Newlyn, W.T. 84 (n8)
Newsweek 351
Nguyen, D.-T. 607, 609
Nickell, S. 525
Nishimura, K. 371
Nobel Prize for Economics 577
Nordhaus, W.D. 639, 642
O’Donnell, R. 97, 98, 99, 100
O’Driscoll, G. 330 (n2)
OECD countries, catch-up 607–9
Ohkawa, K. 584
Okun, A. 147, 344, 492, 511, 516, 522
Olson, M. 344, 585, 591, 598
OPEC oil shocks 431, 453, 468–9, 517,
557
optimizing behaviour 266–7, 280–1,
282, 291 (n6)
output: actual/trend 368; adjustment
513; capital/labour shares 373–4,
649, 651; and employment 45, 336,
366; fluctuations 15, 16, 366, 385;
investment 649; long-run movement
6; money 434; per capita 404, 609–10,
630–1; and prices 232, 298–300; real
404, 408; Solow residual 430
Pack, H. 20
paradigms 109
parameter shifts 340–4, 390–1, 534–5
Park, Y.C. 183
Parkin, M. 453, 499, 540 (n2)
Pasinetti, L. 68
Patinkin, D.: classical dichotomy 426;
effective demand 58, 60; General
Theory 30–1, 58, 60, 63–4, 68, 78;
interpretation 103; markets, non-
clearing 502; on Meltzer 71–2, 73, 75,
76; money neutrality 221; Pigou effect
148–9; quantity theory of money
182, 218
Pen, J. 43
Pesando, J. 303
Phelps, E.S.: economic schools 1;
equilibrium model 285; inflation/
unemployment 222, 224, 335, 337;
modern market theory 440; natural
rate hypothesis 114; price stickiness
344, 478; rigidities 454; staggered
contracts 514; structuralism 456;
wages 9, 17, 526
Phelps-Winter theory 516, 522
Philippines, output per worker 630–1
Phillips, A.W. 84 (n8), 200, 222, 485
Phillips curve 7, 434, 498; expectations-
augmented 159–60, 161, 216, 222–4,
230, 336; failures 114–15; inflation-
unemployment trade off 200–1, 204,
335; price inertia 485–7; rational
expectations 225–6; supply shocks
123; vertical 10–11, 113–14, 233,
296, 396–7, 449; wages 178 (n5)
Pieper, P. 324
Pigou, A.C. 4, 201
Pigou effect 6, 33, 148–9, 152
Plosser, C. L: equilibrium cycle 141;
interest rates 325; money/output 434;
real business cycle 16, 361, 362, 399,
412, 414, 432; shocks 15, 404
policy: activist 233–4; aggregate demand
467; credibility 348, 350, 354 (n19);
economic growth 611–14; Keynesian
278; price stickiness 481–5; rational
expectations 232; rules/discretion
449–50
politics: influence 203; and
macroeconomics 78, 81, 120–1, 236,
467–9
Poole, W. 261 (n3)
population growth 604–5, 606, 610,
611, 612–13, 616
portfolio theory 556, 565–6
post-Keynesian school 2, 67–70, 461
Prescott, E.C.: calibration 363, 415;
equilibrium 361–2, 366, 371, 400;
leisure 413; output/employment 366;
prices 392–3; rational
expectations 11; real business
cycle 16, 361–3, 389–94; rules/
672 Index
discretion 13, 471–2; stability 120;
technological disturbances 429–30;
utility maximizing 417
Price, S. 1
price adjustment 486–7; barriers 508;
Blanchard 486, 514, 519; costs 517,
564; cross-country comparisons 495;
cuts 126; gradual 499–500; inertia/
rate-of-exchange/level effects 529–30;
menu cost 513; rational expectations
151–2; rises 567; speed 228; stability
150
price indexation 505, 518–19
price level: aggregate demand 148, 150–2,
191, 298–300, 336, 496–7; effects
479, 494; and individual prices
189–92; and quantity of money 72,
176; sociological view 223
price mechanism 29, 38, 350–1, 453
prices: administered price hypothesis
490; aggregate demand 191; and
consumption 6; controls 234–5;
cyclical responsiveness 490, 511;
detrending 534–5; employment 286;
equilibrium 50; failure 50; FIFO 511;
flexibility 18, 145, 152, 154 (n5),
336, 553-4; fluctuation 342; General
Theory 58, 67, 191; GNP 481–2,
538–9; hysteresis effect 488–90;
industrial diversity 490–2; inertia
127–8, 147, 479, 485–7, 494, 500,
524; input-output approach 480; as
money prices 38, 183; nominal
144–5, 147; and output 298–300,
232; persistence 531; Prescott 392–3;
rates of change 150–2, 479, 494; real
business cycle 392–3; reductionism
44; regression results 533–4, 535;
replacement cost 511; rigidity 394,
464–5, 470, 491, 506–15, 530, 553,
554, 563–5, 571 (n20); Romer, D.
441, 462, 513, 514, 515, 517, 528,
553; S, s state-dependent rules 509–11;
staggered 514, 515, 517, 519, 520,
528; sticky 17, 113, 145, 336, 344–6,
354 (n20), 454, 462, 478–9, 481–90,
501–2, 508–9, 521, 530, 565; Taylor
344, 481; time/space 492–6; wages
67, 79–81, 85 (n13), 144–9, 280,
344–5, 520, 563; see also price
adjustment; price level
principal-agent theory 556
private sector, stability 187, 199, 222
product cycles 594, 624–5
product differentiation 492
product market 146–7, 479, 505–20
production 68, 384, 416–17, 556, 558;
household 373–3, 375, 384, 390–1;
mass 597
productivity: agricultural/ industrial 612;
competition 599; cross-country
comparisons 596, 606; disturbances
400–1; growth over time 404–5,
407–8, 582, 587, 588–91, 620;
investment 620; labour 141, 583,
587, 588, 650, 651; leadership 591–2;
learning by doing 656; marginal 146,
147–8, 620–1; multi-factor 608, 615
(n2); pro-cyclical 466; real wage
525–7; resource-rich countries 593;
total factor 392, 429–30, 577; worker
525
productivity ethics 68
profit-maximizing 502, 508
protectionism 594, 624
public choice analysis 348
purchasing power parity 518, 588
Purvis, D.D. 160
quantity of money 180, 181–2, 216,
217–22; Friedman 7–8, 159–60, 182,
204, 216; inflation 337; monetarism
181–2, 194, 335; price level 72, 176;
total spending 165, 168–9
Quarterly Journal of Economics 1936
symposium 61
Radcliffe Report 240 (n5, n11), 242
(n30)
Ranis, G. 616
rational expectations 11–14, 296–8,
304–6, 346–8; agents’ maximizing
266–7, 281, 339, 349–50; arbitrage
301–2; Blinder 113; econometrics
117, 119, 124, 282; employment 298;
equilibrium 346–7, 348; errors 300–1,
302, 303; Fischer 113, 455; Gordon
306, 454; impotence result 298–9,
309, 311 (n42); Lucas 336, 397, 453;
McCallum 295, 297; Mankiw 458–9;
Muth 397, 453; new classical
economics 11–14, 346–8, 349–50;
Index 673
new Keynesian economics 567;
persistence 285–7; Phillips curve
225–6; policy 232; price adjustment
151–2; Sargent 303, 304; Taylor 12,
113; testing 303–4; unemployment
13–14
real balance effect 148–9, 563
real business cycle theory 2, 396–405,
566; classical dichotomy 426–7;
empirical assessments 414–15;
employment fluctuations 431–2, 433;
exchange mechanism 393–4;
fluctuations 14–16, 393–4, 425,
427–8; goods/leisure substitution
428; government intervention 410–12;
labour market 412–13; Mankiw 16,
362, 465–6; model 14–19; money
growth slowing 434; multiple sectors
432–3; new Keynesian perspective
425, 466; parameters 390–1; Prescott
16, 361–3, 389–94; prices 392–3;
recession 431, 433, 434; research
agenda 412–15; shocks 391–2, 397,
552, 571–2 (n23); spending 411;
taxation 411; Tobin 15; US 1954–85
405–10
Rebelo, S. 413, 622, 623
recession: deflation 14; disinflation 17;
firms 557; Keynesians 111–12; labour
supply 561; Mankiw 468; new
classical economics 122, 342–3; prices
482; real business cycle 431, 433,
434; social costs 512
Reddaway, W.B. 61, 62
reductionism 30, 36–7, 38–9, 44, 46–51
Reid, M. 218
Reinelt, D. 654
representative agent 399–402, 405,
508–9, 553, 569 (n1), 652, 653
research and development 578, 620–3,
640–1, 655, 656–7
resource allocation 236
resource constraints 417
returns to capital, constant 653–4
returns to scale 651, 653
Ricardian equivalence theorem 13, 266,
330 (n2)
Ricardo, David: budget deficits 315–16,
325–9; criticised 316–25; interest
rates 325–6; interpreted 104–5
Ricoeur, P. 103–4
rigidity: nominal 462, 506, 519, 527;
prices 394, 464–5, 470, 491, 506–15,
530, 553, 554, 563–5, 571 (n20);
product market 505; real 454, 462,
515–17, 519, 521–3, 530, 563–5;
wages 138, 455, 470, 487–8, 519–20,
522, 523
risk aversion 473, 555–60, 563
Roberts, J.M. 514
Robertson, D.H. 61, 65, 474
Robinson, A. 78
Robinson, Joan: equilibrium 39;
heterogeneity of capital goods 40;
imperfect competition 461;
Introduction to the Theory of
Employment 61; IS-LM 69, 70; on
Keynes 42, 81–2, 96, 97; Modern
Cambridge School 67; neo-Ricardians
38
Robson, M. 620, 622
Rogerson, R.D. 379, 413
Romer, C. 3, 436 (n5)
Romer, D.: coordination failure 17, 439;
economic growth model 19, 634, 639,
640, 642, 644, 649, 655; fluctuations
18; inflation 436 (n5); new Keynesian
economics 445, 454, 456, 460; prices
441, 462, 513, 553; staggered prices
514, 515, 517, 528
Romer, P.M.: cross-country convergence
629; economic growth 19, 20–1, 579;
endogenous growth 413, 633, 635;
government policy 411; knowledge
spillovers 632, 635; productivity
620–1; technological change 621,
622; world trade 614 (n1)
Rosenberg, N. 330 (n2), 602 (n19)
Rosovsky, H. 584
Ross, S.A. 556
Rostow, W.W. 616
Rotemberg, J. 521
Royal Economic Society 55, 160–1
rules/discretion 13, 72, 73, 76–7, 81–2,
449–50, 468, 471–2, 514–15
Rush, M. 500
Ryan, C. 363
Rymes, T.K. 78
Sachs, J.D. 492
Saffra, P. 67
Sahasakul, C. 324
Sala-i-Martin, X. 2, 18, 19, 633, 634
Samuelson, P.A. 6, 7, 30, 43, 66
674 Index
Santamero, A.M. 7
Sargent, T.J.: on Keynesian economics
114, 283, 347, 453, 4 98–500; atural
rate hypothesis 303, 486; rational
agents 12–13, 297, 337; rational
expectations 303, 304
savings: budget deficit 317–18, 326–7;
consumption 326–7; investment
447–8; private 123, 316, 317; rates
650; ratios 616
Say’s Law 142
Schleiermacher, F.E.D. 102
Schlesinger, J.R. 64
Schultz, T.W. 623
Schultze, C. 492, 493
Schumpeter, J.A. 97, 102–3, 638, 655–6
Schwartz, A. 3, 8, 159, 217, 219, 433,
485
search theory 208 (n12), 287, 521–3
Seater, J.J. 7
Shackle, G.L.S.: fundamentalist
Keynesianism 38, 42; general
equilibrium 39–40; General Theory
50, 65, 67, 79, 96; interest 64; IS-LM
model 79
Shapiro, C. 454, 526
Shaw, E.S. 79
Shaw, G.K. 21, 29, 578–9
Shaw, K. 354 (n22)
Sheffrin, S. 441
Shell, K. 638, 657 (n2)
Sheshinski, E. 510, 618
Shiller, R. 302, 311 (n31)
Shleifer, A. 318
shocks: demand 15, 143, 402, 517;
monetarism 224; random exogenous
339; real 339, 361; real business cycle
theory 391–2, 397, 552, 571–2 (n23);
risk aversion 558; stochastic 289;
supply 14, 16, 19, 114, 115, 117,
122, 123, 361, 401–2, 499, 505–6;
unpredicted 404
Siegel, J. 324
Sijben, J.J. 310 (n15)
Simons, H. 72, 178 (n1), 281
Sims, C. 275, 282
Sinai, A. 394
Singleton, K.J. 374, 390, 414
Skidelsky, R. 29
Skinner, Q. 57, 72
Slutsky, E. 288, 420 (n4)
Slutzky, E. 367
Smith, Adam 29
Smith, R. 492
Smithies, A. 167
Smulders, S. 20
Snowdon, B.: economic schools 1, 439,
440, 473; employment 7, 160; on
Friedman 10; government
intervention 29; monetarism 159;
technical sophistication 32
Snower, D.J. 146, 454, 524, 525
social capability 584–6, 597–9, 600
social costs 512–13
social welfare 126, 130
sociology of economics 117–18
Solow, R.M.: economic growth analysis
579; General Theory 4, 5; labour
market 464; monopoly 523, 639;
Phillips curve 7; technological change
375–6, 403, 405, 407, 413
Solow growth model 14, 15, 19–20, 399,
447, 577, 616–18, 638
Solow residual 430–1, 435 (n2), 577
Sonnenschein, H. 279
spillovers: Barro-Grossman model 521;
endogenous growth 639, 642;
knowledge 618, 622, 632, 639;
markets 502–3; technological 604,
612
Spulber, D.F. 510
stability: income 199; monetary policy
175; money demand 183–4, 218–20;
Prescott 120; private sector 187, 199
stabilization: expectations 153;
government 295–6, 302, 304;
monetary/fiscal policy 113–14, 248,
255–6, 306, 363, 385; new Keynesian
economics 439–40; by price
mechanism 453
Stadler, G.W. 16
stagflation 296
staggered contracts 147, 514–15, 527–8
stagnation 19, 141
steady state values 400, 402
Stern, N.H. 579, 621, 623
stickiness: prices 17, 113, 145, 336,
344–6, 354 (n20), 454, 462, 479,
481–90, 501–2, 508–9, 521, 530, 565;
wages 17, 137–8, 147, 279, 344–5, 354
(n22), 454, 480–90, 501–2, 527
Stigler, G. 99–100, 105, 490, 491
Stigler’s Law 330 (n2), 645 (n5)
Stiglitz, J.E.: aggregate supply/
demand 474; economic growth
model 642, 653; equity 556; market
imperfections 553; new
Index 675
Keynesianism 17, 18, 439, 441–2,
472; real balance effect 563; rigidities
454; risk aversion 473, 559, 560;
technological changes 562; wages/
prices 526, 561, 567
stochastic models 402, 404
Stokey, N. 655
structuralism 456
substitution: capital/labour 652;
elasticities 366–7, 373
Summers, L.H. 16, 318, 326, 362, 469,
481, 525, 526
Summers, R. 605, 628, 654
Sumner, M.T. 236
supply and demand dichotomy 480–1;
see also aggregate demand; aggregate
supply
Swan, T.W. 19, 369, 616
Tarshis, L. 488
Taussig. F.W. 65
taxation: budget deficits 315–16, 320,
323; finite horizons 316–17; future
321–2; income 322, 620; lump-sum
322; neoclassical model 411–12;
policy 579, 622–3; progressive 300;
real business cycle theory 411;
reduction 122–3; smoothing 323–4;
stabilization 13; timing 322–4
Taylor, J.B.: market-clearing 285; new
Keynesianism 17, 478; prices 344,
481; rational expectations 12, 113;
staggered contracts 147, 514, 527–8;
wages 454, 492
technology: catch-up 607–9, 615 (n3);
change 375–6, 403, 405, 407, 413,
579, 616, 618–22, 637–8, 651–2,
655, 657 (n2); disturbances 428–31;
economic growth 403, 404, 405, 407;
equlibrium business cycle 141, 142;
flows 623; frontiers 594; new classical
economics 32; as non-rival input
621–2; and productivity growth 582,
583–4, 590–1; as public good 20; and
social capability 584–6; spillovers
604, 612
technology gap 586, 596, 633
technology shocks 362, 381–3, 391–2,
402, 552, 566, 572 (n23)
term structure theory 184
theories, as concept 124, 434–5
Third World economies: see developing
countries
time allocation 298
Tobin, J.: business cycle 14, 374; effective
demand 33, 136, 138–40;
employment 286, 449; on Friedman
9, 160; government intervention 29;
on Keynes 6, 32–3, 275; and Mankiw
471; monetarism 161, 181; money
demand 151, 218, 502; new
Keynesianism 441, 452–3, 464; real
business cycles 15; Ricardian
equivalence 317; risk 565–6
Topel, R.H. 433
total factor productivity 392, 577
Townsend, H. 37–8
trade policy 593–4, 623–5
trade unions 480, 523–4
transaction costs, fixed 128
transmission mechanism, monetarism
182–7, 193, 210 (n29), 217, 225
transnational corporations 604, 623
trend 6, 14–15, 367–8, 403
Turnovsky, S. 303
UK: see Britain
uncertainty: competitive equilibrium 371;
employment 40; future taxes 321–2;
Keynes 65–6; stochastic models 402
underemployment 149
unemployment: aggregate demand 17;
business cycle 226; classical 138, 139;
equilibrium 6–7, 17; European 463,
488, 525; frictional 138; Friedman
9–10, 12, 224–5; General Theory 68;
GNP 301; government intervention 9;
hiring/training costs 559; human
capital 17–18, 129, 412; hysteresis
theories 11, 17–18; inflation 9–10,
112, 113, 115–16, 172, 200–1, 222,
230, 277, 304–5, 335, 337, 341,
396–7, 449; insiders/outsiders 146;
involuntary 5–6, 18, 111–12, 126–7,
130, 141, 138, 453, 455, 463, 561;
Keynes 5–6, 42, 139; market 171–2;
natural rate hypothesis 9, 17, 129,
159–60, 170, 171, 178 (n3), 304,
309 (n4), 447, 449, 463, 486, 488,
522; new Keynesian economics 449;
persistence 524–5; prewar figures 3;
rational expectations 13–14; search
theory 287; structural 17–18; US 60;
voluntary 139, 305, 307, 522;
wages 7, 59–60, 138–9; workers’
mobility 433
676 Index
US: Bretton Woods 10, 160, 229, 231;
budget deficits 327; capital/labour
shares of output 373–4; capital stock
596–7; consumption and leisure 366;
cyclical economy 370; decline, price/
production 490; economic growth 18,
618; Employment Act (1946) 450;
exchange rates 176; Federal Reserve 3,
8; GNP 369, 402–3, 497, 537; Great
Depression 3, 5, 8, 19, 149, 164,
165–6, 173, 450, 490, 563; income
by state 629, 634; incomes increase
597–8; inflation 165, 497; mass
production 597; monetary policy
164–7, 194; money demand shifts
219–20; pre-/post-war 492, 493, 535,
537; as productivity leader 582, 591,
592–3, 601 (n11); real business cycles
405–10; responses across time/ space
492–6; summary statistics 406–7; tax
reductions 122–3; wages 60, 171,
373–4, 536, 538; unemployment 60
US Treasury Department 325
utility maximizing 398, 400, 401, 417
Uzawa, H. 413, 621, 623
value 68
Van de Klundert, T. 20
van Els, P.J. A. 16
Vane, H.R. 10, 32, 160, 440
variables: exogenous/endogenous 274,
276, 282; real/nominal 426
Veblen, T. 585, 596
Vercelli, A. 1
Verdoorn effect, technology 594
Vernon, R. 594
Viner, J. 65
Volcker, P. 433–4, 459, 554
wage contracts 147, 514, 527–9
wages: adjustment 564; controls 234–5;
efficiency models 18, 126–7, 462–3,
525–7, 531–2, 561, 571 (n16);
employment 45, 560–1; high 570
(n10); increased flexibility 553–4;
indexation 505; inertia effect 494;
inflation 9, 235; Keynes 147–8, 487–8,
572 (n26); labour markets 479; level
effect 494; money 224, 227–8; new
Keynesian economics 466; nominal
59–60, 170–2, 178 (n4), 299; by
performance 527; Phelps 9, 17, 526;
piece rates 527; prices 67, 79–81, 85
(n13), 144–9, 280, 344–5, 520, 563;
and quit rates 526; rate-of-change
effect 494; real 145, 170, 201, 224,
299, 410, 429, 466, 488; rigidity 138,
455, 470, 487–8, 519–20, 522, 523;
stability 572 (n26); sticky 17, 137–8,
147, 279, 344–5, 354 (n22), 454,
480–90, 501–2, 527; Stiglitz 526,
561, 567; Taylor 454, 492; time and
space 492–6; unemployment 7, 59–60,
138–9; unions 129
Wagner, R. 321
Wallace, N. 11, 12–13, 337
Walras, L.: as auctioneer 50, 146, 154
(n5); equilibrium 141, 170, 343,
425–7, 429; expenditure 210 (n31);
flexibility of wages/prices 142; general
equilibrium 343; market-clearing 12,
141; price changes 152; tatonnement
224
wealth, actual/perceived 324
wealth distribution 149
Weil, D. 634
Weintraub, S. 64–5, 66–7, 79–81, 85
(n13), 88 (n31)
Weiss, A. 560
Weiss, Y. 510
welfare: competitive 371; economic
growth 18; Keynesianism 125;
reduced 16
welfare state, growth inhibiting 598–9
Wickens, M. 363
Wicksell, K. 9, 170, 171, 198, 452
Wilcox, J.A. 3
Williams, J.H. 166
Winter, S.G. 645 (n6)
Winter, S.G. Jr. 516
workers, performance pay 527
world trade 604, 614 (n1)
Wulwick, N. 7
Yellen, J.L. 18, 125–6, 454, 455, 464,
512, 563
Yotsuzuka, T. 321
Young, A. 655, 656
Young, A.A. 619
Young, W. 6, 31, 62
Zeldes, S. 322
Zervos, S. 654
Book Cover
Title
Copyright
Contents
Preface
1 The development of modern macroeconomics
Part I Keynesian economics and the Keynesian revolution
Introduction
2 Keynesian economics
3 On different interpretations of the General Theory
4 Keynes’s General Theory
5 The fall and rise of Keynesian economics
6 Price flexibility and output stability
Part II The monetarist counterrevolution
Introduction
7 The role of monetary policy
8 The structure of monetarism
9 Monetarism
10 The monetarist controversy revisited
Part III The challenge of rational expectations and new classical macroeconomics
Introduction
11 After Keynesian macroeconomics
12 A child’s guide to rational expectations
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=oaef20
Cogent Economics & Finance
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/oaef20
The relative choice over destiny in a country’s
long-run economic growth and economic affluence
Esubalew Tadele & Teshome Sirany |
To cite this article: Esubalew Tadele & Teshome Sirany | (2021) The relative choice over destiny
in a country’s long-run economic growth and economic affluence, Cogent Economics & Finance,
9:1, 1949133, DOI: 10.1080/23322039.2021.1949133
To link to this article: https://doi.org/10.1080/23322039.2021.1949133
© 2021 The Author(s). This open access
article is distributed under a Creative
Commons Attribution (CC-BY) 4.0 license.
Published online: 19 Jul 2021.
Submit your article to this journal
Article views: 1122
View related articles
View Crossmark data
https://www.tandfonline.com/action/journalInformation?journalCode=oaef20
https://www.tandfonline.com/loi/oaef20
https://www.tandfonline.com/action/showCitFormats?doi=10.1080/23322039.2021.1949133
https://doi.org/10.1080/23322039.2021.1949133
https://www.tandfonline.com/action/authorSubmission?journalCode=oaef20&show=instructions
https://www.tandfonline.com/action/authorSubmission?journalCode=oaef20&show=instructions
https://www.tandfonline.com/doi/mlt/10.1080/23322039.2021.1949133
https://www.tandfonline.com/doi/mlt/10.1080/23322039.2021.1949133
http://crossmark.crossref.org/dialog/?doi=10.1080/23322039.2021.1949133&domain=pdf&date_stamp=2021-07-19
http://crossmark.crossref.org/dialog/?doi=10.1080/23322039.2021.1949133&domain=pdf&date_stamp=2021-07-19
GENERAL & APPLIED ECONOMICS | REVIEW ARTICLE
The relative choice over destiny in a country’s
long-run economic growth and economic
affluence
Esubalew Tadele1* and Teshome Sirany2
Abstract: This paper examines an in-depth and systematic review of why some
nations are so rich, while others remain so poor taking into account temporal and
spatial dynamics applied for economic growth covariants. Growth literature under-
scores direct and indirect causes for economic growth. Likewise, economic and non-
economic dynamics are thoroughly examined for countries’ long-run economic
growth and relative wealth accumulation. Endogenous growth theories emphasized
that investment in human capital, innovation, and knowledge are major contributors
to economic growth. Empirics confirmed that time-variant (such as well-established
institutions and their prominent role in devising property rights and policies) gives
Esubalew Tadele
ABOUT THE AUTHOR
Esubalew Tadele is a lecturer and researcher in
the Department of Agricultural Economics, Debre
Markos University. He did Master’s degree in
Development studies, Major in Economics of
Development at International Institute of Social
Studies (ISS), The Netherlands. He has taught
various Economics courses at Jimma University
and Debre Markos University, Ethiopia since
September 2009 (Such as Development
Economics, Institutional economics, Natural
Resource and Environmental Economics,
Microeconomics and Macroeconomics,
Econometrics, statistics, statistics for Economics).
His area of research interest is the economics of
poor countries, Econometric analysis of develop-
ment policies, privatization and investment,
regional development and geopolitics and regio-
nal integration of countries, development and
environment, agriculture and development.
PUBLIC INTEREST STATEMENT
Searching the „black box „ of historical roots of
development on one side and the main center of
gravity for the sluggish economic outcomes of
poor economies on the other side is not such an
easy task until recently. It requires decisive factors
before merely discredited the older explanations of
underdevelopment such as poor resource endow-
ments, or lack of potential markets or, the absence
of economic rationality. However, one who travels
across the globe from West to East; North to South
can simply observe facts on the ground that
diverse tracks of economic progress of nations.
Some countries are gifted with plenty of natural
resources, mountains, oceans, plain lands, diverse
agroecology but failed to realize and exploit eco-
nomic potential while others lack many of its fea-
tures but achieved miracle growth (Example:
Japan, and S. Korea). Some countries seem to have
their geography imprisoned them and stay in
poverty and in conflict trap for longer periods of
time and failed to build a strong economy and a lot
of messes and unstructured system persists (like
many of Sub-Saharan African and Asian countries).
For thus, economics, non-economics issues, direct
and indirect causes, resource and administrative
issues, historical events, and geographic reasons
are considered for such diverse track experiences
of nations and thoroughly examined what to do
with it in this article. Following that, identifying the
sharp insight into the ways how countries deter-
mine their fate in the ever-complex, chaotic, and
interlinked world with its destiny.
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 1 of 17
Received: 20 November 2020
Accepted: 24 June 2021
*Corresponding author: Esubalew
Tadele, Department of Agricultural
Economics, Debre Markos University,
Debre Markos, Ethiopia
E-mail: esu360@gmail.com; esuba-
lewtadele@dmu.edu.et
Reviewing editor:
Christian Nsiah, School of Business,
Baldwin Wallace University, United
States
Additional information is available at
the end of the article
© 2021 The Author(s). This open access article is distributed under a Creative Commons
Attribution (CC-BY) 4.0 license.
http://crossmark.crossref.org/dialog/?doi=10.1080/23322039.2021.1949133&domain=pdf
http://creativecommons.org/licenses/by/4.0/
more emphasis towards shaping a country’s long-run economic growth than time-
invariant exogenous attributes (like geography). Meanwhile, some nations did not
industrialize being geographically advantageous, and the location of the country
does not exclusively determine the fate of the nation’s economic success. Moreover,
state capacity is vital to determining relative wealth accumulation and economic
prosperity. The incidence of routine war undermines the fiscal capacity and leads to
an extractive form of government and weakens public and private investments, and
this sets a country into undesirable outcomes. In a nutshell, time-varying attributes
become more flexible to adjust the fate of countries’ economic growth and destiny.
Furthermore, it requires an intense investigation of what governs a nation’s economic
successes or failures focusing on country-specific concerns. It needs a close and
continuous rectification to reconsider the country’s institutional setup and policy
frameworks towards endogenously rooted economic growth and development for
the relative economic affluence.
Subjects: Development Policy; Economics and Development; Economics
Keywords: economic growth; time-variant factors; institutions; natural resources;
geography; endogenous growth theories
1. Introduction
Economic growth is vital for the state and welfare of society in the first place. Growth is also the
key indicator of the successful government and pinnacle for the national wellbeing in general. It
can be a guarantee of the country’s autonomy. Likewise, stable economic growth is an indicator of
a country’s robust economic system and, thus, its society is safely protected from potential
economic risks and other external economic threats (Wajeetongratan, 2020).
In academic researches, the realization of various correlated factors of economic growth and its
dynamic progress has always been taken among the top priorities (Wajeetongratana, 2020; Curzio
et al., 1994; Zahka, 1990). Analysis of economic growth, its stages, cycles, required resources for
growth, and related policies underpin scholars to continuously engage in the system and investi-
gate the central causes of it over a long period.
Similarly, the primary focus of many developing countries is to have high and sustainable growth.
Nonetheless, to achieve and keep up a high growth rate, there has to be comprehending the
determinants of growth as well as how various policies affect growth. Since World War II, the pattern
of growth of real GDP has become a key policy strategy for all nations practically (see Crafts, 2000).
Several investigations have been carried out to find the path of long-run economic growth. The
earliest studies were suggested by Solow (1956) and Swan (1956) based on the neoclassical theory.
Its simple structure and presumptions—a well-behaved neoclassical production function, a single
homogenous good, exogenous labor-augmenting technical progress, full employment, and exogen-
ous labor force growth—have been used by economists for the past four decades.
One of the fundamental issues raised with long-run economic paths of countries, let’s say prior
to 1800, the living standards of world economies were roughly constant over time: per capita
wage, income, output, and consumption did not grow (Hansen & Prescott, 2002). However, modern
industrial economies, on the other hand, enjoy unprecedented and seemingly endless growth in
nations’ living standards, and hence countries experience diverse growth performances.
Even though it has been realized that economic growth is a continuous and dynamic process,
some countries built a strong economy in a long-lasting and sustained manner and were able to
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 2 of 17
enjoy a high standard of living, while others failed to build a strong economy and were unable to
compete with the advanced nations and even stay in a place of struggling their daily life. Several
studies investigate causes that enhance or hinder economic growth, and it has been placed one of
the fundamental queries for theoretical and empirical growth researchers, but little consensus has
been reached to date (Chirwa & Odhiambo, 2016).
This review article tries to comprehend various perspectives that make countries around the
globe achieve and pursue such a diverse track of economic experiences and why countries perform
as such differently? Thus, it states the identification of principal causes or the missing links for the
long-run economic growth of countries. How some countries achieve a high standard of living and
why others lagged and struggle in daily life with a lot of messes and unstructured systems. Do
institutions matter for everything? Does the existence of abundant natural resources have a good
or bad outcome for long-run economic outcomes? Does the location of the country or being
landlocked determines the fate of the country exclusively? Is low economic trap due to by
misfortune? or does it because of bad government? or its people’s sociocultural setting?
Hence, this review mainly circles to identify the key correlate factors for the existing economic
divergent experience of countries looking at a time changing and unchanging features for various
economic progress.
1.1. The objective of this review
General objective
● To review what determines the relative choices over destiny for long-run economic growth and
economic affluence of countries.
Specifically,
● To review main covariates of long-run economic growth of countries in spatial and temporal
dimensions.
● To review both time-variant and time-invariant dynamics of economic growth parameters.
● To review whether geography (location of a country) matters for long-run economic growth
and economic affluence or not.
2. Review methodology
For this article, a comprehensive literature review is carried out based on theories and empirical
findings, and extracted saturated information. It has been used both temporal and spatial dimen-
sions and able to filter information including the recent works that reflect country-wide verdicts of
economic growth parameters.
2.1. Search engines used
This literature searches mainly emphasized economic growth and significant contributing factors.
The search engine sets in the information offered by Google scholar, and archives of qualified
publishers of academic databases like the Web of Science and Scopus that are mainly focused on
peer-reviewed journal articles, books, reports, and special issues. The search has been filtered
using keywords such as “Economic growth”, ‘time-variant factors for economic growth, “economic
and noneconomic growth factors”, “endogenous and exogenous”, “Technology and economic
growth”, ”institutions and economic growth”, “geography and economic growth”. More than 75
journal articles, books, proceedings, and thesis works, agencies reports related to the topic were
browsed and around 40 materials were prioritized specific to the topic used for this investigation.
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 3 of 17
3. Basics of economic growth theories
Economic growth is a complicated process; however, the main theories of economic growth are
reasonably basic and conceptually simple. The traditional growth model was advanced by Robert
Solow (1956), a Nobel Prize winner, perhaps the most famous one. The basic foundation of this
framework is that growth is caused by autonomous technological change and capital accumula-
tion. Solow views the world as one in which output, Y, is generated by the production function
(Gould & Ruffin, 1993).
There are two basic classifications of economic growth theories—those grounded on the traditional
Solow (1956) growth model and those based on the concept of endogenous growth. Solow model
emphasizes capital accumulation and exogenous population change and technologically innovative
progress. This model predicts that all market-based economies will eventually arrive at a similar pace
of growth rate if they have the same rate of technological progress and population growth. Moreover,
the model expects that the long-run rate of growth is out of the reach of policymakers.
Solow verifies two factors into the model (see equation a): technological and labor growth.
Specifically, at time t, output Yt is determined by two inputs (capital and labor) as well as technolo-
gical progress, and using the Cobb–Douglas function, the model is specified in the following ways.
YtKαt AtLtð Þ
1� α (a)
The variable At denotes technology level at time t. Both labor input Lt and technology level At
are expected to grow at constant rates. This design of technology is referred to as labor-
augmenting and labor become more “effective” so that the effective labor input equals A(t) L(t).
This evolution leads to the central equation of the Solow model involves how the capital per
effective labor k(t) arises as explained in the following equation.
Δk ¼ sfðkÞ � ðþn þ gÞk (b)
In the Solow model, the change in the capital stock Δkequals investment sfðkÞ minus breakeven
investment (δ þ n þ gÞk. Now, however, becausek ¼ K= LxEð Þ, break-even investment consists of
three terms: to keep k constant, δkis needed to replace depreciating capital, nk is needed to
provide capital for the new workers, and gk is the amount of capital needed for the new “effective
workers” formed by technological progress. Here, s is the rate of savings, the fraction of total
output Y(t) saved for investment, δ is capital depreciation rate, the fraction of capital stock K(t)
that turn into obsolete, n is the growth rates of labor L(t), and g is the growth technology A(t). All of
the four parameters s, n, g, δ plus α are exogenous, its values are not deliberately chosen by its
economic agents but rather determined by factors outside of the model. This equation has a very
intuitive explanation: the rate of change in capital per unit of effective labor, Δk, equals the
amount of saving per unit of effective labor minus the dilution of capital stock per unit of effective
labor due to population growth, technology growth, and depreciation (Mankiw, 2010; Zhao, 2019).
Hence, Solow model states that if countries have equal savings rates, population growth, technical
progress, and depreciation rates, then regardless of their initial per capita outputs, they will converge
to a similar balanced-growth path and become the same per capita income in the long run. It
expresses that poor nations should grow faster and catches up with rich nations rather quickly.
To set and see the Romer model, it is essential to realize the Solow model that sustained growth
is cannot achieve with the accumulation of capital alone because of the diminishing returns to
capital: the extra output produced by new additions to capital stock will fall to nil in the long run so
growth must stop when capital stock is sufficiently large. In the long run, to generate persistent
growth, it must be assumed that the continuous rise of effective labor, so the marginal product of
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 4 of 17
capital (MPK) can stay above zero. Without technological change, population growth is the only
way to achieve growth, but then per capita income will stay constant in the long run. This shows
why technology is needed for output per capita to the long-run economic growth.
Moreover, the convergence prediction is not supported by data: there are persistent and large
differences in the rate of growth and per capita income levels among countries. One can argue that
countries differ in technological changes (different countries have different g values) and hence they
do not necessarily converge to the same growth path, at least not unconditionally. The fundamental
challenge in this perception is that there is no explanation why countries differ in technology because
changes in technology are exogenously given in the Solow model (Zhao, 2019).
The ongoing expansion of endogenous growth models started with the work of Paul Romer
(1990), he saw that traditional hypothesis neglected to accommodate its forecasts with the exact
perceptions that, as time goes on, nations seem to have accelerated growth rates and, among
nations, growth rates differ substantially. Romer needs to find a way to make the technology
parameter A(t) come out of decision-making by for-profit firms instead of exogenously given as
explained in the Solow model. He needs to explicitly model the R&D process. Thus, Romer unlike
other neoclassical economists emphasized that economic growth is an endogenous outcome of an
economic system, it’s not the result of forces that impose from outside.
Endogenous growth hypotheses depend on the likelihood that long-run growth is controlled by
various economic incentives. The most mainstream models of this sort keep up that inventions are
deliberate and produce technological spillovers that bring down the expense of future innovations.
Basically, in these models, an informed workforce plays a unique role in deciding the pace of
technological innovation and its long-run economic path (Mankiw, 2010; Zhao, 2019)
As Zhao (2019) describes, Romer’s model is noticeably more complex due to the presence of
three sectors (R&D, intermediate capital goods, and final goods) and the need to explicitly analyze
the inner workings of these sectors. After analyzing decisions in the R&D and the sector of
intermediate good and with some calculations. The model arrives at the final goods production
function is proportional to the following inferences (see equation c)
Kαt AtL
F
t
� �1� α
(c)
One can immediately realize the similarity with the Solow model. The central difference is that
change in technology At is endogenously determined as workers choose between working in the
final goods sector and the R&D sector in the market equilibrium. In market equilibrium, Romer
derives the share of labor input devoted to R&D and technological change g. Hence, g is the result
of the decisions set by workers, consumers, and entrepreneurs not assumed as in the Solow model.
To this end, Solow’s exogenous growth rate g is endogenized (Zhao, 2019).
One of the imperative implications of the Romer model concerns population growth. In the
Solow model, population growth does not contribute to the growth of per capita income, which
only depends on (exogenous) technology growth. While in Romer’s model, population can be
a source of growth through more labor working in the R&D sector will enhance the rate of
technological change.
Across the board, Romer’s analysis has far-reaching policy inferences. It suggests that govern-
ment can do more, for example, using subsidies to correct the market failure and spur economic
growth by promoting the development of science and technology. In recent years, technology has
transformed people’s lives throughout the globe, compared to just a few decades ago. However,
the impacts of technology do not fall equally on everyone’s life. Let’s say when a new technology
arrives, it interrupts the old ways of doing things and can have undesirable effects on some
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 5 of 17
people’s lives, at least for an early time. Society must confront the negative impacts of new
technology while embracing its life-enhancing potentials.
4. Fundamentals of economic growth and some stylized facts
In the new growth theories, research, and development through the production of goods captured
by K (t) and labor L (t) with the full stock of knowledge, A (t) verify long-run economic growth. On
the flip side, a nation cannot maintain its long-run growth by simply accumulating more capital or
labor due to its diminishing returns. Accordingly, endogenous technological progress is the main
driver of rigorous long-run economic growth as stated in the Romers model.
Furthermore, empirical verdicts confirmed that various covariate factors determine the long-run
economic growth of nations. It has been characterized as a direct and indirect cause of economic
growth. Direct factors include human resources (active population, investing in human capital),
natural resources, physical capital, or technological advancements. Indirect causes comprise
institutions (well-devised institutions, financial institutions, private administrations, property
rights), the size of the aggregate demand, saving rates and investment rates, the efficiency of
the financial system, budgetary and fiscal policies, migration of labor and capital, and the effi-
ciency of the government. Generally, it can be consolidated into four major determinants of
economic growth: human resources, natural resources, capital formation, and technology, but
the importance that researchers had given for each determinant has been always different.
Similarly, the economic growth of a nation is influenced by various inter-related factors. As
(Boldeanu & Constantinescu, 2015), states that it can be further sorted into supply-side, and
efficiency, and demand-side factors. On the supply side, natural resources (both renewable and
non-renewable), physical capital goods, human capital, and technology have a direct effect on the
value of goods and services supplied. In addition, socio-political factors and events have a major
influence on the economic progress of a country. Acemoglu et al. (2009) mainly emphasized that
the main determinants are categorized into economic and non-economic classes. “Proximate” or
economic determinants stand for capital accumulation, technological progress, labor, and likewise
“ultimate” or non-economic sources refer to government efficiency, institutions, political and
administrative systems, cultural and social factors, geography and demography can surrogate
economic progress of countries.
Scholars confirmed that different positive economic sources of economic growth like Schooling,
education investment (Barro, 1991); level of human capital (Gould and Ruffin (1993); capital savings
and investment (Mankiw et al., 1992); Equipment investment (De Long & Summers, 1991); open to
trade (Barro, 2003; Romer, 1990). FDI inflows (Lensink & Morrissey, 2006; Li & Liu, 2005).
According to (A. v. Cooray, 2009) who confirms that using a cross-section of 71 economies; both the
size and quality of the government are vital to economic growth. Such as investing in the capacity for
enhanced governance is a priority for the improved growth performance of the countries.
Likewise, non-economic sources for economic growth include high governance (A. Cooray, 2009),
institutional frameworks (property rights, regulatory institutions, institutions for macroeconomic
stabilization, social insurance, and conflict management) (Abrams & Lewis, 1995; Mauro, 1995;
Rodrik, 2000b; Acemoglu et al. 2002).
Some scholars stated that corruption can also be sometimes beneficial because it can make the
economy more efficient and facilitate for investors a way to pass more restrictive and ease
bureaucratic hurdles (Acemoglu et al., 2000; Kaufman; Wei, 2000). While others state that it is
undesirable and it has negative influences and bad outcomes on economic growth (Murphy et al.,
1993; Mauro, 1995).
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 6 of 17
In addition, negative economic influences extend to factors like government consumption
spending (Barro, 1991), Trade barriers (Gould & Ruffin, 1993; De Long & Summers, 1991); Military
spending (Benoit, 1978; Pieroni, 2009; Ho and Chen, 2014). Furthermore, non-economic factors
that have a negative impact include political and social instability (Barro, 1991), socio-cultural
factors, ethnic diversity, and fragmentation, language, religion, civic norms, beliefs are among the
socio-cultural features that may have a negative influence on economic growth. Ethnic diversity
can have a destructive outcome on education (low schooling), creates political instability. As
W. Easterly and Levine (1997), Easterly & Levine (1999) argue that using cross-country differences
in ethnic diversity in Africa, low level of economic development is associated with a high degree of
ethnic heterogeneity and ethnic fractionalization and it affects the choice of public policies. The
existences of political unrest follow discouraging public investment in the sense of implanting low
educational policy, poor infrastructure and growth become collapse and end up with rent-seeking
and benefits towards the competing bodies at the expense of massive population living in poverty.
Thus, this existing behaviour among groups follows in competition in position rather than following
good policies which are productive and public sector investments.
Nevertheless, ethnic diversity also brings new insights, tolerance, appreciation of diversity, and an
opportunity to share and learn from others. It can be beneficial by enhancing productivity through
innovation, skill complementarities, increased creativity, trade, and product varieties. Furthermore, as
García Montalvo and Renal-Querol (2019) state that the association between economic growth and
ethnic heterogeneity is complex. In cross-country data, it confirms a negative or statistically insig-
nificant result. However, at city level analysis—data from small geographical areas-ethnic diversity
create a positive effect on wages and productivity (Alesina et al., 2016; Ottaviano & Peri, 2005). There
is a trade-off between economic benefits and the costs of heterogeneity.
4.1. Conceptual hypothesis related to economic growth
4.1.1. Convergent hypothesis
One of the fundamental economic issues is whether poor nations or regions tend to grow faster
than rich ones: are there automatic forces that lead to convergence over time in the levels of per
capita income and product? (Barro, 1992). From a theory postulating that a convergent occurs
between poorer economics and wealthy economies of national income due to an acceleration of
growth as poorer economies “catch-up” in their use of technologists (Barro, 1991, 1992; Barro and
Sala-i-Martin, 1992; Dowrick & Nguyen, 1989; Mankiw et al., 1992). Advanced economies with high
levels of productivity tend to grow slower than poor economies that require much less capital to
make significant gains. What is behind this circle of convergence? The rationale is somewhat
straightforward—when a country is poor, subsequently that there is little production; hence,
there are few factories and labor is cheap. Building a new factory plant in such a country can be
quite worthwhile: if the product is easily tradable, then with low labor costs and earn higher rates
of return on investment. Alternatively, we can consider the new manufacturing as supplying for
the local market an item that did not exist previously. Once more, benefits will be high a direct
result of the absence of rivalry. Likewise, if the country is starting poor then it has various
opportunities for copying and importing knowledge. Hence, the level of economic growth rates
becomes higher. It can be noted that the initial conditions provide the potential for catching up.
Thus, poor countries can grow faster when they set on a convergence path to the rich economies.
China today, Singapore in the 80s.
4.1.2. Institution hypothesis
For all investment activity, there are expected benefit and risk that is determined by a long list of
factors that we put together under the label of institutions, such as legal institutions (the rule of
law, property rights), political institutions (stability of policy, decision-making), economic institu-
tions (regulation, taxes, customs duties, and procedures), social norms (that will determine how
issues like income inequality will be resolved, which will affect policy variables such as tax rates),
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 7 of 17
culture (entrepreneurial spirit, risk-taking behavior, attitudes to work). In short, the environment for
doing business matters, and it matters a lot. This environment we call it institutions.
“Institutions are the humanly devised constraints that structure political, economic, and social
interaction. They consist of both informal constraints (sanctions, taboos, customs, traditions, and
codes of conduct), and formal rules (constitutions, laws, property rights). Throughout history,
institutions have been devised by human beings to create order and reduce uncertainty in
exchange” (North, 1990).
Robust economic growth and quality of institutions have been recorded. Advanced countries
have institutions of high quality, while the opposite is true about poor countries. This measure of
institutional quality provides countries and governments with clear guidelines on what to do to
speed up reforms and growth. The goal for the government should be to set up the right environ-
ment for business rather than managing investment. Once the environment exists, once it is
rewarding to save and a big chunk of the uncertainty about future payoffs is gone, individuals
and firms start putting aside more money for investment and growth picks up.
In recent years, the “new institutional economics” has emphasized the possible effects on
growth rates and income levels of property rights, contract enforcement, and rent-seeking activ-
ities (e.g., North, 1990; Weingast, 1993).
In principle, the change in institutions and economic growth are interrelated and possibly further
institutional changes might set in only after countries develop further. But although growth itself
determines the willingness to change, it is still valid to say that in most of the poor countries there are
still ample opportunities for improvements in the business environment. The good institution plays the
best way to ensure sound macroeconomic policies (i.e., stability) and political stability is to build
institutions that create incentives for stability: No country has become rich with poor-quality institutions.
The role of institutions for economic growth has been emphasized by many scholars. According
to Acemoglu et al. (2002), considerable attention is given to the functionality of institutions that
highly influence the economic growth of countries. Emphasizing the previously productive area of
the hot climate later fails to compute with the temperate area where the introduction of advanced
technology in the agricultural systems of production and type of improved varieties are released.
Thus, institutions matter more for economic progress.
In growth literature, particular attention is given to endogenous variant factors like policy and
institutional frameworks rather than paying attention to exogenous “unchanging” factors such as
geography. Thus, in the success ladder of the country’s economic growth and development, the
existence of well-performed and managed institutions is vital. It helps to secure property rights and
directs to productive investments and wealth of nations. In this perspective, the country’s economic
prosperity is determined through regulating economic actors and creates a sense of ownership to the
asset properties owned and this can be manifested by achieving better economic development
through directing the use of resources in an intended way (Acemoglu, Simon Johnson et al., 2001).
Hence, evidence assured that different economic performance between North and South Korea, West
and East Germany growth divergence of economy confirmed the existence of well-performed institu-
tions and well-equipped and inclusive social welfare systems. Thus, along with secure property rights
and a market-led economy promotes investments, reinvestment opportunity and creates more space
for human capital development and able to facilitate more economic growth.
Unlikely, those unsuccessful economies follow policies whose interests and benefits are directed
towards only the elites and no property right established for the welfare of the whole society. So that
there is no guarantee for the majority of the society’s assets and then it discourages further
investments and ended up with an unsuccessful economy. Hence, when the institutions are
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 8 of 17
extractive, the system established is serving for the hands of few, and this leads to “risk of expropria-
tion” for a vast number of the society, which discourages the people to invest and finally ends up with
collapse.
According to Acemoglu et al. (2002), in his “institutions hypothesis,” the presence of a good
institution is vital for economic growth and development through adopting new processes and
a new style of desirable economic outcomes and able to promote investment. Those countries that
already established good institutions are more likely to induce and accelerate industrialization
than those that did not set good institutions. By the same token, from the perspectives of time and
nature of the establishment of institutions, during the time of colonization, Europeans used to set
up good institutions in their settlement and developed secured property rights and thus positively
influenced the economic growth of a country of their colony. The reverse has been recorded when
they set up bad/extractive institutions.
4.1.3. Geography hypothesis
In another dimension, scholars argue that once institutions and other factors controlled the role of
geography matters for economic growth and it leads to the type of policy intended to follow. As
Gallup et al. (1998) argued that climate conditions and geographical location is vital for the
countries to determine economic destiny Different Scholars looking the “geography hypothesis,”
in the perspective of explaining countries’ diverse economic progress. Such as differences in
resource endowments, climate (hot, cool environments, temperate, tropics) or ecological variances
across countries and prevalence of diseases, transportation cost—covered the distance from the
core market—and the transfer of knowledge and all its influences on the development of human
capital and productivity of agricultural produces (Diamond, 1997; Gallup et al., 1998, 1998).
Developing countries of South East Asia, sub-Saharan Africa, and South America with a hot climate
and humidity cause decline labor productivity, widely spread and the prevalence of diseases, low
crop productivity due to insect infestation.
People and ideas influence events, but geography largely determines them, now more than ever
(Kaplan, 2009). According to Kaplan, realism is about recognizing and embracing those forces
beyond our control that constrain human action. And of all the unsavory truths in which realism is
rooted, the bluntest, most uncomfortable, and most deterministic of all is geography.
In the discussion of geographical location, Arvis et al. (2007) highlight that the cost of a country
being landlocked. They show that a landlocked country bears not only the high cost of freight
services but also the cost of unpredictable transportation time, widespread rent activities, and
severe flaws in the implementation of the transit systems. An ocean in the immediate surround-
ings of a country has a positive impact on its GDP per capita.
According to (Marshall, 2016), in his book “Prisoners of geography: ten maps that explain
everything about the world” argued that societies are inevitably shaped by the land upon which
they exist. The features of natural resources and geographical topographies can provide safety and
affluence or leave a country’s citizens exposed and struggling. Geography has been a determining
factor in the wars humans fight, as well as the rate of our economic growth. Although modern
technology, the distance between us both in mental and physical space now allows us to bend the
rules of geography, it remains crucial to understanding why nations have turned out the way they
are today and it looks fundamental in geopolitics interaction of different nations.
Furthermore, the near ocean areas have more intense economic activity going on and in the
case of the inland countries, economic activity appears more intense along navigable rivers where
transportation by ship is feasible. One other criterion accompanying intense economic activities is
a temperate climate with adequate rainfall; most likely because it is favourable for the productivity
of agriculture and for mitigating disease. The following image (Figure 1: a collection of images from
the DMSP) (Défense Meteorological program). This highlights that the level of different activity at
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 9 of 17
the global coverage level whether inland or coastal access countries. It inspects darkness and the
presence of light for those countries on the globe. Sub-Saharan Africa (SSA) countries except for
some of it, the level of darkness at the night dominates regardless of whether the country has
coastal access or not. The same is true for some Asian and Latin American countries.
After World War II, there was a surge in the empirical analysis of geography. Braudel (1981–
1984), Crosby (1986), and Diamond (1997) analyzed the impact of geography and climate change
in Europe and its dominance over the colonies. North-Atlantic and Mediterranean Europe were the
creative centers of the world after the Middle Ages ended. Acemoglu (2009) affirmed that geo-
graphy can affect in many ways’ economic growth. Soil quality can influence agricultural produc-
tivity. Natural resources directly contribute to the industrialization of a country by essential
components for production. Climate has a direct impact on production and attitudes regarding
consumption. The topography of a region or state can have a positive or negative impact on
transport costs and communication. And not least, diseases can affect health care, production,
and the accumulation of human and physical capital.
According (Diamond, 1997), in his book “Guns, Germs, and steel” stated that farm-based societies
conquered populations of other areas and maintained dominance, despite sometimes being vastly
outnumbered—superior weapons provided immediate military superiority (guns); Eurasian diseases
weakened and reduced local populations, who had no immunity, making it easier to maintain control
over them (germs); and durable means of transport (steel) enabled imperialism. Diamond argues
geographic, climatic, and environmental characteristics which favoured the early development of
stable agricultural societies ultimately led to immunity to diseases endemic in agricultural animals
and the development of powerful, organized states capable of dominating others.
4.2. The geography of land locked with bad neighbours
Paul Collier (2010) in his book “The bottom Billion” argues that being landlocked in a poor geographic
neighbourhood is one of the four major development “traps” by which a country can be held back. In
general, he found that when a neighbouring country experiences better growth, it tends to spill over
into favourable development for the country itself. For landlocked countries, the effect is particularly
strong, as they are limited in their trading activity with the rest of the world. He states, “If you are
coastal, you serve the world; if you are landlocked, you serve your neighbours.” Others have argued
that being landlocked has an advantage as it creates a “natural tariff barrier” which protects the
country from cheap imports. In some instances, this has led to more robust local food systems.
Indeed, being landlocked does not necessarily condemn a country either to poor or slow growth,
but 38 percent of the people living in bottom-billion societies are in landlocked countries. Expenses for
Figure 1. a) Dark and light
inspection: source: eco browser,
analysis of current economic
conditions and policy, http://
apod.nasa.gov/apod/ap001127.
html; collected from the images
sent by DMSP satellites: http://
heasarc.gsfc.nasa.gov/docs/hea
sarc/missions/dmsp.html, 2016.
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 10 of 17
http://apod.nasa.gov/apod/ap001127.html
http://apod.nasa.gov/apod/ap001127.html
http://apod.nasa.gov/apod/ap001127.html
http://heasarc.gsfc.nasa.gov/docs/heasarc/missions/dmsp.html
http://heasarc.gsfc.nasa.gov/docs/heasarc/missions/dmsp.html
http://heasarc.gsfc.nasa.gov/docs/heasarc/missions/dmsp.html
landlocked countries in terms of costs in shipping and trade of their goods but furthermore the ability
to start manufacturing industries that require a lot of transport is bound to be challenging. Along with
the above is the type of neighbours that the landlocked country has as some of them may provide the
point of first trade or contact such as the differences seen between Switzerland and Uganda in this
respect. This can be married with the point that most of the landlocked countries in Africa have
completely inward-looking or toward the world market policies and not meant to reap the spill-over
benefits of their neighbours which in any case may also be limited or non-existent. Spill-over benefits
for every 1% growth the world average is 0.4% the Non-African Landlocked country average is 0.7%
and that for African landlocked country is very low at 0.2% (Collier, 2010)
While others stated that the location of the country affects growth through productivity and
proximity to the potential market, this leads to shape the choice of policy that a country is
supposed to follow. Meanwhile, countries nearest to the world market benefit to follow a policy
that leads to open their economy than a country far away from the world’s potential market. Thus,
it’s the main finding of growth and geography indicates that domestic trade, as well as interna-
tional trade and productivity, are influenced by access to seas and ports, this is assured by those
countries nearest to the coastal area become more urbanized. But, many developing countries
especially in Africa, majority of the population are living in landlocked, and it extremely affects
transportation and trading capacity and problematic to integrate into the core economies and
industrialize per se (Gallup et al., 1998).
In addition, many more challenges besides the geographical location in the tropics are that
productivity is affected by the weak labor force injured by disease and malaria as we approach the
tropics, disease prevalence becomes so sophisticated and productivity declines. Thus, tropical and
temperate agricultural productivity shows a great variation due to these circumstances (Ibid).
4.2.1. Natural resource and its paradox
The existence of natural endowments serves as an engine of growth where it is supported by the
quality of the institution unless it becomes a curse. Countries with abundant natural resources with
wise management and target on the production of the technological-based production system and
high intensity to productive assets investment make the country more benefited. Nevertheless, there
is a negative effect associated with its abundance due to price instability and challenges the financial
sectors of the economy (Van der Ploeg, 2011). Empirical studies showed that there is a negative
association between natural resources abundance and economic performance when countries
depend on exporting their excess abundance of the natural resources and ended up being a curse
with it (Mavrotas et al., 2011). Similarly, the existence of natural resources (new oil discoveries Sachs,
2003) leads to the appreciation of the real exchange rate and it crowds out the other sectors and
harms other export sectors of the economy Sala-i-Martin and Subramanian (2013). Accordingly,
Hirschman 1958 as cited in Sachs & Warner (1999), countries with plenty of natural resources are
failed to complement with supportive sectors and end up with poor performance and consume more
rather than reinvesting it. Abrams and Lewis (1995) presented the effect of natural resources on long-
term economic growth and confirmed that resource-rich countries tend to grow more slowly than
resource-scarce countries, and they are labeled as the “natural resource curse.” Similarly, natural
resource richness crowds out human and physical capital, causing slower growth in the long term
(Gylfason, 2001; Gylfason & Zoega, 2006). Moreover, resource-rich countries more emphasized on
service sector than manufacturing sectors, and this tends to lag economic progress by the declining
exporting capacity of manufacturing sectors. Not only this but resource-rich countries are also prone
to corruption and established low institutional quality, and lead to negative economic growth.
Paul Collier in his book, “The Bottom Billion” argues natural resource traps. This looks at what
happens when the countries rely on unearned rents. In addition to this is the issue of the “Dutch
disease” where the reliance on a particular natural resource affects the other sectors. Boom and Bust
cycles are a problem that comes along with Dutch disease in that there is normally uncontrolled
spending and borrowing as a result of booms and in the periods of busts it’s difficult to reprioritize
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 11 of 17
effectively but rather politically which creates an even bigger problem. Hence, abundant natural
resources cannot exclusively explain and determine the fate of the country’s economic growth and
development since resource abundance in a country leads to either of the curse or blessing outcomes.
Nevertheless, the existence of resources1 is not a curse by itself but the type of policy they
followed, institutions, and human capital accumulation matter for the achievements. With plenty
of natural resources, only a few countries become successful such as Botswana, Norway, Australia,
and Canada behind this success the establishment of the institution is immense with human
capital development. Different bad experiences are linked with the existence of plenty of natural
resources, Zambia, Nigeria, Sierra Leone, Angola, Venezuela, and Saudi Arabia. But resource-poor
countries of Asian Tigers: Taiwan, Korea, Hong Kong, and Singapore showed progress amazingly.
This lesson should be taken as the best experiences of being a winner without resources and type
of policy they followed with resources and its institutional quality substances more (Acemoglu
et al., 2002; Mehlum et al., 2006; Sachs, 2003). Likewise, abundance resources become curses due
to incompatible institutions (Mavrot et al. 2011).
Source: (Mehlum et al., 2006); Economic growth with resource and institutions.
Besides all the above factors in the process of economic and wealth accumulation of countries,
other factors should explicitly have accounted to know the real effects of the choice of policies
intended to follow and its governing systems (Figure 2). In line with this, the argument of the
existences of democracy, there is an unclear and controversial sight on which comes first, in essence,
does democracy lead to growth or the reverse leads? But, most of the empirical investigation pointed
out that the role of democracy and the existence of political stability for the sake of solving commu-
nal problems and designed a project for the interest of the public and against self-interested gain
from public resources. Democracy is appropriate for the benefit of the majority and maximizing
output by fairly allocating the existing resources efficiently and it discourages corrupt officials. It
links with the establishment of better institutions and demand multi-dimensional aspects and later
enhances development further. Besides, Tavares & Wacziarg (2001) argues that democracy helps the
poor by lowering inequality by redistributing income but at the expense of physical capital accumula-
tion; thus, it affects growth negatively. The latter when there is more democracy, it increases the
Figure 2. Institutions vs
resources (a) performance of
rich countries (b) associated
with bad institutions, (c)
Countries with good institu-
tions. thus, it is the self-
explanatory figure above about
the resources-rich countries
perform negatively, in
a situation a) and bad institu-
tion aggravates in case b, and
good institution compliments
for the growth as indicates in
c).
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 12 of 17
government’s consumption thus ultimately reduces growth. Moreover, Collier (2007) states that
resource-rich democracies not only under-invest but invest badly, with too many white-elephant
projects and focus on elections. This can be seen as the reason why democracy’s ability to harness
resource surpluses is poor based on the findings of the research that was carried out. They have
generally seen the problem of resource rents as being proneness to autocracy: oil induces Saddam
Hussein. There is good evidence for this, but the real problem is even worse.
While others look at the process of economic growth and accumulation of wealth, accounting
social, political, and cultural development becomes more an imperative phenomenon for better
achievements. Thus, fighting poverty and the pace of economic growth takes cultural and moral
dimensions in the poor society at this time. At the beginning of the industrial revolution and capital
development, the set of cultural values gives a clue for economic development (Weber, 1905, as
cited in Cuesta, 2004). Understanding the culture of the society and creating a link with the past
and future economic development is ideal to account for culture as an influencing parameter of
economic development. Thus, development expresses in terms of the cultural dimension of giving
a reserved value to the system beliefs, values, working habit, truthfulness, readiness to accomplish
a given complicated task, and interaction with other new people and structures of social interac-
tions. Hence, religion expresses the cultural habits of nations. For instance, attendance into
religious institutes negatively affects economic growth, but beliefs, trust, and associated things
are positively linked with economic growth (Barro, 2003). While arguing that interims of nations’
awareness in the process of productivity and competitiveness, belonging in certain circumstances
speed up national production and global competencies.
4.2.2. War and state capacity
In the process of building a strong economy, other factors should be accounted for the nature of the
government and its size in which adequately serve the wellbeing of the society. In the sense of giving
protection to its societies and security and administer the economic systems. Thus, a country that builds
its economy mainly depends on the capacity of the state, where state assurances for the provision and
development of publicly available goods and direct in an intended way. According to (Tilly, 1992), in
Europe, in the process of state-building its capacity there were a lot of wars, and it is undisputable liked
to the creation of internal and external sovereignty and end up with strong government and hence
state capacity. The country’s economic growth process highly depends on the capacity of the state in
extending its economic power, enforcement to build and implementing appropriate policies interims of
imposing and securing property right to direct investments in the right direction.
Meanwhile, lack of state capacity induces civil wars, internal instability, unfair and unequal
distribution of resources; which many developing countries lack this capacity at this time, and
the share of the government in the economy is very small and incapable to afford and provide
adequate funding to the provision of the public good. Thus, building a strong economy requires,
being strong in building state capacity, having a big share of government expenditure in a state-
led economy. Moreover, to strengthen state capacity and fulfill nation’s demand’ make sure
internally stable and build an externally competitive economy.
5. Summary and possible suggestions
Growth is a complex phenomenon that depends on many factors. But a quick look at a few stylized
facts has shown us that both economic and non-economic, endogenous and exogenous factors
are considered.
Divergent growth experiences of countries across different regions take decisive factors for long
periods and it accounts for both time-variant and invariant factors, and recent findings more
emphasized the changing parameters for countries economic growth and economic affluence.
Strong institutional frameworks play a pivotal role in achieving robust economic progress via
developing faiths and reduce transaction costs, and encouraging productive (re)investment
options along with human capital development.
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 13 of 17
Consequently, well-functioned institutions are used as a bridge for better economic progress and
stimulate forward and backward linkages of different economic sectors but not everything. The
role of geography and resource endowments should not be underestimated. Similarly, geography,
the location of the country, does not regulate growth performance exclusively as well, even if
countries in the tropics disfavour by climate condition for the prevalence of disease and produc-
tivity, it is not a sufficient condition to be poor. Moreover, countries with coastal, access to the sea
did not industrialize first being an advantage for it (such as Indonesia, the Caribbean, India). Thus,
the mystery is just allocating all changing and unchanging factors together, and designing good
policy, openness, administrative capacity and establishing well-performed institutions, developing
human capital, more emphasis on changing factors. By saying this, sub-Saharan Africa and other
regions struggling today for improved economic development require much more than lectures
about good governance and institutions.
Finally, state capacity is vital to determine the relative wealth of state economic affluence.
Meanwhile, the state should be more strong and powerful in securing their sovereignty and
internal affairs and to enforce and securing property rights, and directing investments in the
desired perspectives.
5.1. Possible suggestions
Along all the discussion for the robust economic growth performance, time-variant factors than
time-invariant parameters shall lead the countries into prosperities and desirable outcomes.
Hence, good institutions by secured property rights, good governance, improve education health
and facilities will lead to countries into desirable outcomes. Geographical disadvantageous and
landlocked countries better involve in a combination of multiple strategies as can be seen below;
● Engage more on creating financial capacity, and human capital development and regional
integrations and promote win–win policies: increase neighbourhood growth spill-overs,
improve neighbours’ economic policies and improve ways for getting coastal access.
● Develop alternative and possible all-rounded infrastructure facilities, telecommunications,
railways and services: do not be Air-locked or E-locked and support and encourage remit-
tances from various nations.
● Develop and create a transparent and all possible investor-friendly environment for resource
prospecting.
● Focus on rural development policies, strategies, and biases towards it.
● Seek alternatives to attract Aid and foreign investments in various sectors.
● Promote research & development and homegrown policies for home-specific problems.
The author is grateful to anonymous reviewers and the
editor for their valuable comments and suggestions on
the earlier version of this paper.
The authors declare that they have not received any fund
for this article.
Esubalew Tadele1
E-mail: esu360@gmail.com
E-mail: esubalewtadele@dmu.edu.et
ORCID ID: http://orcid.org/0000-0001-7519-2738
Teshome Sirany2
1 Department of Agricultural Economics, Debre Markos
University, Debre Markos, Ethiopia.
2 Department of Rural Development, Debre Markos
University, Debre Markos, Ethiopia.
Citation information
Cite this article as: The relative choice over destiny in a
country’s long-run economic growth and economic afflu-
ence, Esubalew Tadele & Teshome Sirany, Cogent
Economics & Finance (2021), 9: 1949133.
1. Even if countries with natural resource scarcity, the
growth process somewhat restricted by these con-
straints, capital can be accumulated and labour pro-
ductivity can be improved through intensive training
and using higher technology. Caution should be taken
when natural resources are scarce in the production
process, it imposes restricts limitation on the destiny of
a country wealth affluence, yet this is not a general
case, for instance, Japan achieve this miracle growth
performance and enjoys high standard of living with-
out plenty of natural resources (Roy et al. 2013).
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 14 of 17
The author(s) declare that they have no competing
interests.
Abrams, B. A., & Lewis, K. A. (1995). Cultural and institu-
tional determinants of economic growth: A
cross-section analysis. Public Choice, 83(3), 273–289.
https://doi.org/10.1007/BF01047747
Acemoglu, D. (1995). Reward structures and the allocation
of talent. European Economic Review, 39(1), 17–33.
https://doi.org/10.1016/0014-2921(94)00014-Q
Acemoglu, D. (2009). Introduction to modern eco-
nomic growth. Princeton NJ: Princeton University
Press
Acemoglu, D., Johnson, S., & Robinson, J. A. (2000) The
colonial origins of comparative development: An
empirical investigation. National Bureau of Economic
Research (Cambridge, MA). Working Paper No. 7771.
http://www.nber.org/papers/w7771
Acemoglu, D., Johnson, S., & Robinson, J. A. (2001). The
colonial origins of comparative development: An
empirical investigation. American Economic Review,
91(5), 1369–1401. https://www.nber.org/system/files/
working_papers/w7771/w7771
Acemoglu, D., Johnson, S., & Robinson, J. A. (2002).
Reversal of fortune: Geography and institutions in the
making of the modern world income distribution. The
Quarterly Journal of Economics, 117(4), 1231–1294.
https://doi.org/10.1162/003355302320935025
Acemoglu, D., Simon Johnson, J. A., & Robinson, J. A.
(2001). The colonial origins of comparative develop-
ment: An empirical investigation. The American
Economic Review, 91(5), 1369–1401. https://doi.org/
10.1257/aer.91.5.1369
Alesina, A., Harnoss, J., & Rapoport, H. (2016). Birthplace
diversity and economic prosperity. Journal of
Economic Growth, 21(2), 101–138. https://doi.org/10.
1007/s10887-016-9127-6
Arvis, J. F., Raballand, G., & Marteau, J. F. (2007). The cost
of being landlocked: Logistics costs and supply chain
reliability (Vol. 4258). World Bank Publications.
Auty, R. M., & Gelb, A. G. (2001). Political economy of
resource abundant states. In R. M. Auty (Ed.),
Resource abundance and economic development (pp.
126–144). University Press.
Barro, R. J. (1991). Economic growth in a cross section of
countries. The Quarterly Journal of Economics, 106(2),
407–443. https://doi.org/10.2307/2937943
Barro, R. J., & Sala-i-Martin, X. (1992). Convergence.
Journal of Political Economy, 100(2), 223–251.
https://www.journals.uchicago.edu/doi/pdfplus/10.
1086/261816
Barro, R. J. (1992). Convergence. Journal of Political
Economy, 100(2), 223–251. https://doi.org/10.1086/
261816
Barro, R. J. (2003). Determinants of economic growth in
a panel of countries. Annals of Economics and
Finance, 4, 231–274.
Benoit, E. (1978). Growth and defense in developing
countries. Economic Development and Cultural Change,
26(2), 271–280. https://doi.org/10.1086/451015
Boldeanu, F. T., & Constantinescu, L. (2015). The main
determinants affecting economic growth.
Transilvania University of Brasov, 8(2), 329–338.
http://webbut.unitbv.ro/bu2015/series%20v/BILETIN
%20I/38_Boldeanu%20Constantinescu
Braudel, F. (1981-1984). Civilization and Capitalism, 15th-
18th Century. Collins (3 volumes).
Chirwa, T. G., & Odhiambo, N. M. (2016). Macroeconomic
determinants of economic growth: A review of
international literature. South East European Journal
of Economics and Business, 11(2), 33–47. https://doi.
org/10.1515/jeb-2016-0009
Collier, P. (2007). Bottom billion. In The Blackwell
Encyclopaedia of Sociology (pp. 1–3).
Collier, P. (2010). The bottom billion and what we can do
to help. In Fairness in international trade (pp. 3–8).
Springer.
Cooray, A. (2009). Government expenditure, governance
and economic growth. Comparative Economic Studies,
51(3), 401–418. https://doi.org/10.1057/ces.2009.7
Cooray, A. V. (2009). The role of education in economic
growth. Available at SSRN 1520160.
Crafts, N. F. R. 2000. Development history. department of
economic history Working Paper No. 54/00. London
School of Economics.
Crosby, A. W. (1986). Ecological imperialism: The biological
expansion of Europe, 900–1900. Cambridge
University Press.
Cuesta, J. (2004). From economists to culturalist devel-
opment theories: How strong is the relation between
cultural aspects and economic development? The
European Journal of Development Research, 16(4),
868–891. https://doi.org/10.1080/
09578810412331332677
Curzio, A. Q., Fortis, M., & Zoboli, R. (1994). Innovation,
resources and economic growth: Changing interac-
tions in the world economy. In Innovation, resources
and economic growth (pp. 3–35). Springer.
De Long, J. B., & Summers, L. H. (1991). Equipment
investment and economic growth. The Quarterly
Journal of Economics, 106(2), 445–502. https://doi.
org/10.2307/2937944
Diamond, J. (1997). Guns. germs and steel: The fates of
human societies, vintage.
Diamond, J. (1998). Guns, germs and steel: A short history
of everybody for the last 13,000 years.
Dowrick, S., & Nguyen, D. T. (1989). OECD comparative
economic growth 1950-85: Catch-up and
convergence. The American Economic Review,
1010–1030. https://www.jstor.org/stable/1831434
Easterlin, R. (1981). Why is not the whole world
developed? Journal of Economic History, 41(1), 1–19.
https://doi.org/10.1017/S0022050700042674
Easterly, W., Kremer, M., Pritchett, L., & Summers, L. H.
(1993). Good policy or good luck? Journal of Monetary
Economics, 32(3), 459–483. https://doi.org/10.1016/
0304-3932(93)90026-C
Easterly, W., & Levine, R. (1997). Africa’s growth tragedy:
Policies and ethnic divisions. The Quarterly Journal of
Economics, 112(4), 1203–1250. https://doi.org/10.
1162/003355300555466
Easterly, W., & Levine, R. (1999). Africa’s growth tragedy:
Policies and ethnic divisions. The Quarterly Journal of
Economics, 112 (4), 1203–1250. MIT Press. https://
www.jstor.org/stable/2951270
Easterly, W., & Levine, R. (2003). Tropics, germs and crops.
How endowments influence economic development.
Journal of Monetary Economics, 50, 3–39. https://doi.
org/10.1016/S0304-3932(02)00200-3
Gallup, J. L., Sachs, J. D., & Mellinger, A. D. (1998)
Geography and economic development, NBER
Working Paper no 6849.
García Montalvo, J., & Renal-Querol, M. (2019). Ethnic
diversity and growth: Revisiting the evidence. The
review of economics and statistics. 2020 Jan 23
Gould, D. M., & Ruffin, R. J. (1993). What determines
economic growth. Economic Review, 2. https://core.
ac.uk/download/pdf/6755147
Gylfason, T. (2001). Natural resources, education, and
economic development. European Economic Review,
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 15 of 17
https://doi.org/10.1007/BF01047747
https://doi.org/10.1016/0014-2921(94)00014-Q
http://www.nber.org/papers/w7771
https://www.nber.org/system/files/working_papers/w7771/w7771
https://www.nber.org/system/files/working_papers/w7771/w7771
https://doi.org/10.1162/003355302320935025
https://doi.org/10.1257/aer.91.5.1369
https://doi.org/10.1257/aer.91.5.1369
https://doi.org/10.1007/s10887-016-9127-6
https://doi.org/10.1007/s10887-016-9127-6
https://doi.org/10.2307/2937943
https://www.journals.uchicago.edu/doi/pdfplus/10.1086/261816
https://www.journals.uchicago.edu/doi/pdfplus/10.1086/261816
https://doi.org/10.1086/261816
https://doi.org/10.1086/261816
https://doi.org/10.1086/451015
http://webbut.unitbv.ro/bu2015/series%20v/BILETIN%20I/38_Boldeanu%20Constantinescu
http://webbut.unitbv.ro/bu2015/series%20v/BILETIN%20I/38_Boldeanu%20Constantinescu
https://doi.org/10.1515/jeb-2016-0009
https://doi.org/10.1515/jeb-2016-0009
https://doi.org/10.1057/ces.2009.7
https://doi.org/10.1080/09578810412331332677
https://doi.org/10.1080/09578810412331332677
https://doi.org/10.2307/2937944
https://doi.org/10.2307/2937944
https://www.jstor.org/stable/1831434
https://doi.org/10.1017/S0022050700042674
https://doi.org/10.1016/0304-3932(93)90026-C
https://doi.org/10.1016/0304-3932(93)90026-C
https://doi.org/10.1162/003355300555466
https://doi.org/10.1162/003355300555466
https://www.jstor.org/stable/2951270
https://www.jstor.org/stable/2951270
https://doi.org/10.1016/S0304-3932(02)00200-3
https://doi.org/10.1016/S0304-3932(02)00200-3
https://core.ac.uk/download/pdf/6755147
https://core.ac.uk/download/pdf/6755147
45(4–6), 847–859. https://doi.org/10.1016/S0014-
2921(01)00127-1
Gylfason, T., & Zoega, G. (2006). Natural resources and
economic growth: The role of investment. The World
Economy, 29(8), 1091–1115. https://doi.org/10.1111/
j.1467-9701.2006.00807.x
Hansen, G. D., & Prescott, E. C. (2002). Malthus to solow.
American Economic Review, 92(4), 1205–1217.
https://doi.org/10.1257/00028280260344731
Hou, N., & Chen, B. (2014). Military spending and eco-
nomic growth in an augmented solow model: A
panel data investigation for OECD countries. Peace
Economics, Peace Science and Public Policy, 20(3),
395–409
Huang, C. J., & Ho, Y. H. (2017). Governance and economic
growth in Asia. The North American Journal of
Economics and Finance, 39, 260–272. https://doi.org/
10.1016/j.najef.2016.10.010
Kaplan, R. D. (2009). The revenge of geography. Foreign
Policy, (172), 96–105. https://www.jstor.org/stable/
20684874
Lensink, R., & Morrissey, O. (2006). Foreign direct invest-
ment: Flows, volatility, and the impact on growth.
Review of International Economics, 14(3), 478–493.
https://doi.org/10.1111/j.1467-9396.2006.00632.x
Li, X., & Liu, X. (2005). Foreign direct investment and
economic growth: An increasingly endogenous
relationship. World Development, 33(3), 393–407.
https://doi.org/10.1016/j.worlddev.2004.11.001
Mankiw, N. G. (2010). Macroeconomics, (7th Int. ed.)
Mankiw, N. G., Romer, D., & Weil, D. N. (1992).
A contribution to the empirics of economic growth.
The Quarterly Journal of Economics, 107(2), 407–437.
https://doi.org/10.2307/2118477
Marshall, T. (2016). Prisoners of geography: Ten maps that
explain everything about the world (Vol. 1).
Mauro, P. (1995). Corruption and growth. The quarterly
journal of economics, 110(3), 681-712. https://www.
jstor.org/stable/pdf/2946696
Mavrotas, G., Murshed, S. M., & Torres, S. (2011). Natural
resource dependence and economic performance in
the 1970–2000 period. Review of Development
Economics, 15(1), 124–138
Mehlum, H., Moene, K., & Torvik, R. (2006). Institutions
and the resource curse. The Economic Journal, 116
(508), 1–20. https://doi.org/10.1111/j.1468-0297.
2006.01045.x
Murphy, K. M., Shleifer, A., & Vishny, R. W. (1993). Why is
rent-seeking so costly to growth?. The American
Economic Review, 83(2), 409–414
North, D. C. (1990). Institutions, institutional change and
economic performance. Cambridge university press.
Ottaviano, G. I., & Peri, G. (2005). Cities and cultures.
Journal of Urban Economics, 58(2), 304–337. https://
doi.org/10.1016/j.jue.2005.06.004
Pieroni, L. (2009). Military expenditure and economic
growth. Defence and peace economics, 201, 327–
339. https://citeseerx.ist.psu.edu/viewdoc/download?
doi=10.1.1.584.5451&rep=rep1&type=pdf
Robinson, J. A., Torvik, R., & Verdier, T. (2006). Political
foundations of the resource curse. Journal of
Development Economics, 79(2), 447–468. https://doi.
org/10.1016/j.jdeveco.2006.01.008
Rodrik, D. (2000a). How far will international economic
integration go? Journal of Economic Perspectives, 14
(1), 177–186. https://doi.org/10.1257/jep.14.1.177
Rodrik, D. (2000b). Institutions for high-quality growth:
What they are and how to acquire them. Studies in
Comparative International Development, 35(3), 3–31.
https://doi.org/10.1007/BF02699764
Romer, P. M. Endogenous technological change. (1990).
J. Of Political Economy, 98(5), S71–S102. Part 2.
https://doi.org/10.1086/261725
Sachs, J., & Warner, A. (1999). The big push, natural
resource booms and growth. Journal of Development
Economics, 59(1), 43–76. https://doi.org/10.1016/
S0304-3878(99)00005-X
Sachs, J. D. (2003). Institutions matter, but not for
everything. Finance and Development, 40(2), 38–41.
https://www.elibrary.imf.org/view/journals/022/0040/
002/article-A012-en.xml
Sala-i-Martin, X., & Subramanian, A. (2013). Addressing
the natural resource curse: An illustration from
Nigeria. Journal of African Economies, 22(4), 570–615.
https://doi.org/10.1093/jae/ejs033
Solow, R. M. (1956). A contribution to the theory of eco-
nomic growth. The quarterly journal of economics, 70
(1), 65-94. URL: https://www.jstor.org/stable/pdf/
1884513
Swan, T. W. (1956). Economic growth and capital
accumulation. Economic Record, 32(2), 334–361.
https://doi.org/10.1111/j.1475-4932.1956.tb00434.x
Tavares, J., & Wacziarg, R. (2001). How democracy affects
growth. European Economic Review, 45(8), 1341–1378.
https://doi.org/10.1016/S0014-2921(00)00093-3
Tilly, C. (1992). Coercion, capital and the European states,
AD 990-1992. Blackwell.
Van der Ploeg, F. (2011). Natural resources: Curse or
blessing? Journal of Economic Literature, 49(2),
366–420. https://doi.org/10.1257/jel.49.2.366
Wajeetongratana, P. (2020). Economic growth and its key
factors: An alternative view on the factors stimulat-
ing agriculture growth. In E3S Web of Conferences
(Vol. 175, p. 13028). EDP Sciences.
Weber, M. (1905). The Protestant Ethic and the Spirit of
Capitalism, reprinted in 2001 (3rd ed.). London:
Routledge Classics
Wei, S. J. (2000). Local corruption and global capital
flows. Brookings Papers on Economic Activity, 2000
(2), 303–354. https://doi.org/10.1353/eca.2000.0021
Weingast, B. R. (1993). Constitutions as governance struc-
tures: The political foundations of secure markets.
Journal of Institutional and Theoretical Economics
(JITE). http://www.jstor.org/stable/40751603
Zahka, W. J. (1990). The Nobel prize economics lectures
as a teaching tool. The Journal of Economic
Education, 21(4), 395–401. https://doi.org/10.1080/
00220485.1990.10844685
Zhao, R. (2019). Technology and economic growth: From
Robert Solow to Paul Romer. Human Behavior and
Emerging Technologies, 1(1), 62–65. https://doi.org/
10.1002/hbe2.116
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 16 of 17
https://doi.org/10.1016/S0014-2921(01)00127-1
https://doi.org/10.1016/S0014-2921(01)00127-1
https://doi.org/10.1111/j.1467-9701.2006.00807.x
https://doi.org/10.1111/j.1467-9701.2006.00807.x
https://doi.org/10.1257/00028280260344731
https://doi.org/10.1016/j.najef.2016.10.010
https://doi.org/10.1016/j.najef.2016.10.010
https://www.jstor.org/stable/20684874
https://www.jstor.org/stable/20684874
https://doi.org/10.1111/j.1467-9396.2006.00632.x
https://doi.org/10.1016/j.worlddev.2004.11.001
https://doi.org/10.2307/2118477
https://www.jstor.org/stable/pdf/2946696
https://www.jstor.org/stable/pdf/2946696
https://doi.org/10.1111/j.1468-0297.2006.01045.x
https://doi.org/10.1111/j.1468-0297.2006.01045.x
https://doi.org/10.1016/j.jue.2005.06.004
https://doi.org/10.1016/j.jue.2005.06.004
https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.584.5451%26rep=rep1%26type=pdf
https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.584.5451%26rep=rep1%26type=pdf
https://doi.org/10.1016/j.jdeveco.2006.01.008
https://doi.org/10.1016/j.jdeveco.2006.01.008
https://doi.org/10.1257/jep.14.1.177
https://doi.org/10.1007/BF02699764
https://doi.org/10.1086/261725
https://doi.org/10.1016/S0304-3878(99)00005-X
https://doi.org/10.1016/S0304-3878(99)00005-X
https://www.elibrary.imf.org/view/journals/022/0040/002/article-A012-en.xml
https://www.elibrary.imf.org/view/journals/022/0040/002/article-A012-en.xml
https://doi.org/10.1093/jae/ejs033
https://www.jstor.org/stable/pdf/1884513
https://www.jstor.org/stable/pdf/1884513
https://doi.org/10.1111/j.1475-4932.1956.tb00434.x
https://doi.org/10.1016/S0014-2921(00)00093-3
https://doi.org/10.1257/jel.49.2.366
https://doi.org/10.1353/eca.2000.0021
http://www.jstor.org/stable/40751603
https://doi.org/10.1080/00220485.1990.10844685
https://doi.org/10.1080/00220485.1990.10844685
https://doi.org/10.1002/hbe2.116
https://doi.org/10.1002/hbe2.116
© 2021 The Author(s). This open access article is distributed under a Creative Commons Attribution (CC-BY) 4.0 license.
You are free to:
Share — copy and redistribute the material in any medium or format.
Adapt — remix, transform, and build upon the material for any purpose, even commercially.
The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms:
Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made.
You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
No additional restrictions
You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
Cogent Economics & Finance (ISSN: 2332-2039) is published by Cogent OA, part of Taylor & Francis Group.
Publishing with Cogent OA ensures:
• Immediate, universal access to your article on publication
• High visibility and discoverability via the Cogent OA website as well as Taylor & Francis Online
• Download and citation statistics for your article
• Rapid online publication
• Input from, and dialog with, expert editors and editorial boards
• Retention of full copyright of your article
• Guaranteed legacy preservation of your article
• Discounts and waivers for authors in developing regions
Submit your manuscript to a Cogent OA journal at www.CogentOA.com
Tadele & Sirany, Cogent Economics & Finance (2021), 9: 1949133
https://doi.org/10.1080/23322039.2021.1949133
Page 17 of 17
1.1. The objective of this review
2.1. Search engines used
4.1. Conceptual hypothesis related to economic growth
4.1.1. Convergent hypothesis
4.1.2. Institution hypothesis
4.1.3. Geography hypothesis
4.2. The geography of land locked with bad neighbours
4.2.1. Natural resource and its paradox
4.2.2. War and state capacity
5.1. Possible suggestions
Acknowledgements
Funding
Author details
Note
Disclosure statement
References
Higher Test Scores or More Schoo
l
ing? Another Look at the Causes of Economic Growth
Author(s): Theodore R. Breton
Source: Journal of Human Capital , Vol. 9, No. 2 (Summer 2015), pp. 239
–
263
Published by: The University of Chicago Press
Stable URL: https://www.jstor.org/stable/10.1086/681911
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access
to Journal of Human Capital
This content downloaded from
������
������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�����
������
��
All use
subject to h
ttps://about.jstor.org/terms
https://www.jstor.org/stable/10.1086/681911
Higher Test Scores or More Schooling?
Another Look at the Causes
of Economic Growth
Theodore R. Breton
Universidad EAFIT
I use a dynamic augmented Solow model to estimate the effect of international test
scores and investment in schooling and tutoring on economic growth rates in
55 countries during 1985–2005. Either test scores or investment in schooling and
tutoring can explain growth rates in the full data set or in countries that had less
than 8 years of schooling in 1985. In countries with more schooling in 1985, in-
vestment in schooling has a small effect and test scores have no effect on growth
rates. In the 24 countries with scores above 470, higher scores have no effect on
growth rates.
I. Introduction
Analyses of the effect of human capital on national income and growth
rates using aggregate cross-country data are valuable because they estimate
the external as well as the direct effects of human capital ðKrueger and Lin-
dahl 2001Þ. Until relatively recently, these analyses relied almost entirely
on school enrollment rates and average years of schooling to represent the
flow and the stock of human capital in an economy.
In a series of articles, Hanushek and Kimko ð2000Þ and Hanushek and
Woessmann ð2008, 2011a, 2011b, 2012a, 2012bÞ use an innovative measure
of human capital, students’ average scores on international tests, to esti-
mate the effect of human capital on rates of economic growth. They ar-
gue that average test scores provide a much more accurate measure of a
nation’s human capital than adults’ average years of schooling attainment
ðhereafter schooling attainmentÞ.
In all of their articles, Hanushek and Woessmann compare the effect of
test scores and schooling attainment on growth rates and obtain similar
results. Hanushek and Woessmann ð2008, 2012aÞ show that over the pe-
I thank Richard Rogerson, Mikael Lindahl, George Psacharopoulos, Michael Jetter,
Andrew Breton, and four anonymous referees for helpful comments on earlier versions of
t
his manuscript.
[ Journal of Human Capital, 2015, vol. 9, no. 2]
© 2015 by The University of Chicago. All rights reserved. 1932-8575/2015/0902-0004$10.0
0
239
This content downloaded from
������������
216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
riod 1960–2000, average test scores explain three times the variation in
growth explained by schooling attainment ð73 percent vs. 25 percentÞ.
They also show that when test scores and schooling attainment are in-
240 Journal of Human Capital
cluded in the same model, test scores explain all the variation in growth.
They conclude from these results that higher cognitive skills at ages 9–15
cause growth and more schooling often does not.
Breton ð2011Þ challenges the validity of these results. He argues that
Hanushek and Woessmann’s ð2008Þ comparison of the effect of test scores
and schooling attainment is flawed. Since Hanushek and Woessmann
ð2008, 2011a, 2011b, 2012a, 2012bÞ use the same methodology to esti-
mate the effect of these measures, his criticism is applicable to the more
recent analyses as well.
The most evident flaw in the methodology is that Hanushek and Woess-
mann compare the effect of students’ test scores from 1964–2003, and
primarily from 1990–2003, to the effect of adults’ schooling attainment
in 1960. These two measures are not remotely comparable. As an example,
students who were tested at age 9 in 2003 and remained in school until
they were 18 would have entered the workforce in 2012. Because of the lag
between the testing of the students and their entry into the workforce and
their subsequent 40-year wo
rking
life, average test scores from 1964–2003
are a proxy for a country’s human capital in about 2010, or 50 years later
than adults’ schooling attainment in 1960. The average scores from 1990–
2003, which they use for most of the less educated countries, are a proxy
for a country’s human capital around 2020.
The less evident flaw in the methodology is that their growth model is
misspecified. The model includes the initial level of human capital, which
is included in some endogenous growth models, but it also includes initial
income, which is included in dynamic neoclassical models to control for
conditional convergence. The empirical results in Hanushek and Kimko
ð2000Þ and Hanushek and Woessmann ð2008, 2011b, 2012a, 2012bÞ sup-
port the lagged income variable and reject the initial level of schooling.
Hanushek and Woessmann ð2012bÞ include the initial level of physical
capital in the model, and this variable is also rejected. So their results con-
sistently reject the initial levels of capital found in some endogenous
growth models and accept the lagged income variable included in the
dynamic neoclassical growth model.
The capital variables in the dynamic neoclassical growth model are the
flow of capital into the economy during the growth period, not the ini-
tial capital stock ðBreton 2011Þ. The implication is that in the Hanushek-
Kimko/Hanushek-Woessmann model, students’ average test scores at
ages 9–15 during 1964–2003 represent the flow of human capital into the
economy during 1970–2010, or about 6–7 years after the testing period.
The comparable schooling measure is the average rate of enrollment or
the rate of investment in schooling during 1964–2003, not the schooling
attainment of adults in 1960. Their model also lacks an analogous flow
of physical capital into the economy. As a consequence, Hanushek and
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
Kimko’s and Hanushek and Woessmann’s estimates of the effects of test
scores and schooling attainment on growth are likely to be severely biased.
In this paper I reexamine whether higher test scores or more schooling
Higher Test Scores or More Schooling? 241
causes growth, using a dynamic augmented Solow growth model, compa-
rable measures for test scores and investment in schooling, and data for
these measures that are appropriate for the period of estimation.1 I also
examine whether private tutoring affects growth and whether there are
nonlinearities in the education-growth relationship that lead to different
results in the complete data set than in subsets of countries with different
levels of schooling.2 As far as I know, these analyses have not been per-
formed in the existing empirical literature.
I begin my analysis by examining the quantitative relationships between
three measures of a nation’s human capital stock: average adult schooling
attainment, the financial stock of human capital per adult, and students’
average test scores. I examine the relationship between stocks rather than
flows because stocks measure the cumulative effect of flows over a long
period.
I show that while these three measures are correlated, they have very
different patterns across countries, which suggests that they quantify dif-
ferent aspects of a nation’s human capital. The measures increase to-
gether in countries with relatively little schooling, but test scores stabilize
once countries have more than 9 years of schooling attainment or have
invested more than $100,000 per adult ð2005 US$Þ in schooling. As a re-
sult, these measures relate to growth rates differently in countries with
different levels of schooling.
Subsequently, I estimate the effects of higher test scores and more
investment in schooling on growth rates, using Mankiw, Romer, and Weil’s
ð1992Þ dynamic version of the augmented Solow model. This model has a
structure that is compatible with Hanushek and Woessmann’s test score
data and their empirical results, and the validity of this model is supported
by considerable recent empirical evidence ðCohen and Soto 2007; Ding
and Knight 2009; Breton 2010, 2011, 2013b, 2013c, 2015; Gennaioli et al.
2013Þ.3 Since the Mankiw et al. model is a well-defined structural model,
the nature, the form, and the vintage of the data required for its estima-
tion are clearly specified. Since most of Hanushek and Woessmann’s test
scores for less educated countries were obtained after 1990, I estimate the
growth model over the 1985–2005 period to ensure consistency with the
vintage of their data.
1 The flow of human capital into the economy is exogenous in the Solow growth model.
Ehrlich and Kim ð2007Þ specify a complex endogenous growth model in which human cap-
ital determines economic growth.
2 Castelló-Climent ð2010Þ finds evidence that human capital inequality affects rates of in-
vestment in human capital differently in high- and low-income countries.
3 Breton ð2013bÞ challenges Klenow and Rodriguez-Clare’s ð1997Þ and Hall and Jones’s
ð1999Þ arguments that Mankiw et al.’s empirical results overestimate the effect of schooling
on national output.
This content downloaded from
������������216.27.113.144 on ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
I confirm Hanushek and Kimko’s and Hanushek and Woessmann’s
findings that average test scores explain cross-country growth rates quite
well in the complete sample of countries. But I find that investments in
242 Journal of Human Capital
schooling ðand private tutoringÞ also explain growth rates quite well. In
both models the estimated parameters for the augmented Solow model
are consistent with theoretical expectations and with estimates in other
cross-country studies. These results reject Hanushek and Kimko’s and
Hanushek and Woessmann’s findings that more schooling is not reliably
correlated with growth.
Perhaps more importantly, when I analyze the effect of higher test
scores and more investment in schooling in countries with different levels
of schooling, I find that these measures explain growth rates well only in
countries with relatively low levels of schooling and test scores. Average test
scores cannot explain growth rates during 1985–2005 in countries that
had more than 8 years of schooling attainment in 1985 or in countries that
had average test scores over 470. These results call into question Hanu-
shek and Woessmann’s ð2011aÞ claim that raising students’ test scores at
ages 9–15 is an attractive growth strategy for OECD countries. In con-
trast, rates of investment in schooling can explain growth rates in coun-
tries with more than 8 years of schooling, but its estimated effect is smaller
than in countries with less schooling.
The paper is organized as follows: Section II examines the quantitative
relationship between the various measures of human capital. Section III
presents the growth model used in the analysis, and Section IV describes
the data used in this analysis. Section V presents the results. Section VI
presents conclusions.
II. Measures of Human Capital
A country’s human capital is analogous to its physical capital but is much
more difficult to measure. A large fraction of human capital is created
through the formal schooling process, particularly in higher-income coun-
tries, but human capital is also created in informal settings, such as in the
home or on the job. Expenditures on formal schooling or on tutoring can
be measured, but historically such data have not been collected as care-
fully or as regularly as expenditures on physical capital. The earnings that
students forgo while in school are an additional, unmeasured investment
in schooling. And some kinds of schooling are an element of consump-
tion rather than an investment in productive capital.4 Owing to all these
complications, estimates of a country’s rate of investment in human cap-
ital or of its human capital stock inherently have more measurement error
than analogous estimates for physical capital.
4 The United Nations system of national accounts classifies education as an element of
consumption.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
There are three measures of human capital that have been used in cross-
country income and growth studies. The first is the average years of school-
ing attainment of the adult population for the ages 15–64, over 15, or
Higher Test Scores or More Schooling? 243
over 25. The second is the average international score on tests of differ-
ent skills for a student population with ages between 9 and 15. The third
is the net cumulative investment in formal schooling of the population
of working age, assuming a 40-year working life after the completion of
schooling.
Most cross-country growth studies use schooling attainment as a proxy
for a country’s human capital because it is the only quantitative measure
of workers’ skills available for most countries for long historic periods.5
Despite its limitations, this measure has acquired legitimacy because the
effect of an additional year of schooling on workers’ incomes ðthe Mincer-
ian returnÞ is relatively consistent across countries ðPsacharopoulos and
Patrinos 2004Þ. The other two measures are available for many fewer coun-
tries and only for recent time periods.
Growth analyses using the average schooling attainment measure al-
most always utilize the Mincerian log-linear relationship between income
and schooling. In these models, each additional year of schooling has an
exponential effect on income. As a consequence, the marginal contribu-
tion of an additional year of schooling to a nation’s productivity and out-
put is much greater in a country with higher average attainment ðlike Ja-
panÞ than in a country with lower average attainment ðlike PeruÞ. These
models implicitly take into account the higher average investment per year
of schooling and the related higher schooling quality in countries with
higher average schooling attainment.
The main weakness in these analyses is that they implicitly assume that
a year of schooling has the same quality in countries that have the same
schooling attainment, for example, in the United States and Canada. One
indicator of how much schooling quality might vary in countries with the
same schooling attainment is the variation in cumulative investment in
schooling per adult in countries with the same average schooling attain-
ment. Countries that invest more in each year of schooling ðadjusted for
differences in purchasing powerÞ are more likely to provide higher-quality
schooling. The cumulative investment measure of human capital could
capture differences in schooling quality to a greater degree than the av-
erage attainment measure, although there are differences in investment
due to institutional characteristics that are not related to schooling quality.
Figure 1 shows the relationship in 2005 between the log of Breton’s
ð2013aÞ estimates of the financial stock of human capital per adult of work-
ing age and Cohen and Soto’s ð2007Þ estimates of the schooling attain-
ment of the population aged 15–64.6 Breton’s measure of human capital
5 Morrisson and Murtin ð2009Þ present average schooling attainment data for 74 countries
for the period 1870–2010.
6 The estimates of average schooling attainment in 2005 are the average of schooling
a
ttainment in 2000 and 2010.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
is analogous to the standard financial measure of the stock of physical
capital. It is created from the sum of the prior 40 years of investment in
schooling and a depreciation rate of 2.5 percent per year. Since the in-
Figure 1.—Logðhuman capital per adultÞ versus average schooling attainment in 2005
244 Journal of Human Capital
vestment is calculated from national income in Penn World Table 6.3
ðHeston, Summers, and Aten 2009Þ, the estimates of the stocks of human
capital are adjusted for purchasing power differences across countries.
The relationship in the figure is clearly linear, and the two data sets are
highly correlated ðr 5 :91Þ. If a nation’s cumulative investment in school-
ing accounts for the quality of its schooling, then the very high correlation
between the log of human capital per adult and average schooling attain-
ment indicates that a log-linear relationship between income and aver-
age schooling implicitly accounts for the higher average quality of school-
ing in more educated countries.7
The data in figure 1 show that South Korea, Japan, and the United King-
dom have invested less per year of schooling than other highly educated
countries, but their investment in schooling does not include their ex-
penditures on private tutoring, which are substantial ðDang and Rogers
2008Þ. As will be addressed later, stocks or flows of human capital calcu-
lated from investment in schooling are underestimated in countries that
spend considerable amounts on private tutoring.
As also shown in the figure, the differences in the financial measure of
human capital per adult can be quite large in countries with the same av-
7 The trend in the relationship shows that countries with 2 years of schooling in 2005 had
invested about $2,000 per adult, and countries with 13 years of schooling had invested about
$130,000 per adult, or 10 times as much per year of schooling.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
erage schooling attainment, and the range is particularly large in coun-
tries in which average schooling attainment is between 4 and 9 years.
These data suggest that the quality of schooling is much higher in Argen-
Higher Test Scores or More Schooling? 245
tina than in the Philippines and much higher in Costa Rica than in Syria.
Breton ð2013bÞ estimates Mankiw et al.’s ð1992Þ static version of the aug-
mented Solow model across countries in 1990 using the financial stock
of human capital per adult and schooling attainment. Both measures ex-
plain the variation in national income quite well, but the financial mea-
sure explains more of the variation, suggesting that across countries it
accounts for differences in schooling quality somewhat better than the
schooling attainment measure.
If the financial stock of human capital per adult is a more accurate mea-
sure of human capital than average schooling attainment, then it could
be a more accurate measure of human capital than average test scores,
particularly in countries with high average levels of schooling. Figure 2
shows the relationship between Hanushek and Woessmann’s measure of
average test scores and the financial stock of human capital per adult in
2005 in 46 countries. These two measures are correlated ðr 5 :70Þ, but the
mathematical relationship between them is not linear or log linear. The
data show that average test scores at ages 9–15 rise as countries raise their
investment in human capital per adult, but only up to about $100,000 per
adult ð2005 US$Þ. Beyond that level of investment, average scores tend to
decline, although not by a substantial amount.
Figure 3 shows the relationship between Hanushek and Woessmann’s
measure of students’ average test scores obtained over the period 1964–
Figure 2.—Hanushek and Woessmann’s average test scores versus human capital per adult in
2005.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
2006 and the average schooling attainment of the adult population in
1985, the midpoint of the testing period. Average scores on internationa
tests at ages 9–15 increase across countries as the average schooling of the
Figure 3.—Average student test score versus average schooling attainment in 1985
246 Journal of Human Capital
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
l
population aged 15–64 rises to a level of about 9 years, and then scores
stabilize at a mean of about 500.
These patterns suggest that a nation’s average test scores at ages 9–15
and its average schooling attainment are measuring different aspects of
its human capital. Average test scores measure students’ competence in
basic skills, while schooling attainment and the financial stock of human
capital measure the overall educational level of the adult population.
These measures rise together in countries with limited postsecondary
schooling, but they diverge in more educated countries. The data in fig-
ures 2 and 3 demonstrate that the test score measure cannot discern the
differences in human capital in countries with widely differing levels of
postsecondary schooling.
The test score measure also has limitations in countries where many stu-
dents do not complete secondary schooling. In these countries, average
test scores are not representative of the skills of the entire school-age pop-
ulation. As an example, in Hanushek and Woessmann’s data, India has rel-
atively high average scores, but in the testing period, its secondary school
enrollment rate was under 50 percent.8 As a consequence, India’s average
test scores overestimate the skills of the school-age population. Although
there is no way to eliminate this measurement error in countries with low
8 World Bank data used are school enrollment, primary ð% grossÞ, secondary ð% grossÞ,
and tertiary ð% grossÞ.
enrollment rates, it can be minimized by using Hanushek and Woess-
mann’s average test scores to estimate growth during the latest possible
period when a larger share of the school-age population attended second-
Higher Test Scores or More Schooling? 247
ary school.
These patterns in the data suggest that students’ cognitive skills at ages
9–15 are an incomplete measure in countries with a financial stock of
human capital per adult above $100,000 or with more than 9 years of av-
erage schooling attainment. As a consequence, test scores may not be a
sufficiently accurate measure to permit estimation of the effect of hu-
man capital on national income or on economic growth in more educated
countries. Since all measures of human capital have their limitations,
which measure best represents cross-country human capital is an empir-
ical issue that can be determined only in a properly specified income or
growth model.
III. Growth Model Specification
In this analysis I utilize Mankiw et al.’s ð1992Þ augmented Solow model to
compare the effect of higher test scores and more investment in schooling
on national output:
ðY =LÞ
t
5 ðK=LÞa
t
ðH=LÞb
t
ðA0egtÞ12a2b: ð1Þ
In this model, output ðYÞ changes in response to changes in physical cap-
ital ðKÞ, human capital ðHÞ, labor ðLÞ, and total factor productivity ðAÞ,
which is assumed to grow at a constant rate gð1 2 a 2 bÞ. Breton ð2013bÞ
shows that when H=L is defined as the financial stock of human capital
per adult, b 5 0:36 and a 1 b 5 0:7. His results support Mankiw et al.’s
assumption that a 1 b < 1, and they are consistent with Mankiw et al.’s
results, in which human capital has large external effects on national in-
come.
Mankiw et al. ð1992Þ derive a dynamic version of their model in which
economic growth is modeled as convergence to the steady state yt 5 y*,
where yt 5 Y =ðegtLÞ and l is the rate of convergence to the steady state:
log ðytÞ 2 log ðy0Þ 5 ð1 2 e2ltÞlog ðy*Þ 2 ð1 2 e2ltÞlogðy0Þ: ð2Þ
They show that y* is a function of the shares of GDP invested in physical
and human capital ðsk and shÞ, the labor growth rate ðnÞ, and the capital
depreciation rates ðdk and dhÞ:
y* 5 a=ð1 2 a 2 bÞ½logðskÞ=ðn 1 g 1 dkÞ�
1 b=ð1 2 a 2 bÞ½logðshÞ=ðn 1 g 1 dhÞ�:
ð3Þ
Substitution of equation ð3Þ into equation ð2Þ and rearrangement creates
a growth model, which contains a lagged income variable, similar to the
variable in the Hanushek-Kimko and Hanushek-Woessmann analyses:
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
log ðY =LÞ
t
2 logðY =LÞ
0
5 c 1 ð1 2 e2ltÞa=ð1 2 a 2 bÞ½logðskÞ=ðn 1 g 1 dkÞ�
248 Journal of Human Capital
1 ð1 2 e2ltÞb=ð1 2 a 2 bÞ½logðshÞ=ðn 1 g 1 dhÞ�
2 ð1 2 e2ltÞlogðY =LÞ
0
1 ε:
ð4Þ
When this model is estimated over a period 0 to t, sk, sh, and n are the
averages of these rates during the growth period. The shares of invest-
ment sk and sh measure the flow of physical and human capital resources
into the economy during this period.
The average test score for a cohort of students aged 9–15 can be
employed as a measure of the human capital flow into the economy in
each country 5–10 years later. Hanushek and Woessmann’s ð2012aÞ data
for average test scores are based on international tests taken between 1964
and 2003, but as described below, most of the scores in the less educated
countries were obtained between 1990 and 2003. As a consequence, their
average scores for developed countries and a few developing countries are
representative of the flow of human capital during 1970–2010, while most
of their average scores for developing countries are representative of the
flow of human capital into the workforce during 1995–2010.
I estimate Mankiw et al.’s growth model over the 1985–2005 period.
This period corresponds relatively well to the period when most of the test
scores were obtained and certainly much better than the earlier 1960–
2000 period that Hanushek and Woessmann use in their analyses. The test
scores also are more representative of the flow of human capital in the less
educated countries in the later period because a much higher fraction of
students remained in school until age 15 in these countries in 1985 than in
1960. As a consequence, there is less measurement error in Hanushek and
Woessmann’s test score measure when it is used to explain growth during
1985–2005 than when it is used to explain growth during 1960–2000.
Figure 4 shows the relationship between average test scores and logðshÞ
in the data set. Although the correlation between these two data sets is not
very high, the pattern in the data indicates that the relationship between
investment in schooling and average test scores could be log-linear. I rep-
resent logðshÞ in equation ð4Þ by the average test score rather than by the
log of the average test score because it provides better results. Hanushek
and Woessmann ð2012bÞ use a linear-exponential relationship between
growth rates and test scores in their analyses for the same reason.
As mentioned earlier, a limitation of the rate of investment in schooling
measure is that some countries expend considerable resources on private
tutoring to improve students’ cognitive skills. Since rates of investment in
schooling do not include these expenditures, they underestimate the rate
of investment in human capital in these countries, and as shown later in
the empirical results, the failure to include the tutoring expenditures in
the growth model substantially changes the estimated effect of investment
This content downloaded from
�������ff:ffff:ffff:ffff:ffff:ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
in schooling. Since cross-country time-series data on expenditures on tu
toring are not available, I control for the effect of tutoring by including
a dummy variable for countries with high expenditures on tutoring.
Mankiw et al.’s growth model assumes that investment in physical capita
Figure 4.—Average test score versus logðinvestment in schooling/GDPÞ
Higher Test Scores or More Schooling? 249
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
–
Figure 5 shows the relationship between the growth rate over the 1985–
2005 period and the average test score variable for the 55 countries in the
data set. The scores exhibit regional patterns that could indicate that un-
known factors affected growth rates. Average test scores are relatively high
in the Asian countries and relatively low in the Latin American countries.
I include dummy variables for these regions in some models to test for
possible omitted variables.
IV. Data Creation and Selection
l
an
d
human capital, growth in workers, and the initial level of income are
the only factors that affect growth. The model also assumes that factors
of production are paid their marginal products. Countries that have cen-
trally planned economies, have income largely determined by oil exports,
have serious civil conflicts, or are tax havens have characteristics affecting
income and growth rates that are not included in the model. These coun-
tries are likely to be outliers in the model’s growth-investment relation-
ships. Particularly in estimates of the model with small data samples, these
outliers can substantially bias the estimated coefficients and reduce their
statistical significance.
Hanushek and Kimko and Hanushek and Woessmann had similar con-
cerns when they estimated their growth models, so they excluded many
countries from their data sample. Hanushek and Woessmann ð2008Þ re-
port that 77 countries participated in international tests of mathematics
andscienceduring the 1964–2003period.Theyexcluded 27ofthese coun-
Figure 5.—Economic growth rate in 1985–2005 versus test the score variable in the growth
model.
250 Journal of Human Capital
tries from their sample because 15 were communist countries; three were
predominantly oil exporters; six were small, were newly created, or lacked
output data; and two were outliers in their growth regressions. Their re-
maining data set includes 50 countries.
Hanushek and Woessmann ð2012bÞ create average scores for an addi-
tional nine Latin American countries based on regional tests of mathe-
matics and reading skills in fourth and sixth grades taken during 1997 and
2006. These scores are less reliable since they correspond to tests of dif-
ferent subjects, correspond to a later period, and had to be adjusted to
merge them with the international scores. I use most of these scores, but I
test whether their use changes the regression results.
I began with these 59 countries and excluded four for the same reasons
as Hanushek and Woessmann. I excluded China and Romania because
they were communist countries and Ven
ezuela
because it was predomi-
nantly an oil exporter.9 I excluded Jordan because it is an outlier in the
growth regressions as a result of the heavy migration of refugees from
Israel and Iraq during the 1985–2005 period. Since these refugees add
human capital to the labor force that is not measured in the test scores
or the investment rates, the inclusion of Jordan in the data set would bias
9 China’s average test scores correspond to tests taken in Shanghai, which is a relatively
educated region.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
the results. So my initial estimates of the effect of test scores on growth are
based on the scores in 55 countries.
Hanushek and Woessmann’s international test scores are the simple
Higher Test Scores or More Schooling? 251
arithmetic average of any available scores on tests of mathematics and sci-
ence for students between 9 and 15 during the 1964–2003 period. The age
distribution of the students tested may be different in each country, but
they argue thatthisisnot a problem because scores ondifferent testswithin
the same country are highly correlated ðHanushek and Woessmann 2008,
2012aÞ. Another concern is that the students participating in international
tests are not always a representative sample of the school population in the
less educated countries. But Hanushek and Woessmann ð2011bÞ present
analyses showing that sample selectivity problems have not unduly biased
their results.
Since the international tests began as a means to compare students’
skills in the more educated countries, there are few scores for less edu-
cated countries prior to 1990. My data set with test scores includes 18 highly
educated countries and 37 less educated countries. Only seven of the less
educated countries have test scores prior to 1990 ðHanushek and Woess-
mann 2008Þ, and eight of the 37 countries have scores only for the period
1997–2006 ðHanushek and Woessmann 2012bÞ.
The shares of GDP invested in physical capital and human capital are
conceptually identical in the growth model, but obtaining estimates of the
investment rate for human capital is much more difficult. Time-series data
on the investment/GDP ratio in non-OECD countries are available only
for public schooling and are intermittent or unreliable in many countries.
In addition, there is a considerable lag between the investment in a stu-
dent’s schooling and the student’s entry into the workforce. This lag var-
ies across countries, depending on the amount of schooling provided, the
structure of the economy, and practices related to child employment.
Forthe averagerateof investment inschoolingvariableðshÞ during1985–
2005, I use the average rate of investment during 1980–2000. I use an
investment period that is 5 years earlier than the growth period to account
for the delay between the expenditures on students’ schooling and the
entry of these students into the workforce.10
Breton ð2013bÞ estimates human capital per adult in 1990 for 61 mar-
ket economies from the shares of GDP invested in schooling from 1950
to 1985, using UNESCO data on expenditures for public education ðper-
centage of GDPÞ, increased by factors to account for private schooling, the
opportunity cost of capital while students are in school, and students’ for-
gone earnings. I use the data elements from these estimates to calculate
10 Five years is a reasonable average lag from a financial standpoint since unit schooling
costs rise with the level of schooling and the delay between expenditures and entry into the
workforce is shorter at higher levels of schooling.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
the average rate of investment in schooling, but with additional UNESCO
data on expenditures in public education for the period 1990–2000.11
Dang and Rogers ð2008Þ survey the extent of private tutoring in 23
252 Journal of Human Capital
countries, including estimates of either total expenditures or shares of the
student population that participate in private tutoring. I include a dummy
variable for tutoring expenditures in the eight countries that expend at
least 0.5 percent of GDP on tutoring or that provide tutoring to at least
25 percent of the student population ðEgypt, Greece, Hong Kong, Japan,
South Korea, Singapore, Turkey, and the United KingdomÞ.
Hanushek and Woessmann’s ð2012a, 2012bÞ data provide test scores for
47 of the 61 countries in Breton’s data set. After Jordan is excluded, the
sample of countries with data for both test scores and investment rates is
reduced to 46. These countries provide the basis for the initial comparison
of the effect of average test scores and investment in schooling on growth
rates.
I use Cohen and Soto’s ð2007Þ data on average schooling attainment
in the population aged 15–64 to calculate average attainment in 1985, and
I then separate the countries into subsets with more and less than 8 years
of schooling attainment at the beginning of the growth period. Four of the
55 countries with test score data ðHong Kong, Iceland, Israel, and TaiwanÞ
are not included in Cohen and Soto’s average schooling attainment data.
I estimate average attainment for these countries from Barro and Lee’s
ð2013Þ data on average attainment in the population over 15.
I use economic data from Penn World Table 6.3 ðHeston et al. 2009Þ. I
use the population over 15 as the proxy for workers, which I estimate from
data on GDP per capita ðrgdpchÞ and GDP per equivalent adult ðrgdpqaÞ.
I then calculate n from the growth in this population over the 1985–2005
period. I use the average investment rate ðciÞ over the period 1985–2004
as the investment share sk during the growth period. I assume g 5 0.01,
dh 5 0:025, and dk 5 0:06. The rate g is the average rate derived from the
Solow residual during 1910–2000 in Breton ð2013cÞ. The depreciation rate
for human capital is based on a 40-year work life, as described in Breton
ð2013bÞ. The depreciation rate for physical capital is from Caselli ð2005Þ.
The data used in the models are shown in the Appendix.
V. Empirical Results
The mathematical structure of the Mankiw et al. ð1992Þ model implies that
a and b are the shares of national income that accrue to the stock of
physical capital ðKÞ and human capital ðHÞ, and the rate of income
convergence l is mathematically related to the values of a, b, n, g, dk, and
dh. One of the desirable features of their model relative to unstructured
models, such as those specified by Hanushek and Woessmann, is that the
11 I use the investment/GDP ratio in 1980, 1985, 1990, 1995, and 2000 to estimate the
average ratio in each 5-year period and then average these four ratios to obtain the 20-year
average.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
validity of the Mankiw et al. model can be evaluated on the basis of whether
the estimated parameters of the model are consistent with its theoretical
predictions.
Higher Test Scores or More Schooling? 253
Table 1 presents ordinary least squares ðOLSÞ estimates of the growth
model in equation ð4Þ and shows the implied values of a, b, and l in the
estimated coefficients. The first two columns in the table present the mod-
el’s results with the test score data. Column 1 shows the effect with the 55-
country data set, and column 2 shows the effect with the 46-country data
set used to estimate the effect of investment in schooling in columns 3–5.
The estimated coefficients for the two models with the test score mea-
sure are very similar, have estimated coefficients that are highly statistically
significant, and have implied parameters for the effect of physical capital,
human capital, and the rate of convergence that are consistent with ex-
pectations for the Mankiw et al. model. The implied values of a are 0.35–
0.37, which are very consistent with Bernanke and Gurkaynak’s ð2001Þ and
Gollin’s ð2002Þ estimates of the share of national income accruing to phys-
ical capital, which is about 0.35 across countries. The implied values of
b are 0.27, which is consistent with, but somewhat lower than, Breton’s
ð2013bÞ estimates. The implied values of l, the rate of convergence to the
steady state, range from 0.016 to 0.018, which are consistent with theoret-
ical expectations and with Barro’s ð2012Þ estimates of actual average con-
vergence rates ð1.7 percentÞ in 80 countries since the 1960s. Although not
shown, the calculated parameter values are statistically significant at the
1 percent level.
TABLE 1
Effect of Human Capital Measures on Growth Rates, 1985–2005
Dependent Variable D logðGDP per AdultÞ
Test Scores Investment/GDP Both
ð1Þ ð2Þ ð3Þ ð4Þ ð5Þ ð6Þ
Countries 55 46 46 46 46 44
ln½sk=ðn 1 g 1 dkÞ� .29* .28** .51** .46** .31** .31**
ð.12Þ ð.09Þ ð.10Þ ð.07Þ ð.08Þ ð.08Þ
ln½sh=ðn 1 g 1 dhÞ� .21* .30** .13 .21*
ð.10Þ ð.07Þ ð.09Þ ð.09Þ
Tutoring dummy .22** .11 .05
ð.08Þ ð.08Þ ð.08Þ
ln ½exptest=ðn 1 g 1 dhÞ� .22** .22** .18** .16**
ð.04Þ ð.03Þ ð.05Þ ð.05Þ
lnðY/L 2 1985Þ 2.28** 2.30** 2.29** 2.32** 2.34** 2.37**
ð.05Þ ð.04Þ ð.08Þ ð.06Þ ð.06Þ ð.06Þ
R2 .49 .60 .42 .52 .62 .64
Implied a .37 .35 .50 .43 .32 .30
Implied b .27 .27 .21 .28 .32 .35
Implied l .016 .018 .017 .019 .021 .023
Note.—Robust standard errors are in parentheses.
* Statistically significant at the 5 percent level.
** Statistically significant at the 1 percent level.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
Columns 3 and 4 present the OLS estimates of the effects of investment
in schooling and private tutoring on growth rates. The results in column 3
without the tutoring dummy are all statistically significant and acceptable
254 Journal of Human Capital
conceptually except the value of a, which is too high. When the tutoring
dummy is added in column 4, the results become acceptable. All the co-
efficients, including the coefficient on the dummy variable, are statistically
significant at the 1 percent level.
The results in columns 2 and 4, which are based on the same 46 coun-
tries, show that the model with test scores and the model with investment
in schooling and tutoring provide very similar empirical results, although
the model with test scores explains somewhat more of the variation in
growth rates ðR2 5 .60 vs. .52Þ.
Columns 5 and 6 show the results when test scores and investment in
schooling ðand tutoringÞ are included in the same model. In column 5 the
effects of investment in schooling and tutoring are positive, but only the
effect of test scores is statistically significant. The effect of tutoring is only
half as large when test scores are included, indicating that tutoring affects
growth in part through its effect on test scores.12
In the 46-country sample, Hong Kong and Singapore are outliers in that
they have high economic growth rates but relatively low rates of investment
in schooling. Hong Kong became a Special Administrative Region of
China in 1997, with additional legal protection for private investment des-
tined for mainland China. Singaporeis considered a tax haven. Sothe high
growth rates in these jurisdictions may be due in part to the reporting of
income that is earned elsewhere.
Column 6 shows the results for both measures in a 44-country sample
that excludes Hong Kong and Singapore. In these results the effect of
both test scores and investment in schooling is large and statistically sig-
nificant. The implied values of the parameters in these models continue
to be consistent with the expectations for the Mankiw et al. model.
Table 2 presents additional tests of the same models. Columns 1 and 4
show the results when dummy variables for the Asian and Latin American
regions are included in the model. Hanushek and Woessmann ð2012bÞ
show that the effect of a Latin America dummy is negative in models that
included adult schooling attainment in 1960, and they argue that this is
due to the low quality of schooling in this region. In the results with the
investmentinschoolingmeasureðcol.4Þ,thecoefficientontheLatinAmer-
ica dummy is negative, but it is small and not statistically significant, sug-
gesting that Hanushek and Woessmann’s ð2012bÞ results were due to the
misspecification of their growth model. The effect of the Asia dummy is
larger and positive but not statistically significant with both measures. The
estimated coefficients on the physical capital and human capital variables
12 Alternatively, large investments in tutoring may be an indicator that the educational
system is test based. If students work harder to raise their skills in these countries, it may be
that the coefficient on tutoring is capturing the effect of the additional effort students
e
xpend in a test-based system rather than just the effect of the tutoring.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
ontinue to be statistically significant, and the parameters continue to be
cceptable when the regional dummies are included in the models.
Column 2 shows the results for the test score measure when the Latin
TABLE 2
Sensitivity Tests: Effect of Human Capital Measures on Growth Rates, 1985–2005
Dependent Variable: D lnðGDP per AdultÞ
Test Scores Investment/GDP
ð1Þ ð2Þ ð3Þ ð4Þ ð5Þ
ountries 55 46a 59 46 47
½sk=ðn 1 g 1 dkÞ� .24* .22* .33* .34** .53**
ð.118Þ ð.10Þ ð.13Þ ð.10Þ ð.09Þ
½sh=ðn 1 g 1 dhÞ� .27* .30**
ð.10Þ ð.07Þ
utoring dummy .15 .22**
ð.09Þ ð.08Þ
½exptest=ðn 1 g 1 dhÞ� .21** .24** .21**
ð.05Þ ð.06Þ ð.04Þ
ðY/L 2 1985Þ 2.22** 2.28** 2.29** 2.24* 2.35**
ð.07Þ ð.05Þ ð.06Þ ð.09Þ ð.07Þ
atin America dummy .05 2.04
ð.09Þ ð.09Þ
sia dummy .14 .14
ð.10Þ ð.12Þ
2 .51 .53 .51 .55 .54
plied a .33 .30 .40 .40 .45
plied b .29 .33 .25 .32 .27
plied l .012 .016 .017 .014 .022
ote.—Robust standard errors are in parentheses.
Excludes Latin America.
Statistically significant at the 5 percent level.
* Statistically significant at the 1 percent level.
Higher Test Scores or More Schooling? 255
c
a
C
ln
ln
T
ln
l
n
L
A
R
Im
Im
Im
N
a
*
*
American countries whose scores were calculated from regional tests are
removed from the data set. These results are almost identical to the results
for the full 55-country data set ðcol. 1 of table 1Þ. Column 3 shows the
model results for Hanushek and Woessmann’s complete 59-country data
set, including the four countries I had eliminated. The effect of test scores
is smaller, but the results continue to provide acceptable, statistically sig-
nificant parameters. Column 5 shows the model results for the investment
in schooling measure when Jordan is included in the data set. The esti-
mated parameters are similar to the parameters in the 46-country data set,
but the effect of investment in schooling is slightly smaller.
The data patterns for the human capital measures in figures 2, 3, and 4
suggest that the estimated effects of test scores in table 1 could be the
average of different effects in countries with high and low levels of school-
ing. Since test scores do not continue to rise once countries achieve 9 years
of schooling attainment or a financial stock of human capital of $100,000
per adult, growth in the more educated countries may be caused by more
investment in schooling rather than by increases in test scores.
To investigate this possibility, I separate the countries into two groups
with less and more than 8 years of schooling attainment in 1985. I split the
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
groups at 8 rather than 9 years to increase the sample size of the more ed-
ucated countries. I use schooling attainment in 1985 to divide the coun-
tries because the flow measures during the growth period do not provide
256 Journal of Human Capital
any indication of the level of education in the countries at the beginning
of the growth period.
Table 3 presents the results for three subsets of the countries, those with
less and more than 8 years of schooling in 1985 and those with test scores
above 470 during 1964–2003. The results in the two subsets of countries
with less and more schooling attainment show that the effects of higher
test scores and more investment in schooling in the complete data set are
due to the effects in countries that had less than 8 years of schooling in
1985. The implied values of the parameters are acceptable in this subset
of countries, and all the human capital measures continue to be highly sta-
tistically significant.
The results for the more educated countries in columns 4 and 5 are very
different. None of the measures of human capital have any statistical sig-
nificance. The 24 countries with more than 8 years of schooling have an
average test score of 498, with a range from 405 to 545, so there should be
enough variation to explain growth rates if there were a strong relation-
ship. As shown in column 4, there is no evidence that higher test scores
affected growth rates in these countries during the 1985–2005 period.
There are 20 countries with more than 8 years of schooling in 1985 that
had data on rates of investment in schooling, and again there is no evi-
dence that these rates affected growth rates.
The growth model explains about 60 percent of the variation in growth
rates in these countries, but this variation is explained by the rate of in-
come convergence, which is quite rapid. Since test scores and investment
rates have no effect, the convergence effect is absolute rather than condi-
tional. In these two subsets of countries, the results show slow conditional
income convergence in the less educated countries and rapid absolute
income convergence in the more educated countries.
In such a small data set, outliers can seriously affect the results. A review
of the residuals in the regressions with investment rates reveals that three
countries, Hong Kong, Ireland, and Norway, are outliers. Hong Kong has
an unusual legal status in China, Ireland is a tax haven for companies in
the European Union, and Norway’s GDP is substantially affected by oil
prices. These characteristics raised reported GDP growth rates beyond
what can be explained by the variables in the model.
Columns 6 and 7 show the results when a dummy variable is included
to control for the omitted factors contributing to higher growth rates in
these three countries. In these two models the rate of investment in school-
ing is statistically significant at the 5 percent level, but the effect is rela-
tively small and investment in physical capital still has no effect. In col-
umn 7 the effect of test scores on growth rates continues to be small and
insignificant.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
T
A
B
L
E
3
E
f
f
e
c
t
o
f
H
u
m
a
n
C
a
p
i
t
a
l
M
e
a
s
u
r
e
s
o
n
G
r
o
w
t
h
R
a
t
e
s
,
1
98
5
–
20
05
D
e
p
e
n
d
e
n
t
V
a
r
i
a
b
l
e
:
D
l
n
ðG
D
P
p
e
r
A
d
u
lt
Þ
S
ch
o
o
li
n
g
<
8
Ye
a
rs
S
ch
o
o
li
n
g
>
8
Ye
a
rs
T
e
st
S
co
re
>
4
7
0
ð8
Þ
ð1
Þ
ð2
Þ
ð3
Þ
ð4
Þ
ð5
Þ
ð6
Þ
ð7
Þ
C
o
u
n
tr
ie
s
3
1
2
6
2
6
2
4
2
0
2
0
2
0
2
4
ln
½s k
=
ðn
1
g
1
d
k
Þ�
.2
7
*
.4
6
*
*
.2
5
*
*
2
.0
4
.0
4
.0
1
2
.0
5
.0
8
ð.1
4
Þ
ð
.0
7
Þ
ð.0
5
Þ
ð
.1
7
Þ
ð.1
8
Þ
ð.1
5
Þ
ð.1
3
Þ
ð.1
8
Þ
ln
½s h
=
ðn
1
g
1
d
h
Þ�
.3
6
*
*
.2
0
2
.1
6
.1
2
*
.1
1
*
ð.1
0
Þ
ð.1
1
Þ
ð.1
6
Þ
ð.0
4
Þ
ð.0
4
Þ
T
u
to
ri
n
g
d
u
m
m
y
.3
2
*
*
.1
7
*
2
.0
5
ð.1
0
Þ
ð.0
7
Þ
ð.1
1
Þ
ln
½ex
p
te
st
=
ðn
1
g
1
d
h
Þ�
.2
2
*
*
.2
1
*
*
.0
7
.0
6
2
.1
4
ð.0
6
Þ
ð.0
4
Þ
ð.0
7
Þ
ð.0
8
Þ
ð.1
7
Þ
H
ig
h
-i
n
co
m
e
d
u
m
m
y
.3
2
*
*
.3
2
*
*
ð.0
5
Þ
ð.0
5
Þ
ln
ðY
/
L
2
1
9
8
5
Þ
2
.2
1
*
*
2
.3
0
*
*
2
.2
6
*
*
2
.5
2
*
*
2
.4
8
*
*
2
.5
6
*
*
2
.5
9
*
*
2
.5
4
*
*
ð.0
6
Þ
ð.0
7
Þ
ð.0
4
Þ
ð.0
7
Þ
ð.1
0
Þ
ð.0
9
Þ
ð.1
1
Þ
ð.0
8
Þ
R
2
.4
9
.6
3
.8
2
.6
4
.5
9
.8
6
.8
6
.6
6
Im
p
li
e
d
a
.3
9
.4
1
.2
7
.0
1
Im
p
li
e
d
b
.3
1
.3
2
.4
5
.1
7
Im
p
li
e
d
l
.0
1
2
.0
1
8
.0
1
5
.0
3
7
.0
3
3
.0
4
1
.0
4
5
.0
3
9
N
o
te
.—
R
o
b
u
st
st
an
d
ar
d
e
rr
o
rs
ar
e
in
p
ar
e
n
th
e
se
s.
*
S
ta
ti
st
ic
al
ly
si
g
n
ifi
ca
n
t
at
th
e
5
p
e
rc
e
n
t
le
ve
l.
*
*
S
ta
ti
st
ic
al
ly
si
g
n
ifi
ca
n
t
at
th
e
1
p
e
rc
e
n
t
le
ve
l.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
Column 8 shows the effect of test scores on growth in countries with av-
erage scores above 470. Again what we see in the results is absolute con-
vergence in income levels, regardless of the level of test scores. Since the
258 Journal of Human Capital
data sets for highly educated countries are so small ð20–24 countriesÞ,
these results cannot be considered definitive, but they call into question
Hanushek and Woessmann’s ð2011aÞ claim that highly educated OECD
countries can raise their growth rates by raising their students’ average test
scores.
It is instructive to examine why Hanushek and Woessmann ð2011aÞ
found a positive effect from test scores on growth in 24 OECD countries
during 1960–2000, while this analysis does not find this effect in two
slightly different sets of 24 countries during 1985–2005. There are several
reasons for the different results, but two stand out. First, two OECD coun-
tries with high test scores, Japan and Switzerland, had much lower growth
rates during 1985–2005 than during 1960–2000. Second, Mexico and Tur-
key, two OECD countries with low test scores and low growth rates during
1960–2000, are not included in the current analysis, while one non-OECD
country, Chile, which had low test scores and a high growth rate during
1985–2005, is included in this analysis. In such small data sets, these
changes are sufficient to completely change the statistical relationships in
the results.
Figure 6 shows the relationship between growth rates in 1985–2005 and
test scores for the 28 countries included in either of these analyses. An ex-
amination of the data in this figure reveals that the more educated OECD
Figure 6.—Economic growth rates versus average test scores during 1985–2005
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
countries, excluding South Korea, had similar growth rates during this
period, which were not related to their average test scores. This same pat-
tern is evident in figure 2 in Hanushek and Woessmann ð2011aÞ during
Higher Test Scores or More Schooling? 259
the 1960–2000 period.In these analyses the statistical relationship between
growth rates and test scores is very sensitive to the inclusion or exclusion of
certain countries that are outliers relative to the traditional OECD group of
highly educated countries.
There is a possible explanation for the lack of correlation between test
scores and growth rates in countries with scores above 470. Experiments
with students at different grade levels in the more educated countries show
that average scores on the same international tests rise by about 32 points
after students complete an additional year of schooling ðWoessmann 2003;
Jürges and Schneider 2004; Fuchs and Woessmann 2007Þ. The implication
is that intensive efforts to raise students’ scores in some of the more
educated countries accelerate by 1–2 years the increase in students’ skills
that otherwise occurs as students continue their schooling. It is not clear
whether the skill advantage at ages 9–15 in the countries with higher aver-
age scores continues later or whether it diminishes with time. Since there
is no noticeable effect of scores above 470 on economic growth in the
results, the skill advantage may be temporary. Alternatively, it may be that
in countries with average scores of at least 470, there are enough students
with high skills to meet the economy’s requirement for highly skilled
workers.
The lack of any effect from test scores on growth rates in the more ed-
ucated countries is not surprising given the data patterns in figures 2 and
3, but the small or negligible effect of investment in physical capital and
human capital is unexpected. It appears that in these countries differences
in other factors not included in the model had a larger effect on reported
growth rates during the 1985–2005 period than differences in capital in-
vestment rates.
VI. Conclusions
Hanushek and Woessmann argue that students’ cognitive skills at ages 9–
15, as measured on international tests, determine a nation’s rate of eco-
nomic growth, and Hanushek and Woessmann ð2008, 2012aÞ show that
increased schooling attainment explains only one-third of the variation in
growth rates explained by higher average test scores. Breton ð2011Þ argues
that their results are severely biased because their methodology is flawed.
In this paper I reexamine the effects of higher test scores and more
schooling on growth rates using a dynamic neoclassical growth model, con-
ceptually appropriate measures of schooling, and a time period more ap-
propriate for the vintage of the test scores. I find that in the complete data
set, either average test scores or more investment in schooling can explain
growth rates over the 1985–2005 period. Test scores explain more of the
variation in growth rates, but the variation explained by the two measures
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
is similar once the effect of private tutoring is taken into account. These
results are consistent with Hanushek and Woessmann’s finding that in-
creases in students’ test scores cause growth, but they reject their finding
260 Journal of Human Capital
that increases in schooling do not reliably cause growth.
When I examine the effect of higher test scores and more investment in
schooling in subsets of countries with low and high levels of schooling
attainment, I find that the effect of these measures during 1985–2005
occurs almost entirely in countries that had schooling attainment below
8 years at the beginning of the growth period. In these countries, either
higher test scores or more investment in schooling and private tutoring
explains a high share of the variation in economic growth rates.
I find that countries that expend considerable resources on private tu-
toring have higher growth rates. The results indicate that investment in
schooling and private tutoring are substitutes for raising students’ cogni-
tive skills and for increasing growth rates in countries with average school-
ing attainment below 8 years. More research should be undertaken to de-
termine whether it is the substantial private tutoring or the greater focus
on testing ðor bothÞ in these countries that raises the scores.
In contrast, I find no evidence that increases in average test scores affect
growth rates in countries with more than 8 years of schooling or in coun-
tries with average scores above 470. These results call into question Ha-
nushek and Woessmann’s ð2011aÞ argument that OECD countries can
raise their growth rates by increasing students’ cognitive skills at ages 9–15.
I find some evidence that more investment in schooling raises growth
rates in countries with more than 8 years of schooling, but the effect is
smaller than in the less educated countries. Over the 1985–2005 period,
income per adult in these countries tended to converge. Countries with
lower income per adult grew faster, regardless of their rates of investment
in physical capital and schooling or their level of test scores. The small ef-
fect of human capital in these results could be due to the diminishing re-
turns to investment in human capital or to the failure of the human capital
measures to adequately represent the human capital characteristics that
are most relevant in highly educated countries.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
Appendix
TABLE A1
Data Used in the Analysis
Country dlnya lnskngdk testngdh lnshngdh lnya85 attain85 Score
Argentina .2116 .6712 6.9127 .7861 9.6127 7.61 3.920
Australia .3970 1.2473 8.0840 1.4971 10.2663 12.48 5.094
Austria
Belgium
Boli
Braz
Can
Chil
Chin
Colo
Cost
Cyp
Den
Ecu
Egyp
El S
Finl
Fran
Gha
Gre
Gua
Hon
Hon
Icela
Indi
Indo
Iran
Irela
Israe
Italy
Japa
Jord
Kore
Mal
Mex
Mor
Neth
New
Nor
Pan
Para
Peru
Phil
Port
Rom
Sing
Sou
Spai
Swe
Swit
Taiw
Tha
Tun
������
.3683
.4098
������216.27
All
1.3124
1.2994
This con
.113.144 on
use subject
8.2997
8.2981
tent downlo
Sun, 06 Feb
to https://abo
1.4167
aded from
2022 23:48:
ut.jstor.org/t
10.2240
10.1438
53 UTC�����
erms
10.63
9.64
��������
5.089
5.041
via
.1302
.1007
5.4342
.8415
8.5179
6.65
2.640
il
.0354
.4415
6.4922
.8347
9.3914
5.40
3.638
ada
.3409
1.1239
8.0651
1.5807
10.3044
11.98
5.038
e
.8398
.9403
6.9869
1.2188
9.1898
8.66
4.049
a
1.3197
1.1755
7.8849
7.6977
4.939
mbia
.1829
.3992
6.9890
.6091
9.0548
5.46
4.152
a Rica
.2895
.8209
7.2695
.9067
9.3285
5.30
4.486
rus
.6528
1.3632
7.5840
9.6342
7.57
4.542
mark
.3742
1.2996
8.2369
1.8099
10.2123
11.29
4.962
ador
2.0464
.7537
5.6132
.6579
9.1042
6.73
2.852
t
.4240
2.1135
6.7946
.8423
8.53
88
3.94
4.030
alvador
.1159
.5338
6.0315
.3590
8.9121
4.07
3.243
and
.3873
1.3393
8.3591
1.4584
10.1044
10.11
5.126
ce
.3108
1.1664
8.2062
1.5522
10.1602
11.46
5.040
n
a
.1311
2.3988
6.3146
.3443
7.7003
4.59
3.603
ece
.3607
1.2178
7.7638
.6352
9.9396
8.22
4.608
temala
.0751
.6498
5.6343
2.0352
9.1219
3.29
2.855
duras
2.0713
.8884
5.1193
8.7115
4.37
2.453
g Kong
.6013
1.1204
8.1316
.5685
10.0968
9.78
5.185
nd
.4277
1.2920
7.9705
10.3711
9.35
4.936
a
.6280
.4913
7.1411
.5421
7.8853
2.88
4.281
nesia
.5324
.7450
6.72
47
8.3088
4.89
3.880
.1959
.9780
6.9689
.5814
9.2600
3.01
4.219
nd
.9117
1.2394
8.0508
1.2472
9.8821
9.24
4.995
l
.3280
1.1021
7.4336
10.0175
11.68
4.686
.3051
1.3504
8.0235
1.2807
10.0810
8.53
4.758
n
.2512
1.5091
8.4690
1.2355
10.1984
11.57
5.310
an
2.4545
.3033
6.7212
.8810
9.3464
4.264
a, Republic
.9410
1.5838
8.3232
.9367
9.2732
9.52
5.338
aysia
.7915
1.0116
7.6196
.9177
9.3186
7.10
4.838
ico
.0306
.8584
6.8286
.7073
9.6084
6.48
3.998
occo
.0539
.4238
6.0775
.9415
8.8684
1.96
3.327
erlands
.4094
1.1822
8.2874
1.4458
10.1867
10.50
5.115
Zealand
.3171
1.0989
8.0367
1.4412
10.0338
10.87
4.978
way
.4857
1.4411
8.0417
1.6584
10.4627
11.94
4.830
ama
.2159
.8375
5.8042
1.0040
9.1277
7.37
2.985
guay
2.1160
.3776
5.7508
.1999
9.0226
5.59
3.03
1
.0387
.6741
5.9410
.6279
8.9932
6.93
3.125
ippines
.1835
.4643
6.4181
.4231
8.5774
6.72
3.647
ugal
.5040
1.4031
7.7078
1.1552
9.5473
5.74
4.564
ania
2.0283
1.3161
7.7802
9.2145
4.562
apore
.8058
1.4634
8.0996
.7043
9.9257
6.12
5.330
th Africa
.0259
2.0539
5.9059
9.5138
5.40
3.089
n
.5488
1.3886
7.9823
1.0710
9.8881
7.95
4.829
den
.3634
1.0459
8.2523
1.7655
10.1585
11.65
5.013
zerland
.1777
1.4119
8.3095
1.3467
10.4738
12.72
5.142
an
.9204
.9159
8.4317
9.3989
9.19
5.452
iland
.6900
1.3383
7.4796
.7715
8.6315
5.19
4.565
isia
.4042
.4778
6.5832
.9768
9.0233
3.03
3.795
TABLE A1 (Continued)
Country dlnya lnskngdk testngdh lnshngdh lnya85 attain85 Score
262 Journal of Human Capital
Turkey .3062 .8486 6.9426 .0459 8.9009 4.69 4.128
United Kingdom .4793 1.0216 8.1928 1.3926 10.0367 11.93 4.950
United States
Uruguay
Ven
Zim
Barro, Robert J.
Paper no. 1829
������������
.3920
.3855
2012. “C
5, NBER
216.27.113
All use
1.0707
.5588
onvergen
, Cambrid
This conten
.144 on Sun
subject to h
7.9706
7.4393
ce and M
ge, MA.
t downloade
, 06 Feb 202
ttps://about.j
1.5204
.7494
odernizati
d from
2 23:48:53 U
stor.org/term
10.4798
9.2012
on Revis
TC�����������
s
12.37
7.26
ited.” Wo
��
4.903
4.300
ezuela
2.0078
.7169
5.3923
9.7069
2.578
babwe
2.7856
.4052
6.9826
9.0107
6.18
4.107
References
rking
Barro, Robert J., and Jong Wha Lee. 2013. “A New Data Set of Educational
Attainment in the World, 1950–2010.” J. Development Econ. 104:184–98.
Bernanke, Ben S., and Refet S. Gurkaynak. 2001. “Taking Mankiw, Romer, and Weil
Seriously.” NBER Macroeconomics Ann. 16:11–57.
Breton, Theodore R. 2010. “Schooling and National Income: How Large Are the
Externalities?” Educ. Econ. 18 ð1Þ: 67–92.
———. 2011. “The Quality vs. the Quantity of Schooling: What Drives Economic
Growth?” Econ. Educ. Rev. 30 ð4Þ: 765–73.
———. 2013a. “The Role of Education in Economic Growth: Theory, History, and
Current Returns.” Educ. Res. 55 ð2Þ: 121–38.
———. 2013b. “Were Mankiw, Romer, and Weil Right? A Reconciliation of the
Micro and Macro Effects of Schooling on Income.” Macroeconomic Dynamics 17
ð5Þ: 1023–54.
———. 2013c. “World Total Productivity Growth and the Steady State Rate in the
20th Century.” Econ. Letters 119 ð3Þ: 340–43.
———. 2015. “Human Capital and Growth in Japan: Converging to the Steady
State in a 1% World.” J. Japanese and Internat. Econ. 36 ð JuneÞ: 73–89.
Caselli, Francesco. 2005. “Accounting for Cross-Country Income Differences.” In
Handbook of Economic Growth, vol. 1A, edited by P. Aghion and S. N. Durlauf, 679–
741. Amsterdam: Elsevier.
Castelló-Climent, Amparo. 2010. “Channels through Which Human Capital In-
equality Influences Economic Growth.” J. Human Capital 4 ð4Þ: 394–450.
Cohen, Daniel, and Marcelo Soto. 2007. “Growth and Human Capital: Good Data,
Good Results.” J. Econ. Growth 12 ð1Þ: 51–76.
Dang, Hai-Anh, and F. Halsey Rogers. 2008. “The Growing Phenomenon of Private
Tutoring: Does It Deepen Human Capital, Widen Inequalities, or Waste Re-
sources?” World Bank Res. Observer 23 ð2Þ: 161–200.
Ding, Sai, and John Knight. 2009. “Can the Augmented Solow Model Explain
China’s Remarkable Economic Growth? A Cross-Country Panel Data Analysis.”
J. Comparative Econ. 37:432–52.
Ehrlich, Isaac, and Jinyoung Kim. 2007. “The Evolution of Income and Fertility
Inequalities over the Course of Economic Development: A Human Capital
Perspective.” J. Human Capital 1 ð1Þ: 137–74.
Fuchs, Thomas, and Ludger Woessmann. 2007. “What Accounts for International
Differences in Student Performance? A Re-examination Using PISA Data.”
Empirical Econ. 32:433–64.
Gennaioli, Nicola, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shlei-
fer. 2013. “Human Capital and Regional Development.” Q.J.E. 128 ð1Þ: 105–64.
Gollin, Douglas. 2002. “Getting Income Shares Right.” J.P.E. 110 ð2Þ: 458–74.
Hall, Robert E., and Charles I. Jones. 1999. “Why Do Some Countries Produce So
Much More Output per Worker than Others?” Q.J.E. 114 ð1Þ: 83–116.
Hanushek, Eric A., and Dennis D. Kimko. 2000. “Schooling, Labor-Force Quality,
Higher Test Scores or More Schooling? 263
and the Growth of Nations.” A.E.R. 90 ð5Þ: 1184–1208.
Hanushek, Eric A., and Ludger Woessmann. 2008. “The Role of Cognitive Skills in
Economic Development.” J. Econ. Literature 46 ð3Þ: 607–68.
———. 2011a. “How Much Do Educational Outcomes Matter in OECD Coun-
tries?” Econ. Policy 26 ð67Þ: 427–91.
———. 2011b. “Sample Selectivity and the Validity of International Student
Achievement Tests in Economic Research.” Econ. Letters 110:79–82.
———. 2012a. “Do Better Schools Lead to More Growth? Cognitive Skills, Eco-
nomic Outcomes, and Causation.” J. Econ. Growth 17:267–321.
———. 2012b. “Schooling, Educational Achievement, and the Latin American
Growth Puzzle.” J. Development Econ. 99:497–512.
Heston, Alan, Robert Summers, and Bettina Aten. 2009. Penn World Table Version
6.3. Philadelphia: Univ. Pennsylvania, Center Internat. Comparisons Produc-
tion, Income, and Prices.
Jürges, Hendrik, and Kerstin Schneider. 2004. “International Differences in Stu-
dent Achievement: An Economic Perspective.” German Econ. Rev. 5 (3): 357–80.
Klenow, Peter J., and Andres Rodriguez-Clare. 1997. “The Neoclassical Revival in
Growth Economics: Has It Gone Too Far?” NBER Macroeconomics Ann. 12:73–103.
Krueger, Alan B., and Mikael Lindahl. 2001. “Education for Growth: Why and for
Whom?” J. Econ. Literature 39:1101–36.
Mankiw, N. Gregory, David Romer, and David Weil. 1992. “A Contribution to the
Empirics of Economic Growth.” Q.J.E. 107 ð2Þ: 407–37.
Morrisson, Christian, and Fabrice Murtin. 2009. “The Century of Education.” J. Hu-
man Capital 3 ð1Þ: 1–42.
Psacharopoulos, George, and Harry Patrinos. 2004. “Returns to Investment in Ed-
ucation: A Further Update.” Educ. Econ. 12 ð2Þ: 111–34.
Woessmann, Ludger. 2003. “Schooling Resources, Educational Institutions and
Student Performance: The International Evidence.” Oxford Bull. Econ. and Statis.
65 ð2Þ: 117–70.
This content downloaded from
������������216.27.113.144 on Sun, 06 Feb 2022 23:48:53 UTC�������������
All use subject to https://about.jstor.org/terms
Coastal Education & Research Foundation, Inc.
Modeling and Analyzing the Dynamic Factors of Economic Growth Evolution in Coastal
Tourism Cities
Author(s): Zhenli Jia
Source: Journal of Coastal Research , SUMMER 2020, SPECIAL ISSUE NO. 103. Global Topics
and New Trends in Coastal Research: Port, Coastal and Ocean Engineering (SUMMER 2020),
pp. 1079-1083
Published by: Coastal Education & Research Foundation, Inc.
Stable URL: https://www.jstor.org/stable/10.2307/48639917
REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/10.2307/48639917?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
Coastal Education & Research Foundation, Inc. is collaborating with JSTOR to digitize, preserve and
extend access to Journal of Coastal Research
This content downloaded from
������������216.27.113.144 on Mon, 07 Feb 2022 00:12:25 UTC�������������
All use subject to https://about.jstor.org/terms
https://www.jstor.org/stable/10.2307/48639917
https://www.jstor.org/stable/10.2307/48639917?seq=1&cid=pdf-reference#references_tab_contents
https://www.jstor.org/stable/10.2307/48639917?seq=1&cid=pdf-reference#references_tab_contents
www.cerf-jcr.org
www.JCRonline.org
Modeling and Analyzing the Dynamic Factors of Economic
Growth Evolution in Coastal Tourism Cities
Zhenli Jia*
School of International Culture and Study
Yuxi Normal University
Yuxi 653100, China
ABSTRACT
Jia, Z., 2020. Modeling and analyzing the dynamic factors of economic growth evolution in coastal
tourism cities. In: Yang, Y.; Mi, C.; Zhao, L., and Lam, S. (eds.), Global Topics and New Trends in
Coastal Research: Port, Coastal and Ocean Engineering. Journal of Coastal Research, Special Issue No.
103, pp. 1079–1083. Coconut Creek (Florida), ISSN 0749-0208.
As an important part of tourism economy, inbound tourism economy has an important impact on the
development of regional economy. In order to improve the economic growth rate of coastal tourism cities,
a modeling analysis method of dynamic factors of economic growth evolution of coastal tourism cities is
proposed. On the basis of the regional economic theory, combined with the spatial analysis tools such as Geo
Da and Arc GIS, through the analysis of the temporal evolution and spatial pattern evolution of inbound tourism
foreign exchange income in coastal city clusters, furthermore, this paper uses the geographical weighted
model (GWR) to analyze the factors that affect the foreign exchange income of inbound tourism of coastal
city groups, and obtains the influencing coefficient. In view of the current situation and influencing factors
of foreign exchange income of inbound tourism in coastal city groups, the corresponding countermeasures
are put forward for the balanced and sustainable development of inbound tourism in coastal city groups. The
comparative experimental results show that the proposed method of dynamic factor modeling and analysis for
economic growth evolution of coastal tourism cities is more efficient and effective than the traditional method
of dynamic factor modeling and analysis for economic growth evolution of coastal tourism cities.
ADDITIONAL INDEX WORDS: Coastal tourism cities, urban economic growth, evolution of economic
growth, dynamic factor modeling.
Journal of Coastal Research SI 103 1079–1083 Coconut Creek, Florida 2020
DOI: 10.2112/SI103-225.1 received 20 August 2019; accepted in
revision 17 January 2020.
*Corresponding author: jiazhenli@126.com
©Coastal Education and Research Foundation, Inc. 2020
INTRODUCTION
In recent years, driven by the national policy of stimulating
domestic demand and increasing investment, China’s coastal
urban agglomerations, as the country’s exporters, have made
rapid development of their national economy with the help of the
spring breeze of reform and opening up. By virtue of its superior
geographical location, convenient transportation location, rich
natural and cultural tourism resources, developed economic
development level provides superior economic support for coastal
tourism development. Under the favorable policy environment
of the national strategy of vigorously developing the ocean,
the coastal tourism industry has maintained steady and rapid
development in general, and the income of inbound tourism has
developed rapidly (Chen, 2018).
This paper designs a dynamic factor modeling analysis of the
economic growth and evolution of new coastal tourism cities.
Starting from the current situation of the inbound tourism
development of coastal cities, it uses the quantitative analysis
method to analyze the dynamic factors of the economic growth
and evolution of coastal tourism cities, finds out the existing
problems through the analysis of the current situation, and then
studies the influencing factors, in order to realize the sharing
of tourism resources and tourism information in coastal areas,
realize the brand effect of urban tourism, better enrich the types
of tourism products, and promote the steady economic growth of
coastal tourism cities (Ma and Liu, 2019).
MATERIALS AND METHODS
Data Analysis of Economic Growth Conditions of Coastal
Tourist Cities
The level of regional economic development is an important
basis for the development of regional tourism industry, which
determines the development of local tourism economy to a certain
extent and affects the spatial pattern of regional tourism economy.
The high level of regional economic development can provide
certain financial and technical support for the development of
regional tourism economy, promote the improvement of tourism
infrastructure, and then affect the attraction of the region to tourists
(Cheng, Xu, and Guo, 2019). Since the Reform and Opening up,
the coastal areas as the mouth of foreign economic development,
with its superior location conditions, perfect transportation
network, suitable climate, the government in policy preferences
and support, so that the coastal areas have a rapid economic
development. Coastal city clusters have perfect infrastructure,
relatively abundant capital flow, convenient conditions to facilitate
the development of foreign trade and maritime transport, rapid
This content downloaded from
������������216.27.113.144 on Mon, 07 Feb 2022 00:12:25 UTC�������������
All use subject to https://about.jstor.org/terms
1080 Jia
Journal of Coastal Research, Special Issue No. 103, 2020
economic development has also brought good economies of scale
(Wen, Wu, and Gong, 2019; Xie et al., 2019).
Regional accessibility plays an important role in improving
the economic level of inbound tourism in a region. Regional
traffic conditions will directly affect the number and frequency
of inbound tourists, and further affect the development of
regional tourism economy. It is found that the level of tourism
economic development is positively related to the convenience
of traffic location conditions. The coastal areas are relatively flat
and economically developed. The roads, railways, waterways
and air traffic networks are perfect. The perfect traffic network
can accelerate the connection between the coastal areas and
between the coastal areas and the mainland (Wang et al., 2017).
The improvement of traffic conditions and the improvement of
traffic network promote the improvement of regional economy.
Efficient and convenient traffic conditions can improve the traffic
conditions of external regions relative to other regions to a certain
extent, and enhance the accessibility and attraction of tourist
destinations in interrelated or unrelated regions. The increase in
the number of tourists will correspondingly expand the market
model of tourists, and then make relevant departments take
measures to optimize the market structure of tourists. With the
gradual improvement of high-speed rail and EMU transportation
network, the transportation in coastal areas is more convenient and
efficient. The improvement of intercity high-speed rail shortens
the distance between cities, and the resources between regions
can be replaced with each other, which shortens the travel time
of tourists, increases their stay time in the destination, and thus
increases their consumption in the tourist destination. To a certain
extent, the tourism carrying capacity of a region affects the traffic
location conditions of the region. If the tourism carrying capacity
of a region is small, the increase of its tourist turnover may
damage the local environment. Therefore, the carrying capacity
of tourism environment should be considered while improving the
regional traffic network (Dong, Sun, and Li, 2018).
Tourism infrastructure is the basic guarantee for the
development of regional tourism economy, which can provide
strong support for the development of regional tourism economy.
The perfect tourism infrastructure itself can also be used as a
tourist attraction to attract tourists. The continuous improvement
of tourism infrastructure also plays an important role in enhancing
tourists’ tourism perception and improving tourism quality.
The coastal economy is relatively developed, and the tourism
infrastructure construction is relatively perfect (Wang et al.,
2020). Star hotels have complete supporting service facilities,
complete entertainment facilities, high quality service personnel,
high level of service, and strong professionalism, so that inbound
tourists can experience intimate services in the process of tourism,
as well as various leisure and entertainment facilities to relieve
the pressure. Multilingual tour guides can better provide good
services for inbound tourists, and the overall business level of
travel agencies is constantly improving, so that inbound tourists
can get better tourism experience in China. The continuous
improvement of tourism infrastructure has provided strong
support for the development of inbound tourism in coastal city
clusters (Wang and Chen, 2019).
Absolute difference is a single index quantization method. It
can only show the difference in quantity of indicators, and can’t
take into account the influence of other indicator factors. It has
poor comparability for different regions and different times.
Relative difference is also a single index difference measure. The
data of relative difference is a ratio, which is not affected by time
factor, space factor, economic factor and other factors, so it is
comparable. However, when the selected index is larger and the
dimension is larger, the result obtained by the relative difference is
relatively small, but the internal difference may be larger (Xiong et
al., 2019). Therefore, in the case of measuring relative difference
and absolute difference, comprehensive consideration should be
carried out, and a more appropriate method should be selected.
Considering both absolute difference and relative difference, the
error of results can be relatively reduced Equation (1):
( )
,i i
x x
S x x NN
v S x
−
=
=
=
(1)
In the equation, S is the standard deviation, ix is the
comprehensive index of the economic development of inbound
tourism in cities. It is the average value of the city group of the
comprehensive index of tourism economic development level; n
is the number of regional samples, V is the coefficient of variation.
The data motion guidance diagram is set accordingly in Figure
1:
It further explains the variation trend of the overall development
speed of inbound tourism in coastal areas, and introduces the
relative development rate here. The relative development rate
index is the ratio of the change of tourism income in one period to
Figure 2. Internal economic data processing diagram.
Figure 1. Data motion guidance diagram.
This content downloaded from
������������216.27.113.144 on Mon, 07 Feb 2022 00:12:25 UTC�������������
All use subject to https://about.jstor.org/terms
Modeling and Analyzing the Dynamic Factors of Economic Growth Evolution in Coastal Tourism Cities 1081
Journal of Coastal Research, Special Issue No. 103, 2020
the change of tourism income in another period. The equation is
as follows Equation (2):
2 1
2 1
i iY YNICH
Y Y
−
=
−
(2)
In this equation, 1iY and 2iY are used to represent the inbound
tourism revenue at the end of the year and the beginning of the
year of the i city, 2Y , 1Y represents the inbound tourism revenue
of the whole region at the end of the study period and the beginning
of the study period. When the value of NICH is greater than 1,
it indicates that the economic development speed of inbound
tourism in a city is greater than that of the whole region. And set
the internal economic data processing chart in Figure 2:
Dynamic Factor Modeling of Economic Growth Evolution of
Coastal Tourism Cities
Exploratory spatial data analysis (ESDA) is a spatial analysis
method, and spatial measurement is its core. Through the research
and analysis of the spatial distribution of economic phenomena, it
is concluded that the global statistics and the local statistics of their
spatial interaction are two different aspects of ESDA. The focus of
global statistics research is to analyze the spatial layout of specific
characteristics of a certain index in an area. Local statistics is to
compare and study the data contained in a small area in the global
area, so as to study whether the regional information changes are
homogeneous or heterogeneous (Tian et al., 2020). The focus of
global statistical research is to analyze the spatial distribution of a
specific index in a region. Local statistics is to study whether there
is homogeneity or heterogeneity in regional information change
by comparing the data contained in a small region in the global.
The local Moran’s I value is positive, which means that regions
with higher economic level are surrounded by regions with higher
economic level Surrounding or low-level areas are surrounded by
surrounding low-level areas. In LISA’s agglomeration diagram,
HH (high agglomeration) indicates that the value of this city
and its surrounding cities is relatively high, indicating high
value agglomeration; LH (low wealth cluster), which means that
the value of the city is relatively low but the observed value of
the surrounding area is relatively rich; LL (low concentration)
means that the observed values of cities and surrounding cities
are relatively low, and represents low concentration. HL (from
Low Agglomeration) means that the observed value of the city
is high and that of the surrounding city is low. LH and HL are
spatial outliers without obvious clustering phenomenon, so they
are called atypical regions. The significance level is judged by
testing the Z value of the normal statistics of the index. And set
the data filtering diagram as follows:
Moran scatter diagram is a method of local spatial autocorrelation
analysis, which can study the local spatial heterogeneity. It is
represented by Cartesian rectangular coordinate system. Abscissa
represents the research area. The research index is the value after
standardization. The ordinate is the average value of the attribute
value of the adjacent element determined by the spatial adjacency
matrix after standardization. M the oran scatter diagram is divided
into four quadrants: high-high (the first quadrant), low high (the
second quadrant), low-low-low (the third quadrant), quotient
low (the fourth quadrant). Compared with local Moran’s I
statistics, Moran scatter-plot can further distinguish which spatial
correlation mode a region belongs to and its adjacent regions. The
results of exploratory spatial data analysis largely depend on the
determination of spatial weight matrix. The commonly used spatial
weight matrix mainly includes: The spatial weight matrix based
on proximity concept, k-value nearest neighbor matrix, distance
based spatial weight matrix and economic and social spatial
weight matrix, in this paper, through the comparative analysis of
several methods of establishing weight matrix, considering the
objectivity, reliability and rationality of the analysis results, the
Figure 3. Data filtering diagram. Figure 4. Data filtering diagram.
This content downloaded from
������������216.27.113.144 on Mon, 07 Feb 2Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
1082 Jia
Journal of Coastal Research, Special Issue No. 103, 2020
k-value nearest neighbor matrix is finally selected. And build the
data filtering diagram
RESULTS AND DISCUSSION
According to the basis, characteristics and trend of China’s
inbound tourism development, this paper selects the tourism
foreign exchange income index reflecting the level of inbound
tourism development as the explanatory variable. The main
factors influencing the development level of inbound tourism,
such as regional economic development level (RIGDP), tourism
traffic conditions (JT), tourism infrastructure (JD), tourism
resource endowment (JQ), are selected as explanatory variables.
Among them, the level of regional economic development (GDP)
is used to measure the economic basis of the development of
inbound tourism in a region. The higher the level of economic
development in a region, the development of other industries
will be driven. Therefore, the development of regional economy
plays a leading role in the development of inbound tourism.
The higher the level of regional economic development, the
higher the per capita disposable income in the region, and then
make the residents here use the per capita national economic
income (RIGDP) to characterize; Tourism traffic condition (JT)
is used to measure the impact of the accessibility of a region
on the tourism attraction of a region. The more convenient the
transportation of a region is, the more tourism development of
the region will be driven to a certain extent, and more inbound
tourists will be attracted. Here, traffic flow is used to represent
the convenience, accessibility and accessibility of tourism traffic;
Tourism infrastructure factor (JD), the perfection degree of
tourism infrastructure in a region plays an important role in the
development of inbound tourism. The perfection degree of tourism
infrastructure will promote or restrict the development of inbound
tourism in a region. Here, the number of star hotels is used to
represent the perfection degree of tourism infrastructure; Tourism
resource endowment factor (JQ), tourism resource endowment is
an important factor affecting the development of inbound tourism
in a region. The uniqueness of tourism resources can attract more
inbound tourists, thus creating more inbound tourism income. The
development degree of scenic spots also affects the consumption
of inbound tourism, so we should pay attention to the appropriate
development and uniqueness of scenic spots in the development
process. Here, the number of scenic spots is used to represent the
tourism resource endowment of a region. And set up the analysis
efficiency comparison diagram of this modeling analysis method
and traditional modeling analysis method as follows:
Compared with the above figure, under the same parameter
conditions, the analysis efficiency of the modeling and analysis
method in this paper is higher and always on top of the traditional
method, which can better serve the experimental research and
promote the development of the modeling system.
The development level of tourism in urban agglomerations
depends on the economic development of the region, which has
an important impact on the improvement of the level of urban
tourism and its spatial layout. The high level of regional economic
development can provide certain financial and technical support
for the development of regional tourism economy, promote
the improvement of tourism infrastructure, and then affect the
attraction of the region to tourists. Its influence on the development
of inbound tourism economy is indirect to some extent and its
direct influence is small.
CONCLUSION
Based on the theory of regional economy, combined with the
spatial analysis tools such as Geo Da and Arc GIS, this paper
analyzes the time evolution and spatial pattern evolution of the
inbound tourism foreign exchange income of coastal city groups,
and further draws the following conclusions: according to the
development situation and influencing factors of the inbound
tourism foreign exchange income of coastal city groups, the paper
uses the geographical weighted model (GWR) to influence the
inbound tourism foreign exchange income of coastal city groups.
Based on the analysis of the factors of foreign exchange income,
the influencing coefficient is obtained, and the Countermeasures
for the balanced and sustainable development of inbound tourism
of coastal city groups are put forward.
LITERATURE CITED
Chen, Z.L., 2018. Spatial econometric analysis of the impact of
R&D input on tourism economy growth. Tourism Research,
10(1), 83-94.
Cheng, H.; Xu, Q., and Guo, Y.Q., 2019. Temporal and spatial
evolution of the coupling coordinated development between
tourism resources development and ecological environment
in China. Economic Geography, 39(7), 233-240.
Dong, Z.W.; Sun, J., and Li, Y.F., 2018. Measurement of
development level of marine tourism in China’s coastal
cities. Statistics and Decision, 34(19), 130-134.
Ma, X.F. and Liu, Y.L., 2019. The evolution characteristics of
urban spatial form and its relationships with the growth of
tourism industry. Economic Geography, 39(5), 226-234.
Tian, X.; Song, Z.; Wang, B., and Zhou, G., 2020. A theoretical
calculation method of influence radius of settlement based on
slices method in tunnel construction. Mathematical Problems
in Engineering.
Wang, H.; Wang, J.; Lu, H.; Bo, G.; Zhang, X.; Cao, Y.; Liu,
L.; Zhang, J., and Zhang, W., 2020. Analysis of coating
electrode characteristic in the process of removing pollutants
from wastewater. Fresenius Environmental Bulletin, 29(2),
715-721.
Wang, P.; Li, J.; Bai, F.; Liu, D.; Xu, C.; Zhao, L., and Wang, Z.,
2017. Experimental and theoretical evaluation on the thermal
performance of a windowed volumetric solar receiver.
Energy, 119, 652-661.Figure 5. Experimental comparison.
This content downloaded from
������������216.27.113.144 on Mon, 07 Feb 2022 00:12:25 UTC�������������
All use subject to https://about.jstor.org/terms
Modeling and Analyzing the Dynamic Factors of Economic Growth Evolution in Coastal Tourism Cities 1083
Journal of Coastal Research, Special Issue No. 103, 2020
Wang, Z.H. and Chen, J.M., 2019. A SWOT analysis on the
integrated development of Liaoning province coastal sports
tourism. Liaoning Sport Science and Technology, 41(5), 10-
14.
Wen, F.; Wu, N., and Gong, X., 2020. China’s carbon emissions
trading and stock returns. Energy Economics, 86, 104627.
Xie, S.Y.; Zhang, Q.; Gong, J.; Han, L., and Wang, X.F., 2019.
Construction and application of comprehensive evaluation
model of the urban scenic spots accessibility: Taking
downtown Wuhan as a case. Economic Geography, 39(3),
232-239.
Xiong, Z.; Wu, Y.; Ye, C.; Zhang, X., and Xu, F., 2019. Color
image chaos encryption algorithm combining CRC and nine
palace map. Multimedia Tools and Applications, 78(22),
31035-31055.
This content downloaded from
������������216.27.113.144 on Mon, 07 Feb 2022 00:12:25 UTC�������������
All use subject to https://about.jstor.org/terms
Determination of Financial Factors in the Latest Theories of Economic Growth … 49
UDC 332.1
doi: 10.15330/jpnu.7.3.49-59
DETERMINATION OF FINANCIAL FACTORS IN THE LATEST
THEORIES OF
ECONOMIC
GROWTH OF TERRITORIAL
COMMUNITIES
HALYNA VOZNYAK, KRYSTYNA PATYTSKA, TARAS KLOBA
Abstract. In today’s modern circumstances ensuring the economic growth of territorial
communities requires theoretical redefining and additional research, since new challenges have
negatively affected the well-being of the population and ensuing progress.
The purpose of the article is to determine financial factors in the latest theories of economic
growth of territorial
communities.
Scientific approaches to the definition of “territorial community” are generalized, own
approach is offered and the main features of territorial community (territorial, natural, social-
psychological, historical-cultural, organizational-functional, political, economic) are substantiated.
Emphasis is placed on the need for a new theoretical approach to the development of territorial
communities based on the synthesis and combination of basic guidelines and principles of modern
scientific concepts and models, their interpretation in the context of local development. The latest
theories of economic growth are studied, including: the theory of endogenous growth, inclusive
development and the theory of sustainable development (E. Ostrom’s concept of community
resource management is highlighted). It is substantiated that the model of endogenous growth is
focused on reducing inequality of economic development through the use of internal factors,
mechanisms of management and management of the territory. There are two vectors of achieving
inclusive development of territorial communities: internal –
maximum involvement of community
members in the process of its development; external –
ensuring access of members of the territorial
community to the opportunities provided by more developed communities. It is emphasized that
the key idea in E. Ostrom’s research is to achieve efficiency and substantiation of ways of
community management of common resources. Within the framework of the theory of inclusive
development, a review of the concept of financial inclusion, which raises the issue of community
development – “banking deserts”. The existence of interdependence of perception of members of
territorial communities – “banking deserts” of opportunities and prospects of obtaining banking
services is indicated, which requires the involvement of theoretical foundations of behavioral
economics in the study of this issue. The factors of economic growth of territorial communities are
systematized, which are united into five groups: natural, factors of human development; social,
production and financial. The financial factors of economic growth of territorial communities
include: budget; investment, inflation, monetary. The expediency of applying the ideas of the latest
theories of economic growth in the process of analyzing financial factors is substantiated.
Keywords: territorial community, financial factors of economic growth, endogenous growth,
inclusive development, sustainable development, common resources, financial inclusion.
JEL Classification: 011, 016.
Journal of Vasyl Stefanyk Precarpathian National University
http://journals.pnu.edu.ua
Vol. 7, No. 3 (2020), 49-59
50 Halyna Voznyak, Krystyna Patytska, Taras Kloba
1. INTRODUCTION
The new challenges caused by the consequences of the pandemic and the socio-economic crisis
had a negative impact on the welfare of the population, exacerbated a number of economic problems of
further functioning and progress of Ukraine as a whole and newly formed communities, that have
entered a fundamentally new phase of development (including through the reform of administrative
and financial decentralization). On the other hand, in Ukraine, ensuring the financial and managerial
capacity of territorial communities is significantly complicated by the deep institutional crisis and the
preservation of established hierarchical relations at all levels. The sensitivity of the economy of
territorial communities to these preconditions is increased by its underdevelopment and the problem of
situational motivation in the behavior and actions of economic entities, which is not focused on
strategic long-term community development and sustainable networking as the basis of its economic
and social
development.
This actualizes the issues that were mentioned above and indicates the urgent need for
theoretical comprehension of the issue of economic growth of territorial communities in the new
conditions.
2. THEORETICAL BACKGROUND
Due to its socio-economic significance, the outlined topics are in the focus of scientific research of
representatives of economic and financial science. Thereby, one of the fundamental models of economic
growth of the territory in modern conditions is the theory of endogenous scientific and technological
progress of P. Romer, in which scientific and technological progress is “a factor of economic growth
generated by internal causes” [1], and economic growth directly depends on the amount of human
capital, concentrated in the field of knowledge acquisition. The factors of economic growth in this
theory are knowledge and information, which determine the development of innovation, scientific and
technological progress and the new state of human capital.
P. Romer’s ideas were developed by World Bank experts D. Chen and H. Ki in the theory of
knowledge and endogenous growth [2]. Using a two-sector model of a closed economy (including the
manufacturing sector and the research and development sector), based on the application of the Cobb-
Douglas function, which reflects the dependence of changes in productivity on certain factors, the
authors link economic growth with increasing accumulation of financial resources in human capital
development.
The issue of using financial resources to increase the level of human capital as a basis for economic
growth of the territory is widely considered in recent decades and is the subject of numerous scientific
discussions. According to conclusions of the World Bank, “empowerment results in the ability of the
poor to influence government institutions that determine their living conditions by strengthening their
participation in political processes and decision-making at the local level. And that means removing
barriers – political, legal and socio-cultural – and increasing the assets of the poor so that they can
effectively enter the markets” [3].
Thus, the hypothesis of the impact of the level of human capital development on the economic
growth of the territory was developed by R. Birochy and M. Pozhebon in their proposed concept of
critical financial education in the context of improving the financial inclusion of the entities of the
community economy as one of its economic growth factors [4]. The empirical study was conducted by a
municipality in Brazil, where providing community residents with access to information and
communication technologies has resulted in significant socio-economic changes. After collecting data
through surveys and using the method of coding economic information, scientists have concluded that
financial education is the driving force of financial inclusion of low-income entrepreneurs and ensuring
their access to financial resources.
Theoretical and methodological development of the issues raised by foreign and domestic scientists
is carried out mainly at the regional, national and international levels. Despite significant
Determination of Financial Factors in the Latest Theories of Economic Growth … 51
developments, the theoretical aspects of the development of territorial communities, the definition of
the basic principles and factors of their sustainable progress in conditions of uncertainty have not been
properly presented, and therefore require in-depth study.
3. RESEARCH OBJECTIVE, METHODOLOGY AND DATA
The purpose of the article is to determine financial factors in the latest theories of economic growth
of territorial communities. The object of research in this article are the latest theories and factors of
economic growth and long-term development of territorial communities. The methodological basis of
this study are the fundamental provisions and principles of the theory of endogenous growth,
sustainable and inclusive development, a number of domestic and foreign empirical studies on this
issue. A set of specific methods of scientific knowledge, namely: logical generalization, analysis,
synthesis, scientific abstraction, historical approach provided the opportunity to realize the integrity of
scientific research.
4. RESULTS AND DISCUSSION
The decentralization reform has enabled territorial communities to become a full-fledged subject of
territorial management by obtaining the appropriate powers and resource base, which provides for the
responsibility of local governments for the socio-economic development of the territory.
The study of the theoretical foundations of the development of territorial communities necessitates
the definition of the essence of this definition. Therefore, the generalization of scientific approaches to
its understanding suggests that the territorial community should be understood as a naturally formed
human community organized in a certain area, characterized by a set of permanent ties and common
interests in their livelihood and socio-economic development. At the same time, emphasizing the role
of the territorial community in the context of its administrative-territorial determination as a subject of
administrative relations in the system of state formation, it should be considered as a primary subject of
local self-government with self-formed governing bodies endowed with the rights and responsibilities
to address issues of local significance, which is the representative of the local community in the political
arena and the owner of communal property in the relevant territory.
This approach to the interpretation of the category “territorial community” allows you to
systematize the set of its explicit features:
territorial – a territorial community is formed on a certain territory;
natural – territorial community is formed and developed on the basis of self-organization, social
and economic activity of its members and the need for constant development, following the principle of
self-regulation;
socio-psychological – territorial community exists subject to self-identification of each member of
the community as part of it and awareness of the commonality of their interests. The community is first
and foremost a system of constant communicative connections between members of the community;
historical and cultural – the community exists and develops over time, so when choosing the
direction of its development it is necessary to take into account the historical aspects of community
evolution, its cultural and customary features that significantly affect the psychological climate in the
community and endogenous relationships;
organizational and functional – territorial community is a system with horizontal organization of
political and social networks, which operates based on democratic principles of development and
subject to the participation of its members in solving their livelihood issues with the possibility of self-
structuring and creation of internal organizational structures;
political – the territorial community is a subject of legal relations and a representative of the local
community in the political arena;
economic – the local community is the owner of communal property located on its territory, and
its members are payers of tax payments to the local budget. The community operates to provide its
52 Halyna Voznyak, Krystyna Patytska, Taras Kloba
members with quality public services, can be a participant in production processes and is a collective
consumer.
This emphasizes the priority of considering the territorial community as an economic system that
operates to achieve social welfare, political goals and economic growth, provided the effective use of
available territorial and spatial resources in the context of ensuring the triad of interests of the local
community, territory and individual members.
It should be noted that the problem of management and rational use of available resources and
opportunities to ensure economic growth of territorial communities, which in domestic conditions is
one of the key targets for the development of the territory, is at the stage of formation and search for
optimal solutions. Despite the significant number of scientific approaches and theories aimed at solving
local development issues, none of them provides a complete and systematic solution to the problems of
local communities and does not take into account current trends and features of their development and
condition. This requires understanding a new theoretical approach to the development of territorial
communities based on the synthesis and combination of basic guidelines and principles of modern
scientific concepts and models of local development.
Popular in recent decades and formed as an independent direction of the theory of economic
growth is the model of endogenous growth, aimed at reducing inequality of economic
development
through the use of internal factors, mechanisms of management and governance. The development of
the theory was ensured not only by scientific research and the work of economists, but also by
representatives of geography and sociology, which resulted in the diversity of trends within the model
(P. Romer’s theory of endogenous scientific and technological progress, “schooling model” and
“learning by doing”). Lucas, the concept of bottom development by K. Weaver, the decentralization
development of B. Planck, the concept of growing development of M. Basand, the theory of rise of
W. Rostow, the theory of convergence of J. Lafontaine and P. Idalo, the theory of local development
developed by F. Bouvet, Yu. Dion, P.-A. Tremblay, B. Pecur, etc.).
Analysis of numerous approaches and scientific concepts within the theory allows us to identify
three key principles that underlie endogenous growth and are common to all areas, namely:
1) territoriality – territory is the basis for endogenous development, which is characterized as integrated
because it is carried out in to a limited extent, and which has certain features of autarky;
2) interdependence – the achievement of endogenous growth is the result of the impact and efficiency of
use of each element of the limited space within which it is provided (natural, cultural, social, economic,
etc.); 3) democracy – endogenous growth is possible only under the conditions of existence and
appropriate level of development of democratic institutions in a given area and is based on meeting the
basic needs of the population (nutrition, education, health, work, etc.) through the use of local
economic potential.
In the context of our study, it is advisable to turn to the analysis of local development concepts that
stand out within the theory of economic growth. Given the rather wide range of concepts and models
of local development, which relate to finding ways to develop industrial areas, innovation circles, the
application of the principles of flexible specialization for economic development of administrative-
territorial formations, etc. [5], and differ primarily in emphasizing the superiority of one principle over
others [6], define the common ideas on which they are based. Thus, the initial conditions of the process
of endogenous economic growth at the local level are the production of innovation, the
ability to adapt
and the ability to regulate. B. Pecur emphasizes this, noting the key role of the dynamism of the actors
in ensuring these conditions [7], as well as M. Bassan, I. Pedrazzini, F. Feinar and R. Peranjake, who
note that local development can be interpreted as a partnership agreement on creating favorable
conditions for the implementation of local initiatives in the context of community capacity building,
adaptation to new conditions, search for new forms and mechanisms of development, which
organizational and production methods will be aimed not only at economic benefits but also at solving
social, cultural and environmental issues character [8].
Interesting in the context of ensuring economic growth at the local level is the model of rural
development developed by DA McGranahan, T. Vojan and D. Lambert in the United States as a
Determination of Financial Factors in the Latest Theories of Economic Growth … 53
synthesis of the theory of endogenous growth and creative economy [9]. The authors of the model
identify three main factors of economic growth in rural areas: entrepreneurship, creative class and
recreational resources. The basis of the proposed model is the development of entrepreneurship and
the involvement of representatives of the creative class through the formation and supply of life in
rural areas. At the same time, scientists note that the application of the model is appropriate for the
local economy in terms of declining employment in traditional industries and reducing production,
which is characterized by the use of low-skilled labor.
Among domestic economists, the issue of endogenous growth has been studied by many scientists.
At the same time, analyzing the theory of endogenous growth, we agree with the conclusions of
J. Zhalil, who, studying the problem of endogenization of economic development, identified the main
areas of economic growth, which are fully consistent with modern conditions of territorial communities
in Ukraine:
investment and innovation policy, in particular the financial tools of their implementation, which
set the institutional mechanisms for investment of financial resources;
business development in the context of increasing the capitalization of economic potential;
development of human capital in order to increase its productivity increase networking (in
particular, as proved by M. Vozhnyak, there is a clear positive relationship between the development of
education in a given area and the level of economic growth, and investment in education and health
have a positive impact on productivity of the community and activate its members to solve problems of
socio-economic nature [10]);
development of communication environment and development of network relationships;
decentralization of processes of identification and involvement in economic circulation of
available and potential resources of economic development [11].
Determining the directions of achieving economic growth of territorial communities requires taking
into account the influence of external conditions, the variability of political and economic conditions
and unforeseen factors, the number of which has increased significantly in the context of globalization.
In this context, it is appropriate to focus on the study of World Bank experts [12], who emphasize the
need to fully support the transformation of local economies during and after the Covid-19 pandemic for
long-term economic recovery and sustainability. Ensuring the economic growth of territorial
communities in rural areas, they propose to carry out in three directions: 1) search and development of
strengths and strengths of a particular area; 2) attracting investment for the development of the local
business environment; 3) development of basic infrastructure and provision of Internet access.
At the same time, due to an in-depth analysis of the economic development of local communities in
different countries during the Covid crisis, the categories of the population most “affected” by the
impact of measures to combat the pandemic were identified. Accordingly, the World Bank
recommends that local governments focus on supporting women, youth, the informal sector, and micro
and small enterprises on a more sustainable basis. That is, the World Bank emphasizes the need to
achieve economic growth while ensuring social justice, which meets the need to implement the Global
Sustainable Development Goals for 2015-2030.
The issue of inclusive development is widely represented in the research of foreign scholars, but
conceptually and theoretically it has not been developed. According to the UN, inclusiveness is based
on the involvement of all marginalized and excluded groups in the development process as
stakeholders [13]. The main principles on which the theory is based are: participation – maximum
involvement of all community members in the process of its development; accessibility and non-
discrimination – ensuring equal access to opportunities for all members of the community.
The issue of inclusive development is widely represented in the research of foreign scholars, but
conceptually and theoretically it has not been developed. According to the UN, inclusiveness is based
on the involvement of all marginalized and excluded groups in the development process as
stakeholders [13]. The main principles on which the theory is based are: participation – maximum
involvement of all community members in the process of its development; accessibility and non-
discrimination – ensuring equal access to opportunities for all members of the community.
54 Halyna Voznyak, Krystyna Patytska, Taras Kloba
It is worth noting that the progress of the theory of inclusive development took place in the context
of the development of the state, international relations and the individual. And only in recent years has
the theoretical foundations of inclusiveness been directed to the problems of local development. To
date, the most complete, in our opinion, definition of inclusive local development has been given by
J. Gupta, N. Pove and M. Ross-Tonen: it is a new dimension of development that focuses on the poorest
and most marginal members of society, taking into account economic, social and environmental aspects
and structural factors that prevent the poorest participants from participating in the development
process [14]. In this context, scholars also identify the components of inclusive development at the level
of territorial communities: providing equal opportunities for development and equitable distribution of
benefits; providing economic opportunities for community members; public participation;
environmental protection; adaptive capacity, which provides mitigation of the shocks of existence for
different groups within the community [15].
Thus, the inclusion of members of the territorial community in the context of economic growth
involves the most effective use of human capital in the direction of enhancing social, labor, managerial
and economic relations and the formation of the business environment. At the same time, it is also
designed to provide quality living space for all residents of the community. Therefore, we can
distinguish two vectors of achieving inclusive development of territorial communities: internal – the
maximum involvement of community members in the process of its development; external – ensuring
access of members of the territorial community to the opportunities provided by more developed
communities.
A component of the theory of inclusive development is the concept of financial inclusion, which
substantiates the problem of “banking deserts” – territorial communities whose members do not have
access to financial services [16]. Historically, the category of territorial communities – “banking deserts”
include poor communities, and lack of access to financial services, despite the factors of slowing
economic development, is an additional factor hindering economic growth and, consequently,
discrimination against the community [17, 18].
An interesting aspect, which was substantiated in the scientific research of researchers, is to identify
the reasons for the inefficiency of the branches of powerful banks in the so-called “banking deserts”. As
M. Baradan notes, banks “do not speak the financial language of the poor”, and therefore do not
understand that poor consumers “can not be offered banking services as if they were just rich people
with less money” [6]. Applying the scientific findings of the psychology of poverty to the theory of
financial inclusion reveals differences in the behavior and prerogatives of people in underdeveloped
(poorer) communities: poor consumers tend to be more community-oriented than wealthier; they are
more concerned with the needs and well-being of their community [17, 19]. This leads to the choice of
appropriate mechanisms for banking services to the population of such communities, including the
development of utility banks, as consumers of banking services in poorer communities often exaggerate
the importance of the utility bank for their community. Accordingly, researchers suggest that banking
institutions that “enter” the “banking desert” attract consumers, emphasizing the link between the
financial interaction of the consumer with the bank and the well-being of the community.
In general, the basis of the concept of financial inclusion is based on solving the following tasks: 1)
determining the role of financial services in an inclusive economy and the development of microfinance
models (rural savings and loan associations, self-help groups, credit unions) to ensure financial
inclusion of the community; 2) research on the financial behavior of low-income groups to develop
more appropriate financial products, especially savings, insurance, payment services, value chain
financing and innovative community-based financing models; 3) search for opportunities and ways to
establish links with private sector financial institutions, use of mobile banking, etc.; 4) development of
approaches to the integration of community-based microfinance models in empowerment strategies for
marginalized community members. The main advantages of financial inclusion include: at the level of
individuals, financial inclusion results in increased savings, investment in education and resilience to
financial shocks; at the community level – reducing financial inequality and economic growth in
general [20].
Determination of Financial Factors in the Latest Theories of Economic Growth … 55
To form a comprehensive approach to economic growth at the level of local communities, in
addition to taking into account the principles of endogenous growth and inclusive development, it is
necessary to focus on the guidelines of sustainable development, in particular the scientific concept of
Nobel Laureate E. Ostrom on rational use of shared resources.
By definition, shared resources are a “natural or artificial resource system” [21], which has the
following characteristics: exclusivity (access to a shared resource can be limited at minimal cost), rivalry
(competition for access to a shared resource), divisibility (the ability of a resource to be divided into
shares) and exhaustiveness. Having conducted numerous empirical studies, E. Ostrom came to a
significant conclusion: the management of shared resources by the community that uses them can be
characterized by much higher efficiency than when transferring them to private ownership or through
public administration; at the same time, methods of managing such resources should differ taking into
account territorial, economic, customary differences. Therefore, on the basis of the conducted
researches, the author has formed eight principles of effective management of common resources:
1) defining clear boundaries of the common resource;
2) formation of clear rules and norms for the use of common resources based on the needs of the
community and the conditions of its
development;
3) democratization of the joint resource management process;
4) control over the use of shared resources;
5) formation of a mechanism of graduated sanctions against violators of the rules of using a
common
resource;
6) formation of a mechanism for effective and rapid resolution of conflicts between users of the
resource;
7) formation of an effective effective system of joint resource management, starting from the local
level to the regional or state [22].
In the context of defining the concept of “common resources” by E. Ostrom and on the basis of
elaboration and supplementation of the classification of common resources of rural communities,
carried out by a team of authors led by V. Nelepa [23], we distinguish the main groups of common
resources, information and resources of the socio-economic sphere (Fig. 1).
Fig. 1. Types of common resources of territorial community and organizationalmethods
of their management. Source: compiled by the authors.
ORGANIZATIONAL MANAGEMENT METHODS OF COMMUNITY COMMON RESOURCES
COMMON RESOURCES OF TERRITORIAL COMMUNITY
NATURAL
RESOURCES OF
THE SOCIO-
ECONOMIC
SPHERE
SOCIAL FINANCIAL INFORMATIVE
Consumer’s
Association of
natural
resources,
environmental
groups
Local media,
radio,
television,
web pages,
information
booklets
Utility services Public
organizations
Local budget Social
enterprises
ECONOMIC
56 Halyna Voznyak, Krystyna Patytska, Taras Kloba
Thereby, the definition in the analysis of scientific theories and concepts of directions and principles
of development of territorial communities allows to systematize the factors of their economic growth,
which can be conditionally grouped into five groups: natural (including natural resources of the
territory); human development factors (including demographic and behavioral (knowledge, skills,
motivation, reactions) characteristics of community members); social (taking into account the
institutions that promote the development of human capital and determine the areas of management
and development of the territory (family, community, enterprises, public organizations, volunteer
organizations)); productive (material goods and fixed assets that contribute to the production process);
financial.
In this study, the main attention is paid to the group of financial factors of economic growth of
territorial communities, and therefore in the process of identifying guidelines for the development of
territorial communities in the analysis of economic growth theories, the main attention should be paid
to substantiating such factors (Fig. 2).
Fig. 2. Selection of financial factors of development of territorial communities in the context of theories of economic
growth. Source: compiled by the authors.
THEORY OF
ENDOGENEOUS GROWTH
The concept of local
development
THEORY OF SUSTAINABLE
DEVELOPMENT
The concept of common resource
management E.Ostrom
THEORY OF INCLUSIVE
DEVELOPMENT
The concept of financial inclusion
THE MAIN IDEA:
reducing inequality of economic
development through the use of
internal factors, mechanisms of
management and administration of
the territory
THE MAIN IDEA:
maximum involvement of community
members in the process of its development;
ensuring access of members of the territorial
community to the opportunities provided by
more developed communities
THE MAIN IDEA:
achieving efficiency and justification of
methods of common resource
management
PRINCIPLES:
territoriality, unity, normality, interdependence,
democracy, public participation, accessibility, non-
discrimination, efficiency, effectiveness, partnership,
subsidiarity
KEY CONDITION:
DYNAMISM OF SUBJECTS OF COMMUNITY
ECONOMY IN THE DIRECTION:
producing innovation
ability to adapt
ability to regulate
FACTORS OF ECONOMIC GROWTH OF TERRITORIAL COMMUNITY
NATURAL
SOCIAL
HUMAN DEVELOPMENT FACTORS
PRODUCTIVE
FINANCIAL
Budget
Monetary
Inflation
Investment
THEORETICAL BASIS OF ECONOMIC GROWTH OF TERRITORIAL COMMUNITIES
Determination of Financial Factors in the Latest Theories of Economic Growth … 57
The financial factors of economic growth of territorial communities include:
budget – includes features of formation and use of the local budget;
investment – involves taking into account the investment policy of the self-government body and
identifying the features of the investment direction of the financial resource in community
development;
inflation – a factor influencing changes in the general level of prices and purchasing power of
money on the economic development of the territorial community and economic entities in its territory;
monetary – a factor that allows you to assess access to financial resources, as well as the turnover,
distribution and redistribution of money capital between economic entities.
5. CONCLUSIONS
The selection of financial factors of economic growth of territorial communities in the context of the
synthesis of the theory of endogenous growth, the theory of inclusive development and the theory of
sustainable development contributes to a comprehensive approach to determining the ways of
development of territorial communities. After all, the study of financial factors of territorial
communities in the context of the analyzed theories will be carried out from different angles: in the
theory of endogenous growth – from the standpoint of stimulating economic development and
overcoming economic and social inequality, in the theory of inclusive development – from the
standpoint of community involvement , which are in the risk group, in the concept of joint resource
management the finances of the territorial community are considered as one of the common resources
of the community, which requires finding an effective approach to its management.
Acknowledgements
The study was conducted under the grant 2020.02 / 0215 “Financial determinants of ensuring
regions and territorial communities’ economic growth based on behavioural economy” with the
support of National Research Foundation of Ukraine.
REFERENCES
[1] Venger V.V. Evolution of views on the theory of economic growth. Economy and Society, 2 (2016), 3–9.
(in Ukrainian)
[2] Chen Derek H.C., Kee H.L. A Model on Knowledge and Endogenous Growth. World Bank Policy
Research Working Paper, 20433 (2004). The World Bank, Washington D.C. Available at:
https://openknowledge.worldbank.org/handle/10986/8878
[3] World development report 2000/2001: Attacking poverty. World Bank, Oxford University Press, New York,
USA, 2000. Available at: https://openknowledge.worldbank.org/handle/10986/11856
[4] Birochi R., Pozzebon M. Improving financial inclusion: towards a critical financial education.
Framework. Rev. adm. empres, 56 (3) (2016). doi: 10.1590/S0034-759020160302
[5] Ryeznikova N.V., Ivashchenko O.A. The problem of economic development and growth in the context
of overcoming global asymmetries. Odesa National University herald, 21 (1) (2016), 55–59. Available at:
http://nbuv.gov.ua/UJRN/Vonu_econ_2016_21_1_14 (in Ukrainian)
[6] Bouvet F. Recherche sur les fondements méthodologiques des politiques économiques environnementales
confrontées à la notion de développement soutenable. Université Panthéon-Sorbonne, Paris, 2001. (in French)
[7] Pecqueur B. Le developpement local: mode ou modele? In: Alternatives économiques. Série Poche.
Syros/Alternatives, Paris, 1989. (in French)
https://economyandsociety.in.ua/index.php/index
58 Halyna Voznyak, Krystyna Patytska, Taras Kloba
[8] Bassand M., Hainard F., Pedrazzini Y., Perrinjaquet R. Innovation et changement social: actions culturelles
pour un développement local. Presses Polytechniques Romandes, Lausanne, 1986. (in French)
[9] McGranahan D., Wojan T., Lambert D. The rural growth trifecta: outdoor amenities, creative class and
entrepreneurial context. Journal of Economic Geography, 11 (3) (2011), 529-557.
[10] Vozhnyak M.H. New approaches to the analysis of the mutual influence of social inequality, human
capital and economic growth. International economic policy, 1–2 (16–17) (2012), 35–54. Available at:
http://nbuv.gov.ua/UJRN/Mep_2012_1-2_4 (in Ukrainian)
[11] Zhalilo Ya. A. On the institutional basis of endogenization of economic development. Paradigmatic
shifts in the economic theory of the XXI century: materials of the III International scientific-practical
conference. Kyiv, November 2-3, 2017. KNEU, 2017, 763. (in Ukrainian)
[12] Sharma S. World Bank says ‘local economic development’ is need of hour; lists out these 3 focus areas.
Financial Express, (2020). Available at: https://www.financialexpress.com/economy/world-bank-says-
local-economic-development-is-need-of-hour-lists-out-these-3-focus-areas-rural-livelihood-
recovery/2090879/
[13] Disability and Development. Inclusive development. Handicap international. AFC. Available at:
http://www.hiproweb.org/fileadmin/cdroms/Handicap_Developpement/www/en_page61.html
[14] Gupta J., Pouw N., Ros-Tonen M. Towards an Elaborated Theory of Inclusive Development. European
Journal of Development Research, 27 (4) (2015), 541–559. doi: 10.1057/ejdr.2015.30
[15] Bos K., Gupta J. Inclusive development, oil extraction and climate change: a multilevel analysis of
Kenya. International Journal of Sustainable Development & World Ecology, 23 (6) (2016), 482–492. doi:
10.1080/13504509.2016.1162217
[16] Mende M., Salisbury L.C., Nenkov G., Scott M. Improving Financial Inclusion through Communal
Financial Orientation: How Financial Service Providers Can Better Engage Consumers in Banking
Deserts. Journal of Consumer Psychology, 30 (2) (2020), 379–391.doi: 10.1002/jcpy.1103
[17] Baradaran M. How the poor got cut out of banking. Emory Law Journal, 62 (3) (2013), 483–548. Available
at: https://scholarlycommons.law.emory.edu/elj/vol62/iss3/1
[18] Morgan D., Pinkovskiy M., Yang B. Banking deserts, branch closings, and soft information. Liberty Street
Economics. Federal Reserve Bank of New York, 2016 Available at:
https://libertystreeteconomics.newyorkfed.org/2016/03/banking-deserts-branch-closings-and-soft-
information.html
[19] Piff P.K., Stancato D.M., Martinez A.G., Kraus M.W., Keltner D. Class, chaos, and the construction of
community. Journal of Personality and Social Psychology, 103 (6) (2012), 949–962. doi: 10.1037/a0029673
[20] Demirguc‐Kunt A., Klapper L., Singer D., Ansar S., Hess J. Global Findex Database 2017: Measuring
financial inclusion and the fintech revolution. DC: World Bank, Washington, 2018. doi: 10.1596/978-1-4648-
1259-0
[21] Ostrom E. Joint management. Evolution of institutions of collective action. Nash chas, Kyiv, 2012.
(in Ukrainian)
[22] Liberto D. Elinor Ostrom. Investopedia. Available at: https://www.investopedia.com/terms/e/elinor-
ostrom.asp
[23] Nelep V., Fursenko M., Vysotsʹka O., Fursenko I. Management of common resources in rural areas: in
the field of Nobel Laureate’ research. Ekonomika APK, 10 (2014) 92–98. Available at:
http://eapk.org.ua/en/contents/2014/10/92 (in Ukrainian)
Address: Halyna Voznyak, Krystyna Patytska, Taras Kloba, SI “Institute of Regional Research named after
M. I. Dolishniy of the NAS of Ukraine”, 4 Kozelnytska St., Lviv, 79026 Ukraine.
E-mail: gvoznyak@gmail.com, pelechata@meta.ua, taras.kloba@gmail.com
Received: September 25, 2020; revised: November 15, 2020.
_____________________
Determination of Financial Factors in the Latest Theories of Economic Growth … 59
Возняк Галина, Патицька Христина, Кльоба Тарас. Детермінація фінансових чинників у новітніх
теоріях економічного зростання територіальних громад. Журнал Прикарпатського університету імені
Василя Стефаника, 7 (3) (2020), 49–59.
В сучасних умовах забезпечення економічного зростання територіальних громад потребує
теоретичного переосмислення та додаткового дослідження, позаяк нові виклики негативно
позначились на добробуті населення та подальшому поступі.
Метою статті є детермінація фінансових чинників у новітніх теоріях економічного зростання
територіальних громад.
Узагальнено наукові підходи до визначення поняття “територіальна громада” запропоновано
власний підхід та обґрунтовано основні ознаки територіальної громади (територіальна, природна,
соціально-психологічна, історично-культурна, організаційно-функціональна, політична, економічна).
Акцентовано на потребі нового теоретичного підходу до розвитку територіальних громад на основі
синтезу і поєднання засадничих орієнтирів і принципів сучасних наукових концепцій і моделей, їх
трактування в контексті місцевого розвитку. Досліджено новітні теорії економічного зростання, серед
яких: теорія ендогенного зростання, інклюзивного розвитку та теорія сталого розвитку (виділено
концепцію Е. Остром про управління спільними ресурсами громади). Обґрунтовано, що модель
ендогенного зростання орієнтована на скорочення нерівності економічного розвитку через
використання внутрішніх чинників, механізмів господарювання і управління територією. Виділено
два вектори досягнення інклюзивного розвитку територіальних громад: внутрішній – максимальне
залучення членів громади до процесу її розвитку; зовнішній – забезпечення доступу членів
територіальної громади до можливостей, якими забезпечені більш розвинені громади. Акцентовано,
що ключовою думкою у дослідженнях Е. Остром є досягнення ефективності та обґрунтування
способів управління громадою спільними ресурсами. В межах теорії інклюзивного розвитку
проведено огляд концепції фінансової інклюзії, якою порушується проблема розвитку громад –
“банківських пустель”. Вказано на існуванні взаємозалежності сприйняття членами територіальних
громад – “банківських пустель” можливостей та перспектив отримання банківських послуг, що
вимагає залучення теоретичних основ поведінкової економіки до вивчення цього питання.
Систематизовано чинники економічного зростання територіальних громад, які об’єднано у п’ять
груп: природні, чинники людського розвитку; соціальні, виробничі та фінансові. До фінансових
чинників економічного зростання територіальних громад віднесено: бюджетний; інвестиційний;
інфляційний; грошово-кредитний. Обґрунтовано доцільність застосування ідей новітніх теорій
економічного зростання у процесі аналізу фінансових факторів.
Ключові слова: територіальна громада, фінансові чинники економічного зростання,
ендогенне зростання, інклюзивний розвиток, сталий розвиток, спільні ресурси, фінансова інклюзія.
Comparative Economic Research. Central and Eastern Europe
Volume 23, Number 4, 2020
http://dx.doi.org/10.18778/1508-2008.23.27
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
An Empirical Study of the Effects
of Demographic Factors on Economic Growth
in Advanced and Developing Countries
Serhii Kozlovskyi
Sc.D., Professor, Vasyl’ Stus Donetsk National University, Vinnytsia, Ukraine
e-mail: s.kozlovskyy@donnu.edu.ua
Mykola Pasichnyi
Sc.D., Associate Professor, Kyiv National University of Trade and Economics, Kyiv,
Ukraine, e-mail: m.pasichnyi@knute.edu.ua
Ruslan Lavrov
Sc.D., Professor, Chernihiv National University of Technology, Chernihiv, Ukraine,
e-mail: lavrus2017@gmail.com
Natalya Ivanyuta
Sc.D., Associate Professor, Donetsk Law Institute of the Ministry of Internal Affairs
of Ukraine, Mariupol, Ukraine, e-mail: natalaivanuta9@gmail.com
Anton Nepytaliuk
Ph.D. student, Vasyl’ Stus Donetsk National University, Vinnytsia, Ukraine
e-mail: anton.nepytaliuk@gmail.com
Abstract
In this article, an updated approach to investigate the effects of demographic fac‑
tors on economic growth is proposed. The initial hypothesis was that these factors
significantly affected production proportions, determining development vectors. The
predictable shifts in production dynamics are considered for the institutional frame‑
work. The article investigates the statistically significant relationships between the
demographic variables and economic growth for the sample of the OECD countries
(excluding Columbia) and Armenia, Belarus, Bulgaria, Croatia, Georgia, Kazakhstan,
Romania, the Russian Federation, and Ukraine, from 1990 to 2017; unbalanced
http://dx.doi.org/10.18778/1508-2008.23.27
https://orcid.org/0000-0003-0707-4996
mailto:s.kozlovskyy@donnu.edu.ua
https://orcid.org/0000-0001-7663-776X
mailto:m.pasichnyi@knute.edu.ua
https://orcid.org/0000-0002-9655-44
67
mailto:lavrus2017@gmail.com
https://orcid.org/0000-0001-9177-9280
mailto:natalaivanuta9@gmail.com
https://orcid.org/0000-0002-7890-3889
mailto:anton.nepytaliuk@gmail.com
46
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
panel data was used. The investigation aimed to highlight the intrinsic interconnec‑
tion between the changes in demographic variables (e.g., the working‑age population
growth rate and the average life expectancy growth rate) and economic growth. Our
investigation focused on the issue of whether demographic influence on economics
was the same for advanced and developing countries in the sample. Over the peri‑
od, a significant increase in life expectancy adversely affected the real GDP per cap‑
ita growth rate. However, the empirical study pointed out that life expectancy was
strongly linked to nominal GDP per capita. In advanced countries, the demographic
indicator was considerably higher than in emerging markets. We found that the rise
in the working‑age stratum of the nation’s population radically reduced the output dy‑
namics as well, but that interconnection was not robust. The institutional framework
should be taken into account in order to achieve a favorable performance of public
governance in the long‑run. The main demographic variables should be properly fore‑
casted and calibrated for potential endogenous economic triggers. Both public and
private investments are important when considering the economic growth rates that
are achieved. We propose a balanced approach to macroeconomic policy regarding
both demographic and institutional determinants.
Keywords: population, human capital, demographic sustainability, institutional
framework, economic growth
JEL: E22, I30, J10, J18, J24, O10
Introduction
There are several indisputable reasons to investigate the population’s impact on eco‑
nomic development – with inherent social and demographic characteristics – regard‑
ing growth theories. Firstly, the actual level of public welfare is sensed and described
only through human consciousness. Secondly, a person with his own needs and desires
creates the initial tasks for public production and directly participates in that process.
So, the nation’s population simultaneously plays the roles of the aggregated producer
and consumer of goods and services. The individual’s economic behavior causes and,
at the same time, is caused by higher interests, which are represented by social groups
(e.g., nations, economic classes, strata, etc.). General changes in the population’s num‑
ber, density, and age structure unquestionably affect public production. After centuries
of relatively slow and uneven growth, the world population reached 1 billion nearly
two hundred years ago. Before the start of the first so‑called demographic transition,
there were countless births and deaths, human life expectancy was short, and the pop‑
ulation was generally young.
Due to the transition, mortality and then fertility seriously declined. The popula‑
tion growth rates accelerated and then – unequally for advanced and the third‑world
countries – they fell again, matching low fertility, extended life spans, and a rather old
population. During the second half of the 20th century, the total population growth
accelerated at an unprecedented rate. The aforementioned global demographic chang‑
47
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
es brought decisive changes, reshaped both the economic and demographic life‑cycles
of the individuals, and restructured communities. As a result, the current world pop‑
ulation exceeds 7.7 billion, and it is expected to increase at a constant rate for at least
the next several decades.
This has raised lots of economic, social, and ecological questions (e.g., the societal
costs of the elderly, the redistribution of responsibility between the generations, the
lack of food provision, global pollution, etc.). The population’s characteristics were
considered to be the main economic growth determinants. So, their complex impact
on the development processes is everlasting and should be evaluated properly. The is‑
sues related to the key factors of economic growth have been at the heart of economic
science since its origin. Nearly five centuries of profound investigations have produced
a plethora of sustainable development theories, but the uncertainty remains.
In the vast majority of those theories, the main demographic variables are regarded
as endogenous determinants. On the one hand, everyone possesses a unique combina‑
tion of productive capacities that should be viewed as a part of human capital. On the
other hand, the population is jointly characterized by an essential economic potential,
which eventually results in output. Thus, modern demography and economics merged
to propose some theoretical and practical statements on production improvement.
Even though demographic issues are traditionally associated with fertility and mortal‑
ity rates, in this paper, we consider demographic variables in a broad sense, including
aggregated knowledge, the purposeful skills of the workforce, the potential of educa‑
tion and public health, etc. Some of the above characteristics overlap, so it is crucial
to identify and separate their influence on the growth processes.
The ongoing shifts in the demographic structure have enabled national econ‑
omies to convert most of the benefits from factor accumulation and technologi‑
cal changes into income per capita growth. Both labor productivity and develop‑
ment processes were generally enhanced via three channels. Firstly, the downtrend
in population growth has limited stock dilution and simultaneously increased the
number of resources per capita. Secondly, the reduced fertility sanctioned the real‑
location of resources from quantity toward the quality of children, intensifying the
human capital formation and total labor productivity. Finally, the reduction in fer‑
tility rates changed the age distribution of the population. So, if the fraction of the
labor force in the population temporarily increased, productivity per capita could
be raised mechanically.
The overall influence of the population change on economic growth and perfor‑
mance is ambiguous. There are alternative possibilities that population growth is sup‑
portive, restrictive, or neutral to economic development. The changes in population
number and density are commonly interconnected with some shifts in the commu‑
nity’s age structure. The latter could be described as the population’s distribution
across different age groups. Human economic behavior varies depending on the dif‑
ferent stages of the individual’s life. Thus, nations with an enormously high propor‑
tion of children should devote most of their resources to childcare programs. That fact
48
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
depresses the pace of economic growth in the short‑run but could be associated with
intensified human capital acceleration in the long‑run.
By contrast, if most of a nation’s population belongs to the working‑age stratum,
the extended productivity of that group can produce the so‑called “demographic div‑
idend.” If the nation’s population consists of the elderly, the effects can be different.
On the one hand, the result can be similar to the case of a very young population,
when a large share of resources is consumed by a relatively less productive popula‑
tion segment, inhibiting economic growth. On the other hand, the elderly, for many
reasons (primarily, effective public health care), can maintain working capacity and
demonstrate significant labor productivity, especially in the tertiary and quaternary
sectors of the economy.
A demographic dividend should be emphasized that considers both productivity
and consumption. It should be specifically mentioned that a demographic dividend
in the modern scientific discourse exists in two different forms. The “first” demo‑
graphic dividend is caused by an increase in the share of countries’ populations that
are concentrated in the working ages. Economically active individuals form the main
factor responsible for development.
The “second” dividend is much more difficult to explain. A rapid growth in the el‑
derly stratum presumably strains the public pension and health care systems. Over
recent decades, this has led to pessimistic forecasts concerning future economic per‑
formance. Nevertheless, an aging population can be a source of a second demographic
dividend rather than an economic decline. While the economically productive pop‑
ulation stratum is declining, a major challenge for aging and aged societies is to pro‑
vide a favorable framework for specific old‑age consumption and to achieve a desira‑
ble level of public welfare. The legislative framework is indirectly connected with the
above problem, but the main task for smart public governance is to provide a favorable
configuration of the financial system.
Demographic dividends do not occur automatically; their scale is largely dependent
on the quality of public institutions. The weaknesses of possible pension programs (e.g.,
an unsustainable increase in public pension benefits or critical tax evasion) could off‑
set many of the potential demographic dividends. For example, if most of the increases
in labor supply are concentrated in the informal sector, which does not contribute to so‑
cial security finances, it can cause imbalances and a decline in public welfare. The most
significant factors for sustainable development demographic variables are represented
by both the qualitative and the quantitative parameters of the working‑age population.
But the age dependency ratio is not the only demographic characteristic that matters.
Fertility and mortality fluctuations affect the average life expectancy and determine the
age distribution between the population strata. Even though increased life expectancy
is interconnected with life quality, it usually reshapes public finances and potentially
induces a tax burden. The quality of human capital depends on aggregate public and
private productive expenses. Thus, the model of sustainable economic growth should
combine the parameters related to human and physical capital creation.
49
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
Development trends in advanced, emerging‑market, and third‑world economies are
incomparable. Moreover, in the above groups, a sub‑group of commodity economies
should be specified regarding a wide range of factors. Even though all economies are
dependent on the same endogenous development triggers, the scale and the proportion
of those growth determinants significantly vary. In this study, we primarily examine
advanced (OECD member‑states) as well as some emerging, post‑Soviet (Armenia, Be‑
larus, Bulgaria, Croatia, Georgia, Kazakhstan, Romania, the Russian Federation, and
Ukraine) economies over the periods of institutional transformation and sustainable
growth. We reveal the overall dual impact on economic development of an expansion
in the working‑age population stratum and the increased life expectancy.
Literature review
Rethinking Romer’s (1990) conceptual model of endogenous technical change, Malm‑
berg (1994) proposed combining it with human capital and the life‑cycle of savings
theories. He also argued that the population’s age structure was crucial. Due to the
profound analysis of the changes in financial behavior and human capital accumu‑
lation over the life‑cycle, a theory of the age pattern of economic growth effects was
highlighted. Bloom, Canning, and Sevilla (2001; 2003) examined the impact of pop‑
ulation change on economic growth, regarding the alternative positions that popu‑
lation growth restricted, promoted, or appeared to be neutral to economic growth
trends. They identified not only the impact of the size and growth rate of the popula‑
tion on economic performance but the effects of the age structure. The agents’ behav‑
ior was described as being highly dependent on the structure. They concluded that,
on the concept of a demographic dividend, the effect of an optimal working‑age pop‑
ulation combined with health care, and educational, financial, and human capital pol‑
icies could initiate decent cycles of wealth creation.
Due to the enormous range of empirical cases, the evidence on the relevance
of the shifts in the age structure for economic growth was highlighted. The concept
of a demographic dividend was further developed by Bloom et al. (2007; 2009). The
age structure was considered to be a crucial determinant of economic growth and
the main forecast objective. Boucekkine, de la Croix, and Licandro (2002) identi‑
fied and empirically proved that endogenous economic growth was caused by the
accumulation of generation‑specific human capital. While preferable shifts in the
survival probabilities resulted in an extended schooling period and later retirement,
their effect on economic growth was ambiguous. Demographic variables had sig‑
nificant medium‑term economic effects, but the numerical interdependencies in the
long‑run did not appear to be robust. Kozlovskyi et al. (2019) pointed out the es‑
sence of economic security management for an emerging economy under conditions
of globalization. The interrelation between security issues and life quality dynamics
was generally revealed.
50
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
Lee (2001; 2003) summarized the main evidence of the demographic transition
and the related issues over the last three centuries. Considering the retrospective data
on the multiple interrelations between population shifts and fiscal policy performance,
he sketched the possible demographic changes and their economic consequences for
advanced and emerging markets. Taking human capital theory into account, numer‑
ous papers are dedicated to the essential social and demographic determinants of both
economic and population growth. Gador (2012) revealed the empirical validity of the
main demographic theories and their relevance for a sound understanding of the tran‑
sition from the stagnation phase to sustainable growth. A significant increase in the
demand for human capital in the development process was suggested as being the main
trigger for fertility reduction as well as the transition to the present growth rates.
Acemoglu and Johnson (2007) and Hansen and Lønstrup (2015) proved that an in‑
crease in life expectancy over the second half of the 20th century simultaneously reduced
the real GDP per capita growth rate and fostered population growth. That dual conclu‑
sion was based on the fact that, due to medical breakthroughs, many advanced coun‑
tries have experienced high growth rates in life expectancy and population size, and low
growth rates in per capita GDP. Based on empirical evidence from Western economies
during the past century, Fernihough (2017) revealed the importance of the demographic
transition as a support mechanism for the growth of human capital. The impact of ed‑
ucation on fertility rates and human capital accumulation was also investigated.
Lucas Jr. (2015) examined the aggregate innovative potential of the nation as a re‑
sult of knowledge creation based on consistent schooling and skills improvement.
Meanwhile, the actual role of knowledge management was dependent on the initial
level of the country’s economic development and the quality of the institutional frame‑
work. Barro and Lee (2013) investigated how output was related to the stock of human
capital, which was determined by the total years of schooling and by the composition
of the workers’ educational attainment. Education had a significantly positive effect
on the output dynamics, optimizing the endogenous interrelations between the main
components of economic growth.
Significant conclusions were made regarding human capital production. Using panel
data, Pelinescu (2015) proved the value of a good education and a flexible training system
for sustainable economic growth. Knowledge diffusion in manufacturing goods and ser‑
vices, creative industries, and concrete efforts to establish a research‑intensive economy
were identified as the main triggers responsible for long‑term development. Hanushek
(2015) examined the possibilities for a tertiary education‑based improvement in public
production. No statistically significant interdependencies between the indicators men‑
tioned were revealed. Nevertheless, reasonable effects of education were observed. Ahsan
and Haque (2017) refuted the hypothesis that the years of schooling were unrelated to eco‑
nomic growth. According to their empirical study, a decisive influence of education could
be discerned after a particular economy exceeds a threshold development level.
Using a growth model with integrated variables from the supply and demand side,
Teixeira and Queirós (2016) assessed the direct and indirect effects of human capital
51
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
on output growth, including the interaction of human capital with the country’s in‑
dustrial specialization. Both human capital dynamics and the country’s productive
specialization were identified as the main economic growth determinants.
Economic development was strongly influenced by the composite effect of human
capital applications and structural change in high knowledge‑intensive industries.
Meanwhile, the sign of the observed effect depended on the type of economic model
and the analyzed period. Over a long‑time period (1960–2011), the cumulative impact
of the interaction between human capital and structural change appeared to be posi‑
tive for OECD countries. Nagarajan, Teixeira, and Silva (2016) reviewed the literature
regarding the aging population and its integral impact on economic growth, and they
discovered the main mechanisms by which aging affected development.
Applying different mathematical methods, Uddin, Alam, and Gow (2016) investi‑
gated population saving behavior regarding age structure, dependency ratio, savings
rate, and real GDP. The negative effect of population aging on advanced economies was
statistically proved. Meanwhile, McGrath (2016) concluded that the indicators of GDP,
capital stock, and human capital were co‑integrated. While the causalities from GDP
to capital stock and from capital stock to human capital were bidirectional, the cau‑
sality from GDP to human capital appeared to be unidirectional, but not vice versa.
As a result, the initial hypothesis that economic growth was caused by human capital
has been generally refuted.
Focusing on the differences in the mortality rate for comparative development,
Cervellati and Sunde (2015) proposed a unified growth theory – covering both demo‑
graphic and economic issues – and investigated the demographic transition’s mechan‑
ics. The results explained an essential part of the differences in economic development
(e.g., the timing of the takeoff ) across countries under study and the worldwide den‑
sity distribution of the main demographic variables.
Acemoglu and Restrepo (2017) disputed the negative effects of an aging popu‑
lation on economic development. The main theoretical statements on the possible
negative effects of an aging population on economic growth were empirically inves‑
tigated. Both the lower labor force participation and productivity decreasing of the
older employees were considered. The hypothesis that aging had a negative impact
on the savings‑to‑investment ratio and led to so‑called secular stagnation was not
supported. It should be mentioned that the authors applied a rather unusual meth‑
odology: all the population over 50 was identified as “aged,” irrespective of the per‑
son’s production activity and economic behavior. Cooley, Henriksen, and Nusbaum
(2019) identified persistent deceleration in economic growth rates of the four largest
advanced economies in Europe caused by a shift in the age‑cohort distribution. De‑
fining the impact of complex demographic factors on economic development, they
revealed some interdependencies between the total factor productivity, capital ac‑
cumulation, labor supply, and population growth. They proved that the effects of an
aging population on economic growth distorted individual factor‑supply choices re‑
garding the pension systems.
52
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
Ahmad and Khan (2019) empirically investigated whether the demographic tran‑
sition and the dynamics of human capital mattered for economic growth for a repre‑
sentative sample of the developing world. The positive lagged contribution of the eco‑
nomically active population and the labor force participation rate in economic growth
were identified. Kozlovskyi et al. (2018) investigated the Ukrainian agrarian sector’s
regional peculiarities in the context of sustainable development. As an essential pre‑
condition for economic growth, they highlighted a strong interconnection between
sound management in the above sphere and the human capital quality.
Regarding the shift in advanced countries from industrial to knowledge economies,
Faggian, Partridge, and Malecki (2017) investigated the underlying causes of endog‑
enous economic development. The main prerequisites for growth were defined as in‑
tensified creativity, an entrepreneurship environment, and expanded human capital;
those factors were linked to the demographic parameters of the nation. While the in‑
terrelation between human capital (measured by educational attainment) and busi‑
ness environment (characterized by the intensity of small and medium‑sized firms)
appeared to be statistically interconnected with subsequent growth, other factors (e.g.,
the share of creative class workers, the share of advanced technology industries, etc.)
were described as insignificant. Meanwhile, Cuaresma et al. (2018) assessed the po‑
tential contribution of future educational attainment to economic growth and income
convergence. They suggested that income convergence dynamics and human capital
acted as important drivers for real income growth.
The aims
This paper investigates the interrelation between selected demographic variables and
the main economic variables regarding OECD and some developing countries. The
possible and predicted demographic dividends and the general character of the im‑
pact of the demographic transition on economic development processes are exam‑
ined. We try to find effective public management measures regarding the highlighted
demographic trends.
Methods and data
Sustainable economic development is dependent on a dynamic interrelation between
economic and demographic factors. Their overall effect is described by a production
function (1):
( )1 2, , , ,nY f x x x= (1)
53
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
where Y – the national production capacity or annual economic growth;
x1, x2, …, xn – the most essential economic and demographic factors.
Those factors are deeply interconnected with the category of human capital. Mean‑
while, all the significant elements of the aforementioned category are mostly insepara‑
ble and overlap. In our investigation, public production is defined by the Cobb‑Doug‑
las function (2).
,ij ij ij ijY A L K
α β= ∗ ∗ (2)
where
Yij – real GDP of country j in the year i;
Aij – the total factor productivity coefficient of country j in year i;
Lij – the labor input of country j in year i;
Kij – the capital input of country j in year i;
α, β – the output elasticities of labor and capital, respectively, while α + β = 1.
In present conditions, all production factors should be regarded as imperfect comple‑
ments. Public welfare could be described by the real annual GDP per capita growth
rate. The latter is dependent on the main productive factors, namely, physical and hu‑
man capital. If the real GDP per capita growth rate is decomposed into several con‑
ditionally independent variables, multiplicative function (2) can be transformed into
an additive one (3):
ij 0 1 ij 2 3growt demogr hum_cap contr ,ij ijγ γ γ γ ε= + + + + (3)
where
growthij – the real GDP per capita growth rate of country j in year i;
demogrij – the demographic variables of country j in year i;
hum_capij – the other human capital variables (indirectly related to the demographic
ones) of country j in year i;
contrij – the economic controls (related to the physical capital) of country j in year i.
The OLS method was applied to evaluate the overall impact of demographic and other
selected determinants on economic development.
Demographic variables are traditionally associated with fertility and mortali‑
ty rates. Broadly, the population’s dynamics depend not only on natural factors but
on mechanical ones as well (e.g., migration). We strongly believe that the overall demo‑
graphic impact on economic growth is represented by the permanent changes in the
working‑age population stratum and expected life span dynamics. So, the demograph‑
ic variables of our study consist of the working‑age population growth rate (WAPopgr)
and the average life expectancy growth rate (LifeExpgr). The other human‑capital‑re‑
lated essential economic growth determinant is represented by the composite pub‑
lic and private expenses on research and development activities (RD%GDP), regarded
as a percentage of GDP.
54
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
In addition, we impose two economic controls: public expenditures (PubExp%GDP)
and total investment (TInv%GDP) as percentages of GDP. Public spending characteriz‑
es the scale of GDP redistribution and the government’s role in welfare creation pro‑
cesses. That variable aggregates both the productive expenses (related to human cap‑
ital formation) and the other expenditures with an ambiguous impact on economic
growth (regarded as unproductive). Aggregating public and private financial activity
simultaneously, the total investment indicator is related primarily to the physical cap‑
ital production of the Cobb–Douglas function.
We used a panel data analysis over the period 1990–2017. The sample included the
economies of the OECD countries (excluding Columbia) and the economies of Arme‑
nia, Belarus, Bulgaria, Croatia, Georgia, Kazakhstan, Romania, the Russian Federa‑
tion, and Ukraine. Due to a critical lack of information on several emerging markets
in the early 1990s, the panel data was unbalanced. Because the vast majority of the
studied emerging economies successfully conducted institutional and structural re‑
forms before joining the EU (and its principal formation ended around 2004–2005),
we examined two periods: 1990–2004 and 2005–2017.
The main sources of our empirical data were the databases of the World Bank and
the IMF. Some essential data were drawn from the databases of the OECD and the
European Commission. Summary statistics data for the sample regarding the three
periods are presented in Table 1.
Table 1. Summary statistics
Variables Period Observations Mean Standard deviation Max Min
GDPpcgr
1990–2017 1032 2.35 3.64 23.99 –14.
56
1990–2004 465 2.89 3.25 15.31 –12.16
2005–2017 567 1.90 3.88 23.99 –14.56
LifeExpgr
1990–2017 1032 0.31 0.38 2.34 –1.
59
1990–2004 465 0.31 0.37 2.34 –1.59
2005–2017 567 0.30 0.38 2.10 –1.03
WAPopgr
1990–2017 1032 0.34 0.94 4.93 –4.08
1990–2004 465 0.48 0.85 4.93 –4.08
2005–2017 567 0.23 1.00 3.03 –2.48
R&D%GDP
1990–2017 1032 1.52 0.94 4.58 0.08
1990–2004 465 1.41 0.81 4.19 0.19
2005–2017 567 1.60 1.03 4.58 0.08
PubExp%GDP
1990–2017 1032 41.39 9.47 68.03 13.79
1990–2004 465 42.22 9.89 68.03 13.79
2005–2017 567 40.71 9.06 65.05 18.
63
TInv%GDP
1990–2017 1032 23.70 4.59 43.81 10.22
1990–2004 465 23.70 4.01 39.02 11.89
2005–2017 567 23.71 5.02 43.81 10.22
Source: authors’ own calculation based on The International Monetary Fund Database (2019), The World
Bank Open Data (2019), The European Commission Database (2019), and The OECD Data (2019).
55
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
Over the period 1990–2017, all the analyzed indicators varied significantly. While
the volatility – characterized by the standard deviation – of the public expendi‑
tures‑to‑GDP ratio reduced slightly, the volatility of the other examined indexes in‑
creased. Meanwhile, the aforementioned ratio was characterized by the highest stand‑
ard deviation, which equaled 9.47%. This was due to the remarkable differences in the
sampled countries’ institutional framework, fiscal policies, and economic models.
Results
Sustainable growth is described as the ultimate and primary objective of an econom‑
ic policy in the long‑run. Different demographic variables are traditionally integrated
into development programs and strategies as their most significant indices. Yet, the
actual role of the population’s characteristics as the growth triggers remains unknown.
Set by the authorities due to electoral obligations and commitments regarding the
mutual interconnection between political and business cycles, some declarative goals
in the distinct fields (e.g., demographics, public finances, etc.) can contradict each oth‑
er and deteriorate the analyzed system’s overall effect. Given the above, a complex nu‑
merical investigation of the contribution of both demographic and non‑demographic
factors to economic growth was carried out.
Sanchez‑Romero, Lee, and Prskawetz (2018) pointed out that differences in life ex‑
pectancy are observed not only between different countries but between high and low
socioeconomic groups as well. That hypothesis is extremely important when societies
with significant inequalities are analyzed. However, in our investigation, both life ex‑
pectancy and economic development indicators are regarded as the universal charac‑
teristics of a particular nation’s population. Figure 1 represents the interrelation be‑
tween the mean GDP per capita (expressed in current US $) and total life expectancy
at birth (expressed in years) in the sample over the period 1990–2017. The observed
interdependency appeared to be statistically significant and quite robust (R2 = 0.58).
Regarding the empirical data on the mean GDP per capita, the sample was divided
into three sub‑samples. The 1st sub‑sample included countries with a mean GDP per
capita lower than US $12,500.00; the 2nd sub‑sample – countries with a mean GDP
per capita from US $12,500.01 to US $37,500.00; the 3rd sub‑sample – countries with
a mean GDP per capita higher than US $37,500.01.
The vast majority of post‑Soviet countries were included in the 1st sub‑sample due
to their rather unfavorable endogenous social and economic conditions in the ear‑
ly 1990s. The 1st sub‑sample also included Chile, Mexico, and Turkey. Meanwhile,
over the entire period, Slovenia appeared to be the only post‑Soviet country with
a sufficiently high average GDP per capita that was equal to US $16,221.94. Con‑
sidering the entire sample, the countries of the 1st sub‑sample were characterized
by the lowest average life expectancy; the indicator varied from 67.47 years in Ka‑
zakhstan to 75.64 years in the Czech Republic. The average life expectancy in Slove‑
56
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
nia (77.03 years) was slightly lower than in Chile (77.08 years). In the 2nd and the 3rd
sub‑samples, the interconnection was generally the same, but its statistical density
appeared to be weaker. The highest average life expectancy at birth was observed in Ja‑
pan (81.60 years). Australia, Italy, and Spain (from the 2nd sub‑sample) as well as Ice‑
land, Sweden, and Switzerland (from the 3rd sub‑sample) formed a group of countries
with an average life expectancy that exceeded 80.00 years. Kazakhstan, the Russian
Federation, and Ukraine formed a group of countries with the lowest average life ex‑
pectancy, which did not exceed 70.00 years. The group was also characterized by the
lowest average GDP per capita.
Figure 1. The average GDP per capita (US $) and total life expectancy at birth (years) in selected
countries over the period 1990–2017
Source: authors’ own calculation based on The World Bank Open Data (2019).
In the entire sample, Luxembourg was characterized by an enormously high mean
GDP per capita that was equal to US $75,070.34. The indicator rose significantly from
US $34,645.14 in 1990 to US $104,498.74 in 2017. The life expectancy indicator varied
from 75.01 years to 82.69 years, respectively. We did not exclude the data on Luxem‑
bourg from the entire sample, but that fact was considered important for subsequent
analysis. It should be specifically mentioned that the standard deviation of average
life expectancy at birth was equal to 3.96 years, while the standard deviation of mean
GDP per capita equaled US $17,630.46. Over the investigated period, the total life ex‑
pectancy at birth in most advanced countries has reached the biological limits. The
level of GDP per capita varied significantly. The results of the above analyses should
be proved in a further investigation.
57
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
As it was numerically proved, demographic variables have fundamentally affected
economic development. Bloom et al. (2007) demonstrated that an increase in popula‑
tion was primarily observed in the non‑working‑age stratum, affecting both consump‑
tion and investment behavior and reducing economic growth. According to Pasichnyi
et al. (2017; 2019), in both advanced and emerging markets economies, an increase
in the total population had a significantly negative impact on their development. That
situation was generally caused by negative shifts in the population’s age structure. The
influence of the human development index on the resultative variable unexpectedly
appeared to be negative as well. That requires further investigations.
Considering the sample and the time scale, the real GDP per capita growth rates
were unsustainable and hugely dependent on the mutual interconnections between the
economic development determinants. Over the period 1990–2017, all the analyzed var‑
iables appeared to be statistically significant (see Table 2, OLS1), while the investigated
demographic variables were negative to economic growth. If the average life expectan‑
cy increased by 1.00%, the decline in real GDP per capita was equal to 1.23%. It should
be specifically mentioned that the life expectancy growth rate was characterized by the
lowest volatility. Its standard deviation equaled 0.31% and showed a slight growth over
the period in almost all countries in the sample. The most rapid decline in the ana‑
lyzed indicator was observed in Iceland in 1995. It was accomplished by a reduction
in the real GDP per capita growth rate. The examined indexes were equal to –1.59%
and –0.43%, respectively. Meanwhile, over the entire period, the highest life expec‑
tancy growth rate was identified in Croatia in 2001. It was associated with rather high
economic growth. The investigated variables equaled 2.34% and 7.51%, respectively.
The interconnection between the indicators was uneven and ambiguous due to the
complex nature of the life expectancy growth rate, which was simultaneously related
to the life quality and the public finances’ architectonics.
Considering the periods of 1990–2004 (OLS2) and 2005–2017 (OLS3), the impact
of the life expectancy growth rate on economic development was negative and stat‑
ically significant in both cases. Thus, an increase in the life expectancy growth rate
by 1.00% reduced the real GDP per capita growth rate by 0.36% and 1.85%, respective‑
ly. This difference could be caused by a lack of information on some emerging econo‑
mies over the period 1990–1995.
Unexpectedly, over the entire period, the working‑age population growth rate ap‑
peared to be negative to economic growth. Between 1990 and 2004, if the working‑age
population growth rate increased by 1.00%, the real GDP per capita growth rates fell
by 0.94%. However, considering the same time‑scale, this variable was statistically in‑
significant. Between 2005 and 2017, if the working‑age population grew by 1.00%, the
real GDP per capita growth rate fell by 0.63%. And the interconnection between the
variables was statistically significant. As a result, between 1990 and 2017, the intercon‑
nection mattered and was negative. If the working‑age population increased by 1.00%,
the respective reduction in GDP per capita growth rate equaled 0.59%. Over the entire
period, the average annual working‑age population expansion equaled 0.34%, while
58
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
the standard deviation equaled 0.94%. Regarding the periods 1990–2004 and 2005–
2017, the average annual working‑age population growth rates equaled 0.48% and
0.23%, respectively. Summarizing the above, it should be mentioned that an insuffi‑
cient increase in the working‑age stratum accomplished by sustainable life expectan‑
cy growth causes population aging.
Table 2. Regressions of economic growth on demographic variables and controls, the sample
of 45 countries, 1990–2017, unbalanced panel
Variables
Period
OLS1 OLS2 OLS3
LifeExpgr
–1.2331 –0.3551 –1.8471
(0.268) (0.366) (0.367)
WAPopgr
–0.5891 –0.940 –0.6311
(0.109) (0.162) (0.147)
R&D%GDP
–0.4521 –0.5511 –0.226
(0.113) (0.177) (0.147)
PubExp%GDP
–0.0741 –0.0781 –0.0901
(0.012) (0.015) (0.018)
TInv%GDP
0.2351 0.1191 0.2891
(0.023) (0.035) (0.030)
R2 0.216 0.195 0.273
N 1032 465 567
Notes: The numbers in parentheses are the standard errors of the estimated parameters.
‘1’ denotes significance at the 1 percent level. R2 represents the adjusted coefficient of determination.
Source: the authors’ own calculation based on The International Monetary Fund Database (2019),
The World Bank Open Data (2019), The European Commission Database (2019),
and The OECD Data (2019).
Research and development (R&D) expenditures denote both public and private pro‑
ductive spending, closely associated with an increase in human capital. Thus, the ex‑
amined interconnection between R&D expenditures and the actual economic devel‑
opment level depended on many determinants. In general, R&D expenditures are
considered to be productive, but their overall effect on the national economy’s devel‑
opment level should be examined properly. The composite structure of R&D expendi‑
tures can contradict the main aims of economic development. Theoretically, if the most
significant economic advantages were received via direct government grants, the na‑
tional economy could be deemed paternalistic.
Meanwhile, the empirical data proved that the achieved economic development
level was indifferent and slightly negatively interconnected with economic growth.
Surprisingly, over the entire period, an increase in the R&D expenditures‑to‑GDP
ratio by 1.00% reduced the real GDP growth rates. And, in that case, the average an‑
nual decline in the resulting variable was equal to 0.45%. Considering the selected
time periods, the dynamic interrelation between the R&D expenditures‑to‑GDP ratio
and the real GDP per capita growth rate was significant over the period 1990–2004.
59
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
When the empirical base of our study was expanded to include some emerging East‑
ern and Central European economies, the statistical significance of the investigated
interconnection rapidly declined and appeared to be insignificant. Regarding the pe‑
riod 1990–2004, an increase in the R&D expenditures‑to‑GDP ratio by 1.00% was
interconnected with a reduction in the GDP per capita growth rate, which was equal
to 0.55%. As previously written, over the period 2005–2017, the observed interrelation
between the R&D expenditures‑to‑GDP ratio and the real economic growth rate ap‑
peared to be statistically insignificant. The standard deviation of the investigated hu‑
man capital‑related variable equaled 0.94%. Regarding the different time periods, that
specific characteristic grew from 0.81% to 1.03%.
According to Barro and Sala‑i‑Martin (2003), the total public expenditures could
and should be divided into two separate groups, productive and non‑productive, con‑
sidering their overall impact on production dynamics. Based on the empirical data,
the dominance of non‑productive public expenditures causes a decline in the real GDP
per capita growth rate. The public spending‑to‑GDP ratio variable was generally neg‑
ative to economic growth regarding the selected time‑scales. Moreover, the negative
impact of the investigated independent variable was observed for both advanced and
emerging market economies. The variable was hugely dependent on the model of the
national economy and the quality of the institutional framework. Due to the extended
time period and the quality of the sample, one can see that the public spending‑to‑GDP
ratio varied widely. Its standard deviation changed from 9.89% to 9.06%.
It should be specifically mentioned that R&D expenditures are hugely dependent
on their inherent structure. If the structure was rigid, it could be characterized as an
intrinsic aspect of the public spending policy. In emerging economies in the early
1990s, the latter was closely interconnected with the doctrine of paternalistic public
finances. Thus, public spending was often determined by the political rather than the
economic cycle. The electoral promises – both at the local and national levels – affect‑
ed the economic performance and quite often deteriorated it.
In general, the total investment indicator – represented by the composite public
and private financial efforts – positively affected the growth processes. Over the pe‑
riod 1990–2017, an increase of 1.00% in the total investment‑to‑GDP ratio was ac‑
complished by the simultaneous increase in the real GDP per capita growth rate that
was equal to 0.24%. The overall effect of investment over the period 1990–2004 (with
a respective coefficient that equaled 0.12%) was less significant compared with the
respective indicator over the period 2005–2017 (with a respective coefficient that
equaled 0.29%). This proved that the composition of the investment recourses real‑
ly mattered. Regarding all the analytical periods in the investigated model, the total
investment‑to‑GDP ratio was the only variable that showed a sustainable positive in‑
fluence on production.
The interrelation between the working‑age population stratum and the out‑
put growth rate should be investigated properly. The interdependency is shown
in Figure 2.
60
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
Figure 2. The average working-age population stratum growth rate and the average real GDP per capita
growth rate in selected countries over the 1990–2017 period, %
Source: authors’ own calculation based on The World Bank Open Data (2019).
In terms of the real GDP per capita growth rates, over the period 1990–2017, Ukraine
maintained a unique position with simultaneous negative average output dynamics
and a decline in the economic active population stratum. Some of the investigated ad‑
vanced countries, namely Germany, Italy, and Japan, were characterized by positive
GDP dynamics accomplished by a decrease in the working‑age population stratum
growth rate. Meanwhile, the influence of the average working‑age population stratum
growth rate on the output dynamics was ambiguous. There were many emerging mar‑
ket economies (Armenia, Bulgaria, Estonia, Georgia, Hungary, Latvia, Lithuania, Ro‑
mania, Russian Federation) with a negative working‑age population stratum growth
rate, while the respective output dynamics was positive. A number of advanced econ‑
omies were characterized by a positive working‑age population stratum growth rate.
Poland, Ireland, and South Korea should be mentioned as the countries with the best
indicators of population and output dynamics. The working‑age population stratum
should be considered the most productive, regarding the structure. Future studies
should focus on the population’s productive capacity.
The legislative framework was essential, as it had an important impact on the agents’
behavior and considering macroeconomic efficiency. If the terms of the national leg‑
islation were regarded as acceptable for the vast majority of the agents involved, pub‑
lic governance would achieve the best performance. Meanwhile, unfavorable national
fiscal legislation fostered migration processes due to the Tiebout hypothesis and de‑
creased the national economy’s final results. So, a significant increase in the popula‑
tion’s quality of life was closely interconnected with public governance and the respec‑
tive formal institutional framework.
61
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
Discussion
Based on the empirical data, one can see that over the past three decades, economic de‑
velopment has been hugely dependent on different demographic variables. Meanwhile,
the examined social and demographic indicators – the working‑age population and the
average life expectancy growth rates – had a rather negative impact on the output growth
rates. In the case of the adverse interrelation between the working‑age population and
the real GDP growth rates, a possible explanation can be derived from the quality of the
labor force. In our investigation, the working‑age stratum of the population was defined
according to the International Labour Organisation methodology; however, people aged
from 15 to 64 objectively possess incomparable working abilities and competencies. The
observed expansions in the above stratum could be caused by increases in the low‑skilled
and unskilled sub‑strata. The latter was described by relatively poor productive capaci‑
ties and a rather insignificant contribution to public production. Further investigation
should cover the structural peculiarities of the working‑age stratum.
The overall negative impact of an extended life expectancy can be explained due
to the same changes in the population’s distribution through the age strata. In ad‑
vanced and emerging market economies, longevity is directly connected to the tax
burden: increased life expectancy induces payments related to social contributions.
A very aged population is characterized by significant medical and recreation spend‑
ing – both public and private – in GDP. Moreover, the dynamics of investment and
consumption behavior are hugely dependent on the population’s age structure. At the
same time, longevity was described as a natural and direct consequence of high eco‑
nomic development.
In this article, the indirect human‑capital‑related economic growth determinant
was represented by the share of R&D expenditures in GDP. In modern economic
discourse, R&D expenditures are traditionally defined as productive. Nevertheless,
in our study, an increase in R&D spending was associated with a downtrend in pub‑
lic production. The possible explanation was interconnected with the structure of the
aforementioned expenditures. In several countries, R&D expenditures were primarily
financed through public funds. If the structure of government spending was infirm,
the efficiency of public spending declined significantly. However, the possible solution
was closely interconnected with the R&D activities and private business convergence.
If scientific decision‑makers were connected to the business programs, their overall
effect could be generally high. If the R&D activities were unconnected to the public
needs, real GDP would be crucially reduced.
Demographic sustainability should be integrated into the national economic doc‑
trine and determined as society’s ability to automatically support and – using implic‑
it compensators – restore its own structure in the context of social stratification. This
would refer to a set of significant parameters, including the level of economic activity
as well as educational, professional, and competence training. Demographic sustaina‑
bility optimizes productivity proportions of intellectual and physical capital, provides
62
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
intensification and continuity of production, and increases the welfare of the popu‑
lation. The complex demographic factors, in particular, the working‑age population
and the average life expectancy growth rates, heavily influence long‑term economic
growth. Demographic sustainability should be defined as a strategic task for nation‑
al socio‑economic policy. To achieve demographic sustainability, tight coordination
of social, fiscal, migration, and cultural policies is required.
Conclusion
Public production can be described as a complicated multidimensional process that
is highly dependent on a set of social, demographic, and economic factors. People in‑
fluence economic dynamics enormously as they are simultaneously producers and
consumers of goods and services. The demographic factors contribute to economic de‑
velopment, and the character of their influence should be investigated properly. We ex‑
amined the OECD countries (excluding Columbia) and Armenia, Belarus, Bulgaria,
Croatia, Georgia, Kazakhstan, Romania, the Russian Federation, and Ukraine over the
period 1990–2017. Our initial hypothesis was that demographic factors significantly
affected production proportions, determining development vectors. However, the ob‑
served demographic variables appeared to be slightly interconnected with the output
dynamics, regarding both advanced and developing countries. The possible and pre‑
dicted demographic dividends and the general character of the demographic transi‑
tion’s impact on the economic development processes in the OECD member‑states and
selected developing countries were examined. Potential public management measures
regarding the highlighted demographic trends were proposed.
Providing comprehensive research, we pointed out some dependencies between the
GDP per capita growth rates and the selected demographic variables (the working‑age
stratum and expected life‑span growth rates). We examined states with relatively low,
medium, and high development levels. Considering the fact that average life expec‑
tancy in the sample was generally dependent on the achieved economic development,
some conclusions were reached. In the emerging market countries (e.g., some of the
post‑Soviet states, as well as Chile, Mexico, and Turkey), the adverse interdependencies
between the extended life expectancy and the output dynamics were primarily caused
by the quality of the institutional framework. In the advanced countries (the OECD
member‑states), this interrelation appeared negative, too. However, the possible ex‑
planation was that the life‑span in those cases had reached biological limits; economic
growth in the developed countries was considerably slower than in developing ones.
In case of the adverse impact of the working‑age stratum on the output dynamics,
it was not only the quantity that mattered but the actual quality, as well. Even though
the impact of both demographic variables that were studied appeared not to be robust
for the entire sample, further research in the aforementioned area with respect to the
national economy’s peculiarities would be mattered.
63
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
In this investigation, we considered there to be three main groups of impact fac‑
tors on economic growth. The 1st group was demographic factors that directly related
to the quality of human capital. This group included the working‑age population and
the average life expectancy growth rates. The 2nd group was closely connected to hu‑
man capital and knowledge management, but indirectly. The investigated variable re‑
ferred to the R&D expenditures‑to‑GDP ratio. The 3rd group was economic controls
that primarily related to physical capital: the public spending‑to‑GDP and total in‑
vestment‑to‑GDP ratios. The entire period was divided into two separate periods, i.e.,
1990–2004 and 2005–2017.
The total sample was divided into three sub‑samples that took into consideration
average GDP per capita and mean life expectancy. We found that the above econom‑
ic and demographic characteristics were directly interconnected: higher life expec‑
tancy was observed in the most developed countries. Moreover, that connection ap‑
peared to bilateral: significantly high real GDP per capita increased life expectancy.
We identified three sub‑samples, regarding low, medium, and high average GDP per
capita. It was proved that countries with the lowest average life expectancy were si‑
multaneously characterized by relatively low real GDP per capita. High life expectan‑
cy was considered the logical and natural consequence of an effective public produc‑
tion structure.
Over the observed periods, the general interdependency – represented by mod‑
el 3 – appeared to be statistically significant and quite robust, while the impact of the
main indicators varied. The vast majority of the investigated variables had a significant‑
ly negative impact on the scale of public production. An increase in the life expectan‑
cy growth rate by 1.00% reduced the real GDP per capita growth rate by 1.23%. If the
working‑age population grew by 1.00%, the output was reduced by 0.59%. Surprising‑
ly, an increase in the R&D expenditures‑to‑GDP ratio by 1.00% slowed down the real
GDP per capita growth rate by 0.45%. An increase in the public spending‑to‑GDP ratio
by 1.00% reduced the output dynamics by 0.07%. The total investment‑to‑GDP ratio
was the only independent variable that had a positive influence on public production:
if the ratio increased by 1.00%, the output was increased by 0.24%.
In the numerous previously mentioned scientific investigations, the impact of de‑
mographic factors was traditionally included in the global influence of human capital
on economic development. In this particular study, we argued that the effect of de‑
mographics on economic growth could not be identified with the category of human
capital effect. Even though the active economic agents produced GDP, both the pro‑
duction and the consumption mattered. Human capital was commonly associated with
production capacity, while demographics determined both the aggregated demand and
supply. Thus, in this article, we considered the direct and indirect influence of demo‑
graphics on developed and emerging market economies. The “overlapping” of the var‑
iables was not critical but should be considered in future studies. We augmented the
above separation in this paper, taking the transformation experience of Central and
Eastern European states into account.
64
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
Regarding the selected time scales, the independent variables had, in general, a sim‑
ilar impact on the output dynamics. Over the period 1990–2004, the impact of the
working‑age population growth rate on economic development appeared to be statisti‑
cally insignificant. The same results were obtained when the entire sample was divided
into two sub‑samples, taking the actual development of the examined economies into
account. The empirical investigation proved there is a robust negative interconnection
between the observed variables. Meanwhile, the actual impact of demographic varia‑
bles still needs to be investigated properly.
References
Acemoglu, D., Johnson, S. (2007), Disease and development: the effect of life expectancy
on economic growth, “Journal of Political Economy”, 115 (6), pp. 925–985. https://
doi.org/10.1086/529000
Acemoglu, D., Restrepo, P. (2017), Secular stagnation? The effect of aging on economic
growth in the age of automation, “American Economic Review”, 107 (5), pp. 174–179.
https://doi.org/10.1257/aer.p20171101
Ahmad, M., Khan, R.E.A. (2019), Does demographic transition with human capital
dynamics matter for economic growth? A Dynamic Panel Data Approach to GMM,
“Social Indicators Research”, 142 (2), pp. 753–772. https://doi.org/10.1007/s11205
‑018‑1928‑x
Ahsan, H., Haque, M.E. (2017), Threshold effects of human capital: Schooling and eco‑
nomic growth, “Economics Letters”, 156, pp. 48–52. https://doi.org/10.1016/j.econ
let.2017.04.014
Barro, R.J., Lee, J.W. (2013), A new data set of educational attainment in the world,
1950–2010, “Journal of Development Economics”, 104, pp. 184–198. https://doi.org
/10.1016/j.jdeveco.2012.10.001
Barro, R.J., Sala‑i‑Martin, X. (2003), Economic growth, 2nd ed., MIT Press, Massachu‑
setts, https://mitpress.mit.edu/books/economic‑growth‑second‑edition (accessed:
21.11.2019).
Bloom, D.E., Canning, D., Sevilla, J. (2001), Economic growth and the demographic
transition, NBER Working Paper Series 8685, National Bureau of Economic Re‑
search, Cambridge, USA. https://doi.org/10.3386/w8685
Bloom, D.E., Canning, D., Sevilla, J. (2003), The demographic dividend: A new perspec‑
tive on the economic consequences of population change, Rand Corporation, Santa
Monica, Canada. https://www.rand.org/pubs/monograph_reports/MR1274.html
Bloom, D.E., Canning, D., Fink, G., Finlay, J.E. (2007), Does age structure forecast eco‑
nomic growth?, “International Journal of Forecasting”, 23 (4), pp. 569–585. https://
doi.org/10.1016/j.ijforecast.2007.07.001
Bloom, D.E., Canning, D., Fink, G., Finlay, J.E. (2009), Fertility, female labor force par‑
ticipation, and the demographic dividend, “Journal of Economic Growth”, 14 (2),
pp. 79–101. https://doi.org/10.1007/s10887‑009‑9039‑9
https://doi.org/10.1086/529000
https://doi.org/10.1086/529000
https://doi.org/10.1257/aer.p20171101
https://doi.org/10.1007/s11205-018-1928‑x
https://doi.org/10.1007/s11205-018-1928‑x
https://doi.org/10.1016/j.econlet.2017.04.014
https://doi.org/10.1016/j.econlet.2017.04.014
https://doi.org/10.1016/j.jdeveco.2012.10.001
https://doi.org/10.1016/j.jdeveco.2012.10.001
https://mitpress.mit.edu/books/economic‑growth‑second‑edition
https://doi.org/10.3386/w8685
https://www.rand.org/pubs/monograph_reports/MR1274.html
https://doi.org/10.1016/j.ijforecast.2007.07.001
https://doi.org/10.1016/j.ijforecast.2007.07.001
https://doi.org/10.1007/s10887-009-9039-9
65
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
Boucekkine, R., De la Croix, D., Licandro, O. (2002), Vintage human capital, demo‑
graphic trends, and endogenous growth, “Journal of Economic Theory”, 104 (2),
pp. 340–375. https://doi.org/10.1006/jeth.2001.2854
Cervellati, M., Sunde, U. (2015), The economic and demographic transition, mortality,
and comparative development, “American Economic Journal: Macroeconomics”,
7 (3), pp. 189–225. https://doi.org/10.1257/mac.20130170
Cooley, T.F., Henriksen, E., Nusbaum, C. (2019), The Growth‑Cost of Demographic
Change in Europe, “Review of Economic Dynamics”, 1352, pp. 1–30. https://econo
micdynamics.org/meetpapers/2019/paper_1352 (accessed: 1.12.2019).
Cuaresma, J.C., Doppelhofer, G., Huber, F., Piribauer, P. (2018), Human capital accu‑
mulation and long‑term income growth projections for European regions, “Journal
of Regional Science”, 58 (1), pp. 81–99. https://doi.org/10.1111/jors.12339
Faggian, A., Partridge, M., Malecki, E.J. (2017), Creating an environment for econom‑
ic growth: creativity, entrepreneurship or human capital?, “International Journal
of Urban and Regional Research”, 41 (6), pp. 997–1009. https://doi.org/10.1111/14
68‑2427.12555
Fernihough, A. (2017), Human capital and the quantity–quality trade‑off during the
demographic transition, “Journal of Economic Growth”, 22 (1), pp. 35–65. https://
doi.org/10.1007/s10887‑016‑9138‑3
Galor, O. (2012), The demographic transition: causes and consequences, “Cliometrica”,
6 (1), pp. 1–28. https://doi.org/10.1007/s11698‑011‑0062‑7
Hansen, C.W., Lønstrup, L. (2015), The rise in life expectancy and economic growth
in the 20th century, “The Economic Journal”, 125 (584), pp. 838–852. https://aca
demic.oup.com/ej/article‑abstract/125/584/838/5077886?redirectedFrom=fulltext
(accessed: 1.12.2019).
Hanushek, E.A. (2016), Will more higher education improve economic growth?, “Ox‑
ford Review of Economic Policy”, 32 (4), pp. 538–552. https://doi.org/10.1093/ox
rep/grw025
Koziuk, V., Dluhopolskyi, O. (2018), Resource curse: The role of weak institutions and
cronysectors, “Ideology and Politics Journal”, 1 (9), pp. 68–104.
Kozlovskyi, S., Baidala, V., Tkachuk, O., Kozyrska, T. (2018), Management of the sus‑
tainable development of the agrarian sector of the regions of Ukraine, “Montenegrin
Journal of Economics”, 14 (4), pp. 175–190. http://dx.doi.org/10.14254/1800‑5845
/2018.14‑4.12
Kozlovskyi, S., Grynyuk, R., Baidala, V., Burdiak, V., Bakun, Y. (2019), Economic Se‑
curity Management of Ukraine in Conditions of European Integration, “Montene‑
grin Journal of Economics”, 15 (3), pp. 137–153. http://dx.doi.org/10.14254/1800‑58
45/2019.15‑3.10
Kozlovskyi, S., Bilenko, D., Kuzheliev, M., Lavrov, R., Kozlovskyi, V., Mazur, H.,
Taranych, A. (2020), The system dynamic model of the labor migrant policy in eco‑
nomic growth affected by COVID–19, “Global Journal of Environmental Science
and Management”, 6, Special Issue (Covid–19), pp. 95–106. http://dx.doi.org/10.22
034/GJESM.2019.06.SI.09
Lee, R.D. (2001), Demographic change and fiscal policy, Cambridge University Press,
Cambridge.
https://doi.org/10.1006/jeth.2001.2854
https://doi.org/10.1257/mac.20130170
https://economicdynamics.org/meetpapers/2019/paper_1352
https://economicdynamics.org/meetpapers/2019/paper_1352
https://doi.org/10.1111/jors.12339
https://doi.org/10.1111/1468-2427.12555
https://doi.org/10.1111/1468-2427.12555
https://doi.org/10.1007/s10887-016-9138-3
https://doi.org/10.1007/s10887-016-9138-3
https://doi.org/10.1007/s11698-011-0062-7
https://academic.oup.com/ej/article‑abstract/125/584/838/5077886?redirectedFrom=fulltext
https://academic.oup.com/ej/article‑abstract/125/584/838/5077886?redirectedFrom=fulltext
https://doi.org/10.1093/oxrep/grw025
https://doi.org/10.1093/oxrep/grw025
http://dx.doi.org/10.14254/1800-5845/2018.14-4.12
http://dx.doi.org/10.14254/1800-5845/2018.14-4.12
http://dx.doi.org/10.14254/1800-5845/2019.15-3.10
http://dx.doi.org/10.14254/1800-5845/2019.15-3.10
http://dx.doi.org/10.22034/GJESM.2019.06.SI.09
http://dx.doi.org/10.22034/GJESM.2019.06.SI.09
66
Serhii Kozlovskyi, Mykola Pasichnyi, Ruslan Lavrov, Natalya Ivanyuta, Anton Nepytaliuk
Lee, R.D. (2003), The demographic transition: three centuries of fundamental change,
“Journal of Economic Perspectives”, 17 (4), pp. 167–190. https://doi.org/10.1257/08
9533003772034943
Lucas Jr., R.E. (2015), Human capital and growth, “American Economic Review”,
105 (5), pp. 85–88. https://doi.org/10.1257/aer.p20151065
Malmberg, B. (1994), Age structure ef fects on economic growth – Swedish evi‑
dence, “Scandinavian Economic History Review”, 42 (3), pp. 279–295. https://
doi.org/10 .1080/03585522.1994.10415889
McGrath, P. (2016), The Relationship between Human Capital and Economic Growth
in Ireland. http://dx.doi.org/10.2139/ssrn.2872739
Nagarajan, N.R., Teixeira, A.A. Silva, S.T. (2016), The impact of an ageing population
on economic growth: an exploratory review of the main mechanisms, “Análise So‑
cial”, 218, pp. 4–35. https://www.jstor.org/stable/43755167
Pasichnyi, M. (2017), Empirical study of the fiscal policy impact on economic growth,
“Problems and Perspectives in Management”, 15 (3), pp. 316–322. http://dx.doi.org
/10.21511/ppm.15(3‑2).2017.01
Pasichnyi, M., Kaneva, T., Ruban, M., Nepytaliuk, A. (2019), The impact of fiscal de‑
centralization on economic development, “Investment Management and Financial
Innovations”, 16 (3), pp. 29–39. http://dx.doi.org/10.21511/imfi.16(3).2019.04
Pelinescu, E. (2015), The impact of human capital on economic growth, “Procedia Eco‑
nomics and Finance”, 22, pp. 184–190. https://doi.org/10.1016/S2212‑5671(15)002 58‑0
Romer, P.M. (1990), Endogenous technological change, “Journal of Political Economy”,
98 (5, Part 2), pp. S71–S102. https://doi.org/10.1086/261725
Sanchez‑Romero, M., Lee, R.D., Prskawetz, A. (2018), Redistributive effects of different
pension structures when longevity varies by socioeconomic status in a general equi‑
librium setting. Population Association of America, Manuscript, pp. 1–56. https://
www.econstor.eu/handle/10419/203628 (accessed: 1.12.2019).
Teixeira, A.A., Queirós, A.S. (2016), Economic growth, human capital and structural
change: A dynamic panel data analysis, “Research Policy”, 45 (8), pp. 1636–1648.
https://doi.org/10.1016/j.respol.2016.04.006
The European Commission Database (2019), https://ec.europa.eu/eurostat/data/data
base (accessed: 21.11.2019).
The International Monetary Fund Database (2019), https://www.imf.org/en/Data (ac‑
cessed: 21.11.2019).
The OECD Data (2019), https://stats.oecd.org/ (accessed: 21.11.2019).
The World Bank Open Data (2019), https://data.worldbank.org/ (accessed: 21.11.2019).
Uddin, G.A., Alam, K., Gow, J. (2016), Population age structure and savings rate im‑
pacts on economic growth: Evidence from Australia, “Economic Analysis and Poli‑
cy”, 52, pp. 23–33. https://doi.org/10.1016/j.eap.2016.08.002
https://doi.org/10.1257/089533003772034943
https://doi.org/10.1257/089533003772034943
https://doi.org/10.1257/aer.p20151065
https://doi.org/10.1080/03585522.1994.10415889
https://doi.org/10.1080/03585522.1994.10415889
http://dx.doi.org/10.2139/ssrn.2872739
https://www.jstor.org/stable/43755167
http://dx.doi.org/10.21511/ppm.15(3-2).2017.01
http://dx.doi.org/10.21511/ppm.15(3-2).2017.01
http://dx.doi.org/10.21511/imfi.16(3).2019.04
https://doi.org/10.1016/S2212-5671(15)00258-0
https://doi.org/10.1086/261725
https://www.econstor.eu/handle/10419/203628
https://www.econstor.eu/handle/10419/203628
https://doi.org/10.1016/j.respol.2016.04.006
https://ec.europa.eu/eurostat/data/database
https://ec.europa.eu/eurostat/data/database
https://www.imf.org/en/Data
https://stats.oecd.org/
https://data.worldbank.org/
https://doi.org/10.1016/j.eap.2016.08.002
67
An Empirical Study of the Effects of Demographic Factors on Economic Growth…
Streszczenie
Badanie empiryczne wpływu czynników demograficznych na wzrost
gospodarczy w krajach rozwiniętych i rozwijających się
W artykule zaproponowano zaktualizowane podejście do badania wpływu czynników
demograficznych na wzrost gospodarczy. Wstępna hipoteza zakładała, że czynniki
te w istotny sposób wpływają na proporcje produkcji, determinując kierunki rozwoju.
Ramy instytucjonalne uwzględniały przewidywalne zmiany dynamiki produkcji. W ar‑
tykule zbadano, wykorzystując niezbilansowane dane panelowe, istotne statystycznie
związki między zmiennymi demograficznymi a wzrostem gospodarczym dla krajów
OECD (z wyłączeniem Kolumbii) oraz Armenii, Białorusi, Bułgarii, Chorwacji, Gruzji,
Kazachstanu, Rumunii, Federacji Rosyjskiej i Ukrainy w latach 1990–2017. Badanie
miało na celu podkreślenie związku między kształtowaniem się zmiennych demo‑
graficznych (np. tempa wzrostu populacji w wieku produkcyjnym i tempa wzrostu
średniej długości życia) a wzrostem gospodarczym. Badanie było próbą odpowiedzi
na pytanie czy wpływ czynników demograficznych na gospodarkę był taki sam dla
badanych krajów rozwiniętych i rozwijających się. W omawianym okresie znaczny
wzrost oczekiwanej długości życia niekorzystnie wpłynął na dynamikę realnego PKB
per capita. Badanie empiryczne wykazało, że oczekiwana długość życia jest silnie po‑
wiązana z nominalnym PKB per capita. W krajach rozwiniętych ten wskaźnik demo‑
graficzny był znacznie wyższy niż na rynkach wschodzących. Okazało się, że wzrost
liczby ludności w wieku produkcyjnym radykalnie zmniejszył również dynamikę pro‑
dukcji, ale związek ten nie był silny. Aby uzyskać pozytywne efekty zarządzania pu‑
blicznego w perspektywie długoterminowej, należy uwzględnić ramy instytucjonalne.
Główne zmienne demograficzne powinny być odpowiednio prognozowane i skalibro‑
wane pod kątem potencjalnych endogenicznych czynników ekonomicznych. Dla osią‑
ganych wskaźników wzrostu gospodarczego ważne są zarówno inwestycje publiczne,
jak i prywatne. Autorzy sugerują wyważone podejście do polityki makroekonomicznej
w zakresie uwarunkowań zarówno demograficznych, jak i instytucjonalnych.
Słowa kluczowe: ludność, kapitał ludzki, równowaga demograficzna, ramy
instytucjonalne, wzrost gospodarczy
Copyright of Comparative Economic Research is the property of Wydawnictwo Uniwersytetu
Lodzkiego and its content may not be copied or emailed to multiple sites or posted to a
listserv without the copyright holder’s express written permission. However, users may print,
download, or email articles for individual use.
Not for Publication
Online-Only Appendix to “Blunt Instruments: Avoiding Common
Pitfalls in Identifying the Causes of Economic Growth”
We provide here additional details on the test statistics and inference procedures used in the paper
to assess the strength of identification in regressions based on instrumental variables procedures.
These weak instruments test statistics are often reported in empirical applications. However, the
inferential implications, particularly for the weak instruments test statistics, are often left unstated.
The first diagnostic tool for assessing the strength of identification is based on a Langrange-
Multiplier (LM) test for underidentification using the Kleibergen and Paap (
2
00
6
) rk statistic. This
test, readily implemented in Stata using the ranktest package, allows researchers to determine
whether the minimal canonical correlation between the endogenous variables and the instruments
is statistically different from zero. Another way of framing the test is by asking whether, after
partialling out exogenous covariates and cross-correlations with the other endogenous variables
and instruments, does the weakest correlation between an instrument and one of the endogenous
variables suffice to contribute enough independent variation to add to the empirical rank of the
instrument matrix? The p-values for this test are readily available after running the 2SLS estimation
using the ivreg2 package for Stata. The LM test for underidentification provides a lower hurdle
than the tests for weak instruments.
The second set of diagnostics are based on the Stock and Yogo (200
5
) characterization of weak
instruments using the first-stage F statistic and its multivariate analogue, the Cragg-Donald Wald
statistic or its robust counterpart, the Kleibergen-Paap Wald statistic. The usual approach in
the applied literature is to conclude that instruments are weak if these test statistics exceed the
critical values tabulated by Stock and Yogo. Much less common is the full use of the testing
procedures detailed in Section
4
of Stock and Yogo (2005), which provides richer probabilistic tools
for characterizing weak instruments. Here, we provide a few practical details on how to construct
p-values for the weak-instruments tests introduced in Section
3
.3 and used throughout our paper.
Adapting the empirical procedures in Gauss deployed in Yogo (2004), the formulation of p-values
in Stata proceeds as follows:
1
. Obtain the asymptotic threshold values for the concentration parameter Λ corresponding to
the weak instruments test (relative OLS bias or t-test size). These values are contained in a
number of Gauss matrices on Motohiro Yogo’s website.1 We have converted these into Stata
datasets (lambfitBias.dta and lambfitSize.dta) and provided them in supplementary
material available online.
1Available WWW: https://sites.google.com/site/motohiroyogo/home/publications/TestingWI_Programs.
zip?attredirects=0.
1
https://sites.google.com/site/motohiroyogo/home/publications/TestingWI_Programs.zip?attredirects=
0
https://sites.google.com/site/motohiroyogo/home/publications/TestingWI_Programs.zip?attredirects=0
2. Select the relevant value Λ̂ from the appropriate column and row of the lambfitBias or lamb-
fitSize matrices based on the number of endogenous variables, the number of instruments K,
and the level of bias or size distortion of interest. The relative OLS bias test is based on the
finite sample distribution of the 2SLS estimator and hence critical values can only be calcu-
lated for cases where there are at least two more overidentifying restritions than the number
of endogenous variables. In all specifications where this condition is not met, we report the
weak instruments test based on the size distortion of the t-test. Critical values for this test,
however, are not tabulated for cases with more than two endogenous variables. Thus, in
cases with more endogenous variables and/or instruments than available in the Stock-Yogo
tabulations, we take the penultimate available critical value in the given row and column of
the table.
3. Obtain the Cragg-Donald (ĈD) and Kleibergen-Paap (K̂P) Wald test statistics after esti-
mating the given 2SLS growth regression using ivreg2 in Stata.
4. Calculate the p-value for the given null hypothesis using the formula: p = 1−nchi2(K,K ×
Λ̂,K × ĈD), where nchi2 is the noncentral χ2 distribution with degrees of freedom K and
noncentrality parameter K × Λ̂. The p-value is valid for the Cragg-Donald statistic, which
assumes homoskedastic error terms. While the K̂P is robust to non-i.i.d. errors, its insertion
in the p-value formula does not immediately follow since the Stock-Yogo diagnostics were not
originally formulated for the non-i.i.d. case. Nevertheless, in characterizing weak instruments,
we follow others in the literature and report ĈD as well as K̂P for each specification. Thus,
while acknowledging that the p-values using the K̂P statistic are not asymptotically correct,
we report it along with that for the ĈD statistic for each of the given bias or size tests.
In Section 4.5, we employ the weak-instrument robust testing procedure of Kleibergen (2002) to
examine 2SLS dynamic panel equations in levels and first differences. This procedure has been
introduced as a higher power alternative to the Anderson-Rubin statistic. Here, we describe the
steps for applying this method. Suppose that the dynamic panel growth equation is given by
equation (3):
gi,t = β ln yi,t−1 + x
′
i,tγ1 + x̃
′
i,tγ2 + ψi + νi,t.
Suppose that x is a j-dimensional vector of endogenous growth determinants and x̃ is a k-
dimensional vector of exogenous growth determinants including indicators for the period t. Af-
ter constructing the appropriate instrument matrices for this equation in levels (LEV) and first
differences (DIF), the method proceeds as follows:
1. Define the j + 1 dimensional grid of possible values for the joint confidence region of β and
γ1. In Figure 3, we restrict attention to a relatively narrow range of parameter values. Our
principle was simply to start from the 2SLS point estimates and ensure that we chose a
sufficiently wide range of values on both sides of that point estimate to encompass many
values above and below zero. In the most general albeit infeasible case, one would want to
examine the whole real line for each of the j+1 parameters. Lastly, one defines the increments
over which to step along the range of values for a given parameter.
2. For the m-th j+1-tuple (βm,γm1 ), define ĝi,t = gi,t−β
m ln yi,t−1−x′i,tγ
m
1 for the LEV equation
and ∆̂gi,t = ∆gi,t −βm∆ ln yi,t−1 − ∆x′i,tγ
m
1 for the DIF equation.
2
3. Regress βm ln yi,t−1 (β
m∆ ln yi,t−1) on all LEV (DIF) equation instruments and the exoge-
nous covariates (∆)x̃′i,t. Obtain the predicted values
̂βm ln yi,t−1 ( ̂βm∆ ln yi,t−1). Repeat the
procedure for each of the j endogenous covariates in x′i,t.
4. Regress ∆̂gi,t on
̂βm∆ ln yi,t−1 and ∆̂x′i,tγ
m
1 . Do the same for the LEV equation.
5. Test the joint significance of the right-hand side variables and store the associated p-values
based on the large-sample χ2(j + 1) statistic for the given j + 1-tuple (βm,γm1 ).
6. Using the resulting dataset comprised of p-values and j + 1-tuples (βm,γm1 ) for the DIF and
LEV equations, plot two-dimensional joint confidence ellipses (using the user-written ellip
in Stata) for those values of j + 1-tuples such that the p-value is greater than 0.05. More
complex three-dimensional ellipsoids can be plotted in Matlab.
C Weak identification of nonlinear effects in Rajan & Subrama-
nian (200
8
)
If there are diminishing returns to capital in an economy, the effect of aid on growth can be nonlinear
and concave. Assuming a linear relationship can easily cloud such a relationship: the best linear fit
to a concave parabola has slope zero (presuming the full parabola is observed). Beyond this clear
theoretical reason to test for nonlinear effects, several important aid-growth regressions published
in the past decade have tested for and found a nonlinear relationship (e.g., Hansen and Tarp, 2001;
Dalgaard et al., 2004). In a small part of one table, Rajan and Subramanian attempt to test for
a nonlinear relationship between aid and growth, but their identification strategy does not allow
this. The instrumentation in these regressions is extremely weak. They do not report this.
Columns 1, 4, and
7
of Table C.1 show the underidentification and weak-instrument test statis-
tics (p-values) for three regressions in Table 4 of Rajan and Subramanian (2008), where the aid
effect is assumed linear. Instrumentation is strong. Columns 2, 5, and 8 show the same statistics
for three regressions in their Table 7 (Panel A), which include a squared aid regressor, and use ar
and its square as the only excluded instruments. The inclusion of the squared term causes instru-
mentation strength to collapse in the periods 1
9
80-2000 and 1990-2000, which is not reported in
RS. Strength is retained in the 1970-2000 period, but solely due to the presence of Guinea-Bissau in
the sample for that period (Guinea-Bissau is omitted from the sample in RS’s other two periods).
Without Guinea-Bissau, in columns 3, 6, and 9 , no useful degree of instrument strength is present
regardless of periodization. In fact, we cannot reject that the structural equation is underidentified.
All instrumentation in these nonlinear regressions, then, depends on a single country in a single
period. The RS instrument does not allow a meaningful test of a nonlinear effect of aid on growth.2
There is no escape from this problem within the RS framework: The instruments independent
of country size (I1–I7) do not explain aid variance, and the only strong instrument (population) is
plausibly invalid, as we demonstrate in Section 3.3 of the paper. A more fruitful way forward is to
find new instruments—better natural experiments to isolate the true effect of aid.
2One alternative procedure would be to carry out two separate zero-stage regressions, with regressands of linear aid
and squared aid, to create two constructed instruments. This does not, however, improve instrument strength.
3
Table C.1: Weak instruments in nonlinear specifications of Rajan and Subramanian (2008, Tables 4A and 7A)
Period 1970-2000 1980-2000 1990-2000
Aid Specification Linear Quadratic Quadratic Linear Quadratic Quadratic Linear Quadratic Quadratic
sans GNB∓ sans GNB∓ sans GNB∓
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Kleibergen-Paap LM test p-value† 0.0004 < 0.0001 0.397 0.0002 0.837 0.837 0.0
14
0.363 0.363
Cragg-Donald Waldstat‡ 31.63
13
.70 0.4
12
29.37 0.012 0.012 8.52 0.14 0.14
H0: t-test size>
10
% (p-value) 0.058 0.085 0.999 0.085 1.000 1.000 0.871 1.000 1.000
H0: t-test size>25% (p-value) 0.001 0.008 0.983 0.001 1.000 1.000 0.285 0.996 0.996
H0: relative OLS bias>10% (p-value) 0.005 0.109 0.999 0.008 1.000 1.000 0.538 1.000 1.000
H0: relative OLS bias>30% (p-value) 0.001 0.0210 0.993 0.001 1.000 1.000 0.275 0.998 0.998
Kleibergen-Paap Wald stat‡ 36.12 13.10 0.279 31.26 0.017 0.017 6.952 0.314 0.314
H0: t-test size>10% (p-value) 0.025 0.105 1.000 0.061 1.000 1.000 0.921 0.999 0.999
H0: t-test size>25% (p-value) 0.0001 0.0
11
0.990 0.001 1.000 1.000 0.388 0.988 0.988
H0: relative OLS bias>10% (p-value) 0.001 0.132 1.000 0.005 1.000 1.000 0.647 1.000 1.000
H0: relative OLS bias>30% (p-value) 0.0001 0.028 0.996 0.001 1.000 1.000 0.376 0.995 0.995
Notes: The estimates in columns 1, 3 and 7 are exact replications of columns 2, 3, and 4 in Table 4A of Rajan and Subramanian (2008). The estimates in columns 2, 4 and
8 are exact replications of columns 2, 3, and 4 in Table 7A of Rajan and Subramanian (2008). ∓ Guinea-Bissau is only included in the 1970-2000 regressions in the original
Table 7A. † The null hypothesis of the Kleibergen-Paap LM test is that the structural equation is underidentified (i.e., the rank condition fails). The test uses a rank test
procedure from Kleibergen and Paap (2006). ‡ The Cragg-Donald and Kleibergen-Paap Wald statistics correspond respectively to non-robust and heteroskedasticity-robust
multivariate analogues to the first-stage F statistics. Below each test statistic, we report the p-values from tests of whether (i) the actual size of the t-test(s) that βaid = 0
(and βaid2 = 0) at the 5% significance level is greater than 10 or 25%, and (ii) the bias of the IV estimates of βaid (and βaid2 ) are greater than 10 or 30% of the OLS
bias. In both cases, the critical values are obtained from Stock and Yogo (2005). Although critical values do not exist for the Kleibergen-Paap statistic, we follow the
approach suggested in Baum et al. (2007) and apply the Stock and Yogo critical values initially tabulated for the Cragg-Donald statistic. The critical values for (ii) are
(less conservatively) based on three instruments since one cannot calculate critical values in the (finite-sample)bias tests for the case of one endogenous variable and fewer
than three instruments.
4
D.1 Rajan & Subramanian Cross-Section Regressions (for 1990-2000)
Table D.1 reproduces the Rajan and Subramanian (2008) results from Table 2 in the paper with
an additional three columns covering their period 1990-2000. The results are similar to those for
the longer periods (1970- and 1980-2000) as discussed in the paper.
D.2 Sources of identification in the Hausmann et al (2007) five-year panel
Using the Hausmann et al. (2007) panel data for the five-year periodization, Table D.2 reports
the same set of specification tests as in Table 4 based on their ten-year periodization. As noted
in Section 3.3, the key result that export diversity (EXPY) increases growth does not hinge on
the excludability of population size in the same restrictive manner that it did in the shorter panel
with ten-year periodization. Although the result becomes null in column 6 when controlling for
country size directly in the second stage, the effect of EXPY on growth is relatively robust to
increasingly relaxing the excludability of the country size instruments in the levels and difference
equation instrument matrices in columns 2-5. Nevertheless, there still remain concerns about the
validity of the size instruments. While we cannot reject the validity of the size instruments on the
basis of the difference-in-Hansen statistic in the specifications of columns 2-4, further unpacking of
the levels and difference equation moment conditions in Section 4.4 revealed the validity of the size
instruments could not be rejected for the levels equation according to the Hahn et al. (2011) test,
the details of which are reported in the notes to Table 8.
D.3 Other measures of financial intermediation in Levine et al (2000)
Tables D.3 and D.4 estimate the same specifications as Table 6 using the two other measures of
financial depth in Levine et al. (2000): private credit/GDP and the ratio of commercial to central
bank credit, respectively. As noted in Section Section 4.2, the weak instruments problem of the
system GMM estimator holds for these additional measures of financial depth. This can be seen
most readily from the p-values for the weak-instruments tests reported in columns 6-9 of each
table. The one slight difference with the liquid liabilities results in Table 6 is that we can reject
the null of underidentification for the levels equation estimated using the collapsed instrument
matrix in column 9 of Tables D.3 and D.4. Although these instruments pass the lower hurdle of
underidentification, they remain weak.
D.4 Simulation results for a larger autoregressive parameter β = 0.8
Using the simulation procedure described in the paper, Figures D.1 and D.2 demonstrate the
performance of the difference and system GMM estimators of γ (the coefficient on the endogenous
growth determinant) when the persistence of the autoregressive parameter β increases from the
baseline value of 0.2 to 0.8 (see equation (4)). The results are qualitatively unchanged from the
baseline presented in Figures 1 and 2.
5
Table D.1: Instrumentation strength in Rajan and Subramanian (2008) cross-section regressions
Period 1970–2000 (N = 78) 1980–2000 (N = 75) 1990–2000 (N = 70)
“Zero-Stage” Specification Replication Colonial Population Replication Colonial Population Replication Colonial Population
vars. only vars. only vars. only vars. only vars. only vars. only
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Point estimate: Aid/GDP 0.096 –15.944 0.078 –0.004 –0.308 –0.028 –0.389 –0.035 –0.294
(0.070) (633.474) (0.067) (0.095) (0.389) (0.084) (0.194) (0.442) (0.144)
CLR confidence set∗: Aid/GDP [–0.027,0.292] (−∞,∞) [–0.039,0.254] [–0.186,0.232] (−∞,∞) [–0.194,0.170] [–1.463,–0.063] (−∞, 7.860] ∪ [8.071,∞) [–0.874,–0.021]
Kleibergen-Paap LM test (p-value)† 0.0004 0.978 0.0001 0.0002 0.282 0.0001 0.014 0.288 0.004
Cragg-Donald Wald stat‡ 31.63 0.001 35.90 29.37 1.41 40.54 8.52 1.69 12.86
H0: t-test size>10% (p-value) < 0.001 0.999 < 0.001 0.001 0.888 < 0.001 0.303 0.865 0.118
H0: t-test size>25% (p-value) < 0.001 0.980 < 0.001 < 0.001 0.341 < 0.001 0.013 0.295 0.002
H0: relative OLS bias>10% (p-value) < 0.001 0.996 < 0.001 < 0.001 0.772 < 0.001 0.
16
1 0.735 0.049
H0: relative OLS bias>30% (p-value) < 0.001 0.987 < 0.001 < 0.001 0.503 < 0.001 0.040 0.455 0.008
Kleibergen-Paap Wald stat‡ 36.12 0.001 31.62 31.26 1.41 39.65 6.95 1.18 9.00
H0: t-test size>10% (p-value) < 0.001 0.999 < 0.001 0.001 0.888 < 0.001 0.407 0.906 0.275
H0: t-test size>25% (p-value) < 0.001 0.984 < 0.001 < 0.001 0.340 < 0.001 0.026 0.385 0.011
H0: relative OLS bias>10% (p-value) < 0.001 0.997 < 0.001 < 0.001 0.770 < 0.001 0.239 0.801 0.142
H0: relative OLS bias>30% (p-value) < 0.001 0.990 < 0.001 < 0.001 0.502 < 0.001 0.071 0.546 0.034
Notes: In all specifications, the instrumental variable is aid/GDP predicted from the zero-stage regression. The dependent variable in all specifications is average annual
growth in GDP per capita over the period. Heteroskedasticity-robust standard errors in parentheses. Following the original paper, we retain the degrees-of-freedom
adjustment to the Kleibergen-Paap F and LM statistics based on robust standard errors. For each of the three periods, the first column is based on exact replication of
the baseline result in Rajan and Subramanian (2008, Table 4); the second column removes donor and recipient population terms from the zero-th stage specification used
to estimate the predicted aid/GDP instrument ar, retaining only the colonial ties indicators; the third column retains only the population terms in the zero-th stage. All
specifications include dummies for sub-Saharan Africa and East Asia. ∗The CLR confidence set corresponds to the weak-instrument robust confidence set obtained using
the conditional likelihood ratio test in Moreira (2003). †The null hypothesis of the Kleibergen-Paap LM test is that the structural equation is underidentified (i.e., the
rank condition fails). The test uses a rank test procedure from Kleibergen and Paap (2006). ‡In this special case of a single endogenous regressor, the Cragg-Donald and
Kleibergen-Paap Wald statistics reduce respectively to the standard non-robust and heteroskedasticity-robust first-stage F statistics. Below each, we report the p-values
from tests of whether (i) the actual size of the t-test that βaid = 0 at the 5% significance level is greater than 10 or 25%, and (ii) the bias of the IV estimates of βaid
reported in the table are greater than 10 or 30% of the OLS bias. In both cases, the critical values are obtained from Stock and Yogo (2005). Although critical values
do not exist for the Kleibergen-Paap statistic, we follow the approach suggested in Baum et al. (2007) and apply the Stock and Yogo critical values initially tabulated for
the Cragg-Donald statistic. The critical values for (ii) are (less conservatively) based on three instruments since Stock and Yogo do not tabulate critical values in the bias
tests for the case of one endogenous variable and fewer than three instruments.
6
Table D.2: The (non-?)excludability of country size in 5-year panels of Hausmann et al. (2007)
Dependent variable Growth Growth Growth Growth Growth Growth
Estimator IV∓ GMM-SYS∓ GMM-SYS GMM-SYS GMM-SYS GMM-SYS
Size Instruments? Yes Yes Yes, lev. Eq. Yes, diff. eq. No Yes
Size Excluded? Yes Yes Yes Yes Yes No
(1) (2) (3) (4) (5) (6)
log initial GDP/capita -0.030 -0.014 -0.015 -0.014 -0.008 -0.005
(4.820) (2.655) (2.764) (2.139) (1.394) (0.748)
log initial EXPY 0.074 0.045 0.046 0.046 0.036 0.016
(5.105) (4.097) (4.204) (3.828) (3.006) (1.112)
log human capital 0.004 0.004 0.003 -0.000 0.000 0.001
(1.781) (0.920) (0.904) (0.088) (0.067) (0.207)
log area 0.014
(3.979)
log population -0.009
(3.233)
Observations 604 604 604 604 604 604
Number of Countries 79 79 79 79 79 79
Number of Periods 8 8 8 8 8 8
Number of Instruments 2 75 75 75 73 75
Hansen J test (p-value) < 0.0001 0.507 0.502 0.467 0.623 0.267
Hansen J test excluding size instruments (p-value) 0.562 0.552 0.537 — —
Difference-in-Hansen test or C statistic (p-value)± 0.173 0.184 0.184 — —
Kleibergen-Paap LM test (p-value)† < 0.001 — — — — —
Cragg-Donald Waldstat‡ 39.09 — — — — —
H0: t-test size>25% (p-value) < 0.001 — — — — —
H0: relative OLS bias>30% (p-value) < 0.001 — — — — —
Kleibergen-Paap Wald stat‡ 34.25 — — — — —
H0: t-test size>25% (p-value) < 0.001 — — — — —
H0: relative OLS bias>30% (p-value) < 0.001 — — — — —
Notes: The dependent variable in all specifications is average annual growth over the period. The size instruments include log population and log area. The internal
instruments refer to the lagged levels and lagged differences of endogenous right-hand side variables in the respective difference and levels equations of the dynamic panel
GMM system of equations. ∓ Columns 1 and 2 are based on Hausmann et al. (2007, Table 9, Columns 2 and 4). The null hypothesis of the difference-in-Hansen test (or
C statistic, see Hayashi, 2000) is that the size instruments are valid. Following the original paper, we report heteroskedasticity-robust standard errors in parentheses and
retain associated degrees of freedom adjustments for the first-stage test statistics. See the notes to Table D.1 for more details on the Kleibergen-Paap and Cragg-Donald
tests, which apply in column 1 to the endogenous EXPY variable.
7
Table D.3: Weak instruments in dynamic panel regressions using private credit in Levine et al. (2000)
Difference Equation Levels Equation
Estimator GMM-SYS∓ GMM-SYS∓ OLS OLS-FD OLS-FE 2SLS 2SLS 2SLS 2SLS
Collapsed IV matrix No No — — — No Yes No Yes
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Private credit 1.522 1.494 0.823 0.807 0.945 0.284 9.291 1.451 2.320
(0.001) (0.001) (0.004) (0.046) (0.040) (0.826) (0.864) (0.033) (0.109)
Log initial GDP/capita -0.364 -0.398 -0.315 -14.016 -7.832 -12.420 0.552 0.593 1.109
(0.001) (0.001) (0.088) (0.000) (0.000) (0.000) (0.989) (0.448) (0.614)
Other parameter estimates omitted
N 359 359 345 323 345 323 323 345 345
Number of countries 74 74 74 74 74 74 74 74 74
Number of instruments 75 75 — — — 40 12 40 12
IV: Lagged levels Yes Yes — — — Yes Yes No No
IV: Lagged differences Yes Yes — — — No No Yes Yes
Kleibergen-Paap LM test (p-value) — — — — — 0.249 0.879 0.635 0.069
Cragg-Donald Waldstat — — — — — 0.73 0.004 0.67 0.51
H0: relative OLS bias>10% (p-value) — — — — — 1.000 1.000 1.000 1.000
H0: relative OLS bias>30% (p-value) — — — — — 1.000 1.000 1.000 0.987
Kleibergen-Paap Wald stat — — — — — 1.08 0.001 1.16 0.47
H0: relative OLS bias>10% (p-value) — — — — — 1.000 1.000 1.000 1.000
H0: relative OLS bias>30% (p-value) — — — — — 1.000 1.000 1.000 0.983
Notes: The dependent variable in all specifications is average annual growth in GDP per capita each period. ∓ Column 1 reproduces the published version of Levine et al.
(2000, Table 5, Column 2), and column 2 reports our best attempted replication using the DPD96 program for Gauss, the publicly available dataset, and a Gauss program
used to generate their results provided by Thorsten Beck. Further details on the difference in sample sizes across columns, our replication efforts, and the associated
differences in the Gauss and Stata programs for dynamic panel GMM regressions can be found in Appendix E.1. The following variables are included in the regressions
but suppressed in the table here for presentational purposes: government size, openness to trade, inflation, average years of secondary schooling, black market premium,
time period dummies and a constant. The first five of these variables are treated as endogenous. Following the original paper, we report p-values in parentheses. See the
notes to Table D.1 for more details on the Kleibergen-Paap and Cragg-Donald tests, which apply in columns 6-9 to the full set of endogenous right-hand-side variables.
8
Table D.4: Weak instruments in dynamic panel regressions using commercial vs. central bank credit in Levine et al. (2000)
Difference Equation Levels Equation
Estimator GMM-SYS∓ GMM-SYS∓ OLS OLS-FD OLS-FE 2SLS 2SLS 2SLS 2SLS
Collapsed IV matrix No No — — — No Yes No Yes
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Commercial vs. Central Bank Credit 2.437 2.293 1.243 2.526 3.397 3.649 -7.670 0.864 4.708
(0.001) (0.001) (0.033) (0.017) (0.010) (0.158) (0.979) (0.536) (0.198)
Log initial GDP/capita -0.117 -0.348 -0.138 -13.896 -7.912 -12.817 -8.985 0.568 1.470
(0.223) (0.015) (0.419) (0.000) (0.000) (0.000) (0.963) (0.347) (0.483)
Other parameter estimates omitted
N 359 359 345 324 345 324 324 345 345
Number of countries 74 74 74 74 74 74 74 74 74
Number of instruments 75 75 — — — 40 12 40 12
IV: Lagged levels Yes Yes — — — Yes Yes No No
IV: Lagged differences Yes Yes — — — No No Yes Yes
Kleibergen-Paap LM test (p-value) — — — — — 0.489 0.963 0.500 0.022
Cragg-Donald Waldstat — — — — — 0.71 < 0.001 0.88 0.76
H0: relative OLS bias>10% (p-value) — — — — — 1.000 1.000 1.000 1.000
H0: relative OLS bias>30% (p-value) — — — — — 1.000 1.000 1.000 0.954
Kleibergen-Paap Wald stat — — — — — 0.84 < 0.001 1.25 0.75
H0: relative OLS bias>10% (p-value) — — — — — 1.000 1.000 1.000 1.000
H0: relative OLS bias>30% (p-value) — — — — — 1.000 1.000 1.000 0.954
Notes: The dependent variable in all specifications is average annual growth in GDP per capita each period. ∓ Column 1 reproduces the published version of Levine et al.
(2000, Table 5, Column 3), and column 2 reports our best attempted replication using the DPD96 program for Gauss, the publicly available dataset, and a Gauss program
used to generate their results provided by Thorsten Beck. Further details on the difference in sample sizes across columns, our replication efforts, and the associated
differences in the Gauss and Stata programs for dynamic panel GMM regressions can be found in Appendix E.1. The following variables are included in the regressions
but suppressed in the table here for presentational purposes: government size, openness to trade, inflation, average years of secondary schooling, black market premium,
time period dummies and a constant. The first five of these variables are treated as endogenous. Following the original paper, we report p-values in parentheses. See the
notes to Table D.1 for more details on the Kleibergen-Paap and Cragg-Donald tests, which apply in columns 6-9 to the full set of endogenous right-hand-side variables.
9
Figure D.1: Power and size properties of GMM estimators in simulation results, β = 0.8
−1
−
.5
0
.5
−1
−.5
0
.5
−1
−.5
0
.5
.1 .5 .9 .1 .5 .9 .1 .5 .9 .1 .5 .9 .1 .5 .9
ω = −0.1, σ2 = 0.1 ω = −0.1, σ2 = 0.5 ω = −0.1, σ2 = 1 ω = −0.1, σ2 = 5 ω = −0.1, σ2 = 10
ω = −0.5, σ2 = 0.1 ω = −0.5, σ2 = 0.5 ω = −0.5, σ2 = 1 ω = −0.5, σ2 = 5 ω = −0.5, σ2 = 10
ω = −0.9, σ2 = 0.1 ω = −0.9, σ2 = 0.5 ω = −0.9, σ2 = 1 ω = −0.9, σ2 = 5 ω = −0.9, σ2 = 10
γ
(
ef
fe
ct
o
f d
o
n
gr
ow
th
)
ζ (persistence of d)
Difference GMM, β = 0.8, Reps = 500
−1
−.5
0
.5
−1
−.5
0
.5
−1
−.5
0
.5
.1 .5 .9 .1 .5 .9 .1 .5 .9 .1 .5 .9 .1 .5 .9
ω = −0.1, σ2 = 0.1 ω = −0.1, σ2 = 0.5 ω = −0.1, σ2 = 1 ω = −0.1, σ2 = 5 ω = −0.1, σ2 = 10
ω = −0.5, σ2 = 0.1 ω = −0.5, σ2 = 0.5 ω = −0.5, σ2 = 1 ω = −0.5, σ2 = 5 ω = −0.5, σ2 = 10
ω = −0.9, σ2 = 0.1 ω = −0.9, σ2 = 0.5 ω = −0.9, σ2 = 1 ω = −0.9, σ2 = 5 ω = −0.9, σ2 = 10
γ
(
ef
fe
ct
o
f d
o
n
gr
ow
th
)
ζ (persistence of d)
System GMM, β = 0.8, Reps = 500
Notes: The graphs show parameter estimates and 95% confidence intervals from simulations of the model in equation (4)
of the paper based on 500 draws of a sample size of 600 with 100 cross-sectional units and 6 time periods, fixed β = 0.8,
varying ζ ∈{0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9}, varying degrees of endogeneity ω ∈{−0.1,−0.5,−0.9}, and alternative
variances of the idiosyncratic shock, σ2 ∈{0.1, 0.5, 1, 5, 10}, where the variance of cross-sectional heterogeneity is fixed at
1. The dashed red line shows the true value of γ = 0.3 in the simulations.
10
Figure D.2: Weak identification in simulation results, β = 0.8
0
.5
1
0
.5
1
0
.5
1
.1 .5 .9 .1 .5 .9 .1 .5 .9 .1 .5 .9 .1 .5 .9
ω = −0.1, σ2 = 0.1 ω = −0.1, σ2 = 0.5 ω = −0.1, σ2 = 1 ω = −0.1, σ2 = 5 ω = −0.1, σ2 = 10
ω = −0.5, σ2 = 0.1 ω = −0.5, σ2 = 0.5 ω = −0.5, σ2 = 1 ω = −0.5, σ2 = 5 ω = −0.5, σ2 = 10
ω = −0.9, σ2 = 0.1 ω = −0.9, σ2 = 0.5 ω = −0.9, σ2 = 1 ω = −0.9, σ2 = 5 ω = −0.9, σ2 = 10
K
le
ib
er
ge
n−
P
aa
p
LM
te
st
(
p−
va
lu
e)
ζ (persistence of d)
Differences instrumented with levels
2SLS underidentification test, β = 0.8, Reps = 500
0
.5
1
0
.5
1
0
.5
1
.1 .5 .9 .1 .5 .9 .1 .5 .9 .1 .5 .9 .1 .5 .9
ω = −0.1, σ2 = 0.1 ω = −0.1, σ2 = 0.5 ω = −0.1, σ2 = 1 ω = −0.1, σ2 = 5 ω = −0.1, σ2 = 10
ω = −0.5, σ2 = 0.1 ω = −0.5, σ2 = 0.5 ω = −0.5, σ2 = 1 ω = −0.5, σ2 = 5 ω = −0.5, σ2 = 10
ω = −0.9, σ2 = 0.1 ω = −0.9, σ2 = 0.5 ω = −0.9, σ2 = 1 ω = −0.9, σ2 = 5 ω = −0.9, σ2 = 10
K
le
ib
er
ge
n−
P
aa
p
LM
te
st
(
p−
va
lu
e)
ζ (persistence of d)
Levels instrumented with differences
2SLS underidentification test, β = 0.8, Reps = 500
Notes: The graphs show p-values from a Kleibergen-Paap LM test for (the null of) underidentification in the levels and
differences equations from simulations of the model in equation (4) in the paper as detailed in the notes to Figure D.1.
See the notes to Table D.1 for details on the Kleibergen-Paap test.
11
We describe here our replications of the empirical growth studies assessed in Sections 3 and 4.
E.1 Levine et al (2000)
Despite the provision by Levine et al. (hereafter, LLB) of a publicly available dataset (Finan-
cial Intermediation and Growth dataset.xls) on a World Bank website (http://go.worldbank.
org/40TPPEYOC0), we faced a few difficulties in obtaining an exact replication of their dynamic
panel GMM results. Nevertheless, on the basis of our replications efforts described here, we are
highly confident that the subsequent OLS and 2SLS results that we report in Tables 6, D.3, and
D.4 are those that LLB would have gotten at the time they wrote, with precisely the same data.
In the process of attempting to replicate the original LLB results using exactly the same version
of their estimator in Gauss (DPD96), the same data, and the same program file provided by one
of the LLB authors (Thorsten Beck), we discovered a bug in DPD96.3 The bug produced different
two-step GMM estimates across consecutive runs of the same program over the same data, even
after reloading the data anew at each run. The result holds for the other two measures as well.
While the estimates do not vary wildly, we believe that this sort of non-deterministic potential
within this program for the deterministic dynamic panel GMM estimator could explain why the
LLB result cannot be reproduced exactly within Gauss (or Stata).
Table E.1 below compares the published parameter estimates in Table 5 of LLB to replications
using the original data and the DPD96 program in Gauss and the xtabond2 program in Stata.
Columns 2, 5, and 8 correspond to the estimates in column 2 of Tables 6, D.3, and D.4, respectively.
The replication based on DPD96 is quite close to the original published estimates. In only one
instance does the sign of the parameter estimate differ (inflation for the private credit outcome).
Turning to the Stata replications in columns 3, 6, and 9, we find larger differences with the estimates
obtained using Gauss despite setting all options in xtabond2 to mimic the DPD96 formulation (see
Roodman, 2009a). Roodman (2009b) reports similar difficulties replicating their results.4 The
other point to notice is that the sample size apparently differs in the Gauss and Stata replications.
This is actually not accurate, though. After inspection of the sample countries and years used in
each, we find that the samples are identical and that DPD96 output does not seem to be reporting
the actual sample size.
E.2 Rajan & Subramanian (2008)
The original Rajan and Subramanian dataset and code were kindly provided by the authors. As
noted in the paper, we exactly replicate their cross-section and dynamic panel results relevant to
our discussion. The analysis in Tables 1, 2, and 5 meanwhile required us to supplement their
original dataset with population data. The original dataset contained population ratios from zero-
stage regressions but not separate figures for period-initial receiving country population. For the
3Before proceeding to the replication, we removed three countries from the excel dataset, which were not listed as
part of the 74 country panel in Table 9 of their published paper.
4Our initial efforts at replication were done in consultation with Roodman. Subsequently, after correspondence with
Thorsten Beck, we obtained additional input into the Gauss replication. Our Stata replication for private credit
slightly differs from that in Roodman (2009b) for two reasons. First, we do not use the Windmeijer (2005) two-step
variance correction since this procedure was not available to LLB at the time of their study in the late 1990s. Second,
we drop three countries from the publicly available excel dataset, which were not listed among the 74 countries in
Table 9.
12
http://go.worldbank.org/40TPPEYOC0
http://go.worldbank.org/40TPPEYOC0
zero-stage regressions, the only database with sufficiently complete country coverage was the In-
ternational Monetary Fund’s online International Financial Statistics (accessed Sept. 9, 2007),
which had populations of all aid recipient countries in the Rajan and Subramanian dataset, except
for Bermuda, Kiribati, Turkmenistan, and Uzbekistan, which come from the World Bank’s World
Development Indicators 2007. In the main regressions, the extreme breadth of country coverage is
not needed and we took population from the Penn World Table 6.1, since real GDP/capita came
from that source. The correlation between the two sources’ population estimates is near unity.
In their dynamic panel GMM results, Rajan and Subramanian include the second through
seventh lags as instruments for the difference equation in both specifications. They note that they
are employing up to eight lags, but given that their panel consists of eight periods and only four
of the five year periods since 1985 are actually used due to missing data on their institutional
quality measure, their specifications naturally do not include eighth lagged levels as instruments
for any of the endogenous regressors. Also, although they claim to include an additional set of
time-invariant, excluded instruments in their main difference-equation specifications (geography,
ethnic fractionalization, Sub-Saharan Africa and East Africa), a Stata coding error results in their
being dropped from the equations regressing differenced endogenous variables on lagged levels.
In Table 7 of the paper, to be consistent with their published results, we exclude these four time-
invariant dummies from the Arellano-Bond regression in column 1 and the difference equation in the
Blundell-Bond regression in column 2, as well as the corresponding 2SLS regressions in subsequent
columns.
E.3 Hausmann et al (2007)
The original Hausmann et al. dataset and code were kindly provided by the authors. In Table D.2,
we exactly replicate their original pooled 2SLS and system GMM estimates for their panel based on
a five-year periodization. In Table 4, despite applying their original code to the original data, we
obtain slightly different estimates from those reported in their published paper for the system GMM
specification on the panel with ten-year periodization. The pooled 2SLS estimates are identical.
Nevertheless, the differences are trivial and in no way affect our main message in Table 4 (or the
key findings in Hausmann et al.’s original paper for that matter).
E.4 DeJong & Ripoll (2006)
The original DeJong and Ripoll dataset and code were kindly provided by the authors. We are
able to obtain exact replications of their dynamic panel GMM estimates in Table 2.
E.5 Hauk & Wacziarg (2009)
The original Hauk and Wacziarg dataset and code were kindly provided by the authors. We are
able to obtain exact replications of their dynamic panel GMM estimates in Table 13.
E.6 Voitchovsky (2005)
The original Voitchovsky dataset was kindly provided by the author. Using the DPD98 package (the
successor to DPD96) for Gauss as originally deployed by the author, we are able to obtain a close
replication of the system GMM estimates reported in Table 2 of the published paper. We could
not obtain an exact replication of the published results likely due to the bug in the DPD96 program
noted above and inherited by the DPD98.
13
Voitchovsky (2005) constructs a non-standard set of instruments, motivated by arguments
against using all the conventional Blundell-Bond moment conditions. For the DIF equation, the
instruments include twice and thrice lagged income per capita, lagged investment, the twice lagged
and difference in schooling rates, and the twice lagged difference in inequality measures. Anderson
and Hsiao (1982) were the first to suggest using twice lagged differences as instruments for the
lagged, differenced dependent variable in a dynamic panel setting (see also Arellano, 1989), though
the typical Arellano and Bond (1991) or Blundell and Bond (1998) applications instrument con-
temporaneous differences with lagged levels, retaining the first lagged difference as an instrument
for contemporaneous levels. For the LEV equation, the instruments include once lagged and dif-
ferenced investment and schooling rates; the inequality measures in levels and lagged income per
capita are treated (rather unconventionally) as exogenous in the levels equation.
14
Table E.1: Replicating Levine et al. (2000)
Original Replication Replication Original Replication Replication Original Replication Replication
SYS-GMM Estimator DPD96 DPD96 xtabond2 DPD96 DPD96 xtabond2 DPD96 DPD96 xtabond2
Collapsed IV matrix No No No No No No No No No
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Liquid liabilities 2.952 2.834 3.176
(0.001) (0.001) (0.000)
Private credit 1.522 1.494 1.451
(0.001) (0.001) (0.000)
Commercial vs. Central Bank Credit 2.437 2.293 1.383
(0.001) (0.001) (0.000)
Log initial GDP/capita -0.742 -0.792 -0.525 -0.364 -0.398 -0.268 -0.117 -0.348 -0.225
(0.001) (0.001) (0.000) (0.001) (0.001) (0.020) (0.223) (0.015) (0.008)
Government size -1.341 -1.419 0.249 -1.987 -1.841 -0.195 -1.13 -1.088 0.555
(0.001) (0.001) (0.481) (0.001) (0.001) (0.576) (0.001) (0.001) (0.000)
Openness to trade 0.325 0.372 -0.047 0.442 0.499 -0.016 0.497 0.620 0.646
(0.169) (0.124) (0.847) (0.010) (0.021) (0.929) (0.002) (0.001) (0.000)
Inflation 1.748 1.675 1.074 -0.178 0.055 -0.598 -1.772 -2.413 -1.802
(0.001) (0.001) (0.000) (0.543) (0.810) (0.007) (0.001) (0.001) (0.000)
Avg. yrs. secondary school 0.780 0.732 0.041 0.639 0.472 0.195 0.638 0.775 0.935
(0.001) (0.001) (0.743) (0.001) (0.001) (0.128) (0.001) (0.001) (0.000)
Black market premium -2.076 -2.014 -2.102 -1.027 -1.109 -1.062 -1.044 -1.395 -1.018
(0.001) (0.001) (0.000) (0.001) (0.001) (0.000) (0.001) (0.001) (0.000)
Constant 0.06 1.061 -4.301 4.239 4.042 1.300 -5.677 -4.001 -5.713
(0.954) (0.195) (0.002) (0.001) (0.001) (0.119) (0.001) (0.001) (0.277)
Observations 359 359 345 359 359 345 359 359 345
Number of countries 74 74 74 74 74 74 74 74 74
Number of instruments 75 75 75 75 75 75 75 75 75
Notes: The dependent variable in all specifications is average annual growth in GDP per capita each period. The following variables are included in the regressions but
suppressed in the table here for presentational purposes: government size, openness to trade, inflation, average years of secondary schooling, black market premium, time
period dummies and a constant. The first five of these variables are treated as endogenous. Following the original paper, we report p-values in parentheses based on
two-step estimates without the Windmeijer (2005) correction, which became available after the LLB study.
1
5
References
Anderson, T. W. and Cheng Hsiao, “Formulation and Estimation of Dynamic Models Using Panel Data,”
Journal of Econometrics, 1982, 18 (1), 47–82.
Arellano, M., “A note on the Anderson-Hsiao estimator for panel data,” Economics Letters, 1989, 31 (4), 337–341.
Arellano, Manuel and Stephen Bond, “Some Tests of Specification for Panel Data: Monte Carlo Evidence and
an Application to Employment Equations,” Review of Economic Studies, 1991, 58 (2), 277–297.
Baum, Christopher F., Mark E. Schaffer, and Stephen Stillman, “Enhanced Routines for Instrumental
Variables/GMM Estimation and Testing,” Stata Journal, 2007, 7 (4), 465–506.
Blundell, Richard and Stephen Bond, “Initial Conditions and Moment Restrictions in Dynamic Panel Data
Models,” Journal of Econometrics, 1998, 87 (1), 115–143.
Dalgaard, Carl-Johan, Henrik Hansen, and Finn Tarp, “On the Empirics of Foreign Aid and Growth,”
Economic Journal, 2004, 114 (496), 191–216.
DeJong, David N. and Marla Ripoll, “Tariffs and Growth: An Empirical Exploration of Contingent Relation-
ships,” Review of Economics and Statistics, 2006, 88 (4), 625–640.
Hahn, Jinyong, John Ham, and Roger Moon, “The Hausman test and weak instruments,” Journal of Econo-
metrics, 2011, 160 (2), 289–299.
Hansen, Henrik and Finn Tarp, “Aid and Growth Regressions,” Journal of Development Economics, 2001, 64
(2), 547–570.
Hauk, William and Romain Wacziarg, “A Monte Carlo Study of Growth Regressions,” Journal of Economic
Growth, 2009, 14 (2), 1381–4338.
Hausmann, Ricardo, Jason Hwang, and Dani Rodrik, “What You Export Matters,” Journal of Economic
Growth, 2007, 12 (1), 1–25.
Hayashi, F., Econometrics, Princeton, New Jersey: Princeton University Press, 2000.
Kleibergen, Frank, “Pivotal statistics for testing structural parameters in instrumental variables regression,”
Econometrica, 2002, 70 (5), 1781–1803.
and Richard Paap, “Generalized Reduced Rank Tests Using the Singular Value Decomposition,” Journal of
Econometrics, 2006, 133 (1), 97–126.
Levine, Ross, Norman Loayza, and Thorsten Beck, “Financial Intermediation and Growth: Causality And
Causes,” Journal of Monetary Economics, 2000, 46 (1), 31–77.
Moreira, Marcelo J., “A conditional likelihood ratio test for structural models,” Econometrica, 2003, 71, 1027–
1048.
Rajan, Raghuram and Arvind Subramanian, “Aid and Growth: What Does the Cross-country Evidence Really
Show?,” Review of Economics and Statistics, 2008, 90 (4), 643–665.
Roodman, David, “How to Do xtabond2: An Introduction To Difference and System GMM in Stata,” Stata
Journal, 2009, 9 (1), 86–136.
, “A Note on the Theme of Too Many Instruments,” Oxford Bulletin of Economics and Statistics, 2009, 71 (1),
135–158.
Stock, James H. and Motohiro Yogo, “Testing for Weak Instruments in Linear IV Regression,” in James H.
Stock and Donald W. K. Andrews, eds., Identification and Inference for Econometric Models: Essays in Honor of
Thomas J Rothenberg, New York: Cambridge University Press, 2005.
Voitchovsky, Sarah, “Does the Profile of Income Inequality Matter for Economic Growth?,” Journal of Economic
Growth, 2005, 10 (3), 273–296.
Windmeijer, Frank, “A finite sample correction for the variance of linear efficient two-step GMM estimators,”
Journal of Econometrics, 2005, 126, 25–51.
Yogo, Motohiro, “Estimating the elasticity of intertemporal substitution when instruments are weak,” Review of
Economics and Statistics, 2004, 86, 797–810.
16
D Further empirical and simulation results
D.1 Rajan & Subramanian Cross-Section Regressions (for 1990-2000)
D.2 Sources of identification in the Hausmann et al (2007) five-year panel
D.3 Other measures of financial intermediation in Levine et al (2000)
D.4 Simulation results for a larger autoregressive parameter =0.8
E Replicating growth studies
E.1 Levine et al (2000)
E.2 Rajan & Subramanian (2008)
E.3 Hausmann et al (2007)
E.4 DeJong & Ripoll (2006)
E.5 Hauk & Wacziarg (2009)
E.6 Voitchovsky (2005)
sustainability
Article
Natural Capital, Domestic Product and Proximate
Causes of Economic Growth: Uruguay in the Long
Run, 1870–2014
Silvana Sandonato 1 and Henry Willebald 2,*
1 PHES, Programa de Posgrado, Facultad de Ciencias Sociales, Universidad de la República,
11200 Montevideo, Uruguay; chivilive@gmail.com
2 Facultad de Ciencias Económicas y de Administración, Universidad de la República,
11200 Montevideo, Uruguay
* Correspondence: hwillebald@iecon.ccee.edu.uy; Tel.: +598-24-131-007
Received: 30 November 2017; Accepted: 27 February 2018; Published: 6 March 2018
Abstract: The debate on the relationship between natural resources abundance and economic growth
is still open. Our contribution to this field combines a long-run perspective (1870–2014) with the
study of a peripheral country in the world economy (Uruguay). The purpose is to build a historical
series of natural capital and contrast its level and evolution with the level and growth of GDP,
as well as the proximate causes of its economic growth (produced and human capital, exports and
terms of trade). We show that natural capital has tended to decline in importance in the economy,
while simultaneously becoming more diversified. Although this evolution is consistent in historical
terms, we do not find a causal relationship between the abundance of natural resources and economic
performance. Instead of a direct relationship, the proximate causes appear to have been important
in explaining the evolution of natural capital when we consider three stages of economic growth:
physical capital and terms of trade during the agro-exporter model; human capital and exports during
the period of import substitution industrialization; and terms of trade from the 1970s afterwards.
These factors cause natural capital but not the other way around, leading us to conclude that an
abundance of natural capital is an endogenous process.
Keywords: natural capital; economic growth; Uruguay
1. Introduction
We identify the concept of “natural resources” with those assets which originate from nature—land,
forests, minerals—and can be exploited for economic purposes. In principle, such assets should provide
at least three advantages for undeveloped economies [1]. First, the income flow from resource
exploitation can finance higher levels of public and private consumption, improving the living
standards of deprived segments of the society. Second, the extraction of natural resources can support
investment in physical and human capital, either directly by use of resource rents, or indirectly because
the natural resources can serve as a guarantee to borrow abroad. Third, since government can place
taxes on rents or directly on the exploitation of the resources, this can provide the revenues needed to
finance fundamental public goods, including infrastructure, health and education.
However, in recent decades, it has been noticed that large national resource wealth does not assure
success and, on the contrary, it may even seriously impinge on economic development. Many Latin
American, African and Asian countries possess large energy, mineral and forestry wealth and yet
their inhabitants continue to experience low quality of life [2]. The literature named this puzzling
phenomenon the “natural resource curse” [3], which refers to the paradox that economies endowed
with abundant natural resources tend to experience deficient economic growth and worse development
Sustainability 2018, 10, 715; doi:10.3390/su10030715 www.mdpi.com/journal/sustainability
http://www.mdpi.com/journal/sustainability
http://www.mdpi.com
http://dx.doi.org/10.3390/su10030715
http://www.mdpi.com/journal/sustainability
Sustainability 2018, 10, 715 2 of 26
outcomes than economies with scarce natural resources. The resource curse thesis has focused,
mainly, on non-renewable assets because such resources have the worst consequences on economic
growth [4,5].
In Section 2, we present a review of different approaches (in conceptual and empirical terms) that
refer to the “curse” and “blessing” of natural resources. The history of economic thought on the matter
has moved from an extended consensus about natural wealth constituting an engine of the economic
growth—a dominant idea from the end of the 19th century to the 1950s—to an extended confidence
about the low probability of creating conditions where resource-based development can be sustainable.
Nevertheless, recently, other authors have criticized this viewpoint and have questioned the intensity
and the causality of the relationship.
Can economic history offer new approaches about this discussion? Can the study of long-run
economic processes contribute with new insights for interpreting this apparent paradox? We answer
both questions affirmatively. The analysis of resource-rich countries is an interdisciplinary field
and “draws on macroeconomics, public finance, public policy, international economics, resource
economics, economic history and applied econometrics. It also benefits from collaboration with
political scientists and historians” [6] (p. 407). As well as, one of the “most interesting aspect[s] of
resource-abundant countries is not their average performance but their huge variation” [7] (p. 242).
In other words, to study particular cases constitutes a valuable effort to understand the dimensions
and the mechanisms behind the curse and the blessing of natural resources.
Our research is part of efforts to include long-run considerations in the debate and to identify
historical specificities that confirm that the relationship is not an immutable process but could respond
to a broad range of conditions, variables and circumstances. In Section 3, we explain that Uruguay is
an excellent case study for evaluating this matter for three reasons: (i) Uruguay is a country historically
described as a natural resources-abundant economy, but this characterization only holds for some
dimensions of the economic process; (ii) Uruguay is a case of abundance of non-mineral natural
resources which allows us to investigate an area where the literature always has paid little attention;
(iii) we develop a novel series of natural capital that represents a better approximation to the concept
of “abundance” in distinction from “dependence”, which has been at the core of recent literature.
This topic is one of the main contributions of our article in two ways. On the one hand, we propose
a long-run estimation of natural capital, annually, from 1870 to 2014, considering five components:
pasturelands, croplands, timber and non-timber forestry and protected areas. On the other hand,
we apply the World Bank’s methodology to account for the natural capital but we propose correcting
and improving upon the original assumptions to capture the historical specificity over such a long
period. Therefore, we devote Section 4 to the description of our empirical strategy and methodology,
including the principal assumptions and details related to each component.
We present the results in Section 5. To start with, we consider the main stylized facts based on
our estimates of the natural capital series (1870–2014) to evaluate the historical consistency of our
results. The growth in natural capital does not cause economic growth nor does the expansion of
GDP cause natural capital growth. Instead of this direct relationship, other channels likely served
as physical capital and terms of trade during the agro-exporter model, human capital and exports
in the industrialization period and terms of trade again from the 1970s afterwards in the so-called
re-globalization period. These factors cause natural capital but the causality in the other direction is
not confirmed, leading us to conclude that abundance of natural capital is an endogenous process.
We discuss the main results in Section 6 and we conclude in Section 7.
2. Theory and Empirical Approaches
We review, briefly, the literature about the relations between abundance of natural resources and
economic trajectories of countries and regions. First, we survey theoretical frameworks considering fives
approaches: natural resource abundance as a blessing; production structure approach; crowding-out
approach; institutional change and factor endowment; and the economic history approach [8].
Sustainability 2018, 10, 715 3 of 26
Second, we refer to empirical tests on the effect of natural resources on real evolution of the economy,
the effects on variables related to economic expansion and a critical view of the curse.
2.1. Theoretical Frameworks
2.1.1. The Blessing of Natural Resource Abundance
Initially, from a theoretical point of view the relationship between abundance of natural resources
and economic performance offered an extended consensus about natural wealth constituting an engine
of economic growth. This was the dominant idea until, at least, the mid-20th century.
In the last decades of the 19th century and up to WWI, several peripheral economies experienced a
real economic boom related to the industrial development in Western Europe and the US. The “core” of
the world economy required low-priced natural resources from the new settlements and the colonized
regions needed to obtain financial resources and labor to increase their ability to provide exports based
on the exploitation of natural resources. Conceptual frameworks such as the “staples theory” and the
“vent for surplus theory” consider the presence of excess resources that are insufficiently exploited in
economies usually small and closed and trade allows to foster exports and growth because natural
resources are used productively [9].
In addition, there is a wide range of literature about the US that underlines the favorable
influence that natural endowments had on welfare levels in the late 19th and early 20th centuries.
Mining promoted the establishment of prestigious educational centers and diffused knowledge to
other activities [10,11] in a sense close to the notion of biased technological transformation encouraged
by the availability of natural resources [11]. Recently, the studies have also deeply analyzed other
successful cases as Australia, Norway, Sweden, Chile, Botswana and Indonesia, where the abundance
and adequate management of natural resources has been able to promote economic growth.
2.1.2. Production Structure Approach: The Difficulties of a Primary Sector Specialization
We identify a couple of views. First, a perspective that considers the allocation of resources
among economic activities with different spillovers and underlines the effects of specialization in
economic growth. Economies whose economic performances are founded on natural resources and
where secondary and tertiary sectors represent a modest participation in the production structure will
expand moderately and evidencing serious limitations to promote the structural change. Within the
mainstream of economic growth research, this field is represented by the Developmentalism of the
1950s and the endogenous growth models of the 1980s [12]. Two alternative visions to this mainstream
literature offer other arguments to interpret the relationship between specialization in primary products
(generated from the exploitation of mineral, forest and land wealth) and low economic growth rates.
One of them is related to the Marxist, Dependency and Structuralist traditions, which consider
the unequal development view [13–15]. The other corresponds to the recent post-Keynesian and
post-Kaldorian theories, which discuss income elasticities and external constraints on growth [16],
as well as the formal characterization of technological dynamics in the Neo-Schumpeterian and
Evolutionary Schools [17].
Second, according to the so-called Dutch disease hypothesis [18–20], economies with abundant
natural resources are subject to successive fluctuations in their levels of economic activity because
commodity prices are volatile and periodically new exploitable natural resources are discovered. As a
result, the economy will become specialized in the exploitation of natural resources and, in consequence,
economic growth will be affected. Recent research, which refers to the endogeneity of resource
dependence, argues that volatility may be the quintessence of the resource curse6.
2.1.3. Crowding-Out Approach: Natural Resources Displace Other Types of Capital
In the formulation of the most extended models, abundance or heavy dependence of natural
resources has effects on some variable “x” which hinders economic expansion. Theorists and empirical
Sustainability 2018, 10, 715 4 of 26
researchers have identified the channels that relate both processes [21]. These mechanisms are
interpreted in terms of a crowding out effect where the abundance of natural capital displaces other
types of capital and hamper economic growth.
Large natural resource rents, combined with property rights poorly defined, imperfect markets
and weak legal structures, may lead producers to adopt uncontrollable rent-seeking actions.
These actions deflect resources away from activities that are socially more productive, causing a
false sense of security and inducing the government to disregard the promotion of high institutional
quality [22–25]. Natural capital may have different effects on economic performance depending
on the conditions of the endowments [26]. “Point resources” (e.g., mineral and energy resources,
activities with intensive use of capital) and “diffuse resources” (e.g., cropland and livestock) contribute
to different development potential. “Point resources” cause greater opportunities for rent-seeking
and corruption than “diffuse resources” and the adverse consequences are more negative [27,28],
especially with the low quality of the institutions and the different levels of resource appropriability [29].
Abundance of natural resources may affect accumulation of human capital because predominates a
high level of non-wage incomes and, usually, these economies tend to underestimate the value of
educating and training the society in the long run [30–32]. Finally, abundant natural resources provide
an ongoing future flow of rents and social welfare seems less bound to intertemporal transferences of
produced capital to the future [21], thereby inducing low saving rates [7].
2.1.4. Institutional Change and Factor Endowment Approach
This approach argues that the main explanation of economic development results from the interaction
of basic exogenous factors (ultimate causes as climate, topography, disease and environmental factors)
and institutional legacy [33].
Colonization strategy was influenced by those natural conditions. Places with worse conditions for
settlement and with high mortality rates among colonists (malaria, yellow fever), were characterized by
the formation of extractive states. But, on the other hand, if colonizers could safely settle in a region they
formed and promoted high quality institutions. The colonial legacy of the institutional matrix persists
in the long run and it results in a fundamental factor in determining economic development [34].
However, the determinant factor endowment was not just the abundance of land and natural resources
in relation to labor but also the type of land, the climatic conditions and native populations in terms
of size and density [35,36]. The extreme differences in terms of wealth and human capital inequality,
as well as political influence, across the New World societies have a fundamental importance in
the explanation. The causal relationship is between natural endowments, conditions of social and
economic inequality and the creation of an institutional matrix capable to generate the requirements
for sustaining the economic development in the long run.
2.1.5. Economic History Approaches
Recent efforts from economic history have offered new evidence, approaches and emphasis that
respond to Van Der Ploeg’s claim referring to the relevance of promoting collaboration between political
scientists, historians and economists to understand, from an economic point of view, the relationships
between natural resources and economic development. Willebald, Badia-Miró and Pinilla (2015) [8]
highlight three answers to the question: what do we learn from history? (i) abundance of natural
resources is non-neutral for economic development; (ii) abundance is an endogenous process;
(ii) institutional quality is the key factor for development of abundant natural resources to have
good economic outcomes.
First, there is evidence of a close relationship between abundance of natural resources and
development. Only 5% of the total world wealth was comprised of natural capital in the beginning
of the 21st century. However, divergence between levels of development was the dominant feature:
this ratio for high and low-income countries was, respectively, 2% and 30%.
Sustainability 2018, 10, 715 5 of 26
Second, abundance of natural resources does not refer to a static concept. “Abundance” is
a process that receives the influence of modifications in the relative prices and the structure of
factor endowments and to progress it requires investment, labor, technological change and proper
institutional arrangements. Therefore, this “abundance” is part of the evolution of the economic
system and, in these terms, we identify the endogeneity of natural capital as a typical historical feature.
In other words, “natural resource abundance was an endogenous, ‘socially constructed’ condition that
was not geologically pre-ordained” [10] (p. 203).
Finally, the relevance of the institutional quality is expressed in the ability of institutions to restrict
opportunities of rent-seeking, to regulate the political competition and participation and to deal with
transactional risk through suitable enforcement of property rights of natural resources.
2.2. The Empirical Evidence of the Resource Curse
This evidence referred to the curse of the natural resources is quite diverse and can be characterized
in three different groups [2]. The first of them regards Sachs and Warner’s cross-sectional study and
deal with different indicators to approximate resource abundance/dependence. The second set of
exercises concentrates on economic elements relevant for growth that can be influenced by natural
capital. The last group sheds doubt about the validity of the econometric exercises and the accuracy of
the empirical tests.
2.2.1. The Impact of Natural Resources on Economic Growth
Empiric on the curse hypothesis begun with a couple of case studies ([3,37]). However, the cross-
sectional analysis of [38] is considered the seminal empirical contribution for a long series of studies
“in the pursuit” of the curse.
Sachs and Warner (1995, 2001) considered a large number of economies for the period 1970–1989 and
verified that natural resource dependence and economic growth are negatively correlated. These articles
inspired subsequent research that has examined the direct and indirect relationships between
dependence of natural resources and economic growth. Multi-country comparisons predominate
and the majority of the studies [2] consider a dependence indicator to proxy the influence of natural
resources on economic growth, for example: primary exports over GDP [38–42]; rents from natural
resources over GDP [43–48]; natural capital on national wealth [21,49]; and mineral exports over total
exports [50–52]. However, other studies have considered a more adequate proxy to abundance of
natural resources as the total natural capital, mineral resource assets in dollars per capita [53,54] and
subsoil wealth [41,53,55] (see [2] for a review of the literature).
In addition, resource curse hypothesis also has been discussed in many single country studies,
with [3,37] being the two main antecedents. Several of these studies explain the failure of many African
economies [25,56], identify cases where the curse was avoided ([57] for the US, [58,59] for Botswana,
for Chile [60], for Norway [61], for Peru [62] and for China [63]) or cases where the evidence confirms
some type of curse (for Angola [64] and Venezuela, for Ghana [56], for Mexico [65] and Venezuela,
for Nigeria [66]).
2.2.2. The Impact of Resources on Factors Linked to Growth
Numerous works additionally obtain evidence of relationships between natural resource
dependence and factors which are tightly linked to economic success. These exercises include
human capital [49,52,67–69], savings rates [21,50,56,70], exports of manufactured products [71],
physical investment, schooling and openness [57], fiscal policy [72] and institutional quality [42,73].
This empirical evidence responds to the crowding-out approach presented previously. According to
this conceptualization, the negativistic effect of natural resource richness on economic growth is seen
as coming from their adverse effect on drivers of economic growth.
Sustainability 2018, 10, 715 6 of 26
2.2.3. Is This Apparent Paradox a Red Herring?
Since previous literature came to understand the specific mechanisms through which the curse
operated, a new trend has emerged. This new approach is to question the whole curse hypothesis as a
“red herring” [54] or a “statistical mirage” [2].
According to [54], a fairly extended way of measuring—the percentage of exports with respect
to GDP– is endogenous, which seriously challenges one of the more fruitful stream of the literature
(included the pioneer Sachs and Warner’s work). The ratio is dependent on a country’s economic
policies and the institutional frame that have effects on both GDP level and growth, which affects both
sides of the equation. An attempt is made to surpass this issue using instrumental variables.
An alternative approach considers the time samples used as the main claim of the critic. In this
sense, [74] assert that the arguments proposed by the abundance of natural resource curse studies are
in many instances due to a deficient interpretation of information. The major part of the studies that
finds a curse uses a time horizon that begins between 1965 and 1970. This period is not convenient
since commercial exploitation in the oil-exporting countries began before 1950, without considering an
important period of analysis. The sample period in question also is considered in [53,75–79].
The time sensitiveness of natural resource curse points to other causal factors involved. Manzano and
Rigobon (2001) assert that it is plausible that the resource curse model of [38] reflects the effects of the
international oil price shocks of the 1970s, instead of an inherent trend for resource-abundant countries to
experience low economic growth.
The better part of these critical academic works demonstrates that if natural resource richness is
used instead of indicators of natural resource dependence, the effect of natural resources is positive.
Also, taking into account the possibility of non-linear or non- monotonic relationships between the
exploitation of natural resources and economic growth, the changing character of the relation in the
long run also appears as an issue [2].
3. Hypothesis and Empirical Strategy
3.1. Historical Overview and Reasons to Study Uruguay
Uruguay is a typical new settlement economy in the sense defined by [80] and constitutes, together
with Argentina, Australia, Canada, Chile, New Zealand, South Africa and the US, the “temperate
economies” that [81] (p. 195) identifies as “the group of non-European countries which in [the beginning
of] twentieth century can be classified as developed.”
Uruguay prospered thanks to its flourishing agrarian activities. Historically, it was characterized
as a country with plenty of natural resources and small population, mostly descendants of European
immigrants [82]. Economic growth was initially supported by exports of agrarian products that met an
expanding international demand. In the final decades of the 19th century Uruguay had achieved levels
of income per capita that exceeded those corresponding to the several leading European economies [83].
These economies had the challenge of going from being a settler society, highly specialized
in primary commodities, to some form of post-settler configuration [84], with a more diversified
production structure.
In the case of Uruguay, literature identifies three historical phases of development [85,86].
During the first globalization, growth combined the progressive consolidation of the domestic
market [87] with an export expansion based on a few primary commodities (over 1870–1879, wool,
hides and preserved meat represented 60% of total exports; in the decade prior to WWI, the same
products represented almost 70%) [88]. In this period, the Uruguayan economy achieved high income
levels; over 1870–1879, the GDP per capita of Uruguay was 90% of the level of GDP per capita of “core”
countries—average of France, Germany and UK—and, in the decade preceding WWI, the same ratio
was almost 80% [89]. The primary activities (agriculture and mining) represented about one-third of
economic structure in 1870–1930 while the proportion of manufacturing averaged 12% of GDP [90].
Sustainability 2018, 10, 715 7 of 26
After some years of economic turmoil in the early 1930s, Uruguay adopted inward-looking
oriented policies focused on a strategy of industrialization by substitution of imports (ISI) to encourage
growth and economic development [91,92]. The industrial sector increased in terms of GDP, reaching
almost one-third of total output, in opposition to the falling share of primary activities. In the decade
following the WWII, the country experienced rapid growth based on the manufacturing industry in
a process that extended until the end of the 1950s [93], when a long period of stagnation and high
inflation started [94]. This situation was not overcome until 1970, when new measures of economic
policy were implemented based on a gradual openness of economy, increasing financial liberalization
and regional trade agreements [86]. The strategy of promoting export of non-traditional goods get
satisfactory economic results, with the expansion of several industrial branches (manufactures of
textiles, tanning and dressing of leather and footwear leather, electrical machinery and apparatus,
transport equipment and paper products) and high economic growth rates for the whole economy.
The liberalization process continued in the 1980s and 1990s and the manufacturing sector declined
drastically as a share of the economy. During these decades, the economy went through two deep crises,
one in the 1980s and another in the 2000s. From 2003–2004 to the present, Uruguay has experienced a
long expansion cycle that suggesting a beginning of an income convergence process with developed
countries [95], as contrasted with the trajectory evidence since the 1960s.
The expansion cycle is founded on high rates of fixed capital formation, with increased
involvement of direct foreign investment in the economy—focused on competitive agro-industrial
sectors—and the increase of exports supported by higher volumes and prices [96]. The Uruguayan
economy has continued to strongly specialize in commodities and services based on natural resources,
which comprise 70% of total exports [95]. The production structure has not experienced major
transformations [96] and the expansion cycle has been driven by agriculture. The major part of
agricultural activities has experienced significant changes in terms of productivity, technological
progress, logistics and transport activities and public policy directed to the sector. In this sense,
historical and “traditional” agricultural sectors have made way to modern activities with important
incorporation of technology and R + D activities.
Considering this long-run economic evolution and the previous theoretical and empirical
antecedents (presented in Section 2), Uruguay is an interesting case to study for three reasons.
3.1.1. Duality of the Structural Change
A strong consensus exists regarding the historical characterization of Uruguay as a natural
resource-abundant economy [97,98] that founded its economic development on the basis of exploitation
of this natural wealth. The influence of natural resources on economic development was diverse and,
to some extent, in opposing directions. Whereas industrial manufacturing increased significantly
as a share of the economic structure at the expense of the primary activities from the 1930s to the
1960s [99] and afterwards services gained a predominate position in total value added, other areas
of the economy did not experience similar transformations. In fact, nowadays, Uruguay is clearly
identified as a natural resource-abundant economy when we evaluate its export structure but when
we consider the structure of the economy as a whole, this characterization is not so evident. This last
assertion is confirmed by World Bank data, where Uruguay does not appear as an economy especially
rich in terms of natural resources.
Therefore, according to the historical evolution of Uruguay, we propose our first working
hypothesis: we expect a high relevance of natural wealth during the agro-exporter model but a
reduced share in the economy after the 1930s, when the structural change led to an increasing role
of manufacturing and services in linkages between economic sectors. This hypothesis would be
an (indirect) expression, in the long run, of the stylized fact of economic growth identified in [100];
i.e., those economies with the highest GDP per capita tend to have shares of natural capital, in their
total wealth, lower than the poor economies.
Sustainability 2018, 10, 715 8 of 26
In this sense, we expect that the abundance of natural resources caused economic growth during
the last third of the 19th century and the first decades of the 20th century; however, this influence faded
in the following decades. In other words, we assume that there would have been a positive relationship
between natural resources and economic growth in the first stages of development and a progressive
reversal of the relationship in the following decades. This is our second working hypothesis.
3.1.2. The Relevance of Studying the Non-Mineral Wealth
The literature has focused on subsoil resources and has underestimated the role of non-mineral
natural wealth. Different types of natural resources may affect differently economic performance [23].
“Point resources” (e.g., mineral and energy resources) can create greater chances for rent-seeking and
corruption than the “diffuse resources” (e.g., cropland and pastureland) representing more serious
consequences for economic growth. Economic performance is usually more affected when natural
resources—and the rents derived from their exploitation [79]—are more easily captured and controlled
by a narrow elite. However, these arguments about economic growth depending on the type of
resources does not necessarily mean that one should leave non-mineral wealth out of the analysis.
In addition, almost a 40% of the total worldwide wealth is comprised of subsoil assets. In other words,
the excessive academic concentration in natural resources such as oil, gas and minerals has probably
impeded advance in knowledge of several dimensions of the process. We contribute new evidence
on this matter by considering the Uruguayan economy, where resource wealth is almost entirely
non-mineral. Implicitly, with the hypothesis presented in the previous section, we are assuming
that the abundance of natural resources is an exogenous variable for Uruguay. Considering that the
main components of the natural capital are pastures and croplands and that the exploitation does not
require—relative with “point resources”—huge capital outlays or special conditions, the hypothesis
seems, at least operatively, plausible.
3.1.3. Natural Capital as a Better Proxy for the Abundance of Natural Resources
Following [38], primary exports over GDP have been the preferred indicator in the natural-
resource-and-growth literature. However, this measure seems an unsatisfactory indicator of
natural resource abundance for two main reasons. On the one hand, the exports-to-GDP ratio as
a representation of the abundance of natural resources is questioned and particularly so for cases
comparable to Uruguay with similar processes of dual structural change. On the other hand, empirical
exercises present problems of endogeneity because they suffer from third factors as fiscal and social
policies, taxation and institutions that have effects in both sides of the equation. According to [53,54],
the measure of natural capital proposed by [100,101] overcomes these criticisms and offers more
acceptable estimations. The indicators of natural resources proposed by the World Bank evaluate
different components of natural wealth considering the net present value of expected rents and we will
apply this methodology for long-run estimates and the historical analysis (see below the discussion).
However, the advantage of this measure is not only from an empirical point of view. World Bank
data is the best proxy to test the natural resource curse hypothesis because the proposal to use natural
capital as a measure for resource abundance is, conceptually, superior. This measure of natural resource
wealth is consistent with a broader concept of an economic wealth indicator—that considers natural,
produced, human and institutional assets—and, consequently, with a more rigorous idea of economic
growth. In this sense, it is possible to distinguish, in a more convincing conceptual and empirical
manner, between resource dependence and resource abundance [102].
3.2. Empirical Strategy
We propose a first attempt to count the natural capital for a periphery economy on an annual basis,
applying the World Bank’s methodology, correcting the initial estimates, adapting the assumptions to
the historical analysis and offering a very long run perspective to cover 144 years of history (1870–2014).
Our efforts to cover so long a period are based on, at least, two kinds of considerations. One, we expect
Sustainability 2018, 10, 715 9 of 26
to contribute to the debate about the mechanisms that link natural resources and economic growth
and, conceptually, this is a matter inevitably concerning the long run. Even, the consideration of long
run periods will allow to identify changes in the relation between these measures of natural resources
and economic growth concepts. Two, this will allow to overcome the previously mentioned limitation
of many resource curse studies referred to the time samples used in the empirical tests.
Therefore, initially, we present, adapt and revise the World Bank’s methodology for applying these
concepts for historical estimates. Then, we present the long-run evolution of the natural capital and
identify how consistent, in historical terms, this trajectory is according to the national historiography.
Finally, we test different causality exercises to find evidence about the direction of the influence
between variables.
4. Materials and Methods
4.1. Natural Capital Estimation
The approach used in our natural capital estimation from 1870 to 2014 is based on the World
Bank methodology presented in [100,101] (in the following, we present the first advances of the Master
Thesis in Economic History of Silvana Sandonato titled: “Capital Natural y crecimiento económico en
Uruguay en el largo plazo (1870–2014),” PHES-FCS-Universidad de la República, Uruguay; which will
be defended in 2018). This methodology rests on the well-established economic principle that asset
values are measured as the present discounted value of economic profits over the life of the resource.
This value, for a particular economy and type of resource, is given by the following expression:
Vt = ∑
t+T−1
i=1
πi·qi
(1 + r)(i−t)
where πi qi is the economic rent at time i, (πi represents unit rent and qi denotes production), r is the
social discount rate and T is the lifetime of the resource.
Natural capital, according to the World Bank dataset, is built from estimates of cropland,
pastureland, forests, metals and other minerals, coal, oil, natural gas and the inclusion of the “protected
areas”. Cropland is represented as a flow of land rents over a 25-years horizon (one generation).
Specific rental rates for cereals fluctuate between 30% and 50% of the gross value of production (wheat,
rice, maize) and they are used to calculate the valorization of cereal lands. Then, the rest of lands are
valued at 80% of this rate to allow the possibility that other types of crops are likely to yield inferior
returns. Pastureland is valued similarly to cropland but the rental rate is 45% of the gross value of
livestock (considering wool, meat and milk). For timber, there are two options for valuation depending
of the sustainability of the exploitation. First, when exploitation is sustainable—i.e., the harvest rate is
lower than the annual growth rate of the forest—the resource is measured according to the present
value of a constant resource rent in a period of 25 years. Second, in the other case, the present value
of a constant resource rent over less than 25 years (depending on the extraction rate) is considered.
Non-timber rents are evaluated by assuming that a tenth of the area destined to forest is exploited for
production different than timber. Protected areas are measured using an opportunity cost approach,
which considers that the value corresponding to pastureland is assigned as a proxy for this type of
resource. Minerals, metals and fossil fuels are considered as the present value of a constant flow of
resource rents for the life of proven reserves (bauxite, copper, iron ore, lead, nickel, phosphate rock, tin,
zinc oil, coal, gas). In cases where there are no data on reserves, a time to exhaustion of 25 years is
assumed. This estimation includes information for over 120 countries for the years 1995, 2000 and 2005.
The immediate antecedent of the present article is [103], which applies the methodology presented
in [100], although it additionally provides a review of assumptions, components and statistical sources
which take into account the specificity of Uruguay within the region and among countries with a
similar development process. This estimate—corresponding to 1990, 1995, 2000, 2005 and 2010—uses
a social discount rate calculated for domestic social investment projects and includes data series for
Sustainability 2018, 10, 715 10 of 26
country-specific mineral resources and uses specific growth rates for Uruguay. This analysis offers
evidence regarding the dynamics of the proximate determinants of natural capital and allows us,
in turn, to recognize the drivers and components of the current path of national growth. In [104],
an annual estimation is offered for 1990–2011.
These estimations consider several fixed parameters: social discount, rental and expected
growth rates. The fixed parameters applied for the estimations are based on contemporary values,
without considering the long-term historical specificity. The historical context of our work leads us to
make changes to this methodology, which means an important contribution of this article.
There are four main modifications with respect to the previous methodology. While the Word Bank:
1. considers only expected production, we consider, whenever it is possible, output actually recorded,
except from 2015 to 2038, where we had to project productions in order to obtain the present
discounted value for 1991 to 2014. When historical data were not available, we use different
estimation techniques to complete the series.
2. considers fixed production rates of return, we consider rents of natural resources actually received,
which are therefore variable over time.
3. considers a fixed social discount rate, we consider an annual variable rate which we estimate.
4. assumes a fixed growth rate of future incomes, we take advantage of the information constructed
and forecast future rents according to past trajectory (144 years) of the rents.
The first and second changes will be explained for each type of natural resource. Estimations of
social discount rates are briefly explained below and described in detail in Appendix A.
The social discount rate or consumption rate of interest (CRI) is defined as the rate at which the
marginal welfare of consumption falls over time, which may be decomposed into the formula:
CRI = δ − L + (µg)
δ = pure time preference
L = changing life chance (negative sign)
µ = marginal utility of consumption
g = expected consumption growth rate
For the estimation of CRI we follow the methodology presented in [105,106] using own
elaboration from data presented in [107]. (See Appendix A).
We assume that the time of exhaustion of resources is 25 years, which roughly coincides with one
generation (the same assumption was made by the World Bank). To consider a longer exhaustion time
would require predictions of total rents for longer time horizons, which have less weight since they are
more heavily discounted. This would mean including predictions further into the future which have
marginal contributions in the final estimation. Finally, the level of uncertainty rises as the agents look
toward a more remote future.
We use 2005 prices to value income from every resource, annual rent rates for each resource,
an annual social discount rate and real values for the total rent up to 2014.
4.2. Natural Capital by Component
4.2.1. Cropland
Cropland wealth is computed as the present discounted value of rents derived from lands used
for cultivation. Annual land return is the sum of returns from: apples, barley, beans, green beans,
broad beans, cabbages and other brassicas, canary seed, carrots and turnips, chilies and peppers,
cottonseed, garlic, grapefruit (included pomelos), grapes, groundnuts with shell, lemons and limes,
linseed, maize, green maize, melons (included cantaloupes), oats, olives, onions, oranges, peaches and
nectarines, pears, peas, plums and sloes, potatoes, pumpkins, squash and gourds, quinces, rapeseed,
Sustainability 2018, 10, 715 11 of 26
rice, seed cotton, sorghum, soybeans, sugar beet, sugar cane, sweet potatoes, tangerines, mandarins,
clementine, satsumas, tobacco, tomatoes, fresh vegetables, watermelons and wheat, in all cases we
assume that the products are sold at prices obtained in international markets.
We gather information about the physical volume of production of each commodity from 1870 to
2014, valued at 2005 prices. To obtain annual income from crops, we estimate an annual rental rate
derived from [107] and additional estimates provided by the authors. In this study, the estimates of
land rents do not distinguish between cropland and pastureland and we use the same rental rates for
both activities, with this rate change over time according to the ratio of land rents and agricultural value
added (in current prices). The rate corresponding to 2005 is 32% and our calculations from national
accounts data—assuming that one-third of the production corresponds to agricultural services—is 30%;
then we consider the rates reported in that work to be roughly correct. Finally, we project the cropland
rents from 2015 to 2038 using standard forecasting methods to represent the sustainability of present
cultivation techniques.
4.2.2. Pastureland
Pastureland wealth is computed as the present discounted value of rents derived from lands
used for livestock. Annual returns from pastureland is the summation of returns from all pastureland
outputs considered: meat, milk, wool, leathers, eggs and honey, valued at international prices.
We gather information about the physical volume of production of each commodity from 1870 to
2014, valued at 2005 prices. In order to obtain annual pastureland incomes, we estimate an annual rental
rate derived from [107] and other estimates provided by the authors (see the previous explanation).
Finally, we project the pastureland rents from 2015 to 2038 to reflect the sustainability of present
grazing practices.
4.2.3. Mineral Resources
Mineral wealth is computed as the present discounted value of rents derived from mineral
exploitation. We assign dollar values to the production of every metal and mineral exploited in
Uruguay: sand, pebble, dolomite, limestone, quartz, feldspar, agate, amethyst, granite, marble, stone
slab, clay, talc stone, boulder, iron, loam, sandstone, filita, bentonite, gold minerals, basalt, cornelian,
greenstone, granodiorite, fluorite, corundum, coarse, rough stone and conchilla. Mineral prices
correspond to the implicit price calculated as the ratio between the production value (BCU website)
and our estimates of the physical volume of total production in 2005.
Data for mineral resources before 1957 is not available. In order to obtain the mineral series from
1870 to 1956, we assume that the evolution of annual mineral income is the same as the mineral value
added [90 and additional estimates provided by the authors]. The 2004 and 2005 World Bank data
on mineral rental rates is a point of reference (World Bank database online); we, however, update
upon this by using the ratio of gross operating surplus and value added for use in our model. We use
data derived from [107]; we also use other estimates provided by the authors. Finally, we project the
mineral rents from 2015 to 2038 assuming that reserves will be exhausted in 25 years.
4.2.4. Timber Resources
Timber wealth is computed as the present discounted value of rents from roundwood production.
We consider three types of production: coniferous industrial logs, non-coniferous industrial logs and
firewood. Also, we consider international trade prices of standing timber for coniferous industrial
logs and non-coniferous industrial logs. As there is not international trade in firewood, we estimate
the price of firewood price as an average of international prices of coniferous industrial logs and
non-coniferous industrial logs.
The time horizon selected to capitalize the annual revenues incorporates the concept of sustainable
exploitation. If logging is less than the net annual growth, then the time horizon will be 25 years and
the exploitation will be considered sustainable. If annual logging is greater than net increases, it means
Sustainability 2018, 10, 715 12 of 26
that the holding is not harvested sustainably and therefore the time horizon for the capitalization of
annual revenues will be equal to the minimum of 25 and:
Stock o f f orest area (ha) × Volume o f wood per hectare × Forest area available
Wood production − Annual increment
The forest area available for timber offer is estimated as forests within 50 km of infrastructure
because not all standing timber is accessible or economically viable.
We use production values from 1870 to 2014 (2005 fixed prices). There is no data for firewood
until 1961, so we estimate this production from 1870 to 1960 using the series of firewood consumption
from [108].
Also, we apply an annual rental rate in order to obtain annual forest income. For this, we use
data derived from [107]; we also use other estimates provided by the authors. The strategy in this case
was to consider the rate proposed in [100] (58%) and for this rate to evolve over time according to the
evolution of the ratio between the gross operating surplus and the value added of the activity. This ratio
is available for 1997–2005, is considered fixed from 2005 to the present and is historically proxied by
the evolution of the inverse ratio of total wages and value added (i.e., considering incomes derived
from work in contrast to the incomes derived from capital, in relation to value-added). This ratio is
available for 1997–2005, 1958 and 1908 and the rest of the intermediate years are lineal interpolation;
for the years before 1908, the level of that year is assumed.
4.2.5. Non-Timber Forest Resources
According to specialists from the General Forest Directorate (Dirección General Forestal), only one-
twentieth of forest areas in Uruguay are reachable for recreation. Therefore, the per-hectare value for
recreation is multiplied by one-twentieth of the stock of forest area to get the benefits of recreation.
We use the World Bank reference price, based on [109], of USD 112 per hectare of forest area for that
year and adjusted it to 2005 using the USD purchasing power index.
Non-timber wealth is then computed as the present discounted value of rents from non-timber
forest benefits. As before, we project the rents from 2015 to 2038 using standard forecasting techniques.
4.2.6. Protected Areas
The protected areas (Sistema Nacional de Áreas Protegidas) are valued at the lowest returns per
hectare to pasturelands and croplands; this decision implies the interpretation of the price as a
quasi-opportunity cost. These returns are then capitalized over a 25 year time horizon. Restricting the
value of the protected areas to the opportunity cost of preservation means reflect the minimum but not
the entire value of this component.
4.3. Comparison with Previous Estimates
As far as we were concerned, the only work that presents a detailed estimate of long-term natural
capital is [106]. This historical work puts forward an estimation of natural capital for Sweden for
the 19th and 20th centuries. The long-term analysis of this article offers a detailed description of the
estimates proposed and its results are an extremely useful point of reference for our research. It shares
the methodology of the World Bank but makes some modifications to estimate the value of land
dedicated to agriculture. They use the current market price of the land instead of updating the future
income of products derived from this kind of asset. The article reports first historical estimates of the
CRI and also an estimation of total wealth according to historical surveys and information derived
from insurance data. We compare some of our indicators with that data to contrast results.
Sustainability 2018, 10, 715 13 of 26
4.4. Contrast with Linear Non-Causality
As we have already stated, one of the most widespread discussions in the literature has been the
relationship between the abundance of natural resources that an economy possesses and the long-term
economic evolution that this economy performs. Therefore, our methodological option is to investigate
the causal relationships between the two variables.
The fact of having long-term series of natural capital and other variables such as GDP makes
it possible to use a concept of specific causality—a la Granger—but taking advantage of techniques
which extend upon the standard Granger approach to overcome some of its limitations.
The Toda-Yamamoto (1995) and Dolado-Lütkepohl (1996) test (from now on TYDL) is an extension
of the classic linear Granger non-causality test, within the framework of a model of auto-regressive
vectors (VAR). Both tests are based on the following operative definition of Granger non-causality.
Let two stationary stochastic processes {Xt}, {Yt}:
Yt does not Ganger cause Xt if E(Xt + 1|Xt) = E(Xt + 1|Xt, Yt) .
The Granger test has the limitation that it can only be applied for stationary series, so prior to
applying it, a series of tests must be done to identify whether unit roots and co-integration are present,
in order to afterward make the appropriate transformations to the series. This practice of pre-testing
may result in significant over-rejection rates of a null hypothesis of no true causality, finding causal
relationships when in fact there are not [110].
The TYDL methodology avoids these preliminary tests and is most useful outside of whatever
integration or co-integration may be present in the system, thus obtaining more robust results. It is a
parametric test that can be contrasted from linear non-causality.
The TYDL test states that a VAR model (p + dmax) must be estimated, where p is the true number
of lags of the model (determined by some criterion for selection of lags) and dmax is the maximum
order of integration that is suspected to occur in the process.
Next, the last dmax lags are ignored—since, as zeros, they are redundant—and tested the
hypothesis of Granger non-causality on the first p lags of the model, using the standard asymptotic
theory. The Wald statistic maintains its χ2 distribution below a certain H0. If the null hypothesis is
rejected, H0 = β1 = β2 =, . . . , β p = 0 we can say that the variable Y Granger causes X.
5. Results
5.1. Some Stylized Facts
In this section, we present the main stylized facts based on the estimation of the natural capital
series, over 1870–2014, for Uruguay. Our first important result is the estimation of the annual series of
natural capital in the long term itself, which makes possible the analysis of its evolution. Over this
time period, natural capital grew at an average annual rate of almost 2% (Table 1), which is less
than the average of the entire economy (3%), resulting in a sustained decline in its share of
GDP
(Figure 1b) (Table 1 includes information about three historical sub-periods; see the explanation of this
periodization below).
Natural capital is seven times the value of annual GDP at the beginning of the period and even
by the WWI, the ratio was around 5. The only antecedent that is comparable—although information
is not available on an annual basis—is the case of Sweden, for which the average over the same
period was around 3. Although a broader comparison would be required, this contrast legitimizes to
conjecture that Uruguay’s historical conceptualization as an economy abundant in natural resources
would be correct. In a previous comparison, the two economies were characterized as “resource-rich
countries” [111] and Uruguay being comparatively more so when the richness is measured in terms of
land. Our evidence is consistent with this last argument.
The long-term evolution of the ratio NK/GDP is an expected result and confirms our first
hypothesis. The Uruguayan economy shows growth of GDP per capita throughout the period and
this would have been accompanied by a decline in the economic relevance of natural capital in the
Sustainability 2018, 10, 715 14 of 26
process of generating wealth (which GDP measures). The verification of the hypothesis is indirect
because the approach of the World Bank is based on the composition of the total wealth; in any case,
the ratio could be a good proxy for the process. The reduction in the ratio was not homogeneous
and the evolution involves a trajectory that is consistent with the long-term trajectory of the economy
presented in Section 3.
Table 1. Natural capital and GDP.
Annual Growth Rates
NK/GDP
Natural Capital (NK) GDP
1870–1909 2.0% 3.6% 527.4%
1910–1959 −0.7% 2.8% 178.6%
1960–2014 4.3% 2.6% 103.5%
1870–2014 1.9% 3.0% 133.2%
Sustainability 2017, 9, x FOR PEER REVIEW 14 of 26
(a) (b)
Figure 1. Natural capital and GDP (2005 million dollars) and NK/GDP ratio. (a) Natural capital and
GDP; (b) Natural capital/GDP.
During the First Globalization, natural capital increased continuously and maintained high
ratios relative to GDP that averaged 5.5 between 1870 and 1913. It is, in fact, a period in which
Uruguay managed to enter in international markets with the export of agricultural commodities and
during which expansion of the domestic market occurred at a rate that, in Latin America, only
Argentina or Chile could approximate [87]. GDP growth is the highest during this stage (3.6%),
coinciding with the highest levels of NK/GDP ratio (527%), which leads us to reject the resource curse
hypothesis in the case of Uruguay. It is after the WWI when the in dicator begins a sustained
downward trajectory, a period that different scholars recognize as the beginning of the transition to
a new pattern of development [112].
Since the 1930s, the economy experienced an ISI process [92,93], which meant clear signs of
structural change and in which activities related to primary production showed clear manifestations
of stagnation [113]. It is in this context that total natural capital reaches levels lower than the annual
GDP (year 1951), coinciding with the booming period of industrialization. It is only at the end of the
1960s that natural capital begins a progressive recovery (Figure 1a), although, in terms of GDP, the
increasing can be seen since the late 1970s, probably due to the commitment to more liberal economic
policy and the promotion of non-traditional exports. Nevertheless, it would not be until the 21st
century, with a new international boom based on commodities, that natural capital would once again
become relevant in the economy, the ratio approaching 2 by 2014.
From a long-term perspective, the loss of economic relevance of natural capital in the economic
structure also meant significant transformations within the natural wealth itself (Figure 2). Three
features can be underlined.
First, the increasing diversification of natural capital is notorious. Indeed, whereas in the first
decade of the period of analysis land wealth represented 94% of total natural capital −85%
corresponding to pastureland and 9% to croplands−, during the first decade of the 21st century land
wealth was still relevant (85%) but with croplands being the predominant component (46% vs. 39%)
and, in addition, forest wealth comprising a significant share (13%). This broader diversification
would have had various linkages with the structural transformation that the economy experienced
in various dimensions, both in considering the production structure as a whole [89] and agricultural
activity in particular [95] and also the energy matrix itself.
Second, one of the outstanding features of the agricultural sector in Latin America in the last
decades of the 20th century was the evidence of the absolute and in particular relative expansion of
the sector that has placed the activity on a renewed path of growth [114]. Uruguay has not been an
exception to this process, which has involved more intensive production development, with higher
requirements for capital, inputs and training [115,116]; industrial crops (rice, soybeans) and dairy are
the two most evident expressions of this transformation. The strong increase in cropland as a share
of natural capital from the 1990s onwards represents this transformation.
0.00
5,000.00
10,000.00
15,000.00
20,000.00
25,000.00
30,000.00
35,000.00
40,000.00
45,000.00
50,000.00
18
70
18
76
18
82
18
88
18
94
19
00
19
06
19
12
19
18
19
24
19
30
19
36
19
42
19
48
19
54
19
60
19
66
19
72
19
78
19
84
19
90
19
96
20
02
20
08
20
14
GDP
Natural capital
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
18
70
18
75
18
80
18
85
18
90
18
95
19
00
19
05
19
10
19
15
19
20
19
25
19
30
19
35
19
40
19
45
19
50
19
55
19
60
19
65
19
70
19
75
19
80
19
85
19
90
19
95
20
00
20
05
20
10
Figure 1. Natural capital and GDP (2005 million dollars) and NK/GDP ratio. (a) Natural capital and
GDP; (b) Natural capital/GDP.
During the First Globalization, natural capital increased continuously and maintained high ratios
relative to GDP that averaged 5.5 between 1870 and 1913. It is, in fact, a period in which Uruguay
managed to enter in international markets with the export of agricultural commodities and during
which expansion of the domestic market occurred at a rate that, in Latin America, only Argentina or
Chile could approximate [87]. GDP growth is the highest during this stage (3.6%), coinciding with
the highest levels of NK/GDP ratio (527%), which leads us to reject the resource curse hypothesis in
the case of Uruguay. It is after the WWI when the indicator begins a sustained downward trajectory,
a period that different scholars recognize as the beginning of the transition to a new pattern of
development [112].
Since the 1930s, the economy experienced an ISI process [92,93], which meant clear signs of
structural change and in which activities related to primary production showed clear manifestations of
stagnation [113]. It is in this context that total natural capital reaches levels lower than the annual GDP
(year 1951), coinciding with the booming period of industrialization. It is only at the end of the 1960s
that natural capital begins a progressive recovery (Figure 1a), although, in terms of GDP, the increasing
can be seen since the late 1970s, probably due to the commitment to more liberal economic policy and
the promotion of non-traditional exports. Nevertheless, it would not be until the 21st century, with a
new international boom based on commodities, that natural capital would once again become relevant
in the economy, the ratio approaching 2 by 2014.
Sustainability 2018, 10, 715 15 of 26
From a long-term perspective, the loss of economic relevance of natural capital in the economic
structure also meant significant transformations within the natural wealth itself (Figure 2). Three features
can be underlined.
First, the increasing diversification of natural capital is notorious. Indeed, whereas in the first decade
of the period of analysis land wealth represented 94% of total natural capital −85% corresponding to
pastureland and 9% to croplands−, during the first decade of the 21st century land wealth was still
relevant (85%) but with croplands being the predominant component (46% vs. 39%) and, in addition,
forest wealth comprising a significant share (13%). This broader diversification would have had various
linkages with the structural transformation that the economy experienced in various dimensions, both in
considering the production structure as a whole [89] and agricultural activity in particular [95] and also
the energy matrix itself.
Second, one of the outstanding features of the agricultural sector in Latin America in the last
decades of the 20th century was the evidence of the absolute and in particular relative expansion of
the sector that has placed the activity on a renewed path of growth [114]. Uruguay has not been an
exception to this process, which has involved more intensive production development, with higher
requirements for capital, inputs and training [115,116]; industrial crops (rice, soybeans) and dairy are
the two most evident expressions of this transformation. The strong increase in cropland as a share of
natural capital from the 1990s onwards represents this transformation.Sustainability 2017, 9, x FOR PEER REVIEW 15 of 26
Figure 2. Natural capital shares by component (in percentage).
Third, since the second half of the 20th century, forest wealth is an increasingly important
component of Uruguay’s natural wealth. Initially, the increase would have had many points of
contact with the industrialization process itself since wood was used as a key source of fuel in many
manufacturing companies [117]. The changes in the energy matrix that have taken place since the
1980s, decidedly toward hydroelectric generation, presumably underlie the decline observed in the
share of this component. However, it is not expectable to return to the levels prior to the expansion
due to a combination of, at least, two types of factors. On the one hand, there are legal norms that
have offered forestry preferential treatment; on the other hand, the demand for wood for pulp is high
in Uruguay and this has sustained production beyond that used for energy.
5.2. Causality Exercises
In the previous section, the stylized facts of the long-term evolution of natural capital were
presented and it was found to display high consistency with Uruguayan economic history. However,
these assessments do not respond to whether the relationship between natural capital and economic
performance is causal. In this section, some standard statistical causality exercises are presented as a
first approximation on this matter.
To start with, causality exercises are proposed between the growth rates of natural capital and
GDP for the entire period of analysis, with no evidence of Granger causality in either direction (Table
2) (a p-value below the 5% significance level indicates rejection of the null hypothesis of non-causality
between natural capital and GDP and vice versa). However, judging by the evolution of the variables,
it is possible to expect the relationship to change over time and so the period was partitioned. The
statistical exercise is sensitive to the length and the extremes of the period, on which many tests were
carried, out and we have decided to apply a periodization representative of development patterns
discussed above: the agro-export period (1870–1909); the end of the previous pattern and ISI (1910–
1959); the end of the previous pattern and liberalization and promotion of non-traditional exports
(1960–2014). In addition to the identification of historical periods, we try to divide our 144 years into
50-year windows to compare periods of similar duration.
Commented [SSG9]: I change the colours. Figure 2. Natural capital shares by component (in percentage).
Third, since the second half of the 20th century, forest wealth is an increasingly important
component of Uruguay’s natural wealth. Initially, the increase would have had many points of
contact with the industrialization process itself since wood was used as a key source of fuel in many
manufacturing companies [117]. The changes in the energy matrix that have taken place since the
1980s, decidedly toward hydroelectric generation, presumably underlie the decline observed in the
share of this component. However, it is not expectable to return to the levels prior to the expansion
due to a combination of, at least, two types of factors. On the one hand, there are legal norms that have
offered forestry preferential treatment; on the other hand, the demand for wood for pulp is high in
Uruguay and this has sustained production beyond that used for energy.
5.2. Causality Exercises
In the previous section, the stylized facts of the long-term evolution of natural capital
were presented and it was found to display high consistency with Uruguayan economic history.
However, these assessments do not respond to whether the relationship between natural capital and
Sustainability 2018, 10, 715 16 of 26
economic performance is causal. In this section, some standard statistical causality exercises are
presented as a first approximation on this matter.
To start with, causality exercises are proposed between the growth rates of natural capital and
GDP for the entire period of analysis, with no evidence of Granger causality in either direction (Table 2)
(a p-value below the 5% significance level indicates rejection of the null hypothesis of non-causality
between natural capital and GDP and vice versa). However, judging by the evolution of the variables,
it is possible to expect the relationship to change over time and so the period was partitioned.
The statistical exercise is sensitive to the length and the extremes of the period, on which many
tests were carried, out and we have decided to apply a periodization representative of development
patterns discussed above: the agro-export period (1870–1909); the end of the previous pattern and
ISI (1910–1959); the end of the previous pattern and liberalization and promotion of non-traditional
exports (1960–2014). In addition to the identification of historical periods, we try to divide our 144 years
into 50-year windows to compare periods of similar duration.
Table 2.
of the non-causality test of TYDL between NK and GDP.
Period
H0: lnNK Does Not Cause lnGDP H0: lnGDP Does Not Cause lnNK
χ2 Statistic p-Value χ2 Statistic p-Value
1870–1909 0.39 0.531 2.56 0.109
1910–1959 0.26 0.607 0.07 0.790
1960–2014 2.13 0.346 7.95 0.047
1870–2014 1.44 0.837 4.45 0.348
As is shown in Table 2, none of the tests allow to reject the null hypothesis of non-causality in
either direction with the only exception of the last sub-period (1960–2014), for causality from GDP to
natural capital.
In consideration of the previous results and the evidence found in other works referring to the fact
that it is important to identify natural capital action channels rather than the direct impact of this on
GDP, additional exercises are proposed. Using the classification of the explanatory factors of economic
growth in terms of proximate, intermediate and ultimate causes [118–121], we choose to work only
with the first. The ultimate and deepest causes of growth and economic development—geographic
conditions, institutions, power and long-run development in science and technology—deserve special
treatment that exceeds the objectives of this article.
To represent the proximate and intermediate factors of economic growth, we propose to work
with two typical factors of supply (physical or produced capital and human capital; the data come
from [122,123]) and demand (exports and terms of trade; the data come from [124]) to cover “both
sides” of the market. The results are presented in Tables 3 and 4 considering, respectively, the causality
from each factor to natural capital and vice versa (exercises for the entire period are not reported; in all
cases the hypothesis of non-causality is not rejected).
Table 3. Results of the TYDL statistic non-causality test from produced capital, human capital, exports
and terms of trade to natural capital.
1870–1909 1910–1959 1960–2014
Statistic p-Value Statistic p-Value Statistic p-Value
Produced capital 11.13 0.004 0.36 0.947 3.51 0.173
Human capital 0.40 0.525 9.86 0.043 2.67 0.102
Exports 0.63 0.429 7.76 0.021 0.76 0.385
Terms of trade 8.65 0.033 0.00 0.985 3.09 0.079
Sustainability 2018, 10, 715 17 of 26
Table 4. Results of the TYDL non-causality test from natural capital to produced capital, human capital,
exports and terms of trade.
1870–1909 1910–1959 1960–2014
Statistic p-Value Statistic p-Value Statistic p-Value
Produced capital 2.6 0.273 3.35 0.340 0.62 0.733
Human capital 0.44 0.506 8.41 0.078 0.27 0.600
Exports 0.96 0.328 3.04 0.219 1.44 0.229
Terms of trade 0.9 0.342 0.94 0.331 0.83 0.364
As shown in Table 3, in the 1870–1909 period, both produced capital and terms of trade cause
natural capital. Both results were expected. On the one hand, the increasing trajectory of the investment
during the last two decades of the 19th century “seems . . . related to the creation of conditions to
support the ‘take-off’ of the agro-exporter economy” [122] such as in the form of railways, harbors
and roads. In these terms, the natural capital would have been endogenous to the physical capital at
least in the initial stages of the expansion. On the other hand, this period was characterized by strong
growth in the prices of exported goods [125,126], in particular of primary sector origin, which would
have encouraged the expansion of the natural capital.
The second period was a phase of decreasing of natural capital in real terms (Table 1) and so human
capital and exports both cause the natural asset. In the case of human capital, the result is consistent
with an important expansion of the social public expenditure in education [127,128] and in particular
the extension of the construction of school along the national territory [128]. Our interpretation of
this process corroborates the expectation of the World Bank about the evolution of the composition of
total wealth in the long run. According to [100]—in the tradition of “weak sustainability”—one
of the stylized facts of economic growth is the substitution of natural capital by other types of
assets, in particular human and intangible capital. Our evidence is consistent with this argument.
Since neither physical capital nor human capital were displaced by natural capital, our evidence
contradicts arguments in the tradition of the crowding-out approach presented above.
In the case of exports, the regulations, norms and, in general, the economic policy of the period,
meant an inward looking development [92] that severely affected the export capacity of the economy [129].
Considering that most of the exports originated from agricultural commodities, the adverse consequences
on the formation of natural capital would have been inevitable. Natural capital did not block the structural
change but several factors that promoted industrialization affected exports, thereby reducing the relative
abundance of natural resources.
Finally, in the third period, when the economy advanced through a progressive process of outward
economic orientation, liberalization and openness, we find the terms of trade to once again be decisive
in the growth of natural capital (if we allow a 10% significance level).
We repeat the exercises but considering the inverse causality. The results, presented in Table 4,
show that natural capital does not cause the factors considered, with the only exception of human
capital, in the second period (10% significance).
6. Discussion
One of the main arguments of the economic history approach presented in Section 2 is that
“rather than being a general pattern, the curse seems subject to the influence of supply and demand
conditions, technological progress and institutional structure with strong historical specificities” [130]
(p. 248). Our long-term analysis provides an adequate framework to analyze how the relationships
change over time and how difficult it can become to sustain the “curse”—or the “blessing”—as an
immutable hypothesis.
Our first result is that natural capital has tended to decline in importance to the economy
(measured as the ratio of natural capital and GDP). This is a historically consistent result. The economic
Sustainability 2018, 10, 715 18 of 26
transformation from a model based on agro-exports to another where the driver of the economy
was industrialization—until the end of the 1950s—meant declining influence of natural capital on
the internal generation of incomes. The conditions of the model began to change from the 1970s
onwards and, especially with the commodity-boom of the 21st century, natural capital has become
more economically relevant again in the case of Uruguay. Additionally, our result constitutes indirect
evidence of the assertion of [100] about the share of natural capital being a smaller share of total
wealth in more developed economies. This evidence is indirect for two reasons. On the one hand,
the approach of the World Bank refers to cross-sectional evidence, i.e., comparisons between economies
with different levels of development at the same time. However, in our case we consider one country
that increased its economic development and wellbeing in the long run. The approach of the World
Bank refers to shares of wealth whereas our estimates refer to shares of GDP.
Our second result is to identify an important long-run diversification of the natural capital since
the second half of the 20th century. This evolution expresses the dual impact of a remarkable process of
change in the agricultural production, based on the intensive use of factors with intensive production
(industrial crops such as rice and soybeans, as well as dairy industry) comprising a larger share of the
subsector and the rising presence of forestry in the use of land.
Our third result refers to the causality relationships; for these exercises, we consider the entire
period and three sub periods: 1870–1909, 1910–1959 and 1960–2014. We do not find causal relations
between the increase in natural capital and economic growth in the long run nor in the sub-periods,
with only one exception (from economic growth to growth in natural capital for 1960–2014). In other
words, we do not find evidence to confirm the presence of a curse or blessing of natural resource
abundance. The evolution of the natural capital does not influence economic growth in the long
run. We thus consider several channels (see Section 2, crowding-out approach) that, potentially,
may connect the two processes. Our stronger evidence shows that proximate and intermediate
explanatory variables for economic growth cause changes in natural capital (with significance varying
by sub-period). We interpret these results according to theoretical positions that conceptualize the
abundance of natural resources as an endogenous process (see Section 2, economic history approach).
7. Conclusions
The debate on the link between natural resources abundance and economic growth is still open.
Our contribution to this field consider a long-run approach (that covers the period 1870–2014) with
the analysis of a periphery country of the world economy—Uruguay—that has three features that
make it an interesting case: (i) the “internal” economy shows evidence of structural change but the
exports have remained highly concentrated in primary products (we identify this process with the
idea of duality of the structural change); (ii) it is, historically, a natural resources abundant economy
intensively positioned in renewable resources (in contrast with most extended analyses in the field
that focus, mainly, on non-renewable resources); (iii) the availability of information makes it possible
to compute using estimations of natural capital over a very long time frame.
The objective is to construct historical series of natural capital based on the World Bank
methodology. This methodology rests on the well-established economic principle that valuation
of assets should be calculated according to the present discounted value of economic rents (or profits)
over the life horizon of the resource. We contrast the level and evolution of the natural capital with
the level and growth of GDP, as well as the proximate and intermediate causes of economic growth
(produced and human capital, exports, terms of trade).
We show that natural capital tends to reduce its importance on the economy and, simultaneously,
increases its diversification. Although this evolution is consistent in historical terms, we do not find a
causal relationship between the abundance of natural resources and economic growth. Instead of a
direct relationship, the proximate and intermediate causes appear to be important in explaining the
evolution of natural capital when we consider three stages of economic growth: physical capital and
terms of trade during the agro-exporter model, human capital and exports during the period of import
Sustainability 2018, 10, 715 19 of 26
substitution industrialization and terms of trade from the 1960s afterwards. These factors cause the
growth of natural capital but not the other way around, which leads us to conclude that abundance of
natural capital is an endogenous process.
In other words, natural capital is not just a matter of endowment. Blessings or curses are “created”
by the effects of multiple factors, which change over time and go hand in hand with the economic
transformation that structural change implies. This notion is not new. “Resources are highly dynamic
concepts; they are not, they become, they evolve out of the triune interaction of nature, man and culture
. . . ” (Quoted in [131], p. 14, from a book of Erich Zimmerman of 1933). Natural resources “should
not be seen as merely a fortunate natural endowment but rather as a form of collective learning” [132]
(p. 186) and also as a return on investments, transportation, knowledge and the technologies of
natural exploitation. Our research has considered only some factors that can affect this process and
our evidence supports the inclusion of additional aspects such as technological and institutional issues
in the future.
Supplementary Materials: The following are available online at http://www.mdpi.com/2071-1050/10/3/715/s1,
Assumptions and sources.
Acknowledgments: We would like to thank Reto Bertoni and Luis Bértola for their comments and contributions,
and Carolina Román and Sabrina Siniscalchi who read our work and shared their knowledge. Thanks to
Nicolás Bonino, Bibiana Lanzilotta and Paola Azar for their help with econometric exercises and to Inés Moraes
and Pablo Castro for answering successive questions about sources. We are grateful to DINAMIGE, Dirección
General Forestal, Dardo Fagundez of OPYPA and to the World Bank also for responding to our doubts. We bear
sole responsibility for any errors that remain. We did not receive additional funding sources in the preparation of
this work or in its publication.
Author Contributions: Henry Willebald designed the research, Silvana Sandonato performed the research and
both authors analyzed the data and wrote the paper. Both authors read and approved the final manuscript.
Conflicts of Interest: The authors declare no conflict of interest.
Appendix. Measuring the Consumption Rate of Interest
In order to calculate total wealth, it is necessary an adequate estimation of the social time
preference rate or consumption rate of interest (CRI). CRI is the rate at which the marginal welfare of
consumption decreases over time. It is usually applied in cost-benefit analyses and environmental
economics (see [133] for a general discussion).
CRI estimate is based on the methodology proposed in [105] and, fundamentally, the decisions
adopted in [106,134]. Our basic components are the following:
• ρ = pure time preference.
• L = changing life chance (negative sign).
• µ = marginal utility of consumption.
• g = expected growth rate of consumption.
According to these definitions, we calculate the CRI as:
CRI = ρ − L + (µg)
ρ cannot be observed historically and a constant time preference—and equal to 0.3% (according
to [104])—is assumed over the full period of analysis.
µ is given by:
µ =
r − ρ
S
Y (r − y) + y
where:
S/Y: investment ratio,
r: expected rate of return on investment,
http://www.mdpi.com/2071-1050/10/3/715/s1
Sustainability 2018, 10, 715 20 of 26
y: expected growth rate of incomes from work.
To estimate the marginal utility of consumption, several considerations are required.
First of all, S/Y is the savings ratio and is calculated from historical national accounts for Uruguay.
Y is GDP and S the difference between GDP and total consumption; all data are derived from [124];
we use items in current prices.
Second, following [106], y is the expected growth rate in incomes from work and it is calculated as
the growth rate of total wages (W). W is also adjusted for change in consumption prices. We suppose
that expected growth rates are entirely based on the history of the variable. Therefore, we calculated a
stochastic but smooth trend in a structural time series framework. The slope of the trend in each year
represents the long-run expected growth rate of wages. This decision implies, implicitly, to suppose
that cyclical and irregular items of the time series correspond to business cycle phenomenon and did
not affect the expectations of the economic agents. Considering r, we observe that it is conceptually
similar to the expected growth rate of capital incomes. We measure r similarly to y but in this case
estimating the growth rate of gross profits (Y–W). Data were derived from [107] and elaborations of
the authors.
L represents that if possibilities for a long life are scarce, then a high interest rate is required to
motivate certain levels of savings. In empirical terms, L is the “crude death rate”: L = −(Total deaths/
Total population). Data were collected from Uruguayan vital statistics (1900–2014: Instituto de Estadística
(online); 1870–1899: Statistical Yearbooks (1899–1900) and authors’ estimates).
Finally, g is the long-term growth rate of consumption. Information derived from historical
national accounts presented in [124] and, as before, we estimated a stochastic but smooth trend in a
structural time series framework where the slope of the trend in a given year is understood as the
long-term expected growth rate.
The CRI is a rate to reflect society’s willingness to give up a unit of current consumption in
exchange for more in future consumption [135]. This rate can be used as a social discount rate as is
demonstrated in [136,137]. Social time preference rate is different to the individual time preference.
The reason is individual time preference rates are revealed from the market decisions, such as lending
and borrowing rates according to the current real interest rate. However, the decision of society’s
willingness to trade off consumption now for later is not solely based on the market but also other
factors. Therefore, the CRI is not equivalent to individual time preference. The manner in which
society is valuing the future depends on multiple factors as individual time preference, life expectancy,
expected incomes of the different social classes and the expectations related to the evolution of
future consumption.
In the case of Uruguay, the evolution of the CRI (Figure A1) presents a changing trajectory that
indicates the historic transformation of a “successful” society to another with huge doubts about the
future. This change was especially evident from the beginning of the second half of the 20th century
when the industrialization process exhausted and inflation started spiraled upward [138]. Referring to
this period, ([139] (p. 82–83), our translation) states:
“In a small economy which is susceptible to face major external shocks given its fragile
international trade relations, the inefficacy of the macroeconomic policy to stabilize the economy caused
very pronounced cyclical fluctuations. Therefore, even though the economy grows, deep crises reverse
some achievements obtained during the boom periods. Within this context, uncertainty increases and
investment decisions are made on a reduced time horizon with expectation of extraordinarily high
returns. In the case of Uruguay, these crisis episodes occur quite often and affect the agent’s behavior
which becomes more impatient and, therefore, less willing to invest in long-term projects, innovate,
develop or incorporate technology.”
Sustainability 2018, 10, 715 21 of 26
Sustainability 2017, 9, x FOR PEER REVIEW 20 of 26
The CRI is a rate to reflect society’s willingness to give up a unit of current consumption in
exchange for more in future consumption [135]. This rate can be used as a social discount rate as is
demonstrated in [136,137]. Social time preference rate is different to the individual time preference.
The reason is individual time preference rates are revealed from the market decisions, such as lending
and borrowing rates according to the current real interest rate. However, the decision of society’s
willingness to trade off consumption now for later is not solely based on the market but also other
factors. Therefore, the CRI is not equivalent to individual time preference. The manner in which
society is valuing the future depends on multiple factors as individual time preference, life
expectancy, expected incomes of the different social classes and the expectations related to the
evolution of future consumption.
In the case of Uruguay, the evolution of the CRI (Figure A1) presents a changing trajectory that
indicates the historic transformation of a “successful” society to another with huge doubts about the
future. This change was especially evident from the beginning of the second half of the 20th century
when the industrialization process exhausted and inflation started spiraled upward [138]. Referring
to this period, ([139] (p. 82–83), our translation) states:
“In a small economy which is susceptible to face major external shocks given its fragile
international trade relations, the inefficacy of the macroeconomic policy to stabilize the economy
caused very pronounced cyclical fluctuations. Therefore, even though the economy grows, deep
crises reverse some achievements obtained during the boom periods. Within this context, uncertainty
increases and investment decisions are made on a reduced time horizon with expectation of
extraordinarily high returns. In the case of Uruguay, these crisis episodes occur quite often and affect
the agent’s behavior which becomes more impatient and, therefore, less willing to invest in long-term
projects, innovate, develop or incorporate technology.”
Figure A1. The consumption rate of interest in Uruguay, 1870–2014. Source: developed by authors.
1. Sachs, J. How to handle the macroeconomics of oil wealth. In Escaping the Resource Curse; Humphreys, M.,
Sachs, J., Stiglitz, J., Eds.; Columbia University Press: New York, NY, USA, 2007; pp. 173–193.
2. Badeeb, R.; Lean, H.; Clark, J. The evolution of the natural resource curse thesis: A critical literature survey.
Resour. Policy 2017, 51, 123–134.
3. Auty, R. Sustaining Development in Mineral Economies: The Resource Curse Thesis; Routledge: London, UK,
1993.
4. Manzano, O.; Rigobon, R. Resource Curse of Debt Overhang? National Bureau of Economic Research:
Cambridge, MA, USA, 2001; pp. 1–37.
5. Leite, C.; Weidmann, J. Does Mother Nature Corrupt? Natural Resources, Corruption and Economic Growth;
International Monetary Fund: Washington, DC, USA, 1999.
6. Van der Ploeg, F. Natural resources: Curse or blessing? J. Econ. Lit. 2011, 49, 366–420.
7. Torvik, R. Why do some resource-abundant countries succeed while others do not? Oxf. Rev. Econ. Policy
2009, 25, 241–256.
8. Willebald, H.; Badía-Miró, M.; Pinilla, V. Introduction: Natural resources and economic development.
What can we learn from history? In Natural Resources and Economic Growth: Learning from History; Badía-
Figure A1. The consumption rate of interest in Uruguay, 1870–2014. Source: developed by authors.
References
1. Sachs, J. How to handle the macroeconomics of oil wealth. In Escaping the Resource Curse; Humphreys, M.,
Sachs, J., Stiglitz, J., Eds.; Columbia University Press: New York, NY, USA, 2007; pp. 173–193.
2. Badeeb, R.; Lean, H.; Clark, J. The evolution of the natural resource curse thesis: A critical literature survey.
Resour. Policy 2017, 51, 123–134. [CrossRef]
3. Auty, R. Sustaining Development in Mineral Economies: The Resource Curse Thesis; Routledge: London, UK, 1993.
4. Manzano, O.; Rigobon, R. Resource Curse of Debt Overhang? National Bureau of Economic Research: Cambridge,
MA, USA, 2001; pp. 1–37.
5. Leite, C.; Weidmann, J. Does Mother Nature Corrupt? Natural Resources, Corruption and Economic Growth;
International Monetary Fund: Washington, DC, USA, 1999.
6. Van der Ploeg, F. Natural resources: Curse or blessing? J. Econ. Lit. 2011, 49, 366–420. [CrossRef]
7. Torvik, R. Why do some resource-abundant countries succeed while others do not? Oxf. Rev. Econ. Policy
2009, 25, 241–256. [CrossRef]
8. Willebald, H.; Badía-Miró, M.; Pinilla, V. Introduction: Natural resources and economic development. What can we
learn from history? In Natural Resources and Economic Growth: Learning from History; Badía-Miró, M., Pinilla, V.,
Willebald, H., Eds.; Routledge Explorations in Economic History; Taylor & Francis Group: London, UK;
New York, NY, USA, 2015; pp. 21–45.
9. Barbier, E. Frontier expansion and economic development. Contemp. Econ. Policy 2005, 23, 165–316. [CrossRef]
10. David, P.; Wright, G. Increasing returns and the genesis of American resource abundance. Ind. Corp. Chang.
1997, 6, 203–245. [CrossRef]
11. Wright, G. Resource-based growth then and now. In Patterns of Integration in the Global Economy; World Bank
Project, Ed.; World Bank: Washington, DC, USA, 2001.
12. Cimoli, M.; Porcile, G.; Primi, A.; Vergara, S. Cambio estructural, heterogeneidad productiva y tecnología
en América Latina. In Heterogeneidad Estructural, Asimetrías Tecnológicas y Crecimiento en América Latina;
Cimoli, M., Ed.; CEPAL-BID: Santiago de Chile, Chile, 2005; pp. 9–39.
13. Presbich, R. El desarrollo económico de la América Latina y algunos de sus principales problemas. El Trimest. Econ.
1949, 16, 347–431.
14. Singer, H. The distribution of gains between investing and borrowing countries. Am. Econ. Rev. 1950, 40,
473–485.
15. Seers, D. A model of comparative rates of growth in the world economy. Econ. J. 1962, 72, 45–78. [CrossRef]
16. McCombie, J.; Roberts, M. The role of the balance of payments in economic growth. In The Economics of
Demand-Led Growth. Challenging the Supply-Side Vision of the Long Run; Setterfield, M., Ed.; Wiley: Hoboken,
NJ, USA, 2002; pp. 87–114.
17. Dosi, G.; Nelson, R. Technical change and industrial dynamics as evolutionary processes. In Handbook of the
Economics of Innovation; Hall, B., Rosenberg, N., Eds.; North Holland: Amsterdam, The Netherlands, 2010;
pp. 51–127.
18. Corden, W.; Neary, J. Booming sector and de-industrialization in a small open economy. Econ. J. 1982, 92,
825–848. [CrossRef]
http://dx.doi.org/10.1016/j.resourpol.2016.10.015
http://dx.doi.org/10.1257/jel.49.2.366
http://dx.doi.org/10.1093/oxrep/grp015
http://dx.doi.org/10.1093/cep/byi022
http://dx.doi.org/10.1093/icc/6.2.203
http://dx.doi.org/10.2307/2228617
http://dx.doi.org/10.2307/2232670
Sustainability 2018, 10, 715 22 of 26
19. Corden, W. Booming sector and Dutch disease economics: Survey and consolidation. Oxf. Econ. Pap. 1984,
36, 359–380. [CrossRef]
20. Neary, J.; van Wijnbergen, S. Natural Resources and the Macroeconomy; MIT Press: Cambridge, MA, USA, 1986.
21. Gylfason, T. Natural resources and economic growth: From dependence to diversification. In Economic Liberalization
and Integration Policy: Options for Eastern Europe and Russia; Broadman, H.G., Paas, T., Welfens, P.J.J., Eds.; Springer:
Berlín, Germany, 2006.
22. Sachs, J.; Warner, A. The curse of natural resources. Eur. Econ. Rev. 2001, 45, 827–838. [CrossRef]
23. Auty, R.
and Overview. In Resource Abundance and Economic Development; Auty, R., Ed.; Oxford
University Press: Oxford, UK, 2001; pp. 3–18.
24. Bulte, E.; Damania, R.; Deacon, R. Resource intensity, institutions, and development. World Dev. 2005, 33,
1029–1044. [CrossRef]
25. Sala-i-Martin, X.; Subramanian, A. Addressing the natural resource curse: An illustration from Nigeria.
J. Afr. Econ. 2013, 22, 570–615. [CrossRef]
26. Auty, R. The political economy of resource-driven growth. Eur. Econ. Rev. 2001, 45, 839–846. [CrossRef]
27. Isham, J.; Woolcock, M.; Pritchett, L.; Busby, G. The varieties of resource experience: Natural resource export
structures and the political economy of economic growth. World Bank Econ. Rev. 2005, 19, 141–174. [CrossRef]
28. Woolcock, M.; Prichett, L.; Isham, J. The social foundations of poor economic growth in resource-rich economies.
In Resource Abundance and Economic Development; Oxford University Press: New York, NY, USA, 2001.
29. Boschini, A.; Pettersson, J.; Roine, J. Resource curse or not: A question of appropriability. Scand. J. Econ. 2007,
109, 593–617. [CrossRef]
30. Birdsall, N.; Pinckney, T.; Sabot, R. Natural Resources, Human Capital, and Growth; Carnegie Endowment for
International Peace: Washington, DC, USA, 2000.
31. Wood, A.; Mayer, J. Africa’s export structure in a comparative perspective. Camb. J. Econ. 2001, 25, 369–394.
[CrossRef]
32. Bravo, C.; De Gregorio, J. The Relative Richness of the Poor? Natural Resources, Human Capital and Economic
Growth; The World Bank Policy Research; World Bank: Washington, DC, USA, 2005.
33. Barbier, E. Scarcity, frontiers and the resource curse: A historical perspective. In Natural Resources and Economic
Growth: Learning from History; Badía-Miró, M., Pinilla, V., Willebald, H., Eds.; Routledge Explorations in
Economic History; Taylor & Francis Group: London, UK; New York, NY, USA, 2015; pp. 54–76.
34. Acemoglu, D.; Johnson, S.; Robinson, J. The colonial origins of comparative development: An empirical
Investigation. Am. Econ. Rev. 2001, 91, 1369–1401. [CrossRef]
35. Engerman, S.; Sokoloff, K. Factor endowments, institutions, and differential paths of growth among new
world economies. In How Latin America Fell Behind: Essays on the Economic Histories of Brazil and Mexico;
Haber, S., Ed.; Stanford University Press: Stanford, CA, USA, 1997; pp. 260–304.
36. Engerman, S.; Sokoloff, K. Factor Endowments, Inequality, and Paths of Development among New World Economies;
National Bureau of Economic Research: Cambridge, MA, USA, 2002.
37. Gelb, A.H. Oil Windfalls: Blessing or Curse? Cambridge University Press: New York, NY, USA, 1998.
38. Sachs, J.; Warner, A. Natural Resource Abundance and Economic Growth; National Bureau of Economic Research:
Cambridge, MA, USA, 1995.
39. Neumayer, E. Does the “resource curse” hold for growth in genuine income as well? World Dev. 2004, 32,
1627–1640. [CrossRef]
40. Arezki, R.; Van der Ploeg, F. Can the Natural Resource Curse be turned into a Blessing? The Role of Trade Policies
and Institutions; International Monetary Fund: Washington, DC, USA, 2011.
41. Beck, T. Finance and Oil: Is There a Resource Curse in Financial Development? Available online: https://papers.
ssrn.com/sol3/papers.cfm?abstract_id=1769803 (accessed on 20 February 2018).
42. Boschini, A.; Pettersson, J.; Roine, J. The resource curse and its potential reversal. World Dev. 2013, 43, 19–41.
[CrossRef]
43. Ross, M. A closer look at oil, diamonds, and civil war. Annu. Rev. Political Sci. 2006, 9, 265–300. [CrossRef]
44. Auty, R. The resource curse and sustainable development. In Handbook of Sustainable Development;
Atkinson, G., Dietz, S., Neumayer, E., Eds.; Edward Elgar: Cheltenham, UK, 2007; pp. 207–219.
45. Collier, P.; Hoeffler, A. Testing the neocon agenda: Democracy in resource-rich societies. Eur. Econ. Rev. 2009,
53, 293–308. [CrossRef]
http://dx.doi.org/10.1093/oxfordjournals.oep.a041643
http://dx.doi.org/10.1016/S0014-2921(01)00125-8
http://dx.doi.org/10.1016/j.worlddev.2005.04.004
http://dx.doi.org/10.1093/jae/ejs033
http://dx.doi.org/10.1016/S0014-2921(01)00126-X
http://dx.doi.org/10.1093/wber/lhi010
http://dx.doi.org/10.1111/j.1467-9442.2007.00509.x
http://dx.doi.org/10.1093/cje/25.3.369
http://dx.doi.org/10.1257/aer.91.5.1369
http://dx.doi.org/10.1016/j.worlddev.2004.05.005
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1769803
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1769803
http://dx.doi.org/10.1016/j.worlddev.2012.10.007
http://dx.doi.org/10.1146/annurev.polisci.9.081304.161338
http://dx.doi.org/10.1016/j.euroecorev.2008.05.006
Sustainability 2018, 10, 715 23 of 26
46. Boos, A.; Holm-Müller, K. The relationship between the resource curse and genuine savings: Empirical
evidence. J. Sustain. Dev. 2013, 6, 59–72. [CrossRef]
47. Bhattacharyya, S.; Hodler, R. Do Natural Resource Revenues Hinder Financial Development? The Role of Political
Institutions; Oxford Centre for the Analysis of Resource Rich Economies: Oxford, UK, 2014.
48. Bhattacharyya, S.; Collier, P. Public Capital in Resource Rich Economies: Is there a Curse? Center for the Study of
African Economies: Oxford, UK, 2014.
49. Gylfason, T. Nature, power and growth. Scott. J. Political Econ. 2001, 48, 558–588. [CrossRef]
50. Dietz, S.; Neumayer, E.; De Soysa, I. Corruption, the resource curse and genuine saving. Environ. Dev. Econ.
2007, 12, 33–53. [CrossRef]
51. Barajas, M.; Chami, M.; Yousefi, M. The Finance and Growth Nexus Re-Examined: Do All Countries Benefit Equally?
International Monetary Fund: Washington, DC, USA, 2013.
52. Daniele, V. Natural resources and the ‘quality’of economic development. J. Dev. Stud. 2011, 47, 545–573.
[CrossRef]
53. Stijns, J. Natural resource abundance and economic growth revisited. Resour. Policy 2005, 30, 107–130.
[CrossRef]
54. Brunnschweiler, C. Cursing the blessings? Natural resource abundance, institutions, and economic growth.
World Dev. 2008, 36, 399–419. [CrossRef]
55. Brunnschweiler, C.; Bulte, E. The resource curse revisited and revised: A tale of paradoxes and red herrings.
J. Environ. Econ. Manag. 2008, 55, 248–264. [CrossRef]
56. Dartey-Baah, K.; Amponsah-Tawiah, K.; Aratuo, D. Emerging “Dutch disease” in emerging oil economy:
Ghana’s perspective. Soc. Bus. Rev. 2012, 7, 185–199. [CrossRef]
57. Papyrakis, E.; Gerlagh, R. Resource abundance and economic growth in the United States. Eur. Econ. Rev.
2007, 51, 1011–1039. [CrossRef]
58. Pegg, S. Is there a Dutch disease in Botswana? Resour. Policy 2010, 35, 14–19. [CrossRef]
59. Fosu, A.; Gyapong, A. Terms of trade and growth of resource economies: Contrasting evidence from two
African countries. In Beyond the Curse: Policies to Harness the Power of Natural Resources; Arezki, R., Gylfason, T.,
Sy, A., Eds.; International Monetary Fund: Washington, DC, USA, 2011; pp. 257–272.
60. De Gregorio, J.; Labbé, F. Copper, The real exchange rate and macroeconomic fluctuations in Chile. In Beyond
the Curse: Policies to Harness the Power of Natural Resources; Arezki, R., Gylfason, T., Sy, A., Eds.; International
Monetary Fund: Washington, DC, USA, 2011; pp. 203–233.
61. Gylfason, T. Natural Resource Endowment: A Mixed Blessing? Available online: https://papers.ssrn.com/
sol3/papers.cfm?abstract_id=1766385 (accessed on 20 February 2018).
62. Loayza, N.; Mier y Teran, A.; Rigolini, J. Poverty, Inequality, and the Local Natural Resource Curse; World Bank,
Policy Research Working Paper; World Bank: Washington, DC, USA, 2013.
63. Liu, Y. Is the natural resource production a blessing or curse for China’s urbanization? Evidence from a
space–time panel data model. Econ. Model. 2014, 38, 404–416. [CrossRef]
64. Hammond, J. The resource curse and oil revenues in Angola and Venezuela. Sci. Soc. 2011, 75, 348–378.
[CrossRef]
65. Rubio, M. Oil illusion and delusion: Mexico and Venezuela over the twentieth century. In Natural Resources
and Economic Growth: Learning from History; Badía-Miró, M., Pinilla, V., Willebald, H., Eds.; Routledge
Explorations in Economic History; Taylor & Francis Group: London, UK; New York, NY, USA, 2015;
pp. 160–183.
66. Marwah, H. Oil as sweet as honey: Linking natural resources, government institutions and domestic
capital investment in Nigeria 1960–2000. In Natural Resources and Economic Growth: Learning from History;
Badía-Miró, M., Pinilla, V., Willebald, H., Eds.; Routledge Explorations in Economic History; Taylor & Francis
Group: London, UK; New York, NY, USA, 2015; pp. 100–118.
67. Stijns, J. Natural resource abundance and human capital accumulation. World Dev. 2006, 34, 1060–1083.
[CrossRef]
68. Blanco, L.; Grier, R. Natural resource dependence and the accumulation of physical and human capital in
Latin America. Resour. Policy 2012, 37, 281–295. [CrossRef]
69. Shao, S.; Yang, L. Natural resource dependence, human capital accumulation, and economic growth:
A combined explanation for the resource curse and the resource blessing. Energy Policy 2014, 74, 632–642.
[CrossRef]
http://dx.doi.org/10.5539/jsd.v6n6p59
http://dx.doi.org/10.1111/1467-9485.00215
http://dx.doi.org/10.1017/S1355770X06003378
http://dx.doi.org/10.1080/00220388.2010.506915
http://dx.doi.org/10.1016/j.resourpol.2005.05.001
http://dx.doi.org/10.1016/j.worlddev.2007.03.004
http://dx.doi.org/10.1016/j.jeem.2007.08.004
http://dx.doi.org/10.1108/17465681211237637
http://dx.doi.org/10.1016/j.euroecorev.2006.04.001
http://dx.doi.org/10.1016/j.resourpol.2009.07.003
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1766385
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1766385
http://dx.doi.org/10.1016/j.econmod.2014.01.023
http://dx.doi.org/10.1521/siso.2011.75.3.348
http://dx.doi.org/10.1016/j.worlddev.2005.11.005
http://dx.doi.org/10.1016/j.resourpol.2012.01.005
http://dx.doi.org/10.1016/j.enpol.2014.07.007
Sustainability 2018, 10, 715 24 of 26
70. Atkinson, G.; Hamilton, K. Savings, growth and the resource curse hypothesis. World Dev. 2003, 31, 1793–1801.
[CrossRef]
71. Wood, A.; Berge, K. Exporting manufactures: Human resources, Natural Resources, and Trade Policy.
J. Dev. Stud. 1997, 34, 35–59. [CrossRef]
72. Bornhorst, F.; Thornton, J.; Gupta, S. Natural Resource Endowments, Governance, and the Domestic Revenue Effort:
Evidence from a Panel of Countries; International Monetary Fund: Washington, DC, USA, 2008.
73. Mehlum, H.; Moene, K.; Torvik, R. Institutions and the resource curse. Econ. J. 2006, 116, 1–20. [CrossRef]
74. Alexeev, M.; Conrad, R. The elusive curse of oil. Rev. Econ. Stat. 2009, 91, 586–598. [CrossRef]
75. Lederman, D.; Maloney, W. Natural Resources: Neither Curse nor Destiny; The World Bank: Washington,
DC, USA, 2007.
76. Boyce, J.; Herbert Emery, J. Is a negative correlation between resource abundance and growth sufficient
evidence that there is a “resource curse”? Resour. Policy 2011, 36, 1–13. [CrossRef]
77. Cavalcanti, T.; Mohaddes, K.; Raissi, M. Growth, development and natural resources: New evidence using a
heterogeneous panel analysis. Q. Rev. Econ. Finance 2011, 51, 305–318. [CrossRef]
78. James, A. The resource curse: A statistical mirage? J. Dev. Econ. 2015, 114, 55–63. [CrossRef]
79. Auty, R. From resource cruse to rent curse: A theoretical perspective. In Natural Resources and Economic
Growth: Learning from History; Badía-Miró, M., Pinilla, V., Willebald, H., Eds.; Routledge Explorations in
Economic History; Taylor & Francis Group: London, UK; New York, NY, USA, 2015.
80. Lewis, W. Crecimiento y Fluctuaciones; Fundación de Cultura Económica: Buenos Aires, Argentina, 1983; p. 209.
81. Foreman-Peck, J. A History of the World Economy. International Economic Relations since 1850; Harvester Press:
Brighton, UK, 1983.
82. Álvarez, J.; Bilancini, E.; D’Alessandro, S.; Porcile, G. Agricultural institutions, industrialization and growth:
The case of New Zealand and Uruguay in 1870–1940. Explor. Econ. Hist. 2011, 48, 151–168. [CrossRef]
83. Willebald, H.; Bértola, L. Uneven development paths among Settler Societies, 1870–2000. In Settler Economies
in World History; Lloyd, C., Metzer, J., Sutch, R., Eds.; Brill: Leiden, The Netherlands, 2013; pp. 105–140.
84. Wellstead, A. The (post) staples economy and the (post) staples state in historical perspective. Can. Political
Sci. Rev. 2007, 1, 8–25.
85. Bértola, L. Overview of the Economic History of Uruguay Since the 1870s. Available online: http://eh.net/
encyclopedia/article/Bertola.Uruguay.final (accessed on 20 February 2018).
86. Oddone, G. El Declive. Una Mirada de la Economía de Uruguay en el Siglo XX; Librería Linardi y Risso:
Montevideo, Uruguay, 2010.
87. Bulmer-Thomas, V. The Economic History of Latin America since Independence; Cambridge University Press:
Cambridge, UK, 2003.
88. Bonino, N.; Tena, A.; Willebald, H. Uruguay and the first globalization: On the accuracy of export performance,
1870–1913. Rev. Hist. Econ. 2015, 33, 287–320.
89. Willebald, H. Desigualdad y especialización en el crecimiento de las economías templadas de nuevo
asentamiento, 1870–1940. Rev. Hist. Econ. 2007, 25, 293–348. [CrossRef]
90. Bonino, N.; Román, C.; Willebald, H. Structural change and long-term patterns. A methodological proposal
for Uruguay in the very long run. In Proceedings of the Seminario de Investigación, Programa de Historia
Económica y Social, PHES-FCS-UdelaR, Montevideo, Uruguay, 6 September 2012.
91. Jacob, R. Uruguay 1929–1938: Depresión Ganadera y Desarrollo Fabril; Fundación de Cultura Universitaria:
Montevideo, Uruguay, 1981; p. 432.
92. Bértola, L. La Industria Manufacturera Uruguaya 1913–1961: Un Enfoque Sectorial de su Crecimiento, Fluctuaciones
y Crisis; Facultad de Ciencias Sociales de la Universidad de la República y Centro Interdisciplinario de
Estudios sobre el Desarrollo Uruguay: Montevideo, Uruguay, 1993.
93. Arnabal, L.; Bertino, M.; Fleitas, S. Una revisión Del desempeño de la industria en Uruguay entre 1930 y 1959.
Rev. Hist. Ind. 2013, 13, 143–173.
94. Astori, D. Estancamiento, desequilibrios y ruptura. 1955–1972. In El Uruguay Del Siglo XX; Instituto de
Economía, Ed.; Ediciones de la Banda Oriental: Montevideo, Uruguay, 2001.
95. Paolino, C.; Pittaluga, L.; Mondelli, M. Cambios en la Dinámica Agropecuaria y Agroindustrial del Uruguay y Las
Políticas Públicas; Serie Estudios y Perspectivas CEPAL; CEPAL: Santiago de Chile, Chile, 2014.
96. Bértola, L.; Isabella, F.; Saavedra, C. El Ciclo Económico de Uruguay, 1998–2012; Documento de Trabajo Facultad
de Ciencias Sociales; Facultad de Ciencias Sociales: Montevideo, Uruguay, 2014.
http://dx.doi.org/10.1016/j.worlddev.2003.05.001
http://dx.doi.org/10.1080/00220389708422502
http://dx.doi.org/10.1111/j.1468-0297.2006.01045.x
http://dx.doi.org/10.1162/rest.91.3.586
http://dx.doi.org/10.1016/j.resourpol.2010.08.004
http://dx.doi.org/10.1016/j.qref.2011.07.007
http://dx.doi.org/10.1016/j.jdeveco.2014.10.006
http://dx.doi.org/10.1016/j.eeh.2010.05.004
http://eh.net/encyclopedia/article/Bertola.Uruguay.final
http://eh.net/encyclopedia/article/Bertola.Uruguay.final
http://dx.doi.org/10.1017/S0212610900000148
Sustainability 2018, 10, 715 25 of 26
97. Reyes Abadie, W. La Banda Oriental: Pradera, Frontera, Puerto; Ediciones de la Banda Oriental: Montevideo,
Uruguay, 1966.
98. Barrán, J.; Nahum, B. Historia Rural del Uruguay Moderno; Ediciones de la Banda Oriental: Montevideo,
Uruguay, 1978.
99. Instituto de Economía. El Proceso Económico del Uruguay; Fondo de Cultura Universitaria: Montevideo,
Uruguay, 1969.
100. World Bank. The Changing Wealth of Nations: Measuring Sustainable Development in the New Millennium;
The World Bank: Washington, DC, USA, 2011.
101. World Bank. Where Is the Wealth of Nations? Measuring Capital for the 21st Century; The World Bank:
Washington, DC, USA, 2006.
102. Ding, N.; Field, B. Natural Resource Abundance and Economic Growth; University of Massachusetts Amherst:
Amherst, MA, USA, 2005.
103. Sandonato, S. Capital Natural en Uruguay. 1990–2010. Propuesta Metodológica, Estimaciones y Ejercicios de
Descomposición. Bachelor’s Thesis, Universidad de la República, Montevideo, Uruguay, 2012.
104. Willebald, H.; Sandonato, S. Indicadores de Capital Natural; Nota Técnica Nº 1, Series Documentos del Reporte
Anual 2014; Recursos Naturales y Desarrollo Red Sudamericana de Economía Aplicada: Montevideo, Uruguay.
105. Pearce, D.; Ulph, D. A social discount rate for the United Kingdom. In Environmental Economics: Essays in
Ecological Economics and Sustainable Development; Pearce, D., Ed.; Edward Elgar Publishing: Cheltenham, UK,
1999; pp. 268–285.
106. Lindmark, M.; Andersson, L. Where Was the Wealth of the Nation? Measuring Swedish Capital for the 19th and
20th Centuries; Centre for Environmental and Resource Economics: Umea, Suecia, 2016.
107. De Rosa, M.; Sinisclachi, S.; Vilá, J.; Vigorito, A.; Willebald, H. La Evolución de Las Remuneraciones
Laborales y la Distribución del Ingreso en Uruguay; Futuro en Foco, Cuadernos Sobre Desarrollo Humano:
Montevideo, Uruguay, 2017.
108. Román, C.; Bertoni, R. Auge y ocaso Del carbón mineral en Uruguay. Un análisis histórico desde fines del
siglo XIX hasta la actualidad. Rev. Hist. Econ. 2013, 31, 459–497.
109. Lampietti, J.; Dixon, J. To See the Forest for the Trees: A Guide to Non-Timber Forest Benefits; Environment
Department Paper; World Bank: Washington, DC, USA, 1995.
110. Giles, J.; Mirza, S. Some Pretesting Issues on Testing for Granger Non Causality; Econometrics Working Paper;
University of Victoria: Victoria, BC, Canada, 1999.
111. Lingarde, S.; Tylecote, A. Resource-rich countries’ success and failure in technological ascent, 1870–1970:
The Nordic countries versus Argentina, Uruguay and Brazil. J. Eur. Econ. Hist. 1999, 28, 77–114.
112. Bertino, M.; Bertoni, R.; Tajam, H.; Yaffé, J. Historia Económica Del Uruguay. Tomo III: La Economía Del Batllismo
y de los Años Veinte; Fin de Siglo: Montevideo, Uruguay, 2005.
113. Moraes, I. La Pradera Perdida. Historia y Economía Del Agro Uruguayo: Una Visión de Largo Plazo, 1760–1970;
Linardi y Risso: Montevideo, Uruguay, 2008.
114. Martín-Retortillo, M.; Pinilla, V.; Velazco, J.; Willebald, H. The goose that laid the golden eggs? Agricultural
development in Latin America in the 20th century. In Agricultural Development in the World Periphery: A Global
Economic History Approach; Pinilla, V., Willebald, H., Eds.; Palgrave Studies in Economic History; Palgrave
Macmillan: Basingstoke, UK, 2018.
115. OPYPA-MGAP. El Desarrollo Agropecuario y Agro-Industrial de Uruguay; Reflexiones en el 50 Aniversario de la
Oficina de Programación y Política Agropecuaria; MGAP-OPYPA: Montevideo, Uruguay, 2015.
116. Castro, P. Distribución Regional de la Producción y Geografía Económica. El Caso Del Agro en Uruguay
(1870–2008). Master’s Thesis, Universidad de la República, Montevideo, Uruguay, 2017.
117. Bertoni, R. Energía y Desarrollo. La Restricción Energética en Uruguay Como Problema: 1882–2000; CSIC:
Montevideo, Uruguay, 2011.
118. Maddison, A. Ultimate and proximate growth causality: A critique of Mancur Olson on the rise and decline
of nations. Scand. Hist. Rev. 1988, 36, 25–29. [CrossRef]
119. Abramovitz, M. Thinking about growth. In Thinking about Growth and Other Essays on Economic Growth
and Welfare; Abramovitz, M., Ed.; Cambridge University Press: Cambridge, UK, 1989; pp. 3–79.
120. Rodrik, D. In Search of Prosperity. Analytic Narratives on Economic Growth; Princeton University Press:
Princeton, NJ, USA, 2003.
http://dx.doi.org/10.1080/03585522.1988.10408114
Sustainability 2018, 10, 715 26 of 26
121. Szirmai, A. Proximate, Intermediate and Ultimate Causality: Theories and Experiences of Growth and Development;
Working Paper Series on Institutions and Economic Growth; Maastricht Graduate School of Governance:
Maastricht, The Netherlands, 2012.
122. Román, C.; Willebald, H. Formación de capital en el largo plazo en Uruguay, 1870–2011. Investig. Hist. Econ.
2015, 11, 20–30. [CrossRef]
123. Fleitas, S.; Rius, A.; Román, C.; Willebald, H. Contract Enforcement, Investment and Growth in Uruguay
since 1870; Instituto de Economía: Montevideo, Uruguay, 2013.
124. Román, C. El PIB Histórico Por el Lado Del Gasto en Uruguay, 1870–2016: Una Primera Aproximación, Unpublished work.
125. Williamson, J. Land, labor and globalization in the third world, 1870–1940. J. Econ. Hist. 2002, 62, 55–85.
126. Bértola, L.; Lorenzo, F. Witches in the south: Kuzntes-like swings in Argentina, Brazil and Uruguay since
the 1870. In The Experience of Economic Growth; van Zanden, J., Heikenen, S., Eds.; Amsterdam University
Press: Amsterdam, The Netherlands, 2004.
127. Azar, P.; Bertino, M.; Bertoni, R.; Fleitas, S.; García Repetto, U.; Sanguinetti, C.; Sienra, M.; Torrelli, M.
De Quiénes, Para Quiénes y Para Qué. Las Finanzas Públicas en el Uruguay del Siglo XX; Editorial Fin de Siglo:
Montevideo, Uruguay, 2009.
128. Azar, P. Public Education Spending: Efficiency, Productivity and Politics. Ph.D. Thesis, Universtat Autònoma
de Barcelona, Barcelona, España, 2017.
129. Roldós, J. A long-run perspective on trade policy, instability, and growth. In The Effects of Protectionism on a Small
Country. The Case of Uruguay; Connolly, M., De Melo, J., Eds.; The World Bank: Washington, DC, USA, 1994.
130. Willebald, H. Land abundance, frontier expansion and appropriability: Settler economies during the first
globalization. In Natural Resources and Economic Growth: Learning from History; Badía-Miró, M., Pinilla, V.,
Willebald, H., Eds.; Routledge Explorations in Economic History; Taylor & Francis Group: London, UK;
New York, NY, USA, 2015; pp. 248–270.
131. Schrijver, N. Sovereignty over Natural Resources: Balancing Rights and Duties; Cambridge University Press:
New York, NY, USA, 2008.
132. Czelusta, J.; Wright, G. Resource-based growth past and present. In Natural Resources: Neither Curse nor
Destiny; Lederman, D., Maloney, W., Eds.; World Bank/Stanford University Press: Stanford, CA, USA, 2007;
pp. 183–211.
133. Angelsen, A. Cost-Benefit Analysis, Discounting, and the Environmental Critique: Overloading of the Discount
Rate? Michelsen Institute Report; Chr. Michelsen Institute: Bergen, Norway, 1991.
134. Lindmark, M.; Acar, S. Sustainability in the making? A Historical estimate of Swedish sustainable and
unsustainable development 1850–2000. Ecol. Econ. 2013, 86, 176–187. [CrossRef]
135. Chua, A.; Choong, W. A review of approaches to construct social discount rate. Sains Humanika 2016, 8,
37–42. [CrossRef]
136. Sen, A. On optimizing the rate of saving. Econ. J. 1961, 71, 479–496. [CrossRef]
137. Kay, J. Social discount rates. J. Public Econ. 1972, 1, 359–378. [CrossRef]
138. Brum, C.; Román, C.; Willebald, H. Un Enfoque Monetario de la Inflación en el Largo Plazo. El Caso de Uruguay
(1870–2010); El Trimestre Económico: Montevideo, Uruguay, 2016.
139. Oddone, G. Restricciones para sostener el crecimiento: Lecciones y desafíos para las políticas públicas. In La
Aventura Uruguaya. El País y el Mundo; Arocena, R., Caetano, G., Eds.; Debate: Montevideo, Uruguay, 2011;
pp. 67–90.
© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
http://dx.doi.org/10.1016/j.ihe.2013.09.002
http://dx.doi.org/10.1016/j.ecolecon.2012.06.021
http://dx.doi.org/10.11113/sh.v8n4-3.1079
http://dx.doi.org/10.2307/2228914
http://dx.doi.org/10.1016/0047-2727(72)90012-6
http://creativecommons.org/
http://creativecommons.org/licenses/by/4.0/.
Theoretical Frameworks
The Blessing of Natural Resource Abundance
Production Structure Approach: The Difficulties of a Primary Sector Specialization
Crowding-Out Approach: Natural Resources Displace Other Types of Capital
Institutional Change and Factor Endowment Approach
Economic History Approaches
The Empirical Evidence of the Resource Curse
The Impact of Natural Resources on Economic Growth
The Impact of Resources on Factors Linked to Growth
Is This Apparent Paradox a Red Herring?
Historical Overview and Reasons to Study Uruguay
Duality of the Structural Change
The Relevance of Studying the Non-Mineral Wealth
Natural Capital as a Better Proxy for the Abundance of Natural Resources
Empirical Strategy
Natural Capital Estimation
Natural Capital by Component
Cropland
Pastureland
Mineral Resources
Timber Resources
Non-Timber Forest Resources
Protected Areas
Comparison with Previous Estimates
Contrast with Linear Non-Causality
Results
Some Stylized Facts
Causality Exercises
References
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/330116952
The Relationship Between Renewable Energy Consumption and Economic
Growth: The Case of
Bulgaria
Article in International Journal of Energy Sector Management · September 2019
DOI: 10.1108/IJESM-11-2017-0005
CITATIONS
40
READS
2,732
2 authors:
Some of the authors of this publication are also working on these related projects:
Credit Bubbles View project
The Relationship Economic Growth and Information Technologies : The Case of Bulgaria View project
Hamit Can
Technical University of Sofia
142 PUBLICATIONS 56 CITATIONS
SEE PROFILE
Özge Korkmaz
Malatya Turgut Özal University
53 PUBLICATIONS 305 CITATIONS
SEE PROFILE
All content following this page was uploaded by Hamit Can on 07 October 2019.
The user has requested enhancement of the downloaded file.
https://www.researchgate.net/publication/330116952_The_Relationship_Between_Renewable_Energy_Consumption_and_Economic_Growth_The_Case_of_Bulgaria?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_2&_esc=publicationCoverPdf
https://www.researchgate.net/publication/330116952_The_Relationship_Between_Renewable_Energy_Consumption_and_Economic_Growth_The_Case_of_Bulgaria?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_3&_esc=publicationCoverPdf
https://www.researchgate.net/project/Credit-Bubbles?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_9&_esc=publicationCoverPdf
https://www.researchgate.net/project/The-Relationship-Economic-Growth-and-Information-Technologies-The-Case-of-Bulgaria?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_9&_esc=publicationCoverPdf
https://www.researchgate.net/?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_1&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Hamit-Can?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_4&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Hamit-Can?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_5&_esc=publicationCoverPdf
https://www.researchgate.net/institution/Technical-University-of-Sofia?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_6&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Hamit-Can?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_7&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Oezge-Korkmaz-2?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_4&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Oezge-Korkmaz-2?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_5&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Oezge-Korkmaz-2?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_7&_esc=publicationCoverPdf
https://www.researchgate.net/profile/Hamit-Can?enrichId=rgreq-b66164f59c4d78d98339e3c397326906-XXX&enrichSource=Y292ZXJQYWdlOzMzMDExNjk1MjtBUzo4MTEyNzU0Mjg4MjcxMzZAMTU3MDQzNDU0MzM2Nw%3D%3D&el=1_x_10&_esc=publicationCoverPdf
The relationship between
renewable energy
consumption
and economic growth
The case of Bulgaria
Hamit Can
Technical University of Sofia, Sofia, Bulgaria, and
Özge Korkmaz
Bayburt Universitesi, Bayburt, Turkey
Abstract
Purpose – The purpose of this study is to investigate the relationship between renewable energy and
economic growth of Bulgaria.
Design/methodology/approach – This study analyzes the relationship between renewable energy and
economic growth of Bulgaria for the period 1990-2016, based on annual data, by using the Toda–Yamamoto
analysis and Autogressive Distrubuted Lag (ARDL) bound test. This period is characterized by the
democratization of the Balkans and several crisis cycles in Bulgaria. Renewable energy consumption (REC,
percentage of total final energy consumption), renewable electricity output (REO, percentage of total
electricity output) and economic growth (GDP constant 2010 US$) were used. The levels or differences of the
variables that are stationary were investigated using the augmented Dickey–Fuller (ADF), Philips–Perron
(PP) and Kwiatkowski-Philips-Schmidt-Shin (KPSS) unit root tests.
Findings – Three different results were obtained from this study. One showed that renewable energy
consumption and renewable electricity output are the causes of economic growth. Another result of this study
is that economic growth and renewable electricity output are the causes of renewable energy consumption.
The last result is that economic growth and renewable energy consumption are not causes of renewable
electricity output. There was no long-term relationship between variables.
Research limitations/implications – The ARDL and Toda–Yamamoto tests were used because of
lack of data sets. Thus, it is estimated that there is no long-term relationship.
Originality/value – This study is an original work for Bulgaria, showing the results of the relationship
between renewable energy and economic growth. In line with the results of this study, renewable energy
projects related to Bulgaria can be predicted.
Keywords Economic growth, Renewable energy, ARDL bound test, Bound tests,
Macroeconomic influence, Toda-Yamamoto causality analysis
Paper type Research paper
Variability in energy economies leads to more frequent use of renewable energy
technologies in the context of sustainable energy. Renewable energy production and
consumption enable the development of new technologies. Therefore, there are many new
opportunities for investors and the economy in general. Renewable energy production and
consumption have a multiplier effect not only on the energy sector but also on all supporting
activities related to this sector. The aim of this study is to analyze the effects of renewable
energy production and consumption on economic growth. Besides, the positive effects of
these technologies in the economy are investigated. Unlike previous renewable energy-
Renewable
energy
consumption
573
Received 13 November 2017
Revised 14 August 2018
19 November 2018
Accepted 20 November 2018
International Journal of Energy
Sector Management
Vol. 13 No. 3, 2019
pp. 573-
589
© EmeraldPublishingLimited
1750-6220
DOI 10.1108/IJESM-11-2017-0005
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1750-6220.htm
http://dx.doi.org/10.1108/IJESM-11-2017-0005
growth studies, this is the first study that examines the relationship between renewable
energy and economic growth of Bulgaria by using the Toda–Yamamoto analysis and ARDL
bound test, which indicates that there is positive long-run causality running from renewable
energy to real GDP. The used data set involves the Republic of Bulgaria’s national statistical
data for the period from 1990 to 2016. In this period, Bulgaria experienced a collapse of the
socialist system. Then, the use of modern energy has been taken into consideration. This
process has been accelerated by the moments of specific crises that change and reduce the
expectations of sustainable growth. This means that democratization of the Balkans and
some crisis cycles in Bulgaria played important roles in the shift to renewable energy
technologies. The data used in econometric analysis have been taken on the website of the
World Bank. Only the natural logarithm of the GDP constant 2010 US$ variable has been
taken from the variables used in the ecometric analysis. Other variables are based on the
ratio of renewable energy consumption (REC, percentage of total final energy consumption)
and renewable electricity output (REO, percentage of total electricity output).
1. Introduction
Consumption continues to increase with the needs of countries with high levels of
development. It is estimated that this increase will continue in line with predictions. Along
with the necessity of meeting the consumption needs of countries and adapting to
technological developments, energy consumption should increase. While energy
consumption is met by a high proportion of fossil energy sources, meeting energy
requirements at low rates from renewable energy sources has increased future concerns
within the scope of sustainable energy. Moreover, the studies on energy economy and the
country’s energy policies are progressing in this direction. Uncertainty of fossil resources,
dependence on countries’ imports, political crises and the negative impact of fossil resources
on the environment constitute the bases of concerns. According to scientific studies, the
findings of positive effects of the use of renewable energy on the economy and the
environment are increasing day by day. Besides, according to the policies after the oil crisis,
distrust on access to fossil fuels has raised the issue of energy diversity, and the dependence
on fossil resources has been sought to be avoided. A great part of the solution proposals has
the tendency to shift to renewable and cheaper energy sources. In the context of sustainable
energy, the use of renewable energy sources is essential for meeting the energy needs of
Bulgaria and other countries.
According to the literature, the effects of renewable energy consumption and production
on economic growth have been examined for Bulgaria. Through literature research, it was
observed that there was no study examining the relationship between the variables used in
the study in this period for Bulgaria. The hypothesis that “the increase in the production and
consumption of renewable energy in Bulgaria can lead to a sustainable growth in the
Bulgarian economy” was analyzed in this study.
The characteristics of the study that may be considered as different from those of the
studies examining the same subject can be stated as follows:
� There is no such empirical work for the case of Bulgaria to study the relationship
between renewable energy and economic growth. The study is aimed to eliminate
the lack of literature in this direction.
� Along with reducing negative effects of fossil fuels and the dependence on them, we
predicted how the influence on the economy of the country would be equal to the
transition in renewable energy use.
IJESM
13,3
574
� To raise awareness about Bulgaria’s policies toward renewable energy, in line with
our results for Bulgaria, we aimed to raise awareness about the energy strategies
and investments for both Bulgaria and other countries.
� Determining the importance of mitigating fossil resource dependence practices and
renewable energy strategies to minimize the adverse effects of possible crises is
important for energy economies. Emphasizing the importance of renewable energies
in energy resources for countries, we aimed to create a perception in terms of the
extent of investments to be made.
2. Literature review
In recent years, there have been factors that need to be considered in the studies
investigating the relationship between energy consumption and economic growth. As each
country has different domestic energy sources, different political regulations, different
institutional arrangements, different cultures and different energy policies, it may be
inappropriate to make general judgments when it comes to analyzing this relationship.
When these methodologies are applied to countries with different economic backgrounds,
conflicting results may be obtained. Energy prices are predicted to have a direct impact on
the economic growth in countries without sufficient energy resources. In this context, the
relationship between renewable energy and economic growth constitutes the theme of many
literature studies within the scope of sustainable energy supply. In this context, the
relationship of energy and growth within the scope of sustainable energy supply has been
the subject of many scientific studies. Some of the empirical studies examining the
relationship between renewable energy consumption and economic growth (GDP) are as
follows.
Some of the literature studies have found that there is a long-term relationship between
economic growth and energy consumption. For example, Al-mulali et al. (2013) used the data
from 1980 to 2009 for 108 countries, examining the relationship between renewable energy
consumption and growth. According to the study results, there was a bi-directional
relationship between renewable energy consumption and growth for 85 countries. For 21
countries, there was no relationship between renewable energy and growth. For two
countries, there was a one-way relationship from growth to renewable energy consumption.
In total, 79 per cent of the countries had a positive, bi-directional long-run relationship
between renewable energy consumption and GDP growth. Shafiei and Salim (2014), using
the data from 1980 to 2011 for 29 OECD countries, investigated the impacts of renewable
and non-renewable energy consumption on economic growth and CO2 emissions. According
to the results, there was two-way causality between economic growth and both renewable
and non-renewable energy consumption. Hassine and Harrathi (2017), from 1980 to 2012 for
the Gulf Cooperation Council (GCC) countries, reported a causal relationship between
renewable energy consumption, real GDP, trade and financial development. It is estimated
that renewable energy consumption, exports and private sector credit have significant
impacts on output. Furthermore, it is foreseen that the renewable energy use and exports
may increase the economic growth of GCC countries.
Fotourehchi (2017), using the data from 1990 to 2012 for 42 developing countries,
analyzed the relationship between renewable energy consumption and economic growth.
According to the results, there was one-way causality between renewable energy
consumption and economic growth. Khobai and Le Roux (2018), using the data from 1990 to
2014 for South Africa, investigated the causal relationship between renewable energy
consumption and economic growth. According to the results, there is a long-term
Renewable
energy
consumption
575
relationship between the variables. The long-run results show that a one-way causality
relationship from renewable energy consumption to economic growth has been identified,
and the short-run results show that there was one-way causality from economic growth to
renewable energy consumption.
Kraft and Kraft (1978) survey has guided many studies examining the relationship
between growth and energy consumption (Belloumi, 2009). Papiez and Smiech (2013), using
the data from 1993 to 2011 for post-communist countries, investigated the relationship
between energy consumption and economic growth. According to the results, there was a
linkage between energy consumption and economic growth in four out of nine countries.
The growth hypothesis was positive for three countries: Bulgaria, Poland and Romania. A
special situation of Poland and Bulgaria – countries confirming the growth hypothesis –
should be mentioned. They rely on coal as the most important source of energy. �Smiech S.
and Papie_z M. (2014) used the data from 1993 to 2011 for European Union (EU) member
states for investigating the relationship between energy consumption and economic growth.
The results indicate that the group of countries with the highest reduction of energy
intensity and share of renewable energy in total energy consumption had casual relations.
There was no causality relationship between variables in 17 countries. For Latvia and
Bulgaria, it was stated that there was a bi-lateral causality relationship between the
variables. Sasana and Ghozali (2017), using the data from 1995 to 2014 for five BRICS
countries, analyzed the effect of the consumption of fossil fuels and renewable energy on the
economic growth. The results show that the consumption of fossil energy had a positive
effect on economic growth, while renewable energy consumption had a negative effect.
Other studies examining the relationship between renewable energy and economic
growth are given below.
Benavides et al. (2017) used data from 1970 to 2012 for Austria, investigating short- and
long-term relationships between CH4 emissions, economic growth, electricity production
from renewable sources and trade openness. According to the results of the long-term
Granger causality, one-way causes were found between CH4 and related variables. Ohler
and Fetters (2014) used the data from 1990 to 2008 for 20 OECD countries, investigating the
relationship between economic growth and electricity generation from renewable sources.
The findings indicate that energy conservation policies would have a positive effect on GDP
in certain circumstances. Aguirre and Ibikunle (2014), using the data from 1990 to 2010 for a
broader sample size of countries, worked on renewable energy growth. According to the
results of the study, it was suggested that weak voluntary approaches may have negative
effects on the growth of renewables to meet public demand.
Apergis and Payne (2014), using the data from 1980 to 2010 for seven Central American
countries, investigated that renewable energy consumption and CO2 emissions are
cointegrated. Zeb et al. (2014), using the data from 1975 to 2010 for Bangladesh, India, Nepal,
Pakistan and Sri Lanka, investigated the relationship among electricity production from
renewable sources, GDP, CO2 emissions, natural resource depletion and poverty. Findings
show that there is two-way Granger causality between CO2 emissions and natural resource
depletion in Nepal and between energy production and poverty in Pakistan. Also, the
results
pointed at Granger causality from energy production to poverty in Bangladesh and India
and from poverty to energy production in Sri Lanka. Fuinhas and Marques (2012), using the
data from 1990 to 2007 for 24 European countries, stated that the high costs of promoting
renewables are probably being placed excessively upon the economy, namely, by increasing
electricity tariffs, thus inducing deceleration in economic activity. Menyah and Wolde-
Rufael (2010a), using the data from 1960 to 2007 for the USA, investigated the relationship
between renewable and nuclear energy consumption and between real GDP and CO2
IJESM
13,3
576
emissions. According to the study, renewable energy consumption does not contribute
significantly to emission reduction. Menyah and Wolde-Rufael (2010b), using data from
1960 to 2007, in the study of the relationship between CO2 emissions, renewable and nuclear
energy consumption and real GDP, found that renewable energy consumption has made a
significant contribution to emission reduction. Chien and Hu (2008), using the data of 2003
for 116 economies, stated that there was a positive relationship between renewable energy
and GDP. Chien and Hu (2007), using the data from 2001 to 2002 for 45 economies, suggested
that increasing the use of renewable energy improves an economy’s technical efficiency.
The relationship between economic growth and renewable energy consumption is the
focus of recent research in terms of the fight against sustainable energy and environmental
pollution. Determining the causality relationship between two variables is of high importance
for policymakers. It is estimated that renewable energy has a direct or indirect effect on
sustainability (Inglesi-Lotz, 2016; Stigllitz, 2002). Moreover, renewable energy can be seen as
the main factor in overcoming the obstacles to sustainable development (energy price shocks,
fossil energy, energy supply security, environmental pollution issues and so on).
Recently, most of the scientific studies have been about renewable energy, such as Payne
(2009), Menyah and Wolde-Rufael (2010a, 2010b), Shahbaz et al. (2012) and Pao and Fu
(2013). Most of the previous empirical studies investigating the effectiveness of renewable
energy policies have focused on policies to promote renewable energy production and
energy investments.
The studies that detect a bi-directional relationship between renewable energy
consumption and economic growth are Apergis and Payne (2010a, 2010b), Apergis and
Payne (2011a,b), Apergis and Payne (2012), Pao and Fu (2013), Al-mulali et al. (2014), Sebri
and Ben-Salha (2014), Lin and Moubarak (2014), Kahia et al., 2016; Saidi and Ben Mbarek
(2016), Amri (2016) and Paramati et al. (2017). Bloch et al. (2015), using the data from 1977 to
2013 for China, investigated the relationship between production and consumption of three
energy sources (coal, oil and renewable energy). Findings show that economic growth
causes the consumption of coal, oil and renewables.
The studies that determine the causality relationship from renewable energy to economic
growth are given below.
Tiwari (2011), using the data from 1960 to 2009 for India, investigated the relationship
between renewable energy and GDP. The findings suggest that a positive shock in
renewable energy source increases GDP. Bobinaite et al. (2011), using the data from 1990 to
2009 for Lithuania, investigated that in the short run, there was one-way causality running
from renewable energy sources gross inland consumption to real GDP. According to the
results, increased consumption of RES will have a positive effect on Lithuania’s real GDP.
Ibrahiem (2015), using the data from 1980 to 2011 for Egypt, investigated the relationship
between renewable electricity consumption, foreign direct investment and economic growth.
According to the results, renewable electricity consumption had a positive effect on
economic growth in the long run and there was two-way causality relationship between
economic growth and renewable electricity consumption. Hamit-Haggar (2016), using the
data from 1971 to 2007 for 11 sub-Saharan African countries, examined the relationship
between clean energy consumption and economic growth. According to the results, it was
stated that there is one-way Granger causality relationship from clean energy consumption
to economic growth. Bhattacharya et al. (2016) investigated the relationship between
renewable energy consumption and economic growth by using data from 1991 to 2012 for 38
countries with high renewable energy consumption. They stated that renewable energy had
a significant impact on economic growth, and that governments and other related
installations and organizations should act together in determining renewable energy
Renewable
energy
consumption
577
policies. According to the research results, there is a one-way relationship from renewable
energy consumption to economic growth for some countries and the long-term increase in
renewable energy consumption has a significant impact on economic output. Naseri et al.
(2016), using the data from 1990 to 2012 for OECD countries, investigated the impact of
renewable energy intermediate consumption on economic growth. Findings show that an
increase in renewable energy consumption leads to increase in economic growth. Bro_zyna
et al. (2017), using the data from 2004 to 2014 for 28 EU countries, investigated whether and
to what extent economic development stimulates the production of energy from renewable
sources. Findings showed that in each group, the production of energy from renewable
sources rose and the production of greenhouse gases fell over the 10-year period. This was
achievable despite the economic crisis. The EU directive from 2009 did not clearly impact
investment in renewable energy in comparison with the five years preceding it. This should
be interpreted positively as it shows investor awareness in this area in different countries,
regardless of top-down regulation. Kahia et al. (2017), using the data from 1980 to 2012 for 24
economies, investigated the effects of renewable energy policies on economic growth in
MENA countries. According to the study, the treatment effect of renewable energy policies
has a significant positive impact on economic growth. Yazdi and Shakouri (2017), using the
data from 1979 to 2014 for Iran, examined the relationship between economic growth,
renewable energy consumption, energy consumption, financial development and trade.
According to the results, there was one-way causality from renewable energy consumption
to economic growth. Taghvaee et al. (2017), using the data from 1981 to 2012 for Iran,
estimated the nexus between economic growth and renewable energy. According to the
results, renewable energy consumption was an insignificant driver to economic growth.
Governors should promote this kind of energy to assign a large part of total energy
consumption to it. Dees and Auktor (2018) estimated that investing in renewables is
beneficial for MENA countries, and that this could be an incentive to intensify the existing
policy toward renewables in the region. To improve the positive economic impact of growth,
policymakers should support local manufacturing and service provision associated with
renewable energy sources. Amri (2017a), using the data from 1980 to 2012 for Algeria,
investigated the relationship between renewable and non-renewable energy consumption,
growth and capital.
According to the findings of this study, there is a one-way relationship from renewable
energy consumption to economic growth. Amri (2017b), using the data from 1990 to 2012 for
72 countries, investigated the relationship between renewable energy consumption, growth
and foreign trade. The results showed that there was a bi-directional relationship between
renewable energy consumption and growth. Bekareva et al. (2017), using the numerical data
from 2000 to 2014, investigated the relationship between renewable energy and economic
growth. According to the results, for the USA and some other states, renewable energy is an
important part of economic growth. Brini et al. (2017), using the data from 1980 to 2011 for
Tunisia, examined the relationship between renewable energy consumption, growth, foreign
trade and oil prices. According to the study, there is a one-way relationship from renewable
energy consumption to growth. Ito (2017), using the data from 2002 to 2011 for 42 developed
countries, examined the relationship between renewable energy consumption, growth, non-
renewable energy consumption and CO2 emission. According to the result, there is a one-
way relationship from renewable energy consumption to growth. Taher (2017), using the
data from 1990 to 2012 for Lebanon, investigated the impact of renewable energy
consumption on economic growth. Results showed a statistically significant impact of
renewable energy consumption on the Lebanese economic growth. Magazzino (2017), using
the data from 1970 to 2007 for Italy, investigated the renewable energy consumption –
IJESM
13,3
578
economic growth nexus. According to the Toda–Yamamoto results, there was a one-way
relationship from renewable energy consumption to aggregate income. Soava et al. (2018),
using the data 1995 to 2015 for 28 countries of the EU, investigated the relationship between
economic growth and renewable energy consumption. Findings showed that there was a
positive impact of renewable energy consumption on economic growth.
The studies that found that economic growth has an effect on renewable energy are
given below.
Sadorsky (2009a), using the data from 1994 to 2003 for 18 emerging countries,
investigated the relationship between real income per capita and renewable energy
consumption per capita. According the findings, real income per capita and renewable energy
consumption per capita had a positive effect for emerging economies. There was a one-way
causality relationship from growth to energy consumption. Burakov and Freidin (2017),
using the data from 1990 to 2014 for Russia, investigated the causal relationship between
renewable energy consumption, economic growth and financial development. The results of
the Granger causality test show that there was two-way causality between economic growth
and financial development. It is also found that renewable energy consumption does not
cause Granger causality on economic growth or financial development.
There is also a study that found no causal link between renewable energy consumption
and economic growth.
Bélaïd and Youssef (2017), using the data from 1980 to 2012 for Algeria, investigated the
relationship between renewable energy consumption, growth, non-renewable energy
consumption and CO2 emission. According to the results of the study, there was no
relationship between renewable energy consumption and growth. Ben Jebli and Ben Youssef
(2015), using the data from 1980 to 2010 for 69 countries, investigated the relationship
between renewable energy consumption, growth, non-renewable energy consumption, labor
force, capital and foreign trade. According to the results of the study, there was no
relationship between renewable energy consumption and growth. Menegaki (2011), using
the data from 1997 to 2007 for 27 European countries, investigated that empirical results do
not confirm causality between renewable energy consumption and GDP.
In this study, we aimed to contribute to the literature by analyzing and strategies of
Bulgaria by examining the renewable energy–economic growth relationship of the country.
At the same time, this study is an example of studies that examine the same subject with
similar and same methods.
3. Methodology and data
This study investigates the relationship between renewable energy and economic growth
for the Bulgarian economy. Annual data from 1990 to 2016 were obtained from the website
of the World Bank. During this period, the democratization of the Balkans and several crisis
cycles in Bulgaria were also considered. REC percentage of total final energy consumption,
REO percentage of total electricity output and economic growth (GDP constant 2010 US$)
were used. Only the natural logarithm of the GDP in constant prices in US dollars is
considered in this study. Other variables are based on the ratio of REC percentage of total
final energy consumption and REO percentage of total electricity output.
It is important to check whether the series are stationary before examining the long-term
relationship in the series. Many unit root tests are available to investigate the stationarity of
the series to determine the existence of regression problems. The levels or differences of the
variables that are stationary are investigated by augmented Dickey–Fuller (ADF), Philips–
Perron (PP) and KPSS unit root tests. However, there might be a structural break in the
global finance crisis in the related period. Therefore, the structural fracture unit root test
Renewable
energy
consumption
579
was also used in the study. The existence of a long-term relationship between variables in
the study was explored by the ARDL bound test. In this cointegration test, some variables
can be used in level values (I[0]) where some variables are stationary in the first difference
(I[1]). In addition to this, other cointegration techniques are sensitive to the periods of the
sample. In this approach, the sample periods are not a problem even if they are short (Harris
and Sollis, 2003, p. 152).
In this context, Pesaran and Pesaran (1999) firstly investigate the existence of a long-term
relationship with the boundary test. We can write an ARDL bound model as follows:
DGDPt5d 0 þ d 1GDPt�i þ d 2RECt�i þ d 3REOt�i þ d 4DGDPt þ d 5DRECt
þ d 6DREOt þ
Xp�1
i51
u ajGDPt�i þ
Xq�1
i51
u bjRECt�i
þ
Xs�1
i51
u cjREOt�i þ « t (1)
The cointegration relation is made by testing the hypothesis (H0: d 1 = d 2 = d 3 = 0).
The F statistic is calculated for any significance level that is possible to make a definite
interpretation without considering the integration scores of the variables (Pesaran
et al., 2001). If the calculated F statistic exceeds the critical value upper limit, the null
hypothesis is rejected and it is decided that there is a long-run relationship between the
variables; otherwise, there is no long-run relationship between the variables.
Furthermore, if the F statistic remains between the upper and lower limits, there is no
definite interpretation as to whether a cointegration relation exists between the
variables (Balcılar et al. (2014: 455).
In the Toda and Yamamoto (1995) causality analysis, the optimum lags length (k) is
determined by Akaike information criteria (AIC) and Schwatz information criteria (SCI) after
the maximum integration level (DMAx) of the series is determined using ADF and PP unit
root tests. Finally, the VAR model (k þ dmax) is estimated by the number of lags and
seemingly unrelated regression method to decide causality relation and direction. The Toda
and Yamamoto (1995) equations for the study are as follows:
GDPt ¼ d 0 þ
Xk
i51
d 1iGDPt�i
þ
Xdmax
j5kþ1
d 2iGDPt�j þ
Xk
i51
b 1iREOt�i
Xdmax
i51
b 2iREOt�j
þ
Xk
i51
b 3iRECt�i þ
Xdmax
j5kþ1
b 4iRECt�j þ v1t (2)
REOt ¼ u 0 þ
Xk
i51
u 1iREOt�i þ
Xdmax
j5kþ1
u 2iREOt�j þ
Xk
i51
a1iGDPt�i
Xdmax
j5kþ1
a2iGDPt�j
þ
Xk
i51
a3iRECt�i þ
Xdmax
j5kþ1
a4iRECt�j þ v2t (3)
IJESM
13,3
580
RECt ¼ v0 þ
Xk
i51
v1iRECt�i þ
Xdmax
j5kþ1
v2iRECt�j þ
Xk
i51
g1iGDPt�i
þ
Xdmax
j5kþ1
g2iGDPt�j þ
Xk
i51
g3iREOt�i þ
Xdmax
j5kþ1
g4iREOt�j þ v3t
(4)
The causality relation and direction Vector Autoregressive (VAR) (k þ dmax) model is
estimated and the first k of the model coefficients are determined by applying the modified
Wald (MWALD) test. In the causality relation, the null hypothesis is that k independent
variables are equal to zero in the group, and the alternative hypothesis is that k independent
variables are not equal to zero as a group. If the MWALD statistic is significant, the null
hypothesis is rejected and the alternative hypothesis is accepted. The acceptance of the
alternative hypothesis implies that there is a causal relationship from the independent to
dependent variables (Dritsaki, 201: 120-129).
4. Empirical analysis results
In the study, first, descriptive statistics related to variables were given. Relevant findings
are given in Table I.
According to the results of Table I, it can be said that all variables have normal
distribution. To obtain reliable results in the study, the levels/differences of the variables
were examined. For this purpose, ADF, PP and KPSS tests were used in this study. The
obtained unit root test results are presented in Table II.
According to Table II, ADF, PP and KPSS unit root test results are contradicted. The
variable that is stationary according to the KPSS test is not stationary at the level value
according to the PP and ADF tests. The conflicting results mean that a structural fracture
unit root test is needed. In such a case, it is necessary to use unit root tests that take into
account the possibility of breakage. We analyzed with breakpoint tests owing to possible
fluctuations in the Bulgarian economy. Perron (1997) used unit root tests to take structural
breaks into account, and the results are reported in Table III.
It can be seen from Table III that all variables are stationary in the first difference.
Following the unit root analysis, the relationship between the variables in the study was
explored by the ARDL boundary test approach.
Table I.
Descriptive stats
GDP REC REO
Mean 8.519012 8.614995 8.170846
Median 8.448206 8.801903 6.994762
Maximum 8.937484 18.15702 17.98859
Minimum 8.183916 1.916849 3.849684
Std. Dev. 0.279915 5.11492 3.74
582
1
Skewness 0.216768 0.365714 1.304164
Kurtosis 1.343287 2.072275 3.831781
Jarque–Bera 3.177038 1.511965 8.119842
Probability 0.204228 0.469549 0.017250
Observations 26 26 26
Renewable
energy
consumption
581
In this test, the appropriate number of lags must first be determined for an unrestricted error
correction model. The diagnostic assumptions were taken into account when determining
the appropriate number of lags. In this context, no autocorrelation or heterodasticity
problem is detected. The appropriate number of lags provided by the normal distribution
assumption is 1. The findings are reported in Table IV.
To be able to decide the existence of a long-term relationship between the variables, the
results in Table IV need to be compared with the critical values of Pesaran and Pesaran
(1997). Critical values are given in Table IV. It can be argued that there is no long-run
relationship between the variables because the F statistic value in the study is below the
critical values for 1, 5 and 10 per cent significance levels. In short, there is no long-term
relationship between REC, REO and GDP. There are studies that find similar results (Bélaïd
and Youssef, 2017; Ben Jebli and Ben Youssef (2015; Menegaki, 2011).
Finally, the existence of causal link between variables is investigated using the Toda–
Yamamoto causality test. The results obtained are reported in Table V. Time graphs related
to these variables are also given in the study. The notation is in Figure 1.
As shown in Table V, three different results were obtained in the study. The findings
show that REC and REO are the causes of GDP. When this result is evaluated together with
Table III.
Perron (1997)
Breakpoint unit root
test
Level
Model A Model B Model C
Variables t-Statistics p-value
Break
point t-Statistics p-value
Break
point t-Statistics p-value
Break
point
GDP �4.1996 0.1877 2003 �3.7056 0.2996 2011 �4.0762 0.4729 1998
REC �3.3114 0.7969 2009 �3.4236 0.4508 2009 �4.2514 0.3636 2006
REO �4.0927 0.3164 2012 �4.0927 0.3164 2012 �4.6917 0.1601 2010
First Difference
GDP �5.8271 0.001*** 2008 �5.3977 0.001*** 2008 �5.0438 0.0711* 2008
REC �6.7242 0.001*** 2013 �6.0962 0.001*** 2014 �6.7125 0.001*** 2007
REO �5.6924 0.001*** 1999 �5.9231 0.001*** 2009 �6.2464 0.001*** 2010
Note: The symbols *, ** and *** denote significance at 0.10, 0.05 and 0.01 levels, respectively
Table II.
ADF, PP and KPSS
unit root tests
Level
ADF PP KPSS
Variables t-Statistics p-value t-Statistics p-value t-Statistics
For constant
GDP 0.8035 0.9920 0.3708 0.9774 0.6914
REC 0.1792 0.9655 07336 0.9905 0.7300
REO �0.3214 0.9083 0.3057 0.9738 0.6265
For constant and trend
GDP �1.7721 0.6849 �3.1184 0.1238 0.1346
REC �2.5690 0.2969 �2.5520 0.3029 0.1118
REO �1.8852 0.6350 �1.7410 0.7023 0.1692
Notes: Critical value at 0.05 significant level in KPSS test is 0.4630 for unit root test with constant term and
0.1180 for unit root test with constant term and trend. The symbols *, ** and *** denote significance at
0.10, 0.05 and 0.01 levels, respectively
IJESM
13,3
582
Figure 1, it should be emphasized that REC and REO increase more than GDP. This result
shows that in the Bulgarian economy, renewable energy and electricity output are expected
to have a positive impact on economic growth. It is obviously aimed at increasing the
economic growth of countries by investing in renewable energy sources in developed
countries. According to these findings, it can be said that for the Bulgarian economy,
policymakers must invest in renewable energy sources. As seen in Figure 1, the use of
renewable energy and the increase in the output of renewable electricity are triggering an
upward trend on economic growth. The increase in renewable energy consumption leads to
an increase in economic growth. It supports the purpose of this work. It can be said that
renewable energy policies play an important role in the country’s strategies, and that a large
part of the investments should be shaped in this direction.
Another conclusion that should be drawn from this study is that GDP and REO are the
causes of REC. It is a fact that the economic growth and renewable electricity output causes
energy consumption. The energy consumption of an energy-intensive country is the
expected result, and this result is in line with the findings of Wolde-Rufael (2004, 2005, 2006,
Table V.
Toda–Yamamoto
causality analysis
results
Hypotheses MWald statistics Probability value Decision
REC and REO do not cause GDP
H0:b 1=b 2=b 3=b 4=0 2.5331 0.0735* Rejection
H0:b 1=b 2= 0 2.4930 0.0763* Rejection
H0: b 3=b 4=0 4.2186 0.0147** Rejection
GDP and REC do not cause REO
H0: a1=a2=a3=a4=0 2.3937 0.0911* Rejection
H0: a1=a2 = 0 1.9240 0.1765 Acceptance
H0: a3 = a4=0 1.5312 0.2446 Acceptance
GDP and REO do not cause REC
H0:g1=g2=g3=g4=0 3.4305 0.0314** Rejection
H0:g1=g2 = 0 3.8121 0.0429** Rejection
H0:g3=g4 = 0 4.9921 0.0197** Rejection
Notes: Jarque–Bera test: 0.7566 Breusch–Godfrey LM test: 2.7455 ** Breusch–Pagan Godfrey test: 1.6704;
Jarque–Bera test: 2.7199 Breusch–Godfrey LM test: 1.2976 Breusch–Pagan Godfrey test: 1.1981; Jarque–
Bera test: 0.4987 Breusch–Godfrey LM test: 1.1950 Breusch–Pagan Godfrey test: 1.8010; The symbols *, **
and *** denote significance at 0.10, 0.05 and 0.01 levels, respectively
Table IV.
Bound test results for
Bulgaria
Unrestricted intercept and no trend case
lags F statistic t statistic
1 2.4566 �2.3737
Pesaran Critical Value
%10 %5 %1
d I(0) I(1) I(0) I(1) I(0) I(1)
2 3.182 4.126 3.793 4.855 5.288 6.309
Notes: Critical boundary values are taken from Table F, Case II, of Pesaran and Pesaran (1997 p. 478); d is
independent variable number
Renewable
energy
consumption
583
2009), Lee (2006) and Squalli (2007). Thus, the country can increase its income by setting its
own taxation and tariffs in this direction. This revenue can be transferred to another project
in need.
The last result is the study that GDP and REC are not the causes of REO. As can be seen
in Figure 1, the consumption of renewable energy in the Bulgarian economy has increased in
the past decade. The subject matter obtained in the study should be evaluated in this
direction. In summary, the results indicate that investment in renewable energy sources in
the Bulgarian economy should be increased.
According to the findings in the study, renewable energy sources contribute to
sustainable economic growth, environmental sustainability, access to basic services,
improvement in human health and income generation activities and provision economic
benefits such as new jobs and industries (IRENA, 2016). It is estimated that policy
preferences for renewables will create an environmental and economic benefit that will
eliminate the ineffective subsidies that support the development and consumption of fossil
fuels (O’Sullivan et al., 2017).
5. Conclusion
Considering that Bulgaria is a country dependent on fuels and natural gas, it can be said that
the contribution of the alternatives provided by renewable resources is quite significant. In the
future, if the cost of establishing renewable energy and its accessibility and cost are competitive
with fossil energy sources, it is foreseen that economic developments will take place in a
positive way. According to the results obtained, it can be said that the country needs to increase
renewable energy resource investments to strengthen long-term relationships between the
variables. In this context, it can be said that the effect of renewable energy consumption on
economic growth will increase if the country’s strategies and investment incentives are
continued and increased. The harmonization laws, incentive premiums and applications for
increasing renewable energy investments can have a positive impact on the country’s economy
by contributing to the energy supply security of the country. The investments to be made in the
field of renewable energy provide significant support for economic growth and development by
ensuring positive externality of these investments to increase domestic production, create more
jobs and reduce import payments. Additionally, it is estimated that with the increase in
employment and employment opportunities, the aging population and immigration rate will
Figure 1.
Graphs for GDP,
REO and REC
IJESM
13,3
584
decrease as the economy grows with sustainable renewable energy. Energy supply should be
directed to renewable energy sources instead of being dependent on imported fossil fuels. The
role of state supporters in this direction is huge. However, it is not possible to mention that
the incentives in practice are sufficient. According to the results of the study, incentives should
be increased. Particularly, the possibility of access to long-term and low-cost loans and the
promotion of industry can be among the solution politics. Attractive credits and incentives are
crucial for an investor who is willing to direct current resources and capital to renewable
energy investments.
Given the competition of renewable energy investments with fossil fuels, the current
fossil fuel incentives are affecting the market in a negative way. This makes renewable
energy costlier. These fossil energy incentives prevent the development of renewable
energy. In addition, fossil fuel incentives prevent the formation of public support for
clean energy. In this context, fossil fuel incentives for Bulgaria should not be applied in the
midterm.
As a result, we can say that the incentives provided in the field of renewable energy in
Bulgaria are limited when compared with world examples. Additional measures need to be
taken to increase the share of renewable energy investments in Bulgaria’s energy production
portfolio faster and more efficiently.
References
Aguirre, M. and Ibikunle, G. (2014), “Determinants of renewable energy growth: a global sample
analysis”, Energy Policy, Vol. 69, pp. 374-384.
Al-mulali, U., Fereidouni, H.G. and Lee, J.Y.M. (2014), “Electricity consumption from renewable
and non-renewable sources and economic growth: evidence from Latin American
countries”, Renewable and Sustainable Energy Reviews, Vol. 30, pp. 290-298.
Al-mulali, U., Fereidouni, H.G., Lee, J.Y. and Sab, C.N.B.C. (2013), “Examining the bidirectional long run
relationship between renewable energy consumption and GDP growth”, Renewable and
Sustainable Energy Reviews, Vol. 22, pp. 209-222.
Amri, F. (2016), “The relationship amongst energy consumption, foreign direct investment and output
in developed and developing countries”, Renewable and Sustainable Energy Reviews, Vol. 64,
pp. 694-702.
Amri, F. (2017a), “The relationship amongst energy consumption (renewable and nonrenewable), and
GDP in Algeria”, Renewable and Sustainable Energy Reviews, Vol. 76, pp. 62-71.
Amri, F. (2017b), “Intercourse across economic growth, trade and renewable energy consumption in
developing and developed countries”, Renewable and Sustainable Energy Reviews, Vol. 69,
pp. 527-534.
Apergis, N. and Payne, J.E. (2014), “Renewable energy, output, CO2 emissions, and fossil fuel prices in
Central America: Evidence from a non linear panel smooth transition vector error correction
model”, Energy Economics, Vol. 42, pp. 226-232.
Apergis, N. and Payne, J.E. (2010a), “Renewable energy consumption and economic growth: evidence
from a panel of OECD countries”, Energy Policy, Vol. 38 No. 1, pp. 656-660.
Apergis, N. and Payne, J.E. (2010b), “Renewable energy consumption and growth in Eurasia”, Energy
Economics, Vol. 32 No. 6, pp. 1392-1397.
Apergis, N. and Payne, J.E. (2011a), “Renewable and non-renewable electricity consumption– growth
nexus: evidence from emerging market economies”, Applied Energy, Vol. 88 No. 12,
pp. 5226-5230.
Apergis, N. and Payne, J.E. (2011b), “The renewable energy consumption–growth nexus in Central
America”, Applied Energy, Vol. 88 No. 1, pp. 343-347.
Renewable
energy
consumption
585
Apergis, N. and Payne, J.E. (2012), “Renewable and non-renewable energy consumption-growth nexus:
evidence from a panel error correction model”, Energy Economics, Vol. 34 No. 3, pp. 733-738.
Balcılar, M., Bal, N., Algan, N. and Demiral, M. (2014), “Türkiye’nin ihracat performansı: ihracat
hacminin temel belirleyicilerinin incelenmesi (1995-2012)”, Ege Akademik Bakis (Ege Academic
Review), Vol. 14 No. 3, pp. 451-462.
Bekareva, S.V., Meltenisova, E.N., A. and Gsysa, J.G. (2017), “Evaluation of the role of renewables
consumption on economic growth of the US regions”, International Journal of Energy Economics
and Policy, Vol. 7 No. 2, pp. 160-171.
Bélaïd, F. and Youssef, M. (2017), “Environmental degradation, renewable and non-renewable
electricity consumption, and economic growth: assessing the evidence from Algeria”, Energy
Policy, Vol. 102, pp. 277-287.
Belloumi, M. (2009), “Energy consumption and GDP in Tunisia: cointegration and causality analysis”,
Energy Policy, Vol. 37 No. 7, pp. 2745-2753.
Ben Jebli, M. and Ben Youssef, S. (2015), “Output, renewable and non-renewable energy consumption
and international trade: evidence from a panel of 69 countries”, Renewable Energy, Vol. 83
No. Supplement C, pp. 799-808.
Benavides, M., Ovalle, K., Torres, C. and Vinces, T. (2017), “Economic growth, renewable energy and
methane emissions: is there an environmental kuznets curve in Austria?”, International Journal
of Energy Economics and Policy, Vol. 7 No. 1, pp. 259-267.
Bhattacharya, M., Paramati, S.R., Ozturk, I. and Bhattacharya, S. (2016), “The effect of renewable
energy consumption on economic growth: Evidence from on top 38 countries”, Applied Energy,
Vol. 162, pp. 733-741.
Bloch, H., Rafiq, S. and Salim, R. (2015), “Economic growth with coal, Oil and renewable energy
consumption in China: prospects for fuel substitution”, Economic Modelling, Vol. 44, pp. 104-115.
Bobinaite, V., Juozapaviciene, A. and Konstantinaviciute, I. (2011), “Assessment of causality
relationship between renewable energy consumption and economic growth in Lithuania”,
Engineering Economics, Vol. 22 No. 5, pp. 510-518.
Brini, R., Amara, M. and Jemmali, H. (2017), “Renewable energy consumption, International trade, oil
price and economic growth inter-linkages: the case of Tunisia”, Renewable and Sustainable
Energy Reviews, Vol. 76 No. Supplement C, pp. 620-627.
Bro_zyna, J., Mentel, G. and Szetela, B. (2017), “Renewable energy and economic development in the
European Union”, Acta Polytechnica Hungarica, Vol. 14 No. 7.
Burakov, D. and Freidin, M. (2017), “Financial development, economic growth and renewable energy
consumption in Russia: a vector error correction approach”, International Journal of Energy
Economics and Policy, Vol. 7 No. 6, pp. 39-47.
Chien, T. and Hu, J.L. (2007), “Renewable energy and macroeconomic efficiency of OECD and non
OECD economies”, Energy Policy, Vol. 35 No. 7, pp. 3606-3615.
Chien, T. and Hu, J.L. (2008), “Renewable energy: an efficient mechanism to improve GDP”, Energy
Policy 2008, Vol. 36, pp. 3045-3052.
Dees, P. and Auktor, G.V. (2018), “Renewable energy and economic growth in the MENA region:
empirical evidence and policy implications”, Middle East Development Journal, pp. 1-23.
Fotourehchi, Z. (2017), “Renewable energy consumption and economic growth: a case study for developing
countries”, International Journal of Energy Economics and Policy, Vol. 7 No. 2, pp. 61-64.
Fuinhas, J.A. and Marques, A.C. (2012), “Energy consumption and economic growth nexus in Portugal,
Italy, Greece, Spain and Turkey: an ARDL bounds test approach”, Energy Economics, Vol. 34
No. 2, pp. 511-517.
Hamit-Haggar, M. (2016), “Clean energy-growth nexus in Sub-Saharan Africa: evidence from cross-
sectionally dependent heterogeneous panel with structural breaks”, Renewable and Sustainable
Energy Reviews, Vol. 57, pp. 1237-1244.
IJESM
13,3
586
Harris, R. and Sollis, R. (2003), Applied Time Series Modelling and Forecasting, Wiley, West Sussex.
Hassine, M.B. and Harrathi, N. (2017), “The causal links between economic growth, renewable energy,
financial development and foreign trade in Gulf cooperation council countries”, International
Journal of Energy Economics and Policy, Vol. 7 No. 2, pp. 76-85.
Ibrahiem, D.M. (2015), “Renewable electricity consumption, foreign direct investment and economic
growth in Egypt: an ARDL approach”, Procedia Economics and Finance, Vol. 30, pp. 313-323.
Inglesi-Lotz, R. (2016), “The impact of renewable energy consumption to economic growth: a panel data
application”, Energy Economics, Vol. 53 (2016), pp. 58-63.
IRENA (2016), Renewable Energy Benefits: Measuring the Economics, International Renewable Energy
Agency, Abu Dhabi, United Arab Emirates.
Ito, K. (2017), “CO2 emissions, renewable and nob-renewable energy consumption, and economic
growth: Evidence from panel data for developing countries”, Int. Econ, Vol. 151, pp. 1-6.
Kahia, M., Aissa, M.S. and Lanouar, C. (2017), “Renewable and non-renewable energy use-economic
growth nexus: the case of MENA net oil importing countries”, Renewable and Sustainable
Energy Reviews, Vol. 71, pp. 127-400.
Kahia, M., Ben Aïssa, M.S. and Charfeddine, L. (2016), “Impact of renewable and non-renewable energy
consumption on economic growth: new evidence from the MENA net oil exporting countries
(NOECs)”, Energy, Vol. 116 No. Part 1, pp. 102-115.
Khobai, H. and Le Roux, P. (2018), “Does renewable energy consumption drive economic growth:
evidence from granger-causality technique”, International Journal of Energy Economics and
Policy, Vol. 8 No. 2, pp. 205-212.
Kraft, J. and Kraft, A. (1978), “On the relationship between energy and GNP”, Journal of Energy and
Development, Vol. 3 No. 2, pp. 401-403.
Lee, C.C. (2006), “The causality relationship between energy consumption and GDP in G-11 countries
revisited”, Energy Policy, Vol. 34 No. 9, pp. 1086-1093.
Lin, B. and Moubarak, M. (2014), “Renewable energy consumption – Economic growth nexus for
China”, Renewable and Sustainable Energy Reviews, Vol. 40 No. Supplement C, pp. 111-117.
Magazzino, C. (2017), “Renewable energy Consumption-Economic growth nexus in Italy”, Int. J. Energy
Econ, Policy, Vol. 7, pp. 119-127.
Menegaki, A.N. (2011), “Growth and renewable energy in Europe: a random effect model with evidence
for neutrality hypothesis”, Energy Economics, Vol. 33 No. 2, pp. 257-263.
Menyah, K. and Wolde-Rufael, Y. (2010a), “Energy consumption, pollutant emissions and economic
growth in South Africa”, Energy Economics, Vol. 32 No. 6, pp. 1374-1382.
Menyah, K. and Wolde-Rufael, Y. (2010b), “CO2 emissions, nuclear energy, renewable energy and
economic growth in the US”, Energy Policy, Vol. 38 No. 6, pp. 2911-2915.
Naseri, S.F., Motamedi, S. and Ahmadian, M. (2016), “Study of mediated consumption effect of
renewable energy on economic growth of OECD countries”, Procedia Economics and Finance,
Vol. 36 No. Supplement C, pp. 502-509.
O’Sullivan, M., Overland, I. and Sandalow, D. (2017), The Geopolitics of Renewable Energy, Columbia
University and Harvard University.
Ohler, A. and Fetters, I. (2014), “The causal relationship between renewable electricity generation and
GDP growth: a study of energy sources”, Energy Economics, Vol. 43, pp. 125-139, doi: 10.1016/j.
eneco.2014.02.009.
Pao, H.T. and Fu, H.C. (2013), “Renewable energy, non-renewable energy and economic growth in
Brazil”, Renewable and Sustainable Energy Reviews, Vol. 25, pp. 381-392.
Papiez, M. and Smiech, S. (2013), “Economic growth and energy consumption in post-communist
countries: a bootstrap panel granger causality analysis”, Dynamic Econometric Models, Vol. 13,
pp. 51-68.
Renewable
energy
consumption
587
http://dx.doi.org/10.1016/j.eneco.2014.02.009
http://dx.doi.org/10.1016/j.eneco.2014.02.009
Paramati, S.R., Mo, D. and Gupta, R. (2017), “The effects of stock market growth and renewable
energy use on CO2 emissions: evidence from G20 countries”, Energy Economics, Vol. 66,
pp. 360-371.
Payne, J.E. (2009), “On the dynamics of energy consumption and output in the US”, Applied Energy,
Vol. 86 No. 4, pp. 575-577.
Perron, P. (1997), “Further evidence on breaking trend functions in macroeconomic variables”, Journal
of Econometrics, Vol. 80 No. 2, pp. 355-385.
Pesaran, M.H. and Pesaran, B. (1997), Working with microfit 4.0, Interactive econometric analysis,
Oxford University Press, Oxford.
Pesaran, M.H. and Pesaran, B. (1999), “An autoregressive distributed lag modelling approach to
cointegration analysis”, in S., Storm (Ed.), Econometrics and Economic Theory in the 20th
Century. The Ragnar Frisch Centennial Syposium, Chapter 11, Cambridge Univ. Press,
Cambridge.
Pesaran, M.H., Shin, Y. and Smith, R.J. (2001), “Bounds testing approaches to the analysis of level
relationships”, Journal of Applied Econometrics, Vol. 16 No. 3, pp. 289-326.
Sadorsky, P. (2009), “Renewable energy consumption and income in emerging economies”, Energy
Policy 2009, Vol. 37, pp. 4021-4028.
Saidi, K. and Ben Mbarek, M. (2016), “Nuclear energy, renewable energy, CO2 emissions, and economic
growth for nine developed countries: evidence from panel granger causality tests”, Progress in
Nuclear Energy, Vol. 88 No. Supplement C, pp. 364-374.
Sasana, H. and Ghozali, I. (2017), “The impact of fossil and renewable energy consumption on the
economic growth in Brazil, Russia, India, China and South Africa”, International Journal of
Energy Economics and Policy, Vol. 7 No. 3, pp. 194-200.
Sebri, M. and Ben-Salha, O. (2014), “On the causal dynamics between economic growth, renewable
energy consumption, CO2 emissions and trade openness: fresh evidence from BRICS countries”,
Renew Sustain Energy Rev, Vol. 39, pp. 14-23.
Shafiei, S. and Salim, R.A. (2014), “Non-renewable and renewable energy consumption and CO2
emissions in OECD countries: a comparative analysis”, Energy Policy, Vol. 66, pp. 547-556.
Shahbaz, M., Zeshan, M. and Afza, T. (2012), “Is energy consumption effective to spur economic growth
in Pakistan? New evidence from bounds test to level relationships and granger causality tests”,
Econ Model, Vol. 29 No. 6, pp. 2310-2319.
�Smiech, S. and Papie_z, M. (2014), “Energy consumption and economic growth in the light of meeting the
targets of energy policy in the EU: the bootstrap panel granger causality approach”, Energy
Policy, Vol. 71, pp. 118-129.
Soava, G., Mehedintu, A., Sterpu, M. and Raduteanu, M. (2018), “Impact of renewable energy
consumption on economic growth: evidence from European Union countries”, Technol. Econ.
Dev. Econ, Vol. 24, pp. 1197-1215.
Squalli, J. (2007), “Electricity consumption and economic growth: bounds and causality analyses for
OPEC members”, Energy Econ, Vol. 29 No. 6, pp. 1192-1205.
Stigllitz, J. (2002), Globalization and Its Discontents, Penguin Books Ltd, London, UK.
Taghvaee, V.M., Shirazi, J.K., Boutabba, M.A. and Aloo, A.S. (2017), “Economic growth and renewable
energy in Iran”, Iran Economic Review, Vol. 21 No. 4, pp. 789-808.
Taher, H. (2017), “Renewable energy consumption impact on the Lebanese economy”, International
Journal of Energy Economics and Policy, Vol. 7 No. 4, pp. 144-148.
Tiwari, A.K. (2011), “A structural VAR analysis of renewable energy consumption, real GDP and CO2
emissions: evidence from India”, Econ. Bull, Vol. 31 No. 2, pp. 1793-1806.
Toda, H.Y. and Yamamoto, T. (1995), “Statistical inference in vector autoregressions with possibly
integrated processes”, Journal of Econometrics, Vol. 66, March-April 1995, Nos 1/2, pp. 225-250.
IJESM
13,3
588
Wolde-Rufael, Y. (2004), “Disaggregated energy consumption and GDP, the case of shanghai,
1952-1999”, Energy Econ, Vol. 26 No. 1, pp. 69-75.
Wolde-Rufael, Y. (2005), “Energy demand and economic growth: the African experience”, Journal of
Policy Modeling, Vol. 27 No. 8, pp. 891-903.
Wolde-Rufael, Y. (2006), “Electricity consumption and economic growth: a time series experience for 17
African countries”, Energy Policy, Vol. 34 No. 10, pp. 1106-1114.
Wolde-Rufael, Y. (2009), “Energy consumption and economic growth: the experience of African
countries revisited. Energy”, Econ, Vol. 31, pp. 217-224.
Yazdi, S.K. and Shakouri, B. (2017), “Renewable energy, nonrenewable energy consumption, and
economic growth”, Energy Sources, Part B: Economics, Planning, and Policy, Vol. 12:12,
pp. 1038-1045.
Zeb, R., Salar, L., Awan, U., Zaman, K. and Shahbaz, M. (2014), “Causal links between renewable
energy, environment degradation, and economic growth in selected SAARC countries: progress
towards green economy”, Renewable Energy, Vol. 71, pp. 123-132.
Further reading
Perron, P. (1989), “The great crash, The oil price shock and the unit root hypothesis”, Econometrica,
Vol. 57 No. 6, pp. 1361-1401.
Phillips, P.C.B. and Perron, P. (1988), “Testing for a unit root in time series regression”, Biometrika,
Vol. 75 No. 2, pp. 335-346.
Corresponding author
Hamit Can can be contacted at: hamitcan88@hotmail.com
For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
Renewable
energy
consumption
589
View publication statsView publication stats
mailto:Further readingPerron, P. (1989), &hx201C;The great crash, The oil price shock and the unit root hypothesis&hx201D;, Econometrica, Vol.57 No.6, pp.1361-1401.Phillips, P.C.B. and Perron, P. (1988), &hx201C;Testing for a unit root in time series regression&hx201D;, Biometrika, Vol.75 No.2, pp.335-346.
https://www.researchgate.net/publication/330116952
1. Introduction
2. Literature review
3. Methodology and data
4. Empirical analysis results
5. Conclusion
References
Received: September 21, 2018; Revised: January 20, 2019; Accepted: February 18, 20
19
1, 2, 3, 4Universitas Indonesia. Jl. Prof. Dr. Sumitro Djojohadikusumo, Kukusan, Beji, Depok, Indonesia
E-mail: 1hari.nugroho@yahoo.com, 2hpasay@gmail.com, 3ariedamayanti@yahoo.com, 4maddaremmeng@gmail.com
DOI: htttp://dx.doi.org/10.15408/ etk.v18i1.8242
Etikonomi
Volume 18 (1), 2019: 13 –
28
P-ISSN: 1412-8969; E-ISSN: 2461-0771
Institutions as the Main Determinant in
Economic Growth
Hari Nugroho1, N. Haidy Ahmad Pasay2, Arie Damayanti3,
Maddaremmeng A.Panennungi4
Abstract. The studies on human capital and technological progress have given incredible
insights on how countries in the world differ from one another. Yet there are more than those
two reasons to account for differences among countries. There is a third reason why a country
would differ in terms of its economic development progress, namely institutional factors.
Hence developing institutional indices would give a deeper explanation than a mere theory.
On the other hand, we can corroborate the institutional index with the general theory that
low-quality institutions will impact an economy negatively. This study seeks to broaden the
understanding of causes of economic growth by incorporating institutional index into a semi-
endogenous growth model and finds a relationship between that index with human capital
and technological progress.
Keywords: institutions, human capital, technological progress, economic growth
JEL Classification: E01, E02, O43.
Abstrak. Penelitian akan topik mengenai modal manusia dan perkembangan teknologi
telah memberi wawasan yang mendalam dan penjelasan yang baik atas perbedaan-
perbedaan yang terjadi dalam hal pertumbuhan ekonomi diantara negara-negara di dunia.
Akan tetapi ada lebih dari 2 alasan mengapa terjadi perbedaan diantara negara-negara
tersebut. Alasan ketiga mengapa terjadi perbedaan adalah adanya perbedaan dalam faktor
institusi. Sehingga dengan membangun indeks institusi, kita akan mendapat pemahaman
lebih ketimbang hanya sekedar memahami teori saja. Selain itu, kita dapat memperkuat
teori umum dengan indeks institusi, yaitu ketika suatu negara memiliki indeks institusi
yang rendah maka pertumbuhan ekonominya pun akan rendah. Penelitian ini berupaya
untuk memahami penyebab-penyebab pertumbuhan ekonomi yang dikaitkan dengan indeks
institusi dan model pertumbuhan semi endogen serta mencari hubungan antara indeks
tersebut dengan modal manusia dan perkembangan teknologi.
Kata Kunci: institusi, modal manusia, perkembangan teknologi, pertumbuhan ekonomi
How to Cite:
Nugroho, H., Pasay, N. H. A., Damayanti, A., & Panennungi, M. A. (2019). Institutions as a Main Determinant in
Economic Growth. Etikonomi: Jurnal Ekonomi. Vol. 18 (1): 13 – 28. doi: http//dx.doi.org/10.15408/etk.v18i1.8242.
Hari Nugroho. Institutions as a Main Determinant in Economic Growth
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
14
Introduction
A question of interest for most of the macroeconomists is a question related to the cause
of the difference in economic growth among countries in the world. According to Acemoglu
& Robinson (2008), economic growth is related to people’s ability to accumulate human
capital, physical capital, and technology. Acemoglu & Robinson (2008) further summarize
the causes of the difference in economic growth to only 2, namely the proximate cause and
fundamental cause. Proximate cause or the most active and dominant cause is human capital,
physical capital, and technology. While the fundamental cause is the most basic, and it will
became the major cause of the resulting proximate cause. So if we want to get a satisfactory
answer from the question at the beginning of the paragraph, then we should focus on the
fundamental cause.
The fundamental cause is the most basic factor and the cause of differences in every
country in the world. Institutions are fundamental causes that cause differences in world
economic growth (Acemoglu & Robinson, 2008). Institutions shape how society behave
and react to certain challenges in their lives. Institutions command a society to react by
creating certain rules and regulations as guidelines. These rules and regulations are sometimes
in the written formal code but most of the times they need not be written. Institutions are
reflections of society and the people who live in therefore different countries possess different
institutions. Therefore institutions are fundamental to every country as they dictate paths to
where a country might progress. Whether it is on the right path or on the wrong one, clearly
rely on institutions at play.
Different countries possess different qualities of institutions. Economic institutions
determine incentives and constraints for economic actors and contribute to shaping the
output of the economy. Economic institutions involve social choices in which social choice
will vary between individuals so that these social choices will lead to conflict. Those who
have a greater political advantage will ultimately win the conflict. Developed countries with
high economic growth are supported by innovation and growth-oriented institutions. So the
difference in economic growth in developed countries within developing countries lies in the
quality of the institution. Therefore optimum human capital and technological advances, as
drivers of economic growth, must go hand in hand with good institutions.
This is interesting because if we understand and know the quality and position of these
institutions from institutions in developed countries so we can map the problems and catch
up quickly with a higher rate of economic growth. By knowing our position in the global
map of institutions, we can map and list our strength and weaknesses and later improve them.
The rapid growth of the stock of knowledge and technological progress begin to flourish once
good institutions are established. Advanced economies all have good institutions in place.
Good institutions can foster the growth of the stock of knowledge and technological progress
by accommodating all possible knowledge spillover from advanced and other countries.
The difference in the rate of economic growth between countries is not merely the issue of
capital accumulation, the role of technology, and human capital; there is a role of institutions
in it (Acemoglu et al., 2001). They suggest that in certain cases, institutions are the main
determinants of economic growth.
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
15
Etikonomi
Volume 18 (1), 2019: 13 – 28
With recent development in semi-endogenous growth model and the availability of
data on institutions, it is imperative to develop an extension to semi-endogenous growth
model with attention towards institutions. Semi-endogenous growth model emphasizes
ideas creation where human capital and technological progress are the key ingredients. There
must exist a relationship between institutions and ideas creation aspects. The first step in
extending a semi-endogenous growth model can be directed toward building an institutional
index. This study attempts to employ PCA methodology popularized by Filmer & Pritchett
(2001) in building an institutional index. Later, a relationship between institutions and
ideas creation is established. The previous study that has attempted to elaborate institutions
in their analysis is of Hall & Jones (1999). The big difference with our study lies in their
methodology in constructing the index. Hall & Jones (1999) use simple mean techniques of
several institutions indicators developed by Knack & Keefer (1995), while our methodology
is based on PCA. Since institutions influence knowledge directly and thus influence the
rate of economic growth, we then build an institutional index in relation to the growth of
knowledge. The technique used to calculate the growth of knowledge follows Jones (2002).
The next section will briefly explain institutions and PCA methodology before we come to
calculate the institutional index.
Methods
This study will use some indicators of the International Country Risk Guide (ICRG)
database between 1983 and 2013. Indicators used in this study are perceived to have direct
impacts on the stock of knowledge and technological progress. The choice of the time span is
merely a case of data availability. The ICRG data that is utilized in this study consists of: (a)
Investment Profile; (b) Internal Conflict; (c) Corruption; (d) Law and Order; (e) Bureaucracy
Quality.
Investment Profile measures risk factors in the business. Risk factors in the business
include contract cancellation factors, the rate at which investors can recover their capital
(repatriation), and the level of government delay in making payments to investors. Internal
Conflict measures the factors of political violence and its influence on government. Corruption
assesses the extent of corruption in the order of the political system. Law and Order measure
the level of strength, independence, and fairness of the legal system. On the point of view of
the government and institutions in relations with knowledge and technological affairs then
the greater the risk signals the more likely knowledge and technological affairs receive less
attention or become the priority. On the point of view of an investor then the greater the risk
the less likely an investor would invest in the projects.
This study includes 100 countries, after that the countries can be grouped according
to values of S and growth of At. S, and At are institutional index and growth of knowledge
consecutively. We then plot this value on quadrant graphs that will give us 4 quadrants of
countries. The use of this classification method is due to the number of countries involved
but also to give better views on the current position of each country. This study is also new
in terms of using 100 countries in the analysis. The original Jones (2002) involves only 6
countries. We will combine the concept of growth of At as is explained in Jones (2002) to
Hari Nugroho. Institutions as a Main Determinant in Economic Growth
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
16
find correlation among sources of growth. The number of studies involving many countries
is expected to provide a more general and comprehensive picture and can represent the entire
sample well.
To avoid collinearity bias, this study used Principal Component Analysis (PCA) to
form a composite index. PCA is used to describe the variance-covariance matrix structure
of a set of variables through a linear combination of these variables. In general, the main
components can be useful for the reduction and interpretation of variables. Let’s say there are
p variables consisting of n objects. Suppose also that from p variable, k main component is
made (with k ≤ p) which is a linear combination of p of that variable. K, the main component,
can replace the p-variables that makeup without losing much information about the whole
variable. Generally, PCA is an intermediate analysis that means the main component results
can be used for further analysis.
To calculate the impact of social infrastructure or institutions we will use a composite
measure defined as “the sum of the weighted components of the political risk measure of the
International Country Risk Guide”. The index is based on the rating of the ICRG on the
6 components as below: (i) 12 points for each variable that includes investment profile and
internal conflict; (ii) 6 points for each variable including corruption and law and order; (iii)
4 points for each variable that includes bureaucratic quality.
Socioeconomic conditions variables are not included, because these variables are
related to economic performance. So, it has a great possibility to influence perceptions of
the institution as described by Jellema & Roland (2011). His replacement uses an additional
open trade. So the PCA model used is:
(12)
Where:
X1 : Bureaucratic Quality
X2 : Investment Profile
X3 : Internal Conflict
X4 : Corruption
X5 : Law and Order
X6 : Trade Openness
Panel data analysis can be used in dynamic models in relation to dynamic dynamics of
adjustment. This dynamic relationship is characterized by the presence of lag of the dependent
variable among the regressor variables.
Some of the criteria used to find the best dynamic model or GMM model are: First,
Not biased. Estimators of pooled least squares are biased upward and estimators of fixed
effects are biased downwards. An unbiased estimate is in between. Second, the instrument
must be valid. The valid meaning is if there is no correlation between the instrument and the
component error. This validity is checked using the Sargan test. The null hypothesis of the
Sargan test states that the instrument has no problem with validity (valid instrument). The
instrument will be valid if the Sargan test cannot reject the null hypothesis. If the result of the
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
17
Etikonomi
Volume 18 (1), 2019: 13 – 28
AB-GMM method indicates the instrument used is invalid, then the SYS-GMM method is
used. Third, the estimation result must be consistent. An autocorrelation test on the GMM
approach is used to determine the consistency of the estimation results. The consistency
properties of the estimators obtained can be checked from Arellano-Bond statistics m1 and
m2, which can be calculated automatically on some software. The estimate will be consistent
if m1 denotes the null hypothesis is rejected and m2 indicates the null hypothesis is not
rejected (Arellano & Bond, 1991)
Result and Discussion
Islam (1995) mentions the importance of the institutional role in explaining differences
in economic growth. Research on the determinants of differences in economic growth and
income between countries can be grouped into 3 broad theoretical groups. The first group of
theories is a group that focuses on factors of input to production processes, such as physical
and human capital, and technological advances that support economic performance. Solow
(1956), Lucas Jr (1988), Romer (1986, 1990), Grossman & Helpman (1991), Jones (1995a,
1995b, 2002), Segerstrom (1998) had started the discussion about this topic. Endogenous
growth models fall into the first group. The second theoretical group is the focus group on
location and geographical location where certain characteristics will support the economy
to reach the highest level of growth while the location or other location is less supportive.
Sachs (2001), Gallup et al., (1999), and others had conducted the study from this group. A
third theoretical group is a group that focuses on institutions as a driver of economic growth.
North (1991) had pioneered the study in the third group.
Good institutions foster the growth of knowledge. Jones (2002) argues that the engine
of economic growth is the creation of ideas. Jones (2002) seeks to explain the stagnant rate of
growth in the United States during 1950 – 1993 and concludes that much of the growth is
attributable to the growth in ideas (almost 70%). Jones (2002) mentions that the differences
among economies are endowment and allocation. This creates opportunities in extending
Jones (2002) by incorporating institutions into the model. Good or bad institutions can be
considered as endowment while effective or ineffective allocations are results of good or bad
institutions.
Jones (2002) proposes to calculate the accumulated knowledge with the following
equation:
(2)
In Nugroho (2018), equation 3 has been modified to become:
(3)
The channel we will use to incorporate institutional index into Jones modified model
is through variable, S, which we limit its value to a maximum of 1. The reason for the
maximum value is to that after a certain country reaches S =1 then it becomes what Jones
(2002) explained in his research, an advanced country. As S ≈ 1, the equation (10) will return
Hari Nugroho. Institutions as a Main Determinant in Economic Growth
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
18
to its original Jones (2002) version, in equation (9). Jones (2002) assumes that countries in
his model are all the same in level of ability to conduct research and to foster new knowledge.
In our model, it then translates to S = 1. We allow for countries’ variabilities in their ability
to conduct research and to foster new knowledge. This is a novelty this study will offer.
Determination of the number of components in the PCA is done by searching for
variables or components that are not correlated, independent of each other, but fewer than
the initial variables. Although it produces a fewer number of variables, it absorbs most of
the information contained in many more initial variables and can contribute to the variance
of all variables. In PCA, the determination of the component refers to the eigenvalue value,
indicating the amount of contribution of the component to the variance or diversity of all
initial variables. In this case, if the eigenvalue value obtained is greater than one, then the
component formed can be maintained. Otherwise, if the eigenvalue value is less than one,
then the component cannot be used.
Table 1 presents the results of eigenvalue calculations for the formation of the corporate
vulnerability index, the percentage of total diversity (Proportion) and the cumulative
total diversity (Cumulative Proportion) capable of being explained by the diversity of
the components formed. Based on Table 1, of the 12 components formed there are three
components that have eigenvalue greater than one. Component 1 has an Eigenvalue of
2.784994, Component 2 of 1.255002, and Component 3 of 1.010487.
Table 1. Eigenvalue Value for Each Component
Eigenvalues: (Sum = 12, Average = 1)
Number Value Difference Proportion Cum. Value Cum Proportion
1 2.784994 1.529992 0.4642 2.784994 0.4642
2 1.255002 0.244515 0.2092 4.039995 0.6733
3 1.010487 0.355627 0.1684 5.050482 0.8417
4 0.654859 0.434570 0.1091 5.705341 0.9509
5 0.220289 0.145920 0.0367 5.925631 0.9876
6 0.074369 — 0.0124 6.000000 1.0000
Meanwhile, in Table 1 there is also a column ‘Proportion’ which shows the percentage
of variance or diversity that can be explained by each component and column “Cumulative
Proportion” which describes the cumulative of each component simultaneously. The
magnitude of diversity capable of being explained by Component 1 is 46.42 percent. The
diversity explained by Components 1 and 2 is 67.33 percent. The diversity explained by
Components 1, 2, and 3 is 84.17 percent. Based on the eigenvalue of the three components
greater than 1, and the cumulative percentage of the three components of 84.17 percent,
it can be concluded that the three components can represent the diversity of the initial
variables.
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
19
Etikonomi
Volume 18 (1), 2019: 13 – 28
Table 2. Component Matrix
Components Loading Factor
Bureaucratic Quality 0.185092
Investment Profile 0.458274
Internal Conflict 0.053452
Corruption -0.534463
Law and Order 0.397554
Trade Openness 0.556030
Table 2 presents the component matrix that shows the magnitude of the correlation
of each variable in the formed component, or loading factor. Based on Table 2 below, it
appears that there are three factors or components that are formed from the six indicators of
vulnerability. This shows that the three components are the most optimal amount to reduce
the six original variables.
We can determine the Factor Equation by comparing the correlation value on each line
within each component formed (see Table 2). We use the general form below to generate a
factor equation.
Where expresses mean and is a standard deviation of indicators (component), .
α expresses weight or loading factor of each indicator, in the first main component. If we
combine the information from Table 2, with the above factor equation, we can get:
Where: X1 is bureaucratic quality; X2 is an investment profile; X3 is internal conflict; X4 is
corruption; X5 is law and order; X6 is trade-openness.
From the results, the most ideal in the determination of the institutional index is the
first group. This is because the nature of the equation is non-negativity that means each
indicator gives a positive contribution to the resulting index. After obtaining the index value
of each country we then do rescaling of value between 0 – 1. We use value 0 – 1 to analyze
institutional index where 0 is the minimum value of the institutional index while 1 is the
maximum value. Before we come to the growth of At, we calculate At using the following
equation:
Hari Nugroho. Institutions as a Main Determinant in Economic Growth
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
20
Thus the growth of At is the log of two consecutive years or we can take a log of the
above equation then differentiate with respect to time. We need to calculate the growth of At
in order to compare each individual country’s institutional indices with the growth of At. In
doing so, we can easily make a conclusion about the relationship between institutional indices
with the growth of At. We can say that to measure productivity, we calculate the growth of At
and the higher the indices, the higher the productivity of a country will be. Hence, countries
with higher productivity also translate to higher economic growth.
Table 3. Calculation Result of Index of Institution in 1984-2013
No. Country Average of S No. Country Average of S
1 Luxembourg 0.937 51 Jamaica 0.479
2 Singapore 0.907 52 Uruguay 0.478
3 Netherlands 0.871 53 India 0.464
4 Finland 0.867 54 Turkey 0.464
5 Switzerland 0.851 55 Dominican Republic 0.463
6 Sweden 0.849 56 Madagascar 0.459
7 New Zealand 0.848 57 Ecuador 0.456
8 Denmark 0.843 58 Gabon 0.449
9 Norway 0.838 59 Panama 0.448
10 Canada 0.838 60 Kenya 0.447
11 Austria 0.836 61 Paraguay 0.444
12 Iceland 0.834 62 Ghana 0.438
13 Belgium 0.808 63 Malawi 0.432
14 Ireland 0.790 64 Nicaragua 0.431
15 Australia 0.787 65 Cote d’ivoire 0.431
16 United States 0.773 66 Burkina Faso 0.428
17 United Kingdom 0.768 67 Iran 0.428
18 Japan 0.747 68 Egypt 0.4
27
19 Germany 0.740 69 Senegal 0.4
24
20 France 0.724 70 Yemen 0.418
21 Malta 0.710 71 Mozambique 0.407
22 Cyprus 0.703 72 Guinea 0.407
23 Taiwan 0.687 73 Suriname 0.404
24 Portugal 0.684 74 Cameroon 0.404
25 Spain 0.650 75 Philippines 0.402
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
21
Etikonomi
Volume 18 (1), 2019: 13 – 28
No. Country Average of S No. Country Average of S
26 Malaysia 0.648 76 Venezuela 0.397
27 Botswana 0.639 77 Algeria 0.396
28 Italy 0.635 78 Liberia 0.386
29 Republic of Korea 0.614 79 Indonesia 0.380
30 Bahrain 0.610 80 Angola 0.378
31 Chile 0.609 81 Sri Lanka 0.375
32 Israel 0.604 82 Honduras 0.374
33 Oman 0.596 83 Peru 0.366
34 Greece 0.580 84 Niger 0.365
35 Costa Rica 0.574 85 Pakistan 0.357
36 Saudi Arabia 0.555 86 Colombia 0.357
37 Kuwait 0.550 87 Zimbabwe 0.356
38 Jordan 0.543 88 Uganda 0.356
39 Tunisia 0.539 89 El Salvador 0.355
40 Trinidad & Tobago 0.538 90 Guatemala 0.354
41 Gambia 0.526 91 Bolivia 0.345
42 China 0.525 92 Togo 0.344
43 Morocco 0.525 93 Mali 0.3
26
44 Thailand 0.524 94 Sierra Leone 0.324
45 South Africa 0.524 95 Nigeria 0.307
46 Ethiopia 0.520 96 Bangladesh 0.284
47 Zambia 0.505 97 Haiti 0.279
48 Mexico 0.496 98 Iraq 0.214
49 Argentina 0.491 99 Sudan 0.206
50 Brazil 0.480 100 Congo 0.153
Table 3 shows that OECD countries dominate index values with the first top 20
countries. It proves that OECD countries have better institutions compared to other
groups of countries. Another interesting finding is Botswana that places no 27, just above
Italy. Perhaps this is so because of the successful and continuous efforts of the Botswana
government in eradicating corruptions. North (1991) defines institutions, as rules of the
game in a society or in more formal definitions do humans that ultimately shape human
relationships within the society create boundaries. In the case of Botswana, its government
has done a great job in defining rules of the game in society hence translating to a better
quality of institutions and economic growth. North (1991) argues that institutions are a
major cause of economic development and have hypothesized that institutions play a role
in both short and long term growth. As we can see most African countries lack good quality
institutions resulting in lower economic growth. The rest of the African countries is at the
bottom of the index.
Hari Nugroho. Institutions as a Main Determinant in Economic Growth
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
22
Figure 1. Quadrant
As was expected, most OECD countries lie in the first quadrant that is characterized
by the value of S near 1 and the above average value of growth of At (See Figure 1). From
the first quadrant, we can also find that Luxembourg (69) has the highest institutional
index (0,9371) of all 100 countries (refer to Table 4). Another interesting finding is that
Botswana, the only African country, made it to the first quadrant with the value of index
0,6393 (refer to Table 5). This can be related to the successful effort of the Government
of Botswana in fighting against corruption in recent years. If this value is compared to
that of Indonesia, then Indonesia still falls behind Botswana (0,6393 > 0,3797). Yildirim
& Golkap (2016) says that institutional factors can increase or decrease productivity.
To achieve high economic growth, the state must have institutions that encourage every
organization within the country to engage in productive activities. In developing countries,
the existing institutions prioritize distribution activities rather than production activities
so that conditions leading to monopolies are created that ultimately inhibit productive
opportunities. In addition to increased productivity, good institutions will increase
efficiency and trust. Complete values of the institutional index of all 100 countries can be
referred to Table 3.
The effect of institutional indexes is estimated using the GMM method, where the
data used is non-OECD data because OECD countries are considered as the maximum
limit of the multifactor productivity value. The estimation result shows that the model
used is valid and consistent. This result is valid and consistent (See Table 6). Variables
that have an impact on the growth of At are human capital, multifactor productivity of
the country, and multifactor productivity of advanced countries, institutional index, and
growth of At in the preceding year. Value of coefficient of the institutional index of 0.0310
describes that whenever there is a rise of 1% in the institutional index then it will raise the
growth of At by 0.03%. Value of growth of At at lag 1 is less than 1 but slightly more than
0 that shows there is convergence among OECD countries being analyzed.
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
23
Etikonomi
Volume 18 (1), 2019: 13 – 28
Table 4. Countries-Quadrants
Quadrant 1 Quadrant 2 Quadrant 3 Quadrant 4
5 Australia 2 Angola 1 Algeria 18 Canada
7 Austria 4 Argentina 10 Bahrain 23 Costa Rica
10 Bahrain 13 Bolivia 11 Bangladesh 52 Iceland
12 Belgium 15 Brazil 16 Burkina Faso 59 Israel
14 Botswana 21 China 17 Cameroon 60 Italy
20 Chile 26 Cote d ‘Ivoire 22 Colombia 62 Japan
27 Cyprus 33 Egypt 28 Congo 64 Jordan
29 Denmark 34 El Salvador 31 Dominican 72 Malaysia
38 Finland 44 Ghana 32 Ecuador 107 Trinidad & Tobago
39 France 48 Guatemala 37 Ethiopia 108 Tunisia
43 Germany 53 India 41 Gabon
47 Greece 54 Indonesia 42 Gambia
58 Ireland 55 Iran 49 Guinea
67 Kuwait 57 Iraq 50 Haiti
69 Luxembourg 61 Jamaica 51 Honduras
74 Malta 73 Mali 65 Kenya
78 Netherlands 85 Pakistan 68 Liberia
79 New Zealand 86 Panama 70 Madagascar
83 Norway 87 Paraguay 71 Malawi
84 Oman 88 Peru 75 Mexico
91 Portugal 90 Philippines 76 Morocco
92 Rep. of Korea 99 Sri Lanka 77 Mozambique
93 Saudi Arabia 101 Suriname 80 Nicaragua
96 Singapore 105 Thailand 81 Niger
98 Spain 113 Uruguay 82 Nigeria
102 Sweden 116 Yemen 94 Senegal
103 Switzerland 117 Zambia 95 Sierra Leone
104 Taiwan 97 South Aftica
111 United Kingdom 100 Sudan
112 United States 106 Togo
109 Turkey
110 Uganda
114 Venezuela
118 Zimbabwe
Hari Nugroho. Institutions as a Main Determinant in Economic Growth
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
24
Ta
b
le
5
. I
n
st
it
u
ti
o
n
al
In
d
ex
a
n
d
G
ro
w
th
o
f A
t
Q
u
ad
ra
n
t
1
A
vg
. o
f S
A
vg
. o
f G
ro
w
th
o
f A
t
Q
u
ad
ra
n
t
2
A
vg
.
o
f S
A
vg
. o
f G
ro
w
th
o
f A
t
Q
u
ad
ra
n
t
3
A
vg
. o
f S
A
vg
. o
f G
ro
w
th
o
f A
t
Q
u
ad
ra
n
t
4
A
vg
. o
f S
A
vg
. o
f G
ro
w
th
o
f A
t
A
u
st
ra
lia
0.
78
66
0.
01
27
A
n
g
o
l
a
0.
37
83
0.
03
53
A
lg
er
ia
0.
39
60
-0
.0
28
1
C
an
ad
a
0.
83
81
0
0.
00
24
6
A
u
st
ri
a
0.
83
59
0.
02
68
A
rg
en
ti
n
a
0.
49
14
0.
05
18
B
ah
ra
i
n
0.
61
01
0.
01
9
2
C
o
st
a
R
ic
a
0.
57
38
8
-0
.0
00
01
B
ah
ra
in
0.
61
0
1
0.
01
92
B
o
liv
ia
0.
34
48
0.
02
41
B
an
g
la
d
es
h
0.
28
41
-0
.0
00
2
Ic
el
a
n
d
0.
83
43
8
0.
00
33
1
B
el
g
iu
m
0.
80
77
0.
01
7
5
B
ra
zi
l
0.
48
00
0.
02
11
B
u
rk
in
a
Fa
so
0.
42
81
0.
00
93
Is
ra
el
0.
60
35
2
0.
00
90
5
B
o
ts
w
an
a
0.
63
93
0.
01
61
C
h
in
a
0.
52
53
0.
03
18
C
am
er
o
o
n
0.
40
37
-0
.0
23
6
It
al
y
0.
63
53
0
0.
01
18
2
C
h
ile
0.
60
91
0.
02
08
C
o
te
d
‘I
vo
ir
e
0.
43
13
0.
01
20
C
o
l
o
m
b
ia
0.
35
71
-0
.0
03
8
Ja
p
an
0.
74
69
8
0.
00
42
1
C
yp
ru
s
0.
70
27
0.
01
96
Eg
yp
t
0.
42
72
0.
04
26
C
o
n
g
o
0.
15
32
-0
.0
33
4
Jo
rd
an
0.
54
34
5
0.
00
54
9
D
en
m
ar
k
0.
84
30
0.
01
92
El
S
al
va
d
o
r
0.
35
55
0.
07
25
D
o
m
in
ic
an
0.
46
26
0.
00
36
M
al
ay
si
a
0.
64
79
2
0.
01
04
4
Fi
n
la
n
d
0.
86
67
0.
02
11
G
h
an
a
0.
43
80
0.
01
85
Ec
u
ad
o
r
0.
45
60
0.
00
91
Tr
in
id
ad
&
T
o
b
ag
o
0.
53
77
0
0.
01
16
4
Fr
an
c
e
0.
72
40
0.
01
56
G
u
at
em
al
a
0.
35
44
0.
01
75
Et
h
io
p
ia
0.
51
95
-0
.0
01
1
Tu
n
is
ia
0.
53
92
1
0.
00
69
2
G
er
m
an
y
0.
74
03
0.
02
67
In
d
ia
0.
46
40
0.
04
02
G
a
b
o
n
0.
44
86
0.
00
96
G
re
ec
e
0.
57
97
0.
01
96
In
d
o
n
es
ia
0.
37
97
0.
02
35
G
a
m
b
ia
0.
52
60
-0
.0
30
3
Ir
el
an
d
0.
79
00
0.
03
57
Ir
an
0.
42
77
0.
04
78
G
u
in
ea
0.
40
65
-0
.0
42
2
Ku
w
ai
t
0.
54
99
0.
01
28
Ir
aq
0.
21
44
0.
03
69
H
ai
ti
0.
27
95
-0
.0
28
1
Lu
xe
m
b
o
u
rg
0.
93
71
0.
02
08
Ja
m
ai
ca
0.
47
91
0.
02
16
H
o
n
d
u
ra
s
0.
37
37
-0
.0
06
9
M
al
ta
0.
70
99
0.
01
54
M
al
i
0.
32
57
0.
02
74
K
en
ya
0.
44
72
-0
.0
05
6
N
et
h
er
la
n
d
s
0.
87
15
0.
01
68
Pa
ki
st
an
0.
35
72
0.
01
89
Li
b
er
ia
0.
38
59
-0
.0
08
8
N
ew
Z
ea
la
n
d
0.
84
80
0.
02
01
Pa
n
am
a
0.
44
82
0.
02
00
M
ad
ag
as
ca
r
0.
45
88
-0
.0
13
0
N
o
rw
ay
0.
83
83
0.
02
89
Pa
ra
g
u
ay
0.
44
41
0.
01
43
M
al
aw
i
0.
43
21
-0
.0
19
0
O
m
an
0.
59
63
0.
01
90
Pe
ru
0.
36
60
0.
01
76
M
ex
ic
o
0.
49
58
-0
.0
13
3
Po
rt
u
g
al
0.
68
41
0.
03
01
Ph
ili
p
p
in
es
0.
40
22
0.
02
60
M
o
ro
cc
o
0.
52
48
0.
00
06
R
ep
. o
f K
o
re
a
0.
61
41
0.
02
71
Sr
i L
an
ka
0.
37
46
0.
04
27
M
o
za
m
b
iq
u
e
0.
40
73
0.
00
22
Sa
u
d
i A
ra
b
ia
0.
55
46
0.
02
19
Su
ri
n
am
e
0.
40
25
0.
02
13
N
ic
ar
ag
u
a
0.
43
13
-0
.0
33
2
Si
n
g
ap
o
re
0.
90
67
0.
02
85
Th
ai
la
n
d
0.
52
39
0.
02
50
N
ig
er
0.
36
49
0.
00
08
Sp
ai
n
0.
65
04
0.
01
99
U
ru
g
u
ay
0.
47
76
0.
02
84
N
ig
er
ia
0.
30
72
-0
.0
11
7
Sw
ed
en
0.
84
87
0.
02
59
Ye
m
en
0.
41
81
0.
05
54
Se
n
eg
al
0.
42
40
-0
.0
08
6
Sw
it
ze
rl
an
d
0.
85
08
0.
01
58
Z
am
b
ia
0.
50
54
0.
03
84
Si
er
ra
L
eo
n
e
0.
32
38
0.
00
70
Ta
iw
an
0.
68
71
0.
01
41
So
u
th
A
ft
ic
a
0.
52
36
-0
.0
06
3
U
n
it
ed
K
in
g
d
o
m
0.
76
77
0.
01
68
Su
d
an
0.
20
58
0.
00
49
U
n
it
ed
S
ta
te
s
0.
77
26
0.
01
38
To
g
o
0.
34
38
-0
.0
10
3
Tu
rk
ey
0.
46
38
0.
00
96
U
g
an
d
a
0.
35
60
0.
00
13
Ve
n
ez
u
el
a
0.
39
68
0.
00
19
Z
im
b
ab
w
e
0.
35
60
-0
.0
46
3
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
25
Etikonomi
Volume 18 (1), 2019: 13 – 28
Table 6. GMM result
Variables Coefficient P>|z|
GA(-1) 0.1079 0.0000
S 0.0310 0.0000
LogH 0.0094 0.0000
LogA 0.0935 0.0000
LogA* 0.0958 0.0000
C -0.4397 0.0000
Arrelano Bond Test
M1 0.0001
M2 0.9125
Sargan Test
Prob > chi2 1.0000
This result is consistent with Siddiqui & Ahmed (2013) that suggest favorable
institutions positively affect economic growth. There is a causal link between a cluster of
good institutions and rapid ling run economic growth (Lin & Chen, 2011). Institutional
is a key role in the process of economic development (Osman et al., 2011; Roy et al., 2014;
Ahmad & Hall, 2017). According to the result, it is imperative that the government should
pay more attention to institutional indices. The improvement of institutional quality
can attract more foreign direct investment (Kandil, 2009). The success of institutions is
largely determined by the degree of accountability and corruption (Sumanjeet, 2015). The
institutional reforms to upgrade the quality of both political and economic institutions are
crucial for the countries (Slesman et al., 2015; Rachdi et al., 2018).
Conclusion
PCA result shows that there is a difference in index values between each quadrant.
The result shows that developed countries have a tendency of higher index value relative
to other countries. Quadrant I characterizes developed coun t ries in which institutional
index is higher than any other quadrant. The growth of At is fairly high in this quadrant
but not the highest. Institutional influence on economic growth is evidenced by the results
of GMM where the influence of institutions has a positive and significant impact on
economic growth. It proves that countries with higher in s titutional indexes have higher
economic growth than those with lower institutional indices. GMM results also prove that
human capital and multifactor productivity have a significant effect on economic growth
that means economic growth is not only influenced by capital and labor but also influenced
by human capital and multifactor productivity variable.
The government must redefine its definition of good institutions as being innovation
and growth-oriented institutions. From the perspecti ve of innovations, we mean that
institutions must provide ways and environment to cu ltivate new ideas. Besides that,
from growth-oriented, we mean that institutions must actively seek new ways to improve
available knowledge and technology. The government can start from very technical research
Hari Nugroho. Institutions as a Main Determinant in Economic Growth
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
26
and development institutions and later make ways to other institutions in the nations. The
improvement of the institutions that can foster and advance knowledge and technology
are prioritized. Improvement can take many forms from increasing budget, increasing
human capital involved, or creating conditions that can sustain continuous research and
development. The conditions that sustain research and development can also be further
supported by means of law and regulations.
References
Acemoglu, D., Aghion, P., & Violante, G. L. (2001). Deunionization, Technical Change
and Inequality. Carnegie-Rochester Conference Series on Public Policy, 55(1), 229–264.
https://doi.org/10.1016/S0167-2231(01)00058-6
Acemoglu, D., & Robinson, J. A. (2008). Persistence of Power, Elites, and Institutions.
American Economic Review, 98(1), 267–293.
Ahmad, M., & Hall, S. G. (2017). Economic Growth and Convergence: Do Institutional
Proximity and Spillovers Matter? Journal of Policy Modelling, 39(6), 1065–1085.
https://doi.org/10.1016/j.jpolmod.2017.07.001
Arellano, M., & Bond, S. (1991). Some Test of Specification for Panel Data: Monte Carl
Evidence and An Application to Employment Equations. The Review of Economic
Studies, 58, 277–297.
Filmer, D., & Pritchett, L. H. (2001). Estimating Wealth Effects without Expenditure Data
or Tears: An Application to Educational Enrollment in States of India. Demography,
38(1), 115–132.
Gallup, J. L., Sachs, J. D., & Mellinger, A. D. (1999). Geography and Economic Development.
International Regional Science Review, 22(2), 179–232.
Grossman, G. M., & Helpman, E. (1991). Innovation and Growth in the Global Economy.
Cambridge: MIT Press.
Hall, R. E., & Jones, C. I. (1999). Why Do Some Countries Produce So Much More Output
Per Worker Than Others? The Quarterly Journal of Economics, 114(1), 83–116.
Islam, N. (1995). Growth Empirics: A Panel Data Approach. The Quarterly Journal of
Economics, 110(4), 1127–1170.
Jellema, J., & Roland, G. (2011). Institutional Clusters And Economic Performance. Journal
of Economic Behavior and Organization, 79(1–2), 108–132. https://doi.org/10.1016/j.
jebo.2011.04.003
Jones, C. I. (1995a). R & D Based Models of Economic Growth. Journal of Political Economy,
103, 759–784.
Jones, C. I. (1995b). Time Series Tests of Endogenous Growth Model. Quarterly Journal of
Economics, 110, 495–525.
Jones, C. I. (2002). Sources of U.S. Economic Growth in a World of Ideas. American Economic
Review, 92, 220–239.
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
27
Etikonomi
Volume 18 (1), 2019: 13 – 28
Kandil, M. (2009). Determinants of Institutional Quality and Their Impact on Economic
Growth in the MENA Region. International Journal of Development Issues, 8(2), 134–
167. https://doi.org/10.1108/14468950910997
Knack, S., & Keefer, P. (1995). Institutions and Economic Performance: Cross-Country
Tests Using Alternative Measures. Economics and Politics, 7(3), 207–228. https://doi.
org/10.1111/j.1468-0343.1995.tb00111.x.
Lin, Y., & Chen, X. (2011). Review of Institutions and Economic Growth Empirical Research.
In Q. Zhou (Ed.), Applied Economics, Business and Development. Berlin: Springer.
Lucas Jr, R. E. (1988). On the Mechanics of Economic Development. Journal of Monetary
Economics, 22(1), 3–42. https://doi.org/10.1016/0304-3932(88)90168-7
North, D. C. (1991). Institutions. Journal of Economic Perspectives, 5(1), 97–112. https://doi.
org/10.1257/jep.5.1.97
Nugroho, H. (2018). Peran Institusi dalam Model Pertumbuhan Semi-Endogen (The Role of
Institutions on Semi-Endogenous Growth Model). (Unpublished Dissteration). Depok:
Universitas Indonesia.
Osman, R. H., Alexiou, C., & Tsaliki, P. (2011). The Role of Institutions in Economic
Development: Evidence from 27 Sub‐Saharan African Countries. International Journal
of Social Economics, 39(1–2), 142–160. https://doi.org/10.1108/03068291211188910
Rachdi, H., Hakimi, A., & Hamdi, H. (2018). Liberalization, Crisis, and Growth in MENA
Region: Do Institutions Matter? Journal of Policy Modelling, 40(4), 810–826. https://
doi.org/10.1016/j.jpolmod.2018.05.001
Romer, P. M. (1986). Increasing Returns and Long-Run Growth. Journal of Political Economy,
94(5), 1002–1037.
Romer, P. M. (1990). Endogenous Technological Change. Journal of Political Economy, 98(5),
S71–S102.
Roy, B. C., Sarker, S., Mandal, N. R., & Pandey, S. K. (2014). Role of Institutions in
Economic Performance: An Empirical Study on Interstate Differences in Indian States
with Reference to Jharkhand. European Scientific Journal, 10(10), 313–335.
Sachs, J. D. (2001). Tropical Underdevelopment. NBER Working Paper No. 8119.
Segerstrom, P. S. (1998). Endogenous Growth without Scale Effects. American Economic
Review, 88, 1290–1310.
Siddiqui, D. A., & Ahmed, Q. M. (2013). The Effect of Institutions on Economic Growth:
a Global Analysis Based on GMM Dynamic Panel Estimation. Structural Change and
Economic Dynamics, 24, 18–33. https://doi.org/10.1016/j.strueco.2012.12.001
Slesman, L., Baharumshah, A. Z., & Raees, W. (2015). Institutional Infrastructure and Economic
Growth in Member Countries of the Organization of Islamic Cooperation (OIC). Economic
Modelling, 51, 214–226. https://doi.org/10.1016/j.econmod.2015.08.008
Solow, R. M. (1956). A Contribution to the Theory of Economic Growth. The Quarterly
Journal of Economics, 70(1), 65–94.
Hari Nugroho. Institutions as a Main Determinant in Economic Growth
http://journal.uinjkt.ac.id/index.php/etikonomi
DOI: htttp://dx.doi.org/10.15408/etk.v18i1.8242
28
Sumanjeet. (2015). Institutions, Transparency, and Economic Growth. Emerging Economy
Studies, 1(2), 188–210. https://doi.org/10.1177/2394901515599272
Yildirim, A., & Golkap, M. F. (2016). Institutions and Economic Performance: A Review on
the Developing Countries. Procedia Economics and Finance, 38, 347–359. https://doi.
org/10.1016/S2212-5671(16)30207-6
Procedia Economics and Finance 30 ( 2015 ) 702 – 709
2212-5671
© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of IISES-International Institute for Social and Economics Sciences.
doi: 10.1016/S2212-5671(15)01319-2
ScienceDirect
Available online at www.sciencedirect.com
3rd Economics & Finance Conference, Rome, Italy, April 14-17, 2015 and 4th Economics &
Finance Conference, London, UK, August 25-28, 2015
Influence of Selected Institutional Factors on the Economic Growth:
Case Open Markets
Pavel Procházkaa, Klára Čermákováa
a University of Economics, nam. W. Churchilla 4, Prague 3, 130 67, Czech Republic
Abstract
This article’s goal is to judge impacts of selected institutional factors on economic growth. Institutional economics, which started
developing its modern approach after 1990 is closely in connection with Index of Economic Freedom presented by The Heritage
Foundation. The article shows comparison of selected institutional variables (Open Market category) with economic development
of particular countries’ both theoretical framework and empiric analysis. Empiric section uses data of The Heritage Foundation on
which the connection between selected indicators and economic progress (represented by income per capita) is compared. Markets)
on the economic growth. The research part concentrates on testing selected basic preconditions that countries striving for economic
growth should meet. These requirements are defined by the Heritage Foundation organization and published under the abbreviation
IEF. After the economic crisis of 2008/09 the index became a target of criticism and its methodology was doubted by many.
Influence of these factors have been tested several economists with positive results. This study, however, did not confirm theory
about positive correlation between trade openness and economic growth in mid-term horizon. Similarly, there is no direct
connection between influence of FDI inflow and economic development. On the other hand, study showed positive correlation
between R&D expenditures and economic growth. Similarly, economics with friendly business environment reach better economic
condition.
© 2015 The Authors. Published by Elsevier B.V.
Peer-review under responsibility of IISES-International Institute for Social and Economics Sciences.
Keywords: Institution economics; Economic growth; Trade, Investment freedom
1. Introduction
Institutional economics in its modern form began to develop after 1990 and related to the works of de Soto (1989)
and North (1991). It offered a new approach to the problematic of economic growth based on so-called rules of
behaviour of institutions (i.e. factors) such as legal framework or religion. Such factors used to be neglected in the
economic theory; however, these very factors may be the cause of uneven economic growth, intensity variation of
convergence and inequality among countries and people, as measured by income or GDP per capita. Compared to the
© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of IISES-International Institute for Social and Economics Sciences.
http://crossmark.crossref.org/dialog/?doi=10.1016/S2212-5671(15)01319-2&domain=pdf
703 Pavel Procházka and Klára Čermáková / Procedia Economics and Finance 30 ( 2015 ) 702 – 709
beginning of the industrial revolution when the difference between the richest and the poorest countries, measured in
GDP per capita, was negligible, in the 1960s it grew 68 times and now it is 456 times higher. In some countries, the
pace of growth also changed. The era of doubled GDP per capita in the first half of the 20th century seemed promising
and lasted from 30 to 50 years. However, in the second half of the century it failed to exceed 20 years (Parente,
Prescott, 2000 in Parente, 2008). S. Kuznets called this development the “modern economic growth” (Parente, 2008,
p. 24).
Although economists have been trying to find causes of growth, factors that influence it and conditions of its
sustainability for a very long time, to this day we have no satisfactory answers. It could be partly caused by
concentration on what is called hard growth factors (e.g. investment and technologies) or omission of the fact that
growth factors are not stock but flow variables. Conditions, under which the economy grows, therefore change. That
means so-called soft growth factors, which include institutions, are important for explanation of the growth as well as
its convergence and inequality among countries and people*.
The aim of the present paper is to demonstrate the influence of selected institutional factors (Trade Freedom,
Investment Freedom). The research proceeded from data published by The Heritage Foundation for 2013.
2. Used Methods and Methodologies
IEF is crucial indicator of potential growth for many economists. To the importance of IEF showed in their studies
Hansson (2009), Joreiman (2004), Skaaning (2010), Xu (2011), Durlauf (2008), Parente (2008), Baumol (2007),
Mitchell (2013), Lewis (2005), Barro (2009, 2013), Chao (2010) or Ortiz (2009).
The aim of the approach adopted is to verify Ortiz’s thesis about positive influence of Open Markets parameters,
especially openness in international trade and capital flow on economic growth. In his research, Ortiz (2009)
demonstrated that country (economy) more involve in international trade and capital flow reaches higher economic
growth. To support his theory we use data of IEF. Indicators will most often be compared with GDP per capita (prices
of 2000 in USD), year-on-year change in growth, or average growth rate in medium-term period (as for this paper it
is a period of 5 years, 2008–2012). The reason for choosing this methodology stems from dependence of the indicators
of IEF and GDP per capita, as indicated in Figure 1, which compares GDP per capita and the degree of freedom in
individual countries (represented by the points in the graph). The 2012 values show that there is a relatively strong
correlation between the two variables, which proves that both these indicators are significant. The relationship between
the two parameters can be demonstrated with the help of regression analysis. The correlation coefficient can also be
used to define the degree of dependence between the observed parameter and economic development. The correlation
coefficient measures the relative significance of two variables (in our case it means one of the ten variables in
correlation with GNI per capita).
3. The Influence of Institutional Factors on the Economic Growth: Open Markets
In this part of the paper we present results of testing selected parameters of the aforementioned Index of Economic
Freedom (IEF). IEF evaluates individual countries according to 10 parameters divided into 4 groups: Rule of law
(Property rights, Freedom from corruption), Limited government (Government spending, Fiscal freedom), Regulatory
efficiency (Business freedom, Labour freedom, Monetary freedom) and Open markets (Trade freedom, Financial
freedom, Investment freedom). These results provide a set of suggestions concerning reforms in fields that need to be
reformed in order to boost the economic growth. From the factors in question we chose group 4, which falls within
Open Markets, as factors of this group have been studied for the longest time.
* Under the term “development” in this text we understand a number of indicators such as improving living conditions, environment, happiness
and satisfaction, security and freedom. Under “growth” we understand GDP per capita and its derivatives.
704 Pavel Procházka and Klára Čermáková / Procedia Economics and Finance 30 ( 2015 ) 702 – 709
This group involves Trade Freedom indicators that measure the scope of barriers on imported goods according to
the extent of tariffs. Non-tariff barriers (quantitative restrictions, requirements for specific quality etc.) are also taken
into consideration and fall within categories of Investment Freedom and Financial Freedom. Investment Freedom
measures restrictions on foreign capital and the possibility of investors to invest in whatever sector they like. Financial
Freedom measures independence of the banking sector, share of the state in the banking sector, influence of the
government on bank’s lending policies etc.
3.1. Trade Freedom
The Trade Freedom indicator, which is probably the oldest of all indicators, is used to deduce wealth of a country
from its integration into international trade. Such deductions were already made by classics of economy, Smith (1776)
and Ricardo (1817). From the historical point of view, the most developed countries were those that were engaged in
trade. As an example we can name prosperous medieval Italian cities or cities of the Hanseatic League. During the
Great Depression the USA adopted the Smooth-Hawley Tariff Act (1932) which increased import tariffs by 60%. Soon
after that, European countries introduced the same countermeasures. As a result, export from the USA to Europe
declined immediately by two-thirds and the crisis even deepened (Chao, 2010).
The Trade Openness indicator can be used for empirical observation of the importance of trade. It measures the
degree of openness of the country, or more precisely, involvement of the country into international trade. It is defined
as a share of exports of goods and services (% GDP) to imports of goods and services (% GDP). The data source was
the database of the World Bank. Economies of countries that have greater trade openness (and therefore higher ratio
of exports to imports) are more likely to grow (Ortiz, 2009).
Fig. 1.Development of the GDP growth rate in the mid-term horizon depending on tariff barriers
Source: 2013 Index of Economic Freedom, our own adjustment
Analysis of openness of the economy can be also assessed through the comparison of tariff barriers extent with
GDP per capita or with average growth rate in the medium-term horizon (Figure 1). The data show that countries that
have lower tariffs achieve higher living standard. As for the relation to the average growth rate, the situation has been
ambiguous in last five years. A regressive analysis has not proven a direct dependence between GDP growth and the
rate of import surcharges, the reason being the existence of many bilateral (or multilateral) agreements on free trade.
Duty-free policy creates a distorted picture of trade barriers.
–
5,
0
0,0
5,0
10,0
15,0
20,0
0,0 5,0 10,0 15,0 20,0 25,0 30,0
G
D
P
G
ro
w
th
(
%
)
Tariff Rate (%)
R2 = 0,462
4
Correlation = 0,18
705 Pavel Procházka and Klára Čermáková / Procedia Economics and Finance 30 ( 2015 ) 702 – 709
3.2. Investment Freedom
Under investment freedom the IEF understands freedom of capital movement and the extent of the state’s
supervision over it. This indicator includes a survey of openness of the country to foreign investors and restrictions
against the capital inflow. The underlying assumption is that an economy that is more open to foreign capital achieves
higher IEF. Such country would be considered safer for investors and there would be greater inflow of capital. Figure
2 shows that there is a moderate positive correlation between these two parameters.
Fig. 2. Correlation between FDI inflow (log scale) and IEF
Source: 2013 Index of Economic Freedom, our own adjustment
Business activities are represented by the third group of the IEF – Business Openness. They are in direct relation
with investment freedom. The influence of entrepreneurs on the economic growth has been known for quite a long
time. For example Schumpeter (1942, p. 83) considered business activities an essential element without which
economy would stagnate: “The fundamental impulse that sets and keeps the capitalist engine in motion comes from
the consumer’s goods, the new methods of production or transportation, the new markets, the new forms of industrial
organization that capitalist enterprise creates.” Schramm (2008) decided to follow Schumpeter’s way and was another
one to study the business environment. He expanded the necessity of business activities to labour market mobility.
Schramm (2008) considered market mobility crucial as it allows individuals to be motivated to start their own
business.†
† This issue is more of a digression from the microeconomic part and represents a key aspect of economic freedom, which Schramm (2008)
called fluidity. The degree of economic freedom (and growth) in any society reflects the ability of institutions to adapt to market conditions. Yet,
Schramm (2008) was not the first one to embed the business sector into micro level. This privilege belongs to Hernando de Sot o (1989), whose
vision of improving the country’s economic situation is a bottom-up process based on creation of business environment from which everyone will
eventually benefit.
0
1
2
3
4
5
6
0,0 20,0 40,0 60,0 80,0
100,0
FD
I I
nf
lo
w
(l
og
s
ca
le
)
Index of Openess
R2 = 0,1245
Correlation = 0,35
706 Pavel Procházka and Klára Čermáková / Procedia Economics and Finance 30 ( 2015 ) 702 – 709
To show the importance of friendly conditions for entrepreneurship we proceed from Doing Business Index. As
you can see bellow (Figure 3), there is mild positive correlation between GDP per Capita and openness for
entrepreneurs.
Figure 3 Correlation between GDP per Capita and Doing Business Index
Source: 2013 Index of Economic Freedom, 2013 Doing Business, our own adjustment
Ability to act freely, to sets a new entrepreneurship leads to nature and causes of the wealth of nations. Innovation
– the creation of new knowledge and applications that flow from it – and entrepreneurship. For importance of
innovations see Schumpeter (1942), Romer (1990), Lewis (2005), Baumol (2007) etc. Common question is what
determines innovation? Schramm (2008) offers an explanation in the term of fluidity: „Fluidity, then, is that condition
of a loose yet stable alignment of institutions, organizations, and individuals that facilitates the exchange and
networking of knowledge across boundaries“ (Schramm, 2008: 17). This fosters both innovation and its propagation
through entrepreneurship (Ibid).
A you can see bellow (Figure 4), there is strong correlation between R&D expenditures (% GDP) and GDP per
Capita (R2 = 0.952). It is obvious that country (economies) with higher share of R&D expenditures generates higher
personal income.
Fig. 4. Correlation between GDP per Capita and share of R&D expenditures (% GDP)
Source: The World Bank, our own adjustment
R² = 0,4082
$0
$10 000
$20 000
$30 000
$40 000
$50 000
$60 000
$70 000
0,0 50,0 100,0 150,0
G
D
P
pe
r C
ap
ita
Doing Business Index
EU
High income
Lower middle
Low & middle
Middle
income
High income:
OECD
Upper middle
R² = 0,9523
0
0,5
1
1,5
2
2,5
3
0 10000 20000 30000 40000 50000
R&
D
(%
G
D
P)
GDP per Capita (USD)
707 Pavel Procházka and Klára Čermáková / Procedia Economics and Finance 30 ( 2015 ) 702 – 709
A similar conclusion can be reached through the Global Innovation Index, published by Cornell University and
World Intellectual Property Organisation.‡ On the basis of the values (Figure 5) can be seen a fairly strong correlation
between the two variables.
Fig. 5. Correlation between GDP per Capita and Potential of Innovation
Source: The Global Innovation Index, The World Bank, our own adjustment
When talking about institutions, the so-called fluidity is a particular importance. It indicates the company’s ability
to adapt to changing market conditions and to absorb these changes efficiently. Therefore it also depicts the openness
of the market to new ideas. Institutions work as “research gateways” or “durable systems of established and embedded
social rules that structure social interactions” (Hodgson, 2004 in Schramm, 2008).
This concept of fluidity is a parallel to the work of Phelps (2007), who dealt with the concept of “dynamism in
economy”. His aim was to clarify differences in economic performance across countries. According to Phelps (2007,
p. 15): “The level of dynamism is a matter of how fertile the country is in coming up with innovative ideas having
prospects of profitability, how adept it is at identifying and nourishing the ideas with the best prospects, and how
prepared it is in evaluating and trying out the new products and methods that are launched onto the market… A
country’s economic model determines its economic dynamism… There are two dimensions to a country’s economic
model. One part consists of its economic institutions… The other part of the economic model consists of various
elements of the country’s economic culture.”
Difficulties of some countries may not necessarily lie in absence of the institutions but in their inability to adapt to
changing market conditions. This is confirmed, among others, by researches carried out by Fairlie (2007) and Baumol
(2002, 2007). These works indicate that on one hand, a high degree of business activity is desirable, but on the other
hand, it is not a condition sufficient enough to stimulate performance of the economy towards high industry.§
4. Conclusion
The aim of the present paper was to assess the influence of selected institutional factors (Open Markets) on the
economic growth. The research part concentrates on testing selected basic preconditions that countries striving for
economic growth should meet. These requirements are defined by the Heritage Foundation organization and published
under the abbreviation IEF. After the economic crisis of 2008/09 the index became a target of criticism and its
methodology was doubted by many.
‡ Více o studii inovačního potenciálů jednotlivých zemí viz The Global Innovation Index 2014 – The Human Factor in Innovation, World
Intellectual Property Organization.
§ There are other theories and factors that may influence the growth potential. For example Hall (1997) considers appropriate infrastructure to
be a crucial element of economic performance as it stimulates production.
R² = 0,6234
0,0
20,0
40,0
60,0
80,0
100,0
0 20 000 40 000 60 000 80 000 100 000 120 000
G
lo
ba
l I
nn
ov
at
io
n
In
de
x
GDP per Capita (PPP)
708 Pavel Procházka and Klára Čermáková / Procedia Economics and Finance 30 ( 2015 ) 702 – 709
Influence of these factors have been tested several economists with positive results. This study, however, did not
confirm theory about positive correlation between trade openness and economic growth in mid-term horizon.
Similarly, there is no direct connection between influence of FDI inflow and economic development. On the other
hand, study showed positive correlation between R&D expenditures and economic growth. Similarly, economics with
friendly business environment reach better economic condition.
Acknowledgement
Written under the institutional support of the Faculty of economics, University of Economics in Prague, Czech
Republic, VŠE IGA F5/3/2015.
References
Baláž, V. (2006), “Politická ekonómia slovenského kapitalizmu: inštitucionálna a evolučná perspektíva.” Politická ekonomie, Vol. 2006, No. 5, pp.
610-631.
Barro, R. J. (1999), “Determinants of Democracy.” Journal of Political Economy, Vol. 107(S6), pp. 158-183.
Barro, R. J. (2013), “Democracy, Law and Order, and Economic Growth.” Washington: The Heritage Foundation and Dow Jones & Company, Inc.
Barro, R. J., Sala-I-Martín, X. (2004), Economic growth, 2nd Ed. Cambridge: MIT Press.
Blaug, M. (1997), Economic theory in retrospect. Cambridge University Press.
Baumol, W. J. (2002), The Free-Market Innovation Machine: Analyzing the Growth Miracle of Capitalism. Princeton: Princeton University.
Baumol, W. J., Litan, R. E., Schramm, C. J. (2007), Good capitalism, bad capitalism, and the economics of growth and prosperity. New Haven:
Yale University Press.
Bavetta, S., Navarra, P. (2011), “Economic Freedom and the Pursuit of Happiness.” Washington: The Heritage Foundation and Dow Jones &
Company, Inc.
Brunetti, A. (1997), Politics and economic growth: a cross-county data perspective. Organisation for Economic Co-operation and Development,
Paris.
Chao, E. L. (2010), “Trading for Prosperity.” Washington: The Heritage Foundation and Dow Jones & Company, Inc.
Drury, A. C., Krieckhaus, J., Lusztig, M. (2006), “Corruption, Democracy, and Economic Growth.” International Political Science Review, Vol.
27, No. 2, pp. 121-136.
Durlauf, S. N., Kourtellos, A., Tan, C. M. (2008), “Are Any Growth Theories Robust? ” The Economic Journal. March 2008, Vol. 118, Issue 527,
pp. 329-346.
Ehrlich, I., Lui, F. T. (1999), “Bureaucratic Corruption and Endogenous Economic Growth.” Journal of Political Economy, Vol. 107, No. 6, pp.
270-293.
Fairlie, R. (2007), “Entrepreneurship in Silicon Valley During the Boom and Bust.” Small Business Research Summary, No. 296.
Feulner, E. J. (2013), “The Rule of Law.” Washington: The Heritage Foundation and Dow Jones & Company, Inc..
Foster, J. D., Miller, T. (2012), “Public Debt, Economic Freedom, and Growth.” Washington: The Heritage Foundation and Dow Jones & Company,
Inc.
Furchtgott-Roth, D. (2005), “What US Labor Laws Can Teach Europe.” Financial Times.
Gersemann, O. (2004), “Cowboy Capitalism: European Myths, American Reality.” Washington, D.C.: Cato Institute.
Grier, R. (1997), “The Effect of Religion on Economic Development: A Cross National Study of 63 Former Colonies.” Kyklos, Vol. 50, Issue 1,
pp. 47-62.
Grier, R. (1999), “Colonial legacies and economic growth.” Public Choice, Vol. 98, No. 3/4, pp. 317-335.
Grossman, G. N., Helpman, E. (1994), “Endogenous Innovation in the Theory of Growth.” The Journal of Economic Perspectives, Vol. 8, No. 1,
pp. 23-44.
Hall, R. E., Jones, C. I. (1999), “Why Do Some Countries Produce So Much More Output Per Worker Than Others?” The Quarterly Journal of
Economics, Vol. 114, No. 1, pp. 83-116.
Hansson, G. (2009), “What Determines Rule of Law? An Empirical Investigation of Rival Models.” Kyklos, Vol. 62, No. 3, pp. 371-393.
Joreiman, S. F. (2004), “Colonization and the rule of law: Comparing the effectiveness of common law and civil law countries.” Constitutional
Political Economy, Vol. 15, Issue 4, pp. 315-338.
709 Pavel Procházka and Klára Čermáková / Procedia Economics and Finance 30 ( 2015 ) 702 – 709
Kadeřábková, K. (2008), “Konvergence ekonomik a prohlubování globálních nerovností jako problém mezinárodních vztahů.” Praha, University
of Economics.
Lewis, William W. (2005), The Power of Productivity: wealth, poverty, and the threat to global stability. Chicago: University of Chicago Press,
2005, 339 s. ISBN 0-226-47698-7.
Mauro, P. (1995), “Corruption and Growth.” The Quarterly Journal of Economics, Vol. 110, No. 3, pp. 681-712.
Miller, T. (2010), “Economic Freedom in Uncertain Times.” Washington: The Heritage Foundation and Dow Jones & Company, Inc.
Miller, T. (2011), “The Limits of Government.” Washington: The Heritage Foundation and Dow Jones & Company, Inc..
Minárik, P. (2013), “Ekonomie náboženství a její relevance pro ekonomy ve střední Evropě.” Politická ekonomie, No. 5, pp. 691-704.
Mitchell, M. (2013), “Economic Freedom and Economic Privilege.” Washington: The Heritage Foundation and Dow Jones & Company, Inc.
Munkhammer, J. (2007), “The Urgent Need for Labor Freedom in Europe – and the World.” Washington: The Heritage Foundation and Dow Jones
& Company, Inc..
Niskanen, W. A. (2008), “Autocratic, Democratic and Optimal Government: A Sketch,” in Niskanen, W. A, Reflections of a Political Economist:
Selected Articles on Government Policies and Political Processes. Washington, D.C.: Cato Institute, pp. 139-165.
North, D. C. (1991), “Institutions.” Journal of Economic Perspectives, Vol. 5, No. 1, pp. 97-112.
Ortiz, J. (2009). “Does Religion Distribution Matter in the Economic Growth of Latin America? ” International Journal of Business and Economics,
2009, Vol. 8, No. 3, 183-199.
Parente, S. L. (2008), “Narrowing the Economic Gap in the 21st Century.” Washington: The Heritage Foundation and Dow Jones & Company, Inc.
Phelps, E. S. (2007), “Entrepreneurial Culture.” The Wall Street Journal.
Ricardo, D. (2001), On the principles of political economy and taxation. 3rd Edition, Batoche Books, Kitchener, ISBN N/A.
Romer, P. M. (1990), “Endogenous Technological Change.” Journal of Political Economy, Vol. 98, No. 5, Part 2, October, 1990, pp. 71-102.
Schramm, C. J. (2008), “Economic Fluidity: A Crucial Dimension of Economic Freedom.” Washington: The Heritage Foundation and Dow Jones
& Company, Inc.
Schumpeter, J. A. (1942), Capitalism, Socialism, and Democracy, 3rd Edition. New York: Harper and Brothers.
Skaaning, S-E. (2010), “Measuring the Rule of Law.” Political Research Quarterly, Vol. 63, No. 2, pp. 449-460.
Smith, A. (2001), Pojednání o podstatě a původu bohatství národů. Liberální institut, Praha, ISBN 80-86389-15-4.
Soto, H. D. (1989), The other path: the invisible revolution in the Third World. New York: Harper & Row.
The Heritage Foundation (2006-2013), Index of Economic Freedom. Washington: The Heritage Foundation
World Bank (2013), World Bank Report. Washington: World Bank.
Xu, G. (2011), “The Role of Law in Economic Growth: A Literature Review.” Journal of Economic Surveys, Vol. 25, No. 5, pp. 833-871.
Экономическая статистика
34 Статистика и экономика Т. 16. № 2. 2019
УДК 330.354
DOI: http://dx.doi.org/10.21686/2500-3925-2019-2-34-44
Влияние производительности труда
на экономический рост
Цель исследования – исследование уровня производительности
труда как важнейшего условия устойчивого экономического
роста и повышения конкурентоспособности национальной
экономики
Материалы и методы. В исследовании были использованы
методы анализа состава и структуры, расчет абсолютных
и относительных показателей динамики, средних величин,
сравнительный анализ, графический метод анализа, а также
общетеоретический метод анализа российских и зарубежных
литературных источников. Основными методами исследования
являются кластерный, стохастический и динамический анализ.
В исследовании нашли применение методы сбора первичной
экономической информации, включая анализ законодательных и
нормативно-правовых актов РФ, официальных статистических
данных, данных публичной отчетности отечественных пред-
приятий, анализ прочих открытых источников информации,
системный подход, методы статистического и сравнитель-
ного анализа. В проведенном исследовании рассмотрены труды
Алексеевой Н.А., Барышевой Г.А., Ивантера В., Идрисова А.,
Кайманакова С.В., Комкова Н.И., Кондратьевой Е.В., Коро-
година И.Т., Косициной, Ф.П., Кувалина Д., Кулькова В.М.,
Мамонтова В.Д., Некипелова А., Никулиной О.В., Одер Д.Е.,
Сапир Ж., Юхачева С.П., Янтовского А.А. и других.
Результаты. Придерживаясь мнения известных ученых-эконо-
мистов, будем рассматривать экономический рост как систему
взаимодействия и воспроизводства элементов и факторов
экономического роста национальной экономики. Экономический
рост предопределяют факторы, определяющие темпы и мас-
штабы долгосрочного увеличения валового внутреннего продук-
та, возможности повышения эффективности и качества роста.
Проведенное исследование позволило определить, что основными
факторами влияющими на экономический рост являются про-
изводительность труда, размер средней заработной платы и
объем инвестиций. В свою очередь производительность труда,
также связана с затратами на технологическими инновации
и индексом изобретательности. Одной из качественных харак-
теристик экономического роста является производительность
труда, которая способствует не только увеличению объема
производства, но и повышение уровня доходов населения, обу-
словленного ростом суммарного потребления товаров и услуг.
Заключение. Реализация принципа опережающего роста произ-
водительности труда над ростом заработной платы, должна
учитывать необходимость обеспечения справедливости в диф-
ференциации доходов, что позволит стимулировать развитие
креативного потенциала трудовых ресурсов. Для повышения
производительности труда на предприятиях необходимо вне-
дрять современные методы процессного управления, принципы
организации бережливого производства. Системный подход к
решению названных проблем позволит уже в ближайшие годы
повысить конкурентоспособность предприятий, а в долгосрочной
перспективе будет способствовать планомерному повышению
производительности труда и обеспечивать достойный эконо-
мический рост.
Ключевые слова: экономический рост, производительность
труда, креативный потенциал, качество жизни населения
The purpose of the study is to study the level of labor productivity
as the most important conditions for sustainable economic growth
and improving the competitiveness of the national economy
Materials and methods. The study used the methods of analysis of
composition and structure, the calculation of absolute and relative
indicators of dynamics, average values, comparative analysis, graph-
ical method of analysis, as well as the general theoretical method of
analysis of Russian and foreign literary sources. The main research
methods are cluster, stochastic and dynamic analysis. The study
found application methods for collecting primary economic infor-
mation, including analysis of legislative and regulatory acts of the
Russian Federation, official statistical data, public reporting data of
domestic enterprises, analysis of other open sources of information, a
systematic approach, statistical and comparative analysis methods.
The study examined the works of N.A. Alekseeva, G.A. Barysheva,
V.Ivanter, A.Idrisova, S.V. Kaymanakova, N.I. Komkova, E.V.
Kondratieva, I.T. Korodina, Kositsina, F.P., Kuvalin D., Kulkova
V.M., Mamontova V.D., Nekipolova A., Nikulina OV, Oder D.E.,
Sapir J., Yuhacheva S.P. , Yantovsky A.A. and others.
Results. Adhering to the opinion of well-known economists, we
will consider economic growth as a system of interaction and re-
production of the elements and factors of economic growth of the
national economy. Economic growth is determined by the factors
that determine the pace and scale of the long-term increase in gross
domestic product, the possibility of increasing the efficiency and
quality of growth. The study made it possible to determine that the
main factors affecting economic growth are labor productivity, the
size of the average wage, and the amount of investment. In turn,
labor productivity is also associated with the cost of technological
innovation and the index of ingenuity. One of the qualitative charac-
teristics of economic growth is labor productivity, which contributes
not only to an increase in production, but also to an increase in
the income level of the population due to an increase in the total
consumption of goods and services.
Conclusion. The implementation of the principle of faster growth of
labor productivity over wage growth should take into account the
need to ensure equity in income differentiation, which will stim-
ulate the development of the creative potential of labor resources.
To increase labor productivity in enterprises, it is necessary to
introduce modern methods of process management, the principles
of the organization of lean manufacturing. A systematic approach
to solving these problems will make it possible in the coming years
to improve the competitiveness of enterprises, and in the long term,
it will contribute to a systematic increase in labor productivity and
ensure decent economic growth.
Keywords: economic growth, labor productivity, creative potential,
quality of life of the population
The effect of labor productivity on economic
growth
Е.Ю. Меркулова
Тамбовский государственный технический университет, Тамбов, Россия
Elena Y. Merkulova
Tambov State Technical University, Tambov, Russia
Economic statistics
Statistics and Economics V. 16. № 2. 2019 35
Введение
Проблема обеспечения эко-
номического роста связана с
повышением производитель-
ности труда национальной
экономики. Проведенное ис-
следование показывает, что
существуют разные подходы
к оценке производительно-
сти труда на макро-, мезо- и
микроуровнях экономики,
что затрудняет их сравнение.
Предложенная гипотеза со-
стоит в том, что повышение
экономического роста и каче-
ства жизни населения России
зависит от темпов развития
производительности труда.
Производительность труда по-
казывает, насколько эффекти-
вен трудовой процесс, а задача
ее повышения на микроуров-
не требует поиска направле-
ний совершенствования биз-
нес-процессов в организации.
В результате ожидается общее
повышение результативности
функционирования организа-
ции. С другой стороны рост
производительности труда
является одним из факторов
экономического роста, обеспе-
чивающим улучшение уровня
жизни населения, так как со-
кращение доли живого труда в
производстве материальных и
нематериальных благ обусла-
вливает рост прибыли и сле-
довательно является одной из
качественных характеристики
экономического роста. Следо-
вательно повышение произ-
водительности труда является
актуальной проблемой, реше-
ние которой позволит повы-
сить конкурентоспособность
отечественного производства,
обеспечить рост уровня жизни
населения страны.
Проблемам экономического
роста и смене технологических
укладов посвящены работы
Алексеевой Н.А. [1], Идрисо-
ва А. [2], Клеймана Ю.А. [3],
Кулькова В.М. [4], Каймана-
кова С.В. [4], Тенякова И.М.
[4], Мамонтова В.Д. [5], Оде-
ра Д.Е. [6], Сапира Ж. [7],
Ивантера В. [7], Некипело-
ва А. [7], Кувалина Д. [7], Ши-
рова А.А. [8], Янтовского А.А.
[8], Потапенко В.В [8], Юха-
чева С.П. [9]. Роль производи-
тельности труда в повышении
конкурентоспособности стра-
ны и основные ее детерминан-
ты отражены в исследованиях
Комков Н.И. [10], Кондратье-
вой Е.В. [11], Барышевой Г.А.
[11], Корогодина И.Т. [12],
Косициной Ф.П. [13], Нику-
линой О.В. [14], Королько-
ва К.Н. [14], Перервы О.Л.
[15], Тучиной Ю.В. [15], По-
ловкиной Э.А. [16], Романцо-
ва А.Н. [17], Серопова Л.М.
[18].
Экономический рост явля-
ется ключевой характеристи-
кой общественного производ-
ства в любой экономической
системе. Экономический рост
позволяет решить проблему
ограниченности экономиче-
ских ресурсов и способствует
росту уровня жизни населе-
ния. Наличие экономического
роста в стране означает каче-
ственное и количественное
улучшение общественного
воспроизводства за определен-
ный промежуток времени. Та-
ким образом, экономический
рост, представленный в виде
системы взаимодействия и
воспроизводства факторов
экономического роста, выра-
женный относительным из-
менением реального валового
внутреннего продукта, в пер-
вую очередь отображает эф-
фективные, наиболее дешё-
вые способы распределения
дефицитных ресурсов, с тем,
чтобы обеспечить устойчивое
и расширенное воспроизвод-
ство товаров и услуг. Иными
словами, экономический рост
выражает количественные из-
менения структуры экономики
и взаимосвязей в ней. Поэтому
возникает необходимость вы-
явления основных факторов,
влияющих на экономический
рост, а также рассмотрение
причин их изменения.
В условиях ограниченно-
сти ресурсов для обеспечения
экономического роста перед
управлением стоит задача со-
кращения доли живого и ове-
ществленного труда в ВВП, но
решение данной задачи свя-
зано обеспечением затрат на
накопление в структуре ВВП,
что в свою очередь приведет к
снижению уровня потребления
на душу населения. В резуль-
тате для преодоления сложив-
шейся дилеммы единственным
решением становится необхо-
димость повышения произво-
дительности труда.
Рост производительности
труда зависит от многих фак-
торов (наличия современных
средств производства, качества
используемых предметов тру-
да, рационального использо-
вания человеческого капитала
и т. д). Однако такое упроще-
ние может привести к невер-
ным решениям, так как на
микроуровне, показатели про-
изводительности чаще всего
определяют через выработку,
то есть отношение выручки от
реализации продукции к сред-
несписочной численности ра-
ботников. Следовательно, про-
изводительность в условиях
рынка зависит от востребован-
ности производимой продук-
ции, конкурентоспособности,
платежеспособности населе-
ния. От уровня производи-
тельности труда зависит себе-
стоимость, ценообразование,
что в свою очередь влияет на
конкурентоспособность пред-
приятия, его прибыльность
и платежеспособность. Рост
деловой активности предпри-
ятий позволяет увеличить ва-
ловую добавленную стоимость,
что в свою очередь обеспечит
экономический рост.
Основная часть
Экономический рост явля-
ется ключевой стратегической
задачей государства. Основ-
ным индикатором экономиче-
ского роста является валовый
внутренний продукт (ВВП),
обеспечивающий долгосроч-
ную тенденцию увеличения
реального выпуска на душу
Экономическая статистика
36 Статистика и экономика Т. 16. № 2. 2019
населения. Рассмотрим пока-
затели динамики и структуры
ВВП России за период с 2002
по 2016 годы (рис. 1). Данные
графика показывают, что объ-
ем ВВП в текущих ценах за 14
лет увеличился с 10830,5 млрд.
руб до 86148,6 млрд. руб. Рас-
чет в ценах 2008 года показал,
что объем ВВП вырос в два
раза, то есть ежегодный при-
рост составил 5,4%.
Рассматривая структуру
ВВП в 2017 году, следует от-
метить, что максимальный
удельный вес приходится на
операции с недвижимым иму-
ществом, аренду и предостав-
ление услуг (13,22%), опто-
вую и розничную торговлю
(13,02%), добычу полезных ис-
копаемых (11,92%) (рис. 2).
За период с 2011 по 2017
годы в структуре ВВП прои-
зошли следующие существен-
ные изменения: доля рыболов-
ства и рыбоводства выросла с
0,16% до 9,35%., доля добычи
полезных ископаемых увели-
чилась на 3,71%, доля госу-
дарственного управления и
обеспечения военной безопас-
ности, социального страхова-
ния выросла с 6,02% до 9,39%.
Таким образом, указанные
изменения не связаны с ре-
альным сектором экономики,
а указывают на продолжение
тенденции наращивания ис-
пользования природных иско-
паемых и прироста ВВП в не-
производственной сферы.
Типологическая группиров-
ка регионов России по объему
ВРП на душу населения (табл.
1), указывает на наличие суще-
ственной их дифференциации.
Наиболее высокий уровень
ВРП сложился в Тюменской
области 1627,9 тыс. руб. на че-
ловека. В группу благополуч-
ных регионов также вошли:
г. Москва, Республика Саха
(Якутия); Магаданская, Саха-
линская область; Чукотский
автономный округ.
Самый низкий уровень на-
блюдается в Республике Ингу-
шетия 106,7 тыс. руб., а также
Рис. 1. Динамика ВВП России в рыночных (текущих и сопоставимых ценах
2008 года)
Рис. 2 Структура ВВП России в 2017 году
Таблица 1
Распределение регионов РФ по среднему размеру ВРП
на душу населения (тыс. руб.)
Уровень ВРП
Регионы
высокий (1265,7 тыс. руб) г. Москва; Тюменская область; Республика Саха (Якутия);
Магаданская область; Сахалинская область; Чукотский
автономный округ
средний (428,0 тыс. руб) Белгородская, Воронежская, Калужская, Курская, Липецкая,
Московская, Тульская, Ярославская, Архангельская,
Вологодская, Калининградская, Ленинградская, Мурманская,
Новгородская, Астраханская, Нижегородская, Оренбургская,
Самарская, Свердловская, Челябинская, Иркутская,
Кемеровская, Новосибирская, Омская, Томская, Амурская
области, Республики: Карелия, Коми, Башкортостан,
Татарстан, Хакасия, Удмуртская Республика, г. Санкт-
Петербург, Краснодарский, Пермский, Красноярский,
Камчатский, Приморский, Хабаровский край (39)
низкий (223,9 тыс. руб.) Брянская, Владимирская, Ивановская, Костромская,
Орловская, Рязанская, Смоленская, Тамбовская, Тверская,
Псковская, Волгоградская, Ростовская, Кировская, Пензенская,
Саратовская, Ульяновская, Курганская области, г. Севастополь.
Республики: Адыгея, Калмыкия, Крым, Дагестан, Ингушетия,
Северная Осетия – Алания, Марий Эл, Мордовия, Алтай,
Бурятия, Тыва, Кабардино-Балкарская, Карачаево-Черкесская,
Чеченская, Чувашская Республики, Ставропольский,
Алтайский, Забайкальский края, Еврейская автономная
область (37)
Economic statistics
Statistics and Economics V. 16. № 2. 2019 37
в г. Севастополь, Чеченской,
Карачаево-Черкесской, Кабар-
дино-Балкарской Республиках.
Таким образом, разница между
самым богатым и бедным ре-
гионами составляет 15,3 раза.
Рассмотрим, какие факторы
определяют экономический
рост. Для этого построим ма-
трицу взаимосвязи парных
коэффициентов корреляции
(табл.2).
Из представленных дан-
ных видно, что за исключени-
ем степени износа основных
фондов, между остальными
индикаторами наблюдают-
ся существенные связи. При
этом наиболее существенное
воздействие на ВРП оказыва-
ет фактор производительности
труда (97,4% общей вариации
r2=0,9872). Вторую позицию
занимает начисленная номи-
нальная среднемесячная зар-
плата( 81% вариации) и на
третьем месте объем инвести-
ций в расчете на душу населе-
ния (79,6 % вариации).
Построим двухфакторную
модель зависимости ВРП в
расчете на душу населения и
производительностью труда.
Как видно из данных табл. 3
модель описывается зависимо-
стью
117682 636,57xy x= − +
В уравнении регрессии па-
раметр а0 показывает усред-
ненное влияние на результа-
тивный признак неучтенных
(невыделенных для исследо-
вания) факторов; параметр
b = 636,57 – коэффициент ре-
грессии показывает, что с уве-
личением производительности
труда на 1000 руб. ВРП в рас-
чете надушу населения увели-
чивается на 636,57 руб.
Параметры данного уравне-
ния являются типичными при
уровне значимости 0,05, так
как таблице t-критерия Стью-
дента tтабл = 1,99, меньше рас-
четных значений параметров
модели. Полученная величина
r = 0,987 означает, что в соот-
ветствии со шкалой Чеддока
установленная по уравнению
Таблица 2
Матрица парных коэффициентов корреляции
х1 х2 х3 х4 х5 х6 у
Численность занятых (тыс.
чел.) – х1
1 ,311** ,122 ,277* -,113 ,875** ,339**
,004 ,275 ,012 ,312 ,000 ,002
Номинальная среднемесяч-
ная з/п (руб.) – х2
1 ,692** ,875** -,271* ,404** ,900**
,000 ,000 ,014 ,000 ,000
Инвестиции на душу населе-
ния (руб.) – х3
1 ,913** -,031 ,355** ,892**
,000 ,779 ,001 ,000
Производительность труда
(тыс. руб.) – х4
1 -,098 ,462** ,987**
,383 ,000 ,000
Степень износа основных
фондов, % – х5
1 -,081 -,093
,471 ,404
Затраты на технологические
инновации, млн. руб. – х6
1 ,500**
,000
ВРП на душу населения,
руб. – (у) 1
Примечание: ** корреляция значима на уровне 0,05 * корреляция значима на
уровне 0,01
Таблица 3
Линейная модель зависимости
Модель
Нестандартизованные
коэффициенты т Знач.
B Стандартная ошибка
1 (Константа) -117681,922 10847,172 -10,849 ,000
Производительность
труда (тыс. руб.) 636,567 11,784 54,021 ,000
Условные обозначения:
Области: 1 – Белгородская; 2 – Брянская; 3 – Владимирская; 4 – Воронежская; 5 – Ивановская; 6 – Ка-
лужская; 7 – Костромская; 8 – Курская; 9 – Липецкая; 10 – Московская; 11 – Орловская; 12 – Рязанская;
13 – Смоленская; 14 – Тамбовская; 15 – Тверская; 16 – Тульская; 17 – Ярославская; 21 – Архангельская;
22 – Вологодская; 23 – Калининградская; 24 – Ленинградская; 25 – Мурманская; 26 – Новгородская;
27 – Псковская; 33 – Астраханская; 34 – Волгоградская; 35 – Ростовская; 51 – Кировская; 52 – Нижего-
родская; 53 – Оренбургская; 54 – Пензенская; 55 – Самарская; 56 – Саратовская; 57 – Ульяновская; 58 –
Курганская; 59 – Свердловская; 60 – Тюменская; 61 – Челябинская; 69 – Иркутская; 70 – Кемеровская;
71 – Новосибирская; 72 – Омская; 73 – Томская; 78 – Амурская; 79 – Магаданская; 80 – Сахалинская;
81 – Еврейская АО; 82 – Чукотский АО. Республики: 19 – Карелия; 20 – Коми; 29 – Адыгея; 30 – Калмы-
кия; 31 – Крым; 37 – Дагестан; 38 – Ингушетия; 39 – Кабардино-Балкарская; 40 – Карачаево-Черкесская;
41 – Северная Осетия – Алания; 42 – Чеченская; 44 – Башкортостан; 45 – Марий Эл; 46 – Мордовия; 47 –
Татарстан; 48 – Удмуртская; 49 – Чувашская; 62 – Алтай; 63 – Бурятия; 64 – Тыва; 65 – Хакасия; 74 – Саха
(Якутия); Города: 18 – г. Москва; 28 – г. Санкт-Петербург 36 – г. Севастополь. Край: 32 – Краснодарский,
43 – Ставропольский, 50 – Пермский, 66 – Алтайский; 67 – Забайкальский; 68 – Красноярский; 75 – Кам-
чатский; 76 – Приморский; 77 – Хабаровский.
Рис. 3. Дендограмма распределения регионов РФ
по производительности труда
регрессии связь между произ-
водительностью труда и объ-
емом ВРП в расчете на душу
населения высокая. Оценка
значимости коэффициента
корреляции осуществляется
по F-критерию. Фактическое
значение этого критерия равно
2918, что существенно выше
Fтабл = 3,96, следовательно
уравнение регрессии значимо
при α = 0,05. Коэффициент
Экономическая статистика
38 Статистика и экономика Т. 16. № 2. 2019
эластичности показывает, что
при росте производительности
на 1% ВРП в расчете на душу
населения увеличится на 1,3%.
Результаты кластерного
анализа на основе метода Вар-
да указывают на целесообраз-
ность выделения трех групп
регионов по производительно-
сти труда (рис. 3).
Из данных дендограммы
видно, что распределение ре-
гионов по производительности
труда повторяет в целом типо-
логию по объему ВРП в расче-
те на душу населения. Постро-
им модель влияния факторов
на производительность труда
с учетом пошагового отбора
(табл.4).
Однако полученные параме-
тры не смотря на существен-
ность модели в целом, с эконо-
мической точки зрения влияют
на ВРП не существенно.
У = 37,002 + 0,015х2 +
+ 0,003х3 + 0,001х6
т.е. при увеличении номиналь-
ной заработной платы на 1%
производительность труда вы-
растет на 0,59%, рост инвести-
ций на 1% обеспечивает 0,34%
производительности труда и
затраты на технологические
инновации приводят к росту
производительности труда на
0,02%.
Из расчетов в табл. 5 сле-
дует, что совокупный коэффи-
циент корреляции равен 0,976.
Т.е. можно сказать, что 95,2%
вариации производительности
труда объясняется вариацией
представленных в уравнении
признаков, что указывает на
весьма тесную связь признаков
с результатом.
Оценим надежность урав-
нения регрессии в целом и
показателя связи с помощью
F-критерия Фишера (табл.6).
Фактическое значение F-кри-
терия больше табличного, сле-
довательно уравнение призна-
ется статистически значимым.
Таким образом, из пред-
ставленных расчетов очевидно
о наличии зависимости меж-
ду производительностью тру-
да и экономическим ростом.
«Востребованность результа-
тов труда на рынке оказывает
влияние как на совокупный
объем потребления произве-
денных в стране товаров, так
и на суммарный доход, по-
лучаемый хозяйствующими
субъектами, производящими
востребованную конкуренто-
способную продукцию. Это
свидетельствует о прямой свя-
зи производительности труда с
понятием валового внутренне-
го продукта, поскольку имен-
но по параметрам ВВП можно
дать оценку сбалансированно-
сти потребления результатов
труда и доходов тех, кто этот
результат обеспечивает» [18].
Для решения пробле-
мы роста производительно-
сти труда в 2014 г. на уровне
Правительства РФ был раз-
работан «План мероприятий
по обеспечению повышения
производительности труда,
создания и модернизации вы-
сокопроизводительных рабо-
чих мест, который включает в
Таблица 4
Модель влияния факторов на производительность труда
Модель
Нестандартизован-
ные коэффициенты
Стандарти-
зованные
коэффици-
енты Бета
т Знач.
B
Стандартная
ошибка
1 (Константа) 389,327 29,019 13,416 ,000
Инвестиции на душу
населения (руб.)
,005 ,000 ,913 20,003 ,000
2 (Константа) 23,861 32,464 ,735 ,465
Инвестиции на душу
населения (руб.)
,003 ,000 ,589 16,449 ,000
Номинальня среднеме-
сячная з/п (руб.)
,016 ,001 ,468 13,054 ,000
3 (Константа) 37,002 31,470 1,176 ,243
Инвестиции на душу
населения (руб.)
,003 ,000 ,578 16,725 ,000
Номинальня среднеме-
сячная з/п (руб.)
,015 ,001 ,444 12,575 ,000
Затраты на технологи-
ческие инновации, млн
руб.
,001 ,000 ,077 2,820 ,006
Таблица 5
Оценка существенности параметров модели
Модель R R-квадрат
Скорректированный
R-квадрат
Стандартная
ошибка оценки
Дурбин-
Уотсон
1 ,913a ,833 ,831 181,58923
2 ,973b ,947 ,946 102,84341
3 ,976c ,952 ,950 98,59576 2,004
Таблица 6
Оценка значимости модели
Модель Сумма квадратов ст.св.
Средний
квадрат F Знач.
1 Регрессия 13193192,669 1 13193192,669 400,101 ,000b
Остаток 2637971,889 80 32974,649
Всего 15831164,558 81
2 Регрессия 14995599,914 2 7497799,957 708,893 ,000c
Остаток 835564,643 79 10576,768
Всего 15831164,558 81
3 Регрессия 15072916,842 3 5024305,614 516,844 ,000d
Остаток 758247,716 78 9721,125
Всего 15831164,558 81
Economic statistics
Statistics and Economics V. 16. № 2. 2019 39
себя несколько направлений:
стимулирование инвестиций
с целью обновления и модер-
низации производства; реше-
ние проблем технологического
обновления; оценка рабочих
мест и переоценка основных
фондов; повышение професси-
онализма работников; увеличе-
ние мобильности работающих
граждан; поддержка малого и
среднего предпринимательства»
[19]. Однако заявленная цель
повышения производительно-
сти труда к 2018 году в 1,5 раза
оказалась не решенной (табл. 7).
Рассмотрим динамику по-
казателей производительности
труда по видам экономической
деятельности. Наилучшие ре-
зультаты за период действия
программы обеспечило сель-
ское хозяйство, добыча полез-
ных ископаемых, производство
и распределение электроэнер-
гии, газа и воды. Более десяти
лет снижается производитель-
ность труда в рыболовстве.
Снижение результативности
использования трудовых ре-
сурсов также наблюдается в
обрабатывающем секторе эко-
номики, строительстве, тор-
говле и других видах экономи-
ческой деятельности.
Падение темпов эконо-
мического роста неизбежно
ведет к сокращению количе-
ства рабочих мест, снижению
размера заработной платы,
потери квалификационных
навыков. «Учитывая, что за-
работная плата является не
только важнейшим средством
стимулирования повышения
эффективности материального
производства, но и ценой ра-
бочей силы, ее низкий уровень
не обеспечивает возможности
удовлетворять различные по-
требности работников и, как
следствие, вызывает падение
мотивации к рационально-
му использованию производ-
ственных ресурсов. Это при-
водит к дальнейшему падению
объемов производ ства и сни-
жению качества производимой
продукции: дешевая рабочая
сила никогда не была высоко-
производительной» [20].
Всем хорошо известный
принцип опережающего темпа
роста производительности тру-
да над темпом роста заработной
платы, в условиях не высокого
уровня выработки приводит к
падению доходов населения.
Выполнение данного принци-
па параллельно требует соблю-
дения справедливого распре-
деления доходов, обеспечения
расширенного воспроизвод-
ства рабочей силы, на основе
удовлетворения потребностей
в пище, одежде жилье, фор-
мировании профессиональных
умений и навыков и т.п.
Реализация национального
проекта по повышению про-
изводительности труда тесно
увязана с показателями роста
высокопроизводительных ра-
бочих мест, при этом критерии
оценки являются достаточно
размытыми. В соответствии с
предлагаемой методикой Рос-
стата, «к высокопроизводи-
тельным рабочим местам отно-
сятся все замещенные рабочие
места предприятия (организа-
ции), на которых среднемесяч-
ная заработная плата работ-
ников (для индивидуальных
предпринимателей – средняя
выручка) равна или превышает
установленную величину кри-
терия (пороговое значение)»
[21]. Динамика создания высо-
копроизводительных рабочих
мест представлена на рис. 4.
Данный показатель не может
быть одинаковым для всех ви-
дов экономической деятельно-
сти, поэтому его необходимо
дифференцировать с учетом
среднеотраслевых показателей
деятельности и региональной
Таблица 7
Индекс производительности труда в экономике России в 2005–2017 гг.
2005 2010 2014 2015) 2016 2017
В целом по экономике 105,5 103,2 100,7 98,1 99,7 101,5
сельское хозяйство, охота и
лесное хозяйство 101,8 88,3 103,3 104,5 103,5 103,8
рыболовство, рыбоводство 96,5 97,0 96,1 99,9 95,6 99,8
добыча полезных ископаемых 106,3 104,3 102,8 98,3 100,3 100,4
обрабатывающие производства 106,0 105,2 102,5 97,1 99,3 99,7
производство и распределение
электроэнергии, газа и воды 103,7 103,0 100,2 99,8 100,5 102,2
строительство 105,9 99,6 98,4 100,8 99,9 98,8
оптовая и розничная торговля;
бытовые услуги 105,1 103,6 98,7 93,4 94,4 101,5
гостиницы и рестораны 108,5 101,7 99,8 96,2 94,3 102,2
транспорт и связь 102,1 103,2 100,4 97,8 99,0 101,7
операции с недвижимым
имуществом, аренда и
предоставление услуг 112,4 104,0 98,6 100,2 100,2 99,6
Рис. 4. Динамика создания высокопроизводительных рабочих мест в России
Экономическая статистика
40 Статистика и экономика Т. 16. № 2. 2019
спецификой, так как было по-
казано ранее межрегиональ-
ная дифференциация весьма
велика.
Одной из составляющих
производительности труда яв-
ляется активное использова-
ние современных инноваци-
онных разработок. Однако как
показывают данные в таблице
8 численность организаций за-
нимающихся инновационны-
ми разработками в 2017 году
по сравнению с 2016 годом
снизилась на 88 единиц – это
самый низкий уровень инсти-
туциональных единиц за пери-
од с 2000 по 2017 годы. Анало-
гичная тенденция наблюдается
в динамике с численностью
работников, занимающихся
инновационными разработка-
ми. Данный факт, возможно,
связан с сокращением расхода
бюджетных средств федераль-
ного бюджета на гражданскую
науку с 2,81% до 2,3%. Соот-
ветственно объем финансиро-
вания данных расходов по от-
ношению к ВВП сократился с
0,53% до 0,41%. Если сравнить
темпы роста снижения органи-
заций и работников занятых
в инновационной сфере с ре-
зультатами в сфере инноваций
можно увидеть следующую
картину количество поданных
заявок на выдачу патентов на
изобретения и полезные моде-
ли сократилось в 2016 году по
сравнению с 2016 годом соот-
ветственно на 12,3% и 4,2%.
Положительным моментом яв-
ляется рост количества заявок
на промышленные образцы на
18,7%.
Положительным результа-
том инновационной деятель-
ности является рост объема
выпущенных инновационных
товаров, работ, услуг с 51316
млрд руб. до 57611 млрд руб.,
то есть прирост составил
12,3%. На фоне сокращения
бюджетного финансирования
и не значительного увеличения
внутренних затрат на научные
исследования и разработки
(темп роста 3,2%) такое при-
рост свидетельствует о повы-
шении эффективности инно-
вационных затрат.
Одной из причин негатив-
ной динамики является санк-
ционная политика стран Ев-
росоюза и США, так как одни
из самых эффективных сек-
торов экономики: авиастрое-
ние, судостроение, космиче-
ская отрасль и электронная
промышленность, оказались в
числе наиболее пострадавших.
При этом ключевой пробле-
мой остается низкий спрос
на инновации, как со сторо-
ны частного, так и со сторо-
ны государственного секторов
экономики. Предпринима-
тельский сектор не желает ри-
сковать и приоритетом с его
стороны является закупка им-
портных инновационных тех-
нологий, доказавших свою эф-
фективность. Поэтому в Указе
Президента РФ от 13.05.2017
№ 208 «О Стратегии экономи-
ческой безопасности Россий-
ской Федерации на период до
2030 года» – основными на-
правлениями инновационного
развития России определены:
– импортозамещение на-
учно-экспериментального и
производственного оборудо-
вания, информационно-ком-
муникационных технологий,
селекционных и генетических
достижений;
– повышение конкуренто-
способности российской эко-
номики на основе интегра-
ции научно-производственных
кластеров;
– создание и устойчивое
развитие перспективных вы-
сокотехнологичных секторов
экономики;
– развитие цифровой эко-
номики, применение техно-
логий глубокой переработки
материалов, производство то-
варов с высокой добавленной
стоимостью и т.д.
Для решения поставленных
задач необходимо активизиро-
вать население, путем созда-
ния условий для реализации
его креативного потенциала.
Таблица 8
Основные показатели инновационной деятельности России за 2000–2017
гг.
Показатели 2000 2005 2010 2015 2016 2017
Число организаций (ед.) 4099 3566 3492 4175 4032 3944
Численность персонала, занято-
го исследованиями и разработка-
ми (тыс. чел) 887,7 813,2 736,5 738,9 722,3 707,9
Расходы на гражданскую науку
из средств федерального бюдже-
та, млрд. руб. 17,4 76,9 237,6 439,4 402,7 377,9
в % к валовому внутреннему
продукту 0,24 0,36 0,51 0,53 0,47 0,41
в % к расходам федерального
бюджета 1,69 2,19 2,35 2,81 2,45 2,3
Внутренние затраты на науч-
ные исследования и разработки,
млрд. руб.
в фактически действовавших
ценах 76,7 230,8 523,4 914,7 943,8 1019,1
к процентах к валовому вну-
треннему продукту 1,05 1,07 1,13 1,10 1,10 1,11
Подано заявок на :
выдачу патентов на изобрете-
ния (ед) 32 254 42 500 45 517 41 587 36454
выдачу патентов на полезные
модели (ед) 9 473 12 262 11 906 11 112 10643
на промышленные образцы (ед) 3 917 3 997 4 929 5 464 6487
Используемые передовые техно-
логии (ед) 70069 140983 203330 218018 232388 240054
Объем инновационных товаров,
работ, услуг (млрд руб.) 25795 45525 51316 57611
Economic statistics
Statistics and Economics V. 16. № 2. 2019 41
В современной статистике
одним из показателей, харак-
теризующих инновационную
активность населения явля-
ется – коэффициент изобре-
тательской активности (отно-
шение числа отечественных
патентных заявок на изобре-
тения поданные в России, в
расчете на 10 тыс. чел. населе-
ния. В связи с этим возникает
интерес к выявлению регионов
с наличием высокого креатив-
ного потенциала.
Для построения типологи-
ческой группировки по коэф-
фициенту изобретательской
активности (табл. 9) проведем
кластерный анализ на осно-
ве метода Варда (расчеты вы-
полнены с использованием
программного продукта SPSS
Statistica. Результаты постро-
ения дендограммы указывают
на наличие трех выраженных
скачков, что позволяет сфор-
мировать три группы регионов:
с высоким уровнем креативно-
сти – 6 регионов; с мезамор-
фным уровнем креативности –
39 регионов; с низким уровнем
креативности – 37 регионов.
Для регионов первой груп-
пы характерны максимальные
результаты инновационной
деятельности, так как в нее
входят два самых крупных ме-
гаполиса России г. Москва и
Санкт-Петербург. В частности
коэффициент изобретательно-
сти составляет 4,14, а иннова-
ционная активность органи-
заций 10,2%. Большая часть
регионов России в среднем
имеет 6,7 тыс. человек, зани-
мающихся инновационными
исследованиями и разработ-
ками, внутренние затраты на
научные исследования и раз-
работки в них в среднем со-
ставляют 8,5 млн.руб., объем
инновационных товаров, работ
и услуг находится в пределах
57,2 млрд.руб. При средней
инновационной активности
8%, в данной группе этот по-
казатель составил 8,9%.
Однако 45% регионов Рос-
сии имеют не высокие харак-
теристики креативности. Ко-
эффициент изобретательской
активности в данной группе
составляет 0,46, и колеблется
от 0 до 0,73, то есть не пре-
вышает 73 патентных заявок
на 1 миллион человек. Сред-
няя численность работников в
сфере научных исследований
составляет 1424 человека, вну-
тренние затраты на научные
исследования и разработки со-
ставляют 1,4 млрд руб. Объем
инновационных товаров, ра-
бот, услуг ниже среднероссий-
ских показателей на 38 млрд
руб., а инновационная актив-
ность организаций в среднем
составляет 6,2%.
Таким образом, проведен-
ный анализ показал, что уро-
вень инновационной актив-
ности в России не достаточно
высок, причем наблюдается
существенная дифференциа-
ция в уровне инновационного
развития на уровне регионов.
Одной из причин сложившей
ситуации является неоднород-
ная концентрация информаци-
онно-коммуникационных, фи-
нансовых, интеллектуальных и
организационных ресурсов.
Заключение
Проведенное исследование
показывает, что экономиче-
ский рост является ключевой
стратегической задачей госу-
дарства, его основной харак-
теристикой является валовый
внутренний продукт. Изучение
структуры ВВП указывает на
необходимость ее трансфор-
мации в пользу материальных
видов производства, а типо-
логическая группировка реги-
онов России по объему ВРП
на душу населения указывает
на наличие существенной их
дифференциации, что вызы-
вает необходимость снижения
их дифференциации. Анализ
факторов влияющих на уро-
вень ВРП показал, что наибо-
лее существенное воздействие
на ВРП оказывает фактор про-
изводительности труда, вторую
позицию занимает начислен-
ная номинальная среднемесяч-
ная зарплата и на третьем месте
объем инвестиций в расчете на
душу населения. Коэффици-
ент эластичности показывает,
что при росте производитель-
ности на 1% ВРП в расчете на
душу населения увеличится на
1,3%. Таким образом доказано,
что производительность труда
является ключевой характери-
стикой экономического роста.
Изучение причин измене-
ния производительности труда
по регионам России показа-
Таблица 9
Типологическая группировка регионов России по уровню креативности
Уровень
креативности
Регионы
низкий Брянская, Костромская, Липецкая, Смоленская, Архангельская,
Вологодская, Калининградская, Ленинградская, Мурманская,
Псковская, Кировская, Оренбургская, Курганская, Тюмен-
ская, Кемеровская, Сахалинская области, Республики: Каре-
лия, Коми, Адыгея, Калмыкия, Крым, Дагестан, Ингушетия,
Мордовия, Алтай, Бурятия, Тыва, Хакасия, (Якутия), Кабарди-
но-Балкарская, Карачаево-Черкесская, Чеченская Республики,
Ставропольский, Алтайский, Забайкальский, Камчатский край,
Чукотский автономный округ (37).
средний Белгородская, Владимирская, Воронежская, Калужская, Ор-
ловская, Рязанская, Тамбовская, Тверская, Саратовская, Туль-
ская, Ярославская, Новгородская, Волгоградская, Астраханская,
Свердловская, Ростовская, Пензенская, Челябинская, Нижего-
родская, Самарская, Ульяновская, Иркутская, Новосибирская,
Омская, Амурская, Магаданская области, Еврейская АО, Ре-
спублики: Татарстан, Северная Осетия – Алания, Марий Эл,
Башкортостан, Чувашская и Удмуртская Республики, Красно-
дарский, Пермский, Красноярский, Приморский, Хабаровский
край, г. Севастополь (39)
высокий Ивановская, Курская, Московская, Томская области, г. Москва,
г. Санкт-Петербург (6)
Экономическая статистика
42 Статистика и экономика Т. 16. № 2. 2019
ло, что при увеличении номи-
нальной заработной платы на
1% производительность труда
вырастет на 0,59%, рост ин-
вестиций на 1% обеспечива-
ет 0,34% производительности
труда и затраты на техноло-
гические инновации приводят
к росту производительности
труда на 0,02%. Следователь-
но для решения поставленных
задач необходимо активизиро-
вать население, путем созда-
ния условий для реализации
его креативного потенциала.
Однако проведенный анализ
показал, что уровень иннова-
ционной активности в Рос-
сии не достаточно высок, при
чем наблюдается существен-
ная дифференциация в уров-
не инновационного развития
на уровне регионов. Одной
из причин сложившей ситу-
ации является неоднородная
концентрация информацион-
но-коммуникационных, фи-
нансовых, интеллектуальных
и организационных ресурсов.
Для активизации инноваци-
онных процессов в регионах
были приняты стандарты по
развитию конкурентоспособ-
ности, одной из основных за-
дач которых является во влече-
ние широких масс населения
и предприятий по созданию
конкурентоспособной продук-
ции, в первую очередь ориен-
тируясь на креативный потен-
циал регионов.
Традиционно предлагаемые
меры активизации инвести-
ций, с целью повышения про-
изводительности труда также
носят противоречивый харак-
тер, так как инвесторы готовы
вкладывать в быстроокупае-
мые, конкурентоспособные
проекты, а виды деятельности
с длительным сроком окупа-
емости продолжают терять
свою инвестиционную привле-
кательность. В данном случае
необходим государственный
подход, определяющий прио-
ритеты развития экономики,
с системой соответствующих
мер поддержки.
Таким образом, современ-
ные процессы общественного
воспроизводства обеспечивают
экономический рост за счет ин-
фляционного сдерживания ре-
альной заработной платы, а не
соблюдения закона роста про-
изводительности труда на осно-
ве инноваций, совершенствова-
ния технологий производства.
Литература
1. Алексеева Н. А., Корняков В. И. Дири-
жизм и воспроизводственный поворот // Теоре-
тическая экономика. 2015. № 1. С. 24–35.
2. Идрисов А. Производительность труда как
ключевой фактор конкурентоспособности рос-
сийской экономики. [Электрон. ресурс] Режим
доступа: http://strategy.ru/research/3/.
3. Клейман Ю.А. Смена технологических
укладов на основе внедрения инноваций как фак-
тор технико-экономического развития // Эконо-
мический вестник Ростовского государственного
университета. 2008. Т. 6. № 1. Ч. 2. С. 164–168.
4. Кульков В.М., Кайманаков С.В., Теня-
ков И.М. Экономический рост в России: на-
циональная модель, качество и безопасность //
Национальные интересы: приоритеты и безо-
пасность. 2014. №38 (275) С. 9–19.
5. Мамонтов В.Д. Современные реалии эко-
номического роста в России // Глобальные
проблемы модернизации национальной эко-
номики. Материалы VII Международной науч-
но-практической конференции, Тамбов, 19–20
апреля 2018 г. Тамбов: Издательский дом «Дер-
жавинский», 2018. С. 21–31.
6. Одер Д.Е. Стратегическое планирование в
области производительности труда и экономи-
ческого роста в России // Вестник Магнитогор-
ского государственного технического универси-
тета им. Г.И. Носова. 2013. № 2. С. 85–87.
7. Сапир Ж., Ивантер В., Некипелов А., Ку-
валин Д. Российская трансформация: 20 лет
спустя. М.: Магистр, 2013.
8. Широв А.А., Янтовский А.А., Потапен-
ко В.В. Оценка потенциального влияния санк-
ций на экономическое развитие России и ЕС //
Проблемы прогнозирования. 2015. № 4. С. 3–16.
9. Юхачев С.П. Экономическая политика и
экономический рост: новые реалии и перспек-
тивы // Саяпинские чтения сборник материа-
лов круглого стола. 19 января 2017 г., Тамбов:
ТГУ им Г.Р. Державина, 2017. С. 437–443.
10. Комков Н.И. Инновационная модерни-
зация и технологическое развитие: отказ или
корректировка стратегии? // Модернизация,
инновации, развитие (МИР). 2013. № 3(15).
С. 3–11.
11. Кондратьева Е.В., Барышева Г.А. Гене-
зис производительного труда. [Электрон. ре-
сурс]. Режим доступа: http://naukovedenie.ru/
PDF/18EVN315 .
12. Корогодин И. Т., Гапонова С. Н. Прин-
ципы анализа соотношения роста заработной
платы и производительности труда // Теория и
практика общественного развития. 2015. № 8.
С. 42–45.
13. Косицина Ф. П. Соотношение роста про-
изводительности труда и заработной платы –
критерий эффективности государственного
регулирования экономики // Бизнес. Образова-
ние. Право. Вестник Волгоградского института
бизнеса. 2010. № 2 (12). С. 19–23.
14. Никулина О.В., Корольков К.Н. Факто-
ры роста производительности труда в экономи-
ке России // Качество. Инновации. Образова-
ние. 2012. № 10. С. 84–91.
15. Перерва О.Л., Тучина Ю.В. Уровень про-
изводительности труда в России // Экономика
и современный менеджмент: теория и практика.
2014. № 40. С. 86–90.
Economic statistics
Statistics and Economics V. 16. № 2. 2019 43
References
1. Alekseyeva N.A., Kornyakov V.I. Dirigisme
and Reproductive Turn. Teoreticheskaya ekonomi-
ka = Theoretical Economics. 2015; 1: 24–35. (In
Russ.)
2. Idrisov A. Proizvoditel’nost’ truda kak kly-
uchevoy faktor konkurentosposobnosti rossiyskoy
ekonomiki = Labor productivity as a key factor in
the competitiveness of the Russian economy. [Inter-
net] URL: http://strategy.ru/research/3/. (In Russ.)
3. Kleyman YU.A. Change of technological struc-
tures on the basis of the introduction of innovations
as a factor of technical and economic development.
Ekonomicheskiy vestnik Rostovskogo gosudarstven-
nogo universiteta = Economic Bulletin of Rostov
State University. 2008; 6; 1 (2): 164–168. (In Russ.)
4. Kul’kov V.M., Kaymanakov S.V., Tenyak-
ov I.M. Economic growth in Russia: the national
model, quality and safety. Natsional’nyye interesy:
prioritety i bezopasnost’ = National interests: pri-
orities and safety. 2014; 38 (275): 9–19. (In Russ.)
5. Mamontov V.D. Modern realities of econom-
ic growth in Russia. Global’nyye problemy moderni-
zatsii natsional’noy ekonomiki Materialy VII Mezh-
dunarodnoy nauchno-prakticheskoy konferentsii =
Proceedings of the VII International Scientific and
Practical Conference, Tambov, April 19–20; 2018.
Tambov: Derzhavinsky Publishing House; 2018:
21–31. (In Russ.)
6. Oder D.E. Strategic planning in the field of
labor productivity and economic growth in Russia.
Vestnik Magnitogorskogo gosudarstvennogo tekh-
nicheskogo universiteta im. G.I. Nosova = Bulletin
of Nosov Magnitogorsk State Technical University.
2013; 2: 85–87. (In Russ.)
7. Sapir ZH., Ivanter V., Nekipelov A., Kuva-
lin D. Rossiyskaya transformatsiya: 20 let spustya
= Russian transformation: 20 years later. Moscow:
Master; 2013. (In Russ.)
8. SHirov A.A., YAntovskiy A.A., Potapen-
ko V.V. Assessment of the potential impact of sanc-
tions on the economic development of Russia and
the EU. Problemy prognozirovaniya = Problems of
Forecasting. 2015; 4: 3-16. (In Russ.)
9. YUkhachev S.P. Economic Policy and Eco-
nomic Growth: New Realities and Prospects. Sayap-
inskiye chteniya sbornik materialov kruglogo stola.
= Sayapinskie readings, collection of materials of
the round table. January 19; 2017. Tambov: Der-
zhavin TSU; 2017: 437-443. (In Russ.)
10. Komkov N.I. Innovative modernization and
technological development: failure or adjustment of
the strategy?. Modernizatsiya, innovatsii, razvitiye
(MIR) = Modernization, innovation, development
(MIR). 2013; 3(15): 3-11. (In Russ.)
11. Kondrat’yeva E.V., Barysheva G.A. Genezis
proizvoditel’nogo truda = The genesis of produc-
tive labor. [Internet]. URL: http://naukovedenie.ru/
PDF/18EVN315 . (In Russ.)
12. Korogodin I; , Gaponova S. N. Principles of
analysis of the ratio of wage growth and labor pro-
ductivity. Teoriya i praktika obshchestvennogo raz-
vitiya = Theory and practice of social development.
2015; 8: 42–45. (In Russ.)
13. Kositsina F: Sootnosheniye rosta proiz-
voditel’nosti truda i zarabotnoy platy – kriteriy
effektivnosti gosudarstvennogo regulirovaniya
ekonomiki. Biznes. Obrazovaniye. Pravo. Vestnik
Volgogradskogo instituta biznesa = Business. Edu-
cation. Law. Bulletin of the Volgograd Institute of
Business. 2010; 2 (12): 19–23. (In Russ.)
14. Nikulina O.V., Korol’kov K.N. Growth fac-
tors of labor productivity in the Russian economy.
Kachestvo. Innovatsii. Obrazovaniye = Quality. In-
novation. Education. 2012; 10: 84–91. (In Russ.)
15. Pererva O.L., Tuchina YU.V. The level of
labor productivity in Russia. Ekonomika i sovre-
mennyy menedzhment: teoriya i praktika = Eco-
nomics and modern management: theory and prac-
tice. 2014; 40: 86–90. (In Russ.)
16. Polovkina E.A. Economic factors of produc-
tivity growth. Vestnik ekonomiki, prava i sotsiologii
= Bulletin of Economics, Law and Sociology. 2008;
2: 45–51. (In Russ.)
17. Romantsov A.N. The economic essence of
labor productivity in solving the problem of its in-
crease. Modeli, sistemy, seti v ekonomike, tekhnike,
prirode i obshchestve = Models, systems, networks
16. Половкина Э.А. Экономические факторы
роста производительности труда // Вестник эко-
номики, права и социологии. 2008. № 2. С. 45–51.
17. Романцов А.Н. Экономическая сущность
производительности труда в решении пробле-
мы ее повышения // Модели, системы, сети в
экономике, технике, природе и обществе. 2016.
№4 (20). С. 44–58.
18. Серопов Л.М. Статистическое иссле-
дование социально-экономических факторов
роста производительности труда в трансфор-
мирующейся экономике России. Йошкар-Ола:
МОСИ, 2009. 258 с.
19. Распоряжение Правительства РФ № 1250-р
от 9 июля 2014 г. 6. Официальный сайт Феде-
ральной службы государственной статистики.
[Электрон. ресурс]. Режим доступа: http://www.
gks.ru/wps/wcm/connect/rosstat_main/rosstat/ru/
statistics/efficiency/
20. Эффективная Россия: производитель-
ность как фундамент роста. [Электрон. ре-
сурс]. Режим доступа: http://gtmarket.ru/files/
news/1986/
21. Методика расчета показателя «прирост
высокопроизводительных рабочих мест, в про-
центах к предыдущему году» [Электрон. ресурс].
Режим доступа: http://de.gov.yanao.ru/doc/o_
effect/ oiv/fsgs_pr449_14112013 .
Экономическая статистика
44 Статистика и экономика Т. 16. № 2. 2019
Сведения об авторе
Елена Юрьевна Меркулова
Д.э.н., профессор, кафедра «Экономическая
безопасность и качество»
Тамбовский государственный технический
университет, Тамбов, Россия
Эл. почта: merkatmb@mail.ru
Information about the author
Elena Y. Merkulova
Dr. Sci. (Economics), Professor, Department of
Economic Security and Quality
Tambov State Technical University,
Tambov, Russia
E-mail: merkatmb@mail.ru
in the economy, technology, nature and society.
2016; 4(20): 44-58. (In Russ.)
18. Seropov L.M. Statisticheskoye issledo-
vaniye sotsial’no-ekonomicheskikh faktorov rosta
proizvoditel’nosti truda v transformiruyushcheysya
ekonomike Rossii = Statistical study of the so-
cio-economic factors of labor productivity growth
in the transforming Russian economy. Yoshkar-Ola:
IOSI; 2009. 258 p. (In Russ.)
19. Rasporyazheniye Pravitel’stva RF № 1250-r
ot 9 iyulya 2014 g. 6. Ofitsial’nyy sayt Federal’noy
sluzhby gosudarstvennoy statistiki = Order of the
Government of the Russian Federation No. 1250-p
dated July 9; 2014. Official website of the Feder-
al State Statistics Service. [Internet]. URL: http://
www.gks.ru/wps/wcm/connect/rosstat_main/ross-
tat/ru/statistics/efficiency/ (In Russ.)
20. Effektivnaya Rossiya: proizvoditel’nost’ kak
fundament rosta = Effective Russia: productivity as
a foundation of growth. [Internet]. URL: http://gt-
market.ru/files/news/1986/ (In Russ.)
21. Metodika rascheta pokazatelya «prirost vy-
sokoproizvoditel’nykh rabochikh mest, v prot-
sentakh k predydushchemu godu» = The method
of calculating the indicator «increase in high-per-
formance jobs, as a percentage of the previous year»
[Internet]. URL: http://de.gov.yanao.ru/doc/o_ef-
fect/ oiv/fsgs_pr449_14112013 . (In Russ.)
