Accumulation, distribution and employment: a structural VAR approach to a Kaleckian macro model
Introduction
This paper investigates the relation between effective demand, income distribution and unemployment empirically. Kaleckians, or if the reader prefers: post-Keynesians, differ fundamentally from neoclassical economists in how they perceive of the interaction between the goods market and the labor market. For Kaleckians unemployment is the result of demand deficiencies on the goods markets. Thus goods market variables determine labor market behavior. For neoclassical economics, at least in its textbook version, unemployment is a labor market phenomenon. There exists a real wage which would clear the labor market. Unemployment will arise if the real wage is too high, which will usually be due to frictions caused by governments, labor unions or insiders. Needless to say, present incarnations of the neoclassical theory are more sophisticated and may at times even give different results. The focus of this paper is on the Kaleckian model and neoclassical arguments will only be discussed in passing.
A Kaleckian macroeconomic model is presented and estimated by means of a structural vector autoregression (VAR) approach. The VAR consists of accumulation, capacity utilization, the profit share, unemployment and the growth of labor productivity. In the VAR methodology, each dependent variable is regressed on the lagged values of all other variables in the system. Due to this flexible modeling strategy the VAR allows for Kaleckian as well as non-Kaleckian effects (of the lagged variables). Only in the contemporaneous interactions of the variables is the post-Keynesian model imposed. The model is estimated for the USA, UK and France.1 The time period under investigation ranges from the early 1960s to the late 1990s.
Hypotheses are derived from the Kaleckian model and explored empirically. Thus the following questions are addressed: Is unemployment driven by goods market variables? What is the effect of changes in the income distribution on accumulation? Does technological progress increase unemployment? In addition to the hypotheses derived from the Kaleckian model, two rather general neoclassical hypotheses are explored. The corresponding questions are: Does an increase in wages cause unemployment? Does an increase in wages lead to substitution effects? Because of the general nature of the VAR methodology, such effects can exist. Since these effects are not derived from an explicit model, but tested ad hoc, no claim is made that a neoclassical model be tested. Rather the empirical relevance of two mechanisms often stressed in the neoclassical literature is evaluated, in addition to the test of the Kaleckian model.
The paper is structured as follows. Section 2 presents the model. Section 3 summarizes the hypotheses to be investigated. Section 4 discusses the econometric method and various data issues. Section 5 presents the econometric results for the tests performed. Section 6 compares the findings with those of earlier research and, finally, Section 7 derives conclusions.
Section snippets
The model
The goods market part of the model consists of behavioral functions for investment, savings, and net exports, and is based on Marglin and Bhaduri (1990), who proposed a flexible neo-Kaleckian model that allows for profit-led as well as for wage led growth regimes. We will explicitly derive the conditions of profit-led vs. wage led growth regimes in our extended model, once we have introduced the full structure. The goods market block of the Marglin-Bhaduri model is complemented by a
Hypotheses
The model estimated is a VAR model, thus past values of all variables are allowed to influence present values of any variable. Thus results that are not in accordance with the structural model outlined above are possible due to lagged effects. The structural model provides the motivation and shapes the interaction of the contemporaneous effects only. Thus it will be useful to summarize the hypotheses to be explored empirically. All hypotheses except the last two, which deal with neoclassical
Econometric method
VAR methodology has become popular among economists since the early 1980s. Originally it had been developed as an alternative to theory-based structural estimation. In a seminal paper Sims (1980) presented VAR analysis as an atheoretical tool because it had no restrictions on the explanatory variables and did not rely on strict exogenous-endogenous distinction. However, few economists and econometricians today hold on to such far reaching claims. The importance of the ordering of variables for
Empirical results
The VAR was estimated with four lags. Lag length tests (see Appendix A) indicated that two lags would be sufficient. However, given that some economic variables, in particular investment, may take longer than a year to respond to changes in economic conditions, it was decided to use a lag length of four. Results hardly differ between the two specifications.
Autocorrelation LM tests indicate that autocorrelation is not a major problem (see Appendix A). However, the null hypothesis of
Comparison with existing empirical literature
The existing empirical macroeconomic literature on Kaleckian model has focused on the goods market. In particular, the model by Marglin and Bhaduri (1990) has inspired empirical research since profit-led as well as wage-led regimes are allowed for by theory.
Bowles and Boyer (1995) focus on the question of how distributional changes affect output. They estimated four equations, investment, profit share, savings and net exports by means of a single equation approach. They find that France,
Conclusion
A structural VAR system consisting of accumulation, capacity utilization, the profit share (as a proxy for income distribution), unemployment and the growth of labor productivity was estimated, based on a Kaleckian macro model. The results suggest that accumulation impacts strongly upon capacity utilization and both, accumulation and capacity utilization, have significant effects on unemployment. Thus the basic Keynesian story is confirmed: goods market variables have a strong impact on
Acknowledgements
The authors are grateful to Dieter Gstach, Werner Hölzl, Werner Müller, Pekka Sauramo, Alfred Stiassny and two anonymous referees for comments that helped to improve the paper. Earlier versions of the paper have been presented at the 3rd Annual Conference of the Association for Heterodox Economics, 7–8 July 2001, London, and the conference ‘Old and New Growth Theories: an Assessment’, Oct. 2001, Pisa, and the present version benefited from the discussions there. The usual disclaimers apply.
References (38)
- et al.
Income shares and capital formation: patterns of recent developments
Journal of Income Distribution
(1997) - Amisano, G., Giannini, C., 1997. Topics in Structural VAR Econometrics, second edition. Springer,...
- Bean, C., 1994. European unemployment: a survey. Journal of Economic Literature XXXII,...
- Bernanke, B., 1986. Alternative Explanation of the Money Income Correlations. Carnegie Rochester Conference Series on...
- et al.
Unemployment and the real wage: the economic basis for contesting political ideologies
Cambridge Journal of Economics
(1990) - Blanchflower, D., Oswald, A., 1994. The Wage Curve. MIT Press, Cambridge,...
International competition, income distribution and economic growth
Cambridge Journal of Economics
(1989)- Blecker, R., 2002. Distribution, demand and growth in neo-Kaleckian macro-models. In: Setterfiled (Ed.), The Economics...
- Bowles, S., Boyer, R., 1995. Wages, aggregate demand, and employment in an open economy: an empirical investigation....
- Boyer, R., 1988. Formalizing growth regimes. In: Dosi, G., Freeman, C., Nelson, R. (Eds.), Technical Change and...
Labour institutions and economic growth: a survey and a ‘régulatioinst’ approach
Labour
Stagnation, income distribution and monopoly power
Cambridge Journal of Economics
Business investment: recent performance and some implications for policy
OECD Economic Studies
Cited by (113)
Growth, cycles, and residential investment
2024, Structural Change and Economic DynamicsThe effects of social infrastructure and gender equality on output and employment: The case of South Korea
2022, World DevelopmentCitation Excerpt :In order to tackle the multi-dimensional complex endogenous relations between wages, output, productivity and social expenditures, we use SVAR as the estimation methodology. SVAR as a methodology to estimate the effect of functional income distribution or wages on output have been previously used by Stockhammer and Onaran (2004), Onaran and Stockhammer (2005), Barbosa-Filho and Taylor (2006), Kiefer and Rada (2015), and Jump and Mendieta-Muñoz (2017). The advantage of this approach is that the interaction between the variables can be incorporated and it allows for tracing the effects through an entire system rather than analysing one equation at a time.
Testing for Granger causality in quantiles between the wage share in income and productive capacity utilization
2022, Structural Change and Economic DynamicsCitation Excerpt :Using a system or aggregative approach, which relies on estimation of a reduced form solution (usually) for output, the following representative results can be mentioned. Stockhammer and Onaran (2004) estimate a VAR model for the U.S., the U.K. and France and find that shocks to the profit share have no significant overall effects on capacity utilization. Nevertheless, Barbosa-Filho and Taylor (2006) detect that the U.S. economy featured capacity utilization varying negatively with the wage share using a VAR model for 1948–2002 and several sub-periods.
International competition, income distribution, and North-South uneven development under the balance of payments constraint
2022, Structural Change and Economic DynamicsThe Macroeconomic Effects of Financialization and the Wage Gap between Blue and White Collar Workers
2021, Structural Change and Economic DynamicsCitation Excerpt :Using a sample for the years 1962-2007, Onaran et al. (2011) find that the US economy is in a modestly wage-led demand regime. Stockhammer and Onaran (2004) find weak evidence for wage-led demand regime for the USA, the UK and France for the 1960s to the late 1990s. Onaran and Galanis (2013) and Stockhammer and Wildauer (2015) are among others who find the US economy is in a wage-led demand regime.
Falling labour share and anaemic growth in Portugal: A post-Keynesian econometric analysis
2023, Economic and Labour Relations Review