Federal Reserve Bank of Minneapolis
Quarterly Review Spring 1992
The Labor Market in Real Business Cycle Theory *
Gary D. Hansen
Professor of Economics
University of California,
Federal Reserve Bank of Minneapolis
and Professor of Economics
University of Pennsylvania
The basic objective of the real business cycle research program is to use the neoclassical growth model to interpret observed patterns of fluctuations in overall economic activity. If we take a simple version of the model, calibrate it to be consistent with long-run growth facts, and subject it to random technology shocks calibrated to observed Solow residuals, the model displays short-run cyclical behavior that is qualitatively and quantitatively similar to that displayed by actual economies along many important dimensions. For example, the model predicts that consumption will be less than half as volatile as output, that investment will be about three times as volatile as output, and that consumption, investment, and employment will be strongly positively correlated with output, just as in the postwar U. S. time series. 1 In this sense, the real business cycle approach can be thought of as providing a benchmark for the study of aggregate fluctuations.
In this paper, we analyze the implications of real business cycle theory for the labor market. In particular, we focus on two facts about U. S. time series: the fact that hours worked fluctuate considerably more than productivity and the fact that the correlation between hours worked and productivity is close to zero. 2 These facts and the failure of simple real business cycle models to account for them have received considerable attention in the literature. [See, for example, the extended discussion by Christiano and Eichenbaum (1992) and the references they provide. ] Here we first document the facts. We
The Editorial Board for this paper was V. V. Chari, Preston J. Miller, Richard Rogerson, and Kathleen S. Rolfe.
then present a baseline real business cycle model (essentially, the divisible labor model in Hansen 1985) and compare its predictions with the facts. We then consider four extensions of the baseline model that are meant to capture features of the world from which this model abstracts. Each of these extensions has been discussed in the literature. However, we analyze them in a unified framework with common functional forms, parameter values, and so on, so that they can be more easily compared and evaluated in terms of how they affect the model’s ability to explain the facts.
The standard real business cycle model relies exclusively on a single technology shock to generate fluctuations, so the fact that hours worked vary more than productivity implies that the short-run labor supply elasticity must be large. The first extension of the model we consider is to recognize that utility may depend not only on leisure today but also on past leisure; this possibility leads us to introduce nonseparable pref-
* This paper is also available in Spanish in Cuadernos Economicos de ICE, a quarterly publication of the Ministerio de Economía y Hacienda. The paper appears here with the permission of that publication’s editor, Manuel Santos.
1 These properties are also observed in other countries and time periods. See Kydland and Prescott 1990 for an extended discussion of the postwar U. S. data, and see Blackburn and Ravn 1991 or Backus and Kehoe, forthcoming, for descriptions of other countries and time periods.
2 Although we concentrate mainly on these cyclical facts, we also mention an important long-run growth fact that is relevant for much of our discussion: total hours worked per capita do not display trend growth despite large secular increases in average productivity and real wages.