Elsevier

Economics Letters

Volume 99, Issue 1, April 2008, Pages 55-58
Economics Letters

Layoffs and lemons over the business cycle

https://doi.org/10.1016/j.econlet.2007.05.028Get rights and content

Abstract

This paper develops a simple model in which unemployment arises from a combination of selection and bad luck. During recessions, the proportion of workers who are laid off due to low productivity declines during recessions, diminishing the adverse signaling effect of an unemployment spell. Wage regressions estimated using the Displaced Workers Supplement support this basic prediction of the model.

Introduction

In the labor economics literature, layoffs are often viewed as a selective weeding process. For example, Laing (1994) and Gibbons and Katz (1991) develop signaling models in which firms may choose to lay off workers when they are discovered to be of low ability.1 In contrast, macroeconomists have typically focused on the type of unemployment that arises because the wage is set above the market-clearing level, but unemployed and employed workers are otherwise indistinguishable. In these models, unemployed workers are simply those who were unlucky enough to be laid off. For example, Gray's (1978) contracting model and efficiency wage models such as Shapiro and Stiglitz's (1984) model have this feature. These models have the advantage that they account more naturally than models in which unemployment is generated solely from asymmetric information for the striking tendency of unemployment to rise in recessions and fall in booms.2 However, they also lack the reasonable feature of asymmetric information models that unemployment falls disproportionately on workers who have unexpectedly low productivity. There are few models that fall between these extremes.

In this paper, I consider what happens when layoffs result from a combination of bad luck and selection. One implication of this model is that the proportion of unemployed workers who are laid off due to low productivity varies over the business cycle. The intuition is simple. Suppose that some workers are laid off in every period due to low productivity. However, the number of workers laid off simply due to bad luck is higher in recessions than in booms. In this case, workers laid off during recessions are, on average, less adversely selected than those laid off in booms. Furthermore, if the workers' productivities are not observable to firms then firms draw inferences about the workers' productivities from their past work experience. Since the signaling effect of an unemployment spell varies over the business cycle, so do the wage losses associated with unemployment.

Section snippets

Model

Consider the following simple model of unemployment over the business cycle. The model consists of firms and workers. There are two unobservable types: good and bad. Only “good” type workers contribute to production. The population fraction of good types is given by θ. The total amount of labor available in the economy is normalized to 1.

Firms produce according to the production function,Yt=AtF(Lt),

where Lt denotes the “productive” labor employed by the firm (defined below), and At is an

Wage change regressions

I estimate a linear regression model of the form,Δlogwt=β0+αu+βx+ε,where ▵ log(wi) is the difference in the individual's real wage before, α and β are vectors of parameters, x is a vector of other covariates, and u is the state-level unemployment at the approximate time of the job displacement, and is ε an error term. This type of specification is also used, for example, in Gibbons and Katz (1991). This specification has the advantage, relative to a specification in levels, that person-specific

Conclusion

This paper shows that in a model with unemployment due to bad luck as well as selection, the composition of displaced workers varies over the business cycle, and presents some empirical evidence supporting this claim. As in Gibbons and Katz (1991), it is difficult to know whether the effect is associated with signaling or adverse selection. The workers may receive lower wages because firms take previous unemployment as a signal of low ability or simply because the firms observe that the workers

Acknowledgements

I would like to thank Daron Acemoglu, Josh Angrist, Larry Katz, Jón Steinsson and an anonymous referee for helpful discussions and comments.

References (9)

  • Henry S. Farber

    “The incidence and costs of job loss: 1982–1991,” Brookings Papers on Economic Activity

    Microeconomics

    (1993)
  • Henry S. Farber

    “The changing face of job loss in the United States, 1981–1995,” Brookings Papers on Economic Activity

    Microeconomics

    (1997)
  • R. Gibbons et al.

    Layoffs and lemons

    Journal of Labor Economics

    (1991)
  • Jo Anna Gray

    Wage indexation: a macroeconomic approach

    Journal of Political Economics

    (1978)
There are more references available in the full text version of this article.
View full text