The current recession has shown, yet again, that theory and evidence are on a collision course. Competitive labor market theory predicts that downturns should have only modest effects on unemployment. A decrease in aggregate demand might lower output and the demand for labor, but this does not necessarily imply higher unemployment. Lower demand for labor leads to lower wages, with the consequence that some individuals might leave the labor market. The remaining workers need only lower their wage demands until suitable employment is found.
One hypothesis, going back to Keynes, is that workers lose their jobs because they are unwilling to accept lower wages. Implicit contract theory, beginning with (Costas Azariadis 1975) and (Martin N. Baily 1974), and subsequently (Paul Beaudry and John DiNardo 1991), suggests rigid wages are the consequence of a contract between firms and risk-averse workers. Yet, (Sherwin Rosen 1985) observes that the hypothesis that wages are the result of an optimal risk-sharing contract implies that laid-off workers are no worse off than employed workers, a prediction that is inconsistent with the evidence.
In this paper we observe that one can build a very simple labor contracting model to explain these facts using three ingredients, two of which pre-date implicit contract theory. The first is the existence of relationship specific investment that both (Gary Becker 1962) and (Jacob Mincer 1962) emphasized as central to the wage setting process.