Empirical identification of the risk shifting aspect of labor market implicit contracts

TR Number
Journal Title
Journal ISSN
Volume Title
Virginia Polytechnic Institute and State University

Much of the recent work in the area of implicit contract theory hypothesizes that firms and workers differ in their attitudes towards risk. The optimal wage and employment contract calls for shifting some of the risk associated with a randomly fluctuating marginal product of labor from the more risk averse party to the less risk averse party. The purpose of this dissertation is to explore the empirical implications of this risk shifting hypothesis. In particular, the following question is addressed: "How can we empirically identify whether risk shifting is occurring in the labor market?”

Chapter 2 explores this question in the context of an implicit contract model with nominal variables and a randomly fluctuating price level. Under the usual assumption of risk neutral firms and risk averse workers the implications of the model are refuted by the industry level nominal wage stylized facts. Under the assumption that risk neutral workers insure risk averse firms the model is capable of explaining the stylized facts about the co-movements in nominal wages and employment.

Chapter 3 explores this question in the context of a long-term implicit contract model with bankruptcy constraints. It is shown that if risk neutral firms insure risk averse workers then the real wage will respond asymmetrically to permanent and temporary revenue function disturbances. In particular, the real wage will respond more to a given permanent shock than to a temporary shock of the same size.

Chapter 4 empirically tests this asymmetric wage response implication. A frequency domain technique is developed for decomposing a measure of revenue function disturbances into its permanent and temporary components and the real wage is regressed on each component. A sample of 12 4-digit SIC code industries are tested. The industry wage responses are estimated separately and as a system of seemingly unrelated regressions. Estimated separately, the results support the asymmetric response implication for 7 of the 12 industries at the .10 level of significance and 6 of the 12 industries at the .05 level. Estimated as a system the joint asymmetric response hypothesis is supported at the .01 level of significance for the 12 industries.