Essays on Repated Moral Hazard
In Chapter 1, I study an infinitely repeated moral hazard problem in which the principal privately observes and publicly reports the agent's output, as in Fuchs (2007). The role of the agent's private strategies, which depend on the history of his private efforts, is examined in providing incentives for the principal to be truthful. I show that in order for his effort history to work as an incentive device, the agent has to use a mixed strategy, since otherwise his efforts are predictable by the principal and thus, in effect, public information. However, hiding the agent's efforts from the principal incurs a non-negligible efficiency loss, which may, or may not be justified by the efficiency gain from the use of the agent's private strategies. Moreover, the agent's optimal strategy is shown to be consistent with empirical studies on how employees respond to subjective performance evaluations.
In Chapter 2, we studies an equilibrium model of the labor market with moral hazard in which jobs are dynamic contracts, job separations are terminations of optimal dynamic contracts, and terminations are used as an incentive device. Transitions from unemployment to new jobs are modeled as a process of matching and bargaining. Non-employed workers make consumption and saving decisions as in a typical growth model, but they must also decide whether or not to participate in the labor market. The equilibrium of the model is characterized. We then calibrate the model to the U.S. labor market to study quantitatively worker turnover, compensation dynamics and distribution. We show that the model can generate equilibrium wage dispersions similar to that in the data. Hornstein, Krusell and Violante (2006) argue that standard search matching models can generate only a very small differential between the average wage and the lowest wage paid in the labor market.