Research

A principal provides incentives to a team of agents working on a risky project with two milestones of unknown feasibility. As agents work unsuccessfully their private beliefs decline, so the principal must pay rents to deter procrastination and free-riding; to reduce these rents she stops the project inefficiently early and gives identical agents asymmetric experimentation assignments, rewarding them with better future contract terms or task assignments rather than monetary bonuses.

Presents counterexamples in which an optimal contract fails to exist even though the first-order approach is valid. Existence turns on the agent's utility at low wealth: if utility diverges to minus infinity at a finite wealth level, existence can fail for some specifications of the cost of effort; if it diverges only as wealth itself goes to minus infinity, existence holds provided the agent continues to dislike risk. The paper also characterizes near-optimal contracts when existence fails.

Empirical analysis of interdependent values in a two-sided matching market. Using Denmark's centralized medical program admissions and an admission reform, we reject the private-values assumption and show that programs favor local applicants; sharing information would raise student persistence and program payoffs.

We define and establish existence of constrained equilibria and trembling hand perfect equilibria in infinite dynamic games with asymmetric and imperfect information, via a new notion of sequential absolute continuity that extends Milgrom and Weber (1985) to dynamic games. In Markov settings the approach also yields existence of Markov equilibria.

Defines notions of trembling hand and sequential equilibrium and shows that both exist in a large class of stochastic games that may feature incomplete and imperfect information, with no need for a public correlating device. Under further regularity assumptions, each stochastic game has a sequence of approximating finite games whose equilibria approximate equilibria of the limit game.

Studies games in which the timing of players' moves is private information but players can disclose their move at a small cost. For coordination games we characterize when the game has a unique prediction: players choose the best coordination equilibrium and disclose nothing. In games of opposing interests we give sufficient conditions for the first-arriving player to disclose.