Ph.D. Candidate in Economics
Curriculum Vitae: PDF
Research Interests: Microeconomic Theory, Information Economics, Corporate Finance
abstract: I study a continuous-time principal-agent model in which a risk-neutral agent protected by limited liability exerts costly efforts to manage a project for her principal. Unobserved risk-taking by the agent is value-reducing in the sense that it increases the chance of large losses, even though it raises short-term profits. In the optimal contract, severe punishment that follows a large loss prevents the agent from taking hidden risks. However, after some histories, punishment can no longer be used because of limited liability. Allowing for risk-taking by the agent is necessary for the project to continue. When the project is close to liquidation, hidden gambling is optimal because of high agency cost in preventing risk-taking. In addition, I explore the roles of standard securities in implementing the optimal contract. The implementation shows that driven by the agency conflicts, incomplete hedging against Poisson risk provides incentives for the agent to take the safe project. Moreover, I study the optimality of "high-water mark" contract widely used in the hedge fund industry and find that "distance-to-threshold" is important in understanding the risk-shifting problem in a dynamic context.
abstract: This paper studies a continuous-time moral hazard model in which the principal hires a team of agents to run the business. The firm consists of multiple divisions and agents exert costly efforts to improve the divisional cash flows. Firm size evolves stochastically based on the aggregate cash flows. I show that exponential utility delivers tractability in the sense that the firm's value can be decomposed into the principal's value and the team's value. The model delivers a negative relationship between firm size and pay-for-divisional incentives, and I characterize conditions under which joint/relative performance evaluation will be used. I also consider the implications of team production on the firm's optimal capital structure.
abstract: In this note, I study a mutli-armed bandits problem with ambiguity. Decision-maker views the probabilities underlying each arm as imprecise and his preference is represented by recursive multiple-priors. I show that the classical "Gittins Index" generalizes to a "Multiple-Priors Gittins Index". In the setting with one safe arm and one ambiguous arm, the decision-maker plays the ambiguous arm if its "Multiple-Priors Gittins Index" is higher than the return delivered by the safe arm.
Working in Progress:
abstract: This paper studies roommate problem introduced by Gale and Shapley (1962) under weak preferences. We aim at strengthening Chung's (2000) "no odd rings" condition for existence and random-paths-to-stability results. By modifying and extending Tan's (1991) stable partition structure to the case with weak preferences, we try to provide a necessary and sufficient condition for the existence of stable matchings. We also attempt to identify a necessary and sufficient condition for convergence of random path to a stable roommate matching, that is, when we start with any initially unstable matching, we can always find a finite sequence of blocking pairs leading to a stable matching.