Essays on Contest Design and Contract Design.

Abstract

This dissertation is broadly on the topic of behavioral economics and mechanism design, using lab experiments to study two settings: contest design and contract design.

The first chapter studies how designers of two-stage elimination contests should allocate prizes and reveal interim performance information to achieve desired outcomes. To investigate the interactive effect of prize allocation and interim information revelation, I design a two-stage contest experiment with different combinations of these two instruments. While my theoretical model predicts a positive effect of awarding a single prize on effort regardless of information structure, I find this effect in the lab only when interim information is revealed. Moreover, revealing information motivates effort only under a single prize. These findings are consistent with a framework that ties together the intuitions of how these two instruments work on contestants: information on others' performance allows contestants to estimate their own probabilities of achieving different ranks, and hence increases their sensitivity to changes in the prize allocation that affects the ranks they are likely to reach. Given this interplay of prize and information, my findings suggest that contest designers should take into account the state of one instrument when optimizing the other.

In the second chapter, we design a real-effort laboratory experiment to investigate how group identity influences decisions in a principal-agent framework with hidden action. Group identity is induced by random assignment to groups, and is further enhanced using a collective puzzle solving task. We find that the principals show ingroup favoritism towards ingroup agents by making more generous revenue-sharing offers. While ingroup agents are less tolerant of low offers from their principals, they exert greater effort in response to higher offers, relative to the control condition. The impact of group identity and incentives on agents' effort also depends on their perceptions of fair offers.