Stimulating Efforts by Coarsening Information and Prize Design (joint with Fabio Michelucci)
We study a principal who jointly designs an information disclosure rule and a prize structure to maximize total effort in a two-agent contest. Agents are uncertain about their own ability and their opponent’s ability. The principal commits to a public disclosure rule — a mapping from ability profiles to messages — and chooses a general output-contingent prize schedule. We show that the optimal prize takes a simple form: a joint success prize k (awarded when both agents succeed) and a sole success prize m (awarded when only one succeeds), with no reward for joint failure. Moreover, pure prizes always dominate mixed prizes: the optimal prize is always either k-only or m-only. This result reflects a fundamental economic distinction: k rewards joint success, so incentives are stronger when the opponent is more likely to succeed — making effort choices strategic complements — while m rewards sole success, so incentives are weaker when the opponent is more likely to succeed — making effort choices strategic substitutes. Under k-only, the optimal disclosure progressively pools ability profiles into a single anonymous message. Under m-only, the optimal disclosure reveals asymmetric profiles to induce the high-ability agent to exert. The optimal regime switches between k and m depending on the budget, the success probabilities, and the ability distribution.
Room A406