Voluntary Disclosure in Asymmetric Contests

This paper studies the incentives for interim voluntary disclosure of verifiable information in probabilistic all-pay contests. Provided that the contest is uniformly asymmetric, full revelation is the unique perfect Bayesian equilibrium outcome. This is so because the weakest type of the underdog will try to moderate the favorite, while the strongest type of the favorite will try to discourage the underdog─so that the contest unravels. Self-disclosure is optimal even though a weak favorite or strong underdog may be induced to raise their efforts, i.e., show "dominant" or "defiant" reactions. To avert Pareto inferior unraveling, the favorite may prefer to shut down communication, but this is never the case for the underdog. We also consider partial information release, cheap talk, Bayesian persuasion, information design, correlation, and continuous types. Applications are discussed. The proofs employ novel arguments in monotone comparative statics and an improved version of Jensen's inequality