Information disclosure in dynamic research contests

We study information disclosure in a dynamic multi-agent research contest, where each agent privately searches for innovations and submits his best to compete for a winner-takes-all prize (Taylor, 1995). Different disclosure policies on the agents' submissions induce different equilibrium behavior, making the design of disclosure a useful instrument for contest sponsors. We analyze and compare various information disclosure policies in the contest with finite or infinite horizons. With an endogenously chosen prize, the public disclosure policy, where submissions are revealed immediately, implements the sponsor's first-best research plan and is an optimal policy in the infinite horizon.