Osaka University

This paper studies the problem of information revelation in a multi-stage tournament where the agents ’ effort in each stage gives rise to a stochastic per-formance signal privately observed by the principal. The principal controls the agents ’ effort incentive through the use of a feedback policy, which transforms his private information into a public announcement. The optimal feedback pol-icy is one that maximizes the agents ’ expected effort. The paper identifies when the principal should use the no-feedback policy that reveals no information, or the full-feedback policy that reveals all his information