Optimal effort incentives in dynamic tournaments

This paper analyzes two-stage rank-order tournaments. A principal decides (i) how to spread prize money across the two periods, (ii) how to weigh performance in the two periods when awarding the second-period prize, and (iii) whether to reveal performance after the first period. The information revelation policy depends exclusively on properties of the effort cost function. The principal always puts a positive weight on first-period performance in the second period. The size of the weight and the optimal prizes depend on properties of the observation error distribution; they should be chosen so as to strike a balance between the competitiveness of first- and second-period tournaments. In particular, the principal sets no first-period prize unless the observations in period one are considerably more precise than in period two