Innovation through competitive experimentation

Abstract

This paper studies optimal contest design for competitive experimentation. A principal wants to implement a project and seeks contributions from multiple agents that increase the value of the project. The agent’s chances of success are independent and their experimentation efforts are unobservable. To induce effort, the principal offers a mechanism that specifies rewards and a termination time which depend on the history of past contributions. I characterize the optimal contest and show that the principal limits the agents’ rent by setting inefficiently tight stopping thresholds.