In this paper, I study a principal-multiagent model in which the principal adopts relative performance evaluation (RPE) as the compensation scheme, and agents are connected in an information network. The ability of each agent is private information, but adjacent agents in the information network...