"Information Design in Moral Hazard Environments with Implicit Incentives and Type-Dependent Costs"

Abstract : We analyze an information design problem in a moral hazard setting with implicit incentives, where effort costs decline with talent. The agent’s competency depends on effort and talent, affecting success in a publicly observed task. The designer selects a signal about talent to maximize expected competency. We show that the designer’s indirect utility depends only on the posterior mean and follows an S-shaped structure, leading to an optimal signal characterized by upper censorship, which is strictly informative when the prior includes sufficiently low-talent agents. This contrasts with existing literature, which assumes talent-independent effort costs and finds full concealment optimal.