Abstract:
This paper studies dynamic mechanism design in the presence of moral hazard. Revelation principle extends to models with moral hazard for both full commitment and limited commitment, but I also identify environments in which the principal doesn’t benefit from eliciting agents’ private information or beliefs. One-shot deviation principle requires the knowledge of agents’ private strategies after deviations, and I characterize the necessary and sufficient condition for all IC constraints that requires only the knowledge of agents’ equilibrium strategies. I also provide two sufficient conditions for smaller set of IC constraints that require obedience after a single-period deviation to be sufficient for all IC constraints. I illustrate how to apply revelation principle and the smaller set of IC constraints with an application allowing for endogenous state.