Implementation in Principal-Agent Models of Adverse Selection

This paper studies implementation in a principal-agent model of adverse selection. We explore ways in which the additional structure of principal-agent models (compared to general implementation models) simplifies the implementation problem. We develop a connection between the single crossing property and monotonicity conditions which are necessary for Nash and Bayesian Nash implementation. We also construct simple implementing mechanisms that rely on the single crossing property and on assumptions about the outcome set frequently made in the principal-agent literature.