Analysis of a Dynamic Adverse Selection Model with Asymptotic Efficiency

Abstract

This paper studies an infinite horizon adverse selection model with an underlying two-state Markov decision process. It introduces a novel approach that constructs the continuation payoff frontier exactly, as the fixed point of a functional operator. If the model supports an incentive-compatible first-best (ICFB) contract, the continuation payoff frontier can be efficiently constructed, and the principal's optimal contracts converge to ICFB contracts over time. The existence of an ICFB contract is implied by the common assumption of private values and is a fairly general scenario. The paper generalizes some key findings in the dynamic adverse selection literature to this scenario.