Robust incentives for risk

Abstract

This paper studies a moral hazard problem wherein the principal is uncertain about the agent’s risk preferences and production technology. The principal, desiring robustness, evaluates contracts according to their worst case payoff against the set of preferences and technologies that satisfy her assumptions about the agent and his capabilities. We show that contracts which exhibit robustness to severe risk taking behavior (i) pay their smallest transfer for all outputs below some lower threshold and (ii) pay their largest transfer for all outputs above some upper threshold. In our main result, we identify a class of environments in which two-wage binary bonus contracts are robustly optimal.