Moral hazard and subjective evaluation

Abstract

This study investigates optimal contracts to solve the moral hazard problem with subjective evaluation in a static environment with multiple agents wherein the principal observes agents' performances privately. Despite the limitations of feasible contracts that the principal can credibly offer, we show the irrelevance theorem that even under subjective evaluation, the principal can be as well off as if agents' performances are objective and verifiable, provided there are at least two risk-neutral agents whose actions satisfy a certain condition on their output distributions. Our irrelevance result encompasses a large class of moral hazard problems, including statistically correlated and technologically interdependent performance signals. We also extend the model to include multidimensional actions, limited liability constraints, and collusion among agents.