A/B Contracts

This paper aims to improve the practical applicability of the classic theory of incentive contracts under moral hazard. We establish conditions under which the information provided by an A/B test of incentive contracts is a sufficient statistic for answering the question of how best to improve a status quo incentive contract, given a priori knowledge of the agent's monetary preferences. We assess the empirical relevance of this result using data from DellaVigna and Pope's (2018) study of a variety of incentive contracts. Finally, we discuss how our framework can be extended to incorporate additional considerations beyond those in the classic theory.