Incentives and performance in the presence of wealth effects and endogenous risk

Two of the most widely tested predictions of agency theory are that there exists a negative association between an agent’s pay-performance sensitivity (PPS) and the risk of output, and that PPS, which is a measure of incentive, enhances performance. Empirical evidence has been mixed. This paper develops a model that introduces a “wealth effect” and also allows the agent to control the risk of output. In doing so, the paper proposes a utility function that has not been employed in the literature. The paper shows that the two predictions, which are formulated in the absence of both wealth effects and endogenous risk, may not hold. Specifically, PPS can be either positively or negatively related to risk, and a higher PPS does not necessarily lead to a higher expected profit for the principal. In other words, even under the linear contract space, PPS is not a sufficient predictor for an agent’s actions and performance. The paper thus presents the need for a new incentive measure that incorporates both PPS and constant compensation.