Search for Optimal CEO Compensation: Theory and Empirical Evidence

Much of the agency literature focuses on effort-inducing while little attention is paid to the participation constraint. Intuitively, it is important to jointly address both for CEOs. This paper achieves this by developing a dynamic search equilibrium model that allows for quitting if a CEO is not satisfied with the incentive contract. The reservation utility is endogenized in equilibrium since the value of the CEO's outside option depends on other firms' contracts. As a result, the equilibrium incentive contract exhibits some new and important features that can explain two long-standing puzzles related to the executive compensation practice. First, the equilibrium pay-to-performance sensitivity negatively depends on the expected aggregate state and a firm's systematic risk, and positively on the firm's specific risk. The separate effects of firms' systematic and specific risks enable us to reconcile the mixed evidence on the relationship between pay-to-performance sensitivity and a firm's risk. Second, the equilibrium salary and total pay depend on the firm's size, which, in turn, increases with the expected aggregate state. This result provides a plausible explanation to the steadily increased compensation paid to executives in the past three decades. These theoretical predictions are broadly supported by our empirical results. Hence, we conclude that the participation constraint is an important determinant for the executive compensation policy.