Most of the empirical work on executive compensation investigates the role of contemporaneous performance measures in setting cash compensation, ignoring the relevance of past performance measures and the structure of cash compensation. In our paper, The Relation between CEO Compensation and Past Performance, forthcoming in The Accounting Review, we focus on the relation between cash compensation components (salary and bonus) and past performance measures as signals of a CEO’s ability.
We first develop a simple two-period principal-agent model with moral hazard and adverse selection. Our model suggests that salary is adjusted to meet the reservation utility and information rent, and is positively correlated over time to reflect ability. Bonus serves to address moral hazard and adverse selection problems by separating agents into contracts with different levels of risk. Agents are screened and receive different bonus arrangements according to their types. The higher an agent’s type, the more sensitive his bonus is to contemporaneous performance. A higher ability agent receives a larger portion of his compensation in the form of bonus and less as salary. For a given agent, salary increases with his past performance and higher current salary predicts higher future performance. Current bonus, however, is negatively correlated with both past and future performance.
We then test theoretical predictions using CEO compensation data from 1993-2006. Consistent with the predictions, we find that salary (bonus) is positively (negatively) associated with past performance for continuing CEOs. We also find that while current salary is positively associated with future performance, current bonus is not.
For newly appointed CEOs, we find that salary is positively related to the past performance of his old firm, but unrelated to the past performance of his new firm. In addition, we find that the future performance of his new firm is positively associated with his current salary (but not his current bonus) at the new firm, after controlling for current and past performance. Our results are robust to controlling for the new CEO’s tenure, old position, internal versus external hire, “ability score,” and industry representation of the old firm. We also find that as a new CEO’s tenure increases, the association between his salary and his new firm’s performance becomes stronger while the association between the CEO’s current salary and his performance at his old firm becomes weaker. Internally promoted CEOs tend to receive larger salaries and smaller bonuses than external CEOs. Finally, we find that the weight on past performance at the old company in the salary regression is higher for new CEOs hired from the same industry.
Our analysis emphasizes the importance of considering past, contemporaneous, and future performance in examining efficient contracting. It also demonstrates the need to distinguish the relation between performance measures and different components of compensation (salary and bonus) because they serve two different though related purposes: (1) to provide the appropriate levels of incentives to screen and motivate agents and (2) to provide agents with their reservation utility based on information on their ability. Third, our study focuses on the relation between new CEO compensation and his performance at both his old and his new firm, which has not been fully investigated in prior research.
The paper is available for download at here.