Moral hazard and internal discipline: Theory and evidence

This paper studies, both theoretically and empirically, how subordinates to CEOs can discipline the CEOs' self-serving activities. I predict that because CEOs' self-serving activities hurt the subordinates through the subordinates' stakes in the firms, the subordinates who observe these activities will take actions that negatively affect the CEOs, and that in anticipation of such reactions by subordinates, the CEOs will limit their own misbehaviors. This disciplinary mechanism will become more effective when the CEOs' self-serving activities are more observable to subordinates. Further, the sensitivity of CEOs' self-serving activities to observability will increase (1) as the agency problem between CEOs and their subordinates intensifies, and (2) when external monitoring is less effective. The incentive pay for the subordinates will also decrease with the strength of external monitoring. Using a series of empirical tests, I find results that are largely consistent with my theoretical predictions.