Third, we examine what factors affect the change in CEO-CFO LSM within each CEO-CFO dyad. For this purpose, we conduct CEO-CFO dyad fixed-effects regressions. If CEO-CFO LSM is mainly driven by CFOs’ ingratiation of CEOs, we expect that CEO power should predict CEO-CFO LSM since CFOs would have stronger incentives to ingratiate CEOs when the CEOs are more powerful. In comparison, if CEO-CFO LSM is mainly driven by CEO-CFO social cohesion, CEO-CFO LSM should increase as CEO-CFO tenure overlap extends. In unreported results, we find that CEO power rather than CEO-CFO tenure overlap predicts CEO-CFO LSM. This combined evidence provides strong empirical support indicating that CEO-CFO LSM is mainly driven by CFO’s ingratiation of CEOs rather than CEO-CFO social cohesion.