Governance and CEO Turnover: Do Something or Do the Right Thing?

We study the efiect of entrenchment on the causes and consequences of CEO dismissal. We begin with the conventional 'costly flring' model of managerial entrenchment, where an entrenched CEO is dismissed only if the subsequent gains from doing so are very large. This model yields a number of intuitive predictions: (a) post-flring flrm performance improvements are greater for entrenched CEOs (b) entrenched CEOs are flred less frequently (c) market reaction is more positive to the flring of entrenched CEOs (d) the pre-flring decline in flrm performance is greater for entrenched CEOs (e) following poor flrm performance, flrms controlled by entrenched CEOs see little improvement, relative to flrms controlled by less entrenched CEOs. We test these predictions using data on dismissals of CEOs of large U.S. corporations from 1980- 1996. While we report results that are strongly consistent with predictions (a) { (c), the data do not support predictions (d) and (e). We suggest an alternative model, built on the premise that shareholders may misattribute poor performance to the CEO rather than circumstance. In this model, an entrenched board may protect 'good' CEOs from dismissal. We suggest that this model is more broadly consistent with the reported flndings and consistent with insights about performance attribution from social psychology.