In our paper, Internal Corporate Governance, CEO Turnover, and Earnings Management, forthcoming in the Journal of Financial Economics, we examine whether executives who manage earnings increase the risk of losing their jobs. We find that earnings management is strongly associated with the subsequent likelihood of forced CEO turnover, but is not significantly related to voluntary turnover. This basic result holds through several test specifications, including multinomial logistic regressions and competing risks hazard models. In the short run, aggressive earnings management in any given year is associated with an increased likelihood of forced ouster the next year. And in the long run, a CEO’s job tenure is negatively related to earnings management over the time he or she is in the CEO position. Similar results hold when we examine the forced turnover of CFOs. A large battery of sensitivity tests reported in the Internet Appendix indicate that these results are robust to alternate measures of earnings management and different model specifications.
We consider several explanations for these results. One explanation is that forced ouster and earnings management both result from poor firm performance. We find, however, that the relation between earnings management and forced ouster is robust to various controls for firm performance, and is similar in firms with strong and poor stock price performance. The relation persists among firms that fail to meet analysts’ forecasts, indicating that it is not explained by earnings shortfalls.
A second potential explanation is that earnings management and forced CEO ouster are endogenous to the firm’s operating environment. Our central results, however, persist in instrumental variable regressions that control for the potentially endogenous nature of earnings management and CEO turnover. Propensity score matched samples indicate that forced turnover is associated with unusually high levels of earnings management, compared to otherwise similar firms that have voluntary turnovers, or firms with no turnover. We also find that earnings management falls after a CEO ouster, implying that the ousted manager – and not the firm’s operating environment – is responsible for the earnings manipulations.
A third potential explanation is that CEO turnover occurs because these firms face external consequences, such as earnings restatements, SEC sanctions, or bad publicity. This explanation also fails, as our main results persist when we control for, or delete from the sample, firms with earnings restatements (from the GAO (2006) database), SEC enforcement actions (from the Karpoff, Lee, and Martin (2008a,b) database), or adverse publicity (from the Dyck, Morse, and Zingales (2010) sample). Finally, we fail to find support for conjectures that managers face an increased risk of ouster only for managing earnings upwards, or that the results occur because of accrual reversals. The data indicate that the magnitude, but not the direction, of earnings management increases the likelihood of a forced CEO ouster. Furthermore, the CEOs who are ousted face negative career consequences compared to CEOs who leave voluntarily. They are less likely to serve on the company’s board, more likely to lose seats on other boards, and more likely to be sued for misconduct.
These findings indicate that at least some boards of directors act to discipline managers who aggressively manage earnings. It is the earnings management itself that is associated with forced ouster, not whether earnings are managed up or down, or whether the firm is performing well or poorly, or whether the firm has to restate earnings or is disciplined by the SEC. We infer that managers are forced out because, by managing earnings, they impose costs on shareholders by decreasing firm transparency.
These results address two important questions about earnings management, managerial opportunism, and firms’ internal governance. First, these results address a puzzle unresolved by previous research. Karpoff, Lee, and Martin (2008a) report that 89% of the CEOs named as culpable parties for manipulations that trigger SEC enforcement actions lose their jobs, and that most of these CEOs explicitly are fired. Similarly, Desai, Hogan, and Wilkins (2006) and Agrawal and Cooper (2008) find that top managerial turnover increases following public announcements of earnings restatements. But are these managers removed because they manipulated earnings or because of the subsequent public and costly consequences? Our findings indicate that it is the earnings management itself that prompts many firms to remove the offending managers.
Second, these findings indicate that boards act proactively to discipline managers who manipulate earnings. Detecting earnings management is costly, and we do not infer that boards detect or discipline all earnings manipulations. As Hermalin and Weisbach (2008) argue, some earnings management may be unavoidable or even optimal. At some level, however, earnings management attracts the board’s attention and prompts a response. This implies that internal governance acts as an important control on managers’ incentives to manipulate earnings. In a recent paper, Dyck, Morse, and Zingales (2010) investigate who detects corporate misconduct using a sample dominated by allegations of earnings misrepresentations. Our results indicate that internal controls play an important role in such investigations. That is, boards of directors act to discipline managers for manipulating earnings, often before the behavior leads to larger problems such as SEC sanctions. This implies that internal governance works in many cases to discipline earnings management problems before they become severe enough to attract public attention.
The full paper is available for download here.