In the paper, Executive Overconfidence and the Slippery Slope to Financial Misreporting, forthcoming in the Journal of Accounting and Economics as published by Elsevier, our detailed analysis of a sample of 49 firms subject to SEC Accounting and Auditing Enforcement Releases (AAERs) suggests two distinct explanations for the misstatements. Just over one quarter of the cases represent many of the well-publicized examples of corporate fraud including Adelphia, Enron, Healthsouth, and Tyco. The nature of the misstatements, their timing, and an analysis of the executives suggest that the activities are consistent with a strong inference of intent on the part of the respondent and consistent with the legal standards necessary to establish fraud.
However, perhaps more surprising, we find that the actions by the executives in the remaining three quarters of the cases are not consistent with the pleading standards required to establish an intent to defraud. Rather, our analysis of the 49 AAER firms suggests that optimistic bias on the part of executives can explain these AAERs. We show that the misstatement amount in the initial period of alleged misreporting is relatively small, and possibly unintentional. Subsequent period earnings realizations are poor, however, and the misstatements escalate. Using a matched sample of non-AAER firms, we show that the misreporting firms did not simply get a bad draw on earnings. Nor does it appear that weaker monitoring relative to the matched sample explains why the misreporting manager’s optimistic bias affects the financial statements.
We further examine whether the optimistic bias for the misreporting firms is associated with the character trait of overconfidence. The evidence from the analysis of the 49 AAER sample is mixed on this question. However, we find evidence of a positive association between proxies for overconfidence and the propensity for AAERs in two larger samples that use alternative measures of overconfidence. The association between overconfidence and AAERs is consistent with the slippery slope explanation in which greater optimistic bias makes it more likely that a manager is in the position that significant misreporting is an optimal choice.
An interesting question raised by the analysis is the importance of monitoring the optimistic bias of executives. Various models predict that overconfidence has desirable effects on the executive’s performance (Goel and Thakor, 2000; Gervais and Goldstein, 2007; Gervais et al., 2010). Our analysis indicates overconfidence can be associated with financial reporting concerns and prior work has documented an association between overconfidence and distorted investment and financing decisions (e.g., Malmendier and Tate, 2005 and 2008 among others). For firms who value the positive aspects of overconfidence, a plausible response is to put mechanisms in place to monitor the executive’s decision-making biases associated with this trait. This response is feasible only if the Board recognizes executive overconfidence. Our evidence indicating that the misreporting firms and matched sample of non-AAER firms have different compensation arrangements suggests that the Board is able to do so at some level. However, our corresponding analysis of monitoring does not indicate that the overconfident managers were better monitored, which explains why they were more likely to end up misreporting. The potential for monitoring to moderate the optimistic bias that characterizes executives remains an interesting open question. Is it that our analysis does not adequately capture the specific mechanisms that would control optimistic bias? Or, is the cost of better monitoring higher than the expected benefits from mitigating the risk of misreporting, which is a significant but unusual event?
The full paper is available for download here.