The frequency of non-GAAP (or "pro forma") reporting has continued to increase in the United States over the last decade, despite preliminary evidence that regulatory intervention led to a decline in non-GAAP disclosures. In particular, the Sarbanes-Oxley Act of 2002 (SOX) and Regulation G (2003) impose strict requirements related to the reporting of non-GAAP numbers.More recently, the Securities and Exchange Commission (SEC) has renewed its emphasis on non-GAAP reporting and declared it a "fraud risk factor." Given the SEC's renewed emphasis on non-GAAP disclosures, we explore the extent to which regulation has curbed potentially misleading disclosures by investigating two measures of aggressive non-GAAP reporting. Consistent with the intent of Congress and the SEC, we find some evidence that managers report adjusted earnings metrics more cautiously in the post-SOX regulatory environment. Specifically, the results suggest that firms reporting non-GAAP earnings in the post-SOX period are less likely to (a) exclude recurring items incremental to those excluded by analysts and (b) use non-GAAP exclusions to meet strategic earnings targets on a non-GAAP basis that they miss based on Institutional Brokers' Estimate System (I/B/E/S) actual earnings. However, we also find that some firms exclude specific recurring items aggressively. Overall, the results suggest that while regulation has generally reduced aggressive non-GAAP reporting, some firms continue to disclose non-GAAP earnings numbers that could be misleading in the post-SOX regulatory environment.

No abstract

We investigate the audit fee response to CEO behavioral integrity (BI). BI refers to the perceived congruence between an individual's words and deeds (Simons 2002). Because low word-deed congruence should result in more explanations when communicating, we use variation in explanations beyond firm fundamentals and CEO-specific characteristics in more than 30,000 shareholder letters to serve as a linguistic-based proxy for CEO BI. We find that audit fees increase as BI decreases, but BI is not associated with financial misstatement or litigation. These findings are potentially consistent with auditors undertaking additional work in response to low BI, which, in turn, mitigates the risk of restatements and lawsuits. The likelihood of option backdating increases as BI decreases, consistent with the contention that auditors lacked incentives to prevent backdating. Finally, BI is increasing in future performance, which suggests that CEOs partially underpin the returns to high-integrity corporate cultures. JEL Classifications: J24; L25; M14; M41; M42. Data Availability: Proprietary data from KRW International cannot be shared because of the terms of a confidentiality agreement. All other data are available from the public sources cited in the text.

No abstract

I investigate whether changes in derivative and hedging footnote disclosures required by SFAS 161 affect investor and analyst uncertainty. My study is motivated by accounting standard setters' and researchers' interest in disclosure effectiveness, and by prior research linking investors' interpretations of public information to measures of uncertainty. For a broad sample of firms, I use textual analysis to measure changes in the amount and salience of derivative and hedging information caused by SFAS 161. Using a difference-in-differences design to study the effects of these changes, Ifind that investor uncertainty is reduced for firms adopting SFAS 161. In addition, I find that for some uncertainty proxies this reduction is greater for firms whose disclosures were more affected by SFAS 161, consistent with the new disclosures improving investor understanding. I also find evidence of a decreased association between bid-ask spread and movements in risk factors, indicating that SFAS 161 reduced uncertainty stemming from these movements.