Information Production, Misconduct Effort, and the Duration of Financial Misrepresentation

Abstract

We examine the link between information produced by auditors and analysts and fraud duration. Using a hazard model, we analyze misstatement periods related to SEC accounting and auditing enforcement releases (AAERs) between 1982 and 2012. Results suggest that misconduct is more likely to end just after fi rms announce an auditor switch or issue audited fi nancial statements, particularly when the audit report contains explanatory language. Analyst following increases the fraud termination hazard. However, increases (decreases) in analyst coverage have a negative (positive) marginal impact on the termination hazard, suggesting that analysts signal whistleblowers with their choice to add or drop coverage. Finally, our results suggest that misconduct lasts longer when it is well planned, more complex, or involves more accrual manipulation. Taken together, our fi ndings are consistent with auditors and analysts playing a key informational role in fraud detection, while managerial effort to conceal misconduct signifi cantly extends its duration.

The Walls Have Ears: Local Information Environments and Corporate Fraud

“…The use of AAER data to identify fraud offers two major advantages. First, the misdetection rate (Type I error) is low in this data set (Dechow et al 2011, Black et al 2018. Budgetary considerations mean that the SEC will select firms for enforcement actions, where there is clear evidence of manipulation.…”

“…2011, Davidson, Dey, and Smith 2015, Black et al. 2018). The use of AAER data to identify fraud offers two major advantages.…”

“…If managers believe they cannot conceal fraud for long, they may be less likely to engage in fraud. Black et al (2018) and Dyck, Morse, and Zingales 11. The Sarbanes-Oxley Act (SOX) requires firms to evaluate the effectiveness of their internal controls and disclose any identified material weakness.…”

“…Enforcement actions follow significant investigations by the SEC and involve material and economically significant violations as well as the allegation that investors were intentionally misled. 5 The initial data set includes 4,012 AAERs issued by the SEC between May 17, 1982 andDecember 31, 2018. AAERs data have been used extensively in prior work to study accounting misstatements and financial fraud (e.g., Dechow et al 2011, Davidson, Dey, and Smith 2015, Black et al 2018. The use of AAER data to identify fraud offers two major advantages.…”

“…Our estimation is based on cross-sectional fraud case data and includes similar controls as in the fraud commission equation. Our sample includes 250 unique fraud cases with detection taking an average of 10 quarters, which is comparable to Black et al (2018) sample. The results are displayed in Table 9.…”