Incremental Information Disclosure in Qualitative Financial Reporting: Differences Between Fraudulent and Non-fraudulent Companies

Abstract

This research investigates how fraudulent companies differ in their narrative disclosure strategies across multiple reporting venues. Fraudulent companies must avoid discussing fraud-relevant items, or they must present a fabricated version of their story. To reduce involuntary information disclosure and attempt to convince their audience that what they say is true, fraudulent firms may have greater language similarity than non-fraudulent firms between their quarterly earnings calls and the subsequent Management’s Discussion & Analysis section of the quarterly financial statement. I measure language similarity between a company’s earnings calls and their corresponding financial statements. I find that the Chief Financial Officers of fraudulent companies have significantly higher language similarity between disclosure venues than their non-fraudulent peers. I also test the predictive ability of these measures using several classification schemes, achieving fraud-detection accuracy levels between 53% and 71%. The paper concludes with a discussion of the findings and avenues for future research.