Irregularities in Accounting Numbers and Earnings Management – A Novel Approach Based on SEC XBRL Filings

Abstract

The SEC XBRL mandate enables the gathering of accounting numbers to be fully automatic in a database-like manner which provides vast opportunities for financial analysis. Using this functionality, this study proposes a simple analytical prescreening measure that uses abnormal digit distributions at the firm-year level to identify firms being suspect of having managed earnings. On average, we find that the constructed measure indicates a greater amount of irregularities in the reported accounting numbers of firms with higher incentives to engage in earnings management. The suggested XBRL-enhanced digit analysis approach may provide the SEC and investors a simple measure to flag financial reports carrying a higher probability of human interaction.

Keywords: XBRL, digit analysis, Benford’s Law, earnings management, earnings benchmarks

JEL Classification: C10, M41, M43