Applying Benford’s law to detect earnings management

Abstract Aim/purpose – This paper analyzes the role of Benford’s law in the detection of earnings management in Poland. Previous research that uses Benford’s law does not split the sample into a fraud and a control group; however, this method is used in logistic regression and data mining analysis. Design/methodology/approach – The sample comprises 126 observations of Polish non-financial companies listed on the Warsaw Stock Exchange for the years 2010-2021. The author uses first, second, and first-two digits analysis as a proxy for earnings management detection. Findings – The results indicate that fraudulent companies have different deviations in the digits than control firms. Accordingly, the statistical test results indicate that control companies have weaker conformity with the Benford distribution than fraudulent companies. Research implications/limitations – The study sample is limited to 126 observations, which is due to the small number of listed firms that received a monetary fine from the Polish Financial Supervision Authority (UKNF Board) for violation of IAS/IFRS principles related to their financial statements during the study period. Originality/value/contribution – The author offers a significant contribution to the accounting literature by proposing the separation of fraudulent and control observations in Benford analysis due to differences in the deviations of digits. Also, analyzing the full sample may lead to the identification of inappropriate areas for further auditor analysis.