Zastosowanie regresji logistycznej do wykrywania fałszowania sprawozdań finansowych

Abstract

This paper aims to detect the financial statements fraud based on the financial data of the US companies accused by the US Securities and Exchange Commission of manipulating financial statements under rule 10(b)-5 of the Securities Exchange Act in 2000–2007. Particular attention was paid to the use of logistic regression in the empirical literature for the detection of creative accounting. Based on empirical articles we selected thirteen financial ratios and two models predicting bankruptcy aimed at verifying the research hypotheses. In addition, we tested collinearity and equality of central tendencies between companies that falsify and do not falsify financial statements. The results confirm that the item of financial statements that is most in danger and should be monitored when attempting to detect the financial statements fraud are current assets.

Keywords

• logit regression
• creative accounting
• falsified financial statements
• american market