Research on Early Warning Model of Financial Report Fraud in China

Abstract

At present, financial report fraud is becoming more frequent with the continuous development of the world economy. How to provide early warning before financial report fraud occurs has become more and more important. The purpose of this paper is to set up a logistic regression model, namely an ex-ante warning model, which can provide early warning before financial report fraud occurs, by comparing the governance of financial report fraudulent companies and non-fraudulent ones. First, this paper uses the “fraud triangle” theory as a framework to find the relevant proxy variables for fraud opportunities, fraud pressure, and fraud rationalization. Second, the study uses T-test, Mann-Whitney test and chi-square test to identify statistically significant differences among these proxy variables. Hypotheses were made about the relationship between the coefficient values and the presence of false behavior in the reports. Finally, an ex-ante fraud warning model is set up from the indicators with statistically significant differences, meanwhile the hypotheses regarding the behavior of the indicators and their impact on the model are tested. The overall accuracy of the ex-ante fraud early warning model developed in this paper is 70.9%. How to further debug the model to make the screening of fraudulent companies more accurate is the difficulty and further research direction of the article.