Financial Reporting Fraud Identification Studies Based On F-Score Model

Financial report of listed companies is the main medium to disclose the company’s financial position and operating results, the important basis for investor to identify companies’quality in capital market. Enterprise financial report fraud conveying untrue information affects the decisions of shareholders, creditors and other report users, which brings huge economic losses to report users. In recent years, although regulators have strengthened the supervision of financial reporting fraud issues, fraud problems still exist. So finding effective ways to identify financial reporting fraud has become the common topic in theory and practice.The paper refers to the research method of American scholars’ F-score fraud identification model study, selecting60listed companies which have the fictional profit fraud behavior and60non-fraud companies as samples from the Shanghai and Shenzhen stock exchange, which from1994year to2013year. And select available16indexes of F-score model and14widely applied indexes of domestic research as variables, by adopting the method of nonparametric test in30indicators in the fraud and non-fraud sample data for significance test and finding indicator variables that have significant differences in the fraud and non-fraud samples as initial variables of regression model. Then use the Logistic regression method to get to financial reporting fraud recognition model. The empirical results show that Change in receivables, Change in cash margin, Asset-liability ratio, Inventory to current assets, Accounts receivable turnover, Asset turnover, Proportion of state-owned shares, Ownership concentration, Audit opinion have significant difference between fraud and non-fraud companies. Model has correct discrimination rate of75.8%. Finally, the paper from the three factor theory of stress, opportunities for fraud and fraud excuse three aspects put forward the related Suggestions in dealing with the financial reporting fraud of listed company.The main innovation point of this paper is selecting variables F-score model, combined with the financial reporting fraud identification research of Chinese scholars, building a new variable system of model. At the same time, using our country listed company fraud data, test the new variable system. Construct the optimization model of F-score which is suitable for China’s listed companies.