The Empirical Study On Financial Reporting Fraud Of Listed Companies In China

Along with the economic development, the increase in the number of listed companies and countries due to financial report produced fraud scandal more and more. The corruption scandals from the companies abroad such as Xerox, Enron WorldCom stroked investors' confidence, and also forced the United States issued a stern Sarbanes-Oxley Act to strengthen the supervision of corruption. The domestic situation is not optimistic. Except Gelinkeer department and Kelon fraud, both sides of the recent incidents of Amoi electronics and Liangmianzhen are also very intriguing. How to better identify financial report of the listed company whether fraud or not has became the most concern of the regulatory structure. Under this context the thesis hope to research financial reporting fraud, construct recognition model to play the investigation role of financial reports. Early recognition of the fraud of the listed company is of great significance. It can ensure the normal order of the securities market to be fair and effective for countries, promote stability and harmonious for society and safeguard the interests of investors and avoid its property from loss for many investors.This thesis introduces the concepts and the research of financial fraud model and motivation theory, analyses the advantages of the BP neural network model adopted and the reasons of our country's financial fraud. Based on the results of the analysis made variables construct the BP neural network recognition model, through the inspection found the recognition has good recognition effect. Finally the thesis issues of the personal view, and illustrates the new development of financial fraud recognition in the new accounting rule.This thesis is divided into five chapters as follows:The first chapter is foreword. First, this chapter introduces the background and significance of the study, then summarizes the financial report fraud recognition, explains the definition related and introduces various research models. On the analysis of the advantages and disadvantages of the present recognition models, this thesis chooses the BP nerve network model and introduces the basic theory.The second chapter analysis our country's financial reporting fraud reasons and select variables. This chapter laid a foundation for the next empirical analysis. Firstly the related theory of foreign corrupt drivers are reviewed, more representative is:iceberg theory, triangle theory, GONE theory and the theory of enterprise risk factor. Based on triangle theory and combining with the actual situation in our country, this chapter analyzed the pressures, opportunities and excuses the Chinese listed companies faced. Through the analysis extracted the variables which the empirical research needed.The third chapter collects the samples needed to construct model and makes a descriptive analysis of the sample.This thesis will be divided into model building stage and application testing phase, corresponding the sample into input model samples and testing samples. The descriptive analysis makes more intuitive understanding of the fraud sample.The fourth chapter constructed the model according to the preparation and tested it. Firstly, pretreat the variables selected. Use the principal component analysis method for variable reduction to favor the next analysis.Test results shows the model constructed has good recognition ability.The fifth chapter makes conclusions and recommendations. This chapter thesis mainly summarizes the gains from the empirical analysis conclusion, and then puts forward the existing shortcomings, insufficiencies and advances.