Empirical Research On China’s A-share Listed Companies’Financial Data Quality Checks And Evaluation Methods

Data is the key reference to realize state condition, understand public opinion, manage the corporation, and regulate the market. With the development of capital market, regulators and investors have increasing demand for high-quality financial data. However, with more and more people casting doubt on the reliability of current data, it is a huge responsibility of statistician to investigate the quality of financial data to make sure that it can reflect the truth and provide accurate information to the public. In this study, we apply statistical method to screen and evaluate the financial data to make reference for audit institution. We make the following contributions:Firstly, we define the concept of data distortion and fraud: data distortion is the diversion between available data and actual data; while data fraud is to manipulate data on purpose. Moreover, our study shows the necessity of analyzing anomaly in financial data because it can either result from actual substantial market progress or due to data fraud.Secondly, we use factor analysis and Benford method to build the index system, and then screen and analyze the actual data from all the listed companies in Shenzhen and Shanghai Stock Market based on theoretical framework of Corporate Finance and Statistics.Thirdly, after pairing the sample data between standard audit companies and nonstandard ones, we use the discriminate analysis and logistic regression to classify the sample data. Discriminate analysis result shows that data from standard audit company has an accurate rate of78%. In addition, While the classification result of logistic regression and classification result shows95.4%accuracy of audit performance.Fourthly, we analyze and evaluate the financial index of the abnormal data. Based on the characteristics of financial report and resource of available information, we use trend simulation method to study the financial index of the balance sheet and evaluate the quality of the index. While for the income statement’s index, we use relevant index to establish econometric models for analysis.