Industrial Discrepancy-Based Financial Distress Diagnosis Evidence Study

Financial distress is a process of gradual deterioration and can be predicted from the point of view of identification and prediction. Early predictive signal can inform governors of adopting measurements so as to shorten the length of loss or improve company's operational and financial conditions for the sake of healthy development. Thus, it is very important to set an effective financial distress prediction model. Meanwhile, listed companies in different industries show their financial features with obvious significance because of their differences of market structures, the correlations with the macroeconomic cycle and the periodic stage they are in their own industries. Therefore, it is essential to make an analysis of the industry which a company lies in.The Chinese listed companies are classified according to CSRC'S Industry Classification Standard in 2001 in this paper. We choose machine-equipment-instrument industry, metal-nonmetal industry and petroleum-chemistry-plastics-plastic cement industry as our research samples and conduct a thorough empirical study of the differences of financial indexes among different industries with the use of statistic analysis and Kruskal-Wallis H method. Based on this, we use factor analysis in order to reduce the correlations among variables and gain factors which are regarded as new variables in our model. Then we set our financial distress prediction models from the perspective of industries using the t-3 data of sample companies. Finally, we test the model's effectiveness.The results of our research show that there are significant differences among such financial indexes in different industries, especially those reflecting corporate ability to pay back debts and capital structure. Furthermore, our model can predict distressed companies with 85.5%, 80.8% and 84.2% accuracy and 83.3% average accuracy. Compared with prediction model not considering industrial differences, the overall average accuracy of our model is 2.5% greater and our model can predict distressed companies with greater than 8.3% average accuracy. Furthermore, it has a more satisfactory fitting degree. The outcomes of the research not only prove that the financial distress prediction model in our research is effective,but also will provide academic and practical support to accounting firm, investors and government departments when they make corresponding industrial analysis and appraise a company by putting it into its own industry.