Empirical Research Of Credit Risk For The Listed Companies

With the rapid development of economic, there are a large number of companies which covering all areas and sectors of social life in China. From the reform and opening up to now, most of the industries need self-sustaining. Objectively speaking, the average life expectancy of Chinese enterprises is five years, the life expectancy of most enterprises is not more than three decades, many companies end in final bankruptcy, collapse or being acquired.So,how to manage the risk of listed companies is crucial.The risk of listed companies includes market risk, operational risk, liquidity risk, legal risk, credit risk, and so on. One of the most important is credit risk and it is difficult to measure and manage.The development of our research abour credit risk is relatively backward, especially in the management compared with the international advanced level. They make advanced methods and techniques for credit checks, aspects rating, asset pricing and prediction, in order to make credit risk in line with the actual judgment. With the rapid development of market economy and the international financial innovation, credit risk management becomes more and more important. The purpose of this paper is to learn from the advanced management based on the international countries. We should build a viable credit risk model for Chinese listed companies quickly in order to provide theoretical and practical foundation in the prediction and management of credit risk.There are many traditional methods, such as a measure of credit risk rating, expert analysis, scoring,especially on qualitative analysis. With the increasingly high demand for risk management, there are a series of credit risk management theory based on modern financial theory.If we study the process of credit risk management, we can find credit risk in the direction of quantitive analysis.In addition, we gradually use the data of financial accounts and stock market.This paper empirically analyses the differences in the identification of credit risk capabilities among KMV model,Naive model and LT model,using60 enterprise in A-share market of China as samples,and decides which one is more suitable for the reality in China. KMV model is exogenous default boundary model, Naive models is simplified Merton model, LT model is endogenous default boundary model.In order to improve the accuracy of the models,this article makes some amendments on the parameters with the samples are separated into ST companies and non-ST companies, we use a poor volatility method to estimate the value of the model equity volatility.Both KMV model and LT model show quite good performances in distinguishing ST companies from non ST companies.Both the mean t test and the value of the Wilcoxon test are significant at the5%significance level.The mean t test of Naive model is significant at the10%significance level and the value of the Wilcoxon test is significant at the5%significance level.This paper uses ROC curve and CAP curve to campare the ability of identity of the three models. By constructing the ROC Logistic model, the paper calculates the area under the ROC curve of KMV model is0.88, the area under the ROC curve of Naive model is0.54, the area under the ROC curve of LT model is0.69. For the CAP curve, we need to measure the precise ratio AR. AR of KMV model is0.81, AR of Naive model is0.39, AR of LT model is0.42.The research of ROC curve and CAP curve shows that KMV model is more suitable for the Chinese market.Through the empirical analysis of KMV model,we collect the frequency statistics for each distance interval of ST companies.Then we draw a stacked column chart based on these and design a primary and a secondary credit warning line. Those listed companies of which warning line are below the secondary credit warning line will have a82%probability of credit crisis and we should concern about these companies.Those listed companies of which warning line are below the primary warning level will have a92%probability of a credit crisis and we should pay enough attention to take the necessary measures.The innovation of this paper is mainly reflected in:using the poor volatility instead of equity volatility in the calculation model. At present,papers based on Naive model and LT model are extremely rare compared with KMV model in China, we use new data to examine the ability of measurement and management. This paper uses ROC curve and CAP curve to campare the ability of identity of the three models. The equation of ROC curve based on the Logistic function can increase the stability of the analysis.