Kmv Credit Risk Measure Based On The Threshold Regression Model Default Point Correction

From the perspective of credit risk management, and base on the model of KMV credit risk measurement, this paper modifies default point of KMV model by using the threshold regression, taking the market information and accounting information of listed companies which default for the first time in the five years from 2005 to 2009 as samples, and using short-term liabilities, long-term liabilities, total liabilities, short-term assets debt ratio, long-term assets liabilities ratio and the total debt ratio as threshold variable in turn. It aims to find out the model default points which are suitable to the actual situation in our country, so as to achieve the purpose of refining the model, and using the model to accurately measure the credit risk of listed companies in China.This study finds that when we use long-term liabilities, short-term asset liability ratio and liabilities rate of total assets as the threshold variable, the default point of China's listed companies have a clear threshold effect. Meanwhile, the paper verifies the model the default of which has been revised using the total A-share listed companies from 2005 to 2009. The results of K-S test and Mann-Whitney U test shows that the revised model can significantly identify the normal business enterprises and the default enterprises, and the results of ROC curve, CAP curve and AR indicate that the capacity and efficiency of the modified model in predicting the default enterprises has improved significantly than before.