Research On Credit Risk Measurement Of Listing Corporation In China Based On Logistic Regression Model

Credit is the foundation of market economy and is also the precondition and foundation of the normal operation of market economy.As the credit transaction has gradually become the main way of daily economic transactions,credit risk has been penetrated into all aspects of daily life.In recent years,as issuers of stocks and corporate bonds,the scale of the listed companies continues to grow in our country and the status of the capital market is so important in our country.However,credit risk events occur among listed company repeatedly,resulting in great loss to our investors and commercial banks.Strengthening the credit risk measurement and management of listed companies in our country can not only provide reliable reference to commercial banks to manage the credit risk of listed companies,but also provide beneficial references for scientific,rational investment information to China’s capital market investors.Through specific researches to the present situation of credit risk in the listed companies in China,we can find that listed company credit risk measurement method in our country is relatively backward and the measures in such aspects as index selection,matching proportion also have some shortcomings.After comparing the assumptions,data requirements,forecasting accuracy and stability,and the applicability of credit risk measurement models we found the Logistic regression model is better in every way to measure and research on the credit risk of listed companies in our country at present stage.According to the above situation,we choose 43 ST companies in 2015 among the Shanghai and Shenzhen a-share listed companies in our country as the credit default sample,and use the sample matching ratio of 5:1 to select 215 non-ST listed companies as the non-default sample.After that we include 24 financial indexes in seven aspects such as the primary company scale,corporate governance,debt paying ability,operation ability,profitability,capital structure and growth ability and combine with the Kolmogorov-Smirnov test,T test and the Mann-Whitney test to filter out 19 indexes which have significant differences in the two groups of samples.Then we use principal component analysis to extract the principal component factors as input parameters of the Logistic regression model to build models for credit risk modeling and prediction.Through the sample test,Logistic regression model was verified the most suitable in data requirements,parameter selection and the prediction accuracy in the listed company credit risk measurement in our country at present.Finally,combining the empirical results and the present situations of our current country,we put forward three suggestions to improve the accuracy of the credit risk measurement of listed companies in our country.