Study On Credit Risk Based On Logit And KMV Model

With the continuous development of science and technology, the proportion of China’s high-tech industries in the national economy is steadily rising.As a symbol of China’s science and technology competitiveness, high-tech industries has become the fundamental power of socio-economic.The high-tech industries is the multiple groups which often occur credit risk,that because it have the characteristics such as the high technical content,serious risk,quite long development cycle and so on which make it has much serious financing problems.Now,the question of financing difficulties have become the restrict which shut down on the development of new and high technology industries.How to solve the supply and demand contradiction between the high-tech industries and commercial Banks,to realize win-win has plagued professional visitors. In order to solve the financing difficulty and help the high-tech industries of China to realize high speed development of economy, to study the credit risk of high-tech industries empirical research has very important significance.For companies,the occurrence of credit risk is a gradual change process,which can’t produce overnight, so if we can control the occer of high-tech industries before it happen and carry on effective management to control credit risk, the incidence of the credit risk will minus.There are so many literature studies of credit risk problem, but the studies of high and new technology industry credit risk literature is very little. Even though the existing literatures have show some measurement models which to study credit risk, but when different models do research on same issue not always have the same conclusion.So in this article we select two credit risk early warning modesl,the Logit regression model and the KMV model,which suitable for the present development of listed companies in our country,and has carried on the empirical researches with the two models to study high-tech industries.The two credit risk early warning models both have practical significance,not only for commercial banks to manage the credit risk of listed companies to provide reliable reference content, also can provide reference information for China’s capital market investors.The study in this essay,we make high-tech industries as the research’s objects,we regard those ST enterprises as the default ones,regard those non-ST enterprises as the normal ones. First of all,we combine factor analysis and Logit regression and then has carried on empirical research with select financial data as the indicator variable, using SPSS to inspect variables,then,carry out common factor by factor analysis method and Logit regression analysis. In the second place, we also apply KMV model to carry on the empirical research of high-tech industries, and compares the Logit regression analysis research.Finally, we compare the result of KMV model results to the Logit regression model.The empirical study is draw the following conclusions:(1) Logit regression model and KMV model have high prediction accuracy,but under different cut value the prediction accuracy of credit risk is vary.When the cut value of Logit regression model is 0.5,KMV model’s prediction is lower,and when the cut value is 0.647,KMV model’s prediction precision is higher.(2)If you want to lower the probability of the type-1’s error and to higher the forecasting precision,you can use Logit regression model with the cut value 0.5.(3) If you want to lower the probability type-2’s error, you can use the Logit regression model with the cutting value of 0.647, but at this time the precision is lower. So we can choose different models according to the demand.The innovation of this article:(1) Make high-tech industries as the research’s objects,and do research on high-tech industries.(2)We choose different threshold,with theLogit model and KMV model to carry on the comparative analysis of high-tech industries’credit risk.The shortcomings of the article:(1)The sample data are limited.Because it is difficult to get high-tech industry customers’internal information from bank.In order to study,we learn form foreign scholars use ST and non-ST companies to decide whether companies default or not, and compare the results with the real situation to observe the effectiveness of the model.(2) Without considering industry factors. As the research object of this article is high-tech industries,there is no study of credit risk in other fields, the determination of model prediction accuracy unavoidably have one-sidedness.