Based On Empirical Study Of Corporate Credit Risk In Zhejiang Ccb Cmis

The financial operations of commercial banks become diversified, but the business loan is still one of the most important operations for the commercial banks. Aim at the current status that the bad loan rate of the national commercial banks is very high, it is important to measure and manage the credit risk. Also, it is practically meaningful to control the increasing of the bad loan.Above all, this study summarizes the literature of firm credit risk, and introduces the structure and application of the logistic regression model. Then by using the loan enterprises in the credit management information system (CMIS ) of China Construction Bank Zhejiang Province subbranch as the research samples to developing the logistic regression model, the study obtains the likelihood that the firms do not reimburse the loan in time, that is the possibility of default(PD).The thesis uses 169 defaulted firms and 338 non-defaulted firms as the samples to develop the model, and the firms involve the four industries: manufacturing, architecture, wholesale & retailing, real estate. The thesis chooses 16 financial variables and 4 dummy variables to participate in the empirical analysis. Obtain the logistic regression model about the 4 industries firms by t test, Mann-whitney U test, collinearity test, data configuration test, backward: wald regression. Then the study tests the predictive validity of the model. The final regression model indicates that: the PD has important relationship with the industries: architecture, wholesale & retailing, real estate. The most significant predictive measures within the model also include: velocity of total assets, the log of total assets, equity ratio, working capital to total assets.To test if using the single industry enterprises to develop the model will obtain better validity or not, the study develops a regression model about the manufacturing firms. The samples are the manufacturing enterprises used in the former study. The step is the same with the former study. Also it tests the predictive validity of the model. The final model indicates that the PD of manufacturing firms is significantly relative to the 4 measures: return on equity, liquid ratio, the log of the total assets, working capital to total assets.Form the test result of the predictive validities of the two models, it can be found that comparing to the predictive validity of the mixed industry firms model that take into account the dummy variables, develop with single industry firm the model' predictive validity is not significant increased.Finally, the thesis gives a detailed analysis and interpretation of the empirical result, and put forward 3 points of inspiration, 4 pieces of suggestion and the future research prospect.