Study On Credit Risk Prediction Of City Commercial Banks In China

Since the financial crisis explosion in2008, many large commercial banks fall into financial crisis in the United States and European countries, and a large number of small and medium commercial banks have been bankrupt. Chinese banking industry is thriving; especially city commercial banks which have the largest single numbers acquire more sturdy development in this period. They not only quickly completed changing the name and the system, but also are developing toward mixing mode. However, Qilu Bank’s outbreak of serious financial fraud in January2011showed that credit risk is becoming one of the biggest risks that Chinese city commercial banks are facing, directly influencing their stability and development.According to this, this article pays attention to Logistic model’s use in Chinese city commercial banks predicting listed companies’ credit risk by using the comparing and empirical methods. The article first compares and analyzes domestic and foreign credit risk measurements and discuss various measurements’ applicability in China, then considers Logistic regression model fit to predict listed companies’ credit risk comparing with modern models such as Credit Metrics model, Credit Risk+model, Credit Portfolio View model and KMV model. That is because Logistic regression model solves nonlinear problem and has higher accuracy. After that analysis, the article chooses100listed and100non-listed companies that have complete financial data and loan in city commercial banks from2008to2010as study samples (totally600groups of data), and uses400groups of data from2008to2009as controlling groups to construct principle component Logistic regression model, which has four principle components influencing listed companies’ default probability. These four principle components are named short term debt paying ability, long term debt paying ability, income quality and operation ability, including liquidity ratio, quick ratio, speed ratio, stockholders’ equity ratio, long term assets for rate, the rate of working capital with loan, equity ratio, equity multiplier, surplus cash protection rate, inventory turnover ratio and equity turnover11key variables. Then the article uses200groups of data in2010as predicting groups to examine the principle component Logistic regression model. The average accuracy is85.57percent. The accuracy of predicting normal customers is85.71percent and the accuracy of predicting defaulting customers is85.42percent. Finally, according to the study result, this article provides same suggestions to push forward credit risk management of Chinese city commercial banks.