Early Warning System Of Financial Crisis On The Base Of Credit Risk Management

The credit risk is one of the oldest risks in the financial market and is also the core in the management of commercial bank. There are lots of corporate be found up and dye everyday. The foundation, development and dying of corporation is bound up with commercial bank, so it has become the focus of theory and demonstration that how to find the sign of financial crisis. The theory of financial crisis be introduced into the credit risk management of commercial bank is very important in the theory and reality.The credit risk managers of commercial bank focus on the profitability and ability of repayment, especially repayment in short-term. By using the early warning system of financial crisis the message of corporation's financial statement can send to the credit managers in time, and bring down the credit risk utmost. Since Altman applied the multiple linear regression model to the prediction of corporate financial risk, the early warning model of financial crisis has been widely used in the field of corporate credit ranking and loan ranking. The early warning model of financial crisis includes single variable model, multiple linear regression model, logistic regression model and ANN model etc.On the basis of illustration of theory of early warning of financial crisis, this paper compares the adaptability, advantage and shortcoming of each model. Using financial data of"special treat"corporations in Chinese stock market we choose line probability model and logistic regression model as the credit risk early warning model. The logistic regression model is excellent than the line probability model, so it should be the first choice in credit risk management of commercial bank.The final line probability model and logistic regression model are made up of cash flow variables. From this we can know that the corporation's cash flow include more information about its financial condition, and the model based on the cash flow has higher correct ratio.