Credit Risk Management Scoring Method For Empirical Studies

Now, commercial banks have focused on the development of credit card business. In current market economy the credit card business is a great potential new profit growth point. So it has become an important target for which domestic commercial banks are competing. However, the objective law of the financial markets tells us that the risks and benefits are proportional. So behind the potential huge profit of credit card business, great risks exist, most notably credit risk. The quality of the prevention and control of credit risk directly affects the profitability of commercial banks. So we must accurately assess the risk in scientific means, so that we can develop the right strategy to effectively control the risk. Now the international successful approach to risk management is based on credit scoring models. However, as China's current conditions for the credit economy development are not fully mature, if we directly use foreign credit scoring models in China's credit card markets, it wouldn't fit. In this paper, based on domestic and foreign credit scoring models, combined with China's actual situation, empirical analysis has been done on the method, content and the principle of credit customer assessment.The thesis has five chapters. The first chapter briefly introduces the domestic and international research on the credit scoring models, and briefly analyzes the background of research. The second chapter introduces the credit card's development history as well as credit card business development condition in our country, discusses the main risk of the credit card business, and generally analyzes the significance of credit scoring models. The third chapter theoretically discusses the cumulative logistic regression model in detail. Taking into account the problem background, three cases of the cumulative logistic regression model are analyzed in order to choose the best model. Then we theoretically elaborate the proportional odds model, and finally discuss threshold selection method. The fourth chapter discusses how to establish credit scoring model which suits present situation in our country. Based on the sample selected from some bank's material of credit card customers, we analyze the determination of response variables, the selection of forecast variables, the quantification of qualitative indicators, the modeling and determination of criteria. Finally we establish the most appropriate cumulative logistic regression model for China's market. The fifth chapter summarizes the real diagnosis result and forecasts possible use of this model in the future.