Customer Credit Scoring And Credit Card Risk Management

Credit card is a global and modernized payment tool, which makes it convenient and safe for credit card holder to consume and increases the sales of authorized merchant. What is more, credit card is an important source of profit to commercial bank. In many foreign commercial banks, the profit of credit card business is general exceed thirty percentage of total profit and so the credit card becomes the core business for most foreign commercial banks. While with the in-depth reform of our financial system, domestic commercial banks have launched credit card out of the profit compression of the traditional business. Starting from the first credit card issued by Zhujiang Branch of China Bank on June, 1985, our credit card industry developed rapidly with the card-issuing institutions up to more than 130, the issuance exceeding 180 million and the issuing value accounting to 13634 billion until the end of the third quarter of 2009, showing a "blowout" type outbreak of the development. However, with the rapid development and the objective law of risks and return, the financial institutions hunt huge profit; meanwhile it have to face tremendous risks, credit risk especially. Therefore, how to avoid credit risk is an important subject to banks. The banks need to make scientifically credit evaluation of information of credit costumer before granting credit card. In the western countries, it is commonly adopted personal credit evaluation of quantitative analysis to evaluat credit status of individual customers. Yet because of the imperfect social credit reporting system and backward credit consumption industry, commercial banks'personal credit models exploiting just start in China and it becomes a great challenge to our bank must how to ensure the quality of assets while expanding the scale of business. Therefore, it is one of important topics of commercial banks solving credit risk that developing a rational and effective method of single consumer credit scoring in commercial banks.Keeping with this credit card status, the paper starts with the reason of credit card risks, divides the current risks of credit card business and describes different methods of risk measurement and control to the four stages of the life cycle. Then the paper introduces the credit scoring which is the most efficient tool in the risk management to all aspects of credit card risk management. Firstly, it is introduced several most advanced methods of building credit scoing model like Discriminant Analysis, Regression Analysis, Neural Networks, Decision Trees, Nearest Neighbor Algorithm, etc and after comparison, it is concluded that Logistic Regression method has better performance in applicability, stability and accuracy. So in the latter empirical part, the research establishes a Logistic regression model including 12 characteristic variables and a new consumer credit scoring model with the credit card holders' information of Hunan Branch of A bank as sample data and fixing index system of credit score. The conclusion shows that the month income, housing condition, other asset and burden of population of single consumer have a significant effect to the model. As to the model itself, the paper makes the effectiveness test afterward and finds the Logistic model fit good and the model's classification accuracy rate of good and bad customers can reach 82.5%. At last, this paper makes a number of appropriate political recommendations for status of credit card risk of China's commercial banks.