Study On Credit Risk Management Of Individual Loan In Commercial Banks

In the past 10 years, with the rapid economic development, China's consumer credit market showed supernormal trends. The rapid growth of loan size has brought great profits for the banks, but also planted a huge risk, particularly with the high housing price recently. The society gradually changes from "All people with deposit" to "All people with debt".We can learn from the U.S. "Subprime Crisis" that in a huge change of macroeconomic and monetary policy, personal loan shows great vulnerability and infectious, and the risk exposure to personal loan tends to occur in 8 years after the agreement signed. With the background of tight monetary policy in China now, how to prevent the risk of individual loans and reduce the possibility of default by the borrowers is an important issue to China's commercial banks. It has also been widespread concerned by many scholars.In the study of the borrower breach, there is a big difference between the "voluntary breach" and the "forced breach" caused by the external environment. Voluntary breach is a behavior from the inside out, determined by the type of occupation, income, educational level and a series of other property of the borrower. It has to make the judgment in the review before loan for the banks. Forced breach of borrower is a behavior from the outside in, due to the great changes of economic environment, such as CPI increasing, higher interest rates, etc. Borrowers have to default to make a living. Therefore, the borrower breach behavior is determined by many complex factors and we need to make a Systematic study.In 1999, the "New Basel Capital Accord" provided a variety of tools to measure the risk accurately for commercial banks. Since then, the worldwide commercial banks can operate in a unified, efficient and fair framework of risk measurement for their assets and the corresponding capital adequacy ratio. BaselⅡprovides many new statistical methods and artificial intelligence algorithms which allow banks to make more accurately calculation of exposure to risk status. Meanwhile, the newly proposed concept of "Stress-Testing" can also help banks determine in advance the possible damage which do not happen at that time to their assets. In this paper, we use the personal loan data from a commercial bank of prefecture-level city in Shandong Province to make an empirical study. First, we combine the national macroeconomic policy changes with the data to explain the fluctuations of the data, and make the single factor analysis. Second, we use a Logit model to analyze how the characteristics of the borrower affect the probability of breach. Third, we introduce the Support Vector Machine, a more advanced neural network model, for more in-depth study of the data. We divide the data into training group and test group. After training and adjusting the Support Vector Machine, we establish a model with high prediction accuracy. Finally, we operate the stress testing by simulating fluctuation of three macroeconomic variables to observe the change of each sample and the overall exposure to risk.The research shows that the reasons which could significantly reduce the borrower's probability of default include:relatively small age, marital stability, high level of education, stable work and income, provide of collateral, short term loans. In the empirical analysis of Support Vector Machine, we can see that this model has a very high prediction accuracy in the individual credit default prediction. In the total 11351 personal loan samples, it achieves 100% prediction accuracy. Compared to traditional neural network models, SVM has great advantages in efficiency, accuracy, computing size and other aspects. It can be widely used in China's banks.Through the micro stress testing on the samples, we found great vulnerability in individual credit assets. They are very difficult to resist large macroeconomic shocks. When faced with falling house price, CPI rise and interest rates rise, the borrowers can not take the repayment pressure. They are prone to be "rational overdue", "insolvent", and have to stop repayment to maintain basic life.In the background of tight monetary policy and house price control, the government should make the economic "soft landing" and pay close attention to the changes of individual loan default in policy implementation to prevent the outbreak of risk.