Individual Housing Mortgage Risk Assessment Of Empirical Research

As the housing system reform progresses during last ten years, residential mortgage loans has become the major financing measure by which Chinese residents buy their houses. The amount of mortgage loans increases at a speed of 40-50% or higher nowadays, but the unperformed loan rate increases significantly as well. Although the mortgage loans belong to the high quality assets in the balance sheet of banks compared with other commercial loans, but their potential risks gradually emerge. On the other hand, as the state-owned commercial banks were long affected by bad debts, they are unusually sensitive to the loan repayment speed while ignoring their profitability. Hence prepayment was often neglected, even advocated. This trend has affected the expected return of the commercial banks. However, empirical studies on default risk and prepayment risk of residential mortgages in China are relatively few. Evaluating credit risk rating has been one of prime content of credit risk management of residential mortgage loans, and there is trend that evaluating credit risk rating with scientific and quantitative methods.Under the background mentioned above, this thesis focuses on the empirical analysis about the major determinants of default and prepayment risk. This study will help banks build theoretic and methodological foundation to control the default risk and prepayment risk at the origination of residential mortgage loans, and providing foundation for bank evaluating credit risk rating of residential mortgage loans.The mortgage loan sample data is from some commercial bank during the period from the year 2000 to the year 2004. After the deleting the samples that have incomplete records or abnormal data, 12911 effective samples are left, in which 6916 samples are of normal status, 3221 of default status and 2774 of prepayment status. Based on the literature review this thesis classifies the variables into three dimensions: borrower characteristics dimension, loan characteristics dimension and property characteristics dimension. 19 variables such as borrower's gender, age, loan-to-value etc enter these three dimensions. Using SPSS10.0, this thesis conducts descriptive statistical analysis, discriminant analysis and Logistic regression. The empirical results indicate that...