;In view of the previous mortgage credit risk assessment models, we have taken the questionnaire or random sampling selected sample data, but in order to avoid error generation. This study done for domestic bank loans up to a branch between 2006 and 2007 for two years, all cases of long-term housing loans actual allocations, including cases of cancellation repayment halfway, total 556 cases, All as research subjects, including the loan amount, number of years of the loan, the loan interest rate, gender, marital status, children number, education, occupation, income, loan-to-income ratio, other assets-real property, other main debt and from debt, the use of revolving credit etc., coupled with the aboriginal for the study variables, use LR model analysis to compare the influence of variables on housing loans in occurred overdue credit, as credit review reference in the bank loans.
14 variables of involved in the empirical estimation of this study, only in the term of the loan, the borrower gender , other assets-real property etc., and these three results show, for credit default less influential. including the remaining interest rate loans, education, loan-to-income ratio, other main debt and from debt, aboriginal etc., significant influence variables, and the loan amount, marital status, child-rearing, occupation, annual income, with or without the use of revolving credit, etc. variable., these variables through empirical estimate the extent of its impact on the occurrence of overdue loans. Banks better grasp of the borrower′s condition and the risk of the borrower′s credit. And then again to confirm where the level of education, loan-to-income ratio, other main debt and from debt, occupation, annual income, the use of revolving credit etc., Bank credit is still the focus of the audit, even heavier weight of these items in the credit risk assessment model, in order to reduce the risk of bank credit and avoid the occurrence of overdue loans.