|
This study aims to assess the consumer mortgage loan default decrease the quantity of influence factors. Commercial banks by the credit business has gradually reduced the proportion of profits, in a competitive environment, in order to seize the market, easing credit review, resulting in low credit quality, settlements, eroded profitability, credit risk management is therefore a financial institutions important aspect of business strategy. Risk management must first make the best of the audit before the loan decision-making; that is provided by loan applicants in accordance with basic information, assessment of credit risk rating Jiekuan households, as balance amount of the loan criteria.
This study of a domestic commercial banks from Tainan, Republic of China Republic of China 98 years 94 years until the housing loans between the loans to the case study sample, the normal payment of interest and normal household items 375, default users have 45 non-normal conditions, giving a total of 420. Variables in this study a total of 15 variables, in order to understand the variables of loans have significant effect on the success or failure, to assist the right to credit officers to efficiently determine the most precise data to correct credit decisions.
On empirical estimates, and use logistic regression model and decision tree analysis to explore the impact of consumer housing loans to credit default risk of the main features of factors, logistic regression models in the empirical results, income, guarantor, the borrower into a few, grace the five a significant explanatory variable of 0.01 standard, and in the decision tree analysis Zeyi node 2 (with grace) of the forecast default rates of 36.8%. To not use the grace of the borrowers concerned, but also by the impact of annual income, annual income by 50 yuan forecast default rate was 21.7% households with annual income of over 50 million borrowers default rate forecast is only 3.1%. To establish as future consumer mortgage credit assessment based on quantitative models, as well as future loans approval or rejection based on reference.
|