The Factors of Default Risk for Mortgage Loan - A Case Study of A Bank

• Main Authors: Rong-Jie Shi, 史榮傑
• Other Authors: Kun-Min Sie
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/75143741422505641909

碩士 === 國立高雄應用科技大學 === 金融資訊研究所 === 101 === The goal of this research is to study the factors that affect the credit risk of mortgage applicants. These have twelve risk variables are used in credit risk model . Logistic Model , Discriminant Analysis and Decision Tree model are used to analyze the relationship between risk variables and defaults. The results show that under Logistic Model the independent variables of Career , Annual Income , Guarantor , Income-To-Expense Ratio , Loan Ratio , and Property Type are statistically significant ; under Discriminant Analysis Model , Property Type and Guarantor have biggest and smallest contribution respectively and level of contribution of Career , Income-To-Expense Ratio , Annual Income , and Loan Ratio are between that of Property Type and Guarantor ; The results Decision Tree Model show that Annual Income , Property Type , Loan Use , Marriage are statistically significant . Comparing the precise of these models , Logistic Model has the best prediction for non-default . However, Discriminant Analysis Model has the best prediction for default . Overall , Logistic Model of 91.3% and Decision Tree Model of 87.6% perform better .