本研究以國內中部某公營銀行承攬之613件不動產抵押貸款案件授信戶為研究對象,透過交叉分析及卡方檢定,進而以羅吉斯迴歸分析模型之建立進行實證,分析影響銀行不動產抵押貸款授信品質之潛在因子,探討借款人之年齡、性別、婚姻狀況、住所是否為自有、教育程度、年所得收入、服務年資、職稱、聯絡人、借款人是否曾更名、保證人數、貸款期限、撥款方式、不動產種類、不動產座落區域、有無其它授信、與本銀行往來期間等相關影響授信品質的變數。實證結果發現,當借款人為男性、借款人曾經更名、借款人向銀行設定的借款期限越長、借款人為未婚、借款人年所得收入越低、借款人未填聯絡人、借款人抵押之不動產座落區域非本地、借款人無其它銀行授信時,預測其成為異常借款戶的機率就越高。
In this study, the domestic central bank of a public contract of the 613 cases the borrower''s real estate mortgage loans for the study, through cross-analysis and chi-square, logistic regression analysis and then to establish an empirical model, analyze the impact of bank mortgage-credit potential quality factor of the borrower''s age, sex, marital status, whether their own residence, education level, annual income, years of service, job title, contact person, whether the borrower has been renamed to ensure that the number of loan, grant way, property type, property located in the region, with or without other credit, and the period between the Bank and other relevant variables affecting credit quality. Empirical results found that when the borrower is a male, was renamed the borrower, the borrower''s loan to the bank to set the longer the maturity, the borrower unmarried, lower income person-years loan, the borrower did not fill in contact person, the borrower''s property mortgage Block off the area of non-local, borrowers, when no other bank credit, as an exception to predict the probability of higher household borrowing.