本研究針對國內T商業銀行的高齡房屋信用貸款為依據,採自 2016 年1月至2020 年12 月間,向該銀行申請個人房屋貸款之中高齡客群總數為抽樣母體,由於本研究抽樣母體採限制對象(針對中高齡借款人),5年內(2016~2020)樣本總數約500~600件,依個案銀行放款資料庫之資料所示,平均每月抽取6~8件、總件數400件為本研究之樣本,以邏輯斯迴歸分析此400件樣本(正常:338件,違約:62件),並進而探討高齡購屋逾期放款的主要影響因素。本研究應用邏輯斯迴歸分析模型,結果顯示在投入 13 種自變數中,性別、婚姻、貸款用途、薪資收入、職業、年資、保證人、信用卡循環等8項具顯著性影響;而不動產座落區位、不動產型態、鑑價金額、借款金額及借款成數等5項不具顯著性影響。本文研究結果可提供銀行未來針對貸款的中高齡客戶調整授信政策的參考依據,以增加違約戶之篩選進而減少金融機構損失。
This study aims at the senior house credit loans of a domestic commercial T bank to collect the data on the total number of old house loan customers who have applied for a personal housing loan from this bank from January 2016 to December 2020 and to use this data as the population. Since the population of this study restricts the targets (to middle-aged and senior borrowers), the total number of samples in 5 years (2016~2020) is approximately 500~600. According to the data of the bank loan database of the case, 6~8 samples are sampled monthly, and the total number of 400 is used as the samples of this study. Logistic regression analysis is used to analyze the 400 samples (338 are normal, and 62 are default) and explore the main factors that affect overdue loans of the elderly. This study applies a logistic regression analysis model, and the results show that 8 out of 13 independent variables including gender, marriage status, loan purpose, salary income, occupation, seniority, guarantor, and credit card debt have significant impacts; and the remaining 5 independent variables including real estate location, real estate type, appraisal amount, loan amount, and loan ratio have no significant impact. The results of this study can be used as a reference basis for banks to adjust their credit policy for middle-aged and senior customers in the future, so as to increase the screening of defaulters and reduce the losses of financial institutions.