摘 要 本研究希望利用各種不同貸款特性,並透過國內中部某銀行5家分行提供的房屋抵押貸款資料進行Logit迴歸分析。希望能藉由分析貸款人提前清償的行為,對往後銀行的風險的控管與承做徵授信業務時,能有更進一步的幫助,利用真正的反映出客戶的信用風險及忠誠度,帶來平均成本降低、費用減少,以及提升風險承受能力…等好處,創造雙贏的市場利益,對於微利時代金融產業之重要性不可言喻。 針對國內中部某銀行5家分行之不動產抵押貸款核貸案件為研究對象,以2003至2007年之個人購置住宅及房屋修繕貸款為研究樣本,其中包含未提前清償戶及提前清償戶,其取得貸款案件都經核准並完成建檔及撥貸程序,樣本數量合計為1,210筆。採用logit迴歸分析,研究結果顯示: 年齡、學歷、有無保人、貸款金融、利率條件、屋齡、所得、買賣價格、所得;婚姻、建築物等因變數會影響房屋抵押貸款提前清償行為。 關鍵字:不動產抵押貸款、提前清償、logit迴歸
Abstract I hope to use numerous mortgage characteristics, and use logic regression model to analyze personal mortgage data offered by 5 branches of bank. The propose of this research is to give suggestion to reveal customers’ loyalty and credibility truly, thus decrease average cost and other expense, and increase industry’s risk bearing capacity. The significance to give that kind of research outcome will be obvious for financial industry. In this paper, I research the cases of mortgage branches of a bank located in central Taiwan, the cases are containing individual purchase of residence and mortgage containing home repair, mainly from 20003 to2007, including the one who has not and have paid off the household ahead of time. All the above 1210 cases are approved, filed, and give loan. The outcome of logic regression model in this research reveals the following elements will influence the behavior of customer whether they have paid off the household ahead of time. They are age, marriage, degree, occupation, revenue, the rate quality, the age of home, income ratio of sale prices, the classes of buildings. Keyword: mortgage loan、repayment、logic regression model