The Analysis of Mortgage Prepayment in Kao-Ping Area

[[abstract]]由於2008年爆發全球金融危機，大部分的銀行對於中小企業融資相當嚴謹，並轉而關注房屋抵押貸款，然而，房貸市場過熱的競爭使得利差大幅減少，利潤也同時下滑。因為房貸的提前清償會衝擊房貸的現金流量，並導致房貸評價的不易，因此找出影響提前清償的因素是非常重要的。本研究應用邏輯斯迴歸模型分出提前清償的特徵變數，進而希望提供銀行制定房貸策略及篩選貸款群組的參考。 本研究蒐集某銀行在高屏地區的1 ,336筆房貸戶資料，（其中包括369筆正常貸款，及622筆提前清償貸款）進行實證分析，從運用正向逐步選取法的邏輯斯迴歸模型對15個特徵變數的實證結果顯示年齡、家戶年收、攤還方式等顯著變數與提前清償為正向相關，教育程度、職業別、地區別、住宅類型、貸款成數、房貸所得比及借戶性質與提前清償為負向相關。[[abstract]]Due to the global financial crisis in 2008, most of banks have become very strict in offering the small and medium enterprises (SME) finance, and turned to focus on the home mortgage loan. However, the overheating competition in the mortgage lending market has considerably decreased interest spreads and profit margins. Because the mortgage prepayment would impact the cash flows of the mortgage loan, resulting in difficulty in evaluating the mortgage loans. Therefore, it is very important to find out the factors influencing prepayment. This study applies the Logistic regression model to discover the key variables of prepayment. We hope to provide a practical reference for banks to make the strategy of mortgage loan and select the appropriate mortgage pool. This study collects 1,033 loans of clients (including 369 normal loans and 662 prepayment loans）from a certain bank in Kao-Ping area for empirically analysis. The empirical results from Logistic regression model applying forward stepwise method on the 15 key variables show that the significant variables of prepayment such as age, household annual income, and type of payment have positive relation with prepayment. Meanwhile, the variables having negative relation with prepayment are education, occupation, location, type of house, loan-to-value, payment-to-income, and borrower's status