房屋貸款逾期還款預警模式之研究

A WARNING MODEL FOR PREDICTING OVERDUE REPAYMENT OF LOAN

林國順(Kuo-Shun Lin)

指導教授 ： 洪育忠

大同大學/經營設計學院/事業經營學系所/碩士(2004年)

摘要

逾期放款比率過高，一直是金融機構欲積極改善之項目。改善逾期放款之方法，事前確實審核申請人之資格，對抵押品確實鑑價，找出影響借款人逾期還款的因素，並以此篩選出高風險的客戶，未雨綢繆防範於未然，實在是最重要的工作。
放新銀行開放營業以來，國內銀行的家數成倍數增加，國內金融業開始進入競爭的戰國時代，激烈競爭的結果，各銀行的存放款利差逐漸縮小。由於競爭者過多，加上近年來房地產景氣持續低迷，金融機構承作房貸的業務風險不斷增加，使得授信品質不良，金融機構的逾放比率不斷上升。
此外由於股、匯市的開放，使得企業籌資的管道大增，銀行已非企業籌措資金的唯一管道，企業授信業務由賣方市場改變為買方市場。為了吸引客戶，調降借貸利率，甚至降低授信標準要求，來爭取客戶等現象亦時有所聞。長期下來台灣金融界遂產生授信浮濫，信用擴張過度，超額借貸與壞帳比例上升等問題。
為爭取放款時效、減少人為疏失及加快審核時間，使用電腦化的審核制度變成為當務之急。
本研究之目的為探討影響房屋貸款申請人信用風險之因素，並建立房屋貸款審核系統。期望金融機構藉由本系統能客觀偵測申請人之房貸風險高低狀況，以作為授信審核之依據，從而降低房屋貸款的逾期比率，提升經營績效增加收益。
本研究採用LR 模型、區別分析模型作比較研究分析。研究結果發現，影響房屋貸款申請人房貸風險的因素，例如職業、教育程度、年收入及貸款成數等都有顯著影響。
研究結果顯示，LR 模型之準確率為71.5%，區別分析模型之準確率為71.7%，相差不大，但以區別分析模型之準確率較高，但兩者皆可以作為授信評量之模式。

關鍵字

羅吉斯迴歸 ； 風險評量 ； 逾期放款 ； 區別分析

並列摘要

The high rate of overdue loans is always an issue that financial organizations want to improve. The solutions to this problem are: thorough examination of an applicant’s qualifications before loan approval, careful appraisal of collateral, identification of factors which influence delinquent payment by a debtor, and exclusion of potential high-risk clients. “Prevention is better than cure.” Early preparation is indeed indispensable.

Since regulations were relaxed, allowing new banks to launch in Taiwan, the number of banks here has increased sharply. Thus, an era of competition in the Financial industry was ushered in. Due to cutthroat rivalry, profit margins of all banks started to narrow. Moreover, owing to too many competitors, and a sluggish economy in real estate trading, risks of home mortgages are spiraling in financial institutions. Consequently, the quality of credit is lowered, and the rate of overdue repayment is skyrocketing, too.

In addition, liberalization of the stock market and foreign exchange broaden the access to capital for enterprises. Banks, therefore, are not the only source of acquiring money any more. The loan business adapts from a sellers’ market to a buyers’ market. To appeal to clients, it is often the case that interest rates are cut, and even standards of loan approval are reduced. This spurs lax rules of loan approval, over-expansion of credit and rising percentages of over-borrowing and bad debt.

With a view to gain efficiency, minimize man-made errors and shorten examination time, the time has come to utilize an automatic evaluation system.

The aims of this research are to investigate factors affecting the credit risk of applicants, and establish an assessment system for mortgages. It is hoped that financial institutes can use this system to quickly and objectively detect the risk status of loan candidates. Furthermore, the system can be taken as a basis of loan approval. Rates of overdue repayment will be brought down, operational performance enhanced, and profit will increase.

This study adopts LR model and discriminant analysis model to compare and diagnose the data. It is found that factors having an impact on risks of loan applicants, such as occupation, educational level, annual income and the ratio of loan approval, are all significant statistics.

The outcome reveals that the accuracy rate of LR model is 71.5%, compared to 71.7% of discriminant analysis model. While the figures are very close, the latter is higher than the former. However, both models can be employed as criteria in loan examination.

並列關鍵字

discriminant analysis ； logistic regression ； overdue repayment ； risk evaluation

參考文獻

Altman, E. I, “ Financial Ratio Discriminant Analysis and the Prediction of Corporate Bankruptcy,” Journal of Finance, Vol.23, No.3, 1968, pp.589-609.

Campbell, T. S. and Dietrich, J. K.,”The Determinants of Default on Insured Conventional Residential Mortgage Loans,” Journal of Finance, Vol. 38,No. 5,1983,pp.1569-1581.

Galitz, L. C. ”Consumer Credit Analysis,”” Management Finance, Vol. 9,No, 1, 1983, pp. 27-33

Martin, D., “Early Warning of Bank Failure: A Logit Regression Approach,” Journal of Banking and Finance, Vol. 1,1997, pp.249-276.

Smith, L. D., Sanchez, S. M. and Lawrence, E. C., “A Comprehensive Model for Managing Credit Risk on Home Mortgage Portfolios,“ Decision Sciences, Vol.27, No.2, 1996, pp.291-317