- 學位論文
消費性貸款授信風險評估之研究—以X銀行為例
The Research on Evaluating the Risk of Consumer Loans－Taking X Bank as an Example
摘要
自1997年亞洲金融風暴後，經濟景氣下滑、產業外移、銀行授信業務趨於保守，逐漸將放款重心由企業金融轉為消費金融，現在成為銀行主要獲利來源。然而，隨著景氣低迷失業率攀升，承作消費性貸款的潛在風險不斷升高，如何評估客戶的信用風險，在提高消費者貸款承做量之同時，如何將違約風險降至最低，應為有志發展消費金融的銀行最迫切的課題。 本論文主要目的以建立消費性貸款審核系統，本研究以某銀行消費者貸款案件為研究對象，以2001年1月到12月底於該銀行辦理消費貸款案件為抽樣母體，為了建立模型，由正常貸款抽取231件，逾期案件抽取231件，利用羅吉斯迴歸(Logistic-Regression)進行資料分析。 以性別、服務機構、年資、薪資、婚姻、房屋、存款、信用狀況、區域及核准利率作為預測變數，結果本模型中對逾期案件的預測準確度為77.9%，對正常的貸款件的預測準確度為79.2%。 以下就本研究實證結果，擇要簡介如下： 一、性別因素為評估違約機率的重要因素：本研究案例及其他研究顯示女性的違約比率較男性低，可將此一因素列入銀行評分表。 二、因科技的快速進步，使得產業變遷造成結構性失業，年資愈長者信用風險未必愈低，宜於評估信用風險時考慮申請戶所從事的產業別。 三、高所得者未必是信用風險較低者，應考量其實質淨收入，故建議增加收支比變數。 四、研究結果，年齡及保證人之多寡、有無，對借戶信用風險的影響並不顯著。 五、已婚者較未婚者之信用風險為低，顯示家庭約朿力使借款人因責任感而降低違約的可能性。 綜上所述良好的消費性貸款審核系統可有效的提升銀行經營效率，降低授信風險，需善用電腦資訊系統，不斷蒐集彙總及分析資料，本研究可供未來銀行承作消費性貸款業務之參考。
關鍵字
消費性貸款並列摘要
In the wake of Asian financial crisis, economy slowed down, industries transferred their techniques and factories to other countries, and the business of bank credit and loan turned more conservative than before. The focus of loan shifted from business finance to consumer finance, now being the main source of profit in banks. However, with economic downturn, climbing unemployment rate and rising risk of consumer loan, how to evaluate the risk of consumer credit should be the most pressing issue for the banks wanting to develop consumer finance. In other words, banks not only increase the number of consumers applying for loans, but also reduce the risk of breaking contracts to its minimum. The main purpose of this thesis is to set up a system of examining consumer loan. This research takes samples from cases of consumer loans in X bank from January 2001 to December 2001. In order to form a model, there are 231 cases of normal loans and 231 cases of overdue loans respectively. Logistic-Regression is used as the research method of this research. Factors such as gender, working organization, seniority, salary, marriage, houses, deposit, credit rating, regions and approved interest rate are used to be predictive variables. The prediction accuracy on overdue loans is 77.9% and on normal loans is 79.2%. The research results are summaried as follows: 1.Gender is an important factor in assessing the chance of breaking contracts: this research and some similar studies all point out that women are less likely in breaking contracts than men are. Thus, this indicator can be included in scorecards of banks. 2.Rapid advances in technology lead to layoffs in some business sectors. It is not true any more that the longer the seniority is, the lower the chance of breaching contracts. In evaluating risk of credit, banks are suggested to consider what the vocations of loan applicants are. 3.The people with high income do not mean the people with low risk of credit. Banks should take substantial net gain into account, so the variable of earnings against expenditure ought to be added to the original scheme. 4.The age of the loan borrower, the number of guarantors, and the qualification of guarantors do not have significant influence on the risk of the borrower’s credit. 5.The married have lower credit risk than the single do. It shows that family bondage makes the borrower feel obligated and therefore reduces the possibility of breaking contracts. In conclusion, a good system of examining consumer loans can effectively enhance the operational efficiency and reduce risks of the bank. Computer information systems should be installed to gather, summarize, and analyze information. This research can be provided for further reference for all banks, which want to launch business of consumer loans