影響消費性貸款逾期因素之實證分析

The Empirical Analysis of Affecting Overdue Factors of Consumer Loans

吳佩芬(Pei-Fen Wu 吳佩芬)

指導教授 ： 王親仁

國立臺中科技大學/商學院/企業管理系事業經營碩士班/碩士(2012年)

https://doi.org/10.6826/NUTC.2012.00002

摘要

銀行業辦理個人消費性貸款案件有核貸速度快、授信期間較長、違約發生損害較小及案件數量多等特點，如何快速篩選出客戶狀況，並且幫助授信人員作出正確決策，以減少貸後產生逾期放款案件，乃當今銀行業辦理消費性貸款案件重要考量因素。本研究評估這些因素，來辨識其隱含的風險，並掌握其屬性及給予可量化的指標，來做為授信策略擬定之依據，希望擴展個人授信業務時，能降低貸款違約的風險與銀行的逾放比率，以增加銀行授信的品質、速度及效率，進而增加銀行實質上的收益及整體上的經營績效。
本研究以2006年~2011年個案銀行於新竹地區所承做之個人消費性貸款授信案件為對象，隨機方式選取355件貸款案件(逾期戶82戶及正常放款戶261戶)，利用羅吉斯迴歸分析發現借款戶之年齡、教育程度、工作年資、平均年所得、負債比、現有信用卡額度及是否使用信用卡循環額度等因素，對於消費性貸款授信案件發生逾期有顯著的影響。本實證模型平均預測之準確率達93.8%，而預測準確率會隨樣本中正常戶與逾期戶之比率不同而異。

關鍵字

消費性貸款 ； 卡方檢定 ； 羅吉斯迴歸模型

並列摘要

Consumer loan have some characteristics, such as fast speed of authorization, longer loan period, less bank loss when contract is broken, and large quantity of cases. Because of these characteristics, there are some important evaluating factors are proved to rapidly screen customer's situation and help credit officers make correct decision. This study will assess these factors to distinguish its implicit risk and know its attribute for giving the quantized indicator and the basis of making decision. In the mean time of expanding individual credit granting can reduce loan risk, non-performing loan ratio, for increasing the quality, speed, efficiency of bank credit granting then improve bank’s physical income and whole operational performance. Therefore, finding out the correct evaluation of risk factor will provide bank officers certain helps for making effective consumer loans decision.
This study were randomly chosen 355 individuals of consumer loans, 271 of normal and 84 of abnormal, at Hsinchu district during 2006~2011. The empirical result of linear Logistics Regression found seven significant factors, such as age, education level, working age, annual income, liability ratio, credit card limits and using revolving credit lines have apparently influence on overdue probability of credit granting for individual customers. The cases of customer loans would be refused if two of the 19 parameter factors, bills of credit and lenders interest payment, are found abnormal. The prediction accuracy rate of empirical model averagely is 93.8%, and will be raised as reducing the number of overdue household sample.

並列關鍵字

Consumer Loan ； Chi-Square Testing ； Logistic Regression Model

參考文獻

17.陳麗珍（2011）。不動產抵押貸款授信品質之研究－以xx銀行為例（碩士論文）。朝陽科技大學。

21.顏柏仁（2011）。銀行對中小企業融資決策之實證分析（碩士論文）。國立台中技術學院。

11.康建民（2008）。個人貸款信用評分模型－以儲蓄互助社為例（碩士論文）。國立暨南大學。

13.張素燕（2009）。消費者信用貸款授信評量模式之研究（碩士論文）。國立中興大學。

1.Cairney, J. & Boyle, M.H. (2004). Home Ownership, Mortagages and Psychological Distress; Housing Studies, 19, 161-174