本文從實務面定義客戶違約原因和客戶違約後果兩項變數，以往文獻很少考慮到。本研究目的主要在探討銀行授信決策與客戶違約之關聯，提銀行有效的識別風險，減少風險損失的參考點。 本文透過對具有實務經驗的銀行授信員工進行問卷調查，其內容包含銀行授信決策、客戶違約原因和逾期放款。回收有效問卷550 份。我們在調查問卷設置的變數都是Likert 變數。首先透過變異數分析，我們發現，逾期放款金額為“1~5仟萬元”與 “控制風險僅僅依賴擔保品” 有顯著差異，因此，授信人員必須更嚴格審核擔保品以降低授信風險。在逾放金額相對客戶違約原因之變異數檢定結果中，也發現類似的現象。再者，從因素分析發現，受訪者認為“客戶短期資金緊俏週轉不靈”、“客戶因重大訴訟案件引發財務危機”、“客戶多元化經營產生過度投資”、“客戶因財務危機謠言引發銀行緊縮銀根”、“客戶偏好風險性理財或投資活動”、“客戶高財務槓桿操作利息負擔過重＂等因素，相對來說是客戶違約的主要原因。最後，從迴歸分析發現，逾放比與“客戶所提供徵信資料的正確性＂和“授信人員判讀正確性＂成負相關。逾放金額與“授信人員判讀正確性＂成負相關。 基於本文之實證分析，我們提出改進客戶識別與授信評價制度的建議，希望這個實證結果能給銀行界之授信決策帶來一些參考指標。 To become a full-fledged commercial bank, the key factor is whether this bank can overcome the risk test to maintain sustainable development capacity. The sustainable development capacity is closely related to acquiring experienced, scientific, and effective credit decision-making. The credit decision-making capacity includes the growing ability to manage risks and the capacity to identify risk so as to reduce losses. In this thesis, a questionnaire survey with 550 effective questionnaires recovered was carried out. The questions included bank credit decision, custom default causes and overdue loans and the set-up variables were Likert variables. The survey objects were bank credit staffs with practical experience. Through the ANOVA testing of overdue loans vs. credit decision-making banks, we found significant difference (pvalue < 1%) between the amount of overdue loans of ＂10 ~ 50 million＂ and ＂the risk control solely rely on collateral＂. In the ANOVA test results of the overdue amount vs. the custom default causes, similar phenomenon was also found. From the factor analysis, we surmised that responders believe that ＂short-term cash flow problems＂, ＂financial crisis triggered by major legal actions＂, ＂excessive investment caused by diversification＂, ＂banks tightening caused by the rumors of finacial crisis＂, ＂customer preferences of risky investment activities＂, and ＂high financial leverage with overloaded interest＂ are the main reasons of customer default. From the regression analysis, the overdue loan ratios present negative correlation with ＂the accuracy of the credit information provided by customer＂ and ＂the accuracy of the data interpretation by credit staff＂. The amount of overdue loan also showed negative relationship with ＂the accuracy of the data interpretation by credit staff＂. On the basis of date obtained by this empirical analysis, we provide the reference point for improving the credit evaluation system.
