本研究利用國內某一家外商銀行於2010年間所承作之1,097筆中小企業授信案件,依銀行內部制定之客戶信用篩選條件及額度風險,進行分析影響銀行授信品質的因素。本研究係以銀行授信人員觀點,分析中小企業主向銀行申請融資時,順利取得融資關鍵要素,希冀能提供銀行業務人員追求業績成長同時找出影響中小企業貸款風險之因素,以建立一套專屬銀行風險控管事前預警模式,以降低未來銀行發生逾期放款之機率。 透過卡方差異性分析、自動互動檢視法及二元羅吉斯迴歸分析模式等實證分析,結果發現影響中小企業取得授信融資之變數包括企業成立年資、供應商照會結果、新業務查詢家數、額度動用率、利息保障倍數、應收款項週轉天數、存貨週轉天數等。由根據建構之混淆矩陣預測模型,對整體樣本、核准戶及拒絕戶之判定正確率分別各為88.6%、99.4%及17.4%。
This study is an empirical analysis of bank credit quality cases, which are developed internally by the customer selection criteria and credit acceptance criteria. This study is based on a foreign bank in Taiwan, and total credit bearing case is 1,097 in 2010. This study investigate from the viewpoint of bank crediting officers, and the purpose is aim for small and mediume enterprises(SMEs) owners to apply for credit. This study will be able to help bank’s staffs to find out the key factors of risk management on SMEs. Moreover, it establishes a early warning models for reducing the probability of the occurrence of overdue bank loans in future. The empirical analysis is based on Chi-Square difference analysis, Chi-Square Automatic Interaction Detection (CHAID) and binary logistic regression analysis. We found the credit indicator variables for influencing the SMEs gaining loan from bank including are “Years in Business”, “Supplier Reference Checking”, “The number of new business inquiries”, “Utilization of all bank’s loan Facilities”, “Interest Coverage Ratio”, “Day of Accounts receivable turnover”, “Day of Inventory turnover”. The correct rate of overall sample, approved cases and rejected cases is 88.6%, 99.4% and 17.4% by means of Confusion Matrix of Prediction Model(CMPM) which is constructed by enterprise Class Index.