- 學位論文
銀行授信實證分析-以A銀行中小企業授信為例
Empirical Analysis of Bank Credit-- for Example The A Bank Credit for SMEs
摘要
中小企業是台灣的經濟命脈，創業艱難，往往得依靠政府政策大力扶持，銀行是否對企業放款關鍵都是藉由財務報表及擔保品的價值來評估。本文的動機是在找到一個可以事先預警的模型，能幫助銀行授信人員作好最佳的授信評估。期望作為中小企業授信決策時重要參考，以降低中小企業因資訊不對稱所帶來的違約風險，建立企業財務危機預警模型，期能發揮真正防患未然的事前預警功能，藉以提供金融機構與在授信貸放與否及額度控管之依據。 經由二元Logistic迴歸模式的預測分類結果，最適的分割值為0.6，整體之準確性有83.2%。而Hosmer-Lemeshow檢定值為8.750 (p為0.364>0.05)未達顯著水準，表示前一年款貨資產比、前二年款貨資產比、前二年流動比率、授信比等四個變數所建立的中小企業信用保證之授信模式適配度良好。因此，銀行授信人員可利用此模式在最短時間內，依申貸企業之資料來預測客戶該筆放款在未來倒閉或正常的機率為何，銀行是否要承作該筆放款，有一預測的準確性。
並列摘要
Small and medium enterprises (SMEs) which are the heart of Taiwan economic were founded very difficult and always need support by government to try to carry out policies. The crucial that banks decided to lend is often based on financial statements and the value of guarantees. To find an early warning model which can assist banks to make the best rule of credit guaranty is the motivation of this study. This study tries to establish a financial cris alert model which provides credit loan and amount of control for financial institution and the SME take precautions against a calamity. And the model also could be an important foundation of the credit loan decision of the SME to decline default risk of SMEs from asymmetric information. The result of Logistic model performance is 0.6 in prediction of bankruptcy, and the aggregate accurany is as large as 83.2%. The coefficient of Hosmer-Lemeshow of 8.750 was not found to be statistically significant at 0.05 (p-value is 0.364), that is, SMEs credit model which was created by these four independent variables had the fitness which is very desirable. Banks can decide credit loan or not and reduce the probability of default risk in the shortest possible time by using this model.