企業經營過程中，透過銀行取得融資為企業發展過程中重要的資金來源，本研究主要調查國內某全國性商業銀行中部區域所屬的7家分行企業申請案件，共計498筆徵、授信及評分等內部審核准駁資料加以估計，首先使用Multinominal logit及Binary model加以評析銀行對企業授信限額的決定，研究結果發現企業已往來銀行家數、是否為中小企業、是否為新往來公司、本次申請授信金額之變化、銀行對企業申貸戶之評分、是否屬分行經理以上階層所核准等變數為申貸企業於銀行授信審核過程中是否被限額決定的主要因素；其次，在其他狀況不變之下，分析重要二元變數之邊際效果；最後，經由樣本選擇，使用Heckman二階段估計方法，對完全未受銀行限額之申貸戶，分析銀行所授予最高放款金額及其相關變數間的關係，其結果顯示目前行庫已貸金額、所提供擔保品價值、銀行對企業申貸戶之評分與所核准授予之放款金額間有非常顯著之正向影響關係，其次，在往來銀行家數、是否為中小企業、是否為新往來公司、是否屬分行經理以上階層所核准與銀行核准授予之放款金額間有顯著之負向關係，上述不同模型中影響決策變數雖然在顯著水準上有些許之差異，但是其意義上相同，結果大多與目前銀行操作之現狀相當吻合且符合預期，並且能夠提供企業及銀行作為授信決定之參考。 It is an important job for enterprise to acquire credit through the borrowing process with bank. This paper using a sample of 498 enterprises in the central aera of 7 different branches of a well-known commercial bank in Taiwan to investigate the determinants of credit rationing in the borrowing processes. Firstly we compared the method of Multinominal logit model with the Binary(Logit and Probit) model. The main results are as follows：variables reflect that the number of banks the enterprise had loan application，if it is a median or micro scale enterprise，if the enterprise is a new applicant，the amount variation of this loan application，the current credit rating，if this application needs to be approved by the upper management are significantly affect the determinants of credit rationing in the borrowing process. Secondly, this study point out the marginal effect of the important binary variables. Finally after sample selection process, through heckit method for enterprises that havn’t been rationed, this study analysis the factors that the maximum amount the bank is willing to grant given the state of knowledge about the applicant and according to the internal criteria of the bank . The main results are as follows：the bank is willing to supply larger loan if the enterprise had been granted more loan amount by other banks，the enterprise can provide more value of collateral and assets ，credit rating of this enterprise with a high score , and these coefficients are positive and significant. Secondly, variables shows that the number of banks the enterprise had applied，if it is a median or micro scale enterprise，if the enterprise is a new applicant，if this application needs to be approved by the upper management are negative significantly with the maximum amount that the lender is willing to grant. This paper has built different models , although there is a little difference in the significant level of credit rationing determinants. These interesting results we obtained are sure to answer the practical operation of the bank and can indeed provide a good explanatory power in the credit determinants for enterprise and bank.
