中小企業信用評等判別程序之構建-以台灣某金融機構為例

Abstract

近年來經濟不景氣與金融市場不穩定使得國內銀行或金融機構皆承受相當大的放款風險。為落實風險管理理念，於2004年6月所公告之新巴塞爾資本協定中，已明文允許銀行使用內部自行創立之評等方法，來有效管理借款企業之風險。目前中、外文獻雖已發展出一些信用評等模型，但大多是針對上市、上櫃公司，鮮少以中、小企業為研究對象，然而根據經濟部之統計，中小企業佔台灣企業數九成以上，借款銀行若將此類信用評等模型直接應用在台灣佔大多數的中小企業上，其評等結果可能會缺乏正確性及合理性。因此，本論文之主要目的則是針對國內之中小企業發展一套有效之信用評等流程，對中小企業進行合理的等級判別，使銀行或金融機構能夠快速地篩選低風險的借款申請者並做出適當的放款決策，進而降低放款風險。本論文所提之信用評等流程主要分三個階段：(1)變數的選擇與資料的收集；(2)利用資料包絡分析(Data Envelopment Analysis，DEA)將中小企業之財務能力及經營狀況作一個整合性的評估，建立信用評等指標(Credit Rating Score，CRS)，並根據 值之大小將客戶進行排序分等，建立等級劃分標準；(3)應用自組性演算法(Group Method of Data Handling，GMDH)建構評等績效值(CRS)預測模型。本論文最後以台灣某金融機構所提供之中小企業借款客戶的實際歷史資料，證實所建構之中小企業信用評等流程確實有效可行。

Many financial institutions have suffered serious loan risk due to economic recession and unstable financial markets. In order to reduce the credit risks due to incorrect loan decisions, Taiwanese governor has required banks and financial institutions to meet the requirements from the New Basel Capital Accord (Basel II). For this reason, developing a reliable credit rating model has become a crucial task for banks or financial institutions. Many studies on credit rating are based on the financial data drawn from the publicly traded companies. However, 90% of enterprises are small and medium-sized enterprises in Taiwan. It is not quite appropriate to apply the credit rating model for publicly traded companies directly to those banks or financial institutions whose customers are mainly small and medium-sized enterprises. Therefore, this study, focusing on small and medium enterprises, proposes an effective and reasonable credit rating procedure to assist banks and financial institutions sifting the lower risk applicants and making appropriate loan decisions quickly. The proposed procedure consists of three stages: (1) selecting variables and collecting data; (2) utilizing DEA to evaluate the credit rating score (CRS) for each loan business and establish a standard for grading the loan business using the CRS; (3) constructing a prediction model by using Group method of data handling (GMDH). Finally, a real case from a Taiwanese loan company is utilized to demonstrate the effectiveness of the proposed procedure.