放款風險評估及信用評等整合流程之構建

Abstract

風險評估及授信評等結果是放款機構或銀行用以評量借款者償債能力的重要依據，然而在現今經濟不景氣的大環境下，節節升高的逾放比率使得越來越多的銀行或放款機構必須檢討其現有評等模式的缺失，以作出更正確有效的放款決策。大部份的評等制度設計不外以借款者的品格、能力、資本、業務狀況及擔保品五個面向加以討論，但要從多個影響企業信用因素中正確的對企業做出信用評等亦非易事。現有中、外文獻雖發展許多信用評等模式來探討此問題，但多以上市上櫃公司為研究對象，這與台灣一般放款公司實際接觸的借款者（中小企業）有所出入，因此文獻上所建議模式其應用價值有限。本研究針對放款公司之實際借款歷史資料，將風險評估納入考量，發展出一個整合風險評估模式及信用評等模式之放款評等流程，此流程主要分三階段：（1）變數之選擇與資料的收集、（2）分別利用判別分析與羅吉斯迴歸構建放款風險評估模式，研究結果發現兩者分類效果相近，考慮應用之便利性後，選取判別分析方法導入模式之中、（3）正常與違約資料存活期預估模式。其中，風險評估模式是用來判斷借款公司是屬正常或違約企業，正常與違約企業的存活期預估模式則可提供放款公司決策者關於借款者可能還款期長短的資訊。放款公司利用本研究發展之放款流程即可快速正確區分出正常與違約資料，能有效幫助放款公司降低放款之風險。

The credit ratio is one of the important indicators for banks to evaluate the payment capability of the loan applicants. However, the raising ratio of the bad loans drives banks to review and reconstruct their credit scoring models. Most of the models are built in five dimensions: applicant’s personality, payment capacity, capital, business condition, and collaterals. In addition to these five factors, many uncertain factors can be utilized the loan applicants. Many literatures proposed credit scoring models for the listed companies, but these credit models would be invalid when they are employed with the real applicants’ data. This study analyze the real data and construct a credit rating procedure. The proposed procedure contains three parts:1. variables selecting and data analysis. 2. constructing my risk analysis models using discriminant analysis method and logistic regression method to discriminate the good and bad loaners. 3. Constructing the survival model for good and bad loaners. Through the survival model, banks can predict the applicants’ survival period. A real case is also provided in this study to demonstrate the effectiveness of the proposed procedure.