信用風險模型的修正與實證分析

An Empirical Study and Modification based on Credit Risk Model

李世鴻(Shih-Hung Li)

指導教授 ： 巫春洲 ； 江長周

中原大學/商學院/國際貿易研究所/碩士(2006年)

https://doi.org/10.6840/cycu200600554

摘要

摘要
違約風險可界定為企業償還負債能力上的不確定狀況，在違約發生之前，往往只能評估公司發生違約的可能性。違約是一種極少發生的事件，但當違約發生時，貸出者或交易對手遭受的損失往往非常的嚴重。近年來因巴塞爾協議的修訂，銀行更致力於建構適當之信用風險模型。而量化模型的發展十分迅速，就信用風險控管上，將有助於增加預測違約風險之準確性。
本研究嘗試在KMV模型的架構之下，針對台灣上市上櫃公司的違約風險進行測試，並且加入Duan（1994,2000）的最大概似估計法估計參數，以作業特性曲線(Receiver Operation Character；ROC)及交叉分類表做為模型的驗證方法，比較Duan模型是否能比KMV模型提高對台灣上市公司違約機率方面的預測能力。樣本期間為1998年至2005年，取違約公司66家與對照樣本正常公司132家。
本研究實證結果發現在作業特性曲線的檢定上，KMV模型的AUC(area under curve)數值是0.7942，Duan模型的AUC數值為0.8194。因此無論是藉由ROC曲線的表現上或是AUC數值大小的指標上，證明Duan模型優於KMV模型。在交叉分類表的檢定上， Duan模型的準確率為84.61%比KMV模型的78.92%為高，也說明了Duan模型的準確率高於KMV模型，因此可以說Duan模型在配適台灣上市和上櫃公司資料時有較好的區別能力，在違約機率的估計上比KMV模型佳。

關鍵字

作業特性曲線 ； 信用風險 ； 最大概似估計法 ； 預期違約機率

並列摘要

Abstract
Default risk is the uncertainty a firm’s ability to service its debts. Prior to default, we only can make probabilistic assessments of the likelihood of default. Default is a rare event. When a firm will default, the loss suffered by a lender or counterparty in the event of default is usually significant and is determined largely. Because of Basel Accord Ⅱ, financial institutions have developed their own credit models to assess various measures for the credit risk. In credit risk management, quantification credit risk models developed rapidly to be contributive to forecast credit risk accurately.
The study test to quantify the default probability of Taiwan’s companies by the KMV model, and try to improve on estimating parameters by the Duan(1994,2000) model. We compare Duan model with KMV model in which default probability of Taiwan’s database by receiver operation character (ROC) and typeⅠerror . We used 66 default companies and 132 undefault companies during 1998-2005 period.
In the empirical result, AUC(area under curve) of KMV model is 0.7942, AUC of Duan model is 0.8194. No matter what in ROC curve or AUC value, Duan model is better than KMV model. In typeⅠerror test, Accuracy of Duan model is 84.61% and Accuracy of KMV model is 78.92%. Duan model is higher than KMV model. So it is better estimated by Duan model than KMV model under Taiwan’s companies data.

並列關鍵字

credit risk ； expected default frequencies ； maximum liklihood estimation ； receiver operation character

參考文獻

Altman, E. I., 1968, “Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy”, Journal of Finance, September, 589-609.

Altman, E. I., G. Marco, and F. Varetto, 1994, “Corporate Distress Diagnosis:Comparisons Using Linear Discriminant Analysis and Neural Networks (the Italian Experience)”, Journal of Banking and Finance, May, 505-529.

Black, F., and J. C. Cox, 1976, “Valuing Corporate Securities:Some Effects of Bond Indenture Provisions”, The Journal of Finance, 31, 351-367.

Black, F., and M. Scholes, 1973, “The Pricing of Options and Corporate Liabilities”, Journal of Poliical Economy ,81, 637-654.

Duan, J. C., 1994, “Maximum Likelihood Estimation Using Price Data of the Derivative Contract”, Mathematical Finance, 4, 155-167.