動態違約門檻下之債券組合信用風險模型：現金流量基礎法

Abstract

探討資產組合信用風險的文獻中，多半運用縮減式模型(reduced form model)或多變量極值理論進行研究，卻鮮有研究運用結構式模型(structural model)進行分析。本研究針對公司債資產組合，建立結構式多期信用風險模型，考量違約相關與動態違約門檻的效應。模型以現金流量基礎法衡量公司資產價值，先運用狀態變化關聯結構(factor copula)建立公司資產價值與景氣因子的連動性。為了縮小評價結果與市場價格的差異，本研究進一步引入不完全資訊下，投資人對違約門檻的預期，以解決Merton類型結構式模型低估接近到期日債券及低違約機率債券之信用風險的問題。本文所提出的模型，相對於既有文獻的最大貢獻，為建立動態風險結構與內生化的回收率(recovery rate)。本模型可以廣泛運用於結構型信用商品的評價與信用風險的衡量。

Most existing studies on portfolio credit analysis adopt reduced form approach or multivariate extreme value theories. Alternatively, this paper proposes an integrated approach to incorporate default correlation and default threshold dynamics into a multi-period structural model. A cash flow based credit model that has a factor structure and the factor copula, a conditional independent framework, are combined to analyze the credit risk of a corporate credit portfolio. To address the issue that Merton type structural models underestimate both short-term credit spread and safe bond credit spread, we suggest employing a dynamic default threshold to account for investors’ imperfect information on the threshold. The proposed approach differs from most existing literatures in that it models the risk structure dynamics and the endogenous portfolio recovery (loss) rate. The model can be applied to valuation of a wide range of structured credit products and credit risk management.