風險性債券組合之多期風險值評估-- 內部價值法以及傅立葉轉換法

Abstract

在進行以公司債為主的風險性債券投資組合分析時，由於所包含之債券具有不同的現金流入點、到期日、評等與產業屬性等，因此也較單一債券複雜許多。本文目的在於：結合內部價值法以及傅立葉轉換法來預測風險性債券投資組合的多期風險值 (VaR)。以個別公司的自由現金流量模型為信用風險評等基礎，並結合傅立葉轉換法與因子模型來處理不同資產間之違約相關問題，據以求得債券投資組合在未來各時間點上的預期回收率。 將投資組合各時間點的現金流量乘以預期回收率，即可將風險性現金流量轉成無風險性現金流量，並據以求得債券投資組合價值。而後，本文採用RiskMetrics的方法求得投資組合之每日之風險值，並與實際市場價格進行實證比較。結果顯示；本研究方法所求得之多期風險值與實際市場價格變動極為接近，因此可為風險性債券投資組合之良好分析工具。

Analyzing a risky bond portfolio formed mainly of corporate bonds with different cash flows, maturities, ratings and industries is much more complicated than analyzing a single bond. The purpose of this study (by combining the intrinsic valuation and the Fourier Transform approaches) is to develop an integrated method for endogenously estimating a credit portfolio’s multi-period expected recovery rates. These are useful in estimating the daily value at risk (VaR) of risky bond portfolios. To be more specific, we adopt a cash-flow based structure form credit model and employ a Fourier Transform method with a factor model to handle the default correlation issues. From this, we get the portfolio’s expected recovery rate for each time vertex. By applying the portfolio’s expected recovery rate to the discount portfolio’s cash flows, the risky bond portfolio can be transferred into one that is risk-free. Consequently, the daily VaR can be calculated according to the RiskMetrics method. This study presents an empirical comparison of VaR and real portfolio daily returns. The results show that the estimation of the VaR is considerable and that the integrated approach can be a useful tool when dealing with risky bond portfolios.