信用與流動性風險研究

Abstract

本篇論文包含兩大主題。第一部分使用非線性濾波器(Nonlinear filter)去同時捕捉違約風險與流動性風險對於信用衍生性商品的影響。第二部份利用政府公債的流動性價差去估算公司債的流動性價差。在第一部分我們提出三因子訂價模型，並且研究流動性風險是否會影響信用衍生商品的定價。實證結果也證實CDS的訂價模型在考慮流動性因子後，可以明顯的改進訂價誤差。此外，為了討論選擇權市場的資訊是否會影響信用衍生性商品的訂價，我們建構了信用衍生性商品價格與選擇權變數的進一步分析。而實證結果證明信用衍生性商品市場與選擇權市場會彼此相互影響，並且驗證了選擇權變數對於解釋信用衍生性商品價格是有幫助的。第二部份為了研究流動性對公司債的相對影響，本篇研究利用政府公債的流動性價差來決定公司債的流動性價差。為了捕捉動態的違約強度與流動性強度的動態過程、克服參數錯誤，以及簡化模型的設定誤差，我們使用半母數(Semi-parametric)的技巧來估計動態的強度過程。實證結果也使我們相信導入流動性因子將有助於我們解釋信用價差低估的現象，而此半母數的模型不但能適當的捕捉違約強度與流動性強度的時間變異，而且也避免模型的設定錯誤。因此本研究的兩篇文章明確的支持雖然違約風險是影響債劵與信用衍生性商品價格的主因，但是流動性風險對於債劵與信用衍生性商品的衡量也有重要的影響。

This thesis has two main articles. The first article uses a nonlinear filter as a way to simultaneously capture default and liquidity effects in the credit derivatives market. The second article uses liquidity spreads in the Treasury bond market to obtain a direct measure of the liquidity factor for corporate bond market, and estimate the liquidity component for corporate bond markets. In first article of this thesis, we propose a three-factor pricing model and investigate whether liquidity is an additional risk factor in determining credit derivatives pricing. Empirical results demonstrated that the introduction of a liquidity risk factor was important to mitigate pricing errors in credit derivatives prices. To discuss whether the information about the options market impacts the pricing of credit derivatives, we conducted further analysis of credit derivatives prices and options variables. Empirical results demonstrated that options variables could be useful in explaining credit derivatives prices. The second article uses liquidity spreads in the Treasury bond market to obtain a direct measure of the liquidity factor for corporate bond market, and estimate the liquidity component for corporate bond markets. To capture time variation on instantaneous spreads and volatility, and reduce modeling bias, semi-parametric techniques are applied to estimate the time varying intensity process. Empirical results demonstrate that the liquidity factor is an important determinant to credit spreads, and the semi-parametric model can simultaneously capture the time effect and reduce modeling bias. The findings of these two articles lead us to believe that the default risk factor can have an enormous effect on credit derivatives and bond prices, but a liquidity risk factor also plays an important role in explaining credit derivatives and bond prices.