摘要
近年來信用衍生性金融商品發展非常迅速，市場上逐一推出不同特色的信用擔保債權憑證。其中，具有時間性相依的信用擔保債權憑證漸漸受到大家青睞，商品內容例如:遠期生效信用擔保債權憑證、信用擔保債權憑證選擇權...等。然而在此篇文章中，我們所研究對象為遠期生效信用擔保債權憑證，其商品的特色在於生效日之前，標的資產若違約，並不構成損失的發生，只會將此商品從投資標的中剔除，故對投資人而言在生效日之前，受到一層信用保護。 在過去市場上，我們廣泛地利用因子相關性模型來評價標準類型的擔保債權憑證，由於它在計算上所帶來便利性廣泛受到市場接受，然而，在因子相關性模型中，當評價具有時間性相依的遠期生效信用擔保債權，卻無法有效衡量不同期間結構下的損失分配。故為了不放棄因子相關性模型所帶來的便利性，我們試著在因子相關性模型中引入跨期相關性因子，並且考量不同分配因子相關性模型，讓評價過程中得以順利進行。目前市場上受到金融危機的影響，在過去評價過程中，假設信用擔保債權憑證與信用指數回復率為40%顯得已不適用。 為了有效解決這個問題，我們引用 Amraoui & Hitier(2008) 所提出的隨機回復率方法，建構在信用擔保債權憑證與遠期生效信用擔保債權憑證評價過程中，相同地，我們也在隨機回復率方法中引入跨期相關性因子，並且考量不同分配因子相關性模型，加以探討其中意涵的經濟意義。
並列摘要
Now the market valuing CDOs with factor copula model is well established. Because the factor copula models remain popular in the pricing of credit portfolio derivatives and their computational efficiency. Recently, however much new credit basket securities with strong timing features have emerged. The prime example is probably the forward starting CDO. Therefore this is generally not true for pricingnon-standard CDO tranches with factor copula model. In this paper, we improve the elaboration on the structure of factor copula model and to consider different factor copula model distribution term structures and to add intertemporal loss correlation in. As part of our analysis, we have faced with the spread widening of the CDO tranches in the main credit indices and recovery assumption of 40%. It becomes impossible to calibrate the factor copula model to certain spread levels during this credit crash period. In order to resolve this problem we employed Amraoui & Hitier(2008) new methodology for CDO and FCDO pricing with stochastic recovery rates. We also numerically test the dependence of forward starting CDO on the correlation of losses across time.