Summary: | 碩士 === 國立政治大學 === 統計研究所 === 103 === The most widely used methods used application of Large Homogeneous Portfolio (LHP) assumption of the one factor copula model for pricing synthetic CDOs. The one factor Gaussian copula model was first used by O'Kane and Schloegl (2001) proposed, however, only in equity tranches get a good evaluation of the results of the fit for each tranches. Kalemanova et al (2007) proposed the application of LHP assumption of one factor NIG copula model. The one factor copula model of NIG distribution evaluation results are far better than normal distribution, but overestimated above mezzanine tranches. The above models are all pricing of Synthetic CDOs before 2008, and select only certain days for analysis. Therefore, this paper in March 2008 to March 2013 to do a complete long-term analysis, comparison of different models for pricing of Synthetic CDOs results. In this paper, one factor Gaussian copula model, one factor NIG copula model and one factor dynamic model do discussion. In Gaussian and NIG one factor copula model, the empirical results of the final analysis, diminishing over time periods n, then will be able to significantly improve the results of Synthetic CDOs pricing. In the part of the one factor dynamic model, since the parameter estimation method is not perfect, so the results of Synthetic CDOs pricing are not in line with expectations.
|