中文摘要：

将广受欢迎的，用于CDO定价的大样本同质投资组合近似方法做了推广，其中涉及到的分布是高斯分布和Variance Gamma分布的混合，即G—VG混合分布．提出了厚尾的G—vG混合Copula模型型和带有随机相关性的混合模型．这些模型可以有效的模拟CDO定价中的“相关性微笑”问题．在这些G—VG混合Copula模型中，得到了损失分布函数和期望分券层损失的解析表达式．并且用实际数据做了实证分析，把新模型和高斯模型的结果做了比较．实证表明，新模型的结果不仅与市场报价更贴近，而且为相关性结构带来了更多的灵活性．

英文摘要：

An extension of the popular large homogeneous portfolio （LHP） approach to the pricing of CDOs is presented, where the involved distributions are mixtures of Gaussian distribution and VG distribution. The fat-tailed G-VG copula model and model with stochastic correlation framework which can effectively model ＂correlation smile＂ in CDO pricing are considered. Loss distribution and expected tranche loss are explicitly derived in the G-VG copula models. The numerical analysis is carried out and the properties of new models with those of the Gaussian models are compared. New models not only provide a closer fit to the market quotes, but also bring more flexibility into the dependencestructure.