论文部分内容阅读
目前,高斯单因子定价模型已成为CDO定价的标准化模型并被广泛采用,但该模型不能准确刻画金融市场数据的“厚尾性”并存在相关性“微笑”现象,模型计量的结果不能很好的拟合市场报价。鉴于此,本文构建了基于正态逆高斯分布的因子模型对CDO估值,并对iTraxx Europe指数分券进行实证研究,通过模型理论估值和市场报价之间的差异来度量模型的定价能力,研究结果表明使用NIG-Copula分布有效改善了定价模型,能够更好地拟合了金融市场数据的“厚尾”现象,且在一定程度上解决了相关性“微笑”现象,能更好地拟合市场报价数据。
At present, the Gaussian one-factor pricing model has become a standardized model of CDO pricing and widely used, but the model can not accurately characterize the “thick tail” of financial market data and has the correlation “smile” phenomenon. The model measures The result can not fit the market quoted very well. In view of this, this paper builds a CDO valuation model based on the factorial model of normal inverse Gaussian distribution, and conducts an empirical study of iTraxx Europe index coupons, measures the model’s pricing power through the difference between the model theory valuation and the market quotation, The results show that using the NIG-Copula distribution can effectively improve the pricing model, better fit the “thick tail” phenomenon in financial market data and to a certain extent solve the “smile” phenomenon, Better fit the market quote data.