基于极值理论和Copula函数的条件VaR计算  

Calculate the condition VaR by extreme value theory and Copula function

作　　者：傅强[1] 邢琳琳[2] 

机构地区：[1]重庆大学经济与工商管理学院,重庆400030 [2]重庆大学数理学院,重庆400030

出　　处：《系统工程学报》2009年第5期531-537,共7页Journal of Systems Engineering

摘　　要：针对金融市场中某种资产不同风险的非线性和非对称尾部的特性,将极值理论和Copula函数应用于资产风险的研究以及条件VaR的估计.经过对深证成指的实证研究表明,极值理论能更好地拟合具有厚尾分布的收益率和日内波幅的边缘分布,Gumbel Copula函数也能更好地反映两者之间的相关关系.由Gumbel Copula函数拟合的联合分布计算出的在一定日内波幅条件下的市场风险VaR能给投资者在进行风险分析以及构建投资组合时提供有用的信息.Aiming at the non-linear and non-symmetrical tail characteristics of different risk of certain aasset, this paper applies the extreme value theory and Copula function to risk assets research and VaR calculation. The empirical result of Shenzhen Component Index shows that the extreme the- ory can better fit the edge distribution of the daily earnings rate and daily volatility which have thick tail. And Gumbel Copula function can also better reflect the relationship between them. The market risk VaR calculated by the joint distributions fitted with Gumbel Copula function can provide the investors in the risk analysis whih useful information.

关 键 词：条件VAR COPULA函数 收益率 日内波幅 POT模型 

分 类 号：F830.9[经济管理—金融学]