基于多维Gumbel Copula函数的投资组合VaR分析

Portfolio VaR Analyses Based on the Multi-dimensional Gumbel Copula Function

基于目前国内有关Copula函数的实证研究主要是研究二种资产的相关性为主,文章根据Copula函数在构建反映随机变量实际分布与相关性的联合分布函数上具有的优势,首先利用GJR模型构建资产的边缘分布,接着利用多元阿基米德Copula函数族中的Gumbel Copula函数构建了反映多个资产收益实际分布和相关性的联合分布函数,并使用蒙特卡罗模拟技术,分析在不同置信度下的投资组合的最小风险价值(VaR)及其资产组成,实证说明根据文章提出的模型度量资产的风险,可以使投资者选择的资产更加稳健,同时也有利于投资者对投资组合整体风险进行分散和监管。

Based on the fact that the current domestic empirical research on Copula is mainly to examine two kinds of assets, this paper first constructs assests＇ marginal distribution with the help of the GJR model, then it constructs a joint distribution with Gumbel Copula Function according the advantage of the Copula function in reflecting the actual distribution of random variables associated. With the use of Monte Carlo simulation technology, the minimum VaR of the different confidence level and the composition of assets related are also analysed.From the evidence, investors cannot only choose more stable assets, but also is conducive to the risks of dispersion and supervision. This paper presents evidence that the use of the model of this paper will lead investors choose more stable assets, which will also be conducive to the overall portfolio risk dispersion and sunervision.