Copula-EVT模型及其投资组合风险分析中应用  

A Copula-EVT Model for Measure a Portfolios Risk

作　　者：余平[1] 史建红[1] 

机构地区：[1]山西师范大学数学与计算机科学学院,山西临汾041004

出　　处：《山西师范大学学报（自然科学版）》2010年第4期41-45,共5页Journal of Shanxi Normal University(Natural Science Edition)

基　　金：山西省青年科技研究基金项目(2008021007);山西师范大学自然科学基金项目(YZ06001)

摘　　要：Copula技术广泛地应用于金融领域,特别是在金融风险、投资组合、资产定价的方面,目前已成为解决金融问题的一个有力工具.本文将Copula技术应用于沪深股市投资组合当中,由于VaR和ES表达式难以求出,于是采用蒙特卡罗模拟的方法模拟组合收益走势,进而计算出在不同置信水平下的风险价值(VaR)和期望不足(ES),其中对数收益边缘分布函数由中心和左尾为Laplace分布,右尾为极值分布组成.从"Basle交通灯法"返回检验的结果来看,Copula-EVT模型能够较好度量组合风险.Copula technology is a powerful implement to solve finance problem,and it is widely used in finance filed especially in finance risk,portfolio,asset price.In this paper,we apply Copula in Portfolio between Shang Hai Stock market Shen Zhen Stock market,because the formula of Value at Risk and Expect short are quite complex,so we measure Value at Risk and Expect short with Monte Carlo method at different confidence levels.Log-return marginal distributions are modelled through Laplace distribution in the centre and in the left tail,while the right tail is built using the principles of Extreme Value Theory.It turns out that Copula-EVT method owns a good result to measure portfolio risk judge by back testing.It turns out that Copula method owns a good result to measure risk judge by Basle traffic light back testing.

关 键 词：COPULA函数 极值理论 拉普拉斯分布 投资组合 

分 类 号：F830[经济管理—金融学]