风险依赖、一致性风险度量与投资组合--基于Mean-Copula-CVaR的投资组合研究  

Risk Dependence,Consistency Risk Measurement and Portfolio:Based on Mean-Copula-CVaR Model

作　　者：张冀[1] 谢远涛[1] 杨娟[2] 

机构地区：[1]对外经济贸易大学保险学院,北京100029 [2]中国科学技术发展战略研究院,北京100038

出　　处：《金融研究》2016年第10期159-173,共15页Journal of Financial Research

基　　金：国家自然科学基金项目(71201029);教育部社科一般项目(15YJA790082);北京市社科一般项目(15JGB090);对外经济贸易大学中央高校基本科研业务费专项资金(CXTD5-04);对外经济贸易大学中国企业“走出去”协同创新中心科研项目(201504YY005B)等资助

摘　　要：本文把风险依赖、一致性风险度量与投资组合纳入到一个分析框架中,结合Coupla-CVaR模型和Mean-var投资组合理论构建Mean-Copula-CVaR的投资组合模型,能有效同时解决风险度量中的一致性和依赖性关系。采用券商指数、银行指数和保险指数实证分析线性依赖和复杂依赖(Copula依赖)情况下金融机构资产配置的差异性和风险度量的充分性,研究结果表明,纳入Copula函数能够更为稳健和准确地预测投资组合的CVaR。然而,本文没有检验出不同形式Copula之间的差异具有显著性。本文的政策含义在于,忽视复杂风险依赖结构可能会造成风险低估,从而影响资产配置的有效性。In this paper,we put the risk dependence,consistency risk measurement and portfolio into an analytical framework,combined with Coupla- CVaR model and Mean- var portfolio theory to construct the investment portfolio model of Mean-Copula-CVaR,and effectively solve the consistency risk measurement and dependence. Using the Securities index,bank index and insurance index,we empirically analyze the differences of asset portfolio and the adequacy of risk measure among financial institutions on basis of linear dependence and complex dependence（ Copula dependence）. The results show that putting Copula model can more robustly and accurately predict the CVaR portfolio. However,there is no significant difference between different forms of Copula in this paper. The policy implication of this paper is that it is possible to underestimate the risk if we ignore the structure of complex risk dependence, thus affecting the effectiveness of asset allocation.

关 键 词：风险依赖 风险度量 Mean-Copula-CVaR 投资组合 

分 类 号：F224[经济管理—国民经济] F842F832