摘要
在计算投资组合的风险价值(Value-at-Risk)时,传统的方法是假设多个金融资产的收益率序列服从多元联合正态分布,但这种假设经常与实际市场的观测不符,并且在极端事件发生时常会低估风险。为了更好地刻画投资组合的风险,本文提出利用Copula函数结合非对称Laplace分布技术来解决这个问题:Copula函数能够很好地刻画多个金融资产间的相依结构关系,另一方面利用非对称Laplace分布来刻画单个金融资产收益边际分布的非对称性和厚尾性。利用Kupiec的失败频率检验方法,本文验证了模型的有效性。实证研究表明Copula函数结合非对称Laplace分布技术能够很好地度量投资组合的风险价值VaR,从而有助于更好地测度和规避金融市场的风险。
In calculating VaR of a portfolio,multivariate normal distributions are often assumed on the returns of the assets. Though simple for calculation of VaR,this assumption often does not conform with what are seen in the real market,and tends to underestimate risks in the extreme cases. To get over this problem,we propose in this paper a new approach by combining copula function technique with asymmetric Laplace distribution,which can better catch the dependence among the returns and the skewness and heavy-tails of the individual return. Empirical studies based on Kupiec's failure test show that this method works reasonably well for measuring portfolio risks,and can help measure and control financial market risks.
出处
《管理评论》
CSSCI
2008年第4期10-16,共7页
Management Review
基金
国家自然科学基金资助项目(70271003
70221001)