摘要
以我国股票收益率为研究对象,分别在正态分布和非正态稳定分布条件下对ES和VaR的凸性、次可加性和有效性进行了实证比较分析,发现:在非正态稳定分布条件下VaR不满足凸性和次可加性,ES满足凸性和次可加性,在正态分布条件下VaR和ES都满足凸性和次可加性;在两种分布条件下ES的有效性都高于VaR的有效性,而在非正态稳定分布条件下ES的优势更加明显。由于本文的收益率分布拟合检验表明我国的股票收益率服从非正态稳定分布,所以在我国股票市场上ES是比VaR更好的风险度量。
This paper empirically compares ES with VaR from aspects of convexity, subadditivity and validity on condition of normal distribution and non-normal stable distribution on China Stock Market, and finds: ES satisfies convexity and subadditivity but VaR do not on condition of non-normal stable distribution, and both of them satisfy convexity and subadditivity on condition of normal distribution; whether on condition of normal distribution or non-normal stable distribution, ES is more effective than VaR, and the problem of underestimating risk of VaR is more serious on condition of nonnormal stable distribution. Because the distribution of rate of return on China Stock Market is non normal stable, ES is a better risk measurement than VaR on China Stock Market.
出处
《技术经济》
2006年第12期1-6,共6页
Journal of Technology Economics
基金
国家自然科学基金(70440003)