摘要
以我国上证综指收益率序列为研究对象,分析了其分布及特性。针对上证指数收益率不服从正态分布,具有"有偏、尖峰、厚尾"的特性,提出了基于g-h分布假设的VaR计算方法,并与正态假设、Logistic假设、Student-t假设和历史模拟法计算的VaR进行了比较,结果表明基于g-h分布假设下的VaR对收益率数据的风险描述最为准确,计算结果优于其它3种分布。
The calculation of value at risk(VaR) on shanghai stock exchange was discussed. By analyzing the empirical data of shanghai stock exchange,it was proved that there exists obvious difference between return rate and normal distribution,and the empirical distribution was characterized by asymmetry,leptokurtosis and heavy tails.A new distribution——g-h distribution was used to calculate VaR,and the calculated results indicate that g-h distribution is better than normal distribution、Logistic distribution and Student-t distribution.
出处
《兵工学报》
EI
CAS
CSCD
北大核心
2009年第S1期175-180,共6页
Acta Armamentarii
