摘要
金融资产收益率不仅具有尖峰厚尾性、异方差性,还具有长记性。基于此,将FIGARCH、EVT和Copula有机融合,建立FIGARCH-EVT-Copula模型来估计组合风险值,利用上证指数、深成指数组合进行实证研究。实证研究表明:我国股市波动确实具有长记忆性;FIGARCH-EVT-Copula模型不仅能够准确刻画边缘分布的尖峰厚尾性、异方差性和长记忆性,而且较之于传统模型,该模型能更准确地测度投资组合风险。
It is well known that financial equity has sharp-peaks, fat-tails, heteroskedasticity and long memory. Considering these three features, this article constructs a risk measure model based on the FIGARCH-EVT-Copula for financial portfolio. The VaR and ES risk measure based on the FIGARCH-EVT-Copula is applied on the portfolio, which is composed by Shanghai Stock index and Shenzhen Component Index equal weight. The empirical results show that there is apparent long memory property in Chinese stock market. The results also show that, the model of FIGARCH-EVT-Copula really can capture the properties of sharp-peaks, fat-tails, heteroskedasticity and long memory, and proves that the model of FIGARCH-EVT-Copula is more efficiency than traditional model in measure the portfolio risk whose marginal distribution has the property of long memory.
出处
《北京理工大学学报（社会科学版）》
CSSCI
2012年第1期44-49,共6页
Journal of Beijing Institute of Technology：Social Sciences Edition
基金
国家自然科学基金资助项目"流动性调整期望损失La-ES和最优变现策略"(70671025/G0115)
