摘要
从原油现货市场收益率的特征分析入手,为了更好地描述原油现货市场收益率的尖峰厚尾、偏态和波动集聚等特征,运用EGARCH对条件波动率进行建模,进而运用极值理论对标准残差序列的尾部分布进行建模,刻画原油现货市场极值风险,同时结合Copula函数和Monte Carlo模拟技术来度量不同持有期相应的VaR值。实证结果表明:原油市场随着置信度的提高和持有期的延长,VaR的绝对值在增大。同时,回测检验结果表明基于EGARCH-EVT-t Copula的模型能够精确有效地度量原油现货市场极端风险。
Based on the characteristics of the rates of return of crude oil spot market, and in order to accurately describe the characteristics of fat-tails, leptokurtosis, skewness and volatility, this paper used EGARCH model to conditional volatility, and then used extreme value theory model extreme tail of standard residuals to calculate the extreme risks of crude oil spot market. Copula function and Monte Carlo simulation method are used to measure the VaR from one day to one month. The empirical results show that for the crude oil spot market, with the increase of confidence and the extension of holding, the absolute value of the VaR increases. Moreover, the back testing results indicate that the extreme risk measurement is feasible and effective based on the EGARCH-EVT-t Copula model.
出处
《北京理工大学学报（社会科学版）》
CSSCI
2012年第4期10-16,共7页
Journal of Beijing Institute of Technology：Social Sciences Edition
基金
中央高校基本科研业务费资助项目(CDJXS11021112)
