基于GARCHSK的铜期货VaR估计方法研究

Study on the Estimation of VaR of Copper Futures  Based on GARCHSK Model

摘要: 基于GARCHSK模型对期货收益序列的条件偏度和峰度进行动态建模，提出了“有偏”和“尖峰厚尾”分布下的VaR估计方法。通过对沪铜期货的实证研究表明，其收益分布存在明显的“有偏”和“尖峰厚尾”。基于不同分布假定下的VaR估计结果的Kupiec检验表明，基于GAR CHSK的VaR估计方法能够有效提高VaR的估计精度。

Abstract: This paper has estimated the conditional skewness and kurtosis using t he GARCHSK model and then proposed an estimation method of VaR based on skewed and fat tailed distribution. The empirical results from Shanghai copper futures data suggest that the conditional skewness and kurtosis are time varying significantly. And the estimation method based on GARCHSK model can make the VaR estimation more accurate after considering the impacts of conditional skewness and kurt osis.