摘要
基于Copula-GARCH模型,以最小VaR为目标函数,考虑了金融资产收益率分布呈现尖峰厚尾情形下的最优套期保值比率的估计问题.首先,基于Cornish-Fisher展开逼近分位数的方法,给出了最小化VaR确定最优套期保值比率的原理;其次,考虑到收益率序列的条件异方差及厚尾性,建立Copula-GARCH模型对波动率及相关系数进行估计;最后,给出了基于沪深300股指期货的实证研究,并将其与最小方差法、正态分布最小VaR法的实证结果进行了比较分析.结果表明:在样本内,3种方法的套期保值效果相差不大,在样本外,非正态分布最小VaR法的套期保值效果显著高于其他模型.
Based on the Copula-GARCH model, taking the minimum VaR as the objective function, estimated the optimal hedge ratio in the case of the return distribution of financial asset has an obvious peak and fat tails. Firstly, based on the Cornish-Fisher ex- pansion method, given the principle of minimizing the optimal hedging ratio of VaR. Secondly, considered the phenomenon of volatility clustering, sharp peaks and fat tail distributions of return series, the Copula-GARCH model is established to estimate the volatility and the correlation coefficient. Finally, given an empirical study on the Hushen 300 index futures, and the results are compared with the least square method, and minimum VaR method based the normal distribution. The result shows that: in the sample, the difference of hedging efficiency among three methods is not large, even can be ignored, besides, in the out of sample, the hedging efficiency of the least VaR method which considered the actual distribution of return series is significantly higher than other models.
出处
《伊犁师范学院学报(自然科学版)》
2017年第4期18-22,共5页
Journal of Yili Normal University:Natural Science Edition
基金
国家自然科学基金资助项目(71501001)
教育部人文社科研究青年基金资助项目(14YJC790133)