基于CAViaR和GARCH模型的沪深300股指期货动态风险测度

发布时间：2018-04-11 01:36

本文选题：CAViaR模型 + GARCH模型　；参考：《系统工程》2017年03期

【摘要】：以我国期货市场上交易最为活跃的沪深300股指期货为例,分别采用CAViaR模型和GARCH模型对多头VaR和空头VaR进行风险建模,深入研究了股指期货的收益分布特征和波动形态规律,并运用严谨的后测检验的方法对比了各个模型的风险预测精度。实证结果表明:(1)沪深300股指期货具有明显的"尖峰厚尾"现象,却没有显著的有偏性和长记忆性;(2)基于杠杆效应的GJR模型和兼具长记忆性和杠杆效应的FIAPARCH模型并没有表现出比传统GARCH模型更高的预测精度,同时,先验GED分布对金融收益分布特征的刻画要优于正态分布和SKST分布;(3)半参数法的CAViaR模型相比GARCH族模型表现出绝对优异的预测能力。总之,CAViaR模型在股指期货的风险预测方面是相对更合理的模型选择。
[Abstract]:Taking CSI 300 stock index futures, which is the most active traded stock index futures in futures market in China, as an example, this paper uses CAViaR model and GARCH model to model the risk of long VaR and short VaR, and deeply studies the distribution characteristics and fluctuation pattern of stock index futures.The risk prediction accuracy of each model is compared with the rigorous post-test test method.The empirical results show that the Shanghai and Shenzhen 300 stock index futures have obvious "peak and thick tail" phenomenon.However, there is no significant bias and long memory. The GJR model based on leverage and the FIAPARCH model with both long memory and leverage effect have no higher prediction accuracy than the traditional GARCH model.A priori GED distribution characterizes the characteristics of financial return distribution better than the CAViaR model based on normal distribution and SKST distribution. Compared with the GARCH family model, the CAViaR model shows absolutely superior predictive ability.In short, CAViaR model is a more reasonable model choice in the risk prediction of stock index futures.
【作者单位】：复旦大学经济学院;
【分类号】：F224;F724.5

【相似文献】