论文部分内容阅读
期货隔夜风险的防范历来是投资者关注的热点,本文以沪深300股指期货为研究对象,采用CAViaR模型对普通隔夜风险进行度量,同时还采用新建的CAViaR-EVT模型对极端隔夜风险进行预测,全面地分析了多头VaR和空头VaR在不同分位数的动态变化特征,最后采用Kupiec似然比检验和动态分位数检验对模型进行后测检验。实证结果表明,隔夜收益序列具有右偏、无长期记忆性和尖峰厚尾等典型特征;CAViaR模型对股指期货的普通隔夜风险具有优异的预测能力,其中AS模型的预测效果最好;加入极值理论后,CAViaR-EVT模型同样能很好地刻画极端分位数下隔夜风险的动态演化过程,且其预测结果比EVT和GARCH-EVT模型要更合理。
The prevention of futures overnight risks has always been a hot spot for investors. This paper takes the CSI 300 stock index futures as the research object, uses the CAViaR model to measure the common overnight risk, and forecasts the extreme overnight risk using the new CAViaR-EVT model. The dynamic characteristics of long VaR and short VaR in different quantiles are analyzed comprehensively. Finally, Kupiec likelihood ratio test and dynamic quantile test are used to test the model. Empirical results show that overnight returns have the typical features of right deviation, no long-term memory and peak-thick tail. CAViaR model has excellent predictive ability for ordinary overnight risk of stock index futures, of which AS model has the best forecasting effect. After the theory, the CAViaR-EVT model can also well characterize the dynamic evolution of overnight risk under extreme quantiles, and its prediction result is more reasonable than the EVT and GARCH-EVT models.