通过实证分析提出了一个全面反映股指期货市场价格风险变化特征的向量空间测度体系,并对香港恒生指数期货收益率序列与基差序列的风险进行了度量,采用R/S分析方法对H指数进行了估计,采用反映市场时变特征的GARCH类模型分别在两种不同的概率分布下预测了未来一日的CVaR值,应用Kupiec准则测试了估计出的CVaR值的准确程度,并对比分析了各模型不同分布下CVaR值的精确程度。结论表明,PARCH模型是计算时变市场风险CVaR值的较佳模型。
This paper presents a vector space measurement system that can make comprehensive response to changes in stock index futures market price risk characteristics.This paper empirically analyzes the risk measurement of the Hang Seng Index futures return series and the basis sequence.It also uses the R/S analysis method to estimate H index,and compute future day CVaR values by using GARCH class model which fully reflect time-varying characteristics at two different probabilities.It applies the Kupiec test to estimate the accuracy of the CVaR value and comparative analysis on CVaR calculation accuracy under different distributions models.Empirical study shows that PARCH model is a better model to calculate the time-varying market risk CVaR value.