基于不同风险特征的跳跃成分识别及其在波动率预测中的应用

跳跃因子的引入能够准确解释波动的非对称特征,同时跳跃中还含有关于波动率的未知信息.为了更有效地改进波动率的预测,利用基于高频数据的非参数波动估计和跳跃检测方法,在波动的非对称性基础上对跳跃作进一步分解,考察具有不同风险特征的跳跃成分对未来波动率的影响,并对2009-2014年上证综指及其行业指数的面板数据进行实证分析.实证研究发现:周期性行业指数的系统性跳跃对其波动率有显著的预测效力,大盘指数与行业指数之间存在高度相关性;而非周期性行业指数几乎没有表现出明显的杠杆效应,与大盘指数的相关性也较低.

Jump tests based on different risk characteristics and the prediction of volatility

The asymmetry of volatility could be correctly explained by jumps which also involve some information additionally.In order to improve the prediction of volatility,by employing the realized volatility and non-parametric jump detection method using high frequency data,this paper discusses the effect of jumps of different risk characteristics on future volatility based on the study of the asymmetry of volatility and conducts an empirical analysis with the SH indexes panel data from 2009 to 2014.The results indicate that systematic jumps of economic cyclical industry indexes bear significant effect on volatility prediction,which means a high degree of correlation between the market index and industry index; while those aperiodic industry indexes almost show no discernible leverage effect,with lower correlation of the market index.