带时变高阶矩的已实现GAS模型及其实证研究

The Realized GAS Model with Time-varying High-order Moments and Its Applications

基于广义自回归得分模型框架,对资产收益率和已实现测度联合建模,假设资产收益率服从时变学生t分布,构建了带时变高阶矩的已实现GAS(TVP-RGAS)波动率模型.TVP-RGAS波动率模型将学生t分布中的自由度参数时变化,可以捕获金融资产收益率序列中的时变高阶矩特征.引入已实现测度,有助于利用日内价格变动信息从而更好地捕获市场波动的急剧变化.TVP-RGAS波动率模型属于观测驱动模型,可以使用极大似然方法估计,具有易于实现的优点.采用上证综合指数进行实证分析,结果表明,上证综合指数的收益率分布存在明显的时变高阶矩特征;TVP-RGAS波动率模型相比已实现GAS(RGAS)波动率模型和已实现GARCH(RGARCH)模型拥有更好的条件方差预测能力以及VaR预测效果.

This paper proposes an extension of the generalized autoregressive score(GAS) model, namely the Realized GAS model with time-varying high-order moments(TVP-RGAS model). This model applied to volatility and Value-at-Risk(VaR) forecasting with Student’s t distributions and a time-varying degrees of freedom.It can capture the time-varying high-order moment characteristics in the return on financial assets sequence.In addition, the model is easy to use in practice. The empirical analysis using the Chinese stock indices, the Shanghai Stock Exchange Composite(SEC), demonstrates that the TVP-RGAS model produces more accurate volatility and VaR forecasts than the realized GARCH(RGARCH) and Realized GAS(RGAS) models.