沪深300股指期货的波动率预测模型研究

Volatility forecasting models for CSI300 index futures

以沪深300股指期货仿真交易的5分钟高频数据为例,运用滚动时间窗的样本外预测和具有Bootstrap特性的SPA检验法,全面对比了基于日收益数据的历史波动率(historical volatility)模型和基于高频数据的已实现波动率(realized volatility)模型对波动率的刻画和预测能力.主要实证结果显示,已实现波动率模型以及加入附加解释变量的扩展随机波动模型是预测精度较高的波动模型,而在学术界和实务界常用的GARCH及其扩展模型对沪深300股指期货的波动率预测能力最弱.

Taking 5-minutes high-frequency mock trading data of CSI300 index futures as example, the out-of- sample daily volatility predictions of these models are calculated by using rolling predicting method, and a bootstrap SPA test is used to evaluate the predicting accuracy for different historical volatility models and real- ized volatility models. The empirical results show that, realized volatility model based on high-frequency data and the extended SV model are superior to other models. However the GARCH and its extended model, which are popular in financial academe and practice, perform worst for volatility predicting of CSI300 index futures.