首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于ARMA-GARCH-SN模型的沪深300股指期货日内波动率研究与预测
引用本文:王苏生,王俊博,许桐桐,余臻.基于ARMA-GARCH-SN模型的沪深300股指期货日内波动率研究与预测[J].运筹与管理,2018,27(4):153-161.
作者姓名:王苏生  王俊博  许桐桐  余臻
作者单位:1.哈尔滨工业大学 深圳研究生院、城市规划与管理学院,广东 深圳 518055; 2.前海金融控股有限公司 博士后创新实践基地,广东 深圳 518052
基金项目:深圳市科技创新委员会知识创新计划项目“基于高频数据的证券市场动力学及其应用研究”资助(JCYJ20140417173156101)
摘    要:运用五个交易日的股指期货高频数据(每秒两笔),本文主要研究了沪深300股指期货日内波动率特征并对日内波动率预测。研究发现高频股指期货日内收益率有明显的波动率聚集和条件异方差现象,但无尖峰厚尾现象,收益率序列分布符合有偏正态分布。因此,我们对时间序列建立了最优的ARMA-GARCH-SN模型,并对模型拟合充分性做了验证,拟合结果发现ARMA(1,2)-GARCH(1,1)-SN模型基本能够刻画股指期货高频日内波动特征,条件方差所受的冲击具有很强的持续性、日内波动也具有长记忆性,最后我们还利用自助法对高频股指期货日内波动率两步预测、利用滚动回归预测方法对样本做了样本内预测。预测结果表明,波动率预测结果能够较好地反映股指期货日内波动特征。

关 键 词:高频数据  ARMA-GARCH-SN模型  沪深300股指期货  日内模式  预测  
收稿时间:2016-10-25

Volatility Research and Forecast on CSI 300 Index Futures byUsing the ARMA-GARCH-SN Model
WANG Shu-sheng,WANG Jun-bo,XU Tong-tong,YU Zheng.Volatility Research and Forecast on CSI 300 Index Futures byUsing the ARMA-GARCH-SN Model[J].Operations Research and Management Science,2018,27(4):153-161.
Authors:WANG Shu-sheng  WANG Jun-bo  XU Tong-tong  YU Zheng
Institution:1.Harbin Institute of Technology Shenzhen Graduate School、School of urban planning and management, Guangdong Shenzhen 518055, China; 2.Postdoctoral iNNOVATION Practice Base of Qianhai Finance Holding Ltd, Guangdong Shenzhen 518052, China
Abstract:By using the high frequency data(two deals per second) of five trading days in CSI 300 index futures, this paper mainly studies the intraday volatility characteristics of the CSI 300 stock index futures and forecasts the intraday volatility. The results show that the intraday yield of high frequency stock index futures has obvious volatility aggregation and AutoRegressive Conditional Heteroscedasticity(ARCH)effect, but there is no peak and fat-tailed phenomenon. The distribution of the return series accords with a skewed normal distribution. So the optimal ARMA-GARCH-SN model is established, and the adequacy of the model fit is verified. The fitting results show that ARMA(1,2)-GARCH(1,1)-SN can well describe the characteristics of ultrahigh frequency intraday volatility of CSI 300 index futures, the impact on the conditional variance has a strong persistence, and long memory effect is found in the intraday volatility. Finally, two steps of the return rate and the volatility are predicted by the bootstrap method. We also use the rolling regression forecasting method to do the in-sample prediction. The prediction results show that the volatility prediction can reflect the intraday volatility characteristics of the stock index futures commendably.
Keywords:high-frequency data  ARMA-GARCH-SN model  CSI 300 index futures  intraday pattern  forecast  
本文献已被 CNKI 等数据库收录!
点击此处可从《运筹与管理》浏览原始摘要信息
点击此处可从《运筹与管理》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号