基于MRS-EGARCH模型的沪深300指数波动预测  

HS300 index volatility forecast based on Markov regime-switching EGARCH model

作　　者：张锐[1] 魏宇[1] 金炜东[1] 

机构地区：[1]西南交通大学经济管理学院,四川成都610031

出　　处：《系统工程学报》2011年第5期628-635,共8页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(70501025;70771097);教育部新世纪优秀人才支持计划资助项目(NCET-08-0826);中央高校基本科研业务费专项资金资助项目(SWJTU09ZT32;SWJTU09CX088);教育部创新团队计划资助项目(PCSIRT0860)

摘　　要：提出了一种新的马尔可夫机制转换EGARCH模型,假定收益残差序列可以服从高斯分布、t-分布或广义误差分布,并允许非高斯分布中自由度与所处机制有关,以刻画可能存在的时变峰度及厚尾特征.以沪深300指数为例进行实证研究,发现新模型能区分隐藏在指数收益序列中的不同机制.预测成功率指标表明设定收益残差服从厚尾分布的MRS-EGARCH比单机制EGARCH具有更好的波动率预测性能.This paper developed a new Markov regime-switching EGARCH model. Gaussian, t, and GED conditional distributions for residuals were assumed and the degrees of freedom of non-Gaussian distribution were allowed to be regime-dependent to depict possible time-varying kurtosis. The results of empirical re- search, which use HS300 index as an example, show that the new model can distinguish different regimes hidden in the return series. Furthermore, the values of standard statistical error functions and the success ratio of directional prediction of volatility indicate that the new model with specification of fat-tailed conditional distributions of residuals outperforms single regime EGARCH model in volatility forecast.

关 键 词：波动率 持续性 EGARCH 马尔可夫机制转换模型 

分 类 号：F224[经济管理—国民经济]