基于MRS-GARCH模型的中国股市波动率估计与预测  

Estimation and Forecasting of Volatility in China's Stock Markets with Markov Regime-switching GARCH Model

作　　者：赵华[1] 蔡建文[1] 

机构地区：[1]厦门大学经济学院,福建厦门361005

出　　处：《数理统计与管理》2011年第5期912-921,共10页Journal of Applied Statistics and Management

基　　金：教育部人文社会科学研究项目(08JC790089);福建省重点科技项目(2009R0079);中国博士后基金(20090450006);中央高校基本科研业务费专项资金资助(2010221055)

摘　　要：基于误差项服从正态分布、t分布、广义误差分布的GARCH族模型和MRS-GARCH模型对中国股市波动的结构变化特征进行了实证研究。结果表明,中国股市存在显著的高、低波动状态,两种波动状态的ARCH和GARCH项系数存在较大差异;高、低波动状态均具有较长的持续时间,低波动状态的持续时间长于高波动状态的持续时间,且中国股市更易于从高波动状态转向低波动状态;MRS-GARCH模型预测效果总体上优于GARCH族模型,基于正态分布的MRS-GARCH模型短期预测效果较好。Based on GARCH models and Markov regime-switching GARCH model with normal, t and generalized error distributions, the paper studied empirically the structural break characteristics of volatility in China＇s stock markets. The results show that in the China＇s stock market there are significant high and low volatility states, and the ARCH and GARCH coefficients differ substantially between the two states. Furthermore, the high and low volatility states all have long durations, the duration for the low volatility state is longer than that for the high volatility state, and China＇s stock markets are prone to transit from low volatility state to high volatility state. Finally in out-of-sample forecast the Markov regime-switching GARCH model outperforms the GARCH models and the Markov regime-switching GARCH model with normal distribution performs best in the short run.

关 键 词：MRS—GARCH模型 DM检验 结构变化 

分 类 号：F830.91[经济管理—金融学] O212[理学—概率论与数理统计]