状态转换和中国股市的独特特征——基于马尔可夫状态转换-自回归模型的分析

Regime Switch and Features of Chinese Stock Market

本文采用马尔可夫转换-自回归模型分析了上证综指的周收益率。通过6个模型的比较,本文指出状态转换模型明显优于普通GARCH模型。研究表明中国股市存在多个独特特征:收益率在不同状态之间的变动规律差异显著;低波动状态的持续时间最短,出现频率也最低,而高波动状态出现次数最多,并且同牛市的相关性显著;中国股市中,低和中等波动状态之间无法直接转换,而是必须通过高波动状态作为媒介而相互转换。这些特征都显著区别于成熟市场,也提供了中国股市缺乏有效性的直接证据。本文结论有助于风险控制、预测等金融实践操作,对于股市制度设计和创新也能提供一定的方向指导。

The paper analyzes Shanghai Stock Exchange Index via Markov Switching AR Models,and points out the superiority of MS model in comparison to normal GARCH models.As conclusion,we point out that the return dynamics between different regimes vary greatly in Chinese stock market;second,the low volatility regime in China has shorter duration and infrequent occurrence,while the high volatility regime appear more frequent and correlates to bull market significantly;third,the low and medium volatility regimes can not switch directly to each other in Chinese stock market.These features differ significantly from the mature markets in Europe and USA,proving the inefficiency of Chinese stock market.These results can help in risk management and forecast in finance practice,and also provide guidance for system design and innovation of Chinese stock market.