中国股票市场的波动率聚集性研究——基于Markov机制转换Copula模型的实证分析

Volatility Clustering in Chinese Stock Markets: An Empirical Analysis Based on Markov Regime Switching Copula Model

波动率聚集性是金融资产收益率序列中的一个重要特征。构建了Markov机制转换Copula模型研究中国股票市场的波动率聚集性(波动率相关性结构)。采用上证综合指数和深证成份指数日内高频数据,构造已实现波动率作为隐波动率的代理变量,对中国股票市场进行了实证分析。结果表明,SJC Copula相比其他Copula能更好地刻画中国股票市场的波动率聚集性,波动率聚集具有明显的尾部非对称特征,高波动率的聚集相比低波动率的聚集发生概率要更高。另外,基于Markov机制转换SJC Copula模型的研究表明,中国股票市场的波动率聚集还具有明显的尾部动态特征。

Volatility clustering is one of the most important stylized facts in financial asset return series. This paper constructs the Markov regime switching copula model to study the volatility clustering(volatility dependence structure) in Chinese stock markets. Using the realized volatility constructed from the intraday high-frequency data as a proxy for the latent volatility, an empirical study of the Shanghai Stock Exchange composite index and the Shenzhen Stock Exchange component index of China is conducted. The empirical results demonstrate that the Symmetrized Joe Clayton(SJC) copula outperforms other copulas. The volatility clustering in Chinese stock markets exhibits asymmetric features in the sense that clusters of high volatility occur more often than clusters of low volatility. In addition, based on the results of the Markov regime switching SJC copula model, it is also found that the tail dependence of volatility clustering in Chinese stock markets exhibits a dynamic behavior.