机构投资者、股权分置改革与股市波动性——基于MCMC估计的t分布误差MS-GARCH模型

从2003年开始, 中国机构投资者占股市流通市值中的比例迅速增长. 论文以这段时期上证指数的日收益率序列为研究对象, 改进了最新的t分布误差MS-GARCH模型, 运用马尔科夫链蒙特卡罗模拟(MCMC)对该模型进行了估计, 为研究股权分置改革、机构投资者对股市收益率波动的影响提供了新的证据. 研究发现, 股权分置改革使股市波动性发生了结构性改变, 股市由低波动风险期转换为高波动风险期; 各类基金的总净值和仓位给股市波动性带来的影响有显著差异, 存款准备金率和利率的调整也会影响股市波动性. 最后, MS-GARCH模型对股市数据的拟合度和预测效率等都优于单状态GARCH模型.

Institutional investors, sub-owned shares reform and stock returns volatility: MCMC estimation of a MS-GARCH model with student-t error

There is a dramatically increase in institutional investors ownership of Chinese stock market from 2003. In this paper, we improve the MS-GARCH model to analyze the daily returns data, and provide empirical evidence on the impact of sub-owned shares reform and institutional investors in stock market. We find that Chinese stock market transform form low volatility state to high volatility state after the sub-owned shares reform. The results also prove that different funds have vary effects on the volatility of the stock market, and the adjusting of interest rate and reserve ratio could also affect the the volatility of the stock market. According to the comparison between two models, MS-GARCH model is much better than single-regime GARCH model in goodness-of-fit and forecasting performance.