The Research Of Stock Market Volatility Based On Markov Regime Switch Stochastic Volatility Model

Volatility is one of the most characteristics in financial market. In order to find out the essence and characteristics, financial-economists, econometrics-economists and many other authors all bend themselves to research on the models with the development of econometrics and computer software. They advanced many models from different aspects. The ARCH-GARCH (Auto-regressive conditional heteroskedaticity and Generalized auto-regressive conditional heteroskedaticity) models and SV models are the most important models. ARCH-GARCH models and SV models both can conquer the limitation that conventional model contains and this two kinds models have become the major models to analyze the volatility of stock market. But the two models still have limitations. They both ignore the regime switch in volatility. In financial market, the regime switch in volatility is ubiquitous and the same in stock market. Therefore, authors all reckon that should introduce the regime switch into the ARCH-GARCH models or SV models.In this paper, Markov regime switch is introduced into SV models and advance a SVMRS (Stochastic volatility with Markov regime switch) model, then this model is used to analyze the Shanghai stock market. In the first, this paper reviews the research on stock market volatility. Then this paper discusses the rudimental theory and characteristics of stock market and the reasons that cause change in the stock market. This paper also introduces the ARCH-GARCH models and SV models and method to estimate the parameters in those models. This paper advances the SVMRS model at last and we apply it to the Shanghai stock market yield series This paper also compare the SVMRS model with SV model and Smith model.The result shows that the Shanghai Stock Market returns serial don't follow normal distribution and the volatility contains regime switch; Due to the above mentioned reasons, SVMRS model can reduce the persistence of models and reveal the real characteristic of stock market, moreover it can also forecast better than the SV model and Smith model.