Theoretical Study Of China's Stock Market Price Volatility With Applications

This dissertation takes a close and deep look at volatility of China's Stock Market and its main subjects are as follows:1. using the data from China's Stock Markets, a Gibbs sampling based analysis and comparison of ARCH-GARCH models and SV models are made in this dissertation. The models-based simulation results suggest that in view of either capturing thick tails, or reproducing the strong autocorrelations of squared returns, SV models are superior to ARCH-GARCH models; furthermore, in comparison of t distribution–based models with N-distribution-based models, the former is better than the latter and thus the returns series of stock market is not of N-distribution.2. having proposed a mean-reverting stochastic volatility model for discrete time. Parameters of the model are then estimated with MCMC method based on Gibbs sampling and the model is used to investigate China's Stock Market. The results show that the mean-reverting time of Shanghai Index volatility is the same as that of Shenzhen Index, about 249 trade days.3. having extended N-distribution-based SV-m model. An A-SV-m model which can catch the effect of asymmetry information and a t distribution–based SV-m model are proposed. These models are used in the empirical studies of the relationship between expected returns and volatility. The results show the relationship between expected returns and volatility is time varying, and the effect of volatility on expected returns is weak. By using previous studies, this dissertation shows that the relationship is positive when markets and investors are rational, and negative when markets and investors are irrational. And when the numbers of rational and irrational investors are almost equal, the effect of volatility on expected returns is weak.4. having extended a Bivariate Mixture Distribution(BMD) model. The model can catch the asymmetry information effect, and is used in the empirical studies of China's stock markets. The results indicate that the impact of shock from positive returns and increased trading volume on volatility is larger than that from negative returns and decreased trading volume in same magnitude, respectively, before 1997 in China's Stock Markets; but after 1997, it was just the opposite. The results also show that the BMD model can capture the persistence of return volatility.5. finally summaring the closed-end fund discount puzzle, and using the stock index data from Shanghai and Shenzhen Stock Markets and the respective closed-end fund discount indices of these two markets, for an empirical study of the relationship among China's Stock Markets volatility, closed-end fund discount and investor's sentiment, based on an augmented SV model. The result shows the relationship between discount and volatility is negative, which indicates that the discount is not a substitute for investors' sentiment, and the Investors' Sentiment Hypothesis can't explain the puzzle of discount completely in China's Stock Market.