Research On Volatility Characteristics Of Stock Markets Based On Bayesian Asymmetric Stochastic Model Using GH Distribution

Volatility is the eternal theme of financial markets. Many factors including macroeconomic policy changes, the financial crisis, oil prices change and important event shocks often have huge impact on the financial markets. Financial market volatility always presents clustering, fat-tailedness and the "leverage effect". Recent related studies indicts that volatility also presents the skewness. In the modern financial studies, if the any characteristic of volatility was not fully considered, it would be easy to make a biased estimation results. On the premise of guaranting data stationarity, traditional stochastic volatility model does not guarantee that the data is stationary, and normality assumption between samples is difficult to set up.So in the view of the existing financial time series modeling are difficult to capture multiple wave characteristics and invalid inference problems, this paper set the conditional distribution of yield obeys the generalized hyperbolic distribution using bayesian technology, and build a asymmetric stochastic model based on bayesian using generalized hyperbolic distribution, in order to describe multiple volatility characterization of financial assets and explore the major impact of system environment of fluctuations.This paper studies from the basic theory to model building and then the empirical study. First of all, this paper reviews the stochastic volatility model, generalized hyperbolic distribution, the characteristics of stock market, the parameter estimation methods and so on; Secondly, it sets the conditional distribution of yield obeys the generalized hyperbolic distribution, using data expansion to the parameter space, and builds a asymmetric stochastic model based on bayesian using generalized hyperbolic distribution with applying the bayesian theory to make statistical inference and MCMC sampling algorithm;Finally, it selects the United States, Japan, Britain and Hong Kong's stock market such as Standard&Poor's 500 index, the Nikkei 225 index, the FTSE 100 index and the Hang Seng index from 2005 to 2015 as the research objects, so as to capture the multiple volatility characteristics of the stock market and explore the major impact of system environment of fluctuations.At the same time, it uses DIC criteria to compare ASV-ST model, ASV-T model and ASV-N model.The research results show that the United States, Japan, Britain and Hong Kong, China, stock market volatility present clustering, fat-tailedness and the "leverage effect" and so on, and Hong Kong's,China, return distribution of stock market is right, the United States, Japan, the UK stock's return distribution of stock market is left. Hong Kong, China, the stock market fluctuations in logarithmic average absolute value and volatility index are the biggest, suggests that Hong Kong as the important part of emerging capital market, the market volatility reaction from outside information is bigger than mature the United States, Japan and the UK stock market;Second, the United States, Japan, Britain and Hong Kong, China, stock market in 2008 and 2011 two times show a strong fluctuation, this suggests that significant impact in the system environment such as the financial crisis, the oil price shocks have a big impact on stock market volatility; Finally, the dynamic iteration trajectory, G-R convergence diagnosis and kernel density estimation curve verify reasonable and effectiveness of the ASV model using generalized hyperbolic model.