Thick Tail And Leverage Sv Model Based On Bayesian Analysis Of China's Stock Market

The finance time series have the universal phenomenon of volatility, but the volatility is a core research question which to describe a finance market. Stochastic volatility model is a good class model to descript volatility. When we descript the volatility of stock market in our country by basic stochastic volatility, it's not satisfied us. One side, compared with standard normal distribution of varibles assumption in heteroskedastic model, the unconditional distribution of finance time series will show bigger kurtosis and thicker tail. In other side, it's correlated between return of assets and volatility. In this paper,we will analyze volatility of stock market in our country through heavy-tailed stochastic volatility model and leverage effect stochastic volatility model.Because the MCMC(Markov Chain Monte Carlo) method is efficient in dealing with problems of high dimension, likelihood function and posterior distribution. In this paper, we will estimate parameters in models by MCMC method. First, the Bayes analysis of basic SV model, SV-T model and leverage effect SV model will be studied by Bayes theory, and we will get posterior distribution of parameters. Then Markov Chain Monte Carlo algorithm procedures with Gibbs sampler and Metropolis-Hastings sampler were designed to estimate the models' parameter through matlab software, in which the sample are Shanghai composite index and Shenzhen composite index.Following , we will analyze volatility characters of Shanghai stock market and Shenzhen stock market, and also we will compare these two market to study volatility of our country. For the comparision with different SV models, most researchers used Bayes Factou.But calculating capacity is required highly to this method. So DIC will be used to compare different SV models with Shanghai composite, and find out the best SV model for our country.