Value At Risk Estimation Of Stock Market Yield Based On MCMC-GARCH Model

During recent years, the globalization of financial market and the liberation of interest rate has been making the structure of financial risk more and more complicated, thus making the function of risk management more important.At present, Va R based on GARCH model is the mainstream of the current financial market risk measurement method.In traditional way, for GARCH models, the estimation of parameters is based on the maximum likelihood theory.In such condition, parameters are subject to many qualifications. So it is difficult to get precise parameters estimation and it is hard to achieve optimal goals. In this paper, Bayesianapproach based on Markov Monte Carlo(MCMC) has been adopted in order to solve such a problem in another way and then a more accurate describe of the volatility is expected, which will be shown in an intuitive and obvious way by estimating the value of Va R based on the GARCH model we have been established.In the empirical analysis in this paper, Shenzhen Composite Index closing price has been collected as the original data, then a simple calculation process has been taken and we ultimately get the form of index returns as the final object.Data in this paper is divided into two sections, one section for establishing and estimating GARCH model, the other section for Va R calculations and back-testing.In addition,ML method respect for the Classical Statistical method realized by Eviews and MCMC method respect for the Bayesian method realized by Open Bugs are respectively used when analysising the GARCH model.Empirical results shows that,GARCH models based on Bayesian theory usinig MCMC method, performs better in characterize the volatility of the market and then get a more accurate estimation of Va R value based on this.