The Study On DCC-MVGARCH Model And Empirical Analysis In Financial Market

During last twenty years, time series econometrics models have been developing quickly, they are widely used in varies fields abroad. Especially in bond market, stock market... They provide very valuable information for the market operator and risk manager to make forecasting and deciding. Also, they become the necessary tool of financial market analyst. At present, the GARCH models are the most common use models in analyzing the return volatility of financial markets. However, with economy internalization and financial liberalization, the relation between each financial market of different areas become more and more close. It's necessary to study the correlation of financial markets. Thus, we should use multivariate GARCH models, such as VECH models, BEKK models, CCC models. But they have several limitations in parameter estimating, calculating andexplaining. A new model--DCC-MVGARCH , proposed by Engle(2001) , candeal with these shortages well.This article first overview some multivariate GARCH models that are in common use. Then, introduce DCC-MVGARCH model which was proposed by Engle, study its algorithm and realize them by MATLAB program. After that, we apply the DCC-MVGARCH model to investigate the effects of the risk-spillover between stock markets. Finally, we compare the VaR which are calculated by DCC and CCC . In the empirical research, we can find out that there would be bias if we simply use constant correlation to measure the VaR of stock market, especially when the market volatility is unstable. However, the dynamic conditional correlation can overcome this problem.