Dynamic Correlation Study

Using the monthly data of the index of 15 industries,which may catch the fundamental of China's economy,from Jan,2003 to Sept,2008,This paper gets to study the conditional dynamic correlation between the rate of return of industry index and the rate of return of market portifolio,by adding stochastic time varying beta and stochastic idiosyncratic volatility on the Fama & French's three-factor model.This paper focuses on the time series characteristics of the dynamic correlation,and its response to market and non-market shocks,and to the change of systemmatic volatility or idiosyncratic volatility.Different from previous literatures which use GARCH type model to study this subject,my model contains stochastic term,thus can incorporate the impact of new coming information;and my model can distinguish idiosyncratic volatility from systemmatic volatility.I use MCMC method through the free software WinBUGS to get the estimation.What I found includes:a),market volatility has"Leverage Effect"to positive market shocks.which is different from western study but consistent with the data used in this paper,while the response of idiosyncratic volatility to non-market shocks are different between different industries.b),The correlation of industries with the market portfolio contains two trends mainly:the first main trend tends to decrease due to the increasing trend of idiosyncratic volatility,and the second trends tends to dance with some other stuffs,for example,the market shocks.c),correlation responds asymmetricly to good and bad market shocks or non-market shocks,this may be induced by the"Leverage effect"of the market volatility or idiosyncratic volatility to different shocks.d),The increase of systemmatic volatility or idiosyncratic volatility tends to decrease the correlation between return of industry index and the market index...