The Research On The Dynamic Correlation Between Stock Market And Fund Market In China Based On Mixed Frequency Data

With the continuous development of the financial market integration,the different financial markets have a more and more closer linkage.The stock market and fund market are the two important parts of the financial market,so investigate the stock-fund dynamic correlation is urgently required.This is not only beneficial for investors to allocate resources efficiently and manage risk,but also help the authorities to understand the market condition and develop appropriate policies.This paper mainly investigates three aspects of the dynamic correlation between the stock market and fund market.The first is about the stock-fund dynamic correlation investigation using the same frequency data;the second is using the mixed data to extract the long-run and short-run dynamic correlation,and studying the impact of macroeconomic on the long-run dynamic correlation;the third is about the long-run dynamic correlation forecast via mixed data.Firstly,this paper uses DCC-GARCH model to investigate the dynamic correlation via same data sampling,and investigates dynamic spillover effect between the stock market and fund market.The empirical results show that the stock-fund dynamic correlation is quite high,the fluctuation range is between 80% and 90%,and volatility is quite small.This indicates the stock market and fund market have strong correlation and stable linkage.Meanwhile the empirical results show that the significant dynamic spillover effect does not exist between the stock market and fund market.Then,this paper investigates the impact of macroeconomic on the dynamic correlation between the stock market and fund market via mixed data sampling,and uses DCC-MIDAS model to extract the long-run and short-run dynamic correlation.On this basis,we directly incorporate the macroeconomic variables in the DCC-MIDAS model to study the impact of macroeconomic on the dynamic correlation.The empirical results show that the DCC-MIDAS model based on the RV can extract the long-run and short-run dynamic correlation between the stock market and fund market well.And residents consumption price index,money supply growth rate as well as interest rate affect the stock-fund long-run dynamic correlation significantly.Finally,this paper uses ADL-MIDAS model and AR model to predict the long-run dynamic correlation between the stock market and fund market separately.ADL-MIDAS model uses the mixed data of daily data and monthly data,while AR model just uses low frequency data of monthly data.We estimate the two models and compare their in-sample precision of prediction using mean square forecast error.The empirical results show that the ADL-MIDAS model's precision of prediction is better than AR model's.Then we use the ADL-MIDAS model's nowcasting method to predict the stock-fund long-run dynamic correlation.The prediction results show that the predicted value of long-run dynamic correlation is more accurate after adopted the nowcasting method.