A Study On The Relationship Between Volatility Of China 's Stock Index Futures Returns And Trading Volume And Position

In order to get the relationship between volatility of returns, trading volume and open interest in stock index futures market, this paper studies the sample of main contract and monthly continuing contract of CSI300index futures (2010.4.16-2013.3.12). We use660daily data to do the regression and use the rest43daily data to do the forecast.Firstly, we use GARCH model to study the influence of open interest and volume on volatility separately. For main contract, we find that the lag of volume has a positive effect on volatility of returns and that the lag of open interest has a negative effect on volatility. For monthly continuing contract, we find that neither volume nor open interest have no effect on volatility.Secondly, we use Granger Causality Test to test the relationship between them. It shows that, there exists dual-Granger causality relationship between trading volume and volatility, and there exists dual-Granger causality relationship between trading volume and open interest, open interest Granger cause volatility, but volatility does not Granger cause open interest.Considering the interrelationship among these three variables, we build Vector Auto-regression model (VAR). From the results of main contract, we find that, trading volume has a positive effect on volatility, and that open interest has a negative effect on volatility, and that there exist a strong correlation between volume and open interest. From the results of monthly continuing contract, we find that volatility is mainly influenced by itself, not by trading volume and open interest.The impulse response analysis gives us the similar conclusion of main contract and monthly continuing contract. We find that volatility influence by itself, but the effect is very short. Trading volume has a positive effect on volatility, and open interest has a negative effect on volatility. And volume and open interest affect each other strongly. But volatility has no effect on open interest and trading volume.Finally, we do the out of sample forecast and evaluate the results. We find that VAR model has a better forecast ability than GARCH model.