The Empirical Research On The Correlation Of Copper Future Between SHFE And LME Future Markets

The futures market is faced with much more risk due to the speculative transactions, the continuity of the price changes and margin leverage compared with the spot market. The research on the correlation of copper prices and the yield fluctuations in China's Shanghai Futures Exchange (SHFE) and London Metal Exchange (LME) will help investors and hedgers establish their own risk Management strategy on a global scale, and will help the government draw the development policies of domestic futures market by learning the foreign markets .This paper first introduces the basic theory and empirical methods of price correlation research, and then chooses the methods needed, in the end, we do some empirical analysis after choosing and dealing with our data. The empirical analysis is implemented as follow: firstly, to test the existence of false return, if not exist, we will determine a long-term balanced relationship and short-term fluctuations deviated from the balanced, and ascertain the adjust force; at last, to abtain the degree of short-term dynamic correlation more exactly, and to apply the dynamic correlation coefficients and series of fluctuations into the use of value at risk (VaR) method to get the maximum of losses in order to afford datas support for the participant for risk management.The empirical results show that it only exists unilateral Granger causality before 2001, but mutual Granger causality after that, i.e. there doesn't exist false return. And the co-integration test comes to the conclusion that there exists long-term equilibrium between their yields, meanwhile the ECM model concludes that when the short-term fluctuations depart from the long-term equilibrium, the adjust force will act more stronger to bring it back to the equilibrium after 2001 than before. Finally, with the help of Dynamic Conditional Correlation Multivariate GARCH (DCC-MGARCH) model, we educe the short-term dynamic correlation coefficients series, apply the datas to the VaR method to test the validity of the model, and we have a good result, so the VaR from this model is appropriate to margin management and risk control.