Study Of Long Term And Short Term Relationship Between Agricultural Commodity Futures

Agricultural commodity futures make up a large proportion of the commodity futures market. The planting and growing of crops lead to the seasonal fluctuation of agricultural commodities. Besides used as food, there is great demand for agricultural commodities in industry. These facts result in the situation that people have the need for investment and avoiding the risk. The prices of different agricultural commodities are related since they can work as substitute for each other, also the geographic and climatic conditions contribute to this kind of relationship. Price of futures are often related when price in spot market are related by supply and demand. Thus it is very meaningful to research the price of agricultural commodity futures.Five different agricultural com-modity futures from Chicago board of trade is chosen to do the research in this article, they are oat, rice,corn, wheat and soybean, the time period is from 2005 to 2014. Long term relation and short term relation is studied.First of all, the ADF test is used to test the stationarity of the time series data; it is found that the time series data of the closing price of the futures are integrated of order one. As time series data, whether the closing prices of agricultural commodity futures have the property of stationarity is of great importance. It is quite important to test the stationarity of time series data, Dickey and Fuller proposed a useful method in 1976, which is called the DF test. By adding lags of variables to eliminate the autorelation of residuals, the ADF test is the augmentation for the DF test. Series that can pass the ADF test have the property of stationarity, which can assure the statistic property of the ordinary least square method, and assure the F test and t test reliable.Secondly, considering the time span of the chosen data, two different methods are used to test the existence of cointegration.For two unstationary time series, cointegration means that there exits a linear combination of these two series that is stationary.In this article, two methods are discussed. The first is EG two steps method proposed by Engle and Granger in 1987.The other method is proposed by Gregory and Hansen in 1996, compared with the EG two step method, the GH method takes into consideration of the possibility that there may exit structural change of the series.It is found that seven pairs of agricultural commodities pass the EG test, and eight pairs pass the GH test.Thirdly, long term equilibrium model is build for those pairs that are cointegrated. The long term equilibrium means that nothing inside the economic system can destroy this equilibrium. If at some time a variable is disturbed and deviates from its point of equilibrium, the mechanism of equilibrium will adjust the situation in the next time to make the variable move back to its point of equilibrium. For pairs that have long term equilibrium, the error correction model of order (1,1) is used to describe the unbalance in short term.Even though there is long term equilibrium relation between variables, in short term the relation between variable can be out of balance. Error correction model functions as a mechanism to correct the unbalance of a time period in the next time period. Another EG two step method is used in build the error correction model. The first step is to test the existence of cointegration. The second step is to build the error correction model by adding the residuals of step one in the model.At last, the forecast of future 48 days’closing prices is made,the predictions are compared with the real values and forecast errors are calculated.The forecast model takes into consideration of both the equilibrium of a economical system in the long run and the disequilibrium in the short time.The forecast ability of the built model is great,the errors are less than 1.5%,and the mean error is about 0.03%,the predicted values are very close to the real values.It means that the model built in this article can describe the long term and short term relationship between different agricultural commodities very well.