Research On The Long-term Memory Of The Agricultural Products Futures

As a supplement to the spot market, futures markets can accurately reflectcommodity supply and demand relationship as well as Future changes in trend.Our country has larger population, price variations of the agricultural futures arerelated to people’s livelihood. As a branch of fractal theory, Long Memorytheory breakthrough the assumption of efficient market theory in asset pricesresearch category, providing a new perspective in asset pricing and riskmanagement. Therefore, this paper uses the nonlinear research method to analyzesugar, natural rubber, soybean meal, soybeans, cotton and corn to truly reflect thecharacteristics of agricultural futures markets.This paper begins from the efficient market hypothesis, then introduces fractal theory.After describing the definition and mechanism of long memory, this paper highlight avariety of testing and estimation methods of long memory. Second, this paper uses thecorrelation coefficient method, modified R/S method and GPH method to analyze theexistence of long memory of agricultural futures. Then, this paper uses FIEGARCHand FIGARCH models to analyze and compare the pros and cons of the two models.Through the study, we find that return series of soybean meal, soybeans and sugarbenefits do not have long memory, but their absolute fluctuation return series do.Natural rubber and cotton show strongly evidences of long memory in both returnseries and absolute return series, but corn has no evidences in either return series orabsolute return series. Relative to FIGARCH model, FIEGARCH model has betterfitting effect.