If markets were efficient, futures prices would be unbiased predictors of future spot prices and a simple prediction model would suffice, but markets are not efficient and such predictions cannot be accurately made. Concerns for inaccurate commodity price predictions spurred this research. Traditional futures forecasts are based on past data from those same markets. Forex futures prices might be better predictors for commodity futures. The purpose of this study was to identify a possible alternative to traditional commodity futures price prediction by finding and using a parsimonious formula that connects Forex futures to commodity futures prices in an equation that conforms to the traditions of Gann, Dow, Fibonacci, and Elliot. An experimental quantitative research design was followed. Archived daily data from the futures market for commodities futures and foreign exchange futures were obtained. From these data, a time-series prediction model was developed to predict wheat prices using U.S. dollar 1 yr. T-Note (TY01) foreign exchange historical records. The model was fitted with a time-series general regression neural networks to predict commodity prices. Statistical t test was used to compare the pattern of prediction to actual prices. Prediction error was only 4.42%, suggesting a well-fitting model. The use of such a model has the potential to stabilize commodity market predictions, which in turn would bring social change that could affect the agribusiness community in planning, planting, banking, financing, buying, selling, and warehousing. The commodity futures market could become more efficient as well. This new future prices information could lead to the betterment of social conditions through better understanding of financial markets.
展开▼
