Growing Season Forecasts of Harvest Time Corn Prices

A procedure is developed for forecasting the probability distribution of harvest-time average corn futures prices. Forecasts are based on a harvest corn futures price model. Conditional forecasts are made at monthly intervals during the 1992 to 1994 corn growing seasons using the harvest corn futures price model with measures of model estimation error, preharvest corn supply and demand estimates, and historic variances and covariances relative to harvest-time estimates together in a Monte Carlo multivariate forecasting procedure. The accuracy and distributional properties of forecasts from this forecasting procedure are compared with alternative options-based and futures forecasts for the 1992 to 1994 period. Results indicate positive skewness in distributions from both the Monte Carlo forecast procedure and the options-based forecasts. There was also a strong tendency for forecast distributions from both sources to become tighter as the growing season progressed with one notable exception during mid-August 1993. The price model forecasts compare favorably with options-based forecasts and with futures prices when analyzing the harvest price forecast accuracy of mid-June and mid-July projections for the 1992 to 1994 period. The significance of this research lies in the development of a procedure for integrating supply and demand forecast uncertainty throughout the corn growing season with a corn price model to derive price forecast probability distributions. Potential limitations include forecast model inaccuracy and incorrect measures of model explanatory variable forecast variances and covariances. If further improvements can be made in accuracy, these findings imply that this procedure has potential to improve the probabilistic content of preharvest grain price forecasts relative to alternative options and futures market-based price forecast sources.