By: Taylor, Mykel R.; Dhuyvetter, Kevin C.; Kastens, Terry L.
This research compares practical methods of forecasting basis, using current market information for wheat, soybeans, corn, and milo (grain sorghum) in Kansas. Though generally not statistically superior, an historical one-year average was optimal for corn, milo, and soybean harvest and post-harvest basis forecasts. A one-year average was also best for wheat post-harvest basis forecasts, whereas a five-year average was the best method for forecasting wheat harvest basis. Incorporating current market information, defined as basis deviation from historical average, improved the accuracy of post-harvest basis forecasts. A naive forecast incorporating current information was often the most accurate for post-harvest basis forecasts.