Incorporating Current Information Into Historical-Average-Based Forecasts To Improve Crop Price Basis Forecasts

Being able to accurately predict basis is critical for making marketing and management decisions. Basis forecasts can be used along with futures prices to provide cash price projections. Additionally, basis forecasts are needed to evaluate hedging opportunities. Many studies have examined factors affecting basis but few have explicitly examined the ability to forecast basis. Studies have shown basis forecasts based on simple historical averages compare favorably with more complex forecasting models. However, these studies typically have considered only a 3-year historical average for forecasting basis. This research compares practical methods of forecasting basis for wheat, soybeans, corn, and milo (grain sorghum) in Kansas. Across most of the multiple-year forecast methods considered, absolute basis forecast errors were slightly higher for the harvest forecasts than the post-harvest forecasts. Using an historical 3-year average to forecast basis for wheat and soybeans was optimal as compared to other multiple-year forecasts. For corn and milo, a 2-year average was the optimal multiple-year forecast method. Incorporating current market information, such as current nearby basis deviation from an historical average, into a harvest basis forecast improves accuracy for only the 4 weeks ahead of harvest vantage point, but improves the accuracy of post-harvest basis forecasts (24 weeks after harvest) from nearly all vantage points considered.