摘要：The forecasting accuracy of five competing naïve and futures-based localized cash price forecasts is determined. The third-week's price each month from 1987-96 is forecasted from several vantage points. Commodities examine include those relevant to Midwest producers: the major grains, slaughter steers, slaughter hogs, several classes of feeder cattle, cull cows, and sows. Relative forecasting accuracy across forecast method is compared using regression models of forecast error. The traditional forecast method deferred futures plus historical basis has the greatest accuracy- even for cull cows. Adding complexity to forecasts, such as including regression models to capture nonlinear bases or biases in futures markets, does not improve accuracy.
