Author
Kastens, Terry L.
This study evaluates agricultural forecasting accuracy in an analysis of responses to the Annual Outlook Survey conducted by the American Agricultural Economics Association from 1983 through 1995. Representative extension and composite, production, and price forecasts for several commodities are constructed from the survey data. These forecasts are compared to each other and to U.S. Department of Agriculture (USDA) and futures-based forecasts. Relationships between forecast features and accuracy are examined. Generally, extension forecasts are more accurate than USDA forecasts for livestock series, but not more accurate for crops. Composite forecasts are often more accurate than either extension or USDA forecasts.
Extension marketing economists commit substantial resources to outlook and market analysis. Producers demand this information and use it to make production and marketing decisions. This study analyzes responses to a marketing survey of producers and extension marketing economists to discern similarities and differences in their perceptions regarding market timing, futures market efficiency, and risk management. Producer and extension perceptions are consistent with regard to several marketing issues, although they are not always consistent with published research results. Both producers and extension economists disagree that producers will receive a lower average price by forwarding contracting, and many do not believe hedging reduces risk and lowers expected return. Extension marketing economists rate risk reduction as a less important goal of marketing strategies than do producers.
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.