Using a Decision Support Framework to Evaluate Forecasts

The paper examines the relationship between statistical measures of forecast accuracy and decision support for determining the most useful forecast method. A decision framework is examined with forecasts used to guide trading decisions of a hog producer, buyer and speculator in the hog market. Out‐of‐sample forecasts from various models are evaluated using mean square error and returns from trading decisions under risk‐efficiency criteria. The results emphasize the importance of evaluating forecasts within specific decision‐making situations.