Predicting Pork Supplies: An Application of Multiple Forecast Encompassing

Conditional efficiency or forecast encompassing is tested among alternative pork production forecasts using the method proposed by Harvey and Newbold. One-, two-, and three-quarter ahead pork production forecasts made by the United States Department of Agriculture (USDA), the University of Illinois and Purdue University Cooperative Extension Service, and those produced by a univariate time series model are evaluated. The encompassing tests provide considerably more information about forecast performance than a simple pair-wise test for equality of mean squared errors. The results suggest that at a one-quarter horizon, the Extension service forecasts encompass the competitors, but at longer horizon, a composite forecast may provide greater accuracy.