Short-Term Forecasting

Pursuant to a congressional request, GAO examined the accuracy of the Department of Agriculture's (USDA) short-term forecasts and estimates of meat production, prices, and inventories.

GAO found that: (1) between 1983 and 1988, overall bias error was less than 3 percent for all the USDA short-term cattle, hog, and broiler production and price forecasts, and overall total error was less than 6 percent; (2) error rates were much higher and more varied when forecast accuracy was evaluated on an annual or monthly level instead of over a multiyear period; (3) droughts, increased consumer demand for broilers, and the federal payment-in-kind, dairy diversions, and dairy termination programs affected the accuracy of USDA forecasts; (4) bias errors and total errors were relatively small for the National Agricultural Statistics Service's (NASS) estimates of cattle and hog inventories and broiler production between 1983 and 1989; (5) the use of other forecasts or estimates as benchmarks for comparison and analysis required some reconstruction, since historical information for other forecasts and estimates was not always retained or accessible; (6) the American Agricultural Economics Association's production and price forecasts were more accurate than USDA forecasts of cattle production and cattle, hog, and broiler prices; (7) although benchmarks showed when improvement in forecast accuracy was needed, they did not indicate what should be done; (8) forecast errors could have implications for government policies, programs, and budget decisions; and (9) between 1983 and 1989, excess meat imports totalled 135.6 million pounds, because USDA used biased forecasts that underestimated actual production and imports.

Recommendations

Our recommendations from this work are listed below with a Contact for more information. Status will change from "In process" to "Open," "Closed - implemented," or "Closed - not implemented" based on our follow up work.

Director: Team: Phone: