USDA Commodity Forecasts:

Inaccuracies Found May Lead to Underestimates of Budget Outlays

PEMD-91-24: Published: Aug 13, 1991. Publicly Released: Sep 20, 1991.

Additional Materials:


Kwai-Cheung Chan
(202) 512-3652


Office of Public Affairs
(202) 512-4800

Pursuant to a congressional request, GAO examined the Department of Agriculture's (USDA) long-term commodity forecasts used in the President's budget process, focusing on the: (1) accuracy of those long-term supply and utilization forecasts, referred to as baselines; and (2) effect that forecast inaccuracies could have on outlay estimates in the President's January 1990 budget submission.

GAO found that: (1) USDA baseline forecasts, particularly those made 3 to 5 years in advance, exhibited both large total error rates and consistent bias error components; (2) weather, macroeconomic factors, and program and policy assumptions contributed to baseline forecast inaccuracies; (3) USDA forecasts for the first 2 years tended to be more accurate than GAO benchmarks, but GAO benchmarks for the third through fifth years showed less bias error than the USDA forecasts; and (4) if the bias error exhibited in long-term forecasts for crop years 1981 through 1988 continued, and if the 1985 farm bill provisions had been extended, costs for the commodity programs could have been $19.5 billion higher than estimated in the President's January 1990 budget submission.

Recommendation for Executive Action

  1. Status: Closed - Implemented

    Comments: USDA has calculated some measures of forecast accuracy and compared its forecasts with benchmark alternatives. Additional results are expected by the end of 1993.

    Recommendation: The Secretary of Agriculture should specifically direct the World Agricultural Outlook Board to measure and report forecast accuracy of 5-year baselines, as well as develop and report on benchmark forecasts.

    Agency Affected: Department of Agriculture


Explore the full database of GAO's Open Recommendations »

Jul 17, 2018

Jul 2, 2018

Jun 5, 2018

Apr 18, 2018

Apr 16, 2018

Apr 10, 2018

Mar 5, 2018

Nov 30, 2017

Nov 27, 2017

Nov 21, 2017

Looking for more? Browse all our products here