Information Rigidity in USDA Production Forecasts

USDA invests significant public resources into developing its crop projection reports. These publications inform decisions across the supply chain. Several previous studies find that revisions to the department’s production and yield forecasts for major agricultural commodities are positively correlated, and conclude that they deviate from what would be observed under rational expectations, possibly due to smoothing on the part of forecasters. Yet correlated revisions may also be explained by information rigidities that cause forecasts to be infrequently or only partially updated. We apply a recently-developed test to these USDA revisions for corn, soybeans, and wheat and find no significant evidence that USDA crop production forecasts are smoothed, strategically. Rather, we show that information rigidities are the more likely culprit, due to production and yield information that is either too costly to obtain or too noisy. Our results demonstrate that data challenges—not models—are the main source of inefficiency in USDA projections, and that the Department can improve the efficiency of its forecasts by making investments that improve its access to crop data, perhaps through crop-monitoring satellite and remote sensing technology