Working paper, second draft, October 2026
Download
Abstract
I propose a novel method to estimate agricultural supply news shocks using the U.S. Department of Agriculture’s (USDA) crop-specific yield forecasts, which are revised monthly over the agricultural season, kept confidential until release, and driven largely by weather. I orthogonalize these yield revisions with respect to their own lags and lagged real crop prices, to purge the predictable direction of revisions between consecutive months within a season and any expected component already priced into commodity prices. The resulting instrument has a robust first-stage F-statistic above 30. Compared with Adjemian and Jo (2025), the orthogonalized revision shocks have a much more muted impact on real activity and a delayed but significant effect on food and core prices.