Down, Then Up: Examining USDA’s Corn Yield Reversals

Executive Summary

In October 2026, the U.S. Department of Agriculture (USDA) increased its projected national corn yield by 2.7 bushels per acre to 181.2, reversing a 2.2-bushel reduction in September and exceeding its August estimate by 0.5 bushels per acre. This reversal, and the subsequent increase in overall corn supply, had substantial impacts on the corn market, raising questions amongst farmers about the accuracy of the report. This has been an unfortunate, continuous trend amongst farmers.

In order to put the October World Agricultural Supply and Demand Estimates (WASDE) report into historical context, this analysis determines how frequently similar reversals to projected corn yields have occurred and what happened in subsequent reports.

An analysis of monthly reports from the USDA WASDE from 2000 through October 2026 found that seven years, or 25.9% of the period, experienced the same three-month reversal in which the USDA lowered its yield estimate before subsequently raising it above the original projection. The 2026 movement had a cumulative absolute revision of 4.9 bushels per acre, the third largest among qualifying reversals.

More specifically, only three previous years - 2001, 2003 and 2017 - experienced the same August-September-October reversal observed in 2026. In all three instances, the USDA further increased its projected corn yield in November.

In 10 of the 15 historical years with qualifying or partial reversals, USDA's January yield estimate exceeded the last month of the reversal, indicating that further upward revisions were common.

While these historical observations provide context for the 2026 revisions, the limited number of comparable occurrences does not establish a reliable prediction for November. Rather, the findings demonstrate that substantial changes in USDA yield projections, including reversals of previous estimates, are relatively uncommon but not unprecedented.

Introduction: USDA's 2026 Yield Reversal

On October 9, 2026, the monthly USDA WASDE report increased the projected 2026/27 corn crop yield by 2.7 bushels per acre, bringing the estimate to 181.2, which contributed to a subsequent decrease in corn prices. Following the report, the December corn futures contract dropped by 30 cents, before closing the day at $4.80, down nearly 22 cents from the October 8 close. The October estimate exceeded the USDA's August estimate by 0.5 bushels per acre, despite the September report lowering USDA's projections by 2.2 bushels per acre to 178.5.

This reversal left many farmers frustrated with the USDA's projections and caught the overall corn market by surprise, as many market participants and commentaries expected the USDA to decrease corn yield. As such, the natural question to ask is how frequently has the USDA lowered its national corn yield estimate in one month, only to increase it the following month above the original projection? Then, how did this reversal compare to the post-harvest January report? This paper analyzes national corn yield estimates from the monthly WASDE reports since 2000 and USDA revisions following previous reversals similar to that which occurred in October 2026.

Data and Methodology

The data consists of monthly national corn yield estimates published in USDA's WASDE reports from 2000 through October 2026. Since USDA begins publishing projections for the upcoming crop in May of each year, this dataset only includes May-December for each calendar year, and January for the years in which a qualifying or partial reversal occurred. The reports for November and December 2026 and January 2027 have yet to be published. Further, October reports were unavailable for 2013 and 2025 due to government shutdowns.

Historical WASDE reports are publicly available in PDF and Excel format. Specific year-by-month compilations are available for the years 2010-2020 but are not available for other years. Thus, this dataset was independently compiled from monthly WASDE reports.

The primary analysis identifies years in which, over a three-month span, the USDA lowers its yield estimates from Month 1 to Month 2, then raises yield estimates from Month 2 to Month 3 above the Month 1 yield estimates. Such an occurrence is referred to in this paper as a qualifying reversal. A qualifying reversal must occur in any three consecutive reporting months, and Month 3 must exceed Month 1. However, the qualifying reversal may occur during any consecutive three-month window from May through December.

Partial reversals are also demonstrated in this analysis for additional context. Partial reversals occur in any three consecutive months in which the USDA lowers its yield estimates from Month 1 to Month 2, then raises yield estimates from Month 2 to Month 3, but Month 3 does not exceed the yield estimates in Month 1.

This analysis also examines subsequent monthly WASDE reports to determine how USDA's national corn yield estimates changed following qualifying movements.

Cumulative absolute revisions (CAR) measure the sum of two absolute monthly revisions to corn yield in bushels per acre, regardless of direction of each revision in order to give context to the 2026 movements.

The USDA's National Agricultural Statistics Service (NASS) develops the in-season corn yield forecasts which are then published in the monthly Crop Production report, which is released simultaneously with the WASDE report. These estimates are then incorporated into the WASDE supply-and-demand projections.

Corn yield was chosen as the variable as it was the data point cited as the largest contributor to the increased corn supply.

Historical Comparisons: How Unusual is 2026?

Table 1 shows the May-December USDA WASDE estimates from 2000-2026. The bottom row shows 2026.

The years with rows highlighted in yellow are years in which a qualifying reversal occurred. From 2000-2026, this occurred seven times, or 25.93% of the time. In these instances, the average CAR was 4.2, meaning the average absolute change in yield in consecutive reports between Month 1 and Month 3 was 4.2 bushels per acre. The 4.9 CAR in 2026 is the third largest CAR, falling just behind 2003 (5.1) and 2008 (7.1).

The years with rows highlighted in green are years in which there was a partial reversal. This occurred 11 times, with an average CAR of 4.58, the largest of which occurred in 2002. In 2008, both a qualifying reversal and a partial reversal occurred. The qualifying reversal occurred from June-August with movement of 148.9-148.4-155 while the partial reversal occurred from August-October in a movement of 155-152.3-154. Likewise, 2021 saw two partial reversals from July-September and September-November. The partial reversal in 2008 and both partial reversals in 2021 are included in the partial reversal CAR.

The most common months for which a qualifying movement occurs (yellow highlight) is August-October, accounting for four of the occurrences. September-November occurred twice, with June-August only once.

What Happened Next?

Table 2 highlights the months preceding and succeeding the qualifying and partial reversals highlighted in Table 1 that extend the reversals.

The orange highlights identify yield estimates immediately succeeding the three-month movements that exacerbate the CARs for that particular year. In nine of the 15 historical years prior to 2026, the USDA further increased its projected corn yield in the next reported month following a qualifying or partial reversal. In 2013, October's estimate was unavailable due to a government shutdown, although November's estimate increased from September.

Additionally, in 10 of the 16 years with highlighted movements, including 2026, the estimate immediately preceding Month 1 was higher than the Month 1 estimate. These findings indicate that some yield revisions extended beyond the original three-month period, although subsequent revisions did not consistently move in the same direction.

There are three instances prior to 2026 in which a qualifying reversal occurred in the August-October time frame - 2001, 2003 and 2017. In all three instances, the USDA increased its estimate again in November to the largest projection of the crop year. However, of the 24 years in which an October report was available, the November report increased corn yields 13 times, approximately 54.2%.

What Happened in January After a Reversal?

Table 3 shows what the subsequent January "final" yield for each year with a qualifying or partial reversal. Each report is from January of the year following a harvest. For example, when looking at the final outcome of the 2001 harvest, the January report in 2002 is used.

The cells highlighted in red indicate years in which the January yield was higher than Month 3. The cells highlighted in blue indicate years in which the January yield was lower than Month 1. Cells unhighlighted in the January column indicate those in which the January yield was higher than Month 1 but less than or equal to Month 3.

The January yield exceeded the Month 3 yield in 10 of the 15 years in which there was a qualifying or partial reversal. The January yield exceeded the Month 1 yield in nine of the years in which a reversal occurred.

There are two instances that are slightly more complicated. In 2008, two reversals occurred - the June-August qualifying reversal and the August-October partial reversal. In this instance, the January yield was larger than Month 1 for the qualifying reversal, but below the Month 1 yield for the partial reversal. Similarly, 2021 had two partial reversals from July-September and September-November. The January yield was below the Month 1 yield for the first reversal, and above the Month 1 yield for the second reversal.

Meanwhile, five of the 15 years saw yields decreased or remain unchanged from Month 3. The January report was less than Month 1 yield in only two years, not including 2008 and 2021's dual reversals.

This analysis establishes that upward revisions occurred from Month 3 in the majority of the historical years in which a reversal occurred. That does not necessarily mean that a reversal causes an upward January revision. These increases also did not always restore the estimate to its original level.

Implications and Conclusion

The USDA's October 2026 corn yield revision was a relatively uncommon occurrence. However, similar reversals have occurred more than one in four years since 2000, and three times in the last decade. Since 2000, seven years have experienced a similar three-month movement in which USDA lowered its national corn yield estimate before subsequently increasing it above the original projection. The cumulative absolute revision of 4.9 bushels per acre in 2026 was the third largest among the occurrences.

Perhaps more interestingly, the three previous years that experienced the same August-September-October movement as 2026 - 2001, 2003 and 2017 - all saw additional increases in November. In all three instances, the November estimate was the largest of the USDA's projections for the year. The October 2026 estimate remains 1.8 bushels per acre below July, raising questions about whether a similar pattern could occur in November. However, with only three historical comparisons, these findings should not be interpreted as evidence that another upward revision is likely.

Further, in 9 of the 15 years in which a reversal occurred, the January report increased yields above the Month 1 yield and exceeded the Month 3 yield in 10 of the 15 years. This implies that not only is an increased yield in the January 2027 report possible, but historically it has happened more often than not. However, this does not imply a statistically established forecasting relationship as discussed in previous analysis on corn production by this author.

The scope of this analysis is not to conclude that the presence of historical reversals establishes that estimates were inaccurate or improperly constructed. This analysis aims to provide historical context to the reversal of the October WASDE. Further, the historical analysis demonstrates why a single monthly estimate should be understood as provisional information that can meaningfully change as the year progresses.

However, while it's beyond the scope of this paper, the author acknowledges that farmer frustration is quite present after the most recent report. Previous analysis by the author has shown that USDA's estimates are not only persistently inaccurate compared to the January report but have gotten worse in recent years. As such, this analysis has a couple of potential caveats that need to be taken into account.

1. Farmer frustration is misplaced and the USDA's reports are accurate.
2. The fall reports, and thus the reversals highlighted in this paper, are inaccurate while the January report is accurate.
3. The fall reports and the January report are inaccurate.

While this analysis does not investigate those possibilities, they must be highlighted in order to provide proper context to the frustrations shared by many farmers.