USDA Reports Review

Rain Makes Grain and All That

Alan Brugler
By  Alan Brugler , DTN Contributing Analyst
The Terra MODIS yield model shows some impressive clusters of 250-to-300-bushel-per-acre yields in Nebraska and Illinois. You can also pick out the flood-stressed areas in Minnesota and Iowa and the drought areas in Kansas and Missouri. (USDA NASS graphic)

USDA threw some bearish numbers at the corn market Wednesday, but pulled their punches a bit on the ending stocks side of the equation.

As you may have heard by now, USDA not only disagreed with the private analysts fading their Aug. 1 crop estimates, they said, "We'll see you one and raise you one." Because the market had been expecting a smaller number and not a larger one, prices dropped by double digits following the report, and December corn futures came within 1/2 cent of the life-of-contract low set back in July.

The national average yield estimate was increased to an all-time-record high of 181.3 bushels per acre (bpa). Let's look at the numbers in a little more detail, to better determine whether the number is likely to still be that high in January. The yield estimate was based on 9,820 farmer surveys and 4,952 objective yield plots measured by USDA personnel. USDA analysts also look at satellite sensing (crop-mapped NDVI and other techniques) to fill in holes in the data. The Terra MODIS yield model shows some impressive clusters of 250-to-300-bpa yields in Nebraska and Illinois. You can also pick out the flood-stressed areas in Minnesota and Iowa and the drought areas in Kansas and Missouri.

The objective yield is essentially ears per acre times grain weight per ear. Ears per acre as of Sept. 1 (they can go either up or down in subsequent reports based on second-ear survivability and weather-related losses) were record large in six of the 10 states where they are counted. Those six states included big acreage heavyweights like Nebraska, Iowa, Illinois and Indiana.

While the National Agricultural Statistics Service (NASS) does have some dried-down samples from the Southern harvest, they don't have much in the way of actual grain weight per ear for the September report in the Corn Belt. Thus, they compute an estimated yield from the farmer surveys and the satellite data, and divide that by the ear count. This results in an implied ear weight, which they determined to be record high for 2018.

Was the September jump to 181.3 bpa unusually large? No, not really. Since 1974, the average August-to-September yield increase -- in years where it was raised -- is 1.66%. This year's hike was 1.62%! They only raised September yields in 22 of the previous 43 years, just slightly over 51% of the time, but this is a typical hike for those years. Several trade axioms come to mind, such as "rain makes grain" and "big crops get bigger."

It is also important to note that raising yield 7 bpa in Illinois from last month increased U.S. production by 75.95 million bushels. That one change moved the national average yield from 180 bpa to 181.3 bpa. Yields were also increased 2 bpa in Nebraska, 4 bpa in Iowa, 6 bpa in Indiana and 8 bpa in Ohio.

What does this tell us about final yield? In the 22 years since 1974, where the September yield was larger than the August yield, the January final yield was 0.499% higher. That would imply a final yield of 182.20 bpa, very close to the polynomial trend yield since 1900 of 182.9 bpa.

We can also see what USDA thinks by digging down to page 42 of the Crop Production report. Their reliability table shows that the September estimate has been below the final numbers in 12 of the past 20 years. The biggest miss (or shift in reality) has been 845 million bushels, with the smallest 14 million. Their root mean square error is 3% over those 20 years, or 5.3% for a 90% confidence interval.

Our bottom line from this exercise is that the final corn yield is highly likely to be within 5% of Wednesday's number, with a bias that the final number in January is another bushel per acre larger based on other years where the September yield was above the August estimate.

Things like Hurricane Florence or an early freeze could still put a dent in the crop. There is immature corn in pockets that were too wet this spring, but much of the crop is ahead of normal maturity and thus less vulnerable. At least half of the North Carolina corn crop had been picked ahead of the storm (43% through Sunday and aggressive efforts this week), but the other part is still in harm's way.

The good news is that the world corn stocks-to-use ratio is very tight, and the tightest since 1973 for coarse grains. That means a strong export environment to soak up these extra bushels. USDA boosted projected exports another 50 million bushels Wednesday. This report (and particularly the selloff after it came out) is also a warning shot to producers considering making wholesale shifts of acreage from soybean into corn for 2019.

Alan Brugler may be reached at


Alan Brugler