Kub's Den

It's been a weird year for a lot of row-crop producers so far -- too dry, too wet, too cold, too hot, or in a few select spots, just perfect. It all depends on where you are; but the overall production expectations for U.S. corn and soybeans in 2022 seems like it's going to be worse than the original trendline expectations. USDA is going to take a stab at projecting just how much worse in Monday's upcoming World Agricultural Supply and Demand Estimates (WASDE) report (Sept. 12), which is always tricky to do before combines are rolling in the heart of the Corn Belt, but particularly tricky in early September of a year when large swaths of the crops got such a late start.

The thing that interests me -- on a day when new-crop corn prices are vaulting up by double digits while new-crop soybean prices are simultaneously falling down by double digits -- is how the damage may play out for one crop more severely than the other.

In conversations with agronomists, researchers and highly observant farmers over the years, I've picked up on a widespread belief that soybeans are more dependable than corn, yield-wise, year over year. We're all just comparing anecdotal evidence, but it's always felt right to me, too, speaking as someone who farms in pretty challenging conditions in the northwestern Corn Belt. If things get planted late, I can still rely on soybeans to get something close to their average yield. If the summer is dry, I feel like I can still rely on soybean fields to fill trucks at harvest, even if the corn suffered and ended up as an insurance claim.

However, this may be just a feeling, and it may not be true everywhere.

With a little statistical sorcery, I figured I could look at the past several decades of yield data for both corn and soybeans in the U.S., conclude that, yes, corn yields are more volatile (having a higher standard deviation) than soybean yields, and -- presto! -- write up the findings and file the column with the editors.

It turned out to be only partly true. On a nationwide basis, yes, corn yields look slightly more volatile than soybean yields. In any given year, we "should" expect to see the nationwide corn yield 5.6% either higher or lower than the original trendline projection, and we "should" expect to see the nationwide soybean yield only 4.6% higher or lower than its original trendline projection.

But nationwide numbers are the product of averaging a lot of averages over very diverse, widespread geographies, so it's hard to see that result and feel confident it would play out that way on any one farm. I decided to drill down into county-by-county yield data for the past 50 years since 1971. The results surprised me: in some of the challenging dryland spots (eastern South Dakota, southern Ohio), where I expected soybeans to shine with their steady, stable predictability, it was actually corn that had less expected variation than soybeans. To be sure, these "challenging" areas had much more volatile yields for both crops than the nationwide data, which is only to be expected once the dulling effect of averaging across a continent-wide geography has been taken away. And interestingly, the privileged counties with reliable irrigation-fed fields had noticeably lower yield volatility than the nationwide data. However, it wasn't universally true that soybean yields are less volatile everywhere than corn yields.

Some of the selected sample data is shown below. For instance, since 1971, corn yields have, on average, tended to be 9.8 bushels off from each year's original trendline projection (either higher or lower). To compare that apples-to-apples to soybeans, in percentage terms, a 9.8-bushel expected variation is 5.6% of 2022's projected trendline yield of 177 bushels per acre (bpa). A 2.3-bushel average expected variation in nationwide soybean yields is 4.6% of 2022's projected trendline yield of 50 bpa.

U.S. corn yield expected variation: 5.6% off trendline

U.S. soybean yield expected variation: 4.6% off trendline

Pickaway County, Ohio, corn yield expected variation: 9.6%

Pickaway County, Ohio, soybean yield expected variation: 10.5%

Beadle County, South Dakota, corn yield expected variation: 13%

Beadle County, South Dakota, soybean yield expected variation: 16.6%

Phelps County, Nebraska, (mostly irrigated pivots) corn yield expected variation: 4.9%

Phelps County, Nebraska, (mostly irrigated pivots) soybean yield expected variation: 4.1%

I have one big caveat to point out -- many of the counties I tried sampling were missing years of yield data because not enough farmers filled out and returned USDA's yield surveys. It might seem like an annoying bit of paperwork, but shared, reliable, public data really is important when it comes to things like this -- having helpful, statistically significant conclusions that improve all our understanding about how crops and markets work. So, the next time you see one in the mailbox or email inbox, please fill out your surveys!

Even with limited sampling, I think it's a fair historical observation to say that irrigation noticeably cuts down on the year-to-year risk of wild yields in comparison to those at the mercy of rainfed dryland farming. It may also still be true that seed companies and university researchers in controlled conditions can observe less volatility in soybean yields than in corn yields. The real-world historical yield data, however, at least from some places, doesn't always validate the prevailing gut feelings about soybeans being a more "reliable" crop than corn.

Now we shall see which one takes the bigger hit, in percentage terms, by the time this wild growing season comes to a close

**

Comments above are for educational purposes only and are not meant as specific trade recommendations. The buying and selling of grain or grain futures or options involve substantial risk and are not suitable for everyone.

Elaine Kub, CFA is the author of "Mastering the Grain Markets: How Profits Are Really Made" and can be reached at masteringthegrainmarkets@gmail.com or on Twitter @elainekub.

(c) Copyright 2022 DTN, LLC. All rights reserved.