Crop Root Zone™

Five of the eight nutrients DTN tracks were priced below the prior month in the second full week of August 2026, and only one of them moved enough to matter by the service's own standard: UAN28 fell 7% to an average $446 per ton (DTN/Progressive Farmer, Aug 19, 2026). That was the headline on the print and on the trade relays that carried it a week later. It is also, read against the rest of the same table, a misleading thing to lead with — because after a 7% decline UAN28 is still the most expensive nitrogen a Corn Belt grower can buy at retail, and the decline moved it $10.79 an acre against a standing gap of $48.34.

1. The board, converted to the unit the decision actually uses

A fertilizer price per ton is not a nitrogen price. The eight-product retail table is quoted in dollars per ton of material, and four of those products are nitrogen sources carrying very different concentrations: urea is 46% N, anhydrous ammonia 82%, UAN28 28%, UAN32 32%. Dividing price by nutrient content converts the invoice into the number that drives a rate decision.

Source: DTN/Progressive Farmer retail fertilizer survey, week of Aug 10–14, 2026, published Aug 19, 2026; per-pound figures reproduced independently from the $/ton prices and standard nutrient contents.

The derivation reproduces all four published per-pound figures exactly to the cent, which is worth stating rather than assuming: it means the price table and the per-pound table in that survey are internally consistent, and that a reader can extend the same arithmetic to any rate without waiting for the service to publish it.

The ranking that falls out is the point. UAN28, the product that just fell 7%, is the dearest nitrogen on the board. Anhydrous, the product that fell only slightly, is the cheapest by a wide margin — and it is also the product up 27% year over year, which is the fact most commentary has fixed on.

Ground Truth: A percentage change and a competitive position are different claims, and August's retail coverage collapsed them. UAN28's 7% decline is real, and it is news about where UAN28 started, not about where it now sits. Nothing in the August board changed the order of the four nitrogen sources by cost per pound of N — it has been anhydrous, UAN32, urea, UAN28 throughout.

2. What the gap is worth on an acre

Rank order is cheap to assert. The size of the gap is the part that decides anything, so it has to be carried to a rate. At 180 lb N per acre — a common Corn Belt corn rate, used here as a stated basis rather than a recommendation — the four sources price out as follows.

Source: derived from the DTN/Progressive Farmer board of Aug 10–14, 2026. Bars scaled to the longest.

$37.55/acre

The nitrogen-cost premium for running UAN28 instead of anhydrous ammonia at a 180-pound rate on the August 2026 retail board — a 35.5% premium on the nitrogen line. (Derived from DTN/Progressive Farmer, Aug 19, 2026)

Set the 7% decline against that. Backing out the prior month from the reported change puts UAN28 near $479.57 per ton, or $0.8564 per pound of N. The fall to $0.7964 is worth $10.79 an acre at 180 pounds. The gap to anhydrous was $48.34 an acre before the move and is $37.55 after it. The largest single move on the August board closed 22.3% of the standing gap between the most and least expensive nitrogen — a real improvement, and not remotely enough to change a sourcing decision.

The comparison inside the UAN family is sharper still, because it inverts. UAN28 at $446 is $12 per ton cheaper than UAN32 at $458. Per pound of nitrogen it is 8.1 cents more expensive, which at 180 pounds is $14.54 an acre. A buyer comparing delivered ton prices across two grades of the same solution gets the answer exactly backwards, and the error is larger than the entire month-over-month move that made the headlines.

3. The premium is not irrational, and pricing it is the useful exercise

None of the above says anhydrous is the right buy. It says the premium is $37.55 an acre and that a grower paying it should know what it purchases. Three things do most of the work.

Equipment and handling. Anhydrous ammonia is applied as a pressurised liquid that flashes to gas, requires a knife toolbar and nurse tanks, and is regulated as a hazardous material with corresponding storage, transport and personal-protective requirements. UAN is a non-pressurised solution that moves through the same equipment a grower already owns for herbicide — a sprayer, a floater, or Y-drops.

Application window. Anhydrous is largely a pre-plant, fall, or early sidedress product, and fall application carries a soil-temperature constraint. UAN is the practical vehicle for in-season and split applications, which is a different agronomic strategy rather than a different price for the same strategy.

Retail availability. In much of the eastern Corn Belt the constraint is not that a grower declines to own a toolbar; it is that the local retailer's fleet, nurse-tank inventory and custom-application calendar determine what can be delivered in the window the grower needs.

Scaled up, the premium is a real capital number: on 1,000 corn acres it is $37,550 a year on the nitrogen line alone. That is inside the range of a used pull-type anhydrous applicator, which is why the ownership question is worth asking rather than assuming — while acknowledging the honest counter, which is that most growers who use anhydrous do not own the tank fleet and are buying an application service, not a machine.

Ground Truth: Price the convenience rather than arguing about it. The right question is not "is UAN worth it" but "is it worth $37.55 an acre this year," and the answer moves with the board. That figure was $48.34 a month ago. A grower who has split-applied UAN for a decade on the reasoning that the premium is small should recheck the number annually, because the premium is a function of two prices that have not moved together — anhydrous is up 27% year over year and UAN32 is down 6%.

4. The finding the year-over-year column contains and nobody printed

The survey reports each product's change against the prior year: UAN28 +6%, UAN32 −6%. Two grades of the same nitrogen solution, 12 percentage points apart in twelve months. Backing out year-ago prices from those changes and converting both to a nutrient basis produces a result neither week's coverage carried.

Source: derived from DTN/Progressive Farmer year-over-year changes as published Aug 19, 2026. Year-ago prices are back-calculated from rounded percentages and are estimates — see the error bound below.

A year ago UAN32 was the more expensive of the two solutions per pound of nitrogen. Today it is the cheaper, by 8.1 cents. The order inverted, and the swing is 9.08 cents per pound of N — $16.34 an acre at 180 pounds.

That figure rests on back-calculation, so it needs an error bound rather than a caveat. If the reported year-over-year percentages are rounded to the nearest whole point, the true year-ago values sit inside a band of roughly 0.7 cents per pound of N for UAN28 and 0.8 cents for UAN32. The swing being measured is 9.08 cents — an order of magnitude larger than the combined rounding band, so the inversion is not an artifact of rounding. It would survive an error four times larger than the one the rounding can produce.

What it is not is explained. UAN28 and UAN32 are the same urea-ammonium nitrate solution at different water dilutions, made in the same plants and moved through the same terminals. A 12-point divergence in twelve months between them is not a nitrogen story — nitrogen is common to both. It points at something grade-specific: regional inventory carried at one concentration, freight economics that favour shipping the more concentrated solution when the water is the freight, or a demand mix that shifted between the two. This publication cannot distinguish those from the retail board alone, and does not.

5. What to watch, and what the board cannot tell you

The next DTN retail print is due in the first week of September, and the board carried here is dated Aug 10–14 — meaning that at the time of writing the most current public retail nitrogen table is more than a fortnight old, and was republished by trade relays on Aug 26 without new underlying data. Anyone treating the late-August relay as a fresh print is reading a two-week-old board.

Three things are worth tracking into fall buying:

1. Whether the UAN28–UAN32 inversion persists or closes. If it is a dilution-freight artifact it should decay; if it holds through a second month it is a demand-mix signal and changes which grade a retailer should be quoting.
2. Whether anhydrous's 27% year-over-year gain flows into fall prepay. The gap analysed above is a retail gap, and fall prepay is where it gets locked.
3. The month-over-month column's declining information content. Five of eight products "lower" with one significant mover is a table in which most cells are noise, and the year-over-year column is where the structure is.

Ground Truth: The retail board's own significance threshold is a percentage test on ton price, which means it flags a 5% move identically whether it is worth $5.29 an acre on anhydrous or $7.17 on UAN28. It is proportionate, not wrong — but it systematically surfaces the products with the highest cost per pound of nitrogen, because a fixed percentage of a larger per-acre bill is a larger number. A reader who takes "UAN28 leads declines" as a buy signal has let a threshold designed to detect movement stand in for an analysis of level.

References

1. DTN/Progressive Farmer, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," Aug 19, 2026 — retail survey for the week of Aug 10–14, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
2. Fertilizer Daily, "U.S. retail fertilizer prices fall for fourth straight week as UAN28 drops 7%," Aug 26, 2026 — trade relay of the same DTN survey week. https://www.fertilizerdaily.com/20260826-us-retail-fertilizer-prices-mid-august-2026/
3. DTN/Progressive Farmer, "Fertilizer Prices Fall as UAN32 Leads Significant Drops Heading Into August," Aug 5, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/05/fertilizer-prices-fall-uan32-leads
4. DTN/Progressive Farmer, "Fertilizers Start August With Mixed Prices as 5 Nutrients Decline, Led by UAN32," Aug 12, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/12/fertilizers-start-august-mixed-5-led

Nutrient contents used throughout (urea 46% N, anhydrous ammonia 82% N, UAN28 28% N, UAN32 32% N) are standard product specifications. All per-acre figures are at a stated 180 lb N/acre basis and are arithmetic, not survey data. Year-ago prices in section 4 are back-calculated from rounded published percentages and are labelled as derived wherever they appear.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Retail phosphate is quoted in dollars per ton of product, and almost nothing a grower decides is denominated that way. DTN's latest published survey puts DAP at $917/ton and MAP at $960/ton — a $43 gap that reads, on the invoice, as DAP being the cheaper phosphate (DTN/Progressive Farmer, Aug 19, 2026). Convert both to the nutrient actually being bought and the gap does not narrow; it inverts. Convert them a second time, crediting the nitrogen that arrives in the same bag, and it inverts again. The two conversions do not merely change the size of the answer. They change which product wins, twice, and the number that decides it is not a phosphate price at all.

1. The board, as published

What's new: DTN's most recent published retail survey covers the week of August 10–14, 2026 and prices all eight major nutrients. DAP and MAP were both slightly higher month over month, and both are well above year-ago levels — DAP +11%, MAP +7% (DTN/Progressive Farmer, Aug 19, 2026).

Evidence: The full board, with the nutrient grades that matter for the arithmetic below.

Source: DTN/Progressive Farmer retail fertilizer survey, week of Aug 10–14, 2026, published Aug 19, 2026.

A note on vintage before anything is built on these numbers. DTN publishes weekly, and a later print dated August 26 is understood to exist. It could not be retrieved from any public-tier source in preparing this piece — the article index returns the August 19 print as the newest retrievable fertilizer article. Every figure here therefore rests on the August 19 print, which is verified, and the arithmetic is presented so that a reader holding a fresher board can re-run it in two minutes.

Ground Truth: A price per ton is a price for a truck, not a price for a nutrient, and the two rank differently often enough that the distinction is not academic. Grade is the conversion factor and it is printed on the bag. That the industry still quotes and compares phosphates per ton — and that the per-ton gap points the opposite way to the per-nutrient gap this week — means the headline comparison in circulation right now is not merely imprecise, it has the sign wrong.

2. Per pound of phosphate, MAP is cheaper

What's new: DAP is 18-46-0 and MAP is 11-52-0. A ton of DAP carries 920 pounds of P₂O₅; a ton of MAP carries 1,040. MAP costs 4.7% more per ton and delivers 13.0% more phosphate in it.

Evidence: The naive unit costs, computed directly from the board above.

Source: Crop Root Zone calculation from DTN/Progressive Farmer retail prices, Aug 19, 2026, and standard product grades. Bars scaled to the higher unit cost.

MAP delivers phosphate 7.4% cheaper than DAP does, on the same day, on the same survey, at a $43/ton premium. At a 150 lb P₂O₅/acre rate — a common corn-soybean replacement level — that is $138.47/acre through MAP against $149.51/acre through DAP, a $11.04/acre advantage to the product with the higher sticker price.

Ground Truth: This is where most of the analysis that gets as far as converting to nutrient stops, and stopping here is worse than not converting at all. The first conversion produces a confident, arithmetically correct, and — as the next section shows — wrong answer. A partial correction that reverses the ranking is more dangerous than the raw per-ton comparison, because it arrives with the authority of having done the math.

3. The nitrogen is not free, and pricing it flips the ranking back

What's new: DAP is 18% nitrogen and MAP is 11% nitrogen. A ton of DAP carries 360 lb N; a ton of MAP carries 220 lb N. That nitrogen is applied to the field and taken up by the crop exactly as nitrogen from any other source is. If it is not credited, section 2 is charging the entire cost of both products against phosphate alone while quietly ignoring that one of them delivers 64% more of a second nutrient the grower is buying anyway.

Evidence: Credit the co-product nitrogen at the cheapest nitrogen on the board — anhydrous, at $0.59/lb N (DTN/Progressive Farmer, Aug 19, 2026) — and recompute the residual cost of the phosphate.

Source: Crop Root Zone calculation. Product prices and the anhydrous nitrogen unit cost from DTN/Progressive Farmer, Aug 19, 2026; grades are standard.

DAP is now the cheaper phosphate, by $0.0324/lb P₂O₅ — a 4.1% advantage running the opposite direction to section 2's 7.4% advantage for MAP. The ranking has reversed twice in three tables, and no price has changed.

$0.4098/lb N

The nitrogen price at which DAP and MAP cost exactly the same per pound of phosphate. Above it DAP is cheaper; below it MAP is. Every nitrogen source on this week's retail board is above it. (Crop Root Zone calculation from DTN/Progressive Farmer prices, Aug 19, 2026)

4. Where the crossover actually sits, and why it is not close

What's new: The reversal in section 3 is not a property of $0.59/lb N specifically. It is a property of the nitrogen price being above a threshold, and that threshold can be solved for exactly.

Evidence: Let x be the nitrogen credit in dollars per pound. Setting the two net phosphate unit costs equal:

(917 − 360x) ÷ 920 = (960 − 220x) ÷ 1,040

Cross-multiplying gives 1,040(917 − 360x) = 920(960 − 220x), or 953,680 − 374,400x = 883,200 − 202,400x. Collecting terms, 70,480 = 172,000x, so x = $0.4098/lb N.

Now compare that crossover against every nitrogen price the same survey publishes.

Source: Nitrogen unit costs as published by DTN/Progressive Farmer, Aug 19, 2026. Margin column is a Crop Root Zone calculation.

The cheapest nitrogen available is 44% above the crossover. The dearest is 95% above it. There is no nitrogen source on the retail board — and, given anhydrous is structurally the cheapest N form and has been all year, no plausible nitrogen source at all — that would put MAP back on top. The conclusion is not marginal and it is not sensitive to which nitrogen the grower actually intends to buy.

Ground Truth: The DAP-versus-MAP decision is a nitrogen decision wearing a phosphate label. What determines the answer is not the phosphate content, not the $43 sticker gap, and not this week's phosphate market — it is whether nitrogen is worth more or less than about forty-one cents a pound. Nitrogen has not been near forty-one cents in this market in years; anhydrous at +27% year over year is nearly half again above it. So the practical form of this finding is a standing rule rather than a weekly call: in any nitrogen market resembling the current one, DAP is the cheaper phosphate, and the per-ton and per-unit comparisons that say otherwise are both incomplete in the same way. The rule inverts only in a genuinely cheap-nitrogen world, which is also the world in which the phosphate decision matters least.

5. The size of the prize, and why it probably does not decide the purchase

What's new: Having established the direction firmly, the magnitude deserves the same treatment — and it is small.

Evidence: At the $0.59/lb N credit, DAP's advantage is $0.0324/lb P₂O₅. Applied at a 150 lb P₂O₅/acre rate:

Source: Crop Root Zone calculation from the section 3 net unit costs. Figures are the phosphate-attributable cost only, after the nitrogen credit.

Against a delivered phosphate bill running $115–153/acre, a $4.86 advantage at the typical rate is roughly 4% of the line. That is real money at scale and it is also comfortably inside the spread between two retailers quoting the same product in the same county, inside the freight differential on a longer haul, and inside the rounding on most blend sheets.

Ground Truth: The honest conclusion is a two-part one and the second part matters more than the first. DAP is the cheaper phosphate at every nitrogen price currently quoted, and the margin is small enough that it should almost never be the reason to change suppliers, re-spec a blend, or take a longer haul. What the arithmetic is genuinely good for is the reverse test: if a retailer's MAP quote is more than about $5/acre cheaper than their DAP quote at the same P₂O₅ rate, something other than nutrient content is driving it — a distressed lot, a freight advantage, a blend-compatibility constraint, or a grade that is not the standard 11-52-0 — and that is worth asking about directly. The unit-cost math here is most valuable not as a purchasing rule but as a detector for when the quoted spread is telling you about something the label does not say.

6. Two limits worth stating

The nitrogen credit assumes the co-product nitrogen genuinely substitutes for nitrogen the grower would otherwise buy. In a fall phosphate application ahead of corn that is a fair assumption. Ahead of soybeans, where the nitrogen requirement is largely met by fixation, the credit is worth substantially less than $0.59/lb and can approach zero — in which case section 2's ranking stands and MAP is the cheaper phosphate. The crossover arithmetic does not change; the applicable nitrogen price does. Growers applying phosphate ahead of soybeans should run the same equation with their own credit and will likely land below $0.4098.

Second, both prices are delivered retail averages from a national survey, and the survey's own week is August 10–14 with a later print not publicly retrievable. Regional basis on phosphate has been wide this year, and the entire DAP advantage established here is smaller than the week-to-week move in either product has been on several occasions since spring. The direction is robust because the crossover is 44% below the cheapest nitrogen; the $4.86 is not robust and should be re-derived from a local quote before it is relied on.

References

1. DTN/Progressive Farmer, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," retail fertilizer survey for the week of Aug 10–14, 2026, published Aug 19, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
2. Product grades (DAP 18-46-0, MAP 11-52-0, urea 46-0-0, anhydrous ammonia 82-0-0, UAN28, UAN32) are standard industry specifications.
3. Crop Root Zone, TRZ-0121, "World Urea Fell 61% Since April. US Retail Urea Fell 18%. Less Than a Third Got Through," Aug 24, 2026 — for the retail pass-through discipline applied to the vintage caveat in section 1.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

The directive had a date on it, and the date is this month. On December 11, 2025 China's National Development and Reform Commission suspended orderly exports of phosphate fertilizers "in principle until August 2026," with the stated aim of concentrating resources on domestic supply ahead of the 2026 spring planting season (SunSirs, Dec 16, 2025). Nine months later the US retail phosphate board sits 11% above year-ago on DAP and 7% on MAP (DTN/Progressive Farmer, Aug 19, 2026), and the supply event that a great deal of that premium was written against is scheduled to expire. What follows is an attempt to size what the expiry is actually worth at a US farmgate, and the answer is considerably smaller than the headline arithmetic suggests — for a reason that has nothing to do with China.

1. What was actually suspended, and how much tonnage it was

What's new: The measure is a suspension "in principle," not a statutory ban, issued by an economic planning body rather than a customs authority. That distinction matters for how it ends: an in-principle administrative suspension can lapse, be quietly extended, or be replaced by a quota without any formal repeal, and none of those three produces a announcement of the kind a market can trade against.

Evidence: Chinese phosphate export volumes had already contracted sharply before the suspension was issued.

Source: SunSirs, Dec 16, 2025, reporting Chinese customs export volumes for January–September 2025. Bars scaled to DAP tonnage.

Domestic Chinese prices moved in the direction the policy intended. MAP (55% granular) was around 3,600 yuan/tonne in May 2025, up about 30%, and DAP around 4,100 yuan/tonne in December, up about 28% (SunSirs, Dec 16, 2025).

Ground Truth: The export decline of roughly 20–24% happened during the first nine months of 2025 — that is, before the December directive existed. The suspension did not cause the contraction it is credited with; it codified one already well underway, and it did so after domestic prices had already risen 28–30%. That sequencing matters for the expiry: a policy that formalised an existing commercial trend will not, on lapsing, reverse a trend it never created. The tonnage came off the export market because Chinese domestic demand outbid it, and the domestic price is still the higher one. Expiry removes a legal constraint; it does not by itself supply a commercial reason to ship.

2. The concentration that makes this a headline at all

What's new: China, Morocco, Russia and Saudi Arabia together account for approximately 80% of global DAP and MAP export trade.

Evidence: That concentration is what converts a single national policy into a world price event. It is also the standing structural fact behind phosphate's persistent premium — the IEA's broader work on mineral supply concentration puts the average top-refining-country share across critical minerals at 72% and describes export restrictions as the mechanism by which concentration becomes a realised risk rather than a latent one (IEA, Global Critical Minerals Outlook 2026).

Sources: export share and volumes, SunSirs, Dec 16, 2025; US retail, DTN/Progressive Farmer, Aug 19, 2026.

Ground Truth: There is a detail inside the US numbers that the concentration story does not predict and should be uncomfortable for it. DAP is up 11% year over year and MAP only 7%, a four-point spread between two products made from the same rock, the same acid and the same ammonia, in the same plants, moving through the same restricted export channel. If the dominant driver were a supply shock at the export source, the two should move close to together. A persistent four-point wedge between them points at something downstream and product-specific — differing domestic blend demand, differing freight, or differing inventory positions at the retail level — which is to say a meaningful part of the US premium is not the Chinese story at all.

3. What a world-price move is worth at a US farmgate

What's new: The critical link is transmission, and it is not one-to-one. This desk measured it directly earlier this month on the nitrogen complex over matched endpoints: world urea fell 61% from April while US retail urea fell 18% — a pass-through of 29.5% (Crop Root Zone, TRZ-0121, Aug 24, 2026).

Evidence: Applying that measured transmission rate to phosphate gives a bounded estimate of what a complete reversal of the 2026 gain would be worth at retail.

Source: Crop Root Zone calculation. Pass-through rate of 29.5% measured on the nitrogen complex over matched endpoints in TRZ-0121 (Aug 24, 2026) and applied here to phosphate as a labelled estimate — see limits. Base DAP price $917/ton, DTN/Progressive Farmer, Aug 19, 2026.

Even the most aggressive case — the entire year-on-year DAP gain reversing at world level, which would require far more than a suspension lapsing — moves the US retail board about $30/ton. On a 150 lb P₂O₅/acre program that is $4.85/acre.

$4.85/acre

What a complete reversal of DAP's entire 11% year-on-year gain is worth to a US grower at a 150 lb P₂O₅ rate, after the 29.5% retail pass-through measured on this desk's own nitrogen work. (Crop Root Zone calculation, from DTN retail prices Aug 19, 2026 and TRZ-0121's measured transmission)

Ground Truth: The number that decides a US phosphate bill is not the Chinese export policy, and it is not the world price. It is the pass-through coefficient, and at roughly 0.3 it absorbs about seven-tenths of whatever happens upstream before it reaches a farm. A grower reading the expiry headline and deferring a purchase in the expectation of relief is making a bet that requires the world price to fall roughly three and a half times as far as the relief they want. That is a demanding condition, and it is demanding in a specific and testable way rather than merely a cautious one.

4. The comparison that puts it in proportion

What's new: The same DTN survey that carries the 11% year-on-year DAP gain also carries a $43/ton spread between DAP and MAP — two products that substitute for each other in most phosphate programs.

Evidence: Set the international story against the choices already available on the board this week.

Source: Crop Root Zone calculation from DTN/Progressive Farmer, Aug 19, 2026. The DAP-vs-MAP after-credit figure is derived in TRZ-0135 (Aug 28, 2026).

Ground Truth: The trade-policy question is the more interesting one to read about and the less useful one to act on, and the ranking is the finding. A supply event with no date certain, transmitted at roughly thirty cents on the dollar, competes for attention against a product-selection decision that is fully determined by numbers already published and available today. The asymmetry runs the wrong way relative to where attention actually goes. The defensible position on the expiry is therefore not a forecast but a discipline: treat it as a reason to keep phosphate purchasing flexible rather than as a reason to defer it, because the deferral only pays in a scenario that requires a large world move and an unusually generous pass-through, and this desk has measured the pass-through and it is not generous.

5. What would actually change the answer

What's new: Two observables would move this analysis materially, and neither is the expiry announcement itself.

Evidence: The first is a Chinese domestic-versus-export price inversion. Exports resumed at scale only if the export netback beats the domestic price, and domestic DAP was around 4,100 yuan/tonne and rising as of the directive. Until the relationship inverts, a lapsed suspension is a permission that nobody exercises.

The second is the DAP-MAP spread in section 2. If that four-point year-on-year wedge closes, it indicates the US premium was in fact export-driven and is unwinding as one. If it persists or widens through the expiry window, it confirms the domestic, product-specific reading and means US phosphate buyers should stop watching Chinese policy for their price signal.

Source: Crop Root Zone analysis. Underlying spread from DTN/Progressive Farmer, Aug 19, 2026; Chinese domestic price from SunSirs, Dec 16, 2025.

Ground Truth: The DAP-MAP spread is the better instrument here precisely because nobody is treating it as one. It is published weekly, it is free, it requires no view on Chinese policy, and it separates the two competing explanations for the US premium cleanly — an export-driven premium cannot easily produce a durable four-point wedge between two products that share a supply chain, while a domestic inventory or blend-demand story produces exactly that. A reader who wants to know whether the expiry mattered will learn it faster from two columns of a US retail survey than from any announcement out of Beijing.

6. Limits

The pass-through coefficient is the load-bearing assumption and it is transplanted. The 29.5% figure was measured on the nitrogen complex over matched endpoints, not on phosphate, and the two complexes differ in ways that plausibly affect transmission: phosphate retail inventory turns more slowly than nitrogen, the US is a larger net importer of phosphate than of nitrogen relative to consumption, and phosphate lacks the domestic ammonia production that damps nitrogen transmission. The direction of the bias is not obvious — slower inventory turnover argues for lower pass-through, higher import dependence for higher — so the figure is used as a labelled central estimate rather than a measured phosphate coefficient, and the scenario table should be read as an order of magnitude. A phosphate-specific pass-through measurement over matched endpoints is the obvious next piece of work and this desk has not done it.

Second, the suspension's status as of publication is taken from the December 2025 directive as reported. No confirmation of expiry, extension, or replacement by a quota was retrievable from a public-tier source in preparing this piece, and readers should not infer from its absence that the suspension has lapsed. "In principle until August 2026" is the language on the record; what happens at the boundary is not yet on it.

Third, the Chinese domestic price figures are May and December 2025 vintage. They establish the direction of the domestic-versus-export incentive at the time the policy was set, not its current level, and the netback test in section 5 requires current data on both legs that this piece does not have.

References

1. SunSirs, "The Logic Behind China's Phosphate Fertilizer Export Suspension," Dec 16, 2025 — NDRC directive of Dec 11, 2025, export volumes for Jan–Sep 2025, and Chinese domestic price levels. https://www.sunsirs.com/commodity-news/petail-28842.html
2. DTN/Progressive Farmer, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," retail fertilizer survey for the week of Aug 10–14, 2026, published Aug 19, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
3. International Energy Agency, Global Critical Minerals Outlook 2026, Executive Summary — supply concentration and export restrictions as a general mechanism. https://www.iea.org/reports/global-critical-minerals-outlook-2026/executive-summary
4. Crop Root Zone, TRZ-0121, "World Urea Fell 61% Since April. US Retail Urea Fell 18%. Less Than a Third Got Through," Aug 24, 2026 — the 29.5% pass-through measurement.
5. Crop Root Zone, TRZ-0135, "MAP Costs $43 More Per Ton Than DAP and Is Cheaper Per Pound of Phosphate," Aug 28, 2026 — the after-credit DAP/MAP unit-cost comparison.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Twice this week the international urea benchmark printed exactly the same number: $417.50 per tonne on Monday, August 24, and $417.50 per tonne again on Wednesday, August 26 (Trading Economics, Aug 26, 2026). Nothing about the level changed. But the year-on-year change reported beside it moved from −4.46% to +0.36% — 4.82 percentage points, across zero, on a market that did not move. A buyer who checks the yearly column to decide whether nitrogen is cheaper than it was a year ago would have gotten one answer on Monday and the opposite answer on Wednesday, from an identical price.

1. The same price, two different years

What's new: The August 26 reading of the Trading Economics urea series shows $417.50/t, up $3.50 (+0.85%) on the session, with a monthly change of −2.91% and a yearly change of +0.36%. This desk recorded the same series on August 24 at $417.50/t, up $8.50 (+2.08%), monthly −2.91%, yearly −4.46% (Crop Root Zone, "Urea Is Up 8% in Ten Days," Aug 25, 2026). The monthly figure is identical between the two readings. The yearly figure is not.

Evidence: The published percentages are enough to recover the levels they are computed against. If the current price is $417.50 and the yearly change is −4.46%, the year-ago reference is $417.50 ÷ 0.9554 = $436.99. If the current price is $417.50 and the yearly change is +0.36%, the year-ago reference is $417.50 ÷ 1.0036 = $416.00.

Source: Trading Economics urea series, readings of Aug 24 and Aug 26, 2026. Implied bases are Crop Root Zone calculations, obtained by dividing the quoted price by (1 + the quoted change).

The month-ago base did not move at all between the two readings — $430.02 on both days. The year-ago base fell $20.99, or 4.80%. That is the whole of the 4.82-point swing, and it happened in a window that closed twelve months ago.

Ground Truth: A year-on-year change that crosses zero while the level is unchanged is not a statement about this year's market. It is a statement about the base period rolling out of the window, and it is arithmetically guaranteed to happen whenever last year's price was falling faster than this year's. Any purchasing rule keyed to "is nitrogen up or down year over year" flipped its answer this week on a market that did nothing. The number to read is the level and the sequence of levels; the trailing percentage is a comparison to a date the reader almost never checks.

2. Where the 4.82 points came from

What's new: The inferred bases say something specific and checkable about last year: between roughly August 24 and August 26, 2025, this benchmark fell about $21 per tonne, or 4.8%, in two sessions.

Evidence: That inference follows directly from the table above — it is the only way an unchanged $417.50 can produce two different yearly percentages. It is worth stating as an inference rather than an observation, because this desk did not read the 2025 series directly; it recovered the 2025 levels from 2026 percentages. The arithmetic is exact, but it inherits any rounding in the published percentages. At the two-decimal precision quoted, the year-ago base is recoverable to within roughly ±$0.50.

Source: Crop Root Zone calculation from the Trading Economics readings of Aug 24 and Aug 26, 2026. Bars scaled so the largest absolute move is ten blocks.

4.82 points

The move in urea's reported year-on-year change across two sessions in which the price was $417.50 on both. (Crop Root Zone calculation from Trading Economics readings, Aug 24 and Aug 26, 2026)

The practical consequence is a timing one. Because the 2025 autumn decline is only now rolling into the twelve-month window, the year-on-year statistic will keep rising through September even if the price never moves again. That is a mechanical prediction, and it is falsifiable: if the level holds near $417.50 into late September and the yearly change stalls or turns back down, then the 2025 base was not declining the way this arithmetic implies, and the inference in this section is wrong.

3. What it does to the benchmark-to-retail ladder

What's new: None of this changes what a grower pays. The most recent US retail print available to this desk remains DTN's survey of the week of August 10–14, published August 19: urea $678/ton, stated at $0.74/lb N (DTN/Progressive Farmer, Aug 19, 2026).

Evidence: Putting the benchmark and the retail print on one ruler requires converting both to dollars per pound of nitrogen. Urea is 46% N. A metric tonne is 2,204.62 lb; a short ton is 2,000 lb.

Source: Trading Economics urea series (Aug 14, Aug 24, Aug 26, 2026); DTN/Progressive Farmer retail survey published Aug 19, 2026. Conversions are Crop Root Zone calculations at 46% N.

The $0.73696 figure reproduces DTN's own published $0.74/lb N, which is the arithmetic control on this table: if our conversion could not recover the number the source prints, the rest of the ladder would not be trustworthy either.

The multiple of retail over benchmark therefore reads 1.936x against the August 14 benchmark and 1.790x against the August 24 and August 26 benchmark — a compression of the multiple with no move in retail at all. This desk reported that compression from the August 24 reading (Crop Root Zone, "Urea Is Up 8% in Ten Days," Aug 25, 2026); the August 26 reading leaves it exactly where it was, because the benchmark did not move.

Sizing the pass-through matters more than the multiple. The benchmark rose $31.50/t from August 14 to August 26, which is $0.03105/lb N. At the 29.5% farmgate pass-through this desk measured earlier this month (Crop Root Zone, "World Urea Fell 61% Since April," Aug 24, 2026), that is $0.00916/lb N, or at 180 lb N/acre $1.65 per acre. Inside the noise of a single weekly print.

There is a further reason the benchmark move should not carry a fall-fill decision on its own: urea is not the cheapest nitrogen on the retail board, and has not been all month. The same August 19 DTN survey prices the four nitrogen sources per pound of N as follows.

Source: DTN/Progressive Farmer retail survey published Aug 19, 2026, covering the week of Aug 10–14. The $/lb N and percentage columns are Crop Root Zone calculations; they reproduce DTN's own published $0.59, $0.72, $0.74 and $0.80 to the cent. Bars scaled so the highest $/lb N is ten blocks.

A benchmark move worth $1.65 an acre at measured pass-through sits inside a product spread worth $26.85 an acre at 180 lb N — the anhydrous-to-urea gap of $0.1492/lb N. (This desk put the same wedge at $26.90 on August 25; the five-cent difference is rounding, not a revision.) The form decision is an order of magnitude larger than the benchmark decision, and it is made on the farm rather than in an international tender.

Ground Truth: The fall-fill decision is not helped by either percentage on that page. The monthly reads −2.91% and the yearly now reads +0.36%, and the honest answer is that the level rose 8.2% from August 14 and has stopped. What a buyer is actually choosing between is the cost of tons already in the shed and the cost of replacing them, and on the retail side neither number has moved yet. The benchmark move is worth about $1.65 an acre at measured pass-through — real, but not a reason to change a purchase date.

4. The forecast on the same page runs the other way

What's new: The same August 26 page that reports a positive yearly change also carries Trading Economics' own model forecasts: $402.60/t at quarter-end and $353.05/t on a twelve-month view (Trading Economics, Aug 26, 2026).

Evidence: Both sit below the current level.

Source: Trading Economics urea series, Aug 26, 2026. Percentage columns are Crop Root Zone calculations.

So one page simultaneously reports that urea is 0.36% more expensive than a year ago and projects that it will be 15.44% cheaper a year from now. These are not contradictory — one is a trailing comparison and the other is a model output — but they are routinely quoted as if they were the same kind of fact. They are not, and a fall-fill argument built on the trailing number is building on the weaker of the two.

This desk does not adopt the forecast. It is carried here specifically because it runs against the reading a positive yearly change would suggest, and a number that cuts against your own piece belongs in the piece.

5. Limits, and what to check next

Four things constrain this read.

The Aug 26 US retail print could not be independently sourced this run. DTN published a retail survey on August 26 covering the week of August 17–21, and this desk could not retrieve it from any public-tier source before filing. Every retail figure above is therefore the August 19 print, clearly dated as such. If the August 26 print moved retail materially, the multiple in Section 3 is stale by one week — the benchmark arithmetic is unaffected.

The 2025 base levels are inferred, not observed. Section 2's claim that urea fell ~$21 in two sessions in late August 2025 is recovered from 2026 percentages, not read off a 2025 series. It is exact arithmetic on rounded inputs.

Basis. The benchmark is an international price per metric tonne; the retail print is delivered US retail per short ton. The gap between them includes ocean freight, import handling, inland distribution, storage and retail margin. The 1.79x multiple is not a markup and should not be read as one.

Two sessions is not a trend. The level rose 8.2% from August 14 to August 24 and then stopped. Whether the stop is a pause or a top is not something two flat prints can answer.

The first checkable test is the next DTN weekly retail survey, which will show whether any part of the benchmark's August move has begun to reach the farmgate. The second is the yearly percentage itself: on the reasoning in Section 2 it should keep climbing into late September on base roll alone, with the level flat.

References

1. Trading Economics, urea commodity series — reading of Aug 26, 2026. $417.50/t, +$3.50 (+0.85%), monthly −2.91%, yearly +0.36%; quarter-end forecast $402.60, 12-month forecast $353.05. https://tradingeconomics.com/commodity/urea
2. Trading Economics, urea commodity series — reading of Aug 24, 2026, as recorded in Crop Root Zone, "Urea Is Up 8% in Ten Days. Its Monthly and Yearly Percentages Both Still Read Negative," Aug 25, 2026. $417.50/t, +$8.50 (+2.08%), monthly −2.91%, yearly −4.46%.
3. Trading Economics, urea commodity series — reading of Aug 14, 2026: $386.00/t, as recorded in Crop Root Zone, Aug 25, 2026.
4. DTN/Progressive Farmer, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," published Aug 19, 2026, covering the survey week of Aug 10–14, 2026. Urea $678/ton, $0.74/lb N; anhydrous $964/ton, $0.59/lb N; UAN28 $446/ton; UAN32 $458/ton; DAP $917; MAP $960; potash $495; 10-34-0 $718. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
5. Crop Root Zone, "World Urea Fell 61% Since April. US Retail Urea Fell 18%. Less Than a Third Got Through," Aug 24, 2026 — the measured 29.5% pass-through used in Section 3.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

The international urea benchmark closed Monday, August 24 at $417.50 per tonne, up $8.50 on the session, a gain of 2.08% (Trading Economics, Aug 24, 2026). Ten days earlier the same series stood at $386.00. That is a move of $31.50, or 8.2%, and it lands in the week that India's 1.7-million-tonne RCF tender stopped accepting offers. On the same page, on the same day, the monthly change for that series reads −2.91% and the yearly change reads −4.46%. Every one of those numbers is arithmetically correct. Two of them describe a market that stopped existing on the fourteenth.

1. The ten-day turn, in levels

What's new: The benchmark bottomed and reversed inside a single fortnight, and it did so with the largest single-session gain in the recent series on the day the tender's offer-validity window closed.

Evidence: Three dated readings on the same benchmark bracket the move. The August 17 weekly print of $390.00 was reported at the time as up 0.52% on the week and down 7.69% on the month (iGrow News, week ending Aug 17, 2026) — a market still being described as falling. One week later the level is $417.50.

Source: Trading Economics urea series, Apr 15, Aug 14 and Aug 24, 2026; iGrow News weekly fertilizer update, week ending Aug 17, 2026. April peak stated as approximate — the series is quoted above $700/t in that window without a single published settlement we could verify to the dollar.

The scale of the round trip is easier to see against the peak than against the trough:

Source: Trading Economics urea series. The July 24 figure is a Crop Root Zone calculation, derived below.

Ground Truth: The bars for late July and late August are the same height, and that is the whole problem with reading this market on percentages. A buyer who checked the board a month ago and checked it again yesterday would conclude nothing had happened. What actually happened in between was a 10% collapse and an 8% recovery, and the fall-fill quote a retailer writes is set by where the market is on the day they replace the ton, not by the two endpoints a monthly statistic happens to connect.

2. Reconstructing what the rolling bases hide

What's new: The monthly and yearly percentages can be inverted to recover the levels they are measured against, and doing so shows the market a month ago was higher than it is today — which is why the monthly figure reads negative through a rally.

Evidence: If today's level is $417.50 and the monthly change is −2.91%, the base is $417.50 ÷ 0.9709 = $430.02. If the yearly change is −4.46%, the year-ago base is $417.50 ÷ 0.9554 = $436.99.

Source: Trading Economics urea series, Aug 24, 2026. Implied base levels are Crop Root Zone calculations from the published percentages; the ten-day figure is computed from the Aug 14 and Aug 24 levels directly.

The July 24 level of $430.02 is the useful one. It says the market fell roughly 10.2% from late July to August 14 and has since recovered 8.2%, and that the monthly statistic nets those two moves into a single small negative. Neither the size nor the direction of the path survives the netting.

This is the same measurement problem this desk documented from the other side in TRZ-0121, where matched endpoints were used deliberately so that a world price and a farmgate price could be compared over the same five months. The rule generalises: on a series this volatile, read the level and the sequence. A percentage is a statement about two dates, and the two dates are almost never the two that matter.

3. What the tender did and did not establish

What's new: The RCF tender's offer-validity window expired on August 24 (Fertilizer Daily, Aug 7, 2026). The benchmark's largest recent up-day is the same date.

Evidence: The tender sought 1.7 million tonnes, split 1 Mt west coast and 700,000 t east coast, with bids opened August 11 and all shipments to complete by September 24. On the west-coast tranche RCF drew roughly 3.1 million tonnes of offers, at a range of $393.65 to $435 per tonne CFR (Bloomberg, Aug 13, 2026). Profercy has separately characterised the tender as lifting global FOB values to four-year highs and sized the prospective award near $2.64 billion.

This desk argued on August 24 that 3.1x oversubscription on one coast was the durable fact of that tender and that the headline low offer was the least informative number in it. The ten days since are a test of that read, and they cut against the simpler interpretation: a tender that drew three times the tonnes it asked for was followed by an 8% rise in the benchmark, not a fall.

Ground Truth: Oversubscription measures how many sellers wanted the business, not how much product exists. Those are different quantities and they come apart exactly when a single buyer is taking a large share of a month's exportable tonnes off the water — 1.7 Mt inside a thirty-day shipment window is roughly 56,700 t/day of loadings. The offers were plentiful because the cargo was attractive; the price rose afterwards because the cargo is now committed. A grower reading "three times oversubscribed" as "supply is abundant" drew the opposite conclusion from the one the market drew, and the market has had ten days to be wrong and has not been.

4. What it does to the farmgate, and when

What's new: The wholesale-to-retail multiple has compressed without retail moving, purely because the wholesale leg rose.

Evidence: Converting both sides to a single ruler — dollars per pound of nitrogen, at 46% N, 2,000 lb/short ton and 2,204.62 lb/tonne — gives:

Source: DTN/Progressive Farmer retail survey published Aug 19, 2026, for the week of Aug 10–14; Trading Economics urea series, Aug 14 and Aug 24, 2026. The $/lb N column is a Crop Root Zone calculation; the retail figure reproduces DTN's own published $0.74/lb N, which is the arithmetic control on the conversion.

The multiple of retail to benchmark was 1.94x on August 14 and is 1.79x on August 24. Retail did not move — the last print available is still the week of August 10–14, and the next one lands Wednesday, August 26.

1.79x

US retail urea as a multiple of the international benchmark, in dollars per pound of nitrogen, on August 24 — down from 1.94x ten days earlier with no change in the retail leg. (Crop Root Zone calculation from DTN, Aug 19, 2026 and Trading Economics, Aug 14 and Aug 24, 2026)

What this is worth per acre is smaller than the percentage suggests, and the reason is the pass-through ratio this desk measured last week. Over the five months from April 1 to August 19, the international market fell 60.6% and US retail urea fell 17.9% — a pass-through of 29.5%. Applying that ratio to the current move: the wholesale gain of $31.50/t is $28.58/st, and 180 lb of N delivered as urea is 0.1957 short tons, so a full pass-through would be $5.59 per acre and a 29.5% pass-through would be $1.65 per acre. Both numbers are inside the noise of a single retail print.

The asymmetry is the part worth holding onto. The same inventory lag that is about to keep this rally out of the retail board for weeks is the lag that kept the 61% collapse out of it for five months. A grower cannot have the slow decline without the slow recovery; they are the same mechanism seen twice.

5. What changes for a fall buyer, and what does not

The distinction that matters this week is between a retailer's inventory cost and their replacement cost. The ton in the warehouse was bought at a price that is now history and no wholesale print can change it. The ton the retailer must buy to refill behind a fall-fill sale is priced off today's board — and today's board is $31.50 higher than it was on the fourteenth.

Fall-fill quotes are written against replacement, not against inventory, which is why a wholesale turn shows up in a fall offer sheet faster than it shows up in a spot retail survey. The practical consequence is narrow and dated: an offer taken this week is being written against a market that has recovered 8%, and an offer taken three weeks ago was not.

Limits, and they are real. The benchmark used here is a single published international series, and it is not the price at which any specific cargo changed hands; the tender offers were CFR and carry freight that an FOB series does not, so a portion of any gap between them is geography rather than price. Trading Economics' own forecast for the series is $402.60 for the current quarter and $353.05 on a twelve-month view, both below today's level — that is a model output, not a market, and it is included here because it runs against the direction of this article rather than with it. The retail leg is a survey of a range, not a transaction price. And ten days is a short window: one more week at this level would make the turn a trend, and one week back at $390 would make it a spike. Wednesday's DTN print is the first checkable test of whether any of it has begun to move downstream, and on the pass-through evidence the honest expectation is that it has not.

References

1. Trading Economics, urea commodity series — readings of Aug 14, Aug 21, Aug 24, 2026, with monthly, yearly and forecast figures as published. https://tradingeconomics.com/commodity/urea
2. iGrow News, "Fertilizer Prices Weekly Update," week ending Aug 17, 2026. https://igrownews.com/fertilizer-prices-weekly-update/
3. Fertilizer Daily, "India issues 1.7 million tonne urea tender through RCF with bids opening on August 11," Aug 7, 2026. https://www.fertilizerdaily.com/20260807-india-rcf-urea-tender-1-7-million-tonnes-august-2026/
4. Bloomberg, India RCF urea tender offer range, Aug 13, 2026.
5. Profercy, "India eyes $2.64bn urea award as tender lifts global fob values to four year highs," Aug 2026. https://www.profercy.com/insights/india-eyes-major-urea-award-as-tender-lifts-global-fob-values
6. DTN/Progressive Farmer, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," published Aug 19, 2026, covering the retail week of Aug 10–14. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
7. Crop Root Zone, TRZ-0120 and TRZ-0121, Aug 24, 2026 — tender dispersion and the 29.5% matched-endpoint pass-through ratio.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

The University of Illinois farmdoc team published its 2027 planting outlook on Aug 4, 2026, concluding that soybeans will be more profitable than corn next year, with elevated nitrogen prices doing much of the work (Schnitkey, Paulson, Zulauf and Zwilling, farmdoc daily, Aug 4, 2026). The finding travelled as a forecast. Read against the same group's own historical series, published two months earlier, it is something less dramatic and more useful: the thirteenth iteration of an established regime, and one whose persistence — not its direction — is the part that has not been explained.

1. Two regimes, thirteen years each, almost exactly mirrored

The companion farmdoc analysis of historical Illinois returns (Paulson, Schnitkey and Zulauf, June 2, 2026) splits the last quarter-century cleanly in two.

Source: farmdoc daily, "Historical Corn versus Soybean Returns in Illinois," Jun 2, 2026 (Paulson, Schnitkey, Zulauf).

The symmetry is the striking part and it does not appear to have been remarked on. Two consecutive thirteen-year windows, each won ten times out of thirteen, with average advantages six dollars an acre apart — a 10% difference in magnitude and an exact match in duration and hit rate. Whatever produced the flip did not merely tilt the comparison; it reflected it.

The first regime has a well-understood cause. The Renewable Fuel Standard came into force in 2005 and expanded in 2007, and the resulting corn demand is the standard explanation for the biofuel-era corn advantage. The second regime's drivers are more diffuse: the farmdoc authors name lower soybean non-land costs — fertilizer and machinery in particular — soybean yields exceeding trend in every year but 2019 since 2012, and biodiesel demand pulling soybean oil.

The exceptions prove informative rather than annoying. Corn won 2020, 2021 and 2022 — the pandemic-assistance and commodity-spike years. In other words the corn-favouring years inside the soybean regime were years in which an external transfer or a price shock, not the underlying cost structure, decided the comparison.

Ground Truth: A 2027 projection that soybeans out-earn corn carries almost no information, because the same statement was true in ten of the previous thirteen years. What would be informative is the opposite claim. Treat "soybeans win again" as the null hypothesis for the 2027 budget and reserve attention for evidence that corn is the exception — which, on this record, has required either a demand shock or a government payment.

2. The magnitudes are small relative to the noise around them

The 2024 figures make the scale concrete. On high-productivity Central Illinois farmland, operator and land return was $281 an acre for corn and $322 for soybeans — a corn-minus-soybean return of −$41 an acre.

−$41/acre

The 2024 corn-minus-soybean operator and land return on high-productivity Central Illinois farmland — about 4% of corn's gross revenue at plateau prices, and roughly one-eighth of either crop's return. (farmdoc daily, Aug 4, 2026)

Against a corn gross of roughly $1,035 an acre at the long-run plateau price of $4.50 and a 230-bushel yield, $41 is under 4% of revenue. The regime average of $53 is about 5%. These are not large signals. They are persistent ones, which is a different property and is why they show up in a thirteen-year average and not in any individual grower's recollection of a season.

The one genuinely large observation in the series runs the other way on interpretation. In 2023 soybeans out-returned corn by $237 an acre — the widest gap in the record — and both crops returned negative numbers that year. A record-wide spread in a year when neither crop paid is a statement about relative loss, not relative profit, and it should not be read as evidence that soybeans were a good year.

3. Acreage responded, but only partly — and the reason is a number

If soybeans have out-earned corn in ten of thirteen years, the obvious question is why Illinois still plants a very large corn acre. Part of the answer is that acreage did respond: the farmdoc authors note Illinois corn acreage has come down proportionally across all regions from its early-2010s peak, which they read as a rational response, and they credit the 1996 farm bill's planting flexibility for making the response possible at all.

But the response was partial, and the reason it was partial is quantifiable. Dropping corn out of the rotation does not simply substitute one crop's economics for another's — it imposes a yield penalty on the soybeans themselves.

The largest empirical study of this effect analysed 748,374 field observations across Nebraska, South Dakota, Minnesota, Iowa, Illinois and Indiana for 2007–2012 (Seifert, Roberts and Lobell, Agronomy Journal 109:541–548, 2017). Its finding is asymmetric in a way that matters here.

Source: Seifert, Roberts & Lobell (2017), Agronomy Journal 109:541–548, as summarised in Ortez & Elmore, UNL CropWatch, Apr 8, 2020.

Continuous soybeans are penalised more than twice as hard as continuous corn, and unlike corn the penalty does not plateau — it climbs monotonically with each additional year.

Now price it. At the plateau soybean price of $11.00 and a 70-bushel high-productivity yield, gross revenue is $770 an acre. The penalties convert directly:

Source: penalties from Seifert et al. (2017); dollar conversion at a stated 70 bu/acre and $11.00/bu basis — arithmetic, not survey data.

Set that against the advantage the grower would be chasing. The average soybean advantage over 2013–2025 was $53 an acre. The first-year penalty for taking corn out is $57.75. The advantage does not clear the penalty in year one, let alone at the multi-year average of $79.31, where it recovers only 67% of the cost. The 2024 gap of $41 clears just 71% of the first-year penalty.

Ground Truth: This is the answer to why a thirteen-year, ten-of-thirteen profitability edge did not empty the corn acre, and it needs no appeal to habit, sentiment or agronomic conservatism. The edge was never large enough. A grower who moved to continuous soybeans on the average advantage would have paid $57.75 to earn $53 in the first year, and the arithmetic worsens every year after. The rotation is not surviving despite the economics — it is winning on them.

4. What would actually change the acre

The break-even follows directly: the corn-minus-soybean gap has to exceed roughly $58 an acre to justify the first year out of rotation, and around $79 to sustain it. Only one year in the record clears that decisively — 2023's $237, which cleared even the year-five penalty twice over.

Two things follow, and they point in opposite directions.

The threshold argues for stability. A gap that has averaged $53 against a $58–79 hurdle is a system in equilibrium, not one on the edge of a shift. Small changes in the cost gap will move budgets without moving acres.

The trend argues for pressure. The mechanism farmdoc identifies for 2027 is nitrogen, and nitrogen is corn-specific. Elevated nitrogen prices widen the corn-minus-soybean cost gap directly, and that is the one input that can push the gap through the rotation hurdle without any change in commodity prices. Retail anhydrous ammonia is up 27% year over year on the current public board (DTN/Progressive Farmer, Aug 19, 2026), and the farmdoc authors expect nitrogen strength to persist into 2027.

Which of those wins is an empirical question with a specific test: does the gap clear $58 an acre, and does it stay there for more than one year? A single year above the hurdle is not enough, because the penalty compounds — a grower who breaks rotation on one wide year and finds the gap back at $50 the following season has bought a climbing penalty with a one-off advantage.

The asymmetry in how the two penalties behave over time deserves more weight than it usually gets, because it makes the two rotations genuinely different bets rather than mirror images. Continuous corn's penalty starts near 2%, rises to roughly 5.5% by years two to four, and then plateaus — a grower who commits to corn-on-corn absorbs a one-time step and thereafter faces a stable, budgetable drag. Continuous soybeans start worse, at 7.5%, and keep climbing to nearly 15% by year five with no plateau in the published range. That means the two decisions are not symmetric in reversibility either. Corn-on-corn reaches a known steady state that can be priced into a long-run budget. Continuous soybeans put the grower on a curve whose end is not established by the data, and the cost of having been wrong grows with every year the position is held. A rotation break justified by one season's economics is a bet that the gap will keep widening — and on this record it has averaged $53 against a hurdle that rises as you stay.

5. Limits on all of this

Three should be stated plainly.

The return figures are Central Illinois high-productivity farmland. They are the best-documented series available and they are not the national experience. A grower in the western Corn Belt with different yields, different nitrogen requirements and different basis faces a different comparison, and the direction can differ, not merely the magnitude.

The rotation-penalty study covers 2007–2012. It is the largest of its kind and it is now fourteen to nineteen years old. Soybean genetics, seed treatments and disease management have all moved since. The direction of any bias is not obvious: better disease packages could reduce the penalty, while sustained continuous cropping and building pathogen pressure could raise it. Nothing here establishes which.

The dollar conversions mix sources. Yield penalties come from a six-state field study; the dollar values come from applying them to a stated Central Illinois yield and a long-run plateau price. That is a sensitivity exercise, not a budget, and it is presented as one. The comparison it supports — advantage against penalty — is robust to reasonable changes in either input, because the two figures differ by less than 10% and would have to move a long way to reverse.

What is not stated anywhere here is the year-by-year distribution of the advantage. The published series gives the period averages, the hit rates and the 2023 maximum. It does not give every year, so this piece cannot say how many individual years cleared the $58 hurdle — only that the average does not. That is a real gap, and a grower with access to the full series should run that count before treating the average as the decision variable.

References

1. Schnitkey, G., N. Paulson, C. Zulauf and B. Zwilling, "Historic Corn and Soybean Profitability and 2027 Planting," farmdoc daily, University of Illinois, Aug 4, 2026. https://farmdocdaily.illinois.edu/2026/08/historic-corn-and-soybean-profitability-and-2027-planting.html
2. Paulson, N., G. Schnitkey and C. Zulauf, "Historical Corn versus Soybean Returns in Illinois," farmdoc daily, University of Illinois, Jun 2, 2026. https://farmdocdaily.illinois.edu/2026/06/historical-corn-versus-soybean-returns-in-illinois.html
3. Seifert, C.A., M.J. Roberts and D.B. Lobell, "Continuous Corn and Soybean Yield Penalties across Hundreds of Thousands of Fields," Agronomy Journal 109:541–548, 2017. doi:10.2134/agronj2016.03.0134
4. Ortez, O. and R. Elmore, "Yield Penalties for Continuous Corn and Soybean in the Midwest, US," University of Nebraska–Lincoln CropWatch, Apr 8, 2020. https://cropwatch.unl.edu/2020/yield-penalties-continuous-corn-and-soybean-midwest-us
5. DTN/Progressive Farmer, "5 Fertilizer Prices Slightly Lower Than Last Month; 3 Higher," Aug 19, 2026. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3

Yield and price bases used for dollar conversions (230 bu/acre corn, 70 bu/acre soybeans, $4.50 and $11.00 long-run plateau prices) are stated where used and are illustrative, not survey values.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

USDA's preliminary 2027 cost-of-production forecast puts total costs at record levels for every major field crop: corn at $952 an acre, soybeans at $701, sorghum at $477 and wheat at $428 (American Farm Bureau Federation Market Intel, Jun 18, 2026; Farm Policy News, Jun 22, 2026). The coverage led on the records and on a second point that reads as relief — that the 2027 increase is not being driven by fuel and fertilizer, but by seed, chemicals, repairs, labour, machinery and cash rent. Both framings miss the number that matters, which is not either crop's cost but the distance between them.

1. The percentage column and the dollar column disagree, and the dollar column decides

Since 2005, total production costs have risen 165% for soybeans, 146% for corn, 106% for wheat and 103% for rice. Soybeans are the fastest-inflating major crop on that list, which is how the figure is usually reported.

Work backwards from the 2027 forecast to what those growth rates imply for the 2005 base, and the picture inverts.

Source: 2027 forecasts from USDA via AFBF Market Intel, Jun 18, 2026; growth rates as published. 2005 values are back-calculated from the two published figures and are derived, not USDA-reported.

Soybeans inflated faster and corn got more expensive. There is no contradiction: corn started from a base $122 an acre higher, so a smaller percentage on a larger number added more dollars — $565 against $436. The gap between the two crops widened by $128.54 an acre and more than doubled.

Ground Truth: A percentage growth rate is the wrong unit for a rotation decision. A grower choosing between corn and soybeans is not comparing inflation rates; they are comparing two budgets in dollars, and on that basis corn has become the more expensive crop by a margin that has grown 105% in twenty-two years. Reporting soybeans as the fastest-inflating crop is accurate and points the reader in exactly the wrong direction.

2. That widening is large enough to explain the whole regime flip

The University of Illinois farmdoc series splits the last twenty-six years into two thirteen-year regimes: corn out-earned soybeans by an average $59 an acre over 2000–2012, and soybeans have out-earned corn by an average $53 an acre over 2013–2025 (farmdoc daily, Jun 2, 2026). The swing between those two regimes is $112 an acre toward soybeans.

The same farmdoc work notes that Illinois corn and soybean prices have sat on long-run plateaus of roughly $4.50 and $11.00 since 2006. If prices are flat and yields trend for both crops, then a shift in relative profitability has to come from the cost side — and the cost side moved $128.54 an acre against corn.

Source: returns from farmdoc daily, Jun 2, 2026; cost gap derived from USDA 2027 forecasts via AFBF, Jun 18, 2026. This is a cross-source reconciliation, not an accounting identity — see the limits below.

The cost divergence accounts for about 115% of the observed swing, leaving a $16.54 residual pointing mildly back toward corn, which is the direction a modest revenue or yield advantage would push it.

This has to be labelled honestly, because it is the kind of result that flatters itself. The two figures come from different datasets covering different geographies and different windows — national USDA cost forecasts spanning 2005 to 2027 against Illinois returns spanning 2000 to 2025. They are not designed to be differenced. What the comparison establishes is a magnitude match: the cost divergence is the right size to have caused the return flip, and no additional mechanism is needed to explain it. It does not establish that it did, and a residual of $16.54 on numbers this loosely matched should be read as "small" rather than as a measurement.

Ground Truth: The usual explanation for the post-2013 soybean advantage reaches for demand — the Renewable Fuel Standard fading for corn, biodiesel arriving for soybean oil. Those are real. But the cost gap alone is the right size to do the whole job, and it is the more tractable explanation because it is the one a grower can look up. Before attributing a rotation decision to biofuel policy, difference the two budgets.

3. The 2027 increase sits in the lines a grower cannot act on

The second reported finding — that 2027's cost growth comes from seed, chemicals, repairs, labour, machinery and cash rent rather than from fuel and fertilizer — was framed as good news. It is closer to the opposite, and the reason is that those two groups of costs have very different response times.

Fertilizer and fuel are the short-cycle lines. A grower can switch nitrogen source, change rate, move application timing, prepay into a dip, or shop retailers between now and spring. They are the lines where information has cash value within a season.

Seed, chemicals, repairs, labour, machinery and cash rent are not. Cash rent is negotiated annually and is sticky downward. Machinery is capital already committed. Labour is a market wage. Seed and chemical prices are set by a small number of suppliers on annual price sheets. Repairs are a function of a fleet already owned and hours already run.

Fertilizer's weight in the bill is well documented and it is smaller than most coverage implies. USDA ERS puts fertilizer at roughly 22% of total corn production expenses over 2006–2016, about 17% over 2017–2021, and around 24% in the 2022 spike year, against a peak of 26% in 2008 (USDA ERS, Charts of Note, Mar 27, 2024). In dollars, fertilizer averaged near $125 an acre over 2006–2021, rose to $225.78 in 2022, and eased to $186.73 in 2023.

Apply the range to a $952 corn budget:

Source: shares from USDA ERS Charts of Note, Mar 27, 2024, applied to the USDA 2027 corn cost forecast. Dollar values are arithmetic, not USDA-reported for 2027.

Now price what "relief on fertilizer" is worth. At a 20% share — mid-range — fertilizer is about $190 an acre of the 2027 corn budget.

Source: arithmetic on a stated 20% fertilizer share of the USDA 2027 corn cost forecast.

$19.04/acre

What a 10% fall in fertilizer prices is worth against a record 2027 corn budget — 2% of the bill, and less than the $41-an-acre 2024 return gap between corn and soybeans. (Derived from USDA ERS share data and the USDA 2027 cost forecast)

A 10% fertilizer decline moves 2% of the corn bill. That is not nothing, and it is a great deal less than the framing "2027 costs are not being driven by fertilizer" invites a reader to assume. The corollary is the uncomfortable one: if the growth is in the other 80%, then the share of the cost stack a grower can act on within a season is falling, and the budget is becoming more fixed and less manageable at exactly the moment it sets records.

4. What this does to the 2027 rotation decision

Three practical consequences follow, and they are not all in the same direction.

The cost gap is now the dominant term. At $251 an acre, the corn-minus-soybean cost difference is larger than any plausible single-season move in fertilizer prices. A grower waiting on nitrogen to decide the rotation is waiting on a variable that can move perhaps $20–30 an acre against a $251 structural gap.

The gap is still not the whole comparison. Corn also generates more revenue per acre, which is why the return gap ($53 an acre on the Illinois series) is far smaller than the cost gap ($251). Confusing the two is the standard error in this analysis and it overstates the case for soybeans by roughly a factor of five.

Cash rent is the line to watch, not fertilizer. It is named among the 2027 drivers, it applies to both crops, and it is the one large cost that is genuinely negotiated rather than quoted. A grower who spends the winter shopping nitrogen and accepts a rent increase without argument has optimised the smaller number.

The other two crops in the forecast are worth a glance for scale, because they reframe what "expensive" means here. Sorghum is projected at $477 an acre and wheat at $428 — half of corn's $952 and roughly 60% of soybeans' $701. Neither is a realistic substitute across most Corn Belt ground, and both carry lower revenue to match, so this is not an argument for planting wheat. What it does establish is that corn's cost structure is not a general agricultural condition but a specific one: corn is the crop that has absorbed the input intensification of the last two decades. A grower whose rotation includes any acreage where a small-grain or sorghum option is agronomically live is holding a cost-side option worth $475 to $524 an acre in avoided expense, against whatever revenue is forgone — and that option becomes more valuable every year the corn budget sets a record.

5. Limits

These are preliminary forecasts. USDA revises in November, and the source reporting says so explicitly. A 2027 forecast made in June 2026 is a planning input, not a budget, and the fertilizer components in particular are being forecast through two more price cycles.

The 2005 base values are derived. USDA published the 2027 forecasts and the growth percentages; this piece back-calculated the 2005 figures from those two. If the published growth rates are rounded, the 2005 values and therefore the $128.54 gap-widening carry that rounding. At whole-percent rounding the gap-widening moves by roughly $2–3 an acre, which does not affect any conclusion drawn from it.

The share data ends in 2023. The ERS fertilizer-share series cited here runs to 2023. The 17–26% range is wide enough to bracket a reasonable 2027 value, which is why the calculation is presented as a range rather than a point, but no 2027 share is published and none is claimed.

Nothing here measures the revenue side. Every comparison in sections 1 through 3 is a cost comparison. The reconciliation in section 2 works only because the cited price series are described as flat over the period, and a reader who believes prices did move should treat that section as unresolved rather than adjusted.

References

1. Parum, F., "Production Costs Expected to Reach New Highs in 2027," American Farm Bureau Federation Market Intel, Jun 18, 2026. https://www.fb.org/market-intel/production-costs-expected-to-reach-new-highs-in-2027
2. Farm Policy News, University of Illinois, "Farm Production Costs to Hit Record Highs in 2027, USDA Says," Jun 22, 2026. https://farmpolicynews.illinois.edu/2026/06/farm-production-costs-to-hit-record-highs-in-2027-usda-says/
3. Paulson, N., G. Schnitkey and C. Zulauf, "Historical Corn versus Soybean Returns in Illinois," farmdoc daily, University of Illinois, Jun 2, 2026. https://farmdocdaily.illinois.edu/2026/06/historical-corn-versus-soybean-returns-in-illinois.html
4. USDA Economic Research Service, "Fertilizer share of expected corn production expenses drops back after 2021–22 spike," Charts of Note, Mar 27, 2024. https://www.ers.usda.gov/data-products/charts-of-note/108828
5. Schnitkey, G., N. Paulson, C. Zulauf and B. Zwilling, "Historic Corn and Soybean Profitability and 2027 Planting," farmdoc daily, University of Illinois, Aug 4, 2026. https://farmdocdaily.illinois.edu/2026/08/historic-corn-and-soybean-profitability-and-2027-planting.html

All 2005 base values, gap calculations and per-acre fertilizer dollar figures in this article are arithmetic derived from the published sources above and are labelled as derived wherever they appear.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Correction, August 31, 2026. This article as originally published on August 28 reported corn dented at 41% against a five-year average of 45%, and built its argument on a dough-to-dent gap that had widened to 45 points against a 37-point average. Those figures were transposed. The USDA table reads 45% dented for the week ending August 23, 2026, against 42% a year earlier and a 2021–2025 average of 41% — four points ahead of average, not four behind. The gap is therefore 41 points against a 41-point average, not 45 against 37. The error was ours, and it inverted the article's central claim; the analysis below has been rewritten against the correct figures rather than patched. How it happened, and why it is worth reporting rather than quietly fixing, is in section 5.

Two numbers in the same USDA table are moving in opposite directions, and the divergence is the story — but it is not the divergence most of the trade press reported. As of Sunday, August 23, corn in the 18 major states was 86% at dough against a five-year average of 82% and 45% dented against an average of 41% (USDA NASS, Crop Progress, released Aug 24, 2026). Both stages are four points ahead. Development is not merely normal, it is marginally early and uniformly so. In the same release, good-to-excellent condition fell three points to 57%, the lowest late-August rating in three years and fourteen points below a year ago. A crop that is running on schedule and deteriorating in the ratings at the same time is telling you something quite specific about which yield component it is losing.

1. What the release actually says

What's new: The development table shows no stage lagging. Dough, dent and maturity are all at or ahead of their five-year averages, and the two that are ahead are ahead by the same margin.

Source: USDA NASS, Crop Progress, released Aug 24, 2026, covering the week ending Aug 23, 2026. Dough, dent and maturity cover 18 states representing 91% of 2025 corn acreage. Figures read from the release's own table.

Evidence: Dough advanced 10 points in the week and dent advanced 16. Both stages sit four points above their five-year averages and dent is three points above where it was a year ago. Soybeans in the same release were 91% setting pods, three points ahead of average, and 6% dropping leaves, two points ahead.

Ground Truth: The uniformity is the useful part. When one stage runs ahead and the next runs behind, the crop has a spread problem — a front end arriving early and a middle stalling — and national percentages conceal it. That is not what this table shows. Dough +4, dent +4 and maturity exactly on average is the signature of a crop moving through its stages at a normal rate, having entered them slightly early. There is no phenological anomaly here to explain, which means any yield concern this year has to be argued on grounds other than timing.

2. The grain-fill window is exactly its usual width

What's new: Dough percentage minus dent percentage is the share of the crop that has reached dough but not yet dented — the population sitting in the grain-fill window on the survey date. It is a stock rather than a rate, and comparing it to its own average is more informative than comparing either stage alone.

Evidence:

Source: Crop Root Zone calculation from USDA NASS Crop Progress percentages, Aug 24, 2026 release. Bars scaled to the widest gap.

The gap is 41 points against a five-year average of 41 — identical to two significant figures, and two points wider than the same week last year. Whatever is happening to this crop, an unusual share of it is not accumulating in the fill window.

41 vs 41

The share of the US corn crop between dough and dent on August 23, against its own five-year average. The grain-fill window is exactly its normal width. (Crop Root Zone calculation from USDA NASS, Aug 24, 2026)

Ground Truth: A normal gap forecloses an explanation that would otherwise be attractive. If fill were being arrested by stress — the crop unable to move kernels toward maturity — the dough-to-dent gap would widen, because corn would accumulate at dough and fail to dent. It has not widened; it is exactly average, and dent is advancing faster than dough entered. So the three-point condition decline cannot be attributed to stalled grain fill, because grain fill is not stalled. That removes the most commonly asserted mechanism linking a late-August rating drop to yield, and it does so without needing any view on the weather.

3. Condition fell anyway, and that is now the whole question

What's new: Good-to-excellent corn fell to 57%, from 60% a week earlier and 71% a year ago. The August series has declined steadily.

Evidence:

Source: USDA NASS Crop Progress weekly releases as reported by DTN/Progressive Farmer, Aug 3, 10, 17 and 24, 2026.

Corn has given up four points across the month and three of them in the final week. Soybeans have declined one point a week, without acceleration. The corn rating is the lowest for this week in three years, and North Dakota's poor-to-very-poor corn share rose ten points in the week.

Ground Truth: Development on schedule plus condition falling is a narrower diagnosis than either observation alone, and it points at kernel weight. Corn sets three yield components in sequence — ear count, kernel number per ear, and kernel weight — and only the last is still open once a crop is past dough. Timing is not being lost, because the stages are on time. Ear and kernel counts were fixed weeks ago. What remains for a deteriorating crop to lose in late August is grams per kernel, and a rating that falls while development proceeds normally is most consistent with a crop filling fast but thin. The corollary for the reader is that the acceleration matters more than the level: three of the four points went in one week, and a rating that is falling at an increasing rate late in fill is a different signal from one that has been drifting since July.

4. Where the crop actually sits, and what September can still reach

What's new: Differencing the cumulative stage percentages converts the table into a position statement — what share of the crop is at each point in the fill sequence right now.

Evidence:

Source: Crop Root Zone calculation by differencing USDA NASS cumulative stage percentages, Aug 24, 2026 release. Components sum to 100%.

45% of the national crop — the dented-and-beyond group — has its kernel weight substantially set. Another 41% is in the window where it is being set now, and 14% has not started.

Ground Truth: The practical consequence cuts against how the market usually treats late-August weather. Nearly half of this crop is past the point where September rain can add weight, and the largest single block that rain could still reach is the 41% in mid-fill. September forecasts, which will get considerable attention over the next fortnight, are arriving to change a number that is already about half written. For a grower the decision-relevant version is that this is close to the last week in which a field-level kernel-depth assessment tells you something a September walk will not, and the yield estimate formed now is close to the one that gets harvested.

5. The transposition, and why it is worth reporting

What's new: The dent figure was widely reported the wrong way round, including by this publication.

Evidence: The USDA table for corn dented, 18 states, week ending August 23, 2026, reads across as: 42% (Aug 23, 2025), 29% (Aug 16, 2026), 45% (Aug 23, 2026), 41% (2021–2025 average). Several trade summaries rendered this as "41%, compared to 29% the prior week and the average of 45%" — the current-week and five-year-average columns swapped. The prior-week figure of 29% was carried correctly in both versions, which is exactly what made the error hard to see: one of the three numbers checks out, and the transposed pair is internally plausible because 41 and 45 are both credible dent percentages for late August.

Two independent checks settle it. First, the same release's dough row reads 81 / 76 / 86 / 82 in that column order, and 86% dough against an 82% average with 76% the prior week is the figure every summary agrees on — so the third column is the current week and the fourth is the average. Applying that order to the dent row gives 45% current against a 41% average. Second, a correctly-rendered summary reports dent as "45%, 3 percentage points ahead of last year's 42% and 4 percentage points ahead of the five-year average of 41%," which is internally consistent and matches the table.

Ground Truth: A transposition that survives because one of three numbers is right is the most durable kind of error, and it is worth more attention than the number itself. The general lesson for anyone consuming secondary agricultural data is that a cumulative progress series has an internal check that costs nothing: stages must be monotonic, and the current week must exceed the prior week. Dent at 41% this week against 29% last week passes that check, which is why the wrong figure propagated. The check that catches it is different — compare the same column across stages. If dough's current-week column is the third and dent's is read as the fourth, the two rows are being read on different bases, and no amount of internal plausibility will reveal it. The habit worth forming is to read one row's column order off a row you already know the answer to.

6. What this does to the tour-versus-USDA gap

What's new: This desk has tracked the divergence between tour-based sampling and USDA's own estimates through August. Pro Farmer's crop tour finished 669 million bushels under USDA (Crop Root Zone, TRZ-0116, Aug 22, 2026), and an earlier eight-state comparison put the sample and USDA about 585 million bushels apart (Crop Root Zone, TRZ-0113, Aug 20, 2026).

Evidence: Those gaps were built on ear counts and kernel rows — measurements taken before most of the crop reached the stage the current progress table describes.

Ground Truth: Tour sampling counts ears and kernel rows; it does not measure kernel weight, which is the component still open in the window this article is about. That is not a criticism of the method — kernel weight cannot be measured in August because it has not happened yet — but it means the 585 to 669 million bushel gaps carry an unmeasured term. The corrected development picture sharpens rather than softens the point. Because development is exactly on schedule, the tour-versus-USDA gap cannot be attributed to phenology: the tours were not sampling an unusually early or late crop. Whatever explains the gap is either sampling method or a yield component the tours did not measure, and kernel weight is the only one of those still moving. The September estimate is therefore not merely a referee between USDA and the tours — it is the first observation that prices the term neither of them measured.

7. Limits

The five-year averages are as published by USDA alongside the current-week figures and are not recomputed here. Stage percentages cover 18 states representing 91% of corn acreage, so the national figure is a weighted aggregate that can conceal offsetting regional patterns — a normal national gap is consistent with several states being genuinely early and several genuinely late, which would have different implications from a uniformly normal crop. This piece does not have the state-level detail to distinguish those, and the distinction matters for basis even where it does not change national yield. North Dakota's ten-point deterioration is the one state signal in the release and it is a condition figure, not a development one.

Second, the diagnosis in section 3 is an inference from two series moving independently, not a measurement of kernel weight. Nothing in Crop Progress measures grams per kernel. The claim that a normally-developing, poorly-rated crop is losing weight rather than time is the most parsimonious reading of the two series; it is not established by them, and a rating decline driven by disease pressure or standability would produce the same table while implying something different for yield.

Third, the condition series in section 3 is compiled from four weekly releases as reported by a single secondary source. Given section 5, that dependence is worth naming rather than glossing: the four G/E values were not independently verified against each release's own table, and only the August 23 figures in sections 1, 2 and 4 were read from the primary document.

References

1. USDA National Agricultural Statistics Service, Crop Progress, ISSN 1948-3007, released August 24, 2026, covering the week ending August 23, 2026. Corn dough, dented and mature percentages for 18 states with 2021–2025 averages; corn and soybean condition; soybean pod-setting and leaf-drop. https://esmis.nal.usda.gov/sites/default/release-files/796031/prog3426.pdf
2. DTN/Progressive Farmer, "USDA Crop Progress: Corn Rated 57% Good to Excellent; Soybeans 60% Good to Excellent as of Aug. 23," Aug 24, 2026. https://www.dtnpf.com/agriculture/web/ag/news/article/2026/08/24/usda-crop-progress-corn-rated-57-60
3. DTN/Progressive Farmer, "USDA Crop Progress: Corn Rated 60% Good to Excellent; Soybeans 61% Good to Excellent as of Aug. 16," Aug 17, 2026. https://www.dtnpf.com/agriculture/web/ag/news/article/2026/08/17/usda-crop-progress-corn-rated-60-61
4. DTN/Progressive Farmer, "USDA Crop Progress: Corn Rated 61% Good to Excellent; Soybeans 62% Good to Excellent as of Aug. 9," Aug 10, 2026. https://www.dtnpf.com/agriculture/web/ag/news/article/2026/08/10/usda-crop-progress-corn-rated-61-62
5. DTN/Progressive Farmer, "USDA Crop Progress: Corn Rated 61% Good to Excellent; Soybeans 63% Good to Excellent as of Aug. 2," Aug 3, 2026. https://www.dtnpf.com/agriculture/web/ag/news/article/2026/08/03/usda-crop-progress-corn-rated-61-63

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

December 2026 corn traded at $5.3325 a bushel on the morning of August 27, down 3¼ cents on the session and within 5½ cents of its 52-week high of $5.3875 (Barchart, 10-minute delayed quote, Aug 27, 2026). The Risk Management Agency's projected price for the 2026 corn crop — the figure that set every Revenue Protection guarantee written this spring — is $4.62 (farmdoc daily, Mar 2026). The gap is 71¼ cents, or 15.4%. That gap is about to be written into a lot of guarantees, and it is being described in a way that overstates what it does.

1. What the projected price is, and what October will do to it

What's new: Nothing about the projected price. It was fixed in February and has not changed since. What changed is the December contract, which has travelled $1.075 off its 52-week low of $4.2575 — a 25.3% rally — and is now well above the price the guarantee was struck at.

Evidence: For corn and soybeans the projected price is the average of daily settlements of the December corn and November soybean contracts across all active trading days in February. The harvest price is the average of the same contracts across October. Both crops came in below their fifteen-year averages this spring.

Source: farmdoc daily, "Projected Prices and Volatility Factors for 2026," March 2026. Discovery is the February average of the December corn and November soybean contracts.

Where the two harvest contracts stand today against those spring prices:

Source: Barchart, 10-minute delayed quotes, Aug 27, 2026 (corn 533-2, −3-2; soybeans 1259-6, −6-2; soybean 52-week high set Aug 26, 2026, low set Oct 1, 2025). Premium column is a Crop Root Zone calculation against the RMA projected prices above.

Note the soybean low: $10.48, set on October 1, 2025 — inside last year's harvest-price discovery window. The October average is not a formality, and it does not always go the grower's way.

2. The guarantee goes up. This is the part everyone reports.

What's new: Revenue Protection sets the guarantee at coverage level × approved yield × the greater of the projected price and the harvest price. If October averages anywhere near today's board, the harvest price wins and every RP guarantee is re-struck 15% higher at no additional premium.

Evidence: On a 200-bushel approved yield, the arithmetic is straightforward.

Source: Crop Root Zone calculation. 200 bu/acre approved yield; guarantee = coverage × APH × price. Harvest-price column uses the Aug 27 December futures quote as a stand-in for an October average that has not been observed.

+$114.00/acre

The increase in an 80% Revenue Protection guarantee on a 200-bushel APH if October averages today's December corn price. It arrives at no additional premium — and, as Section 3 shows, it is not worth $114. (Crop Root Zone calculation from RMA's 2026 projected price of $4.62 and the Aug 27, 2026 December futures quote)

3. The trigger yield does not move, and that is the whole story

What's new: The harvest price appears on both sides of the indemnity calculation. It raises the guarantee, and it raises the revenue that guarantee is measured against. When the harvest price exceeds the projected price, the two effects cancel exactly, and Revenue Protection collapses into yield protection denominated at the higher price.

Evidence: An indemnity is owed when revenue-to-count falls below the guarantee. Revenue-to-count is actual yield × harvest price. So, writing C for coverage, A for approved yield, HP for harvest price and taking HP ≥ projected price:

• Guarantee = C × A × HP
• Revenue to count = actual yield × HP
• Indemnity is owed when actual yield × HP < C × A × HP

Divide both sides by HP, which is positive and identical on both sides, and the condition reduces to actual yield < C × A. The price drops out entirely.

At 80% coverage on a 200-bushel APH, the trigger yield is 160 bu/acre — the same 160 bushels whether corn settles October at $4.62 or $5.3325 or $6.00. Here is what the provision is actually worth at a range of yields:

Source: Crop Root Zone calculation. 80% coverage, 200 bu approved yield, trigger yield 160 bu. The $4.62 column is counterfactual — it is what the same shortfall would have paid had the harvest price not exceeded the projected price. Value of the provision = bushels short × ($5.3325 − $4.62).

Soybeans behave identically, with smaller absolute numbers. On a 60-bushel approved yield at 80% coverage, the guarantee rises from $532.32 to $604.68 an acre — an increase of $72.36 — while the trigger yield stays at 48 bu/acre in both cases. A grower ten bushels short collects $125.98 instead of $110.90, a gain of $15.08 an acre, which is simply 10 × the $1.5075 premium of the November board over the $11.09 projected price. A grower who makes 60 bushels collects nothing under either price.

Ground Truth: The higher harvest price does not make a claim one bushel more likely. It makes each missing bushel worth 71¼ cents more. For the majority of growers who will make or beat their approved yield, the entire 15.4% rally in the guarantee is worth exactly zero, and the "+$114 an acre" figure that will circulate this autumn is a gross number attached to an event most policies will never trigger. The honest statement of what the harvest price provision delivers in 2026 is: $0.7125 per bushel, on bushels you did not grow.

4. The marketing implication runs opposite to the intuition

What's new: Because the provision pays only on a shortfall, it is not a substitute for pricing bushels that are actually in the field.

Evidence: A grower who reads "my guarantee just went up $114 an acre" and slows cash sales has swapped a certainty for a contingency. The extra guarantee is conditional on losing at least 40 bushels off a 200-bushel APH at 80% coverage. Price protection on the bushels actually harvested does not come from the policy at all in a rising-price year — it comes from the market, and it is available at $5.3325 today.

The reverse case is the one worth thinking about, because it is where the provision earns its keep: a grower who has forward sold bushels and then loses the crop faces buy-out or roll costs at the higher price. That is the classic replacement-cost exposure, and the harvest price provision is a partial offset to it — it pays more per missing bushel precisely in the scenario where missing bushels are most expensive. The instrument is well designed for that case, and poorly described for every other one.

Ground Truth: The harvest price provision is replacement-cost insurance for forward-sold bushels, not a revenue bonus. Read that way, the correct response to a 15% rally in the guarantee is not to sell less — it is to recognise that the policy now covers a larger share of the cost of failing to deliver on sales already made, which makes an incremental forward sale less risky than it was in February, not more.

5. What the February decision looks like in August

What's new: The 2026 volatility factors were the lowest in several years — 0.15 for corn against 0.18 in 2025, and 0.13 for soybeans, described by farmdoc as the lowest for soybeans since the 2023 crop year.

Evidence: The volatility factor is an input to premium: lower volatility, lower premium, holding coverage and price constant. So growers who bought up coverage in February 2026 did so in a comparatively cheap year, and then watched the harvest contract rally 25% off its low. The buy-up decision looks materially better in August than it did when it was made — but the reason is the price path, not the premium, and nobody chose the price path.

There is a limit to how much comfort to take from this. The 2026 projected price was itself $0.14 below the fifteen-year corn average of $4.76 and $0.23 below the soybean average of $11.32. A low projected price means a low guarantee floor. Growers whose October average ends up below $4.62 revert to the projected price and get none of Section 3's arithmetic — and the December contract's own 52-week low of $4.2575 is below the projected price, so that outcome is inside the contract's demonstrated range for this crop year.

6. Limits

Today's board is not the harvest price. Everything in Sections 2 and 3 uses the August 27 December quote as a placeholder for an October average that will not exist for another five weeks. If October averages $5.00, the value of the provision on a 20-bushel shortfall falls from $14.25 to $7.60 an acre. If it averages below $4.62, the provision is worth nothing at all.

Quotes are delayed. Both futures figures are 10-to-15-minute delayed quotes read on the morning of August 27, not settlements.

The yield and coverage figures are illustrative. A 200-bushel corn APH at 80% coverage is a common Corn Belt combination, not a survey average, and every number in Section 3 scales linearly with both. The trigger-yield conclusion does not depend on either — it is algebraic and holds at any coverage level and any approved yield.

Unit structure, enterprise units, trend-adjusted APH and the Supplemental/Enhanced Coverage Options are all outside this piece. They change the numbers. They do not change the cancellation in Section 3, which is the point.

Anyone acting on this should confirm their own approved yield, coverage level and unit structure with their crop insurance agent. This is a piece of arithmetic about how the instrument works, not advice about a particular policy.

References

1. farmdoc daily (University of Illinois), "Projected Prices and Volatility Factors for 2026," March 2026. 2026 corn projected price $4.62, volatility factor 0.15; 2026 soybean projected price $11.09, volatility factor 0.13; 2025 comparisons $4.70/0.18 and $10.54/0.14; 15-year averages $4.76 and $11.32; February discovery period using the December corn and November soybean contracts. https://farmdocdaily.illinois.edu/2026/03/projected-prices-and-volatility-factors-for-2026.html
2. Barchart, December 2026 corn futures (ZCZ26), 10-minute delayed quote read Aug 27, 2026: 533-2 ($5.3325), −3-2 (−0.61%); 52-week high 538-6 ($5.3875), 52-week low 425-6 ($4.2575). https://www.barchart.com/futures/quotes/ZCZ26
3. Barchart, November 2026 soybean futures (ZSX26), delayed quote read Aug 27, 2026: 1259-6 ($12.5975), −6-2 (−0.49%); 52-week high 1270-0 ($12.70) set Aug 26, 2026; 52-week low 1048-0 ($10.48) set Oct 1, 2025. https://www.barchart.com/futures/quotes/ZSX26
4. USDA Risk Management Agency, Commodity Exchange Price Provisions — projected and harvest price discovery periods for corn and soybeans. https://www.rma.usda.gov/

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

USDA cut the national corn rating three points to 57% good-to-excellent in the crop progress report for the week ending August 23, released Monday afternoon — the largest single-week decline of the 2026 season, against 61% a year ago (Brownfield Ag News, Aug 24, 2026). The week that produced it was the week the eastern Corn Belt was under water: more than five inches across parts of central Illinois and more than a foot in eastern Indiana. It is a natural inference that the flood took the rating down. It did not. Sort the report by state and every eastern state in it is at or above the national average.

1. Sorting the states against the national number

What's new: The four states that absorbed the heaviest rainfall all rate above 57%, and the states that dragged the average are dry, not wet.

Evidence: State-level good-to-excellent for corn in the same report:

Source: USDA NASS Crop Progress for the week ending Aug 23, 2026, as reported by RFD-TV, Aug 24, 2026, and The Sabetha Herald, Aug 24, 2026, for Kansas. Colorado is reported as more than half poor or very poor without a good-to-excellent figure given; North Dakota is reported as roughly one-third rated good. We were unable to retrieve the NASS release PDF directly this run — the file path returned HTTP 404 — so state figures here are taken from broadcast and print summaries of it rather than from the release itself.

The dispersion is the finding, and it is enormous. Iowa and Kansas are both corn states in the same report and they are 36 rating points apart. Illinois, the state that ran the wettest headlines of the week, sits two points above the national average. Iowa alone rates 21 points above the national figure, and Iowa is the largest corn-producing state in the country — which means the single largest block of production in the survey is in the half of the map that did not deteriorate.

The moisture side of the same report corroborates it. Topsoil moisture was reported 87% short or very short in Wyoming and more than 75% short or very short in Oklahoma, while the eastern Belt was reporting surplus moisture and suspended fieldwork. Those are not marginal observations attached to a wet-week narrative; they are the condition of the ground in the states whose ratings actually fell. A national number built from both halves is an average across two opposite weather regimes, and averaging opposite regimes produces a figure that describes neither of them.

An independent, production-weighted index says the same thing from a different direction. Pro Farmer's Crop Condition Index for corn fell 2.52 points to 353.20, and Pro Farmer attributes the declines to western regions specifically, naming Nebraska (41.67 to 40.96) and North Dakota (11.19 to 10.35) as the movers (Pro Farmer, Aug 24, 2026). Illinois, Indiana and Ohio appear in that analysis only under soybeans, and only as minor declines.

Ground Truth: Two different surveys, one simple and one production-weighted, agree that the deterioration is western. That agreement matters more than either number alone, because it rules out the most common explanation for a surprising rating move — that a single state's crop reporters recalibrated. The trade spent last week pricing an eastern flood and the ratings moved on a western drought, and those two crops have opposite forward risk profiles: the flooded crop's damage is already done and will show up at the header, while the dry crop's damage is still accruing daily until the grain-fill window closes.

2. Why a three-point cut in late August is not a three-point cut in July

What's new: The same report puts 45% of the corn crop at dent and 6% at maturity, with 86% at dough.

Evidence: Development stage as of August 23:

Source: USDA NASS Crop Progress, week ending Aug 23, 2026, via Brownfield Ag News, Aug 24, 2026.

By dent, the corn plant has finished setting kernel number. Ear count and kernels per ear were determined weeks earlier, at silking and early grain fill. What remains is kernel weight — the depth of starch deposition in kernels that already exist. Stress at dent reduces test weight and can shorten the fill period; it cannot remove a kernel that is already there.

That has a direct consequence for how a rating change should be read. A three-point good-to-excellent decline registered in mid-July, when the crop is setting kernels, is a statement about yield potential. The same three-point decline registered when 45% of the crop is dented is a statement about the tail of grain fill, and the arithmetic leverage on final yield is much smaller. The rating scale does not change with the calendar. What it is measuring does.

3. The survey is accurate about the West and blind to the East

What's new: The eastern damage this desk sized last week is, by construction, invisible to a condition survey.

Evidence: Corn lodged and downed by the eastern rainfall does not necessarily change how a field looks to a crop reporter scoring plant health — and more importantly, the loss from lodging is not a growing loss at all. It is a harvest loss: grain that formed normally and then fails to enter the header. This desk put that at 2.25 to 3.00 times the machine hours per acre at the reduced ground speeds lodged corn requires, and derived roughly 5 bu/ac of loss per downed ear per thousandth of an acre from published assessment guidance (TRZ-0123, Aug 24, 2026).

None of that appears in a condition rating, because none of it has happened yet. It happens in October, at ground level, in front of a snout.

Ground Truth: The two halves of the Corn Belt are failing in ways this survey measures with opposite accuracy. Western drought stress is exactly what a condition rating is designed to capture, and it is being captured — late, at the stage where it costs the least, but accurately. Eastern lodging is a mechanical loss that the survey cannot see by design, and it is being missed entirely. Anyone treating 57% as a summary of the crop's problems is holding a number that is precise about the smaller risk and silent about the larger one. The correct reading of this week's report is not "the crop got worse by three points." It is "the crop got worse by three points in the West, and we still have no instrument pointed at the East."

4. Why soybeans only lost one point

What's new: Soybeans fell to 60% good-to-excellent, down one, against corn's three (Brownfield, Aug 24, 2026). Pro Farmer's soybean CCI fell 2.96 to 358.99 — a slightly larger index move than corn's, on a smaller headline decline.

Evidence: The crop calendar explains most of it. Soybeans are 91% setting pods and only 6% dropping leaves, which means essentially the entire seed-fill window is still ahead of the crop. August rainfall in the eastern Belt arrives inside the soybean yield-determining period and after most of the corn's. The two crops received the same weather and are at different points in their response function.

Source: USDA NASS via Brownfield Ag News and Pro Farmer, Aug 24, 2026. Spring wheat figures are Pro Farmer Crop Condition Index values, not good-to-excellent percentages; North Dakota, at 55.32% of HRS production, fell from 182.00 to 176.47.

Illinois is the clean local test: corn 59% and soybeans 59%, identical, in the state that took the rain. Nationally soybeans lead corn by three points. In the flooded state they are level — which is what you would expect if the rain was worth something to beans and nothing to corn.

5. What the market did with it

December corn settled Monday at $5.15½, up seven cents, after trading to a contract high earlier in the session; November soybeans lost 15¼ cents to $12.24¼ (Pro Farmer, Aug 24, 2026). The corn bid came alongside the Pro Farmer Crop Tour's national estimate of 15.344 billion bushels on an average yield of 173.2 bu/ac, released the prior Friday after the close, while the tour's soybean estimate read bearish.

The trade, in other words, bought the crop that lost three rating points and sold the crop that lost one. That is coherent if — and only if — the market is pricing the tour's production estimates rather than the condition series, which is a reasonable thing for it to do in the last week of August. It is a poor week to conclude anything about how much a rating point is worth.

Limits, stated plainly. The state-level figures here come from summaries of the NASS release rather than the release itself, and one of them (Colorado) is a qualitative description rather than a percentage; anyone acting on a specific state number should pull it from the source table. Good-to-excellent is a share, not a yield, and the relationship between the two is unstable across years and states — none of the arithmetic above converts rating points to bushels, deliberately. The claim that lodging is invisible to condition scoring is a structural argument about what the survey asks, not a measured comparison, and it would be falsified if the eastern states show sharp rating declines in next week's report. And the western deterioration, while accurately captured, is being captured in states whose combined share of national production is far smaller than Iowa's alone — which is the other reason the national number moved less than the western states did.

References

1. Brownfield Ag News, "57% of U.S. corn, 60% of soybeans good to excellent," Aug 24, 2026. https://www.brownfieldagnews.com/news/57-of-u-s-corn-60-of-soybeans-good-to-excellent/
2. Pro Farmer, "Corn crop conditions fall for third consecutive week," Aug 24, 2026 — Crop Condition Index values and state attribution. https://www.profarmer.com/news/corn-crop-conditions-fall-third-consecutive-week
3. RFD-TV, "Weekly Crop Progress and Livestock Production Update — Monday, August 24, 2026" — state-level conditions and topsoil moisture. https://www.rfdtv.com/rfd-tv-weekly-crop-progress-and-livestock-production-update-monday-august-24-2026
4. The Sabetha Herald, "Kansas crop progress, condition reported," Aug 24, 2026. https://www.sabethaherald.com/2026/08/24/kansas-crop-progress-condition-reported-10/
5. Pro Farmer, "After the Bell" market summary, Aug 24, 2026 — December corn and November soybean settlements and the Crop Tour production estimates.
6. Crop Root Zone, TRZ-0123, Aug 24, 2026 — harvest machine-hour arithmetic for lodged corn; and TRZ-0116, Aug 22, 2026 — the Pro Farmer-versus-USDA production gap.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Downbound grain barge rates on the Mississippi system have now fallen for three to four consecutive weeks at every origin the USDA tracks. The week ending Aug 25, 2026 put St. Louis at 579.7% of tariff, Twin Cities at 727% and Cairo-Memphis at 537.5% — down from peaks set in the weeks of Jul 28 and Aug 4 (USDA AMS, downbound grain barge rates, via the Agricultural Transport open data platform, retrieved Aug 31, 2026). That is the wrong direction for the calendar, and it is the opposite of what the same series did a year ago.

1. Seven origins, one direction

The decline is not one river reach or one product. It is the whole system.

Source: USDA AMS downbound grain barge rates, weekly, via agtransport.usda.gov, retrieved Aug 31, 2026. Cincinnati and Lower Ohio peaks are summer peaks; both series also carry a higher February value.

Two notes on reading the table. The unit is percent of tariff, the standard USDA publication basis — rates are quoted against a 1976 benchmark tariff rather than in dollars, which is why the numbers run in the hundreds. It is a ratio, and it is directly comparable week to week and year to year, which is what matters here. And Cincinnati and Lower Ohio are effectively one series: across the fifteen months examined they carry identical values in every week but two. Treating them as independent confirmation would be double-counting, and they are not counted twice in anything below.

2. Last year the same weeks went the other way

The reason a three-week decline in late August is worth writing about is that late August is when this series normally turns up. Harvest pulls barges into position, demand for downbound capacity rises, and rates climb into September.

Source: USDA AMS downbound grain barge rates, 2025 weeks, via agtransport.usda.gov.

Every origin rose from late August into mid-September 2025, most of them by 15% or more, with St. Louis up nearly 30%. Rates then eased through October once the peak movement passed. That is the textbook shape, and 2026 is not tracing it.

Ground Truth: A falling barge rate three weeks before harvest is not a small anomaly in a seasonal series — it is the seasonal signal running backwards. Two readings are possible and they have opposite implications: either equipment repositioned early and capacity is ample, in which case harvest freight will be cheap; or downbound demand is softer than the tonnage forecasts imply, in which case the rate is telling you something about export commitments that the movement data has not yet shown. The series alone cannot distinguish them, and anyone claiming it can is reading a conclusion into a price.

3. The year-over-year premium is real and it is evaporating

The level is still high. That is the fact that has driven most of this summer's freight commentary, and it is true.

Source: USDA AMS downbound grain barge rates, via agtransport.usda.gov.

But the premium three weeks ago was far larger, and the collapse is the more informative number.

Source: derived from USDA AMS weekly rates; peak-week YoY compares each 2026 peak with the corresponding 2025 calendar week.

Cairo-Memphis has surrendered 34 of its 47 percentage points of premium in three weeks. Lower Illinois is down to +9.3% and is within a normal year's noise of flat. The origins that have held their premium best — Cincinnati/Lower Ohio and St. Louis — are the Ohio-system and mid-river points, not the upper Mississippi.

4. What the move is worth at the farmgate

Barge freight is not an abstraction to a grower; it is deducted from the elevator's bid. A falling downbound rate widens the basis in the grower's favour, and the size is calculable to an order of magnitude.

The public series is in percent of tariff, so it needs one dollar anchor. Trade reporting in early August put St. Louis spot at $25.62 per ton (RFD-TV, Aug 4, 2026, citing the USDA Grain Transportation Report). Anchoring that figure against the corresponding week's percent-of-tariff value implies a benchmark tariff base near $3.99–4.02 per ton, and applying that base to the Aug 25 reading of 579.7% gives roughly $23.13–23.29 per ton.

6.5–7.0¢/bu

The improvement in the St. Louis downbound freight deduction between the early-August peak and the week of Aug 25, converted at 35.714 bushels of corn per short ton. (Derived from USDA AMS percent-of-tariff rates and a single reported $/ton anchor — see the method note)

The method note matters. That conversion rests on one reported dollar figure whose exact reporting week is not stated in the source, which is why both plausible anchor weeks are carried and the answer is given as a range rather than a point. The two anchors differ by half a cent per bushel, so the range is tight, but the whole calculation inherits whatever basis the original $25.62 was quoted on. Treat it as an order-of-magnitude conversion of a solid percentage move, not as a freight quote. The percentage decline is the measurement; the cents per bushel is an illustration of its scale.

Seven cents is not trivial against a corn basis that typically moves in a range of thirty to fifty cents through harvest, and it is running in the grower's favour at the moment of maximum delivery.

Ground Truth: Two of this year's freight stories point in opposite directions and only one of them has moved recently. Rail costs have been rising — record carloads, a record 61-cent-per-mile fuel surcharge in July — while barge has been falling for three to four weeks. A grower with truck access to both a river terminal and a rail-served shuttle should be repricing that choice now rather than defaulting to last year's answer, because the relative economics have moved and the barge leg is the one that improved.

5. What the series cannot tell you, and what to watch

It cannot tell you why. Percent-of-tariff is a price. Low water, ample equipment, soft export demand and early repositioning would all produce a falling rate, and this dataset distinguishes none of them. The 2022 and 2023 low-water episodes produced the opposite signature — sharply rising rates on reduced drafts and tow sizes — so a falling rate is at least weak evidence against a serious draft restriction at these origins, but that is an inference, not an observation, and no river-stage data is presented here to support it.

It is a spot rate, not a harvest rate. The week of Aug 25 is three to six weeks ahead of peak corn movement at most of these origins. In 2025 the largest weekly increases came in the second week of September. A decline through late August has, on that precedent, said little about where September settles.

One anchor is thin. Everything in section 4 hangs on a single publicly reported $/ton figure. A reader with access to the full USDA Grain Transportation Report can replace it with the published dollar series directly and should.

Three things would resolve the ambiguity in section 3, and all are observable within a month:

1. Whether the decline continues past Labor Day. Two more down weeks into mid-September would break the seasonal pattern decisively rather than merely departing from it.
2. Whether downbound tonnage confirms or contradicts the rate. Falling rates against rising tonnage is a capacity story and is good news. Falling rates against falling tonnage is a demand story and is not.
3. Whether the upper-river origins keep underperforming. Twin Cities and Mid-Mississippi lose their navigation season first; a premium that erodes fastest at the top of the system behaves differently from one eroding at the bottom.

That last point has a hard deadline built into the dataset itself, and it is visible as an absence rather than a number. The Twin Cities series simply stops for part of each winter — the weeks around February and early March carry no rate at all, because the upper Mississippi closes to navigation and there is no market to quote. Every bushel moving out of the upper river between now and the close has to move inside a window whose end is fixed by ice rather than by price. That is the structural reason upper-river rates normally firm into autumn: the seller has a deadline and the buyer knows it. A Twin Cities rate falling 11% from its peak in the last three weeks of August is therefore a more unusual observation than the same decline at Cairo-Memphis, where the river runs year-round and a shipper who dislikes September's rate can wait for November. Read that way, the upper river is the part of this series carrying the most information, and it is currently saying that capacity is arriving ahead of a deadline rather than tightening into one.

References

1. USDA Agricultural Marketing Service, weekly downbound grain barge rates (percent of tariff), Agricultural Transport Open Data Platform, dataset deqi-uken, series retrieved Aug 31, 2026. https://agtransport.usda.gov/Barge/Downbound-Grain-Barge-Rates/deqi-uken/data
2. USDA Agricultural Marketing Service, Grain Transportation Report. https://www.ams.usda.gov/services/transportation-analysis/gtr-datasets
3. RFD-TV, "Record Grain Movement Meets Rising Harvest Freight Costs," Aug 4, 2026 — source of the $25.62/ton St. Louis spot anchor and the July rail fuel-surcharge figure. https://www.rfdtv.com/record-grain-movement-meets-rising-harvest-freight-costs
4. USDA Agricultural Marketing Service, Barge Dashboard, Open Ag Transport Data. https://agtransport.usda.gov/stories/s/Barge-Dashboard/965a-yzgy/

All percent-of-tariff values are as published by USDA AMS. Peak identification, consecutive-decline counts, year-over-year comparisons and the dollar and cents-per-bushel conversions are this publication's own arithmetic on that series and are labelled as derived where they appear.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

The number that describes US grain storage is not the capacity, it is the derivative. Between 2019 and 2025 the country added 337 million bushels of grain storage — a figure that reads as growth until it is set against the period before it, when the industry added an average of 349 million bushels every single year (farmdoc daily, University of Illinois, Feb 9, 2026). Six years of building produced less new space than one ordinary year used to. That is not a slowdown in a growth rate; it is a near-complete stop, and it arrived in the same decade as the largest crops the country has ever produced.

1. The build rate collapsed, and the level hides it

What's new: Total US grain storage capacity was just over 25 billion bushels in 2019 and approximately 25.3 billion in 2025. Measured as a level, capacity looks stable and adequate. Measured as a rate of change against its own history, it has fallen off a cliff.

Evidence:

Source: farmdoc daily (Joe Janzen, Department of Agricultural and Consumer Economics, University of Illinois), "US Grain Storage Capacity Growth Has Stopped," Feb 9, 2026. Per-year figure for 2020–2025 is 337 ÷ 6. Bars scaled to the 2000–2019 rate.

The build rate fell 83.7% (Crop Root Zone calculation). Had the pre-2020 trend continued, capacity would stand at approximately 27.5 billion bushels today rather than 25.3 — a shortfall against trend of about 2.2 billion bushels (farmdoc daily, Feb 9, 2026).

Ground Truth: A stock variable that stops growing looks fine right up until the flow through it rises, and that is exactly the configuration here. Capacity fell only 8% short of trend in percentage terms, which is why the level reads as unremarkable — but the missing 2.2 billion bushels is roughly the entire buffer, and buffers are not consumed proportionally. They are consumed all at once, in the eight weeks of harvest, in whichever region had the big crop. The percentage understates the operational consequence because storage is a local, seasonal constraint being measured with a national, annual number.

2. How long the gap takes to close

What's new: The shortfall against trend and the current build rate are both known, which makes the closure time a division problem rather than a forecast.

Evidence:

Source: Crop Root Zone calculation from farmdoc daily figures, Feb 9, 2026. The 2.2 billion bushel gap is capacity against its own pre-2020 trend, not against a stated requirement.

39 years

How long it takes to close the gap between US grain storage capacity and its own pre-2020 trend, at the rate the industry has actually been building since 2020. (Crop Root Zone calculation from farmdoc daily, Feb 9, 2026)

Ground Truth: Thirty-nine years is not a forecast and should not be read as one — build rates respond to prices, and a sufficiently painful harvest would move it. What the number does is convert a vague concern into a statement about what would have to change. Closing this gap on any timescale that matters to a currently-operating farm requires the build rate to return to something near its historical level, and there is no evidence in the six-year record that it is doing so. The correct planning assumption for the rest of this decade is therefore that the shortfall is permanent, not that it is being worked off. That is a different assumption from the one most storage decisions are being made under, and it changes the sign on the return to owning a bin.

3. The cushion is a third of its normal size

What's new: Surplus capacity — the share of total storage not needed to hold production and carry-in — has fallen to 5%, against a historical average since 2000 of about 15%.

Evidence: Utilization on December 1, 2025 ran 80% on-farm and 65% off-farm, and crop production came closer to total capacity than in any year since 1988 (farmdoc daily, Feb 9, 2026).

Source: farmdoc daily, Feb 9, 2026, drawing on USDA NASS grain stocks and capacity data as of Dec 1, 2025.

On a 25.3 billion bushel base, 5% surplus is roughly 1.27 billion bushels of slack against a 15% norm of about 3.8 billion — the system is operating about 2.5 billion bushels below its customary cushion (Crop Root Zone calculation).

Ground Truth: The split between the two utilization figures is more informative than either number and it points somewhere unexpected. On-farm bins ran 80% full while commercial elevators ran 65%. The tighter constraint is on the farm, not at the elevator — which inverts the usual harvest narrative about elevator queues and dump pits. What that combination describes is grain being held on farm at high occupancy while commercial space sat a third empty, and the reason is that on-farm storage is not primarily a capacity decision, it is a marketing decision: a full bin in December is a farmer declining to sell. So the binding constraint in the current system is not physical space in aggregate. It is space at the location where the marketing option lives, and the two are not substitutes even though the capacity table adds them together.

4. Temporary storage is not the pressure valve, it is the margin

What's new: Ground piles, bunkers and bags are routinely described as the flexible relief for a big harvest. On the arithmetic above they are better described as the structural marginal unit of US grain storage.

Evidence: If permanent capacity is 2.2 billion bushels below trend, is being added at under 57 million bushels a year, and the cushion has fallen from 15% to 5%, then the difference between what the crop needs and what permanent structures provide is being closed by something. Temporary storage is what is closing it, and it is doing so every year rather than in exceptional ones.

Source: Crop Root Zone analysis. Capacity and utilization context from farmdoc daily, Feb 9, 2026.

Ground Truth: Treating temporary storage as a pressure valve rather than as permanent marginal capacity produces a specific and expensive error, and it is an error about the carry rather than about the grain. A ground pile cannot be held indefinitely, which means the grain in it has to be priced against a delivery window that the storage itself is setting — not against the window the market is paying for. This desk established earlier this month that the corn spread most often quoted as a storage carry is not one, and that the relevant figure is considerably smaller (Crop Root Zone, TRZ-0122, Aug 24, 2026). Combine the two findings and the position is sharper than either alone: a grower is being asked to capture a carry that is smaller than advertised, using storage that cannot hold long enough to capture it. The marginal bushel in the US system is stored in the one form that structurally cannot execute the marketing strategy the storage exists to enable.

5. What it means for basis at this harvest

What's new: A storage system running at a 5% cushion transmits a large crop into basis rather than into inventory.

Evidence: When space is abundant, a bigger-than-expected crop is absorbed and basis weakens modestly. When the cushion is 5% and on-farm bins are already the tighter constraint, the same surprise has to clear through price at the point of delivery, because there is nowhere to put it.

Source: Crop Root Zone analysis from the capacity and utilization figures in farmdoc daily, Feb 9, 2026.

Ground Truth: This is where the storage arithmetic becomes a decision rather than a statistic, and the decision is about when rather than whether. In a 5%-cushion system, harvest basis is not a market signal about demand — it is a rationing mechanism for space, and it is at its widest exactly when the most grain needs a home. The grower most exposed is the one whose only alternative to selling at harvest is temporary storage, because that grower faces the widest basis and holds the option least able to wait it out. The practical read is that space secured before harvest is worth more this year than the storage rate on it suggests, and that commercial space — running a third empty in December while farm bins ran 80% full — is the underused half of a system that is described as full.

6. Limits

The capacity, utilization and trend figures come from a single analysis (farmdoc daily, Feb 9, 2026) drawing on USDA NASS data as of December 1, 2025. That is an eight-month-old vintage at publication and it precedes the 2026 harvest entirely. The 2.2 billion bushel trend gap is a counterfactual — capacity against an extrapolation of its own 2000–2019 growth — and not against any stated requirement; a reader who believes the pre-2020 build rate was itself excessive would regard the gap as smaller or as no gap at all. Nothing here establishes that 15% surplus is the correct cushion, only that it is the historical one.

Second, the 39-year closure figure divides a counterfactual gap by a six-year average build rate. Both terms are uncertain and the rate in particular is a short series that includes pandemic-era construction disruption, which plausibly depresses it relative to an underlying trend. The finding that the build rate collapsed is robust; the specific 39 is not, and a reader should treat it as an order of magnitude.

Third, section 5's basis argument is a mechanism, not a measurement. This piece does not quantify how much harvest basis widening a one-point reduction in surplus capacity produces, and that relationship is regional, crop-specific and not established here.

References

1. farmdoc daily (Joe Janzen, Department of Agricultural and Consumer Economics, University of Illinois), "US Grain Storage Capacity Growth Has Stopped," Feb 9, 2026. https://farmdocdaily.illinois.edu/2026/02/us-grain-storage-capacity-growth-has-stopped.html
2. USDA National Agricultural Statistics Service, grain stocks and storage capacity data as of Dec 1, 2025, as compiled in reference 1.
3. Crop Root Zone, TRZ-0122, "The 25-Cent Corn Spread Everyone Is Quoting Is Not a Storage Carry. The 14¾-Cent One Is," Aug 24, 2026.
4. Crop Root Zone, TRZ-0119, "Corn Left the Country at a Record August Rate on 34 Fewer Barges. Every Other Mode Is Bid Over," Aug 24, 2026 — for the transport side of the same harvest logistics constraint.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Canada's counter-tariffs on US goods take effect at 12:01 a.m. on September 8, 2026, announced August 25 and published August 26, at rates of 15%, 25% and 50% set to mirror the US rate on the same good (GHY International, Aug 26, 2026). Agricultural equipment is on the sector list. But the tariff-item detail says something the sector list does not: tractors, combines, planters, tillage equipment and sprayers were not broadly included. What was included is narrower and stranger — certain harvesting machinery such as balers at 25%, parts at 15%, and livestock and other semi-trailers at 25% (DTN/Progressive Farmer, Aug 25, 2026). Canada exempted the machine and taxed the machine's parts, three weeks before harvest.

1. What was announced, and two different totals

What's new: The measures match US Section 232 and Section 338 actions rate-for-rate rather than applying a single retaliatory rate, so the schedule is a patchwork rather than a bloc.

Evidence: The two sources this desk read give different totals for the trade covered, and both are reported here rather than one being chosen.

Source: GHY International trade-compliance summary, published Aug 26, 2026; DTN/Progressive Farmer Washington Insider, Aug 25, 2026. This desk did not reconcile the two figures — see Limits.

The agriculture and food lines where a dollar value was published:

Source: DTN/Progressive Farmer, Aug 25, 2026. Bars scaled so the largest export value is ten blocks. No dollar value was published for the machinery or parts lines in either source read this run.

2. What farm equipment actually got hit

What's new: The exclusions are the story. Both sources agree that the September 8 block does not broadly cover tractors, combines, planters, tillage equipment or sprayers.

Evidence: What is covered, on the two independent readings available:

Source: DTN/Progressive Farmer, Aug 25, 2026; RealAgriculture, Aug 2026, via published summary. See Limits on the scope of "parts."

The scope of the parts line is the one thing this desk could not pin down. Both readings place "15% on parts" immediately after the harvesting-machinery entry, which supports the narrow reading — parts of the tariffed harvesting machinery. The broad reading, parts of farm machinery generally, would be a materially bigger measure. Canada's own tariff-item list is the authority and it did not serve to this desk this run. Section 3 is written to the broad reading, because that is the conservative one: if the narrow reading is correct, every parts figure below is an overstatement.

3. Over a machine's life, the exemption is worth about seven times the parts tariff

What's new: The intuitive read — "a small rate on parts you must buy forever beats a big rate on a machine you buy once" — is wrong by roughly a factor of seven, and the ASABE repair factors used by every extension machinery budget say so.

Evidence: ASABE publishes accumulated repair cost as a share of new list price. For a two-wheel-drive tractor up to 150 hp, average repair cost runs 0.84% of new list price per 100 hours of use (Mississippi State University Extension, Farm Machinery Cost Calculations). At the standard budgeting assumption of 300 hours a year over a 10-year life — 3,000 hours, or thirty 100-hour blocks — accumulated lifetime repairs come to 25.2% of list price.

Now put the two tariffs on that one ruler, expressed as a percentage of the machine's list price:

Source: repair factor from Mississippi State University Extension, Farm Machinery Cost Calculations, citing ASABE; tariff rates from DTN/Progressive Farmer, Aug 25, 2026. Percentages are Crop Root Zone calculations. *The avoided-tariff figure is shown as 25.20% rather than 25.00% only to place it on the same axis as a 10-year total; the tariff itself is a flat 25% of list, paid at purchase.

The third row exists because a repair bill is not all parts. If parts are 60% of a repair invoice and labour the rest — a common workshop split, and an assumption, not a sourced figure — the tariffed base shrinks to 15.1% of list and the lifetime cost to 2.27% of list, about 0.23% a year.

21 percentage points

The gap between a 25% tariff on the machine and a 15% tariff on ten years of its parts, both expressed as a share of list price. The exemption is much the larger relief. (Crop Root Zone calculation from ASABE repair factors via Mississippi State University Extension and the Sept 8 rate schedule)

Ground Truth: This desk went into the calculation expecting the opposite result and is reporting the one it got. A parts tariff feels heavier than a machine tariff because it never stops, but "never stops" is doing less work than it seems: accumulated repairs over a full ten-year service life are only about a quarter of what the machine cost, so a 15% rate on them cannot approach a 25% rate on the whole thing. The headline that most farm equipment escaped Canada's list is substantively true, not cosmetic — and any analysis arguing the parts line quietly undoes the exemption needs to show a repair factor several times the ASABE one to make that case.

4. Why the parts line still matters more than 0.23% a year suggests

What's new: The annual incidence is trivial. The timing and the elasticity are not, and they run in opposite directions to the rate.

Evidence: Three structural points, none of which the rate captures.

The base is captive; the exempt base is not. A Canadian farmer facing a 25% duty on a new US combine has real options — defer a year, buy used, or buy from a non-US manufacturer. That is exactly why exempting whole machines protects Canadian buyers cheaply: the demand was elastic anyway. A farmer whose US-built combine needs a specific part in October has none of those options inside the relevant time window. The rate on the captive base is 15%; the rate on the elastic base is zero.

The date is three weeks before it hurts most. September 8 is not a neutral start date for a parts measure. Parts demand is not spread evenly across the year; it spikes during harvest, when a machine down for a day costs a great deal more than the part. A tariff that begins the week harvest ramps is a tariff that begins at the moment its base is least able to defer.

Balers at 25% is its own season. Hay and forage equipment does not share the row-crop calendar, and a 25% duty on balers lands after most of the North American hay season rather than before it — which makes the baler line, unlike the parts line, mostly a 2027 purchase-decision event.

Ground Truth: Read the two measures by what they can change rather than by what they cost. The whole-machine exemption is worth 21 points and changes a decision a Canadian farmer was free to make anyway. The parts tariff is worth about a quarter of a point a year and changes no decision at all — it is simply collected, from a fleet that cannot substitute, at the season when substitution is hardest. Small, unavoidable and well timed is a different instrument from large, avoidable and deferrable, and the sector-level headline conflates them.

5. What it means on the US side of the line

What's new: For US row-crop growers, this round is not a demand event. Corn, soybeans, wheat, beef, pork and poultry were largely spared.

Evidence: The published breakouts put US agricultural exposure in seafood, dairy ingredients, cheese, honey, molasses, bakery mixes and wood — not in the row-crop complex. A US corn or soybean grower's exposure to the September 8 schedule is therefore on the cost and equipment side, through the North American machinery and parts supply chain, and not through the price of what they sell.

That supply chain is genuinely integrated, which is the part worth watching. Aftermarket parts are the higher-margin end of the equipment business, and a duty applied at the Canadian border to US-origin parts is a duty on the segment manufacturers depend on to carry them through weak new-equipment years. Whether that shows up as lower Canadian volumes or as absorbed margin is a question the next set of quarterly disclosures will answer, not one this desk can settle in August.

6. Limits

The primary list did not serve. Canada's Department of Finance tariff-item list returned an HTTP 403 to this desk on repeated attempts, as did RealAgriculture's report. Every rate and inclusion above rests on two secondary summaries that agree with each other. The tariff-item level governs, and importers should read the official schedule rather than any of this.

The two published totals disagree — $27.6 bn against about $20 bn — and this desk did not reconcile them. They may measure different baskets or different base years. Both are reported; neither is adopted.

No dollar value for the machinery lines was found, so nothing here sizes the measure in absolute terms. All of Section 3 is expressed as a share of a machine's list price for that reason.

The repair factor is one class of machine. ASABE's 0.84% per 100 hours is for a two-wheel-drive tractor up to 150 hp. Combines and high-horsepower four-wheel-drive tractors carry different factors, and a combine's accumulated repair share of list price is generally higher than a small tractor's. If the true factor for the relevant fleet is double the one used, lifetime parts rise to about 50% of list and the tariff to 7.6% of list — the conclusion narrows from a factor of seven to a factor of about three, and survives.

The 300 h/yr, 10-year life is a budgeting convention, not a measurement of any particular farm, and the 60% parts share of a repair invoice is an explicit assumption.

The checkable test is the tariff-item list itself: whether "parts" in the September 8 schedule is scoped to harvesting machinery or to farm machinery generally decides whether Section 3's figures are right or roughly a third of what is written.

References

1. GHY International, "Canada to Impose New Counter-Tariffs on U.S. Goods Effective September 8, 2026," published Aug 26, 2026. Rates of 15/25/50% mirroring US Section 338 rates; ~$27.6 bn of imports covered; effective 12:01 a.m. Sept 8, 2026. https://www.ghy.com/trade-compliance/canada-counter-tariffs-us-goods-september-2026/
2. DTN/Progressive Farmer, Washington Insider, "US-Canada Trade Fight Escalates With New Canadian Tariffs on a Range of Products," Aug 25, 2026. ~$20 bn covered; fish and aquaculture 25% ($926 m); cheese and curd 25% ($135 m); whey 50% ($82.6 m); wood products $1.97 bn; harvesting machinery such as balers 25%, parts 15%, livestock trailers 25%. https://www.dtnpf.com/agriculture/web/ag/columns/washington-insider/article/2026/08/25/canada-hits-back-tariffs
3. RealAgriculture, "Most farm equipment left off Canada's $27.6 billion retaliatory tariff list," August 2026 — tractors, combines, planters, tillage equipment and sprayers not broadly covered. Read via published summary; the page returned HTTP 403 to direct retrieval. https://www.realagriculture.com/2026/08/most-farm-equipment-left-off-canadas-27-6-billion-retaliatory-tariff-list
4. Canada, Department of Finance, "List of products from the United States subject to counter-tariffs effective September 8, 2026." Primary source; returned HTTP 403 to this desk. https://www.canada.ca/en/department-finance/news/2026/08/list-of-products-from-the-united-states-subject-to-counter-tariffs-effective-september-8-2026.html
5. Mississippi State University Extension Service, Farm Machinery Cost Calculations (P3543), citing ASABE repair factors: two-wheel-drive tractor up to 150 hp, average repair cost 0.84% of new list price per 100 hours of use; standard assumption of 300 hours per year over a 10-year life. https://extension.msstate.edu/publications/farm-machinery-cost-calculations

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Mont Belvieu propane settled at $0.723 per gallon on August 18, up from $0.651 on August 5 — an 11.1% gain across eleven trading sessions, arriving in the weeks immediately before the drying season (EIA daily spot series). Two weeks ago this desk recorded the same benchmark at $0.655 for the week ending August 7 and noted it was roughly 9% below the year-ago level, concluding that the exposure this fall was the delivery calendar rather than the price (TRZ-0094, Aug 13, 2026). The price half of that read has been overtaken. The discount is gone. What has not changed is that the price is still the least important number on the page.

1. The move, and what it did to the year-ago comparison

What's new: Eleven sessions took the benchmark up 11.1% and closed a nine-point discount to year-ago.

Evidence: The daily series, as published:

Source: EIA Mont Belvieu, TX propane spot price FOB, daily series, as compiled and published by PropanePricePerGallon.com. The bar column is scaled from zero, which is why the differences look modest — that is deliberate, and the point of section 3.

The year-ago comparison follows arithmetically. If $0.655 was approximately 9% below the year-earlier level, the year-ago level was approximately $0.720. At $0.723, the benchmark is now level with where it stood a year ago, having been nine points cheaper eleven sessions earlier.

$0.723/gal

Mont Belvieu propane on August 18, 2026 — up 11.1% from August 5 and, on the implied year-ago level, no longer carrying any discount at all. (EIA daily spot series; year-ago level is a Crop Root Zone calculation from the 9% figure recorded Aug 13, 2026)

One caution belongs here rather than in a footnote at the bottom: Mont Belvieu is a hub, not a farm. No grower buys propane at this price. A delivered farm price sits materially above it and moves with a lag. The hub is useful as the leading indicator of what a fall contract will be offered at, and as the input to the arithmetic below — it is not a quote.

2. What a point of moisture costs at this price

What's new: With a published energy coefficient and a published cost formula, the hub price converts directly into cents per bushel per point.

Evidence: High-temperature drying consumes roughly 0.018 to 0.020 gallons of propane per bushel per point of moisture removed, with a range of 0.010 to 0.025 depending on the system and the outdoor temperature; a two-year Iowa on-farm case study measured 0.019 (Iowa State University Extension Ag Decision Maker; University of Minnesota Extension; North Dakota State University Extension). The physical basis is 0.72 pounds of water per point per bushel, about 2,500 Btu to remove a pound of water in a high-temperature dryer, and 91,500 Btu per gallon of propane.

Iowa State's cost rule for propane in high-temperature drying is cost per bushel-point = 0.022 × propane price per gallon — the coefficient above 0.019 accounting for system inefficiency.

Source: coefficients per Iowa State University Extension Ag Decision Maker, UMN Extension and NDSU Extension; prices per EIA. The per-bushel-point figures are Crop Root Zone calculations at the hub price, and a farm-delivered price would raise every number in the right-hand columns proportionally.

On a 220 bu/ac crop dried five points, the entire 11.1% rally is worth $1.74 per acre: 1,100 bushel-points × $0.00159.

It is worth setting that against a full-system benchmark rather than a fuel-only one. A 2025 extension estimate put total high-temperature drying cost at roughly $0.04 per bushel-point with propane at $1.50 per gallon and electricity at $0.14 per kilowatt-hour — a propane price more than double today's hub level. Working backwards, propane at $1.50 contributes $0.033 per bushel-point on the 0.022 rule, leaving roughly $0.007 for electricity and everything else. At today's $0.723 hub price the propane share falls to $0.0159 and the non-fuel share does not move, so fuel drops from about 83% of the system cost to roughly 69% of a much smaller total. The practical implication is that the cheaper propane gets, the less of the drying decision the propane market controls — and it has been getting cheaper on a two-year view even after this month's rally.

3. The two numbers that dwarf it

What's new: Points removed and final moisture target are each worth multiples of the price move, and both are decisions rather than market outcomes.

Evidence: At $0.723 propane and the 0.022 rule, one point on a 220 bu/ac crop is 220 × $0.01591 = $3.50 per acre. That is 2.0 times the whole rally.

Over-drying is larger still, because it costs weight as well as fuel. Water is sold as grain at the market moisture of 15.5%; drying to 13% instead of 15% removes 2 points of sellable weight, and the shrink is (15 − 13) ÷ (100 − 13) = 2.299%.

Source: Crop Root Zone calculation. Propane cost rule per Iowa State Extension at the Aug 18 Mont Belvieu price; corn valued at the Dec 2026 settlement of $5.15½ (Pro Farmer, Aug 24, 2026); shrink computed on the physical water basis, excluding any elevator handling shrink, which would increase it.

Ground Truth: Two points of over-drying costs nineteen times what an 11% rally in the propane benchmark costs, and roughly four-fifths of that is not fuel at all — it is grain weight given away. The propane market is the number that will be discussed at every elevator counter this fall because it is the one that moves visibly and gets quoted. The moisture meter on the dryer is the one that decides the bill. A dryer running two points hot on 1,000 acres is a $33,000 decision made by a setting, against a $1,700 decision made by the market.

4. How eastern lodging becomes a propane bill

What's new: The flood damage in the eastern Belt transmits into fuel consumption through a channel that has nothing to do with the propane market.

Evidence: Lodged and downed corn requires reduced ground speeds — published guidance runs 1.5 to 2.0 mph against normal harvest speeds — which this desk converted to 2.25 to 3.00 times the machine hours per acre, turning 7.6 harvest days per thousand acres into 17.2 or 22.9 against an eastern-Belt window of 30 to 40 days (TRZ-0123, Aug 24, 2026).

An operation facing that arithmetic has one lever: start earlier. Starting earlier means harvesting wetter, and harvesting wetter means more points through the dryer. The chain runs lodging → compressed effective window → earlier start → higher harvest moisture → more gallons, and not one link in it involves the price of propane.

The relevant question is whether it is worth it, and at these numbers it is not close. Four extra points on 220 bu/ac costs $14.00 per acre in fuel. An illustrative 5% harvest loss from waiting — 11 bushels at $5.155 — is $56.71 per acre, four times the fuel penalty. That 5% is an illustration, not a measurement, and the real figure varies enormously with how far the corn is down; what does not vary is the sign, and the sign is the part a grower has to act on in the first week of October.

Ground Truth: For a grower in the lodged corridor, the fuel arithmetic and the harvest-loss arithmetic point the same way and the fuel side is the smaller of the two by a wide margin. Start early, pay the propane, accept the wetter corn. The exposure this creates is not financial but physical: an eastern Belt that all starts early at once puts its drying demand into a narrower window than usual, on a delivery system that this desk already flagged as calendar-constrained rather than supply-constrained. The 11% price move is a distraction. The queue at the dryer is not.

5. Inventory says comfortable; the price says something else

US propane inventories stood roughly 31% above their five-year average for the week ending August 7, 2026 (EIA), and Midwest stocks had built 13% over five weeks to 25.0 million barrels as of early August. That is the setup in which prices ordinarily sit flat or drift lower into the fall build.

They did the opposite. An 11% rally against a 31% inventory surplus is not a domestic scarcity signal, and we are not going to assert a cause we have not verified — export pull and fractionation economics are the usual candidates, and neither was checked this run. What can be said is narrower and still useful: the price is not being set by how much propane is sitting in the Midwest, which means Midwest inventory is a poor forward indicator of what a fall contract will cost.

Limits. Every dollar figure here is built on the hub price and a 220 bu/ac yield, and neither is a farm's actual number — a delivered propane price and a different yield scale the whole table. The 0.019 and 0.022 coefficients are extension figures with a stated range from 0.010 to 0.025 gallons per bushel-point; at the low end the fuel numbers halve and at the high end they rise by a third, which changes the magnitude of every fuel comparison above but not the conclusion that shrink dominates it. The shrink calculation is physical water only and excludes elevator handling shrink, so the over-drying figure is conservative. The 5% harvest-loss illustration in section 4 is explicitly not a measurement. And the price series used here ends August 18 — the most recent daily reading we could retrieve — so the rally may have continued or reversed in the week since, and Wednesday's EIA weekly report is the next checkable point.

References

1. US Energy Information Administration, Mont Belvieu, TX propane spot price FOB, daily series, Aug 5–18, 2026, as compiled by PropanePricePerGallon.com. https://propanepricepergallon.com/mont-belvieu-spot-price/
2. US Energy Information Administration, weekly propane stocks, week ending Aug 7, 2026.
3. Iowa State University Extension and Outreach, Ag Decision Maker A2-31, "Estimating the Cost for Drying Corn." https://www.extension.iastate.edu/agdm/crops/html/a2-31.html
4. University of Minnesota Extension, "Energy costs for corn drying and cooling." https://extension.umn.edu/corn-harvest/energy-costs-corn-drying-and-cooling
5. North Dakota State University Extension, "Drying High-Moisture Corn Can Be Tricky" — 0.72 lb water per point per bushel, 2,500 Btu/lb, 91,500 Btu/gal. https://www.ag.ndsu.edu/news/newsreleases/2009/oct-26-2009/drying-high-moisture-corn-can-be-tricky/
6. Pro Farmer, "After the Bell," Aug 24, 2026 — December 2026 corn settlement of $5.15½.
7. Crop Root Zone, TRZ-0094, Aug 13, 2026 — propane supply and the delivery calendar; and TRZ-0123, Aug 24, 2026 — machine-hour arithmetic for lodged corn.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Sulfur has been the growth story in Midwestern nutrition for a decade. Atmospheric deposition fell, grain removal rose, and sulfur-bearing products — ammonium sulfate, gypsum, sulfur-fortified MAP, elemental sulfur, thiosulfates — moved from specialty to routine. A peer-reviewed Illinois dataset published in Agronomy Journal offers the largest single test of whether that spending finds its target, and the answer is uncomfortable in a specific and useful way: the fertilizer did what fertilizer is supposed to do to a soil test, and did almost nothing to a yield.

1. The trial set, and what it found

Preza Fontes and colleagues ran 40 field experiments across 23 Illinois counties over three seasons — 18 replicated small-plot trials and 22 on-farm strip trials (Preza Fontes et al., Agronomy Journal 117(5):e70169, Sep 9, 2025, open access). Rates ran from zero to 52 kg S/ha in the small plots and zero against 34 kg S/ha on-farm. Six sources were compared at a common rate.

Source: Preza Fontes, Jones, Greer, Schaefer, Kaiser & Fernández, Agronomy Journal 117(5):e70169, 2025. Conversions at 0.8922 lb/acre per kg/ha and 15.93 bu/acre per Mg/ha (corn, 56 lb/bu).

The six sources tested — ammonium sulfate (21-0-0-24S), elemental sulfur (0-0-0-90S), gypsum, MAP-10S, MAP-10S+Zn and MAP-15S — cover essentially the whole commercial sulfur shelf, including both the immediately available sulfate forms and the elemental form that requires microbial oxidation before a plant can use it.

The outcome tally is the thing to sit with.

Source: Preza Fontes et al. (2025) — 18 small-plot trials (no increases, three significant decreases) plus 22 on-farm trials (two increases, two decreases).

Significant yield decreases outnumbered significant increases by two and a half to one. The two responsive sites were genuinely responsive — Site 24 gained 3.2 Mg/ha (about 51 bu/acre, +30.5%) and Site 32 gained 1.3 Mg/ha (about 21 bu/acre, +14.7%) — while the decreases ran 0.3 to 0.8 Mg/ha, roughly 5 to 13 bu/acre.

Ground Truth: Do not read the five decreases as sulfur toxicity. Read the whole distribution instead. When a treatment produces a handful of significant results scattered on both sides of zero across 40 trials, the honest description is that the average effect is indistinguishable from zero and the tails are field-year noise. That is a more damaging finding than a clean negative, because it means the two spectacular successes are not evidence of a general effect either — and the 51-bushel site is exactly the result a product brochure would carry.

2. The diagnostic responded to the fertilizer and not to the need

Here is the part that is not a null result, and it is the reason this study matters more than its yield table suggests.

Sulfur application worked on every measurement except the one that pays. Soil test sulfur in the top 15 cm rose from 7.0–10.7 mg/kg in the untreated controls to 10.2–16.5 mg/kg at the 52 kg/ha rate. Earleaf sulfur concentration was significantly affected by rate at 7 of 13 measured small-plot sites. The nutrient went into the soil, the soil test detected it, the plant took it up, and the tissue test detected that too.

And none of it predicted yield. The authors state the conclusion plainly: soil sulfur and earleaf sulfur concentration "are unreliable in predicting yield response in the upper US Midwest."

That combination has a specific meaning that is worth spelling out, because it is easy to file this as "the soil test is imprecise" and move on. A test that reliably moves when you apply the input, but does not correlate with the response to that input, is not an imprecise test. It is a test of the wrong thing. It measures how much sulfur is present. The decision requires knowing whether the crop is short of sulfur — which depends on mineralization from organic matter through the season, on rooting depth reaching sulfate that has leached below the sample zone, and on the timing of supply against demand. None of those are captured by a single concentration in the top six inches in spring.

5% vs 82.5%

Trials in which sulfur fertilizer significantly raised corn yield, against trials in which it did nothing measurable — while raising the soil test in essentially all of them. (Preza Fontes et al., Agronomy Journal, 2025)

The failure mode this creates in practice is self-reinforcing. A grower who soil-tests low, applies sulfur, and re-tests will see the number improve. That feels like confirmation. It confirms only that sulfate was added.

3. Which half of this study travels to 2026, and which does not

The obvious objection is vintage, and it deserves to be taken seriously rather than waved at. The trials ran in 2009–2011. Sulfur's whole agronomic case rests on atmospheric deposition having collapsed, so trials run fifteen-plus years ago may simply have been conducted in a more sulfur-rich environment than today's.

Checked against the deposition record, that objection has real force. CONUS-average total sulfur deposition fell from 5.3 kg S/ha/yr in 2002 to 1.8 in 2017 — about 66% over fifteen years (Benish et al., Atmospheric Chemistry and Physics 22:12749–12767, 2022). The decline decelerated after 2010 relative to 2002–2009, but it did not stop. Interpolating to the trial midpoint puts deposition near 3.4 kg S/ha/yr during the study, roughly twice the 2017 national average, with further decline since.

So the two findings have to be separated, because they age differently.

This table is this publication's assessment, not a claim made by the cited authors.

Ground Truth: The useful takeaway is not "don't apply sulfur." It is that the industry has expanded sulfur use for fifteen years while running on a diagnostic that its own largest validation study says does not work — and that lower deposition, which makes the application more likely to be justified, does nothing at all to fix the targeting. More deficiency plus an invalid test means more sulfur applied to fields that do not need it, alongside the fields that do. The waste and the shortfall grow together.

4. What to use instead, honestly stated

This is where the article has to stop short of what a reader wants, because the literature does.

Extension guidance has moved to rate-of-thumb rather than test-driven rates. The common recommendation is roughly 15 lb S/acre on fine-textured soils and 25 lb/acre on coarse or sandy ground, and Iowa State says directly that soil tests are not a reliable way to predict sulfur deficiency in Iowa soils. That is an admission that the decision is being made without a diagnostic, not with one.

Texture and organic matter are doing the predictive work. Sandy, low-organic-matter, well-drained soils mineralize less and lose sulfate more readily. That is a field property a grower already knows and does not need to test for annually.

Source matters, and one distinction is settled. Elemental sulfur requires microbial oxidation to sulfate before roots can use it, which takes time and depends on temperature and moisture. Sulfate-based sources are available immediately. This is chemistry rather than a field result, and it means an elemental product applied in-season to a visibly deficient crop is unlikely to arrive in time regardless of rate.

The strip trial is the only diagnostic with a clean logic. If the test cannot tell you whether your field responds, leaving an unfertilized strip and weighing it can. The study's own on-farm arm is 22 instances of exactly that design, and it is the reason the dataset can say anything at all.

The authors' own recommendation is narrower and worth quoting in substance: future work should incorporate other organic and inorganic soil sulfur fractions, because the fractions currently measured are not the ones that matter. Until that exists, sulfur is a nutrient bought on inference.

5. Limits

One state, three seasons. Forty trials is large for this kind of work and it is still Illinois in 2009, 2010 and 2011. Illinois soils are on the higher-organic-matter end of the Corn Belt, which is where mineralization would be expected to make sulfur least limiting. The finding may travel poorly to sandier ground in Wisconsin, Michigan or the eastern Dakotas — and the direction of that bias favours sulfur, not against it.

"Significant" is a statistical threshold, not an economic one. A trial can fail significance and still carry a yield effect worth money at the rates involved, particularly in the strip trials where replication is weaker. The 82.5% "no significant effect" group is not 82.5% of fields with zero response; it is 82.5% where the response could not be distinguished from noise.

No economics are attempted here. A cost-per-acre for each source would let this be carried to a break-even, and no public-tier price for ammonium sulfate at a defensible retail basis could be independently sourced this cycle. The 08-19 DTN retail board carries eight products and ammonium sulfate is not among them. Rather than substitute a wholesale or aggregate assessment on a different basis, that calculation is left undone and flagged — it is the obvious next piece, and it needs a price this desk can stand behind.

The deposition comparison is national. The 5.3 → 1.8 kg S/ha/yr series is a CONUS average, and the Upper Midwest carried among the highest sulfur concentrations in that dataset. An Illinois-specific deposition series would sharpen section 3 considerably and is not reproduced here. The interpolation to the 2009–2011 trial window is also linear, while the underlying paper reports a steeper decline before 2010 than after it — which means the true deposition level during the trials was probably somewhat below the 3.4 kg S/ha/yr figure used, and the gap to today correspondingly smaller than stated. That error runs in the direction of making the trials more applicable, not less, and it is noted here rather than left to favour the argument.

References

1. Preza Fontes, G., J. Jones, K.D. Greer, D. Schaefer, D. Kaiser and F.G. Fernández, "Corn response to sulfur fertilizer rate and source in Illinois," Agronomy Journal 117(5):e70169, Sep 9, 2025 (open access). https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.70169
2. Benish, S.E., J.O. Bash, K.M. Foley, K.W. Appel, C. Hogrefe, R. Gilliam and G. Pouliot, "Long-term regional trends of nitrogen and sulfur deposition in the United States from 2002 to 2017," Atmospheric Chemistry and Physics 22:12749–12767, 2022. https://doi.org/10.5194/acp-22-12749-2022
3. Zhang, Y., R. Mathur, J.O. Bash, C. Hogrefe, J. Xing and S.J. Roselle, "Long-term trends in total inorganic nitrogen and sulfur deposition in the US from 1990 to 2010," Atmospheric Chemistry and Physics 18:9091–9106, 2018. https://doi.org/10.5194/acp-18-9091-2018
4. Iowa State University Extension, Integrated Crop Management, "Who needs sulfur? You might need sulfur." https://crops.extension.iastate.edu/post/who-needs-sulfur-you-might-need-sulfur
5. Purdue University Extension, "Sulfur Deficiency in Corn," AY-379-W. https://www.extension.purdue.edu/extmedia/AY/AY-379-W.pdf

Unit conversions throughout use 0.8922 lb/acre per kg/ha and 15.93 bu/acre per Mg/ha for corn at 56 lb/bu. Converted values are labelled alongside the metric figures as published. The assessment in section 3 of which findings survive the study's vintage is this publication's own and is identified as such.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

An input that returns six times its cost, has been commercially available since the 1970s, and is still not used on most fall-applied nitrogen acres is telling you something — and what it is telling you is usually about the number rather than about the growers. A peer-reviewed meta-evaluation of nitrapyrin across Midwestern corn trials found that anhydrous ammonia plus nitrapyrin produced a yield advantage of over 6.4% compared with anhydrous ammonia alone (Wolt, Nutrient Cycling in Agroecosystems, 2004). With anhydrous at $964/ton and December 2026 corn near $5.36 a bushel, that percentage converts into a per-acre figure large enough to be implausible as a standing offer. The figure is not wrong. It is conditional, and the condition is the entire decision.

1. The arithmetic that should raise the question

What's new: Converting the published percentage into dollars at current prices produces a return that no established input market sustains.

Evidence: At a 200 bu/acre yield goal:

Source: Crop Root Zone calculation. Yield advantage from Wolt (2004); corn price Barchart, Aug 28, 2026; inhibitor cost is a stated range, not a sourced quotation — see section 5.

For comparison, the same anhydrous nitrogen the inhibitor is protecting costs $964/ton, or $0.59/lb N (DTN/Progressive Farmer, Aug 19, 2026). A 180 lb N/acre rate is $106/acre of nitrogen. The claimed inhibitor benefit is 65% of the entire nitrogen bill.

Ground Truth: Any input showing a sustained 5× return in a competitive, well-informed, fifty-year-old market is mispriced, mismeasured, or misread — and the first two are far rarer than the third. Fertilizer retail is not a market in which a five-fold return sits unexploited for five decades. The productive response to this arithmetic is not to buy more inhibitor; it is to work out what the 6.4% is an average of, because a number that survives fifty years of commercial scrutiny while implying an unexploited bargain is almost always an average across states of the world that the reader is silently collapsing into one.

2. What the average is an average of

What's new: Nitrapyrin does not increase yield. It inhibits ammonia monooxygenase, the microbial enzyme catalysing the first step of nitrification from ammonium to nitrite, which delays the conversion of applied nitrogen into nitrate (Wolt, 2004; and see the mechanism literature at reference 3). Nitrate is the form that leaches below the root zone and that denitrifies under saturated conditions. Ammonium does neither.

Evidence: The product's entire mechanism is loss prevention. It follows that in a site-year where no nitrogen loss occurs, the inhibitor's yield effect is approximately zero — there was nothing to prevent. In a site-year with a substantial loss event, the effect is large.

Source: Crop Root Zone analysis of the mechanism described in Wolt (2004) and the nitrification-inhibitor literature at reference 3.

Ground Truth: The 6.4% is a conditional expectation being reported, and read, as an unconditional one. It is the average of a large number and a number near zero, weighted by how often each occurred across the trial set — and the trial set was assembled from research locations across diverse environments and many years, which is to say it embeds a particular frequency of loss years that has no reason to match the frequency on any specific farm. This is not a flaw in the meta-evaluation, which did exactly what a meta-evaluation should. It is a flaw in how the resulting number travels: a grower on well-drained ground reading "6.4%" is being quoted a probability-weighted average in which their own probability weight is much lower than the one used to compute it.

3. Inverting it: the probability that break-even requires

What's new: If the benefit only accrues in loss years, the decision is an insurance decision, and insurance decisions are solved for the probability that makes the premium fair.

Evidence: Let p be the probability that a given site-year produces a genuine nitrogen loss event, and B the yield benefit in such a year. Then expected benefit = p × B, and break-even requires p × B = cost.

The meta-evaluation gives the unconditional average, 6.4% of yield, or $68.61/acre. That average already equals pₘ × B for the trial set's own loss frequency pₘ. So B = $68.61 ÷ pₘ. Break-even for a grower whose own loss frequency is p requires:

p × ($68.61 ÷ pₘ) = cost, so p = pₘ × (cost ÷ $68.61).

The grower's break-even probability is the trial set's loss frequency scaled by the cost-to-average-benefit ratio.

Source: Crop Root Zone calculation from the Wolt (2004) average benefit and the stated cost range.

~15%

A grower needs a loss-year frequency roughly 15% of the frequency embedded in the published trials for a $10/acre inhibitor to break even. Below that it does not pay; above it, it pays and pays well. (Crop Root Zone calculation from Wolt, 2004)

Ground Truth: This is the finding, and it is more favourable to the product than section 1's scepticism implied. The break-even threshold is not 100% of the trial frequency — it is about 15% of it. So a grower whose ground is far less loss-prone than the average research site still clears break-even comfortably, and the apparent 5× return in section 1 is real for anyone at or near the trial average. What the inversion actually establishes is that the decision has a wide margin of safety and is therefore not close for most fields. The rare grower for whom it genuinely does not pay is on ground that loses nitrogen less than a sixth as often as the research network did — well-drained, low-rainfall, spring-applied — and that grower can identify themselves from their own field history without needing a trial.

4. Why fall application changes the weight, not the arithmetic

What's new: The probability p is not a constant of the farm. It is a function of application timing, and fall application moves it substantially.

Evidence: Fall-applied nitrogen sits in the soil through the entire overwinter and early-spring period before crop uptake begins — the window in which saturated soils and leaching-driving precipitation occur. Spring pre-plant or sidedress nitrogen is exposed for a fraction of that time. The mechanism is the same; the exposure is not.

Source: Crop Root Zone analysis. Exposure windows are the interval between application and active crop nitrogen uptake under Corn Belt practice; the loss mechanism is as described in Wolt (2004).

Anhydrous at $964/ton is 27% above year-ago (DTN/Progressive Farmer, Aug 19, 2026), and the value protected scales directly with the nitrogen price. A 180 lb N/acre fall application at $0.59/lb represents $106/acre of nitrogen sitting in the ground over winter.

Ground Truth: The insurance framing produces a conclusion that runs against the way this product is usually sold, and the direction is the useful part. Inhibitor value is highest exactly where nitrogen price is highest and exposure is longest — which means the fall-applied acre in an expensive nitrogen year is the strongest case for the product, and the sidedress acre in a cheap nitrogen year is the weakest. Those are not marginally different cases; on the p × B structure they can differ by an order of magnitude. Selling the same 6.4% into both is what makes the number look implausible in one context and conservative in the other. The grower question is not "does nitrapyrin work" — the mechanism is established — but "how long will my nitrogen sit unprotected, and what is it worth while it sits."

5. What this desk could not source, and why it matters

What's new: The single number that would convert this analysis from a framework into a recommendation is the current per-acre cost of the inhibitor, and it could not be sourced to a public-tier origin in preparing this piece.

Evidence: The $10–15/acre range used throughout is a stated range, not a quotation. No 2026 extension budget or retail price list carrying a current per-acre nitrapyrin cost was retrievable from a public-tier source this cycle. Every figure in the break-even table therefore inherits an unsourced input, and the table is presented as a scenario grid rather than as a costing.

Separately, the efficacy figure is 2004 vintage. Twenty-two years is a long time in soil microbiology, formulation chemistry and hybrid genetics, and it predates the widespread adoption of both the current generation of inhibitor formulations and substantially higher yield goals.

Source: Crop Root Zone. Provenance status of each input used in this article.

Ground Truth: The unpublished term is the more interesting gap, and it is the same class of gap this desk found in a seed-treatment trial earlier this week: the research reports an effect and omits the parameter that converts the effect into a decision (Crop Root Zone, TRZ-0133, Aug 27, 2026). There the missing parameter was the trial's coefficient of variation; here it is the loss-year frequency in the underlying trial set. Both are known to the researchers, neither is expensive to report, and without them a grower cannot rescale a published average to their own conditions — which is the only operation that makes a published average useful. The recommendation that follows is not to a grower but to the extension and journal system: report the frequency of the condition the treatment addresses, not only the average effect across conditions. An average benefit without its base rate is an answer to a question nobody is asking.

6. Limits

The break-even inversion in section 3 assumes the yield benefit in a loss year is constant and that loss years are binary. Neither holds: nitrogen loss is continuous in severity, and the benefit scales with the amount lost. Expressing break-even as a fraction of the trial set's own loss frequency is what keeps the arithmetic valid despite that simplification — the fraction is a ratio of expected values and does not require the binary assumption — but the interpretation of that fraction as "a loss-year frequency 15% of the trials'" does, and readers should treat it as a heuristic rather than a probability statement.

Second, the 6.4% figure is specifically for anhydrous ammonia plus nitrapyrin in Midwestern corn. It does not transfer to urea, UAN, other inhibitor chemistries, other crops, or other regions, and this piece makes no claim about any of them.

Third, nothing here addresses the environmental case. The literature also finds that inhibitors reduce nitrous oxide and nitric oxide emissions and nitrate leaching, and there is work assessing costs to ecosystems and environmental health alongside the agronomic benefits (references 3 and 4). Those effects are real and are not part of the private break-even computed here; a grower facing nutrient-loss regulation or a watershed program may face a different and stronger case than the one this article prices.

References

1. Wolt, J.D., "A meta-evaluation of nitrapyrin agronomic and environmental effectiveness with emphasis on corn production in the Midwestern USA," Nutrient Cycling in Agroecosystems, 2004. https://link.springer.com/article/10.1023/B:FRES.0000025287.52565.99
2. DTN/Progressive Farmer, retail fertilizer survey for the week of Aug 10–14, 2026, published Aug 19, 2026 — anhydrous $964/ton and $0.59/lb N. https://www.dtnpf.com/agriculture/web/ag/crops/article/2026/08/19/5-fertilizer-prices-slightly-lower-3
3. "A meta-analysis to examine whether nitrification inhibitors work through selectively inhibiting ammonia-oxidizing bacteria," PMC9343776. https://pmc.ncbi.nlm.nih.gov/articles/PMC9343776/
4. "Widespread Use of the Nitrification Inhibitor Nitrapyrin: Assessing Benefits and Costs to Agriculture, Ecosystems, and Environmental Health," PubMed 33433195. https://pubmed.ncbi.nlm.nih.gov/33433195/
5. Barchart, December 2026 corn futures (ZCZ26), quote retrieved Aug 28, 2026. Delayed quote. https://www.barchart.com/futures/quotes/ZCZ26
6. Crop Root Zone, TRZ-0133, "The Seed Treatment Needs 0.6 Bushels to Pay. Proving It Would Take About 500 Replicates," Aug 27, 2026.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

North Carolina State University's Extension soil management programme ran two commercial microbial seed treatments against untreated soybean seed at six locations in 2025, in randomised blocks with sixteen replicates per treatment, and reported the result plainly: "Soybean yield did not differ between microbial seed treatments and non-treated" (NC State Extension, 2025 Microbial Seed Treatment Soybean Field Trials). That is a well-built trial and an honest null. It is also, by the arithmetic below, incapable of detecting a yield response small enough to pay for the product — and the trial write-up does not address cost at all. Both things are true at once, and the second is the one a buyer needs.

1. The trial, and what it found

What's new: Two products were tested. BioWake (AMVAC GreenSolutions) is a single bacterium, Methylobacterium hispanicum. Buncha Bugs DF ST (Concept AgriTek LLC) combines multiple bacterial strains with mycorrhizal fungi. The control was non-treated seed. Locations spanned all three of North Carolina's ecoregions — Tidewater, Coastal Plain and Piedmont — as part of a multi-year study covering 2024, 2025 and 2026.

Evidence: Yields varied enormously by location and not at all by treatment.

Source: NC State Extension, 2025 Microbial Seed Treatment Soybean Field Trials. Letters denote statistical grouping among locations. Bars scaled so the highest yield is ten blocks.

The spread across locations is 30.1 bushels — from 49.9 at Union to 80.0 at Pasquotank, a 60% range on the low end. The mean across the six sites is 64.55 bu/acre. Two locations, Robeson and Rowan, showed a localised advantage to Buncha Bugs, and NC State explicitly declined to build on them: those "observations lack the statistical support to make broader claims." That is the correct call, and it is worth noting why. With six locations and two products there are twelve location-by-product comparisons; at a 5% significance threshold, chance alone would be expected to produce roughly half a positive. Two is not a signal, but it is not far enough from expectation to be treated as one either.

It is worth stating the two signals side by side, because the ratio between them is the whole problem this article is about. The trial's environmental signal — the difference between its best and worst site — is 30.1 bushels. The economic signal it is being asked to resolve, developed in the next section, is between a quarter and two-thirds of a bushel.

Source: location yields from NC State Extension, 2025 Microbial Seed Treatment Soybean Field Trials; break-even figures are Crop Root Zone calculations developed in Section 2.

An experiment in which the nuisance variation is forty-seven times the effect of interest is not a badly designed experiment — it is an experiment designed for a different question. Replication is what separates the two, and Section 3 puts a number on how much would be needed.

2. What the treatment costs, and what it has to return

What's new: The trial reports no cost figures. Filling that in from the seed trade: adding biologicals to a soybean seed order runs roughly $3 to $8 per bag.

Evidence: A soybean unit is conventionally 140,000 seeds. At typical Mid-Atlantic and Midwest seeding rates of 120,000–160,000 seeds per acre, a bag covers roughly 0.875 to 1.17 acres, so the per-acre cost is close to the per-bag cost. Taking one bag per acre as the central case and pricing the output at the November 2026 soybean board of $12.5975 (Barchart, delayed quote, Aug 27, 2026):

Source: cost range from published seed-trade figures for biological additions to soybean seed; price from Barchart November 2026 soybean futures, delayed quote read Aug 27, 2026. Break-even = cost ÷ price. Crop Root Zone calculation.

0.24 – 0.64 bu/acre

The yield response a $3–8/acre microbial seed treatment must deliver to break even at $12.60 soybeans. Under one percent of the trial's mean yield. (Crop Root Zone calculation)

3. The smallest difference the trial can see

What's new: A randomised block design with sixteen replicates resolves differences of roughly two to five bushels, depending on how variable the field is. It cannot resolve six-tenths of a bushel.

Evidence: The least significant difference between two treatment means scales as t × CV × mean × √(2/n), where CV is the coefficient of variation of the trial and n the number of replicates. Soybean yield trials commonly run CVs between 5% and 12%. At n = 16 and t ≈ 1.96:

Source: Crop Root Zone calculation. LSD ≈ 1.96 × CV × 64.55 bu × √(2/16). The trial's own CV was not published in the summary read; the range shown brackets typical values for replicated soybean yield trials.

Running that backwards gives the replication the question actually requires. To resolve a 0.635 bu/acre difference at a CV of 8% takes roughly 508 replicates per treatment — about thirty-two times what the trial ran. At the $3 end of the cost range it is over 3,500.

Ground Truth: "No significant effect" and "does not pay" are different findings, and this trial produced only the first. Its design answers the question is there a large effect? — the answer is a well-supported no — and is structurally silent on is there a six-tenths-of-a-bushel effect?, which is the only question a buyer at $3–8 an acre is asking. The null is not evidence the product fails to pay. It is also not evidence it pays. The honest summary is that public replicated variety-trial infrastructure, at the replication levels it can afford, cannot resolve the economics of a low-cost input, and no amount of further trials at this scale will change that.

4. Why more years will not close the gap

What's new: Pooling the multi-year study helps, and not nearly enough.

Evidence: Standard errors shrink with the square root of replication. Pooling three seasons of six locations rather than one triples the site-years, which cuts the detectable difference by about √3, or 1.73×. On the 8% CV case that takes the resolution from 3.58 bu/acre to roughly 2.07 bu/acre.

Source: Crop Root Zone calculation, scaling the single-season LSD by √3. Treats site-years as independent, which is generous — locations repeat across years, so the true gain is smaller than shown.

So the completed three-year study, when it reports, will be a substantially better answer to the large-effect question and still roughly three times too coarse for the economic one. That scaling is generous, because the same six locations recur and are therefore not fully independent.

Ground Truth: The practical consequence is that a grower will never get this decision settled by the trial data, and should stop waiting for it. The decision has to be made on grounds the trials can support: that the effect, if it exists, is smaller than roughly two bushels — because a larger one would already have shown up — and that the product costs a known $3–8. That bounds the bet from above without confirming it from below. A $5 input with a demonstrated ceiling of two bushels and no demonstrated floor is a defensible small trial on your own acres and an indefensible whole-farm program.

5. What would actually answer it

Three designs would resolve what a yield trial cannot.

On-farm strip trials at scale. Field-length strips replicated across many farms trade experimental control for replication, and replication is exactly the binding constraint here. Yield-monitor data from a few hundred paired strips gets closer to the 500-replicate requirement than any small-plot programme can.

Measuring the mechanism instead of the outcome. If a treatment's claim is improved nutrient access or seedling vigour, those are measurable directly and at far higher precision than a yield difference of one part in a hundred. A trial that shows a root-colonisation or early-biomass effect and no yield effect at least locates where the chain breaks.

Publishing the CV. The single most useful number NC State could add to a summary like this one is the trial's own coefficient of variation, because it converts a null result into a stated detection limit. "No difference" tells a reader nothing about what size of difference was ruled out. "No difference, and we could have detected 3.2 bushels" tells them a great deal.

6. Limits

Two products, one state, one season. These findings apply to BioWake and Buncha Bugs DF ST on soybeans in North Carolina in 2025. They do not generalise to other microbial products, other crops, or the much broader biologicals category — and specifically say nothing about nitrogen-fixing products for corn, which are a different claim with a different economic threshold.

The trial's CV was not published in the summary this desk read, so every figure in Sections 3 and 4 is a scenario across a plausible range rather than a calculation on the trial's actual variance. If the true CV is at the low end, the gap narrows to roughly 3.5×; it does not close.

The cost figure is a trade range, not a survey. $3–8 per bag for a biological addition is what the seed trade reports; a specific grower's invoice may sit outside it, and treatment cost may be bundled into a seed package in ways that make the marginal cost hard to isolate.

The break-even ignores everything but yield. No credit is given for stand establishment, harvestability, or any effect that shows up somewhere other than the yield monitor — and no charge is made for handling or application.

The 1 bag ≈ 1 acre convention drives the per-acre cost directly. At 120,000 seeds/acre the cost per acre falls about 14% and the break-even falls with it; at 160,000 it rises about 14%.

References

1. NC State Extension, Soil Management, "2025 Microbial Seed Treatment Soybean Field Trials." Six locations (Beaufort 56.7, Pasquotank 80.0, Robeson 69.4, Rowan 71.5, Sampson 59.8, Union 49.9 bu/acre) across Tidewater, Coastal Plain and Piedmont ecoregions; randomised blocks, 16 replicates per treatment; products BioWake (AMVAC GreenSolutions, Methylobacterium hispanicum) and Buncha Bugs DF ST (Concept AgriTek LLC) against non-treated control; multi-year study covering 2024, 2025 and 2026. https://soilmanagement.ces.ncsu.edu/agricultural-biologicals/biological-product-trials/2025-microbial-seed-treatment-soybean-field-trials/
2. Barchart, November 2026 soybean futures (ZSX26), delayed quote read Aug 27, 2026: 1259-6 ($12.5975). https://www.barchart.com/futures/quotes/ZSX26
3. Published seed-trade figures for the incremental cost of biological seed treatments on soybeans, $3–8 per 140,000-seed unit, and conventional soybean seeding rates of 120,000–160,000 seeds per acre.
4. Crop Root Zone, "Microbial Nitrogen Is Not Free Nitrogen. It Is Nitrogen Bought With Sugar, and the Exchange Rate Is Publishable," Aug 22, 2026 — the corn nitrogen-fixation case, which is a different product class and a different economic threshold from the seed treatments examined here.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.

Michigan State University Extension's guidance on manganese deficiency in soybeans ends with a sentence that reads like agronomy and functions like a budget rule: the best way to avoid the problem is to grid-sample or zone-sample, apply lime by variable rate, and avoid raising soil pH above 6.5 (MSU Extension). On lakebed and glacial outwash soils, crossing that line is what turns a soybean field manganese-deficient. And lime is the one routine amendment on the farm that is cheap to add and very expensive to take back — which makes the error one-directional in a way almost no other input decision is.

1. Where the threshold sits, and which soils it binds on

What's new: The pH trigger is not one number. It depends on the soil.

Evidence: Manganese deficiency becomes common above pH 5.8, and the threshold varies by soil type in a way that matters for which fields carry the risk.

Source: Michigan State University Extension, "Identifying and correcting manganese deficiency in soybeans," and MSU Extension nutrient management guidance for high-yield soybean production.

The muck and dark-sand case is the awkward one. A 5.8 threshold sits below the pH most liming programmes are aiming at, which means on those soils the agronomic target for pH and the agronomic target for manganese availability are in direct conflict — you cannot satisfy both, and the lime decision is choosing which nutrient problem to have.

Ground Truth: Liming is the only routine amendment whose principal cost may be incurred years later, in a different budget line, on a different crop. The lime invoice lands in the amendment column in one year; the foliar manganese it makes necessary lands in the micronutrient column in the soybean years that follow. Those two lines are never compared, because they never appear on the same page of any budget. That is not an accounting quirk — it is why the over-liming decision keeps getting made.

2. What the lime side costs

What's new: Published state surveys put delivered-and-spread agricultural lime in a fairly tight band this year, which makes the one-time side of the trade easy to price.

Evidence: Three 2026 survey figures and the general trade range:

Source: 2026 Missouri farm custom rate survey; 2026 Maryland and Delaware custom rate survey; 2026 Texas agricultural survey (application only); general trade range for delivered-and-spread agricultural lime. Bars scaled so $48.70 is ten blocks; the trade range is shown as a span.

The Texas figure is on a different basis and is included only to show the split: at $15.31 for application alone against $30.58 all-in in Missouri, roughly half the delivered-and-spread cost is the spreading, not the rock.

At soil-test rates of 1 to 3 tons per acre, that puts a lime pass between about $31 and $146 per acre, effective for several years.

3. What the manganese side costs, and the number we could not source

What's new: The correction is well specified in the extension literature. The price of the product is not, at least not at US agricultural retail, and this desk is not going to invent one.

Evidence: The agronomy is unambiguous. Foliar application of 1 to 2 lb of actual manganese per acre is described as the most economical and effective correction. Equivalent guidance gives 0.5 lb applied multiple times, or 0.7–1.0 lb applied once, and — stated as product rather than element — 4 to 8 lb of manganese sulfate at the V2 to V3 stage will usually correct minor deficiencies.

Those two forms of the recommendation agree, which is a useful internal check. Commercial agricultural manganese sulfate runs about 31% Mn (range roughly 28–32%), so 4–8 lb of product carries 1.24 to 2.48 lb of actual Mn — squarely inside the 1–2 lb element recommendation.

Source: MSU Extension manganese guidance; Crop Root Zone conversion at 31% Mn.

The price of agricultural-grade manganese sulfate could not be sourced to a US retail series this run. The quotations this desk could reach — roughly $590 to $1,180 per metric tonne — are industrial and battery-grade Chinese and Indian assessments, and battery-grade manganese sulfate is a materially different, higher-purity product sold into a different market. Using it as a farmgate proxy would be wrong. So the product cost is presented as a scenario grid rather than a figure:

Source: Crop Root Zone scenario calculation. The $/lb column is a grid, not a market quotation — see Limits.

Across that whole grid the product is a small number. The recurring cost is dominated by the application pass, not the manganese, and a foliar pass has to be paid for whether it carries $1.60 or $10.00 of product. That is the structural point: the manganese bill is really a trip across the field bill, and trips across the field do not get cheaper with the micronutrient market.

There is one product-selection nugget in the extension guidance worth carrying, because it is worth real money and is frequently got backwards. If you are not tank-mixing with glyphosate, manganese sulfate works as well as the chelated forms for foliar feeding, so cost per pound can simply decide the choice. If you are tank-mixing with glyphosate, the chelate earns its premium. Band application of chelated manganese is not recommended at all, on cost grounds.

4. The error is asymmetric, and that should move the target

What's new: Raising pH is a routine, priced, widely-offered service. Lowering it is not.

Evidence: There is no custom-rate survey line for "reduce soil pH," because it is not a service farms routinely buy. Bringing pH back down means elemental sulfur at rates measured in hundreds of pounds per acre, on a multi-season response curve, and this desk did not find a published custom rate for it to put beside the $30.58 lime figure. That absence is the finding: one direction of the pH decision is a commodity service with a survey price, and the other is not a normal purchase at all.

Source: 2026 Missouri and Maryland/Delaware custom rate surveys for the lime line; the absence of a comparable elemental-sulfur pH-reduction line in those surveys is this desk's observation.

Ground Truth: When the cost of overshooting is much higher than the cost of undershooting, the right response is not better monitoring — it is a lower target. On Mn-responsive lakebed, outwash, muck and dark sand soils, the lime target should deliberately sit below the agronomic optimum, because an acre that comes in half a point light can be topped up next year at survey rates, and an acre that comes in half a point heavy cannot be brought back at any rate you will find quoted. Most liming programmes are set to hit a number. On these soils the programme should be set to approach it from underneath.

5. Where this changes the variable-rate calculation

What's new: MSU's remedy — grid or zone sampling with variable-rate lime — is usually justified on lime saved. On responsive soils the larger return may be somewhere else entirely.

Evidence: Variable-rate lime is normally pitched as an input-efficiency purchase: put the tons where the pH is low, skip where it is not, bank the difference in rock and spreading. That saving is real but bounded by the lime bill, which Section 2 puts at $31–146 an acre depending on rate.

The unbudgeted return is the over-limed zone that never happens — and with it, the annual foliar pass on that zone for as long as the elevated pH persists. This desk has previously set out the general case for and against variable rate (Crop Root Zone, "The Variable-Rate Bill vs. the Variable-Rate Payoff," Jul 18, 2026). The specific case on Mn-responsive soils is stronger than the general one, because the avoided cost is recurring and the avoided error is effectively permanent, neither of which is true of the ordinary nutrient-placement argument.

Ground Truth: Zone-based liming on responsive soils should be evaluated as insurance against an irreversible error, not as an efficiency gain on a divisible input. Those are different investment cases with different hurdle rates, and the second one — the way it is normally sold — undervalues it on exactly the fields where it matters most.

6. Limits

No US agricultural retail price for manganese sulfate was obtained. Section 3's cost grid is a scenario, explicitly. The industrial and battery-grade quotations available are not a farmgate proxy and were not used as one.

No custom rate for a foliar micronutrient pass was sourced this run either, so the application cost that Section 3 argues is dominant is asserted structurally rather than priced. That is a real gap: the piece establishes that the pass matters more than the product without establishing what the pass costs.

The pH thresholds are Michigan-derived. They come from MSU Extension guidance developed on Michigan soils. Lakebed and outwash soils elsewhere in the northern Corn Belt behave similarly, but a grower outside that region should check thresholds against their own state's recommendations rather than these.

Lime survey figures are state averages from three different states on two different bases, one of which — Texas at $15.31 — excludes the material entirely and is shown only to decompose the cost.

Soybeans are the responsive crop here. Corn is considerably less sensitive to manganese, so on a corn-soybean rotation the recurring cost falls in roughly half the years, which halves the annual average and does not change the direction of the argument.

The persistence of an elevated pH is not quantified. How long an over-limed zone stays over-limed depends on soil buffering, rainfall, nitrogen programme and cropping. It is years, not one season, which is the load-bearing part — but this piece does not put a number on it.

References

1. Michigan State University Extension, "Identifying and correcting manganese deficiency in soybeans" — Mn deficiency common above pH 5.8, especially on muck and dark-coloured sands above 5.8 and lakebed or outwash soils above 6.5; grid or zone sampling with variable-rate lime; avoid raising soil pH above 6.5. https://www.canr.msu.edu/news/identifying_and_correcting_manganese_deficiency_in_soybeans
2. Michigan State University Extension, "Nutrient management recommendations for high-yield soybean production" — foliar application of 1 to 2 lb/acre actual manganese as the most economical and effective correction; 4 to 8 lb manganese sulfate at V2–V3; chelate versus sulfate selection and the glyphosate tank-mix caveat; band application of chelated manganese not recommended on cost grounds. https://www.canr.msu.edu/news/nutrient_management_recommendations_for_high_yield_soybean_production
3. Crop Protection Network, "Manganese (Mn) Deficiency of Soybean." https://cropprotectionnetwork.org/encyclopedia/manganese-mn-deficiency-of-soybean
4. 2026 Missouri farm custom rate survey — lime delivered and spread, including material, average $30.58/ton.
5. 2026 Maryland and Delaware custom rate survey — lime and application, average $48.70/ton.
6. 2026 Texas agricultural survey — lime application rate $15.31/ton, material excluded.
7. Crop Root Zone, "The Variable-Rate Bill vs. the Variable-Rate Payoff," Jul 18, 2026.

Disclosures

Crop Root Zone is an independent publication. Nothing here is investment advice. Market calls, where included, reflect the editors' own read and are not a recommendation to buy or sell any security.