Oracle Bone #009 · Filed September 27, 2026 · Judge by January 31, 2028
The Harvest Has Less Margin
A reading on the harvests of 2027, the reserves behind them, and the difference between expensive food and scarce crops.
Claim. For at least one of wheat, corn, or soybeans, USDA’s January 2028 World Agricultural Supply and Demand Estimates will show global production below global domestic consumption in 2027/28, and a global ending-stocks-to-use ratio at least 15 percent lower than in 2026/27.
Falsifier. The specified report supplies the comparable data, but none of these three crops meets both tests. Higher grocery prices, regional crop failures, or a commodity-price rally cannot substitute for the stated result.
The question
Will the harvests of 2027 leave materially less reserve capacity in a major food or feed crop?
This bone follows The Weather Is Not the Trade. That essay considers whether adverse weather, strained farm finances, disrupted grain routes, and expensive fuel could reinforce one another. Here we move from a possibility worth investigating to a prediction that can be wrong.
System: global wheat, corn, and soybean supply. Window: the 2027/28 crop marketing year, assessed in January 2028—not a claim that every relevant harvest or shortage falls within calendar 2027. Attribution: Matthew Langenkamp’s hypothesis, operationalized and recorded with Thea. The counterargument below is part of this filing, not a claim of an independent multi-agent Court review. No numerical probability is assigned.
Where the reading begins
The September 11, 2026 WASDE, pages 18–19, 22–23, and 28, supplies the starting point. These are global figures in million metric tons; use means the report’s Domestic Total, not feed alone or soybean crush alone.
| Crop | 2026/27 production | 2026/27 use | 2026/27 ending stocks | Stocks/use, 2024/25 | 2025/26 | 2026/27 |
|---|---|---|---|---|---|---|
| Wheat | 822.43 | 826.75 | 276.29 | 32.23% | 34.05% | 33.42% |
| Corn | 1,290.95 | 1,320.23 | 272.10 | 23.60% | 22.74% | 20.61% |
| Soybeans | 442.35 | 442.44 | 124.02 | 30.56% | 29.09% | 28.03% |
All three already have projected production below use in 2026/27. That alone would be too weak a test of our claim. The bone therefore requires a further material decline in the reserve-to-use ratio in 2027/28.
The 15 percent relative-decline threshold exceeds each one-year ratio decline in this small, three-crop, two-transition comparison; the largest is about 9.4 percent for corn from 2025/26 to 2026/27. This is a baseline sanity check, not a long-run backtest or a calibrated probability. Fifteen percent is our declared materiality threshold, not an official USDA definition of scarcity.
The reading
Brazil’s importance to world crop supplies makes a serious harvest disappointment there consequential well beyond its borders. Adverse weather across more than one supplier, combined with pressure on farm inputs and finance, could leave production unable to keep up with use. Drawing down reserves would then leave less protection against the next disruption. Export restrictions and damaged transport routes could make available supplies still harder to obtain.
But the steps in that argument must remain separate. Farm auctions do not prove disappearing acreage. Drought in the Amazon does not establish drought across Brazilian soybean fields. An El Niño forecast does not establish a failed harvest. Expensive diesel does not guarantee higher farm selling prices. We are predicting a measurable tightening of crop buffers, not treating these signals as proof that it has already occurred.
The environmental, economic, and geopolitical signals are the convergence behind the call. The blind spots are local weather at critical growing stages, acreage and yield responses, accessible exporter inventories, substitution, and demand destruction. Market positioning may tell us the story is already popular; it does not establish a physical shortage.
What this measure cannot tell us
Global stocks are not all available to importers. Their location, ownership, quality, and export restrictions matter. Global marketing-year totals also combine different national crop calendars. This deliberately reproducible test measures a narrowing global buffer, not universal shortages, famine, or the conditions of every household. Regional scarcity could worsen while this prediction fails.
Nor is this a prediction of investment returns. Grain can become more valuable while a farmer loses a crop, a food processor loses margin, or an investor loses money on an expensive position.
The Jester’s counter
The shortage story may arrive before the shortage—and then remain only a story. Good harvests elsewhere, expanded acreage, improved yields, restored shipping, or weaker demand could replenish reserves. A regional disaster can coexist with adequate global supply. Farmers can remain under financial strain while harvests stay large.
The global measure might also miss the very mechanism we find most troubling: a crop that exists but cannot reach its buyer. That would be important, but it would not make this particular bone correct. We do not get to change the test afterward.
How the bone will be scored
Use the January 2028 WASDE, taking both 2026/27 and 2027/28 from that same release, including any official correction issued by January 31, 2028. For each named crop:
- Divide global ending stocks by global Domestic Total for each crop year.
- The 2027/28 ratio must be no more than 0.85 times the 2026/27 ratio.
- That same crop’s 2027/28 global production must be strictly below its global Domestic Total.
A ratio falling from 20 percent to 17 percent qualifies: a 15 percent relative decline, not a 15-percentage-point decline. Calculate from the source figures without rounding intermediate ratios. Do not mix crops or releases to satisfy the two conditions. Today’s baseline documents what we knew; the scoring comparison uses the revised prior-year figures in the specified January report.
- Fulfilled: at least one of the three crops meets both conditions.
- Refuted: comparable data are available for all three and none meets both conditions.
- Untestable: missing comparable data prevent a verdict by January 31, 2028. If a crop with complete data already meets both tests, the claim is fulfilled even if another crop’s data are missing. Do not silently substitute a later report.
This scores USDA’s estimates at a fixed date, not final harvest figures. Later revisions belong in a separate note and do not rewrite the original verdict. A mid-window review is due June 30, 2027; it may record new evidence but must not revise the claim.
Scoring slot
Verdict: Open. Judge by January 31, 2028.
Reasoning audit: To be completed with the verdict. Did weather, input costs, trade disruption, or demand changes drive the result? A correct numerical call does not establish that our causal account was correct.
← Oracle Bones · The Weather Is Not the Trade · Return to Dictionary