Why does a grain market live on a calendar?
Energy is produced every hour and metals are produced every day. A grain crop is produced once a year, in one hemisphere. Everything between harvests is not production — it is inventory management. That single fact reorganises the whole market.
The two calendars
The US corn and soybean year runs roughly like this:
- Late April to May — corn planting. May to June — soybean planting.
- July — corn pollination, the single most weather-sensitive window of the year. Heat and drought during pollination cannot be repaired later.
- August — soybean pod fill, the equivalent critical window for that crop.
- September to November — harvest.
The Southern Hemisphere runs on the opposite clock. Brazil plants soybeans from September to December and harvests January to March, then plants safrinha — second-crop — corn into the same fields.
So the world gets two harvests a year to argue about, and a "weather market" every few months, alternating hemispheres. A trader who understands only the US calendar is blind for half the year.
Old crop and new crop
This is the concept that separates people who have read about grains from people who understand them. Grain contract months are not interchangeable, because they refer to different crops.
For US corn, the marketing year turns on 1 September, and the contract months sit either side of it. July is the last clean old crop month — grain already harvested and sitting in a silo. September straddles the boundary: its delivery period falls inside the new marketing year, and by late September new-crop corn from the southern Corn Belt is deliverable against it. December is the main new crop contract: grain still in a field, or not yet planted. For soybeans, the new-crop month is November.
Now consider a July drought. It hits the December contract hard, because December's corn is being damaged in real time. It hits the July contract far less, because that corn was harvested last autumn and is already in storage. The two contracts can move in opposite directions on the same news, and the curve between them can bend sharply mid-year.
That bend is not carry, storage or interest. It is two different crops inside one price series. Any analysis that treats a grain futures curve as a single continuous carry structure will misread it every summer.
The contract arithmetic
CBOT corn, soybeans and wheat all trade in 5,000-bushel contracts, quoted in cents per bushel. Bushel weights differ by crop: corn is 56 lb, wheat and soybeans 60 lb.
Corn at 450 cents:
- $4.50 per bushel × 5,000 = $22,500 of notional per contract
- A 10-cent move = $500 per contract
- A 1-cent move = $50
Keep those handy — grain quotes are in cents and are easy to misread by a factor of a hundred.
Wheat is three commodities
There is no such thing as "the wheat price". The three main US contracts are:
- CBOT soft red winter — lower protein, for cakes, pastries and crackers.
- KC hard red winter — higher protein, the bread-flour benchmark.
- MGEX hard red spring — highest protein, milled for blending and specialty flour.
Three exchanges, three growing regions, three planting calendars, three prices — and the spreads between them are traded in their own right. Bread flour and cake flour do not come out of the same field, and the market prices them accordingly. Unit 1's rule, once more.
The number everyone actually reads
Grain analysis converges on one government publication and one ratio. The USDA's monthly WASDE (World Agricultural Supply and Demand Estimates) publishes a balance sheet per crop; Prospective Plantings at the end of March gives intended acreage; quarterly Grain Stocks gives the physical count.
The output that matters is ending stocks, and better still the stocks-to-use ratio:
Ending stocks ÷ annual use.
Example: ending stocks of 1.5 billion bushels against annual use of 14.5 billion bushels:
1.5 ÷ 14.5 = 10.3%, or about 38 days of cover (0.103 × 365).
That ratio does exactly the job that spare capacity and inventories do in oil. When stocks-to-use is comfortable, a 3% production shortfall is absorbed by the buffer and barely registers. When it is tight, the same 3% must be rationed by price — and because food demand is highly inelastic in the short run, rationing takes a large move. It is the elasticity arithmetic from the energy unit, running on an annual clock instead of a weekly one.
In the data
The global corn price is one number a month, in dollars per metric tonne, and it has no contract months inside it: the old-crop/new-crop boundary this lesson is about simply cannot appear in it.
The contract-level view is the CBOT future, the second row below, which prints in the hundreds because it is quoted in cents per bushel.
On 29 September 2026 the future read 521.5, or $5.215 a bushel, which at 39.37 bushels to the tonne is about $205 a tonne, the same order as the monthly series' $213 for July. Two units, two calendars, one grain.
Try it now
- Five years of corn futures is below, in cents per bushel. Mark the June-to-August windows with a Level at the top of each, and note where the biggest ranges sit. Weather markets are visible in a price series carrying no weather data at all.
- Compute the notional value of one soybean contract at 1,150 cents per bushel. ($11.50 × 5,000 = $57,500.) Then the cash value of a 25-cent move. ($1,250.)
- Compute stocks-to-use for ending stocks of 2.1 billion bushels against use of 15.0 billion. (14.0%, about 51 days.) Say in one neutral sentence how that market would absorb a 3% crop shortfall compared with the 10.3% case — as a description of buffering, not a price prediction.