To see what a bushel cost on each of your fields, add up what that field cost and divide by the bushels weighed off it. What the field cost is its invoices plus its share of the hours of your own machines and operator. Do this after harvest, one row per field. A published cost per bushel will not give you the answer, because it describes a typical farm and not your fields. In the invented example, the farm averages $4.55 a bushel, yet one field costs $3.91 and another $5.65.
The worked example is invented: three soybean fields of 300, 250 and 150 acres. The sheet has six columns. Four hold the result: acres, bushels per acre, dollars per acre and dollars per bushel. Two hold the numbers each division starts from: the bushels weighed off the field and the dollars charged to it. This is the actual cost of each field, measured after harvest from the farm’s own records. It differs from the budget written before planting, which the crop budget article covers.
Read the cost per bushel only with the cost per acre and the yield in bushels per acre beside it, because it swings much more than the cost per acre does. This shows in the two published sources that report both costs: soybeans in Illinois and in Pergamino, Argentina. In each, the highest cost per bushel or tonne sits above the lowest by about 24 percent in Illinois and 28 percent in Pergamino. The highest cost per acre or hectare sits only 8 percent above the lowest. So the per-bushel gap is three to three and a half times as wide.
Why isn’t a published cost per bushel your cost?
Every public unit cost is built for a typical farm, or averaged over other farms, and none is read off yours. Four sources show how. Conab, the Brazilian federal public company that tracks crop supply and costs, multiplies the quantity of each input by its price to cost the most common way of farming in a region. Conab calls that setup the modal unit. Embrapa, Brazil’s public agricultural research company, studied soybeans in Mato Grosso do Sul in 2018/2019. It assumes one typical farmer with 840 hectares, about 70 percent owned and 30 percent rented, about 2,075 acres. INTA, Argentina’s national agricultural technology institute, takes its crop management from its own experimental field and prices field work from local contractors in its soybean estimate for Pergamino.
The University of Illinois series is built from real farm records. Its author, B. Zwilling of the University of Illinois, averaged the records of farms enrolled in an Illinois farm-management association, a membership group whose farms share their records. The average covers only farms above 500 acres (about 200 hectares) without livestock, grouped into four sections of the state.
None of the four calculates what a field of yours cost. Each calculates what a model, or an average of other farms, cost. They work as a yardstick once your own figure exists, and that figure comes from the sheet in this article.
Which invoices belong to which field?
Sort the season’s costs into three groups. The first goes straight to a field, because the invoice says where it went: seed, fertilizer, chemicals and hired work, matched by application date and field. The second is shared by several fields: fuel, owned machines and the operator. You split these by the hours worked in each field, because no invoice says which field they served. The third is land and general expenses. Leave them out of the shared costs, and add them only if you write at the head of the sheet that they are in (see the section on which costs are inside the number).
Costs of the season
Sorted by how each one reaches a field
01
Straight to a field
Seed, fertilizer, chemicals and hired work
02
Shared by several fields
Fuel, owned machines and operator, split by hours worked
03
Land and general expenses
Left out unless you write that they are in
FAO, the United Nations food and agriculture agency, has a handbook on cost-of-production statistics. For the shared costs, FAO says in it that there is no right or wrong way to split them, only a rule that is stated and applied the same way to every field. Its examples of such a rule are the time the tractor worked each crop or the number of field passes and, when neither is known, the share of land. The shares must add to 100 percent, so every dollar of the shared costs lands somewhere.
The hourly cost of an owned machine comes from the machine-hour worksheet, and you can check the fuel against the hour meter with the fuel and hour-meter check. In the example, the shared costs are $48,000 for fuel, machines and operator: an all-in $120 for each of the 400 hours on the hour meter, split by the hours each field took.
Field C costs $21 an acre more than field A, 8.4 percent: $36 more in the shared costs and $15 less in invoices. Land and general expenses are left out of the example on purpose, because land can change what the number means. In the Illinois soybean costs for 2024, land was 31 percent of the total. In Pergamino, INTA prices soybeans at a middle yield of 3.4 tonnes a hectare, about 50 bushels an acre, the average of the last 10 seasons at its experimental field. On owned land its expense total comes to $175.41 a tonne. On rented land, with its rent line added, the total comes to $315.65 a tonne. INTA builds this total on a gross-margin sheet, which adds up only the costs that vary with the crop (field work, inputs, harvest and marketing). For that reason its sheet holds no general expenses or depreciation.
Which harvest do you divide by?
The bushels that were weighed off each field, from the delivery tickets and the bin readings. FAO sets the rule for the divisor: the production in a cost per unit should be the marketable output, with waste left out. When a truck carries grain from two fields, write the field on each ticket. When a bin mixes fields, read its level before and after each one.
If you use a yield from memory, or the farm’s average for every field, you cannot see the difference between fields, and that difference is what the sheet is for. With the weighed bushels in hand, divide each field’s dollars by its bushels, one row per field.
| Field | Bushels per acre | Bushels weighed | Cost | Per acre | Per bushel |
|---|---|---|---|---|---|
| A, 300 acres | 64 | 19,200 | $75,000 | $250.00 | $3.91 |
| B, 250 acres | 55 | 13,750 | $66,850 | $267.40 | $4.86 |
| C, 150 acres | 48 | 7,200 | $40,650 | $271.00 | $5.65 |
| Whole farm, 700 acres | 57.4 | 40,150 | $182,500 | $260.71 | $4.55 |
Behind the farm’s $4.55 sits a $1.74 gap between fields A and C. Field A grew 16 more bushels an acre than field C and cost $21 less per acre. The sheet does not tell you why; you find out in the field itself: soil, planting date, weeds, or a spot that flooded. With it you see which field to look at first, and whether the dollars or the bushels moved.
Why does the cost per bushel swing more than the cost per acre?
Because dollars per acre tend to change little from field to field, while bushels per acre change a lot. FAO points the same way, between years: the cost per unit of land is likely to be more stable in the short term, since technology varies less from year to year than yields do. For comparing sources, it adds that a cost per unit of output might not come from a normal year, and that cost per area may serve better.
The published numbers point the same way, but from other countries and in dollars. In Illinois in 2024, soybean costs ran from $863 to $931 an acre across the four sections, and from $12.42 to $15.41 a bushel, which Zwilling relates to weather, yields and land quality. The INTA model for Pergamino has three yield levels for first-crop soybeans (soybeans sown as the season’s first crop, not after wheat), from 2.9 to 4.0 tonnes a hectare. On owned land, the cost per hectare rises from $575.10 to $622.00. Only the marketing costs (freight, warehouse commission and others) grow with the tonnes; harvest costs the same at all three levels. The cost per tonne falls from $198.31 to $155.50.
| Source | Per acre or hectare | Per bushel or tonne |
|---|---|---|
| Illinois soybeans, 2024, four sections | $863 to $931 an acre, 7.9% apart | $12.42 to $15.41 a bushel, 24.1% apart |
| Pergamino soybeans, 2024/25, three yield levels, owned land | $575.10 to $622.00 a hectare, 8.2% apart | $155.50 to $198.31 a tonne, 27.5% apart |
| Example farm, three fields (hypothetical) | $250.00 to $271.00 an acre, 8.4% apart | $3.91 to $5.65 a bushel, 44.5% apart |
The three comparisons are different things: four sections of a state that differ in weather and land, three yield levels of one model, and three fields. In each, the per-acre or per-hectare figure moves about 8 percent, and the per-bushel or per-tonne figure moves 24 to 45 percent. Only Illinois is data from real farms; Pergamino is one model’s estimate, and the example was written with invented numbers to have this shape, so it proves nothing. What it illustrates is the arithmetic: with the cost per acre almost still, the cost per bushel follows the bushels per acre. When you read your own sheet, take one column at a time. If a field’s cost per bushel rises, check whether its bushels per acre fell, its dollars per acre rose, or both.
Neither published source follows one field across years, so neither shows how your fields move from one season to the next. That takes two seasons of your own sheets side by side.
Which costs are inside the number?
Whoever writes a cost per bushel chooses which costs go into it, and the figure changes with the choice. Conab separates the explicit costs, those actually paid, from the implicit ones, which involve no payment. Depreciation is one: the yearly loss in value of a tractor you already own. Opportunity cost is another: what the money tied up in the land or machines would have earned elsewhere. Conab’s operating cost already includes depreciation; its total cost adds to that a charge for the land and for the money tied up in machines and buildings. FAO lists the usual cuts: cash costs alone, cash costs plus land rent, or total costs.
How much the cut matters shows in Embrapa’s soybean study for Mato Grosso do Sul in 2018/2019 (2018 values, which serve here only to show the size of the cut). For glyphosate-tolerant (RR) soybeans at 55 sacks a hectare, Embrapa puts the operating cost at R$ 2,559.59 a hectare and the total cost at R$ 3,143.90. A sack weighs 60 kilograms, about 2.2 bushels of soybeans. Divided by the 55 sacks, that is R$ 46.54 and R$ 57.16 a sack, the break-even prices Embrapa itself gives (the lowest selling price that still covers the cost). The total is 22.8 percent higher. The whole difference of R$ 584.31 a hectare is a charge for what the farm already owns. It is R$ 459.36 for land, R$ 93.42 for machines and R$ 31.53 for buildings.
So write at the head of the sheet which costs the dollars hold: paid invoices only, or those plus the owned machines and the operator, or those plus land. Two sheets with different sets cannot be compared. The same division, with the costs chosen by date, gives the farm’s break-even, covered in break-even price per acre: which costs count, and when.
Which paper do you pull out first?
The delivery tickets of the last harvest, because the bushels weighed are the only number the sheet cannot rebuild later from the invoices. Plan on about three hours for a first sheet, most of it sorting invoices by field (an estimate, not a measured time). Start with the crop that has the most fields, since that is where a single average leaves the most differences out of sight.
The sheet is a method built from cost-accounting handbooks and extension models. No study cited here tested it on working grain farms, or showed that farms keeping it earn more.
After a second season, lay the two sheets side by side, field by field. A field whose cost per acre held while its bushels fell is a different question from one whose bushels held while its cost per acre rose. Only two seasons side by side let you tell them apart. The cost per acre and the yield are also the two columns you write the next budget from.
Keeping one sheet per field and season is part of farm management, and the finance articles take up the questions that come next. Related: How do you rank crops by margin and trust the order? and Did the crop make money? Work out three results, not one. Also: How to register each field: code, area and history and Is your farm’s margin really behind a public reference?.