Storing maize costs 18% of its value. Two-thirds of that is interest.
Grain bought at the harvest trough and sold into the lean season has to earn its cost of carry before it earns anything else. Over five months that cost is roughly 18 per cent of the grain's value at Ghanaian commercial rates. The warehouse is the small part of it. The seasonal spread that makes storage a business is set, more than by anything else, by the price of money.
Basis and date
This note is a cost build-up, not an assessment, and the seasonal spread it is tested against is a planning range, not a measured distribution. The archive covers one harvest-to-lean cycle, January 2023 to April 2024. Values are cedis per kilogram, 2026.
The trade
Ghanaian maize is cheapest in the weeks after harvest, when everyone sells at once, and dearest in the lean season before the next crop. The gap between the two is the return on storing it, and storage is the aggregation margin that makes grain trading a business anywhere in the world.
The question is not whether the seasonal move exists. It does, in this market and in every market with one harvest and twelve months of eating. The question is how much of it survives the cost of holding the grain, and which cost line does the damage.
The cost of carry, decomposed
Five months, harvest to lean, on maize bought at a harvest reference of GH¢4.60/kg. That reference is a planning figure for a normal year and is not an assessment.
| Cost of holding one kilogram, five months | GH¢/kg | Share of purchase price |
|---|---|---|
| Working capital at 30% p.a. | 0.58 | 12.5% |
| Warehouse and fumigation | 0.18 | 3.9% |
| Additional shrink over the stored period | 0.09 | 2.0% |
| Total cost of carry | 0.85 | 18.4% |
Table 1 — Shrink is the two percentage points a stored lot loses beyond what a prompt lot loses. Working capital is a Ghanaian commercial overdraft rate; see the closing section.
- Working capital, 30% p.a.GH¢0.5868.2%
- Warehouse and fumigationGH¢0.1821.2%
- Additional shrinkGH¢0.0910.6%
Interest is 68 per cent of the carry. The warehouse, the fumigation and the weight lost in the bag together come to under a third. A storage business in this market is only nominally a warehousing business.
Cost build-up, not an assessment. Five months, harvest to lean, on a harvest reference of GH¢4.60/kg — a planning figure for a normal year. Working capital at a Ghanaian commercial overdraft rate of 30% p.a.; the assumptions are set out under “Conditions that would alter this conclusion” below.
Interest is 68 per cent of the total. The warehouse, the fumigation and the weight lost in the bag together come to under a third. A storage business in this market is only nominally a warehousing business; it is a lending business that happens to hold grain as the asset.
What the spread has to be
Break-even is therefore a seasonal move of about 18 per cent. Against a planning range of +10 to +35 per cent, with a centre near +25:
| Harvest-to-lean move | Lean-season price | Cost of the position | Result |
|---|---|---|---|
| +10% — flat curve | GH¢5.06 | GH¢5.45 | −0.39/kg (−$33/t) |
| +18% — break-even | GH¢5.45 | GH¢5.45 | nil |
| +25% — planning centre | GH¢5.75 | GH¢5.45 | +0.30/kg (+$25/t) |
| +35% — strong carry | GH¢6.21 | GH¢5.45 | +0.76/kg (+$63/t) |
Table 2 — Purchase at GH¢4.60/kg plus carry of GH¢0.85. Dollar figures converted at GH¢12/USD.
Two things follow from the shape of that table. The first is how narrow the working band is: the difference between a good storage year and a losing one is around eight percentage points of seasonal move, and the low end of an ordinary planning range sits below break-even. The second is that the position is convex in the wrong direction for anyone who has to size it in advance. It is decided by a price fifteen weeks after the money is committed.
That is a reason to bound the volume held rather than a reason not to store. But it is not a trade that survives being entered at full size on an assumed spread.
The term that actually sets the answer
Because interest dominates the carry, the break-even spread is largely a function of the cost of money, and moves with it almost one for one.
| Working capital rate | Interest over five months | Break-even seasonal spread |
|---|---|---|
| 15% p.a. — concessional or offshore | 6.3% | ~12% |
| 20% p.a. | 8.3% | ~14% |
| 25% p.a. | 10.4% | ~16% |
| 30% p.a. — Ghanaian commercial overdraft | 12.5% | ~18% |
| 35% p.a. | 14.6% | ~20% |
Table 3 — Warehouse, fumigation and shrink held constant at 5.9 per cent of the purchase price.
At fifteen per cent money, the +10 per cent year that loses GH¢0.39/kg at commercial rates is marginally profitable. Nothing about the grain, the warehouse or the weather has changed. The whole difference is the rate at which the position is financed.
The seasonal spread is not the storer's margin. It is the amount the storer's lender has already claimed.
This connects directly to the third condition set out in RN 001 — that there is no working capital at the point of surplus, so grain is sold when it is harvested rather than when it is worth most. That note stated the condition. This one prices it: at Ghanaian commercial rates, five months of carry consumes roughly two-thirds of a normal seasonal move before the storer sees any of it. The grower who sells at the trough is not making a mistake. They are declining to borrow at 30 per cent against a price fifteen weeks out.
The 2026 caveat, stated separately
The current crop year is a glut. Carry-over stocks flatten the seasonal curve, because the lean season is met partly out of the previous harvest rather than out of scarcity. A planning centre of +25 per cent is not the right expectation for a season entered on top of unsold stock, and the +10 per cent column is the relevant one until the carry-over clears. Storage sized on a normal-year spread in a glut year is the specific way this trade is usually lost.
Conditions that would alter this conclusion
- The spread range is a planning range. The +10 / +25 / +35 range reflects the general pattern in this market, not a distribution estimated from the archive, which spans one cycle. A monthly wholesale series of eight to ten years, by market, would show how often the move fails to clear break-even, and whether it fails in the same markets each time.
- The warehouse charge of GH¢0.18/kg over five months is a planning figure and will vary widely with store condition, whether the space is owned or rented, and fumigation frequency. It is the cheapest line to verify and the least consequential to the conclusion.
- Thirty per cent is a representative commercial working capital rate, not a quoted term. A storer with a genuine term sheet at a materially different rate is running a different business, and Table 3 is the map of how different.
The first carries the most weight.
Acting on this
Where the seasonal spread will not reliably pay for the carry, the alternative is not to take the position at all: a forward supply contract establishes the price before the season, marked against the Ghana Commodity Exchange, and removes the fifteen-week window in which the outcome is decided by something nobody controls.
The data behind this note is licensed.
With its audit trail, under the Birkdal Methodology.
Data & API →