Why commodity prices swing so far (inelastic demand and supply meeting supply shocks), what volatility does to farmers, consumers and exporting countries, how a buffer stock holds a price band, and why such schemes usually fail. Through the coffee price of 2025 and 2026.
In February 2025 arabica coffee futures reached $4.41 a pound, the highest ever, after drought cut Brazil’s crop and prices rose 150% in a year. Roasters shrank packs and raised shelf prices; a jar of instant coffee became a thing people noticed on the receipt.
Then the rains came. In February 2026 Brazil’s crop agency Conab forecast a record 2026 harvest of 66 million bags, 17% up on the year before, and traders sold. By June 2026 arabica was about $2.50 a pound, the lowest since late 2024. Brazilian growers who had held back beans hoping for better prices were left holding them. Coffee drinkers, meanwhile, had changed almost nothing about their mornings.
Why do the prices of coffee, cocoa, wheat and oil swing so much more than the prices of cars or haircuts, and what can be done about it?
Source: ICE arabica futures via Trading Economics (all-time high 440.85 US cents per pound, February 2025; 249 cents, 4 June 2026); Conab, 5 February 2026.
Primary commodities are unprocessed goods from farming and mining: coffee, cocoa, wheat, copper, oil. Their prices are volatile, swinging far more than the prices of manufactured goods, for three reasons:
Steep demand meeting steep supply means any shift moves the price a long way. The cobweb adds to it: farmers plant on this year’s price and harvest into next year’s.
Governments respond with buffer stocks (an agency buys when the price falls to a floor and sells when it rises to a ceiling), guaranteed minimum prices, and diversification away from one crop.
Arabica coffee cost $4.41 a pound in February 2025, an all-time record. By June 2026 it was about $2.50.
World coffee consumption barely changed over those sixteen months. People drank roughly the same number of cups.
Why did the price nearly halve?
Brazil’s 2026 crop is forecast at a record 66 million bags, 17% more than the drought-hit year before. Buyers do not drink 17% more coffee just because it is cheaper, so the extra supply forces the price down a long way. Inelastic demand turns small quantity changes into large price changes.
The coffee market in a normal year. Demand is **steep**: habit, few substitutes, a small share of budgets, so buyers barely respond to price. Supply **S1** is also steep: this year’s crop is already on the trees, and new trees take four years. They cross at **E1**: 6 million bags at $8 a kilo.
Check: Both curves are drawn steep. What does the steep demand curve represent?
A steep demand curve means buyers barely change quantity when the price moves: coffee drinkers keep drinking.
Drought in Brazil. Supply shifts left to **S2**. Because buyers will not give up their coffee, the price has to rise a long way to ration the smaller crop: to **E2**, $12, up 50%, while quantity falls only a third. With steep curves, a modest shift in quantity becomes a large move in price. That is the arithmetic of volatility.
Check: Drought cuts the harvest and supply shifts to S2. The price rises from $8 to $12. What is the percentage change in price?
(12 − 8) ÷ 8 × 100 = 50%. Quantity fell by a third; the price rose by a half.
Compare with a good whose demand is elastic, the flatter grey line. The same harvest failure raises the price only to $10.67, because buyers switch away and less of a price rise is needed to clear the market. Elasticity is the whole difference between a commodity and a manufactured good: cars and haircuts have substitutes and can be made more of next month.
Check: If demand had been elastic (the flatter grey curve), the same harvest failure would have raised the price to $10.67 instead of $12. Why?
With elastic demand, a small price rise removes enough buyers to clear the market. The steeper the demand curve, the bigger the price move for the same shift.
Now the good year. Supply shifts right to **S3** and the price collapses to **E3**, $4. Farmers sell more bags for less money: revenue falls from $48m to $32m. Volatility hurts in both directions: a bad harvest leaves growers with little to sell, a good one leaves them with a price that may not cover costs. Countries that depend on one crop see their export earnings and tax revenue swing with it.
Check: A bumper harvest moves supply to S3 and the price falls to $4. Farmers sell 8 million bags instead of 6. What happens to their total revenue?
6 × 8 = 48; 8 × 4 = 32. With inelastic demand a good harvest lowers farm income: the paradox of plenty.
The classic remedy: a buffer stock. An agency sets a price band, here a floor of $6 and a ceiling of $10. In the bumper year the market price would be $4. At the $6 floor buyers want 7 million bags but farmers offer 9. The agency buys the 2 million surplus and stores it. The price stays at $6 and farmers earn 9 × $6 = $54m instead of $32m.
Check: A buffer stock agency sets a floor of $6 and a ceiling of $10. In the bumper year, at $6, farmers supply 9 million bags and buyers want 7 million. The agency must:
At the floor there is excess supply. Unless the agency buys it, the price falls below $6. It stores the beans for a bad year.
The poor year. The market price would be $12. At the $10 ceiling buyers want 5 million bags but the harvest is 3. The agency sells 2 million from stock and the price stops at $10. Steadier income for farmers, steadier prices for consumers, and in principle it pays for itself: buy at $6, sell at $10.
Check: In the poor year, at the $10 ceiling, buyers want 5 million bags and farmers supply 3 million. The agency:
At the ceiling there is excess demand. The agency meets it from the stock it built up in the good year, and the price stops at $10.
Other responses. A guaranteed minimum price without a stock: the government pays farmers the difference (deficiency payments), which costs taxpayers but creates no mountain. Futures contracts let a grower sell next year’s crop at today’s price, moving the risk to a trader. Diversification: Vietnam added pepper, cashews and rubber to coffee. And income support targets the farmer rather than the price.
Check: Which policy reduces the harm from volatile prices without trying to control the price itself?
Diversification spreads the risk; a futures contract fixes the price a farmer will receive before the harvest. Neither needs an agency to hold the market price.
The floor is drawn at $3, below the market price of $4. A floor below the equilibrium does nothing: the market clears at $4 on its own and there is no surplus for the agency to buy. A floor must be above the market price to bite.
Explain why the price of coffee is so volatile, and evaluate whether a buffer stock scheme is an effective response.
Coffee prices are volatile because both demand and supply are price inelastic. Demand is inelastic because coffee is habit-forming, has few close substitutes and takes a small share of household budgets, so a cheaper cup does not make people drink many more. Supply is inelastic within the year because the crop is fixed once harvested and new trees take about four years to bear. Both curves are steep, so when weather or disease shifts supply, the price has to move a long way to clear the market. In February 2025, after drought in Brazil, arabica reached a record $4.41 a pound; after a record 2026 harvest was forecast it fell to about $2.50, while consumption barely changed.
On the diagram a poor harvest shifts supply from S1 to S2 and raises the price from $8 to $12 while quantity falls by a third; a bumper harvest shifts supply to S3 and the price falls to $4. Because demand is inelastic, the bumper year lowers farmers’ total revenue, from $48m to $32m, so a good harvest can mean a bad year. Countries that depend on one crop see export earnings and tax revenue swing with the price, and investment in farms becomes a gamble.
A buffer stock scheme tries to hold the price in a band. In the bumper year the agency buys the surplus at a floor of $6 and stores it; in the poor year it sells from stock at a ceiling of $10. Farmers get a steadier income, consumers a steadier price, and in principle the agency profits by buying low and selling high.
**However,** the record is poor. Floors are set under pressure from farmers and end up above the long-run average price, so the agency buys every year and never sells: the EU’s butter mountains, and the tin buffer stock that went bankrupt in 1985. Storage is expensive and coffee deteriorates. A guaranteed price also encourages farmers to grow more of the crop already in surplus, and the scheme needs a large fund that speculators will test. Alternatives tackle the harm rather than the price: futures contracts let growers lock in a price, diversification into other crops spreads the risk, and income support protects smallholders without distorting planting. A buffer stock can smooth a price for a while, but it works only with a floor set realistically and a government willing to sell as well as buy, which is why most such schemes have failed.
Coffee prices: ICE arabica futures as reported by Trading Economics and Barchart; Brazil harvest forecasts from Conab, 5 February 2026. The curves on the diagram are for an invented market, chosen so the numbers are clear. Diagrams drawn with Rough.js and perfect-freehand, both MIT licensed.