Prices and cycles: How a commodity price is made

1.1 The price is set at the margin, by the last unit of supply that is needed

In a competitive market the price at any moment sits where short-run supply meets demand. In the short run capacity is fixed, so the supply curve is just the variable costs of existing mines, wells and farms ranked from cheapest to dearest. Analysts call this the cash-cost curve. With spare capacity the curve is flat and extra demand barely moves the price. Near full capacity it turns steep and then vertical, and small shifts in demand produce large price jumps.

The long run is different. Given time, capacity can be built, so the long-run supply curve shows the full cost, capital included, of the marginal new project. Radetzki and Wårell stress that nobody ever trades at this long-run price; it is the level the market price is pulled towards. Above it, investment floods in and eventually drags the price down. Below it, investment dries up and prices recover. Tilton and Guzmán add a useful division of labour: short-run swings are mostly about demand, long-run trends mostly about supply costs.

The real cost curve is blurry. Closure and restart are expensive, so a mine may run at a loss through a slump it expects to be short, and state firms may chase jobs rather than profit. Above all, prices feed back into costs. High prices breed slack and pull high-cost producers in; low prices force cost cutting and closures. Humphreys shows copper's curve rising and steepening between 2004 and 2007 as wages, equipment, energy, royalties and producer currencies all rose with the copper price.

• Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 102-109
• Tilton & Guzmán (2016), Mineral Economics and Policy · pdf pp. 112-117
• Humphreys (2015), The Remaking of the Mining Industry · pdf pp. 33-34, 108-110
• Yergin (1991), The Prize · pdf pp. 1288-1289

1.2 Exchanges discover the price, and most physical trade borrows it

An exchange gathers many buyers and sellers of a standard grade, with a clearing house standing between every pair of trades. Its central function is price discovery. Humphreys describes the London Metal Exchange setting its official prices for aluminium, copper, lead, nickel, tin and zinc in short open-outcry sessions called rings; those prices then serve as references for metal and ore traded worldwide. LME turnover runs at roughly 40 times annual metal production (Humphreys), yet little metal changes hands there: Radetzki and Wårell report that in 2018 LME physical deliveries were about 2.2% of world aluminium output, 3% of copper and 3.6% of nickel.

Not every commodity can support a futures market. Radetzki and Wårell's conditions: many buyers and sellers, hedgers willing to use it, enough instability to make hedging worthwhile, easy grading, and storability so spot and futures are linked. Tea fails on grading; niobium, with what Tilton and Guzmán count as two significant producers, on concentration.

Once an exchange price is established it tends to absorb rivals. Aluminium joined the LME in 1979 and within a decade its quote had displaced the Alcan producer price. Platinum pricing moved to NYMEX in 1956. After crude oil futures opened on NYMEX in 1983 and in London in 1988, OPEC stopped posting official prices and tried to defend a price band by adjusting supply instead. Exchange prices react to every piece of news, so they are more volatile than other prices. Radetzki and Wårell call this a feature: a smoothed price can hide an imbalance until it breaks.

• Humphreys (2015), The Remaking of the Mining Industry · pdf pp. 101-102
• Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 117-118, 121, 143, 147-148, 153-156
• Tilton & Guzmán (2016), Mineral Economics and Policy · pdf pp. 95-97, 102

1.3 Where no exchange works, prices are announced, negotiated or assessed

Most international commodity trade, Radetzki and Wårell judge, runs on bilateral contracts. Where an exchange price exists the contract simply references it. Copper concentrate, for example, is priced off the LME copper price less treatment and refining charges negotiated with the smelter. Without an exchange price, each pair must agree a price, slowly and privately. In manganese and, for decades, iron ore, one large supplier and one large buyer settled first and everyone else adopted that settlement as the benchmark, adjusting for grade.

Producer pricing is the second arrangement. A few sellers announce a take-it-or-leave-it price, as De Beers long did for rough diamonds and Amax did for molybdenum. Tilton and Guzmán list diamonds, platinum group metals, potash and borates as still largely sold this way. Producer prices change slowly, so in booms sellers ration and in slumps they offer quiet rebates; a series of producer prices misstates what was paid.

The third is the assessed price. Where prices are negotiated privately, trade journals and price reporters such as Metal Bulletin, Platts and Industrial Minerals survey traders in confidence and publish a reference price or range. Tilton and Guzmán note this for many minor metals, ferroalloys and lithium carbonate; since 2010 seaborne iron ore has been priced this way too.

The last is the transfer price inside an integrated firm, which can be whatever the tax position favours; bauxite is the clearest case. Where private arrangements dominate, small sellers with poor information tend to get worse terms.

• Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 119-123
• Tilton & Guzmán (2016), Mineral Economics and Policy · pdf pp. 96-98
• Humphreys (2015), The Remaking of the Mining Industry · pdf pp. 121-126

See it in the table

Check your understanding

According to Radetzki and Wårell, what is the long-run equilibrium price of a commodity?

Why did the major iron ore producers give up on annual benchmark pricing around 2010?