# Supply, demand and the cost curve: Market structure: competition, cartels and the state

Source page: https://www.materialinsights.org/learn/commodities/supply-demand-costs/p5/

[Understanding Commodities](https://www.materialinsights.org/learn/commodities/) / [Module 3: Supply, demand and the cost curve](https://www.materialinsights.org/learn/commodities/supply-demand-costs/) Part 5 of 5

## Market structure: competition, cartels and the state

Most commodity markets are competitive in the sense that matters: no seller has both the ability and the incentive to hold the price up for long. Some are not, and cartels and governments have often tried. This part covers how to judge market power, the arithmetic of cartels, why buffer stocks failed, and how state producers change supply.

### 5.1 Market power needs both the ability and the incentive

Concentration measures the ability. The four-firm ratio adds the top four shares; the Herfindahl-Hirschman index (HHI) sums all firms' squared shares, on a scale of 0 to 10,000. US merger guidelines of 1982 treated an HHI under 1,000 as unconcentrated and over 1,800 as highly concentrated. On Tilton and Guzmán's 2013 figures, niobium, platinum, rhodium, lithium and palladium exceeded 1,800, potash, diamonds and uranium sat in between, and iron ore, copper, nickel, zinc, gold and silver were below 1,000. Ownership matters: De Beers controlled about half of world diamond output through its marketing arm, though Alrosa had the larger equity share.

Ability is not enough. A producer that restricts output gains for a while, but the higher price draws in rivals and pushes buyers to substitutes. If the later losses outweigh the early gains in present-value terms, it has no incentive to restrict and behaves competitively. Chile, with a third of world copper, is no Saudi Arabia, Tilton and Guzmán argue: its costs are not far below others', so a Chilean cut would mostly hand market share to Peru, Indonesia and the rest.

Producer pricing is a clue, not proof; Radetzki and Wårell's classic case is De Beers selling rough diamonds take-it-or-leave-it. Potash is another: in 2013 four firms held 67 per cent of output, and an Eastern cartel operated from 2005 to 2013 alongside a Western one. Yet the Potash tile shows the cycle intact, from $736 a tonne in September 2021 to a ten-year high of $1,202 and down to $367.5 in September 2026. Conway ties the 2022 shortage to the war in Ukraine.

- Tilton & Guzmán (2016), Mineral Economics and Policy · pdf pp. 99-103, 178-185, 191-193
- Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 120-121
- Conway (2023), Material World · pdf p. 140

### 5.2 Cartel arithmetic explains why most commodity cartels fail

A cartel raises revenue by cutting supply only if demand for its own output is inelastic. Radetzki and Wårell, and Tilton and Guzmán, give the same formula. In absolute values, the elasticity facing the cartel equals world demand elasticity plus outsiders' supply elasticity times the outsiders' share, all divided by the cartel's share. Below 1, cuts raise revenue. Radetzki and Wårell keep the minus sign on demand elasticity, so their tables show negative values.

All three ingredients worsen with time. Buyers substitute, outsiders build mines, and members cheat.

The producer-power wave of the 1970s tested this. Jamaica sharply raised bauxite levies in 1974; its share of non-socialist output fell from 22 per cent to 12 per cent by 1982 as Australia, Guinea and Brazil gained. Morocco raised its phosphate rock price from $14 to $42 a tonne in January 1974 and to $63 that July, but by 1977 the real price was back to pre-cartel levels. CIPEC's members held only 37 per cent of world mine supply in 1975 (Radetzki and Wårell); Tilton and Guzmán cite estimates of 1.08 to 2.34 for the elasticity facing its output. It was dissolved in 1988. Uranium was the brief exception: near-zero demand elasticity took the spot price from under $7 a pound in late 1973 to over $40 by mid-1976, until new deposits and excess utility stocks reversed it.

OPEC divides the authors. Tilton and Guzmán write of OPEC pushing up oil prices in 1973, while noting that its influence is debated. Radetzki and Wårell deny it much effect: demand growth might have sustained the 1970s rises anyway, quotas since 1983 were routinely breached (output beat ceilings by 6.9 per cent on average to 2001), and OPEC's share of world oil fell from 51.6 per cent in 1973 to 41.5 per cent in 2018. Stagnant state-run capacity, and Saudi Arabia acting alone, mattered more.
 Worked example

#### Running the cartel formula

Take a group with 60 per cent of world supply. With world demand elasticity of 0.2 and outsiders' supply elasticity of 0.4, the cartel faces 0.2 plus 0.4 times 0.4, divided by 0.6: an elasticity of 0.60. That is below 1, so cutting supply raises revenue. A few years later buyers have adjusted to an elasticity of 0.4 and outsiders to 0.7. Now the cartel faces 0.4 plus 0.4 times 0.7, divided by 0.6: about 1.13, and further cuts lose revenue. Same share, different decade (Radetzki and Wårell).

- Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 216-236
- Tilton & Guzmán (2016), Mineral Economics and Policy · pdf pp. 14, 174, 192-195

### 5.3 Buffer stocks can smooth prices only around the right price

Not every attempt to manage prices was a producer cartel. From the 1950s to the 1980s producing and consuming governments ran international commodity agreements for sugar, wheat, coffee, cocoa, rubber and tin. Their stated aim was to keep prices within a band, using export quotas and buffer stocks: buy when the price is low, sell when it is high.

Radetzki and Wårell identify two flaws. The first is that stabilising price need not stabilise income. When prices swing because demand swings, as for metals, a fixed price can lower average revenue over the cycle. When prices swing because harvests swing, a high price in a bad year compensates for low volume; fixing the price removes that cushion and makes farm income less stable, not more.

The second flaw proved fatal. Stabilisation needs someone to know the long-run equilibrium price. In practice agreements drifted into defending a price above equilibrium, and a buffer stock defending too high a price must keep buying until its funds run out. Government stockpiles complicated matters further. In 1962, with the tin buffer stock at 51,000 tonnes, the US declared 150,000 tonnes of its 350,000-tonne strategic tin stock surplus, and its sales in 1963 to 1966 added 10 per cent to world mine supply. Gilbert's review of six agreements with economic clauses, cited by Radetzki and Wårell, found that they generally failed to reduce volatility. By the early 2000s none of the surviving agreements tried to manage prices; they had become forums for statistics and consultation.
 Case

#### The tin crisis of 1985

The International Tin Agreements held the tin price up for two decades. In the early 1980s the International Tin Council, dominated by producers, defended a price well above the long-run level. Export restrictions were not enough, so the buffer-stock manager kept buying, then borrowed against the stockpile to pay futures margins. In October 1985 the money ran out; he defaulted, leaving debts above $1 billion. LME tin trading was suspended until June 1989 and reopened far lower (Radetzki and Wårell). Tilton adds that tin stayed depressed for five to ten years.

- Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 31, 151, 274-276
- Tilton (2003), On Borrowed Time? · pdf p. 58
- Tilton & Guzmán (2016), Mineral Economics and Policy · pdf p. 101

### 5.4 State producers and export controls change the supply curve

Governments shape supply as owners as well as regulators. Between 1956 and 1974 roughly a quarter of all foreign direct investment in developing countries was nationalised, about 60 per cent of it without compensation, with mines and oilfields prominent. The state-controlled share of world copper mining peaked at 58.7 per cent in 1984 and fell to 25 per cent by 2009 (Radetzki and Wårell).

State firms tend to have higher costs, for three reasons: disorder after nationalisation, the cost of social goals such as jobs and regional development, and weaker pressure to cut costs when performance has many yardsticks. Recovery after takeover took five years in Venezuelan iron ore, over twenty in Indonesian tin, and never finished in Zambian copper. Governments often drained their firms' profits, starving reinvestment; in oil, Radetzki and Wårell argue, the resulting capacity stagnation mattered more to prices than OPEC's quotas. State firms may also keep producing through slumps, which blunts supply's response to low prices.

Export controls are the other lever, and a dominant share makes them bite. China's rare earth export quotas fell from 65,000 tonnes in 2005 to 50,000 in 2009, and 2010 exports were only about 30,000 (Pitron). Tilton and Guzmán reckon that cut removed about 16 per cent of world supply, and prices surged. The power faded as high prices revived mines elsewhere and even made Ukrainian uranium tailings worth processing, adding some 3,000 tonnes in 2011. For any dominant state seller the question is the same one a cartel faces: how fast will others respond?
 Case

#### Rare earths in 2010, cobalt in 2025

On 22 September 2010, two weeks after a Chinese trawler clashed with the Japanese coastguard near the Senkaku Islands, Chinese rare earth deliveries to Japan stopped without any declared embargo, for six months by one Osaka importer's account (Pitron). In 2025 Congo, with about 73 per cent of mined cobalt that year (USGS; the Cobalt tile shows 74.5 per cent for 2024), banned exports in February over low prices, then set quotas of 96,600 tonnes a year for 2026 and 2027, including 9,600 tonnes for a state reserve, against 230,000 tonnes mined in 2025. The LME cash price rose from $11.84 a pound in 2024 to about $15 in 2025 (USGS).

- Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 235-236, 243-252
- Pitron (2020), The Rare Metals War · pdf pp. 73-76
- Tilton & Guzmán (2016), Mineral Economics and Policy · pdf pp. 69-71, 190-191
- USGS Mineral Commodity Summaries 2026, Cobalt [web ↗](https://pubs.usgs.gov/publication/mcs2026) · Salient Statistics; Events, Trends, and Issues
- USGS Mineral Commodity Summaries 2026, front matter [web ↗](https://pubs.usgs.gov/publication/mcs2026) · Significant events affecting U.S. mineral commodities in 2025

### See it in the table
[U Uranium](https://www.materialinsights.org/energy/uranium/)[K Potash](https://www.materialinsights.org/commodities/potash/)[P Phosphate rock](https://www.materialinsights.org/commodities/phosphate/)[Cu Copper](https://www.materialinsights.org/commodities/copper/)[Zn Zinc](https://www.materialinsights.org/commodities/zinc/)[Ni Nickel](https://www.materialinsights.org/commodities/nickel/)[Fe Iron ore](https://www.materialinsights.org/commodities/iron_ore/)[Nb Niobium](https://www.materialinsights.org/commodities/niobium/)[Au Gold](https://www.materialinsights.org/commodities/gold/)[Ag Silver](https://www.materialinsights.org/commodities/silver/)[Pt Platinum](https://www.materialinsights.org/commodities/platinum/)[Pd Palladium](https://www.materialinsights.org/commodities/palladium/)[Rh Rhodium](https://www.materialinsights.org/commodities/rhodium/)[Dm Diamonds](https://www.materialinsights.org/commodities/diamond/)

### Check your understanding

Under the 1982 US merger guidelines described by Tilton and Guzmán, how is an industry with an HHI of 800 classed?
 Below an HHI of 1,000 the guidelines treat an industry as atomised enough to allow mergers; most large mineral industries, including copper and iron ore, fell in this group in 2013.

A cartel holds 60 per cent of supply, world demand elasticity is 0.2 and outsiders' supply elasticity is 0.4. What elasticity does the cartel face, and can it raise revenue by cutting output?
 0.2 plus 0.4 times 0.4 is 0.36; divided by 0.6 that gives 0.60, below 1, so a cut raises revenue, at least until elasticities rise.

Why did the International Tin Council's buffer stock collapse in 1985?
 Defending a price above equilibrium forced ever larger purchases financed by borrowing against the stock, until the manager defaulted in October 1985.

[← Part 4: Inelastic supply: spikes, lead times, by-products and harvests](https://www.materialinsights.org/learn/commodities/supply-demand-costs/p4/)[Mark read · Wrap-up →](https://www.materialinsights.org/learn/commodities/supply-demand-costs/review/)

## Other sections

- [Module overview](https://www.materialinsights.org/learn/commodities/supply-demand-costs/index.md)
- [Demand: derived, cyclical and tied to development](https://www.materialinsights.org/learn/commodities/supply-demand-costs/p1/index.md)
- [Substitution, technology, and demand for food, feed and fuel](https://www.materialinsights.org/learn/commodities/supply-demand-costs/p2/index.md)
- [Supply: the cost curve, the marginal producer and rent](https://www.materialinsights.org/learn/commodities/supply-demand-costs/p3/index.md)
- [Inelastic supply: spikes, lead times, by-products and harvests](https://www.materialinsights.org/learn/commodities/supply-demand-costs/p4/index.md)
- [Key terms, quiz and sources](https://www.materialinsights.org/learn/commodities/supply-demand-costs/review/index.md)