Reserves, stocks and depletion: Inventories: the buffers between harvest, mine and market

5.1 Commercial and exchange stocks set the shape of the forward curve

Firms hold inventory because having material on hand has value: it covers a surprise order or keeps a plant running through a supply break. Kaldor and Working called this benefit the convenience yield; net of storage costs, Geman notes, it can turn negative when stocks are plentiful.

The theory of storage ties spot and futures prices together through these stocks. In a well-supplied market futures can trade above spot, a contango, but only by the cost of carrying the metal. Radetzki and Wårell give the arithmetic: with storage and deterioration at 3 per cent a year and interest at 4 per cent, the 12-month price cannot exceed spot by much more than 7 per cent, or traders would buy spot, store it and sell forward. That trade also makes contango a cheap way to finance surplus stock, which is why Blas and Farchy describe traders filling tanks and warehouses in the 2008-09 glut. When supply is short, spot can rise above futures, a backwardation, and there is no arbitrage ceiling, because nobody can borrow metal from the future.

So prices and volatility move inversely with stocks. Geman shows this for six LME metals from 1995 to 2003, and cites work finding soybean volatility an inverse function of inventory. Pirrong adds that spreads swing more when stocks are low.

Stocks also amplify demand shifts. In Radetzki and Wårell's example, if three downstream stages each hold half a period's sales in stock, a 5 per cent rise in final demand raises demand for the raw material by almost 17 per cent as each stage rebuilds. Tiles report the visible part: LME warehouses held 217,000 tonnes of lead at the end of October 2025, and US refined copper stocks rose from 123,000 to 450,000 tonnes during 2025.

• Geman (2005), Commodities and Commodity Derivatives · pdf pp. 36-37, 195-196
• Radetzki & Wårell (2021), A Handbook of Primary Commodities in the Global Economy · pdf pp. 110-111, 157
• Pirrong (2014), The Economics of Commodity Trading Firms · pdf p. 14
• Blas & Farchy (2021), The World for Sale · pdf pp. 81-82, 239
• USGS Mineral Commodity Summaries 2026, Lead web ↗ · pdf p. 119
• USGS Mineral Commodity Summaries 2026, Copper web ↗ · pdf p. 76

5.2 Government stockpiles are cheap insurance that is hard to manage

Tilton and Guzmán argue that stockpiles are the most cost-effective defence against temporary disruptions such as wars, embargoes and export bans, which strike suddenly but rarely last more than a few years. Against true depletion, which would come slowly and persist, they are useless.

The US built the largest programme. After the Second World War and Korea it held some 80 commodities; by 1973 the stockpile was worth $6.7 billion, and the tin alone equalled a year of world consumption. Radetzki and Wårell list the problems. Sizing required war scenarios that kept changing; in 1973-74 about 90 per cent was declared surplus; disposals depressed prices; buying tended to happen when prices were high and selling when they were low. By 2010 it held 28 commodities worth $1.4 billion. Pitron likens 1990s sell-offs to spending savings in good times, and Scheyder notes that the National Defense Stockpile had sold its rare earths by 2010. The USGS now reports 54 commodities at 10 sites, with fiscal 2025 potential purchases including 300 tonnes of neodymium-praseodymium oxide and 450 tonnes of magnet block. For fiscal 2026 the annual plan was not released.

Oil shows a better design. IEA importers must hold at least 90 days of net imports. In mid-2019 members controlled about 4.5 billion barrels, against a loss of about 1 billion barrels in the largest disruption since 1973, the Iranian revolution.

Humphreys describes China's strategic metals stockpile, held by the State Reserve Bureau (SRB), which in 2009 reportedly bought 590,000 tonnes of aluminium, 235,000 of copper and 159,000 of zinc while prices were low. Congo replaced a temporary 2025 cobalt export ban with quotas, expected at 96,600 tonnes a year in 2026 and 2027, of which 9,600 tonnes go to a government stockpile.

• Tilton & Guzmán (2016), Mineral Economics and Policy · pdf pp. 221, 238
• Radetzki & Wårell (2021), A Handbook of Primary Commodities in the Global Economy · pdf pp. 210-213
• Pitron (2020), The Rare Metals War · pdf p. 64
• Scheyder (2024), The War Below · pdf p. 113
• USGS Mineral Commodity Summaries 2026, Introduction, Cobalt and Rare Earths web ↗ · pdf pp. 13, 75, 157
• Humphreys (2015), The Remaking of the Mining Industry · pdf pp. 13, 68, 76

5.3 Crop stocks: the stocks-to-use ratio and where the grain sits

Crops have no reserves in the ground. Their buffer is grain carried from one harvest to the next. Geman sets out the balance sheet: supply is beginning stocks, production and imports; use is domestic consumption, for food and feed, plus exports; what remains is ending stocks, the carry-over that links prices across crop years.

The key ratio is ending stocks divided by use. Geman's example is 540 million bushels of wheat against use of 2.7 billion, or 20 per cent, and she treats 20 to 40 per cent as the normal range. Below it the market is tight, and a harvest shock against steady demand can send prices very high. That is why the USDA's monthly WASDE report moves markets when it surprises.

The October 2026 WASDE projects 2026/27 world ending stocks of 276 million tonnes of wheat against use of 828 million, a ratio of 33 per cent; corn, 280 against 1,320, or 21 per cent; rice, 199 against 539, or 37 per cent. But China is projected to hold 119 million tonnes of the wheat (43 per cent), 165 million of the corn (59 per cent) and 107 million of the rice (54 per cent), stocks kept for domestic food security that rarely reach the world market. Without China the ratios fall to 23 per cent for wheat, 12 per cent for corn and 23 per cent for rice.

For exporters the denominator should include exports. The major wheat exporters (WASDE's group, which excludes the US) are projected to end the year with 48 million tonnes against domestic use of 192 million and exports of 154 million, about 14 per cent. US corn, on the same basis, is projected at about 11 per cent. These are the stocks that can actually reach a buyer.

• Geman (2005), Commodities and Commodity Derivatives · pdf pp. 154-155, 173
• USDA, World Agricultural Supply and Demand Estimates, October 2026 web ↗ · WASDE pp. 8, 18-19, 22-25

5.4 Thin stocks turned modest shocks into the food price spikes

The early 1970s set the pattern. Morgan records that world grain stocks fell by 40 million tonnes in 1972, and US wheat stocks from 23 to 6 million tonnes; grain prices then ran out of control in 1973 and 1974. He argues that countries panicked because the grain was under no government's control and went to the highest bidder. A US proposal for a 30 million tonne world wheat reserve collapsed in 1979 over who would pay and when stocks would be released.

The 2007-08 spike followed a long run-down. Geman reported in 2004 that world wheat consumption had exceeded production for six consecutive years and stocks-to-use was at its lowest in a long time, with China drawing on its own reserves. Blas and Farchy record soybeans more than doubling from $5.40 a bushel in mid-2006; by 2008 China bought over half of world soybean trade. Soybeans peaked above $16 before halving in the crash. Radetzki and Wårell summarise Baffes and Haniotis on causes: no sign that Asian food demand accelerated, a biofuel effect too small to be decisive, mixed evidence on speculation, and a dominant role for the tighter link between energy and farm prices.

The 2010-11 spike had a clear trigger. Drought hit Russia's grain belt in summer 2010. On 5 August Moscow banned exports, two days after the head of Glencore's Russian grain business publicly urged it; wheat rose 15 per cent in two days, and Glencore had already taken long positions in wheat and corn. From June 2010 to February 2011 wheat more than doubled. Egypt bought half its wheat from Russia, and panicked importers bought more. Blas and Farchy count food inflation among the sparks of the Arab Spring.

In each case the shock was modest relative to world output. Thin stocks in trading countries, and the hoarding and export bans they provoke, made it violent.

• Morgan (1979), Merchants of Grain · pdf pp. 486-489
• Geman (2005), Commodities and Commodity Derivatives · pdf p. 163
• Blas & Farchy (2021), The World for Sale · pdf pp. 231-238
• Radetzki & Wårell (2021), A Handbook of Primary Commodities in the Global Economy · pdf pp. 136-138

See it in the table

Check your understanding

Storage costs 3 per cent a year and interest is 4 per cent. Spot copper is $10,000 a tonne. Which 12-month futures price is inconsistent with the theory of storage?

World corn stocks-to-use for 2026/27 is projected at about 21 per cent, but about 12 per cent without China. Why is the lower figure more relevant to an importer such as Egypt?

What did Radetzki and Wårell identify as a central weakness of the US strategic metals stockpile?