Reserves, stocks and depletion: Reserve life and why it misleads
Understanding Commodities / Module 4: Reserves, stocks and depletion Part 2 of 5Reserve life and why it misleads
The number most often derived from reserves is reserve life: reserves divided by a year's production. It reads like a countdown. This part works the arithmetic for four tiles, shows how growth, concentration and joint production change the picture, and uses oil to show how reserve figures are revised and sometimes declared for reasons unrelated to geology.
2.1 Static reserve life on four tiles, and how it has moved
Static reserve life, or the R/P ratio, divides reserves by one year's output. With USGS figures for 2025:
Copper: 980 million tonnes of reserves against mine output of 23 million gives about 43 years. Zinc: 240 million against 13 million gives about 18 years. Lithium: 37 million tonnes against 290,000 tonnes (world output, excluding withheld US data) gives about 128 years. Phosphate rock: 73 billion tonnes against 250 million gives about 290 years.
Read literally, zinc runs out in the early 2040s and phosphate lasts into the 2300s. History says neither reading is sound. Tilton computed the same ratios from 1999 data: copper then had 28 years and zinc 25. Tilton and Guzmán repeated the exercise with 2015 reserves and found copper at 39 years and zinc at 17. After a quarter of a century of record mining, copper's clock shows more years than in 1999. Zinc's has hovered below 20 years for a decade while the mines kept running, and identified zinc resources, about 1.9 billion tonnes, are eight times reserves. Lead, at 16 years in 2015, was the lowest in Tilton and Guzmán's table; in 2025 it stands at 21.
The ratios differ across metals mainly because of the cost and benefit of proving ore, not because some metals are nearer exhaustion. A low R/P often means that firms see little reward in drilling further ahead. A high one can simply mean that a few very large deposits have been measured. The useful question to ask of any tile is not when the number reaches zero but whether it has been falling for decades despite exploration.
- USGS Mineral Commodity Summaries 2026, Copper web ↗ · pdf p. 77
- USGS Mineral Commodity Summaries 2026, Zinc web ↗ · pdf p. 217
- USGS Mineral Commodity Summaries 2026, Lithium web ↗ · pdf p. 123
- USGS Mineral Commodity Summaries 2026, Phosphate Rock web ↗ · pdf p. 147
- USGS Mineral Commodity Summaries 2026, Lead web ↗ · pdf p. 119
- Tilton (2003), On Borrowed Time? · pdf pp. 31-33
- Tilton & Guzmán (2016), Mineral Economics and Policy · pdf p. 226
2.2 Growth, concentration and joint production change the picture
Static life assumes output stays flat. If primary output grows at a rate g, a stock that would last N years at constant output lasts ln(1 + N x g) / ln(1 + g) years. Tilton and Guzmán show their tables at 0, 2 and 5 per cent growth for this reason. Applied to 2025 data, copper's 43 years fall to 31 at 2 per cent growth and 23 at 5 per cent. Lithium is the striking case: its mine output rose from 222,000 to 290,000 tonnes between 2024 and 2025, about 31 per cent in a year. At 10 per cent sustained growth its 128-year static life shrinks to 28 years; at 20 per cent, to 18.
The formula also shows the limits of the exercise. Tilton and Guzmán apply it to copper's whole crustal resource base and get 84 million years at constant output but only 723 years at 2 per cent growth. No demand grows at 20 per cent for decades; both are extrapolations, not forecasts.
Two other traps sit on the tiles. Concentration: Morocco holds 50 of the world's 73 billion tonnes of phosphate rock reserves, about 68 per cent, so a 290-year world figure says little about supply security for a country that cannot buy from Morocco. Joint production: many metals are by-products. Tilton and Guzmán explain that the molybdenum recovered from copper ore cannot exceed what that ore contains. Once all of it is recovered, a higher molybdenum price cannot raise supply, because nobody mines copper ore for its molybdenum. A by-product's reserves are really a statement about the host metal's mine plans.
2.3 Oil shows how reserve figures are revised and declared
M. King Hubbert, a geophysicist, argued that output from an oil province traces a bell curve, so once growth slows the peak can be dated. His forecast of a US peak around 1970 came true. Tilton reports a later follower, Deffeyes, arguing that world output could peak as early as 2003 and could not plausibly be delayed past 2009. Tilton's objection is that the method ignores price, technology and politics. Radetzki and Wårell describe what followed: the shift to deep water, oil sands and shale proved economically seamless. US crude output rose by more than 125 per cent between 2008 and 2018, almost all from shale, and the 2014 price collapse ended the peak-oil debate.
The Energy Institute series puts world proved reserves at 683 billion barrels in 1980, 1,001 billion in 1990 and 1,732 billion in 2020. The 1980s gain was OPEC's, in a few jumps: the UAE from 33 to 97 billion barrels in 1986, Iran from 59 to 93 the same year, Iraq from 72 to 100 in 1987, and Saudi Arabia from 170 to 255 in 1988. Non-OPEC reserves stayed near 247 billion through those years. The jumps came while OPEC was negotiating production quotas, and they are state declarations that, the Energy Institute warns, need not meet company reporting standards. Some never move: the UAE has reported 97.8 billion barrels every year from 1996 to 2020. Yergin, drawing on a different compilation, gives 615 and 917 billion barrels for 1985 and 1990, against 774 and 1,001 billion in the Energy Institute series.
In the Energy Institute series OPEC held about 70 per cent of proved reserves at the end of 2020, including Venezuela's 304 billion barrels. Yet Radetzki and Wårell note that OPEC's share of world output fell from 51.6 per cent in 1973 to 41.5 per cent 45 years later. Reserves held by states that do not invest are not supply.
- Tilton (2003), On Borrowed Time? · pdf p. 72
- Radetzki & Wårell (2021), A Handbook of Primary Commodities in the Global Economy · pdf pp. 176, 183-185, 235
- Energy Institute, Statistical Review of World Energy (data workbook) web ↗ · Oil - Proved reserves and Oil - Proved reserves history sheets
- Yergin (1991), The Prize · pdf pp. 675-676, 703, 1322
See it in the table
H₂O WaterP Phosphate rockCu CopperZn ZincMo MolybdenumLi LithiumCheck your understanding
Zinc reserves are 240 million tonnes and 2025 mine output was 13 million tonnes. What is the best reading?
240 divided by 13 is about 18. The ratio has stayed in a narrow band for decades because firms replenish reserves as they mine, and identified resources of about 1.9 billion tonnes are eight times reserves.Lithium's static reserve life is about 128 years. If output grew at 10 per cent a year indefinitely and reserves never changed, roughly how long would they last?
Using ln(1 + 128 x 0.10) / ln(1.10), the life is about 28 years. Compound growth shortens any fixed stock dramatically, though reserves would not in fact stay fixed.Why can a higher molybdenum price fail to raise by-product molybdenum supply much?
By-product output depends on the host metal. Once recovery from the copper ore being processed is complete, only more copper mining, driven by the copper price, can add supply.