Editorial Article

Sixty-Three Per Cent Is Not Idle Capacity

Updated October 19, 2026

Two sentences, one million tonnes, opposite conclusions

Chinese yellow phosphorus capacity is 1.58 Mt. Output in 2025 was about 1.0 Mt. Divide one by the other and you get 63 per cent — a number that appears in every bear note on the thermal phosphate chain, and the number behind the widely repeated claim that this market is running at roughly a third idle.

The same million tonnes can be described a second way. Effective capacity is being administered down to a band of 1.18–1.22 Mt, a 24.05 per cent reduction from nameplate. Against 1.20 Mt, that identical output is 83.33 per cent.

Denominator Output Utilisation
1.58 Mt — installed nameplate 1.00 Mt 63.29%
1.20 Mt — effective capacity band 1.00 Mt 83.33%

Both are true. Only one of them describes the constraint. And the difference between them — 580,000 tonnes of apparent headroom against 200,000 tonnes of real headroom — is the difference between a market with slack and a market with a queue.

Why the denominator is an instrument and not a measurement

Yellow phosphorus is not made in a chemical plant in the sense that a phosphonate is. It is made in an electric furnace that reduces phosphate rock with coke at roughly 13,000 kWh per tonne of P₄. The arithmetic of that is unforgiving and it is entirely bought in.

At RMB 0.42/kWh, the power is USD 769.01 a tonne. The rock requirement is stoichiometry rather than estimation: P₂O₅ carries 61.95 g of phosphorus in 141.94 g of molecule, so a tonne of contained P needs 2.2912 t of P₂O₅; at the 30 per cent concentrate grade a furnace is actually fed that is 7.637 t; with an 8 per cent process loss it is 8.248 t of rock. At USD 143 that is USD 1,179.46.

Together, rock and electricity are 52.80 per cent of a USD 3,690.14 tonne before an electrode is bought or a wage is paid. The industry’s theoretical margin on that price is 16.37 per cent — and neither of the two inputs that dominate the cost is set by the producer. Rock is priced on a benchmark. Power is priced on a tariff, and furnaces that are energy-cost residual suppliers run when the grid is cheap and idle when it is not.

So the only durable lever in this business is the one that governs who is permitted to run. Capacity policy is not an overlay on the supply story. It is the supply story, and it is why the denominator moves while the numerator does not.

What the shrinking denominator is worth to the chain above it

Roughly 9.2 Mt of the world’s 250 Mt of phosphate rock goes into an electric furnace rather than a fertiliser plant. That is 3.68 per cent of the tonnage and 8.70 per cent of the market’s value — and it is one hundred per cent of the industrial chemistry. The phosphite, the phosphonate, the phosphorus trichloride behind an electrolyte salt, the flame retardant: all of it sits on top of this one thermal branch, and the branch cannot be expanded by planting a mine.

Everything above it therefore prices as a conversion fee on an administered commodity. The phosphonate market is a USD 855m to USD 1,340m business growing at 5.78 per cent; purified wet-process acid supplies a different route entirely. Both are downstream of a denominator that policy controls, which is a strange thing to be downstream of and a useful thing to know when the utilisation number is quoted at you.

Two price series that disagree, and the disagreement is the point

The World Bank’s Morocco rock benchmark sat flat at USD 152.50 through 2024 and 2025, then reached USD 170.00 in August 2026 — up 11.48 per cent year on year. Yellow phosphorus did something else. It went from RMB 26,200 on 13 March 2026 to RMB 32,429.33 on 19 May: +23.78 per cent in nine weeks and +44.16 per cent on the year, after a fire in Guizhou on 14 May took 40,000 tonnes and four furnaces offline. By 8 October it was back to RMB 27,600–28,900.

The rock barely moved. The power and the availability did. That is the signature of a market whose price is set at the top of the cost stack rather than at the bottom, and it is exactly what you would expect if the binding variable is whether a furnace is permitted and able to run rather than whether rock exists.

What would falsify this

One: a real utilisation disclosure. If a provincial or industry source shows furnace operating rates consistent with output materially above 1.0 Mt against the 1.18–1.22 Mt band, the constraint is looser than this reading claims.

Two: the band being restored upward. An administered denominator that starts growing again is a supply expansion, and it would show up in exactly the way the cut showed up — in utilisation, before it showed up in price.

Three: two quarters below the theoretical margin. A sustained price under USD 3,690.14 a tonne would say the furnaces are running for cash rather than for margin, which is demand destruction dressed as administered scarcity, and it would invert the argument.

One caveat belongs in print rather than in a footnote. Because 81.71 per cent of this market’s value is fertiliser and feed phosphate, the headline 3.98 per cent growth rate is a fertiliser number with an industrial premium attached. The industrial reader should follow the 2.92 per cent elemental phosphorus volume line, not the value line. The full chain — rock, the thermal branch, the acid, the phosphonate — is set out in the accompanying phosphate rock and yellow phosphorus market report, and the molecule-level argument is in One Oxygen Is the Whole Market Boundary.

Frequently Asked Questions

Is the yellow phosphorus industry really running at only 63 per cent utilisation?

Against installed capacity of 1.58 Mt, yes — about 1.0 Mt of 2025 output is 63 per cent. Against the effective capacity band of 1.18–1.22 Mt that policy is steering the industry toward, the same output is 83.33 per cent. Both statements describe the same tonne of output. Only the second describes the constraint that binds.

Why does the denominator matter more than the output number?

Because in an electrothermal business neither input is set by the producer. Rock is priced on a benchmark and power is priced on a tariff — 13,000 kWh per tonne at RMB 0.42 is USD 769.01, and 8.248 tonnes of 30 per cent concentrate at USD 143 is USD 1,179.46, so 52.80 per cent of a USD 3,690.14 tonne is bought in before an electrode or a wage. The only lever left is who is permitted to run, and that is capacity policy, not demand.

How much of this market is actually industrial chemistry rather than fertiliser?

About 9.2 Mt of roughly 250 Mt of phosphate rock goes to an electric furnace rather than a fertiliser plant — 3.68 per cent of tonnage carrying 8.70 per cent of value. Because 81.71 per cent of the market's value is fertiliser and feed phosphate, the headline 3.98 per cent growth rate is a fertiliser number with an industrial premium attached; the industrial reader should track the 2.92 per cent elemental phosphorus volume line.

What evidence would show the utilisation argument is wrong?

Three things: a provincial disclosure of actual furnace operating rates that shows output materially above 1.0 Mt; the effective capacity band being restored upward rather than cut; or yellow phosphorus trading below the theoretical margin for two consecutive quarters, which would indicate demand destruction rather than administered scarcity.