Every chokepoint has a half-life

On November 10, China’s suspension of its broadest rare-earth export controls expires. The International Energy Agency estimates that full implementation could put $6.5tn of annual downstream production outside China at risk. Yet a blanket embargo remains unlikely because Beijing appears to value the network it can exploit more than a rival order it would have to sustain. If that judgement is wrong, restraint means something darker: preparation, not accommodation.

China imposed export controls on seven categories of medium and heavy rare-earth items in April 2025, widened the regime in October, then suspended the expansion in negotiations with Washington; automakers cut output, but Beijing demonstrated the weapon without exhausting it.

This power did not begin as a master plan. It grew from a cheaper processing method, disorderly competition, pollution that western economies willingly outsourced and later state consolidation. Its coercive value became unmistakable after the 2010 dispute with Japan. American sanctions and chip controls then helped turn a commercial concentration into an action-reaction spiral. Chokepoints are often built for efficiency or security before governments discover their political use.

A chokepoint’s half-life is the period before adaptation, evasion or the coercer’s own costs remove half its usable leverage. It depends on substitution speed, allied enforcement and trust in restored access. Selective licensing can extend it: a blanket embargo gives every injured state reason to finance an alternative, while a licence denied to one company leaves the rest trading. Applications identify end users and exposed production lines, preserving the instrument while improving the coercer’s map. Precision carries Bismarck’s risk: a system dependent on constant calibration may not survive a less disciplined successor.

Japan shows why adaptation may still stop halfway. China supplied 85 per cent of Japanese rare-earth imports in 2009. By 2020, Tokyo had cut the share to 58 per cent through new mines, conservation and a trade challenge. But China now accounts for 91 per cent of rare-earth refining and 94 per cent of permanent-magnet production. Projects outside China could produce more than 50 kilotonnes of mined rare earths by 2035, but less than 40 kilotonnes of refined material and only 18 kilotonnes of metals, alloys and magnets. A new mine is not an alternative supply chain if its ore still travels to the old refiner.

Chart 1  The bottleneck comes after the mine

China’s share of world output in 2024 and projected capacity outside China in 2035. The IEA thresholds are more than 50kt mined, less than 40kt refined, and around 18kt of metals, alloys and magnets. Source: International Energy Agency.

The IEA puts the capital needed for a diversified rare-earth chain at about $60bn over a decade, while annual production exposed to full Chinese controls exceeds one hundred times that sum. Markets underinvest because processing is dirty, difficult and vulnerable to price falls before new plants reach scale. Shortening this half-life requires patient capital, purchase commitments and processing capacity, rather than another announcement about mineral partnerships.

Hormuz presents a harder limit because geography cannot be reproduced. Oil flows through the strait averaged 21.6mn barrels a day in the final quarter of 2025 and 4.9mn in the second quarter of 2026. Washington offered a $40bn maritime reinsurance facility in April; by mid-May, it had placed no cover. The offer could absorb financial loss, but it could not make a ship master accept physical danger. Here the half-life is set by how long reserves, rerouting and political patience can carry consumers.

Chart 2  Hormuz oil flows fell by more than three-quarters

Quarterly averages. The EIA cautions that AIS signals from ships transiting the strait have been especially unreliable since the end of February 2026 and says its 2026 estimates are revised frequently. Source: US Energy Information Administration, Global Energy Security Data.

Six months into the war, the strait is neither simply open nor closed: tracking recorded seven vessel transits on September 10, against a prewar daily average above 130. The EIA cautions that AIS signals from ships in the strait have been especially unreliable since the end of February; Washington, meanwhile, has circulated higher counts. Each side, working from a different picture, may assume its leverage has a longer half-life than it does.

Financial chokepoints decay more slowly. Washington sanctioned Hengli Petrochemical on April 24, having targeted four other Chinese refineries in earlier actions. On May 2 Beijing issued its first blocking order covering all five, forbidding Chinese entities to recognise, enforce or comply with the American measures. In the first quarter, the dollar represented 57.13 per cent of official reserves, while dollar credit to non-bank borrowers outside the US reached $14.7tn by end-March, up 7.3 per cent in a year. The expanding network lengthens its half-life; rival law makes its authority less automatic. Sanctions become a contest over which market a company can afford to lose.

That is the deeper political test. A stable order needs enough power to deter defection and terms that enough states regard as legitimate. From Lebanon, where access to fuel, dollars and shipping is often decided elsewhere, the difference between consent and necessity is difficult to miss. Compliance under pressure should not be mistaken for belief in the rules.

The holder of a chokepoint should therefore treat denial as a reserve, not a revenue stream. Every use must be weighed against the processing plant, payment route, legal defence or technical standard it may finance. Denial may win a concession; it cannot found an order. That depends on keeping access credible and the system valuable enough that others choose to remain within it.