The Taiwan Strait is China's chokepoint too

1. Two straits, and the bigger one is the one nobody prices

Per CSIS 2024 figures, the Taiwan Strait carries about $1.3tn of trade against Malacca's $963bn. The Taiwan figure represents 33% of Chinese imports and 16% of exports, including $986bn of electrical and machinery goods. Malacca carries 21% of imports and 14% of exports. Between them the two straits handle roughly 21% of all global maritime trade.

One caution before those numbers travel further, because the pair does not quite reconcile. Per our own recomputation of the CSIS figures, $1.3tn against $963bn computes to 35% more, not the 33% usually quoted. For exactly 33% the Taiwan figure would need to be about $1,281bn. The $1.3tn is a rounding, so quote "about a third more" or quote the precise pair, but not both in the same sentence.

Malacca gets the attention because it is the classic oil argument and has been the subject of strategic writing for decades. The "Malacca dilemma" has a name. The Taiwan Strait equivalent does not.

On the numbers, though, the Taiwan Strait is larger by value, carries more machinery, and doubles as domestic coastal infrastructure. It is the more consequential of the two and the less modelled.

2. Why an import shock becomes a logistics collapse

The distinction rests on something mundane: what flag the ships fly.

Malacca is an international sea lane. If it closes, cargo reroutes. Longer, more expensive and slower, but it moves. Ships go around Indonesia through the Lombok or Sunda straits, and the world absorbs the extra days.

The Taiwan Strait carries Chinese-flagged vessels moving goods between Chinese ports. Coal from the north to the industrial south. Ores and metals along the seaboard. Refined fuel between coastal refineries and inland distribution. That is domestic freight infrastructure that happens to travel on water, and China moves an enormous share of its bulk cargo that way because it is far cheaper than rail or road.

Close the strait and you have not merely interrupted trade with the world. You have severed a domestic artery while simultaneously cutting a third of imports.

Our read, not a sourced finding: that simultaneity is what makes the scenario asymmetric. A country can ration imports, and every wartime economy has. Rationing imports while your internal freight network is also degraded is much harder, because the rationing mechanism itself depends on being able to move things around. No source we found models the two together.

3. What China actually imports, and from whom

Per UN Comtrade, the import base is narrow and physical rather than broad and consumer-facing. Fuels, ores and metals are about 34% of imports. Seaborne oil exposure alone runs to roughly 7.8 million barrels a day, from around 14 million consumed daily, about 70% imported, and about 80% of that arriving by sea.

Then there is the detail that makes the picture close on itself. The single largest import source is Taiwan, at $218bn. It appears in trade data as "Other Asia, nes", a customs-reporting convention rather than an accident, and the content is integrated circuits and electronic components.

So the strait that carries a third of China's imports also leads directly to the supplier of its largest single import line. The chokepoint and the dependency point at the same island, which means a contingency does not present China with a choice between two problems. It presents one event that causes both.

4. The export figures are not what they look like

Two accounting artefacts need stripping out before anyone draws conclusions from China's bilateral trade numbers. Both are well known to trade economists, and both routinely appear in strategic analysis as though they were demand signals.

The first is Hong Kong. It is the number-two export destination at $291bn, against a population of about 7.5 million people. Work that out and it is roughly $38,800 of imports per resident per year, which is not final demand under any reading anyone would defend. Hong Kong is an entrepot. Goods pass through and are re-exported onward, and the destination data records the first stop rather than the last.

The second is round-tripping. China itself appears as a $117bn import source in its own data, goods leaving and returning through Hong Kong or bonded zones for VAT-rebate arbitrage and processing-trade documentation rather than any economic activity.

Neither figure represents demand. Both inflate the apparent scale of China's trade relationships, and anyone building an exposure model should remove them before starting.

5. The processing signature, in one ratio

One number reveals what China's largest export sector actually does. Electrical machinery is the biggest export chapter at $927bn, and the same chapter imports $584bn.

Imports equal 63.0% of export value inside a single product category.

That ratio is the arithmetic fingerprint of assembly rather than manufacture. High-value components arrive, finished electronics leave, and the domestic value added is the difference between them: substantial in absolute terms, but a minority of the final value.

It is also why the Taiwan dependency is structural rather than commercial, and the distinction matters for anyone thinking about substitution. A commercial dependency can be re-sourced at a price. You find another supplier, pay more, accept a quality difference. You cannot substitute the supplier of the component that constitutes most of the value of your largest export line, not because it would be expensive but because the alternative capacity does not exist at that scale anywhere on earth.

6. The customer base has already moved

While the chokepoint exposure stayed where it was, the destination map changed underneath it.

Per China Customs data, ASEAN has overtaken both the US and the EU as China's top export destination, reaching 16.7% by 2024 and about 18% by 2025. The US share contracted to somewhere between 10.8% and 14.7%, and that spread is a source-variance artefact between exports-only and total-trade reporting, with both figures in circulation. We carry the range rather than picking one.

Transshipment through Vietnam and Mexico tracks US tariff escalation closely enough to be measurable in the data, which suggests a meaningful part of the shift is routing rather than genuine reorientation of demand.

For the chokepoint argument this runs in an uncomfortable direction. More ASEAN trade means more traffic through exactly the waters under discussion, so commercial diversification away from Western markets has increased China's maritime concentration rather than reducing it.

The trade strategy that reduced political exposure to Washington increased physical exposure to a strait. Those two hedges were designed by different people solving different problems, and they work against each other.

7. The slower constraint underneath

Two structural facts sit beneath all of this and move on a much longer clock. Neither is a chokepoint risk. Both determine how much shock absorption exists when one arrives.

The first is demography. Per World Bank projections, China's working-age cohort falls from 984 million in 2024 to 745 million by 2050, a decline of 24.3% and a loss of 239 million people. Fertility of 1.0 to 1.2 is 48–57% of the replacement rate, and the population peaked in 2021 and has now declined for three consecutive years.

That is not a forecast in the ordinary sense. The people who will be of working age in 2050 have already been born, and the ones who have not been born cannot be conjured by policy.

The second is the balance sheet. Per IMF and CFR figures, property sales have contracted 47% by value and 50% by volume since the July 2021 peak, against household wealth that is 59–70% housing. Local government debt runs 35–38% of GDP on-budget, with off-budget financing-vehicle debt estimated at a further 44–50%, taking total public-sector debt to roughly 125% of GDP on the broadest measure.

Put those together and the implication for a chokepoint scenario is specific. A state absorbing a major trade disruption normally does so with fiscal expansion and by drawing on labour flexibility. China's capacity to do the first is more constrained than its headline growth suggests, and its capacity to do the second falls every year for reasons no policy reverses.

A shock arriving in 2030 lands on a materially thinner cushion than the same shock in 2015 would have.