Two closures, one bypass

Bab el-Mandeb's traffic reappeared at the Cape of Good Hope — 97.9% of it. Hormuz's traffic reappeared nowhere.

Published 2026-08-09 · analysis

Two chokepoints have closed to traffic since 2023, and the transit counts say they closed in two different ways. Bab el-Mandeb emptied out from late 2023 under the Houthi campaign. The Strait of Hormuz emptied out in February 2026. Both are measured here the same way and out of the same file: AIS-observed daily transit calls from IMF PortWatch, averaged over a window before the closure against a window after it. The difference between the two is not the size of the loss. It is whether the ships turned up somewhere else.

Round the Red Sea, the ships went round Africa. Comparing 2023-01-01 to 2023-11-30 against 2026-01-01 to 2026-02-27, Bab el-Mandeb fell from 74.5 transits a day to 34.9 (-53.2%). Over the same two windows the Cape of Good Hope rose from 48.8 to 87.7 (+79.5%) — a gain of 38.8 a day against Bab el-Mandeb's loss of 39.7, so the Cape picked up 97.9% of what the Red Sea lost. Suez, which counts most of the same voyages a second time, moved -45.6%. Malacca, which no Houthi missile can reach, moved +9.7%.

The Persian Gulf has no round-Africa. Comparing 2025-09-01 to 2026-02-27 against 2026-03-07 to 2026-08-09, Hormuz fell from 63.8 transits a day to 5.9 (-90.7%), a loss of 57.9 a day. The Cape gained 3.2, which is 5.5% of it — and the Cape's before-figure here is already 89.3, the elevated level it reached absorbing the Red Sea, so it was not starting from a quiet baseline and had room to show a second surge. Suez moved -0.7% and Bab el-Mandeb +0.2%: flat. Malacca, the largest eastward route by which Gulf cargo reaches Asia, fell 20.4 transits a day (-8.5%) — the opposite of what rerouting looks like, and the same strait that sat at +9.7% through the Red Sea episode.

Four routes in this file could physically carry cargo that stopped transiting Hormuz: east through Malacca, west through Bab el-Mandeb and Suez, or round the Cape. Their net change across the Hormuz windows is -17.4 transits a day. That is a net and not a sum of the gainers: adding only the two routes that rose would report +3.3 a day of "absorbed" traffic while discarding 20.7 a day of losses on the routes the cargo would actually have to use. The other 23 chokepoints in PortWatch are not counted, and that is a limit of this piece rather than a closed argument: Lombok, Sunda and Makassar are real eastward alternatives to Malacca, and a complete accounting would test them too.

The rhyme is exact until the last line. A strait closes; its transit count collapses by 53.2% in one case and by 90.7% in the other; the world keeps wanting the cargo. Then Bab el-Mandeb's traffic reappears at the Cape of Good Hope, and Hormuz's reappears nowhere — not at the Cape, not at Suez, not at Bab el-Mandeb, and least of all at Malacca, which fell. The detail that differs is geography, and it is the whole story: every voyage through Bab el-Mandeb can be sailed round Africa at the cost of days and fuel, and the Persian Gulf has exactly one sea exit. When the Red Sea closed the cargo moved. When Hormuz closed the cargo stopped.

† Caveats. Transit counts are AIS-observed and are a LOWER BOUND: GPS jamming, AIS spoofing and dark sailings mean real traffic is higher than counted, and the incentive to sail dark is highest in exactly the waters and months this piece is about — so the true collapse at Hormuz is smaller than -90.7%, and so is the true gain anywhere else. The two comparisons are also not equally clean. The Red Sea one leaves 109 weeks between its windows (334 days before, 58 after); the Hormuz one leaves 7 days (180 before, 156 after). Years of unrelated shipping growth can accumulate inside the first gap and cannot inside the second, so treat the Red Sea decimals as the looser pair. The Cape of Good Hope carries ships that never went near either strait, so "97.9% absorbed" is a match of daily counts, not ship-for-ship identity. That window deliberately stops on 2026-02-27, the day before Hormuz closed: run it to the present and it swallows a second closure, which pushes the absorbed share past 100% and makes the Cape look like it gained more than the Red Sea lost. Every window that avoids the overlap lands between 96% and 98%. PortWatch revises history retroactively, so these are the current vintage. And this is descriptive, not causal — the Malacca decline is an out-of-sample check that held, not a proof of mechanism.

Open this piece in the interactive site (charts, timeline, related events) →