Bab el-Mandeb traffic halves as Middle East shipping risk becomes a dual bottleneck

The Analysis

Clarksons Research published an update at 10:00 on 27 July 2026. The Bab el-Mandeb Strait averaged about 31 transits a day on 25 and 26 July. That is roughly 50 percent below the second-quarter average measured by tonnage. Over the same two days Hormuz was near standstill. Houthi forces had declared a naval blockade on Saudi vessels on 20 July. Two COSCO-managed VLCCs, Xin Long Yang and Cosnew Lake, cleared Bab el-Mandeb on 23 July carrying about 4 million barrels of Saudi crude to China, after pausing two days and switching their AIS destination fields before transit. Those are the published facts. No adjectives, only the numbers.

My habit as a rate analyst is to distrust the headline before I trust the number. The headline says traffic halved and a chokepoint seized. The number underneath says 31 ships a day still went through. That is the gap between a closed gate and a toll booth that just got expensive. We price the toll. We do not panic about the gate.

So the rest of this is arithmetic, not commentary. I will take 31 a day apart, show which costs it adds to your freight, tell you when those costs hit your quote, and then give you the exact line to lock before they do. The line is not a slot. It is a ceiling.

Start with the base figure, 31 transits a day. A transit is one ship moving through the strait. Across 25 and 26 July that is about 62 passages total. Clarksons measured the drop by tonnage, not by ship count, and put it at roughly 50 percent below the second-quarter mean. Fifty percent below the mean means the current level is about half the mean, so the Q2 baseline was on the order of 62 against today's 31. Read the caliber before the drama. A tonnage-based 50 percent fall can sit on top of a smaller fall in actual ship calls if the missing tonnage was carried by bigger or fuller hulls. The importer paying per box does not care how many hulls passed. The importer cares how many boxes those hulls carried, and the tonnage frame is the one that bites.

Now read the same number from the other end. Hormuz near standstill means the Gulf exit is mostly shut. Bab el-Mandeb at 31 a day means the Red Sea alternative is open but thin. Saudi Arabia's contingency was to push crude out through Yanbu on the Red Sea instead of through the Gulf. Yanbu crude now has to run the Bab el-Mandeb gauntlet. The Houthi blockade declared on 20 July covers Saudi vessels. So the route built as the relief valve is now under the very threat it was meant to dodge. That is the dual bottleneck the headline names, and the numbers confirm it is real rather than rhetorical. One box is nearly shut, the other is half open, and the half-open one is the one Saudi was counting on.

Notice the caliber of the 20 July wording. It names Saudi vessels, not all Red Sea traffic. That targeting is why Yanbu is exposed while general Red Sea commerce keeps moving at 31 a day. If the declaration had named only Gulf-flagged hulls, Yanbu would sail clean and the Red Sea lane would be the cheap escape, not the threatened one. The text of the blockade, not its existence, is what moved the number. Watch the wording. That is a rate analyst's job more than a strategist's.

Hormuz near standstill deserves its own line. Standstill is not a number Clarksons gave in tonnage for Hormuz here; the update only said near standstill alongside the 31 a day for Bab el-Mandeb. So I will not invent a Hormuz figure. I will only note the pairing: one box nearly shut, the other half open, reported together. The report's own structure tells you the bottleneck is dual by design of the data, not by my interpretation. When two chokepoints are reported in one breath, the routing math for every Gulf-origin cargo changes at once, because the two exits are no longer substitutable.

The chain from announcement to your invoice is short and mechanical. The 20 July blockade text starts it. Underwriters then reprice war-risk cover for vessels inside the Red Sea box; war-risk is set per voyage, so the cost lands on the next fixture, not the last one. Owners either add a war-risk surcharge line to the rate or route around the Cape to avoid the box. Routing around the Cape adds roughly ten to fourteen days of round-trip sailing and the fuel that rides with it, recovered through a bunker or reroute surcharge. Even ships that do transit absorb friction: the two VLCCs paused two days and changed AIS destination fields before passing. The pause is idle vessel cost. The AIS change is compliance and operations overhead that never appears on the freight line but shows up in schedule and paperwork.

One more word on the underwriter repricing. War-risk cover is priced per voyage and per declared corridor. The moment a corridor is named in a blockade text, the load jumps on the next fixture. This is why the two VLCCs paused two days; they were likely waiting on cover terms and on the AIS edit before committing to the transit. The pause is the visible cost of repricing. You will not see it on your freight line, but your carrier's cost basis moved during those two days, and that basis is what the next quote carries. Repricing is not a rumor. It is a dated, observable event.

Those two VLCCs are the cleanest evidence that the channel is open but priced. They moved about 4 million barrels through on 23 July, after the 20 July declaration, after a two-day pause, and only after masking their destination. They did not cancel. They adapted. Four million barrels is a real cargo on a real transit. The fact that they made it proves 31 a day is a working lane, not a sealed one. The fact that they paused and masked AIS proves the cost of using it rose before they ever loaded a barrel of premium onto the buyer. The premium is the story. The closure is not.

Two VLCCs at 4 million barrels also tell you the channel is not moving dinghies. A very large crude carrier is among the biggest trading ships afloat; two of them clearing the strait in one transit window is meaningful tonnage, not token traffic. The 31 a day average therefore includes real volume, not just coastal craft dodging the box. When you hear halved, remember halved off a base that still carried two of the largest ships on the water in a single window. The dramatists count the percentage. The rate analyst counts the deadweight still moving.

Put the buyer's cost stack on the table. Layer one is base ocean freight, the number from a calm market. Layer two is the war-risk insurance loading, charged inside the rate or passed through as its own line. Layer three is the reroute or Cape surcharge if the box is avoided. Layer four is schedule friction, the two-day pause, the AIS admin, the slower transit, felt as inventory carrying cost and demurrage risk rather than a printed line. Layer five is slot scarcity premium; when 50 percent less tonnage flows, the boxes that do flow get bid up, most sharply on lanes that still need Red Sea passage.

The headline compresses all five layers into one sentence about traffic. A rate analyst unpacks them. The first three are negotiable and contractible today. The fourth and fifth are where money leaks if you only watch the freight line. Most buyers fixate on layer one and meet layers two through five for the first time when the quote arrives already loaded. By then the loading is published and the negotiating window is closed. The fix is to contract the ceiling on layers two and three before the carrier prints them.

Translate layers four and five into working capital so the CFO reads the same sheet. A two-day pause on a transit is two extra days of inventory in motion, which is two days of capital not yet sold. On a lane running a shipment every week, that is a permanent two-day bulge in transit inventory, funded at your cost of capital, not the carrier's. Slot scarcity premium layer five does the same from the price side. Neither shows as freight. Both show as cash. The freight line is the smallest number on the page once the channel is thin. The cash behind it is the larger one.

When does the premium reach your quote. Clarksons data is 25 and 26 July, published 27 July at 10:00. War-risk loadings reprice fast because underwriters read the same feed, often within the week. Carriers announce surcharges on a rolling basis; a reroute surcharge can appear on quotes seven to ten days after a blockade declaration. The two VLCCs transited on 23 July, already inside the post-20 July risk box, which is why they paused and masked AIS. The gap between announcement and fully repriced quote is measured in days, not months. Book spot in the first week of August and assume the surcharge is in. Book in the last days of July and you are racing the repricing clock.

Why wait and see fails here specifically. In a calm market you can watch the rate drift and catch it on the way down; that is the other article's logic, not this one. Here the drift is up and the repricing is dated to a 20 July event with a 25 to 26 July data print. The clock started before you read the headline. Every day you wait, a carrier prints one more surcharge you could have capped. The asymmetry is the whole point. Upside of waiting is a cheaper rate that will not come. Downside of waiting is a surcharge you could have capped. Lock.

Here is the part the front pages skipped. They wrote 'blockade' and 'halved' because those words sell. The buyer's real problem is not that ships stopped. Thirty-one a day still moved. The real problem is the premium on those 31 a day. So the move is not 'find a slot before they vanish', because slots are not vanishing; 31 a day is still a working channel. The move is to lock the premium, not the slot. Concretely, write a cap into the contract: war-risk surcharge capped at X, reroute surcharge capped at Y, AIS compliance friction absorbed by the carrier up to Z. You are not buying space. You are buying a ceiling on the extra cost the next quarter will throw at you. That is the arithmetic the headlines never print because it is dull and it is exact.

A cap is not a discount. It is a number above which you stop paying. Say war-risk prints at 8 percent of base and you capped it at 4 percent. Every dollar above 4 percent is the carrier's problem, not yours, for the life of the contract. Same for reroute at 2 percent cap. The AIS friction you push to the carrier as a disclosed line, not buried in transit time, so you can see it and challenge it. None of this requires the strait to reopen. It requires the strait to stay open at 31 a day, which it is. The cap works precisely because the channel is not closed.

A cap also protects the contract's other clauses. Once war-risk prints at 8 percent uncapped, the carrier has leverage to hold the base and the reroute too, because you are suddenly desperate for any slot on a 31 a day lane. Capping early removes that leverage before it forms. You are not just saving the model shipment's amount. You are denying the carrier the anchor to push layers one and three after layer two prints. The cap is a position, not a discount. Treat it like one.

Another reason the cap beats the slot. A slot is a position on a sailing. If the sailing reroutes around the Cape, your slot still exists but your transit does not, and the surcharge lands anyway. A cap travels with the cargo, not the sailing. You can change carriers, change sailings, reroute yourself, and the cap on war-risk and reroute still binds because it is written against the corridor, not the ship. That portability is why the ceiling, not the berth, is the thing to lock when the channel is open but priced.

Test the threshold with a counterfactual. If Bab el-Mandeb had gone to zero like Hormuz, the calculus flips. No channel, only reroute, and reroute cost would dominate, making a cap nearly meaningless because the base would explode. At 31 a day it has not flipped. The channel exists, so the premium is the variable you control. The line that matters is not some geopolitical red line. It is the daily transit count. Below about 20 a day for three straight days, the cap strategy stops working and you move to full reroute with a Cape fuel budget. Above 20, you cap.

Second counterfactual. If the 20 July text had named only Gulf-flagged vessels, Yanbu would be the safe lane and the Red Sea the cheap relief valve. The Saudi crude would flow clean, the premium would sit on Gulf-origin boxes only, and your cap would be narrow. The actual text named Saudi vessels, so Yanbu is exposed and the premium sits on the very cargo Saudi was rerouting to dodge the Gulf risk. The caliber of the words moved the number. That is why I read the announcement before the headline. The number is downstream of the wording.

Now a number with the assumption said out loud. Assume a buyer moves one shipment of 100 containers on a Red Sea-touching lane. Base ocean freight is 2,000 dollars per container, with war-risk bundled at zero in a calm market. Assume the repricing adds a war-risk surcharge of 8 percent and a reroute-or-absorb loading of 5 percent on boxes that cannot avoid the box. Total risk premium is 13 percent on the 2,000 base, equal to 260 dollars per container. On 100 containers that is 26,000 dollars of new cost on one shipment. Assume the buyer locks a cap today holding war-risk at 4 percent and reroute at 2 percent, total 6 percent, or 120 dollars per container, saving 140 dollars per container and 14,000 dollars on the shipment. That 14,000 is the value of locking the ceiling before the surcharge publishes.

The 2,000, the 8, the 5, the 4, the 2, and the 100 are assumptions for illustration. The only published figures I anchor to are 31 a day, 50 percent below Q2 tonnage, 20 July blockade, 23 July two VLCCs, 4 million barrels, two-day pause, AIS change. Everything on top is a model, not a quote. If your real base and real surcharge differ, plug them in; the structure holds. The point is not the exact saving. The point is that the saving exists only if you cap before the print, and the print is days away.

One more way to read the 50 percent below Q2 tonnage. The Q2 mean was on the order of 62 passages a day, so the lane normally ran full and the 31 a day is a discount event, not a new normal. That matters for the negotiation because the carrier cannot claim the thin flow is structural; it is a risk event with a start date of 20 July. You anchor your cap to the pre-event baseline, not to the discounted present. A cap sized off 62 a day assumes the lane refills when the blockade text changes; a cap sized off 31 a day pays the carrier for a thin lane it did not build. Read the baseline, not the dip.

The AIS destination change is worth a line of its own because it is a recurring cost, not a one-off. Changing the declared destination before transit is what the two VLCCs did, and it is not free. It triggers sanctions screening review, charterer liability checks, and a paper trail on every passage. On a lane moving 31 ships a day that is 31 sets of extra admin a day, every day the risk box stays open. None of it prints on your freight line, but all of it sits in the carrier's cost basis and eventually in the quote. The cap should name it, because unnamed friction is friction the carrier can bill later as a surprise line.

Close the loop on Hormuz. Near standstill at the primary Gulf exit means every contingency reroute lands on the secondary exit, which is Bab el-Mandeb. That is why the half-open strait is stressed rather than relaxed: it is absorbing the spillover from the shut one. The dual bottleneck is not two independent problems. It is one problem with two mouths, and the open mouth is now carrying the closed mouth's load. Price the open mouth accordingly, and cap it, because the moment the closed mouth reopens the open mouth's premium does not automatically fall; the contracts written during the squeeze stay written.

The strait is still open. It is just politely asking for more money on the way through. Lock the ceiling this week, before the first reroute surcharge prints. If you wait for the gate to close before you negotiate, you will be haggling in a parking lot with everyone else's truck already parked. The line to watch is 20 transits a day. Above it, cap. Below it for three days, reroute. Do not confuse the two moves, and do not wait for the drama to tell you which one you need.

↑ Back to the key points

— By Vivian Zhao