Forty Years of Threats Never Closed the Strait of Hormuz. 2026 Did.

For four decades the most durable shortcut in energy geopolitics was a base rate: threats to close the Strait of Hormuz recur, and the strait stays open. It survived the Tanker War of the 1980s, two documented mine-laying campaigns, and the largest convoy escort operation since the Second World War. Analysts used it to discount headline risk, and for forty years it paid.

It stopped paying in 2026. The U.S. Energy Information Administration estimates that crude oil and petroleum liquids moving through the Strait of Hormuz averaged 4.9 million barrels per day in the second quarter of 2026, against 21.6 million b/d in the fourth quarter of 2025 — a decline of roughly 77%. In June 2026 the agency described the strait as remaining effectively closed in the near term.

The useful question is not whether the base rate was wrong, but what it was measuring. A regularity that holds for forty years and then fails is rarely a law. It is a set of conditions stable enough to be mistaken for one — and conditions are observable, so they can be tracked before they break.

The Surface Issue: A Chokepoint Read as a Switch

Coverage of Hormuz has long treated the waterway as binary: open or closed, priced or not priced. That framing suits a headline and fails as risk management, because the strait has never operated as a switch.

The geometry explains part of this. At its narrowest point the strait is about 21 miles (33 km) wide, but the usable channel is far narrower: EIA describes the shipping lane in either direction as only two miles wide, separated by a two-mile buffer zone. That cuts both ways. A few hazards in a narrow lane change the calculation for every master and every underwriter, yet traffic willing to accept the risk can still move.

The 2026 record shows that gradient rather than a switch. Transits in the period after the June 17 memorandum of understanding were reported at 513 ships between June 18 and July 5, roughly 28 vessels per day against a pre-conflict average near 100. Volumes at 23% of normal are not a legal closure and are a commercial catastrophe. Both are true, which is why the binary frame fails.

The Structural Cause: Four Conditions the Base Rate Was Tracking

The forty-year record of non-closure rested on four conditions. None was a permanent feature of the region. Each is measurable, and each behaved differently in 2026.

Condition one: the interdicting party's own dependence on the same water

The oldest argument against closure was self-harm: a state that seals Hormuz seals its own exports, having no meaningful alternative outlet. That remains quantifiable. The Goreh-Jask pipeline, inaugurated in July 2021 as a bypass to the Gulf of Oman, carries an effective capacity of roughly 0.3 million b/d in EIA's chokepoint analysis, and EIA has recorded it as largely unused. Against export volumes measured in millions of barrels per day, that does not change the arithmetic.

What changed is not the arithmetic but its weight. Self-harm deters when export revenue is the dominant objective. Where those revenues are already impaired or forgone, the marginal cost of closing falls sharply. The deterrent was never pipeline capacity but an assumption about what was being maximised.

Condition two: convoy escort worked, historically, against a specific threat type

The Tanker War is the reference case, and its numbers deserve stating precisely. Naval History and Heritage Command records that by the end of 1987 Iraq had conducted 283 attacks on shipping and Iran 168, with 116 merchant sailors killed, 37 missing and 167 wounded. Operation Earnest Will ran from 24 July 1987 to 26 September 1988, the largest naval convoy operation since the Second World War.

The outcome is the part usually compressed into a clause. Research collated by the Strauss Center notes that 239 petroleum tankers were attacked, of which 55, or 23%, were sunk or declared constructive total losses; that commercial shipping initially dropped about 25%; and that at its peak the campaign still failed to disrupt more than about 2% of ships transiting the Gulf. Hundreds of attacks over four years suppressed a fraction of traffic. That is the empirical foundation of the base rate.

It is also the foundation's limit. Escort works against attacks that are above the waterline, attributable and repeatable, where a warship alongside changes the attacker's expected cost. It works far less well against a hazard that sits in the water and does not care which hull arrives first.

Condition three: the asymmetry between laying mines and clearing them

Mines were always the exception inside the historical record. The tanker Bridgeton struck a moored contact mine on 24 July 1987, the first day of Earnest Will. On 21 September 1987 the vessel Iran Ajr was intercepted laying mines, carrying 18 Sadaf-02 moored contact mines. On 14 April 1988 the frigate USS Samuel B. Roberts struck a mine and was nearly lost; Operation Praying Mantis followed on 18 April 1988.

The economics of that exchange have not improved for the clearing side. Mines are cheap and quick to deploy, and effective in a two-mile lane. Clearance is slow, specialised, and must be verified before traffic and its insurers accept the route. The June 17 memorandum reportedly committed to demining within 30 days; Council on Foreign Relations analysis published after that window put an estimated 80 mines still in the main navigation areas, timeline unsettled. The gap between an agreement to clear and a channel certified clear is the most underweighted variable in the file.

Condition four: buyer-side reroute capacity

Two pipeline systems bypass the strait at scale, which is the condition markets could always point to. Saudi Arabia's East-West line, the 745-mile Petroline from Abqaiq to Yanbu on the Red Sea, has a nameplate capacity of about 5.0 million b/d, temporarily expanded toward 7.0 million b/d in 2019. The UAE's line to Fujairah is listed at 1.8 million b/d. EIA puts the two together at roughly 4.7 million b/d of usable bypass, and in mid-2025 estimated only about 2.6 million b/d as spare capacity actually available during a disruption.

Set that against 21.6 million b/d of normal throughput and the ceiling is obvious: bypass infrastructure protects a share of exports, not the flow. The 2026 data confirms both use and limit. EIA reports Saudi crude rerouted through the East-West pipeline to Yanbu, and crude and liquids through the Bab el-Mandeb rising to 8.1 million b/d in 2Q26 from 5.4 million b/d in 4Q25 — about 2.7 million b/d more, inside the range EIA had flagged as available. The bypass performed as specified. The specification was never large enough.

Condition	Measurable proxy	2026 reading
Interdictor's export dependence	Non-Hormuz outlet capacity	~0.3 million b/d, largely unused
Escort effectiveness	Share of traffic disrupted historically	~2% at Tanker War peak; not the 2026 constraint
Mine clearance lag	Verified mines remaining in lanes	~80 estimated, timeline unsettled
Bypass capacity	Available spare pipeline capacity	~2.6 million b/d against 21.6 million b/d of flow

The Asymmetry Nobody Built Infrastructure For: LNG

About one-fifth of global LNG trade transited the strait in 2024. EIA puts Qatari volumes at roughly 9.3 billion cubic feet per day and UAE volumes at about 0.7 Bcf/d, with 83% of that gas moving to Asian markets. There is no East-West equivalent for liquefied gas: a pipeline cannot substitute for a liquefaction train and a loading berth, and the terminals that could receive rerouted molecules sit on the wrong side of the geography.

LNG exposure is therefore closer to binary than oil exposure — the opposite of the intuition most risk frameworks carry. EIA notes LNG vessel traffic slowed considerably after strikes on vessels resumed on 7 July 2026. Damage to the Ras Laffan complex, which CFR describes as roughly 5% of world natural gas and 20% of global LNG supply, carries reported repair timelines of up to five years. Oil faces a constrained detour; gas faces a rebuild.

What Markets Have Been Slower to Price: Reopening Is Not Closing Run Backwards

The price path through 2026 has been read as a reassurance signal, and that reading deserves scrutiny. Brent averaged about $103 per barrel in March 2026, peaked near $118 on 29 April, and fell to roughly $72 on 26 June. Daily swings in April and May averaged about $4 per barrel, against roughly $1 in the same months of 2025.

That move happened while a very large volume of regional production was still shut in. EIA estimates put Persian Gulf shut-ins near 10.05 million b/d across March to May, 7.48 million b/d in June and 5.46 million b/d in July. Supply did not do the balancing. Global consumption fell roughly 1 million b/d year over year, with 2026 demand projected down about 1.1 million b/d. Inventories absorbed the rest: global stocks fell an average of 4.2 million b/d in 2Q26, with a further 3.8 million b/d draw expected in 3Q26 — OECD inventories at their lowest reported level since 2003.

A price that falls because demand contracted and stocks drained is not the same signal as a price that falls because supply returned.

Inventory is a buffer that spends itself. Each month of drawdown removes the cushion that suppressed the last spike, making the distribution of outcomes asymmetric even where the central forecast is benign. EIA's outlook carries the tension openly: Brent near $85 in 3Q26, $78 in 4Q26 and $69 in 2027, but with residual disruption of about 0.6 million b/d assumed through end-2027.

Reopening carries frictions that closing did not. Mines must be cleared and the clearance verified. War-risk cover must be quoted at a level shipowners accept. Crews must be willing to sail; roughly 11,000 seafarers were reported stranded in the region. Damaged facilities must be repaired, with one estimate of physical damage to Gulf oil infrastructure at about $58 billion. The legal frame is contested: the transit passage regime holds that passage "shall not be impeded" and that there "shall be no suspension of transit passage," but the relevant coastal state signed the convention in 1982 without ratifying it. The memorandum that restored partial traffic expired on 17 August 2026.

Where This Frame Breaks Down

Several objections to the argument above deserve full weight.

The base rate was not wrong; it was conditional. One closure in roughly forty-five years is still a low unconditional frequency, and anyone using it to discount headline threats between 1984 and 2025 was right almost every time. The failure was the omission of the conditions attached to it, not the estimate. The corrective is to track four observable variables, not to assume the next threat is real.

Partial disruption was always the real cost channel. Insurance, routing, freight and delay generate losses long before any closure — the Tanker War produced hundreds of attacks and a 25% drop in shipping without one. A framework pricing only the closure tail underpriced the far more common outcome, an error that predates 2026 by decades.

"Closed" remains a spectrum. At 4.9 million b/d, the strait in 2Q26 still moved roughly what the Suez Canal and SUMED system have carried since the Red Sea diversions began. Language that treats this as zero will misread the recovery as badly as the old language misread the risk.

The benign scenario deserves equal weight. A path in which demining completes, war-risk premiums normalise and flows recover toward pre-conflict levels through 2027 is fully consistent with the data — it is the central case in EIA's published outlook, with Brent easing to roughly $69 in 2027. Shut-ins fell from about 10 million b/d in spring to 5.46 million b/d in July.

One observation is not a regime. The temptation after a base rate breaks is to invert it and assume chokepoints are fragile everywhere. That does not follow. One transfer does deserve attention: the alternate route is itself a chokepoint. EIA notes prices rose in July partly on a new blockade threat aimed at Red Sea exports.

What to Watch Next Week

• Demining verification, not announcements. The operative number is mines confirmed cleared from the main lanes against the roughly 80 estimated to remain. Underwriters price the verified figure.
• Daily transit counts against two anchors: roughly 28 vessels per day from 18 June to 5 July, and the pre-conflict norm near 100. Movement toward 50 is a different signal than movement toward 20.
• Bab el-Mandeb throughput. Holding above 8.1 million b/d says rerouting is still load-bearing; a decline without a matching Hormuz recovery says volume is being lost outright.
• LNG loadings on a cargo-count basis. Gas has no bypass, so loadings are the cleanest single indicator of whether passage is genuinely usable or merely legally open.
• Inventory direction rather than price level. With OECD stocks at reported multi-decade lows, a continued draw matters more for the risk distribution than any single week of Brent.

Concrete Framework

Each item pairs a condition with a threshold that turns commentary into an observation.

1. Track the interdictor's outside option quarterly. Non-chokepoint export capacity as a share of that state's total exports. Below roughly 15% the self-harm deterrent looks intact — but the live question is whether export revenue is still being maximised.
2. Separate above-water from sub-surface threats in every incident log. Escort answers the first and not the second. A shift in incident mix toward mines is the highest-information change available, and it precedes the volume data.
3. Convert bypass capacity to a coverage ratio. Spare pipeline capacity divided by normal throughput. Hormuz entered 2026 at roughly 2.6 against 21.6, about 12%. Any chokepoint under 20% coverage is capacity-constrained, not hedged.
4. Hold gas and liquids in separate models. Liquids have a partial detour; LNG has none. Hormuz-transiting LNG exposure is closer to binary, and a rebuild timeline in years belongs in the base case, not the tail.
5. Measure recovery in verified clearance and insurance quotes, not agreements. The gap between a 30-day demining commitment and 80 mines still estimated in the lanes is the operative variable for a restart.
6. Read price against the inventory path. A falling price alongside a 4.2 million b/d quarterly stock draw is a buffer being consumed, not a supply problem solved. Log the draw rate beside the price monthly.
7. Re-underwrite the alternate route. When rerouting moves 2.7 million b/d into a second chokepoint, that chokepoint's threat profile becomes part of the original exposure.

The checklist is not a forecast of the next closure. Its purpose is to make the conditions visible while they are still conditions. A base rate that quietly stops describing the world is more dangerous than none, because it keeps feeling like evidence.

This article is macroeconomic and geopolitical analysis, not investment or financial advice. Figures are as published by the cited sources on the dates noted and are subject to revision.