Delivery squeeze

A delivery squeeze is controlling the certified stocks, warrants or logistics needed to satisfy a futures contract, so that short sellers cannot deliver even where the commodity exists elsewhere.

Also called warehouse squeeze, delivery point squeeze. Observed in futures, commodities. One of the corners and squeezes techniques. No enforcement actions yet in the library.
Updated 2026-09-07

How does a delivery squeeze work?

A delivery squeeze attacks the narrowest point in a futures contract: the physical act of delivery.

A futures contract is not a promise to hand over a commodity. It is a promise to hand over a specified grade, at a specified location, within a specified window, in specified units, certified by an approved inspector, into an approved warehouse. Each of those qualifiers shrinks the universe of what counts. By the time all of them are applied, the quantity capable of satisfying the contract is typically a rounding error against world production.

That is the pressure point, and there are three ways to squeeze it.

Take the stocks. Buy the certified material in exchange-approved warehouses. It is a finite, published quantity, and it is usually small.

Take the warrants. In metals markets, deliverable material is represented by warrants — the transferable title documents. Cancelling warrants removes metal from the deliverable pool without moving a single tonne. The metal is still in the shed; it is simply no longer available to satisfy a contract.

Take the logistics. Exchange-approved warehouses have minimum load-out rates but no cap on how long a queue may become. If material can only leave at the minimum rate and the queue is months long, the metal is nominally deliverable and practically inaccessible. Nothing has been removed; the exit has just been narrowed.

The third route is the most elegant and the hardest to characterise, because at no point does anyone refuse to sell anything.

A worked example with real numbers

A metal contract, 25 tonnes per lot, delivery at approved warehouses in one region.

World production is 21 million tonnes and completely irrelevant. What matters is 46,000 tonnes against 3,900 lots of obligation — 97,500 tonnes required against 46,000 tonnes available. More than half the shorts structurally cannot deliver before anyone does anything.

The squeeze. A participant acquires warrants over 34,000 of the 46,000 tonnes and declines to sell or lend them.

Available to the remaining shorts   46,000 − 34,000 = 12,000 tonnes = 480 lots
Shorts needing to deliver                             3,900 lots
Shortfall                                             3,420 lots

The effect on price. The spot month, previously trading at a $14 discount to the three-month contract, moves to a $260 premium — a swing of $274 a tonne in a market where nothing about supply or demand changed.

For a short holding 200 lots (5,000 tonnes) who must buy back:

5,000 tonnes × $274 = $1,370,000 of additional cost

For the holder of the warrants across 3,420 lots of shortfall, the corresponding gain is very large.

The countervailing risk, which is real. The squeezer now owns 34,000 tonnes of metal at a price they created. Once the delivery month passes there is nobody forced to buy it, and the spot premium collapses. They must either hold it — paying storage, insurance and financing indefinitely — or sell it into a market that has just watched what happened.

Meanwhile the exchange is not a spectator. Facing an open-interest-to-stock ratio above two, a venue will typically raise margins, impose delivery-month position limits, order liquidation-only trading, or change warehouse load-out rules. Any of those turns a 3,420-lot advantage into a forced liquidation.

Why is a delivery squeeze illegal?

Section 9(a)(2) of the Commodity Exchange Act, codified at 7 U.S.C. § 13(a)(2), makes it a felony to manipulate or attempt to manipulate the price of a commodity, or to corner or attempt to corner one. Section 6(c)(1) and CFTC Rule 180.1 supply the modern fraud-based route.

The traditional price manipulation claim has four elements, and the third is where cases go to die:

1. the accused had the ability to influence prices
2. they intended to create an artificial price
3. an artificial price existed
4. they caused it

Element three obliges the regulator to establish what the price would otherwise have been. In a commodity market with genuine supply constraints, weather, transport disruption and shifting demand, that is a contest between econometricians rather than a question of fact — and it has defeated cases that looked strong on the other three elements.

This is why the CFTC has increasingly preferred Rule 180.1. Modelled on Rule 10b-5, it asks about deception rather than artificiality, which is a more tractable question. Where a participant represented to a venue that warehouse material was available while arranging that it was not, the deception is identifiable in a way that price artificiality is not.

Position limits under 7 U.S.C. § 6a are the structural defence, and the aggregation rules — treating commonly controlled accounts as one holder — do more practical work than the limits themselves, because historic squeezes were built through accounts nominally belonging to others.

Warehouse regulation is the other lever. After sustained queueing controversies in metals markets, exchanges imposed load-out rules tying the rate at which material must leave a warehouse to the rate at which it arrives. That is a structural remedy aimed precisely at the logistics variant, and it did more than any enforcement action to address it.

How does a delivery squeeze get detected?

This family is unusual in that regulators can watch it form.

The stock ratio. Delivery-month open interest against certified deliverable stocks, published by the exchange. A ratio well above one means most shorts structurally cannot deliver, and it is public before any price move.

Warrant concentration. Who holds title to the deliverable material. Exchanges track this directly, and concentration approaching the whole certified stock is visible immediately.

Large trader reporting. Participants above thresholds report positions daily, so the futures side is transparent to the regulator in a way that most manipulation is not.

Spread and basis analysis. Spot-month dislocation from deferred months, and cash prices at the delivery location diverging from the same grade elsewhere, are the price signatures. Neither proves manipulation, because genuine shortage produces both.

Load-out and queue monitoring. Where material is leaving at the minimum permitted rate while queues lengthen, and where warehouse ownership is connected to a holder of the material, the logistics variant becomes visible.

Account aggregation. As always in this family, establishing that positions in several names are one position is the difficult and decisive step.

• Certified warehouse stocks concentrating in a small number of holders as a delivery month approaches.
• Registered warrants cancelled or withdrawn from the deliverable pool without a corresponding physical movement.
• Spot-month prices dislocating from deferred months in a way no supply fundamental explains.
• Cash prices at the delivery location diverging sharply from prices at other locations for the same grade.
• Load-out rates from a warehouse falling far below its stated capacity while queues lengthen.

What are the red flags?

• Delivery-month open interest that substantially exceeds certified deliverable stocks.
• Sharp backwardation in the spot month with normal deferred spreads.
• Exchange emergency action — margin increases, liquidation-only orders, or extended delivery periods.
• Warehouse queues measured in months where the commodity itself is not scarce.

For a commercial hedger, the practical warning sign is published and available in advance: open interest in the delivery month against certified stocks. When that ratio goes above one, rolling early costs a spread and staying costs whatever the squeeze demands.

What a delivery squeeze is not

It is not a shortage. Harvest failures, mine strikes, transport disruption and shipping delays all produce identical price behaviour, and they are the more common explanation.

It is not holding inventory. Commercial firms hold physical commodities as their ordinary business, and being long the physical into a tight market is a position, not a plan.

It is not backwardation. Spot prices above deferred prices is a normal state in markets with genuine near-term tightness.

It is not a corner. A corner pairs supply control with a long futures position used to demand delivery. A delivery squeeze can operate on the supply side alone.

Frequently asked questions about delivery squeeze

What techniques are related to delivery squeeze?

• Cornering — Cornering is acquiring control of the deliverable supply of an asset while holding a large long derivative position in it, so that short sellers cannot obtain the asset and must settle on the cornerer's terms.
• Engineered short squeeze — An engineered short squeeze is deliberately acquiring the tradable supply of a security and withdrawing it from the lending market, so that short sellers who must cover cannot find stock and are forced to buy at the engineer's price.
• Settlement price manipulation — Settlement price manipulation is trading to move an official settlement price, because that price determines margin, portfolio marks and the payoff on contracts many times larger than the trading itself.
• Box squeeze — A box squeeze is holding both a long position and the lendable supply of a security, so that recalling the borrow forces short sellers to buy back at a moment and a price the holder chooses.

Terms defined on this page

Deliverable Supply · Delivery Month · Open Interest · Contango And Backwardation · Corner Squeeze Distinction · Position Limit

Sources

1. Commodity Exchange Act § 9 — manipulation — Cornell Legal Information Institute
2. Commodity Exchange Act § 4a — position limits — Cornell Legal Information Institute
3. CFTC market surveillance programme — Commodity Futures Trading Commission

Reviewed September 7, 2026. Every statute link points at the primary text. If something here is wrong, tell us — corrections are logged in public.