- The LME’s 2022 nickel crisis showed that executed trades can still be cancelled under the rules governing an exchange.
- Trade execution does not always mean irreversibility, particularly when transactions sit within broader exchange, clearing and contractual frameworks.
- Businesses should understand who can suspend, cancel, defer or otherwise change a trade before execution, not after problems arise.
On the morning of 8 March 2022, thousands of nickel trades on the London Metal Exchange (LME) had been agreed and entered into the exchange’s systems. Within hours, the affected trades were cancelled.
At 12:05, the LME published a notice cancelling, under its own rules, every nickel trade executed from midnight – around $12 billion in value – and deferring physical delivery contractually due the following day. Members were instructed to cancel or reverse the affected transactions in their books, and trades that had been struck and recorded were, pursuant to the rules of the venue, cancelled and reversed.
For anyone whose work life involves moving goods and meeting obligations, that sequence is worth sitting with. It’s easy to treat execution as the moment a deal becomes real. But the nickel episode is a clean, well-documented reminder that execution and irreversibility are not the same thing, and that the difference lives in the rules governing the venue rather than in the commercial bargain itself.
LME’s existing cancellation provisions
Against a low-stock backdrop and following Russia’s invasion of Ukraine, the price of three-month nickel rose sharply. It closed on Monday, 7 March, at around $50,000 a tonne, already an extraordinary surge. Then in the early hours of Tuesday, 8 March, with the exchange’s automated price bands suspended, the price skyrocketed to a peak past $101,000 – more than double the previous day’s close.
Once senior management concluded that the market had become disorderly, the LME suspended nickel trading on orderly-market grounds. Trading did not resume until Wednesday, 16 March.
The part that matters here is not the suspension but the cancellation, and where the power to cancel came from. The LME did not invent a remedy in the heat of the morning. Its rulebook already contained a provision: under rule 22, the Exchange may cancel, vary, or correct any agreed trade or contract where it deems appropriate.
The LME exercised that existing power. The related decision to defer physical delivery was approved separately through the LME’s special committee. Participants who were not themselves LME members were still bound through contracts incorporating the LME’s rules. The power to cancel was therefore already part of the architecture governing the affected trades before 8 March, not a measure improvised after the price spike.
That point was tested in court. Elliott Associates and Jane Street – among those whose positions were affected by the cancellation – challenged the decision by judicial review.
In R (Elliott Associates LP and others) v London Metal Exchange and LME Clear [2023] EWHC 2969 (Admin), the Divisional Court dismissed the challenge. It read the cancellation power as one available only in exceptional circumstances and held that the LME had not acted unlawfully. The Court of Appeal dismissed a further appeal, and in January 2025, the Supreme Court refused permission to appeal.
Separately, in March 2025, the Financial Conduct Authority (FCA) fined the LME around £9.24 million for weaknesses in the systems and controls intended to keep the market orderly. The regulatory action concerned the LME’s oversight mechanisms enforced during the disorder, not that the cancellation itself was unlawful.
It helps to separate ideas that are easily merged. A trade can be agreed and executed. It can be entered in the venue’s records. It can carry delivery and payment obligations. And, under the venue’s rules, it can still be open to cancellation or adjustment.
What makes metals different
In the nickel case, the trades had been executed and recorded, and delivery on the nearest prompt date was deferred rather than performed. What the parties did not have was a position the venue could no longer alter under its rules. They had agreed on a price – they had not obtained an outcome beyond the reach of the rulebook.
It’s worth being careful about how far this travels. Exchange-traded, centrally cleared metals are not ordinary bilateral trade contracts, and the specific power the LME used does not exist in most agreements between a buyer and a seller. The nickel cancellations should therefore not be read as evidence that executed contracts are generally revocable.
The transferable point is narrower, and more useful. Almost every transaction sits inside some architecture beyond the bargain itself: exchange rules, clearing arrangements, standard-form terms, financing conditions, operational procedures, and the platform through which the deal is booked. Most of the time that architecture is invisible because it only asserts itself when things stop behaving normally.
The nickel episode made that layer visible. Participants could focus, reasonably, on price, size, and counterparty. Yet when conditions became exceptional, a pre-existing governance power outside those commercial variables became decisive.
Operationally, this is a new, sharp illustration of an old lesson. Anyone who has spent time in port and cargo operations knows the gap between an agreed position and an operationally irreversible one. A booking is not a loaded container. A signed contract is not a released cargo. Things that look settled on paper can still move because some other party, rule, or procedure retains the ability to move them.
That suggests a question worth asking before execution rather than after. Not simply whether a trade has been agreed, but what still has the power to change that agreement. Who can suspend, cancel, defer, reprice, or decline to perform, under what rules, and have those rules already been incorporated into the transaction?
—
In the nickel case, the answer lies in a rule that had been in place all along. The lesson is not that a completed trade was simply undone. It’s that execution did not put those trades beyond the reach of the architecture under which they had been made.
