Switzerland Didn't Rescue Credit Suisse. It Wrote a Law Over a Weekend and Hoped It Held.

Over a single weekend in March 2023, a small group in Bern did something no market could: they ended a 167-year-old bank, exchanged it for CHF 3 billion of UBS stock, erased CHF 16.5 billion of bonds, and added a new clause to Swiss law to make all of it possible.¹³ The clause - Article 5a - was written days before it was used.³ When the lawyers needed a power that did not exist, the state simply created one and reached for it before the ink had dried. This was not a rescue negotiated between two banks. It was a deal authored by a government and handed to UBS to sign.

The official story is that the market collapsed and UBS stepped in to save a rival. The truer story is that the discussions were initiated jointly by the Federal Department of Finance, FINMA and the Swiss National Bank¹ - and that the two most celebrated features of the deal, the bond wipeout and the giant government backstop, have since been annulled in court and voluntarily thrown away. What looked like the boldest piece of crisis management of the decade looks, two years on, like improvisation that the courts and UBS itself have been quietly unwinding ever since.

The trigger nobody pulled on purpose

It is tempting to start with a villain, and the usual one is the Saudi National Bank. On 15 March 2023, its chairman Ammar Al Khudairy was asked on Bloomberg TV whether SNB would invest more in Credit Suisse, and answered: 'The answer is absolutely not.' Credit Suisse shares fell more than 30% in a day.⁶ The legend is that a Gulf backer slammed the door. The transcript says something duller and more important. He was answering a hypothetical about raising SNB's stake above its 9.9% holding, and he gave a regulatory reason: crossing 10% triggers a supervisory burden the bank did not want.⁶ SNB later clarified that Credit Suisse had never asked them for fresh capital at all.⁶ The remark was technically accurate and catastrophically timed - a true sentence that the market read as a death sentence. The point isn't that one man broke a bank. It's that the bank was already so fragile a precise answer to a hypothetical could finish it.

Which is the setup for the real subject: not why Credit Suisse fell, but how Switzerland caught it. The answer is the thesis of this whole episode.

A merger written by the state, not the market

Strip away the speed and the drama and you find a deal that ordinary corporate law could not have produced. A normal acquisition needs both sets of shareholders to vote. This one bypassed them. It needed a legal basis to extinguish Credit Suisse's AT1 bonds while leaving equity holders with something - inverting the usual order of who loses first - and that basis did not exist, so the Federal Council added Article 5a to its emergency ordinance on 19 March, granting FINMA the authority to order the write-off, and FINMA issued the decree the same day.³ The whole architecture - the all-share price of CHF 3 billion, the CHF 16.5 billion bond erasure, the CHF 9 billion loss guarantee, the CHF 100 billion liquidity backstop - was assembled and detonated inside roughly 48 hours.¹²³ That is not a transaction. It is a sovereign act wearing a transaction's clothes.

The court that called it unconstitutional - two and a half years later

The improvisation held just long enough to stop the panic, and then it started coming apart in slow motion. The AT1 wipeout was the deal's most defended feature - regulators insisted the bonds' own prospectuses allowed it the moment a 'viability event' occurred. About 3,000 complainants disagreed and filed roughly 360 appeals.³ On 1 October 2025, the Swiss Federal Administrative Court agreed with the complainants and revoked FINMA's decree.⁴ The court's findings are devastating precisely because they are unglamorous: the contractual viability trigger had not legally been met, because Credit Suisse was still adequately capitalised when its bonds were erased; Article 5a was unconstitutional and inapplicable, lacking any clear legal basis to permanently extinguish private creditor claims; and the bondholders' property rights under Article 26 of the Swiss Constitution were unlawfully infringed.⁴ Translated: the law invented to justify the write-down was too vague to be law, and the event it was meant to cover hadn't actually happened.

The backstop UBS handed back before it could be used

Now the second showpiece. The headlines warned that Swiss taxpayers had been put on the hook for as much as CHF 109 billion - the CHF 9 billion loss guarantee plus the CHF 100 billion liquidity backstop.² It is the part of the story most people still believe, and it is the part that aged worst. On 11 August 2023 - less than five months after signing - UBS voluntarily terminated both facilities without drawing a franc from either.⁵ The Confederation walked away with roughly CHF 200 million in guarantee premiums and not a single loss; UBS even paid CHF 40 million to compensate the state for setting up the arrangement in the first place.⁵ The enormous public exposure that justified the emergency turned out to be a fee-generating formality. The state was paid to insure a risk that never materialised - which raises an awkward question about how much the emergency machinery was ever truly needed, versus how much it was simply the price of moving in a weekend.

Wasn't this still a brilliant rescue?

The honest objection is strong, and it deserves a straight answer: the deal worked. No bank run spread, no taxpayer money was lost, a systemic failure was contained over a weekend, and UBS has already booked $9.1 billion in annualised cost savings - 70% of its CHF 13 billion 2026 target.⁵⁸ By the only test that mattered in March 2023 - is the system still standing on Monday? - it passed. All true. But 'it worked' and 'it was sound' are different claims. The annulment matters not because the panic should have been left to burn, but because the state borrowed against the rule of law to buy speed, and the bill came due in 2025. The trigger that wasn't met, the clause that was too vague to be law, the backstop priced and returned like a parking ticket - these aren't footnotes. They're evidence that the rescue's legal scaffolding was always provisional, held up by the assumption that no court would ever look closely. A court did. And the integration itself is now lagging: UBS sits near 105,000 staff against an internally planned 85,000, shedding only ~1,300 roles a quarter against more than 3,500 in late 2023.⁸ The emergency was easy. The marriage is the hard part, and it is running slow.

Run the counterfactual and the deal's real nature comes into focus. Without Article 5a, the bonds could not have been written down. Without the joint hand of the Finance Ministry, FINMA and the SNB, no shareholder vote would have been bypassed. Without a state willing to invent a power overnight, there is no weekend deal at all - only a slower, messier, more lawful unwinding that Switzerland was unwilling to risk.³ The genius of March 2023 was never financial engineering. It was administrative nerve: the willingness to act first and legalise later. That nerve stopped a panic. It also wrote a CHF 16.5 billion cheque the Constitution refused to honour, and handed UBS an integration the emergency could not finish. The state bought the weekend. UBS is still paying for the years.