Credit Suisse Didn't Die in a Weekend. It Died of a Decade, and Was Buried Under an Unconstitutional Law.

On March 15, 2023, a journalist asked the chairman of the Saudi National Bank whether he would put more money into Credit Suisse. He said, "absolutely not." The stock cratered, the wires lit up, and within four days a bank that had stood for 167 years was sold to its oldest rival for CHF 3 billion — about CHF 0.76 a share.¹ The tidy version of history now has its villain and its murder weapon: three words at the wrong moment. It is a satisfying story. It is also wrong in the way that matters most.

The official story is that a confidence shock killed a healthy bank. The real story is that the patient had been dying for a decade, the death certificate was signed under an emergency law that a court would later call unconstitutional, and the 'clean resolution' everyone praised is still being litigated in 2026. Credit Suisse did not fall in a weekend. It fell in slow motion, and the weekend was only when the cameras arrived.

The bank was already bleeding out before anyone said a word

A bank does not collapse from one quote. It collapses when the people whose money it holds stop believing it will be there next year — and at Credit Suisse, they had stopped believing well before March 2023. In 2022 alone, affluent clients pulled CHF 123.2 billion out of the bank, dragging total assets under management down roughly 20% to CHF 1.294 trillion; in wealth management, the heart of the franchise, AuM fell 27%.⁸ That is not a run triggered by a headline. That is a clientele leaving for the exits over twelve unhurried months. By the time the Saudi chairman spoke, the outflows were already disclosed in the bank's own annual report, published the day before on March 14.⁴ The match found a building already soaked in fuel.

And the fuel had a name: serial risk-management negligence dressed up as a one-off. When the family office Archegos blew up in 2021 on notional exposure north of $20 billion, Credit Suisse first told the world it had lost $4.7 billion, then quietly revised that to $5.5 billion within weeks.⁷ A first-rate prime broker does not take that hit; a bank that has stopped watching its own counterparties does. The 2022 net loss of CHF 7.3 billion — including CHF 3.7 billion in deferred-tax-asset impairments tied to yet another 'strategic review' — was not bad luck.⁴ It was the running tab of a culture that had decided, again and again, that controls were optional.

What the Swiss state actually risked, and what it actually lost

Here is where the second great myth lives: that this was a taxpayer bailout. The headline numbers invite it — a CHF 9 billion federal loss-protection guarantee for UBS, a CHF 100 billion liquidity backstop from the Swiss National Bank.³ But a guarantee is a promise to cover losses, not a loss. Both guarantees were terminated on August 11, 2023, and the Swiss Confederation assumed no losses at all.³ The state did not write a cheque; it lent its credibility, and got it back intact five months later. The real cost of Credit Suisse was paid by its shareholders, who got pennies, and by one group nobody expected to be standing in the front of the line.

The day Switzerland rewrote the law to wipe out a creditor

On March 19, 2023, FINMA ordered Credit Suisse to write off its entire stack of Additional Tier 1 bonds — roughly CHF 16.5 billion in nominal value — to zero.⁵ In the normal order of a failing bank, bondholders rank above shareholders; here, the shareholders walked away with UBS stock while the AT1 holders got nothing. To make that possible, the Federal Council amended its own Emergency Ordinance that same day, inserting Article 5a to authorize the move.⁵ Read that sequence again: the legal basis for the largest AT1 write-down in history was created on the morning it was used. The 'contractual viability event' the world was told about was, in fact, a law written to order.

In October 2025, the Swiss Federal Administrative Court annulled FINMA's decree. It found the emergency provision unconstitutional, found the legal basis insufficient, and — most damning — found that Credit Suisse had told FINMA the contractual conditions for a write-off had not actually been triggered.⁶ The bonds were destroyed not because the contract said so, but because the regulator decided they should be and changed the rules to fit. FINMA and UBS have appealed to the Federal Supreme Court, which granted suspensive effect, so the bonds sit at zero while the lawyers fight.⁶ The 'clean resolution' that central bankers held up as a model is, as of 2026, an unresolved legal dispute.

But wasn't speed the whole point?

The fair objection is that this is hindsight luxury. On the night of March 19, regulators did not have the option of a leisurely, fully-litigated process — they had hours before Asian markets opened and a globally systemic bank with no buyers left but UBS. A messy, lawful resolution that triggered cross-border contagion might have cost far more than an unconstitutional-but-contained one. By the only test that mattered that weekend — was the fire put out? — the Swiss authorities passed. The honest answer is that both things are true at once. The intervention was probably the least-bad action available in the room, and the conditions that made it the least-bad action were a decade of failures by regulators who let Credit Suisse stagger from Archegos to outflow to loss without forcing a real reckoning. The emergency was genuine. So was the long negligence that manufactured it. Praising the firefighting does not absolve the arson.

It is easy to misdate a death. The headlines say Credit Suisse died on March 19, 2023, but the company was only legally absorbed into UBS on June 12, when its shares ceased to trade and 167 years of name ended on the SIX exchange.² The truer date is harder to pin, because Credit Suisse did not die on any single day. It died of a decade — of controls treated as optional, of a clientele quietly voting with CHF 123 billion, of regulators who watched it stagger and acted only when it fell. And the law that buried it was so thin that a court would later strike it down. The lesson is not that a Saudi chairman said the wrong thing. It is that a bank can be insolvent in trust long before it is insolvent in capital — and that when the state finally moves, the question of whether it was even allowed to can outlive the bank itself.