Credit Suisse and the UBS Rescue of March 2023

Key Takeaways

• The bank passed its published tests and failed anyway. It reported a 14.1 percent common equity tier 1 ratio and a 144 percent liquidity coverage ratio for the end of 2022, and on 15 March 2023 FINMA and the Swiss National Bank stated jointly that it met the requirements imposed on systemically important banks.
• The capital ratio was partly an accounting permission. The Swiss Parliamentary Investigation Committee found that without a regulatory filter FINMA granted in 2017, the bank would have failed capital adequacy requirements significantly in 2022.
• The run began in October 2022, not March 2023. Customer deposits fell CHF 138 billion in the fourth quarter and a further CHF 67 billion in the first quarter of 2023, against a closing balance of CHF 233.2 billion.
• The rescue was enormous and was repaid. Credit Suisse drew about CHF 168 billion from Swiss National Bank facilities, and every loan under the federally guaranteed backstop was repaid by the end of May 2023.
• The loss allocation ran against the usual ranking. FINMA ordered a complete write-down of all Credit Suisse Additional Tier 1 debt, around CHF 16 billion, while shareholders received UBS shares valued at CHF 3,223 million.
• The write-down was later held unlawful and is not yet settled. The Federal Administrative Court revoked FINMA's order on 1 October 2025 in case B-2334/2023, and FINMA said a fortnight later that the judgment was not final and that it would take the case to the Federal Supreme Court.
• Swiss bank credit was repriced by the rescue, not by the crisis. On 20 March 2023, the first trading day after the announcement, the yield gap between eight-year Swiss commercial bank bonds and eight-year Confederation bonds reached 134.7 basis points, the widest of 2,513 daily observations from 2015 through 2024, against 70.8 on 8 March.

How Did Credit Suisse Lose Its Independence in One Week?

Credit Suisse AG was incorporated on 5 July 1856; the holding company above it, Credit Suisse Group AG, dates only from 1982. Both were registered at Paradeplatz 8 in Zurich, the address on the cover of the last annual report either of them filed. At the end of 2022 the group employed 50,480 people. Within a single week in March 2023 it stopped being an independent institution.

The sequence begins with a telephone call. On 9 March 2023 Credit Suisse said it would delay publication of its 2022 annual report following, in its own words, a late call on the evening of March 8, 2023, from the U.S. Securities and Exchange Commission about open comments on previously disclosed revisions to the consolidated cash flow statements for 2020 and 2019. The report appeared on 14 March, carrying management's conclusion that internal control over financial reporting had not been effective at the end of 2022 and, on reassessment, at the end of 2021 either.

On 15 March, FINMA and the Swiss National Bank issued a joint statement saying that Credit Suisse met the capital and liquidity requirements imposed on systemically important banks, and that if necessary the SNB would provide it with liquidity. The Parliamentary Investigation Committee later dated the onset of the acute crisis to that same Wednesday. On 16 March the bank said it would borrow up to CHF 50 billion from the SNB. On Sunday 19 March the Federal Department of Finance, the SNB and FINMA asked Credit Suisse and UBS to sign a merger agreement, which an emergency ordinance allowed to proceed without shareholder approval on either side.

From the SEC's phone call to the merger, day by day

Dated events with two Swiss market readings: the spot yield on two-year Confederation bonds, and the yield gap between eight-year Swiss commercial bank and Confederation bonds. Both computed by Swoopr Investment from the Swiss National Bank data portal.

The two columns disagree, and the disagreement is the story. Government yields fell hard, which is what a flight to safety looks like. The bank yield gap widened most on the Monday after the rescue rather than during the panic that preceded it. Something in the resolution itself made Swiss bank credit more expensive.

Why Did the Run Actually Begin in October 2022?

The most common error about this episode is treating it as a March event. Credit Suisse's own annual report is explicit that it is not. Under the heading of liquidity issues in the fourth quarter of 2022, it records that the bank began experiencing significantly higher withdrawals of cash deposits as well as non-renewal of maturing time deposits, and that customer deposits declined by CHF 138 billion in that quarter alone.

The distribution inside the quarter matters more than the total. Group net asset outflows came to approximately 8 percent of assets under management as of the end of the third quarter, with approximately two-thirds concentrated in October 2022. Wealth management, where the wealthiest and best advised clients sat, lost approximately 15 percent in a single quarter, against approximately 2 percent at the Swiss retail and corporate bank. The people closest to the institution left first and fastest.

It also contains an admission that is easy to read past. While the group liquidity coverage ratio and net stable funding ratio were maintained at all times, the bank states that it fell below certain legal entity-level regulatory requirements. A group can satisfy a consolidated ratio while entities inside it do not, and depositors bank with entities, not with consolidations.

The Parliamentary Investigation Committee, reporting in December 2024 after reviewing more than 30,000 pages, put it without hedging: at the beginning of October and again in late December, Credit Suisse experienced massive liquidity outflows and came close to insolvency. The authorities switched to crisis mode in October. By the time the public learned there was an emergency, the state had been managing one for five months.

Over 2022, customer deposits fell CHF 159.6 billion, or 41 percent, to CHF 233.2 billion, and total assets fell from CHF 755.8 billion to CHF 531.4 billion. A bank does not shrink by 30 percent in a year because of a headline. It shrinks because its funding is leaving continuously and it is selling assets to keep pace.

What Did the Annual Report of 14 March 2023 Disclose?

The document that reached the market five days before the merger was not reassuring. Three of its disclosures explain why the last week went the way it did.

The control failure. Management concluded that as of 31 December 2022 the group's internal control over financial reporting was not effective, reassessed 2021 and reached the same conclusion, and concluded that disclosure controls and procedures were not effective either. The weaknesses related in part to the failure to maintain an effective process to identify and analyse the risk of material misstatement. For a bank asking clients to leave money with it, an audited statement that its own reporting controls do not work is close to the worst possible disclosure.

The loss. Credit Suisse reported a net loss attributable to shareholders of CHF 7,293 million for 2022, against CHF 1,650 million in 2021. Inside that number sat a CHF 3,655 million valuation allowance against deferred tax assets, a write-off that only makes sense if the bank no longer expected enough future profit to use them.

The valuation the market had already reached. The report put market capitalisation at CHF 11,062 million against total shareholders' equity of CHF 45,129 million. The shares were changing hands at roughly a quarter of the book value the accounts claimed. That is not a market anticipating a difficult year. It is a market saying it does not believe the balance sheet, and it had been saying so for months.

On the same day the board published its April meeting agenda, proposing to re-elect every director and pay a dividend of CHF 0.05 per share. Four days later the shareholders who would have voted on it had their company merged away without being asked.

Why Did a Bank That Met Its Requirements Fail Anyway?

This is the part of the episode that generalises furthest, so it is worth stating precisely. Every published prudential test Credit Suisse faced was passed on the way to being sold.

The common equity tier 1 ratio was 14.1 percent at the end of 2022. The three-month daily average liquidity coverage ratio was 144 percent against a regulatory minimum of 100 percent, on average high quality liquid assets of CHF 120.0 billion. On 14 March 2023 the quarter-to-date reading was still approximately 153 percent. FINMA's own December 2023 review states that Credit Suisse satisfied the regulatory capital requirements, while noting that the parent company, CS AG, had the weakest capital adequacy within the group, which made it the weakest link in the chain.

The liquidity coverage ratio is where the mismatch is clearest. It is calibrated to a stress scenario of defined severity over thirty days, assuming particular runoff rates for particular kinds of funding. A bank holding enough liquid assets to survive that scenario has demonstrated exactly one thing: that it survives that scenario. No ratio built on a thirty-day window describes a six-month, continuous withdrawal by clients who have decided to bank somewhere else.

FINMA's report reaches for a phrase that captures the mechanism better than any ratio: the loss of confidence led to rapid, extensive liquidity outflows further exacerbated by digital communication channels, a digital bank run. That is the same acceleration that made the Silicon Valley Bank failure a thirty-six hour event, except that Credit Suisse's version ran in slow motion for five months first.

What did carry information was the deposit trend, the share price relative to book value, the credit rating downgrades through 2022, and the fact that the state had been in crisis mode since October. None of those is a capital ratio.

What Was the Regulatory Filter, and Why Does It Matter?

There is a second answer to why the capital ratio said nothing, and it did not become public until the Swiss Parliament published its investigation in December 2024. The committee found that FINMA granted Credit Suisse extensive easing of its capital adequacy requirements in 2017 in the form of a regulatory filter, letting the bank switch from a portfolio to an individual valuation method for its participations without building up substantial capital straight away.

The committee's assessment of that decision is worth quoting rather than paraphrasing, because the wording is unusually direct for a parliamentary document. It states that the filter was legal, that the committee questions its usefulness, that its impact turned out to be far greater than expected, and that without the filter Credit Suisse would have failed to meet capital adequacy requirements, just marginally in 2021 but significantly in 2022.

Follow that through. The 14.1 percent ratio published on 14 March 2023, and cited implicitly by the regulator the next day, rested on a permission granted six years earlier whose effect nobody had sized correctly. The supervisor's public assurance and its private accommodation were the same number viewed from two ends.

This matters beyond Switzerland. A capital ratio is not a measurement in the way a temperature is. It is the output of a rulebook containing discretionary elements, transitional arrangements and firm-specific permissions rarely legible from outside. The committee's response was blunt: it sees an urgent need for action in the granting of alleviations to systemically important banks, because the alternative is a system in which the published number is not the number.

How Much Did the Swiss National Bank Actually Lend?

The support came in layers, and separating them is the only way to see who was exposed to what.

The first was ordinary collateralised central bank lending, available to any solvent bank against acceptable security: the facility announced on 16 March 2023, up to CHF 50 billion, fully collateralised by high quality assets.

The second was emergency liquidity assistance under the Federal Council's emergency ordinance of 16 March 2023, amended on 19 March: up to CHF 100 billion of additional SNB liquidity for UBS AG and Credit Suisse AG combined, the loans carrying preferential rights in bankruptcy. That ranking is the tell. It exists because the collateral was no longer sufficient on ordinary terms, so the central bank was given a statutory jump up the queue instead.

The third was the public liquidity backstop: up to a further CHF 100 billion backed by a federal default guarantee. The Federal Department of Finance records the Confederation's side as a CHF 100 billion guarantee in favour of the SNB, alongside a separate CHF 9 billion loss protection guarantee for UBS.

The amounts used were large and were repaid quickly. Credit Suisse reported net borrowings of CHF 108 billion at 31 March 2023, after CHF 60 billion of repayments during the quarter. Adding the two puts gross drawings at about CHF 168 billion, roughly seven tenths of the CHF 233.2 billion of customer deposits the bank still held at the end of 2022, and close to the whole of what was left once a further CHF 67 billion had gone in the first quarter. UBS records that all backstop loans were repaid by the end of May 2023, and terminated both the loss protection agreement and the backstop on 11 August.

The Confederation's ledger closed in the black, with roughly CHF 200 million of fee receipts and no losses assumed. That is a good outcome and it is not the same as the guarantees having been costless, because a guarantee that is never called still transferred real risk while it was live.

Why Were AT1 Bondholders Written Off While Shareholders Were Paid?

Additional Tier 1 instruments are subordinated bank capital designed to absorb losses while the bank is still a going concern. They sit above equity and below senior debt, and pay a higher coupon precisely because of where they sit. On 19 March 2023 FINMA stated that the extraordinary government support would trigger a complete write-down of the nominal value of all AT1 debt of Credit Suisse in the amount of around CHF 16 billion. The same weekend, shareholders were given UBS stock.

FINMA set out its reasoning four days later. The instruments contractually provide that they will be completely written down in a viability event, in particular if extraordinary government support is granted, and the emergency ordinance authorises FINMA to order that write-down. On 19 March Credit Suisse received liquidity assistance loans secured by a federal default guarantee, which FINMA held satisfied the contractual condition.

The mechanism is the thing to understand, and it is not that Swiss law inverted the creditor hierarchy. Nobody was placed below equity in a liquidation, because there was no liquidation. The AT1 instruments were extinguished by a regulatory order acting on a contractual trigger. The shares were exchanged in a merger, and merger consideration is negotiated rather than distributed down a priority waterfall. Ranking is enforced in insolvency, and this outcome was engineered to avoid one.

Three figures for the amount written down appear in three official documents, and this page does not reconcile them. FINMA said around CHF 16 billion. Credit Suisse's own first quarter accounts describe CHF 15 billion of AT1 notes written to zero. The Federal Administrative Court put the nominal value at approximately CHF 16.5 billion.

For anyone who owns subordinated debt, the lesson is about documents rather than about Switzerland. A viability clause that triggers on the granting of extraordinary government support converts a credit instrument into something closer to a bet on whether a state will intervene, and states intervene when things go badly. That is credit analysis, not legal trivia.

What Did the Federal Administrative Court Decide in 2025?

Roughly 3,000 bondholders filed around 360 appeals against FINMA's order. On 1 October 2025 the Swiss Federal Administrative Court issued a partial decision in the lead case, B-2334/2023, and revoked it.

The court's reasoning went to both legs of FINMA's justification. On the contract, it found the conditions for a write-off were not fulfilled because the viability event had not been triggered. On the statutory basis, it found the bondholders' property rights were seriously interfered with, which would have required a clear and formal legal basis that did not exist, and held that the ordinance provision relied on, Article 5a, was unconstitutional.

Nothing has changed for the bondholders yet, and it may not. FINMA announced on 15 October 2025 that the judgment was not final and that it would contest it before the Federal Supreme Court inside the thirty-day appeal window. The Administrative Court has not yet ruled on the bondholders' second request, which was that the write-off be reversed rather than merely the order revoked, and the remaining cases are suspended until the revocation itself becomes final. The instruments remain written down and no compensation has been ordered.

Two things are worth carrying away while it remains unresolved. First, a legal outcome delivered under emergency conditions is not final at the moment it is delivered, and the interval before review can run to years. An investor who assumed in March 2023 that the matter was closed was wrong, and one who assumes today that it has been reopened for good is assuming something a court has not yet said. Second, this is now a live question about the credibility of the whole loss-absorbing capital regime, because the instrument only works as designed if the write-down is enforceable when it is needed.

Legal status as at the research date of this page, 26 August 2026. This is an active proceeding and the position can change. Check the Federal Supreme Court and FINMA directly before relying on it.

What Did UBS Pay, and What Did It Book?

The headline number announced on 19 March 2023 was CHF 3 billion, delivered as one UBS share for every 22.48 Credit Suisse shares, which UBS described at the time as equivalent to CHF 0.76 per share. The final accounting figure moved with the UBS share price: on completion, 3,949 million Credit Suisse shares converted into 176 million UBS shares at CHF 18.35, giving consideration of CHF 3,223 million, or USD 3,547 million. Credit Suisse shareholders ended up holding 5.1 percent of UBS.

Set that against the seller's last published accounts. Total shareholders' equity was CHF 45,129 million at the end of 2022, so the price was about 7 percent of reported book equity, and market capitalisation on the same date was CHF 11,062 million, so it was about 29 percent of what the market had valued the equity at ten weeks earlier.

What UBS booked is the more striking half. In its second quarter 2023 accounts it recognised negative goodwill of USD 28,925 million, the accounting consequence of acquiring net assets whose measured fair value exceeded the consideration paid. That single line produced a quarterly net profit of USD 28,875 million and a return on equity of 160.7 percent. UBS itself published the number that matters more: excluding negative goodwill, integration expenses and acquisition costs, return on equity for the quarter was 3.9 percent.

Negative goodwill of that scale is not a windfall in cash, and treating it as one is a mistake. It reflects a price agreed over a weekend, under state pressure, for a business nobody else could bid for, whose eventual value depended on litigation, restructuring costs and client attrition that had not yet happened. Read it as a measure of the uncertainty the buyer was being paid to absorb, not of how cheaply it bought. UBS reviewed the assets covered by the Confederation's CHF 9 billion guarantee, decided it was no longer required, and terminated it on 11 August 2023, less than two months after closing.

Why Did Credit Suisse Report a Record Profit in Its Final Quarter?

On 24 April 2023, five weeks after agreeing to be absorbed, Credit Suisse published first quarter results showing pre-tax income of CHF 12,764 million and net income attributable to shareholders of CHF 12,432 million. The comparative quarter had produced a loss of CHF 273 million. The press release explains the reversal in its own headline: the results reflect the write-down of CHF 15 billion of AT1 capital notes.

The write-down flowed through treasury results in the corporate centre, and it flowed in as revenue, because extinguishing a liability without paying it produces a gain. Strip it out and the underlying business had an adjusted pre-tax loss of CHF 1.3 billion for the same three months, alongside a CHF 1.3 billion goodwill impairment almost entirely in wealth management.

The capital ratio moved the same way and for the same reason. The common equity tier 1 ratio rose to 20.3 percent from 14.1 percent, and the bank stated plainly that the increase was mainly driven by the write-down of the AT1 capital notes as ordered by FINMA. The strongest capital ratio Credit Suisse ever published was the one it published after it had already ceased to be a going concern in any meaningful commercial sense.

Nothing here was hidden and this is not an accounting scandal. It demonstrates something more useful: a reported profit and a regulatory capital ratio are constructions, and both can move sharply in the right direction for reasons unconnected to the business improving. Anyone screening banks on trailing profitability or capital strength would have seen Credit Suisse's numbers improve dramatically in the quarter it died. Reading the reconciliation between reported and adjusted results, which is what fundamental analysis is for, was the only defence. The outflows, meanwhile, had not stopped: CHF 61.2 billion left in the quarter and the bank said on 24 April that they had moderated but not reversed.

What Did the Swiss Bond Market Do After the Rescue?

Most retellings reach for the Credit Suisse share price, which by March 2023 had stopped being informative because everyone already agreed the equity was close to worthless. The Swiss bond market is the better instrument, and the SNB publishes the daily series needed to read it.

Take the government leg first. The spot yield on two-year Confederation bonds was 1.638 percent on 8 March 2023 and 0.511 percent on 20 March, a fall of 112.7 basis points in eight sessions. The ten-year fell 74.9 basis points, from 1.514 to 0.765 percent. The largest one-day decline in the two-year yield across the whole of 2022 and 2023 came on 13 March, two days before any Swiss authority said anything in public about the bank.

The credit leg is where the interesting result is. The SNB publishes eight-year Swiss franc yields by borrower category on a consistent maturity basis, one series for the Confederation and one for Swiss commercial banks. The gap measures what the market charged Swiss banks to borrow relative to their own government. Swoopr Investment computed it daily from 2015 through 2024, 2,513 observations.

On 8 March 2023 the gap was 70.8 basis points, unremarkable, and through the crisis week it drifted between 80 and 94. On 20 March, the first trading day after the merger was announced, it reached 134.7 basis points, the widest reading in the entire ten-year window, and it was still above 100 at the end of the month.

That inverts the usual expectation. The rescue was announced on a Sunday evening specifically to restore confidence, and the cost of Swiss bank credit hit a decade high the next morning. The plausible explanation, and it is an explanation rather than a proof, is that the resolution taught the market something new about the terms on which Swiss bank capital could be extinguished. Removing one bank's failure risk does not lower the price of bank debt if the method repriced everyone else's subordination.

Was Credit Suisse Brought Down by Silicon Valley Bank?

The two failures were eleven days apart and the sequencing invites a causal story that the chronology does not support.

The evidence for a connection is real. The Parliamentary Investigation Committee states that the United States regional banking crisis and its immediate impact on Credit Suisse took the federal authorities by surprise before they had completed their analysis, and dates the acute Swiss phase from 15 March. FINMA and the SNB felt the need to say publicly, that same day, that American bank problems posed no direct contagion risk.

The evidence against treating it as the cause is stronger. Credit Suisse had already lost CHF 138 billion of deposits in the fourth quarter of 2022, the authorities had been in crisis mode since October, and the committee found the bank came close to insolvency twice before any American bank failed. A bank that has lost 41 percent of its deposits over a year and trades at a quarter of book value does not need an external trigger. It needs a date.

The mechanisms were also different. Silicon Valley Bank failed on interest rate duration: government and agency securities that had fallen in price as rates rose, funded by uninsured corporate deposits from a single industry, with no credit losses anywhere in the chain. Credit Suisse failed on reputation after a decade of scandals and control failures, with an asset side nobody was seriously alleging to be impaired and a capital ratio that never breached its minimum. One was a duration accident inside an otherwise conventional bank; the other was the terminal stage of a franchise clients had been leaving for years.

March 2023 supplied a market willing to believe the worst about any bank, and Credit Suisse was the bank about which the worst was most plausible. The American failures set the clock. They did not build the mechanism.

What Had Been Going Wrong at Credit Suisse for a Decade?

A bank does not lose 41 percent of its deposits in a year over one bad quarter. The supervisory record FINMA published in December 2023 is the most compact description of the underlying problem available.

Since 2012, FINMA conducted 43 preliminary investigations of Credit Suisse for potential enforcement proceedings, and fourteen enforcement proceedings followed, eleven of them in 2018 or later. Between 2018 and 2022 it carried out 108 on-site reviews and recorded 382 points requiring action, of which 113 were classed as high or critical risk. That is a supervisor with its hands on the institution continuously for a decade, and a bank that kept generating findings anyway.

Two failures were large enough to have their own financial statement line. Credit Suisse recorded a net charge of CHF 4.8 billion in 2021 for the Archegos matter, where an independent report found a failure to effectively manage risk in the investment bank's prime services business by both the first and second lines of defence. Separately, the supply chain finance funds matter was carved out of the discharge shareholders granted the board and executive board for 2021 at the 2022 annual meeting, and when the board asked that same meeting to grant discharge for the 2020 financial year on the same carved-out terms, shareholders voted it down.

The parliamentary committee assigned responsibility without ambiguity: it lies with the Board of Directors and Executive Board, who had defied numerous interventions by FINMA in the preceding years. It also found FINMA's supervisory activities, while intensive, lacked sufficient impact, and finds it regrettable that FINMA did not withdraw recognition of the guarantee of proper business conduct, the Swiss supervisory power that goes furthest.

What matters is what the sequence produced. Each episode was individually survivable. What accumulated was not a solvency problem but a reputational one, and reputation is an input to a wealth manager's revenue in a way it is not for a lender. When clients pulled approximately 15 percent of assets in a single quarter, they were not making a credit judgment about the balance sheet. They were declining to keep paying a fee to an institution that had stopped looking competent, a business risk that risk management frameworks built around market and credit exposure do not naturally capture.

What Has Changed in Swiss Banking Regulation Since?

The rules described here can and probably will change. What follows is the position as researched on 26 August 2026, and anyone acting on it should verify the current state directly.

Switzerland's too-big-to-fail framework was built after 2008 on the premise that a systemically important bank could be recapitalised through a resolution in which subordinated creditors absorbed losses and the bank kept operating. In March 2023 that framework existed, was available, and was not used. The parliamentary committee's judgment is that the current regulations are not designed to deal with a crisis of confidence and overlook some important market indicators.

The Federal Council published its own report on banking stability on 10 April 2024, proposing 22 measures for direct implementation with seven more to be examined. Two bear directly on this case. Capital requirements for systemically important banks should be tightened in a targeted way and supplemented with a forward-looking component. And the public liquidity backstop, improvised by emergency ordinance, should be enshrined in ordinary law.

The parliamentary committee, reporting in December 2024 after 79 hearings across 45 sessions, addressed twenty recommendations to the Federal Council along with six postulates, four motions and one parliamentary initiative. It recommended restricting future easing of capital and liquidity requirements, the direct descendant of the 2017 filter finding.

The structural fact underneath all of it is arithmetic. Switzerland now has one remaining globally systemically important bank, and the committee observed that UBS is many times larger relative to Swiss gross domestic product than comparable institutions are relative to their own countries' output. March 2023 was also the second time the Swiss state had intervened to save such a bank. A country that has done this twice and has one candidate left for a third is not in a comfortable position.

Why Will the Next Bank Failure Not Look Like This One?

Four things about this particular failure were peculiar enough that preparing for a rerun would be preparing for the wrong event.

The failing business was a wealth manager, and its inventory was trust. Most bank failures run through credit losses, as in the 2008 financial crisis, or through interest rate duration, as in the savings and loan crisis. Credit Suisse's assets were not the problem. Its clients were leaving because they no longer wanted the relationship, and a wealth management client can move assets in an afternoon with no penalty. That funding is more mobile than retail deposits and far more mobile than a loan book.

The state had a five-month head start and still improvised. The authorities were in crisis mode from October 2022 and had analysed liquidation, emergency liquidity assistance, temporary public ownership and a takeover. When the moment came they still legislated by emergency ordinance over a weekend, with fallback options pursued in parallel because the negotiation might have failed. Assuming the next episode will produce a clean, pre-planned resolution is not supported by the one episode where the planning had been done.

The buyer was uniquely available. A domestic competitor large enough to absorb a globally systemic bank, and willing to do so under a state guarantee, is not a standing feature of most banking systems. The committee understood that a solution with a foreign bank was no longer feasible by that point, even if it might have been better for Switzerland in the longer term. That constraint produced the concentration Switzerland now has to regulate around.

The loss allocation is under judicial review. The AT1 write-down that made the transaction work is the specific action a Swiss court has since found unlawful. Whatever the Federal Supreme Court decides, no authority can now use this precedent with confidence, so the next resolution will not look like this one even if the next bank does.

Common Myths About the Credit Suisse Collapse

"Credit Suisse was insolvent." Nothing published says so. Its CET1 ratio was 14.1 percent, its liquidity coverage ratio 144 percent, and FINMA's crisis review states that the bank satisfied the regulatory capital requirements. What the parliamentary committee added is separate: without the 2017 filter it would have failed capital adequacy requirements significantly in 2022. Failing a capital requirement is not insolvency, and neither is what caused the sale. It was sold because its funding left.

"Bondholders were ranked below shareholders." They were not ranked at all, because there was no insolvency in which to rank them. The AT1 instruments were extinguished by a FINMA order acting on a contractual viability trigger; the shares were exchanged in a negotiated merger for CHF 3 billion of UBS stock. The outcome offended the intuition that subordination protects you, and it did so because subordination is enforced in bankruptcy and this was designed to avoid bankruptcy.

"Swiss taxpayers bailed out Credit Suisse." The Confederation guaranteed CHF 100 billion in favour of the SNB and CHF 9 billion of UBS losses, collected roughly CHF 200 million in fees, and states that it assumed no losses. Every guaranteed loan was repaid by the end of May 2023. Whether a state should have written those guarantees is a legitimate argument. Describing the outcome as taxpayers paying is factually wrong.

"The rescue calmed the market." It stopped the run at Credit Suisse. It also coincided with the widest gap between Swiss commercial bank and Confederation bond yields of the decade from 2015 to 2024, on the first trading day after the announcement, and the SNB raised its policy rate from 1.00 to 1.50 percent five days later.

"Nobody could have seen it coming." The shares traded at roughly a quarter of book value at the end of 2022, the deposit flight was disclosed in the bank's own filings, and the ratings had been downgraded through the year. What was genuinely unforeseeable was not the failure but the resolution, and specifically that a subordinated instrument would go to zero while the equity beneath it received consideration.

What a Reader Can Actually Carry Forward

The tempting conclusion is a rule about avoiding European bank shares, which would have been both unprofitable and irrelevant. The useful material concerns how to read what an institution publishes about itself.

What generalizes

• A regulatory ratio is the output of a rulebook, not a measurement of health. Credit Suisse met its requirements in the week it was sold, and its capital ratio rose to 20.3 percent afterwards because it had destroyed its own bondholders. Any screen built on a published prudential metric would have ranked this bank adequate the whole way down.
• Funding that can leave without cost will leave without warning. Wealth management clients pulled roughly 15 percent of assets in one quarter, with no penalty and no reason to wait. Ask what it would cost a funding provider to walk away, because that cost is the only thing slowing them down.
• Read the trigger, not the rating. The AT1 instruments went to zero because their documentation converted on the granting of extraordinary government support. Nothing in a credit rating or a yield describes that clause, and for subordinated bank debt it is the term that decides the outcome.
• Stress tests answer the question you asked them. A ratio calibrated to thirty days told the truth about thirty days and said nothing at all about the six months of continuous withdrawal that actually killed the bank. Credit Suisse never failed the test it was set; it failed a test nobody had written down. Choosing which scenario to run is the part that decides everything, and it is the whole subject of stress testing and scenario analysis.

What does not generalize

• The write-down of subordinated capital while equity was paid. A Swiss court has since found that order unlawful and the appeal is unresolved, so treating it as the standard handling of loss-absorbing instruments means reading precedent from a decision under review.
• The clean repayment. Around CHF 168 billion of central bank lending was repaid within months and the state took no loss, which depended on the collateral being genuinely good. That is a fact about this bank, not about rescues.
• The availability of a domestic buyer. Very few countries have a second bank capable of absorbing the first one. Switzerland used up its only one.

The one question worth asking now

Not whether a given bank is sound, which almost nobody can assess from outside, but a narrower question with a findable answer: for anything you own, what would have to be true for you to be repaid in full, and is that condition written in a contract or dependent on a decision somebody makes under pressure? Credit Suisse depositors were repaid because a state chose to arrange it. AT1 holders were not. Both groups thought they held a claim on a bank. One held a claim on a policy decision, and the documentation said so.

Applied to a bank funded by ordinary deposits, the Silicon Valley Bank case study covers the deposit insurance half of the question, and the collapse of Long-Term Capital Management covers what happens when the rescue is privately organised rather than legislated. The 2022 rate shock supplies the macro setting all three shared.

References

Every figure on this page was verified against the following sources, retrieved on 26 August 2026:

• Credit Suisse Group AG: Annual Report on Form 20-F for the Year Ended December 31, 2022: every 2022 figure quoted, including the CHF 7,293 million net loss, the 14.1 percent CET1 and 144 percent liquidity coverage ratios, the CHF 138 billion and CHF 159.6 billion deposit declines, equity of CHF 45,129 million, market capitalisation of CHF 11,062 million and the CHF 4.8 billion Archegos charge.
• Credit Suisse Group AG: Report on Form 6-K, March 9, 2023: the late call on the evening of March 8, 2023 from the SEC, and the resulting delay to the annual report.
• Credit Suisse Group AG: Report on Form 6-K, March 14, 2023: publication of the 2022 Annual Report, the 4 April agenda and the proposed CHF 0.05 dividend.
• Credit Suisse Group AG: Report on Form 6-K, March 16, 2023: the intention to borrow up to CHF 50 billion under a covered loan facility.
• Credit Suisse Group AG: Report on Form 6-K, March 20, 2023, Credit Suisse and UBS to Merge: the merger agreement, the 22.48 exchange ratio and the ordinance permitting it without shareholder approval.
• Credit Suisse Group AG: Report on Form 6-K, April 24, 2023, First Quarter 2023 Results: pre-tax income of CHF 12,764 million, net income of CHF 12,432 million against a CHF 273 million loss a year earlier, CHF 15 billion of AT1 notes written to zero, the CHF 1.3 billion adjusted loss and the CHF 1.3 billion goodwill impairment, the 20.3 percent CET1 ratio, the approximately 153 percent quarter-to-date liquidity coverage ratio on 14 March 2023, CHF 61.2 billion of net asset outflows, CHF 67 billion of deposit outflows and CHF 108 billion outstanding after CHF 60 billion of repayments.
• UBS Group AG: Second Quarter 2023 Report: negative goodwill of USD 28,925 million, the 160.7 and 3.9 percent returns on equity, the CHF 3,223 million purchase price, the 5.1 percent shareholding, the loss protection structure, the two CHF 100 billion liquidity layers and the 11 August 2023 termination.
• UBS Group AG: Report on Form 6-K, March 20, 2023, News Release of 19 March 2023: the CHF 3 billion headline consideration and the CHF 0.76 per share equivalent.
• FINMA and the Swiss National Bank: Statement on Market Uncertainty, 15 March 2023: that Credit Suisse met the requirements imposed on systemically important banks, and that the SNB would provide liquidity if needed.
• FINMA: FINMA Approves Merger of UBS and Credit Suisse, 19 March 2023: the crisis of confidence and considerable outflows of client funds, and the write-down of all AT1 debt of around CHF 16 billion.
• FINMA: Basis for Writing Down AT1 Capital Instruments, 23 March 2023: the viability event triggering on extraordinary government support, and the ordinance authorising the order.
• FINMA: Lessons Learned from the CS Crisis, 19 December 2023: 43 preliminary investigations since 2012, fourteen enforcement proceedings, 108 on-site reviews, 382 points requiring action of which 113 were high or critical, and CS AG as the weakest link.
• Swiss Federal Administrative Court: Unlawful Write-Off of AT1 Capital Instruments, Judgment B-2334/2023 of 1 October 2025: the revocation, approximately CHF 16.5 billion of nominal value, around 3,000 complainants and 360 appeals.
• FINMA: FINMA to Appeal Partial Decision of the Federal Administrative Court Concerning AT1, 15 October 2025: FINMA contesting the judgment before the Federal Supreme Court.
• Swiss Federal Department of Finance: UBS Takeover of Credit Suisse: the CHF 9 billion and CHF 100 billion guarantees, the 11 August 2023 termination, the roughly CHF 200 million of fees and the statement that the Confederation assumed no losses.
• Swiss Parliament: Lessons from the Credit Suisse Crisis, PInC Identifies Need for Action, 20 December 2024: the 2017 regulatory filter and the finding that without it the bank would have failed capital requirements significantly in 2022, the near insolvency in October and December 2022, the acute phase dated from 15 March 2023, and the attribution of responsibility to the board and executive board.
• Swiss Federal Department of Finance: Federal Council Report on Banking Stability, 10 April 2024: the 22 measures with seven more to be examined, the tightening of capital requirements, and the proposal to enshrine the public liquidity backstop in ordinary law.
• Swiss National Bank data portal, for every Swiss yield and policy rate quoted: Swiss National Bank: Yields on Bond Issues, Cube rendoblid, for the Confederation spot rates and the eight-year Swiss bank yields, and Swiss National Bank: Money Market Rates, Cube snbgwdzid, for the policy rate of 1.00 percent through 23 March 2023 and 1.50 percent from 24 March.

Figures deliberately not stated. This page gives no Credit Suisse share price on any dated day, no percentage decline for the shares or any bank index, no size for the supply chain finance funds, no Swiss deposit protection limit, no dollar figure for the Archegos trading loss as distinct from the CHF 4.8 billion charge, and no daily outflow figure for the week of 15 March 2023, because no verified source supplied them.

Method note: the bank yield gap is the eight-year Swiss franc yield for Swiss commercial banks minus the eight-year yield for the Confederation, both from the same SNB table on the same maturity basis, over 2,513 daily observations from 1 January 2015 to 31 December 2024. Gross drawings of about CHF 168 billion are the CHF 108 billion outstanding at 31 March 2023 plus the CHF 60 billion repaid in the same quarter. The two comparisons of the UBS price divide UBS's stated consideration by Credit Suisse's reported equity and market capitalisation at 31 December 2022, two documents with two measurement dates.

Several rules described here can change and one is actively disputed. The legality of the AT1 write-down is on appeal to the Swiss Federal Supreme Court and the Administrative Court's judgment is not final. Swiss capital and liquidity requirements for systemically important banks, and the statutory basis for the public liquidity backstop, were under legislative revision at the research date. Verify the current position before relying on any of it.

This page describes an episode that has already run its course. Nothing in it is investment advice, none of it is a prediction, and no sentence here is a statement about how any bank stands today.

Frequently Asked Questions