Credit Suisse AT1 ruling: don’t get your hopes up, bondholders
Prasad Gollakota
The rally in Credit Suisse’s written down Additional Tier 1 bonds as a result of a shock court judgment has tempted some to call a turn. This is not investment advice, but that view is likely misplaced. The Federal Administrative Court’s decision to revoke Swiss market regulator Finma’s March 19 2023 decree to write off all CS AT1s ahead of its takeover by UBS leaves enough uncertainty in its reasoning that it should not be treated as final, and it is certainly a shaky basis for buying the bonds.
The court’s decision, announced on October 1 (which both Finma and UBS are appealing), turns on a single question: was CS "non-viable" on March 19 2023 when Finma ordered the writedown of SFr16.5bn (now US$20.6bn) of AT1s in accordance with the terms of the bonds? Finma and UBS argued yes; AT1 holders argued no.
Competing readings
Switzerland implemented Basel III standards through the Banking Act and the Own Funds and Risk-Distribution Ordinance. Article 29 of the latter requires that AT1 capital be written off "at the latest before recourse to public assistance" or "if Finma orders it to avoid insolvency".
The terms in the CS AT1s reflect this and define a "viability event" in two ways:
- Limb A: Finma has notified CS that it has determined that a writedown of the notes is an essential requirement to prevent CS from becoming insolvent, or;
- Limb B: CS has received extraordinary public sector support that has the effect of improving its capital adequacy and without which, in Finma’s determination, it would become insolvent.
The court made clear that once a viability event occurs, the writedown is automatic and irreversible.
Finma and UBS argued that without the Swiss National Bank’s liquidity facilities and liquidity backstop, CS would have collapsed within days; hence, support improved its capital situation indirectly by averting losses. Investors argued the opposite: the bank still met all regulatory capital ratios, and the support was purely liquidity assistance and did not alter CS’s capital position.
Under Swiss civil law, a contract is interpreted according to the parties' real common intention or, failing that, by the principle of trust (ie, how a reasonable counterparty would read the text in context).
Intention of the parties
The court relied heavily on CS's email on March 19 2023 to Finma that explicitly stated that "the [liquidity] measures currently assessed are designed to create confidence and obtain liquidity, not capital", and that it remained compliant with all capital requirements. This, the court said, reflected the parties' shared understanding that public support would trigger a viability event only if it was needed to remedy an insufficient level of capital, not a temporary liquidity shortfall.
This reasoning provides a sound basis to challenge the court's conclusion. It is akin to asking a cardiac patient how long they have left while the surgeons are performing open-heart surgery. Worse still, it assumes they will say something insightful after having ignored persistent and serious cardiovascular symptoms for years.
The prospectus and risk factors in the AT1 documents repeatedly warned that Finma could order a writedown at its discretion as part of a going-concern loss-absorption regime. That broader context, combined with the market's understanding of AT1s since 2013, arguably forms part of the parties' shared intention.
Investors knew the instruments were designed to absorb losses well before insolvency, irrespective of the bank’s capital ratios, and that Finma's judgment governed that process. The real question is one of viability, not merely capital adequacy.
A broader reading could go further. In a crisis, Finma's will is effectively CS's will. Under Swiss law, Finma has sweeping powers to direct the management of supervised institutions. Decisions taken by the bank in such circumstances are not independent but shaped, and sometimes compelled, by Finma's oversight. On that view, the "real will" of the parties cannot be separated into regulatory and contractual spheres. Finma's determination was not external to the contract but the very mechanism through which the contract was carried out.
The trust principle
Even if "intent" was unclear, the court held that an "objective" reading leads to the same result. The phrase improving "capital adequacy" can only mean strengthening equity. A liquidity loan increases both assets (cash) and liabilities (borrowings) by the same amount, leaving capital unchanged. Therefore, such support cannot "improve" capital adequacy within the plain language of the clause.
Yet under Basel's own framework, capital adequacy is not a static balance-sheet measure but a forward-looking assessment of whether the bank has sufficient capital to withstand stress in its operating environment. It explicitly calls for supervisory and managerial judgment, not mere arithmetic. This represents another basis to challenge the court's decision.
Any "reasonable" investor informed about AT1s knew these were going-concern instruments whose operation depended on Finma's supervisory judgment, often before a capital ratio breach. The average investor would not draw fine distinctions between liquidity and capital; both were part of the same continuum of viability. Reading "improvement" as a capital accounting concept risks rendering the preventive trigger devoid of value.
Firesale is "speculative"
The court concluded CS was "always sufficiently capitalised", according to the government's statement of March 29 2023. The liquidity facilities were short term and reversible. They did not alter capital ratios, nor were they intended to do so. Moreover, Finma's argument that such support indirectly improved capital by preventing firesales and resulting losses was dismissed as "speculative".
This represents a further basis to challenge the court's decision. In practice, the link between liquidity and capital is anything but speculative. Without SNB liquidity, CS would likely have been forced to sell assets into collapsing markets, crystallising heavy losses that would have reduced Common Equity Tier 1 capital and triggered the capital shortfall the clause was designed to pre-empt. Liquidity support prevented that spiral, stabilising the balance sheet. By rejecting that causal link, the court underestimated this dynamic supervisory trigger.
No "technical discretion"
The court rejected Finma's claim that its technical expertise and supervisory role entitled it to interpret the trigger flexibly. Switzerland had chosen a contractual, not statutory, model of Basel III adoption: Finma approves the language of AT1s before issuance, but once sold, AT1s are governed by contract law between issuer and investors.
That distinction proved decisive. After issuance, interpretation falls under contract law, not ongoing supervisory discretion. Finma's expertise may shape approval before issuance, but not interpretation after the fact.
Finma's determination
Finma also argued that it had determined the writedown was essential to avoid insolvency. The court held there was no evidence of such a determination before the order was issued. The March 19 decision referred only to "public assistance", corresponding to Limb B, not to any formal insolvency finding required by Limb A.
This also represents a basis to challenge the court's decision. The court appeared to gloss over why emergency public assistance was provided at all. What other purpose could it have served other than to prevent the financial collapse of Credit Suisse Group?
The absence of a separate written "determination" is irrelevant. The act of ordering the writedown is in itself a determination that it was essential to prevent insolvency. That is precisely what the contractual clause envisages. In supervisory practice, form follows substance: Finma's instruction necessarily implies that it had reached the required judgment, even if not expressed in a standalone notice. Insisting on a separate document elevates form over function.
Bringing it all together
The court's decision ultimately rests on its view that extraordinary state liquidity support by itself does not trigger either of the contractual non-viability limbs.
For the reasons outlined, that argument is perplexing.
As a rule of thumb, a banking crisis unfolds much like a cardiac event. Illiquidity is the first chest pain, an early warning that bloodflow is constrained. At this point, intervention becomes essential and the patient is only viable with life support. Left unattended, the condition progresses to cardiac arrest; the equivalent of insolvency. By then, other health metrics may still look reassuring, but the patient is dead. Viability isn't judged by those metrics; it's defined by whether the heart can beat without life support.
Finma has confirmed it will contest the court's judgment and appeal to the Federal Supreme Court. It will likely take years for this saga to reach its endpoint. As it stands, no outcome is certain. But what does seem certain to me is that, irrespective of the administrative court's ruling, without the life support from the SNB, CS was dead. And that means Finma was entirely within its rights to do what it did.
Prasad Gollakota worked at UBS from 2003 until 2012 and was co-head of the global capital solutions group during the global financial crisis. He is now chief content and operating officer at educational technology company xUnlocked. For his take on the CS AT1 writedown at the time it took place, see here.