Bailout Blues: The Write-Down of the AT1 Bonds in the Credit Suisse Bailout

The recent bailout of Credit Suisse is noteworthy for many reasons. One of them is that, while AT1 bondholders were wiped out, shareholders were not. This violates the ‘absolute priority rule’ which is central to corporate reorganisation and bank resolution regimes. In this article, we analyse the motives and mechanics of the write-down and argue that, given the bond terms, the prospect for a legal challenge by the bondholders is slim. At the same time, we question the merits of the write-down. Bondholders should fare no worse than common equity, regardless of whether a financial institution is put in an insolvency proceeding or bailed out, and the applicable bond terms should reflect this. We also raise the issue of a more principled approach to bailouts more generally.


The Bailout of Credit Suisse
A key principle of Chap. 11 corporate reorganisations, which can be considered as the blueprint for reorganisation regimes across Western jurisdictions, is the 'absolute priority rule'. 1 It requires that the claims of a dissenting class of creditors be paid in full before any stakeholders in a class junior to the dissenting class may receive or retain any property in satisfaction of their claims. As a consequence, creditors cannot be forced to accept cuts if shareholders are not completely wiped out. This principle is also central to legal frameworks governing the restructuring of banks. For example, Article 34(1)(a) of the European Bank Recovery and Resolution Directive 2 (the 'BRRD') stipulates that 'Member States shall ensure that … resolution action is Extended author information available on the last page of the article 1 See 11 U.S.C. § 1129. taken in accordance with the following principles: (a) the shareholders of the institution under resolution bear first losses …'.
Against this background, one would have expected that the write-down of the Additional Tier 1, or AT1, bonds (the 'CS AT1 Bonds') in the Credit Suisse ('CS') bailout would not have happened without its shareholders also being wiped out. 3 But the unexpected did happen: the bonds were written down, and CS's shareholders received UBS shares worth USD 3.25 billion under the bailout deal. 4 In this article, we analyse the motives and mechanics of the write-down and argue that, given the bond terms, the prospect for a legal challenge by the bondholders is slim. At the same time, we question the merits of the write-down. Bondholders should fare no worse than common equity, regardless of whether a financial institution is put in an insolvency proceeding or bailed out, and the applicable bond terms should reflect this. We also raise the issue of a more principled approach to bailouts generally. Bankruptcy and bank resolution are two heavily regulated fields of the law. Ad hoc bailouts are almost completely unregulated. This needs to change.
CS's demise can be summarised by the exchange between Bill and Mike in Hemingway's 'The Sun Also Rises'. 'How did you go bankrupt?' Bill asked. 'Two ways', Mike said. 'Gradually and then suddenly.' When Silicon Valley Bank ('SVB') failed on 10 March 2023, regulators, supervisors and investors across the globe started to consider what else could go wrong and whether the international financial system was really in better shape than it was immediately before the 2007-2008 financial crisis. In the United States, eyes turned to regional banks, anticipating that the failure of SVB might trigger a 'flight to safety' response and result in depositors deciding to move funds to larger banks, expecting to benefit from better governmental protection in the event of a run. 5 JP Morgan and other leading institutions attempted to assuage the markets by making available up to USD 30 billion to First Republic Bank, the bank that many consider to be the next of the dominos to fall. 6 In Europe, eyes turned towards CS, a global, systemically important bank that had been at the centre of everyone's mind whenever problems in the European financial system were considered -particularly after its involvement in the Archegos and Greensill scandals. 7 CS had been the subject of a probe by the SEC involving accounting errors for eight months before the bank's collapse. 8 Attempts had been made to propose and implement various restructuring plans, with the last iteration launched in October 2022. 9 4 See Massoudi et al. (2023). 5 See, e.g., Hughes et al. (2023). 6 See Aliaj et al. (2023). 7 See, e.g., Walker and Morris (2023). 8 SEC (2022), White (2023) or Halftermeyer (2023). 9 See Walker (2022).
The nail in CS's coffin came on 15 March 2023, when the chairman of its main shareholder, the Saudi National Bank, 10 uttered the now infamous 'absolutely not' 11 when asked about an additional liquidity injection into the bank. Despite an immediate liquidity backstop by the Swiss National Bank ('SNB') of up to CHF 50 billion, 12 market pressure mounted, and during the weekend of 17-19 March, SNB, the Swiss Financial Market Supervisory Authority ('FINMA'), and the Federal Council ('the Swiss government') engineered a takeover of CS by rival bank UBS. 13 CS's shareholders received CHF 3 billion in UBS shares, the AT1 bondholders were wiped out, UBS received an additional CHF 100 billion liquidity line from SNB backed by a federal default guarantee, and the Swiss government also provided a conditional loss guarantee to UBS of up to CHF 9 billion.

The Exceptionality of the Bailout
The bailout deal is noteworthy for many reasons. First, Switzerland has a bank resolution regime comparable to the BRRD. 14 But when confronted with the distress of a global, systemically important bank, the relevant authorities did not consider that the regime would work in an emergency. Indeed, the Swiss finance minister believed that following the existing protocols 'would have triggered an international financial crisis'. 15 The Swiss government proceeded to deploy enormous public funds in an ad hoc rescue effort. If it was thought that resolution according to a dedicated regime set up after the 2007-2008 financial crisis would not work in Switzerland, what can we expect in other jurisdictions with similar regimes? 16 The Swiss government might have bailed out CS, but at the cost of triggering an international regulatory crisis, disrespecting a decade-long international regulatory effort to create a resilient resolution system for systemically important financial institutions such as CS.
10 Saudi National Bank became CS's major shareholder as recently as December 2022 after a USD 4 billion capital round, and held 9.88% of the share capital in CS. 11 See Elder (2023). 12 See Credit Suisse (2023a). 13 See Massoudi et al. (2023). 14 See, for example, Schelo (2020) (particularly pp 216-223), IMF (2019), or FINMA's website at https:// www. finma. ch/ en/ enfor cement/ recov ery-and-resol ution/ or FDF (2023). 15 Jones (2023) for the statement of the Swiss finance minister that the use of the (resolution) procedure under Swiss too-big-to-fail regulations 'would have triggered an international financial crisis', before noting that 'I have come to the realisation in recent weeks that a globally active, systemically important bank cannot simply be wound up according to the too-big-to-fail plan.' Also, see FDF (2023) (Swiss Federal Department of Finance), noting that the resolution plan (including the spin-off of CS's Swiss business arm) was not viable and that 'an alternative with significantly lower economic and financial risks existed, i.e. the takeover of CS by UBS.' 16 As to the strength of the Swiss system, see, on the too-big-to-fail regulation, FDF (2023)  Second, to carry out the rescue, the Swiss government passed an Ordinance of the Swiss Federal Council on 16 March 2023 17 (amended on 19 March) 18 (the 'Emergency Ordinance'). The new provisions introduced by the Emergency Ordinance included: (i) the establishment of priority rights in insolvency for the additional liquidity assistance provided by SNB, as well as the granting of default guarantees for such liquidity assistance; 19 (ii) the granting of powers to FINMA to bypass the need for general meetings to approve transactions involving systemically relevant banks; 20 (iii) the granting of powers to FINMA to require the write-down of 'additional core capital'; 21 and (iv) the granting of guarantees for loss coverage in relation to potential losses arising from certain assets of the acquired bank. 22 As a consequence, the bailout was implemented by administrative fiat, bypassing both parliament and the shareholder assemblies of the affected banks. Unsurprisingly, Switzerland's Lower House later conducted a symbolic vote against providing state guarantees for UBS's takeover of CS, reflecting a high degree of public discontent with the deal. 23 Third, based on Article 5a of the Emergency Ordinance and the applicable bond terms, the holders of CHF 16 billion of AT1 bonds in CS were completely wiped out while equity holders, despite being materially diluted, 24 were not. The Swiss authorities apparently believed that it was more important to placate CS shareholders (particularly its major international shareholders 25 and domestic retail investors) than the AT1 bondholders. Shareholders could have initiated blocking litigation, anchor investors are needed to meet future financing needs, and employee shareholders must be motivated to come to work. By contrast, the AT1 bondholders -primarily non-Swiss sophisticated institutional investors -were apparently seen primarily as interchangeable contributors of risky capital or even speculators. Indeed, some came in very 'late in the game', i.e., days before CS's collapse. 26 Massoudi et al. (2023). 25 Particularly the Saudi National Bank (~ 10%), Qatar Holding LLC (~ 5%) and Olayan Group (~ 5%). The nature of these investors as quasi-governmental and the relevance of the Middle Eastern jurisdictions where they are based are relevant. Ultra-high-net-worth (UHNW) clients from these jurisdictions are important for CS's 'crown jewel' private wealth management division. See, for reference, Credit Suisse (2023b). 26 There was significant trading in the CS AT1 bonds during the weekend of 17-19 March -these investors acquired the instruments fully aware of the situation of distress. See Vossos et al. (2023) (as reported in Wigglesworth (2023)).
into equity -and offering the bondholders shares in UBS -might have disrupted the current governance structure. 27

The Write-down of the AT1 Bonds
For sure, the bondholders are upset. But their chances of succeeding in litigation are slim because the contractual terms of the bonds 28 are quite clear. 29 They stipulate a write-down to zero following the occurrence of a 'Write-down Event' (Condition 7). This can be either of two types: a 'Contingency Event' or a 'Viability Event'. Under the former, the CET1 ratio (Common Equity Tier 1 Capital / Risk-Weighted Assets) of CS must fall below 7% at any reporting date. Under the latter, 'the Regulator' must determine that a write-down is essential to prevent CS from becoming insolvent or from ceasing to carry on its business (scenario 1); alternatively, CS must have received an 'irrevocable commitment of extraordinary support from the public sector' (scenario 2). The relatively ambiguous framing of the 'Viability Event' scenarios contrasted with the more clearly defined 'Contingency Event' which is based on a specific metric.
These triggers, especially the 'Contingency Event,' are designed to kick in before the bank faces insolvency and to restore the specified capital ratio. In other words, the function of the AT1 bonds is to absorb losses before an existential crisis wipes out the bank's equity. 30 FINMA claims that the write-down of the bonds was done on the basis of the extraordinary support received by CS 31 (a scenario 2 'Viability Event'), and this seems right. Under the bond terms, the complete write-down of the bonds is the automatic consequence of the irrevocable commitment of public support. The FINMA powers granted by the Emergency Ordinance to require the writedown of 'additional core capital' (see Sect. 2 supra) are superfluous, at least if such 27 To be clear, there was no provision for a conversion into equity in the contractual documentation (as it is in other AT1 bonds issued by other banks), which only contemplated a 100% principal write-down. Such a conversion could only have taken place as an exercise of the Swiss government's discretion in the design and implementation of the bailout. See details in Sect. 3 (The Write-down of the AT1 Bonds) below. 28 A detailed list of the bonds that were written down was published by FINMA (see FINMA (2023)). For the purposes of our discussion, we will refer to the terms and conditions of the $1.650 billion 9.750% Perpetual Tier 1 Contingent Write-down Capital Notes (ISIN US225401AX66/USH3698DDQ46), as per their 'Information Memorandum' of 16 June 2022 (https:// www. credit-suisse. com/ about-us/ en/ inves torrelat ions/ debt-inves tors/ bonds-secur ities/ capit al-instr uments. html?t= 925_0. 20920 91420 61675 73). These terms and conditions are typical of the AT1 bond issues of CS. 29 Even if the courts were to end up siding with the bondholders regarding their contractual claims (which we do not think should be the case), the actions by the Swiss authorities were probably lawful, based on the provisions of the Emergency Ordinance, and an appropriate exercise of their discretion in matters of fiscal policy (and financial regulation). 30 Typically, they would be considered 'going concern' when they rely on a 'high trigger' (such as a 7% capital requirement), or 'gone concern' when they are triggered at a very low capital requirement or at the 'point of no viability'. For a comprehensive empirical analysis of bank CoCo issues, see Avdjiev et al. (2020). 31 See FINMA (2023). a (contractual) commitment has already been made. 32 That the only 'loss absorption' mechanism provided for in the contractual terms of the CS AT1 Bonds was a complete write-down is a specific choice made by CS as issuer and was accepted by the bondholders and the regulator.
To understand the frustration of the bondholders, one must consider other terms of the AT1 bonds. A section on 'Events of Default' (Condition 12) stipulates that, inter alia, the commencement of an involuntary bankruptcy case against CS constitutes such an event. The consequences are set out in a section on 'Subordination of the Notes' (Condition 4). The Notes shall rank 'senior to the rights and claims of all holders of Junior Capital'. While this seems to suggest that the bondholders would rank before shareholders in a bankruptcy proceeding, the following paragraph qualifies this by providing that all bondholders' claims shall be subject to and superseded by a 'Write-down Event … irrespective of whether the relevant Write-down Event has occurred prior or after the occurrence of an Event of Default.' So, if there is no (superseding) 'Write-down Event,' the bondholders will rank before the bank's shareholders. But a superseding 'Write-down Event' can occur even in a bankruptcy proceeding. Could this create a legitimate expectation by the bondholders that their instruments would be treated at least no worse than equity under all circumstances? We do not think so. The AT1 bonds were held by sophisticated investors with access to top-tier advisers and the warnings of credit rating agencies 33 and central banks 34 who reviewed this type of instrument.
But there is no denying that the full write-down of the AT1 bonds is an unusual sanction if shareholders are not also wiped out. By contrast, Eurozone AT1 bonds often contain an equity conversion feature 35 which lets the bondholders rank at least pari passu with the shareholders (post-conversion) if a trigger event occurs. Alternatively, bonds may be of the (reversible) partial write-down type. 36 Further evidence for the 'unusual' character of the complete write-down is provided by statements from the European Central Bank 37 and the Bank of England. 38 Immediately after the bailout, both central banks rushed to declare that what happened with CS would not happen elsewhere in Europe. The goal was to calm investors, 39 assuming that this was an unexpected and unusual development that could create turbulence in the wider AT1 market and thus have an effect on financial stability. 33 As discussed by Avdjiev et al. (2013), particularly p 50 and footnote 11. 34 See Bank of England (2013). 35 Jiménez et al. (2018). 36 See Glasserman and Perotti (2018). 37 ECB, SRB, EBA (2023). 38 Bank of England (2023). 39 See, e.g., Asgari et al. (2023). 32 It appears that the Swiss government adopted a 'belt and suspenders' approach, adding another layer of protection to its bailout plan with the Emergency Ordinance. See, for example, Gollakota (2023) (Prasad Gollakota was Co-Head of the Global Capital Solutions Groups at UBS in the aftermath of the 2007-2008 financial crisis).
Unquestionably, the CS bailout could have been structured differently. Had no 'Write-down Event' within the meaning of Condition 7 of the bond terms occurred, the bondholders would not have been wiped out. A deal could have been structured without an 'irrevocable commitment of extraordinary support from the public sector' in the technical sense of a 'Viability Event' (scenario 2) of the bond terms. A complete write-down of the AT1 bonds could have been avoided. But that is very different from saying that it had to be avoided. In the absence of principles or guidelines applicable to government bailouts which require that shareholders suffer first losses before creditors have to take cuts, there is no basis for complaint on the part of bondholders.

Writing AT1 Bond Terms Differently
The simplest way to achieve a different result would be to include different terms in AT1 bonds. And there are good reasons for writing AT1 bond contracts such that bondholders fare no worse than shareholders in a bailout, resolution, or bankruptcy. Somewhat ambiguous contractual language, as in the CS case, encourages different expectations as to what will happen in a crisis. Against the background of these ambiguities and the international prevalence of the 'absolute priority rule' or at least a pari passu regime, the bondholders probably hoped that no (complete) write-down would occur, and the price of the bonds apparently reflected this. 40 CS 'traded' on this ambiguity to minimise its financing costs, and regulators such as FINMA relied (and continue to rely) on it to strengthen banks' financial position and financial stability. 41 Resolving the ambiguity is important to reduce unnecessary transaction costs, as evidenced by the litigation 42 triggered by the current 'Write-down Event'. It is also important to eliminate the perverse incentive for the bondholders to accelerate the demise of the issuer by filing for insolvency in an attempt to preserve their presumed rank above shareholders. Moving to (at least) a (post-conversion) 43 pari passu regime could avoid the need to significantly increase the price for AT1 bonds with a clear 'total write-down' feature. The pricing of these bonds would continue to be based on the principle that debt should not rank lower than equity -a cardinal principle and focal point 44 of bankruptcy laws world-wide. If we accept that the market was not appropriately pricing the AT1 instruments beforehand, 45 it is clear that it 40 See Asgari et al. (2023). 41 Levine (2023). 42 See Wiegmann (2023). 43 By this we mean that, as bonds get converted into equity, they rank pari passu to old equity after the conversion, but the rate at which bonds get converted may mean that old equity is in fact diluted (thus departing from a 'strict' pari passu). 44 Schelling (1960). is now widely aware of the risks. 46 If the yield on these instruments increased significantly to reflect this, the cost of funds for financial institutions in Switzerland (or institutions that are perceived to operate in a similar framework) would increase, 47 potentially creating additional hazards for (international) financial stability. Finally, CoCos which convert to equity rather than being written down have been shown to offer a superior design from the point of view of reducing bank fragility. 48

The Need for Principles for Bailouts
The implications of the CS bailout go beyond the (future) treatment of the AT1 bondholders. It is remarkable that the Swiss government resorted to an ad hoc bailout outside the applicable framework for bank resolution to restructure CS. Similar ad hoc bailouts of 'critical' non-financial firms 49 have occurred and will continue to occur in the aftermath of geopolitical or macroeconomic shocks such as the COVID-19 pandemic 50 or the (European) energy crisis following the war in Ukraine. 51 Arguably, the applicable bankruptcy laws suffer from certain structural limitations which can make an ad hoc bailout the better restructuring tool under certain circumstances. 52 Bankruptcy and bank resolution are heavily regulated. In contrast, ad hoc bailouts are not regulated at all. The potential for abuse (of taxpayer money) and arbitrary decision-making is significant. When 'critical' firms face existential difficulties, governments resort to whatever measures are available in the moment to overcome the immediate problem and prevent a perceived catastrophe. What should concern us is the inappropriate use by governments of their fiscal capacity to pursue bailouts that are not in the best interest of society -because they are unjustifiable on a cost/ benefit basis, or because they have unjustifiable distributive consequences.
Of course, some flexibility to react to these difficult situations on a case-by-case basis is important. At the same time, establishing some regulatory guideposts could assist regulators and governments to act consistently and based on principles in a bailout scenario. For example, the principle that shareholders should bear losses first is a cornerstone of corporate bankruptcy laws and policies worldwide, as elaborated in the first section of this article. International institutions such as UNCITRAL or 46 Globally, the yield on AT1s reached 10.2% on 20 March after hovering in the 7 to 8% range until the start of the month, according to the benchmark ICE BofA Contingent Capital Index, see Shinozaki (2023). It is also telling that on 3 April Mitsubishi UFJ decided to postpone an issue of AT1 bonds (see Kato (2023) Bolton and Kartasheva (2023), with reference to Avdjiev et al. (2020). 49 For the concept of a 'critical' firm, see Eidenmüller and Paz Valbuena (2021), especially pp 520-521. 50 Eidenmüller and Paz Valbuena (2021). See, for example, the case of Eurostar (Keohane and Picard (2021)). 51 Eidenmüller (2023). See, for example, the case of Uniper (Storbeck et al. 2022).
UNIDROIT should begin work on 'Principles on Ad Hoc Bailouts of Critical Firms' that would fill the current regulatory void and provide some certainty for all affected parties. This would increase both the efficiency and legitimacy of such bailouts.