Post-turmoil bank failure management: some reflections

Remarks by Mr Fernando Restoy, Chair, Financial Stability Institute, at the second edition of the Global Conference on Financial Resilience, organised by the College of Supervisors of RBI, Mumbai, 21 June 2024.

BIS, FSI speech  | 
21 June 2024

1. Introduction 

Let me first thank Rabi Mishra and his team at the RBI College of Supervisors for the kind invitation to participate in this Global Conference on Financial Resilience.1

In my intervention, I plan to share with you some reflections on bank crisis management inspired by last year's banking turmoil. As you all know, one of the most important policy reforms that emerged from the Great Financial Crisis (GFC) was the creation of a new bank resolution framework. Under the slogan "avoid the perception of too-big-to-fail banks", the Financial Stability Board established new standards aimed at reducing the impact of systemic bank failures.

The FSB's Key Attributes of Effective Resolution Regimes for Financial Institutions (KAs) contain the main elements of the new framework. The KAs aim to facilitate orderly resolution of systemic entities without exposing public funds to losses. A key component of the new resolution regime is the bail-in tool that would allow resolution authorities to write down liabilities or convert them into equity in order to absorb losses and, in some cases, recapitalise a firm in resolution.

In some jurisdictions that have applied the KAs, a new resolution framework for systemic banks co-exists with domestic failure management regimes which are applied to the entities which do not meet the criteria to be subject to the new procedures. In the European banking union, a common resolution framework for significant banks runs in parallel to national insolvency procedures for less systemic institutions. In the US, the new resolution legislation (contained in Title II of the Dodd-Frank Act) is only applied to large, complex bank holding companies categorised as systemically important financial institutions (SIFIs). The new arrangements complement the existing regime, contained in the Federal Deposit Insurance (FDI) Act, under which banks are subject to an administrative insolvency regime managed by the Federal Deposit Insurance Corporation (FDIC). Under this regime, the FDIC typically manages banks' failures through purchase and assumption (P&A) transactions, with possible support from the Deposit Insurance Fund (DIF). That support is subject to strict restrictions which can only be waived in exceptional circumstances (the so-called systemic exception).  

During the 2023 bank turmoil, crisis management frameworks in both the United States and Switzerland were directly tested. In the US, the failure of two regional banks, Silicon Valley Bank (SVB) and Signature Bank (SB), required the use of a systemic exception as authorities felt that preserving financial stability justified waiving the restrictions on FDIC support in order to protect all of those banks' deposits. Moreover, a special liquidity facility was established by the Federal Reserve to ease potential system-wide funding pressures.

In Switzerland, the crisis of Credit Suisse (CS), a global systemically important bank (G-SIB), was not managed under the new resolution framework but rather through a series of ad hoc measures taken to facilitate the absorption of CS by UBS without formally declaring CS to be a failing institution. Moreover, although the measures adopted outside resolution included a substantial bail-in of some creditors, they also entailed the provision of public guarantees to support the liquidity and solvency of the resulting institution.

Arguably, the actions taken by authorities met the primary objective of preserving financial stability. At the same time, those actions did not follow the usual procedures and, contrary to the objectives of the post-crisis reforms, required different forms of external support.

Recent events therefore provide good motivation for a general reflection on possible gaps and flaws in the prevailing bank failure management frameworks in different jurisdictions (Carstens (2023)). My remarks will discuss some of the issues that the recent turmoil and other significant bank failures have raised in relation to the current policy framework for bank crisis management.2 

2. The management of recent bank failures

Let's start with a quick overview of recent bank failures. In the US, actions taken by authorities upon the failure of SVB and SB were very much influenced by intense bank runs and the signs that contagion was affecting regional banks. This prompted the FDIC and the Federal Reserve Board to propose to the US Treasury that it activate a systemic risk exception, allowing the FDIC to extend its protection to all deposits of those two banks. Despite this, another bank, First Republic Bank (FRB), subsequently failed.

The approach followed by the FDIC to resolve SVB and SB consisted of a transfer strategy entailing the creation of a bridge bank for each institution, composed of all their deposits and most of their assets, and the subsequent sale of those bridge banks' assets and liabilities to suitable acquirers. The bank acquiring SVB's assets and liabilities3 benefited from a loss-sharing agreement with the FDIC.

In the case of CS, the strategy followed by the Swiss authorities was quite unique. Rather than triggering statutory resolution, a commercial transaction was orchestrated under which UBS would take full control of CS. That transaction was supported by a number of ad hoc actions by the authorities outside a formal resolution procedure. These included: (i) the adoption of emergency legislation that allowed the Swiss National Bank (SNB) to provide privileged liquidity facilities4 and the Swiss government to waive the obligation under Swiss law for the merger operation to be endorsed by the shareholders of both institutions; (ii) the provision of a public second-loss guarantee to UBS for certain CS assets;5 (iii) a partial write-off of equity that preserved a residual value of CHF 3 billion; and (iv) the activation of contractual clauses that allowed AT1 instruments (around CHF 16 billion) to be written off once the bank received public support;

The issues raised by these resolution cases are quite different. In the case of the failure of US banks, the measures authorities took did not require the adoption of new rules, procedures or any sort of emergency legislation. Indeed, all actions were performed in accordance with the FDI Act, including the invocation of a systemic risk exception. The only innovation was the Fed's adoption of a new and highly flexible liquidity facility to help contain the threat of additional bank runs and the pressure on weak banks to (fire) sell their assets.6

The case of CS is much more complex. Like all G-SIBs, the bank already had in place a detailed resolution plan that entailed the application, in this case, of a resolution strategy that would allow the bank to continue performing its critical functions once resolved and recapitalised through creditors' bail-in. This strategy was thoroughly prepared by the bank's crisis management group, composed of authorities in those jurisdictions where CS had material subsidiaries.

However, despite the preparations to execute the resolution plan, authorities chose to follow a different route outside resolution. Swiss authorities felt that a deep restructuring involving massive bail-in would have been risky and that its failure could have led to the liquidation of the bank's domestic business in accordance with regular insolvency procedures. Therefore, Swiss authorities opted for a strategy based on a commercial transaction – outside the resolution framework – that was deemed, over the critical weekend, to be less disruptive than the one developed in the resolution plan.

Arguably, rather than relying on a commercial transaction supported by contractual bail-in and public support outside resolution, in principle Swiss resolution authorities could have implemented a largely similar bank failure management approach by employing their statutory powers, upon declaring that the bank met the conditions for resolution.

It is true, though, that resolution involving creditors' bail-in would have been incompatible with preserving any residual value for equity holders, and it would not have allowed UBS shareholders' rights to approve the acquisition of CS to be waived. Moreover, the provision of liquidity by SNB after resolution would still have required the ad hoc provision of public guarantees, as Switzerland – like many other jurisdictions – had not yet established a specific regime to provide funding for banks in resolution.

In any case, the policy discussion following the CS failure should not focus just on the pros and cons of adopting specific measures, but should also consider why the failure was managed through the activation of contractual clauses and emergency legislation supporting a commercial transaction rather than through the use of existing resolution statutory powers. That discussion could help inform the debate about what aspects of the new resolution framework might need to be revised.

Let me now cover some of those features that may be worth revisiting in the light of recent experience. 

3.  Some issues stemming from the recent turmoil

Resolution planning

First, the speed with which apparently solvent banks became failing banks, particularly in the US, points to the need to strengthen resolution planning (FDIC (2023a)). This should first be achieved by enlarging the scope of application of meaningful resolution planning obligations to all banks that can be systemic in failure – something that is not yet the case in all jurisdictions, notably the US.

In addition, resolution plans for international banks should address practical issues related to the making resolution actions – particularly bail-in – operational in a cross-border context. Given that securities qualifying as total loss-absorbing capacity (TLAC) are typically issued in international financial centres, it is important that resolution decisions – such as a conversion of debt securities into equity – be effective in all relevant jurisdictions.

Resolution plans should contemplate different options and not focus on just a single resolution strategy (FSB (2023a)). As the case of CS shows, the preparatory work conducted around the development of the entity's resolution plan proved very useful for managing the failure of the bank, even if the plan was not ultimately implemented. Yet the process would have been facilitated if, in addition to contemplating a massive bail-in, the plan had included provisions for a possible full or partial sale of business.

Loss absorbency

One of the main ingredients of the new resolution framework – and the new resolution planning and resolvability requirements – that emerged from the crisis is the availability of sufficient resources within systemic banks' balance sheets to absorb losses and, if needed, recapitalise the institution after resolution is triggered. In particular, the FSB has issued standards for TLAC that all G-SIBs should satisfy. The standards establish a minimum amount of bail-in-able liabilities that would be written off or converted into equity in resolution. Eligible liabilities could be both equity and debt securities that satisfy certain conditions.

In jurisdictions where the new resolution framework is being applied beyond G-SIBs (like the European Union (EU)), there is a version of the TLAC standard that is also binding for less systemic institutions: the minimum requirements for eligible liabilities (MREL). In other jurisdictions, such as the US, no TLAC-type requirement is applied for non-GSIBs. Therefore, most US banks – including those failing in the recent turmoil – had no specific obligation to hold liabilities that could absorb losses in resolution beyond the capital requirements established in prudential regulation.

However, a recent proposal by the FDIC (Gruenberg (2023) and FDIC (2023b)) would require banks with more than USD 100 billion in assets to satisfy minimum long-term debt requirements. The counterpart of those debt instruments on the asset side could be transferred to the acquirer, but the debt instruments themselves would be left in the residual entity to be liquidated. This would make those debt instruments act as gone-concern capital supporting the transfer transaction.

MREL obligations in the EU are substantially larger on average than the long-term debt requirements now considered in the US.7  However, while the proposed US requirements can only be met with debt, MREL targets in the EU can be met with a variety of eligible liabilities that include equity, debt and even some non-covered deposits. In reality, many small and mid-sized institutions in the EU cover a large part of their MREL requirements with equity instruments.8  This is probably due to the fact that, given those banks' lack of experience and specific business model, it is difficult for them to tap regulated debt markets.

From a conceptual point of view, there is merit in, at least, limiting the eligibility of equity to satisfy gone-concern capital requirements. Experience shows that, unlike long-term debt, equity instruments tend to disappear quite quickly as a bank approaches the point of non-viability and during the resolution process itself as hidden losses emerge in the balance sheets.9  Therefore, equity, being the most powerful loss-absorbing instrument in going-concern, might simply not be available in gone-concern.

How to calibrate loss absorbency requirements to facilitate transfer strategies

Appropriately calibrated loss-absorbing requirements could facilitate the success of resolution strategies involving the transfer of assets and liabilities of failing banks to other institutions. Recent, and not so recent, experience shows that this resolution tool can be very effective for managing the failure of different types of banks.

The success of a transfer strategy consisting of the assumption by a suitable entity of a failing bank's sensitive liabilities (such as deposits) requires that the acquirer be offered sufficient compensation for taking over the new obligations. This compensation normally takes the form of a transfer of assets from the failing bank to the acquirer and some external support either in the form of cash or different types of loan loss guarantees.

The ability of asset transfers to compensate acquirers depends on the difference between the value of unencumbered assets and that of the transferred liabilities. External support is normally provided by the DIF, subject to a financial cap (or least cost constraint) determined by the cost (net of recoveries) that the DIF would have assumed if it had to pay out covered deposits if the bank were liquidated. Naturally, the more senior the DIF claim in insolvency, the tighter the least cost constraint. In that regard, jurisdictions where DIF claims are "super-protected" and rank above other deposits (such as the EU) implicitly impose a particularly tight financial cap for DIF support. By contrast, in jurisdictions where both DIF claims and non-covered deposits rank pari passu (such as the US), there is more room for the DIF to facilitate transfer transactions through funding.

Recent evidence (Restoy (2023)) confirms that gone-concern capital requirements can facilitate transfer transactions. However, the feasibility of that resolution strategy would also require appropriate constraints on the proportion of banks' non-covered deposits and, crucially, that the available support from the DIF is not overly restricted. The latter condition would not normally hold if DIF support is subject to a rigid least cost constraint and the DIF enjoys an excessively privileged status in insolvency.

Public support

Finally, a word on public support. The foundational principles of the new resolution framework developed after the GFC included the objective to minimise the cost of bank failure management actions for taxpayers. However, experience – including the recent bank turmoil – shows that there are instances in which some form of external support is required to preserve financial stability and the continuity of the systemically critical functions of failing banks. Thus, it cannot be guaranteed that the failure of any bank can be managed in all circumstances by just transferring deposits and other sensitive liabilities to an acquirer or maintaining the critical operations of the bank by converting liabilities into equity. In the first case, external support may be needed if the available assets are not sufficient to convince any suitable acquirer to assume the sensitive liabilities. Likewise, when the failing bank's business model cannot easily accommodate the issuance of large amounts of bail-in-able liabilities and there are no suitable acquirers, external support might be required, at least temporarily, in order to provide the bank with the resources required to continue conducting its critical operations.

Regular support for resolution actions is often provided by the DIF. As discussed earlier, that support is normally capped by a least cost restriction that prohibits the DIF from committing funds exceeding the expected cost (net of recoveries) of paying out covered deposits if the bank were liquidated (Costa et al (2022)). Additional support aimed at protecting public interest could be provided directly by the national treasury or by dedicated funds contributed by the industry. In the US, extraordinary support for failing large systemic institutions can be provided by an orderly liquidation fund, as provided for in Title II of the Dodd-Frank Act. Moreover, under the FDI Act, the least cost restriction for FDIC support can be waived if a systemic risk exception is applied. In both cases, extraordinary external support can only be authorised through a special procedure requiring the endorsement of the regulatory agencies and the Treasury after consulting the US president.

A completely different model is in place in the EU, where external support can be provided by the Single Resolution Fund (SRF), built up with contributions from the industry. However, the conditions for access and the available amounts are highly restrictive.[10]  Moreover, beyond the SRF, the possibility of the state directly supporting resolution is almost non-existent. Since national insolvency regimes are less restrictive and allow for the provision of public liquidation aid, the failure of some European banks that could have systemic implications was in fact managed through national insolvency procedures, thereby effectively reducing the scope of application of the common resolution framework.

Recent developments show that minimising public support should remain a key objective. However, there should be no ambition to establish a resolution framework that can eliminate any possible need to use external funds to support the orderly resolution of any systemic bank. In fact, the FSB Key Attributes do recognise that, even where resolution regimes are in place, there may be (extreme) circumstances where public funds are needed. Therefore, there could be merit in envisaging a robust regime that would ensure that sufficient extraordinary external support is available in extreme circumstances where it is required to preserve systemic financial stability. That regime should contain sufficiently clear and restrictive – but not overly punitive – conditions for access, rigorous approval procedures that would ensure the preservation of the public interest and reliable processes for recovering the costs incurred.

A specific situation in which some sort of public support would normally be required is the provision of liquidity in resolution. Once a bank has been resolved, there is no guarantee that it will immediately recover the trust of its clients and other fund providers. Therefore, an effective funding-in-resolution facility needs to be put in place, backed by some sort of public indemnity that would allow a bank in resolution to obtain funding from the central bank even when it does not hold all the required collateral.

4. Conclusions 

Let me conclude. In line with the famous Churchillian statement "Never let a good crisis go to waste", the international community has traditionally reacted to different banking crises, albeit with various degrees of ambition, by adopting reforms to the relevant regulatory framework. The spring 2023 banking turmoil should be no exception. Although financial stability was ultimately preserved, the events unveiled deficiencies in the bank failure management regime, despite the far-reaching improvements adopted after the GFC.

In this presentation I have covered several possible reforms of bank failure management regimes. In general, adjustments to the current setup should aim to satisfy two basic objectives: first, improving the resolution framework and resolution tools to make them more effective and therefore reduce the need to provide government support to failing banks in order to preserve financial stability; and, second, embedding sufficient flexibility and pragmatism in the arrangements as regards the use of different tools and the availability of external funds.

In particular, there are strong reasons to extend resolution planning obligations to all banks whose failure could have adverse effects on the financial system. Crucially, resolution plans should include well defined requirements for a minimum amount of loss-absorbing liabilities in resolution. Those requirements should be calibrated to directly support the feasibility of the envisaged resolution strategy and ideally be composed primarily of debt instruments rather than equity, as the latter might well largely disappear before resolution is triggered.

In addition, as there is no way to foresee all the possible conditions that might occur in a resolution weekend and affect the feasibility of resolution measures, planned resolution strategies should be more an array of options for the deployment of different tools than a rigid playbook. Very importantly, experience shows that it is wise to put in place well defined procedures for delivering extraordinary external support in extreme circumstances. 

Many thanks. 


Acharya, A, E Carletti, F Restoy and X Vives (2024): "Banking turmoil and regulatory reform", IESE Banking Initiative and CEPR, June.

Basel Committee on Banking Supervision (BCBS) (2023): Report on the 2023 banking turmoil, October.

Carstens, A (2023): Some lessons for crisis management from recent bank failures, speech at the High-level meeting on banking supervision of the Association of Supervisors of Banks of the Americas, the Basel Committee on Banking Supervision and the BIS Financial Stability Institute, Panama City, 19 October.

Costa, N, B Van Roosebeke, R Vrbaski and R Walters (2022): "Counting the cost of payout: constraints for deposit insurers in funding bank failure management", FSI Insights on policy implementation, no 45, July.

Expert Group on Banking Stability (2023): Report of the group of experts on banking stability, Swiss Federal Council, September.

Federal Deposit Insurance Corporation (FDIC) (2023a): Options for deposit insurance reform, May.

--- (2023b): Fact sheet on proposed rule to require large banks to maintain long-term debt to improve financial stability and resolution, August.

Financial Stability Board (FSB) (2023a): 2023 bank failures: preliminary lessons learnt for resolution, October.

--- (2023b): 2023 Resolution Report: Applying lessons learnt, December.

Restoy, F (2023): "MREL for sale-of-business resolution strategies", FSI Briefs, no 20, September.

Single Resolution Board (SRB) (2023): "SRB MREL Dashboard Q2.2023", November.

1 This speech is based on my contribution to Acharya et al (2024). I am grateful to the very insightful comments of Rodrigo Coelho, Elke König, Zack Thor, Rastko Vrbaski and Ruth Walters.

2 Thus, it complements the work already conducted, or which is still ongoing, by different national and international organisations. See eg BCBS (2023), FSB (2023a,b) and Expert Group on Banking Stability (2023).

3 The UK subsidiary of SVB was resolved by the Bank of England following a resolution procedure under which the bank's shares were transferred to an acquirer after writing down all equity, AT1 and Tier 2 capital instruments.

4 These included a liquidity facility from the SNB of up to CHF 100 billion with no collateral requirement (but granting SNB enhanced creditor status) and another facility of CHF 100 billion guaranteed by the state.

5 The guarantee would cover losses up to CHF 9 billion for certain assets after the first CHF 5 billion in losses to be assumed by UBS. The federal loss protection guarantee agreement was terminated by UBS on 11 August 2023.

6 The Bank Term Funding Program offered loans of up to one year to lenders pledging collateral including US Treasuries and other "qualifying assets", which were valued at par.

7 According to SRB (2023), the average recapitalisation amount for banks following a resolution strategy based on transfer transactions is around 10% of risk-weighted assets, ie 4 percentage points above the long-term debt requirement stipulated in the US.

8 SRB (2023) shows that for (significant) banks under the SRB remit classified as "non-Pillar 1", equity instruments represent on average more than 60% of the resources used to meet MREL requirements.  

9 In fact, this is partially recognised in the FSB TLAC Term Sheet, as it contains an expectation that at least 33% of the TLAC requirements will be met with debt securities.

10 Access to the SRF for banks under resolution should be preceded by a bail-in of at least 8% of total liabilities of the institution in resolution. Moreover, SRF support cannot exceed 5% of that institution's total liabilities.