A test of resolve: credible resolution following the 2023 banking turmoil
Speech by Mr Agustín Carstens, General Manager of the BIS, at the FSB-FSI-IADI Resolution conference, Basel, 29 October 2024.
Introduction
I would like to welcome you all to the Resolution Conference 2024, the first that has been co-organised by the BIS Financial Stability Institute (FSI), the Financial Stability Board (FSB) and the International Association of Deposit Insurers (IADI). This event is motivated by the banking turmoil in March 2023. The 18 months that have passed since those events have given time to reflect seriously on it and derive some lessons. This conference provides an opportunity to take stock, compare notes and try to identify a productive way forward.
Scene-setting
It is now commonplace to say that the March 2023 failures of several US regional banks, followed a week later by the near failure of Credit Suisse, were the first meaningful test of the international resolution framework that was put in place following the Great Financial Crisis (GFC).
The headline message is that large bank failures did not lead to a systemic crisis. Authorities managed them in an orderly manner with no ultimate loss to public funds. Creditors and shareholders bore losses. In the case of Credit Suisse, there was a significant writedown of loss-absorbing instruments. This is a noteworthy achievement, and stands in stark contrast to the GFC.
The extensive work to put in place cross-border cooperative arrangements has demonstrably strengthened the financial system. The outcomes might have been very different without the planning and coordination that took place between home and host authorities, and the understanding and trust that have been developed.
However, work remains to be done. The reports published last year by the FSB and IADI set out lessons learned for resolution and deposit insurance.1 They include the risk of faster failures, accelerated by digital technologies; the scope of resolution planning and requirements for loss-absorbing capacity (LAC); and flexibility in resolution strategies. Other reports, including by the Basel Committee on Banking Supervision, elaborate on the supervisory shortcomings and the vulnerabilities arising from large quantities of uninsured deposits. Work on all these issues is ongoing.
In any case, I would like to concentrate my remarks on two elements of the bank resolution framework that I think must be tackled as we go forward. The first is the power to bail-in creditors as a key element of resolution strategies. The second is the need to put in place effective facilities for providing liquidity in resolution. The events of March 2023 highlighted the importance of both. They are also among the elements of a resolution framework that are most challenging to implement.
The credibility of bail-in
Bail-in powers are core to the resolution framework adopted after the GFC. Bail-in allows a systemically important bank to be recapitalised without the need to find a buyer for its business or to split up its operations, at least in the short term. Appropriate liabilities absorb losses without putting a failing bank into insolvency. Crucially, it is designed to ensure that a bank's owners and investors, rather than depositors or taxpayers, bear the costs of resolution costs.
In practice, a bail-in is a highly complex transaction involving multiple parties, and a huge amount of work has been carried out on how to execute it. A typical bail-in would involve multiple valuations; a mechanism to write down and cancel instruments, which are likely to be traded; and the issuance of new shares to the bailed-in debt holders. The process has been mapped out in detail by resolution authorities. However, a full bail-in strategy remains untested.
Credit Suisse had a resolution strategy based on bail-in, and FINMA and key host authorities had prepared extensively to execute that strategy.
In the end, the Swiss authorities chose not to follow the resolution playbook because they had another option that achieved their objectives: a state-brokered commercial merger of Credit Suisse and UBS. Nevertheless, the contractual writedown of all the outstanding Additional Tier 1 (AT1) capital instruments issued by Credit Suisse was a key element of the transaction. The writedown extinguished liabilities amounting to CHF 16 billion from the bank's balance sheet.2
Although the writedown was more limited than that planned under the full bail-in strategy for Credit Suisse, it demonstrates that bail-in is a core instrument in the crisis management framework. Contrary to what some commentators have feared, a substantial debt writedown is possible and can be executed without significant systemic disruption.
Nevertheless, there are a few lessons to draw from this to reinforce that bail-in is credible and feasible.
Flexible resolution toolkits
First, authorities need flexibility. Planning is essential, but it cannot be prescriptive. We cannot know with absolute confidence in advance how a failure will happen and what actions will best safeguard financial stability. Accordingly, authorities need options so that they can shape their response to the circumstances of a failure. This implies a toolkit approach under which authorities can combine the use of different tools.
The Credit Suisse transaction demonstrated that, even in the case of a global systemically important bank (G-SIB), bail-in may not be an exclusive strategy, but debt writedown could be a core element. Moreover, bail-in is not a tool exclusively for G-SIBs. For other banks, the writedown of liabilities in resolution can finance transfers of business and reduce the demands on industry-funded sources such as deposit insurance funds.
Flexibility of this kind brings operational complexities. A toolbox approach means that authorities and firms need to accommodate different options in resolution planning. Banks will need the systems and capabilities to support those options. Key aspects of resolvability such as structure and LAC may become even more complex. However, an effective toolbox approach will further reduce the residual risk that public funds will be needed in crisis management.
Loss-absorbing capacity
Second, for bail-in to be credible banks must have liabilities that can be written down with legal certainty and without systemic impact. The FSB's TLAC standard ensures this for G-SIBs. Some jurisdictions have extended similar requirements for LAC to other banks that could be systemic in failure.
For example, the EU requirement for resolution-related LAC, the minimum requirement for own funds and eligible liabilities (MREL) – applies to all banks. The amount above the regulatory minimum required for individual banks is based on their resolution strategy. It aims to ensure that any bank that is expected to be resolved rather than wound up maintains LAC in sufficient quantity and quality to absorb losses and recapitalise it in resolution.
The US financial regulators have consulted on a proposal to require banks with $100 billion or more in assets to maintain a layer of long-term debt. This additional LAC would be used, in the event of a bank's failure, to absorb losses and increase the resolution options. It should also foster depositor confidence among uninsured deposits.
The three US regional banks that failed in 2023 had little or no outstanding long-term debt. It has been observed elsewhere that if the proposed requirement had applied to Silicon Valley Bank and Signature Bank, they might have been resolved within the FDIC's normal funding constraints, without a systemic risk exception being required.3
If bail-in is to help fund resolution transfers, there need to be instruments that can be written down. The amounts are lower than that needed to recapitalise the bank and finance restructuring in a "pure" bail-in. Nevertheless, calibrating those requirements may be challenging.
Moreover, meeting LAC requirements should not put banks' legitimate business models in jeopardy. This is particularly relevant for banks that are predominantly deposit funded. A pragmatic way to alleviate the challenges for those banks is to take account of the resolution funding available from external sources, such as deposit insurance or resolution funds, when setting LAC requirements.4
Liquidity for crisis management
Let me turn now to liquidity for crisis management. Resolution powers can recapitalise a failing bank through bail-in. However, capital is not enough on its own. Without liquidity, the resolution will fail.
Market funding will almost certainly not be available to a bank following its resolution until counterparty confidence can be restored. Resolution frameworks therefore require a credible source of liquidity, at the necessary scale and for a sufficient period of time to allow a resolved firm to return to market-based funding.
This is recognised by the FSB, which has published two sets of guidance on funding in resolution. However, the arrangements in place vary considerably across jurisdictions and in many cases are not designed for the resolution needs of systemically important banks.
The liquidity arrangements that were needed in the case of Credit Suisse support this point. The Swiss government had been working on a public liquidity backstop, but this was not yet in place in March 2023. Accordingly, the authorities had to adopt emergency legislation to enable the Swiss National Bank (SNB) to provide a liquidity facility of up to CHF 200 billion. Part of that lending was uncollateralised and coupled with a privileged bankruptcy status for the SNB and part was backed by a guarantee from the Swiss state.
This case illustrates that ordinary central bank lending arrangements, including emergency liquidity assistance, may not be sufficient for resolution. The amount of liquidity needed by a systemically important bank will be considerable and required over an extended period. Moreover, lending may need to be secured against a wider range of assets or, in extreme circumstances, be uncollateralised. Arrangements for resolution funding must meet these needs. This implies a fiscal backstop to increase the firepower where that is needed.
A fiscal backstop might appear to introduce a risk to public funds, something that the framework for ending "too big to fail" was designed to avoid. But the risks of loss to public funds should be low. It's worth noting that all lending in relation to Credit Suisse was repaid, and no losses were incurred by the SNB or the Swiss state under its indemnity. If resolution is effective, the bank will be viable and the borrowing should be repaid.
Concluding remarks
I will end where I began. Financial crises provide a good opportunity to identify flaws or shortcomings in the policy framework. The March 2023 banking turmoil was the most significant banking crisis since the GFC and the subsequent policy reforms. Therefore, we should grasp this opportunity to draw lessons.
Overall, authorities managed to preserve financial stability. In Switzerland, that was accomplished, despite the failure of a G-SIB, without any cost to the taxpayer. This was a remarkable achievement, and the resolution framework developed after the GFC contributed to that.
But we also need to take note of the obstacles encountered in the process. In particular, it is clear that maximising the potential of bail-in and the provision of liquidity in resolution are pending tasks that need to be addressed.
Work to do that is ongoing, and this conference is a small but significant part of that process. I am delighted that so many people have come to Basel to participate, and I expect productive discussions during the day.
1 FSB, 2023 Bank Failures – Preliminary lessons learnt for resolution, October 2023 and IADI, The 2023 banking turmoil and deposit insurance systems: potential implications and emerging policy issues, December 2023.
2 See R Coelho, J Taneja and R Vrbaski, "Upside down: when AT1 instruments absorb losses before equity", FSI Briefs, no 21, September 2023.
3 See M Gruenberg, "Remarks by Martin J Gruenberg, Chairman of the FDIC, on the resolution of large regional banks – Lessons Learned", August 2023 and FSB, op cit.
4 See F Restoy, "MREL for sale of business strategies", FSI Briefs, no 20, September 2023.