Source: Bank for International Settlements
Introduction1
Good morning, and many thanks for inviting me to speak to you today. The last time I addressed the EFDI International Conference was two years ago in Budapest. This time you are meeting in my home city, and I hope that you all have an opportunity to enjoy it.
In 2023, the conference took place shortly after the banking turmoil in March of that year. That episode focused international attention on structural risks posed by high volumes of uninsured deposits and the need to review prevailing assumptions about their “stickiness” in an environment where technology allows rapid movement of funds and social media can vastly amplify rumours and, potentially, misinformation. The 2023 failures set an agenda for international work that is still being pursued. Moreover, weeks before that 2023 conference, the European Commission had released its legislative proposal on reforms to the EU crisis management and deposit insurance framework (CMDI). Two years later, this is still in the legislative process.
So, in a way, not much has changed in the intervening two years. But, arguably, we are now in a riskier environment. The global landscape is changing. Geopolitical risks are more pronounced than they have been in at least a generation. Technological innovation is moving faster than governments and regulators can keep pace, and the opportunities and risks it will bring remain in flux.
The theme of this conference is how to lay the solid ground that will allow us to embrace the future – to take advantage of its opportunities and contain its risks. This is a wide, almost daunting question. What’s more, we need to confront it against a backdrop of rising calls to reduce the costs of regulatory compliance. In some cases, the demand is to directly cut regulation. In others, it is framed in terms of simplification. The debate is obviously welcome. It is an obligation for regulators to avoid, to the greatest extent possible, imposing disproportionate costs on the industry. But by the same token, any alleviation of regulatory obligations should be fully compatible with the preservation of key social objectives, including financial stability.
I am not going to talk about this bigger picture today. Rather, I wish to focus on one dimension that is of particular relevance to this European audience of deposit insurers and financial crisis managers. That is the question of funding for bank failure management. Here, too, there are calls on us to revisit our standards and alleviate disproportionate burdens. It is claimed that the Minimum Requirement for Own Funds and Eligible Liabilities (MREL) framework in the banking union requires banks to hold materially more loss-absorbing capacity (LAC) than comparable US banks. It is argued that this both undermines the competitiveness of European banks and affects their ability to fund the EU economy and function as an engine for growth.
I will come back to this comparison shortly, but first let’s go back to basics.
The background
Following the 2008 financial crisis, the international community put in place a resolution framework designed to provide credible alternatives to bailing out failing firms. Of course, this includes resolution powers and tools, an institutional architecture and cooperative arrangements. But LAC requirements are core to the effectiveness of the framework. LAC allows the costs of resolution to be internalised so that losses are absorbed by shareholders and creditors, as they would be in liquidation, and a failed bank’s critical functions can continue to operate without resort to public funding.
The Financial Stability Board’s Total Loss-absorbing Capacity(TLAC) Standard therefore supplements the resolution framework and aims to ensure that the world’s most systemically important banks – the G-SIBs – hold sufficient LAC that will be readily available in the event of their resolution and allow authorities to execute their resolution strategy. Expressed as a comprehensive requirement that includes regulatory capital, the FSB standard requires G-SIBs to maintain TLAC of at least 18% risk-weighted assets or 6.75% of total leverage (whichever is higher).
The decision to limit TLAC requirements to G-SIBs was based on their cross-border nature and the fact that they are sufficiently comparable to support an internationally coordinated, one-size-fits-all approach. However, these are not the only banks that would be systemic in failure. The US regional bank failures of 2023 highlighted that systemic risks can flow from the failure of banks that are significantly smaller and less international than G-SIBs.
In the case of Silicon Valley Bank (SVB) and Signature Bank, those risks were considered sufficiently serious to justify invoking the systemic risk exception that allowed the Federal Deposit Insurance Corporation (FDIC) to execute a more costly resolution strategy than would otherwise have been permitted and thereby protect the uninsured deposits. Those cases also illustrated the importance of adequate LAC for banks other than G-SIBs. Under the current US framework, only G-SIBs are required to hold liabilities designed to absorb losses in resolution, beyond prudential capital requirements. In the wake of those failures, it has been observed that the depositor runs and costs of failure management might have been less severe, and the FDIC might have had more options, if SVB and Signature Bank had had additional LAC which could have protected uninsured deposits from loss.2 This is an important lesson.
The arrangements for LAC in the EU are more stringent than in the United States. This is so for two reasons:
First, unlike the TLAC requirement, MREL is set for each bank based on its expected resolution strategy and considerations of resolvability, and may be adjusted within parameters set out in the framework. While TLAC is the baseline for EU G-SIBs, they are subject to an MREL add-on to align with the general EU approach to resolvability.
This approach has resulted in EU banks being subject to higher LAC requirements than similar banks in other jurisdictions. For G-SIBs, the EU requirements result overall in a surcharge of roughly 4% of risk-weighted assets for systemically important EU banks compared with those in the US.3
Second, the scope of application of MREL requirements is wider than resolution-related LAC requirements in the US. The comparison is stark. Approximately 300 EU banks, including more than 80 large banking groups in the banking union, must meet MREL, calibrated to their expected treatment in the event of failure. It should be stressed that not all those banks are required to hold LAC in excess of own funds. That is the case of banks that are expected to be liquidated in the event of failure. However, in the US only the eight G-SIBs are subject to TLAC requirements. As a result, the EU banking sector is much less at risk of the insufficiencies identified in the US during the banking turmoil.
I should note at this point, however, that there have been moves in the US to expand that scope with a proposal to require banks with $100 billion or more in assets to issue long-term debt sufficient to recapitalise the bank in resolution.4 The benefits put forward for the proposed measure include improving the resolvability of those large banks, reducing the risk of loss by uninsured depositors, giving greater scope for the FDIC to transfer all deposit liabilities and creating additional resolution options. The rule has not yet been adopted, and it is not clear what will emerge from the consultation process, but the arguments made to support the proposal are, I would argue, irrefutable. LAC is central to the credibility of resolution.
Is there scope to revise MREL requirements downward?
However, irrespective of the way in which US requirements may evolve, we want the EU framework to be a ground for stability and economic prosperity. It is therefore valid to ask whether the EU LAC requirements impose burdens on EU banks that could be alleviated by reducing MREL or simplifying how it is set.
It is hard to argue that the MREL framework is not complex, and since it is impossible to calibrate LAC requirements with scientific precision, is that complexity necessary? Simplification is always desirable provided that the ultimate objective is not compromised.
To answer this question, we need to look at the broader framework. Resolution is not credible or feasible without funding, but that funding can come from a range of sources. In addition to the LAC on the balance sheet of individual banks, those sources include resolution funds, deposit guarantee schemes (DGS) and governments. LAC pushes the costs of failure management on to the bank’s creditors and shareholders. Industry-sourced funding mutualises those costs across the banking sector, while public backstop funding channels costs to taxpayers, although they may ultimately be recovered from industry through levies or taxes.
The balance between these possible sources of funding is a policy decision based on a number of political and technical considerations. The latter include the need to minimise moral hazard. In the EU, a clear decision has been taken that LAC should be the first line of defence and that access to mutualised funding sources should be significantly restricted.
The Single Resolution Fund (SRF) is positioned as the second line of defence, after banks’ LAC. This resource can only be used in a resolution after prior loss-sharing by a bank’s shareholders and creditors. This principle is hard-wired into the framework through a mandatory writedown of at least 8% of the failing bank’s liabilities and own funds, and there is no flexibility to override this requirement.
Moreover, while in principle feasible, there is little scope for the deposit guarantee funds to be used to support resolution. This is due to a least cost constraint that, as you all know well in this audience, becomes quite restrictive as a consequence of the super-preference of insured deposits within the EU.
Finally, in resolution there is minimal scope for direct government support. Indeed, the Single Resolution Mechanism Regulation does not provide for the use of a government stabilisation tool, despite this option being included in the EU resolution recovery and resolution directive (BRRD) and therefore being available for EU countries outside the banking union.
As a result of those tight constraints for the use of external funds to support resolution of banks in the banking union, comparatively more internal resources (ie LAC) may be required to execute their resolution strategy than would otherwise have been the case.
In other jurisdictions, a different policy decision has been taken about the balance between the possible funding sources for bank failure management. For example, in the US the least cost constraint for the deposit insurer to fund resolution actions is, in practice, more flexible than in the EU because insured deposits rank equally with uninsured deposits under general depositor protection. Moreover, I have already mentioned that in March 2023, it was possible to override the US restrictions on resolution funding by the deposit insurer to address the systemic risk posed by those failures and compensate for the shortfall in the banks’ own LAC. The use of the systemic override is subject to a high procedural hurdle, designed to ensure that it is used only in exceptional circumstances. However, this “safety valve” is available under the US legal framework, backstopped by the US Treasury, and similar mechanisms are present in other jurisdictions.
We could argue about the merits of introducing a degree of flexibility also in the EU, but the current situation was a deliberate decision by European legislators which is rooted in the institutional context of Europe. While the current EU framework is not necessarily optimal, it is internally consistent.
This is not to say that change should not be considered. The Commission’s original CMDI package contained, in my view, some moderate and balanced proposals which, if adopted, could expand the resolution options for banks whose preferred resolution strategy is a sale of business.
The original CMDI proposes to replace the current super-preference for DGS claims with a general depositor preference. Moreover, it suggests allowing the DGS contribution to resolution funding to count towards the 8% bail-in requirement. Those adjustments would directly alleviate the current constraints on the availability of external funding. Consistently with that, although this is not directly recognised in the CMDI text, there could be grounds to lower calibration of MREL with respect to current levels for a large portion of the banks that are subject to LAC in excess of own funds.
I hope that the final agreement on the CMDI package will achieve the original purpose of facilitating funding for resolution. Yet the ongoing difficult debates that surround this legislation clearly illustrate the policy considerations that underlie the way in which funding sources are balanced in any jurisdiction’s legal framework.
Conclusion
This brings me back to the comparisons that are made with the US framework and the calls for MREL levels to be aligned. As a matter of principle, any material reduction in MREL would need to be compensated by other sources of funding, and in the EU those are comparatively limited at present.
It could imply, for example, greater reliance on the SRF, by relaxing the conditions of access. To spell this out, reducing the amount of prior loss absorption increases the amount of external funding needed to execute the resolution strategy. This would mean that more costs are mutualised across the national banking sectors. Alternatively, lower MREL could be balanced by greater flexibility in public funding, including through the use of government stabilisation tools.
Naturally, any of those measures would entail a significant modification of the political agreement that supported the current resolution regime. Therefore, MREL calibration may appear to be a technical exercise, but any material adjustment would require sensitive political choices.
Many thanks.