The Basel capital framework, and robust stress testing (including by the ECB and EBA), has ensured that banks have sufficient capital to absorb extreme losses, and in most states of the world, if bank losses were to deplete capital below the regulatory minima, the capital would not be completely exhausted (even before BaU pre-provision profits) . We also observe that a bank’s loss absorbing capacity would extend beyond instruments that qualify as regulatory capital (eg legacy capital) or MREL (eg unsecured term debt with less than 12 months remaining maturity). We therefore believe it would be sufficiently prudent to base the assumption of capital to be lost on the minimum capital requirement (rather than that plus buffers), consistent with the FSB’s TLAC proposal.
Regarding the individual elements above the minimum, we note that Pillar 2 is set at the discretion of supervisors and, while this may be tied to expected loss, where it is not (eg for non-Pillar 1 risk types such as interest rate risk in the banking book, lease tail risk, insurance risk etc, and where there is a resolvability component to any Pillar 2 add-ons) it seems inappropriate to include it in the loss absorption component. Other types of capital buffers – such as systemic risk buffers and countercyclical buffers – reflect metrics such as size and interconnectedness, or macroprudential tools used to lean against increases in system-wide risk, rather than expected loss.
The leverage ratio is not designed as an indicator of loss in resolution, but is intended to act as a backstop based on balance sheet size (and is not calculated in a consistent manner across jurisdictions). We therefore do not believe leverage is the best determinant of MREL, but if it is used (and we acknowledge it will form part of the TLAC requirement for G-SIBs), we would ask that leverage is defined on a BCBS consistent basis.
Yes, we welcome the proposed flexibility for the resolution authority to allow downward adjustments. This could be appropriate if, for example, a firm’s recovery plan (deemed credible by its home regulators) involved the sale of businesses, or the post-resolution group was expected to be smaller and / or less interconnected.
No, the loss absorbency benchmarks seem comprehensive.
We agree that the amount of capital required post-resolution should in part depend on the resolution strategy. We believe that in bail-in a firm should be recapitalised to the minimum regulatory capital, with buffers rebuilt over time (within a timeframe determined by the prevailing circumstances), whereas the MREL proposals imply instant replenishment of all buffers in a bail-in, which we are strongly opposed to.
We believe it would be inappropriate to include any Pillar 2 TLAC component for a G-SIB (as translated into MREL) in the post-resolution requirement, since that would reflect any firm-specific impediments to resolvability and in a post-resolution world the firm would have been resolved.
Given the vast potential differences in G-SIBs’ balance sheet composition, minimum capital ratios, position in the economic cycle, fiscal and monetary policy being pursued by their home and host governments, recapitalisation requirements will always be different. While a peer-group approach (across all G-SIBs) would provide one comparator, we believe post-resolution capital requirements would more appropriately be set by regulators on a case-by-case basis depending on the firm’s circumstances. We are therefore opposed to the proposal in Article 3(7) to require recapitalisation to at least the median CET1 ratio of a G-SIB peer group, and also consider it inappropriate to base recapitalisation requirements on O-SIIs in a jurisdiction .
If, however, a peer-group approach is adopted, we believe peer groups divided by their precise G-SIB capital buffer would be too narrow (it could be only 1-2 other firms).
For SPE firms, where HoldCo is the resolution entity, we believe MREL should be set primarily on a group consolidated basis, with firms able to apply for a waiver from requirements for non-material subsidiaries. Where subsidiary requirements are applied, the subsidiary should be able to meet this wholly or partially through internal MREL issued to the parent, or through external MREL that meets the eligibility criteria.
For MPE firms, with multiple resolution entities, MREL should be held at the resolution entity (subsidiary) level.
Where (internal) MREL is ‘downstreamed’ to material subsidiaries via an intermediate entity, and where the applicable regulatory regime includes solo capital requirements, the intermediate entity will attract solo capital charges on funding exposures to its material subsidiaries. This could give rise to level playing field issues if rules are not applied on a consistent basis internationally. We would therefore ask the EBA to work with the BCBS to establish whether a waiver from intra-group capital requirements, Large Exposure rules and leverage ratio requirements would be appropriate for internal MREL exposures (where not covered by the Basel capital framework).
Finally, we note that certain of the elements of the proposed loss absorption amount and recapitalisation amount are applicable only at the consolidated level, for example the G-SIB buffer and Pillar 2 requirements. Furthermore, leverage requirements are not necessarily applied at the level of subsidiaries. It follows that these requirements should be disregarded when calculating the MREL for subsidiaries.
Yes, we believe a de minimis derogation is appropriate, given the existence of excluded ‘administrative’ liabilities (such as tax liabilities and utility bills), as well as liabilities which the resolution authority may use its discretion to exclude.
Although it is difficult to pin-point precisely the right level, a derogation threshold of at least 10% seems appropriate, since we believe that expected losses for senior creditors in insolvency (which constitutes the No Creditor Worse Off (NCWO) counterfactual) would typically be at least 10% higher than in a resolution which preserved the firm as a continuing business (given the legal and other costs of insolvency and likely firesale of assets). Moreover, if debt is converted to equity (at book value), then this provides more comfort against successful NCWO claims. Alongside the need to take into account the ‘ordinary course of business liabilities’, we would supportive of setting a higher threshold of 15-20%.
No. Some national authorities may determine that it is not desirable to rely on resolution funds (for a loss ‘tranche’ of 8-13% of liabilities), since this would imply transferring the loss from the firm’s creditors to the banking industry more broadly (which could exacerbate contagion), and increase moral hazard.
We note that the 8.0% refers to allocated losses and not to MREL-eligible liabilities, so it should be possible to bail-in 8% of liabilities with an MREL requirement lower than that (ie some liabilities would not be eligible because they had fallen below 12 months remaining maturity).
No, we believe a longer transition period (of 5-6 years) would be more appropriate, given the significance of the requirements, likely market capacity constraints and intention to align MREL requirements with TLAC requirements for G-SIBs (where we would welcome alignment of the conformance periods). In determining the transition period, resolution authorities should also have regard to the timing of structural reform programmes for individual institutions, and the challenges of implementing such structural reforms simultaneously with MREL and TLAC requirements.
We urge that any requirement to issue contractual bail-in instruments as part of a subsidiary’s internal MREL (BRRD Article 45, 13-14) is not implemented during the transition phase and that authorities have regard to the relative ranking of Holdco/Opco creditor rights throughout the transition phase.
Article 9(2) could also provide that resolution authorities should be required to notify each institution of its planned MREL within a certain period of time before each 12 month period begins, to enable institutions to prepare adequately for the implementation of MREL. We believe the final sentence of Article 9(2) should be amended so that resolution authorities can only subsequently revise planned MREL for future 12 month periods, and so that resolution authorities cannot reduce the length of the transition period or retrospectively increase planned MREL during any 12 month period which has already commenced. Otherwise, there will be significant market uncertainty and issuance planning will be difficult.
Once MREL is depleted in resolution, we would expect the minimum capital requirements to be restored immediately and the buffers to be rebuilt over time, with the timeframe dependent on prevailing circumstances.
While we acknowledge that some aspects are firm-specific (eg the preferred resolution strategy), overall it seems that – especially if a minimum 8% of total liabilities is required – the MREL proposals would impose a high minimum requirement across the industry, without taking full account of the risk profile of individual firms’ balance sheets.
Where resolution authorities are required to make judgements about whether certain instruments are expected to be excluded from bail-in in setting MREL (eg corporate deposits, structured notes), we feel a consistent approach would assist investors in understanding and pricing risk.
We consider that a key risk to consistency of MREL requirements across Member States is that resolution authorities could specify that, in order to constitute MREL, instruments must contain particular terms in addition to meeting the criteria of Article 45(4) of the Directive. We would ask that the EBA encourages resolution authorities not to impose additional criteria beyond those in Article 45(4) of the Directive, at least until the FSB’s TLAC requirements have been confirmed. Otherwise, there may not be a level playing field across institutions in the Union.
The impact assessment should take into account the proposed TLAC requirements for G-SIBs (including market capacity issues), and likelihood of this being extended to D-SIBs in due course.
