• The MREL calibration should not lead to more stringent requirements than those of TLAC for banks subject to both requirements (G-SIBs)
The proposed logic for the MREL calibration method is greatly inspired by the principles which were originally considered for the TLAC calibration: regenerate enough capital on the assumption that existing capital will have been completely lost at entry into resolution. This rationale was abandoned during FSB discussions and is no longer included in the final TLAC principles.
The TLAC rules are not finalised and are still subject to probable modifications. We strongly recommend the EBA to accommodate enough flexibility in the proposed standard so that MREL calibration does not lead to higher requirements for G-SIBs than TLAC will. This is an essential issue for EU banks in terms of ensuring an international level playing field.
• In particular, the two steps approach proposed by the EBA leads to an inflationary calibration
The proposed method is highly questionable. In our opinion, the MREL requirement should be considered as a whole (as it is described in BRRD) and not built in two subsequent parts (first setting the loss absorption amount –LAA, then the recapitalisation amount). As currently proposed, the method will inevitably lead to precautionary approaches at each stage of the analysis, leading to overall disproportionate requirements.
The two step approach does not, in particular, enable proper consideration of the various options included in the recovery plan. Some of them may either be performed during the recovery phase (thereby reducing the loss absorption amount) or partly, if not in full, in the resolution plan (thereby reducing the recapitalisation amount). It is impossible, of course, to say whether recovery plan options should reduce the loss absorbency amount or the recapitalisation amount, but they should be taken into account in the overall amount. The proposed two steps approach makes this difficult.
• The assumption of loss of all own funds is too extreme
If the approach in 2 steps is nevertheless retained, we believe the EBA’s method to assess potential loss in resolution (total loss of all capital) is too simplistic and mechanical:
o It does not take into account higher banking sector capitalisation subsequent to the implementation of CRR/CRD IV nor banks’ ability to absorb shocks as demonstrated by the stress tests performed within the ECB’s Comprehensive Assessment;
o It does not take into account the benefits of diversification. Banks that have a significant diversification of their activities in terms of geography and variety of business lines are more resilient to crisis: this should be taken into consideration in the MREL determination; the current drafting of the RTS seems to ignore these diversification benefits;
o The scenario of losing 100% of minimum own funds (including buffers and Pillar2) is highly unlikely. It is more realistic to assume that there will be a certain amount of own funds left, as it seems unrealistic to assume that a supervisor, given the lessons of the last crisis and their increased powers, would allow a bank to lose all buffers without requiring any recovery or early intervention measures.
• The leverage ratio is not adequate for assessing the LAA
The EBA retains the highest of all regulatory constraints: capital ratio including buffers and Pillar 2 and leverage ratio. We are opposed to the inclusion of the leverage ratio as part of the MREL assessment.
The leverage ratio has been conceived as a backstop to capital requirements; it only includes Tier 1 items, and should not be used for other purposes such as a loss absorption measure. In addition this ratio has very different impacts depending on banks business models, its calibration is still being finalised, its implementation is underway and it is not an appropriate measure at subsidiaries’ level.
We would recommend that the leverage ratio be ignored for MREL purposes, and at least until it has been finalised, its impacts have been assessed, and its suitability to be used for MREL assessment has been properly thought through.
• The implied calibration is overestimated and can exceed FSB requirements
This is a consequence of article 2 of the EBA proposal, even if paragraph 5 leaves a certain leeway to authorities to adjust downwards the LAA. The RTS as currently drafted will incentivise authorities to stick to this rule which is too extreme. In particular the Pillar 2 can be very high in some countries where the supervisor for example requires high buffers to protect deposits. That is why the most adequate starting point for an MREL assessment should be the minimum capital requirement (please see article 92 of Regulation 575/2013).
We note that the FSB has taken 8% of RWA only as a starting point of its calibration and we do not see the rationale behind setting a higher constraint for European banks.
• An impact study should be carried out
We note that banks’ historical loss data are explicitly excluded from being relevant to the calibration of MREL, which we find surprising, particularly at a time when the FSB is preparing to conduct an analysis of historic losses. In addition, a comprehensive impact study should be conducted (including market absorbing capacity, costs …). We advise EBA to seize the opportunity of the FSB/BCBS study to investigate the impact of its RTS on the full scope of concerned entities.
The fact that minimum own funds as loss absorbing components could be adjusted by the resolution authority is an indication that the proposed starting point of a two-step approach is not appropriate and that a more global approach should be studied.
If the 2-step approach is retained, we are concerned by the suggestion that the resolution authority could increase the loss absorption amount (LAA) above capital requirements. As already stated in our response to question 1, we believe that the implied calibration retained by the EBA is already over-conservative and we do not see additional vulnerabilities that may not be already covered by capital requirements. We believe the LAA should be set at 8% of RWA for all banks and that the resolution authority discretion to add macro prudential (pillar I) or micro prudential (pillar II) buffers should only be used in exceptional circumstances and in relation to the feasibility of the resolution plan. As mentioned above, recovery options must be able to reduce the overall MREL assessment, on the grounds that they will be exercised, at least partially, either before resolution, or after resolution as part of the restructuring plan.
Besides, the wording of article 5.2 is not clear. It should be clarified that the resolution authority is only empowered to define the overall MREL, but it does not have the capacity to challenge the supervisor’s assessment of a bank’s ongoing solvency, and as such is not provided with any power regarding the entity’s capital requirements determination. This must remain the competence of the supervisory authority. We believe that the wording of this paragraph should be modified in this regard.
Finally, in order to foster transparency and increase understanding of applicable MREL requirements, we suggest that the resolution authority be obligated to share bilaterally its calculation approach and assessment with each bank.
We do not support article 2.6 which links the resolvability of the banks and the LAA assessment. We are unsure as to why resolution impediments should impact the LAA rather than the recapitalisation amount, and are in some circumstances equally unsure as to whether they should impact either amount. If a resolution plan cannot ensure that an institution will be resolvable (absence of sales options or wind down possibilities for instance), then the action required to remove such impediments may be something completely different from simply increasing MREL. The rule should specify that an increase in MREL is justified exclusively in the case of financial impediments that can be removed by such action.
Moreover, using the regulatory capital as a unique reference to calibrate the LAA is overly simplistic. As mentioned above, both recovery options and diversification effects should be taken into account. Diversification (geographical and/or product mix) makes banks more resilient to a crisis in one particular activity or market.
As stated above, we believe that the EBA should take a more holistic view instead of splitting the MREL in two parts. As such, the recapitalisation amount should not be determined in isolation nor as an addition to the LAA.
If the 2 steps approach is however retained, we appreciate that the EBA has indicated that the recapitalisation amount can be based on a reduced scope of activities according to the resolution plan (we would like to remind the EBA that article 45.6a of BRRD provides that MREL must be determined to ensure that the institution can be resolved, not revived as it was before resolution).
Nevertheless, according to article 3.3 as currently drafted, authorities are incentivised to use bank data before resolution to calibrate the recapitalisation amount, unless certain conditions are met. We oppose this approach and advocate for using the resolution plan to calibrate the recapitalisation amount as the overarching rule with no other condition required. It is not logical to use the metrics of a bank that is operating effectively and may be pursuing a growth strategy, to fix an amount of post-resolution recapitalisation. We should not forget that the recapitalisation amount will be fixed immediately at, not in the weeks preceding resolution.
In particular we are opposed to the condition of article 3.3c, as it fundamentally repeats the conditions of article 3.3b, (credibility of restructuring) and serves no useful purpose other than to introduce an unjustifiably higher hurdle for GSIB’s which will encourage authorities not to make use of the option outlined in article 3.3. This condition should be deleted.
Regarding the proposed level of own funds required after resolution, we believe that the RTS should limit the requirement to the minimum capital (8% x RWA). This level is already significant and should be sufficient to meet the purpose of article 45.6b of BRRD (see also our answer to question 5 below). We recommend therefore to delete article 3.6(b-d) and 3.7 of the draft RTS.
Moreover, the rationale of including Pillar 2 capital requirements in the recapitalisation calibration is not appropriate in our view. Any pre-resolution Pillar 2 capital requirement imposed by the supervisor would be based on the bank’s risk and vulnerabilities (mainly based on the SREP outcomes). Once the institution enters into resolution, authorities would impose a tough restructuring plan which may eliminate, or at least significantly reduce, all the previous capital, liquidity, business and governance uncertainties. As a consequence the restructured entity will likely have reduced or non-existent Pillar 2 capital surcharges.
In the same vein, the capital buffers should not be used to calibrate the recapitalisation amount. As a comparison, CRR provides that institutions have time to restore their buffers after a regulatory breach; similarly banks should be given time after resolution to rebuild all buffers.
Should the pillar 2 or capital buffers be maintained, the recapitalisation amount should be calibrated after taking into account the impact of losses, or the entry into resolution, and the impacts of the restructuring plan on the restructured entity. In particular systemic buffers should not be assumed to be required in principle.
Regarding the reference to the leverage ratio, please refer to our comments above (question 1) with respect to its deletion.
We believe restructuring is bound to reduce a bank’s systemic nature, hence it seems inappropriate to recapitalise that bank at the same level as its pre-crisis peers. We believe the peer group approach is unnecessarily inflationary, and therefore suggest that a bank under resolution may be recapitalised at the minimum level –i.e. 8% of RWA. The systemic nature of a restructured bank will be dealt at the resolution plan level.
Beyond this suggestion, we are opposed to the use of a market reference to determine a regulatory requirement. Indeed, G-SIIs from different jurisdictions may be subject to significantly different Pillar 2 and buffer requirements, affecting their total CET1 ratio. Any comparison should be adjusted for differences due to national discretions and reflecting competent authorities’ appreciation of risk.
Moreover, there might be a risk that anchoring the recapitalisation amount to the peer group median value would push the institutions’ capital levels into a rising spiral. All institutions manage capital requirements with management buffers, and any increase in such buffers would then automatically translate into increased MREL requirements for all other banks in the same peer group.
Finally, we believe that the requirement is disproportionate in relation to the goal of article 45.6(b) of BRRD which provides that the level of own funds after resolution should be sufficient to restore market confidence. Market confidence after resolution does not only depend on the level of own funds, but also on the quality of management, the way the resolution is conducted by the resolution authority, the availability of liquidity, including from central banks, and the circumstances which originally brought the bank into resolution. We would also like to point out that BRRD has the impact of considerably enhancing the quality of capital of banks in resolution. Given that, beyond write down, the conversion tool only allows for conversion into shares (no AT1 or T2 conversion possible), any recapitalisation amount will by definition be made up of CET1 only. An 8% CET1 requirement should be more than sufficient to re-start a bank after resolution, given that it will then be subject to close supervisory oversight and be the object of a demanding restructuring plan.
We are concerned that the RTS does not take into account the implementation of MREL in banking groups, and do not see in what way the dispositions of articles 2 and 3 reflect any differences between consolidated and subsidiary capital requirements.
As highlighted in BRRD’s recital 80, MREL should reflect the appropriate resolution strategy and be set at the appropriate level in a banking group. As provided in article 45 of BRRD, the MREL level of group entities should be discussed within resolution colleges, and should be fixed at levels that are appropriate to the nature of an individual entity, whether it is a resolution entry point or not.
Our initial comment is that, if there is to be a threshold for liability exclusion, it should be made clear that the purpose of the threshold is to dispense resolution authorities from conducting complex ex-ante NCWOL analyses, and no more than this. If the threshold is exceeded, then the effect should be that the resolution authority conducts necessary analyses to determine if there is likely to be a problem in meeting the NCWOL principle.
Article 44.3 of BRRD provides that a resolution authority may decide to exclude liabilities from bail-in on a discretionary basis (in exceptional circumstances listed in paragraphs a to d). In that case, losses that would have been borne by excluded liabilities are carried over non-excluded liabilities as long as the NCWOL safeguard is respected. When this safeguard is in danger, the resolution authority has a series of options available: it may either draw on the resolution fund (but only if an 8% bail-in has been applied), it may reconsider the scope of discretionary exclusions, and it may choose to use resolution tools other than bail-in, in addition to or in replacement of bail-in.
It should also be noted that the class of liabilities where NCWOL problems are most likely to arise is the class of senior unsecured liabilities, and that NCWOL problems will most likely arise when assessing the conversion of debt into equity, because senior debt is most likely to be called upon as part of the recapitalisation process, not for loss absorbency. When assessing the conversion of debt into equity, losses will arise only if the value of equity issued in exchange for debt deteriorates as a result of additional losses that were not anticipated at the start of the resolution process. Losses on conversion cannot be assumed to be any more than a portion of the value of converted liabilities, and in many cases will be non-existent.
We believe that losses in liquidation are, by definition, significantly higher than in resolution. Resolution has been introduced because a classic insolvency process is agreed by all to be inappropriate for large banks, as it tends to generate delays, fire sales, and allows an extended loss of confidence to accelerate and deepen the losses suffered by the failing bank. The bigger and more complex is a bank, the greater is the gap between losses in liquidation and in resolution.
This gap, or ‘resolution bonus’, needs to be carefully analysed by resolution authorities, and confronted with a forecast of actual losses suffered by converted liability-holders, in order to determine whether a potential NCWOL problem may arise, and to decide how they would address it.
The resolution bonus is of particular benefit to senior unsecured creditors, because in resolution they will be converted at par into capital instruments, rather than facing a long, uncertain and probably partial recovery of their claims. They may even benefit from an upside on their equity where resolution is successful.
Therefore, the proposal for a threshold of 10% can be seen as an administrative simplification to avoid these calculations in circumstances where a NCWOL issue is improbable, but should have no impact on the assessment of the MREL requirement, which should only be affected after detailed NCWOL analysis.
If 10% reflects an EBA assumption of the difference between losses in resolution and losses in a classic insolvency process, then we believe that this is an excessively conservative stance that takes an overly pessimistic view of the advantages of resolution, and should therefore be significantly increased.
We agree with the proposal that systemically relevant institutions, and indeed all institutions, should have sufficient MREL. However, as already said above, the MREL requirement should be calibrated according to the resolution plan of each institution and not according to a uniform approach like a minimum level for systemic relevant institutions, which is not mentioned under the specific criteria in article 45 of the BRRD.
Moreover, the level 1 text underlying this RTS does not provide that the MREL should be set high enough to draw the resolution fund, and contains no legal requirement to set the MREL for all systemic relevant institution at a uniform minimum. Indeed a uniform minimum would also inadequately reflect institutions’ different business models/ risk profiles (retail, universal versus investment banking) and diversification effects.
Systemic institutions (G-SIIs and D-SIIs) have various profiles and calling on the resolution fund should not be regarded as the single answer to achieve a satisfactory resolution.
As article 7 of the RTS is written, resolution authorities would be given no other choice than determining the MREL in a manner enabling to reach 8% bail-in because condition 2(b) would be extraordinarily difficult to comply with (there are no plausible circumstances..). This article should be modified to give more flexibility to the resolution authority to determine the appropriate MREL.
We support the phasing in of the requirement. It should be a general rule and not a national discretion in order to ensure a level playing field. We consider that a comprehensive QIS exercise in Europe is vitally important to set a workable transition period.
In our opinion, the main issue arises from the additional volume of capital required to comply with the MREL and its availability in the market under conditions making it sufficiently attractive for investors. This may limit institutions' capacity to swiftly adopt the MREL.
We therefore believe there should be a sufficiently long transition period, ideally aligned to the TLAC one not only along the run up phase but also in the implementation phase post 2019.
Yes, it makes sense to set a relevant transitional period during or after the resolution process, depending on the resolution outcome. It can be assumed that market expectations will reflect the situation of an institution in or shortly after resolution; hence the institution will not be expected to reach the pre-crisis MREL target immediately. Once MREL is depleted in resolution, we would not expect the minimum capital requirements to be restored right away and the buffers to be rebuilt over a period of time dependent on prevailing circumstances. The same transitional period that may apply for the TLAC should be allowed for restoring the MREL over the initial post-resolution period.
In our opinion the draft RTS has several failings in terms of promoting consistency:
• It penalises banking sectors and banks which are highly capitalised and do not need funding from the secondary market;
• In attempting to introduce consistency between MREL and TLAC, it creates the potential for inconsistent requirements on banks according to whether they are subjected to TLAC or similar requirements or not;
• It exacerbates differences between jurisdictions.
There are currently significant differences across jurisdictions with respect to prudential capital requirements set at individual institution level, either through the exercise of national discretions and methodological differences in the determination of Pillar 2 requirements or through the imposition of different buffer levels. Rather than creating a harmonised level of MREL for similar institutions in the EU, the current RTS actually amplifies the existing divergent practices by using the prudential requirements as the basis for both the loss absorption and recapitalisation amounts (thereby effectively increasing the differences).
We would like to draw the EBA’s attention to differing ability by banks to draw on DGSs and the potential for an unlevel playing field this may create. We are concerned that a two-speed regime may be implemented whereby small and medium-sized banks would be able to take into account the bail-in of the DGS’s resources in their MREL while G-SIBs and D-SIBs would not. We understand article 109 of the BRRD permits the use of deposit guarantee funds in resolution, but limits their contribution to the resolution of a given bank to the lesser of a) the amount of losses covered depositors would have borne in insolvency, or b) 50% (or a higher percentage set by the member state) of the target level of the deposit guarantee fund. However, the draft RTS incorporates this term in an entirely different manner which would allow banks, whatever their size, to rely on the DGS for up to 50%, or an even higher percentage, of the total resources of the DGS. We believe this provision transforms a BRRD measure designed to protect DGS resources from being overstretched into an incitation to consider that any bank may benefit from 50 % or more of a DGS. This may bias the resolution authority assessment and would clearly create moral hazard.
Finally, we are not fully convinced that the MREL, as currently set, properly deals with subsidiaries that may be locally systemic; we are concerned this may create a potential unlevel playing field between subsidiaries of a systemic bank and local banks in a given country.