Before we respond to the question, we would like to highlight the particularities of Euroclear Bank profile and to share with you our key considerations with regard to the consultation.
I. Special treatment for CSDs with a banking licence
Euroclear Bank is an International Central Securities Depository (ICSD) that also holds a banking licence. Therefore, it needs to comply with the Bank Recovery and Resolution Directive.
However, CSDs with a banking licence are not only subject to the regulatory requirements for credit institutions, but also to special requirements related to their CSD status (EU Regulation 909/2014 [CSD Regulation]). The banking services of an ICSD are ancillary (as defined by the CSD Regulation) to its core functions; cash positions are related to its settlement operations and short term. The unique capital adequacy requirements for ICSDs is set down in the CSD Regulation. The CSD Regulation includes the requirement to have an orderly wind-down plan and requires the necessary capital to support such an orderly wind-down of the CSD’s activities. There is therefore, a “conflict” of legislation for CSDs with a banking licence. The Bank Recovery and Resolution Directive recognises this by explicitly noting the need for a special framework adapted to the limited nature of banking services provided by such CSDs in recital 12.
Furthermore, at the global level, recommendations for the recovery and resolution framework applicable to CSDs with a banking licence differ from those applicable to banks. It is not clear yet if the European Commission will issue a special framework for recovery and resolution of CSDs. We, however, ask EBA to take into account the specificities of CSDs explicitly, which we discuss below, in its final technical standards.
II. Summary of key points
We believe that, given the specific nature of a CSD with a banking licence and the specific regulatory and capital regime for such entities under the CSDR, the EBA should propose a derogation to these rules for such CSDs. This derogation should be based on that suggested by the European Commission in its delegated act for ex ante contributions to the resolution financing arrangements (recital 11 of the delegated act).
If, despite the special profile, different regulatory treatment of CSDs with a banking licence and the precedent of the delegated act on contributions to resolution financing, this was not possible, other measures should be adopted as listed below.
• The loss absorption amount should be based on appropriate stress tests. If it is based on capital requirements, it should be based on Pillar I requirements and exclude buffers that are not entity-specific. In the case of Euroclear Bank, intraday credit risk drives its economic capital determination, unlike many other banks. Therefore, to base MREL on Pillar II would by far overestimate the necessary amount required and would distort the level playing field.
• The recapitalisation amount should not impose recapitalisation of business lines that could be wound-down nor assume that the bank’s activities after resolution would consume similar amounts of capital (e.g. it is uncertain whether Euroclear Bank would, after resolution, offer similar amounts of credit). It should be based on the expected capital needs in a steady state, after reorganisation, otherwise the resulting bank would be overcapitalised.
• The requirement for a recapitalisation amount is particularly inappropriate for ICSDs, which are required to establish an orderly wind-down plan and to hold sufficient amounts of capital to make such a wind-down possible under the CSD regulation.
• The MREL requirements should not result in an issuance of additional liabilities where an institution has a stable and sizeable excess of common equity tier 1.
• Liabilities related to the use of payment and settlement systems should be removed from the total amount of liabilities to be covered by the MREL.
Response to Q1: Do you consider that any of these components of the overall capital requirement (other than the minimum CET1 requirement) are not appropriate indicators of loss in resolution, and if so, why?
Using the sum of all components of existing capital requirements as the required loss absorption capacity under MREL has a number of drawbacks:
- Generally, it overestimates the loss absorption capacity requirement of individual institutions. Indeed, not all buffers have been designed to enhance loss absorption capacity. The countercyclical and systemic risk buffers are not institution-specific, and their main objective is to rein in pro-cyclicality. The former is expected to vary and be higher when the markets are thriving.
- Specifically, it is inappropriate to determine the loss absorption amount of institutions that conduct very specific activities, like Financial Market Infrastructures (FMIs) and which have specific ways to determine capital requirements. Euroclear Bank’s capital requirements under Pillar II take into account intraday credit risks. Furthermore, additional capital adjustments are also required under the CSD Regulation. As this differs strongly from market practice, it would lead to level playing field concerns from an MREL perspective.
- More fundamentally, statistical methods that rely heavily on risk-weighting do not provide information on the level of actual losses in resolution. Stress tests would provide a better indication of such amounts, with supervisory and resolution authorities agreeing on what levels of stress an institution should be able to face before entering into resolution. The recovery and resolution plans already identify stress situations in which the viability of institutions is threatened and could be used for that purpose.
While we believe an approach based on plausible and agreed stress scenarios would allow to reflect optimally individual institutions’ loss absorbency needs, we understand that such a method presents other difficulties (including that of ensuring harmonised expectations).
In view of those difficulties, if the EBA chooses to retain capital requirements for the purpose of determining the loss-absorption amount, we would propose a more harmonised Pillar I approach and an exclusion of buffers that are not entity-specific (as referred to under article 2 c) for the reasons outlined above.
The word “adjusting” should not be modified. Indeed, it may be desirable to revise an individual institution’s loss absorption amount downwards to reflect:
a. measures that may ensure loss allocation to other stakeholders (e.g. guarantees, commitments, insurance, contractual liability limitations). Such measures would allow a higher loss amount to be absorbed before entry in resolution. It would be inadequate if the expected MREL for loss absorption purposes would be identical for institutions facing similar initial amounts at risk, but with very different risk absorption measures. This is particularly true for FMIs; and
b. the risk profile of the institution, including its history of losses and recapitalisation events, as well as any other regulatory frameworks it must comply with. Resolution authorities should be able to decrease the loss absorption amount to take into account the fact that CSDs with a banking license should be treated differently based on their special role in the market, different risk profile and distinct regulatory framework.
Institutions’ recovery and resolution plans and their resolvability assessment should not only serve as an input to the determination of the recapitalisation amount, but also of the loss absorption amount. Indeed, the scenarios described in such plans give an indication of the levels of market or idiosyncratic stress that are necessary in order to bring the institution down.
They also show which recovery measures would need to fail before entry in resolution. In that sense, they are an indicator of the overall risk appetite of institutions and of resolution authorities. The easier resolution – the more acceptable it may be to let an institution fail. The lower the institutions’ risk appetite in terms of resistance to stress events (taking into account all loss-absorption and risk transfer measures) – the lower any MREL requirements should be.
Determining a recapitalisation amount gives rise to a number of conceptual and practical issues:
• The timing by which such capital should be available leads to very different conclusions as to the amount of capital needed. The paper implicitly assumes that it is immediately after the “resolution weekend” and that the relevant “bail-inable” liabilities should be part of resources that need to be immediately available for stabilisation purposes. This is emphasised by the mention that any changes to the capital needed for complying with authorisation requirements should reflect the post-resolution situation but without “adversely affecting the provision of critical functions by the institution”.
However, the resolution process may lead to a transfer or sale, or even to an orderly wind-down of some activities that are deemed critical for the market, over a period of time. This means that on Monday morning, the institution may well be required to re-open and offer such activities itself, but transfer them or close them down over the course of several months. Does it mean that the full recapitalisation requirement for MREL should be applicable? Or that a lower amount could do, with the perspective that some activities may be sold/transferred/wound down? We believe the latter, as capital should only be needed to support the institution’s activities in a steady state.
This argument is particularly true for ICSDs. Most of ICSDs’ activities are critical. However, winding-down is not unthinkable, provided sufficient time is available to ensure an orderly transfer of participant cash, assets and business to alternative providers.
• Furthermore, the requirement for a recapitalisation amount is particularly inappropriate for ICSDs in view of other regulatory requirements applicable to them. Indeed, CSDs are required to establish an orderly wind-down plan and to maintain procedures ensuring the timely and orderly settlement and transfer of participants’ assets to another CSD in the event of a withdrawal of authorisation (art. 20.5 of CSD regulation), and to hold sufficient amounts of capital to make such a wind-down possible under the CSD regulation (art. 47 of CSD regulation). Forced recapitalisation is at odds with the wind-down requirement.
• Finally, and though we believe this argument to be stronger for FMIs than for other institutions, it remains applicable to all: if the market does not recapitalise an institution (if it is unable to raise capital before resolution), would that not mean that the market does not support a return to viability of this precise infrastructure? Then, why should recapitalisation be forced ex ante upon the market? Would it not create moral hazard and disincentivise support in recovery?
In view of the preceding points, we believe that the recapitalisation amount, if any, should be based on the resolution strategy and on the resolvability assessment. It should be the lowest of:
• The amount necessary to conduct the business in the least capital-intensive way during the resolution process. For example, in the case of Euroclear Bank, capital requirements are mainly driven by intraday credit provision to participants. In resolution, this source of risk would likely be reduced, automatically leading to lower capital requirements.
• The amount necessary in the new steady state, taking into account the shape of the institution after resolution has been completed, including any transfer/sale/wind-down of business that may take place, which may entail the closing down of critical activities over time.
The peer group approach will allow resolution authorities to inform their decision based on a benchmark. However, we believe that the proposed “automatic” approach has a major drawback: whatever authorities decide will influence the median that serves as benchmark. This may lead to an upward drift in MREL requirements for G-SIFIs. This should be addressed.
Due to that major drawback, we do not believe that this approach should be extended in a mechanical way (i.e. forcing authorities to match a median) for other types of institutions. We nevertheless acknowledge that comparisons are informative and we encourage authorities to share information in that respect to ensure a level-playing field. This should not prevent authorities from taking into account the specificities of the institutions under consideration. For example, any set of O-SIIs (domestic or cross-border) would be rather different from Euroclear Bank, which, as an FMI, should only be benchmarked against similar types of institutions.
As also explained below (Q11), we believe that “sufficient MREL to access resolution funds” should, in the case of FMIs, at least exclude liabilities related to the use of payment and settlement systems.
The draft RTS leaves a high level of discretion to resolution authorities in setting the MREL and, in that sense, it does allow authorities to determine MREL in a way that would be appropriate for individual institutions.
However, by adding up a loss absorption and a recapitalisation amount, and assuming that such recapitalisation amount should be set such as not to disrupt the provision of critical activities, the draft RTS may encourage authorities to set an excessively conservative MREL, in particular for FMIs.
This could force institutions to issue bail-inable liabilities they do not need. For example, even though Euroclear Bank has a capital ratio of some 40%, a strict interpretation of the draft RTS may lead to a doubling of the requirements in terms of eligible liabilities and own funds. Would it be reasonable for Euroclear Bank to issue large amounts of debt that it does not need for the purpose of conducting its business?
Finally, the expression of MREL in terms of eligible liabilities is not entirely appropriate for FMIs, as their liabilities are not sought after for funding purposes, but are the consequence of their business model. Euroclear Bank’s liabilities are heavily driven by overnight participant balances. Euroclear Bank invests this cash short-term in reverse repurchase operations, but does not seek such funding actively. The fact that Euroclear Bank does not conduct maturity transformation and that such funding does not serve to fund medium or long-term assets, shields Euroclear Bank from the adverse consequence of a funding withdrawal in a crisis. Euroclear Bank’s balance sheet would automatically shrink, with limited impact on the banks’ activities.
We would therefore propose to exclude liabilities related to the use of payment and settlement systems from the total amount of liabilities to be covered by the MREL.