The limits of central bank financing in resolution
Speech by Yves Mersch, Member of the Executive Board of the ECB, IMFS Distinguished Lecture Series Goethe Universität Frankfurt, 30 January 2018
1. Introduction
My talk today will focus on the provision of central bank liquidity to entities that are close to or in resolution – a topic that has attracted much public interest.
This discussion is taking place in the broader context of the completion of Europe’s banking union. But a single European Deposit Insurance Scheme (EDIS), alongside an effective, common backstop for bank resolution, remains one of the missing pieces.
While we are not there yet with respect to EDIS and indeed common backstops, the need remains for banks to be able to plan ahead and obtain liquidity, even if determined as failing or likely to fail or once the resolution process is activated.
The question is whether and to what extent this liquidity should be provided by central banks.
The ECB’s position on the matter has been constant: the provision of central bank liquidity – be it through monetary policy credit operations or emergency liquidity assistance, should not be automatically assumed in resolution planning. Resolution measures should be financed by contributions from shareholders and creditors of the bank, or by the State or at Union level, but not by central banks.
This obviously does not mean that an entity otherwise in compliance with central bank requirements for the provision of liquidity cannot access such liquidity.
2. Guiding principles and different objectives behind the provision of liquidity
Before going into the specifics, it is important to distinguish between the actors providing the liquidity; the objectives for which the liquidity is provided; and the beneficiaries of the liquidity.
Starting with the actors, they can be grouped into two categories: on the one hand, the central banks, in the performance of both Eurosystem and national tasks; and, on the other hand, governmental entities and supranational authorities.
If we now look at the first group of actors, the central banks, they have been granted independence from the political instruction of the government. This means that they cannot be instructed to provide liquidity – the provision of liquidity must be their free and independent decision. This applies regardless of whether the central bank provides liquidity in order to pursue the objectives assigned to it by primary law or under domestic legislation to fulfil its lender of last resort function.
Moving on to the objectives, liquidity can be provided to serve the primary objective of maintaining price stability. This is without prejudice to supporting national central bank objectives or to ensuring financial stability to which the European System of Central Banks (or the ESCB) only contributes. But it can also be provided as part of other economic or fiscal policies. While the first set of objectives are the responsibilities of Eurosystem central banks, economic and fiscal policy objectives are firmly within the remit of governmental entities. Central banks can exercise powers only within their mandate and therefore provide liquidity for an objective that has been assigned to them by law.
Concerning the beneficiaries, liquidity can be provided to a solvent institution (this is a typical central banking function and therefore of a temporary nature) or to an insolvent institution, also for a longer period (this, by contrast, is a government task) under the scrutiny of competition authorities.
Having said this, I will focus first on liquidity provided by central banks.
There are two main sources of overnight central bank liquidity generally available to credit institutions. One is in the form of Eurosystem monetary policy operations. The other is in the form of emergency liquidity assistance or “ELA”. These two sources of liquidity meet different objectives and are provided under different legal frameworks.
Our constitutional mandate requests that we at all times to pursue the Eurosystem’s price stability objective in line with a functioning open market economy. The corollary is the provision of liquidity must ultimately serve the functioning of the real economy. Eurosystem monetary policy has to be provided to an entity that participates in the monetary policy transmission mechanism. Participating in monetary policy transmission means in principle channelling liquidity to the real economy. Extending life-lines to institutions outside the monetary policy transmission mechanism is structural or economic policy or financial stability. But if there are financial stability considerations the central bank can only support or participate within its mandate.
In addition, the provision of Eurosystem monetary policy liquidity must also comply with other requirements set down by primary law, notably with the requirement that lending is based on adequate collateral and the prohibition of monetary financing. The Eurosystem should be adequately protected against any risks associated with its lending activity. I will come back to this point later. And as I have said, central bank independence also plays an important role in the sense that a central bank cannot be instructed to provide liquidity, even in a crisis. It has to assess the situation of each entity and decide independently.
I will make a few remarks on the monetary financing prohibition – an important requirement flowing from the Treaty with which the provision of central bank liquidity (including in resolution) has to comply. Any assessment of liquidity provision against the prohibition of monetary financing requires a careful case-by-case analysis.
The ECB has repeatedly stated in its – publicly available – Convergence Reports and in its opinions that financing by central banks, even when granted independently and at their full discretion, of credit institutions other than in connection with their central banking tasks, in particular the support of insolvent credit and/or other financial institutions, is incompatible with the monetary financing prohibition. Financing insolvent institutions is a government task – indeed the ECB has identified criteria to distinguish between government tasks and central banking tasks. The ECB has also repeatedly clarified that while central banks may be involved in administering resolution measures, they should not finance them. National central banks could also administer resolution measures on a genuine agency basis on behalf and for the account of a third party, that is, the government or one of its entities. In this case, however, the central bank would not itself provide liquidity, but simply carry out these tasks like an agent.
It is a primary law requirement that lending can only take place against “adequate collateral”. On the basis of this fundamental requirement, the Eurosystem defines the rules and conditions for collateral to be acceptable, or “adequate”, for monetary policy purposes. This framework is set down by means of various ECB legal acts and instruments (in particular in the General Documentation guideline). Adequate collateral for the Eurosystem’s monetary policy framework means, first, that collateral must fully protect the central bank against losses in credit operations as these losses are in principle, shared. Consequently the Eurosystem is required, under primary law, to act as a secured creditor. In concrete terms this is ensured through pledge or repurchase agreements adequately secured by an asset capable of being appropriated, transferred and/or sold at first notice. A key point here is that a standalone guarantee has never been recognised as adequate collateral under our framework. Guarantees can only play a limited role, namely to make up for the lesser credit quality of a given collateral provided by a financially sound counterparty. In no circumstances would a guarantee “cure” the lack of financial soundness of a given counterparty, or the lack of collateral or a combination of the two, which is often the case in a resolution scenario.
If a public guarantee can prevent insolvency, it is only after having tested its value in the markets. If a public guarantee serves to cut out the markets and replace them with central bank liquidity it would contravene the principle of an open market economy with free competition and could be seen as a circumvention of the monetary financing prohibition.
Indeed it is already questionable to what extent the counterparty is able to fulfil the criterion of financial soundness. Finally, as I just mentioned, it is an important requirement in our framework that collateral can be freely realised and sold at a market price. In the case of a mere guarantee this is not obvious and thereby offers a standard of protection which is legally not comparable to appropriable collateral. If all prudential and regulatory shortcomings could be replaced by a guarantee, it would undermine the whole framework of prudential and monetary intervention and, in the case of public guarantees, the competition principles underlying an open market economy. This is particularly true if a guarantee in rem would be replaced by a blanket guarantee ad personam to a bank. Why then not allow a guarantee for all the banks to maintain financial stability. A cure all guarantee that a central bank must accept is a handover of money creation to the Government at EU or national level. The acceptance of own used government-guaranteed bank bonds in the past was therefore rapidly be curtailed by the Governing Council of the ECB.
Alongside “adequate collateral”, the second important layer of protection against the Eurosystem suffering potential loss is the requirement to interact only with financially sound counterparties.
Without devoting too much time to the technical details, counterparties to Eurosystem credit operations must fulfil certain eligibility criteria. In essence, an entity must not only be a licensed credit institution or a branch but also needs to be financially sound. Article 18.1 of the Statute of the ESCB also allows the provision of liquidity to “ credit institutions and market participants”, for example financial market infrastructures such as central counterparty clearing houses. But I will not reflect on this particular case if not to question that a credit institution in or close to resolution is still participating in the market if it would have the central bank as sole or predominant counterparty. The financial soundness of a given counterparty is assessed by the Eurosystem on an ongoing basis, notably taking into account each entity’s reported capital, liquidity and leverage ratios. The specific eligibility criteria are to be found in our framework and in particular the General Documentation guideline.
I should add at this point that, where necessary, the ECB’s Governing Council may also take discretionary measures in individual cases. So, in some justified instances where there are prudential concerns, it may reject certain assets as collateral or the counterparties despite their being eligible.
Let me turn now to the second source of central bank liquidity: emergency liquidity assistance, or ELA.
Emergency liquidity assistance can be provided by national central banks on the basis of their national competences and to pursue national objectives, namely to preserve financial stability when a solvent entity is facing temporary liquidity issues.
The provision of ELA must likewise be backed by sufficient collateral. It must also comply with primary law requirements, such as the prohibition of monetary financing and it should not interfere with the tasks and objectives of the ESCB (in which case the Governing Council can object to it).
I will make only a few more comments on the provision of central bank liquidity. Central banks can only provide liquidity in the context of pursuing their objectives and to carry out the tasks within their mandate. In addition, with regard to the monetary financing prohibition, the ECB has repeatedly stated that the financing of resolution measures is a government task. Does this mean that the Eurosystem would be prevented from providing liquidity in the context of resolution? The answer is, not necessarily, as long as the provision of liquidity complies with the requirements of any of the two sources of central bank liquidity I just described. These requirements are even more strict in the case of monetary policy liquidity than for emergency liquidity assistance. For example, in the case of the asset separation tool, liquidity can be provided to the solvent part of the bank that is participating in monetary policy transmission and not in order to finance the separation itself. Liquidity can be provided under the generally applicable monetary policy rules and respecting the limits set out in the Treaty and any Governing Council decisions. Yet, whilst the provision of central bank liquidity should not be ruled out in resolution, it should not be assumed either.
This is precisely the reason why the Single Resolution Fund was established: financing of resolution (setting aside the case of recent national insolvency proceedings) should no longer come from the taxpayers, but from the banks themselves. The provision of liquidity by the Single Resolution Fund is of key importance in the euro area. But since, for the time being, this source of liquidity is yet to be fully operationalised, the question is where should the backstop come from? I argue that this source cannot come from the central banks, as resolution financing is a government task.
This is nothing new, as the idea that central bank liquidity should not be assumed nor ruled out is laid down in the Bank Recovery and Resolution Directive (BRRD) and has also been underlined by the ECB on a number of occasions.[1]
Let us now turn to the second group of actors providing liquidity in resolution – governmental entities and supranational authorities – and the liquidity provided by the Single Resolution Fund on which I will be relatively brief.
The SRF ensures uniform practice in the financing of resolutions within the Single Resolution Mechanism (SRM). It also ensures that the financial industry, as a whole, finances the stabilisation of the financial system, as it pools contributions raised at national level from entities within the Single Resolution Board’s remit in each of the 19 Member States in the banking union. SRF support may take the form of guarantees or loans if certain conditions are met and to the extent necessary to ensure the effective application of the resolution tools, within the resolution scheme.
But, as you may know, the SRF is not yet fully funded. It will be gradually built up during its first eight years of existence and should reach the target level of at least 1% of the amount of covered deposits of all credit institutions within the banking union by the end of 2023. So it does not have sufficient funding at this stage to provide liquidity on a large scale.
On the other hand, there is the question whether the fund will be sufficient, even in the final stage, in the event of a very big and widespread financial crisis. And if not, where would the backstop come from? Currently the Fund also lacks a common fiscal backstop as there are still dissenting views concerning an eventual mutualisation of recapitalisation needs resulting from so-called “legacy assets”. In this context, one of my key messages is that the monetary financing prohibition would prevent such fiscal mutualisation through the backdoor of the Eurosystem.
Of course, there might be situations where a “funding gap” could arise. How to fill that funding gap will have to be assessed on a case-by-case basis, also looking at the responsibilities of governments which are the other source of financial support in resolution through State aid.
But it is not my intention to reflect on how such future situations could be dealt with; I have come here to discuss the limits of central bank funding in resolution. In this respect, it is important to repeat that, ultimately, resolution planning should not assume that central bank liquidity will fill the gaps.
3. Concrete examples
Let me now give you some concrete examples. We should clearly distinguish between two types of cases depending on where we are in the resolution process: at the time the bank is determined as failing or likely to fail, or when it is put into resolution.
At the time the entity is determined as failing or likely to fail
As a precondition for a resolution action to be taken the entity needs to be determined as failing or likely to fail. This is one of the three cumulative conditions for triggering resolution. The other two are: No private funding in the market and public interest, in the sense that resolution is necessary to achieve its objectives (art 31 BRRD and 14 SRMR).
At that stage though, it will be too early to know whether there is a reasonable prospect of an alternative private sector solution or supervisory intervention and indeed whether resolution is in the public interest. Nonetheless, it is inescapable that the entity’s risk profile will be worsened by the very fact of the “failing or likely to fail” determination.
Since with “failing or likely to fail” the supervisor hands over to resolution, how can it be financially sound since the respect of prudential criteria is by definition controlled by a microprudential authority?
In principle, such an entity could still have recourse to Eurosystem monetary policy liquidity provided that it complies with the counterparty eligibility criteria, in particular that it is financially sound and has sufficient eligible collateral. However, to address the uncertainty and the associated risk in the Eurosystem’s counterparty framework, our rules provide that the entity’s access to Eurosystem monetary policy credit instruments is frozen at the level prevailing at the time it is determined as failing or likely to fail. And this decision can obviously only be revisited through a new decision once it is assessed that all conditions for an unlimited access have been restored.
Moreover, if necessary, the Governing Council can also impose further restrictions on the grounds of prudence, including suspending access to monetary policy credit operations for an entity with very low capital ratios. Such decisions would be taken on a case-by-case basis to address the specific risk presented by the counterparty.
Now let me elaborate further. Let us assume that following the “failing or likely to fail” determination, the entity is facing cash outflows. Since the terms of its access to Eurosystem monetary policy credit operations will have been frozen and perhaps even further restricted, it will not be able to meet its increased liquidity needs through this means.
What will the bank do? It is likely to seek recourse to ELA from its national central bank in order to meet these additional liquidity needs. So if the entity is assessed as “solvent” by the competent supervisory authority and it has sufficient clearly identified eligible collateral it will most probably receive ELA from its national central bank.
Let me pause a moment on these two concepts. The microprudential authorities’ assessment of “solvency” under the current ELA agreement is based on a number of backward-looking indicators, such as the reported Common Equity Tier 1, Tier 1 and total capital ratios with a possible grace period in the case of a credible prospect of recapitalisation within a limited timeframe. We will probably change these indicators in the future in order to take account of forward-looking elements, in line with the evolution of supervisory practices, so as to have a more accurate picture about the solvency of the bank and the appropriateness of continuing to provide ELA, and in view of the fact that FOLTF means a handover to resolution authorities.
What constitutes sufficient collateral for the purposes of ELA is currently for the different national central banks (NCBs) to define. It is, however, essential that sufficient information on such collateral is made available to the relevant NCB and, in turn, to the Eurosystem. If there isn’t enough information, higher haircuts will need to be applied to make up for the increased risk until the necessary details are provided.
But ultimately, the provision of emergency liquidity assistance depends on each central bank’s national framework, since ELA is carried out as a national function which national central banks may perform unless the Governing Council finds with a 2/3 majority that it interferes with the objectives and tasks of the ESCB.
Since the provision of ELA can interfere with implementation of the single monetary policy (as it is a parallel source of liquidity to the single monetary policy), the Governing Council needs to be informed of the ELA operation in a timely manner. It will assess the situation and, if an interference with Eurosystem tasks is indeed established, object to the granting of ELA. The specific form of an eventual objection is at the discretion of the Governing Council. The Governing Council could, for instance, set certain limits or conditions on the provision of ELA, like the preservation of the financial independence of the NCB.
So central bank liquidity could be rolled over within the limits of the entity’s frozen access and through ELA, but this is clearly not warranted in all scenarios for an entity determined as failing or likely to fail.
At the time the entity is put into resolution
I’ve talked about central banking funding at the time the bank is determined as failing or likely to fail. Let us now examine what can happen later on in the process.
First of all, there are some clear-cut situations where no liquidity can be provided by a central bank, be it under the monetary policy framework or ELA. This is the case for entities put into insolvency or liquidation proceedings.
Second, let us take the example of an entity that is put in resolution following a “failing or likely to fail” determination. Let us assume that it is no longer contributing to monetary policy transmission, because it no longer channels liquidity to the real economy and is in the process of being wound down.
Wind-down entities, whose main purpose is the gradual divestment of their assets and the cessation of their business, have been excluded from access to monetary policy credit operations. This applies for asset management vehicles, too.
As regards their access to emergency liquidity assistance, the relevant central bank assesses the situation of each entity according to that central bank’s national framework. However, in most cases, such entities are unlikely to obtain access to ELA if there are doubts as to their solvency, or if this would raise monetary financing concerns.
Third, let us take the situation where the application of one of the resolution tools will lead to the credit institution becoming financially sound again. This would happen through the sale of business tool, the bridge institution tool, the asset separation tool, or the bail-in tool.
The use of these resolution tools will give rise to a number of individual situations which I will not detail here. I can say, though, that in situations in which the failing counterparty continues to exist, its restored solvency will first have to be confirmed by the competent supervisory authority before the Governing Council will decide to “unfreeze” its access to Eurosystem monetary policy liquidity. Such confirmation of the entity’s solvency will also be needed for the provision of ELA. The entity would need to be handed back to the relevant competent supervisory authority to provide information on the required observation of ratios.
If the entity is a newly created counterparty, such as a bridge bank, it will first have to be licensed, or obtain a licence, to operate as a credit institution. It will then have to be accepted as a Eurosystem counterparty. This means it will have to fulfil all the eligibility criteria set down in our framework, including the requirement to comply with the relevant capital, liquidity and leverage ratios.
Therefore, if applying one of the resolution tools results in the entity’s financial soundness being restored, or if combining its business with that of an absorbing entity results in the emergence of a financially sound entity or group, access to central bank liquidity could resume after the Eurosystem has made its assessment. But this will depend on the elements I have just described.
Time and planning will be of the essence. Indeed, in this stage, the flow of information will be crucial and the competent supervisory authority will have to provide all relevant information on whether the counterparty is fulfilling its capital requirements, so that the necessary decisions on access can be taken.
4. Conclusion
Let me draw to a close.
As I have already indicated, liquidity provision by central banks in the event of resolution must not be assumed ex ante, even though the possibility is not excluded provided the specific rules and objectives of the Eurosystem are followed. The provision of central bank liquidity will be the independent and ad hoc decision of the Eurosystem under the respective frameworks for monetary policy and potential emergency lending not interfering with monetary policy.
Resolution financing is a government task, now complemented by the rules and procedures applied by the Single Resolution Board and the national resolution authorities within the framework of the SRM. Central banks provide liquidity, not solvency support. And funding gaps that cannot be addressed by the industry or through the Single Resolution Fund should be filled, ultimately, by Member States.
Thank you for your attention.
[1] Opinion CON/2012/99 on the proposal for the BRRD: “The ECB wishes to underline that this provision should not in any way affect the competence of central banks to decide independently and at their full discretion on the provision of central bank liquidity to solvent credit institutions, both in standard monetary policy operations as well as emergency liquidity assistance, within the limits imposed by the monetary financing prohibition under the Treaty”.
European Central Bank
Directorate General Communications
- Sonnemannstrasse 20
- 60314 Frankfurt am Main, Germany
- +49 69 1344 7455
- media@ecb.europa.eu
Reproduction is permitted provided that the source is acknowledged.
Media contacts