5. Banking Resolution and Its Key Concepts and Tools

The reader should be aware that over the last few decades alternative forms of intermediation have risen alongside traditional banking. It has been already observed that these intermediaries may also pose systemic risk (Bengtsson, 2016; Aramonte et al., 2022). This could potentially require additional regulatory attention, even in fields that traditionally were an exclusive domain of banks, such as the transmission of monetary policy (Dabrowski, 2017; Cappiello et al., 2021). From the pure perspective of resolution, some non-bank intermediaries, such as financial market infrastructure, are already subject to resolution regimes similar to those for banks. In the present chapter we focus on bank resolution only.

The literature on the effects of bank liquidation is abundant and spans over decades. The specific point we want to make here is that – in addition to the risks of financial contagion – a bank liquidation is likely to affect the real economy as a whole. For example, the performance of firms depending on banks’ lending is affected by capital shocks in banks (as found in Chava & Purnanandam, 2011). Another well-known consequence of an unsterilised bank failure – especially under a systemic crisis – is a steep reduction in the lending activity, which also affects households and small and medium enterprises. This phenomenon, which usually goes under the name of ‘credit crunch’, was well documented after the 2007–2008 financial crisis (Mizen, 2008; Brunnermeier, 2009; Hull, 2009, on secondary effects on investments and wedges Buera & Moll, 2012; Cingano et al., 2016, with an EU perspective Iyer et al., 2014, for Japan Watanabe, 2007). These negative consequences on real economy are normally experienced on top of much more intuitive immediate effects, such as the unavailability (or even not full repayment) of deposits and the suspension of certain services provided by the bank, such as for cash or payments.

We want to note for the sake of transparency that models in the literature also point towards the opposite conclusion. According to a model developed in 2015 (Klimek et al., 2015), considering three possible solutions for a banking failure (bailout, bail-in, and sale of business, which is also a resolution tool) as well as different levels of economic growth and employment, in no scenario a bailout is able to outperform resolution tools in terms of efficiency.

Literature often refers to bank bailouts as forms of implicit guarantee of banks’ debt. Within this chapter, we refer to two contributions providing also estimations on the extent of this implicit guarantee. Toader measured this effect as the difference between credit rating scores with and without the expectation of public supports (note that the expectation of public support is indeed a factor considered in the credit rating). Interestingly, writing in 2015 after the adoption of the Bank Recovery and Resolution Directive (BRRD), the author noted a consistent reduction of this spread across the sample of 56 large European banks from 17 countries of the European Union, suggesting that the adoption of the European resolution regimes has had an impact on the expectations of market participants.

The expression was first popularised by the US congressman Stewart Brett McKinney in 1984, in the context of a hearing concerning the intervention of the Federal Deposit Insurance Corporation to rescue Continental Illinois. Since then, the expression has entered both the public and the academic debate on the situations of moral hazard of the largest credit institutions. A review of the literature concerning the too-big-to-fail issue would probably deserve a separate chapter. However, we urge to provide the reader with some ‘classical’ references in case he were interested to gain additional knowledge on this topic. One of the earlier contributions is the one by O’hara and Shaw (1990). Appeared few years after the speech of McKinney, this article found that the declarations by the Comptroller of the Currency concerning deposit insurance for largest banks boosted their equity value, when compared to the smaller banks. While today deposit insurance is commonly considered as a key element of the safety net protecting financial stability, neutral in terms of moral hazard provided that it is clearly limited to covered deposits and excludes any other creditors (Gropp & Vesala, 2004; Schich, 2009; Demirgüç-Kunt et al., 2015), the article well outlines the relation between banks’ size and moral hazard. A now classical reference in the literature concerning the too-big-to-fail issue is the book by Stern and Feldman (2004), with a foreword by Paul A. Volcker. This book is an ample monographic study on the too-big-to-fail issue, which provides the common definitions and concepts that are now used in the literature. With a historical and comparative approach, a useful reference can be found in the 2011 article by Morris Goldstein and Véron (2011), which also outlines the key differences between the too big-to-fail-issue in the United States and in Europe. For a reconstruction of the debate on the too big-to-fail-issue after the 2007–2008 financial crisis, the reader may also refer to the 2014 article by Kaufman (2014), which also reflects on the previous literature to improve the definitions of the issue.

In the recent case of Silicon Valley Bank, the depreciation of government bonds was caused by an increase of monetary interest rates in the market and not by a deterioration of the rating of the US Treasury bonds. However, we refer to this case to show in concrete what are the consequences of a bond depreciation on banks’ balance sheet. It has to be noted that most of the literature on the doom loop was published in a period of low interest rates. Raising interest rates coming on top of a doom loop depreciation would most likely act as an amplificatory effect.

The KAs have a scope (KA1) that goes beyond banks, including also other types of financial institutions. In the present chapter, wherever we refer to the KA, we only select those parts pertaining to the resolution of banks, which are those reflected in the various bank resolution regimes. Hence, the reader should not be surprised to find a longer list of objectives within the KA.

The KAs also make a distinction between the objectives of resolution regimes, included in the Preamble, and objectives of resolution authorities in KA2, with the second list being a subset of the first. As a matter of fact, the legal texts do not make this difference and refer to the ‘resolution objectives’. See for example Art. 31 of Directive EU 2014/59 of the European Parliament and of the Council of 15 May 2014 establishing a framework for the recovery and resolution of credit institutions and investment firms, the so-called Bank Recovery and Resolution Directive (BRRD). The terminology used in this list should be seen as a simplification of the wording used in the KA for illustrative purposes.

To be precise, in the wording of the KA the contagion prevention is not listed as an objective of resolution regimes or resolution authorities, but it is itself an overarching objective included at the beginning of the Preamble and constituting a motivation for the adoption itself of the KA. We choose to include it along the others, because this is the approach normally taken by the legal texts reflecting the KA.

European Commission Implementing Regulation (EU) 2018/1624 of 23 October 2018, laying down implementing technical standards with regard to procedures and standard forms and templates for the provision of information for the purposes of resolution plans for credit institutions and investment firms pursuant to Directive 2014/59/EU of the European Parliament and of the Council and repealing Commission Implementing Regulation (EU) 2016/1066.

Directive 97/9/EC of the European Parliament and of the Council of 3 March 1997 on investor compensation schemes.

Securities Investor Protection Act of 1970, 15 U.S.C. §78aaa et seq. The initial level of coverage was USD 100.000, then increased for securities, and expanded to cash in 1980.

We are not referring to shares or bonds issued by the failing entity but shares and bonds owned by clients of that entity and kept in custody at that entity.

For instance, the liquidation of Lehman Brothers was initiated in September 2008 and closed 14 years later in September 2022.

More precisely the KA refers to significant financial institutions, thus including large insurance companies but excluding smaller banks. This choice reflects the compromise in the KA for which the resolvability assessment is to be carried at least for the largest groups. We choose to refer to banks instead, in line with the approach taken in the European Union and other jurisdictions, where the resolvability assessment is carried for all credit institutions.

We feel that in order to make the present section properly understood, we need to provide some clarifications about what we mean with policy operationalisation. References to policy operationalisation can be found across very different fields, from access to dental care (Harris 2013) to natural hazards prevention (Kumar 2020), passing through public transport (Milne et al. 2020), marine areas (Galparsoro et al. 2021b), health (Huber et al. 2016; Stucki et al. 2020), water (Lawson et al. 2020), homeland security (Murphy 2014), or sustainable aquaculture (Osmundsen et al. 2020).

Some of these authors (Harris 2013; Huber et al. 2016; Stucki et al. 2020) use the world ‘operationalisation’ to refer to the operationalisation of a variable at a given moment in the policy cycle, which is not the process we aim to study in this research. Even among those who use the word ‘operationalisation’ to describe a step in the policy cycle, there is no consensus around what it represents. For Kumar (Kumar 2020) this process covers all the aspects relating to the policy implementation and evaluation of its effects. Lawson et al. (2020) and Murphy (2014) intend operationalisation as a synonym of ‘phase-in’ or ‘implementation’ of the regulators’ expectations. As in Galparsoro et al. (2021a), the policy operationalisation is a sub-step of the formulation phase. A definition of policy operationalisation that is comparable to what we intend was provided by Dinica (2004), who describes it as the process by which (p.2) ‘the policy goals, means and schemes are specified in a way that implementing actors are able to work with them’. Other authors (Verschuere & Vancoppenolle, 2012) agree with the link to the implementation phase but emphasise the role of implementing actors. In brief, ‘operationalisation forms a blind spot in the current literature’. Taking note of this void, Mastenbroek, Treib, and Versluis interpret policy operationalisation as the passage between the legislation and the practical implementation, and they describe it as the process ‘making instruments applicable’. In this chapter we adhere to this approach and define policy operationalisation as the process by which legislation is transformed into implementable provisions. This process involves inter alia guidance documents, public statements, consultations, and informal interactions with the addresses, accompanied by the adoption of internal instructions and procedures.

Yet another ‘-ility’ word in the resolution lexicon. Separability refers the possibility of separating different group entities without causing a disruption of their critical functions. To visualise this concept, the reader can think about different types of cookies. Some can be fractioned in parts and shared with friends. Others, if we attempt, will just break apart in many pieces. We can say that the first type is ‘separable’, whereas the second one is not.

Note that we referred to years of planning, checks, and testing exercises. Ideally, all of that needs to take place before the resolution policy is effectively implemented in a crisis case. We want to stress this aspect, because when thinking about resolution regimes, the immediate association would be with crisis cases. These cases luckily appear only seldom, whereas resolution regimes entail an intense operationalisation process in a business as usual environment. We want to conclude this part on the concept of resolvability by emphasising this too often neglected aspect. We may in fact be prone to ask ourselves what are the consequences of having a resolution regime when a crisis materialises. Less evident, but certainly not less important, are however the changes in the market expectations and in the level of resilience to stress that years of crisis preparation have already produced in banks.

Differently from the on-site inspections carried in a business as usual situation, the BRRD clarifies that in the context of early intervention measures on-site inspections are to inform the resolution action and the valuation of assets and liabilities. In this particular case, such early intervention measure is likely to be performed in a case in which the situation is gradually heading towards resolution.

On this point see in particular the conclusions arising from Judgment of the General Court (Fourth Chamber, Extended Composition) of 12 October 2022, Francesca Corneli v European Central Bank, ECLI:EU:T:2022:627.

We need however to acknowledge that this is rather a hypothetical case. In reality, by definition systemically important banks (e.g. the O-SIIs recognised by the EBA) are institutions that pose financial risks to the system. It is therefore to be expected that for most of these banks resolution authorities anticipate to start a resolution action in case of failure. If willing, the reader can test this statement by checking what systemically significant banks disclose in their annual report in terms of the resolution strategy envisaged by resolution authorities.

It should be noted however that in the past it has not been infrequent that creditors of a liquidated bank received some form of compensation ex post at public expenses. This is however an element that cannot be anticipated at the time a resolution authority needs to assess the public interest and that rather pertains to the domain political considerations.

We refer here to the PIA Approach, but this was already the case in the first wave of resolution plans (see SRB 2016).

Note that the preliminary PIA is performed annually. Hence, whatever it is the timing at which a bank fails, this failure will happen no later than one year after the last preliminary PIA.

The KA also outlines the resolution tools for insurance, which are not covered by this chapter.

For this fictional example, we do not express any currencies, as it does not matter. The reader is free to assume that we reason in the currency he prefers. Oranges or apples would work equally well.

For this fictional example, we do not express any currencies, as it does not matter. The reader is free to assume that we reason in the currency he prefers. Oranges or apples would work equally well.

Furthermore, the difference between liability and instrument consists in the fact that a bank holds liabilities that are not necessarily issued financial instruments (e.g. deposits).

We give in this subsection some examples of what is considered within the outstanding amount for explanatory purposes. However, for the exact definition of outstanding amount, we recommend the reader to refer to the applicable legal framework and guidance by resolution authorities.

There is an extensive literature concerning the need for deposit guarantee, which spans over multiple decades starting from the debate of the adoption of the Glass-Stegall Act. The reader interested to get a better view on this topic may refer to the by now classical contributions by Diamond and Dybvig (1983), Diamond and Dybvig (1986), outlining an economic model of bank runs and discussing the role of deposit insurance in mitigating their effects. Costs and benefits of deposit insurance have been analysed for example by Carlstrom (1988), Hein (1992), Blair et al. (2006), or Anginer and Demirgüç-Kunt (2018).

Eligible deposits are the part of deposit that is not covered by a deposit guarantee scheme.

Please note that the term sale (or transfer) of business, now largely predominant in resolution regimes, is based on the EU BRRD language. American readers may instead be more familiar with the notion of purchase and assumption (see for example Giliberto & Varaiya, 1989; McGuire, 2012).

In reality, to convince existing shareholders to sell, an acquirer must be ready to put more on the plate and offer an additional mark up, but to keep this example simple we assume all shareholders are ready to sell at the market price.

In this case however the undepleted capital should be considered in the consideration paid by the buyer.

This limitation is made explicit in the EU BRRD framework. However, as the reader will see from its functioning, common sense would also arrive to the same conclusion.

It is worth noting that in most resolution regimes the staff members of the failing bank cannot be appointed in the bad bank, under the assumption that they (or at least those of them who remain) will need to keep the critical functions alive. For the record, given the limited life span of a bad bank, it should be expected that it would be rather hard to staff it, unless generous salaries are paid. This comes as an additional reason to use caution in recurring to this resolution tool.

This is a commonly accepted conclusion based on the fact (a) that deposits are remunerated at a lower interest than securities, (b) that their remuneration rates adapt more slowly to changes in market interest rates (Driscoll & Judson, 2013; Duquerroy et al., 2020), and (c) that they have been observed to display a relative stability in the medium-to-long term, with covered deposits also being relatively stable in times of crisis (Martin et al., 2016). For a descriptive comparison of deposits with other sources of funding, see also (Beau et al., 2014).

Attributable to the parent means in proportion to the percentage ownership of the parent. For example, if the parent owns the 75% of the resolution entity, only the 75% of surplus will be considered in this calculation.

To complement, lower intermediate requirements were set for 1 January 2019, corresponding to 16% of risk-weighted assets and 6% of leverage ratio exposures.

In this case also the Term Sheet considers lower intermediate targets, to be met by 1 January 2025. The intermediate minimum target levels are the same as for the rest of G-SIBs in 2019. The Term Sheet also included a clause on the anticipation of the timeline if the outstanding bonds of corporates and financial institutions reached an amount corresponding to the 55% of the GDP of the relevant jurisdiction. As of 2019 the threshold was not met. However, the FSB did not undergo an extensive review, considering the commitments of the Chinese G-SIBs to meet their minimum TLAC requirements ahead of 2025 (FSB, 2019).

An exception exists for debt instruments issued by a directly and fully owned funding entity prior to 1 January 2022, should resolution authorities conclude that they are fit to absorb losses.

Note however that in EU on the back of a common legal framework to be considered as instruments issued in a different jurisdiction are only those issued in a country that is not a Member State of the Union (i.e. third countries).

More correctly, to the instruments that the Term Sheet itself identifies as exclusions, which are covered deposits, sight and short -erm deposits, derivatives, derivative-linked debt instruments, liabilities that do not arise from a contract, such as tax liabilities, secured liabilities, and any other liabilities to which a jurisdiction confers special protection in insolvency.

A key example is the United Kingdom, where the BRRD architecture was maintained in their legislation after the exit from the European Union. The United Kingdom also unilaterally implemented the update of the BRRD of 2019, which effectively concerned MREL mainly. On the developments of UK regulation after Brexit, refer to James and Quaglia (2020).

The reader interested to get a more in-depth knowledge of the MREL can refer directly to the BRRD or read the SRB MREL Policy, updated annually and available on the institutional website. At the time this chapter is drafted, we refer to the 2023 version, available at https://​www.​srb.​europa.​eu/​en/​content/​mrel.

Note however that this is only an expectation. In a real case, a bank earmarked for liquidation can ultimately be resolved in case new evidence appears to support this conclusion. A recent and notable example is the one of the UK subsidiary of Silicon Valley Bank. Although considered a liquidation entity in resolution planning, in March 2023 – when the parent entity failed – it was subject to resolution on a standalone basis and acquired by HSBC.

Pillar 2 requirements are disclosed by the European Central Bank on the institutional website. This average has been calculated by us based on the 8 February 2023 update.

As a side note for the reader, we want to highlight that one should not be surprised by the difference in the percentages, almost 20.5% for the risk-based requirement in addition to buffers and 6% for the non-risk-based requirement. Risk-weighted assets are typically (and should be) very low compared to total assets, as some assets are accounted with weights lower than 100%, in some cases even equal to 0%, for exposures with very low credit risk. By contrast, leverage ratio exposures are roughly equivalent or a bit higher than total assets. In practical terms, except very specific cases, the absolute amounts of the risk-based and non-risk-based MREL requirements are typically not too far away one of the other.

Note that the amount calculated based on prudential requirements and lower values of risk-weighted assets and leverage exposures will be added on top of the loss absorption amount and then expressed as a percentage of reported risk-weighted assets and leverage ratio exposures without adjustments.

The reader will be familiar in these times with the depreciation of bonds following a rise in interest rates. As their market value decreases, an effective loss will materialise only if the holder is forced to sale. The sale of widely depreciated long-term US Treasury bonds to cover the withdrawn of deposits was the main cause of the failure of Silicon Valley Bank in the State of California. However, if a holder is not forced to sale a depreciated bond, this will maintain by contract its book value and will be repaid at maturity.

Maintain by contract its book value and it will be repaid at maturity, which means that risk-based and non-risk-based exposures used for the calibration of the recapitalisation amount will be multiplied by (1 – the scaling factor), i.e. by a value between 75 and 85%.

Pay attention to the terminology. Restructuring plan has a precise meaning in relation to the State Aid Framework. It is not a synonym of neither recovery plan, resolution plan, nor business reorganisation plan. The latter, in particular, is the plan prepared for the reorganisation after the implementation of a bail-in tool. In practice, restructuring and reorganisation may be often confused in the use by non-specialists or people not familiar with the EU legislative framework.

We omit in this section the specific adjustments due to strategies entailing multiple points of entry, given their inherent level of complexity. At a high level, we can however say that these adjustments reflect the same logic of the TLAC deductions.

Lower intermediate requirements had to be met as of 1 January 2022. Resolution authorities have however faculty of deviating from the default transition period for specific reasons, mainly related to the timing for market access and the availability of MREL resources.

The reader interested to have a precise view on MREL-eligible liabilities can refer to Article 45b BRRD.

Note however that for the internal MREL, due to the need of performing an internal write down and conversion limiting potential no-creditor-worse-off issues, subordinated instruments are better placed to comply with the requirement.

The minimum level of subordination is also referred to as Pillar 1 subordination, where Pillar 1 stands for G-SIBs and Top-Tier banks. We prefer however not to use this definition, as it may create confusion with Pillar 1 prudential requirement. Furthermore, the definition of Pillar 1 subordination has no legal relevance, but it is only a practice.

As a practical consequence, the amount corresponding to 8% of total liabilities and own funds will be expressed in full as a percentage of leverage ratio exposures, but as a percentage of risk-weighted assets the same amount will be deducted of the amount corresponding to the combined buffer requirement. This way, also on the risk-weighted leg, it is possible to conciliate the provision of an overall level of subordination corresponding to 8% of total liabilities and own funds with the subordination being an integral part of the MREL requirement, to be met in addition to capital buffers.

We need to add that the shares of a cooperative bank are not profit-oriented, which means that profits are not distributed, and they are not listed. In principle, to be client of a cooperative bank, a person needs also to be a member of the bank, by the purchase of a share. The holdings of shares of cooperative banks are nominative and limited in number (in some cases even to a single share per client). Some jurisdictions also have (or had in the past) limitations as to the residence of the members, which should not be outside a certain maximum radius from the headquarters of the cooperative bank. This is because cooperative banks, in the tradition opened in the XIX by Friedrich Wilhelm Raiffeisen (cooperative rural and artisanal banks) and Franz Hermann Schulze-Delitzsch (people banks for the middle class), rather than profit-making organisation are treated by the legislation as a form of self-organisation of certain social classes to access credit services. In many EU countries people banks have been gradually assimilated by law to ordinary commercial banks and lost their cooperative corporate form. Cooperative banks coming from the rural and artisanal tradition are still in existence in many EU countries, with notable examples in Austria, France, Germany and Italy, although they have formed (or have been forced to form) large confederations at a national level. It is worth noting that the French Crédit Agricole is also classified as a G-SIB.

It should be noted that when this treatment is granted, it is applied strictly to affiliated entities. Subsidiaries of these entities that are not themselves directly affiliated to the cooperative group, will in any case need to comply with internal MREL requirements, should the minimum thresholds for a requirement be met.

This goes hand in hand with the possibility of applying a coordinated bail-in simultaneously to all the entities in the cooperative group, where within each rank own funds or eligible liabilities are written down or converted on a pro rata basis throughout the whole group.