Survey of Resolution and Restructuring in Europe: Pre- and Post-BRRD

What motivated policymakers to resolve the institution?

In all of the cases in our survey, both pre-BRRD and post-BRRD, the authorities had determined that the targeted firm was likely to fail and had exhausted other available measures. In response to this situation, the authorities settled on direct intervention with the view that the targeted firm should not be handled through a traditional insolvency. In the European context, Article 32 of the BRRD, adopted in 2014, codified these three primary conditions for the commencement of a resolution action and provided standards for meeting them:

     (a)  The institution has failed or is likely to fail (the FOLTF determination).

     (b)  There is no reasonable prospect that any alternative private sector measures will prevent the failure of the institution within a reasonable time frame.

     (c)  Resolution is in the public interest.

Under the BRRD, in most cases, the supervisory authority makes the first of these determinations, in consultation with the resolution authority. The resolution authority makes the second and third.

A potential pitfall in weighing whether to initiate resolution action is the risk of costly delay. Early action can improve the likelihood of maintaining critical functions and minimizing the impact of an institution’s failure on the financial system and broader economy. It also reduces the amount of time during which an institution’s bail-in-able base can erode through depositor withdrawals and other redemptions, as occurred in Cyprus in 2011–2013 due to a delay between attempted recapitalizations and resolution (World Bank Group 2016a).

FOLTF Determinations and the Role of Valuation

To be placed into resolution under the BRRD, an institution must have failed or been likely to fail. Per Article 32(4) of the BRRD, this FOLTF determination requires that one or more of the following conditions be met:

• The institution is facing a loss of its banking license or other official authorization.
• The institution’s assets are or will soon be less than its liabilities.
• The institution is or will soon be unable to pay its debts as they come due.
• The institution requires extraordinary public financial support (subject to certain exceptions).

These conditions are more flexible than, for example, a specific capital ratio trigger, providing authorities with discretion to intervene early and avoid the potential for costly delays as mentioned above (World Bank Group 2016b).

A key component of the decision to initiate resolution action is an understanding of the magnitude of the losses the financial institution could face in the near future. In the words of Gleeson and Guynn (2016), two leading bankruptcy attorneys, “you can’t know how to fill a potential hole if you can form no reasonable estimate at all of the size of the hole” (Gleeson and Guynn 2016, 15). For resolution to be successful, the entities that emerge as going concerns—for instance, any financial institutions that have acquired assets and liabilities of the failed firm—must be solvent and financial market participants must not be uncertain as to their solvency (Gleeson and Guynn 2016). Effective valuation of a firm’s assets and liabilities is thus an important early step in determining whether and how to resolve that firm.

The European Banking Authority—the EU-wide supervisory authority for banking—has created a framework for distinguishing among the valuations required by the BRRD (SRB 2019):

Valuation 1—to inform the determination of whether the conditions for resolution are met (BRRD Article 36[4][a]).

Valuation 2—to inform the choice of resolution option to be adopted (BRRD Article 36[4][b]).

Valuation 3—to inform the “no creditor worse off” determination, a concept described in more detail in Key Design Decision No. 9, Treatment of Creditors and Equity Holders (BRRD Article 74).^FNFor further discussion of the relationship between the three valuations, see World Bank Group (2016b).

Valuations under the BRRD must be “fair, prudent and realistic” and be carried out by “a person independent from any public authority, including the resolution authority, and the institution or entity” (Article 36[1]). In exigent circumstances, a resolution authority may conduct a provisional valuation subject to subsequent confirmation by an independent valuation expert (Article 36[2]).

A full discussion of the technical approaches to performing valuations in preparation for resolution is beyond the scope of this paper (see, for example, EBA [2017]) for a discussion of the technical standards for valuation in the BRRD context). However, the challenge associated with performing such valuations in the condensed time frame of a rapidly failing institution should not be ignored.^FNIn its Valuation 2 report for Banco Popular, Deloitte noted that it had been given 12 days to perform a valuation exercise that would typically take six weeks to complete (SRB 2017). This is especially so when the source of an institution’s trouble is improper lending or previously undisclosed liabilities, as was the case with Portugal’s BES. Not surprisingly, then, allegations of improper valuations are often at the heart of creditor and shareholder litigation in response to resolutions, as was the case in Banco Popular.

Mechanisms for addressing the differences between estimated valuations and the actual economic outcomes that are achieved as a resolution unfolds are discussed in Key Design Decision No. 9, Treatment of Creditors and Equity Holders. There are also steps that could potentially be taken to provide additional time for valuations to occur. It has been proposed that a “moratorium tool” that would prevent creditors from running on an institution approaching resolution could provide for a less rushed valuation process (see, for example, Lastra and Olivares-Caminal [2018]). However, the potential benefits of such a tool must be weighed against the risk that the threat of a moratorium could accelerate the run on an institution. While Heta issued a moratorium on its debt to provide authorities additional time to assess the situation before initiating bail-in, Heta was the asset management vehicle created out of an earlier bank resolution and had no deposits. Additionally, the announcement of the moratorium resulted in a collapse in the price for Heta bonds and a wave of litigation and contributed to a protracted resolution. In Iceland, three failed banks were split into new banks with government capital support, which continued to operate, and the legacy old banks, which were put under moratoriums. Still, uncertainty about further obligations the new banks may have had to the old banks remained for a year, affecting their ability to conduct business.

Other Options

The BRRD requires that a resolution action be initiated only where alternative private sector solutions are not available. More broadly, given its challenges and risks, resolution is typically something authorities do after exhausting other options. Key Design Decision No. 2, Part of a Package, discusses interventions authorities undertook before resolution in an attempt to stabilize the institutions that are the focus of the case studies reviewed here. Resolutions also often occur following unsuccessful attempts by private parties to inject capital into or acquire troubled institutions, as was the case in Roskilde, Parex, and JAK.

Public Interest

As noted in the Introduction, resolution is generally intended to prevent harm to the financial system and broader economy. Within the BRRD framework, resolution actions are to be initiated only when in the public interest. Article 31(2) of the BRRD defines public interest to include:

     (a)  The continuity of critical functions;

     (b)  The avoidance of significant adverse effects on the financial system, particularly via contagion;

     (c)  The protection of public funds by minimizing extraordinary public support;

     (d)  The protection of depositors; and

     (e)  The protection of client funds and assets.

In the case studies reviewed, both those pre- and post-BRRD, authorities typically cited one or more of these considerations as justifying the need to resolve the institution in question. A determination that the resolution of a given institution is in the public interest is a context-specific analysis. It can depend on factors such as the size and composition of a country’s banking system, the primary activities of the institution and how they fit into that system, and whether the activities can be substituted for. In the case of Banco Popular, for example, the nature of the small and medium-sized enterprise (SME) lending market (in both Spain and Portugal) and the payments system (in Spain) were such that authorities feared that Popular’s collapse would significantly undermine these functions.

Was the resolution accompanied by other interventions intended to address the problems that caused the institution to fail?

Authorities typically resort to resolution only after prior attempts to support the institution using other tools have failed. As noted in the Introduction, new resolution regimes have been accompanied by the strengthening of tools to be used in the recovery phase, before an institution has reached the point of needing to be resolved. The Key Attributes call for recovery measures including private sector recapitalization, capital conservation via suspension of dividends and variable compensation, the sale of subsidiaries and the spin-off of business units, voluntary conversion of debt into equity, and diversification of funding sources. All of these measures appeared at least a few times in the cases studied for this paper.

Additionally, interventions often accompany resolution itself, particularly when the authorities are concerned about the potential systemic implications of a disorderly failure. In about half of the cases we reviewed, the central bank also provided emergency lending to the troubled bank in order to ensure access to adequate liquidity. The resolution may even give rise to the need for such interventions. In the Roskilde case, a transfer of assets triggered an early termination right resulting in creditor withdrawals that required emergency liquidity assistance (see Key Design Decision No. 10, Treatment of Clients).

A key consideration when pairing other interventions with a resolution is how the other interventions will interact with the creditor hierarchy. Given the desire to ensure creditors and shareholders absorb losses before taxpayers, the public is often given priority over certain other creditors in recouping amounts from interventions accompanying resolution. However, structures can vary. In the case of Dunfermline, for example, the British Treasury contributed approximately 1.6 billion pounds sterling (GBP) as part of the resolution, splitting its claim into two so that GBP 0.5 billion ranked equal to other unsecured creditors while the remaining GBP 1.1 billion was subordinated to unsecured creditors (but senior to other subordinated note holders).

Most of the cases studied involved some amount of government recapitalization. Although one of principal purposes of resolution as articulated in the Key Attributes is to provide an alternative to taxpayer exposure to losses for solvency support, authorities continued to see such support as sometimes required to achieve a necessary outcome of resolution—the elimination of uncertainty about an institution’s solvency. Under the BRRD, the use of public funds for recapitalization is not prohibited but rather is subject to a requirement that shareholders and creditors first absorb losses equal to 8% of the institution’s total liabilities. We review recapitalization cases in a separate survey.

What was the legal basis for the resolution and how did it influence how the resolution was done?

The resolution of failed institutions typically affects the property rights of individuals, as when assets and liabilities are transferred out of the institution or the claims of shareholders or creditors are written down (Gleeson and Guynn 2016). Thus, to a degree that is even more true than with other crisis interventions, a clear and complete framework providing policymakers with the necessary legal authorities to accomplish resolution is essential.

As noted in the Introduction, in May 2014, the EU adopted the BRRD to provide a common set of rules for resolution among member states.^FNThese common rules represent a “minimum harmonised set of resolution tools and powers.” (European Parliament and Council of the European Union 2014, 197). As the word “minimum” suggests, under the BRRD, member states have the ability to adopt additional tools and powers at the national level, provided that such additional tools and powers are consistent with the principles and objectives of the BRRD and other EU legal requirements. The degree to which individual countries have gone past the minimum requirements of the BRRD is beyond the scope of this paper (World Bank Group 2016b). This common set of rules sits alongside existing national laws providing for “normal insolvency proceedings.” (European Parliament and Council of the European Union 2014, 197). Under the BRRD, such normal insolvency proceedings remain the default approach to dealing with failed institutions, with resolution per the BRRD considered only when it is in the public interest to deviate from this default approach, as discussed in Key Design Decision No. 1, Purpose.

Although member states had until December 31, 2014, to transpose the Directive into their national law and until January 1, 2016, before bail-in requirements had to become effective, the transition from a pre-BRRD framework to a post-BRRD framework was even more gradual. Some member states had already adopted certain elements of the BRRD (including loss absorption requirements) even before its enactment. This change resulted from best practices, the ongoing negotiations over the BRRD, and EU State Aid rules (World Bank Group 2016a). The case studies reviewed thus represent a mix of resolutions undertaken at various stages in the BRRD’s implementation and resolutions undertaken outside the BRRD (whether because in non-EU countries or in EU countries before any BRRD implementation had taken place). Figure 2 summarizes the legislation authorities relied on in the case reviewed—whether existing or new and whether adopted pursuant to the BRRD or not.

Discussion of how legal frameworks influenced the design of specific elements of resolution in the case studies reviewed is included across the various Key Design Decisions analyzed in this paper. The discussion in this Key Design Decision focuses on case studies in which the legal framework proved inadequate. An inadequate legal framework lacking powers necessary for effective resolution can require hurried amendments to provide missing powers or result in delays and changes to the design of an intervention. Cyprus, for example, had to pass a new resolution law on an urgent basis in March 2013 to provide authorities with the ability to resolve Laiki Bank.

In several of our cases, authorities lacked the ability to impose losses on stakeholders, tying their hands in the resolution process. In Belgium, authorities lacked the necessary authority to override shareholder rights, necessitating costly and time-consuming negotiations with shareholders before the restructuring could proceed. In the cases of SNS Reaal and BES, the fact that the BRRD had not yet been adopted left Dutch and Portuguese authorities without an explicit legal basis for such powers as the ability to impose losses on creditors. In Denmark, the absence of a legal and organizational framework for orderly resolution resulted in authorities’ needing to rely on existing legislation to resolve Roskilde while concurrently undergoing financial sector reform.

Who was responsible for administering the resolution?

One of the lessons of the GFC is that existing entities—bank supervisors, central banks, and ministries of finance—weren’t necessarily best positioned to handle failed institutions (World Bank Group 2016b). When the objectives associated with resolution are layered on top of these entities’ other mandates, the possibility for conflicts of interest arise. For instance, a bank regulator may not want to initiate proceedings that would call into question the effectiveness of its supervisory function. In various cases, the government, or specifically ministry of finance, sought to delay resolution due to short-term political considerations (World Bank Group 2016b). For example, in the Cyprus case, both the government and the Eurogroup, which was considering official support for the country, sought to delay agreement on the resolution of their banking crisis for several months, until after a scheduled election.

Key Attribute 2.5 calls for the designation of a resolution authority with operational independence (yet still subject to transparency and accountability mechanisms). Under Article 3 of the BRRD, member states are required to designate a resolution authority that can include the central bank, the ministry of finance, or (exceptionally) bank supervisors, provided that adequate structural arrangements must be in place to ensure operational independence and avoid conflicts of interest. In the eurozone, the Single Resolution Board (SRB) acts as the central resolution authority. Globally, resolution powers for deposit insurers have also significantly increased, with the role played by the FDIC in the United States a model (Defina 2021).

Even where the resolution authority is not housed within the central bank, ministry of finance, supervisory authority, or deposit insurer, these entities have a role to play in resolution. As noted in Key Design Decision No. 1, Purpose, the supervisory authority is generally responsible for the FOLTF determination that triggers resolution. The central bank and ministry of finance may be needed to provide additional assistance as described in Key Design Decision No. 2, Part of Package.

How was the authority responsible for administering the resolution governed and what mechanisms existed for holding it accountable?

As noted in Key Design Decision No. 4, Administration, the Key Attributes provide for a balance between a designated resolution authority with operational independence on the one hand and mechanisms to ensure that this resolution authority is accountable for its decisions on the other. Ex post reviews of resolution decisions and their implementation were one common approach to promoting accountability in the cases reviewed. Such reviews often occurred at the behest of the legislature. Once again, the US is something of a model, with the Federal Deposit Insurance Act requiring a review any time the Deposit Insurance Fund incurs a loss of USD 50 million or more (FDIC 2018, 38[k]). That said, if the legislature (or other body mandating the review) is not committed to undertaking a sufficiently comprehensive review, the results can give a false impression of accountability. In Latvia, the commission tasked by parliament with investigating the resolution of Parex ultimately issued an incomplete report after parliament declined to extend the length of the investigation. If not handled appropriately, reviews can also raise questions of independence, objectivity, and political motivations. In Cyprus, a report by the central bank in the immediate aftermath of the resolution of Laiki Bank absolved the central bank’s leadership of wrongdoing. At around the same time, Cyprus’s president launched his own investigation after being critical of the central bank.

What did authorities communicate about the need for resolution and how it would unfold?

The fundamental goal of resolution is to promote credibility and confidence in the banking system. A resolution may result in the closure, sale, or reopening of the failed institution, or the sale of its assets to a bridge bank on the way to one of those outcomes. But whatever path the authorities take, insured depositors of the failed institution must have confidence that their claims will be protected. Throughout the process, the institution must be able to perform its essential functions for the economy. And if the authorities do impose losses on some creditors, they must do so in a way that doesn’t erode the confidence of participants in the broader financial system. Given the role communication can play in either bolstering or undermining credibility and confidence, effective communication is essential in the resolution context. Oftentimes, effective communication will require balancing the objectives of credibility and confidence. Leaning too far in either direction may cause problems, as when Hungarian authorities insisted that the resolution of MKB was being undertaken to improve profitability despite clear evidence of insolvency, thereby potentially jeopardizing their credibility. Similarly, Portuguese authorities suggested that BES was being split into a good bank and a bad bank at no risk to the taxpayer. Continued losses at the so-called good bank and the decision to transfer some senior debt holders back to the “bad” bank (as discussed in more detail in Key Design Decision No. 9, Treatment of Creditors and Equity Holders) resulted in criticism from the European Commission and ultimately undermined authorities’ credibility. Opposition politicians also accused Portuguese policymakers of misrepresenting Portugal’s potential exposure under a capital backstop provided in connection with the sale of the “good” bank, resulting in the issuance of a clarifying statement.

Even within robust resolution frameworks, significant discretion is inherent in the process. Providing transparency into the decision-making involved in the use of this discretion can provide greater clarity for market participants about how similar situations might be handled in the future. A lack of such clarity has the potential to undermine confidence. This is especially true when it comes to issues such as the creditor hierarchy. If policymakers’ communications leave market participants uncertain as to the treatment of their assets, panic may grip the institution and spread to the broader financial system. In Cyprus, for example, rumors of potential depositor haircuts and a failed legislative bill to impose such haircuts resulted in further runs on an already struggling Laiki Bank. A lack of communication and disclosure may also hinder accountability. This circumstance can be a particular challenge in the case of vehicles used to wind down the remaining assets of a failed institution over a period of many years. Limited disclosure around the assets transferred to wind down vehicles and their ultimate disposition occurred in several countries, including Cyprus, Hungary, and Portugal.

How was the resolution funded?

Although the new resolution regimes adopted in the wake of the GFC are intended to provide policymakers with a third option beyond chaotic insolvency and taxpayer-funded bailouts, resolution has often required a commitment of significant official funds. Determining where those funds will come from is thus an essential consideration in any resolution.

The objective of an effective resolution regime as defined by the Key Attributes is to “make feasible the resolution of financial institutions without severe systemic disruption and without exposing taxpayers to loss [emphasis added].” Key Attribute 6.2 provides that any funding extended by the authorities to accomplish orderly resolution should be temporary in nature, with losses covered by equity holders and unsecured creditors or, if necessary, the broader financial system. Under Key Attribute 6.3, broader funding can be accomplished ex ante, via a privately financed deposit insurance or resolution fund, or ex post, via a funding mechanism with recovery from the financial sector. Article 103 of the BRRD requires member states to establish a national resolution fund with the ability to raise ex ante contributions from banks based on their size and risk profile. These funds are to meet a minimum funding threshold of 1% of insured deposits by December 31, 2024. If the proceeds of ex ante contributions are insufficient to cover losses resulting from resolution funding, the national resolution funds must require extraordinary ex post contributions from banks to make up the difference. Although distinct from national deposit insurance funds, national resolution funds may use the same administrative structure per Article 100 of the BRRD.

In the eurozone, the requirement to establish a national resolution fund is satisfied by the Single Resolution Mechanism (SRM) and its Single Resolution Fund (SRF). The SRF set a funding target of 1% of covered deposits to be reached by December 31, 2023. In July 2023, the SRF announced that it would reach this target by year’s end, with a balance of more than EUR 77 billion (SRB 2023).

Adequate ex ante funding can be a challenge early in the life of a resolution fund, before it has been able to build up its balance. In the Italy case, for example, the resolution fund had to rely on market-rate borrowings from three major Italian banks to meet its needs in connection with the resolution of four banks in 2015. The resolution fund then repaid these borrowings using both ordinary contributions and extraordinary ex post contributions, a process that was not complete until 2021.

In addition to having potential administrative roles in resolution as described in Key Design Decision No. 4, Administration, deposit insurance systems also provide a potential source of resolution funding. If instead of being used to reimburse depositors once losses have already occurred, the same funds could be deployed earlier to minimize the extent of the losses, deposit insurance systems should be able to expend funds in this way. This approach is consistent with, the International Association of Deposit Insurers’ Core Principles for Effective Deposit Insurance Systems (the Core Principles), which provide that the contribution of deposit insurance funds to resolution be “restricted to the costs the deposit insurer would otherwise have incurred in a payout of insured depositors in a liquidation net of expected recoveries” (IADI 2014).

In addition to minimizing losses, the proactive use of deposit insurance may also promote other objectives such as providing depositors continued access to their money. In the case of Dunfermline, for example, UK authorities allowed the use of the Financial Services Compensation Scheme (FSCS) to facilitate a sale of business to a purchaser, thereby ensuring immediate access to deposits as compared with a potentially drawn-out reimbursement process. Under this approach, the UK Treasury made a GBP 1.6 billion payment to the purchaser representing the difference in the assets and liabilities it would take on. The FSCS in turn agreed to reimburse the UK Treasury for any amounts not recovered from the liquidation of the remaining business (subject to a cap equal to the amount the FSCS would have had to pay to insured depositors in liquidation net of recoveries).

Consistent with the Core Principles, Article 109 of the BRRD limits the amount that a deposit insurance system can contribute to a resolution to the amount that it would have had to pay out to insured depositors under a normal insolvency proceeding. However, because insured deposits receive priority in normal insolvency proceedings under the BRRD and are thus unlikely to sustain losses, as a practical matter, deposit insurance systems typically would not be a viable source of resolution funding under the BRRD (Gleeson and Guynn 2016). One exception among the case studies reviewed is JAK, where losses were significant enough to affect insured deposits, requiring a contribution from the deposit insurance fund.

Roskilde presents a unique approach among the cases reviewed, with a private sector banking group collaborating with the central bank on the resolution and providing a source of funds.

In four of the cases we reviewed, the International Monetary Fund or other international bodies provided significant financial support to the sovereign. However, in Cyprus, the funds were explicitly not to be used directly to support bank resolution.

What tools were used to accomplish the resolution?

Given the wide range of financial institutions, no one approach to resolution will be appropriate in all situations. Rather, policymakers should have at their disposal a suite of tools to be combined as needed based on the particular case. As will be discussed, this analysis is highly context dependent.

The primary tools to be used for resolution are summarized in Figure 3. Chapter IV of the BRRD establishes and defines these tools.

In this paper, we group “bail-in” and “write-down” into the same category as each involves a reduction in principal or conversion into equity, but it is important to note that under the BRRD, they are separate powers with key differences. A full discussion of these differences is beyond the scope of this paper, but in brief, write-down can also occur outside of resolution (and indeed, under Article 59[3], must occur once an institution reaches the point of nonviability) and is specific to capital instruments, whereas bail-in occurs in resolution and can be applied to liabilities generally. Thus, under the BRRD, write-down can be used as a recovery tool before resolution. This differs from the Key Attributes, which focus on the use of bail-in as a resolution tool.

As shown in Figure 4, authorities typically use these tools in combination. One tool can also sometimes be used to accomplish the outcome typically associated with another tool. For example, if a given liability cannot be bailed in or written down (for example, because the legal framework and instruments’ terms do not allow for it), the use of other tools to transfer assets out of the institution while leaving the liability behind can result in the liability being bailed in (that is, there being nothing remaining in the institution to satisfy the liability). For example, in the case study of the four Italian banks, of approximately EUR 800 million in total subordinated debt, only about EUR 500 million was eligible to be written down directly pursuant to Article 59 of the BRRD. To impose losses on the remaining EUR 300 million, authorities left it behind in the residual entities for liquidation. Similarly, Iceland split three failed banks into new banks, with performing domestic assets and domestic deposits, and the old legacy bank, whose liabilities vastly exceeded the value of assets. The banks’ original equity and subordinated debt holders were wiped out, while legacy senior creditors ultimately lost about three-quarters of their investment as the legacy banks sold assets at a loss.

Source: Authors’ compilation.

The specific tools that are appropriate for a given resolution are highly dependent on the institution and the extent of its difficulties. As outlined in Key Design Decision No. 1, Purpose, a valuation of the institution is necessary to inform the approach to be taken. For instance, whether the use of bail-in or write-down is itself sufficient to recapitalize a failed institution depends on whether the institution has enough bail-in-able debt relative to the size of its losses. Finding this balance may be particularly difficult to achieve in economies where banks are heavily dependent on deposits and capital markets are not sufficiently developed for the issuance of bail-in-able debt (Richards 2020). As discussed in Key Design Decision No. 9, Treatment of Creditors and Equity Holders, who holds this debt also matters. It is important to note that among the case studies reviewed, the use of bail-in or write-down was relatively concentrated among institutions in what may have been a “sweet spot”—large enough to issue bail-in-able debt, not so large that creditor bail-ins or write-downs would threaten financial stability.

The approach to resolution is also dependent on the broader context of the financial system. For example, the use of the sale of business tool requires the existence of viable buyers. So too does the bridge institution tool when used in anticipation of an eventual sale (which is typically done to provide a longer time frame in which to market the failed institution to buyers and to give them an opportunity to conduct diligence). Authorities were able to resolve Banco Popular overnight with no direct cost because of Santander’s willingness and ability to purchase it. In a financial system with no viable buyers, the sale of business tool becomes unworkable—either no sale will occur (as was the case, for example, with JAK and BES) or, perhaps even worse, the sale may take place to a buyer that is itself troubled (as was the case, for example, with the sale of Laiki Bank’s businesses to Bank of Cyprus and Piraeus and the sale of ATE to Piraeus). Among the case studies reviewed, it is interesting to note the significant role of private equity firms as purchasers in systems that may have lacked domestic institutions capable of acquiring resolved banks.

Conditions in the broader financial system can also drive the use of the asset separation tool. Whereas bridge institutions are intended to be short term (under the BRRD, allowed to exist for no more than two years), asset management vehicles typically allow for a longer time frame. This longer horizon may be important in a situation in which rapid asset sales could further destabilize markets or result in lower recoveries than would otherwise be possible. Asset separation also offers greater flexibility to authorities because asset management vehicles typically don’t require banking licenses or minimum capital levels given that they exist solely to dispose of assets.

How were the claims of creditors and equity holders treated during the resolution?

A major question any time an institution is facing failure is how the institution’s losses will be absorbed and thus what will happen to the claims of its creditors and equity holders. One of the chief motivations behind post-GFC resolution regimes such as the BRRD was the desire for stakeholders in a failed institution—rather than taxpayers— to bear the losses associated with that failure. Under the BRRD, government recapitalization and temporary public ownership tools may be used only after shareholders and creditors bear losses equivalent to 8% of the bank’s total liabilities.

As discussed in the Introduction, a fundamental difference between financial firms and nonfinancial companies when it comes to insolvency involves the nature of their creditors. Nonfinancial companies have both commercial and financial creditors. Traditional insolvency proceedings can enable an insolvent nonfinancial company to continue operating by allowing payments to commercial creditors while delaying payment to financial creditors. Given the nature of financial firms, there is no such distinction between commercial and financial creditors, and thus no ability to keep firms in operation during a traditional insolvency proceeding.

One function of resolution regimes is to divide the financial liabilities of a firm into two categories: operating liabilities such as deposits and capital structure liabilities such as subordinated debt (Gleeson and Guynn 2016). Resolution regimes generally seek to protect the former type of liabilities because of their nature (for example, because authorities typically want to safeguard depositors from loss) and because losses on those types of liabilities could destabilize the system (say, because depositor losses may result in a broader run on the financial system). The latter type of liabilities includes those intended to absorb losses through bail-in or write-down if a firm becomes troubled.

Problems can arise when the demarcation between these categories becomes blurred or cannot be adhered to. In Italy, a significant amount of bail-in-able debt was held by retail investors, which made authorities reluctant to trigger bail-in.^FNItalian authorities bailed-in retail investors but ultimately set up a fund to compensate them for their losses. In Austria, the existence of a state guarantee on debt subject to bail-in complicated the resolution of Heta and ultimately imposed significant public costs despite the bail-in.^FNAlthough the BRRD itself does not provide for the bailing in of liabilities secured by third-party collateral such as a state guarantee, the legislation transposing the BRRD into Austrian law did allow for this outcome. Bail-in can also be more difficult if bail-in-able debt is held by other financial market participants in amounts that make it a vector for contagion. An additional potential source of contagion may occur if the holders of liabilities thought to be in the protected category face losses. In Cyprus, the bail-in of depositors at Laiki Bank and Bank of Cyprus undermined confidence in the broader banking system and caused some depositors to default on their own obligations, resulting in new nonperforming loans. The treatment of Credit Suisse’s additional Tier 1 (AT1) bondholders as junior to equity holders resulted in a freezing up of the market for such instruments (although it now seems to have rebounded) (Vossos, Ramnarayan, and Gutscher 2023). In the pre-BRRD case of SNS Reaal, authorities had the ability to bail-in senior debt holders but decided not to, citing contagion risk and the heavy reliance of Dutch banks on unsecured funding markets. Bail-in can also create a challenge when aggrieved former creditors become involuntary equity holders, as was the case with the Bank of Cyprus.

The valuation process described in Key Design Decision No. 1, Purpose, drives the initial treatment of creditors and equity holders during resolution. Given the uncertainty inherent in these valuations, an important consideration in determining the treatment of creditors and equity holders is the mechanism by which adjustments are made for the differences that emerge between estimated valuations and the actual economic outcomes that are achieved as a resolution unfolds. If creditors suffer losses based on an estimation of the value of an institution’s assets that turns out to be too low, the BRRD seeks to establish a mechanism for compensating them when the assets are ultimately worth more than expected. This is especially true in government recapitalizations, when valuations are conservative by their nature given the need for the entity to emerge from resolution solvent and credible. In the case studies reviewed, any adjustments were almost always in favor of creditors. (An exception being BES, where certain senior bonds had to be transferred from the bridge bank back to BES, where they would suffer unexpected losses, triggering litigation and creating uncertainty among senior bond holders about the different treatment of creditors at the same level of the creditor hierarchy). In addition to providing equity in new entities pursuant to bail-in or write-down, authorities used a variety of other mechanisms to accomplish these adjustments, including earn-out provisions (Roskilde), contingent bonds (Iceland), and write-backs (Heta).

One key consideration regarding the treatment of creditors is the requirement contained in many resolution regimes, consistent with Key Attribute 5.2 and with the BRRD, that they be left no worse off under resolution than they would have been in a traditional insolvency proceeding. This “no creditor worse off than in liquidation” (NCWOL) principle is set forth in Article 73 and is evaluated using Valuation 3 as described in Key Design Decision No. 1, Purpose. An assumption of resolution regimes is that they will generally result in less loss of value; and thus, the “no creditor worse off” principle will not typically be in danger of being violated, partly because models assume that the process of selling assets in liquidation would deflate their value. This was true in the case studies reviewed, with the exception of JAK, where a small number of creditors received NCWOL compensation.

The treatment of creditors and equity holders in resolution can also depend on whether and how a jurisdiction has adopted a system of depositor preference. Under such systems, depositors received priority over other unsecured senior creditors in insolvency proceedings, either in unlimited amounts or up to the level of deposit insurance coverage. On the one hand, depositor preference can facilitate resolution by enabling the transfer of all deposits out of a failed institution without violating the NCWOL principle because such deposits would have received priority anyway (Gleeson and Guynn 2016). On the other hand, knowing that they will rank behind depositors in the event of insolvency could cause other senior unsecured creditors to run, reducing the pool of creditors that can absorb losses (Clifford Chance LLP 2023). As noted in Key Design Decision No. 7, Source and Size of Funding, it may also limit the ability of a deposit insurance fund to contribute to resolutions because such contributions typically cannot exceed what the fund would have expended in liquidation, an amount that depositor preference reduces or eliminates altogether.

Finally, it is important to note that one of the potential outcomes of a credible resolution regime is that by creating clarity about the losses to which creditors and equity holders may be exposed, the regime incentivizes stakeholders to agree to solutions before an institution enters resolution. Knowing what will happen to their claims in resolution, creditors may agree to a voluntary writing down of their holdings through “liability management exercises.” This happened in the case of Piraeus, where creditors faced with the prospect of a total loss via bail-in agreed to exchange their debt for shares worth 100% of par (for Tier 2 and senior notes) or 50% of par (for Tier 1 notes). Such “consensual bail-ins” may have advantages including the ability for them to unfold on a less urgent time frame and the reduced likelihood of legal challenge (World Bank Group 2016a).

How were the institution’s clients treated during the resolution?

Authorities must also consider the effects of a resolution on an institution’s clients. Entering a resolution proceeding could result in contractual accelerations or the exercise of early termination rights that allow counterparties to demand immediate repayment, which can further stress the institution. The triggering of such rights in normal insolvency proceedings can be one of the chief reasons why such proceedings are a poor fit for financial institutions, as when the Lehman Brothers bankruptcy resulted in widespread terminations by derivatives counterparties. These terminations cost the bankruptcy estate an estimated USD 30 billion to USD 75 billion and helped spread contagion (Wiggins and Metrick 2019a). Among the case studies reviewed, Roskilde’s transfer of senior liabilities to a bridge institution resulted in creditors’ exercising acceleration rights and demanding the return of their funds. This development required the Danish central bank to provide almost USD 8 billion in credit facility support to fund redemptions.

The Key Attributes include mechanisms to help authorities avoid such scenarios. Key Attributes 4.2 and 4.3 provide that entry into resolution should not constitute an event that triggers acceleration or early termination and that if it nonetheless does, resolution authorities should have the ability to temporarily stay such rights (FSB 2014). New resolution regimes have incorporated these attributes. For example, Article 71 of the BRRD requires member states to provide their resolution authorities with the power to temporarily suspend termination rights.

How were the institution’s assets treated during the resolution?

Even for an institution that finds itself on the brink of failure, not all its assets are problematic. Thus, a common theme in the resolution context is the need to separate good assets from troubled assets and to come up with a strategy for dealing with the latter. Among the cases studied for this paper, about half involved so-called good bank/bad bank resolutions. Typically, assets identified as good are transferred from the existing entity directly to a purchaser (via a sale of business tool) or to a bridge institution for eventual sale (via a bridge institution tool). The existing entity retains the troubled assets and is placed into normal insolvency proceedings. In some cases, a new entity may be established to take on those troubled assets (via an asset separation tool). This option allows for the pooling of troubled assets from multiple failed institutions, as was the case in Italy.

One notable exception to this approach to transferring assets was Iceland, where the division of assets was based not solely on good versus bad but also on domestic versus foreign. Following the failure of Iceland’s three largest banks, authorities established new entities to take on their domestic performing assets and domestic insured deposits. Foreign assets and foreign deposits, together with nonperforming domestic assets and uninsured domestic deposits, remained with the failed entities, which were then wound down. This strategy allowed for the continuation of Iceland’s domestic banking business, with the help of government capital injections, separate from its foreign operations, which had become too large relative to the size of Iceland’s economy for the state to rescue. This distinction between domestic and foreign was highly controversial and resulted in years of legal challenges but, given the magnitude of the challenge posed by the foreign operations, it is not clear that Iceland had other options.

A full treatment of the challenges associated with managing troubled assets is beyond the scope of this paper. (For some discussion on this topic, see McNamara et al. [2021]). However, it is important to note that such challenges can hinder the ability to bring a resolution to a successful conclusion in a timely fashion. In an attempt to avoid this issue and dispose of assets quickly, Roskilde attracted loan purchasers by providing them with a window during which to return any loans the purchasers ultimately did not want. More than 50% of purchased loans were ultimately returned, prolonging the resolution. The body tasked with disposing of troubled assets can also itself become troubled, as was the case with Heta, the asset management vehicle established to facilitate the resolution of Hypo Alpe Adria.

The mechanisms available for transferring assets can present challenges. During the resolution of Dunfermline, property transfer orders used to move assets and liabilities could be amended only via legislation. This constraint added complexity to the resolution. Given the haste often associated with resolution and the possibility that subsequent due diligence will reveal the need for adjustments to initial transfers, a system that relies on contracts that can be more easily amended to transfer assets and liabilities may reduce complexity. The nature of the assets themselves can also create difficulties. Before its resolution, Dunfermline had entered into several swap contracts, subject to netting arrangements, to hedge parts of its business. These netting arrangements required that swaps be kept together even when the underlying assets and liabilities being hedged were separated.

How were the institution’s board and management treated during resolution?

A fundamental premise underlying post-GFC resolution regimes is that the consequences of an institution’s failures should be borne first and foremost by its own stakeholders. In that vein, the very first power included in Key Attribute 3.2’s list of general resolution powers is the ability to remove and replace an institution’s senior management and directors and to recover funds, including via clawback of incentive compensation. This authority is followed by the ability to appoint an administrator to take control of and manage the institution. Article 61(l) of the BRRD requires member states to provide resolution authorities with the power to remove or replace the management body and senior management of an institution in resolution, but the BRRD does not specifically require the power to recover funds (Coleman, Georgosouli, and Rice 2018). In several of the cases studied, including Heta and BES, authorities charged former members of the board and/or management with crimes associated with the failure of their institutions, typically for fraud contributing to the collapses.

Did other countries participate in the resolution and if so, how?

The challenges associated with handling the failure of a financial institution operating across multiple national borders were one of the primary motivations behind the new resolution regimes adopted by policymakers post-GFC. In Europe before the introduction of the BRRD, the rescue of failed multinational financial institutions was often a source of significant conflict among different national authorities. In the case of Fortis, an initial agreement among authorities from the Netherlands, Belgium, and Luxembourg fell apart after the Dutch unilaterally nationalized Fortis’s Dutch assets, highlighting the threat that incomplete information sharing and differing priorities and assessments can pose to multilateral resolution. Critics have argued that strategic differences also marred the rescue of Dexia by Belgium and France, with Belgium focused more on domestic financial stability and France on maintaining a key source of municipal finance (with both seeking to minimize the costs to their respective countries).

As described in more detail in Key Design Decision No. 11, Treatment of Assets, Iceland’s rescue of its banking system involved favoring domestic creditors over foreign ones (although, as noted, the magnitude of the challenge foreign operations posed may have left Iceland with no other options). Similar allegations of differing treatment between domestic and foreign creditors arose during the resolution of BES, as did operational difficulties stemming from the treatment of assets and subsidiaries located in different countries that might not have recognized Portugal’s resolution decisions.

The Key Attributes contain an entire section devoted to the legal framework for cross-border resolution. This section calls for countries to create statutory mandates that empower and strongly encourage resolution authorities to cooperate with their foreign counterparts. It also calls for countries to avoid provisions that trigger automatic action upon the initiation of resolution or insolvency proceedings in other jurisdictions and to treat foreign creditors the same as domestic creditors. In the European context, Title V of the BRRD implements the cross-border Key Attributes across the European Union. For eurozone countries, the Single Resolution Mechanism established the SRB in 2015 as the central resolution authority, further facilitating cross-border resolution.

Many of the post-BRRD/post-SRM case studies reviewed involved institutions with little to no global footprint. However, the sale of Banco Popular to Santander in 2017 represented the first resolution undertaken by the SRB. Based in Spain, Banco Popular also had a subsidiary in Portugal that accounted for a significant share of Portugal’s SME lending, which the SRB took into account when deciding to resolve Popular by selling its business to Santander. Although the Banco Popular resolution is widely seen as having been a success, it did lead the European Central Bank (ECB) to propose clarifications in the ECB framework. In June 2019, the European Parliament adopted an updated BRRD II that clarified, among other things, the role of cross-border resolution colleges.

Did the resolution require the institution to limit its activities moving forward in any way?

Crisis interventions are frequently accompanied by restrictions on what recipient institutions can do while those interventions are pending. Motivations include protecting taxpayer funds, restricting the activities that got the institutions into trouble in the first place, and preventing institutions under government control from gaining an unfair competitive advantage. Depending on the nature of a given resolution, authorities may not perceive restrictions as necessary to achieve any of these outcomes. For example, in a resolution such as that of Banco Popular where the failed institution was immediately absorbed by another bank with no direct cost to the state, it is not clear what purpose restrictions would serve. Indeed, placing ongoing restrictions on a failed institution after its acquisition may discourage buyers from undertaking the purchase in the first place. Conversely, where a bridge institution continues to operate in anticipation of an eventual sale, restrictions may be imposed, as was the case in Italy, where authorities restricted the bridge banks’ activities, compensation, acquisitions, and advertising. And in cases where wind-down vehicles remain to be liquidated, it is typical for such vehicles to be restricted to only those activities necessary for the wind-down.

How long did the resolution last and was this consistent with any initial timeline proposed by authorities?

Given that the approach to a resolution can vary widely based on the institution and context, as described in Key Design Decision No. 8, Approach to Resolution and Restructuring, the duration can also be quite different, as reflected in Figure 5. Authorities were able to resolve Banco Popular in one day by selling it to Santander. Where viable buyers failed to materialize, as in the cases of JAK and BES, resolution may take longer than anticipated. In cases where an asset management vehicle has been used to deal with troubled assets, resolutions can be ongoing years later (as was the case, for example, in Italy, where the disposition of bridge banks took place in 2017 but the asset management vehicles took until 2023 to dispose of bad loans). Among the case studies reviewed, the UBS restructuring is noteworthy for having concluded a year before its scheduled end date.

Source: Authors’ compilation.

The establishment of preexisting resolution plans may assist in the timely completion of resolution. This eventuality is not necessarily because the resolution will strictly adhere to the preexisting plan (Banco Popular had a preexisting plan, but authorities deviated from it in resolving the bank). Rather, the act of planning prepares institutions and authorities for more efficient resolutions by requiring them to become familiar with considerations such as institutions’ complex corporate structures. The planning process may even encourage institutions to streamline that complexity, with Key Attribute 11.6 calling for institutions to include in their resolution plans the identification of barriers to resolution and steps for mitigating those barriers (Armour 2014).