Developing an effective framework for cross-border resolution is a key priority in international regulatory reform. Large bank failures during the global financial crisis brought home the lack of adequate tools for resolving “too-big-to-fail” institutions. In cross-border cases, misaligned incentives and lack of robust mechanisms for resolution and cross-border cooperation left some country authorities with little choice but to take unilateral actions, which contributed to the high fiscal costs of the crisis and resulted in disorderly resolution in some cases


Developing an effective framework for cross-border resolution is a key priority in international regulatory reform. Large bank failures during the global financial crisis brought home the lack of adequate tools for resolving “too-big-to-fail” institutions. In cross-border cases, misaligned incentives and lack of robust mechanisms for resolution and cross-border cooperation left some country authorities with little choice but to take unilateral actions, which contributed to the high fiscal costs of the crisis and resulted in disorderly resolution in some cases


1. Cross-border bank resolution is at the forefront of the international regulatory reform agenda. In the wake of the global crisis, a central feature of the international community’s efforts to make the global financial system safer has been to end “too-big-to-fail” financial institutions. This effort has focused on identifying global systemically important financial institutions (G-SIFIs), making them more resilient through increased capital and tighter regulation and supervision, and strengthening the resolution regime to minimize risks of financial instability and taxpayer loss should they fail. On this last dimension, the international response has centered on the work of the Financial Stability Board (FSB), in particular, the Key Attributes of Effective Resolution Regimes for Financial Institutions (Key Attributes)—a new international standard endorsed by the G-20 leaders in 2011. The IMF has played an active role in the development of the Key Attributes, in particular through the 2010 Board paper on Resolution of Cross-Border Banks—A Proposed Framework for Enhanced Coordination. 1

2. Consensus has been reached on a framework, but important details need to be worked out—as yet, orderly resolution of systemic cross-border banks is not a feasible option. Despite ambitious legislative reforms, more needs to be done at the country level to fully implement the Key Attributes. Moreover, international agreement and further guidance is needed in a number of areas. These include the development of mechanisms to facilitate the allocation of losses among private stakeholders, and to manage the residual risks that public funds may be needed. Greater consideration of the applicability of the Key Attributes beyond the G-SIFIs and their main countries of operation is also needed.

3. This paper reviews recent progress in the development of an effective framework for cross-border bank resolution and identifies the key gaps that need to be addressed. As it has been two years since the Executive Board was last given a comprehensive briefing on cross-border resolution, an update on developments is timely.2 The paper begins with an examination of the principal lessons learned from the recent financial crisis and the weaknesses in the international framework that the crisis exposed. It then discusses how the Key Attributes seek to address these weaknesses, and countries’ progress in implementation. The paper then discusses the gaps that remain to be filled in the design of the international framework and ways to address implementation challenges.

Cross-Border Resolution in Practice

4. Cross-border banks have benefits—and costs. 3 The experience of the past decade has demonstrated the benefits and potential risks presented by the rapid growth of cross-border banking. By pursuing growth opportunities beyond their borders, successful national banks export capital and banking technologies, and realize economies of scale in managing large, diversified operations. Risk is pooled across economies, and the banks’ “deep pockets” often provide a credible backstop for host authorities. However, in the face of a systemic shock, these entities are also instant transmitters of financial stress. The capacity to safeguard global financial stability and minimize losses through an efficient resolution process then becomes essential.

5. The vast majority of cross-border finance is intermediated by a small, core set of financial institutions that are interconnected. In 2008, 145 banks globally accounted for 85 percent of the assets of the world’s top 1,000 banks. Some of the largest cross-border groups have assets that account for several times their home country GDP. With $10 trillion of total assets, the top five largest cross-border banking groups account for about 10 percent of global banking industry assets. These five groups, all of which are headquartered in Europe, have over 50 percent of their credit risk exposures and employ roughly 60 percent of their staff outside their “home” country, with subsidiaries in more than 60 countries.

6. Resolving a cross-border bank involves enormous challenges. Financial groups develop complex international structures in order to exploit new business opportunities and maximize profits. These structures are often made up of hundreds of different entities in a web of financial and operational interdependency. When such an institution fails, authorities in different jurisdictions have to confront significant uncertainty and institutional constraints, and potentially misaligned incentives. Without an effective arsenal of resolution tools and cooperative arrangements between key jurisdictions, orderly resolution of a large cross-border financial group will prove impossible.

7. Incentives for cooperation in cross-border resolution are typically weak. Resolution authorities are answerable to domestic constituencies; their first responsibility is often to protect domestic financial stability and minimize any risk to public funds. Although preserving international business lines, financial links, and operational dependencies in resolution may minimize overall economic losses from a bank’s failure, individual country perspectives may differ from globally efficient solutions. Unilateral “ring-fencing” operations may allow protecting systemic operations domestically at a cost or risk known with some margin of error, versus an uncertain gain from an orderly international intervention, with risk of a worse outcome. Where public funds are put at risk, bearing a share of the burden that is later perceived to be “unfair” or disproportionate could carry important political costs. “Home”—where the parent company is located—and “host”—where international affiliates are located—jurisdictions of cross-border banks can face starkly different incentives to cooperate where affiliates are either systemic for the host, but immaterial for the group; or material for the group, but not systemic for the host. In both cases, one party may have strong incentives to take unilateral action, even if this has a material impact on the group and generates spillovers for other countries.

8. Ex ante agreements on public sector burden sharing may not amount to credible commitments and have had little impact in practice to date. Prior to the recent crisis, even the exchange of information proved difficult, given legal constraints and sometimes conflicting national interests (Box 1). Formal agreements on sharing the public costs of resolution have been difficult to achieve in practice, as incentives to depart from their provisions in the light of actual circumstances are high. Moreover, the existence of formal burden sharing agreements that suggest that public funds will sometimes be available for resolution may increase moral hazard.

Why did MoUs Fail to Secure Cooperation in the Global Financial Crisis?

MoUs grew in importance as supervisory cooperation intensified. Cross-border cooperation in banking supervision has existed since the 1970s, with Basel Committee work on consolidated supervision. Through the 1990s, standards focused on information flows between home and host supervisors, with cooperation in crisis times generally deemed unrealistic. Supervisory cooperation intensified in the 2000s, with work on MoUs and cooperation between banking supervisors embodied in Basel II stimulating the development of supervisory colleges. A 2006 Financial Stability Institute (FSI) survey found that 80 percent of bank supervisors worldwide used MoUs and informal contacts to support supervision in normal times.

Much less thought was given to cross-border arrangements on crisis management. The FSI survey found that only 19 percent of supervisors would make use of college arrangements to address financial stress, while 49 percent would resort to ad hoc meetings, and 40 percent had no formal arrangements at all. In 2003, EU national banking supervisors and central banks did agree on a multilateral MoU on high-level principles for cooperation in crisis management, later revised and expanded to include Ministries of Finance. The goal was to enable timely engagement and adequate information for all parties involved in a crisis, committing signatories to: (i) information exchange; (ii) joint public communication; (iii) joint contingency planning; (iv) optionally, forming cross-border stability groups; and (v) voluntary specific cooperation agreements on burden sharing. The agreement was not legally binding.

Supervisory MoUs failed to enable cooperation in the crisis. As MoUs were nonbinding and supervisory colleges were not empowered to make decisions, many MoU commitments—such as timely sharing of information—were quickly abandoned as domestic financial stability concerns became paramount. Supervisory colleges could not restrain unilateral action. MoUs did not provide for specific timelines or escalation procedures and had many exemptions. Differences in the systemic (and political) importance of banks for home and host countries resulted in misaligned incentives—which came to the fore when authorities found themselves negotiating over the distribution of significant financial commitments under intense time pressure and with no agreed process; or took unilateral action, leading in some cases to disorderly resolution and higher overall costs.

9. A key lesson from the 2008 crisis was that national authorities lacked both the legal tools and the cooperative arrangements needed to resolve cross-border financial groups. As is demonstrated by the studies set out in Annex I, national authorities faced huge challenges in responding to bank failures, both of large, systemic financial institutions and smaller ones. Unilateral responses were the norm, leading in some cases to the breakup of groups into national components and the commitment of large amounts of public funds. In particular,

  • National resolution frameworks were inadequate. Resolution authorities did not have the legal powers to restructure failing domestic or cross-border institutions or to promptly write down the claims of stakeholders. Many jurisdictions relied upon their general corporate insolvency frameworks where restructuring operations required slow, court-based proceedings and, in some cases, the consent of shareholders and creditors.

  • Arrangements for cross-border cooperation were lacking. Mechanisms in national legal systems to give effect to foreign resolution measures were inadequate. The cooperative arrangements in place between the competent authorities in key jurisdictions were insufficient.

  • Concerns for domestic financial stability impeded cooperative solutions. Faced with domestic financial stability concerns and significant uncertainty about risks from international operations, national authorities generally did not pursue cooperative approaches. In many cases, there was no legal authority or appetite to design group-wide resolution strategies.

  • Sovereign financial strength emerged as a key factor in determining national strategies. Financially strong governments restored stability by bailing out parent banks in their jurisdictions, leaving group structures largely intact. Benefits were transmitted across borders through group access to liquidity support, in particular, jurisdictions. This minimized cross-border spillovers, but exposed authorities to high fiscal cost and political backlash—such a solution will not be feasible in future for some countries.4 Where credible sovereign backstops were lacking (such as in Cyprus and Iceland), authorities had to take exceptional steps to restructure banks while stemming depositor runs.

10. In sum, the crisis exposed three major gaps in the international framework for resolution:

  • Inadequate resolution powers and tools through which to resolve a failing financial institution or group in an orderly fashion and through a least-cost approach;

  • Inadequate frameworks for the enforcement of resolution measures on a cross-border basis and for coordination between authorities in different countries in the event of a cross-border failure;

  • Inadequate mechanisms for loss allocation, including the ability to credibly subject bank creditors to loss in a bank failure without endangering financial stability.

Improving Resolution Frameworks

11. The Key Attributes attempt to address these gaps. They aim to enable orderly resolution while ending the distortions created by the perception that some financial institutions are “too-big-to-fail” and will be bailed out by the state. By making orderly resolution credible for large financial groups, these reforms seek to change incentives for bank creditors and contain moral hazard. Further progress is essential to maintaining the benefits from open financial markets and the efficient allocation of resources and risk-sharing globally.

A. Overview of the Key Attributes

12. The Key Attributes establish a new international standard for the design of resolution regimes for systemically important financial institutions and groups. 5 They set out the powers and tools necessary to enable effective resolution and call for coordinated approaches to resolving cross-border institutions, and for allocation of losses to private creditors. They are a non-binding standard, and countries will be encouraged to implement their provisions over time.

13. The Key Attributes call for resolution authorities to be given a comprehensive toolkit of powers. These include the power to take control of a failing financial institution, merge ailing banks with stronger institutions, and transfer assets and liabilities to a healthy institution or a so-called “bridge bank.” A key tool is the power to engage in the mandatory debt restructuring or “bail-in” of a financial institution. This is essentially a statutory power to restructure the liabilities of a distressed financial institution by writing down its unsecured debt and/or converting it to equity while the relevant institution remains open. In order to ensure the prompt and effective restructuring of the institution, all of these powers are to be exercised without the need to obtain the consent of shareholders or creditors, subject to appropriate safeguards.

14. The Key Attributes aim to enhance the effectiveness of cross-border resolution. They recognize that cross-border cooperation requires the ability of resolution/supervisory authorities to share information with their foreign counterparts, and mechanisms through which resolution measures taken in one jurisdiction may be enforced in another jurisdiction. To address these obstacles, the Key Attributes call for national resolution regimes to: (i) empower and strongly encourage the resolution authorities, wherever possible, to act to achieve a cooperative solution with foreign resolution authorities; (ii) empower the resolution authorities to share information with their foreign counterparts, provided that arrangements are in place for the protection of confidential information; and (iii) provide for transparent and expedited processes to give effect to foreign resolution measures.

15. The Key Attributes contemplate the establishment of cooperative arrangements between national resolution authorities for the resolution of SIFIs. Jurisdictions are required to put in place Recovery and Resolution Plans (RRPs), covering, at a minimum, domestically incorporated firms that could be systemically significant or critical if they fail. RRPs enable the preparation of concrete strategies to resolve each failing bank while safeguarding financial stability. Strategies currently under development focus on “single point of entry” (SPE) and “multiple point of entry” (MPE) approaches (Box 2). For G-SIFIs, the Key Attributes further require home and key host authorities to maintain Crisis Management Groups to prepare for and facilitate the management of a cross-border failure.6 Relevant national authorities are also required to put in place for all G-SIFIs, at a minimum, institution-specific Cooperation Agreements that enable the execution of the agreed resolution strategy and ensure information exchange.

Institution-Specific Resolution Strategies

Progress in developing operational strategies for the resolution of G-SIFIs has focused on two approaches: single and multiple points of entry. These strategies apply to financial groups in their entirety, which may consist of hundreds of different legal entities located in multiple jurisdictions, and give the lead in resolution to the home authority under SPE, and to individual home and host authorities under MPE. Losses of the institution in resolution will accrue and be managed at the parent level under SPE, or at the parent and subsidiary levels under MPE. Both approaches can imply the use of a range of resolution powers, including bail-in, which plays a central role, particularly in SPE approaches.

  • Under SPE strategies, resolution occurs at the very top of the financial group while operating parts of the group are preserved. The home resolution authority intervenes and restructures the holding company that sits at the top of the financial group without the need for host authorities to resolve operating subsidiaries under their respective control. Shareholders and creditors of the apex institution absorb losses of the entire group through a write-down or restructuring of their equity and/or debt claims against the apex entity. Capital freed up from this exercise is “passed down” to loss-making operating subsidiaries and used to recapitalize and provide liquidity for such subsidiaries. SPE requires the ability to downstream loss-absorbing capacity to other parts of the group; and the capacity and willingness to provide liquidity support in resolution—both of which require a high level of cooperation and trust among authorities during resolution.

  • Under MPE strategies, operating affiliates are resolved separately in different jurisdictions by the respective resolution authorities. Individual parts of the group are resolved in separate proceedings and losses are dealt with at the subsidiary level. Given the risks of disruption and inadequately coordinated actions, MPE requires extensive preparation and coordination to facilitate orderly resolution and preserve essential services and financial functions.

  • Hybrid strategies are possible—for example, core operations in key jurisdictions where the G-SIFI is active might be resolved under SPE, while stand-alone operations in other countries are resolved following an MPE approach, or closely linked regional operations might be resolved as a group in an overall MPE strategy.

The choice of resolution strategy and the structure of cross-border banks must be consistent. MPE approaches can work well where operating entities can operate on a stand-alone basis. Arrangements need to be made in order to ensure critical financial and operational services provided by other parts of the group can be reproduced, and that adequate loss-absorbing capacity exists locally. Where intra-group operations are more intensively inter-linked, or operations benefit from the size and liquidity provided by a “global” balance sheet (some capital market activity, for example), an SPE approach may be warranted. RRPs can help develop a picture of where losses might fall or liquidity needs arise in a crisis, providing input for cross-border dialogue on burden sharing and appropriate resolution strategies.

B. Implementation of the Key Attributes

16. The Key Attributes have not been fully implemented by many countries and political commitment is necessary to make meaningful progress. A 2013 peer review conducted by the FSB found that implementation within the FSB’s membership was at an early stage.7 Resolution regimes were generally more developed for banks than for other types of financial institutions, but, even for banks, lacked important powers such as bail-in or the ability to temporarily suspend early termination rights under financial contracts. The range of powers available for non-banks (e.g., insurance companies) was, in many cases, very limited. With respect to cross-border cooperation, only a few jurisdictions had established robust mechanisms to give effect to foreign resolution measures, or gave their resolution authorities clear statutory mandates to cooperate with or to share information with foreign resolution authorities. The peer review report called for more action on the part of national authorities and the FSB to promote implementation of the Key Attributes.

17. Some progress in the implementation of the Key Attributes has been made since 2013. A major step was recently taken by the European Union with the adoption in May 2014 of the Bank Recovery and Resolution Directive by the European Parliament (Annex II). The Directive was explicitly intended to bring national resolution frameworks in line with the Key Attributes.

18. Achieving progress with recovery and resolution planning for individual cross-border groups has proven challenging. As contemplated under the Key Attributes, firm-specific Crisis Management Groups have been put in place for all global systemically important banks designated as such by the FSB, and preferred resolution strategies (SPE, MPE, or hybrid) are being discussed within those Groups and made operational. Firm-specific Cooperation Agreements are at various stages of development—their coverage remains limited to procedural aspects of cooperation, such as exchange of information, and does not include details on resolution strategies to be pursued.

19. Political commitment is needed to promote implementation. All FSB members have committed to fully implement the Key Attributes by the end of 2015—yet this will require much work. The FSB will support the process through guidance, regular monitoring,8 and peer reviews.

20. The Fund works with its members on strengthening resolution regimes. In the context of program support, surveillance, and technical assistance (including the Financial Sector Assessment Program, FSAP), the Fund has recommended enhancing resolution and crisis management frameworks in line with the Key Attributes (Box 3). Moreover, it is expected that the Fund will play an important role in assessing members’ implementation of the Key Attributes: they will be presented to the Fund and Bank Executive Boards for endorsement as a standard under the Report on the Observance of Standards and Codes (ROSC) program, once the associated assessment methodology is finalized.

Examples of Fund Advice to Member Countries on Resolution Frameworks

The Fund provides policy advice to strengthen resolution frameworks in the context of Fund-supported programs, surveillance, and technical assistance (including the FSAP). Fund advice typically addresses some of the following issues:

  • Resolution Authority: Providing for a dedicated and autonomous resolution authority with a mandate to preserve financial stability and ensure continuity of systemically important services;

  • Early Intervention: Strengthening powers for intervention before a firm is balance-sheet insolvent;

  • Resolution Powers and Tools: Giving the resolution authorities a broad range of restructuring powers, including transferring a bank’s assets and liabilities to another institution without shareholder or creditor approvals;

  • Safeguards: Introducing safeguards for creditors to ensure that the hierarchy of claims is respected and creditors would be no worse off under restructuring than under liquidation;

  • Legal Protection: Providing for legal protection for officers, staff, and agents of resolution authorities for good-faith actions in resolution;

  • Resolution Funding: Providing explicitly for emergency liquidity assistance with appropriate safeguards to protect the central bank’s balance sheet, and the use of deposit insurance funds to facilitate resolution;

  • Judicial Review: Ensuring that the legal framework does not endanger the effective implementation of resolution measures and legal certainly through appeals processes in court; and

  • Information Sharing and Cooperation: Providing adequate legal authority for information sharing with foreign authorities and providing clear legal mandates for cross-border resolution cooperation.

Enhancing Mechanisms for Loss Allocation

21. Beyond the need for countries to implement the Key Attributes, much more work is needed to achieve a feasible framework for cross-border resolution. Remaining gaps pertain to how losses will be allocated to private sector stakeholders in a resolution, and managing residual risks that public funds will be needed in resolution. Both issues could undermine orderly cross-border resolution by leaving resolution authorities with incentives to act unilaterally in a crisis.

A. Shifting the Burden to the Private Sector

22. In determining how to allocate losses among private stakeholders, two particular questions require further work by the international community. These concern the development of requirements for loss-absorbing claims and the need for greater international harmonization of creditor hierarchies.

Making bail-in a credible option

23. Holders of claims targeted for bail-in must be able to absorb potential losses without generating systemic risk themselves as a consequence of their financial losses. Financial groups need to carry sufficient buffers of such claims to ensure that bail-in is not only legally feasible but also credible as a policy option. Further work at the international level is needed to ensure that this is the case and also to promote uniformity as to which claims may be bail-inable. Doubts about whether losses can, in fact, be absorbed across a financial group, as specified in a resolution strategy, could increase incentives to ring-fence to minimize risks from the strategy failing, particularly with respect to SPE strategies, where losses are passed back to one entity.

24. The FSB is working on a proposal to develop “gone-concern loss-absorbing capacity” (GLAC). The purpose of GLAC is to ensure a credible buffer of liabilities beyond equity that can be subject to mandatory restructuring in a resolution and to promote a uniform international approach to this issue. A set of principles on GLAC will be presented at the G20 summit in November 2014, which is expected to cover the following issues:

  • Definition of claims that would be regarded as GLAC: While any unsecured liability could, theoretically, be bailed in, the risks that may arise with imposing losses on uninsured deposits, interbank loans, or selected other liabilities in specific cases imply that the most likely liabilities to be eligible as GLAC are subordinated debt, senior unsecured bonds, and intra-group loans. To the extent that investors perceive the bail-in of such claims to be credible, funding costs from these sources may rise to reflect actual risks.9 At the same time, the funding costs could fall for liabilities that are significantly less likely to suffer loss.

  • Amount of GLAC. Enough “bail-inable” liabilities are needed to help ensure that a mandatory debt restructuring would be sufficient to resolve distressed institutions in an orderly manner, potentially forestalling runs by short-term creditors concerned about counterparty solvency. This will depend partly on past experience with bank losses, which suggests that fiscal costs of crises averaged 13 percent of financial system assets;10 as well as other considerations, such as the target capital ratio after resolution to foster market confidence, and any additional GLAC requirement (or discount) to reflect the degree of resolvability. Issuing GLAC on a significant scale may prove challenging, given the historical size of markets in unsecured debt instruments.

  • Location of GLAC: Any difference between the place where losses are generated and the location of loss-absorbing liabilities within a group’s structure can complicate the resolution of cross-border banks. This may well be the case, as financial depth and tax considerations are critical factors in deciding on where marketable bank liabilities are issued. If GLAC is not located in the jurisdictions (and within a jurisdiction, the entity) where losses are most likely to arise in distress, there must be a mechanism to move losses to where GLAC exists, or vice versa (see Box 2). The choice of jurisdictions where GLAC is located depends on the structure of the bank and the resolution strategy to be pursued.

  • Appropriate investor base: Ensuring GLAC is held by the “right” creditors—capable of absorbing losses without socially undesirable consequences—may be critical for the credibility of bail-in. A particular example could be measures to mitigate contagion risks to other financial institutions by limiting their cross-holdings of senior unsecured debt. Additionally, as claims are traded in secondary markets, there is a risk they might quickly migrate to investors that are perceived to need protection from bail-in. Another considerations is bank governance after resolution, as conversion of GLAC into equity may create new controlling shareholders (or a more dispersed shareholding structure), that may or may not meet regulatory suitability requirements.

  • A uniform approach: Some jurisdictions have limited the application of mandatory debt restructuring and GLAC requirements to systemic banks (e.g., the United States), while other countries plan on applying requirements more broadly (e.g., the European Union)—a common approach would better ensure a level playing field. Any GLAC framework should not discriminate on the basis of the nationality of investors when making decisions on triggering bail-in.

  • Disclosure: Publication by banks of how much GLAC they have issued and how it would be used could help enhance fair market pricing and creditor discipline.

25. The amount and location of GLAC determine ex ante how losses accrue in resolution, aligning incentives of home and host authorities with respect to private burden sharing. This includes intra-group GLAC, such as debt of subsidiaries held by the parent. Resolution authorities must also understand how losses will be absorbed and transmitted within financial groups. In MPE approaches, each point of entry will need to issue sufficient GLAC to allow for orderly resolution. In SPE approaches, all parties must be clear on exactly how losses will be imposed on the entities within the group that issue GLAC, which will require intra-group loss-transfer agreements. These may provide for the conversion of loans from the parent company to the subsidiary into equity, or collateralized guarantees provided by the parent on borrowing by the subsidiaries. How, when, and by whom such mechanisms will be activated will need to be addressed in loss-transfer agreements.

Creditor hierarchy

26. The lack of harmonization of creditor hierarchies between jurisdictions is a key obstacle to the allocation of losses to private stakeholders. Creditor hierarchies determine the ranking/order in which losses will be allocated against the claims of various stakeholders in the liquidation of a firm. Losses are first allocated against shareholders, before subordinated debt holders and other types of creditors in ascending order. Claims falling in the same class rank “pari passu” with each other and are treated in the same way in a liquidation (i.e., it is generally not possible for the liquidator to pay one but not the other).

27. The creditor hierarchy is an important instrument for securing the appropriate allocation of losses amongst stakeholders in both the liquidation and resolution of a bank. The creditor hierarchy for the liquidation of a bank will generally resemble that applicable in ordinary corporate insolvency proceedings, except that it will need to address creditors specific to banks, such as small depositors, the deposit insurance scheme, and the public sector providing liquidity or solvency support. In the restructuring of a bank as a going concern, the restructuring of the bank’s claims will generally take into account the order of priorities applicable in liquidation, and treat all of the claims that fall into a particular class in the same manner. However, there may be instances where the resolution authority must depart from those general principles. This may be achieved by exempting certain claims ex ante from the imposition of losses in the going-concern resolution, even though those claims would be likely to suffer losses under the creditor hierarchy for liquidation. Another way is to empower the resolution authorities, in the resolution of a specific institution, to treat equally ranking claims in an unequal manner.11 Unequal treatment of creditors or deviation from the established creditor hierarchy in liquidation raises the risk of litigation by affected creditors who may challenge a restructuring on the basis that their interests have been harmed. Adequate safeguards, such as a “no creditor worse off” requirement, can mitigate this risk.

28. Significant differences in the design of creditor hierarchies between jurisdictions can be an obstacle to cross-border resolution of branches of an international bank. If the ranking in the home country differs considerably from the ranking in the host country, the host authorities may decline to cooperate. For instance, if uninsured depositors are preferred to other senior creditors in the host country but not under home country law, the host authorities may have an incentive to ring-fence instead of contributing all local assets and liabilities to a resolution per home country law.12 Incentives for cross-border cooperation will also be undermined if the home country’s creditor hierarchy discriminates against foreign or nonresident creditors or creditors of foreign branches.

29. Further guidance at the international level on creditor hierarchies is needed, particularly in the treatment of deposits. The Key Attributes provide some useful guidance on the hierarchy to be employed in the context of a bank resolution: Key Attribute 5.1 requires that “resolution powers should be exercised in a way that respects the hierarchy of claims,” in particular, that “equity should absorb losses first, and no loss should be imposed on senior debt holders until subordinated debt has been written off entirely.” The Key Attributes also permit departures from the pari passu treatment of claims in order to contain systemic risk or to maximize the value for creditors. With respect to cross-border cooperation, Key Attribute 7.4 provides that national resolution frameworks should not discriminate on the basis of the nationality of creditors, the location of their claim, or the jurisdiction in which it is payable. However, the Key Attributes leave important questions unanswered—in particular, on the treatment of insured and uninsured deposits in the creditor hierarchy. Greater specificity on this question would facilitate the resolution of banks, particularly on a cross-border basis.

30. The crisis has shown that giving preference to depositors is a useful element to mitigate the likelihood of contagion (Box 4). Depositor preference can mitigate the likelihood of contagion by providing depositors with a higher probability of recovering their claims in a bank failure. Where a deposit insurance scheme is in place, a framework that gives a preference to insured deposits mitigates the risk that the deposit insurance scheme will suffer losses; extending preference to all deposits (including uninsured) decreases the likelihood that depositors might need to be bailed in, or otherwise suffer losses. A significant number of jurisdictions already have some form of depositor preference. In providing advice on bank resolution frameworks, Fund staff typically addresses the issue of creditor hierarchy explicitly, and advocates providing depositor preference, although the exact form will depend on country circumstances, including whether or not a deposit insurance scheme is in place.

The Role of Depositor Preference in Bank Resolution

Depositor preference (DP) gives depositors a preferential claim on the assets of a failed bank. It is reflected in the hierarchy of claims in the liquidation of a bank, and is also of relevance for the treatment of creditors in the resolution of banks. DP has attracted renewed interest in the context of international regulatory reform, as it affects loss allocation among various private sector stakeholders.

Existing and proposed forms of DP vary across countries, but three specific approaches are common:

  • Insured depositor preference provides preferential treatment for insured deposits (and the deposit insurance scheme (DIS) through subrogation), and ranks uninsured deposits pari passu with the senior unsecured creditors.

  • General depositor preference gives preference to all deposits, including balances above the deposit insurance limit, over senior unsecured creditors. The DIS is subrogated for insured deposits and, thus, ranks pari passu with uninsured deposits;

  • Tiered depositor preference prefers insured deposits (and the DIS through subrogation) over uninsured deposits, and prefers both over senior unsecured creditors.

    Different Types of Deposit Preference

    article image
    Note: A higher position in the table and darker shading (within the same column) indicates a higher creditor claim.

DP can mitigate the likelihood of contagion by providing depositors with a higher probability of recovering their claims in a bank failure. For countries with a DIS that is subrogated for insured deposits, insured or tiered DP mitigates the risk of the DIS suffering losses, thereby also contributing to shifting the burden of resolution to the private sector. By creating clear legal grounds for the preferential treatment of all deposits over other unsecured senior creditors, and signaling transparently which bank liabilities will be protected in resolution, general or tiered DP may help reduce legal challenges in case of a transfer of all deposits to another institution or the bail-in of non-deposit creditors only.

Potential benefits of changing the creditor hierarchy under all forms of DP need to be weighed against the potential impact on bank funding costs, or the availability of unsecured wholesale funding.

A significant number of jurisdictions already have some form of depositor preference. For example, Argentina, Australia, China, India, Hong Kong SAR, Indonesia, Mexico, Russia, Singapore, Switzerland, Turkey, the United States, and several European countries afford preferential treatment to at least some depositors, and similar reforms were recently introduced in the United Kingdom and the European Union.

B. Minimizing the Potential Burden for the Public Sector

31. International regulatory reform holds out the prospect of reduced probabilities of bank failure and lower fiscal costs, should failures occur. Following the recent crisis, there is both less capacity and little appetite for any future fiscal resources to be put at risk to stem the consequences of a large financial institution’s failure. Both increased capital and GLAC requirements will help impose market discipline and create better incentives for managers. Should a bank become distressed, thicker buffers of equity capital will bolster confidence and give more time for remedial action and recovery plans to work, while GLAC—should resolution ultimately be needed—can absorb losses before public funds may be needed, at least with respect to solvency support.

32. Nonetheless, there will always be a residual risk that public funds will be needed to preserve financial stability, at least temporarily. Emergency liquidity assistance may have to be offered on terms or against collateral that leave the central bank with some risk of loss. In some cases, state guarantees of bank liabilities (or capping losses on specific asset portfolios) may be needed to ensure continued market access. Finally, should it become apparent that GLAC cannot cover potential losses (or do so without causing fresh contagion) solvency support for systemic institutions may be needed—even if this is to be recovered later from the industry.13 Once GLAC is exhausted, it is unlikely that private sector funding can be mobilized in the very short time periods in which a bank resolution will need to be implemented—in particular, because many legal frameworks do not enable private-sector “debtor in possession” financing for a bank resolution (whereby post-insolvency liquidity providers gain priority in the creditor hierarchy over pre-insolvency creditors).

33. Both liquidity and solvency support raise challenges for cross-border resolution. Liquidity support can entail significant “risk transfer” while the potential need for solvency support will create incentives to ring-fence to minimize uncertainty and fiscal risks.

  • Liquidity support: Central banks can provide unlimited support only in their own currency— where other currencies are needed, swap arrangements have to be put in place. While the central bank that provides the other needed currency retains the risk for repayment of the swap, the counterpart central bank is exposed to foreign exchange risk. Additionally, central banks may need to consider accepting foreign collateral, leaving them exposed, should that jurisdiction ring-fence or take other unilateral action.

  • Solvency support: A realistic possibility of having to put public funds at risk may create incentives to ring-fence to minimize the uncertainty arising from lower visibility on foreign operations, and the prospect that foreign creditors will be the main beneficiaries of taxpayer funds.

34. Ex-ante agreements defining how public support would be committed to resolve a weak or failing institution are difficult to achieve in practice. 14 The ex-ante burden sharing framework among the Nordic-Baltic countries is an important example of a non-binding agreement for public sector burden sharing in a resolution.15 Beyond such tightly knit countries, however, the development of criteria that would guide public sector burden sharing remain elusive, and national authorities have been reluctant to make binding commitments for the use of public funds. As a consequence, the regulatory reform agenda has been directed toward the avoidance of public solvency support.

35. Shifting decisions and actions to the pre-crisis phase can reduce the risks that public funding for resolution may be needed. Recovery and resolution planning can help develop a picture of structural imbalances in the distribution of assets and liabilities in a financial group from a national perspective. Information exchange alone will not change incentives for unilateral action in a crisis, and this information will be highly incomplete and risk exposures can change very quickly. Nonetheless, knowledge about where liquid and other assets are concentrated within cross-border groups, the existence of internal guarantees, and how and where losses can be absorbed with GLAC can help generate a common understanding of the implications of the group’s failure for each party. This knowledge can be used in an incremental process to require financial groups to adjust their structure, regulate the placement of some assets and liabilities (as with GLAC), or put in place prudential limits on specific exposures.16 Such a procedure can also help individual hosts to manage potential risks up to acceptable levels, considering also the benefits that accrue from hosting cross-border banks. In this manner, incentives will be better aligned to ensure resolution plans will be executed in practice as agreed. An effective framework that meets these objectives would include the following elements (see Figure):

  • Assessing structural financial imbalances between countries embedded in cross-border financial groups’ balance sheets; this process can help generate a common understanding of the implications of the bank’s failure for each party;

  • Determining if structural changes to cross-border financial groups are needed such that each country is satisfied with the balance between advantages of cross-border activities and the downside of potential spillover risks; and

  • Assessing the potential for cross-border sovereign risk transfer or outright public costs arising from policy responses in the event of failure.

36. Through this process, the costs and benefits of different bank structures or resolution strategies can be assessed. The perspective of potential costs if a cross-border bank fails can help inform a “risk-based” assessment of the benefits of hosting institutions, and in what form (e.g., as a subsidiary or branch), or what risks need to be addressed by the bank, the supervisor, or in recovery and resolution planning. Three questions will help determine whether a particular configuration for a bank and resolution strategy makes sense for any one country:

  • To what extent can the burden of a bank’s failure be contained within the private sector?

  • What burden could remain with the public sector and how can this be minimized or eliminated?

  • What type of coordinated policies can be put in place to minimize ex-post ring-fencing or unilateral decisions?

37. With a clearer picture of the risks, countries can assess whether the imbalances are acceptable. Key factors will be the willingness and capacity to cope with the risks of providing support in case of failure, including the fiscal and monetary capacity of each country to implement resolution strategies for failing banks. The assessment will also be guided by broader policy considerations taking into account the benefits attributed to cross-border banking, and the credibility of cooperation pledges by other countries.17 Ex-ante measures may be taken to ensure that imbalances are at a mutually acceptable level. Potential measures include:

  • Reducing upfront exposure by limiting cross-border risk transfer. Measures could include requiring banks to adjust their structure, regulating placement of assets and liabilities or prudential limits on specific exposures (e.g., foreign exchange risk, liquidity) or financial flows.

  • Increasing acceptability of imbalances through agreement. Agreement could be sought on the provision of emergency liquidity at the group level or any residual solvency support.

38. A system-wide perspective is important. There are gains to be derived from looking beyond individual banks where there are multiple channels of contagion between financial systems. For example, the joint support expressed by U.K. and U.S. authorities for resolution of SIFIs based in their jurisdictions using SPE-type strategies is predicated on the likely high costs across both financial systems of a failure to cooperate, as both countries are home and host of systemically relevant banks active in both jurisdictions.18

39. Cross-border simulation exercises can provide a forum for discussing alternatives to unilateral actions. Important elements include: (i) Would an ailing bank be allowed to keep up cross-border capital and funding flows? (ii) Which sources of funding could be made available in crisis? Under what conditions? (iii) Which claims would a country accept to be bailed in? (iv) What size of imbalances would a country be comfortable accepting, and under what circumstances?

40. Several outcomes of cooperation strategies are possible. Considering asset and liability positions in isolation, full ex ante ring-fencing and maximum resolvability will always be an equilibrium strategy. However, this ignores the efficiencies that may accrue from sharing critical functions. Where efficiencies appear small (for example, where locally funded operations undertake mainly basic banking functions of taking deposits and lending), the benefit of ease of resolution may weigh highly. Where the loss of shared platforms or a global operation has higher costs, mechanisms to offset imbalances with prior commitments to burden sharing actions could be explored (Box 5). For example, country A may be willing to accept a liability overhang in the local operations of a cross-border bank, if B is willing to provide counterbalancing resolution funding.

Incentives to Cooperate and Subsidiarization

Cross-border resolution can be helped or impeded depending on group structures—particularly banks’ preference for subsidiaries versus branches. From the perspective of financial stability, a branch structure unequivocally puts responsibility for soundness on the parent institution, while a subsidiary structure can limit losses at the host level, should the parent come under stress (given local capital and liquidity requirements). The subsidiary approach may also limit the overall cost of resolution in the event of a failure, because spinning off the relatively healthy parts of the group may be easier when the group is structured as a network of independent subsidiaries rather than a fully integrated branch network.

Subsidiarization ex ante is likely to be a superior strategy to ring-fencing ex post when cross-border cooperation is unlikely to emerge. Ring fencing during a crisis has significant operational, financial, and legal risks and can of itself worsen the situation. However, if cooperative strategies are not credible and countries fear that their counterparts will act unilaterally in crisis, the host authorities may choose to protect their interests through requiring cross-border banks to hold local capital and liquidity ex ante. While subsidiarization does not preclude ring-fencing in a crisis, stand-alone-subsidiarization can reduce incentives for hosts to take such action, as they have a more complete understanding and control of local assets and liabilities, and transparently requiring capital and liquidity beforehand can reduce negative effects of unilaterally impeding moving funds during a crisis.

Isolating pools of capital and liquidity could have efficiency costs and impede cross-border financial intermediation (although the extent of these costs versus the benefits of increased “resolvability” has not been established). Where operations are locally funded and assets consist of relatively straightforward loans to local counterparts, the net costs of insisting on locally resolvable structures may be low. Costs may rise if branch structures are very integrated with other parts of the group, or as the foreign banks play a more active role in intermediating international financial flows (e.g., providing significant funding to a host economy on a long-term basis). In these cases, host authorities will have to weigh the advantages of intermediation through integrated branches against a greater need for cooperation in resolution.

For heavily interconnected markets with systemic players, formal cooperation may be a better strategy. In these cases, contagion risk and the greater reciprocal damage from severing financial flows makes it less likely that unilateral action will pay off. As an example, in December 2012 the United States and the United Kingdom stated their preference for a cooperative SPE strategy for their systemic banks. With significant interconnectedness between these financial markets, the failure of a systemic bank in either country would have a severe impact on both countries, whether the bank has a subsidiary or branch structure, so that cooperative support appears to be the least-cost solution for both.

Subsidiarization might reduce risks, but could hamper financial integration, for example, where cross-border banks provide substantial foreign funding to the economy and local sources of alternative funding are not well developed. Equally, the benefits of cross-border banking in good times need to be carefully weighed against the benefits of being able to resolve institutions at low risk and low cost in a crisis. Increased cooperation would help countries find a better balance between constraints on cross-border capital transformation and mitigating resolution risks. Having effective dialogue, information sharing, and cooperation agreements for resolution in place would make authorities indifferent about specific bank structures. This would allow financial groups to take advantage of structures that best fit their business models without limiting their ability to manage risks during normal times and support affiliates in stressful times.

See J. Fiechter, İ. Ötker-Robe, A. Ilyina, M. Hsu, A. Santos, and J. Surti, Subsidiaries or Branches: Does One Size Fit All? IMF Staff Discussion Note 11/04, March 2011; and D’Hulster, K. and Ï. Otker-Robe, Ring-fencing cross-border banks: An effective supervisory response? Forthcoming in Journal of Banking Regulation.

Cross-Border Resolution Beyond Globally Systemic Institutions

41. Many cross-border banks are not globally systemically important. The Key Attributes were designed primarily with G-SIFIs and advanced financial systems in mind, and set a high standard in terms of the powers and infrastructures that countries should introduce. While they, nonetheless, have broad application beyond G-SIFIs, embodying best practice and lessons from experience from a range of countries and circumstances, not all elements may be relevant to all countries. In addition, many home and host countries to cross-border banks are not FSB members or party to the fora advancing work on G-SIFI resolution. In some cases, there may be a significant asymmetry of power in interactions between G-SIFI home jurisdictions and smaller host jurisdictions where the operations are not material to the institution’s overall health. This section looks at the resolution reform agenda from the perspective of small host countries and emerging and developing economies.

A. The Key Attributes in Proportion

42. Reforms to enhance resolution regimes should be tailored to the complexity of financial systems in different jurisdictions (Box 6). While the Key Attributes embody best practice in resolution on many core issues that are relevant for all jurisdictions, local circumstances may justify tailoring the adoption of reforms—an approach explicitly recognized in the introduction to the Key Attributes. Elements addressing some capital market functions, or the complex structures and multiple business lines of the largest institutions may not be relevant to countries with less sophisticated financial systems. In many countries insurance companies or securities firms may not present a systemic threat. In some countries it may be sufficient to replicate the economic effects of statutory bail-in powers using other resolution tools (e.g., a bridge bank). While sound resolution planning and cross-border coordination are of equal importance for regionally systemic firms (e.g., a pan-African bank), the detail required in a recovery and resolution plan would be significantly less than for a large, complex G-SIFI.

Key Attributes Beyond G-SIFIs

Many of the Key Attributes could apply to any financial institution, including those that could be systemically important in the jurisdiction concerned. However, Key Attributes 8 to 11, which set out best practice for resolution planning and cross-border cooperation, are only required per se for G-SIFIs,1 while the majority of Fund members are neither key host nor home jurisdictions for G-SIFIs.

The impact of the failure of a bank of systemic regional importance will likely be severe for those countries directly affected, even if not felt globally. Further guidance is needed as to what best practice resolution planning and cooperation arrangements should be implemented by the resolution authorities of those countries in which such an institution is systemic. Fund and Bank staff have proposed the following approach:

  • RRPs and resolvability assessments (Key Attributes 11 and 10) should, as a minimum, be required for all those firms designated as domestic SIFIs in the jurisdiction concerned, using relevant sector-specific methodologies. For banks, for example, this would mean domestically systemically important banks identified using the Basel Committee for Banking Supervision criteria, which include the impact that the failure of the bank will have on the domestic economy.2 RRPs and resolvability assessments should be tailored to the complexity of the firm and its operations.

  • Crisis management groups (Key Attribute 8) or comparable arrangements such as “resolution colleges” should be required for firms that are systemic in more than one jurisdiction and where resolution of entities in more than one jurisdiction is material for resolution of the firm. Comparable arrangements could include less formal structures that provide a framework for the cooperation and coordination necessary to support resolution planning and resolvability assessments for the firm in question. For example, if the firm is systemic and its resolution is material in only two jurisdictions, existing arrangements for cooperation between the resolution authorities in those jurisdictions could be sufficient without the need to formally establish a separate crisis management group.

  • Cooperation Agreements (Key Attribute 9) should be required for any firm for which a crisis management group or comparable arrangement is required under Key Attribute 8. There would also be significant benefits in further work to define best practice cooperation agreements for regional SIFIs.

1 While some of these Key Attributes leave open the possibility of extending to other institutions, no further guidance on which firms and what type of arrangements are specified.2 Comparable frameworks may be developed for other sectors in due course.

43. Further consideration will need to be given to loss-absorbing requirements in banking systems that are primarily deposit funded. As is the case with respect to GLAC for G-SIFIs, banks of regional or domestic systemic importance should have adequate loss-absorbing capacity (LAC) to reduce the risks to the deposit insurance scheme and/or the public sector arising from their potential failure. Further consideration will need to be given as to what form LAC requirements should take in jurisdictions with less developed capital markets, where authorities may be reluctant to allow uninsured deposits to incur losses. In such regimes, it may be appropriate to allow for the substitution of extra capital in a way that recognizes the relative cost and loss-absorbing characteristics of different liabilities. For instance, equity will absorb losses under a wider set of scenarios than bond debt, so that less may be required to substitute for GLAC, while it typically will be more expensive. While retail deposits may be more stable than wholesale deposits, they will not be suitable for absorbing losses.

B. The “Small Host” Perspective

44. For cooperative solutions to cross-border resolution to be credible, incentives need to be aligned between countries. When the operations of a cross-border institution are of systemic importance in a host country, but immaterial to the G-SIFI as a whole, it may not be realistic to expect an alignment of interests between jurisdictions. While legal, business, and reputational concerns can be an important factor in inducing G-SIFIs and their home authorities to assist in finding solutions to problems in host countries, this concern will be in the background if the whole institution’s stability is in question. Furthermore, these small hosts often are not party to crisis management groups or core supervisory colleges where, for practical reasons, participation is limited to countries in which the G-SIFI is most exposed. Therefore, and particularly in a crisis where attention is focused on the most significant exposures and concerns, the authorities of smaller host countries may not have timely access to information that is critical for their domestic financial stability, or to a forum where their perspectives on resolution will be considered.

45. In these circumstances, the resolution strategies of small hosts may have to prioritize domestic financial stability. For small hosts and home jurisdictions of G-SIFIs alike, the authorities’ first priority in a bank failure will be to ensure contagion is contained, critical financial services continue to function, and that economic costs are minimized. Where small hosts cannot ensure their capacity to achieve these outcomes through ex ante agreements on cooperative resolution strategies and burden sharing, measures ensuring local operations are resolvable through structural requirements, such as requiring local capital and liquidity via subsidiarization or asset maintenance requirements for branches, may be appropriate.19 This approach would be consistent with MPE (or hybrid) resolution strategies, and small hosts should work with home countries to ensure this fits into the comprehensive resolution plan for the bank. While this approach may potentially have costs from a global perspective—inefficiencies from requiring local pools of capital and funding, or limitations on the scope for international bank financing of structural external deficits—these costs may be low where operations are “plain vanilla,” collecting local deposits to fund lending, and also bring benefits in reducing prudential risks and increasing resolvability. Cooperation with the home jurisdiction will still have important payoffs with respect to deciding on the location and magnitude of GLAC, how critical services will be maintained, and in building trust to enable smooth resolution, should a crisis eventually occur.


46. Recent initiatives in international regulatory reform have made progress toward achieving an effective cross-border resolution framework. Widespread implementation of the Key Attributes will provide resolution authorities involved in cross-border bank failures with a broad set of tools for executing resolution strategies, minimizing risks to financial stability, and placing the costs of failure on shareholders and private creditors.

47. However, orderly cross-border resolution is still far from assured. Should a large cross-border bank fail today, it appears unlikely that the pitfalls and misaligned incentives that undermined cooperation in the global financial crisis could be avoided. Progress in the implementation of national legal and institutional reforms consistent with the Key Attributes has been mixed. More fundamentally, implementation of the Key Attributes would only partially align divergent national interests and incentives that work against cooperative cross-border resolution strategies. Additional work is required at the international level on remaining gaps including:

  • Defining specific proposals for GLAC (amount, eligible claims, location, and investor base; and

  • Better aligning creditor hierarchies and depositor preference between countries.

48. Progress on minimizing residual risks that public funds will be needed to preserve financial stability is necessary to ensure that resolution strategies are credible. While a credible bail-in option would reduce the contingent need for public financial support, a residual risk remains, creating scope for misaligned incentives to undermine cooperation in resolution. Moving up decisions and actions to the pre-crisis phase could ensure incentives are aligned and authorities can credibly commit to carry through agreed resolution strategies. Such arrangements could cover (i) identifying structural imbalances on banks’ balance sheets in various countries; (ii) assessing costs and benefits of different bank structures and resolution strategies, and taking action to make these consistent with country authorities’ objectives; and (iii) assessing and monitoring the potential for cross-border sovereign risk transfer or outright public costs arising from policy responses in the event of failure, and updating resolution strategies accordingly.

49. A great deal of work remains to be done, and the Fund will continue to support the international reform agenda. Fund staff will continue to promote the adoption of the Key Attributes in its engagement with Fund members. Staff will also continue to participate in various FSB work streams to help address current gaps in the international regulatory agenda, and will keep the Executive Board apprised of future developments.

  • Do Directors agree that progress has been made toward a framework for an effective cross-border resolution regime? Do they agree with the gaps identified in the paper and the key areas that need to be worked on to avoid potential repeats of the misaligned incentives and unilateral actions seen during the global financial crisis?

  • Do Directors agree that greater harmonization of creditor hierarchies, including with respect to depositor preference, is desirable/necessary?

  • Do Directors agree that progress in agreeing on standards for GLAC would help remove important obstacles to effective cross-border resolution? What are the challenges to reaching agreement?

  • Do Directors agree that the approach outlined in paragraphs 35–40 to move up decisions to the pre-crisis phase, including possible consideration of ex-ante changes to bank business models and structure, would help support resolvability and minimize the likelihood of non-cooperative outcomes in a cross-border resolution?

  • How relevant are recent advances for “small host” economies? Do Directors agree that further work to tailor the Key Attributes to the complexity of financial systems in different jurisdictions would benefit the broader Fund membership?