AVOIDING FINANCIAL CRISES*

David G MayesUniversity of Auckland and Bank of Finland

Paper prepared for the Estonian Economists Association 2nd annual conferencePärnu 12th January 2007

Abstract

This paper proposes two major improvements to the process of financial regulation in the EU countries including Estonia to improve the ability to handle financial problems that run across international borders. The first is to implement a credible means of resolving individual cross-border financial institutions in difficulty rapidly and at a low cost. This would prevent bank exit from threatening financial stability. The second is to have a programme of prompt corrective action so that financial institutions can be turned round short of insolvency. Effectively this would involve a system of structured intervention and early resolution comparable to that introduced by FDICIA in the US. While much of it can be achieved by extensive co-operation and agreement among the authorities in the different EU countries involved with each important cross-border bank through a college of supervisors it ultimately requires both the creation of a resolution agency and a switch from private to public law to handle bank resolution.

I Implementing A Framework for Structured Early Intervention and Resolution (Prompt Corrective Action) in Europe

Large pan-European and regional banks are developing in the European Union (EU). However, the existing institutional framework for dealing with cross border crisis has thus far largely neglected the coordination among prudential supervisors, deposit insurance regulators and reorganization authorities that is needed in an explicit drive to try to ensure the minimization of the potential loss to the taxpayer. Indeed, the present safety net framework across borders not only does not have minimization of taxpayers losses as a goal but has embedded in it incentive conflicts that are likely to substantially increase taxpayer losses.

Academics and policy makers alike have made proposals on how to reform the EU safety net in order to reduce the problems of asymmetric information and create an incentive compatible regulatory structure. However, most of these proposals have focused on mechanisms to reduce asymmetric information between prudential supervisors and central banks, and much less attention has been paid to propose mechanisms to align the incentives among prudential supervisors and between them and deposit insurance and resolution authorities.

The importance of this topic was recognized by the European Shadow Financial Regulatory Committee, which devoted its very first report (ESFRC, 1998) to a proposal for dealing with problem banks, in which it recommended establishing a Structured Early Intervention Resolution (SEIR) regime that called for predictable supervisory action for undercapitalized banks culminating in the withdrawal of the bank’s charter before its regulatory capital reaches zero. More recently, the ESFRC (2005) argued that implementation of a version of SEIR called Prompt Corrective Action (PCA) in each individual Member State would contribute to host country supervisors´ trust in home country supervisors. Benink and Benston (2005) also propose SEIR as a mechanism for protecting * This paper draws together material from my previous work and a joint article with Maria Nieto, Bank of Spain, and Larry Wall, Federal Reserve Bank of Atlanta. The views expressed here are my own and should not be attributed to my co-authors or the institutions for which any of us work.

1

deposit insurance funds and taxpayers from losses in the EU, as part of a more broad based regulatory reform. Along similar lines, Mayes (2004) proposes intervention at prescribed benchmarks (ideally above economic insolvency) as a means of offering a plausible policy for coping with the exit of banks whose failure poses systemic risks in the EU.

While PCA is, in our view, one reasonable approach, there are two issues to be addressed before it could be used to set minimum standards in Europe. First, PCA was designed to work with the institutional structure of U.S. bank regulation. Nieto and Wall (2006) identify several institutional changes that would be needed in European bank regulatory institutions in order for PCA to be effective (described in the second section of this article). Secondly, PCA was designed to reduce principal-agent problems in a purely domestic setting where the supervisor as agent is ultimately accountable to his principal, the voters and taxpayers. While the basic structure of PCA would be helpful in an international setting, explicit consideration of cross-border issues would make PCA more effective in addressing the principal-agent problems that arise from the supervision of a cross-border banking group.

The focus of this part of the paper is on making PCA more effective for cross-border banking groups in the EU.1 We take as given that all Member States have adopted a uniform system of PCA that complies with the requirements set out by Nieto and Wall (2006). In recognition of the political problems in implementing an EU-level supervisor, we take as granted the existing supervisory and other regulatory institutions in the EU to the extent feasible. However, in some cases we identify gaps between what exists and what is needed for effective prudential supervision, deposit insurance and reorganization of cross-border banking groups that can only be covered by substantial changes to existing legislation in the Member States. While we believe the general approach to disciplining large cross-border banking groups advocated in this paper provides the best opportunity for an effective system in the absence of EU-level institutions, this paper does not consider the desirability of EU-level institutions and arrangements should they become politically feasible.

The discussion is organized as follows. The first section analyzes the potential problems with the current institutional framework of bank supervision. The second section evaluates the potential contribution of adopting a PCA type regime in setting minimally acceptable supervisory responses. As the second section discusses, PCA was developed for banks operating in the US and, as such, does not address some important cross-border concerns. Thus, the third section considers additional measures that may be taken to supplement PCA and make it more responsive to cross-border issues. The concluding remarks are provided in the last section.

1. Supervisory Discretion and Cross-border BankingCross-border groups increasingly operate as integrated entities with provision of services such as risk management, liquidity management, data processing, and loan evaluation each centralized in one part of the group (though not all services are necessarily centralized in the same country). They often do not have a neat structure of a parent and free-standing locally incorporated subsidiaries but a complex interweaving of branches and subsidiaries that cannot survive on their own. In this context, bank supervisory structures must also be structured for efficient cross-border operations. The need for efficient cross-border prudential supervision implies someone has to be clearly responsible, it needs a clear objective whose attainment can be transparently and objectively assessed and, most importantly, it needs the tools and powers to undertake the tasks efficiently and effectively in practice and in prospect. This has long been recognized in the work of the Basel Committee of Banking Supervisors (Basel Core Principles for Effective Banking Supervision, 1997)2. Some authority has to take the lead, normally one in the 'home' country where the bank or holding company is headquartered, and the other, 'host' country authorities have to co-operate with them and with each other if the system is to work. Moreover, since there are multiple authorities in

1 The related question of the relationship of the bank supervisor to the lender of last resort when dealing with cross-border banking groups is also important but it is beyond the scope of this paper. See Repullo (2004), and Kahn and Santos (2002, 2004).2 The Basel Core Principles for Effective Banking Supervision have been revised in 2006.

2

each country, whose range of powers and competences often do not match, this coordination is very difficult to achieve.3 Each country remains responsible for its own financial stability yet, where there are large cross-border institutions, that can only be delivered realistically if substantial reliance is placed on foreign authorities. In a crisis, national authorities will tend to put their own national interests first, so any process of recognition of international claims in advance needs to be very carefully structured so that the joint actions match an agreed means of addressing and, where necessary, trading off the possibly conflicting interests of the countries involved.4

The present structure of supervision, deposit insurance coverage and bank resolution in the EU largely follows the legal structure of banking groups. As shown in Table 1, prudential supervision, deposit insurance and resolution are generally the responsibility of the regulators of each country in which a bank is incorporated. The principal exceptions are that: (1) a bank subsidiary is supervised as part of the consolidated group by the home country supervisor and as a "solo" entity may be supervised by the home supervisor of the parent bank if the host country supervisor of the subsidiary delegates its responsibility,5 and (2) the host country deposit insurer of a branch may supplement the coverage provided by the insurer of the home country of the bank to bring it up to the host country's level.

The problem with supervising banking groups as collections of separate legal banking charters is that the legal approach does not reflect how these organizations function in practice. A well-known example of cross-border banking regional integration is Nordea (see Table 2), which is currently organized in the form of subsidiaries that operate with a highly integrated operation. This is set to go further if Nordea changes to a branch structure across the whole region under the European Companies Act as currently planned. Indeed if Nordea does change from its current subsidiary structure to one with branches, its legal form will become a much closer match to the actual structure of its current operations. It is actually an illusion that many subsidiaries can somehow be cut off from their parent in the event of difficulty and asked to function on their own, with or without statutory management (Mayes, 2006). As Schmidt Bies (2004) puts it 'entities can be created within the structure of the group to transfer and fund assets [that] may or may not be consolidated for accounting purposes, depending upon their structure.' (p.1). The idea that the various deposit insurers or supervisors can take independent decisions to minimize their losses in these circumstances is thus not realistic.

The interdependence of prudential supervision of banks operating across borders creates a principal-agent relationship between the society (voters and taxpayers) of one country as principal and the various supervisors of the rest of the banking group as the agents.6 The delegation approach has also been used recently to debate financial supervisory issues (Bjerre-Nielsen, 2004). The standard set of principal agent problems are made substantially worse when some of the principals have no direct authority over the agent, such as is the case when supervisors in one country may expose the taxpayers in another country to losses. The problem is, that to the extent the agent follows the interests of the principals, the agent’s incentives will be to follow the goals of the principal that has some direct authority over the agent. That is, when conflicts arise among the principals, the supervisor (agent) is likely to follow the perceived interests of their own country’s government and voters (principle). Eisenbeis and Kaufman (2006) describe the agency problems and conflicts of cross-border banking in general and, in particular, in the EU.

3 This mismatch of responsibilities relates to the different financial sectors – insurance, banking, securities markets – to the different functions – prudential supervision, deposit insurance, crisis resolution – and to the powers each holds under the variety of legal and regulatory systems that currently exist.4 If there is a threat to the financial system as a whole from bank failure or distress, countries tend to permit special measures to be taken, as in the case of the systemic risk exemption in the United States (Mayes, 2006a).5 This delegation is contemplated in Article 131 of the CRD. In addition, according to Article 44, the home country authorities are responsible for the prudential supervision of consolidated banking groups including bank subsidiaries and affiliates in other Member States (Directive 2006/48/EC of the European Parliament and of the Council of 14 June 2006 relating to the taking up and pursuit of the business of credit institutions (recast)).6 See Alessina and Tabellini (2004, 2005) for a discussion of the conditions for the delegation of the tasks to agents.

3

2. Structured Early Intervention and Resolution /Prompt Corrective Action as a Limit on Prudential Supervisors’ DiscretionSEIR was first laid out by Benston and Kaufman (1988) to minimize deposit insurance losses by requiring a series of mandatory supervisory interventions as a bank’s regulatory capital ratio falls.7

One way that this proposal could work is illustrated in table 2 of Benston and Kaufman (1988, p. 64) in which they propose that banks be placed in one of four categories or tranches: 1) “No problem”, 2) “Potential problems” that would be subject to more intensive supervision and regulation, 3) “Problem intensive” that would face even more intensive supervision and regulation with mandatory suspension of dividends and 4) “Reorganization mandatory” with ownership of these banks automatically transferred to the deposit insurer. Although the deposit insurer would assume control of the bank, Benston and Kaufman (1988, p. 68) ordinarily would have the bank continue in operation under the temporary control of the FDIC, or be sold to another bank with liquidation only as a “last resort”. The deposit insurer would remain at risk under SEIR, but only to the extent of covering losses to insured depositors. However, Benston and Kaufman did not expect such a takeover to be necessary, except when a bank’s capital was depleted before the supervisors could act, perhaps as a result of a massive undetected fraud. Because the bank’s owners would realize that the supervisors were mandated to take over a bank while it was solvent (3 percent market value of capital-to-asset ratio), the owners had strong incentives to recapitalize, sell, or liquidate the bank rather than put it to the FDIC.8

A version of SEIR was adopted under the title prompt corrective action (PCA) with the 1991 passage of the Federal Deposit Insurance Corporation Improvement Act (FDICIA). PCA deals with prudential supervisors´ agency problem by first allowing and then requiring specific intervention by the supervisory authorities on a timely basis.

Whereas SEIR sketches out how supervisors would respond to a drop in capital adequacy, PCA provides a list of actions the supervisors must take and another set of actions the supervisor may take to further the goals of PCA (minimizing losses to the deposit insurance fund). While PCA reduces supervisory discretion as a bank’s capital level falls, supervisors retain substantial discretion over almost all banks. Even the “mandatory provisions” often include a significant element of supervisory discretion. For example, while an undercapitalized bank must submit a capital restoration plan, the supervisors have discretion over whether the plan will be approved as “acceptable.”

PCA may appear to be simply a set of supervisory corrective measures that should be taken as a bank’s capital declines that any country could easily adopt. However, PCA is unlikely to work as intended if a country has not accepted PCA’s underlying philosophy or lacks the necessary institutional prerequisites. Focusing specifically on the EU, Nieto and Wall (2006) identify three important aspects of the philosophy underlying PCA: (1) “that bank prudential supervisor’s primary focus should be on protecting the deposit insurance fund and minimizing government losses,” (2) “that supervisors should have a clear set of required actions to be taken as a bank becomes progressively more undercapitalized,” and (3) “that undercapitalized banks should be closed before the economic value of their capital becomes negative.” The four institutional prerequisites identified are: (1) supervisory independence, and accountability; (2) adequate authority, (3) accurate and timely information; and (4) adequate resolution procedures. They find that European countries currently comply with these institutional requirements to varying degrees.

The adoption of a version of PCA would provide the EU with a set of minimum supervisory responses to violations of the Capital Requirement Directive (CRD).9 The definition and level of the capital ratios that would trigger mandatory supervisory action and eventually intervention is a 7 See Benston, George J. and George G. Kaufman (1988). For a discussion of the intellectual history of PCA see Benston, G., and Kaufman, G. (1994). 8 Table 2 in Benston and Kaufman (1998) gives “Illustrative Reorganization Rules” with mandatory reorganization at a 3 percent market value of capital-to-asset ratio. However, the text talks about the possibility that this ratio should be revised upwards.9 Directive 2006/49/EC of the European parliament and of the Council of 14 June, 2006 on the capital adequacy of investment firms and credit institutions (recast). Official Journal of the European Union L177/201 30 June, 2006.

4

relevant subject that is beyond the scope of this paper. Moreover, the original PCA was designed to address principal-agent problems in the supervision in the US and does not explicitly contemplate the complications introduced by cross-border banking groups. A number of authors discuss the merits of adopting PCA in the EU, including in some cases the recognition of the gains from using PCA in supervising cross-border groups. However, none of these authors (Nieto and Wall, 2006; Benink and Benston 2005; Mayes 2004 and 2006) and policy analyst recommendations (ESFRC, 2005) explicitly consider the changes needed in the EU if PCA is to be effective in resolving the cross-border agency problems that arise in supervising cross-border banking groups.

3. A Prompt Corrective Action for Cross-Border Banking Groups in the EUBanks operating under PCA can fall into one of three categories: (1) adequate capital, (2) undercapitalized but still having a good chance of rebuilding its capital, and (3) sufficiently undercapitalized that the bank should be placed into resolution to minimize the losses. Cross-border banking groups that are being supervised by national banking supervisors introduce additional supervisory challenges in each of these three categories. The following subsections consider those challenges and recommends additions and modifications of PCA adopted with the 1991 passage of the FDICIA to address the challenges of cross-border groups in the EU.

3.1 Assuring accurate and timely information of banking groups financial condition

In order for bank supervisors to use their powers effectively, they must have an accurate understanding of the bank’s and banking group’s financial condition. A potential problem for a prudential supervisor of a cross-border banking group is that of determining the status of those parts of the group outside its supervisory control.

The need for information sharing among the supervisors is recognized in the CRD, Article 132, which establishes that the "competent authorities shall cooperate closely with each other. They shall provide one another with any information which is essential or relevant for the exercise of the other authorities' supervisory tasks under this Directive. In this regard, the competent authorities shall communicate on request all relevant information and shall communicate on their own initiative all essential information. […] Information shall be regarded as essential if it could materially influence the assessment of the financial soundness of a credit institution or financial institution in another Member State. In particular, competent authorities responsible for consolidated supervision of EU parent credit institutions and credit institutions controlled by EU parent financial holding companies shall provide the competent authorities in other Member States who supervise subsidiaries of these parents with all relevant information. In determining the extent of relevant information, the importance of these subsidiaries within the financial system in those Member States shall be taken into account." This obligation for information expands to encompass also: "(c) adverse developments in credit institutions or in other entities of a group, which could seriously affect the credit institutions; and (d) major sanctions and exceptional measures taken by competent authorities in accordance with this Directive, including the imposition of an additional capital charge under Article 136 … ."

These provisions for information sharing have also been strengthened with the adoption of Pillar 3 of the new Capital Accord.10 For example, banks are required to report total and Tier 1 capital ratio for the consolidated group and for significant bank subsidiaries. In this case, the host supervisors of the subsidiaries could use this information that would be reflected in a market indicator, as justification for triggering consultations with the home country supervisor and/or for triggering a special examination of the banking group.11

10 Pillar 3 aims to encourage market discipline by developing a set of disclosure requirements which will allow market participants and foreign supervisors to assess relevant pieces of information on the scope of application, capital, risk exposures, risk assessment processes, and hence the capital adequacy of the institution. Since domestic supervisors typically request additional information from the banks it is unlikely that this public disclosure will be thought sufficient.

5

While the information sharing mandated by the CRD should provide national supervisors with the information they need, ad hoc sharing on a banking group by banking group basis is likely to be inefficient and leave room for gaps in information sharing. A better alternative would be the creation of a single database on banks´ financial condition so that all prudential supervisors can understand the condition of the group as a whole and its relationship to the bank they each supervise. The European Central Bank (ECB)or the Committee of European Banking Supervisors (CEBS) could harbour that database. In the case of the ECB, this responsibility would be consistent with article 105.5 of the EC Treaty: "the ESCB shall contribute to the smooth conduct of policies pursued by the competent authorities relating to the prudential supervision of credit institutions and the stability of the financial system." The Mayes (2006b) and Vesala (2005) proposal of a common data base would minimise the information asymmetries between home and host prudential supervisors although their proposal is limited by the professional secrecy imposed by article 44 of the Directive 2006/48/EC of the European Parliament and of the Council of 14 June, 2006 relating to the taking up and pursuit of the business of credit institutions (recast).12

While measures may be adopted at the EU level to enforce the sharing of hard information (verifiable facts) such as financial statements among prudential supervisors, the sharing of soft information (or nonverifiable judgments) cannot be compelled. For example, a national supervisor may consider that a bank’s financial statements overstate the true capital of the bank. But if the supervisor wants to exercise forbearance, it can do so by doing nothing, neither compelling the bank to revise its financial statements nor sharing the additional information with other supervisors.

Nieto and Wall (2006) note that the enforcement of PCA depends on the accuracy of reported capital adequacy ratios. They survey several studies suggesting that market signals, primarily subordinated debt spreads, provide useful information about banks’ financial conditions and that in some cases these signals have proven more accurate than the banks’ reported Basel I capital ratio. They find some authors (Sironi, 2001; Evanoff and Wall, 2002; Llewellyn and Mayes, 2004) who are persuaded that the information is sufficiently reliable to be of use at least in setting a backstop for critically undercapitalized organizations. We concur that the use of market risk measures would provide a valuable supplemental measure for PCA.

However, supervisors have proven reluctant to use market signals to determine the capital category of banks operating under PCA. A less controversial and perhaps easier approach to implement would be to use market risk measures as triggers for closer supervisory scrutiny of a bank. These measures could include subordinated debt spreads but they could also include other measures such as the pricing of credit derivatives or equity based measures, such as Moody’s KMV Expected Default Frequency. The measures could be used informally by individual supervisors to trigger closer scrutiny of the various parts of the group. The use of such market measures would be consistent with Pillar 2 of the new Capital Accord, which requires supervisory review of bank’s reported capital adequacy and with Pillar 3, which seeks to encourage market discipline. Market risk measures could further be used to trigger a mandatory meeting of the college of supervisors (see definition in p.9) to review the group’s condition and, when appropriate, for triggering a coordinated special examination of the banking group.

3.2 Co-ordination of PCA disciplinary measures short of resolutionAlthough PCA reduces supervisory discretion, some element of discretion is inevitable. While a supervisor can be compelled to employ some measures, the choice of what limits the risk best and reduces any impending loss is bound to be substantially case specific. For example a measure, such as replacing existing management, which might be essential to restore the banks´ financial health in some cases, could be counterproductive in other cases.13

11 The required level of disclosure is both limited in its relevance and its timeliness (Mayes, 2004). In this author's view, the requirements fall well short of what has been required of banks in New Zealand since 1996, where disclosure statements are required quarterly, have to reveal peak exposures and where bank directors are legally liable for their accuracy.12 L 177/ 1 OJ of 30 June, 2006.

6

The existence of supervisory discretion raises the possibility of a supervisor taking or failing to take a variety of actions that are harmful to the overall banking group but which yield net benefits to the supervisor’s particular country. An example where taking action could be harmful to the group but benefit the supervisor's country would be that a supervisor could impose draconian limitations on a bank that is small relative to its financial system, but where the bank provides valuable services to the rest of the group elsewhere. An example, where failing to act could be harmful would be forbearance on the part of the home country supervisor of a banking group that has a large presence in its country. Such forbearance could take the form of a supervisor accepting inadequate capital restoration plans and imposing only the minimum disciplinary measures required under PCA even though additional measures are likely to be necessary to rebuild the bank’s capital. The consequences could be that weakness at the group level that would adversely impact subsidiaries (even the banking systems) in other countries and may substantially raise the cost of resolving the group should it become insolvent.

The EU has some mechanisms that could be extended to provide an element of coordination in the use of discretionary measures. The CRD provides for some coordination of banks supervision and allows for the delegation of some supervisory responsibilities to another Member State’s prudential supervisor. Article 131 establishes that "in order to facilitate and establish effective supervision, the competent authority responsible for supervision on a consolidated basis and the other competent authorities shall have written coordination and cooperation arrangements in place. Under these arrangements additional tasks may be entrusted to the competent authority responsible for supervision on a consolidated basis and procedures for the decision-making process and for cooperation with other competent authorities, may be specified. The competent authorities responsible for authorizing the subsidiary of a parent undertaking which is a credit institution may, by bilateral agreement, delegate their responsibility for supervision to the competent authorities which authorized and supervise the parent undertaking so that they assume responsibility for supervising the subsidiary in accordance with this Directive." Thus, the CRD provides for a general mechanism of coordination and cooperation among supervisors and it also envisages a stronger form of coordination, which is the possibility that the host supervisor of a subsidiary may delegate its responsibility to the home country prudential supervisor of the subsidiary’s parent.

The primary problem with using the authority provided by the CRD is that delegating supervisory responsibility to the home country supervisor of the parent bank is likely to worsen the principal-agent conflict between the parent’s supervisor and the subsidiary’s country’s taxpayers and voters as principal. The parent’s supervisor would be responsible for the impact of its supervisory action on the deposit insurance fund and possibly the financial stability of the host country of the subsidiary, but the parent’s supervisor would not be directly accountable to the government and the tax payer of the subsidiary’s host country, hence, increasing the agency problem.

Another mechanism for the coordination of discretionary PCA actions would be that of a college of the prudential supervisors of the banks in the group.14 The college would be fully compatible with Article 129 of the Directive 2006/48/EC of the European Parliament and of the Council of 14 June, 2006 relating to the pursuit of the business of credit institutions (recast), which envisages the cooperation of the consolidating supervisor with the competent authorities of the subsidiaries.15 The coordination mechanisms could be merely advisory, leaving the final decision up to the national supervisors of each bank, or binding upon the members. In some cases allowing each supervisor to take disciplinary action may be acceptable, especially if the action would be unlikely to have adverse consequences on other group members. However, leaving the final decision in the hands of

13 It is assumed that the authorities in each of the EU countries would have a similar if not identical range of powers available for PCA, when faced by the same circumstances. Currently this is far from the case and, although the toolkit may be similar, what can or must be done in each circumstance varies considerably.14 While there is common cause about the importance of the role of the consolidating, lead or coordinating supervisor, the meaning of these terms varies considerably across the CRD, the European Financial Services Roundtable (2004) and CEBS (2006). 15 L 177/48 Official Journal of the European Union of 30 June, 2006.

7

each bank’s national supervisor may not result in effective coordination to the extent that different supervisors reach different conclusions about the appropriate actions either because the supervisors have different incentives or because they have reached different judgments. Thus, for an effective implementation of a PCA policy as a coordination mechanism between supervisors, a better solution may be to give the authority to make decisions about discretionary actions that will be binding on all prudential supervisors to the college (see Appendix for a description of different scenarios of collegial binding decision). The idea behind such a grouping is that the supervisors can become in some sense jointly responsible for the actions the group takes. In such a case it may then be easier to agree remedial actions and even burden sharing in the event of loss.

Ideally, a college of supervisors for each cross-border banking group will have been formed before the need arises to invoke PCA’s disciplinary provisions. However, the formation of a college with authority to make discretionary decisions within the PCA policy framework should at the least be mandatory as soon as a bank owned by a cross-border banking group falls below the capital standard.16 The formation of the college does not entail that decisions will always be made in a timely and harmonious fashion. Even the best of colleges is likely to be an inefficient mechanism for addressing most issues that require consultation or negotiation with the banking group. For example, if a cross-border banking group with capital below the minimum capital requirements is required to develop a capital restoration plan that is acceptable to its supervisors, having the bank negotiate the plan with each of the college members would be slow and inefficient. Where such consultation or negotiation is required, a better alternative would be for the committee to select one supervisor as the primary contact (typically the parent’s supervisor) with the bank unless the problems are focused in particular subsidiaries/markets. The role of the college would then be to review and approve the contact supervisor’s agreement with the bank.

For a variety of reasons, a college of supervisors may at times have problems reaching a decision. One way of forcing timely action would be for PCA to establish a presumption that a certain action will automatically be effective within 30 days (or 60 or 90 days) after a bank violates one of the PCA triggers unless the college determines that taking the action will not further the purposes of PCA. Similar provision is envisaged in Article 129 of the Directive 2006/48/EC of the European Parliament and of the Council of 14 June, 2006 relating to the pursuit of the business of credit institutions (recast), which foresees that the consolidating supervisor will decide in a time framework in the absence of a joint decision. This would prevent a subset of the college from using committee deliberations to stall effective action. Additionally, the colleges may somewhat reduce the scope for relatively unimportant disagreements to stall decision making by adopting decision rules which give greater weight to the judgments of supervisors of the larger banks in the group and the supervisors from countries where the banking group is systemically important.

Although a college provides a mechanism for all affected Member States to have a voice in the corrective measures´ decision taken under PCA, the college does not completely solve the agency problem caused by the mismatch between supervisory powers and supervisory accountability to voters. Giving each country’s supervisor a say in a coordinating college is not equivalent to the power that the supervisor would have to protect its country’s interests as would be possible with a purely domestic bank. However, the inability of supervisors in each country to have the same control as they would over a purely domestic group is an unavoidable consequence of groups operating as integrated entities in more than one Member State. Corrective measures taken (or left untaken) will have sometimes different consequences for different countries.17 The best that can be said is that a college structure will typically provide better representation of each of the affected countries than would a system that gives all of the power to a single supervisor, hence, reducing the agency problem by increasing supervisor's accountability to the government and the tax payer.

16 There is a clear complexity if responsibility for ongoing supervision and resolution (whether or not least cost) belong to different agencies17 Giving every supervisor a veto over taking an action would not prevent problems if failure to act would have large adverse consequences for some country. Similarly, giving every supervisor a veto over failing to act would not help if taking a given action would have large adverse consequences for some countries.

8

3.3 Coordination of resolutionPCA requires timely resolution, which is to say it sets a hard boundary which, when crossed by the bank, requires that the bank be forced into resolution.18 Timely resolution of banks can enhance financial stability in a variety of ways. First, the lack of a deposit insurance subsidy to risk taking and the threat of losing the bank’s charter may deter the bank from taking excess risk. If problems should arise, the bank has an incentive to quickly rebuild its capital or sell itself to a stronger bank before the supervisors withdraw the bank’s charter.19 Moreover, in the case of the license withdrawal, depositors and other non-subordinated creditors are less likely to run on a failed bank if they know that deposits are backed by assets or guarantees. Last but not least, the smaller the losses to deposit insurance and governments, the less of a problem it will be to allocate those losses across the various insurance schemes and the less likely any deposit insurance is to renege on its obligations. In a PCA cum closure rule at a positive level of regulatory capital, losses will be by definition smaller than in the absence of PCA to the extent that deposits would be backed by assets of at least the same market value, except in the case of rapid decline in asset value, massive fraud or inadequate monitoring by the regulatory agencies.

If this hard boundary is to be credible, Nieto and Wall (2006) argue that it must be accompanied by a credible process for resolving insolvent banks. Supervisors will justifiably be reluctant to force a bank into resolution if forcing a bank into resolution is almost certain to cause systemic impacts that will adversely impact their country’s real economy because of inadequate resolution procedures.

In the EU, there is no a framework of commonly accepted standards of bank resolution practice including a common definition of bank insolvency and a fully-fledged single legal framework or a common decision-making structure across Member States. Hadjiemmanuil (2004) argues that a single pan-European legal and administrative framework for bank resolution is not only still lacking but also it is unlikely to emerge in the foreseeable future. As a result, bank resolution procedures largely depend on national laws. These national laws often fail to meet many of the requirements for a credible, efficient resolution system. Even if consideration is limited to the requirements for a large domestic bank group operating in a single country, most EU countries lack an adequate system. Nieto and Wall (2006) highlight two requirements that are generally not met by EU national resolution systems: (1) the need for special bankruptcy provisions for banks in which a banking authority is given authorization to create and operate a 'bridge' or similar bank,20 and (2) a requirement that depositors be provided prompt access to their funds. These weaknesses in EU national resolution systems are likely to give policymakers little choice but to recapitalize a large, deeply insolvent banking group.

Additional problems arise when the banking group subject to resolution operates across borders. Goodhart and Schoenmaker (2006) analyze these problems associated with the resolution of cross-border banks. Although their analysis focuses on the issues associated with recapitalizing a distressed bank operating across Member States boundaries, many similar issues are likely to arise with a bank forced into resolution.21 The following subsection summarizes their key findings and the next subsection discusses how the issues would be addressed in a PCA framework.

18 SEIR simply calls its lowest category “mandatory reorganization.” Banks in PCA’s “critically undercapitalized” category are to have a receiver or conservator appointed within 90 days unless the supervisor can show that another action would better meet PCA’s goal of minimizing deposit insurance losses.19 Kane, Bennett and Oshinsky (2006) find evidence that distressed banks are more likely to recapitalize or sell themselves in the period after the adoption of PCA than in a prior period.20 In the US the most obvious way to do this in the case of a large bank is to form a 'bridge bank', which is a national bank newly chartered by the Comptroller of the Currency under the control of the FDIC.21 Goodhart and Schoenmaker (2006, p. 37) note that early closure of a bank as provided for by the U.S. version of PCA would “reduce the problem.” Their focus on recapitalization presumably reflects their views about the political viability of adopting PCA in Europe for the foreseeable future rather than its economic merits.

9

3.3.1 Recapitalizing a cross-border banking group in the absence of PCAThe withdrawal of the charter of a cross-border banking group, especially a large group, could have severe adverse consequences for the financial stability of one or more Member States. Given the limitations of other existing EU resolution options, the only option that is likely to forestall financial instability may be for the affected Member States to recapitalize the bank at taxpayer expense. However, disagreements about whether a bank should be recapitalized and, if so, how the burden should be apportioned are likely to delay action until the market losses confidence in the bank.

By the time confidence is lost, the time for organizing a recapitalization will be very short (likely only a few hours) and the costs of recapitalization are likely to be a substantial fraction of the bank’s assets. Without any "ex ante" agreement on sharing the cost of recapitalization, the country most affected may be forced to decide whether to bear all of the recapitalization cost or to let the bank be forced into bankruptcy proceedings where liquidation is possible. While this may be the largest country this is by no means certain. Nordea, for example is more important in Finland than it is in the home country, Sweden. Small countries may simply not have the resources for such a recapitalization and will hence be forced into having the crisis.

An alternative to negotiating an agreement during a crisis would be for an "ex ante" agreement on burden sharing involving the various national ministries of finance. There are several ways in which such an ex ante agreement could be structured. Goodhart and Schoenmaker (2006), for example, recommend that all countries in which the bank operates share the burden according to some measure of the operations that the bank has in their country, assets being their preferred measure. However, obtaining agreement on any single measure (a proxy) for a fair distribution may be difficult. For example, assets may not be a good proxy for the real and financial impact of a bank’s failure. Such impact may depend, for example, on the structure of the local deposit market or on the bank’s role in the country’s securities and derivatives markets.

It is also not clear how decisions would be taken. Access to pubic funds is presumably a matter for the relevant ministries of finance. However, ministries of finance would no doubt want to be advised by supervisors, deposit insurers and central banks. Whether they should all sit round the table or whether different parties should meet for different purposes during the process of managing the problems is an open issue. Goodhart and Schoenmaker (2006) recommend that all three parties from each of the countries being there in addition to EU level representation from the Committee of European Banking Supervisors (CEBS), the European Central Bank, ECOFIN and the European Commission, subject to a ‘de minimis’ threshold of 5 percent of the group’s assets and 15 percent of the country’s banking assets.

3.3.2 Resolution of a cross-border banking group under PCAA version of PCA that was effective for groups operating only in one country would by itself substantially reduce the problems of resolving a large cross-border banking group. PCA provides for early resolution (charter withdrawal) before a bank can incur losses substantially in excess of its regulatory capital.22 At best, such a PCA would give supervisors time to organize an orderly resolution of a problem bank because it would result in the bank’s charter being withdrawn while creditors were confident the bank had sufficient assets to honor their claims. More likely, given the U.S. experience, some bank runs will occur because at least some uninsured creditors are likely to take losses in bank resolutions and will act to protect themselves. However, even if market participants control the timing of the bank resolution, PCA will still reduce the problems of resolving a failing banking group. PCA’s requirement that bank charters be withdrawn at positive values of bank´s regulatory capital should substantially reduce the losses to tax payers and significantly reduce any conflicts over how best to share the burden. The losses may even be sufficiently low so that they can be absorbed by the banking industry through payments to their deposit insurer.

An EU version of PCA designed to resolve cross-border banking groups could significantly improve the efficiency of bank resolution, further lowering the overall cost of resolution. The first 22 Such a PCA would include a credible resolution mechanism as advocated by Nieto and Wall (2006).

10

part of cross-border resolution version of PCA would require that that the parties to the process start meeting as soon as a bank falls below the minimum capital standard required by the CRD. Market participants will look for signals that supervisory intervention is likely so that they can reduce their credit exposure before the supervisors act. Thus, the formation of a college to decide the status of a bank (resolution college) should occur either automatically well before resolution becomes likely or it will have to be kept secret (although it is unlikely that it will remain so). Early formation of the college would allow all concerned safety net regulators to plan for the possibility that the bank will need to be recapitalized or resolved, without sending the signal that the supervisors consider such action likely. Thus, should a run occur at a banking group, the relevant safety net regulators and finance ministries will be prepared to take appropriate action.

The college that decides how best to resolve a bank will need to reflect the views of most, if not all, of the participants in the Goodhart and Schoenmaker (2006) proposal. Even if the bank is closed without any losses to the taxpayer, at least some finance ministries/national central banks may need to advance funds to the deposit insurer to cover the insurer’s share of the losses, in part because some deposit insurers collect funds on an "ex post" basis. In theory, such support by national governments is limited by the Directive 94/19/EC on deposit insurance, which discourages governments from providing funding to their deposit insurer and support by the central bank is limited by EC Treaty (article 101). In practice, these restrictions may not prove viable given the importance of giving depositors immediate access to their funds discussed in Nieto and Wall (2006) and the limited funds available to many deposit insurers. The burden sharing proposals of Goodhart and Schoenmaker (2006) implicitly recognize this possibility.

While the college planning the potential resolution cannot know for certain whether or how much losses will be incurred in resolving the bank, there could be disagreements about how to share any costs that do arise. One method of allocation would simply be to assess for each insurer the amount needed to cover losses to insured depositors in the bank or banks covered by the insurer. The losses allocated under this procedure, however, will depend in part on the gains from keeping the banking group together so that the group retains any going concern value and so that the group can be sold to its highest value. However an ex ante agreement on burden sharing may turn out to be more workable in practice, as suggested by Goodhart and Schoenmaker (2006).

It is likely that the balance of interests needed to be taken into account in deciding whether to intervene will also be appropriate for decision-making about the subsequent resolution of the bank. The fact that a bank had to be put into resolution suggests that a quick sale of the entire group is unlikely. The group is likely to have arranged such a sale before resolution if that were possible. Thus, the resolution of almost all large cross-border groups is likely to involve their being operated as some equivalent of a bridge bank (or bridge banking group) pending the return of its assets to the private sector. A bridge banking group would be roughly equivalent to a governmental recapitalization except that the shareholders in the failed group would lose their claim on the group and losses may be imposed on some classes of creditors (especially the subordinated creditors). Someone will have to have managerial authority over the bank and in almost all cases the home country supervisor will be the logical party to appoint the new management. The bank's management should be overseen by a college with representatives from all of the affected Member States, perhaps reduced by the same de minimis rule used before the bank went into resolution. In effect, the college would serve as a replacement for the group's board of directors as the original directors would no longer have any role in the governance of the group. Whether each nation needs to be represented by its banking supervisor, its ministry of finance and its national central bank may depend on the circumstances. If the respective national ministries of finance or national central banks are not making an important contribution to the resolution then they should probably be dropped from the oversight college to help keep its size manageable.

11

The conflicts between different stakeholders will not end after the formation of a bridge bank.23

The managers and college of overseers of the bridge bank will have a variety of decisions to make that could provoke sharp controversies. One such decision is where the banking group should continue lending and where it should reduce or stop lending. Those countries and industries facing reduced lending may be concerned about the impact of the cuts on their domestic economic activity. However, having the bank continue to lend to loss making geographical areas and industries is likely to provoke concerns from some committee members about the likely losses to the bank. Another potentially controversial decision is that of closing some branches and subsidiaries. The managers may also recommend these closures to improve the efficiency of the surviving organization. Again, those Member States that face the cuts may view the situation differently from those that are concerned about further losses. A third potential source of controversy is the weight given to various considerations when the group’s assets are returned to the private sector. Many in the college overseeing the bridge bank (formerly resolution committee) will likely favor accepting the highest bid for the group (or parts of the group) but others on the college may want to include other considerations, such as any labor force reductions planned by the prospective acquirers, or keeping the national charter of the bank.

4. ConclusionPCA was designed to improve the prudential supervision of banks in the U.S., most of which operate in a single market. An EU version of PCA could also improve the prudential supervision of banks operating in more than one Member State. However, to be as effective as possible, the EU version should address a number of cross-border issues that are compatible with the existing decentralized structure of the EU safety net.Bank supervisors need to understand the overall financial condition of a banking group and its various individual banks if they are to effectively anticipate problems and take appropriate corrective measures. The EU could use PCA to enhance the availability of information to prudential supervisors as well as supervisor's use of market information. Availability could be improved by enhancing information sharing requirements on individual bank's financial condition as a part of the adoption of PCA. The use of market based risk measures could be mandated in the supervisory process. At a minimum, this would include requiring additional examinations of banking groups whose reported capital exceeds minimum required levels but which are identified as high risk by financial markets and mandating that the relevant banking supervisors meet to share their evaluations of the group.PCA reduces supervisors’ ability to exercise forbearance, but it by no means eliminates supervisory discretion. Supervisors retain substantial discretion in their implementation of PCA so long as a bank’s regulatory capital exceeds the critical level at which it is forced into resolution. If the consequences of bank supervision in one country can have large consequences for the group’s banks in other countries, then deciding how best to exercise this discretion should be decided by the supervisors of all the banks (or at least all of the significant banks) on a collegial form. However, even if a satisfactory means of deciding what to do can be implemented, the actual powers of supervisors in the EU are not identical. Some may not be able to implement the actions others wish to vote for. Hence, effective implementation would require as a precondition that prudential supervisors be given the same authority to take the corrective measures in PCA.

Finally, should a bank that is part of an integrated cross-border banking group reach the point where PCA mandates resolution, its resolution could have implications for a number of Member States. The timing of the resolution is unlikely to remain in the supervisor’s hands, so the process of making these decisions needs to begin before markets perceive that the bank must be resolved. The

23 The same sorts of conflicts are likely to occur under the current system if the national ministries of finance decide to recapitalize a distressed bank. To the extent the various ministries hold a sizeable part of the bank’s stock, they will likely expect to participate in the decisions of the bank before privatization and also in the decisions on how best to privatize the bank.

12

parties from each country that will play a role in the resolution (the banking prudential supervisor, the ministry of finance and the national central bank) should begin planning for the resolution with the appropriate EU institutions and the ECB no later than the time the bank first falls below the minimum capital adequacy requirements set in the CRD. In a PCA cum closure rule at a positive level of regulatory capital, losses will be by definition smaller than in the absence of PCA to the extent that deposits would be backed by assets of at least the same market value, except in the case of rapid decline in asset value, massive fraud or inadequate monitoring by the regulatory agencies. In almost all cases, the best resolution of a large cross-border bank will involve the creation of the equivalent of a bridge bank or bridge banking group. This would require special bankruptcy provisions for banks in the EU. A number of additional decisions will then be needed as to how to run the bridge bank(s) until its assets are returned to the private sector as well as decisions about how best to return the assets to private owners. Thus, on-going oversight of the bridge bank should be provided by a board with safety net regulators from all of the affected Member States (banking prudential supervisor, ministry of finance and national central bank) perhaps reduced by the same de minimis rule used before the bank went into resolution.

II A Route for Allowing the failure of Large Complex Cross-Border Banks Without Causing Systemic Problems

In the United States, Canada and a number of other countries (LaBrosse and Mayes, 2006) it is the duty of the deposit insurer to minimise the losses to itself from a bank failure. This entails not simply that the insurer should apply a technique that results in the lowest cost in the event of failure but that it should manage its risk exposure all the time and ensure that the system of regulation, supervision, corrective action prior to failure and the procedures set up for handling failure all contribute to ensuring that the losses are minimised.

If the deposit insurer is in effect a junior claimant on the bankruptcy estate, by virtue of succeeding to the insured claims of depositors then in minimising its own losses it will be working in line with the rest of the creditors and ensure that the return they get is maximised. It is normally the duty of the receiver to maximise the value of an insolvency estate, which should be equivalent to minimising the loss provided the value of the estate is measured after the various charges including those from the receiver. If the deposit insurer is higher up the hierarchy of claims then it has an incentive to make sure that the more junior claimants bear all the loss after the shareholders have been wiped out. However, this presupposes that all those encountering losses from the failure can join in the claims on the estate. Many concomitant losses will not be covered – borrowers may have to pay more for new loans, if assets are concentrated then their sale may reduce their price to the detriment of other asset holders and so on.

There is therefore a case for considering minimising some wider concept of loss when dealing with the framework for bank failure and the event of failure itself. In the case of the United States, this is only spelt out if the external loss is considerable,24 then the FDIC is entitled to request that the ‘systemic risk exemption’ be invoked. In this case it could then take into account the wider impact on society.25 In practice this has not been exercised but other countries, for example Finland, Norway and Sweden in their crises about 15 years ago, did take steps to limit the contagion to other parts of the economy, by, inter alia, issuing blanket guarantees, or allowing a public agency to become the ‘owner of last resort’ (Moe, 2004).26 This fear of contagion is one of the fundamental

24 The Act uses the words ‘serious adverse effects on economic conditions and financial stability’ (12 USC 1823(c)(4)(G)).25 FDICIA makes a further important proviso in this regard, namely, that the systemic risk exemption can only be used if not only is there a risk of serious harm but that the FDIC can mitigate the effect by using techniques not available to it under the ordinary working of the Act.26 The techniques the three countries used differed, as set out in Moe et al. (2005), which explains in some detail why in the Norwegian case a blanket guarantee was not thought the best way to go. In Norway the government ended up owning 60% of the banking system in 1993.

13

explanations for having deposit insurance in the first place (Diamond and Dybvig, 1986).27 If depositors know they will be safe then they will not start a run on the bank, thereby rapidly sparking a crisis and – more importantly – they will not start runs on other solvent banks in the pursuit of rumours that they too may be under threat.

This wider concern for contagion forms part of what has been formalised more thoroughly as the desire for financial stability (Schinasi, 2005). There is an industry growing up in defining what this concern for stability is (Allen and Wood, 2006) but their definition is a good description: ‘financial stability is a state of affairs in which an episode of financial instability is unlikely to occur, so that fear of financial instability is not a material factor in economic decisions taken by households or businesses. … Episodes of financial instability are episodes in which a large number of parties, whether they are households, companies or (individual) governments, experience financial crises which are not warranted by their previous behaviour, and where these crises collectively have seriously adverse macroeconomic effects.’ (pp.8-9).

A narrow interpretation would suggest that, even for the claimants themselves, taking the spill overs into account would reduce their losses, as it would reduce the fall in market values that would otherwise occur without the intervention. Both they and the rest of society will benefit from the greater stability. Since stability is a rather soft concept and highly nonlinear in the way it operates it will be difficult to judge how extensive the public sector response should be. Here I therefore suggest approaches that are clearly defined and hence strictly limited in any cost that might be imposed on the insurance fund or the taxpayer.

The discussion is easiest to delimit if the circumstances are simply the failure of a single institution for ‘idiosyncratic’ reasons. However, it is often difficult to determine whether the failing institution is simply the most exposed of several institutions and hence any discussion of the resolution method employed must effectively allow for the possibility of a stream of actual or feared insolvencies in other financial institutions. It is therefore not possible to organise this on a highly discretionary basis or the result will be both misleading and inequitable. In the same way ‘constructive ambiguity’ will not be helpful as banks and their creditors will expect, in the absence of explicit history or a plan, that they will be favourably treated on the basis that the problem will be too difficult or too complex to allow them to fail.

This part of the paper deals with how the problems of failure can be dealt with in a predictable manner to handle the wider worries about financial system stability without exposing the taxpayer or those not directly exposed to extensive cost and giving strong incentives to those owning and managing banks to run them prudentially.

1 The issueIf a financial institution that is small compared to the markets it operates in fails, the ripples that spread out to the rest of the financial sector and the economy at large are also likely to be small, largely irrespective of the methods used for resolving the problem. Only ‘largely irrespective’, because methods which revealed that insolvency procedures were inefficient or inequitable would lead people to reappraise the risks they were taking. Such spill overs need not be negative as methods could reveal a better set of outcomes than expected. The spill overs can also be large if the economy is in a period of weakness when many institutions are under threat. However, it is not correct to consider these effects on a case by case basis as the expectation of what will happen under insolvency will be built up by the sequence of cases.

If an institution that is large in its markets fails then the methods used in resolution can have very different effects on the size of the ripples. This part of the paper explores methods that can be used to minimise the size of the external consequences, particularly the use of what are known as bridge banks in the United States, and builds on the ideas in Mayes et al (2001) and Mayes and Liuksila (2003). The concern is to have methods of resolution that enable functions of the bank that are crucial to the maintenance of financial stability to remain in operation without a break (Hüpkes,

27 The other main reason normally advanced is to protect the interests of ordinary depositors, who are not in a position to be adequately informed about the risks they face or avoid them readily.

14

2004; Harrison, 2005) while nevertheless allowing the authorities to effectively take over the running of the bank from its previous owners. Thus the business continues even if, as in the US case, the bank is failed as a legal entity and reborn again as a new institution under FDIC control. 28

Approaches to this problem have been developed in both Switzerland and New Zealand and the necessary structures and rules exist in the US but have fortunately not been called into use since the passing of FDICIA.

It is not at all clear in many countries round the world whether they have such a coherent and prima facie effective system for keeping systemically essential functions of the banking system running in the event of a failure, without following some strategy of open bank assistance. Even in the US the FDIC has made it clear that it requires regulatory changes if it is to be able to handle the failure of a large bank satisfactorily, whether or not it is subject to the systemic exemption (FDIC, 2006). Given the choice between the financial crisis and support most governments are likely to opt for support. Banks know that and hence this opens the system up to moral hazard. However, with some large cross-border banks, particularly in Europe, it is not clear whether the authorities would be able to keep such a bank open if it is large relative to the economy or if a large number of countries would be involved in deciding what to do in a case of difficulty.

Here we begin by setting out some principles that might govern any effective resolution method before going to consider the problems of the particular methods and finally what can be done in the cross border case. For such a resolution system to work it needs to have the following five characteristics.

Key Characteristics of an Effective Framework for Handling Systemic Banks1. A rapid implementation so that the systemically essential functions can be maintained without a

material break2. A predetermined means of deciding what losses are to be taken into account and how they are to

be assessed3. A means of allocating losses that respects the hierarchy of claimants under insolvency and does

not make the claimants any worse of than they would be under normal insolvency procedures.4. the method should not reduce the risk that shareholders and those responsible for the losses

would otherwise bear, in particular it should not introduce a moral hazard that would encourage institutions to take on increased risk

5. the method needs to be equitable across financial institutions irrelevant of their size or the sequence in which insolvencies occur.

These characteristics presuppose being able to define what the systemic functions are in addition to giving insured deposit holders uninterrupted access to their funds. It clearly involves ensuring that confidence in the system needs to be maintained as well as the actual avoidance of serious fluctuations. This implies that not merely must financial markets continue to operate, along with the payment and settlement systems but that their structures must be such that people expect them to work in the face of shocks. This confidence comes not just from the systems themselves but from credible commitment of the authorities to step in the event of market failure. Of course, it may very well be that the confidence is misplaced and the authorities cannot or are unwilling to deliver if a problem hits.

There is a clear problem in deciding what is ‘systemic’ in advance. In terms of retail banking, this may simply be in terms of market share. In the New Zealand case this was straightforward as there was a clear gap between the smallest of the major banks and the largest of the smaller banks. While a line has to be drawn somewhere there is no practical difference between two banks with a dollar difference in size. For the proposals suggested here to work such banks need to be determined in advance and a sufficiently close relationship built up that they could be taken into statutory management or its equivalent at short notice. The authorities in each country need to make

28 Murton (2005) spells out how the FDIC would propose to act in the event of the failure of a large bank including how a bridge bank would be set up.

15

careful assessments of the nature of their markets to determine which institutions might be systemic in them. As evidenced by the actions of the UK in expanding the number of banks in the primary market it may be possible to solve the problem by removing the systemic position of the bank in advance.

However, there is an important distinction between a problem in a single bank when the system is under no other threat and one where the whole system is very fragile and a problem with one could easily bring down many others. In these circumstances the number of institutions and events that could lead to systemic problems will be greatly increased, thus making prior identification more difficult, as it is context specific. One problem that needs to be addressed with some care is how to be equitable across all the banks during a problem period. Mishkin (2005) suggests allowing the first to fail but intervening to limit the knock on consequences. In the event of a large external shock this could be rather a lottery and encourage masking of problems to avoid being first in the queue. Perhaps more likely the first will be saved as the total potential call on taxpayers is underestimated initially.

2 Too big to failOne of the main concerns of this paper is distinguish circumstances where the concern is wider than purely minimising the loss to the deposit insurer from the concept of ‘too big to fail’. The concept of too big to fail implies that the only technique available to the resolution agency in the event of failure is open bank assistance, namely that the institution has to be kept open and running without a break in business via some form of capital support. The discussion that follows distinguishes keeping the functions of the bank in continuous operation from continuing the ownership and legal personality of the bank. Indeed, the contention is that if the moral hazard is to be avoided and the appropriate pressure put on the bank to resolve its own problems, then the resolution method should involve the owners losing their claim on and control of the bank.29

There is of course a prior concern articulated clearly by Stern and Feldman (2004) that in a large economy like the United States it may be possible to organise the structure of the financial system in such a way that although most of the economies of scale can be reaped no financial institution is actually essential to the operation of the financial system. They also argue that it is possible to reduce the interdependencies in the system, for payments for example. Hüpkes (2004) takes this further by introducing the idea of ‘replaceability’ – a function is systemic if it cannot be replaced from other sources during the value day. If such replaceability is achieved, any bank could be closed in an ordinary manner without systemic consequences. In most countries, however, internationally competitive banks can readily be large compared to their financial systems and hence too large to close without using methods that take into account their possible systemic impact.30

In many respects the problem is largely that it is difficult to act quickly enough with a large and complex institution (Evanoff and Kaufman, 2005). If the actions could be sorted out in time then the sheer size of the loss may not be the difficulty. However, it is probably mistaken to suggest that problems are likely to be proportionate to the size of the institution. Problems that are of the dimension to affect markets will tend to emerge, whether they are large or small relative to the size of the institution. In one sense this is reassuring as it also implies that unexpected shocks are less likely to bring down the largest institutions. Where countries are small compared to the size of international banks (as outlined in Hüpkes (2003) for Switzerland in the case of Credit Suisse and UBS) there is a real danger that the banks will be ‘too big to save’ in the sense that neither the fiscal capacity nor the technical capacity exists to avert a disorderly failure and hence financial instability

29 It is important not to imply that currently large banks are indeed exhibiting the moral hazard. As Mishkin (2005) points out the evidence in the US tends to suggest that there is little indication. See also Wall (1993).30 Stern and Feldman (2006) go further than this and suggest that the authorities should say in advance for each institution that has systemic functions how they propose to handle the insolvency without invoking the idea of ‘too big to fail’. The implication is that either alternative service providers can be lined up so that contracts can be switched or detailed arrangements are in place for how the bank’s systems can be handled that they can be permitted to fail as an entity in the manner described in this article.

16

will ensue.The main difficulty therefore is likely to come from banks that may not necessarily be large from

a global point of view but are large compared to some of the markets they operate in.31 They will indeed be too large compared to those markets to be allowed to stop operating but they may be small enough to manage a very rapid resolution without disrupting the market. It is noticeable that the FDIC cites being able to buy time (Reidhill et al., 2005) as an important point in favour of bridge banks. However, the time which needs to be bought has four components:

– time to decide what is the best course of action– time to establish the condition of all of the parts of the banking group– time to implement the resolution and reorganisation process– time to line up buyers for the bank as a whole or in parts.

However, simple delay, as with forbearance does not address the issue. The swift action needs to be final and to provide the confidence to keep trading in place without a rush to exit.

3 The nature of the interventionIf the authorities want to keep key activities in a bank operating they have a straightforward range of options available:(a).They can provide sufficient support to the bank directly to stop the threat of closure, while

insisting on strong conditions for action and reorganisation to prevent the losses worsening and start the process of recovery. This would probably involve a change in key staff and the appointment of someone to run the bank in whom the authorities had trust. This would be typical open bank assistance.

(b).They can provide sufficient support to enable a competitor to take over either whole bank or at least the functions that the authorities wish to see continue unbroken – purchase and acquisition

(c). They can acquire the bank themselves as receivers, conservators or other court appointed agents and continue to keep either the whole bank or the appropriate functions operating. This in itself can take a number of forms. A typical approach is to split the organisation into a ‘good bank’ where the healthy and required activities continue and a ‘bad bank’ where the losses are concentrated and where these activities will be terminated and liquidated in the manner normal for an insolvency. The question at issue in the present discussion is the legal form for the continuing bank, whether in whole or in part. The term bridge bank relates to a particular form used in the US on a number of occasions by the FDIC (Krimminger, 2006, Marino and Shibut, 2005) but this is a generic type and in what follows the discussion is not restricted to the particular characteristics of the US structure. In the scheme described in Mayes et al. (2001) and implemented in New Zealand, Switzerland and Norway, inter alia, the authorities can step in and reorganise the existing legal institution.

Although the over-riding aim should be some concept of minimising the loss, the principal driving force of the choice of resolution method is going to be the practicality of what can be done in the time available. However, having a purely ad hoc approach will in itself affect expectations and the moral hazard inherent in the system. The nature of the decision on what is going to be applied needs to be clear in advance and something that will be less desirable for the current owners and managers of the bank in difficulty than the normal sorts of private sector solution, whether or not assisted by the ‘good offices’ of the authorities. One of the other noteworthy features of the US system, which has not been universally adopted, is the concept of Structured Early Intervention and Resolution and within that the idea of Prompt Corrective Action, which is designed to try to prevent problems getting as far as insolvency and direct intervention by the authorities. This entails the early involvement of the resolution agency not just when things are starting to deteriorate but in normal times. This enables the authorities to have in place some of the key ingredients of a successful 31 This article does not consider the problem that some small countries face, that their largest banks could take decisions that are fairly minor for the bank yet have systemic consequences. For example, it is argued that the largest banks operating in Estonia (which are Swedish) could write off their Estonian operations completely and still not make a loss for the parent group in the quarter this occurs. They therefore need tools for avoiding systemic actions by banks in more normal operations, far short of insolvency.

17

resolution that can be applied in a hurry. If there has been no preparation it is rather unlikely that anything could be done to keep a large and complex organisation in business other than open bank assistance.

Many countries have powers to act in the way set out under FDICIA in the US but they do not have the all important compulsion to act in a particular way with only limited opportunity to delay at a series of prescribed benchmarks that trigger actions of increasing firmness. Even fewer have opened a dialogue with their banks to ensure that they know enough about their computer systems, structures and other operating procedures to be able to implement a helpful resolution process in a hurry.

It is still typically assumed that the time for action is the proverbial weekend, with the bank appearing with its problems at the close of business on Friday and everything being sorted out before the markets reopen on Monday morning. This is however an increasingly unrealistic view of the world. The key decisions may very well have to be taken within the course of a day. Knowledge of the problem is likely to leak quickly and turn a problem of manageable proportions into one where the authorities have real difficulty avoiding just the systemic event they are trying to prevent. The emphasis in any of these procedures therefore needs to be on having as much as possible in place that can be done without imposing undue costs on the banks or indeed on the authorities as that could in itself result in something that is not a ‘loss minimising’ arrangement. The regulatory costs of the system can impose a loss, if they do not act to reduce moral hazard, the risks and the costs in any failures that do occur adequately. It is very difficult to estimate what these parameters are.

The approach of the Reserve Bank of New Zealand, with its ‘Outsourcing Policy’, seems as good a description as any of what might be done. Here the objective is to try to ensure that whatever operational difficulties a bank may encounter it can be up and operating again before the end of the value day. Insolvency is only one of the things that could act as a trigger, key service providers could fail or a natural disaster could terminate normal operations. In any of these circumstances the authorities need to be convinced that the bank’s vital operations can be restarted promptly without causing harm to the financial system. The outsourcing policy (RBNZ, 2006) is explicit in what the performance should be. The arrangements must be such that: The bank’s clearing and settlement obligations due on a day can be met on that day The bank’s financial risk positions on a day can be identified on that day The bank’s financial risk positions can be monitored and managed on the day following any

failure and on subsequent days The bank’s existing customers can be given access to payments facilities on the day following

any failure and on subsequent days.A statutory manager, the New Zealand term for the official administrator appointed to takeover the running of the institution in the event of failure, must also be able to achieve these objectives. While a bank can only offer simulation evidence that it meets these requirements, it would nevertheless have to satisfy the authorities that its systems were compliant in this regard.

These four provisions cover all of the main aspects that a bank would need to offer to meet the needs of financial stability, namely, that:

– Responsibilities with respect to the rest of the financial system are met in clearing and settlement

– Customers have access to their accounts– It is possible to identify and manage the risks

What it does not cover explicitly, although it was referred to in the draft, is the ability of the statutory manager to be able to make a summary valuation of the net position of the bank before the end of the value day. If a bank is going to reopen it has to be in a position to offer creditors, depositors and counterparties a credible assurance that they will suffer no further losses and will not be advantaged by attempting to exit their positions rapidly. This requires either recapitalisation, a writing down of claims or the issuing of some form of guarantee with state backing.

Kaufman (2006) offers a simple four point programme that has to be met in the case of failure:

18

1. Prompt legal closure when the bank’s capital declines to some prespecified and well publicised minimum value greater than zero (legal closure rule),

2. Prompt estimates of the recovery value and assignment of any credit losses (haircuts) to de jure uninsured bank claimants,

3. Prompt reopening (e.g., the next business day), particularly of larger banks, with full depositor access to their accounts on their due dates at par value for insured deposits and recovery value for uninsured deposits and full borrower access to their pre-established credit lines, and

4. Prompt re-privatisation and re-capitalisation of the bank in whole or in parts at adequate capital levels.

all characterised by the word ‘prompt’. These requirements are similar in character to the scheme set out in Mayes et al. (2001).

The first of his points emphasises that action needs to occur early, preferably before capital is exhausted. It is preferable that ‘failure’ for a bank be insufficient capital and not insolvency or negative value. Pozdena (1991) emphasises the virtues of the Danish system that prevailed in the 1980s; see also Mølgaard (2003). Then, despite having the largest stock of non-performing loans of the Nordic countries, Denmark was the one that avoided the financial crisis that swept Finland, Norway and Sweden. It is arguable that their closure policy and associated features of their regulatory system helped in this by making sure that banks did not become highly undercapitalised. They set a floor of 8% for capital adequacy. If capital fell below 6% then the bank has at most until its next shareholders’ meeting to raise at least 75% of the shortfall. Closure, in the form of withdrawing the banking licence follows almost immediately upon failure. Even above the 8% ratio, any bank with a capital ratio below 15% had to devote a proportion of its profits to reserves.

On top of that the Danish authorities required mark to market accounting, the writing down of non-performing loans and, for stocks and bonds, a provision for volatility. The experience, which may not be the result of the regime, was that almost all Danish banking problems were solved within the market and troubled banks were bought by larger competitors before the difficulties mounted too far. Only two banks were subject to compulsory closure; one (6th July Bank) was immediately sold, while the other (C&G Banken) liquidated. Pozdena (1991) argues that this was all achieved might imposing any due burden on the Bank in question. It is worth noting that deposit insurance was only introduced late in the day as a result of EU directive requiring it in all member states, between the two failures. 6th July Bank was closed in March 1987 and C&G Banken the following October. Pozdena suggests that it was only some unusual features of the loan portfolios that prevented these institutions from finding private sector ‘voluntary’ solutions as well.

However, Pozdena conducted his analysis before the main problems erupted in Denmark. Between 1987 and 1995, 122 financial institutions got into difficulty. 102 ceased to exist. Most of the normal resolution methods, merger, capital injection, split into good and bad banks and liquidation were applied. Neither of the two largest ‘systemic’ banks got into serious difficulties. Nevertheless, early action by the authorities kept the problems manageable and avoided a ‘crisis’.

4 Who should be responsible?Although the deposit insurers in the US and Canada, the FDIC and the CDIC, and in some other countries have the responsibility for minimising losses (to themselves) and hence need to take an active role in ensuring that the risks in the financial system are managed well all of the time – a task that involves them inter alia in the supervision of the institutions holding insured deposits – this is not the norm; see the IADI survey, whose results are discussed in Su (2006). It is more common for deposit insurers to be subservient organisations, often just simple payboxes under the direction of a national supervisor, bankers’ association – as is the case in Finland – or other responsible body such as the central bank. In these cases it is likely that the responsible organisation will face a conflict of interest if it is responsible both for the continuing supervision and health of the banking system as well as for resolving problems. The supervising arm will tend to have an interest in seeing the problem resolved without a failure, as will the central bank if it has emergency lending at stake

19

whose collateral may be of doubtful value in the event of a sudden sell off. Although some of the concern about forbearance is largely theoretical there are some egregious examples of its use as documented by Mishkin (2005) in the case of the US Savings and Loan crisis. This motivation for forbearance will conflict with the pressure for early intervention and acknowledgement of the problem by an organisation responsible for minimising its losses.

It is relatively difficult – but not impossible - for a single organisation to supervise an institution for two purposes at the same time. The resolution agency or arm of the wider organisation is interested in understanding the business so it can step in. It is interested in getting reasonable valuations for the assets and liabilities and being able to manage aspects of the computer system in an emergency. This is very different from ensuring compliance. The area of clear overlap is in the assessment of the risks. Having an agency that is well versed in how to handle insolvency step in as soon as the bank encounters any difficulty is likely to make the incentive to achieve a private sector solution much stronger. Mayes and Liuksila (2003) argue that such a resolution agency, even at a European level, does not necessarily have to be heavily staffed if it can draw rapidly on experts who are normally employed elsewhere, as failures are relatively unusual in most western European countries.

5 The power to actBeing able to create and run a bridge back or otherwise take a failing institution into a version of public ownership in order to keep functions of systemic importance running implies an appropriate basis in law. It could be applied under normal bankruptcy law by giving appropriate powers to receivers or other administrators of the bankruptcy estate but in most countries this is not the case. In general it requires a special bankruptcy law for banks. Even this is not straightforward as the ‘owner’ of the bank may be a holding company. It may therefore be the owner who is insolvent rather than simply one or more banking components of the group. Other parts of the financial group may not be banks and hence subject to different legislation. Hüpkes (2006) stresses that the legislation needs to enable the banking authorities to act at whatever level of a complex group is appropriate in order to resolve the banking problems. Her conclusion is rather negative ‘In view of the limited intervention powers of the lead or co-ordinating supervisor, there is reason to question whether it would be possible to achieve effective early resolution of a crisis affecting globally active institutions with diverse financial activities carried out through separate legal entities.’ (p.25).

To some extent traditions and experience matter in this regard, as both Hadjiemmanuil (2003) and Blowers and Young (2003) argue that the ‘London approach’ of rapid actions through the courts can work when the judiciary well understand the need for speed and accept that errors can be put right or compensated for after the event. The resolution agency needs to be able to act early. Krimminger (2006) reports that Mexico introduced even stronger powers for its deposit insurer Instituto para la Proteccion al Ahorro Bancario (IPAB) in April 2006. Not only can IPAB form a bridge bank but it can intervene earlier with a strong effect on the shareholders. Once the capital ratio falls below 8% it can insist on the development of a capital restoration plan, one of whose conditions is that 75% of the bank’s equity is deposited in a trust for the benefit of IPAB. If the conditions are not met, IPAB can impose losses on the shareholders equivalent to the losses and even gain control of the bank. Once the capital ratio falls to 4% it can implement resolution procedures so that it stands a much better chance of avoiding serious loss. Forming a bridge bank in these circumstances has to offer the prospect of lesser loss than simply closing the bank; such a bridge may only last for a year.

6 The bridge bank approach

A crucial distinction between the ‘bridge bank’ concept and some of the alternatives is that the existing bank is terminated and its banking licence withdrawn, and a new institution is created. Under the Swiss system, for example, it is the existing institution that is reorganised Hüpkes, 2003).

If an appropriate framework is not in place then the authorities will have no alternative to open

20

bank assistance if they want to avoid the systemic consequences, as the FDIC found in the case of the Continental Illinois failure in 1984 (Reidhill et al, 2005). A bridge bank is likely to be the way the FDIC chooses to handle the failure of bank with systemic implications if only because getting grips with the bank and sorting out the nature of the problems, what can and should be done is likely to take weeks rather than the hours that may be available.

Bridge banks have not been used widely in the US and not at all in the last 12 years, in part because the scale and complexity of failures have been not been such as to necessitate it.32 They are deliberately a temporary arrangement to allow all or part of a failing institution to continue in operation, for a period of up to two years, extendable in one year increments to five if it has not yet proved possible to sell it or close it in a manner that minimises the losses. While the FDIC is responsible for creation of the institution and can control it, it does not own it in the sense of holding the common stock or voting shares. The bridge bank is a new national bank, chartered by the Comptroller of the Currency. The ability to create bridge banks was given in the Competitive Equality Banking Act of 1987.

The key feature of the arrangement, which is some respects unique to the US environment is that it provides the authorities with a means of cutting the bank off from its holding company. In that way any further assistance given to the bank does not accrue to the shareholders of the holding company or to claimants on the holding company rather than on those with a claim directly on the banking subsidiary. It is a device that puts the failed bank in a form more suitable for its sale to the private sector, in particular by keeping it operating. It enables the FDIC to benefit from the observation that a bank is usually worth more alive than dead even if it is under water Guttentag and Herring (1986).

Interestingly, the bridge bank form has been used to put together a whole set of subsidiaries into a single banking entity (40 in the case of First Republic in 1988, 19 in the case of MCorp in 1989 and 3 in the case of the Bank of New England). However, in the case of First City in 1992 the FDIC decided that the most effective way to proceed was to create 20 individual bridge banks and then it could sell them individually or as a group. Within the 20 subsidiary banks, 16 were formed from the whole deposit base of the failed institution and four only from the insured deposits. Clearly it may depend upon the circumstances whether all subsidiaries should be treated the same. If there are large outstanding claims before the courts there may be too much risk in acquiring all the liabilities. It is a difficult judgement in advance to decide whether the subsidiary banks will perform better and be more attractive as a group or individually. No doubt this will be affected be the degree of independence of the various subsidiaries prior to failure. The greater their common systems and synergies from being part of the same holding company then the more sense a joint operation is likely to make. Even if in a sense forming separate bridge banks always leaves the option of recreation of the group open.

First City was not only the last bridge bank to be formed but the only one after the passing of FDICIA in 1991. Reidhill et al. (2005) in their assessment of the usefulness of bridge banks, based on the 10 occasions they have been used is that they help to shorten the period of resolution and that they can be consistent with the least cost resolution requirement. Clearly they buy time, which will enable potential purchasers to complete their investigations and enable the FDIC to restructure the banks to make them more profitable, while meeting the requirements of access by depositors to their funds. They imply that the first choice of immediate sale (at a realistic price) was not available and that simply paying out the depositors and disposing of the assets piecemeal over time as ‘ordinary’ insolvencies did not appear a lower cost route.

Handling a large bank failure is going to be difficult whatever system is used both because of the complexity of the organisation and the lack of experience anyone from the outside is going to have in running it or something similar. Thus, while some help will be available from the recently retired in particular, it is inevitable that there will be a heavy reliance on existing staff, some of whom may turn out to be responsible for the problems in the first place. One likely cause of failure is the

32 They have been used just 10 times in the period between 1988 and 1994. However, the case of Superior Bank, a thrift institution, is functionally similar (Eisenbeis, 2006).

21

inability of the previous administration to manage the integration of a previous expensive merger so the systems to be managed may themselves be somewhat in disarray. As Marino and Shibut (2005) point out, the importance of preparedness for the resolution agency is crucial. Techniques can be evaluated on smaller cases and in simulations but the main source of strength will come from a good understanding of the small number of systemically important institutions that might have to be handled.

7 Alternatives to the bridge bank approachIn Mayes et al. (2001) we proposed a scheme that had many of the characteristics of the US scheme, emphasising the importance of early intervention under a form of Prompt Corrective Action with specific predetermined benchmarks. This entailed a close relationship between a resolution agency and the supervised institutions before there was any concern over the condition of the bank otherwise procedures could not be put in place. They differed from the US arrangements in two main respects. The first was that although we thought it desirable to intervene finally before capital was totally eroded our interpretation of EU Law was that it would be very difficult for the authorities to intervene before net worth of the institution in mark to market terms reached zero. Otherwise it would be over-riding the rights of shareholders. In the light of experience since then, including the discussions of the proposal included in Mayes and Liuksila (2003) we have become more convinced that our original idea was correct and that the authorities should seek a change in the law to permit earlier intervention. Simple withdrawal of the banking licence might not meet the need as that would imply a sudden stop to activity and not the smooth resolution we envisaged. The second major difference was in the form of the reorganisation. We did not prescribe the exact method but suggested that the most important issue was to be able to restore the bank to viability in short order so that there would be no break in the business. We saw this entailing four steps:– The appointment of a receiver, conservator or other form of statutory manager to take over the

running of the bank– A summary valuation to establish the extent of the deficiency– Where insolvency was either not the lowest cost method of resolution or where it would lead to

unacceptable financial instability, a writing down of the claims or other restructuring to return the bank to positive net worth in a form that would give confidence to markets

– we argued that this latter would probably entail a guarantee against any further loss that had government backing.

We argued that the writing down of the claims needed to follow the normal sequence in an insolvency, with the shareholders taking the initial loss, then the subordinated debt holders and then the other creditors in order, with each class being treated equitably. The proposal put forward by Aghion et al. (1992) that this take the form of a debt for equity swap with the claimants now having an equity claim on the new institution might be a suitable way to go. If the bank had positive value compensation would have to paid to the existing shareholders, otherwise this would amount to expropriation.

One of the problems with the FDIC’s approach to bridge banks is that involves being able to distinguish clearly between insured and uninsured deposits. The FDIC has no duty to protect uninsured deposits and its proposals suggest they would not be taken into the bridge bank. It is not all clear that unless adequate arrangements are made for such deposits and other claims, so that at least the deposit holders can have immediate access to their written down claims that a loss of access on a large scale might not trigger the systemic problems. Clearly it will help to insist that the deposit insurer can identify the accounts (LaBrosse and Mayes, 2006) and that the banks systems should be transparent and known to the resolution agency in advance as a result of its continuing interaction with the bank in normal circumstances. Banks in the US were not surprisingly rather unenthusiastic about altering their systems to achieve this (Feldman and Stern, 2006) although it is not clear how large the costs might be in practice.

In any case focusing simply on deposits does not identify the systemic features of the bank’s operations. As Hüpkes (2004) points out, these cannot really be ‘carved out’ from the rest of the

22

operation of the bank. They may not be legally or administratively distinct. One could of course require this as a regulator. The area where carve outs do work is in the collateralisation of transactions in many markets. In these cases there cannot normally be losses unless collateral values fall so far that they are insufficient to cover the claims.

The Swiss arrangements, set out in Hüpkes (2003) have the advantage of more flexiblility than the Mayes et al (2001) proposals but they require time to implement the reorganisation and apply any reorganisation haircut necessary to bring the new institution to adequate capitalisation. Mayes et al. suggest that taking the whole organisation as it stands is the only thing that can be done in a hurry, even if subsequently it is clear that a reorganisation that involves sale of some of the more profitable parts might maximise a bankruptcy estate. The Swiss scheme can in many ways be thought of as more akin to the later steps in prompt corrective action, and has the advantage that they can be implemented without the agreement of the existing shareholders, although approval of the reorganisation by the courts allows the stakeholders to express their views before the decision is taken. It is a process with considerable merit if it can be implemented before the point of insolvency.

8 Handling the various parts of the businessHüpkes (2004) helpfully points out that even if one wants to keep a bank open in order to prevent an adverse impact on the financial system this does not imply that all parts of the bank or the wider group need to be kept running if they are a source of actual or potential future losses. Indeed a degree of shrinkage is likely even if the bank is treated as whole. Peripheral activities may well be sold in order to improve the liquidity of the bank if a fair price can be obtained. New lending is likely to be conducted on a conservative basis etc. Typically it is necessary to offer (insured) depositors virtually uninterrupted access to their funds. However, where a bank is a key player in financial markets it needs to behave reasonably normally if markets are not going to be disturbed. It is not simply that the bank needs to avoid breaking a whole set of open contracts if its rating and credibility are to be maintained but it needs to offer new contracts on a scale that provides adequate liquidity to the market. This is particularly important where a bank is a key counterparty. Such contracts entail risk, which has to be managed, hence the skilled market players in the bank have to be retained in their posts.

If the restructuring process for the bank requires a haircut or some other means of writing down claims, then the speed at which that has to be done will tend to vary with the time to maturity of the contracts. It is only the transactions that take place on the value day on which the resolution process occurs that have to be handled immediately. However, certainty is required about what will happen to the remaining contracts otherwise the holders of the claims may try to advance them, causing the systemic difficulties it is hoped to avoid.

9 The cross-border issueWhile the rapid resolution of a large bank may be possible as described here, it is already more complicated when they are part of a larger group, much of which may be involved in insurance or other non-banking activities. Not only may they not be covered by the same laws that permit intervention of the authorities before insolvency but their insolvency might have much more limited implications for financial stability. Where the bank or group runs across regulatory borders the problem becomes much more complex (Lastra, 2003).

If there are only systemic implications in the country of the parent (home) then the complexity makes resolution more difficult but the spin off problems are largely internalised in the home country, even if the problems causing the insolvency have been incurred in foreign operations. However, once there are systemic implications in more than one country or only in host countries then the problem becomes a serious threat to systemic stability as there may be a conflict of interest. The home country may be content see an orderly closure, whereas the host with the systemic problem will want some form of resolution that keeps the business going. It depends on the structure of the group as to whether this can be achieved. If the group is divided neatly into largely

23

free-standing subsidiaries that are locally incorporated and supervised then having a system where each country can step in and apply its own processes may make sense even if this does not necessarily maximise the recovery value for the group as a whole. However, if the activities themselves run across borders and the banks have branches in other countries that are not free-standing and not locally supervised then this neatness does not apply.

It is this last case that is increasingly applying in the EU,33 where operation through branches in other member states is actively encouraged in the pursuit of economic integration and the completion of the single market. Operating directly or through branches does not require local approval in the host countries, although local conduct of business rules apply, and supervision (and deposit insurance) remains the responsibility of the home country. Furthermore, under the Winding Up Directive, the parent and branches are supposed to be treated in single proceedings under the insolvency regime of the home country. The host country then does not have control over what will affect its financial stability where the branches are of systemic importance, described as ‘responsibility without power’ in Mayes (2006b). It cannot institute PCA, it cannot intervene before insolvency nor finally afterwards. It has to hope that the home country will do that job for it. That might work if the bank were systemic in the home country as well but this is not guaranteed. Mayes et al. (2006) offer a way of implementing PCA in such cross-border instances. The home country may choose a resolution approach that does not foster stability in the host. Simply the anticipation that there is going to be a difficult wrangle about ‘who pays’ in the event of a failure would tend to trigger an exit of the smart money from the bank. The current position is clearly unacceptable (Srejber, 2006).

It is highly tempting therefore to follow the approaches applied by the US and New Zealand and cut through the international complexity so that a bank with systemic implications can be treated like a domestic institution whatever the nationality of its ownership (Baxter et al, 2004). Thus New Zealand has decided that all banks of systemic importance have to be locally incorporated. Not only that but the New Zealand operation has to be viable in its own right to the extent that a statutory manager can take over something that can be run independently of the parent within the value day as described above.

Since that option is not open in the EU a more collaborative approach is required as set out in Mayes (2006a), for example, where a college of supervisors under the leadership of the home country jointly supervises the financial group and implements SEIR and PCA. Nevertheless there is still a problem in applying a means of resolving an insolvency or unacceptably low capitalisation. Losses have to be assigned, guarantees given and even loans advanced. This burden could be too great for the home country and might require joint action by the component countries of the college. This cannot be done at the time. Srejber (2006) argues for a fixed prenegotiated key for burden sharing, however it seems difficult to see foreign governments being willing to finance something which is largely a concentrated problem even if there would be a theoretical quid pro quo in a future crisis.

A likely solution therefore seems to be a supranational organisation to handle cross-border banks that are systemically important in at least one member state. This could be done on a case by case basis as the number of such banks is small (between 30 and 50) and the range of countries involved decidedly variable. The alternative is to have a European resolution agency – it could be a European Deposit Insurance Corporation (EDIC) to match the framework of the FDIC, CDIC etc. but it does not need to handle the vast bulk of banks in Europe that are largely national in character and not of systemic importance. Hence it can be a much smaller organisation. However, this does not solve the problem for such banks that have substantial operations outside the EU. Here the territoriality approach seems likely to be applied in systemic cases and a New Zealand style approach where rapid intervention along the lines suggested in this paper is possible seems a sensible way to go. The particular form of the intervention, whether as a bridge bank or another form of reorganisation haircut, is less important than simply ensuring that the systemic functions can be continued without

33 The term EU embraces the wider European Economic Area, that adds Iceland, Liechtenstein and Norway to the 25 EU member states, as they are all governed by EU law in this regard.

24

a break. Nevertheless, some form of acknowledging the loss and removing the bank from the existing owners seems likely to be both lower cost and more equitable. However, with the current state of preparedness in many countries it is not so likely to be the route chosen. Either some form of open bank assistance or the disorder that results in the systemic problem seems more likely.

References

Aghion, P, Hart, O. and Moore, J (1992) ‘The Economics of Bankruptcy Reform’, Journal of Law, Economics and Organization, vol.8, pp.523-46.

Alesina, A. and Guido Tabellini (2004), Bureaucrats or Politicians?, CEPR Discussion Paper Series, WP n. 5242.

Alesina, A. and Guido Tabellini (2005), Why Do Politicians Delegate?, NBER Working Paper Series, WP n. 11531.

Allen, W A and Wood, G (2006) ‘Defining and Achieving Financial Stability’, Journal of Financial Stability, forthcoming.

Basel Committee on Banking Supervision (1992) Minimum Standards for the Supervision of International Banking Groups and their Cross-border Establishments', Basel, July.

Baxter, TC, Hansen, J and Sommer J H (2004) ‘Two Cheers for Territoriality: An Essay on International Bank Insolvency Law’, American Bankruptcy Law Journal, vol. 78(1), pp.57-91.

Benink, H. and George J. Benston (2005) The future of banking regulation in developed countries: Lessons from and for Europe. Mimeo.

Benston, George J. and George G. Kaufman (1998) Risk and Solvency Regulation of Depositor Institutions: Past Policies and Current Options. New York: Salomon Brothers Center, Graduate School of Business, New York University.

Benston, G., and Kaufman, G. (1994) The Intellectual History of the Federal Deposit Insurance Corporation Improvement Act of 1991. In G. G. Kaufman (ed.) Reforming Financial Institutions and Markets in the United States, p 1-17. Boston: Kluwer.

Bjerre – Nielsen, H. (2004) The Financial Regulatory and Supervisory Authority: A Principal and an Agent, The Danish Financial Services Authority (mimeo).

Blowers, B and Young, G (2003) ‘The Economic Impact of Insolvency Law’, pp.164-179 in Mayes and Liuksila (2003).

Diamond, D and Dybvig, P (1986) ‘Bank Runs, Liquidity and Deposit Insurance’, Journal of Political Economy, vol.91(3), pp.401-19.

Eisenbeis, R (2006) ‘Home Country versus Cross-Border Negative Externalities in Large Banking Organization Failures and How to Avoid Them’, chapter X in this book.

Eisenbeis, R. A. and G. G. Kaufman (2006) Cross border banking: Challenges for deposit insurance and financial stability in the European Union (mimeo).

European Shadow Financial Regulatory Committee Statement No. 1, Dealing with problem banks in Europe. Center for Economic Policy Studies, 22 June, 1998.

European Shadow Financial Regulatory Committee (ESFRC, 2005) Reforming Banking Supervision in Europe. Statement No. 23 Frankfurt am Main, 21 November 2005.

Evanoff, D and Kaufman, G (2005) Systemic Financial Crises: Resolving Large Bank Insolvencies, Singapore: World Scientific.

Evanoff, Douglas D. and Larry D. Wall, (2002) Subordinated Debt and Prompt Corrective Regulatory Action in Prompt Corrective Action in Banking: 10 Years Later edited by George Kaufman, pp. 53-119. This book is volume 14 of Research in Financial Services: Private and Public Policy published by JAI.

FDIC (2006) ‘Large-Bank Deposit Insurance Modernization Proposal’.Garcia, G. and María J. Nieto (2005) Banking Crisis Management in the European Union: Multiple

Regulators and Resolution Authorities Journal of Banking Regulation Vol. 6 N 3 pp.215-219.Goodhart, C. A. E. and Dirk Schoenmaker (2006) Burden Sharing in banking crisis in Europe

LSE Financial Markets Group. Special Paper Series No 163 September 2005.

25

Guttentag, J and Herring, R (1986) ‘Disaster Myopia in International Banking’, Essays in International Finance, no.46, Princeton University, September.

Hadjiemmanuil, C (2003) ‘Bank Resolution Policy and the Organisation of Bank Insolvency Proceedings: Critical Dilemmas’, pp. 272-330 in Mayes and Liuksila (2003).

Hadjiemmanuil, Ch. (2004) Europe’s Universalist Approach to Cross-Border Bank Resolution Issues, paper presented at the Federal Reserve Bank of Chicago Conference on Systemic Financial Crises: Resolving Large Bank Insolvencies, September 30-October 1.

Harrison, I (2005) ‘The Reserve Bank of New Zealand’s Creditor Recapitalization (BCR) Project’, in (D. Evanoff and G. Kaufman, eds., Systemic Financial Crises: Resolving Large Bank Insolvencies, pp.397-406, Singapore: World Scientific.

Hüpkes, E (2003). ‘Learning Lessons and Implementing a New Approach to Banking Insolvency in Switzerland’, pp.242-71 in Mayes and Liuksila (2003).

Hüpkes, E (2004). ‘Protect Functions not Institutions’, The Financial Regulator, vol.9(3), pp.43-49.Hüpkes, E (2006) ‘Piercing the Corporate Veil: Managing Financial Failure in an Evolving

Economic and Financial Environment’, Western Economics Association International Conference, San Diego, 29 June-2 July.

Kane, E.J., R, L. Bennett and R. Oshinsky (2006) “Evidence of Improved Monitoring and Insolvency Resolution under FDICIA.” (mimeo)

Kaufman, G (2006) ‘Using Efficient Bank Insolvency Resolution to Solve the Deposit Insurance Problem’, ch. 8 in A Campbell, R LaBrosse, D Mayes and D Singh, eds., Deposit Insurance, Basingstoke: Palgrave-Macmillan, forthcoming.

Khan, Ch. M. and Santos, J.A.C. (2002) Allocating Lending of Last Resort and Supervision in the Euro Area in (V. Alexander, J. Melitz and G.M. von Furstenberg eds.) Monetary Union: Why, How, and What Follows?, chapter 19, London, Oxford University Press.

Khan, Ch. M. and Santos, J.A.C. (2005) Allocating bank regulatory powers: Lender of last resort, deposit insurance and supervision  European Economic Review, Volume 49, Issue 8, November 2005, Pages 2107-2136.

Krimminger, M (2006) ‘Controlling Moral Hazard in Bank resolutions: Comparative Policies and Considerations in System Design’, mimeo, FDIC, May.

LaBrosse, J R and Mayes, D G (2006) ‘Promoting Financial Stability through Effective Depositor Protection: The Case for Explicit Limited Deposit Insurance’, ch 1 in A Campbell, R LaBrosse, D Mayes and D Singh, eds., Deposit Insurance, Basingstoke: Palgrave-Macmillan, forthcoming.

Lastra, R (2003) ‘Cross-border Bank Insolvency: Legal Implications in the Case of Banks Operating in Different Jurisdictions in Latin America’, Journal of International Economic Law, pp.79-110.

Llewellyn, D and Mayes, D G (2004) 'The role of market discipline in handling problem banks', pp. 183-210 in G. G. Kaufman ed., Market Discipline in Banking: Theory and Evidence, Amsterdam: Elsevier

Marino, J and Shibut, L (2005) ‘Resolution Strategies for Large US Commercial Banks’, mimeo, FDIC, November.

Mayes, D G (2004) 'Getting the Best out of Basel II for the European Banking System', pp.64-70 in The New Basel Capital Accord and the Future of the European Financial System, CEPS Task Force Report no 51, Brussels: CEPS.

Mayes, D G (2006a) ‘Cross-border Financial Supervision in Europe: Goals and Transition Paths’, Sveriges Riksbank Economic Review, 2006/2, pp.58-89.

Mayes, D G (2006b) ‘Responsibility Without Power: Two Solutions to the Problem of Foreign Owned Systemic Bank Branches’, Journal of Banking Regulation forthcoming.

Mayes, D G, Halme, L and Liuksila, A (2001) Improving Banking Supervision, Basingstoke: Palgrave.

Mayes, D G and Liuksila, A (2003) Who Pays For Bank Insolvency?, Basingstoke: Palgrave-Macmillan.

Mayes, D G, Nieto, M and Wall, L (2006) ‘Multiple Safety Net Regulators and Agency problems in

26

the EU: Is Prompt Corrective Action Partly the Solution?’, Bank of Spain.Mishkin, F S (2005) ‘How Big a Problem is Too Big to Fail?’, NBER working paper 11814,

December.Moe, T G (2004) ‘Norway’s Banking Crisis – How Oslo Got it Right’, The Financial Regulator,

vol.9(3), pp.26-33.Moe, T, Solheim, J and Vale, B (2005). The Norwegian Banking Crisis, Oslo: Bank of Norway

Occasional Paper 33.Mølgaard, E (2003). ’Avoiding a Crisis: Lessons from the Danish Experience’, in Mayes and

Liuksila (2003), pp.222-241.Murton, A (2005) ‘Resolving a Large Bank: The FDIC’s Perspective’, in D Evanoff and G

Kaufman, eds., Systemic Financial Crises: Resolving Large Bank Insolvencies, pp.415-20, Singapore: World Scientific.

Nieto, Maria J.and Larry D.Wall (2006) Preconditions for a successful implementation of supervisors´ Prompt Corrective Action: Is there a case for a banking standard in the EU? LSE Financial Markets Group Special Paper Series No 165 March.

Pozdena, R J (1991) ‘Danish Banking: Lessons for Deposit Insurance Reform’, Journal of Financial Services Research, vol. 5, pp.289-98.

Reidhill, J, Davison, L and Williams E (2005) ‘The History of Bridge Banks in the United States’, Washington DC: FDIC.

Repullo, R (2004) “Liquidity, Risk-Taking, and the Lender of Last Resort”, International Journal of Central Banking, 1 (2005), pp. 47-80.

Reserve Bank of New Zealand (2006) ‘Outsourcing Policy’, Document BS11, January, Wellington.Schmidt Bies, S., (2004), "Enterprise Perspectives in Financial Institution Supervision", University

of Connnecticut School of Law, October 21, http://www.federalreserve.gov/boarddocs/speeches/2004/200410212/default.htm.

Schnasi, G (2005). Safeguarding Financial Stability: Theory and Practice, Washington DC: International Monetary Fund.

Sironi, A. (2001) Testing for Market Discipline in the European Banking Industry: Evidence from Subordinated Debt. The Financial Safety Net: Costs, Benefits, and Implications for Regulation. The 37th Annual Conference on Bank Structure and Competition. Federal Reserve Bank of Chicago, pp. 366 - 384.

Srejber, E (2006) ‘Are We Ready to Deal With A Cross-border Banking Crisis in Europe?’, Gdansk, 12 May, available from http://www.riksbank.se

Stern, G and Feldman, R (2004) Too Big To Fail: The Hazards of Bank Bailouts’, Washington DC: Brookings Institution.

Stern, G and Feldman, R (2006) ‘Managing Too Big to Fail by Reducing Systemic Risk: Some Recent Developments’, The Region, June

Su, W (2006) ‘General Guidance for the Resolution of Bank Failures’, ch 10 in A Campbell, R LaBrosse, D Mayes and D Singh, eds., Deposit Insurance, Basingstoke: Palgrave-Macmillan, forthcoming.

Vesala, J., (2005), Prudential Supervision and Deposit Insurance Issues Raised by the European Company Statute, Colloquium on the European Company Statute, ECB, Frankfurt, 23 February.

Wall, L (1993) ‘Too Big to Fail After FDICIA’, Federal Reserve Bank of Atlanta Economic Review, January/February, pp.1-14.

27