March 18, 2010
Toward an Effective Resolution Regime for Large Financial Institutions
Governor Daniel K. Tarullo
At the Symposium on Building the Financial System of the 21st Century, Armonk, New York
Two years ago this week, Bear Stearns succumbed to severe liquidity stress. It was rescued, and eventually absorbed by JPMorgan Chase, with financing assistance provided by the Federal Reserve. Although it would take another six months before the accumulating stress and uncertainty posed an immediate threat to nearly all of our major financial institutions, it is clear in retrospect that this arranged marriage, and its accompanying dowry of government financing, set off an expansion of the universe of firms perceived as too big to fail.
During the financial crisis, government authorities in the United States and elsewhere believed they had only two realistic options in the face of serious distress at a large financial firm. First, they could try to contain systemic risk by stabilizing the firm through capital injections, extraordinary liquidity assistance, a subsidized acquisition by a less vulnerable firm, or some combination of these supports. Second, they could allow the firm to fail and enter generally applicable bankruptcy processes, risking in those times of fear and uncertainty a run on similarly situated firms.
The Bear Stearns deal was an example of the first policy option. Lehman Brothers was an example of the second. When its bankruptcy set off a firestorm in the exceedingly dry tinder of financial markets in the fall of 2008, the U.S. government decided that further failures of large, interconnected financial institutions risked bringing down the entire financial system. It responded to the situation with the Troubled Asset Relief Program to provide capital, and the Temporary Liquidity Guarantee Program to extend debt guarantees, to large financial firms.
Indeed, faced with the possibility of a cascading financial crisis, most governments around the world selected the bailout option in most cases. But if the costs of this approach are less dramatic during a crisis, they are no less significant afterward. Entrenching too-big-to-fail status obviously risks imposing significant costs on the taxpayer. It undermines market discipline, competitive equality among financial institutions of different sizes, and normal regulatory and supervisory expectations.
The desirability of a third alternative to the Hobson's choice of bailout or disorderly bankruptcy is obvious--hence the prominence during the regulatory reform debate of proposals for a special resolution process that would allow the government to wind down a systemically important firm in an orderly way while still imposing losses on shareholders and creditors. The crisis has also focused attention on the special problems created by the failure of a large, internationally active financial firm. In my remarks I will elaborate on the relationship between resolution regimes and an effective overall system of financial regulation and supervision, both in the international and domestic spheres. 1
At the risk of some oversimplification, I would state that relationship as follows: First, an effective domestic resolution process is a necessary complement to supervision that would bring more market discipline into the decisionmaking of large financial firms, their counterparties, and investors. At the same time, even a well-designed resolution mechanism is no substitute for reformed regulatory rules and strengthened supervisory oversight.
Second, the high legal and political hurdles to harmonized cross-border resolution processes suggest that, for the foreseeable future, the effectiveness of those processes will largely depend on supervisory requirements and cooperation undertaken before distress appears on the horizon. I would further suggest that the importance of proposed requirements that each large financial firm produce a so-called living will is that this device could better tie the supervisory and resolution processes together.
A Resolution Regime for Large, Interconnected Firms
As compelling as the case for such a process is, the debate around resolution proposals has highlighted the challenge of crafting a workable resolution regime for large, interconnected firms. The basic design problem is that such a regime must advance the goals of both financial stability and market discipline. While these goals are usually complementary, they can at times be competing--especially in periods of high financial stress, when time consistency problems can loom large. In the midst of a crisis, governments fearful of financial upheaval can be tempted to provide assistance to supposedly uninsured creditors, even at the cost of increasing moral hazard in the post-crisis period. Despite the consequent design difficulties, I think there are certain essential features of any special resolution process.
First, any new regime should be used only in those rare circumstances where a firm's failure would have serious adverse effects on financial stability. That is, the presumption should be that generally applicable bankruptcy law applies to nonbank financial firms--even large, interconnected ones. One way to help ensure that the regime is invoked only when necessary to protect the public's interest in systemic stability is to use a "multi-key" approach--that is, one that requires the approval of multiple agencies and a determination by each that the high standards governing the use of the special regime have been met.
Second, once invoked, the government should have broad authority to wind down the company in an orderly way. This authority should include--among other things--selling assets, liabilities, or business units of the firm; transferring the systemically significant or viable operations of the firm to a new bridge entity that can continue these operations; and repudiating burdensome contracts of the firm, subject to appropriate conditions and compensation.
Third, there should be a clear expectation that the shareholders and creditors of the failing firm will bear losses to the fullest extent consistent with preserving financial stability. Shareholders of the firm ultimately are responsible for the organization's management (or, more likely, mismanagement) and are supposed to be in a first-loss position upon failure of the firm. Shareholders, therefore, should pay the price for the firm's failure and should not benefit from a government-managed resolution process.
To promote market discipline on the part of the creditors of large, interconnected firms, unsecured creditors of the firms must also bear losses. Here is where the potential conflict of policy goals is obvious. While losses imposed on creditors will increase market discipline in the longer term, the immediate effect could be to provoke a run on other firms with broadly similar positions or business strategies. Thus the extent of these losses and the manner in which they are applied may need to depend on the facts of the individual case. At the very least, however, subordinated debt, or other financial interests that can qualify as regulatory capital, should be fully exposed to losses.
Fourth, the ultimate cost of any government assistance provided in the course of the resolution process to prevent severe disruptions to the financial system should be borne by the firm or the financial services industry, not by taxpayers. The scope of financial institutions assessed for these purposes should be appropriately broad, reflecting that a wide range of financial institutions likely would benefit, directly or indirectly, from actions that avoid or mitigate threats to financial stability. However, because the largest and most interconnected firms likely would benefit the most, it seems appropriate that these firms should bear a proportionally larger share of any costs that cannot be recouped from the failing firm itself. To avoid pro-cyclical effects, such assessments should be collected over time.
Establishing a resolution regime with these characteristics is, I would suggest, one of the most important financial regulatory reforms for every country that does not already have such a mechanism in place. It would lend substance to the idea that market discipline can be a solid third pillar of financial regulation, along with stronger prudential requirements and improved supervisory oversight. Still, as is implicit in the foregoing discussion, an untested regime will probably not acquire complete credibility until it is actually applied successfully. For this reason, among others, it is important to ensure that other regulatory tools will help compensate for the uncertainties associated with an essentially untested mechanism.
International Efforts on Resolution Issues
The looming or actual failure of a large, internationally active financial firm inevitably complicates the already challenging process of resolution. Mismatches in the amounts and maturities of assets and liabilities held by the firm in the various countries in which it operates can lead host governments to take special action to protect the interests of depositors and creditors. And different insolvency regimes apply to separately incorporated subsidiaries across the world. Some of those regimes may be substantively inconsistent with one another, or may not account for the special characteristics of a large international firm.
A natural response, which one can find peppered through various law journals over the years, is to propose an international treaty that would establish and harmonize appropriate insolvency regimes throughout the world. Just to state the proposition is to see the enormous hurdles to its realization. The task of harmonizing divergent legal regimes, and reconciling the principles underlying many of these regimes, would be challenge enough. But an effective international regime would also likely require agreement on how to share the losses and possible special assistance associated with a global firm's insolvency.
Despite the good and thorough work being undertaken in both the Basel Committee on Banking Supervision (Basel Committee) and the Financial Stability Board, we must acknowledge that satisfyingly clean and comprehensive solutions to the international difficulties occasioned by such insolvencies are not within sight. 2 It would certainly be useful if jurisdictions could at least broadly synchronize both standard bankruptcy and any special resolution procedures applicable to a failing financial firm. But even this significant advance would not settle many of the nettlesome problems raised by a cross-border insolvency.
It thus seems reasonably clear that effective management of these problems will, at least for the foreseeable future, require regulatory coordination and supervisory cooperation before a large firm's failure becomes a real possibility. In one sense, this observation reinforces the importance of the international agenda for strengthening capital and liquidity standards. It also counsels continued attention to efforts to ensure that globally active institutions are subject to effective consolidated supervision, and that information-sharing arrangements among home and host country supervisors are well designed and implemented. To this end, the key supervisors and central banks for each of the largest global banks will begin to meet regularly to discuss crisis planning, with particular attention to contingency liquidity planning.
The crisis demonstrated that issues around cross-border liquidity support are difficult. Liquidity pressures may arise in unexpected places, time for coordination will be short, and failures in one jurisdiction likely will spread quickly to other jurisdictions. The Basel Committee and the Committee of European Banking Supervisors are each working on definitions of liquid assets, common stress-testing metrics, and structural balance sheet measures. We are actively discussing the appropriate division of responsibility between home and host authorities to provide liquidity support and the related issue of how to approach cross-border branch operations. Some have called into question the traditional assumption that home country authorities will be willing and able to support all of the worldwide operations of a banking group headquartered in its jurisdiction. It is not clear what approach might work better, but there is an obvious need for broad international consistency and careful calibration with other prudential requirements.
One of the key issues identified by the Basel Committee's Cross-Border Bank Resolution Group is the complexity and interconnectedness of the largest organizations. Often the complexity is motivated by tax or regulatory factors, rather than a clear business purpose. Given the way these firms are structured and their linkages to key systems and other institutions, resolution of such an organization will carry significant risk of spillovers to other key markets, payments systems, or systemically important institutions. The Cross-Border Bank Resolution Group consequently recommended developing initiatives that would result in simpler, less interconnected organizational structures.
This point leads us to one much-discussed idea, that of firm-specific resolution plans--sometimes referred to more colorfully, though not wholly accurately, as living wills. This proposal provides a good opportunity to advance the aim of linking resolution mechanisms to other regulatory tools, both domestically and internationally.
In one variant of the idea, each internationally active bank would be required to develop, and potentially to execute, its own resolution plan--literally, to plan for its own demise. Such a requirement could doubtless be helpful to some degree, but it has notable limitations.
Most obviously, it is very difficult to predict in advance of a crisis which parts of the firm will be under greatest stress, what geographical regions may be affected most severely, and what the condition in various markets and economies will be, as well as the stability of counterparties and similarly situated institutions. Furthermore, governments may be understandably reluctant to rely too much upon a wind-down plan developed by an internationally active financial firm that so mismanaged itself that it is on the brink of failure, placing other institutions at peril. Finally, management of an institution can be expected to seek to preserve as much value for shareholders as possible in its planning, whereas the supervisors' objective in a crisis is to achieve an orderly resolution, which will often entail winding down or restructuring the insolvent firm in ways that effectively wipe out shareholder interests.
The living will requirement could be broadened so as to make it into a potentially very useful supervisory tool for healthy firms, as well as a resource in the event that resolution became necessary. Under this approach, the firm would, in addition to developing a resolution plan, be required to draw up a contingency plan to rescue itself short of failure, identify obstacles to an orderly resolution, and show it can quickly produce the information needed for the supervisor to orchestrate an orderly resolution should the need arise. These plans will need to evolve as the organization's business and economic conditions evolve, and accordingly, the plans will need to become a regular part of normal supervisory processes.
A living will of this type could remove some of the uncertainty around a possible resolution. It would force firms and their supervisors to review contingency plans regularly. As part of their ongoing oversight, supervisors could target the areas where a firm's planning falls short of best practices. Focusing on the legal, contractual, and business relationships among the firm's subsidiaries could yield significant benefits for prudential supervision in normal, as well as stressed, times. The various elements of the regulatory system could thus be better integrated by identifying mechanisms and connections for the transmission of risk and liability between affiliates and by identifying relationships that may present an obstacle to the ready sales of businesses, the proceeds from which might allow the firm to avoid failure.
Central to the success of a living will as a supervisory tool is the quality of information it would make available in a crisis. Some of the information would be relatively static. A firm would have to inventory all of its legal entities, along with the legal regimes applicable to each one, and map its business lines into legal entities. A firm also would have to document interaffiliate guarantees, funding, hedging, and provision of information technology and other key services. This information would be needed to deal with any crisis, no matter what its specific form.
Once the centrality of accurate, comprehensive information is understood, it becomes apparent that a very significant upgrade of management information systems (MIS) may be the only way for the firm to satisfy living will requirements, just as we at the Federal Reserve found when we led the Supervisory Capital Assessment Program--popularly known as the bank stress tests--that improved MIS are needed for ongoing risk management at the institution.
Supervisory demands for improved MIS could have another benefit. Just as a homeowner has an incentive to shed belongings to reduce the expense of moving, so a financial firm may have a powerful incentive to simplify its organizational structure and rationalize relationships among its corporate entities to reduce the cost of developing comprehensive MIS that enable an organization to retrieve information in multiple formats across jurisdictions, business lines, and legal entities. Simpler structures can also be encouraged by reemphasizing existing supervisory guidance requiring banking organizations to measure and manage their risks not only on the global, consolidated level, but also on a legal entity basis. Together, the information requirements of living wills and the need to measure and manage risks at the legal entity level can help create the right incentives for firms to simplify their structures without necessarily requiring a supervisor to delve into the details of a banking group's structure.
All of this work on resolution, both domestic and international, is important and necessary. But we must be realistic about what it can accomplish. In light of what has happened over the past two years, it is imperative that governments convince markets that they can and will put large financial firms into a resolution process rather than bail out creditors and shareholders. Yet no one can guarantee that future resolutions of systemically important firms will proceed smoothly or predictably. Resolution mechanisms must be understood not as silver bullets, but as critical pieces of a broader agenda directed at systemic risk and the too-big-to-fail problem.
2. See Basel Committee on Banking Supervision, Cross-Border Bank Resolution Group (2009), Report and Recommendations of the Cross-Border Bank Resolution Group (Basel, Switzerland: Basel Committee, September); and Financial Stability Forum (2009), FSF Principles for Cross-Border Cooperation on Crisis Management (Basel, Switzerland: FSF, April). (The Financial Stability Forum subsequently was renamed the Financial Stability Board.) The Basel Committee's Cross-border Resolution Group released its report and recommendations today, and the FSB will present a final report and recommendations to the G-20 in October. Return to text