William C Dudley: Ending too big to fail

Remarks by Mr William C Dudley, President and Chief Executive Officer of the Federal Reserve Bank of New York, at the Global Economic Policy Forum, New York City, 7 November 2013.

The views expressed in this speech are those of the speaker and not the view of the BIS.

It is a great pleasure to have the opportunity to speak here today. My remarks are going to focus on what is called the "too big to fail" problem. As you are aware, this problem arises when the failure of a large, complex financial institution threatens to cause such significant disruption to the financial system and the economy that these potential costs are judged as too severe to bear, leading to government intervention to prevent the failure. As a result, the firm, by being too big to fail, gains an implicit guarantee at the taxpayers' expense that it does not have to pay for. I think there is broad agreement that such a regime is unacceptable in several respects. The first problem is that it creates an uneven playing field between large and small financial firms, with larger financial firms gaining a funding advantage from the perception that they may be too big to fail. The second problem is that this funding advantage creates incentives for financial firms to become bigger and more complex. The third problem is that there is a positive feedback loop. As the banking system becomes more concentrated and complex, that just increases the financial stability risks, making the too big to fail problem even more acute.

So what should we do about it? Today, I will evaluate three broad sets of choices: 1) Building a credible resolution regime and more resiliency in the financial system that together reduce the systemic costs of failure sufficiently so that large, complex firms can be allowed to fail; 2) taking steps, such as tougher prudential standards, that further reduce the probability of failure of such firms; and 3) breaking up the too big to fail firms so that no firm is so large that its failure would threaten financial stability in the first place.

To summarize, I conclude that building a credible resolution regime is necessary but not sufficient. Even if the single-point of entry resolution framework proposed by the Federal Deposit Insurance Corporation (FDIC), which I very much endorse, is fully perfected, the costs of resolution for the largest systemically important financial institutions (SIFIs) will still be significant. I will argue that at least as much effort should be made to lower the risk of failure of such large, complex firms. Not only does this include higher capital and liquidity requirements, which we are implementing, but also building incentives into the system so that firm managements will act more forcefully and much earlier to put their firms on more solid ground before they encounter greater difficulties.

Finally, I am not yet convinced that breaking up large, complex firms is the right approach. In particular, these firms presumably exist, in large part, because there are scale or network effects that allow these firms to offer certain types of services that have value to their global clients. These benefits might be lost or diminished if such firms were broken up. In addition, the costs incurred in breaking up such firms need to be considered. Finally, the breakup of such firms would not necessarily result in a significant reduction in overall systemic risk if the resulting component firms were still, collectively, systemic.

As always, what I have to say reflects my own views and not necessarily those of the Federal Reserve System.

Defining the too big to fail problem

The root cause of the too big to fail problem is the fact that the failure of a large, complex financial firm is likely to generate significant, undesirable externalities. These include disruption to the ability of the financial system to provide credit and other essential financial services to households and businesses. When this happens, not only is the financial sector adversely affected, but its troubles spill over and harm the real economy.

Although there are negative externalities associated with the failure of any financial firm, these externalities are disproportionately more damaging in the case of large, complex and interconnected firms. However, despite the label, too big to fail, the magnitude of these externalities does not depend simply on size. The extent of the externalities also depends on the particular mix of business activities and the degree of interconnectedness with the rest of the financial industry. One significant element is the importance of the services the firm provides to the broader financial system and the economy, as well as the ease with which its customers can move their business to other providers. Another element is the extent to which the firm's structure and activities create the potential for contagion - that is, the potential that a problem at one firm could spread more broadly through the financial system. Contagion might occur along several different pathways, including the direct losses imposed by failure on the firm's counterparties, the impact of the failure on the asset prices held by other leveraged financial institutions, or a loss of confidence spreading to other firms with similar business models.

The presence of large negative externalities creates a dilemma for policymakers when such firms are in danger of failing, particularly at a time when the wider financial system is also under stress. At that point in time, the expected costs to society of failure are very large compared to the short-run costs from providing the extraordinary liquidity support, capital, or other emergency assistance necessary to prevent catastrophic failure.

The market's belief that a too big to fail firm is more likely to be rescued in the event of distress than other firms weakens the degree of market discipline exerted by capital providers and counterparties. Since the government does not charge for this implicit guarantee, this reduces the firm's cost of funds and incents the firm to take more risk than would be the case if there were no prospect of rescue and funding costs were higher. The fact that firms deemed by the market to be too big to fail enjoy an artificial subsidy in the form of lower funding costs distorts competition to the detriment of smaller, less complex firms. This advantage, in turn, creates an unfortunate incentive for firms to get even larger and more complex. As a result, the funding benefit of being seen to be too big to fail causes the financial system to become skewed toward larger and more complex firms in ways that are unrelated to true economies of scale and scope.

Thus, the too big to fail problem consists of two intertwined issues. The first is that the negative consequences to the financial system and to the economy from failure for a given set of firms are unacceptably high. This is the financial stability risk. The second is that anticipated interventions to prevent catastrophic failure create an uneven playing field. Not only is this outcome unacceptable from an equity or fairness perspective, but it is also undesirable because it can increase the incentives for firms to become even bigger and more complex. Over time, this may lead to a greater number of systemically important firms and expose the financial system to greater systemic risk.

This suggests that we need to maintain our focus on two goals: (1) Making the financial system more stable by reducing the degree of disruption when failures occur - that is, shrinking the size of the externalities - and by lowering the risk of failure in the first place, and (2) eliminating the artificial advantages that large, complex firms might have that create incentives for them to become bigger and more complex.

One more point deserves emphasis. Only eliminating the competitive advantage of large, complex firms from the too big to fail subsidy will not necessarily make the financial system more stable. For example, consider a regime in which interventions to prevent the failure of large, systemically important firms were impossible. In that case, the funding advantage that comes from being perceived as too big to fail would be eliminated because such firms could fail putting the firm's equity and debt holders at risk. But, this would not be a good outcome from a financial stability perspective because the failure of such firms could still be catastrophic for the financial system and the economy. Thus, we need to solve the "fairness/subsidy" issue, but do so in a way that ensures that the resulting regime is more stable, resilient and robust.

The fairness/subsidy issue

There is considerable debate about the size of the funding advantage of large banks that is due to too big to fail and how much this funding advantage creates incentives to become larger and more complex. One reason is that it is difficult to measure the size of these effects. For example, consider funding costs. We can document that larger banks' unsecured debt funding costs tend to be lower than those for smaller banks. But the funding costs for larger firms across a broad range of industries also tend to be lower than for smaller firms in the same industries. This reflects, in part, the fact that the debt issues of larger firms tend to be more sizeable and more liquid. In addition, larger firms may be more diversified on average and thus less prone to failure, everything else equal.

The measurement challenge, then, is to identify what portion of the lower funding costs at the largest financial firms, after accounting for these other factors, is due to a perception among investors that such firms are too big to fail.

Despite these measurement difficulties, most of the evidence is consistent with a too big to fail funding advantage in banking. First, during the financial crisis, we observed that the funding cost advantage of the larger banks grew substantially relative to the smaller banks as the crisis became more acute. This is noteworthy because if the funding cost advantage is due to a too big to fail premium, then that premium should rise when the economic environment worsens and the potential risk of failure increases. Of course, the Dodd-Frank Act has changed the landscape.

Second, ongoing research by Federal Reserve Bank of New York staff shows that the funding advantage of large versus small banks is higher than the funding advantage for large versus small non-bank financial firms and non-financial firms when other factors are held constant.

Third, the major rating agencies add an uplift to their credit ratings for the largest banks due to the prospect of government support. While it is possible that the rating agencies are wrong in their assessment, what matters is perception. If investors share this view or simply follow the ratings, then this should create a too big to fail funding advantage.

With respect to the second consequence of the subsidy - the incentive for management to make their firms larger and more complex so as to be perceived as too big to fail and thus able to take advantage of any funding subsidy - this is harder to evaluate. On one hand, this is not the only incentive management may have to get bigger. For example, the fact that chief executive officer compensation and prestige are related to size may be just as important. On the other, this still is a problem because it works to distort the structure of the financial system and makes - by encouraging the proliferation of large, complex firms - the too big to fail problem worse.

I conclude that while the funding advantage issue is very relevant to the debate on too big to fail, it may not be the most important issue. To me, the unacceptable risk posed to financial stability from the failure of large, systemic firms is at least as important.

Solving too big to fail

As I see it, there are two broad ways of solving the too big to fail problem. The first is to create a more robust financial system so that the failure of a large, systemically important financial firm does not threaten to take down the rest of the financial system. In this case, the authorities can let the troubled firm fail. Because failure is now credible, this should also eliminate the funding advantage from being perceived as too big to fail.

The second is to take steps to prevent the failure of such large, complex firms in the first place. This also reduces any funding advantage for firms perceived as being too big to fail, but without some of the negative financial stability consequences associated with failure. I view these two strands as complements rather than substitutes, with more work necessary on both.

Turning first to the issue of reducing the consequences of failure, there are numerous elements of this effort underway. Some are focused on making the financial market infrastructure more robust so that when a failure occurs the shock is dampened, not amplified and propagated throughout the financial system. For example, considerable effort has been made to create incentives for firms to standardize over-the-counter derivative trades and to clear those trades through central counterparties. Similarly, extensive work has been undertaken by the Federal Reserve and the major clearing banks to make the tri-party repo system more stable. Money market mutual fund reform is also part of this ongoing effort.

Other initiatives are focused on reducing the financial stability consequences from the failure of a systemically important financial firm. The major initiative here is the single point of entry framework for resolution proposed by the Federal Deposit Insurance Corporation. Under this framework, if a financial firm is to be resolved under Title II of the Dodd-Frank Act, the FDIC will place the top tier bank holding company into receivership and its assets will be transferred to a bridge holding company. The equity holders will be wiped out and sufficient long-term unsecured debt will be converted into equity in the new bridge company to cover any remaining losses and to ensure that the new entity is well capitalized and deemed creditworthy. Subsidiaries would continue to operate, which should limit the incentives for customers to run. By assigning losses to shareholders and unsecured creditors of the holding company and transferring sound operating subsidiaries to a new solvent entity, such a "top-down" resolution strategy should ensure continuity with respect to any critical services performed by the firm's subsidiaries and this should help limit the magnitude of any negative externalities.

I very much endorse the FDIC's single point of entry framework for resolution. I think it is the best plan for implementing Title II given the complexity and scope of large, global financial institutions, and I also think it is well suited to the U.S. bank holding company framework. For this regime to work properly there needs to be a sufficient amount of debt outstanding at the parent company that can be converted by the FDIC into equity to ensure that the new bridge company will be demonstrably well-capitalized. We don't yet have a long-term debt requirement - this is an area where we are still working out the details. My own view is that a holding company needs a substantial amount of long-term debt to ensure that the newly created bridge company will be considered fully viable by its counterparties.

While the Title II single-point-of-entry strategy holds tremendous promise, there are important implementation issues that still need to be worked out. Implementing the resolution regime on a cross-border basis remains one of the most significant challenges. In this regard, I am worried about two current shortcomings:

● The reach of Title II's one-day stay on the close out of qualified financial contracts is incomplete and does not extend to contracts governed by non-U.S. law with non-U.S. counterparties.

● The placement of the parent company into receivership may be treated by these counterparties as an event of default due to the presence of a parent guarantee or other cross-default provisions triggered by the parent-level insolvency.

Unless market participants make the appropriate contractual changes that will ensure that the placement of the parent company into Title II will not trigger the close-out provisions of over-the-counter derivatives and other qualified financial contracts that are outside the reach of Title II's U.S. application, foreign counterparties to a SIFI will tend to exercise this right whenever it is in their individual economic interest to do so. This would create significant difficulties. Such actions could greatly complicate the operations of the firm during a time when it is already under considerable stress and could propagate stress more broadly in financial markets.

There are two main courses of action for addressing this concern, and they are not mutually exclusive. First, existing derivative contracts need to be amended and future contracts need to provide that the parent's entry into the Title II proceeding does not trigger the close out option. Second, legal changes need to be implemented abroad so that the one-day stay that applies to qualified financial contracts governed by U.S. law is enforceable against those contracts governed by foreign law. Only by making these changes can we avoid the potential for disruptive close outs. I strongly encourage the ongoing efforts to address this critical issue.

A second issue with respect to Title II resolution on a cross-border basis is that we cannot be certain how foreign authorities will react when the parent holding company is put into the Title II proceeding. While the U.S. authorities have been in discussion with our colleagues abroad to enable the coordination needed for a smooth cross-border resolution process, uncertainties remain regarding the circumstances under which host authorities may either choose to take or be required to take actions such as unilateral "ring-fencing" that might disrupt the implementation of the single point of entry approach. Thus, we need to continue to work with foreign regulators to iron out any issues ahead of time to remove these uncertainties so that the resolution regime will work well for global, systemically important firms.

Another issue with respect to the Title II resolution regime is the residual uncertainty about whether a particular firm in fact would go through a Title II resolution. Recall that under the Dodd-Frank Act, Title II is not the default approach for dealing with the failure of a large complex firm. For a firm to go through Title II, a determination must first be made that the failure of the firm and its resolution would have serious adverse effects on U.S. financial stability under the insolvency law that would otherwise apply. This means that market participants will not know for certain which path U.S. authorities will ultimately take until this determination is made - allowing resolution under ordinary insolvency law, as contemplated under Title I, versus initiating a Title II resolution. Faced with this uncertainty, investors in the short-term obligations that would likely be protected only under the Title II regime could decide to run.

Reducing the likelihood of failure

Even with an appropriate resolution process in place that ends too big to fail, the consequences of failure will still be messy. Also, market discipline is only effective to the extent that those investors with their funds at risk perform the appropriate due diligence on an ongoing basis. This suggests that a resolution regime by itself is not sufficient. Further efforts should also be made to reduce the probability of a default.

A number of steps have already been taken along these lines that further reduce default risk. The Basel III framework significantly raises both the quantity and quality of capital required of internationally active bank holding companies. This risk-weighted capital standard is also being augmented by a higher leverage ratio requirement. The Basel framework will also require SIFIs to hold additional capital due to their complexity (more commonly referred to as the SIFI surcharge). As a result, the capital buffer for the more systemic firms should be higher based on size, complexity, interconnectedness, global exposure and substitutability, so that their expected probability of failure will be lower than for less systemic firms.¹

The SIFI surcharge acts as a penalty for size and complexity, leaning against any remaining funding cost advantage that might remain for a large complex firm even after a viable resolution regime is in place. This should further level the playing field for smaller firms and reduce the incentive to become systemically important in the first place. However, I expect that it will take considerable time for the surcharge to have a meaningful impact on size and complexity.

We also need to create new mechanisms and incentives for bank management to act early, well before resolution becomes necessary. Early intervention is likely to be much more successful in preventing failure as compared to last-ditch efforts. These actions can take many forms - cutting capital distributions earlier, raising new capital faster, restructuring businesses sooner, and reorganizing senior management and boards of directors more radically when the firm is not performing well.

In this regard, the Comprehensive Capital Analysis and Review (CCAR) should bolster the incentives for early action. The CCAR is a significant supervisory innovation intended to ensure that large, complex bank holding companies have robust internal capital planning processes and sufficient capital to withstand stressed economic conditions and to continue to be viable financial intermediaries. Importantly, the CCAR takes a forward-looking perspective in assessing a bank's capital adequacy, including ensuring that its internal stress testing processes address its unique vulnerabilities and exposures. The CCAR should push banks to proactively manage their capital so as to be well prepared for the type of market stresses that could cause significant declines in income and capital.

Other steps could be taken to complement the CCAR. One approach would be to implement a long-term debt requirement in a way that further enhances market discipline and provides incentives for regulators to promptly intervene if necessary. Another might be to structure compensation practices to strengthen senior bank managers' incentives to proactively manage risk. For example, imagine how incentives would change if a significant portion of senior bank management's compensation each year were deferred to be available to cover future capital losses. I suspect that over time this would meaningfully alter management's risk tolerance. Also, in addition to the quantity of the deferred compensation, the form it takes could also affect incentives. If most of the deferred compensation were in the form of debt rather than equity, I suspect this would also affect management's risk tolerance and the appetite to cut dividend payments, reduce share repurchases or raise more capital more promptly when the firm began to become stressed. The terms and the form of deferred compensation are an important tool that could be further developed to generate a better set of incentives consistent with our financial stability objectives.

We must also ensure that firms can respond to financial or operational weakness of any type by having robust recovery plans that get triggered long before resolution becomes necessary. They should develop and track metrics that not only identify when they are on the verge of insolvency or default, but that trigger appropriate remediation actions whenever the firm begins to become stressed. These plans should identify actionable options that the firm can take in response to financial weakness that will restore the confidence of the firm's counterparties without the need for extraordinary official sector support. They must also have disciplined processes that analyze the root causes of their problems and identify longer-term strategies that will need to be employed as other recovery options to restore capital and liquidity are being executed. And firms must support the viability of their contingency and recovery plans by implementing the internal governance necessary to develop, test, update and implement them credibly.

Some argue that what I have proposed - higher capital requirements and better incentives that reduce the probability of failure combined with a resolution regime that makes the prospect of failure fully credible - are insufficient. Perhaps, this is correct. After all, collectively these enhancements to our current regime may not solve another important problem evident within some large financial institutions - the apparent lack of respect for law, regulation and the public trust. There is evidence of deep-seated cultural and ethical failures at many large financial institutions. Whether this is due to size and complexity, bad incentives or some other issues is difficult to judge, but it is another critical problem that needs to be addressed. Tough enforcement and high penalties will certainly help focus management's attention on this issue. But I am also hopeful that ending too big to fail and shifting the emphasis to longer-term sustainability will encourage the needed cultural shift necessary to restore public trust in the industry.

One alternative strategy for dealing with too big to fail is to reduce the size of large banks by imposing size limits. While I understand the motivation, it is important to recognize that imposing size limitations might not be a very efficient means of making the financial system more stable. First, it could sacrifice some socially beneficial economies of scale and scope, especially to the global clients of large financial institutions, depending on how the size limits were implemented. Second, the resulting benefits in terms of reducing systemic risk might be quite small. As demonstrated by the thrift crisis, if many firms are all vulnerable to the same shock, ensuring that these firms are small and numerous won't necessarily reduce the systemic risk they collectively pose to the financial system. Third, the costs incurred in breaking up such firms also need to be considered.

Others have argued in favor of separating commercial and investment banking activities of the large banks. In my opinion, there are shortcomings to re-imposing Glass-Steagall-type activity restrictions. It is not obvious to me that the pairing of securities and banking businesses was an important causal element behind the financial crisis. In fact, independent investment banks were much more vulnerable during 2008 than the universal banking firms which conducted both banking and securities activities. More important is to address the well-known sources of instability in wholesale funding markets and to give careful consideration to whether there should be a more robust lender of last resort regime for securities activities.

Evaluating the socially optimal size, scope and organizational structure of financial firms is a complicated exercise, and so is establishing a viable transition path to a system of much smaller banks. It would be helpful in this regard if advocates of the break-up solutions would develop detailed proposals that address the essential questions of how such downsizing or functional separation would be accomplished, and what the expected benefits and costs could be from such a restructuring.

The recent financial crisis underscores the importance of financial stability as a necessary condition for a vibrant economy. An important element in improving financial stability is to reduce the expected loss associated with the failure of any financial firm to the point where default does not imperil the financial system and the real economy. A holistic approach is needed that both provides a credible resolution process for large, complex and interconnected banks, uses enhanced prudential standards and initiatives to further reduce the probability of default and the social losses associated with a default, and that incents management to intervene early to address incipient problems before they threaten the viability of the firm. Relying solely on resolution is not sufficient. Until the Title II resolution process is used, there will remain uncertainties regarding how well this approach will work in practice - especially in a time of market stress. For this reason it is also important to continue to pursue a number of alternative approaches.

Thank you for your kind attention. I would be happy to answer a few questions.

---