Skip to main content

Better Late Than Never: Addressing Too-Big-To-Fail

June 1, 2009


Gary H. Stern Former President (1985 - 2009)
Better Late Than Never: Addressing Too-Big-To-Fail

Editor’s note: This column is based on remarks presented at the Brookings Institution, Washington, D.C., March 31, 2009.

Destiny did not require society to bear the cost of the current financial crisis. To at least some extent, the outcome reflects decisions, implicit or explicit, to ignore warnings of the large and growing too-big-to-fail problem and a failure to prepare for and address potential spillovers. While I am, as usual, speaking only for myself, there is now I think broad agreement that policymakers vastly underestimated the scale and scope of too-big-to-fail and that addressing it should be among our highest priorities.

From a personal point of view, this recent consensus is both gratifying and disturbing. Gratifying because many initially dismissed our book,1 published five years ago by Brookings, as exaggerating the TBTF problem and underestimating the value of FDICIA in strengthening bank supervision and regulation. In turn, I would point out that we identified:

  • virtually all key facets of the growing TBTF problem, including the role that increased concentration and increased organizational and product complexity, as well as increased reliance on short-term funding, played in creating the current TBTF mess; and
  • important reforms which, if taken seriously, could have reduced the risk-taking that produced the crisis.

But belated recognition of the severity of too-big-to-fail is also disturbing because it implies that inaction raised the costs of the current financial crisis, as our analyses and prescriptions went unheeded. Despite our warnings, important institutions, public and private alike, were unprepared. And I am quite concerned that policymakers may double-down on previous decisions; some ideas presented in the current environment to address TBTF are unlikely to be effective and, if pursued, will waste valuable time and resources.

In the balance of these remarks, I will principally cover three subjects: (1) the nature of the current TBTF problem; (2) policies essential to addressing the problem effectively; (3) policies that, although well intentioned, are unlikely to make a material difference to TBTF at the end of the day.

The current TBTF problem

As matters stand today, the risk-taking of large, complex financial institutions is not constrained effectively by supervision and regulation or by the marketplace. If this situation goes uncorrected, the result will almost surely be inefficient marshaling and allocation of financial resources, serious episodes of financial instability and lower standards of living than otherwise. Certainly, we should seek to improve and strengthen supervision and regulation where we can, but supervision and regulation is not a credible check on the risk-taking of these firms. I will go into this issue in more detail later and will simply note at this point that the recent track record in this area fails to inspire confidence.

Similarly, market discipline is not now a credible check on the risk-taking of these firms; indeed, a critical plank of current policy is to assure creditors of TBTF institutions that they will not bear losses. Given the magnitude of the crisis, I have supported the steps taken to stabilize the financial system by extending the safety net, but I am also acutely sensitive to the moral-hazard costs of these steps and have no illusion that losses experienced by equity holders and management will somehow resurrect market discipline.

How did we arrive at such a bleak point in terms of TBTF? Let me make just two observations. First, the crisis was made worse, in my view significantly worse, by the lack of preparation I mentioned above. To provide some examples, policymakers did not create and/or execute (1) an effective communication strategy regarding government intentions for uninsured creditors of firms perceived as TBTF; (2) a program to systematically identify the interconnections between these large firms; and (3) systems aimed at reducing the losses that these large firms could impose on other firms. I raise these examples, not surprisingly, because we identified these steps as critical to addressing TBTF in the book and related analysis.2

Second, addressing the TBTF problem earlier could have avoided some of the risk-taking underlying the current crisis. To be sure, many small institutions have failed as a result of the crisis in housing finance but, nevertheless, the bulk of the losses seem concentrated in the largest financial institutions. And creditors of these large firms likely expected material support, thereby facilitating excessive risk-taking by such institutions. Policymakers should correct problems at credit-rating agencies with off-balance-sheet financing, mortgage disclosures and the like. But if, fundamentally, TBTF induces too much risk-taking, then these firms will continue to find routes to engage in it, other things equal.

Addressing sources of spillovers

I have spoken and written about TBTF concerns and policy proposals with sufficient frequency that some observers characterize my views on the topic as “boilerplate,” a backhanded compliment I presume. Nonetheless, it suggests I only judiciously review the key points of the reforms we have long endorsed. The logic for our approach is clear.

  • In order to reduce expectations of bailouts and reestablish market discipline, policymakers must convince uninsured creditors that they will bear losses when their financial institution gets into trouble.
  • A credible commitment to impose losses must be built on reforms directly reducing the incentives that lead policymakers to bail out, that is, provide significant protection for uninsured creditors.
  • The dominant motivation for bailouts is to prevent the problems in a bank or market from threatening other banks, the financial sector and overall economic performance. That is, policymakers intervene because of concerns about the magnitude and consequences of spillovers.

Thus, the key to addressing TBTF is to reduce the potential size and scope of the spillovers, so that policymakers can be confident that intervention is unnecessary. What specifically should policymakers do to achieve this outcome? To answer this question we have taken reforms proposed in the book and combined them in a program we call systemic focused supervision (SFS), which we have discussed in detail elsewhere. In general, SFS, unlike conventional bank supervision and regulation, focuses on reduction of spillovers; it consists of three pillars: early identification, enhanced prompt corrective action (PCA) and stability-related communication.

Early identification. As we have described in detail elsewhere, early identification is a process to identify and to respond, where appropriate, to the material direct and indirect exposures among large financial institutions and between those institutions and capital markets. We anticipate valuable progress simply by having central banks and other relevant supervisory agencies focus resources on, and take seriously, the results of failure simulation exercises, for example. Indeed, such exercises appear to have identified the precise type of issues—around derivative contracts, resolution regimes and overseas operations—that have plagued policymakers’ ability to adequately address specific TBTF cases.3

In fact, it appears that the policy failure was not primarily in identification of potential spillovers, but rather in making corrective action a sufficiently high priority. One constructive option related to early identification would require the relevant TBTF firms to prepare documentation of their ability to enter the functional equivalent of “prepackaged bankruptcy.”4 The appropriate regulatory agencies should require TBTF firms to identify current limitations of the resolution regime they face and the spillovers that might occur if their major counterparties entered such proceedings.

Without doubt, implementing early identification will prove challenging. That said, recommendations from other knowledgeable observers suggest that the task is possible and worthwhile. The G-30 recommendations, for example, would have firms continuously monitor and report on the full range of their counterparty exposures, in addition to reviewing their vulnerability to a host of potential risks, many related to spillovers.5 These reports are precisely the key supervisory inputs to early identification.

One might reasonably wonder about a plan that seems to give center stage to supervisors, when I earlier noted reservations about supervision and regulation. I would point out, however, that here we are emphasizing a role for supervision where it in fact has a comparative advantage. In particular, we would focus supervision on collection of private information on financial institutions, looking across institutions, and worrying about fallout that potentially affects the public, rather than asking supervisors to try to tune risk-taking to its optimal level. Other entities have neither the incentive nor the access to carry out the role we envision for supervision.

Enhanced prompt corrective action. PCA works by requiring supervisors to take specified actions against a bank as its capital falls below specified triggers. One of its principal virtues is that it relies upon rules rather than supervisory discretion. Closing banks while they still have positive capital, or at most a small loss, can reduce spillovers in a fairly direct way. If a bank’s failure does not impose large losses, by definition it cannot directly threaten the viability of other depository institutions that have exposure to it. Thus, a PCA regime offers an important tool to manage systemic risk. However, the regime currently uses triggers that do not adequately account for future losses and give too much discretion to bank management. We would augment the triggers with more forward-looking data, outside the control of bank management, to address these concerns.

Communication. The first two pillars of SFS seek to increase market discipline by reducing the motivation policymakers have for protecting creditors. But creditors will not know about efforts to limit spillovers, and therefore will not change their expectations of support and in turn, their pricing and exposures, absent explicit communication by policymakers about these efforts. This recommendation highlights a key distinction between our approach and that advocated by others: Our approach does not simply seek to limit systemic risk, but takes the next step of directly trying to address TBTF by putting creditors at risk of loss. If we do not do this, we will not limit TBTF.

Now let me turn to some alternative reforms that have received significant attention recently.

Reducing the size of (TBTF) financial institutions

This proposal is straightforward: If financial institutions raise systemic concerns because of their size, make them smaller. We intend to discuss this suggestion at some length in a separate document, but suffice it to say that we have serious reservations about the ultimate effectiveness of such an approach. And I would note, in passing, that it is an idea born of desperation since it seems to admit that large, complex organizations cannot be supervised effectively. (See “Addressing TBTF by Shrinking Institutions”)

To provide a flavor for our concerns about this proposal, consider the government’s ability to keep the firms “small” after dismantling has occurred. There might, for example, be tremendous pressure in the direction of expansion if, in the future, the smooth resolution of the failure of a major institution required the sale of assets to other significant institutions. Even if this situation can be avoided, these firms could still engage in behavior that increases the risk of significant spillovers. They could do so, for example, by shifting their portfolios to assets that suffer catastrophic losses only when economic conditions deteriorate dramatically, thus making themselves and the financial system vulnerable to cyclical outcomes.

Reliance on supervision and regulation and/or FDICIA

The two broad approaches discussed to this point seek both increased market and supervisory discipline to better constrain the risk-taking of large financial institutions. But some observers do not believe that policymakers can credibly put creditors of these firms at risk of loss. And some analysts do not believe that creditors can effectively discipline these oft-sprawling firms even if they had an incentive to do so. As a result, some proposals to better limit the risk-taking of firms perceived TBTF focus primarily on strengthening conventional supervisory and regulatory discipline.

Policymakers could pursue this approach in many ways. After identifying TBTF firms, a more rigorous supervisory and regulatory regime would be applied to them. The tougher approach might include, for example, (a) higher capital requirements, (b) requirements that the firms maintain higher levels of liquid assets, (c) additional restrictions on the activities in which the firms engage, and (d) a much larger presence of on-site supervisors monitoring compliance with these dictates.

My concerns about this approach, and they are considerable, center on the heavy reliance on supervision and regulation but are not a wholesale rejection of S/R per se. Given the distortion to incentives caused by the explicit safety net underpinning banking, society cannot rely exclusively on market forces to provide the appropriate level of discipline to banks. We must have a system of supervision and regulation to compensate. And naturally we should learn from recent events to improve that system, a process under way.6

But we must recognize the important limitations of supervision and regulation and establish objectives that it can achieve. The owners of systemically important financial institutions provide incentives for firm management to take on risk, which is the source of the returns to equity holders (risk and return go hand in hand). Under a tougher S/R regime, these firms have no less incentive than formerly to find ways of assuming risk that generates the returns required by markets and that does not violate the letter of the restrictions they face. By way of example, research on bank capital regimes finds ambiguous results regarding their ultimate effect, as firms can offset increased capital by taking on more risk.

And, as I noted earlier, the track record of S/R does not suggest it prevents risk-taking that seems excessive ex post. True, long shots occasionally come in, and perhaps a regime dependent on conventional S/R would succeed, but it is NCAA tournament time, and we know that a 15 seed rarely beats a number two. To pick just one example from the current episode, supervisors have been unable once again to prevent excessive lending to commercial real estate ventures, a well-known, high-risk, high-return business which contributed importantly to the serious banking problems of the late 1980s and early 1990s.

I recognize that creating a new regulatory framework for a small number of very large institutions differs from supervising thousands of small banks. But I forecast the same disappointing outcome for two reasons. First, we have already applied a version of the suggested approach; right now, we have higher standards and more intensive supervision for the largest banking firms. Second, the failure of supervision and regulation reflects inherent limitations. Supervisors operate in a democracy and must follow due process before taking action against firms. This means that there is an inevitable lag between identification of a problem and its ultimate correction. As previously noted, management has ample incentive to find ways around supervisory restrictions. Further, the time inconsistency problem frequently makes supervisory forbearance look attractive.

A truly draconian regulatory regime could conceivably succeed in diminishing risk-taking, but only at excessive cost to credit availability and economic performance. As Ken Rogoff, a distinguished economist at Harvard who has considerable public policy experience as well, put it: “If we rebuild a very statist and inefficient financial sector—as I fear we will—it’s hard to imagine that growth won’t suffer for years.” (See interview in the December 2008 Region.)

Just as we should not rely exclusively, or excessively, on S/R, I do not think that imposing an FDICIA-type resolution regime on systemically important nonbank financial institutions will correct as much of the TBTF problem as some observers anticipate. To be sure, society will be better off if policymakers create a resolution framework more tailored to large financial institutions, in particular one that allows operating the firms outside of a commercial bankruptcy regime once they have been deemed insolvent. This regime would take the central bank out of rescuing and, as far as the public is concerned, “running” firms like AIG. That is a substantial benefit. And this regime does make it easier to impose losses on uninsured creditors if policymakers desire that outcome.

But I am skeptical that this regime will actually lead to greater imposition of losses on these creditors in practice. Indeed, we wrote our book precisely because we did not think that FDICIA put creditors at banks viewed as TBTF at sufficient risk of loss. We thought that when push came to shove, policymakers would invoke the systemic risk exception and support creditors well beyond what a least-cost test would dictate. We thought this outcome would occur because policymakers view such support as an effective way to limit spillovers. I don’t think a new resolution regime will eliminate those spillovers (or at least not the preponderance of them), and so I expect that a new regime will not, by itself, put an end to the support we have seen over the last 20 months.


I recognize the limits of any proposal to address the TBTF problem. We will never avoid entirely the financial crises that lead to extraordinary government support. But that is a weak excuse for not taking the steps to prepare to make that outcome as remote as we can. It is with deep regret for damage done to residents of the Red River Valley that I note the return of flood season to the Upper Midwest. Many residents have noted that the “100-year flood” has come many more times to this part of the country than its designation implies. And these residents have rightly focused on preparing to limit the literal spillovers when this extraordinary event becomes routine. In contrast, policymakers did not prepare for the TBTF flood; indeed, they situated themselves in the flood plain, ignored the flood warning and hoped for the best. We must now finally give highest priority to preparation and take the actions required before the next deluge.



1 See Gary H. Stern and Ron J. Feldman, 2004, Too Big to Fail: The Hazards of Bank Bailouts, Washington, D.C.: Brookings Institution.

2 See Gary H. Stern, 2008, “Too Big to Fail: The Way Forward,” Nov. 13, 2008, at

3 “Remarks by Sheila C. Bair, Chairman, Federal Deposit Insurance Corporation at the Exchequers Club; Washington, D.C.,” March 21, 2007, and June 18, 2008. Available at

4 For a similar suggestion, see page 62 of Markus Brunnermeier, Andrew Crockett, Charles Goodhart, Avinash D. Persaud and Hyun Shin, 2009, “The Fundamental Principles of Financial Regulation,”
Geneva Reports on the World Economy.

5 See Group of Thirty, 2009, “Financial Reform:
A Framework for Financial Stability,” p. 41.

6 For a discussion of improvement efforts under way for both the banking industry and bank supervisors, see Roger T. Cole, 2009, Risk Management in the Banking Industry, before the Subcommittee on Securities, Insurance, and Investment, Committee on Banking, Housing, and Urban Affairs, U.S. Senate, Washington, D.C., March 18, 2009.