Skip to main content

Addressing TBTF by Shrinking Institutions: An Initial Assessment

June 1, 2009


Gary H. Stern Former President (1985 - 2009)
Ron J. Feldman Senior Vice President
Addressing TBTF by Shrinking Institutions: An Initial Assessment

“If financial institutions raise systemic concerns because of their size, fix the TBTF problem by making the firms smaller.” A number of prominent observers have adopted this general logic and policy recommendation.1 While we’re sympathetic to the intent of this proposal, we have serious reservations about its likely effectiveness and associated costs. Our preferred approach to addressing the too-big-to-fail problem continues to be better management of financial spillovers.2

In this essay, we review our concerns about this “make-them-smaller” reform. We also recommend several interim steps to address TBTF that share some similarities with the make-them-smaller approach but do not have the same failings. Specifically, we support (1) imposing special deposit insurance assessments for TBTF banks to allow for spillover-related costs, (2) retaining the national deposit cap on bank mergers and (3) modifying the merger review process for large banks to provide better focus on reduction of systemic risk. If our suggested reforms prove less effective than we believe, policymakers will have to take the make-them-smaller approach seriously.

The reform

While its proponents have not provided details, this reform—if taken literally—seems straightforward. Policymakers would demark some firms as TBTF through the use of a specific measure, such as share of a given market(s), asset size or revenue. Policymakers would then force those firms to (1) shrink their balance sheets organically (that is, not replacing loans or securities after repayment), (2) divest certain operations or assets and/or (3) split them into smaller constituent parts such that the resulting firms fall below a specified threshold. (We distinguish such measures from short-term efforts to wind down the operations of a targeted, insolvent financial institution to position it for resolution, a reform we support.)

Rationale for reform

On its surface, the proposal has two attractive features, both related to simplicity. First, size seems to offer an easily measured and verifiable means of identifying financial institutions whose financial or operational failure would raise systemic concern. After all, firms that are frequently identified as posing TBTF concerns are large in some important, obvious way.

Second, implementing this reform appears to be fairly straightforward. The government could simply order across-the-board shrinkage of balance sheets for certain firms. Since many larger financial institutions came about through mergers of smaller institutions, and because the popularity among corporate leaders of creating and then destroying conglomerates tends to wax and wane, a simple “unbundling” would merely return the financial world to a period when the TBTF problem did not loom as large.

A third rationale for the reform appears rooted in desperation. Recent events suggest profound failure in the supervision and regulation of large and complex financial institutions. Likewise, a number of observers have long seen the TBTF problem as intractable because policymakers will always face compelling incentives to support creditors at the time systemically important firms get into trouble. Society therefore appears to have no way to impose meaningful restraint on large or complex financial institutions. An option that makes firms neither large nor complex may appear to offer the only real means of imposing either market or supervisory discipline.

The reform’s weaknesses

Shrinking firms so they don’t pose systemic concern faces static and dynamic challenges that seem to seriously limit its effectiveness as a potential reform.

The static challenge involves the initial metric used to identify firms that need to be made smaller. Given the severity of the punishment (that is, breakup), policymakers will have to use a simple standard they can make public and defend from legal challenge. They might consider using, for example, the current limit on bank size that can be achieved via merger: 10 percent of nationwide deposits. Importantly, we assume (and again, because of the high-stakes nature of the reform) that policymakers would make only a few firms subject to forced contraction. This “high bar” raises the stakes in getting the “right” firms cut down to size.

But such a metric will not likely capture some or perhaps many firms that pose systemic risk. Some firms that pose systemic risk are very large as measured by asset size, but others—Northern Rock and Bear Stearns, for example—are not. Other small firms that perform critical payment processing pose significant systemic risk, but would not be identified with a simple size metric. We believe that a government or public agent with substantial private information could identify firms likely to impose systemic risk, but only by looking across many metrics and making judgment calls. Policymakers cannot easily capture such underlying analytics in a simple metric used to break up the firms.

The dynamic challenge concerns both the ability of government to keep firms below the size threshold over time and the future decisions of firms that could increase the systemic risk they pose.

On the first point, we anticipate that policymakers would face tremendous pressure to allow firms to grow large again after their initial breakup. The pressure might come because of the limited ability to resolve relatively large financial institution failures without selling their assets to other relatively large financial firms and thereby enlarging the latter. We would also anticipate firms’ stakeholders, who could gain from bailouts due to TBTF status, putting substantial pressure on government toward reconstitution. These stakeholders will likely point to the economic benefits of larger size, and those arguments have some heft. Academic research has typically found economies of scale exhausted before banks reach the size of the largest banking organizations, although some recent analysis suggests such economies may exist at these large sizes as well.3 (Indeed, policymakers will have to consider the loss of scale benefits when they determine the net benefits of breaking up firms in the first place.)

Prominent examples suggest our concern about reconsolidation is not theoretical. Consider the breakup of the original AT&T and the subsequent mergers among telecommunication firms. Scholars have also highlighted the historical difficulty in limiting the long-run market share of powerful financial firms, including those found in the “zaibatsus” of Japan.4

Even if policymakers could get the initial list of firms right and were able to keep the post-breakup firms small, this reform does nothing to prevent firms from engaging in behavior in the future that increases potential for spillovers and systemic risk. Newly shrunken firms could, for example, shift their portfolios to assets that suffer catastrophic losses when economic conditions fall off dramatically. As a result, creditors (including other financial firms) of the “small” firms could suffer significant enough losses to raise questions about their own solvency precisely when policymakers are worried about the state of the economy. Moreover, funding markets might question the solvency of other financial firms as a result of such an implosion. Such spillovers prompted after-the-fact protection of financial institution creditors in the current crisis, and we believe they would do so again, all else equal. One might call on supervision and regulation to address such high-risk bets. But the rationale for the make-them-smaller reform seems dubious in the first place if such oversight were thought to work.

These dynamics of firm risk-taking mean that the make-them-smaller reform offers protection with a Maginot line flavor. That is, it appears sensible and effective—even impregnable—but in fact it provides only a false sense of security that may lull policymakers into inaction on other fronts. In our experience, policymakers would likely view this reform as a substitute for other desirable actions, including some of the key reforms we think necessary to address spillovers. In the past, policymakers have thought—mistakenly—that the strong condition of banks, the FDICIA resolution regime or initiatives around new capital rules all provided rationales for not addressing the underlying sources of spillovers and the TBTF problem. If we exclusively embrace a reform that misleadingly promises victory over TBTF by constraining the size of large financial firms, we may squander the time and resources needed to address the problem at its roots.

Interim steps

While we would not move forward with a plan to make large financial firms smaller, we take seriously its intent to put uninsured creditors at risk of loss and to address concerns over size, spillovers and government support. In that vein, we recommend three interim steps that address concerns that might lead to support for the make-them-smaller option. They are (1) modify the FDIC insurance premium to better allow for spillover-related charges, (2) maintain the current national deposit cap on bank mergers and (3) modify the merger review process for bank holding companies to focus on systemic risk. We conclude this section with a brief discussion on when the make-them-smaller option might make sense.

Expand FDIC insurance premiums. First, we recommend expanding the ability of the FDIC to charge banks (through the deposit insurance premium it levies) for activities that increase potential for spillovers.5 The presence of spillovers makes it more likely that policymakers will resolve bank failures in a manner outside of the FDIC’s mandated “least-cost” resolution, because those spillovers impose broader costs on society. Premiums offer an established mechanism by which society can force banks to internalize potential costs.6

We use the term “expand” in referring to the FDIC’s ability to charge banks, because the FDIC has already created an infrastructure to facilitate spillover-related charges. In particular, the current premium structure allows under certain conditions for a “large bank [premium] adjustment.” The FDIC offers several rationales for the adjustment, including the need “to ensure that assessment rates take into account all available information that is relevant to the FDIC’s risk-based assessment decision.”7

The FDIC lists the types of information it would consider in setting the adjustment, and several of them provide reasonable proxies for potential spillovers. For example, the FDIC would review (1) potential for “ring fencing” of foreign assets (which would limit the FDIC’s ability to seize and sell those assets to pay off insured depositors, for example), (2) availability of information on so-called qualified financial contracts (which include a wide range of derivatives) and (3) FDIC ability to take over key operations without paying extraordinary costs.8 We might propose that the FDIC include other proxies of systemic risk, including measures of organizational complexity (such as number and type of legal entities) and a supervisory “score” of each bank’s contingency plan for winding down operations while minimizing spillovers.

The FDIC apparently believes it can price spillover risk without having to rely on size per se (although it limits this assessment adjustment to large institutions). Not having to rely on size of financial institutions seems desirable, as it more directly targets activities causing spillovers. And imposing a price on these activities would discourage them, which is the point.

However, the FDIC has limited its ability to fully incorporate such spillover-related factors into its premium. It can, for example, only adjust large bank premiums by 100 basis points or less (recently increased from 50 basis points).9 We recommend that the FDIC remove such artificial restrictions so that it can fully price the potential costs of spillovers.

Keep the cap. Second, we recommend retaining the current national deposit cap. In general terms, Congress forbids authorities from approving mergers or acquisitions if it would result in the acquiring bank holding more than 10 percent of U.S. bank deposits. This cap, which applies to M&As across state lines, was put in place by the Riegle-Neal Banking Act of 1994. Note that a bank can exceed the national cap if its deposit growth comes from a non-M&A source (that is, so-called organic growth).

Why keep the cap at the current level? We see some serious downsides to lowering the cap as a way of addressing TBTF. A lower cap could cause the bank to increase its funding from nondeposit sources, which, all else equal, could increase its susceptibility to a run. Or a firm could meet the target by jettisoning its retail banking operations and increase its securities, payments or wholesale operations. This outcome, too, would seem to increase systemic risk.

Lowering the cap effectively taxes deposits, thereby directing energies at the wrong target. While this argument might suggest abolishing or increasing the cap, we would keep it at its current level at least for the foreseeable future because its costs do not seem large. In particular, the cap has not prevented the creation of extremely large and diversified financial institutions through mergers. Thus, we doubt it has had significant scale or scope costs.

Moreover, we think the cap offers some benefits. It provides a binding limit on size growth that may offer a marginal contribution to managing TBTF. The cap may also have the salutary effect of keeping policymakers’ attention on the TBTF issue over time. Because the costs of keeping the cap seem quite low, we feel comfortable with our recommendation, even though the benefits seem low as well.

Reform the merger review process. Third, we recommend implementing a reform to the merger reviews that the Federal Reserve conducts for large bank holding companies. In 2005, we proposed that “for mergers between two of the nation’s 50 largest banks, the Federal Reserve, the Federal Deposit Insurance Corporation (FDIC) and the U.S. Treasury should report publicly on their respective efforts to address and manage potential TBTF concerns.”10 Such a requirement, which needn’t be restricted to the 50 largest banks if policymakers favor another cutoff, would highlight the key policy issues raised by the merger itself and provide a communication focus for spillover-reduction efforts. We could envision this as an interim approach if spillover reduction does not prove possible to achieve. The Federal Reserve may find it appropriate over time to support changes to the statutes governing merger reviews to allow for explicit consideration of potential spillover costs created or made worse by the merger.11

We have confidence in our preferred approach of tackling spillovers directly by putting TBTF creditors at credible risk of loss. But others with equally strong convictions have been proven wrong when it comes to financial instability, and we could be wrong as well. In that case, we must go with an alternative, and the proposed reform to make firms smaller may offer the only promising choice.

Moreover, we view addressing spillovers as the primary motivation for providing after-the-fact protection to uninsured creditors. To the degree that other motivations drive provision of such protection in the United States (for example, to reward “cronies” of elected officials or other entrenched interests), our reforms may not adequately address the TBTF problem, and other reforms might. That said, we continue to strongly believe that spillovers are the salient motivation that policymakers must address to fix TBTF (and our prior writings comment extensively on why we do not think other motivations have equal weight).


There is no easy solution to TBTF. Our longstanding proposal to put creditors at risk of loss by managing spillovers will prove challenging to implement effectively. Cutting firms down to size may seem easy by comparison. It is not. The high stakes of making firms smaller will make it difficult to determine which to shrink, and even then, the government will not have an easy time managing risk-taking by newly shrunken firms. We do take the aims of the make-them-smaller reform seriously and in that vein suggest options in this regard that we think would be more effective, including a spillover-related tax built on the FDIC’s current deposit insurance premiums.



1 Examples include Robert Reich in an Oct. 21, 2008, blog post (“If they’re too big to fail, they’re too big period”), George Shultz in the Aug. 14, 2008, Wall Street Journal (“If they are too big to fail, make them smaller”), Gerald O’Driscoll in the Feb. 23, 2009, Wall Street Journal (“If a bank is too big to fail, then it is simply too big”), Meredith Whitney in a Feb. 19, 2009, CNBC interview (reported to advocate “disaggregating” market share of largest banks) and Simon Johnson in a Feb. 23, 2009, blog post (“Above all, we need to encourage or, most likely, force the large insolvent banks to break up”).

2 The Minneapolis Fed Web site (minneapolis provides access to our fairly extensive prior writing on TBTF.

3 For literature that did not find economies of scale for large banks, see Allen N. Berger, Rebecca Demsetz, and Philip E. Strahan, 1999, “The Consolidation of the Financial Services Industry: Causes, Consequences, and Implications for the Future,” Journal of Banking and Finance 23 (2–4), pp. 35–94; and Group of Ten, 2001, “Report of Consolidation in the Financial Sector,” p. 253. For summaries of more current research finding economies of scale for larger institutions, see Joseph P. Hughes and Loretta J. Mester, 2008, “Efficiency in Banking: Theory, Practice and Evidence,” Chap. 18 in Oxford Handbook of Banking, Oxford University Press. See also Loretta J. Mester, 2008, “Optimal Industrial Structure in Banking,” in Section 3 of Handbook of Financial Intermediation and Banking, Elsevier.

4 See Raghuram G. Rajan and Luigi Zingales, 2003, “The Great Reversals: The Politics of Financial Development in the Twentieth Century,” Journal of Financial Economics 69, July, pp. 5–50.

5 More generally, George Pennacchi argues that premiums for banks should incorporate a “systematic risk” factor to account for links between a bank’s specific condition and overall economic conditions. See George G. Pennacchi, 2009, “Deposit Insurance,” paper for AEI Conference on Private Markets and Public Insurance Programs, January.

6 Some observers have outlined a broader reform along the same lines that would charge all systemically important financial firms an assessment. We focus on banks in the short term because the infrastructure for such charges already exists; charging other systemically important financial firms should have similar benefits. For a discussion of the broader change, see Viral Acharya, Lasse Pedersen, Thomas Philippon and Matthew Richardson, 2008, “Regulating Systemic Risk,” Chap. 13 in Restoring Financial Stability: How to Repair a Failed System, Wiley.

7 See Federal Register, Oct. 16, 2008, p. 61568.

8 See Federal Register, May 14, 2007, p. 27125.

9 See Federal Register, March 4, 2009, p. 9525.

10 See Gary H. Stern and Ron J. Feldman, 2005, “Addressing TBTF When Banks Merge: A Proposal,” The Region, September, Federal Reserve Bank of Minneapolis.

11 For discussions of how policymakers should or should not consider TBTF in the antitrust review process, see statements by Deborah A. Garza and Albert A. Foer before the House Judiciary Committee, Subcommittee on Courts and Competition Policy, March 17, 2009.