Ending "Too Big to Fail": What's the Right Approach?

In a recent speech at the Hutchins Center at the Brookings Institution, Neel Kashkari, the new president of the Federal Reserve Bank of Minneapolis, argued that we need new strategies to tackle the problem of “too big to fail” (TBTF) financial institutions. On Monday, I’ll be on a panel at the Minneapolis Fed on the issue. This post previews my comments. In short, it seems to me that a lot of progress has been made (and more is in train) toward reducing the risks that large, complex financial institutions pose for the financial system and the economy. To say that “nothing has been done” is simply not correct. That said, because it’s really important to get this right, thoughtful debate on the issue is necessary and welcome.

At the 50,000-foot level, a key question is the extent to which structural change in the financial industry is needed to end TBTF, and, to the extent it is, what that change should look like. The argument of this post is that, while substantial and even fundamental changes may ultimately be necessary, we don’t yet know exactly what they will be. Instead, the legacy of the Dodd-Frank Act, the Basel agreements, and other reforms is a sensible process which, with sustained effort, will help us solve the problem. A key element of the strategy is that it gives banks strong incentives to shrink or otherwise restructure themselves to reduce the risk they pose to the financial system.

Why not just break up big banks? Of course, some find the ongoing process too slow or ineffective. If some banks are “too big to fail,” critics argue, why not take a more direct approach and make them smaller—for example, by putting stringent limitations on the assets or liabilities that any one bank is allowed to hold?1 Let’s stipulate that modest size reductions are not going to accomplish much—after all, Lehman Brothers was only about a third the size of the largest banks when its failure in September 2008 almost brought down the global financial system. So, for this more-direct strategy to be credible—that is, for people to be convinced that no institution will be bailed out if it gets into trouble—it would have to involve very substantial reductions in the scales of the largest banks.

As I’ve already suggested, if major structural changes in the banking system are necessary to avoid another crisis, to promote financial stability, and to control moral hazard and excessive risk-taking, then we should all be for making the changes. However, the forced breakup of large firms, say through the imposition of arbitrary limits on assets, doesn’t seem to be a smart way to achieve those objectives, at least until reasonable alternatives have been tried. A strategy focused on breaking up big banks has at least two important drawbacks:

First, having large firms in the financial industry, as in other industries, has benefits as well as costs. Certainly, TBTF is a (big) problem, and there are other concerns associated with size, like undue political influence on the part of large firms.2 But, as in other industries, large financial firms also have cost advantages in some areas and can provide services that smaller firms cannot. Potential advantages of size for banks (not associated with TBTF status) include the ability to exploit economies of scale (in technology, in establishing networks, in branding), greater diversification of risks, the spreading of fixed overhead costs over many activities, the ability to offer combinations of complementary products, and global reach. Even putting aside the short-term costs and disruptions that would likely be associated with breaking up the largest banks, in the long run a US financial industry without large firms would likely be less efficient, providing fewer services at higher cost. From a national perspective, this strategy could also involve ceding leadership in the industry, and the associated jobs and profits, to other countries.

It’s true that economists disagree on the returns to size in banking.3 There are no doubt diseconomies of scale as well (e.g., “too big to manage.”) But uncertainty about the costs and benefits of size can itself be a reason for caution in applying arbitrary size limits, particularly in a world in which firms and business models are heterogeneous. Jeremy Stein, in a speech given when he was a Fed governor, provided a simple but insightful illustration of this point:

Consider the following example. There are three banks: A, B, and C. Banks A and B both have $1 trillion in assets, while C is smaller, with only $400 billion in assets. Bank A actually generates significant economies of scale, so that it is socially optimal for it to remain at its current size. Banks B and C, by contrast, have very modest economies of scale, not enough to outweigh the costs that their size and complexity impose on society. From the perspective of an omniscient social planner, it would be better if both B and C were half their current size.

Now let’s ask what happens if we impose a size cap of say $500 billion. The size cap does the right thing with respect to Bank B, by shrinking it to a socially optimal size. But it mishandles both Banks A and C, for different reasons. In the case of A, the cap forces it to shrink when it shouldn’t, because given the specifics of its business model it actually creates a substantial amount of value by being big. And in the case of C, the cap makes the opposite mistake. It would actually be beneficial to put pressure on C to shrink on the margin…but since it lies below the cap, it is completely untouched by the regulation.


Given that size in banking can have social benefits as well as social costs, and that banks and their business models are (and need to be) heterogeneous, it’s important to ask whether we can achieve a simpler and safer banking system in a way that does not disrupt the desirable functions of that system more than necessary.

A second objection to a strategy based primarily on breaking up large banks is that size is not the only relevant attribute of banks; in particular, size is only one among a number of determinants of whether the failure of a particular firm is likely to be broadly destabilizing. Financial panics can occur in systems dominated by small banks as well as by large ones, as was the case in the United States in the Great Depression. (In contrast, Canada, which has only large banks, did relatively well in both the Depression and in the recent crisis.) The basic elements of a financial panic—broad-based loss of confidence in banks, runs by providers of short-term funding, fire sales of bank loans and other assets, disruption of credit flows—can arise even without large banks. Yes, the problems of some large institutions, notably the collapse of Lehman Brothers in September 2008, intensified the recent crisis. But, factors other than size, including complexity, opacity, illiquidity, and interconnectedness with other firms, contributed to the catastrophic effects of Lehman’s collapse, as did the fact that the government did not have the legal authorities it needed to manage Lehman’s demise in a more orderly way. A more nuanced approach to ensuring financial stability and ending TBTF should take size into account, but other factors as well.

So what is the current strategy for tackling TBTF? This is not the place to review in any detail the extensive reforms that have been made to financial regulation since the crisis. But broadly, the government’s current approach to ending TBTF has two main parts:

First, to make the financial system as a whole more resilient to shocks, there has been increased regulatory attention to firms that may pose systemic risks. For example, on top of measures to strengthen capital for all banks, regulators have imposed a capital surcharge (that is, a higher required level of capital) on systemically important banks, which rises according to regulators’ assessment of the systemic risks associated with the firm. Systemically important banks must also pass tough stress tests and meet higher standards for liquidity and risk management. More generally, systemically important firms (including nonbanks thus designated by the Financial Stability Oversight Council) are subject to heightened oversight and additional restrictions by the Fed, FDIC, and sometimes other regulators as well. The intent of these rules is to make firms that pose systemic risks less likely to fail, obviously, but as I’ll discuss, they have incentive effects as well.

The second part of the program for ending TBTF is the ongoing development of a new failure resolution regime, including the so-called orderly liquidation authority (OLA), created by the Dodd-Frank Act. The OLA directly attacks the TBTF problem by giving regulators (the Fed and the FDIC, in most cases) the legal tools necessary to resolve a systemically important financial firm on the brink of failure, in a manner that involves no taxpayer funding (shareholders and creditors bear all losses) and which mitigates the spillovers to the broader financial system and the economy. Resolving a complex firm, particularly in a situation in which the financial system as a whole is in turmoil, poses substantial challenges. But, as discussed here and here, for example, a great deal has been accomplished. The single most important development has probably been the enunciation by the regulators of the so-called single-point-of-entry strategy, which both simplifies and makes more predictable any intervention in a failing firm. Other key elements include the extensive planning taking place at the Fed and the FDIC; requirements that firms supplement their equity capital with long-term debt that can be converted to equity when needed (“bail-in-able debt”); protections for short-term creditors and other measures to forestall runs; and the requirement that firms present plans for their own resolutions—so-called “living wills”—that must be approved by the regulators. In order to receive approval of living wills, firms have already begun to make structural changes to enhance their resolvability—simplification of legal structures, for example—and more such changes are likely.

A work in progress. As I said earlier, I see a lot of progress on the TBTF issue, more than the public appreciates. But more remains to be done to fully implement and calibrate the anti-TBTF program. For example, as the recent rejection of living will submissions by five large banks illustrates, large firms have more work to do, probably including some substantial structural changes, to give regulators confidence that they can be safely resolved. More consultation with foreign regulators will also be needed to ensure appropriate coordination in resolving multinational firms. In the area of capital regulation, controversy still dogs some key issues, including the fundamental question of what constitutes an adequate level of capital. These are important debates.

My takeaway is not that the problem is solved—that will take more time—but rather that the current approach amounts to a process that will help us find the solution. In particular, the government’s strategy for ending TBTF addresses the deficiencies, noted above, of imposing arbitrary limits on bank size. Most obviously, the strategy does not make the mistake of treating size as the only determinant of systemic risk (e.g., capital surcharges depend on a variety of criteria). Particularly interesting, though, is that the government’s strategy, by enlisting market forces and making good use of incentives, is more likely to lead to “right-sizing” of banks, where by “right-sizing” I mean getting right (from a social perspective) not only size (assets and liabilities) but also other dimensions of firm structure and organization that bear on systemic risk.

As Jeremy Stein’s example above illustrated, breaking up banks through regulatory fiat is likely to be quite inefficient, especially when the reasons that banks choose to be large and complex are heterogeneous. If, as seems probable, bank managers and shareholders better understand the institution’s motivations for size and complexity than regulators do, it makes sense to use that knowledge. To do that, the right incentives need to be provided: The privately perceived benefits of TBTF status need to be reduced and the costs increased, so that bank managers and shareholders are considering something closer to the social costs and benefits of size and complexity when they think about how to organize their business. In terms of Stein’s example, we should be okay with a bank (Bank A) staying large if size allows it to provide substantial economic (non-TBTF) benefits, while at the same time inducing banks (Banks B and C) whose size mostly reflects TBTF motivations to shrink.

To a first approximation, that’s what the government’s approach aims to do. For example, the capital surcharge and similar regulations directed at systemically important institutions act like taxes on size and complexity. This should induce managers and shareholders of systemically important institutions to take into account, in their decisions about firm scale and structure, the implications for financial stability and TBTF. That is, the extra costs that regulators impose on systemic institutions force their decisionmakers to “internalize the externality” that their firms create for the financial system.4 Similarly, the development of the liquidation authority (which raises the probability that creditors will take losses) and improvements in the overall resilience of the financial system (which would reduce any incentive that future regulators might have to try to engineer a bailout) should reduce the perceived benefits associated with TBTF status, as measured in terms of funding costs, for example. Putting creditors at risk also brings market discipline back into play, putting additional pressure on managers not to take excessive risks. Together with the requirements imposed by the living will process, better incentives for managers, shareholders, and creditors should lead, over time, to a banking system that is safer, but also more competitive and efficient.


1 Some restrictions on bank size are already in place. As required by the Dodd-Frank Act, the Fed has approved a rule prohibiting mergers that result in a financial institution whose liabilities exceed 10 percent of US financial liabilities. A previously existing rule prohibited bank mergers that created an institution with deposits exceeding 10 percent of the national total. Although these safeguards are aimed at preventing banks from getting bigger, rather than making them materially smaller, the Dodd-Frank Act also created authorities that regulators could use to break up banks: Notably, Section 165(d) of the Dodd-Frank Act empowers the Fed and the FDIC to break up or otherwise restructure a firm that does not provide credible plans (a “living will”) for its own resolution; and Section 121 of the Act empowers the Financial Stability Oversight Council, a council of financial regulators, to break up a firm that it deems a “grave threat” to U.S. financial stability.

2 Firm size does not appear to be the only source of influence, though. Community banks, tiny individually and a small part of the sector even taken together, often wield substantial political power. One could also imagine addressing the issue of political influence more directly, through greater transparency about corporate contributions for example.

3 See here for recent research that finds significant scale economies for even the largest firms.

4 Taxes intended to force managers to take account of the external effects of their decisions—pollution charges, for example—are called Pigouvian taxes. I’m arguing, with Stein, that higher capital charges and other more onerous regulations on systemically important firms serve as a kind of Pigouvian tax.

© Brookings Institute

Read more commentaries by Brookings Institute