The Case for Breaking Up Too-Big-To-Fail Banks


The presidential campaign has brought new attention to the problem of banks that are too big to fail (TBTF).

ByEd Dolan

As everyone agrees, the largest banks are bigger than ever. As the following chart shows, the share of all bank assets held by the four largest banks rose from 33 percent in 2007 to 41 percent by 2015. Over the same period, the combined assets of the four largest banks, as a share of GDP, grew from 28 percent to 40 percent.

Image placeholder title

The major candidates disagree, not on whether the largest banks are too big to fail, but on what to do about it. Senator Bernie Sanders has made breaking up the banking giants a centerpiece of his campaign. Hillary Clinton favors a continuation of the regulatory approach embodied in the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010. The GOP candidates favor an approach that combines deregulation with market discipline.

Sanders’ anger at the banks seems to resonate well with voters, but influential voices in the media skeptical. The Editorial Board of the Washington Post has argued against breaking up the big banks. The New York Times has done likewise, prominently featuring an opinion piece written by Steve Eisman, a managing director of the investment firm Neuberger Berman. Politico also thinks breaking up the banks would be a bad idea.

I find the arguments of these critics unpersuasive. In what follows, I will examine the three approaches to dealing with the problem of TBTF and explain why I think Sanders is right to think that a reduction in the size and influence of the largest banks should be a part of any comprehensive plan to improve the stability of the financial system.

The Regulatory Approach and Its Limitations

Let’s begin by looking at the strengths and weaknesses of the regulatory approach.  to financial stabilization, as embodied in  The objectives of Dodd-Frank, the most important piece of legislation to come out of the 2008 crisis,  are to reduce the risk of future bank failures by curbing risk-taking and to provide for a more orderly resolution of insolvent institutions when and if failures do occur. The original legislation left it to regulators to fill in many details, but they are making progress. Federal Reserve Governor Lael Brainard summed up both the progress and work that remains to be done in a speech last summer:

  • Capital is the first line of defense against bank failure. Banks that are designated as systemically important financial institutions (SIFIs), will face a “capital surcharge” that requires them to hold up to twice as large a capital buffer as smaller institutions. Final rules have been issued and the surcharge will become fully effective in January 2019.
  • Inadequate liquidity is another source of failure. In times of crisis, banks may be forced to sell illiquid assets at fire sale prices and may be cut off from sources of short-term funding on which they normally rely. To protect against excessive liquidity risk, regulators now require systemically important banks to hold extra protection in the form of high-quality liquid assets. Before the crisis, large banks relied more on potentially risky wholesale funding than smaller banks did. Now they are less reliant.
  • Before the crisis, senior bank executives and risk managers received much of their compensation in the form of bonuses and stock options that gave them an incentive for short-term risk taking. Now banks are required to defer payment of some of that compensation for up to three years, and to adjust payment for losses that occur.
  • The above measures do not guarantee that no large banks will fail. Accordingly, systemically important banks are required to submit plans for orderly resolution in case of failure, informally known as “living wills.” In theory, these plans would allow regulators to minimize the collateral damage from any failures of large banks that did occur.

Unfortunately, these regulatory efforts may not be enough. Part of the reason is that the banks themselves, with armies of lobbyists and tame members of Congress, have been fighting them all the way, whether by trying to amend the law section by section, stalling efforts to finalize rules, or working to place the effective date of those rules as far in the future as possible. But that is just the beginning.

Politics aside, there are serious technical difficulties in defining and measuring basic concepts like capital and liquidity. Fed Governor Brainard’s generally optimistic assessment of Dodd-Frank is full of cautions and qualifications, such as, “What may seem like thick capital cushions in good times may prove dangerously thin at moments of stress, when losses soar and asset valuations plummet.” Or, “As we learned from the crisis, risk modeling and risk weighting are subject to considerable uncertainty, and stressed financial markets can make even the most rigorous risk assessments look optimistic in hindsight.” These technical difficulties mean that rules setting minimum standards for capital, liquidity, and other important risk indicators are necessarily complex. Complex rules, in turn, make it easier for banks to game the rules with strategies that comply with the targets in form but evade them in substance.

Most importantly of all, regulations that aim to reduce risk taking of any kind are vulnerable to offsetting behavior or risk homeostasis. The idea is that people have a level of risk they are comfortable with, so that if regulations force them to behave more safely in one regard, they compensate by taking on risk in some other way. This phenomenon was highlighted in the 1970s by economist Sam Peltzman, who argued that requiring drivers to wear seatbelts would encourage them to drive faster. As a result, fatalities would not fall as much as safety regulators hoped, and might even fail to decrease at all.

By its nature, banking is open to offsetting behavior. For example, up to a point, banks can increase their expected profit by making riskier loans, which carry higher interest rates. However, doing so increases the chance that an unlucky spate of defaults could exhaust their capital. The cost of carrying additional capital to protect against insolvency thus limits their appetite for risky loans. However, if regulations mandate a higher capital ratio than the banks would normally carry, they have every incentive to review their portfolio of assets, moving back toward their equilibrium tolerance for risk by adding high interest loans that they would otherwise have passed up.

Risky loans are only one of many possible form of offsetting behavior. Banks might, instead, choose to take on added risk through derivatives trading, or increased exposure to exchange rate risk, or changes in liquidity management.

Unfortunately, the fact that the largest banks are subjected to more burdensome regulations than their smaller competitors may be making the situation worse. In Congressional testimony last year, Federal Reserve Chair Janet Yellen noted that the largest banks could opt out of extra regulatory burdens like capital surcharges and living wills by reducing their systemic footprints, but there has been little sign of that to date. Meanwhile, those and other extraordinary measures mean that the largest banks also have the largest incentives to engage in offsetting behavior.

Sanders Breakup Proposals Also Have Their Limitations

If there is no way to eliminate the risk of bank failure, then the only way to solve the too-big-to-fail problem is to make sure that no banks are too big. That is the economic logic behind the Sanders proposal for breaking up the largest banks. His proposal for doing so has two parts.

First, he proposes reinstating the Depression-era Glass-Steagall Act in a 21st century version. The original purpose of that law was to separate commercial banking (the business of taking deposits and making loans) from investment banking (the business of underwriting and making markets for securities). Glass-Steagall was gradually weakened over time and was finally repealed in 1999.

Second, Sanders proposes to apply Section 121 of the Dodd-Frank Act, sometimes described as the “nuclear option.” Section 121 gives regulators the power to restrict the scope of any institutions that “pose a great threat to the financial stability of the United States.” As a last resort, it authorizes them to force such institutions to “sell or otherwise transfer assets or off-balance-sheet items to unaffiliated entities”—bureaucratese for breaking them up.

But Sanders probably overpromises when, in his campaign appearances, he says he will break up the big banks during his first year in office. There would be some hard work in store, both political and technical, before the job could be finished.

Politically, the problem is that the president can’t do much by executive action. Bringing Section 121 to bear would require majority approval both by the Fed’s Board of Governors and by the Financial Stability Oversight Council. That clumsy interagency committee that has ten voting members, including the Fed Chair, the Secretary of the Treasury, the head of the Securities and Exchange Commission, and the Comptroller of the Currency, among others. If the incumbent members of these bodies dragged their feet, replacing them would require Senate confirmation, something that could be problematic for nominees who announced in advance that it would be their intention to break up large and politically powerful banks. A reincarnation of the Glass-Steagall Act would require action by both houses of Congress.

On the technical side, someone would have to figure out just how many pieces of what size and shape the big banks should be broken into. Splitting them along the lines of investment banking vs. commercial banking would be a start, but the resulting entities might very well still be too big to fail, in some cases. The living wills that banks are supposed to submit may eventually include provisions for spinning off healthy pieces of a failed institution, but the Fed has not been very satisfied with the early versions of these plans. Officials who have reviewed submitted drafts complain, among other things, that the corporate structures of the big banks do not closely follow their functional structures, making a hypothetical breakup that much harder.

In short, breaking up the big banks, while not impossible, would require more than issuing a few executive orders.

The GOP Solution: Market Discipline

While Democrats argue over regulation vs. breakup, Republicans face a political dilemma of their own. On the one hand, Republicans are at least as friendly to the banking industry, and at least as willing to take Wall Street campaign contributions, as is the Clinton wing of the Democratic Party. On the other hand, the rural and working class conservatives whose votes the GOP courts are no less angry at the big banks than are supporters of Sanders, the democratic socialist.

The idea Republicans invoke to square this circle is market discipline. In part, market discipline means deregulation. To please their allies on Wall Street, the Republican candidates have uniformly pledged to repeal Dodd-Frank in its entirety. At the same time, to please grass roots conservatives, they promise that market discipline will also mean no more bailouts. Instead, they propose to treat banks like lawn services or Christmas tree farms: Let them risk their capital as they would like, and if they go under, they are on their own.

There is definitely a need for market discipline in banking. If banks know they will be rescued if they fail, they have less incentive to avoid excessive risks—a problem known as moral hazard. Moral hazard affects not only managers and shareholders, but also other creditors of banks. If bondholders or parties that make short-term loans to banks think their claims will be guaranteed in case of default, they will supply funds at lower interest rates. If investors think big banks are more likely to be bailed out than small ones, the cost of funds to big banks will be lower, giving them a competitive advantage that allows them to grow even bigger. Moral hazard played a significant role in the 2008 crisis and has not yet been brought fully under control.

Still, the superficially appealing formula, “No more regulation—no more bailouts” is unrealistic. It fails adequately to recognize some fundamental differences between the financial structure of banks and that of nonfinancial companies.

One difference is that banks are more interdependent than other businesses and more prone to contagion. If one lawn service in a town fails, that does not harm others in the same business. In fact, it probably helps them, because they can pick up customers from their failed competitor. The financial system operates differently. The failure of a major player like Lehman Brothers sends our shockwaves that undermine the soundness of competitors, threatening collapse of the entire system.

Another important difference is that nonfinancial firms, whether large or small, typically have a more generous cushion of equity on their balance sheets than do those in the financial sector. There is no one-size-fits all rule, but broadly speaking, analysts consider it healthy for a nonfinancial business to have equity capital equal to or larger than its total debt. Such a business can sustain losses equal to half or more of its assets before it becomes insolvent. If it does fail, the sale of remaining assets in bankruptcy can often repay a substantial part of what is due to creditors.

Banks are much more thinly capitalized. Before the 2008 crisis, many of them operated with capital equal to 5 percent of their liabilities, or even less. Although that has now risen to an average of 6 to 10 percent for US banks, depending on how it is measured, it is still far less equity than is normal for nonfinancial corporations.

Because they hold less capital, banks can quickly become insolvent in a crisis. Once they do, regulators have two choices—a bailout, at taxpayer expense, or a bail-in, which means shifting losses to those who hold the bank’s liabilities. Small depositors are protected from bail-ins by deposit insurance, so the losers are bondholders, holders of unsecured short-term debt, and sometimes, unsecured depositors. (For a detailed discussion of bailouts, bail-ins, and bank rescues, see this tutorial.)

As the 2013 Cyprus crisis showed, bail-ins of depositors and other creditors can be very messy. The losers can include not just other financial institutions, but pension funds, small businesses, and even not-for-profit entities like university endowments. These parties may not be well positioned to absorb losses. They may have to curtail their own operations, and in doing so, shift the losses to still other firms and households. From 2012 to 2014, the economy of Cyprus shrank by more than 10 percent.

The thin capitalization of the financial system, combined with its susceptibility to contagion, means that applying the Republican formula of “No regulation, no bailouts” would both increase the likelihood of bank failures and maximize the potential damage when they do occur. The larger the biggest banks are when the next crisis comes, the more likely it will be that Congress, Republican majority or not, will authorize a bailout. A bailout of an unregulated system headed by fragile, undercapitalized giants would be even more expensive and chaotic than the TARP bailout of 2008.

The Bottom Line

The bottom line is that the too-big-to-fail problem is still with us, and no one approach will be enough to protect the financial system when the next crisis comes. However, breaking up or shrinking the largest banks, as candidate Sanders proposes, should be part of the mix.

Some of those who argue against breakup say the regulatory regime is working well enough that such drastic action is no longer necessary. It is true that Dodd-Frank includes tools that are intended to deal with excessively large banks. Those include capital surcharges, living wills, and the “nuclear option” of Section 121. However, those special provisions for SIFIs are a classic example of a kludge—engineering slang for a work-around solution that is clumsy, inelegant, inefficient, and difficult to maintain. The whole regulatory system would work more smoothly if there were no banks or other financial institutions large enough to pose a systemic threat in the first place.

The same goes for market discipline. Yes, more market discipline would be useful—essential, in fact—as a way to curb moral hazard and excessive risk taking. However, if the banks at the top of the system remain too big to fail, any promise of “no more bailouts” is simply not credible, either economically or politically. The largest banks will know that despite any pledges that might have been made, the White House and Congress—whether controlled by Republicans or Democrats—will blink when faced with the choice between another TARP or a systemic meltdown.

That said, I would add that breaking up or shrinking the largest financial institutions is not in itself a complete program for financial stability. For one thing, it is not a ready-to-go option that a president could invoke by executive action. As critics point out, a lot of groundwork, both technical and political, remains to be accomplished before it could actually happen.

Furthermore, the nature of banking is such that even a financial industry with less concentration at the top would still be subject to contagion and systemic failure if poorly regulated. The savings and loan crises of the 1980s, which took place when banks were smaller and Glass-Steagall was still in force, is a case in point.

Senator Sanders deserves credit for using his campaign to keep too big to fail in the public eye. For the most part, he has focused on political arguments for breaking up the big banks, but there are sound economic arguments for doing so, too. Let’s hope that he, or whoever else becomes president, is able to make some real progress toward resolving the problem of TBTF before the next financial crisis strikes.


Emerging Markets