Good afternoon. Thank you for that warm introduction. I appreciate the Minnesota Chamber of Commerce hosting this event today and inviting me to address its members. Before I begin, I would just like to remind everyone that the views I express today are my own and not necessarily those of the Federal Open Market Committee or the Board of Governors, which sets supervision and regulatory policy for the Federal Reserve System.
In today’s talk, I will cover two main items. First, I will explain why we launched a major initiative to end too big to fail (TBTF). I continue to think that the largest banks1 in the country are too big to fail, and I am skeptical that current efforts to fix that problem will ultimately work. I will provide a few key examples from the recent financial crisis, when I ran the Troubled Assets Relief Program (TARP), which explain why I am skeptical. Second, I will summarize where our initiative stands today. We just had our first symposium,2 which focused on two transformational reforms to end TBTF. I will summarize key points I took away from that event. I will then conclude with issues that I continue to wrestle with as we seek the right reform framework. Then I will be happy to take your questions.
Current Efforts to Address TBTF
I come at the too-big-to-fail problem from the perspective of a policymaker who was on the front line responding to the 2008 financial crisis. In the past six years, legislators and regulators have worked hard to address the TBTF problem. My colleagues in the Federal Reserve System, working closely with other financial regulators, have implemented important tools and regulations that are making the financial system stronger under the reform framework Congress established.
I agree that many current reform efforts are headed in the right direction, particularly those that make banks stronger with additional capital, deeper liquidity and stress testing. But I am not sure those measures go far enough. More importantly, we know that banks will still get into trouble. So we need a way to deal with failing large banks. The current reforms attempt to address this problem with a new legal framework to resolve failing banks combined with a plan to bring new investors into banks who agree to absorb losses when the bank gets into trouble. The idea is that by having these new investors take the hit, taxpayers will not be on the hook. This sounds good, but it has not worked in practice in prior crises, and I doubt it will work in the future. I fear policymakers will have to turn to taxpayers rather than impose losses on creditors.
My experiences from the 2008 crisis highlight this problem. One of the toughest challenges we faced was dealing with risk spreading between large banks: Multiple large banks were under stress at the same time, and actions we might have taken to recapitalize one bank (by haircutting creditors or counterparties) could actually lead to increased stress at other banks as their creditors worried they too might face losses. In effect, a “run” at one bank could quickly become a run at multiple banks. This risk is one reason the Treasury Department asked Congress in 2008 to authorize the TARP, giving the government the ability to inject capital rather than haircut creditors as a means of recapitalization. In fact, Treasury’s announcement on October 13, 2008, to simultaneously inject capital into nine of the largest American banks was done in part to stop the spreading of risk from one bank to another.
The challenge of sticking debt holders with losses also occurred with the government-sponsored enterprises (GSEs) Fannie Mae and Freddie Mac, providing another relevant example of the failure of debt as a loss absorption instrument during a crisis. In 2001, the GSEs began issuing subordinated debt with the specific intention that it would be available to absorb losses if they ran into trouble. The GSEs compensated investors with higher yields than senior debt holders for the additional risk they were taking. During the 2008 crisis, because of widespread risk in financial markets, Treasury chose to protect the GSEs’ subordinated debt rather than haircut it. The risky economic environment prevented debt that had been specifically issued to absorb losses from being harmed. Instead, taxpayer money again had to be used to recapitalize the GSEs.
Let me point to one more example from the crisis about the difference between intentions to impose losses and actual behavior. Some large banks had sponsored off-balance-sheet structured investment vehicles (SIVs) before the crisis. They had no legal obligation to provide a financial backstop if their SIVs ran into trouble. Yet during the crisis, some banks chose to rescue their SIVs because of reputational risks they would incur if they had allowed the SIVs to collapse. The SIVs’ liabilities became the sponsoring banks’ liabilities. During the crisis, reputational risk broke through what had appeared to be a strong legal firewall.
Perhaps it is easier to impose losses on creditors today than it was during the crisis, but a recent case gives me pause. Some large European banks have issued contingent convertible debt, which is designed to convert to equity if a bank runs into trouble. This debt, however, may actually be a mechanism that spreads uncertainty from one big bank to another. We saw this in recent months as concern arose that some banks might have insufficient capital to be eligible to pay coupons on their contingent convertible debt securities. Proponents of such contingent debt argue that it will be cheaper for banks to issue than common equity. Given the volatility of this debt in recent months, it is hard to see bond investors not demanding equity-like returns for the risks they are taking. If it isn’t significantly cheaper than equity as a funding source, the benefits of the additional complexity and unpredictability are not justified. And I have serious concerns that such contingent convertible debt will add to uncertainty in a crisis environment rather than reduce it. At a minimum, regulatory structures must not make a crisis worse.
Our ability to impose losses on creditors of large banks is just as important now as it was prior to the 2008 crisis. The recent decision by the Federal Reserve and FDIC to deem the resolution plans of some of the largest banks in the country not credible suggests that their complexity, operations and structures continue to make their potential failure a real challenge for all of us. To me, this means we must work even harder to reduce the likelihood of large bank failures because resolving them in ways that does not trigger widespread economic harm is proving so difficult.
What do I take away from these examples? First, as we analyze the current regulatory framework and other potential solutions to TBTF, we must attempt to understand whether they would really allow creditors and not taxpayers to take losses. Second, we must work to assess whether these plans increase the transmission of risk from one bank to another. Third, complex securities and legal protections may not behave as expected in a future crisis. If they don’t, taxpayers will likely be on the hook.
Update on Our Initiative to End TBTF
Our initiative includes a series of public symposiums to be held throughout this year. Experts from around the country will help examine these issues and will contribute to informing our plan to end TBTF, which we will release to the public by the end of the year. Today we published a summary of the first symposium,3 and all presented materials and video recordings are available on our website at minneapolisfed.org.
When I gave my initial speech on ending TBTF in February,4 I highlighted three potential options experts have offered to address systemic risks posed by large banks: (1) breaking them up, (2) substantially increasing capital requirements and (3) taxing leverage across the financial system. But we also asked for other ideas from experts and the public at large. Experts have pointed us to another potential option: alternative resolution mechanisms that could address some of the perceived shortcomings of current resolution plans. We have added this fourth option for consideration and could potentially add others as we move forward with our initiative.
In our first Ending TBTF symposium on April 4, we heard from experts about the need to assess the costs and benefits of proposals, and we explored the first two potential solutions to TBTF: substantially increasing capital requirements and breaking up large banks. I will now offer my perspective on the ideas that were discussed at the symposium.
Costs and Benefits of Regulation
Former Federal Reserve Governor Randy Kroszner made a strong case for the need for carefully considering costs and benefits in assessing current and potential future regulatory reforms. He argued that we need to consider the costs of new regulations against the benefits of increased safety in our banking system. How much safety do we want, and what price are we willing to pay for that safety?
A simple analogy to illustrate the trade-offs between increased safety and increased costs comes from airport security. We understand that air travel includes some inherent risk of falling victim to a terrorist attack. The odds are very low, but not zero. One cost of safety is waiting in airport security lines. Of course, we could always be safer. For example, we could discard x-ray machines and instead hand-search all bags individually. But what is the cost of this added safety? Instead of 15 minutes to go through security, it might take an hour. Instead of the ticket prices we are familiar with, routine flights might cost tens or hundreds of dollars more to pay for additional TSA agents. As a society, we have decided that the costs of extra safety provided by hand-searching all bags outweigh the benefits. As the terror threat environment changes, for better or for worse, we may adjust the amount of safety we are willing to pay for in both time and money.
Preventing banking failures that could trigger a widespread economic downturn poses similar trade-offs between safety and costs. The benefits of a more resilient financial system are clear: less-frequent and less-severe financial crises, resulting in reduced loss of income, jobs, savings and foreclosures for Main Street. What are the costs? Potentially lower average economic growth between financial crises. For example, increasing regulatory requirements as a result of Dodd-Frank has forced many banks to hire more compliance officers and auditors. Requiring banks to issue more equity may increase their average funding costs. These cost increases would have to at least in part be passed on to customers in the form of higher borrowing costs. Increased borrowing costs for businesses to fund a new factory, or for families to fund a home or car purchase, could lead to lower economic growth for society. However, we must consider not only the expected growth rates between crises, but also the expected growth rates including the effects of financial crises. An illustration of two simplified economies highlights the choice: One economy has a less-regulated financial system and experiences a higher average growth rate between financial crises. But it also experiences more-frequent and more-severe financial crises than the more-regulated, slower-growth economy. Depending on the costs and benefits of the various regulations, Main Street could be better off in one economy or potentially the other.
I found the discussion of costs and benefits compelling—but, as Governor Kroszner noted, we must not suffer from “analysis paralysis.” Weighing of costs and benefits will be an important part of our work to develop our plan to end TBTF.
Substantially Increasing Capital Requirements
Professor Anat Admati of Stanford presented her recommendation to substantially increase capital requirements for the largest banks. Before the financial crisis, large banks averaged 3 percent shareholder capital. Today, with the implementation of the Dodd-Frank and Basel III standards, large banks have approximately 6 percent shareholder capital.5 The presenter made the case that is far too low and should be closer to 25 percent, noting that nonbanks typically have at least 30 percent equity funding.
The advantages of substantially increased capital requirements are that banks are much better positioned to withstand unknown, and perhaps unforeseeable, shocks in the future. Increased capital will force banks to internalize the negative externality of the costs to Main Street of a financial crisis.
Discussants noted that substantially higher capital requirements could lead to higher borrowing costs for businesses and families, which then have the potential to slow economic growth. There was a wide range of views about the magnitude of such effects.
Discussants also noted that by substantially increasing capital requirements of large banks, risk may be encouraged to migrate from the highly regulated banking sector to the unregulated shadow banking sector. For example, if just as many subprime loans were written leading up to the financial crisis, but rather than being held by large banks, they were held by other highly leveraged financial firms, would the financial system have been any safer? Perhaps not. I found this to be an important consideration as we work to reduce systemic risk.
As I listened to the wide range of views on the merits and costs of higher capital, it struck me that we could increase capital requirements in a straightforward way to address TBTF. As we saw in my earlier comments about the behavior of legal structures in a crisis, there is a strong argument that simpler solutions are more likely to be effective than complex ones, so I see virtue in focusing on increasing common equity to assets, which seems the simplest and potentially the most powerful in terms of safety and soundness.
More capital has another advantage. We have seen extraordinary structural changes in our financial system over the past several decades. We have to be prepared for more change. More capital will absorb losses even from activities we cannot anticipate today.
Thinking back to my earlier comments about risk spreading between banks in the 2008 crisis, I realize that capital by itself is not a direct firewall to contain risk, but at a minimum, it does not intensify the spreading of risk from bank to bank. If all of the largest banks had enough capital that investors were confident in their strength, even during an economic shock, such concerns about shared risks could be substantially reduced. We saw this in 2008, as noted above, when Treasury injected capital into the nine large banks simultaneously.
Altering the Organizational Structure of Large Banks
Professor Simon Johnson of MIT and senior fellow at the Peterson Institute for International Economics presented his recommendation to impose an effective cap on bank size of 2 percent of GDP, or approximately $350 billion. His proposal is for capital requirements to dramatically increase if banks exceed that size requirement. The presenter noted that the failure of Lehman Brothers, which had in excess of $600 billion in assets, triggered widespread damage to financial markets in 2008, while the eventual failure of CIT Group, with $80 billion in assets, but potentially up to $120 billion in total exposure, did not.6
The presenter argued that having large banks at the center of the U.S. financial system is a fairly recent phenomenon, taking off in the past couple of decades, and yet economic growth was strong in the decades before banks became so large and concentrated. He argued that there is little evidence this recent growth of large banks has led to real economic benefits for the U.S. economy. I had not heard this argument before the symposium and found it compelling.
Discussants appropriately noted that simply having numerous smaller banks is not necessarily less risky than having one large bank, depending on the underlying risks each bank takes. For example, if a $3 trillion bank were to split into 10 $300 billion banks, and each of those new banks made identical investments, there is little enhancement to financial stability. If a shock were to hit the economy that imposed large losses on the original $3 trillion bank, the 10 smaller banks would simply share the identical losses, each coming under similar distress.
I believe this argument, while logically true, must be tempered. It is unlikely that 10 smaller banks, each with its own management team and board of directors, would all make identical investment decisions to one another and to the original consolidated bank long into the future. To the extent their business strategies differed, financial stability should be enhanced.
Discussants also expressed doubts about how easy it would be to actually implement a “break-up” plan if it were enacted. An image of government analysts arbitrarily deciding how to divide a trillion-dollar bank into smaller pieces, deciding, for example which branches and which operations went with which new entity, gives many people concern—including me. The presenter noted that rather than the government deciding how to divide a large bank, incentives should be put in place such that banks decide how to optimally divide themselves to meet a new regulatory cap. In fact, the new requirements put in place after the financial crisis have already encouraged some large financial institutions to shed some operations and assets. While the magnitude and speed of their spin-offs may not satisfy financial stability concerns, they do demonstrate firms’ ability to downsize in response to new requirements. I believe that given sufficient incentive, banks would be able to restructure themselves.
Discussants presented analysis that large banks benefit the economy by providing economies of scale. While there is a wide range of estimates of such benefits, they should lead to lower operating costs compared to relatively smaller banks. I believe these scale arguments need to be thought of in the benefit and cost framework I mentioned before, as banks with scale may also pose significant risks to stability. But accepting it at face value, I worry that Dodd-Frank is adding to the advantage large banks have over small banks, given that complying with regulatory costs likely exhibits scale economies. We need a regulatory system that does not add to the advantage large banks have over small ones.
Reflecting on our first symposium, I am left with some important questions for further consideration:
- Should we view proposals to address TBTF in isolation, or could we combine them?
- How do the costs and benefits of proposed solutions line up relative to the status quo?
- How much confidence do we have that the solutions will perform as expected in a crisis environment?
- Will markets think the proposal credibly puts creditors at risk of loss?
- Will the solutions merely push risk into unregulated areas of the financial markets?
- Will the solutions promote fairness between the regulatory burdens imposed upon large, medium and small banks?
- How likely are the solutions to remain effective over decades?
The Minneapolis Fed’s initiative to end TBTF is still in its early stages. Our next symposium is scheduled for May 16. We will hear from Professor John Cochrane of the Hoover Institution, who will present his plan to tax leverage across the financial system as a means of reducing systemic risk. In addition, we will hear from John Bovenzi of the Bipartisan Policy Center and Oliver Wyman, who will present an alternative resolution mechanism. We will once again have a group of experts bring a wide range of experiences and perspectives to respond to those proposals and offer their own ideas.
Like the first one, all of our symposiums will be open to the press and live-streamed so that we can inform the public while we learn ourselves. We believe engaging the public in an open and transparent process is an important part of building confidence in our regulatory system. Thank you.