To Fail or Not to Fail? A Discussion of Banking's "Too Big to Fail" Problem
Tom Heintjes: Welcome to another Economy Matters podcast. I'm Tom Heintjes, managing editor of the Atlanta Fed's Economy Matters magazine. Today, we're joined by Larry Wall, an economist at the Atlanta Fed and the executive director of the Atlanta Fed's Center for Financial Innovation and Stability.
Larry has spent much of his career studying the complex dynamics of the financial system, and in the course of his work he's also done a great deal of research into the concept of too big to fail, its impact on the financial system, and how to deal with the implications of too big to fail—often referred to simply as TBTF. And he's agreed to discuss this topic with us today. Larry, thanks for joining us.
Larry Wall: You're welcome.
Heintjes: Larry, let's talk a little about the background of the concept of "too big to fail." What gave rise to that particular term?
Wall: The term was first popularized with the bailout of Continental Illinois in 1984, but former Federal Reserve Bank of Minneapolis President Gary Stern and his colleague Ron Feldman wrote a book on "too big to fail" and they show the concept actually goes further back in history.
But it's important to understand what "too big to fail" means. It doesn't necessarily mean that the shareholders and bank management are going to be protected. Continental Illinois didn't go through bankruptcy, but the shareholders were wiped out and management was replaced. What did happen at Continental is that all of the creditors—both the depositors and all the other depositors—were protected from taking any losses.
Heintjes: You mention shareholders. How do they benefit from this protection?
Wall: Shareholders benefit from this protection in at least two ways: First, the creditors won't charge them as much for risks taken by the bank as they otherwise would, which results in the bank having the potential to earn higher profits.
And second, because creditors know they are likely to get bailed out, they won't be as quick to run a bank. As a result, distressed banks get extra time before facing a bank run—they can use that extra time to recover. Of course, they might also end up digging an even deeper hole if they fail to recover.
An example from the 1980s is the forbearance shown towards other large banks that suffered very large losses on their Latin American loans—the so-called "LDC debt crisis"—but these banks were not run. They were ultimately recovered, and their shareholders were able to retain control.
Heintjes: Right. Larry, the U.S. has always had very large financial institutions. Haven't we always essentially had institutions that were TBTF, even if we didn't perhaps refer to them as such?
Wall: That depends upon how far back in history one wants to go. The presumption that the government was responsible for bailing out financial crises didn't exist even as late as 1907.
Heintjes: The last big crisis before the Federal Reserve was founded.
Wall: Correct. That left the private sector to deal with financial crises. This was done through clearinghouses and—in the 1907 case, most famously—by J.P. Morgan's actions. However, the 1907 [crisis] was to be, as you said, the last crisis which was primarily a private sector responsibility. The problems that arose then helped lead to the creation of the Fed.
Indeed, it's important, I think, to understand that the Fed was not originally created to conduct monetary policy as we currently understand it. That was taken care of by the gold standard. Rather, the Fed was created to provide an elastic currency, which could be expanded during periods of bank runs, albeit the Fed was supposed to take good collateral to back its loans and hence wasn't expected to absorb losses.
The Federal government got into the loss-absorption business with the creation of the FDIC [Federal Deposit Insurance Corporation] in 1933 in response to the bank runs at the start of the Great Depression.
Heintjes: Let me ask you: Why are some financial firms treated as too big to fail?
Wall: I think the fear on the part of policymakers is that the failure of a major financial firm or firms could lead to a breakdown in the provision of financial services to the rest of the economy. This breakdown could lead to a deeper recession, or even a depression. For example, several estimates find that the collapse of the financial system after Lehman's failure [in 2008] resulted in trillions of dollars of lost GDP. Moreover, these are not just abstract numbers; many individuals and families were financially devastated by the Great Recession that followed Lehman.
Confronted with the risk of such a collapse, and typically having little time to craft a solution, policymakers often perceive the lowest cost way to avert a financial system collapse is to bail out the distressed banks.
Heintjes: Well, if TBTF policies avoid the risk of such large losses, why are so many opposed to TBTF? Is it just because such policies seem unfair to the rest of society?
Wall: TBTF policies certainly seem unfair to those who are paying the costs. The people who made the decisions that resulted in the banks getting into trouble receive the bailout, while the bill is paid by people who had no say in the decisions.
Heintjes: So, in one sense, TBTF is seen as a fairness issue.
Wall: Yes, it is a fairness issue, but it's more than just a fairness issue. TBTF also distorts competition by giving larger firms a competitive advantage. Too big to fail also distorts investment decisions by encouraging banks to fund higher-risk projects.
A common way of expressing this concern is that the banks can take the attitude of "heads, I win and my bank earns big profits; tails, the taxpayers take big losses."
But I think the problem is actually a bit more subtle, because I don't think bank managers think that way when they're making decisions most of the time. I think it's more like the dog that didn't bark in Sir Arthur Conan Doyle's The Hounds of the Baskervilles.
Bank depositors and other creditors should charge higher risk premiums to the more risky banks. But if the bank is TBTF, they aren't so much at risk so they don't charge the premiums. The people that then are running the banks observe that they can invest in higher-return assets without paying for the risks those assets generate. If the bank does so, they can boost shareholder returns and so they take advantage of the opportunity. Thus, one consequence of too big to fail policies is that it results in larger, riskier banks that would make a future crisis more likely.
Heintjes: Larry, you say that policymakers perceive the lowest-cost solution to be a bailout, but given the bad consequences of "too big to fail," should we ask if bailouts are really needed to protect the economy—and if not, maybe we would be better off by banning all bailouts?
Wall: Answering whether bailouts really are needed is difficult. As a practical matter, it's very hard to prove beyond all doubt they are needed, but it's often just as difficult to prove they are not. One needs to forecast what would have happened had the government not taken the action, such as what would have happened if the government didn't provide a bailout when actually it did provide a bailout.
The only clean-cut case we have is where no bailout is provided, and the financial system remains stable. Then it's pretty clear no bailout was needed. Otherwise, there's lots of moving parts that could determine whether the bailout was needed to preserve financial stability.
Heintjes: Right...and it's always hard to prove a hypothetical. What does, for example, the decision not to bail out Lehman Brothers show us?
Wall: Lehman, arguably, shows the cost of getting the decision wrong. There were enough moving parts so that it's possible to argue—and some do—that Lehman was not the cause of the collapse of other parts of the financial system and that allowing Lehman to fail was the right decision. But I think most policymakers around the world take the opposite view: that the failure to provide Lehman with a bailout was an enormously costly mistake.
Heintjes: And even if you think bailouts are unnecessary?
Wall: Even if you think they are unnecessary, banning them won't necessarily solve the problem because existing policymakers impose binding commitments on themselves or their successors. So as long as policymakers think there's a reasonable chance that the financial system might collapse, they'll have a strong incentive to undertake a bailout.
And moreover, market participants, when they're going to judge how big of a risk premium they need to demand from the bank, know that the policymakers cannot bind themselves. So, if the people that are supplying the money to the bank think the bank is going to get bailed out, they won't charge the appropriate risk premiums, and the large banks will have those competitive advantages and incentives to take too much risk that we had talked about earlier.
Heintjes: So, where does the solution to TBTF come from?
Wall: I would argue that the solution to too big to fail has to come from reducing the risk that the failure of a large bank would lead to the collapse of the financial system.
Heintjes: If we can't guarantee an end to future bailouts of large banks, should we seek to make banks so safe that they cannot fail, or so small that they wouldn't pose a systemic risk to the financial system?
Wall: Reducing the risk of failure is an important part of the solution, but I'm skeptical that we can go so far as to prevent all systemically important institutions from failing. Now, we might be successful in preventing banks from failing, but that would likely result in the risks going to nonbank firms and/or foreign banks, into which we have less insight but which would nevertheless remain very important to our economy.
A similar problem arises with proposals to break up the banks: They might be helpful, they might make the banks a little easier to resolve. But any bank that's large enough to be efficiently serving large institutions is likely itself to be systemically important. My view, which I expressed at a Minneapolis Federal Reserve Bank too big to fail forum, is that we need to continue the hard work of creating a safe structure for failure.
Heintjes: All of this can start to get a little depressing, when you think about it. Is there nothing we can do?
Wall: I think our best bet is to create a system in which failure is unlikely to have systemic consequences, and then rely on voters' aversion to bearing those costs.
Heintjes: You mean something like what the Dodd-Frank Act did—giving the FDIC the power to resolve failing banks?
Wall: Enhancing our ability to resolve large banks is a critical part of the puzzle, and the FDIC gained important new powers under the title of orderly liquidation authority. But it's important to understand exactly what "orderly liquidation authority" involves. The FDIC has always had the authority to resolve the large banks, the piece with the bank charter. What Dodd-Frank gave the FDIC was the power to also resolve their nonbank affiliates, and that's important for several reasons.
One of these is that the largest banking groups consist of banks, and nonbank affiliates—subsidiaries, that is—that are tightly interwoven on the operational side and so cannot be easily broken into separate pieces.
Additionally, the current bankruptcy law was written for nonfinancial firms that typically have longer-lived assets and liabilities. So, this provides the courts with more time to evaluate how to resolve these typically nonfinancial bankrupt firms. Additionally, bankruptcy law doesn't look at the effect of the bankruptcy court's decision on the overall financial system. It just is concerned with the creditors of the firm.
Heintjes: Dodd-Frank even takes further steps to help resolve banks, doesn't it?
Wall: Yes, it certainly does; Dodd-Frank requires banks to develop a so-called "living will," which will detail how the bank thinks it should be resolved, and that should be an enormous help in resolution. Now, you need to be a little careful here—no one who has given this serious thought believes that these living wills will be implemented exactly as written. Too much depends on the circumstances at the time. But they do provide an incentive for banks to make themselves easier to resolve, and they can also provide the FDIC and Federal Reserve with valuable insights into how the firm works in practice, which can be helpful in resolving it [the firm] with minimal consequences to the rest of the financial system.
Heintjes: Well, this begs the question: That sounds good, but did Dodd-Frank really solve the problem?
Wall: Unfortunately, we don't yet have a complete solution for a variety of reasons. One of the problems is that all of the large bank groups operate across national boundaries, and the non-U.S. parts of these groups are going to be subject to actions taken by the host country supervisors. And of course, these foreign supervisors might reasonably take actions to protect their own citizens, even if this makes it more difficult to resolve the overall entity. The second problem is that resolving one bank is a lot easier than resolving multiple large banks at almost the same time.
And a third reason—and this a problem that's written into Dodd-Frank—is that it requires the risk of financial instability be significant before early liquidation procedures can be implemented. This means we would be trying out a new procedure at a time when unexpected consequences could have severe implications.
Heintjes: Wow! Well, are these the only problems with Dodd-Frank?
Wall: There have been some other concerns expressed. One of these is that it created a new set of too big to fail firms. That is, Dodd-Frank said that the supervisors are supposed to designate certain firms as being too big to fail, and they argue that once a firm has been designated that it will be considered as systemically important and the government will be pretty much forced to use orderly liquidation authority. These people would abolish orderly liquidation and replace it with a new bankruptcy code, specifically designed for financial firms. That may be the case, but I tend to view this as a second-order problem—one that's smaller than the gains from orderly liquidation.
Heintjes: And what's your take on that concern?
Wall: I don't agree with the concern for several reasons: First, designation is primarily about identifying nonbank firms that should be subject to stricter prudential regulation. It's not a guarantee that the firms will go through orderly liquidation. That decision must be made on a case-by-case basis involving the Treasury, the Fed, and another federal agency—for banks, it would be the FDIC—based on the threat failure would pose given the conditions at that time.
Second, Dodd-Frank doesn't give the FDIC a resolution fund for nonbank firms, but rather requires that it first be paid out of the failed institution's assets. There are provisions for the FDIC to borrow money to provide liquidity, but this isn't the same as having a fund readily available to bail out all the creditors.
The third reason is that if policymakers are sufficiently concerned that a nonbank firm's failure could cause a systemic crisis, then they'll likely find a way to provide a bailout whether or not the FDIC retains its orderly liquidation authority.
Heintjes: So, what are we really talking about here?
Wall: We're really talking about the increased risk that in marginal cases a designated firm will fail and be put under orderly liquidation, and the FDIC will abuse its borrowing authority to provide funds for the failing firm, which cannot be repaid from that firm's assets. Each of the steps in these chains is possible, but I think that risk pales compared with the additional gain from giving the FDIC orderly liquidation authority.
Heintjes: Larry, what can be done about the problems you've identified with resolution?
Wall: The FDIC is working with foreign supervisors to develop agreements on how to handle the foreign subsidiaries of U.S. banks and the U.S. subsidiaries of foreign banks. This is important work that needs to continue. Indeed, these discussions probably need to be a permanent part of the international dialogue among bank supervisors.
Paradoxically, while I disagree with those advocating the repeal of orderly liquidation authority, I strongly agree with their proposal to create a new bankruptcy chapter for financial firms. Ideally, this procedure would be close to the one planned by the FDIC but of course would remain under the control of the bankruptcy court rather than the FDIC.
Heintjes: Why would you support that approach?
Wall: One reason for supporting it is, it would reduce the incentive to invoke orderly resolution and get the FDIC involved. But more importantly, hopefully the first time this new chapter could be invoked is when the financial system is not on the brink of systemic crisis, so that we could get a chance to get a better understanding of how this new procedure would work in practice without having the risk of making a bad situation even worse.
Heintjes: We've talked about a lot of the problems and a lot of the challenging dynamics at work here. Do you think we're on our way to solving the too big to fail problem after all?
Wall: We're on our way, but we still face one major obstacle that really hasn't been addressed. It seems that I've saved the really bad news for last.
Heintjes: Oh, goodie!
Wall: To me the deeper problem is summarized by a quote form Shakespeare's Hamlet: "When sorrows come, they come not as single spies but in battalions."
Heintjes:: [laughter] Well said!
Wall : It's relatively rare for large banks to get into serious trouble while the rest of the financial system is doing well. Rather, if one big bank is in trouble, it's likely the other large ones are also in trouble due to exposures to a common risk factor. In the 1980s, it was Latin American loans. Earlier this century, it was to residential mortgages.
Heintjes: So, if you are able to sum it up: What do we need?
Wall: We either need a credible way to resolve multiple important financial institutions at the same time, or we need to find a way to prevent them from having such large exposures to a common risk factor. Arguably, through the stress test we made a bit of progress in requiring these large exposures to common risk factors to be met by more capital so failure is less likely. But I'm not aware of anyone that's articulated a feasible strategy for resolving multiple important financial institutions at the same time without risking systemic consequences.
Heintjes: Larry, thanks for spending time with us today. This has been a really great conversation.
Wall: You're welcome, Tom. I've enjoyed it very much.
Heintjes: I want to note on our website, frbatlanta.org/economy-matters, we'll have a link to Larry's research into too big to fail and the financial system in general. I encourage you to visit to get a fuller idea of what we've talked about today.
And that brings us to the end of another Economy Matters podcast episode. Join us again next month when we'll talk to Atlanta Fed President Dennis Lockhart, who is preparing to step down as president of the Atlanta Fed after 10 very eventful years. He'll look back at his tenure here and share some of the insights he's gained along the way. It's going to be a very interesting conversation, and I hope you'll join us for it. Thanks for listening.