Managing Too Big To Fail by Reducing Systemic Risk: Some Recent Developments
Recent developments include three important initiatives.
Published June 1, 2006 | June 2006 issue
Central banks and like-minded entities spend considerable time and resources trying to reduce systemic risk, the chance that disruptions in one financial market or failure of a large, important financial institution will spill over to other institutions. In contrast to the Federal Reserve's very public efforts to fight inflation, for example, these risk-mitigation efforts typically go unnoticed by the public at large. We believe, however, that there is substantial value in publicly explaining the steps that are taken to mitigate systemic risk; indeed, we think that public knowledge of such efforts can itself reduce the chance of spillovers. In this essay, therefore, we describe three important risk-reducing initiatives.1
Now, there is good reason why systemic risk mitigation is not apparent. Typically, these efforts yield highly technical changes in rules and institutions that occur over a long period of time and thus do not lend themselves to regular and frequent action by central banks. It might sound like a nice idea, but having central banks vote every six weeks to reduce systemic risk would not be very meaningful.
So where is the gain in notifying the public of progress? Consider the case of systemic risk caused by large bank failure. When confronted with the failure of a very large bank whose insolvency could spill over to other banks and markets, policymakers face the temptation to bail out the bank's creditors. This temptation will probably grow, we believe, as banks continue to consolidate, and as the organization and activities of these banks become more complex.
But while bailouts may halt a crisis, they also have a serious downside, as Federal Reserve Governor Donald Kohn recently made clear:
Although policy action may be able to reduce the odds of adverse effects or alleviate their impact, some policy responses to a crisis can themselves have important costs that need to be balanced against their possible benefits. In short, intervening in the market process can create moral hazard and weaken market discipline. If private parties come to believe in the possibility of policy actions that will relieve them of some of the costs of poor decisions or even just bad luck, their incentives to appropriately weigh risks in the future will be reduced and discipline on managers watered down. Weaker market discipline distorts resource allocation and can sow the seeds of a future crisis.2
Putting bank creditors at risk today—that is, convincing creditors that their bank is not too big to fail (TBTF)—is the key step to avoiding the problem pinpointed by Kohn, and to preventing unnecessary waste of society's resources.3 And here is where broader public knowledge of risk-mitigation efforts yields large benefits. Since policymakers bail creditors out to avoid systemic risk, the most persuasive way to convince creditors that they truly are at risk of loss is for policymakers to take steps to reduce systemic risk and to communicate those steps to the public.
In this essay, we describe three such advances. First, we discuss a proposal from the Federal Deposit Insurance Corp. (FDIC) to improve its ability to limit taxpayer bailouts when a large bank fails. We devote much of this essay to explaining our general support for the FDIC's proposal. This initiative cuts to the core of the government's ability to limit bailouts. Second, we report on the so-called NewBank reform to create a backup for certain banks that played a heretofore irreplaceable role in key financial markets. Finally, we comment on a private-sector initiative to better equip financial firms to understand and manage the risks posed by modern financial transactions. We see all three efforts leading to better management of systemic risk, reducing creditors' expectations of TBTF bailouts and thus making society better off.
Getting a grip on uninsured deposits: The FDIC's effort4
In December 2005, the FDIC requested early feedback on potential changes to the way it determines which deposits are insured at a failed large bank. Simply put, the FDIC is concerned that due to technical limitations in its ability to identify and isolate insured deposits at large failing banks, it may have to support uninsured bank creditors. Given the costs of such bailouts to society as a whole, we believe that serious effort should be devoted to finding a cost-effective way of preventing such an outcome.
The problem. In simple terms, when a bank fails, the FDIC has responsibility for responding to the claims made against it by those owed money by the bank. In that process, bank depositors who are insured receive the full value of their claim. All other creditors receive a value determined by the collateral they held and/or how much the FDIC can sell the bank for, either in whole or parts. (For example, certain bank assets such as loans may be sold separately.)
In carrying out this "resolution" process, the FDIC must settle claims in a manner that minimizes government costs. In practice, this means the FDIC will not use government funds to make good on deposits, for example, that exceed the insured limit or are otherwise not covered by deposit insurance. The law does, however, give the FDIC an out. If extraordinary government protection would prevent a systemic crisis, bailouts are allowed. Several government agencies, including the Federal Reserve Board, the FDIC and the Treasury, are involved in making this determination of a "systemic risk exception."
In its proposal, the FDIC identified a serious limitation in its current ability, under certain circumstances, to resolve a large institution in the least-cost manner. The problem facing the FDIC results from the interaction of two distinct issues—isolating insured from uninsured deposits and the use of "bridge banks" during the resolution process.
First, the FDIC faces serious technical challenges in determining which deposits at large banks are insured and which are not. This difficulty arises from the absence of two variables in banks' deposit tracking systems. Banks may not have a single, unique identifier for all deposit accounts associated with a particular depositor (making it difficult to determine if those deposits are covered by the FDIC). Banks also do not have systems that readily allow the FDIC to determine the insurance category in which each deposit account falls (making it difficult to tell how much of the account under review is insured). Even more serious, perhaps, bank systems can't segregate the insured from the uninsured.
These problems are most acute at large banks with multiple deposit-tracking systems that do not necessarily "talk to each other," a category that may include many systemically important institutions. To be clear, noting these limitations is not to criticize banks. While there are social and policy reasons for wanting banks' systems to have these features, banks have little if any business rationale to implement them.
The second issue relates to the specific tool the FDIC expects to use when confronting a large bank failure. Specifically, the FDIC intends to resolve a large bank by creating a temporary bank—appropriately called a "bridge bank"—and transferring certain of the failed bank's assets and liabilities to the bridge after the bank fails.5 The FDIC would keep the bridge bank open until it can sell the entity or its parts in a manner that minimizes cost to the government. Importantly, those liabilities that get moved from the failed bank to the bridge bank, as a rule, would be protected while those not moved would suffer losses.
Why use a bridge bank? A bridge bank is a potentially effective tool for minimizing the chance of spillover failures—and government bailouts to contain spillovers—by maintaining the operations of the failed bank and providing depositors with access to all or most of their funds. Bridge banks also allow the FDIC to maximize the value obtained from the failed bank's assets and operations by keeping them running. In short, using a bridge bank allows the FDIC to keep the operations of the failed bank up and running without providing assistance to a wide array of creditors.
Difficult interaction. Unfortunately, the interaction of these two factors—difficulty in determining the insurance status of deposits and the use of bridge banks—creates a problem for policymakers. Consider the FDIC's potential inability to segregate insured and uninsured deposits. Recall that the FDIC does not want to transfer deposits or other liabilities to a bridge bank if it does not plan to protect them. Unfortunately, some large banks' deposit-tracking systems would not facilitate the FDIC's need to put "holds" on uninsured deposits transferred from a failed bank to a bridge bank. Absent such holds, depositors would have free access to uninsured deposits transferred to the bridge bank.6 What should have been uninsured deposits subject to losses could then end up being protected.
Now, the response to these problems may seem straightforward. The FDIC could restrict depositors' access to their funds at the bridge bank in order to ensure that only insured deposits receive government backing. In the extreme, without a mechanism to place automatic holds on uninsured deposits, the FDIC could decide to release deposits to depositors only after manual review of their accounts. Given the time a manual review would consume, such an approach could lead to the lapse of significant time between bank closure and bridge bank opening.
Because of this delay, imposing tight restrictions on access to depositor funds is neither attractive nor particularly practical. After all, the whole point of using the bridge bank was to get the failed bank back in operation as quickly as possible with depositors having access to their cash. In the face of insufficient technology to segregate deposits or inadequate information to determine the insurance status of deposits, therefore, the FDIC would likely prefer to provide depositors with access to deposits even if they might be uninsured. This preference, even if understandable, undercuts least-cost resolution and puts pressure on policymakers to invoke the systemic risk exception.
And herein lies the problem. Invoking the systemic risk exception should be, indeed, an exceptional event, one that occurs only to prevent a true systemic crisis. The last thing policymakers should do is invoke it simply because of limitations in the resolution process. Such a decision would signal uninsured creditors of large banks that the FDIC cannot easily stick them with losses when called for. Such uninsured creditors would then have less reason to monitor their banks, the unmonitored banks would have greater incentive to take on risk and the economy would suffer. A vicious cycle would be in full swing.
Potential solutions. The FDIC offered several early draft reforms that directly address the root causes of the problem. Again, to simplify, these reforms would, to differing extents, require banks to modify their information systems to segregate insured liabilities from uninsured liabilities on an ongoing basis, and better identify which liabilities are insured. These reforms would be limited to the largest banks, potentially covering on the order of the 150 largest (although the FDIC set no expectation that reforms would have to cover this maximum number or that each bank in this group would face a similar reform process).
Responses to the proposed solution. Comments on the proposal from banks and their representatives were almost entirely unsupportive of the proposals. Their major concern was the potential cost of changing information systems. Many banks objected to applying the reforms to banks in good condition given the very remote chance of failure among large banks. Several comments suggested that the FDIC target the reforms at banks in weaker condition, but before failure. Indeed, because large banks fail so rarely, some commentators doubted that the reforms offered much benefit at all.
Banks are well positioned to help quantify the costs of the proposed reform, a key input for policymakers. Moreover, we agree that the FDIC should be flexible in developing options to address its concerns. To the degree policymakers offer any alternatives, they should focus on effective, lower-cost suggestions provided by banks. If additional conversations between the FDIC and the banking industry can facilitate such reforms, they should proceed as soon as possible.
But it would be unwise to reject the FDIC's proposals and retain the status quo solely because of the cost of implementing changes; benefits matter as well. In fact, in their comments on the proposal, several banking trade associations made clear why more of the same is not acceptable. They noted that:
Each association recognizes that the Federal deposit insurance system's viability depends on the principle that no financial institution is either too big or too small to fail. The development of prudent systems to prepare for and respond to the failure of any size institution is an important component of the Corporation's [FDIC's] receivership functions. Further, the Corporation should demonstrate that it is fully prepared to handle even a large bank failure, quickly and efficiently.7
The FDIC's analysis of technical limitations of the current resolution process raises serious doubt that creditors of large banks will be as likely to suffer losses as creditors of small banks upon bank failure. Any reader of the FDIC's proposal should also be concerned about its current ability to respond effectively and quickly to a large bank failure. In other words, absent some reform, the conditions for credibly imposing a least-cost resolution on a large bank across several plausible scenarios do not exist.
The environment just described fosters the TBTF problem; in such a climate, a creditor of a large bank can rationally expect to benefit from at least some government protection and can therefore reduce its monitoring (or "market discipline") of banks. We have already noted that reduced market discipline—and related increased bank risk—leads to general economic loss, as resources are wasted. (It is important to emphasize that such economic losses occur whether or not a bank actually fails. What changes creditors' behavior toward their bank's risk-taking is their rational belief that their deposits are protected.) Avoiding these economic losses offers potentially large benefits that the FDIC board of directors, acting on behalf of taxpayers and society as a whole, should surely consider. (Also worth emphasizing is the need for credibility of any reform. If banks oppose these reforms when they are in strong financial condition, would the FDIC really be able to impose them when banks are financially troubled, shedding staff and focusing on recovery?)
Backing up critical payment providers
It is well known that the U.S. government issues lots of debt. Less well known is the fact that only two banks—the Bank of New York and J.P. Morgan Chase—offer major market participants the full set of services they need to "settle" transactions involving U.S. Treasury securities (that is, have the proper exchange of cash for securities). Serious spillovers could result if either of these banks were suddenly unable to offer its settlement services for a sustained period. The U.S. government could have a challenge funding itself. The Federal Reserve's tool for implementing monetary policy could be impaired. And markets for fixed-income securities could be seriously strained. Faced with such systemic risk, policymakers would have incentive to do whatever it takes to keep the two banks up and running, and providing these services.
To avoid such a meltdown and after-the-fact government intervention, a private-sector group—with involvement from staff of government financial regulators, including the Federal Reserve—has recently continued the development of a plan to back up the two clearing banks.8 The plan involves creation of an entity called NewBank that has the legal, technical and financial infrastructure necessary to quickly take over the clearing business of an existing clearing bank. Absent its activation, NewBank would stay in a form of hibernation.
As the private-sector group's report makes clear, getting from "NewBank the plan" to "NewBank the reality" faces real challenges and will require significant work. Substantial time and effort went into the creation and endorsement of the NewBank plan; reports released in 2002 laid the groundwork for the current effort. But the plan and its public-sector and private-sector partnership offer tangible evidence that seemingly insurmountable causes of TBTF can, in fact, be managed.
Limiting the systemic risk of new, complex financial transactions
After the collapse and rescue of the hedge fund Long-Term Capital Management and other market disruptions, a private-sector group called the Counterparty Risk Management Policy Group (Policy Group) was formed to assess key features of risk management at large financial firms. E. Gerald Corrigan, a senior official with the securities firm Goldman Sachs and former president of the Federal Reserve Banks of New York and Minneapolis, chaired the group, which included other senior officials at financial firms. In 1999, the Policy Group issued a report with a number of recommendations on how the firms could better understand, track and manage the risk posed in financial transactions.
In 2005, a second Policy Group report reviewed progress on the original recommendations, made new recommendations and analyzed developments in financial markets that could alter systemic risk.9 The Policy Group reported significant progress on the original recommendations, but also highlighted where additional progress is needed and urged joint public/private remedial action in new areas of concern.
Progress has already been made on some of the new concerns. For example, the 2005 report was critical of certain practices in the market in which firms buy and sell insurance against the default and/or failure of a firm (the "credit default swap" market). A consortium of financial firms-observed with keen interest by the New York Federal Reserve Bank and other supervisors—has begun to address those concerns. 10
While the recommendations focused on specific practices that could benefit private firms, we read the report as an argument that financial instability arises from spillovers and that such financial instability can harm even sophisticated financial firms. Private firms benefit from a reduction in systemic risk, this perspective suggests, and can take actions to reduce it. If a firm effectively manages its exposure to movements in financial markets and instruments, and to changes in the credit quality of counterparties, the chances for instability fall materially. To the degree that the Policy Group's recommendations reduce systemic risk, they create the base on which policymakers can more confidently impose losses on creditors of large banks and cut to the heart of TBTF expectations.
More generally, the second report made clear that the battle against systemic risk is an ongoing one. The Policy Group suggested that a risk management cycle characterizes the behavior of financial firms.
- Over time, the firm recognizes that staff and management are not adequately applying the basics of risk management to their activities.
- Then the firm takes action to ensure that staff and management apply the basics (perhaps before a loss, but certainly after one).
- But the firm falls behind the cycle as it moves into new activities in which the basics have yet to be applied.
- Once again, the firm takes action to apply the basics (hopefully, before a loss).
Given this description, the ongoing progress made by the Policy Group and related efforts are critical to addressing the TBTF problem.
Only on very rare occasions does the failure of one bank or the drop in the value of assets in one financial market lead other banks to fail or other markets to fall precipitously. But the chance is not zero, and the costs from such systemic risk and spillovers can be quite high. Unprepared policymakers can make the costs even higher in the long run by responding with excessive government support. Indeed, if bank creditors come to expect such support, they will alter their behavior today by providing less discipline to banks than they would otherwise, which can lead banks to take on too much risk.
Taking steps to limit the potential fallout from systemic risk, therefore, has its own public return but also gives creditors the incentive to properly discipline what had been TBTF banks.11 This article has focused on three recent examples in which central banks, related government agencies and private-sector initiatives have made progress against systemic risk. Such efforts indicate that policymakers, bankers and others in the financial industry are both concerned about the potentially devastating losses that systemic failures pose and convinced of the importance of taking steps to prevent them. Continued efforts to mitigate risk must remain a central bank priority. And ensuring that the public understands these initiatives should be a goal of the central bank.
1 Others have also discussed the role of the central bank in managing systemic risk. For an excellent recent example that also supports disclosure of and discusses progress made against systemic risk, see Roger W. Ferguson Jr.'s "Thoughts on Financial Stability and Central Banking," remarks at the Conference on Modern Financial Institutions, Financial Markets, and Systemic Risk, Federal Reserve Bank of Atlanta, April 17, 2006.
2 Donald L. Kohn. "Crisis Management: The Known, the Unknown, and the Unknowable." Remarks at the Wharton/Sloan/Mercer Oliver Wyman Institute Conference, "Financial Risk Management in Practice," Philadelphia, Pa., Jan. 6, 2005.
3 We discuss TBTF in detail in our book: Stern, Gary H., and Ron J. Feldman. 2004. Too Big to Fail: The Hazards of Bank Bailouts. Washington, D.C.: Brookings Institution Press.
4 This section draws heavily on Gary H. Stern's comments on the FDIC's Advance Notice of Proposed Rulemaking for a "Large-Bank Deposit Insurance Determination Modernization Proposal," Feb. 7, 2006.
5 Murton, Arthur J. 2005. "Resolving a Large Bank: The FDIC's Perspective." In Douglas D. Evanoff and George G. Kaufman, eds., Systemic Financial Crises: Resolving Large Bank Insolvencies. Singapore: World Scientific, pp. 415-20.
6 The FDIC would not have to apply holds on deposits to the degree that other creditors absorb all resolution losses. While not implausible, history suggests that large bank depositors will likely face some loss-albeit a potentially smaller one relative to that experienced by depositors of failed small banks-after resolution.
7 Mark J. Tenhundfeld, Robert Davis and Rich Whiting. Comments on the FDIC's Advance Notice of Proposed Rulemaking for a "Large-Bank Deposit Insurance Determination Modernization Proposal," March 13, 2006, p. 3. [PDF]
9 "Toward Greater Financial Stability: A Private Sector Perspective." Report of the Counterparty Risk Management Policy Group II,
July 27, 2005. [PDF]
10 New York Fed Welcomes New Industry Commitments on Credit Derivatives. Press release, March 13, 2006.
11 But concern about TBTF need not drive concerns about systemic risk. As Mishkin notes, "Even if the too-big-to-fail problem is not as serious as [Stern and Feldman] contend, the policies they outline can make it less likely that a banking crisis will occur even if driven by other factors." Frederic S. Mishkin. 2005. "How Big a Problem Is Too Big to Fail?" Abstract. Working Paper 11814. National Bureau of Economic Research.