Too Big to Fail and the Definition of Systemic Significance

“Too big to fail” has been a useful shorthand for what is now termed “systemically important.” [1] The usage was always far from literal: it described institutions thought to require protection in support of the public interest, not those unable to fail, and size was never the only criterion (for example, Bear Stearns was not noticeably larger than Drexel, Amaranth, or Lehman). Even in the case of AIG, sheer size was less the issue than the firm’s centrality to a key derivatives market. While policymakers have traditionally avoided categorizing firms as too big to fail, the Obama administration has proposed designating certain firms as systemically important via the label Tier 1 Financial Holding Company. This raises many issues, but two particularly demand attention. How will Tier 1 institutions be defined, and is such a binary sorting mechanism optimal?

At one level, the definition of “systemically important” is fairly simple. A firm is considered systemically important if its failure would have economically significant spillover effects that could destabilize the financial system, with potential negative impact on the real economy. Unfortunately, this definition provides little guidance in practice. A practical definition will enable regulators to distinguish firms that are systemically important from those that are not, the object of which is to allow for differential regulatory treatments. Treating every firm as systemically important would burden both supervisors and the majority of firms, whose failure would not have systemic implications.

The ability to make these distinctions requires careful thought as to what makes a firm systemically important, even if some element of discretion is ultimately involved. Size may certainly be a starting point, and indeed there has been some implicit recognition of this in the designation of the 19 financial firms chosen for the supervisory capital assessment program, commonly referred to as the stress test. [2] Beyond that, four other factors encapsulate the attributes that, individually or collectively, make a financial institution systemically important. These are the four C’s of systemic importance: Contagion, Correlation, Concentration and Conditions (Context). Systemically important firms will typically have one or more of these characteristics.

Contagion : Contagion is when the losses at one institution are transmitted to other institutions through credit and counterparty risk exposures or through inter-firm connections like payments systems. The two classic cases of contagion as a source of systemic importance are Herstatt Bank in 1974 and Continental Illinois in 1984. Despite being a relatively small institution, Herstatt’s closing disrupted the international payments system and imposed nontrivial losses on the bank’s counterparties. For Continental Illinois, the stated rationale for the FDIC bailout of all creditors was the prospect of losses to some 2,300 community banks that had correspondent relationships with Continental.[3] Most recently, the justification for the Federal Reserve Bank of New York’s assisted acquisition of Bear Stearns by JPMorgan Chase would seem to be the potential for contagion, this time through the credit default swaps market.[4]

Correlation : Correlation as a source of systemic importance is also known as the too-many-to-fail problem. Penati and Protopapadakis show how correlated risk exposures contributed to the overexposure of large U.S. banks to less-developed-country (LDC) borrowers during the 1980s. [5] Two aspects of correlation risk are important. First is that there are incentives to take on correlated risks because policymakers are less likely to close an institution when many others are in similar trouble. This can lead to herding behavior, such as when financial institutions overexposed themselves to subprime mortgages, mortgage-backed securities, and derivative securities. Second is the potential for largely uncorrelated (low correlated) risk exposures to become highly correlated during periods of financial stress. Andrew Lo calls this phenomenon phase-locking behavior.[6] A group of institutions not typically considered systemically important might in certain conditions become so. This second form of correlation-driven systemic importance is actually an example of condition- or context-driven systemic importance.

Classifying institutions as systemically important because of correlated risks will require the development and estimation of risk models, the use of stress testing and scenario analysis, and a set of fundamental risk exposures that financial institution portfolios can be mapped into. Fortunately, this type of risk modeling and scenario analysis is being done in some form by large financial institutions for their own risk profile evaluation, so there is a good foundation to work from. Moreover, academic economists have begun thinking about modeling macro-financial risks in the economy, a step towards modeling and quantifying correlated-risk exposure.[7]

Concentration: A firm may be systemically important because it dominates a key market or activity. Consequently, its failure could materially disrupt a financial market or payments system that in turn impedes the functioning of broader financial markets and/or the real economy. Concentration has two important aspects: the size of the firm’s activities relative to the market and the contestability of the market. That is, concentration is less likely to result in a financial institution being systemically important if another firm can easily assume its activities.

Conditions/Context: As discussed above, during times of financial market distress, phase-locking behavior results in situations where otherwise low-correlated-risk exposures can becoming highly-correlated. As a result, institutions that would not pose a systemic threat during normal economic or financial market conditions could become systemically important. Because regulators may be reluctant to allow the official failure (and closure) of such institutions during those times, conditions (context) are a source of systemic importance. A case in point is the handling of the failure of Long Term Capital Management. Intervention by the Federal Reserve Bank of New York to prevent Long Term Capital Management from sliding into bankruptcy was driven by concerns about this hedge fund’s failure on already fragile financial markets at that time – following the Southeast Asian currency crises and the Russian default.[8] This might explain, in part, why LTCM was regarded differently than Amaranth, which was over twice the size of LTCM.[9] Moreover, context could have played a key role in the intervention to prevent the bankruptcy of Bear Stearns by merging it (with assistance) into JPMorgan Chase in early 2008, while in early 1990 Drexel Burnham Lambert entered bankruptcy.

Firms that might be systemically important due to conditions or context are likely to be the most difficult to identify ex ante. Certainly, stress testing and scenario analysis can help identify them.

The Treasury’s white paper on financial reform[10] proposes designating systemically important firms as Tier 1 financial holding companies. In light of the various ways a firm can be systemically important, moving beyond a simple binary classification system offers some useful advantages. An alternative, suggested by the Geneva report, is to create several categories based on the reasons for systemic importance.[11] Institutions may be systemic on their own, systemic as part of group, or systemic in a particular context (or state of the economy). This scheme would likely have four or five categories, and financial institutions could migrate between the categories as their activities and risks evolve. An alternative might be a three-tiered framework sorting financial firms into “noncomplex,” “moderately complex,” and “systemically important.” [12]

Sorting firms into categories applies two modern tax principles to the regulation: horizontal and vertical equity. Horizontal equity in this case means that institutions in each tier would be subject to similar regulatory treatment, even though the exact forms of regulation may not be the same. Vertical equity means that more systemically important firms would be subject to higher levels of regulatory treatments and increased supervisory attention, analogous to the prompt corrective action provisions of the Federal Deposit Insurance Corporation Improvement Act of 1991 (FDICIA).

This “progressivity” in regulatory taxes and supervisory scrutiny follows naturally from the distribution of costs and benefits. Economic efficiency dictates increasing taxes up to the point where the cost of the last increment of these taxes equals the benefits of imposing them. Economies of scale suggest that costs are more easily borne by larger, more complex firms. Furthermore, as they become more systemically important, the benefits of effective regulation increase. The difference between the private social costs of failure across firms almost by definition varies with the systemic importance of the firm. This demands a set of graduated regulatory taxes designed to internalize the externalities.

In a system with five categories, only three would contain institutions that are considered systemically important. The rationale for a five-category system is that it allows for more consistent application of regulatory taxes and supervisory oversight across categories – under the notion that differential supervision and regulation can level the playing field by mitigating any advantages systemic importance gives a financial institution.

We offer an example of how this five-tiered system might be operationalized. Under a philosophy of progressive systemic mitigation, financial institutions in category 5 would be subject to a basic level of safety and soundness regulation and supervisory oversight. [13] It would most likely consist of small community banks. Category 4 institutions would be subject to additional reporting requirements and expected to implement risk management systems and more sophisticated risk controls than category 5 institutions. Moreover, category 4 financial institutions would be subject to more vigorous supervision than their brethren in category 5.

At a minimum, category 3 institutions should be subject to periodic stress tests and be required to have contingency plans in place. Regulatory agencies would conduct routine scenario analysis and simulations to ascertain the vulnerabilities of the financial system to a correlated-risk event and establish the appropriate regulatory treatment. Such treatments might include regulatory actions like portfolio limits, add-on capital requirements, and loss reserves tied to the activities driving the correlated risks. Scenario analysis and risk simulations would also be used to develop contingency plans for handling correlated risk events. Stress tests, scenario analysis, risk simulations, and contingency plans would also be used as part of the operational regulatory system for dealing with institutions that are systemic due to conditions or context.

Category 2 and category 1 firms, the most systemically important, would face even stricter supervision and additional regulations. Category 2 institutions would report both direct and indirect interbank/interfirm exposures. They would be subject to limits on exposures to counterparties and specific reserves and capital charges designed to limit contagion across firms. Category 1 institutions would be subject to mandatory debt structure requirements, which should enhance market discipline. These could include a mandatory subordinated debt requirement, reverse convertible debentures, or both.[14]

How much information is made public about the categories, including the criteria used, the firms included, and the regulations imposed, depends on several factors. Most importantly, it depends on the extent that the supervisory regime utilizes market discipline, as opposed to direct supervision.

The choice of disclosure regime would seem to be between transparency (publication of the list of firms in each category) and some version of constructive ambiguity, where only selected information is released. The term constructive ambiguity has been attributed to former Secretary of State Henry Kissinger during the Nixon Administration.[15] In the diplomatic context, constructive ambiguity refers to the use of ambiguous statements as part of a negotiating strategy. However, in a central banking and in a financial market context, constructive ambiguity refers to a policy of using ambiguous statements to signal intent while retaining policy flexibility. In the context of the federal financial safety net, many have argued for a policy of constructive ambiguity to limit the de facto expansion of the federal financial safety net.[16] If market participants are uncertain about whether a firm is too big to fail, they will exert more risk discipline on the firm. Uncertainty about a firm’s systemic importance, or which category it falls under, may result in a higher level of market discipline than if the list were made public.

However, supervisory transparency offers a number of advantages over constructive ambiguity. First, it is unlikely that constructive ambiguity would result in sufficient ambiguity to be of much value. Markets are likely to be able to observe which firms are on the list through observed differences in capital structure, balance sheet entries (including footnotes), and the intensity of regulatory scrutiny. In addition, it would seem that any constructive ambiguity resulting from a published list would affect only a small number of firms at the margin. The advantage of avoiding the certification effect for these marginally systemic firms is likely to be swamped by losses associated with withholding information from the market. Hence, the list of firms including categories and criteria for inclusion should be made public along with a watch list of financial institutions whose status might change.

An effective system of supervisory transparency entails more than simply disclosing lists. Supervisory transparency must also include the production of information and the dissemination of this information in a form that is useful. [17] This means that all information collected, supervisory risk models used, and any results from these models to assign financial institutions a category should be disclosed. Furthermore, disclosures should include stress tests and contingency plans for financial institutions in distress.

One of the most important issues that regulators, legislators, and other policymakers need to come to grips with is systemically important financial institutions. Deciding which firms are systemically important and for what reasons is obviously a key part of their regulation. But it is only a first step. What form that regulation takes is equally important. Fighting Too Big to Fail can proceed on many levels, from regulation that reduces systemic risk, taxes that discourage firms from becoming systemically important, to procedures that once again make failure a viable option.

We envision a financial market supervisory infrastructure where systemically important financial institutions are identified, separated into categories based on nature of their systemic importance, and subject to specific regulatory treatments designed to address the systemic risk these firms impose. The ultimate objective is to promote socially compatible risk incentives for systemically important financial institutions and fairness in the financial system by leveling the playing field – by reducing or removing through regulatory taxes the advantages associated with being systemically important.

* Office of Policy Analysis, and Research Department, Federal Reserve Bank of Cleveland. The views expressed herein are those of the authors and not necessarily those of the Federal Reserve Bank of Cleveland, the Board of Governors of the Federal Reserve System, or their respective staffs.

[1] For a more comprehensive treatment of identifying and supervising systemically important financial institutions, see James B. Thomson, “On Systemically Important Financial Institutions and Progressive Systemic Mitigation,” Federal Reserve Bank of Cleveland Policy Discussion Paper 27 (August 2009).

[2] The supervisory capital assessment program was conducted from February 25, 2009, through the end of April to determine the viability and capital needs of the 19 largest financial firms in the United States. Details on the stress tests can be found at http://www.federalreserve.gov/newsevents/press/bcreg/bcreg20090424a1.pdf .

[3] Walker F. Todd and James B. Thomson (with W. Todd), “An Insider's View of the Political Economy of the Too Big to Let Fail Doctrine,” Public Budgeting and Financial Management: An International Journal 3 (1991): 547-617.

[4] Concerns about contagion were expressed by then New York Federal Reserve President Timothy Geithner in defense of the actions taken to prevent the failure of Bear Stearns. See Timothy F. Geithner, President and Chief Executive Officer, NY Fed, Testimony before the U.S. Senate Committee on Banking, Housing and Urban Affairs, Washington, D.C., April 3, 2008.

[5] See Alessandro Penati and Aris Protopapadakis, “The Effect of Implicit Deposit Insurance on Banks’ Portfolio Choices with an Application to International Overexposure,” Journal of Monetary Economics 21 (1988): 107–26. For a discussion of too many to fail, see Janet Mitchell, “Strategic Creditor Passivity, Regulation, and Bank Bailouts,” CEPR discussion paper no. 1780, 1998.

[6] See Andrew W. Lo, Hedge Funds: An Analytic Perspective (Princeton, NJ: Princeton University Press, 2008).

[7] See for example, Dale F. Gray, Robert C. Merton and Zvi Bodie,“A New Framework for Analyzing and Managing Macrofinancial Risks of an Economy,” NBER working paper no. 12637, October 2006, available at http://www.nber.org/papers/w12637

[8] See Joseph G. Haubrich, “Some Lessons on the Rescue of Long-Term Capital Management,” Federal Reserve Bank of Cleveland Policy Discussion Paper no. 19, April 2007.

[9] See House Committee on Banking and Financial Services, Testimony of Alan Greenspan at the Hearing on Hedge Fund Operations: before the House Committee on Banking and Financial Services, 105th Cong., 2d sess., October 1, 1998, serial 105–80.

[10] U.S. Department of Treasury, 2009, “Financial Regulatory Reform: a New Foundation,” available at http://www.financialstability.gov/docs/regs/FinalReport_web.pdf..

[11] Markus Brunnermeier et al., “Fundamental Principles of Financial Regulation,” Geneva Reports on the World Economy 11 (2009).

[12] Recently, Federal Reserve Bank of Cleveland President Sandra Pianalto outlined a new regulatory scheme, “tiered parity” where financial firms would be separated into three classes or tiers based upon their complexity. Regulatory treatment of a firm would be determined by the tier to which it is assigned (with equal regulatory treatment of firms within a tier). Sandra Pianalto, “Steps toward a New Financial Regulatory Architecture,” Ohio Banker’s Day address, April 1, 2009, available at http://www.clevelandfed.org/For_the_Public/News_and_Media/Speeches/2009/....

[13] These institutions would continue to be subject to consumer regulation.

[14] For a discussion of mandatory subordinated debt requirements see Rong Fan et al., “Getting the Most Out of a Mandatory Subordinated Debt Requirement,” Journal of Financial Services Research 24 (2/3): 149-179 (2003); Reverse convertible debentures are discussed in Mark J. Flannery, “No Pain, No Gain? Effecting Market Discipline via ‘Reverse Convertible Debentures,” in Capital Adequacy beyond Basel: Banking, Securities, and Insurance, ed. Hal S. Scott (Oxford: Oxford University Press, 2005).”

[15] See http://en.wikipedia.org/wiki/Constructive_ambiguity.

[16] For a discussion of constructive ambiguity as a tool to limit conjectural government guarantees of bank creditors, see Frederic S. Mishkin, “Financial Consolidation: Dangers and Opportunities,” Journal of Banking and Finance 23(2-4): 675-91 (1999). For a discussion of constructive ambiguity in the context of lender of last resort policies, see Marvin Goodfriend and Jeffery M. Lacker, “Limited Commitment and Central Bank Lending,” Federal Reserve Bank of Richmond Economic Quarterly 85(4): 1-27 (1991).

[17] Another example of providing useful information can be found in the recommendations of the 2001 Working Group on Public Disclosure, who suggest that supervisors release information that provides a consistent view of a bank’s risk management approach – such as information about risk exposures. See Board of Governors of the Federal Reserve System, Division of Banking Supervision, SR 01-6: Enhancement to Public Disclosure. April 2001.