Certain financial institutions are so central to the American financial system that their failure could cause traumatic damage, both to financial markets and the larger economy. These institutions are often referred to as “systemically important financial institutions” or SIFIs. The Dodd-Frank Act, the comprehensive reform legislation signed into law during the summer of 2010, requires financial regulators belonging to the Financial Stability Oversight Council (FSOC) to name those financial institutions that it believes are systemically important. Such SIFIs are to be supervised more closely and potentially required to operate with greater safety margins, such as higher levels of capital, and to face further limitations on their activities.
Throughout Dodd-Frank the focus is principally on banks, particularly commercial banks, and the act effectively designates all commercial banking groups with $50 billion or more in assets as SIFIs. However, it requires regulators to consider whether other financial institutions are systemically important, leaving the decision about which non-bank financial institutions should receive that designation up to the FSOC, with advice from the Federal Reserve Board (Fed). The FSOC is in the process of determining what non-bank institutions it will designate as SIFIs, but it seems clear that several large life insurance groups and at least one large finance company (GE Capital) will be named. Eight “financial market utilities” have already been designated. (These are firms such as clearing houses that do the back office transactions that make many financial markets function.) Other financial institutions may be added as well, such as hedge funds or money market funds.
Dodd-Frank also authorizes the FSOC to designate certain types of activities as systemic regardless of what institution is conducting them, giving the regulators greater powers to control those activities. There is some potential for this to be invoked in regard to money market funds and that possibility has given the FSOC greater leverage in pushing for changes to the rules governing money market funds even if the systemic activities designation is never used. This paper will generally not discuss the activities clause, but will focus instead on the regulation of entire institutions designated as SIFIs.
Once a non-bank financial institution has been designated as a SIFI, very real questions arise as to how best to regulate these institutions. The Fed becomes the regulator for SIFI purposes, alongside the existing primary regulator. However, the Fed has little previous experience of overseeing some of these types of institutions, particularly insurers. Therefore, it needs to figure out how to evaluate their safety and how to coordinate with existing supervisors. Doubtless, the Fed will end up falling somewhere on a spectrum between simple reliance on existing regulatory paradigms and procedures and developing an entirely separate approach that may rely excessively on its prior experience as a banking supervisor.
The Fed should not simply defer to existing regulators and view non-bank SIFIs as safe if they say so. It has a legal obligation to form its own conclusions. Further, viewing the institutions systemically may provide a different perspective, perhaps pointing to systemic risks that would not be given adequate attention by traditional industry regulators who are not responsible for the safety of the financial system across the country or concerned about linkages to the rest of the world. This could be particularly true in insurance, which is regulated at the state level and therefore has not historically had any body whose primary responsibility was to look at national systemic risks. The National Association of Insurance Commissioners (NAIC) acts as a coordinator for the state insurance commissioners and works to ensure high standards across the country. However, these standards are aimed at ensuring the safety of individual institutions with little emphasis on the linkages between these institutions that could lead to systemic problems.
On the other hand, there is a real risk that the Fed will give insufficient deference to the extensive experience and knowledge residing with the existing regulators, particularly in regard to insurance, which has so many differences from banking. Decision-makers at the Fed would be only human if they relied excessively on the tools with which they were already familiar and if they were more comfortable starting from scratch in designing regulation and supervisory tools, instead of relying on the experience of others.
There are multiple dangers in taking an idiosyncratic Fed perspective that pays too little attention to existing regulatory approaches:
The Fed may simply get a decision wrong, out of an insufficient level of understanding of the new industry. It is one thing to study an industry intensively, it is another to have lived with it for many years, as the primary regulators have.
The Fed could be “right” from the point of view of reducing systemic risk, but the economic cost of eliminating or reducing a particular source of risk may far exceed the benefit. Dodd-Frank did not call for the elimination of systemic risk, but rather appropriate control over it. As with so many areas of life, absolute elimination of risk would require forbidding a great deal of beneficial activity. The bureaucratic peril here is that the Fed’s mandate from Dodd-Frank may bias the organization towards elimination or sharp reduction of systemic risk, with insufficient regard to the economic costs that would show up in day-to-day operations.
New Fed regulations could effectively force “relitigation” of a myriad of issues that have already been decided by the primary regulators. Sometimes there are multiple legitimate ways to approach an issue and it may be better to stay with the existing decision than to go through the industry upheaval of adopting to a new approach that simply has a different set of pros and cons, but may not be substantially better.
Lack of sufficient coordination with existing regulators could result in contradictory requirements that hamper operations. The Fed and the primary regulators will presumably manage to avoid outright contradictions, although there is definitely the possibility of temporary stand-offs as the two sides feel their way to a working arrangement. Beyond that, though, there is the risk that the approach of the Fed and of the primary regulators will be incompatible in practice, even if this is not obvious on the surface of the written regulations. One side or the other may believe it is possible to meet their requirements without infringing the rules issued by the other, but it may not in fact be feasible.
Pointing out these dangers of inappropriate regulation is not intended to argue against the designation of non-bank SIFIs, which I do favor and which is clearly the intent of Dodd-Frank. There are legitimately differing views on whether insurers, for example, are ever systemically significant, but I am among those who believe that a few very large life insurance groups likely do merit this designation. The key message of this paper, however, is that non-banks are not just funny looking banks, but operate in truly different industries, providing different services, and facing a different balance of risks and opportunities than do banks. Therefore it is very important that Fed regulation of non-bank SIFIs is tailored to each distinct industry and is managed with appropriate humility about the Fed’s level of understanding and with appropriate deference to primary regulators, while meeting the Fed’s obligations to develop their own independent judgments. This is a difficult balancing act, but not fundamentally different than the balancing acts that all regulators face between the risks of action and inaction. The bulk of this paper delves deeper into these issues in the context of life insurers.
The Fed is most definitely aware of the dangers and is intent on avoiding them. However, it is virtually certain that mistakes will be made in an area of this complexity where there are at least two sets of perspectives and experiences coming together, especially given the novel nature of the task of regulating systemic risk. One concerning point is that there is not a clear agreement yet on what systemic risk is and how it ought to be measured, adding still more uncertainty about how best to regulate it.
There is some disagreement about the best definition of systemic risk. A report by the International Monetary Fund and two global financial regulatory bodies defined systemic risk as:
“a risk of disruption to financial services that is (i) caused by an impairment of all or parts of the financial system and (ii) has the potential to have serious negative consequences for the real economy. Fundamental to the definition is the notion of negative externalities from a disruption or failure of a financial institution, market or instrument. All types of financial intermediaries, markets and infrastructure can potentially be systemically important to some degree.
Three key criteria that are helpful in identifying the systemic importance of markets and institutions are: size (the volume of financial services provided by the individual component of the financial system), substitutability (the extent to which other components of the system can provide the same services in the event of a failure) and interconnectedness (linkages with other components of the system).”
Dodd-Frank defines systemic risk in terms of a situation in which “material financial distress at the [financial institution], or the nature, scope, size, scale, concentration, interconnectedness, or mix of the activities of the [financial institution], could pose a threat to the financial stability of the United States.”
There is substantially more disagreement about how to measure the level of systemic risk in the aggregate. Breaking this down to the contribution from individual institutions is yet trickier still. As a further important complication, systemic risk arguably varies over time. An entity could be systemically significant under some circumstances and not others.
The FSOC’s evaluation process to decide which institutions to designate as SIFIs relies heavily on subjective judgments of the relative importance and inter-relationships of the relevant qualitative and quantitative factors. This is not a criticism. Objective, quantitative criteria will require both a detailed analytical model of how the financial system works that is well beyond the current state of research and considerably more and better quality data than currently exists. Many academics and official researchers are working to create those prerequisites, but it will be years before they can hope to succeed, if they ever fully do.
There are multiple ways in which a financial institution can be systemically important – by its size, the degree to which to which it is “interconnected” with other parties, or conceivably by its reputation and thus influence on financial markets. The central concern is that a SIFI’s failure would cause serious damage to the financial system, and thereby to the rest of the economy. The sources of that damage could be any one or more of the following, and perhaps others as well:
Counterparty and other credit risks. One of the most obvious concerns is that when a SIFI goes under it may impose substantial, if not crippling, losses on other financial institutions and parties who are owed money by the institution. This could cascade throughout the financial system with knock-on damage to the wider economy.
Contagion. Sometimes the principal damage from the collapse of a financial institution comes from serving as a “bad example” that causes the market to reassess which other organizations might wind up in the same difficulties.
Problems with deposit-taking activities. One of the key reasons that banks are regulated so highly in the first place is that consumers and businesses place deposits with them which they count upon to be readily available and riskless. There can be severe economic disruptions if depositors find their funds suddenly unavailable.
Maturity mismatches. Financial institutions often operate by “borrowing short and lending long”, since the interest rates on short-term borrowings are typically below the interest rates earned on longer-term loans and other assets. This strategy usually is exposed to the risk of a sudden liquidity freeze that makes it highly expensive or impossible to “roll over” short-term liabilities. Excessive maturity mismatches become a systemic problem if they are too widespread or concentrated at one or more SIFIs.
Market utility interruptions. Some institutions play a central role in the day-to-day functioning of financial markets, resulting in the potential for widespread damage if they fail.
Types of non-bank SIFIs
There are several major categories of non-banks that could be systemically important; the considerations that could lead to their designation are discussed briefly below. (A fuller review of the issues is available in the paper I wrote with Robert Litan, referenced in footnote 1, which focuses more on the issues surrounding designation of SIFIs.) The discussion excludes banks of all types and their close affiliates, which are effectively already designated as SIFI’s under Dodd-Frank.
Finance companies. Until the recent crisis, there were a number of major lenders to consumers and small businesses that financed themselves by issuing short to intermediate term debt in the wholesale financial markets, in contrast to commercial banks that raise their funds primarily with insured deposits. When financial markets froze, this finance company business model proved to be too risky, except in special circumstances, since it exposed the firms to the danger that they would be unable to “roll over” their debts. Borrowing short-term and lending long-term only works if the ability to borrow short-term is not interrupted for any extended period. The recent crisis showed once again that such liquidity freezes occur too frequently to be assumed away.
Smaller finance companies may not pose a systemic risk if they fail, since in a crisis the markets may still be willing to fund their larger competitors. However, when large finance companies are threatened with failure, they may indeed pose systemic risks. Because of the risks of the finance company business model that were revealed in the recent crisis, a number of the solvent finance companies that have survived have converted to bank status in order to have access to insured deposits even in difficult economic conditions.
Securities firms. Investment banks and brokerages can clearly create risks to the financial system, as demonstrated by Bear Stearns, Lehman, Merrill Lynch, and others in the recent financial crisis. However, the most important of these firms are affiliated with commercial banks and are therefore already considered SIFIs for that reason. It appears unlikely that any of the stand-alone securities firms based in the US will be designated as SIFIs, but one or more could expand over time to the point where they might be designated in the future. It is also possible that a large US subsidiary of a foreign securities firm could be designated as a SIFI.
Life insurers. Some life insurance entities are so large that their sheer size makes them obvious candidates for designation since other financial institutions will have major credit exposures to them. On the other hand, the types of activities they undertake tend not to be as risky for the system, especially since they are generally funded by quite long-term liabilities, such as life insurance policies and annuities that have substantial fees for early surrender. In general, the systemic risk created by a life insurer is likely to be considerably less per dollar of asset size than would be true for a bank, taking into account probabilities rather than just worst cases. However, each case must be examined on its own merits and regulators must watch out for the development of activities at one or more life insurance groups that might spawn greater systemic risk in the future. Life reinsurers, which provide wholesale insurance protection to life insurers, have greater risk per dollar of assets because they are interconnected with many other insurers and reinsurers. However, none of the US-based life reinsurers are of sufficient scale to be likely to be designated as SIFIs.
Property/casualty insurers. Insurers providing protection against accidents and lawsuits are important financial institutions and sometimes very large. However, the nature of traditional property/casualty insurance creates little risk for the financial system as a whole. The investments of these firms tend to be very conservative and liquid, since they could be needed quickly in the event of a natural catastrophe. As a result, the big risks to these insurers are on the claims side, which has little correlation with financial crises. (Financial crises do not spawn natural disasters and even extremely large hurricanes and earthquakes are too small to trigger a financial crisis.) There is no indication that any property/casualty insurers will be designated as a SIFI, with the exception of AIG. That firm will be designated for political and historical reasons more than anything else, although the stated rationale will doubtless refer to its life insurance business and activities outside of traditional insurance.
Hedge funds. These funds cover a very wide range of activities, most of which would not warrant SIFI designation. If any do, it would almost certainly be because they operated with quite significant amounts of financial leverage and were of considerable size (as was LTCM in the late 1990s before the Fed helped arrange a private sector reorganization). The combination of size and leverage could generate sufficiently large credit exposures for other SIFIs to merit inclusion of these funds or they might exacerbate other potential sources of risk, including contagion.
Other fund models. Two other important fund business models are venture capital (VC) and private equity (PE) funds. Neither would appear to create any significant systemic risk when they are run in a traditional manner. However, the legal structure could be used to operate more like a highly leveraged hedge fund, in which case there is at least the theoretical possibility of being a SIFI. In practice, it is unlikely that the FSOC will designate any of these funds as SIFIs for some years, if ever.
Mutual funds. These fund groups are an interesting case, since some of them are of very large size, yet they are essentially pass-through entities and seldom use very much in the way of leverage. The small amount of leverage employed means correspondingly less credit exposure to lenders. There may be significant credit exposures for trading counterparties, but the lack of leverage makes it hard for the funds to go broke and therefore fail to be able to meet their obligations. Given their importance in the financial system as a whole, regulators may wish to know what these funds are up to and thus possibly demand additional information beyond what they are required to submit now, but because of their pass-through nature they are likely to be small contributors to systemic risk. Here, too, it is unlikely that the FSOC will designate any mutual funds or their management companies as SIFIs anytime soon.
Money-market mutual funds. Consumers often use money market funds almost as if they were bank accounts, including writing checks against them in order to make day-to-day transactions or to easily withdraw cash from them. These funds are also large purchasers of commercial paper (CP) issued by both financial and non-financial corporations. In the midst of the recent financial crisis when the main alternative to CP financing — bank loans — was often unavailable, the continued viability of these funds was (and remains) especially important.
It was for both these reasons that the federal government felt compelled to guarantee money market funds in the recent crisis. The government feared that a potential major run on many, if not all, money market funds constituted a substantial risk to the financial system.
A number of changes have already been made to the regulation and operation of money market mutual funds in order to reduce their systemic risk, including a shortening of the maximum maturities of their investments and the creation of expanded disclosure. However, it remains an open issue as to whether one or more money market funds will be designated eventually as SIFIs.
Other institutional investors. There are numerous other categories of institutional investors whose members could theoretically be designated as SIFIs, but where this is unlikely to occur in practice. These include pension funds, endowments, and sovereign wealth funds, among others. In general, these share the characteristics of very low leverage, long-term funding, and the absence of a primary role as a financial intermediary. As a result, even the largest of these organizations is unlikely to represent sufficient system risk to be designated as a SIFI.
Financial market utilities. There are many entities that operate behind the scenes to implement financial transactions, such as stock and commodities exchanges, clearing houses for derivatives transactions, etc. Some of these, such as the largest clearing houses, will definitely present enough systemic risk to qualify as SIFIs, in part because of their combination of sheer size and their volume of counterparty credit risk, as well as their overall centrality to important markets. In fact, the FSOC has already designated eight financial market utilities as systemically important and may designate more.
Once SIFIs have been identified, it is almost certain that they will then be regulated differently from other financial institutions. An important underlying decision is whether the Fed’s regulation should focus solely on sources of systemic risk, holistically on the entirety of safety and soundness issues, or somewhere in between. Dodd-Frank does not clearly answer this question. On the one hand, federal regulation is imposed on non-bank SIFI’s precisely because of systemic risk issues, suggesting that such issues should be at the core of the Fed’s supervision. On the other hand, Dodd-Frank calls for heightened prudential standards for SIFI’s of all kinds, presumably on the theory that the failure of a SIFI, no matter what the cause, would have systemic repercussions.
Blending these two viewpoints, the Fed is almost certain to look at a wide range of prudential concerns, but perhaps with a sharper focus and tougher rules for those aspects that appear to increase systemic risks. For example, the Fed would be particularly inclined to be concerned about maturity mismatch and liquidity issues because they are significant safety and soundness issues in their own right while also bearing the potential to make the system as a whole riskier by triggering the equivalent of a “run on the bank”, with all the potential for contagion that would bring. On the other hand, operational issues that carry idiosyncratic risk may be given a lower priority and left largely to the primary regulators. For example, internal accounting weaknesses could help to sink a single entity, but might not have any larger systemic significance. Similarly, issues that are likely to arise at a time of wider financial crisis may garner more attention than items that are random or more likely to surface during good times, when any potential systemic problems would be easier to handle.
What can the Fed do as a supervisor? There are at least five ways additional regulation of SIFIs could occur:
Regulating at least certain non-bank SIFIs in a manner consistent with banks. One of the hardest questions in financial regulation is where to place the “perimeter of regulation.” In this case, the key question is which entities should face the heavy regulation that banks and their close affiliates do. (Banks also benefit from special privileges, such as access to deposit insurance and the Fed’s discount window, but regulation of other SIFIs may not bring such advantages in the current environment.) One of the concerns expressed in the Dodd-Frank debates was how to prevent some institutions from acting very similarly to banks, but retaining the advantage of lighter regulation. Dodd-Frank provides quite considerable powers that could be used to add many bank-like regulations (such as activity restrictions) for certain non-bank SIFIs.
If such a broad scope of regulation is applied, it is likely only to be for institutions regulators view as acting like banks. Finance companies could be caught in this net and it is theoretically possible that a large hedge fund that went after banking type business could also be brought in. This is unlikely to be an issue for most categories of non-bank SIFIs, such as insurance groups that do not already own deposit-taking institutions. That said, Dodd-Frank does provide that certain restrictions should apply to all SIFIs even though the specifics appear to have been designed primarily with banks in mind.
Information reporting. SIFIs will doubtless be mandated to provide a great deal of information, with particular emphasis on aggregate credit and counterparty exposures to other SIFIs and near-SIFIs. Other information requirements will likely include exposures to particular asset classes, capital levels, and the results of stress tests. It is also likely that many non-SIFIs will be subject to some additional reporting obligations as well, both to determine whether they qualify at some point as SIFIs themselves and also for the FSOC and its new agency in the Treasury, the Office of Financial Research, to better monitor overall system-wide financial risks.
Counterparty exposure limits. Dodd-Frank requires that banking groups limit their total exposure to individual counterparties. Non-bank SIFIs could be faced with similar requirements.
Activity limits. Banking groups are also limited by the “Volcker Rule” included within Dodd-Frank, which requires them to limit or eliminate certain types of proprietary trading and investment activity. Similarly, provisions pushed by Senator Lincoln created restrictions on the ability of banking entities to act as derivatives dealers. Non-bank SIFIs might be placed under similar restrictions on activities that are perceived as being particularly risky and not at the core their business models, or at least the business models policymakers view as being in the public interest.
Capital requirements. One of the most important ways that regulators can encourage safety at financial institutions is to require appropriate levels of capital as a margin for error against losses that might come through bad luck or errors. Banking groups already face substantial capital requirements that are being tightened significantly through the so-called Basel III process, coordinated by the Basel Committee on Banking Supervision. Insurers also have substantial capital requirements imposed by their regulators for similar reasons. Dodd-Frank specifically calls for SIFIs to face higher capital requirements than non-SIFIs, with the details to be determined by the regulators.
Capital requirements are such a universal, and important, element of the regulatory approach to banks that there is a strong likelihood that non-bank SIFIs will be subjected to similar requirements. This is most likely for SIFIs that perform a classic intermediation function and have large balance sheets, such as finance companies, which play a role fairly similar to banks. Some sort of capital regulation might also be extended to hedge funds, although these funds may be able to argue that their differences from banks justify an exemption from any capital regulation. Other asset managers, such as mutual funds or venture capital management companies, are the least likely to have this requirement, because their business models create little need for capital. As discussed below, capital requirements already exist for insurers and may be expanded or altered by the Fed in its role as a regulator of SIFIs.
Capital regulation is an extremely powerful tool to affect the behavior of financial institutions, since it very directly alters their ability to provide an adequate return to their shareholders. This is even more powerful since top managers in financial institutions almost invariably hold a considerable amount of their net worth in company stock. If this powerful tool is applied too widely, such as to funds managers that act as pass-through entities and not true intermediaries, it could substantially change the ability of otherwise valid business models to work. Ironically, adding an unreasonable burden to, say, mutual funds could push financial assets into the hands of financial intermediaries instead that present greater systemic risks.
Liquidity requirements. The recent financial crisis underlined the importance of liquidity, the ability to come up with cash, potentially on short notice, to cover deposit withdrawals, debt redemptions, and other needs. Banks will have quite extensive liquidity requirements going forward and the Fed will certainly consider appropriate liquidity requirements for other SIFIs.
Principles for regulating non-bank SIFIs
Some key principles should guide the Fed’s regulation of non-bank SIFIs.
Carefully balance the costs and benefits when designing regulation and supervision. This is important for all regulators and is so basic that it probably needs little further elaboration. However, it will be critical not to lose sight of this key principle. It will always be tempting for the Fed to add yet further constraints and safety margins on non-bank SIFIs, in its pursuit of systemic stability, particularly as the Fed will take the blame if a serious future crisis develops. However, safety margins come with costs and it would be harmful to the economy if those costs were excessive compared to what may be only a modest increase in stability from a given regulation. For example, equity capital is significantly more expensive, in practice if not always in theory, than other sources of funding. Requiring more capital therefore adds a cost that will have to be absorbed by some combination of customers, employees, stockholders, and others who deal with the firm. Deciding what regulations to impose and choosing which firms they are imposed upon must be a balancing act between the improvements in safety and the economic costs of achieving the improvements.
Defer to primary regulators as appropriate while ma