One cycle is invariable in securities regulation from the time of the South Seas Bubble, circa 1715-1720 (when this author was only an assistant professor) to today: market crashes produce securities legislation. Last week, the Senate Banking committee heard testimony from eight witnesses on what policy changes were required to address the current financial crisis. 1 The first witness, a modest, humble law professor from Columbia, argued that the recent collapse of investment banks, insurance companies, and hedge funds showed the need for a “systemic risk regulator.” 2 He attributed the origins of the current crisis to two causes: (1) the excesses of an “originate-and-distribute” model under which financial institutions – including loan originators, mortgage banks, and investment banks – lost any incentive to screen borrowers for creditworthiness because they knew they would not hold the loans for long enough to matter; and (2) deregulatory policies that permitted financial institutions to self-regulate with respect to capital adequacy and risk management. Under these conditions, a moral hazard problem developed; lending discipline evaporated; and asset-backed securitizations deteriorated into toxic waste. Based on his view that self-regulation would not work in an environment of intense competition, he called for legislation creating a “systemic risk regulator.”
Normally, the views of the academic witness at congressional hearings are politely ignored by the various industry representatives who follow. But this time, six of the next seven witnesses also agreed that a systemic risk regulator was needed – a level of consensus sufficient to suggest a conspiracy. The committee also seemed receptive to the concept of such a risk regulator. Only the last witness disagreed. Lynn Turner, the former chief accountant of the SEC, told the committee that the only response that would work was to “appoint good people to the SEC.” He also called for transfer of the CFTC’s authority over financial futures to the SEC, but otherwise dismissed structural reform as irrelevant.
So that was the range of opinion: The Old Guard said “stick with the SEC,” while virtually everyone else favored a new regulator for this special task of prudential financial oversight. Few at the hearing (including the senators) thought that the SEC could successfully perform such an oversight role. Some pointed to its limited jurisdiction and the failure of its Consolidated Supervised Entity Program (all of whose participants have by now either failed, merged into a bank, or become bank holding companies in 2008). Others (including this author) believe that the SEC is at its best as an enforcement agency, but, as a lawyer-dominated agency, lacks the culture, competence and personnel to undertake the task of prudential financial supervision across the financial landscape. Still, senators of both parties were equally skeptical of the Federal Reserve as the other logical candidate for the role of “Systemic Risk Regulator” (or “SRR”). The Federal Reserve, they agreed, is not adequately politically accountable. An aristocrat that does not depend on congressional funding, it tends to ignore Congress (as shown by its recent refusal to identify AIG’s counterparties to Congress). 3 More importantly, banking regulators in general tend to prefer secrecy to transparency and sometimes form a conspiracy of silence with the regulated to hide and suppress problems with financial institutions that might, if disclosed, cause depositors and other suppliers of capital to flee. In short, while the culture of securities regulators tends to stress openness, transparency, and full disclosure, the culture of bank regulators starts from a fear that full disclosure might produce a “run on the bank.”
Against this backdrop, this column will examine what an SRR should do and the policy tradeoffs surrounding who might best perform this role. Although the concept of “systemic risk regulator” is not self-defining, most who use the term envision a body that would, at a minimum, have responsibility for the capital adequacy, safety and soundness, and risk management practices of all institutions that are too “big to fail.” The two key advantages of a unified systemic risk regulator with jurisdiction over all large financial institutions are that (a) its broader jurisdiction makes it less subject to regulatory capture, 4 and (b) the common standard that it would enforce solves the critical problem of regulatory arbitrage. AIG, which has already cost U.S. taxpayers over $160 billion, presents the paradigm of this latter problem because it managed to issue billions of dollars in credit default swaps without becoming subject to regulation by any regulator at either the federal or state level. Predictably, if there are material differences in the intensity of regulatory supervision, entrepreneurs will game the system by electing to conduct their activities through entities subject to the least regulation. Given the convergence of financial institutions, almost any entity can today conduct almost any activity. Hence, a systemic risk regulator needs universal jurisdiction.
But more is involved than simply supervision of capital adequacy. Systemic risk is most easily defined as the risk of an interconnected financial breakdown in the financial system – much like the proverbial chain of falling dominoes. The closely linked insolvencies of Lehman, AIG, Fannie Mae and Freddie Mac in the fall of 2008 illustrates the plausibility of this scenario of an interconnected crescendo of financial failures. Were some of these firms not bailed out, other financial institutions were likely to have also failed. The key idea here is actually not that one financial institution is too big to fail, but rather that some institutions are so interconnected that the failure of one could mean the failure of all.
What should a system risk regulator be authorized to do? Among the obvious powers that it should have are the following:
a. Authority to Limit the Leverage of Financial Institutions and Prescribe Mandatory Capital Adequacy Standards. At a minimum, Congress would empower the SRR to prescribe minimum levels of capital and ceilings on leverage for all categories of financial institutions, including banks, insurance companies, hedge funds, money market funds, pension plans, and quasi-financial institutions (such as, for example, G.E. Capital). These standards need not be identical for all institutions and should be risk-adjusted. The SRR should be authorized to require reductions in debt to equity ratios below existing levels, to consider off-balance sheet liabilities (including those of partially owned subsidiaries and also contractual agreements to repurchase or guarantee) in computing these ratios (even if generally accepted accounting principles would not require inclusion of these liabilities).
The SRR would focus its monitoring on the largest institutions in each financial class, leaving smaller institutions to be principally regulated and monitored by their primary regulator. For example, the SEC might require all hedge funds to register with it under the Investment Advisers Act of 1940, but hedge funds with a defined level of assets (say, $50 billion in assets) would be subject to the additional and overriding authority of the SRR. Because there are over 5,000 registered broker-dealer firms in the U.S. today, it would be infeasible to ask the SRR to exercise close oversight over smaller firms, which would continue to look to the SEC as their primary regulator. The radar screen of the SRR should only pick up the larger flying objects.
b. Authority to Approve, Restrict and Regulate Trading in New Financial Products. By now, it has escaped no one’s attention that one particular class of over-the-counter derivative (the credit default swap) grew exponentially over the last decade and was outside the jurisdiction of any regulatory agency. This was not accidental, as the Commodities Futures Modernization Act of 2000 deliberately placed over-the-counter derivatives beyond the general jurisdiction of both the SEC and the CFTC. A highly debatable question is whether the SRR should be required to give advance approval for new financial products. This might slow innovation undesirably. More important are the trading practices associated with these products, in particular the adequacy of the collateral or margin that counter-parties are required to post. This is what truly sank AIG, which contracted to post additional collateral simply on a rating downgrade. Arguably, the SRR should be authorized to limit those eligible to trade such instruments and could bar or restrict the purchase of “naked” credit default swaps (although the possession of this authority would not mean that the SRR would have to exercise it, unless it saw an emergency developing).
c. Authority to Mandate Clearing Houses. Securities and options exchanges uniformly employ clearing houses to eliminate or mitigate credit risk. In contrast, when an investor trades in an over-the-counter derivative, it must accept both market risk (the risk that the investment will sour or price levels will change adversely to the investor) and credit risk (the risk that the counterparty will be unable to perform). Credit risk is the factor that necessitated the bailout of AIG, as its failure could have potentially led to a cascade of failures by other financial institutions if AIG defaulted on its swaps. Use of the clearing house should eliminate the need to bail out a future AIG because its responsibilities would fall on the clearing house to assume and the clearing house would monitor and limit the risk that its members assumed.
At present, several clearing houses are in the process of development in the United States and Europe, and ICE just began clearing swap contracts this month. The SRR would be the obvious body to oversee such clearing houses (and indeed the Federal Reserve was already instrumental in their formation). Otherwise, some clearing houses are likely to be formed under the SEC’s supervision and some under the CFTC’s, thus again permitting regulatory arbitrage to develop. A final and complex question is whether competing clearing houses are desirable or whether they should be combined into a single centralized clearing house. This issue could also be given to the SRR.
d. Authority to Mandate Writedowns for Risky Assets. A real estate bubble was the starting point for the 2008 crisis. When any class of assets appreciates meteorically, the danger arises that on the eventual collapse in that overvalued market, the equity of the financial institution will be wiped out (or at the least so eroded as to create a crisis in investor confidence that denies that institution necessary financing). This tendency was palpably evident in the failure of Bear Stearns, Lehman, Fannie Mae and Freddie Mac. If the SRR regulator relies only on debt/equity ratios to protect capital adequacy, they will do little good and possibly provide only illusory protections. Any financial institution that is forced to writedown its inventory of overpriced mortgage and real estate assets by 50 percent will necessarily breach mandated debt to equity ratios. The best answer to this problem is to authorize the SRR to take a proactive and countercyclical stance by requiring writedowns in risky asset classes (at least for regulatory purposes) prior to the typically much later point at which accountants will require such a writedown.
Candidly, it is an open question whether the SRS, the Federal Reserve, or any banking regulator would have the courage and political will to order such a writedown (or impose similar restraints on further acquisitions of such assets) while the bubble was still expanding. But Congress should at least arm its regulators with sufficient power and direct them to use it with vigor.
e. Authority to Intervene to Prevent and Avert Liquidity Crises. Financial institutions often face a mismatch between their assets and liabilities. They typically invest in illiquid assets or make long-term loans, but their liabilities consist of short-term debt (such as commercial paper). Thus, regulating leverage ratios is not alone adequate to avoid a financial crisis, because the institution may suddenly experience a “run” (as its depositors flee) or be unable to roll over its commercial paper or other short-term debt. This problem is not unique to banks and can be encountered by hedge funds and private equity funds (as the Long Term Capital Management crisis showed). The SRR thus needs the authority to monitor liquidity problems at large financial institutions and direct institutions in specific cases to address such imbalances (either by selling assets, raising capital, or not relying on short-term debt).
Failed Financial Technology
The foregoing are the minimum duties to be assigned an SRR. But it can also be asked to do more. Unless the credibility of asset-backed securitizations can be restored, housing finance in the United States will remain moribund. The blunt truth is that asset-backed securitizations represent a financial technology that failed. This failure seems principally attributable to a “moral hazard” problem that arose under which both loan originators and underwriters relaxed their lending standards and packaged non-creditworthy loans into portfolios, because both found that they could sell these portfolios at a high profit and on a global basis – at least so long as the debt securities carried an investment grade rating from a major credit rating agency. Put differently, in a bubble market, the major participants in the securitization process all expect that they can pass the assets on to the next buyer in the chain – “before the music stops” (in Mr. Prince’s memorable phrase). Thus, they tend to economize on due diligence and ignore signs that the assets are not creditworthy, because none expect to bear the costs of holding the financial assets to maturity. As a disclosure agency, the SEC is poorly positioned to exercise paternalistic oversight or impose prophylactic rules (and clearly it did not do so).
What could an SRR do instead? When asset-backed securitizations began in the early 1990s, the promoter usually issued various tranches of debt to finance its purchase of the mortgage assets, and these tranches differed in terms of seniority and maturity. Typically, the promoter would sell the senior most tranche in public offerings to risk-averse public investors and retain some or all of the subordinated tranche, itself, as a signal of its confidence in the creditworthiness of the underlying assets. Over time, this practice of retaining the subordinated tranche withered away. In part, this was because hedge funds would take the risk of buying this riskier debt; in part, it was because the subordinated tranche could be included in more complex CDOs (where overcollateralization was the investor’s principal protection), and finally it was because, in a bubbly market, investors no longer looked for commitments from the promoter.
Given this definition of the problem, the answer seems obvious: require the promoter to retain some portion of the subordinated tranche. This would incentivize it to buy only creditworthy financial assets and end the “moral hazard” problem. But such a prophylactic rule could never be adopted by the SEC, because it goes well beyond a disclosure or antifraud rule. But this is precisely what an SRR could do to minimize systemic risk. 5
Finally, the last and most sensitive responsibility that could be given to an SRR is to shift financial regulation from a pro-cyclical bias to a counter-cycled one. Many rules and policies today are emphatically pro-cyclical. When the SEC in 2007 eliminated the uptick rule, 6 which bars short selling when the stock price has fallen on its most recent price movement, this step arguably increased market volatility because short sellers will predictably be more active in a declining market and less so in a rising market. As a result, many are lobbying the SEC to bring back the uptick rule. But such a rule might be easily evaded now that stock prices trade in pennies and not eighths of a point. More importantly, a restored uptick rule might deter short selling in a bubbly market when policy makers should want stock overvaluations to be challenged. After all, it was the short sellers who first detected the excessive valuation of Enron and began to counter it. Thus, a better rule might be to employ a “circuit-breaker” approach that would simply bar short-selling after a substantial stock price decline (either in the market as a whole or in the individual stock) for a defined period, but that would freely permit short-selling (without any tick test) in a rising market (at least after some defined percentage increase occurred). Short selling would then become a counter-cyclical force.
Arguably, another example of a pro-cyclical policy is mark-to-market accounting. Predictably, financial institutions would lobby an SRR to suspend such an accounting rule, because it typically requires writedowns in a declining market. But here we again encounter the classic tension between the bank regulators’ desire to hide bad news and securities regulators’ preference for the truth. The truth is often ugly and pro-cyclical. If the value of mortgage assets has declined, it cannot fairly be assumed that they will recover with the business cycle, as housing prices are not apt to recover that quickly and mortgagors are unlikely to pay off mortgages that exceed the fair market value of their properties. Hence, a writedown is necessary, even if it is arguably pro-cyclical.
This illustration shows the need for assuring the independence of the investor protection agency (presumably the SEC) if an SRR is created. Under a “twin peaks,” both the SRR and the consumer protection agency are independent, and one is not subordinated to the other. Around the world today, this is not the most common pattern. In the U.K., the Financial Services Authority is a centralized regulator that is responsible both for prudential oversight and consumer protection. But these often inconsistent purposes may lead to low-visibility trade-offs within the single agency. To some degree, these trade-offs are inevitable. Securities litigation may protect investors, but it necessarily threatens bank solvency (as the $7.5 billion recovered in the Enron case, mainly from banks, proves). The “twin peaks” model at least makes the tensions visible and invites Congress to resolve them, thereby assuring greater political accountability.
Within the U.S. context, the Federal Reserve Board is the most obvious candidate to be assigned the responsibilities of an SRR. As our central bank, it is, after all, the body that, if preventive controls fail, will be called on to provide the bail out. Thus, it should be properly motivated to reduce risks, as it largely bears the costs of those risks. But it is not committed to transparency and is subject to only limited political accountability.
As a result, some have suggested that a new commission be created whose commissioners would be the chairs of the various federal financial regulatory agencies and whose own chair would be the chair of the Federal Reserve Board. This body would have its own staff, and congressional funding would support it (thus making it more amenable to congressional supervision). Interesting as this idea is, Congress should remember that a camel is an animal designed by committee. The sensible debate for the near future should be whether it is better and more realistic to refashion the Federal Reserve into a more accountable body or to design a new body to serve as SRR. But this question must be answered soon – if only provisionally.
John C. Coffee Jr. is the Adolf A. Berle Professor of Law at Columbia University Law School and director of its Center on Corporate Governance.
1. See Hearings Before the United States Senate committee on Banking, Housing and Urban Affairs on “Enhancing Investor Protection and the Regulation of Securities Markets” on March 10, 2009 (available on LEXIS).
2. This humble professor’s testimony is available at the Columbia Law School Web site. I have analyzed the same issues in greater detail in a draft article with Professor Hillary Sale, which will appear in the 75th Anniversary of the SEC volume of the Virginia Law Review. See Coffee and Sale, “Redesigning the SEC: Does the Treasury Have a Better Idea?” (available on the Social Science Research Network at http://ssrn.com/abstract=1309776).
3. See Brady Dennis, “Senators Call AIG ‘Lost Cause,’” Washington Post, March 6, 2009 at D-2 (describing senatorial anger at Federal Reserve’s refusal to name AIG counterparties). Because the Federal Reserve is not part of the executive branch, it would seem unentitled to raise any claim of executive privilege (at least absent presidential directions not to respond).
4. See Jonathan Macey, “Organizational Design and Political Control of Administrative Agencies,” 8 J. Law, Economics and Organizations 93 (1992) (arguing that agencies with “narrow” missions are more easily captured than those with broad missions). Debatable as this issue is, an SRR with authority over all financial institutions seems less easily captured than one with jurisdiction only over investment banks. In case of the Federal Reserve, which at least partially funds bailouts, the incentive to resist capture is even stronger.
5. To make this proposal truly effective, however, more must be done. The promoter would have to be denied the ability to hedge the risk on the subordinated tranche that it retained. Otherwise it might hedge that risk by buying a credit default swap on its own offering through an intermediary.
6. The “uptick rule” was formerly codified in Rule 10a-1. The SEC proposed to replace this test in late 2006. See Securities Exch. Act Rel. No. 34-54891 (Dec. 7, 2006). It adopted its proposal in 2007. See Securities Exch. Act Rel. No. 34-55970 (July 3, 2007).