The Dodd–Frank Act was the principal response of the United States to the 2008 financial crisis. 1 Despite the fact that the crisis actually had little to do with asset connectedness, many of the most important provisions of the Act are addressed to this form of systemic risk. This is not to say these policies are bad and may not address potential problems in the future, but as we will see, contagion was the major problem in the crisis, and Dodd–Frank made this problem worse, not better.
There are three key provisions of Dodd–Frank addressed to asset connectedness: (1) the requirement for central clearing of over-the-counter derivatives,2 (2) the imposition of counterparty exposure limits,3 and (3) the designation of systemically important nonbank financial institutions.4
Central clearing of derivatives and other financial contracts may reduce the magnitude of asset connectedness. Under central clearing procedures, counterparty exposures are guaranteed by a central clearing counterparty (“CCP”), whose sole business is to stand between parties and assume the credit risk of buyers and sellers. Some scholars have concluded that participants in centrally cleared markets will have reduced incentives to flee from a weak counterparty.5 By making “counterparty runs” less likely, central clearing may forestall the failure of a weak financial institution,6 and in the unlikely event of a financial institution’s collapse, “[e]ffective clearing mitigates systemic risk by lowering the risk that defaults [will] propagate from counterparty to counterparty.”7 As indicated by the experience of counterparties holding centrally cleared exchange-traded derivatives against Lehman—which suffered no losses from the bank’s collapse8—central clearing can completely insulate market participants from losses associated with the default of a dealer. Moreover, to the extent that it might have mitigated liquidity pressures on Lehman by reducing the likelihood of derivatives counterparty runs, a greater degree of central clearing might have also benefited Lehman. Given the potential risk-reducing effects of central clearing, it is not surprising that the Dodd–Frank Act mandates central clearing in certain cases. Specifically, swaps and security-based swaps not used to hedge commercial risk must be cleared if the relevant regulator so determines.9 In making this determination, the regulators are instructed to examine, inter alia, contract liquidity, operational capacity, and the effect of clearing on the mitigation of systemic risk.10 The more standardized a contract, the more likely it is to be subject to mandatory clearing.
The efficacy of central clearing is, however, subject to certain limitations. One key limitation of central clearing is that CCPs are ill-suited to handling highly customized derivatives contracts. For example, a CCP would likely be unable or unwilling to clear AIG’s notorious CDS portfolio on multi-sector CDOs.11 Some might argue that the inability of CCPs to clear highly customized derivatives contracts is a rationale for the abolition of such contracts, yet, as the Treasury has recognized, such derivatives can play a “legitimate and valuable role.”12 Nevertheless, even if all contracts were standardized and subject to clearing, systemic risk would not be entirely eliminated and, under certain circumstances, could even increase. As derivatives increasingly head toward central clearing, CCPs themselves may cross the “too-big-to-fail” threshold.13 As Fed Chairman Bernanke has observed, “[T]he flip side of the centralization of clearing and settlement activities in clearinghouses is the concentration of substantial financial and operational risk in a small number of organizations, a development with potentially important systemic implications.”14 While clearinghouses do minimize the risk of bilateral connectedness, they increase the risk of multilateral connectedness because they connect multiple participants who, but for the clearinghouse, might not be connected at all.
A second limitation of central clearing is that counterparty credit risk transfers to CCPs may not be one for one. The nature of the transfer depends on the extent to which market participants are able to net exposures across asset classes and parties, which in turn depends on the number and nature of CCPs. The introduction of a CCP for a particular asset class will be risk reducing only “if the opportunity for multilateral netting in that class dominates the resulting loss in bilateral netting opportunities across uncleared derivatives from other asset classes.”15 In other words, by using a CCP for a particular asset class, market participants may obtain the benefit of netting in the asset class across parties (multilateral netting) but will lose the benefit of netting against another party outside the asset class (bilateral netting). One might wonder why such a trade-off between multilateral and bilateral netting need exist in the first place. Indeed there would not necessarily be as striking a trade-off if a CCP itself were to net multiple asset classes or CCPs each netting one (different) asset class were linked to one another in a manner allowing netting across asset classes. However, at present, such a “first-best” solution does not exist.16
A third and final limitation of central clearing is that derivative participants will lose initial margin as a source of funding, since under central clearing procedures margin is instead posted to a CCP. As a result banks will likely seek alternative sources of short-term funding. At the same time, increased central clearing of derivative products will cause a corresponding increase in demand for safe collateral. Studies have estimated that demand for such collateral could increase by $5 trillion as a result of the migration of swaps to CCPs. 17 Such increased demand will likely lead to rising costs of safe collateral, and hence rising funding costs for financial institutions. 18
The single-counterparty credit concentration limits required by the Dodd–Frank Act are also designed to address the systemic risks of asset connectedness.19 Banks commonly monitor and limit their exposures to individual counterparties and have long been subject to state and federal laws limiting the amount of credit that may be extended to a single borrower. The Dodd–Frank Act requires the Federal Reserve to establish limits to prevent covered companies from having credit exposures to any unaffiliated company in excess of 25 percent of the capital stock and surplus of the covered company.20 The Federal Reserve is authorized to reduce this limit if “necessary to mitigate risks to the financial stability of the United States.”21 In January 2012 the Federal Reserve proposed rules to implement this provision,22 and in fact chose to lower the counterparty exposure threshold to 10 percent for entities with greater than $500 billion in consolidated assets.23
Implementation of counterparty exposure limitations poses numerous challenges. First, a 25 percent limit may simultaneously be overly generous or overly restrictive, depending on the counterparty in question. Critics of the Federal Reserve’s approach have argued that the proposed rules’ method of calculating counterparty exposure is extremely narrow and inconsistent with the approach the Fed itself has taken in other contexts.24 The Federal Reserve has not explained why a 10 percent restriction for large banks is more appropriate than a 25 percent limitation, particularly given the potential increase in connectedness as large banks are required to “spread their exposures across more and smaller, potentially less stable counterparties.”25 The lack of an exemption for a bank’s exposure to central clearing parties is also troubling, as it works at cross-purposes to the Dodd–Frank Act’s central clearing requirements.26 But the fact remains that the members of a clearinghouse are all ultimately at risk if clearinghouse losses exceed collateral and clearinghouse capital.
Exposure limitations are also a potential response to problems of liability connectedness. Liability connectedness involves a firm’s exposure to the risk of a withdrawal of credit by a significant funding source. In theory, regulators could impose caps on the level of a financial institution’s dependence on any single funding source, thus checking the funding domino effect in the event of the failure of a large financial institution. As argued throughout this book, liberal recourse to the central bank’s lender-of-last-resort facilities is better approach to addressing funding shortfalls.
Section 113 of Dodd–Frank gives the Financial Stability Oversight Council the power to designate nonbank systemically important financial institutions (SIFIs), with the consequence that such firms shall be regulated by the Federal Reserve and be subject to enhanced supervision and capital requirements. In making such designations the connectedness of potential SIFI designees plays a major role. Section 113 requires the FSOC to consider the “connectedness of the company”27 and the “extent and nature of the transactions and relationships of the company with other significant nonbank financial companies and significant bank holding companies”28 In its final rule the FSOC finds that the most important factor for determining the systemic importance of a nonbank financial company is the “exposure of creditors, counterparties, investors, or other market participants to a nonbank financial company.”29 This focus on connectedness is underscored by the fact that three insurance companies, AIG, MetLife, and Prudential, have been designated as nonbank SIFIs despite the fact that none of these nonbanks depends to a significant extent on short-term funding, and is therefore not susceptible to contagion risk.
Although all US banks with assets over $50 billion are designated as systemically important under Dodd–Frank, the additional capital surcharges applicable to global systemically important US banks (G-SIBs) is also premised on the theory of connectedness. This is plainly obvious from the relevant measures in the Federal Reserve’s recent rule, which is largely consistent with the Basel Committee standards.30 For example, the Basel Committee requires the consideration of five broad categories: (1) size, (2) connectedness, (3) cross-jurisdictional activity, (4) substitutability, and (5) complexity. However, the relevant data inputs for factors other than connectedness frequently relate to connectedness, as demonstrated below for the category of size.31
Thus the Dodd–Frank Act has strong measures to combat connectedness despite the lack of evidence that this was a real problem in the crisis. That said, however, it could be a problem in the future. We now turn to the real problem in the crisis, contagion.
Federal Reserve: Banking organization systemic risk report