Comment Text:
Mr. David Stawick
Secretary
Commodity Futures Trading Commission
Three Lafayette Center
1155 21st Street NW
Washington, DC 20581
1 February, 2011
Dear Mr Stawick
Re: Comments of the United States Department of Justice on: 17 CFR Parts 1, 37, 38, 39, and 40 Requirements for Derivatives Clearing Organizations, Designated Contract Markets, and Swap Execution Facilities Regarding the Mitigation of Conflicts of Interest
The CFTC is currently considering rules regarding the ownership of, and membership requirements for, Designated Clearing Organizations (“DCOs”), Designated Contract Markets (“DCMs”), and Swap Execution Facilities (“SEFs”). As a scholar who has written extensively on (a) the organization and governance of financial exchanges, (b) clearing, and (c) competition and market power in trading exchanges and clearing, I have serious concerns about proposals that would impose aggregate caps on the limits on the aggregate fraction of DCOs, DCMs, and SEFs that classes of firms (e.g. dealer banks) can own. I also have concerns about regulation of the capital requirements of members of DCMs. Based on my extensive research—including much published in peer-reviewed academic journals—I conclude that such limits and regulations (a) may actually undermine the objectives of the Dodd-FrankWall Street Reform and Consumer Protection Act (the “Dodd-Frank Act”), and (b) are unlikely to enhance competition in trading and post-trade services. I further conclude that the recent analysis of these issues by the Antitrust Division of the Department of Justice is fundamentally flawed and is a very poor guide to policy.
The issues are somewhat different for DCOs on the one hand, and DCMs and SEFs on the other, so I will address each separately in turn starting with DCOs.
Aggregate Ownership Limits on DCOs
Mandated use of clearing is a central element of the Dodd-Frank Act’s plan to reduce systemic risk arising from derivatives transactions. As a result of the Dodd-Frank Act clearinghouses (or central counterparties) will be the foundation of the infrastructure of the derivatives markets. As such, ensuring their safety and security is paramount. One shudders at the thought of the consequences of financial distress at a large clearinghouse.
To operate effectively, efficiently and most importantly safely, clearing entities must have access to sufficient capital to absorb member firm defaults. Moreover, members themselves must have incentives to set margins that are adequate to control efficiently counterparty risk. Furthermore, the ability of central counterparties (“CCPs”) to operate effectively, particularly under conditions of stress, depends crucially on the efficiency of their governance. Aggregate ownership limits can interfere with all of these essential functions of clearing entities.
CCPs are a mechanism for pricing (via collateral) and sharing counterparty risk. Any financial institution that agrees to participate in the sharing arrangement will reasonably expect to have control rights in the organization that oversees this sharing mechanism. No financial institution is going to agree to share in a large fraction of the risk borne by a clearinghouse if it is restricted to having a small share in control. This means that a restriction of ownership and control rights will inevitably exert a decisive influence on the allocation of counterparty risk through a CCP. If swap dealers and other major swap market participants are limited in the amount of control they can exercise over CCPs, they will limit the amount of risk they are willing to bear. That is, they will limit the amount of capital that they contribute to absorb the risks of loss arising from the default of member firms. Thus, any restriction on the fraction of a DCO that a particular class of firm can own will limit the amount of capital this class of firm will be willing to supply to that DCO.
It should also be noted that CCPs must be able to manage defaults by replacing the positions of defaulting members. In doing so, they rely on non-defaulting member firms to step in and bid on the defaulted positions. Restrictions on ownership are likely to reduce both the willingness of the restricted entities to participate in this essential process, and the scale of their participation.
Given that large dealer firms (a) are the largest sources of risk capital, (b) have expertise in pricing and managing counterparty risk, and (c) have expertise in pricing and managing the risks of large positions, they have a comparative advantage in supplying capital to CCPs, bearing the default risks in CCPs, and absorbing and managing the risks of defaulted members. This means that they are the most efficient providers of these vital services. Constraining their ability to exercise control over CCPs will limit the amount of these services that such firms will supply. Other firms can offset this to some degree, but the fact that they will be less efficient at performing these functions means that with restricted ownership, CCPs are likely to be less well-capitalized, and less able to absorb and manage member defaults. This result would be directly contrary to the Dodd-Frank Act’s objective of enhancing the soundness of the financial system.
Nor is this all. If they bind, aggregate ownership limits will increase the heterogeneity of CCP membership. In particular, it is likely that under the restrictions, smaller, less well-capitalized firms will represent a larger fraction of CCP ownership than would be the case absent restrictions. Thus, under the restrictions CCPs will have both large and small firms as members.
It is well known that entities with heterogeneous owners are more vulnerable to conflict and governance difficulties than entities with homogeneous owners. Conflicts of interest are more likely when owners are heterogeneous. Such conflicts lead to disputes over decisions and management. For instance, in a CCP, members with different levels of capital at stake can have very different incentives with respect to the setting of margin levels. Less well-capitalized members may benefit from the greater volume of business that results when margins are reduced, but do not bear the bulk of the costs when margins are inadequate to cover a member firm’s losses and that member defaults. Greater heterogeneity therefore creates a real risk that CCPs will choose lower than optimal margins.
Thus, aggregate ownership rules (or rules on minimum capital) that result in the admission of smaller, less-well capitalized members can lead CCPs to make decisions that compromise their safety and soundness.
Moreover, forced heterogeneity will affect the rules and governance of CCPs. I have shown in my academic research that member heterogeneity leads to the adoption of cumbersome governance and decision-making mechanisms. Elaborate constraints on the discretion of management are necessary to reduce the likelihood that rules and decisions are used to benefit one type of member at the expense of another type. Organizations with more homogeneous membership can dispense with such constraints because the interests of members are well-aligned due to their similarity. Aggregate ownership restrictions that forcibly increase the heterogeneity of membership will result in organizations that are more prone to conflict, more cumbersome to manage, less effective at responding to changes in the marketplace, and less effective at responding to crises that are likely to have disparate impacts on different types of firms.
It should also be remembered that the owner-members of CCPs typically use clearing services: they are customers as well as owners. Heterogeneity can create conflicts between owners over the pricing of CCP services if ownership shares differ from the distribution of the demand for CCP services among members. For instance, if smaller firms own a fraction of a CCP that exceeds the share of business they contribute to the CCP, they would like the CCP to charge high prices as a way of diverting profits from the bigger firms: in this situation, the smaller firms more than make up for the higher fees that they pay with a bigger share of profits from the CCP. This can result in the CCP charging excessively high fees that harm not just the larger members, but the ultimate customers of CCP members. Aggregate ownership limits are therefore likely to affect the pricing of CCP services, and in particular can cause CCP fees to be higher than they would be in the absence of these limits.
In sum, aggregate ownership limits are very costly, and create risks that undermine the very purpose of CCPs.
This is not to say that CCPs operating without any restriction or regulatory oversight will have the optimal number of members. Indeed, I have demonstrated in my academic research that, given the extensive economies of scale and scope in clearing, it is possible for a suboptimally small group of financial firms to form a CCP without fear of competition. A CCP that is smaller than optimal, but has just enough members to achieve scale and scope economies that no other potential competitor could match, generates super-competitive profits for those members. This means that CCPs can exercise market power by adopting excessively restrictive membership requirements.
Thus, there is a fundamental tension. On the one hand, imposing restrictions on the membership policies of CCPs, either through aggregate ownership limits or through limitations on their ability to impose membership requirements, can fundamentally compromise their financial safety and soundness. On the other, granting them complete discretion over membership could lead to the creation of CCPs that are suboptimally small and which can therefore exercise market power through the adoption of excessively onerous membership requirements.
The Antitrust Division is concerned about the potential for CCPs to exercise market power, but treats the safety and soundness concerns dismissively. Its dismissal is cavalier, and exhibits far too little concern and appreciation for the very real financial and systemic risks that can result when CCP membership and ownership are constrained by regulation. Given that the Division’s focus, and indeed its reason for being, is on controlling market power, its one-sided treatment is perhaps to be expected. However, the Commission has the responsibility under the Dodd-Frank Act to promulgate rules that promote the safety and soundness of the financial system without imposing undue burdens on competition. Thus, it cannot treat concerns about the effects of regulations on the safety and soundness of CCPs as cavalierly and superficially as has the Antitrust Division. Moreover, a fair reading of Dodd-Frank would support the view that ensuring safety and soundness is the Commission’s paramount duty to which concerns about competition are subordinate.
The trade-off between safety and soundness on the one hand, and competition on the other, is not an easy one to make. I am deeply skeptical, however, that aggregate ownership limits will strike the appropriate balance between these competing considerations.
Based on my years of studying the economics of clearing and the organization and governance of mutual and mutual-like financial entities such as CCPs, I conclude that aggregate ownership restrictions would jeopardize their safety and soundness by dramatically inflating the costs that the most efficient bearers of counterparty risk incur to participate in clearing arrangements.
Restrictions on the composition of ownership and membership (in the form of aggregate ownership limits) are unlikely to mitigate the exercise of market power. The ability to exercise market power through inefficient restrictions on the size of CCPs derives from the economies of scale and scope in clearing. Aggregate membership limits in no way affect these underlying sources of market power. Furthermore, due to these economies of scale and scope, a CCP can exercise market power by limiting the number of members while still complying with constraints on the fraction of the CCP that can be owned by any given type of firm. Market power relates to size, not composition, meaning that regulating composition is an inappropriate way to control market power. In fact, regulations of the composition of CCP ownership are likely to lead to costs and no benefits - costs in the form of inefficient risk bearing and compromises in safety and soundness ¬ and no benefit in the form of increased competition.
The Antitrust Division and others have also argued that bank/dealer control of CCPs is problematic because these large financial institutions will attempt to protect their profitable OTC derivatives businesses by refusing to make certain products available for clearing. For several reasons, these concerns are vastly overblown.
First, even assuming that the argument is correct, if clearing is indeed more cost-efficient than bilateral dealings, mandated clearing of products that are undeniably suitable for clearing would lead to a decline in the costs of trading these products, relative to the costs of trading the kinds of customized products that dealers would arguably attempt to continue to trade bilaterally. This would lead customers to substitute away from the more customized products to the cleared ones. Dealers cannot compel their customers to use the higher-cost products. This ability of customers to substitute away from products that dealers find more profitable - and hence customers find more costly - undercuts the benefits that dealers realize by keeping some products bilateral.
Second, it is by no means evident that clearing would materially undercut dealer market power in more complex products, even if such market power exists. Clearing per se does not increase an end user’s choice of counterparties with the expertise in structuring more complicated products tailored to his particular needs; large dealers almost certainly would retain a strong competitive advantage over smaller banks and other financial intermediaries in the marketing of complex products, even if those products were cleared. Clearing per se does not increase pre-trade transparency in a way that might - might - increase competition between dealers. Thus, moving more complex products to a CCP does not undercut these potential sources of dealer market power. Consequently, clearing is unlikely to subject dealers to more intense competition in the design and marketing of more complex products, meaning that fear of increased competition is unlikely to induce even a dealer-dominated CCP to refuse to clear them.
Third, under the Dodd-Frank Act non-cleared derivatives are likely to be subject to higher capital and margin requirements than cleared ones. This will provide an incentive to move more products to clearing. It will also induce additional substitution away from non-cleared products to cleared ones. Differential margin and capital requirements are a far more direct way to affect the incentives to utilize clearing than aggregate ownership limits, and do so without affecting CCP capitalization, management and governance.
DCMs and SEFs
The Division’s analysis of the relation between ownership structure and market power for DCMs and SEFs also flies far wide of the mark. It is not reliable guidance for the Commission in its deliberations.
Market power is a fact of life in the execution of financial transactions on centralized markets. This market power arises from the nature of liquidity. Liquidity attracts liquidity, and as a result, trading activity tends to “tip” to a single execution venue unless regulatory policy effectively socializes order flow. The exchange to which order flow tips can exercise market power, by admitting an inefficiently small number of members, or charging super-competitive prices for its services.
In equity markets, multiple trading venues have long existed. Many of these are satellite markets that do not contribute to price discovery, but instead permit verifiably uninformed traders to reduce execution costs. In the United States, multiple venues contribute to price discovery in individual equity options and common stocks. This does not contradict the tipping result I describe above. In both instances, Securities Exchange Commission (“SEC”) regulation has effectively socialized order flow by mandating linkages between execution venues. In stocks, for instance, RegNMS requires a market receiving an order to route it to the venue offering the best price. This has created a single market (except in turbulent periods, when linkages can break) that is not subject to the tipping effects that would occur when execution venues not obligated to route orders to better-priced markets compete for order flow. Note that the market share of the New York Stock Exchange plummeted from the mid-80 percent range (and virtually 100 percent of the orders that contributed to price discovery) pre-RegNMS to the low-20 percent level post-RegNMS. In its analysis, the Antitrust Division completely ignores these basic features of competition in trading markets, - an omission that renders its comparisons between derivatives and equity markets meaningless.
In fact, there is no connection between trade execution venue ownership structure and market power. Mutual, not-for-profit exchanges (the dominant form in the open outcry era) had market power. In the US, these mutuals were not dominated by large financial institutions, or by any single type of firm. For instance, US futures exchanges had diverse members, including locals, floor brokers, and Futures Commission Merchants (“FCMs”) of varying sizes, some owned by large financial institutions, but many not. Presently, in both the US and around the world, demutualized, investor-owned exchanges have market power. These entities typically do not have ownership dominated by large financial institutions.
Competition between trading venues is impeded by the tipping phenomenon discussed above. Ironically, many of the major attempts to compete with incumbent exchanges that have occurred in recent years were driven by groups of large banks or single financial firms. These include: (a) BrokerTec which attempted to compete with the Chicago Board of Trade in trading Treasury futures; BrokerTec was founded by 14 large financial institutions; (b) The Cantor Futures Exchange, an initiative of interdealer broker Cantor-Fitzgerald, also entered the Treasury futures market, (c) ELX, an exchange launched by a consortium of large financial institutions and market making firms, and (d) Turquoise, a European equity trading platform backed by investment banks to compete with incumbent exchanges like the London Stock Exchange.
Most of these attempts were failures, due to the difficulties of wresting order flow from incumbent exchanges. However, the most notably successful displacement of an incumbent market by a competitor involved a bank-dominated exchange (Deutsche Terminbörse, now Eurex) wresting the Bund futures market from LIFFE.
Thus, nascent exchanges predominately owned by large financial intermediaries have been the main source of competition in derivatives and equity markets in the past decade and more. It would particularly perverse for the Commission to adopt a policy in the name of promoting competition that would hamstring the most likely source of competition for dominant execution venues.
I reiterate that the market power that exists in the market for the execution of transactions in derivatives and equities derives from the nature of liquidity. It does not derive from the ownership structure of these venues. The most likely effect of aggregate limits on ownership would be to affect who captures any profits relating to market power, not the amount of market power that exists.
That is, aggregate limits are not an effective policy for encouraging greater competition in derivatives markets. Indeed, they can actually hamper competition.
Summary and Conclusions
In summary: limits on the aggregate share of DCOs, DCMs, and SEFs that any type of firm can own are extremely inadvisable. With respect to DCOs, they threaten the safety and soundness of these vital elements of financial market infrastructure. They will not materially increase competition in clearing or derivatives execution. With respect to DCMs and SEFs, aggregate limits will not increase the competitiveness of derivatives markets, and are in fact more likely to reduce competition.
Sincerely yours etc.
Craig Pirrong
Professor of Finance
Energy Markets Director, Global Energy Management Institute
