Jonathan G. Katz,
Securities and Exchange Commission,
450 Fifth Street, NW,
Washington, DC 20549,
Your Ref: Release No. 34-39884; File No. S7-12-98
28th July 1998
Proposed Rules on the Regulation of Exchanges and Alternative Trading Systems
Dear Mr. Katz,
Please find below three very brief comments on the SECs proposed new rules on the Regulation of Exchanges and Alternative Trading Systems (ATSs). They examine, respectively, the topics of exchange governance, transparency, and the appropriate strategy for regulating trading systems. These comments reflect my personal opinions, and not necessarily those of the Oxford Finance Group, its clients, or any co-authors of mine. The comments draw heavily on "What is an Exchange? The Automation, Management and Regulation of Financial Markets", Ruben Lee, Oxford University Press (forthcoming 9/1988), and "The Legal Basis for Stock Exchanges: The Classification and Regulation of Automated Trading Systems", Ian Domowitz & Ruben Lee, Working Paper, Oxford Finance Group and Northwestern University (5/1996). Please do not hesitate to contact me for elaboration of any of the points made below.
A key problem with the proposals concerns how the fair representation requirement should be imposed on trading systems classified as exchanges. Only two organizational outcomes are possible. On the one hand, a trading system may be free to adopt whatever corporate structure it chooses and take decisions on whatever commercial grounds it believes appropriate. This does not imply that the trading system would not take account of the views of all the diverse participants on its market; indeed not to do so is likely to lead to a loss in business. It does imply, however, that the system would not be obliged to take such views into account. On the other hand, a trading system may be required to provide some form of fair representation to the various participants on its market. This does not imply that all decisions relevant for the management of the trading system would be taken on a non-commercial basis. It does imply, however, that there may be circumstances when as a result of the fair representation requirements, the trading system would be required to pursue a course of action that it would not choose to undertake on commercial grounds. The problem is that either the governance constraints imposed on a trading system bind, or they do not. To impose constraints on the manner on which trading systems should be governed may be to undermine the very factor which leads to their efficiency and innovativeness.
While the statutory membership and governance requirements the SEC proposes to maintain on exchanges may lead to efficient market structures, there is no a priori reason that they should do so. On the contrary, the imposition of a specified form of governance and membership structure is likely to obstruct market developments, and restrict competition. In a market with rivalry between trading systems, regulation of the governance of exchanges that is imposed in order to influence their market structure is unnecessary. If, as a result of its governance structure, an exchange takes a decision concerning its market that is not in the interests of the traders using its systems, they can divert their order flow to a competing exchange. The owners of an exchange therefore face the correct incentive to create a suitable governance structure, even if this means giving up some of the rights associated with the control of the exchange to other market participants. Any regulatory constraints on governance can therefore only restrict an exchange's freedom to deliver an optimal governance structure. Even when trading systems do not face competitors, market participants will normally have appropriate incentives to agree a governance structure that maximizes efficiency. Only if they do not, is regulatory interference warranted.
The SEC has consistently been one of the most influential proponents of mandatory price and quote transparency in the world. The SEC believes, in essence, that transparency promotes almost all the key goals that regulation of the securities markets should seek to deliver, namely investor protection, fairness, competition, market efficiency, liquidity, market integrity and investor confidence. It therefore argues that mandatory price and quote transparency should be a fundamental and universal cornerstone of the regulation of exchanges and markets, and the current proposals extend its historical policy of intervening in market structure in order to achieve this. I believe that such a policy is inappropriate.
Contrary to what the SEC implicitly claims, there may be significant trade-offs between the attainment of different regulatory goals. An assessment of the relative importance of such objectives is therefore necessary. While the SEC has not explicitly stated a preference between different regulatory goals, and indeed is statutorily required to follow competing objectives, many of its statements and its actions indicate that foremost amongst all the goals it pursues is that of investor protection, and particularly the protection of retail investors. Not only is this a worthy goal, mandatory transparency may further its realization, by helping retail investors monitor the quality of their executions, and by reducing the inequalities in access to information that they face compared to other types of market participants.
Despite this, I believe that mandatory transparency is in general an inappropriate regulatory policy for the central reason that it can compromise the delivery of the two key regulatory goals of efficiency and liquidity. Private incentives as opposed to regulatory fiat, will lead to the level of transparency that best furthers these goals, unless competition is impaired. It is notable in this context that the SEC itself allows for the possibility in the rule proposal that institutional investors might be best placed to determine whether their orders on an ATS should be publicly disseminated, or should be shown only to those participants using the ATS. If the regulatory goals of investor protection and fairness are thought more important than efficiency and liquidity, however, or if it is believed that there are no significant trade-offs in the attainment of different goals, a policy of mandatory transparency may be viewed as optimal.
Given the difficulties of enforcing a regulatory regime against the interests of market participants, the ignorance and confusion we have about the effects of transparency, and most importantly given the private incentives that exist to disseminate information, market participants themselves should be left to decide the appropriate level of transparency, unless competition is impaired. Contrary to what the SEC argues, therefore, price and quote transparency should be viewed as a mechanism that may or may not deliver the desired regulatory goals, rather than being seen as a regulatory objective per se. It should not be mandated.
The strategy recommended here as the best way of regulating trading systems, is composed of two prongs: the separation of the regulation of market structure from the regulation of other areas of public concern, and the employment of competition policy to regulate market structure. Given the frequent confusion that exists about the meaning of the phrase "market structure", it is worth repeating that the phrase is used here to refer to the combination of the three market functions of order routing, order execution, and information dissemination.
The first prong of the joint strategy proposed here is that the regulation of market structure should be separated from the regulation of other areas of public concern. The most significant of these are market conduct, in particular the activities of financial intermediaries and other market participants, and market integrity, which relates to issues such as price manipulation and insider trading. The second prong is that competition policy should be employed to regulate market structure. The reason for proposing this joint strategy is to accommodate two beliefs I think are central in determining the most appropriate way of regulating and classifying trading systems. They are, first, that competition is the single most important factor in effecting a market structure that will best deliver the desired regulatory goals, and second, that the regulation of competition between trading systems cannot be well executed via a self-regulatory process.
The claim that competition is the most important factor in effecting an optimal market structure is controversial. There is first of all debate as to whether private incentives are best placed to deliver the goal of efficiency. Although a full examination of their merits in effecting optimal market structure is not presented here, the advantages of allowing market participants to determine appropriate levels of transparency are highlighted above. If the merits of competition in delivering an optimal market structure are accepted, however, a key premise on which the SECs rule proposals are based is undermined. In particular, it cannot then be argued that the perceived flaws associated with ATSs should best be addressed by integrating alternative trading systems more effectively into national market system mechanisms.
A further controversial aspect of using competition as a regulatory policy is that even if it is accepted that private incentives do deliver efficiency, there is debate about the extent to which other regulatory goals should be sacrificed in the pursuit of efficiency. No logical resolution of this issue is possible, as it is essentially a matter of preferences. Unlike what the SEC implicitly accepts, however, I believe that significant trade-offs between the attainment of different regulatory goals are commonplace. In order to select any regulatory policy, it is therefore critical to make a ranking by importance of the different regulatory goals. Efficiency is taken to be a key goal here, not least because its delivery is thought to enhance that of fairness and the other desirable goals.
The justification for the second belief underlying the proposed separation, namely that the regulation of competition between trading systems cannot be well monitored via a self-regulatory process, is that the conflict between the interests of the owners and operators of one trading system and those of other trading systems is deemed too direct for self-regulation to be effective. It is certainly possible that an increased amount of trading in one asset on one system may lead to increased trading volume in that asset on other competing systems. More commonly, however, trading systems are viewed as substitutes rather than complements, implying that if one trading system does not compete as aggressively as possible, orders will flow away from it to its competitors. The operators of a trading system therefore typically have an incentive to act anti-competitively, if such an option is available, in order to attract order-flow. Self-regulation is thus insufficient to enforce fair competition between trading systems.
A central implication of the separation approach proposed here is that there should be no distinction in the regulation of any market structure questions between institutions which are classified as exchanges, ATSs, or brokers. If the approach were implemented, ambiguities in the "exchange" definition with regards market structure would no longer matter because the definition would no longer be applicable. Problems concerning market structure arising from the re-classification of trading systems, from the incoherence of the SEC's past approach for classifying trading systems, from the enhanced ATS framework it has proposed, from any competitive advantages accruing to broker-dealers that arise as a result of regulatory diversity, and from the regulatory rigidities placed on exchanges, would also all no longer be relevant.
Two approaches for allocating those regulatory responsibilities apart from market-structure concerns that are currently assigned to SROs may be considered. The first, which is similar to the status quo, would maintain an institutional definition of an SRO, possibly still called an "exchange", and would assign appropriate regulatory duties to any institutions satisfying the definition. Trading systems that did not satisfy the definition would be required to join one of the statutorily defined SROs. Depending on the nature of the definition, the various classificatory difficulties identified above with the current exchange definition may still, however, continue to arise. One way of minimizing them would be to define an SRO using some form of limited volume terminology.
A second more radical approach would be to allow a trading system the freedom to choose which organization should act as its regulator, but to require it to be regulated by one SRO. This would place competitive pressure on competing regulators to provide cheap and efficient regulatory services to the competing markets they served. There would need to be SEC oversight of any institution choosing to be a regulator of trading markets in order to ensure minimum regulatory quality, in much the same manner as the current SEC oversight of existing SROs. Given the economies of scale associated with regulation, both approaches are likely in practice to lead to much the same result.