March 30, 2001
Dean Joel Seligman
Washington University School of Law
Dear Dean Seligman:
As requested of members of the SEC Advisory Committee on Market Information (the "Advisory Committee"), Reuters Group PLC ("Reuters") is pleased to respond to the questions put forth regarding implementation of a new market data dissemination system.
As outlined in our previous submission to the Advisory Committee on December 4, 2000, Reuters advocates a framework for developing a new market data dissemination system that will fulfill the mandate of Congress as established in 1975, while at the same time reap the benefits of the vast technological improvements of the last twenty-five years. While we propose an "alternative" to the existing monopoly consolidation system that favors competition, innovation and interplay with market forces, our proposal seeks to advance - and not to serve as an alternative to - the policy mandate established by Congress in enacting the National Market System provisions of the Securities Exchange Act of 1934.
Central to Reuters' proposal is the introduction of competition into the marketplace for data dissemination, without abandoning regulatory structures viewed as essential to ensure adequate consolidation. The remainder of this letter seeks to identify, within the framework of this objective, our specific responses to the questions forwarded by the SEC staff regarding our proposal. As noted in our prior submission, Reuters has sought to identify key elements of a constructive framework that could serve as a starting point for further reform; we recognize that the exposition of the details of any change to the current system, whether under an "incremental" or "alternative" approach, will require additional careful consideration by the Advisory Committee.
Your model contemplates broker-dealers making their data available to at least one consolidator, which would then have an obligation to make its data available to all, including competing consolidators, on a non-discriminatory basis. To complete its consolidated data feed, would a consolidator be required to purchase from its competitors their entire data feeds, or just the broker-dealer information it does not have? If the former, isn't that inefficient? If the latter (i.e., unconsolidated broker-dealer data), how would the non-discriminatory price be determined? Would there be a fixed charge for each broker-dealer feed, or would the charge vary depending on the level of trading of the broker-dealer? And regardless, wouldn't the SEC still need to play an active oversight role to ensure that pricing abuses don't occur?
Under the proposal outlined in our previous submission, broker-dealers would be required to submit their market data to a consolidator, or to several consolidators, based on arrangements negotiated by the parties through the market. Each data consolidator would, in turn, be required to generate a consolidated steam of data. This stream would be generated from the data that the consolidator receives directly from broker-dealers, combined with data that is acquired from other consolidators.
In order to complete its consolidated data stream, each consolidator would be obligated to make data available to other consolidators on a non-discriminatory basis. As a threshold matter, as suggested in the question above, it does not appear necessary in this context to require each consolidator to purchase the entire stream from other consolidators, but rather only the data that it would not otherwise be able to include in its data stream.
In our view, the most effective approach would be to establish a general non-discrimination requirement comparable to the obligations that currently arise in other contexts. Regulation ATS, most notably, sets out a non-discriminatory access requirement applicable to ATSs that exceed a specified market share, without explicitly creating a rigid formula. As in the case of Regulation ATS, the SEC could require each consolidator to establish standards for access by competing consolidators - and provide for case-by-case SEC oversight in the event of non-compliance with this mandate.
In the first instance, a general non-discrimination requirement would look to individual consolidators to negotiate appropriate pricing based on a full range of market considerations. In addition to ordinary market considerations, the regulatory obligation of each consolidator to acquire others' data would establish strong incentives to price data in a reasonable manner - since no consolidator would be permitted to disseminate data without acquiring every other consolidator's data.
At the same time, a general non-discrimination requirement could be effectively supplemented through SEC oversight. The SEC would play an important role in ensuring that particular consolidators do not "hold up" competitors in circumstances where market and regulatory incentives fail to produce a privately negotiated price. The SEC could, moreover, establish sufficiently severe penalties for failure to reach a non-discriminatory price in private negotiations (e.g., disqualification from consolidator status) that consolidators would rarely need to resort to it for resolution of disputes.
Assuming that each consolidator has an obligation to acquire only the data it has not acquired directly from broker-dealers and other market centers, the determination of a non-discriminatory price could occur in a variety of ways. For example, as suggested in the question above, the Securities and Exchange Commission (the "SEC") could establish a fixed charge or a formula upon which different consolidators would interact with each other. In practice, however, Reuters considers it likely that overly detailed formulas could lead to market distortions and excessively complex regulations, especially if all the nuances of different consolidator's services are taken into account.
Regardless of whether a general or detailed approach is taken in establishing a non-discrimination requirement, the SEC may wish to seek to establish certain minimum time periods over which consolidators would be expected to obtain commitments for access to the data necessary to generate a consolidated feed. These periods presumably should have a length sufficient to provide the stability necessary to justify capital investment, as well as to minimize unnecessarily frequent changes in consolidation mechanisms and any potential burden on SEC resources.
With consolidated data being generated through individual feeds from broker-dealers to various consolidators, and then among the various consolidators, won't there be practical difficulties in ensuring that high-quality, accurate and timely consolidated data is available from all sources? And how would you avoid double-counting?
The introduction of competition will not introduce, in our view, increased risks with regard to the quality, accuracy or timeliness of data dissemination. As stated by the New York Stock Exchange (the "NYSE"), "today's technology permits multiple entities to receive simultaneously multiple streams of data and to create consolidated outputs that sequence prices and quotes in the same order." A competitive regime will not raise any novel dangers that do not arise in the current scheme for consolidation - particularly if supplemented by SEC "backstop" oversight (e.g., as to standard-setting).
Indeed, experience suggests that the existing monopoly regime poses a much more substantial threat to the development of high-quality, accurate and timely data than any competitive approach. Under existing rules, monopoly providers have little incentive to update their systems effectively or to take steps to reduce unnecessary use of available vendor systems. Recent experiences in the options market - where plan participants have failed to act on a timely basis to update systems and have sought to shift their problems to vendors such as Reuters - only confirm the dangers associated with a monopoly mechanism. Indeed, the technology of many facets of the National Market System has perennially failed to keep up with the needs of industry participants, notwithstanding active efforts by the SEC to push that technology forward. This should not be surprising - monopolists have limited incentives to provide superior service and vendors cannot recoup capital investments due to regulations that preclude them from competing effectively with their suppliers.
Moreover, competition - and not regulatory monopoly - has become the rule, rather than the exception, in other securities markets around the world, as well as in commodities and other financial markets in the United States. Those markets demonstrate overwhelmingly the feasibility of data dissemination based on competitive market forces.
On the specific point relating to double-counting, each consolidator would be expected to develop a system of tracking individual broker-dealer feeds. These systems are already available and in use. Further development of existing technology will address double-counting and ensure that no part of the feed is missing from the consolidated product.
In your model, the role of the SROs appears limited to competing as potential consolidators, with market data fees going to broker-dealers. Do you believe the exchanges, for example, have no ownership rights in the data produced in their markets?
In our model, the ownership rights in data would be established through contractual bargaining between market participants. Broker-dealers would have the ability to bargain with multiple consolidators and presumably would be able to obtain some payment or other consideration for their data from one or more consolidators. Self-Regulatory Organizations ("SROs"), like other market participants, could compete to act as consolidators. In addition, even if an SRO did not elect to act as a consolidator, it could acquire an interest in the data aggregated in its marketplace and offer that aggregated data to one or more consolidators for value. For example, an exchange could negotiate with its members for the right to disseminate their data on an individualized or aggregated basis. Of course, SROs would continue to receive data necessary to perform their regulatory function, but it would not presumed that they could use members' data for commercial purposes.
Outside the context of National Market System securities, market participants routinely enter into contractual arrangements allocating property rights in data among multiple entities. For example, electronic trading systems commonly acquire property rights in aggregated, anonymous data for trading through their systems, even while individual users of the system retain complete rights in data relating to their individual trades. In our experience, individual negotiations in a competitive marketplace tend to increase the efficiency and breadth of data dissemination, as well as to grant appropriate value to the rights of respective market participants.
If SRO market data revenues are substantially reduced, how will the SROs continue to fund their regulatory function?
SROs, in our view, should fund their regulatory function through direct charges to users of the SRO's services. Under existing rules, SROs already levy a variety of fees on their members, both directly and in connection with transactions effected through the facilities of the SROs. The subsidization of regulatory functions with market data revenues imposes the cost of regulation in an unpredictable and inefficient way. Moreover, Congress has not given the SEC adequate standards for measuring - or even a clear mandate for creating - subsidies to SROs.
In opposing the use of market data to subsidize SROs, Reuters does not dispute or minimize the significance of their regulatory functions. Nevertheless, market data subsidies tend to reduce the accountability of SROs to those paying for their services. In addition, the effective performance of regulatory functions provides SROs with substantial reputational value that should not be supported by hidden subsidies. Especially in an environment of increased competition among for-profit SROs and other market participants, the SEC should strive to eliminate regulations that could effectively benefit the shareholders of one class of entities at the expense of shareholders of another.
Your proposal is a bit unclear on the degree of consolidation to be required by the SEC, and at one point you refer to the SEC requiring consolidated data only from the "principal" trading markets for a security. Please explain what you mean by this.
In its previous submission, Reuters wrote that market data should be consolidated from "principal" trading markets. This was intended to be a reference to the fact that in any new data consolidation system - as in the current system - not all market data can actually be consolidated. For example, some transactions occur after regular trading hours or directly between market participants that are not broker-dealers; others occur overseas in venues outside the jurisdiction of the SEC. Any new system, like the existing system, should contain exceptions for data whose collection exceeds the SEC's jurisdiction or whose value is de minimis.
You also indicate that, in time, there may no longer be a need for mandatory consolidation. Please elaborate. Also, please respond to each of the questions asked the NYSE under the heading, "If you believe that consolidated data need not be provided."
As set out in our prior submission to the Advisory Committee, Reuters proposes an approach that would introduce competitive forces for dissemination of market data, but would continue in the near term to mandate consolidation for the NBBO only. In our view, the evolution of the market may lead to a number of developments that ultimately render the obligation to consolidate either redundant or counterproductive. For example, once competition has been introduced to the market, a high degree of marketwide transparency may arise within a framework of multiple competing data sources, and the SEC may conclude that the costs of imposing regulation are no longer necessary. Moreover, potential changes in market structure and the rapid growth of the amount of market data could make consolidation of all available data less useful or even harmful. If, for example, market participants adopt increasingly small trading increments, and the volume of message traffic grows very quickly, market participants may greatly prefer a system that excludes data whose costs (in terms of its use of available bandwidth and proprietary computer systems) exceeds its value (because it may represent a stream of data of fundamentally little importance, such as 100 share quotes that improve the best bid or offer by only a fraction of a penny).
In practice, we consider it preferable to focus at present on how to introduce competitive forces even while retaining the requirement of mandatory consolidation. The SEC and market participants will have adequate opportunity going forward to assess whether the consolidation requirement can be relaxed or eliminated, and can do so in a manner that will take into account the questions cited above that have been asked of the NYSE.
* * *
We appreciate the opportunity to respond to questions forwarded by the SEC staff and look forward to discussing the Reuters proposal, and the proposals of other Advisory Committee participants at upcoming meetings. If you should have any questions or would like to discuss any aspect of this letter further, please do not hesitate to contact me at (212) 593-5550 or Mitchell Feuer, Reuters America Vice President for Government and Regulatory Affairs, at (202) 898-8343.
cc: Annette Nazareth, Esq.