October 28, 2002

Jonathan G. Katz, Secretary
Securities and Exchange Commission
450 Fifth Street, N.W.
Washington, D.C. 20549-0609

    Re: October 29, 2002 Hearing on Market Structure Issues, File No. 4-466

Dear Mr. Katz:

Market Data Revenue - Financing the SRO Function

The Commission has a limited range of choices in dealing with market data issues. The recent federal budget discussions and the Commission's expanding list of mandates virtually require that the regulatory functions which have traditionally fallen to the SROs continue to be financed in the traditional manner - by the markets. Actual SRO budgets will almost certainly be at a higher level than in the past because the SROs will be called upon to perform additional duties. While we have the NASD/NASDAQ model for increasing independence of the SRO from the market structure and operations before us, NASD and NASDAQ have taken several years to get to their current degree of separation. It would take a number of additional years before anything comparable might be achieved at the NYSE and the AMEX - should SRO independence be a goal.

The issue of SRO independence is not integral to the more fundamental issues of what regulation is to be done and how it is to be financed. It is my understanding that any increase in fees levied directly by the Commission would go to the Treasury and would not be available to the Commission for use as it might direct. It is also my understanding that an SRO or a marketplace could be directed by the Commission to pay for compliance, surveillance and other regulatory functions undertaken by it or on its behalf. Obviously, the SRO and the marketplace need revenue to pay for this regulation.

There are only four significant sources of revenue available to an SRO or a marketplace. Setting aside market data for a moment, we need to examine the other three sources - listing fees, transaction and access charges, and fines.

It is clear from recent history that listing fees (which are partly a payment by the listing entity to support the issuer regulatory functions of the exchange) have been under pressure at all of the exchanges. The market which has the greatest flexibility in this respect - the New York Stock Exchange - has substantially reduced its listing fees and extracted compensating payments from the floor, specifically the NYSE specialist units. In a highly competitive market environment for primary listings and related order flow, the scope for raising listing fees to cover SRO regulatory functions related to listed companies would be limited without a Commission mandate setting listing fees for each primary listing exchange.

Transaction-related charges (TCs) are of two basic types: (a) fixed charges which may take the form of periodic payments for member access or some other service performed by an exchange for its members and (b) variable TCs levied on the basis of number of trades or number of shares or some combination that reflects the volume in a marketplace. As with listing fees, transaction charges of both the fixed and variable variety have been under pressure and appear unlikely to serve as a source of significant SRO revenue growth, given the already extensive competition among SROs for primary market listings and, more importantly, for order flow.

Fines have not been a significant source of SRO revenue in the past. Under any circumstances, they would be unreliable in timing and magnitude as a revenue source and any large fine or systematic increase in fines levied by an SRO would be subject to appeal to the Commission and to the courts.

The only significant remaining source of revenue which is available to the SROs is market data revenue. Setting aside the niceties of who owns market data, the effective functioning of the SRO system seems to depend in substantial measure on the ability to finance SRO activities with market data revenues. I offer for the Commission's consideration some incomplete and slightly inchoate thoughts on how market data problems might be resolved in a constructive way.

I suggest the Commission explore market-oriented solutions to the collection of market data revenue and its distribution to SROs/producers of market data. I have long been appalled at the structure and the impact of the Consolidated Tape Association Agreement on the entire range of market structure issues we will discuss today. If this dysfunctional agreement had not been implemented as part of the National Market System or if it had existed in a different form, many of the controversies we will be discussing would never have occurred or would have been resolved quickly with a market-based solution. The fact that the revenue collected from the NYSE, AMEX and NASD tapes bears no meaningful relationship to the volume of business or the value of the market information available from each participating marketplace is ample evidence of some flexibility in setting pricing policies. This is not the classic pricing of perfect competition. Consequently, there is some risk in adopting a totally free market policy. The principal reason market data has value is that it is a true monopoly (or oligopoly) product. Its value comes from an economic rent which is available only to the organization(s) which provide the market data. There would have to be limitations on any price increases that could be extracted from market data. The Commission would undoubtedly be under pressure to impose a ceiling on charges for certain types of market data users. If such constraints became too numerous or complex, the whole notion of opening this process up to market forces would be scuttled very quickly.

The Commission is probably best advised to set an overall revenue ceiling with a budgeted share for each SRO to cover its regulatory mandates and let the remaining market data revenue be apportioned to each marketplace by market forces. Market forces will determine what data will be purchased by each user and at what price. A wider range of services provided by various traditional "vendors" and by services distributing market data via the Internet could permit vendors and users to customize data selection in a variety of ways. There would be no exclusives granted to any market data consolidator or vendor, but everyone providing or offering a specific level of service would pay the same rate per similar user to the market data producing entity. If there was no demand for data from crossing sessions or a particular type of ECN, for example, these markets would not share in the revenue. The markets might determine that there was no significant price discovery and no utility in having certain price information as part of an information package. A wide range of service levels could be achieved by, for example, having some price reporting services that would cover only the primary market and others that would provide delayed quotes with a directional change indicator to indicate that the price or quote had gone up or down since the posted figure was generated.

A primary-market-only service could indicate if a non-reported market currently had a better quote or if a substantially larger size was available at the primary market quote or better, permitting the market data user to escalate his inquiry to a higher level and higher cost source of market data to get more information. It is clearly in the interest of the markets - as opposed to the SROs - to keep consumer costs of market data as low as possible, though many markets will continue to depend on market data for a significant fraction of their operating expenses.

Changing the approach to distributing market data is a complicated process, and we are at a disadvantage in starting a long way from market equilibrium. Nonetheless, something like this will be necessary eventually. The sooner we start, the less painful the transition. Integrating some variant of this proposal with some of the changes necessary to deal with transaction charges implemented by various exchanges and ECNs will add complexity, but an attempt at integration is probably worth the effort.

I have not described how innovation in market structure will be accommodated. Actually, integrating innovative market mechanisms will be relatively simple. A new market will attract market data revenue by persuading a vendor or a group of end users that the new market's data has value. The days of silly deals between markets and regional exchanges acting as pass-through agents to get a piece of the market data revenue stream will be over.

                Sincerely,

                Gary L. Gastineau
                Managing Director