CONCEPT RELEASE #34-42208 (FILE #S7-28-99):



Summary of American Stock Exchange Position

The American Stock Exchange ("AMEX") provides below in Part I a brief summarization of its position vis a vis the Securities and Exchange Commission's ("SEC" or "Commission") questions in its Concept Release No. 34-42208; File No. S7-28-99: Regulation of Market Information Fees and Revenues.

An elaboration of these views can be found in the "Exposition of American Stock Exchange Views" in Part II. A question-by-question series of replies, and explanations of those responses, to each of the Commission's 62 questions can be found in Part III.

The SEC's questions fall into six broad categories, and the AMEX' comments in Parts I-III follow that general pattern.


The AMEX strongly believes the method of setting SRO market information fees, proven effective for over 24 years, should not be changed.


Moving to a cost-based system of setting SRO market information prices will deliver few benefits and will cause many problems, including (but not limited to) an enlarged bureaucracy for all parties, vast amounts of paperwork, and continual acrimony over the veracity of the data and the accuracy of the conclusions reached by the SROs or the Commission.


If the SEC mandates a move to cost-based pricing for SRO-generated market information, the AMEX will functionally be able to participate.


If the SEC mandates a move to cost-based pricing for the SROs' market information, additional categories of costs beyond those of market regulation and market operations that the SEC believes appropriate must be included in any fee-setting formula, specifically, member regulation, administration and related occupancy costs, and a cost-of-capital allowance.


If the SEC mandates a move to cost-based pricing for SRO market information, the percentage of costs to be included as it relates to the creation of market information may differ from (and be higher than) those the SEC believes appropriate; any percentage chosen should be standard across all SROs in any given product, but will have to be different for options than for equities due to the unique characteristics of options costs relative to those of equities.


If the SEC mandates a move to cost-based pricing for SRO-generated market information, a rate of return similar to the returns on capital ("ROC") or returns on equity ("ROE") earned by broker-dealers, or some other appropriate proxy must be permitted.


Market information fees charged by SROs for professionals are extremely good value, both absolutely in terms of the amount of product delivered for the price, as well as relative to the costs of broker-dealers in general.


On any basis examined, whether relative to other financial industry service fees or relative to the amount of time market information is monitored by either securities industry professionals or retail investors, recently-instituted SRO market information fees for retail (or online) investors are in the range of prices charged to professionals, represent extremely good value, and do not inhibit the investors themselves or their brokers from providing needed market information.


Premium prices for SRO-generated market information paid by parties that are not members of the SROs in a Plan, compared to prices charged to those that are members, are justified and should be maintained.


SRO fees to vendors are much too low and should be raised substantially.


All volume discounts, or tiered pricing, should be dropped; alternative pricing methods to monthly per-service fees, such as per-query or enterprise license fees, however, should be permitted.


The method of distributing market information revenue among the SROs should be changed to compensate SROs that offer the "quality market" that does the most to advance all of the objectives of the Exchange Act.


The new market information revenue distribution method should, in order to promote the fair and competitive markets for which the Exchange Act calls, ignore the number of trades, which is a meaningless statistic; Plan revenues should be distributed based on a formula that equally weights the market value of a security traded and the quality of a market's quotations in that security (weighted by the market value of the shares or contracts represented in the bid or offer and the amount of time the bid or offer is the National Best Bid ("NBB") or National Best Offer ("NBO").


No costs of an SRO should be recovered through the Direct Distribution of a Plan's market information revenues, except those now recovered by a Plan's Administrator; if cost-based market information fee-setting for SROs were mandated, limited costs related to the additional audits that would be necessary to produce the needed figures might be considered.


Non-SRO parties should have no vote in the Operating Committees of the Plans.


If a vote for non-SRO parties were mandated by the Commission, the AMEX would recommend that one representative for broker-dealers, one for professional investors, and one for retail (or online) investors should be allowed a vote; vendors should have no vote.


Non-SRO parties, including vendors, could participate in the Operating Committees of the Plans as observers, providing ideas and communicating the discussions and results of the meetings back to their constituencies.


No new committees should be mandated for either formal or informal oversight of the SROs' market information business.


The market information administration process (initiating and canceling services, contract terms and administration, and auditing to ensure that users are being properly billed for services used) should be simplified and streamlined on an industry-wide basis.


Pilot programs should be permitted, but on a more restricted basis regarding the number in effect, and their duration; pilot programs should go through an "effective on filing" process, which gives the SEC an opportunity to halt or rescind a pilot program if objections are raised that warrant such action.


Member rebates should be permitted, because determining whether a fee reduction, elimination or rebate, was appropriate as a legitimate competitive response to the changes in the industry, and the degree to which the SEC should interfere in an SRO's business strategies, are all questions that will result in serious, and probably unintended, consequences for the Commission, the SROs, and investors.


If the SEC mandates a move to cost-based pricing for SRO-generated market information, member rebates might be permitted only after an SRO has "earned back" (as described in #13 above) all the market information-related costs it submitted as part of the fee-setting process, though this limitation could entail the same problems as highlighted in #21 above.


The amount of public disclosure required of SROs and the Plans should not be increased; any degree of increased disclosure should be strictly for the use of the SEC and should be protected in order to not disclose business strategies of an SRO to its competitors, though whether that is possible in practice appears problematic.


Exposition of American Stock Exchange Views

The SEC requested comment on 62 separate, but often related, questions. Individual replies, and the reasons for those answers, are provided in Part III of this document. This section, Part II, provides a less specific look at, but deeper explanatory narrative for, the positions the AMEX is taking, as outlined in Part I, on the Concept Release's questions.

Market Information Revenues Are Crucial to the SROs

As the SEC stated clearly in the Release, a position with which the AMEX agrees, market information is the property of the SROs that produce it. It plays a vital role in funding many of the activities and investments that allow the SROs to fulfill their functions as SROs and to provide the U.S. with the world's largest, most liquid, and best-regulated capital market. The AMEX does not believe there is any dichotomy in the principle of market information's being the private property of the SROs, but whose prices are regulated for the public good.

Moving to Cost-Based SRO Market Information Pricing Would Be a Mistake

In the Release, the Commission asks for comment on ten questions relating to the basis for setting the prices for SRO-generated market information, focusing specifically on moving to a cost-based approach. The AMEX believes strongly that the fee-setting mechanism currently employed has worked exceedingly well for nearly 25 years and urges the SEC not to change the method for setting market information fees.1

Although a cost-based approach may appear to offer benefits, especially by providing a more specific justification for whatever price is set by each of the Plans, it is just this specificity that is likely to cause major difficulties. Thus, the "improvements" in the SRO-market information fee-setting process are likely to prove chimerical. Instead of helping to quell debate, a cost-based route will take the Commission and the capital markets down a road already abandoned by other industries and their regulators.2

As will be reviewed later in Part II, the SEC is focusing its concern on the wrong part of the market information industry when questioning the costs to broker-dealers and investors. Over 95% of the costs of market information derive from the vendor side of the industry, but vendors' pricing does not fall under the Commission's oversight.3 A fundamental question of fairness and equity is raised if the SEC decides to pursue a course negative to the SROs when whatever problem may exist is not due to the SROs, but to vendors, and is not within the SROs' capacity to control.

Cost-Based SRO Market Information Pricing Will Be a Bottomless Pit of Trouble

It is hard to imagine that the Commission wishes to institute in the securities industry an era similar to that seen, for example, in the electric utility or telephone industries before deregulation.4 That time was marked by acrimony, enormous paperwork, lengthy hearings and appeals, and not infrequently, court battles.5 Though the specific issues would be different, the outcome of moving to a cost-based system of setting SRO market information fees is likely to be the same.

Using costs to set market information rates will, in all probability, force the SEC to hold hearings annually. Otherwise, at least one of the myriad contending factions will claim that the revenues it is paying (or receiving) are not justified by the cost figures submitted. Those figures, both as to what should be included and how they were derived, will be an endless source of friction and disagreement.

The end result is that, contrary to the efforts and desires of Congress, the public, and the political process in general over the last decade, the Commission risks initiating a process that represents a quantum leap in the degree of regulation and the size of the SEC's bureaucracy. It is impossible at this juncture to estimate how much the Commission's staff will have to grow in order to hold what may prove to be almost literally endless hearings, to analyze the mountains of data that all sides will produce to support their positions, and to deal with appeals and probable court challenges to decisions the SEC hands down. It will surely be more than minimal.

The Cure May Prove Worse Than the Illness

However, the problems go far beyond an increase in staff for all concerned, or philosophical debates over "to regulate or not to regulate." All stages of trying to set SRO market information fees based solely on a set of pre-determined cost categories lend themselves to fomenting arguments: determining what expenses should be included in the gross common cost pool, how they are being accounted for, which of them relate to equity and which to options, what percentage of them should filter down to the net common cost pool, how equity charges are divided among the three equity plans, etc. These issues, and those cited in the immediately preceding paragraphs, represent only a partial list of the problems the SEC will be confronting if it decides to open the Pandora's Box that heretofore it has wisely declined to open.

In view of all the risks inherent in moving to cost-based pricing for SRO market information (which appears to have minimal support from either producers or users of market information),6 there seems little reason to change a system that has worked smoothly and with only minor disputes for 24 years. This is particularly true when, as the Concept Release notes, SRO-produced market information accounts for less than one quarter of one percent of the costs of the securities industry, by far the user that bears the greatest part of the cost for the product.

Cost-Based Market Information Pricing Is Functionally Possible

Though not desirable, nor desired, cost-based pricing of SRO market information is functionally possible, and the AMEX will be able participate in such a system if the Commission mandates it. The AMEX firmly believes, however, that if the SEC forces a move to cost-based pricing of SRO-generated market information, even if it were not precise "down to the last penny," the method of allocating the revenues the SROs receive from market information should not be related to the costs the SROs submit in the rate-setting process. The Administrator of each Plan, of course, should still recoup its costs of administering the Plan, and the charges from the Processor, from the Direct Distribution portion of the Plan's revenue. The SROs in general, however, should "earn" any revenue they receive, as described in detail below.

Additional Cost Categories Must Be Included

In its discussion in the Release about setting SRO market information fees in a cost-based environment, the SEC indicated that it feels that only the cost of market regulation and market operations are justifiable SRO expense categories, and only to the degree that they relate to market information. The AMEX believes such a narrow prescription is not correct, and that any cost-based formula, if it were imposed, should include member regulation, as well as administrative and occupancy costs as they relate to market regulation, member regulation, and market operation. In addition, the AMEX believes a cost-of-capital and rate of return allowance must both be allowed.

Member Regulation Is Indistinguishable from Market Regulation

Member regulation appears to have been excluded solely because it has the word "member" in it. However, member regulation is not an activity that benefits the SRO's, members more than investors in any way, and is not different from market regulation in its ability to protect all market participants from abuses that can cost them money. At the AMEX, those in member regulation monitor the financial condition of members and examine their sales practices. The services performed by member regulation are no less central to the maintenance of a fair and honest market, and the ability to produce high quality and reliable market information, than is true for the services performed by market regulation.

Imagine a market in which members trade, knowing they cannot financially complete the transaction. That market's counter-parties are no less being defrauded than those who trade in a market where the participants engage in insider trading, front-running, painting the tape, etc. Member regulation prevents the former (financial fraud); market regulation prevents the latter (market fraud). However, fraud is fraud, and the damage done to investors by either type has the same practical outcome. To say that market information fees can be justified by one (market regulation) but not by its mirror twin (member regulation) is not to carry out the mandate of the Exchange Act to its fullest.

Similarly, examining the sales practices of its member firms allows an SRO to prevent fraudulent or deceptive practices from being perpetrated on the public. Sales is only the early part of the trading continuum, and to permit unlawful or unethical behavior in any part of the process, sales as well as trading, will have the same negative effect on the investor. If it is the job of the SROs under the Exchange Act to prevent behavior deleterious to investors, then to disallow the expense of member regulation from counting toward the cost of market information would appear to be less than a full exploitation of the powers at hand to enforce the Act's provisions.

Administrative Costs Support Functional Areas

As a simple matter of good accounting practice, the costs associated with the areas whose work ensures high quality and reliable market information must include an appropriate allocation of occupancy and administrative overhead. The following paragraphs of this section, and Appendix AA, break Occupancy costs out separately from Administration; in fact, many organizations may include an Occupancy and Administration charge as part of the cost of operating divisions (like Market Regulation, Member Regulation and Market Operations), or may include either Occupancy or Administration as part of the operating divisions, and break out the other. If one or the other is so included, then the breakout of those categories shown in Appendix AA can be ignored; otherwise, they should be used.

Any functional area (market regulation, member regulation, and market operations) included in the cost equation for setting market information fees must be supported by an administrative infrastructure. If there were no organization, including human resources, finance, legal, communications, purchasing, and the senior executives, to support and provide guidance to these areas, it would not be possible for market regulation, member regulation, and market operations to function.

Similarly, if there were no occupancy costs, such as rent or mortgage (or an imputed per-foot cost to represent them); cleaning, janitorial, and other building services; and real estate taxes, none of the functional areas could operate, as there would be no locale from which to operate. The principle governing the need for these administrative costs applies equally to floor-based, non-floor, and virtual SROs, even if the dollars involved are not equal.

A Cost-of-Capital Allowance Should Be Included

The last major element that must be included in constructing a gross common cost pool is a cost-of-capital allowance. Just as changes have occurred in the SRO environment, so, too, have there been changes in the accounting profession's recognition of what are "best practices" for a company to use when allocating capital in its subsidiaries. In fact, it has become common for companies and their auditors to recognize that investment of capital, whether debt or equity, carries its own cost, and on an internal (parent/subsidiary) basis that cost of capital must be charged as part of "best practices" management.7

Capital costs are acknowledged, of course, through depreciation of items purchased and capitalized because of the length of their useful lives. However, there is also another cost that must be recognized: the initial cost of the invested capital itself and the rate of return that it could otherwise generate. Thus, there is not just the cost of the asset to be depreciated; there is also the cost of debt or equity used to finance the acquisition of the capitalized item, or a foregone return by having used cash that had previously accumulated.

Not Including All Suggested Items Will Lead to Inaccurate Pricing

For the reasons detailed above, the AMEX urges the Commission to include in the gross common cost pool market regulation, member regulation, market operations, all related administrative costs, and the cost of invested capital employed in those areas. If all these are not included in a cost-based system of setting SRO market information fees, proper levels of charges will not be set, and SROs will be asked to provide reliable and untainted market information while not being granted adequate resources to do so.

Appendix AA Provides Some Guidance

It may now be useful to turn to Appendix AA. Were the SEC to determine that cost-based pricing of market information must be mandated, this is a template that could be used to implement such a system. It illustrates the apparent relative simplicity of using a cost-based mechanism to set market information fees, but it also allows a focused discussion on the real dangers to which the SEC will expose itself, the SROs, and the entire market system of the U.S. by imposing cost-justified pricing.

Appendix AA is not a representation of any one SRO, but could stand for almost any that exist or that are contemplated. The one addition that some SROs will need is a set of columns at the end of the spreadsheet allowing an SRO to apportion its equity costs among the Plans whose securities it trades if it trades securities on an Unlisted Trading Privilege ("UTP") basis in more than one equity Plan. However, Appendix AA is instructive as an exercise to understand the few positives and the many negatives of pricing market information on a cost-justified basis.

Even Assigning Costs to Equity vs. Options Is Not Easy

Assuming agreement among all parties can be reached on what categories should constitute the gross common cost pool, a prime example of an issue that could lead to considerable friction is how to take the gross common cost pool and divide it between equity-related costs and those that are options-related.

Three formulas offered that could be used to do so are shown in Appendix AA. These formulas are probably as simple as can be constructed, but nonetheless leave more than enough room for subjective decisions and imprecision. Unfortunately, whatever formulas or algorithms are used are likely to spark disagreements among the various parties to the pricing debate.

Developing the Net Common Cost Pool May Be an Achilles' Heel

However, the difficulty of assigning costs between equity and options pales in comparison with another issue likely to prove to be a source of endless argument, acrimony, appeals, and actions in court.

The SEC has postulated, correctly in theory, that any item included in the gross common cost pool is a function of producing market information, attracting transactions, and developing and maintaining listings. The primary problem, of course, is how to assign the proper percentage allocations to each of these three efforts by the SROs as it relates to regulation (both member and market) and market operations.

Once the "correct" percentages have been developed for those three functional areas, assigning the appropriate net common cost pool percentages to administrative costs and to the cost of capital is relatively more straightforward, if not less subject to dispute.

Listings Have Little Tie to Regulation and Operations

In all cases in Columns 8-10 and 11-13 of the template in Appendix AA, a very important rule has been applied. The question is first asked to what extent are market regulation, member regulation, and market operations crucial to attracting and keeping listings. As can be seen in Appendix AA, it is low for equities for a number of reasons.

One is that most companies do not care that much about either regulation (market or member) or the operation of the market unless the SRO does a very poor job in one of the areas. If that were the case, the Commission itself might well have put the SRO out of the SRO business.

It is also crucial to note that with listing fees' falling, listing standards' and rules' changing, and SROs' positioning themselves on the basis of prestige or market segment emphasis, equity listings have little relationship to the broader regulatory and operational issues as far as the public or company managements are concerned. Though one could rightly argue that securities industry professionals understand these issues better (and, therefore, care more about them), it must also be emphasized that their primary goal is to list the company either where its stock can be traded for the most profit, or to best protect the investment banking relationship with the company.

The listing question is irrelevant for options, for any SRO that trades options can list any option it wishes to, at any time.

Transaction Flow Is No Longer Natural

The question is then asked regarding how much of market regulation, member regulation, and market operations relate to attracting transactions. On the equity side, it is somewhat more important in all areas, but again, if an SRO were doing a particularly poor job in these areas, it would likely have resulted in SEC action. Market operations, especially in options, is significantly more important to bringing transactions to an SRO than is true for equities. Unfortunately, however, a trend of payment-for-order flow has developed in equities, and more recently in options, that has a very important role in attracting transactions to an SRO, and it is not related at all to the quality of an SRO's regulation or market operations.

Payment-for-Order Flow Distorts Normal Practice

Were it not for payment-for-order flow's existence, one might argue plausibly that market regulation, member regulation, and market operations would all be weighted more heavily that they are in Appendix AA in the Transactions columns. However, besides depriving investors of whatever payment has been made to a broker-dealer to route an order to the SRO, the pernicious practice of payment-for-order flow has inevitably reduced the importance of regulation (market or member) and market operations in the decision about whether to send orders to an SRO.

Of course, if an SRO were to consistently transact at an inferior bid or offer, even payment-for-order flow would not attract business to that SRO, nor could it under the Commission's rules. However, if an SRO's quotations are equivalent, or if an SRO has made certain customers a pledge that it will meet the NBB or NBO being displayed on another SRO even while it displays an inferior one, the monetary inducement to a broker-dealer to bring order flow to an SRO can override all other issues, assuming there are no gross regulatory or operational problems.

Another Major Problem Looms in Starting Cost-Based Pricing

In setting any percentage of the costs of regulation and market operations that are expected to be offset by transaction fees, the Commission will immediately be hitting a hornet's nest with a very large stick. SROs today charge very differently for transactions in the same securities, ranging from some cost to the party executing, to a negative cost (paying the party to deliver order flow to the SRO).

If a constant percentage of the extent to which regulation and market operations are related to transaction revenues (or, the extent to which transaction revenues are expected to offset regulation and market operations costs) is applied, then those with no (or negative) transaction charges will find their business models adversely affected. They then may be forced to institute fees, which would have very negative consequences for their order flows and exacerbate further the problem caused by applying this constant percentage. On the other hand, if varied percentages are applied in order to help those with no (or negative) charges, those that now charge must have the opportunity to move to the no (or negative) charge model. Doing so, however, would throw more net costs onto market information and could result in considerable higher prices than would otherwise appear likely if looking at the world as it exists now. In other words, the very functioning of a cost-based model could act to drive up SRO market information fees, not push them down as appears to be the intent of trying to impose cost-based pricing for SRO market information.

Market Information Revenues As a % of an SRO's Total Revenue Is Not Relevant

Having asked and answered the questions regarding the importance of regulation and market operations to listings and transactions leaves producing market information that is reliable, that is untainted by fraud or undesirable business practices, and that is useful, as the primary relationship for the costs associated with market regulation, member regulation, and market operations. In the Release, the Commission suggests a range of 20-40% of these costs be applied in any formula that sets market information fees on a cost-justified basis. The stated reasoning in the Release is that 20-40% has been the historic level that market information revenues have been as a percentage of SRO total revenues.

However, this figure is not meaningful. To date, there has been no accurate study of SRO costs as they relate to market information. Current market information fees may be well below or well above those costs as might be dictated by the degree of the relationship.

Most importantly, in neither the current, nor the proposed, method of distributing market information revenues among the SROs is there any relationship between an SRO's market information-related costs and the market information revenues it can "earn" to recoup those costs. It would, thus, have no meaning to try to apply a 20-40% figure as useful in designating the degree of costs that should apply, when that figure has only had significance as a revenue figure as it related to an SRO's entire revenue (not cost) picture.

Template Methodology Lends Credibility to the Process

For the reasons discussed above, the important role of and relationship between market information and market regulation, member regulation, and market operations is greater than the Commission (and some others) might imagine. It is important to re-emphasize that the relatively high percentages assigned to market information for regulation and market operations are a residual, solely a function of determining first the importance (or lack of it) in the Listings and Transactions columns of Appendix AA.

There is another way to approach the process, and that would be to assess first what the allocation to market information should be regarding market regulation, member regulation, and market operations. It is likely that doing so would result in a higher percentage allocation in each category being attributed to market information. That is why the AMEX has suggested the reverse approach.

Administration, Occupancy, and Capital Costs Flow Easily

Once the determination has been made regarding the importance of the functional areas in market information, it is easier to develop the appropriate figure for the associated support and investment areas. By weighting the percentages of each of the functional areas, the composite percentage of the support areas that should be assigned to market information, transactions, and listings is a natural and logical fallout.

The last six columns of Appendix AA merely translate the percentages found in Columns 8-13 and relate them to the final equity and options gross common cost pools shown in Columns 6 and 7.

The Playing Field Has Developed a Tilt

In recent years, the lines between SROs and broker-dealers have blurred. The SEC has encouraged new entities like ECNs and ATSs to develop, and has sanctioned internalization of order flow and payment-for-order flow.8 All these trends have led to a fragmentation of the market, which the SEC is now preparing to examine with a new Concept Release.9

No judgment will be made here about whether these developments are positive for the market or not in terms of volatility, liquidity, etc. What can be said, though, is that these new market participants are for-profit broker-dealers. Their cost structures, different from the SROs, do not bear the same burdens of unprofitable areas that guarantee a fair market where quotations and transactions are reliable and untainted by fraud or other abuses of the public's trust.

To ask SROs to compete in this new climate without having the ability to also make a profit is unfair, unreasonable, and unrealistic. A for-profit entity, because of the profit it makes or because of the capital that is invested in the expectation that profits will be earned at some point, will be able to make technological or marketing investments that an organization that does not, cannot, or is forbidden to, earn a profit will never be able to make.

The Playing Field Must Be Leveled Again

If that situation were permitted to continue, SROs would wither and die. In turn, that would force the regulatory role and provision of good quality market information onto the SEC. That outcome is not one that the Commission, Congress, or the public appears to want; indeed, the trend in the political process in recent years has been toward less regulation and governmental bureaucracy, and the assumption of more responsibility by the private sector.10

The only way the SROs will be able to continue doing their job is to be able to earn an ROC or ROE, just as their new competitors do, and to make the requisite investments to remain competitive. Unfortunately, though, the sources of revenue that will permit this to happen are rapidly shutting down because of the changes that have been taking place in the markets.

Profits on Market Information Are the Key

Transaction revenue is no longer a viable source that can be counted on to fund SRO activity. As the market fragmentation increases, both the new competitors and some traditional SROs are using reduced-priced transactions, free transactions, or negative cost transactions (payment-for-order flow), as well as internalization of order flow, to capture market share.11 Even listing fees have trended down on a per-unit basis in recent years.

This leaves market information as the only remaining significant revenue source for the SROs. Thus, if the SEC mandates cost-based market information fees for SROs, the Commission must also allow the SROs to earn a rate of return (either on ROC or ROE) on those revenues. These profits can then be reinvested in whatever manner needed to permit the SROs to remain competitive and to do their job of ensuring that the U.S. capital markets remain the world's fairest and most equitable for investors.

A Lack of Profits in a For-Profit World Carries Other Risks

To the degree that the SROs no longer exist in a genteel, not-for-profit world, but must compete with for-profit entities, their members are faced with choices they never previously had to make. In the future, an SRO member will have a choice: 1) leave capital tied up in an SRO on which no return is earned because it is not permitted, 2) invest in a non-SRO that competes in the same market and makes a profit, or 3) invest in a for-profit SRO. It is pretty clear which choice the member, over time, is likely to make, and it will not be #1.

Therefore, in order to protect the public, and ensure that the SROs are able to survive and fulfill their functions, the SEC must allow an attractive rate of return to be earned on market information if it forces a move to cost-based market information pricing for SROs. This should take the form of an ROC or ROE equivalent to that earned by the major users of market information, the broker-dealers.12

In Totality, SRO Market Information Fee Structures Are Fair

In the Concept release, the Commission asks 14 different, but related, questions about the value and fairness of various SRO market information fees and the overall fee structure. Examined at length below and in Section III, these questions are generally answered by the AMEX in the affirmative. That is, SRO (in contrast to vendor) market information fees are very low and offer good value for all participants, and with one exception, the fee structures in place today are appropriate and should be maintained.

Professional SRO Market Information Fees Are Not a Problem

It is just this good value that provides another reason for the SEC to agree that SRO market information should be priced with an attractive ROC or ROE imbedded in it. A prime example of how inexpensive is SRO market information is the Commission's own assertion that SRO-generated market information charges represent less than one quarter of one percent of the costs of securities industry participants. By any measure, that represents extremely good value for both professional and retail (traditional and online) investors.

As is detailed in Part III of this response, and as the Release admits, prices have been steady or fallen in equity market information over the past decade while the number of trades and quotations have risen dramatically, 250% in the past five years alone. In options, though market information prices have risen by about 30% in the past five years, the number of transactions and quotations reported have soared by over 2,660%.13

Per-Trading Day and Per-Hour Cost of SRO Market Information Is Low

Further evidence of the very reasonable, and probably inadequate, level of SRO market information fees for professionals is the cost per trading day. The securities industry pays its professionals, on average, about $80,000 a year,14 or about $320 per trading day. By contrast, it pays the SROs only $0.50 to $2.40 per trading day15 for the information that allows these professionals to do their jobs.

Finally, on a cost-per-hour of use (actually monitoring the market information), Part III has computations showing that the cost for professionals ranges from $0.11 to $0.74, compared to the hourly cost of the professionals to their firms of $40.00 or more.16 None of these figures would argue that SRO market information fees are too high.

Since on a daily or hourly-use basis SRO market information fees average under one fifth of a percent, to just under two percent of the cost of securities industry compensation,17 respectively, and compensation costs are 30-35% of the costs of the securities industry,18 the daily and hourly cost computations are generally consistent with the SEC's comment in the Release that SRO market information is less than one quarter of one percent of securities industry costs.

A final note on this topic: as detailed in Part III, SEC figures themselves show that SRO market information revenues (and revenues in general) have risen more slowly than either revenues or profits of the broker-dealers, which is the group that is the largest consumer of SRO market information. Thus, SRO market information pricing has itself contributed to a widening in the profitability of broker-dealers despite the large increase in the number of services being ordered and the huge jump in the amount of data being delivered.

Retail SRO Market Information Fees Are Not a Problem

Much of the Commission's concern appears to be directed to the possibility that retail investors (specifically online retail investors, or more correctly, their broker-dealers) may be incurring SRO market information charges that are too high. As Part III makes clear, there is no evidence this is the case.

Based on conservative figures, a mainstream online investor makes 18 trades a year19 and pays (or the brokerage firm pays) $24-48 a year20 in SRO market information fees. This leaves a cost for a transaction that averages $10,000+21 in value at $1.33 to $2.67 per transaction, or 1 to 3 basis points of the value of each transaction. This, in turn, compares very favorably to the cost of $1.00 to $3.00 for using a non-host ATM machine,22 or 75 to 150 basis points for every such transaction that averages $200 per use.

Another way to look at the cost to a retail (or online) investor is on a cost-per-hour of use basis. Using calculations found in Part III,23 the retail online investor (or the investor's brokerage firm) pays about $0.18 to $0.52 for SRO market information, well within the range of what professionals in the securities industry pay.

Since day traders spend far more time monitoring the market, and average far more trades, in some cases approaching or exceeding 3,000 trades a year,24 if their numbers of trades are inserted into the calculations, the per-transaction or per-hour cost to the average retail (online) investor drops substantially.
The sum of all these analyses is that there would appear to be no reason to say that the SROs are pricing their market information on a retail basis in a way that would deprive such an investor, online or not, from being able to pay for (or be provided for free by the investor's firm) necessary market information.

There May Be Other Agendas at Work Here

Online broker-dealers, ATSs, and others that have grown rapidly in the past few years are investing heavily in their businesses to attract and retain customers in order to gain critical mass to become profitable or increase current profitability. Some of the investment is in technology, but much has been the enormous amount of money poured into marketing. For many such firms, this has taken the form of expensive advertising, but it also has meant offering extremely low transaction costs, as well as various services to customers, including real-time market information paid for by the broker-dealer.

To the extent that the discount, online broker-dealers are trying to control their costs of acquiring customers, it makes sense that they would de-emphasize advertising and offer more "free" services such as real-time SRO market information. Figures show that the cost to acquire a customer can be as high as $325 per account,25 against this, a cost to the firm of $24-48 for SRO market information to be provided for free to attract a customer would appear to be a much lower cost and more efficacious way to bring in accounts.

However, there is a fundamental problem with offering to supply customers with free market information, and then calling for lower SRO market information fees. By doing so, firms that are using the lure of free market information to attract customers are trying to do so, not at their own expense, but at the expense of the SROs. Another way to look at it is that such firms are trying to reduce their marketing costs on the backs of the SROs and their competitors at the risk of putting the integrity of the world's most reputable capital market system in jeopardy purely for their own advantage. It is hard to imagine that the SEC would sanction that type of activity.

Discount Broker ROEs Put the Lie to the Plea of Poverty

A final note is worthwhile on this subject. Despite the concerns that have been expressed to the Commission about how SRO market information fees are disadvantaging discount brokers and/or retail online investors, evidence is to the contrary. Even before the recent filing of lower rates for non-professional monthly and per-query fees, SIA data show that discount brokers, in recent years, have averaged ROEs around 35% compared to about 20% earned by the industry as a whole.26 If anything, those figures (and the ones in footnote 27) might argue that retail SRO market information fees have been too low and that this has given the discount brokers an unfair advantage over full-service brokers.

As for the SROs, though they are currently technically not-for-profit organizations under the law, they do make money. Yet, in its best year ever, its last as an independent entity (1998), the AMEX had an ROE of only about 13%, despite significant cost-cutting in its non-regulatory areas and a 15%+ cut in headcount in the prior two years. This rate of return is far below that earned by the broker-dealers, especially those in the discount segment of the industry. To argue that they should lower their ROEs even more, and therefore have even less to invest to remain competitive with discount brokers and others that are moving closer to being SROs, is difficult to rationalize.

Look Beyond SRO-Generated Market Information Fees

If there is an issue with market information costs for the securities industry, it is not with the SROs, but with the vendors who use the data as one part of their raw materials. A recent Waters Magazine survey27 indicated that in the U.S. there are over 130 re-distributors (vendors) of market information (excluding firms that receive information from various sources and re-distribute it internally only).

In 1999, these 130+ firms had revenues from the sale or leasing of equipment and the sale of market information of about $19 billion in the U.S, with about $12.5 billion being for the information alone. The SROs in Plans under the jurisdiction of the SEC and that of the CFTC, by contrast, generated revenues of only about $650 million.28 Thus, the SROs' income from market information is about 3% of the amount paid by the securities industry and investors to vendors for market information (not including equipment charges).

In view of these facts, and that SRO-generated market information is less than one quarter of one percent of the costs of the securities industry, the SROs' market information charges are not the problem in the world of market information. However, to the extent that market information vendor charges may represent 8-10% of the securities industry's costs, that could certainly be an issue. On the other hand, as noted previously in footnote 3, the prices that vendors charge does not appear to fall under the regulatory purview of the SEC. In sum, in view of the evidence, it does not appear to be a good use of the Commission's time or manpower to try to construct an elaborate new system to solve a problem at the SRO level where one does not exist.

Some SRO Market Information Fees Could Change

Having examined at length above and in Part III the level of SRO market information fees for both securities industry professionals and retail investors, the AMEX is very firm in its belief that such fees are quite reasonable, represent excellent value, and if anything, may be low. This is particularly true in options. The economics of the options business are such that without higher options market information fees, the ability of the SROs to fulfill their functions will become problematic.

However, there are three issues pertaining to the fee structure that were examined in detail as a result of the SEC's Concept Release. In the end, there is one type the AMEX feels should be maintained, one the AMEX believes should be increased in magnitude, and one the AMEX believes should be modified. These are discussed fully below.

Premium Prices for Non-Members of SROs Are Justified

The Plans covering Networks A and B, and that covering OPRA, currently carry higher monthly fees for non-member professionals than for members. This differential should be maintained, as it provides a level playing field between members and non-members, and therefore, is just and equitable.

Members of an SRO have monthly and annual fees they must pay that help to defray the costs of running the SRO and ensuring that it is able to adequately perform its function as an SRO. Non-members pay no such fees, and so upon entering the market to compete for business, have a built-in cost advantage which, if not rectified, would constitute an unfair competitive edge.

This is especially true since the fees the members pay inure to the benefit of all those using the SRO's capabilities and facilities, members and non-members alike. To allow the non-members such an advantage would be akin to having a retailer build two stores, occupy one and then provide the other rent-free to a competitor. The lower cost enjoyed by the second retailer will allow the second retailer to either charge less for the same goods or to reap higher profits. The result will be the same-the first retailer will lose in the end and will have provided its rival with the instrument used to win the war.

Having non-members pay higher fees for market information generated by the Plan of which the SRO is a participant is a way to make sure the costs between members and non-members are kept more equal. Such a system also prevents non-members from benefiting from the structure that members have helped put in place without paying any usage fee.

Consideration Should Be Given to Raising Vendor Access Fees

As noted previously, the market information provided by the SROs to those vendors that then re-distribute that information for a fee to external parties is one of the most elementary and important pieces of the raw materials that allows those vendors to have a business. The fees they pay for access to this information appear to be far too low.

On average, the vendor access fee for Network B is about $1,100 per month. Vendor access fees paid by those that re-distribute the information externally amounted to about $1.3 million of Network B's total revenue in 1999, or about 1% of Network B revenues. In view of the commercial importance of the information and the increasing inability of the SROS to generate adequate revenues from transaction or listing fees due to the increasing number of entrants in a fragmenting market, as well as the rising burden of collecting information and monitoring the market to ensure its integrity, the AMEX is considering proposing a substantial increase in the monthly access fee for vendors of Network B market information, to about $5,000 per month.

SRO Volume Discounts and Tiered Pricing Should End

After extensive consideration and analysis, the AMEX generally agrees with the Commission's view that price tiers based on the volume of services taken is not warranted commercially or competitively. It could be argued that commercially there is widespread precedence in American business, even in public utilities, for granting volume discounts. Even though minor, even in market information, it is a fact that the per-unit cost of market information goes down the more it is produced and disseminated. However, the difference is so small, and the argument is so much more true for vendors than for SROs, the AMEX is willing to concede the point.

More important is the competitive issue. To the extent that volume discounts and tiered pricing provide a cost advantage to larger firms in the securities industry, while burdening smaller firms that are less able to absorb higher costs, concentration of the business may be encouraged. Since the SEC has been encouraging more competition and less concentration in all segments of the securities industry, such pricing plans fly directly in the face of the the direction the Commission has clearly indicated it wants.

For these reasons, the AMEX will encourage other participants of Plans of which the AMEX is a member to eliminate promptly any fees that are teired or offer discounts based solely on the number of services taken.

Alternative Pricing Methods Can Be Valid

Despite the foregoing, there is no absolute requirement that a service taken be paid for in just one manner. For this reason, the Plans have developed per-query fee structures, and more recently, enterprise license types of charging methodologies applying to both professional and nonprofessional investors.29 These appear to have met with generally positive receptions. It is necessary for the Plans to have the flexibility to continue to develop and institute such innovative billing structures in the future to be able to meet the evolving changes in the needs and makeup of the customer base.

Current Methods of Distributing Market Information Revenues Are Flawed

One of the most important, and perhaps the most meaningful, of the topics addressed in the Concept Release is that of how to best distribute market information revenues among the SROs. There are 17 separate, often related, questions on this subject in the Release. The AMEX' replies are shown below and in Part III.

Today, Networks A and B and OPRA distribute revenue among participating SROs based solely on the number of trades each SRO does. The NASDAQ Plan uses a variant of this theme. This methodology has led to abusive efforts to "game the system" and has done little to encourage what are surely the real goals of the Exchange Act and the SEC, namely, more real price competition among marketplaces, and greater liquidity.

Dividing Market Information Revenues by Trade Counts Is Meaningless

The problems that "counting trades" has caused particularly include payment-for-order flow, executing (and counting for market information purposes) riskless principal trades, and printing trades on a second market when they have already been printed on a different market ("double-printing"). In the first instance, regardless of how competitive may be the price at which a security is traded by the SRO that is paying for a broker-dealer's order flow, the fact is that the investor is being deprived of whatever money is going, instead, to the broker-dealer in return for that order.

Printing riskless principal trades and double-printing are even more insidious outcomes of distributing market information revenues among the SROs based solely on counting the number of trades a market executes. Though falling short of "painting the tape," which is prohibited, both practices still have the ability to mislead the public regarding the real activity and liquidity in an issue.

Improved Markets Should Be Encouraged When Distributing Revenue

The Commission may be able to kill two birds with one stone by changing the method of distributing market information revenues among SROs. It may be able to both end or sharply curtail abusive and useless practices and promote improved markets in terms of narrower spreads and greater liquidity.

The route to take in getting to these destinations does not involve a complicated roadmap, though some fruitful side trips that would be useful in furthering the SEC's full achievement of the Exchange Act's objectives will involve a more complex effort.

Start Making the Quality of Quotations Count

Knowing the price of the last trade is useful for having an historic frame of reference. Knowing the current bid and offer is at least as useful, as it tells a buyer or seller where the market for the security is now and what the liquidity is. Making quotations part of the formula for distributing market information revenues will help tighten spreads, improve liquidity, and improve investor knowledge about the real state of the market in a security.

There are those who will argue that trades are more important, and those who will argue the same on behalf of quotations. At least to get the new system of distributing market information revenues among the SROs underway, and to begin to reap the benefits to the marketplace of doing so, the SEC should weight trades and quotations equally.

Trades Should Still Count, But Differently

As the formula and its description in Appendix BB indicate, trades should be the metric for awarding 50% of the Proportional Distribution of a Plan's revenue each quarter. However, the focus will shift to the market value of what is traded, not the number of trades as is the case currently. The former, counting market value traded, will measure the liquidity that various SROs have brought to the trading of the shares or contracts of a security, and this will encourage SROs to try to maximize the liquidity brought to bear. The latter, counting trades, measures nothing of value.

All trades, regardless of size and regardless of whether they are executed during regular trading hours or in off-hours' trading, should be included in the calculation. Trades marked as "sold sale," because they are out of order and do not contribute to providing good quality information about the market, should not be included in the calculations.

Quotations Will Also Carry Weight

As the description of the formula in Appendix BB also indicates, quotations should be the basis for awarding the other 50% of the Proportional Distribution of a Plan's revenue each quarter. Again, the way the formula is constructed, the focus will be on the liquidity an SRO is willing to provide, and it is likely to result in more liquidity's being promised.

One suggestion, sensible on the surface, is not likely to work in practice, and the AMEX urges the Commission not to adopt it: rewarding markets for being the first at a new NBB or NBO. An obvious potential for abuse exists, as a market participant could "game the system" by entering a minimally-sized bid or offer at the next best price for a fraction of a second in order to establish time priority until the next trade or a better NBB or NBO is provided. Though a new price was certainly established by this first small bid or offer, the extent of its real value is low because it did not exist long enough for it to be accessed, and its minimal size made it minimally useful, in any case.

Time Is of Value in Quotations

Unlike trades, where time is of no importance, the amount of time a bid or offer was either the NBB or NBO until either a trade took place or a better bid or offer were made is very important. Thus, one part of the formula to calculate the value of bids and offers is to weight them by the number of seconds they exist. This is shown in Appendix BB.

In order to make the quotation sizes relate to trading activity, to account properly for the importance of off-hour quotations, and to prevent attempts to "game the system" by having an SRO put in large size bids or offers in high-priced, but inactive securities, both the NBB Pool and the NBO Pool will have each security's cumulative bid and offer calculations weighted at the end of a quarter by the percentage of market value traded that that security accounted for relative to all securities in the Plan in which it is traded.

Bids and Offers Are Monitored Separately

There has never been a requirement that an SRO trading a security be both the best bid and the best offer, only that it maintain a two-sided market. There is no reason to change that prescription. Therefore, as the formula in Appendix BB makes clear, bids and offers are handled separately in order for the SROs to "earn" their portion of market information revenues based on the quality of their quotations.

One issue with which the Commission must deal is the inability to access quotations, or very high numbers of expirations and cancellations of ITS commitments. Fourth quarter ITS statistics show that, with rates of 25-43%, SRO's other than the NYSE and the AMEX were refusing to execute ITS commitments on the grounds that the bid or the offer had already been taken, or that the commitment had expired.30 Though the consistency of these occurrences has been at least irritating to date, if market information revenues are to ride on an SRO's having the NBB or NBO, the SEC must take steps to monitor the reality of an SRO's bids and offers, and to penalize unseemly behavior in the future.

A More Advanced Wrinkle Is Possible

Although the AMEX has not formally incorporated it into the formula proposed in Appendix BB, rewards for SROs that provide price improvement might also be included when apportioning revenue from the Proportional Distribution. Some suggestions for doing so would include: extra credit's being given for trades where either the price or the size of the NBB or NBO were improved ("price improvement"), and no credit's being given for trades in which the price was outside the NBB or NBO ("price dis-improvement").

In any effort to reward price improvement, several issues will have to be decided. One is that a final and consistent methodology for measuring price (and size) improvement will have to be determined by the Commission, and a neutral source must be found to measure both improvement and dis-improvement. In addition, the revenue source from which any extra "credit" will come must be determined, and the amount of any such credit must be decided. These are issues the AMEX would be pleased to work on with the Commission, but they add considerable complexity to the straightforward proposal and formula the AMEX has advanced for changing the method of distributing market information revenues among the SROs, and they are beyond the scope of this document.

No Discrimination Among SROs Should Occur

All SROs should be able to enter a Plan, should they choose to do so, upon payment of a fee that has a pre-set formula. This fee should be used to reimburse the existing members of the Plan for the undepreciated assets and goodwill the Plan has built up through its years of operation prior to the new SRO member's joining it.

Similarly, no Plan should have restrictions on the amount of revenue that a new SRO member can "earn" if they produce the trades and quotations as called for in the formula in Appendix BB.

SROs Should Not Receive Direct Distribution Revenues

Historically, no SRO has received any payments from a Direct Distribution of market information revenue. The one exception to this rule has been that a Plan's Administrator has been able to recoup the costs its organization bore that were directly a function of administering the Plan. Payments from the Direct Distribution have also been made, through the Administrator, to the Processor (SIAC).

The AMEX strongly advocates leaving that situation unchanged, particularly in view of its recommendation for a new, improved method of distributing market information revenue among the SROs. It would be contrary to the thrust of this new method that tries to encourage improved market pricing and liquidity if the SEC were then to sanction paying SRO expenses from the Direct Distribution, without the SRO's having "earned" the right to the money.

One exception might be considered, if despite the lack of justification for doing so and all the potential negatives previously enumerated, the Commission decides to impose a cost-based system for setting SRO market information fees. Since such a methodology is likely to cause significant additional auditing costs, the SEC might consider putting some money aside from the Direct Distribution, as these costs would clearly be related only to market information. However, even this plan is likely to spark controversy between those who believe they are not being sufficiently reimbursed for such expenses, and those who feel the opposite.

Governance of the Plans Works well Now

In the Release, the Commission addresses 13 questions to the issue of Plan governance. Though some obvious questions appear to have been omitted, the AMEX believes a minimum of tinkering along the lines hinted at by the SEC should take place. The AMEX' overall thoughts are provided below, with more elaboration in Part III.

Non-SRO Parties Should Have No Voting Power in the Plans

Although the SROs play a role in protecting the public from fraudulent and abusive actions by their members or others, they are also part of the private enterprise system. This was never more true than today, and is likely to be more so in the future as, with the SEC's encouragement, more parties file to become SROs. Many of these are for-profit companies, and many of the current not-for-profit SROs may well convert to for-profit status in order to generate the capital and reinvestable funds needed to remain competitive.

The generation of market information revenues is an integral, in fact a crucial, part of that business model. Though it is clear that because of the industry-wide cooperative nature of the Plans' policy-setting and rate-making structures (which have long been mandated, sanctioned and overseen by the Commission), the Plans need to be overseen by an outside body and demonstrate the justification for policy or price changes, that body exists. That body is the SEC, and the method of oversight used for more than 24 years has functioned well.

Most Market Information Users Have Input into SROs Already

Besides having the Commission watching how the Plans operate and price their products, most SROs have representatives of the broker-dealer and investment management segments of the securities industry, and representatives of the public, on their Boards.31 Any pricing action that the Plans are contemplating must be approved by those Boards first. Thus, the two largest groups of market information users are represented prior to any pricing action's being initiated.

Another check on the Plans' ability to exercise free rein in pricing is the role the SROs have historically voluntarily asked the SIA to play prior to any pricing moves. When such an action is being considered by the Plans, the Administrator of the Plan involved meets with the SIA, the body the securities industry has chosen to represent it in various forums. There, the industry has a chance to comment on, dissent from, or approve of the coming pricing change.

Finally, and very potently, there is the appeal process an interested party has via the SEC itself. If, despite all the opportunities listed above to try to stop or modify a pricing action, a user of market information feels strongly that an action it opposes is about to be taken, a direct approach to the Commission to block, suspend, or rescind a pricing action can be made. The efficacy of this system of checks and balances has been demonstrated consistently, and recently.

Vendor Representation Is Another Matter

The one group of users not represented on the SRO Boards, nor a formal part of the consultative process, except on an ad hoc or information basis, is the vendor community. However, through such bodies as the Financial Information Forum, and under the auspices of the SEC itself when necessary, vendors have an ability to provide input formally to the SROs regarding their views.

On the other hand, it does not seem appropriate for vendors to sit on the Operating Committees of the Plans as voting members. The Operating Committees of the Plans must run the Plans as businesses that are extensions of the SROs themselves, developing technological capabilities to deal with changing product and volume needs, and pricing everything appropriately. There are reasons to question whether vendors would have those as their objectives.

By the nature of their business, the vendors would look not only to minimize price changes by the SROs, leaving themselves more flexibility in that regard, but also would potentially look to sanction only those products whose characteristics (maximum demand, minimum traffic) would do the most to help their revenues and restrain their costs. This focus is not likely to be in the best interests of investors, or of producing the best possible market information.32

There Should Be Limitations If Non-SRO Votes Are Mandated

Should the SEC mandate that non-SRO parties be allowed to vote on matters that come before the Operating Committees of the Plans, the AMEX strongly believes that vendors should not get such a vote. This is particularly the case in view of how vendors have already reacted to changes that could put downward pressure on their returns, as well as due to the antithetical business relationship of vendors to the SROs.

If others are to get a vote, the AMEX believes one representative each for broker-dealers, investment managers, and the public should vote. If the number were not limited to one vote for each of these groups, it is not clear what cutoff point could be found that would be acceptable to these three groups and the disparate elements in them. In fact, though it would appear that the SROs would still have a majority of votes on the Operating Committees based on "the number of votes at the table" in each of the Plans, such would not be the case in practice.

Were one or more of the three non-SRO parties to strongly oppose something the SROs on the Operating Committee were in favor of, particularly pertaining to prices, that opposition would carry far more influence than a strict vote count might otherwise indicate. As has been experienced in the past in a far less formal structure, were the SROs on the Operating Committee to believe that a non-SRO representative were going to bring an entire constituency, and possibly the SEC, to bear in a dispute, it would be "advice" the SROs, who have not ignored it in the past, would also pay attention to in the future.

Non-Voting Operating Committee Participation Might Be Positive

In the absence of some break in the anti-trust issues that have come to the fore in the last year, the AMEX questions how effectively the Plans can be governed in the future. One solution would be to have the SEC, by a formal order, provide both for the consultative participation on the Operating Committees of the Plans by representatives of all users of market information (broker-dealers, investment managers, the public, and vendors), and provide SRO anti-trust immunity under the aegis of the Commission. If this latter is not done, the SEC may eventually find itself having to directly administer the Plans itself.

A potential benefit of having participation of non-SRO parties on the Operating Committees of the Plans be purely on a consultative basis is that they would not only provide a faster way to convey to the SROs the ideas and views of their constituencies, but could act as a conduit in the other direction. As the SROs are formulating plans or taking decisions that will affect one or more of the parties, the non-SRO representatives will be able to carry that intelligence back to the groups on whose behalf they are working so that there are no breaks in the information chain, and they would also be able to provide relatively rapid feedback on their constituencies' reactions to different issues for the benefit of the SROs.

More Committees Will Not Be Helpful

One solution suggested regarding how to provide more non-SRO input into or oversight of the Plans has been to create a series of new committees, for example audit, pricing, technical, etc., that would give non-SROs a greater voice. Doing this will only add a very inefficient layer on top of the structure of Operating and Technical Committees that now exists.

Besides adding considerably to time and travel burdens for all concerned, it is not clear what would be the relationship of an entirely new set of committees to those that are already charged with, and have proven their ability in, running these Plans so that they fulfill their function of consistently delivering good quality market information at reasonable prices, uninterrupted by technical failures. Would these new committees only advise the current SRO-run committees, or would the new committees be able to mandate policies to the current committees? Would the new committees meet before or after the current committees meet? Would there be a voting structure in the new committees? What would be the method of resolving disputes among those on these new committees whose interests might be sharply divergent?

With the potential for so much wasted effort and unnecessary conflict, the idea of creating new committees, then, appears to provide little positive to the process. If it felt that non-SRO parties should have more active input into the process before the fact, it should be done within the current committee structure, as suggested earlier.

Contract and Billing Administration Should Be Streamlined

One area of market information that has been the subject of nearly-constant, if low-level, dissatisfaction for a long time has been the overly-complicated process of administration. This ranges from customers' having to deal with many different forms of contracts and agreements, the sheer number of them that exist, the need to execute most of them in hard copy rather than electronically, the various parties that are involved at different points in the chain, the process of actually initiating and terminating services, and the complexity of the entire billing cycle.

None of the aforementioned issues are the core business of any of the parties, SROs, vendors, or ultimate users of market information. Nonetheless, they are each forced to create groups within their organizations to wrestle with the complexities presented by the opacity of the market information administration process. Since each such group is purely a support function and relatively small in size, the capabilities of those involved vary widely, there is no ability to generate critical mass efficiencies, and so real and opportunity costs are inevitably wastefully high.

SEC Guidance Could Be Helpful

Despite the long-standing unhappiness of the parties with the problems that exist in administering market information, little has changed because of both inertia and the various proprietary interests of the organizations involved. In order to get an effort underway to reform and simplify the system, it may be necessary for the SEC to take a lead role.

Recognizing that many of the firms may not come directly under the jurisdiction of the Commission, the SEC can probably be most helpful in bringing the parties together and helping them begin to frame the outline of a solution, rather than trying to craft the solution itself. Elements that should be part of the effort could include standardization of contracts and agreements, especially as it relates to language and terms; an industry-wide central billing authority, possibly by a third party; converting much of the paper process into one that can be done electronically; and developing an industry-wide auditing function to ensure that all parties are upholding their parts of the agreements.

The desired outcome of this effort would be twofold. First, the administration of market information could be much simplified for all concerned. In addition, by putting the process into the hands of one participant-neutral, central administrator to the extent possible, the ability to create a more efficient, less-costly system of administration would benefit all concerned.

Pilot Program Allow Innovation and Have Their Place

All industries need the ability to experiment on a small scale with new technologies, programs, and pricing schemes, either with limited numbers of participants, or for limited periods of time. In order to safeguard the health of both the sponsoring organizations and their customers, there are times when beginning an effort that immediately and irrevocably applies to all may not be the wisest course to take.

For this reason, the AMEX suggests that, perhaps with some modifications, the SEC should continue to allow pilot programs to be initiated by the Plans. One important change that might reduce the concern that pilot programs are not being instituted for the experimental purpose for which they are intended would be to have pilot programs use an "effective on filing" process. The pilot program would, thereby, be visible to all who read the Federal Register. Then, if strong and valid objections to the pilot program were raised, the SEC would be able to halt, delay, or request modifications in it.
Another change that might be instituted could include these pilot programs' having a formal stated, and limited, duration that could only be extended after a presentation of the reasons for such an extension, and agreement of the Commission to grant it.

A More Open Process Has Some Risks

As indicated above, a key reason to allow pilot programs is to limit their impact if they prove not to have the intended outcome on either the sponsoring organization or the clients of that organization and limitations can be either time (as proposed above), or restrictions on the number of participants. In the latter case, though it may make sense for a Plan to file "effective on filing" with the SEC the outline of the pilot program and those that will be part of it, it is not clear how to restrict those that might want to be part of the program if they find out it is in effect because of such filings with the Commision.

For this reason alone, the SEC must decide whether pilot programs in the future can be restricted only as to duration. If participant restrictions are also to be allowed, it will be important for the Commission to help the Plans craft some guidelines regarding how to do that without giving the appearance of there being intentional discrimination for or against some party.

Member Rebates Take Many Forms and Will Be a Sticky Subject

The SEC also addresses five questions to a very thorny issue-SRO member rebates. The AMEX explains, below and in Part III, its responses given in Part I.

Member rebates are, by their nature, hard to identify as such. They can consist of payment-for-order flow, reductions in or elimination of trading fees, dividends to members of a for-profit SRO, lower market information fees for members than for non-members of an SRO, forgiveness of past or future dues or fees, and the like. Were the SEC to consider disallowing member rebates, there is at least one major difficulty in doing so.

All enterprises manage, to the extent possible, their sources of revenue generation, and their costs, as part of their business strategies for maintaining or maximizing their competitive positions. Member rebates are clearly part of this equation. Thus, were the SEC to try to dictate the extent or timing of SRO member rebates, they would begin to insert themselves directly into determining the business strategy and tactics each SRO could employ at any given itme. The implications of doing so could prove dangerous for the Commission, the SROS, and the markets themselves.

There are certainly potential positives to trying to address limiting or eliminating some member rebates, for instance, ending the practice of payment-for-order flow. However, since, as noted, there are a variety of forms that member rebates can take, and since at heart they all have the same objective of trying to increase the competitiveness of an SRO, the SEC's limiting or forbidding one kind of member rebate while allowing another would be a form of potential business discrimination in which the Commission would not likely want to engage.

Cost-Based SRO Market Information Fees May Introduce an Exception

Were the SEC to mandate cost-based pricing for SRO market information fees, despite all the potential negatives of doing so that have been discussed, there may appear to be a reason to limit, though not eliminate, member rebates. Because market information fees for SRO data would have been set specifically to allow the SROs as a group to offset the costs related to market information which they must bear (and after doing that, to earn a profit), permitting unfettered rebating of any fees would be contrary to the intent of the Commission.

It is important to remember that submitting costs on which SRO market information fees will be set will have no relation to the ability of an SRO to "earn" the revenues to offset those costs. The costs can only be "earned back" if an SRO is willing and able to supply or find liquidity or provide tighter markets as demonstrated by the formula for the distribution of market information revenues contained in Appendix BB. Some SROs may never contribute enough liquidity or quality quotations to the market to fully "earn" their costs back, while others will earn more than their portion of the costs because they have an outsized positive effect on liquidity and market quality.

It is only when a market has fully "earned back" the costs that it said it had when market information fees were being set, at least partly on that SRO's cost submissions, that the SRO should then be allowed to rebate any of its fees, whether in payment-for-order flow, reduction of transaction charges, or any type of action with regard to any type of fee.

The SEC May Still Face a Firestorm of Protests

Despite the logical allure of this line of reasoning in a system where SRO market information fees are cost-based, going down this path will most probably embroil the Commission in controversy and have a number of unintended effects. (This, then is another reason not to move to trying to cost-justify SRO market information prices.) There are markets today that levy no transaction charges, while others do. It would appear unfair for the SEC to take the position that the markets that currently charge transaction fees should never have the ability "to catch up" by having the opportunity to cut or eliminate their fees unless they were to "earn" their costs back first.

To take that stance, the Commission would clearly be interfering in some SROs' ability to remain competitive. That, in turn, could lead some SROs, which might otherwise be performing well in making markets and providing liquidity, to exit their roles as SROs. This would be an outcome the SEC would not have intended, nor would it be consistent with the objectives of the Exchange Act.

SRO And The Plans' Disclosure Requirements Should Not Be Increased

Although some may feel that the SROs, by their nature as guardians of the public trust, should be completely open books, financially and otherwise, the fact is that both potential new, and existing traditional, SROs are businesses, with competitive strategies that can be revealed by those who know how to read financial statements. The degree of disclosure now required appears to strike a satisfactory regulatory/business balance, especially as only one part of the SROs' business can be considered a monopoly, while all the rest is fiercely competitive.

To some extent the argument against more disclosure applies to the Plans, also. If the detail that is discussed in the Release regarding costs that might be used to set fees were included in new disclosure rules for the Plans, it would be essentially the same thing as providing an unwarrantedly detailed look inside each SRO in the SROs' own filings.

It could be argued that more disclosure could be made, and only the SEC would have access to it. However, experience has shown that once information is filed with the Commission, other parties are prone to file Freedom of Information Act requests in order to look at the data. For whatever reason, the SROs have not been able to prevail, when such requests are made, in protecting the confidentiality of previously-submitted material. There is no reason to believe this will change in the future. Therefore, the AMEX would oppose greater disclosure on any basis that would potentially reveal its current or future business strategies.


Appendix AA

Template to Develop Gross and Net Common Cost Pools
  (1) (2) (3) (4) (5) (6) (7)
Cost Category Gross
of Cate-
gory (g)
Am't. (g)
% of Gross
Cost Pool
Attributable to
Equity        Options
Amount of Gross
Cost Pool
Attributable to
Equity        Options
Market Regulation (a) 100 8.0 800 40.0% 60.0% 320 480
Member Regulation (a) 100 1.0 100 60.0% 40.0% 60 40
Market Operations (b) 100 32.0 3,200 30.0% 70.0% 960 2,240
Administration (c) 100 4.0 400 32.7% 67.3% 131 269
Occupancy (d) 100 4.0 400 30.0% 70.0% 120 280
Cost of Capital (e) 100 2.0 200 33.0% 67.0% 66 134
TOTAL (f) 100 51.0 5,100 32.5% 67.5% 1,657 3,443

Template to Develop Gross and Net Common Cost Pools
(8) (9) (10) (11) (12) (13)
  % of Equity Gross
Common Cost Pool
Amount Attributable to
% of Options Gross
Common Cost Pool
Amount Attributable to
Cost Category Mkt.
Market Regulation (a) 65.0% 25.0% 10.0% 85.0% 15.0% 0.0%
Member Regulation (a) 65.0% 25.0% 10.0% 85.0% 15.0% 0.0%
Market Operations (b) 55.0% 35.0% 10.0% 55.0% 45.0% 0.0%
Administration (c) 57.8% 32.2% 10.0% 60.7% 39.3% 0.0%
Occupancy (d) 57.8% 32.2% 10.0% 60.7% 39.3% 0.0%
Cost of Capital (e) 57.8% 32.2% 10.0% 60.7% 39.3% 0.0%
TOTAL (f) 57.8% 32.2% 10.0% 60.7% 39.3% 0.0%

Template to Develop Gross and Net Common Cost Pools
  (14) (15) (16) (17) (18) (19)
  Amount of Equity Net
Common Cost
Pool Attributable to
Amount of Options Net
Common Cost
Pool Attributable to
Cost Category Mkt.
Market Regulation (a) 208 80 32 408 72 0
Member Regulation (a) 39 15 6 34 6 0
Market Operations (b) 528 336 96 1,232 1,008 0
Administration (c) 76 42 13 163 106 0
Occupancy (d) 69 39 12 170 110 0
Cost of Capital (e) 38 21 7 81 53 0
TOTAL (f) 958 533 166 2,088 1,355 0


(a) Columns (8) through (13) are AMEX estimates; in all cases, the figures for Transactions and Listings were derived first, leaving the figure for Market Information as a residual to be such that the three numbers would add to 100%.
(b) The figure in Column (1) includes all operating costs, both personnel and non-capitalized systems costs.
(c) The figure in Column (1) is determined by the formula:

((HcMkR+HcMmR+HcMO)/HcSRO) x $AdSRO

HcMkR = the headcount of the Market Regulation staff
HcMmR = the headcount of the Member Regulation staff
HcMO = the headcount of the Market Operations staff
HcSRO = the total staff headcount of the SRO
$AdSRO = the total Administration cost of the SRO (human resources, finance, legal, all communications and purchasing costs, and the senior executive staff)

The figure in Column (4) is determined by the formula:

((MkR3 x MkR4)+(MmR3 x MmR4)+(MO3 x MO4)) / (MkR3+MmR3+MO3)

MkR3 = Market Regulation, Column (3) in Appendix AA (or Column (1) if actual dollars were being shown)
MkR4 = Market Regulation, Column (4) in Appendix AA
MmR3 = Member Regulation, Column (3) in Appendix AA (or Column (1) if actual dollars were being shown)
MmR4 = Member Regulation, Column (4) in Appendix AA
MO3 = Market Operations, Column (3) in Appendix AA (or Column (1) if actual dollars were being shown)
MO4 = Market Operations, Column (4) in Appendix AA

The figure in Column (5) is determined by the formula:

((MkR3 x MkR5)+(MmR3 x MmR5)+(MO3 x MO5)) / (MkR3+MmR3+MO3)

MkR3 = Market Regulation, Column (3) in Appendix AA (or Column (1) if actual dollars were being shown)
MkR5 = Market Regulation, Column (5) in Appendix AA
MmR3 = Member Regulation, Column (3) in Appendix AA (or Column (1) if actual dollars were being shown)
MmR5 = Member Regulation, Column (5) in Appendix AA
MO3 = Market Operations, Column (3) in Appendix AA (or Column (1) if actual dollars were being shown)
MO5 = Market Operations, Column (5) in Appendix AA

The figures in Columns (8) through (10) are derived by multiplying the respective column figures in Market Regulation, Member Regulation, and Market Operations by the figure in the same line in Column (6), then dividing the sum by the total of the first three lines of Column (6).

The figures in Columns (11) through (13) are derived by multiplying the respective column figures in Market Regulation, Member Regulation, and Market Operations by the figure in the same line in Column (7), then dividing the sum by the total of the first three lines of Column (7).

(d) The figure in Column (1) is determined by the formula:

((HcMkR+HcMmR+HcMO)+HcMm)/HcSRO+HcMm) x $OcSRO

HcMkR = the headcount of the Market Regulation staff
HcMmR = the headcount of the Member Regulation staff
HcMO = the headcount of the Market Operations staff
HcMm = the headcount of the Members and related staff making markets or acting as brokers for customers trading with the SRO
HcSRO = the total staff headcount of the SRO
$OcSRO = the total occupancy cost of the SRO (rent/mortgage/imputed cost, building services, real estate taxes)

The figure in Column (4) is determined by the formula:


HcMkRE = the headcount of Market Regulation staff primarily involved with equities
HcMmRE = the headcount of Member Regulation staff primarily involved with equities
HcMOE = the headcount of Market Operations staff primarily involved with equities
HcMmE = the headcount of the Members and related staff making markets or acting as brokers for customers trading with the SRO in equities

The figure in Column (5) is determined by the formula:


HcMkRO = the headcount of Market Regulation staff primarily involved with options
HcMmRO = the headcount of Member Regulation staff primarily involved with options
HcMOO = the headcount of Market Operations staff primarily involved with options
HcMmO = the headcount of the Members and related staff making markets or acting as brokers for customers trading with the SRO in options

The figures in Columns (8) through (10) are derived by multiplying the respective column figures in Market Regulation, Member Regulation, Market Operations, and Administration by the figure in the same line in Column (6), then dividing the sum by the total of the first four lines of Column (6).

The figures in Columns (11) through (13) are derived by multiplying the respective column figures in Market Regulation, Member Regulation, Market Operations, and Administration by the figure in the same line in Column (7), then dividing the sum by the total of the first four lines of Column (7).

(e) The figure in Column (1) is determined by the formula:

(($CIMkR+$CIMmR+$CIMO+$CIAd+$CIOc) x %ROR) + ($DMkR+$DMmR+$DMO+$DAd+$DOc)

$CIMkR = the dollars of capitalized investment in Market Regulation
$CIMmR = the dollars of capitalized investment in Member Regulation
$CIMO = the dollars of capitalized investment in Market Operations
$CIAd = the dollars of capitalized investment in Administration
$CIOc = the dollars of capitalized investment in Occupancy
%ROR = the designated rate of return that would have been earned on the capital invested (or the debt or equity cost of the capital invested)
$DMkR = the dollars of depreciation attributable to capital invested in Market Regulation
$DMmR = the dollars of depreciation attributable to capital invested in Member Regulation
$DMO = the dollars of depreciation attributable to capital invested in Market Operations
$DAd = the dollars of depreciation attributable to capital invested in Administration
$DOc = the dollars of depreciation attributable to capital invested in Occupancy

The figure in Column (4) is determined by the formula:


$CIMkRE = the dollars of capitalized investment in Market Regulation primarily involved with equities
$CIMmRE = the dollars of capitalized investment in Member Regulation primarily involved with equities
$CIMOE = the dollars of capitalized investment in Market Operations primarily involved with equities
$DMkRE = the dollars of depreciation attributable to capital invested in Market Regulation primarily involved with equities
$DMmRE = the dollars of depreciation attributable to capital invested in Member Regulation primarily involved with equities
$DMOE = the dollars of depreciation attributable to capital invested in Market Operations primarily involved with equities

The figure in Column (5) is determined by the formula:


$CIMkRO = the dollars of capitalized investment in Market Regulation primarily involved with options
$CIMmRO = the dollars of capitalized investment in Member Regulation primarily involved with options
$CIMOO = the dollars of capitalized investment in Market Operations primarily involved with options
$DMkRO = the dollars of depreciation attributable to capital invested in Market Regulation primarily involved with options
$DMmRO = the dollars of depreciation attributable to capital invested in Member Regulation primarily involved with options
$DMOO = the dollars of depreciation attributable to capital invested in Market Operations primarily involved with options

The figures in Columns (8) through (10) are derived by multiplying the respective column figures in Market Regulation, Member Regulation, Market Operations, Administration, and Occupancy by the figure in the same line in Column (6), then dividing the sum by the total of the first five lines of Column (6).

The figures in Columns (11) through (13) are derived by multiplying the respective column figures in Market Regulation, Member Regulation, Market Operations, Administration, and Occupancy by the figure in the same line in Column (7), then dividing the sum by the total of the first five lines of Column (7).

(f) In this example, Column (1) is not added; rather, the index of 100 is multiplied by the cumulative relative sizes in Column (2); were Column (1) actual dollars, it would be added, and Columns (2) and (3) would not exist.

The totals for columns (4) and (5) are calculated by multiplying each line of Column (3) by the same line of Column (4), or Column (5), respectively, and then dividing by the sum of Column (3); were actual dollars used, Column (1) would be substituted for Column (3) in the calculations.

(g) In practical application, this column is not needed if Column (1) shows actual dollars.


Appendix BB

1. Every security traded is evaluated separately in each of the three formulas below; all data for each security are added up over the course of a quarter, with each of the formulas below treated separately, across all SROs.
2. An SRO will be awarded the proportion of the revenue pool that pertains to each of the formulas below that it earns for its activity relative to that formula over the course of the quarter.
3. The Transaction Pool will account for 50% of a Plan's distributable revenue each quarter; the National Best Bid ("NBB") Pool will account for 25% of a Plan's distributable revenue each quarter; the National Best Offer ("NBO") Pool will account for 25% of a Plan's distributable revenue each quarter.
4. At the end of a quarter, every security in the NBB Pool and the NBO Pool will be weighted according to the percentage its market value traded relative to the market value traded in total for all stocks in its Plan on a quarterly basis.
5. Transaction Pool Formula: 50% of each Plan's quarterly distributable revenue
  - each set of parentheses represents a separate transaction in a security
  - all transactions in a security are added cumulatively during a quarter
  - the cumulative figures for all securities traded within a Plan are added for each SRO trading those securities
  - the sums of all SROs' Transaction Pool data are added
  - each SRO will be awarded the percentage of the Transaction Pool revenue that the sum of its transaction data is as a percentage of all SROs' Transaction Pool data

(v1 x p1) + (v2 x p2)....+(vn x pn) =

v1 = volume in the first trade in a security in a quarter
p1 = price in the first trade in a security in a quarter
v2 = volume in the second trade in a security in a quarter
p2 = price in the second trade in a security in a quarter
vn = volume in the last trade in a security in a quarter
pn = price in the last trade in a security in a quarter

6. National Best Bid Pool: 25% of each Plan's quarterly distributable income
  - each set of parentheses represents a NBB for a security until a better NBB replaced it or a transaction occurred
  - all of an SRO's NBBs in a security are added cumulatively during a quarter
  - the cumulative figures for all securities traded within a Plan are added for each SRO that has had NBBs for those securities
  - the sums of all SROs' NBB Pool data are added
  - each SRO will be awarded the percentage of the NBB Pool revenue that the sum of the value of its NBBs data is as a percentage of all SROs' NBB Pool data

((v1 x t1) + (v2 x t2)....+(vn x tn)) x (MVs/MVtp) =

v1 = volume of the bid when it was the NBB for bid #1 of the quarter
t1 = number of seconds the bid was the NBB for bid #1 of the quarter
v2 = volume of the bid when it was the NBB for bid #2 of the quarter
t2 = number of seconds the bid was the NBB for bid #2 of the quarter
vn = volume of the bid when it was the NBB for the last bid of the quarter
tn = number of seconds the bid was the NBB for the last bid of the quarter
MVs = market value traded of the security by all SROs during the quarter
MVtp = market value traded of all securities in the Plan by all SROs during the quarter

7. National Best Offer Pool: 25% of each Plan's quarterly distributable income
  - each set of parentheses represents a NBO for a security until a better NBO replaced it or a transaction occurred
  - all of an SRO's NBOs in a security are added cumulatively during a quarter
  - the cumulative figures for all securities traded within a Plan are added for each SRO that has had NBOs for those securities
  - the sums of all SROs' NBO Pool data are added
  - each SRO will be awarded the percentage of the NBO Pool revenue that the sum of the value of its NBOs data is as a percentage of all SROs' NBO Pool data

((v1 x t1) + (v2 x t2)....+(vn x tn)) x (MVs/MVtp) =

v1 = volume of the offer when it was the NBO for offer #1 of the quarter
t1 = number of seconds the offer was the NBO for offer #1 of the quarter
v2 = volume of the offer when it was the NBO for offer #2 of the quarter
t2 = number of seconds the offer was the NBO for offer #2 of the quarter
vn = volume of the offer when it was the NBO for the last offer of the quarter
tn = number of seconds the offer was the NBO for the last offer of the quarter
MVs = market value traded of the security by all SROs during the quarter
MVtp = market value traded of all securities in the Plan by all SROs during the quarter



I. Cost of Market Information

(1) SEC Requests Comment on: "the concept of a flexible, cost-based approach to evaluating the fairness and reasonableness of [market information] fees and revenues." (p. 62).

AMEX Reply: It is possible to develop such a system in the hope it will improve how the Plan, now over 24 years old, responds in the future to changes that have occurred in the markets. The drawbacks are significant and should not be ignored, however, and probably outweigh the putative benefits.

AMEX Reasoning and Comments: The AMEX agrees with the Commission that such an approach should "not [be] a strict mathematical calculation..." (p. 63), and that the "SROs would...need to provide sufficient periodic financial disclosures to demonstrate their compliance [with the ground rules established for such an approach]..." (p. 65). However, providing such disclosures, and the need for the SEC to analyze them, illustrates one serious risk of a cost-based system for setting market information fees-the likelihood of the need for a much larger staff at both the Commission and the SROs just to deal with the new requirements.

The AMEX also agrees with the Commission's view that "SROs required to file a proposed fee change with the Commission when necessary to maintain compliance with the limit." (p. 65). This is another of the negatives of a cost-based system for setting market information fees, as will be discussed more fully in Request for Comment (54) below--fee filings will probably need to be done annually, regardless of whether or not fees are to be changed. This will put a substantial burden not only on the SROs, but also on the SEC.

Certainly, if cost-based pricing were imposed, the AMEX would agree with the Commission's assertion that the "[n]etworks would be required to adjust...fees charged to different categories of vendors and subscribers so that they..[do] not that would exceed [an established cost-based] limit." (p. 64). Making these adjustments is likely to be both time-consuming and, potentially, contentious. Parties subject to price reductions will be happy; in those instances in which costs dictate higher rates, the consumers of the data are likely to be unhappy.

As touched on above, the drawbacks to a cost-based system for setting market information rates are numerous, and additional problems are examined in the replies to Request for Comment (34)-(40), (42) and (54) below. Despite the conceptual appeal of such a system, in fact these drawbacks probably far outweigh the few benefits.

* * * * * * * * * * *

(2) SEC Requests Comment on: "[whether the costs of market operations and market regulation as] categories [to be included in calculating what the fees to be charged for market information should be] are sufficiently distinct to provide the basis for a workable internal cost allocation." (p.65).

AMEX Reply: They can be identified, but may not be distinct enough to prevent major clashes among all parties; they are certainly not inclusive enough.

AMEX Reasoning and Comments: As will be discussed in Request for Comment (7), there are other categories of "nontraceable costs" (p. 56) that should, at least in part, be included; the reasons for doing so will be analyzed at that time.

The AMEX has spent considerable time and effort analyzing the costs attendant on its various SRO and commercial roles. Costs for its various functions are able to be identified readily. The problem that will plague a cost-based system for setting market information rates will be separating the portion of those costs attributable to market information, listing, or transaction revenue. The potential for serious discord exists and is discussed in Request for Comment (5).

* * * * * * * * * * *

(3) SEC Requests Comment on: "specific types of costs that should, and should not, be classified as substantially contributing to the value of market information." (p. 66).

AMEX Reply: Beyond market operations and market regulation, there are several other categories of costs that, at least in part, should go into the determination of market information fees. These include, but may not be limited to: member regulation, cost of capital, and administration (broadly defined).

AMEX Reasoning and Comments: The reasons why at least a portion of the costs of member regulation, cost of capital, and administration should be included, and how to determine the appropriate percentages, are addressed at length in Request for Comment (7).

* * * * * * * * * *

(4) SEC Requests Comment on: "whether the cost of market information should include an allowance to provide a fair rate of return and, if so, how a fair rate of return should be determined." (p. 66).

AMEX Reply: If cost-based pricing for SRO-generated market information were to be imposed, a fair rate of return must be allowed.

AMEX Reasoning and Comments: This is a business. Current SROs and their potential new competitors are increasingly for-profit entities,33 or planning on becoming34 and to allow some competitors the opportunity to make a profit through the generation or use of market information, and others not to do the same, is not in the best interests of the long-term health of the markets.

Just as broker-dealers, vendors, or investors can decide where to put their money in order that it earns the maximum rate of return, so, too, should an SRO and its members be able to make the same choice. To the degree that the world of SROs and their competitors, the latter with seeming SEC encouragement, have moved to a for-profit model, where reinvestment is crucial to long term survival and continued improvement in the markets, earning an attractive rate of return on market information is extremely important.

If an attractive rate of return on the investment required for the creation of market information were not to be allowed, then the members of the SRO may determine that they want to invest less in that part of the SRO. By doing so, they are able to reduce the SRO fees they pay. Lower fees will then allow them to put the money they have saved to use in areas of their own business where they are able to earn an attractive rate of return.

The actual rate of return that should be allowed should be related to the rates of return earned by the two major end users of market information: broker-dealers and investment managers. Since market information is important to both types of users, a 3-year or 5-year rolling average of their rates of return might be computed, and then these might be averaged equally across the three to arrive at the rate of return the SROs should be allowed to earn.35

In order to avoid annual adjustments to the rate the SROs would be allowed to earn, and to minimize the need to change pricing every year, a band of plus or minus 10-20% around the allowed rate of return computed as described above might be permitted.

* * * * * * * * * * *

(5) SEC Requests Comment on: "what would be an appropriate standard allocation percentage [for "nontraceable costs"]." (p. 67).

AMEX Reply: The AMEX believes that costs in a particular category should be allocated to market information at the same percentage across all SROs, and that the Commission's proposal that there should be one standard allocation percentage for all "nontraceable costs" is a legitimate way to proceed, as long as that standard allocation percentage is developed both transparently and by sub-category first.36

AMEX Reasoning and Comments: The use of a single percentage is attractive in concept because it would theoretically reduce the amount of paperwork needed to compute the correct level for market information fees. However, a number must be developed first for each of the underlying, appropriate categories. Those figures then can be combined in the right proportions to develop a single percentage figure to reduce the gross common cost pool to the net common cost pool. If this process of building from the bottom up is not undertaken, the resultant figure is likely to be too opaque to satisfy some parties.

Since the figures for the underlying categories, such as market regulation, will have to be computed anyway in order to get to a final, consolidated number, it would be better to show the finer gradations than just the bulk number.

The Commission noted, in an effort to simplify the matter, that such a standard allocation "could be derived from the historical experience of the SROs (on average the SROs appear to fund 30-40% of their market operation and market regulation costs through market information revenues." (p. 66).

Were that a figure that could be applied to Administration and Member Regulation costs also, it might be simpler to use the SEC's formula. However, the Commission appears reluctant to accept such a sweeping figure for all these categories. In addition, because the 30-40% figure refers to the percentage of revenues that market information has been of the SROs total revenues, it is not clear this is an appropriate figure to use when trying to determine an accurate percentage of costs.

Although the Release states without qualification that "the value of a market's information is dependent on the quality of the market's operation and regulation" (p.8), the SEC seems prepared to grant only 30-40% of the costs of market operation and regulation in setting market information fees.

The Commission appears to justify picking this level by asserting that "...[the common] costs [as listed in the Release, i.e., market regulation and market operation] are incurred... also to provide listing and transaction services..." (p.66).

Unfortunate, but true, the operation and regulation of a market have little impact on listings. Most companies have little understanding of how the markets work. Companies are generally more heavily influenced in their listing decisions by other factors such as fees, prestige, advice from their investment banker or others, etc.

The same might be said regarding the role market regulation and market operation have in generating transaction services. A much greater influence on generating transactions than market regulation and operation is the personal relationship between the transacting parties, but especially the level of fees (or negative fees, in the case of "payment-for-order flow").

The bottom line is that market operation and market regulation costs in equities are more likely 55-65% related to market information, 25-35% to transaction services, and 10% to listing services.

* * * * * * * * * * *

(6) SEC Requests Comment on: "whether...[the percentage allocation of the gross common cost pool's being the same for all SROs] would be consistent with the Exchange Act objective of fair competition or whether there are appropriate reasons for allocation percentages to vary from SRO to SRO." (p.66).

AMEX Reply: On an SRO-to-SRO basis, the percentage allocation should be the same unless the product line dynamics are sufficiently different to warrant different treatment.

AMEX Reasoning and Comments: Once a number has been developed for each of the "nontraceable cost" categories, that number should apply equally to all SROs. Trying to tailor a number for each SRO in each category would be too difficult, likely to lead to inequities, and not be consistent with the percentage being based on a perceived, rather than an exact, figure.

In addition, since the cost figure should have nothing to do with the distribution of revenues, as is discussed below, the SROs should not care what the percentage is, since it will not disproportionately or directly affect one compared to another SRO.

However, there may need to be a difference between the percentages applied as they relate to OPRA on the one hand, and Network A, Network B, and NASDAQ on the other. The cost dynamics in the options business are quite distinct from those in equities in some areas, and that probably should be accounted for when deciding on the percentages to be allocated against the gross common cost pool when setting market information fees in options.

* * * * * * * * * * *

(7) SEC Requests Comment on: "a conceptual approach for setting a cost-based limit on a Network's total market information revenues...[that] would...[require]... each SRO...[to]

[a] calculate the amount of its direct market information costs (p. 65);

[b] calculate a gross common cost pool made up of the total amount of its costs that are appropriately classified as contributing substantially to the value of market information (p. 65);

[c] apply a standard allocation percentage to its gross common cost pool to determine its net common cost pool (p. 66);

[d] allocate its total cost of market information [direct costs, plus the net common cost pool] to the various Networks whose securities it trades." (p. 67).

AMEX Reply: If the SEC insists on imposing cost-based pricing for SRO market information, the AMEX believes the Commission's suggestions regarding the calculation of direct market information costs or allocating them among the appropriate Networks as embodied in [a] and [d] above could be accomplished. The ease with which that can be done, and the potential for disagreement and conflict over the numbers that results, however, are likely to be 180 degrees apart.

The AMEX agrees that, should the SEC choose to proceed with cost-based pricing, despite its risks of becoming a much more cumbersome process, the common cost pool should include market regulation and market operations as the Commission embodied in [b] above, but believes there were a number of cost categories that were specifically excluded that should be included. These are Member Regulation, Administration, and Cost of Capital.

For the reasons discussed previously, the AMEX does not believe a standard percentage as embodied in [c] above should apply to the whole pool of common costs, unless it has been built up in a fully-revealed way from less-broad categories.

AMEX Reasoning and Comments: [a] Calculating Direct Market Information Costs The AMEX already calculates its direct costs as Administrator of Network B and would continue to do so under the proposal as embodied in the Release.

[b] Additional Cost Categories to Be Included When considering what is "contributing substantially to the value of market information," one must include timeliness, accuracy, accessibility, being fairly-priced, and deriving from a quality market. If all these are factors, then to acknowledge only market operations and market regulation as contributing to the value of market information is far too narrow a construction.

Member Regulation, Administration, and Cost of Capital are also important contributors to the value of market information.

Member Regulation: Financial Examinations The SEC clearly wants to limit consideration of SRO costs to be included in setting market information fees to the costs of market operation and market regulation (p. 65). However, the cost of member regulation should also be included, for just as "[m]arket regulation by the SROs helps assure that the information on which investors rely is not tainted by fraud or manipulation and that market participants comply with trading rules designed to enhance the efficiency and fairness of the SRO's markets..." (p. 76), so, too, does member regulation.

Financial examinations conducted as part of member regulation assure that counter-parties are able to meet their obligations, and thus that the quotations and trades that occur are real. If a member is not able to stand up to a bid or an offer, or is not able to complete a trade, due to financial problems, then the quotation or trade was not real. In that case, the counter-party was just as much defrauded by the market as would be the case in an infraction of the rules of market regulation.

The Commission itself asserts that "[t]here is little value in market information that is tainted by fraud, deception, or manipulation." (p. 57). It is impossible to see the difference between fraud, deception, or manipulation caused by a party's not following equitable rules for trading, and a party's knowingly offering to enter into transactions it is not able to complete.

In the same vein, excluding member regulation costs from a formula used to set market information fees would also, for the same reasons, appear to violate "Section 11A(c)(1)...[under which]...the Commission is directed to prevent deceptive or fraudulent information (subparagraph A)," (p. 35). It is as much incumbent on an SRO to prevent parties that are not financially able to stand behind their bids, offers, or trades from participating in the market as it is to prevent fraudulent trading. Both will have the same end result-investors will be harmed. Therefore, both should be treated the same way by inclusion in any cost-based formula used to set market information fees.

The SEC states in the Release that member regulation involves the "SROs...promulgating and enforcing rules that govern all aspects of their members' securities business, including their financial condition, operational capabilities, sales practices, and the qualifications of their personnel [emphasis added]" (p. 47). That being the case, there would appear to be no reason for excluding the costs of member regulation relative to those for either market regulation or market operations.

It is unlikely that the Commission is any more anxious to take on financial and sales practice regulation than it is to assume responsibility for market regulation. The SEC has left these types of regulation in the hands of the SROs because Congress has wished to minimize the costs and size of the bureaucracy at the SEC, and because the SROs have done well in their appointed role.

Given the similar damage that can be done to investors by violations that would be uncovered by the SROs in either their market regulatory or member regulatory efforts, and the similar way the SEC has chosen to delegate the responsibility, it is inconsistent to attempt to distinguish between the two types of regulation when trying to develop a cost-based formula for setting market information fees.

The Commission claims that "...although the financial soundness of broker-dealers is undoubtedly an essential factor in the overall integrity of the markets [emphasis added], the connection between this regulatory function and the quality of market information is much more attenuated than in the case of market operation and market regulation." (p. 58). There is no justification presented for this degree of "attenuation," and the arguments presented above in favor of including member regulation as part of any formula that uses costs as the basis for setting market information fees would appear to be quite compelling.

Member Regulation: Sales Practice Examinations Similarly, sales practice examinations are undertaken by the SROs at the Commission's direction in order to assure that deceptive or fraudulent practices are not taking place. The existence of fraudulent behavior in the sales process is only one step prior to entry to the market itself, and may induce an investor to enter the market in a way that should not take place. There is no logical reason to distinguish between fraud in one part of the process compared to another part of the process.

Sales practice examinations also ensure that broker-dealers are not permitting the churning of customer accounts, not allowing trades to be done in bulk and then divided among customers in a way that disadvantages some relative to others, etc. Again, there is no reason to view these deceptive, dishonest, or potentially manipulative practices as anything other than regulatory. To try to distinguish member regulation from market regulation purely because of the word "member" fails to recognize that the two have the same role, and that "member" regulation is a term whose usefulness and reality has been overtaken by changes in the marketplace.

Administration and Occupancy To the degree that any costs such as market operation, market regulation, or member regulation, are included in a proposed cost-based formula for setting market information fees, then so, too, should all costs of administration (human resources, occupancy, executive, etc.) related to running those areas, as well as a related occupancy cost. If those expenses were not present, the areas themselves could not function, and if they did not function, there would be no costs associated with them.

If an SRO does not allocate these administrative and occupancy costs to operating areas as a matter of course in its accounting, including these costs involves structuring a formula that has three steps. First, the cost of all administrative areas would be added up. Then, to arrive at the gross cost pool for Administration, the same percentage that the areas like market regulation, etc. account for relative to the total costs of the SRO would be applied to the total of the administrative costs.

Finally, the blended percentage that results from developing a specific percentage for determining net costs in areas like market regulation, etc. would be calculated. It is this blended percentage that would then be applied to the derived administrative gross cost pool to find the administrative net dollar cost pool.

Cost of Capital The final cost that must be taken into account is the cost of capital. All businesses must make intelligent decisions about where to invest their resources. Capital is a prime, if not the foremost, resource most organizations can use to remain competitive. However, cost of capital is not usually captured directly in the financial statements.

Nonetheless, capital invested in various parts of a company's business can be calculated. Since there is no denying that as the return the organization expects to earn on capital invested is a prime determinant of where assets will be built up (or not), the cost of that capital must be taken into account when allowing prices to be set. Therefore, to omit cost of capital (as opposed to the return on capital) from the formula that would use costs to set market information prices could leave a significant gap between the theory and the reality of the cost of what the SROs are investing in the areas that are crucial to producing valuable market information.

[c] Standard Allocations See Request for Comment (5) for this discussion.

[d] Cost Allocation Among Networks The AMEX has determined with some degree of accuracy what its costs are in its various products as they apply to Network B and OPRA, whose securities the AMEX now trades. Thus, the AMEX could comply with the proposal as embodied in [d] above to allocate costs into the proper Network whose securities are relevant. One modification, however, is suggested.

The Commission asserts that "[t]his allocation can be done directly (for those costs that can be associated with a particular Network), with the remainder allocated based on the proportion of the SRO's total trading volume represented by a Network's securities." (p.67)

Although the concept of using trading volume on different Networks as a proxy for where to put costs may work when an SRO trades only Network A, Network B, and/or NASDAQ securities, it probably does not work if costs must be allocated between OPRA and any of the equity Networks. The problem is that the transaction fees and the notional value being traded in equities is totally different from that being traded in options. A common ground must be found.

Furthermore, a definition of "trading volume" needs to be determined, both for equities and for options. It probably cannot relate to transaction fees, as these have been declining sharply (or for some SROs, are non-existent), and such a relationship to market information revenue might interrupt what has been a very positive and meaningful trend for investors. That is, transaction fees might slow their descent or stop declining altogether.

One suggestion for a definition of "trading volume" would be to use "market value traded" for equities and "notional underlying market value" for options traded.

Finally, should the AMEX begin trading Network A or NASDAQ securities at some point, an allocation of costs in the equity portion of the business at the AMEX would have to be made between or among the equity Networks then relevant to the AMEX.

* * * * * * * * * * *

(8) SEC Requests Comment on: "the advisability and practicality of...[a cost-based] approach [to setting market information fees]" (p.68).

AMEX Reply: Remember that "the only thing worse than not getting what you want is to get it."

On the positive side, in the short run, those who assert that market information costs are not justified will have an opportunity to see specific numbers.

The SROs will be able to comply with the data requests.

Almost all other effects may prove to be negative.

AMEX Reasoning and Comments: A cost-based approach to setting market information fees, even if it tries not to be too exact, will embroil the Commission in what is likely to be nearly endless internecine warfare in an area that for 25 years operated with only 3 notable disputes (pp. 37-38). Arguments, and potential court challenges, will be over such things as:

a) the correct percentage allocations to reduce the gross common cost pool to the net common cost pool;

b) higher prices when the data call for them;

c) possibly much higher market information prices long term across the board if a significant number of securities firms convert to SRO status to claim a piece of the market information revenue pie and load their costs into the system;

d) much more paperwork for the SROs and the Commission staff to make sure that the terms are being adhered to.

Even worse, the SEC is setting itself up to become bogged down in an area that represents "a very small portion of the securities industry's total expenses-less than one-quarter of one percent [emphasis added] in 1998." (p.52).

* * * * * * * * * * *

(9) SEC Requests Comment on: "whether a single [cost-based] approach is appropriate for each of the different Networks and for different types of securities." (p. 68).

AMEX Reply: The structure of the analysis could be the same, but some of the internal applications (as in allocation percentages to be applied to gross common costs) will have to be modified from plan-to-plan, especially with regard to OPRA.

AMEX Reasoning and Comments: Different Networks have different cost structures in the products they trade. For instance, though market operations and regulation in equities is probably 55-65% related to market information, in options that figure could be 75-80%, since there are practically no listing expenses or benefits as there are in equities.

* * * * * * * * * * *

(10) SEC Requests Comment on: "...[provide] suggestions for any alternative approaches to setting a fair and reasonable limit [emphasis added] on market information revenues." (p.68).

AMEX Reply: Leave the current system alone.

AMEX Reasoning and Comments: As embodied by the use of the word "limit" as highlighted above, the Commission seems to have become convinced that SRO-generated market information is not fairly priced. Therefore, the SEC seems determined to find a new methodology to formulate market information fees. However, consideration should be given to leaving the system alone since it has functioned well for 24 years.

As evidence, note that prices have been steady or fallen in equity market information, while product delivered per user has risen sharply. For example, the SRO market information industry has delivered 250% more quotations and transactions in equities, and nearly 2,700% more in options, per user in the past five years.37

It also must be remembered market information revenues paid to vendors are about 20 times greater than those paid to the SROs and yet there has not been a call to regulate their pricing.

II. Fee Structures for Market Information

(11) SEC Requests Comment on: "the fairness and reasonableness of all...the monthly fees applicable to professional subscribers and the fees applicable to retail investors...([including] both monthly nonprofessional subscriber fees and per-query fees)...[and] fees for vendor access and a variety of other services." (p. 69).

AMEX Reply: In general, they are fair and reasonable, and in some cases are probably under-priced.

AMEX Reasoning and Comments: Rationale for each of the fees applied to broker- dealers and retail investors, and the fairness of those fees, is discussed in other sections of this comment. Vendor fees are noted immediately below.

If anything, the "vendor access fee" is far too low. The market information vendor industry is a huge business, now over $12 billion in the U.S. for the data alone. Much of that is based originally on the existence of the SROs and their market information.

The vendors depend on the SROs to generate the market information that is the basis for much of their industry. Being able to buy the raw material for a business for 0.05% of the resultant revenue stream is unheard of in American business. This is a gross margin of 99.97%, which puts to shame that of industries such as pharmaceuticals, where gross margins have at times been over 80%.

The Network A and Network B vendor access fee should be raised from its extremely low level averaging about $1,100 per month, to $5,000+ per month. The higher fee should apply only to vendors that redistribute the data to other firms and individuals outside their own organizations, i.e., use the data directly to generate additional revenues and profits.

* * * * * * * * * * *

(12) SEC Requests Comment on: "the fairness and reasonableness of professional subscriber fees." (p.70).

AMEX Reply: They are extremely good value.

AMEX Reasoning and Comments: The Release asserts that monthly fees range "from $18.50 to $50," or "$0.90 to $2.40 per trading day for market information." (p.70) The source of these figures is not identified.

Daily Cost Modest However, for reference, Network B professional fees run between $0.64 and $1.24 per trading day, depending on the number of services taken, while OPRA fees range from $0.50 to $1.29 per trading day. It is hard to believe that any legitimate firm cannot afford this, when the pay for professionals in financial services averages over $80,000, or $320 per trading day.38

Broker-Dealer Profits Grew 100%+ Faster Than Market Information Revenues Some of the figures the Commission itself introduces, and statements it makes, support the AMEX' view that professional fees for market information are very good value. The SEC comments that "SRO revenues [including those for market information] and costs...have been outpaced by the growth in revenues, costs, and profits [emphasis added] of the securities industry as a whole," (p. 10). This clearly implies that market information costs have lagged securities industry cost increases, and therefore, have contributed positively and incrementally to profit increases for broker-dealers.

Other figures cited by the Commission include those showing that SRO market information revenues rose 67% from 1994 to 1998, only slightly above the rate of growth in SRO overall revenues (64%) and expenses (60%). During that period, securities industry revenues and profits rose 129% and 139%, respectively. (pp. 51-52). The point here should not be missed: revenues and profits of the major user of market information rose 93% and 107% faster that the SROs' market information revenues. Those figures would appear to make it hard to claim that SRO market information fees have been a drag on the profitability of broker-dealers, or put another way, have been unfairly priced, and therefore have hurt the major users.

Growth in Services, Not Prices, the Driver In addition, the SEC correctly notes that most of the rise in SRO market information revenues was due to increases in the number of services ordered when it states that "[t]he fees themselves have remained essentially the same over the last five years. It is an increase in the number of professional subscribers that has produced the increase in revenues." (pp. 69-70).

Further supporting the reasonableness of the fees charged to professionals is the SEC admission that "[f]ees for professional subscribers...produced revenues of $351.1 million in 1998, compared to $231.1 million in 1994, for an increase of 52%..." (p.69), while Network A, and Network B services taken rose 44% and 43%, respectively.

Even more important, the Commission says that "monthly fees for professionals have remained steady despite a substantial increase in the amount of information provided." (p. 70). Statistics cited on page 70 of the Release show that SIAC processed 634 million transaction reports and quotations in 1998 for Networks A and B, compared to only 188 million in 1994, an increase of 237%. The increase for OPRA over that period was over 600%, and with changes in the structure and trading patterns in the options industry, the increase from the full year 1994 through the first two months of 1999 has been nearly five times the 600% seen through 1998.

SRO Market Information Costs Immaterial to Broker-Dealers Already cited, but worth repeating, is that "the SRO's market information revenues represent a very small portion of the securities industry's total expenses-less than one quarter of one percent in 1998." (p. 52)

* * * * * * * * * * *

(13) SEC Requests Comment on: "...the relevant Exchange Act question[:]...whether the fees for particular classes of subscribers, given their economic circumstances [emphasis added]..., are at a sufficiently high level that a significant number of users are deterred from obtaining the information or that the quality of their information sources is reduced." (p. 69).

AMEX Reply: Based on the lack of a significant amount of evidence to date, there is no case to be made to support the concerns that have been raised to the SEC that, in turn, led to this Request for Comment.

AMEX Reasoning and Comments: The major concern appears to be on the part of online brokers on behalf of their clients. A "mainstream" online broker-dealer customer makes 18 trades a year averaging about $10,000 per trade in market value, or $180,000 per year in market value traded.39 It is hard to imagine that a cost of $24-48 per year in SRO-based market information charges is a significant deterrent to such a client.

* * * * * * * * * * *

(14) SEC Requests Comment on: "whether...[retail investor] fees now are low enough and structured in such a way that they do not significantly limit the availability of real-time information to retail investors..." (p. 71).

AMEX Reply: Though a case might have been made prior to recent price changes that, depending on the method by which a retail investor was taking market information, some rates were at the high end of the range professionals pay, that is most certainly not the case now, and at no time were they significantly above professional rates.

AMEX Reasoning and Comments: Recent actions to re-price certain segments of market information is further proof that the current system is responsive to the need for changes. When customers, particularly, broker-dealers acting on behalf of their clients, ask for rate reviews and adjustments, and they are warranted by the evidence, they occur.

* * * * * * * * * * *

(15) SEC Requests Comment on: "whether [retail investor] fees now are low enough and structured in such a way that they do not significantly limit the availability of real-time information to retail investors..." (p. 71).

AMEX Reply: Relative to most fees in the brokerage and financial service industries, SRO-generated market information fees are very low and do not inhibit retail investors, or any other investors, from obtaining real-time information.

AMEX Reasoning and Comments: The heart of the Commission's request for com-

ment is contained in its question about whether "a monthly fee of $0.50 or $1.00 per Network deter[s] a significant number of retail investors from using real-time market information or preclude[s] broker-dealers from providing enhanced information services to their retail customers..." (p. 71).

The SROs provide basic information streams; it is the vendors who provide the "enhanced information services" and whose charges are, in total, 20 times those of the SROs. Thus, if there is a problem with pricing so that retail investors cannot afford such enhanced information, the problem lies with the vendors, not with the SROs.

Furthermore, as is addressed by the SEC's Request for Comment (16) below, many retail investors with online accounts trade through brokers that pay the charges on behalf of their clients. Though some professional day traders may pay those charges because they are set up as independent business people, the volume of their transactions is such that these charges are diminimus relative to the other costs of trading they incur. (And they should probably be paying the professional fee, anyway.)

When compared to other fees in the financial services industry, SRO market information fees compare very favorably. In a competitive environment, bank fees have generally been rising at such a rate that they have even sparked ballot initiatives (that have been successful at the polls, if not in the courts) to limit them.

As an example, unless a saver keeps $10,000 or more in an account, many banks will charge monthly service, or per-check, fees. These can often run a minimum of $15-$50 per year for those who do not have the funds necessary to avoid them. Thus, the SRO market information fees of $24-$48 per year cited by the SEC for individuals whose financial resources are much greater would appear to be very reasonable.

Another example from the banking industry is the imposition of a charge of $1.50-$3.00 every time a consumer uses an ATM at a bank that is not his own. If the average withdrawal is $200, this represents a fee of 0.75%-1.5% (75-150 basis points) on each transaction so effected.40 By contrast, to make one trade of $10,000 in market value per month costs a retail investor only $2.00-$4.00 in SRO market information charges, or 0.01%-0.03% (1-3 basis points).

Finally, an important question must be asked. If a retail investor cannot afford $2.00-$4.00 per month in SRO market information charges, from a financial standpoint, is it appropriate for that individual to be investing directly in the equity or options market in the first place (as opposed to in a managed mutual fund, where a "load" can be 3-5%)?

* * * * * * * * * * *

(16) SEC Requests Comment on: "whether the current fee schedules could inappropriately restrict the information services that broker-dealers provide to their retail customers."(p. 72).

AMEX Reply: Though as discussed below, the Commission postulates a number of areas where individuals may be disadvantaged, none of the examples are realistic. Fee schedules from vendors may be a problem, but those for SRO-generated market information are not.

AMEX Reasoning and Comments: It is true that "...revenue from fees applicable to retail investors (which include monthly fees for nonprofessional subscribers and per-query fees) have grown exponentially in recent years. In 1994, such revenues amounted to $3.7 million [,but by] 1998, they amounted to [$]38.9 million, for an increase of 951%." (p. 71). However, these figures are totally misleading because they are stated in a relative vacuum and not put in context.

There were few online trading accounts in 1994, and it is the online trading accounts that have generated most of the growth in monthly nonprofessional and per-query fees. There were an estimated 3.7-7.5 million such accounts by the end of 1998.41 A rise in accounts from even 500,000 to 7.5 million would dwarf 951% as a rate of increase.

At the same time, CSFB has estimated that in just two years, commissions from online equity trades rose from $268 million in 1996 to $1.3 billion in 1998, an increase of 385%. If it is assumed that 75% of the increase in monthly nonprofessional and per-query fees the SEC cites occurred from 1996 to 1998, those fees would have been risen from $12.5 million to $38.9 million from 1996 to 1998.

That would mean the increase of 211% in market information fees that online brokers would have incurred on behalf of their retail customers actually would have trailed the 385% rise in commissions the brokers received from those customers. Furthermore, as a percent of online trading commissions, SRO monthly nonprofessional and per-query market information fees slipped from 4.7% to 3.0%. With profitability of the online trading business rising relative to the cost of SRO-generated market information, it is curious that anyone could hold a view that online broker-dealers would have to restrict the amount or quality of SRO-generated market information their clients receive.

As indicated above, to the degree the availability of market information is a problem, the difficulty may well be centered on the vendor side of the equation. It is certainly not on the SRO side.

* * * * * * * * * * *

(17) SEC Requests Comment on: "the number of hours in a month that retail investors, on average, could be expected to monitor real-time information." (p.73).

AMEX Reply: Somewhere between 20 and 30 minutes a day for online investors.

AMEX Reasoning and Comments: First it must be pointed out that investors who are customers of full-service firms, as the Commission notes in the Release, receive their market information from their brokers, and those charges are professional monthly fees borne by the firms. Therefore, the relevant subset of retail customers to discuss are online retail customers.

Second, there are no reliable, provable statistics on the degree to which online retail investors monitor real-time market information. The answer of 20-30 minutes a day was developed using the following assumptions:

a) that 95% of online investors are not day traders, and therefore probably monitor real-time market information no more than 7.5 and 15 minutes per day

b) that 5% of online investors are day traders and monitor real-time market information between 5 and 6 hours per day

* * * * * * * * * * *

(18) SEC Requests Comment on: "whether the fees applicable to retail investors are unreasonably discriminatory compared to those of professional subscribers." (p. 72).

AMEX Reply: No; they are solidly in the range of those paid by professionals in the securities industry, which themselves, as discussed above, represent very good value.

AMEX Reasoning and Comments: The Commission's worry is that "retail investors are unlikely to use real-time market information as much as professional subscribers." (p. 72). The question posed in this Request for Comment (18) and the concern expressed about relative usage are addressed more fully below in the reply to Request for Comment (19).

* * * * * * * * * * *

(19) SEC Requests Comment on: "...[does] the difference in rates between professional and nonprofessional subscribers adequately reflect...this difference in use?" (p. 73).

AMEX Reply: Yes; on a time-use adjusted basis, nonprofessionals are paying no more than professionals.

AMEX Reasoning and Comments: The amount of usage by retail investors (and the assumptions used to get there) were provided in the answer to Request for Comment (17). They will be referred to again shortly. The SEC indicates that professional investors "are likely to monitor the stream of real-time market information for a substantial portion of each trading day during the month..., as many as 136 hours..." (p. 72). This assumes that professional users stare at their screens for every minute of every trading day in a 21-day trading month.

A much more realistic assumption would acknowledge marketing travel, meetings, phone calls, reading, lunches, machine-sharing, and other similar interruptions and distractions. Once these are accounted for, the average professional probably monitors real-time information 50-70% of the time (or less), or 68-95 hours a month.42

The other input needed to develop some meaningful cost comparisons of the "time use cost" for securities industry professionals as compared to that of online retail investors is the monthly cost for the various Plans' market information:

SEC figures: $18.50-$50.00 per month
Network B: $13.60-$30.20 per month
OPRA: $10.50-$27.00 per month

Using these figures, and those developed in Request for Comment (17) above, some estimates of per-hour market information costs are possible for both professional and retail investors:

Professional @ 68 hrs (SEC figures): $0.27-$0.7443
Professional @ 95 hrs (SEC figures): $0.19-$0.5344
Professional @ 68 hrs (Network B): $0.20-$0.4445
Professional @ 95 hrs (Network B): $0.14-$0.3246
Professional @ 68 hrs (OPRA): $0.15-$0.4047
Professional @ 95 hrs (OPRA): $0.11-$0.2848
Online Retail @ 7.7 hrs. and $2-4/ month: $0.26-$0.5249
Online Retail @ 11.3 hrs. and $2-4/ month: $0.18-$0.3550

The averages of the four bolded sets of figures are: SEC figures, $0.47; Network B, $0.29; OPRA, $0.26; and Retail, $0.35. These numbers certainly would appear to indicate that on a time-weighted basis, online retail investors have a very competitive rate for market information relative to professional investors.

It should also be pointed out that 1) the active and day trader portions of the online retail investor group are the most rapidly growing segments; 2) it is possible that more than 5% of online retail investors are now made up of these two segments; 3) if that is true, then the time used to monitor the market is understated in the above calculations, and thus, the cost of monitoring the information is overstated.

* * * * * * * * * * *

(20) SEC Requests Comment on: "whether fees for on-line access to market information by retail investors are warranted by the degree of use and the quality of service provided." (p. 73).

AMEX Reply: Yes.

AMEX Reasoning and Comments: As demonstrated in Request for Comment (19), on a time-weighted basis, retail investors have as good a price for market information as the professional does.

Seldom can those clients that buy small quantities of a product receive both an increase in convenience, and have the ability to lower their costs for related services. The SRO market information industry is one of the few to offer this: retail investors get the lowest prices for market information, do not have to make a phone call and talk to a broker, and as a result, can execute trades at a lower price online than with a traditional full service broker.

* * * * * * * * * * *

(21) SEC Requests Comment on: "[whether] any fee applicable to retail investors for on-line access to market information constitutes unreasonable discrimination against on-line investors and their broker-dealers [since] comparison, traditional broker-dealers pay the monthly professional fee and provide market information to their customers by personal telephone call without incurring additional fees." (p. 73).

AMEX Reply: At the hourly usage levels demonstrated in Request for Comment (19) a- bove, the answer is "No."

AMEX Reasoning and Comments: The Orrick, Harrington petition cited in the Release assumes falsely that because their client has found a way to reduce its costs by eliminating a live broker from the cost equation, all the other costs associated with preparing for and doing a trade, such as market information, should disappear also.

Online broker-dealers are able to reduce costs by eliminating people, their associated occupancy and support costs, and the market information expenses that would otherwise have been incurred by those people that were eliminated. However, someone still has to get the market information (and pay for it) so that the online client still has access to it.

In this instance, the online broker will make a decision that either the online investor himself, or the broker-dealer on behalf of the online investor, should pay for the market information. (For competitive reasons, seldom does the broker-dealer pursue the former course.) In either case, however, on a time-use-adjusted basis, the cost of SRO market information is about the same whether paid for at professional rates (a broker-dealer paying it on behalf of an employed broker) or paid for at non-professional rates (a broker-dealer paying it on behalf of a retail client). Thus, there is no discrimination against either the online retail investor or his broker-dealer compared to the traditional full-service model.

Another way to look at this issue is to look at the cost to the broker-dealer of providing market information directly to online retail clients compared to the cost of paying a traditional broker to provide a full-service client with quotations. The equalizer is trades per customer. Data show that the average full-service client executes about 0.5 trades a month. Online clients execute 1.5 trades a month (mainstream online), 10+ (active onstream), or 300+ (day traders).51

It is assumed that a full-service broker spends, over a year, 30 seconds a day on average with a client providing quotations and prices; thus, the broker-dealer will have spent about $0.44 to provide that service. This is about a 25% premium to the cost an online broker-dealer incurs to provide the same to its customer, based on the figures developed in Request for Comment (19).

* * * * * * * * * * *

(22) SEC Requests Comment on: "whether...discounts [offered by Network A, Net- work B, and OPRA as described on page 74 of the Release] are consistent with the Exchange Act objective that exclusive processors of information should remain neutral in their treatment of firms and customers." (p. 74).

AMEX Reply: As it regards discounts offered to "members," while none are granted to "non-members," the AMEX believes these are justified and should be maintained.

After considerable study, the AMEX has come to the conclusion that volume discounts, while justified in "normal" commercial practice, and while not discriminatory in a significant way to any single user (or group of users) of market information, should be eliminated. However, the AMEX also believes that clients should have a variety of choices about how to pay for SRO market information. In this regard, differing modalities of payment, such as per-query or "enterprise license" fees, as compared to the historic per-service monthly fees, are both justifiable and to the benefit of different investor groups whose demands for market information differ in degree and intensity.

AMEX Reasoning and Comments: The AMEX believes the discounts offered to members of the SROs in each of the Plans are justified. The members of the SRO pay annual and monthly fees to help run the SRO and have it carry out its responsibilities as an SRO. Non-members pay no such fees. To that extent, they are getting a "free ride" on the costs assumed by and paid for by the SRO's members. To allow this situation to exist without having a form of redress would be discriminatory against the members of the SRO.

Therefore, having non-members pay an incremental amount to receive market information, which supports so much of the SRO that is vital to their being able to trade in a fair and orderly market, knowing the market information is not tainted by inaccuracies or fraud, seems to be a good solution.

Such a difference in fees, slight as it is, between members and non-members would not appear to entirely satisfy the SEC's belief "that disparities in fees should be justified by such legitimate factors as differences in relevant costs, degree of use, or quality of service." (p. 74). However, to the degree that "differences in relevant costs" can also be seen as "differences in offsetting relevant costs," the disparity appears to fall within the Commission's guideline.

The Commission appears to agree when they say "the benefits of market regulation extend directly to all those who use an SRO's information..." (p. 76). Unless there is a difference in fees for market information, there is no way to correct the imbalance between members who have paid to put in place the structure from which non-members benefit, and non-members who use the structure but do not incur the monthly and annual fees to implement or support it.

* * * * * * * * * * *

(23) SEC Requests Comment on: "whether the size...of discounts should be strictly limited to differences in administrative costs." (p. 75).

AMEX Reply: For the reasons discussed in Request for Comment (22), the answer is "No."

AMEX Reasoning and Comments: Certainly, administrative costs might be one of the factors that determine costs of market information generally if the SEC decides, despite all the potential problems, to proceed with cost-based pricing for SRO market information. However, the premium rate for market information paid by non-members compared to members is to offset the SRO fees that members pay. Again, it is a form of discrimination to have non-members pay no SRO fees, yet receive all the benefits that the payment of those SRO fees by the members bring to the non-members.

* * * * * * * * * * *

(24) SEC Requests Comment on: "any other issue [besides all those previously discussed regarding the cost and fee structures of market information] relating to the effect of market information fee structures on broker-dealers conducting different types of business." (p. 73)

AMEX Reply: After-hours trading, being led by online brokers, will likely raise the costs for the SROs to run their markets, not just the market information central processing systems.

AMEX Reasoning and Comments: This raises the possibility of an increase in prices for market information. The question is whether that increase should be partly paid by those who continue to trade within traditional market hours, or whether those who choose the convenience of trading at non-traditional hours should pay all the incremental costs.

III. Distribution of Market Information Revenues Among SROs

(25) SEC Requests Comment on: "a conceptual approach to distributing the Network's revenues to the individual SROs that could reflect more fully the Exchange Act's national market system objectives." (p.62).

AMEX Reply: The current methodology should be changed.

AMEX Reasoning and Comments: The AMEX believes it positive that "[t]he Commission is considering a conceptual approach to distributing Network revenues that could reflect more fully the objectives of the Exchange Act." (p. 75)

The AMEX has long held the view that the current distribution formulas for Networks A and B, NASDAQ, and OPRA, are seriously deficient in furthering the goal of improving the quality of the markets, rather than just increasing the number of markets. In effect, this helps to encourage fragmentation of the markets without bringing forth the benefits that fragmentation should provide. Having lower transaction fees more than offset by less favorable transaction prices is not a good tradeoff.52

The AMEX fully supports some of the approach the Commission has put out for comment, and in fact, discussed some aspects of it several years ago with staff. The AMEX has tried to refine and provide some useful formulas for those portions with which it agrees, as is reviewed below. However, there are some aspects of the SEC's thoughts that are of concern, primarily due to worry about their feasibility.

One area of worry is the Commission assertion that "...[under] approach... [setting a cost-based limit on a Network's total market information revenues]...[t]he total amount of the costs allocated to each Network from the individual SROs would represent a limit on the amount of revenues that could be generated by each Network's fees..." (p. 67).

This appears to be at odds with the Commission's appearing to leave the door open to allowing a rate of return to be earned on the investment, which would leave prices at some level higher than a system based just on costs would seem to imply. In a period when both the major current, and the new, players in this arena are now (or are moving toward being) for-profit entities, permitting a return on investment will help ensure that the proper investments and reinvestments are made.

An area in which the AMEX is in total agreement with the SEC is the Commission's statement that "separate rules would govern the distribution of Network revenues..." (p.67).

* * * * * * * * * * *

(26) SEC Requests Comment on: "...[whether] the current practice of allocating revenues based solely on an SRO's proportion of transaction volume adequately further[s] the Exchange Act objectives of maintaining the quality of market information and encouraging competition." (p. 78).

AMEX Reply: No. As stated above, it should be changed radically.

AMEX Reasoning and Comments: The drive to get "tape prints" onto a particular market has led to payment-for-order-flow, which takes money that would have gone to the investor and, instead, gives it to a market-maker. Worse, it has also led to double-printing of riskless trades, which at the least, creates a false impression of real demand, and therefore, could be misleading to those who do not understand the basis for the apparent activity in a security.

Even more seriously, the current system has no built-in mechanism that rewards markets that provide price discovery for the investor. Nor has there been a way to make sure that those who put up the most capital benefit from market information revenues. Finally, the current system does nothing to encourage improved markets, that is, those with tighter spreads.

If there is no way for market participants to be rewarded for providing tighter spreads, capital commitment, or price discovery opportunities, the quality of market information is degraded. The current system has encouraged additional participants. Now, more should be done to design a new way to distribute market information revenues among the SROs to encourage real competition on prices, not just ways to get transactions on the tape in order to claim market information revenues.

With this Release, the SEC has the opportunity to correct what has been a serious flaw in the historic system and an impediment to the maximum achievement of the Commission's goals of fully meeting the broadest objectives of the Exchange Act.

* * * * * * * * * * *

(27) SEC Requests Comment on: "whether...transaction volume accurately represents the value of an SRO's quotations, or whether some other basis should be found for distributing a portion of Network revenue based directly on the value of quotations." (p. 79).

AMEX Reply: Transaction volume alone is a poor proxy for the quality of an SRO's bids and offers. This applies to price, size, and the time the bids or offers are left in place to give the market a chance to react. Any new basis for distributing market information revenues should include quotations that are the NBB or NBO directly in a new formula, as all the above items will then be captured, as will (at least indirectly) the value of having provided an engine of price discovery.

AMEX Reasoning and Comments: The specific thinking behind this reply is enlarged upon in the section that pertains to Requests for Comment (28)-(30).

However, it is worth noting here that "one of the fundamental policy decisions made by Congress and the Commission in the mid-1970s was to require all SROs to make their market information, particularly their quotations [emphasis added], available to the public." (p. 78). Though that has surely been done, because quotations were not included in any formula for rewarding the markets that provide the NBB or the NBO, their value was denigrated. That problem should be fixed now.

* * * * * * * * * * *

(28) SEC Requests Comment on: "whether the formula for making [the] Proportional Distribution should be revised to compensate the SROs more in accordance with the value of the information they contribute to the stream of consolidated information.""(p. 76).

(29) SEC Requests Comment on: "whether the formula for making the Proportional Distribution should be revised to reflect more directly the value that each SRO's information contributes to the stream of consolidated information made available to the public." (p.78).

(30) SEC Requests Comment on: "whether the formula for the Proportional Distribution should be revised to reflect...[the] objective [of] generat[ing] high quality market information that enhances the value of the stream of consolidated information..." (p.78).

AMEX Reply: The answer to all three of these questions is "Yes." There are some very specific items that should be part of any formula used in the future; these include the relation of bids and offers to the NBB or NBO and the willingness to commit capital. In the long term, a way should be found to include in a formula the ability to provide price improvement.

AMEX Reasoning and Comments: In order to encourage all market participants to pay more attention to the substance, rather than just to the form, of competition, a formula should include the following:

a) Market value-weighted trade size (which will serve as a measure of the amount of capital that a market committed or brought together; this will help to make sure the markets are as liquid as possible, a factor likely to become even more important as the U.S. markets begin to move to decimals starting on July 1);

b) Market value-weighted bids and offers that are the NBB or NBO (which will show one aspect of the amount of capital that a market is willing to commit or is able to bring together; again, this will help to ensure the markets are as liquid as possible, a factor likely to become even more important as the U.S. markets begin to move to decimals starting on July 1);

c) Time-weighted bids and offers that are the NBB or NBO (which will show another aspect of the amount of capital that a market is willing to commit or is able to bring together; again, this will help ensure the markets are as liquid as possible, a factor likely to become even more important as the U.S. markets begin to move to decimals starting on July 1);

The degree of price improvement a trade received relative to the NBBO, measured either as the price's being inside the NBBO bid or offer or the size of the trade's being larger in size than the relevant bid or offer in the NBBO, could also be part of the equation. However, this will add considerable complexity and may be better left for a second stage modification of the current, totally inadequate system.

When the time comes to include price improvement in the formula, there are a number of issues that should be built into the formula. When credit is being given for trades, besides having them be weighted for size, those that fall outside the NBBO as measured only by price, should receive no credit toward receiving market information revenues. Those that can demonstrate either type of price improvement might receive some increment relative to the credit that would be given to a trade that occurs only at the NBBO.

In order to make this system work, and to prevent its being "gamed," the SEC will have to make it clear that those who do not make real bids or offers as to price, time or size, will suffer significant sanctions.

In order for the system to work in options, the processor and the SRO systems will have to be able to measure time in seconds. This will require that a system that provides a central time stamp, now done in CTA for equities, be set up in options. Putting in place such a central time stamp would further assure the quality of the markets, as well as allow revenue to be awarded to those SROs that are truly providing the best markets.

* * * * * * * * * * *

(31) SEC Requests Comment on: "...[whether it is] possible to devise a pragmatic formula or algorithm...that would reward markets that provide "price discovery" to which other market participants look to set their own prices." (p. 79).

AMEX Reply: For equities, it is possible, but there are complexities that need to be resolved. There are some interim steps that can be taken, however. For options, it will depend on the degree of linkage that exists, and the ability to look at quota- tions in seconds, rather than just minutes, across all markets.

AMEX Reasoning and Comments: As is clear from earlier replies, the AMEX believes that rewarding those markets that provide price discovery, that is, provide real NBBs and NBOs, should be rewarded.

The formula the AMEX has suggested that will significantly improve on the rationality of allocating market information revenues among the SROs can be found in Appendix BB at the end of Part II.

* * * * * * * * * * *

(32) SEC Requests Comment on: "...[whether there is] a way to reward markets that are the first to publish quotations at the best prices and in the largest sizes." (p. 79).

AMEX Reply: Previous answers make it clear that the AMEX believes that rewarding SROs that provide quotations that are the NBB and the NBO, with credit given for the largest sizes, is a very important change in the formula for distributing market information revenues. The AMEX believes such a system should be instituted as soon as possible.

However, the AMEX does not believe that rewarding markets with the first quotation at the best price would be a good idea. Though in theory rewarding the "pricing leader" is an attractive concept, there are risks that such a system could be severely abused.

AMEX Reasoning and Comments: Were the "first" bid or offer that was the NBB or NBO to be rewarded, specific and strict rules would have to be put in place regarding the amount of time such quotations had to remain in place, as would minimum sizes that would qualify such quotations to be counted as meaningful (true, liquid) price discovery. Furthermore, strong and rapid enforcement action would have to be taken by the SEC against those SROs that do not stand behind their quotations, or that too frequently have excuses for why a significant portion of their quotations are not being honored.

If these steps are not taken, there is great likelihood that a plethora of quickly-entered and quickly-withdrawn quotations in small size will make a mockery of the attempt to improve the quality of the market, worsening it, instead. Unfortunately, it is not clear the needed effective policing and enforcement procedures can be put in place. Thus, the AMEX is not in favor of focusing on the "first" NBB or NBO, but rather on all that are the NBB or NBO, weighted for time and price to reward those that are truly beneficial to the market.

* * * * * * * * * * *

(33) SEC Requests Comment on: "how the results [of] a assess the value of quotations in an individual security...[could] be aggregated for all of the securities that are included in a Network." (p. 79).

AMEX Reply: The formula in Appendix BB provides the specifics. See the section immediately below for a brief synopsis.

AMEX Reasoning and Comments: In the new proposed system of distributing market information revenue among the SROs, the AMEX postulates that trades and quotations should be weighted equally in the formula. It is certainly possible to argue that one or the other may be more important, and therefore should receive greater weight than the other, but a 50/50 weighting between quotations and trades would appear to be a good starting point for the new system.

Counting Quotations The bid and the offer sides of each quotation would then be accounted for separately in the formula, since it should not be incumbent on an SRO to be both the NBB and the NBO in order to get credit on one side. The concept is to multiply the NBB price times the number of shares (or contracts), to create a value figure. That value figure is then multiplied by the number of seconds the bid is the NBB until a better one is created. This creates a time/value figure which is used to calculate, in part, how market information revenues are to be apportioned among the SROs.

The same process occurs for offers. In effect, time/value-weighted NBBs will represent 25% of the determination to which SRO the revenue from market information will go, time/value-weighted NBOs will represent 25%, and value-weighted trades will represent 50%.

In order to protect the integrity of the system, all figure will be adjusted at the end of the quarter to align them with the relative percentage that the security traded during the quarter relative to all other securities traded in that Plan.

Counting Trades Trades should not be counted based on the number of trades. A marketplace that provides a venue for doing 1,000 trades of 100 shares each arguably has provided much less liquidity, and value, than a marketplace that does one trade of 100,000 shares.

Payment from the Proportional Distribution as it relates to trades should be based on the market value of a security (or all in aggregate) traded. At some point in the future, this simplistic formula could be modified by giving different credit weights to trades that exceed certain percentages of normal market value traded in that stock. In addition, there might be no credit given when a trade takes place outside the price of the NBB or NBO, and extra credit given when improvement is provided, i.e., a trade occurs inside the price or volume of the NBB or NBO. However, until a number of issues cited in Part II and earlier in Part III are resolved, the SEC should move ahead with the simpler formula, which will do much more than the current system of dividing market information revenue among the SROs to encourage tighter markets and greater liquidity.

* * * * * * * * * * *

(34) SEC Requests Comment on: "whether certain individual SRO costs that most directly enhance the integrity of market information (principally, the cost of market regulation) should be funded as part of the Direct Distribution in addition to Plan costs..." (pp. 75-76).

(35) SEC Requests Comment on: "...whether a portion of market information revenues should be earmarked in the Direct Distribution to fund, in addition to Plan costs, SRO costs that directly enhance the integrity and reliability of market information." (p. 76)

(36) SEC Requests Comment on: " whether allocating market information revenues directly to fund specified market oversight and information integrity and reliability costs would further Exchange Act objectives." (pp. 76-77)

AMEX Reply: Although there is some conceptual appeal to the thought embodied in the three Requests for Comment above, the AMEX believes the answer to all should be "No."

AMEX Reasoning and Comments: The negative opinion regarding using the Direct Distribution to pay for some costs that are now paid in the Proportional Distribution (even though on the surface it might be a major benefit to SROs that en-gage in extensive regulation, such as the AMEX), is based on a variety of factors.

First, in general, people are going to go to a market where the basis of trading is fair and honest. With all equity and options securities' being traded on any and/or all SRO, ECN and ATS systems, any venue that does not maintain a pristine reputation regarding the quality, integrity, and reliability of its market will not be in business for very long.

Obviously, the SEC does not want to have a market, even if it were not to survive for long, be tainted, corrupt, regulated poorly, or operationally suspect. The argument that "...the funding would be shared among all users of market information, rather than falling on the particular SRO that incurs the particular costs..." (p. 77) has some truth in the theory, but is likely to be flawed in practice.

Fixing Spending Levels Is Difficult One of the difficulties is that if the Commission believes, as it seems to (a stance with which the AMEX would agree), that market regulation ought to be funded, at least in part, by a distribution of market information revenues, there should be some relation to the amount of regulatory cost and both the market value and quality of bids, offers, and trades that occur on that SRO.

By allowing a payment for regulation from the Direct Distribution, it is quite possible that the Commission will be a party to either over- or underpayment of costs that are real for any specific SRO. The fact is that without some objective measure, such as a formula similar to the one the AMEX proposes in Appendix BB, there will be no way for the SEC to determine whether a market is submitting too much in costs for which it is claiming reimbursement.

In today's environment, the SEC sends in teams of inspectors to the relevant SROs to study what they are doing in their regulatory arms and to make suggestions about how to improve the efforts. However, the SEC generally has not mandated specific spending levels, or even relative spending levels (say, to the number of bids, offers, trades, people, firms, etc.) that is expected. Instead, the Commission watches the performance of the SRO as it practices its craft and looks at the results that issue from that.

It is highly unlikely that the SEC will want to try to set the specific spending levels that each SRO can claim in order to try to ensure that rates of reimbursement from the Direct Distribution will be fair. In fact, there is no way to tell before the fact whether an SRO is spending too much or too little on regulation, much less whether the claims it submits for reimbursement from the Direct Distribution are accurate.

Market Quality Should Determine Revenue Allocation Second, if a payment were to be made from the Direct Distribution to pay for regulation, markets that are providing good quality quotations and liquidity as measured by trades would earn less money for their effort. This is true because once a payment were made in the Direct Distribution, there would be less money available for the Proportional Distribution to pay those that have, in their systems or otherwise, generated an improvement in the quality of the market (and therefore, of market information).

Though ensuring that regulation is done in the highest quality manner possible is certainly one task of the Commission, ensuring that the overall quality of the market is as high as it can be is surely its ultimate responsibility. It is illogical to give money to a market that may never provide a quote that is the NBB or NBO, the NBB or NBO in an acceptable size, or small trades (meaningless to the market's liquidity) that are not frequently the result of payment-for-order flow that has deprived investors of that part of the payment that went to the broker-dealer for delivering the order to that SRO. Yet, that could be the result of trying to fund regulation through the Direct Distribution.

Third, the SEC is (perhaps rightly) worried that in a for-profit world, those areas that do not provide a return, such as regulation, will be neglected. That is an important reason for the earlier argument the AMEX evinced that there should be an attractive ROE allowed on SRO market information. If an adequate ROE is permitted, then, to the degree that regulatory costs are included in setting market information fees, by definition, profit will be earned on those regulatory costs. They would then be less likely to be neglected.

In any system in which the costs of regulation are paid for by SRO market information costs, one SRO is effectively funding another SRO's costs of having the function. The question must, then, be asked whether it is fairer to have those revenue transfers be made solely as a result of the claims of an SRO regarding what its regulatory costs are, or whether the revenue transfer should be the result of superior performance of an SRO's market in totality.

The answer to that should clearly be that if one SRO is to effectively take money from another SRO, that money must be "earned." For that reason alone, the AMEX believes that the payment of market information revenues should come from the Proportional, and not the Direct, Distribution.

Member Regulation Should Be Included in Costs A final brief point, made earlier, is that though the Commission states the costs of regulation that might be paid from the Direct Distribution "could include primarily the costs incurred by the SROs in performing their market regulation function (as opposed to member regulation)..." (p. 76), such a distinction should not be made. Though because the word "member" in the term "member regulation" carries the implication that it benefits only the SRO involved, "member regulation," in fact, is a form of regulation that is very valuable in terms of assuring the integrity of the market and its associated information for all participants in the market.

As was brought out earlier, the level of member regulation costs that might be included in setting market information fees is, to some degree, a function of the fees charged to the members themselves at one SRO compared to those at another. The risk that the Commission will get drawn into mandating member fees at separate SROs (in order to assure a level playing field when determining what market information fees should be) is certainly present if a cost-based system of setting the market information fees is imposed. To compound that risk by allowing a payment for the regulatory fees before any trading takes place and an SRO demonstrates that its market has earned the right to such fees does not appear to be the wisest decision that might be made on this question.

* * * * * * * * * * *

(37) SEC Requests Comment on: "...[whether] identification of the cost of market regulation [can] be a reasonably objective task that could be accomplished without excessive accounting and auditing costs..." (p.77)

AMEX Reply: Theoretically, "Yes," but in practice, "Probably Not."

AMEX Reasoning and Comments: There are likely to be numerous points of conflict among the parties in a cost-based system of setting fees. That should be obvious from the experience of the utility industry prior to deregulation in the 1990s. Rate cases did not drag on for months, or even years, and end up frequently in the courts, because all the parties agreed to the submitted information and the final decision of the utility commission.

To assume that anything will be different in trying to set market information fees based on costs (or estimated costs, or quasi-costs, or any variant thereof) is being extremely optimistic. It is hoped that the SEC has in place the funding for a much larger staff to examine all the data from the contending parties that is likely to come pouring down the pipe when the first case comes up for a decision.

Most of the SROs are probably able to determine their costs in broad categories now. However, there is likely to be significant dissension among various sides of what at times may look like a triangle (SROs, vendors, and users having one side of the triangle each) and at others a square, pentagon, decagon, or centigon as the sides fragment into finer coalitions as their interests are threatened.

The arguments will, in most cases, center on things such as "what are the correct categories to include?", "what are the right percentages to reduce these gross cost categories down to net cost categories?", "should member fees already imposed that offset these costs be considered in allowing costs to be included in setting market information rates?", "what should happen to market information fees if member or transaction fees are raised or lowered?", "are the benchmarks for allowable ROEs correct?", "how should new costs submitted by in-coming SROs be counted relative to the existing SROs?", etc.

None of these problems will be disposed of quickly, nor without calls to "prove" the figures provided or disprove opposing figures and assertions.

* * * * * * * * * * *

(38) SEC Requests Comment on: "...[whether] there are pragmatic methods that could simplify...[the]task [of identifying the cost of regulation] while still achieving the goal of adequately funding appropriate costs." (p. 77).

AMEX Reply: Theoretically, "Yes," but in practice, "Probably Not."

AMEX Reasoning and Comments: It is not likely the Commission will be able to avoid some, if not all, of the battles enumerated in Request for Comment (37) above once it embarks on a path of trying to tie market information fees to a specific target level of costs.

There are two ways the SEC could try to do so, though it is not clear how successful the effort will be. There could be a spending guideline imposed that would relate trading and quoting activity to the costs that will be accepted. However, that effort could well bog down on arguments over the costs of running different technology platforms; whether costs escalate in a straight line as activity rises, or at a greater or lesser rate; and how effective the dollars that are accepted as part of the market information cost base are being spent.

The other way the SEC could attempt to bring some simplicity to the process would be to require each SRO that wants to include appropriate costs in the market information fee-setting exercise to submit an annual report on the specifics of its spending in each of the areas. This, unfortunately, will still not finesse a great number of the contentious issues and areas of disagreement that will arise as those who advocate various positions jockey to put forth their own data or their own "spin" on the data of others.

* * * * * * * * * * *

(39) SEC Requests Comment on: "whether direct funding would create an inappropriate incentive for the SROs to increase these costs beyond reasonable levels." (p. 77)

AMEX Reply: Yes.

AMEX Reasoning and Comments: Any system that uses costs as a benchmark for setting any sort of fee will carry the risk that to set the fees as high as possible, the costs will be added at higher rates than would otherwise have occurred. Allowing these costs to be defrayed "off the top" without in any way "earning" them (as, for example, has been suggested through the mechanism of the Proportional Distribution, using the new, proposed formula in Appendix BB) can only exacerbate the risk.

Though allowing an ROE to be earned may throw a further level of risk into the mix, the danger posed by not allowing any ROE in a for-profit world of SROs, is even greater and probably not worth it.

* * * * * * * * * * *

(40) SEC Requests Comment on: "whether any other category of SRO costs that directly enhance the integrity and reliability of market information should be funded in the Direct Distribution." (p. 77).

AMEX Reply: No.

AMEX Reasoning and Comments: All the reasons the AMEX would disagree with the idea of funding regulatory costs through the Direct Distribution, as elaborated on in Requests for Comment (34)-(36) would apply as it pertains to any other cost category. That would include " systems with sufficient capacity and reliability to handle the highest volume days..." (p. 77).

Even though technology spending might seem more objective as a measure when thinking about paying for a cost through the Direct Distribution, it is probably not. How would the costs of running and maintaining different hardware platforms be handled? The cost of developing software applications? Expansion of an SRO well beyond any reasonable expectation of its percentage share of the capacity, even on the highest volume days? The expense and risk of write-offs inherent in an older SRO's moving current technology that may be higher cost than that of some newer competitors to the most advanced and lowest cost (at least in theory, if everything works) technology? The use of market information revenues by an SRO to fund competitive technology efforts it would otherwise not be able to afford?

* * * * * * * * * * *

(41) SEC Requests Comment on: "whether some portion of technology costs that directly relate to the integrity and reliability of information (such as costs incurred to comply with policies set forth in the Commission's ARP Releases) should be funded in the Direct Distribution." (p. 77).

AMEX Reply: No.

AMEX Reasoning and Comments: Since ARP-required expenditures are part of being an SRO, the same problems as were highlighted in Request for Comment (40) probably make paying for these costs through the Direct Distribution unwise.

IV. Plan Governance, Administration and Oversight

(42) SEC Requests Comment on: "whether [current] governance structures [that do not provide for broader securities industry or public participation] should be broadened to include such parties as vendors, broker-dealers, and investors." (p. 82).

AMEX Reply: Yes, in some cases.

AMEX Reasoning and Comments: To some extent, the concern expressed by the Commission that vendors, broker-dealers, and investors have no input into the governance process of the Networks is overstated.

All those parties have the ability to directly participate during the comment periods that are available before any of the Plans' filings are made final, if not effective. In addition, both broker-dealers and the public are represented on the Boards of Directors of the SROs that are part of the Plans. Thus, they have the ability to vote against any effort on the part of the Plan(s) they participate in even before a final vote at the Plan level is taken.

All three parties also have an indirect voice in influencing the governance of the Plans, as they all have access to the SEC if there is an issue they feel is negative to their interests that they would like to have addressed.

It is not clear what would be gained by giving any of these parties a more direct voice in the governance process, such as having a vote in the Plans' Operating Committees. If any of the representatives of the firms or the public that now sit on the SRO Boards felt there were a problem, that is a very powerful forum in which to bring up any question.

A particular concern is whether this would be appropriate in the case of the vendors. The vendors collect 95%+ of all market information revenues, while the SROs collect the rest. This sets up an immediate conflict of interest if the vendors are part of the process of setting fees for SRO market information. They might well take the view that to the degree they can reduce SRO market information fees, they have more room to raise the fees for their own data streams, thereby enhancing their profits at the expense of the health of the markets.

Despite these reservations, and the AMEX' view that broker-dealers and the public are adequately represented in the governance of the Plans, the AMEX would support having all the parties be part of the Operating Committees on a consultative basis.

Again, it is important to emphasize that, as the Commission points out on pages 37-38 of the Release, the process for ensuring that market information fees are set at levels that are fair, equitable, and reasonable, worked well for 24 years. When a strong objection was raised in 1997 to a change in Network A nonprofessional fees to a level that had been accepted previously by all parties for NASDAQ and OPRA market information fees, all three Plans then began a process that has resulted in lower fees across the board.

It certainly can be said that these reductions were accomplished only with a great deal of contention. In the course of commerce, that should not be considered surprising, shocking, or a reason to replace a system that has worked in as simple and as efficient a manner as one could wish, and in a way that with almost no exceptions has been considered satisfactory by all parties.

The AMEX reiterates that it enthusiastically embraces changes in the method of distributing market information revenues among the SROs in order to provide real incentives to improve the quality of bids, offers, trades, and liquidity in the market. However, though the AMEX has said it will certainly participate in a move to cost-based pricing, there remain real concerns about replacing a system of setting SRO market information fees that has worked well with one that could well lead to much more paperwork and many more disputes down the road.

* * * * * * * * * * *

(43) SEC Requests Comment on: "...[whether] non-SRO parties [should] be included on the Operating Committee[s, the governing bodies of the Plans]?" (p. 82).

(44) SEC Requests Comment on: "In what capacity should such representatives be allowed to participate (for example, voting or non-voting)?" (p. 82)

AMEX Reply: There may be some reasons to have non-SRO parties on the Operating Committee, but it should probably be in a non-voting capacity.

AMEX Reasoning and Comments: The participation of these parties on the Operating Committees could be helpful in a number of respects. A major problem that has arisen in the last year is concern over anti-trust issues. If the SEC provided a formal order that, having broker-dealer, investor, and vendor representatives sitting in the Operating Committee as observers and to provide input would resolve the anti-trust concerns and risks, their presence should be welcomed. In the current climate, the Operating Committees, and thus, the Plans themselves, are becoming much less effective as decision-making entities than they once were, and anything that improves that situation should be considered positive.

Having non-SRO representatives at these meetings also would simplify lines of notification and communication, as these representatives should be charged with reporting back to their various constituencies coming changes, including those for which each constituency must prepare pro-actively. The representatives could also provide ideas and suggestions that would allow the SROs to react more quickly to the needs of those who use market information.

The AMEX sees no reason for non-SRO representatives' having a vote on the Operating Committee. The Operating Committee must be able to manage the policy direction of the Networks, and to have entities whose goals and objectives may not be coincident with those of the Plans is not good management practice or theory. If the vendors, broker-dealers, or investors want to vote on actions contemplated by the Networks, they have forums to do it in the Boardrooms of the SROs or with the SEC when proposals are filed with the Commission.

* * * * * * * * * * *

(45) SEC Requests Comment on: "Should additional committees with broad participation be established to address the particular issues of most direct concern to parties that are not SROs (for example, a committee for establishing or reviewing fee structures)?" (p. 82).

AMEX Reply: No.

AMEX Reasoning and Comments: There are already mechanisms in place for reviewing fee structures and changes. When fees are to change, the SIA meets with the relevant Network Administrator, and the fees are filed with the SEC and can be commented upon in that forum. If representatives of vendors, broker-dealers, and investors were to be added as observers/commentators to the Operating Committee meetings, they could provide input at that point. Adding an entirely new committee (or set of committees) would not seem to add much in the way of efficiency or oversight.

There is also a Technical Committee for the Plans, which consists of representatives of the SROs and SIAC. In certain specific instances, and on a regular semi-annual basis, representatives of the vendors and broker-dealers are included. In addition, the Technical Committee publishes any new technical specifications and gives time frames for upcoming changes.

Inclusion of vendor and broker-dealer representatives in every Technical Committee meeting is not recommended. The discussions are very specific and highly technical in nature, and the risk that vendor and broker-dealer representatives will bring in extraneous, business-policy issues is high. The end result would be less efficient and effective meetings, which would not be positive.

To the extent a decision is made to include vendor, broker-dealer, and investor representatives on currently-existing committees, or to create new committees, the AMEX believes these representatives, or the new committees, should not have a vote regarding decisions for which the Plans have been, and in the future should be, ultimately responsible.

* * * * * * * * * * *

(46) SEC Requests Comment on: "What should be the mechanism for selecting non- SRO representatives to a committee?" (p. 82)

AMEX Reply: They should be selected by their industry trade groups.

AMEX Reasoning and Comments: Such organizations as the Securities Industry Association, the Investment Company Institute, and various organizations that purport to represent individual investors could be brought together by the SEC and asked to choose one or two people to represent those organizations collectively. (This may be another problem area for the Commission: taking on this role will take staff time to organize and hold meetings, and the staff may end up arbitrating disputes over representation among the various constituencies.)

* * * * * * * * * * *

(47) SEC Requests Comment on: "If [non-SRO parties were] given the power to vote, what should be the relative proportion of voting weight between the SRO and the non-SRO representatives?" (pp. 82-83).

AMEX Reply: "One group, one vote" should be the rule.

AMEX Reasoning and Comments: Since all the SROs have one vote each in each of the Plans, the question would arise why, if non-SRO parties were granted voting representation on the Committee, they were not granted the same proportionate level as the SROs. That is a fight that could be avoided by giving vendors, broker-dealers, and investors one vote each.

That would mean that, if the three ATSs that have indicated some interest in, or applied for, membership in CTA were to join, the non-SRO parties would have 20% of the vote in CTA. Assuming the ISE were part of OPRA, the non-SRO parties would have 33% of the vote in OPRA.

Current practice is that the non-SRO parties have an opportunity to comment to the SEC on filings by the Plans. If these parties are given a vote in the Plans' Operating Committees, they would, in effect, have a second vote by using the comment or appeal process.

Furthermore, the broker-dealers have an additional vote, viz, the ability to move business from one SRO to another to express a viewpoint. That, in and of itself, would seem to constitute a powerful vote, both now and in the future.

* * * * * * * * * *

(48) SEC Requests Comment on: "whether, as an alternative to formal participation in Plan governance, the creation of an industry advisory committee on market information arrangements would constitute a more efficient and flexible vehicle to convey a broad range of views to the Plans and to the Commission." (p. 83).

AMEX Reply: This idea has some appeal, perhaps, as it pertains to flexibility, though it still may be cumbersome to set up and hold meetings, and so may not be all that efficient.

AMEX Reasoning and Comments: If there were going to be committees to deal with market information "arrangements," there should be a specific definition of what types of "arrangements" these committees would concern themselves.

Some obvious areas might lend themselves to separate advisory committees: fee-setting and structure, distribution arrangements among the SROs that might achieve more efficiency and liquidity in the market, technology, and product development.

One problem with moving forward with the suggestion, however, is that to ensure that these committees do not constantly go over ground that has been covered in the Operating Committee meetings, representatives from these committees may have to sit in on the Operating Committee meetings. This could result in very large and unwieldy Operating Committee meetings.

Alternatively, representatives from the Operating Committee could sit in on the new committees' meetings. That, however, will mean a substantial commitment of time on the part of Operating Committee members who are designated to do this.

* * * * * * * * * * *

(49) SEC Requests Comment on: "whether the Plans should establish industry-wide standards for administering their fee structures, and if so, the most appropriate means for the Plans to act jointly in developing such standards." (p. 83).

AMEX Reply: There should be industry-wide standards, and all interested parties, working with the Commission as coordinator, should have input into the process.

AMEX Reasoning and Comments: In fact, the Plans have all begun moving, or considering moving, to vendor billing, which will help simplify part of the adminis-trative process. To the extent that both this trend, as well as changes in (and simplification of) fee structures that result from decisions made in response to this Release, move forward, many (if not all) of the differences in contracts and other agreements that now make the process so difficult may disappear.

In any case, the SEC should take the lead and coordinate meetings of SROs, vendors, broker-dealers, and investors, along with others such as the Financial Information Forum. In the end, customer satisfaction on this issue is crucial, particularly on issues such as standardizing contract language.

There should also be consideration given to establishing, at the vendor level, or with some outside entity, a central billing authority where real economies of scale in the billing and administration process can be brought to bear. If that were done, the headaches and costs related to this entire aspect of the market information industry could be sharply reduced for all parties.

* * * * * * * * * * *

(50) SEC Requests Comment on: "...[whether] pilot programs [s] eliminated entirely or should the Plans have some flexibility to experiment with innovative services and fee structures without first going through the process of a Commission filing and public comment." (pp. 83-84).

AMEX Reply: Pilot programs should be permitted, with some tighter controls.

AMEX Reasoning and Comments: The SROs need some way to experiment with various programs on a limited and restricted basis. In this way, it is easier to assess whether a proposal will have its intended effects, advancing the cause of further improving market information quality or the process in some way, and not doing major harm to either an SRO or a user of market information if the program does not have the intended effects.

* * * * * * * * * * *

(51) SEC Requests Comment on: "If pilot programs should continue in some form,...should [they] be limited to a specified time period (for example, one year), after which the program could not be continued unless it was filed with the Commission." (p. 84).

AMEX Reply: There should be limitations, but with some flexibility.

AMEX Reasoning and Comments: There are two examples of the type of flexibility that would be useful to the SROs and not harmful to the users of market information. One might include those instances in which changes in the industry are occurring rapidly, and in order to obtain an accurate reading of the effect of a program, it might be necessary to extend the program for a further period of time.

The effect of doing so, however, could be mitigated if an "effective on" filing were again made. Should there be a significant number of adverse comments, consideration would have to be given to dropping it, either as a pilot, or in its entirety.

This leads to the second type of flexibility, which would involve the SRO with the pilot's going to the SEC after a year, and in a private meeting, asking for a continuation of the pilot for a specified additional time without a filing. If the SRO could defend adequately why that is in the public interest, it would be allowed to continue it as requested.

* * * * * * * * * * *

(52) SEC Requests Comment on: "whether the terms and conditions of all pilot programs should be made available to the public in some fashion prior to the initiation of the program." (p. 84).

AMEX Reply: Yes.

AMEX Reasoning and Comments: Just because pilot programs are a good way to test an idea does not mean they should not be done in a public way. They do not reveal the business strategies of the SROs themselves, so there is not the issue of trade secrets' being given away.

For this reason, the AMEX suggests that consideration be given to having pilot programs commence with an "effective on filing" notice to the Commission. In this way, if there are to be objections, they can be brought out in the open quickly. However, because pilot programs are supposed to be experimental, objections should be given less weight (or have a greater burden of proof of great harm during the experimental period) than would be the case with the filing related to a program meant to be permanent. Otherwise, many worthwhile pilot programs might never take place solely due to the forces of inertia and self-interest.

One advantage to having pilot programs on a "silent" basis is that there is no resultant clamor to be part of the experiment on the part of organizations not included in the experiment. Thus, the SROs have been able to avoid the issue of discriminating against a party that uses market information by refusing to include it in the experiment even when requested.

Were a "sunshine" system to be initiated, the Commission must put in place, or agree to, some scheme by which the SRO will have the ability to limit the number of participants (if that is part of the pilot) until the impact and the feasibility of the program is clear. This will involve being able to set those limits in a way that is not subject to challenge. It is not clear whether it will be possible to do so.

* * * * * * * * * * *

(53) SEC Requests Comment on: "whether participation in the process of setting and administering fees should be broadened to include vendors, broker-dealers, and users of market information." (p. 62).

AMEX Reply: Only in cases where they are observers providing input; otherwise, "No."

AMEX Reasoning and Comments: The AMEX's views on this subject have been embodied in the answers given for Requests for Comment (42) - (48) above. Only if it is felt the inconvenience and inefficiencies resulting from having more people and more meetings as part of the effort to set and administer fees will be more than offset by the peace of mind that will develop if those people feel they are part of a more participatory process, would it be worth considering.

* * * * * * * * * * *

(54) SEC Requests Comment on: "whether there should be specific requirements relating to the frequency and timing of proposed fee changes." (p. 65)

AMEX Reply: If a cost-based system is going to be put into place, "Yes." Were it determined that the current fee-setting mechanism were sufficient, "No."

AMEX Reasoning and Comments: A cost-based program of setting market information fees will almost certainly require annual meetings. These will be used to determine cost changes, percentage allocation changes, etc. Once the relevant items have been examined, a determination will have to be issued, either that fees will be held steady or that they will change. If the latter is the decision, notification will have to be given regarding the extent to which they will change.

V. SRO Member Rebates

(55) SEC Requests Comment on: "whether the SROs should be permitted to rebate market information revenues to their members." (p. 76).

AMEX Reply: Yes.

AMEX Reasoning and Comments: Perhaps the major issue that will have to be addressed if any limits are to be placed on rebates is: what constitutes a rebate of market information fees versus a reduction of other SRO fees? It may prove difficult to limit any fee reduction, including transaction fees.

There will be questions regarding "which bucket are they in?" that cannot be answered easily. In a period in which SROs are competing fiercely for order flow based on the cost to trade in the SRO's market, to limit the ability to cut transaction fees, provide relif from fees during a profitable year, etc. would constitute a perhaps-dangerous interference by the SEC in the competitive arena.

If a cost-based solution to setting market information fees is instituted, an SRO should probably be able to provide such rebates only after it has earned back through the new proposed distribution mechanism all the costs that it submitted prior to an allowed rate of return. Effectively, then, only the SRO's rate of return profit (if earned) is permitted as a member rebate.

* * * * * * * * * * *

(56) SEC Requests Comment on: "whether...rebates...given to...members responsible for effecting...transactions that resulted in a Network's revenues being distributed to the SRO are consistent with the Exchange Act objectives of fair competition." (p. 80).

AMEX Reply:. Certainly, rewarding members when the SRO or the members pay for order flow or engage in practices such as double-printing riskless trades is not consistent with the objectives of the Exchange Act. However, the problem still comes down to determining what constitutes a rebate versus a "legitimate" fee reduction.

AMEX Reasoning and Comments: Generating revenue for an SRO in and of itself is not the problem. The problem comes when the rebate to the member is solely a function of buying order flow, but doing nothing to improve the NBB or NBO, or liquidity.

Not to be repetitive, but however repugnant the policy of payment-for-order flow may be to the AMEX, limiting it is difficult. A determination would then have to be made about how lowering or eliminating fees to members or order flow providers fits the definition of paying for order flow. That, in turn, risks involving the SEC in dictating in which competitive business strategies an SRO can engage. Doing so, that is, taking the decision about what is required from the marketplace and putting that decision in the hands of someone who is not close to (or may not understand) the market, puts great risk into the system in terms of its long term health and competitiveness.

* * * * * * * * * * *

(57) SEC Requests Comment on: "...[whether] rebate programs constitute an equitable allocation of an SRO's charges among its members when only selected members receive a rebate based on their transaction volume in a particular type of security..." (p. 80)

AMEX Reply: No.

AMEX Reasoning and Comments: As has been stated, the AMEX does not believe SROs should be rewarded with market information revenues based on the number of transactions going through the marketplace, but rather on the value of the transactions, as that tells something about the SRO's having provided liquidity. Similarly, the AMEX does not believe a member should receive a rebate unless all members are eligible for rebates in all securities that are traded in that Network.

* * * * * * * * * * *

(58) SEC Requests Comment on: "whether changing the rules for distribution of Network fund information integrity and reliability costs directly and to reward the SROs that provide the highest quality market information would address the extent to which rebates could constitute unfair competition." (p.80).

AMEX Reply: Changing the rules governing the distribution of market information revenues will address to a great extent the problem that rebates pose for the fairness issue. Funding costs directly would likely make the problem of rebates even worse.

AMEX Reasoning and Comments: As was indicated in the answers to Request for Comment (34) - (36), (39), and (40), funding cost recovery through the Direct Distribution would have many pernicious effects. The result would likely be even more abuses of the market information revenue distribution mechanism than exist today. Changing the rules for the Proportional Distribution pool after taking out a large amount of money through the Direct Distribution pool will not help.

The problem is that the money taken from the Direct Distribution pool would then be used to fund rebates and would result in the new distribution rules governing the Proportional Distribution's having little positive impact in correcting the shortfalls that exist in today's rules that are failing to encourage the creation of a more liquid market with tighter spreads.

However, leaving the Proportional Distribution as the mechanism to distribute most market information revenue will result in improved spreads and liquidity provision, for the first time, having a measurable meaning in (and, thus, a measurable impact on) market behavior when it comes to awarding market information revenue.

Payment-for-order flow in those cases where the NBB or NBO (and so the quality of the market and its information) are not improved will be much reduced because the funds will not be available to do so. Payment-for-order flow in cases where all the positive things listed above occur will be justified because a service to the investor has really been provided.

* * * * * * * * * * *

(59) SEC Requests Comment on: "...[whether] rebate programs indicate that market information revenues exceed self-regulatory funding requirements." (p. 80).

AMEX Reply: Not necessarily.

AMEX Reasoning and Comments: No accounting has been done that could prove or disprove what appears to be the Commission's concern in view of this Request for Comment. Furthermore, there has not been advanced any proposal that would set market information fees solely as a function of regulatory costs. However, a high level of rebates may indicate that investment in areas such as regulation and technology are not receiving the attention or investment that perhaps they should.

VI. SRO and Plan Disclosure

(60) SEC Requests Comment on: "whether the Plans and SROs should provide greater public disclosure concerning their fees, revenues, and costs..." (p.62).

AMEX Reply: The SRO disclosure issue should be left up to the SROs. The Plan disclosure issue should follow the pattern of other similar associations.

AMEX Reasoning and Comments: To force the SROs to change or increase the amount of disclosure they provide may not be appropriate. Every business makes disclosures in a manner that allows those who need information about the entity to receive an accurate picture of the firm's financial condition and results of operations. However, all businesses do so in a manner that protects them from having their competitors obtain too accurate a view of the firms' business strategy, strengths, and weaknesses. There is no reason the SROs, particularly as the for-profit world develops, should be forced to operate any differently.

There are many other associations operating in the U.S. If the Plans' disclosure is significantly less that the norm, that should be addressed. If not, unless there is a reasonable allocation of costs that the Plans are allowed to incorporate into the cost of market information generation, more encompassing (or even today's) statements would be very misleading.

* * * * * * * * * * *

(61) SEC Requests Comment on: "whether the Plans should be required to make annual filings for the Networks that would be available to the public...[to] include (1) a complete listing of their fees,...(2) the number of users participating in each of their different fee programs, and (3) audited financial statements setting forth their revenues (including an itemized listing of revenues attributable to their different fees), expenses, and distributions." (p. 81).

AMEX Reply: In general, the AMEX would have no problem as Administrator of Network B in doing (1), though (2) and (3) are more problematic.

AMEX Reasoning and Comments: The risk is that due to the "derived" nature of the market information generation cost structure, the statements could be very misleading. However, the ability to do what the SEC suggests is there. All incremental costs, and they could be substantial, should be passed on in the Direct Distribution.

"...all their fees..." should be all filed and current fees. Pilot program fees should probably be omitted, at least in the initial implementation.

The "number of users participating in each...program..." is not relevant unless the users reveal how many brokers, investors, money managers, traders, etc. are affected. Unless these data are also reported, which is likely to be fiercely resisted by the broker-dealers, the figures the Commission has suggested are relatively meaningless, and may even prove misleading.

If "(3)" in its entirety refers to the SRO, the comments on disclosure and business strategy that were made in Request for Comment (60) above would apply here, as well.

* * * * * * * * * * *

(62) SEC Requests Comment on: "whether the SROs should be required to provide greater disclosure of their financial condition, including...costs associated with the performance of various SRO functions." (p. 81)

AMEX Reply: No, as they now have extensive disclosure; more Plan disclosure will either prove misleading or will reveal too much about the SROs. The one exception is the extent of spending on regulation.

AMEX Reasoning and Comments: In a world of for-profit SROs, it will be necessary for the SEC to more closely monitor the overall financial condition of each SRO. In addition, the need to make sure that all are being accurate in the amount of money they claim they are spending on regulation and technology costs also assumes greater importance, particularly if a cost-based SRO market information fee-setting mechanism is imposed.

If the Commission decides to go forward with cost-based pricing of market information fees, more disclosure may be necessary, particularly as the SEC states that, "at the very least, the SROs will need to provide financial disclosures that are sufficient to support whatever approaches ultimately are adopted for the evaluation of fees and distribution of revenues." (pp. 81-82).

To the degree there are audit costs associated with these greater disclosures, they could come from the Direct Distribution. The formula that might be used would aggregate all SROs' audit costs and distribute them in the same percentages as the Proportionate Distribution percentages for the year. This formula will prevent a minor market (or one that does a poor job) from loading audit costs into the Direct Distribution in order to get a larger payout than its size or performance would warrant.

Exactly what breakdown of "SRO function costs" is needed must be better defined. Should it be by area, personnel, technology operation and development, etc.?

Most importantly, any further disclosure that is made in order to allow the SEC to monitor costs being submitted for the purpose of setting market information fees, must be non-public. It is not reasonable to require an SRO to disclose detail publicly that could provide their competitors with information that could damage the SRO in the short- or long-term. However, history has shown that once information is submitted to the SEC, even on a "confidential" basis, it has often ended up in the public domain via Freedom of Information Act requests. Unless a solution to this problem can be found, more mandated disclosure could cause material commercial harm to SROs in the developing for-profit world.

1 The evidence is provided by the SEC's recitation of the history of the Plans and disputes that have arisen on pp. 37-43 of the Release.
2 The best examples are the electric utility and telephone industries. The reins of price regulation were considerably loosened for most electric utilities and telephone utilities in the 1990s.
3 The "Vendor Display Rule," which does set guidelines regarding what vendors must display if they are going to be in the market information business, would appear to be silent on the SEC's ability to rule on vendor pricing.
4 See footnote 2
5 Famous rate cases include numerous units of Southern Company, whose battles with state regulators in the 1970s went on for years, and frequently ended up in court appeals.
6 Based on feedback from February meetings of market information users.
7 To the extent that a Plan is a subsidiary of the SROs that are members of that Plan, the Plan stands in the same relation to those SROs that a subsidiary or division of a company would stand in relation to its corporate parent. It is commonly acccepted by auditors that as part of the process of good capital allocation monitoring, a corporate parent will hold a subsidiary to a Risk Adjusted Rate of Return on Capital (RAROC) hurdle to ensure that the subsidiary earns at least a minimum, required rate of return on that capital. It is that hurdle rate that is the cost-of-capital allowance for which the AMEX is arguing in this section.
8 Though the SEC has frequently examined internalization and payment for order flow, the SEC to date has taken no definitive stance. Regarding ATS/SRO issues, the SEC issued a Concept Release (#34-38672) on May 23, 1997 and then adopted much of it in Release #34-40760 on December 8, 1998.
9 The SEC issued Concept Release 34-42450 on February 23, 2000, requesting comment on market fragmentation issues within a 60-day period.
10 Examples include electric utilities, telephones, and the meatpacking industry.
11 Various SROs charge no transaction fees and then rebate a portion of transaction charges to their market - makers, with that rebate then used to pay for order flow to increase market share, while others selectively rebate market information revenue earned on Network B securities. Many discount brokers use internalization of order flow, payment-for-order flow, and layoff arrangements that have the effect of directing order flow to specific SRO's.
12 For the last half of the 1990s (1995-1999), the figure has been close to 25% based on the numbers for All Firms' ROE 1995-98 in the 1999 Securities Industry Fact Book issued by the Securities Industry Association ("SIA"), as well as preliminary reports of 1999 results for the industry reported through 1/31/00
13 AMX data shows average daily quotations and trades 1/3/00 - 2/23/00 of 11, 733, 218, vs. 298,517 for all of 1994, an increase of 3,830%; SIAC data show an increase of 2,660%.
14 Wall Street Journal, February 1, 2000; the figure is understated by as much as 50% because it excludes stock options and bonuses, which are a substantial portion of securities industry compensation. If the $80,000 figure were raised to $120,000, the relative value of SRO market information would appear even greater.
15 The release gives a figure of $18.50 - $50.00 per month, which is $0.86 to $2.40 for a 21-day trading month; however, as Part III, Request for Comment (12) indicates, Network B and OPRA fees have low ends of $13.60 and $10.50, or $0.65 and $0.50, respectively.
16 Based on an 8 hr day and $320 per day; however, if compensation were 50% higher due to bonuses, the per-hour costs would be $60.00.
17 These figures are based on $0.11 / $40 = 0.28% to 0.74 / $40 = 1.9% for the per-hour of use calculation; the daily (and, thus, hourly) cost is $0.50 / $320 = 0.16% to $2.40 / $320 = 0.75%. If the compensation were closer to $120,000, the range of the percentages, rather than being 0.16% - 1.90%, would be 0.11% - 1.27%.
18 Figures are both from the SIA 1999 Securities Industry Fact Book for NYSE member firms and the reported income statements of the firms that are publicly held.
19 Hambrecht and Quist February 4, 2000 report, "Charles Schwab Corporation"
20 See Part III, SEC Request for Comment (13); the figures are derived from p. 71 of the Release.
21 The average discount broker trade is 500+ shares; the average U.S. stock is priced at $40+, so the calculation assumes a heavier-than-average-weighting of low-priced stocks.
22 Data from June, 1999 Board of Governors of the Federal Reserve System "Annual Report to the Congress on Retail Fees and Services of Depository Institutions."
23 SEC Request for Comment (17) and Request for Comment (19)
24 See footnote 20
25 As reported in various articles in the Wall Street Journal and the trade press, as well as discussions with discount firm executives.
26 SIA 1999 Securities Industry Fact Book shows that in the period 1994-98, the ROE for discounters ranged from 26.4% to 42.6% and averaged 34.9%. The industry as a whole had an ROE range during that time of 3.3% to 29.1% and averaged 19.8%. Note that the industry figures are biased upward by inclusion of the higher returns of the discounters, so the performance of the discounters relative to all others in the industry is even better than the figures indicate.
27 All figures cited for vendors in this section of Part II are from Waters, February 2000, pp. 47-51.
28 It is estimated that Networks A and B, NASDAQ, and OPRA had revenues of $425 - $500 million, and the futures exchanges had revenues of $125-$200 million; the figure of $600 million is in the upper end of the range.
29 On November 15, 1999, the SEC approved Network B's filing of a $1.00 maximum monthly fee for nonprofessionals; on January 4, 2000 the SEC approved a Network B $500,000 enterprise license fee for professionals. On October 7, 1999, the SEC approved both these fees for Network A.
30 In the "Reported Trades Sent Through ITS" report matrices for October - December, 1999 (issued by the NYSE), the figures show that the NYSE and AMEX had non-execution rates of 6% and 10%, respectively. All other SROs had rates of 25-43%, and these were quite consistent from month to month.
31 For example, on the Board of the AMEX, there are 9 public, 5 broker-dealer, and 1 investment advisor representatives; on the NASD Board there are 12 public, 10 broker-dealer, and 5 investment advisor representatives. The NYSE has 12 public, 10 broker-dealer, and 2 investment advisor representatives on its Board.
32 This may seem to be a relatively harsh indictment, but it is grounded in fact. The results of a 1998 meeting called by the SEC at the request of the vendors to try to force the SROs to reduce options quotation activity (prior to the recent sharp increase caused by multiply listing options that at the time were solely-traded) is illustrative. The meeting was chaired by Richard Lindsay, then Director of Market Regulation at the SEC. Under questioning, the vendors admitted that rather than make the necessary investments that would be very costly (thereby hurting their returns) that would allow them to carry all the options quotation traffic, they had begun arbitrarily to not publish quotations. That is not the sort of policy move that will provide the best quality market information to the public, and it should not earn the vendors the right to sit as voting members on the Operating Committees of the Plans.
33 The Island ECN, Nextrade, and Archipeligo have all expressed interest in one form or another in becoming SROs, according to letters from Allan A. Bretzer, Chairman of ITS (dated January 10, 2000) and Thomas E. Haley, Chairman of CTA (data January 26, 2000) to Annette L. Nazareth, Director, Securities & Exchange Commission; Archipeligo may have resolved the issue with its PCX merger.
34 The NASD is in the process of a vote, due in April, 2000, to spin off NASDAQ as a for-profit entity; the NYSE has announced its intentions of going public, which would necessitate a for-profit structure.
35 According to the SIA Security Industry Fact Book , the pretax ROE earned by the broker-dealers that report figures to it was 26% and 24% in the 1995-1997 and 1995-1998 periods, respectively
36 See Appendix AA, Part II for a way to accomplish this.
37 See footnote 14.
38 See footnotes 15 and 16.
39 See footnotes 20-22
40 See footnote 23
41 Forrester Research and Gomez Advisors estimates, respectively as contained in a report on the online trading industry on the SIA website in January, 2000.
42 This lower estimate was provided from discussions with salespeople and traders at major brokerage firms, and by observing behavior of traders on the floor of the AMEX.
43 $18.50/68 - $50.00/68
44 $18.50/95 - $50.00/95
45 $13.60/68 - $30.20/68
46 $13.60/95 - $30.20/95
47 $10.50/68 - $27.00/68
48 $10.50/95 - $27.00/95
49 $2.00/7.7 - $4.00/7.7
50 $2.00/11.3 - $4.00/11.3
51 February 4, 2000 Hambrecht & Quist report of the Charles Schwab acquisition of CyBerCorp.
52 Wall Street Journal article, Section C page 1, February 29, 2000.