U.S. Securities & Exchange Commission
SEC Seal
Home | Previous Page
U.S. Securities and Exchange Commission

Federal Advisory Committee on Market Information:
New Stock Exchange
Recommendation for an Alternative Model

Robert G. Britz
New York Stock Exchange, Inc.
Group Executive Vice President
11 Wall Street
New York, NY 10005
tel: 212.656.6057
fax: 212.656.5547
rbritz@nyse.com

December 1, 2000

Mr. Joel Seligman
Dean and Ethan A. H. Shipley University Professor
Washington University School of Law
1 Brookings Drive
Campus Box 1120
St. Louis, MO 63130

Re: Competing Consolidators Model

Dear Dean Seligman:

By your October 26 email to the members of the SEC Advisory Committee on Market Information, you requested that members who have ideas for alternatives to the current "consortium consolidator" model describe their ideas in writing. By this letter, the New York Stock Exchange responds to your request. We describe below how a model involving competition among multiple, non-exclusive consolidators (the "competing consolidators" model) might work.

First, we explain why we believe the competing consolidators model is a better way to address the problems with the consortia than the proposals in the SEC's concept release on market data.1 We then provide an overview of how the competing consolidators model might work. In the balance of this letter, we respond to the specific questions (somewhat re-ordered for the sake of logical progression) that you posed in your October 26 email.

INTRODUCTION

As we described at length in our submission responding to the SEC's concept release on market data,2 over the last 25 years, CTA/CQ has established a solid track record of making a high-quality data service ubiquitously, inexpensively and reliably available to securities professionals and consumers. As we also pointed out, we believe that, because proposed CTA/CQ fees are vetted with the industry, approved by the constituent governance processes of the nine exchanges3, and subjected to SEC review and public comment, the market data revenue that the NYSE receives reflects a fair allocation of the NYSE's costs. In short, it is not at our initiative that we propose to end our joint effort with the other exchanges, any more than it was at our initiative that we began that joint effort in the first place.

But the concept release's proposals to add additional layers of bureaucracy and to embrace a utility ratemaking model, as well as the growing dysfunction of CTA/CQ in the face of evolving market structure, led our Board last April to determine that we should withdraw from the CTA/CQ Plans. We are confirmed in our course by the subsequent crystallization of the legal perils of joint activity even when taken within the framework of an SEC-approved plan,4 and by the contradiction between continued joint dissemination and the prescription by the Antitrust Division and the SEC that each exchange must acquire and manage its own data transmission capacity independently.5

Our decision recognizes that technology and market forces provide a workable way for the NYSE and the rest of the industry to assure the display of consolidated data after the NYSE's exit. It also recognizes that we have the right under the CTA/CQ Plans to withdraw so long as we demonstrate to the satisfaction of the SEC that we will continue to meet the requirements of the applicable SEC rules.

But, to be clear, the NYSE is not calling for the abolition of CTA/CQ. Nor do we believe our withdrawal will trigger any crisis in data consolidation and dissemination. Similarly, we are not trying to promote a new market data distribution model for the industry, which would be both presumptuous and well beyond our ambition to exit CTA/CQ.

Nevertheless, we recognize that, by definition, our withdrawal will create the potential for multiple consolidators for the securities we list, and will somewhat alter the market data landscape. We also recognize that our action will create a template that other exchanges may wish to use or extend. That could change the market data landscape more significantly.

While the NYSE is not advocating any particular model for the industry, we have thought through what the market data industry structure might look like following our withdrawal from CTA/CQ. In response to your request, we are pleased to share our thinking with the Committee.

OVERVIEW

Current Model (Consortium Consolidator)

Under the CTA/CQ Plans, the nine exchanges (1) establish and collect common charges, (2) jointly enter into and administer contracts regulating the terms of receipt of last sale prices and quotation information and (3) oversee a common processing and data distribution facility. The nine exchanges select an exclusive processor to operate their common data facility.6 For the purposes of this letter, the consortium processor's functions may usefully be delineated as three: (1) collecting each of the nine exchanges' data output (the "single-exchange data feed function"), (2) consolidating the nine exchanges' data (the "consortium consolidation function") and (3) providing a consolidated data feed to anyone who wishes to avail itself of it (the "consolidated data feed function"). Diagram 1 shows the current data distribution arrangement.

As is depicted in Diagram 1, the CTA/CQ consortium processor provides a consolidated feed directly to 62 entities that can be divided into two categories: (1) those that provide data feed services to others (depicted as "Data Feed Providers" in Diagram 1) and (2) those that solely service end-users or their own internal use (depicted as "Vendors" and "Information Users" in the diagram). 7 The 62 direct recipients comprise eight exchanges, three press organizations, 18 vendors and 33 broker-dealers. Most of the 18 vendors (Bridge, ILX, Reuters, S&P Comstock, etc.), as well as a few broker-dealers, offer data feed services. These services typically supplement the consortium feed with consolidated options and OTC stock data, as well as data from domestic futures exchanges and from non-U.S. exchanges.

In addition, approximately 1400 vendors and information users receive consolidated data indirectly, through one or more of the data feed providers. For the most part, these indirect recipients are broker-dealers, institutions and others that redistribute data internally to their employees and, more recently in the case of broker-dealers, to their retail customers.

Competing Consolidators Model

The competing consolidators model unbundles several functions currently performed at the consortium level. Upon the NYSE's withdrawal from the CTA/CQ Plans, the NYSE would (1) separately establish and collect its own fees, (2) separately enter into and administer its own contracts regulating the terms of receipt of its last sale prices and quotation information and (3) provide its own data distribution facility. The NYSE will use the services of SIAC (its data processing subsidiary) to operate this NYSE-only facility. However, its facility would only perform the single-exchange data feed function. Diagram 2 illustrates this revised arrangement.

As depicted in Diagram 2, to disseminate consolidated data, direct data feed recipients would have to:

  • Take two data feeds - the consortium feed (which no longer includes NYSE data) and (2) the NYSE-only feed, and

  • Perform the consolidation function and the consolidated data feed function.

Any data feed provider (including SIAC8) or other data processing firm could enter this non-exclusive consolidation business. (Such data feed providers and firms are separately delineated as "Competing Consolidators" in Diagram 2.)

Any data feed provider that does not enter the consolidation business, as well as any of the other 62 direct recipients that choose not to enter the consolidation business, would have to contract to receive consolidated data from a competing consolidator.

Similarly, any of the 1400 indirect data feed recipients that does not wish to avail itself of the consolidation services provided by the competing consolidators could choose to self-consolidate (depicted in Diagram 2 by the arrows from CTA/CQ and from the NYSE that bypass the competing consolidators box). The rest of the indirect recipients would continue to receive consolidated data feeds. Professional and nonprofessional subscribers (depicted in the diagrams as "Information Users") would also continue to receive consolidated data services.

Appendix A traces the flow of data from the input of an order into the NYSE systems through display under the competing consolidators model.

Extended Competing Consolidators Model

The model also works if other exchanges -- or all exchanges -- withdraw from the CTA/CQ Plans. We anticipate that it would also work if they withdraw from the OTC/UTP Plan and the OPRA Plan as well (in which case consortium consolidation would disappear altogether). Diagram 3 illustrates the model as extended to all CTA/CQ exchanges. Note that the schematic is actually simpler than that of Diagram 2, since the consortium consolidator box (CTA/CQ) disappears altogether.

Comparative System Costs

Most of the costs of market data are internal to the nine exchanges, taking the form of developing, deploying and operating the technical infrastructure to support the production and dissemination of the data. The competing consolidators model implies no changes to these costs. However, the NYSE's withdrawal would affect the portion of the dissemination costs that are currently pooled.9

The nine CTA/CQ exchanges currently share the approximately $6.6 million in costs annually incurred by the consortium consolidator in providing the single-exchange data feed functions, the consolidation function and the consolidated data feed function. While these costs are tiny when compared to the overall costs of the market data industry, we believe these costs could well decrease, at least under the extended competing consolidators model.

Under the extended model, each exchange would incur the cost of its single-exchange data feed function individually, and competing consolidators would take on the costs of the consolidation function and the consolidated data feed function.

  • Single-Exchange Data Feed Costs - Predicting the outcome as to the costs of the single-exchange data feed functions under the extended model is highly speculative since each exchange would, for the first time, independently enjoy a variety of insourcing and outsourcing alternatives that would tend to drive costs down. However, one can bound the costs at the upside by observing that, if the NASDAQ Exchange simply used its existing data processing facilities and the other exchanges simply continued to use SIAC for this purpose, these costs would change little.

  • Consolidation and Consolidated Data Feed Costs - As noted, the competing consolidators model will completely subject these aspects of the dissemination system to market forces. Moreover, among most likely entrants into the competing consolidators business are the existing data feed providers, which already process the consolidated feed (i.e., they would require little incremental capacity); already handle scores of data feeds from the consortium consolidators, the several domestic futures exchanges and exchanges around the world; and already provide data feed services to 1400 vendors, broker-dealers and others. Thus, it is unclear whether they would have material incremental costs in taking on the consolidation function. As a result, the extended model might take the portion of the $6.6 million attributable to consortium consolidation and consortium consolidated data feeds completely out of the overall system, rather than merely shifting them from the consortia to the consolidators.

RESPONSES TO QUESTIONS

Who will act as competing consolidators?

Under the competing consolidators model, neither SIAC nor any other entity would serve as the exclusive source of consolidated trades and quotes. Instead, any entity could pick up the consortium and NYSE data streams, consolidate the streams and make the resulting consolidated stream available to vendors, broker-dealers and investors.

The broad universe of possible non-exclusive consolidators includes current industry participants (i.e., SIAC and the data processing subsidiaries of other exchanges, as well existing data feed providers). The non-exclusive, competing consolidators would compete to provide consolidation services, just as they compete today to provide data feeds.10

How will competing consolidators obtain market information?

Today, SIAC and the nine exchanges install lines by which each exchange sends its trades and quotes to SIAC for validation, consolidation and dissemination. Pursuant to the CTA/CQ Plans and bilateral contracts, SIAC acts as each exchange's agent in this activity. It neither enters into contracts with, nor collects fees on, the vendors, broker-dealers and others that gain access to the consolidated data from SIAC. Instead, the exchanges compensate SIAC for its work from the pool of revenues that they jointly collect from data users.

Under the competing consolidators model, CTA/CQ would provide the consortium feed (no longer including NYSE data), and the NYSE would provide the NYSE-only feed, to the competing consolidators and to vendors, broker-dealers and others that self-consolidate. As CTA/CQ does today, the NYSE would use IP multicasting protocols (or comparable distribution mechanisms) and require consolidators to receive the data over at least a T1 communications line (thereby minimizing any latency delays that might result from lower line speeds).11

How will competing consolidators make market information available to users?

Competing consolidators would provide consolidated data to the approximately 1400 broker-dealers, institutions and other indirect data feed recipients in the same way that that they provide data feed services today. In turn, the indirect data feed recipients would disseminate consolidated data to end-user devices in the same ways as they do today.

How will brokers satisfy their best execution responsibilities under the competing consolidators model?

As noted, the competing consolidators model would make no change in the ability of end-user devices to continue to receive consolidated data. Thus, the model implies no change in the manner in which brokers satisfy their best execution responsibilities and, in particular, in the availability of the NBBO and the other consolidated data necessary to comply with the recently-adopted rules on disclosure of execution quality.12

How will terms, including fees, for information be determined?

Each competing consolidator would administer its own business. It would assume responsibility for assuring that it has sufficient capacity to receive and process all nine exchanges' data streams. It would also establish its own charges for its creation and provision of a consolidated feed under competitive conditions, just as data feed providers charge for the services they offer today. Thus, we answer this question from the perspective of terms that the NYSE is likely to establish after withdrawal.

Terms other than fees - Today, the market data consortia disseminate data under three sets of contracts.13 Under the competing consolidators model, both CTA/CQ and the NYSE could continue to establish terms and conditions through contracts.14 They simply would not do so jointly.

Thus, CTA/CQ and the NYSE could separately enter into a contract with each consolidator, other data feed provider, vendor and professional subscriber.15 Each contract would establish terms and conditions that are fair and reasonable and not unreasonably discriminatory.16 Contracts with competing consolidators and self-consolidating vendors might include technical specifications.

Fees - As noted above, today, the nine exchanges establish fees jointly though a multi-layered process that involves vetting fee proposals with customers and industry groups, approval by each exchange's constituent finance committee or board, approval by CTA/CQ, and SEC review (including an opportunity for public comments). Under the competing consolidators model, except for the consortium-approval step, the process would be the same for the NYSE. Thus, the NYSE would establish its fees for NYSE-only data in the same way as it does today as to fees for listing, trading, regulatory and other services. Each fee would be subject to SEC review under the fair-and-reasonable and not-unreasonably-discriminatory standards noted above and the fee-filing requirements of Rule 11Aa3-2. The NYSE would bill its customers and collect amounts due.

Should market information be made available in a standardized format? If so, what should that format be? How will investors be assured of receiving accurate, real-time consolidated information under the competing consolidators model?

While today's technology permits multiple entities to receive simultaneously multiple streams of data and to create consolidated outputs that sequence prices and quotes in the same order, a multi-consolidator environment will introduce at the consolidation level four types of risk that are present today among data feed providers, vendors and other recipients.

  • Different Consolidator Operating Environments - Hardware, software and line differences could cause differences in the ways consolidators process the CTA/CQ and NYSE outputs, which could lead to sequencing differences.

  • Validation - Different validation tolerances would result in sequencing differences.

  • Capacity Considerations - A particular consolidator might have insufficient capacity.

  • Protocols/Data Formats - Consolidators would have to manage any differences between the CTA/CQ and NYSE in line protocols and data formats.

Today, the CTA/CQ Plans resolve many of these issues, and it would seem likely that those resolutions would carry forward without any intervention.17 In addition, because the four types of risk are present today among data feed providers, vendors and other recipients, the industry is experienced in managing those risks. Moreover, continuing differences among competing consolidators in these risk areas would result in differences in service levels; one would expect these differences to self-correct under the discipline of market forces. However, to the extent that intervention is necessary,18 it could be effected through standards-setting under the auspices of a trade group, through contractual undertakings and, if necessary, through direct SEC regulation of non-exclusive securities information processors ("SIPs").19

Why is the competing consolidators model preferable to the existing consortium model?

As noted above, our decision to withdraw from the CTA/CQ Plans is not based on any view that CTA/CQ has fallen short in delivering a high-quality dissemination service ubiquitously, inexpensively and reliably. Rather, it is based on the view that the "cure" to the consortium model that the SEC concept release and the options antitrust settlements prescribe is worse than the "illness" its critics perceive. Thus, we answer this question from that perspective.

Costs - The costs implied by the competitive model appear relatively small. As noted above, the joint costs of the collection, consolidation and dissemination functions are modest relative to the overall costs of producing and disseminating market data, and it appears that the unbundling of consortium costs may not raise them at all. Indirect costs could arise from failing to effectively manage the loss of today's "one-stop shopping" at a single consolidator that has a track record of high quality and cost-effective service and that acts as a neutral leader in the arena of setting standards and coordinating industry activity. But, as noted above, market forces and, if necessary, intervention in the area of standards can minimize this risk.

Benefits - In contrast, the benefits appear substantial. As described above, the competing consolidators model provides a data dissemination process that, operationally, is at least as good as that which exists today. In one stroke, the model ends joint pricing, and thereby the inter-exchange subsidies, cross-consortia fee distortions and other market distortions that currently exist.20 Extended, it solves the problem in the OTC/UTP consortium resulting from the undifferentiated internal/inter-market systems of the NASDAQ Exchange21 and the operational dysfunction now plaguing all the consortia.22 The extended model also achieves the capacity independence among the exchanges sought by the Antitrust Division and the SEC.

Affirmative benefits include that the NYSE would compete with the other eight exchanges in the market data arena, just as its competes today with the other eight exchanges individually in attracting listings and liquidity. Similarly, data feed providers would compete as to consolidation services, just as they do today as to other services. The general benefits from competition in any industry are well understood, and need not be repeated here.23

* * *

I hope our description of how a model involving competition among multiple, non-exclusive consolidators might work, as well as our identification of some of its costs and benefits, proves useful. I look forward to discussing our proposal with you and the other Committee members at our meeting on December 14.

Sincerely yours,

/s/ Robert G. Britz

Robert G. Britz

cc: Annette Nazareth
Members of the SEC Advisory Committee on Market Information

Appendix A

This Appendix A contains two flow charts. The first (Basic Systems Flow - Current) shows the current routing and processing steps of an order sent by a member firm to NYSE from the moment NYSE receives it to its distribution to data recipients and the attendant flow of market information that emanates from such orders. The second chart (Basic System Flow - Competing Consolidator) illustrates the modest changes that would occur in the routing of market information under the competing consolidator model.

Basic System Flow - Current (See chart below)

Beginning at the upper left-hand corner, member firms send orders to NYSE via the CAP network into CMS. CMS provides reliable communications between member firms and NYSE host systems operated by SIAC. CMS supports over 450 circuits, handling a variety of communications speeds and protocols. CMS receives, edits and validates orders and formats associated reports back to the member firms. CMS also handles queuing, flow control and priority routing. From CMS, the order flows to SuperDOT, where it receives a reference number and a "safe store" record is created and placed in the database of record. Based on member firm parameters, SuperDOT then routes the order via PSS to its appropriate destination on the trading floor (which could be the specialist's display book ("DB"), a member's booth, a handheld device or the like).

Market orders received by DB are generally executed within a matter of seconds. When a trade execution occurs, DB sends a report of it back to the member firm and reports the details of the trade to MDS via PSS along with an update of the quotation. Limit orders entered into DB have the potential to generate a new quotation, which DB would also generate and send to MDS via PSS. MDS validates the trade and/or quote, captures it in a database and passes it on to CTS or CQS, as appropriate.

CTS and CQS receive trades and quotes from NYSE and from each of the eight other U.S. exchanges. These trades and quotes are validated, captured in databases, consolidated into two separate data streams (i.e., one for trade reports and one for quotes), and disseminated to data recipients via high speed networks.

IGS is a NYSE system that receives the CTS and CQS consolidated data streams just like any other data recipient. It then provides the data streams to NYSE applications such as crowd displays and booth support.

Basic System Flow - Competing Consolidator Model (See chart below)

The left side of this chart is the same as previous chart. The channels through which member firms send orders to NYSE, and the several processing steps that are performed by NYSE host systems in handling those orders, remain unchanged in the competing consolidator model. Similarly, the trade report and quote generation process also remains the same.

Under the competing consolidator model, NYSE would no longer make its trade and quote information available solely to one exclusive consolidator, but to as many self- consolidating data recipients and/or competing consolidators as might wish to receive it. Under the current model, data recipients have no choice but to receive a consolidated stream of data. Under the competing consolidator model they could choose to receive individual streams of data from one or more exchanges and to consolidate those data streams themselves or they could choose to receive a consolidated stream of data from any competing consolidator.

NYSE anticipates that it would continue to support SIAC providing a consolidation facility, but not an exclusive processor facility. NYSE, as well as the other exchanges, will have a continuing need to receive and display data from all markets in order to fulfill their regulatory obligations. The chart below illustrates the making available of both unconsolidated and consolidated streams of data.

The box on the chart entitled "Exchange Data Distribution" depicts the output of unconsolidated streams of data received from any other exchange that wishes to distribute its data on an unconsolidated basis. It also reflects the fact that, even if an exchange chooses to utilize another facility to distribute its data, NYSE would, of necessity, make appropriate arrangements with all exchanges to receive their data and to consolidate such data for its own purposes.

The box on the chart entitled "Consolidated Data Distribution" reflects SIAC's continued performance of the consolidation function on a nonexclusive basis for NYSE and for the benefit of any other data recipient that might wish to receive a consolidated feed from SIAC.

The functions depicted in both of these boxes could potentially be duplicated by a number of other organizations. Each would receive trades and quotes from NYSE and from each of the other exchanges, and each would provide to data recipients (1) separate feeds consisting of each exchange's data (e.g., a NYSE-only feed), (2) a consolidated stream of last sale prices and/or (3) a consolidated stream of quotes.

Footnotes

1 Release No. 34-42208 (December 9, 1999).
2 See "Comments to Securities and Exchange Commission Concept Release on `Regulation of Market Information Fees and Revenues'" (April 10, 2000; the "NYSE Concept Release Response") at pp. 11-13 and Appendix D.
3 For the sake of simplicity, in view of NASDAQ's anticipated registration as a national securities exchange, this letter generally does not address the continuing role of the NASD in reporting internalized executions and other off-board trades.
4 Recently-settled proceedings of the Antitrust Division of the Department of Justice and the SEC alleged that a decision taken by the Amex, CBOE, PCX and PHLX to limit the capacity of OPRA effectively constrained multiple trading in certain options classes. United States v. American Stock Exchange et al., No. 1:00CV02174 (D.D.C. filed Sept. 11, 2000), Complaint at 10; In the Matter of Certain Activities of Options Exchanges, AP File No. 3-10282, Exchange Act Release No. 43268 (Sept. 11, 2000), Order Instituting Public Administrative Proceedings Pursuant to Section 19(h)(1) of the Securities Exchange Act of 1934, Making Findings and Imposing Remedial Sanctions at 4, 6.
5 The settlement in the options antitrust matter imposed the prescription on the Amex, CBOE, PCX and PHLX. See United States v. American Stock Exchange et al., Competitive Impact Statement of United States at 21. At the May 10, 2000 CTA meeting, the SEC staff requested that the constituent exchanges file amendments to the CTA and CQ Plans that would meet this independent capacity standard as well.
6Since the CTA/CQ Plans' inception, the constituent exchanges have contracted with the Securities Industry Automation Corporation ("SIAC") to act as the consortium processor.
7 Note that the diagram simplifies the market data industry by depicting as discrete several functions that are often combined within the same entity (e.g., a data feed provider typically also provides interrogation services to its own employees or customers). This simplification is not material to the changes suggested by the competing consolidators model.
8 Outside of its consortium role, SIAC becomes just one more data feed provider serving the information industry.
9 Note that we are focused in this section solely on a steady-state analysis of the system changes necessitated by the NYSE's withdrawal, asking whether proliferating the number of connections with exchanges and the number of consolidators implies the creation of repetitive facilities and, therefore, the incurring of increased costs. We address fees separately below.
10 While we have not generally assessed the interest of data feed providers in entering the competing consolidator business, SIAC has indicated its strong interest in doing do.
11 Under the "concurrent use" provisions of the CTA/CQ Plans, SIAC already performs single-exchange data feed functions on behalf of the several exchanges that make available through SIAC data that is not consolidated, e.g., index values and trades and quotes in local securities. In performing these single-exchange data feed functions, SIAC does not act for all nine exchanges collectively, but solely for the exchange whose data it processes. SIAC is compensated for providing those bilateral services under the general arrangements of the CTA/CQ Plans. (As noted above, SIAC will perform the single-exchange data feed function on behalf of the NYSE.)

Any other exchange withdrawing from any of the consortia would have to either create its own capability to make its data available to consolidators or enter into bilateral arrangements with a data processing firm to provide a feed of its data to consolidators. While we have not generally assessed the interest of data processing firms in performing this function, SIAC has indicated its strong interest in doing so.

12 Rules 11Ac1-5 and 11Ac1-6. The question of whether to continue to require vendors and broker-dealers to provide consolidated data displays to all customers (rather than requiring them only to offer consolidated displays) deserves careful consideration by the Committee. However, the competing consolidators model works with either approach.
13 The three sets consist of one set under the CTA and CQ Plans, one set under the OPRA Plan and one set under the OTC/UTP Plan. Each set includes forms of agreement between each market and the consolidator, between the exchanges and vendors, and between the exchanges and professional subscribers. It also includes a form of nonprofessional subscriber agreement which, under the CTA/CQ Plans, runs between the subscriber and the vendor acting on behalf of the nine exchanges. The NYSE expects to continue this approach as to NYSE-only data.
14 Under the competing consolidators model, the NYSE would have to receive SEC approval of its own transaction reporting/national market system plan. (See Rules 11Aa3-1 and 11Aa3-2.)
15 Under the extended competing consolidators model, the resulting change from joint contracts to individual exchange contracts could theoretically result in an increase from today's three sets to nine (11, taking into account the NASD and the International Securities Exchange, which is the sole options exchange that is not also a CTA/CQ Plan participant). However, if the industry's experience regarding the high level of conformity among the exchanges' rule books is predictive, those three sets would serve as models for all the exchanges, and the regional exchanges would tend to follow subsequent modifications to the three sets that the primary exchanges make.
16 See 1934 Act sections 11A(c)(1)(C) and (D).
17 If the NYSE were the only exchange to withdraw from the CTA/CQ Plan, uniform standards would almost certainly prevail so long as SIAC remained the consortium processor.
18 For example, data feed providers currently manage protocol and data format differences in receiving feeds from the several data consortia, from the several domestic futures exchanges and from scores of exchanges around the world. On the other hand, we think that the issue of validation differences would almost certainly have to be addressed through uniform standards.
19 The SEC has the authority to regulate non-exclusive SIPs under 1934 Act section 11A(c). It uses this authority to require consolidated displays. (See Rule 11Ac1-2.)
20 SEC Market Data Concept Release, p. 34; NYSE Concept Release Response, pp. 3, 20-21 & 37. In brief, the joint activity under the plans distorts pricing and market efficiency because it creates (1) a revenue sharing scheme that ignores the value of each market's data (i.e., it creates inter-exchange subsidies) (see pp. 3, 20 and 21 of the response); (2) disincentives to quote competition (see pp. 20-21 of the response); and (3) cross-plan fee distortions (i.e., Network A receives 23 cents per message; Network B, $1.69) (see p. 37 of the response).
21 See generally Letter dated October 17, 2000, from Robert G. Britz, Group Executive Vice President, New York Stock Exchange, to Richard G. Ketchum, President and Chief Operating Officer, NASDAQ Stock Market (the "October Letter to NASDAQ"). We have attached a copy of the October letter, which included the NYSE Concept Release Response as an exhibit, as Attachment A to this letter.
22 October Letter to NASDAQ, p. 2 & n. 6; NYSE Concept Release Response, p. 21 & n. 37.
23 Indeed, it seems to us that the consortium model can no longer be sustained as a regulatory matter. A core principle of the 1934 Act is that exchanges may not act in a way that imposes any burden on competition not necessary or appropriate in furtherance of the purposes of the 1934 Act. (See, e.g., 1934 Act section 6(b)(8).) In the face of the workable, pro-competitive alternative that today's technology makes possible, it is not clear to us how those who might espouse continuing today's consortium model could demonstrate that the burdens on competition imposed by the consortium model continue to be "necessary or appropriate". Thus, when subjected to the criteria of the 1934 Act, the question that you pose must be reformulated as, "What benefits of the consortium model over the competing consolidators model justify the burdens on competition created by the consortium model?"

Attachment 1

Robert G. Britz New York Stock Exchange, Inc.
Group Executive Vice President 11 Wall Street
New York, NY 10005
tel: 212.656.6057
fax: 212.656.5547
rbritz@nyse.com

October 17, 2000

Richard G. Ketchum
President and Chief Operating Officer
The NASDAQ Stock Market
1735 K Street N.W.
Washington, D.C. 20006

Re: Request to Exclude Stocks Listed on the NASDAQ Exchange from Systemically-
Independent and Neutral NMS Facilities

Dear Rick:

In seeking to join the CTA, CQ and Intermarket Trading System ("ITS") Plans, the NASDAQ Exchange1 proposes to amend the CTA Plan to exclude stocks listed on the NASDAQ Exchange from these three plans.2 If your amendment were adopted, inter-exchange3 competition in the over-the-counter ("OTC") market4 pursuant to unlisted trading privileges ("UTP") would be relegated to facilities undifferentiated from the intra-market facilities of the NASDAQ Exchange.

Because of the resulting anti-competitive effects, we cannot agree to your amendment. However, we believe that much of what you seek can be achieved if each exchange is permitted to report its trades and quotes pursuant to whichever joint data plan it chooses. Therefore, until such time as the data plans and the ITS linkage5 give way to market-driven substitutes, we agree that the CTA and CQ Plans should effect this choice so long as ITS is preserved as the vehicle for inter-exchange linkage.

CONSOLIDATION TECHNOLOGY AND INTER-EXCHANGE COMPETITION

As you well know, a generation ago -- under very different technological and competitive circumstances -- the SEC mandated exchanges to create a joint venture for the dissemination of trade reports. As you also know, in April, our Board authorized the NYSE to seek regulatory approval to withdraw from the CTA and CQ Plans. Our Board believes that dissemination through these legacy consortia imposes unwarranted burdens on inter-exchange competition and that market forces provide a better solution.6 It also believes that the Federal government should not compel exchanges - vigorous competitors in all arenas except market data -- to jointly arrange how they distribute and charge for their "output" of market data.7

The competitive bankruptcy of the very concept of the data consortia is exacerbated by the way in which the OTC/UTP Plan structures inter-exchange competition in the OTC market. As to stocks listed on the NASDAQ Exchange, your CTA Plan amendment would require other exchanges to report trades and quotes exclusively through systems of the NASDAQ Exchange that are undifferentiated from its internal systems. For the purposes of both data collection and inter-exchange linkage, these undifferentiated systems position other exchanges as if they were members of a competitor exchange -- the NASDAQ Exchange.

BARRIERS TO ENTRY: NASDAQ'S HOME-COURT ADVANTAGE

To assure that consolidated reporting in stocks listed on the NASDAQ Exchange occurs through the exclusive facilities of its information technology ("IT") subsidiary,8 you propose to amend the CTA Plan's definitions of "Eligible Security" and "Network B Eligible Security" to exclude your listed stocks.9 As you well know, those definitions were designed to include in the national market system all exchange-traded securities meeting the listing requirements of the NYSE or Amex. In turn, the exchanges and the NASD carefully negotiated the two plans -- as well as the ITS Plan -- to assure that they conferred no competitive advantage on any market.10

The OTC/UTP Plan affords no such "level playing field". By extending the internal systems of the NASDAQ Exchange to the regional exchanges in order to effect joint data dissemination and trading linkage, the plan melds into undifferentiated systems (1) its intra-market functions (i.e., the provision of facilities for the display, routing and execution of broker-dealer quotations and orders) and (2) its inter-market functions (i.e., the provision of facilities enabling each competing exchange to disseminate its internal quote and trade data and to protect its market from trade throughs, locked markets, etc., resulting from activity on another exchange). 11

Thus, a regional exchange accedes to "second class" citizenship by the very way in which it reports trades and quotes under the OTC/UTP Plan: the trades and quotes emanating from its auction market are treated in the same way as the interest of a single OTC dealer. In the CTA/CQ Plan context, the analogy would be to include the internal BBO of the NASDAQ Exchange on the NYSE limit order book.

The failure of the OTC/UTP Plan (1) to conceptually sort out whether the NASDAQ Exchange is a competing exchange or an operator of a linkage among competing exchanges and (2) to provide discrete systems for each of those two roles has the same anti-competitive consequences for the OTC trading linkage. It provides neither a neutral inter-market linkage like ITS12 nor rules governing trade throughs, locked markets or block trades. Nor does it provide for participation in openings at prices that are away from the previous day's close.13 Your proposal would codify these anti-competitive effects and prolong the void in inter-exchange price protection by precluding inter-exchange competition in the OTC market from occurring within the framework of ITS.14

Said another way, the OTC/UTP Plan and OTC linkage preclude another exchange from contesting the primacy of the NASDAQ Exchange through an auction/agency market model engaged in independent price discovery.15

SYSTEMS CAPACITY AS A BARRIER TO COMPETITION

Year in and year out, NYSE's internal market systems and the national market facilities managed by SIAC perform at "five nines": they are "up" 99.999 percent of the time. Our capacity planning, development and implementation has enabled us to keep pace with the explosive growth in volume during the 1990s and to commence trading in decimals last August.

In contrast, the undifferentiated internal/inter-market systems of the NASDAQ Exchange have been plagued by frequent queues in quote and trade reporting, other message queuing,16 and outright system failures and outages. Inadequate capacity also prevents the OTC market from trading in decimals. Thus, the effect of your amendment is to subject other exchanges' trading of OTC stocks pursuant to UTP to the "single point of failure" of your undifferentiated systems. Given those systems' track record, this is much more than just a theoretical or prudential concern, especially in light of the allegations by the Antitrust Division and the SEC that agreement by the settling options exchanges to limit OPRA capacity violated the antitrust laws (as well as SEC Rule 19c-5).17

A FIRST STEP TO EXITING THE PLANS: PLAN CHOICE

For all the reasons cited above, we cannot agree to amend the CTA Plan as you propose. At the same time, we believe that the NASDAQ Exchange should have the right to choose to continue to report trades and quotes in OTC stocks -- and to begin to report trades and quotes in Amex- and NYSE-listed stocks should it so desire -- pursuant to the plan that it administers and whose capacity it determines. Accordingly, we wish to put forth for your consideration an alternative approach that would govern until we receive regulatory approval to withdraw from the NMS plans.18

We propose to preserve the CTA Plan's definition of "Eligible Security". Thus, until such time as ITS gives way to market-driven substitutes, inter-exchange linkage would continue to occur through ITS.

However, as to any stock, any exchange could "opt out" of the trade and quote reporting provisions of the CTA and CQ Plans (but, as noted, not the linkage, price-protection and pre-opening provisions of the ITS Plan). The exchange would have to (1) report the stock's trades and quotes pursuant to another joint plan19 and (2) demonstrate that non-exclusive processor(s) will make available a consolidated stream of trades and quotes in the stock.20

* * *

We believe our interim approach gives the NASDAQ Exchange autonomy and choice in its quote and trade reporting while reducing the anti-competitive effects inherent in the joint data plans, avoiding expansion of the operational dysfunction now plaguing the plans, and containing the anti-competitive effects peculiar to the OTC/UTP Plan. It also assures that inter-exchange competition in stocks listed on the NASDAQ Exchange occurs within the framework of a neutral linkage that affords inter-market price protection.

Our representatives will be prepared to discuss the details at the October 18 CTA/CQOC meeting. Should you determine that amendments along the lines that this letter describes are worth pursuing, we are prepared to work through the details with you and the other exchanges over the next several weeks. This would enable our Board to consider the necessary plan amendments at its next meeting (December 7), thereby assuring that these issues do not interfere with the timely completion of the NASDAQ Exchange's reorganization and registration as a national securities exchange.

Sincerely yours,

/s/ Robert G. Britz

Robert G. Britz
Group Executive Vice President

C: CTA/CQ Participants
Robert Colby
Annette Nazareth
Joel Seligman

Footnotes

1 This letter refers to the NASDAQ Stock Market as the "NASDAQ Exchange", which assumes that NASDAQ will shortly complete its announced reduction of its ownership by the National Association of Securities Dealers, Inc. (the "NASD") to 21 percent and will be registered as a national securities exchange.
2 See July 25 Memorandum to CTA/CQ Participants from Robert E. Aber, Senior Vice President and General Counsel, NASDAQ Stock Market, Inc. (the "Aber Memo").
3 The focus of this letter is inter-exchange competition, which recognizes that the NASDAQ Exchange expects to be registered as a national securities exchange. However, although this letter generally does not speak to the continuing role of the NASD in the CTA, CQ and ITS Plans, we recognize that internalized executions and other off-board trades will continue to be reported through the NASD and, therefore, that the NASD will continue as a full participant in these NMS facilities.
4 This letter refers to the market for stocks listed on the NASDAQ Exchange as the "OTC market".
5 The NASDAQ Exchange indicated at the August 3 meeting of the ITS Operating Committee that it would be sending a letter seeking to become a Participant in the ITS Plan. We have not yet received that letter.
6 We discuss these reasons at length in our response to the Commission's 1999 concept release on market data. (April 10, 1999, Letter to Jonathan G. Katz, Secretary, SEC, from James E. Buck, Senior Vice President and Secretary, NYSE.) A copy of our April 1999 letter is enclosed for your convenience. The entirety of our extensive submission is available at http://www.sec.gov/rules/concept/s72899.shtml. I will briefly summarize a few salient points of our Board's rationale here.

After noting the general benefits of extending inter-exchange competition to the market data arena, our April 1999 letter listed a series of anti-competitive effects and market distortions arising from the revenue-sharing provisions of the CTA and CQ Plans. Our list included (1) sharing in revenues without regard to either the costs of an exchange's systems development and operations or the value of its data (i.e., inter-exchange subsidies) (p. 3 & n. 1; p. 20 & n. 35; p. 21); (2) disincentives to quote competition (p. 20 & n. 36); and (3) cross-plan fee distortions (i.e., Network A receives 23 cents per message; Network B, $1.69) (n. 37).

We also listed eight controversies and other phenomena that contribute to the growing dysfunction of the market data consortia. (P. 21 & n. 37.) We noted disputes about the eligibility of securities for revenue sharing, about the fee liability for devices on exchanges with virtual floors, and about the non-consolidated dissemination by an options exchange of its information to its virtual floor. We mentioned the meeting gridlock resulting from the interjection of antitrust lawyers. We also observed how revenue sharing is motivating ECNs to register as national securities exchanges, which would increase the SEC's direct oversight burden and, therefore, would contribute to the call for separating NASDR from NASDAQ. And we pointed out the concerns of the Antitrust Division of the Department of Justice as to the payment of an entry fee by an exchange that joins one of the data consortia.

Our list continues to grow. New additions include (1) the further crumbling of "one-exchange, one vote" by the addition of the NASDAQ Exchange to the existing affiliation between the NASD and the American Stock Exchange, Inc. (the "Amex") (the affiliation between the Chicago Board Options Exchange, Inc. and the Cincinnati Stock Exchange, Inc. having provided the first breach); (2) a vendors' commencement of arbitration proceedings contesting the requirement that vendors must disseminate all options data, regardless of intrinsic value or customer choice; (3) the SEC's enforcement of its prohibition against unconsolidated dissemination of data against exchanges, but not against ECNs; and (4) the debate over whether to exclude an ECN's access fees from its quotes (reminiscent of the "net" versus "gross" debate over OTC quotes and prints that bedeviled CTA and the SEC for years during the 1970s and 1980s). Perennial issues that our April 1999 letter did not mention include (1) distortion of revenue sharing by the significant double counting in the OTC market due to the much higher levels of dealer interpositioning and (2) revenue sharing among exchanges based on prints rather than shares or trade value.

7 If, as we proposed in April, we can effect our withdrawal by the end of this year, many of the issues presented by your requested amendment would become moot.
8 By referring to the NASDAQ Exchange's "IT subsidiary", this letter means to describe whichever internal division or corporate affiliate of the NASDAQ Exchange emerges from its corporate reorganization to conduct its systems development and operations.
9 The Aber Memo also (1) proposed non-controversial, ministerial amendments designed to effect the participation of the NASDAQ Exchange and (2) sought to excuse the NASDAQ Exchange from paying the standard entry fee. From the discussion of the entry fee at the joint meeting of Consolidated Tape Association ("CTA") and the Consolidated Quotation Operating Committee ("CQOC") on August 16, we understand that the NASDAQ Exchange no longer seeks excuse from its payment obligation.
10 Both the CQ Plan and the ITS Plan incorporate the CTA Plan's definition of "Eligible Security". Note in this regard that the current trading by the regional exchanges of OTC stocks pursuant to UTP without reporting on Network B and without use of ITS violates the CTA, CQ and ITS Plans. Statutory, rule and contractual violations aside, the NYSE has never acquiesced in that result because of the "sphere of influence" implications, as well as the specific "unlevel playing field" reasons discussed in this letter.
11 In the listed market context, this would be the equivalent of the NYSE and its subsidiary, the Securities Industry Automation Corporation ("SIAC"), combining the NYSE's internal Market Data System (which collects trades and quotes from our order entry systems and the Floor) with the Consolidated Tape System and the Consolidated Quotation System (which SIAC operates as the facilities manager under the CTA and CQ Plans). On the linkage side, this would be the equivalent of the NYSE and SIAC combining SuperDOT®, Display BookTM, NYSE Direct+TM and Institutional ExpressTM (which route and execute NYSE orders) with ITS (which SIAC operates as the facilities manager under the ITS Plan).

Of course, the NASDAQ Exchange is entitled to exercise complete dominion over its intra-market functions. But a neutral facilities manager must conduct inter-market functions through a separate system. Indeed, use of a single, undifferentiated system to simultaneously effect both intra-market and inter-market functions would appear to violate the provisions of the Securities Acts Amendments of 1975. Congress specified that exclusive processors must be "neutral"; i.e., not dominated by a particular exchange and, therefore, not likely to engage in practices that placed other exchanges at a competitive disadvantage. Senate Committee on Banking, Housing & Urban affairs, Report to Accompany S.249: Securities Acts Amendment of 1975, S. Report No. 94-75, 94th Cong., 1st Session (1975) at 11-12. The situation is further exacerbated by the fact that the NASDAQ Exchange also uses the same system in yet a third role, that of a vendor of market data to end-users, creating the potential for competitive advantage vis-à-vis unaffiliated vendors.

12 The SuperSOES and SuperMontage proposals only exacerbate this problem by forcing regional exchanges to enter their customer orders into NASDAQ in order to receive meaningful price protection. This "Hobson's choice" means that the regional exchanges have to make their markets available for simultaneous execution in two places: within the SuperMontage and within their own markets. Moreover, the proposed algorithm for establishing execution priority among OTC market maker quotes, proprietary orders emanating from ECNs and proprietary orders emanating from regional exchanges discriminates against the last category of proprietary orders. This discrimination would appear to violate section 15A(b)(6) of the 1934 Act (section 6(b)(5) after the NASDAQ Exchange's registration as a national securities exchange). In its Market 2000 Report, the SEC urged the Amex to amend or withdraw a similar proposal that discriminated against OTC market makers and regional exchange specialists. (Study III, p. 11.)
13The absence of these types of NMS attributes from the OTC linkage has precluded the SEC from ordering permanent approval of the OTC/UTP Plan. (See n.15, infra.). In contrast, the listed market enjoys all these attributes. ITS provides inter-market access independent of any one market's internal facilities; protection against trade throughs, locked markets and block trades; and inter-market participation in openings that occur away from the previous day's closing price. (To be sure, as our public (non-industry) Directors recommended in their March report on market structure, with regulatory approval, the NYSE will lead the way in dismantling ITS and substituting market-driven alternatives to provide inter-market connectivity among broker-dealers. (Market Structure Report of the New York Stock Exchange Special Committee on Market Structure, Governance and Ownership (Match 23, 2000), pp. 41-44.))
14 The OTC/UTP Plan is also replete with structural, ministerial and technical provisions conferring other competitive advantages on the NASDAQ Exchange. Elements of the OTC/UTP Plan that compare unfavorably to their CTA/CQ counterparts include (1) a more limited role for the plan's operating committee (on which the regional exchanges are represented), and a concomitantly greater role for the NASDAQ Exchange and its IT subsidiary, in fee, contract and administrative matters; this greater level of control is bolstered by a more expansive list of matters that the NASDAQ Exchange (or any regional exchange) can veto, (2) absence of "size" in calculating the OTC national best bid and offer ("NBBO"), (3) the NASDAQ Exchange's unilateral right to extend hours, (4) a 4-1/2 year payout cap, and (5) preferential and confidential treatment regarding links to non-U.S. exchanges. We also understand that the OTC/UTP Plan processor does not report developments to the operating committee either orally or in writing and, indeed, that the operating committee has never met. In addition, the SEC's short sale rule, which otherwise applies to listed stocks, does not apply to stocks reported under the OTC/UTP Plan.
15 We are mindful that protracted negotiations over processor neutrality, development costs and revenue sharing unduly delayed the competitive entry of the Chicago Stock Exchange ("CHX") into the OTC market and postponed revenue sharing for more than a decade. In September 1985, the SEC ordered the NASD and the exchanges that wished to trade OTC stocks to submit a joint reporting plan by December 1, 1985. (See Rel. No. 34-22412.) Wrangling about the plan delayed CHX's entry until May 1987, when it commenced to trade OTC stocks under an interim plan. Three more years passed before a superseding plan was approved on a pilot basis (June 1990). (See Rel. No. 34-28146.) It then took another six years to reach an agreement on revenue sharing (March 1996) and, to this day, no regional exchange has received its full share of formula revenues for any calendar year. Other unresolved issues, such as a priority rule (BBO calculation) at variance with SEC Rule 11Ac1-2, the mandatory use of ACT as a trade comparison facility, the absence of a trade-through rule and the lack of effective inter-market linkage, continue to keep the SEC from permanently approving the OTC/UTP Plan.
16 Indeed, today's Wall Street Journal reports that another instance of queuing that occurred as recently as yesterday. (P. C7, col. 6.)
17 United States v. American Stock Exchange et al., No. 1:00CV02174 (D.D.C. filed Sept. 11, 2000), Complaint at 10; In the Matter of Certain Activities of Options Exchanges, AP File No. 3-10282, Exchange Act Release No, 43268 (Sept. 11, 2000), Order Instituting Public Administrative Proceedings Pursuant to Section 19(h)(1) of the Securities Exchange Act of 1934, Making Findings and Imposing Remedial Sanctions at 4, 6. The issue of capacity may turn out to be the "straw that breaks the camel's back" of data dissemination through consortia. Prior to the settlement of the antitrust proceedings, the Antitrust Division had already made clear its concerns about the options exchanges' sharing of capacity and related information. These concerns formed the basis for the following core requirement in the options antitrust settlements: "each [settling options] exchange [must] acquire and manage its own data transmission capacity independently." United States v. American Stock Exchange et al., Competitive Impact Statement of United States at 21. One has to wonder how an exchange can both meet that requirement and continue to engage in data dissemination through consortia. Indeed, the requirement would seem impossible for the NASDAQ Exchange to meet because of its undifferentiated intra-market and inter-market data system.
18 While we hope our proposal will lead to our near-term exit from the CTA/CQ Plans, we note that, as a transitional phase, our interim proposal accords with the SEC staff's wish to defer the proliferation of fee filings that could result from our withdrawal.
19 We recognize that an exchange will be able to choose between the OTC/UTP Plan and the CTA/CQ Plans in an effort to maximize its share of revenue. While choosing the plan(s) likely to provide the greatest percentage of revenues would be a factor in such a choice, a greater incentive is likely to be the inter-plan and inter-network distortions noted above. For example, as the Aber Memo points out, a shift to Network B from the OTC/UTP Plan would severely reduce the revenue share of the Amex and the other Network B Participants. However, we note that such a reduction is salutary, since the revenue per print on Network B is an order of magnitude higher than the revenue per print on Network A.
20 As you know, at the request of the Advisory Committee on Market Information, we will shortly be providing to you and the other committee members with a writing that describes the technical and practical basis for our view that consolidated dissemination via competing, non-exclusive processors will assure both (1) uniform sequencing and calculating of NBBOs and (2) the provision of back-up and retransmission capabilities. In this connection, we note that, whether acting as the CTA and CQ Plans' exclusive processor or as a non-exclusive, competing consolidator, SIAC can accommodate any increased capacity needs through the straightforward addition of hardware. Moreover, software releases scheduled for December will separate the three streams (i.e., the two streams for which SIAC acts as the exclusive processor (Network A and Network B) and the stream for which SIAC already acts as a non-exclusive processor (index values and other concurrent use)), permitting independent adjustments in capacity.

Attachment 2

  • The comments of James E. Buck, Senior Vice President and Secretary, New York Stock Exchange, Inc., April 10, 2000. (File name: buck1.htm)

http://www.sec.gov/divisions/marketreg/marketinfo/nysemi.htm

Home | Previous Page
Modified: 03/22/2001