Instinet
A REUTERS Company  

Mark Nienstedt
Acting President and CEO
and Chief Financial Officer

  Instinet Corporation
3 Times Square
New York, NY 10036
 
Tel     212.310.9500
Fax     212.302.7766

August 26, 2002

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

      Re:   Release No. 34-46249; File No. SR-NASD-2002-97

Dear Mr. Katz:

Instinet Corporation ("Instinet")1 appreciates the opportunity to share with the Securities and Exchange Commission (the "Commission") its comments on the Alternative Display Facility pilot (the "ADF Pilot") proposed by the National Association of Securities Dealers, Inc. (the "NASD") and recently approved by the Commission.2 Instinet welcomes the ADF Pilot as an essential step in meeting the Commission's policy goals set out in the SuperMontage Order3 and in maintaining the statutory requirements necessary for competition between and among exchange and non-exchange venues.

In Instinet's view, however, it is important not to confuse progress - even significant progress - with success. The Commission's stated goal in mandating the creation of the ADF was to ensure that "participation in SuperMontage will be entirely voluntary."4 Plans, proposals, and pilots are all important steps on the road, but only a fully functional and commercially viable ADF will address the need for a "meaningful display alternative to Nasdaq for OTC securities" as the Commission intended.5

The Commission specifically - and rightly, in Instinet's view - left open the question of whether the ADF Pilot was sufficient to satisfy the conditions laid down in the SuperMontage Order.6 It remains essential that the ADF be fully operational before SuperMontage begins trading, and as a matter of common sense and sound policy the ADF should not be considered fully operational without potential participants having a reasonable opportunity to establish linkages to the ADF and without having been tested on a system-wide basis in both test and live environments by potential ADF participants and other industry participants. Because of delays in clarifying the ADF proposal and providing the necessary technical specifications to potential participants, such development work and testing has not yet taken place and market participants are not in a position to rely upon the ADF.

As matters stand, if SuperMontage were to begin trading today there would be no practical alternative for participants in the OTC market to meet their obligations under Commission regulations (i.e., quotation display, order access, and trade reporting requirements), which is precisely the result the Commission sought to avoid in the SuperMontage Order.7 Furthermore, the absence of a viable ADF would leave no non-exchange trading venue after Nasdaq's exchange application is approved, contrary to the requirements of Section 11A of the Exchange Act.

It is vital that the ADF provide a real, commercially viable alternative to Nasdaq. Nasdaq's counterargument that "[n]owhere in the 2001 conditions is there a requirement that ADF must be commercially viable'" is so strained and superficial as to border on the ridiculous.8 One need only state the logical counterpart of Nasdaq's proposition - that the Commission's intended policy goals in the SuperMontage Order could be achieved by the creation of an ADF that technically complied with Commission regulations but was commercially unusable - to see its fallacy. It is clear from both the language of the SuperMontage Order and its context that the Commission did not intend to create a lose-lose choice between dependence on Nasdaq and reliance on an ADF with crippling flaws, but a real choice that would prevent Nasdaq's gaining anticompetitive advantages over NASD member- operated trading systems in the OTC market and satisfy the Congressional mandate to preserve and promote competition among the OTC market and exchange markets following Nasdaq's conversion to a for-profit exchange.9 We do not propose that the ADF could or should be required by regulation to become a commercially successful facility; the market will determine that. However, the competition provided by a genuine effort to provide commercially viable alternative will benefit not only users of the ADF, but users of Nasdaq, Nasdaq itself, and other venues that will be disciplined and motivated by competitive pressure. The Commission is therefore right to focus on the need "to gauge whether the ADF in practice fulfills the NASD's statutory obligations."10

We believe that satisfying the Commission's policy goals requires three additional things: a reasonable opportunity for coding to and testing (on an individual and market-wide basis and in a test and production environment) of the ADF before SuperMontage is permitted to launch; Commission monitoring of and readiness to intervene in the long-term fee schedule of the ADF; and a Commission mandate that the NASD take seriously the question of the ADF's commercial utility and consider those changes to its structure and operation that will improve its viability for a broad range of OTC market participants. While progress has been made, Instinet submits that these conditions have not yet been met.

To fully achieve these goals, Instinet respectfully suggests that the ADF Pilot be required to pass through a testing and evaluation phase before trading active securities on SuperMontage is permitted. Specifically, potential users should be given ninety days from the date of the ADF Pilot Order (July 24, 2002) to complete coding and bilateral testing. Next, the ADF Pilot should operate for a further thirty-day period to permit system-wide testing of the production environment and testing of linkages with the broader market so that non-ADF participants can assured of access to ADF participants' quotes under actual trading conditions. In addition, the NASD should be required to certify to the Commission that the thirty-day testing period has been successfully completed. Finally, and simultaneously with the testing periods above, concerns commenters have raised regarding the ADF's economics and functionality should be more fully addressed by the NASD.

I. Potential ADF Users Should Be Given Adequate Time to Code to and Test the ADF

The ADF has been a moving target. Although an initial partial proposal was published in the Federal Register in January 2002, the proposed fee structure was not available until March 2002, and the NASD did not publish a response to comments on its proposals until June 2002.11 Instinet, along with other industry participants, duly commented on each formal release.12 There has been considerable uncertainty regarding what form the ADF would take and whether it would meet the needs of a sufficient fraction of the industry to be a viable entity.

More importantly for present purposes, technical specifications needed to code to the ADF have been slow to arrive. Certain initial generic specifications to coding to XTP, the proprietary communications protocol the NASD has selected for the ADF, were made available in late February 2002. These draft specifications did not provide potential ADF participants the ability to begin coding to XTP. It was mid-July 2002 before potential users were provided with a users' guide for coding to XTP. The July 2002 users' guide identified the relevant aspects of the initial draft specifications, revised some of those specifications in material respects, and most importantly, contained the first, albeit incomplete, technical specifications for the ADF real-time order feed. The NASD subsequently provided Instinet with an updated version of the July 2002 users' guide on August 5, 2002. In addition, the question of whether and when an industry-standard FIX protocol will be available also remains open.13

Instinet has previously stressed the importance of timely provision of this critical technical information orally and in writing, and this issue cannot come as a surprise to any interested party.14 Seen in this light, Nasdaq's claim that the necessary information has been available since March and "[a]pparently, no NASD member took up NASD's offer to code or test" is highly misleading.15

These uncertainties and gaps explain why no market participant to our knowledge has yet completed coding to or testing with the ADF. It is vital that potential users be given an adequate opportunity to do so. For firms shifting quotation and/or trade reporting activity to the ADF, their ability to conduct business will depend on their ADF connection; it would be irresponsible to leave inadequate time for the coding effort or to begin ADF operations without full individual and industry-wide testing. Indeed, SuperMontage provides an instructive example; in that context it was deemed essential to provide for a detailed and extensive schedule of testing long after full technical specifications were made available.

This process must be completed before SuperMontage is allowed to become operational. If SuperMontage is permitted to begin trading now, market participants will face severe pressure to join it whether or not they want to. The ADF is being rushed into service with its technical work only partially completed and without any testing, and its potential users have not been given enough time to make the necessary preparations. As market participants cannot simply suspend operations while waiting for the ADF to come up to speed and are unlikely to bet their ability to operate on a hastily-constructed connection to an unproven system, they will tend to be forced into SuperMontage. If SuperMontage is conferred the advantage of liquidity and incumbency because of the absence of a viable alternative at its launch, the resulting barriers to entry may foreclose any later effort to launch a tested ADF and eliminate any chance the ADF had of preserving competition. The Commission was right to insist that the ADF "must be implemented prior to or at the same time as the SuperMontage," and we submit that the current state of the ADF falls far short of full implementation in any meaningful sense of the term.16

Instinet therefore suggests that the ADF Pilot be required to complete the following schedule before SuperMontage is cleared to trade. First, firms should be allowed ninety days from the date of the ADF Pilot Order to complete coding and bilateral testing with ADF systems. Second, a further thirty days should be allotted for system-wide testing of the ADF production environment and testing of linkages with the broader market so that non-ADF participants can be assured of access to ADF participants' quotes under actual trading conditions.17 If, after completing the testing schedule above, the NASD can certify to the Commission that the ADF has successfully completed such testing, then the ADF should be considered fully operational from a technical perspective.

II. The Economics of the ADF Must Be Viable on a Long-Term Basis

The Commission has noted and relies upon the NASD's temporary fee waivers and discounts as providing a preliminary solution to commenters' concerns regarding ADF fee levels. However, a rational firm evaluating whether to make a commitment to the ADF will not look only at the introductory fees that by their own terms expire in a few months. Rather, potential users are rightly concerned with the long-term cost of participating in the ADF. It is therefore critical to address the flaws in the NASD's fee proposal as soon as possible and before the ADF is declared fully operational in order to permit firms to make an informed decision on participation.

The ADF should not be required to be the lowest-cost service provider. However, particularly in light of its very limited functionality and the connectivity and other costs its users are required to bear, neither is there any apparent justification for the ADF's being a relatively high-cost provider. The NASD's commercial disinterest in running a market center cannot be allowed to justify an uncompetitive fee structure. The users the ADF is intended to service still must compete vigorously in a difficult market, and a fee structure that is not competitive with other venues would create commercial disadvantages for its users that would render it unusable as a practical matter.

In particular, there are three cost-related issues that should be re-examined during the pilot period. These issues were examined in greater detail in our previous comment letter and will be summarized briefly here.18

We also note that the NASD's argument that high fees are justified by the need to recoup its investment in the ADF fundamentally misapprehends the nature of the transactions that have taken place. The OTC trading community has already paid for the necessary OTC infrastructure through Nasdaq fees that built the existing Nasdaq systems. The NASD has chosen to convert that infrastructure to private gain through the Nasdaq restructuring, and it has done so in full knowledge of its statutory obligation to support OTC trading after Nasdaq becomes a for-profit exchange. While this decision is not improper, the transaction was meant to take place on arm's-length terms. Nasdaq should already have paid the NASD fair value for the assets it received, which should be enough for the NASD to build the limited replacement facilities it needs to fulfill its statutory obligations; indeed, Nasdaq should have paid much more given the higher value of Nasdaq's systems in comparison to the stripped-down ADF. NASD members should not see the public market infrastructure built using their fees given to a private party for less than their fair value and then be asked to pay to replace them.

III. The NASD Should Be Encouraged to Address the Commercial Concerns of ADF Participants

The position taken so far by the NASD, that the ADF is a regulatory exercise whose commercial viability is irrelevant, has deeply colored its response to comments and concerns submitted by potential users. In our view, the most important thing that the Commission can do at this stage is to impress upon the NASD its responsibility to create an ADF that provides a reasonable basis for commercial operations.

Instinet's previous letters describe the issues in detail, and we summarize only one issue here that has an immediate and significant impact on facilitating potential ADF market participants' access to the ADF. The status of the FIX interface, which has been expected imminently for nearly three months now, should promptly be clarified. While Instinet itself is working to code to the proprietary XTP protocol, there is substantial industry interest in a FIX interface because it is far less expensive and resource-intensive. However, as of the date of this letter, the NASD has not announced who will develop such an interface or when. One of the important reasons for delay in takeup of the ADF is that many market participants naturally do not want to incur the expense and difficulty of coding to XTP if an industry-standard FIX interface will be available soon. By delaying clarification of its FIX plans, the NASD is further delaying adoption of the ADF by creating uncertainty over whether coding to XTP will be a wasted investment in light of the availability of FIX. Instinet supports early development and deployment of a FIX interface to maximize the accessibility of the ADF to the widest possible universe of potential users.

IV. Conclusion

The Commission will not achieve its goals in requiring the creation of the ADF unless the NASD makes a bona fide effort to create a genuine alternative to Nasdaq's facilities that gives its users a fair chance to compete successfully. To that end, Instinet respectfully requests that the Commission delay the commencement of SuperMontage trading until the ADF has adequately been tested and its operational issues addressed. Specifically, the Commission should require:

  1. A 90-day period for firms' bilateral coding to and testing of the ADF interface, running from the date upon which the ADF Pilot was approved (July 24, 2002);
     
  2. A further 30-day period for testing of system-wide operations and interaction with other market participants (e.g., non-ADF members attempting to access ADF quotes); and
     
  3. The NASD's certification to the SEC that the 30-day system- and industry-wide testing period has been successfully completed.

Please do not hesitate to contact the undersigned, Peter D. Rich of our Washington office (202-789-8550), or Edward J. Rosen of Cleary, Gottlieb, Steen & Hamilton (212-225-2820), counsel to Instinet, if you should have any questions regarding this matter.

 

Sincerely yours,

Mark Nienstedt
Acting President and CEO
and Chief Financial Officer

 

cc:   The Honorable Harvey L. Pitt, Chairman
The Honorable Paul S. Atkins, Commissioner
The Honorable Roel C. Campos, Commissioner
The Honorable Cynthia A. Glassman, Commissioner
The Honorable Harvey J. Goldschmid, Commissioner
Annette L. Nazareth, Director, Division of Market Regulation
Robert L.D. Colby, Deputy Director, Division of Market Regulation
John Polise, Senior Special Counsel, Division of Market Regulation

 


1 Instinet is a registered broker headquartered in New York City. Together with its affiliates, Instinet currently trades in forty global markets and is a member of twenty exchanges around the world. Instinet is a pure agency broker, serving its global client base by consistently reducing transaction costs, and thereby increasing investment performance for investors and their proxies. More information about Instinet and its activities is available on Instinet's web site at www.instinet.com.
2 Securities Exchange Act Release No. 46,249 (July 24, 2002), 67 Fed. Reg. 49,822 (July 31, 2002) (File No. SR-NASD-2002-97) (the "ADF Pilot Order").
3 NASD Rulemaking: SuperMontage, Exchange Act Release No. 43,863 (Jan. 19, 2001), 66 Fed. Reg. 8019, 8054 (Jan. 26. 2001) (the "SuperMontage Order").
4 Id. at 8,054.
5 Id. at 8,052.
6 ADF Pilot Order at 49,848.
7 Instinet is aware that The Island ECN has developed the ability to post quotations on the Cincinnati Stock Exchange ("CSE"), although Island is providing access to its CSE quotations through Nasdaq's SelectNet system. However, until the CSE's proposed rule amendments to support decentralized trading and order delivery are approved, and the necessary systems changes to effectuate these rule changes are implemented by the CSE and ECNs, CSE is constrained in its ability to offer multiple ECNs a truly effective alternative in meeting their quotation display, order access, and trade reporting obligations under SEC regulations..
8 Letter from Hardwick Simmons, Chairman and Chief Executive Officer of The Nasdaq Stock Market, Inc., to Harvey Pitt, Chairman of the Commission (July 1, 2002) ("Nasdaq ADF Letter") at 4.
9 See Exchange Act Section 11A.
10 ADF Pilot Order at 49,848.
11 Exchange Act Release No. 45,156 (Dec. 14, 2001), 67 Fed. Reg. 388 (Jan. 3, 2002); Exchange Act Release No. 45,501 (Mar. 4, 2002), 67 Fed. Reg. 10,942 (Mar. 11, 2002); Exchange Act Release No. 45,991 (May 28, 2002), 67 Fed. Reg. 39,476 (June 7, 2002).
12 See Letter from Douglas M. Atkin, President and Chief Executive Officer of Instinet Corporation, to Jonathan G. Katz, Secretary of the Commission (Feb. 13, 2002) ("ADF Comment Letter"); Letter from Douglas M. Atkin, President and Chief Executive Officer of Instinet Corporation, to Jonathan G. Katz, Secretary of the Commission (Apr. 1, 2002) ("ADF Fee Comment Letter"); Letter from Jon Kroeper, First Vice President, Regulatory Policy/Strategy of Instinet Corporation, to Jonathan G. Katz, Secretary of the Commission (July 1, 2002) ("Amended ADF Proposal Comment Letter").
13 The ADF Pilot Order's observation that no other marketplace uses a FIX protocol is not quite accurate. For instance, the New York Stock Exchange ("NYSE") currently enables member firms to access NYSE via FIX. Regardless, the selection of XTP imposed an unnecessary barrier to the adoption of the ADF, but more importantly (as noted below) uncertainty over the status of the FIX interface delayed technical preparations for the ADF. Potential participants were naturally reluctant to undertake burdensome and expensive efforts to code to XTP if these efforts would soon be obviated by provision of a FIX interface, which would enable firms to interface with the ADF in a much shorter timeframe and at a reduced cost.
14 See Amended ADF Proposal Comment Letter.
15 Nasdaq ADF Letter at 4.
16 SuperMontage Order at 8,054 (emphasis added).
17 The industry's recent experience with the American Stock Exchange's initiation of trading in Nasdaq-quoted securities clearly illustrates the importance of avoiding potential market disruption by providing market participants with adequate testing opportunities to evaluate the impact of the introduction of a new market system on their order access needs and to make any necessary adjustments to their business practices.
18 See Amended ADF Proposal Comment Letter.
19 Allowing for the possibility of negotiation of course does nothing to resolve the problem, since the firm being subsidized has no incentive to agree to any other arrangement.