May 12, 2000

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

Re: Release No. 34-42450; File No. SR-NYSE-99-48

Commission Request for Comment on Issues Relating to Market Fragmentation

Dear Mr. Katz:

Primex Trading N.A., LLC ("Primex Trading") appreciates the opportunity to comment on the above-referenced Commission Release ("Release").1 In the Release, the Commission solicits comment on a broad range of market fragmentation issues in connection with the elimination of NYSE Rule 390.2

Specifically, the Commission outlined six possible approaches in its Release to address fragmentation concerns in the event it is determined that some regulatory action is warranted. These range from simple disclosure of execution quality and order routing processes by market centers and broker-dealers, to a strict system of inter-market price/time priority. Primex Trading is submitting its views with respect to one particular proposal in the Release relating to the exposure of market orders to price competition ("Market Order Exposure Proposal").3

Executive Summary

Primex Trading believes that an increase in the exposure of market orders to price competition in some form is desirable, feasible, and inevitable, given the recent evolution in market structure, the proliferation of new technologies, and the pending move toward decimals. Of the proposals suggested by the Commission, we believe the Market Order Exposure Proposal can provide the greatest benefit with the least amount of market dislocation. If properly implemented, the proposal can enhance the interaction of trading interest and maximize opportunities for price improvement for customer orders. Most importantly, this can be accomplished in a manner that strengthens the overall market without sacrificing the ability for market centers and their participants to innovate and compete for order flow. Any action by the Commission to advance this proposal should be in a form that encourages exposure of market orders to the greatest extent possible, but should not be strictly mandatory.

Response to Commission Request for Comment on Market Order Exposure

Primex Trading is an innovative developer of auction trading systems recently formed by several leading financial services firms for the purpose of developing new platforms for trading financial products.4 We are in the process of implementing a unique trading system within Nasdaq that incorporates an electronic auction application in a multi-dealer environment. Because our system relies, in large part, on fundamental principals of order exposure and price competition, we feel the Commission may be interested in the views of Primex Trading as they relate to the Market Order Exposure Proposal as set forth in the Release.

We offer the following responses to the Commission's questions:

1. What would be the advisability and practicality of this option?

Primex Trading supports the basic notion of exposing market (and marketable limit) 5 orders to the market at large. The benefits that continue to accrue to customers as a result of the Order Handling Rules' display requirements demonstrate the efficacy of exposure with respect to limit orders. These benefits include greater transparency and access to available trading interest, increased order interaction, and, most importantly, improved execution quality for customer orders.

As markets continue to evolve, Primex Trading believes the Commission should continue its efforts along these lines by supporting structures and technologies that extend similar benefits to market orders. While market orders represent more than one-third of the total universe of individual investor orders, they currently do not, or cannot, enjoy the maximum benefits of systematic exposure via the conventional means of the public quote stream. While some exchanges already expose market orders to a limited audience via "open outcry," we believe that technology is available to automate this process much further, removing inefficiencies and expanding the exposure well beyond the confines of a physical trading floor. This can be particularly instrumental in achieving the Commission's goals when applied to a multi-dealer environment.

In addition to improving execution quality for customer orders, an exposure process also could yield a new source of public market data relating to the flow of exposed orders, depending on the manner in which market orders are expressed to the broader market. New technology can and should be used to more broadly disseminate this information to the general public and interested participants. This will provide another dimension of information, beyond currently available quotes and trade reports, that the public could use as a tool in assessing the market and making investment decisions, thus reducing unfair informational advantages.

2. Would it effectively address the problems presented by market fragmentation?

We believe that exposure of market orders would address two of the most significant issues that many believe require renewed attention.

First, exposure can provide a greater degree of order confluence, particularly in a multi-dealer market, by making a larger pool of trading interest available to a broader range of participants. Discrete pools of market orders that do not interact today can be more effectively linked or aggregated, increasing the potential for greater interaction. This in turn can provide improved prices and enhanced liquidity for customer orders.

We believe this can be accomplished in a manner that does not sacrifice the ability of market centers and their individual participants to continue to innovate and provide technological advances in the handling of, and competition for, that order flow, thus addressing age-old conflicts involving centrality and monopolistic tendencies of markets. Market centers and their individual participants can participate in an order exposure mechanism on both an intra-market and inter-market basis as a way to provide additional bases on which to compete (e.g., in terms of the collection and dissemination of this type of market information, available liquidity, price improvement opportunities, access and trade management tools, and other value-added services that markets and participants may offer to compete for order flow). This is consistent with Chairman Levitt's view that the industry needs to search for new approaches that use technology to strengthen the overall market and seek technological innovation that "garners the benefits of centrality while allowing markets to vigorously compete."6

Second, and perhaps even more important from a regulatory perspective, a systematic exposure mechanism that makes heretofore undisclosed market orders more transparent and available to a wider audience can provide enhanced abilities for broker-dealer fiduciaries to obtain price improvement for their customers as part of the firm's best execution obligations. These obligations have taken on new significance recently due to the proliferation of ECNs/ATSs, which has resulted in an increase in the number of potential execution destinations. Best execution obligations will become even more complicated in a decimal environment when liquidity, wherever it may reside, becomes distributed across many more price points. Exposing more market orders to a broader universe of participants in an organized fashion can enable fiduciaries to obtain better prices and enhanced liquidity for their customers. In addition, such exposure provides a meaningful and objectively-determined process for arriving at the price ultimately provided to the customer. This is particularly important in situations where the fiduciary supplements the exposure process by committing its own capital to customer orders.

3. Would exposure of market orders through the quote mechanism provide a viable means of allowing the holders of undisclosed orders to interact with market orders?

While Primex Trading believes in the value of exposing market orders to a broad audience for price competition, careful consideration should be given as to how this should be accomplished. In the Release, the Commission suggested two approaches as to how the Market Order Exposure Proposal for market orders could be implemented: (1) exposure of orders in a system that provided price improvement to some specified percentage of orders over time; and (2) exposing orders in the public quote for some specified time whenever the NBBO spread is greater than the minimum variation.

As explained below, the use of publicly displayed quotes as a vehicle for exposing market orders, particularly where there are many quote contributors (multiple market centers or in a multi-dealer environment such as Nasdaq), may not maximize the price competition for the orders and may not be the most viable method going forward, regardless of the order size, although it certainly can be one alternative means. Instead, we believe that there are other methods (discussed in questions 4 and 5 below) of communicating available interest to a broad, public audience, without being limited to the use of the public quote. These methods can be more effective in eliciting liquidity not otherwise disclosed, which can interact with market orders on a more competitive basis. This liquidity may be in the form of other market orders, large orders being "worked," or capital supplied by market makers and specialists.

Limit orders, by virtue of their priced nature, can be exposed relatively easily by simply integrating the price and size of the order directly into the public quote. Unpriced market orders, on the other hand, raise unique issues, particularly in a decimal environment. Specifically, market orders do not contain a price and thus must be "converted" into a static, absolute-priced order for purposes of display in a market maker's or specialist's quote, at least for some period of time.

Merely exposing a market order as a priced order one minimum variation better than the NBBO (as the Commission suggests in the Release) may not serve to maximize the benefits of exposure for the order, particularly in a decimal environment. For example, assuming a minimum price variation ("MPV") of a nickel, if the quoted spread is 20 x 20.10 and a broker-dealer displays a customer's market order to buy by posting an improved bid of 20.05, the customer is most likely precluded from ever getting an execution better than 20.05, even though there may have been other participants willing to sell to that buyer at a lower price (e.g,, between 20 and 20.05) if given the opportunity (nickel MPV notwithstanding). Thus, while it is possible that such an exposure can result in the relatively better price of 20.05 for the customer, such price would not truly be the result of any "competition" for the order, but would result only where another market participant sees the 20.05 bid and is able to execute against it.

These flaws can become even more pronounced if the MPV is a penny. For example, assume that instead of immediately executing the market order to buy at the offer price, the firm displays the buy order by improving the bid by one penny. The market quote of 20 x 20.10 now becomes 20.01 x 20.10. This may not be enough enticement to attract another participant to sell to the order, delaying an execution for the life of the exposure. More problematic is the likelihood of the "penny jumping" phenomenon, where other participants or active traders seeking price priority (or other customer market orders being displayed in the quote for exposure purposes) may easily "penny jump" the original order by posting an improved price at relatively low cost to the penny jumper before the original order can get executed. This may nullify the benefits of the exposure and delay a fill for what was intended to be a market order to buy seeking prompt execution. Alternatively, the firm could display the order at a point one MPV away from the opposite side of the market, in effect displaying the market order to buy at the highest price the quotes will allow. For example, if the market is 20 x 20.10, the buy order can be displayed as a bid for 20.09. This, however, would preclude the customer from receiving an execution from a willing seller at better prices between 20 and 20.08. Thus while the order may have an improved opportunity for execution, it comes with the cost of foregoing an even better price. Either way, true price competition for the order is lacking.

It becomes clear that any price chosen for purposes of displaying an unpriced market order in the quote is almost completely arbitrary, and may not maximize price improvement opportunities for the order. This arbitrariness is reflected in many price improvement algorithms that market centers and individual participants use today. In fact, the primary motivation for forming Primex Trading was to develop a trading model where price improvement opportunities can be defined by dynamic and competitive forces, rather than static, ad hoc assumptions of what is or should be the "best" price.

Furthermore, attempts at using the public quote stream for the exposure of market orders would surely compound the traffic and capacity problems that the industry is wrestling with today. The exposure of market orders via quotes would be in addition to the increased quote traffic that already has resulted from the required display of limit orders, and which is expected to increase even further in a decimal environment. While the exposure of market orders in the quote may have been feasible historically when there was less volatility and quotes moved in increments no smaller than 1/8th or 1/16th, the smaller increments that are expected going forward will only make the NBBO more volatile, placing further strains on the capacity of systems already taxed.

Many of these concerns are not new. As noted in the Release, the Commission's suggestion on the use of the public quote as a vehicle to display market orders is similar to prior exposure proposals, the most recent of which was contained in the original proposal for the Order Handling Rules published in 1995.7 That exposure proposal was deferred pending the implementation of the Order Handling Rules and assessment of its effects.8 Even at that time, many involved in the process indicated their concern of how the proposed exposure rule would work in practice, including then-Commissioner Wallman, who was reported as expressing concerns about "flickering quotes" as a result of such exposure.9 We believe these are legitimate concerns today, and will be even more of a concern in the future decimal environment.

4. Are there effective means of representing undisclosed orders in markets in which trading interest is divided among many different market centers?

While Primex Trading does not completely foreclose the possibility of employing the public quote as one possible vehicle for exposing market orders, we do suggest that the Commission consider new and alternative approaches that may be more efficient and provide a greater degree of price competition and order interaction for market orders. Such approaches need not be mutually exclusive, and can complement other methods, including use of the quote. Quotes will surely continue to play an important role as both a vehicle for trading and as a barometer of the condition of the market, but it also must be recognized that not all participants need, or want to be confined to a flat model where trading interest can only be expressed and accessed in the form of a displayed quote. Only recently has new technology been introduced to supplement customary quote-based trading with a variety of mechanisms tailored to the needs of some participants. Examples include new types of crossing and matching systems that allow undisclosed interest to interact while minimizing costs and market impact. The Commission should allow these new systems to develop to their full potential and ensure that all participants, particularly individual investors, can benefit from them. We must not lose sight of the overriding goal of better executions for customers, regardless of how such executions come about.

To that end, Primex Trading believes that new technologies will be able to more effectively represent or expose heretofore undisclosed market orders to a broader universe of participants. We believe that the first step is to provide more transparency to this interest in new and dynamic ways, and by providing participants with the tools to access such interest on a fair and equal basis.

For example, one way to accomplish this is to employ new technology to emulate and expand upon the traditional floor trading model of exposing orders to a trading crowd for price competition. Primex Trading is finalizing development of such a system that facilitates the interaction of buyers and sellers through an electronic auction process. The system is predicated primarily on providing price improvement opportunities for market orders. These orders are announced to an electronic "crowd" who "bid" for them within the context of the NBBO in a fair, orderly, and highly automated manner using a variety of new and sophisticated tools designed for this purpose.

This will allow trading interest that would otherwise remain isolated or undisclosed to be represented to a broader audience. Because the platform is completely electronic, the exposure process is not bound by the confines of the physical space surrounding a floor trading post, or the number of traders that may be physically present at such location at any given time. Thus the exposure and execution mechanisms can be accessed by a much larger and virtually unlimited universe of participants, providing the greatest potential for price competition and crowd interaction. While we don't presume our particular model should be the only way the Commission's goals can be met, we do believe that new technologies providing greater exposure and "dynamic pricing" are inevitable and cannot be ignored.

5. What other means to facilitate the interaction of undisclosed and disclosed orders is feasible and practical?

The next step would be to ensure that the aforementioned mechanisms are fully integrated and linked into existing market centers so that they can work in tandem with other display and execution systems. This has already occurred to some degree. For example, SROs are enhancing their systems to incorporate features such as multiple quote level display, reserve size, and automatic execution. Other participants are deploying sophisticated order management tools coupled with so-called intelligent routers that seek out market centers and liquidity pools with the greatest likelihood of a quality execution based on factors that not only include price, but size and speed of execution as well. These systems incorporate displayed interest and other liquidity pools, providing more robust interaction between those participants who choose to display their interest and those who do not.

These types of advances are sure to continue. Primex Trading also is integrating different forms of trading interest to maximize the liquidity available to customer orders. In addition to our exposure mechanism, we provide participants with the ability to have their customer orders automatically forwarded to displayed quotes or supplemented with additional liquidity supplied by the participant. The integration of these elements helps elicit latent, undeclared trading interest for the benefit of investors' orders, but does not neglect displayed interest.

6. Would requiring the exposure of market orders to price competition unwarrantedly delay the execution of those orders? If so, should order exposure be offered as a choice to customers?

The delay, if any, would be a function of the method by which the order was exposed and executed. Existing systems employing Application Program Interfaces ("APIs"), as well as technology that Primex Trading has developed, allow for meaningful price competition to take place in an environment where we believe there may be virtually no delay at all in a majority of situations.

In some circumstances, depending on the characteristics of the stock and market conditions, there may be brief periods of time during which a market order may not quickly elicit contra-side interest while seeking opportunities for price improvement or enhanced liquidity. Even in these instances, however, we share Chairman Levitt's suspicion that "most retail investors would willingly sacrifice 5 or 10 seconds in order to get a chance for price improvement."10

Where a customer has determined that waiting a few seconds for the opportunity of a better price is inconsistent with his or her trading strategy, or during certain market conditions where price improvement opportunities may not be meaningfully available, we believe customers and their fiduciaries should have some flexibility to opt out of the exposure mechanism, or rely on alternative mechanisms that provide the benefits of exposure without any attendant delay.

7. How would implementation of this option affect the opportunity for execution of displayed trading interest at the NBBO?

Broker-dealers who handle market orders have a fiduciary and regulatory responsibility to consider opportunities for price improvement for their customers. This has long been a part of their best execution obligations. Existing market centers have long competed for order flow on the basis of their trading model's ability to provide these price improvement opportunities.

A natural consequence of this is that, in some cases, it is possible that a publicly displayed order may not receive an execution as quickly is it would have, where an incoming market order is price improved by a specialist or other market participant instead of being matched with the displayed order. While the incoming customer order receives the benefit of a better price, this also could be construed as an adverse selection cost imposed on the limit order customer.

Any such adverse selection costs of using limit orders should not, however, be viewed as a dysfunctional characteristic of the market structure. It is merely a practical consequence of the tradeoff between limit orders and market orders, recognizing the fiduciary duty owed to the market order. By displaying its order in the quote, the limit order customer chose to define its price parameter and remain passive, in contrast to the customer who chose to use a market order to actively extract liquidity "at the best available price."

What is critically important, however, is that the motivations and ability for participants to interact with the incoming market order in this way (forestalling an execution against the displayed order) should be based on independent and non-discriminatory competition for that incoming order. It should not result from arbitrary determinations, or unfairly opportunistic choices made by participants with exclusive or privileged information regarding market orders. This underscores the need for broad dissemination of market order exposure information, and fair and equal access to that order flow by any participant interested in interacting with it.

Other factors, unrelated to the exposure of market orders, will have a greater influence over any impact on the execution opportunities for displayed limit orders. Decimalization, for example, will further transform the value proposition that limit orders offer to investors. Historically, limit orders were used to obtain "standing" or priority. As we have seen, this value has shrunk with the minimum quote increment. As participants are able to quote in finer increments, it becomes less expensive to gain standing by simply posting a marginally improved quotation, resulting in a "race to the front of the line." While this may benefit the participant who can post a limit order or quote the fastest, it does not necessarily benefit the average investor whose previously-displayed limit order falls further back in the price/time priority queue as it is leapfrogged by others. We believe this will reach critical mass with decimals, particularly in a penny MPV environment.

As a result, we believe that displayable limit orders will be used and valued for other reasons. Specifically, limit orders will be increasingly relied upon as more of a tool to limit the boundaries of an order to avoid transacting at a price unacceptable to the investor. This has already been observed with an increase in the use of limit orders for commencement of trading in IPOs. Limit orders will also continue to be used by investors who want to define a point at which they want to be in or out of the market in an easy and inexpensive way (e.g., by leaving a limit order in a book away from current prices, the investor need not actively monitor the markets or the status of their order). As long as the market center in which the limit orders are posted provide some form of priority among similar types of orders (e.g., price/time priority among displayable limit orders) and as long as investors and participants understand the rules of engagement, they can make informed choices as to the market center and type of order best suited for them in that market.

It is conceivable that investors dissatisfied with the value of using limit orders may seek alternative order types to achieve their goals. This may include an increase in the use of market orders, in which case an exposure mechanism with a competitive pricing element will become even more crucial to ensure the proper handling of customer orders.

Irrespective of the market structure or exposure mechanism, we believe that basic, existing protections for customer orders will continue to play an important role in providing an incentive for customers to continue to use limit orders. From a regulatory perspective, this includes the "Manning" rules, which will ensure that investors who use limit orders are protected when the fiduciary handling their order trades for its own account at prices equal to or better than the customer's limit price.11 Distinct from these Manning obligations, and from a more competitive perspective, firms or market centers can (and do in fact today) offer "competitive print protection." This gives the customer additional protection of obtaining an execution when another market center, or participant other than the firm handling the order, trades at the customer's limit price or better.

In addition, limit order display mechanisms such as ECNs and exchange books will always compete and seek to provide value-added services for the handling and execution of limit orders. Factors such as liquidity, speed, cost, reliability, and service already are used to define and distinguish the marketing of ECNs and exchanges, and will no doubt continue. These systems should be encouraged to compete for order flow by developing the better market mechanism and providing value. The market itself will ensure that those systems that do not provide value will not survive.

Advisability of a Mandatory Exposure Requirement

While the Commission should foster a market structure environment that facilitates the broadest exposure possible for all types of orders, it is unclear at this time whether such exposure, as it applies to unpriced market orders, should be mandatory. At least initially, we believe that participants and their customers should be given some latitude as to how best to handle at least some portion of their market orders. As customers and other participants become more knowledgeable and discriminating as to how their orders are handled, we believe that natural, competitive forces will dictate the manner and degree of market order exposure.

This is not to imply that those who may rely on a particular exposure mechanism for the benefit of their customers should not be held to certain fundamental standards of fairness so as not to undermine the benefits that exposure can provide to the market as a whole. For example, a market maker or specialist with access to a system that facilitates the exposure and interaction of customer market orders should not have the advantage of being able to trade for its own account against the customer orders of their competitors in such a system, unless it is willing to provide a commensurate degree of liquidity to the system in the form of its own customer order flow as well. In other words, a market maker or specialist who is not willing to expose its own customer orders on a fair and consistent basis to other participants should not be able to selectively trade as principal for its own account with the customer orders of other participants who have made such a commitment. Firms unwilling to provide this level of commitment could still represent their customer orders as agent in such a system, at the firm's discretion, but firms who choose to use the system in this manner should not have the benefit of being able to selectively access the system's liquidity solely for their proprietary accounts.

We thank you for considering our views and would be delighted to discuss them further with the Commission and staff. We can be reached at (212) 230-2475.

Sincerely, Sincerely,

Glen Shipway
CEO
Primex Trading N.A., LLC

Peter B. Madoff
Manager
Primex Trading N.A., LLC

cc: Hon. Arthur Levitt, Chairman, SEC
Hon. Norman S. Johnson, Commissioner, SEC
Hon. Isaac C. Hunt, Commissioner, SEC
Hon. Paul R. Carey, Commissioner, SEC
Hon. Laura S. Unger, Commissioner, SEC

Mark J. Ready, Chief Economist, SEC
Annette Nazareth, Director, Division of Market Regulation, SEC
Robert L.D. Colby, Deputy Director, Division of Market Regulation, SEC
Elizabeth King, Associate Director, Division of Market Regulation, SEC

Footnotes

1 See Exchange Act Release No. 42450 (February 23, 2000); 65 FR 10577 (February 28, 2000).

2 As the Commission is aware, NYSE Rule 390 served to prohibit NYSE members from effecting transactions in NYSE-listed securities away from an exchange.

3 See Release at 10586 (Section IV.C.2.c.). While Primex Trading is not providing any detailed views at this time with respect to the NYSE portion of the Release, we generally support the elimination of any barrier to a free and open marketplace, and believe that market centers should compete vigorously on the merits of their respective trading models and services.

4 More information can be found at www.primextrading.com.

5 Unless otherwise noted, the term "market orders" throughout our comments generally refers to both marketable limit orders as well as unpriced market orders.

6 Dynamic Markets, Timeless Principles, speech by Arthur Levitt given at Columbia Law School, New York, NY, September 23, 1999.

7 See Exchange Act Release No. 36310 (September 29, 1995); 60 FR 52792 (October 10, 1995) ("Proposed Order Handling Rules").

8 See Exchange Act Release No. 37619A (September 6, 1996); 61 FR 48290 (September 12, 1996) (Approval of Order Handling Rules) at 48321 (Section III.C.). This proposal did not require orders to be exposed per se, but rather required only that they be provided with some type of price improvement opportunity. The Commission proposed a non-exclusive safe harbor that would satisfy this requirement, which involved "stopping" or guaranteeing the best price publicly displayed and then exposing the order at an improved price in the public quote for 30 seconds.

9 SEC Commissioner Voices Concern About Nasdaq Access Rule Plan, by Neil Roland, Bloomberg (August 14, 1996).

10 Best Execution: Promise of Integrity, Guardian of Competition, remarks by Arthur Levitt before the Securities Industry Association, Boca Raton, Fla., November 4, 1999.

11 See NASD Rule IM-2110-2 (Trading Ahead of a Customer Limit Order).