The International Securities Exchange, Inc.

September 12, 2003

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

Re: File No. SR-BSE-2002-15, Amendment No. 3; Release No. 34-48355

Dear Mr. Katz:

The International Securities Exchange, Inc. ("ISE") appreciates the opportunity to comment on Amendment No. 3 (the "Amendment") to the Boston Stock Exchange's ("BSE") proposed Boston Options Exchange ("BOX") rules package. We previously commented on the original proposal, arguing that the filing did not meet the requirements of the Securities Exchange Act of 1934, as amended (the "Exchange Act").1 Among other problems, we viewed BOX merely as a vehicle for the internalization of customer order flow, rife with conflicts of interest, and operating in a manner inconsistent with a broker-dealer's best execution obligations.

While the Amendment addresses a few of the issues we raised in our February Letter, overall we believe that the current proposal is a step backwards, actually raising more legal and policy issues than the original filing. As we will explain in detail below, the driving forces behind BOX are to limit competition, and to foster internalization, preferencing and payment for order flow ("PFOF"). If approved, competitive pressures will force other markets to adopt provisions similar to BOX. This will create a cascading degradation of the options markets, undoing the significant accomplishments that the industry and the Commission have achieved in developing a strong, competitive national market system for options trading.

We urge the Commission to institute proceedings to disapprove the filing. The primary deficiencies in the current proposal, discussed in more detail below, are:

  • Assault on the Auction Market: The central focus of BOX is the Price Improvement Period ("PIP"), which could drain the life-blood from the auction process. Order flow providers ("OFPs") will use the PIP to internalize their order flow or to sell their order flow to market makers. Competitive pressures will force other exchanges - including the ISE - to offer similar services, further spreading BOX's ill effects. Trading that occurs in an exchange's general auction (that is, outside of PIP and its progeny) will consist mainly of unattractive orders: orders that OFPs do not want to internalize; or Directed Orders (discussed below) that a market maker decides are not sufficiently profitable to enter into the PIP. Overall, this will result in wider quotations and harm to investors.

  • The "Directed Order" and Anticompetitive "Preferencing": The Amendment introduces the concept of a Directed Order. By allowing OFPs to preference their orders to market makers of their choosing, BOX not only would be fostering PFOF, but also would be eliminating the major incentive market makers have for disseminating competitive quotations. As proposed, preferenced market makers will receive Directed Orders regardless of the quality of their quotations, and will be able to examine the orders and decide whether it would be profitable to trade against them in the PIP. If not, the market maker simply could throw the orders back into the pond.

  • Conflicts of Interest and Best Execution Concerns: There are two concerns. First, the PIP and preferencing together create a structure that exacerbates the inherent conflicts of interest in internalization and PFOF. Second, any OFP who sends a non-marketable customer limit order to the BOX is violating its best execution obligation to that customer. This is because BOX will not provide a customer order with priority over professional trading interest at the same price.

  • Lack of Competition and Wide Quotations: BOX bans true market orders. Also, there likely will be few customer limit orders in BOX. Accordingly, market makers will know that the only orders actually traded in the BOX "auction" are those rejected by other market makers. With these likely being unprofitable orders, there will be a strong incentive to keep quotations wide. Furthermore, BOX proposes significant restrictions on a market maker who is at the national best bid and offer ("NBBO") and chooses not to enter a Directed Order into the PIP. Such a market maker must guarantee the execution at its quotation after other market participants have a chance to trade against the order. Since a market maker need not be quoting at the NBBO to receive a Directed Order, there will be strong incentives to quote outside the NBBO in order to provide greater flexibility in handling Directed Orders.

  • Inconsistency with Intermarket Linkage Requirements: The Amendment provides us with the first glimpse of BOX's Linkage rules. These rules are deficient both with respect to the sending and receiving of Linkage orders. Under the Linkage Plan,2 only a market maker with agency responsibility for customer orders may send Principal Acting as Agent Orders ("P/A Orders"). The BOX rules are inconsistent with that provision. The Linkage Plan also requires automatic execution of certain orders. By providing for a three-second "flash" of non-marketable incoming Linkage Orders, BOX is not in compliance with that requirement and inappropriately is giving its participants a "second look" at Linkage Orders.

I. The PIP is a Serious Threat to the Integrity of the Auction Market

The basic structure of the PIP remains unchanged from the original filing. OFPs and market makers will be able to enter orders into a mini three-second auction at a price at least a penny better than the then-current NBBO. The originator of the PIP is guaranteed at least 40 percent of an order if it at least matches the best price in the PIP auction. There also continues to be only a limited ability of other market participants to compete for the order in penny increments, and there is no minimum size requirement for PIP orders. The amended PIP thus continues to be an internalization vehicle that well exceeds the bounds of what the Commission previously has permitted.

The three most recent Commission chairmen all have raised significant concerns with internalization.3 Consistent with those concerns, the Commission has placed strict limits on systems that options exchanges offer that could result in internalization. These restrictions are intended to help ensure that there is a proper balance between preserving the quality of the auction market and permitting firms to provide additional liquidity when it would benefit the customer. Before the Commission should consider approval of the PIP as being consistent with the Exchange Act, we believe that the Commission should impose on the PIP the following restrictions: (1) the PIP should be limited to orders of 50 or more contracts; (2) all BOX participants should be permitted to participate in the PIP auction in pennies; (3) BOX must disseminate prices in the PIP auction to all market participants; and (4) BOX should require its participants to post the best price at which they are willing to trade against a customer order at the start of the PIP. We discuss the first three issues below and the last issue in Section III of this letter.

A. PIP Will Lead to Inferior Prices

The central problem with the PIP is that, ultimately, any supposed "price improvement" for a customer order in the PIP will be measured against an NBBO driven wider by a market structure focused on internalization, not price competition. This will result from the PIP removing small customer orders from the central auction, not just on BOX, but also on other exchanges as competitive pressure drives all exchanges to provide PIP-like functionality. Customer order flow is the foundation for the auction pricing mechanism, with market makers posting the best prices they are willing to pay for such orders. Removing orders from the auction and into PIP will reduce incentives for market makers to quote competitively.

We recognize that there are some orders for which a simple auction may not be appropriate. For example, there may not be enough liquidity in the market to handle large block orders. Thus, exchanges have developed special rules or functionality for such orders, like the ISE Block Mechanism.4 In contrast, the BSE is not arguing that the regular auction market is ill-suited to handle small customer orders. Rather, the BSE argues that the PIP will provide customers with price improvement over the NBBO and thus is a benefit to the market. We disagree with the BSE and, consistent with previous Commission rulings, the BSE should limit the PIP to orders of at least 50 contracts.

In attempting to justify the PIP, the BSE focuses on a specific customer order that will receive a price at least a penny better than the NBBO, rather than on the macro effect of the PIP on the market place. As the PIP and its progeny on other exchanges drain order flow from the auction there will be a cascading widening of quotations. Far from providing price improvement to customers, this will harm investors. Indeed, as discussed in more detail below, the BSE's proposed "Directed Order" will further exacerbate this problem.

B. PIP is a Private Auction

Another core problem with the PIP is the private nature of the auction. While BOX trumpets itself as an "open" market, it is selective in applying that openness. The PIP auction will be conducted in pennies. While BOX would disseminate PIP auction information to all BOX participants, few BOX participants will be able to compete effectively: Only market makers in the options series subject to the PIP can enter responses in pennies. All other professionals must respond by entering orders in standard increments, and thus cannot effectively compete for the order.

The BSE would open the auction as widely as possible if it actually sought to maximize the potential for customer price improvement. Rather, the BSE has restricted the ability of members to participate in the auction, presumably to limit the possibility that PIP responses would interfere with a member's opportunity to internalize its order flow. Contrast this with our proposed competitive response to PIP, in which all market participants can compete for an order in pennies.5 If the Commission approves the BSE's closed system, it is likely that our proposal will not be attractive to those firms seeking the maximum internalization potential. Thus, we are currently considering amending our proposal to match the restrictions of the PIP. This is the type of "race to the bottom" that the BOX filing has launched.

The Amendment does add the concept of the "Customer PIP Order" ("CPO"), which supposedly permits customers to participate in the PIP in penny increments. However, the CPO is not available to non-customers, who continue to be barred from competing for PIP order flow in pennies unless they are a market maker in that options series. Even for customers, the CPO process is severely limited and awkward to use. For the following reasons, we doubt that it will provide significant opportunities for customer participation in PIP auctions:

  • The CPO process requires that customers enter a new order type specifying a price both in standard increments and in pennies, with the hidden penny price to be used only in a PIP. While BOX market makers have the ability to see a PIP and decide whether to compete for the order, customers must make that decision in advance.

  • Even to be eligible to participate in the PIP auction, the customer limit order must be on the BOX book at a price equal to the NBBO. There is no such requirement that a market maker's quotation be at the NBBO to participate in the PIP.

  • While the filing is unclear on the point, it appears that the OFP handling the CPO must enter the CPO into the PIP, and that BOX does not handle this automatically.6 This will require OFPs to determine when a PIP begins and then reenter the CPO into the PIP auction within three seconds. It will be impossible for an OFP to enter the CPO manually within three seconds. Therefore, this approach will require OFPs themselves to automate the process. We also note that the Commission required us to increase our proposed time frame to expose orders in our Facilitation Mechanism from five seconds to 10 seconds. We fail to understand why the Commission would believe it necessary for us to expose ISE facilitation orders for ten seconds, yet permit the BSE to provide OFPs only three seconds to fulfill their agency duties with respect to CPOs. Indeed, it makes no sense to us that BOX even requires this extra step to enter a CPO into a PIP, especially when there is no requirement that market makers automate this process. BOX should automatically include all CPOs in the PIP.

For all these reasons, it appears to us that this provision is mere "window-dressing" to give the appearance of a more open PIP auction. In reality, the BSE continues to exclude non-market-maker professionals from the PIP, while providing functionality that makes customer access impractical.7

In the past few months we have had discussions with the Commission staff regarding specialized auctions in non-standard increments. In all those discussions the staff has made clear that they would not view favorably any such auction unless all market participants generally have the ability to compete for orders on the same pricing terms. We fully agree with the staff's position on this issue and we are prepared to meet this condition for any non-standard auctions that we might propose. If the BSE truly believes that competition in pennies is appropriate, it either should conduct its entire auction in pennies or should open up the penny PIP to all market participants, on a fair and even playing field.

C. PIP Violates the Firm Quote Rule

Section (a)(4) of the Firm Quote Rule,8 defines a bid or offer as a price at which (1) an exchange member is willing to buy or sell a security, and (2) that is communicated to any other broker-dealer or customer. All PIP orders are prices communicated by and to BOX participants. As such, section (b)(1)(i) of the Firm Quote Rule requires the BSE to collect these bids and offers and to make them available to quotation vendors. However, the BSE is proposing to limit the dissemination of PIP orders to BOX participants, and thus will not comply with the rule. We are not aware of any exemption that the BSE may be requesting from this requirement.

The Commission should require the BSE to comply fully with the Firm Quote Rule for all PIP orders. The purpose of this rule is to provide the investing public with the full and true sense of the market for a security. Hiding auctions so only BOX professionals have access to this information is contrary to both the letter and spirit of the law. There is a strong public policy goal of ensuring that all market participants have equal access to market information. Thus, the Commission should require the BSE to comply with all the requirements of the Firm Quote Rule.

II. The "Directed Order" - "Preferencing" - Further Threatens Market Integrity

The most significant change in the Amendment is the addition of the "Directed Order," which would introduce the controversial concept of "preferencing" into the listed options market.9 At the outset, we are troubled that in the more than seven months between the filing of the original proposal and the Amendment, the BSE has not concentrated its efforts on addressing the myriad problems with the BOX proposal. Rather, its focus appears to have been on how to add even more anticompetitive and damaging features to its proposed market structure.

The Directed Order permits OFPs to direct their order flow to a specified BOX market maker. The directed market maker then can use the PIP to trade against selected orders. When preferencing is coupled with the PIP, it is clear that the BSE is actually proposing a market in which: (1) OFPs that have a market making or proprietary trading capability can internalize their order flow; and (2) OFPs without such trading capability can capitalize on their order flow by selling it to market makers to whom the OFPs will preference their orders. The BSE has gone to great lengths to ensure that there will be few impediments to interfere with this captive, non-competitive form of trading.

For the reasons discussed below, the Commission should not approve any BOX proposal that includes Directed Orders. However, if the Commission determines to permit preferencing, it should: (1) prohibit preferencing to a market maker not quoting at the inside market; and (2) prohibit an OFP from preferencing a market maker for a size larger than the market maker's then-current quotation.

A. Assumptions Regarding Preferencing

It is difficult for us to understand some aspects of the BSE's filing. Thus, we base our analysis of preferencing in BOX on the following six important assumptions:

  • Market makers "opting out" of receiving directed orders. We assume an OFP can direct orders to any market maker of its choosing. The BOX rules do not provide an ability for a market maker to "opt out" of receiving Directed Orders. However, a BOX technical document implies that registration is necessary in order to receive Directed Orders.10 We assume that this reference is to registration as a market maker, and not a further registration as a permitted Directed Order recipient;11

  • Directing orders to OFPs. We assume that an OFP can direct an order only to a BOX market maker. That would be consistent with the definition of Directed Order in Section 1(a)(21) of the BOX rules. However, the Design Document, at Section 4.11.4, states that an OFP can direct a PIP order to "any other BOX Participant," which would include another OFP. If that is the case, it would appear to provide functionality to effect solicited transactions of all order sizes, which are pre-arranged trades effected outside of the auction market process. This raises issues similar to those raised by internalization. The ISE has had a proposal for solicited order functionality pending at the Commission for over two years. If the BSE seeks to permit OFPs to direct orders to other OFPs, a full description of that functionality needs to be included in the BOX rules and explained in detail in its filing;

  • Anonymity of preferencing. We make two assumptions here:

    • Anonymity in sending orders: When an OFP directs an order to a market maker (either affiliated or unaffiliated with the OFP), we assume the OFP will not have access to that market maker's individual quotation. We further assume that OFPs generally will not have access to information regarding the prices at which market makers may be willing to trade against orders in the PIP. Chapter VI, Section 10 of the BOX rules, dealing with informational barriers, appears to preclude such sharing of information among affiliates, and the anonymity of the system appears to preclude the sharing of information among non-affiliates. If the OFP has access to this information, it would raise significant issues both of fairness and of pre-arranged trading.

    • Anonymity in receipt of orders: When a market maker receives a Directed Order, we assume that the market maker does not know the identity of the sending OFP, and thus cannot limit from whom it receives Directed Orders. This is based on Chapter V, Section 14(e) of the BOX rules, which states that the "identity of Options Participants who submit orders to the Trading Host will remain anonymous to market participants at all times," except for error resolution and clearing;

  • Refusing to accept Directed Orders. We assume that a market maker cannot refuse to accept Directed Orders. In this regard, Chapter I, Section (a)(21) of the BOX rules gives OFPs the ability to direct Customer Orders to a specified market maker, and Chapter VI, Section 5(c) specifically states that a market maker cannot reject a Directed Order; and

  • Quoting as a condition to receiving Directed Orders. We assume that a market maker must be quoting in the specific series in order to be eligible to receive a Directed Order and therefore in order to use the PIP. Chapter VI, Section 5(c) of the BOX rules present only two scenarios when a market maker receives a Directed Order: either the market maker's quotation is equal to the NBBO or the quotation is not equal to the NBBO. There is no scenario discussed where the market maker is not quoting in the series.

If any of these six assumptions is incorrect, the BOX filing is materially misleading and would need to be amended. In such a case we would have concerns not only with the overall effect of preferencing on market quality, discussed below, but even greater concerns regarding the fundamental fairness of the BOX trading system.

B. Preferencing Will Harm Market Quality

A cornerstone of the ISE market model is that market makers never trade against non-block-sized orders unless they are quoting at the inside market. This market model encourages competition and keeps quotations narrow. By having narrow quotations, we attract additional order flow from OFPs, while also encouraging participants on competing exchanges to narrow their quotations. On other exchanges, a market maker must at least be competing in a crowd in order to participate in the auction. This is the case both for executions on a "wheel," and in the developing electronic or hybrid execution systems that require a market maker to be at the best price to interact with marketable orders. To date, the Commission has required that market participants be actively participating in an auction in order to be eligible to trade against incoming options order flow. Such a requirement is an obvious key to competition among markets and a successful national market system for options.

Compare the ISE model to the BOX internalization/preferencing model: An OFP can trade against their small customer orders simply by using the PIP. There is no requirement that the OFP have any pre-existing interest in the market at the NBBO (or at any other price). OFPs that do not have market making or proprietary trading capability can sell their orders to BOX market makers. A preferenced market maker would be able to examine a Directed Order and decide if it wanted to trade against it. The Rules contain no time limit on how long a market maker may review the order, thus giving rise to the possibility that the market maker could "play" the market in handling the order.12

Even more important, the market maker need not be quoting at the inside market, or even quoting in competitive size to be preferenced, thus diminishing the incentive to maintain narrow quotations with size.13 This is an integral part of structure of BOX that provides little incentive to quote competitively, and even provides incentives to maintain wide quotations. If the market maker sees a profitable trading opportunity, it will use the PIP, with the only price it pays being the cost of paying (or receiving) a penny more (or less) than the NBBO. If the market maker decides against the trade, it simply throws the fish back into the water.

C. Preferencing Can Lead to Game-Playing

The BOX rules give a market maker the choice of entering a Directed Order into the PIP or sending the order back into the BOX auction. If the market maker returns the order to the auction, the market maker's knowledge of this order raises regulatory concerns. The market maker can change its quote either to avoid trading against the order or to attempt to intercept the order ahead of other market participants. In an apparent attempt to guard against this potential abuse, the BOX rules would require a market maker quoting at the NBBO to couple any Directed Order redirected to the BOX auction with a "Guaranteed Directed Order" ("GDO") at least equal to the price and size of the market maker's NBBO quotation. We assume that the intent of the GDO is to prohibit the market maker from fading its quotation to avoid trading against the order. However, we do not believe that this is sufficient protection.

First, if a market maker is not quoting at the NBBO when it sends a Directed Order back to BOX, there appear to be no restrictions on the market maker's quoting or trading activities.14 This opens the door to abuse. For example, when the market maker sends the order back to BOX, the system will enter the order on the BOX book if it is not executable against the NBBO. The market maker then can immediately "pick off" that order by changing its quote or entering an order. While all market participants theoretically will have the opportunity to try to intercept the order, the entering firm's knowledge of the order will ensure that it is first to enter a counter-order.15

Even the GDO itself can lead to abuse. For the reasons discussed in detail in Section IIV of this letter, we generally do not expect market makers to be quoting at the NBBO. However, that situation could occur. Assume that a market maker is quoting at the NBBO at a size at least equal to a Directed Order, and the market maker decides not to price-improve the order by entering it in the PIP. It then must send the Directed Order to BOX, coupled with a GDO. There would be a three-second exposure of the customer order. In theory, any BOX participant can trade against the order in that three-second period. However, if there is no response to the "Directed Order Broadcast," the entering market maker gets to trade against the entire order.

These provisions effectively give a market maker a three-second facilitation mechanism for all-sized orders. While the market maker does not receive any guaranteed portion of the execution, unless BOX truly develops into a competitive trading system, the GDO will be a new vehicle for market makers to internalize directed order flow. The BSE provides no explanation as to why it is appropriate to require a 30-second exposure for OFP customer orders, but only three seconds for market makers handling customer orders.

D. Preferencing Goes Hand-in-Hand with PFOF

In its marketing presentations, BOX attempts to distinguish itself by stating that its market model does not contain an exchange-sponsored PFOF program. However, BOX's structure effectively removes competition from the market place. Thus, market makers freely can pay for order flow knowing that they can rely on guaranteed participation in trades, regardless of the quality of their quotations. Indeed, PFOF will be a more critical component on BOX than on the existing exchanges due to the lack of quotation competition on BOX. Without quote competition, firms will compete for order flow through payment deals.

Under the BOX structure, if a market marker decides to trade against a Directed Order, it is guaranteed 40 percent of the order; with the limits on participation in PIP, it will likely result in a much greater participation rate. Again, this guarantee is present regardless of the quality of the market maker's quote. This will provide a market maker with sufficient assurance that it can trade against Directed Orders that it can confidently pay for such order flow. The reason BOX proposes preferencing is to facilitate PFOF deals, not to protect against conflicts of interests when market makers both trade for their own accounts and handle customer orders.

Contrast BOX with a market model such as the ISE's, where a market maker trades against incoming order flow only to the extent it is quoting at the inside market, and only in a size proportionate to its quotation. Because this competitive model does not assure a market maker that it can trade against a specified percentage of incoming orders, it is not feasible for a market maker to pay for order flow on its own. Rather, the only way we can compete against PFOF on less competitive market models like BOX is to have all market makers pool their payments into a fund administered by the Primary Market Maker.16

E. If the Commission Permits Preferencing, it Should be Limited to Market Makers Actively Competing in the Auction

Taken as a whole, the BOX market model is inconsistent with the quotation-based model of competition underlying exchange registration requirements and the national market system mandated by the Exchange Act. If the Commission decides to approve this market structure as proposed, other markets - including the ISE - will have no choice but to seek to offer members the same internalization opportunities as customer order flow is diverted from the auction markets to be internalized on BOX. This will lead to snow-balling deterioration in the quality of the markets available to the investing public as quotes widen across the markets. The net effect will be worse prices for customer.

For all these reasons, we believe preferencing is bad for the options markets and bad for investors. Thus, the Commission should not approve any BOX proposal that includes Directed Orders. However, if the Commission determines to permit preferencing, at the very least, the BSE should amend the proposal in two important respects: First, the BSE should prohibit market makers from internalizing order flow in the PIP - or being preferenced with order flow - unless the market maker is quoting at the inside market at the time it seeks to use the PIP. Second, an OFP should not be permitted to preference a market maker for a size larger than the market maker's then-current quotation. At least this would help maintain some incentive for market makers to maintain competitive quotations.

III. BOX Raises Conflict of Interest and Best Execution Concerns

There are inherent conflicts when an intermediary seeks both to provide the best price possible to a customer and to maximize its profitability in handling an order. These conflicts are heightened when the intermediary either internalizes the order or sells it as part of a PFOF program. Through the PIP and preferencing, the BOX market structure takes these concerns to yet another, and higher, level. In addition, BOX's "flat book" structure gives rise to best execution concerns in the handling of customer limit orders.

A. The PIP and Preferencing Together Present Conflicts of Interest

It is an OFP's agency obligation to seek the highest price for a customer selling an option and the lowest price for a customer buying an option. Internalization and PFOF lead to obvious conflicts of interest for an OFP seeking to discharge that obligation, while also seeking to profit itself from the order it holds. The PIP and preferencing introduce an even more virulent strain of this conflict. As currently proposed, when an OFP or directed market maker initiates a PIP, it does not disseminate the highest price it is willing to buy the option or the lowest price at which it is willing to sell. If that were the case, the auction would offer others the possibility of providing even greater price improvement to the customer. In contrast, by allowing an OFP or directed market maker to improve its price in the PIP, the BOX structure anticipates that the OFP or market maker is willing to provide a better price to the customer than the initial bid or offer in the PIP.17

This leads to clear violations of an OFP's fiduciary obligations. An OFP who does not put forward its best price in initiating a PIP is not initiating the auction at the best price for the customer and thus does not provide best opportunity for price improvement. There is a similar violation of fiduciary obligations when an OFP directs the order to a preferenced market maker. The OFP knows that market makers seek to maximize their trading profits by buying at the lowest possible price and selling at the highest possible price. It is thus in a market maker's best interest to improve the price of the customer's order by the smallest increment possible. An OFP selling its order flow to a directed market maker thus knows that the recipient will do exactly the opposite of what is best for the customer! The Commission should require that an OFP or market maker start the PIP auction by disseminating the best price they are willing to provide the customer.

B. BOX Cannot Provide Best Execution of Customer Limit Orders

The Commission long has emphasized the importance of a broker-dealer's duty to provide its customer with best execution of its orders. The Commission has defined this duty as requiring "a broker-dealer to seek the most favorable terms reasonably available under the circumstances for a customer transaction."18 The Commission has emphasized that the scope of a broker-dealer's best execution obligations must evolve as changes occur in the market, and that broker-dealers must modify their order-routing procedures as price opportunities become "reasonably available."

In detailing best-execution obligations, the Commission has focused specifically on a broker-dealer's handling of limit orders, providing guidance that is critical in the evaluation of the BOX rules. Specifically, the Order Handling Release states:

Where material differences exist between the price improvement opportunities offered by markets or market makers, these differences must be taken into account by the broker-dealer. Similarly, in evaluating its procedures for handling limit orders, the broker-dealer must take into account any material differences in execution quality (e.g., the likelihood of execution) among the various markets or market centers to which limit orders may be routed.19

BOX will not provide customer limit orders with priority over professional trading interest. Thus, if a broker-dealer sends a non-marketable customer limit order to BOX, and there exists professional trading interest on the book at the same price, the customer order will stand behind the professional interest in priority. In contrast, if the broker-dealer sends that customer order to the ISE - or to any options exchange other than the BSE - the customer will have priority over any pre-existing professional interest on the book.

Since a broker-dealer will be well aware of these different priority rules when it makes its customer order-routing decisions, it is clear that the broker-dealer will violate its best execution responsibility to its customers any time the broker-dealer sends a customer limit order to BOX without first confirming that there is no professional trading interest already on the BOX book at the same price. If the BSE is able to resolve the other substantial issues with its filing and the Commission determines to approve the BOX proposal, we urge the Commission to include in its approval order a reminder to broker-dealers that BOX's unique trading algorithm gives rise to this inherent best execution problem.

IV. BOX Has Multiple Incentives to Maintain Wide Quotations

BOX provides a strong incentive to maintain wide quotations. As discussed, OFPs can internalize small customer order flow even though they are not market makers disseminating quotations. Moreover, market makers can receive small customer Directed Orders - the most attractive order flow available - regardless of the quality of their quotations. This eliminates one of the major reasons for market makers to disseminate competitive quotes.

This market structure actually encourages market makers to widen their quotes and to keep their quotes small. They know that most order flow entering the non-PIP auction will be orders that internalizers or preferenced dealers reject. These generally will be unprofitable orders. In order to avoid having to trade against these orders, market makers will keep their quotations wide and limit the size of their quotes.

BOX even penalizes market makers that quote at the NBBO. As discussed, if a market maker receives a Directed Order, it can either enter that order into the PIP or send it back into BOX. However, if the market maker is quoting at the NBBO at the time it returns the order to BOX, it is required to couple the Directed Order with a GDO. This effectively requires the market maker to guarantee the execution of the Directed Order at the then-current NBBO for at least the size of the market maker's quotation. Moreover, the market maker's quotation loses any priority it may have had at the NBBO, and the market maker will trade only when all other trading interest is exhausted.

We assume that the BSE is proposing these provisions to prevent a market maker that is quoting at the NBBO from sending a Directed Order back to BOX and then fading its quote to avoid trading against the order. However, imposing these restrictions severely penalizes any market maker who quotes competitively to attract order flow and is, at the same time, preferenced. To avoid this harsh result, market makers are likely to ensure that they are not at the NBBO, allowing them to accept preferenced orders without penalty.

Another disincentive to competitive quoting is the fact that OFPs will not have the ability to send customer market orders to BOX. The Amendment specifically provides that "the BOX Market will not have a 'market order,' as that term is typically used, that can be executed at successive price levels." Rather, in the "BOX-Top Order," once such an order is executed in part, the remainder of the order is converted into a limit order on the BOX book. BOX is pointedly telling its OFPs not to send market orders to this "open" auction. Knowing that BOX will not be attracting market orders eliminates a significant incentive market makers would have to maintain competitive quotations.20 Taken as a whole, BOX seems uniquely structured to avoid intramarket quotation competition.

V. The BOX Linkage Rules Fail to Comply with the Intermarket Linkage Plan

The BOX rules are inconsistent with the Linkage Plan with respect to both the sending and receipt of Linkage orders.

A. The BOX Rules Do Not Adequately Control the Sending of Linkage Orders

The Plan specifically states that only a market maker can send a P/A Order, and that the order must be for the account of a market maker acting as agent for an underlying customer order. While the section of the BOX rules specifically devoted to Linkage parrots the required language when describing the sending of P/A Orders, the rest of the BOX rules, and the BSE's description of the process, are inconsistent with these requirements.

In describing its order-handling process in the filing, the BSE states that there will be a three-second "flash" of an unexecutable order at the NBBO. If a customer order is not executed in this process, the filing states that "a P/A will be generated and sent to the away exchange that is displaying the NBBO." This passive sentence structure, repeated in the rule text, leaves ambiguous who is sending the P/A Order. However, the Design Document resolves this ambiguity by stating that "a P/A Order is generated by BOX, in the name of a designated BOX Market Maker."21 In case that is not sufficiently clear, the document further states: "Note that the P/A order is initiated either manually or automatically by BOX, not by BOX Participants themselves."22 The BOX web site further describes the process of routing P/A Orders when there is a better market away:

the BOX trading engine software will automatically generate a PA order for the remaining quantity and transmit it automatically to the BOX Market Operations Center for routing to one or more of the away exchanges posting a better price. BOX MOC will decide to which exchange the PA order will be routed and process any trade executions received. Any quantity remaining after this process will be reintroduced to the BOX trading engine.23

It is clear that no BOX market maker has agency obligations with respect to P/A Orders. The BOX market maker does not decide to which exchange to send the order if there are multiple exchanges at the NBBO. In addition, the Linkage Plan requires that market makers exercise discretion on how best to handle large-sized P/A Orders. There are multiple ways to send these orders, and a market maker must use its discretion, in acting as agent for the customer, on how best to execute the order. Since BOX gives its market makers no discretion on how to handle P/A Orders, there is no indication how BOX is addressing this.

Given the overall structure of BOX, we are concerned that the BSE is planning to use Linkage simply to off-load orders to other exchanges when participants in its own non-competitive system decide not to trade against an order. With a market structure offering few incentives for competitive quotations, we anticipate that many Directed Orders rejected from PIP and "thrown back into the water" will pass through the three-second flash and sent through the Linkage to other exchanges. It is not the intent of the ISE (nor, we believe, the intent of the other linked exchanges) for the Linkage to be used as an automated dumping ground of unattractive orders.

B. BOX Should Not "Flash" Unmarketable Incoming Linkage Orders

Under the Linkage Plan, a market maker can send a Principal Order ("P Order") or P/A Order only when the receiving exchange is disseminating a quotation at the NBBO. Upon receipt, an exchange must execute smaller orders in its automatic execution system, and has 15 seconds to respond to larger orders. The BSE is proposing that if it receives a P or P/A Order of any size where it is not disseminating a quotation at the current NBBO, it will "flash" that order for three seconds at the limit price. If there is trading interest at that price it will execute the order; otherwise it will reject the order.

The Linkage Plan does not give the BSE the right to hold up orders eligible for automatic execution for three seconds while it flashes the order internally. In proposing to do so, the BSE misunderstands the purposes of Linkage and gives its market participants an inappropriate second look at an order after they have faded their quotations from the NBBO. Specifically, the purpose of Linkage is to give market makers on one exchange limited access to other exchanges either to access superior prices for their customers (using P/A Orders) or to clear away markets (using P Orders). Thus, P and P/A Orders are limited price orders, priced at the NBBO, with strict controls on their use.

1. P/A Orders

The Linkage Plan states that a market maker may send a P/A Order to another exchange only when that exchange is at the NBBO. The market maker must price the order at the NBBO, the so-called "Reference Price." Section 7(a)(ii)(A) of the Linkage Plan states that the receiving exchange "must execute the P/A Order in its automatic execution system, if available, if its disseminated quotation is equal to or better than the Reference Price when that order arrives...." Implicit is the requirement that the receiving exchange reject the order if it is not then at the NBBO. However, if BOX receives a P/A Order after it has faded from NBBO (and thus is no longer quoting at the Reference Price), it will neither automatically execute nor reject the order, as required by the Plan. Rather, it will "flash" the order for three seconds.

The Linkage Plan does not give BOX the right to provide its members with a "second look" to see if they want to trade this order at a price from which they faded just moments ago. During this period, the market could fade even further, leading to an inferior customer execution when BOX finally does reject the order and the sending market maker finally must price the order for the customer. Contrast this with the 15-second time period the Plan provides to allow receiving exchanges to execute larger orders. In that case, the receiving exchange's quotation presumably has remained at the NBBO (or the exchange should reject the order when its quote moves) and the customer is not harmed. However, during BOX's three-second flash, by definition the market is moving against the customer and the quality of his or her execution is at risk.

2. P Orders

The Linkage Plan provisions for P Orders are similar to the provisions for P/A Orders, and do not give receiving exchanges the ability to "work" orders when the receiving exchange has faded from the NBBO. In practice, a market maker will send a P Order to BOX to clear away a quotation that it believes is out of line. Once that quotation is clear, the market maker will have greater latitude in servicing its customers without trading through the BOX quotation. If BOX has already faded from the NBBO by the time it receives a P Order, there no longer is a reason for the order to be executed. The market maker sending the order is not looking for price improvement for its principal trading interest. Rather, the goal is to clear the market, which already has been accomplished. Also, as with P/A Orders, this three-second flash gives BOX market makers an opportunity to hold up an away market maker when the market may be moving adversely to the sending party's interests. Thus, there again is no reason to "flash" the order.

VI. The BSE Inappropriately Attempts to Provide Market Makers with Specialist Privileges

The BOX rules contain a provision stating that BOX "market makers are designated as specialists on BOX for all purposes under the Exchange Act or Rules thereunder."24 Being designated as a specialist provides a market maker with certain benefits, such as the ability under Regulation M to continue to trade an option on a security when the market making firm is involved in the underwriting of the security underlying the option.25 However, there are costs in being a specialist. Specifically, under Section 11(b) of the Exchange Act and Rule 11b-1 thereunder, specialists are subject to both "affirmative" and "negative" trading obligations.

One of the most important negative obligations on a specialist is the requirement that a specialist "restrict his dealings so far as practicable to those reasonably necessary to permit him to maintain a fair and orderly market. . . ."26 While this broad limitation on trading has many facets, a bedrock principle of exchange trading always has been that specialists should not trade ahead of customer orders at the same price. Indeed, it is hard to imagine how a specialist can "restrict his dealings" to maintain a fair and orderly market for the investing public when the specialist is trading ahead of customers.

As discussed above, on BOX, public customers do not have priority over professionals - including BOX's so-called "specialists" - at a given price. Rather, BOX, is a flat-book, first-in, first-served trading system. If a BOX "specialist" places a quote or order on the book ahead of a public customer limit order, that "specialist" is entitled to trade first. In addition, a market maker trading a Directed Order in the PIP could trade ahead of customer orders entered into the auction during the three-second exposure period. In contrast to this market structure, on the ISE professionals, including market makers, always yield priority to customer orders, even if a member enters such an order during the exposure period of our Facilitation Mechanism.27

We do not quarrel with BOX's flat-book system. We believe that the market place should determine whether such a market structure is viable. Nor do we quarrel with BOX seeking to designate its market makers as specialists. We do so on the ISE.28 However, BOX cannot both have a flat-book system allowing market makers to trade ahead of the public and also deem these market makers to be specialists.

The Commission currently is considering a number of other applications that call into question the very nature of what constitutes an "exchange" and how it should regulate the persons who trade on those markets. These applications will require careful consideration of who is a specialist and the permissible trading algorithms of markets that employ a specialist system. These are important decisions for the national market system and the Commission should consider them carefully. It should not inadvertently create an adverse precedent by approving the current BOX filing.

VII. Incorporation of Previous Comments

Finally, we again note that we incorporate into this letter the comments from our February Letter. To the extent that the BSE has not addressed our comments, we continue to believe that the filing is deficient, even as amended. Our February Letter addresses a number of issues addressed above, including the effect of the PIP on market structure and the inappropriate labeling of BOX market makers as specialists. In addition, that letter raises other issues that we believe that BOX must address, including, but not limited to, the general applicability of BSE rules to BOX and the standard of care market participants must give to orders during the PIP. We believe that those issues remain relevant to the Commission's consideration of the amended filing.

VIII. Conclusion

Our comment letters detail a large number of specific problems with the BOX proposal. As such, it can be viewed as a description of various "trees." We urge the Commission to review carefully each of these trees. We also urge the Commission not to lose sight of the "forest" that results from these various trees. That is, while each element is a problem, taken together they magnify our market structure concerns:

  • The PIP and preferencing, taken together, remove the incentive to quote competitively in order to attract order flow;

  • The PIP and preferencing will result in a dramatic increase in internalization and PFOF;

  • The PIP and preferencing will remove from the auction small customer orders that provide the basic foundation for the competitive market maker system;

  • The PIP and preferencing exacerbate the inherent conflicts of interest in internalization and PFOF;

  • The BSE proposes many other features that will encourage wide quotes, such as the prohibition on market orders, the best execution conflicts in sending limit orders to the exchange, and the restrictions BOX imposes on market makers who are quoting at the NBBO when they receive Directed Orders; and

  • The BSE proposes to misuse the Linkage as a vehicle for BOX itself, and not its market makers, to dump on other exchanges unattractive orders.

The Commission should not approve the revised proposal. BOX would seriously damage the auction market in options trading while ushering into that market the concept of preferencing. It would lead other markets to copy this model to remain competitive, harming market participants and the investing public by degrading the quality of the overall market in options. The filing is seriously deficient when measured against the standards in the Exchange Act. We urge you to institute disapproval proceedings. If you have any questions on our comments, please do not hesitate to give us a call.

Sincerely,

Michael J. Simon
Senior Vice President and Secretary

cc: Chairman Donaldson
Commissioner Atkins
Commissioner Campos
Commissioner Glassman
Commissioner Goldschmid
Annette Nazareth
Robert Colby
Elizabeth King
Stephen Williams

____________________________
1 Letter dated February 12, 2003 from Michael J. Simon, Secretary, ISE, to Jonathan G. Katz, Secretary, Commission (our "February Letter"). Since the BSE has addressed only a few of the issues we raised in February, we specifically incorporate by reference that letter into this comment letter.
2 Plan for the Purpose of Creating and Operating an Intermarket Options Linkage, as last amended on January 31, 2003.
3 See Securities Industry News, September 1, 2003, at page 18, quoting Commission Chairmen Donaldson, Pitt, and Levitt referring to internalization as, among other things, posing "a conflict of interest for broker-dealers," and as "compromis[ing] a broker's fiduciary obligation to achieve best execution of its customer orders."
4 See ISE Rule 716.
5 See File No. SR-ISE-2003-06.
6 In an e-mail dated September 2, 2003, Thomas Peterffy, Chairman of Interactive Brokers ("IB," who Mr. Peterffy describes as a "major sponsor" of BOX), seeks the support of IB's customers for the BOX filing. In that e-mail, he describes the CPO as follows:

[c]ustomers may also participate in the PIP, in direct competition with market makers. You could, for example, tell us that you are bidding $2.20 for a particular option but are willing to go up to $2.22 if a PIP of a sell order in that option takes place. In this case, IB would bid $2.20 and if a PIP started we would match the initial PIP bid price of $2.21. If the bid for the PIP order is raised to $2.22, we would match that bid on your behalf. Alternatively, you could tell us to bid $2.22 in a PIP right from the start, so as to give you time priority. The end result is that you have a chance, for the first time in the U.S. options market, to trade against other orders between the quotes.

Neither the BOX rules nor BSE filing describes this type of flexibility in the CPO, or this level of discretion in a broker's handling of CPOs. If this is how the CPO really works, the filing is materially deficient. The BSE must describe exactly how the PIP auction will work, and the public should have an opportunity to review and comment on those procedures.

7 The BOX proposal also would bring the concept of "penny jumping" to the options market. When a public customer has a limit order at the inside market on the BOX book, a market maker could jump ahead of that order by a penny. Without the true ability of public customers to compete in the PIP in pennies, this raises serious policy issues. See "Stepping on Traders' Toes?" Active Trader, Volume 4, No. 10, October, 2003 at page 32.
8 The Firm Quote Rule is Rule 11Ac1-1 under the Exchange Act.
9 "Preferencing" exists in the listed equity market through the third market (the over-the-counter market for listed stocks), as well as on the Cincinnati Stock Exchange ("CSE") and BSE. The Commission permitted preferencing on the CSE and BSE in 1996 only after conducting an exhaustive review of the practices, including analyzing data regarding the actual effects of the CSE and BSE pilot programs on the quality of their markets and the national market system. Moreover, unlike the current state of the options market, internalization by market makers of listed equities in the third market was a longstanding practice at the time the Commission considered whether it was appropriate to allow preferencing on equity exchanges. See Exchange Act Release Nos. 37045 and 37046 (March 29, 1996).
10 BOX Business Design Guide (Specifications for FIX access to BOX) Version 1.0, August 2003 ("Design Guide") at Section 4.11.4.6, "Registration of Executing Participants."
11 The Design Guide also differentiates between "Directed Orders" and "Routed Orders." The BOX rules contain no such differentiation. We assume that these are technical terms and that all aspects of the "Routed Order" fit within the legal definition of the "Directed Order" in the rules.
12 The Design Document at Section 4.11.4 states that if a market maker does not respond "after a period = n (n to be set by the exchange), the [Directed Order or Routed Order] will automatically be booked on BOX." We believe that this "n" period should be included within BOX's rules so that market participants can understand how long market makers may examine Directed Orders and play the market.
13 As noted above, we assume that the market maker must be at least quoting in the series in order to be preferenced. If this assumption is incorrect, then the issues we raise are exacerbated because it not only discourages competitive quoting, but discourages quoting at all.
14 This is yet another example of a rule that will drive market makers to maintain wide quotations. See Section IV of this letter.
15 The Amendment does contain a revised provision in Chapter V, Section 18(c) of the BOX rules that requires OFPs to wait 30 seconds before trading against their own agency orders they place on the book unless they use the PIP. However, that 30-second provision does not apply to market makers; in fact, it specifically excludes "Market Makers handling Directed Orders...."
16 The ISE reiterates its support for abolishing PFOF, but only if the Commission addresses the overall issue of both exchange-sponsored and member-administered programs. See Letter dated February 7, 2003 from David Krell, President and CEO, ISE, to Harvey Pitt, Chairman, Commission.
17 As opposed to the BOX market structure, the ISE Facilitation Mechanism requires the entering firm to specify the best price at which it is willing to fill its customer's order.
18 This quotation is from the Commission's release adopting the "order handling rules," which provides a detailed discussion of the best execution obligations. Exchange Act Release No. 37619A (September 6, 1996), (the "Order Handling Release"), at footnote 349 and accompanying text.
19 Id. at footnote 357 and accompanying text (emphasis added).
20 Not only will BOX prohibit true market orders, but, as discussed below, OFPs will be violating their best execution obligations if they send customer limit orders to BOX.
21 Design Document at Section 4.14.2.2.
22 Id. (Emphasis in original.)
23 Although undefined, we assume that "MOC" stands for Market Operations Center, or some equivalent BOX function.
24 Chapter VI, Section 1 of the BOX rules.
25 See 17 CFR 242.100 et seq.
26 Rule 11b-1(a)(2)(iii).
27 This is another area that will result in a "race to the bottom" if the Commission permits market makers to trade ahead of customers. Competitive forces will require that we reconsider providing customers with priority if it will result in OFPs directing order flow to other exchanges that have fewer impediments to their ability to internalize order flow.
28 ISE Rule 800(a).