Release No. 34-51808

File No. S7-10-04

Regulation NMS

Section II: Order Protection Rule

II. Order Protection Rule

The Commission is adopting Rule 611 under Regulation NMS to establish protection against trade-throughs for all NMS stocks. Rule 611(a)(1) requires a trading center (which includes national securities exchanges, exchange specialists, ATSs, OTC market makers, and block positioners)⁵² to establish, maintain, and enforce written policies and procedures that are reasonably designed to prevent trade-throughs on that trading center of protected quotations and, if relying on an exception, that are reasonably designed to assure compliance with the terms of the exception. Rule 611(a)(2) requires a trading center to regularly surveil to ascertain the effectiveness of its policies and procedures and to take prompt action to remedy deficiencies. To qualify for protection, a quotation must be automated. Rule 600(b)(3) defines an automated quotation as one that, among other things, is displayed and immediately accessible through automatic execution. Thus, Rule 611 does not require market participants to route orders to access manual quotations, which generally entail a much slower speed of response than automated quotations.

Rule 611(b) sets forth a variety of exceptions to make intermarket price protection as efficient and workable as possible. These include an intermarket sweep exception, which allows market participants to access multiple price levels simultaneously at different trading centers – a particularly important function now that trading in penny increments has dispersed liquidity across multiple price levels. The intermarket sweep exception enables trading centers that receive sweep orders to execute those orders immediately, without waiting for better-priced quotations in other markets to be updated. In addition, Rule 611 provides exceptions for the quotations of trading centers experiencing, among other things, a material delay in providing a response to incoming orders and for flickering quotations with prices that have been displayed for less than one second. Both exceptions serve to limit the application of Rule 611 to quotations that are truly automated and accessible.

By strengthening price protection in the NMS for quotations that can be accessed fairly and efficiently, Rule 611 is designed to promote market efficiency and further the interests of both investors who submit displayed limit orders and investors who submit marketable orders.⁵³ Price protection encourages the display of limit orders by increasing the likelihood that they will receive an execution in a timely manner and helping preserve investors' expectations that their orders will be executed when they represent the best displayed quotation. Limit orders typically establish the best prices for an NMS stock. Greater use of limit orders will increase price discovery and market depth and liquidity, thereby improving the quality of execution for the large orders of institutional investors. Moreover, strong intermarket price protection offers greater assurance, on an order-by-order basis, to investors who submit market orders that their orders in fact will be executed at the best readily available prices, which can be difficult for investors, particularly retail investors, to monitor. Investors generally can know the best quoted prices at the time they place an order by referring to the consolidated quotation stream for a stock. In the interval between order submission and order execution, however, quoted prices can change. If the order execution price provided by a market differs from the best quoted price at order submission, it can be particularly difficult for retail investors to assess whether the difference was attributable to changing quoted prices or to an inferior execution by the market. The Order Protection Rule will help assure, on an order-by-order basis, that markets effect trades at the best available prices. Finally, market orders need only be routed to markets displaying quotations that are truly accessible. Accordingly, as discussed in detail below, the Commission finds that the Order Protection Rule is necessary and appropriate in the public interest, for the protection of investors, and otherwise in furtherance of the purposes of the Exchange Act.

A. Response to Comments and Basis for Adopted Rule

Rule 611 as adopted reflects a number of changes to the rule as originally proposed. As discussed below, the Commission has made these changes in response to substantial public comment on the proposed rule and on the issues arising out of the NMS Hearing that were addressed in the Supplemental Release. In addition, the adopted rule includes a new exception for certain "stopped orders" in response to the suggestions of commenters on the reproposal. The public submitted more than 2200 comments addressing the trade-through proposal and reproposal.⁵⁴ Although the comments covered a very wide range of matters, they particularly focused on the following issues:

(1) whether an intermarket trade-through rule is needed to promote fair and efficient equity markets, particularly for Nasdaq stocks which have not been subject to the current ITS trade-through provisions;

(2) whether only automated and immediately accessible quotations should be given trade-through protection and, if so, what is the best approach for defining such quotations;

(3) whether intermarket protection against trade-throughs can be implemented in a workable manner, particularly for high-volume stocks;

(4) whether the exception in the original proposal allowing a general opt-out of protected quotations is necessary or appropriate, particularly if manual quotations are excluded from trade-through protection;

(5) whether the scope of quotations entitled to trade-through protection should extend beyond the best bids and offers of the various markets; and

(6) whether the benefits of an intermarket trade-through rule would justify its cost of implementation.

In the following sections, the Commission responds to comments on the trade-through proposal and reproposal and discusses the basis for its adoption of Rule 611.

1. Need for Intermarket Order Protection Rule

Commenters were divided on the central issue of whether intermarket protection of displayed quotations is needed to promote the fairest and most efficient markets for investors.⁵⁵ Many commenters strongly supported the adoption of a uniform rule for all NMS stocks to promote best execution of market orders, to protect the best displayed prices, and to encourage the public display of limit orders.⁵⁶ They stressed that limit orders are the cornerstone of efficient, liquid markets and should be afforded as much protection as possible.⁵⁷ They noted, for example, that limit orders typically establish the "market" for a stock.⁵⁸ In the absence of limit orders setting the current market price, there would be no benchmark for the submission and execution of marketable orders. Focusing solely on best execution of marketable orders (and the interests of orders that take displayed liquidity), therefore, would miss a critical part of the equation for promoting the most efficient markets (i.e., the best execution of orders that supply displayed liquidity and thereby provide the most transparent form of price discovery). Commenters supporting the need for an intermarket trade-through rule also believed that it would increase investor confidence by helping to eliminate the impression of unfairness when an investor's order executes at a price that is worse than the best displayed quotation, or when a trade occurs at a price that is inferior to the investor's displayed order.⁵⁹

Other commenters, in contrast, opposed any intermarket trade-through rule.⁶⁰ These commenters did not believe that such a rule is necessary to promote the protection of limit orders, the best execution of market orders, or efficient markets in general. They asserted that, given public availability of each market's quotations and ready access by all market participants to such quotations, competition among markets, a broker’s existing duty of best execution, and economic self-interest would be sufficient to protect limit orders and produce the most fair and efficient markets. They therefore believed that any trade-through rule would be unnecessary and costly. These commenters also were concerned that any trade-through rule could interfere with the ability of competitive forces to produce efficient markets, particularly for Nasdaq stocks.

Commenters on the original proposal who were opposed to any trade-through rule also expressed their view that there is a lack of empirical evidence justifying the need for intermarket protection against trade-throughs. They noted, for example, that trading in Nasdaq stocks has never been subject to a trade-through rule, while trading in exchange-listed stocks, particularly NYSE stocks, has been subject to the ITS trade-through provisions. Given the difference in regulatory requirements between Nasdaq and NYSE stocks, many commenters relied on two factual contentions to show that a trade-through rule is not needed: (1) fewer trade-throughs occur in Nasdaq stocks than NYSE stocks;⁶¹ and (2) trading in Nasdaq stocks currently is more efficient than trading in NYSE stocks.⁶² Based on these factual contentions, opposing commenters concluded that a trade-through rule is not necessary to promote efficiency or to protect the best displayed prices.

The Commission has carefully evaluated the views of these commenters on both the original proposal and the reproposal. In addition, Commission staff has prepared several studies of trading in Nasdaq and NYSE stocks to help assess and respond to commenters' claims. The studies and the Commission's conclusions are discussed in detail below. In general, however, the Commission has found that current trade-through rates are not lower for Nasdaq stocks than NYSE stocks, despite the fact that nearly all quotations for Nasdaq stocks are automated, rather than divided between manual and automated as they are for exchange-listed stocks. Moreover, the majority of the trade-throughs that currently occur in NYSE stocks fall within gaps in the coverage of the existing ITS trade-through rules that will be closed by the Order Protection Rule. Consequently, the Commission believes that the Order Protection Rule, by establishing effective intermarket protection against trade-throughs, will materially reduce the trade-through rates in both the market for Nasdaq stocks and the market for exchange-listed stocks.

In addition, the commenters' claim that the Order Protection Rule is not needed because trading in Nasdaq stocks, which currently does not have any trade-through rule, is more efficient than trading in NYSE stocks, which has the ITS trade-through provisions, also is not supported by the relevant data.⁶³ This conclusion is particularly evident when market efficiency is examined from the perspective of the transaction costs of long-term investors, as opposed to short-term traders. The data reveals that the markets for Nasdaq and NYSE stocks each have their particular strengths and weaknesses. In assessing the need for the Order Protection Rule, the Commission has focused primarily on whether effective intermarket protection against trade-throughs will materially contribute to a fairer and more efficient market for investors in Nasdaq stocks, given their particular trading characteristics, and in exchange-listed stocks, given their particular trading characteristics. Thus, the critical issue is whether each of the markets would be improved by adoption of the Order Protection Rule, not whether one or the other currently is, on some absolute level, superior to the other. The Commission believes that effective intermarket protection against trade-throughs will produce substantial benefits for investors in both markets and, therefore, has adopted the Order Protection Rule for both Nasdaq and exchange-listed stocks.

a. Trade-Through Rates in Nasdaq and NYSE Stocks

The first principal factual contention of commenters on the original proposal who were opposed to a trade-through rule is premised on the claim that there are fewer trade-throughs in Nasdaq stocks, which are not covered by any trade-through rule, than in NYSE stocks, which are covered by the ITS trade-through provisions.⁶⁴ One commenter asserted that, outside the exchange-listed markets, competition alone had been sufficient to create a "no-trade through zone."⁶⁵ To respond to these commenters, the Commissions staff reviewed public quotation and trade data to estimate the incidence of trade-throughs for Nasdaq and NYSE stocks.⁶⁶ It found that the overall trade-through rates for Nasdaq stocks and NYSE stocks were, respectively, 7.9% and 7.2% of the total volume of traded shares.⁶⁷ When considered as a percentage of number of trades, the overall trade-through rate for both Nasdaq and NYSE stocks was 2.5%. When considered as the size of traded-through quotations as a percentage of total share volume, the overall rates for Nasdaq and NYSE stocks were, respectively, 1.9% and 1.2%.⁶⁸ In addition, the staff study found that the amount of the trade-throughs was significant – 2.3 cents per share on average for Nasdaq stocks and 2.2 cents per share for NYSE stocks.⁶⁹

The staff study also revealed that a large volume of block transactions (10,000 shares or greater) trade through displayed quotations. Block transactions represent approximately 50% of total trade-through volume for both Nasdaq and NYSE stocks.⁷⁰ Importantly, many block transactions currently are not subject to the ITS trade-through provisions that apply to exchange-listed stocks. Broker-dealers that act solely as block positioners are not covered by the ITS trade-through provisions if they print their trades in the over-the-counter ("OTC") market. In addition to not covering the trades of block positioners, the ITS trade-through provisions include an exception for 100-share quotations. They therefore often may fail to protect the small orders of retail investors. When block trade-throughs and trade-throughs of 100-share quotations are eliminated, the overall trade-through rate for NYSE stocks is reduced from 7.2% to approximately 2.3% of total share volume.⁷¹ The two gaps in ITS coverage therefore account for most of the trade-through volume in NYSE stocks. The Order Protection Rule, by closing these gaps in protection against trade-throughs, will establish much stronger price protection than the ITS provisions.

Commenters opposed to the trade-through reproposal offered a number of criticisms of the staff study. Such criticisms generally fall into two categories: (1) possible reasons why the staff study might have overestimated trade-through rates, particularly for Nasdaq stocks; and (2) even assuming the estimated trade-through rates were accurate, arguments for why such rates do not support a conclusion that the Order Protection Rule is needed or will benefit the markets, particularly for Nasdaq stocks. These criticisms are evaluated below.

i. Accuracy of Estimated Trade-Through Rates

Several commenters asserted that the staff study overestimated trade-through rates because it failed to consider the existence of reserve size and sweep orders in the Nasdaq market, which could have caused "false positive" trade throughs.⁷² In theory, order routers could intend to sweep the market of all superior quotations before trading at an inferior price, but if they did not effectively sweep both displayed size and reserve size, the superior quotations would not change and the staff study would report a false indication of a trade-through when the trade in another market occurred at an inferior price. In practice, however, those who truly intend to sweep the best prices are quite capable of routing orders to execute against both displayed and estimated reserve size, thereby precluding the possibility of a false positive trade-through. Indeed, although commenters asserted that the staff study failed to consider the existence of reserve size for Nasdaq stocks, the validity of their own argument is premised on the failure of sophisticated market participants to consider the existence of reserve size when routing sweep orders.

It currently is impossible to determine from publicly available trade and quotation data whether the initiator of a trade-through in one market has simultaneously attempted to sweep better-priced quotations in other markets.⁷³ The data can reveal, however, the extent to which false-positive indications of a trade-through were even a possibility by examining trading volume at the traded-through market. If the accumulated volume of trades in that market did not equal or exceed the displayed size of a traded-through quotation, it shows that a sweep order, even one attempting to execute only against displayed size, could not have been routed to the market that was traded-through. Commission staff therefore has supplemented its trade-through study to check this possibility and to help the Commission assess and respond to commenters' criticisms. It found that this possibility rarely occurs – a finding that fully supports an inference that market participants are capable of effectively sweeping the best prices, both displayed and reserve, when they intend to do so.⁷⁴ Thus, it is very unlikely that the existence of reserve size and sweep orders caused a significant number of false positive trade-throughs in Nasdaq stocks.⁷⁵

One commenter asserted that the staff study was flawed because its sample trading days involved unusual trading activity.⁷⁶ Commission staff chose the sample trading days, however, only after affirming that they were representative of normal trading. To respond to this commenter's claim, Commission staff reaffirmed that all four days were well within the norms for trading volume and price volatility.⁷⁷ In addition, the trade-through rates remained quite stable across the four days (e.g., ranging only from 2.3% to 2.6% for Nasdaq stocks).⁷⁸

Two commenters asserted that, even if the staff study's estimate of trade-through rates was correct for the trading days chosen in the Fall of 2003, such rates are now outdated for Nasdaq stocks because of structural changes in the market.⁷⁹ In particular, they cited the merger of the Island and Instinet ECNs and Nasdaq's acquisition of the BRUT ECN. Nasdaq also presented statistics indicating that the trade-through rates for Nasdaq stocks in some trading centers had dropped from the Fall of 2003 to the Fall of 2004. The staff study used data from the Fall of 2003, however, because it was prior to the Commission's proposal of a trade-through rule and its public announcement that the staff was reviewing trade-through rates. While the conduct of market participants may have changed in certain respects when they were a focus of regulatory attention, the Commission cannot be assured that such behavior would continue if the Commission did not adopt the proposed regulatory action to address trade-throughs.

Indeed, Nasdaq's own data illustrates this possibility.⁸⁰ Although Nasdaq asserts that the reduction in trade-through rates from 2003 to 2004 is a result of fewer independently operating ECNs, its data undercuts this explanation. For example, Nasdaq's data shows that the trade-through rate at internalizing securities dealers dropped from 3.2% in 2003 to 1.4% in 2004.⁸¹ It is unlikely that ECN consolidation could have caused such a major reduction in trade-through rates at securities dealers when they execute their customer orders internally.⁸² The great majority of internalized trades are the small trades of retail investors. The fact that, in 2003, nearly 1 of 30 of these millions of trades appears to have been executed at a price inferior to an automated and accessible quotation is troubling. Given that one of the primary benefits of the Order Protection Rule is to backstop a broker's duty of best execution on an order-by-order basis, Nasdaq's data appears to indicate a continuing need for regulatory action to reinforce the fundamental principle of best price for all NMS stocks.

Nasdaq also criticized the staff study for failing to address whether large block trades "intentionally avoid interacting with the posted quotes."⁸³ Far from demonstrating a flaw in the staff study, however, the fact that large trades intentionally avoid interacting with displayed quotations was one of the primary reasons identified in the Reproposing Release supporting the need for intermarket order protection.⁸⁴ The opportunity for displayed limit orders to begin interacting with this substantial volume of block trades is likely to be one of the most significant incentives for increased display of limit orders after implementation of the Order Protection Rule. Moreover, the Order Protection Rule will promote a more level playing field for retail investors that currently see their smaller displayed orders bypassed by block trades.

Two commenters did not believe the staff study should have included trades larger than quoted size, asserting that "[e]ven in a hard CLOB environment, orders larger than the inside quote would still 'trade through' the inside quote in effect at the time the order was received."⁸⁵ These commenters do not appear to have understood the methodology of the staff study or the operation of a central limit order book ("CLOB"). As discussed above, large trades would not have been identified as trade-throughs in the staff study if orders simultaneously had been routed to sweep displayed quotations with superior prices. To exclude such trades from its analysis, the study used a three-second quotation window in which the lowest best bid or the highest best offer during the three-second period must be traded-through before a trade was identified as a trade-through. The 3-second quotation window particularly was designed to allow sufficient time for quotations to update to reflect the arrival of sweep orders (just as in a CLOB environment, the execution of a large order simultaneously would eliminate all superior-priced quotations). In sum, large orders would trade with, rather than trade through, the superior-priced displayed quotations, thereby leaving only quotations that did not have superior prices to the trade price. Such large orders therefore would not have been identified as trade-throughs in the staff study.

Commenters also criticized the staff study for allegedly failing to consider the effect of locked or crossed quotations for Nasdaq stocks.⁸⁶ By using a 3-second quotation window, however, the staff study excluded any trade-throughs that would have been caused by short periods of locking or crossing quotations. The staff analysis appropriately did not exclude longer periods of locked quotations. Indeed, locked quotations do not qualify for an exception from the Order Protection Rule – both the best bid and best offer are readily accessible at the same price and should not be traded through. Quotations rarely are crossed for three seconds and therefore are unlikely to have caused a material number of false trade-throughs.⁸⁷

Finally, commenters asserted a variety of arguments relating to timing latencies in the quotation and trade data that might have caused the staff study to include false trade-throughs, including delayed trade reports, flickering quotations, stale quotations, manual quotations, and poor clock synchronization.⁸⁸ The staff study, however, used a variety of means to minimize the effect of these factors on the data, as well as to check for the extent to which timing latencies might affect its results. The goal of the staff study was to obtain a reasonable estimate of the true trade-through rates for Nasdaq and NYSE stocks. It is important to recognize that, in designing a methodology to achieve this goal, the more conservative the methodology used to eliminate potentially false indications of trade-throughs, the greater the number of true trade-throughs that are likely to be eliminated. Thus, a methodology designed simply to assure the elimination of every conceivable false indication of a trade-through would not have been useful to the Commission in assessing its policy options because it would have severely underestimated true trade-through rates. The staff study's conservative methodology was designed to produce reasonable estimates of true trade-through rates, but still is more likely to have resulted in an understatement of trade-through rates than an overstatement, particularly for Nasdaq stocks. Nasdaq stocks are traded primarily on automated markets, and the data for such stocks therefore should be less affected by timing latencies than the data for NYSE stocks, which is produced by both automated and manual markets.

For example, the staff study used a three-second quotation window for both Nasdaq and NYSE stocks to minimize the effect of possible timing lags between trade data and quotation data. Given that in Fall 2003 the overwhelming proportion of trades in Nasdaq stocks were executions of automated orders against automated quotations, with automated reporting of trades to the relevant Plan processor, three seconds is a conservative time frame to assess overall trade-through rates. But even when the quotation window is extended to an overly conservative eight seconds and thereby clearly excludes a large number of true trade-throughs, trade-through rates remain significant – 1.7% of trades and 6.8% of share volume in Nasdaq stocks.⁸⁹

In addition, the trade execution time derived from audit trail data for Nasdaq stocks, rather than trade report time, was used when it was supplied and whenever the two times differed to minimize timing latencies in the data caused by delayed reporting. Separate times derived from audit trail data are not reported for NYSE stocks, and delayed trade reports therefore could have contributed to false reports of trade-throughs in NYSE stocks. Similarly, for Nasdaq stocks, the quotations of Amex – the only market that displays manual quotations – were excluded from the staff study. Because the NYSE currently displays primarily manual quotations in NYSE stocks, while other markets display automated quotations, the difficulties of integrating data from manual and automated markets could have caused false indications of trade-throughs for NYSE stocks.⁹⁰ The occurrence of false indications of trade-throughs caused by manual quotations in exchange-listed stocks is addressed in the Order Protection Rule by protecting only automated quotations that are immediately accessible and immediately updated.

Fidelity incorrectly believed that the staff study failed to use the time of trade execution derived from audit trail data when analyzing trade-through rates in Nasdaq stocks.⁹¹ Fidelity also attached to its comment letter a paper prepared by two academics, Robert Battalio and Robert Jennings, which included a variety of criticisms of the staff study and the Reproposing Release in general ("Battalio/Jennings Paper").⁹² Among other things, the Battalio/Jennings Paper cited an academic paper which, for trading in Nasdaq stocks in 1996 and 1997, found significant delays between the time of trade execution reflected in proprietary trading center data and the time of trade report in public data disseminated by Nasdaq as Plan processor.⁹³ The authors of the Battalio/Jennings Paper, however, did not account for significant improvements in the quality of trade data for Nasdaq stocks since 1997. In particular, the NASD developed and implemented a new order audit trail system ("OATS").⁹⁴ As summarized in a 1998 NASD Notice to Members, OATS specifically was designed, among other things, to address the discrepancies between proprietary trade data and trade data reported to Nasdaq's Automated Confirmation Transaction Service ("ACT") :

OATS is designed to provide NASD Regulation, Inc. (NASD Regulation) with the ability to reconstruct markets promptly, conduct efficient surveillance, and enforce NASD and SEC rules. The SEC has directed that OATS must provide an accurate, time-sequenced record of orders and transactions from the receipt of an order through its execution. To accomplish this, NASD Regulation will combine information submitted to OATS with transaction data reported by members through ACT and quotation information disseminated by Nasdaq. . . . The ACT trade data and the OATS order information will be used to construct an integrated audit trail. Under the amended rules, all trade reports for OATS-eligible securities entered into Nasdaq's ACT system will be required to have a time of execution expressed in hours, minutes, and seconds.⁹⁵

To obtain the most accurate analysis of trade-through rates in Nasdaq stocks, the staff study used the audit trail record of the time of trade execution, rather than the time of trade report, whenever it was supplied and whenever the two times differed.⁹⁶ The Battalio/Jennings Paper therefore was mistaken when it stated that "[w]ith the data OEA chose to use, we simply cannot conclude anything about actual trade-through rates" and when it "urge[d] the OEA to revise their methodology and conduct a trade-through analysis using audit-trail data."⁹⁷ The staff study did indeed use audit trail data when available for Nasdaq stocks and therefore provides a reasonable basis for estimating true trade-through rates for Nasdaq stocks.

As noted above, however, the data for exchange-listed stocks may be more affected by timing latencies because it is generated by both automated and manual markets. The trade-through rates estimated in the staff study therefore may somewhat overstate the true trade-rates for NYSE stocks. Given that the ITS trade-through provisions currently apply to exchange-listed stocks, however, the Commission does not believe that the possibility that true trade-through rates potentially are lower than estimated in the staff study detracts from the strong policy reasons to maintain and strengthen trade-through protection for exchange-listed stocks. Rather, eliminating any trade-through protection for exchange-listed stocks could lead to rates that are as high, or higher, than were conservatively estimated for Nasdaq stocks, which have not been subject to any trade-through restrictions.

Moreover, the evidence from the staff study itself indicates that the concerns about delayed trade reporting discussed at length in the Battalio/Jennings Paper with respect to historical data have largely been resolved. For example, if delayed trade reporting were truly a serious problem that caused the staff study to be flawed, one would expect to see significant rates of trade-throughs by a single trading center's trades of its own quotations – the two data feeds would be out of synchronization with each other because trades were reported slower than quotation updates. In fact, however, the staff study found very low trade-through rates for single trading centers of their own quotations.⁹⁸ The primary exception is for trades reported on Nasdaq that trade through Nasdaq quotations, but Nasdaq, unlike the other major markets, does not consist of a single trading center. Rather, it includes the NASDAQ Market Center, several ECNs, and many market makers that trade, to a great extent, separately. Thus, the trade-through rates for Nasdaq reflect true trade-throughs among different trading centers, not false trade-throughs of a single trading center of its own quotations.

Finally, problems with clock synchronization at the various trading centers are unlikely to have materially detracted from the accuracy of the staff study. The great majority of time stamps were assigned to quotations and trades as the data was received by a single entity – Nasdaq as the Plan processor for Nasdaq stocks and SIAC as the Plan processor for NYSE stocks.⁹⁹ One commenter, however, asserted that the two Plan processors themselves had major clock synchronization problems between quotation data and trade data.¹⁰⁰ If this were in fact the case, the staff study likely would have found a high rate of trade-throughs by a single market of its own quotations, because the Plan processor's time stamps for the market's quotations would have been out of synchronization with its time stamps for the market's trades. As noted in the preceding paragraph, the staff study found few trade-throughs by a single market of its own quotations, thereby indicating that the Plan processors' quotation data and trade data are not materially out of synchronization.

ii. Significance of Trade-Through Rates

Some commenters questioned whether the trade-through rates found by the staff study were significant enough to warrant adoption of the trade-through reproposal.¹⁰¹ They believed, for example, that the rates were low, particularly when considered as a percentage of total trades (2.5% for both Nasdaq and NYSE stocks) and as the percentage of total share volume represented by the total displayed size of quotations that were traded through (1.9% and 1.2%, respectively, for Nasdaq and NYSE stocks).¹⁰² They therefore asserted that the rates did not demonstrate a serious problem or a need for regulatory action to address trade-throughs.

The Commission does not agree that the trade-through rates found in the staff study are insignificant, nor does it believe that the total number of trade-throughs is the sole consideration in evaluating the need for the Order Protection Rule. A valid assessment of their significance and the need for intermarket protection against trade-throughs must be made in light of the Exchange Act objectives for the NMS that would be furthered by the Order Protection Rule, including: (1) to promote best execution of customer market orders; (2) to promote fair and orderly treatment of customer limit orders; and (3) by strengthening protection of limit orders, to promote greater depth and liquidity for NMS stocks and thereby minimize investor transaction costs. The staff study examined trade-through rates from a variety of different perspectives, including percentage of trades, percentage of total share volume, percentage of share volume of trades of less than 10,000 shares, and percentage of total share volume of traded-through quotations.¹⁰³ In evaluating the need for the Order Protection Rule, the different measures vary in their relevance depending on the particular objective under consideration.

For example, the percentage of total trades that receive inferior prices is a particularly important measure when assessing the need to promote best execution of customer market orders. The staff study found that 1 of every 40 trades (2.5%) for both Nasdaq and NYSE stocks have an execution price that is inferior to the best displayed price, or approximately 98,000 trades per day in Nasdaq stocks alone.¹⁰⁴ As discussed above,¹⁰⁵ investors (and particularly retail investors) often may have difficulty monitoring whether their orders receive the best available prices, given the rapid movement of quotations in many NMS stocks. The Commission believes that furthering the interests of these investors in obtaining best execution on an order-by-order basis is a vitally important objective that warrants adoption of the Order Protection Rule.

The percentage of total trades that receive inferior prices also is quite relevant when assessing the need to promote fair and orderly treatment of limit orders for NMS stocks. Many of the limit orders that are bypassed are small orders that often will have been submitted by retail investors. One of the strengths of the U.S. equity markets and the NMS is that the trading interests of all types and sizes of investors are integrated, to the greatest extent possible, into a unified market system. Such integration ultimately works to benefit both retail and institutional investors. Retail investors will participate directly in the U.S. equity markets, however, only to the extent they perceive that their orders will be treated fairly and efficiently. The perception of unfairness created when a retail investor has displayed an order representing the best price for an NMS, yet sees that price bypassed by 1 in 40 trades, is a matter of a great concern to the Commission. The Order Protection Rule is needed to maintain the confidence of all types of investors that their orders will be treated fairly and efficiently in the NMS.

The third principal objective for the Order Protection Rule is to promote greater depth and liquidity for NMS stocks and thereby minimize investor transaction costs. Depth and liquidity will be increased only to the extent that limit order users are given greater incentives than currently exist to display a larger percentage of their trading interest. The potential upside in terms of greater incentives for display is most appropriately measured in terms of the share volume of trades that currently do not interact with displayed orders. It is this volume of trading interest that will begin interacting with displayed orders after implementation of the Order Protection Rule.

The share volume of trade-throughs, rather than the number of trade-throughs, is most useful for assessing the effect of the Order Protection Rule on depth and liquidity because very small trades represent such a large percentage of trades in today's markets, but a small percentage of share volume. For example, the staff study found that, for Nasdaq stocks, 100-share trades represented 32.7% of the number of trade-throughs, but only 0.8% of the share volume of trade-throughs.¹⁰⁶ Thus, the number of trade-throughs is useful for assessing the number of investors, particularly retail investors, affected by trade-throughs, while the share volume of trade-throughs is useful for assessing the extent to which depth and liquidity are affected by trade-throughs. For example, 41.1% of the share volume of trade-throughs in Nasdaq stocks is attributable to trades of greater than 1000 shares that bypass quotations of greater than 1000 shares.¹⁰⁷ Addressing the failure of this substantial volume of trading interest to interact with significant displayed quotations is a primary objective of the Order Protection Rule.

In contrast, the share volume of quotations that currently are traded through grossly underestimates the potential for increased incentives to display because it reflects only the current size of displayed quotations in the absence of strong price protection. As a result, the share volume of quotations that currently are traded through is a symptom of the problem that the Order Protection Rule is designed to address – a shortage of quoted depth – rather than an indication of the benefits that the Order Protection Rule will achieve. For example, when many Nasdaq stocks can trade millions of shares per day, but have average displayed size of less than 2000 shares at the NBBO, it will be nearly impossible for trade-throughs of displayed size to account for a large percentage of total share volume – there simply is not enough displayed depth.¹⁰⁸ Small displayed depth is evidence of a market problem, not market quality.

Every trade-through transaction in today's markets potentially sends a message to limit order users that their displayed quotations can be and are ignored by other market participants. The cumulative effect of such messages over time as trade-throughs routinely occur each trading day should not be underestimated. When the total share volume of trade-through transactions that do not interact with displayed quotations reaches 9% or more for many of the most actively traded Nasdaq stocks,¹⁰⁹ this message is unlikely to be missed by those who watched their quotations being traded through. Certainly, the routine practice of trading through displayed size is most unlikely to prompt market participants to display even greater size.

Thus, the Commission believes that the percentage of share volume in a stock that trades through displayed and accessible quotations is a useful measure for assessing the potential increase in incentives for display of limit orders after implementation of the Order Protection Rule. In particular, the dual measurements of percentage of share volume of traded-through quotations (an overall 1.9% for Nasdaq stocks) and the percentage of share volume of trades that bypass displayed quotations (an overall 7.9% for Nasdaq stocks) likely represent the lower and upper bounds for a potential improvement in depth and liquidity after implementation of the Order Protection Rule.

Commenters opposing the trade-through reproposal questioned whether protection against trade-throughs would lead to any increase in the use of limit orders, particularly given the many reasons militating against display (e.g., displayed limit orders give a free option to all other market participants to trade at the limit order price).¹¹⁰ The Commission is aware of a variety of reasons that currently deter market participants from displaying their trading interest in full. Indeed, it is the existence of these negative factors, combined with a shortage of positive incentives for display, that have contributed to the relatively small displayed depth at the best prices that characterizes the market for many NMS stocks today. A large investor interested in buying 50,000 shares of a stock is unlikely to suddenly decide to display all of its trading interest simply because its order is given trade-through protection. The objective for the Order Protection Rule is more modest. The Rule is designed to increase the perceived benefits of order display, against which the negatives are balanced. As a result, the market participant that currently displays only 500 shares of its 50,000-share trading interest might be willing to display 1000 shares. The collective effect of many market participants reaching the same conclusion would be a material increase in the total displayed depth in the market, thereby improving the transparency of price discovery and reducing investor transaction costs.

Moreover, because of the enormous volume of trading in NMS stocks, even a small percentage improvement in depth and liquidity could lead to very significant dollar benefits for investors in the form of reduced transaction costs. As discussed in section II.A.6 below, for example, the annual implicit transaction costs of large institutional investors are estimated at more than $30 billion in 2003.¹¹¹ As a result, even a small percentage reduction in these costs because of improved depth and liquidity would result in very substantial annual savings for millions of mutual fund and pension fund investors. The Commission therefore believes that the estimated trade-through rates in the staff study support the need for enhanced protection of limit orders as a means to promote greater depth and liquidity in NMS stocks.

b. Efficiency of Trading in Nasdaq and NYSE Stocks

A few commenters on the original proposal submitted empirical data to support their claim that trading in Nasdaq stocks currently is more efficient than trading in NYSE stocks.¹¹² Specifically, they submitted tables asserting that effective spreads in Nasdaq stocks in the S&P 500 are significantly narrower than effective spreads in NYSE stocks in the S&P 500.¹¹³ To help assess and respond to the views of commenters on market efficiency, the Commission staff analyzed Rule 11Ac1-5 reports and other trading data to evaluate the markets for Nasdaq and NYSE stocks.¹¹⁴

In its comment on the reproposal, Nasdaq argued that the staff studies contained flaws in their methodologies.¹¹⁵ With respect to the S&P Index Study, Nasdaq stated that the execution quality statistics were drawn from an atypical month and that the methodology for analyzing effective spreads favored higher-priced NYSE stocks over lower-priced Nasdaq stocks. The S&P Index Study presented statistics from January 2004, however, because this was the month selected by Nasdaq in the comment letter that it submitted on the proposal in July 2004. Moreover, the general statistics reported by Nasdaq for later months do not appear to differ materially from those for January 2004.¹¹⁶ In addition, the S&P Index Study analyzed investor transaction costs in terms of a percentage of investment rather cents per share because, as discussed below, the percentage of investment methodology most reflects economic reality for investors.¹¹⁷

With respect to the Matched Pairs Study, Nasdaq asserted that it largely examined small stocks. Nasdaq noted, for example, that more than 25% of the stocks included in the Matched Pairs Study were not eligible for NYSE listing and that only 10% of the stocks were included in the Nasdaq-100 Index. The purpose of the Matched Pairs Study, however, was to compare execution quality in Nasdaq and NYSE across a broad range of stocks, not solely for large stocks or those that were eligible for NYSE listing. Although 25% of the stocks may not have been eligible for NYSE listing, the staff analysis used matching criteria more directly designed to produce an "apples-to-apples" comparison – market capitalization, price, average daily dollar volume (adjusted downward by 30% for Nasdaq stocks to reflect trade reporting practices in such stocks), and relative price range. The Commission therefore believes that the staff studies provide a valid basis to compare trading in Nasdaq stocks and NYSE stocks.

The staff studies indicate that the execution quality statistics submitted by commenters on the original proposal are flawed. The claimed large and systematic disparities between Nasdaq and NYSE effective spreads disappear when an analysis of execution quality more appropriately controls for differences in stocks, order types, and order sizes.¹¹⁸ The staff studies reveal that both the market for Nasdaq stocks and the market for NYSE stocks have significant strengths. But, as discussed below, both markets also have weaknesses that could be reduced by strengthened protection against trade-throughs.

First, the effective spread analyses submitted by commenters do not, in a number of respects, reflect appropriately the comparative transaction costs in Nasdaq and NYSE stocks.¹¹⁹ They were presented in terms of "cents-per-share" and therefore failed to control for the varying level of stock prices between Nasdaq stocks and NYSE stocks in the S&P 500. Lower priced stocks naturally will tend to have lower spreads in terms of cents-per-share than higher priced stocks, even when such cents-per-share spreads constitute a larger percentage of stock price and therefore represent transaction costs for investors that consume a larger percentage of their investment. By using cents-per-share statistics, commenters did not adjust for the fact that the average prices of Nasdaq stocks are significantly lower than the average prices of NYSE stocks. For example, the average price of Nasdaq stocks in the S&P 500 in January 2004 was $34.14, while the average price of NYSE stocks was $41.32.¹²⁰

The effective spread analyses submitted by commenters also were weakened by their failure to address the much lower fill rates of orders in Nasdaq stocks than orders in NYSE stocks. The commenters submitted "blended" statistics that encompassed both market orders and marketable limit orders. The effective spread statistics for these order types are not comparable, however, because market orders do not have a limit price that precludes their execution at prices inferior to the prevailing market price at time of order receipt. In contrast, the limit price of marketable limit orders often precludes an execution, particularly when there is a lack of liquidity and depth at the prevailing market price. For example, the fill rates for marketable limit orders in Nasdaq stocks generally are less than 75%, and often fall below 50% for larger order sizes.¹²¹

Accordingly, investors must accept trade-offs when deciding whether to submit market orders or marketable limit orders (particularly when the limit price equals the current market price). Use of a limit price generally assures a narrower spread by precluding an execution at an inferior price. By precluding an execution, however, the limit price may cause the investor to "miss the market" if prices move away (for example, if prices rise when an investor is attempting to buy). Effective spreads for marketable limit orders therefore represent transaction costs that are conditional on execution, while effective spreads for market orders much more completely reflect the entire implicit transaction cost for a particular order. Market orders represent only approximately 14% of the blended flow of market and marketable limit orders in Nasdaq stocks (reflecting the fact that ECNs now dominate Nasdaq order flow and limit orders represent the vast majority of ECN order flow).¹²² In contrast, market orders represent approximately 36% of the blended order flow in NYSE stocks.¹²³ Accordingly, the effective spread statistics for marketable limit orders, and particularly for orders in Nasdaq stocks, must be considered in conjunction with the fill rate for such orders – while a narrow spread is good, the benefits are greatly limited if investors are unable to obtain an execution at that spread. The analyses presented by the commenters, however, did not address the respective fill rates for Nasdaq stocks and NYSE stocks or reflect the inherent differences in measuring the transaction costs of market orders and marketable limit orders.

The analyses prepared by Commission staff are designed to provide appropriate evaluations of comments on the efficiency of trading in Nasdaq and NYSE stocks. In particular, they are more finely tuned to evaluate trading for different types of stocks with varying trading volume, different types of orders, and different sizes of orders. These analyses indicate that the markets for Nasdaq and NYSE stocks each have weaknesses that an intermarket price protection rule could help address. By "weakness," the Commission simply means that there appears to be considerable room for improvement. For example, the effective spread statistics for large, electronically-received market orders in NYSE stocks show significant "slippage" – the amount by which orders are executed at prices inferior to the national best bid or offer ("NBBO") at the time of order receipt.¹²⁴ Slippage often results in effective spreads for large orders that are many times wider than the effective spreads for small orders in the same NYSE stocks. By protecting automated quotations, the Order Protection Rule should enhance the depth and liquidity available for large, electronic orders in NYSE stocks and thereby improve their execution quality.

For Nasdaq stocks, the Rule 11Ac1-5 statistics reveal very low fill rates for larger sizes of marketable limit orders (e.g., 2000 shares or more), which generally fall below 50% for most Nasdaq stocks. Contrary to the assertion of some commenters,¹²⁵ certainty of execution for large marketable limit orders clearly is not a strength of the current market for Nasdaq stocks. Certainty of a fast response is a strength, but much of the time the response to large orders will be a "no fill" at any given trading center.¹²⁶

Two commenters on the reproposal disputed whether low fill rates for marketable limit orders in Nasdaq stocks indicate any weakness that needed to be addressed.¹²⁷ Instinet, for example, believed that "the Commission is misplaced in its contention that low fill rates in Nasdaq stocks are a weakness of that market," and that they are a phenomenon "intrinsic to electronic markets in which market participants are free to cancel and replace orders."¹²⁸ Instinet also noted that many market centers in Nasdaq stocks have significant reserve size in addition to displayed size and that market participants commonly routed oversized marketable limit orders to attempt to interact with reserve size.¹²⁹ Similarly, Nasdaq stated that the staff studies "erroneously conclude that differential fill rates for large marketable limit orders in Nasdaq-listed and NYSE-listed stocks are evidence of a defect in Nasdaq's market structure," and that they failed "to consider a widely used order routing technique of intentionally sending oversized orders at displayed quotes searching (also known as "pinging") for reserves within the many limit order books trading Nasdaq-listed securities."¹³⁰ Nasdaq also asserted that marketable limit orders are "exceedingly popular in electronic venues where they have effectively supplanted market orders as the order of choice in accessing availability liquidity at the current price."¹³¹

The Commission continues to believe that fill rates for large marketable limit orders are a useful measure of order execution quality for Nasdaq stocks. They are especially useful because they measure the availability of both displayed and undisplayed liquidity, whereas simply measuring displayed size would understate the total liquidity readily available for Nasdaq stocks. Indeed, the existence of "pinging" orders searching for reserve size in Nasdaq stocks at electronic markets is widely known. Such oversized orders (i.e., orders with sizes greater than displayed size) could as aptly be labeled "liquidity search" orders as "pinging" orders. Given the relatively small displayed size in nearly all Nasdaq stocks (i.e., significantly less than 2000 shares),¹³² orders with sizes of 2000 to 4999 shares and 5000 to 9999 shares (the two largest Dash 5 size categories) generally will exceed the displayed size. Thus, low fill rates demonstrate that the total displayed and reserve liquidity available for Nasdaq stocks at any particular trading center typically is small compared to the demand for liquidity at the inside prices. Moreover, increased displayed liquidity – a principal goal of the Order Protection Rule – would promote market efficiency by reducing the uncertainty and costs associated with the need for market participants to "ping" electronic markets for liquidity that is held in reserve.

The Reproposing Release did not suggest, however, that the differential fill rates for large marketable limit orders in Nasdaq and NYSE stocks were useful in comparing the liquidity and depth available in each market. Instead, the Reproposing Release focused on the most relevant Dash 5 statistic for each market, given its particular trading characteristics. As noted above, the significant amount of "slippage" in the execution of electronically-received large market orders in NYSE stocks suggest that improved incentives for display of automated trading interest will help improve execution quality for NYSE stocks. Notably, Instinet and Nasdaq agreed that slippage rates for automated market orders represented a problem in the market for NYSE stocks.¹³³ Because market participants generally choose not to submit market orders to electronic markets in Nasdaq stocks, however, the fill rates for marketable limit orders are a more relevant Dash 5 statistic to assess depth and liquidity in Nasdaq stocks.

Accordingly, the Commission's concern with fill rates for larger orders in Nasdaq stocks is not that they are lower than those for NYSE stocks, but that they are very low in absolute terms – often falling well below 50%.¹³⁴ Moreover, the larger order sizes typically account for a small percentage of executed shares compared to the executed shares of smaller order sizes.¹³⁵ When considered in conjunction with one another, the low fill rates and small percentage of executed shares indicate substantial room for improvement in depth and liquidity in many Nasdaq stocks. An important objective for Regulation NMS as a whole is to facilitate more efficient trading in larger sizes, an objective that has become much more important to large investors since decimalization.¹³⁶ An improvement in fill rates for larger sized orders (or an increase in their percentage of executed shares) would evidence progress toward this objective.

Fill rates for marketable limit orders, however, offer only indirect evidence of the total transaction costs incurred by investors. They indicate that no execution was obtained for an investor order at a particular trading center, but do not indicate how the investor subsequently fared in obtaining an execution. As discussed above, there are significant trade-offs between marketable limit orders and market orders. The use of a restrictive limit price at the NBBO precludes any slippage in execution price, but also may cause an investor to miss the market if prices subsequently move away from the order (i.e., rise when an investor is attempting to buy or fall when an investor is attempting to sell). To evaluate the total transaction costs associated with an order that goes unfilled or receives a partial fill, it is necessary to know the price at which the investor ultimately obtained an execution for its full order.

To help the Commission evaluate and respond to commenters' criticisms and, in particular, to supplement its analysis of fill rates as a measure of depth and liquidity for Nasdaq stocks and to evaluate the extent to which missed fills may lead to higher investor transaction costs, Commission staff also examined execution quality statistics for marketable limit orders in Nasdaq-100 Index stocks that are executed outside the best quotes at the Inet ATS and the NASDAQ Market Center.¹³⁷ By definition, such orders have been placed with liberal limit prices that give more flexibility for executions away from the NBBO than orders with limit prices that are restrictively set at the NBBO. Accordingly, the slippage rates for such orders give another indication of available liquidity for Nasdaq-100 stocks.

The statistics for outside-the-quote executions in marketable limit orders buttress a conclusion that there is significant room for improved depth and liquidity in Nasdaq stocks. For example, the Inet ATS did not fill 83.0% of its large marketable limit orders.¹³⁸ Of the orders it executed, 19.5% of shares were executed outside the quote by an average of 2.7 cents. Thus, while the overall quoted and effective spreads for executed shares for large orders were, respectively, 1.6 cents and 2.5 cents, the spread for outside the quote executions was 7.0 cents – 438% wider than the narrow quoted spread. The statistics for the NASDAQ Market Center are similar. It did not fill 68.4% of its large marketable limit orders.¹³⁹ Of the orders it executed, 14.7% were executed outside the quote by an average of 2.3 cents. The overall quoted and effective spreads for large orders were, respectively, 1.6 cents and 2.5 cents, compared to 6.2 cents for outside the quote executions – 388% wider than the narrow quoted spread. The outside-the-quote spreads provide the best available indication of execution quality that otherwise would have been obtained at the time orders were placed for the 83.0% and 68.4% of shares that were not filled due to their restrictive limit price. The outside-the-quote spreads also are relevant in assessing the reasons why market participants most often use marketable limit orders with limit prices at the NBBO rather than market orders when trading Nasdaq stocks.

In addition, the supplemental staff study separately examined fill rates and executed share volume for types of Nasdaq-100 stocks where liquidity for orders with large share sizes can reasonably be expected to be highest.¹⁴⁰ These stock groupings were selected primarily to assess whether low fill rates for large marketable limit orders are an inherent part of the structure of the market for Nasdaq stocks. Specifically, the supplemental staff study calculated fill rates and executed share volume for the three Nasdaq stocks with the largest capitalization – Microsoft, Intel, and Cisco. These three stocks are widely recognized among all Nasdaq stocks as having markets with significant depth and liquidity. In addition, the supplemental staff study examined the seven Nasdaq-100 stocks with share prices of less than $10 per share. Liquidity for orders with large share sizes in these stocks can be expected to be higher than for stocks with higher prices because the dollar sizes are much smaller (e.g., a 5000 share order in a $5 stock totals $25,000, whereas a 5000 share order in a $30 stock totals $150,000). In terms of economic reality, therefore, large orders in a low-priced stock generally are easier to execute than large orders in a higher-priced stock, assuming the stocks are otherwise comparable. Finally, the supplemental staff study separately examined the other 90 stocks in the Nasdaq-100 Index (i.e., stocks with prices of at least $10 per share other than Microsoft, Intel, and Cisco).

The supplemental staff study reveals that low fill rates for large marketable limit orders are not an inherent feature of the market for Nasdaq stocks. For example, the NASDAQ Market Center fill rates for large orders are 76.7% for the three large-cap stocks, 70.1% for the low-priced stocks, and 27.1% for the other 90 stocks in the Nasdaq-100 Index.¹⁴¹ Similarly, the Inet ATS fill rates for large orders are 58.5% for the three large-cap stocks, 55.0% for low-priced stocks, and 12.6% for the other 90 stocks in the Nasdaq-100 Index.¹⁴²

The order execution quality measures included in Dash 5 reports do not, of course, reflect all types of investor transaction costs. They generally focus on the execution price of individual orders in comparison with the best quoted prices at the time orders are received. As a result, they do not capture transaction costs that are associated with the short-term movement of quoted prices. To further assist the Commission in evaluating the views of commenters, Commission staff has analyzed price volatility for trading in Nasdaq and NYSE stocks.¹⁴³ This analysis particularly focuses on transitory volatility – short-term fluctuations away from the fundamental or "true" value of a stock. Transitory volatility should be distinguished from fundamental volatility – price fluctuations associated with factors independent of market structure, such as earnings changes and other economic determinants of stock prices. The staff analysis found that on average both intraday volatility and transitory volatility are higher for Nasdaq stocks than for NYSE stocks.¹⁴⁴ Excessive transitory volatility indicates a shortage of depth and liquidity that otherwise would minimize the effect of short-term order imbalances. Such volatility may provide benefits in the form of profitable trading opportunities for short-term traders or market makers, but these benefits come at the expense of other investors, who would be buying at artificially high or selling at artificially low prices. Retail investors, in particular, tend to be relatively uninformed concerning short-term price movements and are apt to bear the brunt of the trading costs associated with excessive transitory volatility.¹⁴⁵ The Order Protection Rule, by promoting greater depth and liquidity, is designed to help reduce excessive transitory volatility in Nasdaq stocks.

Finally, an important measure of depth and liquidity for NMS stocks is the transaction costs actually incurred by institutional investors when they trade in large size. These costs are not readily available for public view because their measurement requires access to a large volume of private order and execution data of institutional investors. One of the leading authorities on institutional transaction costs uses an extensive database of such data obtained from its clients to calculate their transaction costs. It recently published calculations of average transaction costs for Nasdaq and NYSE stocks during the fourth quarter of 2003 as, respectively, 83 basis points and 55 basis points.¹⁴⁶ Given the significant differences in the overall nature of Nasdaq and NYSE stocks, these figures cannot be used to assess the relative efficiency of the two markets. The figures for both, however, suggest room for improved depth and liquidity, particularly when compared with the average quoted spreads in NMS stocks, which generally are less, and often much less, than 10 basis points for large capitalization stocks that dominate trading volume.¹⁴⁷

c. Need for Intermarket Rule to Achieve Effective Protection Against Trade-Throughs

As discussed in the preceding section, the relevant data, as well as the policy choices the Commission has articulated above, supports the need for strengthened protection against trade-throughs in both Nasdaq and exchange-listed stocks. Some commenters argued, however, that competitive forces alone would achieve the fairest and most efficient markets.¹⁴⁸ In particular, they asserted that reliance on efficient access to markets and brokers' duty of best execution would be sufficient without the need for an intermarket rule against trade-throughs. This argument, however, fails to take into account two structural problems – principal/agent conflicts of interest and “free-riding” on displayed prices.

Agency conflicts may occur when brokers have incentives to act otherwise than in the best interest of their customers. For example, brokers may have strong financial and other interests in routing orders to a particular market, which may or may not be displaying the best price for a stock. Moreover, the Commission has not interpreted a broker's duty of best execution for retail orders as requiring that a separate best execution analysis be made on an order-by-order basis.¹⁴⁹ Nevertheless, retail investors generally expect that their small orders will be executed at the best displayed prices. They may have difficulty monitoring whether their individual orders miss the best displayed prices at the time they are executed and evaluating the quality of service provided by their brokers.¹⁵⁰ Given the large number of trades that fail to obtain the best displayed prices (e.g., approximately 1 in 40 trades for both Nasdaq and NYSE stocks), the Commission is concerned that many of the investors that ultimately received the inferior price in these trades may not be aware that their orders did not, in fact, obtain the best price. The Order Protection Rule will backstop a broker's duty of best execution on an order-by-order basis by prohibiting the practice of executing orders at inferior prices, absent an applicable exception.

Just as importantly, even when market participants act in their own economic self-interest, or brokers act in the best interests of their customers, they may deliberately choose, for various reasons, to bypass (i.e., not protect) limit orders with the best displayed prices. For example, an institution may be willing to accept a dealer's execution of a particular block order at a price outside the NBBO, thereby transferring the risk of any further price impact to the dealer. Market participants that execute orders at inferior prices without protecting displayed limit orders are effectively “free-riding” on the price discovery provided by those limit orders. Displayed limit orders benefit all market participants by establishing the best prices, but, when bypassed, do not themselves receive a benefit, in the form of an execution, for providing this public good. This economic externality, in turn, creates a disincentive for investors to display limit orders and ultimately could negatively affect price discovery and market depth and liquidity.

Fidelity's comment letters on the reproposal questioned whether large trades that bypass displayed quotations should be considered as free-riding on the price discovery provided by displayed limit orders.¹⁵¹ It emphasized that the price-formation process reflects information stemming from all trading interest and that institutional trading interest is an important part of the process. As evidence, it noted that almost one-third of reported volume on the NYSE in 2004 was of block size, typically representing undisplayed institutional trading interest.

Institutional trading interest, both displayed and undisplayed, undoubtedly is an important part of the price discovery process. Notably, the large volume of block trades currently executed on the NYSE is subject both to the NYSE's order interaction rules and the ITS trade-through rules. Accordingly, NYSE block trades cannot be considered as free-riding on displayed limit orders, in contrast to block trades reported by block positioners in the OTC market that currently do not interact with (and thereby are free-riding on) displayed liquidity and are not covered by the ITS provisions.

Moreover, the Order Protection Rule does not require that all institutional trading interest be displayed. Rather, the Rule strengthens the incentive for the voluntary display of a greater proportion of latent trading interest by assuring that, when such interest is displayed, it is protected against most trade-throughs. In these circumstances, institutions will choose to display when they determine it is in their own interests, not because it is mandated by Commission rule. Greater displayed size will improve the quality and transparency of price discovery for all market participants.

Fidelity also asserted that "an institutional investor, seeking to acquire or dispose a large block of stock will be put to a distinct and unfair advantage if it is deprived of the ability to negotiate, at one time and at a specified price, an all-in price for its block trade with a dealer."¹⁵² Similarly, the Battalio/Jennings Paper suggests that, for large marketable limit orders of institutions, "it might be better to ignore a penny quote for a few hundred shares in order to get a large order done quickly rather than try to chase the small quote and risk losing the ability to fill the size desired."¹⁵³ These contentions do not recognize that the Order Protection Rule does not, in fact, preclude institutions from negotiating "all-in" prices for their trades with dealers or immediately routing orders to access larger-sized depth-of-book quotations. Rather, the Rule simply requires a dealer, at the same time as executing a large institutional order at an all-in price, to route an intermarket sweep order to execute against the displayed size of protected quotations with superior prices to the institution's trade price. Similarly, the Rule allows an institution to simultaneously route intermarket sweep orders to execute against both small-sized quotations at the best prices and larger-sized depth-of-book quotations. The Rule therefore does not require institutions to parcel out their block orders in a series of transactions over time.

Fidelity and the Battalio/Jennings Paper also incorrectly asserted that the Commission's concern about free-riding on displayed quotations related only to the limit orders of retail investors, citing a number of academic studies indicating that institutional trades and quotations are important contributors to price discovery.¹⁵⁴ In fact, however, the Reproposing Release did not distinguish between the limit orders of retail investors and those of institutions when discussing the problem of free-riding.¹⁵⁵ Rather, the Order Protection Rule is designed to promote displayed liquidity from all sources, and institutional limit orders clearly are a significant source of such liquidity. Indeed, the Battalio/Jennings Paper itself notes that "institutions dominate price discovery via quoting" and that "the preponderance of quote-based discovery for NYSE-listed securities takes place at the NYSE" where "institutions dominate trading."¹⁵⁶ Many institutional investors and the NYSE are strong supporters of strengthened limit order protection for all NMS stocks.¹⁵⁷ For example, the ICI, whose members manage assets that account for more than 95% of assets of all U.S. mutual funds, stated that it "strongly supports the establishment of a marketwide trade-through rule. . . . [S]uch a rule represents a significant step in providing protection for limit orders. By affirming the principle of price priority, a trade-through rule should encourage the display of limit orders, which in turn would improve the price discovery process and contribute to increased market depth and liquidity."¹⁵⁸

Another commenter asserted that the reproposal overly emphasized the importance of displayed limit orders in the price discovery process.¹⁵⁹ It stated that the interaction of displayed limit orders with marketable orders is only one aspect of price discovery, which is "a dynamic process that operates in the context of other transactions that have recently been made, current quotes, and a richer tapestry of the expressed and latent interest of a broader array of market participants."¹⁶⁰ The Commission generally concurs with this characterization of the price discovery process, but believes that displayed limit orders are a critically important element of efficient price discovery that deserve greater protection against trade-throughs. Publicly displayed and automated limit orders are the most transparent and accessible source of liquidity in the equity markets. Moreover, displayed limit orders provide price discovery on a going forward basis – they indicate the prices at which trades can be effected in the future. Trade reports, in contrast, look backward at the prices of trades that already have occurred, which may or may not be still be available.

There are, of course, other sources of liquidity, including: (1) reserve size (limit orders with undisplayed size); (2) "not held" institutional orders that are worked by floor brokers on an exchange; (3) automated matching networks that allow large buyers and sellers to meet directly and anonymously; and (4) securities dealers that are willing to commit capital to facilitate customer orders. Displayed limit orders, however, give anyone the ability to trade when they want to trade on a first-come, first-served basis. They thereby act as a vital reference point for all other sources of liquidity. Specifically, reserve size, undisplayed floor interest, automated matching, and dealer capital commitments all are facilitated by displayed information concerning the price and size of stock that is available for immediate trading in the public markets.

As demonstrated by the current rate of trade-throughs of the best quotations in Nasdaq and NYSE stocks, the problems of agent/principal conflicts and the free-riding externality often can lead to executions at prices that are inferior to displayed quotations, meaning that limit orders are being bypassed. The frequent bypassing of limit orders can cause fewer limit orders to be placed. The Commission therefore believes that the Order Protection Rule is needed to encourage greater use of limit orders. The more limit orders available at better prices and greater size, the more liquidity available to fill incoming marketable orders. Moreover, greater displayed liquidity will at least lower the search costs associated with trying to find liquidity. Increased liquidity, in turn, could lead market participants to interact more often with displayed orders, which would lead to greater use of limit orders, and thus begin the cycle again. We expect that the end result will be an NMS that more fully meets the needs of a broad spectrum of investors.

2. Limiting Protection to Automated and Accessible Quotations

The original trade-through proposal sought to strengthen protection against trade-throughs, while also addressing problems posed by the inherent differences in quotations displayed by automated markets (which are immediately accessible) and quotations displayed by manual markets (which are not), by distinguishing between automated and non-automated markets with respect to trade-through protection. The proposal included an exception that would have allowed automated markets to trade through manual markets, but only up to certain amounts that varied depending upon the price of the security. Under the proposal, a market would have been classified as "manual" if it did not provide for an immediate automated response to all incoming orders attempting to access its displayed quotations.¹⁶¹

At the NMS Hearing, a significant portion of the discussion of the trade-through proposal addressed issues relating to quotations of automated and manual markets. Representatives of two floor-based exchanges announced their intent to establish "hybrid" trading facilities that would offer automatic execution of orders seeking to interact with their displayed quotations, while at the same time maintaining a traditional floor.¹⁶² These representatives acknowledged the difficulties posed in developing an efficient hybrid market, but emphasized that they were committed to developing such facilities and that such facilities were likely to become operational prior to any implementation of Regulation NMS.

Other panelists at the NMS Hearing strongly believed that manual quotations should not receive any protection against trade-throughs and that the proposed trade-through amounts should be eliminated.¹⁶³ They noted, however, that existing order routing technologies are capable of identifying, on a quote-by-quote basis, indications from a market that a particular quotation is not immediately and automatically accessible (i.e., is a manual quotation). Using this functionality, a trade-through rule could classify individual quotations as automated or manual, rather than classifying an entire market as manual solely because it displayed manual quotations on occasion.

To give the public a full opportunity to comment on these issues, the Supplemental Release described the developments at the NMS Hearing and requested comment on whether a trade-through rule should protect only automated quotations and whether the rule should adopt a "quote-by-quote" approach to identifying protected quotations.¹⁶⁴ The Supplemental Release also requested comment on the requirements for an automated quotation, including whether the rule should impose a maximum response time, such as one second, on the total time for a market to respond to an order in an automated manner. Comment also was requested on mechanisms for enforcing compliance with the automated quotation requirements.

Nearly all commenters on the original proposal believed that only automated quotations should receive protection against trade-throughs and that therefore the proposed limitation on trade-through amounts for manual markets should be eliminated.¹⁶⁵ In response to these commenters, the Commission modified the proposed Rule in the Reproposing Release to protect only those quotations that are immediately and automatically accessible. As noted above in Section II.A.1, a substantial number of commenters supported the reproposed Order Protection Rule, with some commenters specifically supporting limiting trade-through protection to automated and immediately accessible quotations.¹⁶⁶

The Commission agrees with commenters that providing protection to manual quotations, even limited to trade-throughs beyond a certain amount, potentially would lead to undue delays in the routing of investor orders, thereby not justifying the benefits of price protection. The Commission therefore is adopting, as reproposed, an approach that excludes manual quotations from trade-through protection. Under the Order Protection Rule as adopted, investors will have the choice of whether to access a manual quotation and wait for a response or to access an automated quotation with an inferior price and obtain an immediate response. Moreover, those who route limit orders will be able to control whether their orders are protected by evaluating the extent to which various trading centers display automated versus manual quotations.

Commenters expressed differing views, however, on the appropriate standards for automated quotations and on the standards that should govern "hybrid" markets – those that display both automated and manual quotations. These issues are discussed below.

a. Standards for Automated Quotations

Nearly all commenters addressing the issue believed that only quotations that are truly firm and fully accessible should qualify as "automated."¹⁶⁷ To achieve this goal, they suggested that, at a minimum, the market displaying an automated quotation should be required to provide a functionality for an incoming order to receive an immediate and automated (i.e., without human intervention) execution up to the full displayed size of the quotation. In addition, they believed the market should be required to provide an immediate and automated response to the sender of the order indicating whether the order had been executed (in full or in part) and an immediate and automated updating of the quotation. A number of commenters advocated requiring a specific time standard for distinguishing between manual and automated quotations, ranging from one second down to 250 milliseconds.¹⁶⁸ Other commenters did not believe the definition of automated quotation should require a specific time standard, generally because setting a specific standard might discourage innovation and become a “ceiling” on market performance.¹⁶⁹

The Commission included in the Reproposing Release a definition of automated quotation that incorporated the three elements suggested by commenters:¹⁷⁰ (1) acting on an incoming order; (2) responding to the sender of the order; and (3) updating the quotation. The proposed definition of automated quotation did not set forth a specific time standard for responding to an incoming order. As noted above, a significant number of commenters on the Reproposing Release supported the reproposed Order Protection Rule,¹⁷¹ with a few commenters specifically supporting the definition of automated quotation.¹⁷² As discussed in detail below, the Commission has adopted the definition of automated quotation as proposed.

In particular, Rule 600(b)(3) requires that the trading center displaying an automated quotation must provide an "immediate-or-cancel" ("IOC") functionality for an incoming order to execute immediately and automatically against the quotation up to its full size, and for any unexecuted portion of such incoming order to be cancelled immediately and automatically without being routed elsewhere. The trading center also must immediately and automatically respond to the sender of an IOC order. To qualify as "automatic," no human discretion in determining any action taken with respect to an order may be exercised after the time an order is received. Trading centers are required to offer this IOC functionality only to market participants that request immediate action and response by submitting an IOC order. Market participants therefore have the choice of whether to require an immediate response from the trading center, or to allow the market to take further action on the order (such as by routing the order elsewhere, seeking additional liquidity for the order, or displaying the order). Finally, trading centers are required to immediately and automatically update their automated quotations to reflect any change to their material terms (such as a change in price, displayed size, or "automated" status).

The definition of automated quotation as adopted does not set forth a specific time standard for responding to an incoming order. The Commission agrees with commenters that the standard should be "immediate" – i.e., a trading center's systems should provide the fastest response possible without any programmed delay. Nevertheless, the Commission also is concerned that trading centers with well-functioning systems should not be unnecessarily slowed down waiting for responses from a trading center that is experiencing a systems problem. Consequently, rather than specifying a specific time standard that may become obsolete as systems improve over time, Rule 611(b)(1) addresses the problem of slow trading centers by providing an exception for quotations displayed by trading centers that are experiencing, among other things, a material delay in responding to incoming orders. Given current industry conditions, the Commission believes that repeatedly failing to respond within one second after receipt of an order would constitute a material delay.¹⁷³ Accordingly, a trading center would act reasonably in the current technological environment if it bypassed the quotations of another trading center that had repeatedly failed to respond to orders within a one-second time frame (after adjusting for any potential delays in transmission not attributable to the other trading center).¹⁷⁴ This "self-help" remedy, discussed further in sections II.A.3 and II.B.3 below, will give trading centers needed flexibility to deal with another trading center that is experiencing systems problems, rather than forcing smoothly-functioning trading centers to slow down for a problem trading center.

b. Standards for Automated Trading Centers

The original trade-through proposal would have classified a market as manual if it did not provide automated access to all orders seeking access to its displayed quotations. Many commenters responded positively to the concept of allowing hybrid markets to display both automated and manual quotations that was raised at the NMS Hearing and discussed in the Supplemental Release. Most national securities exchanges believed that focusing on whether individual quotations are automated or manual would permit hybrid markets to function, thereby expanding the range of trading choices for investors.¹⁷⁵ For example, Amex stated that hybrid markets would offer investors the choice to utilize auction markets when advantageous for them to do so, while at the same time offering automatic execution to those investors desiring speed and certainty of a fast response.¹⁷⁶ A majority of other commenters also believed that the application of any trade-through rule should depend on whether a particular quotation is automated.¹⁷⁷ They believed that such a rule would achieve the benefits of encouraging limit orders and improving market depth and liquidity, while avoiding indirectly mandating a particular market structure.

Although generally supportive of the concept of hybrid markets, several commenters on the original proposal expressed concern about how the "quote-by-quote" approach to protected quotations would operate in practice.¹⁷⁸ The ICI noted that "[w]e are concerned that if it is left completely up to an individual market’s discretion when a quote is 'automated' or manual, that market could base its decision on what is in the best interests of that market and its members, as opposed to the best interests of investors and other market participants."¹⁷⁹ These commenters suggested that the Commission should provide clear guidelines as to when and how a market could switch its quotations from automated to manual, and vice versa, so as to prevent abuse by the market.

After considering the views of commenters, the Commission included in the reproposed Rule certain requirements for a trading center to qualify as an "automated trading center," one of which requires that a trading center adopt reasonable standards limiting when its quotations change from automated quotations to manual quotations (and vice versa) to specifically defined circumstances that promote fair and efficient access to its automated quotations and that are consistent with the maintenance of fair and orderly markets. The reproposed Rule also provided that only a trading center that met all of the requirements could display protected quotations. Although a substantial number of commenters supported the reproposed Rule,¹⁸⁰ a few commenters continued to express concern with the ability of a trading center to switch from automated to manual quotations.¹⁸¹

The Commission recognizes the concerns of commenters regarding the ability of a trading center to change from automated to manual quotation mode, but believes that the requirements necessary to qualify as an automated trading center will sufficiently mitigate this concern. Any standards established by an SRO trading center to govern when its quotations change from automated to manual will be subject to public notice and comment and Commission approval pursuant to the rule filing process of Section 19(b) of the Exchange Act. If a non-SRO trading center intends to display both automated and non-automated quotations, it will be subject to the oversight of the SRO through whose facilities its quotations are displayed with respect to the reasonableness of its procedures, as well as Commission oversight.

The Commission therefore is adopting the definition of automated trading center as reproposed. The adopted approach offers flexibility for a hybrid market to display both automated and manual quotations, but only when such a market meets basic standards that promote fair and efficient access by the public to the market's automated quotations. This approach is designed to allow markets to offer a variety of trading choices to investors, but without requiring other markets and market participants to route orders to a hybrid market with quotations that are not truly accessible.

To qualify as an automated trading center, the trading center must have implemented such systems, procedures, and rules as are necessary to render it capable of displaying quotations that meet the action, response, and updating requirements set forth in the definition of an automated quotation.¹⁸² Further, the trading center must identify all quotations other than automated quotations as manual quotations, and must immediately identify its quotations as manual quotations whenever it has reason to believe that it is not capable of displaying automated quotations.¹⁸³ These requirements will enable other trading centers readily to determine whether a particular quotation displayed by a hybrid trading center is protected by the Order Protection Rule. Finally, an automated trading center must adopt reasonable standards limiting when its quotations change from automated quotations to manual quotations, and vice versa, to specifically defined circumstances that promote fair and efficient access to its automated quotations and are consistent with the maintenance of fair and orderly markets.¹⁸⁴

These requirements are designed to promote efficient interaction between a hybrid market and other trading centers. The requirement that automated quotations cannot be switched on and off except in specifically defined circumstances is particularly intended to assure that hybrid markets do not give their members, or anyone else, overbroad discretion to control the automated or manual status of the trading center's quotations, which potentially could disadvantage market participants that must protect these quotations. Changes from automated to manual quotations, and vice versa, must be subject to specific, enforceable limitations as to the timing of switches. For a trading center to qualify as entitled to display any protected quotations, the public in general must have fair and efficient access to a trading center's quotations.

Some commenters on the Reproposing Release expressed a concern about the scope of the exception for single-priced reopenings in Rule 611(b)(3), particularly in the context of a trading center switching back and forth from automated quotation to manual quotation mode.¹⁸⁵ They asserted that the applicability of the exception to the recommencement of trading after a non-regulatory trading halt in one market (such as a trading halt due to an intra-day order imbalance) could lead to disruptive trading activity and provide an unfair competitive advantage for the trading center that halted trading. They believed this could create a significant loophole in the protections provided by the Rule. For instance, one commenter expressed concern that a trading center could halt trading and reopen solely to enable it to trade-through other trading centers.¹⁸⁶ Another commenter expressed concern regarding the interplay of the proposed exception and the operation of the NYSE’s proposed hybrid trading system, stating that it is unclear what would be considered a reopening under NYSE’s proposal, particularly with respect to when a liquidity refreshment point is reached or when the quotation is gapped.¹⁸⁷ Two commenters suggested that the exception apply only to reopenings after regulatory trading halts.¹⁸⁸

The Commission recognizes the commenters' concern, but emphasizes that the exception will not permit a trading center to declare a trading halt merely to be able to circumvent the operation of the Order Protection Rule upon reopening. The exception applies only to single-priced reopenings and therefore requires that a trading center conduct, pursuant to its rules or written procedures, a formalized and transparent process for executing orders during reopening after a trading halt that involves the queuing and ultimate execution of multiple orders at a single equilibrium price.¹⁸⁹ In addition, the trading center must have formally declared a trading halt pursuant to its rules or written procedures. Thus, the exception would not include a situation where a trading center merely spread its quotations or switched back to automated quotation mode from manual quotation mode.¹⁹⁰

3. Workable Implementation of Intermarket Trade-Through Protection

Several commenters expressed concern that the original proposed trade-through rule could not be implemented in a workable manner, particularly for high-volume stocks.¹⁹¹ Morgan Stanley, for example, asserted that an inefficient trading center might have inferior systems that would delay routed orders and potentially diminish their quality of execution.¹⁹² Instinet emphasized that protecting a market's quotations "confers enormous power on a market. . . Such power can and will be abused either directly (e.g., by quoting slower than executing orders) or indirectly (e.g., not investing in more than minimum system capacity or redundancy)."¹⁹³ Hudson River Trading noted that markets sometimes experience temporary systems problems and questioned how a trade-through rule would address these scenarios.¹⁹⁴ Nasdaq observed that quotations in many Nasdaq stocks are updated more than two times per second. It said that these frequent changes could lead to many false indications of trade-throughs and that eliminating these "false positives" would greatly reduce the percentage of transactions subject to a trade-through rule.¹⁹⁵ Finally, many commenters noted that market participants need the ability to sweep multiple price levels simultaneously at different trading centers. They emphasized that a trade-through rule should accommodate this trading strategy by freeing each trading center to execute orders immediately without waiting for other trading centers to update their better priced quotations.¹⁹⁶

The Commission agreed with these commenters that intermarket protection against trade-throughs must be workable and implemented in a way that promotes fair and orderly markets, and therefore amended the original proposal in the reproposal to better achieve this objective in a variety of ways. As discussed below, commenters were generally supportive of the measures included in the reproposal as providing necessary flexibility, although several commenters made specific recommendations as to how to improve the operation of the exceptions. In response to these comments, the Commission has made additional modifications to the Order Protection Rule that, in conjunction with the reproposed measures, will further promote its workability.

First and most importantly, as included in the reproposal and as adopted today, only automated trading centers, as defined in Rule 600(b)(4), that are capable of providing immediate responses to incoming orders are eligible to have their quotations protected. Moreover, an automated trading center is required to identify its quotations as manual (and therefore not protected) whenever it has reason to believe that it is not capable of providing immediate responses to orders.¹⁹⁷ Thus, a trading center that experiences a systems problem, whether because of a flood of orders or otherwise, must immediately identify its quotations as manual.

The Commission will monitor and enforce the adopted requirements for automated trading centers and automated quotations. Nevertheless, it concurs with commenters' concerns that well-functioning trading centers should not be dependent on the willingness and capacity of other markets to meet, and the Commission's ability to enforce, these automation requirements. The adopted Order Protection Rule therefore provides a "self-help" remedy that will allow trading centers to bypass the quotations of a trading center that fails to meet the immediate response requirement. Rule 611(b)(1) sets forth an exception that applies to quotations displayed by trading centers that are experiencing a failure, material delay, or malfunction of its systems or equipment. To implement this exception consistent with the requirements of Rule 611(a), trading centers will have to adopt policies and procedures reasonably designed to comply with the self-help remedy. Such policies and procedures will need to set forth specific objective parameters for dealing with problem trading centers and for monitoring compliance with the self-help remedy, consistent with Rule 611. Given current industry capabilities, the Commission believes that trading centers should be entitled to bypass another trading center's quotations if it repeatedly fails to respond within one second to incoming orders attempting to access its protected quotations. Accordingly, trading centers will have the necessary flexibility to respond to problems at another trading center as they occur during the trading day.

Most commenters that addressed the self-help exception supported the exception as providing necessary flexibility to trading centers to avoid inaccessible quotations.¹⁹⁸ Some commenters, however, objected to a statement in the Reproposing Release that a trading center must attempt to contact the non-responsive trading center to resolve a problem prior to disregarding its quotations.¹⁹⁹ They believed that such a requirement would not be practicable or workable, especially during real-time trading.²⁰⁰ One commenter recommended that, instead of requiring notice as a “condition precedent,” the Commission require the trading center electing the self-help exception to contact the slow or non-responding trading center immediately after it elects self-help.²⁰¹

The Commission agrees with the concerns of the commenters that a prior notice requirement may not be practicable or workable in real-time, and that a trading center should be allowed simply to notify the non-responding trading center immediately after (or at the same time as) electing self-help pursuant to objective standards consistent with Rule 611 that are contained in its policies and procedures. An electing trading center must also assess, however, whether the cause of a problem lies with its own systems and, if so, take immediate steps to resolve the problem appropriately.

Another commenter suggested that third-party vendors that provide connectivity among trading centers should be allowed to determine when a trading center has failed to meet the immediate response requirement.²⁰² The Commission agrees that a third-party vendor could perform such a function, but, as with use of the intermarket sweep order exception, the responsibility for compliance with the exception remains with the relevant trading center that uses the services of the third-party vendor. Thus, a trading center is responsible for compliance with the requirements of the exception, including the obligation to establish, maintain, and enforce written policies and procedures and to surveil for their effectiveness, regardless of whether it routes orders using its own systems or a third-party vendor’s systems.

Some commenters believed that the trading center experiencing a problem should have primary responsibility for notifying other trading centers and market participants when such problems occur and when they are resolved.²⁰³ The definition of automated market center in both the reproposed and adopted rule directly imposes this responsibility on the trading center experiencing difficulties.²⁰⁴ It requires such a trading center immediately to identify its quotations as manual whenever it has reason to believe that it is not capable of displaying automated quotations. The trading center must continue to identify its quotations as manual until it no longer has reason to believe that there will be a problem with its quotations. A trading center that continues to identify its quotations as automated when it has reason to believe otherwise would make a material misstatement to other trading centers, investors, and the public.

One commenter believed that, in the absence of an opt-out, the material delay exception was too narrowly drawn, and that market participants should be allowed to avoid trading with trading centers for any objective, reasonable basis as they do today in the context of fiduciary and best execution obligations, and not just for slow response times.²⁰⁵ The Commission does not believe that the scope of the exception should be expanded to give a trading center the ability to avoid another trading center for reasons not related to reliable and efficient accessibility because to do so would be inconsistent with the objectives of the Rule. The exception in paragraph (b)(1) of Rule 611, however, covers any failure or malfunction of a trading center's systems or equipment, as well as any material delay. The Commission believes that there may be certain limited instances where repeated, critical system problems, even those that do not necessarily cause a delayed response time during trading (such as systems problems that repeatedly result in the breaking of trades), would justify use of the exception by other trading centers until the problem trading center has provided reasonable assurance to all other trading centers that the problems have been corrected.²⁰⁶

In many active NMS stocks, the price of a trading center's best displayed quotations can change multiple times in a single second ("flickering quotations"). These rapid changes can create the impression that a quotation was traded-through, when in fact the trade was effected nearly simultaneously with display of the quotation.²⁰⁷ To address the problem of flickering quotations, the Commission included in the reproposal a proposed exception from Rule 611 that would allow trading centers a one-second "window" prior to a transaction for trading centers to evaluate the quotations at another trading center. Specifically, the Commission proposed that pursuant to Rule 611(b)(8) trading centers would be entitled to trade at any price equal to or better than the least aggressive best bid or best offer, as applicable, displayed by the other trading center during that one-second window. For example, if the best bid price displayed by another trading center has flickered between $10.00 and $10.01 during the one-second window, the trading center that received the order could execute a trade at $10.00 without violating Rule 611.

Most of the commenters that addressed this exception supported it.²⁰⁸ The SIA noted that the exception would provide "much-needed practical relief."²⁰⁹ Several commenters, however, raised issues regarding the time frame for the exception, with some supporting a longer window²¹⁰ and some questioning whether it was necessary to establish a specific time frame in the rule, rather than through interpretive guidance.²¹¹ One commenter opposed the exception because it believed that it would create an arbitrage opportunity that could be taken advantage of by computerized market participants.²¹² Another commenter expressed concern that the exception would enable trading centers to execute trades internally and route orders using the worst quotation during the one second window.²¹³

After reviewing the response from commenters, the Commission is adopting the exception as proposed. Allowing a one-second "window" prior to a transaction for trading centers to evaluate the quotations at another trading center will ease implementation of and compliance with the Order Protection Rule by giving tradin

// This display the document title in the html page document.write(document.title);

Regulation NMS

Section II: Order Protection Rule

Table of Contents