Special Study:
Report Concerning Display
of Customer Limit Orders

May 4, 2000

Table of Contents

I. Executive Summary

II. Background: Limit Order Display

A. Market 2000 Report

B. Adoption of the Display Rule in the Equities Markets

III. Importance of Limit Order Display to Market Transparency

IV. Limit Order Display in the Equities and Options Markets

A. Equities Markets

1. Order Routing, Execution and Display Systems

a. Manual Order Handling

b. Automated Display

2. Limit Order Display Data

3. SRO Surveillance Programs

a. Surveillance Procedures

b. Improvements Needed in Surveillance Programs

4. SRO Disciplinary Programs

B. Options Markets

1. Need for Rules Requiring Immediate Display of Limit Orders

2. Need for Enhancements to Automated Order Routing, Execution and Display Systems

3. Limit Order Display Data

4. Need For Enhancement to the Options Last Sale and Quotations Reporting System

5. Surveillance Programs

6. Disciplinary Programs

V. Conclusion

I. Executive Summary

On March 16, 2000, Chairman Levitt requested that Commission staff in the Office of Compliance Inspections and Examinations ("OCIE" or "Staff"), together with the Office of Economic Analysis ("OEA"), prepare a public report analyzing the display of limit orders in our equities and options markets and the adequacy of the markets' surveillance and disciplinary programs for limit order display.¹ This report responds to Chairman Levitt's request.

Investors use two principal types of orders to buy securities: market orders and limit orders. When an investor uses a market order, which is an order to buy or sell a security at the prevailing market price, a broker executes the order at the best price available. In contrast, when an investor uses a limit order, a broker enters an order to buy or sell a security for the investor at a price specified by the investor or better. Thus, limit orders permit investors to compete for better prices than the market is offering. Limit orders serve a critical market function by increasing the information available to the overall market and by allowing all market participants to better determine prices. Further, limit orders have begun to level the playing field between dealers and the investing public by promoting the ability of investors to trade without the intervention of dealers.

In recent years, limit orders have become a powerful tool to enhance investors' role in setting prices. Numerous economic studies confirm the benefits of limit orders. As discussed more fully in the body of this report, key research findings indicate that:

• Limit orders constitute two-thirds of all orders on Nasdaq, and two-thirds of all system orders on the NYSE.

• Limit orders constitute three-quarters of all automated orders on two options markets.

• Most quotes on the NYSE are set by limit orders.

• Spreads appear to be narrowest when set by limit orders.

• Limit orders supply additional liquidity to the market.

• Spreads in Nasdaq stocks have narrowed by 30% following implementation of the Order Handling Rules. More than half of the decrease in spreads was due to the Display Rule.

In view of the importance of limit order display in encouraging competition, transparency, and better executions, it is critically important that limit orders be properly displayed, and that surveillance and enforcement for the display of limit orders be effective. To evaluate that effectiveness, the Staff undertook a series of inspections of SRO surveillance and disciplinary programs in the equities markets following the adoption of the Display Rule. Those inspections revealed serious deficiencies. The Staff also conducted follow-up reviews during March and April 2000. In addition, the Staff reviewed the quality of limit order display by several large OTC market makers and in the options markets.³

Overall, the Staff found that, while significant improvements have been made in some markets since the Staff's initial inspections, there are still problems that must be addressed. Improvements have resulted from automation of the markets' order routing and surveillance programs. This automation has allowed more of the markets to automate the immediate display of eligible customer limit orders, which provides less opportunity for specialists and market makers to manually delay the display of eligible limit orders. The Staff concluded, however, that many exchange specialists and OTC market makers should take steps to improve the prompt display of customer limit orders, and that many SROs can take steps to ensure better compliance with limit order display requirements. The significant weaknesses found during inspections are summarized below and described in detail in this report.⁴

Limit Order Display Weaknesses

• Not all order display systems are fully automated. In fact, many OTC market makers handle all customer limit orders manually. In addition, most automated exchanges and OTC market makers allow some limit orders to be excluded from automated display and execution and are handled manually by specialists and traders. While manual handling in some situations is appropriate, manual handling also creates the potential for limit order display problems.⁵ Significant violation rates were observed with respect to certain manually-handled limit orders. For example:

  • Samples of limit orders received by three larger-sized OTC market makers revealed evidence that significant limit order volume was manually handled, resulting in Display Rule violation rates of 92%, 58%, and 46% of the samples reviewed. Samples of eligible limit orders⁶ received by a fourth larger-sized market maker revealed an apparent Display Rule violation rate of 26.5% of the samples reviewed. The violations included failures to display proper order size, failures to display orders within 30 seconds after receipt, and failures to properly transfer the order display obligation to another exchange system or member.

  • One large OTC market maker's traders turned off the firm's automated display system for an entire day, which resulted in the manual handling of over 1,000 customer limit orders. One trader on that day failed to properly display 83% of the eligible customer limit orders that he manually handled. The violations included failures to display proper order size, and failures to display orders within 30 seconds after receipt.

  • An examination of another OTC market maker revealed that a firm employee turned off the automated display feature for the firm's entire OTC trading desk for a period of several months without detection by the firm.⁷ A sample of eligible limit orders received during this period revealed an apparent Display Rule violation rate of 46%. The violations included possible failures to display proper order size and failures to display orders within 30 seconds after receipt.

  • An earlier examination of the same market maker revealed that, prior to the time the firm implemented an automated display system, the firm failed to properly display 78% of a sample of customer limit orders. Subsequently, the firm implemented a display system, which, although automated, provided traders with extensive opportunities for manual intervention. Thereafter, an examination revealed an apparent Display Rule violation rate of 22%. The violations consisted of failures to display orders within 30 seconds after receipt.

  • On one exchange, a sample of 400 manually-handled customer limit orders eligible for display revealed that approximately one in six were not executed or displayed appropriately, in violation of the Display Rule. The violations included failures to display proper order size, failures to display orders within 30 seconds after receipt, and failures to properly transfer the order display obligation to another exchange system or member.

  These and other findings described in this report indicate that specialists and OTC market makers need to take steps to improve their compliance with display rules and should increase supervisory efforts to ensure compliance.

• While automated display systems that are properly programmed typically result in a near 100% eligible limit order display rate, some systems are not programmed to fully comply with the Display Rule requirements. Data reviewed by the Staff of samples of eligible limit orders received by two of the larger and more fully automated OTC market makers revealed programming deficiencies and apparent Display Rule violations of 19% and 11% of the samples reviewed. The violations included failures to display proper order size, and failures to display orders within 30 seconds after receipt.

• Most market makers reviewed were unable to provide basic data on the display of customer limit orders critical to an effective supervisory and compliance program. Most SROs were also unable to provide complete, accurate data on the display of customer limit orders by their members. For example, in many instances, firms and SROs could not identify whether limit orders were eligible for display, or whether they subsequently became eligible for display. Many firms and SROs were also unable to identify limit orders that were unexecuted or re-routed to another market. The lack of complete, accurate data, as well as synchronized clocks and audit trails, impedes surveillance and makes determining overall compliance rates impossible.⁸

Surveillance for the Display of Limit Orders

• Some SROs conduct no limit order display surveillance. Complaints serve as their only sources to identify customer orders that are not displayed.

• Some SROs do not conduct any automated surveillance for compliance with the Display Rule or SRO rules or policies requiring the display of limit orders. Other SROs conduct random surveillance that, while partially automated, remains manually intensive and inadequate to detect all limit order display violations. Some SROs surveil only for egregious patterns of violations. This surveillance often covers only a small sample of potential violations and is extremely manually-intensive. For example, one exchange, during a seven day period, sampled only 129 of 28,408 (0.45%) manually-excluded customer limit orders. These manual reviews often take many hours and involve the compilation and analysis of data from various sources.

• Several SROs that allowed their specialists and traders to routinely override their automated display systems lacked any surveillance review to determine whether these overrides were appropriate.

• Some SROs were slow in building surveillance systems or suspended surveillance for the proper display of limit orders due to technology development. One SRO completely suspended surveillance for six months, and another SRO severely limited its surveillance for six months.

• Most SROs that did conduct automated surveillance failed to surveil for the immediate display of eligible customer limit orders. Instead, they allowed specialists and traders to routinely display eligible customer limit orders at the 30 ^th second after receipt without flagging such trading for review.

Disciplining Members for Violations of Display Rules

• Sanctioning guidelines for violations of limit order display rules vary greatly, and some SROs impose fines that may not be adequate to deter violations. For example, while one SRO may impose a $1,000 fine for a single violation, another may send a cautionary letter.

• In some cases, the disciplinary process for straightforward Display Rule violations is not conducted in a timely manner. One SRO often imposed sanctions up to 18 months after the occurrence of the violative conduct.

Limit Order Display Rules in the Options Markets

• The options exchanges currently do not have specific rules requiring immediate limit order display. Options markets are taking steps to adopt rules and enhance surveillance. In addition, the options markets currently lack the capacity to publicly display the sizes of their quotes.

Because of the importance of limit order display to our markets, and indications that many limit orders are not being handled appropriately, the examination staff will focus efforts on ensuring that specialists and OTC market makers comply with limit order display obligations, and when appropriate, will refer serious violations to the Division of Enforcement. In addition, the Commission's Office of Economic Analysis will undertake a broad study of limit order display and execution quality in the equities markets.

Top   

Investors generally enter two types of orders: limit orders or market orders. A limit order is an order to buy or sell a security at or better than a specified price. When an investor places a limit order, an investor is competing for a better price than the market is offering or limiting the price that the investor will accept. In this way, the investor is a price setter. In contrast, a market order is an order to buy or sell a security at the current market price. When an investor places a market order, a broker executes the trade at the best current price in the market. In this way, the investor is a price taker.⁹

A. Market 2000 Report

In 1994, in its Market 2000 report,¹⁰ the staff of the Commission's Division of Market Regulation raised concerns about order handling practices in the U.S. securities markets. The staff noted a concern by market participants and economists that OTC market makers were concealing from the public limit orders that would have improved the market price for a security.¹¹ In addition, the report recognized that the failure to display limit orders that improve [both price and size] current quotes raised at least three regulatory concerns.¹² First, the failure to display limit orders could artificially widen spreads. Second, the failure to display limit orders raises fair competition concerns, and third, the failure to display the best quotes results in inferior executions for small customer orders. Accordingly, the report recommended that all market makers display limit orders that better the best intermarket quotes.¹³ The report concluded that the display of these limit orders would provide a more accurate picture of trading interest, result in tighter spreads, and contribute to improved price discovery.¹⁴

B. Adoption of the Display Rule in the Equities Markets

During the Commission's 1996 investigation of the Nasdaq market, the Commission found a number of practices by Nasdaq market makers that, among other things, served to cause a failure to display customer limit orders that would have improved market makers' quotes. To improve the transparency of the markets, including the quality of the handling of customer limit orders, the Commission adopted order handling rules, including the Display Rule, in 1996.¹⁵

The Display Rule is designed to improve the handling of customer limit orders. It requires that exchange specialists and OTC market makers display in their quotes eligible customer limit orders that improve the price or add to the size of their quotes.¹⁶ To comply with the Display Rule, specialists and market makers must either: 1) display the price and full size of a customer limit order in their quote; 2) execute the limit order; or 3) send the limit order to another entity (a broker-dealer, an exchange, or an electronic communications network), which will display it. The Display Rule requires that the specialist or market maker display the order immediately upon receipt. The Commission has stated that, to comply with the requirement that display take place "immediately," specialists and market makers must display (or execute or re-route) eligible customer limit orders "as soon as is practicable after receipt which, under normal market conditions, would require display no later than 30 seconds after receipt."¹⁷ Therefore, specialists and market makers who routinely rely on automated quoting systems to display customer limit orders at the 30 ^th second after receipt would not be deemed to be in compliance with the Display Rule's immediate display requirement.¹⁸ The Commission has further clarified that "this 30 seconds is an outer limit under normal market conditions and is not [intended] as a 30-second safe harbor."¹⁹ The Display Rule specifies that a "customer" limit order does not include an order for the account of a broker or dealer, but does include an order transmitted by a broker or dealer on behalf of a customer.²⁰

In adopting the Display Rule, the Commission identified two reasons why OTC market makers would be reluctant to immediately display eligible customer limit orders. First, a market maker has an incentive to not publicly display a particular customer limit order because the market maker might want to execute the order on a proprietary basis without allowing any other market participants to interact with the order. With regard to this incentive, the Commission stated that:

[A] market maker that holds a customer limit order has, in effect, a private option' to execute the order as principal. The longer this option' remains open, the more time the market maker has to determine whether it can profit from executing the order as principal. This private market maker option,' however, is potentially detrimental to the execution opportunities for the limit order. The Display Rule will limit this option' and expose the order to market-wide trading interest.²¹

These points illustrate that limit orders, if displayed, allow the investors who submit them to compete with dealers for trades and in setting better prices than the market would otherwise provide. Limit orders compete for trades because limit orders that are displayed have a greater chance of attracting other orders in the market. Limit orders compete with dealers in setting prices because limit orders, if displayed, can improve the market price otherwise set by the dealer's quote. The result is a more level playing field between the dealer and the investing public. But in order to level the playing field, the dealers must actually display the limit orders they receive.

Top   

Limit order display has been found to be a key element in promoting competition, providing liquidity, and increasing transparency. Findings from recent studies support the importance of limit order display:

• Limit orders constitute two-thirds of all orders on Nasdaq, and two-thirds of all system orders on the NYSE.²⁴

• Limit orders constitute three-quarters of all automated orders on two options markets.²⁵

• Most quotes on the NYSE are set by limit orders.²⁶

• Spreads appear to be narrowest when set by limit orders.²⁷

• Limit orders supply additional liquidity to the market.²⁸

• Spreads in Nasdaq stocks have narrowed by 30% following implementation of the Order Handling Rules. More than half of the decrease in spreads was due to the Display Rule.²⁹

Top   

A. Equities Markets

1. Order Routing, Execution and Display Systems

a. Manual Order Handling

All exchanges and most large OTC market makers have automated order routing systems, through which orders are received and handled electronically (see Appendix B for a description of order routing systems). The automated routing systems of all exchanges and most OTC market makers are designed to exclude or kick out certain types of orders upon initial receipt for manual handling by specialists or OTC market makers' traders.³⁰ For example, orders that are typically excluded from automated display or execution include larger-sized orders, orders that could create locked or crossed markets³¹ if displayed upon receipt, and orders that could result in a violation of short sale rules upon execution of the orders.³² These types of limit orders generally are routed to the exchange specialist or OTC trader for manual handling.

In addition, some exchanges allow specialists to remove any incoming order from the exchange's automated routing system for manual handling, including limit orders that improve their quotation. Some regional specialists routinely exclude customer limit orders from their own exchanges' automated execution and display systems in order to "lay off" the orders to other exchanges for display and/or execution.³³ During the period in which specialists are trying to access other markets to execute or display a customer limit order, they often prevent the order and subsequent incoming orders from being placed in their limit order book or being displayed in their own exchange's quotes to avoid "double executions" simultaneously on their own exchange and the exchange to which they sent the lay-off order.

Furthermore, some OTC market makers' systems do not provide for the automatic display or execution of customer limit orders. Instead, the responsibility to properly display the eligible customer limit orders in the quote, to execute the orders, or to route those orders falls upon the individual trader, who is required to perform those functions manually.

Although the routing of orders for manual handling is necessary in some circumstances,³⁴ the manual handling of orders also provides opportunities for abuses of the Display Rule. This is because, at some of the exchanges and market makers, during the time that specialists and traders are manually handling eligible customer limit orders, the orders may not be placed on the limit order book or displayed in the quotation and, therefore, are not available for interaction with other orders. Therefore, a specialist or trader, by removing a limit order from the automated display system, may prevent a recently arrived eligible customer limit order from narrowing the spread or preserve a "private option" to trade with the limit order.³⁵

Because the manual handling of customer limit orders by specialists and traders increases the potential for violations of the Display Rule, the Staff believes that exchanges and market makers should take steps to more fully automate the order routing, execution and display processes and, as discussed more fully below, in instances in which manual intervention by specialists or traders is required, to implement surveillance programs that accurately, reliably, and comprehensively surveil for the proper display of customer limit orders.

b. Automated Display

The Staff found that automated display systems at some exchanges and a few of the OTC market makers reviewed provide for the immediate display of certain eligible customer limit orders upon receipt.³⁶ Many of these automated display systems were developed after the adoption of the Display Rule, and they enhance compliance with the rule. As exchanges and market makers increase the use of automated display and execution systems, compliance with the Display Rule should be enhanced.

Some exchanges and most OTC market makers' automatic display systems provide their specialists and traders with a predetermined amount of time to "interact" with eligible customer limit orders after receipt but prior to their execution, display, or routing of the orders to other market centers.³⁷ During this time, the orders are not displayed. Some exchanges and market makers use automated display systems that contain internal clocks that begin a countdown for display purposes that starts upon the specialist's or trader's receipt of the customer limit order. The countdown period ranges from 0 to 30 seconds, depending on the exchange or OTC market maker. If, at the end of the countdown period, the specialist or market maker has not yet interacted with the customer limit order (i.e., executed, displayed, or routed the orders to another market center), the system then automatically displays the customer limit order in the specialist's or trader's quotation. This type of automated display feature ensures that limit orders are displayed within 30 seconds, and improves compliance with the Display Rule.³⁸

The Staff found that all exchanges allow their specialists to institute some type of override of their automatic display systems. Overrides are generally allowed during "fast markets," and on some exchanges are automatic during trade reporting. Some override features shut off the automated display feature altogether, which means that all eligible customer limit orders received while the automatic display system is shut off must be manually entered into the specialist's quote. One exchange requires that specialists obtain and document floor official approval within three minutes of instituting a type of override of the automatic display feature. The same exchange permits specialists to institute a second type of override that halts the order routing and display system, which also prevents the automatic display and surveillance clock from initiating its countdown. Other exchanges allow specialists to manually withdraw individual orders from the automated display system. Similarly, some OTC market makers permit their traders to institute comparable system overrides. The Staff found that OTC market makers' policies regarding supervisory approval for the institution of any automated display system override varied from firm to firm. The Staff believes that the use of overrides to automated display systems by specialists and market makers increases the likelihood of Display Rule violations.

A number of OTC market makers, rather than display the customer limit order in their own quotations, automatically route certain of their orders to another market center for display and/or execution. The routing may occur immediately, as is sometimes the case with non-marketable limit orders, or automatically at the end of the system's countdown period.

2. Limit Order Display Data

In connection with this study, the Staff requested that each SRO and a limited number of OTC market makers submit numbers and percentages of eligible customer limit orders received during December 1999 that were executed, routed to another market center or displayed in the quotation within 30 seconds, 60 seconds, 90 seconds and 120 seconds. The Staff requested this information to assess current rates and timeliness of limit order display.

Most of the SROs and market makers were unable to provide relevant data in response to the Staff's request for a variety of reasons. For instance, four of the exchanges' computer systems were unable to calculate the number of limit orders received that were eligible for display upon receipt. Some of these SROs provided the Staff with data based on a sample of total limit orders instead of a sample of eligible limit orders.³⁹ Therefore, the numbers and percentages they submitted were not relevant in determining rates of compliance with the Display Rule for eligible limit orders. Some SROs also were unable to determine the time that any particular limit order became eligible for immediate display after receipt, and thus, these SROs were unable to produce accurate numbers regarding the number and rate of Display Rule violations. In addition, some of the exchanges were not able to provide any numbers for orders that had been routed to other exchanges.

Of the two SROs that were able to provide complete data in the form requested, one indicated that, for December 1999, 88.3% of its eligible customer limit orders were executed or displayed within 30 seconds of order receipt. The second represented that for a one month period, it executed, displayed, or routed to another market center, 99.57% of its eligible customer limit orders within 30 seconds of receipt.

To adequately surveil for compliance with the Display Rule, an exchange must be able to track whether an incoming customer limit order is eligible for display. It also must have synchronized clocks, audit trails and relevant market information to determine whether the customer limit order was displayed properly based upon market information and the order's price and size. Therefore, those exchanges that were not able to provide the requested information on limit orders are less likely to be able to conduct adequate surveillance for Display Rule violations.

In any event, the limited data provided by the equities SROs and OTC market makers, as well as other data reviewed by the Staff, indicates problems in limit order display. For example:

• Samples of limit orders received by three larger-sized OTC market makers revealed evidence that significant limit order volume was manually handled, resulting in Display Rule violation rates of 92%, 58%, and 46% of the samples reviewed. Samples of reviewed eligible limit orders received by a fourth larger-sized OTC market maker revealed an apparent Display Rule violation rate of 26.5% of the samples reviewed. The violations, resulting from the use of automated display systems that failed to display customer limit orders within 30 seconds after receipt and the use of manual override features, included failures to display proper order size, failures to display within 30 seconds after receipt, and failures to properly transfer the order display obligation to another exchange system or member.

• One large OTC market maker's traders turned off the firm's automated display system for an entire day, which resulted in the manual handling of over 1,000 customer limit orders. One trader on that day failed to properly display 83% of the eligible customer limit orders that he manually handled. The violations included failures to display proper order size, failures to display orders within 30 seconds after receipt.

• An examination of another OTC market maker revealed that a firm employee turned off the automated display feature for the firm's entire OTC trading desk for a period of several months without detection by the firm.⁴⁰ A sample of eligible limit orders received during this period revealed an apparent Display Rule violation rate of 46%. The violations included possible failures to display proper order size and failures to display orders within 30 seconds after receipt.

• An earlier examination of the same market maker revealed that, prior to the time the firm implemented an automated display system, the firm failed to properly display 78% of a sample of customer limit orders. Subsequently, the firm implemented a display system, which, although automated, provided traders with extensive opportunities for manual intervention. Thereafter, an examination revealed an apparent Display Rule violation rate of 22% The violations consisted of failures to display orders within 30 seconds after receipt.

• Sample data reviewed by the Staff from two of the larger, more fully automated OTC market makers revealed that system programming deficiencies resulted in apparent Display Rule violation rates of 19% and 11%. The violations included failures to display proper order size, and failures to display orders within 30 seconds after receipt.

• Until the Staff's inspection in early 1999, one SRO allowed its specialists to fail to display eligible customer limit orders for up to 65 seconds after receipt of such orders. The use of the 65 second time parameter resulted in the SRO failing to discover 13% of its specialist's Display Rule violations during a one month period in 1999.⁴¹ The violations included failures to display proper order size, failures to display orders within 30 seconds after receipt, and failures to properly transfer the order display obligation to another exchange system or member.

• One market maker's policies provided that for orders that improved the firm's quote, but did not equal or better the NBBO, the firm was allowed to improve its proprietary quote to match the price of the customer limit order and represent the newly priced order as its own.⁴² However, the Staff found that the market maker's quotation sizes often failed to reflect the entire sizes of the customer orders. The Staff reviewed 454 eligible customer limit orders received by the market maker and found that, as a result of the policies, 45 (10%) of those limit orders were improperly displayed. The violations consisted of failures to display proper order size.

• During one Display Rule inspection, the Staff reviewed a sample of customer limit orders received by an equities exchange. In a sample of approximately 400 customer limit orders that were manually removed from the exchange's order display system by the exchange's specialists, 15.8% of those customer limit orders were handled in violation of the Display Rule. The violations included failures to display proper order size, failures to display orders within 30 seconds after receipt, and failures to properly transfer the order display obligation to another exchange system or member.

• Sample data provided by one SRO indicated that for the first 5 trading days of December 1999, the SRO's specialists failed to execute or display 3,640 (5%) of the eligible customer limit orders handled on the floor within 2 minutes after receipt.⁴³

• Another SRO reported that in excess of 99.9% of eligible limit orders were executed or displayed within 30 seconds of order receipt. The data, however, do not include eligible limit orders received during the times that specialists effected two types of commonly-used overrides of the automatic display function. The Staff found that in instances in which one type of manual override was instituted, 8.55% of customer limit orders received after 10 minutes following the imposition of the override were not handled in compliance with the Display Rule.⁴⁴ The violations included failures to display orders within 30 seconds after receipt.

3. SRO Surveillance Programs

a. Surveillance Procedures

The Staff found a wide disparity among the SROs in their Display Rule surveillance programs and procedures. Some SROs surveil for compliance with the Display Rule using automated surveillance systems, while other SROs conduct manually intensive, and therefore less efficient and comprehensive, surveillance programs. The limited amount of data captured by some order routing and surveillance systems also inhibits Display Rule surveillance. The different types of surveillance are described below.

i. Sampling/Manual Surveillance

Some SROs have not fully automated their Display Rule surveillance, and rely on a manual sampling of undisplayed orders. The "sampling method" of surveillance typically consists of a manual review of a random sample of customer limit orders that an exchange's automated review has identified as orders that were either: 1) manually withdrawn from the automated display system; and/or 2) not displayed within 30 seconds from time of receipt by the specialist or market maker.

Some of the exchanges' Display Rule exception reports produce so many kickouts that they can result in a hundred pages or more of exceptions per day.⁴⁵ Many of these reports use parameters that may be overinclusive. For instance, the reports may include customer limit orders with prices that are not eligible for display because they are away from the market, and/or limit orders that are excluded from the Display Rule.

Exchanges whose specialists route a significant number of orders to other market centers typically employ an additional manual review to ensure that these layed-off orders were properly routed to the other market centers. Analysts review samples of orders that were withdrawn from the exchanges' systems to verify that those orders were actually routed to other market centers in a timely manner, at the proper limit order prices, and in the proper sizes.

ii. Automated Surveillance

Other SROs have more fully automated Display Rule surveillance. Several SROs create exception reports that capture and track each customer limit order received by a specialist or market maker and produce, on a monthly basis, surveillance reports that reveal both numbers and percentages of all eligible customer limit orders displayed and/or executed by individual specialists in accordance with the requirements of the Display Rule. Specialists and member firms that fail to achieve display rates in certain predetermined percentages typically receive some form of disciplinary action from the SROs.

b. Improvements Needed in Surveillance Programs

i. Inadequacies in the Sampling Method

The sampling method of limit order display surveillance is extremely time-consuming and manually intensive, and because it often includes only a small portion of eligible limit orders, it results in an incomplete review. Examples of weaknesses in this type of surveillance are as follows.

• One SRO, during a seven day period in early 1999, sampled only 129 of the 28,408 manually excluded orders (0.45%) that were captured in its daily exception reports.⁴⁶

• One SRO represented that it performs limit order display surveillance for approximately 5% of its specialists per day. For each specialist chosen, only a sample of limit orders captured by the limit order display exception report is analyzed. Thus, only a small fraction of total customer limit orders are reviewed for compliance with the Display Rule.

• Until the Staff's on-site inspection, one SRO surveilled only a small percentage of executed customer limit orders, and did not surveil for unexecuted, undisplayed limit orders.⁴⁷

• Three SROs are unable to determine whether their specialists lay off particular orders. Therefore, for the sample of orders reviewed, analysts must request off-floor routing data directly from the specialist or from the off-floor routing data providers and manually attempt to match and verify layed-off orders with those orders that were excluded from the exchange's automated display system.⁴⁸

The Staff also found that the SROs that use the sampling method of surveillance sometimes fail to refer even the most egregious single violations of the Display Rule to their investigatory and/or enforcement staff. Instead, the SROs may instruct their analysts to pursue violations of the Display Rule only after the analysts determine that a specialist or market maker had established a "pattern" of Display Rule violations. Because these SROs review only a small sample of total limit orders, it may be difficult to identify a pattern of violations.⁴⁹

ii. Over-Reliance on Automated Display

The Display Rule provides that eligible customer limit orders must be displayed immediately, but no later than 30 seconds under normal market conditions. Therefore, market makers or specialists that routinely allow an automated order routing or display system to display customer limit orders at the 30 ^th second after receipt would not be deemed in compliance with the Display Rule because they may be able to display the eligible customer limit order more quickly.⁵⁰ The Staff found that only two SROs surveil for instances in which individual specialists over-rely on automated quotation systems to display eligible customer limit orders. Each month, these systems capture the number and percentage of times that specialists rely on their automated display systems to display eligible customer limit orders at 30 seconds. Specialists may then be sanctioned for failing to display immediately. Many SROs lack surveillance review for overreliance on automated display systems.

iii. Inadequate Surveillance Policies and Procedures

In inspections in the past several years, the Staff found that a number of SROs used incomplete or deficient surveillance policies and procedures. Examples of inadequate surveillance policies and procedures are outlined below.

• Until the Staff's on-site inspection, one SRO's procedures provided for surveillance of customer limit orders at 65 seconds.

• One exchange's procedures provide that no Display Rule surveillance is required to be performed for orders received by the exchange prior to 9:45 AM. The Display Rule, however, contains no such exemption.

• Some SRO surveillance staff are inexperienced or are not adequately trained in Display Rule surveillance and are not provided with clear and concise policies and procedures with which to conduct proper surveillance reviews.

Orders on the limit order book may become displayable during the course of a trading day as the market moves. The Display Rule requires that these orders then be displayed immediately. Two SROs and numerous market makers do not conduct surveillance for the proper display of customer limit orders that reside on the limit order book once they become eligible or re-eligible for display.⁵¹ The staff found a number of instances in which the burden of requoting falls on the manual efforts of specialists and traders, which increases the likelihood of Display Rule violations. The following are some examples.

• Two SROs have no automated requote feature. Instead, specialists must manually input the new quotes. These specialists, when changing the quotations, must take into account any customer limit order on the limit order book.⁵²

• A number of market makers use automated display systems that often fail to automatically display limit order book orders once they become eligible/re-eligible for display. Thus, the individual traders become responsible for manually inputting all customer limit orders on the market maker's limit order book as they become eligible for display.

• Some market makers fail to surveil traders for over-reliance on automated display systems in instances in which the systems initiate a new countdown (up to 30 seconds) when the customer limit orders on their limit order books become re-eligible for display.

Many SROs do not monitor or track instances in which specialists manually override or turn off automatic display systems. While some exchanges and market makers require approval by supervisory personnel, not all types of manual overrides are captured and tracked by SRO surveillance. One SRO, for example, does not monitor Display Rule compliance for the first 10 minutes of a certain type of override. In addition, one SRO's surveillance countdown clocks start when an eligible customer limit order is received by the limit order book and not when the order is received by the specialist at the post. Therefore, any delay that occurs after order receipt by a specialist but prior to the customer limit order being placed on the limit order book is not measured by the surveillance exception reports.

In addition, some market makers and specialists fail to monitor or track instances in which traders override or turn off automatic display systems. For example, an individual at one market maker turned off the automated display feature for the firm's entire OTC trading desk for a period of several months, apparently without the surveillance staff detecting the deactivation. A sample of reviewed eligible limit orders received during this period revealed an apparent Display Rule violation rate of 46%. In addition, the Staff found that prior to its on-site inspection of another OTC market maker, traders had turned off the firm's automated display system, also apparently without detection by the firm. A sample of reviewed eligible limit orders received during this period revealed an apparent Display Rule violation rate of 83%.

vi. Inadequate Surveillance for Locked and Crossed Markets and Short Sales

The Display Rule "does not require an exchange specialist or OTC market maker to immediately display a customer's limit order that would lock or cross the market,"⁵³ or to display a customer's short sale limit order "where doing so would likely result in a violation of short sale rules."⁵⁴ However, in instances in which a locked or crossed market "unlocks" or "uncrosses," or a short sale order is eligible for display because it would not violate a short sale rule, a specialist or market maker would then be required to display the limit order in accordance with the Display Rule. Until the Staff's inspections in early 1999, some SROs deemed all of these customer limit orders to be exempt from the Display Rule and, therefore, failed to review for instances in which the orders became eligible for display.

vii. Suspension of Exchange Surveillancefor Display Rule Violations

The Staff found that two exchanges each suspended or sharply reduced their surveillance for Display Rule violations for six month periods. Both exchanges stated that they suspended or limited their Display Rule surveillance programs during these periods because they were in the process of developing and implementing more automated Display Rule surveillance systems.

4. SRO Disciplinary Programs

Most violations of the Display Rule are included in a SRO's equivalent of a minor rule violation plan. More serious or multiple violations may result in a referral to the SRO's enforcement staff for more formal disciplinary action. The Staff found, however, that sanctioning guidelines for Display Rule violations vary greatly among the SROs. For instance, first and second violations of the Display Rule at one SRO may result in a cautionary letter, whereas another SRO may impose a $1,000 fine for a single violation. Another SRO surveils for Display Rule violations on a percentage basis by month and (automatically) imposes a $1,000 fine or more for a failure to display a preset percentage of eligible customer limit orders.

The Staff found that some SROs have not been aggressive in imposing sanctions for Display Rule violations. Until the Staff's inspection of two SROs, they had never imposed monetary sanctions on their members for failures to properly display eligible customer limit orders. In fact, one SRO official admitted that there were many violations of the Display Rule in excess of three minutes, for which the SRO imposed no disciplinary actions.⁵⁵ In addition, the Staff found that some SROs fail to conduct their disciplinary processes in a timely manner. One SRO has imposed sanctions on its members up to 18 months after the violative conduct. All of the SROs, however, have taken steps to enhance discipline for Display Rule violations.

The Staff generally found the disciplinary processes for sanctioning violations of the Display Rule by SROs that conduct less-automated surveillance generally to be more discretionary and, therefore, more likely to be inconsistently applied. As described above, some SROs review for patterns of Display Rule violations without adequately setting forth the requirements to establish a pattern of violative conduct. These SROs also tend to consolidate Display Rule violations for purposes of issuing cautionary letters or fines and limit "lookback" periods for considering similar past violations when imposing sanctions.

Conversely, the Staff found that those SROs that use more fully-automated surveillance procedures and practices tend to establish more objective and predictable disciplinary practices. These SROs are more likely to be able to produce monthly order display percentages by specialist and specialist firm.

B. Options Markets

1. Need for Rules Requiring Immediate Display of Limit Orders

As noted, the Display Rule under the Exchange Act applies only to equities. There is currently no comparable rule under the federal securities laws that applies to trading in options. Each of the options markets has rules or policies governing the handling of limit orders. Two exchanges' rules provide that options Designated Primary Market Makers ("DPMs") or Order Book Officials ("OBOs") continuously display, in a visible manner, the full size of the highest bid and lowest offer. Some exchanges require their members to use due diligence to ensure that the best available bid and offer is displayed, or consider the timely display of customer limit orders in assessing specialists performance. Importantly, none of the options exchanges' rules currently require the immediate display of customer limit orders. Each options exchange has indicated that it is formulating rules requiring the immediate display or execution of limit orders by members.

2. Need for Enhancements to Automated Order Routing, Execution and Display Systems

The Staff's recent inspections of the options exchanges revealed several instances in which the design of the automated execution and automated order routing systems may serve to disadvantage some customer limit orders. The Staff found that three options exchanges' automated execution systems are programmed to route most incoming orders that are eligible for execution against an order on the limit order book, including marketable limit orders, to manual handling instead of routing them for automatic execution against the order in the limit order book.⁵⁶ This is because the auto-ex systems are set to execute incoming customer limit orders only against market maker or specialist quotes, but not against the limit order book, even in instances when the limit order book has priority.⁵⁷ The Staff also found that three options exchanges' automated order routing systems prevent many orders that improve the quote from being included in the quote automatically for immediate display without manual intervention.⁵⁸ As with equities limit orders, the manual handling of customer limit orders may be appropriate in some circumstances, but also increases the potential for limit orders to be concealed.

The Staff also found that only two of the options exchanges' automated order routing systems have features that allow for member firm booths on the floor to enter all types of system-eligible orders that were sent to the booth by telephone or wire into the exchange's automated order routing and execution system. In response to the Staff's recent inspections, two exchanges have stated to the Commission that they plan to implement technology, and rules requiring the use of that technology, that will allow the system to capture options orders directed to floor members for floor booth order entry. Systems that permit floor brokers to enter orders that they receive by phone or by wire into the exchange's automated order routing and execution system will afford the order the greater protections of the systems. These include automatic executions, automatic placement on the book, automatic display, and a more reliable audit trail for use by surveillance personnel.

3. Limit Order Display Data

In connection with this study, the Staff requested that each options exchange submit numbers and percentages of eligible customer limit orders received during December 1999 that were executed, routed to another market center or displayed in the SRO's quotation within 30 seconds, 60 seconds, 90 seconds and 120 seconds. The Staff requested this information to assess current rates and timeliness of limit order display.

None of the options exchanges were able to provide complete, relevant data in response to the Staff's request. The exchanges conduct very limited surveillance for limit order display and do not compile data on statistics relevant to the display of limit orders. In fact, three of the options exchanges were unable to produce any limit order data. These three exchanges stated that either they were unable to produce such a report, or that current systems constraints prevented them from providing the data prior to the date of this report.

The one exchange that did provide data produced information that was limited to only those customer limit orders that were routed through their automated routing system and ultimately executed on the exchange. No data was available on limit orders that were not executed, even if they were eligible for display. The exchange stated that for the executed limit orders routed through their automated system, the average length of time from order receipt until display or execution, whichever came first, was 43 seconds. In addition, the exchange represented that 96% of these limit orders were displayed or executed within two minutes of receipt.

Just as with equities limit orders, to adequately surveil for the proper display of customer limit orders, an exchange must be able to track whether an incoming limit order is eligible for display or not. To do this, it should have synchronized clocks, audit trails and relevant market information to determine whether the customer limit order was displayed properly based upon market information and the order's price and size. Therefore, the options exchanges should take steps to enhance the availability of this data in order to conduct adequate surveillance for the proper handling of limit order display and execution.

4. Need for Enhancement to the Options Last Sale and Quotations Reporting System

To enhance the display of limit orders, improvement is needed to the system that displays the best quotations among the options exchanges and disseminates them publicly. That system, the Options Price Reporting System ("OPRA System"), collects last sale reports and quotation information from the four operating options exchanges and provides for the uniform dissemination of the information over a network to vendors, subscribers, and other approved persons. Each exchange is required to transmit the quotation information so as to reflect the current market at the exchange in the security. The quotation information includes the premium bid (or offered), the option's series, the market in which the quotation was entered, and other relevant information. However, the OPRA system does not include the size of the quotation in the information it collects and disseminates. As a result, market participants, including other options exchanges and members of the public, are not able to see the size that is bid or offered on a specific exchange. As discussed previously, the dissemination of this trading interest is a requirement of a truly transparent market. Therefore, while some options exchanges' rules require full display of the customer order size on the book, because of the limitations of the OPRA system, the full size of the order is not publicly disseminated, and therefore is only available for view on the floors of these exchanges. OPRA plan participants have expressed their intent to develop a system that will disseminate quotations with order size information.

5. Surveillance Programs

The Staff's recent inspections of the options exchanges covered, among other things, compliance with the applicable limit order display requirements of each exchange. The inspections revealed that the options exchanges should improve their surveillance for limit order display. In fact, three of the four options exchanges conducted very limited surveillance for display of limit orders, relying solely on customer complaints to identify failures to display eligible limit orders.

The fourth options exchange's surveillance procedures for limit order display are limited to labor-intensive reviews of order tickets received and processed by options specialists. The order ticket reviews are conducted twice yearly, and consist of reviews of four randomly chosen days of trading in three options classes. As stated above in connection with the discussion of equities surveillance, a review of order tickets is cumbersome and inefficient.

In response to the Staff's 1999 inspections, all of the options exchanges have stated that they intend to enhance their surveillance of limit order display. One exchange has begun to conduct daily manual surveillance using a random sample of the limit orders of one specialist in one particular issue. In addition, all four of the options exchanges have indicated that they plan to develop and implement automated limit order display surveillance programs.

6. Disciplinary Programs

The Staff's recent inspections revealed that all of the options exchanges should improve discipline for the display of limit orders.⁵⁹ The Staff found that several investigations conducted by options exchanges, which stemmed from customer complaints rather than surveillance programs, appeared to have been improperly closed without action. The following are examples.

• One exchange declined to impose any sanctions for failing to properly display a limit order in instances in which the specialist provided a price adjustment or execution to the customer after the customer complained.

• Another exchange closed one investigation into a customer complaint based on the judgment of options surveillance department personnel that "five minutes was a reasonable amount of time to disseminate a bid or offer."⁶⁰

Top   

Limit orders serve a critical function by increasing information available to the market and by allowing all market participants to better determine prices. The Commission's OEA will undertake a broad study of limit order display and execution quality in the equities markets. Economic data strongly indicates that limit orders are increasingly driving market prices and narrowing spreads. Accordingly, proper limit order display and handling is critical. The findings described in this report indicate that, as the markets' order routing and surveillance programs have become more automated, compliance with the limit order display and handling obligations is enhanced. The findings of violations however, indicate that specialists and market makers should take steps to enhance compliance with display rules, and that SROs can increase surveillance and discipline of their members. In light of the importance of limit orders, limit order display and handling will remain a priority for Commission examiners, and where appropriate, serious findings will be referred to the Division of Enforcement.

    Appendix A – Improving the Handling of Limit Orders

    Appendix B – Equities and Options Trading: Background

Footnotes

¹ Visible Prices, Accessible Markets, Order Interaction, Remarks of Arthur Levitt, Northwestern University School of Law, Kellogg Graduate School of Management (March 16, 2000).

² The Display Rule is codified in Rule 11Ac1-4 under the Securities Exchange Act of 1934 ("Exchange Act").

³ This report includes summaries of information from inspections of equities and options exchanges and several OTC market makers conducted since 1997, and recent information provided by SROs and market makers on limit order display. OCIE appreciates the cooperation of the SROs and market makers in the preparation of this report. Because information was obtained as part of the Commission Staff's and SRO's non-public oversight, individual SROs and market makers are not identified by name. The examination findings described in this report are those of the Staff, and are not the findings of the Commission.

⁴As a result of these inspections, some SROs and market makers have improved their rates of limit order display and have implemented improvements to their surveillance and disciplinary procedures.

⁵ In contrast, those systems that are designed to handle eligible customer limit orders automatically and are programmed properly, typically have a near 100% limit order display rate.

⁶ An "eligible" customer limit order is a customer limit order that is received or held by an exchange specialist or market maker that either: (1) is superior in price to the specialist's or market maker's existing quotation; or (2) adds to the size of the specialist's or market maker's quotation, is at the NBBO, and represents more than a de minimis change in relation to the specialist's or market maker's quotation.

⁷ These are apparent violations and the matter is still being reviewed by the SRO.

⁸ The Staff believes that the full implementation of the NASDR's Order Audit Trail System ("OATS") will enhance compliance and surveillance with the Display Rule by OTC market makers.

⁹ See Remarks of Chairman Levitt, Northwestern University School of Law, supra note 1.

¹⁰ Securities and Exchange Commission, Division of Market Regulation, Market 2000: An Examination of Current Equity Market Developments (January 1994) ("Market 2000 Study").

¹¹ See Market 2000 Study at IV-5.

¹² See Market 2000 Study at IV-6.

¹³ See id.

¹⁴ See id.

¹⁵ Securities Exchange Act Release No. 37619A (Sept. 6, 1996), 61 Fed. Reg. 48,290 (Sept. 12, 1996) ("Adopting Release").

¹⁶ Exchange Act Rule 11Ac1-4, 17 C.F.R. 240.11Ac1-4 (2000). An exchange specialist or OTC market maker is not required to immediately display an eligible customer limit order if any of the following seven exceptions to the Display Rule applies: 1) the order is executed immediately upon receipt; 2) the customer expressly requests that the order not be displayed; 3) the order is an odd-lot order; 4) the order is a block-sized order (10,000 or more shares or $200,000 or more in total market value); 5) the order is delivered immediately to an exchange or association-sponsored system that displays limit orders in compliance with the Display Rule; 6) the order is delivered immediately to another exchange member or market maker that handles the order in accordance with the Display Rule; or 7) the order is an "all-or-none" order. In addition, specialists and market makers are not required to display customer limit orders that match the price but increase the size of the quotation if the specialist or market maker is not quoting at the national best bid and offer ("NBBO").

¹⁷ See Adopting Release at 48,304.

¹⁸ See Letter from Richard R. Lindsey, Director, Division of Market Regulation, Securities and Exchange Commission, to James F. Duffy, Executive Vice President and General Counsel, American Stock Exchange (Apr. 1, 1997).

¹⁹ See Securities Exchange Act Release No. 40066 (June 4, 1998), 63 Fed. Reg. 31,817, 31,818 n.14 (June 10, 1998).

²⁰ Exchange Act Rule 11Ac1-4(a)(6), 17 C.F.R. 240.11Ac1-4(a)(6) (2000).

²¹ Adopting Release at 48,299.

²² Exchange specialists and OTC market makers generally are prohibited from trading at prices equal or superior to customer limit orders they hold without executing the customer limit orders.

²³ Adopting Release at 48,300 n.123.

²⁴ William J. Atkinson and Peter G. Martin, Halving the Minimum Tick Size on the NYSE, Office of Economic Analysis, U.S. Securities and Exchange Commission (1999).

²⁵ The Commission's OEA's preliminary review of orders entered for a four-day period in September 1998.

²⁶ Kee H. Chung, Bonnie F. Van Ness, and Robert A. Van Ness, Limit Orders and the Bid-Ask Spread, 53 J. Fin. Econ. 255 (1999). See also, Michael A. Goldstein and Kenneth A. Kavajecz, Eighths, Sixteenths and Market Depth: Changes in Tick Size and Liquidity Provision on the NYSE, NYSE Working Paper 98-01 (1998).

²⁷See Chung, Van Ness, and Van Ness, supra note 26.

²⁸ Goldstein and Kavajecz show that on the NYSE for three months in 1997, 75% of depth at the quote is provided by limit orders. See Goldstein and Kavajecz, supra note 26. This finding is confirmed by Kavajecz using TORQ data from 1990-91. Kenneth A. Kavajecz, A Specialist's Quoted Depth and the Limit Order Book, 54 J. Fin. 747 (1999). Kavajecz also finds that limit orders provide all the depth (at the quote) about 30% of the time. See id.