Appendices:
Report on the Practice of Preferencing

Appendix A: CSE Preferencing Pilot Program (1991 - 1996)

A. Commission Approval of the Initial CSE Preferencing Pilot Program in 1991

B. Extensions of the CSE's Preferencing Pilot Program

C. CSE Request for Permanent Approval of the Preferencing Pilot Program

Appendix B: BSE Competing Specialist Initiative Pilot Program (1994 -1996)

A. Commission Approval of the Initial BSE CSI Pilot Program in 1994

B. Extensions of BSE's CSI Pilot Program B-4

Appendix C: Development of the Cincinnati Stock Exchange's National Securities Trading System (1977 - 1990)

Appendix D: Detailed Data Description D-1

Appendix A

CSE Preferencing Pilot Program (1991 - 1996)

A.Commission Approval of the Initial CSE Preferencing Pilot Program in 1991

In April 1990, the CSE proposed to adopt its preferencing program on a six month pilot basis. [See Securities Exchange Act Release No. 27910 (April 17, 1990), 55 FR 15311 (April 23, 1990) (File No. SR-CSE-90-06) (notice of CSE proposal to adopt a preferencing program on a six-month pilot basis). The CSE amended the proposal in November 1990 to propose the following restrictions on the program: (1) restricting to 60 the number of stocks each preferencing dealer could trade; (2) preventing a preferencing dealer from making direct cash payments for preferenced order flow; (3) allowing preferencing dealers to preference their customer order flow which is related to index arbitrage only on plus or zero plus ticks when the Dow Jones Industrial Average declines by fifty points or more from the previous day's closing value; and (4) clarifying that its policy on the voluntary withdrawal of Designated Dealer status was applicable to preferencing dealers. See Letter from Frederick Moss, Chairman of the Board of Trustees, CSE, to Richard G. Ketcham, Director, Division of Market Regulation, SEC, dated November 14, 1990. ] In its proposal, the CSE recounted its previous efforts to attract retail order flow to the Exchange through the development of electronic interfaces with retail order delivery systems and increasing the number of issues covered in the NSTS by creating the primary Designated Dealer category of market maker. [See Securities Exchange Act Release No. 27220 (September 5, 1989), 54 FR 37856 (September 13, 1989) (File No. SR-CSE-89-1) (order approving CSE's adoption of primary Designated Dealer status). See Appendix C, at notes 24-26 and accompanying text for a description of the primary Designated Dealer status. There are currently no stocks traded on the CSE on a primary Designated Dealer basis.] The Exchange contended, however, that such efforts were not sufficient to overcome the lack of incentivein the CSE's multiple market maker environment for a market maker to direct its retail order flow to the Exchange. In this regard, the CSE stated that its dealers were subject to losing all or a portion of their public orders entered into the NSTS to other market makers on the Exchange. [See supra Part I.B.2.] Further, the Exchange cited its need to compete with unitary specialists affiliated with order flow firms on the other regional exchanges who did not face the price competition of other specialists on their floor. As a result, the CSE claimed that altering the priority rules between professional trading interests was necessary to put the CSE dealers on par with other regional specialists and consequently attract retail order flow and enhance liquidity on the Exchange.

The CSE stressed that the public would not be disadvantaged by the proposed rule change because either side of a paired order trade entered into NSTS would be replaced with pre-existing same-sided customer interest on the CSE's central limit order book up to the size of the proposed trade, if any existed. Accordingly, if the dealer's side is replaced, the trade would occur between two customer orders without the intervention of the preferencing dealer. The replacement of the customer's side by a pre-existing customer limit order maintains time priority for public agency orders on the CSE limit order book. In addition, the CSE asserted that the public would benefit from the increased liquidity supplied by CSE dealers, because public limit orders in the book which are priced at the Intermarket Trading System best bid or offer("ITS/BBO") would be executed in order for CSE dealers to interact with their own public order flow.

In February 1991, the Commission approved the CSE's proposal on an initial six-month pilot basis. [See Securities Exchange Act Release No. 28866 (February 7, 1991), 56 FR 5854 (February 13, 1991) (File No. SR-CSE-90-06) (initial approval of CSE preferencing pilot program on a six- month basis, through August 7, 1991).] In its order initially approving the preferencing pilot program, the Commission stated its belief that the proposal addressed the CSE's legitimate desire to attract additional business to the exchange, while at the same time providing adequate protection for public agency orders sent to the exchange. Moreover, the Commission noted that it was appropriate to approve the proposal on a pilot basis in order to allow the CSE and the Commission to better analyze the effect that dealer preferencing would have on the market.

B. Extensions of the CSE's Preferencing Pilot Program

After the approval of the initial preferencing pilot program, the Commission granted approval to a series of extensions of the program between 1991 and 1995. [Securities Exchange Act Release Nos. 29524 (August 5, 1991), 56 FR 38160 (August 12, 1991) (File No. SR-CSE-91-03) (extension of pilot to February 7, 1992); 30353 (February 7, 1992), 57 FR 5918 (February 18, 1992) (File No. SR-CSE-92-01) (extension of pilot to August 7, 1992); 31011 (August 7, 1992), 57 FR 38704 (August 26, 1992) (File No. SR-CSE-92-10) (extension to May 7, 1993); 32280 (May 7, 1993), 58 FR 28424 (May 13, 1993) (File No. SR-CSE-93-03) (extension of pilot to May 7, 1994); 33975 (April 28, 1994), 59 FR 23243 (May 5, 1994) (File No. SR-CSE-94-03) (extension of pilot to August 6, 1994); 34493 (August 5, 1994), 59 FR 41531 (August 12, 1994) (File No. SR-CSE-94-06) (extension of pilot to May 18, 1995); 35717 (May 15, 1995), 60 FR 26909 (May 19, 1995) (File No. SR - CSE-95-06) (extension of pilot to October 2, 1995); and 36324 (September 29, 1995), 60 FR 52436 (October 6, 1995) (File No. SR-CSE-95-07) (extension of pilot to March 29, 1996).] Along with these extensions, theCommission approved various changes to the preferencing pilot program. [See Securities Exchange Act Release Nos. 29885 (October 31, 1991), 56 FR 56676 (November 6, 1991) (File No. SR-CSE-91-03) (increase in number of stocks from 60 to 125); 30809 (June 15, 1992), 57 FR 27990 (June 23, 1992) (File No. SR-CSE-92-04) (increase in number of stocks from 125 to 250); and 31011 (August 7, 1992), 57 FR 38704 (August 26, 1992) (File No. SR- CSE-92-10) (increase in number of stocks from 250 to 350).]

In March 1994, the CSE filed a request for permanent approval of its preferencing pilot program, including removal of the restriction on the number of stocks a preferencing dealer could trade and the lifting of the ban preventing a preferencing dealer from making direct cash payments for preferenced order flow. [See Securities and Exchange Commission File No. SR-CSE-94-01 (initial request for permanent approval of CSE preferencing program).] The CSE withdrew its proposal for permanent approval of the preferencing program in June 1994, [In the interim, the Commission approved a further three month extension of the pilot. See Securities Exchange Act Release No. 33975 (April 28, 1994), 59 FR 23243 (May 5, 1994) (File No. SR-CSE-94-03) (order approving extension of pilot to August 6, 1994).] and submitted a proposal to extend the program until May 1995. [See Securities Exchange Act Release No. 34258 (June 24, 1994), 59 FR 33992 (July 1, 1994) (File No. SR-CSE-94-06) (notice of CSE request to extend pilot to May 18, 1995). ] Along with this request, the CSE submitted a number of revisions and clarifications to its rules. These were intended to codify the five principals that theCommission's Division of Market Regulation recommended in its Market 2000 Study regarding all market makers in listed stocks, including third market makers and firms internalizing order flow. [See Division of Market Regulation, SEC, Market 2000: An Examination of Current Equity Market Developments ("Market 2000 Study") (January 1994), Study III, at III-14. The five principles that the Division recommended all market makers adhere to were the following: (1) dealers should expose customer limit orders that are better than the existing ITS best bid or offer unless a customer expressly requests that the order not be exposed; (2) dealers should not trade ahead of customer limit orders; (3) if a dealer holds a customer buy order and a customer sell order that can be crossed, the dealer should cross them without interposing itself as dealer; (4) dealers should establish fixed standards for queuing and executing customer orders; and (5) dealers should not trade at a price outside the ITS/BBO without satisfying the market interest at that price in accordance with ITS trade-through and block policies. See Id. at Appendix II (description of ITS trade-through and block policies).]

First, the CSE proposed to add interpretive language to CSE Rule 12.10, the Exchange's best execution rule, in order to require exposure of a limit order priced either at or between the ITS/BBO. [CSE Rule 12.10, Interpretation and Policy .01 required a member to expose to the national market system all or a representative portion of any limit order which is priced either at or between the ITS/BBO, unless (1) such order is immediately executed; or (2) the customer expressly requests that the order not be exposed. In determining the amount of a limit order to expose, a member, as part of its fiduciary duties, was charged with ensuring the best execution of a customer's order. The member could meet the requirement by displaying the order or a representative part in its CSE quote, the CSE central limit order book, or by forwarding the order to either another firm or exchange. The CSE's limit order display policy was amended in conjunction with the permanent approval of the preferencing program. See infra notes - . This amended policy was itself superceded by the recent adoption of SEC Rule 11A1c-4, 17 CFR 240.11A1c-4 ("Display Rule"). ] Second, to encourage members to place public agencylimit orders on the CSE's central limit order book, the CSE proposed eliminating the transaction charge on public agency limit orders. Third, the CSE proposed to codify its interpretation of CSE Rule 12.6 that if a dealer holds a customer buy order and a customer sell order that can be crossed, the dealer should cross those orders without interposing itself as dealer. Fourth, the Exchange proposed to codify that dealers should establish and adhere to fixed standards for queuing and executing customer orders. Fifth, the Exchange restated that, consistent with ITS trade-through and block policies, dealers would continue to be prohibited from trading at a price outside the ITS/BBO without satisfying the market interest at that price.

In approving the CSE's request for a further extension of the preferencing program, [See Securities Exchange Act Release No. 34493, supra note .] the Commission noted its concerns about the implications of the CSE program; namely, its effect on quote competition and the risk of market fragmentation through increased internalization of order flow. The Commission believed, however, that the proposed revisions and clarifications sought to be adopted by the CSE were substantive steps that addressed the Division's concerns discussed in the Market 2000 Study regarding minimum order handling principles to ensure that customer orders are treated fairly. In particular, the Commission believed that the CSE's limit order display policy would ensure that customer orders were exposed, eligible for price improvement, and added to the overall transparency of the marketplace. Accordingly, the Commissionbelieved that the adoption of this policy addressed a number of the concerns made by commenters regarding the CSE's preferencing program with respect to order exposure on the CSE. [In particular, commenters had alleged that CSE dealers did not place their limit orders on the CSE book, but rather held such orders "upstairs" until the orders became marketable, and thus disadvantaged customers. The Commission noted that even if limit orders were not placed on the CSE book, the limit order display policy would require the member to expose all or part of a customer limit order to the market, unless it was executed immediately or the customer expressly requested that the order not be exposed. See Id.]

As a condition of extending the pilot program, the Commission enumerated six categories of data that the CSE was to submit quarterly in order to facilitate the Commission's evaluation of the CSE preferencing program. [The six categories of data were the following: (1) a list indicating how many preferencing dealers are on the exchange and the issues in which they make markets; (2) the volume of the preferenced trades and shares and the percentage of total CSE trade and share volume that preferenced activity represents; (3) the CSE's volume attributable to ITS commitments sent by other ITS participants and the percentage of such commitments in issues which have preferencing dealers; (4) the number of preferenced orders effected against the NSTS limit order book; (5) the percentage of time that the CSE improves and/or matches the NBBO, and the percentage of time that the CSE achieves price improvement for customer orders when the spread is wider than 1/8 point; and (6) the CSE's share of Consolidated Tape trade reports, as compared to its share on August 5, 1994. Id.] Moreover, the CSE was required to submit a report detailing how the program affected the quality of the CSE's market, including its effect on quote competition, market transparency, depth and liquidity. In this regard, the CSE was instructed to analyze the effects of the program, as amended by the enhancements incorporated along with the extension of the program, on the quality of market making by CSE preferencing dealers, anddemonstrate that the program resulted in added depth and liquidity to the CSE market and improved quotations. The Commission made clear that if this showing could not be made, it would not be inclined to further extend the preferencing program. [Id.]

Consequently, the Commission approved the extension of the CSE preferencing pilot to May 18, 1995, to coincide with the expiration of the BSE's competing specialist pilot, thus allowing the Commission to consolidate its future review of the preferencing programs of the various exchanges. The CSE subsequently filed the required quarterly data to the Commission, as well as the report analyzing such data. [See Letters from David Colker, Executive Vice President and Chief Operating Officer, CSE, to Arthur Levitt, Jr., Chairman, SEC, dated January 18, 1995 ("CSE January 1995 Report"); to Jonathan Katz, Secretary, SEC, dated June 19, 1995 ("CSE June 1995 Data"); to Richard Lindsey, Director, Division of Market Regulation, SEC, dated January 31, 1996 ("CSE January 1996 Data"). In June 1995, the CSE also submitted a report that analyzed CSE activity for a six day period in 237 stocks that only had preferencing dealers. See Letter from David Colker, CSE, to Jonathan G. Katz, Secretary, SEC, dated June 15, 1995 ("CSE June 1995 Report").]

C.CSE Request for Permanent Approval of the Preferencing Pilot Program

In March 1995, the CSE submitted a second proposed rule change to have its preferencing program approved on a permanent basis. This proposal subsequently was published for notice and became the subject of extensive commentary. [See Securities Exchange Act Release No. 35448 (March 7, 1995), 60 FR 13493 (March 13, 1995) (File No. SR-CSE-95-03).] In August 1995, the CSE amended the proposed rule change to propose the adoption of threeorder handling policies designed to increase order exposure on the Exchange and ensure the timely execution and display of limit orders held by CSE dealers. [See Securities Exchange Act Release No. 36092 (August 11, 1995), 60 FR 42209 (August 15, 1995) (File No. SR-CSE-95-03) (notice of Amendment No. 1 to CSE request for permanent approval of its preferencing program).]

The first proposed order handling policy, applicable in greater than minimum variation markets, requires CSE preferencing dealers to either immediately execute a market order routed to it for execution on the CSE at an improved price or expose the order on the Exchange for a minimum of 30 seconds to allow other market participants an opportunity to provide an improved price. [See CSE Rule 11.9(u), Interpretation and Policy .01. In exposing a market order on the Exchange for price improvement, a dealer "stops" the order ( i.e. , guarantees the execution of the order at the then-best inter-market price) in the event that the order does not receive price improvement. As part of its Order Execution Obligations proposals, the Commission had proposed a rule requiring that in markets where the difference between the best bid and offer is greater than the minimum trading variation, all market orders receive an opportunity for price improvement. See Securities Exchange Act Release No. 36310 (September 29, 1995), 60 FR 52792 (October 4, 1995) (File No. S7-30-95) (Order Execution Obligations Proposing Release) ("OEO Proposing Release"). At the present time, however, the Commission has deferred taking action on the price improvement proposal. See Securities Exchange Act Release No. 37619A (September 6, 1996), 61 FR 48290 (September 12, 1996) at n.357 (File No. S7-30-95) (Order Execution Obligations Adopting Release) ("OEO Adopting Release").] To comply with the policy, a preferencing dealer may either represent the order at an improved price in his or her CSE quote, or place the order on the CSE's central limit order book at an improved price. Second, the CSE proposed to adopt a limit order price protection policy applicable to preferencing dealers, designed toensure that limit orders routed to CSE dealers for execution on the CSE receive timely executions relative to same-priced limit orders on the primary market for the security, [See CSE Rule 11.9(u), Interpretation and Policy .02. Under this policy, a public agency limit order routed to a CSE preferencing dealer for execution on the CSE would be filled if (i) the bid or offer at the limit price has been exhausted in the primary market; (ii) there has been a price penetration of the limit order in the primary market; or (iii) the stock is trading at the limit price on the primary market unless it can be demonstrated that such order would not have been executed if it had been transmitted to the primary market. The customer and the CSE dealer may agree to a specific volume related or other criteria for requiring execution of limit orders. ] and substantially similar to limit order price protection policies that have been adopted by other regional exchanges. [The CHX and the BSE currently have limit order price protection policies nearly identical to the CSE's. See BSE Rules, Ch. II, Sec. 33; CHX Art. XX, Rule 37(a).]

Third, the CSE proposed to amend its public agency limit order display policy, contained in Interpretation and Policy .01 to CSE Rule 12.10. [For a description of the CSE's then-existing policy, see Securities Exchange Act Release No. 34493, supra note . See also supra note . The CSE's policy as amended subsequently has been superceded by the adoption of the Display Rule. See Id. ] Under this amended policy, all CSE dealers were required to display on the CSE all or a representative portion of public limit orders that they represented as agent for execution on the CSE priced at or better than the ITS/BBO. [If the limit order was for 500 shares or less, a CSE dealer was required to display the entire order; however, if the limit order was for more than 500 shares, a dealer was required to display a representative portion equal to at least 500 shares plus 10% of the size of the order.] A dealer couldsatisfy this requirement either by representing the orders in their CSE quotes or placing the orders on the CSE's central limit order book. [In addition, if a representative portion of an order was executed, the CSE dealer was required to display all or a representative portion of the remainder of the order until the order was filled in its entirety.]

Generally, commenters opposed to the CSE's preferencing program argued that the program (1) decreases order interaction and consequently has a negative effect on execution quality on the CSE, and (2) detrimentally affects the quality of both the CSE market and the larger NMS. Specifically, commenters argued that, by altering the CSE's priority rules to facilitate internalization, the CSE preferencing program had the result of discouraging interaction among CSE dealers, thereby denying customer orders the opportunity to interact with other trading interest in the market. [In this regard, the NYSE conducted its own analysis of trading on the CSE for five consecutive trading days in March 1995, and concluded that as a result of preferencing CSE members can internalize order flow with minimal probability of that order flow interacting with the orders of other CSE members. See Securities Exchange Act Release No. 37046 (March 29, 1996), 61 FR 15322 (April 5, 1996) (File No. SR-CSE-95-03) (order permanently approving CSE's preferencing program) ("CSE Approval Order"). In response, the CSE indicated that an average of 8,000 trades per month resulted from interaction among CSE dealers. See CSE January 1995 Report, supra note . In addition, the CSE believed interdealer activity had been increasing due to efforts by the Exchange to encourage quote competition among its dealers through its Quality of Markets proposal, which had not yet been approved by the Commission, but which the CSE asserted that many of its dealers began to comply with voluntarily in January 1994. See Securities Exchange Act Release No. 35268 (January 24, 1995), 60 FR 5951 (January 31, 1995) (File No. SR-CSE-95-01) (notice of CSE Quality of Markets proposal). The CSE's Quality of Markets proposal would establish a one-year pilot program to: 1) require that specialists maintain quotation spreads that do not exceed 125% of the average of the three best (narrowest) quote spreads provided by all markets that participate in the national market system; and 2) prohibit dealers from furnishing quotations that are generated by automated quotation tracking systems. Id.] Commenters further argued that preferencing underminesthe proper price discovery and market transparency functions of the agency auction market and makes best execution of customer orders less likely. [See CSE Approval Order, supra note 25. In this regard, commenters asserted that preferencing makes it more profitable for dealers to internalize orders by maintaining limit orders on their internal proprietary systems until they become marketable, rather than placing them on the Exchange's central limit order book where they would be displayed and have an opportunity to interact with other customer orders. Id. In addition, commenters charged that because orders on the CSE's central limit order book must be satisfied before a dealer can internalize orders at the same price, it is to the advantage of dealers that seek to internalize customer order flow to discourage the placing of limit orders on the CSE's book. Id. The CSE asserted that it had encouraged dealers to place limit orders on the NSTS book, but that no exchange has the authority to dictate firm order handling practices by requiring that firms place their limit orders in the exchange's book. In addition, the CSE reported that in the first quarter of 1995, 2104 preferenced orders interacted with pre-existing public agency limit orders on the CSE's book, while in the fourth quarter of 1995, 4802 preferenced orders interacted with agency limit orders on the CSE's book. See CSE June 1995 and January 1996 Data, supra note . ] The CSE responded that, to the contrary, the rate of price improvement on the Exchange compares favorably to that on other exchanges. [The CSE asserted that 59% of CSE executions in greater than minimum variation markets were printed between the ITS/BBO in the fourth quarter of 1995, and that an additional 4% of the orders received price improvement after being exposed at prices that narrowed the ITS/BBO. See CSE January 1996 Data, supra note . The CSE determined that the respective figures for the first quarter of 1995, were 57% and 3%. See CSE June 1995 Data, supra note . The NYSE maintained, however, that based on its own analysis for the month of December 1994, the CSE's rate of price improvement was only 48.7% in 1/4 point markets, 48% in 3/8 point markets, and 37.7% in 1/2 point markets. See CSE Approval Order, supra note 25, at n.50.]

Commenters also asserted that the CSE's preferencing program should not be permanently approved because the Exchange had not made the demonstration previously required by the Commission [See Securities Exchange Act Release No. 34493, supra note .] that preferencing results in added depth and liquidity to the CSE market, or improved quotations; in fact, some commenters even asserted that the existence of preferencing dealers in a stock actually diminished the quality of CSE quotes. [See Id. at 19-20. But see also Letter from Robert Jennings, Faculty Fellow and Professor of Finance, Indiana University School of Business, to Jonathan G. Katz, Secretary, SEC, dated June 30, 1995; and Letter from Robert Battalio, Asst. Professor, University of Notre Dame, to Jonathan G. Katz, SEC, dated March 6, 1996 (collectively, "IU Study"). The IU Study analyzed the short term effects of preferencing on market quality. Its preliminary results indicated that, while internalization results in significant volume redistribution, the preferencing program does not appear to have had an adverse effect on the bid/ask spreads and liquidity premiums ( i.e. , the measurement of the closeness of transaction prices to the mid-point of the spread). In addition, the IU Study noted that to the extent retail brokers that internalize trades reduce (or even eliminate) commissions, investor welfare is improved. Some commenters maintained that there was no evidence to indicate that the purported efficiencies of internalization were passed along to CSE dealers' customers. See CSE Approval Order, supra note 25, at n.73.] The CSE, however, maintained that preferencing dealers add depth and liquidity to the market through quotes at the ITS/BBO. In this regard, the CSE conducted two separate studies, one including allCSE trading and the second limited to preferencing dealer trading, in which it analyzed average quote spread, average quote size, the relation of CSE quotes to the ITS/BBO, total trade activity, ITS inbound trade activity, and customer order price improvement. [See CSE January and June 1995 Reports, supra note .] The CSE reported that in every category, the performance of both the CSE as a whole and preferencing dealers in particular either equaled or exceeded that of the other regional exchanges. [Specifically, for a subset of 237 stocks which were traded only by preferencing dealers, the CSE found that (1) the average quotation spread was 1/4 point, which the CSE reported was narrower than any other regional exchange, (2) 60% of preferencing dealers' quotes matched one or both sides of the ITS/BBO, while none of the other exchanges exceeded 30%; and (3) the CSE generated 4% of all quotes that established a new ITS/BBO, which exceeded the performance of all other regional exchanges. See CSE June 1995 Report, supra note . The CSE asserted that it had the highest average quote size (700 shares) of any regional stock exchange in the 237 stocks; moreover, the CSE claimed that the depth provided by preferencing specialists when their quotes establish or match the ITS/BBO actually contributed more depth to the national market than all the other regional exchanges when viewed in conjunction with the lower rates at which the regional exchanges quote at the ITS/BBO. Id. In addition, the CSE maintained that preferencing dealers executed almost half of all ITS inbound activity in those issues that have at least one preferencing dealer, and that preferencing dealers' ITS/total trade ratio of 3.5% compared favorably to the NYSE's ITS/total trade ratio of 2.8%. Id. In its later data submissions to the Commission, the CSE reported that in the first quarter of 1995, 73% of the CSE's quotations matched at least one side of the ITS/BBO, while the respective number for the fourth quarter of 1995 was 71%. See CSE June 1995 and January 1996 Data, supra note . The CSE also reported that in each of these two quarters it was responsible for generating 6% of all quotes that established a new ITS/BBO. Id.]

The NYSE and Amex asserted, however, that any liquidity that may be provided by the CSE is artificial, claiming that CSE quotes change too quickly for other market participants to have a meaningful opportunity to send ITS orders to the CSE. [See CSE Approval Order, supra note 25, at n.65. These commenters believed that the rapid quote changes are caused by a combination of multiple dealers in the single market and the use by some CSE dealers of automated systems that generate quotations. Id. In addition, the NYSE criticized the market order price improvement policy by contending that (1) it could affect as few as 8% of preferenced orders; (2) preferencing dealers could continue to trade with customer orders at the improved price, negating the opportunity for two customer orders to meet without dealer intervention; and (3) that the 30 second exposure period was unlikely to provide other market participants sufficient time to trade with exposed orders. See Id. at 25.] The CSE maintained, however, that 87% of quote changes that result in cancellations of inbound ITS commitments were displayed to other market participants for over one minute, and more than 50% of these quotes had been displayed for over five minutes. [See Letter from David Colker, Executive Vice President and Chief Operating Officer, CSE, to Jonathan G. Katz, Secretary, SEC, dated April 26, 1995.]

In response to the differing assertions made by the commenters and the CSE, the Commission's OEA evaluated CSE quotations and transactions before approval of the CSE filing. OEA's analysis found that CSE preferencing dealers often matched the NYSE BBO, and

that the percentage of time that the CSE quotes matched those on the NYSE was greatest for those stocks in which preferencing takes place. [During the time period considered in OEA's study, preferencing dealers accounted for more than 90% of trades and 66% of share volume on the CSE. See CSE Approval Order, supra note 25, at 26-27. For the 281 stocks in which preferencing dealers accounted for 80% to 99% of total CSE trades, the CSE quote on average matched the NYSE best bid approximately 54% of the time and the NYSE best offer nearly 61% of the time. Id. When matching at least one side of the ITS/BBO, the CSE's quotation depth in these 281 stocks averaged over 720 shares. For all quotes in the 281 stocks, the CSE quotation depth averaged close to 900 shares. Id.]

Appendix B

BSE Competing Specialist Initiative Pilot Program (1994 - 1996)

A.Commission Approval of the Initial BSE CSI Pilot Program in 1994

The BSE submitted its proposal to adopt the CSI in June 1993. [Securities Exchange Act Release No. 32549 (June 29, 1993), 58 FR 36229 (July 6, 1993) (File No. SR-BSE-93-12) (notice proposing initial adoption of BSE's CSI pilot program).] The Commission approved the CSI on an initial one year pilot basis in May 1994. [Securities Exchange Act Release No. 34078 (May 18, 1994), 59 FR 27082 (May 25, 1994) (File No. SR-BSE-93-12) (order approving BSE's CSI on an initial one year pilot basis). Concurrently with its approval of the CSI pilot program, the Commission approved Chinese Wall procedures and added a BEACON liability interpretation for competing specialists. See Securities Exchange Act Release Nos. 34076 (May 18, 1994), 59 FR 26822 (May 24, 1994) (File No. SR-BSE-93-17) (order approving adoption of Chinese wall procedures for competing specialists); 34075 (May 18, 1994), 59 FR 26821 (May 24, 1994) (BSE File No. SR-BSE-94-03) (order approving adoption of BEACON liability interpretation).] In doing so, the Commission cited its support for exchange efforts to provide increased market making and competition on their trading floors, maintaining that such efforts should increase the provision of liquidity services by an exchange and enable the exchange to compete more effectively with other markets. The Commission acknowledged that the CSI could assist the Exchange in this regard, but noted its sensitivity to the concerns of the commenters as to the structural implications of the proposal. Accordingly, the Commission believed that it was important that the CSI "provide real quote competition for the benefit of investors, not simply a means for firms to internalizetheir customer order flow while receiving specialist designation and treatment." [See Securities Exchange Act Release No. 34078, supra note 2.]

The Commission specifically noted the Exchange's assertion that the proposal would provide an incentive for BSE specialists desiring to attract order flow to quote at the ITS/BBO more frequently. The Commission stated that if this assertion were correct, the proposal potentially could add depth at the ITS/BBO if BSE specialists begin to quote at the best ITS price more frequently, and thereby improve the overall quality of the BSE market.

With regard to the proposal's effect on the internalization of order flow, and the potential for resultant market fragmentation, the Commission recounted the Division's findings in the Market 2000 Study that multiple, varied market centers and competitors for listed stocks had provided many benefits to users of the markets without impairing market quality or price discovery, and that concerns about fragmentation had been overstated. [See Market 2000 Study, supra Appendix A at note 10, Study III at III-1 to III-3.] The Commission was concerned that the implementation of a preferencing system at one or more exchanges might alter the analysis. However, the Commission stated that it was not willing to conclude at that time that the increased internalization that might result from the BSE proposal would be detrimental to the quality of the equity markets. The Commission emphasized that the proposal was limited in scopeand that customer orders would continue to be able to interact and retain the opportunity for price improvement. As a result, the Commission believed that the BSE proposal was appropriately structured to be implemented on a one year pilot basis.

However, the Commission believed that before the pilot could be extended or approved on a permanent basis, the BSE would have the burden to prove that the CSI resulted in added depth and liquidity to its market and improved quotations. Accordingly, the Commission requested data submissions from the BSE on a quarterly basis. [The quarterly data submissions encompassed the following: (1) the number of competing specialists on the Exchange; (2) the issues in which each competing specialist is specializing; (3) the volume of "preferenced" trades in each stock, and the percentage of total volume of trades sent to the BSE for execution in each issue represented by "preferenced" volume; (4) the volume of "preferenced" share and transaction volume, and the percentage of total BSE share and transaction volume the "preferenced" volume represents; (5) BSE's volume attributable to ITS commitments sent by other ITS participant markets, and the percentage of such commitments attributable to preferenced stocks; (6) the number of preferenced orders executed by the non-designated specialist(s); (7) the number of preferenced orders effected against the BSE limit order book, both in total and separately for each preferenced stock; (8) the number of limit orders placed on the book, as compared to before the commencement of the pilot, both for the BSE as a whole and separately for the stocks for which there are competing specialists; and (9) BSE's share of Consolidated Tape trade reports, as compared to before the commencement of the pilot.] In addition, the Commission requested the BSE to provide the Commission with a report on the operation of the pilot discussing such data in terms of the effects of the CSI on the quality of the BSE market with respect to depth, liquidity, and quote improvement. The Commission emphasized that if it was dissatisfied with the data or the market effects of the pilot, itwould not be inclined to extend the operation of the CSI.

The BSE subsequently filed its required quarterly data reports with the Commission, as well as the report analyzing such data. [See BSE, Competing Specialist Initiative Report, submitted to the Commission on February 13, 1995 ("BSE February 1995 Data and Report"); Letter from Karen Aluise, Assistant Vice President, BSE, to N. Amy Bilbija, Attorney, SEC, dated April 28, 1995 ("BSE April 1995 Data"); and Letter from Karen Aluise, BSE, to Glen Barrentine, Senior Counsel, SEC, dated February 14, 1996 ("BSE February 1996 Data").]

B.Extensions of BSE's CSI Pilot Program

On February 6, 1995, the BSE submitted to the Commission a proposal to adopt its CSI on a permanent basis. While considering the BSE's permanent approval request, the Commission approved two interim extensions of the CSI pilot program. [See Securities Exchange Act Release Nos. 35716 (May 15, 1995), 60 FR 26908 (May 19, 1995) (File No. SR-BSE-95-07) (extension of BSE CSI pilot through October 2, 1995); and 36323 (September 29, 1995), 60 FR 52440 (October 6, 1995) (File No. SR-BSE-95-14) (extension of BSE CSI pilot through March 29, 1996 with clarification that BEACON limit orders are to be executed in strict time priority, allowing up to four competing specialists per stock, and increasing to 100 the maximum number of stocks in which each competing specialist could register).] In the second extension order, the Commission noted that while the BSE's data submissions indicated that the volume of orders directed to the BSE and the depth of the BSE limit order book in competing specialist stocks had increased during the pilot, the data as to quote competition among competing specialists was mixed. [See Securities Exchange Act Release No. 36323, supra note 7. ] In this regard, while the data showed an increase in the volume of ITS commitments to the BSE in competing specialist stocks, it alsoindicated a lack of interaction by the regular specialist with incoming orders directed to a competing specialist. [In this regard, orders directed to a particular competing specialist were almost always executed by that competing specialist. See BSE April 1995 and February 1996 Data, supra note 6 (data indicating regular specialist intercepts less than one percent of the order flow directed to competing specialists). Further, if BSE specialists were quoting at the ITS/BBO, they would at times have priority over the designated specialist and intercept some of the directed order flow. See supra Part V at Table V-5 (BSE specialists quote either one or both sides of the ITS/BBO only five percent of the time). ] Consequently, the Commission remarked that while this data was inconclusive as to the program's effect on quote competition, it believed that certain systems enhancements could facilitate quote competition on the BSE. [See supra Part I.C.2.]

Appendix C

Development of the Cincinnati Stock Exchange's

National Securities Trading System (1977 - 1990)

In June 1977, the CSE submitted to the Commission a proposal to operate the Multiple Dealer Trading System ("MDTS"), the predecessor of the current system, for up to 200 stocks on a nine month pilot basis. The MDTS accepted limit orders in system-designated stocks from two categories of system users, Approved Dealers and access non-members of the Exchange. [Market orders could not be entered into the MDTS, as the system algorithm required that a price had to be assigned to an order prior to its entry into the system. The System was enhanced in the early 1980s to permit the entry of market orders.] The execution of orders on the MDTS was governed strictly according to price and time priority, except that public agency limit orders, [A public agency order was defined as "any order for the account of a person other than an Approved Dealer, a Member, or a person who could become an Approved Dealer by complying with this Rule with respect to his use of the [CSE] System, which order is represented, as agency, by a User."] regardless of their time of entry, were granted priority over other limit orders at the same price.

In its order initially approving the MDTS pilot, the Commission noted its responsibilities under the Act to ensure the development and maintenance of a fair field of competition among brokers and dealers and among securities markets, and concluded that the CSE System pilot was the type of "competition enhancing" development that should be permitted in light of the Congressional mandate in the 1975 Amendments to the Act for the Commission tofacilitate the development of a national market system ("NMS"). [See Securities Exchange Act Release No. 14674 (April 18, 1978), 43 FR 17894 (April 26, 1978) (File No. SR-CSE-77-1) (order approving MDTS pilot program on an initial nine month basis). Trading in the CSE System did not commence until June 6, 1978.] The Commission recognized that the MDTS provided new opportunities for large retail firms, by making markets on the system, to deal directly with their customers on a principal basis, but nonetheless noted that it was possible that the MDTS could be used by dealers to internalize order flow without affording other users an adequate opportunity to interact with such order flow. However, the Commission noted that the MDTS offered a continuing opportunity for users to achieve priority over dealers by entering buying or selling interest at a given price earlier than a dealer and provided customers of users with absolute priority over all dealer interest.

In December 1978, the Commission approved a further extension of the MDTS pilot through January 31, 1980. [See Securities Exchange Act Release No. 15413 (December 15, 1978), 44 FR 129 (January 2, 1979) (File No. SR-CSE-78-2) (order extending MDTS pilot through January 31, 1980).] In response to commenters' concerns that MDTS offered CSE dealers an opportunity to internalize their order flow, the Commission concluded that based on the data available at that time, it had not been able to discern any adverse impact to investors or the markets with respect to stocks traded in the MDTS. Subsequently, the CSE filed a proposal in July 1979 to extend the operation of the MDTS for anadditional three years, through January 31, 1983. [See Securities Exchange Act Release No. 15999 (July 6, 1979), 44 FR 40745 (File No. SR-CSE-79-3) (notice of proposal to extend the MDTS pilot through January 31, 1983).] The CSE indicated that extending the pilot for an additional three years would provide a more meaningful opportunity for participation in the MDTS. [Id. In this regard, the CSE indicated that a three year extension of the pilot would allay concerns on the part of the CSE, its systems administrator, and potential users as to the possibility of the abrupt termination of the pilot before the costs of enhancements to, or participation in, the MDTS could be recouped.]

In approving the extension of the CSE System through January 31, 1983, [See Securities Exchange Act Release No. 16215 (September 21, 1979), 44 FR 56074 (September 28, 1979) (File No. SR-CSE-82-1) (order approving extension of MDTS pilot program for up to 200 stocks through January 31, 1983).] the Commission stated that the MDTS possessed many basic features that were consistent with the statutory objective of an NMS, including increased speed and efficiency in trading, the ability to generate an audit trail for surveillance purposes, and an automated mechanism for linking upstairs market makers with one another, as well as with exchange specialists. However, the Commission also stated its intention to continue to monitor the operation of the CSE System and analyze both its trading and the market structure implications, if any, of firms utilizing the CSE System to internalize their retail customer order flow. The Commission further stated its readiness "to take remedial action should such trading have adverse effects upon the markets or uponinvestors." [Id.]

In addition, the Commission noted that the development of a NMS continued to be an evolutionary process, and if the CSE sought to secure indefinite approval of the MDTS as a feature of the NMS, it must "continue to make improvements, changes and adaptations to meet the needs of various market centers and to accommodate Commission regulatory requirements designed to improve order interaction and price protection between and among markets." [The Commission stated its intention to continue consideration of the significant market structure questions resulting from: (i) the ability of firms, in a variety of contexts, to transact business on a principal basis with their own retail customers, or to otherwise combine principal and agency functions for a particular security; and (ii) the need to ensure that order flow in every particular market center is exposed to buying and selling interest represented in other market centers. Id. The Commission noted that such questions should be addressed in a broad and generic context rather than in connection with a proposed rule change filed by a single SRO; however, the Commission added that such deliberations may well lead to regulatory initiatives which have a significant impact on such systems as the MDTS. Id.] In this regard, the Commission indicated that at a minimum, the CSE should become a participant in the Consolidated Quotation Plan ("CQ Plan"), [As of September 1979, quotations from all market centers (including third market makers) except the CSE were being made available in a single consolidated data stream processed by the Securities Industry Automation Corporation ("SIAC"). See Securities Exchange Act Release Nos. 15009 (July 28, 1978), 43 FR 34851 (order declaring CQ Plan temporarily effective); 15511 (January 24, 1979), 44 FR 6230 (order extending CQ Plan for an additional 12 months). Although at this time CSE System quotations were available on CSE System terminals and available to vendors pursuant to Rule 11Ac1-1, the Commission indicated that some vendors had been reluctant to include CSE System quotations in displays provided to subscribers because of the higher costs to such vendors of processing quotations which are not transmitted to them in a single data stream. See Securities Exchange Act Release No. 16215, supra note 7.] and take immediate steps in implementing a linkagebetween the ITS and MDTS in order to allow access between the ITS market centers and MDTS. [The Commission explained that for the CSE System to be consonant with the objectives of a fully-operational NMS, it must be linked to traditional exchange markets in order to create the maximum degree of order interaction between the two different types of markets. Moreover, the Commission believed that a linkage between ITS and the CSE System was essential to achieving nationwide protection for public limit orders, and potentially the only feasible means of implementing the Commission's limit order protection proposal then under consideration (subsequently withdrawn). See Securities Exchange Act Release No. 15770 (April 26, 1979), 44 FR 26692 (File No. S7-32-92) (Rule 11Ac1-3 Proposing Release). ] The CSE became a participant in the CQ Plan on November 15, 1979 [See Letter from K. Richard R. Niehoff, President, CSE, to George A. Fitzsimmons, Secretary, SEC, dated November 16, 1979.] and an ITS participant in February 1981. [See Securities Exchange Act Release Nos. 17532 (February 10, 1982), 46 FR 12919 (February 18, 1982) (File No. 4-208) (implementation of a manual interface between NSTS and ITS); 17531 (February 10, 1981), 46 FR 12920 (February 18, 1981) (File No. SR-CSE-81-1) (order approving rules to implement and govern CSE participation in ITS Plan). Additionally, in 1980 the CSE implemented an automatic interface between the NSTS and the National Securities Clearing Corporation ("NSCC") for the clearing of all NSTS trades. See Securities Exchange Act Release No. 16883 (June 10, 1980) (File No. SR-NSCC-77-5) (order approving implementation of automatic interface between the CSE and the NSCC).]

The MDTS was renamed the National Securities Trading System ("NSTS") by the CSE in January 1981. In May 1981 and September 1982, the Commission released monitoring reports on the operationof the NSTS from its inception in June 1978 through mid-1982. [SEC, Monitoring Report on the Operation of the Cincinnati Stock Exchange National Securities Trading System, May 1981 ("1981 Monitoring Report"); SEC, A Report on the Operation of the Cincinnati Stock Exchange National Securities Trading System: 1978 - 1982 ("1982 Monitoring Report"). The 1981 and 1982 Monitoring Reports stated that the number of stocks traded in the NSTS increased during the pilot as follows: 38 in June 1978; 41 in December 1979; 49 in December 1980; 73 in December 1981; and 88 in August 1982. Moreover, the number of participants in the NSTS increased from: eight in December 1980; 11 in December 1981; and 12 in August 1982. ] These reports indicated that although CSE trade and share volume had increased since the inception of NSTS, it still remained only a small percentage of consolidated volume in NSTS-traded stocks.

Table C-1: NSTS Volume and Composite Volume in NSTS

Stocks - 1979 to June 1982 (in Thousands of

Shares) [ Source: 1982 Monitoring Report, supra note 14, at Table 1.]

1,483,186

Further, the 1982 Monitoring Report indicated that the NSTS had not had any discernable effect on the primary market spread of NSTS securities and that NSTS securities were characterized by a higher average quote size in the primary market relative to non-NSTS securities.

On December 9, 1982, the Commission terminated the status of the NSTS as an experimental program by extending its duration for an indefinite period and eliminating the 200 stock limitation on the number of stocks which could be traded in NSTS. [See Securities Exchange Act Release No. 19315 (December 9, 1982), 47 FR 56236 (December 15, 1982) (File No. SR-CSE-82-1) (order approving NSTS for an indefinite period).] The CSE contended in its proposal that the experimental status of the NSTS had hurt the Exchange in competing on an equal basis with other market centers. [The CSE indicated that new users may have been discouraged from incurring the start-up costs inherent in participating in the NSTS in view of the System's temporary status. Id. The CSE further contended that its experimental status had made it difficult for both the Exchange and its systems administrator to plan effectively for long term NSTS enhancements such as expansion of System capacity and an automated interface between NSTS and ITS. Id.] In its approval order, the Commission stated that NSTS represented a constructive approach to integrating trading in physically dispersed locations, and characterized the data in the 1982 Monitoring Report as indicating that the NSTS provided its users, particularly NSTS market makers, with an effective mechanism for executing orders and attracting order flow. [Id. The 1982 Monitoring Report had indicated that in a sample of five months during the initial period of the NSTS - ITS interface, the CSE had been a net importer of ITS share volume, executing orders for 5.9 million shares from other market centers versus receiving executions for orders amounting to 2.9 million shares from other markets. See 1982 Monitoring Report, supra note 14, at Table 5. ] Further, the Commission noted that the 1982 Monitoring Report had indicated that the degree to which NSTS firms executedorders on an intra-firm basis [Intra-firm trading consists of trading in which a single upstairs Approved Dealer firm appears on both sides of the same transaction, either as agent for one side and dealer on the contra-side, or as agent for both sides.] had declined, [The 1981 and 1982 Monitoring Reports analyzed the amount of intra-firm activity on the NSTS. The 1982 Monitoring Report indicated that for a six month period from October 1981 to March 1982, 75% of total NSTS agency share volume resulted from intra-firm executions, 47% of this amount consisted of principal/agency transactions, and the other 28% consisted of agency crosses. See 1982 Monitoring Report, supra note 14, at 18. This represented a decline from the 1981 Monitoring Report's review of four sample weeks of trading that found that 89% of total NSTS agency share volume resulted from intra-firm transactions, 63% principal/agency transactions and 26% agency crosses. See 1981 Monitoring Report, supra note 14, at Table 5. The SEC did not find these results surprising given that the largest upstairs dealer accounted for 44% of NSTS volume in 1982 and 83% in 1980, and the NSTS had not been able to attract significant agency order flow from other NSTS participants. See 1982 Monitoring Report, supra note 14, at 20.] and, referencing its then-outstanding order display proposal, stated that internalization should be addressed in a broad and generic context rather than in connection with a proposed rule change of a single SRO. [See Securities Exchange Act Release No. 18738 (May 13, 1982), 47 FR 22376 (May 24, 1982) (File No. S7-930) (Order Exposure Rule Proposing Release).]

In 1986, the ITS participants agreed to an amendment of the ITS Plan that provided for an automated interface between ITS and NSTS. [See Securities Exchange Act Release Nos. 22988 (March 7, 1986), 51 FR 8927 (March 14, 1986) (File No. 4-208) (Joint Industry Plan; Filing and Temporary Summary Effectiveness of Amendment to the ITS Plan); 23365 (June 23, 1986), 51 FR 23865 (July 1, 1986) (File No. 4-208) (Joint Industry Plan; Approval of an Amendment to the ITS Plan Relating to an Automated Interface with the Cincinnati Stock Exchange, Inc.); and 22430 (September 30, 1985), 50 FR 39209 (September 27, 1985) (File No. SR-CSE-85-4) (order approving amendment to CSE rules to implement an automated interface with ITS). The automated interface allows orders entered into NSTS, processed in accordance with CSE rules, but not fully executed in NSTS, to be formatted automatically into an ITS commitment and routed to the ITS participant market displaying the ITS/BBO at that time. The interface also automatically accepts an ITS commitment routed to the NSTS from another market if the NSTS quotation is at the same price as the ITS commitment when that commitment is received. Further, in response to ITS participant comments, the CSE modified the NSTS to automatically provide an incoming ITS order a better execution price than the stated order price if the CSE quotation improved after the order was transmitted to the NSTS. ] Later that same year, the CSE entered into an agreementto provide the Chicago Board Options Exchange ("CBOE") with a proprietary interest in the CSE. [See Securities Exchange Act Release No. 24090 (February 12, 1987), 52 FR 5225 (February 19, 1987) (File No. SR-CSE-86-6) (order approving changes in CSE rules to accommodate CSE-CBOE affiliation).] The CSE subsequently moved its operations to the CBOE's facilities in Chicago, Illinois.

In 1989, the Commission approved a CSE proposal to create a primary Designated Dealer status in stocks that were not traded in NSTS prior to the adoption of the proposal. [See Securities Exchange Act Release No. 27220 (September 5, 1989), 54 FR 37856 (September 13, 1989) (File No. SR-CSE-89- 01) (order approving adoption of CSE primary Designated Dealer status).] This status entitles a Designated Dealer to receive all of the guaranteed portion of market and marketable limit orders in such stocks (i.e., up to the lesser of 2099 shares or the quoted size of the stock at the ITS/BBO at the time the order is entered into the System), in conjunction with its agreement to maintain "competitive" quotationsthroughout the trading day, [The CSE represented that the criteria for determining whether to appoint a CSE dealer as primary designated dealer in a non- NSTS stock included a commitment to quote tight or large-sized markets in order to attract order flow to the Exchange, but in no way required the applicant's firm to direct order flow to the CSE. Id. at n.7.] regardless of whether other Designated Dealers subsequently become registered in the stock. The Exchange stated that the proposal would provide an incentive for CSE dealers to become Designated Dealers in stocks not yet traded in NSTS, allowing the Exchange to achieve broader coverage of issues in NSTS, while improving its markets. In approving the proposal, the Commission noted that the primary Designated Dealer status would result in the grafting of elements of a specialist system onto the NSTS for such stocks, which may facilitate the execution of customer orders and provide tight and liquid markets. [Id.]

In proposing its preferencing pilot program in 1990, the CSE stated that the primary Designated Dealer status initiative had not been successful either in attracting trading in additional issues on NSTS or overcoming the lack of incentive in the CSE's multiple dealer environment for a dealer to direct its own retail order flow

to the Exchange. Indeed, the CSE remained the smallest of the national securities exchanges through the latter part of the 1980s, as measured by the percentage of both Consolidated Tape trades and volume transacted on the Exchange.

Table C-2:CSE Consolidated Tape Network A Volume and

Trades (NYSE-listed stocks) - 1985 to 1990 [Source: NYSE 1995 Fact Book (1995) at 27-28.]

Appendix D: Detailed Data Description

General

Order and transaction data were obtained from the NYSE and from the BSE, CHX, CSE, Phlx and PSE for the four week period from October 28 to November 22, 1996. Included in the data is information such as the order arrival time, the size of the original order, a buy/sell indicator, an indicator or means to identify market and limit orders, and limit order prices. Information on the execution of the order includes the time of the execution, the execution price, and the fill quantity. For orders filled in more than one piece the data contain information for each of the fills. The datasets vary in the information they provide regarding price qualifiers, settlement conditions, and tick sensitive conditions. However, each of the data were screened to the extent possible to eliminate orders with non-standard conditions attached to the orders.

An important difference in the datasets analyzed is that some exchanges' data generally include all orders received and executed on the exchange, whereas the NYSE data, PSE data and the Phlx data used in this study do not contain orders entered from the floor. In addition, the regional exchange data analyzed exclude all ITS trades because, generally, key data such as the exchange on which the trade was printed, was not identified.

The limit order analysis is based on day limit orders except for the CHX, whose data did not distinguish day limit orders fromother limit orders. However, in evaluating limit order trading costs and time from order entry to execution, only limit orders executed on the day the order was received were included in the analysis. Day limit orders expire at the end of the trading day if not filled, and account for the majority of limit orders submitted.

SIAC Data/National Best Bid and Offer

Data from the Securities Industry Automation Corporation ("SIAC") were used to establish the NBBO used in the calculations. The NBBO was merged with data from individual exchanges, and executions were evaluated against the NBBO taken from the SIAC data. If more than one market was at the NBBO, the NBBO depth represents the greatest depth of those markets at the NBBO.

NYSE Data

The NYSE data are taken from the System Order Database ("SOD") file. This file contains all orders entered via the SuperDot system, both market and limit orders. SOD does not contain orders entered into the auction from the floor or by other means. According to the NYSE 1995 Fact Book, in 1995 85% of all orders and 33% of volume went through the SuperDot system.

Because this study primarily focuses on retail orders, this limitation should not be severe, as relatively few retail-sized orders are entered from the floor.

BSE Data

BSE data are taken from the BEACON system. This system records the entry and execution of market orders. BEACON also records information about the entry and execution of limit orders that subsequently execute on the system. However, it does not retain a record of limit orders that expire unexecuted, nor does it have detailed quote records of the interaction between competing specialists in the BSE's CSI program. To get the limit order data, the BSE sent the Commission hard copy records of all unexecuted day limit orders for the period under study. These data were hand entered and merged with the BSE’ electronic data.

CHX Data

The CHX data file combines orders and executions in a single observation. Orders can be entered either electronically or from the floor. Since floor orders are not entered electronically, the order information is unavailable. [The relevant order information includes: limit price, order submission time, and variables indicating whether the trade was a buy or a sell. This information is used to affix quotes to the order, and to enable calculation of price improvement.] Additionally, order information is incomplete for ITS trades. Therefore, ITS trades and floor orders are excluded, as there is no means available to reliably calculate price improvement. The last problematic issue regarding the CHX data file concerns the instance where market orders are paired off with limit orders. In this case, the order information pertains to the limit order side of the trade. Accordingly, the order information for the market order side of the trade is absent. In this situation, the market order is considered to have no price improvement and the effective spread is the bid-ask spread.

CSE Data

The CSE is a multiple market-maker exchange. The data in this study are taken only from six of the seven CSE preferencing dealers: Prudential, Olde, Pershing, Fidelity, Piper, and Redwood. Data for the seventh preferencing dealer, Dain Bosworth, were unavailable; however, Dain Bosworth accounts for a relatively small proportion of CSE activity.

The CSE’s preferencing dealers pair orders, subject to exchange rules, at their trading desks and send the paired trades to the CSE’s facilities in Chicago for execution. The CSE recently improved its audit trail to allow limit orders to be distinguished from market orders. The CSE data contain orders and transactions information for those trades that execute on the CSE, as well as for unexecuted limit orders. The CSE data also include "marker" orders and other proprietary orders sent to the NYSE. Market quotation data for the CSE was obtained from the SIAC data. The CSE also provided data on individual dealer quotes.

PSE Data

The PSE data file contains both order and execution information for orders received on both of the PSE’s trading floors. Trades used in our analysis consist primarily of trades processed through the PSE’s P\COAST system. Records for floor trades and manually reported trades do not contain an order entry time and, therefore, were excluded from the analysis. ITS orders sent to other exchanges from the PSE, and ITS orders received on the PSE are both included in the file, but there is no indication on which exchange the ITS trade was printed. Therefore, ITS trades were also excluded from the analysis.

Phlx Data

The Phlx's Market Surveillance Department retains separate trade and order files. The order file consists of all electronically placed orders by brokers. This file does not include orders phoned in by brokers. [On March 3, 1997, 95% of the orders on PHLX were entered electronically (10,591 of 11,113). The electronic orders represented approximately 57% of the volume for the day (3,560,299 of 6,236,998 shares).] The trade file includes all Phlx prints, including ITS trades. Note, however, statistics can only be calculated for those trades with complete order information.