April 13, 1998 Mr. Jonathan Katz Secretary United States Securities and Exchange Commission Judiciary Plaza 450 5th Street, N.W. Washington, D.C. 20549 Re: File No. SR-NASD-98-21 Dear Mr. Katz: I. Introduction This comment letter is submitted on behalf of the Electronic Traders Association ("ETA"), an association of order-entry and other related firms spread around the country. ETA provides an opportunity for thousands of individual investors to trade electronically through The Nasdaq Stock Market's ("Nasdaq") Small Order Execution System ("SOES") and SelectNet Service ("SelectNet") and, in some cases, through other electronic communications networks ("ECNs") and electronic systems sponsored by the listed exchanges. The views of ETA should mirror the concerns of all Nasdaq investors whose orders are not covered by a preferenced order flow arrangement. ETA generally supports the Order Execution Rules designed and approved by the Commission in 19961, which have benefitted investors greatly, but continues to believe that the Actual Size Rule, which was included in the package of rules ("Implementing Rules") proposed by The National Association of Securities Dealers, Inc. ("NASD") purportedly to implement the Order Execution Rules2, undermines the objectives of the Order Execution Rules. ETA reiterates the views expressed in its four comment letters submitted to the Commission to date3 and emphasizes in this letter that the data drawn from the Pilot Program demonstrates that the Actual Size Rule harms Nasdaq depth and liquidity and adversely impacts the use of SOES without any compensating investor protection or public interest benefit. Thus, it remains the view of ETA that the Pilot Program should be terminated and the Actual Size Rule should be rejected. 1 Securities Exchange Act Release No. 34-37619A (September 6, 1996), 61 FRR 48290 (September 12, 1996) ("Order Execution Rules"). 2 Securities Exchange Act Release No. 34-38156 (January 10, 1997), 62 FR 2415 (January 16, 1997) ("Actual Size Rule"). This release provided that Nasdaq market makers, in the first 50 stocks subject to the Order Execution Rules, were allowed to display a minimum quotation size of one normal unit of trading when quoting solely for their own proprietary account under a pilot program ("Pilot Program"). The Pilot Program was extended several times and is now slated to run until June 30, 1998. The Pilot Program was also subsequently expanded so that it encompassed a total of 150 stocks. Securities Exchange Act Release No. 34-39285 (October 29, 1997), 62 FR 59932 (November 5, 1997). 3 See letter from Richard Y. Roberts, Reid & Priest, to Jonathan Katz, Secretary, Commission, dated May 12, 1997, on behalf of ETA ("First ETA Comment Letter"); letter from Richard Y. Roberts, Reid & Priest, to Jonathan Katz, Secretary, Commission, dated July 2, 1997, on behalf of ETA ("Second ETA Comment Letter"); letter from Richard Y. Roberts, Reid & Priest, to Jonathan Katz, Secretary, Commission, dated August 13, 1997, on behalf of ETA ("Third ETA Comment Letter"); letter from Richard Y. Roberts, Reid & Priest, to Jonathan Katz, Secretary, Commission, dated November 26, 1997 ("Fourth ETA Comment Letter"). II. Studies Dispute the Justification for the Actual Size Rule In support of the Actual Size Rule, the NASD has made many arguments; and ETA challenges all of them. One argument advanced by the NASD is that the empirical analyses submitted by the staff of NASD Economic Research demonstrate that the Actual Size Rule does not affect the market quality -- in terms of spread, volatility, depth, or liquidity -- of the stocks included in the Pilot Program4 . According to the NASD, since the empirical evidence suggests that the Actual Size Rule does not cause any material difference in market quality in these four areas for Nasdaq stocks, then the Actual Size Rule should be approved because artificial regulatory minimums are no longer necessary to ensure market quality in the Nasdaq marketplace. ETA agrees that the Actual Size Rule has had no favorable impact on spreads, has not been associated with any change in intraday volatility, and does not impact the number of Nasdaq market makers. However, ETA disputes the proposition that the Actual Size Rule does not lead to a substantial reduction in market depth and liquidity, especially during times of market stress. In support of this position, ETA submits a study of its own5 . For additional support, ETA points to data contained in the Second NASD Study which indicates, as did the Simaan-Whitcomb Study, that the Actual Size Rule does result in a substantial reduction in Nasdaq depth and liquidity6. In favor of the Actual Size Rule, the NASD also argues that "economic theory" indicates that permitting dealers to quote in size commensurate with their true trading interest could further narrow quoted spreads and improve the pricing efficiency of the Nasdaq marketplace. Furthermore, the NASD argues that minimum quote size requirements could limit the ability of market makers to manage risk effectively and could thereby present a barrier to market making for small firms. Unfortunately, two things that the data submitted to the Commission by the NASD did establish was that the Actual Size Rule does not contribute to the narrowing of market maker spreads and does not encourage any increase in market maker participation7. Moreover, in support of the Actual Size Rule, the NASD argues that a minimum quotation size requirement may reduce the benefits associated with Nasdaq's reduction of the minimum increment for quoted prices to 1/16th on June 2, 1997, by limiting the ability of market makers to use the reduced increment. While the NASD uses the word "may" in this statement, the argument is still confusing since even the NASD's studies established that the Actual Size Rule had no favorable effect on spreads. ETA is astounded by the use of these arguments by the NASD in support of the Actual Size Rule. ETA is unaware of any data supporting the proposition that the Actual Size Rule would remove a disincentive for market makers and, in any way, would thereby attract more liquidity and pricing efficiency on Nasdaq. 4 The NASD submitted a study to the Commission dated June 3, 1997, on the Pilot Program for the Actual Size Rule from inception until April 18, 1997. See Effects of the Removal of Minimum Sizes for Proprietary Quotes in The Nasdaq Stock Market, Inc., prepared by the Staff of NASD Economic Research (June 3, 1997)("First NASD Study"). In order to examine the effects of an actual size rule, the First NASD Study compared four measures of market quality -- spread, volatility, quoted depth, and liquidity -- for a matched sample of 40 stocks subject to the Pilot Program ("First 40") and 40 stocks subject to the Order Execution Rules but not the Actual Size Rule ("Second 40"). The NASD submitted a second study to the Commission dated March 4, 1998, on the expansion of the Pilot Program for the Actual Size Rule which began on November 10, 1997. See The Effect of the Actual Size Rule on The Nasdaq Stock Market, Inc., prepared by the Staff of NASD Economic Research (March 4, 1998) ("Second NASD Study"). In order to examine the effects of the expansion of the Pilot Program for the Actual Size Rule, the Second NASD Study compared four measures of market quality -- spread, volatility, quoted depth, and liquidity -- for the expanded stocks subject to the Pilot Program ("Next 103 Stocks") and for the Nasdaq National Market stocks not subject to the Actual Size Rule ("Non-ASR 3,207 Stocks"). 5 In an attempt to provide the Commission with even more information concerning the impact of the Actual Size Rule on Nasdaq market depth and liquidity, ETA submitted a study which was prepared jointly by Professors Yusif Simaan and David Whitcomb of Rutgers University ("Simaan-Whitcomb Study"). The Simaan-Whitcomb Study was commissioned by ETA and is included as a part of this comment letter as Attachment I. The Simaan-Whitcomb Study compared various trading data drawn from the 50 stocks subject to the Pilot Program for the Actual Size Rule ("First 50") with similar data drawn from a sample of 50 stocks subject to the Order Execution Rules but not the Actual Size Rule ("Second 50"), first in the latter half of September of last year and then during the "turbulent market" of October 20-31, 1997. According to the Simaan-Whitcomb Study (p. 11), during the September period, the average market maker inside size for the First 50 was about 15 percent less than for the Second 50, a material difference. When the stocks being compared are those that the NASD considers comparable (the size-matched sample of 40 stocks being drawn by stratified random sampling, the "First 40" and "Second 40" stocks), the differences are even more striking. The average market maker inside size of the First 40 stocks in September of last year ranged from 15 percent to 20 percent less than that of the Second 40 stocks. Further, for the October time period, the Simaan-Whitcomb Study (pp. 11-13) demonstrated that mandatory minimum quotation sizes do cause market makers to provide more liquidity to the market, especially during times of market stress. The Simaan-Whitcomb Study shows that market makers clearly reduced the average aggregate size of their quotations when the market was under substantial stress in October of last year, with the greatest decrease observed for the First 40 stocks on the bid size. More specifically, during certain times of the October stress period, the First 40 bid average aggregate market maker sizes tend to be quite small relative to the Second 40, whereas the First 40 ask sizes generally remain at or above their September percentages. During these times, the First 40 average aggregate market maker quote size is as little as 63 to 67 percent of the Second 40 quote size. More generally, in the same side, same group comparisons, according to the Simaan-Whitcomb Study (p. 12), the October quotation sizes range from 30 to 10 percent less than the September quotation size counterparts. All of this data is consistent with the logical hypothesis that market makers will provide reduced liquidity on the side of the market under the most stress if they are free to do so. The Simaan-Whitcomb Study corroborates what ETA has stated for some time -- an actual size rule will adversely impact Nasdaq market depth and liquidity, especially when the market is under stress. 6 Normalized effective depth (dollar volume required to move bid-ask midpoint one percent by movement size) is a sound liquidity measure, and the normalized effective depth for the Next 103 Stocks fell to a much greater degree than for the Non-ASR 3,207 Stocks. For example, normalized effective depth for 1.0 percent BAM (bid-ask midpoint) movements fell by 15.0 percent for the Next 103 Stocks and by only 3.9 percent for the Non-ASR 3,207 Stocks. Normalized effective depth for 2.0 percent BAM movements fell by 19.2 percent for the Next 103 Stocks and by only 7.3 percent for the Non-ASR 3,207 Stocks. See Table III.8 of the Second NASD Study on p. 83. The NASD attempts to explain away this difference by utilizing a regression analysis, but even the NASD grudgingly admits that there is "weak evidence of a reduction in liquidity for stocks that became subject to the Actual Size Rule." Second NASD Study, Appendices, at p. 107. Another measure of depth and liquidity, although an imperfect one, is the average trade size. The average trade size for the Next 103 Stocks decreased by 14.7 percent, while the average trade size for the Non-ASR 3,207 Stocks increased by 3.7 percent, which is a startling difference. See Second NASD Study, Appendices, at p. 125. This data further suggests that the Actual Size Rule adversely impacts Nasdaq depth and liquidity. Moreover, the Second NASD Study examined the market quality and market maker accessibility of stocks included in the Pilot Program during the market stress of October 27 and 28, 1997. For the October part of the study, the NASD utilized stock groups based on the 36 ("First 36") of the First 40 stocks and the 36 ("Second 36") of the Second 40 stocks that still trade on Nasdaq. The quoted depth for the stocks subject to the Actual Size Rule declined by much more during the October market stress period. For the First 36, the quoted depth fell an average of 142 shares; while for the Second 36, the quoted depth fell an average of only 6 shares. This translates into a decline of quoted depth of 8.6 percent for the First 36 stocks, while there was only a decline of 0.4 percent for the Second 36. See Table III.12 of the NASD Second Study on p. 95. This data essentially confirms the finding established by the Simaan-Whitcomb Study that the Actual Size Rule adversely impacted Nasdaq depth, especially during times of market stress. See supra note 5. 7 The First NASD Study set forth in Table A.3 on before and after average quoted spreads (as well as the data in Table A.5 on "effective" spreads) demonstrates that an actual size rule does not cause market makers to quote narrower spreads than they do on stocks still subject to mandatory quote size minimums. Further, data from the First NASD Study appearing in Table IV.6 shows that the mean number of market makers has increased for both the First 40 and Second 40 stocks subsequent to the inception of the Order Execution Rules and by almost an identical amount. The Second NASD Study essentially confirmed the findings of the First NASD Study that the Actual Size Rule did not contribute to the narrowing of market maker spreads and did not encourage any increase in market maker participation. The Second NASD Study data set forth in Table III.2 shows that the quoted dollar spread decreased by 4.8 percent for the Non-ASR 3,207 Stocks, while it decreased only 3.8 percent for the Next 103 Stocks. The dealer dollar quoted spread (Table III.3) increased 6.3 percent for the Next 103 Stocks, while it increased only 5.8 percent for the Non-ASR 3,207 Stocks. The effective dollar spread for all trades (Table III.4) decreased 5.7 percent for the Non-ASR 3,207 Stocks, while it decreased only 2.6 percent for the Next 103 Stocks. Further, the number of market makers for both the Next 103 Stocks and the Non-ASR 3,207 Stocks (Table III.6) increased by an identical amount from a percentage standpoint during the period examined in the Second NASD Study. While the Actual Size Rule has been the subject of rigorous analyses, many of the NASD's strongest arguments in support of its existence rely on the use of "economic theory."8 This speaks volumes about the data contained in all of the analyses and attests to the lack of justification for the Actual Size Rule. The fact is that ETA is unaware of data indicating that the Actual Size Rule improves in any manner the quality of the Nasdaq marketplace or otherwise offers any benefit to the investing public. 8 The NASDıs "economic theory" depends, as set forth in the First NASD Study (p. 12), on the assertion that "[c]ompetition among market makers is the fundamental principle underlying the operation of Nasdaqıs market structure . . . this principle implies that . . . market makers should be left to post quotes freely and negotiate trades." However, there is a flawed key assumption in this "economic theory." The NASDıs "economic theory" assumes that quotation price competition exists among market makers; but if that were true, it would indeed follow that removing mandatory minimum quote size requirements would increase market maker participation and would reduce spreads. Why do the facts not fit this theory? Because, as the Simaan-Whitcomb Study establishes (pp. 13-17), Nasdaq market makers do not really engage in quotation price competition. This finding is consistent with that of other studies, as well as with the findings contained in the Commissionıs Section 21(a) report. See Collusion in the stock market, The Economist (January 17-23, 1998), at 1. The circumstance that severely undermines the "economic theory" utilized by the NASD is that systems executing paid-for or internalized order flow (proprietary autoexecution systems) constitute a significant portion of the trading activity on Nasdaq. The fact that so much order flow is preferenced to dealers means that quote price competition is significantly limited. There is little reason for market makers to quote tighter prices since most order flow comes either from preferenced retail orders or from institutionalized orders where previous performance in filling large orders plays a stronger role in attracting orders than a quote for 1,000 (or 100) shares. Indeed, ETA is of the view that the absence of meaningful price competition and the presence of a "pricing" convention among dealers were major reasons for establishing the Order Execution Rules in the first place. If the main venue for competition among Nasdaq market makers continues to be in the price paid for purchased order flow, the effect of removing mandatory minimum quote size requirements will be to increase the prices paid for order flow. That will further weaken quotation price competition and will undermine what the Order Execution Rules have accomplished. This cannot be the result sought by the Commission. III. Actual Size Rule Adversely Impacts Use of SOES Another of the NASD arguments in favor of the Actual Size Rule is that the implementation of the Actual Size Rule has not altered the ability of investors to access market maker capital. Yet, one of the more potent dangers inherent in the Actual Size Rule is the deleterious impact of the Actual Size Rule on SOES, especially during times of market stress. SOES was designed to enable investors to achieve immediate and efficient executions. Historically, SOES became operational in 1985 and was established to permit small orders in Nasdaq securities to be automatically executed at the best bid or ask price. Market maker participation in SOES became mandatory in 1988 because, as the staff of the Commission explained in Market 2000: During the October 1987 Market Break, due to record volume, unreliable quotations, and delayed transaction reports, market makers received an unusual high volume of calls both to verify quotes and to execute agency orders. Increasingly overwhelmed with calls, market makers were unable and, in the face of volatile market conditions, perhaps unwilling to answer the telephone and provide market prices. Indeed, even when reached, market makers were only willing to provide prices for a nominal amount of shares. Thus, it was necessary, for example, to make several calls to execute a single order of 1000 shares. Dealer participation -- and hence, market liquidity -- also suffered as a record number of market makers withdrew from the NASDAQ system. . . .9 Thus, in 1988, in response to the above-described issues raised by the market break of October 1987, the Commission approved an NASD rule proposal that mandated participation as a SOES market maker for all market makers10 . Under this rule, Nasdaq/NMS securities were assigned tiered maximum order sizes of 200, 500, or 1,000 shares depending upon the trading characteristics of the security.11 The Commission approved the rule proposal because it would "further the purposes of sections 15A and 11A of the [Securities Exchange] Act, by enhancing the liquidity of the OTC market, improving the accuracy of its pricing systems, and promoting the timeliness of trade reporting."12 9 Commission, Division of Market Regulation, Market 2000: An Examination of Current Equity Market Developments (January 1994), at note 57 (Study II) ("Market 2000"). 10 See Securities Exchange Act Release No. 25791 (June 9, 1988), 53 FR 22594 (June 16, 1988) ("Mandatory SOES Participation Rule"). 11 Id., at 22595. 12 Id., at 22596. The concern that the Actual Size Rule poses a problem for SOES is readily apparent from the data displayed in the Second NASD Study13. During the compared time periods, SOES dollar volume increased by 94 percent for the Second 36 stocks, but only by 16.7 percent for the First 36 stocks, an eyepopping difference14. Moreover, during these compared time periods, the SOES average trade size increased 2.0 percent for the Second 36 stocks, while it decreased 6.7 percent for the First 36 stocks15. Clearly, the Actual Size Rule negatively impacts the utility of SOES at all times, but especially during times of market stress. This adverse effect of the Actual Size Rule on SOES negatively impacts the ability of investors to achieve automatic executions which is harmful to Nasdaq market quality. 13 In the Second NASD Study, there are also presented a comparison of market quality measures for the First 36 and Second 36 stocks during three different time periods: (1) before implementation of the Order Execution Rules, (2) immediately after implementation of the Order Execution Rules and the Actual Size Rule for the First 36 stocks and of the Order Execution Rules only for the Second 36 stocks, and (3) a recent time period consisting of the 18 days before November 10, 1997 (this time period would include the October market stress period). The Second NASD Study specifically compares the mean values for the third time period to the mean values for the first time period. 14 See Table B.2 of the Second NASD Study, Appendices, at p. 117. 15 Id. IV. Actual Size Rule Harms Nasdaq Depth and Liquidity The lone remaining argument advanced by the NASD is that the "more" order-driven nature of Nasdaq engendered by the Order Execution Rules obviates the regulatory justification for minimum quote size requirements because ECN and customer limit orders are now displayed in the marketplace. ETA challenges this proposition. There is little evidence that the number of customer limit orders has increased materially. For example, if one assumes that market makers in the Non-ASR 3,207 Stocks (subject to the Order Execution Rules and to the 1,000 share mandatory proprietary quote minimum) never misrepresent a smaller proprietary quote as a customer quote, the Second NASD Study shows that only 18.5 percent of dealer quotes are limit orders smaller than 1,000 shares16. ETA believes that the only aspect of the Nadsaq market that has become more "order-driven" is that all market makers and some SOES day trader "layoff" orders, which always existed, have now become visible and can affect the best bid and offer. However, ETA does agree that an important public policy question confronted by the Actual Size Rule is whether it is still appropriate to impose, as the Order Execution Rules have moved Nasdaq toward a "more" order-driven market, mandatory minimum quotation size requirements on market makers to ensure an acceptable level of Nasdaq market depth and liquidity. The Pilot Program was initiated to provide the answer to this question, and all of the measures of both NASD studies purporting to deal with the provision of depth and liquidity show a reduction of depth and liquidity provision for the stocks subject to the Actual Size Rule.17 The Simaan-Whitcomb Study also examined the impact of the Actual Size Rule on Nasdaq market depth and liquidity; and, as described earlier, the Simaan-Whitcomb Study corroborates what ETA has stated for some time -- the Actual Size Rule adversely impacts Nasdaq market depth and liquidity.18 16 See Table III.7 of the Second NASD Study, at p. 76. 17 See supra note 6. 18 See supra note 5. The Order Execution Rules were designed to increase market depth and liquidity. As the First NASD Study stated "[l]iquidity is an important characteristic of an equity market; in a sense, it is synonymous with market quality."19 The Simaan-Whitcomb Study demonstrates that the Actual Size Rule adversely harms Nasdaq depth and liquidity and even both NASD studies admit that this is the case20. Since the Actual Size Rule creates lower depth and liquidity and fails to narrow the spread, the Rule's continued existence undermines the objectives of the Order Execution Rules and is harmful to Nasdaq market quality. Simply stated -- the Actual Size Rule has produced no discernible positive impact on Nasdaq market quality but has harmed Nasdaq depth and liquidity. Further, the Actual Size Rule has adversely affected the ability of investors to achieve automatic executions through SOES which also negatively impacts Nasdaq market quality. Therefore, ETA believes that the Pilot Program should be terminated and the Actual Size Rule should be rejected. 19 First NASD Study, at 35. 20 The Second NASD Study stated that there is "weak evidence of a reduction in liquidity for stocks that became subject to the Actual Size Rule." Second NASD Study, Appendices, at p. 107. See supra note 6. The First NASD Study stated that "there is statistically significant evidence, . . . that the quoted depth of the First 40 declined to a greater degree due to the removal of the 1,000-Share Quote Size Rule." First NASD Study, Appendix B, at 78. V. Conclusion For the reasons expressed hereinabove, ETA urges the Commission, after carefully considering all of the comments and based on the Commission's experience and knowledge of current market practices and conditions, to terminate the Pilot Program and reject the Actual Size Rule. Thank you for your consideration.                                          Sincerely,                                                                   Richard Y. Roberts Attachment I (Simaan-Whitcomb Study) cc: Honorable Arthur Levitt Chairman Honorable Norman Johnson Commissioner Honorable Isaac C. Hunt, Jr. Commissioner Honorable Laura S. Unger Commissioner Honorable Paul R. Carey Commissioner ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ James H. Lee (713)706-3300 x227 (713)977-7975 FAX jlee@soes.com ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ President, Momentum Securities Management Company, Houston, TX http://www.soes.com President, Electronic Traders Association, Washington, DC http://www.electronic-traders.org ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~