Speech by SEC Commissioner:
Remarks before the Open Meeting: Reproposal of Regulation NMS

by

Cynthia A. Glassman

Commissioner
U.S. Securities and Exchange Commission

Washington, D.C.
December 15, 2004

When we first went out for comment on proposed Regulation NMS, I said that I considered that "Round 1." I consider today's reproposal of Reg NMS as "Round 2"and we may even need "Round 3" to get to the right result. I do not necessarily agree with the assumptions or the analysis in the reproposal, but I do support going out for additional comment. Given the magnitude of the issues we are dealing with in the area of market structure, and the possible consequences of this rulemaking on U.S. markets and investors, it is incumbent upon us to make our thinking as transparent as possible and to seek informed comment. I note that, in contrast to the ABS proposal, I have had a virtual parade of concerned stakeholders visiting my office on this proposal.

The draft reproposing release states that "[v]igorous competition among markets promotes more efficient and innovative trading services, while integrated competition among orders promotes more efficient pricing of individual stocks. Together, they produce markets that offer significant benefits for investors and listed companies." Given my understanding of the proposal and my review of the comments, I am not convinced that the proposal achieves these goals. I am concerned that the proposal may disincent competition and may not achieve the benefits of integration.

The reproposing release seeks comment on two alternatives: a trade-through rule designed to protect a market center's "top of book" or a trade-through rule designed to protect full "depth of book" on a voluntary basis. Framing the question this way presupposes that we are going to have a trade-through rule, which the Commission has not decided. Moreover, limited to the two choices in the release, commenters' preference for one alternative over the other will not be instructive for our purposes if they are simply choosing the lesser of two evils. To find out what commenters think is the best market structure solution, we need to acknowledge that there is a third alternative - not having a trade-through rule -- which many commenters supported in Round 1.

Throughout our work on market structure, I have urged consideration of the entire spectrum of possible alternatives - from having no trade-through rule all the way to building a Central Limit Order Book or CLOB. Since one of the purposes of proposed Reg NMS is to protect limit orders, we need to look at how various alternatives along the spectrum accomplish this goal. I understand the drawbacks of a CLOB, notably that it would stifle innovation and competition, but a CLOB would provide the ultimate in limit order protection. Short of a CLOB, today we are focusing on the voluntary depth of book alternative. Some view this alternative negatively as being a "virtual" CLOB. However, it would protect more limit orders than the "top of book" alternative. But if protecting limit orders is the goal, why aren't we considering other changes that would further protect limit orders?

For example, under current rules and under our current proposal, broker-dealers are not required to execute orders at the displayed NBBO price. Broker-dealers can internalize the trade and simply "match" the limit order price, leaving the limit orders representing the NBBO unexecuted. I would be interested in getting comment on a rule requiring broker-dealers to take out posted limit orders up to their displayed size before matching. In other words, limit orders would trump internalization. While it may sound radical, such a change would be more intellectually consistent with limit order protection than the alternatives we are currently considering. If protecting "top of book" is good, and if protecting "depth of book" is better, why isn't limiting price-matching even better - which brings us very close to a CLOB, where we don't seem to want to go?

So where are we on the spectrum? My concern is that at this point we are in the middle of the spectrum, thinking about imposing a trade-through rule that - even with depth of book protection -- does not protect as many limit orders as it could and that may be burdensome, expensive and difficult to enforce. One way out of this dilemma is to focus instead on the accessibility of quotes. Without good linkages and immediately accessible quotes, the trade-through rule will not work. On the other hand, if we have good linkages and accessible quotes - together with best execution requirements -- doesn't that lead to the same increased liquidity result without the regulatory interference with the market?

I have indicated all along that, even if we adopt a trade-through rule for the listed market, I would be reluctant to expand it to the Nasdaq market. The premise of the staff's recommendation to expand the trade-through rule seems to be the current level of trade-throughs on Nasdaq and a desire for uniformity of regulation across markets. The basis of the staff's recommendation is studies by our Office of Economic Analysis ("OEA") indicating that 2% of all trades on Nasdaq are trade-throughs and between 2 - 8% of volume on Nasdaq is traded through. I note that these studies will be available on the website along with the reproposing release. I urge commenters to look at them very closely and let us know if the interpretation of the data matches reality and is really a concern.

Finally, on the topic of market data, I want to make clear that while the proposed changes to the formulas for allocating market data revenue may enhance incentives for better quoting and reducing abuses, in my view they are only an interim step. The market data revenue pie is only going to get larger if there is depth of book protection, and we need to consider a longer-term solution in the context of our proposed SRO governance rule and concept release.

Questions:

• In addition to the other issues I've highlighted, I would like commenters to focus on whether the costs and benefits in the proposal are realistic. I have a question for the staff along these lines. On page 53 of the latest version, you state that "intermarket trade-through protection could improve depth and liquidity for NMS stocks by at least 5% or an average reduction of 1.87 basis points in price impact and liquidity search costs for large investors" and you go on to say that the impact would be about $755 million in 2003. How did you come up with these numbers? And, even if you're right, could an improvement of only 1.87 basis points possibly justify the overall costs of the rule?
   
• What would be the impact of the voluntary depth of book alternative on market data revenues? How would we deal with this newly bloated pie?
   
• It strikes me that there will be many practical issues in implementing the trade-through rule and monitoring compliance with it. One concern that I've heard - particularly with respect to the voluntary depth of book alternative -- has to do with time lags between datafeeds available through private linkages and the slower datafeed from the SIP. How will trading centers deal with this discrepancy?
   
• Won't the trade-through rule slow down the market? Doesn't the one second window in the self-help exception imply that "immediate" means one second or less? My understanding is that one second is slower than current response times.
   
• If we don't have a trade-through rule, would we need a cap on access fees? Why?
   
• In thinking about enforcement of a trade-through rule, how will we distinguish between trade-throughs that are inadvertent and trade-throughs that are intentional? Where will we draw the line?
   

Thank you. I appreciate the staff's continued efforts on these very complex issues and look forward to the next round of comments.

http://www.sec.gov/news/speech/spch121504cag.htm