Thank you for inviting me to testify before you today. I appreciate the opportunity to provide the Committee with my thoughts on important market structure issues.
The publication of proposed Regulation NMS by the Securities and Exchange Commission has given the securities industry a unique opportunity for improvement. If we are thoughtful enough, if we are committed enough to the idea of competition, then we can reshape the regulatory environment in a way that will have significant positive benefits for the public investor.
One such opportunity for improvement is the Intermarket Trading System’s trade through rule. Originally this rule was put in place to require the New York Stock Exchange to honor better prices that were being displayed on non-primary exchanges. Since the establishment of the trade through rule in 1982, the non-primary exchanges have led the way in automating the trading process, which has reduced trading costs, enhanced market efficiency, and created opportunities for new types of electronic participants. Because of these innovations, however, the trade through rule has now become a barrier to competition, an unintended instrument for unfair government protection of the NYSE’s manually intensive monopoly. Because of the trade through rule, exchanges that provide automatic executions are required to slow their service and send an order to an NYSE quote that, though it may appear to be the best price in the country, may not actually be available. Even if that NYSE quote is available, a more efficient marketplace may have to wait up to thirty seconds to get an NYSE execution, and that thirty second wait could result in losses due to changes in market valuations during that time. Even worse, the NYSE’s execution price is often inferior to the NYSE quoted price that induced the order to be sent to New York in the first place.
The SEC’s own execution quality statistics (Rule 11Ac1-5) clearly reveal that the trade through rule is a major barrier to competition: NYSE-listed stocks subject to the trade through rule have wider spreads, get slower executions, and impose higher costs on investors. In contrast, Nasdaq-listed stocks, which are not subject to the trade through rule, provide investors with increased transparency, greater access to liquidity, faster executions, higher fill rates, and better-priced executions. In addition, experience with the SEC’s exception to the trade through rule for exchange-traded funds demonstrates that quote spreads narrowed and trade volume grew significantly after the trade through rule was modified to allow trade throughs in the top three ETF products.
The trade through rule has become unnecessary and counterproductive as a result of easy access to complete market data, technological advancements in trading systems, the increase in market competitors, and the implementation of decimal trading. The obvious solution, then, is to allow an investor to forego the trade through rule and let competition between exchanges drive best execution. It is time for the government to stop telling the American public where it has to conduct its business.
Although a consensus is building about the necessity for a trade through opt-out for manual or “slow” markets, there is still a lot of debate about whether the choice of an investor to opt-out of the trade through rule should extend to automated or “fast” markets. For two reasons, I believe that it is important to preserve opt-out flexibility for fast markets as well as slow markets. First, there are definitional concerns. The definition of “fast” will be arbitrarily determined based on the lowest common denominator solution that will be acceptable to the most politically powerful marketplace. Defining “fast” will become a slippery slope that will force the SEC to regulate more and more aspects of market technology. Second, fiduciary duty and economic self-interest obviate the need for any government-imposed trading rules, and therefore the Commission does not need to condition the flexibility given by an opt-out provision, even when a market is characterized as “fast” for regulatory purposes. If there really is no incentive to avoid a fast market, then a rule is not needed to require an investor to act in its economic self-interest. The SEC should allow exchanges and brokers the opportunity to compete, to operate without the constraints of intermarket order routing requirements, until it is empirically proven that the imposition of a trade through rule is necessary.
If the Commission decides not to extend the trade through rule to the Nasdaq market when it trades Nasdaq-listed securities, then, in order to ensure equal regulation, the SEC needs to grant exchanges the ability to trade Nasdaq-
listed securities without intramarket or intermarket trade through requirements. To do otherwise would be to allow Nasdaq to retain a significant regulatory advantage over its competitors. NSX has had for over two years a rule change proposal in front of the SEC – its “voluntary book” proposal – that would eliminate this regulatory advantage. Now is the time to either approve NSX’s voluntary book filing or to eliminate the ability of broker-dealers to trade through better-priced orders in Nasdaq’s marketplace. Only in this manner can the mandate of the Securities and Exchange Act of 1934 that self-regulatory organizations be subject to equal regulation be fulfilled.
On the issue of market data revenue, I would like to challenge the assumption in proposed Regulation NMS that the current system for market data revenue distribution needs to be changed, and I would like to caution the Commission against making a change that could prove to be far worse than any problem that exists today.
It is important to place the issue before us in historical context. The current
market data revenue distribution method has been in place for over 25 years. For most of that time, the method has been perceived to be fair, easy to administer,
and effective. It was first questioned two years ago by Nasdaq for competitive reasons because the National Stock Exchange had captured significant order flow in Nasdaq-listed securities as a result of NSX’s cost initiatives that involved the sharing of market data revenue with members. The fact is that, because other market centers that trade Nasdaq and Amex-listed issues were forced by NSX’s competitive initiative to share their market data revenue, investors have been saving over $100 million annually, and they now have the opportunity today to choose from multiple broker-dealers offering automated, price-improved executions for under ten dollars. These benefits were created through competition, not regulatory mandate. If, as Regulation NMS suggests, the SEC is not going to address the explicit cost of market data by lowering the overall size of the market data revenue pool, then it is important that the SEC preserve a competitive environment among SROs in order to continue to indirectly bring down the cost of market data for the brokerage community and the public investor.
The premise that a trade-based formula creates economic distortions, regulatory distortions, or inappropriate incentives to engage in fraudulent behavior has not been sufficiently proven to warrant the proposed change to market data revenue distribution. Even if one believes that fraudulent actions are
encouraged by the current distribution model, such actions represent rule violations that are already being regulated by effective SRO enforcement
programs. The potential for such malfeasance no more justifies the adoption of Reg. NMS’s proposed costly solution than the potential for speeding justifies shutting down the highway system.
The proposed formula amendment is unnecessarily complex, misguided in its price discovery value judgment, and expensive to administer. The complexity of the formula has made it a poster child in the industry for the inherent limitations of regulation. The position that trades less than $5000 have no price discovery value favors the exchange with the most block activity, and it defies logic. For example, the new formula includes a 100-share trade of a $56 stock
like JNJ but excludes a 1200-share trade of a $4 stock like SUNW, even if
the SUNW trade creates a new high or low of the day. The fact is that all trades have price discovery value. Finally, in a world that is generating eighteen million
NBBO quote changes daily across eight market centers, imagine the annual cost of determining how many thousands of quote credits each particular quote is due for each of the 23,400 seconds in every trading day. The calculation becomes particularly ludicrous if you consider that, in one example provided in the SEC release, a single quote that equaled the national best bid or offer for three seconds would be entitled to 12,000 credits. While NSX respects the desire of the Commission to encourage quote competition, the benefits of doing so through the proposed formula amendment simply do not come close to outweighing the new formula’s administrative costs.
Given all of the reasons above, NSX does not believe that the adoption of Reg. NMS’s proposed market data reallocation formula would be sound public policy.
Finally, whatever action the SEC ultimately takes, it is important to combine such action with changes that require NBBO and trade report market data to become real-time. This means that the Consolidated Quote Plan’s sixty-second quote update provision and the Consolidated Tape Plan’s ninety-second trade reporting provision must be significantly reduced. Given the electronic nature of the trading world today, real-time market information – which is the stated purpose of the CQ and CT Plans – means automatic execution, automatic quote updating, and automatic trade reporting. Therefore, no quotes or trades should be given credit for market data revenue unless they emanate from a “fast” automated market.
Thank you again for the opportunity to present my views on these important subjects. I look forward to answering any questions that you may have.
