Mr. Chairman and members of the subcommittee: I am pleased to have this opportunity to offer my views on the SEC’s proposed Regulation NMS and, if implemented in current form, its likely effects on investors.

Regulation NMS is a complex proposal, with elements that address different aspects of the national market system. In my testimony, I will discuss only the basic question of market structure, which is implicated by the Regulation’s proposed changes in the applicability of the so-called trade-through rule. 

The US securities market today consists of two entirely different structures--a centralized market for the trading of New York Stock Exchange listed securities, and a set of competing market centers for the trading of Nasdaq securities. One of these models--and only one--is likely to be the best for investors, and hence the best market structure, but Regulation NMS does not help us decide which it is. In fact, by allowing some investors and markets to trade through prices on the NYSE, and by attempting to impose the trade-through rule on the trading of Nasdaq securities, Regulation NMS further confuses the issues. 

The fundamental question of market structure is whether investors are better off when securities trading is centralized in a single dominant market, or when it spread among a number of competing market centers. If the SEC is interested in reforming securities market structure it must address this question, and Regulation NMS does not do so. Accordingly, I believe the Regulation should be withdrawn until the SEC has done sufficient study and analysis of market structure to make an appropriate recommendation.  

There are two basic models for organizing a securities trading system. In the first, trading in specific securities is centralized, so that to the maximum extent possible all orders to buy and sell meet each other in a central market. In economic theory, this produces the greatest degree of liquidity and thus the best prices and the narrowest spreads. 

This model has two potential large-scale deficiencies: it forces all trading into a single mode--one size fits all--and thus will not meet the trading needs of some investors, and it does not create incentives for innovation or encourage accommodation to changing investor needs and demands.

The second model is a decentralized structure that contemplates competing market centers. Any security can be traded in any market. The advantages of this structure are that it can potentially meet the trading requirements of the greatest number of investors, and because the markets are in competition with one another it provides adequate incentives for innovation and change. 

The disadvantage of this structure is that it breaks up liquidity and thus could potentially interfere with price discovery; it also could result in investors getting different prices for the same security executed at the same time, which some regard as unfair.

In most sectors of the US economy--other than the securities markets--the second model prevails, and in fact industry structures have been moving in that direction because of technology and the benefits of competition. The same products, from automobiles to computer software, are sold through a variety of retail outlets, including stores, websites and catalogues, at varying prices and in combination with a bewildering variety of associated services such as warranties, technical support, and money-back guarantees. No one seems to find this strange or in any sense troubling, and consumers treat it mostly as a game, boasting to friends about the bargains they were able to obtain through shrewd negotiating or extensive research. 

For some reason, however, this market structure is not deemed to be satisfactory for securities trading. In the securities markets, SEC regulation has over many years attempted to protect investors against the possibility that they might not get the best price available in the market when they want to trade. This long-standing policy has been justified as assuring investors that the market is fair, or--on a more practical level--encouraging them to place limit orders in trading venues because they have some assurance that the offer will not be ignored. Because Americans seem quite content with receiving prices for identical products that differ on the basis of how much research or effort went into the purchase process, it is not at all clear that the SEC’s rationale for its policy is sound. Nevertheless, that rationale is the basis for the trade-through rule, which--as discussed below--is the most significant current determinant of securities market structure.

That market structure is characterized today by only limited competition among trading venues. Although SEC officials describe the US system as one that consists of competitive markets, this is true in only two limited senses. The Nasdaq market is indeed competitive. There, many market centers vie for investor attention and trading interest, and investors are getting the benefits of the pricing and innovation that results. But this is only half the system and probably a good deal less than half the volume of daily trading in US securities. The other half consists of the trading in New York Stock Exchange-listed securities, and competition in this market is severely limited. The NYSE is a centralized market, where investors may get the benefits of concentration and resulting liquidity, but not competition among markets. The only sense in which Nasdaq and the NYSE are competing with one another is that both of them seek listings from the same issuers. This is competition of a kind, but not competition that might provide benefits to investors.

The reason for the difference between competitive conditions in the two markets is probably the trade-through rule, which is applicable to NYSE-listed securities, but for historical reasons not to those listed on Nasdaq. The trade-through rule requires that customer orders to buy or sell NYSE-listed securities be forwarded to the market center where the best price for those securities has been posted. The purpose of the rule--in conformity with the SEC’s long-standing policy--is to increase the chances that buyers and sellers of a security will get the best price available in the market at the time they want to trade, even if the security is traded in different markets. 

When all markets were human-mediated, the rule did not have much effect on the market structure of, or on competition in, the markets for NYSE and Nasdaq securities. In both market centers, it was routine and expected that investors would wait for the NYSE specialist or the Nasdaq market maker either to execute a trade or to report that the trade could not be executed because prices had changed between the time the trade was sent and the time it was acted upon. However, with the rise of electronic communications networks (ECNs), the applicability of the rule in the case of the NYSE--and its non-applicability in the case of Nasdaq--had significant effects on the competitive structure of each market. 

ECNs are capable of matching buy and sell orders virtually instantaneously; this is the source of their attractiveness as trading venues. Where the trade-through rule was applicable--where  it was necessary to forward a trade to a human-mediated market (usually the floor of the NYSE) and wait for a decision by a specialist--ECNs cannot function efficiently or effectively. As a result, the ECNs have not been able to compete effectively for trading in NYSE securities. The liquidity and depth that has resulted from the centralization of trading on the floor of the NYSE still pulls in volume, and the trade-through rule prevents the ECNs from developing the degree of volume and liquidity in NYSE-listed securities that is required to compete effectively. As long as the trade-through rule continues to apply in the market for NYSE securities, it will not be possible to test whether ECNs provide a more efficient trading venue--at least for some investors--than the specialist system that currently prevails at the NYSE. 

In the case of Nasdaq securities, however, where the trade-through rule does not apply, ECNs apparently have apparently been able to provide better overall pricing than Nasdaq market-makers. This is true even though some trades took place at prices that would--if the trade-through rule had been applicable--have required the trade to be forwarded to a Nasdaq market-maker for execution. What seems clear is that in the Nasdaq market, where there has been a real world test because of the inapplicability of the trade-through rule, ECNs offered such strong competition for Nasdaq’s dealer-market structure that Nasdaq was compelled to convert itself into an electronic market--in effect, an ECN. Only in this form could it compete for trading in Nasdaq-listed securities. 

The rise of ECNs is a classic example of how technological change can completely upend settled economic arrangements, and the trade-through rule is an example of how regulation can create and sustain market structures that--without it--might not have existed at all. 

When the national market system for securities was first mandated by Congrss in 1975, there were a number of regional securities exchanges functioning in the United States, and Congress seems to have contemplated that they would all be in competition with one another. A comprehensive national quotation and trade reporting system would inform investors and their brokers where the best prices in specific securities were to be had, and it was assumed that trading would naturally flow in that direction. 

In the normal course, if this market had been allowed to develop as other markets do, trading in particular securities would tend to centralize in a single market--just as all the auto sales showrooms, antique stores and flower shops tend to locate in the same area--because customers will go to the place where the most choices and the greatest competition among sellers are thought to prevail. All other things being equal, investors will generally place their orders to buy and sell in the market with the greatest liquidity, since that is where the chances that the trade will be executed are greatest, and where the prices are likely to be best.

This sorting out process might have resulted in a securities market structure where certain regional exchanges--say, Philadelphia or Boston--would have become market centers for specific securities or the securities of certain industries, concentrating the trading in those securities in a single place and thus producing the advantages of high liquidity. If such a structure had developed, it would have provided some of the benefits of true competition among markets, since the existence of potential competitors would have spurred each market to innovate and operate efficiently. 

However, this structure did not have a chance to develop. The SEC--pursuing its policy of assuring that investors have access to the best price anywhere in the market--pressed for a linkage among all the existing markets that would have assured the execution of orders in each listed security on the basis of time and price priority. In other words, if an investor was the first to offer to buy 100 shares of US Steel at 30, this order would have to be executed before anyone anywhere could buy US Steel shares at 30 and 1/8.

The existing regional markets resisted this plan, and a compromise was developed, now known as the Intermarket Trading System, or ITS. In this system, each market center had reciprocal trading privileges in every other market. To meet the SEC’s demand that investors always have access to the best price posted anywhere, the ITS included a “trade-through” rule, which required that orders in particular securities be sent to the market where the best price was posted. In securities market lingo, to “trade through” a price is to ignore it and execute a purchase or sale at an inferior price.

In practicality, this meant that the New York Stock Exchange, which at that time had the deepest and most liquid market in virtually all listed securities, would become the dominant and unchallenged market for listed securities. Although interposed for other reasons, the trade-through rule had the effect--probably unintended, at least by the SEC--of eliminating the competition among markets that Congress had originally contemplated. 

There is good reason to believe that in the absence of the trade-through rule ECNs would be able to provide substantial competition for the NYSE in NYSE-listed securities. One of the peculiarities of the securities markets is that they are in a sense both wholesale and retail markets. Buyers and sellers of 100 shares are mixed in with buyers and sellers of 100,000 shares, even though their needs and demands are very different. One of the principal differences is that big buyers and sellers have an effect on the price of a security, while purchases and sales by small investors do not. Even information about the trading interest of big buyers and sellers can have an effect on the market, and thus has a value in itself, while information about the trading interest of small investors has no special value.

Accordingly, the best price available for institutional buyers and sellers is far different from the best price available for retail investors. A retail investor can often buy or sell 100 or 1000 shares at a price that is better than the spread on the NYSE, but that opportunity is not available to the institutional buyer. The fact of the institutional buyer’s trading interest can cause the price to rise or fall. Thus, the best price for an institutional buyer may be obtained in markets where institutions can trade anonymously, without intermediaries and without revealing the full scope of their trading interest--in other words where their trading has the least market impact.

Where markets are rising or falling quickly, the ability to achieve quick execution for a large order is also a factor in whether the institutional investor receives the best price that could be obtained in the market at that time.

These examples illustrate that it is a vast oversimplification to suggest that the current dispute about market structure and the role of ECNs is a question of price versus speed of execution.

Because of the material differences between the trading needs of the institutional investor and the retail investor, the institutional investor’s definition of best price is different from that of the retail investor. To the institutional investor, the best price is the price that can be obtained with what is called low market impact. Speed of execution is a factor, too, because large orders simply cannot be filled at a single price, and the market is frequently moving away. These are not considerations that the retail investor must keep in mind, so it should not be surprising that the definition of best price for retail investors and institutional investors turn out to be different. 

Thus, it appears that the current structure of the NYSE is suitable for the transactions of millions of retail investors, whose trades and trading interests do not on an individual basis have market impact, but--because it provides few mechanisms for limiting market impact--is not entirely suitable for institutional investors.

With this background, we come to the question that the SEC is now required to answer: whether, in a world in which technology has made it possible for investors--especially institutional investors--to have access to electronic trading venues, the SEC should continue to apply the trade-through rule for transactions in NYSE-listed securities. A more general statement of this question is whether the best structure for the US securities markets is one in which there is a single centralized market like that for trading NYSE securities, or one--like the current market for Nasdaq securities--in which a variety of market centers and trading venues compete with one another. 

Regulation NMS does not answer this question. Instead, it simply attempts to tweak the existing structure so as to address some of the complaints of institutional investors, ECNs and others, without considering whether its support of centralization of trading should be modified or abandoned. This approach is internally inconsistent, reflects a confusion of policy at the SEC, and fails to address the fundamental issues involved. Indeed, it has many of the aspects of dividing the baby: it seems superficially fair or evenhanded by offering each group what it seems to want, but in every real sense it does not produce a satisfactory or workable result.

For example, the Regulation would permit institutional investors and possibly others to opt-out of the trade-through rule for NYSE securities. It would also allow electronic or automated markets, under certain conditions, to trade through non-automated markets. This is clearly an effort to address the desire of institutional investors to trade NYSE securities on ECNs, but it completely ignores the traditional rationale for the trade-through rule, and the arguably beneficial effects of the rule in centralizing trading in the NYSE.

The basis of the rule is fairness; it is intended to provide an opportunity for everyone who is trading securities at a given time to have access to the best posted price available in the market.  Supporters of the rule, including SEC officials, argue that investor confidence in the fairness of the market would be compromised if investors did not get the best price available at the time of their trade, or if their limit orders were ignored and traded through in some other trading venue.

As noted above, in markets other than the securities market it is not unusual for buyers and sellers to find that they have not gotten the best price, and they seem to survive this information with their emotional stability intact. But if we assume that the securities market is somehow different from other markets, and that investors cannot rise above the unfairness associated with not getting the best available price at a given time, on what principled basis does the SEC exempt automated markets from the trade-through rule or allow investors to ignore the posted prices in non-automated markets by opting out of the rule? Fairness is either important or it is not. 

Similarly, it can be argued that the trade-through rule protects the NYSE against the kind of competition that has reduced Nasdaq to just one of a number of automated markets competing for trading interest in its own listed securities. If so, in the view of supporters, the rule is also preventing the “fragmentation” of trading in NYSE securities that would reduce liquidity and widen spreads. If centralization is indeed a benefit, and fragmentation is a danger, on what basis is the SEC allowing automated markets to trade through the NYSE, or allowing traders to opt out of the trade-through rule and trade elsewhere in NYSE securities? 

The answer to all these questions is that the SEC provides no answer, either in Regulation NMS or in the accompanying material. It seems simply to be seeking to mollify ECNs, and institutional investors who want to use ECNs for trading NYSE securities, by providing a limited opportunity to trade through NYSE prices. There is no indication in the proposed regulation or the accompanying background material that the SEC has made a judgment about the fundamental question involved--whether US investors as a group would be better off if market centers compete or better off if trading were centralized in one place. 

A similar set of questions could be raised about the Regulation’s proposal to impose the trade-through rule across the board to all NMS securities, including Nasdaq securities. What evidence is there that the absence of a trade-through rule in the Nasdaq market--and the migration of trading to ECNs--has harmed investors in Nasdaq-listed securities? Is there evidence that investors who have been traded-through are disgruntled or disheartened, or that they are no longer placing limit orders on ECNs or with Nasdaq market-makers? If there is no evidence that these things are happening, why impose the trade-through rule where it is not already applicable? The reason can’t be the usual reason for support of the trade-through rule--that it is necessary to assure investors of fairness and access to the best prices. That reason has already been given away by the SEC’s proposal in the Regulation to allow an opt-out for some investors and an exemption for automated markets. Nor can the reason be that the SEC wants to centralize trading in Nasdaq securities; the opt-out provisions in the Regulation preclude that.

Although it would be tempting to do so, the Regulation cannot be viewed as an experiment. If it were, it would not have been proposed as a regulation. A proposed regulation can be put into effect after the comment period and is permanent until modified or withdrawn. Market participants act in reliance, assuming that the regulation will not change. When the SEC has done experiments, as it has several times in this area, it describes them as such, and notes that they are temporary. 

Some experiments might have been helpful in framing a regulation that effectively addresses the central questions of market structure. For example, the Commission could have selected a sample of NYSE-listed securities, and permitted investors to opt out of the trade-through rule with respect to those securities, or permitted automated markets to trade through non-automated markets in trading those securities. The overall effect on investors could then have been assessed. Was the result  substantially to reduce liquidity and increase the spreads in those securities? Did some investors in fact find pricing so much better on the ECNs that they moved substantial amounts of their trading to automated markets? The answers to these questions might have provided the SEC with some of the data necessary to make a judgment about whether the trade-through rule should continue to apply to the trading in NYSE-listed securities, and whether or not a market consisting of competitive market centers is better for investors than a centralized market.

Thus, the only apparent rationale for Regulation NMS is that the SEC is temporizing. The agency has no answer to the question whether centralization of trading is better than competitive market centers--although this is the fundamental point at issue--and no plan for finding an answer. Instead of doing nothing, it is proposing a regulation that will keep everyone engaged in the debate. By permitting some investors to opt out of the trade-through rule, and exempting automated markets from the rule in certain circumstances, the SEC is suggesting that it thinks competing market centers are good policy and that the fairness issues traditionally associated with the trade-through rule are not important anymore. On the other hand, by applying the trade-through rule to trading in all NMS securities, the SEC is suggesting just the opposite--that fairness is important and enhanced competition between market centers is not. In other words, Regulation NMS is less than a half-measure; it leaves the major issues of securities market structure unexplored and unresolved.

Accordingly, there is a complicated answer to the question of how investors will fare if Regulation NMS is adopted as proposed. Because it does not resolve the basic question whether securities trading should occur in a centralized market or in multiple market centers that are competitive with one another, Regulation NMS--if it is imposed at all--will inevitably be only a temporary stopping point. 

Because I believe that competitive market centers will provide the most choices for investors and ultimately produces the most efficient markets, I expect that the Regulation’s opt-out provisions and its exemption for automated markets--if they go into effect--will over time cause substantial amounts of trading in NYSE securities to move to the ECNs. On the other hand, the imposition of the trade-through rule on trading in Nasdaq securities will not have much effect, since most participants in the automated Nasdaq market already commit to forward customer orders to other market centers where they can be immediately executed. 

Under these circumstances, for competitive reasons, the NYSE will eventually elect to become an electronic market, and will privatize for this purpose. Investors will fare better, overall, in this environment than they will in the current system, but the transition from the current structure to a structure consisting entirely of competing market centers will be messy and costly to all concerned. To avoid this, it would be far better for the SEC to step back from Regulation NMS and consider the fundamental issues involved--running experiments and doing analysis before leaping ahead with a poorly-considered and temporary plan.
Mr. Chairman, that concludes my testimony.

Peter J. Wallison is a resident scholar at the American Enterprise Institute.