'I felt the earth move under my feet*'

R ‘Tee’ Williams

*Apologies to Carole King.

In the rural United States of times past, fairs and other public events were often marked by a contest in which participants competed for fame and prizes by attempting to catch a pig that had been doused in grease. The afternoon after I had delivered the first version of this article to the editor (and spent the day trying to correct the typos for which I am famous) the NYSE announced its merger with the ArcaEx. A rewrite was the only alternative. I now have great empathy for the participants in greased-pig contests.

R ‘Tee’ Williams, 20 April, 2005

At the beginning of the week, it was rumoured that Nasdaq would purchase Instinet, currently controlled by Reuters and parent of Inet, Island and the Instinet broker. Three days later it was announced that the NYSE would buy ArcaEx, the exchange that evolved from the Archipelago ECN when it purchased the trading activities of the Pacific Exchange (PCX). Two days after the NYSE announcement, the Nasdaq/Instinet merger was confirmed. In one week, we had gone from what seemed to be a highly fragmented market to what seems to be a return to two primary marketplaces competing for listings and market share. (By Monday, April 25, another group headed by Kenneth Langone, a former director of the NYSE and friend of former chairman Richard Grasso, had announced a competing takeover bid for the NYSE.)

In the remainder of this article we look at these events critically and try to see what they mean for both markets. We will assess what has made the two markets successful, and what may result from the proposed mergers. Most importantly, we will investigate how these events relate to and are influenced by the enactment of Regulation National Market System. It has been a hell of a month.

To understand what is happening a little history is required. If you know the history of the US trading markets you may want to skip the next section.

A little history

In the time before the changes created by the Securities Acts Amendments of 1975 (the ’75 Amendments), there were three markets of primary listing in the United States. The National Association of Securities Dealers (NASD) was the self-regulatory organisation for the over-the-counter (OTC) market run by a group of dealers that bought and sold securities from inventory. Typically, companies that traded in the OTC market were new companies just listing for the first time and smaller companies without a national following. Some rather large banks were included because they were mostly regional issues in a time before nationwide banking.

When a firm’s shares became more seasoned they moved to the American Stock Exchange where they were traded on an exchange floor with a specialist trading system. At some point as they grew, firms usually elected to move from the AMEX to the New York Stock Exchange. For firms choosing a market in which to list, the critical difference between these three markets was increasingly stringent requirements. Each market imposed obligations with regard to different number of shares outstanding, number of shareholders and reporting requirements that were least demanding for the OTC^[1] market and most demanding for the NYSE. However, there were also important differences in trading style.

The critical feature of the Nasdaq trading system is that a dealer or market maker is in the middle of every trade. Before the 1970s, the NASD market had a reputation for being a bit wild. In 1973, under pressure from reforms sweeping the markets and leading up to the ’75 Amendments, the NASD created the NASD Automated Quote system or Nasdaq which system displayed all bids and offers from market makers. In the early 1980s, the SEC required Nasdaq to publish all last sale prices. Dealers initially fought the display of trade prices, but increased transparency caused the market to grow dramatically. Dealers eventually embraced last-sale reporting. In the mid 1990s, in response to a scandal among market makers, the SEC required dealers to include any public order they had that was priced at or better than their own quote as a part of the quote the dealer was obliged to display on Nasdaq. A number of firms created systems for displaying public orders that were in effect automated limit order books. These systems became Electronic Communications Networks or ECNs.

The AMEX and NYSE had a specialist trading system, or auction as it is known. The fundamental difference between a specialist system and the Nasdaq dealer system is that a specialist is not permitted to participate when two public orders can be executed without help. On the floors of the AMEX and NYSE a specialist is responsible for a list of securities. The specialist acts as an agent for orders in his securities when requested, or as a market maker if a customer order demands immediate execution but there is no other customer order against which to execute.

In the period after the ’75 Amendments there were a number of reforms to the Nasdaq market. Most of the abuses that created the OTC market’s reputation for wildness were eliminated and Nasdaq fought hard to be known as ‘the Nasdaq Stock Market’ instead of the OTC. Moreover, the technology firms such as Apple and Microsoft that grew in the 1980s elected to stay with Nasdaq rather than move to the AMEX and NYSE, because exchange floors seemed more antiquated than Nasdaq’s computerised quotes. The AMEX was eclipsed and Nasdaq became a formidable competitor to the NYSE.

The Nasdaq marketplace

The Nasdaq marketplace has undergone profound changes since the enactment of the Order Handling Rules in 1997. The change is so dramatic that I must be clear what I mean when I say the ‘Nasdaq marketplace’. I mean all of the different trading venues that execute trades in Nasdaq securities. The venues include:

• the Nasdaq SuperMontage;
• Nasdaq market makers trading in a traditional dealer role;
• the ECNs; and
• regional exchanges trading Nasdaq listings under Unlisted Trading Privileges (UTP).

To understand the Nasdaq marketplace it is important to understand that each execution venue predominantly trades specific types of orders for different constituencies. In effect, while there is significant overlap, most of the trading venues are not direct competitors.

The most immediate change following the Order Handling Rules was the emergence of ECNs. Instinet^[2] had existed as a type of electronic broker trading Nasdaq securities on a pure agency basis since the mid 1980s. After the Order Handling Rules, other ECNs sprang up quickly, each catering to a different constituency. Instinet had always catered to institutions. Island and Attain grew from firms that wanted low-cost executions for active retail traders, and in particular day traders, who were increasingly entering the market directly using the Internet. Archipelago also had roots servicing day traders but took on an institutional flavour. Archipelago (now ArcaEx) is headquartered in Chicago and was adopted by market makers on the Chicago Board Options Exchange (CBOE) to hedge options positions. Brut, RediBook and Strike were created by or for large dealers and wholesalers that wanted to capture some of the agency business they were beginning to lose to other ECNs. Bloomberg created TradeBook as a means for Bloomberg’s customers, particularly those outside the US, to trade in US securities. Each ECN has charges and execution rules that catered to its target market. The most important feature of ECNS from the perspective of the marketplace is that when an ECN executes an order, the ECN is only permitted to be an agent^[3].

The SuperMontage is the system that Nasdaq created to provide all dealers with a venue to post customer limit orders in order comply with the Order Handling Rules. SuperMontage is the first system created by Nasdaq^[4] that actually executes trades.

Dealers in their traditional role continue to trade for their customers. Most often, dealers have automated execution systems (AutoEx) where a customer can be guaranteed an immediate execution at the dealer’s posted bid or offer for the quantity of securities in the quote. Dealers typically pay for orders, particularly from retail brokers.

The UTP exchanges are each different, but generally can be thought of as being analogous to mutualised dealers. Individual market makers band together on regional markets to share infrastructure costs and divide up the securities each market maker (member or specialist) is permitted to trade.

Anyone trying to understand the market shares of the various trading venues is reminded of the musical ‘The Producers’, because summing up the claimed share of all participants equals well over 100%. Several important conclusions can be drawn, however. First, Nasdaq is no longer a purely dealer market. Second, ECNs have dramatically reduced the importance and profitability of market makers. (When statistics are used to imply that Nasdaq has a small percentage of its own marketplace, what is really meant is that the dealers’ share is dramatically lower. Actually, Nasdaq the firm now has a share of executions that it never had before.) Third, for small trades in liquid securities no dealer is required, and trades on ECNs and SuperMontage are primarily small orders in liquid securities. Fourth, dealer quotes on SuperMontage are effectively limit orders, similar to those entered by customers. Quotes from dealers that are pure advertisements only exist in the dealers’ own systems.

At present, Nasdaq marketplace has become the most competitive marketplace with the most execution alternatives. A pattern developed within the Nasdaq marketplace in which different users select different trading venues as their preferred place to trade. Until now, the combinations among ECNs primarily involved two ECNs targeting the same market segment. Even Nasdaq’s purchase of the Brut ECN from Sunguard combined the SuperMontage, created primarily to support Nasdaq dealers, with an ECN created to support the dealers that used Sunguard’s preeminent Brass dealer order management system.

The NYSE marketplace

There are certain entities – sports teams, companies and well-known organisations – that defy objective assessment. Obvious examples include Manchester United, the University of Notre Dame, Microsoft, and the New York Yankees. These organisations generate either fanatic loyalty or unfettered contempt. All are well known and have a history of success that generates both partisan praise and scorn.

To this list I would add the New York Stock Exchange (NYSE). The goal of this article is to try to provide an objective assessment of prospects for the NYSE in light of the recently passed Regulation National Market System (Reg NMS) and the Exchange’s declared intent to create a ‘Hybrid Market.’ A number of commentators have argued about Reg NMS and the Hybrid Market. Most articles and comments are either in complete support of the NYSE and the hybrid proposal (and assure rosy prospects for the Exchange), or demand removal of rules that ‘unfairly protect’ the NYSE (with the assurance that on a level playing field the NYSE would go the way of the dinosaur.) Typically in these articles each commentator’s views and the ‘public interest’ are described as identical.

There seem to be four interrelated reasons why the NYSE has remained dominant in its marketplace while the Nasdaq market fragmented. The reasons are:

• Benign opacity.
• Initial market share.
• The placement of agents and dealers.
• A loose linkage of markets.

I describe these characteristics as follows:

Benign opacity

A characteristic of electronic markets is that their rules are applied rigidly by the code of the trading system. To be sure, the NYSE has explicit rules, but these rules are applied as rigidly as humanly possible. The human element inserts ambiguity into the process – NYSE supporters would call it flexibility and detractors would call it slop or wiggle room – that means a trade on the NYSE is less predictable than trades on electronic markets. This unpredictability means that regulators are frustrated, those demanding certainty of execution are angered, and most traders cannot afford to ignore the NYSE even if they prefer another trading method. I will call this ambiguity ‘benign opacity’ even though many would say that it is malignant.

Initial market share

The opacity of the NYSE would provide no benefit if the Exchange’s market share were lower. With a high market share, most traders are forced to find some means of gaining a look at the NYSE if they have a trade that is large, complicated, or both. Any trader that is not in a hurry is likely to try the NYSE simply because there is a chance to improve on the price of an order. The process feeds on itself like a reactor that has achieved critical mass. The dark side of this phenomenon is that if the market share of the Exchange were ever to fall dramatically, the fall would be hard to stop.

The placement of agents and dealers

Electronic markets give the illusion that distance does not matter. In point of fact distance matters a great deal. This is witnessed by the fact that traders that execute electronically using strategies where speed of execution is critical are besotted by the idea of minimising physical distance and thus ‘latency’^[5]. Some traders have sought to host trading models at exchange facilities, even on the same computers.

A dealer or agent on the NYSE floor can interact with the market on the floor and modify strategies. By contrast, a dealer or agent using an automated system must effect his or her strategy either by a series of conditional order types or by issuing cancel/replace orders as he or she reacts to market events. In essence there is a level of information that exists on the floor and an opportunity to react to that information that does not exist in an electronic market.

There is a real question of propriety for the information and trading opportunities that exist on the floor. There is a strong potential for manipulation and abuse. Moreover, a floor is inherently unfair. Those who are not on the floor are deprived of information that is available to traders on the floor. The inequity is balanced by the fact that those who have the information advantage are required to provide services (i.e., they add liquidity and represent customers) that benefit the whole market.

A loose linkage of markets

One feature of the NYSE floor that is not often discussed is the fact that multiple ‘markets’ coexist and interact within the same venue. While the ECNs and other markets develop a following based on their specific fees and rules (as discussed above), differing trading constituencies come into the floor through different routes and with different motivations. The floor permits the interaction of (a) large institutions that often arrive directly through the Designated Order Turnaround (DOT) system under the sponsorship of a direct-access broker; (b) retail orders sent to market makers on regional exchanges (often attracted to the regional market maker through payment for order flow) that arrive by the Intermarket Trading System; and (c) hedge funds and other quantitative traders that use the Direct+ system to interact with the floor. On the floor itself, (d) floor traders with blocks and regular customer orders; and (e) specialists with both orders they have been given as agent and their own capital are physically present. These five major groups, representing many diverse subsegments, meet. Orders from the physical crowd and electronic order flow intermingle. Each group can interact with the others in ways that are not currently possible in stand-alone electronic markets.

I maintain, therefore, that the NYSE has been successful by blending a combination of natural attributes and some luck. To continue to thrive, the Exchange must find a means to sustain this successful blend in spite of the fact that several basic components must change.

What has happened

Against the background just described, three significant events have occurred in one month. The SEC passed rule Reg NMS; Nasdaq will acquire Instinet; and the NYSE and ArcaEx are proposing to merge. I describe each in turn.

Reg NMS

The current turmoil facing Nasdaq, the NYSE, and the other markets in the US is the result of a new SEC rule known as Regulation National Market System or Reg NMS. The rule has a complex history and the reasons for its enactment are complex as well. The short explanation is that the SEC believes in two overarching principles of regulation. First, it believes that regulation should be uniform across markets. Second, in an increasingly institutional market, the SEC believes that the interests of the retail investor must be protected.

Reg NMS has components that are largely technical, such as how market data is controlled, the level and type of fees and a prohibition against ‘minimum price variations’ of less than $0.01 for most securities. The one component that is changing the market structure in the US regulates the interaction among markets.

For the first time, the SEC is requiring that all US equity markets must abide by a ‘trade-through rule.’ The new trade-through rule says that no trading venue can execute a trade in a US-listed security at a price that is lower than the best price (bid or offer) on any other market trading the same security. The new rule now applies to all markets including Nasdaq-listed and NYSE-listed securities. (Previously, the NYSE market had a trade-through rule but Nasdaq–listed securities were not covered.) The trade-through rule only applies to the ‘top-of-book’ (i.e., the best bid and offer on each market, not the other orders on books at inferior prices (i.e., lower than the best bid or higher than the best offer.)

The new trade-through rule is significant for the NYSE because it only applies to ‘fast markets’. This means that, to be protected from trade-throughs, a market must be automated. This is significant for the NYSE and for the SEC. For the NYSE, the floor, which is the cherished feature of the NYSE’s historical method of trading, is no longer protected. For the SEC, there is a realisation that for some types of orders, speed of execution may be as important as the price.

In general, beyond requiring that the NYSE embrace automation, there is a general belief that the NYSE came out better than competing markets. Other markets must now honour trades on the NYSE or match the prices. Moreover, a controversial feature was rejected that would have allowed traders that prefer speed to ‘opt out’ of trade-through protection.

For the Nasdaq marketplace, brokers can no longer choose a trading venue and route all orders there without regard to conditions in other markets, a practice known as ‘preferencing.’ Preferencing evolved in Nasdaq’s wilder days when the quality of executions among dealers was very uneven. A broker could choose to ‘prefer’ a dealer with a history of quality executions. A broker is now obliged to get the best price for a customer at the ‘best’ bid or offer displayed among all automated markets trading a security. The obligation can be satisfied by routing the order to the best market, or by a guarantee from a market maker in another market to automatically generate an order equaling the best quoted price.

The Reg NMS proposals are controversial. The rules were passed by Chairman Donaldson and two Democratic commissioners. There are Republican lawmakers who are threatening to overturn Reg NMS in Congress. The future of the rule is therefore uncertain.

Controversy aside, the rule is remarkable. The Commission has harmonised rules among the various marketplaces. The SEC has forced the NYSE to adopt automation to a greater degree than would have been acceptable before the scandals over former chairman Grasso’s salary and the specialists’ performance. The SEC has also forced a degree of linkage among trading venues without requiring a single market mechanism such as a consolidated limit order book (CLOB). (I had personally predicted that the SEC would not be able to develop rules that would balance the interests of electronic and floor markets, and that as a result, uncontrollable chaos would ensue. We must see what happens, but I am surprised by the apparent quality of Reg NMS.)

One critical factor remains to be determined – the means by which the markets will be linked. The trade-through rule creates system-wide price protection, and the obligation for a market to either match a better price in another market or send the order to that market. The question of how the markets interact must be determined.

Currently, the markets that trade NYSE-listed securities are linked by the Intermarket Trading System (ITS). ITS has many problems that all relate to the rules under which it operates. A key feature is that ITS is free to market makers on linked exchanges. Therefore one alternative would be a more functional version of ITS that would be an ‘official’ market element free to authorised participants. A second alternative would be bilateral linkages among the exchanges themselves. The ArcaEx ECN/exchange links to other markets and encourages traders to send orders to ArcaEx with the promise that orders that are not instantly filled will be sent to the next market. A third alternative would be for brokers introducing orders to be responsible for finding the best market and/or moving orders among markets as conditions change. Firms such as Rosenblatt Securities provide this service on an agency basis. Also, brokers such as Speer Leads and Kellogg (the RediPlus products) and Lek Securities (the ROX system) could employ their proprietary routing systems to enable their institutional clients to manage orders directly. Finally, third party service companies such as Lava Trading (recently purchased by Citi Group), Sonic Trading or royalblue [sic] could provide routing facilities for a fee.

The Nasdaq/Instinet merger

Nasdaq proposes to purchase Instinet and keep Inet (the Instinet and Island trading systems). The Instinet Broker (the agency broker that was required for Instinet to qualify as an ECN) and Lynch Jones and Ryan (a specialty agency broker that Reuters purchased and combined with Instinet) would be spun off.

The effect of the Nasdaq/Instinet merger will depend largely on how the combined company is managed. First, a substantial portion of the trading in Nasdaq-listed securities will now occur in systems Nasdaq controls for the first time. Second, the trade-through obligations will be less complicated for Nasdaq since it will control four possible destinations – Brut, Instinet, Island and SuperMontage.

The open question is how the four trading systems will be managed. One possibility is that Nasdaq has simply purchased market share and will ultimately fold all of the acquired companies into a single trading system. The other alternative is that all or most of the trading systems will continue in a loose linkage similar to the linkage on the floor of the NYSE.

The NYSE/ArcaEx merger

There are two important elements to consider in evaluating the NYSE/ArcaEx merger. First there are the characteristics of the proposed merged entity. Second is the NYSE’s proposed Hybrid Market.

The NYSE and ArcaEx

At first blush, it is tempting to call the combination of the NYSE and ArcaEx ‘OrcaEx’. However, there are subtle weaknesses as well as strength.

The NYSE dominates trading in NYSE-listed securities. The NYSE has significant listing revenues and listing on the NYSE is considered to be very prestigious for listed companies. It also has perhaps the strongest market data franchises in the world even though it is currently tied to the regional exchanges. The Exchange is a major self-regulatory organisation (SRO). Finally, the NYSE is a majority owner of the Securities Industry Automation Corporation that provides a majority of the systems support for the NYSE and the AMEX as well as acting as the service provider for:

• the Consolidated Tape Association and the Composite Quote Operating Committee that consolidates market data for NYSE-listed securities;
• the Options Price Reporting Authority that consolidates options prices; and
• the Intermarket Trading System described above.

ArcaEx was among the most important ECNs but has become an exchange by purchasing the trading rights of the PCX for both equities and options. ArcaEx has significant market share in the Nasdaq marketplace, now trading as an UTP exchange. ArcaEx has been very successful in trading exchange-traded funds (ETFs). ArcaEx success is in part the result of a practice known as ‘outbound preferencing.’ It linked to most other ECNS and made the promise to prospective traders that if they sent their orders to ArcaEx and no trade was immediately possible, ArcaEx would route the order to other markets using the traders’ instructions. (Outbound preferencing was successful enough for other ECNs to copy the strategy.) Finally, ArcaEx believes that it has a real chance to get technology listings from Nasdaq.

The proposed merger would create the NYSE group. Apparently ArcaEx will continue as a separate entity unrelated to trading at the NYSE. The NYSE regulatory activities would be spun off into an independent regulatory organisation analogous to NASDR that was separated from Nasdaq. Significantly for the members of the NYSE, the new entity would be a for-profit company and seat holders would get shares in exchange for their current ownership of ‘seats’.

The current Hybrid Market proposal

The response of the NYSE to Reg NMS was a proposal for a ‘Hybrid Market’ – a combination of the traditional floor and an automated execution facility. At least for the moment this proposal is unchanged by the merger. The concept of a blend of a floor and automation is not new. The NYSE and all of the floor-based regional exchanges have had some form of automated execution primarily for smaller orders for years. More recently, the Chicago Board Options Exchange declared their combination of a floor and automated execution to be a hybrid market. Therefore, the NYSE’s proposal is evolutionary rather than revolutionary. Nevertheless, it represents a dramatic change for the Exchange.

The proposal has two primary components. First, the NYSE has developed a series of rule changes that will allow customers to specify that their orders to NYSE be executed purely electronically. Second, the Exchange has created a number of new order types, rules and procedures to ensure the opportunity for orders on the floor to interact with the electronic market and vice versa.

The new electronic trading facility is the most dramatic break with the past at the Exchange. Trading will take place on the Direct+ system. Direct+ has been in operation for four years but its functionality has been limited. The largest order currently permitted in the system is 1099 shares and one firm can only add a new order for a single security to Direct+ every 30 seconds. The reason that has been given for this limitation is to stop algorithmically generated orders from flooding the market. Even with these limitations, Direct+ averages more than one million trades per day for more than 160 million shares, making it a significant automated market. Interestingly, Direct+ was created to be a system to facilitate small retail orders but it quickly began to be used primarily by quantitative traders. In addition to providing broader appeal to those preferring automated trading and to providing the facility for the NYSE to comply with Reg NMS, the changes forever remove the presupposition that the floor is the dominant and preferred method of trading on the NYSE.

The flip side of the proposal is to provide new order types and procedures that permit the floor and the automated market to interact. Significant among these new facilities are:

• an auction limit order that permits a limit order sent to Direct+ to be executed on the floor if a better price exists there;
• the capacity for floor traders to provide orders that can participate with the crowd and the automated market in an automated fashion yet remained hidden from the specialist;
• new tools that permit the specialist to generate quotes that are controlled by algorithms (presumably one algorithm could be to always equal a better price in an away market);
• the facility for an institution to conduct a limited sweep of limit orders on the specialist’s book; and
• the insertion of Liquidity Replenishment Points (LRP) that act as circuit breakers in a period of rapid price movement.

All of these features were submitted to the SEC for approval in November and will be revised in the near future now that the final details of Reg NMS are known. Moreover, competitors and members will get to comment on the final proposal and official approval is likely to take a while. (Nasdaq’s SuperMontage took more than two years. Like the Nasdaq submission, approval of the Hybrid Market is likely to be a tough fight against competitors that try to fight the Exchange by challenging technical details of the proposal.) Both the timing and details of the final rule are certain to change.

Biases concerning future success

I have four assumptions or biases about what makes a market successful that I want to state explicitly. You may disagree, but you will at least be able to understand the basis of my reasoning. Also, if the assumptions are explicit, you will be better prepared to judge the validity of the conclusions.

Incorrect fact: Most trading globally is automated

If you ask most people, even knowledgeable professionals in the securities industry, how trading happens in most markets globally, they will tell you that with the exception of the NYSE substantially all markets are automated. They cite the Deutsche Börse, Eurex, Euronext, Liffe, LSE, Nasdaq, the exchanges in Asia and most of the markets in the third world. All of these markets seem to be automated. In the third world organisations such as the Asia Development Bank, the British Know-How Fund, the Peace Corps, USAID, the World Bank/IFC and other first world assistance groups have paid an army of systems houses and consultants to create and automate the emerging markets over the last decade. (I have been one of these troops.)

This belief fails to take into account the fact that the vast majority of automated trading takes place for small order sizes, where ‘small’ depends on the liquidity of the securities involved. In most cases large trades (again large is relative) take place in a dealer environment and are reported either late or not at all. This means that the majority of all trades (ie, transactions) are likely automated, but the majority of shares traded are probably still traded by some other means (ie, on floors or by phone.)

In fact, the most successful automated markets occur in those places where there is not much difference between large and small trades. Most trading in futures, options and ETFs has migrated to automated markets. In relative terms, large trades and small trades in these instruments are only a few hundred units different while the variation between a large trade and a small trade in many equity securities can be millions of shares in the normal course of trading.

New systems like Liquidnet and Pipeline are beginning to offer mechanisms for trading large quantities of securities using automation, but these systems still account for small proportions of all large trades.

Questionable assumption: The NYSE’s hegemony is protected by rigged rules

For the last thirty years people have argued that rules protect the NYSE from competition and if the protection was removed the Exchange could not survive. Chief among these was NYSE Rule 390 that prohibited NYSE members from trading NYSE-listed securities anywhere but on an exchange registered with the Securities and Exchange Commission (SEC). That rule was dropped in 2000, and there has been no noticeable impact on NYSE market share.

There are certainly other rules that provide the NYSE with distinct advantages, but these advantages would not matter if there were overwhelming benefits for its competitors compared to the Exchange. The best case in point is Instinet. Instinet was founded in 1969 and traded for about 15 years trying a variety of strategies to compete with the NYSE. Instinet was never successful. When Bill Lupien bought Instinet in the early 1980s, he switched the strategy to begin competing in the Nasdaq marketplace. The new strategy was quickly successful, and Instinet became the prototype for the ECN. I believe that the advantages protective rules provided to the NYSE were no more significant than the advantages that dealers had in the Nasdaq marketplace. The primary difference was that in the Nasdaq market, for many trading situations the interposition of a dealer between a customer buy order and a customer sell order was not necessary. Instinet provided a mechanism for agency executions in Nasdaq securities and this proved to be attractive, overcoming protective rules.

In a parallel example, the Chicago Board Options Exchange (CBOE) had a firm market share of about 40% of the US options market with the American Stock Exchange (AMEX) its nearest competitor having about 30%. These shares had been very stable for about twenty years when the International Securities Exchange (ISE) began trading in 2001. Currently the ISE is larger than the CBOE in most of the contracts they jointly list. This shift in market share occurred during unfavorable market conditions, and in spite of strong actions by the CBOE to defend its position.

I assert the following: When the benefits of a new system are overwhelming and/or the disadvantages of an old system are pronounced, no set of rules, however biased, can protect the old from the new.

Observed phenomenon: Trading style drives market preference

The automated markets have sets of rules and policies that are particularly adapted to certain types of trading strategy. For example, some ECNs reward putting liquidity on the book with fee discounts or rebates and/or charge for taking liquidity (ie, hitting a standing order on the book.) Markets have different types of reserve orders, execution rules and policies for accepting, sequencing and managing orders. These small differences can have a large impact on the desirability of different venues based on the specific goals of a would-be trader. Moreover, since the rules and charges are automated, they are applied rigidly.

Observed phenomenon: Traders are being isolated from market management

At both Nasdaq and the NYSE traders – floor traders, market makers and specialists – are becoming increasingly isolated from decisions about the strategy of markets. In some ways this is understandable. Traders can be a particularly ‘stroppy’ lot. They tend to be imperious and condescending. The staffs at trading venues all have countless stories about how members or traders blocked strategically important initiatives because of self-interest.

Unfortunately, the technologists that are replacing traders in managing markets are vulnerable to their own vanities. To be fair, some of the most successful systems are strongly influenced by technologists. But typically, most technologists believe that any trading situation can and should be automated. This may be true but it is not yet proven. More dangerous, they tend to believe that their particular conception of an automated market will work in any situation. As a result, when markets combine and technologists are in charge, they want to standardise on a single trading platform to reduce costs and simplify the technical architecture.

Even given the foibles of traders and the fact that technologists can be very smart, we have to be concerned if the strategy for our markets does not properly value the input of traders. Technologists, market staff members, the SEC and the author of this paper have never ‘written an order ticket in anger.’ However much we may know, it is Aristotelian science. We must value understanding born of experience.

The future of the Nasdaq

I believe that the success of the Nasdaq marketplace will depend on its ability unify the marketplace without homogenising the trading mechanisms. Nasdaq should continue to allow different traders with different styles to select the execution venue that best meets their needs. What it must do is find a method to allow each different venue to interact with other venues when an attractive opportunity exists. If Nasdaq can create this linkage it will be successful.

If Nasdaq tries to consolidate all trading on a single platform, then traders will become frustrated by the lack of execution models that meet their needs. We know now that building a new trading venue and developing a viable following can be done. The necessary regulatory infrastructure is in place. Following these mergers there are likely to be a small army of former executives with the required skills to build new markets who will be unemployed and free to create competitors to Nasdaq and the NYSE Group.

The future of the NYSE

For the NYSE the future is less clear. The new Hybrid Market proposal and the changes under Reg NMS are as conducive to success as the Exchange could have hoped for, given the pace of change. Of the four primary features of the NYSE that I believe account for its success, three remain unchanged under the proposed rule. The only uncertainty is whether the rules that finally emerge from the SEC will permit diverse constituencies to interact as effectively as they have in the past.

The single most likely threat to the NYSE is its own members who will soon be just investors. In past market structure fights, floor members (primarily specialists and two-dollar brokers) had very personal reasons to support the Exchange and the floor. For most members the floor was the source of their personal incomes. Now the remaining specialists are part of larger companies that trade on the NYSE because it is a profitable place to do business. Two-dollar brokers have become direct access members who are agents trading on the floor of the NYSE and any other trading venue that is attractive to their customers. They are loyal to the floor, but their business will not end if the floor goes away. Therefore the support of members now is less intense than in the past. Significantly, more than 950 of the seat holders on the NYSE do not actively trade, and lease their seats to others. For these holders the seat is a simple investment.

In fact, in the week following the passage of Reg NMS, 15 individual specialists received criminal indictments for their activities and five others were charged with offences that would result in civil penalties if proven. These charges are in addition to the USD240m settlement agreed last year by their firms for alleged abuses over several years ending with charges that were made public in 2003. Whatever the validity of the criminal charges, the following fact remains: in a fast-paced trading environment with penny increments and competing against algorithmic trading strategies, trading on the floor is more financially risky than it ever was before. Now it is obvious that trading on the floor is personally risky as well. The floor environment is very open to scrutiny and actions can be subject to second-guessing. A floor trader is more likely to be caught in an abuse, either intended or accidental, than a trader in an upstairs firm. If the floor-trading environment does not remain very profitable, firms and individuals will have to ask themselves if the risks are justified.

Finally, with a for-profit status, members of the exchange are less tied to the Exchange than they were in the past. They will be free to take their capital and leave and many will. Investors, who are only interested in the best financial opportunity will replace members, who have a commitment to the floor. Investors will be less inclined to see the subtle values of the floor process. Investors are more likely to demand short run profit that can be realised by closing the floor to save cost.

There is a second threat that was stated above. If its competitors are better able to compete with the NYSE in the wake of Reg NMS, I do not believe that the NYSE can survive as one of a group of roughly equivalent market centres with a share of 30% to 40% of the total market volume. The NYSE is a high cost producer. Without significant volume to absorb the costs it will be uncompetitive. The NYSE also has the advantage that it is the listing market and listing fees are significant. Its dominance as a trading centre and the prestige of the floor – the ability to ring the opening bell – represent significant draws for listed companies. If trading dominance declines the NYSE’s hold on listings is problematic.

Looking more specifically at the changes implicit in the Hybrid Market and the linkage to ArcaEx, the Exchange can no longer force people to trade on the floor. If traders prefer automated execution they are free to choose it. The NYSE has been very clever in developing new rules, systems and procedures as part of the Hybrid Market proposal. These changes may sustain the Exchange’s historic advantages such as benign opacity – the elements that have made the NYSE successful. It is hard to guess whether these elements will survive competitive assaults or if they will be enough to preserve the Exchange’s preeminent position.

Summary

A director of a major broker-dealer once argued with me, after the SEC’s Order Handling Rules were announced in 1995, that the new rules meant the end of the Nasdaq market. The market did not end, but it sure is different. The events of April 2005 are unlikely to mean the end of either market or a return to the status of the past. However, both the Nasdaq and the NYSE markets will be fundamentally different. Whatever happens, firms will still have to raise capital and they will find a way to raise it. Firms and individuals with securities or cash positions will still need to trade and they will find a way to do it. Changes to Nasdaq and the NYSE will not be the end of capitalism, but it is likely to change the face of the markets in the United States.

In conclusion, this much is certain. If Nasdaq, the Exchange or both decline in importance some of the central functions they perform will have to be replaced. Among these required functions are:

• Large orders must find a place to be transacted. They may revert to a purely dealer environment, or negotiating systems such as Liquidnet and Pipeline may become the primary means of block execution.
• A new approach to market data collection and dissemination will have to be developed. If nothing is done, market data fragmentation will escalate.
• A listing entity may need to be created. We could develop a system like that in the United Kingdom where listing is a governmental instead of a marketplace function. If we created such a system, any market would be free to trade securities properly listed with the SEC. Significantly, the markets would no longer receive listing fees.

The end of the dominance of Nasdaq or the NYSE will not be the end of capitalism. Their functions will still be required. Whatever transpires, the future will be very different to the past.

As a confirmed observer, it will be fascinating to see this Brave New World unfold and watch the wondrous people (and things) in it.