April 5, 2000,

Jonathan G. Katz
Secretary
U.S. Securities & Exchange Commission
450 Fifth Street, NW
Washington, DC 20549-0609

Re: Release No. 34-42450; File No. SR-NYSE-99-48; Market Fragmentation

Dear Mr. Katz:

I am a professor at the Zicklin School of Business, Baruch College, CUNY. I am also a consultant to, and an investor in, AZX, Inc. I welcome this opportunity to comment on the Commission's release seeking public views on market fragmentation.

In an earlier release, "Regulation of Market Information Fees" (Release No. 34-42208; File No. S7-28-99), the Commission emphasized that retail investors must have access to market information, that information must be consolidated, and that transparency must be enhanced:

One of the most important functions that the Commission can perform for retail investors is to ensure that they have access to the information they need to protect and further their own interests (page 7).

Although it intended to rely on competitive forces to the greatest extent possible to shape the national market system, Congress also recognized that the Commission would need ample authority to achieve the goal of providing investors and broker-dealers with a central source of consolidated market information (page 33).

The consolidated, real-time stream of market information has been an essential element in the success of the U.S. securities markets. It is the principal tool for enhancing the transparency of the buying and selling interest in a security, for addressing the fragmentation of buying and selling interest among different market centers, and for facilitating the best execution of customers' orders by their broker-dealers. Broad public access to consolidated market information was not the fortuitous result of private market forces, but of planning and concerted effort by Congress, the Commission, the SROs, and the securities industry as a whole (page 5).

The current release adds the thought that limit orders and quotes may be isolated from full interaction and that price competition may be insufficient:

The Commission is concerned, however, that customer limit orders and dealer quotes may be isolated from full interaction with other buying and selling interest in today's markets (page 3).

... a market structure that provides a full and fair opportunity for interaction of investor trading interest may, by enhancing price competition, reduce the transaction costs of investors who submit market orders (page 9).

The consolidation of orders and trading information is indeed desirable. Consolidation focuses liquidity, lowers trading costs, and enhances participants' ability to transact at prices that reasonably reflect the broad market's desire to hold shares of a stock. Unfortunately, however, the Commission cannot become involved with the production, distribution, and pricing of information without interfering with the natural formation of a marketplace. Issues concerning consolidation are complex and I believe the Commission would be best advised to let the order flow consolidate naturally. Legislated requirements or prohibitions concerning the consolidation of orders can have consequences that are difficult to predict and very hard to control. I believe it preferable to leave this important aspect of market architecture up to the market centers to resolve as they fight for order flow in a competitive environment.

I further believe that the Commission's consideration of order flow consolidation in the current release is inappropriately restrictive. The release addresses the spatial dimension of the issue, but completely ignores the temporal dimension. The restricted focus is particularly serious because, in my view, the two dimensions are not independent.

The Commission is not alone in ignoring the time dimension. Unfortunately, virtually all of the public debate concerning order flow fragmentation/consolidation has focused on the spatial distribution of orders across multiple markets. However, the time dimension is every bit as important. To achieve trades at reasonable prices, orders must meet. As with any meeting, two dimensions must be specified: place and time. In continuous trading environments (e.g., the New York Stock Exchange and Nasdaq), trades can be made any time two counterpart orders meet or cross during normal trading hours (9:30 a.m. to 4:00 p.m.). Assume one participant submits a buy order at 10:50 a.m., and that a second participant submits a sell order at 10:55. If these orders cannot be brought together, the market is fragmented. How do orders meet in time?

In a quote driven market, a market maker solves the time problem by selling to the buyer at 10:50 and buying from the seller at 10:55. In an order driven market, the limit order placed by one participant enables another participant at another moment in time to trade with immediacy by market order. A third alternative is the call auction. The call enables a large number of buyers and sellers to meet because it establishes a predetermined meeting point in time.

I suggest that the introduction of electronic call auction trading along with continuous trading would be the most useful innovation that could be made in U.S. equity market structure. The call auction is an excellent facility for consolidating market information, achieving the full interaction of investor trading interest, and enhancing price competition. More attention is currently being given to electronic call auction trading in European markets but, unfortunately, this trading facility is not widely known or understood in the U.S. Consequently, it is not surprising that the Commission made no mention of it in the current release and, as noted above, ignored the temporal dimension of order flow consolidation/fragmentation.

The temporal consolidation of the order flow would also facilitate spatial consolidation. In continuous trading, a satellite market that does not undertake the surveillance and other self-regulatory activities of a major market center can give less expensive, faster executions at prices established in the major market center, and the continuous posting of quotes facilitates its doing so. This free-riding problem does not exist in call market trading because the only way to trade with certainty at a call market price is to participate in the call. During the order accumulation period before the call is held, only indicated clearing prices are shown (the actual price has not been determined yet) and, after the call has been completed and the actual clearing price determined, the process is complete and the call auction is closed.

A continuous market encourages both the fragmentation of orders across markets (because of free riding) and the fracturing of orders over time. The fracturing of orders occurs for two reasons. First, large orders are broken up and fed to the market in smaller pieces to reduce market impact. Second, large orders are broken up and fed to the market over the course of a trading day in an attempt to obtain a stock's volume weighted average price (VWAP). This order fracturing is serious: it thins out liquidity, encourages momentum trading, makes it more difficult to realize trades at reasonable prices, and accentuates intra-day price volatility. Moreover, it appears that order fracturing has increased in the current environment. I urge the Commission to consider the issue.

The Commission's focus on best execution addresses the problem a participant faces when coming to a market to trade. One could view the process as the "production of a trade." Alternatively, one could consider many investors' orders coming together in the marketplace, and view the process as the "production of a price" (price discovery). Order flow fragmentation can certainly impede the production of a trade; it also disrupts price production, with serious consequences for the broad market. I believe that with respect to price discovery, the more serious form of fragmentation is temporal, not spatial.

In conclusion, I believe that the Commission should recognize problems concerning the fracturing of orders over time, as well as the fragmenting of orders across markets. I suggest that problems concerning both temporal and spatial fragmentation can be mitigated by the introduction of electronic call auction trading. However, I urge that the Commission not take a prescriptive position with regard to this market structure development, but instead allow it to occur naturally.

Parts of this letter have been drawn from my paper, "The Call Auction Alternative," which is Chapter 1 of a book I am editing, Building a Better Stock Market: The Call Auction Alternative, Kluwer Academic Publishers, 2000, forthcoming. A copy of this paper is appended. Please do not hesitate to contact me if you have any additional questions.

Sincerely yours,

Robert A. Schwartz





The Call Auction Alternative

by

Robert A. Schwartz
Marvin M. Speiser Professor of Finance
And University Distinguished Professor
Zicklin School of Business
Baruch College, CUNY

Chapter 1 for

Building a Better Stock Market:
The Call Auction Alternative

Robert A. Schwartz, Editor
Kluwer Academic Publishers, 2000, forthcoming

Revised Draft, May 9, 2000



The Call Auction Alternative1

In just the past few years, the U.S. equity markets have experienced, among other things, the growth of electronic communications networks (ECNs), the breathtaking expansion of the internet's use for order routing,2 and the NASD's acquisition of the American Stock Exchange. Around the world, we have also witnessed the disappearance of trading floors.3 The astonishing pace of change is attributable to the convergence of three powerful forces: advances in computer technology, intensified competition, and regulatory intervention in market structure.

The ECNs are electronic trading platforms that match customer orders anonymously without the intervention of middlemen. They have made substantial inroads into the Nasdaq market, and their threat of capturing NYSE order flow is forcing the Big Board to rethink both its trading system and organizational structure. What should be the competitive responses of Nasdaq and the New York Stock Exchange? What effect will these profound changes have on the provision of dealer services and on the quality of our markets?

This chapter surveys the current scene and addresses these questions. In so doing, we give particular attention to issues concerning the price discovery function of a market place, and to problems concerning liquidity, volatility, and order flow consolidation that have not been resolved after a quarter of a century of extensive analysis and public debate. We also consider the role of a call auction in a hybrid market structure.4

A call auction is a form of trading that died out in the pre-computer age but is making its reentrance today as an electronic marketplace. In contrast with a continuous market where a transaction is made any time a buy and sell order meet in price, in call auctions orders are batched together and executed in multilateral trades at specific points in time.5 Recent advances in computer technology have considerably expanded the call auction's functionality.6 We suggest that the problems we are facing concerning liquidity, volatility and price discovery are largely endemic to the continuous market, and that the introduction of electronic call auction trading in the U.S. would be the most important innovation in market structure that can be made. Use of the call auction in extended trading hours, a time when problems of market quality are more accentuated, would also be desirable. Whichever market first makes the innovation, be it the NYSE, Nasdaq or an ECN, that market will appreciably increase its ability to attract order flow.

Countless design possibilities exist for building a stock market. On a fundamental level, however, we have only three generic structures: (1) the continuous agency/auction (order driven) market, (2) the continuous dealer (quote driven) market, and (3) the periodic call auction (which is also order driven).7 Currently, much attention is focused on continuous, electronic limit order book markets (e.g., the ECNs). With a limit order book, public buyers and sellers meet without the intervention of a dealer because the limit orders of some public participants establish the prices at which other public participants can trade by market order.8

An order driven platform may work well for retail order flow in liquid stocks under non-stressful conditions. But when markets come under stress, additional structural support is needed. The advent of news puts pressure on price discovery, as does the daily opening of the market. Market closes and the expiration of derivative contracts are generally stressful times. The arrival of a 100,000 share buy or sell order is stressful. Investors occasionally panic and prices go into free fall. One day the current bull market will end, a flat or bear market will ensue, and the problems of maintaining a fair and orderly market will become more apparent. Thus, the other two generics - market makers and the call auction - are essential for today's marketplace.

Dealers, specialists, and upstairs block positioners have historically been important providers of immediate liquidity in the U.S. equity markets, and the broker-dealers must be appropriately compensated. The bid-ask spread has traditionally been considered the source of their compensation. Broker-dealer operations, however, have been forcefully impacted in recent years by the tremendous transformation of our securities markets. After the dust has settled, will their services be appropriately applied? Perhaps not. If our markets increasingly become disintermediated, the call auction will acquire enhanced importance. As a predetermined point-in-time meeting place, it enables the natural buyers and sellers to provide liquidity directly to each other.

Its Not a Revolution, its an Earthquake

As we have noted, the equity markets in the U.S. and indeed around the world are being reshaped by the simultaneous convergence of three powerful forces: technology developments, intensified competition (both domestic and global) and regulation. Each of the three may be viewed as desirable or well intentioned, but their joint impact is producing results that are difficult to predict, hard to control, and not easy to understand. What we are witnessing does not resemble a revolution. It is an earthquake.

Over the years, we have witnessed the re-engineering of new financial products, the emergence of new investment styles, and the employment of new quantitative trading techniques. Commissions have dropped sharply. Three decades ago, mutual funds, pension funds and other institutional investors were in their infancy; today they dominate the market. Nasdaq has become a powerful market center along with the New York Stock Exchange. And now, alternative trading systems (including the ECNs) have emerged, that are attracting significant order flow away from market centers.

Thirty years ago, daily price and volume data had to be collected by hand. As a by-product of computer technology, we now have electronic transaction records for individual stocks that include all quotes, trades, and volumes complete with times stamped in fractions of a second. The data have enabled users to see how execution costs can erode portfolio performance. Not surprisingly, investors are seeking to control these costs. From the individual's viewpoint, this can be done by one's selection of a broker and by careful order handling. Execution costs can also be controlled by asset selection (avoid illiquid stocks) and by market selection (avoid illiquid marketplaces). But ultimately, the only way to reduce costs across the board for all issuers and all investors is by building a better stock market.

The Long Arm of Regulation

Congress's involvement with market architecture issues started in the 1960s when it ordered the SEC to sponsor the Institutional Investor Study. The Report, filed with the SEC in 1971, focused largely on the competitive structure of the securities markets, the profitability of specialist operations, and the behavior of institutional investors (a group that, while far larger now, was at the time big enough to prompt the study).

Next, in a major overhaul of the Securities Exchange Act of 1934, Congress passed the Securities Acts Amendments of 1975. The Amendments precluded the securities exchanges from imposing fixed commission rates and mandated the development of a national market system (NMS). The two motivations for the NMS mandate were: "the maintenance of stable and orderly markets" and "the centralization of all buying and selling interest so that each investor will have the opportunity for the best possible execution of his order, regardless of where in the system it originates."9 But best execution cannot be achieved without information, and this led the SEC to believe that the development of a central market system requires that all price, volume, and quote information be available to all investors, for all securities, in all markets.10 Recently, the SEC stated:11

One of the most important functions that the Commission can perform for retail investors is to ensure that they have access to the information they need to protect and further their own interests (page 7).

Although it intended to rely on competitive forces to the greatest extent possible to shape the national market system, Congress also recognized that the Commission would need ample authority to achieve the goal of providing investors and broker-dealers with a central source of consolidated market information (page 33).

The consolidated, real-time stream of market information has been an essential element in the success of the U.S. securities markets. It is the principal tool for enhancing the transparency of the buying and selling interest in a security, for addressing the fragmentation of buying and selling interest among different market centers, and for facilitating the best execution of customers' orders by their broker-dealers. Broad public access to consolidated market information was not the fortuitous result of private market forces, but of planning and concerted effort by Congress, the Commission, the SROs, and the securities industry as a whole (page 5).

However, a regulatory authority cannot become intimately involved with the production, distribution and pricing of market information without, at the same time, interfering with the natural formation of a marketplace. For this reason, despite the SEC's statements that market structure should be determined by competition and not by the regulators, the Commission in recent years has become ever more deeply involved in the design of U.S. trading platforms. The effectiveness of this involvement has been questioned by a number of industry participants and academicians.12

In 1994, a paper by two professors, William Christie and Paul Schultz, was published in The Journal of Finance that had far reaching consequences for the U.S. equity markets in general, and broker-dealer firms in particular.13 Christie and Schultz had found that Nasdaq dealers were commonly avoiding odd-eighth quotes (e.g., 60 5/8), and the authors suggested that dealers were "implicitly colluding" to keep spreads artificially wide. An investigation ensued of the Nasdaq market and Nasdaq dealers by the Department of Justice and the SEC, and a class action lawsuit was filed. Approval of the court was issued on November 9, 1998 of a settlement in the aggregate amount of $1,027 million.14 Further, communications between broker-dealers about market conditions were also discouraged by a Department of Justice requirement that their conversations be taped. This has made price discovery, particularly at market openings, a good deal more difficult.

Continuing to press for greater transparency of price and quote information, the SEC in 1997 instituted new order handling rules. First, the Commission required that any market maker holding a customer limit order must display that order in his or her quote. Second, the new rules stated that if a market maker has placed a more aggressively priced quote in an ECN, the market maker is okay if that ECN displays the top of its book in the Nasdaq quote montage. However, if that ECN's own best quotes are not in the quote montage, then the market maker must update his or her own quote in Nasdaq to match its ECN quote.

What have these changes accomplished for investors? Liquidity pools have fragmented, not consolidated. Intermediary profitability has been impaired for many market maker firms. Spreads have tightened, but it is not clear that total execution costs have fallen for institutional customers. Price volatility has been high. Price discovery is more difficult, especially at market openings. Enormous price movements are commonly observed at initial public offerings. Will the markets be resilient in the face of a protracted flat or bear market? Will customers get best execution for their orders if prices head south? Are they getting best execution now?

It is apparent that a better trading system is needed. With this objective in mind, I suggest that the introduction of electronic call auction trading in the U.S. would be the most useful innovation that can be made.

The Current Scene at Nasdaq

In a continuous agency/auction market, the rules of order execution generally lead specialists and limit order traders to compete for incoming market orders by raising their bids, lowering their offers and, in the process, narrowing the inside spread. In a quote driven market, two dealer practices, preferencing and quote matching, diminish the incentive for market makers to do this.15 Instead of simply tightening their quotes, the market makers compete for order flow by establishing on-going relationships with their customers and by quote matching when orders are preferenced to them, even though their own quotes may not be on the inside market. In this environment, dealers do not welcome competition from public limit orders and, indeed, public limit orders have historically been disadvantaged in the Nasdaq market.16

The SEC's new order handling rules have made it easier for the ECNs to capture order flow, and the ECNs have provided a limit order facility for Nasdaq stocks. As of September 1999, the ECNs had roughly 29 percent of the dollar volume, and 22 percent of the share volume in Nasdaq trading. As shown in Table 1, Instinet is the largest ECN with 56 percent of both dollar and share volume, and Island is the second largest with 27 percent and 25 percent of dollar and share volume, respectively.

Narrowly viewed, the ECNs may appear to be taking market share away from Nasdaq. Broadly considered, this is not the case. The ECNs have added a desired feature to the market for Nasdaq stocks - a public limit order facility - and this has presumably helped Nasdaq retain listings that might otherwise have been lost to the NYSE. On the other hand, one ECN, Archipelago, announced in March 2000 that it is combining with the Pacific Stock Exchange, and two others (Island and Nextrade) have filed with the SEC for exchange status. These changes will diminish the reach of the Nasdaq market.

The ECNs are being viewed as a competitive threat to the traditional markets both in the U.S. and abroad. Even more striking, the preponderance of stock exchanges around the world now have electronic trading platforms. The electronic systems are fast, inexpensive, eliminate intermediaries, and provide anonymity. Electronic technology has made it possible to combine systems in new ways so as to give users some of the options they want for working their orders. However, as noted, because of order preferencing, quote matching, and the inapplicability of time priorities in a quote driven market, limit order trading does not co-exist smoothly, side-by-side, with dealer trading, regardless of the technology employed. Fortunately, an alternative exists.

The electronic call auction should be used at Nasdaq's daily openings (currently, no special facility is available). Further, Nasdaq's continuous dealer market could periodically be punctuated by a call. Rather than operating in parallel with the dealer market, the public limit orders would be at "right angles" to the continuous market.17 Public limit orders would naturally be attracted to the calls, and dealers would remain the dominant source of immediate liquidity in the continuous market.

Aggregate order flow for Nasdaq shares could increase appreciably if an electronic call auction were made available on a voluntary basis for Nasdaq issues. The calls would reduce immediacy costs for investors, would be an important price discovery mechanism for the broad market (especially at market openings), and would provide a means of laying off inventory positions for dealers by facilitating inter-dealer trading. With direct verbal communication about market conditions between dealer firms now discouraged, it is imperative that an alternative communication device be established. The electronic call auction would be such a device. And with public limit orders drawn to the calls, dealers would be freer to compete with each other in the continuous market in whatever manner they deem to be most natural, given the inherent nature of their market.

The Current Scene at The NYSE18

Competition is also hitting the Big Board with the force of a one hundred-year storm and the Exchange, as we know it, is at risk of going under. There are the new alternative trading systems and ECNs. Also on the scene are the dramatic growth of electronic brokerage and day trading, the extension of trading hours, and the new acquisitions and alliances that are being formed. Responding to these developments, the NYSE is considering entering the off-hours market and is formulating plans to go public. Where will all this leave the Exchange?

The NYSE, of course, is by no means a perfect market. Institutional investors complain about the string of intermediaries both on and off the trading floor who might handle an order - sales traders, position traders, floor brokers, and, of course, the NYSE specialists. The intermediaries have to be compensated. The process takes time. Even worse, institutions lose anonymity and fear others trading ahead of their orders, thereby increasing prices when they are buying, and decreasing prices when they are selling. Little wonder that the alternative markets are attractive.

The contest resembles a chess match between a grand master and a computer. The more sophisticated the computer, the more it can do what the grand master does, the greater is the computer's chance of winning. But price discovery is a complex game to play and the grand master is pretty good. Thus far, the ECNs are trading only about five percent of NYSE volume. We should not be too quick to count the NYSE's trading floor out.

The Exchange's trading floor is informationally rich because it receives a substantial percentage of the order flow for its listed stocks - 84% of listed share volume was executed on the NYSE in 1998. NYSE specialists and other floor traders facilitate price discovery by handling orders, both large and small, when the orders are matched together to make trades and establish transaction prices. Large trades are commonly negotiated in the upstairs market with reference to NYSE floor information and are brought to the floor for execution. A large order may also be given to a floor trader who can flexibly feed it to the market in smaller pieces in light of current conditions and customer needs.

Nevertheless, the Exchange has historically lost order flow to other markets, and this leakage is threatening to increase. As noted above, a critical cause of the fragmentation has been the ability of others to free-ride on NYSE-discovered prices. Now more than ever, the NYSE is threatened by order flow being sucked away by electronic markets that can provide cheaper, faster and anonymous order handling while making trades based on prices set on the NYSE. If the NYSE does not solve the free-rider problem and keep order flow consolidated on its trading floor, it will loose the technology battle.19

Armed with electronic technology, the NYSE's competitors are also challenging the Exchange to enter a new arena - the extended hours market for both retail and institutional order flow. The extended hours market will likely remain illiquid, without good price discovery, and the Exchange's trading floor will not be there. The NYSE should be very cautious about following its competition into these unfamiliar waters.

The alternative electronic markets are privately owned business firms, not membership organizations. As such, they are free of impediments to technological innovation that can beset a membership exchange. They are better able to raise capital, form alliances, and make acquisitions without resistance from members. The exchanges of Stockholm, Australia, Amsterdam and elsewhere have demutualized. The London Stock Exchange is considering doing the same. In December 1999, Deutsche Börse announced its plans for an IPO and, on January 4, 2000, Nasdaq announced its intention to become equity based. Should the NYSE follow suit? Despite all the obvious positives, the governance change could have serious implications for the NYSE's self-regulatory responsibilities and the operations of its trading floor. Demutualization, which may work for some other markets, could be the end of the NYSE's dominance.

Whether because of information consolidation, flexible order handling, the Exchange's self-regulatory status, or whatever else might be making it a viable ecology, the NYSE's trading floor is currently receiving the lion's share of the order flow for Exchange-listed stocks, and its floor is performing meaningful price discovery. But with the threat of fragmentation looming large, what can the NYSE do to survive?

The NYSE will not beat the alternative markets at their own game. It must use technology differently. The electronic systems are providing fast order handling and cheap executions in a continuous market environment where a trade is made whenever a buy and a sell order match in price. The NYSE should counter by focusing liquidity at specific points in time. It can do this by periodically punctuating its own continuous market with electronic call auctions (perhaps as frequently as once an hour for its big stocks at least). Doing so will enable the NYSE to retain its order flow. This is because it is not possible to free ride on a call auction.

Thus far, however, the call innovation has been resisted, and it is not at all clear that, as a membership organization, the NYSE would approve this change in its market structure. The Exchange is indeed on the horns of a dilemma. Nevertheless, the bottom line is, a sea change is shaping up. Let's hope the Big Board will appropriately strengthen its trading system and weather the storm.

The ECN's

Technology developments have enabled innovations in security trading systems that could not have been made just a few years ago. In particular, technology has accelerated the speed with which orders are handled, translated into trades, and reports made. Technology has also extended connectivity and offered enhanced computational power.20 Speed and connectivity, in particular, have played to the ECNs' advantage, and their growth in the past couple of years has been eye catching.

Regulation has also facilitated the ECNs' inroads into Nasdaq trading. The SEC's new order handling rules (the limit order display rule and the market maker quote rule) have made it much easier for a new ECN to capture public order flow. All the new ECN needs is to be a gateway that will attract some customers to place limit orders on its electronic book. Connectivity with other markets (either directly via linkages or indirectly through Nasdaq's SelectNet system) will enable market orders from other firms' customers to reach its book and trigger trades.21

But technology and regulation are not enough to produce good markets. Both can have unintended consequences that are difficult to undo once they occur. Currently, the SEC enforced consolidation, transparency, and accessibility of price information are causing the flow of limit orders to fragment onto multiple books, and the ECNs' cheep, fast, anonymous, and extended hours trading is forcing Nasdaq and the NYSE to alter their trading systems and organizational structures. After the dust settles, what will be the quality of the U.S. equity markets? The situation is questionable. To date, liquidity pools have fragmented, market maker operations have been perturbed, price discovery is more difficult and intra-day price volatility is accentuated. The current environment has encouraged momentum trading and this in turn is destabilizing. New solutions to old problems are needed. To find them, let us consider the call auction.

The auction's potential is sizable. The first U.S. market to properly incorporate a well structured call will attract the order flow. As this occurs, the broad market - including both investors and listed companies - will benefit from the innovation. We now turn to a more detailed assessment of this trading facility.

The Call Auction

An electronic call auction has been incorporated in recent years in a number of market centers around the world, most notably the ParisBourse and Deutsche Börse.22 These electronic calls are not being used as stand alone systems, but have been combined with continuous trading to create hybrid markets. The reason is that, when it comes to handling participants' orders, one size does not fit all. With a hybrid trading system, an investor can more easily select among alternative trading venues depending on the size of his or her order, the liquidity of the stock being traded, and the investor's own motive for trading.

Four proprietary trading systems in the U.S. are based on call auction trading principles: Reuter's Instinet Crossing Network, Investment Technology Group's Posit, AZX's Arizona Stock Exchange, and OptiMark Technologies. Additionally, most electronic markets such as the Toronto Stock Exchange's Computer Assisted Trading System (CATS) and most floor-based markets, including the New York Stock Exchange, open trading with a call auction.23 Along with opening the market, calls could also be held twice during the trading day (e.g., at noon and at the close) or more frequently if desired (e.g., every hour).24 In the Far East, the Tokyo and Korea Stock Exchanges open and close both morning and afternoon sessions with call auction trading. The Taiwan and Kuala Lumpur Stock Exchanges go even further; they offer no continuous trading at all, but instead run calls every several minutes for their stocks.

Over one hundred years ago, the New York Stock Exchange was a call market (non-electronic, of course). In some respects, the non-electronic call was a fine system for participants on the exchange floor, but it certainly had deficiencies for anybody away from the floor. Investors not physically present had little knowledge of what was happening (the calls offered no transparency) and access to trading was limited because shares of a stock could be exchanged only periodically (when the market for the stock was called). On May 8, 1869, the call procedure was abandoned when the NYSE merged with a competing exchange, The Open Board of Brokers, and became a continuous trading environment. The Tel Aviv Stock Exchange through the 1970s and the ParisBourse before the 1986 introduction of its electronic market, CAC,25 also were non-electronic call auctions that did not survive. But in recent years, tremendous advances in information technology and a slew of other developments in the industry have paved the way for the call's reentry.

Batching orders for simultaneous execution at a single moment in time is the essence of call auction trading. The principle of bringing orders together to determine a clearing (consensus) price is discussed in more detail in the Appendix and is illustrated with the diagrams it contains (see Figure A1). Because its determination is based on the full set of orders, the clearing price in a call auction can be thought of as a "consensus value." While the discussion of Figure A1 clarifies this and the principle behind call market trading, it is important to note that all call auctions are not alike.

Economides and Schwartz identify four different auction designs.26 (1) In a price-scan auction, an auctioneer announces tentative prices and traders state their buy/sell responses until the price that best balances the buy and sell orders is found.27 (2) In a sealed bid/ask auction, traders submit priced orders that are not disclosed to one another.28 (3) In a crossing network, traders submit orders that are matched at a price determined in some other market (i.e., trades are priced for the Posit crosses using mid-spread values established in the major market center at the time of a cross). (4) In an open order book auction, traders follow the market as buy and sell quantities are cumulated and displayed at each price, along with a continuously updated indicated clearing price, until the market is called.

I first learned of call auction trading in the mid-1970s, just prior to participating in a conference in Israel. The talk I planned to give contained a discussion of bid-ask spreads. The Tel Aviv Stock Exchange, however, was using non-electronic call auction trading at the time, which meant that this part of my presentation had to be eliminated - there is no bid-ask spread in a call auction. It has subsequently become apparent to me that the differences between the call and continuous environments extend well beyond the bid-ask spread.

Call auctions, because they focus liquidity, have commonly been thought appropriate mainly for small cap, less frequently traded stocks. However, they also have particular appeal for the large caps because they cater to the needs of institutional participants whose portfolios are mostly comprised of these issues. Market impact is reduced for the institutional investor because the call is a point in time meeting place where orders are batched together for a multilateral trade. Commissions may be lower due to the greater ease of handling orders and clearing trades in the call auction environment. For the broad market, call auctions can reduce price volatility, unreliable pricing, unequal access to the market, and various forms of manipulation and abuse.

One feature of call auction trading in particular deserves attention.29 The call auction is an explicit price discovery facility. That is, batching many orders together for simultaneous execution at a single price produces a consensus value that better reflects the broad market's desire to hold shares. Consequently, the call auction is a good opening facility for both the order driven (exchange) market and the quote driven (dealer) market.30 Moreover, because it is an explicit price discovery facility, call auction trading can be used to dampen intra-day price volatility. We consider implications of call auction trading for price discovery and volatility in the next two sections of the paper.

Price Discovery and Volatility31

Roughly fifteen years ago, while considering with a small group of academicians what the unique service is that is provided by an exchange, William Batten, who at the time was Chief Executive Officer of the New York Stock Exchange said, "We produce the price." The insight is of major importance.

Little regulatory or academic attention has been given to price discovery, presumably because no obvious benchmark exists against which to assess the accuracy of a transaction price. Regulators and academicians have given far more attention to bid-ask spreads (which are visible and easily measured). However, sharpening price discovery may be considerably more important than tightening spreads.32 The relative importance of these two objectives - tight spreads and accurate price discovery - is suggested by the share price movements shown for Pfizer, Inc., in Figure 1.

Figure 1 presents a half-hour (11:00 - 11:30) and a full day (9:30 - 4:00) price chart for Pfizer for January 20, 2000. In the half-hour chart (Figure 1a), Pfizer's price can be seen bouncing within a one-sixteenth range over a succession of trades. For two different time intervals in the half-hour period, the bounce is particularly pronounced, and it has presumably traced out Pfizer's bid and ask prices. For these intervals, we have placed dash lines that denote the higher ask and the lower bid. It appears that, except for the price changes around 11:20 which show that the location of the spread has moved up, the spread is the sole cause of volatility in an otherwise stable environment.

We can still discern tiny price movements that presumably reflect the bid-ask bounce in the full day price chart (Figure 1b), which includes the half-hour period discussed above. Notable in the full day chart are the considerably larger price swings that have occurred over the trading day as Pfizer moved from an opening price of 36.81 to close down 2.72% at 35.81. What explains the multiple intra-day swings seen in Figure 1b? Could it be a succession of news releases from Pfizer? No. News releases from other companies and Washington? Possibly. Changing investor sentiments independent of informational change? Plausible. A hunt and search process that characterizes price discovery in continuous trading? Most likely. Price discovery no doubt does explain some of the observed volatility, and achieving reasonable price discovery may indeed be more important than reducing the bid-ask spread.33

The difficulty of achieving accurate price discovery has not been adequately appreciated. The fact is, no one knows the right price for a stock until investors' orders have been submitted and trades made. However, because both order entry and execution are staggered over time, many trades are made at prices that are not consensus values. Additionally, the execution of large institutional orders can dislocate prices. Inaccuracies in price discovery inflate intra-day price volatility, raise trading costs for participants, and shake investor confidence in the reasonableness of the executions they get. Accordingly, in our considerations of market structure and regulation, we should shift emphasis from the size of the spread to the location of the spread (i.e., to the accuracy of price discovery).

Consolidating order flow for individual shares, all else equal, is expected to improve the accuracy of price discovery. Because of the network externalities in trading, order flow naturally attracts order flow, and thus consolidation may occur naturally in an equity market.34 As orders are pulled into a single market center, however, the regulatory authorities, despite their above noted objective of consolidating market information, become increasingly concerned that the market center will acquire a monopoly position.

The tendency for orders to consolidate may be countered by a satellite market free-riding on price discovery. That is, a satellite market that does not undertake the surveillance and other self-regulatory activities of a major market center can give less expensive, faster executions at prices established in the major market center. As this occurs, the regulatory authorities become increasingly concerned about fragmenting order flow.35

Another important dimension is the temporal batching of orders for simultaneous execution in periodic call auctions. As we have discussed above, a great deal of regulatory attention has been given to spatial pooling of orders in a single market or order book. Little has been given to temporal consolidation. Temporal consolidation is the essence of call market trading.

Interestingly, the temporal consolidation of orders in call auction trading can also facilitate spatial consolidation. The reason is that it is not possible to free-ride on call market prices. During the order accumulation period, indicated clearing prices are shown (the actual price has not yet been determined) and, after the call has been completed and the actual clearing price determined, the process is complete and the call auction is closed. This contrasts with a continuous market where the quotes at which trades can be made are posted continuously on the market.

In a continuous environment, trading progresses while prices are being discovered. The process might be thought of as discovering prices with live bullets. In call auction trading, on the other hand, tentative, indicated clearing prices are displayed and the batched trades are not set until the price discovery process has been completed. Thus, not only are prices more accurately discovered in a call auction, but all investors participating in the auction trade at the better-discovered prices.36

Price Discovery in a Market Under Stress

From time-to-time, a continuous market loses its ability to discover prices appropriately, especially under conditions of major informational change. As prices start to move, momentum trading can kick in and accentuate any adjustment. The standard regulatory response to a market under stress has been to make trading more difficult - i.e., "to put sand on the wheels" or even to stop trading altogether. However, it is not traders in general who cause the problem; specifically, it is the momentum players. The objective should not be to hinder trading, but to diffuse momentum. Proper use of a call auction would accomplish this.

When prices in the continuous market de-couple from underlying consensus values, trading should be halted briefly so as to make the transition from the continuous trading modality to the call auction modality. Stopping the market under conditions of stress is referred to in the United States as a circuit breaker.37 The circuit breaker was activated twice in the U.S. equity markets on October 27, 1997, a day the Dow Jones Industrial Average fell a total of 554 points from 7715 to 7161. Was the procedure effective? Not as it was implemented on October 27 but, if properly used, it could be.38

It has been thought that closing the market would give traders time to cool off. So what happened on the afternoon of October 27, 1997 after a 350 point drop in the Dow triggered the first circuit breaker? Upon reopening, traders sensing that the market would be closed again, shot in their sell orders and, as a self-fulfilling prophecy, the market plunged another 200 points in half an hour. A second circuit breaker then kicked in that stopped trading for the rest of the day. Clearly, it is not easy to control the emotions of traders and keeping them from the market is not the way to do it. Neither is it a good idea to prevent the market from producing much needed price information.

Circuit breakers are desirable if closing the market enables it to be reopened, and if the reopening procedure is a good price discovery mechanism (i.e., it is a well structured call auction). Order batching, in and of itself, produces a price that better reflects the broad market's desire to hold shares of a stock. Further, because of its affect on order placement, the call auction procedure has a second benefit of paramount importance. To see this, let us contrast the thought processes of participants in continuous markets and in call auctions.

As investors, participants consider how prices might change over a relatively long span of time (e.g., weeks, months, or years). As traders, however, participants take account of how prices might change over a relatively brief span of time (e.g., a trading day, an hour, a minute or less). In so doing, they make strategic order placement decisions. If a participant expects a stock's price to fall in the short run, he or she will place a buy order for the stock at a depressed price when seeking to acquire shares, or will rush an aggressively priced sell order to the market when seeking to lay off shares. Similarly, if a participant expects the stock's price to rise in the short run, he or she will rush an aggressively priced buy order to the market when seeking to acquire shares or will place a sell order at an elevated price when seeking to lay off shares. These order placement tactics accentuate price movements in a market that is under stress. Additionally, the very reality of participants making these tactical order placement decisions is itself prima facie evidence of pricing inefficiency.

In a continuous market, participants know that, with few exceptions, if a limit order sitting on the book executes, it does so at the price at which it has been placed. In a call auction, on the other hand, all orders execute at a common clearing price, not at their own limit prices (unless the two values happen to coincide). This change in the trade execution procedure has a critically important effect - it encourages participants to reveal the prices they would be willing to trade at if a market crash does not occur. Consequently, a crash may be avoided.39

Other Public Policy Issues

Stock markets around the world are predominantly continuous trading environments (even though most open with a call) and the continuous trading modality is in the mindset of most people when discussing structural, competitive, and regulatory issues concerning the markets. Electronic call auctions are not well understood or even widely known in the United States, and they have not as yet been incorporated in our major market centers (although the NYSE does use a non-electronic call procedure to open its market). Perhaps the call auction will deliver some new answers to a spectrum of old questions.

We have previously considered the call auction in relation to price discovery, price volatility and order flow consolidation. Other market structure issues that have been widely discussed over the past two decades have included the introduction of electronic trading, extended trading hours, the competitiveness of our markets in an increasingly global environment, and the capital raising efficiency of our markets. What does the call offer with regard to these issues? It makes excellent use of electronic technology. The call could easily be activated off-hours if there is demand. Calls would strengthen our markets' global competitiveness, and facilitate capital raising by providing enhanced secondary market liquidity.

There also is the 1975 congressional mandate to develop a National Market System so as to achieve four broad goals:

The four goals boil down to two: (1) the integration of customer orders in the marketplace and (2) the consolidation of market information. The integration of orders produces the quotes and prices, and this is the information that is to be consolidated and broadly distributed. If public buy and sell orders do not meet in an orderly fashion, however, noisy information results, short-run price volatility is accentuated, and the NMS goals are not met. Accordingly, for both of our major equity markets, the orderly integration of buy and sell orders in a call market environment should be seen as furthering the attainment of the four NMS goals. Unfortunately, too much attention has been given by the regulatory authorities to the consolidation and dissemination of information, and too little to the production of information. The reality is that poor information is produced if orders are not properly integrated in trading. With respect to this, for investors individually and for the market in aggregate, a call auction would be very desirable.

Still, one might question whether or not investors would hold back their orders and wait for the periodic call auctions if both call and continuous modalities were available to them. We turn to the issue of investor trading patience in the next section of the paper.

Investor Demand for Immediacy

An important way in which institutional investors can control trading costs is by trading patiently. Yet, there is a belief that investors, when they decide to trade, want to trade as quickly as possible (i.e., that they demand immediacy). Thus, an often-cited disadvantage of call auction trading is that it denies participants continuous access to the market. This is not a problem, of course, if call auction trading is combined with continuous trading in a hybrid system. Moreover, institutional participants may find that the call environment actually provides them with greater immediacy than the continuous market. This is because a call is a point-in-time meeting place that enables participants to find each other more easily, and thus to fill their large orders faster and at lower cost.

Whether or not institutional participants actually require immediate execution of their large orders has rarely been questioned, and relatively little has been known in the past about institutional investors' demand for immediacy. Because of this lack of information, Nicholas Economides and I conducted a survey of institutional trading practices.40 The responses received from 150 equity traders at funds with over $1.5 trillion of equity under management indicate that institutions commonly work their orders patiently to control their execution costs. The survey revealed that: (1) portfolio managers typically give their traders more than a day to work large orders, and (2) more than one day is indeed commonly required to execute fully a large order that has been broken up and fed to the market in smaller pieces. These findings were confirmed by three other surveys of institutional investors in Europe, in France only, and Australia.41

We conclude the following. While distinctions are commonly drawn between informed traders and liquidity traders, and between institutional investors and retail customers, the critical difference may actually be between participants who wish to trade with immediacy and those who are willing to trade patiently. Certainly, one of the most important options to give users is the choice between trading in the continuous market and paying the price of immediacy, and trading in a call auction and avoiding this cost. Integrating call and continuous markets would give investors the choice. Effectively, the hybrid market would unbundle "trading" from the "immediacy" of a trade.

How will investors respond? The survey responses, as well as cumulating evidence from France and Germany, suggest that the two alternative trading modalities would work well together in the U.S. Call auctions would receive substantial order flow, while an appreciable proportion of orders would continue to be directed to the continuous market as well.42 And so, with a properly structured call auction, a win-win situation could be achieved for both buy-side investor and sell-side broker-dealer participants.43

Why Has the Call Been Resisted?

If the call is such an excellent facility, why is it that electronic call auction trading has not been more widely used in the United States? Two reasons are paramount. First, sell-side participants resist the innovation, fearing that it would dis-intermediate the market and, in so doing, hurt them economically. The second reason is ignorance.

The concern of the sell-side is understandable and hardly surprising. Broker-dealers, both on and off the floor, provide an array of services (information provision, order handling, transactional immediacy, account management, etc.) that help bring the customer to the market. However, because their compensation for these services is trading related (i.e., it comes from commissions and the bid-ask spread), technology that enables electronic trade execution and dealer dis-intermediation directly impacts broker-dealer revenues. To resolve the problem, a minimum fee could be established for all orders entered into the call with broker-dealer firms retaining part of this fee as their compensation for having brought the customers to the market, although such an arrangement may conflict with the 1975 ban on fixed commissions.44 In any case, sell-side participants should recognize that any improvement in market structure will spur further growth of the aggregate order flow and, consequently, that the pie which everyone shares will be bigger.

Ignorance is a formidable problem. On the surface, the call auction may seem to be a simple alternative to continuous trading. Upon first hearing about it more than two decades ago, I viewed the call through the eyes of a micro-economist. I visualized orders being cumulated to form downward sloping buy curves and upward sloping sell curves that look like the demand and supply curves we teach our students in economics 101 (as does Appendix Figure A1d). The simple elegance of the price determination procedure was captivating, and I thought the call was an innovation that could not be resisted. I have subsequently learned that far more is involved. The very subtlety of the procedure and the complexity of the array of issues it has implications for have impeded its acceptance. Among the issues involved are:

Technology: While information technology (IT) can be used advantageously in continuous trading, it is essential for call auction trading. Moreover, the call auction is an extremely good environment for the application of IT.45 In a continuous market, IT speeds up the rate at which orders can be submitted, displayed, and turned into trades; in so doing, it accentuates the importance of nano-seconds. In a call auction environment, on the other hand, IT is used to sort and cumulate orders and to find the clearing prices. It is apparent from systems such as AZX and OptiMark that, in a call auction environment, the computer is used to do one thing in particular that it was created to do - namely, to compute.

Call Auction Structure: Considerable choice exists in call auction design, and a poorly designed system will fail.46 Further, if a market center runs a call auction along with its continuous market, the two trading modalities must be properly interfaced.

Achieving Critical Mass: Much as a car cannot run without gas, a securities market cannot operate without sufficient order flow. Even a well-designed call auction may not succeed in initially capturing critical mass order flow and so may fail.

Dealer Compensation: As discussed above, broker-dealers must be appropriately compensated when operating in a hybrid environment that includes a dis-intermediated trading modality. In the short run, if the intermediaries do not accept that an innovation will improve market structure and benefit them, the innovation will be blocked. In the long run, the continued participation of the intermediaries in bringing customers to the market must be assured.

Regulatory Issues: In considering regulatory issues concerning market structure, we have tended to focus on the spatial consolidation of orders, not on temporal consolidation. We have also tended to focus on the size of the bid-ask spread and to ignore the accuracy of price discovery. Realigning the regulatory focus toward temporal consolidation and price discovery would facilitate the resolution of various regulatory issues, and could lead to market structure developing naturally rather than by regulatory edict.

Buy-Side Traders' Demand for Immediacy: The conventional view persists that buy-side traders demand immediacy. However, many institutional investors are more concerned about anonymity and keeping trading costs low than about obtaining immediate executions, per se. It is also widely believed that continuous markets provide immediacy. Nevertheless, large, institutional sized orders cannot be executed immediately at reasonable cost in our continuous markets. Ironically, for some customers, periodic call auctions could provide more immediacy than the continuous market.

Order Specification: Limit and market orders have different properties in call auctions than in continuous markets. In a continuous market, limit orders placed on the book generally execute at their limit prices. In a call auction, limit orders execute at the common clearing prices, not at the prices at which they are written. In continuous trading, market orders are generally executed immediately at the counterpart market quote. In a call auction, market orders are held, along with limit orders, until the next call. At the next call, depending on the structure of the auction, market orders may be matched against contra-side market orders or treated as aggressively priced limit orders. Either way, the distinction between market and limit orders is minimal in call auction trading.47 As a consequence of limit orders and market orders having different properties in a call auction, participants should write their orders differently in this trading environment. Assuming no market impact, a buyer in a call auction should state the highest price he or she would be willing to pay rather than not trade at all. Similarly, a seller in a call auction should state the lowest price he or she would be willing to receive rather than not trade at all.48

Information Release: Information events are costly to limit order traders. Consequently, the prospect of news occurring discourages the placement of limit orders, thereby making markets less liquid. The release of some kinds of information is controllable (e.g., announcements of corporate earnings or unemployment statistics) while the release of other kinds is not (e.g., earthquakes or ice storms). With continuous trading, controllable information releases are commonly (but by no means always) made outside of regular trading hours. With call auction trading, controllable information releases that are timed in relation to the calls could be more readily made during the trading day (the more complex the information, the earlier it should be released before the start of a call). We expect that having predetermined times at which the market will be called would facilitate the intra-day release of information without discouraging the placement of the limit orders that make a market more liquid.49

Extended Hours Trading: Prior to 1999, Instinet, a few other ATSs and the NYSE were offering after hours markets to institutional customers. In 1999, competitive pressures strengthened for the Big Board, Nasdaq, and others to offer extended hours trading for retail customers. Yet it has been expected that the after hours market will remain relatively illiquid and volatile. Further, after hours trading blurs the meaning of a market's closing price, which can disrupt the various legal uses to which a closing price is put: the assessment of margin requirements in the cash market, marking-to-market in the derivatives market, mutual fund purchases and redemptions, and so forth. Call auction trading would certainly provide useful price discovery and liquidity provision to the after hours market. Further, call auction prices could be used for the various legal purposes just noted.

Over time, resistance to change and ignorance break down. Increasingly, it is becoming apparent that interfacing periodic call auctions and continuous trading modalities is a step in market structure development that must be taken. It would certainly be desirable for our major equity markets to follow this approach to developing their hybrid structures. As with plywood, having the grain go in different directions would indeed improve the performance of the NASDAQ Stock Market. It would strengthen the Big Board as well.

Note: Figure 1 and Table 1 available upon request. Please contact Antoinette_Colaninno@baruch.cuny.edu.


Appendix

Order Batching in a Call Auction

The principle of order batching in a call auction is illustrated with reference to Figure A1. In each of the four diagrams of the Figure, we show share price on the vertical axis and, by simply letting all orders be for the same number of shares (e.g., one round lot), we show the number of orders at each price on the horizontal axis.

The first diagram (Figure A1a) displays a set of buy and sell orders, with the orders arrayed at each price according to the sequence in which they arrived at the market. The next three diagrams show how the orders are batched together to determine which execute, and the price at which they execute. Only buy orders are shown in Figure A1b. At 51, the highest price at which any buy order has been placed, there is one order to buy. Two additional buy orders have been entered at 50 and thus, at 50, we have a cumulative total of three orders to buy. Looking at yet lower prices, one order has been placed at each of the prices, 49, 48, and 47. Thus, the cumulative number of orders at these prices is four, five, and six, respectively.

Only the sell orders are shown in Figure A1c, and we can cumulate them much as we did the buys. The sell orders, however, are cumulated from the lowest price (48) up to the highest price (52).50 Because there is one sell order at each of the prices in this range, the cumulative number of sell orders increases by one order as we move from the single order at 48 to the five orders at 52.

The cumulative buy and sell orders are brought together in Figure A1d. The two curves intersect at the point were price is 50 and the number of orders is three. Thus three buy orders (one placed at 51 and the two at 50) are matched with three sell orders (one placed at 48, one at 49, and one at 50). At 50, the maximum number of shares trade (there is only one buy order at the higher price of 51, and only two sell orders at the lower price of 49).

Note that the most aggressively priced buy orders are matched with the most aggressively priced sell orders, and that three of the executed orders receive price improvement (the buy at 51, the sell at 49, and the sell at 48). The unexecuted orders are the buys at 49, 48, and 46, and the sells at 51 and 52. These less aggressively priced orders may be cancelled, rolled into the continuous market, or held for the next call, depending on the wishes of the trader.

Note: Figure A1 available upon request. Please contact Antoinette_Colaninno@baruch.cuny.edu.


  
Footnotes
1 I thank Paul Arlman, Antoinette Colaninno, Miriam Humbach, Kris Monaco, Deniz Ozenbas, Michael Pagano, and Steve Wunsch for the much appreciated contributions they have made to this chapter.
2 Forrester Research Inc. estimates that Americans held less than 4 million online brokerage accounts at the beginning of 1999, with a total value of $106 billion. The report predicts that by 2003, there will be 20.3 million U.S. online accounts worth $3 trillion. See BusinessWeek Online, e.biz, Data Mine, May 24, 1999.
3 At the time of this writing, there are active trading floors in Germany and Amsterdam, although they are receiving a diminishing portion of the order flow.
4 My own advocacy of call market trading is expressed in this chapter. However, the views are my own and do not necessarily represent those of the other contributors to this book.
5 The principle of order batching and price determination in a call auction is described more fully in the appendix to this paper.
6 This is particularly apparent with regard to OptiMark Technologies and State Street's Bond Connect.
7 The generics are typically not employed in their pure forms, however. For instance, agency/auction continuous markets commonly open with a call. The NYSE's order driven system includes intermediaries (the NYSE specialist, other floor traders, and upstairs market makers). A dealer can be included in a call auction (dealers, known as designated sponsors, are included in both the call and continuous trading modalities in Deutsche Börse's Xetra). And so forth.
8 The dynamics of the price process must, of course, compensate public traders for placing limit orders or else that market structure would not be viable. Research that I have undertaken with Puneet Handa suggests that limit order traders in a continuous market are compensated by a stock's transaction price mean reverting after the price impact of a liquidity event has caused a limit order to execute (with the mean reverting process implying accentuated volatility in the short run). The intuition behind the Handa-Schwartz model is the following. A limit order obtains an undesirable execution if the arriving market order that caused it to transact was motivated by informational change. However, liquidity events also trigger price changes and, if sufficient relative to the price changes caused by informational change, the liquidity driven changes can cause limit order trading to be profitable. This is because price reverts back to its previous level following an execution that was caused by an imbalance between liquidity motivated buy and sell orders. In the process, price volatility in short trading intervals is elevated. For further discussion, see Chapter 21 (Handa and Schwartz) of this book and P. Handa, R. Schwartz, and A. Tiwari, "The Ecology of an Order Driven Market," Journal of Portfolio Management, Winter 1998, pp. 47-55.
9 See S. Rep. No. 94-75, 94th Cong., 1st Sess. 7 (1975) ("Senate Report"). For further discussion, see SEC Market Data Concept Release, Release No. 34-42208, December 9, 1999.
10 See, for instance, Statement of the Securities and Exchange Commission on the Future Structure of the Securities Markets (February 2, 1972), 37 FR 5286. Also See SEC Market Data Concept Release, Op. Sit.
11 SEC Market Data Concept Release, Op. Sit.
12 See, for instance, Chapter 20 (Wunsch) of this book and R. Schwartz and A. Colaninno eds., Who's Controlling our Equity Markets: The Industry or the Regulators?, manuscript in process, 2000.
13 W. Christie and P. Schultz, "Why do Nasdaq Market Makers Avoid Odd-eighth Quotes?" Journal of Finance 49, 1994, pp. 1813-1840.
14 Opinion by the Honorable Robert W. Sweet, 94Civ.3996, U.S.D.C., Southern District of New York.
15 Preferencing refers to the practice of a customer directing an order to a particular dealer even if that dealer is not posting the most aggressive quote on the market. Quote matching refers to that dealer's practice of matching the best bid or offer on the market when receiving a preferenced order. Dealers receive preferenced orders because of their on going relationships with other brokers and public customers. They typically accept these orders and quote match because, to do otherwise, would lower the likelihood of receiving further orders in the future.
16 As noted above, the new order handling rules introduced by the SEC in 1997 have required that public limit orders be represented in the Nasdaq quotes.
17 To see this, picture continuous trading in the dealer market as being represented by a horizontal time line. Periodic trading in the call auction would then be represented by vertical hash marks at fixed intervals along the horizontal time line.
18 This section of the chapter is a modified version of my piece, "How the NYSE Can Save Itself," Global Investment, December 1999, Volume 6, Number 1.
19 Although the ECNs are less differentiated from Europe's electronic bourses, their threat of entering the European markets is being taken seriously by the European Exchanges. If the ECNs were to start trading the exchange listed stocks of Europe, problems of free-riding on exchange provided services such as regulation would arise.
20 A computationally sophisticated system such as OptiMark could not have been run on the computers available in the early 1990s.
21 A market maker can use SelectNet to send an order it has received to another market maker or to broadcast the order to all market makers. As quote providers, an ECN can also connect directly into SelectNet. SelectNet includes a negotiation feature that allows a participant (market maker or ECN) to accept, reject, or counter a received order.
22 For further discussion, see Chapter 8 (Samaran, S. Thomas, and Demarchi), Chapter 9 (Reck), and Chapter 25 (Schmidt Oesterhelweg, and Treske) of this book. The electronic call auction has also been used by the Tel Aviv Stock Exchange; see Chapter 10 (Bronfeld) of this book.
23 Peake, Mendelson, and Williams's proposed electronic system, which has been the prototype for most electronic continuous markets including Toronto's CATS, Paris's CAC, and Tokyo's CORES, incorporates a call as its opening procedure. See J. Peake, M. Mendelson, and R. T. Williams Jr., "The National Book System: An Electronically Assisted Auction Market," Proceedings of the National Market Advisory Board of the Securities and Exchange Commission, April 30, 1976.
24 See Chapter 4 (Cohen and Schwartz) and Chapter 5 (Economides and Schwartz) of this book.
25 The acronym stands for "Cotation Assistée en Continu."
26 See Chapter 5 (Economides and Schwartz) of this book.
27 For further discussion, see Chapter 4 (Cohen and Schwartz) of this book.
28 OptiMark is a sealed bid/ask auction. For further discussion, see Chapter 22 (Clemons and Weber) of this book.
29 For further discussion of the properties of call auction trading, see Chapter 4 (Cohen and Schwartz) and Chapter 5 (Economides and Schwartz) of this book.
30 The call auction is also a good procedure for closing these markets.
31 Parts of this and the following section have been modified from "Order Flow Consolidation with Multiple Trading Modalities," a paper I presented at the Deutsche Börse AG Symposium, Equity Market Structure for Large-and Mid-Cap Stocks, Frankfurt, December 1997.
32 For further discussion, see P. Schreiber and R. Schwartz, "Price Discovery in Securities Markets," Journal of Portfolio Management, Summer 1986, pp. 43 - 48.
33 Over the years I have used many different price charts to illustrate the point just made that intra-day volatility is attributable in part to price discovery being a dynamic process in continuous trading. It has not been difficult to find price charts that suggest this story.
34 Network externalities exist when an environment can be viewed as a network, and when the value of the network is more valuable to individuals, the greater the number of individuals in it. For instance, the broad ownership of fax machines can be viewed as a network, and each individual fax machine is more useful to its owner, the larger the number of other people who also have a fax machine.
35 A market center will be similarly concerned. The issues involved can justify a market center such as the New York Stock Exchange having imposed its order handling Rule 390, which requires that a member firm that has received a public order for a listed stock bring that order to an exchange for execution. For further discussion, see E. Bloch and R. Schwartz, "The Great Debate over NYSE Rule 390," Journal of Portfolio Management, Fall 1978, pp. 5 - 8. In December 1999, the NYSE board voted to remove Rule 390.
36 A call may also be structured to identify multiple clearing prices at an auction, as does OptiMark. The system allows customers to submit orders that represent more complete expressions of the satisfactions they would receive from alternative possible executions (the complex orders are referred to as satisfaction profiles). See Chapter 22 (Clemons and Weber) of this book for a description of OptiMark.
37 In Germany, the procedure is referred to as a volatility interruption.
38 Circuit breakers were put into effect in the U.S. markets following the crash on October 19, 1987, to keep the markets from going into free fall.
39 Chapter 6 (Beiner and Schwartz) of this book considers in more detail how the properties of limit orders placed in a call auction differ from the properties of limit orders placed in a continuous market.
40 See Chapter 5 (Economides and Schwartz) of this book.
41 See Chapter 13 (Schwartz and Steil), Chapter 14 (Demarchi and S. Thomas), and Chapter 15 (Douglas and C. Thomas) of this book.
42 See Chapter 13 (Schwartz and Steil), Chapter 14 (Demarchi and S. Thomas), and Chapter 15 (Douglas and C. Thomas) of this book.
43 See Chapter 4 (Cohen and Schwartz) and Chapter 7 (McCormick) of this book for further discussion of the design details of a call auction.
44 See Chapter 28 (Schwartz) of this book.
45 There is a tendency to use new technology to do better and faster what has been done in the past with an older technology (e.g., to mimic the procedures of continuous market, floor based systems), and non-electronic calls have been very inefficient. However, the successful introduction of electronic trading requires the design of new order handling procedures, and the call auction in an electronic environment becomes an excellent vehicle.
46 See Chapter 4 (Cohen and Schwartz), Chapter 7 (McCormick), Chapter 22 (Clemons and Weber), and Chapter 24 (Domowitz and Madhaven) of this book for further discussion.
47 See Chapter 6 (Beiner and Schwartz) of this book.
48 In the terminology of microeconomics, participants without market power should submit their reservation prices. Note that participants in a continuous market should never reveal their reservation prices because doing so would result in the total elimination of the gains to trading.
49 This thought is expressed in Chapter 6 (Beiner and Schwartz) of this book.
50 Buy orders are cumulated from the highest price to the lowest because the price limit on a buy is the highest price the trader is willing to pay (a lower price would gladly be accepted). Sell orders are cumulated from the lowest to the highest because the price limit on a sell is the lowest price the trader is willing to receive (a higher price would gladly be accepted).