Floor vs. Automated Trading Systems
By Lawrence HarrisProfessor of Finance & Business Economics

Traders say automated trading systems are more fair, more efficient and cheaper to operate than floor-based systems. Then why is the NYSE building more trading floors?

Introduction
This summer, the Bangladeshi Stock Exchange replaced its trading floor with a new fully automatic computerized trading system. At the same time, the New York Stock Exchange considered locations to build a new trading floor.

The continued commitment of the New York Stock Exchange to its trading floor may be its most important decision of the 20th century. The men and women who are making this decision are fully aware of its significance. They believe that the tremendous success of the NYSE is due in large part to its floor-based market structure. They also know that they may lose much, if not all, of their franchise if they are wrong.

The New York Stock Exchange is the world's most liquid stock market. If the liquidity comes from its floor-based market structure, eliminating the floor would be foolish. However, if the liquidity is due to some other factor, a switch to electronic trading may be feasible and perhaps even desirable.

Are automated trading systems better than floor-based trading systems? Traders widely perceive them to be more fair, more efficient and cheaper to operate than floor-based systems. Why then would the New York Stock Exchange consider building a new trading floor?

This article examines the arguments for and against these two trading systems considering aspects of fairness, convenience, capacity, exchange of information and cost.

Fairness

Operational Fairness
In operationally fair markets, trading rules are uniformly applied and no cheating occurs. Traders widely believe that fully automated trading systems are the fairest of all market structures. Automated systems do only what they are programmed to do. They implement their trading rules exactly and without exception. They expose orders only as instructed and only to those traders to whom orders may be exposed.

In contrast, fairness in oral auctions depends on the skill and honesty of the traders who arrange the trades. Traders must follow trading rules honestly even when doing so may cause them to lose an advantage. Although most oral auctions are quite fair, all such markets have suffered from well-documented trading scandals involving front-running, inappropriate order exposure, fraudulent trade assignment or prearranged trading.

Front running occurs when a broker improperly allows one order to trade ahead of another order; the order that goes first usually profits from the price impact of the following order. Inappropriate order exposure occurs when a broker shows an order to another trader for the other trader's benefit rather than for his client's benefit; the other trader will typically act on the information, either by trading ahead of the order or by refusing to trade with it. Fraudulent trade assignment can occur when a broker executes orders on the same side of the market for more than one client; a dishonest broker may assign the best prices to his favored clients. Prearranged trading occurs when a broker arranges a trade without properly exposing her client's order to other traders who might be willing to offer better prices; under such circumstances, the client often receives a worse price than he might have received if the broker properly exposed the order.

Fair Access
Markets have fair access when all traders have equal access to the market. In such markets, no traders have special advantages over other traders. Few trading systems provide pure fair access.

In floor-based trading systems, floor traders have an advantage over off-floor traders. Floor traders can see and react to market developments well before off-floor traders can.

Off-floor traders must obtain their information through market data systems, and they must respond through an order routing system. Floor traders also can observe all market information revealed on an exchange floor and not just what the market data systems report. In particular, they observe who is trading, which can be valuable if you can guess why they want to trade or who they represent. In some floor-based markets, large and loud traders have some advantage over others because they can control the spots within the pit that offer the best sight lines.

Automated markets favor traders with good keyboard skills and abstract visualization skills. Although most people would not consider this unfair, it does cause anxiety among some traders when they first start using electronic systems. Automated markets also favor traders who use automated systems to generate their orders. Such systems can monitor electronic data feeds and respond instantly to new information. Although this advantage is a natural consequence of faster trading technologies, many manual traders resent competing with such automated traders.

Convenience
A primary advantage of automated trading systems is that they allow traders to trade from desks in their offices rather than on an exchange floor. This convenience allows traders to sit next to their telephones and to consult any data systems they want in support of their trading.

Such facilities are often difficult or impossible to arrange on the floor of an exchange. Floor traders who want instant access to telephone and data services must carry cell phones and portable data terminals with them onto the trading floor. These instruments are often cumbersome, and some markets do not permit them. The NYSE recently introduced an innovative mobile electronic communications system for its floor brokers.

System Capacity
More traders can directly participate in an automated auction than in an oral auction. The number of traders who can effectively communicate with each other at the same time limits the size of an oral auction. When too many traders try to participate in the same auction, they exceed its capacity to process information in an orderly fashion. As the number of traders bidding and offering is large, traders cannot easily keep track of who is quoting the best prices. Traders may then mistakenly arrange trades that violate time precedence or even price priority.

Great numbers of traders can simultaneously interact with each other in automated trading systems because these systems process order messages much more quickly than people can. Traders who use automated systems do not have to keep track of the best bid and offer. Instead, they let the system do it for them. They also do not have to arrange their trades. The system does it for them according to the market's trading rules. By supporting these functions, automated trading systems allow traders to focus their attention exclusively on the creation and submission of their orders.

Negotiation Speed
Some floor traders believe they can trade more quickly on a floor than in an electronic market. They claim that they can more quickly shout a bid or offer, or accept a bid or offer than they can enter this information into a computer. Although this probably is true, modern trading systems with graphical interfaces allow traders to enter information almost as quickly as traders can shout it out.

In any event, automated trading systems can arrange trades much more quickly than can individual traders. In an oral auction, traders must manually record the price, the size, the counterparty and the item traded for each of their trades. In many markets, they must also record the time of the trade. Since trading requires both order entry and trade record keeping, traders can complete trades more quickly in electronic markets than in oral markets.

Oral trading, however, is undoubtedly faster than screen-based trading when traders want to negotiate their trade sizes. In such negotiations, traders often will not reveal the full size of their order unless they are sure that the other trader will trade the same size. Accordingly, after agreeing on a price, they often take turns proposing successively higher sizes. Negotiations stop when one trader proposes a greater size than the other trader will accept. The trade size is then the last agreed-upon size. This back and forth negotiation moves very quickly in oral auctions. When conducted through a computer that only accepts firm orders, the negotiation is much slower, since the traders must split their orders into pieces to avoid displaying their full sizes.

Some electronic trading systems, such as Reuters' Instinet, solve this order display problem by providing messaging systems that allow a buyer and seller to negotiate their trade size through the exchange of messages on their screens. Others allow large traders to place hidden limit orders to restrict the displayed size of orders.

Exchange of Information
Floor-based trading systems dominate electronic trading systems when brokers need to exchange information about buyers and sellers to arrange their trades. Most electronic trading systems do not provide this information.

Such information is especially important for large traders who want to avoid trading with well-informed traders. When a well-informed trader wants to sell, prices are probably too high. It is a poor time to be a buyer. It is always better to not trade than to trade at a poor price. Therefore, large traders often direct their brokers to trade only with institutions they deem to be uninformed. They prefer to trade with institutions that have no research staffs and that trade only to invest and disinvest rather than to speculate. They often refuse to trade with the proprietary trading groups of investment banks for fear of losing to them. Large traders can issue these instructions to their brokers because they can afford the significant commissions necessary to obtain the personal attention necessary to provide these services.

Traders also do not want to trade with a large trader if that trader intends to continue trading on the same side afterwards. The market impact of the large trader's subsequent trades will generate immediate losses for the first traders. Traders therefore instruct their brokers to ask how much more size the other side wants to do. On an exchange floor, a broker who has no additional size to fill will freely offer this information to obtain a better price. In an electronic trading system, no similar mechanism allows traders to credibly indicate to each other how much additional size remains in their orders.

Market Data Reporting
Automated systems report market data to the public much faster and much more accurately than do floor-based trading systems. In floor-based systems, traders or market reporters must manually enter quotes and trades into the market information system to report them to the public. In automated systems, the data is already in electronic form.

Cost
Fully automated trading systems have high initial creation costs. They require extensive data networks and data processing systems that validate users, accept orders, process orders, report trades and report order status. These functions must all be reliable, secure and fast. Several vendors sell fully automated trading systems off-the-shelf for about $5 million for purchase and set up. Once established, electronic trading systems have small operating costs because everything is automated.

Floor-based trading systems cost less to set up than electronic trading systems. To set up a floor-based trading system, an exchange needs to acquire and/or build a suitable trading floor. A floor-based market must also have adequate telecommunications systems to route orders in and market information out as well as various information display systems to assist their floor-based trade.

However, floor-based trading systems are often quite expensive to operate. They require substantial labor to run. Well-trained brokers must communicate with clients, and arrange trades. Reporters must report the trades. Officials must watch for trading abuses. And runners often must carry messages to and from brokers. The brokers must know and follow the trading rules and procedures, and they also must know how to best expose their orders and negotiate their trades.

In contrast, electronic trading systems do not require well-trained brokers to operate. In many such systems, brokerage clients can access the market themselves without the involvement of their brokers.

Conclusion and Predictions
Floor-based trading systems and automated trading systems have different strengths and weaknesses. Consequently, they appeal to different clienteles. It is unlikely that one market structure will dominate all trading.

Fully automated systems are fast and generally cheap to use and operate. These characteristics ensure that active markets and markets that serve small traders will use automated trading systems extensively. In the U.S. equities markets, Bernard L. Madoff Investment Securities and other dealers that offer automated execution systems provide excellent service to high volumes of small traders. Options markets tend to have high order volumes and small transaction sizes. These markets therefore will automate their trading.

Fully automated systems also allow traders to exercise direct control over their orders. They therefore appeal to traders who do not trust their brokers or to traders who do not want to pay for brokerage. Large institutions that are concerned about how brokers expose their orders often favor automated systems if they are willing to employ their own traders.

Floor systems work best when traders need to exchange information about each other before they trade. They also work best when brokers need to actively search for traders to fill their orders. Since these advantages are most important to large traders, floor-based markets will serve primarily large institutional traders. Increasingly, the NYSE is an institutional market. Although people have been predicting the demise of the NYSE floor since the mid-1960s, it will not disappear as long as the brokers there provide services that cannot be provided elsewhere.

About the Author
Lawrence Harris serves as professor of finance and business economics. His research centers on security market structure, economic modeling and forecasting. He was the first visiting scholar at the NYSE from 1989 to 1990. During the 1988-89 academic year, Harris served as Economic Fellow in the Office of the Chief Economist of the SEC in Washington, D.C. He earned an MA and PhD from the University of Chicago.