Market Microstructure and Exchanges
Chapter for World Federation of Exchanges
50th Anniversary Commemorative Book
Larry Harris
USC Marshall School of Business
Draft: December 25, 2009
1. Introduction
During the last 50 years, academic research into trading and exchanges has advanced substantially. Researchers have carefully considered the determinants of transaction costs; the role of information in the markets; how trading rules affect liquidity, price efficiency, and volatility; the competition among exchanges, brokers, and dealers; and how best to regulate the markets. This chapter provides a brief overview of insights into trading processes that have come out of the academic literature.
Many of the perspectives presented in this chapter are familiar to practitioners who often are not consciously aware of them. Indeed, much of what academics know about the markets has come from practitioners who generously shared their practical knowledge with academics. Academics took this knowledge and organized it to help people recognize and better understand the issues important to them. The resulting body of knowledge provides a “big picture” that helps practitioners make better trading decisions and helps exchanges and regulators better understand how to organize and oversee trading.
The discussions in this chapter present the essentials of this knowledge. To ensure that the presentation reads easily, the text discusses only the principles and not also the history of each contribution. Readers who wish to learn more should consider reading my introductory text, Trading and Exchanges: Market Microstructure for Practitioners. For excellent surveys of theoretical and empirical studies in market microstructure, readers also may consider consulting Maureen O’Hara’ book, Market Microstructure Theory and Joel Hasbrouck’s book, Empirical Market Microstructure: The Institutions, Economics, and Econometrics of Securities Trading.
Practitioners and academics are very interested in various dimensions of market quality, the most important of which are liquidity, price efficiency, and volatility. Liquidity refers to the cost of trading. Markets are liquid when traders can easily arrange a trade without affecting prices much. Price efficiency refers to the extent to which prices reflect information about fundamental security values. Prices are efficient when traders cannot use publicly available information to predict future price changes. Volatility refers to the rates at which prices change. Prices are volatile when fundamental values change quickly, when traders trade foolishly, or when exchange mechanisms do not work well.
Market quality depends on primarily on traders. Traders make markets liquid when they allow other traders to trade. They make prices efficient when they trade securities based on analyses of information that they conduct. They may make prices excessively volatile if their demands to trade become excessive. Accordingly, our discussion starts with an identification of the major reasons why traders trade.
Market structure—how exchanges, brokers, and dealers arrange trades—helps determine market quality by shaping the opportunities and incentives that different types of traders face. The second half of this chapter describes how different market structures affect market quality.
2. Traders
Economists identify three broad groups of traders when describing the origins of various market quality dimensions: Utilitarian traders, informed traders, and dealers. Utilitarian traders trade because they expect to obtain some benefit besides trading profits from their trades. Informed traders trade to profit from accurate predictions of future prices. Dealers profit by selling liquidity to other traders. Although people trade for many different reasons—and often for more than one reason at a time, these three types of traders represent most trading strategies.
2.1 Utilitarian traders
Utilitarian traders include many different types of traders whose trading generally is unrelated to fundamental security values. The best known utilitarian traders are investors and borrowers who trade to move money from the present to the future or vice versa. Other utilitarian traders include hedgers who trade to offset or insure against risks that they scare them; gamblers who trade for entertainment; tax avoiders who arrange trades to lower or defer their taxes; and asset exchangers who exchange one asset for another that is of greater immediate value to them.
In each of these cases, the traders trade because they hope to obtain some benefit from trading besides trading profit. Investors and borrowers move money through time. Hedgers reduce their net risk exposure. Gamblers obtain entertainment. Tax avoiders lower the present value of their tax liabilities, and asset exchangers obtain assets of greater immediate value to them. In all cases, these traders would like to profit from their trading, but if they are trading purely for the reasons noted, they do not expect to obtain returns in excess of the normal expected returns associated with holding (or shorting) securities.
Economists often call these utilitarian traders noise traders because their trading is unrelated to fundamental values. Accordingly, if their trading pushes prices around, prices will be less informative. Statisticians say that such prices are noisy estimates of fundamental value.
2.2 Informed traders
Informed traders collect information that allows them to predict futures prices. They buy when they expect prices to be higher and sell otherwise. Although they often are wrong, successful informed trades are right more often than they are wrong and therefore profit from their efforts. Informed traders base their trades on analyses of data that they believe will help them predict future values. Their information may include data about fundamental security values, or it may include data about the trades that other traders will likely do in the future.
Informed traders employ three main trading strategies. They may base their trades on estimates of fundamental values, on news about changes in fundamental values, or predictions about what other traders will do.
Value-motivated traders estimate fundamental security values. They then buy if their value estimates are sufficiently greater than market prices, and they sell if their value estimates are sufficiently lower. They estimate fundamental values by collecting and analyzing all information that their research budgets allow them to obtain. Since data collection and analysis are expensive, value-motivated traders can only profit if their trading gains are greater than their research costs.
News traders trade on news about events that change fundamental values. They buy when they believe that values will rise in response to an event, and they sell when they believe that values will fall. Since the effect on values of many events are quite obvious, news traders must trade very quickly to profit from new information. They therefore invest in systems that allow them to obtain and act upon information quickly. Since these systems can be expensive to build and operate, news traders can only profit if their trading gains are greater than their information collection costs.
Order anticipators try to predict the trades that other traders will do. If they expect that other traders will buy substantial quantities of securities and thereby increase prices, they try to buy first to profit from the price increases. They likewise try to sell before other traders sell. Order anticipators generally increase the costs of liquidity for the traders before whom they trade. When order anticipators trade in front of orders that they know about, they are called front-runners. Fiduciary duty prevents brokers from front-running their clients’ orders or from knowingly allowing others to do so. In most markets such activities are illegal. However, informed traders often try to anticipate the orders that other traders intend to submit by using pattern analyses and psychological models. The other traders must be utilitarian traders because traders cannot predict the trades of value-motivated traders or news traders without the information upon which these other informed traders base their trades.
2.3 Dealers
Dealers are traders who supply liquidity—the ability to trade when you want to trade—to other traders. They generally post quotes or limit orders that give other traders options to trade. Dealers profit by selling to buyers at offer prices that are slightly higher than the bid prices at which they buy from sellers. If the average spreads that they obtain between the prices at which they sell and the prices at which they buy are sufficiently large, their dealing will be profitable.
Dealers risk that prices will fall after they have bought or that prices will rise after they sell. When such price changes occur, they may not be able to trade out of their positions at a profit. They therefore hope to sell immediately after they buy and vice versa.
Dealers generally do not know much about fundamental values. They do know that if they set quotes too low, they will receive many buy orders and few sell orders. Likewise, if they set their quotes too high, they will receive many sell orders but few buy orders. Since they want to sell after they buy and buy after they sell, they try very hard to find the prices at which buyers and sellers are willing to trade equal volumes. These prices are called market clearing prices. Market prices differ from fundamental values when traders do not know fundamental values well, as is generally the case.
2.4 Adverse selection
Dealers and other traders who supply liquidity lose on average to well-informed traders. When well informed traders expect prices to fall, they sell. The dealers who buy from them lose if prices fall before they can sell their positions. Dealers likewise tend to lose to well-informed traders when the informed traders expect prices to rise.
Economists call this problem the adverse selection problem. The problem results because informed traders do not trade at random. Rather, their trading is correlated with future price changes: They buy when they expect prices to be higher and sell when they expect prices to be lower. If they realize their expectations, whoever traded with them loses. Dealers obtain adverse results—trading losses—when trading with well informed traders because the well informed traders more often than not select the profitable side of the market when they trade.
Dealers must quote bid/ask spreads large enough to recover from utilitarian traders what they lose to informed traders. Utilitarian traders thus lose to informed traders (through the intermediation of dealers) even if they trade at random because they must pay wide spreads. Adverse selection thus makes liquidity expensive. Utilitarian traders can avoid these losses by trading infrequently. If few utilitarian traders trade, informed traders will be a substantial fraction of the total order flow and bid/ask spreads will be very large. The informed traders will not be able to trade much, which will decrease their profits and decrease the resources that they spend on research, which will make prices less informative.
In the extreme, if informed traders are a substantial fraction of the total order flow, the spreads that dealers must quote to survive may be to large to support a viable market and the market will fail to operate. Averse selection thus explains why small stocks for which information is not widely disseminated tend to trade in less liquid and less informationally efficient markets than do larger stocks. Many exchanges that have started special markets for emerging companies have discovered that these markets often do not operate well due to the adverse selection problem.
The adverse selection problem has pervasive effects throughout all aspects of trading. Consider some examples of how adverse selection affects trading strategies, exchange design, and the design of financial products:
· Dealers try to avoid trading with well informed traders if they can do so. Many dealers internalize retail order flows to avoid the losses that they would incur trading with well informed institutions at exchanges.
· Dealers rarely will offer to trade large size to anonymous traders. They want to know with whom they are trading so that they can avoid losing to well-informed traders.
· Since dealers who can avoid adverse selection offer more liquidity, exchanges are very keen on creating trading mechanisms that protect dealers from adverse selection. For example, trading halt rules generally halt trading when dealers would be most vulnerable to informed traders.
· Many block brokers permit their clients to specify with whom they are willing to trade so that they do not trade with traders such as hedge funds that are likely to be well informed.
· Exchanges and regulators generally ban trading on inside information. These bans protect dealers from losing to well-informed traders. With this protection, the exchanges hope that dealers will offer more liquidity to other traders.
· Exchanges and others have created many index products that help investors purchase exposure to market-wide risks through a single transaction rather than through multiple transactions in many individual securities markets. Liquidity generally is cheaper in index markets than in individual security markets because adverse selection is a smaller problem in the index markets than in securities markets. Generally, few if any people reliably have deep insights into market-wide valuations whereas some traders often are well informed about individual security values.
2.5 The zero-sum game
Trading is a zero-sum game with respect to trading profits. The profits that buyers make are profits foregone by sellers, and the losses that buyers incur are losses avoided by sellers.
If traders were only motivated by expected trading profits, no rational traders would trade. The least well-informed traders would recognize that they would lose on average when trading so they would not trade. The least well-informed traders among the remaining traders also would not trade because they too would recognize that they would lose on average. The logical end to this argument is that no traders will trade with each other if they are only motivated by expected trading profits. In which case, markets would have no liquidity and prices would not be informative.
Fortunately, trading is not a zero-sum game when utilitarian traders are included. Utilitarian traders trade because they obtain benefits besides expected trading profits when they trade. As noted above, these benefits include moving money through time, risk management, tax avoidance, and entertainment. The losses that utilitarian traders sustain when trading with informed traders are the costs they must bear to obtain the various services that they seek from using the markets. The equivalent gains that the informed traders make fund their efforts to acquire information, and their trading on that information makes prices more informative. Utilitarian trading thus indirectly funds price efficiency. Markets exist only when traders are interested in trading for reasons besides trading profits. Without utilitarian traders, markets fail.