Is the market microstructure of stock markets important?

• Author: Naes, Randi

The market microstructure literature studies how the actual transaction process--i.e. how buyers and sellers find one another and agree on a price--can affect price formation and trading volumes in a market. This article provides an introduction to the concepts, frameworks and most important themes in this literature. The market serves two functions: one is to provide liquidity for buyers and sellers; the other is to ensure that prices reflect relevant information about fundamental value. Microstructure models differ from traditional financial models by recognising that legitimate information about companies' fundamentals may be unequally distributed between, and differently interpreted by, market participants. We can therefore no longer assume that prices will reflect information immediately even if all participants are rational. The microstructure literature argues that both information risk due to asymmetric information and differences in liquidity over time and between companies impact on long-term equilibrium prices in the market.

1 Introduction

If participants in the stock market behave rationally and have the same information, share prices will at all times reflect all available information about companies' fundamental value. Since it was first advanced in the 1960s, this has been one of the most important hypotheses in financial economics. However, over the last 20 years, both the theoretical foundation for this hypothesis and the previously strong empirical support for it have been challenged.

The microstructure literature challenges the hypothesis of efficient markets by studying how prices can deviate from (or converge towards) informationally efficient equilibrium prices as a result of rational participants behaving strategically (Biais et al., 2004). (1) Strategic behaviour can be put down to unequal access to information (2) or to limited liquidity (3) in the secondary market. While the efficient market hypothesis abstracts from the actual process which leads to buyers and sellers finding one another and agreeing on a price, the microstructure literature focuses on the functions performed by the marketplace.

Themes in the microstructure literature divide naturally into three: (i) the actual transaction process, (ii) the effects of market structure and trading rules on the transaction process, and (iii) the transaction process's implications for fundamental economic decisions. This subdivision also largely reflects the chronological development of this research field.

Models of the transaction process are described in section 2 below. There are two main groups of model. The first (inventory models) studies how an intermediary (hereinafter referred to as dealers, see figure 1) can solve the problem of buyers and sellers not being present in the market simultaneously. The second (information models) analyses how information which is asymmetrically distributed between participants in the market is reflected in the prices of securities.

Research into the significance of market structure and trading rules is the subject of section 3 below. The importance of the organisation and design of the stock market came to the fore in the wake of the crash of 1987 and the revelation of collusion among the dealers on NASDAQ in 1994. There has since also emerged a considerable body of literature on the effects of market fragmentation and competition from new electronic trading systems.

Microstructure research rejects the hypothesis that the transaction process and the organisation of markets have no effect on the prices of securities. However, this does not necessarily mean that microstructure is important for our understanding of fundamental economic decisions. In section 4 we discuss a group of studies which look at whether the stock market's microstructure can also have long-term effects on prices and returns. Section 5 then sums up the most important contributions from the literature and highlights key themes and challenges in ongoing research.

2 The transaction process

2.1 Dealer markets versus limit order markets

A fundamental function of a market is to ensure that buyers and sellers find one another and have the opportunity to trade when they want to. One way of resolving the problem of coordination between buyers and sellers is to involve a dealer who undertakes to sell when somebody wants to buy, and to buy when somebody wants to sell. A trading system of this kind is illustrated in Figure 1 (a). (4) To be able to perform this function, the dealer must ensure that he has an adequate inventory of shares. In return for providing this liquidity for buyers and sellers in the market, the dealer earns the difference between the bid price and the ask price (spread).

[FIGURE 1 OMITTED]

Another way of resolving the coordination problem is to gather together all buy and sell orders in a limit order book. Figure I(b) illustrates a market of this kind. Buyers and sellers choose themselves whether they wish to provide liquidity by placing limit orders (orders to buy or sell at a given price) or demand liquidity by placing market orders (orders to buy or sell at the current price in the limit order book). In other words, a limit order market is not dependent on dealers. Trades are generated by electronically matching orders on the basis of ser rules, orders typically being prioritised first by price and then by the time they were submitted to the market.

Some markets, known as hybrid markets, have come to include elements of both types of market. One example of a market of this kind is the New York Stock Exchange (NYSE), which has evolved from a dealer market into a hybrid market where the bulk of trading goes through the limit order book but where dealers (known as specialists) have to set prices if liquidity in the stocks for which they are responsible is too low. In limit order markets, there are solutions where brokerage houses enter into agreements with listed companies to act as dealers in these companies" shares. Among other things, the broker must then ensure that the spread between bid and ask prices is not too large. (5)

2.2 Inventory models

Demsetz (1968) was the first to point out that there are costs associated with transacting shares. Besides explicit costs (such as stock exchange fees and brokers' commissions), there is also an indirect cost associated with getting to trade when you want to. As buyers and sellers do not necessarily need to trade at the same time, Demsetz argues that investors wanting to buy quickly must pay a higher price to motivate patient sellers to sell (and vice versa). Another important implication of his analysis is that the price at which you trade depends on whether or not you wish to buy or sell quickly, i.e. that there are two equilibrium prices rather than one.

The first microstructure models assumed optimal dealer behaviour. Garman (1976) looks at how a risk-neutral monopolistic dealer will set bid and ask prices in order to maximise expected profit per unit of time. The dealer wants to set prices to avoid bankruptcy, but must also ensure that prices are not set in such a way that his inventory empties. In Garman's model, the dealer sets prices once, after which buyers and sellers arrive in the form of two independent Poisson processes (6). Garman shows that it is optimal for the dealer to set different bid and ask prices, and that both prices will be functions of the frequency at which buyers and sellers arrive. Thus his model explains why there is a positive spread in a dealer market.

Amihud and Mendelson (1980) expand Garman's model into a multi-period model where the dealer balances his inventory over time by changing his prices in each period. This model shows that optimal bid and ask prices fall monotonically with the size of the dealer's inventory. In other words, the dealer lowers both bid and ask price in response to a growing inventory (and vice versa when his inventory shrinks). This behaviour is known as quote shading. Thus Amihud and Mendelson's model also means that the dealer sets a positive spread; what is new in this model is that the optimal pricing strategy also takes account of the dealer wanting to keep his inventory of shares at a given level. Madhavan and Smidt (1991, 1993) and Hasbrouck and Sofianos (1993) find empirical support for dealers actually having such a desired inventory level, but also for them appearing to be willing to move away from this desired position for long periods. One empirical implication of inventory effects and quote shading is that they lead to a return towards "normal" stock returns (mean reversion).

The main outcome of these inventory models is that dealers set bid and ask prices in such a way as to cover their order-processing and inventory-keeping costs.

2.3 Information models

The information models are to a great extent inspired by the insight of Bagehot (1971) that trading also entails a cost associated with some investors having better information than others. Like all other investors, informed investors can choose whether they want to trade or not, unlike the dealer who must always trade at the prices he sets. This means that, in cases where an informed investor wishes to trade, the dealer will always lose money. Copeland and Galai (1983) show that a dealer who cannot distinguish between informed and uninformed investors will always set a positive spread to compensate for the expected loss that he will incur if there is a positive probability of some investors being informed.

By expanding Copeland and Galai's model into a sequential trade framework, Glosten and Milgrom (1985) show how private information will be incorporated into prices over time. In their model, the dealer and other uninformed investors learn what the correct price is by observing the order flow. Thus the dealer takes account of information in the order flow when setting his prices. In this way, prices converge towards informationally...