سفارشات محدود و رفتار روزانه ی نقدینگی بازار: شواهد از بورس تورنتو

This paper examines the intraday behavior of market liquidity on the TSE. It shows that spread follows U-shaped intraday pattern, depth displays opposite pattern, while volume is low at the open, stable during the day and increases at the close. The paper finds evidence that spread and depth are negatively correlated, suggesting that limit-order traders actively manage both price and quantity dimensions of liquidity to protect themselves from informed trades. Furthermore, it finds that there is price improvement on the TSE. Finally, it shows that liquidity is inversely related to volatility but directly related to volume.

Major stock markets in the United States, such as the New York Stock Exchange (NYSE), the National Association of Securities Dealers Automated Quotation System (Nasdaq), and the American Stock Exchange (AMEX), employ a hybrid trading system in which both market makers (specialists and dealers) and limit order traders play a pivotal role in providing liquidity to the market. However, the trading systems with market makers are not quite popular outside the United States. Indeed, Ahn and Cheung (1999) document that among the top 37 stock markets outside the U.S. only three exchanges employ the trading systems with market makers. The rest operate under the pure order-driven mechanism in which all liquidity is provided by limit-order traders alone. Even in markets with a market-maker trading system, the reliance on the market makers for liquidity has been diminished due, in part, to the introduction of various computerized trading systems that automatically match buy and sell orders. Although the pure order-driven trading system is used in the majority of markets in the world, researchers have mainly focused on hybrid markets, namely, the NYSE and Nasdaq, because they are major stock markets both in the U.S. and in the world. Only a few studies in the market microstructure literature have so far empirically investigated the pure order-driven trading system and relatively little is known about this market microstructure1. This paper intends to bridge this gap by conducting an empirical analysis on the Toronto Stock Exchange (TSE), an order-driven stock exchange. More specifically, I attempt to examine the intraday behavior of the market liquidity on the TSE to see if it has similar intraday patterns observed in other markets such as the NYSE or Nasdaq. The TSE is chosen for the analysis because it provides an ideal setting to examine the behavior of limit order traders. For instance, the market entirely relies on limit orders for liquidity, and it is so transparent that traders can use detailed information in the limit-order book to formulate their trading strategy. Limit orders play a very important role in providing liquidity to the world stock markets of various market architectures. In an order-driven market, such as the TSE, the Paris Bourse, or the Tokyo Stock Exchange, limit orders provide all liquidity to the market. In a hybrid specialist market, such as the NYSE, a large amount of liquidity comes from limit orders2. Even in a dealer market, such as Nasdaq, limit order trading is also allowed. In a pure order-driven market, investors can submit limit orders or market orders. Limit orders are kept in a limit-order book, waiting for execution. The advantage of limit order is that traders can control the price at which the order might be executed. However, there is risk associated with it. First, a limit order might not be executed. Second, as limit price is fixed, there is adverse selection risk due to the arrival of informed traders3. Market orders, on the other hand, are executed with certainty. However, traders have no control over price. Glosten (1994) examines an equilibrium model in which there are two types of traders. The patient traders place limit orders and therefore supply liquidity to the market. The urgent traders, on the other hand, place market orders and consume liquidity. Informed traders are more likely to be urgent than patient because they want to exploit their superior information4. Glosten shows that patient traders would not place limit orders unless the expected gains from trading with liquidity traders exceeded the expected loss from trading with informed traders. However, his model does not endogenize the traders' choice between market order and limit order. Handa and Schwartz (1996) extend Glosten's analysis by examining the investors' rational choice between market and limit orders. The choice depends on the investors' beliefs about the probability of their orders being executed against an informed or a liquidity trader. Handa and Schwartz show that in an order-driven market, if price volatility is high, investors submit more limit orders than market orders because the expected gain from providing liquidity to the market exceeds the potential loss from trading with informed traders. Foucault (1999) argues that price volatility determines the mix between market and limit orders. He shows that when market is volatile, the probability of trading against informed investors increases; therefore, the expected loss to them is larger. To protect themselves, limit buy (sell) order traders have to post lower (higher) bid (ask) price and/or reduce their order sizes. This establishes a direct relationship between bid-ask spread and volatility, and an inverse relationship between market depth and volatility. Like Handa and Schwartz (1996), Foucault (1999) also predicts that when price is volatile, market orders become less attractive than limit orders; as a result, more limit orders than market orders are placed. Seppi (1997) develops a microstructure model of liquidity provision in which a specialist competes against limit-order traders for profit. Seppi shows that hybrid specialist markets, such as the NYSE, provide better liquidity to small and large investors, but pure order-driven markets, such as the TSE or the Paris Bourse, offer better liquidity to mid-size order traders. Bondarenko and Sung (2003) show that a specialist has incentives to trade against market trend if limit-order book is thin, and to trade with the market trend if it is deep. In addition, the participation rate of the specialist also depends on the uncertainty of the depth of limit-order book, price volatility, and trading volume. Harris (1990) defines liquidity as the willingness of some traders to take the opposite side of a trade at low cost. Thus, Harris argues that market liquidity should have two dimensions: the price dimension, represented by spread, and the quantity dimension, represented by depth. A market is liquid if it is deep and spread is narrow. In a market, a complete quote comprises a bid price, an ask price, and depth which is the number of shares available at each price. Harris argues that if there is an indication that the probability of informed trade in the market has increased, limit buy (sell) order traders may respond by posting lower (higher) bid (ask) price and/or reduce their order sizes. This implies a negative correlation between spread and depth. On the empirical side, numerous studies examine the intraday variation in bid-ask spread. Brock and Kleidon (1992), McInish and Wood (1992), Lee, Mucklow, and Ready (1993), Chan, Chung, and Johnson (1995) find that the spreads of NYSE stocks exhibit a U-shaped intraday pattern. These studies attribute this U-shaped pattern to specialists' market making behavior. In contrast, Chan, Christie, and Schultz (1995) find that for Nasdaq stocks, spreads decline throughout the day but narrow significantly during the last 30 min. They attribute these divergent patterns to the structural differences between specialist and dealer markets. Chung and Van Ness (2001) compare intraday variations in the spreads of Nasdaq stocks before and after the market reforms. They find that spread has declined since the Securities and Exchange Commission changed the order handling and tick size rules on the Nasdaq in 1997. Kavajecz (1999) compares the spreads of the limit-order book with that of specialists. Kavajecz finds that specialists play a vital role in narrowing the bid-ask spread, especially for less frequently traded stocks. Lee et al. (1993) show that on the NYSE, wide spreads are accompanied by low depths and liquidity falls in response to high volume and anticipation of earnings announcements. Harris (1987) finds that there is an inverse relationship between price volatility and liquidity. However, in his study, liquidity is not measured directly by spread or depth but by a set of hypothesized explanatory variables. Madhavan (1992) suggests that given trading volumes, the number of trades may be positively related to liquidity. Chordia, Roll, and Sabrahmanyam (2001) analyze the aggregate market liquidity and the trading activity on a day-to-day basis of the NYSE stocks over an eleven-year period. They find that aggregate market spreads, depths, and volumes are more volatile than returns and they are influenced by short-and long-term interest rates, market volatility, the days of the week and major announcements. Interestingly, contrary to intuition, they find that market volatility tends to reduce spreads. Biais et al. (1995) show that on the Paris Bourse, investors tend to place limit (market) orders when the spread is wide (narrow) or the depth is low (high). To compete for price and time priority, investors quickly place limit orders within the quotes when the depth at the quote is high or the spread is wide. Chung, Van Ness, and Van Ness (1999) show that the U-shaped intraday pattern of spread on the NYSE largely reflects the intraday behavior of the spread established by limit-order traders. However, the quote classification method used by Chung et al. fails to take into account the fact that specialists may manipulate traders' orders for price improvement5. This could lead to misleading results. Chung and Van Ness (2001) also find that limit-order traders play a significant role in the quote-setting process for Nasdaq stocks; however, spread on this market exhibits a different pattern. It declines throughout the day and narrow significantly during the last 30 min. Chan, Christie, et al. (1995) explain the difference in intraday spreads between the NYSE and Nasdaq by the structural differences between the two markets. Given that the structure of a pure order-driven market differs from that of a hybrid market, could it be the case that spread on a pure order-driven market display another different intraday pattern? To investigate this issue, this paper examines the intraday behavior of market liquidity (both spread and depth) on the TSE, a pure order-driven market. The primary finding in the paper is that spread (both quoted and effective spreads) displays a U-shaped intraday pattern. It is widest at the opening, narrows down monotonically in the first two hours of trading and then becomes stable. It slightly picks up at the market close. Depth exhibits the opposite pattern. The market is very thin at the beginning, becomes deeper and deeper during the first one and a half hours of trading and then becomes stable. In addition, the paper finds that spread and depth are negatively correlated. It also examines whether there is price improvement in an order-driven market6. It shows that as in other markets, price improvement still exists in a pure order-driven market. This could be due to several factors such as, marketable limit orders, the upstairs market, the Intermarket Trading System, among others. In contrast to the theoretical prediction of Easley and O'Hara (1992) and the empirical findings of McInish and Wood (1992), Lee et al. (1993) on the NYSE, the paper discovers that liquidity is directly related to volume of trade. However, this finding can be explained by the theory of increased trading induced by differences in opinion among investors, proposed by Harris and Raviv (1993). Finally, it provides evidence to substantiate the theory of Handa and Schwartz (1996), and Foucault (1999) that volatile market reduces liquidity. This paper proceeds as follows. Section 2 describes the market and the dataset. Section 3 examines the intraday pattern of depth, spread and volume of TSE stocks. Section 4 presents the empirical relations among liquidity, volume, price volatility, and trading activity. Section 5 provides some concluding remarks.