Liquidity, Volume, and Price Behavior: The Impact of Order vs. Quote Based Trading

We provide a three way theoretical comparison of dealer, limit order, and hybrid markets and analyze the impact that the organization of trading has on volume, liquidity, and price efficiency. We find, in particular, that trading volume is highest in the limit order market and lowest in the dealer market. Small order price impacts are lowest and large order price impacts are highest in limit order markets. Prices are most efficient in the hybrid market and least efficient in the dealer market, except when the level of informed trading is very high. Post-trade market transparency in a hybrid market hampers price efficiency for thinly traded securities. We further identify that traders behave as contrarians. We thank Bruno Biais, Ryan Davies, David Goldreich, Ingrid Lo, Rob McMillan, Christine Parlour, Jim Pesando, Avanidhar Subrahmanyam, and seminar participants at the CEA2008 meetings and the Karlsruhe Symposium on Finance, Banking and Insurance. Financial support from SSHRC and the Connaught Foundation is gratefully acknowledged. Email: [email protected]; web: http://individual.utoronto.ca/kmalinova/. Email: [email protected]; web: http://www.chass.utoronto.ca/∼apark/. Intra-day financial market trading is organized using two major mechanisms: order driven and quote driven trading. Order driven markets typically employ a public limit order book. In quote driven markets, all trades are arranged by designated institutions that post quotes. The latter markets are commonly referred to as dealer markets. Many real world markets are hybrids, combining both organizational forms. The coexistence of competitive limit order and dealer markets and the differences in their trading outcomes have long been challenged by academic research. Madhavan (1992) shows that, with competitive liquidity provision, a quote driven system and a uniform price order driven system lead to identical outcomes. Glosten (1994) and Back and Baruch (2007) argue that a quote driven system that competes with a discriminatory limit order book in an anonymous market would mimic the limit order book. Our paper builds on this line of research but serves a different purpose. We posit nonanonymity, in the sense that repeat order submissions are identified, and thus effectively take the coexistence of the mechanisms as given. Our goal is to describe the relative advantages and disadvantages of the three trading systems: a discriminatory limit order book, a dealer market, and a hybrid market. Our major contribution is twofold. First, we provide an integrated theoretical framework that admits a three-way comparison. The differences in trading outcomes of the three trading mechanisms in our setting highlight, in particular, the significance of the discriminatory order book and post-trade market transparency. Second, we employ our framework to derive novel empirical predictions for the impact of the organization of trading on volume, liquidity, and price efficiency. In comparing competitive markets, we complement the literature on markets where liquidity providers have market power (e.g., Seppi (1997)), which we discuss below. Limit order and dealer markets differ in many aspects. One defining feature of the systems is the level of market transparency enjoyed by the liquidity providers. Limit orders are posted prior to the liquidity demand realization, whereas dealers’ quotes account for the order size. We show that this difference in the liquidity providers information both yields new empirical predictions and explains a large share of the previously noted heterogeneity among the trading outcomes. The liquidity providers in the dealer market are able to determine the information See Harris (2003) for a comprehensive list. For example, on the NYSE and the Toronto Stock Exchange, traders can either send orders directly to the limit order book (the “downstairs” market) or arrange trades via floor brokers (NYSE) or “upstairs” dealers (Toronto); on Nasdaq, trades can be arranged on INET or through a dealer. Finally, a hybrid structure also arises when a limit order book competes with a dealer market. An example is Paris Bourse (a limit order market) and London Stock Exchange (a dealer market, until recently), which compete for the order flow in cross-listed stocks.