Does quoted depth predict intraday stock returns?: theory and evidence

We generate and test new implications of Kavajecz’s (1996a) market microstructure model for cross-sectional variation in intraday expected stock returns. In this model, the specialist presents a price schedule consisting of bid and ask prices and a bid and ask size. We document that in his model, the specialist reveals through the bid-ask size spread what she believes to be the expected return on the risky asset. We then examine whether the ability to act on this revealed information is o¤set (if only partially) by movements in the remaining choice variables (namely, the bid-ask spread). We prove that when the intraday expected return on the asset is relatively high or low, the specialist sets a lower bid-ask spread, allowing rational traders to use the revealed information concerning the likely movement of the stock in their trading decisions. We test these implications using 1993-1994 intraday quote data from the NYSE Trade and Quote (TAQ) database. We con...rm our theoretical ...ndings that the relative sizes of the bid and ask quotes given by NYSE specialists provide information about future price movements. Our empirical evidence has major implications for several important groups of investors, including index arbitrageurs and fund managers. These traders must buy (sell) a speci...c type of stock (perhaps one possessing a particular degree of riskiness or industry classi...cation). Our theoretical and empirical results indicate that these traders should buy (sell) the stock among the stocks of the type in question with the largest (smallest) bid-ask size spread. Additionally, though we model a one-period economy, our results suggest potentially pro...table strategies for transactions timing in a particular stock. Back-of-the-envelope calculations suggest that a trader with 100 percent monthly portfolio turnover can earn an additional 2.5% per year using this strategy. *Brown ([email protected]) is from the Department of Finance, University of Texas at Austin, CBA 6.222, Austin, TX 78712-1179. Cohen ([email protected]) and Polk ([email protected]) are from the Graduate School of Business, University of Chicago, 1101 E. 58th Street, Chicago, IL 60637. The authors wish to thank Eugene Fama and Doug Diamond for helpful discussion. We generate and test new implications of Kavajecz’s (1996a) market microstructure model for cross-sectional variation in intraday expected stock returns. In this model, the specialist presents a price schedule consisting of bid and ask prices and a bid and ask size. We document that in his model, the specialist reveals through the bid-ask size spread what she believes to be the expected return on the risky asset. We then examine whether the ability to act on this revealed information is o¤set (if only partially) by movements in the remaining choice variables (namely, the bid-ask spread). We prove that when the intraday expected return on the asset is relatively high or low, the specialist sets a lower bid-ask spread, allowing rational traders to use the revealed information concerning the likely movement of the stock in their trading decisions. We test these implications using 1993-1994 intraday quote data from the NYSE Trade and Quote (TAQ) database. We con...rm our theoretical ...ndings that the relative sizes of the bid and ask quotes given by NYSE specialists provide information about future price movements. Our empirical evidence has major implications for several important groups of investors, including index arbitrageurs and fund managers. These traders must buy (sell) a speci...c type of stock (perhaps one possessing a particular degree of riskiness or industry classi...cation). Our theoretical and empirical results indicate that these traders should buy (sell) the stock among the stocks of the type in question with the