David Cushing, ITG, Inc., and Ananth Madhaven, University of Southern California, "Stock Returns and Institutional Trading at the Cross"
Discussant: Simon Gervais, University of Pennsylvania
Lawrence Harris, University of Southern California, and Vankatesh Panchapagesan, Washington University, "The Information Consent of the Limit-Order Book: Evidence from the NYSE Specialist Actions"
Discussant: Mark Lipson, New York Stock Exchange
Clifford Ball and Tarun Chordia, Vanderbilt University, "True Spreads and Equilibrium Prices"
Discussant: Charles Cao, Pennsylvania State University
Thierry Foucault, HEC; Ailsa Roell, Princeton University; and Patrik Sandas, University of Pennsylvania, "Imperfect Market Monitoring and SOES Trading"
Discussant: Paul Schultz, University of Notre Dame
Bruno Biais and Sebastien Pouget, Toulouse University, "Learning to Play Equilibrium Strategies in Experimental Financial Markets: Does Microstructure Matter?"
Discussant: Utpal Bhattacharya, Indiana University
Closing prices are used to calculate portfolio returns, to tally the net asset values of mutual funds, and as a basis for certain types of contracts and after-hours trading. Consequently, many institutional traders seek to trade at or near the close, which in turn gives rise to concerns about associated imbalances and possible gaming behavior. Cushing and Madhaven empirically analyze the behavior of stock returns at the market close for stocks of the Russell 1000, using both transaction-level data for June 1997 to July 1998 and the complete record of all market on close order imbalance indications issued by New York Stock Exchange specialists. The authors show that the last five minutes of the trading day explain a disproportionate fraction of the variation in daily returns - almost 18 percent in portfolios - although the closing period constitutes only 1.3 percent of trade time. This return phenomenon reflects both a higher fraction of nonblock trades and higher sensitivity of price to the flow of nonblock orders in the closing period. Finally, they find that systematic return reversals following order imbalance publications are consistent with temporary price pressure that is related to liquidity trading.
Specialists compete with limit-order traders to provide liquidity at the New York Stock Exchange. Since specialists see all system limit orders, they enjoy a unique advantage in this competition. Harris and Panchapagesan examine whether the limit-order book is informative about future price changes and whether specialists use this information when trading. The authors consider three actions that specialists can take when a market order arrives: stop the order; fill the order immediately at the quoted price; or fill the order immediately at an improved price. Using SuperDOT limit orders in the TORQ database, the authors find that the limit-order book is informative, especially about short-term price movements. They also find that the specialists use this information in a way that favors them (and sometimes the floor community) over the limit-order traders. The results are more evident for active stocks, in which the competition between specialists and limit-order traders is more intense. The authors also show that specialists in lower-priced stocks are less likely to initiate such actions because of the binding tick size.
Stocks and other financial assets are traded at prices that lie on a fixed grid determined by the minimum tick size permitted in the market. Consequently, observed prices and quoted spreads do not correspond to the equilibrium prices and the true spreads that would exist in a market with no minimum tick size. Ball and Chordia model the equilibrium movements of two latent variables, equilibrium price and spread, using a bivariate autoregressive process with correlated errors. They estimate the parameters governing the movements of these variables using transaction prices and information on quoted bid-ask spreads. Because of the econometric complexities created by rounding to a discrete grid, the authors use Monte Carlo Markov Chain methods to estimate the parameters. Analyzing a selection of large, heavily traded U.S. stocks, they find that most of the quoted spread is attributable to the rounding of prices, and that the adverse selection component is nonexistent or very small.
Foucault, Roell, and Sandas develop a theoretical model of price formation in a competitive dealership market. The model is designed to match some of the key institutional features of the Nasdaq's Small Order Execution System (SOES). Marketmakers post firm quotes and choose how intensively to monitor the arrival of news. Because monitoring is costly, marketmakers do not monitor the arrival of news continuously. Imperfect monitoring thus creates profit opportunities for speculators (comparable to SOES bandits), who also monitor the arrival of news in order to pick off "stale" quotes. The fact that market monitoring is a public good for the marketmakers provides further profit opportunities for the speculators. The presence of speculators widens the spread as marketmakers protect themselves against the risk of being picked off. Prices are more likely to reflect news when speculators enter the market and the aggregate level of monitoring increases. Thus, the presence of speculators in the authors' model affects the trade-off between the bid-ask spread and price discovery.
Biais and Pouget characterize the strategies played in the perfect Bayesian equilibriums of trading games with differential information and experimentally analyze deviations from and convergence to these equilibrium strategies. They consider three different market microstructures: 1) a continuous, oral double auction market; 2) a call market followed by a continuous market; and 3) a pre-opening period followed by a call market and then a continuous market. The proportion of actions inconsistent with equilibrium is significantly lower for a market in which there is a pre-opening period; this suggests that offering a platform to the agents to communicate about preplay, such as the pre-opening period, enhances their ability to learn equilibrium strategies. Focusing on the orders placed during the continuous trading phases of the market (while considering the three market structures), the authors find that the proportion of orders inconsistent with equilibrium decreases significantly as subjects become more experienced.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||National Bureau of Economic Research meeting|
|Date:||Jun 22, 1999|
|Previous Article:||Asset pricing.|
|Next Article:||International Finance and Macroeconomics.|