Price movements, quote revisions and asymmetric information in securities markets: An empirical investigation

This dissertation develops two empirical market microstructure models to analyze transaction price movements, quote revisions and asymmetric information in securities markets. The first model develops generalized spread-based measures for the autocovariances of price changes and quote revisions which nest previous specifications in the literature and provide a metric for the degree of asymmetric information. Model parameters are estimated from intraday data for a cross-section of NYSE stocks using both a cross-sectional methodology and the Generalized Method of Moments on an over-identified system. Specification tests broadly support the formulation proposed here, but reveal mis-specification in previous approaches. Inferences about the impact of order flow on price and quote revisions are found to be very sensitive to the particular assumptions placed by earlier researchers. The second model provides a generalized framework for analyzing the degree of asymmetric information, assessing trading costs, and investigating the sources of short-run price volatility. The model is estimated using transaction level data for a cross-section of NYSE stocks at different times during the day. The results suggest that a large fraction of transaction price volatility is explained by market frictions, including asymmetric information. Asymmetric information itself is found to decline over the course of the day, consistent with price discovery. The temporal variation in the parameters provide an explanation for the observed U-shaped pattern in quoted bid-ask spreads and in price volatility. Execution costs are found to be substantially smaller than the bid-ask spread and to increase over the day, perhaps providing an explanation for the concentration in trade at the open