The impact of preferencing on execution quality

We examine the impact of preferencing on execution quality for NASDAQ and NYSE-listed stocks. Our theoretical model demonstrates that realized spreads are more reliable than effective spreads in the presence of preferencing, but even realized spreads are a poor measure of execution quality if the stocks being compared have different degrees of information asymmetry. We provide a new measure of the costs of preferencing that is independent of asymmetric information. Using data from the SEC 11Ac1-5 reports for marketable orders of up to 2000 shares, we find that both realized spreads and our preferencing measure are lower for NYSE-listed stocks.