The accuracy of trade classification rules: Evidence from Nasdaq

Researchers are increasingly using data from the Nasdaq market to examine p ricing behavior, market design, and other microstructure phenomena. The val idity of any study that classifies trades as buys or sells depends on the a ccuracy of the classification method. Using a Nasdaq proprietary data set t hat identifies trade direction, we examine the validity of several trade cl assification algorithms. We find that the quote rule, the tick rule, and th e Lee and Ready (1991) rule correctly classify 76.4%, 77.66%, and 81.05% of the trades, respectively. However, all classification rules have only a ve ry limited success in classifying trades executed inside the quotes, introd ucing a bias in the accuracy of classifying large trades, trades during hig h volume periods, and ECN trades. We also find that extant algorithms do a mediocre job when used for calculating effective spreads. For Nasdaq trades , we propose a new and simple classification algorithm that improves over e xtant algorithms.