How Stock Splits Affect Trading: AMicrostructure Approach

Abstract

Extending an empirical technique developed in Easley,Kiefer, and O'Hara (1996), (1997a), we examinedifferent hypotheses about stock splits. In linewith the trading range hypothesis, we find thatstock splits attract uninformed traders. However, wealso find that informed trading increases, resultingin no appreciable change in the information contentof trades. Therefore, we do not find evidenceconsistent with the hypothesis that stock splitsreduce information asymmetries. The optimal ticksize hypothesis predicts that stock splits attractlimit order trading and this enhances the executionquality of trades. While we find an increase in thenumber of executed limit orders, their effect isovershadowed by the increase in the costs ofexecuting market orders due to the larger percentagespreads. On balance, the uninformed investors'overall trading costs rise after stock splits.