Tick Size and Adverse Selection: Spurious Effects Arising from Serial Correlation

Abstract

In this paper, we demonstrate using a simple model that reducing tick size may either reduce or increase adverse selection. Therefore, the effect of tick size on adverse selection is an important empirical question. At the same time, we demonstrate that the standard asymmetric information models used to estimate effects of informed trading rely on econometric techniques that may give spurious results. Using simulations from an equilibrium model with no asymmetric information but endogenously determined market orders, we demonstrate that standard techniques may generate a positive asymmetric information component even when the true one is zero, and that this upward bias in the adverse selection component is affected by tick size. This bias occurs in part from serial correlation of order flow that may increase under sufficiently small tick sizes due to orders being split.