Institutional trading costs, Robinhood activity, and expected returns
Abstract
Using high-frequency TAQ data, I decompose the permanent price impact in U.S. equities into retail and institutional components. Institutional trades have substantially larger permanent price impact than retail trades and are more closely related to effective spreads, information, and proxies for limits-to-arbitrage. The institutional price impact predicts future stock returns in a size-neutral trading strategy and the predictive power is not absorbed by standard proxies for trading costs, adverse selection, and limits-to-arbitrage. An orthogonalized measure of institutional price impact continues to predict future returns, indicating that it contains a priced component beyond existing lower-frequency measures. Using Robinhood trading and platform outages as shocks to retail order flow, I show that Robinhood activity primarily drives this residual component and provides an economic driver of institutional price impact.
Keywords
Price impact; Trading costs; Cross-sectional asset pricing; Retail trading; Robinhood