We examine the informational efficiency of size-based US exchange traded funds (ETFs) and comparable Center for Research in Security Prices portfolios. ETFs are better suited for market efficiency tests since they avoid potential asynchronous trading problems, and their negligible bid-ask spreads greatly diminish noise due to the bid-ask bounce. Variance ratio analysis demonstrates that return autocorrelations have diminished significantly over the past decade. Granger causality tests reject the presence of lead-lag effects among size-based ETFs. However, volatility spills over from large firm ETFs to those of smaller firms, and these spillovers extend to ETF option implied volatilities.