Dark Trading and Post-Earnings-Announcement Drift

Both theory and evidence are mixed regarding the impact on prices of trading on "dark" venues partially exempt from National Market System requirements. Theory predicts that price discovery improves as dark venues siphon noisy uninformed trades, but increased adverse selection reduces liquidity. Empirical studies, which focus on intraday inefficiency, also find contradictory results. We extend that literature to investigate the impact of dark trading on a long-standing inefficiency based on under-reaction to quarterly earnings. We study a randomized controlled trial created by the "trade-at" rule of the Securities and Exchange Commission's Tick Size Pilot Program that exogenously shocks dark trading. We supplement that with ordinary least squares and two-stage least squares regressions on amore representativeCompustat/Center for Research in Security Prices sample. Allourresultssuggestthatunder-reactionincreaseswithdarktrading,consistentwithreduced liquidity limiting arbitrage. We contribute to the literature on dark trading and inefficient processing of accounting disclosures, highlighting the role of advances in trading technology.