This paper presents the first evidence that publicly observable retail order flow predicts short-term return reversals at market open, indicating that competitive traders fail to anticipate foreseeable uninformed demand in opening auctions. A long-short trading strategy using public trade data generates significant abnormal returns. This return pattern resulted in an estimated $16 billion transfer from retail investors to wholesalers. A difference-in-differences analysis around the 2020 NYSE floor closure shows that variation in auction design accounts for much of the pricing inefficiency. To improve market quality, I propose that wholesalers be required to report an additional measure: the trade-to-close return.

Using both public and proprietary order execution data, we show that retail trades executed by wholesalers outside typical exchanges receive significantly more favorable prices, on average, than comparable trades executed on exchanges. Contrary to the claim that brokers sacrifice execution quality for payment for order flow (PFOF), trades executed via Robinhood, a high PFOF broker, receive better-than-average prices. Additional analysis shows this advantageous pricing arises because retail orders exhibit lower levels of adverse selection, and yields no evidence market power drives up retail trading costs. Our estimates and conclusions differ from related research because we analyze a broader sample and weight observations by dollar volume rather than equally. Overall, the zero-commission PFOF model for executing retail trades has lowered total costs below 6bp –- far smaller than they were previously. 

We find that levels of competition both at the local level and at the national level have meaningful effects on patient level outcomes in the dialysis market. New patient survivability decreases by an additional 1% if the acquisition occurs in a county with no other facilities, compared to facilities acquired in competitive landscapes. Additionally, in counties with 2 facilities, we find that the acquisition has similarly poor outcomes if both facilities are owned by separate large chains as if both are owned by the same large chain (~5% decrease in survivability compared to competitive 2 facility counties). Our results contradict those of the previous literature who find no effect on competition between facilities. We conclude that competition improves patient quality, and that competition only between large chains is insufficient to drive competitive outcomes.