Payment for Order Flow and Market Quality: A Field Experiment

Abstract

The success of so-called neo-brokers has re-sparked the regulatory debate about potentially det- rimental effects of payment for order flow, culminating in a recent proposal by the Commission of the European Union to ban such arrangements. This study presents results of a field experiment conducted in cooperation with a large German neo-broker. On treatment days, large amounts of retail orders from randomly selected stocks were routed to the main market, Xetra, instead of being executed at a trading venue with payment for order flow. We observe various standard measures of liquidity and informational efficiency before, during and after the treatment, both for the treatment and a control group of similar stocks. Our difference-in-differences analyses allow for clean identification of the causal effect of payment for order flow on stock market quality. We do not observe a significant change in any of the market quality measures we consider. The analysis thus does not lend support to the claim that payment for order flow negatively affects market quality in the main market.

Keywords: Neo-broker, market quality, payment for order flow, field experiment, difference-in- differences

JEL Classification: G10, G14, G19