Working Paper: NBER ID: w29883
Authors: Thomas Ernst; Chester S. Spatt
Abstract: The paper documents important differences in payment for order flow (PFOF), spreads, and price improvement across asset classes. In stocks we show that PFOF is small. While many retail trades are executed off-exchange, we find that they receive meaningful price improvement, particularly when spreads are at their minimum. In single-name equity options, we show that PFOF is large. While all option trades are executed on-exchange, option exchanges have rules that facilitate internalization. We exploit variation in the Designated Market Maker (DMM) assignments at option exchanges to show that retail traders receive less price improvement, and worse prices, from those DMMs who pay PFOF to brokers. Current debate concerning PFOF has focused on equity routing. We show that option routing is comparatively worse, and this gives rise to a second potential conflict of interest of brokers: encouraging customers to trade assets offering higher PFOF. As fintech has eliminated retail commissions, these cross-asset differences in PFOF have become far more consequential to broker incentives.
Keywords: No keywords provided
JEL Codes: G1; G12; G13; G18
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| PFOF-paying DMMs (G35) | wider bid-ask spreads for retail option traders (G19) |
| PFOF (L29) | execution quality (C69) |
| DMMs (E13) | execution quality (C69) |
| execution quality for internalized orders in equity markets (C69) | smaller bid-ask spreads (G19) |
| PFOF impacts execution quality (C69) | brokers earn more from routing options trades (G24) |
| execution quality for options routing (C69) | lower than for equity routing (G12) |