Matching Technology and Competition in Ride-hailing Marketplaces

Département Information Systems et Operations Management

Intervenant: Kaitlin Daniels (Washing U in St Louis Olin)

Salle Bernard Ramanantsoa

Abstract

We study how operational differences influence competition between taxis and Uber. Taxis and Uber differ in their matching technology – taxis use street hailing while Ubers use central dispatch – and in their supply elasticity – taxi supply is fixed by regulation while Uber’s supply is determined by the number of drivers who choose to drive. We provide a novel characterization of the distribution of street hailing wait times, which we compare to the established wait distribution with central dispatch. While central dispatch delivers substantially shorter wait times on average in practical applications, street hailing more frequently offers short wait times. We construct a model where operational frictions from matching are the primary driver of competition between taxis and Uber for two types of wait-sensitive passengers. The first type – constrained passengers – requires a ride within a limited time frame and prefers the ride-hailing service more likely to meet this constraint. The second type – unconstrained passengers – does not adhere to a rigid schedule and prefers the ride-hailing service with shorter expected wait. We derive the status quo equilibrium where taxis use street hailing and Uber uses central dispatch and ask whether taxis can capture more market share by centralizing their dispatch. Doing so is not definitively beneficial for taxis. Taxis sacrifice their advantage
with constrained passengers and change the incentives for Uber drivers to drive, which can lead to larger Uber supply and smaller taxi market share than the status quo.