Document Type

Working Paper

Publication Date


JEL Codes

C21, D22, D43, H41, R40


C21 - Single Equation Models; Single Variables: Cross-Sectional Models; Spatial Models; Treatment Effect Models; Quantile Regressions, D22 - Firm Behavior: Empirical Analysis, D43 - Market Structure and Pricing: Oligopoly and Other Forms of Market Imperfection, H41 - Public Goods, R40 - Transportation Systems: General


Since 2011, the private ride-hailing companies Uber and Lyft have expanded into more and more US cities. We use regression discontinuity design to examine the impact of Uber and Lyft’s entry on public transportation use in the US’ largest urban areas. In most cases, entry into cities by the two ride-hailing companies was staggered: Uber entered first followed some months later by Lyft. We find that public transportation use increased in an urban area, all else equal, immediately following the first entry. However, we find that the spike in public transportation use after first entry disappeared following the entry of the second company. In fact there is some evidence that monthly public transportation ridership levels fell below their pre-first entry levels. In other words, the joint presence of the two major private ride-hailing services transformed ride-hailing services from a public transportation complement to a public transportation substitute, at least in the studied urban areas. We speculate that the first entrant complemented public transportation use for some in an urban area by solving the “last-mile” problem and by providing a potentially safer option at night when public transportation service has been reduced. However, we speculate the second entrant is likely to have spurred price competition in the urban area’s ride-hailing duopoly market and an increase in ride-hailing car supply. This competitive effect could have tipped the scales, making an entire trip with a ride-hailing service more cost-effective and convenient than splitting a trip between a ride-share company and public transportation.