Abstract Most of the current literature on transportation network companies (TNCs) or ride-hailing services (e.g., Uber, Lyft) has focused on major metropolitan areas, though TNCs are currently extending their services to smaller cities. This point is particularly salient when examining their impacts on alternative transportation modes, such as public transit which operates very differently in small or medium-sized cities (e.g., less frequent services, shorter routes). This study is the first to use micro-level public transit data to examine the effects of Uber on bus ridership in a medium-sized urban area. Uber entered the neighboring cities of Eugene and Springfield, Oregon on July 23, 2014, but abruptly stopped operating on April 5, 2015 after being sued by the City of Eugene. Consequently, this study leverages a natural experiment since the period during which Uber was active is plausibly exogenous to other factors that affect bus ridership. Results show that when Uber operated in treatment cities, bus ridership reduced by 5.4 percent (p < 0.01) relative to control cities and when Uber was not active, though the magnitude of these reductions varied within and across days, as well as across neighborhoods. Interestingly, our results also suggest that even after Uber exited, these reductions in bus ridership persisted. Results highlight the importance of utilizing micro-level bus data and have important policy implications for how smaller cities could stem possible reductions in public transit after TNCs enter the market.

Highlights • First study to examine bus ridership and ride-hailing services in medium-sized city • One of the few studies to use micro-level bus data • Results show Uber substitutes for buses with ridership decreasing by 5% • Size of effect varies within and across days and across neighborhoods • Persistent reduction in bus ridership even after Uber exits