查看文章

Between 2010 and 2019, Uber and Lyft launched in hundreds of cities across the United States. During that decade, these transportation network companies (TNCs) frequently asserted that their ridesourcing services brought increased jobs, wages and economic growth to the cities they served. Many such claims emphasized job flexibility for drivers, increased mobility for passengers, and the combined potential of both populations to stimulate economic activity. We test these claims by leveraging the staggered entry of Uber and Lyft across 167 service regions nationwide to estimate the aggregate effects on jobs, wages, and regional GDP. Given that the timing of treatment is staggered, we deploy three difference-in-differences approaches that generate consistent estimates in the presence of heterogeneous treatment effects: the Callaway & Sant’Anna estimator, the Sun & Abraham estimator, and stacked regression. Across all three methods, we find that Uber and Lyft entry caused (1) an increase in employment per working age population, especially for seasonal, temporary, or otherwise intermittent jobs;(2) an increase in earnings per working age population for intermittent jobs; and (3) an increase in gross domestic product (GDP) per capita. Our results are broadly consistent in sign, magnitude, confidence intervals, and statistical significance, across all three estimation methods. In addition to reporting point estimates and 95% confidence intervals, we discuss possible mechanisms of TNC economic impact with particular attention to challenges in estimating and interpreting the magnitudes of treatment effects.