Seminar with Chandra Krishnamurthy: “Do ride-hailing services worsen congestion and air quality? Evidence from Uber in California”

Abstract

Ride-hailing services, such as Uber, have become a significant component of urban transportation over the past decade, and have affected many aspects of transportation, including transit, air pollution, traffic patterns and conges­tion. The effects of Uber, in particular, upon traffic patterns and pollution can be ambiguous, since it depends upon its relation to transit and private vehicle usage. We empirically identify the effect of Uber upon traffic and air pollution in California using a panel difference-in-differences design, ex­ploiting variation in the timing of entry of Uber into different counties in California, and hourly traffic data on California's freeways along with daily pollution data from all pollution monitors in California. Our empirical esti­mates suggest that congestion is reduced following Uber's entry to an aver­age county. There is however, significant variations over space and time in this effect, with congestion relief restricted to the lesser populated counties (and to the off peak afternoon periods), with the most populated ones (and the morning peak hours) seeing an increase in congestion (and VMT). We also find a reduction in PM2.5 following Uber's entry into an average coun­ty, but no change in the concentration of other pollutants associated with vehicle emissions, specifically NO2, ozone and CO. Similarly spatial­ly heterogeneous effects upon pollutants are seen: the most populated coun­ties see little change in PM2.5, while all of the pollution reduction is seen in the less densely populated counties.