Prof. Jordan James Norris, Ph.D. (Aarhus University)
Exploiting sharp geographic and temporal variation in the entry of Uber across US cities, we find that ride-sharing availability causes a significant increase in sales, employment, and entry primarily in two industries: gyms and sports centers (NAICS 713), and restaurants and bars (NAICS 722). Using consumption microdata, we show this is driven by a demand-side response to a reduction in transportation costs caused by Uber: millenial consumers substitute from home to away-from-home consumption of food, alcohol and fitness. Ride-sharing attenuates spatial frictions, reducing the relative economic price of away-from-home consumption. These effects are polarized between core and peripheral neighborhoods, yet opposite for each industry: Uber causes greater agglomeration in 713, but dispersion in 722. We show this is theoretically consistent with demand substitution across products within 713 being greater than within 722: agglomeration forces are maximal at lower transportation costs when substitutability is higher. We provide empirical evidence supporting this demand heterogeneity by verifying theoretical implications on the distance traveled for consumption and the extent of initial agglomeration.