Willingness to Travel with Endogenous Distance: Evidence from the Changing Retail Landscape

Economists often use variation in consumers’ distance from services as a source of demand variation. This approach typically treats consumer-supplier distance as exogenous, despite suppliers strategically choosing locations. We develop a novel class of instruments to address this endogeneity. These instruments exploit the spatial distribution of consumer demographics and can be constructed from standard cross-sectional data, making them useful in a variety of spatial applications. Our preferred instruments use the income composition of concentric discs centered on the Central Business District. Applying these instruments to smartphone geolocation data for millions of devices across 18 metropolitan areas, we estimate consumer preferences for general merchandise chains across income groups. We show that accounting for distance endogeneity significantly alters willingness-to-travel estimates, distorting welfare conclusions. Contrary to a prevailing “retail apocalypse” narrative, consumer surplus per trip remained stable from 2010 to 2019. Ignoring endogeneity falsely suggests substantial welfare declines for lower-income households.