Spatial frictions in consumption and retail competition

In this paper, we empirically quantify spatial consumption frictions and the degree
of local retail competition. We exploit a unique data set including 1.5 billion daily
transactions in combination with detailed characteristics of more than 3 million
households. Our estimates are based on a quasi-experimental approach to estimate
the causal effect of store openings. We find that a same-chain store opening in the
proximity of households' residences reduces their expenditures at incumbent stores
by 30% in the first month. Smaller effects for competitors suggest imperfect
substitutability between retail chains. Exploiting more than 350 openings, we
identify causal consumption gravity functions, which allow us to quantify spatial
consumption areas. We document significant heterogeneities across regions and
socio-demographic groups, indicating substantial inequalities in consumption
access.