Spatial differentiation in the supermarket industry: The role of common information

In this paper, I investigate the geographic location decisions of supermarkets to infer their tradeoffs between locating close to favorable demand conditions and differentiating themselves geographically from rivals. The model is based on a discrete-choice game between two types of supermarkets, and incorporates firm uncertainty arising from firm- and location-level private information as well as researcher uncertainty arising from location-level common information. Thus the model addresses the concern that firms' actions may be based on factors that are unobservable to the researcher, thus correlated conditional on observables. The estimates reflect a significant level of common information. Importantly, I find that ignoring unobserved location heterogeneity results in biased estimates of both the competitive effects and the effects of location-specific observables on profits. Counterfactual predictions are therefore misleading if unobserved location heterogeneity is unaccounted for.