I develop a model of spatial competition between multi-branch firms in which consumers value the price of services, spatial proximity to their home branch, and the number of other branches in the firm's network. The model delivers within and across market predictions on the pattern and density of branching, the relationship of concentration to market size, the price-concentration relationship, and price dispersion. I consider the applicability of this model to the commercial banking market for retail deposits. I test the model's predictions by utilizing variation in the timing and extent of within-state branching restrictions on banks and bank holding companies and fi nd results broadly consistent with the model. These restrictions constitute a plausible natural experiment with which to identify the role of spatial externalities on the equilibrium deposit-market structure of commercial banks. I also identify and parsimoniously model difficult-to-resolve conceptual issues in separating loan-side and deposit-side factors in the determination of bank market structure.