Entry and location choice in retail markets

I develop an empirical model to analyze differentiation and market structure over large geographic areas based on firms' entry and location choices. While other entry models have been applied to product differentiation, their use of a reduced-form profit function makes them difficult to apply to geographic areas composed of many small interrelated markets. I begin by examining different reduced-form approaches to modeling profits in such markets. While the simplicity of these reduced-form models makes there more tractable in determining the equilibrium configuration of the market and the techniques to estimate the models, they do not capture the more complex ways in which firms compete for customers in markets where differentiation plays a significant role. I resolve this problem by modeling the benefits from differentiation more directly, incorporating consumers' preferences for location into a structural profit function.{09}Using only aggregate demographic data and firms' location choices, the model allows me to calculate more complex competitive effects among firms---competitive effects that vary with differentiation, the number of competitors and potentially the intensity of price competition. An additional benefit of this approach is that it yields the geographic market definition in which firms compete. I adopt a simulations approach to estimate the structural model, where parameters of the profit function are chosen to best fit the simulated entry patterns to observed entry patterns. Both the reduced-form and structural models are applied to San Francisco dry cleaning establishments, revealing that competition is fairly localized; for the average firm, the structural estimates show that demand decreases by 16% when an entrant enters nearby and 5% when it enters 1/2 mile away.{09}The average firm draws its demand primarily from within a 1/2 mile radius; yielding the geographic market definition for most firms in the market.