Larger cities typically give rise to two opposite effects: tougher competition among firms and higher production costs. Using an urban model with substitutability of production factors and pro-competitive effects, I study product market responses to an increase in city population, land-use regulations, and commuting costs. I show that those responses depend on the land intensity in production. If the input share of land is low, a larger city attracts more firms setting lower prices, whereas for an intermediate land share, city expansion increases both the mass of firms and product prices. For a high land share, the mass of firms decreases with city size while product price increases. Softer land-use regulations and/or lower commuting costs reinforce pro-competitive effects, making city residents better-off via lower product prices and broader diversity.
We document the geographic concentration patterns of Russian manufacturing using detailed microgeographic data. About 80% of three‐digit industries are significantly agglomerated, and a similar share of three‐digit industry pairs is significantly coagglomerated. Industry pairs with stronger buyer–supplier links—as measured using Russian input–output tables—tend to be slightly more coagglomerated. This result is robust to instrumental variable estimation using either Canadian or US instruments. Using Canadian ad valorem transport costs as a proxy for transport costs in Russia, we further find that industries with higher transport costs are more dispersed, and industry pairs with higher transport costs are less coagglomerated.
Standard measures of competitive toughness fail to capture the fact that, as consumers optimize intertemporally, firms operating today compete with (yet non-existent) businesses which will be started tomorrow. We develop a two-tier CES model of dynamic monopolistic competition in which the impact of product differentiation on the market outcome depends crucially on the elasticity of intertemporal substitution (EIS). The degree of product differentiation per se fails to serve as a meaningful indicator of competitive toughness: what matters is its cross-effect with EIS. We also extend the model to the case of non-CES preferences to capture variable markups.