An extensive set of land use regulations shape the geography of most major US cities. I study how the distribution of these restrictions affects the location of industries within the city, aggregate productivity, and welfare. In a theoretical framework, the welfare effects of density restrictions depend on the strength of spillover effects, which I estimate using multiple datasets on economic activity within New York City. The estimates based on variation in the historical zoning reform imply higher productive externalities in the office sector and negative amenity externalities in the industrial sectors. Counterfactual policy simulations show that current density restrictions are very costly — increasing allowed density in the city by 5% leads to a welfare gain of 1%. Heterogeneity across sectors magnifies the welfare effects of restrictions on the office sector and in the city center.
This paper studies the welfare implications of input tariff liberalization in a model featuring firm heterogeneity in both exposure and responses to input tariff shocks. We show that this setting gives rise to a new anti-competitive effect that dampens the benchmark welfare gains from trade liberalization. We derive a sufficient statistic for this anti-competitive effect in a general environment: the correlation between firm market shares and their cost shares of imported inputs. The interplay between import intensiveness and market power implies that larger firms benefit more from input tariff cuts while passing through less of the cost reductions, leading to increased aggregate markup dispersion. We empirically test this framework using granular regulatory data from Colombia around its 2010 trade reform. The data covers the universe of firm-level inputs, both imported and domestically sourced, allowing us to disentangle firm-specific exposure and responses to input tariff shocks. Our quantitative analysis demonstrates that input tariff liberalization induces a substantial anti-competitive effect, comparable in magnitude to the pro-competitive effects in the studied episode.