We estimate the causal impact of communication infrastructure quality on growth and structural transformation. We use variation across US counties’ Internet speeds in 2018 and build an instrument using ARPANET, a military network that preceded the modern Internet, with its location documented in historical government reports. We find that doubling Internet speeds increases the 4-year employment growth by 3.3–6.1 percentage points. Faster Internet shifts economic activity toward high-skilled services and away from non-tradeable services while increasing inequality. Industry linkages, capital-skill complementarity, and information and communication technology workers’ sorting rationalize our results. Medium and small cities and rural areas drive our results.

We study the impacts of a colonial segregation policy on modern-day spatial population patterns and residential sorting by human capital in Mexican cities. After the Conquest, the Spanish aimed to segregate Indigenous individuals into settlements called pueblos de indios. While the segregation policy lasted until the end of the colonial era, we use present-day census data at the block level on population, schooling and access to medical services to understand the persistent effects of pueblos on within-city structure. First, we document a spatial non-monotonic correlation between the location of the pueblos and population deagglomerations. Second, we study the causal impact of the pueblos on sorting by human capital by exploiting quasi-experimental variation created by pueblos' catchment area defined by colonial laws. Using a Regression Discontinuity Design, we find a slight increase in the blocks' share of households lacking access to medical services near the boundary of all pueblos. We further exploit the degree of success of the policy: it effectively isolated the Indigenous population only in a fraction of the settlements. We show that blocks near to the pueblos where the segregation policy was successful experience discontinuities on the blocks' measures of schooling of minus 0.38 years and the blocks' share of individuals with no access to medical services of 2.2 percentage points. Worse urban amenities and lower land prices explain our results. Our findings show that colonial segregation institutions have persistent impacts for centuries on urban sorting, even when the originally targeted group becomes a very small share of the modern-day population.

I quantify the impacts of road infrastructure improvements on a country's comparative advantage. First, I build an international trade and internal geography model with heterogeneous shipping routes and input-output linkages. Then, I simulate how different roads projects that improve the connection between the largest agricultural and manufacturing regions of the country with seaports, impact the comparative advantage of Colombia, a developing nation with concentrated exports in commodities. I consider roads that improve the access of heterogeneous regions to global markets. My results show that roads can shift the comparative advantage of a country to manufacturing, and away from mining. Industry linkages and access to inputs are key to understand these shifts. Hence, I confirm that a country's comparative advantage is shaped by domestic trade costs, in addition to classical determinants like endowments, technology, and institutions. Lastly, my results suggest that road infrastructure influences the structural transformation of a nation.

We provide evidence that new residential construction can lead to gentrification. Between 2006 and 2008, developers in NYC rushed to claim expiring property tax incentives and built more housing. We instrument the number of new units built by the number of vacant parcels available before these incentives expired. Using novel building-level data, we find that existing buildings’ rents increased by 1.8% in response to a 1% increase in rental stock within a 150-meter radius. We document that the new buildings generated new amenities, thus attracting to nearby blocks high-income, college-educated, young and white households who are new to the neighborhood. Moreover, new housing units reduce school enrollment. These findings indicate that new construction can spur gentrification in a dense city with strong supply frictions, which might explain the opposition to new housing by working class residents. Our results do not imply local governments should restrict the construction of new housing. Rather, our findings suggest that local governments might need to consider the political economy of tax breaks for new residential investment.

I study how commodity booms impact the structural transformation of regions in a developing country context. To do this, I quantify the impact of commodity boom shocks on regional manufacturing and welfare outcomes through the lens of a spatial model with both domestic and international trade. My framework includes heterogeneous geography and sectoral linkages. I use truck shipments data, customs administrative datasets, and local records of commodity extraction to calibrate my model. My results show that two local different commodity booms have different impacts on local manufacturing, but similar effects on the manufacturing output of neighboring regions. I provide evidence that commodity booms can have both positive and negative spillovers on regional manufacturing output. My findings are partially in line with Albert Hirschman's hypothesis on the industrialization of commodity-dependent developing nations. I provide light on the structural transformation of regions, and the limitations of empirical studies regarding the impact of commodity booms on local economic outcomes, such as the local resource curse.