Chugging Along: The Long-Run Effects of Historical Railroads in Mexico

<p dir="ltr">This dissertation examines how nineteenth-century rail expansion shaped the long-run spatial distribution of population and patterns of industrial specialization in Mexico. Drawing on extensive archival research, I digitized a series of historical rail maps to construct a panel of network expansion from 1874 to 1910 and assembled a harmonized dataset of locality-level population from 1900 to 2020. This work builds on the economic history literature on Mexican railroads and enables the application of modern empirical methods to newly digitized historical data.</p><p dir="ltr">The first chapter investigates the long-run effects of historical rail access on population across Mexico. Using highly detailed and harmonized census data at the locality level, I measure changes in both locality population and population density at the grid-cell level. While grid-cell density remains greater near historical rail lines, locality-level associations weaken over time, consistent with rural-to-urban migration and broader diffusion. Persistence in the spatial distribution of locality population is observed only among urban localities that gained access to the Mexican Central Railway between 1882 and 1884. This conditional persistence remains evident until 1970, when modern infrastructure, often built near or alongside the rail network, appears to have exerted greater influence on population. Results suggest that the long-run influence of historical rail is closely tied to the layering of new infrastructure and reflects a process of substitution and institutional path dependence.</p><p dir="ltr">The second chapter examines the relationship between historical rail access and industrial development in Mexico’s cotton textile sector. I combine my spatial rail dataset with mill-level data from Razo and Haber (1998) to assess how the network influenced textile production. In the short run, rail access increased the value of output, but these gains dissipated within 20 years of initial connection. In the long run, municipalities with rail access before 1880 and a documented history of cotton textile production exhibit greater specialization in textiles and weaker specialization in services during the late twentieth and early twenty-first centuries. This pattern reflects sectoral persistence rather than structural transformation or a shift toward services. Overall, the findings suggest that infrastructure can spur structural change only when supported by favorable initial conditions and reinforced through complementary investments and policies.</p>