Research

In many slums, especially in Latin American cities like Rio de Janeiro, crime rates are particularly high, which results in many negative outcomes, including educational disruptions and acute stress. We find that the neighborhood shootings close to the school shortly before the exam have a robust negative effect on students’ exam performance on the language and math university entrance exam tests. The effects are robust to the inclusion of school and year fixed effects, as exposed students perform significantly worse (by 0.1-0.3 st. dev.) than non-exposed students.

We examine the links between urban and rural economic performance in India, the country with the largest rural population in the world, using geo-spatial socio-economic data on 1,000 cities and 500,000 villages. Rural measures of development fall almost universally with distance from towns; the gradients are shallow and the urban-rural gaps are very large even at very short distances from town centers. A Bartik estimation of the effects of urban demand shocks reveals a clear pattern of urban-to-rural spillovers. Spillovers are larger and wider for manufacturing than for service-sector jobs, and extend further along major highway networks. Surprisingly, urban demand raises population in proximate rural areas but not in the urban areas themselves, perhaps suggesting constraints on densification.

The most urbanized continent, Latin America is lagging in economic development expected of its high urbanization rates. Such ‘urbanization without growth’ has been linked in some contexts to upper classes benefitting from natural resource extraction while having a preference for urban living. Across Latin America, I find that while industrialization and resource rents do explain some variation, they are insufficient to explain the exceptionally high urbanization rates. I present suggestive evidence that the colonial past may have created an urban system that was later conducive to speeding up urbanization without growth.