Agriculture is the primary technology which allows us to meet one of our most basic needs, nutritional sustenance, and changes in agricultural practices can have profound consequences. In this dissertation, I study how changes in agricultural production, via technology adoption, infrastructure development, and environmental shocks, affect society. In particular, I study the spatial distribution of these effects. Because agricultural commodities are traded over large distances and agricultural production generates pollution that can travel vast stretches of space, the effects can be widespread. At the same time, changes in agricultural production have local impacts on farmers and markets that experience the changes. Each chapter of the dissertation studies a different change to agricultural production along with its causal local and distant consequences.
The first chapter, co-authored with Bram Govaerts, explores the effects of conservation agriculture adoption in Mexico. Sustainable agriculture practices which allow for increased productivity while limiting environmental impacts are seen as key to achieving a number of sustainable development goals. However, there has been limited evaluation of such practices at scale, particularly with respect to their environmental impacts. Using the roll-out of an agricultural extension effort with a partial focus on conservation agriculture as a source of variation in conservation agriculture adoption, we show that locations see a decrease in agricultural burning after being reached by extension technicians. As a consequence, urban localities located downwind up to 50 kilometers away see reductions in air pollution and infant mortality. These results suggest that the value of the environmental benefits of conservation agriculture are large, at least in the case of Mexico.
The second chapter, co-authored with Mame Mor Anta Syll, Abdoulaye Cisse, Alain De Janvry, Marco Gonzalez-Navarro, Samba Mbaye, and Elisabeth Sadoulet, tackles a classic question of development economics: what are the consequences of a large increase in agricultural productivity? We study this question in the context of the Senegalese portion of the Senegal River Valley. Since the damming of the river in the early 1980s, the region has seen massive investment in irrigation infrastructure in the form of canals, pumps, and land preparation. We document, using over three decades of satellite imagery, that the development of an irrigation project leads to a large increase in agricultural production on average, but with substantial heterogeneity. We then explore the consequences of these agricultural productivity shocks in the long run by analyzing changes in urban footprint and night time luminosity of towns from 1995 to 2018. Towns with greater increases in agricultural production within 15 kilometers also experience more rapid urbanization, providing new evidence on the interaction between agricultural development, structural transformation, and urbanization.
The third chapter begins from two widely documented empirical regularities. The first is that weather shocks appear to causally alter the likelihood of civil conflict incidence in low-resource settings, particularly in sub-Saharan Africa. The second is that international price shocks generate variation in a number of economic variables which are predictive of conflict. Together, these regularities suggest that distant weather realizations may affect civil conflict incidence through general equilibrium effects. To test this hypothesis, I develop and calibrate a quantitative spatial equilibrium model of trade with risky production due to stochastic weather realizations. Once calibrated, the model can be used to answer the question of interest using standard design-based causal inference techniques. I find that decreases in labor's share of income and real wages due to distant weather shocks cause increases in civil conflict incidence. This finding allows for more nuanced predictions of how climate change may affect civil conflict in the region.
