Three essays in economic development: Migration networks, village-level risk sharing, and agriculture in development

This dissertation consists of three essays on migration networks, village-level risk sharing and agriculture in development. In the first essay, we examine the relative importance of family and community migrant networks in a household's decision to send migrants from Mexico to the United States. The migration decision depends on the size and characteristics of the migrant networks since networks provide information that alter the perceived returns from migration. We find that households with family or community members currently in the United States are more likely to send migrants. Family and community networks are found to be substitutes in the production of information, implying that community (family) networks are less valuable to those with substantial family (community) networks. Finally, we observe that in the presence of extensive community networks, investment in certain types of local infrastructure may promote, not inhibit, migration.;The second essay proposes an explanation as to why village-level risk sharing is imperfect. We argue that less than full insurance at the village level is the result of a combination of ex ante fixed and marginal costs associated with informal insurance arrangements. Such costs limit both the size of the insurance group and the level of insurance within that group. Five potential outcomes are identified depending on the combination of fixed and marginal costs: (1) full insurance at the village level, (2) full insurance within village subcoalitions, (3) partial insurance at the village level, (4) partial insurance within village subcoalitions and (5) no insurance (autarky). Rejection of the village-level, full insurance hypothesis may be due to any of the four other outcomes.;In the third essay, we develop a methodology for characterizing the role of agriculture in economic development. Following economic history and theory, a number of contributions agriculture to economic development are identified. Starting from a social accounting framework and using computable general equilibrium modeling, we develop a method for quantifying and decomposing the contributions of agriculture. Computable general equilibrium models for three archetype economies, based on stylized portrayals of poor net food-importing countries in Africa, Asia, and Latin America, are used to measure the various contributions of agriculture. Agricultural surplus is found to always be positive, implying a net flow of resources from the agricultural to non-agricultural sector, and is greatest in the African context and smallest in the Latin American context. The price system is the main form of surplus extraction.