Essays on conflict, diversity, and economic development

This dissertation examines the relationships between ethnolinguistic diversity, civil conflict, and economic development. The first chapter examines the relationship between regional ethnolinguistic diversity and conflict in Sub-Saharan Africa. We use census subsamples and large-scale household surveys to construct a new subnational-level dataset on ethnic inequality capturing group-level differences in education, asset ownership, and access to basic amenities, covering several hundred regions in thirty-five countries. We distinguish between deep-rooted and contemporary cleavages by incorporating ethnolinguistic groups from our sample into Ethnologue’s linguistic tree model and constructing multiple diversity measures at the level of ancestral language families. Our analysis, based on within-country variation and accounting for numerous regional characteristics, shows that there is no robust pattern linking ethnic inequality to conflict frequency or severity. In contrast, there is a clear positive relationship between ethnolinguistic fractionalization and conflict, and it is generally stronger for metrics based on deep-rooted cleavages. The second chapter measures local ethnolinguistic diversity in Sub-Saharan Africa using Hill numbers. We measure three types of Hill numbers – (i) ordinary Hill numbers that account for the shares of different groups in the regional population, (ii) functional Hill numbers that incorporate the linguistic distances between each pair of groups in a region, and (iii) phylogenetic Hill numbers that incorporate language tree information to reflect lineage diversity in a region. Phylogenetic Hill numbers and ordinary Hill numbers that reflect deep-rooted diversity are associated with worse education and health outcomes. The third chapter investigates the role of a civil conflict on internal migration in Nepal. We use household surveys and censuses between 1995 and 2018 to measure migration flows across rural and urban sectors of five administrative regions of Nepal. Using a pseudo-panel of bilateral migration flows over time, we find that greater conflict was associated with significantly more emigration. This pattern was more pronounced for rural-urban migration. Disaggregated analyses show that conflict resulting in civilian deaths was associated with greater emigration while conflict resulting in combatant deaths was not. Migrants with low education and those in elementary occupations were more responsive to conflict than those with higher education or skilled occupations.