Strength in Numbers: Ethnic Group Size and Economic Development 

Do agglomeration gains operate within ethnic groups? We study whether larger coethnic populations lead to higher levels of group economic development in the long run. To identify this effect, we exploit the arbitrary partition of African ethnic homelands caused by the Scramble for Africa, which generated exogenous variation in the size of the same ethnic group across adjacent countries. We find that a 10 percent increase in pre-colonial population causes a roughly 2 percent increase in nighttime light density today. Results are robust to a battery of geographic controls and to alternative measures of pre-colonial population, including machine learning-based predictions and land area. We trace the mechanisms using geocoded survey data: larger populations have generated denser urban agglomerations, better transportation infrastructure, and a greater likelihood of hosting the national capital. Residents of larger-group territories have attained higher education, work in more formalized and occupationally diversified labor markets, exhibit better anthropometric outcomes (a marker of cumulative long-run development), and display fertility patterns consistent with a demographic transition. These gains are place-based rather than ancestry-specific. While larger groups are also more politically powerful, absolute population size (economic gains of scale) outperforms population share (relative political power) in predicting long-run development.