1 Does Diversity Impair Development? A Multiple-level Analysis John Gerring Boston University Dept of Political Science 232 Bay State Road Boston, MA 02215 jgerring@bu.edu Strom C. Thacker Boston University Dept of International Relations 152 Bay State Road Boston, MA 02215 sthacker@bu.edu Wei Huang Boston University School of Public Health 715 Albany St, Talbot Bldg Boston, MA 02118 whuang@bu.edu Yuan Lu Harvard University School of Public Health Dept of Global Health and Population Boston, MA 02115 yuanlu@hsph.harvard.edu Draft: November 12, 2011 ABSTRACT. Social diversity is generally thought to have negative repercussions for development, though the thesis has been widely tested only at national levels in the developing world. In this study, we provide a broad and systematic test of this hypothesis. Our analyses, based on Demographic and Health Surveys, incorporate a wide range of diversity measures and outcomes, a global sample of countries, multiple levels of analysis (national, regional, and district), and individual- level data relevant to development outcomes. With this set of analyses, we find that the Diversity Deficit thesis does not apply to development outcomes at subnational levels in the developing world. Indeed, we find that ethnic and religious fractionalization is often associated with desirable outcomes. Moreover, smaller groups as well as recent migrants within a region or district often demonstrate superior outcomes relative to larger groups and native inhabitants of that area, supporting our general claim that greater diversity does not harm societal outcomes. In the concluding section of the paper we speculate on the reasons why diversity might not negatively impact development at subnational levels.