Essays on endogenous fractionalization and its impact on economic performance

This dissertation is a collection of essays that study the effects of ethno-racial diversity or fractionalization on economic performance when diversity is a function of public policy. The first chapter analyzes how city governments expenditure level on public goods affect ethno-racial fractionalization growth in U.S. cities. Using U.S. city level data across decades, I find that public policy has a significant impact on fractionalization growth. Cities in which expenditure on public goods is higher experience a significantly lower growth of ethno-racial fractionalization. I use instrumental variables to identify the direction of causality. The negative impact of public goods spending on fractionalization growth is caused by a significant differential reaction by race to public goods spending. Cities with higher spending on public goods experience a higher growth of the non-Hispanic White population, but a lower growth of the population of other ethno-racial groups. These findings are novel because they indicate that fractionalization is neither a static characteristic of a locality, nor it is an exogenous cause of public policy design. Treating it as endogenous, I find that fractionalization has no significant effect on public goods provision.;The second chapter investigates the effect of ethno-racial fractionalization on income growth in U.S. cities. I find that fractionalization has no significant impact on income growth. This directly contradicts the conclusion of established researches that fractionalization causes significant negative impact on income growth of a locality. I consider fractionalization and income growth in U.S. cities to be jointly determined, and use instrumental variables to establish the direction of causality between fractionalization and income growth. My results show that when fractionalization is assumed to be an exogenous characteristic, it does have a significant negative impact on growth. However, instrumenting fractionalization in a 2SLS framework shows no significant impact on growth.