Noémie Zurlinden

We conjecture that mistrust and social conflict in a society may depend on ethnic stratification, i.e., the extent to which the hierarchy in socio-economic positions across individuals follows ethnolinguistic lines. We define and axiomatically characterize an index of ethnic stratification that generalizes the idea of between-group inequality to situations where data on economic and ethnolinguistic distances between pairs of individuals is available. We use Afrobarometer survey data to measure ethnic stratification at the level of towns and villages in 26 ethnically diverse African countries. We find that ethnic stratification is negatively related to trust in relatives, neighbors and other acquaintances, and positively related to nearby conflicts. These findings shed new light on the debate about the merits of conflict and contact theory.

This dissertation comprises three contributions in development economics. Using geocoded data, ethnic stratification, social conflict, favoritism and local aid effectiveness in Africa are studied. Chapter 1 conjectures that mistrust and social conflict in a society may depend on ethnic stratification, i.e., the extent to which the hierarchy in socio-economic positions across individuals follows ethnolinguistic lines. This chapter defines and axiomatically characterizes an index of ethnic stratification that generalizes the idea of between-group inequality to situations where data on economic and ethnolinguistic distances between pairs of individuals is available. It uses Afrobarometer survey data to measure ethnic stratification at the level of towns and villages in 26 ethnically diverse African countries. It shows that ethnic stratification is negatively related to trust in relatives, neighbors and other acquaintances, and positively related to nearby conflicts. These findings shed new light on the debate about the merits of conflict and contact theory. Chapter 2 conducts a systematic study of favoritism by cabinet members in Africa with continent-wide coverage (47 countries). For this purpose, birthplace information for all cabinet members between 2001 and 2014 was hand-collected. This chapter provides causal evidence of favoritism by health ministers. First, administrative regions receive more World Bank health aid when a region-born health minister is in office. Second, neonates and infants are less likely to die when the current health minister originates from their region. However, the reduction in mortality is not associated with increased health aid, implying that health ministers' favoritism also occurs through other channels. Chapter 3 studies whether World Bank health aid can improve health outcomes at the local level. Exploiting geocoded data on the location of projects and on individual-level health outcomes, it investigates whether health-related World Bank projects reduce infant mortality at the local level in 25 African countries. This chapter shows that children born close to health-related World Bank projects are less likely to die before their first birthday than their siblings born before any health projects are implemented in their vicinity. It studies whether this effect is higher where the need for aid is higher and institutions stronger. The results suggest that health aid is more effective in more disadvantaged subnational areas but not in poorer countries. No evidence is found that the effect of health projects is higher in subnational regions or countries with stronger institutions.

Noémie Zurlinden

Filters

Settings

Sort By

Results per page