Zusammenfassung:
Does foreign aid target the poorest people in developing countries? We examine this question in the context of World Bank foreign aid projects at the local level in 50 countries worldwide. We combine georeferenced data on World Bank foreign aid projects with remote-sensing data on nighttime light intensity, micro-geographic conditions, and household survey data to analyze the determinants of the spatial distribution of World Bank foreign aid projects. We employ an inner-outer buffer matching approach to identify treatment and control locations and estimate the effect of the determinants of sub-national foreign aid allocation. Our results suggest that the allocation of World Bank foreign aid projects is driven by a combination of distance-based measures, bio-geographic variables, remote-sensing data, and socio-economic variables based on household survey data. We find that locations with better living conditions, such as higher nighttime light intensity, better access to water and sanitation, or higher educational attainment, are generally more likely to receive aid projects. Our results further suggest that substantial heterogeneity exists in the factors that are relevant to the allocation of aid projects across sector classifications. This suggests that the allocation decisions on aid projects are driven rather by efficiency considerations than the needs of the local population. Additional robustness analyses further confirm our main findings.