Ethnic Homogenization and Public Goods: Evidence from Kenya’s Land Reform Program

Little is known about the effects of ethnic homogenization policies despite being a logical corollary of the literature that links ethnic fractionalization to negative development outcomes. In this paper, we examine the effects of ethnic  homogenization on public goods provision using a natural experiment that took place in Kenya. We study a large-scale land reform program that took place in 1964-1974 and led to a significant reduction in ethnic diversity, the settlement schemes program. Using a novel dataset about the precise location of program area boundaries (Lukalo et al., 2019) that we combine with archival, survey, census, and satellite data, we implement a spatial regression discontinuity design. We argue that the border between program areas (treatment) and neighboring areas (counterfactual) is plausibly random at the local level and confirm that there are no observable differences in pre-treatment characteristics. We find a strong discontinuity in ethnic diversity but no differences in school provision between program areas and counterfactual areas in the short run as well as in the long run. We hypothesize that, as individuals were resettled to the program areas, they lack, in spite of co-ethnicity, the dense social networks that favor collective action to either hold politicians accountable or to provide public goods throughout cooperation at the community level. Our results are not driven by spillovers from treatment to counterfactual areas. A mediation analysis indicates that income effects are unlikely to drive this null result.

Speaker: Juliette Crespin-Boucaud (EthnicGoods project, Institut Barcelona d’Estudis Internacionals (IBEI))