Insights from a Geospatial Dataset on Uganda Water Projects
Miguel, Edward
UC Berkeley
2017
UC Berkeley
2017
Developing countries have long relied on foreign aid and international NGOs to providesupplemental public goods such as communal drinking water points. Exploitinga geospatial dataset of water projects in Uganda, I consider both eectiveness andtargeting of projects. To evaluate eectiveness as practiced at scale, I use a retrospectivedesign that links placement of water projects with Demographic andHealth Survey (DHS) clusters, nding that over a ve year span, protected sourcesproduce statistically signicant improvement of weight-for-age, height-for-age, fever,and hemoglobin levels, but not of weight-for-height and diarrhea. In exploring povertytargeting, I link water data with spatial poverty data and argue for a relativetargeting measure that compares dierent entity types to one another in terms ofhow progressive their allocation of projects is. Finally, I use a similar procedure toanalyze placement in terms of its correlation with election outcomes.