A New Perspective on Spatial Heterogeneity in African Development

Access to basic infrastructure is critical for quality of life and economic development. However, data about infrastructure access in low-income countries are sparse and costly to collect. We leverage satellite imagery and survey data to train a machine learning model to predict infrastructure access for each 6.72 kilometer by 6.72 kilometer area of Africa. To show this dataset's value, we use spatial regression discontinuity to study how much infrastructure heterogeneity across countries comes from differences in institutional quality, finding a positive effect of modest magnitude, reconciling previous contradicting results.