Improving Estimates of Mean Welfare and Uncertainty in Developing Countries

Reliable estimates of economic welfare for small areas are valuable inputs into the designand evaluation of development policies. This paper compares the accuracy of point estimates and confidence intervals forsmall area estimates of wealth and poverty derived from four different prediction methods: linear mixed models, Cubistregression, extreme gradient boosting, and boosted regression forests. The evaluation draws samples fromunit-level household census data from four developing countries, combines them with publicly and globallyavailable geospatial indicators to generate small area estimates, and evaluates these estimates against aggregatescalculated using the full census. Predictions of wealth are evaluated in four countries and poverty in one. All threemachine learning methods outperform the traditional linear mixed model, with extreme gradient boosting and boostedregression forests generally outperforming the other alternatives. The proposed residual bootstrap procedurereliably estimates confidence intervals for the machine learning estimators, with estimated coverage rates acrosssimulations falling between 94 and 97 percent. These results demonstrate that predictions obtained using tree-basedgradient boosting with a random effect block bootstrap generate more accurate point and uncertainty estimates thanprevailing methods for generating small area welfare estimates.