Regression-based Imputation for Poverty Measurement in Data Scarce Settings

Working Paper 2022-611

Abstract

Measuring poverty trends and dynamics are important inputs in the formulation and design of poverty reduction policies. The empirical underpinnings of such exercises are often constrained by the absence of suitable data. We provide a broad, generalist, overview of regression-based imputation methods that have seen widespread application to estimate poverty outcomes in data-scarce environments. In particular, we review two imputation methods employed in tracking poverty over time and estimating poverty dynamics. We also discuss new areas that promise of further research.

Authors: Hai-Anh Dang, Peter Lanjouw.