Survey-to-Survey Imputation for Estimating Poverty : Operational Decision Tree for Assessing Feasibility, Risk, and Needed Validations

Overview & Synopsis

Survey-to-survey imputation (S2S) is widely used to estimate poverty when a target survey lacks the consumption or income aggregate needed for direct welfare measurement. This guidance note provides a practical framework for deciding when S2S is credible, when it is high risk, and when alternative approaches should be preferred. It emphasizes that S2S depends on two sets of conditions: comparable source and target survey populations with consistently measured predictors, and a welfare model that can be transported across surveys or periods. Drawing on background notes which provide simulation evidence, empirical work using Peruvian data, and lessons from a global S2S imputation challenge, the note illustrates why violations of these assumptions can generate large prediction errors, particularly during shocks or structural change. It offers a decision tree, diagnostic checks, validation practices, and documentation standards to help practitioners make assumptions explicit, triangulate predictions, and communicate uncertainty in operational poverty work.