Prevalence of international migration: an alternative for small area estimation

TL;DR

This paper proposes a new small area estimation method for international migration prevalence in Colombian municipalities, utilizing Fay-Herriot models with census and survey data, resulting in more reliable and consistent estimates.

Contribution

It introduces an alternative estimation procedure using empirical best linear unbiased predictors within a Fay-Herriot framework for small area international migration prevalence.

Findings

Estimates are consistent with demographic data.

Coefficient of variation below 20% for key municipalities.

Method produces reliable prevalence estimates.

Abstract

This paper introduces an alternative procedure for estimating the prevalence of international migration at the municipal level in Colombia. The new methodology uses the empirical best linear unbiased predictor based on a Fay-Herriot model with target and auxiliary variables available from census studies and from the Demographic and Health Survey. The proposed alternative produces prevalence estimates which are consistent with sample sizes and demographic dynamics in Colombia. Additionally, the estimated coefficients of variation are lower than 20% for municipalities and large demographically-relevant capital cities and therefore estimates may be considered as reliable.