최신자료
Fusing Data Sources to Measure Multidimensional Poverty
Multidimensional poverty metrics offer a more complete picture of poverty than monetary poverty, yet their adoption has been limited by data constraints. Often not all dimensions in a multidimensional measure are observed in the same survey, leading to countries being excluded from global measures. This study introduces a novel data fusion approach to estimate multidimensional poverty metrics by combining summary statistics from multiple data sources. The proposed method is validated by simulating typical missing data scenarios using 571 household surveys from 112 countries (1989-2024) to estimate the World Bank’s Multidimensional Poverty Measure. The fusion method accurately predicts various multidimensional poverty metrics across these scenarios, even in the presence of simulated sample bias reflecting differences in survey design. It enables estimates of multidimensional poverty for a greater number of countries, gap-filling in time series, and the inclusion of previously unavailable indicators and dimensions in global multidimensional poverty measures.
많이 본 자료
