How can we measure multidimensional poverty when the data we need are scattered across different surveys?
A
@WorldBankGroup Data Blog by Ben Brunckhorst, Minh Cong Nguyen , Nishant Yonzan, Hai-Anh H. Dang, and
@ChristophLakner explores a promising approach: combining indicators from separate datasets to estimate how monetary poverty, education, and infrastructure deprivations overlap.
The challenge is significant. While individual indicators used in the World Bank’s Multidimensional Poverty Measure cover at least 85% of the global population, complete data are available to estimate the full measure for only about two-thirds of the world.
Using more than 570 surveys spanning 1989–2024, the authors show that data fusion can produce credible estimates—particularly when datasets share some overlapping indicators. The approach could help extend multidimensional poverty measurement to countries and years where complete survey data are unavailable, and potentially incorporate additional dimensions such as health and security.
Better poverty measurement depends not only on collecting more data, but also on finding rigorous ways to make better use of the data that already exist.
Read it here:
blogs.worldbank.org/en/opend…