The Paradox of within-Country Convergence: Evidence from Colombia

Karina Acosta, Cornell University
María Dávalos, The World Bank
Sandra Segovia, The World Bank

The aggregate trends in economic growth and poverty reduction obscure the diverse development trajectories occurring at the local level. This study produces a unique dataset comprising poverty, household income, and GDP per capita maps, covering 1,122 municipalities in Colombia—the most granular level of administrative geographic disaggregation—over a 13-year period, utilizing small area estimation techniques to test the hypothesis of convergence within the country. The study draws on an extensive range of data from the 2005 and 2018 censuses, household surveys, and population projections. Furthermore, the analysis decomposes poverty convergence elasticity to assess whether the observed income convergence has been accompanied by a parallel convergence in poverty levels. The findings reveal evidence of income convergence across municipalities, whereas poverty trends exhibit a contrasting divergence. This outcome is based on estimates that account for population size; however, when population is excluded from the analysis, the results suggest apparent convergence. These patterns are likely influenced by worsening poverty conditions in wealthier, urban areas, as opposed to the improvements seen in poorer, rural regions.

Keywords: Population and Development, Small area estimation, Data and Methods, Econometrics

See extended abstract.