Is regional poverty converging across Indonesian districts? A distribution dynamics and spatial econometric approach

Using a novel district-level dataset of three poverty indicators and related determinants, this paper attempts to examine the convergence process in terms of poverty reduction across 514 districts over the 2010-2018 period. The main findings of this study are fourfold. First, results of both distributional convergence and spatial convergence analyses suggest that poverty convergence takes place in all three poverty indicators. Second, there is an overall clustering pattern and a persistent West-East polarization in terms of poverty indicators. In addition, decreasing patterns of poverty inequality are significantly associated with increasing patterns of spatial dependence. Third, spatial effects are significantly low in accelerating convergence rates in poverty reduction. However, substantial differences in the speed of convergence among three poverty indicators is apparent. Under specifications of the Spatial Durbin Model as the best model, convergence regression in poverty rate model appears to generate the longest half-life for the catching-up process in poverty reduction across districts. Fourth, regression results from the spatial convergence framework suggest that mean years of schooling tend to contribute the largest effect to the catching-up process in poverty reduction.