Spatial Analysis of Poverty: The case of Peru

The concept of Multidimensional Poverty traditionally was used for comparative analysis across regions or countries. Nevertheless, in this paper, we use the concept of Multidimensional Poverty calculated for each Peruvian region and analyze the spatial patterns and spatial autocorrelation observed across the country, and, later, identify the spatial spillovers in poverty across the country. We find evidence of spatial autocorrelation across the regions which are statistically significant across models; in other words, it means poverty has spatial effects. Additionally, we find strong and significant evidence of spatial spillovers originated in the error terms rather than the endogenous variable, which has an unstable effect. Finally, the set of covariates we use in our regressions are statistically significant and stable across the models