Identificación de la pobreza en Colombia

This study examines poverty identification in Colombia by comparing the multidimensional poverty index and the social program beneficiary identification system. The results are analyzed in Colombia and the City of Medellín using measures of agreement and a contingency table. A statistical learning model is employed to estimate the classification of the social program beneficiary identification system, achieving a balanced accuracy of 78%. Qualitative analysis reveals the multidimensional poverty of four families not identified as poor by the beneficiary identification system. These findings highlight a low and insignificant agreement between the two measuring systems. The results contribute to improving the definition and identification of poverty in the country, supporting public policy design.