Resumen

HERNANDEZ FRANCO, Daniel; OROZCO CORONA, Mónica  y  VAZQUEZ BAEZ, Sirenia. Métodos de focalización en la política social en México: Un estudio comparativo. Econ. mex. Nueva época [online]. 2008, vol.17, n.1, pp.101-128. ISSN 1665-2045.

This paper tries to describe and analyze some of the tools used in social policy to target poverty reduction programs in Mexico. Targeting strategy programs center public actions to a reasonable well defined population or territory, in order to achieve a more focused efficiency in the resource allocation stage. Using data from the National Household Survey of Income and Expenditure (ENIGH) 2002, three econometric models are used to identify the target population: Discriminant Analysis, a Logit Model and a Hierarchical Logit Model. The results show that Discriminant Analysis, which is the main tool used by Sedesol to identify individuals in poverty conditions, has the lowest undercoverage error for the three measures of the Foster-Greer-Thorbecke poverty index. The result leads to conclude that this method proves to be the most efficient in identifying population in poverty conditions, given the available information.

Palabras llave : targeting; discriminant analysis; hierarchical models; Logit model; leakage error; undercoverage error.