Texto para Discussão, Instituto de Pesquisa Econômica Aplicada (IPEA) 1517
Estimates based on household surveys often differ considerably from administrative records. In the PNADs, for instance, the estimated number of beneficiaries of the Programa Bolsa Família (PBF) and of the Benefício de Prestação Continuada (BPC) are always lower than the official figures. This paper presents a simple methodology, based on the sampling design of household surveys, to explain these differences, decomposing them in three terms: the representativeness bias (due to selection of locales, census tracts or municipalities for the survey), the selection bias (due to selection issues in the chosen locales) and the interaction among them. The application of this methodology to the PBF and the BPC reveals that, regarding the former, the representativeness bias is accountable for a good part of the problem: the selection of the municipalities to be surveyed is responsible for 40% of the difference between official records and the PNAD. In the case of the BPC, the representativeness bias tends to act in the opposite direction. Thus, the selection bias is entirely responsible for the observed difference. Also, the erroneous reporting of the BPC as a Social Security benefit in the PNAD seems to occur mostly in the years prior to 2004 and in any case does not explain by itself the low number of beneficiaries identified in the PNAD.