As estimativas de pobreza no Brasil estão superestimadas ?

Abstract

This article aims to verify the possibility of overestimation in the calculation of the poverty rate, proportion of poor people, in Brazil, as compared to other results found in the literature. For doing this, two simulations are accomplished, where the first one considers the per capita household income of the Brazilian states accounted for both the urban and rural areas, according to the division adopted by the National Household Survey - PNAD (2002), which is taken as reference in the analysis. In the second simulation, also at state level, it is analyzed the poverty rate for the urban population only. Distinct poverty lines were adopted in each simulation, so that it allowed comparing the two estimates with the results in general accepted by the literature, as the one of Rocha (2004) and IPEADATA. The methodology was conducted under rigid statistical criteria in order to guide to an optimum choice of the density function with the best fitness of the income distribution of the population. Integration of such function to the limit of poverty line should provide the correct poverty rate. The inference on the results indicates that there have been overestimations in great magnitude of the poverty rate as pointed out by other studies. Also, the rate is negatively and positively correlated to PIB and income distribution, respectively, of the state.