Aplicação da teoria do valor extremo e suavização por núcleo estocástico na distribuição de renda dos pobres no Estado do Ceará

Abstract

The research of estimating of poverty in Brazil have been concentrated using the tools of statistical inference or the inefficient use of some ad hoc distributions, or through studies of convergence as the b-convergence and s-convergence. This work contributed to a discussion of different methods of non-parametric statistical inference, with the aim of estimating the evolution of the density of the poor of Ceará by smoothing by stochastic kernel (Kernel Density), based on data collected by the National Survey Sample of Households - PNAD (2001, 2003, 2005 and 2007) can be concluded that the average family income per capita has been increasing during the analysis. There was a contrast between two methodologies for estimating the rate of poverty in Ceará concerning the year 2007, which are: The traditional method (using Discrete Uniform formulation) and an application based on the theory of extreme values (TEV) that is commonly applied to VaR (Value at Risk) of financial assets. Contrast the results we can conclude that there is strong indication that there is underestimation of poverty rates by using the traditional methodology.