Artigo publicado na revista da Academia Brasileira de Letras. 

Resumo: A desigualdade de renda no Brasil é ainda maior do que o imaginado. O índice de Gini resultado da imputação aos microdados da Pesquisa Nacional por Amostra de Domicílios Contínua (PNADC) interpolados de tabulações padronizadas do Imposto de Renda da Pessoa Física (IRPF), chegou a 0.7068 em 2020, bem acima dos 0,6013 calculados a partir da apenas da PNADC. O IRPF consegue captar melhor a renda proveniente do ganho de capital, como os lucros no mercado financeiro ou das empresas, por isso traz mais realismo para o rendimento dos mais ricos.

This paper develops a new imputation methodology applied to missing income values in PNAD. PNAD was the main Brazilian household survey but with no income imputation. The imputation process starts fitting regression models considering complex sampling design applied to different income sources. Later this procedure is combined with stochastic methods. In 2015, 2.5% of the sample had per capita incomes imputed resulting in slightly higher levels of inequality due to higher frequencies of imputations in the two extremes of the distribution.

Abstract: This paper evaluates the impacts of combining household surveys with income tax return files, in terms of growth, inequality, and social welfare in Brazil from 2007 to 2015. This exercise holds the promise of adding more realistic top income values to traditional surveys. While the previous literature focused on the impacts of these data combination exercises on income inequality, we assess their cumulative welfare implications.

Abstract: This paper assesses causes and consequences of fiscal redistribution in Brazil. The framework proposed allows evaluating in an integrated manner the impacts of government-sponsored actions in inequality and mean income changes on social welfare, addressing both static and dynamic implications. To the best of our knowledge, this is the first microsimulation attempt to gauge actual fiscal policy redistribution changes over time in Brazil. The study develops an empirical methodology that allows consistent comparisons among the years 1995, 2003, 2009, and 2015.