Inconsistencies and consequences of the race variable for the measurement of inequalities

• Autor(a) principal: Muniz, Jeronimo Oliveira
• Data de Publicação: 2016
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Civitas - Revista de Ciências Sociais (Porto Alegre. Online)
• Texto Completo: https://revistaseletronicas.pucrs.br/index.php/civitas/article/view/23097
• topic: Schooling. Skin color. Consumption. Inequality. Racial distribution.; Escolaridade. Cor. Consumo. Desigualdade. Distribuição racial; Civitas - Revista de Ciências Sociais

“It is not possible to separate people by race in Brazil” (Fry et al., 2007). This common and long-standing argument is part of the Brazilian racial classification system at least since the early 20th century. The inability to identify target populations has also been a reason to be skeptical about the success of affirmative action programs implemented to mitigate persistent racial inequalities in Brazil. This article investigates the association between inconsistencies in racial classification and the level and composition of inequalities in education and consumption. Using unique data from the 2002 Brazilian Social Survey (PESB), we show that individuals whose race/color was inconsistently reported account for at least half of total inequality and more than a third of inter-racial inequality. After excluding inconsistently classified individuals from our sample, however, the share of inter-racial inequalities does not substantially change, demonstrating that racial uncertainty has little influence on racial inequality in Brazil.