Brazil in black and white? Race categories, the census, and the study of inequality
The recent introduction of race-targeted affirmative action policies in Brazil unleashed a contentious national debate over whether, how, and toward what ends the government should engage in the racial classification of citizens. In its general contours, this national conversation in Brazil echoes analogous discussions in a growing number of countries around the globe. Voices of opposition to the Brazilian government’s long-established practice of official racial classification raise principled arguments in defence of the liberal ideal of individual equality before the law. They also question the morality
and scientific legitimacy of government use of racial categories, pointing to ignominious historical examples of abuse enabled by official racial classification and citing the unscientific nature of racial categories per se (Fry et al. 2007). Proponents of government initiatives to draw racial or ethnic distinctions among citizens in certain contexts counter with principled and pragmatic arguments of their own. They note that official racial categories can capture social distinctions without implying that ‘race’ is valid as a biological category. They also argue that official racial classification can facilitate government efforts to promote true equality of citizens through reparation or remediation of historical injustices to racially defined minority populations (Manifesto 2008). As in other contexts, the precise configuration of the Brazilian
debate over government use of racial categories is shaped by the country’s particular historical approach to dealing with racial and ethnic difference. The Brazilian state has a long history of recognizing racial and colour distinctions within the population in censuses as in other administrative domains (Nobles 2000; Loveman 2009). In stark contrast to the United States, however, the explicit use of such categories in public policy or law was rare. Though some have argued aggressively against government recognition of racial or colour distinctions, the crux of debate in Brazil does not appear to pivot on whether the state should classify citizens by race or colour at all, but on the specific categories it should use to do so, and the legitimacy of public policies that attach material consequences to categorical membership (Bailey and Peria 2010). For social scientists, policy makers, and activists, one key stake in
this broader debate is the format of the question used by Brazil’s census agency, the Instituto Brasileiro de Geografia e Estatı´stica (IBGE), to collect official race/colour statistics. The IBGE has always included a ‘mixed-race’ or ‘brown’ category as one option when race/ colour is asked in the census. This corresponds with Brazil’s selfdefinition over the course of the twentieth century as a nation that is, in its ‘essence’, mixed. Critics contend that the official distinction between ‘black’ and ‘mixed’ Brazilians on the census perpetuates an ideological myth of a fluid racial order; they argue that in reality Brazilian racial dynamics are essentially binary, and official categories should reflect this state of affairs. These debates also extend to the classification schemes employed for the administration of racetargeted policy. Categories used to collect racial/colour statistics have wide-reaching
repercussions, determining which lines of distinction become socially visible and amenable to statistical analysis and policy intervention. This article examines how a change in the classification scheme used to collect official statistics could affect social scientific understanding of
racial dynamics in Brazil. If the ‘mixed-race’ category on the national census were eliminated, would the Brazilian population end up looking lighter or darker? Would racial inequality appear attenuated or more severe? We begin with a critical overview of the contemporary debate over
racial classification in Brazil. We identify the primary criticisms of the classification scheme used in Brazil’s national census, and point to sources of momentum toward adopting a binary lens for social analysis and public policy. Drawing on a nationally representative dataset that includes race/colour questions in multiple formats, we then evaluate how a switch to a binary classification scheme on the census would likely affect social scientific understanding of racial dynamics in Brazil. Our empirical analysis examines how a change to a binary classification scheme for data collection affects: (1) the descriptive picture of Brazil’s racial composition; and (2) statistical estimates of income inequality between and within racial categories. We conclude with a brief discussion of the implications of our findings for contemporary debates over the state’s collection and use of racial statistics.