UC San Diego

Brazil is Latin America's largest democracy and home to the largest African descendant population of any country outside of Africa. Despite comprising a majority of the Brazilian population, though, Afro-Brazilians hold less than 10 percent of the elected positions in Brazil's Congress. In this dissertation, I seek to answer two central questions about the political marginalization of African descendants. First, why do Afro-Brazilians not attain representation in Brazil's Congress commensurate with their numerical strength? And second, how do racial disparities in electoral outcomes affect the representation of Afro-Brazilian' policy interests?

To answer the two main questions, it is first necessary to examine how Brazilian elites racially identify themselves. I argue the absence of official rules for racial group membership and the political salience of racial group membership encourage political candidates strategically present themselves as members of the racial group that maximizes their chances of winning public office. Consistent with this argument, I find using an original dataset assembled from official electoral documents that over a quarter of Brazilian politicians who ran for office in 2014 and 2016 changed their self-reported race from one election to the next. My analysis reveals these changes are by no means random, but instead reflect strategic electoral calculations. This suggest that candidates view their racial groups as electoral vehicles and indicates that measures of self-identified race are likely endogenous to electoral outcomes in Brazil.

Building on evidence that politicians strategically present themselves, I use data from the 2014 congressional elections and an original measure of candidate ascribed race to explore why phenotypically nonwhite candidates are rarely elected to public office. I show that party opportunity structures and socioeconomic differences between white and nonwhite candidates hinder the latter from winning office. Nevertheless, when I control for theoretically important differences between phenotypically white and nonwhite candidates using regression methods, I continue to find that nonwhites receive significantly fewer votes their their white competitors. This finding suggests that voters discriminate against candidates that they perceive as Afro-Brazilian.

With respect to the second question, I demonstrate that racial disparities in electoral outcomes have substantial policy implications. Through the first quantitative analysis to explore the relationship between legislator race and agenda-setting behavior in Brazil, I show that Afro-Brazilian legislators are more likely to introduce legislation that reflect the economic, social, and political preferences of nonwhite citizens. Nevertheless, Afro-Brazilian legislators limited numbers constrain their ability to enact the legislation their propose. These results suggest that Afro-Brazilians are disproportionately on the losing side of Brazilian democracy and raises concerns about the nation's future.

By offering a systematic account of how racial group membership shapes political outcomes in Brazil, this dissertation confirms long-held suspicions about the political significance of race in Brazil. It presents a bleak picture of Afro-Brazilian political marginalization and identifies a series of challenges to achieving racial inequality. The future of Brazilian democracy and the country's unity, rests on the ability of its leaders to acknowledge and redress the racial disparities that are reflected and perpetuated by Afro-Brazilian political underrepresentation.