UC Berkeley

In this dissertation, I examine how skin color, class status, and gender shape labor market outcomes in Brazil. In spite of overwhelming evidence that racial inequality pervades almost every aspect of Brazilian life, there is still much disagreement among researchers and policy-makers as to the root causes of this form of inequality. According to one school of thought, racial inequality in Brazil is not the product of active contemporary discrimination, but related to the legacies of slavery as well as class-based inequalities, such as differential access to education. Others, however, argue that in spite of class-based factors, skin color has an independent effect on shaping racial inequalities. This dissertation, adjudicates between these competing explanations. Moreover, I examine whether perceived class status moderates the effects of skin color in shaping labor market outcomes. However, I take it two steps further. First, I test for the possibility that the effect of race is moderated by gender. Second, I test the validity of the “mulatto-escape hatch” hypothesis, which predicts that lighter skinned Afro-descendents (browns) suffer less discrimination than their darker skinned counterparts (blacks).

In this dissertation, I use a field-experiment to adjudicate between these competing theoretical perspectives and related contingencies. Whereas existing research on racial inequality in Brazil relies on indirect methods to estimate racial discrimination (e.g. “wage regression” analysis), I deploy a field experimental methodology to directly test for discrimination. After identifying job vacancies in newspapers, I sent 1,200 fictitious resumes to employers of entry-level jobs in two major metropolitan areas in Brazil: São Paulo and Rio de Janeiro. I signaled the skin color of job candidates through photographs attached on resumes, and class status through class-specific first and last names. There were several important findings from the field experiment. First, I found that skin color does shape labor market outcomes, but it is moderated by class status. Second, I found that the effect of skin color is also gendered. While skin color did not affect the likelihood of receiving a callback among men, it did so among women, in particular women with “poor” class status and with dark skin. Moreover, I found that the “mulatto escape hatch” hypothesis is also contingent on the intersection of skin color, class status, and gender as well as labor market contexts. The implications of these findings for the literature on racial inequality in Brazil are considered.