This article examines skin color discrimination in two Brazilian labor markets using a field experimental approach. Fictitious resumes including photographs of job candidates were randomly assigned one skin color category via photo manipulation and submitted to entry-level job openings. In addition to assessing the extent of skin color discrimination, this article adopts an intersectional framework to examine how the effect of skin color in employment is moderated by class status and varies by gender. I found mixed results about the role of skin color in predicting the employment outcomes at the initial stages of the hiring process. Results from logistic regression and Linear Probability Models show that skin color is a weak predictor of hiring outcomes (e.g. receiving a callback from employers) among male applicants and for female applicants with brown skin. However, I find strong evidence that having dark skin is causally associated with hiring outcomes among female applicants. I also found that having a higher-class status erases skin color differences, thus identifying a potential mechanism that mitigates the effects of skin color in hiring.

本文使用现场实验方法研究了两个巴西劳动力市场中的肤色歧视。虚假简历（包括应聘者的照片）通过照片处理被随机分配一种肤色类别，并提交给入门级职位空缺。除了评估肤色歧视的程度外，本文还采用交叉框架来研究肤色在就业中的影响如何通过阶级地位和性别而变化。在招聘过程的初始阶段，我发现肤色在预测就业结果方面的作用参差不齐。Logistic回归和线性概率模型的结果表明，肤色对招聘结果的预测能力较弱（例如，男性申请人之间以及棕色皮肤的女性申请人中收到雇主的回覆。但是，我发现有力的证据表明，皮肤黝黑与女性申请人的招聘结果有因果关系。我还发现拥有较高级别的身份可以消除肤色差异，从而确定一种减轻肤色影响的潜在机制。