The rainbow reality: income difference and discrimination based on sexual orientation and occupations

Purpose - The article aimed to identify and analyze the income differences across the income distribution between homosexuals and heterosexuals by occupational groups in Brazil. Design/methodology/approach - PNAD-C microdata was used in two periods (2013–2015 and 2016 to 2019), highlighting the different economic, social and political contexts in Brazil. Recentered influence function and quantile income decomposition were estimated to verify the difference and income discrimination according to the guideline the worker’s sexuality. Findings - For some cases homosexual workers earn more and in others, homosexuals earn less than heterosexuals. The differences in remuneration according to sexual orientation were smaller in positions that demand low qualification and competence. The quantile income decomposition between 2013 and 2015 revealed the positive effect of discrimination was the generator of income disparities between homosexuals and heterosexuals, with greater impact for the 10th and 90th quantiles and on groups that require small levels of complexity and education. Between 2016 and 2019, there was the presence of both effects, but the explained effect was the promoter of wage disparities in most occupational groups. Research limitations/implications - Given the dataset, it was only possible to work with proxies of homosexual couples. Originality/value - There has been little research linking the themes of discrimination based on sexual orientation and occupational groups, and so far, it does not exist similar in Latin America. This study found that sexual orientation influences remuneration according to the occupational group that the worker belongs to, affecting income and, consequently, occupational choice. This connection of issues will contribute to new insights into discrimination based on sexual orientation, as well as more effective public policies aimed at reducing discrimination against homosexuals.