Race, education and occupation patterns in the relationship between assortative marriage and earnings inequality in Brazil

Melissa Brandão, Universidade Federal de Minas Gerais (UFMG)
Ana Hermeto, Cedeplar, UFMG

Some theoretical perspectives suggest that assortative mating marital education should be higher in more unequal contexts, as Brazil. Our objective is to investigate how patterns of assortative mating and marital family arrangements have changed over the last three decades in Brazil and interpret these changes in terms of their economic and social implications, as expressed in income inequality. Marital assortative mating estimates based on individual level of education took into account differences in educational distributions of men and women, who also changed differentially over time. We also explored the changing patterns of marital assortative mating in terms of ages and cohorts over time. Data source is the microdata from PNAD, IBGE for 1986 to 2012. Log-linear and log-multiplicative methods are estimated in order to compare patterns of racial and educational assortative mating over time in Brazil. It is further investigated to what extent the factors that led to changes in the patterns of assortative mating and marital family arrangements have contributed to changes in income distribution between 1986 and 2012, using a modified version of a semi-parametric decomposition method, which allows the construction of counterfactual densities, allowing to isolate the effects on income inequality. Groups considered in the decomposition of income inequality are related to: labor force participation, family characteristics, marital assortative mating patterns, educational levels and demographic characteristics. We found substantial variation in the strength of specific barriers to educational intermarriage, and a close association between these barriers and the earnings gaps across educational categories. Educational marital sorting is remarkably symmetric across gender in spite of the arguably different resources that men and women bring to the union. We highlight the limitations of using single aggregate measures of spousal educational resemblance to capture variation in assortative mating and its relationship with socioeconomic inequality