In 1999 a working mother in Brazil with 11 years of schooling earned 3.05 reais per hour while her male counterpart, with the same qualifications, earned an average salary of 4.52 reais. The case of Brazil is not an exception in Latin America. Although in some countries, such as Bolivia, the wage gender gap is not so evident, men receive better salaries than women for the same job. The good news is that empirical evidence suggests that the gap is slowly closing; the bad news is that the narrowing of the discrimination gap has been slower in Latin America than in industrial countries and its size may have been underestimated.
Conventional estimates of labor market gender discrimination in Latin America in the 1990s have put the unexplained portion of the male-female earnings gap at around 10-25 percent, depending on the country and time period selected. However, the preliminary results of a new IDB study indicate that these estimates understate the true degree of the unexplained gender gap by not fully accounting for selection of women into the labor force.
The hypothesis of this study is that traditionally the male-female earnings gap has been calculated comparing the salaries of men and women with the same education doing similar jobs, without taking into consideration that women who work outside the home are not - in contrast with men - average women.
To put it simply, women who join the workforce do not represent a random draw of all women. Rather, they have different characteristics than stay-at-home women, and many of these characteristics can be expected to affect their labor market productivity and earnings. The implication of this non-random selection is that earnings comparisons between men and women, which are based on observed earnings only for women who work, will likely misrepresent the true magnitude of labor market discrimination.
Using household survey data from Bolivia, Brazil and Nicaragua, the study finds preliminary evidence that the unexplained portion of the gender gap is significantly larger than previously estimated, at least in the case of married women.
The results of the study are driven by a new methodology that accounts for women's participation in the labor force. Namely, since teenage girls substitute for their mothers in performing household chores, the study uses child gender to predict women's labor force participation. “Male-female wage gaps based on these predictions indicate that the unexplained portion of the earnings gap increases by roughly 50 percent for Brazil, and more than doubles for Bolivia and Nicaragua,” says Yuri Soares, who is with the IDB's Office of Evaluation and author of the study.