Labor market inequality and the changing life cycle profile of male and female wages

Summary

We provide new estimates of how the gender wage gap in the US varies over the working life for cohorts of workers born between the 1920s and early 1990s. Using 43 years of data from the Current Population Survey Annual Social and Economic Supplement, our results show striking convergence in gender wage gaps across cohorts at a given age for those born before the 1960s. At the same time, we find little convergence over the working life among college educated women born after 1960 in the upper half of the wage distribution, and indeed some divergence among high-educated, high-wage Millennials. These patterns help account for the observed decline in the overall gender wage gap from the 1970s to the 1990s, but little progress over the last quarter century. We explore possible mechanisms for these results, and find that time factors such as technological change that increased returns to education, and legislative, judicial, and cultural changes that removed barriers to women’s work eliminated much of the cross-cohort differences in gender wage gaps. However, these changes were more neutral across gender starting with the 1960s birth cohort. Just as these later cohorts of women were entering the labor force, they also started to delay fertility, with the peak age of raising a child increasing by at least 5 years between the 1940s and 1960s birth cohorts. This delayed child rearing likely pushed gendered “child penalties” of wages later in the working careers of women, which helps account for the stalling of the gender wage gap in recent cohorts. Methodologically, our approach corrects for the fact that people who decide to work may be different from those who choose not to, in both observable and unobservable characteristics. In addition, our empirical strategy allows this labor market participation decision, and the associated wages, to potentially differ across gender and education groups when they are identical in terms of other demographic characteristics, such as age and marital status.