Women’s Evolving Careers Helped Shrink the Gender Pay Gap

Between 2005-2015, men in the U.S. earned about $18 per hour in average wages, while women earned about $15 per hour (the first figure below), which was about 82% of the male wage (the second figure below). The same figure reveals that the gender wage gap was even wider in the 1970s, when women made only about 68% of the male wage. Note that the gender wage gap shrunk most rapidly between 1980 and 1995.

Wages by Gender

Gender Wage Gap: Female Wage as Share of Male Wage

SOURCES FOR BOTH FIGURES: U.S. Census, Current Population Survey and authors’ calculations.

NOTES FOR BOTH FIGURES: The first figure shows the average wage among working men and working women; an individual is considered to be working when he or she works a positive number of hours and earns at least $1,200 a year. The average wage is calculated as total labor earnings divided by total hours. The second figure shows the ratio of these wages.

Of course, equality requires that men and women performing equally well on the same job be compensated equally. So, why is there a gender wage gap, and why did it shrink so rapidly starting in the 1980s—the very decade that saw the most dramatic growth of overall wage inequality in the U.S.? Francine Blau and Lawrence Kahn argued in a 1997 article that it is partly a result of government antidiscrimination activity in the 1960s and 1970s, which encouraged women to train for and enter traditionally male fields.See their article “Swimming Upstream: Trends in the Gender Wage Differential in the 1980s.” Journal of Labor Economics, January 1997.

In the 1997 article, they further explained that the “increasing labor force commitment of women likely contributed to a reduction in statistical discrimination against them.” This means that as employers learned that women were becoming more committed to lifetime careers, they became more willing to hire them in traditionally male-dominated career tracks and compensate them accordingly.

In this post, we present a more detailed picture of how the gender wage gap evolved, cohort by cohort, and occupation by occupation. We show that, consistent with the hypothesis by Blau and Kahn, the closing of the gender wage gap is indeed closely associated with the closing of the gender gap in terms of labor market attachment, and that, over time, women became more likely to select into higher-paying occupations.

Women’s Labor Market Attachment Increased over Generations

First, we considered labor market attachment, cohort by cohort. In the next two figures, we plotted life-cycle labor supply profiles for cohorts who were 21-25 years old in 1940, 1945, and so forth until 1985. We measured average labor supply as average annual hours worked divided by 5,600. (The denominator is our proxy of total available productive time per year.)

The third figure in this post (see below) shows the labor supply profiles for women, while the fourth figure (see below) depicts them for men. While the male labor supply profiles look similar across cohorts, the labor supply profiles for women change dramatically, showing increasing labor market attachment over time.

The oldest cohort—women who were ages 21-25 in 1940—spent 9% of their productive time at work in 1940. By the time they were 31-35 (in 1950), a lot of them dropped out of the labor force or reduced their hours to raise children, implying that only 7% of their time was spent at work. By the age of 51-55 (in 1970), as child care responsibility eased off, labor supply peaked at 14%.

The labor supply profile looks very different for the youngest cohort—women who were 21-25 years old in 1985. These women began their careers by allocating nearly 20% of their time endowment to market work and did not drop out of the labor force to raise a family; their labor supply peaked around 23% of available time. This cohort of women is much more similar to the corresponding cohort of men whose labor supply peaked at only a slightly higher level of 32%.

Time Spent at Work: Women

Time Spent at Work: Men

SOURCES FOR BOTH FIGURES: U.S. Census, Current Population Survey and authors’ calculations.

NOTES FOR BOTH FIGURES: These figures depict the age profiles of labor supply for cohorts of men and women who were 21-25 years old in 1940, 1945, and so forth until 1985. Labor supply was measured as average annual hours worked divided by 5,600, our proxy of available productive time per year. Data for cohorts who have not yet reached the older age ranges are based on estimates created using average rates of time spent at work across the different age groups.

We see that, consistent with early antidiscrimination efforts taking place in the 1960s and 1970s, the most dramatic change in labor attachment is between the cohort of women who were 21-25 years old in 1965 and those who were five and 10 years older. A well-known result in economics proposed by Ben Porath states that, if a worker plans to work more in the future, he or she will allocate more current labor supply toward obtaining new skills and becoming more productive in the future.See Yoram Ben-Porath’s article “The Production of Human Capital and the Life Cycle of Earnings.” Journal of Political Economy, August 1967.” Journal of Labor Economics, January 1997. This will result in a steeper wage profile.

In a 2022 working paper, Oksana Leukhina and Guillaume Vandenbroucke use the labor supply profiles discussed here in a model that captures the Ben-Porath-type mechanism to show that the profiles imply that the wage gap dynamics are consistent with the data.

Greater Shares of Women Moved into Higher Wage Growth Occupations

How exactly workers allocate their on-the-job hours between simply doing the job and investing in their future productivity is not something that we can measure directly. However, we can look at the type of jobs held. Some jobs/occupations involve a lot of training early on and exhibit steep wage growth over the life cycle (e.g., physicians start as medical residents, spending most of their time training).

What we show next is that women indeed increasingly sorted into occupations that involve more training early on and more wage growth over time—such occupations also exhibit higher average wage levels.

The fifth and final figure (below) shows how the distribution of working women across 10 broadly defined occupations evolved over time. We are looking at women in the beginning of their careers (in their late 20s). The occupations (top to bottom) are ranked according to the average male wages in 2015, with technicians, managers and professionals having the highest average wages that year, and machine operators and those in farming and services ranking near the bottom of the pay distribution.

Female Workers in Each Occupation

SOURCES: U.S. Census, Current Population Survey and authors’ calculations.

NOTES: For each year, the figure represents the distribution of 26- to 30-year-old working women across 10 occupations. The occupations are ranked (top to bottom) by the average wage among 36- to 40-year-old men in 2015, with technicians, managers and professional workers having the highest-average wages that year.

In 1970, women entered into lower-paying occupations, working predominantly in administrative jobs. Over time, women clearly shifted toward better-paying occupations. For example, we see that the percentage of women working in the professional occupations went up from about 19% in 1970 to about 26% in 2015. The fraction of women working in the top four high-paying occupations increased from around 30% to 55%.

This evidence provides additional support for the main hypothesis that, as women increased their labor force attachment and as young women planned to work more over their life cycles, they increasingly entered career tracks that involved more training early on and higher wage growth over time, thereby narrowing the gender pay gap.

Notes

1. See their article “Swimming Upstream: Trends in the Gender Wage Differential in the 1980s.” Journal of Labor Economics, January 1997.
2. See Yoram Ben-Porath’s article “The Production of Human Capital and the Life Cycle of Earnings.” Journal of Political Economy, August 1967.