The Gender Pay Gap: What’s Actually Closing It, and What Isn’t

Every spring, the U.S. Census Bureau marks a date it calls Equal Pay Day — the point in the following year by which the average woman working full-time has finally earned what the average man earned by December 31 of the previous one. In 2026, that day fell on March 26. The gap behind it has narrowed substantially since the government first started tracking it in the early 1970s, when women earned roughly 57 cents for every dollar men earned. By 2024, that figure had climbed to about 83 cents among full-time, year-round workers. Progress, by any measure — and also, depending on how you read the last two decades of data, a story of progress that has mostly stalled.

What Is the Gender Pay Gap, Precisely?

The gender pay gap is the difference between the median earnings of men and women, typically expressed as the percentage of men’s earnings that women’s earnings represent. The U.S. Census Bureau’s standard figure compares full-time, year-round workers using American Community Survey data; by that measure, women earned about 83% of what men earned in 2024. A separate, broader figure from the Pew Research Center, which includes part-time workers using Current Population Survey data, put women’s earnings at about 85% of men’s the same year. Both are legitimate, commonly cited numbers; they differ because they measure slightly different populations, which is itself a useful reminder that “the gender pay gap” is not one single, universally agreed-upon statistic.

Specialty Digest’s companion explainer on the gender pay gap covers the basics; this piece goes further into how the number has moved over time, what actually drives it, and where the real disagreement lies.

How Much Has the Gap Actually Closed Over Time?

The long-term trend is real and large. Per Pew Research Center analysis of Census Bureau data, the overall gap went from a 35-cent difference in 1982 to a 15-cent difference in 2024. But that progress has been uneven. Women earned 81% of men’s earnings back in 2003 — meaning the overall ratio moved only a few percentage points across the following two decades, after moving much faster in the prior two. The pattern looks different for younger workers: among employees ages 25 to 34, the gap shrank from 26 cents in 1982 to just 5 cents in 2024, suggesting that whatever is keeping the overall gap from closing further is concentrated later in careers — not at the point of hiring a new graduate.

What Actually Explains the Gap?

Economists who study this generally sort the explanation into two buckets: factors that are measurable and factors that aren’t. On the measurable side, differences in educational attainment, the occupations and industries men and women are concentrated in, and years of work experience account for a substantial share of the overall gap — and closing those gaps over past decades is what drove most of the long-term narrowing described above. Occupational segregation — women remaining overrepresented in lower-paying fields and underrepresented in higher-paying ones relative to their share of the overall workforce — continues to contribute measurably to the remaining gap. So does what researchers call the motherhood penalty: a body of research finds that becoming a mother is associated with reduced earnings for women, while becoming a father is, on average, associated with a modest earnings increase for men — a divergence that shows up clearly once researchers track individual earnings trajectories around the birth of a child, and one explanation for why the gap widens later in careers rather than at entry level.

What’s Left Over Once You Account for All of That?

This is the contested part. After researchers control for education, occupation, industry, and experience, a residual gap remains that is harder to explain with observable data — and economists disagree sharply about how much of that residual reflects outright discrimination versus unmeasured factors like negotiation differences, hours and schedule flexibility, or career interruptions not fully captured in the data. The Census Bureau’s own 2026 Equal Pay Day analysis found the widest occupational gaps concentrated in physically demanding fields — Natural Resources, Construction, and Maintenance occupations showed women earning about 74% of men’s pay; Production, Transportation, and Material Moving occupations showed about 75% — though the agency noted the difference in some of these categories was not statistically significant given smaller sample sizes. Public opinion tracks this ambiguity, too: in a 2022 Pew survey, half of U.S. adults named unequal treatment by employers as a major reason for the gap, while 42% pointed to women’s own work-family choices and 34% cited women being concentrated in lower-paying jobs — with women themselves (61%) far more likely than men (37%) to cite discrimination as a major driver.

Why Did the Gap Widen Slightly in Recent Years?

After decades of narrowing, several analyses of 2023-2024 Census data found the gap ticking modestly wider for two consecutive years — a reversal researchers have described as unusual against the longer historical trend. The explanations offered center on uneven patterns in post-pandemic labor-force participation and hours worked across industries, rather than any single new cause, and researchers studying the shift have been careful to note that a two-year wobble in an annual government survey is not the same as a confirmed reversal of a forty-year trend. It is, at minimum, a sign that the remaining gap is not on an automatic path to zero.

Evidence Strength

Well-supported: That a measurable gender pay gap exists in the United States, that it has narrowed substantially since the 1970s, and that occupational segregation and the motherhood penalty each contribute measurably to the remaining difference are all well established through decades of consistent government survey data and peer-reviewed labor economics research.

Contested: How much of the residual gap — the portion left after accounting for education, occupation, industry, and experience — reflects direct discrimination versus unmeasured factors like negotiation, scheduling flexibility, or self-selection into roles is genuinely disputed among economists, and is not resolved by the topline percentage figures that get the most public attention.

Key Takeaways

• The U.S. gender pay gap has narrowed from a 35-cent difference in 1982 to a 15-cent difference in 2024 among full-time workers, per Census Bureau data analyzed by Pew Research Center.
• Different methodologies produce different headline numbers: the Census Bureau’s full-time, year-round figure put women’s 2024 earnings at about 83% of men’s; Pew’s broader figure including part-time workers put it at about 85%.
• Progress has been uneven: the gap barely moved between 2003 and the early 2020s after narrowing faster in prior decades, and ticked slightly wider again in 2023 and 2024.
• The gap is much smaller for younger workers — a 5-cent gap for ages 25-34 in 2024, versus 26 cents for that age group in 1982 — suggesting it widens later in careers rather than at hiring.
• Occupational segregation and the “motherhood penalty,” where having children is associated with reduced earnings for women but often increased earnings for men, are both measurable, well-documented contributors.
• A residual gap remains even after accounting for education, occupation, and experience; economists are genuinely divided on how much of it reflects discrimination versus unmeasured factors.
• Public opinion splits along similar lines to the research: more women than men attribute the gap primarily to unequal treatment by employers rather than to choices or occupational differences.

Frequently Asked Questions

What is the current gender pay gap in the U.S.?
By the Census Bureau’s full-time, year-round measure, women earned about 83% of what men earned in 2024. A broader Pew Research Center figure that includes part-time workers puts it at about 85%. Both are accurate; they measure slightly different groups.

Has the gender pay gap actually been closing?
Yes, substantially, over the long run — from a 35-cent gap in 1982 to a 15-cent gap in 2024. But progress slowed dramatically after the early 2000s, and ticked slightly backward in 2023 and 2024.

Is the gap the same at every age and in every job?
No. It’s much smaller for workers in their twenties and early thirties than for the workforce overall, and it varies significantly by occupation — wider in some physically demanding fields and narrower elsewhere.

Is the pay gap caused by discrimination?
Partly, according to most economists who study it, but the measurable, provable share is smaller than the headline percentage. Education, occupational choice, and work experience explain a large portion of the gap; what’s left over after accounting for those factors is where the genuine disagreement about discrimination’s role lives.

What is the “motherhood penalty”?
It’s the research finding that having children is associated with reduced earnings for women on average, while having children is associated with a modest earnings increase for men — a divergence that shows up clearly in earnings data tracked around the birth of a child and that helps explain why the gap widens later in careers.

Does comparing men’s and women’s pay mean comparing identical jobs?
Not in the headline figures. Those compare median earnings across the whole workforce, not pay for the exact same role, which is why occupational segregation — which jobs men and women are concentrated in — is such a large part of the explanation.

The Bottom Line

The gender pay gap is one of the rare economic statistics where the topline number and the real story diverge the most. The topline — 83 cents to 85 cents on the dollar — is accurate, but it compresses four decades of genuine, hard-won progress, a stall over the last twenty years, and a two-year uptick that may or may not be a blip, into a single figure. The more useful question than “what’s the number” is “what’s still driving it”: occupational concentration and the motherhood penalty are well-documented and measurable; the share left over for discrimination and other unmeasured factors is real but genuinely contested. Both can be true, and neither should be flattened into the other.