Alerts
-
3 minutes, 11 seconds
The U.S. gender pay gap narrowed in 2025 according to new Census Bureau data, but the Bureau of Labor Statistics reported earlier this year that it widened. Both agencies are measuring real things; they simply measure them differently. Here is what each number means, why they diverge, and which figures to trust for the long view.
The Census Bureau reported Tuesday that women earned 83.9 cents for every dollar men earned in 2025, up from 80.6 cents in 2024. That puts the gap at 16.1 cents per dollar, down from 19.4 cents in 2024 and nearly matching the all-time low of 16 cents in 2022.
In January, the Bureau of Labor Statistics reported the opposite direction: women earned 82.1 cents per dollar, down from 82.7 cents in 2024. By that measure, the gap was 17.9 cents per dollar in 2025.
The two agencies define earnings and worker groups differently, so their results are not directly comparable.
The Census measures annual earnings among people who worked full-time, year-round. The BLS measures weekly earnings among full-time workers, including those who worked only part of the year. Someone who moved into full-time work midway through 2025 could appear in the BLS data but be excluded from the Census count.
Another key difference: the Census includes self-employed workers, while the BLS pay gap figures exclude them. A strong year for self-employed women would shift the Census number but not the BLS one.
Collection methods also differ. The BLS combines monthly surveys of about 15,000 households across the year. The Census surveys about 89,000 households each spring about the previous year's income.
One further complication: because of the 2025 government shutdown, the BLS collected only 11 months of data last year. That can make even year-to-year comparisons within the BLS series difficult.
Neither number is wrong. Both are widely cited in research and news reports, and each answers a slightly different question.
What matters more is not reading too much into a single year. The reported gap moves around from year to year in both series, so longer-term trends are more informative. The table below shows how much the figures fluctuate.
Data from the Census Bureau Table 12 and the Bureau of Labor Statistics.
According to the Census, women first reached 60 cents on the dollar in 1976, 70 cents in 1990, and 80 cents in 2016. Nearly a decade later, the two federal measures put women's earnings at roughly 82 to 84 cents per dollar. The gap has shrunk considerably over the past half-century, but the pace has been slow.
An overall figure can also hide large differences among groups of women. Census data show that Black women earned 68 cents for every dollar earned by white men in 2022. By 2025, that had fallen to 64 cents. BLS data show a similar decline, from 71.2 cents per dollar in 2022 to 69.6 cents in 2025.
Both agencies use medians rather than mean averages, which may understate the true gap.
The median is the middle number when earnings are listed from lowest to highest. Take two sets of salaries: $50, $100, $120 and $50, $100, $1,000,000. Both have the same median of $100, even though one includes a million-dollar salary.
Because the median gives little weight to exceptionally high salaries, and those are disproportionately earned by men, it can miss part of the gap at the top. One analysis using mean averages instead of medians revealed a gender pay gap five percentage points larger.
For 2025, the Census and BLS figures disagree on whether the gap widened or narrowed. The long-term trend is clearer: women have made substantial progress, but equality remains distant, and the gains have not been shared equally among all women.
8.4K articles
1.6K articles
34 articles
28 articles
Comment