The single best number in the American labor market right now belongs to women. In March 2026, the share of women aged 25 to 54 who were working or looking for work hit 78.5 percent — the highest reading since the government started measuring it in 1948. It has drifted down since, to 77.6 percent in June, but every month since January 2024 has matched or exceeded anything the country managed in the 1990s boom, in the 2000s, or in the long expansion that ended in February 2020.
Over the same stretch — February 2020 to June 2026 — prime-age men went almost nowhere: 89.1 percent in February 2020, 89.2 percent in June 2026, and an annual average that rose about half a point between 2019 and the first half of 2026 against more than two points for women. Whatever has been lifting people into paid work has been mostly a women's phenomenon, and I think the biggest single ingredient is the least glamorous one: they no longer have to be in a particular building at a particular hour to hold the job.
Why this is remote work and not just a hot labor market
A chart cannot prove causation. Lots of things were going on: a very tight labor market, later childbearing, women's continued education gains, and — as the Hamilton Project at Brookings puts it — pandemic-era federal money for childcare, which it offers as one possible contributor alongside telework itself. But how little men moved is a problem for the "it was just a strong economy" story. A hot labor market pulls in everyone at the margin. This one pulled in women roughly four times as hard as men.
What changed asymmetrically for women was the price of showing up. Before 2020, a job was a bundle: the work, plus a commute, plus a fixed schedule that had to be reconciled with school pickup, a sick child, or a parent's appointment. Claudia Goldin's research on the gender pay gap — part of the body of work that won her the 2023 Nobel — showed that the modern gap is driven less by unequal pay for equal work than by the premium employers pay for long, rigid, unpredictable hours, and by the fact that the person who absorbs the family's scheduling shocks is usually the mother. Remote and hybrid work cut the price of that bundle. They do not make anyone a better economist or a faster coder; they just remove ninety minutes of commuting and make a 3 p.m. emergency survivable.
The most credible evidence that this actually moves employment is quasi-experimental rather than anecdotal. Emma Harrington and Matthew Kahn, in a 2025 NBER paper, use the fact that technology made working from home feasible in some college fields well before the pandemic and not in others. In the fields where remote work became more common, the employment gap between mothers and other women narrowed: a 10 percent rise in remote work was associated with mothers' employment rising about 0.78 percentage points relative to comparable women without children. Their panel data point to the mechanism — women who could work from home before having a child were less likely to leave the workforce afterward. It is a difference-in-differences design on survey data, not a randomized trial, and the authors are careful to say so.
The people doing it are the people the theory predicts. In the second quarter of 2026, 24.3 percent of women at work did some of their paid work at home during the survey week, against 19.4 percent of men. Workers with children under 18 at home teleworked at 25.7 percent, against 20.0 percent for everyone else. That last gap is small, and telework is not overwhelmingly a parents' benefit. It is a benefit that mothers, in particular, appear to convert into staying employed.
Compare 2019 with 2025 in the Bureau of Labor Statistics' annual family employment release: participation among women with a child under six rose from 66.4 percent to 68.0 percent, among women whose youngest is 6 to 17 from 76.8 to 78.2 percent, and among married mothers of children under 18 from 69.9 to 72.1 percent — the biggest gain of the lot, in the group whose participation is most sensitive to whether the household can cover the day. Women with no children under 18 went the other way, from 52.2 to 51.9 percent.
That last comparison needs a caveat: women "with no children under 18" includes everyone from 16 to 96, so an aging population drags their rate down for reasons that have nothing to do with laptops. It is a check on the story, not the proof.
The gain is real, and it is narrow
Nobody should be declaring the problem solved. Working from home did not arrive for American women. It arrived for some of them.
Among workers aged 25 and over with a bachelor's degree or higher, 37.8 percent teleworked in the second quarter. Among high school graduates with no college, 9.4 percent did. Among those without a high school diploma, 4.6 percent. The education gap in access to flexibility is roughly six times the gap between women and men — so the flexibility is reaching one slice of women, and any participation gain running through it can only reach that slice.
You can see the same thing by occupation. Four groups — management and finance, professional occupations, office and administrative support, and sales — have telework rates between about 24 and 43 percent, and they employ about 56 million women. The other six groups in the chart below, from service work to production to transportation, have telework rates between about 2 and 6 percent, and they employ about 22 million women: more than one in four of the women in these ten groups. A home health aide, a line cook, a warehouse picker and a school bus driver have gotten nothing at all out of the flexibility revolution.
Twenty-two million is a floor, because the group averages hide their own inequality. Inside the "professional" group, which averages 31.7 percent, computer and mathematical occupations run at 63.3 percent — while the two largest occupations inside that group, both about three-quarters female, do not. Healthcare practitioners and technicians — nurses, therapists, pharmacists — 7.9 million women, telework at 13.1 percent. Education, training and library occupations — teachers, mostly — 7.0 million women, at 14.8 percent. Both sit inside the flexible groups on the chart and have almost none of the flexibility.
None of this is an accident of who wanted it. It follows the technology: Jonathan Dingel and Brent Neiman's count of jobs that can be done at home, published in the first weeks of the pandemic, has held up well as a map of where remote work actually landed. The jobs that could go remote were disproportionately the ones that already paid well.
What could take it away
Telework has not collapsed. Comparing like with like in an unadjusted series — second quarter to second quarter — it has been flat: 21.9 percent of workers in 2024, 21.6 in 2025, 21.7 in 2026. A separate survey by WFH Research, which measures the share of paid working days done at home rather than the share of people, tells a slightly different story: rather than holding flat it has drifted down each year, to just under 26 percent of days in 2026 so far from 28.8 percent in 2023. Its pre-pandemic comparison, about 7 percent, comes from the 2019 American Time Use Survey rather than from its own panel, which did not exist before May 2020.
But this is an arrangement employers grant and can withdraw, and the withdrawal has started at the top. In January 2025 a presidential memorandum directed federal agency heads to end remote-work arrangements and return employees to in-person work full time, subject to whatever exemptions each agency head deems necessary; implementation has been messy, and several large private employers have made similar moves.
The case against them, on the employer's own terms, is strong. The most careful study of hybrid work we have — Nicholas Bloom, Ruobing Han and James Liang's randomized trial, later published in Nature — assigned 1,612 engineering, marketing and finance staff at a large travel-technology firm to either two days a week at home or five days in the office, by whether their birthday was odd or even. Attrition fell by about a third. There was no significant effect on performance ratings or promotions. It is one firm — Trip.com, in Shanghai, in 2021 and 2022 — with a graduate workforce and a co-author who chairs the company, so it does not settle the question for a warehouse. But firms treating two days at home as a perk to claw back in a soft labor market are throwing away a retention tool that costs them roughly nothing — and, in the aggregate, are quietly deciding how many mothers stay in the workforce.
Treat flexibility as infrastructure
Three things follow, and none of them requires new spending.
Employers should price flexibility as compensation, because that is what it is. Workers give up real wages for it, and the retention evidence says it is close to free. A firm withdrawing hybrid work is cutting pay without saying so — and cutting it hardest for the employees most likely to leave rather than negotiate.
Governments should measure what their own mandates do. It is hard to think of a larger single withdrawal of workplace flexibility than the federal return-to-office order, and it is running without an evaluation. Separation rates by sex and parental status, for affected agencies against comparable unaffected ones, would answer within a year whether the participation gain carries a policy risk. The agencies already hold those records; someone needs to publish them.
And nobody should mistake this for a childcare policy. The 22 million women in jobs that cannot go remote — plus the teachers whose jobs technically can and don't, and the nurses whose jobs never could — need the older agenda: affordable childcare, predictable schedules, paid leave. The Hamilton Project at Brookings found that in the first quarter of 2025, caregiving still accounted for more than two-thirds of non-participation among women aged 25 to 44. Remote work did not touch that. It only rearranged who has to choose.
The record set in March 2026 was decades in the making and is, in part, an accident — an emergency measure that turned out to solve a problem nobody had designed a policy for. It would be a strange kind of failure to let it slip away by treating it as a perk.