Four-Day Week Statistics 2026: 28% of Employers Offer One, Only 10% Actually Cut the Hours

1. The single biggest problem with four-day week statistics

Almost every number you will read about the four-day week is technically accurate and practically misleading, because at least three different schedules share the name. A statistic collected about one is routinely quoted as evidence about another.

Employer figures: CIPD, Flexible and hybrid working practices in 2025, YouGov survey of 2,050 UK HR professionals, fieldwork 14 April – 5 May 2025.

2. What employers actually offer

The CIPD's 2025 flexible working survey is the best available employer-side dataset because it asks about each variant separately rather than lumping them together. Its headline is that 28% of UK organisations offer some kind of four-day week. Its detail is that the version doing the heavy lifting is the one that does not reduce hours.

The gap between the 28% headline and the 10% reality is the whole story of this category. A journalist, a vendor or a candidate reading "more than a quarter of employers offer a four-day week" will picture a three-day weekend on full pay. What most of those employers offer is four ten-hour days.

The split by sector is sharper still, and it is almost entirely driven by the compressed variant, which is a rota-management tool in operations-heavy public services rather than a productivity experiment.

In the United States there is no equivalent breakdown, because no federal statistic tracks it. The most widely cited number comes from the American Psychological Association's Work in Americasurvey: 22% of employed adults said their employer offered a four-day workweek in 2024, up from 17% in 2023 and 14% in 2022 (Harris Poll, n=2,027, fielded March–April 2024). That is a genuine and steep trend, but it is employee-reported and the question does not distinguish compressed from reduced hours — so it is not comparable with the CIPD's 10%, and the two should not be placed in the same table without that caveat.

3. Demand is not the constraint

Every demand-side dataset points the same way, and the size of the gap is unusual. This is not a benefit employees are lukewarm about.

Two details are worth pulling out. First, the reduced-hours version is wanted almost twice as much as the compressed one (45% versus 25%) — employees are not asking for a longer day, they are asking for less work. Second, the four-day week is at the very top of the CIPD's unmet-demand list, ahead of compressed hours and flexi-time. That ordering matters for anyone designing a benefits package: the thing people want most is the thing they are least likely to be offered.

The US attitudinal data is even more lopsided. In the same APA survey, 81% of workers said they believed they could be just as effective working four days, 79% said they would be happier, and 67% expected the four-day workweek to become the American norm within their lifetime. Treat the last one as a mood reading rather than a forecast — expectations about the distant future are cheap to express — but the first two are consistent with everything the trials found about how employees experience the change.

Employees want fewer hours, not longer days. Employers mostly offer longer days. That single mismatch explains why a benefit with 45% stated demand sits at 10% availability.

4. What the largest trial actually measured

In July 2025, Nature Human Behaviour published the largest peer-reviewed study of the four-day week to date, led by Wen Fan and Juliet Schor at Boston College. It followed 2,896 employees at 141 organisations across Australia, Canada, Ireland, New Zealand, the UK and the US through a six-month trial of reduced hours with no reduction in pay, preceded by a period of deliberate work reorganisation. Twelve organisations that did not adopt the schedule served as a comparison group.

The direction of every result is positive and consistent. The magnitudes are more modest than the coverage suggested, and it is worth putting them on a common footing. Each outcome below is expressed both as the raw change the paper reports and as a share of the full range of the scale it was measured on — which is the only way to compare a 1–5 burnout scale with a 0–10 satisfaction scale.

Burnout moved most and job satisfaction moved least, which is a coherent and slightly deflating pattern: a four-day week is much better at removing strain than at making people like their job. The authors identify the mechanism as improved work ability, better sleep and less fatigue, and they note a dose–response relationship — benefits were larger for employees whose hours fell by eight or more than for those who lost one to four. On average, hours fell by roughly five per week, not the eight a literal four-day week implies.

None of this makes the finding wrong. It makes it a well-evidenced result about what happens when a willing professional-services employer redesigns its work and then cuts hours — which is a much narrower and more useful claim than "the four-day week works."

5. The cross-trial scoreboard

Trials have now run in enough places to compare. The column that deserves the most attention is the last one: what was measured objectively rather than asked about.

The German pilot, run by the University of Münster with Intraprenör and 4 Day Week Global from February 2024, is the methodologically strongest of the set and the least quoted. It is the only one that collected physiological data: cortisol from hair samples, and activity and sleep from wearables, alongside within- organisation control groups who kept standard hours. Participants slept about 38 minutes more per week than controls, took more daily steps, and showed measurably fewer high-stress minutes. Company turnover and profit were slightly up but not significantly different from the prior year — which the researchers read as consistent with productivity gains offsetting the lost hours, rather than as proof of them.

It is also the trial with the lowest continuation rate: more than 70%, against 90–100% elsewhere. Two of the 45 organisations dropped out during the trial, both large companies, citing economic pressure or a lack of internal support. A lower number from a better-instrumented study is exactly the pattern you would expect if enthusiasm is doing some of the work in the self-report trials.

6. The most-quoted four-day week statistic is not what you think

If you have seen one number about the four-day week, it is this one: 86% of Iceland's working population now works shorter hours or has the right to. It comes from a 2021 report by Autonomy and the Icelandic research organisation Alda, analysing two public-sector trials that ran from 2015 to 2019 and covered around 2,500 workers — a little over 1% of Iceland's workforce.

Iceland is still a genuinely important case: it is the only place where shorter hours moved from pilot into national collective bargaining, and that is the harder problem. But it is evidence about modest, negotiated hour reductions at national scale, not about a three-day weekend. Anyone citing it as "Iceland proved the four-day week works" has not read past the headline.

7. Why 90% retention and 0.018% adoption are both true

The apparent paradox in this field is that trials report retention rates of 90–100% while actual adoption stays in the low single digits. Both are correct, and the reason is a selection funnel that filters out almost everyone before the trial begins. Here is the arithmetic, using the two places where the numbers on both sides of the funnel are published.

Every published success rate in this category is measured at the narrow end of that funnel. Companies self-select on confidence, then are screened, then go through weeks of guided work reorganisation before the trial starts — and only then is the "did you keep it?" question asked. A 90% retention rate among firms chosen for their fit is not a forecast of what would happen if the schedule were imposed on a random employer, and the researchers do not claim it is.

8. Where things actually stand in 2026

The centre of gravity has shifted from campaign-run pilots to state-funded ones, which is the more consequential development. Poland is running the largest publicly funded trial in this part of Europe.

The Polish scheme is the one to watch, for three reasons. It is publicly funded rather than campaign-run, so the evaluation is not being conducted by an organisation whose purpose is to promote the outcome. It is deliberately model-agnostic — the ministry has been explicit that a four-day week is only one of the permitted ways to cut hours, alongside shorter days and extra leave — which means it will generate a comparison between models rather than a single verdict. And it has a phased dose (10%, then 20%), which is the closest thing yet to a dose–response design at national scale.

One caveat on the money: when the pilot was announced in April 2025 the ministry cited PLN 10 million for the first year. The final awards — 90 employers at an average of over PLN 500,000 — imply at least PLN 45 million. The budget grew; treat any single figure you see quoted for it with care.

9. Benchmark cheat sheet

Every number below is attributed to a named study with its sample size. Copy the qualifier along with the figure.

10. If you are one of the 90% who will not do this

For most employers, and for almost all of them outside professional services, a reduced-hours four-day week is not on the table in 2026. It is worth being honest about what that does and does not leave you.

The trials are consistent about why the schedule works when it works, and it is not the extra day in itself. It is recovery: better sleep, less fatigue, restored work ability. That is a specific mechanism, and it is the reason the usual substitutes are weaker than their proponents claim. Below is a candid assessment of what actually addresses the same mechanism.

That last row deserves the honesty. If your people are exhausted, a reward will not fix it, and presenting one as though it might is how recognition programmes lose credibility. Our write-up of the most common recognition programme mistakes covers this failure mode in detail: recognition offered as a substitute for a structural fix reads as an insult, not a thank-you.

Where rewards genuinely earn their place is the fairness problem the four-day week creates. Reduced hours are overwhelmingly available to desk-based staff and structurally hard for shift, frontline and customer-facing teams. If you introduce a four-day week for part of the workforce, you have just handed your most visible benefit to the group that already had the most flexibility — a dynamic we covered in frontline employee recognition. Something tangible and immediate for the teams that cannot take the day is not a consolation prize; it is the minimum required to keep the change from reading as a two-tier one. The same reasoning runs through the non-monetary rewards employees actually want, where time is consistently the most valued and the least evenly distributed.

For wider context on how this sits alongside the other big workplace shifts, see our hybrid and remote work statistics, employee burnout statistics and employee productivity statistics for 2026. If you are building a benefits package rather than a schedule, the perks and benefits platform comparison and our guide to what belongs in benefits versus rewards are the better starting points. Polish employers running the pilot should also read our guide to tax-free employee gifts in Poland before adding any reward on top of reduced hours.

Methodology and sources

Every figure on this page was checked against the original study or the publishing organisation's own release before use. Where a widely circulated number could not be traced to a primary instrument, it was either excluded or flagged in the text. Percentages are reported as the source reports them; where a statistic depends on a definition that varies between studies, the definition is stated alongside the number rather than left implicit.