← Articles
Read in another language
Society

The Four-Day Week: From Experiment to Institution?

Jayden

Analyzes global supply chains, industrial policy, and technology issues.

Published

Key points

  • The four-day week that trials actually test is the 100-80-100 model — 100% of pay, 80% of hours, 100% of output — and it is not the same thing as Belgium's compressed 4×10 week, which packs 38 hours into four days without cutting total working time.
  • The results are consistent and they stick: 56 of the 61 UK pilot companies (92%) carried on after the 2022 trial, and a year later at least 54 (89%) were still on a four-day week with 31 (51%) having made it permanent.
  • The largest peer-reviewed study — 2,896 workers at 141 organisations across six countries, published in Nature Human Behaviour in 2025 — found burnout falling from 2.83 to 2.38 on a five-point scale and job satisfaction rising from 7.07 to 7.59 on a ten-point scale, with no such pattern in 12 control companies.
  • What the evidence does not yet show is company performance. The Nature authors did not analyse firm-level productivity at all, and Germany's Münster-run pilot found no clear change in absenteeism, financial performance or productivity — improved well-being alongside unchanged output.
  • Every large trial so far has been run with self-selected, mostly white-collar volunteers and no random assignment — the Nature authors themselves call for a randomised controlled trial — while shift-based, continuous and face-to-face sectors face structural obstacles that a simple hour cut does not solve.

For a century the standard workweek has been treated as a fixed law of economic life: five days, roughly forty hours, everyone in at the same time. In the last few years that assumption has been quietly stress-tested. Dozens of companies and a handful of governments have run pilots of a shorter week with no cut in pay, and the results have started to pile up. In 2025 the largest study of the idea yet was published in a peer-reviewed journal, and in April that year Tokyo's metropolitan government began offering a four-day option to roughly 160,000 employees. What was a curiosity is now a small mountain of data.

So this is a good moment to ask the harder question: what do those pilots actually show? The honest answer requires keeping three distinctions in view throughout. First, results measured on a few dozen volunteer companies are not the same as an economy-wide effect. Second, workers feeling better is not the same as a firm performing better, and the two are measured very differently. Third, a report from the advocacy group that organized a trial is not the same as independent peer review. Hold those apart, and a nuanced picture emerges — genuinely encouraging on well-being, much thinner on hard output, and far from settled as policy.

Table of Contents

  1. What "100-80-100" actually means
  2. The evidence so far: what the pilots found
  3. Why it seems to work — and where the methods wobble
  4. The hard cases: where four days strains
  5. From experiment to policy
  6. What to watch

What "100-80-100" actually means

The model at the center of the current wave has a shorthand: 100-80-100. Employees keep 100% of their pay, work 80% of their previous hours — typically dropping from five days to four — and are expected to sustain 100% of their prior output. The framing was popularized by 4 Day Week Global, a non-profit that has organized many of the coordinated trials, on the premise that the lost day can be recovered through sharper focus and less wasted time [source: 4 Day Week Global, 2023].

It is worth pausing on what this is not, because the term "four-day week" hides two very different things. A genuine 100-80-100 week reduces total hours worked. A compressed week — the model Belgium wrote into law in 2022 — squeezes the same 38 or 40 hours into four longer days of roughly ten hours each, leaving total working time unchanged [source: Reuters, 2022]. The two produce different fatigue, different childcare logistics, and different economics. Most of the well-being findings below come from the reduced-hours version; conflating it with mere compression is one of the most common errors in the debate.

The evidence so far: what the pilots found

The UK trial and its follow-up

The best-known experiment is the UK pilot of 2022, run by 4 Day Week Global with the Autonomy Institute and researchers at the University of Cambridge and Boston College. From June to December, 61 companies and about 2,900 employees moved to a four-day week on full pay. When the organizers reported results, 56 of the 61 firms (92%) had chosen to continue and 18 had made the change permanent. Staff turnover fell 57% against the same period a year earlier, and 71% of employees reported reduced burnout [source: Autonomy, 2023].

Two caveats belong right next to those numbers. The companies volunteered, so they were probably predisposed to make it work, and the organizers are advocates for the policy — this is participant-reported data from an interested party, not a neutral audit. The widely repeated claim that revenue rose about 35% is especially soft: it compares the trial period to the same months a year earlier, a gap that could be explained by ordinary growth or market conditions rather than the schedule [source: Autonomy, 2023]. A follow-up a year later found the change durable: at least 54 of the 61 firms still had the four-day week in place and 31 had made it permanent, with the earlier well-being gains largely sustained [source: Autonomy, 2024].

The peer-reviewed study

The strongest evidence to date arrived in 2025, when researchers led by Boston College sociologists Wen Fan and Juliet Schor published in Nature Human Behaviour. Pooling 2,896 employees across 141 organizations in six countries — the US, UK, Ireland, Canada, Australia, and New Zealand — they compared surveys before and after a six-month four-day trial. Work-related burnout fell from 2.83 to 2.38 on a five-point scale, job satisfaction rose from 7.07 to 7.59 on a ten-point scale, and mental and physical health improved — changes not seen in a set of 12 control companies. The gains were driven partly by better sleep, less fatigue, and improved work ability [source: Nature Human Behaviour, 2025].

This is real peer-reviewed evidence, and it matters. But note carefully what it does and does not establish. Every one of those outcomes is self-reported well-being, gathered from survey questions such as how workers rate their own mental health. The study did not analyze company-wide productivity or revenue at all. And because the firms self-selected, the authors themselves flag the likely overestimate and explicitly call for randomized controlled trials to pin down the true effect [source: Nature Human Behaviour, 2025]. In other words, the best study we have is strong on how people feel and silent on how firms perform.

Germany and Portugal

Two 2024 government-adjacent trials add texture. In Germany, a nationwide pilot organized with the University of Münster ran 45 companies through six months; 73% chose to keep the schedule. Employees slept about 38 minutes more per week and were more physically active than a control group, and self-reported productivity rose. Tellingly, the researchers found no notable change in absenteeism or financial performance — well-being improved while hard performance held roughly steady rather than surging [source: University of Münster, 2024]. Portugal's government-backed trial, coordinated through Birkbeck and Henley Business School, ran about 41 companies who cut average hours from 41.6 to 36.5 per week. Exhaustion and negative mental-health symptoms fell against a control group, 93% of workers wanted to continue, and only four firms returned to five days [source: Henley Business School, 2024].

Why it seems to work — and where the methods wobble

The mechanism advocates propose is intuitive. Given one less day, teams cut the low-value parts of the week first: fewer and shorter meetings, less multitasking, tighter agendas. Better-rested workers concentrate more and burn out less, and the recovered focus offsets the lost hours. The consistent finding across pilots — that people report lower burnout and better sleep while output does not obviously fall — is at least consistent with that story.

But the methods have real limits, and they run in one direction. Companies that join these trials volunteer, which means the sample is tilted toward firms already confident it will work — classic self-selection that inflates apparent success. Most outcomes are self-reported, and employees who want to keep an extra day off have an obvious incentive to answer generously. Follow-up windows are short, samples are often a few dozen firms, and — crucially — hard productivity is rarely measured at all; the Nature team, with the largest dataset, did not measure it [source: Nature Human Behaviour, 2025]. The participating firms also skew heavily toward white-collar and knowledge work, where a wasted hour is easy to trim. None of this means the benefits are illusory. It means the evidence is strong for well-being, weak for output, and not yet the kind of randomized test that would let anyone generalize to the whole economy.

The hard cases: where four days strains

The pilots' composition is itself a warning. A software team can compress its week by killing meetings; a hospital ward, a factory line, or a fire station cannot. Where the work is continuous, shift-based, or face-to-face, the hours are the service, and cutting them means either hiring more people or covering less. Healthcare has to staff around the clock under safety and regulatory constraints; manufacturing depends on continuous production and expensive equipment utilization, so a shorter week can mean idle capital or an extra shift to arrange [source: APA Monitor, 2025]. There is a safety dimension too: compressing hours into longer days can raise fatigue in exactly the high-risk settings — construction, energy, transport — where tired workers are dangerous.

That does not make a four-day week impossible in these sectors; some hospitals and plants have experimented with rota redesigns. It does mean the headline results from knowledge-work pilots should not be read as a promise to a nurse or a machine operator. The cost question is also real and often skipped: if output per hour does not rise enough to cover the lost fifth day, someone — the employer, the customer, or the worker through slower raises — pays the difference. Fairly stated, the case for a four-day week is strongest where the bottleneck is attention, and weakest where the bottleneck is coverage.

From experiment to policy

Even with those caveats, the idea has begun crossing from corporate pilots into public policy — though "institution" still overstates it. Iceland is the most cited example: public-sector trials between 2015 and 2019 covering more than 2,500 workers led to agreements under which roughly 86% of the workforce gained shorter hours or the right to them [source: Autonomy, 2021]. But Iceland also illustrates the framing trap. Critics note the reduction was modest — often one to three hours, not a true four-day week — and that the "overwhelming success" headline rested partly on managers' own assessments, so the case is more measured than the slogans suggest [source: The Conversation, 2021].

Governments have taken different routes. Belgium legislated a right to request a compressed four-day week — same hours, four days — rather than a reduction [source: Reuters, 2022]. Spain went further toward the reduced-hours model, using EU-backed public funds to subsidize small and medium manufacturers that cut hours without cutting pay, a notable attempt to test the idea in industry rather than offices; a parallel experiment in Valencia found improved self-reported health [source: World Economic Forum, 2023]. And in Japan, where overwork is a chronic policy concern, the Tokyo metropolitan government began offering a four-day option to its employees in April 2025, explicitly linking it to work-life balance and the country's low birth rate, with other prefectures following [source: NBC News, 2025]. These are mostly public-sector or subsidized moves and opt-in rights, not economy-wide mandates — real momentum, but a long way from a settled norm.

What to watch

The four-day week has earned a serious hearing. Across independent settings, the well-being signal is consistent: less burnout, better sleep, high rates of firms choosing to continue. That is not nothing, and for knowledge-work employers competing for talent it may already be reason enough. What the evidence does not yet support is the stronger claim — that shorter hours reliably hold or raise output across the economy — because the studies that measure well-being mostly do not measure productivity, and the firms in them chose to be there.

So watch three things. First, randomized trials of the kind the Nature authors called for, which would separate the schedule's effect from the enthusiasm of volunteers. Second, results from outside the white-collar core — the manufacturing subsidies in Spain, any serious pilot in healthcare or continuous operations — because that is where the model's limits will show. Third, whether the government moves in Tokyo, Belgium, and elsewhere stay opt-in perks or harden into durable entitlements. Until then, the fair verdict is the careful one: a promising redesign of work with strong evidence for how it feels, thinner evidence for what it produces, and a real but unfinished path from experiment to institution.

Charts

How many of the 61 UK pilot companies kept the four-day week

How many of the 61 UK pilot companies kept the four-day weekStill four-day, end of pilot (2023-02) 56companies, Made permanent, end of pilot 18companies, Still four-day, one year later (2024-02) 54companies, Made permanent, one year later 31companies56companiesStill four-day, end of pilot (2023-02)18companiesMade permanent, end of pilot54companiesStill four-day, one year later (2024-02)31companiesMade permanent, one year later
All four bars are counts out of the same 61 companies. The first two come from the pilot results published in February 2023, the second two from Autonomy's one-year follow-up "Making It Stick" (February 2024), which reports at least 54 still operating a four-day week. Both reports are from the organisers, who advocate for the policy.Autonomy Institute, Making It Stick (2024-02) (opens in a new tab)

Burnout before and after a six-month trial (six countries)

Burnout before and after a six-month trial (six countries)Before 2.83pts, After 2.38pts2.83ptsBefore2.38ptsAfter
Self-reported burnout on a 1–5 scale, where lower is better, among 2,896 workers at 141 organisations in the US, UK, Ireland, Canada, Australia and New Zealand. The same pattern was not observed in the 12 control companies. This is a survey measure, not an output measure.Fan, Schor et al., Nature Human Behaviour (2025) (opens in a new tab)

Job satisfaction before and after the same trial

Job satisfaction before and after the same trialBefore 7.07pts, After 7.59pts7.07ptsBefore7.59ptsAfter
Self-reported job satisfaction on a 0–10 scale, where higher is better, from the same 2,896 respondents. The authors attribute the gains partly to better work ability, less fatigue and improved sleep — and state explicitly that company-wide productivity was not analysed.Fan, Schor et al., Nature Human Behaviour (2025) (opens in a new tab)

Portugal's government-backed trial: weekly hours worked

Portugal's government-backed trial: weekly hours workedBefore 41.6hours, After 36.5hours41.6hoursBefore36.5hoursAfter
Average weekly hours including overtime across about 41 companies over six months — a 12% reduction. Measured against a control group, participants reported better mental and physical health and significantly less exhaustion; 93% wanted to continue and only four of the 41 companies went back to five days.Henley Business School / Birkbeck, Portugal four-day week trial results (2024) (opens in a new tab)

Timeline

  1. Iceland runs public-sector trials covering more than 2,500 workers in healthcare, education and administration, cutting the week from 40 hours to 35–36 — a reduction of one to three hours, not a four-day week.

  2. Icelandic collective agreements extend shorter hours, or the right to them, to roughly 86% of the workforce.

  3. Autonomy and Alda publish "Going Public" on the Icelandic results; The Conversation argues the "overwhelming success" framing is greatly overstated, since the cut was small and productivity claims rest on managers' own assessments.

    The Conversation (2021) (opens in a new tab)
  4. Japan's government begins recommending optional four-day weeks, framed as a response to overwork and the falling birth rate.

  5. Belgium's federal labour reform gives workers the right to compress a 38-hour week into four days, subject to employer approval — total hours are unchanged.

  6. The UK pilot runs at 61 companies with about 2,900 workers on the 100-80-100 model, organised by 4 Day Week Global, Autonomy, the University of Cambridge and Boston College.

  7. UK results published: 56 of 61 companies continue and 18 make the change permanent; resignations are down 57% year over year; 71% report less burnout and 39% less stress.

    Autonomy Institute, UK pilot results (2023) (opens in a new tab)
  8. Spain's industry ministry opens a subsidy programme for small and medium manufacturers to cut hours without cutting pay, using EU funds; Valencia turns four consecutive Mondays in April and May into regional holidays, affecting about 360,000 people.

  9. Germany's national pilot runs at 45 companies with roughly 900 employees, organised with the University of Münster.

  10. Autonomy's one-year follow-up finds at least 54 of the 61 UK companies still on a four-day week and 31 permanently converted.

  11. Portugal's government-backed trial reports results: weekly hours fall from 41.6 to 36.5, with significant reductions in exhaustion measured against a control group.

  12. Münster reports German results: 73% of companies decide to continue, two large firms drop out, workers sleep 38 minutes more a week and take 1,848 more steps — while absenteeism, financial performance and productivity show no clear change.

    University of Münster (2024) (opens in a new tab)
  13. Nature Human Behaviour publishes the six-country study of 2,896 workers at 141 organisations, the largest peer-reviewed evidence to date.

    Nature Human Behaviour (2025) (opens in a new tab)
  14. The Tokyo Metropolitan Government offers an optional four-day week to roughly 160,000 public employees, a model that then spreads to Osaka, Chiba and Kanagawa, with Miyagi planning to expand by fiscal 2026.

Analysis

Well-being is measured; company performance mostly is not

The strongest finding in the literature is about how workers feel: burnout down from 2.83 to 2.38, satisfaction up from 7.07 to 7.59, less exhaustion in Portugal, 38 minutes more sleep a week in Germany. The weakest link is what happens to the business. The Nature authors state plainly that firm-level productivity was not part of the analysis, and the German pilot — which did look — found absenteeism, financial performance and productivity essentially unchanged. "Workers are better off and output did not fall" is a defensible reading of the evidence. "Productivity rises" is not.

The volunteer problem nobody has solved yet

Companies that join a four-day-week trial are companies that already think it might work for them, and they skew heavily towards knowledge work. That self-selection runs through every large result quoted in this article, and it is why the authors of the biggest peer-reviewed study call for a randomised controlled trial rather than treating their own findings as settled. Until someone runs one, the honest summary is that the model works for the kind of firm that opts into it.

The revenue number that gets quoted the most is the weakest one

The widely repeated figure of roughly 35% higher revenue from the UK pilot is a simple comparison with the same period a year earlier. It does not isolate the effect of the four-day week from everything else that changed in those twelve months, and the organisers' own framing of revenue during the trial was that it held broadly steady. Treat it as a confounded indicator, not as a return on the policy.

Iceland is the most cited case and the least representative

The Icelandic trials cut the working week from 40 hours to 35–36 — one to three hours, not a fifth of the week — so they are not a four-day-week test at all, and the claim that productivity was maintained rests on managers' own assessments. Critics have argued the "overwhelming success" framing is greatly overstated. What Iceland does show, and it is not nothing, is that a negotiated hours reduction can be extended to most of a national workforce: roughly 86% ended up with shorter hours or the right to them.

Where the model runs into structure, not attitude

The obstacle in healthcare, continuous manufacturing, emergency services and other shift-based work is not managerial reluctance but coverage: patients, production lines and call-outs do not pause for a fifth day, so cutting hours means hiring, paying more, or lengthening shifts — and longer shifts bring fatigue and safety questions of their own. Spain's subsidy programme for small manufacturers is one of the few interventions aimed squarely at that gap.

Comparison

Two different policies share one name. Mixing them is the single most common error in four-day-week coverage.
100-80-100Compressed week (4×10)
Total working hoursCut by about 20% — the week itself gets shorterUnchanged — the same hours are packed into fewer days
PayHeld at 100%Unchanged
Working dayRoughly the same length, one day fewerLonger: about 10 hours a day
ExampleThe UK, German, Portuguese and six-country trialsBelgium's 2022 federal labour reform, which requires employer approval
What the trials measureWhether output holds up when hours fallNot applicable — hours did not fall
The same headline — "the four-day week works" — rests on evidence of very different strength. Read the tier before reading the number.
StudyScaleWhat it measuredEvidence tier
UK pilot (2022, results 2023)61 companies, ~2,900 workersRetention, resignations, self-reported burnout and stressOrganiser-run trial by advocacy bodies; self-selected participants; self-reported outcomes
Nature Human Behaviour (2025)141 organisations, 2,896 workers, six countriesSelf-reported burnout, job satisfaction, mental and physical health, with 12 control companiesPeer-reviewed — but self-selected, not randomised, and firm productivity was not analysed
Germany (2024, Münster)45 companies, ~900 workersSleep, steps, physical activity and self-reported productivity against a control group, plus absenteeism and financial performanceUniversity research team with control comparison; large self-reported component
Portugal (2024, Henley/Birkbeck)~41 companies, six monthsActual weekly hours, exhaustion and health against a control groupGovernment-backed trial with control comparison; self-reported well-being
Iceland (2015–2019)2,500+ public-sector workersHours, well-being, productivity maintained per managers' assessmentPublic-sector trial, but a one-to-three-hour cut rather than a four-day week, and the success framing is contested

Process

  1. Which model is it?

    100-80-100 cuts total hours; a compressed 4×10 week does not. A claim about one says nothing about the other.

  2. Who ran it?

    4 Day Week Global and Autonomy advocate for the policy they are testing. That does not make their numbers wrong, but it belongs in the description.

  3. Who took part?

    Every large trial so far used companies that volunteered, and most of them were white-collar. Firms that expect the change to fail do not sign up.

  4. Self-reported or measured?

    Burnout and satisfaction come from questionnaires. Hours worked, resignations and absenteeism are records. Both are useful; they are not the same kind of evidence.

  5. Was there a control group?

    The six-country study compared against 12 control companies, and the German and Portuguese trials measured their results against control groups. Where no comparison group is named, a before-and-after gap is not yet an effect.

  6. Was output measured at all?

    This is where most trials stop. The Nature authors state they did not analyse company-wide productivity; Germany measured it and found no clear change either way.

Sources

  1. Autonomy Institute — The results are in: the UK's four-day week pilot (2023-02).View source (opens in a new tab)
  2. Autonomy Institute — Making It Stick: The UK Four-Day Week Pilot One Year On (2024-02).View source (opens in a new tab)
  3. 4 Day Week Global — The 4 Day Week UK Pilot Programme Results (2023).View source (opens in a new tab)
  4. Nature Human Behaviour — Fan W., Schor J., et al. Work time reduction via a 4-day workweek finds improvements in workers' well-being (2025).View source (opens in a new tab)
  5. University of Münster — Wellbeing increases when working hours are reduced (2024-10).View source (opens in a new tab)
  6. Henley Business School — Results from Portugal's government-backed four-day week trial released (2024).View source (opens in a new tab)
  7. World Economic Forum — Four-day work week trial in Spain leads to healthier workers, less pollution (2023-10).View source (opens in a new tab)
  8. Autonomy Institute & Alda — Going Public: Iceland's journey to a shorter working week (2021).View source (opens in a new tab)
  9. The Conversation — The success of Iceland's four-day week trial has been greatly overstated (2021).View source (opens in a new tab)
  10. Reuters — Belgium agrees four-day week and gives employees right to ignore bosses after hours (2022-02).View source (opens in a new tab)
  11. NBC News — Tokyo government to introduce four-day workweek for its employees (2024).View source (opens in a new tab)
  12. APA Monitor on Psychology — The rise of the 4-day workweek (2025-01).View source (opens in a new tab)

Tags

  • #four-day-workweek
  • #work-life-balance
  • #productivity
  • #labor-policy
  • #future-of-work