The question that refuses to die
Every few months, someone publishes a headline declaring the four-day work week an unqualified success. A few months later, someone else publishes one saying it collapsed. Both can’t be right. Let’s look at what actually happened in the trials, not the press releases.
Truza’s job here is not to tell you whether a shorter week is good or bad. It’s to lay out what the strongest sources actually found, grade those sources honestly using our 5-star ranking system, and let you decide.
The Iceland trial (2019-2021)
This is the one most people cite. Between 2015 and 2019, roughly 2,500 Icelandic workers across 66 workplaces shifted to a 35-36 hour week without a pay cut. The trials covered police, care homes, preschools, city maintenance crews, and office workers.
The findings, published by the Icelandic Association for Democracy and Equality and Autonomy (a UK think tank), reported that productivity stayed the same or improved in most workplaces. Worker wellbeing went up. Sick days went down. No workplace reverted to the old schedule after the trial ended.
Source rating: 3 stars. The report was co-authored by the organization that helped design the trials, which is a structural bias. The sample was limited to public-sector jobs where output is harder to measure than in, say, a factory or a sales floor. The report itself acknowledges it was “not a controlled scientific study.” That doesn’t mean it’s wrong. It means you should weight it accordingly.
The UK four-day week pilot (2022-2023)
61 companies in the UK took part, with about 2,900 workers. They cut hours to 80% of a normal week while keeping pay at 100%. The results were published by researchers at Cambridge, Boston College, and the 4 Day Week Global foundation.
Headline numbers: 92% of companies said they’d continue. Revenue stayed broadly flat, rising 1.1% on average. Hiring went up. Sick days dropped by 65%. None of this is trivial.
But here’s the part that got less coverage: the companies that signed up were self-selected. They chose to participate because they already believed it might work. There was no control group. And “revenue stayed flat” is not the same as “revenue would have stayed flat without the change.” In a growing market, flat revenue under a shorter week could still mean you left money on the table.
Source rating: 3 stars. The academic team is credible. The methodology is well-documented. But self-selection bias and the lack of a control group are real limitations that the press coverage largely ignored.
What Belgium actually did (2022)
Belgium passed a labour reform in 2022 that is often reported as “Belgium introduced a four-day week.” That headline is misleading. What Belgium actually did was give workers the right to compress a full 38-hour week into four days without overtime penalty. That is a compressed week, not a shorter one. You still work 38 hours. You just do it in longer days.
The reform also gave employees a legal right to request a shift to a four-day pattern. Employers can refuse but must justify the refusal. This is a flexibility measure, not a productivity experiment.
Source rating: 4 stars. The legislation itself is public and verifiable. The misreporting came from secondary sources that confused “compressed” with “reduced.”
The Microsoft Japan experiment (2019)
Microsoft Japan tried a four-day week for one month in August 2019. They reported a 40% productivity increase compared to the same month the previous year. The result was widely cited as proof that shorter weeks boost output.
The problem: August in Japan is a low-productivity month due to summer holidays (including Obon). Comparing one August to the previous August, when you’ve also introduced other efficiency measures, tells you less than the headline suggests. A single month at a single subsidiary is not a trial. It’s an anecdote with a spreadsheet.
Source rating: 2 stars. The company reported its own numbers, for one month, with no peer review and no control. Interesting, but not evidence.
What nobody mentions: the sectors where it doesn’t apply
Nearly every trial and pilot has been conducted in office, care, or public-sector work. There is a reason for this. In jobs where output is task-based rather than time-based, compressing or reducing hours is theoretically possible. But consider:
- Manufacturing lines run on machine hours, not person-hours. A 20% reduction in staff hours means 20% less line time unless you hire more people or run more shifts.
- Hospital nursing already runs on 12-hour shifts. Compressing further is not a safety question you want to test casually.
- Small retail with one or two staff cannot close for a day and still serve customers.
This does not make the four-day week wrong. It means the evidence base is narrower than the claims built on top of it. A care home in Reykjavik and a foundry in Sheffield are not the same operating environment.
The honest summary
Here is what the evidence, graded honestly, supports:
- In self-selected, mostly public-sector or knowledge-work settings, a reduced-hour week has not measurably reduced output in the trials conducted so far.
- Worker satisfaction and reduced sick absence are the most consistently reported benefits across studies. These are real and should not be dismissed.
- No trial has been conducted with a proper control group in a private-sector, output-measured, for-profit context over a period longer than 12 months.
- The headline productivity gains from individual company experiments (like Microsoft Japan) are not reliable evidence because of confounding variables.
If you are a manager considering this for your team, the honest reading of the evidence is: it might work for you, the trials are encouraging but not conclusive, and the only way to know for your specific context is to try it carefully with measurement. If you are a policymaker, the evidence does not yet support a blanket mandate.
If you want to explore more topics like this one, browse the full alphabetical list of issues Truza is tracking.