Iceland's 2015–2019 hours cut was run and evaluated by the public employers themselves. The UK's widely cited 2022 pilot was organised by the campaign for the policy, using companies that volunteered. Both found gains — but the two evidence bases are not the same strength.
Iceland's trials ran from 2014 to 2019, led by Reykjavík City Council and then the Icelandic national government, and covered about 2,500 public employees across 66 workplaces — schools, hospitals, offices, a police station.12 Pay was unchanged. But the cut itself was smaller than “four-day week” implies: in 61 of the 66 workplaces, hours fell by one to three hours a week, and the largest reduction anywhere was four hours.7 Most participants kept a five-day week with shorter days, or rearranged shifts, rather than dropping a day outright.2
Both trials were designed, run and initially measured by the employers themselves — the city council and central government tracked productivity, service satisfaction, sick days and stress internally throughout.1 The international profile came afterwards: UK think tank Autonomy and Icelandic research group Alda analysed the results and published them as Going Public in 2021, two years after the trials ended, and it is that report most non-Icelandic coverage draws on.3
The trials fed into collective bargaining. Between 2019 and 2021, Icelandic unions renegotiated agreements covering 170,200 of the country's roughly 197,000 union members — about 86% — giving most workers either shorter hours or the right to negotiate them.2 That coverage figure is real and large. But the hours actually banked look nothing like the trial: the agreements that followed delivered around 35 minutes a week off in the private sector and 65 minutes in the public sector — a fraction of the up-to-four-hour cuts tested.7 A wide-reach statistic and a small-effect statistic are both true at once; reporting only the first misleads on the size of the change.
The UK's most-cited four-day week evidence comes from a six-month pilot, June to December 2022, in which 61 companies and about 2,900 employees moved to a four-day week on full pay.4 It was organised by 4 Day Week Global and the UK's 4 Day Week Campaign — both advocacy bodies for the policy — working with Autonomy, the same think tank behind the Iceland write-up, and academic teams at Cambridge and Boston College who ran surveys and interviews.5 The design was self-selected throughout: companies volunteered, each chose its own version of a shorter week, and outcomes — including revenue and staff attrition — were self-reported by the participating firms. No matched group of similar companies staying on five days was tracked alongside them.5
The clearest self-reported shifts in the UK pilot were on staff wellbeing and retention, not on the finances. Comparing before and after the six months, 71% of employees reported lower burnout and 39% reported feeling less stressed; sick days fell 65% and the number of staff leaving participating companies fell 57%, against the same period the previous year.4 Revenue moved least of all — up 1.4% on average — and even that figure covers only the 23 of 61 organisations able to supply usable revenue data, not the full pilot.4
Neither trial had a matched control group of employers who did not change their hours, so it's not possible to fully separate the effect of shorter hours from the effect of volunteering to be watched.5 The UK pilot's wellbeing and attrition figures are company-reported, not independently audited, and its one solid financial figure — revenue up 1.4% — is drawn from only 23 of the 61 companies, so it says little about the other 38.4 And it remains unclear whether Iceland's largest cuts — up to four hours a week, tested only in public-sector workplaces that opted in — would hold up more broadly: the union agreements that actually followed delivered far less.7
Both trials found gains and no adverse effect on output. What's disputed is how much weight that evidence should carry.
Backers point to convergent findings: Iceland's employer-run data, the UK pilot's academically-run surveys, and later trials elsewhere that did use control groups, which researchers involved say found similar wellbeing gains with no comparable change among the companies that kept a five-day week.5
Leisure-studies academic A.J. Veal argues Iceland's results are routinely over-read as a four-day week when they were a modest hours cut, most of which did not survive into national bargaining.7 Michael Sanders, professor of public policy at King's College London, argues self-selected trials like the UK's mainly show what happens when already-enthusiastic employers try the policy — not what a reluctant one would find.8
Held out of this comparison to keep it to two trials: