The Case for Four-Day Work Weeks: What Trials in Different Countries Have Found

The five-day, 40-hour working week wasn’t the product of scientific optimisation—it emerged from 20th century labour negotiations between industrial manufacturers and trade unions as an improvement over longer hours that had been standard during industrialisation. The eight-hour day and five-day week were established as norms in a context of manufacturing economies where inputs (hours worked) translated more directly to outputs (things produced) than in knowledge work, where the relationship between time and productive output is considerably more complex.

The interest in four-day working weeks over the past decade reflects both a re-examination of whether the five-day norm is optimal for contemporary knowledge work and a series of controlled trials that have produced evidence about the outcomes. The results are broadly positive and more consistent across different national contexts than might be expected—but also more nuanced than the headline claims typically convey.

The Iceland Experiments: Large-Scale and Long-Running

The most frequently cited evidence comes from Iceland, where two large-scale trials of shorter working hours were conducted between 2015 and 2019 by the Reykjavik City Council and the Icelandic national government. The trials involved approximately 2,500 workers—representing about 1% of Iceland’s entire working population—who worked either 35 or 36 hours per week rather than the standard 40, with no reduction in pay.

The trial results, published in 2021 by the Autonomy think tank and the Association for Sustainability and Democracy, found that productivity and service delivery in the participating workplaces either maintained at the same level or improved. Worker wellbeing improved substantially across multiple measures: reported stress and burnout decreased, work-life balance improved, health outcomes were better, and job satisfaction increased.

Following the trials, the vast majority of Iceland’s workforce shifted to reduced hours arrangements—approximately 86% of workers either moved to shorter hours or gained the right to do so through collective bargaining agreements. Iceland is a small, high-income country with a particular labour market structure, which limits how directly its experience translates elsewhere, but the scale and duration of the trials provide stronger evidence than most workplace experiments.

The UK Trial: 61 Companies, Multiple Sectors

The UK’s 4 Day Week Campaign ran a large-scale trial in 2022 involving 61 companies and approximately 2,900 workers across multiple sectors, in partnership with researchers from Cambridge, Oxford, and Boston College. The design was a 100-80-100 model: 100% of pay for 80% of the time while maintaining 100% of productivity. Companies were asked to find efficiency gains through process improvement, meeting reduction, and focused work practices rather than simply compressing five days of work into four.

The results published in early 2023 were broadly positive: revenue was stable or slightly increased on average, 71% of employees reported reduced burnout, sick days decreased by 65%, and 57% of employees said they were less likely to quit than before the trial. Of the 61 participating companies, 56 continued with the four-day week after the trial period; 18 made it permanent immediately.

Critically, the trial was not a controlled experiment with a comparison group: participating companies self-selected (which biases toward companies that believed it would work), the baseline was measured at the start of the trial, and there was substantial variation in results across different company types and sectors. Manufacturing and customer-facing service businesses had more difficulty maintaining output than knowledge work and professional services companies. The results are suggestive rather than definitive.

Other International Evidence

Several other national contexts have generated evidence:

Japan: Microsoft Japan ran a one-month trial in August 2019 in which all employees took every Friday off. Productivity (measured as sales per employee) increased 40% compared to the same month the previous year. The trial was widely reported but represents a short-duration, single-company experiment with several confounds (August is a historically slower business month; sales is a coarse productivity measure for a technology company).

Spain: The Spanish government ran a €10 million pilot involving volunteer companies in 2021–2022. Results were more mixed than the Nordic experience—take-up was lower than expected, many service sector companies found the transition difficult, and the evidence on productivity was less clear-cut than in the UK trial.

Belgium: In 2022, Belgium made the four-day week a legal right for workers—not a mandate, but an option. Employees can request to compress their standard 38 hours into four days with no change to pay. The compressed schedule (same hours, fewer days) is different from the reduced hours model tested in Iceland and the UK, and the evidence on outcomes from Belgium’s right-to-request model is less developed.

The Productivity Mechanism

The finding that requires most explanation is why productivity is maintained or improved when working time is reduced. Several mechanisms have been proposed and have varying degrees of empirical support:

Parkinson’s Law compression. Parkinson’s observation that work expands to fill available time has empirical support in experimental settings. When workers have 32 hours to complete what previously took 40, they eliminate low-value activities, reduce meeting time, and focus more intensely during work time. Several trial participants reported that the four-day week forced a deliberate reduction in unproductive time—long meetings, excessive email, administrative work—that improved focus even in the fifth day equivalent when they returned to five-day weeks.

Wellbeing and cognitive performance. The evidence linking fatigue and stress to cognitive performance impairment is robust. Workers operating at high stress and low rest produce lower-quality work and make more errors. If the fifth day’s work is being done by workers who are tired, stressed, and cognitively depleted, the effective productivity of that fifth day may be lower than it appears in hours-worked terms.

Recruitment and retention effects. Companies offering four-day weeks report improved ability to recruit and substantially lower turnover. In labour markets where recruitment costs are significant, the total cost advantage of lower turnover can offset the cost of paying the same total wages for fewer hours.

Where It Doesn’t Work Easily

The evidence is less favourable for certain sectors. Manufacturing with continuous production lines can’t simply stop for an extra day per week without significant operational restructuring. Customer-facing services with long opening hours require the same total coverage regardless of individual schedules—which means the four-day week requires overlapping shifts or hiring additional staff, potentially offsetting any cost savings. Healthcare and emergency services have constraints that make reduced individual hours complex to implement without service degradation.

The case is strongest for knowledge work—software development, professional services, writing, design, research—where individual output is relatively independent of colleagues’ schedules, where cognitive performance matters more than raw hours, and where the work is less dependent on simultaneous presence. The sectors most likely to adopt four-day weeks are also the sectors where self-selection bias in trials is strongest, which means the evidence base needs to be interpreted with care.

The overall picture from the international trial evidence is that a four-day working week is feasible and beneficial in a wider range of contexts than the sceptical prior might suggest—but that the magnitude of the benefit varies significantly by sector, company type, and implementation approach, and that the headline claims about universal productivity maintenance overstate the certainty of the evidence.