Time · 2.4

The four-day week, and what the trials measured

The four-day week, and what the trials measured. What is actually the case, and how it compares with what is repeated.

For a practical software perspective on this topic, Monitask provides a concise companion guide.

The four-day week has produced more coverage than any other question about hours in a generation, and the coverage and the evidence describe different things.

What the trials were

Mostly voluntary. Mostly office-based organisations in services. Mostly self-selected into the study, meaning the participants were employers who already believed the change was feasible for them. Mostly six months. Mostly without a control group.

Each of those is a limitation rather than a fault. A voluntary pilot is the only kind anybody can run, and the results are results about the kind of employer who volunteers.

The Icelandic case, which is not what it is said to be

The trials conducted in Iceland between 2015 and 2019 reduced weekly hours from forty to around thirty-five or thirty-six, with pay maintained. They were not a four-day week.

The widely repeated claim that most of the country subsequently moved to a four-day week is a description of union agreements giving workers the right to negotiate reduced hours. That is a substantial change and it is a different one.

This entry is not correcting a fringe error. The misstatement has appeared in national broadcasters, in policy documents and in academic introductions, which is what makes it worth setting out.

Compressed and reduced are different arrangements

A compressed week is the same hours in fewer days: four tens instead of five eights. A reduced week is fewer hours for the same pay.

They have different effects on fatigue, on childcare, on coverage and on cost, and they are routinely discussed as one thing. Belgian legislation of 2022 gave workers a right to request the first; most coverage described it as the second.

What the results showed

Improvements in self-reported wellbeing, sleep and burnout, which were substantial and consistent. Retention improved. Revenue in participating firms was generally maintained.

The weak link is output. The model claims full output in eighty per cent of the time, and output in knowledge-based service firms is the hardest thing in this subject to measure, which the part of this site on measurement is about.

Where the question is different

In any work where the requirement is coverage rather than output. A ward needs a nurse present; a shop needs somebody at the till; a bus needs a driver. Reducing each worker's hours by a fifth while keeping coverage means employing more people, and the cost of that is the whole question.

None of the well-publicised trials answers it, and it is the setting in which most people work. Trials in health and care settings have been run and have generally reported the same finding: reduced hours require additional staff and the arrangement stands or falls on whether the additional staff can be found and afforded.

What would settle it

Randomised assignment within a sector, a control group, a measured output variable defined in advance, and two years rather than six months. That study has not been done and would be expensive.

Until then the honest summary is that the arrangement clearly works for a particular kind of employer, that the mechanism is plausible, and that generalising it to employment as a whole is unsupported.

The novelty problem

Six months is short enough that attention, goodwill and the effort people make when observed are plausible contributors to the result. That is not a suggestion of dishonesty; it is a standard difficulty with short workplace interventions.

The trials that have run longest report the improvements persisting, which is the most reassuring evidence available on this point and comes from a small number of organisations.

What to take from it

That slack exists in office work and can be removed without losing output, which is a real and useful finding. That saying so is not the same as saying it about work.

What the trials do establish

That in a range of office-based organisations, a fifth of the working time was not producing output. That is a substantial finding about how such work is organised, and it is consistent with what anybody who has worked in such an organisation would tell you privately.

The interesting question it raises is not about the length of the week but about what the fifth was being spent on, and the trials generally answered it the same way: meetings.

What happened afterwards

Most participating organisations in the published pilots continued with the arrangement after the trial ended, which is a meaningful signal because it is a decision taken with their own money and without anybody observing.

For broader context, consult four-day-week research library.