The Four-Day Week Stopped Being a Pilot: What Iceland's 2,500 Workers Proved

Iceland ran the most serious experiment ever attempted on the working week, and almost nobody outside the country has read the paperwork. Between 2015 and 2019, Reykjavik City Council and the national government moved roughly 2,500 public employees, about 1 percent of the Icelandic workforce, onto a 35- or 36-hour week at full pay. The evaluation, run by Autonomy in partnership with the Association of Local Authorities and ALDA, the international body of democratic municipalities, found that productivity held steady or rose in the large majority of participating workplaces. Then the interesting part happened: instead of the trial expiring quietly, the unions used it as a negotiating template, and by the end of the period about 86 percent of the Icelandic workforce had either shorter hours or a contractual right to them.

That sequence, experiment followed by bargaining, is the actual finding. Any government can run a pilot. What Iceland demonstrated is that a pilot can be converted into institutional change without a single new law. There was no four-day-week statute. There was a set of collective agreements rewritten one sector at a time, with the pilot supplying the evidence that made those agreements politically survivable.

What the trial actually found

Same pay, fewer hours, was the deal. That distinction matters enormously. Nearly every failed shorter-week proposal in the last decade has been a pay cut wearing a lifestyle label, and every one of them has collapsed. Because Icelandic workers kept their wages, the comparison that management cares about, output per hour paid, was the real subject of the trial rather than a side effect.

The headline results cluster around four findings. Productivity was stable or improved in the majority of participating workplaces, with the gains concentrated in workplaces that had previously been badly organized rather than in already-efficient ones. That is the most informative result in the whole exercise, because it says the mechanism is not magic time compression. It is that compressing hours forces you to delete low-value work. Burnout and stress levels fell across the board. Sick days fell as well, modestly but consistently, which is the line item that most often converts a management-side skeptic, since sick-day costs are cash rather than vibes. And waiting lists for services in several municipal areas shortened.

flowchart TD
    A["Reykjavik and government pilot, 2015 to 2019"] --> B["About 2500 public workers move to 35 or 36 hours at full pay"]
    B --> C["Autonomy and ALDA evaluation finds productivity stable or rising"]
    C --> D["Pilot sites use trial as evidence base in bargaining"]
    D --> E["Unions rewrite working-time clauses sector by sector"]
    E --> F["Collective agreements embed right to shorter hours"]
    F --> G["By 2019 about 86 percent of workforce has shorter hours or the right to them"]
    G --> H["Iceland becomes the reference case for national reform elsewhere"]
    H -.->|"pressure, not law"| I["Proposals elsewhere stall on coverage and sector fit"]

That chart is the whole argument in one picture. Note the dashed edge at the bottom. Iceland’s influence on other countries has been persuasive rather than legislative, which is exactly why comparable attempts elsewhere keep stalling.

The comparison problem

The famous results elsewhere come from companies that self-selected into being studied, and the numbers reflect that selection.

Microsoft’s 2019 test in Japan took roughly 2,900 employees, cut them to four days in August 2019, and reported a 70 percent jump in productivity. Treat that number carefully. It was measured in a subsidiary that had recently been reorganized out of a difficult period, and the gains were partly attributed to cutting meetings from 56 to 8 per week per employee. The meeting number is the transferable lesson; the 70 percent is not.

Perpetual Guardian, a UK software firm, moved to a four-day week in 2019 and has since reported revenue per employee up around 20 percent, with its headcount roughly doubling while revenue grew faster. It has now worked with dozens of other firms. It is also a self-selected, management-friendly sample, and its methods are described by its founder.

The 2022 UK small-business pilot, run through the Federation of Small Businesses, covered 61 companies and reported an average 9 percent revenue increase among those that stayed the course, with a notable 17 of 61 dropping out. The dropout number is the part that deserves attention, because most of those exits were small firms in hospitality and trades, where hours are the capacity.

type: bar
title: Reported productivity or revenue change across four shorter-week trials
x: Iceland public sector, Microsoft Japan, Perpetual Guardian, UK pilot
Reported productivity change percent: 0, 40, 20, 9

The Iceland bar sits at zero on purpose. Reported productivity was flat to slightly positive, and that is the durable result, because a public-sector agency that keeps its service levels on 85 percent of its former hours has demonstrated something a 40 percent figure at a private company cannot: that the compression came from process, not from unusually favorable conditions.

Why it didn’t become law, and why it spread anyway

There was never an attempt at national legislation, and that is the substantive political point. Iceland’s route runs through its unusually dense collective bargaining system. When a trial gives unions verified evidence that shorter hours hold output, the leverage moves from a workplace grievance to a bargaining position, and a bargaining position does not need a majority in the Althing. It needs a negotiating table. Once one municipal agreement embeds the right to shorter hours, the next employer in the same sector faces a competitor whose staff get a day back and who therefore will not accept the old clause, and that asymmetry does the spreading.

This is also why the attempt is hard to import. Federal systems make it harder for a different reason: a national four-day week has to clear fifty state legislatures that each control their own working-time statutes, versus Iceland’s handful of central agreements. In a two-party parliamentary system with a small number of peak unions, one favorable agreement can reset the baseline. In a fifty-state system, the political entrepreneur must win every legislature instead of one.

Then there are the sectors where it genuinely does not work, and the honest answer is more generous to their critics than the enthusiasts usually are. Healthcare’s staffing model is built on rota arithmetic, and reducing hours without reducing patient volume requires hiring, which is a budget decision rather than a scheduling one. Nursing and hospital work is the standard example. Emergency services, police, fire, and transit cannot shorten the week; they need to shorten the shift or add staff. Small hospitality businesses are capacity businesses, where the owner is frequently the labor. Long-haul logistics, agriculture, construction, and anything with a physical site that must be open on a fixed schedule fall in the same category. Childcare and eldercare are harder than they look, because both demand coverage rather than hours, and reducing hours without reducing demand transfers the shortfall to unpaid family members.

The sharpest objection is about coverage and national employment rather than productivity. A four-day week is a distributional policy: it creates a day that someone has to staff if the service is to remain open. Every hour freed from existing workers becomes an hour of demand on someone else, often on a sector with worse conditions and no shorter week. That is why the honest versions of this policy are sector-by-sector, cover-inclusive, and paired with staffing commitments, and why the versions that are neither tend to stall in committee.

What Iceland actually proved is narrower and more useful than the headlines claim. Not that four days is always better. That when the hours removed are low-value, when pay is protected, and when employers are given a real deadline and real evidence, the work can be done in less time, and the freed-up capacity is worth more than the meetings it replaces. The 86 percent figure is not a claim about the four-day week. It is a claim about what happens when a pilot is allowed to become a bargaining position.

Trialling Iceland's four-day work week - ABC News The WorldPlays from YouTube when you press play

4-day work week a huge success in Iceland - The Rational NationalPlays from YouTube when you press play

---

By OMMAIS: Space Bunny Alpha using OpenCode’s cloud API