OMXUS Press
2026
This paper exists because of Goal 2 (22-hour work week) (22-hour work week):
This paper examines the economic, health, social, and environmental evidence for a 22-hour standard work week. We synthesise data from large-scale reduced-hours trials (Iceland, N=2,500; United Kingdom, 61 companies; Microsoft Japan; Perpetual Guardian, New Zealand), productivity-compensation divergence data (Economic Policy Institute), automation displacement research (Frey & Osborne, 2013; OECD, 2016), the Results-Only Work Environment experiment at Best Buy, and meta-analytic health data on overwork mortality (Kivimaki et al., 2015; WHO/ILO, 2021).
Our central finding is that the barrier to a sub-25-hour work week is distributional, not technological. Labour productivity has increased approximately 400% since 1950 across OECD countries. Between 1979 and 2023, net productivity in the United States grew 64.7% while hourly compensation for typical workers grew 14.1%. The gap -- roughly 50 percentage points of productivity growth -- represents value created by workers and captured by capital. Multiple large-scale trials confirm that reducing hours by 20-25% maintains or improves productivity in knowledge and service work. Overwork is associated with 745,000 deaths per year globally. Care work is structurally incompatible with a 40-hour paid work week. Shorter work weeks correlate with lower per-capita carbon emissions.
The 22-hour work week is not utopian. It is the redistribution of gains that have already been produced.
Keywords: working hours, productivity-pay gap, four-day week, ROWE, automation, Keynes, Parkinson's Law, overwork mortality, care economics, bullshit jobs
This paper exists because of Goal 2 (22-hour work week) (22-hour work week):
Work 22 hours max. Keep your pay. Choose your hours. Work from home.
That goal did not emerge from policy analysis. It emerged from watching people die on schedule. From watching a man work sixty hours a week to pay a mortgage on a house he never sees, to feed children he never holds, to maintain a marriage that collapses under the weight of his absence. He is doing everything right. He is loyal. He works hard. He provides. And the system he is loyal to, works hard for, and provides through is extracting roughly fifty percentage points of his productivity growth as profit while handing him back fourteen. He is not lazy. He is being robbed in broad daylight by an arrangement so familiar it looks like nature.
The automation already did the work. Labour productivity in OECD countries has increased roughly 400% since 1950. If those gains had been shared as reduced hours rather than captured as shareholder returns, the standard work week would already be under 20 hours. The 22-hour target is not aspirational. It is the hours-equivalent of the productivity that already exists, if gains were shared.
But this is not merely an argument about hours. The extra 20+ hours per week freed by reduced work is the precondition for everything else. You cannot run a direct democracy (Goal 1 (direct democracy) (direct democracy)) with citizens who work 40 hours, commute for 10, and have nothing left. You cannot build community emergency response (Goal 13 ($29 emergency ring) ($29 emergency ring)) with neighbours who never see each other. You cannot have mateship when every structure that let you see your mates has been replaced by a subscription service. The 40-hour week is not just inefficient. It is the operating system of isolation. Reduce the hours, and the capacity for self-governance, mutual aid, and human connection is restored.
Results-Only Work Environment. The gains go to workers, not shareholders. The extra 20+ hours per week is what makes community governance possible.
This paper presents the evidence. All of it. And the evidence says what it says.
This paper examines the economic, health, social, and environmental evidence for a 22-hour standard work week. We synthesise data from large-scale reduced-hours trials (Iceland, N=2,500; United Kingdom, 61 companies; Microsoft Japan; Perpetual Guardian, New Zealand), productivity-compensation divergence data (Economic Policy Institute), automation displacement research (Frey & Osborne, 2013; OECD, 2016), the Results-Only Work Environment experiment at Best Buy, and meta-analytic health data on overwork mortality (Kivimaki et al., 2015; WHO/ILO, 2021).
Our central finding is that the barrier to a sub-25-hour work week is distributional, not technological. Labour productivity has increased approximately 400% since 1950 across OECD countries. Between 1979 and 2023, net productivity in the United States grew 64.7% while hourly compensation for typical workers grew 14.1%. The gap -- roughly 50 percentage points of productivity growth -- represents value created by workers and captured by capital. Multiple large-scale trials confirm that reducing hours by 20-25% maintains or improves productivity in knowledge and service work. Overwork is associated with 745,000 deaths per year globally. Care work is structurally incompatible with a 40-hour paid work week. Shorter work weeks correlate with lower per-capita carbon emissions.
The 22-hour work week is not utopian. It is the redistribution of gains that have already been produced.
Keywords: working hours, productivity-pay gap, four-day week, ROWE, automation, Keynes, Parkinson's Law, overwork mortality, care economics, bullshit jobs
In 1930, John Maynard Keynes published "Economic Possibilities for Our Grandchildren," predicting that by 2030, technological progress would make a 15-hour work week sufficient. The essay imagined that the "economic problem" -- scarcity -- would be solved within a century, and that the main challenge would be figuring out what to do with all the leisure.
Keynes was right about productivity. Labour output per hour in OECD countries has increased roughly 400% since 1950 (Bureau of Labor Statistics; OECD Productivity Statistics Database). In the United States, economic output per hour of labour more than quadrupled between 1948 and 2024. If those gains had been distributed as reduced working hours rather than increased output and profit, the standard work week would be well under 20 hours.
But productivity gains were not distributed as reduced hours. They were captured almost entirely as increased profits and executive compensation. The Economic Policy Institute's annual "Productivity-Pay Gap" report documents the divergence: from 1948 to 1973, productivity and typical worker compensation grew roughly in tandem. After 1973, they diverged sharply. Between 1979 and 2023, net productivity grew approximately 64.7%, while hourly compensation for typical workers grew only 14.1% (Bivens & Mishel, 2023). The gap -- roughly 50 percentage points of productivity growth -- represents value created by workers and captured by capital.
The 22-hour work week is not a utopian proposal. It is the hours-equivalent of the productivity that already exists, if gains were shared.
This paper asks a simple question: what does the evidence say about radically reducing the standard work week? The answer, drawn from national-scale trials, meta-analyses, historical data, and automation economics, is that the evidence is overwhelming, consistent, and largely ignored.
The reason it is ignored is not evidentiary. It is political. Reduced working hours threaten the extraction mechanism that has defined post-1973 capitalism: the capture of productivity gains as profit. Understanding why the evidence is ignored requires understanding whose interests are served by the current arrangement. That analysis is the subject of this paper's later chapters, and of the companion research on economic servitude and cooperative capitalism.
The history of working hours is a history of struggle, not of natural evolution.
Pre-industrial agricultural workers had irregular schedules, but during peak seasons worked 12-16 hours daily. E.P. Thompson's "Time, Work-Discipline, and Industrial Capitalism" (1967, Past & Present) documented how industrial capitalism imposed clock-time discipline on populations accustomed to task-oriented work. Factory owners needed workers to arrive at fixed hours, work continuously, and accept the rhythm of the machine rather than the rhythm of the body.
The imposition of clock-time was not a neutral technological development. It was an act of power. Thompson documented the destruction of pre-industrial work patterns -- seasonal, task-oriented, interspersed with rest and socialising -- and their replacement with the factory clock. The bell that rang at 6am did not measure time. It measured obedience.
Early industrial work weeks were 60-80 hours. The fight for shorter hours was one of the central labour struggles of the 19th and early 20th centuries:
The 40-hour week was not discovered through optimisation. It was not the output of a productivity study. It was not recommended by economists. It was won through decades of strikes, organising, and political struggle. People died for it. And then the trend stopped.
Since the 1970s, average working hours in most OECD countries have declined only modestly, and in the United States have been roughly flat or increased for many workers. The standard full-time week remains 38-40 hours despite the fact that productivity per hour has more than doubled since the 40-hour standard was established.
Consider what this means. In 1938, the 40-hour week represented a political settlement: this many hours of your life belong to the employer, the rest belong to you. Since then, the productive output of each of those hours has more than doubled. The employer is getting twice the output. The worker is getting the same number of hours back. The settlement has been violated, and the violator has been rewarded.
Why did the trend stop? Several factors converge:
The 40-hour week is not a law of nature. It is a political artefact from 1938, based on productivity levels that have since quadrupled, maintained by the interests of those who capture the difference.
The central economic fact of the past fifty years, the fact that explains more about contemporary life than any other, is the divergence between productivity and compensation.
The Economic Policy Institute has tracked this divergence since 1948. The data is publicly available, updated annually, and not seriously disputed by any school of economics. The disagreements are about interpretation, not about the numbers.
The numbers: between 1948 and 1973, net productivity and hourly compensation for typical (non-supervisory) workers grew in approximate tandem -- productivity up 96.7%, compensation up 91.3%. Workers received roughly what they produced.
After 1973, the lines diverge. Between 1979 and 2023:
The 22-hour target is, in this context, a compromise. It asks for less than what the productivity data says workers have already earned. It accounts for the fact that not all productivity gains can translate perfectly into reduced hours, that some sectors have different productivity profiles, and that the transition would need to be managed. But make no mistake: the 22-hour week is not a demand for something new. It is a demand for the return of something taken.
Multiple large-scale trials of reduced work weeks have now been conducted. The results are remarkably consistent. They are consistent enough to constitute a robust evidence base, and they say the same thing: reducing hours by 20-25% does not reduce productivity proportionally, and in many cases does not reduce it at all.
The largest and longest trial to date. Reykjavik City Council and the Icelandic government ran two trials covering approximately 2,500 workers (more than 1% of Iceland's working population) across offices, hospitals, schools, and social services. Workers moved from 40-hour to 35-36 hour weeks with no reduction in pay.
The results, analysed by Autonomy and the Association for Sustainability and Democracy (ALDA) and published in 2021, found that productivity was maintained or improved across nearly all workplaces, while worker wellbeing improved substantially on measures of stress, burnout, and work-life balance.
Following the trials, Icelandic unions negotiated shorter working weeks for 86% of the country's workforce.
This is not a pilot. This is a national-scale structural change. Eighty-six percent of a country's workers now work shorter hours because the evidence said it worked. The trial lasted four years, covered multiple sectors, and was evaluated by independent researchers. It is as close to a natural experiment as labour economics gets.
61 companies and approximately 2,900 workers participated in a six-month trial coordinated by 4 Day Week Global, with research led by Boston College, Cambridge University, and Autonomy. Companies moved to a four-day work week (32 hours) at full pay.
Results:
These are not marginal effects. A 65% reduction in sick days. A 57% reduction in turnover. These represent massive cost savings to employers -- savings that, in the current arrangement, are achieved through overworking people into illness and then replacing them when they leave.
A five-week trial of a four-day week at Microsoft Japan found productivity (measured as sales per employee) increased by 39.9%. Electricity consumption dropped 23.1%. Pages printed dropped 58.7%.
The trial was short and in a single company, so the generalisability is limited, but the magnitude of the productivity effect demands explanation. A 40% productivity increase from a 20% reduction in hours means the lost day was not just unproductive -- it was actively counterproductive. Workers were more effective in four days than in five.
A financial services firm of 240 employees trialled a four-day week over two months. Researchers from the University of Auckland found that productivity was maintained, work-life balance scores improved from 54% to 78%, and stress levels decreased from 45% to 38%. The company made the change permanent.
The Spanish government launched a pilot program offering companies subsidies to move to a 32-hour, four-day work week without reducing pay. Results from the initial phases showed improved worker satisfaction and no decline in output, though the full evaluation is pending.
The pattern across these trials is consistent enough to draw firm conclusions. Reducing hours by 20-25% does not reduce productivity proportionally, and in many cases does not reduce it at all.
The reason is straightforward -- in most knowledge and service work, the binding constraint is not hours but attention, energy, and motivation. Workers in a 40-hour week waste substantial time on low-value activity, presenteeism, and inefficiency driven by fatigue. When hours are cut, workers compress their productive work into fewer hours, cut the waste, and maintain output.
This is not surprising to anyone who has worked in an office. It is only surprising to managers who measure input (hours at desk) rather than output (work completed). The 40-hour week persists not because it is optimal, but because it is measurable, controllable, and familiar.
The most dramatic evidence for radically flexible work may come from Best Buy's corporate headquarters. In the early 2000s, Cali Ressler and Jody Thompson, HR innovators at Best Buy, created what they called a Results-Only Work Environment (ROWE). Under ROWE, employees had complete freedom over when, where, and how they worked. The only requirement was that they delivered their agreed results. No mandatory meetings. No set hours. No expectation of physical presence. You could work at 2am from a coffee shop or take Wednesday off to go fishing, as long as your work got done.
The results, documented in Ressler and Thompson's 2008 book Why Work Sucks and How to Fix It, were striking:
These are internal company metrics, not peer-reviewed findings, and should be treated with that caveat. But the scale of the effects attracted significant academic and media attention. A 35% productivity increase is not a rounding error. A 90% reduction in turnover is not a mood swing. These are structural effects that suggest the standard model of work -- fixed hours, fixed location, fixed schedule -- is not just suboptimal but actively destructive.
ROWE was cancelled in 2013 when new CEO Hubert Joly eliminated the programme, requiring employees to return to the office. The stated reason was a need for "all hands on deck" during a difficult business period.
The pattern -- a successful flexibility programme killed by a new leader who prefers control -- has repeated across organisations. Yahoo's Marissa Mayer eliminated remote work in 2013 under similar reasoning. The evidence suggests that the decision to end flexibility programmes is driven by managerial preference for visibility and control, not by evidence about productivity.
This is worth sitting with. The data says flexibility works. The people who cancel flexibility programmes do not cite data. They cite "culture," "collaboration," and the need for "all hands on deck." These are not empirical claims. They are assertions of authority dressed in business language. The manager who requires butts-in-seats despite evidence that remote workers are more productive is not making a business decision. They are making a power decision. The office is not a productivity tool. It is a control apparatus.
ROWE represents the most radical version of the reduced-hours thesis: do not reduce hours to 32 or 22. Eliminate the concept of hours entirely. Measure results. Let adults manage their own time. The evidence from Best Buy says this works better than any fixed-hours arrangement. And it was killed anyway.
In 1955, C. Northcote Parkinson published a short essay in The Economist that contained one of the most durable insights in organisational theory: "Work expands so as to fill the time available for its completion."
Parkinson's Law was presented humorously, but the underlying observation is empirically supported by every reduced-hours trial in the literature. When workers move from a 40-hour week to a 32-hour week, they do not produce 80% of their previous output. They produce 95-100% or more. The reason is that the 40-hour week contains substantial slack -- extended meetings, unnecessary email chains, social media browsing, recovery time from fatigue, context-switching overhead, and the general inefficiency of trying to sustain focused attention for eight hours in an environment designed for control rather than output.
The corollary is equally important: work contracts to fill the time it is given. When the constraint tightens, waste is eliminated first. Meetings become shorter or disappear. Email becomes concise. Decisions are made faster because there is no time for the performative deliberation that characterises most organisational decision-making.
This is not a theory. It is what happened in Iceland, in the UK trial, at Microsoft Japan, at Perpetual Guardian. In every case, the mechanism was the same: workers, given fewer hours, identified and eliminated the waste that had been invisible when time was abundant.
The implications are profound. If Parkinson's Law is correct -- and the trial evidence says it is -- then the standard 40-hour week is not 40 hours of work. It is some lesser number of hours of work expanded to fill a 40-hour container. The container is the problem, not the contents.
This connects directly to David Graeber's "bullshit jobs" thesis (see Chapter 9). If 37-40% of workers consider their jobs meaningless, and if the meaningful work that remains can be done in fewer hours than are currently mandated, then the standard work week is doubly inflated: inflated by meaningless roles and inflated by the expansion of meaningful work to fill unnecessary hours. Reduce the roles and reduce the hours, and you arrive at something close to 22 hours of actual, productive, meaningful work per week.
In 2013, Carl Benedikt Frey and Michael Osborne published "The Future of Employment: How Susceptible Are Jobs to Computerisation?" estimating that 47% of US employment was at high risk of automation within one to two decades. The paper has been enormously influential and also extensively critiqued -- the OECD's 2016 reanalysis using a task-based approach estimated only 9% of jobs at high risk, while acknowledging that many more would be significantly transformed.
The exact figure matters less than the direction. Automation is replacing human labour in an expanding range of activities -- manufacturing, logistics, data entry, basic legal work, customer service, and increasingly cognitive tasks through AI. The arrival of large language models in 2022-2025 has accelerated this timeline dramatically. Tasks that were considered safely "human" -- writing, analysis, code generation, customer interaction, even medical diagnosis -- are now performed by AI systems at lower cost and, in many cases, comparable quality.
The question is not whether automation reduces the total labour needed to produce current levels of goods and services. It does. The question is who captures the value that freed labour represents.
Under current arrangements, automation primarily benefits capital owners. When a factory replaces 100 workers with machines, the productivity gain flows to shareholders as profit, not to workers as leisure. The workers are simply displaced -- forced to find new employment, often at lower wages, or to exit the workforce entirely. This is the mechanism by which Keynes's productivity prediction came true while his leisure prediction did not.
The standard economic response -- "new jobs will be created" -- is historically accurate but increasingly strained. The jobs created by the digital economy are often lower-paid, less stable, and more precarious than the jobs they replaced. The gig economy is not a new form of work. It is an old form of exploitation with a smartphone interface. A worker who drives for Uber has less security, lower pay, and fewer rights than the taxi driver they replaced. The productivity gain went to Uber's shareholders. The worker got an app.
The 22-hour work week is, in this framing, a distributional question, not a technological one. The technology to support it already exists. What does not exist is the institutional mechanism to distribute automation's gains as reduced hours rather than as unemployment for some and intensified work for others.
The mechanism is simple in principle: as automation increases output per hour, the hours required to produce a given standard of living decrease. If those gains are shared as reduced hours at the same pay, automation becomes a liberating force. If they are captured as profit, automation becomes a displacement force. The current arrangement chooses displacement. A 22-hour work week chooses liberation.
The companion research on cooperative capitalism (Applebee & Combe, 2026) and the $19 trillion solution (Applebee & Combe, 2026) describes the ownership and distributional mechanisms that would make this choice possible. The present paper establishes that the productivity exists. The question of who captures it is addressed there.
Overwork is not merely unpleasant. It is measurably dangerous. It kills people. The evidence is extensive, meta-analytic, and published in the highest-tier medical journals.
Kivimaki et al. (2015) published a meta-analysis in The Lancet covering over 600,000 individuals across multiple cohort studies. The findings: working 55 or more hours per week, compared with standard hours (35-40), was associated with a 33% increased risk of stroke (95% CI: 1.11-1.61) and a 13% increased risk of coronary heart disease (95% CI: 1.02-1.26). These are dose-response relationships -- risk increases progressively with hours worked beyond the standard.
This is not correlation fishing. This is a meta-analysis in The Lancet covering more than half a million individuals. The confidence intervals exclude one. The effect is real.
The WHO and ILO published a joint estimate in 2021 (Environment International) finding that long working hours (55+ per week) were associated with an estimated 745,000 deaths per year globally from stroke and ischemic heart disease -- making overwork one of the largest occupational risk factors for mortality worldwide.
Seven hundred and forty-five thousand deaths per year. From working too much. This is not a footnote. This is a catastrophe on the scale of a major disease, caused not by a pathogen but by an economic arrangement. If a virus killed 745,000 people a year, we would mobilise global resources to stop it. When an economic system kills 745,000 people a year, we call it "the economy" and discuss GDP growth.
Japan has an official term for death from overwork: karoshi. The Japanese government recognises karoshi as a cause of death in workers' compensation claims, primarily from stroke and heart attack associated with excessive working hours. The government's own White Paper on Karoshi (published annually since 2016) documents hundreds of recognised cases per year, with many more suspected but unrecognised.
The existence of an official government category for "death from overwork" should be sufficient evidence that the 40-hour week -- let alone the 55-hour weeks common in many countries -- is a health hazard. That it is not sufficient, that the category exists and the practice continues, tells us that the arrangement is maintained by something other than evidence.
Beyond cardiovascular disease, overwork is associated with increased rates of depression, anxiety, sleep disorders, substance abuse, and relationship breakdown. A systematic review by Virtanen et al. (2018, The Lancet) found that long working hours were associated with depressive symptoms, with the relationship being strongest for employees working more than 55 hours per week.
The mental health effects are compounding. Overwork causes stress. Stress disrupts sleep. Disrupted sleep reduces cognitive function. Reduced cognitive function requires more hours to complete the same work. More hours cause more stress. The cycle is self-reinforcing, and it terminates in burnout, breakdown, or death.
The evidence is extensive enough that the health argument alone would justify reduced working hours, independent of the productivity and distributional arguments. If the 40-hour week were a new drug seeking regulatory approval, the adverse event profile would prevent it from reaching the market. The fact that it is an inherited economic arrangement rather than a pharmaceutical product means it is subject to no such scrutiny.
A 22-hour work week would place the standard well below the 35-hour threshold where cardiovascular risk begins to increase. It would leave sufficient time for sleep, exercise, social connection, and the other determinants of health that are currently crowded out by paid employment. It would, in the most literal sense, save lives.
In 2013, David Graeber published a short essay titled "On the Phenomenon of Bullshit Jobs" that became the most-read article in the history of STRIKE! Magazine. The essay argued that a significant proportion of paid employment in advanced economies is meaningless -- that the workers themselves know it is meaningless, and that this knowledge produces profound psychological damage.
Graeber expanded the essay into a full book in 2018, developing a five-part taxonomy of bullshit jobs:
A YouGov poll found that 37% of British workers said their job made no meaningful contribution to the world. A Dutch survey using a more conservative methodology (Sofie Graeber) estimated approximately 5%. The discrepancy reflects question framing, not a refutation of the phenomenon -- the YouGov question asked about meaningful contribution, while the Dutch study asked whether the job was "useless," a higher bar.
The full analysis of bullshit jobs is presented in the companion paper (Applebee & Combe, "Bullshit Jobs: Quantifying Meaningless Work," 2026). For the purposes of this paper, the relevant finding is this: a substantial proportion of working hours are devoted to tasks that the workers themselves recognise as meaningless.
Combined with Parkinson's Law (the meaningful work that remains expands to fill the available time), the bullshit jobs thesis supports the 22-hour work week from the demand side: not only can productive work be done in fewer hours (the trial evidence), but a significant proportion of current hours are not productive work at all. They are performative labour maintained by managerial ideology, institutional inertia, and the absence of any mechanism to redistribute the freed hours.
The standard 40-hour work week assumes a worker with no other responsibilities. This assumption was never true and is less true today than ever.
The International Labour Organization estimates that globally, women perform 76.2% of total unpaid care work -- childcare, eldercare, housework, emotional labour. In OECD countries, women spend on average 4.5 hours per day on unpaid work compared to 2.3 hours for men (OECD Time Use Surveys). This work is invisible in GDP calculations, uncompensated in the labour market, and structurally incompatible with a 40-hour paid work week.
The 40-hour week was designed for a worker who had someone at home doing the care work. In the mid-20th century, this was the explicit social model: the male breadwinner worked, the female homemaker cared. The model was never universal -- working-class women and women of colour have always worked -- but it was the ideological basis for labour market design.
That model has collapsed. In most OECD countries, both adults in a household now work. But the labour market has not been redesigned. It still assumes a worker with no caring responsibilities. The result is that women (primarily) perform a double shift: 40 hours of paid work plus 4.5 hours per day of unpaid care work. This is approximately 71.5 hours of total work per week. The "work-life balance" discourse is a euphemism for an arrangement in which one life's worth of work has been distributed across one-and-a-bit people's waking hours.
A 22-hour work week creates space for care work to be integrated into adult life without the current binary: either you work full-time and outsource care (if you can afford it), or you provide care and forfeit economic participation. Reduced working hours are, implicitly, a feminist economic policy -- not because they are designed specifically for women, but because they break the structural assumption that the "standard worker" has no caring responsibilities.
The 22-hour week also makes visible what is currently hidden. If paid work occupies 22 hours, the remaining time is available for care, community, governance, and leisure. These are not secondary activities. They are the substance of human life. The current arrangement treats them as residual -- what's left after the employer has had their share. A 22-hour week reverses the hierarchy: work serves life, not the other way around.
Knight, Rosa, and Schor (2013, "Could working less reduce pressures on the environment?") found a statistically significant relationship between working hours and ecological footprint across OECD countries. Countries with shorter work weeks had lower per-capita carbon emissions, controlling for GDP per capita. The mechanisms are multiple:
Juliet Schor's broader body of work, including Plenitude (2010, later retitled True Wealth), argues that reducing work hours is one of the most effective climate interventions available -- not because it directly reduces emissions, but because it shifts the economy from a growth-maximising to a sufficiency-oriented model. The current arrangement -- in which people work long hours to earn money to consume goods they don't have time to enjoy -- is both ecologically destructive and personally unsatisfying. Reducing hours addresses both problems simultaneously.
Microsoft Japan's trial provides concrete data: a four-day week reduced electricity consumption by 23.1% and paper printing by 58.7%. These are facility-level effects; the commuting and consumption effects would add to them.
This is not to say that reduced work hours automatically solve climate change. But the current model is a heat engine: it converts human time into economic output into consumption into waste, at a rate that exceeds the planet's capacity to absorb. Slowing the engine -- by working less, consuming less, and living more -- is not sufficient, but it is necessary.
If the evidence for reduced working hours is this strong -- and it is -- why has nothing changed? Why is the 40-hour week still the standard in most countries, more than 85 years after it was codified?
The answer is not ignorance. The trials have been covered extensively in mainstream media. The Iceland results were global news. The UK pilot was featured in every major outlet. The evidence is available. It is not hidden.
The answer is structural. Reduced working hours threaten several interlocking interests:
1. Employer control. The office is a control apparatus. Hours at desk are a proxy for obedience, not productivity. Managers who measure input (presence) rather than output (results) lose their primary metric when hours are reduced. The resistance to reduced hours is, in many organisations, resistance to being measured on what you produce rather than how long you sit there. This is why ROWE was killed at Best Buy: not because it failed, but because it succeeded in a way that made management's traditional function visible as unnecessary.
2. The extraction mechanism. The productivity-pay gap is the central extraction mechanism of post-1973 capitalism. If workers received the productivity gains as reduced hours, the gap would close -- not through higher wages, but through less labour for the same output. This would redistribute value from capital to workers on a scale not seen since the post-war social contract. The beneficiaries of the current arrangement -- shareholders, executives, the financial sector -- have every incentive to prevent this.
3. Work-as-identity ideology. The most insidious resistance comes not from employers but from workers themselves. "Hustle culture," the equation of long hours with virtue, the Protestant work ethic secularised into LinkedIn posts about "grinding" -- these are ideological formations that serve employer interests while being experienced as personal values. The worker who says "I define myself by my work" is saying "I have internalised the employer's claim on my time as my own identity." This makes reduced hours feel threatening rather than liberating -- not because the evidence is wrong, but because identity is at stake.
4. The consumption treadmill. The current arrangement requires high income to fund high consumption. Long hours generate income that funds the consumption that compensates for the misery of long hours. Reduce the hours, and the cycle breaks -- but breaking it requires confronting the possibility that the consumption was not satisfying in the first place, that the purchases were substitutes for the time and connection they displaced.
5. Housing and debt. In countries where housing costs have decoupled from wages -- Australia at 14x median income, the UK, Canada, much of the US -- workers cannot reduce hours without reducing income, and cannot reduce income without losing their homes. The housing market is a mechanism for locking workers into long hours regardless of productivity evidence. This is addressed in the companion research on economic servitude (Applebee & Combe, 2026) and housing-first policy.
6. Political capture. The industries that benefit from long working hours are also the industries that fund political campaigns, staff regulatory agencies, and shape policy discourse. The revolving door between corporate management and government ensures that policies promoting reduced hours face structural opposition from the people responsible for implementing them.
None of these barriers is evidentiary. The evidence is clear. The barriers are political, structural, and ideological. Overcoming them requires not more evidence but different power -- which is why the 22-hour work week is not a standalone policy proposal but part of an integrated system that includes direct democracy (Goal 1 (direct democracy) (direct democracy)), cooperative ownership (cooperative capitalism research), and the redistribution of national wealth ($19 trillion solution).
The evidence establishes that a 22-hour work week is feasible, beneficial, and supported by trial data. The question is how to get there. The existing trials provide several implementation models:
Iceland's success was driven by public-sector trials followed by union negotiation. The government ran the trial, the evidence accumulated, and unions used it to negotiate shorter hours across the workforce. This model works where unions are strong and public-sector employment is significant.
Applicable to: Nordic countries, countries with strong union movements. Limited by: union decline in Anglophone countries.
The UK pilot relied on voluntary employer participation. Companies that tried it mostly continued. This creates a competitive dynamic: companies offering four-day weeks attract better talent, have lower turnover, and save on sick days. Over time, holdouts lose talent to adopters.
Applicable to: knowledge-economy sectors with talent competition. Limited by: low-wage sectors where workers lack bargaining power.
The historical precedent is the Fair Labor Standards Act of 1938, which mandated the 40-hour week. A similar legislative act could mandate a 32-hour week (as a first step) or a 22-hour week (as the target), with overtime protections ensuring that pay does not decrease.
Applicable to: any country with functioning legislative processes. Limited by: political capture (see Chapter 12).
The most radical model: do not set a number of hours. Measure results. Let workers manage their own time. This is what ROWE demonstrated. It requires a fundamental shift in managerial culture from input measurement to output measurement.
Applicable to: knowledge work, creative work, any role where output is measurable. Limited by: roles with coverage requirements (healthcare, emergency services, retail).
The OMXUS model: citizens vote directly on work-hour policy. No representatives. No lobbyists. No campaign donors. If 86% of Icelandic workers preferred shorter hours after the trial, and the trial evidence supports shorter hours, then a direct vote would produce shorter hours. The barrier is not public opinion. It is the intermediation of representatives who are captured by employer interests.
This is Goal 1 (direct democracy) (direct democracy) (direct democracy) enabling Goal 2 (22-hour work week) (22-hour work week) (22-hour week). The goals are not a list. They are a system.
The transition from 40 to 22 hours need not be immediate. A phased approach:
The transition is funded by the productivity gap: the 50 percentage points of productivity growth that have been captured as profit since 1979 are the funding source. Reduced hours do not require new productivity. They require the redistribution of existing productivity.
The evidence base for reduced working hours is now large enough to draw firm conclusions:
The 22-hour work week is not utopian. It is the hours-equivalent of the productivity that already exists, if gains were shared. The automation already did the work. The trials proved it works. The health data proves the current arrangement kills. The only question is whether the gains belong to the people who produced them or to the people who captured them.
The answer to that question is not economic. It is moral. And the evidence makes the moral question very simple.
| Source | Confidence | Notes |
|---|---|---|
| Bivens & Mishel / EPI | HIGH | Updated annually, publicly available data, standard reference |
| Iceland trial (Autonomy/ALDA) | HIGH | Large-scale (N=2,500), multi-year (4 years), multi-sector, public sector |
| UK pilot (4 Day Week Global) | HIGH | 61 companies, 2,900 workers, academic research partners (Boston College, Cambridge) |
| Microsoft Japan | MEDIUM | Single company, 5-week trial, internal metrics. Magnitude (39.9%) striking but generalisability limited |
| Perpetual Guardian (NZ) | MEDIUM-HIGH | Academic evaluation (University of Auckland), made permanent, but single company |
| ROWE / Best Buy | MEDIUM | Internal company data, not peer-reviewed. Scale of effects (+35%/-90%) extraordinary but unverified externally |
| Frey & Osborne 47% | MEDIUM-HIGH | Widely cited, published in peer-reviewed journal (2017). OECD reanalysis found 9% using different methodology. Directional finding robust |
| Kivimaki et al. | HIGH | Meta-analysis in The Lancet, N=603,838. Gold-standard evidence |
| WHO/ILO 745K deaths | HIGH | Joint institutional estimate, peer-reviewed journal (Environment International) |
| Keynes (1930) | HIGH | Primary source, prediction verifiable against actual productivity data |
| Graeber 37% bullshit jobs | MEDIUM | YouGov poll, not academic survey. Dutch replication (5%) used different methodology. Directional finding robust |
| 400% productivity increase | MEDIUM-HIGH | Approximate aggregate across OECD. Exact figure varies by country, sector, base year, and measurement method. BLS data for US is precise; cross-country aggregate is estimated |
| Knight, Rosa & Schor (2013) | HIGH | Peer-reviewed, cross-national panel data, published in Global Environmental Change |
| ILO 76.2% care work | HIGH | Institutional estimate, comprehensive global survey |
| Spain pilot | LOW-MEDIUM | Ongoing, preliminary results only. Watch for final evaluation |
This paper is part of an integrated research programme. The 22-hour work week cannot be understood in isolation -- it connects to the distributional mechanisms that keep hours high, the meaningless work that inflates them, the economic structures that could replace them, and the governance systems that could implement them.
| Paper | Location | Relationship to This Paper |
|---|---|---|
| Bullshit Jobs: Quantifying Meaningless Work | ../bullshit_jobs/ | If 37-40% of jobs are meaningless (Graeber taxonomy: flunkies, goons, duct-tapers, box-tickers, taskmasters), eliminating them reduces the total labour requirement. Combined with Parkinson's Law, this brings the productive core close to 22 hours. ~16,000 words, complete. |
| The $19 Trillion Solution | ../nineteen_trillion/ | Models the redistribution of Australia's $19.4T national wealth through a Sovereign Equity Fund. The universal basic dividend ($800/week) supplements reduced wages during the transition to shorter hours. Without a wealth redistribution mechanism, workers locked into mortgage debt cannot reduce hours. ~113,000 words, complete. |
| 2000 Years of Economic Servitude | ../economic_servitude/ | Traces extraction mechanisms from ancient slavery to modern debt-based finance. The productivity-pay gap (Chapter 3 of this paper) is the current form of the same extraction. The 40-hour week is the modern equivalent of tied tenancy: you cannot leave because your housing depends on your income, and your income depends on your hours. Includes UBI paper (~15,000 words). Complete. |
| Cooperative Capitalism | ../cooperative_capitalism/ | Proposes the ownership model (cooperative, distributed) that would distribute automation gains as reduced hours rather than as profit. Mondragon (81,000 worker-owners, 65 years) demonstrates that cooperative structures sustain employment and distribute gains more equitably. ~69,000 words, complete. |
| Paper | Location | Connection |
|---|---|---|
| Housing First | ../housing_first/ | Housing at 14x median income locks workers into long hours. Housing-first policy is a precondition for reduced hours, not a separate issue. |
| Drug Policy Reform | ../drug_policy_reform/ | Body sovereignty. The worker who self-medicates with alcohol after a 50-hour week is treating a symptom of an economic arrangement, not a personal failing. |
| Swiss Direct Democracy | ../democratic_voting_mechanisms/ | The governance mechanism for implementing reduced hours. Citizens vote directly; no representative captured by employer lobbying stands between the evidence and the policy. |
| Emergency Response | ../emergency_response/ | Goal 13 ($29 emergency ring) ($29 emergency ring) ($29 ring, community response in 60 seconds) requires people who are present -- physically in their community, not at a desk 40km away. The 22-hour week is the precondition for community presence. |
| Education (Prussian Model) | ../education_prussian_model/ | The school system was designed to produce factory workers who accept clock-time discipline (Thompson, 1967). Redesigning education (Goal 12 (play-based education) (play-based education)) and redesigning work (Goal 2 (22-hour work week) (22-hour work week)) are the same project. |
| Sanctuary Design Thesis | ../sanctuary_design_thesis/ | The grief-to-design framework that generated all 14 goals, including Goal 2 (22-hour work week) (22-hour work week). The man who works 60 hours and dies of a stroke is not an abstraction. He is the referent. |
These papers are not a collection. They are a system. The 22-hour work week requires:
Remove any one component and the others are weakened. The 22-hour week without wealth redistribution leaves workers unable to pay their mortgages. Wealth redistribution without cooperative ownership leaves the gains captured by new owners. Cooperative ownership without direct democracy leaves the policy vulnerable to political reversal. The system is the solution.
| Trial | Country | N (workers) | Duration | Hours Reduction | Productivity | Revenue | Continuation Rate |
|---|---|---|---|---|---|---|---|
| Reykjavik/Government | Iceland | 2,500 | 4 years (2015-2019) | 40 -> 35-36 hrs | Maintained/improved | N/A (public sector) | 86% of national workforce |
| 4 Day Week Global | UK | 2,900 | 6 months (2022) | 40 -> 32 hrs | Maintained | +1.4% average | 92% of companies |
| Microsoft | Japan | ~2,300 | 5 weeks (2019) | 5-day -> 4-day | +39.9% (sales/employee) | N/A | Trial only |
| Perpetual Guardian | New Zealand | 240 | 2 months (2018) | 5-day -> 4-day | Maintained | N/A | Made permanent |
| Government pilot | Spain | TBD | Ongoing (2023-) | 40 -> 32 hrs | No decline (preliminary) | N/A | Ongoing |
| ROWE (Best Buy) | USA | ~4,000 | ~8 years (2003-2013) | No fixed hours | +35% | N/A | Cancelled by new CEO |
| Period | Productivity Growth | Compensation Growth (typical workers) | Gap |
|---|---|---|---|
| 1948-1973 | +96.7% | +91.3% | ~5 points |
| 1979-2023 | +64.7% | +14.1% | ~50 points |
| Post-1948 cumulative | ~400%+ | ~150% | ~250+ points |
Source: Economic Policy Institute, 2023
| Outcome | Hours Threshold | Risk Increase | Source |
|---|---|---|---|
| Stroke | 55+ hrs/week | +33% (95% CI: 1.11-1.61) | Kivimaki et al., 2015 (Lancet) |
| Coronary heart disease | 55+ hrs/week | +13% (95% CI: 1.02-1.26) | Kivimaki et al., 2015 (Lancet) |
| Global deaths (stroke + IHD) | 55+ hrs/week | 745,000/year | WHO/ILO, 2021 |
| Depressive symptoms | 55+ hrs/week | Significant association | Virtanen et al., 2018 |
| Year | Standard Work Week | Key Event |
|---|---|---|
| Pre-1800 | 60-80 hrs (industrial) | Factory system imposes clock-time |
| 1817 | N/A | Robert Owen proposes 8-hour day |
| 1886 | N/A | Haymarket affair (8-hour day rally) |
| 1914 | 48 hrs (Ford) | Ford adopts 8-hour day, 5-day week |
| 1938 | 40 hrs (US law) | Fair Labor Standards Act |
| 1970s-present | 38-40 hrs | Trend stalls despite 2x+ productivity growth |
| Base Year | Productivity Multiple (to 2024) | Implied Hours (at shared gains) |
|---|---|---|
| 1938 (40-hr standard) | ~4x | ~10 hours |
| 1950 | ~4x | ~10 hours |
| 1979 | ~1.65x | ~24 hours |
| Target | N/A | 22 hours (compromise accounting for transition) |
The 22-hour target is conservative. The raw productivity data supports a work week well below 20 hours.
This unified thesis compiled from the labor economics research module of the OMXUS Research Programme. All source materials preserved in the labor_economics_22hr_week/ directory. Cross-references: ../bullshit_jobs/, ../economic_servitude/, ../cooperative_capitalism/, ../nineteen_trillion/.
The automation already did the work. The trials proved it works. The health data proves the current arrangement kills. The only question is who the gains belong to.