Introduction
A “four-day work week” can mean fewer paid hours at unchanged pay, forty hours compressed into four long shifts, or a partial reduction spread across the week. These are different interventions. A claim that one improves health or output cannot automatically be transferred to another. The distinction matters because the most visible recent pilots generally promise less work time, whereas much of the older four-day literature tested four ten-hour days. Historical studies examined employees' acceptance, reactions, and public-sector attitudes toward those compressed arrangements (Martin, 1981) (Fottler, 1977) (Lacapra, 1973) (Kenny, 1974) (Hodge & Tellier, 1975). The recent growth of four-day schedules in labor-force data also says little by itself about which version workers received (Hamermesh & Biddle, 2022).
This narrative review asks what the research evidence says about productivity, wellbeing, and retention under four-day work weeks. Its central conclusion is uneven: evidence of better short-term wellbeing under paid hours reduction is comparatively persuasive; reports of maintained output are encouraging but less well measured; evidence of fewer employee exits is still limited. The intervention's effect also depends on work intensity, staffing and the use employees can make of freed time. Direct trial evidence is read alongside occupational-health research and studies of compressed schedules, but the latter are identified as different designs rather than pooled into one effect. (Fan et al., 2025) (Utzet et al., 2026) (Bernstrøm et al., 2025)
Methodology
This is a narrative literature review. Searches through Crossref and OpenAlex covered “four-day workweek,” “shorter working week,” “working time reduction,” “compressed workweek,” productivity, health, and turnover, with follow-up queries for named studies and mechanisms. Fifty distinct DOI records were resolved against Crossref or DataCite. Their abstracts were retrieved where available, and records without accessible content were used only to map the literature, not to report findings. A UK pilot report without a retrieved DOI was read separately as grey literature and is identified by name and link when discussed. Selection prioritized direct interventions and reviews; older work on hours and output supplied context. The synthesis compared schedule design, evidence type and outcome measure, without treating a survey, administrative series and vignette experiment as interchangeable observations.
One global HR-outcomes evaluation was retained in the scope map, although its content was unavailable for outcome synthesis (Adediran, 2025). The mapped literature includes chronological and sustainability reviews (de Araujo & Venz, 2025) (Campbell, 2023), work-time policy arguments (Frayne et al., 2024) (Chung, 2022) (Hong et al., 2024), economic analyses of employment and distribution (Booth & Schiantarelli, 1987) (Corneo, 1995), and environmental analyses of reduced hours (Zhang et al., 2026) (Cieplinski et al., 2021). These works define the field's scope rather than establish effects on the three outcomes examined here. Some accounts discuss public services and labor-law implementation without evaluating individual outcomes (Blackham, 2025) (Kümmerling, 2026) (Hopkins et al., 2026). This distinction between background scholarship and outcome evidence governs the claims below.
Schedule Definitions and Evidence Map
Reduced-hours studies generally retain pay while cutting paid time; the intervention may shorten each day, remove a day, or rotate time off. Spanish manager interviews document several such designs, including shorter daily hours, Friday closure and rotating schedules (Gallego-Gómez & Parente, 2026). The Australian early-adopter study concerns a 100:80:100 model and reports firm accounts of redesigning workflows, while a Portuguese leadership study examines managerial perceptions of organizational feasibility (Hopkins et al., 2026) (Fontes et al., 2026). Neither estimates the effect of one standard schedule across all employers. A United Arab Emirates survey studied a four-and-a-half-day policy, with perceptions of both recovery and denser weekdays; it is another distinct case (Alawad & Smail, 2026).
By contrast, compression preserves weekly hours and extends shifts. A Belgian event study measures investors' reaction to the announcement of a compressed-week law, not subsequent worker output (Dutordoir & Struyfs, 2024). A Philippine survey models perceptions of stress, work–life balance and productivity under a compressed schedule (Paje et al., 2020); a front-line service study relates satisfaction with compressed hours to later absenteeism (Deery et al., 2016). The longitudinal review of compressed weeks found mixed health and work outcomes, including greater shift satisfaction but possible higher sickness absence. Shift length and the comparison schedule mattered (Bernstrøm et al., 2025). This is a direct reason to resist a single pooled “four-day” estimate.
The wider literature also asks different questions. Studies on weekly hours and job satisfaction (Nanda & Browne, 1977), long hours and health (Hulst, 2003), and work-time control among overtime employees (Yu & Leka, 2022) help explain why time might matter, but none by itself tests a paid four-day reduction. A paper on work-time reduction during pandemic furlough reported gender-differentiated mental-health patterns in that exceptional context (WANG et al., 2022). It should inform caution about distributional effects, rather than serve as a trial of a modern four-day policy.
Productivity
A work week can become more productive per hour while producing the same, more, or less in total. This distinction is often blurred. In a historical munitions dataset, Pencavel found a nonlinear output-hours relationship: output rose at a diminishing rate beyond an hours threshold (Pencavel, 2014). A separate study modeled both fatigue and learning effects across working-hour thresholds (Lee & Lim, 2017). These studies make it plausible that cutting low-yield hours can preserve output, but neither measured a contemporary four-day-week rollout. Recovery-based theory similarly proposes that a shorter week may improve subsequent performance through rest (Pencavel, 2016). The cognitive-science argument for recovery and motivation is a mechanism hypothesis, not a trial result (Rae & Russell, 2025).
Direct employer evidence is more provisional. In Lewis and colleagues’ UK pilot report, revenue among the firms supplying start-to-end data rose by 1.4% on a company-size-weighted basis; the participating organizations had adopted a variety of shorter-hours arrangements. Revenue is an imperfect proxy for productivity, only a subset supplied the figures, and the pilot lacked a comparable nonadopting-firm series. The Australian early-adopter study identified reported productivity gains and workflow changes, but its qualitative design cannot isolate an output effect (Hopkins et al., 2026). Spanish managers described efficiency as a motivation for adoption and recommended clear performance indicators; an intention to become more efficient is not observed efficiency (Gallego-Gómez & Parente, 2026). A four-day-week review reports favorable productivity findings among the small number of organizational papers it examined while noting the shortage of organization-level research (Araújo et al., 2024).
Another distinction is between expectations and realized performance. Belgian listed firms had positive average stock-price reactions around the announcement of a compressed-week law, particularly among firms with more knowledge work or lower prior employee productivity (Dutordoir & Struyfs, 2024). Investors' expectations do not establish an eventual production gain. A policy analysis of Australian productivity advocates a four-day week, but an argument for adoption is also not a measured treatment effect (Blackham, 2025). Nor should a study of changing national days-of-work patterns be read as an evaluation of firms adopting reduced hours (Hamermesh & Biddle, 2022).
Taken together, the observed firm reports are consistent with output being maintained in some workplaces. They do not justify a general claim that a four-day week raises aggregate productivity. Service quality, staffing costs, work shifted to nonparticipants and output per paid hour need to be measured together. The case for a gain is strongest where organizations can remove low-value meetings or reallocate work and weakest where continuous coverage requires replacement labor; the retrieved employee interviews identify coordination and workload concerns in such roles (Jain et al., 2025). Contextual work on hours, productivity and job satisfaction points to mechanisms worth measuring, not a universal effect size (Nanda & Browne, 1977) (Lee & Lim, 2017).
Wellbeing
The wellbeing evidence is broader and methodologically stronger. Fan and colleagues compared before-and-after data for 2,896 employees in 141 organizations adopting a paid four-day reduction with data from twelve comparison companies. Burnout, job satisfaction and self-reported mental and physical health improved in the intervention organizations, a pattern not observed in the comparisons. Larger individual reductions in hours were associated with larger wellbeing improvements, and changes in work ability, sleep problems and fatigue statistically mediated the association (Fan et al., 2025). The comparison firms strengthen inference, but employer participation and worker outcomes were not randomized.
A controlled Swedish public-sector intervention reduced weekly hours by a quarter at unchanged pay. Across 33 randomized workplaces, diary data from 580 participants showed better sleep, less sleepiness and less perceived stress at eighteen months; the reported effects were small (Schiller et al., 2016). A related randomized time-use analysis found lower total workload and more recovery activities during workdays (Schiller et al., 2017). These analyses draw on a related intervention population and should not be counted as independent replications. A separate Swedish social-service quasi-experiment reported improvement in stress, exhaustion, sleep and work intrusion into private life after hours and workload were reduced together (Barck-Holst et al., 2015). Mixed-methods follow-up found less emotional exhaustion but persistent high caseloads, making it difficult to attribute all benefits to hours alone (Barck-Holst et al., 2019). Interviews in that setting emphasized improved off-work recovery and family relationships (Barck-Holst et al., 2020).
Review evidence agrees on direction while limiting certainty. A scoping review of paid hours reduction mapped improvements in work–life balance and mental health, but emphasized heterogeneous schedules, few long-term studies and limited sector coverage (Utzet et al., 2026). A four-day-week people-management review found favorable satisfaction and work–life-balance patterns among the small set of papers it examined, while calling for more organization-level evidence (Araújo et al., 2024). These syntheses overlap with primary studies above, so their positive conclusions should not be treated as additional independent trials. An exploratory Italian study also reported positive individual and organizational wellbeing assessments, while stressing the need to examine how work is evaluated and organized (Antonelli & Luciano, 2025).
Qualitative and survey research exposes important exceptions. A New Zealand financial-services case study found that staff welcomed the policy yet experienced stronger performance monitoring and productivity pressure (Delaney & Casey, 2021). A small Hungarian study found that wellbeing gains depended on time for personally meaningful activities, not merely nominal release from work (Venczel, 2024). Interviews with women in a childcare nonprofit described how freed time could be absorbed by unequal care obligations (Cruz & Hoth, 2026). In the UAE's four-and-a-half-day survey, many respondents reported better work–life balance, but more than a third reported increased weekday stress; the sample was concentrated in public education and was not representative (Alawad & Smail, 2026). A qualitative UK study of workers still on five-day schedules raised similar anticipated concerns about workload and coordination; these were expectations, not measured implementation effects (Jain et al., 2025). Across these studies, hours saved, work intensity and household demands jointly shape wellbeing.
Retention
Retention has to be measured as actual employee departures, not willingness to stay, enthusiasm for a schedule, or an employer's decision to keep the policy. Lewis and colleagues’ UK pilot report reported resignations per hundred employees falling from 2.0 in a comparison period to 0.8 during the pilot, a 57% decline. The report itself cautioned that small counts and labor-market conditions prevented a statistically firm causal conclusion. The participating firms were self-selected, and the comparison was historical rather than a matched control group. The figure is therefore suggestive, not an established treatment effect.
Other evidence is about attraction or commitment rather than exits. A conference study of full-time, compressed four-day schedules found that burnout risk predicted preference more clearly than job satisfaction, and that a four-day-week offer increased organizational commitment in a vignette experiment (du Bois et al., 2025). That result identifies a plausible retention pathway for compressed schedules but does not show that fewer employees resigned under a paid hours reduction. Australian early adopters described the schedule as a recruitment and retention tool (Hopkins et al., 2026). Spanish managers likewise named talent attraction and retention among adoption motives (Gallego-Gómez & Parente, 2026). Portuguese leaders perceived strategic value but disagreed about how feasible and durable the model would be (Fontes et al., 2026). These are managerial accounts, not independent turnover estimates.
A four-month pilot among volunteer hospital leaders reported lower burnout and higher job satisfaction, with no observed deterioration in selected patient-safety measures; it did not establish an effect on actual turnover (Hayakawa et al., 2025). A study of front-line service workers linked satisfaction with compressed hours to lower subsequent absenteeism, which is relevant to staffing stability but is not retention and concerns a different schedule (Deery et al., 2016). Early research on compressed weeks documented employee acceptance and favorable reactions (Fottler, 1977) (Hodge & Tellier, 1975). Those historical attitudes show that schedule preference can be measured; they cannot substitute for personnel-record evidence about current paid hours reductions. The research basis for a causal retention benefit is consequently thinner than the basis for short-term wellbeing gains.
Discussion
The evidence is most coherent when the treatment is defined by paid hours, pay, workload and shift length. The favorable reduced-hours results can coexist with mixed compressed-week results because four longer days may increase fatigue even when a fifth commuting day disappears (Utzet et al., 2026) (Bernstrøm et al., 2025). Mechanisms also vary: recovery and motivation may protect performance (Rae & Russell, 2025) (Pencavel, 2016), but intensified monitoring and workload can offset some of the time benefit (Delaney & Casey, 2021). Studies of social workers demonstrate that cutting workload alongside hours is part of the intervention, not a nuisance detail (Barck-Holst et al., 2015) (Barck-Holst et al., 2019). A case study of a small Hungarian firm adds that usable free time depends on other people's schedules (Venczel, 2024).
The wider policy literature raises questions beyond the firm-level average. Social-policy scholarship frames the four-day week as a distributional question (Chung, 2022); research on shorter full-time standards examines gender equality in paid and care work (Kümmerling, 2026). A recent interview study suggests that a day away from paid work can be filled with unpaid obligations unequally (Cruz & Hoth, 2026). Studies of environmental consequences (Cieplinski et al., 2021) and employment effects (Booth & Schiantarelli, 1987) address different endpoints. An SSRN analysis of four-day-week discourse (Lehmann et al., 2024) and broader accounts of implementation (Frayne et al., 2024) (Hong et al., 2024) help explain why advocacy, law and organizational experiments often use the same label for different arrangements. None of these contextual works should be mistaken for direct evidence on output or exits.
Limitations
This review is narrative, with no formal screening or pooled effect estimate. Thirty-eight DOI sources were read at abstract level and twelve at metadata level; none was marked as full-text read in the source corpus. Metadata-only works inform the scope map only. The UK pilot report has no DOI in the retrieved record and is linked by name; its bibliographic status was not checked by the DOI script. A resolved DOI establishes that a work exists, not that a claim is supported; claim checks here were limited by accessible abstracts and the separately read pilot report. The literature itself is constrained by self-selection, short follow-up, heterogeneous interventions and limited objective output and turnover measures. No unresolved DOI or unverifiable empirical claim is carried into the outcome synthesis.
Conclusion
Paid reductions to a four-day week are associated with better short-term wellbeing in multi-organization pilots and controlled hours-reduction studies. Evidence that output can be maintained in some adopting firms is plausible but relies heavily on incomplete firm reports and indirect productivity measures. The proposed retention benefit is promising yet rests mainly on pilot turnover comparisons, organizational accounts and commitment measures rather than controlled personnel-record evidence. Compressed forty-hour weeks have a different, mixed evidence base. Future evaluations should record actual paid hours, workload, output quality, costs and employee exits against comparable nonadopters, with attention to sector and household care demands.