Executive summary

Develop a common right to genuinely shorter paid working time through different funded sector arrangements. Keep the goal of a four-day week at unchanged ordinary weekly pay, while specifying how coverage, staffing and earnings protection would work. Start personal AI access independently; an employee should not have to wait for national hours reform, and a carer or unemployed person should not be excluded because an employer cannot release their time.

The Australian starting point is not one standard per industry. Official examples include 35 hours in black coal mining, 37.5 at PM&C and 38 in several major awards; the Clerks award also recognises shorter full-time workplace hours. These are instrument examples, not population averages. The recommended design uses a verified workplace baseline, an initial 10% reduction where staging is needed, and a final 20% reduction by 2035, with a four-day pattern or a worker-agreed equivalent for other rosters. All targets remain proposals for review.

The practical choice is between a declaration of fewer hours and an arrangement that actually gives people time. Office redesign can remove unnecessary work. Care, education, transport, retail and continuous operations also need people present. Additional staffing may be necessary even where administrative AI is useful. Do not finance the programme by assuming service users can accept less care or by counting the same productivity saving twice.

For a constant-output scenario, reducing hours by 20% requires 25% more staff when output per hour is unchanged. A measured 10% gain reduces that additional requirement to about 13.6%; a 25% gain removes it arithmetically. These are conditions to examine, not productivity forecasts. National costs need sector wage bills, skill mix, actual coverage and funding responsibilities. This report supplies reproducible cost factors and a sector ledger, not an invented national total.

Recommendation: pursue WF4 through phased, negotiated sector arrangements, with a funded staffing route where efficiency cannot preserve services; treat WF3 as a useful delivery position while those arrangements develop. Require explicit part-time/casual offers and measurable service protection. The current judgement is DESIGN PLAUSIBLE; DELIVERY UNRESOLVED. A3 can develop an honest public proposal from this design after review of the funding direction.

Reader orientation, premises and confidence

A1 appraises selected trials and postwar hours reforms. A2 translates their limits into Australian options; A3 sets out a campaign with successive political and sector-delivery commitments. B2 supplies the companion individual AI funding design. P1–P3 remain ASSUMED FOR DESIGN: some people use released time with free AI, personal learning by doing is the foundation, and that capability contributes to employment and cohesion resilience. Participation in learning remains voluntary.

Confidence is high in the cited official rules and conditional arithmetic, medium in the usefulness of the sector design distinctions, and unresolved for national net cost, attainable coverage and resilience effects. This is a bounded implementation review, not legal advice on a particular agreement, an independent audit or an exhaustive sector study.

Pass 0 — Define the offer and the continuing-policy baseline

A genuine four-day week means fewer actual work hours at unchanged ordinary weekly pay. It is distinct from four longer days, working invisibly on the fifth day, or closing a service for a day. A target of 80% of a 38-hour baseline gives 30.4 hours, whereas four eight-hour days give 32. Define the baseline in each agreement. A 10% hours reduction can be a transitional step; it is not reported as the final 20% reduction.

The implementation promise should protect contracted ordinary earnings, recurrent remuneration affected by changed rosters, leave and superannuation treatment. Unusual overtime cannot simply become a permanent guaranteed windfall, but a worker who routinely depends on penalties must see the actual before/after pay comparison. Separate genuine savings from costs transferred to workers.

The dated WF1 inventory, at 9 September 2026, includes existing employment instruments, flexible-work requests, voluntary employer trials, current public-service staffing obligations and existing training/adoption programmes. Flexible-work rights are not a general entitlement to 20% fewer paid hours. The Fair Work Ombudsman describes eligibility, employer responses and dispute mechanisms; changing hours still requires the applicable arrangement. [A2-02]

ABS August 2025 data identify 2.5 million employees without paid holiday or sick-leave entitlements, or 20% of employees; among part-time employees, 47% lacked these entitlements. These statistical categories must not be treated as a legal classification of every worker. They show why a salaried full-time office model is an inadequate national design. [A2-01]

Baseline costs are not zero. Existing wages, public services and programmes continue unless an explicit change removes them. Record incremental spending separately; do not deduct hypothetical savings from programmes that remain necessary. The baseline inventory here identifies institutions and obligations. A complete reconciled national spending baseline remains an input needed before a fiscal submission.

Verified Australian starting hours: instruments, not one number per industry

Under the national-system framework, the NES maximum is generally 38 weekly hours for a full-time employee plus reasonable additional hours. Awards and registered agreements specify ordinary hours and averaging arrangements. That is not a rule that everyone ordinarily works 38, or an absolute cap on all actual work. Check jurisdiction, coverage, classification, the operative agreement and contract before making an offer; state public-sector arrangements cannot be inferred from a federal award. SH01: Fair Work Ombudsman

Verified 9 September 2026. These eight examples establish variation and important roster distinctions. They are not an exhaustive industrial survey or an estimate of each industry's average actual hours. The proposals in the final column are this report's design inferences, not existing entitlements.

Research evidence and design table 1
Industry/setting and instrument example Current full-time ordinary-hours reference Meaning for a reduced-hours offer
Black coal mining — Black Coal Mining Industry Award 2020, clauses 9 and 14 35 per week, or average 35 over a roster cycle Audit actual shifts/overtime first; 20% of the 35-hour standard means 28, not 32. Do not generalise to every mining activity. SH02
Commonwealth public administration — PM&C Enterprise Agreement 2024–27, clause 91 37.5 per week, ordinarily 7.5 per day A five-day pattern becomes four 7.5-hour days at 30 hours under the proportional design. This is a named agency example, not every public employer. SH03
Private clerical work — Clerks–Private Sector Award 2020, clauses 9 and 13 38, or fewer hours the employer treats as full-time; averaging provisions apply Use the established lower full-time baseline where applicable. Do not reset a 35-hour workplace to 38 before calculating its benefit. SH04
Manufacturing day workers — Manufacturing and Associated Industries and Occupations Award 2020, clause 17.2 Average 38, not exceeding 152 in 28 days under this provision Separate daily attendance and cycle hours; other shift provisions and enterprise arrangements need their own check. SH05
On-site construction — Building and Construction General On-site Award 2020, clauses 16.1–16.2 38 averaged over a four-week cycle; eight-hour working days accrue time towards an RDO The ordinary pattern is not automatically a 40-hour baseline. Existing rostered days off are already part of the starting entitlement. SH06
Award-covered nursing — Nurses Award 2020, clauses 9 and 13.1(c) 38 per week, 76 per fortnight or 152 over 28 days Design reduced cycle hours with funded skill coverage. Four existing long shifts are not necessarily an hours reduction; public-hospital instruments may differ. SH07
Retail — General Retail Industry Award 2020, clauses 9 and 15.6 Average 38; multiple averaging/roster options Protect weekly earnings and trading coverage; compare whole cycles and existing days off. This is not a claim that every retail employee works 38. SH08
Hospitality — Hospitality Industry (General) Award 2020, clauses 9 and 15.1 Average 38, with work-arrangement options Establish paid-hours, penalties and seasonal demand before a predictable shorter roster. Other hospitality occupations/instruments require separate coverage checks. SH09

For teachers, face-to-face teaching hours are not the whole working week: the baseline audit must include preparation, assessment and administration. This revision does not assign one verified national teacher standard. For every setting, report ordinary contractual hours, usual total hours and actual diary/payroll hours separately, including unpaid work and other jobs.

ABS statistics distinguish actual from usual hours; their full-time category generally uses 35 hours or more across jobs, with specified reference-week treatment. That statistical label is different from full-time status under an employment instrument. A successfully reduced full-time contractual job can therefore move below a statistical threshold without becoming a failed job or an involuntary loss of employment. Do not use a falling statistical full-time count alone as a programme failure measure. SH10: ABS July 2026 methodology, glossary

A staged destination calculated from the person's starting arrangement

Let H0 be verified ordinary weekly hours, averaged over the applicable roster cycle. The recommended proportional route is H0 → 0.9H0 → 0.8H0, with ordinary weekly pay protected throughout. The intermediate step is optional where immediate full implementation is funded and workable. It is not two successive 10% cuts: those would leave 81%, rather than 80%, of original hours.

Research evidence and design table 2
Verified example baseline H0 First 10% reduction Final 20% reduction Four equal days at final target Reduction if 32 hours were used instead
35 hours 31.5 28 7 hours/day 8.57%
37.5 hours 33.75 30 7.5 hours/day 14.67%
38 hours 34.2 30.4 7.6 hours/day (7h 36m) 15.79%

These are arithmetic illustrations, not negotiated rosters or new national legal standards. A common 32-hour rule is a legitimate alternative to assess, but it distributes different percentage gains and cannot be called a universal 20% reduction. Preserve any already shorter arrangement; never raise hours to manufacture a common starting point. The proposed proportional target carries forward the existing programme scenario; whether it becomes a public commitment remains subject to human review.

For a regular five-day worker, a nine-day fortnight at unchanged daily length can implement the initial 10% reduction, followed by four days a week. For someone already working a nine-day fortnight, four long shifts or an RDO cycle, calculate fresh cycle totals: the name of the roster proves nothing. Existing leave, RDOs and accrued time must not be relabelled as a new benefit. Shorter days or equivalent predictable blocks can suit some workers; record these as reduced-hours equivalents, and separately report how many actually gain a four-day pattern.

Where usual actual work exceeds ordinary hours, first identify excessive or unpaid work and the pay/coverage remedy. Keep workload correction and the new ordinary-hours reduction as separate accounts. An illustrative employee with 38 ordinary hours and 44 actual hours needs both the ordinary-hours target and evidence that overtime or unpaid catch-up has not swallowed the gain. Reducing overtime alone is not proof that the 20% ordinary-hours target was met. Historical Japan and Korea underline why these measures must be distinct; A1 retains their source appraisal.

Before setting a timetable, each sector agreement should specify H0, current attendance days and RDOs, actual-hours evidence, pay components, the first and final target, funding/skills, service floors, worker time control, dates and remedies. This instrument inventory must precede a national weighted hours or cost estimate; none is invented here.

Pass 1 — What the implementation evidence changes

A1's mixed trial record supports examining work redesign and retained pay, with particular caution about small, self-selected firms and staffing-intensive sectors. It supplies no universal productivity rate. The discontinued Australian arrangement and UK withdrawals remain part of the evidence; a resilient arrangement needs a downturn and exit plan. [A2-06]

The fuller South Cambridgeshire 27-month evaluation is now available for this review. Its COVID-adjusted comparison identifies nine statistically improved indicators and three worsened indicators; the latter concern rent collection, satisfaction with responsive repairs and re-letting time. Other indicators had no statistically detectable difference. That is not proof of equivalence. The unadjusted analysis gives a different pattern, and the authors explicitly limit causal attribution. The key-findings and target tables were appraised; raw-data/statistical replication was not performed. [A2-05]

This closes A1's employer-summary-only gap at the level of direct evaluation findings. It does not establish that a four-day week caused all changes, that every service met its target, or that this council's staffing savings would recur nationally. Australian demonstrations need both absolute service standards and changes relative to an appropriate comparison. Averaging several improvements cannot compensate for an unacceptable loss of a critical service.

In residential aged care, the government specifies a sector-wide average of 215 direct-care minutes per resident daily, including 44 registered-nurse minutes; individual service targets depend on their residents. Administrative efficiency does not itself remove these direct-care obligations. That makes care a clear example of a coverage constraint that must be funded and staffed rather than assumed away. [A2-04]

Historical analysis adds a reason to combine a destination with intermediate commitments: Australia's award route, German bargaining and Swedish legislation all required implementation machinery. France and Portugal warn that subsidies, wage terms and flexibility affect outcomes. A2 therefore treats the first step as a funded, measurable gain with a subsequent deadline, rather than evidence that job creation or final delivery is automatic. A1's historical section and the unchanged background retain the primary citations and contested findings.

Pass 2 — Six practical sector arrangements

The following are design options, not claims that entire industries have already adopted them.

Research evidence and design table 3
Setting Proposed reduced-hours route Evidence and resource requirement Proposed accountable parties
Office and professional services Remove low-value meetings/rework; stagger days off; retain client access Actual hours, client outcomes, workload and full cost; measured efficiency rather than a blanket assumption Employers and worker representatives
Health, aged care and disability services Four-shift or equivalent reduced-hours rosters with maintained skilled coverage Funded replacement capacity, care/skill-mix requirements, continuity and clinical/service quality Service operators, workforce representatives and existing public/private funders
Schools and other face-to-face education Separate teacher workload reduction from student service hours; redesign administration and staff coverage Preserve pupil learning/contact provision; cost release and additional teachers where required Education systems, schools and workforce representatives
Retail, hospitality and small customer-facing firms Staggered rosters, predictable paid shifts and shared coverage where practical Opening/service hours, peaks, casual earnings, small-team constraints and consumer prices Employers, workers and relevant sector bodies
Manufacturing, logistics and continuous operations Redesign shifts, handovers and maintenance; use relief teams Safety, fatigue, skill availability, machinery utilisation and coverage costs Operators, workforce representatives and safety functions
Public administration and local services Retain public opening/service availability through staggered arrangements Processing times, accessibility, unmet demand, procurement and staffing accounts Existing public employers and accountable funders

For each setting, compare a redesign-led offer, a staffing-funded offer and a phased combination. Do not nominate voluntary office adoption as the only practical route and abandon the people most constrained by time. Equally, do not promise a simultaneous national start before skilled staffing and funding exist.

Part-time, casual and self-employed people

For regular part-time work, the preferred comparison is the same proportional time reduction at protected earnings. A 22.8-hour week reduced by 20% becomes 18.24 hours at the same agreed pay. Some workers would rather retain hours and receive the corresponding higher hourly rate; examine that voluntary alternative, but report it as income benefit rather than time released. Audit equity against new full-time hours and benefits.

For casual work, establish a representative paid-hours/earnings baseline that accounts for seasonality and periods without work, then negotiate predictable rosters and a protected equivalent offer. A rate increase without stable hours may leave the worker with less weekly income. A shorter roster imposed on someone seeking more hours is not automatically a benefit. No single retrospective baseline period is selected here without sector evidence.

Self-employed people and contractors lack an employer who can guarantee unchanged pay. Personal AI access, optional business support and service improvements may help, but do not claim they receive the same paid day off. Any separate income/time support requires its own funding justification. People outside employment retain the full personal entitlement; optional accessible support and respite/community-service options should be assessed separately from compute funding.

Three individual experiences

A designer can receive an actual day back if meetings and rework fall without evening catch-up. A nurse can receive the same opportunity only if roster coverage and skill mix remain funded. An unemployed carer already outside payroll may still have almost no discretionary time. All three can access B2's personal offer; their different time constraints require different responses. These are illustrative journeys, not participant findings.

Pass 3 — Cost the hours and distinguish who pays

Let W be the current annual remuneration cost of a defined workforce, q the share covered, h the proportional hours reduction and g the independently supported gain in effective output per hour. For constant required output and unchanged remuneration per worker:

additional payroll = W × q × max(0, 1 / ((1 − h) × (1 + g)) − 1).

The zero floor means this model does not finance the proposal with assumed redundancies. Add redesign, recruitment, training, capital, payroll effects and delivery overhead separately; subtract only independently evidenced, non-overlapping savings. Effective output must reflect service quality, not merely throughput.

Research evidence and design table 4
20% hours reduction Additional staff/payroll factor Per A$100m of covered baseline payroll
No output-per-hour gain 25.0% A$25.0m annually
10% supported gain 13.6% About A$13.64m annually
25% supported gain 0% No added payroll under these assumptions; redesign may still cost money

For the proposed initial 10% reduction, the same no-gain coverage formula gives 11.11% additional payroll; moving from 90% to 80% of original hours then requires a further 12.5% of the intermediate payroll, reaching 25% above the original. Do not add 11.11 and 12.5 as if they had the same denominator. A fixed 32-hour destination instead gives no-gain factors of 9.38% from 35, 17.19% from 37.5 and 18.75% from 38. Its lower cost is accompanied by a smaller time benefit. The accompanying calculation file reproduces both offers and keeps them separate.

For a heterogeneous rollout, sum W_i × q_i × max(0, H0_i / (H1_i × (1+g_i)) − 1) across distinct covered workforces. Use the applicable target and participation dates for each; prorate only genuine partial-year implementation. Whole-person staffing and capability constraints still apply. This is an extension of the existing sensitivity method, not a revised national total.

These are sensitivities, not estimated sector gains. Where fixed direct-staffed hours must be maintained, use the no-gain coverage case unless a change in required staff time is independently justified. Small teams need whole people and suitable skills, so fractional staffing calculations can understate practical costs. If demand grows, the denominator changes and these figures must be recalculated.

Maintain separate public and employer ledgers. Publicly funded care or education may require recurrent government funding. Private employers may cover costs through measured efficiency, margins, prices or new staffing; each has distributional consequences. A transitional public contribution is an option to appraise, not an automatic permanent wage subsidy. Grants to operators for an agreed time/service transition serve workers and service users; they are separate from B2's individual AI entitlement.

No defensible national payroll total by eligible workforce, service obligations and pathway coverage was established in this bounded review. The model therefore leaves T2, T3 and T4 as explicit national time-cost terms and supplies reproducible factors. A3 must not present the AI allowance's cost as the entire four-day-week programme budget. Before a national fiscal claim, reconcile workforce counts and compensation with sector accounts, workforce supply, funded service contracts and overlap between public and private payments.

Pass 4 — Four pathways, fair reach and adoption mechanisms

Research evidence and design table 5
Pathway Time arrangement Fiscal/service question 2035 interpretation
WF1 — Business as usual Existing rules and voluntary changes evolve What improvement and exclusion occur at existing spending? Credible comparator, not assumed failure
WF2 — Incremental reform Expanded voluntary trials and institutional flexibility Price selected additional coverage, T2, without extrapolating every trial Useful gains with continued unequal time
WF3 — Individual opportunity, uneven time Personal AI entitlement operates while only some sectors deliver shorter hours Price partial additional coverage, T3; identify who lacks usable time Personal opportunity delivered, integrated proposition incomplete
WF4 — Integrated success by 2035 Funded sector arrangements achieve real time and protected pay with maintained services Price full intended coverage, T4, plus workforce and service contingencies Time, access and measured resilience criteria jointly achieved

The legal route should begin with negotiated employment arrangements and public-service funding agreements, while A3 compares the case for broader statutory change. The Fair Work Commission's better-off-overall test examines the agreement as a whole against the relevant award; the programme's individual earnings/time promise can be more demanding than that legal comparison. Passing the test alone does not prove this promise has been kept. Applicable national/state coverage and specific instruments need verification before implementation. [A2-03]

Common stress cases are a staffing shortage, revenue downturn, demand surge, faster AI disruption and constrained public finance. For every pathway show who absorbs the shock. WF4 needs funded continuity and renegotiation provisions; it cannot assume a permanently favourable labour market. If a critical service requirement cannot be met, pause expansion and address staffing rather than conceal deterioration in an average indicator.

Proposed readiness and success criteria

Before a demonstration, require a specified reduction/pay baseline, an agreed roster, a funded coverage plan, meaningful worker participation and service-specific safety/quality thresholds. These are proposed design requirements. Legal service floors cannot be relaxed by an AI assessment. Set improvement/non-inferiority margins and measurement periods with the relevant service expertise before analysing outcomes, rather than treating a non-significant result as success.

For WF4, report population coverage, actual discretionary time, ordinary and usual total pay, workload, service outcomes and distribution by working arrangement. Record non-users of AI without penalising them. B3 must then examine transfer, livelihoods and cohesion; A2's time delivery cannot establish that entire causal chain. Numerical national coverage targets require a costed sector account and human judgement, not an invented universal adoption percentage.

Pass 5 — WF4 backcast and handoff

Research evidence and design table 6
Window Required output and responsible role Evidence / response if missing
Late 2026 Research team develops sector templates; human decision-makers review phased funding/earnings direction Explicit service and pay definitions; revise an offer that hides income loss
2027–2028 Seek funded first-step and final-destination commitments; employers, workers and funders develop instrument-specific priced agreements Publish H0, proposed first/final targets, pay, service and workforce accounts; distinguish a political promise from an operative agreement
2029–2030 Separately approved demonstrations and first-step arrangements include care and shift work as well as offices Evaluate actual time dose and coverage; direct-to-final demonstrations may proceed where funded; pause unsafe expansion
2031–2032 Review initial gains and implement the next funded reduction toward the final target Publish delayed sectors, corrective workforce/funding plans and revised dates; preserve viable earlier gains
2033–2034 Negotiated expansion addresses remaining groups Document funded reach and exclusions; retain viable arrangements if wider adoption is delayed
2035 Existing accountable bodies report delivered time and services; B3 evaluates integrated outcomes Partial coverage or unmeasured effects remain partial success, not full WF4

These are proposed roles, not appointments. Working-time reform and personal access can progress at different speeds. If national reform is delayed, preserve WF3's useful access offer and publish the unresolved time barriers. If an arrangement fails, use an agreed worker/service transition process rather than silently reducing pay or counting a terminated trial as continuing success.

A3 handoff: develop a multi-step campaign around a verified starting point, funded initial reduction, dated final destination, protected earnings, essential services, inclusion and transparent incremental costs. Compare voluntary/negotiated and statutory routes; do not claim a voter threshold or a cost-free national conversion. B2 handoff: keep AI and time costs separate and disclose their eventual combined funding burden. B3 handoff: pre-specify service and outcome comparisons, including null/adverse/subgroup results.

Objections and current judgement

Would this trap essential services in permanent delay? A funded staffing route is essential to WF4; voluntary office adoption alone is WF2/WF3. The remedy is a credible workforce and funding commitment, not a paper entitlement without coverage.

Does higher hourly pay solve the problem? It can improve income, but does not release time unless hours fall. Keep both outcomes visible and preserve worker choice.

Could ordinary reform perform better? Yes, WF1/WF2 may deliver important gains at lower incremental cost. Compare their actual reach and public value fairly under common conditions. The programme's accepted premises do not permit suppressing that possibility.

Does the proposal already have a national price? No. It has a reproducible costing method, sector requirements and named missing inputs. The funded, sector-specific route remains an unresolved design decision. The evidence judgement remains DESIGN PLAUSIBLE; DELIVERY UNRESOLVED.

References and source locations

Retrieved 9 September 2026. The implementation source register records appraisal depth and transfer limits.

The SH01–SH10 sources above were checked on 9 September 2026. The public source register records clauses and coverage limits. The public calculation and targets files reproduce the sector-hours arithmetic; the current reports incorporate the selected historical findings.

Glossary

BAU: business as usual, with existing approaches continuing to evolve. Coverage: people/services included in an arrangement. FTE: full-time equivalent, which need not correspond to a whole recruit. Non-inferiority margin: the largest predefined deterioration considered acceptable for a specified comparison. Backcast: work backwards from the intended future to required actions.

Authorship and method

Research by Rick Molony, developed with AI assistance. This is bounded critical research and implementation analysis, not an independent audit, field trial, procurement quote, approved budget or legal opinion.

Suggested citation

Rick Molony (2026). “Making time available fairly”. Report A2, What If Research, Our Resilient World. Research dated 9 September 2026. CC BY 4.0.

How this page was derived: the full substantive source body was rendered after the declared public-context transformations, removal of private delivery/workflow material and omission of the private change log.