Time-use surveys — where participants record what they do across a day — have been conducted in many countries and produce consistent findings.
Women perform substantially more unpaid work than men: childcare, care of adults, cooking, cleaning, laundry, and the organisational work of running a household. Estimates of the ratio vary by country, and the direction is consistent everywhere it's measured.
Why it isn't counted
National accounts measure market production. Unpaid household production is excluded by convention, on grounds including difficulty of valuation and the boundary problem of what counts.
The consequence is a significant distortion. If a person is paid to care for children, that's economic output. If a parent does the same work, it isn't.
Which means an activity's inclusion in the measured economy depends on whether money changed hands rather than on whether anything was produced.
Attempts to estimate the value of unpaid work through satellite accounts — supplementary calculations alongside standard national accounts — have produced figures representing a substantial proportion of GDP, with estimates varying considerably by method.
Why the measurement matters
Not merely a technical point.
Policy attention follows measurement. Activities that appear in economic statistics attract policy interest. Those that don't are invisible to the frameworks used for decisions.
It shapes what counts as work. Someone providing full-time care is described as economically inactive, which is an odd description of somebody working long hours.
It affects entitlements. Social insurance systems built around employment can exclude people whose work is unpaid, with consequences for pensions and benefits.
It obscures the trade-off. Policy discussions about labour force participation frequently ignore that the care work has to happen regardless, and that participation requires either paid substitution or a redistribution within households.
The composition of the gap
The imbalance isn't uniform across activities.
Routine, daily, non-discretionary tasks — cooking, cleaning, laundry, daily childcare — show the largest disparities. These are the activities that must happen every day and cannot be deferred.
Occasional and discretionary tasks show smaller differences, and some show the reverse. Maintenance, repairs and outdoor work skew the other way in many surveys.
The distinction matters because it isn't only about hours. Work that must be done at particular times, every day, constrains other activity in a way that occasional tasks don't.
The organisational layer
Time-use surveys capture activity and struggle to capture the cognitive and organisational work of anticipating, planning and delegating.
Research on this dimension has found it distributed even more unevenly than the physical tasks, and that it's associated with distress independently of hours spent.
The distinction between doing a task when asked and holding responsibility for noticing it needs doing is substantial and largely invisible in any measurement.
The lifetime consequences
Unpaid work has effects that persist long after the caregiving period.
Reduced hours and career interruptions affect earnings trajectories, as discussed extensively in labour economics.
Pension contributions are foregone, with compounding effects.
Skills and networks depreciate during absence from paid work.
And the effects are largest for caring for adults, which is less predictable than childcare, frequently arrives during peak earning years, and has no equivalent of children eventually starting school.
What shifts it
Evidence on interventions is limited and a few things have support.
Non-transferable parental leave for fathers. Where leave is reserved and lost if not taken, take-up rises substantially, and there's evidence of persistent effects on subsequent division of care.
Affordable childcare. Reduces the total unpaid load and enables participation.
Flexible working available to everyone. Where flexibility is available only through mechanisms used predominantly by women, it reinforces the pattern.
Measurement itself. Countries conducting regular time-use surveys have data available for policy in a way that others don't.
At household level
Individual arrangements matter and the general advice — communicate, divide fairly — is thin.
What appears to work better is making the invisible visible. Listing everything that happens in a week, including the organisational work, produces a document that's frequently surprising to whoever is doing less.
Allocating ownership of domains rather than tasks addresses the organisational layer. Being responsible for meals means noticing, planning, shopping and cooking, rather than cooking when asked.
And reviewing periodically, since arrangements drift and default to previous patterns after any disruption.
None of that resolves a structural issue at household level. It does address the specific problem of one person carrying an invisible load that nobody has acknowledged exists.
Care of adults, specifically
The component growing fastest and receiving least attention. Ageing populations mean more people providing unpaid care to parents, partners and other adults, and surveys consistently find this falling disproportionately on women.
It differs from childcare in ways that make it harder. It arrives unpredictably rather than on a known timeline. There is no equivalent of children starting school. It frequently coincides with peak earning years and with having children still at home. And it is emotionally harder, because the trajectory is towards decline rather than independence.
Carer's allowances and entitlements exist in many systems and are frequently unclaimed, partly because people providing care do not identify as carers. Checking eligibility is worth doing, and the assessment thresholds are frequently lower than people assume.