The Unpaid Ledger

Essay · economics · September 2026

Over a lifetime, women in Europe and America now do more paid work than they did twenty years ago. Did it cost them leisure, or did it replace work at home? The time-use surveys give a fairly clear answer: mostly it replaced housework, which fell for women in every country measured, often by more than paid work rose. But women still do more work in total than men in most countries, and have less leisure in all of them.

Reading time ≈ 8 min Paid → unpaid → leisure A follow-up to The Fifteen-Hour Week

The question

The Fifteen-Hour Week counted the hours of paid work in a whole life and found that, in most rich countries, they had hardly fallen: the extra time people gained came mostly as longer lives. In Portugal they had risen, from about 46,000 to 62,000 hours over a lifetime, because women moved into paid work. That essay left a question open, because it counted only paid work. When women took paid jobs, did they give up leisure, or did they stop doing some of the work at home that no statistic of paid hours records?

This essay answers it with the time-use surveys, in which people record what they did in each ten-minute slot of a day. It builds the same lifetime ledger as before, but for every activity and for women and men separately, and compares it across survey rounds twenty years apart.

13 of 13countries where women’s expected hours of unpaid work between 20 and 75 fell between survey rounds

10 of 13countries where women’s total work, paid and unpaid together, fell over the same years

11 of 13countries where women still do more total work than men over those years of life

1999the last time Portugal ran a time-use diary survey

The method

A ledger of the whole day

The ledger works like life expectancy. Take the time people of each age spend on each activity in a given year, averaged over everyone and every day of the week, including weekends, holidays and people without jobs. Weight each age by the years a newborn can expect to live at that age under that year’s death rates. The sum is the number of hours a person would spend on each activity if they lived through that year’s patterns of time use and mortality. The surveys cover adults in comparable age groups only between 20 and 75, so the ledger counts those years: fifty-five years of adult life, about 480,000 hours for someone who lives through all of them, and about 460,000 on average once early deaths are allowed for.[1][2]

The European data come from the Harmonised European Time Use Surveys, run around 2000, 2010 and 2020; the American data from the Bureau of Labor Statistics’ annual survey, pooled for 2003–2005 and 2022–2024.[1][3] The day is split into paid work, unpaid work (housework, shopping and caring for children and adults), leisure, study, travel and personal care, which includes sleep. When comparing rounds, the life table is held at the later year, so that the changes shown come from how people spend their days, not from longer lives.

The day

Who works, and at what

In every country surveyed, women do more unpaid work than men and less paid work (Figure 1). In Germany in 2020, a woman could expect about 47,000 hours of paid work between 20 and 75, 67,000 hours of unpaid work and 104,000 of leisure; a man about 64,000, 42,000 and 111,000. Add paid and unpaid together and women do more work in total in 11 of the 13 countries, by 8,000 hours in Germany and the United States and by 25,000 to 35,000 in Italy, Spain, Greece and Bulgaria. Only in Norway and the Netherlands do men do slightly more. Men have more leisure in all of them.[4]

Fig. 1 — Expected hours between the ages of 20 and 75 spent on paid work (excluding commuting), unpaid work (household and family care, including childcare and shopping) and leisure, women (W) and men (M), latest time-use survey round; the label gives paid plus unpaid work in thousands of hours. Personal care (including sleep), study and travel make up the rest of the roughly 460,000 hours. Data: Eurostat, HETUS 2010 and 2020 rounds (tus_00age, tus_20); BLS, American Time Use Survey microdata 2022–2024; life tables from Eurostat and the UN; author’s calculation.

The division is sharpest in southern and eastern Europe. An Italian woman in 2010 could expect 95,000 hours of unpaid work between 20 and 75, three times as many as an Italian man; a Norwegian woman in 2020, 66,000, a third more than a Norwegian man.[4]

The change

Housework gave way

Between survey rounds, women’s paid work rose in 12 of the 13 countries with comparable data, and their unpaid work fell in all 13 (Figure 2). In ten of them the fall in unpaid work was larger than the rise in paid work, so women’s total work fell. In Germany between 2000 and 2020, holding the life table fixed, women’s paid work rose by about 9,000 hours between 20 and 75 and their unpaid work fell by 14,000; in Estonia unpaid work fell by 33,000 hours while paid work rose by 5,000. In the United States between 2003–2005 and 2022–2024, paid work rose by 5,000 hours and unpaid work fell by 8,000.[4]

Fig. 2 — Women: change in expected hours of paid work, unpaid work and leisure between the ages of 20 and 75, between the first and last time-use survey round available, holding the life table at the later year. The 2020 round was partly carried out during the pandemic. France and the United Kingdom are left out because their travel time jumps between rounds in a way that suggests a change in coding. Data: as Figure 1, plus the HETUS 2000 round and ATUS 2003–2005; author’s calculation.

Leisure is the mixed part of the picture. It rose for women in Germany, Estonia, Austria, Spain and Italy, fell in Poland, Bulgaria, the Netherlands, Hungary, Norway and the United States, and barely changed in Finland and Belgium. The three countries where women’s total work rose, Poland, the Netherlands and Hungary, are among those where their leisure fell: there, part of the move into paid work did come out of free time. Elsewhere the time freed from housework went partly into paid work and partly into sleep, personal care and study, which rose in almost every country. In the United States, the largest change for women aged 25 to 54 was not in work or leisure but in sleep and personal care, which rose by about half an hour a day.[3]

The fall in housework is not only because more women have jobs. Among employed American women aged 25 to 54, time on housework fell from 170 to 149 minutes a day between 2003–2005 and 2022–2024.[3] Less housework is being done, by everyone, in the time the surveys can see.

The men

A slower shift

For men the changes since 2000 are smaller and point in different directions (Figure 3). Their unpaid work rose by a few thousand hours in Norway, Finland, Poland and Spain, and fell in Estonia and Bulgaria; their paid work fell in some countries and rose in others. In the United States it barely moved.[4]

Fig. 3 — Men: change in expected hours of paid work, unpaid work and leisure between 20 and 75, same countries and method as Figure 2. Data: as Figure 2; author’s calculation.

The larger shift happened earlier. In the United States, between 1965 and 2003, women aged 21 to 65 increased their market work from 20 to 26 hours a week and cut their non-market work from 35 to 22; men cut their market work from 51 to 40 hours and increased their non-market work from 10 to 14 (Figure 4).[5] Across seven countries with surveys from the 1970s to the 2000s, adults’ total work, paid and unpaid, fell only a little, while its division between women and men changed a great deal.[6] The measured rise in American leisure over those decades is disputed: Mark Aguiar and Erik Hurst find it large; Valerie Ramey and Neville Francis, using a different age range and treatment of the 1965 survey, find it close to flat for people of working age.[5][7]

Fig. 4 — United States: hours per week of total market work (including commuting) and non-market work (housework, shopping, care), women and men aged 21 to 65 and not retired, 1965 and 2003. Data: M. Aguiar and E. Hurst (2007), Table 2.

Portugal

The missing survey

Portugal cannot be put in this ledger. Its only diary survey was carried out by INE in 1999, and its results were published by employment status rather than by age. They show the same division at its starkest: employed women spent 310 minutes a day on paid work and 237 on housework and family care, employed men 402 and 54, so that working women did about an hour and a half more work a day in total.[8] A 2015 survey by CESIS for the equality commission asked about the previous weekday instead of keeping a diary; it found women doing about four and a half hours of unpaid work on a weekday and men about two and a half, but its averages cover only those who did each activity and cannot be compared with the diaries.[9] Portugal did not take part in the European round of 2020.[1]

That matters for the question The Fifteen-Hour Week left open. Portuguese women’s move into paid work has been large, and the Portuguese lifetime ledger of paid work rose because of it. Whether it replaced housework, as in the countries above, or came on top of it, no survey since 1999 can say.

The balance

What the ledger shows

The answer to the question is mostly reassuring, with a qualification. Across Europe and the United States, women’s move into paid work over the last twenty years was more than matched by a fall in their housework in most countries, so their total work fell and their leisure, on balance, held steady. Where it did not, in Poland, the Netherlands and Hungary, part of the extra paid work came out of free time. The qualification is the level: women still carry more of the total work in most countries and have less leisure in all of them, and men’s share of the work at home has moved little since 2000.

For the Keynes question this adds a line to the ledger. Part of the productivity dividend of the last half-century was taken as less work at home, a gain that no statistic of paid hours records. It was taken mostly by women, and it paid for much of their move into paid work.

On method and tools

This piece was written collaboratively with Claude Opus 5.5 (Anthropic): human specification, editorial direction and critical review; machine data research, analysis and drafting. The figures and derived numbers are computed by scripts/unpaid_ledger.py from Eurostat’s published HETUS tables and life tables, and from the American Time Use Survey microdata, which are public; the harmonised European microdata, which need registration, were not used. The ledger is a period measure for ages 20 to 75. The European tables use broad age groups, and two approximations were needed: the 2010 round’s figures for people aged 65 and over stand in for ages 65 to 74, and the 2020 round’s figures for ages 15 to 24 stand in for 20 to 24. Survey rounds differ in their years of fieldwork, and part of the 2020 round took place during the pandemic. Commuting is counted as travel, not paid work, throughout. The results are in docs/unpaid-ledger-results.json and the figures in docs/unpaid-ledger-figures.html; the downloaded sources, with a table or page reference for every number, are kept with the script’s data.

The cover photograph is Portogallo2007, Rua de Santa Cruz do Castelo, Lisbon, by “ho visto nina volare”; CC BY-SA 2.0, via Wikimedia Commons, cropped.

Authored by: Luis Matos Ferreira — Physicist, Developer, Writer

Related essays on this blog

1. The Fifteen-Hour Week — Keynes’s prediction, and the lifetime ledger of paid work this essay extends.
2. Where the Hours Went — income, consumption, unpaid work and screens.
3. The Fifth Day — Portugal’s four-day-week pilot.
4. The Portuguese Basket — the 1980 household basket repriced in hours of work.

Sources

1. Eurostat, Harmonised European Time Use Surveys: time spent on main activities by sex and age group, 2000 and 2010 rounds (tus_00age) and 2020 round (tus_20 tables), minutes per day averaged over all persons and all days; retrieved September 2026. The 2020 round covers eleven countries.
2. Eurostat, life tables (demo_mlifetable), person-years lived by single year of age; United Nations, World Population Prospects 2024, abridged life tables (United States).
3. U.S. Bureau of Labor Statistics, American Time Use Survey, multi-year activity summary and respondent files, 2003–2024, with final weights (TUFNWGTP; TU20FWGT for 2020), as recommended in the ATUS User’s Guide; author’s tabulation by sex and age.
4. Author’s calculation from sources 1–3 (scripts/unpaid_ledger.py).
5. M. Aguiar and E. Hurst, “Measuring Trends in Leisure: The Allocation of Time over Five Decades”, Quarterly Journal of Economics 122(3), 2007; NBER Working Paper 12082, Table 2.
6. J. I. Giménez-Nadal and A. Sevilla, “Trends in time allocation: A cross-country analysis”, European Economic Review 56, 2012, Tables E3–E4.
7. V. A. Ramey and N. Francis, “A Century of Work and Leisure”, American Economic Journal: Macroeconomics 1(2), 2009, pp. 210–211.
8. INE, Inquérito à Ocupação do Tempo 1999, “O Uso do Tempo”, press release of 19 June 2000, table of average duration of activities by labour status and sex.
9. H. Perista et al., Os Usos do Tempo de Homens e de Mulheres em Portugal, CESIS/CITE, 2016 (Inquérito Nacional aos Usos do Tempo 2015), Figure 33, p. 60.