The Invisible
24 Hours

Paid participation and the rest of the day

The comparison becomes clearer when we place women's and men's paid and unpaid time side by side.

Source: MoSPI's 2019 Statement 26 and 2024 Statement 25, all India, rural and urban combined. Each average includes everyone in the stated age and sex group. The reports allocate time across all recorded activities within each time slot. Published components and totals are rounded separately, so a total can differ by one minute from their sum. 2019 report package; 2024 report

Men's average paid time also increased. Their unpaid time declined, and their published combined total fell slightly. Women's combined total increased. The difference between the female and male totals widened from 54 to 67 minutes a day, calculated from the published rounded estimates.

The comparison does not explain the causes of those changes. Shifts in population composition, household circumstances, economic conditions and reporting may all matter. It does show how differently the same national development story can look when both paid and unpaid activity enter the frame.

The participation measure also deserves care. It records whether someone did an activity during a particular 24-hour period. A person could have a job and be off work that day; someone outside regular employment could perform paid activity. These percentages answer a different question from a labour-force participation rate.

There is another revealing distinction. Among working-age women who did paid activity, average time in that activity fell from 332 to 305 minutes. Yet paid time averaged across all women rose, alongside higher participation. A changing participant group can have a lower average while the population as a whole records more time. To understand the change, we need to know both how many people participate and how long participants spend. MoSPI, 2019, Statement 25; MoSPI, 2024, Statement 24

For policy, the larger issue is how the different parts of a day fit together. A rise in paid activity can expand opportunity while leaving the organisation of unpaid work largely unresolved. Measuring the first change alone cannot tell us how manageable the whole day has become.

To see what that unresolved work contains, we need to separate housework from care.

Housework and caregiving moved differently

The surveys allow us to look inside unpaid responsibilities rather than treating them as one undifferentiated block.

For this comparison, the population changes to everyone aged six and over. It therefore includes girls and boys as well as adults. Across that population, females' average time in unpaid domestic services fell from 243 to 236 minutes a day between 2019 and 2024. Their caregiving time rose from 37 to 47 minutes. Male domestic-service time moved from 25 to 24 minutes, while caregiving increased from 11 to 13. These are averages per person, including nonparticipants. MoSPI, 2019, Statements 5–7; MoSPI, 2024, Statements 1–3

Adding the two rounded female averages gives 280 minutes in 2019 and 283 in 2024. A headline about falling housework time would miss the increase in caregiving. Looking at both categories together changes the interpretation.

Caregiving participation also rose for both sexes, with females continuing to report higher participation. The survey cannot tell us from these totals alone whether this reflects different care needs, changing household arrangements, greater recognition of care, or other influences. It gives us a clearer question to investigate.

That matters because activities placed under the broad heading of unpaid work may respond to quite different conditions. In a hypothetical household, a shorter meal-preparation routine could coexist with more time caring for an older relative. The household could save time in one activity while facing greater demands in another. An aggregate total would conceal that rearrangement.

Care can carry affection and obligation together. Its duration leaves questions about choice and support open. A decline in one activity is therefore not an automatic improvement: what changed alongside it, for whom, and under what circumstances?

Once those tasks are separated, another question follows: whose average day are we looking at?

A national average can hide opposite movements

Even the finding of unchanged unpaid time needs a closer look.

In the surveys' separate paid-and-unpaid classification, average unpaid time for females aged six and over was 305 minutes a day in both rounds. Within that national figure, rural unpaid time fell from 317 to 314 minutes, while urban unpaid time rose from 276 to 283. The national estimate combines groups with their survey weights; it is not a simple average of the rural and urban figures. MoSPI, 2019, Statement 22; MoSPI, 2024, Statement 21

A flat headline can conceal differences by age, household composition and circumstance.

The word “unpaid” has a specific scope here. The classification includes goods produced for household use and other unpaid production, alongside domestic services and care. Its 305-minute average is different from the domestic-service and caregiving figures above. It also differs from the working-age unpaid average of 366 minutes. Each figure answers a question about a particular activity grouping and population.

These distinctions are part of understanding the economy itself. If a policy is intended to help people combine paid work with care, the relevant population and responsibilities need to be visible in its evidence. A national mean can start the discussion. Decisions usually require a more specific view of whose day is constrained.

Before asking policy to respond to these differences, we need to know how a day becomes a national statistic.

What a time-use survey measures

India's 2024 survey covered 139,487 households containing 454,192 people aged six and over. The full report lists 3,735 people for whom time-use information could not be collected, leaving 450,457 reporting cases. These are sample counts, not estimates of India's population. MoSPI, 2024 report, Chapter 1, Table 1, p.3

Interviewers collected accounts of the previous day, from 4am to 4am, in 48 half-hour slots, using the International Classification of Activities for Time Use Statistics 2016. For multiple-activity slots, up to three activities performed for at least ten minutes were recorded. Proxy informants were permitted, so these were not all first-person diaries. MoSPI, 2024 fact sheet, pp.2–4; 2024 report, §2.10, p.55

Three measures help turn those records into a useful account. Participation tells us how common an activity was on the reference day. Time per participant tells us how long people who reported it spent in it. Time per person spreads the activity's time across everyone in the specified group, including those who did none of it.

The last measure allows a population's day to be described across all the activity categories. Participant averages cannot be assembled in the same way: the average participant in paid work and the average participant in care need not be the same person.

The method also shapes the apparent day. For the estimates used here, if more than one activity was recorded in a half-hour slot, the slot's time was divided equally among them. Cooking and caregiving recorded together would each receive 15 minutes. The result is an allocated account that fits within 24 hours, rather than a full measure of every concurrent demand. MoSPI, 2024, Chapter 1, pp.2–3

This is particularly relevant to activities such as supervision and care, which can overlap with other tasks. Attention can remain committed even while a person is doing something else. A diary's categories and allocation rules determine how much of that overlap its summary can show.

India already measures these activities. The gap this article investigates is how to connect those records to a fuller account of progress. The 2024 round used computer-assisted interviews and an explicit question about simultaneous activities. We have not adjusted the comparison for possible instrument, interview-mode or informant differences, or tested its statistical significance. MoSPI, 2024 report, Highlights and §1.2

The diary explains the method. Behind it are the people and institutions that turn records into estimates.

Who produces India’s time-use statistics

The 2024 Time Use Survey was produced by the National Statistics Office within the Ministry of Statistics and Programme Implementation, or MoSPI. Its report credits the Household Survey Division with developing the methodology, processing the data and preparing the report. Field Operations Division carried out the fieldwork, while Coordination and Quality Control Division coordinated survey activities. It also acknowledges guidance from the Technical Advisory Group, the Steering Committee for NSS surveys and the National Statistical Commission. MoSPI, 2024 report, Preface

The route from a household to a national estimate has several steps. The survey uses a stratified sample, with villages or urban sampling units selected first and households selected within them. Its year is divided into four quarters to spread collection through the seasons. Interviewers collect the diary information, activities are coded, and the estimation procedure uses sampling multipliers to produce estimates for the relevant population groups. A household’s record is therefore an input into a weighted estimate, rather than a census of everyone’s day. MoSPI, 2024 report, Appendix D

Each step can affect what the final number means. Who is selected, who answers for a household member, which short activities get recorded, and how simultaneous tasks share a slot all matter. Published questionnaires, definitions and estimation methods allow readers to examine those choices. The methods behind a time-use statistic belong alongside the headline it supports.

That makes the time visible in a survey. Whether it also enters GDP depends on a different boundary.

Unpaid work and the boundary of GDP

The distinction between paid and unpaid work also differs from the boundary used in national economic accounts.

Under the 2008 System of National Accounts framework used in India's survey categories, producing goods for a household's own use falls inside the production boundary. Most services that household members produce for their own consumption fall outside it. There are explicit exceptions, including services from owner-occupied housing and those produced by paid domestic staff. System of National Accounts 2008, paragraphs 6.25–6.29

The practical implication is easy to recognise. Preparing a meal for one's own household and preparing a meal as paid domestic work can involve similar human activity but receive different treatment in the accounts. Informal production can still fall inside the production boundary. Saying that all unpaid or informal work is absent from GDP would obscure these differences.

The accounting boundary serves a defined purpose. It does not establish which activity matters most to a family, which experience is meaningful, or whose contribution deserves recognition. A monetary production account and a record of daily life describe different dimensions of the same society.

That is why a broader measurement agenda need not begin by pricing every hour. We can examine how responsibilities are distributed, how much time people can choose for themselves, and whether essential services change those conditions. Those questions remain valuable even when an activity has no assigned market price.

Activity, duration, monetary production, contribution and experience answer different questions. A single score could hide the distinctions this inquiry needs to preserve.

Once that boundary is clear, the next question is who chooses which measures guide public decisions.

Who decides what counts as progress

The institutional arrangement has several kinds of responsibility. MoSPI’s National Statistics Office produces official statistics. Within the ministry, National Accounts Division prepares national accounts, including GDP. The National Statistical Commission has functions concerning statistical priorities, standards, quality and coordination. Technical guidance, statistical production and policy choices are connected, but they are different responsibilities. MoSPI, Roles and Responsibilities, Introduction and sections 1, 7, 26

International classifications provide a common language for measurement. Their use in an Indian survey does not make the UN or World Bank the administrator of its fieldwork. Nor does a national survey automatically require a particular welfare programme or budget. A statistical office can make unpaid activity visible; deciding what public support follows involves government choices and debate.

This is where the political economy of measurement enters our argument. The selection of a headline shapes which questions receive attention. An account centred on paid participation directs us toward employment opportunities. An account that also follows care directs us toward the conditions that make participation possible. Both can be useful. They invite different questions about who supplies essential services, who bears their costs, and whose time adjusts when support is missing.

Households also make decisions within constraints. A family may need an activity performed regardless of whether it carries a wage. If one person takes on more paid work, the responsibility may be shared, transferred, purchased, performed differently or left unmet. A national increase in employment cannot tell us which arrangement emerged. That is a question for evidence about households, services and bargaining as well as work.

The same distinction matters when valuing care. Counting its hours, estimating a replacement wage, deciding who should provide it and deciding who should pay are separate steps. A monetary estimate can inform a debate without settling entitlement, funding or the quality of the care provided. Greater visibility creates an opportunity for political choice; it does not make that choice disappear.

This interpretation of the evidence points toward a practical standard for public discussion: state the measurement boundary, identify the population carrying the responsibility, and make the proposed policy trade-off explicit. It also gives citizens a way to question a reassuring national average without assuming that the statisticians fabricated it. A number can be correctly produced while answering a narrower question than the public needs to ask.

Those choices are being debated internationally, too. India’s diaries offer one concrete way into the Beyond GDP discussion.

The 2026 Beyond GDP agenda

The United Nations' independent High-Level Expert Group on Beyond GDP places this broader measurement problem on the international agenda. Its 2026 report, Counting What Counts, proposes a dashboard of 31 indicators, bringing well-being, inclusion and sustainability into a framework that complements GDP. It is a proposal for how progress could be measured more fully. UN report, Executive Summary, p.viii

One of its indicators concerns time spent on unpaid domestic and care work. The report's annex says that data for this indicator are produced for only 46% of countries and, on average, once every five years. India has a substantial source of evidence in this area; the report identifies a wider international problem of coverage and frequency. This is the report’s coverage assessment; the annex does not provide an underlying country count. UN report, Annex, p.33

The policy implication I draw is that a more complete account of progress requires both broader questions and the statistical capacity to answer them repeatedly. An indicator cannot guide a timely response if the underlying evidence is missing or too infrequent to show change.

The World Bank's Changing Wealth of Nations 2024 approaches complementary measurement from another direction, examining the assets that support production, including produced, human and natural capital. Wealth measures ask whether the productive base is being maintained. Time-use evidence asks how activities are distributed through people's lives. These approaches can inform a richer discussion without being interchangeable. World Bank, Changing Wealth of Nations

A broader account of progress also changes how we read a promise of greater productivity. What happens to any time a new technology might release?

AI productivity and the destination of time

The World Bank's October 2026 South Asia Economic Update, Adopting AI for Growth, supplies a timely reason to ask where productivity gains might lead. Its research compares formal firms' AI adoption and their expectations of future gains. In the surveyed population, 23.4% of Indian firms reported using some form of AI, compared with 42.7% in the United States. World Bank, p.33, Figure 1.16

The study's scope matters. Its AI survey included 1,355 Indian and 401 US formal firms with at least five employees. It re-interviewed firms from the Enterprise Surveys sample and reported weighted estimates. Unpaid household responsibilities sit outside that firm-survey population. World Bank, p.50 and Annex 2.1, p.72

Indian firms' mean expected three-year gain in sales per worker from faster AI development was 3%. The report defines this as the difference between fast- and slow-development scenario expectations, with employment held constant in each scenario. It is not observed growth or recovered time. World Bank, p.55, Figure 2.6; Annex 2.1, p.73

Suppose, as a hypothetical example, that a task becomes quicker. Its consequences could take several forms. A business might produce more within the same hours. A worker might finish earlier, take on additional tasks, or spend time learning a new process. Changes in staffing could affect someone else's income and time. At home, a shorter paid working day might allow rest, or it might make more unpaid responsibilities possible.

Distinguishing these outcomes requires observing working hours, income, household activity and experience alongside the production measure.

The IMF's June 2026 Finance & Development article by Amory Gethin and Emmanuel Saez makes a related measurement boundary explicit. Its account of working hours includes unpaid agricultural work but excludes unpaid home services such as cleaning, cooking and caring for family members. Reading a working-hours measure requires knowing which work it contains. Gethin and Saez, How Much Does the World Work

For South Asia's AI debate, the resulting research question is concrete: as productive activity changes, who gains usable time, whose responsibilities expand, and how do those changes relate to security and well-being? India's earlier survey rounds help establish the distribution we need to understand. Their differences cannot identify an AI effect.

Even if we observed where the hours went, we would still need to ask how they were experienced by the people living them.

Experience and choice within the recorded day

Consider two hypothetical hours recorded as caregiving. In one, someone chooses to spend unhurried time with a relative. In the other, someone struggles to arrange necessary care while worrying about losing paid work. Duration and activity category might look similar. Choice, support and experience could differ considerably.

Researchers already study these dimensions. ADB's March 2026 report, It Is About Time, discusses adding subjective measures to time-use studies and describes regional examples connecting time with well-being and social norms. It provides a starting point for further work on the relationship between activities and experience. ADB, pp.18–19

At Internet of Value Research Foundation, we want to investigate whether brief, voluntary personal observations could add useful context to such records. A person might describe what they noticed during an activity, how much choice they felt they had, or what support would have changed the experience. The useful contribution would depend on whether these questions produce reliable information at an acceptable burden.

More frequent observation also creates trade-offs. Repeated prompts could interrupt the day they seek to understand. People who are comfortable recording their lives may differ from those who decline. A tool that works for one language, literacy level or pattern of device access may miss others. ADB's discussion of digital and passive collection highlights the privacy concerns that such methods can introduce. ADB, pp.38–40

Any personal-observation study would need clear consent, control over disclosure, and a way to evaluate whether the added information is useful. Turning voluntary records into evidence about a country would require a separate sampling and validation strategy. Individual usefulness and national representativeness are different tests.

The test is whether that added context helps us ask better questions about services, work and everyday choice.

Bringing the whole day into policy

The most immediate policy application is to make time an explicit outcome to investigate.

For care policy, the questions include how responsibilities are shared, whether support is available when needed, and what changes for people combining paid work with care. Employment counts alone cannot answer them. Nor can a care-time total establish whether the care provided is adequate or whether its provider has meaningful alternatives.

For transport and infrastructure, evaluation could examine changes in the time required for everyday tasks, alongside access, cost and reliability. If a service saves time, the next question is how that time is used and who benefits. Its value cannot be read solely from a shorter task duration.

For AI and workplace policy, studies could pair productivity measures with working hours, earnings, predictability and household responsibilities. The relevant outcome might be additional output, greater security, less strain, more choice, or a mixture. These possibilities need observation rather than assumption.

India is a starting point for this inquiry. Regional work must retain differences in survey years, age coverage, activity classifications and diary design before comparing countries. A South Asian account of human time will be stronger when it builds from those specific national records.

The opening figures leave us with a question that is both ordinary and consequential. A larger share of working-age women recorded paid activity, while average unpaid time remained near six hours a day. What would make that greater participation part of a more manageable, secure and freely chosen life?

Sources and methods

Primary estimates belong to MoSPI. This essay reproduces selected published tables and interprets what they can—and cannot—tell us.