The Participation That Is Not Yet Employment

Pritiraj BrahmaFounding Partner, Pritiraj & Partners · 8 February 2026

India’s female labour force participation rose eighteen percentage points in six years — and two-thirds of employed women were self-employed, in a category that requires nobody to pay them. Then the survey was rebuilt, and in its first year the composition finally moved: self-employment down 2.3 points, regular salaried work up 1.6. The number that matters has started to move. It is not the number the country quotes.

ProvenancePublished 8 February 2026; revised 28 August 2026 to argue from the Ministry’s primary releases rather than secondary analysis, and to incorporate the Periodic Labour Force Survey Annual Report 2025. Participation and worker-population figures to 2023-24 are from the Ministry of Statistics and Programme Implementation press note of 23 September 2024. Status-in-employment shares for 2024 and 2025 are from the press note on the Annual Report 2025, the first edition on a calendar-year basis. The Ministry cautions that results from January 2025 are not comparable with earlier publications because of methodological change; readers should treat the eighteen-point series and the 2025 composition figures as two series, not one.

India’s female labour force participation rose eighteen percentage points in six years — the fastest improvement of any major economy. Almost none of it arrived as a wage: two-thirds of employed women were self-employed. Then, in the first year the survey was rebuilt, the composition finally moved — female self-employment down 2.3 points, regular salaried work up 1.6. The number that matters has started to move. It is not the number the country quotes.

India has a genuine achievement to report on women’s work, and for most of a decade it was reported in the wrong unit. According to the Periodic Labour Force Survey, female labour force participation for those aged 15 and above rose from 23.3 per cent in 2017-18 to 41.7 per cent in 2023-24. The Ministry of Statistics and Programme Implementation’s press note of 23 September 2024 records the male rate over the same period moving from 75.8 to 78.8 per cent — three percentage points against eighteen.

Labour force participation counts anyone working or seeking work. It does not ask what the work pays, whether it pays at all, or whether it will exist next quarter. When the composition of those eighteen points is examined the achievement narrows considerably — and when the survey’s 2025 edition is examined, something more interesting happens.

Six findings anchor this analysis:

1. The rise was real and steep. Female participation on the usual-status basis went from 23.3 per cent in 2017-18 to 37.0 per cent in 2022-23 to 41.7 per cent in 2023-24. The worker population ratio, which counts only those actually employed, rose from 22.0 to 40.3 per cent over the same period. This was not women queuing fruitlessly; on the Ministry’s own figures they were being counted as employed.
2. It was overwhelmingly rural. Rural female participation rose from 24.6 per cent to 47.6 per cent across those six years, taking rural female participation above the national average — an inversion of the pattern in most economies, where urbanisation drives women’s employment. Rural female worker population ratio held at 44.9 per cent in 2025.
3. It was overwhelmingly self-employment, and the primary figures now confirm it. The Ministry’s press note on the Annual Report 2025 records the self-employed share of employed women at 66.5 per cent in 2024, against a regular wage or salaried share of 16.6 per cent. Self-employment in the survey includes own-account workers and helpers in household enterprises, whether paid or unpaid. Two-thirds of employed Indian women were in a category that does not require anyone to pay them.
4. In 2025 the composition moved, for the first time, in the direction that matters. The same press note records the female self-employed share falling to 64.2 per cent and the female regular wage or salaried share rising to 18.2 per cent. The shift is visible for men too — self-employment 52.9 to 52.0 per cent, regular salaried 25.4 to 26.5 per cent — and casual labour was broadly stable at about a fifth of all employment. A 1.6-point rise in one year is not a transformation. It is the first year the number this essay argued should be the headline actually improved.
5. The series was rebuilt in January 2025, and the Ministry says so. From January 2025 the survey moved to a calendar-year basis with methodological changes, and the press note explicitly cautions users against comparing results from that point with estimates published up to December 2024. The 2025 report covers 22,594 first-stage units, 2,70,472 households and 11,48,634 persons. The eighteen-point trend line and the 2025 composition figures therefore belong to two different series, and anyone quoting them as one line is misreporting.
6. The annual and weekly measures still disagree. On the Current Weekly Status basis, which asks about the preceding seven days rather than the preceding year, female participation rose from 21.1 per cent to 35.6 per cent over the earlier period — roughly six points below the usual-status figure. The gap is itself a finding: a substantial number of women counted as in the labour force across a year are not in it in any given week.

Enrolment in the labour force is not a wage

The survey classifies a person by the activity pursued for a majority of the reference period, with a subsidiary activity captured alongside. Including unpaid family labour is a defensible choice, consistent with international practice: work on a household farm or shop is economic activity, and excluding it historically erased a great deal of women’s work from the record.

But a category built to capture work has been read as a measure of livelihoods. A woman who spends part of the year assisting on a family plot without wages, and who would take paid work if it were available, appears in the same aggregate as a woman in a salaried post. Both are in the labour force. Only one has an income of her own. Between 2017-18 and 2023-24 the female workforce grew and became less waged: regular salaried employment — the category associated with social protection, written contracts, predictable hours and the possibility of promotion — lost share, casual labour lost share, and self-employment absorbed the difference.

Two readings of that were available and not mutually exclusive. One, argued in a finance ministry paper led by the Chief Economic Adviser, is that the feminisation of agriculture reflects a structural shift in which surplus male labour moves out and remaining female labour is used more efficiently. The other, argued by economists including Santosh Mehrotra, is distress participation: households under income pressure putting more members to work, with women absorbing the least remunerative end. The 2025 movement is mild evidence for the first reading and does not close the question, because the survey still does not publish net earnings for the self-employed in a form that would settle it.

The series that cannot be quoted as one line

The break matters more than it appears. For seven years India had a single comparable series on the most-cited social statistic in the country, and the eighteen-point rise was its headline. From January 2025 that series was replaced by a better one on a calendar-year basis, and the Ministry told users plainly not to bridge the two. That is correct statistical practice and it has a cost: the country’s most quoted number about women’s work now has a discontinuity running through the middle of it, and no published bridge estimate.

This is a recurring pattern in Indian statistics, and it is worth naming as a design failure rather than a technical footnote. A series is improved; the improvement is announced with a warning; and no back-cast is published. The result is that the improved series cannot be used to argue about the past, and the old series cannot be used to argue about the present. Whoever wants to claim progress or deny it can now pick a base year on either side of the break and be technically correct.

The counter-case, honestly stated

Three arguments cut the other way, and each has force.

First, part of the increase was better measurement, not new work. Survey questionnaires were revised over the series to probe more carefully for activities women often do not self-report as work — kitchen-garden cultivation, livestock care, unpaid family labour. If the survey is now capturing labour that was always occurring, the earlier figures were understated and the trend line overstates the change. That is a criticism of the 23.3 per cent, not of the 41.7 per cent, and it does not weaken the compositional point at all.

Second, participation may be a necessary precursor. In most economies women’s employment follows a U-shape against development: high in subsistence agriculture, falling as household incomes rise, rising again with education and services. Entry into the labour force, even unwaged, can be the step that makes later waged work thinkable within a household. The 2025 shift toward salaried work is consistent with exactly that sequence, and dismissing the participation rise as illusory would have been its own error.

Third — and this is the objection to our own framing — a 1.6-point rise in the female salaried share is a single year’s movement on a rebuilt series, and one year is not a trend. It could reverse. It is also, on the Ministry’s own caution, not strictly comparable with anything before it. An essay that spent five findings warning against over-reading a number should not over-read this one either. The honest statement is narrow: the composition moved the right way in the first year it was measured properly, and it needs three more years before it means anything.

What we would do

1. Split the self-employed line in the headline release. The survey already distinguishes own-account workers from helpers in household enterprises, and paid from unpaid helpers. That split belongs in the press note and the Economic Survey table, not only in the annexures. A single published column ends the ambiguity between two-thirds of working women being entrepreneurs and two-thirds being unpaid family labour.
2. Make female regular salaried share a published headline, state by state. The 16.6 to 18.2 per cent movement is the most important number in the 2025 release and it appears in a paragraph about status distribution rather than as a headline. Published annually by state, it is the figure that moves only when a woman gets a job with a wage attached, and therefore the figure worth being accountable for.
3. Publish a bridge estimate across the January 2025 break. Not a revision of the old series, and not a claim of comparability — a published, documented back-cast of the new methodology onto two or three earlier years, with its confidence range stated. Without it, the country cannot say whether the composition shift began before 2025 or with it, and every participant in the argument gets to choose a base year.
4. Publish earnings for the self-employed. Until net earnings from self-employment are published as regularly as wages for salaried and casual workers, the distress-participation reading and the efficient-reallocation reading cannot be adjudicated. The survey collects enough to make a start.
5. Give the childcare question an owner. Anganwadi timings, creche provision under the Maternity Benefit Act, and the childcare component of state labour policy sit with three different departments in most states and are nobody’s reported outcome. Name a secretary, set a coverage target for children under three in the ten districts with the widest gap between female participation and female salaried employment, and publish attendance monthly.

India moved a number that had not moved in thirty years, and then, in the first year it rebuilt the instrument, moved the harder number underneath it. Both facts deserve saying plainly. What does not deserve saying is that the problem is solved: 64.2 per cent of employed women remain in a category that requires nobody to pay them, the series that would show whether this is a turn or a blip has a discontinuity through the middle of it, and no state is required to publish the one figure — women in waged work — that would make the claim testable.

Companion essay. Read alongside The Half That Does Not Show Up — India's Female Participation Problem.