Jobs and incomes: the labour survey, formal payrolls and rural work

The Jobs dimension asks “Are people finding work and earning more?” This note covers the data behind /economy/jobs/ and its sub-pages.

Code: pipeline/tipsheet/compute/economy_jobs.py; publish/economy_households.py (publish_jobs); step economy_jobs in steps/economy_households.py (needs macro, economy_households; about 2 seconds). Tests: pipeline/tests/test_economy_households.py. Bundles: economy/jobs/monthly, quarterly, payrolls, feelings, mgnrega_years, c86, c87, facts. Pages: /economy/jobs/ (overview), survey/, formal/, rural/. Pre-registration and results: C86, C87 in docs/research/site_calculations_spec.md.

The labour survey (PLFS)

Source. MoSPI’s PLFS through the Data bank (mospi_plfs, latest vintage).

• Monthly from April 2025 (all India, rural, urban) and quarterly (urban from April–June 2018; rural and all India from April–June 2025). Current weekly status: a person is employed if they worked at least an hour on any day of the week before the interview, paid or in a household enterprise. Persons 15 and over; youth is 15–29.
• Rates: unemployment rate (% of the labour force), labour force participation rate and worker population ratio (% of the population), for persons, women and men. The gender gap is men’s participation less women’s.
• Checks. Persons’ rates come from /macro/’s identity-checked build (macro.build_jobs): rows where UR ≠ 1 − WPR/LFPR beyond rounding are blanked, with MoSPI’s all-ages UR deciding which rate is wrong (two urban quarters so far). Rates by sex and age come from the same table without that check.
• Breaks. MoSPI redesigned the sample from January 2025; the quarterly urban series is marked at the join. MoSPI publishes no standard errors. Usual status (the annual measure over the past year) is not used.

Formal payrolls

• EPFO net new subscribers (joiners less exits plus rejoiners), monthly, by age and sex, via India Data Hub (publisher EPFO). The newest month held is July 2025; no later EPFO release was found (labour ministry and press, searched 24 October 2026), and the pages say the releases stopped.
• ESIC new insured workers, by sex, via India Data Hub (MoSPI’s payroll reporting), to July 2026. February and March 2024 are missing at source, so twelve-month totals are blank for the year after.
• Measures: twelve-month sums and their growth on a year earlier; the share of 18–25-year-olds and of women in twelve-month EPFO additions; women’s share of ESIC’s.
• Not used: India Data Hub’s weekly EPF contributing-members series, which is sparse after May 2025 and jumps by millions between weeks.
• NACH bulk credits: NPCI’s NACH credit statistics, “other credits” (credit_306__other_credits plus credit_156__other_credits), successful value, in ₹ crore: salaries, pensions, dividends and interest, excluding the government’s benefit transfers. Shown as a flow of income, never a jobs count (C87).

Rural work

• MGNREGA by financial year from the scheme’s national dashboard (mgnrega_at_a_glance, newest capture): persondays, households and individuals that worked, average days per household, average wage rate per day, households given the full 100 days, women’s share of persondays, expenditure. Past years are complete; the current year is partial and kept apart.
• VB-G RAM G, the successor from 1 July 2026 (125-day guarantee), from its own dashboard (vbgramg_at_a_glance). The two are never added or spliced: the current year shows both side by side.
• Real wage: the wage rate’s growth over the year deflated by rural CPI inflation, taken as the mean of the twelve monthly year-on-year rates of the linked rural CPI (macro_cpi_monthly.cpi_rural_linked_yoy), an approximation to the growth of the year’s average index. CPI for agricultural and rural labourers is not yet held.

What households say about jobs

RBI’s urban and rural confidence surveys, net responses on employment and income (see economy_households.md), charted inside the no-download wrapper.

C86 and C87

• C86 (feelings and the unemployment rate): four-quarter changes in the urban net response on employment against changes in PLFS urban unemployment, lockdown quarters out. Result: 13 changes, Spearman +0.10, permutation p 0.75 (+0.31, p 0.38, without the redesign span); levels −0.24. Does not pass.
• C87 (NACH credits and hiring): three-month totals’ growth on a year earlier from April 2022. Against EPFO, r −0.36 over 40 months, circular-shift p 0.59; against ESIC, −0.24 over 44 months, p 0.27. Does not pass.

What these numbers cannot show

Low unemployment on current weekly status counts an hour’s work, unpaid family work and self-employment as employment, so it misses underemployment and low pay. Most women of working age are outside the labour force on these measures, and unpaid work at home is not counted. Payroll data count enrolments, including formalised existing jobs. No monthly market wage is held; the MGNREGA wage is a guaranteed minimum. Labour Bureau rural wages and PLFS earnings are candidates to add.