Labor Force Surveys

India’s Jobs Data Just Got a Major Upgrade More than just a survey change. PLFS, India’s Periodic Labour Force Survey tracks who is working, who isn’t, and who’s looking for work. It’s the government’s main tool to understand employment trends across the country. Old job data left rural India in the dark We were planning policy on partial truths. The old PLFS had big flaws: Rural data missing in quarterly reports Out-of-sync annual cycle Too small a sample size Result? We missed seasonal trends, migration shifts, and real-time crises. From April 2025, things changed: Monthly job data for both rural & urban areas Sample size doubled — 1 lakh to 2.72 lakh households Continuity — 75% households repeated monthly, 50% across quarters Richer insights — landholding, income sources, vocational training Why it matters: April 2025 LFPR: 56% overall (Rural 58%, Urban 51%) Female LFPR shocker: Rural 38%, Urban 24% Urban female unemployment: 9% vs 6% for men Youth unemployment: 14% overall; urban young women 24% What monthly data enables: Targeted interventions for women’s employment Reviving failed schemes like Mahila Shakti Kendra Youth-specific job programmes in high-unemployment areas Dynamic rural support like extra MGNREGA days during slumps Think of it as switching from yearly to weekly pantry checks. You spot shortages sooner. You fix them faster. Bottom line: The new PLFS moves India from reactive to proactive in tackling jobs. The only trade-off? We can’t compare it neatly to older datasets. But that’s a small price for sharper, faster, better jobs intelligence. Jobs data isn’t just numbers. It’s livelihoods. And now, it’s real-time. Do you think monthly jobs data will actually lead to better policy? #Jobs #Economy #PLFS #IndiaData #Employment

It's the 'NO SHlT SHERLOCK' heard round the world: The July Jobs Report. Not the July numbers themselves - which came in at 73,000 new jobs (just half of what was expected) - but the revision to May and June's numbers. They represent two of the largest single month revisions since 1979: 125k and 133k respectively, for a combined total of 258,000 jobs. It's a hell of a headline. But it's not the whole story, because taken in isolation, these numbers lack context and nuance. So. I'm here to provide both. I'm not even going to address the elephant in the room, which is the President's dismissal of the BLS commissioner, an act of unabashed political grandstanding, because the Bureau of Labor Statistics is an independent statistical agency and the work they do is both non-partisan and non-political. Full Stop. But it's also imperfect. I've written on this topic ad nauseum and will link to my lengthy article outlining how of the Employment Situation Summary actually works, but here is the abridged version: Each month, during the week that includes the 12th day, the Census Bureau collects survey data from 121,000 employers representing 631,000 worksites, to measure wages, hours worked and most-notably, job creation. The super smart data scientists at BLS then use this data to estimate, extrapolate, model and make sense of what's happening in the labor market, all under the mandate to report their findings by 8:30am on the first Friday of the following month. This is a behometh undertaking, the scale of which normies like me can scarcely wrap our heads around. And because of this, the results are - by definition - preliminary and subject to multiple revisions. The first two happen in each of the two months immediately following initial publication. And why is that? It's because only about 70% of participants in the Current Employment Statistics (CES) - AKA: "The Establishment Survey," actually submit their data on time. The remaining 30% of the data trickles in over the next several weeks. Some never trickles in at all, because guess what: participation in the CES is voluntary. In addition to two monthly revisions, BLS conducts a benchmark revision twice annually, reconciling the monthly numbers with the results of the Quarterly Census of Employment and Wages (QCEW). The preliminary benchmark revision is published with the August numbers in early September, while the final benchmark revision is published with the January numbers in early February. You may recall a lot of hoopla last September when 2024's preliminary benchmark revision estimated that the jobs numbers had been overstated by 818,000. Which is A LOT, but not actually the highest ever. What you probably don't recall is that 2024's *final* benchmark revision (published in February of this year) came in much lower, at 598,000. Also a lot, but NOT abnormally so. For context, 2019's benchmark revision came in 501,000 lower than initially reported. (Con't).

Every policymaker in India appears worried about job growth. But jobs are still not a big part of RBI's policymaking process. In fact, we don’t have reliable estimates of how monetary policy impacts job creation. Indian policymakers have long been content with once-in-five-year updates of workforce numbers provided by the quinquennial rounds of the National Sample Survey (NSS). Economists held that most Indians were too poor to be unemployed. In other words, the informal labour market was considered to be a residual sector, always offering ‘work’ to those who lacked regular ‘jobs’. The Indian definition of employment included informal work arrangements. Official unemployment rates remained low and stable, even as ‘disguised unemployment’ levels stayed high. It was only in the late 2000s that the idea of a high-frequency labour-market tracker gained ground. The Indian economy had moved into a faster lane by then and the salaried class had grown substantially. The issue of jobs—as opposed to work—had started gaining political traction. The National Statistical Commission (NSC) appointed a committee headed by Amitabh Kundu to prepare a roadmap for conducting a periodic labour force survey (PLFS). The Kundu committee report, submitted in 2010, recommended quarterly updates for urban India and annual updates for the countryside. After field trials in three states and several rounds of deliberation, the PLFS survey was finally launched in 2017-18. Unfortunately, the survey report got mired in controversy even before it was released. The PLFS has come to be used widely by policymakers since then. The lags with which the data was released have declined over time. The National Statistical Office (NSO) is planning to come up with a monthly PLFS series soon. This series could be a useful input for the central bank’s policy decisions in the months and years to come. However, the new series should be subjected to adequate quality checks and cross-validation before it is used as a policy target. High-frequency updates from other surveys such as the ASI, ASUSE and the upcoming services survey (ASSSE) could be useful complements to the monthly PLFS series in the future. The NSO should also review the reliability of administrative datasets such as EPFO. To provide a clear and comprehensive view of the labour market, the NSO and NSC should jointly evaluate how far different labour-market indicators track each other, and how far they track other high-frequency indicators. Jobs have become an important part of India’s electoral and fiscal policy discourse. They will become a part of the country’s monetary policy calculus only when the official statistical system provides high-quality labour-market data at high frequency. Full article: https://lnkd.in/gTwJVQMy

The bottom line for the August Employment Situation is that hiring is trending lower and is concentrated in a few narrow sectors. Unemployment is rising bit-by-bit but the restraint is due more to reshaping the demographics of available labor in the job market. Wage growth continues. However, it may be an artifact of hiring taking place for skilled workers who can command bigger compensation but even that seems to be ebbing. When the FOMC meets on September 16-17, it will have more and more convincing evidence that the mandate for maximum employment could use a little support from a rate cut. However, that will have to be balanced against the next round of inflation data for August – the PPI on Wednesday the 10th and the CPI on Thursday the 11th. If the price increases associated with tariffs are more visible in the data, anything larger than a 25basis point cut is unlikely. Nonfarm payrolls were up 22,000 in August and well below the market consensus. There was a net downward revision of 21,000 to the prior two months. Private payrolls rose 38,000 in August while government payrolls were down 16,000. For the third quarter to date, monthly payroll increases averaged 51,000, down somewhat from the average of 55,000 a month in the second quarter and lower than the average of 111,000 in the first quarter. Overall hiring has slowed substantially with mixed conditions across industries. Goods-producers payrolls were down 25,000 in August with declines in all major categories. Private sector service-providers’ payrolls were up 63,000 in August, but the majority of that was from a 46,800 increase in health care and social assistance and 28,000 in leisure and hospitality. Average hourly earnings are up 0.3 percent in August from July and up 3.7 percent year-over-year. In the aggregate, workers continued to receive moderate raises but the trend is slowing. The unemployment rate rose one-tenth to 4.3% in August and was in line with market expectations. This was the highest since 4.5% in October 2021. The labor force was up 436,000 to 170.8 million in August, with the number of employed up 288,000 and unemployed up 148,000. A 4.3% unemployment rate is not a bad one but it needs to be read in the context of underlying chances in the labor market due to the massive influence of policies from the Trump administration. There were indications of a less positive situation for workers in August. The number of people working part time for economic reasons was up 65,000 to 4.749 million. Job losers were up 32,000 to 3.437 million. The number of job leavers was unchanged at 784,000 in August. New entrants to the labor force are down 199,00 to 786,000 while reentrants are up 107,000. #nonfarmpayrolls #unemploymentrate Please do not use without attribution. Prepared without use of AI. Copyright © Theresa A Sheehan

At Data For India, our Measurement vertical (https://lnkd.in/ga5e7eDD) has gained an engaged following, and with good reason. Indian data is impressive in its history, scope and quality, but writing around the history of datasets, changes in methodology and their implications is sparse. Our latest publication by Pramit Bhattacharya and Nandlal Mishra attempts to map India's Periodic Labour Force Surveys, which must be one of the most widely used and studied national labour datasets in the world. We look at the evolution of the PLFS, changes in methodology, key uses of the data and debates around the survey. We hope it is useful to anyone who works with or wants to work with labour data on India. https://lnkd.in/gz2JznE9

The Bureau of Labor Statistics Monthly Job Report: it’s based on two surveys the establishment & household. The survey in question is the former. It’s based on a 121K survey from non-farm private employers & government agencies comprising 631K worksites. The response rate has recently slipped below 60% down from pre pandemic rate of 70% and well below the 80% rate typical over a decade ago. The survey has a +/- of 100K each month. Typically with a three month window the response rate for a single monthly survey time improves to 90% which is why there are ongoing rolling revisions that are rarely as large as the downward 258K revision in June. The numbers are not rigged and they are the gold standard internationally.