Latest news with #PeriodicLabourForceSurveys
Mint
6 days ago
- Business
- Mint
Himanshu: India's economy is too complex to afford less than robust statistics
Recent months have seen a flurry of data and report releases from the National Statistics Office (NSO). While some of these are routine surveys conducted by it, in some cases the NSO has made significant changes in the nature of data made available and the frequency of releases. India's statistical system had come under criticism for denying or delaying access to survey data. We have seen delays in updating the base year for key variables like the Consumer Price Index (CPI) and National Accounts, both of which have a base which is more than a decade old. Improved survey coverage and data-release frequency would help generate confidence in our statistical system. But it is also essential for statistics to serve as inputs for economic analysis and policy formulation. Regular base-year updates for many macro variables are necessary, given the economy's dynamism. Also Read: Headline labour force survey data masks a pressing employment problem The NSO has expanded the coverage of its annual Periodic Labour Force Surveys (PLFS) and also increased the frequency of data releases from quarterly to monthly, starting with April 2025. Going monthly has meant that the NSO had to increase the PLFS sample size by 2.65 times to 272,304 households, together with changes in sampling design for the generation of monthly estimates. A larger sample also lets the NSO release quarterly estimates for rural areas (done only for urban areas so far). PLFS data has been a valuable tool to track trends and patterns in India's workforce structure since 2017-18. Along with their precursor Employment-Unemployment Surveys (EUS), available since 1972-73, PLFS results provide a comparable data series on employment patterns. However, these also remain the only credible source of information on the quality of job and earnings from them. While there are several sources for the wages and earnings of casual workers, EUS-PLFS data-sets are the only source of information on the earnings of regular workers, who account for almost one-third of all workers in the economy. The monthly report for April 2025 released on 15 May is the first of the monthly series. It presents estimates of the Labour Force Participation Rate (LFPR), Workforce Participation Rate (WPR) and Unemployment Rate (UR) by 'weekly status' for rural and urban areas. Also Read: Himanshu: What consumption data reveals of India's economy For the latter, the findings are broadly similar to what was revealed by the quarterly report for October-December 2024. But its rural estimates are significantly at variance with findings of the annual report for 2024, with a lower LFPR and WPR in April 2025, and significantly higher unemployment rate for the 15-29 age group (and also for the country's 15-plus population). While monthly estimates of basic indicators are useful, these are of limited relevance for an economy whose employment structure is very diverse and complex. Unlike rich countries where most workers have regular payroll jobs in non-farm sectors, the Indian workforce relies mostly on informal employment. Even today, almost half of all Indian workers are engaged in agriculture, compared to less than 5% in most developed countries. Variations in the LFPR, WPR and UR are less relevant in an economy with a large proportion of the population vulnerable in terms of job quality and income assurance. If the purpose of the monthly series is to provide meaningful insights into our labour market, it requires detailed data on sectoral shares, the nature of enterprises and earnings from employment. Fortunately, the data lets us generate most of these estimates. The PLFS's re-introduction of land information allows rural analysts to delve deeper. Also Read: TCA Anant: How India's statistical system could win the ongoing war of narratives The principal challenge now relates to how we understand and characterize the labour market. Given the emergence of new employment categories such as gig work and new forms of labour arrangements interlinked with land and credit markets, we need a better understanding of what holds back the creation of quality jobs in the economy. The revamped PLFS series also expands its questionnaire on education and skills, which have emerged as important drivers of changes in the economy's employment structure. While the NSO has stepped up to provide the basic data necessary for us to analyse and understand the complexity of the country's labour market, deeper research is now needed for this move to spell meaningful policy engagement. Research and policy must look beyond basic estimates of the WPR and unemployment rate. Expanding the monthly release to include wages/earnings, job quality and other relevant co-variates would aid the process of analysing India's employment challenge. The author is associate professor at Jawaharlal Nehru University and visiting fellow at the Centre de Sciences Humaines, New Delhi.
The Hindu
22-05-2025
- Business
- The Hindu
Analysing poverty levels in India by comparing various surveys
Himanshu et al, 'Poverty Decline in India after 2011–12', The Economic and Political Weekly, Vol 60, Issue No: 15, April 12, 2025 A recent paper has estimated that poverty reduction in India slowed down significantly after 2011-12. While poverty levels of 37% in 2004-05 fell to 22% by 2011-12, it has since fallen only by 18% in 2022-23, the paper finds based on its own calculations. The paper, titled 'Poverty Decline in India after 2011–12: Bigger Picture Evidence', authored by Himanshu of Jawaharlal Nehru University, and Peter Lanjouw and Philipp Schirmer of the Vrije University in Amsterdam, noted that India hasn't had an official poverty estimate since 2011-12. In the absence of an official estimate, a number of unofficial and often contradictory estimates have been made, of which this one is the latest. Three methodologies The paper notes that the various contradictory estimates can essentially be clubbed into three broad buckets based on their methodology. The most common approach, it noted, has been to use alternative socio-economic surveys of the National Sample Survey Office (NSSO), since there are significant comparability issues between the Household Consumption Expenditure Survey (HCES) of 2022-23 and 2011-12. There are no intervening surveys, either. The HCES for 2017-18 was scrapped by the government, citing 'methodological issues'. In the NSSO's 71st round, which covered the January-June 2014 period, the government introduced a consumption expenditure measure that was derived from a single question in the survey called the Usual Monthly Per Capita Consumption Expenditure (UMPCE). This UMPCE was used for all subsequent rounds of the NSSO surveys as well as in the Periodic Labour Force Surveys (PLFS). However, as the authors correctly note in their paper, this measure can't be compared to earlier estimates of consumption because it is based on a single question 'with no clear definition of what it comprises'. According to this method, poverty estimates range between 26-30% for 2019-20. The second approach has been used by the economist Surjit Bhalla and his colleagues in 2022 in a paper in which they used Private Final Consumption Expenditure (PFCE) estimates from the government's National Accounts Statistics (NAS) to derive consumption aggregates after 2011-12. This method basically scaled the consumption expenditure data from the HCES 2011–12 based on the implicit growth rate of PFCE after 2011-12. The third broad approach — and the one used by the authors themselves — is to use survey-to-survey imputation methods. This basically means data gaps in one survey can be filled using information from a related base survey. This method, the authors note, has occasionally been used by World Bank researchers to update the World Bank's Poverty and Inequality Platform (PIP) database. Looking at different surveys This approach is significantly prone to somewhat divergent results, based on the different surveys used to complement each other, but are useful in revealing trends in data. For example, the paper notes that one estimate by David Locke Newhouse and Pallavi Vyas used the 2011-12 HCES and the 2014-15 survey on Consumption of Services and Durables to estimate that poverty in India declined from 22% in 2011-12 to 15% in 2014-15. Similarly, Ifeanyi Nzegwu Edochie and their colleagues in 2022, used the 2017-18 survey on Social Consumption on Health to estimate poverty at 10% for 2017–18, which confirmed the trend that poverty had reduced since 2011-12. In 2025, Sutirtha Sinha Roy and Roy van der Weide used a radical approach to apply the survey-to-survey imputation using a private sector survey. They used the Consumer Pyramid Household Survey (CPHS) for 2019 by the Centre for Monitoring Indian Economy (CMIE) along with the 2011-12 Consumer Expenditure Survey (CES). Their estimate was that poverty was around 10% in 2019. Himanshu et al also use this survey-to-survey imputation method. However, the authors note that their strategy differs from previous attempts in three aspects. First, they have used the Tendulkar Committee's poverty lines as opposed to the World Bank's poverty lines. Second, they have used the employment surveys of the NSSO for imputation. The Employment-Unemployment Survey (EUS) is a companion survey to the 2011-12 CES, and is based on similar sampling design and survey implementation procedures. Further, the PLFS, which replaced the EUS in 2017-18, is modelled on the EUS, the authors note. What this essentially means is that the two surveys Himanshu and his colleagues used to impute data are similar in their methodology and parameters, yielding a more accurate fit in the data. Third, the authors note that, unlike the World Bank studies, their own imputation models are estimated at the State level or include State-fixed effects when estimated at the sector level. Their methodology shows that while poverty based on the Tendulkar Committee poverty lines fell sharply between 2004-05 and 2011-12 — from 37% to 22% — it subsequently has fallen only to around 18% by 2022. Based on these estimates, the authors add, the number of poor persons in India fell only slightly since 2011-12, from 250 million persons to about 225 million in 2022–23. Different trends across States State-level trends derived from their methodology suggest differing trends across States over this period. Notably, the authors find that Uttar Pradesh, India's most populous State, seems to have markedly reduced its poverty rate. 'However, in other historically poor States, such as Jharkhand and Bihar, progress was much slower,' they added. 'It is noteworthy that in several of the large central and southern States, such as Maharashtra and Andhra Pradesh, poverty reduction appears to have stagnated.' Importantly, the authors do acknowledge that 'a full resolution of the present debate' on poverty is unlikely to be forthcoming without new government data that can be compared with previous years' data. However, they also try to back up their findings using other data sources that point to the same conclusions. For example, they noted that the growth of India's Gross Domestic Product (GDP), which averaged 6.9% per annum between 2004-05 and 2011-12, slowed to 5.7% between 2011-12 and 2022-23. This, they said, is consistent with a slower decline in poverty reduction after 2011-12. Similarly, they point out that the Wage Rates in Rural India (WRRI) data compiled by the Labour Bureau on real wages points to a slowdown in wage rates. It shows that the annual growth rate of wages fell from 4.13% per year between 2004-05 and 2011-12 to 2.3% per year between 2011-12 and 2022-23. Thirdly, the authors point out that while the absolute number of workers in agriculture declined by 33 million between 2004-05 and 2011-12, and by a further 33 million by 2017-18, this trend has reversed since then with 68 million workers being added to the agriculture sector since 2017–18. One consequence of the rising workforce in agriculture, the authors point out, has been the decline in the growth of agricultural productivity in recent years. Lower productivity leads to lower wages, which leads to higher poverty levels. This paper is hardly going to be the last word on poverty estimates, with many more sure to follow. However, as the authors themselves conclude, there's more than enough evidence to show that poverty reduction efforts need to be accelerated.
