In a recent interview, Prime Minister Narendra Modi apparently said, “more than a lack of jobs, the issue is a lack of data on jobs”.
For those of us who have been using the National Sample Survey Organisation (NSSO) data on employment and unemployment for decades now, such a statement comes as a surprise as almost all research on several aspects of Indian labour markets has been based on data provided by the NSSO employment-unemployment surveys (EUS). Even field-based studies use the NSSO-EUS for their background research.
There, of course, exists other sources of data on employment and unemployment, those available from the economic census (establishment surveys) – Census of India, Annual Survey of Industries, the Directorate General of Employment and Training (DGET) – but these suffer from limited coverage and therefore are not representative. The Census of India although provides exhaustive estimates of workers under the broad categories of rural-urban, male-female, main-marginal and a few others once in a decade, it does not provide detailed information like the household surveys.
Unfortunately, in a recent move by the government, the NSSO-EUS has been officially discontinued since the last available figures for the 68th Round (2011-12). It is thus a fact that post-2011-12, we do not have any extensive information on the labour market situation in India.
If we consider the six rounds of annual employment-unemployment surveys conducted by the Labour Bureau under the aegis of the Ministry of Labour and Employment since 2010-11, which have been somewhat similar in size and methodology to the NSSO-EUS from its third round (2013-14), even those have not been continued after 2015-16.
These surveys have been scrapped based on the recommendations of a task force headed by a well-known economist and the former head of NITI Aayog, Arvind Panagariya. The committee pointed out that since the surveys are carried over a period of one year and are made available at a lag of another year, it makes the data lose its relevance and desirability for informing policymaking. The data is then unable to take into cognizance the rapid changes taking place in the economy. It has also been pointed out by the task force that the NSSO-EUS is low-frequency data since it is carried out only once in five years and hence is unable to inform policy at regular intervals. The committee thus recommended developing a data system that would be replacing the existing five-yearly surveys, and would be addressing the issues of representation, periodicity, timeliness and hence improve the reliability and usability of the data. The report of the task force is available here.
In the meantime, due to a lack of statistics on employment and unemployment, several sources are being explored in order to get an idea of the current scenario. One such among them has been a research commissioned by the government whose results are being widely discussed and contested. This study by Ghosh and Ghosh used information based on payroll data which included using the figures from the Employees’ Provident Fund Organisation, Employees’ State insurance, State Employment Registers and so on in order to get an idea of the employment situation. The study reportedly claimed an employment generation of seven million jobs in the FY2017-18. These sources of payroll data are entirely based on self-reporting mechanisms and are not survey-based information. It thus fails to provide any adequately reliable information on the current situation.
Researchers, economists and statisticians have already written several newspaper articles identifying the problems of over or underestimation associated with such sources.
In addition, the above-mentioned study has convoluted the overall context of the debate and discourse on the prevailing levels of joblessness in the economy and has led to unfounded conjectures and speculations. Survey-based information whether enterprise based surveys or household surveys, work on the principle of recording and enumeration by trained investigators and are based on detailed questionnaires. Hence they form a much more reliable tool for collecting and disseminating information and thus inform policy. It is in this overall context, that the government needs to be reminded about the importance of the NSSO-EUS, the purpose it served and the opportunity cost to be borne by the economy for an overall scrapping of the surveys.
Significance of the NSSO-EUS
The NSSO-EUS, like all the other surveys conducted by the NSSO, embodied the idea with which the organisation was instituted in 1950 by professor Prasanta Chandra Mahalanobis, widely regarded as the father of Indian statistics. His vision for the organisation was to obtain and quantify comprehensive information on an annual basis on the socio-economic, demographic, sectoral and other profiles of the country, both at the national and state levels.
Thus NSSO developed a methodology based on multi-stage, multi-subject and multi-purpose cross-sectional surveys, conducted annually for some and five-yearly for others (such as EUS and consumer expenditure) and covers a wide range of subjects. The NSS data are collected and disseminated by the National Sample Survey Organization (NSSO) of India, which is a part of the Ministry of Statistics and Programme Implementation (MoSPI).
Researchers and organisations ranging from the ILO, World Bank and the erstwhile Planning Commission of India to academic and private institutions have been using these data extensively. In addition to the data published in its reports, the NSSO also provides unit-level (household and individual) data, which are a vast source of information on several aspects of individual households at detailed disaggregated levels, which are not available in published reports.
The first nation-wide comprehensive survey on employment, unemployment and labour force was conducted in the 9th round (May-November 1955) of the NSSO. These surveys were regularly carried out up to the 17th round in the rural sector and 22nd round in the urban sector with different concepts and approaches tried in different rounds making the comparability of data difficult and limiting the validity of long-term trends. Based on the recommendations of the Committee of Experts on Unemployment Estimates set up by the Planning Commission in 1969 (Dantwala Committee), the NSSO standardised the concepts and definitions of labour force, employment and unemployment, which were then adopted in quinquennial surveys (large sample sizes of over 100,000 households) on employment and unemployment and were conducted since 1972-73 (27th round) in rural as well as urban areas.
This first quinquennial survey made a marked departure from the earlier more simplistic employment surveys conducted by the Census and NSSO in terms of procedure and content. The concepts and procedure followed in this survey were based on the recommendations of standardisation drawn from the contemporary global discourse on strengthening the databases of the macroeconomic indicators of national economies. India adopted some of the most important resolutions passed by the International Conference of Labour Statisticians (ICLS), the world’s recognised standard-setting body in the area of labour statistics since the 1960s and thereafter adopted the international standards of classifications and categories in the realm of labour statistics, thus setting the stage for building a rigorous, comprehensive and internationally comparable employment-unemployment data system.
The quinquennial surveys, based on such standardised methods served as an extensive source of information pertaining to not only the levels of labour force participation rates, work participation rates and unemployment rates, but have also been a source of rich information on the levels of under-employment or disguised unemployment, different categories of workers, their living conditions and status of employment, level of wages and earnings, occupational structures, status of informal labour and several other indicators of the quality of the workers and the non-workers. These surveys remained a substantial source of employment data disaggregated across religious and social groups as well as by the socio-demographic trends, for the past 42 years. In all, there have been a total of ten rounds of these surveys which do provide long-term insights to the changes in the Indian labour markets. There exists a continuity as well as comparability across these employment-unemployment data at several broad and disaggregated levels. The data are also internationally comparable.
Albeit, the NSSO survey architecture and the sampling designs are complex and technical and are not easily accessible to the general public. However, the purpose with which it was designed in order to be as representative as possible, primarily catered to experts and thus served its ends adequately. Some of the most complex surveys such as the EUS would also be the ones supplying the minutest details of labour market situation in India. Although the sampling methodology adopted by the NSSO is described in detail for the users, it needs to be recognised that given its complexities, replicating such a methodology has not been feasible for individual researchers or research organisations. Further, the expertise required for reading and interpreting as well as extracting raw unit level NSSO figures have also remained limited over a long period of time. However, in the last few years, several initiatives in the form of increasing online resources, conducting workshops and training for students, researchers, practitioners and others to facilitate the interpretation of this data increased its accessibility and usefulness substantially, even for the non-technical audience.
The validity of the NSSO data has also been one of the best in the world. Not only due to the fact that the NSSO trained its neighbouring countries in building a robust data system, but also because of the several stages of validation which the sampling methods ensured. The fieldwork for the national sample is carried out by teams of experienced and highly trained investigators from the Field Operations Division (FOD) of the NSSO.
Second, each state conducts a parallel survey, supervised by the Directorate of Economics and Statistics. This survey uses a larger sample size and the sample overlaps with that of the nationally commissioned survey. Third, the FOD scrutinises each questionnaire for logical correctness and the questionnaires found to be invalid are surveyed again. These steps ensure the validity of the survey and make it the most valuable source of information on India’s economic and social parameters that are of critical importance to policymakers and researchers.
However, despite such rigorous methodologies, the NSSO data have not been free from critical appraisals. Apart from periodicity and timeliness, the NSSO has been criticised on several grounds such as the estimated size of the population. In NSSO-EUS, it has always been on the lower side and hence even for the quinquennial rounds, the data have been approached with caution while using for an in-depth analysis at the lowest level of industries or occupations. The data have not been completely free from the errors of identification, enumeration and tabulation. There have been occasions where discrepancies were found among the codes recorded against sample observations/individuals for status of employment/industrial classifications/occupational classifications.
Such critical examinations of the NSSO data have in fact led to major improvisations and corrections in revising the sample methodology which in turn made the database more robust, reliable and representative. The methods also underwent changes in altering the questions of the survey schedules so as to adopt, accommodate and be able to supply detailed information pertaining to the Indian labour markets in the context of the rapid changes that took place across the globe over the period. Properly extracted and tabulated, NSSO-EUS helped immensely in the analysis of trends and also in the evaluation of the impact of government interventions, thus playing an important role in influencing the overall macro policy framework of the country.
The NSSO-EUS data has been a dynamic process evolving steadily over time to take into account the factors important to analyse the structure of the Indian labour market and has kept itself updated with the global data systems. India has been a proactive member of several ILO committees and task forces that have looked into redefining several categories of labour market indicators and is working towards integrating the labour statistics with the other domains of the data systems. The NSSO under the MoSPI, with its years of experience and expertise have been leading such initiatives effectively.
New labour force surveys
Currently, in order to increase global compatibility and comparability, the MoSPI has also undertaken the initiative to provide high-frequency labour-force statistics from 2017-18 under a new series, the Periodic Labour Force Surveys (PLFS), data of which are expected to be released towards the end of 2018.
The current PLFS by MoSPI would be providing annual employment-unemployment figures for rural areas and similar quarterly figures for urban areas. Undoubtedly it would provide high-frequency data on the broad employment-related indicators that would be much useful for the immediate concerns related to ‘joblessness’ of the economy.
However, the methodology adopted for the PLFS, so far made public, does also indicate departures from those of the earlier EUS rounds and therefore would not be fully comparable with the earlier rounds of NSSO-EUS. This is to say that while we may get high-frequency information on the broad indicators of the labour markets beginning 2017-18, they may not provide the necessary comparability over the past. So a major setback would be to lose out on the long-term oversights of the labour statistics in India. While the headline figures would be available, the dynamics of the labour markets which could be studied historically from the rich disaggregated data of the several EUS rounds will no longer be possible.
Thus the overall scrapping of the EUS surveys and discontinuing an important source of labour statistics based on grounds of certain limitations of sampling adequacy, periodicity, timeliness and relevance – i.e. limitations which were easier to correct and revise rather than discard in its entirety – is akin to chopping off ones head when one gets a headache, rather than taking a medicine.
In the process, what it has essentially created is a situation where there would be no data system to compare the present with the past. Presently, there exists no comparable information on the status of the labour force, workforce and unemployment rates of the country since 2011-12.
With a new series of labour force data, the current problem of data-lessness would be addressed, but the lack of comparability with the previous years would still continue. Such a situation in a way justifies the above-mentioned statement made by the PM. The irony, however, is that it has been created by none other than the government itself. While it becomes most important to publish the results of the PLFS at the earliest, which may cater to the immediate requirements, it is absolutely necessary that the government consider producing long-term comparable data in order to inform policies for sustained employment generation in an economy like India, which has been witnessing ‘jobless growth’ for some time now.
Sona Mitra is a Senior Consultant with the Centre for Budget and Governance Accountability and can be reached at firstname.lastname@example.org. Views expressed are personal.