ISSN (Online) - 2349-8846
-A A +A

Employing the Right Measures

Better employment data will have to rely on large and frequent surveys, not misleading proxies.


Prime Minister Narendra Modi, in a recent televised interview, misleadingly claimed that an “independent agency” found that 70 lakh Employees’ Provident Fund Organisation (EPFO) accounts were opened for persons aged 18–25 years. He said this without even bothering to indicate the period during which this additional employment was presumably generated. “Does not this show new employment?” he asked. It was clear that this claim came out of a study submitted to the Niti Aayog, which provides employment projections for this financial year, that is, till 31 March 2018. Most available evidence on employment creation (and associated economic indicators) contradicts the possibility of higher employment growth during 2017–18 over the previous year.

The study “Towards a Payroll Reporting in India” (a summary is available online), submitted to the Niti Aayog, is ostensibly a product of the government think tank’s larger push to change how employment is tracked in India. Its rationale for change, shared by some labour economists as well, is that the measurement of employment in India has certain weaknesses: small samples, infrequent surveys, and lagged data release. There has been a long-standing need to have larger surveys as well as quicker, “real-time” indicators. The “Report of the Task Force on Improving Employment Data” (2017) mentions, among other possible methods, the use of “administrative datasets” such as the EPFO, the Employees’ State Insurance Corporation (ESIC), the National Pension Scheme (NPS), and other similar schemes to gather information on the labour market. It uses these data sets to estimate employment generation and concludes that 5.9 lakh persons will be added to the “payroll” every month during the present financial year. It is clear that the study conflates new enrolment into these social security schemes with new employment. The set of workers covered by the administrative data sets used in the study in any case comprise a small proportion of all workers. Therefore, it is meaningless to make a claim of new employment without accounting for employment figures for the rest of the labour market. At best, this data represents the rate of ­enrolment into some social security schemes in the workforce—a kind of formalisation—and not employment creation.

The taskforce report also speaks of some disadvantages of administrative data sets. The specificities of these databases make them vulnerable to being misleading proxies. Biases creep into these data sets due to government policy changes. The willingness and intensity of enforcement may vary from year to year. Government pressure to enrol workers in these schemes may lead to inflated estimates if these data sets are used to measure employment. The EPFO has been particularly proactive in recent times and has directed various government bodies to enrol their contractual workers in its provident fund and pension schemes (this is apart from the Employees’ Enrolment Campaign of early 2017). When a non-permanent staff becomes permanent, it too adds to the “payroll” without creating new employment. The report also warns that there will be a high chance of duplication between administrative data sets (like EPFO, ESIC, NPS), as they lack a common identifier to de-duplicate entries. Even if one accounts for de-duplications, the data might not be error-free.

Both 2016–17 and 2017–18 were unusual years for the economy. With two far-reaching actions—demonetisation and the goods and services tax (GST)—the state clamped down on informal economic activities. This left many of those in the informal sector with no option but to accept more state regulations, often on terms that proved detrimental to their enterprises. This “forced formalisation” would probably reflect in higher social security scheme ­enrolment, but will not necessarily mean more employment overall. Enrolment into these schemes, therefore, serves as a poor proxy for survey-based methods of measuring employment.

Curiously, the authors of the study that was submitted to the Niti Aayog also talk about how their method will help build something similar to monthly non-farm payroll reports by the Bureau of Labour Statistics (BLS) in the United States (US). The BLS, however, relies on large-scale quick enterprise surveys supplemented by household surveys to compile information for its non-farm payroll reports. The sample is large, covering more than a third of non-farm employment. These surveys are somewhat similar to India’s Labour Bureau enterprise surveys, but are larger, conducted more regularly, and with systems that allow quick generation of information. Clearly, administrative databases designed for specific purposes are no substitute for household and enterprise surveys. Measuring the extent of employment growth is only one aspect of studying the labour market. It also involves studying the nature and conditions of work. Importantly, even as the US is a largely formal economy, the main “real time” labour market indicator is survey-based.

In this sense, the experiment with administrative databases, though academically interesting, may be useful for measuring formality in sections of the organised workforce, but is no substitute for surveys. The focus should be to build on the far richer history of the National Sample Survey Office and Labour Bureau employment surveys. Over the years, they have designed methods that adapted survey techniques to Indian conditions to better understand difficult phenomena. These, for instance, have attempted to measure informality of work in the formal sector, various kinds of employment in the informal sector, the seasonality of employment, and even underemployment.

What is most disappointing about this study and its dissemination is that neither the full paper nor the administrative data sets are in the public domain. With no peer review of these untested methods and new data, it is troubling that its findings are popularised for political gain.

Updated On : 29th Jan, 2018


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top