Why Payroll Data Dissemination in the Country Needs to Improve

The time is ripe to draw up a better depiction of employment numbers by using all available data sources

Why Payroll Data Dissemination in the Country Needs to Improve
OLM Desk 28 May 2021

The pandemic has given rise to multiple deep, defining scares in labour markets all over the world. As per the latest data from World Bank, there have been persistent reports based purely on employment surveys of permanent job losses and a significant spike in unemployment, particularly among the youth, and a compression of India’s labour force from 495 million in 2019 to 472million in 2020.

Some of these concerns, and perhaps specifically the crunch in employee expenses, are genuine concerns. Soumya Kanti Ghosh, Group Chief Economic Adviser, State Bank of India, states that employee expenses data for 284 listed companies for FY21 reveal the largest decline was in the smallest firms, indicating that employees of smaller firms were impacted most due the pandemic. It is also true that one of the more pressing concerns of our economy at the present moment id the creation of jobs. Diamond, Mortensen and Pissarides (2010 Nobel winners) point out in one of their classic papers that unemployment remains high when jobs are available (particularly relevant in the context of India’s shift in the unemployment age).

What are the observations from India’s payroll data for FY21?

Firstly, as per EPFO and NPS, India created 100.4 lakhs payroll, nearly unchanged from 102.3 lakhs in FY20, indicating Indian labour market did not do that badly in FY21, despite the obstacles posed by the pandemic. However, it should be mentioned these are low quality jobs.

Second, when we breakup 95.4 lakhs jobs created by EPFO payroll, 41.2 lakhs were through second jobs, 44 lakhs through first jobs and 9.3 lakhs were through formalization. The important point to note is that there was a decline in first time jobs in FY21 by 16.9 lakhs, though the number of second time jobs/members who re-joined the payroll rose by 17.9 lakhs. This clearly indicates that people were coming back to the labour market in later part of FY21. The rate of formalization however declined by 1.2 lakhs, reflecting the disruptions in the MSME space.

Third, NPS data indicates that there is a decline of 1.74 lakh new subscribers in 2020-21, of which State Government payrolls lost 1.06 lakh, followed by non-Government ones of 36,416 and 31,420 in Central Government.

Fourth, the ratio of women enrolment to total enrolment in EPFO data has remained at 23% in FY20 and has not changed significantly in FY21.

Fifthly, we must use payroll data and not employment surveys simultaneously to have a more meaningful debate of India’s unemployment, lest it could be significantly biased and unidirectional as is now.

Additionally, the data suffers from the intrinsic deficiency of reactivity (respondents reacting to the initial questioning). Another issue is the issue of attrition (i.e., respondents dropping out of the study).

We are not sure whether the surveys done in India live up to all such challenges. A case in point is that the University of Michigan, in their Consumer Surveys, always asks consumers about their anticipation of unemployment rate changes, and that is subsequently validated. We are not sure whether the employment survey in India which is modelled along the lines of the University of Michigan Consumer Survey, addresses such an issue!

Lastly, there are also problems with EPFO data. The classification of the EPFO job data leaves a lot be desired. For example, Expert Services that constitute the largest segment of payroll creation do not tell us anything at granular level. Another problem is regarding retiring employees. Given that retirees are also netted out, this may imply a downward bias to net EPFO numbers as retirees mean a new vacancy and hence a new hire. This can be revised in later disclosures, but this is a suggestion that EPFO might well consider, apart from giving out the payroll creation across all industry groups as the US does. EPFO should start releasing non-farm productivity (as in the US) estimates at least for those sectors for which we have output data from CSO’s GVA database. This will fill a huge lacuna in productivity estimates in India.

Clearly, the employment debate in India is perhaps haplessly one sided and only talks about problems and not the solutions! The time has now come to have a better depiction of employment numbers by using all available data sources, including payroll.