The recent State of Working India report on jobs uses data from a survey that is marred by several limitations.
India has been lagging behind in the way it estimates employment, and the recent report ‘State of Working India/SWI’ is no exception.
The irony is that the measurement of job creation, employment and unemployment in India is usually done through statistical estimates from household surveys, enterprise/establishment surveys, administrative data and data from government schemes.
And, the Quarterly Employment Survey (QES) that was initiated as a short-term response to monitor disruptions after the global financial crisis in 2008 continues till date. The QES, in fact, is marred by several limitations as it elicits responses from only eight sectors (against a vast universe of 190 sectors). The results are released with a nine-month lag, all information is provided on a voluntary basis and not verified and does not capture employment data on new units.
Unfortunately, these reports have been quoted and used extensively by policy makers, including labour economists, to decipher India’s employment problem. Even the SWI report uses the QES as its basis.
To address the criticism against QES, the government also appointed a Task Force on Improving Employment Data under NITI Aayog a year ago. In January this year, we had also pointed out, in a joint study by the Indian Institute of Management, Bangalore and the State Bank of India, a new benchmark for the total organised payroll in India.
We suggested looking at the administrative data of the Employees’ Provident Fund Organisation (EPFO), the National Pension Scheme (NPS) and the Employees’ State Insurance Corporation (ESIC). Consequent to that, the government started an initiative to report the progress made in formal employment using measurable data from administrative records.
Recent data from the EPFO indicates that the net payroll generated during September 2017-July 2018 was around 61.8 lakh. In fact, during July 2018, EPFO payroll registrations jumped by 9.5 lakh, the highest in any month since inception. The EPFO has also published industry-wise break-up and state-wise break-up of payroll data. The top 10 industries in the country constitute around 85 per cent share in the total payroll of the EPFO. As per the NPS data, around 6.7 lakh new payrolls were generated during the period of September 2017- August 2018, with an average generation of 60,000 per month.
The purpose of the release of these datasets is to overcome the limitation that is primarily contained in survey methods. Although the results of the EPFO and the NPS have opened up a new system of real-time payroll data reporting, the data series is yet to stabilise. Furthermore, the ESIC data is still undergoing structural changes because it has not been mandated to link such data with Aadhaar.
In its recent press releases, the EPFO had released data on number of members who ceased subscribing during the month along with the number of new EPF subscribers. The difference between the two indicates the net payroll generated. This data was mistakenly reinterpreted in public domain as job loss. In the 20 September 2018 release, the EPFO clarified the data and also provided data on the number of exited members who have re-joined and re-subscribed during the concerned month.
The 61.8 lakh net payroll can thus be divided into two parts: existing payroll and new payroll.
We define existing payroll as the sum of re-joined and re-subscribed members and the extent of formalisation within the economy in terms of payrolls being counted for existing organisations that migrated from less than 20 jobs. This comes out to be 25.1 lakh during September 2017-July 2018 period (formalisation at 10.25 lakh and re-joining at 14.84 lakh).
So, the net of existing payroll, 36.7 lakh, is the count for new payroll generated during September 2017 to July 2018 (total 11 months). Ideally, the number of people who have taken up new jobs after quitting existing one (at 14.84 lakh) should also be counted as new payroll. If this is the case, the annualised new payroll is coming at 56.2 lakh.
We reiterate here that employment surveys, unless properly and scientifically designed, will give wrong results. For example, as per surveys by the CMIE, it was claimed that there were job losses during the period of January-April 2017 (when compared with concomitant figures from September-December 2016). However, as per the average unemployment rate published by the CMIE, the average unemployment rate declined to 4.9 per cent in January-April from 7.05 per cent in September-December period.
It appears that even the surveys done by the CMIE may reveal completely divergent trends. The EPFO data may have still some limitations, but we must allow the database to evolve over time in terms of trials and tribulations rather than merely dismiss the data for pecuniary gains.
The authors are Group Chief Economic Advisor, SBI and Professor, IIM Bangalore. Views are personal.