Making sense of EPFO data

Payroll statistics are India's only officially released employment source but they suffer from several gaps

Historically, India’s employment estimates have been generated using household and establishment surveys. Following the clamour for real time high frequency employment data, the task force set up to review India’s employment data last year recommended an additional data source, namely, administrative datasets. Since April 2018, the government has started reporting the progress made in payroll employment using membership data from three schemes — Employees’ Provident Fund Organisation (EPFO), Employees’ State Insurance Scheme (ESIC) and National Pension Scheme (NPS) on a monthly basis. The EPFO payroll numbers are most widely cited as they address problems of duplicate and inoperative accounts by Aadhaar seeding the data and examining only non-zero contributory members. Based on this data, it has been argued that 4.47 million formal jobs were added in the economy between September 2017 and May 2018. Several economists have doubted such claims of a surge in job growth. Reams of newsprint have been spent discussing whether a new enrolment on the EPFO database does indeed reflect creation of a new formal job. However, even if we put aside this debate and believe that the data reflects growth in new formal jobs, several doubts and concerns persist.

So far, the government has released the EPFO payroll series four times and each time it has revised its previous estimates. These results are reported in the table. It is worth noting that barring two instances, in each subsequent release the net monthly payroll enrolment numbers have been revised downwards. For instance, enrolment numbers for the month of September 2017 when released originally in the first series stood at 5,96,483. By the fourth, and most recent release, the net enrolment number for this month had been revised downwards to 529,432. Similarly, the net addition to payroll from September 2017 to February 2018 was initially reported as 3.27 million. However, in the recent release the enrolment numbers for the same time period were revised downwards to 2.68 million. The reasons for the revisions are not well understood and warrant greater clarity given the claims of significant job creation made on a monthly basis.  

A footnote in the first EPFO data release mentioned that the estimates include temporary employees whose contributions may not be continuous throughout the year. This is a matter of concern as it is likely to bring in volatility into the numbers reported. We know that with the EPFO pulling up firms for failing to ensure that workers hired on a contract basis received their provident fund, increasingly contract workers are getting enroled on the EPFO database. Given the inclusion of workers whose EPFO contributions are not continuous, we can expect to see fluctuations in monthly enrolment numbers. Additionally, it has been noted that there could be irregularity or delay in remitting contributions by establishments into the EPF accounts. In fact, in a recent note dated June 19, 2018, the EPFO has cited an increase in the number of defaulting establishments over the past few months[1]. The number of defaulting establishments for the wage month of May 2018 (payable in June 2018) was as high as 54,885. This raises a question mark on the reliability of monthly enrolment numbers and whether a monthly time period is indeed appropriate for analysing payroll data.

Another striking observation vis-à-vis the EPFO database is that whilst we are seeing net additions of new members with non-zero contributions in the payroll data releases, the total number of contributory members as reported by the live dashboard of the Ministry of Labour and Employment are, in fact, declining. For instance, the dashboard reports that the total number of contributory members in March 2018 stood at 47 million. This then declined to 46 million in April 2018 and 45 million in May 2018. However, for the same months the payroll data in the table shows positive additions. The contradiction between the stock and flow numbers is indeed baffling and needs to be understood.

Furthermore, an analysis of the EPFO database needs to be done in the backdrop of the existence of the Pradhan Mantri Rojgar Protsahan Yojana (PMRPY). By paying the full employer’s contribution towards Employees’ Provident Fund and Employees’ Pension Scheme (EPS) for new employees for the first three years of their employment, the government has attempted to incentivise formal employment generation through this programme. The terminal date for registration of a beneficiary under the scheme is March 31, 2019. The programme is likely to have driven up enrolment in the EPFO database. Data from the Ministry of Labour and Employment’s website shows that between April 23 and May 29, 720,000 members enroled under the PMRPY scheme. The payroll data also shows an addition of 743,608 new members in May 2018 suggesting that the increase in payroll data may well have stemmed entirely from PMRPY. This raises the question of whether the EPFO data is presenting an accurate picture of the extent of formalisation or is simply reflective of a temporary boost to EPF enrolment under PMRPY.

Following the discontinuation of the employment-unemployment surveys undertaken by the National Sample Survey Organisation and Labour Bureau, payroll data is India’s only officially released employment data. Whilst we await the results of the Periodic Labour Force Survey, payroll data can potentially provide a sense of what is happening in the labour market in the interim period. However, as evident from the above discussion, there are several issues and gaps in this data. Much needs to be clarified before we can make any conjectures regarding job creation using this database.

Kapoor is senior fellow, Indian Council for Research on International Economic Relations (Icrier); Ghai is a research assistant at Icrier

[1] https://www.epfindia.gov.in/site_docs/PDFs/Circulars/Y2018-2019/defaulting_establishments_caiu.pdf