The Centre is learnt to have held discussions at the highest level to look into the delay in frequency of the release of government employment-unemployment data, with one of the main concerns pertaining to the time lag in the release of the official Periodic Labour Force Survey (PLFS) as against privately-conducted surveys such as the one brought out by Centre for Monitoring Indian Economy (CMIE).
In recent meetings held at the Cabinet Secretariat attended by statistical department and Union Labour Ministry representatives, discussions converged on the need to improve the frequency of the labour market data, especially in the aftermath of the Covid-19 pandemic, a senior government official said.
“There was a meeting held in the Cabinet Secretariat. PLFS is taken out by the statistical department and they have to follow international norms. There is also more scrutiny: whether random sampling is happening or not, urban-rural population parameters, agriculture-industry proportion, and so there are limitations. We have asked to increase the frequency (of the official data release),” this official told The Sunday Express.
The person further said that the discussions have been held in the context that any delay in the release of the government data on employment-unemployment tends to raise questions about its credibility, even as PLFS data is more robust than the privately-conducted surveys.
“PLFS is a much more robust survey. In the private surveys, there is a greater skew towards urban areas and the methodology is also different, you have to keep changing the set otherwise biases creep in. But they haven’t done it. There are gaps in these surveys. Why should we find fault … if our data doesn’t go public, they think we are hiding or there’s something wrong. But that’s not true. Look at EPFO payroll data: whether increase or decrease, it comes out at 20th of every month. Data release should not be delayed,” the official said, asserting that these points were flagged at the Cabinet Secretariat meeting.
The National Statistical Office (NSO) under the Ministry of Statistics of Programme and Implementation (MoSPI) conducts the PLFS for rural and urban areas. Labour data for rural areas is published annually, while that for urban areas is released every quarter. There is a gap of at least a year in the release of the annual PLFS report. In recent times, PLFS for 2020-21 (July-June) was released in June 2022, while the PLFS for 2019-20 was released in July 2021 and that for 2018-19 was released in June 2020. On the other hand, Mumbai-based CMIE comes out with more frequent data for unemployment on a daily, weekly or even monthly basis.
In its latest report presented to Parliament in August, the Standing Committee on Finance too flagged the time lag in the release of the PLFS, citing that the quarterly PLFS for July-September 2020 was released in August 2021, for October-December 2020 in September 2021 and for January-March 2021 in November 2021.
Software, database issues
The PLFS data, considered to be among the most vital social indicators that policymakers depend on, faced delays in release in recent times due to glitches in a software application developed by the World Bank as well as database-related issues. These delays got even longer after Covid.
The Committee pointed out that the employment data provided by PLFS, being one of the “most important social-indicators, vital to policymakers”, should be produced periodically without time lag. The MoSPI, in its reply, said the PLFS has increased the frequency to annual data as against once in five years earlier. It cited the glitches in the software application developed by the World Bank (WB) and database-related issues for the delay. PLFS was initially conducted using the Computer Assisted Personal Interview (CAPI) application developed by the World Bank with high expectations to improve the time lag to a great extent. While a number of problems encountered in the WB application were sorted out, a database-related major problem could not be. Multiple revised versions of the CAPI application provided by the WB for addressing this problem failed to clear the security audit, which is a mandatory requirement for any software application for use by the government, the report said quoting the MoSPI’s response.
The ministry then discontinued the WB’s CAPI application and in late 2018 had to resort to the conventional paper-assisted personal interview mode for PLFS. The manual mode of data collection and data processing resulted in considerable lag of PLFS reports. “The deterioration in the working conditions in the field and otherwise due to the Covid-19 pandemic only aggravated the issue,” the report said. It, however, added that the NSO has released recent PLFS bulletins such as the one for January-March 2022 with a time lag of only two-and-half months which was earlier 8-10 months.
The discussions within the government now have also focused on the need for more frequent data for rural areas as well. “We are insisting that the gestation period should be reduced for the PLFS survey. We had data for 2019-20, now we have for 2020-21. We should improve the frequency. Rural data should come quarterly like urban data but it is still coming annually. They (CMIE) take out weekly, monthly, daily,” the official said.
Experts said that the CMIE being a panel survey has limitations in providing detailed analysis beyond the headline numbers. “Theirs is a panel survey. CMIE survey asks very limited questions so it doesn’t have the richness and analytical richness that the PLFS has. PLFS is much more complex. If you are looking at finer characteristics rather than broad numbers like unemployment-employment rate, CMIE can’t give it, but PLFS can,” Pronab Sen, former Chief Statistician of India, said.
For the gap in rural-urban data, Sen said that the decision that was taken earlier was based on the view that the rural job scenario doesn’t change very quickly whereas the urban does. “Urban is much more dynamic while rural is more stable. But during a crisis like we had, the rural situation also changed very quickly. Under normal circumstances, they are working according to a budget and quarterly data collection is more demanding. They are doing it on the phone now, which in rural India is a problem unless you simplify the questionnaire a lot. It is easier to do a high-frequency survey in a panel because the respondents are already familiar with the questions. Whereas if you are doing a cross-sectional survey, you have to educate people about what you are asking. So it is more time-consuming. The delay in the release earlier for PLFS was also because of some glitches in the software. They have straightened it and results are now coming on time,” he said.
Newsletter | Click to get the day’s best explainers in your inbox
As per the CMIE, it estimates the size of the labour force in India and the unemployment rate in India by directly interviewing a large sample of randomly selected households, to find the employment/unemployment status of all members of over 15 years of age. The full sample of 1,78,677 households is surveyed over a four-month period, with the monthly sample being well-distributed over rural and urban regions to enable a weekly estimation of unemployment at the all-India level.
Under PLFS, a total number of 12,800 first sampling units are covered for the survey at the all-India level during July 2020-June 2021. A rotational panel sampling design was used in urban areas, where each selected household is visited four times in urban areas in the beginning and thrice periodically later, while no revisit happened for rural areas with 25 per cent of sampling units covered in every quarter.