India is facing a jobs data crisis: Here's how it can be fixed

Take the US for example. Payroll data by US Bureau of Labor Statistics measures new job creation accurately by releasing data each month to present the health of the economy, something that India lacks.

Conflicting data sets

Varied sets of data have led to confusion in India. A recent UBS Securities report said that there are two sets of data with conflicting outputs.

“A study based on EPFO (employee provident fund) data suggests a likely 7 million formal jobs created in FY18. This number is a positive surprise to us and markets. The BSE-CMIE index, which is household survey-based, suggests that not many jobs have been created in this same period and unemployment levels (formal + informal) may have started to go back up,” the report said.

In April 2018, the EPFO, Employees’ State Insurance Corporation (ESIC) and the Pension Fund Regulatory and Development Authority (PFRDA) declared that they had released payroll data for the first time. Since then, data is being released each month.

EPFO data showed that 3.11 million new additions were made in payroll between September 2017 and February 2018 across all age groups. However, EPFO said that the data for recent months was provisional due to continuous updation of employee records, hinting that the data could be a conservative estimate.

From PFRDA, the New Pension Scheme (NPS) data indicated generation of 4,20,000 new payroll during the given period, that too only from Tier-I account. NPS currently manages the corpus of around 5 million employees in state and central government offices.

From the above two organisations itself, 3.53 million new payrolls were generated during the six-month period.

However, questions have been raised about whether this number is the actual number of jobs created. The UBS report quoted earlier said that members of the Employees' Provident Fund are those who have a PF balance in their PF account.

“The investment corpus per member has been growing less than inflation. This seems to suggest that the quality of jobs may not necessarily be improving,” it said.

BSE-CMIE index

The BSE-CMIE index suggests that unemployment levels may have started to climb post July 2017, following a previous decline. The survey suggests a declining trend in the labour participation rate, and paints a worrisome picture.

In 2016, BSE and CMIE joined hands to launch regular data on unemployment and consumer sentiments. The effort was to produce a 30-day moving average measure of unemployment rate in India, including daily and monthly unemployment rate at the end of each month.

The data does not measure how many jobs were created or added every month, but is a representation of general joblessness in the country in a given period.

The panel of households contains over 5,00,000 adults from 1,58,000 households, spread around 325 cities and 2,900 villages across India. Panel of households is utilised to compute the index. The size of the labour force and the unemployment rate in India is measured by taking all adult members of the panel of randomly selected households. A comprehensive survey is conducted over three rounds, each of which spans four months.

The survey provides unemployment/employment status of 1,25,000 adults per month. The estimates are computed using the sample of responses. Data is collected from the field on hand-held GPS enabled phone devices. It is uploaded from immediately after completion of survey and cross-checked centrally to ensure plausibility of the entries.

Lack of single source of data

There is no one single annual survey that captures the exact number of formal and informal jobs created in a particular year. Online surveys by job-sites like Monster and Naukri released monthly give a partial idea of how many new jobs were posted on the internet. However, this does not take offline jobs into account.

These surveys give a rough estimate of how many absolute number of job roles were added on their websites.

For instance, the Naukri Job Speak Index is calculated based on job listings added month on month to the site. Data is sourced from Naukri.com and it reflects job listings and hiring trends on the site. Jobs are categorised by location, functional area, industry, and experience.

US model of data collection

Similar to the Union Budget season in India, the US jobs data crunching process is considered a top secret. The location where the data is processed is out of bounds for certain days before data is released so that it is not tampered with or leaked.

According to the Bureau of Labor Statistics, about 60,000 households are selected and informed by the US Labor Bureau on a rotational basis. Each respondent is asked a fixed set of questions including whether they have a job, looking for a job and the type of job they have.

The basic concept of identifying employed and unemployed is quite simple. People with jobs are employed, while those who are jobless, looking for a job and are available for work are unemployed.

Those who are neither employed nor unemployed are not in the labour force.

Here, the total unemployment rate is the number of unemployed people as a percentage of the total labour force in the country during a given period.

Fewer jobs in the manufacturing sector

A large job creator in the country is the manufacturing sector. However, with the sector trailing, job additions have been impacted. A CRISIL report on the four years of the Modi government, titled ‘4 years through 6 lenses’, said that GDP in construction has trailed overall growth, while manufacturing has not shown any notable improvement.

The report pointed out that while creating employment for a fast-growing workforce is a key challenge, there is no robust way to assess the progress on this front. This, it said, was due to paucity of reliable statistics and often-conflicting interpretations.

NSSO to be scrapped?

One data source that has been referred to for welfare schemes and official purposes, is the National Sample Survey Office (NSSO) employment and unemployment survey that is released every five years. Similar to the US jobs data, this is a sample survey of select households across India to get detailed information about the number of people employed, duration of employment and type of employment.

Similarly, India's Labour Bureau publishes quarterly surveys, which covers just about 30 million people employed in the country. The estimated labour in the country (farm and non-farm labour) is about 480 million.

Though initially it was anticipated that NSSO could be given the mandate to conduct surveys on a regular basis, a PM Modi-appointed taskforce in 2017 recommended that NSSO surveys should be scrapped.

The panel suggested that there should be a more periodic survey to account the exact number of jobs being created and jobs lost. It suggested that there be a quarterly survey commissioned officially by the government. However, no decision has been taken on this until now.

Will India get a reliable, regular jobs data soon?

The Modi government is in its last year of its term as general elections are set to be held in 2019. Even as government appointed panels and officials have suggested that new data will be released, giving details of both farm and non-farm labour, the EPFO/NPS data seems to be a good start.

If this data can be extrapolated to the entire economy and data on new jobs based on company registrations with the Ministry of Corporate Affairs could be taken into account, that would be a better reflection of new jobs.

Rather than different bodies bringing out complete sets of conflicting data, the need of the hour is for research organisations to come together and bring out statistics that are reliable and capture the demographic differences of the country.