The COVID-19-induced lockdown is an unprecedented event. It has rendered millions of people unemployed for an extraordinarily long period in the interest of saving lives from an uncontrollable virus. The lockdown began slowly and hesitatingly in mid-March as many state governments announced local lockdowns till the end of March. But, the nation-wide lockdown that began on 25 March was unparalleled. It was severe, prolonged, confusing and unpredictable in multiple ways.
The impact of this lockdown on the economy was devastating. The index of 8-core industries fell by 38.6% in April. Three months later in June 2020, the index was still 15% lower than it was a year ago. Professional forecasters’ estimates compiled by Reserve Bank of India (RBI) during late July–early August 2020 showed that real GDP was expected to shrink 22.8% in the quarter ended June 2020 compared to the corresponding quarter a year ago.
What could the impact of this draconian lockdown be on labour markets and employment? India has had a unique distinction of maintaining high real GDP growth with no growth in employment for a very long time. Could its economy also shrink without impacting employment? While India has improved the frequency with which it releases most macroeconomic indicators, its official machinery is still incapable of providing corresponding employment and unemployment estimates. And, the nature of the lockdown is such that historical data from past labour surveys could be a poor guide to guess its impact.
CMIE’s Consumer Pyramids Household Survey (CPHS) is best suited to guide us on the impact of the lockdown on employment and labour markets because of its currency. CPHS began in January 2014 and has continued uninterrupted since then. Its execution did not stop during the lockdown. This is a longitudinal panel household survey with a total sample size of 174,405 households. The entire sample is surveyed during a Wave that lasts four months. Three Waves of the survey are completed every year.1
The CPHS survey execution is designed to enable fast-frequency estimates. This is possible from the subsample of 43,500 households surveyed per month and 11,500 households surveyed per week. The CHS execution machinery ensures that every major stratum is surveyed uniformly across the four months and the 16 weeks of the survey. This is what enables the generation of reliable fast-frequency measures from CPHS.
CPHS was a CAPI survey conducted using GPS-enabled handheld devices before the lockdown. During the lockdown, this survey execution system switched to a telephonic survey administered on the same panel sample and using the same questionnaire that was deployed before the lockdown. The telephonic survey was rolled out on 28 March.
The sample size in the telephonic survey was curtailed to about one-third of the original size. This was necessary because the execution was escalated to supervisors to ensure respondent participation and the quality of response in the face of reduced controls of peace times. The average responses accepted per day before the lockdown were 1285 in February 2020. This fell to 446 per day in April 2020. This count increased from the second half of May when the lockdown started to ease and face-to-face interviews could resume in some regions.
While the sample size was curtailed, the sample distribution across states and across rural and urban regions was kept the same as it was before the lockdown. A post hoc analysis of the survey execution during the lockdown reveals that the distribution of the sample across rural and urban, across states, across household income groups and across occupation groups has remained stable over pre- and post-lockdown period (CPHS execution during lockdown of 2020, Mahesh Vyas, 10 August 2020).
Before the lockdown, CPHS published weekly rates of labour participation, employment and unemployment. It continued producing these throughout the lockdown. Also, as before the lockdown, it continued to produce several more detailed estimates of the Indian labour markets at a monthly frequency when the sample size permitted greater disaggregation.
What Happened to Labour During the Lockdown?
CPHS showed that the unemployment rate shot up sharply immediately upon the imposition of a nationwide lockdown. In the week ended 29 March 2020, the unemployment rate catapulted to 23.8%. It had averaged at 7.6% in the preceding two weeks but it was volatile—ranging from 6.7% to 8.4%. This volatility reflects the various partial lockdowns imposed by state governments during March 2020. Nevertheless, compared to these levels, 23.8% was an entirely new order of magnitude. The rate persisted at 23.4% in the week ended 5 April and then inched up to 24% in the week of 12 April. Then, it shot up to 26.2% in the week of 19 April.
There was some confusion of the lockdown being relaxed selectively from 20 April. This was again reflected in volatility in the unemployment rate. The rate which had risen to 26.2% in the week ended 19 April fell to 21.1% in the week ended 26 April and then escalated again very sharply to 27.1% in the week ended 3 May.
The unemployment rate rose during those weeks of April even as the labour participation rate was falling sharply. This fell from 42.6% in the week ended 22 March to 39.2% in the week ended 29 March. It then fell to 36.1%, 35.5% and 35.4% in the following weeks.
CMIE releases these weekly labour market metrics estimates publicly immediately upon the completion of the week. Data for a week ended on a Sunday is released in the wee hours of the very next day. The impact of the lockdown was therefore revealed in near real time by the data generated from CPHS every Monday morning. More detailed data is released in the early hours of the first of every month. The data for April 2020 was therefore released on 1 May 2020.
The unemployment rate in April 2020 turned out to be 23.5%. This was much higher than the 8.8% unemployment rate in March 2020 and even higher than the 7.6% recorded in fiscal 2019–20.
The labour participation rate fell from 41.9% in March 2020 to 35.6% in April 2020. The immediate shock of the lockdown was, therefore, a double whammy of a sharp fall in the labour participation rate and a simultaneous sharp increase in the unemployment rate. The result was a correspondingly steep fall in the employment rate, which is the ratio of employed person to the total population of 15 years of age or more. This fell from 38.2% in March 2020 to 27.2% in April 2020.
In May 2020, the unemployment rate remained elevated at 23.5%. But, the labour participation rate improved a tad to 38.2%. This led to an improvement in the employment rate which now scaled up to 29.2%.
The labour participation rate had bottomed out in the last week of April and was rising steadily in May. It climbed up from 35.6% in the week ended 26 April to 38.8% by the week ended 17 May and then stayed around that level till the end of the month. The return of labour to the labour markets as seen in this steady rise of the labour participation rate was a strong positive sign. Sure enough, by the last week of May, even the unemployment rate came down to 20.2%.
So, while the May 2020 unemployment rate was the same as in April 2020, the steady increase in the labour participation rate and then the fall in unemployment in the last week May were indications that things were improving in May. This was eventually borne out by several other indicators such as a 56% increase in year-on-year comparison of MGNREGS person-days. The 8-core index reversed from its sharp fall in April with a 30% increase in May over April.
The improvement in labour conditions seen during May strengthened in the early weeks of June and then showed some fatigue towards the end of the month. The labour participation rate continued to rise, and the unemployment rate continued to fall till the week ended June 2020. The weekly LPR reached 42%, and the weekly UER reached fell sharply to 8.5% by then. Monthly LPR in June was 40.3% and UER was 11%. Weekly trends suggested that the worst was over. This again ties well with official data that showed a doubling of MGNREGS employment and also of Kharif sowing in June 2020 compared to a year ago.
By July 2020, the UER at 7.4% had repaired with the pre-lockdown levels. But, while the LPR has also recovered, at 40.7%, it was still far below the 42% or more levels before the lockdown.
Who Suffered During the Lockdown?
CMIE releases monthly estimates of absolute values related to the labour markets in India along with their distribution by region, gender, age, occupation, etc. These give us insights about the absolute job losses, their recovery and their distribution.
We compare the employment estimates of a month during the lockdown to the average monthly estimates of employment during 2019–2020. The average monthly estimate of employment in 2019–2020 was 403.7 million. Estimates of employment provided by CMIE from CPHS are systematically lower than those derived usually from Periodic Labour Force Survey or Employment / Unemployment Survey of the NSSO because of differences in definitions. CPHS estimates of employment are based on status as of the day of the survey and not during the seven days preceding the date of the survey as is done in PLFS or EUS. The latter are relaxed, while the CPHS definitions are far more stringent in considering a person employed. Compared to this estimate, employment in April 2020 at 282.2 million was 121.5 million lower. We claim therefore that 121.5 million jobs were lost in April 2020, the first month of the lockdown. The first month of the lockdown took away 30% of all employment.
Of this, 91.2 million jobs lost were those of small traders, hawkers and daily wage labourers. This category of employed persons accounted for 32% of all employment but they accounted for 75% of all jobs lost immediately upon the imposition of the lockdown. These people are largely self-employed and possibly were the most vulnerable to the shutdown.
Around 18.2 million business persons and another 17.7 million salaried persons lost jobs during the month.
The count of farmers went up by 5.7 million. The increase in the count of farmers is interesting. We see this increase further to 7.1 million in May and then to 19 million in June before declining to 15 million in July. This, initially, in April or part of May could have been disguised unemployment but the sharp increase in June reflects the increase in Kharif sowing.
Business persons have slowly recovered their lost jobs or rather, their lost enterprises. By the end of July, the count of persons employed as business persons was almost the average count in 2019–2020.
The lasting impact of the lockdown seems to have been on the salaried employees. The immediate impact of the lockdown was a loss of 17.7 million jobs in April. They lost a little more in May but recovered 3.9 million jobs in June. However, July reported a setback again with a loss of 5 million jobs. As a result, in July 2020, salaried jobs were still 18.9 million less than they were in 2019–2020.
Four months into the lockdown, salaried jobs are the worst hit. As of July 2020, the estimated employment in India was 392.7 million. This is 11 million less than it was in 2019–2020. But, salaried jobs are 18.9 million short. Small traders, hawkers and daily wage earners are down by 6.8 million and business persons by 0.1 million. The count of farmers has bloated by 14.9 million.
Conclusion
Fast-frequency estimates from CPHS were able to provide early signals of the impact of the lockdown on labour. These also showed the severe stress in April and then the gradual improvement since then. While the overall impact of the lockdown has waned, 4 months later, it still leaves behind a hit of 11 million jobs. While the immediate impact was very severe on daily small traders, hawkers and daily wage earners, a more lasting impact has been on the salaried employees.
Footnotes
See consumerpyramidsdx.cmie.com for more details about this survey.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.