An IMF Paper Says ‘Extreme Poverty’ Didn’t Worsen In India Amid Covid. Not Many Are Convinced
An IMF paper claims that the share of extreme poor was largely unchanged amid the pandemic in India. Economists are skeptical.
An IMF working paper has once again ignited debate over poverty in India.
“Extreme poverty”, those falling below $1.9 on purchasing power parity basis per person a day in India, fell to below 1% of the population in 2019, according to the paper, titled ‘Pandemic, Poverty, and Inequality: Evidence from India’.
This is a substantial reduction in poverty rates from 2011 levels. It remained below 1% even in 2020, the pandemic year, said the paper—authored by Surjit Bhalla, executive director at IMF for India, Bangladesh, Bhutan and Sri Lanka; Karan Bhasin of University at Albany; and Arvind Virmani, chairman at EGROW Foundations.
These estimates include the effect of in-kind food subsides on poverty and inequality that other estimates of poverty do not incorporate, overestimating poverty. The doubling of entitlements under the Pradhan Mantri Garib Kalyan Ann Yojana prevented any substantial increase in extreme poverty, the paper said. Post-food subsidy inequality as measured by the Gini coefficient was found to be at 0.294—very close to its lowest level of 0.284 in 1993-94.
According to the more appropriate low-middle income poverty line of PPP$3.2 a day, 14.8% of the population was poor in pre-pandemic 2019. That rose to 18.1% in 2020. This was after adjusting for food transfers, 18.5% of the population would be poor in 2019, rising to 26.5% in 2020.
The paper, also, clarified that the “views expressed in IMF working papers are those of the authors and do not necessarily represent the views of the IMF, its executive board, or IMF management”.
But not many agree.
Santosh Mehrotra
Visiting Professor, Centre for Development, University of Bath,
Research Fellow, IZA Institute of Labour Economics
Ex- Director-General, Institute of Applied Manpower Research, Planning Commission
“The findings of the report do not stand up to logic.”
Methodology Issues:
Globally, poverty is estimated based on private final consumption expenditure derived from household surveys, not national accounts. Consumption expenditure in national accounts statistics is a residual. Unlike exports, gross fixed capital formation and government consumption expenditure, which are hard numbers, private final consumption expenditure that comprises 56-58% of the GDP by expenditure, is an estimate.
“We know from experience that NAS overestimates PFCE and poverty based on this estimate would be lower than what it would be if based on CES. That is how Bhalla et al come up with unbelievable poverty ratio for India of 0.9% of the population, which the 5 kg of cereal transfer alone reduced even further to 0.8%.”
The latest Consumption Expenditure Survey of 2017-18 was junked by the government. The survey was preceded by sudden demonetisation in 2016 and unplanned implementation of the untested GST—“both of which knocked the bottom out of India’s informal sector.” It slowed the economy for nine full quarters, all the way to end 2019. As such, the findings of the CES indicated a reduction in PFCE in 2017-18 compared to 2012.
“The Consumption Expenditure Survey of NSSO in 2012 gave us a poverty estimate of 22%. What could have happened between then and 2020, for poverty to have reduced by that much?”
Substantive issues
While the authors have considered the inflation in food items supplied as part of the food subsidy scheme, what about aggregate food inflation? The rise in food inflation is so much in that year it almost negates the provision of food subsidy. One can’t be expected to survive on just the subsidised food items. Food inflation, according to the RBI Bulletin in mid 2021, was 31% between July 2020 and July 2021.
Household expenditure also saw a rise because of private out-of-pocket expenses borne because of health expenditure. How does the survey take that into account?
According to Mehrotra’s own estimates, poverty was 21% of the population in 2020, compared to 22% in 2012. In absolute terms, the number of poor by the (Tendulkar) national poverty line rose from 270 million in 2012 to 285 million in 2020 (just before Covid). These estimates are derived from the PLFS for 2020, which also provides data on consumption though it is not detailed like in the CES. Post Covid, poverty worsened further because unemployment, already at a 48-year high in 2019, rose further.
Mahendra Dev
Director and vice chancellor at Indira Gandhi Institute of Development Research
It's difficult to believe that inequality declined during the pandemic, said Dev. It's known that the pandemic had a K-shaped impact on the economy. Corporates, including listed companies and stock markets, have done well, while informal workers, migrants and MSMEs lost incomes and jobs, he said.
There’s no official consumer expenditure data for India after 2011-12. The data collected in 2017-18 was not used because of ‘low quality’. Bhalla et al use 2011-12 data and extrapolate with the private consumption data of National Accounts Statistics data up to 2020-21. “We do not know whether this methodology is correct or not. This may underestimate poverty as growth of consumption in surveys is lower than national accounts. Second, they also state domestic growth for extrapolating consumption data. This also has limitations.“
The 1% extreme poverty relates to the $1.99 poverty line. For India, it is better to use the $3.2 poverty line. “If we take that poverty increased from 18.5% in 2019-20 to 26.5% in 2020-21 (for MMRP method) without transfers. With transfers, the poverty increased from 14.8% in 2019-20 to 18.1% in 2020-21.“ Therefore, in the pandemic period, poverty increased in spite of food transfers. Much more fiscal stimulus should have been there given the fiscal deficit constraints.
$1.9 and $3.2 poverty lines are at the national level. Although they did some estimates using CPI data at state level, but international poverty lines are taken. “In the poverty estimates, we first estimate at state level and aggregate to national level.“
In the pre-pandemic period, economic growth declined from 8% in 2016-17 to 3.6% in 2019-20. Unemployment increased. Real wage growth declined. “In this context it is difficult to believe so much decline in poverty during 2011-12 to 2019-20,“ Dev said.
People gesture for food while waiting in a bus during a lockdown in New Delhi, India. (Photographer: Prashanth Vishwanathan/Bloomberg.)
Amit Basole
Head, Centre for Sustainable Employment at Azim Premji University
The first assumption is that consumption as recorded in survey data grew at the same rate as PFCE. This happens to be true for the period 2004-05 to 2011-12, as seen by the fact that the ratio of NSSO consumption to PFCE remained constant around 50%. But it was not true for the period 1993-94 to 2004-05, when PFCE grew faster. It may not be true of the period 2011-12 to 2017-18 either given shocks like demonetisation and goods and services tax that occurred in this period.
The second assumption is that the impact of GDP growth is constant across sections—meaning that incomes of the poor grew at the same rate as the rich. This has not been demonstrated in the paper.
The consumer expenditure survey used is as of 2011-12. In a CES survey, households are asked about the quantity of the food grains that they receive. But since no such survey data are available after 2011-12, how does one determine the share of households that received PDS, what income brackets they belong to and how much they received? The paper does not adequately delve on this.
The authors have extrapolated survey results from the survey conducted in 2011-12. The latest one was conducted in 2017-18 and was debunked by the government, Basole said.
The share of those below the poverty line, according to Basole, would have been higher if the 2017-18 CES was used as the data recorded a fall in consumption for the first time since records are being kept (this was the main reason it was junked).
With a higher poverty line at $3.2 a day (that the authors recommend should be used), poverty rose around 8 percentage points from 18.5% to 26.5% during the pandemic).
To be sure, the World Bank uses the poverty line at $1.9 per person a day, and the authors have done the same.
One has to be careful that the figures apply to ‘extreme poverty’. “For our research, we studied the percentage of people under a higher poverty line that we found more meaningful,” Basole said.
The State of Working India report for 2021, that Basole co-authored, had estimated that the first wave of the Covid-19 crisis pushed nearly 23 crore Indians into poverty.
A Pew Research Centre Analysis had estimated that the Covid-19 recession led to nearly 7.5 crore Indians being pushed into poverty, accounting for nearly 60% of the global increase in poverty. The think tank benchmarked income to $2 or less a day.