(This essay is the first part of a two-part explainer from the Centre for New Economics Studies (CNES) InfoSphere's latest edition on poverty assessment in India: a pre-covid and post-covid difference in difference perspective. Review InfoSphere’s work here for more details.)
A contested academic debate has surrounded the measurement (and state) of poverty in India in recent years, due to the government's decision to not release the results of the National Sample Survey's (NSS) 2017-18 Consumer Expenditure Survey (CES). The lack of credible government data has further created uncertainty, compounded by the absence of NSS data since 2011-12.
Recently, two papers from the International Monetary Fund (IMF) and the World Bank attempted to estimate India’s poverty estimates while giving rise to different conclusions.
The resulting debate, often referred to as the "Great Indian Poverty Debate 2.0," mirrors a previous discussion on the impact of India's economic liberalisation in the 1990s (called the Great Indian Poverty Debate 1.0).
There are various studies which have been published on analysing poverty estimates since 2012 (see Figures below for a reference guide).
This first essay from our Centre – InfoSphere’s research on poverty assessment in India focuses on two papers and their observations.
The first paper is titled ‘Pandemic, Poverty and Inequality’ authored by Surjit Bhalla, Karan Bhasin, and Arvind Virmani, referred to as BBV. It is a part of the working paper series of the International Monetary Fund published in 2022. The second paper titled ‘Poverty in India Has Declined over the Last Decade But Not As Much As Previously Thought’ authored by Roy and Weide (referred to as RW) is a part of the working paper series of the World Bank published in 2022.
Both studies use alternative data sources, methods, and assumptions to derive estimates of the poverty headcount ratio for India in the recent years and reach contrasting conclusions.
'Pandemic, Poverty and Inequality'
The paper by Bhalla, Bhasin, and Virmani (BBV) uses household survey data to estimate the impact of the pandemic on poverty and inequality in India. It considers Private Final Consumption Expenditure (PFCE) to obtain the poverty line, MMRP (Modified Mixed Reference Period), and URP (Uniform Recall Period) to collect data on household consumption expenditure. To account for the cash equivalent of food subsidies, they added the cash equivalent of food subsidies (inclusive of leakage) to eligible households' expenditures. They also adjusted the cash equivalent for leakage if applicable.
The BBV paper also uses State Domestic Product (SDP) as a proxy variable for measuring economic activity at the State-level.
Their study considers the poverty line of USD 1.9 PPP (Purchasing Power Parity) which is the international poverty line to compare poverty levels across different countries, as well as the poverty line of USD 3.2 PPP which is commonly used as a poverty threshold specifically for lower-middle-income countries.
The BBV paper finds that SDP (State Domestic Product) has a significant impact on poverty levels. States with higher SDP tend to have lower poverty levels
The paper highlights that extreme poverty in India remained at a low level of 0.8 percent in 2020, due to government food transfers
The BBV paper also finds that the Subsidy to the Poverty Line (SPL) has increased in India in recent years (higher in rural areas than in urban areas)
The rise in SPL is due to several factors, including population growth, urbanisation, and an increase in government assistance programs. The SPL (Subsidy to the Poverty Line) is a measure of the extent to which government subsidies help to reduce poverty– a lot of this spending by the government (on subsidising food to fuel) is directly relevant in keeping poverty levels low. A higher SPL means that government subsidies are more effective at reducing poverty.
The BBV paper finds that the SPL has had a significant impact on reducing poverty in India. It has reduced the proportion of the population living in extreme poverty by 1.5 percentage points since 2011-12.
200However, the paper also notes that the SPL is not a perfect measure of poverty reduction as it does not consider the quality of the goods and services that are subsidised.
Although the paper provides key insights into the poverty scenario in India, there are certain limitations to it. The BBV poverty paper uses the CES (Consumer Expenditure Surveys) method to estimate poverty in India.
The CES method may not sufficiently cover wealthy households. It also excludes individuals living in institutions and those without permanent housing. Similarly, there are other shortcomings in the CES method.
The Roy and van der Weide Paper
The next poverty assessment paper by Sutirtha Sinha Roy and Roy van der Weide uses two approaches to align the Consumer Pyramids Household Survey (CPHS) with the National Sample Survey (NSS) Consumer Expenditure Survey (CES) 2011-12 and impute NSS consumption expenditure into CPHS.
Their paper uses data from the Consumer Pyramids Household Survey (CPHS) to estimate poverty at three different levels: USD 1.9 PPP which is the international poverty line, USD 3.2 PPP which is a poverty threshold for lower-middle-income countries and the USD 5.5 PPP which measures poverty for higher income countries.
The paper finds that poverty at all three levels has declined significantly since 2011. Poverty at USD 1.9 PPP has declined from 22.5 percent in 2011 to 10.2 percent in 2019. Poverty at USD 3.2 PPP has declined from 61.7 percent in 2011 to 44.9 percent in 2019. And poverty at USD 5.5 PPP has declined from 87.4 percent in 2011 to 80.9 percent in 2019.
The paper also finds that the decline in poverty has been greater in rural areas than in urban areas. For instance, poverty at USD 1.9 per capita per day has declined from 32.8 percent in rural areas to 14.6 percent in urban areas.
The paper also finds that extreme poverty in India has declined by 12.3 percentage points since 2011, but that the pace of reduction has been slower than previously thought. The paper highlights that poverty is more prevalent in rural areas than in urban areas and that it has increased in some states in recent years.
In the absence of a more suitable alternative, the World Bank has incorporated the findings as its official estimate of poverty in India.
Although the paper provides a comprehensive analysis of the poverty scenario in India, it has certain shortcomings. The paper uses the Consumer Pyramids Household Survey (CPHS), which has some limitations. The CPHS samples are primarily taken from main streets, which means that it under-represents the poorest and richest households. Also, the mean consumption per person in the CPHS is much lower (33-35 percent) than the official national accounts data, which suggests that the CPHS may not accurately represent the true consumption levels in the country.
Methodological Differences and Comparative Perspectives
The BBV and RW poverty papers have some key differences in the rate at which poverty has declined in India. These differences in the findings of the two papers can be attributed to the different data sources used:
The NSS CES is a household survey that is conducted by the government of India and is a larger survey. The CPHS is a market research survey that is conducted by the Consumer Pyramids Foundation. Thus, the NSS CES is a more reliable source of data on household consumption expenditure than the CPHS.
However, the CPHS has some advantages over the NSS CES. The CPHS collects more detailed data on household consumption expenditure. This allows the RW paper to make more nuanced estimates of poverty in India. The BBV poverty paper has been used by the government of India to justify its poverty reduction policies. Whereas the RW poverty paper has been criticized for its methodology.
However, the two papers complement each other, and they aim to provide an insight into the picture of poverty in India.
Leaked NSS CES 2017-18
The Indian government did not release the NSS CES expenditure for 2017-18. This makes the task of keeping methodological coherence in poverty estimation extremely difficult. As a result, the above two working papers in discussion were crucial in understanding the poverty trends, while using different proxy (substitutive measures).
However, Arvind Subramanian, a former chief economic advisor to the Government of India, used a leaked draft version of the National Sample Survey (NSS) for 2017-18 to estimate poverty levels in India. He used the poverty line recommended by the Rangarajan Committee which is about 25 percent higher than the poverty line established by the Tendulkar Committee.
Some Important Findings:
Subramanian's estimates showed that the poverty rate in India increased from 31percent in 2011-12 to 35 percent in 2017-18. This makes his poverty claim differ from those of the RW and BBV papers
The increase in extreme poverty from 12.2 percent in 2011-12 to around 17 percent in 2017-18 is a cause for concern
Arvind Subramanian's estimates have been challenged by some economists, who argue that the NSS data is not reliable and that the Rangarajan poverty line is too high. This can be examined by using the NSS Survey to National Accounts Ratio.
NSS Survey to National Accounts Ratio is a measure of the consistency between the estimates of consumption from the NSS survey and the official national accounts data. The lower the ratio, the more divergent the two sources of data are. In 2017-18, the NSS survey to National Accounts Ratio was 39.8 percent, which is the lowest level recorded in Indian history. This suggests that the NSS survey data may be underestimating consumption in India.
On the contrary, the official national accounts data may be overestimating consumption. Subramanian's estimates have also been supported by other economists, who argue that they provide a more accurate picture of the state of poverty in India.
However, a key question to raise at this point (going beyond the jargon) is whether consumer expenditure is the best indicator of poverty. Is there any other index that can comprehensively measure poverty?
These are critical theoretical questions that require a more critical dialogue and intellectual scrutiny. Unfortunately, in a hyper-polarised, political debate on poverty estimates, social policy analysis and its theoretical assumptions/questions, have lacked clarity and reason, which has confounded more confusion than clarity amongst economists and other social scientists.
(Deepanshu Mohan is a Professor of Economics and Director, Centre for New Economics Studies (CNES), Jindal School of Liberal Arts and Humanities. He is currently a Visiting Professor to the School of International Development and Global Studies, Faculty of Social Sciences, University of Ottawa. Yuvaraj Mandal, Amisha Singh, Nishit Patel, Aditi Desai, and Amisha Mittal are all Research Analysts with CNES and members of the InfoSphere Team. This is an opinion piece and the views expressed are the author’s own. The Quint neither endorses nor is responsible for them.)