Poverty is bad – but is vulnerability worse?

Official data reveal that poverty in India has declined significantly over time. In this post, Kamila and Wadhwa make the case for policy discourse to shift towards the phenomenon of ‘vulnerability’. Presenting their view on alternative methods of quantifying vulnerability, they highlight the challenges of leveraging existing data for this purpose and the kinds of data that would be needed to capture the essence of vulnerability.

The discourse on development has always been suffused with discussions around poverty and standards of living. As early as 4^th century BC, in Kautilya’s Arthashastra – a treatise on statecraft and policymaking – there is mention of the imperative for the State to support the needy. As social reformers advocated for equitable living standards, poverty became a major public focus during the colonial period. Dadabhai Naoroji’s seminal work Poverty and Un-British Rule in India (1901) is not only distinguished by the rich data-driven, evidence-based discussion on the lopsided development policies of the British in India, it also captures snapshots of appalling poverty in 19^th century India. Consequently, poverty alleviation was a key theme in the struggle for Independence. Post-independence, policymaking has understandably focused on poverty reduction, with varying degrees of success.

However, there is a discernible trend of secular decline in poverty. Estimates of poverty as per the accepted methodology for the years since 1973-74 show that at the all-India level, poverty (as a percentage of population below the poverty line) declined from about 55% to 36% in 1993-94, and further to 22% by 2011-12. Since official estimates of poverty are based on the Household Consumption Expenditure Surveys conducted by the National Sample Survey Office (NSSO), and the latest results from these surveys are available for 2011-12, more recent official data are missing. Although exact numbers might differ, the trend of decline is palpable for various definitions of poverty – like US$1.90 per person per day in 2011 PPP (purchasing power parity) terms, or US$2.15 per person per day in 2017, the International Poverty Line estimated by the World Bank. Notably, multidimensional poverty¹ also declined significantly, with India reducing the proportion of multidimensionally poor from 24.85% in 2015-16 to 14.96% in 2019-21 – effectively lifting 135 million people out of poverty (NITI Aayog, 2024).

Policymakers appear to have identified certain key factors in poverty reduction, including (i) universal education for girls; (ii) declining fertility rates; (iii) widespread maternal and child health services; (iv) women’s economic mobility; (v) formation of women’s self-help groups (SHGs); (vi) provision of credit to SHGs; and (vii) pro-poor welfare schemes providing basic amenities like gas, electricity, and bank accounts. As a result, debates around poverty eradication per se have become less prominent in contemporary India.

Lest one might be tempted to think that we have completed the long march, let us be cognisant of the fact that fresh thinking has emerged on the subject. For instance, in an increasingly volatile, uncertain, complex and ambiguous world, it is imperative to ensure that those who have risen above poverty do not slip back below the threshold and those below it continue to maintain or improve their standard of living. Recent observations of poverty resurgence following the Covid-19 pandemic reveal that, notwithstanding substantial progress in poverty alleviation, a considerable segment of the population in numerous countries continues to be at risk of descending into poverty (World Bank, 2020). Some surveys have, in fact, highlighted that for India, decline in poverty has occurred with increasing vulnerability (Desai 2024). In view of recent developments suggesting that India is contemplating the formulation of new poverty lines, it is instructive to highlight the policy priority of tackling vulnerability as a supplementary strategy to eradicating income-poverty.

Defining vulnerability: Why it is different from poverty

Vulnerability, in economic terms, refers to the likelihood of individuals or households falling into poverty due to exposure to risks or shocks. It is a forward-looking measure that assesses potential declines in well-being on account of unpredictable events like natural disasters, economic downturns, or health crises. Unlike poverty, which is often static, vulnerability captures the uncertainty and dynamism of future risks (Dutta and Mishra 2023).

The two related concepts are very different from each other. A household may not be below the poverty line today but still could be vulnerable to falling into poverty in the future. Understanding this distinction is crucial for designing effective interventions to address vulnerability. Some dimensions along which the two differ are given in Table 1 below.

Table 1. Dimensions along which poverty differs from vulnerability

Dimension	Poverty	Vulnerability
Temporal	Static indicator of current deprivation in income or consumption	Forward-looking, in terms of risk of future poverty and potential impact of shocks on welfare over time
Predictive	Involves identifying households currently below a certain threshold (for example, living on less than US$1.90 a day)	Predicts who might fall below a certain threshold due to various risks – taking into account frequency, intensity, and impact of potential shocks
Policy approaches	Focus is on providing relief and lifting households above the poverty line, through cash transfers, subsidies, and basic goods and services	Focus is on building resilience and reducing exposure to risks, through improving access to insurance, promoting diversified income sources, enhancing social safety nets, and investing in infrastructure to mitigate impact of shocks
Measurement complexity	Relatively straightforward, measured through surveys	Requires comprehensive data on dimensions of risk and coping mechanisms as well as assumptions about likelihood of households falling below poverty line in future

Types of vulnerability

Households can be vulnerable to a variety of shocks, including:

Economic: A large part of India's workforce, especially in rural areas, relies on informal sector jobs that lack security, benefits, and minimum wage protection. Economic downturns, seasonal fluctuations, or illness can lead to job losses and income disruptions, pushing families into poverty (OECD and ILO, 2019).

Health: India's healthcare system has significant gaps in insurance coverage, especially for the poor. Even minor illnesses can lead to substantial medical bills, forcing families to sell assets, borrow money, or forgo necessities (World Bank, 2005).

Climate: A large portion of India's population depends on rain-fed agriculture. Droughts, floods, and erratic weather patterns can destroy crops, leading to income loss, food insecurity, and forced migration. Making the agriculture sector resilient to climate change is crucial, given its severe impact on India (Prasad 2024).

Fiscal: The government often subsidises essential goods and services like food and fuel to make them affordable for the poor. For instance, the Public Distribution System (PDS) provides subsidised grains. However, reducing these subsidies in the pursuit of fiscal consolidation can push many households below the poverty line.

Some shocks, like the Covid-19 pandemic or natural disasters, are aggregate, while others (for example, job loss) are more idiosyncratic. A household’s vulnerability also depends on their resilience against shocks, which may vary based on their assets and informal networks. In the absence of formal mechanisms, these networks often serve as the backbone of resilience.

While social networks can help mitigate individual risks, these mechanisms often fail during aggregate crises (Gaiha and Imai 2008). These complexities make vulnerability measurement difficult, and call for more research for effective policymaking.

Measuring vulnerability

There have been multiple attempts at measuring vulnerability, with approaches of varying complexity (Dutta et al. 2011, 2019, Gaiha and Imai 2008). To illustrate the idea, we focus on one specific measure discussed in Gaiha and Imai (2008), that uses ICRISAT (International Crop Research Institute for the Semi-Arid Tropics) panel data to measure vulnerability. This intuitive measure, known as Vulnerability as Expected Poverty (VEP), was proposed by Chaudhuri, Jalan and Suryahadi (2002), who applied it to Indonesian household data. To estimate the probability of a household falling below the poverty line in the future, historical household data is used to estimate a distribution of household consumption. Once this distribution is available, one can estimate the probability of consumption falling below a threshold in the future. As the authors outline, empirically, both consumption and variance in consumption can be assumed to be ‘linear’ functions of household characteristics. Data are used to estimate the parameters of these linear functions. By further assuming the distribution of consumption to be ‘normal’, one estimates the probability of consumption falling below the poverty line in the future.²Evidently, this measure of vulnerability imposes a number of (arguably strong) assumptions. However, as discussed earlier, given the nature of vulnerability, any measure will need to make certain assumptions about future shocks and the ability to deal with them.

While VEP can be estimated with cross-sectional data, panel data can offer deeper insights by tracking the same cross-section of households over time, revealing fluctuations in their consumption and income, as well as their ability to deal with such variation. Specifically, we propose collecting the following panel data to better quantify vulnerability:

Household consumption: Regular, detailed data on consumption patterns and fluctuations.
Income and employment: Information on income sources, employment status, and income variability.
Assets and liabilities: Data on asset ownership, liabilities, and savings.
Health: Data on health expenditures, illness incidence, and healthcare access.
Agriculture: Information on crop yields, livestock, and agricultural inputs, especially in climate change-prone areas.
Demographics: Data on household composition, education, and other socioeconomic indicators.
Amenities: Data on access to amenities like clean water, cooking fuel, public transport, and affordability thereof.

One example of such data is CMIE’s (Centre for Monitoring Indian Economy) Consumer Pyramids data, which follows the same households every four months³, collecting much of this information (except agricultural yields). However, the dataset over-samples urban populations, which are arguably less vulnerable to begin with.

To effectively measure vulnerability, it is imperative that efforts are initiated to start collecting data more holistically across the country. It is important to stress here that data required for getting a grip on vulnerability need to be recorded in a way that there are no concerns about the comparability and compatibility of data when drawn from a diversified set of sources. Given the scope and scale of this undertaking, it would be prudent to begin at a smaller level, such as at the state level, before expanding to the national level. This approach allows for the refinement of data collection methods and the assessment of initial findings, which can guide larger-scale implementation.

We can also leverage existing data sources in some states or specific districts that already have a foundation of relevant data. By starting with regions where data infrastructure is relatively well-developed, the methodology can be piloted, identifying challenges and making necessary adjustments.

Conclusion and way forward

Policymakers need to engage with the concept and perception of vulnerability and move towards a national consensus on how it should be defined and managed. To be sure, there have been sectoral initiatives to acknowledge the need to build resilience against certain kinds of vulnerability (such as in agriculture (Sharma 2024) and industry (Dora and Mishra 2024)), but there is a need for a comprehensive approach, which recognises and addresses multiple sources of vulnerability. In this regard, it would be helpful to pilot a few initiatives of panel data collection in states where datasets on profile of residents already exist, such as in Madhya Pradesh, Haryana or Karnataka. With growing uncertainty, we need to plan for vulnerability assessment in a more structured manner, and start taking steps to identify households vulnerable to a variety of shocks.

The views expressed in this post are solely those of the authors, and do not necessarily reflect those of their organisations or of the I4I Editorial Board.

Notes:

In 2010, the Multidimensional Poverty Index (MPI), developed by Alkire and Foster, was adopted by the United Nations Development Programme (UNDP) in their Human Development Report. It captures overlapping deprivations in health, education and living standards. In India, NITI Aayog, in collaboration with UNDP, and Oxford Poverty and Human Development Initiative (OPHI), developed the National Multidimensional Poverty Index (MPI) that offers a multi-dimensional perspective on poverty. The national MPI largely follows the global methodology, retaining 10 indicators from the Global MPI, but has added two new indicators, namely Maternal Health (in the dimension of Health) and Bank Account (in the dimension of Standard of Living).
A linear function is a mathematical expression where a change in one variable is directly proportional to a change in another variable. For example, if the size of a household increases, the total consumption might increase in a predictable way based on a constant rate. A normal distribution, sometimes called a 'bell curve,' is a statistical pattern where most values cluster around a central point, with fewer values appearing as you move further away from this centre. Assuming this distribution makes it easier to estimate probabilities, such as the chance of consumption falling below the poverty line.
The advantage of collecting data at a high frequency is that it allows us to study the fluctuations in consumption within a year, that is, due to different seasonal patterns. It also allows us to study the ability of households to deal with these shocks within a short period of time. In surveying households after many years, we may miss out on a lot of information. It would also allow for many other household factors to change, making it difficult to study the support system they have to deal with temporary shocks.