One thing is clear from the recent academic, media, and policy discussions on poverty in India: there are many reasonable truths and few consensus facts.
To what extent have the number and proportion of the poor in India changed over time? The picture depends on how one adjusts for differences in the survey instrument used to measure household consumption. That is, if one accepts that consumption is even the appropriate way to measure poverty. To what extent do poverty levels and rates vary across and within states? The answer depends on whether one is using the official Government of India numbers, unofficial but widely noted expert estimates, state government estimates, or looking at broader measures such as the human development index. Who is poor? It depends on how one defines poor. Is it inability to afford a minimal level of calories? Or inability to afford some notional minimum consumption basket? Or vulnerability? Is it a relative or absolute state, and if relative, relative to whom?
With the scope of the National Food Security Bill still up for some debate and the ongoing Socio-Economic Caste Census (SECC 2011) about to deliver a new trove of household information, debates on how many and who are the poor will be in the news for some time to come.
The sixth edition of Prayas provides some background for understanding the numbers as well as the logic and data generating processes behind poverty counting. The issue underscores the extent of the blind spots in understanding poverty as a condition for the purposes of directing policies. These blind spots are inevitable, given the difficulty of reducing a multi-dimensional, socially and contextually influenced, evolving state of existence, into a number or at best a few numbers, but their impact is also exacerbated by the way that poverty alleviation programs are designed.
In highlighting the near-impossibility of statistically “seeing” poverty for directing policy, we hope to raise questions about whether we should concentrate on correcting the blind spots or adapting to them through institutional re-design. Counting the number of poor and identifying poor households is a means to the larger end of improving their opportunities and circumstances; it is not an end in and of itself. Are there ways to reduce our dependence on the numbers?
Counting and Identifying
The first set of articles in this issue highlight the variety of criteria for identifying the poor, the range of estimates of the number of poor that emerge from these techniques, and the importance of these debates for broad policy choices as well as individuals’ and families prospects. Rathish’s Balakrishnan’s essay provides some insight into the range of defensible estimates of poverty in India, as well as the underpinnings for the variation across these summary statistics. Charis Idicheria’s article, “Identifying with States” surveys the landscape of state actions in actually identifying the poor who will be recognized as eligible for national and state anti-poverty programmes.
The issue contains a series of vignettes based on particular experiences of being poor and working on the ground to try to identify the poor. In addition to highlighting the challenges of living anywhere close to the much-maligned 32-rupee a day urban poverty line, the articles about households and communities in rural Gujarat, Karnataka, and Mumbai highlight another dimension of the challenge of identifying the poor: actually conferring and recognizing this definition. The featured individuals are obviously poor by any socially acceptable definition of the number but do not succeed in achieving the official designation for one reason or the other. Practitioner perspective on identifying the poor reinforces the picture of inaccuracy in capturing the location and identity of the deprived.
Looking Forward: Vision Correction?
The next set of essays explores possible avenues for improving the state’s “vision” when it comes to identifying and counting the poor.
Himanshu & Rinku Murgai’s, “SECC 2011: Can it Deliver?” discuss the potential use of the SECC information for delivery of social services. They emphasize several aspects of the critical interplay between information and institutional design: First, that the SECC will make it possible to rank households in various ways, but that programmes must be designed to take advantage of these rankings. Second, that the survey’s importance as a tool for identifying particular families as beneficiaries requires it to focus on verifiable indicators, with some potential loss of accuracy in quantifying poverty. It could produce a more nuanced picture of deprivation by looking into other factors that might not be as easy to audit.
Reetika Khera’s piece on “Escaping the Poverty Line” highlights the problems created by the institutionally embedded interactions between counting and identifying the poor. She argues that the overly optimistic under-counting of the poor is likely to continue affecting the adequacy of anti-poverty efforts, no matter how much the SECC contributes to identification of the poor.
Swathi Shivanand’s article on Delhi’s vulnerability survey, part of the State Government’s “Mission Convergence”, documents an ongoing experiment in conceptualizing and measuring poverty. The shift from “vulnerability” as a guiding concept for identifying those entitled to public support marks a significant philosophical shift in what “poverty” means. It includes “potential poverty” as well as “actual current destitution” as something that is detrimental to quality of life. It also imposes significant new challenges for information collection, since many of the roots of vulnerability are intertwined with absence of recognition from (and visibility to) the state.
Learning to Cope with Blurry Vision
Even with new estimates and re-estimates, perfect vision is unlikely. We must also explore ways to redesign anti—poverty policy and practice to cope with blind spots more effectively.
Perfect vision is unlikely; every approach to measuring poverty has its flaws. Calorie-based approaches conflate energy with nutrition, fall down on generalizations about the true monetary equivalent of calories (given not only food prices but costs of preparation), and ignore the other expenses that households incur as a matter of choice and circumstances. Expanded definitions based on “basic needs” try to capture more of the baseline of a reasonable life, but impose this statistical basic life on actual people who may have other preferences. Income-based metrics have an admirable foundation in the notion of agency and choice – that it’s important to make sure that people have the capacity to live reasonably, but that people should be free to choose how to use their income – but then how does one define “typical living costs” when many of the significant ones may be statistically invisible? Wealth-based metrics start to get at households’ potential resilience to shocks, but then the question is how liquid are the assets? Cattle may have an average value, for example, but the specific price depends on the season and market. Land may make people rich on paper, without being readily convertible to cash for medical expenses.
Moreover, some means of counting are impossible to use for identification. Households being counted have little incentive to lie or misrepresent their economic status. They may be careless, or have limited memories, but they neither gain nor lose from giving false information. Households being assessed for identification of the poor, on the other hand, have strong incentives to misrepresent their status and the collection process has to have to build in safeguards to limit these errors. These safeguards, in turn, can limit the accuracy of the assessment by forcing the data collectors to have to rely only on observable information rather than subjects’ memory or personal experience.
Much of the discussion has focused on better measurements of poverty, but we should also think seriously about how to adjust to the inevitability of imperfect visibility.
Innovate in Identification
One pragmatic step would be to separate identification of the poor from the debate over aggregate counts. Counting and identifying the poor are necessarily intertwined – every effort to identify the poor necessarily produces a count, and one of the counts in turn affects the numbers who can feasibly be identified as poor – but they fulfill different functional purposes. One determines broad allocations, the other specific lifelines for particular people.
Innovations in identification are especially important since the consequences of errors in identification are different than errors in counting. Counting errors influence the allocation of resources between levels of government and across programs, but they are invisible on a human scale. The population of a state may receive less social spending, but it is hard for any particular individual to perceive how much they personally were affected by the margin of error in the count of the poor. Mistakes in identification, on the other hand, have concentrated and identifiable impacts on particular people. Exclusion can be catastrophic for an individual. Errors in counting also cancel each other out in some senses – one household missed, another wrongly included leaves the overall count correct and unbiased. For identification, one missed and another included means two errors, albeit with different impacts.
States already adjust around the edges imposed by budgets and central anti-poverty program design to set their own criteria for identifying the individuals behind the aggregate numbers. Encouraging further experimentation could generate lessons for future national efforts to identify poverty as well as support constructive competition in efforts to ensure that resource reach the neediest. Diverse approaches in identifying the poor would confound efforts to create a uniform national safety net, but wide dissemination of identification criteria and transparent implementation could safeguard against many exclusionary practices.
Use Other Ways to Hone in on Poverty
Poverty counts and poverty designations are important parts of efforts to direct resources to the most vulnerable. However, individuals’ status is not the only indicator of need. If we can’t identify poor people exactly, why not shift some of the attention to poor contexts and make a concerted effort in ensuring that every person has access to an infrastructure and services baseline? It is easier to spot the absence of a road, the dilapidation of a school, or the strength of a pharmacy’s stock than to define the state of being and vulnerability of a person. Headcounts are a means of assessing relative levels of poverty for the purposes of distributing funds across nations, states, and districts, but if we know that they are inaccurate or misleading about special circumstances of deprivation, is there a way to set up a mechanism for corrective transfers, some kind of arena for states or local governments to make a public case for additional resources? Such an exercise might also motivate lower levels of government to devote more effort to understanding and documenting the conditions that their residents live in. It might also help the country understand more about the variety of circumstances across states.
Adjust Efforts to Minimize Consequence of Mistakes
Strong reliance on the BPL card as the single point of access to a multitude of services also increases the consequences of errors in designation. The BPL designation is convenient as a summary statement of eligibility for programmes, but the cost of this convenience should be considered and more diversity in eligibility criteria for specific programs encouraged.
Finally, tightly targeted anti-poverty policy places enormous weight on the quality of the government’s eyesight. The fact that the blur is inevitable, especially for aspects of poverty and vulnerability such as hunger and ill-health, calls for more attention to strategies that loosen criteria for lifeline benefits where the consequences of mistakes are most significant. Universalisation of one programme is sometimes seen as the start of a slippery slope toward universalisation of everything, but it would be more humane to rank programs in order of the likelihood of mistakes, the consequences of mistakes and adjust the targeting accordingly.
In summary: the debate about the poverty numbers will go on. The efforts to reach the poor should too.
Dr. Jessica Seddon is Head of Research at the Indian Institute for Human Settlements, Bangalore. Jessica has advised a number of state and national government initiatives in India as well as published book chapters and articles in international academic and policy journals. Her work has been supported by grants from Google.org, Australian Development Research Awards, ICICI Foundation for Inclusive Growth, Stanford University, and University of California, San Diego, among others. She is a member of the Advisory Board of the Wilderhill Global Clean Energy Index (NEX). She is also a Research Scholar, Global Information Industry Center, University of California, San Diego.
Jessica is a term member of the Council on Foreign Relations. Prior to joining IIHS, she served as Director of the Centre for Development Finance and as Professor at the Institute for Financial and Management Research in Chennai. Jessica has also taught as Assistant Professor, Political Economy, at UC San Diego and worked at the World Bank, Asian Development Bank, and Inter-American Development Bank.