In the second of a three-part series on the poverty line, Prof. S. Subramanian, former National Fellow, Indian Council of Social Science Research, argues that there is a built-in incentive for official poverty lines to be pitched ‘low’.
It was our contention, in the previous article on this subject, that in commonly understood terms which are compatible with a responsible use of language, the term ‘poverty line’ should refer to “the estimated minimum level of income needed to secure the necessities of life” (the definition suggested by Oxford Dictionaries.) The language of the ‘poverty line’ therefore suggests a view of income in which it is a means to an end, rather than an end in itself. (Not that there is anything intrinsically objectionable to a view of income as an end in itself: it is just that the notion of a ‘poverty line’ is incompatible with such an interpretation of income.) The ‘poverty line’ approach to conceptualising income deprivation therefore sees income (or resources in general) as a means to the end of avoiding deprivation in the space of what the economist Amartya Sen has called human functionings. A ‘functioning’, in Sen’s view, is ‘a state of being or doing’, such as is reflected in a person’s nutritional status, or status in respect of access to shelter, or clothing, or good health, or knowledge, or the ability to move about freely, or the ability to ‘appear in public without shame’ (a desideratum much stressed by the 18th century political economist Adam Smith, and repeatedly endorsed by Amartya Sen.)
There are certain unavoidable implications for identifying the poor which a poverty-line approach to conceptualising poverty entails. Implication 1 is that the logical way of deriving a poverty line is the following one: first, identify a set of human functionings in respect of which escaping deprivation is necessary to qualify for not being judged poor; second, agree on minimally acceptable levels of achievement for each of these functionings; third, assess the commodity requirements for attaining these minimal levels of achievement; fourth, compute the reasonable cost of the commodity requirements for each valued functioning; fifth, add up these costs to arrive at the poverty line.
Implication 2 is that individual-specific poverty lines derived in accordance with Implication 1 are most unlikely to be identical across individuals. This is because of heterogeneities that must be expected to obtain, across both individuals and ‘contexts’ or ‘environments’. What does one mean by this? Some examples should help to illustrate the point. Consider two individuals A and B who have the same income but who have one important non-income characteristic in respect of which they are very different: A is physically handicapped, but B is not. Typically, to achieve the same level of functioning in the matter of ‘mobility’, the physically handicapped individual A will require more resources (in the form of a wheel-chair, say) than individual B. This is another way of saying that it is much more difficult for A to convert income (or resources in general) into functionings than it is for individual B. But since it is functionings that ultimately matter - as we have argued earlier - A must be judged poorer than B even when both persons have the same income. Or putting it differently, the income poverty line for A must be judged to be higher than for B.
Consider another example involving a variation in ‘environment’. Suppose A lives on the plains of Dindigul district in Tamil Nadu, where it is always warm (or at least never particularly cold), while B lives in the hills of Kodaikanal, where it is almost always cold. For both individuals to achieve the same level of functioning with respect to ‘adequacy of clothing’, clearly B will require access to more warm clothes than A. Even if both the income and the non-income characteristics of A and B were identical, the income poverty line for A must be judged to be lower than for B. Variations in the objective environment, or ‘context’ in which life is lived, again ensure that there are interpersonal differences in the ability to convert resources into functionings. Implication 2 of adopting a meaningful poverty-line approach to conceptualising poverty therefore goes against the grain of postulating a unique income poverty line for all individuals and all environments. Implication 2 can be pithily stated in Amartya Sen’s terms as the proposition that it is sensible to view poverty as an absolute concept in the space of human functionings, but - because of interpersonal variations in the ability to convert incomes or resources into functionings - as a relative concept in the space of incomes or resources.
It should be mentioned here that there is one logical situation in which deferring to Implication 2 is still compatible with violating Implication 1. This is the situation in which we advance a unique poverty line, but one that is yielded by the largest of the individual-specific poverty lines derived through deference to achievements in functionings space: we can be sure, in such a situation, that no-one who is poor will be left out of the count of the poor. This is a cautious - and therefore liberal - approach to identifying the poor. By contrast, a conservative (or crabbed and niggardly) approach would advance a poverty line that is yielded by the smallest of the individual-specific poverty lines: we can be sure of excluding all non-poor persons from the count of the poor in such an approach, but we cannot be sure of including all poor persons within the count. It should be noted that identifying the poor is only the first step in measuring the poor. The second step is aggregation: the process of combining information on the poverty line and the distribution of incomes with a view to coming up with a number which is intended to signify the amount of poverty that obtains in a society. The simplest such measure of poverty is the headcount ratio, or proportion of the population below the poverty line. It does not require rocket science to deduce that, other things equal, the headcount ratio will increase as the poverty line is increased. There is therefore a built-in incentive for official poverty lines to be pitched ‘low’.
How does the World Bank global poverty line measure up to Implications 1 and 2, as discussed above? First, the Bank proposes a unique ‘dollar-a-day’ poverty line for all countries of the world and over time. (The line is approximately US$1 per day, expressed in ‘purchasing power parity terms’, to correct for variations in country-specific currencies’ purchasing power.) This is by no means the maximum of country-specific poverty lines, derived in accordance with functionings-inspired calculations. Rather, it is the minimum of country-specific poverty lines, based on the actual poverty lines (some of which have in fact been recommended by the World Bank itself!) employed by a few of the world’s poorest countries. The global poverty line employed by the Bank is less than 10% of the US poverty line. It is with reference to such a low poverty line that the Millennium Development Goal of poverty reduction has been set. The philosopher Thomas Pogge has demonstrated that global achievements in poverty reduction become altogether less impressive if the poverty line is raised to a less modest, but nevertheless modest, US$2.50.
Briefly, the World Bank’s global poverty estimates are based on a conceptually dubious and ungenerously Spartan reckoning of the poverty line.
This article first appeared in the national daily newspaper The Tribune under the title ‘Identifying the poor’ on 24 July 2015. The author is indebted to the editor of The Tribune, Dr Harish Khare, for permission to reproduce the article here.
The first part can be accessed here. The third and last part will be posted on I4I on Friday, 27 May 2016.