Jump to ContentJump to Main Navigation
On Economic Inequality$

Amartya Sen

Print publication date: 1973

Print ISBN-13: 9780198281931

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198281935.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 25 February 2017

A.6.1 Poverty: Identification and Aggregation

A.6.1 Poverty: Identification and Aggregation

On Economic Inequality
Oxford University Press

While OEI‐1973 considered the varying importance of inequality in different parts of the distribution of incomes, it did not specifically investigate poverty. Concentration on poverty in particular as opposed to inequality in general would require a very specific focusing on the predicament of the poor. In addition, any assessment of poverty cannot be entirely ‘relative’, since absolute incomes (and absolute opportunities in general) must have a bearing on what we take to be the prevalence of poverty in a particular society. Thus, the study of poverty cannot really be seen to be a matter of studying inequality only.

Yet there is a close connection between evaluating poverty and assessing inequality (including inequality among the poor), and this connection has received much attention recently. In Sen (1976b), there was an attempt to integrate the two sets of concerns (i.e., poverty and inequality), and this line of investigation has been very extensively explored in the subsequent literature.72 We shall present some of the main issues that have emerged. Most of the work has been conducted in the context of a unidimensional indicator of individual income, with poverty being seen as inadequately low income; this may be called ‘income poverty’. This emphasis is in line with the view of advantage and deprivation investigated in OEI‐1973, and its reflection in the distribution‐sensitive measure of poverty proposed in Sen (1976b). Recently this view of poverty has been seriously questioned by the argument (p.165) that the exclusive reliance on ‘income poverty’ can hide crucial aspects of economic deprivation (in Sen 1980, 1983, and other works). An alternative view will be discussed in section A.7, after considering the more general issue of the relevance of ‘space’ in judging inequality as well as poverty. The present section focuses on ‘income poverty’ only.

Evaluation of poverty can be broadly divided into two steps:

  1. (1) identification: we have to identify the poor among the total population in the community;

  2. (2) aggregation: the diverse characteristics of the poor would have to be put together to arrive at an assessment of the level of aggregate poverty in that community.

In the context of ‘income poverty’, the identification exercise is primarily one of choosing a ‘poverty line income’ below which people are counted as poor. The aggregation exercise would consist, in this case, of choosing a way of ranking communities with different vectors of individual incomes, and—more ambitiously—of choosing a functional form that maps different income vectors (and poverty lines) into a numerical index of aggregate poverty.

It is the latter exercise—aggregation—that relates most immediately to issues of inequality evaluation, and this will occupy much of the discussion. But distributional concerns can be important even in the determination of the poverty line income. The identification of the level of income at which people can be cogently described as poor may well depend on the pattern of affluence and deprivation that others experience, and this can be affected by both the mean income and the actual distribution around the mean. Indeed, a ‘relativist’ view of income poverty can take us forcefully in the direction of making the poverty line responsive to the distribution of incomes as well as the mean income (e.g., the poverty line may be fixed at half the median income level of that community).73 The sensitivity of the poverty standard to the pattern of (p.166) distribution relates inter alia to taking a ‘relative’ view of poverty. The extent to which a person falls behind other individuals can be checked only with the detailed pattern of distribution—not just the mean income. The issue depends, thus, on the merit of seeing poverty in mainly relative rather than absolute terms, and this issue will reappear in the discussion that follows (and again in section A.7).

There is another question of some practical importance in the choice of the ‘poverty line’. Is the choice to be viewed primarily as a ‘descriptive’ exercise (e.g., what is the level of income below which a person would be regarded as ‘poor’ and seriously deprived in a given society?), or mainly as a ‘prescriptive’ exercise (e.g., what is the level of income below which no one should be allowed to fall in that society?)? While the two types of questions are interconnected (since the prevention of significant deprivation can plausibly be regarded as one of the ethically important objectives of the society and the state), they are not identical questions and need not obtain the same answer (on this see Sen 1979a, 1981).

One reason for a difference between the two interpretations of poverty (aside from the somewhat ‘foundational’ distinction that a description need not by itself induce a prescription) is that the overall ethical objectives of a society can include concerns other than the elimination of economic deprivation. For example, no decision to ‘supplement’ the incomes of all the people below a descriptively relevant line of deprivation (or ‘poverty line’) follows automatically from the fixing of that poverty line, which merely recognizes that some people are poor. Even if a country did not have the means to supplement the incomes of the people thus identified as poor, that would not wash away their ‘poverty’.74

(p.167) While descriptive poverty lines (‘who are the poor?’) have been officially specified—and have been periodically revised—in a number of countries (this includes the United States, even though the actual lines chosen there have been subjected to much criticism), in other countries (such as the United Kingdom or Italy) there is no specified descriptive line at all—only an identification of the level of income below which a person has the legal right to receive assistance from the state (‘who are eligible to receive assistance?’).75 Equating the two can lead to some confounding of distinct issues. For example, a country's ability to pay for income support and the competing demands on scarce resources may radically restrict the number of people who can be assisted, even when some of the people not assisted are recognized as being seriously deprived and poor.

There are, thus, strong arguments for distinguishing between (1) the diagnostic poverty line and (2) the immediately imperative income‐support line. The latter exercise is, of course, clearly ethical and value based, but even the former—mainly descriptive—subject cannot be seen as being ‘value free’. Indeed, valuation does come into deciding what is to count (or not count) as ‘serious deprivation’ or ‘poverty’. However, in so far as the former exercise takes the form of recording the values that happen to be prevailing in a particular community at a given time, the principal task for the investigator is one of description of what values are actually and widely held. This is an old subject, which has been much discussed, and need not be pursued further here.76 But the nature of social understanding and public discussion can be (p.168) helped by giving a specific role to the description of poverty (in terms of contemporary standards), even when the prescriptive links with remedying may not be immediate.77

If the specification of a poverty line has an inescapable dependence on values, the same applies to the exercise of aggregation in ‘putting together’ the diverse information on the deprivation of different people in a community, into an aggregate index of overall poverty in that community. The aggregation exercise, which has received much analytical attention in the literature on poverty in recent decades, involves the use of various competing value systems implicitly present in distinct formulae of composition. The axioms and properties of different aggregation procedures reflect values of various kinds which may be worth scrutinizing explicitly.


(72) The literature on distribution‐sensitive measures of poverty is now quite large; see the critical assessments (in addition to new results) presented in Foster (1984), Kakwani (1984a), Seidl (1988), Atkinson (1987), Ravallion (1994), and Zheng (1996).

(73) On this and related issues, see the classic papers of Victor Fuchs (1965, 1976).

(74) The distinction was extensively discussed in Sen (1979a, 1981). It may be true that even in most cases of serious famine, the victims can be helped to survive within the aggregate means of the poorest of economies through sensible economic policy (on this see Sen 1981 and Drèze and Sen 1989), but there is still an important conceptual distinction between (1) the recognition of economic deprivation and (2) the political and economic feasibility of eliminating it.

(75) On this issue, see Atkinson (1996).

(76) The distinction has been important in the classical writings on ‘subsistence’ and ‘necessities’ (see, for example, Smith 1776 and Marx 1887) on what counts as ‘necessity’ in a particular country at a given moment. As Marx (1887) put it, discussing the notion of ‘subsistence’, the concept of ‘the so‐called necessary wants’ have ‘a historical and moral element’, but ‘nevertheless in a given country, at a given period, the average quantity of the means of subsistence is practically known’ (p. 150). The relation between values and description, and in particular the need to see ‘description as choice’, are discussed in Sen (1982a, Chs. 19 and 20).

(77) Recently, Atkinson (1996) has argued persuasively for the need to specify an ‘official poverty line’ in the UK separately from the requirements of ‘income support’. This is part of his argument for having regular reporting on the state of poverty in the UK, in ways comparable to the regular ‘inflation report’ issued by the Bank of England.