Jump to ContentJump to Main Navigation
Poverty and Famines$

Amartya Sen

Print publication date: 1983

Print ISBN-13: 9780198284635

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198284632.001.0001

Concepts of Poverty

Chapter:
(p.9) Chapter 2 Concepts of Poverty
Source:
Poverty and Famines
Author(s):

Amartya Sen (Contributor Webpage)

Publisher:
Oxford University Press
DOI:10.1093/0198284632.003.0002

Abstract and Keywords

Problems in the conceptualization and measurement of poverty are discussed. Two requirements are identified as (1) a method of identifying a group of people as poor (identification), and (2) a method of aggregating the characteristics of the set of poor people into an overall image of poverty (aggregation). As a foundation for these exercises, a study is made of the kinds of approaches that can be used. These include the biological (minimum nutritional requirement) and inequality approaches to poverty, the concept of relative deprivation, value judgement, policy definition, common standards for comparisons between communities, and the relative scaling of deprivation as a means of aggregation.

Keywords:   aggregation, common standards, conceptualization, definition, deprivation, identification, inequality, methodology, nutritional requirement, poverty

2.1 Requirements of a Concept of Poverty

On his deathbed in Calcutta, J. B. S. Haldane wrote a poem called ‘Cancer's a funny thing’. 1 Poverty is no less funny. Consider the following view of poverty:

People must not be allowed to become so poor that they offend or are hurtful to society. It is not so much the misery and plight of the poor but the discomfort and cost to the community which is crucial to this view of poverty. We have a problem of poverty to the extent that low income creates problems for those who are not poor. 2

To live in poverty may be sad, but to ‘offend or [be] hurtful to society’, creating ‘problems for those who are not poor’ is, it would appear, the real tragedy. It isn't easy to push much further the reduction of human beings into ‘means’.

The first requirement of the concept of poverty is of a criterion as to who should be the focus of our concern. The specification of certain ‘consumption norms’, or of a ‘poverty line’, may do part of the job: ‘the poor’ are those people whose consumption standards fall short of the norms, or whose incomes lie below that line. But this leads to a further question: is the concept of poverty to be related to the interests of: (1) only the poor, (2) only the non‐poor, or (3) both the poor and the non‐poor?

It seems a bit grotesque to hold that the concept of poverty should be concerned only with the non‐poor, and I take the liberty of dropping (2)—and the ‘view’ quoted in the first paragraph—without further ado. Alternative (3) might, however, appear to be appealing, since it is broad‐based and unrestrictive. There is little doubt that the penury of the poor does, in fact, affect the well‐being of the rich. The real question is whether such effects should enter into the concept of poverty as (p.10) such, or whether they should figure under the possible effects of poverty. I believe a good case can be made for choosing the latter alternative, since in an obvious sense poverty must be a characteristic of the poor rather than of the non‐poor. One can, for instance, argue that, if one considers a case of reduction of real income and increase in the suffering of all the poor, it must be described as an increase of poverty, no matter whether this change is accompanied by a reduction in the adverse effects on the rich (e.g. whether the rich are less ‘offended’ by the sight of penury).

This conception of poverty based on (1) does not, of course, imply any denial of the fact that the suffering of the poor themselves may depend on the condition of the non‐poor. It merely asserts that the focus of the concept of poverty has to be on the well‐being of the poor as such, no matter what influences affect their well‐being. Causation of poverty and effects of poverty will be important issues to study on their own rights, and the conceptualization of poverty in terms of the conditions only of the poor does not affect the worthwhileness of studying these questions. Indeed, there will be much to say on these questions later on in the book.

It is perhaps worth mentioning in this context that in some discussions one is concerned not with the prevalence of poverty in a country in the form of the suffering of the poor, but with the relative opulence of the nation as a whole. 3 In those discussions it will, of course, be entirely legitimate to be concerned with the well‐being of all the people in the nation, and the description of a nation as ‘poor’ must obviously relate to such a broader concept. These are different exercises, and so long as this fact is clearly recognized there need not be any confusion.

Even after we have identified the poor and specified that the concept of poverty is concerned with the conditions of the poor, much remains to be done. There is the problem of aggregation—often important—over the group of the poor, and this involves moving from the description of the poor to some over‐all measure of ‘poverty’ as such. In some traditions, this is done very simply by just counting the number of the poor, and then expressing poverty as the ratio of the number of the poor to the total number of people in the community in question.

(p.11) This ‘head‐count measure’—H for short—has at least two serious drawbacks. First, H takes no account of the extent of the short‐fall of incomes of the poor from the ‘poverty line’: a reduction in the incomes of all the poor without affecting the incomes of the rich will leave this head count measure completely unchanged. Second, it is insensitive to the distribution of income among the poor; in particular, no transfer of income from a poor person to one who is richer can increase this head count measure. Both these defects make the measure H, which is by far the most widely used measure, quite unacceptable as an indicator of poverty, and the conception of poverty that lies implicit in it seems eminently questionable.

In this chapter I am not concerned with problems of measurement as such, which will be taken up in the next two chapters and in Appendix C . But behind each measure lies an analytical concept, and here I am concerned with the general ideas on the conception of poverty. If the preceding argument is right, then the requirements of a concept of poverty must include two distinct—but not unrelated—exercises, namely (1) a method of identifying a group of people as poor (‘identification’); and (2) a method of aggregating the characteristics of the set of poor people into an over‐all image of poverty (‘aggregation’). Both these exercises will be performed in the next two chapters, but before that we need to study the kinds of considerations that may be used in choosing the operations (both identification and aggregation). The rest of the chapter will be concerned with these issues.

The underlying considerations come out most sharply in the alternative approaches to the concept of poverty that one can find in the literature. Some of these approaches have been subjected to severe attacks recently, while others have not been examined sufficiently critically. In attempting an evaluation of these approaches in the following sections, I shall try to assess the approaches as well as their respective critiques.

2.2 The Biological Approach

In his famous study of poverty in York, Seebohm Rowntree ( 1901 ) defined families as being in ‘primary poverty’ if their ‘total earnings are insufficient to obtain the minimum necessities for the maintenance of merely physical efficiency’. It is not surprising that biological considerations related to the requirements of (p.12) survival or work efficiency have been often used in defining the poverty line. Starvation, clearly, is the most telling aspect of poverty.

The biological approach has come under rather intense fire recently. 4 There are indeed several problems with its use. First, there are significant variations related to physical features, climatic conditions and work habits. 5 In fact, even for a specific group in a specific region, nutritional requirements are difficult to define precisely. People have been known to survive with incredibly little nutrition, and there seems to be a cumulative improvement of life expectation as the dietary limits are raised. In fact, physical opulence seems to go on increasing with nutrition over a very wide range; Americans, Europeans and Japanese have been growing measurably in stature as their diets have continued to improve. There is difficulty in drawing a line somewhere, and the so‐called ‘minimum nutritional requirements’ have an inherent arbitrariness that goes well beyond variations between groups and regions.

Second, the translation of minimum nutritional requirements into minimum food requirements depends on the choice of commodities. While it may be easy to solve the programming exercise of a ‘diet problem’, choosing a minimum cost diet for meeting specified nutritional requirements from food items sold at specified costs, the relevance of such a minimum cost diet is not clear. Typically, it turns out to be very low‐cost indeed, 6 but monumentally boring, and people's food habits are not, in fact, determined by such a cost minimization exercise. The actual incomes at which specified nutritional requirements are met will depend greatly on the consumption habits of the people in question.

Third, for non‐food items such minimum requirements are not easy to specify, and the problem is usually solved by assuming that a specified proportion of total income will be spent on food. With this assumption, the minimum food costs can be used to derive minimum income requirements. But the proportion spent on food varies not merely with habits and culture, but also with relative prices and availability of goods and services. It is not (p.13) surprising that the assumptions made often turn out to be contradicted by actual experience; for example, Lord Beveridge's estimate of subsistence requirements of income during the Second World War proved to be far from correct, since the British were spending a much lower proportion of their income on food than was assumed (see Townsend, 1974 , p. 17).

In view of these problems, one may well agree with Martin Rein's ( 1971 ) assertion that ‘almost every procedure in the subsistence‐level definition of poverty can be reasonably challenged’ (p. 61). But the question that does remain is this: after we have challenged every one of the procedures used under the biological approach, what do we do then? Do we simply ignore that approach, 7 or do we examine whether something remains in it to be salvaged? I would argue that there does remain something.

First, while the concept of nutritional requirements is a rather loose one, there is no particular reason to suppose that the concept of poverty must itself be clear‐cut and sharp. In fact, a certain amount of vagueness is implicit in both the concepts, and the really interesting question is the extent to which the areas of vagueness of the two notions, as commonly interpreted, tend to coincide. The issue, thus, is not whether nutritional standards are vague, but whether the vagueness is of the required kind.

Second, to check whether someone is getting a specified bundle of nutrition, one need not necessarily go through the procedure of examining whether that person has the income level that would generate that bundle. One can simply examine whether the person is, in fact, meeting that nutritional requirement or not. Even in poor countries, direct nutritional information of this type can be collected through sample surveys of consumption bundles and can be extensively analysed (see, for example, Srinivasan and Bardhan, 1974 , especially the paper by Chatterjee, Sarkar and Paul, and Panikar et al., 1975 ); and the ‘identification’ exercise under the nutritional approach need not go through the intermediary of income at all.

(p.14) Third, even when we do go through the intermediary of income, the translation of a set of nutritional norms (or of alternative sets of such norms) into a ‘poverty line’ income (or poverty‐line incomes) may be substantially simplified by the wide prevalence of particular patterns of consumption behaviour in the community in question. Proximity of actual habits and behaviour makes it possible to derive income levels at which the nutritional norms will be ‘typically’ met. (This question is discussed further in Chapter 3 .)

Finally, while it can hardly be denied that malnutrition captures only one aspect of our idea of poverty, it is an important aspect, and one that is particularly important for many developing countries. It seems clear that malnutrition must have a central place in the conception of poverty. How exactly this place is to be specified remains to be explored, but the recent tendency to dismiss the whole approach seems to be a robust example of misplaced sophistication.

2.3 The Inequality Approach

The idea that the concept of poverty is essentially one of inequality has some immediate plausibility. After all, transfers from the rich to the poor can make a substantial dent on poverty in most societies. Even the poverty line to be used for identifying the poor has to be drawn with respect to contemporary standards in the community in question, so that poverty may look very like inequality between the poorest group and the rest of the community.

Arguments in favour of viewing poverty as inequality are presented powerfully by Miller and Roby, who conclude:

Casting the issues of poverty in terms of stratification leads to regarding poverty as an issue of inequality. In this approach, we move away from efforts to measure poverty lines with pseudo‐scientific accuracy. Instead, we look at the nature and size of the differences between the bottom 20 or 10 per cent and the rest of the society. Our concern becomes one of narrowing the differences between those at the bottom and the better‐off in each stratification dimension. 8

There is clearly quite a bit to be said in favour of this approach. But one can argue that inequality is fundamentally a different (p.15) issue from poverty. To try to analyse poverty ‘as an issue of inequality’, or the other way round, would do little justice to either. Inequality and poverty are not, of course, unrelated. But neither concept subsumes the other. A transfer of income from a person in the top income group to one in the middle income range must ceteris paribus reduce inequality; but it may leave the perception of poverty quite unaffected. Similarly, a general decline in income that keeps the chosen measure of inequality unchanged may, in fact, lead to a sharp increase in starvation, malnutrition and obvious hardship; it will then be fantastic to claim that poverty is unchanged. To ignore such information as starvation and hunger is not, in fact, an abstinence from ‘pseudo‐scientific accuracy’, but blindness to important parameters of the common understanding of poverty. Neither poverty nor inequality can really be included in the empire of the other. 9

It is, of course, quite a different matter to recognize that inequality and poverty are associated with each other, and to note that a different distribution system may cure poverty even without an expansion of the country's productive capabilities. Recognizing the distinct nature of poverty as a concept permits one to treat it as a matter of interest and involvement in itself. The role of inequality in the prevalence of poverty can then figure in the analysis of poverty without making the two conceptually equivalent.

2.4 Relative Deprivation

The concept of ‘relative deprivation’ has been fruitfully used in the analysis of poverty, 10 especially in the sociological literature. Being poor has clearly much to do with being deprived, and it is natural that, for a social animal, the concept of deprivation will be a relative one. But within the uniformity of the term ‘relative deprivation’, there seem to exist some distinct and different notions.

One distinction concerns the contrast between ‘feelings of (p.16) deprivation’ and ‘conditions of deprivation’. Peter Townsend has argued that ‘the latter would be a better usage’. 11 There is indeed much to be said for a set of criteria that can be based on concrete conditions, so that one could use ‘relative deprivation’ ‘in an objective sense to describe situations where people possess less of some desired attribute, be it income, favourable employment conditions or power, than do others’. 12

On the other hand, the choice of ‘conditions of deprivation’ can not be independent of ‘feelings of deprivation’. Material objects cannot be evaluated in this context without reference to how people view them, and even if ‘feelings’ are not brought in explicitly, they must have an implicit role in the selection of ‘attributes’. Townsend has rightly emphasized the importance of the ‘endeavour to define the style of living which is generally shared or approved in each society and find whether there is . . . a point in the scale of the distribution of resources below which families find it increasingly difficult . . . to share in the customs, activities and diets comprising that style of living’. 13 One must, however, look also at the feelings of deprivation in deciding on the style and level of living the failure to share which is regarded as important. The dissociation of ‘conditions’ from ‘feelings’ is, therefore, not easy, and an objective diagnosis of ‘conditions’ requires an objective understanding of ‘feelings’.

A second contrast concerns the choice of ‘reference groups’ for comparison. Again, one has to look at the groups with which the people in question actually compare themselves, and this can be one of the most difficult aspects of the study of poverty based on relative deprivation. The horizon of comparison is not, of course, independent of political activity in the community in question, 14 since one's sense of deprivation is closely related to one's expectations as well as one's view of what is fair and who has the right to enjoy what.

These different issues related to the general notion of relative deprivation have considerable bearing on the social analysis of (p.17) poverty. It is, however, worth noting that the approach of relative deprivation—even including all its variants—cannot really be the only basis for the concept of poverty. A famine, for example, will be readily accepted as a case of acute poverty no matter what the relative pattern within the society happens to be. Indeed, there is an irreducible core of absolute deprivation in our idea of poverty, which translates reports of starvation, malnutrition and visible hardship into a diagnosis of poverty without having to ascertain first the relative picture. Thus the approach of relative deprivation supplements rather than supplants the analysis of poverty in terms of absolute dispossession.

2.5 A Value Judgement?

The view that ‘poverty is a value judgement’ has recently been presented forcefully by many authors. It seems natural to think of poverty as something that is disapproved of, the elimination of which is regarded as morally good. Going further, it has been argued by Mollie Orshansky, an outstanding authority in the field, that ‘poverty, like beauty, lies in the eye of the beholder’. 15 The exercise would, then, seem to be primarily a subjective one: unleashing one's personal morals on the statistics of deprivation.

I would like to argue against this approach. It is important to distinguish between different ways in which the role of morals can be accommodated into the exercise of poverty measurement. There is a difference between saying that the exercise is itself a prescriptive one and saying that the exercise must take note of the prescriptions made by members of the community. To describe a prevailing prescription is an act of description, not prescription. It may indeed be the case that poverty, as Eric Hobsbawm ( 1968 ) puts it, ‘is always defined according to the conventions of the society in which it occurs’ (p. 398). But this does not make the exercise of poverty assessment in a given society a value judgement. Nor a subjective exercise of some kind or other. For the person studying and measuring poverty, the conventions of society are matters of fact (what are the contemporary standards?), and not issues of morality or of subjective search (what should be the contemporary standards? what should be my values? how do I feel about all this?). 16

(p.18) The point was brought out very clearly by Adam Smith more than two hundred years ago:

By necessaries I understand not only the commodities which are indispensably necessary for the support of life, but what ever the custom of the country renders it indecent for creditable people, even the lowest order, to be without. A linen shirt, for example, is, strictly speaking, not a necessary of life. The Greeks and Romans lived, I suppose, very comfortably though they had no linen. But in the present times, through the greater part of Europe, a creditable day‐labourer would be ashamed to appear in public without a linen shirt, the want of which would be supposed to denote that disgraceful degree of poverty which, it is presumed, nobody can well fall into without extreme bad conduct. Custom, in the same manner, has rendered leather shoes a necessary of life in England. The poorest creditable person of either sex would be ashamed to appear in public without them. 17

In a similar vein Karl Marx ( 1867 ) argued that, while ‘a historical and moral element’ enters the concept of subsistence, ‘nevertheless, in a given country, at a given period, the average quantity of the means of subsistence necessary for the labourer is practically known’ (p. 150).

It is possible that Smith or Marx may have overestimated the extent of uniformity of views that tends to exist in a community on the content of ‘subsistence’ or ‘poverty’. Description of ‘necessities’ may be very far from ambiguous. But the presence of ambiguity in a description does not make it a prescriptive act—only one of ambiguous description. One may be forced to be arbitrary in eliminating the ambiguity, and if so that arbitrariness would be worth recording. Similarly, one may be forced to use more than one criteria because of non‐uniformity of accepted standards, and to look at the partial ordering generated by the criteria taken together (reflecting ‘dominance’ in terms of all the criteria). 18 But the partial ordering would still reflect a descriptive (p.19) statement rather than a prescriptive one. Indeed, the statement would be rather like saying, ‘Nureyev may or may not be a better dancer than Nijinsky, but he dances better than this author, according to contemporary standards’, a descriptive statement (and sadly non‐controversial).

2.6 A Policy Definition?

A related issue is worth exploring in this context. The measurement of poverty may be based on certain given standards, but what kind of statements do these standards themselves make? Are they standards of public policy, reflecting either the objectives of actual policy or views on what the policy should be? There is little doubt that the standards must have a good deal to do with some broad notions of acceptability, but that is not the same thing as reflecting precise policy objectives—actual or recommended. On this subject too a certain amount of confusion seems to exist. For example, the United States President's Commission on Income Maintenance ( 1969 ) argued thus for such a ‘policy definition’ in its well‐known report, Poverty amid Plenty:

If society believes that people should not be permitted to die of starvation or exposure, then it will define poverty as the lack of minimum food and shelter necessary to maintain life. If society feels some responsibility for providing to all persons an established measure of well‐being beyond mere existence, for example, good physical health, then it will add to its list of necessities the resources required to prevent or cure sickness. At any given time a policy definition reflects a balancing of community capabilities and desires. In low income societies the community finds it impossible to worry much beyond physical survival. Other societies, more able to support their dependent citizens, begin to consider the effects that pauperism will have on the poor and non‐poor alike. 19

There are at least two difficulties with this ‘policy definition’. First, practical policy‐making depends on a number of influences, going beyond the prevalent notions of what should be done. Policy is a function of political organization, and depends on a variety of factors including the nature of the government, the sources of its power, and the forces exerted by other organizations. In the public policies pursued in many countries, it is, in fact, hard to detect a concern with the elimination of deprivation in any obvious sense. If interpreted in terms of actual (p.20) policy, the ‘policy definition’ may fail to catch the political issues in policy‐making.

Second, even if ‘policy’ is taken to stand not for actual public policy, but for policy recommendations widely held in the society in question, there are problems. There is clearly a difference between the notion of ‘deprivation’ and the idea of what should be eliminated by ‘policy’. For one thing, policy recommendations must depend on an assessment of feasibilities (‘ought implies can’ 20 ), but to concede that some deprivations cannot be immediately eliminated is not the same thing as conceding that they must not currently be seen as deprivations. (Contrast: ‘Look here, old man, you aren't really poor even though you are starving, since it is impossible in the present economic circumstances to maintain the income of everyone above the level needed to eliminate starvation.’) Adam Smith's notion of subsistence based on ‘the commodities which are indispensably necessary for the support of life’ and ‘what ever the custom of the country renders it indecent’ for someone ‘to be without’ is by no means identical with what is generally accepted as could and should be provided to all as a matter of policy. If in a country suddenly impoverished, say, by war it is agreed generally that the income maintenance programme must be cut down to a lower level of income, would it be right to say that the country does not have any greater poverty since a reduction of incomes has been matched by a reduction of the poverty line?

I would submit that the ‘policy definition’ is based on a fundamental confusion. It is certainly true that with economic development there are changes in the notion of what counts as deprivation and poverty, and there are changes also in the ideas as to what should be done. But while these two types of changes are interdependent and also intertemporally correlated with each other, neither can be defined entirely in terms of the other. Oil‐rich Kuwait may be ‘more able to support their dependent citizens’ with its new prosperity, but the notion of what is poverty may not go up immediately to the corresponding level. Similarly, the war‐devastated Netherlands may keep up its standard of what counts as poverty and not scale it down to the level commensurate with its predicament. 21

(p.21) If this approach is accepted, then the measurement of poverty must be seen as an exercise of description assessing the predicament of people in terms of the prevailing standards of necessities. It is primarily a factual rather than an ethical exercise, and the facts relate to what is regarded as deprivation, and not directly to what policies are recommended. The deprivation in question has both absolute and relative aspects (as argued in Sections 2.2 and 2.4 above).

2.7 Standards and Aggregation

This still leaves two issues quite untouched. First, in comparing the poverty of two societies, how can a common standard of necessities be found, since such standards would vary from society to society? There are actually two quite distinct types of exercises in such inter‐community comparisons. One is aimed at comparing the extent of deprivation in each community in relation to their respective standards of minimum necessities, and the other is concerned with comparing the predicament of the two communities in terms of some given minimum standard, e.g. that prevalent in one community. There is, indeed, nothing contradictory in asserting both of the following pair of statements:

  1. (1) There is less deprivation in community A than in community B in terms of some common standard, e.g. the notions of minimum needs prevailing in community A.

  2. (2) There is more deprivation in community A than in community B in terms of their respective standards of minimum needs, which are a good deal higher in A than in B. 22

It is rather pointless to dispute which of these two senses is the ‘correct’ one, since it is quite clear that both types of questions are of interest. The important thing to note is that the two questions are quite distinct from each other.

Second, while the exercise of ‘identification’ of the poor can be based on a standard of minimum needs, that of ‘aggregation’ requires some method of combining deprivations of different people into some overall indicator. In the latter exercise some relative scaling of deprivations is necessary. The scope for (p.22) arbitrariness in this is much greater, since conventions on this are less firmly established and the constraints of acceptability would tend to leave one with a good deal of freedom. The problem is somewhat comparable with the criteria for making aggregative descriptive statements in such fields as, say, comparisons of sporting achievements of different groups. While it is clear that certain circumstances would permit one to make an aggregative statement like ‘Africans are better at sprint than Indians’ (e.g. the circumstance in which the former group keeps winning virtually all sprint events over the Indians), and other circumstances would force one to deny this, there are intermediate cases in which either of the two aggregative descriptive statements would be clearly disputable.

In this context of arbitrariness of ‘aggregate description’, it becomes particularly tempting to redefine the problem as an ‘ethical’ exercise, as has indeed been done in the measurement of economic inequality. 23 But the ethical exercises involve exactly similar ambiguities, and furthermore end up answering a different question from the descriptive one that was originally asked. 24 There is very little alternative to accepting the element of arbitrariness in the description of poverty, and making that element as clear as possible. Since the notion of the poverty of a nation has some inherent ambiguities, one should not have expected anything else.

2.8 Concluding Remarks

Poverty is, of course, a matter of deprivation. The recent shift in focus—especially in the sociological literature—from absolute to relative deprivation has provided a useful framework of analysis (Section 2.4 ). But relative deprivation is essentially incomplete as an approach to poverty, and supplements (but cannot supplant) the earlier approach of absolute dispossession. The much maligned biological approach, which deserves substantial reformulation but not rejection, relates to this irreducible core of absolute deprivation, keeping issues of starvation and hunger at the centre of the concept of poverty (Sections 2.2 and 2.4 ).

To view poverty as an issue in inequality, as is often recommended, seems to do little justice to either concept. (p.23) Poverty and inequality relate closely to each other, but they are distinct concepts and neither subsumes the other (Section 2.3 ).

There is a good case for viewing the measurement of poverty not, as is often asserted, as an ethical exercise, but primarily as a descriptive one (Section 2.5 ). Furthermore, it can be argued that the frequently used ‘policy definition’ of poverty is fundamentally flawed (Section 2.6 ). The exercise of describing the predicament of the poor in terms of the prevailing standards of ‘necessities’ does, of course, involve ambiguities, which are inherent in the concept of poverty; but ambiguous description isn't the same thing as prescription. 25 Instead, the arbitrariness that is inescapable in choosing between permissible procedures and possible interpretations of prevailing standards requires recognition and appropriate treatment.

Notes:

(1) Oxford Book of 20th Century English Verse, ed. P. Larkin, Oxford, 1973, p. 271.

(2) Rein ( 1971 ), p. 46. I hasten to add that here Professor Rein is describing one of the three ‘broad concepts’ of poverty, viz. (1) ‘subsistence’, (2) ‘inequality’, and (3) ‘externality’; the view quoted corresponds to ‘externality’.

(3) See, for example, Paul Streeten, ‘How Poor Are the Poor Countries and Why?’ in

(4) See, for example, Townsend ( 1971 , 1974 ) and Rein ( 1971 ).

(5) See Rein ( 1971 ), Townsend ( 1974 ), Sukhatme ( 1977 , 1978 ), and Srinivasan ( 1977a , 1979 ).

(6) See, for example, Stigler's ( 1945 ) astonishing estimates of ‘the cost of subsistence’. See also Rajaraman ( 1974 ).

(7) Much depends on what the alternatives are. Rein ( 1971 ) himself recommends that ‘other’ conceptions ‘deserve more attention and developments’ (p. 62). Since ‘subsistence’ is one of his three ‘broad concepts’ of poverty, we are left with ‘externality’ and ‘inequality’. Inequality—though related to poverty in terms of both causation and evaluation—is, however, a distinct issue from poverty, as will be presently argued (see Section 2.3 ). ‘Externality’, in terms of the effects of poverty on the non‐poor, is an approach that we have already discussed (in Section 2.1 ), critically.

(8) Miller and Roby ( 1971 , p. 143). Also Miller, Rein, Roby and Cross ( 1967 ). See Wedderburn ( 1974 ) for discussions of alternative approaches.

(9) It is also worth noting that there are many measures of inequality, of which the gap ‘between the bottom 20 or 10 per cent and the rest’ is only one. See Atkinson ( 1970 ), Sen ( 1973a ), Kolm ( 1976a , 1976b ), and Blackorby and Donaldson ( 1978 , 1980b ). Also, inequality is not just a matter of the size distribution of income but one of investigating contrasts between different sections of the community from many different perspectives, e.g. in terms of relations of production, as done by Marx ( 1859 , 1867 ).

(10) See Runciman ( 1966 ) and Townsend ( 1971 ), presenting two rather different approaches to the concept.

(11) Townsend ( 1974 ), pp. 25–6.

(12) Wedderburn ( 1974 ), p. 4.

(13) Townsend ( 1974 ), p. 36.

(14) For example, Richard Scase ( 1974 ) notes that Swedish workers tend to choose rather wider reference groups than British workers, and relates this contrast to the differences in the nature of the two trade union movements and of political organization generally.

(15) Orshansky ( 1969 ), p. 37. For a critique of this position, see Townsend ( 1974 ).

(16) This does not, of course, in any way deny that one's values may implicitly affect one's assessment of facts, as indeed they very often do. The statement is about the nature of the exercise, viz. that it is concerned with assessment of facts, and not about the way it is typically performed and the psychology that lies behind that performance. (The doctor attached to the students' hostel in which I stayed in Calcutta would refuse to diagnose influenza on the powerful ground that ‘flu shouldn't be a reason for staying in bed’.) The issue is, in some respects, comparable to that of one's interests influencing one's values; for an important historical analysis of various different aspects of that relationship, see Hirschman ( 1977 ).

(17) Smith ( 1776 ), pp. 351–2.

(18) Sen ( 1973a ), Chapters 2 and 3 .

(19) US President's Commission on Income Maintenance ( 1969 ), p. 8.

(20) Cf. Hare ( 1963 ), Chapter 4 .

(21) For an account of that predicament, see Stein, Susser, Saenger, and Marolla ( 1975 ).

(22) There is also no necessary contradiction in asserting that community A has less deprivation in terms of one community's standards (e.g. A's itself), while community B is less deprived in terms of another community's standards (e.g. B's).

(23) See Dalton ( 1920 ), Kolm ( 1969 ), and Atkinson ( 1970 ).

(24) See Bentzel ( 1970 ), Hansson ( 1977 ), and Sen ( 1978b ).

(25) The underlying methodological issues have been discussed in Sen ( 1980a ).