What is poverty? - Sakiko Fukuda-Parrsakikofukudaparr.net/wp-content/uploads/2013/01/Poverty... · 2013. 2. 7. · Lord Meghnad Desai finds the definitions of absolute poverty static,

PovertyInternational Poverty Centre

United Nations Development Programme December 2006

What is poverty?Concepts and measures

2 United Nations Development Programme

T he international development community has had poverty in focus for more thana decade. At summit meetings and other occasions, world leaders have stated and

reconfirmed their agreement that poverty must be reduced and eventually eradicated.

The political commitment is a necessary, but not sufficient, condition for this tohappen. Analysts, policy-makers and practitioners need appropriate concepts anddedicated measures to enable progress from rhetoric and general policy statementsto action and results on the ground.

In this issue of IPC’s journal Poverty in Focus we present ten articles intended to throwlight on the question of how best to define and measure poverty.

Robert Chambers outlines five clusters of meanings and reminds us of the importance ofthe analysis and views of poor people themselves and their many meanings. When theyget to express their views, we get a case for changing language, concepts andmeasures in development. The key issue is whose reality counts – theirs or ours?

Peter Townsend provides an historical perspective of the poverty concept and the settingof poverty lines. Three poverty concepts have evolved, based on ideas of subsistence,basic needs and relative deprivation. Since material needs are socially determined, weneed a new international poverty line based on what is required in different countriesto surmount material and social deprivation.

Sakiko Fukuda-Parr describes the multidimensional poverty measures developed byUNDP’s Human Development Reports since 1990, especially the Human Poverty Index(HPI). It shows a large spread of human poverty among countries with similar levels ofincome poverty and thus, HPI is only weakly correlated with income poverty. Recent HPItrends are also presented and discussed.

Caterina Ruggeri Laderchi, Ruhi Saith and Frances Stewart analyse empirical evidence to seeif and why the definition of poverty matters. They also report on field testing in twodeveloping countries of four different approaches. These are shown to have differentimplications for policy and also for targeting, since they identify different causes andeffects of poverty, and different people as being poor.

Gustav Ranis, Frances Stewart and Emma Samman review the various listings of humanwellbeing and poverty elements, thus identifying a comprehensive set of dimensionsin order to empirically explore whether UNDP’s Human Development Index is adequateor needs to be supplemented. They show that assessing human development fullyrequires a broader set of indicators.

Peter Edward outlines a moral concept of absolute poverty and defines an EthicalPoverty Line derived from globally standardised and ethically justifiable wellbeingindicators. Applying it to actual income data shows that world poverty by a moraldefinition is much larger than by current measures, and so is the required globalincome redistribution.

Lord Meghnad Desai finds the definitions of absolute poverty static, calorific, asocialand atheoretical. He proposes a new poverty line to be based on the need tomaintain individual labour capacities intact, thus connecting to health, nutritionand monetary measures.

Ravi Kanbur considers the conundrums of measuring poverty when populations changeand analyses three population size scenarios – increased, decreased and unchanged, butwith churning around the poverty line. He delivers some remarkable points to consider.

Nanak Kakwani proposes a multidimensional poverty concept that is causally linked tocommand over economic resources. He argues for an income poverty line that reflectsthe cost of achieving basic human needs.

Sabine Alkire in response to Kakwani argues that it is not the cause of poverty thatmatters, but what is actionable by public policy. There are many ways to measurecapability deprivation. The debate ends, for now, with a rejoinder by Kakwani.

We wish you an informative and valuable reading of this issue of Poverty in Focus.

F R O M T H EE D I T O R

Dag Ehrenpreis

Poverty in Focus is a regular publication of theUNDP International Poverty Centre (IPC). Itspurpose is to present the results of research onpoverty and inequality in the developing world.

The International Poverty Centre (IPC) is ajoint project between the United NationsDevelopment Programme and Brazil to promoteSouth-South Cooperation on applied povertyresearch. It specialises in analysing poverty andinequality and offering research-based policyrecommendations on how to reduce them. IPCis directly linked to the Poverty Group of theBureau for Development Policy, UNDP and theBrazilian Government’s Institute for AppliedEconomic Research (Ipea).

IPC Director (acting)Terry McKinley

EditorDag Ehrenpreis

International Advisory Board

Desktop PublishersRoberto Astorino and Lucas Moll

Front page: Photo by Fabio Veras, IPC, ofan African cobweb to illustrate both thefragility and the tenacity of the life situationsof people living in or close to conditions ofpoverty. The web metaphor is also applicableto the interlinkages between the variousdimensions of poverty and wellbeing, as in thefigure on page 3.

Editor’s note: The sources of graphs andtables are the reference publications atthe end of each article concerned. Thanks forpermission to use this material to RobertChambers, Institute of Development Studies,Sussex, UK; UNDP Human DevelopmentReport Office, New York; Journal of HumanDevelopment, Routledge; Third WorldQuarterly, Routledge. And thanks not least toall the authors contributing their time andintellectual products without any monetaryremuneration.

United Nations Development ProgrammeInternatinal Poverty CentreSBS – Ed. BNDES, 10º andar70076-900 Brasilia DF Brazil

[email protected]

The content of this publication does notnecessarily reflect the official views of UNDP.

Oscar Altimir, CEPAL, Santiago de ChileGiovanni A. Cornia, Università di FirenzeNora Lustig, UNDP/BDP Poverty GroupGita Sen, Indian Institute of Management, BangaloreAnna Tibaijuka, UN Habitat, NairobiPeter Townsend, London School of EconomicsPhilippe van Parijs, Université de Louvain

International Poverty Centre Poverty In Focus December 2006 3

The flood of development rhetoricon poverty, the primacy accorded bylenders and donors to the MillenniumDevelopment Goals, of which the reductionof extreme poverty is the first and usuallyconsidered the most important, and thefrequency with which reducing, alleviatingor eliminating poverty is seen as a primegoal and measure of development –these factors make it matter more thanever to know what poverty is. What it istaken to mean depends on who asks thequestion, how it is understood, and whoresponds. From this perspective, it has atleast five clusters of meanings.

The first is income-poverty or its commonproxy (because less unreliable to measure)consumption-poverty. This needs noelaboration. When many, especiallyeconomists, use the word poverty theyare referring to these measures. Povertyis what can be and has been measured,and measurement and comparisonsprovide endless scope for debate.

The second cluster of meanings ismaterial lack or want. Besides income, thisincludes lack of or little wealth and lackor low quality of other assets such asshelter, clothing, furniture, personalmeans of transport, radios or television,and so on. This also tends to include noor poor access to services.

A third cluster of meanings derives fromAmartya Sen, and is expressed ascapability deprivation, referring to what wecan or cannot do, can or cannot be. Thisincludes but goes beyond material lackor want to include human capabilities,for example skills and physical abilities,and also self-respect in society.

A fourth cluster takes a yet more broadlymulti-dimensional view of deprivation,with material lack or want as only one ofseveral mutually reinforcing dimensions.

These four clusters of the meanings ofpoverty have all been constructed by“us”, by development professionals.They are expressions of “our” education,training, mindsets, experiences andreflections. They reflect our power, asnon-poor people, to make definitionsaccording to our perceptions. And theprimacy we accord to povertyalleviation, reduction or eliminationimplies that these meanings that wegive are fundamental to whatdevelopment should be about.

One expression of this has twelvedimensions, each one potentially havingan impact on all of the others, andvice versa, thus emphasising theinterdependence of the dimensions ofpoverty as we see them (see figure below).

Another expression of this has fivedimensions to illustrate development as

What poverty is takento mean depends onwho asks the question,how it is understood,and who responds.

Our common meaningshave all been constructedby us, non-poor people.

They reflect our powerto make definitionsaccording toour perceptions.

Whose reality counts?Ours, as we construct itwith our mindsets and forour purposes? Or theirsas we enable themto analyse and express it?

What is Poverty?Who asks? Who answers?

by Robert Chambers,Institute of Development Studies,Sussex, UK


good change. Development thus can beseen as shifting from illbeing to wellbeingwith equity, with interventions toenhance wellbeing possible at any ofthe five points (see figure on the right).

But these dimensions are all abstractions,to varying degrees reductionist, based onour analysis and views. They tend tooverlook and ignore the analysis and viewsof the objects of the definition anddescription – “the poor”, that is people whoare in a bad condition variously describedas poor, marginalised, vulnerable, excludedor deprived. There is then a fifth cluster,which is the multiplicity of their meanings.

These dimensions have been elicitedin many contexts, most extensivelyperhaps in the World Bank’s participatoryresearch programme Voices of the Poor,in which over 20,000 poor women andmen from 23 countries were convened insmall groups and facilitated to analyseand express their realities. Questionshad to be confronted concerning words,translations, languages and concepts.

The word poverty translated into otherlanguages carries different connotations.This was one factor in deciding to seekbetter insights and comparability byinviting the local analysts to use their ownwords and concepts for illbeing or badquality of life, and wellbeing or goodquality of life. Even allowing for the pitfallsof analysing and imposing outsiders’categories on their diverse responses,values and realities, it was striking howcommon and strong the same dimensionswere across cultures and contexts.

There were many poverties or deprivations.Dimensions of the bad life included notonly income-poverty and material lack, butmany others, some of them represented inthe web of poverty’s disadvantages in thefigure, for example poverty of time, livingand working in bad places – “the placesof the poor” and bad social, especiallygender, relations. Others were the body asthe main asset of many poor people,indivisible, uninsured, and vulnerableto flipping from asset to liability; manyaspects of insecurity, worry and anxiety;and pervasively powerlessness.

The many ideas of wellbeing and thegood life to which people aspired had

striking commonalities – materialwellbeing, having enough; bodilywellbeing, being and appearing well;many aspects of social wellbeingincluding being able to settle children,and being able to help others; security;and freedom of choice and action.Both these commonalities and localdifferences make a case for changinglanguage, concepts and measures indevelopment.

The case is for the language of illbeingand wellbeing to be widely used inaddition to poverty and wealth, whichare only one part of them. It is forrepeated participatory processes toenable local people, especially thepoorest, most marginalised and mostvulnerable, to analyse and monitor thequality of their lives, and for this to befed back regularly to policy-makers.It is for policy-makers to spend timeliving in poor communities andappreciating their conditions andrealities firsthand.

If we are seriously pro-poorprofessionals, the answer to “What ispoverty?” is “That is the wrong question.”It is our question, not theirs. Thequestion of those who are poor,marginalised and vulnerable is morelikely to be, in varied forms and manylanguages with different nuances:

“What can you do to reduce our badexperiences of life and living, and toenable us to achieve more of the goodthings in life to which we aspire?”Policies and actions that follow wouldthen be designed to reduce illbeing andenhance wellbeing in their own terms.The MDGs may help, but are far fromenough for this, and may at times evenmisdirect effort.

Direct actions towards their achievementmay often not present the best prioritiesand paths. For they narrow andstandardise vision, leave out muchthat matters, and do not allow for themultifarious ways in which people canbe enabled to enjoy a better life.Policies and actions need to be informedmuch less by top-down targets andmuch more by the diverse bottom-uprealities of the powerless.

The questions are then: whose realitycounts? Ours? Or theirs? Or moreprecisely: ours, as we construct it withour mindsets and for our purposes?Or theirs as we enable them to analyseand express it?

Robert Chambers: “Power, knowledge andpolicy influence: reflections on an experience”in Karen Brock and Rosemary McGee (eds)Knowing Poverty: Critical reflections onparticipatory research and policy, London 2002.


What is Poverty?An historical perspective

by Peter Townsend,London School of Economics

Three poverty conceptshave evolved, basedon ideas of subsistence,basic needs andrelative deprivation.

Material needs aresocially determined.

Globalisation connectspeoples and livingstandards, whileinequality is increasing.

Poor people are deniedthe resources to fulfilsocial demands andobserve customs.

We need better povertymeasures and regularUN monitoring reports.

Historically, poverty has been relatedto income, which remains at the core ofthe concept today. However, “income” isitself no less problematic a concept than“poverty”; it too has to be carefully andprecisely elaborated. Other resources suchas assets, income in kind and subsidies topublic services and employment shouldbe imputed to arrive at a comprehensivebut accurate measure of income.

People can be said to be in poverty whenthey are deprived of income and otherresources needed to obtain the conditionsof life—the diets, material goods,amenities, standards and services—that enable them to play the roles, meetthe obligations and participate in therelationships and customs of their society.

The determination of a poverty line cannotbe based on an arbitrary selection of a lowlevel of income. Only scientific criteriaindependent of income can justify wherethe poverty line should be drawn. Themultiplicity and severity of differenttypes of deprivation can constitute thosecriteria. The key is therefore to define athreshold of income below which peopleare found to be thus deprived.

The measure of multiple deprivationsmust be decided on the basis of evidenceabout each and every sphere of the rangeof social and individual activities peopleperform in fulfilment of individual andfamily needs, and social obligations. Thedegree of material and social deprivationrelative to income is the basis forascertaining the threshold amount ofincome ordinarily required by householdsof different compositions to surmountpoverty. The application of this methodpermits analysis of trends in poverty inand across different countries.

The understanding and relief of povertyhas been a major human preoccupationfor many centuries. Since the 1880s, threealternative conceptions of poverty haveevolved as a basis for international and

comparative work. They dependprincipally on the ideas of subsistence,basic needs and relative deprivation.

The subsistence idea was a result of workprompted by nutritionists in VictorianEngland. Families were defined to be inpoverty when their incomes were not“sufficient to obtain the minimumnecessaries for the maintenance ofmerely physical efficiency”. A family wastreated as being in poverty if its incomeminus rent fell short of the poverty line.Although allowance was made incalculating the income level for clothing,fuel and some other items, this allowancewas very small, and food accounted formuch the greatest share of subsistence.

These ideas have influenced scientificpractice and international and nationalpolicies for over 100 years. Examples arethe statistical measures adopted todescribe social conditions, at first withinindividual countries but later with wideapplication by international agencies suchas the World Bank. The idea of subsistencewas freely exported to member States ofthe former British Empire, e.g. for settingthe wages of blacks in South Africa andframing development plans in India andMalaysia. In the United States, “subsistence”remains the basis of the official measureof poverty.

The use of “subsistence” to definepoverty has been criticized because itimplies that human needs are mainlyphysical rather than also social needs.People are not simply individualorganisms requiring replacement ofsources of physical energy; they aresocial beings expected to performsocially demanding roles as workers,citizens, parents, partners, neighboursand friends. Moreover, they are notsimply consumers of physical goods butproducers of those goods and are alsoexpected to act out different roles intheir various social associations. They aredependent on collectively provided


utilities and facilities. These needs applyuniversally and not merely in the richindustrial societies.

The lack of elaborate social institutionsand services in low-income countries andtheir scant resources direct our attentionto whether or not the most basic materialsubsistence needs can be met in thosecountries. Meeting such needs as thesatisfaction of hunger is widely acceptedas a priority. Such needs have beenincluded in the categorisation of“absolute” poverty, which however wouldbetter be labelled “extreme” or “severe.”

Physical needs are subject to rapidchange because of shifts in social activityand demand patterns. The need formaterial goods, their relevance to thesociety of the day, and even the goodsthemselves, are not, after all, fixed orunvarying. And the amount and kind,and thus the cost, of food depend onwork, climate and social customs. Somaterial needs turn out to be sociallydetermined in different ways.

By the 1970s a second formulation—thatof “basic needs”—began to exert wideinfluence, supported strongly by the ILO.Two elements were included. First,minimum consumption needs of afamily: adequate food, shelter andclothing, as well as certain householdfurniture and equipment. And second,essential services provided by and for thecommunity at large, such as safe water,sanitation, public transport and healthcare, education and cultural facilities. Inrural areas, basic needs also include land,agricultural tools and access to farming.

The “basic needs” concept is an extensionof the subsistence concept. In additionto material needs for individual physicalsurvival and efficiency, there are the facilitiesand services—for health care, sanitationand education—required by localcommunities and populations as a whole.

The attractions to some of the“subsistence” concept included its limitedscope and therefore limited implicationsfor policy and political action. In the pastand into the present, it seemed easier torestrict the meaning of poverty to materialand physical needs than also to includethe non-fulfilment of social roles, giventhe overriding emphasis of individualism.

The “basic needs” concept, on the otherhand, aimed at establishing at least someof the preconditions for communitydevelopment. It played a prominent partin national development plans fosteredby the international community,especially UN agenices.

In the late 20th century, a third socialformulation of the meaning of povertywas developed: relative deprivation.“Relativity” as suggested above, applies toboth income and other resources and alsoto material and social conditions. In the21st century societies are passing throughsuch rapid change that a poverty standarddevised at some historical date in the pastis difficult to justify under new conditions.People living in the present are not subjectto the same laws, obligations and customsthat applied to a previous era.

Globalisation is connecting peoples andtheir standards of living, while inequalitieswithin and between countries are growing.There are, therefore, major objectionsto merely updating any historicalbenchmark of poverty on the basis ofsome index of prices. Over many yearsthe “relativity” of meanings of povertyhas come to be recognized, in part if notcomprehensively. Adam Smith, for example,recognized the ways in which “necessities”were defined by custom in the early partof the 19th century, citing the labourer’sneed to wear a shirt as an example.

It is not enough to describe poverty asa condition applying to those whosedisposable income is low relative to thatof others. This is to fail to distinguishconceptually between inequality andpoverty. Poor people are not just thevictims of a maldistribution of resourcesbut, more exactly, they lack, or aredenied, the resources to fulfil socialdemands and observe the customs aswell as the unfolding laws, of society.This criterion lends itself to scientificobservation, measurement and analysisof multiple deprivations.

However, as with any formulation, thereare problems in defining povertyoperationally. Under the “relativedeprivation” approach, a threshold ofincome is envisaged, according to sizeand type of family, below whichwithdrawal or exclusion from activemembership of society is common.

Establishing that threshold depends onaccumulating the available evidence, andwhether sociological and economicapproaches may be reconciled.

Such reconciliation is some distance away.Despite the influence of Amartya Sen’scontributions to development studies fortwo decades, his ideas on capabilities havenot reached the mainstream of povertyanalysis among economists, and havebeen said to leave important gaps.

There are forms of impoverishment, forexample through social exclusion, whenindividual capabilities to overcomepoverty are not at issue. Those capabilitiesare also identified as originating withinthe individual rather than with groups ornations collectively or being determinedexternally by market conglomerates. Again,the capabilities approach does not seemto address the structural sources of thecapabilities of the rich and powerful.Capabilities are different from perceptions.These sometimes offer a valuablecorrection to independent analysisof behaviour and living conditions.

In an attempt to define povertyoperationally, the World Bank in 1990adopted a rule-of-thumb measure ofUS$ 370 per year per person at 1985prices (the “dollar a day” poverty line) forpoor countries. This crude indicator mayhave been a convenient interim measurefor practical purposes, a short-termexpedient, but has not turned out tobe of continuing value.

Eliminating poverty requires betterdefinition and measurement. We need(i) an international poverty line thatdefines a threshold of income (includingin kind) required in different countries tosurmount material and social deprivation;(ii) regular reports on the extent of povertyin every country, based on measures ofboth “absolute” – i.e. “extreme” – and“overall” poverty, as agreed in 1995 at theWorld Summit for Social Development.Thus, antipoverty policies must bemonitored and evaluated regularly andon a much larger scale by governments,by the United Nations, by theinternational financial institutions and byother relevant international agencies.

Townsend, P.: Introduction, Compendiumof best practices in poverty measurement;Expert Group on Poverty Statistics (RioGroup), Rio de Janeiro, September 2006.


Poverty as a public policy concern,whether at the global, national orcommunity level, is now widelyconsidered to be a multidimensionalproblem. Over the last few decades, newperspectives on poverty have challengedthe focus on income and consumption asthe defining condition of poor people.Studies of the problems of poor peopleand communities, and of the obstaclesand opportunities to improving theirsituation, have led to an understanding ofpoverty as a complex set of deprivations.

These alternative perspectives haverefocused the concept of poverty as ahuman condition that reflects failures inmany dimensions of human life – hunger,unemployment, homelessness, illnessand health care, powerlessness andvictimisation, and social injustice; they alladd up to an assault on human dignity.

Strategies to eradicate poverty requirenot only economic growth andredistribution but also direct interventionin many areas such as expandingeducation, removing discrimination andsecuring social injustice; different typesof deprivations in human lives areinterrelated and reinforce one another.

For example, lack of education oftendefines the condition of a poor person butit is also an obstacle to other importantaspects of a person’s wellbeing, e.g.employment and income, good healthand health care and other basic amenitiessuch as clean water and sanitation. Theseare also interrelated with discriminationand lack of access to justice.

These ideas are not new but what isrelatively new is their emerging as aconsensus among policy makers, thepublic and development specialists. This isreflected in the adoption of the MillenniumDevelopment Goals, and as a conceptualshift in the treatment of poverty by theWorld Bank in their World DevelopmentReports from 1980 up to the WDR 2000/01,

which analysed poverty in terms ofopportunities, empowerment andvulnerability. This consensus builds on alively debate that challenged the incomefocus and argued for a human focus.

This debate included variousperspectives on poverty, such as basicneeds, participatory assessment, humanrights, social exclusion, and capabilities;all of which are discussed elsewhere inthis journal. The analytical applicationof the capability approach has beendeveloped and diffused through UNDP’sHuman Development Reports (HDR). Theyview poverty as reflecting the lack ofchoices and opportunities in the key areasof education, health, and command overresources, as well as voice related todemocratic processes. The first HDR in1990 introduced the HumanDevelopment Index (HDI).

Despite this shift to a multidimensionalpoverty concept, monitoring has continuedto rely on the income measure. At theglobal level, the $1/day (PPP) measuredeveloped and updated regularly by theWorld Bank is the one that is consistentlyused to monitor the size and trends inglobal poverty. At the national level,most governments define povertythreshold lines by household income.

Measures of poverty used by researchersalso rely on the income- and consumption-based national and internationalmeasures. For sure, a wide range of otherindicators are used at both global andnational levels. For example, the eightMillennium Development Goals have 49indicators to monitor progress. But it isthe income measure that is most usedto gauge trends overall.

One reason why the $1/day measureis relied upon for overall monitoringpurposes is the need to look at onenumber rather than 49 different ones tomake an overall assessment of progress.It is useful to have focused measures of

Human poverty isa complex set ofdeprivations inmany dimensions.

It has to bemeasured andanalysed withthis in mind.

Thus UNDP hasdeveloped thethree-dimensionalHuman Poverty Index.

It shows a largespread of humanpoverty amongcountries with similarlevels of income poverty.

Trends show goodprogress: 40 out of 44covered countries,including many inAfrica, have reducedhuman poverty levels.

The HumanPoverty Index:A multidimensional measure

by Sakiko Fukuda-Parr,The New School University


critical areas of human wellbeing such aschild mortality or access to clean water.But it is difficult to decide which one touse in making an overall assessmentabout whether poverty overall isimproving or deteriorating. A compositemeasure therefore is needed to make thisoverall assessment that can aggregatethe different features of deprivation.

The HDR 1996 introduced the HumanPoverty Index (HPI) to fill this gap. It is acomposite measure set in the capabilityand human development space, drawingon the several important perspectivesthat have enriched our understanding ofpoverty. In this framework, poverty is thedeprivation side of human development –the denial of basic choices andopportunities to lead a long, healthy,creative and free life; to enjoy a decentstandard of living; and to participatein the life of the community includingpolitical freedom and cultural choices.

HPI is a measure of capability deprivation;it aims to capture ‘human poverty’ asdistinct from ‘income poverty’, i.e. failuresto achieve the basic capabilities neededfor human functioning rather than anygiven level of consumption or income.But like the HDI, the HPI is a highlyincomplete measure and does notcapture many of the dimensions of afull life, nor the social conditions that arenecessary. Sen refers to five ‘instrumentalfreedoms’ as essential to a life of dignity,namely economic facilities, socialopportunities, political freedom, securityand transparency guarantees. Neither theHPI nor HDI include indicators of politicalfreedom, security and transparency.

The challenges of measuring humanpoverty are considerable, starting withthe selection of key dimensions, theirweighting, and a search for appropriatedata sets. Important aspects of humanpoverty, notably those relating to

Some very low-incomecountries reduced HPImore than many withhigher income levels.

HPI trends 1990-2002

Country Change in HPI HPI Rank Rank Change HPI-value (%) 1990 2002 1990 2002 in ranks*

India -12.1 43.5 31.4 33 26 -7

Viet Nam -11.7 31.7 20.0 25 22 -3

Ghana -10.5 36.5 26.0 26 24 -2

Tunisia -10.0 29.1 19.2 23 21 -2

Nicaragua -9.2 27.5 18.3 22 20 -2

Uganda -8.9 45.3 36.4 35 30 -5

Morocco -8.6 43.1 34.5 31 27 -4

Indonesia -8.3 26.1 17.8 20 18 -2

Guatemala -8.3 30.8 22.5 24 23 -1

Senegal -7.6 51.7 44.1 40 37 -3

Madagascar -4.8 40.7 35.9 29 29 0

Malawi 3.4 43.4 46.8 32 40 8

Botswana 19.8 23.7 43.5 17 36 19

Zimbabwe 25.8 26.3 52.0 21 41 20

* Negative value = rank improvement

participation such as political freedomand cultural choices, are simply notquantifiable or lack data and thus notincluded. Still, HPI is a more adequatemeasure of deprivations in human livesthan the income poverty measure.

HPI focuses on three of the four keydimensions of HDI: the capability to(i) survive, measured by vulnerability toearly death defined as before 40 years;(ii) be knowledgeable, measured by theadult illiteracy rate; and (iii) have accessto private income as well as publicprovisioning, measured by thepercentage of malnourished childrenunder five and by the percentage ofpeople without access to safe water.

These particular measures are not sorelevant in the richer industrial societies.Thus, an adjusted index (HPI-2) wasdeveloped with indicators more suitedto high income OECD countries. The HPIestimates are published in the annualHDRs. The new HDR 2006 shows someinteresting trends that contrast it fromthe $1 a day measure. Although HPI andthe $1 a day headcount measure arebroadly correlated, there are considerabledivergences between them (see page 9).

The correlation between income povertyand human poverty is weak, and there isa large spread in levels of human povertyamong countries with similar levels ofincome poverty. Some countries withhigh levels of human poverty around40% such as Pakistan and Yemen haverelatively low levels of income povertyaround 15-17% like Ecuador and China,which have reduced human poverty toaround 10%.

These data highlight both achievementsand challenges for governments. Somecountries with very low GDP per capitahave achieved significantly morereduction in human poverty thancountries with higher levels of income.Tanzania for example reduced humanpoverty to 36%, which is still high,though not when compared to theirincome poverty of 58% and per capita


Country designations

GDP of only PPP$ 674. This level of humanpoverty is similar to or lower thancountries with much higher incomes,such as Pakistan (36%) and Bangladesh(44%), where per capita GDP isrespectively PPP$ 2225 and PPP$ 1870.

Trends in human poverty are available inHDR 2004 for 12 years up to 2002 for 44countries. They show important progress;40 out of 44 countries show a netreduction in HPI, with India, Viet Nam,Ghana, Tunisia, Nicaragua, Uganda,Morocco, Indonesia and Guatemala havingreduced poverty by over 8 percentagepoints. It is encouraging to note themany African countries having madeprogress. However, dramatic increasesin human poverty were registered inZimbabwe (26%) and Botswana (20%).

HPI estimates so far have showna number of strengths:• Policy priorities: Countries that show

higher levels of human poverty inrelation to income poverty or percapita income point to a need forgreater policy attention on low socialachievements, especially publicprovisioning related to health andeducation. Although HPI does notcapture the full extent ofdeprivations in human lives,it covers the essential aspectsof life that relate to social andeconomic policies, and that gobeyond income alone.

• Feasibility: The HPI demonstratesthe relevance and the feasibility of amulti-dimensional aggregate measureto monitor human as opposed toincome poverty.

• Data availability: HPI usesinternationally comparable data.While there are many gaps, the gapsin these data are fewer than the gapsin the PPP $1 a day estimates. HPI isestimated for 102 developingcountries while the PPP $1 a dayestimate is available for 94 countries.

Further work that could develop a moreuseful measure to monitor poverty include:• Coverage: HPI covers 95 countries as

compared with 174 countries for HDI.• Global and regional estimates. HPI is

of particular interest for monitoringglobal trends in the way that the$1/day income poverty estimates do.Global and regional estimates of HPIshould be attempted.

HPI estimates show the relevance ofaggregate, internationally comparablemeasures of multidimensional poverty.Intercountry comparisons of HPI sharpenthe focus on human poverty that requirespolicy attention, and should be furtherdeveloped to provide a globalmonitoring tool.

UNDP Human Development Report 2006:Beyond Scarcity: Power, poverty and theglobal water crisis, New York, 2006.

BD = Bangladesh

BO = Bolivia

BR = Brazil

BF = Burkina Faso

CF = Central African Republic

CN = China

DO = Dominican Republic

EC = Ecuador

ET = Ethiopia

GH = Ghana

IN = India

ID = Indonesia

MG = Madagascar

ML = Mali

MN = Mongolia

MA = Morocco

NA = Namibia

NI = Nicaragua

NE = Niger

NG = Nigeria

PK = Pakistan

ZA = South Africa

TZ = Tanzania

TH = Thailand

TN = Tunisia

YE = Yemen

ZM = Zambia

ET

EC CN PK

TZ

YE

BDBF

NI

ID

TN

BOMN

MA

ZA

IN

MG

NA

TH

DO

ML

BR

NE

ZM

NG

GH

%

CF


There is now world-wide agreementon poverty reduction as an overridinggoal of development policy, but not onthe definition of poverty. Does thismatter for policy and practice? Wereviewed the monetary, capability, socialexclusion and participatory approachesto defining and measuring poverty; theydo have different implications for policy,and also for targeting since they identifydifferent people as being poor.

There are a number of issues involved indefining and measuring poverty. Is itconfined to material aspects of life, ordoes it also include social, cultural andpolitical aspects? Is it about what may beachieved, given the resources availableand the prevailing environment, or whatis actually achieved? Should definitionsand measurement methods be applied inthe same way in all countries and used forcomparisons? Are there “objective”methods, or are value judgementsinvolved? What is the rationale fordefining a poverty line? Should it beabsolute as in the Millennium DevelopmentGoals and in most developing countries,or relative as in the rich OECD countries?

Some further issues: Should poverty bedefined and measured at the individual,household or geographic area level; andfor what respective purpose? Themultidimensionality issue—consideringthat individual well-being/povertymanifests itself in multiple dimensions,should an aggregate index bedeveloped, and how? Finally, the timehorizon over which poverty is identifiedneeds to be specified—many peoplemove in and out of poverty over seasonsand years, hence the longer the timeperspective the less poverty will appear.

The monetary approach to theidentification and measurement ofpoverty is the most commonly used.

It identifies poverty with a shortfall inconsumption (or income) from somepoverty line. The valuation of thedifferent components of income orconsumption is done at market prices,which requires identification of therelevant market and the imputation ofmonetary values for those items that arenot valued through the market (such assubsistence production and social servicesand, in principle, other public goods).

Various technical solutions have beensuggested for differentiating the poorand non-poor using the monetaryapproach, although there is no theory ofpoverty that would clearly differentiatethe poor from the non-poor. Relativepoverty lines can be determined bypolitical consensus. Attempts to find anobjective basis for an absolute povertyline aim at identifying behaviouralbreaks between the poor and non-poor.The nutritional needs for survival providethe most common basis for such a break.

Yet there are problems about nutritionally-based poverty lines. Differing metabolicrates, activities, size, gender and ageamong people mean that what is adequatevaries among them. Moreover, differingtastes, food availability and prices affecthow much money income is needed tosecure any particular level of nutrition.Moreover, poverty lines are often drawnup at the level of the household,disregarding how the intra-householddistribution affects individual nutritionlevels. All this suggests that it is notpossible to draw up a unique povertyline based on nutritional requirements.

The three other approaches to deprivationaddress some of the perceived defects ofthe monetary approach.

The capability approach (CA) rejectsmonetary income as its measure of well-

Does the Definition ofPoverty Matter?Comparing four approaches

by Caterina Ruggeri Laderchi,*Ruhi Saith & Frances Stewart

Queen Elizabeth House,Oxford University

Poverty definitions andmeasurement have importantimplications for targetingand policy.

Empirical evidence shows thatpoverty rates vary a lot whendifferent concepts andmeasures are used.

A study of India and Perufound different peopleidentified as poordepending on whichapproach was adopted.

Clearer definitions andmeasures are essential forpoverty-centreddevelopment policies.

* The World Bank, Washington, D. C.


being, and instead focuses on indicatorsof freedom to live a valued life. Poverty isdefined as failure to achieve certainminimal or basic capabilities, i.e. theability to satisfy adequately certaincrucially important functionings. Manyanalysts have proposed generic lists ofessential capabilities, but they tend toreflect current western conceptions of the“good life”, raising doubts on its ability toreflect an “overlapping consensus”.Defining operational CA measures posesa number of methodological issues.

The concept of social exclusion (SE)describes the processes of marginalizationand deprivation that can arise even inrich countries with comprehensive welfareprovisions. It is a reminder of themultiple faces of deprivation. SE isperhaps the most difficult to interpretof the poverty concepts under review.Still, SE is the only approach thatfocuses intrinsically on the processesand dynamics that allow deprivationto arise and persist.

Moreover, the analysis of exclusionlends itself to the study of structuralcharacteristics of society and thesituation of marginalised groups (e.g.ethnic minorities or the landless), whereasthe monetary and capability approachestend rather to focus on individualcharacteristics and circumstances.SE leads to a focus on distributionalissues—the situation of those deprivedrelative to the norm generally cannotimprove without some redistribution ofopportunities and outcomes. SE alsopoints to excluders as well as excludees,again a contrast to the previousapproaches that describe a world withoutanalysing or attributing responsibility.

The participatory approach (PA) aims toget people themselves to participate indecisions about what it means to bepoor and the magnitude of poverty. Themajor advantage is that it largely avoidsexternally imposed standards. PA alsoprovides a way of solving some of theproblems encountered with the othermethods. For example, they help todefine: an appropriate minimum basketof commodities for the monetaryapproach; a list of basic capabilities in thecapability approach; and whether theconcept of SE can be applied in a

particular society and what its mainelements might be. The PA method iscomplex and invariably containsmultidimensional analysis. Like the SE itincludes processes, causes and outcomesof poverty, as perceived by the poor. ThePA is apparently cost-effective, but thecommunity spends much more time onthese exercises.

A critical issue for our comparison iswhether the four approaches identifybroadly the same people as poor; as ifthey do the theoretical differences maybe unimportant in policy or targetingterms. Despite its theoretical deficiencies,monetary poverty could be used as aproxy for other types of poverty ifbroadly the same people are identifiedas poor under the different measures.

Empirical evidence indicates that povertyrates differ significantly according to theapproach adopted, for countries as awhole and for regions of the world.Country ranking differs in comparingcapability poverty and both internationaland national monetary poverty lines.Different measures of deprivation areassociated, and indeed the differentmeasures do not seem independent(see page 9). What is striking, however,is that low levels of poverty according toone measure are compatible with highlevels of poverty according to another.

It is this variability which points tothe potential lack of overlap in practicebetween different ways of measuringpoverty, and it is this variability whichcalls for in-depth empirical assessmentof the underlying causes. Suchempirical tests can also show whetherdifferent measures are capturingdifferent populations.

A study of India and Peru, drawing onboth national data sets and micro-surveys, found that significantly differentpeople were identified as poor in the twocountries according to whether themonetary, capability or participatoryapproach was adopted.

The national data sets showed that inIndia, using the national poverty line,monetary poverty, at 38%, was belowcapability poverty: 52% of adults wereeducation poor (illiterate); and 26% of

children were education poor (notattending primary school); 70% ofchildren less than 13 years old wereundernourished, 44% severely; but only7% of individuals between 7 and 59suffered from chronic illness.

In Peru, in contrast, monetary poverty at54% (again using a national poverty line)was greater than capability poverty: 20%of the adults and 7% of the children wereeducation poor; 10% of adults werehealth poor and 29% of the childrenbelow 5 years were undernourished.

The India/Peru study had problems inestimating social exclusion. It had beenintended that the participatory focusgroups would define social exclusion, andthis definition would then be applied tothe data set. But none of the participatoryactivities generated a definition of socialexclusion—none of the groups sawthemselves as socially excluded.

Poverty definitions and measurementhave important implications for targetingand policy. The considerable lackof overlaps between the differentapproaches means that targetingaccording to one type of povertywill involve serious targeting errorsin relation to other types. Moreover,definitions also have implicationsfor policy. While a monetary approachsuggests a focus on increasing moneyincomes (by economic growth, orredistribution), a capability approachtends to lead to more emphasis on theprovision of public goods. Socialexclusion draws attention to the needto break down exclusionary factors,for example, by redistribution andantidiscrimination policies.

This suggests that identification andtargeting of the poor with combinedmethods should be more widelyadopted, reflecting the current concernsfor a broad characterization of poverty.Definitions do matter. Clearer and moretransparent definitions of povertyare an essential prerequisite of anydevelopment policy that puts povertyreduction at its centre.

Caterina Ruggeri Laderchi, Ruhi Saith &Frances Stewart: Does it Matter that wedo not Agree on the Definition of Poverty?A Comparison of Four Approaches. OxfordDevelopment Studies, September 2003.


Human Development:Beyond the HDI

by Gustav Ranis,* Frances Stewartand Emma Samman,

Queen Elizabeth House,Oxford University

Enlarged conception of HD

1. The HDI itself, which includes health, education and a measure of income 2. Mental wellbeing – an individual’s psychological state 3. Empowerment 4. Political freedom 5. Social relations 6. Community well-being 7. Inequalities 8. Work conditions 9. Leisure conditions10. Political security – freedom from political violence or instability11. Economic security – freedom from economic fluctuations12. Environmental conditions

Human Development (HD) hasbeen defined as ‘a process of enlargingpeople’s choices’. Although HD is oftenequated with the UNDP’s HumanDevelopment Index (HDI) in populardiscourse, it has long been recognizedthat this is an incomplete measure of HD,leaving out many important aspects oflife; it encompasses only three ratherbasic elements of human wellbeing.Our aim is to identify a broader set ofdimensions of wellbeing constituting HDand indicators to measure them. We thenexplore whether the HDI in fact is a goodmeasure of the broader dimensions orwhether additional indicators are needed.

Defining what makes for a fulfilled lifehas been a central theme of philosophersand politicians throughout history.Aristotle’s Ethics, for example, wasdevoted to identifying the conditionsneeded to achieve eudaimonia, commonlyinterpreted as ‘the best life’.

Our first task is to identify which aspectsof life might reasonably qualify as part ofa broader conception of HD. Manyattempts have been made to identifywhat makes for a flourishing life. Here wewill consider six lists, each of whichadopts a different philosophicalapproach and justification (see page 13).

The lists are not in total agreement; someemphasize certain aspects more thanothers for example regarding materialand mon-material aspects such asfriendship and emotions, andenvironmental issues.

Surveying these attempts to define thefull life, we find that they are in broadagreement about the main dimensions tobe included. In the light of this, we drewup a list of the categories of life that arecandidates to be included as part of anenlarged conception of HD (see box below).

Any list of categories is inevitably bothsubjective and ethnocentric. Hence, anyonefinding this type of approach helpfulshould be able to amend the categorizationto reflect different views. This appliesespecially across different cultures.

Unlike some of the reviewed literaturelistings of human flourishing, our list doesnot include spiritual well-being, due toproblems of definition and measurement.We separated social relations, whichconcerns individuals having satisfactoryrelations with others, from communitywellbeing, which concerns the communityas a whole and includes such elements aslow crime rates and a thriving civil society.We also separated empowerment from

UNDP’s HDI and HPImeasures include only a few basic aspects ofhuman development.

This article reports ona study that identifiedmany more relevantindicators.

Correlation analysisfound 31 indicators withdifferent trends fromeach other and alsofrom HDI.

Income per capitaproved a poor indicatorof human development.

Assessing humandevelopment fullyrequires a broad setof indicators.

* Yale University


Requirements for human flourishing

Authors Rawls (1972) Finnis, Grisez, Doyal and Nussbaum Narayan Camfieldand Boyle (1987) Gough (1993) (2000) -Parker (2000) (2005)

Definingconcepts

Bodilywellbeing

Materialwellbeing

Mentaldevelopment

Work

Security

Socialrelations

Spiritualwellbeing

Empowermentand politicalFreedom

Respect for

other species

Primarygoods

Income andwealth

Freedom ofoccupation

Social basesof self-respect

Rights,liberties,opportunitiesPowers &prerogatives ofoffice & positionsof responsibilityFreedom ofmovement

Basic humanvalues

Bodily life –health, vigourand safety

KnowledgePracticalreasonableness

Skillfulperformancein work and play

Friendship

Self-integrationHarmony withultimate source

of reality

Basic needsand Intermediateneeds

Physical health-Nutrition: foodand water-Health care-Safe birth controland child bearing-Safe Physicalenvironment

Protective housingEconomic security

Basiceducation

Work

Physical security

Significantprimaryrelationships

Autonomyof agencyCivil andpoliticalrightsPoliticalparticipation

Central humanfunctionalcapabilities

LifeBodily healthBodily integrity

Senses,Imagination,ThoughtEmotionsPractical reasonPlay

AffiliationSocial bases forself-respect

Controlover one’senvironment

Other species

Qualityof life

FoodShelter

Education(Bangladeshand Ethiopia,vs. Thailandand Peru)

Family

Religion(important inBangladeshand Thailand)

Dimensionsof wellbeing

Bodily well-beingAccess to healthservicesGood physicalenvironment

Material well-being FoodAssets

Work

Civil peacePhysically safeenvironmentLawfulness(access to justice)Personalphysical securitySecurity inold age

Social well-being-Family-Self-respectand dignity-Communityrelations

Freedom ofchoice and

action


political freedom, as the former focuseson the lack of power of particular groups,while the latter relates to liberal politicalconditions more generally. We includedinequalities as a general category, becausethey affect wellbeing independently,especially that of the poor.

We then identified indicators ofperformance in each of the categoriesfrom international data sets. Analysingcorrelations among the indicators withineach category across countries, anddiscarding those that are highly correlatedwith others in the same category, left 39indicators to encompass the eleven non-HDI categories. Of these, eight indicatorsare highly correlated with the HDI andmay therefore be represented by it. But 31are not highly correlated, suggesting thata full assessment of human development

The Ethical Poverty Line:A moral definition ofabsolute poverty

by Peter Edward,Judge Business School,Cambridge, UK

The first Millennium DevelopmentGoal (MDG) to halve ‘extreme’ povertyby 2015 has been justified as a moralduty. However, this morality is onlypartial if absolute poverty is definedby the $1-a-day poverty line. We needa morally defensible poverty line.

Poverty defined as a lack of well-being isclearly multidimensional. The MDG andWorld Bank $1-a-day poverty line, beinga solely monetary measure, has thereforerightly been criticised as inadequateand monodimensional. Yet, it is the mostwidely recognised poverty estimate inuse today. The objective of this article isto demonstrate that, even if we accepta monetary measure of poverty, the$1-a-day poverty line is unrealisticallylow thus misleading policy makersand the public on the extent of globalpoverty and the scale of redistributionneeded to remove it.

The basis for the $1-a-day poverty line issimply that it is the median of 10 of thelowest national poverty lines in the world.It is not derived from any considerationof well-being or basic needs. The WorldBank economists most involved in thisarea, recently called it ‘frugal’, statingthat it ‘must be deemed a conservativeestimate whereby aggregate poverty inthe developing world is defined by theperceptions of poverty found in thepoorest countries’. The implication is thatthe $1-a-day poverty line is unreasonablylow. Almost certainly it is lower thandeveloped world populations wouldconsider morally justifiable. Indeed, theWorld Bank does increasingly quotepoverty indices for a $2-a-day line thatseems to be based simply on a doublingof the $1-a-day poverty line.

World income distribution curves showthe highest population density between

requires a much broader set of indicatorsthan the HDI alone. We found that under-five mortality rates do as well as HDI, butPPP income per capita is less representativeof other dimensions of humandevelopment. The HDI (and the other twobroad indicators) are shown to be worseindicators of the extended conception ofhuman development for OECD countriesthan for developing countries.

Our research thus explores empiricalcorrelations, but does not attempt toinvestigate causality. We recognize thatour procedures are somewhat arbitraryand a change in the dimensions used,the thresholds adopted, etc., would yieldsomewhat different results. Our basicpurpose, however, is not to be definitivebut to show that extending the conceptand measurement of Human Development

to a broader set of dimensions seriouslyaffects the way one should measure andassess country performance.

In future work in this area, we intendto identify typologies of developingcountries/regions according to theirsuccess or failure over time with respectto the different dimensions of HD andrelate this to potential policy choices.To the extent that data are available,we would also like to trace the historicalprogress of the current OECD countriesin the various categories, which mayhelp in drawing conclusions abouttransitions over time.

Human Development: beyond theHuman Development Index by GustavRanis, Frances Stewart and Emma Samman.Journal of Human Development 7, 3, 323-358.

“We are one moral universe.And the shared moral sensecommon to us allmakes us recogniseour duty to others.”

Gordon Brown, UK Chancellorof the Exchequer, advocatinga “Marshall Plan” forthe world’s poor.


the $1-a-day to $2-a-day poverty lines.Thus, the poverty headcount is verysensitive to changes in the poverty line,especially around those levels. If thepublic and policy makers are to recognisejust how much poverty there is in theworld, a clear, relevant and morallyjustifiable basis for setting the povertyline becomes essential.

The ethical poverty line (EPL) proposedhere is derived from globally standardisedand ethically justifiable wellbeingoutcomes for which sufficient dataalready exist in the World DevelopmentIndicators (WDI). The EPL does notovercome the inherent problem of allincome poverty lines, namely that theyoversimplify and reduce the complexityof global poverty to a monodimensionalmonetary measure. However, the EPL atleast provides a morally defensible basisfor setting the line.

The intellectual background to the EPLdraws on two specific areas of research;first by economists on wellbeingmeasures and their determinants; andsecond on public health, relatingindividual life-expectancy to individualincome levels. These two areas can becombined to derive an internationalpoverty line directly from aggregate life-expectancy (i.e. wellbeing) outcomes. Theassociation between life-expectancy andincome was first explored by healtheconomists who showed how nationalaverage life-expectancy falls rapidly whenaverage income levels fall below a criticallevel (see figure). A similar correlation wasfound by public health researchersbetween individual income and life-expectancy, indicating a maximum life-expectancy around 73 to 75 yearsbeyond which increases in income havelittle or no further effect.

If poverty is best defined in terms of lowlevels of wellbeing rather than income,and life-expectancy at birth is theindicator of well-being, then an ethicalpoverty line can be developed bymodelling the relation between nationalaverage life-expectancy outcomes andtheir underlying source in individualabsolute poverty levels. The modelassumes the observed fact of a “kink” inthe correlation curve indicating thecritical consumption level.

Applying the model to nationalpopulations disaggregated byconsumption level, the statisticalcorrelation was established betweenmodelled life-expectancies and the actualones published in WDI. Similar correlationswere achieved from the best of thefunctions investigated for the model.Collectively these results set the kinkconsumption at between $2.7-a-day and$3.9-a-day, and probably around $3-a-day. To err on the side of caution, thefollowing discussion of the results isentirely based on the function that setskink-consumption at its lowest and henceleast challenging level, namely $2.7-a-day.

The results can be interpreted asshowing that, given the current state ofworld development, it is reasonable toexpect to live to 74 years if you have‘adequate’ consumption. Based onaverage performance across the world(excluding the worst distorting effect ofAIDS in Africa), $2.7-a-day (in 1993 PPPprices) should be ‘adequate’.Consumption above this level adds onlynominal years to expected lifespan.Consumption below this level reducesexpected lifespan dramatically.

Applying an ethical dimension to thekink model assumes that none of uswould wish to be born into such a lowconsumption level that our lifespan

risked dramatic shortening solely as aresult of that poverty. Invoking theGolden Rule—‘Do unto others as youwould have them do unto you’—wemight argue that every communityshould be entitled to achieve a minimumconsumption level sufficient for allindividuals to expect to live a reasonablyfull lifespan. On this basis 74 years couldbe considered to be a reasonablelifespan and $2.7-a-day would be theglobal ethical poverty line (Global EPL).In this sense the kink consumptionbecomes not only an absolute povertyline but also an Ethical Poverty Line (EPL).

This Global EPL is obviously significantlyhigher than the established $1-a-day line.However, it is not unreasonable. TheWorld Bank increasingly uses a $2-a-daypoverty line which is considered to bemore representative of the poverty linesof middle-income countries. The averagenational poverty line in the developingworld is is reported to be close to $2 perday , providing a further argument forraising the international poverty line toaround this level.

This analysis still includes a significantnumber of countries in sub-SaharanAfrica, where problems such as AIDS andcivil war lead to premature death. Whilethese problems are undoubtedlyassociated with poverty, they are not

Note: Circles are proportional to population.


50% of the world’s population. To removepoverty at the Minimum EPL level, theshare of the world’s output that thesepoor consume would need to increase by70%. Yet this is still only a poverty gap of5% of global consumption.

This may look like a fairly smallproportion of global consumption.It is put more clearly into perspective,however, by estimating the extent ofredistribution that would be required ifwe wanted to eliminate global povertytoday. For example, the cost of ‘merely’removing $1-a-day poverty would beequivalent to a 30% global tax on theconsumption of roughly the richest 1%of world population, affecting one in 10people in the USA and one in 20 in the UK.

Raising the poverty line to the EPL bothputs many more people in poverty andraises the threshold that each poorperson has to climb over to get out ofpoverty. The redistribution implications

directly caused by a lack of consumptionand they certainly cannot be resolvedmerely by increasing the consumption ofthe poor. Consequently their inclusion inthe analysis introduces an upward biasinto the EPL calculation that some mightargue was unreasonable.

Recognising this, the EPL has also beenrecalculated separating out all of sub-Saharan Africa. The results are striking.As expected, this Minimum EPL is lowerthan the Global EPL at around $1.9-a-day.The kink lifespan (72 years) is little alteredand stands in stark contrast to sub-Saharan Africa. There, the kink lifespanfalls to 48 years, while the very lowpoverty line ($0.6-a-day) may wellindicate that, only for those in themost extreme poverty, do the risksof premature death from lack ofconsumption outweigh the very highrisks of death from other causes. Thismay well be striking evidence of thedramatic impact of AIDS in the region.

What then are the implications of usingthe EPL to define absolute poverty? Thetable summarises a number of povertyindicators for the $1-a-day poverty line,the $1.9-a-day Minimum EPL and the$2.7-a-day Global EPL. Compared withthe $1-a-day line, the Minimum EPL morethan doubles the number of peopleconsidered to be in poverty to 2.5 billion,or 40% of the world’s population, whilethe Global EPL lifts this to 3 billion, or

of the EPL are huge. For example, thecost of removing ethical poverty today(conservatively based on the MinimumEPL line) would be comparable to anadditional global tax of 30% on allconsumption above US median levels.As a tax levied on anyone, anywhere inthe world, it would affect 6% of worldpopulation, including half the USpopulation and one in three people inthe UK. If we wanted to remove ethicalpoverty at the higher Global EPL level,this tax would extend to four-fifths ofthe US population and almost three-quarters of the UK population.

The EPL, therefore, reveals that the priceto be paid for accepting a moral duty toremove poverty today is one that wouldcut deeply into the pockets of developedworld populations.

Peter Edward: The Ethical Poverty Line:a moral quantification of absolute povertyThird World Quarterly, Vol. 27, No. 2, 2006.

A New Measure of thePoverty Line

by Lord Meghnad Desai,the London School of Economics

There is now a huge literature onpoverty, its measurement and its cure.This article concentrates on the questionof drawing up the poverty line, theabsolute level of income or consumptionbelow which people are said to bein poverty.

Two measures are considered as basicreferences, (i) the UNDP Human Poverty

Index (HPI, see article by Sakiko Fukuda-Parr, page 7), which includes lifeexpectancy, literacy, access to safe waterand child malnutrition as the criteria fordefining the proportion of populationwho are poor, and (ii) a new method fordetermining the poverty line usingutility theory as proposed by NanakKakwani, International Poverty Centre(see page 20).

This article suggests a new measurewhereby the two approaches canbe reconciled.

It does so by using the Hicksian conceptof income as the maximum that can bespent from a stock of capital whileleaving the value of the stock intact. Itlinks therefore income generation andconsumption to define the poverty line.

Poverty Line World Bank

$1-a-day

Indicators from poverty lines

Counting the poor

Poverty head-count (millions) 1050 2500 3100

Poverty rate (head-count as % of global population) 17% 41% 51%

Poverty gap ($billions PPP) 140 900 1980

Poverty gap as % of global GNI 0.7% 4.7% 10.4%

Minimum EPL $1.9-a-day

Global EPL $2.7-a-day


How to draw up thepoverty line, belowwhich people aresaid to be in poverty?

A new measure isproposed, based onthe need to maintaina person’s labourcapacity intact.

Definitions of absolute poverty havebeen static, calorific, asocial andatheoretical. Typically a basket of goodsis defined which will, at minimum cost,afford a required calorific standard to anindividual or a household. That cost isthen augmented to reflect the share offood in the total budget of a sampleof the population for which the povertyline is being drawn.

The thus derived poverty line is typicallystatic since it is taken to be a constantover time, except for inflationadjustment. It is asocial since it is uniformfor individuals of all ages and both sexes,regardless of health status and specialneeds. It is atheoretical because theabsolute poverty level concept fixes apriori an arbitrary level of calories withno choice involved.

Yet the biggest criticism of such measuresis that they focus on a passive variable,consumer expenditure, while ignoringthe income generation process.Inadequate income is after all the rootof poverty. A composite measure suchas the HPI looks at the family as a wholeand factors in several dimensionsof human poverty. These two ways oflooking at poverty should be combinedinto a single one.

We assume a household with someadults of working age and capacity aswell as children. The adult labour poweris an asset which will yield an income ifemployment can be found. Althoughthe household may also have access toland and/or credit or some form of non-human capital, we first focus on labourpower alone in the case of a personseeking paid work as hired employee.This capacity to work crucially dependson food intake, health and literacy.

From these assets the individual canexpect a flow of income, which willdepend on the demand for work and thewage rate. For someone self-employed,for example a peasant cultivator, it willdepend upon expectations of yield andthe price of output.

We can define the poverty line as thatlevel of expected income which allowsthe individual to consume enough foodwhile maintaining his or her labour

power intact. In particular, we can saythat enough food should be purchasablewhich will prevent any deterioration inhealth. The poverty level will thencrucially depend on the availability ofwork and the wage rate. If the individualcannot earn enough to be able to eat soas to prevent deterioration in health, and,in general, labour power, then he or sheis poor.

There are several advantages to thisdefinition when compared to thestandard definition which relies oncalories. For one thing it is dynamic;as availability of work or wages change,e.g. due to seasonal changes or becausethe individual’s health may deteriorate,he or she will move in and out ofpoverty. It also relates the consumptionof food to income generation whileavoiding imposing any specific calorificrequirements. It is firmly based in aneconomic theoretic definition of income.

The definition connects in an obviousway to issues of health and nutrition andthus builds a bridge between monetarymeasures and the HPI. In principle,though data demanding in practice,we can define this critical food intake atindividual level allowing for age, gender,health status.

There are of course several caveatsto be added. It is at an individuallevel and only for a wage labourer.It also simplifies the income generationprocess by taking the wage as givenand letting the uncertainty attach tothe hours worked. It will need to beextended to other income generationprocesses such as peasant cultivationwhether as a sharecropper or tenantfarmer or small landholder. There is alsothe need of aggregating over membersof a household. No doubt others willpoint out many more caveats whichneed to be made.

But the new approach presents in ahighly simplified fashion a way ofconnecting consumption and incomegeneration via the notion of labourpower/capacity to work. It allows fordifferent levels of required foodconsumption for different individualsand thus breaks away from any uniform“dollar-a-day” notion.


Measuring Povertywith a ChangingPopulation

by Ravi Kanbur,Cornell University

Poverty incidence may fallwhile the number of poorpeople grows. Has povertythen fallen or risen?

Most poverty measuresshow a decrease whenthe mortality amongpoor people is higherthan that of the rest.

Should those who wouldhave been alive but fortheir poverty be includedwhen counting the poor?

Is poverty unchangedwhen as many peoplehave become poor asthose who haveescaped poverty?

The basic structure of povertymeasurement has been built onintuitions that are strong when thepopulation whose poverty is beingmeasured is fixed and given. But when itis changing, standard measures of povertycan lead to uncomfortable or dissonantconclusions. To illustrate and explain thispoint, I want to consider incomedistribution at two points in time whenthe population has also changed in someway. I will look at three population sizescenarios—increased, decreased, andunchanged but with churning around aconstant income distribution.

Consider a country where the incidenceof poverty is falling by 1 percentage pointper annum but population is rising by 2percentage points per annum. Poverty asnormally measured—the number of poorpeople as a proportion of totalpopulation—has fallen. But the absolutenumber of the poor has risen. This is notjust a statistical curiosum, but it canoccur, and has occurred in countries, likethose in Africa, where population growthis relatively high.

Such an outcome can lead, and has led, todissonance between the claims ofeconomists that poverty has fallen, andthe ground level, lived experience of thosewho work with the poor and of the poorthemselves. “How,” the latter group exclaim,“can you economists claim that povertyhas gone down, when the soup kitchensare fuller than ever before and there aremore street children than ever before?”

The answer is found in some variant orother of the “population replicationaxiom”, which undergirds everycommonly used measure of poverty.Imagine two societies identical ineveryway. There is poverty in each, but itis identical poverty. Now imagine that wecombine these two societies into one.Would you say that poverty has stayed

the same, or doubled? The populationreplication axiom asserts categoricallythat poverty is unchanged. From thiscomes the mathematical form of most ofour poverty indices, where total populationis found in the denominator of theformula. But if you would say that in thisexample poverty has doubled, becausethere are twice as many poor people,then you would not buy into thepopulation replication axiom. I believethat most ground level activists, thosewho actually deal with poverty, wouldindeed be concerned about the largernumbers of the poor.

Since the issue is about an axiom, it is notnecessarily about being “right” or “wrong”.At the very least it is about seeing thingsin different ways, and it is about askingstatistical agencies to produce both sortsof numbers—normalized by populationsize, and not so normalized. It is goodthat the World Bank has started doingthis. With this information, we mightbetter understand the incomprehensionof those who see official poverty figuresgoing down while experienced poverty,the number of the poor, goes up.

Let me now construct another scenario.Starting from a given distribution ofincome, suppose that some of the poordie because of poverty. Of course, theabsolute number of poor goes down. It iseasy to check that even the proportion ofpoor will go down. Thus the conflictdiscussed above is no longer present.

Poverty has gone down whichever wayyou measure it, but only because poverty haskilled the poor! And this is not just acuriosum either. This phenomenon will bepresent to some extent whenever mortalityis higher among the poor than among therich, which is of course a very commonlyobserved phenomenon. And it will arisefor a wide variety of poverty measures, notjust the incidence of poverty. Moreover,


the issue arises also in the measurementof health status of a population, and inthe discussion of whether AIDS deathscan deliver a dividend in the form ofan increase in per capita income.

This cannot be right. We cannot have ourstandard poverty measures rewardingpolicies that reduce poverty by eliminatingthe poor. Put another way, we cannothave our standard poverty measurespenalizing policies that prevent thedeaths of the poor. And yet, the Number1 Millennium Development Goal, ofhalving the incidence of poverty over agiven length of time, is wide open to thisobjection. We have been led, somewhatunthinkingly, into this trap because theintuitions that work well when populationis fixed go awry when it is changing.

The answer to this conundrum is not easy,because among other things it forces usto ask how far back we want to go inevaluating today’s poverty. Had somepoor children who died yesterday beenalive today, poverty would have beenhigher. But how would we count thosewho died a hundred years ago becauseof poverty? If a year ago is too short, anda hundred years ago is too long, wherein between is the right cut-off?

An alternative is to have measures ofincome poverty stay as they are, but alsoinclude measures of mortality (amongthe poor) in our evaluation. This is thedirection taken by UNDP’s Human PovertyIndex, which has the proportion of peoplelikely to die before the age of 60 as acomponent along with an income povertymeasure. But the specific assumptionsunderlying any measure need to be madeclear; in particular how to define thepopulation whose wellbeing is beingevaluated. Those currently alive, which isthe present empirical method, and whichgets us into the difficulties noted above?Or do we include some of those whowould have been alive but for theirpoverty? An explicit derivation ofmeasures that satisfy key assumptionsand basic intuitions is some ways away.

Let us finally turn to a situation whereboth the population and the incomedistribution is unchanging, but thedifferent income slots are occupiedby different individuals at different times.In other words, there is mobility. Consider

the simplest case, where there are twoincome levels, one below the poverty lineand one above the poverty line. In onesociety, the rich stay rich and the poorstay poor. In another, the rich and poorswap places every period. To keep tingsconceptually simple, let there be nosaving or investment, so that income isconsumption. The snapshot poverty(income or consumption) is unchangingperiod to period, of course, and is thesame in the two societies. But are these twosocieties really the same in terms of poverty?

The answer depends on our intuitionsof what the experience of poverty means,and whether the experience of poverty inone period for an individual can be“washed away” by an above poverty lineexperience in another. Imagine yourself atthe end of Shakespeare’s play, King Lear.You see a once powerful man broughtlow and you heart goes out to him. Thensomeone says, “Well, you know, the firstfifty years of his life weren’t all that bad,and all in all….” To aggregate over Lear’slife is to miss the point and thepoignancy of his downfall.

To aggregate poverty experiences withabove poverty line experiences for anindividual is equally, in my view, to missthe point about poverty—at least if wetake the idea of poverty seriously enoughto set it apart from non-poverty, as wedo in the way that we measure poverty.On the King Lear view, therefore, the twosocieties I described above have the samepoverty. If this makes you uncomfortable,and at the same time you find somethingcompelling about the Lear example, thenyou will agree that this issue has notbeen fully addressed in our measurementof poverty.

The three conundrums exposited here aremeant not to paralyze the measurement ofpoverty but to invigorate the discussionof the assumptions that underlie thestandard measures of poverty. I lookforward to the response.

Chakravarty, Satya R., Ravi Kanbur andDiganta Mukherjee. 2006. “PopulationGrowth and Poverty Measurement.” SocialChoice and Welfare, Vol.26, pp 471-483.

Kanbur, Ravi and Diganta Mukherjee. 2006.“Premature Mortality and PovertyMeasurement.” Processed. CornellUniversity: <http://www.arts.cornell.edu/poverty/kanbur/Pov&Death.pdf>.

Can the experience ofpoverty in one periodfor an individual be“washed away”by an above poverty lineexperience in another?


Poverty and Wellbeingby Nanak Kakwani,International Poverty Centre,

UNDP

Economic growth isnot a good indicatorof poverty reduction.

Poverty means lowlevels of wellbeing,not just low income;measures must relateclosely to people’s lives.

Poverty is failure of sociallydetermined basic capabilities,but only when caused by lackof resources.

A new international povertyline is needed. It should reflectthe cost of basic human needs.

According to the 2000 WorldDevelopment Report (WDR), “poverty ispronounced deprivation in wellbeing”.But what does wellbeing mean? Howdo we define it? What are the elementsnecessary to ensure a decent level ofwellbeing? These are not easyquestions to answer.

In the socioeconomic literature, severalapproaches have been used to describewellbeing; important among them arebasic needs, economic growth, qualityof life and welfare. How do theseapproaches differ? And, whichapproach is the most appropriate todescribe wellbeing? This article deals,briefly, with these broad questions.

In any society, some people obviouslyenjoy higher levels of wellbeing thanothers. Poverty is viewed here as thelowest level of wellbeing, which isexperienced by those people in societywho are so deprived that they areunable to function with dignity. Asstated in the 2002 WDR, to be poor “is tobe hungry, to lack shelter and clothing,to be sick and not cared for, to beilliterate and not schooled”.

Economic growth provides greatercommand over goods and servicesand thus, on average, gives peoplegreater choices. However, this doesnot necessarily imply higher wellbeingfor everyone; the benefits of economicgrowth are never shared equally.Increasing per capita income is notan appropriate indicator of changesin the aggregate wellbeing. Economicgrowth is a means and not an endof development and thus must besupplemented with indicators that aremore closely related to individual lives.

People want income because it givesthem the possession of commodities,

which they consume. The higher theincome the greater the command peoplehave over commodities. The possessionof commodities, which include services,provides people with the means to leada better life; thus, the possession ofcommodities or opulence is closelyrelated to the quality of life peoplelead. But it is only a means to an end.As Amartya Sen points out, “ultimately,the focus has to be on what we can orcannot do, can or cannot be”. Thus,wellbeing must be seen in terms ofindividual achievements and not interms of means that individuals possess.

The wellbeing or standard of living is notabout the possession of commodities,but it is about living. This line ofreasoning led Sen to develop theideas of “functionings and capabilities”.Thus, a ‘functioning’ is an achievement,and a ‘capability’ is the ability to achieve.Functionings are directly related to whatkind of life people actually lead, whereascapabilities are connected with thefreedom people have in the choice oflife they lead, which is their functionings.

Income allows an individual to purchasecommodities with which he or shegenerates various functionings. But allindividuals cannot convert commoditiesinto functionings to the same degree.For instance, a disabled person maynot be able to do many things that anable-bodied individual can do with thesame commodity. Thus, in measuringwellbeing or standard of living, ourfocus should be on the achievementsof people and not on the commoditiesthey possess.

It may seem obvious that the higherthe income people have, the greaterwill be their capabilities. After all, it is anobserved fact that rich countries do havea higher standard of living than poor


It would be oddto call a disabledmillionaire “poor”.

Wellbeingis not happiness.It is about capabilitiesand achievements.

countries. But the relationship betweenthe two is not as simple as it appears. Forinstance, consider a country, which hassucceeded in reducing mortality so muchthat its per capita GDP falls because ofthe resulting increase in population; hasthe country’s living standard improved ordeteriorated? The answer is not clear. Thefall in per capita income shows that theaverage citizen has become poorer, whileat the same time the country has extendedthe capability of its citizens to live alonger life. This example demonstratesthe complex nature of the relationshipbetween the income and the capabilitiespeople possess.

It is not uncommon to think thatultimately it is the happiness of peoplethat really counts. Could we thenmeasure wellbeing by the degree ofhappiness enjoyed by the people?The answer is no. Sen argues that livingstandards cannot be the same thingas happiness. People can be happy orunhappy irrespective of how they liveor do not live. In India many people livein severe poverty with no access to thebasic amenities of life, but they may stillbe happy because of the religious beliefthat in their next life they will be betteroff. Indeed, they believe that the morethey suffer in this life, the better off theywill be in their next life. These religiousbeliefs may help people to suffer happilybut they do not improve their livingconditions. Wellbeing is aboutcapabilities and achievements.

The level of happiness can still be a goodindicator of people’s wellbeing, but it isnot the same thing. Happiness can onlybe measured by people’s perceptions,which may change drastically in a shortperiod. Thus, the level of happiness alonewould not be a stable or suffcientindicator of wellbeing. It may changewithout any change in people’s actuallevels of living.

The income approach views poverty asdeprivation of income (or consumption).

Poverty is defined by a poverty line, i.e.the minimum income needed to be ableto satisfy minimum basic needs. Butincome is not the only kind of deprivationpeople may suffer. Although incomedeprivation may give rise to several otherkinds of deprivations, people may sufferacute deprivation in many aspects of lifeeven if they possess adequate commandover commodities. It is the low level ofwellbeing which is important rather thanlow level of income.

Thus, poverty should be viewed as thedeprivation of basic capabilities ratherthan merely as low level of income.Poverty encompasses not only materialdeprivation (measured by income orconsumption) but also many other formsof deprivations in different aspects of lifesuch as unemployment, ill health, lack ofeducation, vulnerability, powerlessness,social exclusion and so on.

Under the capability deprivationapproach, an individual may be definedas poor if he or she lacks basiccapabilities. What are these basiccapabilities? How do we identify them?This is an issue of value judgment. Itdepends on how a society prioritizesdifferent capabilities. These priorities mayalso depend on a country’s economicresources. There is no clear-cut formulafor determining basic capabilities.

Can we describe poverty purely in termsof capability deprivation? Suppose thata millionaire, who has all the economicmeans to buy anything, has a diseaseor disability which prevents him fromachieving some basic functionings. Heor she is surely suffering from a seriouscapability deprivation in spite of havingall the best medical facilities at his or herdisposal. Yet, it would be odd to call thismillionaire “poor”.

Poverty is present when basic capabilityfailure arises because a person hasinadequate command over resources,whether through market or non-marketsources. By examining capabilitydeprivation alone, we cannot alwaysidentify persons who are poor in thisspecific sense. Accordingly, a distinctionshould be made between capabilitydeprivation in general and povertyspecifically. Poverty results from the

inadequacy of command over resourcesneeded to generate socially determinedbasic capabilities, whereas capabilitydeprivation is more general and maybe caused by a host of factors.

Among them, income or entitlement toresources may not be the most important.Thus, a person may suffer capabilitydeprivation but still may not always bepoor. Defining poverty from the capabilityperspective cannot be done independentlyof income. The capabilities to function,which are linked to poverty, are thosederivable from income. Income orcommand over recourses cannot beseparated from capability, but at thesame it must be recognized that the linkbetween them is far from simple.

Individuals have different needs and,therefore, differ with respect to theirability to convert incomes or resourcesthey have into capabilities to function.Thus different individuals will requiredifferent resources to achieve basiccapabilities. It is best that any proposedincome measure of poverty be constructedfrom capabilities. The choice of a povertyline should reflect the cost of achievingbasic human needs.

However, that is not the case with themost frequently used internationalpoverty measure, the PPP $1/day perperson. It was constructed by WorldBank researchers in 1990 as the medianof the lowest ten national poverty linesavailable in a sample of 33 countries.It has been shown that a person maynot attain income-dependent basiccapabilities even with an incomesufficient to count as non-pooraccording to the $1/day yardstick.If the reduction of poverty, as properlyunderstood, is to be achieved, then newmeasures of its extent and distributionin the world will be required.

Nanak Kakwani and Hyun H. Son: “NewGlobal Poverty Counts”, Working paper no. 29,UNDP International Poverty Centre,September 2006.

22 United Nations Development Programme22 United Nations Development Programme

I N R E S P O N S Eby Sabina Alkire, Oxford University

Kakwani’s broad viewof the possible elementsof poverty seems exactly right.

But then he turns back toresource-based definitionsof poverty and toincome-based definitionsof capability poverty.

Basic capabilities can oftenbe expanded by public actionregardless of causation.

There are numerousways to measurecapability deprivation.

They should be furtherdeveloped rather thanconfining interest toincome poverty lines alone.

Re: Poverty and Wellbeing

Nanak Kakwani’s careful and clear article acknowledges the conceptualinadequacies of aggregate economic growth and happiness as representations ofwellbeing or poverty. It then argues that poverty should be viewed as the deprivationof basic capabilities rather than merely as low level of income. Moreover it states thatthe selection of basic capabilities is a value judgement and that there is no clear-cutformula for determining basic capabilities. It further states: “Poverty encompasses notonly material deprivation (measured by income or consumption) but also many otherforms of deprivations in different aspects of life such as unemployment, ill health, lackof education, vulnerability, powerlessness, social exclusion and so on.” This broad viewof the possible elements of poverty seems exactly right.

At the close the article takes two quite interesting turns, first back to resource-baseddefinitions of poverty and second back to income-based definitions of capabilitypoverty. Although I am not entirely sure I have interpreted these correctly, I would liketo draw attention to these turns, and also to other possible avenues for exploration.

Having established that “poverty should be viewed as the deprivation of basiccapabilities”, Kakwani then appears to argue that this is the case only when poverty iscaused by resource deprivation: “Poverty is present when basic capability failure arisesbecause a person has inadequate command over resources, whether through market ornon-market sources.” He poses the case of a millionaire with an illness that preventshim or her from achieving a basic functioning. The millionaire is capability deprived –despite having splendid medical care – but should not, he argues, be called ‘poor’.

The article closes with a discussion of the most frequently used poverty line – $1 a dayper person. Here the article takes a turn, arguing that despite the complexity of therelationship between capabilities and income, “defining poverty from the capabilityperspective cannot be done independently of income”. An alternative reading ofKakwani’s last argument is that capability poverty might be defined in other ways, butin those situations in which income poverty lines are to be constructed, the basis of theirconstruction should be some kind(s) of capability deprivation.

The article contains three quite distinct lines of thought:• Poverty should be viewed as the set of deprivations of basic capabilities.• Poverty refers to the subset of basic capability deprivations that are caused by

inadequate command over resources.• Poverty refers to basic capability deprivation insofar as this can be represented

by income poverty lines or in income space.

I would warmly endorse the first of these, but reframe the second and third, to arguea) that poverty should be defined normatively not according to its cause but accordingto whether it can be influenced by public action, and b) that multidimensional povertycomparisons should be developed to displace income poverty lines (although whenthe latter are used they should be pegged to basic capabilities).

Consider the ill millionaire, and compare her with a vivacious millionaire. Perhaps the‘basic functioning’ that her illness prevents is the physical ability to be nourished; thusissues of energy and indeed survival reemerge constantly. Clearly this capabilitydeprivation is caused by some condition other than resource deprivation. Should shebe considered poor? The category of ‘capability deprivation’ allows Kakwani to describethe difference between our rail-thin millionaire and a plump and vivacious millionaire.In defining poverty as the subset of capability deprivations that are caused by


inadequate resources, it seems that Kakwani wishes to define poverty not descriptively but normatively as capability deprivationsthat can (and should) be addressed by public action (by which we might mean not only action of the state sector but also of NGOs,business, and civil society). The problem with defining poverty according to its cause is not only that it is difficult to establishcausality, but it may not be necessary: basic capabilities can often be expanded by public action regardless of causation.

When we use the term poverty as a moral or normative concept that serves as a trigger for action (rather than as a description, whichone could also do), one might instead define poverty as the subset of basic capability deprivations that are “socially influenceable.”In his article “Elements of a Theory of Human Rights,” Sen argued that capabilities could be selected for special attention on the basisof two criteria. First, they were of special importance (that is, they were judged to be basic capabilities). Second, they were sociallyinfluenceable. For example, tranquility of mind or serenity of heart may be tremendously important to many – both the destitute andthe wealthy – but it is not terribly amenable to change by concerted public action. This definition might achieve Kakwani’s objectivewithout requiring neither the establishment of causality, nor the definition of what counts as command over resources.

A third point concerns how to establish income poverty lines. Kakwani acknowledges the difficulties in using income to indicatebasic capabilities because of people’s very different personal conversion factors (as well as, we might observe, different institutionaland market contexts), but argues that income poverty should be grounded on capability deprivation rather than on absolute incomelevel such as $1/day. The only peculiarity with this argument is that it neither defends nor critically examines the focus on incomespace. Income poverty lines adjusted to reflect some basic functioning, such as nutritional status, or (and this is quite different) toreflect a resource, such as a caloric line, have a conceptual advantage over current PPP lines. But there are numerous ways to measurecapability deprivation.

Foster and Sen describe four. The first is the ‘direct’ approach that employed indicators of functionings either alone, in a disaggregatedway or aggregated into a multidimensional index such as the Human Poverty Index (HPI) or the Bourguignon-Chakravarty Index,which specifies a ‘poverty line’ for each of multiple dimensions of poverty. The second supplemented income measures informallywith other measures, perhaps akin to contemporary poverty profiles. The third sets income poverty lines with respect to capabilities,for example through equivalence scales; the fourth examines partial orderings of ordinal intensity (for example DALYs and theirsuccessor indicators).

Research has broken open multidimensional poverty measurement and is actively developing multidimensional measures such asindices, poverty profiles, and stochastic dominance comparisons. Given their promise, measurement of capability poverty shouldinvolve and further develop these measures rather than confining interest to income poverty lines alone.

Sen, A. K., On Economic Inequality with a substantial annexe ‘after a Quarter Century’ by J Foster and A Sen. 2nd Edition. (Oxford: ClarendonOxford Press, 1997).Sen, A. K., ‘Elements of a Theory of Human Rights’, Philosophy and Public Affairs, 234 (2004).

Rejoinder by Nanak Kakwani


Sabina Alkire’s gracious response even strengthens my argument that a distinction should be made between capability deprivation ingeneral and poverty specifically. Poverty is concerned with the inadequacy of command over resources needed to generate sociallydetermined basic capabilities, whereas capability deprivation is more general and may be caused by a host of factors. It seems thatAlkire accepts our view that poverty is a subset of capability deprivations, which is generally not recognized. We define this subsetas the one arising due to inadequate command over resources, whereas Alkire seems to agree with the one that can be influencedby public action. Any public action intended to reduce poverty basically provides (directly or indirectly) additional command overresources to those people who lack them. Thus, if we understand her position correctly, there is but little difference in our views.

There are two distinct problems in measuring poverty. The first is identifying the poor, and the second is aggregating poverty. Myarticle is mainly concerned with the first issue, while for example the HPI is a response to the second. The identification problem isclearly prior to aggregation. It would be odd to identify a millionaire as poor even when suffering from a basic capability deprivation.Thus, defining poverty from the capability perspective cannot be done independently of income. In this context, the issue ofestablishing causality is not all that difficult. Moreover, in the formulation of public policies, we have to confront causal relationshipseven if they are difficult to establish.

A poverty line should not be a single number defined in income space. Since different individuals differ with respect to their abilityto convert income or resources into capabilities to function, they require different resources to achieve basic capabilities. It meansthat different individuals will have different poverty lines to be able to enjoy the same basic capabilities. In our view the concept ofpoverty line is multidimensional because it takes into account people’s individual differences in the enjoyment of a set of basiccapabilities. Thus, the poverty lines properly constructed can reflect different interpersonal conversion rates as well institutionaldifferences. Sabina Alkire’s criticisms are mainly directed to a unitary income poverty line, which obviously is not our view.

International Poverty CentreSBS – Ed. BNDES, 10º andar70076-900 Brasilia DFBrazil

povertycentre@undp-povertycentre.orgwww.undp-povertycentre.orgTelephone +55 61 2105 5000 D

ecem

ber

2

00

6

Documents

What is poverty? - Sakiko Fukuda-Parrsakikofukudaparr.net/wp-content/uploads/2013/01/Poverty... · 2013. 2. 7. · Lord Meghnad Desai finds the definitions of absolute poverty static,