Download - PLAAS

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South AfricaContentsAbstractiiAuthor’s noteiii1. Introduction 12. Imagining and understanding chronic poverty 2Conceptualising and measuring chronic poverty 2Some chronic problems with poverty measurement 2Capabilities and multidimensionality 33. From chronic to structural poverty 6Vulnerability, agency and structural poverty 6From ‘distributions’ to relationships 64. Poverty measurement and the government of poverty 10Beyond Q-squared 10Power, knowledge methodological dissent 115. Conclusion 13References 14FigureFigure 1: PLAAS and CPRC’s research sites in South Africa 7i

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South AfricaAbstractThis paper explores the challenge of understanding chronic and structural poverty in South Africa,and questions the dominance of the econometric imaginary in present-day development andpoverty studies. It argues that measurement-based, econometric approaches to chronic poverty aredependent upon mystifying narratives about the nature of poverty and how it can be known, thatthey direct attention away from the underlying structural dimensions of persistent poverty and thatunderstanding structural poverty in turn requires a theorised engagement with the complexitiesof social relations, agency, culture and subjectivity. Valuable as the recent re-recognition ofthe need to connect qualitative and quantitative research has been, attempts at ‘qual-quant’integration often remain tied to positivist assumptions – bringing the risk of a new ‘ordering’ ofmethodological dissent that leaves problematic aspects of the econometric imaginary unchanged.Underlying this process is the entanglement of poverty research with the ‘government ofpoverty’: the attempt to constitute poverty as something objectively measurable and scientificallymanageable. The paper closes with a consideration of the ethical and political challenges thisposes for critical researchers and intellectuals in post-colonial contexts.ii

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africa1. IntroductionWe don’t want complicated stories.What we need is a number. One number,if possible. One indicator that tells uswhere the poor and vulnerable are.That’s what we need.(Member of the Regional VulnerabilityAssessment Committee for Botswana,at a planning meeting of the SouthernAfrican Vulnerability Initiative, October2004).Discussions about method and methodologyin applied social research are often framed asif the central differences are those betweenquantitative and qualitative methods, and as ifthe key issue to be decided is the value of oneor the other – or the best way of ‘integrating’them (see Kanbur 2002). This paper arguesthat it is necessary to go further. It considersthe difficulties that arise out of the dominationof development studies and poverty researchby what is here called the ‘econometricimaginary’: an approach that frames questionsof social understanding essentially as questionsof measurement. Although the limitations ofthe econometric imaginary clearly illustratethe need for qualitative modes of research andunderstanding, I argue here that, although itis important to think about various methodsof combining or ‘integrating’ qualitative andquantitative approaches, there is another,additional and more daunting challenge thatcannot be avoided. Also pertinent is the largerexplanatory meta-narratives: the paradigmsand theoretical frameworks that guide theprocess of integration. Meeting this challengeis, however, impossible without an engagementwith the ways in which applied social scienceresearch in the 21 st century is shaped by thearchitectures of power and knowledge inmodern states and donor institutions. In SouthAfrica, these limitations, I argue, are part ofa fertile yet hazardous terrain for engagementand contestation by critical scholars andresearchers.These threads of argument are hung fromthe rather humble edifice of a considerationof some years of ‘chronic poverty’ researchconducted in South Africa (see Aliber 2001;De Swardt 2004a, 2004b; Du Toit 2004, 2005a;Du Toit, Skuse & Cousins 2005; Arnall et al.2004) and the attempt to link the findings ofthis research to mainstream debates on chronicpoverty. In the first place, the paper argues thatdominant approaches to the conceptualisationof chronic poverty are undermined by theirreliance on a mystificatory theoreticalmeta-narrative that tries to imbue povertyjudgements with a spurious aura of objectivityand by the fact that they direct attention awayfrom structural aspects of persistent poverty.Secondly, it argues that if the analysis ofstructural poverty is to avoid reductionism orabstraction, we need to come to grips with theextent to which the structural configurationsof poverty are socially meaningful; shapedthrough and through by the complexitiesof culture, identity and agency. Thirdly, itproposes that this implies that more is neededthan the simple addition of qualitative data toexisting measurement-based accounts: instead,critical theory allows a re-imagining and reframingof the way in which inequality andpoverty are conceptualised in the first place.The paper closes with a consideration of someof the obstacles and limitations to be facedin an attempt to bring these alternative waysof imagining poverty into the mainstream ofapplied poverty work in South Africa.1

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africa2. Imagining and understandingchronic povertyConceptualising and measuringchronic povertyOur research on persistent poverty in SouthAfrica is essentially framed by the organisingconcept of chronic poverty. This is often givena fairly broad meaning – in the work of theChronic Poverty Research Centre (CPRC),for instance, it refers inter alia to poverty oflong duration, the poverty of those who arepoor for most of their lives and ‘transmit theirpoverty’ [sic] to subsequent generations, to thesituation of those caught in poverty traps andto those who number among the ‘hard-to-reachpoor’, and so on (see Hulme & Shepherd 2003;CPRC 2004). Ultimately, however, chronicpoverty is usually understood, in its canonicaleconometric sense, in contradistinction totransitory poverty. Though the econometricanalysis of chronic poverty is possible onthe basis of ‘static’ indicators that are robustto change over time (for example, Chauduri& Ravallion 1994; see also McKay & Lawson2003), a preferred strategy is to aggregatestatic snapshots in a way that might allowa composite ‘moving’ picture to emerge. Atypical approach is to run a panel dataset and touse a poverty line (most commonly monetaryin nature) to develop a dichotomous indicatorwhich is then used to divide the individualsin the population in each wave of the panelstudy into two groups – usually ‘the poor’ and‘the non-poor’. Those who move above (ordip below) the poverty line are held to have‘escaped poverty’ (or to have ‘entered’ it);those who are counted as poor in every waveof the survey, or who on average remain belowthe poverty line are counted as the ‘chronicallypoor’ (see Bane & Elwood 1986; Baulch1996; Baulch & Masset 2003). This approachdominates the ways in which ‘the chronicpoor’ are identified; although other ways ofapproaching persistent poverty exist they areoften treated simply as complementary.In this paper I argue that, important as thedistinction between chronic and transitorypoverty can be, it is also very limited, focusingattention away from other matters critical to theunderstanding of persistent poverty. It is alsotied up with some deeply problematic – indeed,thoroughly mystificatory – underlying metanarrativesabout poverty itself, what it is andhow it can be scientifically known. To gobeyond the limitations of the econometricconcept of chronic poverty, then, it isnecessary to engage with the ways in whichthe econometric imaginary dominant in appliedsocial science frames the concept of povertyitself.Some chronic problems withpoverty measurementLet us begin this engagement by consideringthe practices of ‘poverty measurement’ uponwhich the definition of chronic poverty– and the identification of ‘the chronic poor’– depend. These involve, as we have seen,two key operations. Firstly, they require theidentification of an ‘indicator’ which stands asa proxy for the state of poverty; and secondlythey involve the division of a ‘population’ intotwo groups on the basis of this indicator.These operations involve three keydifficulties. Firstly, poverty judgements– judgements as to whether someone is poor,and about what it is that constitutes theirpoverty – are ordinarily moral and politicaljudgements: they derive their import and areinvested with significance and consequenceby virtue of being embedded in underlyingdiscourses about the nature of society, theidentity of its members, and the nature of theclaims and counter-claims that membershipenables. Furthermore, poverty judgementsare always made by particular social actors,and are therefore always part of some largersocial and political agenda. Any judgementabout whether or not a particular person ispoor – or about what the ‘essentials of life’are, the lack of which constitutes poverty– is always a political judgement and is2

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africaoften contested (Noble, Ratcliffe & Wright2004). This means that there is no objective,uncontroversial, value-free and unitary conceptof poverty directly available for transparentoperationalisation by ‘social science’.Scholarly and applied research aboutpoverty cannot disregard this. The claimspoverty experts make to truth, resources, timeand attention are dependent – even parasitic– upon these broader and essentially contestedpolitical and moral meta-narratives. Tryingto impart a spurious cut-and-dried ‘objective’‘scientificity’ to poverty measurement is not tomake it rigorous, but to mystify it.This is not simply an abstract point.Consider the role played by poverty linesin the attempt to make poverty judgementsrigorous and objective. As should be evident,this immediately raises the issue of just wherethe poverty line should be set. (For a SouthAfrican discussion see Leibbrandt & Woolard2001.) Some have developed interestingapproaches that attempt to ground this decisionin local consensus(es) about ‘socially acceptednecessities’ (Noble, Ratcliffe & Wright2004), but quite often this decision seemsto be informed by the assumption that valuejudgements can be avoided altogether and thatit is possible to develop a ‘scientific’ standardbased on some ‘objective’ reality – for exampledietary needs, caloric intake requirements andthe like (see Baulch & Masset 2003). Almostinevitably this leads not to an uncontroversialbut to a punishingly conservative povertyline – one in which only those who are at riskof starvation or malnutrition will ever reallyformally count as poor – and a situation where,paradoxically, there is widespread povertyabove the poverty line.Secondly, one important consequence ofthe inherently political and moral character ofpoverty judgements is that they involve a widespace for nuance and indeterminacy. It is partof the logic of the concept of poverty that wecan speak of someone as being, for example,‘not very poor’, ‘almost poor’ or ‘poor – fora white person’. The econometric habit ofdividing ‘populations’ into ‘poor’ and ‘non’-poor – a distinction absolutely central to theway in which chronic poverty is distinguishedfrom transitory – involves a misrecognitionof this essential feature. Though somehave attempted to recognise the space forindeterminacy in poverty judgements, forexample, by using fuzzy set theory (Qizilbash2002), these involve a doomed attempt toshoehorn them into a binary, two-tailed form.Thirdly, poverty judgements are complex,theory-rich and layered interpretations, notsimply of one aspect of a person or group’sexistence (how much they earn, for instance)but of complex and dynamic states of wellbeingor suffering. Though those ‘states ofbeing’ typically involve aspects of deprivation,some of which may be quantifiable, these aremoments in a complex non-linear interactiveprocess – ‘transient elements in the movingnow,’ as Bevan (2004:28) puts it – a process inwhich they figure both as momentary outcomesof complex interactions and as determinantsof further interactions. What is central inunderstanding people’s prospects and situationsis not any particular aspect of deprivationbut how all the facets of their existence andexperience come together in a complex, andalways historically situated, way to produce astate of lack, powerlessness or need which canthen (always in a particular context and alwaysby particular people) be called poverty.Econometric definitions of poverty on theother hand are, as Bevan (2004) has pointedout, measurement-based, relying on theinterpretation of ‘indicators’ which in turnare created through abstracting and isolatingparticular elements of people’s overallsituation from the broader context in whichthey exist and assigning meanings to themin their own right. This is a tricky enterprise,in which a lot depends on the ability to usethose indicators in an informed way; and it isparticularly dangerous in flagging a conditionsuch as ‘poverty’, which is highly complex,comprising a number of different determinants,mechanisms and long term trajectories. Inpractice the definition of poverty is essentiallycollapsed into its indicator – and the indicatorthen taken for the condition it tries to measure:a circular operation that directs attention awayfrom a concern with the complex underlyingcausal dynamics that link particular aspects ofdeprivation with the social experience of lackdisempowerment, need and suffering.Capabilities andmultidimensionalityThe problems pointed out here apply mosttrenchantly – and most obviously – to thatmost familiar of ‘poverty indicators’: incomeor expenditure measured at household level.3

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South AfricaOne approach that attempts to transcend someof the limitations of this approach involvesa focus, deriving from the work of AmartyaSen, on ‘multidimensional’ poverty and onpeople’s ‘capabilities’. Sen famously arguedthat the study of poverty should focus, not onattempting to measure income and expenditure,but on the underlying capabilities withoutwhich it is not possible to live a fully humanlife (Sen 1999; Nussbaum 1999). This offersthe potential for an account of poverty that isalive to the complex and time-bound dynamicsof deprivation, suffering and need. Althoughthe capabilities approach has fundamentallychallenged some of the underlying assumptionsof welfare economics, its implications haveonly been followed through in limited ways.Sen’s framework is notoriously hard tooperationalise (see Chiappero-Martinetti2000), and many attempts at operationalisationhave fallen afoul of similar problems tothose described in the previous section.Typically, attempts to put it into practice haveinvolved identifying various capabilities (forexample, health, nutrition, education, politicalparticipation), matching these to quantifiableindicators (longevity, anthropometricmeasurements, school enrolments, democraticinstitutions), and then trying to assess whetheror not people are deprived according to thesecriteria (see UNDP 2002; Barrientos 2003;Klasen 2000; Qizilbash 2004; McGillivray2003). This can shed valuable additional lighton the extent and nature of poverty, makingvisible aspects of deprivation not discernablefrom an income perspective alone – butultimately the underlying problem has notbeen transcended. McGillivray (2003), forinstance, has endeavoured to use correlationsbetween ‘non-economic dimensions of wellbeing’(life expectancy, adult literacy, grossschool enrolment) to empirically identify ‘thevariation not accounted for by income percapita’, and then taking this variation as an‘aggregate measure of non-economic wellbeing’– assuming, in other words, that thereis some abstract thing called ‘non-economicwell-being’ which all these indicators partlymeasure. Another, less extreme exampleis again Baulch and Masset (2003), whounderstand the idea that ‘monetary and nonmonetaryindicators of poverty tell differentstories about chronic poverty’ to mean thatthere are ‘different subgroups’ of the chronicpoor, or even different kinds of chronic poverty(for example ‘nutritional poverty’, ‘chroniceducation poverty’ – Baulch & Masset 2003:449–50).Aside from the conceptual difficultiesinvolved in describing capability deprivationin this way (how can hunger, for example, bedescribed as ‘non-economic’?) this approachproduces intractable problems when used totry to identify ‘the chronic poor’ on the basisof panel studies. Are ‘the chronic poor’ onlythose who show up as deprived every timealong every dimension measured? If we donot wish to adopt such a rigorous criterion,should we disaggregate ‘the chronic poor’ into‘the chronic monetary poor’, ‘the chronicallymalnourished’, and so on? And how are weto understand the difference between thosewho are deprived in ‘only one’ dimensionand those who suffer multiple forms ofdeprivation? (Is someone who is educationallydeprived, chronically sick and food-insecurethree times as poor as someone who is justfood-insecure? Is someone who is deprivedin two ‘dimensions’ as poor as someone whois deprived in two others?) These seem likesilly questions, but they are precisely theones that arise in any attempt to develop anaggregate multidimensional poverty score, orto rank poor people – activities that are routinein econometric approaches to poverty (seeAtkinson 2003; Bourguignon & Chakravarty2003).Surely all this misses Sen’s point? Therelationship between human capabilitiesand the ‘full human life’ that they enable iscomplex and dynamic. To treat the absenceof a particular capability, or the lack ofaccess to the resources required for it, as an‘indicator’ of ‘poverty’ is to reify it and to missits significance. Those who lack educationare not suffering from ‘education poverty’;and those who have poor health are not ‘thehealth poor’. They are caught in a process oflack, deprivation or suffering which may ormay not lead to a severe impairment of theiragency and functioning in the world – and thedifferent dimensions of their deprivation reflectthe diverse material roots and determinants ofthat state. It may well be that those who aredeprived in more dimensions than one are lesslikely to escape poverty – but this dependson the local structural context and the actual,empirical ways in which different aspects of4

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africadeprivation play into and feed into one another.The significance of a variation, for example,in literacy or access to water lies not in thefact that they are ‘indicators’ or transparentreflections of ‘non-economic well-being’, butin their implications and consequences forwhat people can do – which are always shapedby a dynamic and complex interplay, andwhich are irreducibly different and thereforenon-substitutable.5

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africa3. From chronic to structuralpovertyVulnerability, agency andstructural povertyIf we want to identify ‘the chronic poor’and understand what keeps them poor,measurement-based approaches, then, offeronly a slippery grasp. A different approachis to recognise that many of those who showup as ‘transitorily poor’ in a panel studymay still be held to be chronically poor iftheir underlying situation – the way they arestructurally inserted in society – means thatthey are unlikely to get out of poverty inthe long run. Such an approach requires anengagement with the causal dynamics andprocesses that drive and shape livelihoodcareers. Understanding who is likely to sinkinto poverty, who is likely to stay out of itfor long periods of time, and who is able tomake the investments required to ensure thata subsequent generation gets out (or stays out)of it requires not only the post-hoc trackingof actual welfare over time, but also anassessment of the underlying factors that shapetheir likely welfare. This means that the studyof chronic poverty – and the identification ofthe chronically poor – is inseparable from thestudy of structural poverty and vulnerability.Development economics and econometricsare not disciplines well geared towardsunderstanding the structural configurationsof vulnerability. Sen’s approach and thepresently popular ‘livelihoods framework’ atleast orient enquiry towards an explorationof the material systems that underlie povertyand well-being – but even they offer scantguidance, partly because they offer abstract anddecontextualised methods of thinking aboutthe particular ways in which individuals andgroups are situated in society.Here, it may be instructive to look at oneof the more innovative attempts in SouthAfrican poverty scholarship to use econometricanalysis to develop an assessment, not simplyof whether or not people are poor, but of theirunderlying ‘structural poverty’: Carter andMay’s (2001) analysis of the KwaZulu-NatalIncome Dynamics Study panel dataset (see alsoCarter & Barrett 2005). Their analysis goeswell beyond the limitations explored above,partly because it uses a component analysisto explore the underlying aspects of people’slivelihood situation. Rather than simply look atincome, expenditure or capability-deprivation,Carter and Barrett look at the assets (land,human capital, financial wealth, social claimsand grain stocks) upon which householdsrely to generate their income. They argue thathouseholds whose assets fall below the levelrequired to generate an income equal to thepoverty line are ‘structurally poor’ even thougha windfall may cause them show up abovethe income poverty line during a particularmeasurement. They further postulate thatthough some people may suffer transitorystructural poverty (in other words, ‘structuralpoverty’ from which is it possible to escapeby accumulating sufficient assets) there maybe a ‘Micawber threshold’ – a level of assetdeprivation so severe it renders escape throughaccumulation impossible.From ‘distributions’ torelationshipsThis is an important corrective to the ‘structureblindness’ in definitions of chronic poverty thatrely on poverty spells. It raises an importantpoint – although the notion of ‘the steadylevel of well-being’ a household ‘can expect’based on a particular level of assets or assetcombination is a useful fiction, a fiction it tosome extent remains. Any attempt to ‘derive’an expected income level from an assessmentof a given household’s asset base will bedogged by uncertainty – particularly if wewant to start including notions like ‘socialcapital’ in that asset base. Although there isa link between the assets over which someonedisposes, and the income one may expect themto generate from it, the link is not linear and ismediated in complex ways by a host of otheroften non-quantifiable factors.6

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South AfricaThis is something not well-recognised in theeconometric approach to poverty and socialunderstanding. For all the innovativenessof their approach, May, Carter and Barrettstill see inequality statistically: as a matterof distribution. This is where the seductivelanguage of ‘household assets’, ‘social capital’and ‘human capital’ becomes dangerouslymisleading. For one thing, households are notnatural units but small, open systems (Bevan,pers. comm.) that are internally contested, thatchange and re-form over time. A household’saccess to resources is powerfully mediated bynetworks and connections that extend outsidethe supposed household boundaries, so thatthere is often not a very clear line betweenhousehold members and non-members (Russel2004; Du Toit; Skuse & Cousins 2005). Foranother, social capital is not a quantifiableresource, like a seed bank or a herd of cattle,which exists in greater or lesser amounts andwhich can be cashed or converted into otherforms of capital in predictable ways. It isa general term for a wide range of variouslystructured human relationships – kinshipnetworks, friendships, affiliation to formal andinformal bodies, patron-client relationshipsand political alliances – that can be used tomake claims and counter-claims (Du Toit,Skuse & Cousins 2005). These, crucially, aremeaningful relationships: deeply informed andshaped by underlying ideologies, moral metanarrativesand cultural paradigms that cometogether to form a more or less consensual orcontested ‘moral economy’ (Thompson 1964;Scott 1985) that defines them and specifieswhich expectations can legitimately be basedupon them.A consideration of the different sociallandscapes explored as part of PLAAS’songoing poverty research in South Africa(see Figure 1) highlights how these complexwebs of relationship and power work in verydifferent ways in different contexts. On thecommercial fruit farms of Ceres, for instance,one very important form of ‘social capital’ isconstituted by the highly racialised patronclientrelationships between the colouredworkers who work on deciduous fruit farmsand the white people who manage and ownthem (Du Toit 2004). These relationshipsare shaped by discourses and practices ofpaternalism that took shape in the courseof a century-and-a-half of slavery and thatFigure 1: PLAAS and CPRC’s research sites in South Africa7

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africaadapted and mutated into new forms inthe course of a century-and-a-half more ofcapitalist modernisation (Du Toit 1993, 1998;Ewert & Hamman 1999; Du Toit & Ewert2005). Paternalist discourse sets in placean underlying ‘moral community’ betweenblack and white that is highly racialisedand hierarchical, which also allows for theformulation of claims for resources andprotection dependent on personal historiesof loyalty and service; and which requiresa complex politics of moral suasion, hiddenresistance and subtle negotiation beneatha façade of racial deference. This racialisedideology shapes relationships among whiteand black, between African and coloured andamong the powerful and the powerless even offthe farms (Du Toit 2004). People with highlysimilar levels of ‘asset endowment’ as thelivelihood framework (or Carter & May 2001)would describe them, will have wildly differentfortunes depending on their ability to negotiatethese relationships and to secure their interests.In Mount Frere in the remote Eastern Cape,‘social capital’ is also central – but here whatmatters are complex traditional networks ofkinship and patronage shaped by the historyof the Eastern Cape and the ways in whichthese have adapted and mutated in response tomodernisation and change (Skuse & Cousins2005a; Du Toit, Skuse & Cousins 2005).Social capital is embedded and embodied bya vast, complex, relational economy involvingdomestic fluidity and ‘stretched’ households’(Spiegel, Watson & Wilkinson 1999) extensivetrade in goods, services, favours, labour andsometimes even money; and shaped by morethan a century of migrant labour. The localcultures that shape this relational economy anddefine people’s expectations about themselvesand one another are thoroughly differentfrom those one would find in Ceres: thoughXhosa culture has not persisted unchangedinto modernity, local traditions about identitygender and status, for instance, play a powerfulrole in shaping aspirations and behaviour.Again, households that look very similar ina livelihood survey can have very differentfortunes depending on their links to local elites,their ability to make claims and to exploitsometimes tangential kinship networks.In Cape Town’s ‘African’ suburbs (DeSwardt 2004b; Skuse & Cousins 2005b)survival also depends on an informalrelational economy, but here things workvery differently again. Kinship is important,but it is only one of a wide range of socialrelations, affiliations, alliances and enmitiesthat structure and are structured by informalexchange. Xhosa cultural forms and practicesare still important, but the ethos is much lessshaped by traditionalism and is infused withan assertive, street-smart urbanity (Skuse& Cousins 2005b). What matters here is theability to ‘work’ the urban system to get accessto social services; the ability to juggle debtsand obligations and the ‘politics of intimacy’ inthe dance of the relational economy; the abilityto manage risk and violence, and the ability tointerface effectively with white society and theformal economy. The ability to insert oneselfin complex local development processes; theability to claim membership of particular subcommunitiesand interest groups; one’s historyof belonging in Khayelitsha and the alliancesand allegiances thus formed all have a majorimpact on the resources one can mobilise.In all three of these contexts, the local logicof social capital leads to the identification ofvery different groups of vulnerable people.In Mount Frere, for instance, women andgirl children bear the brunt of the impact ofgender roles that assign them most of theresponsibility for household reproduction (DuToit, Skuse & Cousins 2005). At the sametime, those gender roles have given them, aftermore than a century of migrant labour, a veryreal centrality in the networks of civil society,while young men are no longer as able to usemigrancy as a path to full adult manhood.A different vulnerable group is comprised byolder people who end up being the heads ofHIV/Aids-affected households. In urban CapeTown, it is young women who are particularlyat risk, partly because gender roles dictate thatthey should be dependent on men. In Ceres,African men and women are disadvantaged bylocal culture that constructs them as outsiders(Du Toit 2005b).Clearly an attempt to deduce ‘expectedincomes’ from ‘asset combinations’ byrunning regressions on household surveydata stands a poor chance of uncovering anyof this complexity. The point is not merelythat there is plenty that does not show up onthe radar of any particular dataset. It is alsothat the incorporation of these additionalfactors involves, not merely their additionto an existing analysis of correlations, but8

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africathe development of a critical theoreticalaccount of power, ideology, culture andinequality in these contexts. The thumbnailsketches provided above derive from ananalysis informed by family of (Geertzian andFoucauldian; agent-centred and structuralist)theoretical frameworks very different fromthe econometric one – a theoretical imaginarythat emphasises the role of structure, agency,antagonism and social change. According tothese the perspectives and stated experiencesof various social actors are not taken simplyat face value but seen as complex socialcreations, shaped by social power relations andin turn impacting upon them. This has crucialimplications for the prospects of building morerobust accounts of the nature of structural andchronic poverty.9

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africa4. Poverty measurement and thegovernment of povertyBeyond Q-squaredIn one sense, of course, none of the abovearguments are very new. Arguments about thelimitations of purely quantitative research areprobably as old as ‘quantitative social science’itself and have recently become commonplaceagain even within the development mainstream(see Kanbur 2002). This recognition has,however, usually taken quite a limited form– being confined, for instance, to the idea thatit is enough for ‘quantitative approaches’ tobe supplemented, corrected or added to insome way by ‘qualitative’ research. This isundoubtedly a good thing: forays into ‘qsquared’,attempts to integrate qualitativeand quantitative work, clearly adds to therigour, depth, reach and accuracy of povertyresearch (see Adato, Lund & Mhlongo 2004).At the same time, this recognition is oftenquite circumspect and the integration between‘qualitative’ and ‘quantitative’ often takesplace in restricted ways. For writers likeThorbecke, for instance, qualitative data seemsto be understood as being equivalent to doingsome PRAs (participatory rural appraisals)– and the role of qualitative data seems tobe limited to generating hypotheses that canbe quantitatively tested (Thorbecke 2004).Others admit of a wider range of methodsand highlight a number of different ways inwhich qualitative and quantitative work canillustrate, confirm, refute, enrich and illuminateone another (Carvalho & White 1997; seealso Howe & McKay 2004; Adato, Lund& Mhlongo 2004). On the whole, however,‘qualitative data’ has been seen to have anessentially supplementary and illustrativerole in accounts of poverty still essentiallyshaped by the econometric imaginary. Evenmore problematically, ‘qualitative data’itself is almost universally understood,very simplistically, as if it is transparentlymeaningful in itself. It is as if what emergesfrom PRAs, life histories, focus groups andthe like can be taken at face value, without anengagement with the need to interpret these astextual artefacts, themselves the products ofconflicts, antagonisms and other encountersthat are shaped by social power relations andconcrete social interests. There is, furthermore,very little reflexive awareness of the processof research itself and how this shapes the way‘qualitative data’ is produced, analysed andinterpreted.There is a danger, therefore, that attemptsto assert the value of qualitative research cansimply take us back to a new positivism, inwhich slightly more methodologically diverseresearch strategies (household surveys plusfocus group interviews; panel data sets plus lifehistories; econometric regressions plus PRAs)figure within accounts of society and socialchange – essentially still caught within the a-historical, power-blind, technicist and rationalchoiceimaginary of econometric analysisand mainstream development economics (seeKothari 2001). What the calls for ‘integration’ignore is that the real issue is not whether weneed to connect qualitative and quantitativeresearch – we obviously do – but that anyattempt at integration is always theory-rich,utterly dependent on underlying narrativesabout the nature of society, agency, power,poverty and social change.Two issues arise out of this observation.The first is that this need not be so. If thepurpose is indeed to understand chronic (andtherefore structural) poverty, and to understandhow social relations shape people’s chancesof getting into or out of poverty, the field ofsocial science and critical social theory offerswide and deep resources. A veritable academicindustry exists in which the links betweenpower, agency, culture, identity and historyare explored and which offers wide space forreflective and incisive accounts of the waysin which these are linked to the distribution ofresources in society.The second is that, in spite of thispromise these critical traditions are to a largeextent marginalised in the field of applied10

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africadevelopment and poverty studies, relegatedto a fairly well-defined circuit of institutionsand journals which development economistsand poverty scholars seem to feel they cansafely ignore. In South Africa, for instance,there is a rich legacy of critical debate andresearch dating from the 1980s and 1990s onthe relationships between capital accumulation,identity, ideology, social change and inequality– a legacy that has been radicalised andextended more recently, in the work ofinstitutions such as WISER (Wits Institute forSocial and Economic research), into searchingreflections on postcoloniality, racism andidentity. At the same time it is possible forscholars (who, to all intents and purposes,are clearly deeply committed to social justiceand the eradication of inequality and poverty)to produce an account of labour marketvulnerability and poverty in South Africaignores the ‘revisionist’ debates of the 1980sthe criticism of liberal orthodoxy of the 1960sand 1970s (Bhorat et al. 2001; for a discussionsee Du Toit 2005a).Power, knowledge andmethodological dissentWhat is the scope for this state of affairs to bechallenged and for applied social science ingeneral (and policy-oriented poverty researchin particular) to become more sensitive tothe need for – and the power of – critical andagent-centred accounts of structural povertyand the prospects for getting out of it? Inmy own recent work (see Du Toit 2005a),I have, to some extent, attempted to nameand problematise the marginalisation I havedescribed here, which is all too often seen asthe natural order of things.One of the most prominent stated reasonsfor the failure of critical social theory toseriously challenge the hegemony of theeconometric imaginary is that there is noclearly hegemonic ‘critical theory’ approach.In contrast to the field of economics andeconometrics, where debates and discussionsare underpinned by a widely shared andhegemonic framework setting the boundariesof a generally accepted ‘normal science’ (andalso, in contrast, to the field of developmentstudies, which lacks its own rigour butis thoroughly governed by the changingorthodoxies and frameworks adopted byleading donor institutions), critical socialtheory and anthropology has since the mid-1980s been characterised by a floweringof increasingly different and sometimescompeting explanatory paradigms andontologies, sub-disciplines and specialities(postcolonial, gender and cultural studies,social constructionism, critical realism, poststructuralisttheory and discourse analysis,actor network theory, agent-centred theories,global value chain analysis, convention theory,to name but a few), with no particular approachsucceeding in establishing itself as centralor dominant. Norman Long has argued that,rather than being seen as fragmentation andcrisis, this diversity should be recognised asa fundamental condition of social enquiry; andwelcomed as an opportunity for innovation(Long 1992). Nevertheless, this diversitymeans that there is no single generally accepted‘qualitative’ or ‘non-positivist’ or ‘postfoundational’approach. Calls by economistsfor examples of generally accepted ways inwhich social theory can help us understandchronic poverty have to be met by the answerthat there is no master paradigm. Any attemptto ‘operationalise’ the insights of qualitativesociology and critical social theory has to bepartial and local, and will require the case-bycasetheoretical concepts and approaches thatcan help illuminate particular problems.This is, of course, only part of the story,the demand for ‘normal science’ in socialresearch – for powerful, uncontroversialand replicable methodologies and schemasthat can be used to produce reliable, policyrelevantknowledge about poverty – has itsown political economy. Paradoxically, manyof the most problematic features of povertymeasurement described in previous pagesare precisely those that make it attractive togovernments and donor institutions. Some ofthe crucial operations I have criticised abovearise to some extent out of the underlyinglogic of the social technologies of knowledgeand power which make poverty measurementnecessary and possible as an enterprise inthe first place. Poverty measurement has acomplex history, but a very important role inthis history has been played by what we mightcall the historical project of the ‘government ofpoverty’. The need for universal measurementsand easily replicable indicators is indissolublylinked to the project of constituting poverty asan object of management and government – assomething whose presence in society needs to11

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africabe recognised in ways that render it subject toregulation; and which can contain and limitits present as a radically disruptive politicalproblematic.As such, the discipline of povertymeasurement is caught on the horns ofa dilemma or a double bind: like the ‘optics’of modern government identified by JamesC Scott (Scott 1998; but see also Foucault1987), it is partly driven by the need to makesociety ‘legible’ in a regular, homogenous anduniversalising way. In order to be useful forthe process of government and planning at all,technologies of measurement and assessmenthave to be developed that can be treated asindependent, or which can be delinked fromthe complexity and non-transparency of localcontext. Economies of scale in governmentjudgement and assessment require thedevelopment of embodied techniques ofregularised knowing and decision-making. Itshould be possible to ‘port’ such techniquesfrom one context to another so that oneindividual (or household or region) can becompared with another and ranked so that gooddecisions can be made about the allocationof resources. Such tools make it possible foroperations to be done in the shadow of theauthority of ‘science’ – apparently free of bias,objective and incontrovertible.The problem, as Scott points out, ariseswhen this process of abstraction and decontextualisationleads not to legibilitybut to misreading. When, for example,imposing the template of monoculture onforestry management destroys the underlyingecological base of biodiversity on which theforest depends, or when dirigiste city plannersmisunderstand the local dynamics that makeneighbourhoods viable. In such cases thepreference for certain kinds of information– information that is readily quantifiableand standardised, that abstracts from localcomplexity and appears to sidestep nontransparency– leads not to an accurate graspof the dynamics of a situation, but to distortedand misleading accounts that miss crucialdynamics.The question is what follows fromthe recognition of these distortions andmisunderstandings. What scope is therefor what Scott called metís – for forms ofknowledge that allow for an understanding ofsome of these complex dynamics and whichare by their very nature more provisional,more embodied and localised, more connectedwith specific histories and relationships, morevalue laden and political? What scope is therefor the state to learn other ways of seeing andimagining poverty and vulnerability?The struggle is an uphill one, if recentattempts to build governmental capacity tounderstand food insecurity and vulnerabilityin South Africa are anything to go by. A casestudy of the development of a Food Insecurityand Vulnerability Information and MappingSystem (FIVIMS) for the ‘social cluster’ ofdepartments in South Africa shows that, inspite of the recognition of the role of localhistory and power relations – and in spiteof the acknowledgement of the importanceof practical local knowledge embedded ininstitutions on the ground – very little couldbe done to shift the perception on the part ofthe officials involved that ultimately, whatwas practical was a geographical informationsystems (GIS)-based system that wouldprovide information about ‘indicators’ of‘structural vulnerability’ in unambiguous,map-able, quantifiable terms. (It is hoped thatthis experience will be explored in a separatecase study – Du Toit, Vogel and Ziervogel;forthcoming.) This institutional inertia seemedto be produced partly by what one could callthe mystique of ‘quantitative data’ – a whollymisplaced faith in what one could learn fromthe ‘quantitative data’ that is available for usein a GIS-based system – but partly also byunderlying totalising narratives about the placeof ‘integrated planning and implementation’and centralised knowledge in the exercise ofstate power (Du Toit, Vogel & Drimie 2005).Asking governments and donor institutionsto make space for critical accounts of socialchange – accounts that are more sensitive tothe nature and dynamics of power relations– seem inevitably to come up against thelimitations that arise out of the present-daylogic of forms of power-knowledge and modesof governmentality that seek to de-link claimsto authority from knowledge from locality;and that depend on spatialised technologies fordecontextualising and homogenising social andpolitical space (for a broader discussion seeKothari 2005; Duffield 2004 & 2005).12

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africa5. ConclusionWhat, then, is the scope for ‘decolonising’methodologies that are so clearly linkedto formations of power and knowledge sodeeply shaped by their links to post-colonialand still-imperial forms of governance andgovernmentality? In the long run, there is onlyone way of finding out: by actually trying tocontest homogenising quantitative narrativesby developing powerful and convincingcounter-hegemonic accounts. In South Africa,at least, it is possible to imagine that the termsof this engagement do not run only one way.Rather than being the stage for a seamless‘ordering of dissent’ in which the institutionsof globalised corporate power are alwaysand inevitably able to contain criticism byincorporating it, the field of applied socialscience research in South Africa seem toembody a fruitful, if hazardous terrain forengagement. Given the urgency of addressingpersistent poverty in South Africa and thedawning recognition by the ruling party thatmodernising narratives about ‘trickle-down’are not working (Mbeki 2003), there is a widescope for critical scholars to interrupt and toproblematise the apparent self-evidence ofnormalising meta-narratives about growth,modernity, security and the like. It is partof both the fertility and the hazard of thisterrain that all such interventions needs mustbe themselves situated and informed by anawareness of their own dependency on andinevitable complicity with a history steeped inconflict and suffering.13

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South AfricaReferencesAdato, M, Lund, F & Mhlongo, P. 2004.Methodological innovations in researchon the dynamics of poverty: A mixedmethod longitudinal study in KwaZulu-Natal, South Africa. Paper presented atQ-squared in practice: A conference onexperiences of combining qualitative andquantitative methods in poverty appraisal.http://www.utoronto.ca/mcis/q2/papers/IV_Adato_Lund_Mhlongo.pdfAliber, M. 2001. Study of the incidence andnature of chronic poverty and developmentpolicy in South Africa: An overview.Manchester: Chronic Poverty ResearchCentre. (CPRC working paper; no.3.)www.chronicpoverty.org/pdfs/sa.pdfArnall, A, Furtado, J, Ghazoul, J & De Swardt,C. 2004. Perceptions of informal safetynets: A case study from a South Africaninformal settlement. Development SouthernAfrica, 21(3): 443–60.Atkinson, AB. 2003. Multidimensionaldeprivation: Contrasting social welfare andcounting approaches. Journal of EconomicInequality, 1:51–65.Bane, MJ & Ellwood, DT. 1986. Slipping intoand out of poverty: The dynamics of spells.Journal of Human Resources, 21(1):1–23.Barrientos, A. 2003. Non-contributorypensions and the well-being of olderpeople: Evidence on multi-dimensionaldeprivation from Brazil and SouthAfrica. Mimeo. Manchester: Institute forDevelopment Policy and Management,University of Manchester.Baulch, B. 1996. Neglected trade-offs inpoverty measurement. IDS Bulletin, 27(1):36–42.Baulch, B & McCulloch, N. 1998. Beingpoor and becoming poor: Poverty statusand poverty transitions in rural Pakistan.Brighton: Institute for DevelopmentStudies, University of Sussex. (IDS workingpaper; no. 79.)Baulch, B & Masset, E. 2003. Do monetaryand non-monetary indicators tell the samestory about chronic poverty? A study ofVietnam in the 1990s. World Development,31(3):441–54.Bevan, P. 2004. Exploring the structureddynamics of chronic poverty: A sociologicalapproach. Bath: Wellbeing in DevelopingCountries ESRC research group,University of Bath. (WeD working paper;no. 06.) www.welldev.org.uk/research/working.htm#wed6Bhorat, H, Leibbrandt, M, Maziya, M, Van derBergh, S & Woolard, I (eds). 2001. Fightingpoverty: Labour markets and inequality inSouth Africa. Cape Town: University ofCape Town Press.Bourguignon, F & Chakravarty SR. 2003. Themeasurement of multidimensional poverty.Journal of Economic Inequality, 1:25–49.Carter M, & May, J. 2001. One kindof freedom: Poverty dynamics inpost-apartheid South Africa. WorldDevelopment, 29(12):1987–2006.Carter, M & Barret, C. 2005. The economicsof poverty traps and persistent poverty:An asset-based approach. Ithaca: CornellUniversity. (Cornell Food and NutritionPolicy Programme working paper; no. 178.)www.cfnpp.cornell.edu/images/wp178.pdfCarvalho, S & White, H. 1997. Combining thequantitative and qualitative approachesto poverty measurement and analysis: Thepractice and the potential. Washington, DC:World Bank. (World Bank technical paper;no. 366.) www.wds.worldbank.org/servlet/WDS_IBank_Servlet?pcont=details&eid=000009265_3971031092544Chaudhuri, S & Ravallion, M. 1994. How welldo static indicators identify the chronicallypoor? Journal of Public Economics, 53:367–94.Chiappero-Martinetti, E. 2000.A multidimensional assessment of wellbeingbased on Sen’s functioning approach.Revista Internazionale di Scienza Soziale,108(2):207–39.14

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South AfricaCichello, PL, Fields, GS & Leibbrandt,M. 2003. Earnings and employmentdynamics for Africans in post-apartheidSouth Africa: A panel study of KwaZulu-Natal. Cape Town: Development PolicyResearch Unit. (DPRU working paper; no.03/77.) www.commerce.uct.ac.za/dpru/WorkingPapers/wp.asp?WP_ID=2003/77CPRC (Chronic Poverty ResearchCentre). 2004. The chronic povertyreport 2004–05. Manchester: CPRC.www.chronicpoverty.org/De Swardt, C. 2004a. Chronic poverty andthe basic income grant. Cape Town:Programme for Land and Agrarian Studies,University of the Western Cape. (Chronicpoverty and development policy workingpaper; no. 2.) www.uwc.ac.za/plaas/publications/poverty.htmDe Swardt, C. 2004b. Cape Town’s Africanpoor. Cape Town: Programme for Landand African Studies, University of theWestern Cape. (Chronic poverty anddevelopment policy working paper;no.3.) www.uwc.ac.za/plaas/publications/poverty.htmDu Toit, A. 1993. The micropolitics ofpaternalism: Discourses of management andresistance on Western Cape fruit and winefarms. Journal of Southern African Studies,19(3):314–36.Du Toit, A. 1998. The fruits of modernity:Law, power and paternalism in WesternCape fruit and wine farms, in South Africain transition: New theoretical perspectives,edited by A Norval and D Howarth.London: Macmillan:149–64.Du Toit, A. 2004. Forgotten by the highway:globalization and chronic poverty inCeres. Cape Town: Programme forLand and Agrarian Studies, Universityof the Western Cape, Centre for SocialScience Research, University of CapeTown and Chronic Poverty ResearchCentre. www.chronicpoverty.org/pdfs/49%20duToit.pdfDu Toit, A. 2005a. Chronic and structuralpoverty in South Africa: Challengesfor action and research. Cape Town:Programme for Land and Agrarian Studies,University of the Western Cape. (Chronicpoverty and development policy series; no.6.)Du Toit, A. 2005b. Beyond Q-squared: Usingqualitative and quantitative approaches tounderstand structural and chronic povertyin South Africa. Unpublished typescript.Du Toit, A & Ewert J. 2005. A deepeningdivide in the countryside: Restructuringand rural livelihoods in the South Africanwine industry. Journal of Southern AfricanStudies, 31(2):315–32.Du Toit, A, Skuse A & Cousins, C. 2005. Thepolitical economy of social capital: Chronicpoverty, remoteness and gender in the ruralEastern Cape. Unpublished typescript;Adelaide and Cape Town.Du Toit, A, Vogel, C & Drimie, S. 2005.Institutional arrangements and utilization ofFIVIMS-ZA. Unpublished typescript.Du Toit, A, Vogel, C & Ziervogel, C.2005. Maps in the dark: Bio-power,‘vulnerability’ and the crisis of integratedplanning. (Forthcoming, provisional title)Duffield, M. 2004. Carry on killing: Globalgovernance, humanitarianism and terror.DIIS Working Paper 2004/23, Copenhagen:Danish Institute for International Studies.www.diis.dk/graphics/Publications/WP2004/duffield_carry_on_killing.pdfDuffield, M. 2005. Getting savages to fightbarbarians: Development, security and thecolonial present. Conflict, Security andDevelopment, 5(2). (Forthcoming)Ewert, J & Hamman, J. 1999. Why paternalismsurvives: Globalisation, democratisationand labour on South African wine farms.Sociologia Ruralis, 39(2):202–21.Foucault, M. 1987 [1977]. Disciplineand punish: The birth of the prison.Harmondsworth: Penguin.Howe, G & McKay, A. 2004. Combiningquantitative and qualitative methods inassessing chronic poverty: The case ofRwanda. Paper presented at ‘Q-Squared inPractice: A Conference on Experiences ofCombining Qualitative and QuantitativeMethods in Poverty Appraisal’, Universityof Toronto, 15–16 May. www.utoronto.ca/mcis/q2/papers/VI_Mckay_Howe.pdfHulme, D & Shepherd, A. 2003.Conceptualizing chronic poverty. WorldDevelopment, 31(3):403–24.15

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South AfricaKanbur, R. 2002. Economics, social scienceand development. World Development,30(3):477–86.Klasen, S. 2000. Measuring poverty anddeprivation in South Africa. Review ofIncome and Wealth, 46(1):33–58.Kothari, U. 2001. Power, knowledge and socialcontrol in participatory development, inParticipation: The new tyranny, edited by BCooke and U Kothari. London: Zed.Kothari, U. 2005. Authority and expertise:The professionalisation of internationaldevelopment and the ordering of dissent.Antipode, 37(3):425–46.Kritzinger, A & Barrientos, SW. 2004.Squaring the circle: Global productionand the informalization of work inSouth African fruit exports. Journal ofInternational Development, 16:81–92.Leibbrandt, M & Woolard, I. 2001. Measuringpoverty in South Africa, in Fightingpoverty: Labour markets and inequalityin South Africa, edited by H Bhorat, MLeibbrandt, M Maziya, S van der Bergh andI Woolard. Cape Town: University of CapeTown Press.Long, N. 1992. From paradigm lost toparadigm regained: The case for an actororiented sociology of development, inBattlefields of knowledge: The interlockingof theory and research in social theoryand development, edited by N Longand A Long. Wageningen: WageningenAgricultural University:17–42.Mbeki, T. 2003. Letter from the President.www.anc.org.za/ancdocs/anctoday/2003/at33.htm#presletMcCulloch, N & Baulch, B. 1999.Distinguishing the chronically from thetransitorily poor: Evidence from Pakistan.Brighton: Institute of Development Studies,University of Sussex. (IDS working paper;no. 97.) www.ids.ac.uk/ids/bookshop/wp/wp97.pdfMcGillivray, M. 2003. Capturing noneconomicdimensions of human well-being.Paper presented at the World Institutefor Development Economics ResearchConference on Inequality, Poverty andHuman Well-Being, Helsinki, 30-31 May.www.wider.unu.edu/conference/conference-2003-2/conference%202003-2-papers/papers-pdf/McGillivray%20280503.pdfMcKay, A & Lawson, D. 2003. Assessingthe extent and nature of chronic poverty inlow income countries: Issues and evidence.World Development, 31(3):425–40.Noble, M, Ratcliffe, A & Wright, G. 2004.Conceptualising, defining and measuringpoverty in South Africa: An argument fora consensual approach. Oxford: Centre forthe Analysis of South African Social Policy,University of Oxford.Nussbaum, M. 1999. Women and equality: Thecapabilities approach. International LabourReview, 138(3):227–45.PASS (Programme of Advisory and SupportServices to DFID). 2003. Review of theVAC methodology for assessing foodneeds. www.passlivelihoods.org.uk/default.asp?project_id=77Qizilbash, M. 2002. A note on themeasurement of poverty and vulnerabilityin the South African context. Journal ofInternational Development, 14:757–72.Qizilbash, M. 2004. On the arbitrariness androbustness of multi-dimensional povertyrankings. Journal of Human Development,5(3):355–75.Ravallion, M. 1996. Issues in measuring andmodelling poverty. The Economic Journal,106:1328–43.Russel, M. 2004. Understanding blackhouseholds in southern Africa: The Africankinship and Western nuclear family systems.Cape Town: Centre for Social ScienceResearch, University of Cape Town. (CSSRworking paper; no. 67.) www.cssr.uct.ac.za/papers/wp67.zipScott, JC. 1985. Weapons of the weak:Everyday forms of peasant resistance. NewHaven: Yale University Press.Scott, JC. 1998. Seeing like a state: Howcertain schemes to improve the humancondition have failed. New Haven: YaleUniversity Press.Sen, A. 1999. Development as freedom.Oxford: Oxford University Press.Skuse, A & Cousins, T. 2005a. Managingdistance: The social dynamics of rural16

Poverty measurement blues: Some reflections on the space forunderstanding ‘chronic’ and ‘structural’ poverty in South Africatelecommunications access and use in theEastern Cape, South Africa. Unpublishedtypescript submitted to Journal of Africanand Asian Studies (under review).Skuse, A & Cousins, T. 2005b. Gettingconnected: The social dynamics of urbantelecommunications use in Khayelitsha,Cape Town. Unpublished typescriptsubmitted to New Media and Society (underreview).Spiegel, A, Watson, V & Wilkinson, P. 1999.Speaking truth to power: Some problemsusing ethnographic methods to influencethe formulation of housing policy in SouthAfrica, in The anthropology of power:Empowerment and disempowerment inchanging structure, edited by A Cheater.London and New York: Routledge:175–90.Thompson, EP. 1964. The making of theEnglish working class. Harmondsworth:Penguin.Thorbecke, E. 2004. Conceptual andmeasurement issues in poverty analysis.Helsinki: World Institute for DevelopmentEconomics Research. (WIDER discussionpaper; no. 2004-004.) www.wider.unu.edu/publications/dps/dps2004/dp2004-004.pdfUNDP (United Nations DevelopmentProgramme). 2002. Human developmentreport 2002. New York: Oxford UniversityPress. http://hdr.undp.org/reports/global/2002/en/17