Anthony Hall From to Bolsa Família: social policies and poverty alleviation under Lula

Article (Published version) (Refereed)

Original citation: Hall, Anthony (2006) From Fome Zero to Bolsa Família: social policies and poverty alleviation under Lula. Journal of Latin American studies, 38 (4). pp. 689-709. DOI: 10.1017/S0022216X0600157X

© 2006 Cambridge University Press

This version available at: http://eprints.lse.ac.uk/21244/ Available in LSE Research Online: September 2012

LSE has developed LSE Research Online so that users may access research output of the School. Copyright © and Moral Rights for the papers on this site are retained by the individual authors and/or other copyright owners. Users may download and/or print one copy of any article(s) in LSE Research Online to facilitate their private study or for non-commercial research. You may not engage in further distribution of the material or use it for any profit-making activities or any commercial gain. You may freely distribute the URL (http://eprints.lse.ac.uk) of the LSE Research Online website. j Lat. Amer.Stud 38, 689-709 ? zoo6 Cambridge UniversityPress 689 doi:Io.IoI7/SoozzzI6Xo6oox 57X Printedin the United Kingdom FromFome Zero to BolsaFamilia: Social Policiesand PovertyAlleviation underLula*

ANTHONY HALL

Abstract.Under the administrationsof Fernando Henrique Cardoso (I995-2002) andespecially President Lula (2003-),conditional cash transfer (CCT) programmes havebecome adopted as mainstreamsocial policy in .This followsa marked trendsince the 1990sin LatinAmerica towards the settingup of targetedsafety nets to alleviatepoverty. Lula consolidatedand expandedCCTs, firstlyunder FomeZero and later Bolsa Familia, now the largest such scheme in theworld. Its four sub-programmes(educational stipends to boost school attendance,maternal nutrition,food supplements and a domesticgas subsidy)benefit some 30 millionof Brazil'spoorest people, with a targetof 44 millionby 2006. Since 2003, spending on BolsaFamilia has risensignificantly to consume over one-third of thesocial assist- ancebudget for the poorest sectors and it remaineda flagshippolicy in therun-up to thepresidential elections of October2006. Although coverage of BolsaFamilia is impressive,however, systematic evaluation of its social and economicimpacts is stilllacking. Evidence from other CCT programmesin LatinAmerica suggests thatpositive results may be achievedin termsof meetingsome immediate needs of thepoor. However, there have been many implementation problems. These include poor beneficiarytargeting, lack of inter-ministerialcoordination, inadequate monitoring,clientelism, weak accountabilityand allegedpolitical bias. Given the heightenedprofile of cash transfersin Brazil'ssocial policy agenda, key questions need to be asked.These concern,firstly, the extentto whichBolsa Familia does indeedcontribute to povertyalleviation; and secondly,whether it createsgreater dependenceof thepoor on governmenthand-outs and politicalpatronage at the expenseof long-termsocial investment for development.

Introduction

When PresidentLuis Inicio Lula da Silva took officein January2003, Fome Zero (Zero Hunger) soon became the defining flagshippolicy of his government.Inspired by the MillenniumDevelopment Goals, which call AnthonyHall is Reader in Social Planning, Department ofSocial Policy, London School of Economicsand Political Science, Houghton Street, London WC2A 2AE, UK. a.l.hall@ lse.ac.uk * Thisarticle is basedon a paperpresented at theLASA 2004 conference,Las Vegas,NV. The authoris indebtedto severalanonymous JLAS reviewerswho providedvaluable commentson a firstdraft. 690 AnthonyHall forhalving hunger and extremepoverty by zoI 5,the ambitious programme was intendedto bringregular supplies of food and cash aid to Brazil's estimated44 millionliving below the officialpoverty line. Building upon policiesalready established under the administrationof President FernandoHenrique Cardoso, Fome Zero sought to expandand consolidate the social safetynet programme.'It also representeda heartfeltpersonal commitmentby Brazil's first working class president to addressthe country's seeminglyintractable problems of povertyand severeinequality. At his inauguration,Lula famouslypledged: 'If, by theend of myterm of office, everyBrazilian has foodto eat threetimes a day,I shallhave fulfilledmy missionin life.' Brazil's enthusiasmfor constructingsocial safetynets has followed a markedinternational trend in policy-makingsince the late I98os.2The freemarket ideology developed under the Reaganand Thatcheradminis- trations,amongst others, was transferredto the developingcountries throughprogrammes of economicstabilisation and structuraladjustment via theIMF andWorld Bank. Under a varyingpackage of measuressome- timesloosely referred to as the 'WashingtonConsensus',3 moves towards marketderegulation and increasedcompetition would lead to the dis- mantlingof statemachinery, growing privatisation of serviceprovision and cutbacksin publicspending, especially in the social sectors.This process helpedto destroymuch of the socialprotection infrastructure in a region whereeffective and universalwelfare systems had been developedin onlyfive countries (Argentina, Chile, Costa Rica,Cuba and Uruguay),and wherethe impactsof social policieshave generallybeen regressivewith negativedistributional consequences. Higher income groups have secured thelion's share of benefitsand subsidiesin keysectors such as education, healthand pensionsprovision. On cost grounds,universalisation has beendismissed as too expensivein thecontext of developingcountries. In particular,it was arguedthat such generousanti-poverty solutions were not affordableunder conditions of economicausterity. However, even more critically,the notionof universalcitizenship rights has (exceptin the case of Cuba) generallybeen eschewedin favourof selectivesocial

1 'Safetynets' is a broadlabel to describeshort-term, targeted interventions for vulnerable householdsdesigned to mitigate the immediate effects of poverty and other risks, providing assistancein the formof cash,food, housing, subsidies, fee waivers,scholarships and publicworks programmes. 2 A. Hall andJ. Midgley, Social Poliy for Development, Sage (London,2004). 3 J. Williamson,'What Washington means by policyreform', in J. Williamson(ed.), Latin AmericanAdjustment: How MuchHas Happened?Institute for InternationalEconomics (WashingtonDC, 1990). SocialPolicies and Poverty Alleviation under Lula 691 rightslinked to the politicallydetermined distribution of benefitsby the State.4 The solutionfor maximising both economicand politicalreturns from thedistribution of scarcesocial funding has been to targetthe poorest and mostvulnerable groups through the use of specificinstruments such as socialfunds implemented by a rangeof institutionsincluding government, civilsociety, international donors and, indeed, poor communities themselves. This reactionwas in parta pragmaticand understandableresponse for rapidlydealing with the increased suffering of some sectorsbrought about by the adversesocial and economicimpacts of structuraladjustment.5 However,it also representeda major ideological shift in thinkingtowards a moreselective and means-testedapproach for addressing mass poverty. This approachwas formalisedand endorsedat the highestinternational levelsthrough the conceptualisation of social safety nets. These nets would notonly 'catch' thepoor and 'protect'them from economic shocks in the shortterm; they would also providea 'springboard'for development, creatingemployment and income-earning opportunities through strategies of 'social riskmanagement'.6 In LatinAmerica, social safety nets have been enthusiasticallyembraced by manygovernments. In particular,conditional cashtransfer (CCT) programmesare increasingly regarded as a majorsocial policyinstrument in a numberof countries.Since the late 1990s, apartfrom Brazil,major CCT programmeshave been introducedin Mexico(Progresa, now knownas Oportunidades),Colombia (Familias en Accidn), Chile (Subsidio UnitarioFamiliar), Nicaragua (Red de Proteccio'n Social), Argentina (Jefes de Hogar) and Ecuador(Bono de Desarrollo Humano).7 Their underlyingprinciple is thathuman capital can be enhancedas a developmentvehicle by providing money to familiesto persuadethem to investin themselvesthrough greater participation in educationand health services.The technicalrationale is that,by focusingon children,long-term humancapital accumulation can help breakthe inter-generationaltrans- missionof poverty. Through stimulating effective demand for social services, it is also designedto counterthe shortcomings of supply-sideinterventions such as schoolsand clinicsthat often do not reachthe poor. This CCT approach,in theoryat least,contrasts with traditionalsocial assistance

4 C. Abeland C. Lewis,'A Diagnosisof SocialPolicy in LatinAmerica in theLong Run', in C. Abeland C. Lewis(eds), Exclusion and Engagement: Social Policy in Latin America (London, 2002), pp. 3-71. 5 See G. Cornia,et al.,Adjustment with a HumanFace (Oxford, 1987), and T. Conway,Social Protection:New Directions forDonor Agencies, DFID (London,2000). 6 WorldBank, Social Protection Strategy Paper: from Safety Net to Springboard (Washington DC, 2000). 7 L. Rawlings,A NewApproach to Social Assistance: Latin America's Experience with Conditional Cash TransferPrograms, World Bank (WashingtonDC, 2oo4). 692 AnthonyHall strategiesthat use short-termredistributive mechanisms to tacklepoverty duringtimes of crisis.It aimsto fosterjoint responsibility between families and thegovernment, placing the onus on parentsto spendcash wisely and ensureattendance at schoolsand healthclinics. Thus, it is seen as con- stitutinga break from clientelism and, in the words of one World Bank study, 'has been heraldedas an alternativeto more traditional,paternalistic approachesto socialassistance and has helpedcounter criticisms of CCT programmesas handouts'.8However, as thecase ofBrazil considered below illustrates,this is verymuch a matterof contention. Yet CCT programmesare seen as offeringother innovationsand advantagesover traditionalsocial assistance.In additionto encouraging humancapital formation and family responsibility, CCTs arereputedly easier to targetthan other social assistance programmes, using geographical and householdlevel criteria, including proxy-means tests to estimatehousehold povertylevels. Providing cash (ratherthan benefits in kind,food stamps or vouchers)is morecost-effective and flexible,and avoids the creation ofdistorting secondary markets. Furthermore, the focus on health,education and nutritionfosters those synergiesamongst diverse components of humancapital considered essential for breaking the vicious circle of inter- generationalpoverty. Several countries adopting CCTs have also carried outevaluations that have provided empirical evidence of theireffectiveness, facilitatingscaling-up and political-administrative continuity.

Socialsector and social assistance in spendingBrazil Governmentfigures show that social sector spending as a wholein Brazilis quitehigh, at 16 per centof GDP, withfunding of directcosts equivalent to 14.1 per cent of GDP (Table i). In 2004, social security,including pensions,accounted for 60o per cent of direct social spending in Brazil, health for13 percent and educationfor just over five per cent. However, Brazil's socialsector investment, while comparing reasonably well with the OECD averageof 25 per centof GDP, is underminedby its highspending on pensions.Whereas pensions in OECD countriesaccount on averagefor 3 3 per centof spending,this figure rises to 44 per centin the case of Brazil, surpassedonly by Italy and Mexico.9 'Social assistance',targeted at the poorestgroups, remains a relatively smallproportion of the total social budget.However, it has expanded steadilyunder the administrationsof both FernandoHenrique Cardoso (1995-zoo2) and, especially,Lula (2003--). Under FernandoHenrique

8 Ibid.,p. 6. 9 Brazil,Orfamento Social do GovernoFederal 2001-2004, Minist&rioda Fazenda,Secretaria de PoliticaEcon6mica (Brasilia, 2005). SocialPolicies and Poverty Alleviation under Lula 693 Table i. Spendingonsocialprogrammes inBrail,.2001oo-2006 Totalsocial Socialassistance Bolsa budget* budget** Familia*** % direct % direct social socialassistance R$(billions) % GDP R$(billions) budget R$(billions) budget

2001 16o 13-3 8.5 5.3 1.5 18.o 2002 182 13.5 10.2 5.6 2.4 23.0 2003 213 13.7 12.9 6.o 3.4 26.0 2004 249 14.1 16.2 6.5 5.8 36.0 200f 17.0 6.5 38.0 2006 22.0 8.3 38.0 Sources:Orfamento Social do GovernoFederal 200oo-2004, Minist&rio da Fazenda, Secretariade Politica Econ6mica, Brasilia (200oo);'Queda da pobreza e maior no campo que nas metr6- poles', Folha de Sdo Paulo, I January(2006); 'QuitaFio da divida do FMI permitiraiinvestir a favordo povo, diz Lula', O Estado de SdoPaulo, 16 January(2006). * Directcosts of socialsecurity and pensions, health, education, social assistance, worker protection,housing and sanitationand agrarianorganization. ** Directcosts of assistanceto childrenand nutritionincluding Bolsa Familia, child labour, school forthe and etc. *** feeding,help elderly physicallydisabled, indigenous groups, BolsaEscola, Bolsa Alimentafdo, Cartdo Alimentafdo and Auxilio Gds.

Cardoso(FHC), therewas a significantincrease in socialassistance spending, reaching5.6 per centof totalsocial spendingby 2002 as protectionpro- grammes,such as BolsaEscola, expanded. By 2004,social assistancewas budgetedat R$16.2 billion,equivalent to 0.9 per centof GDP and 6.5 per cent of the totalsocial budget.In nominalterms, therefore, total social assistancespending leapt by 60 per cent underLula from2002-2004. However,as a proportionof the totalsocial budget the increase was less dramatic,although still significant, moving from 5.6 to 6.5 percent. Increasedspending for Bolsa Familia specifically has been moremarked. As Table I shows, spending on the four major safetynet programmes, comprisingwhat has becomeknown as BolsaFamilia, grew from 23 per centof thesocial assistance budget (R$2.4 billion)in 2oo02under FHC to 36 per cent(R$5.8 billion) under Lula in 2004. This rose to R$6.5billion in 2005 and willreach a predictedR$8.3 billionin 2006, or 38 per centof overallsocial assistance.Figure i underlinesthe heavieremphasis on safetynet spendingunder Fome Zero and BolsaFamilia within the social budgetunder Lula after2003. Duringthe Lula administration,spending on BolsaFamilia has risenfrom I.i to 2.5 per centof totalgovernment expenditure,increasing from 0.2 to 0. 5 percent of GDP.10However, to keep

10 K. Lindert,'Bolsa FamiliaProgram - Scaling-UpCash Transfersfor the Poor', MJDR PrinciplesinAction: Sourcebook onEmerging Good Practice, World Bank (Washington DC, 200zoo). 694 AnthonjHall 40 ___-ac------a 35 35 U?----

30

25

20

15

10

5-

0 2001 2002 2003 2004 2005 2006 -+- Social assistance as % of total social budget --- Safetynets (Fome Zero-BolsaFamilia) as % of social assistancebudget

Fig. i. Trendsinfunding for social assistance and safety nets, BraZil, 2001-200oo6. Sources: Orpamento Social do GovernoFederal z001-2004, Ministirio da Fazenda,Secretaria dePolitica Econdmica, Brasilia(2ooy); 'Queda dapobreza 6maior no campo que nas metropoles', Folha de SdoPaulo, IJanua~y (2oo0);'Quitapio da dividado FMIpermitird investir a favor do povo, diz Lula', 0 Estado de Sdo Paulo, ;6Januay(20oo). mattersin perspective, itshould be bornein mind that Bolsa Familia accounts forjust 2.3 percent of directmonetary transfers in Brazil,far outweighed bypensions at 82 percent, which are far more regressive.11

FomeZero

'Zero Hunger'marked out theWorkers' Party (PT) governmentas socially progressiveand seemingly more serious about dealing with poverty than any otherprevious regime. The programmeitself was inspired by Jose Graziano da Silva,former professor of agrarianstudies at theUniversity of Campinas in SIo Paulo,appointed by Lula to headthe newly created Ministry of Food Securityand FightAgainst Hunger. Fome Zero was in practicean umbrella programmefor initiatives already developed under the FHC administration. Thesefederal initiatives had in turn developed from localised projects started duringthe I990s, replacingan earlierprogramme of distributing food parcels (cestasbaisicas-PRODEA) which operatedfrom 1993 to 2000 and was designedto providefor the needs of a familyfor one month.12However, n Brazil,Orcamento Social. 12 SmallerCCT projectshad already been implemented atmunicipal level from 1995 onwards in Campinas,Brasilia, Blumenau, Belo Horizonte,Victoria, Salvador and RibeirdoPreto. See J. Grazianoda Silva,W. Belikand M. Takagi,'The Challengesof a Policyof Food SocialPolicies and Poverty Alleviation under Lula 695 foodparcels were heavily criticised for being unashamedly used to capture votes,with 30 millionbeing distributed in thegeneral election year of 1998, twicethe normal number.13 Fome Zero thus brought together under one label severalexisting initiatives; namely, the conditional cash transfer (CCT) pro- grammesof Bolsa Escola for boosting school attendance, Bolsa Alimentafdo for maternalnutrition and the PETI programmeagainst child labour, along with theAuxilio Gds cooking gas subsidy.However, it also added a new food entitlementscheme, the CartaoAlimentafdo, based on the use of a special creditcard for the purchase of selectedfood items. The flagshipwithin this 'flagship' programme, however, was alwaysBolsa Escola (School Grant) that providesmothers with a monthlystipend (currentlythe equivalent of US$7 per monthper child)in returnfor their childrenattending school at least 85 per centof the time.Started in the FederalDistrict of Brasilia in 1995,by then governor Cristovam Buarque, as wellas in Campinasand Sdo Paulo, it was adoptednationally in 2001.By the endof 2003, thescheme had been implemented in almostall of the country's 5,561municipalities and had distributedalmost US$5 oo millionin grantsto overfive million families with 8.6 millionchildren. According to thelatest availablefigures, Bolsa Escola accounts for around half the total spending on thefour major safety net programmes under Bolsa Familia.14 In consolidatingexisting safety net programmes,however, the Lula administrationdid introducesome major changes.Firstly, in line with PresidentLula's electoralpromises, Fome Zero was underpinnedby a clear and vociferouspolitical commitment to benefitthe very poorest sectors of Braziliansociety. Secondly, while implementation of the programmewas decentralisedto municipallevel, Graziano's ministry attempted, in some measure,to bypasslocal political interests to createan alternativedistribution network;for example, by substitutingfood stampswith credit card style food cardswhich were in theoryat leastless vulnerableto incorrectuse. However,as notedbelow, decentralisation of programmeimplementation has broughtits own politicalproblems. Third, appeals were made to major companiessuch as Fordand Unilever,as wellas to supermarketchains, for contributionstowards Fome Zero that would signify a new sense of corporate socialresponsibility and alliancebetween public and privatesectors. Over 00oocompanies are now involved in thescheme. Internationaldonors such as the IMF, theWorld Bank and the Inter- AmericanDevelopment Bank moved quickly to endorsethe Lula govern- ment'santi-poverty stance. After all, it testifiedto thenew administration's

Securityin Brazil',in A. Cimadamore,H. Dean andJ. Siqueira (eds), The Poverty ofthe State: therole in the reconsidering of thestate struggleagainst global poverty (Buenos Aires,200oo), 13 14 pp. 157-78. Ibid. Brazil,Orfamento Social. 696 AnthonyHall prudentsocial policies, belying its radicalleft-wing image. It also fittedin verynicely with the increasingly popular view held by international financial institutions,led bythe World Bank, of socialpolicy as theconstruction of safetynets.15 During the 1980s and 90s, theintroduction of socialfunds to deal withthe adverseimpacts of structuraladjustment had becomewell established.16Subsequently, the collapse of theSoviet Union and compre- hensivesocial security schemes in theformer communist countries, together withthe East Asia crisis, led to theintroduction oflarge-scale, targeted social protectionmeasures with World Bank assistance.As notedabove, these emergencymeasures were now couchedin the languageof 'social risk management',intended to go beyondthe provision of short-termrelief to helpstrengthen people's livelihoods and their ability to copewith shocks and risksin thelong-term. At the same time,the Bank set up a specialSocial ProtectionUnit within its HumanDevelopment network to overseethis policy.In 2005,zoosocial protection accounted for some 12 per centof total Banklending, with US$2.5 billioninvested in 45 projects.17 Yet despitethe initial enthusiasm which accompanied the inauguration of FomeZero in early2003, seriousproblems soon became apparent. Many were due to thefact that each programmeoperated independently of the other withno overallcoordination. Each had separateadministrative structures, beneficiaryselection processes and bankingcontracts for payments. The Lula governmentinherited a partiallyunified database, or CadastroUnico, fromthe previous administration but this covered only 70 percent of poor families.This made effective implementation more difficult, leading to high implementationcosts and targeting errors.18 Problems included the adoption of politicalcriteria for selectingbeneficiary families, an over-centralised managementsystem, exclusion of someeligible poor families, duplication of benefitsand a lackof updated information. Even the then Minister Graziano publiclyadmitted to a targetingerror under the programme of up to 30 per cent,which led to longdelays in implementation.19

BolsaFamilia Bymid-zoo3, just a fewmonths into the Lula administration,such problems werealready creating disillusionment with Fome Zero, which was perceivedas increasinglyineffective in fulfilling its mission to fighthunger and absolute

15 Hall andMidgley, Social Poliy for Development, Sage (London,2004). K. Subbarao, yNetPrograms and Reduction:Lessons CrossCount~y 16 Safet Poverty from Experience, WorldBank (Washington DC, 1997). 17 WorldBank, Putting Social Development to Work for the Poor: An OED Reviewof World Bank Activities(Washington DC, 200oo5). 18 IPEA, PoliticasSociais: Acompanhamento eAnilise, no. 7, August(2003). 19 na area fica deSdo 'UnificapAo social para2004', Folha Paulo,Io June(2003oo). SocialPolicies and Poverty Alleviation under Lula 697 poverty.The newlylabelled Bolsa Familia (Family Grant) scheme was thus announcedin October2003 to integratethe actions and budgets of thefour main CCT programmes;namely Bolsa Escola,Bolsa Alimentafdo,Cartdo Alimentafdoand Auxilio Gds. Bolsa Familia is nowthe largest conditional cash transferprogramme in theworld. The distinctiveFome Zero label was retained as a widerumbrella term to encapsulatethe overall safety net programme in Brazilinvolving a diverse range of over30 socialinterventions in the areas of foodsecurity, farming and otherforms of support,including Bolsa Tamilia, althoughthe label is regardedas cosmetic.20 The unificationof safetynet programmesunder Bolsa Familia involved significantchanges. Under the old set-up,the components of Fome Zero were administeredseparately: Bolsa Escola (Ministryof Education),Bolsa Alimentafdo(Ministry of Health),Cartdo Alimentafdo (Fome Zero) and Auxilo Gas(Ministry of Mines and Energy). Although providing benefits to roughly thesame target population, they were difficult to administer,each scheme havingits own bureaucraticstructure, data collection,fiduciary responsi- bilitiesand reporting systems. Not onlywas thisfragmentation costly, but it sacrificedpotential benefits in termsof synergiesand complementaritiesat thefamily level in schooling,health and nutrition.By contrast,integration has helped reduce administrativecosts, improve targetingefficiency, standardiseprocedures and resultsindicators and coordinatefederal with statelevel safety net programmes. Conditionalities for education, health and nutritionwere merged to producegreater synergy while unit transfer benefits werealso increased. There is a newemphasis on targetingthe household unit ratherthan the individual. However,many aspects of BolsaFamilia implementation remain decen- tralisedto municipallevel. These includedata collection,registration of potentialbeneficiaries under a singleregister (Cadastro Unico) and monitoring of adherenceto conditionalities.Under Brazil's decentralisation law, each municipalityis required to setup a socialcouncil (Conselho de Controle Social) forthis purpose, whose members are chosen by the mayor from public and civilsociety sectors. In January2004, President Lula announcedthe integration of theMinistry of Food Securityand Fightagainst Hunger with the Ministryof Social Welfareto formthe new Ministry of SocialDevelopment and Fightagainst Hunger(MDS). Grazianowas replacedby PatrusAnanias as Ministerof Social Development.The Inter-ministerialManagement Committee and ExecutiveSecretariat for Bolsa Familia,originally linked directly to the President'sOffice, were transferred to thenew MDS in orderto facilitate betterintegration of social programmes.

20 http://www.fomezero.gov.br/ 698 AnthonyHall The schemeis targetedat twogroups, the 'verypoor' withhousehold incomesof up to R$5o (US$23 permonth), and the'poor' withhousehold incomesof R$51-IOO (US$23.50-46)per month.The upperlimit was increasedin Aprilzoo6 to R$2zo (US$55) in orderto facilitateattainment of theoverall target numbers. Very poor familiesare awardeda fixedmonthly stipendof R$5o regardlessof familysize, while families earning between R$51-120 permonth receive no basicpayment. Both groups are eligible for variable of child of school to payments R$I1 (US$7) per age (6-15) a maximumof three,or up to sixyears of age underBolsa Alimentafdo. The maximumbenefit per household was setat R$95 (US$43).These cash grants are conditionalupon proof of regularpresence at school, children's vaccination,attendance at healthclinics and, where relevant, participation in nutritionand vocationaltraining courses. The gas subsidyprovides R$I5 everytwo monthsfor the same social groupsto purchasecooking gas cylinders.The CartaoAlimentafdo, or food card system, was initiated in March 2003 forthe poor, semi-arid Northeast as partof Fome Zero. It givesfamilies to halfthe minimum a of earningup wage monthlycash supplement R$50 forfood purchases. In June2004, the World Bank approveda US$57z millionsector-wide loan (SWAP) to supportBolsa Familia. This will providefunds for cash transfers(96 percent) as wellas fortechnical assistance to developa unified database,improve targeting mechanisms, develop a systemfor monitoring and evaluation,strengthen the institutions involved and the programme's implementationcapacity within the MDS andto developa disseminationand communicationsstrategy.21 In December of the same year,the Inter- AmericanDevelopment Bank (1DB) approveda loan of US$I billionfor BolsaFamilia, with the promise of up to US$z billion,depending on satis- factoryprogress.22 These twomultilateral commitments, totalling US$2.57 billion,are the equivalentof one-quarterof the estimatedtotal funding forBolsa Familia (R$24 billion)during the Lula governmentand provide an unequivocalexternal endorsement of the safetynet/CCT approach to povertyreduction in Brazil.The Bankhas also been verygenerous in its publicpraise for Bolsa Familia, whose instrumentsare seen as havingan applicabilitybeyond Brazil itselfto othercountries implementing CCT schemes.23

21 World Bank, ProjectAppraisal Document on a ProposedAdaptable Program Loan to theFederative Republicof Brazil in Supportof the First Phase of the Bolsa FamiliaProgram (Washington DC, May 25, 2004). 22 cIDB approvesloan of US$i billionfor expansion and consolidationof social protectionin Brazil based on the Bolsa Familia program',Press Release, December (2004). 23 15 K. Lindert,'Bolsa Familia Program'. 'Banco Mundial faz elogio ao Bolsa-Familia',Folha de SdoPaulo, 15 February(2006). SocialPolicies and Poverty Alleviation under Lula 699 Assessmentofprogress The firstgeneration of CCT programmesin Latin America(Mexico, Colombia,Honduras and Nicaragua)incorporated evaluation strategies usinga varietyof techniques,allowing preliminary conclusions to be drawn regardingtheir effectiveness.24 Education programmes in Mexico,Colombia and Nicaraguahave boosted primary school enrolment rates for both boys and girls.However, results on regularattendance are mixedwhile the issue of qualityof schoolingprovided is notusually addressed. Child health and nutritionhas also improvedin thesecountries as a resultof CCTs, with greaterparticipation in growthmonitoring and immunisation campaigns. At the sametime, food securityhas improved,with cash transfersleading to higherlevels of consumptionand nutrition.Evaluation results show that over80 per centof CCT programmebenefits in thesecountries reach the poorest40 percent of families.25 Whilethis first round of evaluationresults is, in generalterms, encourag- ing,evidence on theperformance ofBolsa Familia has yet to be systematically gathered.In termsof numerical achievements, the record of Bolsa Familia has been impressivein manyrespects. First, the numberof beneficiarieshas morethan doubled in threeyears to over30 million.This is theequivalent of roughlythree-quarters of those living below the povertyline, who are concentratedmainly in thepoorest North and Northeastregions. Second, accordingto officialestimates, the average level of benefit paid per family has almosttripled from R$28 (US$I 3) to R$75 (US$34) per month.26By the end of 2005,some 8.7 millionfamilies had been includedand thisfigure was expectedto riseby late zoo6 to 1.2 millionfamilies or 44 millionpeople, its eventualtarget population.27 At themoment, based on evidencefrom other Latin American countries, sweepingassumptions are beingmade that safety net programmes such as thesecan reduce inequalities, strengthen human capital and improve people's well-being.Brazil's Ministryof Finance,for example,has confidently declaredthat, 'programmes such as BolsaFamilia are highlyefficient instru- mentsin thefight against poverty.'28 Official government statements to the pressregularly make similar claims. The reportedreduction in ruralpoverty levelsin Brazilfrom 39.5 per cent in 2003to 35.4 per cent in 2004has been attributedin largemeasure to cash transferssuch as pensionsand Bolsa

24 25 Rawlings,A NewApproach toSocialAssistance. Ibid. 26 Brazil,Orfamento Social, p. 24. 27 Ibid. World Bank, Brazil: Equitable,Competitive, Sustainable: Contributions for Debate (WashingtonDC, 2004). 28 Brazil, OrfamentoSocial, p. 16. 'Bolsa-Familiaalivia pobreza, diz governo', Folhade Sdo Paulo, 25 December(2005). 700 AnthonyHall Familia.29However, critics express doubts about the ability of Bolsa Familia to helpgenerate income or employment.Questions have also beenraised over a numberof keyoperational dimensions concerning, for example, targeting and selectionprocedures, the effectivenessof conditionalities,the role of politicsand clientelism, weak local institutional capacity, the participation of civilsociety and thedegree of transparencyand accountability. Althoughthere has notyet been any comprehensive impact evaluation of BolsaFamilia, some investigationshave been carriedout. A longitudinal studyby the Ministry of Healthand theInternational Food PolicyResearch Institute(IFPRI) in fournortheastern municipalities, over an 18 month periodfrom 2002-2003, concludedthat the Food Grant(Bolsa Alimentadio) whichmakes cash to womenof component, monthly grants R$25 (US$I i) conditionalupon regular health clinic attendance, had succeededin boosting thefood security of mothers and children. The quantityand diversity of food consumedimproved, while the growth and nutritional status of childrenhad progressedcompared with non-participants in the programme."3A com- prehensiveimpact evaluation of BolsaFamilia is underwayduring 200 5-06, fundedby the World Bank loan mentionedabove, that will no doubtshed lighton theconsequences for school attendance, nutrition status and other socialindicators.31 An ex-anteevaluation of BolsaEscola using simulation techniquesconcluded that it would increaseschool enrolmentsamongst poorten- to I 5-yearolds by 60o per cent but would reduce poverty by just one percent.32 An earlyevaluation of thePETI childlabour scheme in selected Northeasternstates showed that it reducedthe probabilitythat children wouldwork by up to 26 percent.33 Althoughcomprehensive impact evaluations of BolsaFamilia as a whole arepresently lacking, some studies have been carriedout of BolsaEscola, its largestcomponent, which was introducedin variousmunicipalities during the1990s beforebeing extended as a nationalprogramme in 2001. One such investigationinto early municipal school stipend programmes, based on 2000 Censusdata, found that they had a positiveimpact on schoolattendance but

29 'Queda da pobrezae maiorno campoque nas metr6poles',Folha de Sdo Paulo, I January (zoo6). 30 Ministryof Health,Avaliafdo Final de Impacto do Programa Bolsa-Alimentafdo (Brasilia, 2004). A similarstudy is beingcarried out by theFederal University of Bahia (UFBA) in three northeasternmunicipalities of BolsaFamilia impacts upon thelocal economy,health and nutritionstatus. 31 This is being undertakenby Centre for Development and Regional Planning (CEDEPLAR) at theFederal University of MinasGerais. 32 F. Bourgignon,F. Ferreiraand P. Leite,'Conditional Cash Transfers, Schooling and Child Labor:Micro-Simulating Brazil's Bolsa Escola Program',The World Bank Economic Review, vol. 17,no. z, pp. 229-54. 33 Citedby Rawlings, 'A NewApproach toSocialAssistance', pp. 9-10. SocialPolicies and Poverty Alleviation under Lula 701 hadno significanteffect on childlabour since children were induced to go to schooland workconcomitantly. That is, thecash transfersprovided were too smallan incentiveto persuadefamilies to forgochild labour.34 More worryingly,research based on Brazil'snational household survey (PNAD) for2003 concludedthat Bolsa Escola, which is aimedat childrenaged from six to 15 yearsof age,is poorlytargeted and does not benefitthose families whosechildren are morelikely to be persuadedto attendschool by a cash stipend.35The studyfound that, for children aged between seven and 13,the cash incentiveeffect is negligiblesince they are likelyto attendanyway as provisionis universal.Yet it is in thisage groupwhere most Bolsa Escola benefitsare concentrated. However, children start dropping out of schoolat age 14 and it is in the 14 to 17 age rangewhere a cashincentive would be mosteffective in retainingchildren, the study concludes. The sameresearch intoBolsa Escola found that the programmeis generallywell focusedon poorersocial groups, with 50 per centof benefitsreaching the two lowest incomedeciles. Yet 18 percent of cashtransfers accrue to 1.5 millionchil- drenin thefifth income decile and higher,suggesting the presence of some distributivedistortions.36 Furtherevidence on the effectivenessof BolsaEscola is providedby a processevaluation of its implementation, based on fieldsurveys carried out in 2004 in 261 randomlyselected municipalities in four states of Northeast Brazil.37Although programme impacts on targeting,educational partici- pationand poverty alleviation will be assessedonly at a laterdate, the study neverthelessprovides important initial insights into the highly variable mu- nicipalrecord of implementationand some of theproblems encountered. The evaluationfound, for example, that beneficiary screening and selection forBolsa Escola is basedon a rangeof objective and diverse criteria set at local and federallevels (such as per capitafamily income, schooling, family size, ages of children,health records, etc.) and thatthe process is on thewhole quitetransparent, with widespread discussion and disseminationof infor- mation.Initial confusion over the respective roles of municipaland central governmentin beneficiaryselection was clarifiedunder Bolsa Familia by a rulingwhich transferred responsibility for this task over to the federal authorities.

34 E. Cardosoand A. PortleaSouza, 'The Impactof Cash Transferson ChildLabor and SchoolAttendance in Brazil',mimeo, 2003. 35 S. Schwartzman,'Education-oriented social programsin Brazil: the impactof Bolsa Escola', mimeo,Instituto de Estudosdo Trabalhoe Sociedade,Sdo Paulo,2005. 36 Ibid.,p. I6. 37 A. de Janvry,et al., 'Brazil'sBolsa Escola Program:The Role of Local Governancein DecentralizedImplementation', SP DiscussionPaper No. 0542, Social ProtectionUnit, WorldBank (Washington DC, 2005). 702 AnthonyHall Anecdotal evidence suggests,however, that targetingerrors occur regu- larlyand that thereis, for example,much benefitduplication due to poor coordinationamongst programmesand an incomplete central database. Researchersmaintain that it is possible to improve targetingthrough the developmentof betterincome-based and proxy means testingdevices.38 Other specialistsare more scepticalof thisapproach as expensive,unwieldy and ultimatelyineffective.39 A current concern is how to improvethe central database (CadastroUnico), the main beneficiaryregistration and selection mechanismfor Bolsa Familia, so thatit may be used to channel,more effec- tively,resources to the poor, in view of persistentproblems such as the underreportingof incomes and politicalbias in clientselection.40 A major conclusionof the same studyof BolsaEscola is thattransparency in beneficiaryselection does not necessarilylead to overall accountability. Unsurprisinglyin the Braziliancontext perhaps, strong evidence was found of politicalmanipulation in programmeimplementation in at least ten per cent of municipalitiessampled. This observationis consistentwith alleged cases of resource misuse and corruptionunder Bolsa Familia documented regularlyin the Brazilianpress. Involvementof the mayor'soffice in ben- eficiaryregistration and selectionwas associated with the highestlevels of clientelismand patronage.The absence of 'social councils' comprisinglocal citizensand authorityrepresentatives, whose functionis to monitorpro- grammeexecution and induce a degree of transparencyand accountability, was found to increasethe likelihoodof such favouritism.Where theyhad been set up, and where theywere actuallyworking (in two-thirdsof cases), councilswere foundto be quite effectiveon the whole and theirimpact on programmeimplementation was generallypositive. Yet one-fifthof munici- palities sampled had no such council, despite the fact that it is a federal requirement.It was also foundthat they often performed poorly, their ob- jectivityin monitoringbeneficiary selection and adherenceto conditionalities (such as school attendance)being compromisedby local connectionsand affiliationto the mayor,to whom council memberswere not infrequently

38 R. Paes de Barrosand M. Carvalho,'Desafios para a PoliticaSocial Brasileira', Textos Para Discussio,No. 985, IPEA (Rio de Janeiro,2003). M. Neri,'Designing a Systemof Social Targetsand SocialCredit', and M. Medeiros,'Conditional Flow Transfers in Brazil:Some Problemsfor Targeting the Poor'. Paperspresented at theconference on OvercomingSocial Exclusion: in Centrefor BrazilianStudies, of Brazil ComparativePerspective, University Oxford,22 June(2004). 39 E. Suplicy,'The Approvalof theCitizen's Basic IncomeBill in Brazil',and G. Standing, 'PromotingIncome Security as a Right'.Papers presented at theconference on Overcoming SocialExclusion: Brazil in ComparativePerspective, Centre for Brazilian Studies, University of Oxford,2z June(2004). 40 B. de la Briereand K. Lindert,Reforming Brazil's Cadastro Unico to Improve Targeting ofthe Bolsa FamiliaProgram (Washington DC, zoo5). SocialPolicies and Poverty Alleviation under Lula 703 related.One officialreport by the federalauditors (Tribunal de Contas da Unido)strongly criticised Bolsa Tamilia for its lackof programmemoni- toring.41 A majorbone of contention in discussionsover the effectiveness of CCTs liesin thesupposedly 'conditional' nature of cash transfers. As notedearlier, thetheory is thatbenefits should be contingentupon mothers and children meetingsocial obligationssuch as regularschool attendance,undergoing healthscreening and receivingvaccinations. Not onlyanecdotal evidence, but also systematicresearch, demonstrates that in practicethe monitoring andenforcement of suchconditionalities is highly problematic. For example, school teachersare oftenreluctant to reportabsent pupils, while social councilsseem unable or unwilling to performa policing role. In thewords of one study: The federalgovernment is unable to supervisethe behaviour of poor families throughoutthe country; local governments and municipalities areeither inefficient or tiedup withlocal elites, or both;and community grassroots organisations are easilycaptured by political parties and movements.42

Imposingconditionalities is often seen not onlyas difficultin practicebut also as paternalisticinprinciple, leading many to recommendits replacement byunconditional support.43 Only time will tell whether such obligations are justifiedin termsof inducingimprovements in economicand social in- dicatorsand in buildinghuman capital. Beyondquestions of operationaleffectiveness, however, lie morefunda- mentalconcerns. Serious doubts have been raised,for example, about the abilityof such safetynets, even when properlyimplemented, to mount a seriouschallenge to povertyin the contextof highlyunequal societies such as Brazil. Some criticsallege that Bolsa Familiais an essentially politically-drivenstrategy for holding down increases in thelegal minimum wage which,they maintain, would actuallybenefit more families working in bothformal and informal sectors, through multiplier effects. In contrast, based on IPEA researchfindings, the governmentinsists that selective incometransfers have had a moresignificant anti-poverty impact than wage policy.44Echoing the more universalisticwage policyapproach, Senator Eduardo Suplicy(Sio Paulo, PT) has for over a decade been fighting

41 'Para o TCU, faltade fiscalizavdoprejudica Bolsa-Familia', 0 Estadode Sao Paulo,30 September(2004). 42 S. Schwartzman,'Education-oriented social programs in Brazil',p. 25. 43 Ibid.and Suplicy,'The Approval...' 44 'Bolsa-Familiasupera o minimo,diz estudo',Folha de Sdo Paulo, 22 June(2004). 'Minimo deve seravaliado junto com politicassociais, diz Palocci', O Estadode Sdo Paulo, 8 June (2004). 704 AnthonyHall forthe introduction of an unconditionalCitizen's Basic Incomein Brazil.45 This was finallyapproved in January 2004 by theBrazilian Congress (Law No. Io,835), committingthe government to introducinga standard basic incomein zoo200for all Brazilians,commencing with the poorest sectors but eventuallyextending it to all citizens.Although Suplicy regards Bolsa Familia as thefirst step towards such a Citizen'sBasic Income, there are clearly major practicaland ideological differences between the safety net approach and the idea of a Citizen'sBasic Income as a right.46 Yet othermore far-reaching criticisms of theBolsa Familia approach to socialpolicy concern Brazil's persistent inequality and relativedeprivation, despiteimprovements in levels of absolutepoverty. With the seventh mostunequal income distribution in the world, the top 20 percent in Brazil earnalmost 64 per centof personalincome and the bottomquintile just 2.3 percent.47 Although the challenges of dealingwith absolute poverty and inequalityare distinctand implyseparate kinds of economicand social policies,frustration is sometimes expressed over the seeminginability of BolsaFamilia to bringabout any redistribution of wealth.However, even strongadvocates of targetedincome transfers, such as the WorldBank, admitthat they cannot on theirown overcomethe poverty generated and sustainedby entrenchedinequities. Ironically, Lula himselfacknowledged thispublicly, declaring that, 'Bolsa Familia is not our salvation,merely an emergencymeasure ... andthe ideal is thatin a fewyears time Bolsa Familia willno longerbe necessary.'48 Yet theblatant politicisation of Bolsa aramiliasuggests that it may retain a long-termrole in Brazil'ssocial policy agenda. It is evidentthat Lula and his governmentcame to dependheavily on theprogramme to strengthenpol- iticalsupport and generate votes in theOctober zoo6 presidentialelections.

45 Suplicy,'The Approval...'. IPEA, PoliticasSociais:Acompanhamento eAnalise, no. 8,February (2004). 46 In a similarvein, human rights activists have for some timeadvocated the idea of a universalchild benefit in cash or kind,designed to guaranteelonger-term minimum livingstandards and havea moresustainable anti-poverty impact. This would be funded throughinternational taxation such as a 'Tobin-style'currency transaction tax to build up an internationalinvestment fund for children. The ILO, forexample, has a plan for the eliminationof childlabour over 20 yearsat a cost of US$95 million,equivalent to less thanten per centof developingcountry debt service payments or one-fifthof their defencespending. See, P. Townsend,'Making Human Rights Stick: Action to Eradicate Seminarson Human London School of Economics, Poverty',LSE R'ghts, 29 June(2004) and 'The Need forDirect Policies to FightChild Poverty', In Focus,UNDP, February (2004). 47 R. Paes de Barrosand M. Carvalho,'Desafios para a PoliticaSocial Brasileira', Textos Para Discussdo,No. 985, IPEA (Rio de Janeiro,2003). do discursode Lula 48 'Integra em cerim6niado Bolsa Famila',0 Estadode Sdo Paulo, 23 December(2005). SocialPolicies and Poverty Alleviation under Lula 705 Relianceon suchelectoral populism and the strengthening ofthe President's officialimage as the'Father of thePoor' (Pai dosPobres) could indeed help compensateat the ballot box fora fundamentalpolitical weakness in the Lula administrationattributable to whatsome criticshave identified as 'the ab- senceof a politicalproject structured within the social movements and mass organisationswhich had sustainedthe PT sinceits foundations'.49Opinion polls have demonstratedquite clearlythat during zoo6, in the wake of corruptionscandals, political support for Lula strengthenedsignificantly in the Northeast,Brazil's poorest region, in whichthree-quarters of Bolsa Familiabeneficiaries are concentrated.Support for Lula has been substan- tiallyhigher amongst beneficiaries than amongst those not taking part in the programme.50 The temptationsare obvious. With average total monthly transfers under BolsaFamilia of almostR$5oo00 (US$230) millionper month, official figures showthat some municipalitiesrely on theprogramme for 40 per centor more of theiroverall budgets. This dependenceis especiallymarked in poorerNortheastern municipalities.51 It was also claimedin zoo200that Bolsa Familiacould be worthup to 22 millionvotes for Lula in the2006 elections.52 An estimatedthree-quarters of Brazil's absolute poor (eightmillion house- holds,or 16 per centof thetotal) now benefitin somemeasure from cash transfersof one kindor another.In theNortheast, one-third of households receivesuch transfers, rising to 44 percent in ruralareas.53 It is thereforenot difficultto see how this could translateinto politicalsupport for the President.Given his humbleorigins and unrelentingpublic commitment to povertyalleviation since his inauguration,he is verymuch personally identifiedwith Bolsa Familia. In pre-electionspeeches, as wellas in general governmentpublicity campaigns, a concertedgovernment effort was mounted,perhaps understandably, to publicise the achievementsof Bolsa Familiaand gainmaximum political advantage.54

49 P. Flynn,'Brazil and Lula, zoo200:crisis, corruption and changein politicalperspective', ThirdWorld Quarterly, vol. 26, no. 8 (200zoo),p. 1245. so 'Bolsa Familiae principalrazio parasubida de Lula',Folba de Sdo Paulo, 26 February(2006). Datafolharesearch showed that while Lula enjoyedan eightpoint lead overhis political rivalsgenerally in theNortheast, this figure rose to 21 pointsin thecase of BolsaFamilia beneficiaries.See also D. Fleischer,Bra#il Focus, special report, z22 February (200oo6). 51 R. M. Marques,'A importnnciado Bolsa Familianos municipiosbrasileiros', Cadernos de EstudosDesenvolvimento Social em Debate, No. i, Ministryof Social Development(Brasilia, 2005). 52 'Bolsa Familiadeve ser "arma" de Lulaem zoo6', Terra,9 October(2005). 53 Nationalhousehold survey (PNAD) datafor z004. 54 In a speechLula emphasizedthe fact that, '77 percent of thoseliving below the poverty lineare receiving benefits from Bolsa Familia'.See 'Lula fazbalanpo de programassociais em Porto Alegre',Folha de Sao Paulo,17 February(2006). SECOM, the department responsiblefor campaign publicity, made no secretof thefact that Bolsa Familia would be 706 AnthonyHall Furthermore,local politicians, especially in theNortheast, have developed theirown vote-catchingstrategies in supportof thePresident, capitalising on federalschemes such as BolsaFamilia to strengthenthe image of President Lula and theWorkers Party.55 Indeed, even the political opposition openly consideredhow, in theevent of winningpower in Octoberzoo6, it would extendand modifysuch CCT programmesto build politicalsupport amongstthe poor.56Partido da Social DemocraciaBrasileira presidential candidate,Geraldo Alckmin, publicly declared while campaigning in the Northeastthat, if elected,he would not onlymaintain Bolsa Familia but expandthe programme.57

Conclusion

The upwardtrend in budgetsupport for Bolsa Eamilia since 2003 continued in 2006, thanksin partto generousassistance from the World Bank and the Inter-AmericanDevelopment Bank. This testifiesto itscontinuing political importancedomestically and ideologicallyas a markof the re-castingof socialpolicy by multilateral organisations. As thePresident himself noted in a televisedspeech to thenation, Brazil's repayment in December2005 of its US$ 5.5billion IMF debt,saving at leastUS$9oo million in interest,would also releasefunds for social programmes.58 Despite such high-profile com- mitments,however, there is stilla longway to go in deliveringcomprehen- sivesocial protection to thecountry's poor. Brazil'sfirst national study of foodsecurity revealed recently that of the 18 million'food insecure'popu- lationonly 5.3 million (29 percent) was benefitedby income transfer pro- grammes.59Yet no matterhow effectiveBolsa Familia might or mightnot eventuallyprove to be in termsof alleviatingabsolute poverty, the govern- ment'sdependence on suchsafety nets as a coreelement of itssocial policy begsmuch larger questions.

a highprofile area in the elections.See 'Governo aceleragastos em publicidadein- stitucional,',Folha de Sdo Paulo, ss 19February (2006). For example,the major of Te6filoOttoni declared that she had sentletters to all ben- eficiariesof BolsaFamilia and the'Electricity for All' (Lu.zparaTodos) programmes in her municipalityreminding people that these were the personal initiatives of PresidentLula. Thishad resulted, she declared to thepress, in increasedpolitical support for the President amongstthe poor. See 'Prefeitarelata pratica de angariarvotos para Lula com programa federal',Folha de Sao Paulo, 13 February(2006). 56 'Tucanospreparam plano para seduzirmais pobres', 0 Estadode Sdo Paulo, 19 February (2006); 'PT e PSDB convergemno diagn6sticosobre o futurodo Bolsa Familia',Valor Econdmico,31 January (2006). 57 'No NE, Alckmindiz que Bolsa Familiando e criap-ode Lula', 0 Estadode Sco Paulo, 22 April,(2oo6). 58 'Quitaraoda dividado FMI permitiraiinvestir a favordo povo,diz Lula', O Estadode Sdo Paulo,i6 January(2006). 59 IBGE, SeguranfaAlimentar 2004 (Brasilia, 2oo6). SocialPolicies and Poverty Alleviation under Lula 707 A major perversetrend is thatthe poor in Brazil must now relyincreas- ingly on governmenthandouts to support theirlivelihoods. Analysis of census data shows thatwhereas in 1995 earningsfrom employment con- tributed89 per centof poor householdincomes, by 2004 thisproportion had dropped to 48 per cent.60In otherwords, over half of the income in very poor householdsin Brazilis now derivedfrom federal cash transfersthrough largelynon-contributory programmes including Bolsa Familia and ruralpen- sions. Some observershave identifieda growingculture of dependence perpetuatedthrough such safetynet mechanisms.Less kind criticshave re- namedBolsa Familia/Bolsa Escola as 'Bolsa Esmola' ('charitygrant' or 'beggar's grant'). There is a strongrisk that the perpetuationof such a hand-out culturethrough safety net policies could lead to increased clientelismas politiciansconsciously use and manipulatethese programmesas part of widerpolitical and vote-capturingstrategies. This would belie the 'social risk management'component of currentsafety net strategiesdesigned to address longer-termdevelopment challenges. Of course,such dependenceof poorer classes on federaltransfers has to be seen togetherwith the massivereliance of the middle classes and formerpublic sector employees upon heavily subsidisedstate pensions, which consumes the lion's shareof Brazil's social budget. A consequence of these politicaldistortions in public expenditurecould be, however,that key areas of social infrastructuresuch as schools and hospitals are starvedof resources.Arguably, it is in these social welfare sectors where more long-terminvestment is necessaryto build up the human capitalnecessary as the basis for economic growth.Perhaps it is no coincidence that,while spendingper annum on education and remainedconstant as a proportionof GDP (0.7 and 0.2 per cent respectively)between 2001 and 2004, the social assistance budget grew from 0.7 to 0.9 per cent of GDP over the same period.61Furthermore, BolsaFamilia's shareof thisgrowing social assistancebudget itself increased significantlyfrom 26 per cent in 200zooto 38 per cent in 2oo00(Table I and Figure i). Indeed, it is becoming larger than all other federal govern- ment education programmes,excluding higher education.62 In a growing number of Latin American countries,CCTs are becoming a cornerstone of social policyand thereis a dangerthat although they do certainlyrepresent a creative approach to providing social assistance, 'they constitutean 'end-run' around the more difficulttask of reforminginefficient public

60 'Pobres se distanciamde ricose dependemmdos do governo',Folha de SdoPaulo, 25 December(2005). 61 Brazil,Orfamento social. 62 S. Schwartzman,'Education-oriented social programs in Brazil',p. 3. In 2004, non-higher federallyfunded education accounted for R$4 billion,compared with a BolsaFamilia budget of R$5.8 billion,around half of whichis allocatedto BolsaEscola. 708 AnthonyHall services.'63There is a riskthat the prioritisationof such schemes may ser- iouslycompromise longer-term investments into basic social infrastructure. Finally,it is frequentlyargued that, in orderto mounta seriouschallenge to mass povertyin Brazil,there will have to be modificationsin asset own- ership,labour marketsand economic policy.Indeed, the World Bank itself now arguesvehemently for stronginterventionist policies to address mass povertyand inequalitywhile promotinggrowth.64 In future,Brazil's federal CCT schemesmay be more closelytied to income-generatingopportunities, as in theWorld Bank's 'social riskanalysis' model of social protection.This approach is alreadybeing triedin state-levelincome transferprogrammes in Brazil and has been a featureof similarschemes in Chile, Mexico and Nicaragua.65At the national level in Brazil, however, strategiessuch as agrarianreform, together with concomitant cross-sector policy support, are essentialto help rebuildand sustainrural livelihoods. In addition,poverty alleviationmust be drivenby thecreation of stableand decentlyremunerated employmentopportunities in both ruraland urban areas. Thus, major pov- ertyreduction in Brazil,it is argued,can onlybe achievedthrough vigorous, job-creatingeconomic growth together with redistributive policies and social investment. Buildingprogressive social policycapable of attackingthe roots of poverty and promotingredistributive welfare remains a major challengeto policy- makers.The New Social Policy model of Latin America,with its growing reliance on targetingand safetynets, is seen by some as an appropriate vehicle for reconcilinggrowth with equity,offering an alternativeto more regressive,traditional, institutionalised welfare service deliverymechan- isms.66Despite some promisinginitial evidence from conditional cash transferschemes in the region,however, it remainsto be seen how well the

63 Rawlings,A NewApproach toSocialAssistance, p. i i. 64 D. de Ferranti,et al., Inequalityin LatinAmerica. Breaking with History? World Bank (WashingtonDC, 2004).World Bank, World Development Report 2006. Equity and Development (WashingtonDC, 2005). 65 NineBrazilian states have their own CCT programmes(Alagoas, Ceara, the federal district ofBrasilia, Goias, Mattos Grosso do Sul,Rio deJaneiro, Rio Grandedo Sul,Sio Paulo and Tocantins).For example,the 'Citizen'sIncome' (RendaCidadda scheme in Goias was developedfrom 2000 to substitutethe food parcel programme. It permitsregistered poor familiesto receivecash compensationfor the purchaseof basic foodstuffs.In Rio de Janeiro,the 'Citizen's Cheque' (Cheque Cidadao) initiative was setup in 1999and distributes shoppingvouchers to approvedneedy families. Several of theseCCT projectsinclude vocationaltraining components. See 'Apenas nove estadostem programas pr6prios de transferenciade renda', Valor Econdmico, 31January (2006) and J. Grazianoda Silvaet al., 'The Challengesof a Policyof Food Securityin Brazil'.For otherLatin American ex- see A New toSocialAssistance. 66 periences, Rawlings, Approach Abeland Lewis, 'A Diagnosisof SocialPolicy'. SocialPolicies and Poverty Alleviation under Lula 709 constructionof safetynets such as BolsaEamilia can liveup to theseam- bitiousexpectations in Brazil'scase. Perhaps the crux of thematter is notso muchthe viability of theconcept itself, but how wellit can be appliedto provideunderprivileged sectors of societywith lasting access to basic ser- vicesand job opportunitieswhich may enhance their life chances.