Cash Transfers for Economic Strengthening of Families affected by HIV and AIDS

1
Cash Transfers for Economic Strengthening of
Families affected by HIV and AIDS

• Michelle Adato
• International Food Policy Research Institute
• Meeting Children's Needs in a World with
  HIV/AIDS
• An International Symposium
• Harvard Medical School, Boston, MA.
• 24 September 2007

2
Overview of Presentation

• What social protection can achieve
• Why social protection is critical for families
  affected by AIDS
• Overview of cash transfers and how they respond
  to needs
• Impacts of cash transfer programs
• Key policy issues
• Conclusions

3
What Social Protection can achieve
Protective Preventative
Promotional Transformational
Secure basic consumption
Reduce fluctuations in consumption and avert
asset reduction
Enable people to save, invest, and accumulate
through reduction in risk and income variation

• Build, diversify, and enhance use of assets
• Reduce access constraints
• Directly provide or loan assets
• Build linkages with institutions

• Transform institutions and relationships
• Economic
• Political
• Social

• Public works
• Insurance (health, asset)

• Unconditional cash transfers

• Livelihoods programs
• Credit

Conditional cash transfers

• Direct feeding
• Subsidies

• Home-based care
• for the ill

• Maternal and Child Health and Nutrition

• Child and adult education/skills
• Early childhood development

4
Why is Social Protection Critical for Children in
families affected by AIDS?

• Children in families affected by AIDS are at-risk
  with respect to their nutrition, health and
  education
• Human capital underinvestment/loss has lifelong
  consequences for individuals, families and
  communities

5
Cycles of Poverty Interactions between
malnutrition, education, and income

• Poor and malnourished children more likely to
  start school late (Zambia 4x, Uganda 10x, Nepal,
  China, Tanzania)
• Stunted children more likely than healthy
  children to attain lower achievement levels and
  poorer grades (10 countries)
• Stunting associated with lower scores on
  cognitive tests (Guatemala, Philippines, Ecuador)
• Studies from 51 countries show that, on average,
  each year of schooling increases wages by 9.7
• Loss in adult income from stunting (55
  countries) 22.2

(Sources Glewwe, Jacoby and King 2001
Grantham-McGregor et al. 2007 Pollitt et al.
1995 Martorell 1995, 1999 Grantham-McGregor et
al. 2007 Behrman, Alderman, Hoddinott 2004
Strauss and Thomas 1997)
6
Types of Cash Transfer Programs

• Cash support usually targeted toward poor and
  vulnerable groups
• Household grants, child support grants, old age
  pensions, foster care grants, disability grants
• Unconditionalgiven without obligations
• Unconditional complementary activities
• Conditionalcash grant to mother, with
  obligations
• School attendance (about 80) gtgirls than boys
• Preventative health care such as pre-natal and
  post-natal check-ups, vaccinations, growth
  monitoring
• Participation in training health, nutrition,
  skills
• Participation in work
• Soft conditions condition not enforced may
  sign or consent to oral agreement to meet
  condition no sanctions

7
Cash Transfers for AIDS-affected households

• Targeted toward poor households
• Often focusing on incapacitated households with
  high child dependency ratios
• Mostly unconditional, but conditionality under
  exploration
• Sometimes integrate activities providing
  information and services

8
Impacts of Cash Transfer Programs
9
Impacts of Conditional Cash Transfers on
Education (Enrollment)
(Sources Schultz 2001 Skoufias 2005 IFPRI
2003 Maluccio and Flores 2005 Filmer and
Schady 2006 Ahmed 2006 Khandker, Pitt, and Fuwa
2003 Ahmed et al. 2007)
10
Impacts of Conditional Cash Transfers on Health
and Nutrition
(Sources Skoufias 2005 Gertler 2000 Hoddinott
forthcoming IFPRI 2003 Maluccio and Flores
2005)
11
Impacts of Unconditional Transfers on Education

• South Africa, Child Support Grant
• 25 reduction in non-attendance gap (national)
• 8.1 point increase in enrolment among 6 year
  olds 1.8 for 7 year olds (KwaZulu-Natal)
• No evidence on school-related expenditureimpacts
  due to health and nutrition impacts?
• South Africa, Old Age Pension
• In households receiving pension, 3 increase in
  enrollment for boys 7 increase for girls
• 33 reduction in non-attendance gap if pension
  received by a female, no impact if received by
  male
• Zambia, Pilot Cash Transfer
• 3 pt increase in enrollment for children 7-18
• Mainly for age 5-6 (10.4) not enough to
  encourage secondary much greater impact for boys
  than girls
• Number of HHs not sending at least one child to
  school decreased by 7.6 pts

(Sources Samson 2004 Case et al 2004 Govt. of
Zambia/GTZ 2006)
12
Impacts of Unconditional Transfers on Health

• Zambia Incidence of illnesses in previous month
  decreased by 7.8
• Malawi (FACT) Second largest item purchased with
  grant after food is health care
• Malawi (DECT) Transfer gave beneficiaries access
  to health services by helping pay transport
  costs, hospital bills, and medicine, reported
  overall improvements in health status
• Ethiopia 75 of beneficiaries used health
  facilities more than prior to program 76 of
  these credited program

(Sources Devereux et al. 2006a Devereux et al.
2007 Devereux et al 2006b Govt. of Zambia/GTZ
2006)
13
Impacts of Unconditional Cash Transfers on
Nutrition

• South Africa CSG in KwaZulu-Natal
• Male child treated before age 1 and receiving
  benefits for two-thirds of the first 3 years of
  life has gains in height-for-age of .4 or approx.
  3.5cm, leading to 2.1 gain in adult height
• No nutritional gain if received for under 20 of
  period age 0-3 years
• Increased future earnings estimated to be 60-130
  higher than the cost of CSG support
• South Africa OAP
• Pension associated with a 5 cm. increase in child
  height (Western Cape)
• Pension associated with 2.23 cm. increase in
  girls height .88 cm for boys, over two year
  period, where pensioner is a woman

(Sources Agüero, Carter and Woolard, 2007 Case
2001 Duflo 2000)
14
Impacts of Unconditional Cash Transfers on Food
Consumption

• Zambia
• HHs consuming only one meal a day - 6 pts.
• HHs reporting hunger pangs after a meal - 21
  pts.
• HHs consuming vitamins and protein 7 days a
  week 12 pts.
• Mozambique
• No effect because transfer size was too small
  (3-6/month)
• Malawi
• FACT transfers protected food consumption in
  beneficiary households against the rationing and
  hunger during 2006 crop failures and food price
  seasonality
• Beneficiaries consumed more than
  non-beneficiaries (10-44 more meals/month),
  though they had started hungry season with lower
  food consumption than non-beneficiaries (8-15
  fewer meals/month)

(Sources Govt. of Zambia/GTZ 2006 Low et al.
1999 Devereux et al. 2006b)
15
Use of Transfer
(Sources Schubert 2007 Govt. of Zambia/GTZ
Acacia Consultants 2007 Devereux 2002 Devereux
et al. 2006a)
16
Policy Issues Conditionality and Targeting
17
Why Condition transfers?

• Externalities parents may under-invest in
  childrens schooling relative to optimal levels
  from societal standpoint, e.g. girls orphans in
  fostering families
• Conditioning may be more effective at increasing
  human capital
• Mexico Conditioning increased likelihood that
  students continue to secondary school by 20
  points.
• Ecuador Conditioning increased enrollment by
  7-13 pts.
• Kenya, Zambia, Uganda, Burkina Faso comparing
  conditional vs. unconditional transfers
• Sociocultural biases may be imposed by more
  powerful groups upon the less powerfulconditional
  ity can provide state legitimation of social
  change
• Can overcome stigma as beneficiaries enter into
  social contract with the state
• Increases public support as policymakers are able
  to show indicators of progress in education and
  health
• CCTs serve as a strong impetus to improve services

18
Why Not to Condition Transfers

• Insufficient access to nearby facilities of
  sufficient quality
• Government or NGOs do not have the financial and
  administrative capacity to manage process,
  coordinate deliveryor if the costs exceed the
  benefits
• Costs in Honduras 9 Nicaragua 3
• Cost to household of meeting conditions may
  exceed benefits
• People may have different preferences from
  development planners or prefer to make their
  own choicesseen as demeaning
• Poverty, culture, social exclusion,
  discrimination, illness, or trauma may prevent
  participant despite benefits, such that
  conditionality punishes those who most need
  assistance

19
Considering conditioning in the context of AIDS

• What are the objectives? Condition on thosedont
  use blueprints
• Conditioning can be used creatively to respond to
  specific needs of AIDS-affected families, e.g.
• Early Childhood Development for vulnerable
  children
• STI testing and health informationpilots
  proposed in South Africa and Tanzania
• Use non-compulsory complementary activities

20
Complementary Activities with Cash Transfers

• Kenya
• Lectures on child and maternal health, prevention
  and treatment of illness (HIV, malaria, STDs),
  nutrition (for CCT and UCT)
• ARV referral and UNIMIX (for malnourished
  children and pregnant and lactating women) under
  consideration
• Malawi
• DECT Information and behavioral messages on HIV
  and AIDS on pay-days
• Social Cash Transfer Pilot Scheme Linkage and
  integration with health services and home-based
  care considered in next phase
• Mozambique
• Psychosocial support and income-generating
  projects

(Sources Government of Kenya 2006 Devereux et
al. 2007 UNICEF Save the Children/HelpAge
International/IDS, 2005)
21
Data-based or community-based targeting?

• Geographic targeting is usually data-based
• Conditionality introduces self-targeting
• Household identification
• Proxy-means tests via surveys (most CCTs)
• Exhaustive list of variables that are proxies for
  poverty
• Proxy-means via application
• E.g. in South Africa, self-reported documentation
  of income, type of dwelling and area of residence
• Community-based (e.g. Zambia, Malawi, Kenya)
• Data community review

22
Effects of Proxy-means test

• Proxy-means test are very effective in
  identifying the poorest but can be expensive and
  management-intensive
• For 3 Latin American countries, averaged between
  12-34 of program activity costs
• But they capture only outsider definitions of
  poverty, without local knowledge
• In Nicaragua and Mexico, process not transparent
  and most people did not understand it
• Widespread perceptions of errors of exclusion
• In Mexico, particularly in early years, this bred
  social conflict

(Sources Caldes et al. 2004 Adato 2000 Adato
and Roopnaraine 2004)
23
Errors of exclusion and inclusion

• Exclusion ( excluded poor/ poor)
• Inclusion ( non-poor / beneficiaries)

Geog. targeting
HH targeting
24
Community-Based Targeting

• Zambia, Kenya, Malawi (variations on approach)
• Community committees lists, interview and rank
  potential eligible households, based on criteria
  representing poverty, incapacitation and high
  dependency rationsselects beneficiaries
• Community reviews ranking, public discussion and
  consensus
• Generally assessed as effective
• Zambia and Malawi approximately 70 of
  beneficiary households estimated to be affected
  by HIV/AIDS
• Problems with community-based targeting
• Very time-consuming (universal more efficient?)
• May be missing some households in need How do
  outcomes and costs compare with proxy-means
  methods?
• Comparisons of alternative targeting methods
  planned

(Source Govt. of Zambia 2006 Schubert 2007
Pearson et al. 2006 Devereux et al. 2006 2007)
25
Who to Target?

• Poverty or AIDS?
• Stigma How to reach AIDS-affected households
  without stigmatizing them
• In a study in Kenya of food assistance
  distributed with ARVs, people reported
  confidentiality problems and stigma
• Households affected by AIDS most in need of
  benefits
• But many other households are also very
  poorequity considerations
• Targeting poverty and AIDS
• In Zambia, e.g. community-based criteria includes
  extreme impoverishment/hunger no able-bodied
  working-age members high dependency ratios no
  valuable assets or regular substantial income
• The household or children via women?
• Extensive evidence that income controlled by
  women translates into higher calorie intake for
  children, and expenditures on food, health, and
  education, than income controlled by men
• Need means of reaching child-headed households

26
Conclusions

• Critical to rapidly scale-up social protection
  for children in families affected by AIDS
• Reduce acute poverty and protect against mass
  loss of human capital with irreversible, lifelong
  consequences
• Unconditional and conditional transfers are
  effective
• More robust evidence on human capital impacts of
  conditional transfers convincing and
  strengthening evidence of impacts of
  unconditional transfers
• Immediately roll-out unconditional transfers,
  while piloting context-specific, creative
  conditionalities, with comparative evaluations
• Need strong emphasis on supply side components
• Start with non-mandatory complementary activities
• Target poverty explicitly, while capturing
  AIDS-affected through effective proxies (maybe
  exception with ARVs?)more work needed on
  identification methods
• Make women the official beneficiary (test further
  in Africa?)
• Pay attention to other complexities and
  specificities of an AIDS-affected context