Presentación de Jere R. Behrman, University of Pennsylvania, durante el Tercer Seminario de Transferencias Condicionadas de Ingresos, realizado en Santiago de Chile el 01 y 02 de Diciembre de 2008.
1. University of Pennsylvania Behrman1
Commentary on IFPRI and IFS
Papers
Jere R. Behrman
University of Pennsylvania
III Seminario Internacional
Transferencias Condicionadas
Erradicació n del Hambre y la Desnutrició n Cró nica
Santiago, Chile
1 December 2008
2. University of Pennsylvania Behrman2
• IFPRI 1: “Conditional Cash Transfer
Programs and Nutrition in Latin America:
Assessment of Impacts and Strategies for
Improvement” by John Hoddinott and Lucy
Bassett
• IFPRI 2: “Designing CCT Programs to
Improve Nutrition Impact: Evidence,
Principles, and Questions” by James
Garrett, Lucy Bassett, Marie Ruel and
Alessandra Marini
• IFS: “Old and New Welfare: their relative
effect on Child Nutrition “ by Orazio
Attanasio, Julieta Trias and Marcos Vera-
Hernández
3. University of Pennsylvania Behrman3
What are CCTs? And why of
interest?• Transfers conditional on behaviors as
reflected in explicit indicators
• So what is new? Are scholarships CCTs?
• Demand-side oriented in practice but could
be conditional on supply behaviors as
Garrett et al. note (and, at a certain level,
subsidies for schools and clinics are CCTs)
• Usually multiple conditionalities (e.g,
school enrollment, health checkups, growth
monitoring – Garrett et al. Table 3)
4. University of Pennsylvania Behrman4
• Why so popular?
– Conditionalities chosen have fairly short term indicators so
apparent progress visible, useful given high discount rate
among most policymakers
– Coresponsibilities mean less stigma? more acceptance
throughout society?
– Paternalistic assurance of “good use” of transfers (e.g.,
Hoddinott & Bassett, p. 3 “Behavioral conditions … important
to encourage …households… that undervalue … or are not
willing to undertake risks associated with a good or
service…”; Garrett et al, p. 27: “Policymakers … must think
how to direct changes caused by [income] increase to the
child’s need”)
– Second-best way of dealing with information imperfections
(but then need to be ongoing?)
5. University of Pennsylvania Behrman5
–Redistribution towards children (equalizing?),
women (Martinelli and Parker 2003)
–IFPRI evaluation of PROGRESA gave added
credibility of “arms-length” evaluation
internationally as well as in Mexico
6. University of Pennsylvania Behrman6
• CCTs and Nutrition
–Most CCTs have nutritional component
–Partly in name (i.e. relatively unconditional
component of Mexican program)
–Partly in practice (supplements, information, growth
monitoring) – though debate about effectiveness
(e.g., supplements shared with others)
–Could be made more effective by tying more closely
to outcomes (i.e., child growth rather than accepting
supplements)?
–Attractive vehicle given popularity, but tradeoffs.
7. University of Pennsylvania Behrman7
• Three studies are nice complements regarding
CCTs and nutrition, with some overlaps but also
some distinctive contributions for each.
8. University of Pennsylvania Behrman8
IFPRI 1: Hoddinott and Bassett
• Summarizes IFPRI work on Mexico, Brazil,
Nicaragua and Honduras
• Presents interesting and new, but in some
way troublesome, analysis of multiple
studies on Mexico.
– Striking how difficult is replication , in part because of
data availability.
– Also striking differences among studies, reinforced by
more recent evidence (after 10 years no significant
effects on stunting – though on behaviors and language
abilities, Behrman et al. 2008).
9. University of Pennsylvania Behrman9
• Useful speculation on CCT modifications to
enhance effectiveness regarding nutrition
• Raises question whether CCTs best way to go
(as does IFPRI 2 and as addressed in sense by
IFS)
• Possible limitations
–Not clear that altitude controlled in anemia
estimates.
–Does inclusion of added variables in (1) change
point estimates?
–Costs not considered much
–Efficiency policy motive not explicitly considered
much
10. University of Pennsylvania Behrman10
IFPRI 2: Garrett, Bassett, Ruel &
Marini• Rich, broad perspective on CCTs and nutrition
– both as part of larger system and on their
own with detailed description of possible
channels
• Systematic summary (particularly in tables) of
a number of experiences in LAC.
• In some cases questions about what we really
know (e.g., girl’s education on p. 17, given
micro estimates that suggest associations
overstate such impacts unless control for
endowments – Behrman & Rosenzweig 2002,
Black et al. 2005, Plug 2004 in American
Economic Review).
11. University of Pennsylvania Behrman11
• More elaboration on nature of community effort
(p. 31)
• Possible additional evidence on inefficiencies
(e.g., externalities mentioned on p. 25, but no
empirical estimates)
• Which of six “Essential Nutrition Actions” (p. 41)
particularly suitable for CCTs and which for
other policies?
• Insufficient attention to costs
12. University of Pennsylvania Behrman12
IFS: Attanasio, Trias & Vera-
Hernández
• Interesting comparison of two programs,
Hogares Comunitarios and Familias en
Acció n
• Technical question of further diagnostics
of estimates (Stock and Yugo weak
instrument and overidentification type
tests)
• Interesting that can not find significant
difference, particularly since extent to
which directed towards nutrition differs
strongly; what about cost differences?
13. University of Pennsylvania Behrman13
All Three
• More clarity about cost side. To make
choices would like internal rates of return
or benefit-costs of alternatives
• More clarity about efficiency motive for
policy – differences between private and
social rates of return. Are these “win-
win” options with both efficiency and
(pro-poor) distributional gains likely?
What are the nature of the inefficiences
empirically (capital markets, insurance
markets, information)?
14. University of Pennsylvania Behrman14
• Time and duration effects may be important:
startup, pioneer versus learning effects, path of
effect as duration changes (Behrman & King
2008, King and Behrman 2008)
• To clarify priorities, what would be top three
experiments that each set of authors would
propose to assess improvement in CCTs with
regard to nutrition or whether other programs
better?