Content uploaded by Wameq A Raza
Author content
All content in this area was uploaded by Wameq A Raza on Jul 11, 2017
Content may be subject to copyright.
Debating Graduation
Volume 14, Issue No. 2 • July 2017
A publication of
The International Policy Centre for Inclusive Growth
Policy in Focus is a regular publication of the
International Policy Centre for Inclusive Growth (IPC-IG).
The International Policy Centre for Inclusive Growth (IPC-IG) is a
partnership between the United Nations and the Government of Brazil to
promote South–South learning on social policies. The Centre specialises
in research-based policy recommendations to foster the reduction of
poverty and inequality as well as promote inclusive growth. The IPC-IG is
linked to the United Nations Development Programme (UNDP) in Brazil,
the Ministry of Planning, Development, Budget and Management of
Brazil (MP) and the Institute for Applied Economic Research (Ipea)
of the Government of Brazil.
Director: Niky Fabiancic
IPC-IG Research Coordinators: Diana Sawyer; Fábio Veras Soares;
Rafael Guerreiro Osorio and Luis Henrique Paiva
The views expressed in IPC-IG publications are solely those of the authors
and should not be taken as representing the views of the United Nations
Development Programme or the Government of Brazil.
Rights and Permissions – All rights reserved. The text and data in this
publication may be reproduced as long as written permission is obtained
from the IPC-IG and the source is cited. Reproductions for commercial
purposes are forbidden.
Some of the photographs used in this publication are licensed under
The Creative Commons license; full attribution and links to the individual
licenses are provided for each. Special thanks to Emily Coppel (BRAC USA)
for her invaluable support.
Editor-in-Chief:
Michael MacLennan, International Policy Centre for Inclusive Growth (IPC-IG)
Specialist Guest Editors:
Fábio Ver as Soares , International Polic y Centre for Inclusive Growth (IPC-IG)
and Ian Orton, consultant.
Publications Manager: Roberto Astorino
Copy Editor: Jon Stacey, The Write Effect Ltd.
Art and Desktop Publishing: Flávia Amaral and Rosa Maria Banuth
Editorial Assistant: Manoel Salles
Cover photograph: BRAC. Once destitute and homeless, Jorina, a BRAC
microfinance client in Bangladesh, after nine years now owns the largest
general store in her area, along with a profitable rice business. She lives
in a large brick house on her own land, and is a proud member of her
community. Her greatest joy is to help and support her two sons and
parents as well as other poor villagers.
Editor’s note: On behalf of the IPC-IG, I would like to extend a special
thanks to the specialist guest editors, Fábio Veras Soares and Ian Orton
for their dedication to the publication of this issue. We would also like
to express our sincere appreciation to all the authors for their generous
and insightful contributions, without which this special edition simply
would not have been possible.
Summary
7 Graduation: an overview
11 What does the future hold for graduation?
17 The Graduation Approach within social protection: opportunities for going to scale
22 The effectiveness of the Graduation Approach: what does the evidence tell us?
29 The labour markets of the ultra poor
33 Can graduation approaches contribute to building social protection floors?
36 What we know about graduation impacts and what we need to find out
40 Responsible graduation
43 (Accidentally) Harvesting higher hanging fruits: addressing under-5
malnutrition using the Graduation Approach
47 Challenges for addressing child poverty in Malawi through graduation
52 Digital inclusion for the ultra poor: the Graduation Approach
55 Caveat emptor: the Graduation Approach, electronic payments and the
potential pitfalls of financial inclusion
58 Resilience and graduation
62 Leaving no one behind: graduation for refugees
67 Private-sector investment capital in graduation: it is time to unlock
sustainable financing at scale
6
Since its inception in Bangladesh in 2002, the Graduation
Approach has received much attention, including in
mainstream media outlets. Beyond this positive media
acclaim, momentum has gathered behind graduation
as an important social policy instrument. There has
been a proliferation in the implementation of new
graduation-inspired programmes.
Primarily, graduation has been advanced as an effective
means to combat extreme poverty and embodies part
of the ‘big push’ to achieve Sustainable Development
Goal 1: “End poverty in all its forms everywhere”.
It is one of the most thoroughly evaluated poverty
reduction programmes ever, and its putative results
are resoundingly positive, which helps explain the
surge in interest. However, the increased enthusiasm
and visibility enjoyed by the Graduation Approach has
not been free from controversy. Significant concerns
linger—centring on targeting efficacy and equity and
what happens post-graduation (i.e. after households
exit the programme)—and impact results have been
vehemently contested.
Nevertheless, the buzz continues to grow, and
thus graduation-type programmes merit further
examination. Given this groundswell of interest,
this special issue of Policy in Focus attempts to capture
the diversity of views that exist in the debate. The
articles feature a veritable smorgasbord of perspectives,
ranging from those of committed proponents and
enthusiastic new implementers, to the cautiously
optimistic who reason that graduation could be a
valid component of wider social protection systems,
to outright contestation.
Today, the Graduation Approach has arguably arrived
at an inflection point. The debate on its role and
effectiveness remains to be settled. There might be
increased take-up, or it might recede into obscurity—
it may even possibly be repurposed into other hybrid
programmatic forms. Whatever its destination, we hope
that this publication contributes to promoting a better
understanding of this significant policy development
and stimulating the debate even further.
Fábio Veras Soares and Ian Orton
Editorial
The International Policy Centre for Inclusive Growth | Policy in Focus 7
Graduation: an overview
Fábio Veras Soares1 and Ian Orton2
According to the Oxford English
Dictionary, ‘graduation’ refers either to
the act of “receiving or conferring an
academic degree or diploma” or “the
action of dividing into degrees or other
proportionate divisions on a graduated
scale” (Oxford Dictionaries 2017). The
so-called ‘Graduation Approach’ thus
may be dened as (i) reaching a state in
which one has exited/escaped (extreme)
poverty, based on a given poverty metric
and, therefore, can be considered ready
to ‘graduate’ from the interventions
dedicated to enable this transition; or
(ii) the act of going through a set of
phased-in and overlapping interventions
meant to improve the well-being of their
participants. As we will see throughout this
special edition of Policy in Focus – Debating
Graduation, the ideas of surpassing
a predetermined threshold3 and of a
continuum of phases are embedded in the
dierent approaches to graduation as well
as to social protection more broadly.
From Bangladesh…
The Graduation Approach was born from
recognition that the poorest households
in rural Bangladesh—invariably referred
to as the ‘ultra poor’—were so marginalised
that they were not in a position to engage
with BRAC’s mainstream development
projects. For instance, they were too
poor to access BRAC’s micronance
programmes, which were intended to
boost their livelihoods in a sustainable way.
To ensure that ultra-poor women could also
directly benet from micronance, a multi-
pronged, and eponymously dual-named,
programme—Challenging the Frontiers of
Poverty Reduction: Targeting the Ultra Poor
(CFPR-TUP)—was designed and tested.
Phase 1 of CFPR-TUP (2002–2006) brought
together a series of innovative features
in poverty reduction programmes.
They feature ve main characteristics:
1. a focus on able-bodied women who
were mostly engaged in domestic work
or begging and did not benet from
micronance and/or other development
projects led by the government or
non-governmental organisations
(NGOs) but who had access to small
plots of land (less than 40.5 m2) and who
lived in extremely poor households.
These households should not have any
economically active male member,
with children in the household involved
in (or at risk of) child labour, and
household members should have no
access to productive assets. Participants
were selected through a participatory,
community-based wealth-ranking
targeting approach taking all these
inclusion and exclusion criteria into
account (Ahmed et al. 2009);
2. one-time asset transfers (livestock in
most cases) combined with time-bound
but regularly paid cash or in-kind
transfers to support the consumption
of families and satisfy their vital needs.
This approach has avoided potential
asset-selling by addressing immediate
consumption needs;
3. income-generating activity training
and regular social worker/case
worker visits focusing on savings
and nancial literacy;
4. links with social protection, particularly
facilitating access to health services; and
5. evidence-building on how well the
graduation intervention worked by
undertaking a series of studies and
impact evaluations.
Positive impacts as documented in
early impact evaluations based on non-
experimental design brought much
attention to CFPR-TUP.4 These evaluations
revealed positive impacts across many
dimensions, including: per capita income;
food security; occupational shifts towards
self-employment; asset holdings; savings;
access to sanitation; and clothing; but
little or no impact on children’s school
enrolment, health-related outcomes and
women’s empowerment.
Nevertheless, the lack of experimental
impact evaluations (using randomised
control trials—RCTs) as well as evidence
from other countries raised concerns that
impacts were biased due to the lack of a
proper counterfactual (as in experimental
designs), and that the approach was too
anchored in Bangladesh’s realities to be
applied worldwide with similar success.
The second phase of CFPR-TUP (2007–
2011) was planned with some design
changes. Potential participants were
classied into two groups—namely, the
‘specially targeted ultra poor’ (STUP) and
the ‘other targeted ultra poor’ (OTUP).
The main dierence in this approach
was that micronance was the main
entry point for the OTUP, and the asset
component of the graduation package
was nanced by BRAC through a
‘soft loan’ model (Bandiera el al. 2013).
The expectation was that the asset
component would be paid back to BRAC
with interest rates at around 20 per
cent, some 5 per cent less than BRAC’s
mainstream micronance loans. The STUP
received the comprehensive package
with all components described above,
but with no expectation of paying BRAC
back later, with a view to building an asset
base for the participants so that after
24 months they could be graduated into
“mainstream development activities such
as micronance” (Raza et al. 2012).
Even among the STUP, two dierent
packages were developed to consider
the geographic and demographic
heterogeneity among the ultra poor. The
dierentiated packages implied a dierent
size of asset transfers and intensity of social
work visits. Higher benets (asset value)
and more frequent visits were oered to
those living in areas with higher poverty
density and depth (ibid.).
This second phase of TUP was also an ideal
opportunity to implement an RCT. Forty
BRAC branch oces covering 1,309 villages
were randomised into treatment or control
groups over a four-year period to allow time
for a rigorous evaluation of the programme.
The results reported in two articles of this
issue of Policy in Focus—Clare Balboni et
al.’s ‘Labour Markets of the Ultra Poor’ and
Wameq Raza’s ‘(Accidentally) Harvesting
higher hanging fruits: Addressing the
8
under-5 malnutrition using the Graduation
Approach’—use the data collected under
this experimental design to assess Phase 2
of the TUP programme.
Their ndings conrm the positive impacts
provided by the non-experimental
evaluations. The labour supply of
participating women increased, mostly
through the allocation of more hours to
rearing livestock (self-employment). This, in
turn, enabled them to experience increases
in earnings (21 per cent), in per capita
expenditure (11 per cent) and in the value
of durable goods owned by the household
(57 per cent). Similar impressive positive
impacts were found for the increased value
of savings (400 per cent), livestock (200
per cent), other productive assets (159 per
cent) and land (82 per cent). In addition,
Raza reports that under-5 children whose
families participated in the programme
experienced a reduction in malnutrition
as measured by wasting (low weight-to-
height ratio) and underweight, but not
for stunting (low height-to-age ratio). He
also reports positive spill-over eects for
children from participating villages whose
parents were not participating in the
programme. These results are important,
as human development impacts on
education and health were not found in
the rst phase of CFPR-TUP evaluations.
These results address the concerns
regarding the ‘buzz’ and alleged hyperbole
around the Graduation Approach.
However, a second concern—related to
its adaptability to other contexts—was
rst addressed through the Graduation
into Sustainable Livelihoods project led
by the Consultative Group to Assist the
Poor (CGAP), with support from the Ford
Foundation, in 2006. The project aimed to
adapt the Graduation Approach through
10 pilots implemented in eight countries:
Ethiopia, Ghana, Haiti, Honduras, India
(2 pilots), Pakistan, Peru and Yemen.
They provide evidence of its adaptability
and suitability beyond Bangladesh. All of
the pilots had embedded robust learning
and evaluation components.
…to the world
Six pilots were individually and
collectively evaluated by researchers
from the Abdul Latif Jameel Poverty
Action Lab (J-PAL) and Innovations
for Poverty Action (IPA ). Banerjee et al.
(2015) summarised their ndings in an
inuential article in Science magazine.
In this Policy in Focus, Nathanael
Goldberg summarises the results of these
evaluations and nds positive impacts
on: income and revenues; total per capita
consumption; assets; food security; nancial
inclusion; mental health; total time spent
working; and political involvement, for
both immediately after the end of the
programme and a period of one year later.
For women’s empowerment and physical
health, impacts were positive just after the
end of the interventions, but they were no
longer statistically signicant one year after.
In Honduras, the programme failed to
achieve most of these positive impacts,
as a disease killed most of the assets
distributed by the programme (poultry),
illustrating the risk of relying on one
income-generating asset only.
Similarly, the evaluation of the pilot
programme in the south Indian state of
Andhra Pradesh undertaken by Bauchet
et al. (2015) failed to nd positive impacts
on income, asset accumulation or
consumption. The authors claim that a
tight labour market (i.e. with high wage-
employment rates)—not observed in other
areas where the Graduation Approach
has been implemented—led to osetting
impacts on income and time use. This
local context seems to explain a drop of 57
per cent in terms of livestock ownership
at the end of the intervention. This raises
the question of the pertinence of asset
transfers in similar contexts.
Goldberg also outlines a future research
agenda in which the need to evaluate the
contribution of individual components of
the graduation package will be a priority.
He argues that the cost of some
components may be beyond the reach
of governments in low-income countries
if they are to be scaled up, particularly
the labour-intensive case management
visits. Testing the impacts in contexts with
fewer home visits, with group rather than
individual sessions and with the use of
tablets for self-taught nancial literacy
tutorials are examples of recent evaluation
processes. Additionally, dimensions such
as women’s empowerment, alternative
targeting methods and testing special
packages for those who fail to succeed with
the standard components and ‘intensity’
of the Graduation Approach are other
important aspects of the research agenda.
The expansion and adaptation of the
Graduation Approach worldwide is the
focus of the two opening articles of this
issue: ‘The Graduation Approach with
Social Protection: Opportunities for
Going to Scale’ by Aude de Montesquiou
and Syed Hashemi, and ‘What does the
future hold for Graduation?’ by Harshani
Dharmadasa et al. With 57 programmes
in almost 40 countries, the Graduation
Approach has been adapted to a diverse
range of contexts, including urban areas,
replacing the focus on assets with an
emphasis on employment opportunities
where appropriate, including men as
explicit participants, targeting beyond
the ultra-poor ‘poverty line’ and including
other categories, such as refugees and
internally displaced people, indigenous
groups, people with disabilities, youth
and elderly people. Moreover, a growing
number of programmes are being
implemented by governments, rather than
by NGOs or donors. Among the poverty-
related dimensions envisioned by the
Graduation Approach, there has been an
evolution towards greater engagement
with: nancial exclusion; child poverty;
climate change; fragile and conicted-
aected regions; and youth employment.
The growing popularity of the Graduation
Approach has also attracted stark
criticism. Stephen Kidd and Diloá Bailey-
Athias’s article, ‘The eectiveness of the
Graduation Approach: What does the
evidence tell us?’, delves deeply into the
details of many of the impact evaluations
of graduation interventions.
The authors highlight the high level of
inclusion errors in the programmes—
the proportion of those living above the
extreme poverty line over the total number
of participants—reaching over 50 per cent
in the cases of Peru and Pakistan, and also
point out the discrepancy between gains
in household consumption—deceptively
high in terms of percentage but in fact
extremely modest in absolute terms.
Echoing Banerjee et al. (2015), the
authors acknowledge that these average
eects would not be large enough
to liberate programme participants
from a poverty trap.
The International Policy Centre for Inclusive Growth | Policy in Focus 9
The article also suggests that the positive
impact of the Graduation Approach is
exaggerated by the advocacy rhetoric and
needs to be put into perspective—even
more so when simple and large-scale
social cash transfers seem to have similar
impacts and cover broader segments
of the population.
Edward Archibald, in his article ‘Challenges
for Addressing Child Poverty in Malawi
through Graduation’, analyses the risks
involved in the adoption of the Graduation
Approach as an ‘exit strategy’ from social
cash transfer programmes. The author
argues that this may lead to unrealistic
expectations regarding the potential of
cash transfer beneciaries to exit poverty
permanently, and questions the costs
involved for governments to implement
a programme that requires intensive use
of social workers for its training and case
management components. This discussion
raises important ethical questions centring
on the implications of ‘graduating’ people
into potential contingency and the void
between social programmes.
The relationship between the Graduation
Approach and social protection is at
the core of a erce debate. In the article
‘Can Graduation Approaches Contribute
to Social Protection Floors?’, Christina
Behrendt argues that in many cases the
concept of graduation from poverty
is misunderstood as graduation from
a specic programme or from social
protection altogether. In her view, this
is highly problematic, as it assumes that
social protection is only for those living
in extreme poverty and/or that social
protection and income generation/
employment are not compatible,
which is certainly not the case.
Similarly, Keetie Roelen et al., in their
article ‘Responsible Graduation’, claim
that it is possible to reconcile the
Graduation Approach with the ‘right
to social protection’. In their view,
responsible graduation would facilitate
this process by focusing on endogenous
graduation from poverty based on
clear welfare improvement measured
by well-dened indicators, rather than
by simply completing an exogenously
dened programme cycle of two
years. Participants should be allowed
to re-enter the programme in case of
need—a revolving door, rather than
a one-way door. Thus, more tailored
responses should be developed for
potential participants, taking into account
household and community needs. In
this vision, graduation could be seen as
a continuous pathway, through which
participants are graduated into other forms
of social protection adequate to their
living standards and their location along
their life course, as proposed by Samson
(2015). Responsible graduation would also
encompass grievance mechanisms; more
focus on stable employment, rather than
just reliance on self-employment; and the
acknowledgement that some extremely
poor households are not suitable for the
graduation programmes and most likely
will need social assistance (social cash
transfers) for indenite periods.
The features put forward by Roelen
and others in their article seem to be in
line with the idea that participants of
responsible graduation programmes
will become more resilient to shocks.
Moving forward, Greg Collins, in his article
‘Resilience and Graduation’, stresses
that it is crucial to overcome the divisive
opposition between investment in
social protection (social cash transfers)
and in graduation options, highlighting
the importance of some components
of the Graduation Approach in making
families more resilient to shocks. One of
the resilience components in graduation
pinpointed by the author is the nancial
inclusion aspect.
The following article by Tatiana Rincón,
‘Digital Inclusion for the Ultra Poor: The
Graduation Approach’, presents some
innovative uses of digital solutions for
training and also some innovations in
the use of an e-payment infrastructure
to facilitate both nancial inclusion and
nancial literacy. She highlights the
‘Microsavings with a purpose’ project,
which is being piloted in Paraguay in a
partnership with the government and a
telecom company, Fundación Capital.
However, as interesting as these
developments are, such approaches
also raise some concerns. In their article,
‘Caveat Emptor: The Graduation Approach,
Electronic Payments and the Potential
Pitfalls of Financial Inclusion’, Paulo dos
Santos and Ingrid Kvangraven argue that
the geographic distance of electronic
banks from their borrowers in low-income
areas makes them far less likely to engage
in lending to new productive enterprises
than traditional micronance institutions.
Traditional micronance applies a ‘social
technology’ approach, in which the social
connection among borrowers (clients) and
between borrowers and lenders serves as
the knowledge base to support informed
loan decisions. Therefore, e-payment
providers involved in micronance end
up having an incentive to rely more on
consumption loans than business-related
loans, leading poor borrowers to be highly
indebted. To the authors, these micro-level
innovations are no substitute for national
industrialisation and broader social
policies that can address structural and
systemic hurdles and deciencies linked
to chronic underdevelopment.
The article ‘Leaving no one behind:
Graduation for Refugees’ by Helene Kuhle
et al. provides another example of an area
where the Graduation Approach looks
poised to feature more prominently in
the future. The authors highlight the work
undertaken by the United Nations High
Commissioner for Refugees (UNHCR)
regarding graduation, with a view to
bridging the gap between humanitarian
and development policies. The Graduation
Approach is seen as one of the possible
ways to increase self-reliance and
resilience among both refugees and host
communities whose members live in
extreme poverty. Given the unprecedented
forced displacement of entire populations
in recent years, this represents a crucial
new demographic addressed by
graduation. The urgency of this approach
is underscored further when one considers
that 93 per cent of the people living in
extreme poverty in the world today live
in a context of humanitarian crisis (Global
Humanitarian Assistance 2015).
Despite a recent push back, the general
trend in the near term is undoubtedly
towards further take-up, expansion
and increased diversication of the
Graduation Approach.
A new era involving greater use of
graduation as a social policy presupposes
adequate nancing. The costs of the model
are not insignicant, and the initial upfront
investment required may be too high for
10
some resource-constrained governments
to bear alone. This is complicated further
by a challenging global nancial context
where nancing for development is at best
uncertain. New sources of or approaches
to nancing might be required.
The role of philanthropic and private
capital in partnership with governments
to pursue development goals is not new.
However, in the article ‘Private-sector
investment capital in Graduation: It is
time to unlock sustainable nancing at
scale’, Shaifali Puri and Anne Hastings
introduce the reader to an emergent
nancing model tasked with combating
extreme poverty through graduation. The
authors suggest that an approach based
on social and development impact bonds
could contribute to solving the nancing
conundrum. They explain the impact
bonds model, its relevance to graduation,
the contributions it could make towards
the eradication of poverty, and how they
can ll the funding gap. Some in the
development space might be sceptical,
arguing that looking for investment returns
in the sphere of development runs contrary
to a rights-based approach, but it might
be worth keeping an open mind to such
innovation and realpolitik pragmatism.
Moreover, this endeavour could open the
door for further development nancing
by bringing in more would-be private
development capital into the graduation
and development fold.
This collection of articles hopes to
contribute to a better understanding of the
origins and the ongoing transformation
of the Graduation Approach, and of its
interactions with social protection. The
idea of a multi-pronged package to
address dierent aspects of poverty is not
alien to social protection itself. Instruments
such as consumption support (social
transfers), training for income-generating
activity and coaching are key aspects
of social protection and can be seen as
essential and classic components of social
assistance and labour market policies
(including support for self-employed
people and entrepreneurship).
The logic of such an approach can also be
found in some components of conditional
cash transfers in Latin America or in broader
strategies that use these programmes
and their monitoring and information
systems to facilitate access to services and
integrate social protection programmes.
The Mexican Prospera, the Brazilian Brasil
sem Miséria strategy nestled in the Bolsa
Família programme, and Chile’s Ingreso Etico
Familiar (formerly Chile Solidario) are all good
examples. While these programmes were
not directly inuenced by the Graduation
Approach, they recognised (around the
same time as BRAC did) the need for
multifaceted approaches to respond to
local challenges and the pressing need
to have a systemic response to poverty.
Over time, the strategies that have hinged
mostly on cash transfers were linked to
other programmatic dimensions beyond
the cash component, encompassing
areas such as health, education, nutrition,
case management, productive inclusion
and, in some cases, access to nancial
services (Dharmadasa et al. 2017). The
adoption of responsible graduation logic
can strengthen existing programmes by
combining complementary mechanisms
and expanding social protection coverage.
The proliferation of graduation-inspired
approaches has the potential to provide
good examples of how to mainstream it into
nationally owned social protection oors.
In any case, what is clear is that the debate
on graduation will rumble on unabated
for some time to come.
Ahmed, A.U., M. Rabanni, M. Sulaiman, and
N.C. Das. 2009. “The impact of asset transfer
on livelihoods of the ultra poor in Bangladesh.”
Research Monograph Series No. 39. Dhaka:
Research and Evaluation Division, BRAC.
Bandiera, O., R. Burgess, S. Gulesci, I. Rasul,
and M. Sulaiman. 2013. “Can entrepreneurship
programs transform the economic lives of the
poor?” LSE Working Paper. London: London
School of Economics. <sticerd.lse.ac.uk/dps/
eopp/eopp43.pdf>. Accessed 7 June 2017.
Bandiera, O., R. Burgess, N. Das, S. Gulesci,
I. Rasul, and M. Sulaiman. 2016. “Labor markets
and poverty in village economies?” LSE Working
Paper. London: London School of Economics.
<sticerd.lse.ac.uk/dps/eopp/eopp58.pdf>.
Accessed 7 June 2017.
Banerjee, A., E. Duo, N. Goldberg, D. Karlan,
R. Osei, W. Parienté, J. Shapiro, B. Thuysbaert, and
C. Udry. 2015. “A multifaceted program causes
lasting progress for the very poor: Evidence from
six countries.” Science 348(6236). <http://science.
sciencemag.org/content/348/6236/1260799>.
Accessed 7 June 2017.
Bauchet, J., J. Morduch, and S. Ravi. 2015. “Failure
vs displacement: Why an innovative anti-poverty
program showed no net impact in South India.”
Journal of Development Economics 116: 1–16.
Das, N.C., and F.A. Misha. 2010. “Addressing extreme
poverty in a sustainable manner: Evidence from
CFPR Programme.” CFPR Working Paper No. 19.
Dhaka: Research and Evaluation Division, BRAC.
Dharmadasa, H., I. Orton, and L. Whitehead.
Forthcoming 2017. Realising the Human Right
to Social Protection through Graduation. Brasìlia:
International Policy Centre for Inclusive Growth.
Emran, M.S., V. Robano, and C.S. Smith. 2009.
“Assessing the frontiers of ultra poverty
Reduction: Evidence from Challenging the
Frontiers of Poverty Reduction/Targeting the
Ultra-poor, an innovative program in Bangladesh.”
mimeo. Washington, DC: Department of
Economics, George Washington University.
Emran, M.S., V. Robano, and C.S. Smith. 2014.
“Assessing the frontiers of ultrapoverty reduction:
Evidence from Challenging the Frontiers of
Poverty Reduction/Targeting the Ultra poor, an
innovative program in Bangladesh.” Economic
Development and Cultural Change 62(2): 339–380.
Global Humanitarian Assistance. 2015. Global
Humanitarian Assistance Report 2015. Bristol,
UK: Global Humanitarian Assistance. <http://
www.globalhumanitarianassistance.org/wp-
content/uploads/2015/06/GHA-Report-2015_-
Interactive_Online.pdf>. Accessed 7 June 2017.
Misha, F.A., W. Raza, J. Ara, and E. Van de Poel.
2014. “How far does a big push really push?
Mitigating ultra-poverty in Bangladesh.”
ISS Working Paper, No. 549. Rotterdam:
International Institute of Social Studies.
Oxford Dictionaries. 2017. “Graduation.”
Oxford Dictionaries website. <https://en.
oxforddictionaries.com/denition/graduation>.
Accessed 7 June 2017.
Raza, W.A, and E. Van de Poel. 2016. “Impact and
spillover eects of an asset transfer programme
on malnutrition – Evidence from a randomized
control trial in Bangladesh.” BRAC Research
Monograph Series 64. Dhaka: BRAC.
Raza, W.A., N. Das, and F.A. Misha. 2012.
“Can ultra poverty be sustainably improved?
Evidence from BRAC in Bangladesh.” Journal
of Development Eectiveness 4(2): 257–276.
Samson, M. 2015. “Exit or Developmental Impact?
The Role of ‘Graduation’ in Social Protection
Programmes.” IDS Bulletin 46(2): 13–24.
1. International Policy Centre for Inclusive
Growth (IPC-IG).
2. Consultant.
3. The notion of passing over a threshold to
determine programmatic exit is not universally
accepted nor without controversy. Resistance
to this notion is a recurrent theme throughout
this publication, raising concerns about the
possibility that this adversely impacts the most
vulnerable people, running contrary to a rights-
based discourse. The passing over a threshold is a
crude indicator of improved well-being, given the
high existential and economic uncertainty that
confronts extremely poor people, leading to a high
possibility of recidivism back into extreme poverty.
4. See Ahmed et al. (2009), Emran et al. (2009),
Das and Misha (2010), Raza et al. (2012), Emran
et al. (2014) and Misha et al. (2014).
The International Policy Centre for Inclusive Growth | Policy in Focus 11
What does the future hold for graduation?
Harshani Dharmadasa,1 Julie Kedroske,1
Nazia Moqueet,1 Sadna Samaranayake,1
Isabel Whisson1 and Lauren Whitehead 1
Since launching the ‘Targeting the Ultra-
Poor (TUP)’ programme in Bangladesh in
2002, the success of this agship ‘graduation’
programme, and adaptations supported
by CGAP and the Ford Foundation, have
sparked a movement to apply BRAC’s
Graduation Approach to ultra-poverty
contexts around the world. To understand
where graduation is heading, and to
accelerate this momentum, it is imperative
to rst address what graduation is and is not.
Graduation is not a ‘silver bullet’ solution
to poverty. It is a programmatic approach
that links social protection, livelihoods and
nancial inclusion, and can be a strong
complement to traditional programmes
in these arenas. It is not an alternative to
national social protection oors but is,
rather, an approach that can be embedded
into them to strengthen their promotive
and transformative functions. Graduation
should not be considered a means to
wean vulnerable households o social
protection interventions. Rather, it is in
combination with these interventions that
graduation approaches can activate latent
economic potential and place households
on a sustainable pathway out of poverty
(Dharmadasa et al. 2016).
The Graduation Approach is a combination
of programming interventions including
asset transfers, consumption support,
savings, enterprise training, hands-on
coaching and mentoring and, in some
cases, health and social integration
support to ultra-poor households. While
there is a tendency to oversimplify or to
focus exclusively on its more visible tenets,
graduation is not about cows or chickens,
or other types of asset transfers, or any
one of its components in isolation, but
about recognising that a complex and
multifaceted problem such as extreme
poverty requires comprehensive solutions.
Implicit in the approach is the recognition
that the last stretch in achieving a world
free of poverty requires a tailored approach
to building resilience among the most poor
and vulnerable people. These populations
are identied by parameters including
income thresholds as well as other
dimensions of vulnerability reected in
the communities, geographies and social
strata in which they live. While denitional
discussions persist around the approach,
we are rapidly moving into a new,
more innovative round of adaptations.
Graduation is at an inection point.
Pioneering the approach
Even in its original iteration, the term
‘graduation’ is often misunderstood.
Graduation is often considered to be an
exogenous exit for participants, beyond
which they graduate out of a need
for services, or a crossing of a pre-set
income threshold or extreme poverty
itself (Devereux and Sabates-Wheeler
2015). It is none of those things. Rather,
graduation is a time-bound, sequenced
set of programmatic interventions that
are designed to boost several drivers of
resilience at the household level. Together,
these interventions achieve milestones
in social and economic advancement
(basic skills, nancial literacy, economic
self-reliance, social integration) and place
participants on an upward trajectory into
sustainable livelihoods.
In countries with strong social protection
systems, graduation interventions better
position households to avail themselves
of the promotive elements of social
protection. In the absence of deliberate
and well-executed social protection
programmes, graduation interventions
may be one of few avenues for households
to rise above the standards of a would-be
social protection oor.
As countries develop and poverty contexts
change, graduation interventions must be
exible to adapt to new conditions and
eectively address systemic contributors
to poverty entrenchment. At BRAC, this
translates into a deepened focus on
questions and intersections we seek
to further explore. These areas include
the persistent nancial exclusion of the
poorest households; food insecurity
that denies children the opportunity to
grow into healthy, active citizens; special
contexts and populations, including
those acutely aected by climate
change, conict and instability; and
youth unemployment. Simultaneously,
an imperative to scale and position the
approach for greater uptake compels us to
explore ways to reduce costs and achieve
the best possible return on investment.
Continuous adaptation
Over the 15-year history of the graduation
programme in Bangladesh, BRAC has
remained committed to its iterative
evolution. BRAC’s Bangladesh operational
sta involved in the nationally scaled-up
TUP programme, having just entered its
Photo: BRAC. Jorina at her store. Bangladesh, 2015.
12
“ In the absence
of deliberate and
well-executed social
protection programmes,
graduation interventions
may be one of few
avenues for households
to rise above the
standards of a would-be
social protection oor.
Photo: BRAC. Woman with her ducks, South Sudan, 2013.
fourth phase of implementation, reviewed
the ecacy of various support packages,
their consumption stipend, the modality
of asset transfers, the regularity of home
visits, market assessments and enterprise
selections. These internal assessments and
resulting shifts are in progress at the time
of writing this article, but a preliminary
simple takeaway is this: there is no one
approach to graduation programming,
and even the original programme—
perhaps especially so—requires innovation
and re-engineering betting the changing
needs of the poorest people.
The last decade and a half has brought
improvements to overall macro indicators
in Bangladesh. There has been a persistent
decline in the numbers of poor and
extremely poor households since 2000,
coupled with an impressive improvement
in the living conditions of poor people,
characterised by the materials used in
the construction of homes and access to
services such as sanitary latrines, electricity,
health care and immunisations. More
modest—though important—gains have
been made in food security and dietary
diversity (World Bank 2013a).
Simultaneously, however, extremely
poor households are increasingly facing
new pressures, including the realities
of population growth, migration and
urbanisation, and shocks related to
climate events. Despite signicant
improvements in access to health and
education services, persistent pockets
of extreme poverty, particularly in rural
contexts, remain (World Bank 2013b).
As BRAC continues to iterate in
Bangladesh and elsewhere, the
Graduation Approach continues to
experience and demonstrate versatility
in new environments and with new
stakeholders. In Kenya, BRAC is providing
technical assistance on the design and
implementation of a graduation pilot
funded by the International Fund for
Agricultural Development (IFAD)
through the Government of Kenya.
The goal is for the pilot to bear insights
for the government and its social transfer
programme, the Hunger Safety Net
Programme. The pilot extends graduation
programming to 2,600 women and
youth across the arid and semi-arid
lands (ASAL) of Kenya, in areas prone
to drought, and in some areas where
households are pastoralists and some
members are on the move with their
herds. As an approach that relies on
frequent interpersonal interaction
between sta and participants, Kenya’s
ASAL regions present new challenges
for graduation programming.
We attempt to address these challenges
in part through the use of technology.
Mobile phones transferred to participants
as part of their asset package provide a
way for front-line workers to check in with
participants on the move, and a way to
leverage mobile money transfers.
The International Policy Centre for Inclusive Growth | Policy in Focus 13
“ To help end
ultra-poverty,
graduation must
continue untangling
and addressing the
multifarious factors
that lead to
entrenched poverty.
Meanwhile, in the Philippines, BRAC is
collaborating with the Asian Development
Bank and the Philippines’ Department of
Social Welfare and Development (DSWD) to
provide technical assistance for the launch
of a graduation pilot. In the Philippines,
as opposed to piloting the approach
anew, a nationally scaled-up social
protection system is attempting to identify,
combine and augment elements of cash
transfer programmes, livelihood support
and community-driven development
programming into a graduation
intervention. This exercise of converging
on a graduation approach within existing
programming is perhaps indicative of
what the future could hold for graduation
programming in contexts where strong
social protection agencies exist.
Breaking down systemic
contributors to extreme poverty
To help end ultra poverty, graduation must
continue untangling and addressing the
multifarious factors that lead to entrenched
poverty. To be a viable and scalable
approach, it must systematically reduce
its drivers of cost and complexity in ways
that optimally position it for uptake among
various types of implementers, spanning
multilateral nancial institutions (MFIs),
non-governmental organisations (NGOs) and
governments. Several key areas where BRAC
endeavours to focus its eorts are as follows:
Tackling the nancial
exclusion of the poorest people
Around 2 billion of the world’s poorest
people still do not have formal nancial
accounts (Demirguc-Kunt et al. 2014).
Lacking access to basic instruments such
as savings and credit, ultra-poor people
are unable to invest in simple market
opportunities, and are frequently left
without recourse when the inevitable
crisis strikes: a sudden health shock,
economic downturn, loss of a household
breadwinner, and so forth. Through
graduation, participants gain access to
a savings pathway—whether through
community-led groups or formal nancial
institutions—nancial literacy training
and eventually the nancial capacity
(and condence) to borrow loans.
While the Graduation Approach was
initially envisioned as a way to reach
those who were too vulnerable to benet
from BRAC’s micronance programme,
BRAC has always incorporated nancial
inclusion platforms such as savings and,
increasingly, microcredit as a facet of the
approach, building savings habits as a key
outcome of the programme and providing
access to credit to participants before
concluding their cycle. In Bangladesh,
in particular, this includes BRAC’s interest-
free credit model, whereby qualifying
participants contribute to the cost of the
asset package. This opens the door for
them to qualify for products and services
in the formal nancial system by building
a credit history and learning the process
of gradual repayment.Meanwhile, digital
nancial services provide an additional
path to greater nancial inclusion for
ultra-poor people. Today, graduates of
the programme who continue to save
with BRAC can access mobile money
accounts and complimentary assistance
to start saving digitally and to make
transfers and payments.
With about 50 per cent of adults from the
poorest households having no access to
nancial services such as credit, savings
or mobile banking (ibid.), graduation
programmes can help to close the gap
in nancial inclusion by building the
capacity of the poorest people to make
use of such services.
Combating child poverty
Children are more than twice as likely to
be in extreme poverty as adults (UNICEF
and World Bank Group 2016). The enduring
eects of extreme poverty on children can
sow the seeds for a lifetime of struggle—
chronic health problems, lack of skills,
limited future productivity—which is
often transmitted to future generations.
Graduation can break that cycle. The
average TUP household in Bangladesh
has two children, who, as a result of the
programme, are much more likely to
benet from a healthier diet, from the
health, hygiene and behaviour-related
messaging and coaching, and to attend
school—a requirement for the household
to be considered ‘graduated’.
The eects of graduation on child
nutrition are signicant, as evaluation
and programme monitoring demonstrate.
The TUP programme in Bangladesh was
found to reduce the likelihood of wasting
and being underweight among children
under 52 (Raza and Van de Poel 2016).
14
Photo: BRAC. Women in BRAC's Targeting the Ultra Poor programme receive health support, often from the
organisation's own community health workers in Bangladesh, who are volunteers from the community who
have been trained to treat common illnesses and refer patients to nearby clinics.
“ BRAC continues to
explore ways in which
graduation can help
build capacity and
resilience to withstand
and recover from conict
and related instability.
Furthermore, the impacts of graduation
spill over into the local community, in terms
of increases in exclusive breastfeeding and
Vitamin A administration (ibid.).3 Graduation
contributes to improved nutrition and
caloric intake through regular messaging
and coaching on the importance of a
balanced dietary intake, food preparation
and hygiene, and by ensuring consumption
by means of time-bound stipends until
households are able to generate earnings
from enterprises to buy adequate food.
Once they do reach this level, households
establish long-term food security well
beyond the length of the programme
(J-PAL and IPA 2015). Moving forward,
BRAC plans to further explore the potential
impact of graduation programming on
children, and the implications for breaking
the cycle of intergenerational poverty.
Building solutions for populations
aected by climate change
Climate models indicate that by 2050,
Bangladesh will experience increasing
temperatures and monsoon precipitation,
intensied cyclones, more severe droughts,
riverbank erosion and rising sea levels.
The potential eects of climate change and
correlated natural disasters on ultra-poor
people are substantial, aecting their
access to fresh drinking water, natural
resources that support livelihoods, and the
ability to accumulate household savings
and partake in modest consumption and
food security (World Bank 2013b).
In response, BRAC adapted the
graduation model and implemented
the Addressing Climate Change-related
Destitution (ACCD) pilot programme
in 2012 to build the resilience of ultra-
poor households in the coastal region
of Bangladesh, with a focus on adaptive
agricultural and non-farm enterprises,
to reduce the climate change-related
vulnerabilities of these households.
Moving forward, BRAC’s graduation
programme continues to experiment
with various interventions that attempt
to build resilience against climatic
challenges, such as: home fortication,
early warning systems, interventions
addressing specic climate change-
triggered health problems, and
climate-adapted enterprises that boost
households’ resilience to shocks and
vulnerabilities resulting from resource
scarcity and environmental degradation.
Addressing fragile and
conict-aected situations
By 2030, nearly half the global share of the
world’s poor people will live in fragile or
conict-aected states (World Bank 2016).
Political conict threatens to paralyse
societies and national economies, in
turn threatening millions of households
worldwide with the prospect of falling
into extreme poverty. For households
already in dire circumstances, exposure to
a conict or post-conict environment can
be too destabilising to overcome without
signicant external support.
The International Policy Centre for Inclusive Growth | Policy in Focus 15
In South Sudan, a BRAC TUP pilot
programme found that the graduation
intervention equipped participant
households with a measure of resilience
against the ethnic conict that broke out
in December 2013. Adapted components
of graduation, such as setting up women’s
groups within which trainings and visits
were conducted, as well as engaging male
partners in trainings, appear to have led to
programme participants being able to better
cope with the shock of instability and limited
market activity than other non-participant
ultra-poor people in the community.
Women showed a 40 per cent increase in
the value of their assets and a 25 per cent
increase in spending relative to non-TUP
households, with over 70 per cent of
women maintaining at least two sources
of income (Zerihun Associates 2015).
Impacts extended to children of TUP
members too: 17 per cent of TUP children
under 5 were underweight, compared to 70
per cent of children under 5 from non-TUP
households (ibid.). Meanwhile, impacts on
the wider community were sustained by
female participants who independently
trained and inuenced community
members by sharing their knowledge,
skills and resources gained throughout
the programme, especially those related
to health and nutrition. Fifty-ve per
cent of women became peer trainers
and reported assisting at least two other
female community members each (ibid.).
BRAC continues to explore ways in which
graduation can help build capacity and
resilience to withstand and recover from
conict and related instability.
Improving youth employment prospects
In countries with a youth bulge4 coupled
with extreme poverty and high rates of
unemployment, failure to provide jobs
for youth risks fuelling social anomie,
squandering productive potential and
thus causing this population to be further
entrenched in poverty as they age. Where
jobs for youth are in limited supply, the
Graduation Approach can oer alternative
avenues through individual enterprises or
provide skills development to eectively
access labour markets.
In Uganda, which is experiencing a
youth bulge, 64 per cent of unemployed
people are under 25 years of age
(World Bank 2015). Due to limited
employment opportunities, even
individuals with basic education and skills
are vulnerable to extreme poverty. Given
this scenario, BRAC is currently shifting
its Graduation Approach accordingly
to serve 1,500 youth throughout the
Luwero district. BRAC designed the
asset package to equip youth with the
basic skills and resources to become
self-employed in livestock rearing, small-
scale trading and other microenterprises
suited to the local environment. The
programme is exploring livelihood
options in combination with mentorship
training, HIV and family planning services,
community integration and linkages
to self-employment opportunities.
Ensuring a strong
cost–benet proposition
As a tested and proven holistic
development model, the long-term
benets of the Graduation Approach
Source: Authors' elaboration.
FIGURE 1: Before and after graduation – Improved outcomes through the Graduation Approach
Producve assets
Food security
Savings
Quality healthcare
Increased skills
Social integraon
Social isolaon
Limited skills
Inadequate healthcare
Begging
Food scarcity
Limited assets
16
far outweigh the short-term costs of
implementation (Bandiera et al. 2016;
Banerjee et al. 2015).
This is especially true given that
programme costs are incurred during
a time-bound implementation period
(of 24–36 months) and produce benets
shared across the household rather than
enjoyed by single individuals. Thus, an
estimated cost of USD600 per participant
for the two-year TUP programme in
Bangladesh translates into just under
USD70 per capita annually for an
average household size of four members.
The interventions and benets provided
to each ultra-poor household are
numerous, including a vital short-term
injection of cash to boost income
generation; access to critical tools for
development; broader support services
(including linkages to government
benets and social protection
programmes); technical training in
enterprise and workforce development;
essential life skills training on issues such
as health, gender parity and condence-
building; customised troubleshooting
and guidance; and increased social
capital and peer community support.
As graduation continues to evolve,
however, BRAC and other innovating
adopters and implementers have
sought to further unpack the associated
complexity and related costs of the
approach, exploring key cost drivers and
cost recovery mechanisms to enable scale
and the greater likelihood of adaptation
by a wider net of implementers, including
governments, NGOs and MFIs.
BRAC is currently experimenting with
several variations. These adaptations
include full and partial credit models,
labour-saving strategies such as aggregating
weekly home visits to each participant
into bi-monthly home visits coupled with
monthly group visits, or substituting hired
sta for community-based peer trainers
and incentivised volunteers. The relative
performance of these models and variations
vis-à-vis commonly understood graduation
programming remain to be evaluated.
Trade-os in impact versus cost are to
be expected in the drive to settle on the
best cost for benet and variations of the
approach that can be put into practice
by governments, multilateral institutions
and other key implementers.
Conclusion
While the momentum around graduation
programming is encouraging, at BRAC
we recognise that there is much yet
to do. BRAC remains the only scaled-up
implementer of the Graduation Approach—
a reality that must change if graduation is
to play a critical role in achieving a world
free of extreme poverty. BRAC is supporting
other NGOs, governments and MFIs through
technical assistance to help spur the global
critical mass of graduation programmes
required to eradicate or signicantly
reduce extreme poverty.
While by no means the only solution
to poverty alleviation for the poorest
populations, graduation programmes
have an impressive record of impact and
sustainable gains, and provide a formula
for unlocking these outcomes in the most
vulnerable households. The evolution
of graduation and its potential success
lies in the commitment of multilateral
donors, NGOs, governments and their
social protection frameworks to invest
in these approaches, and corresponding
research agendas, in ways that strike
a balance between adaptation and
scale, technology and the human
touch, impacts and cost. We are just
getting started.
Bandiera, O., R. Burgess, D. Narayan, S. Gulesci,
I. Rasul, and S. Sulaiman. 2016. Labor markets and
poverty in village economies. London: London
School of Economics.
Banerjee, A., E. Duo, N. Goldberg, D. Karlan,
R. Osei, W. Pariente, J. Shapiro, B. Thuysbaert,
and C. Udry. 2015. “A multifaceted program
causes lasting progress for the very poor:
Evidence from six countries.” Science 348: 6236.
Demirguc-Kunt, A., L. Klapper, D. Singer,
and D. Van Oudeusden. 2014. “The Global
Findex Database 2014: Measuring Financial
Inclusion around the World.” World Bank Policy
Research Working Paper 7255. Washington,
DC: World Bank.
Devereux, S., and R. Sabates-Wheeler. 2015.
“Graduating from Social Protection? Editorial
Introduction.” IDS Bulletin 46(2). Brighton, UK:
Institute of Development Studies.
Dharmadasa, H., I. Orton, and L. Whitehead.
2016. “Mainstreaming graduation into Social
Protection oors.” IPC-IG One Pager 324. Brasília:
International Policy Centre for Inclusive Growth.
<http://www.ipc-undp.org/pub/eng/OP324_
Mainstreaming_graduation%20_into_Social_
Protection_oors.pdf>. Accessed 1 June 2017.
J-PAL and IPA. 2015. “Building Stable Livelihoods
for the Ultra Poor.” Policy Bulletin, September.
Cambridge, MA: Abdul Latif Jameel Poverty
Action Lab and Innovations for Poverty Action.
<https://www.poverty-action.org/sites/default/
les/publications/building-stable-livelihoods-
ultra-poor.pdf>. Accessed 1 June 2017.
Raza, W.A., and E. Van de Poel. 2016.
“Impact and Spillover Eects of an Asset
Transfer Programme on Malnutrition.”
Research Monograph Series No. 64. Dhaka:
BRAC Research and Evaluation Division.
UNICEF and World Bank Group. 2016. “Ending
Extreme Poverty: a Focus on Children.” New
York: UNICEF and Washington, DC: World Bank
Group. <https://www.unicef.org/publications/
les/Ending_Extreme_Poverty_A_Focus_on_
Children_Oct_2016.pdf>. Accessed 1 June 2017.
World Bank. 2013a. Bangladesh Poverty
Assessment, Assessing a Decade of Progress in
Reducing Poverty, 2000–2010. Washington, DC:
World Bank. <http://documents.worldbank.
org/curated/en/109051468203350011/
pdf/785590NWP0Bang00Box0377348
B0PUBLIC0.pdf>. Accessed 1 June 2017.
World Bank. 2013b. Turn Down the Heat:
Climate Extremes, Regional Impacts, and the
Case for Resilience. A report for the World Bank
by the Potsdam Institute for Climate Impact
Research and Climate Analytics. Washington,
DC: World Bank.
World Bank. 2015. “Empowering Uganda’s
Youth to Be Job Creators.” World Bank
website. <http://www.worldbank.org/en/
news/feature/2015/08/04/empowering-
ugandas-youth-to-be-job-creators>.
Accessed 1 June 2017.
World Bank. 2016. “Helping Countries
Navigate a Volatile Environment.” World Bank
website. <http://www.worldbank.org/en/
topic/fragilityconictviolence/overview>.
Accessed 1 June 2017.
Zerihun Associates. 2015. BRAC’s Targeting
the Ultra Poor (TUP) Project in Yei, South
Sudan. Evaluation Report. Arlington, VA:
Zerihun Associates.
1. All authors are members of the BRAC
USA Ultra-Poor Graduation Initiative.
2. Wasting for children under 5 decreased by
8 percentage points, and the likelihood of being
underweight decreased by 19 percentage
points by the end of the programme.
3. Among non-treated households,
there was a 49 per cent increase in the
duration of exclusive breastfeeding, and
a 20 percent increase in the probability
of Vitamin A administration.
4. A phenomenon whereby a large share
of the population is composed of children
and young adults, often as a result of a stage
in the country’s development where it
achieves lower mortality rates while
fertility rates are still high.
The International Policy Centre for Inclusive Growth | Policy in Focus 17
Source: CGAP (2016).
The Graduation Approach within social
protection: opportunities for going to scale1
Aude de Montesquiou 2 and Syed M. Hashemi 2
Governments, donors and non-
governmental organisations (NGOs)
attempting to reduce extreme poverty
are increasingly implementing graduation-
type interventions as part of their social
protection strategies, to create sustainable
livelihoods for many of the world’s
poorest and most vulnerable people.
The global commitment to Sustainable
Development Goal (SDG) 1: “End poverty
in all its forms everywhere” by 2030,
the rigorous evidence-based proof of
concept,3 the adoption in varied regional
contexts, the successful scale-up in many
countries, the adaptation to dierent
vulnerable groups, and the extensive
coverage in academic literature and the
popular press have all contributed to the
heightened interest of policymakers and
donors in graduation approaches. These
are growing fast, with 57 graduation
programmes now implemented in nearly
40 countries,4 of which a third are being led
by national governments. Figure 1 shows
the time-lapsed proliferation of graduation
programmes and the diversication of
dierent implementers since 2002.
CGAP (2016) provides a rich set
of information drawn from these
programmes that allows us to make
general observations on the global trends
of graduation.5 Over 2.5 million households
have been reached to date through
graduation programmes. The average size
of a programme is approximately 42,475
households, and the median size is 1,350
households, indicating that programmes
widely vary in size, ranging from a mere
150 households in Nicaragua to 675,000
households in Ethiopia. Programmes
included in the CGAP inventory are
projected to reach an additional
1.2 million households in the near future.
The Graduation Approach
Graduation programmes have been
adapted to specic objectives and
contexts. However, they share some
common characteristics (see Figure 2):
(i) they are time-bound (generally 24
to 36 months), household-level
interventions deliberately targeting
extremely poor people—either those
living under the international poverty
line of USD1.90 per day and/or those
identied as the poorest, the most
marginalised or the most vulnerable;
(ii) they are a carefully sequenced, holistic
eort, combining social assistance,
livelihoods and nancial services to tackle
the multifaceted constraints of extreme
poverty; (iii) they represent a ‘big push’
(seed capital and/or training for jobs
based on the idea that a large investment
to kick-start an economic activity is
necessary to make a meaningful change;
and (iv) they are interventions that
include some form of mentoring and
sta accompaniment to help participants
overcome not only their economic
constraints but also the many social
barriers they face, to essentially take
control of their future. Additionally, many
of these programmes facilitate access to
other social protection initiatives and to
nancial services to build resilience and
improved economic conditions beyond
the duration of the programmes.
Reviewing the expansion of graduation
Over the past 18 months, graduation
programmes have shifted their focus from
predominantly targeting rural households
(73 per cent in 2015 to 53 per cent in
2016) to mixed programmes, operating in
both rural and urban areas (7 per cent to
31 per cent), and purely urban areas
(2 per cent to 7 per cent).6 This represents
a fourfold increase in mixed and purely
urban programmes since 2015 (see Figure 3)
and reects an increased concern with
urban poverty. In fact, the extension of
graduation approaches to urban areas has
led programmes to recognise that linkages
to employment opportunities (rather than
seed capital for micro enterprises) would
be a far better option, especially for the
youth, to combat extreme poverty.
Targeting has also shifted from a
predominant focus on women and the
poorest people to a broader range of
beneciaries. Approximately a third of
projects solely targeted women in 2016,
and 60 per cent targeted populations living
on less than USD1.90 a day, down from 73
per cent of programmes in 2015 (targeting
FIGURE 1: Scaling up the Graduation Approach
BRAC Bangladesh (1) – since 2002
NGO-implemented (30) – since 2010
Donor-implemented (7) – since 2010
Government-implemented (20) – since 2010
18
“ Graduation
programmes have
shifted their focus from
predominantly targeting
rural households to
mixed programmes,
operating in both rural
and urban areas and
purely urban areas.
Source: CGAP (2016).
Source: CGAP (2016).
Source: CGAP (2016).
populations living on under USD1.25 a day—
the extreme poverty line at that time). There
is a growing eort to adapt the Graduation
Approach to other vulnerable and
marginalised segments of the population,
such as indigenous groups (31 per cent),
refugees and Internally Displaced Persons
(IDPs—9 per cent), youth (18 per cent),
people with disabilities (22 per cent) and
elderly people (9 per cent) (see Figure 4).7
Graduation programmes cost on average
USD550 per household. This includes all
costs (e.g. sta costs, programme overhead
costs, transfers to participants etc.) for the
entire duration of the programme.
While the sticker price of graduation
programmes tends to be high, it is
important to recognise that the potential
benets are also high. Randomised control
trials (RCTs) conducted by the London
School of Economics about BRAC schemes
found that the total programme costs
of USD365 would yield total benets of
USD1,168 over a projected span of 20 years
(the discounted sum of consumption and
asset gains in 2007 U.S. dollars). This would
amount to a benet–cost ratio of 3.2—or
USD3.20 in benets for every USD1 spent
on the BRAC programme (Bandiera et al.
2016). Recent research also shows that
among programmes targeting extremely
poor people (livelihood development,
lump sum cash transfers or graduation)
and for which there is long-term evidence
available, the Graduation Approach has the
greatest impact per dollar, with a positive
impact on economic indicators that
persists over time (Sulaiman et al. 2016).
Figure 5 shows that a third of ongoing
graduation projects are being implemented
by governments, reecting the growing
interest in this carefully sequenced,
multipronged approach as part of their
national social protection initiatives.
However, implementing such a holistic
approach is particularly challenging for
governments, who are the least likely
to oer the ‘full graduation package’ to
their beneciaries (compared to NGOs or
FIGURE 2: The Graduation Approach
The graduaon into sustanaible livelihoods approach
Extreme
poverty
Market/value
chain analysis
Targeng
Mentoring
Training
Access to financial services
Consumpon assistance
Seed capital/
employment
Start
Sustainable
livelihoods
Month
3
Month 6
Month 24
Month 36
FIGURE 3: Percentage by site location
Rural Urban Mixed N/A
60%
50%
40%
30%
20%
10%
0%
The International Policy Centre for Inclusive Growth | Policy in Focus 19
“ Unless there are
concomitant efforts to
expand markets through
value chain analysis,
market studies or local
economic investments,
the Graduation Approach
will be ineffective for
large numbers of the
poorest people.
Source: CGAP (2016).
donors). In fact, 36 per cent of all graduation
programmes do not oer the complete
suite of interventions. While sta costs are
part of the reason, limited governmental
capacity in stang front-line workers
with the people skills necessary to work
closely with beneciaries is an important
prohibitive reason. Governments, therefore,
often exclude case management, technical
training and the more labour-intensive
activities from the graduation package.8
Some governments are exploring ways to
make the sta accompaniment or mentoring
component of programmes easier to deliver.
In Ethiopia, for example, the government
is layering the mentoring component onto
its pre-existing social worker infrastructure.
The Government of Colombia is exploring
the use of technology with videotaped
mentoring sessions delivered via tablets or
social informational text messages sent to
participants’ mobile phones. The Peruvian
government’s Haku Wiňayprogramme is
not implementing the sta accompaniment
component at all, since it does not have
the capacity to do so eectively. Rather, it is
promoting the use of community volunteers
as a resource for technical assistance, instead
of paid programme sta.
In a number of countries, government
commitment to scaling up graduation
coincides with national initiatives to
increase the availability and use of digital
and non-digital nancial services. Digital
transfers hold strong potential to make
delivery more convenient for recipients,
while reducing the costs for project
implementers. Eighteen per cent of
projects have digitised part of the transfers,
but more research is required on digitising
components of graduation programmes
and determining which components result
in the greatest cost savings.
Challenges of scaling up
The key to successful implementation
of the Graduation Approach is contingent
on the following:
ycareful design to consider existing and
potential livelihood opportunities,
markets and prevailing cultural norms;
yparticipatory and transparent targeting9
to avoid confusion and conict as
well as to ensure that the appropriate
beneciaries are identied and included;
yclear messaging around time-bound
assistance to help catalyse and accelerate
the planning for economic livelihoods.
While safety net guarantees for those
facing crisis and slipping back is integral
to the social protection commitment
we advocate for all, the Graduation
Approach is designed as a time-bound
‘big push’ for participants to quickly
launch their livelihood activities and
stay on course towards sustainability;
yappropriate linkages to other social
protection interventions as well as
health care, schooling and nancial
services, so that participants can
continue to improve their social and
economic conditions beyond the
duration of the programme; and
yhiring sta to ensure close sta–
participant interactions and build
the agency of the poorest and
marginalised people.
The proliferation and scale-up of the
Graduation Approach is, more importantly,
conditional on easing many of the meso-
and macro-level constraints. Market size
represents an important bottleneck. Too
many people pursuing the same micro-
businesses or with the same employable
skills will soon reach the absorptive
capacity of the local economy and be out
of work. Unless there are concomitant
eorts to expand markets through value
chain analysis, market studies or local
economic investments, the Graduation
Approach will be ineective for large
numbers of the poorest people. Natural
resources, the local ecology, climate
variability and physical terrain all represent
other major constraints. Arid or semi-arid
FIGURE 4: Percentage of programmes by targeting specic characteristics
0%
10%
20%
30%
40%
50%
60%
70%
Indigenous
Refugees
Disabled
Eldery
Youth
IDPs
Women only
1.90 or less
20
“ The long-term success
of the Graduation
Approach is contingent
on a comprehensive
social protection regime
that offers a range of
risk-mitigating measures
to address the many
vulnerabilities faced by
poor people.
zones, mountainous or sparsely populated
regions, droughts or ooding all create
conditions for low economic dynamism
and low livelihood opportunities.
The absence of basic social services
such as health care facilities and schools
increases morbidity and mortality rates
and restricts the possibilities of preventing
the intergenerational transmission of
poverty. A low level of participation in local
government reduces the chances of local
budgetary expenditures for poor people.
In addition, macro-level variables, such
as economic mismanagement, corrupt
governance, violence and conict, as well
as the vagaries of the global economy
contribute to the creation of fragile States
and severely constrain possibilities for
reducing extreme poverty.
It is important to note that ‘graduation’ is
only one pathway—a rigorously tested
model—to reduce extreme poverty.
Others, such as wage employment,
may often be more eective. And
additionally, the long-term success of
the Graduation Approach is contingent
on a comprehensive social protection
regime that oers a range of risk-
mitigating measures to address the many
vulnerabilities faced by poor people.
In fact, what we advocate for are national
social protection policies with graduation
as an integral component.
The learning agenda
There is still much to learn. The graduation
‘Community of Practice’ is actively trying
to do so—for instance, 71 per cent of
graduation programme implementations
include a research component to study
their impact on beneciaries.10 Thirty
interventions include RCTs as part of
their research agendas.
Research focuses on a variety of topics:
42 per cent of research projects11 will
assess programme components, including
adaptations and method of delivery,
27 per cent of projects will assess long-
term impacts of graduation, and 11 per
cent will conduct cost–benet studies
to provide policymakers with robust
evidence to determine the eectiveness of
graduation relative to other programmes.
There is keen interest in innovations
and knowledge-sharing to: 1) adapt
the Graduation Approach to additional
vulnerable segments of the population,
including refugees, extremely poor urban
households and disadvantaged youth or
children; 2) expand the range of income-
earning options beyond rural livelihoods
to safe and decent employment; 3) provide
better linkages to meso-level interventions,
such as schooling, health care, and
climate change and disaster mitigation
programmes; 4) improve cost-eectiveness
through measures such as digitisation
of transfers and nancial services, case
management and advice through volunteer
groups or existing social support, and
leveraging existing government services;
and 5) ensuring closer convergence of
formal government social protection and
other programmes for vulnerable people.
Looking ahead
The Graduation Approach is expected to
continue to grow in scale and inuence,
with strong demand from donors and
governments for nationally scaled
programmes. CGAP is actively working
with the World Bank’s Social Protection and
Jobs Global Practice as well as members
of the global Graduation Community
of Practice to develop a dedicated,
autonomous platform for graduation
eorts, as part of the broader social
protection agenda. The platform will serve
and support governments and other
stakeholders, bringing them together so
that they can implement household-level,
holistic, income-earning interventions
for extremely poor households and other
vulnerable populations by focusing on
ve key activities:
yAdvocacy: Scaled adoption
of eective graduation-style
programmes for extremely poor
people and better integration
within social protection systems.
FIGURE 5: Programme implementer share
Government
35%
NGO
53%
Donor
12%
Government implementaons
are increasing over me,
including 70 per cent of
government implementaons
occurring since 2015
Source: CGAP (2016).
The International Policy Centre for Inclusive Growth | Policy in Focus 21
“ The Graduation
Approach is expected to
continue to grow
in scale and inuence,
with strong demand from
donors and governments
for nationally
scaled programmes.
yKnowledge generation and innovation:
A set of proven scaled models by
context and levels of resourcing/social
protection budgets; an established
process for ongoing innovation;
clarity on adaptations to additional
vulnerable segments; and proven
eciencies through digitisation
and other innovations.
yKnowledge management: A robust
global database of intervention design
and implementation guidance, technical
tools, best practices, technical service
providers and ongoing/completed
research that is accessible and used for
building graduation pathways.
yCompliance with standards:
An established and widely accepted
set of standards and methodologies
for evaluating programmes and any
new interventions.
ySustainable resourcing: A sucient
and diversied set of funders and
funding tools to support countries and
programmes to ensure that extremely
poor and vulnerable populations have
access to eective programmes and
ongoing support needed for resilience.
The platform will serve the Community
of Practice with critical public goods, test
new solutions and magnify the impact
of targeted investments in economic
inclusion by governments, development
agencies and others.
Bandiera, O. R. Burgess, N. Das, S. Gulesci, I.
Rasul, and M. Sulaiman. 2016. “Labor Markets
and Poverty in Village Economies.” Quarterly
Journal of Economics.
Banerjee, A., E. Duo, N. Goldberg, D. Karlan, R.
Osei, W. Parienté, J. Shapiro, B. Thuysbaert, and
C. Udry. 2015. “A multifaceted program causes
lasting progress for the very poor: Evidence from
six countries.”Science348(6236): 1260799.
CGAP. 2016. “State of Graduation Programs
2016.” CGAP Micronance Gateway website.
<https://www.micronancegateway.org/
library/status-graduation-programs-2016>.
Accessed 24 April 2017.
Hashemi, S., A. de Montesquiou, and K. McKee.
2016. “Graduation Pathways: Increasing Income
and Resilience for the Extreme Poor”. CGAP Brief.
Washington, DC: Consultative Group to Assist
the Poor. <http://www.cgap.org/publications/
graduation-pathways>. Accessed 3 May 2017.
Sulaiman, M., N. Goldberg, D. Karlan, and
A. de Montesquiou. 2016. “Eliminating
Extreme Poverty: Comparing the Cost-
Eectiveness of Livelihood, Cash Transfer,
and Graduation Approaches. Comparative
analysis of three strands of anti-poverty social
protection interventions.” Washington, DC:
Consultative Group to Assist the Poor and
Innovations for Poverty Action. <https://
www.micronancegateway.org/library/
eliminating-extreme-poverty-comparing-cost-
eectiveness-livelihood-cash-transfer-and>.
Accessed 24 April 2017.
Source: CGAP (2016).
1. The data reported in this article are based on
the information summarised in CGAP (2016).
2. The Consultative Group to Assist the Poor (CGAP).
3. See Banerjee et al. (2015) and Bandiera
et al. (2016).
4. Although we have identied 57 graduation
programmes globally, we received completed
questionnaires from 55 programmes. The analysis
in this article is based on these 55 programmes.
5. Five of the 55 programmes were not included
in the report due to limited information. See:
<https://www.micronancegateway.org/library/
status-graduation-programs-2016>.
6. Eighteen per cent of programmes in 2015 and
11 per cent of programmes in 2016 did not provide
complete information on regional distribution.
7. Seventeen per cent of the sample did not
report information on targeting; some reported
targeting more than one group.
8. Total supervision costs (salaries of implementing
organisation sta, training materials, training,
travel costs and other supervision expenses)
amount to 44 per cent of total programme costs,
on average, in Ethiopia, Ghana, Honduras, India
and Peru (Banerjee et al. 2015).
9. Some in the development community have
critiqued ‘targeting’ as ineective, exclusionary
and expensive, opting to support universal
coverage. However, we feel strongly that
while targeting is not 100 per cent accurate
in avoiding all inclusionary and exclusionary
errors, it is the most eective methodology
for including specic segments of the poorest
population in graduation approaches in a cost-
eective, open, participatory manner and in a
budget-constrained environment.
10. Only 2 per cent of the sample did not report
research practices.
11. Calculations are based on 36 projects that
provided specic impact assessments questions.
See CGAP (2016).
FIGURE 6: Programmes implemented with all graduation components
80%
70%
60%
50%
40%
30%
20%
10%
0%
Government NGO Donor
22
The effectiveness of the Graduation
Approach: what does the evidence tell us?
Stephen Kidd 1 and Diloá Bailey-Athias 1
No one should be against giving families
living in poverty a few goats, chickens or
cattle. Indeed, development projects have
been doing these things for decades, with
variable results. However, in recent years,
major claims have been made about the
impacts of livestock schemes known as
graduation programmes. According to the
World Bank’s Consultative Group to Assist
the Poor (CGAP 2017), the Graduation
Approach “holds signicant purpose if
implemented at scale to move people out
of extreme poverty and into sustainable
livelihoods”. Banerjee et al. (2015) argue
that it “causes lasting progress for the very
poor”, while in an article for the Guardian
newspaper, Emma Graham-Harrison
(2016) claims that it “has transformed
the lives of more than a million of the
world’s poorest families”. Indeed, the
name Graduation was chosen because
it was believed that the approach will,
in fact, ‘graduate’ people out of poverty.
A diagram produced by CGAP to explain
the programme (see Figure 1) shows how
beneciary families are expected to be on
an ever upward trajectory out of poverty.
So, are graduation schemes the ‘silver
bullet’ that the world’s poorest families
have been waiting for?
As Figure 1 indicates, the Graduation
Approach combines the provision of
assets to families—usually livestock such
as cattle, goats, chickens or, in the case of
Peru, guinea pigs—in addition to a regular
cash or food transfer for a few months
combined with intensive coaching. In some
cases, beneciaries are encouraged to save
regularly and are given access to health
services. The value of the assets provided is
not particularly large: across ve countries
Banerjee et al. (2015) found it to be the
equivalent of between 3.8 and 8 goats
(and 17.1 goats in Ethiopia). In Bangladesh,
most beneciaries appear to receive two
cows. Yet the programme is believed by
many to be life-transforming.
What is the actual evidence on the
impacts of the Graduation Approach?
Has it really achieved its stated objective
of ‘graduating’ the ‘ultra poor’ out of
poverty? This article hopes to answer
these questions by examining whether:
the beneciaries of the programme are,
in fact, the poorest people; the impacts
of the programmes are as signicant as
claimed; the impacts are sustainable;
and the approach is cost-eective.
Programme beneciaries:
are they the poorest people?
The Graduation Approach attempts to
target the poorest members of society, in
the belief that they are excluded from more
mainstream development programmes
and nancial services. However, attempts
to reach the poorest people do not appear
to be particularly successful.
As Figure 2 indicates, when measured
against a poverty line of USD1.25
(Purchasing Power Parity—PPP)—which
was used by the Millenium Development
Goals to benchmark extreme poverty—a
high proportion of recipients had higher
levels of consumption when selected for
the graduation programme.2
In Peru and Pakistan, for example,
over 80 per cent of recipients were
above the USD1.25 poverty line.
Even in Bangladesh, around 45 per cent
of beneciaries were above this line,
indicating that the vast majority would
not have been considered ‘ultra poor’
when they entered the programme (since
the ‘ultra poor’ are expected to be in the
poorest 6 per cent of their communities).
Furthermore, the Bangladesh graduation
programme is known for selecting
female-headed households, yet, in fact,
58 per cent of households were male-
headed, and only 41 per cent of
women were the sole earner in the
family (Bandiera et al. 2012; 2016).
Therefore, the claim that graduation
programmes are helping only the very
poorest is not supported by the evidence.
While some recipients are indeed living
in extreme poverty, many are not. As we
will see in the next section, this has a
major inuence on the eectiveness
of the Graduation Approach.
Source: Hashemi and de Montesquiou (2011).
FIGURE 1: The graduation model as depicted by CGAP
24 months
Poverty line
Extreme poverty
Regular
coaching
6 months3 months0 months 21 months
Targeng
Consumpon
support
Savings
Skills training
Asset transfer
Sustainable
livelihoods
Market
analysis
Access to credit
Targeng to ensure only the poorest
households are being selected.
Consumpon support to
stabilise consumpon.
Savings to build assets and insll
financial discipline.
Skills training to learn how to care for
an asset and how to run a business.
Asset transfer of an in-kind good
(such as livestock) to help jump-start
a sustainable economic acvity.
The International Policy Centre for Inclusive Growth | Policy in Focus 23
Source: Authors’ elaboration based on Banerjee et al. (2015).
FIGURE 2: Proportion of the beneciaries of graduation programmes that
were not living under USD1.25 PPP at the time of the baseline
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Proporon of beneficiaries above the USD1.25 (PPP) poverty line
India
(W.Bengal)
Ethiopia
Honduras
Bangladesh
Ghana
Pakistan
Peru
Source: Authors’ elaboration based on Balboni et al. (2015) and Bandiera et al. (2016).
Source: For Bangladesh, see Bandiera et al. (2013); for India (Andhra Pradesh), see Bauchet et al. (2015);
for all remaining impact evaluations, see Banerjee et al. (2015).
TABLE 1: Claims about the impacts of the Bangladesh graduation programme
and the evidence in absolute values (in 2007 nominal USD)
Claim Absolute values
Earnings increase by 37% USD0.06 per day
Value of household durables increases by 110% USD9.90
Value of cows increases by 208% USD20.27
Household savings increased by 863% USD13.00
Value of land owned increases by 220% USD87.64
Business assets rise by 283% USD17.36
TABLE 2: Impacts of graduation programmes on consumption,
per capita per day
USD 2014
(PPP) per day
USD 2014
(nominal) per day
Percentage of consumpon
of 30th percenle
Bangladesh 0.17 0.07 7.7%
Ethiopia 0.24 0.09 13.4%
Ghana 0.11 0.04 5%
Honduras -0.15 -0.08 -5%
India (West Bengal) 0.20 0.06 10%
India (Andhra Pradesh) -0.10 -0.03 -5%
Pakistan 0.20 0.06 6.9%
Peru 0.20 0.12 3.1%
Programme impacts:
are they signicant?
Advocates of the Graduation Approach
claim signicant impacts for the
programmes. For example, in Bangladesh
it is claimed that recipients’ earnings
increased by 37 per cent, the value of cows
increased by 208 per cent, and business
assets rose by 283 per cent. While these
gures are correct, the actual impacts
are, in reality, much less impressive when
expressed as absolute values. As Table 1
shows, earnings only increased by USD0.06
per day, the value of cows rose by a mere
USD20.27, and business assets increased
by just USD17.36. The claim by Banerjee et
al. (2015) that the Bangladesh programme
is “very eective” does not seem to be
substantiated by the evidence.
The principal aim of the Graduation
Approach, however, is to improve
household consumption (ibid.). Yet, even
with this indicator, impacts are relatively
modest or even non-existent. Table 2
shows the per capita impacts on the
expenditure of beneciary households.
Household consumption only increased—
across six countries—by between
USD0.04 and USD0.12 per capita per day.
In Honduras, beneciaries ended up poorer
than when they had started, due to their
assets (chickens) dying as a result of disease,
while similar negative impacts were also
found in Andhra Pradesh, India. A further
aspect of the scheme in Bangladesh has
been an increase in child labour among
beneciary households of around 60 hours
per year, although the exact nature of this
work is unclear (Bandiera et al. 2013).
This level of impact on consumption can in
no way ‘graduate’ families out of poverty.
As Figure 3 indicates, for rural India, the
average impact from the West Bengal
experiment would only move a family at
the 10th percentile to the 16th percentile
in the consumption distribution, so they
would still be living in extreme poverty.
In fact, buried within Banerjee et al.’s (2015)
paper is the observation that: “ the average
eects are not very large and do not
correspond to our intuitive sense of what
it would mean to be liberated from the trap
of poverty”. This contrasts markedly with
the same study’s headline nding that “the
Graduation programme’s primary goal, to
substantially increase consumption of the
very poor, is achieved by the conclusion
24
take advantage of the opportunity
presented to them. This raises the
question of whether the programme is,
in fact, appropriate for its target group—
the ‘ultra poor’—since they do not appear
to have gained much from it.
Programme benets:
are they sustainable?
Is the claim by Banerjee et al. (2015) that
graduation programmes bring about
lasting progress for very poor people
correct? First of all, as we have seen
already, those beneting most from the
programme are unlikely to be ‘very poor’.
Nonetheless, are the impacts sustainable?
The authors make this claim despite
measuring programme impacts only one
year after the programme ended, which
seems rather premature.5
There are strong indications that
household productive assets begin to
diminish among many beneciaries
either during or just after the programme
nishes. Banerjee et al. (2015) note
that beneciaries sold o some of their
productive assets soon after receiving
them, so that, one year after the
programme had been nalised, the value
of the assets held by families was less
than those they had received. In fact,
in Honduras, Pakistan and Peru, it was
less than 20 per cent of the value initially
Source: Authors’ elaboration based on data from PovcalNet 6 for Rural India in 2012.
FIGURE 3: The movement of a household in rural India up the consumption distribution,
if it receives the average impact from the graduation programme in India
0
2
4
6
8
10
12
14
16
18
20
22
24
1611 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96
Daily per capita consumpon in PPP
s
Percenle
received in assets. Across the six countries
in their study, the reduction in asset
values was greatest among the poorest
households. Similarly, in Bangladesh,
between the end of the programme
and two years afterwards, the average
number of cows owned by beneciaries
had reduced slightly from 1.22 to 1.21,
the number of poultry from 4.15 to 3.1,
and the number of goats from 0.83 to 0.57
(Bandiera et al. 2013). Also, in Bangladesh,
Misha et al. (2014) found that the
beneciaries, nine years after joining the
scheme in 2002, owned an average of only
0.72 cows/bulls and 1.95 poultry, which
does not suggest a ‘sustainable livelihood’.
The reduction in assets is not surprising.
Households continually face risks, any of
which may oblige them to sell o assets
as a coping strategy. Since the regular cash
transfer component of the schemes only
lasts from a few months to a maximum
of one year, households do not have a
guaranteed minimum level of income
security that they can draw on when hit by
a crisis. Therefore, most participants are left
exposed to risk, without any form of social
protection to support them, unless they
are lucky enough to have entered into a
government social security scheme.
In fact, in Bangladesh, there is evidence
that the change in the lives of beneciaries
of the programme” (ibid.). Therefore,
this nding does not appear to be
substantiated by the evidence either.
Furthermore, the evidence indicates that
the largest impacts are among those
who were least poor when entering the
programme.3 Banerjee et al. (2015) found
that the increase in consumption was
greater among households in the top
quantiles. As Figure 4 shows, across six
countries, the impact on consumption
at around one year following the end of
the programme was approximately four
times higher at the 90th percentile of the
consumption distribution than at the 10th
percentile. Unsurprisingly, those who
were in a stronger position nancially at
the beginning of the programme—
in particular those who were not poor—
were better able to take advantage of it.
Indeed, Misha et al. (2014) found that the
most sustainable impacts of the BRAC4
graduation scheme were among the
people who were entrepreneurs prior to
entering the programme; the impacts
on those who were not entrepreneurs
were negligible in the long term. One
explanation could be that, psychologically,
people who were not poor were better
prepared to prot from the programme,
since those living in extreme poverty—
or the ‘ultra poor’—would probably have
been less condent and less able to
The International Policy Centre for Inclusive Growth | Policy in Focus 25
was not particularly large nine years
after entering the graduation programme:
in fact, it was deteriorating year over
year. For example, around 50 per cent
of beneciaries continued to depend
on day-labouring as their main source
of income7 (Misha et al. 2014). Many
other initial benets from the programme
also diminished over this period: for
example, the number of animals owned
by beneciaries decreased consistently
between 2005 and 2011, indicating that
the initial gains from the programme
were not sustainable.8 Similarly,
while the probability of engaging in
entrepreneurship had increased by
9 percentage points by 2005, it had fallen
to 7 percentage points by 2008 and only
4 percentage points by 2011, which, as
stated by Misha et al. (2014) “renders the
long-term eect to be rather limited”.
Indeed, they conclude that: “While the
programme caused an initial shift to more
entrepreneurial employment activities,
by 2011 many treated households reverted
back to their baseline occupations.”
Of particular concern is the nding that
the women who had initially been maids
or beggars when joining the programme
in 2002 had reverted to these same
occupations by 2011, indicating no real
change among the most vulnerable
people. A similar pattern occurred among
those who had been day-labourers in 2002:
by 2011 they had reverted to being day-
labourers, and, indeed, some had become
maids or even beggars.
Similarly, Banerjee et al. (2015) report that
a range of positive impacts found at the
end of the programmes they studied had
disappeared after only one year. These
included gains in nancial inclusion, time
spent working, income and revenue,
mental and physical health, and women’s
empowerment. There is no way of knowing
whether the situation has deteriorated
further in later years, but it would not be
surprising, since families face more and
more crises over time.
The assertion that households are on a
continuously upward path out of poverty,
as indicated by the advocates of the
Graduation Approach (see Figure 1),
is unrealistic. Figure 5 shows the degree
of consumption volatility found in
Uganda over a period of two years and
the extent to which household rankings
changed signicantly. In fact, 45 per cent
of households living in poverty in 2013
had not been living in poverty in 2011,
and a similar volatility is found across
many—if not all—developing countries.
Households are frequently hit by shocks
and crises or, alternatively, are able to take
opportunities to improve their livelihoods.
Without access to regular and predictable
FIGURE 4: Combined eects of the graduation programmes on
per capita consumption across Ethiopia, Ghana, Honduras,
India (West Bengal), Pakistan and Peru
Total per capita consumpon (standardised)
10th
percenle
25th
percenle
50th
percenle
75th
percenle
90th
percenle
0.25
0.20
0.15
0.10
0.05
0.00
Source: Authors’ elaboration based on Banerjee et al. (2015).
“ The Graduation
Approach combines
the provision of assets to
families—usually livestock
such as cattle, goats,
chickens or, in the case
of Peru, guinea pigs—in
addition to a regular
cash or food transfer for
a few months combined
with intensive coaching.
26
Source: Authors’ elaboration.
social security transfers, beneciaries
of graduation programmes are just as
exposed to risk as other households,
and, over time, it is likely that the assets
of most will be signicantly eroded.
Cost-eectiveness of the
Graduation Approach
Graduation programmes are not cheap.
Costs range from USD379 per household
in India to USD2,865 in Peru, not counting
the cost of health services.9 Banerjee et
al. (2015) used a simplistic technique to
argue that the benet–cost ratio of the
programme is positive: they assumed that
the estimated consumption gains only one
year after the programme would continue
indenitely into the future, although a
discount rate was included. Thus, they
were able to report benet–cost ratios
of between 133 per cent and 433 per
cent (although it is minus 198 per cent in
Honduras). Similarly, Bandiera et al. (2016)
used the highly optimistic assumption
that the gains from the programme would
last “year-on-year”—presumably until the
death of the participants—to derive an
average benet–cost ratio of 5.4 for BRAC’s
Graduation scheme.
They also appear to use a cost of USD280
per recipient for the entire two-year
programme, whereas in an earlier paper
they put the cost of the programme at
USD300 per year, which would result in
an overall cost of USD600 (Bandiera et al.
2011). Since the USD600 cost is more in
line with other estimates, it suggests that
Bandiera et al. (2016) have signicantly
underestimated the true cost of the
TABLE 3: Value of monthly cash transfers that could be
provided to beneciary households
Country Value of monthly transfers
over 5 years (USD)
Value of monthly transfers
over 10 years (USD)
PPP value Nominal value PPP value Nominal value
Bangladesh 19 8 9 4
Ethiopia 52 20 26 10
Ghana 59 20 29 10
Honduras 39 22 20 11
India (West Bengal) 18 6 9 3
India (Andhra Pradesh) 40 12 20 6
Pakistan 78 24 39 12
Peru 74 43 37 21
“ The claims made
about the success of
Graduation programmes
are both misleading
and exaggerated, since
they give the impression
that impacts are much
greater than they
actually are.
programme and, consequently, signicantly
overestimated the benet–cost ratio.
Nevertheless, as indicated above,
the average level of productive assets
in graduation programmes declines over
time, which would most likely result in
progressively lower incomes. Indeed,
across the six countries studied by Banerjee
et al. (2015), consumption among the
poorest beneciaries was already falling
after one year. Therefore, the assumptions
of Banerjee et al. (2015) and Bandiera et al.
(2016) appear to be awed, and the nding
that these programmes are cost-eective
is highly questionable. It is certainly not
based on robust evidence.
A more interesting question would be
whether oering families a regular and
predictable transfer—and nothing else—
over a longer period but at the same cost
would result in greater impacts. Table 3
indicates the value of monthly transfers
that could be provided over ve and
10 years for the same cost as graduation
programmes across seven countries.
Compared to many social security cash
transfers in developing countries, the
values of the benet are relatively high.
Given that there is a great deal of evidence
that providing families with regular and
predictable transfers enables them to
build productive assets and engage in the
labour market, it is likely that the benets
of a long-term regular transfer would be
signicant (see DFID 2011; Kidd 2014;
Bastagli et al. 2016; Davis et al. 2016; and
Handa et al. 2016). Even among old-age
pensioners in Uganda, the proportion
purchasing livestock within a period of one
year increased to 46 per cent, compared to
26 per cent prior to the introduction of the
pension, alongside a 42 per cent increase in
the value of their purchases (Kidd 2016). Yet
many Graduation programmes do not allow
older people to participate, mistakenly
regarding them as unproductive.
Regular and predictable transfers oer
families income security, which enables
them to plan for and invest in the future.
Furthermore, if they are hit by shocks,
they have this transfer to fall back on, as
an alternative to selling their productive
assets. However, this essential safety net
oered by social security transfers is not
available to graduation beneciaries,
unless they manage to access a national
social security scheme. Indeed, the
absence of a regular and predictable social
security transfer is likely to undermine the
investment in Graduation programmes,
since the assets given to families could
be lost as a result of exposure to even
relatively small shocks.
Until a proper evaluation is undertaken
to compare graduation programmes
with regular and predictable transfers that
endure for longer periods of ve or even
10 years, we will not know the actual cost-
eectiveness of the Graduation Approach
compared to other viable options. Those
donors nancing graduation schemes
may nd that using their funds to provide
a regular and predictable transfer for up
to 10 years may have greater and more
sustainable results. There is, of course, value
in oering active labour market programmes
The International Policy Centre for Inclusive Growth | Policy in Focus 27
Photo: MGLSD Uganda. Ugandan SCG beneciary at her pig farm.
to recipients of social security schemes, but
Graduation programmes should never be
seen as a replacement for comprehensive
national social security systems. At best, they
are a complement, albeit an expensive one.
Conclusion
There is clear evidence that the Graduation
Approach does not achieve its purpose of
‘graduating people out of poverty’. It may
improve the consumption of beneciaries,
but this is hardly surprising, given that
they have recently received a range of
productive assets as gifts. Moreover, those
who are most able to take advantage
of the Graduation Approach are not the
so-called ‘ultra poor’ but, instead, those
who are better-o (a reection of the
programme’s inclusion errors).
The claims made about the success
of graduation programmes are both
misleading and exaggerated, since
they give the impression that impacts
are much greater than they actually
are. It would be much more realistic if
programme implementers were to set a
target of, for example, about 20 per cent
Source: Kidd and Gelders (2016).
FIGURE 5: Movement of households in Uganda across wealth quintiles between 2011/12 and 2013/14
Welfare quinles in 2011/12 Welfare quinles in 2013/14
Richest
Richer
Middle
Poorer
Poorest
Richest
Richer
Middle
Poorer
Poorest
of beneciaries having long-term success
due to the intervention. Indeed, that would
be a rather successful outcome, given
the low rate of success associated with
employment programmes and start-up
companies in developed countries.
The Graduation Approach is based on
the traditional belief in the international
development community of heroic
individuals dragging themselves
out of poverty by their bootstraps.
Indeed, Bandiera et al.’s (2013) claim
that Graduation programmes turn
beneciaries into ‘entrepreneurs’ is,
perhaps, a slight exaggeration, given
that they only possess a few goats,
chickens or guinea pigs.
The Graduation Approach alone does
not engage with the more fundamental
challenge of addressing social injustice
and ensuring that people living in
poverty can access State-funded social
services, including social security (such as
tax-nanced old-age pensions and child
and disability benets), although BRAC,
at least, implements other programmes
that support people to access their
rights. All NGOs delivering graduation
programmes should ensure that they also
focus on advocating for comprehensive
national social security systems, as this
is the best means of bringing about
28
fundamental and long-lasting change.
And, in the absence of such systems, NGOs
may nd that the provision of a regular
transfer for up to 10 years may serve
beneciaries much better than oering
a few animals alongside some training.
Graduation programmes considered in
isolation are certainly not social protection,
since they neither protect beneciaries
against risk—except in the short term
when people are able to sell o their
assets—nor do they provide regular
and predictable transfers (apart from
for a few initial months).
They most denitely should not be
regarded as innovative programmes,
as heralded by their advocates. Instead,
they are—as indicated earlier—just rather
expensive versions of good old-fashioned
livestock schemes, following a long tradition
of such programmes, which have had mixed
success. They are not the ‘silver bullet’ for
widespread poverty reduction, and, in
contrast to the claims of their advocates,
do not even achieve the aim of moving
most of their beneciaries into ‘sustainable
livelihoods’ or even out of poverty.
As stated at the beginning of this article,
there is no harm in giving people a few
animals, and it is even better to oer them
complementary training. Nonetheless, a
more strategic activity for those engaged
in graduation programmes would be
to advocate for the introduction of
comprehensive social security systems
accompanied by active labour market
support to the beneciaries. Only through
the redistribution of national wealth
across all citizens through tax-nanced
social security schemes—ensuring
the realisation of the basic right of
everyone to social security—can there
be fundamental social transformation.
Unfortunately, this cannot be achieved
solely by graduation programmes.
Balboni, C., O. Bandiera, R. Burgess, and
U. Kaul. 2015. “Transforming the economic
lives of the ultra-poor”. IGC Growth Brief.
London: International Growth Centre.
Bandiera, O., R. Burgess, N. Das, S. Gulesci,
I. Rasul, R. Shams, and M. Sulaiman. 2012.
“Asset Transfer Programme for the Ultra Poor:
A Randomized Control Trial Evaluation.”
CFPR Working Paper No. 22. Dhaka: BRAC
Research and Evaluation Division.
Bandiera, O., R. Burgess, N. Das, S. Gulesci,
I. Rasul, and M. Sulaiman. 2013. “Can Basic
Entrepreneurship Transform the Economic
Lives of the Poor?” CFPR Working Paper No. 23.
Dhaka: BRAC Research and Evaluation Division.
Bandiera, O., R. Burgess, N. Das, S. Gulesci,
I. Rasul, and M. Sulaiman. 2016. “Labor Markets
and Poverty in Village Economies.” Quarterly
Journal of Economics.
Banerjee, A., E. Duo, N. Goldberg, D. Karlan,
R. Osei, W. Parienté, J. Shapiro, B. Thuysbaert, and
C. Udry. 2015. “A multifaceted program causes
lasting progress for the very poor: Evidence from
six countries.”Science 348(6236): 1260799.
DOI: 10.1126/science.1260799.
Bastagli, F., J. Hagen-Zanker, L. Harman, V. Barca,
G. Sturge, T. Schmidt, and L. Pellerano. 2016.
Cash transfers: what does the evidence say?
A rigorous review of programme impact and
the role of design and implementation features.
London: Overseas Development Institute.
CGAP. 2017. “Graduation into Sustainable
Livelihoods.” Consultative Group to Assist the
Poor website. <http://www.cgap.org/topics/
graduation-sustainable-livelihoods>.
Accessed 11 April 2017.
Davis, B., S. Handa, N. Hypher, N.R. Rossi,
P. Winters, and J. Yablonski (eds). 2016. From
Evidence to Action: The Story of Cash Transfers
and Impact Evaluation in Sub-Saharan Africa.
Oxford: Food and Agriculture Organization
of the United Nations, the United Nations
Children’s Fund and Oxford University Press.
DFID. 2011. “Cash Transfers.” Evidence Paper,
Policy Division 2011. London: UK Department
for International Development.
Graham-Harrison, E. 2016. “Has a Bangladesh
charity found a way to banish extreme
poverty?” Guardian, 6 December 2016.
<https://www.theguardian.com/global-
development/2016/dec/06/has-bangladesh-
charity-found-way-to-banish-extreme-
poverty>. Accessed 11 April 2017.
Handa, S., L. Natali, D. Seidenfeld, G. Tembo,
and B. Davis. 2016. “Can Unconditional
Cash Transfers Lead to Sustainable Poverty
Reduction? Evidence from two government-
led programmes in Zambia.” Office of
Research—Innocenti Working Paper WP-2016-21.
Florence: UNICEF.
Hashemi, S.M., and A. de Montesquiou.
2011. “Reaching the Poorest: Lessons from
the Graduation Model.” CGAP Focus Note,
No. 69. Washington, DC: Consultative
Group to Assist the Poor.
Kidd, S. 2014. Social protection: an eective and
sustainable investment in developing countries.
Frankfurt: KfW Development Bank,.
Kidd, S. 2016. Uganda’s Senior Citizens’ Grant:
A success story from the heart of Africa.
Kampala: Ministry of Gender, Labour
and Social Development, Expanding
Social Protection programme.
Kidd, S., and B. Gelders. 2016. Inclusive
lifecycle social security: an option for Uganda?
Kampala: Ministry of Gender, Labour and
Social Development, Expanding Social
Protection programme.
Misha, F.A., W. Raza, J. Ara, and E. Van de Poel.
2014. “How far does a big push really push?
Mitigating ultra-poverty in Bangladesh.”
ISS Working Paper, No. 549. Rotterdam:
International Institute of Social Studies.
1. Development Pathways.
2. We use the USD1.25 PPP poverty line rather
than the current USD1.90 PPP poverty line of
the Sustainable Development Goals, as this is the
line that was used by the evaluators themselves.
Nonetheless, given that the Graduation
programme is aimed at the ‘ultra poor’ rather
than ‘poor’ people, it seems appropriate to
use the USD1.25 PPP poverty line.
3. Misha et al. (2014) came to the same
conclusion among beneciaries of the
Bangladeshi Graduation programme.
4. BRAC is a non-governmental
organisation (NGO) based in Bangladesh.
See <http://www.brac.net>.
5. Unfortunately, it is not possible to use the
evaluation results from the study by Bandiera
et al. (2016) of the eects the Bangladesh
Graduation programme seven years after
programme commencement, since there
was no longer a control group, and impacts
could not be measured.
6. PovcalNet is an online analysis tool for global
poverty monitoring. See <http://iresearch.
worldbank.org/PovcalNet/home.aspx>.
7. The actual reduction in the proportion of
those depending on day labour was, in fact,
modest, falling from 60 per cent to 50 per cent
after nine years (Misha et al. 2014).
8. For example, while beneciaries of the BRAC
Graduation programme had seen an increase in
the number of cows/bulls owned of 1.5 by 2005,
by 2011 that number had fallen signicantly
to a net gain of only 0.4 cows/bulls.
9. These gures are in nominal USD (2014
values). In USD PPP terms, costs range from
USD1,257 per household in India to USD5,150
in Pakistan. It is also likely that the administrative
costs of implementing NGOs and donors are
not included in these gures.
“ The assertion that
households are on a
continuously upward
path out of poverty,
as indicated by
the advocates of the
Graduation Approach
is unrealistic.
The International Policy Centre for Inclusive Growth | Policy in Focus 29
Source: Authors’ elaboration.
The labour markets of the ultra poor1
Clare Balboni,2 Oriana Bandiera,3
Robin Burgess 3 and Imran Rasul 4
Why labour markets?
One in 10 people live in extreme poverty
today (World Bank 2016). Bringing this
number to zero by 2030 is the central
tenet of the Sustainable Development
Goals agreed in 2015. This will involve
moving 700 million people out of
extreme poverty. Can it be done?
The answer, we argue, requires
understanding poor people’s labour
markets. Labour is all that poor people
have, and employing it in a productive way
is key to enabling them to exit poverty.
In simple terms, poor people need to
work more hours and/or earn more per
hour to lift themselves out of poverty.
Here we report on results from the
evaluation of a programme designed to
achieve this goal. The programme is BRAC’s
Targeting the Ultra Poor (TUP), also known
as ‘Graduation’. This consists of a transfer
of livestock assets and skills training to
the poorest women in the poorest villages
of Bangladesh. The programme targets
women, as they are the most vulnerable
among ultra-poor people. Previous
approaches to poverty reduction typically
involved providing access to capital (e.g.
assets, nance) or human capital (e.g. skills,
education). The programme we study is
innovative in that it combines both these
approaches with a view to making labour
supplied by poor people more productive.
To evaluate this programme, we ran a large-
scale and long-term randomised control
trial in Bangladesh, covering over 21,000
households in 1,309 villages surveyed
four times (in 2007, 2009, 2011 and 2014).
The evaluation design allows us to look
at eects for a full seven years after the
programme was introduced. This helps us
to see beyond short-term eects and to
assess whether the programme is capable
of achieving its aim of lifting people out of
poverty or only provides a short-term boost.
Our baseline survey revealed a very strong
correlation between poverty and labour
market choices. We found that the poorest
women engage in low-paid and seasonal
casual wage labour, while wealthier
women engage in livestock rearing.5
The programme enables poor women
to start engaging in livestock rearing,
increasing their aggregate labour supply
and earnings. Further, we show that this
leads to the accumulation of livestock, land
and business assets and poverty reduction,
both sustained after four and seven years.
Our results show that poor people are able
to take on the occupations of the wealthier
but face barriers to doing so, and that one-
o interventions that remove these barriers
lead to sustainable poverty reduction.
Who are the ultra-poor people?
To identify who is eligible for the
programme, BRAC asks the village
residents to classify all households into
four or ve wealth groups, broadly
corresponding to three ‘social classes’—
namely, the poor class, the middle class
and the upper class. BRAC ocers then
visit households in the two lowest ranked
groups—which correspond to the poor
group—and determine eligibility for the
programme based on a set of criteria
meant to identify the most vulnerable poor
people and those of them who are able to
participate in the programme.6 The eligible
poor group is called ‘ultra poor’, and the
non-eligible poor, ‘near poor’.
Baseline data show that the able-bodied
women who are identied as being ultra-
poor are disadvantaged across a whole
range of dimensions relative to other
women in these villages. Not only are
more of them living beneath the global
poverty line, they are also almost entirely
illiterate, and a much larger proportion
of them are the sole earner in their
household. These women are also largely
landless and lack livestock and business
assets. Indeed, this is the key feature that
dierentiates them from those women
who are better o in the village.
The randomisation of the villages into
treated and control villages, and the
detailed data collection undertaken
in both types of villages, allowed for a
thorough description of the nature of
poverty before the implementation of the
programme and a thorough analysis of
how the programme has aected living
standards and vulnerability across a range
of dimensions beyond monetary poverty.
How do they employ their labour?
Figure 1 shows that, across all social classes,
women devote 80 per cent of their labour
hours to just three labour market activities—
maid work, agricultural labour and livestock
FIGURE 1: Occupation by wealth class at baseline
1.2
0.8
0.6
0.4
0.2
0
1
Ultra poor Near poor Midle class Upper class
Other
Livestock
Maid
Ag labour
30
Source: Authors’ elaboration.
Source: Authors’ elaboration.
rearing. The choice of where women can
employ their labour is, therefore, extremely
limited. What is even more striking in Figure 1
is that poor people are predominantly
engaged in low-paid and seasonal casual
wage labour, whereas middle- and upper-
class women are engaged in higher-paid
and less itinerant livestock rearing. Poverty
is thus associated with those labour market
activities that require unskilled labour and
no inputs of capital. 7
Because of seasonal demand for their
labour, ultra-poor women work for many
fewer days per year than their wealthier
counterparts—two months less over the
course of a year. This is consistent with
evidence from other parts of the developing
world that rural landless poor people are
often underemployed. Indeed, across the
globe, the people who remain in extreme
poverty are often characterised by having
an almost complete dependence on low-
paid and insecure activities where they
are on daily wage contracts in unskilled
occupations. The baseline data we gathered,
therefore, make it very clear that any serious
attempt to lift these women out of poverty
will have to address the labour market
activities in which they are engaged.
Would beneciaries be able
to take on rich people’s
occupations and escape poverty?
BRAC’s TUP programme gives participating
women a menu of income-generating
assets, all valued at USD560 in purchasing
power parity (PPP) terms, from which all
participants select a bundle of livestock
assets. The assets are accompanied by a
support package of similar value to train
and assist recipients in working with
livestock over two years, including via
weekly home visits.
The fundamental question this study is
asking is whether poor people can take
on the labour market activities of wealthier
women in the same villages. The answer
to this question depends on whether ultra-
poor women can successfully employ the
assets they receive as part of the programme
and combine this with the skills training
and support they receive to successfully
operate small livestock businesses. The
history of asset transfer programmes in the
developing world is a chequered one, with
many cases of poor people being unable
to operate businesses that are typically
run by wealthier households. The illiterate,
asset-less group that the programme
targets may be expected to be the least
well positioned in these communities to
become successful entrepreneurs.
Yes they can
Figures 2 and 3 answer the question of
whether the poorest women in the treated
villages can take on the occupations of their
wealthier counterparts and escape poverty.
The short answer is that they can. Four years
after the asset transfer, they work 17 per
cent more hours and 22 per cent more days
per year, as shown in Figure 2. This Figure
also shows that this change is driven by
a 217 per cent increase in hours devoted
to livestock rearing, while hours devoted to
agricultural labour and maid work both fall.
As shown in Figure 3, this shift in working
hours from casual wage labour towards
livestock rearing means that the earnings
of women in the programme after four
years are 21 per cent higher, their per
capita expenditure is 11 per cent higher,
and the value of durable goods they own
is 57 per cent higher. There is, therefore,
clear evidence that ultra-poor women in
the villages which received the programme
FIGURE 3: Four-year eect on earnings and expenditure (percentage increase) 8
Earnings
21
PCE
11
Durables
57
FIGURE 2: Four-year eect on hours and days worked (percentage increase) 8
217
-17
17
22
-26
Livestock
hoursTotal hoursTotal days
Ag labour hours Maid hours
The International Policy Centre for Inclusive Growth | Policy in Focus 31
Source: Authors’ elaboration.
Note: Household assets include jewellery, sarees, radio, television, mobile phones, furniture etc. Productive assets
include livestock, land, agricultural equipment and other machinery used for production.
Source: Authors’ elaboration.
experienced a very signicant improvement
in living standards. This is an important
nding, in particular because the results
presented in Figure 3 are four years after
the baseline, and two years after the
programme ended, demonstrating that
welfare gains were maintained even
without further assistance. This is supported
by evidence from a survey conducted in
2014—seven years after the programme
began—that seven-year impacts were at
least as large as the four-year impacts.9
In Figure 4, we examine nancial and
productive assets, which are what
dierentiate the poorest women from
richer women at baseline. Here we see
very large eects. Moving from a very low
base, ultra-poor women in these villages
experience a 400 per cent increase in
savings, a 200 per cent increase in the
value of livestock held, a 159 per cent
increase in the value of other productive
assets and an 82 per cent increase in the
value of land owned after four years.
These eects demonstrate that women
are moving from a position of being largely
asset-less, with limited participation in
nancial markets, to one of greater asset
ownership and nancial inclusion.
Our study design allows us to test whether
these gains come at the expense of
other households in villages where the
programme operates and that do not
receive the programme. We tracked 15,100
households from wealth classes other than
the ultra poor (near poor, middle class
and upper class), and found no evidence
that this is the case: the livestock-rearing
businesses of richer women are not
crowded out, and the agricultural and maid
wages they receive increase, particularly for
those who are near poor, as participating
women withdraw their labour.
Are these big eects?
There is no doubt that these eects are
very large relative to baseline values and
relative to ultra-poor women in control
villages. But these comparison benchmark
values are very small, so even small
changes for the treated group lead to
large percentage dierences. Given that
the absolute eect is not informative,
we need to look for alternative benchmarks
to gauge how large these eects are.
One such benchmark is given by similar
programmes in other settings. The eects
we estimate are larger than those achieved
by pilots of similar programmes across six
dierent countries and similar to those
achieved in neighbouring West Bengal
(Banerjee et al. 2015).
Our sample design also allows us to
compare the size of the impact with the
gap between the ‘ultra-poor’ households
and those in the next wealth class, the ‘near
poor’—who, as described above, were
not eligible for the programme because
they had productive assets, an able
working male or were already receiving
assistance from the government. Figure 5
compares the ultra poor to the near poor
in these communities. In terms of per
capita expenditure (PCE) and household
assets, ultra-poor people close the gap
between them and those who are near
poor. Even more surprising is the fact that
in terms of savings and productive assets,
ultra-poor people actually overtake the
near-poor group. These ndings point to
the programme’s transformative eect, not
only on the welfare of the poorest women
in these villages as measured by earnings
and per capita expenditure, but also in
terms of their asset holdings. The nding
that women are acquiring more productive
FIGURE 4: Four-year eect on asset value (percentage increase) 8
Savings
400
Livestock
200
Land
82
Other producve assets
159
FIGURE 5: Treatment eect as a share of the gap 8
PCE
1
HH assets
0.98
Producve assets
2.1
Savings
3.8
32
assets and saving more is important
because these provide a future income
stream, thus cementing a better economic
future for them within the village.
Another useful benchmark is the cost
of the programme, which leads to the
next question.
Is it worth it?
The ultra-poor programme that we studied
is expensive. The combined value of the
assets, training and support provided over
two years is USD1,120 per household in PPP
terms. Therefore, this is really a ‘big push’ but
a time-limited programme. As we know the
returns to the programme, we can carry out
a cost–benet analysis. We estimate that
the programme’s benet–cost ratio is 3.2
if the estimated consumption benets in
the fourth year are repeated over 20 years.
In short, based on this assumption, the
benets from the programme far exceed
the costs. We also estimate the internal rate
of return of the programme to be between
16 per cent and 22 per cent, which is very
high relative to existing market rates. This
demonstrates that it would have been
worthwhile for households to have invested
in these types of activities if they could
have aorded to do so. The highly positive
benet–cost ratios and internal rates of
return underline the fact that this is a
highly cost-eective programme.
Things we know
The ndings of this study demonstrate
that the programme gives previously
underemployed women a labour market
activity across the whole year. This
allows them to dramatically expand
the hours they work by putting many
additional hours into raising livestock.
More importantly, rather than being
entirely consumed, the extra earnings
are used to purchase more productive
assets, including land. This bodes well for
poverty reduction in the long term. As the
new assets start producing income, the
increase in per capita expenditure, and the
corresponding reduction in the share of
beneciaries below the poverty line, will be
larger than the gures estimated here.
Thus, to have a permanent eect on the
living standards of very poor people,
interventions need to change the labour
market activities in which they can engage.
Benets and transfers that do not change
poor people’s labour market activities
are essential as insurance policies but are
not going to allow them to permanently
escape extreme-poverty.
The fact that the living standards of ultra-
poor peole continue to improve and that
they continue to accumulate assets long after
the end of the two-year programme makes it
very clear that as a result of the programme
they are in a very dierent position from
ultra-poor women in the control villages.
Although the programme is very expensive,
these costs are time-limited, whereas the
benets continue to accrue over time.
Things we do not know
The study helps to open up a whole new
area of research focused on combining
capital and human capital in large amounts
in one-o, ‘big push’ interventions to
counter extreme poverty. However, it
leaves several questions unanswered. One
key question relates to the heterogeneity
of eects: while no households are harmed
by the programme, some benet much
more than others. Understanding why
some households are able to generate very
large returns from the assets and skills that
are transferred to them while others are
not is a key area for further research.
Another key question is why, if the internal
rate of return from acquiring livestock and
skills is so much higher than going market
rates, poor people do not invest in these
activities themselves. This raises fundamental
questions about the ultimate causes of
poverty traps in these contexts, a topic on
which we hope to conduct future research.
Finally, this study only looks at the transfer
of assets and skills designed to enable poor
people to set up small livestock businesses.
It is possible that a better approach might
be to give them cash and let them choose
how to invest this money. We are engaged
in another randomised evaluation in the
Punjab in Pakistan which compares the
eectiveness of cash and asset transfers,
and we look forward to reporting on
those results in the near future.
Bandiera et al. 2017 (forthcoming). “Labor
Markets and Poverty in Village Economies.”
Quarterly Journal of Economics.
Banerjee, Abhijit, Esther Duo, Nathanael
Goldberg, Dean Karlan, Robert Osei, William
Parienté, Jeremy Shapiro, Bram Thuysbaert, and
Christopher Udry. 2015. “A multifaceted program
causes lasting progress for the very poor: Evidence
from six countries.” Science 348(6236: 1260799.
World Bank. 2016. Poverty and Shared Prosperity
2016: Taking on Inequality. Washington, DC:
World Bank. <http://www.worldbank.org/en/
publication/poverty-and-shared-prosperity>.
Accessed 11 May 2017.
1. This article draws on Bandiera et al. (2017).
2. London School of Economics (LSE).
3. LSE and International Growth Centre (IGC).
4. University College London (UCL) and IGC.
5. We focus on the labour markets of women
because the programme targets them.
Poverty and labour market choices are also
strongly correlated for men, with the poorest
men engaging in casual wage work, and the
wealthiest in land cultivation.
6. Three out of ve of the following inclusion
criteria have to be met: household owns less
than 10 decimals of land; adult women in the
household work outside the homestead; no
active male adult (female household head);
school-aged children working for pay; no
productive or income-generating assets.
Households are excluded if they satisfy any
of the following exclusion criteria: no active
female member in the household; micronance
participants; household members receiving
government benets such as old-age pensions.
7. The gure reports the share of total work hours
devoted to each of the three occupations and to
a residual ‘other ’ category at baseline by all the
sample women divided into four wealth classes.
The ‘other ’ category is distributed across several
other activities, which typically account for less
than 1 per cent of hours each (where work on
the household’s own land is counted as own
cultivation, not agricultural labour). The activities
that account for more than 1 per cent for ultra-
poor people are begging, tailoring, casual day
labour outside agriculture and land cultivation.
For the near poor, they are begging, tailoring,
casual day labour outside agriculture and land
cultivation. For the middle classes they are tailoring
and land cultivation. For the upper classes they are
tailoring, teaching and land cultivation.
8. The gure reports the coecient of the
interaction between a treatment status dummy
and a dummy for four years after baseline (two
years after the end of the programme: 2011)
from a specication that includes treatment,
survey dummies and sub-district xed eects.
The sample is all ultra-poor women in treatment
and control villages.
9. By 2014, every control BRAC branch oce had
treated some villages within its radius, and 20
per cent of the originally selected beneciaries
were treated overall. The challenge in identifying
the eect of the programme in 2014 is that the
selection of the late treated is correlated with
the outcome of interest: poverty. In the paper
(Bandiera et al. 2017) we perform a bounding
exercise that exploits our quantile treatment
eect estimates on the original treated to
create counterfactuals of the eect of the
programme on the late treated, and we nd
that the eects are stable.
The International Policy Centre for Inclusive Growth | Policy in Focus 33
Can graduation approaches contribute
to building social protection oors?1
Christina Behrendt 2
Graduation approaches3 have attracted
intense interest for helping poor
and vulnerable households develop
sustainable livelihoods (e.g. Hashemi and
de Montesquiou 2016). As both graduation
approaches and social protection oors
aim to lift people permanently out of
poverty, it is not surprising that observers
have raised the questions of whether
and how graduation approaches can
contribute to build social protection oors
and realise the right to social protection
for people living in extreme poverty
(e.g. Dharmadasa et al. 2016).
Social protection oors play a key
role in eradicating poverty, reducing
vulnerabilities and promoting social
inclusion, as highlighted in the 2030
Agenda for Sustainable Development—
namely, Target 1.3 of Sustainable
Development Goal 1 (“end poverty in all its
forms everywhere”): “implement nationally
appropriate social protection systems and
measures for all, including oors, and by
2030 achieve substantial coverage of the
poor and the vulnerable” (United Nations
2016a) and the ILO Social Protection Floor
Recommendation No. 202, which was
adopted in 2012 by the ILO’s 187 Member
States (ILO 2012a; 2012b; Kaltenborn 2015).
Social protection oors can be considered
as the ‘ground oor’ of a national social
protection system, which should guarantee
at least a basic level of social security to all
with a view to realising the human right to
social security (ILO 2014; United Nations
2016b). They should guarantee at least
access to essential health care and basic
income security for all members of society
throughout the course of their lives.
Guided by the ILO Recommendation, it is
the responsibility of each country to dene
their social protection oors in accordance
with their national circumstances,
ensuring that the dierent elements of
their national social protection systems
guarantee at least a basic level of social
security for all. The Recommendation also
highlights the importance of coordinating
social protection policies with labour
market, employment and other policies,
and emphasises the need to support in
particular disadvantaged groups and
people with special needs, including
those in the informal economy.
These features—a focus on reducing
poverty and vulnerability, outcome
orientation and linkages to employment
promotion—highlight the potential
contribution of graduation approaches to
nationally dened social protection oors.
Yet there are many questions
surrounding the theory and practice
of graduation approaches which make
their contribution to social protection
oors less straightforward than it
may seem at rst glance. This article
focuses on three core questions, and
sketches out some considerations for
policymakers and practitioners, based on
ILO’s Recommendation No. 202, on how
graduation approaches could be upgraded
to better and more broadly contribute to
nationally dened social protection oors.
Graduation from what?
The rst and most fundamental of these
questions is: ‘Graduation from what?’
While the stated objective of graduation
approaches is ‘an exit from poverty’
(which is fully in line with the social
protection oor concept), they are often
operationalised as ‘an exit from a particular
social protection programme’. Despite
eorts to problematise this operational
denition issue, ‘graduation from social
protection’ continues to be a commonly
used term (e.g. in Devereux and Sabates-
Wheeler 2015). This notion seems to
suggest that (a) social protection is only
for those living in poverty, and (b) social
protection and employment/income
generation are not compatible. Both
assumptions are problematic (Kidd 2013).
While recognising the importance of
employment, income generation and
labour market integration for people of
working age, Recommendation No. 202
emphasises the principle of universality
of protection and sets out that the social
protection oor guarantees should cover all
members of society throughout the course
of their lives. This implies that everyone
should be protected throughout their life
course—it does not necessarily mean that
everyone should receive a benet at every
point of their lives. In practice, this can imply
that people who move out of a means-
tested programme because they no longer
full the eligibility criteria are still eectively
protected through other elements of the
social protection system (should they
come to need it), including in case of ill
health (health insurance or public health
service), disability, maternity etc., to prevent
poverty. It also implies that they can reapply
for means-tested benets at any time in
case of need. The conditions under which
benets are provided should be set out in
national legislation, ensuring that benets
are adequate and predictable, respecting
people’s rights and dignity and including
adequate complaint and appeal procedures.
Such continued universal protection
through the social protection system (not
necessarily through the same programme)
is key to enabling people to engage in
productive and decent employment
(including entrepreneurial risk-taking)
in a sustained way, to address poverty
dynamics and prevent people from being
pushed back into poverty due to a lack of
social protection (Chronic Poverty Advisory
Network 2014; ILO 2014). The combination
of non-contributory (tax-nanced)
and contributory (social insurance)
mechanisms can play an important role
in ensuring universal coverage—including
poor people—through equitable and
sustainable nancing mechanisms.
Some graduation programmes include
elements of continued social protection
coverage (e.g. health insurance—see de
Montesquiou et al. 2014), yet in many cases
such coverage is limited or incomplete.
If graduation approaches are to contribute
to a social protection oor, more systematic
eorts would be needed to shift from
the notion of ‘graduation from social
protection’ to ‘an exit from poverty through
34
continued social protection’.4 This would
imply that people receive cash transfers
and other benets for as long as necessary,
based on clear eligibility conditions, and
continue to be protected by appropriate
social protection mechanisms (including
eective access to health care, income
security in case of maternity disability,
unemployment or loss of livelihood, as well
as child benets and old-age pensions)
after having started employment or
other income-generating activities.
This continued social protection would not
only contribute to realising their right to
social security but also protect their fragile
livelihoods and enable them to follow a
more solid upward trajectory out of poverty.
Graduation for whom and at what cost?
The second question is: ‘Graduation for
whom and at what cost?’ While the basic
premise of graduation approaches—of
helping poor people overcome the
multiple obstacles to exiting poverty—
is undisputed, the question of how
resources are allocated has been subject
to extensive debate, especially regarding
targeting mechanisms. Many graduation
programmes use targeting mechanisms
that are not suciently eective,
transparent or equitable, often leading
to both exclusion and inclusion errors
(Mkandawire 2005; Kidd 2013; 2015; Brown
et al. 2016).5 A package of productive
assets, cash transfers and other types of
support is provided—often at a signicant
cost—to a small group of programme
participants, while other people in similar
situations are excluded from accessing
any (or most) of the benets and services
provided (but may have to serve as a
control group for programme evaluations).
Within communities, this may lead to
frustration, tensions and stigmatisation.
The provision of expensive graduation
packages to a subset of poor people (often
not the poorest) may be justied where
the programme is run and nanced by a
non-governmental organisation, often as a
response to a lack of eective and inclusive
public services. However, governments
have dierent obligations and are
accountable to their entire population
(United Nations 2012). As these obligations
have important implications for the
provision of public services, governments
may conclude that, instead of allocating
a cost-intensive graduation package to
a very limited group of people, it may be
more eective, ecient and equitable to
provide a dierentiated set of benets and
services to a larger share of the population
in a way that ensures the inclusion of the
most vulnerable. Such a universal and
inclusive approach can exploit economies
of scale in the provision of such benets
and services, and make use of innovative
approaches to facilitate access to public
services—particularly in rural areas, such as
one-stop shops (‘single window services’),
integrated and individualised support
(as in the Chile Solidario programme and
its successor, Ingreso Ético Familiar) and
mobile communication services. Such an
approach also helps to redress inequalities,
promote social inclusion and enhance trust
in eective and ecient public institutions.
Graduation to where?
The third question is: ‘Graduation to
where?’, and focuses on the capacity of
graduation approaches to foster true
transformation in people’s lives. Many
observers have pointed to the limited
capacity of graduation programmes to
ensure the quality and sustainability of
employment and income generation,
and to their relative blindness towards
demand-side barriers (McCord and Slater
2015; Daidone et al. 2015). In addition,
despite some positive eects, many
graduation programmes do not pay
sucient attention to children’s current
and future needs with regard to access to
nutrition, education, care and any other
goods and services (Roelen 2015). As a
result, many ‘graduations’ are relatively
short-term, and fail to break the vicious
cycle of poverty in a sustained way.
While graduation approaches focus on
endowing poor people with the capacity
to improve their lives through micro-level
and supply-side interventions, they are not
(or not suciently) equipped to address the
structural issues that perpetuate poverty
and deprivation, and that prevent people
from exiting poverty in a sustained manner.
A more comprehensive approach
combines micro-level and supply-side
interventions with macro-level and
demand-side interventions to address
the full range of barriers to employment
and income security, and to achieve
a real transformation of poor people’s
livelihoods. Recommendation No. 202
sets out an ambitious agenda for such a
policy approach: the progressive extension
of social protection systems—including
oors—should be coordinated with
other policies “that enhance formal
employment, income generation,
education, literacy, vocational training,
skills and employability, that reduce
precariousness, and that promote secure
work, entrepreneurship and sustainable
enterprises within a decent work
framework” (ILO 2012a). This reects
a broad commitment to tackle the
wide range of issues that trap people
in precarious or informal employment
conditions and hold back sustainable
development, which has been further
elaborated in Recommendation No. 204,
concerning the transition from the
informal to the formal economy (ILO 2015).
Social protection oors, embedded in
such a comprehensive policy framework,
contribute to lift people out of poverty in
a sustainable way. Graduation approaches
can play a role in this by shifting their
focus from ‘graduation out of a social
protection programme’ to ‘graduation into
a social protection system’. This could be
done, as previously mentioned, through a
concerted eort to directly link participants
with existing mechanisms (such as health
protection, cash transfers etc.) or, where
these do not exist, support for building up
eective and inclusive mechanisms that
can ensure continued social protection
for all—in other words, universal social
protection. This is indispensable for the
reduction and prevention of poverty, the
promotion of productive employment and
decent work, human rights and inclusive
growth and development, and, ultimately,
for realising the commitments of the 2030
Agenda for Sustainable Development.
BRAC. 2016. The PROPEL Toolkit: An
Implementation Guide to the Ultra-Poor
Graduation Approach. Dhaka: BRAC.
<https://issuu.com/bracultrapoor/docs/brac_
propel_toolkit>. Accessed 25 April 2017.
Brown, C., M. Ravallion, and D. Van De Walle.
2016. A Poor Means Test? Econometric Targeting
in Africa. Washington, DC: World Bank.
<http://documents.worldbank.org/curated/
en/484991481639919564/pdf/WPS7915.pdf>.
Accessed 25 April 2017.
Chronic Poverty Advisory Network. 2014.
The Chronic Poverty Report 2014–2015: The
road to zero extreme poverty. London: Overseas
Development Institute. <https://www.odi.org/
The International Policy Centre for Inclusive Growth | Policy in Focus 35
Photo: Conor Ashleigh/AusAID. Children take part in an exercise in groups inside a BRAC primary school in
Manikganj, Bangladesh, 2012 <https://goo.gl/cm0h76>.
sites/odi.org.uk/les/odi-assets/publications-
opinion-les/8834.pdf>. Accessed 25 April 2017.
Daidone, S., L. Pellerano, S. Handa, and B. Davis.
2015. “Is Graduation from Social Safety Nets
Possible? Evidence from Sub-Saharan Africa.” IDS
Bulletin 46: 93–102. <https://opendocs.ids.ac.uk/
opendocs/bitstream/handle/123456789/7288/
IDSB_46_2_10.1111-1759-5436.12132.
pdf?sequence=1>. Accessed 25 April 2017.
Devereux, S., and R. Sabates-Wheeler. 2015.
“Graduating from Social Protection? Editorial
Introduction.” IDS Bulletin 46: 1–12. <https://opendo
cs.ids.ac.uk/opendocs/bitstream/handle/1234567
89/7273/IDSB_46_2_10.1111-1759-5436.12124.
pdf?sequence=1>. Accessed 25 April 2017.
Dharmadasa, H., I. Orton, and I. Whitehead.
2016. Mainstreaming graduation into social
protection oors. Brasìlia: International Policy
Centre for Inclusive Growth. <http://www.
ipc-undp.org/pub/eng/OP324_Mainstreaming_
graduation%20_into_Social_Protection_oors.
pdf>. Accessed 25 April 2017.
Hashemi, S., and A. de Montesquiou. 2016.
Graduation pathways: Increasing income and
resilience for the extreme poor. Washington, DC:
Consultative Group to Assist the Poor. <http://
www.cgap.org/sites/default/les/Brief-Graduation-
Pathways-Dec-2016.pdf>. Accessed 25 April 2017.
ILO. 2012a. Social Protection Floors Recommendation,
2012 (No. 202). Geneva: International Labour Oce.
<http://www.ilo.org/dyn/normlex/en/f?p=NOR
MLEXPUB:12100:0::NO::P12100_INSTRUMENT_
ID:3065524>. Accessed 25 April 2017.
ILO. 2012b. Social Security for All: Building social
protection oors and comprehensive social security
systems. The strategy of the International Labour
Organization. Geneva: International Labour Oce.
<http://www.socialsecurityextension.org/gimi/
gess/RessFileDownload.do?ressourceId=34188>.
Accessed 25 April 2017.
ILO. 2014. World Social Protection Report 2014/15:
Building economic recovery, inclusive development
and social justice. Geneva: International Labour
Oce. <http://www.ilo.org/wcmsp5/groups/
public/---dgreports/---dcomm/documents/publica
tion/wcms_245201.pdf>. Accessed 25 April 2017.
ILO. 2015. Recommendation concerning the
Transition from the Informal to the Formal Economy,
2015 (No. 204). Geneva: International Labour Oce.
<http://www.ilo.org/dyn/normlex/en/f?p=NOR
MLEXPUB:12100:0::NO::P12100_INSTRUMENT_
ID:3243110>. Accessed 25 April 2017.
Kaltenborn, M. 2015. “Global Social Protection.
New impetus from the 2030 Agenda for
Sustainable Development.” Global Governance
Spotlight. <http://www.sef-bonn.org/leadmin/
Die_SEF/Publikationen/GG-Spotlight/ggs_2015-
07_en.pdf>. Accessed 25 April 2017.
Kidd, S. 2013. The Misuse of the Term ‘Graduation’
in Social Policy. London: Development Pathways.
<http://www.developmentpathways.co.uk/
resources/wp-content/uploads/2014/06/14-
PathwaysPerspectives-Misuse-of-graduation-
pathways-perspectives-14.pdf>.
Accessed 25 April 2017.
Kidd, S. 2015. The Political Economy of ‘Targeting’
of Social Security Schemes. London: Development
Pathways. <http://www.developmentpathways.
co.uk/resources/wp-content/uploads/2015/10/
Political-Economy-of-Targeting-PP19-4.pdf>.
Accessed 25 April 2017.
McCord, A., and R. Slater. 2015. “Social Protection
and Graduation through Sustainable Employment.”
IDS Bulletin 46: 134–144. <https://opendocs.ids.ac.
uk/opendocs/bitstream/handle/123456789/7277/
IDSB_46_2_10.1111-1759-5436.12136.
pdf?sequence=1>. Accessed 25 April 2017.
Mkandawire, T. 2005. Targeting and universalism
in poverty reduction. Geneva: United Nations
Research Institute for Social Development.
<http://www.unrisd.org/unrisd/website/
document.nsf/240da49ca467a53f80256b4f005
ef245/955fb8a594eea0b0c1257000493eaa/$
FILE/mkandatarget.pdf>. Accessed 25 April 2017.
De Montesquiou, A., T. Sheldon, F.F. DeGiovanni,
and S. Hashemi. 2014. From Extreme Poverty
to Sustainable Livelihoods: A Technical Guide
to the Graduation Approach. Washington, DC:
Consultative Group to Assist the Poor. <http://
www.cgap.org/sites/default/les/graduation_
guide_nal.pdf>. Accessed 25 April 2017.
Roelen, K. 2015. “The ‘twofold investment
trap’: children and their role in sustainable
graduation.” IDS Bulletin 46:, 25–34. <https://
opendocs.ids.ac.uk/opendocs/bitstream/
handle/123456789/7281/IDSB_46_2_10.1111-
1759-5436.12126.pdf?sequence=1>.
Accessed 25 April 2017.
United Nations. 2012. Guiding principles on
extreme poverty and human rights, submitted by
the Special Rapporteur on extreme poverty and
human rights, Magdalena Sepúlveda Carmona.
New York: United Nations. <http://www.
ohchr.org/Documents/Publications/OHCHR_
ExtremePovertyandHumanRights_EN.pdf>.
Accessed 25 April 2017. United Nations. 2016a.
“Sustainable Development Goal 1: End poverty
in all its forms everywhere.” United Nations
Sustainable Development Knowledge Platform
website. <https://sustainabledevelopment.
un.org/sdg1>. Accessed 25 April 2017.
United Nations. 2016b. Report of the Special
Rapporteur on extreme poverty and human rights,
Philip Alston on Social Protection Floors (A/69/297).
New York: United Nations. <http://socialprotection-
humanrights.org/resource/report-of-the-special-
rapporteur-on-extreme-poverty-and-human-
rights-a69297/>. Accessed 25 April 2017.
Whitehead, L. 2016. “The Ultra Poor Graduation
Approach.” UNRISD Seminar. <https://www.
youtube.com/watch?v=_lK8NG524ko>.
Accessed 25 April 2017.
1. This contribution is written in a personal
capacity and does not necessarily reect
the position of the International Labour
Organization (ILO). Comments from Uma Rani
Amara, Stephen Kidd, Ian Orton, Ippei Tsuruga,
Fábio Veras Soares and Veronika Wodsak are
gratefully acknowledged.
2. International Labour Organization (ILO).
3. While there is a signicant variety of denition
of graduation approaches, this article follows the
following denition: “graduation programmes
aim to provide a sequenced and intensive
package of support to very poor people, with
the objective of facilitating their movement out
of poverty towards resilient and sustainable
livelihoods” (Devereux and Sabates-Wheeler
2015, 1). These packages typically include a
social protection component (e.g. cash transfer),
in addition to a transfer of productive assets
(often livestock), access to savings and credit
facilities, and in many cases also livelihood
training or coaching (for BRAC’s approach,
see BRAC 2016).
4. The importance of continued social protection
coverage has been recognised also within
the graduation community. Recently, Lauren
Whitehead (2016) of BRAC emphasised that
“graduation aims to enhance the visibility of the
poorest, providing a ‘big push’ and placing them
on an upward trajectory from extreme poverty
and towards, not out of, formal government
protection and social services”.
5. To be clear, this is not exclusive to graduation
approaches; similar targeting mechanisms are
also used by cash transfer programmes, which
do not follow an explicit graduation approach.
36
What we know about graduation impacts
and what we need to nd out
Nathanael Goldberg1
Rarely has an antipoverty strategy been
evaluated so thoroughly as the Graduation
Approach, the holistic livelihood
development programme popularised by
BRAC.2 The Graduation Approach includes
ve or more components designed to
ensure that beneciaries—typically
people living in extreme poverty—are
able to manage or avoid new shocks
while nding a pathway out of poverty.
Targeted households are provided with
consumption support (cash or food
assistance) to meet basic daily needs, an
income-generating asset (or a combination
of assets, usually livestock) along with
training in managing the asset, a savings
account (or savings groups where banking
is unavailable), and coaching or mentoring
over a two-year period to reinforce lessons,
monitor households’ progress, provide
moral support and help to overcome any
challenges along the way.
In 2006 the Consultative Group to Assist
the Poor (CGAP) and the Ford Foundation
teamed up to determine whether
BRAC’s Ultra-Poor Graduation Approach
could be adapted successfully outside
Bangladesh. They identied ten partners
in eight countries around the world
and, with much foresight, invested in an
evaluation strategy that would provide an
impressive body of evidence once these
programmes had run their course nearly a
decade later. The eight sites were subject
to randomised evaluations, which were
complemented with rigorous qualitative
research. Innovations for Poverty Action
(IPA) conducted seven of these evaluations,
while one of the research sites in India
was managed by our partners at J-PAL.
We pooled the data from six3 of the
randomised evaluations and published
the results in Science (Banerjee et al. 2015).
What we know now
The results showed positive impacts on
every outcome we looked at, including
income and revenues, total per capita
consumption, assets, food security,
women’s empowerment, physical health,
nancial inclusion, mental health, total time
spent working and political involvement.
Most of the outcomes were remarkably
stable from year 2, when the programme
completed, through year 3, a full year after
households stopped receiving any services
from the programmes. Two results, physical
health and women’s empowerment, were
no longer statistically signicant by year
3, though the direction of the impact
remained positive. There was very little or
no decline in the impact of the programme
after 36 months on the key outcomes,
including consumption, household assets
and food security.
The magnitudes of the individual
impacts are relatively modest (per capita
consumption increases by about 5 per
cent compared to the control group), but
they do add up. We conducted a thorough
cost-benet analysis, which showed a
benet-cost ratio of 166 per cent across
all the sites, with the highest result being
over 400 per cent. The benets were based
primarily on a projection of the three-year
impact of the programme on per capita
consumption. This calculation rested on
an assumption that the impacts would
continue into the future, based on the
stability of results from year 2 to year 3.
This assumption was bolstered by results
from a separate randomised evaluation of
BRAC’s original programme, which showed
strong impacts after four years (Bandiera
et al. 2016) and later greatly strengthened
by a long-term follow-up at one of the
sites (Bandhan, India), which showed that
impacts actually increased after seven
years (Banerjee et al. 2016).
The Ford Foundation and CGAP’s
investment in research paid o.
The publication of the results drew wide
attention, including coverage in the
New York Times and The Economist and
on US National Public Radio.
Today, CGAP counts 55 graduation sites
implemented by governments and non-
governmental organisations, and is further
extending to new populations, including
the urban poor and refugees. The United
Nations High Commissioner for Refugees
(UNHCR), with technical assistance from
Trickle Up, is piloting Graduation for
refugees in ve countries, with plans
to expand to 22. CGAP and BRAC have
released guides to help new implementers
plan graduation programmes. However,
we still know relatively little about how
best to design and deliver graduation.
The CGAP–Ford Foundation evaluations
primarily tested the full graduation package
compared to control households who
received nothing, so we know less about
the impacts of individual components.
The evidence for the impact of that package
remains strong, but the median cost to
deliver the package was around USD1,100—
beyond the reach of many governments.4
Is the full package necessary, or might a
reduced form enable more households
to benet from the programme?
As graduation gets further embedded into
safety net programmes, governments will
need to nd ways to identify those who
should receive the programme and can
benet from it. An analysis of targeting in
Honduras and Peru showed the three-step
graduation targeting method (geographic
targeting, participatory appraisal and
verication check) failed to perform much
better than random sampling within a
poor community (Karlan and Thuysbaert
2016). But community targeting can
still have benets that may make such
approaches worthwhile. Alatas et al. (2012)
found that while community targeting
does not outperform proxy means tests
on objective measures of poverty, it results
in greater community satisfaction.
Not all graduation programmes explicitly
target women, and it remains unclear
whether within-family targeting can
improve women’s empowerment or other
outcomes. Is the standard Graduation
Approach sucient, or could something
like gender-specic training improve
outcomes for women? Future research
may address how graduation can increase
female empowerment, as measured by
asset and land ownership, social networks
and decision-making power.
The International Policy Centre for Inclusive Growth | Policy in Focus 37
“ The Graduation
Approach in isolation,
without income
generation, does not
achieve the stated goals
of the programme, and
income generation by
itself is also insufcient.
Photo: Conor Ashleigh/AusAID. A women's group, graduated through BRACs TUP programme, gather for a
meeting, Rodrodha village, Bangladesh, 2012 <https://goo.gl/nMu7E9>.
Going forward: unpacking the pieces
Two lessons from the CGAP–Ford
Foundation sites shed light on the holistic
nature of the Graduation Approach, which
has livelihood creation at its core. In one of
the sites (Honduras) the income-generating
activity did not pay o for beneciaries. Most
participants chose poultry as their income-
generating activity; however, the chickens
they received died o in large numbers from
disease, and household consumption was
not higher among programme participants.
In Ghana, we were able to test just the
income-generating assets (goats), with
no other supporting services: no training,
coaching or savings accounts. In this case,
the households that only received goats
had more of them three years later, but
despite the windfall in assets (approximately
USD250 in goats) they had no more net
worth in livestock and consumed no
more per capita than control households.
From these examples, we learn that the
Graduation Approach in isolation, without
income generation, does not achieve the
stated goals of the programme, and income
generation by itself is also insucient.
However, these are extreme cases.
IPA has a research agenda designed to
optimise the Graduation Approach by
learning more about poverty traps and
what it takes to move households out of
extreme poverty. Some of this, such as
nding ways to improve psycho-social
outcomes for beneciaries, is frontier
research, and some will require only
simple tests to determine how much of
each component is necessary to create
a positive impact. In some cases, simply
trying the programme with and without
individual components can shed light on
the nature of poverty traps. For example,
household visits are often the most
expensive component of the programme
to deliver, representing 30–40 per
cent of the total programme budget.
Naturally, programme implementers
will be interested to know whether the
coaching component is essential to the
success of the programme. Testing the
programme with and without coaching
will help reveal the constraints faced by
extremely poor people: do they primarily
lack capital and technical skills, or are
behavioural constraints (such as lack
of condence) more binding?
Optimising component levels
Interesting as such a test with and without
coaching might be, perhaps the right
answer is somewhere in between. The classic
version of the programme calls for weekly
coaching visits over a two-year period. Are
weekly visits necessary? At the Peru site,
in the mountains surrounding Cusco, the
beneciaries were simply too remote to
visit every week. Households were visited
every six weeks and fared reasonably well,
though not as well as in the top-performing
sites. Whether this is because of the limited
coaching or another explanation such as
limited access to markets is impossible
to say. Blattman et al. (2016), evaluating a
package of cash, business skills training and
supervision among extremely poor people
in Uganda, varied the number of follow-
up visits provided to participants. Some
households received two visits to ensure
beneciaries invested the cash, while others
received ve visits, for both commitment
to invest and business advice. Those who
received visits were more likely to have a
surviving business but did not have more
income or consumption.
Yet were two or ve the right numbers?
Five is much closer to zero than the 104
visits a beneciary would expect in the
Graduation Approach. Perhaps ve is too
few, and 104 more than necessary. A robust
research agenda would require testing
many permutations of the graduation
components to determine the optimal
intensity of each component, measured
by cost-eectiveness analysis: the greatest
impact per dollar spent by the programme.
Such a research agenda would not only
consider the classic graduation model as
it is now, but would allow for variations
to determine which ones can improve
the cost-eectiveness of the programme.
For example, Fundación Capital has
been working with tablets to replace
face-to-face coaching in Colombia.
E-coaching is likely to reduce costs, but
the cost-eectiveness of the programme
will improve only if the tablets perform
suciently well to do better than the
cost-eectiveness of face-to-face coaching.
This includes designing software that can
be used by illiterate beneciaries, and
solving the last-mile problems of making
sure the tablets remain charged and in
working order. If the e-coaching initiative
is designed well enough, it may even
38
“ The question remains
why some people
were able to seize the
opportunity to sharply
change their trajectory
out of extreme poverty,
while others improved
just a bit.
Photo: BRAC/Alison Wright. Girls carrying vegetable on their heads, Tanzania, 2014.
work better than traditional coaching by
ensuring consistency in messaging and
allowing the households to work through
materials at their own pace.
Group approaches may also increase the
cost-eectiveness of the graduation model.
In Kenya, the BOMA project5 provides
cash grants of approximately USD300 to
a group of three women—the amount
other graduation programmes typically
spend on assets for a single beneciary.6
Group accountability and support may
allow BOMA to create impact at lower
cost. Nevertheless, is the investment in
productive assets per beneciary sucient
to put them on the path out of extreme
poverty? The eect of group versus
individual delivery of graduation has yet
to be rigorously evaluated, and must be
weighed against the additional burden
of forming and following small groups.
Adjusting the model for
those who don’t succeed
Analysis of the distribution of impacts of
the CGAP–Ford Foundation sites showed
that while everyone benets from the
programme on average, some benet
much more than others. Even those at the
10th percentile on many outcomes—such
as consumption, assets, and income—
show some gains, but the dierence
was quite small compared to those at
the upper end of the distribution. The
question remains why some people were
able to seize the opportunity to sharply
change their trajectory out of extreme
poverty, while others improved just a
bit. Those who do least well or backslide
in the traditional programme may need
more support or a dierent programme
altogether. We are working with Trickle
Up to test a version of the programme
whereby eld sta would identify the
households that need the most support
and allocate their time to ensure that those
with the greatest need get additional
coaching. Or perhaps livelihood choice
is the critical factor. Evidence suggests
that some livelihood choices were more
protable than others, but does that reect
the inherent protability of the livelihoods
or the types of participants who selected
them? Households could potentially be
nudged into selecting livelihoods with
greater income potential.
Perhaps some beneciaries struggle
to engage with new livelihoods due to
underlying psychosocial capabilities.
In Ghana, we are testing the addition
of group-based cognitive behavioural
therapy (CBT) to reduce depression and
improve the forward-looking aspirations
among poor people before they enter the
graduation programme. The CBT has been
shown to reduce depression and improve
productivity among patients in India and
Uganda and may help ‘ultra-poor’ people
engage more productively with their new
livelihoods (Thomas and Haushofer 2015).
Challenges of scaling-up
An ideal programme would be customised
to the individual needs and potential of
each household, with those requiring
fewest resources given only what they
need, and others receiving more intensive
services. Designing an evaluation to learn
how to do that would be unrealistically
expensive, but we can learn a lot from
evaluations of individual scale-up
solutions. For maximum scalability,
households could simply be given cash
grants rather than any of the support
services in the Graduation Approach.
An evaluation of cash grants7 provided
by GiveDirectly in Kenya showed positive
impacts on consumption, assets and
psychological well-being (Haushofer
and Shapiro 2016), but the follow-up
period for the endline survey was much
shorter (four months) than in graduation
evaluations, and the targeting was
performed dierently than by graduation
programmes. New studies directly
comparing graduation to cash grants will
help determine whether the holistic nature
of the Graduation Approach outperforms
cash, in cost eectiveness, for the poorest
and most vulnerable households.
A hybrid approach providing much of
the support structure of graduation while
easing procurement challenges is to give
cash rather than in-kind assets, along
with other graduation services. But will
households invest the cash well? In this
sense, programmes could provide cash at
the marketplace, with assistance in selecting
protable assets and healthy livestock.
Technology solutions, such as e-coaching
as previously mentioned, can potentially
boost the impact of lighter-touch
programmes at low cost. Some potential
The International Policy Centre for Inclusive Growth | Policy in Focus 39
“ As the programme
increases in reach and
density, several scale-up
questions about the
impact of the approach
will increase
in importance.
Photo: Conor Ashleigh/AusAID. Woman received a cow and goat from BRAC through its TUP programme,
Kazipara, Bangladesh, 2012 <https://goo.gl/h8u7Wf>.
examples include providing information
on market prices, or accountability
mechanisms to help beneciaries stick to
their plans. These can be readily tested
with randomised evaluations.
The Graduation Approach is now being
brought to scale in several countries,
including Ethiopia, Pakistan and
Colombia, with interest and activity in
many more. As the programme increases
in reach and density, several scale-
up questions about the impact of the
approach will increase in importance,
starting with the general equilibrium
eects of the programme. What happens
to goat prices when so many new entrants
are given livestock? Neither Banerjee
et al. (2015) nor Bandiera et al. (2016)
nd evidence of crowding out among
non-beneciaries, though the ultra-poor
people targeted by BRAC represent only
the bottom 6 per cent of the population.
There is strong interest in adapting the
Graduation Approach to serve the urban
poor and livelihood options will have to
be adjusted for urban contexts, including
more choices beyond livestock.
So far, most graduation programmes are
working within the constraints of existing
value chains. The standard graduation
model identies a menu of livelihood
activities for beneciaries and works with
households to match them to appropriate
activities. Programmes could potentially
improve revenues by facilitating group
input purchases and market price
information. It may be possible to push out
the curve of potential livelihoods by linking
poor people to markets or supporting the
creation and expansion of local and national
value chains. The BRAC enterprise model
promotes quality products, fair producer
pricing and market literacy across entire
value chains in Bangladesh. For example,
their poultry operations include a feed mill,
chick distribution, processing plants and
package printing. Is such a vertical approach
possible outside BRAC and Bangladesh?
The CGAP–Ford Foundation programme
was set up to answer just such a question
for the original BRAC model. The research
questions mentioned in this article could
be answered with a similarly ambitious
investment in the delivery and evaluation of
next-generation graduation programmes.
Alatas, Vivi, Abhijit Banerjee, Rema Hanna,
Benjamin A Olken, and Julia Tobias. 2012.
“Targeting the Poor: Evidence from a Field
Experiment in Indonesia.” American Economic
Review 102(4): 1206–40. doi:10.1257/aer.102.4.1206.
Bandiera, Oriana, Robin Burgess, Narayan
Das, Selim Gulesci, Imran Rasul, and Munshi
Sulaiman. 2016. “Labor Markets and Poverty
in Village Economies.” LSE Working Paper.
London: London School of Economics.
<http://sticerd.lse.ac.uk/dps/eopp/eopp43.
pdf>. Accessed 20 April 2017.
Banerjee, Abhijit, Esther Duo, Raghabendra
Chattopadhyay, and Jeremy Shapiro. 2016.
“Long Term Impact of a Livelihood Intervention:
Evidence from West Bengal.” Working Paper.
Banerjee, Abhijit, Esther Duo, Nathanael
Goldberg, Dean Karlan, Robert Osei, William
Parienté, Jeremy Shapiro, Bram Thuysbaert, and
Christopher Udry. 2015. “A Multifaceted Program
Causes Lasting Progress for the Very Poor:
Evidence from Six Countries.” Science 348(6236):
1260799. doi:10.1126/science.1260799.
Blattman, Christopher, Eric P. Green, Julian
Jamison, M. Christian Lehmann, and Jeannie
Annan. 2016. “The Returns to Microenterprise
Support among the Ultrapoor: A Field
Experiment in Postwar Uganda.” American
Economic Journal: Applied Economics 8(2): 35–64.
Haushofer, Johannes, and Jeremy Shapiro.
2016. “The Short-Term Impact of Unconditional
Cash Transfers to the Poor: Experimental
Evidence from Kenya.” Quarterly Journal
of Economics 131(4).
Karlan, Dean, and Bram Thuysbaert. 2016.
“Targeting Ultra-Poor Households in Honduras
and Peru.” The World Bank Economic Review.
Thomas, Catherine, and Johannes Haushofer.
2015. “Get Happy, Get Rich.” Foreign Aairs,
April 21. <https://www.foreignaairs.com/
articles/africa/2015-04-21/get-happy-get-rich>.
Accessed 20 April 2017.
1. Innovations for Poverty Action.
2. A non-governmental development
organisation based in Bangladesh.
3. The Yemen evaluation was delayed due to
conict, but the data were eventually collected,
and data analysis is being completed.
4. Where potential beneciaries already
receive cash transfers under pre-existing safety
net programmes, the cost of implementing
Graduation will be reduced, as there is no
need to include consumption support.
5. A US non-prot and Kenyan
non-governmental organisation.
6. Village Enterprise also provides grants to
groups of women, though in slightly larger
amounts (USD500).
7. Cash grants dier from other social cash
transfer programmes in that they are explicitly
unconditional, large and concentrated in time.
In the case of Kenya, the experiment relied on two
payment modalities: monthly instalments over
nine months vs. a one-time lump sum transfer;
and two transfer magnitudes: USD404 PPP vs
USD1,525 PPP (Haushofer and Shapiro 2016).
40
Responsible graduation
Keetie Roelen,1 Stephen Devereux 1 and
Rachel Sabates-Wheeler 1
The ‘Graduation Approach’ has enjoyed a
surge in policy and academic attention in
the past few years. Graduation programmes
provide a comprehensive mix of sequenced
support that includes consumption
transfers, asset transfers, access to savings
and credit, training and coaching (Hashemi
and Umaira 2011), oering a holistic
approach to poverty reduction. They are
grounded in the theory that extremely
poor households require a big push to
set in motion a positively reinforcing
cycle of income generation and asset
accumulation (Carter and Barrett 2007).
Graduation programmes have been
applauded for their success in ‘graduating’
people out of extreme poverty in
Bangladesh and elsewhere (Devereux and
Sabates-Wheeler 2015). A recent multi-
site evaluation found that graduation
programmes can improve food security,
assets, income and consumption, and that
impacts are sustained one year after the
end of the programme (Banerjee et al.
2015). Evidence from BRAC’s Targeting the
Ultra Poor (TUP) programme in Bangladesh
indicates that the programme sets poor
women on a positive trajectory of asset
accumulation and poverty reduction
following increased earnings after
participation in the programme (Bandiera
et al. 2016). Findings in Burundi and
Rwanda show that programmes also have
positive eects on non-material outcomes
such as hygiene practices and social
relations (Devereux et al. 2015; Devereux
and Sabates 2016). Evidence from
Bangladesh and Burundi also indicates
that impacts stretch beyond direct
beneciaries and lead to positive changes
for the wider community (Devereux et al.
2015; Raza and Van de Poel 2016).
Notwithstanding this evidence of success,
not all of the attention that graduation
programmes have attracted has been
positive. Questions have been raised about
the long-term sustainability of graduation
and the cost of implementing these
programmes, especially if they scale up from
pilot projects to national programmes, and
whether governments—unlike grassroots
non-governmental organisations—have
the human resources needed to deliver
the intensive support (coaching) that
is considered critical to graduation
success stories. Estimates suggest that
the overall cost per participant ranges
from roughly USD1,500 to USD6,000 (for
two-year programmes), with the labour
requirements for the implementation of
programmes representing a considerable
share of those costs (Banerjee et al. 2015).
While cost–benet analyses point out
that benets outweigh the costs and that
programmes are cost-eective (ibid.), the
nancial implications nevertheless give rise
to discussions about the value-added of
individual components and the extent to
which support can be stripped down or
built up while still achieving positive change.
Challenges have also been raised at an
ideological level, with critics arguing that
graduation as a concept is antithetical
to the drive towards rights-based social
protection and a minimum ‘social protection
oor’ for all, which has been endorsed not
only by rights-based agencies such as the
International Labour Organization (ILO)
and the United Nations Children’s Fund
(UNICEF) but even by market-oriented
agencies such as the World Bank (see
the joint statement by the ILO and the
World Bank in June 2015). Graduation
implies withdrawing support—a ‘one-way
door’ out of the programme—and can,
therefore, be considered the epitome of
neoliberal social policy, oering a pathway
out of poverty but no support for those
needing continued or renewed protection
against poverty (Kidd 2013 in Devereux
and Sabates-Wheeler 2015). Rights-based
social protection, however, stipulates the
provision of support for anyone who
needs it at all times.
Thus, are graduation programmes just
another manoeuvre by governments and
donors who favour narrowly targeted time-
bound programmes, and who are always
looking for exit strategies to minimise their
spending on poor people?
We at the Centre for Social Protection
support the vision driving graduation
programmes, but we also believe in rights-
based approaches—specically, of course,
the right to social protection. This apparent
ambivalence explains the question
mark in the title of a recent Institute
of Development Studies IDS Bulletin:
‘Graduating from Social Protection?’
(Devereux and Sabates-Wheeler 2015).
Can we reconcile these apparently
contradictory stances?
Yes, we can. The solution lies in
‘responsible graduation’. This means
implementing graduation in such a
way that the livelihoods of participants
are strengthened and supported while
ensuring the right to social protection is
not violated. Our response to graduation
critics is that the challenges to graduation
programmes stem primarily from linear
and short-term understandings of theories
of change, feeding into inadequate
funding, short policy cycles and donor
time-frames and the lack of appropriate
monitoring and evaluation. These elements
coalesce, so that too often recipients
are being removed from programmes in
‘irresponsible’ ways, in the sense that there
is limited or no consideration given to
their human welfare and future well-being.
Graduation implemented in irresponsible
ways is antithetical to a rights-based
approach, yet it does not have to be.
We advocate the following principles on
how to ‘graduate’ programme participants
‘responsibly’ rather than ‘irresponsibly’.
yDon’t confuse ‘exit’ with ‘graduation’:
Most social protection programmes
include exit strategies—i.e. provisions
that stipulate the discontinuation of
programme support. These strategies
can be considered ‘exogenous’ or
‘endogenous’, with exogenous
provisions basing decisions on exit
independent of programme participants’
poverty status or living conditions, while
endogenous strategies do consider
changes in socio-economic criteria
(Samson 2015). The two terms are often
conated to denote ‘graduation’, yet only
the endogenous approach refers to a
discontinuation of programme support
following movements out of poverty.
The International Policy Centre for Inclusive Growth | Policy in Focus 41
Photo: BRAC. A member of BRAC's CFPR—TUP programme, taking care of her nursery, Bangladesh, 2008
<https://goo.gl/NB6gJr>.
“ Programme design,
delivery and provision
should be appropriate to
individual contexts that
inuence the extent to
which participants are
able to capitalise on
graduation opportunities.
Referring to exit from programmes
without actual changes in living
conditions as ‘graduation’ is misleading
and directly opposes a rights-based
approach to graduation.
yA ‘revolving door’, not a ‘one-way door’:
Most graduation programmes register
participants only once, and see that they
leave the programme permanently after
the graduation cycle is complete—a
‘one-way door’ (see Pritchard et
al. (2015) on the Chars Livelihood
Programme in Bangladesh). Similarly,
other programmes—including BRAC’s
pioneering TUP—support participants
for a xed period, typically about two
years, after which participants exit the
programme, and support is terminated.
This contradicts the rights-based
principle that social protection should
be available to whoever needs it,
whenever they need it. We believe that
people should be allowed to move in
and out of programmes when they are
eligible and in need of their support.
The new agenda on ‘shock-responsive’
social protection systems speaks to
this, whereby ‘potential’ programme
recipients are registered within a social
protection system so that when shocks
occur and support is required, people
are able to register and make claims
on social provisioning. This would
apply equally to past graduates who
subsequently fall back into programme
eligibility—they should be allowed
to re-register. While recognising that
implementers face resource constraints,
programmes often lack the degree of
exibility to allow ‘re-entry’ but move to
new cohorts of eligible groups instead.
yAppropriate programmes: Many
graduation programmes operate a
single income-generating model
or oer participants a limited set of
livelihood options to choose from.
Programme design, delivery and
provision should be appropriate to
individual contexts that inuence the
extent to which participants are able to
capitalise on graduation opportunities.
This includes acknowledgement of
individual situations such as household
composition, dependency ratio and
ability to do physically demanding
work, as well as community-level
factors such as availability of markets.
Programmes can provide more tailored
responses by introducing ‘full-family
targeting’ or including child care
services for households with children,
for example (Roelen 2015).
yGraduation should be based on applying
clear and consistent eligibility criteria
to determine whether each participant
has exceeded graduation thresholds.
Households that have not reached
these thresholds should not be ‘exited’
but should remain on the programme
for another cycle, or until they are
assessed as ready to graduate, however
long that takes. This principle requires
that robust and transparent targeting,
monitoring and evaluation systems
and indicators are in place.
yGraduation should facilitate movement
into other support as needed:
‘Developmental’ graduation (Samson
2015) sees graduation as a continuous
pathway rather than a ‘threshold’
outcome. Instead of being abandoned,
graduates should move from social
assistance into social services, gain
access to micronance and so on.
The principle is graduation ‘into’ other
forms of support, rather than ‘out of ’ all
forms of support. This is a rights-based
principle that ensures ‘rights across the
life-course’.
yGraduation programmes require
accountability mechanisms: Issues
such as inappropriate targeting,
premature graduation (either
individually or en bloc) and inadequate
support post-graduation require
strong and transparent accountability
mechanisms. The establishment of
grievance and complaints procedures,
including an adequate response
mechanism, is crucial for ensuring that
graduation is underpinned by rights.
yGraduation programmes should
be accompanied by labour market
policies addressing structural
barriers to employment. Graduation
programmes focus on the capacities
of individuals, households and
communities to build assets, increase
productivity and diversify their
livelihoods. The large majority of
those initiatives are premised on a
model of entrepreneurship and self-
42
“ Instead of being
neglected or
abandoned once
the programme cycle
ends, these households
should be moved out of
graduation programmes
and into social
assistance programmes,
either permanently
or temporarily.
Photo: BRAC/Alison Wright. Woman runs a small tailoring business, Tanzania, 2012.
employment, raising broader questions
of how graduation programmes can
support moves into stable employment
(McCord and Slater 2015). Formal jobs
not only represent a cornerstone of
sustainable employment but may
also be favoured over self- or informal
employment (Sumberg et al. 2015).
Graduation programmes should be
implemented within a wider remit of
structural policy, to avoid governments
being absolved of their responsibilities
and individuals and communities
bearing the brunt of creating their
own employment opportunities.
yGraduation is not for everybody:
Some households will not be able to
reach graduation thresholds within
the time-frame of the programme
because they lack the necessary
capabilities for generating self-reliant
livelihoods (due to health constraints
or old age, for example). Instead
of being neglected or abandoned
once the programme cycle ends,
these households should be moved
out of graduation programmes and
into social assistance programmes,
either permanently or temporarily.
In Ethiopia’s graduation-oriented
Productive Safety Net Programme
(PSNP), for example, women who fall
pregnant are moved out of public
works and into direct support until
their newborn child is one year old.
Those unable to work receive direct
support without an expectation to
‘graduate’ out of programme support.
Even if most graduates remain better o
than before they joined the programme,
there will inevitably be some who fall
back into poverty after they stop receiving
programme suppor t. ‘Responsible
graduation’ requires making appropriate
arrangements for these people. Either
they should be allowed to rejoin the
programme if they can demonstrate their
eligibility, as argued above, or assisted
into other programmes that oer poor
people social assistance. Follow-up
assessments would help to determine
whether graduation was sustained and
to oer appropriate support to those for
whom it was not. Some thought about
reconceptualising and redesigning
graduation programmes along the lines
suggested in this article may help to
ensure that livelihoods are supported
over the long term and in a way that
adheres to a rights-based understanding
of development.
Bandiera, O., R. Burgess, N. Das, S. Gulesci,
I. Rasul, and M. Sulaiman. 2016. Labour Markets
and Poverty in Village Economies. London:
London School of Economics.
Banerjee, A., E. Duo, N. Goldberg, D. Karlan, R.
Osei, W. Parienté, and C. Udry. 2015. “A multifaceted
program causes lasting progress for the very poor:
Evidence from six countries.” Science 348(6236).
Carter, M.R., and C.B. Barrett. 2007. “Asset
Thresholds and Social Protection: A ‘Think-Piece’.”
IDS Bulletin 38(3): 34–38.
Devereux, S., and R. Sabates. 2016. Final
Evaluation Report. Enhancing the Productive
Capacity of Extremely Poor People in Rwanda.
Dublin: Concern Worldwide.
Devereux, S., and R. Sabates-Wheeler. 2015.
“Graduating from Social Protection? Editorial
Introduction.” IDS Bulletin 46(2): 1–12.
Devereux, S., K. Roelen, R. Sabates, D.Stoelinga,
and A. Dyevre. 2015. Final Evaluation Report.
Concern’s Graduation Model Programme in
Burundi. Dublin: Concern Worldwide.
Hashemi, S., and W. Umaira. 2011.
“New pathways for the poorest: the graduation
model from BRAC.” CSP Research Report 10.
Brighton: Centre for Social Protection,
Institute of Development Studies.
ILO and World Bank. 2015. A Shared Mission
for Universal Social Protection. Concept Note.
Geneva/Washington, DC: International Labour
Organization and World Bank.
Kidd, S. 2013. “The Misuse of the Term ‘Graduation’
in Social Policy.” Pathways Perspectives 14.
Banbury, UK: Development Pathways.
Pritchard, M., S. Kenward, and M. Hannan.
2015. “The Chars Livelihoods Programme in
Bangladesh: Factors that enable, constrain and
sustain graduation.” IDS Bulletin 46(2): 35–47.
Raza, W., and E. Van de Poel. 2016. Impact and
Spillover Eects of an Asset Transfer Programme
on Malnutrition. Dhaka, Bangladesh: BRAC.
Roelen, K. 2015. “The two-fold investment
trap: children and their role in sustainable
graduation.” IDS Bulletin 46(2): 25–34.
Samson, M. 2015. “Exit or Developmental Impact?
The Role of ‘Graduation’ in Social Protection
Programmes.” IDS Bulletin 46(2): 13–24.
Sumberg, J., T. Yeboah, J. Flynn, and N.A.
Anyidoho. 2015. “Perspectives on jobs and
farming: Findings from a Q study with young
people, parents and development workers in
rural Ghana.” FAC Working Paper 109. Brighton:
Institute of Development Studies.
1. Co-directors of the Centre for Social Protection.
The International Policy Centre for Inclusive Growth | Policy in Focus 43
(Accidentally) Harvesting higher hanging
fruits: addressing under-5 malnutrition
using the Graduation Approach1
Wameq A. Raza 2
Despite signicant improvements in
various social and health indicators
over the past decades, children under
ve years of age (under-5s) are still
disproportionately aected by extreme
poverty and, consequently, malnutrition.
A recent paper by the World Bank
Group and UNICEF (2016) suggests
that the largest proportion of the 767
million people living under the extreme
poverty line of USD1.90 per day are
under-5 children (15.9 per cent).3 The
consequences of such a high burden
of malnutrition are far-reaching and
detrimental, not only to these children,
but also to the societies in which they
live. A number of targeted and general
social protection programmes have been
implemented to combat both poverty
and, in some cases, also malnutrition, with
varying results. Although the innovative
‘Graduation Approach’ focuses more
on poverty reduction, socio-economic
empowerment and food security
outcomes than malnutrition per se, this
article explores whether this approach has
contributed to this issue in Bangladesh.
Drivers and consequences of malnutrition
The UNICEF framework suggests that
child and maternal malnutrition are
exacerbated by a lack of access to maternal
and child health care, inappropriate health
practices and a lack of knowledge about
and access to safe water and sanitation
services, all of which are exacerbated by
poverty (Gartner et al. 2005; WHO 2013).
Malnutrition is further compounded by
factors such as poor housing conditions
(Odunayo and Oyewole 2006), access to
mass media (Rahman, Chowdhury, and
Hossain 2009) and cultural practice—for
instance, in Bangladesh, the administration
of formula milk to infants was considered
a sign of wealth and a higher social status
(Egata, Berhane, and Worku 2014; Hien and
Kam 2008; Rayhan and Khan 2006; Bill &
Melinda Gates Foundation 2007).
Such a high burden of childhood
malnutrition is an acute public health
problem, over both the short and the
long terms. Early stages of malnutrition
lead to extreme weight loss, stunting
and susceptibility to infections and
chronic illnesses. Prolonged exposure,
especially within the rst 1000 days, can
permanently limit cognitive development
and productivity (Brown and Pollitt 1996;
Black et al. 2008). Malnourished children
living in poverty are signicantly more
likely to remain trapped in their current
socio-economic status in later life, thereby
precipitating an intergenerational cycle
of poverty (Black et al. 2008).4 Moreover,
malnourished girls are more likely to have
poorer pregnancy outcomes, leading to
higher probabilities of children with low
birthweight and culminating in higher rates
of maternal and neonatal mortality (ibid.).
Global attempts at curbing
child malnutrition
The Millennium Development Goals are
considered one of the initial drivers of global
eorts to curb malnutrition, especially for
under-5s. A variety of programmes exist with
the exclusive goal of reducing malnutrition
among children, varying from supplemental
feeding to nutritional supplements such as
micro-nutrient packets and encouraging
exclusive breastfeeding for the rst six
months. However, the results have thus
far been mixed (Allen and Gillespie 2001).
For instance, the Bangladesh Integrated
Nutritional Project (BINP), implemented at
scale by the World Bank, aimed to improve
maternal and child nutritional status by
improving the knowledge of caring practices
via training, education and the provisions for
supplementary feeding. While the mid-term
eects were modest, the impact all but
evaporated by the end of the programme
(White 2005; Pelletier et al. 2005).
The late 1990s saw the spread of social
assistance interventions such as conditional
cash transfer (CCT) programmes. The receipt
of the CCTs is typically predicated on the
beneciaries participating in predetermined
conditionalities set out by the programme
such as sending children to school or using
certain forms of health care (Manley, Gitter,
and Slavchevska 2012). A rising number
of positive evaluations have provided
further impetus to their popularity and
growth—while nearly all countries in Latin
America have implemented some sort of
CCT programme, they have also spread
substantially across Asia (Benhassine et al.
2015; Haushofer and Shapiro 2016; Manley,
Gitter, and Slavchevska 2012). Though the
positive eects of these social assistance
programmes on core outcomes such as
education and greater use of health care
are generally beyond reproach, as are
their impacts on consumption to
some degree, their ecacy in reducing
malnutrition remains mixed (Fiszbein and
Schady 2009; Angelucci and De Giorgi
2006; Angelucci 2015).
Evidence from the Graduation Approach
Ushering in a new era of social
protection programmes, the concept
of the Graduation Approach has taken
o since the early 2000s when BRAC,
an international non-governmental
organisation from Bangladesh, pioneered
the Targeting the Ultra Poor (TUP)
programme through its Graduation
initiative.5 Enrolled after a rigorous three-
step targeting procedure, participants
receive a host of interventions over a two-
year period, including income-generating
asset transfers (such as livestock),
comprehensive business development
training and bi-weekly mentorship sessions
to embed and apply essential life-skills
knowledge. While TUP’s primary mandate
is to socio-economically improve the lives
of the ‘ultra-poor’, the programme is also
nutrition and health sensitive. Participation
within the TUP programme facilitates the
participants’ access to the ‘Essential Health
Care’ package, which includes health and
nutrition education, covering topics such as
the importance of exclusive breastfeeding,
child immunisation, pregnancy care, oral
44
“ Participation within the
TUP programme facilitates
the participants’ access
to the ‘Essential Health
Care’ package, which
includes health and
nutrition education.
Photo: UN Photo/Kibae. A mother and her newborn baby, Bangladesh, 2010 <https://goo.gl/TyP4nL>.
rehydration therapy, and the provision of
basic curative care for common illnesses
by BRAC Community Health Promoters. If
the patient is unable to pay for advanced
care, TUP sta arrange the funds through
community mobilisation eorts. Lastly,
the programme guarantees access to
sanitary latrines and safe drinking water
for the participants through the existing
BRAC WASH programme, either through a
direct transfer to beneciaries or to nearby
facilities via negotiations with their owners.
Evidence from the rst phase of the
programme (2002–2006) shed some
light on its eects on adult calorie
consumption and child nutritional status.
Haseen and Sulaiman (2007) reported
that participation led to both an increase
in mean calorie intake from 1,750 to
2,138 calories per day and in the quality
of the calories consumed.6 Studying the
causal eects on the nutritional status of
under-5 children, Jalal et al. (2009) were
able to identify positive eects on the
weight-for-height z-scores of children
aged 24–36 months (0.32 SD).7 This study,
however, used a non-experimental design
and relied on a control group of children
who live within the districts targeted by
TUP. Spill-over eects could, therefore,
have been responsible for the nutritional
improvement among the control groups
and hence the seemingly small impact of
the programme on TUP participants.
The second phase of TUP was rolled out
between 2007 and 2011 across 13 districts
in Bangladesh. Using the randomised
rollout of the programme, Raza and van
de Poel (2016) identify its causal eects
on the nutritional status of under-5
children living in participant households
using data amassed from 8,000 ultra-poor
households. By comparing trends across
12,500 non-participant poor households
living in treated and control areas, the
authors next identify the spill-over eects
of the programme on under-5 children.
Lastly, the authors report the most
important pathways through which the
eects are channelled. They report a marked
improvement in the nutritional status of
children living in ultra-poor households
(see Table 1). The weight-for-height (WHZ)
indicator increases by 0.78 SD, representing
a 40 per cent increase over the baseline
level. Similarly, ultra-poor under-5 children
experienced a 0.52 SD increase in the
weight-for-age (WAZ) indicator, a 31
per cent increase over the baseline. This
translates into a reduction in the likelihood
of wasting (WHZ<-2 SD) by 8 percentage
points, followed by a reduced probability
of being underweight (WAZ<-2 SD) of
19 percentage points. Signicant eects
on the height-for-age indicator (stunting)
were not detected, suggesting that there
is some room for improvement.
Results also suggest that TUP has
signicant spill-over eects among under-5
children living in poor non-participant
households in treated communities. These
poor households, also known as ‘other-
poor’, were not invited to take part in the
programme because they were disqualied
during the nal verication survey after
the initial village wealth ranking.8 Similar
to the eects on the treated households,
the WHZ and WAZ indicators are signicant
and positive (0.45 SD and 0.28 SD,
respectively). The likelihoods of wasting
or being underweight among children in
these households were reduced by
9 percentage points and 12 percentage
points, respectively. The fact that the
eects on being underweight is lower
among the ‘ultra-poor’ than among the
‘other-poor’ is probably driven by a greater
proportion of the latter group falling near
the underweight cut-o point of -2 SD;
a smaller push, therefore, allows non-
participant children to be classied as
not underweight.
Raza and van de Poel continued to analyse
the eects of the TUP programme on
underlying determinants of childhood
malnutrition: likelihood of breastfeeding,
its exclusive duration, administration of
vitamin A and food security. Given its high
prevalence, while the authors found no
eects on its likelihood, they did report
large increases in the duration of exclusive
breastfeeding (75 per cent increase over
the baseline for the ultra-poor and 49 per
cent among the other-poor). Similarly, the
probability of a child receiving a vitamin A
supplement increased by 26 percentage
points and 20 percentage points among
ultra-poor and other-poor households,
respectively. The combination of these
factors is expected to have a larger eect
on children younger than 24 months, and
this is reected in the results where the
The International Policy Centre for Inclusive Growth | Policy in Focus 45
Note: *, **, *** indicate signicance at 1 per cent, 5 per cent and 10 per cent, respectively.
Source: Raza and Van der Poel (2016).
non-participants.9 Explicit targeting
of pregnant women and children in
the programmatic framework and the
deliberate use of behavioural interventions
merit further testing and incorporation
into the programme. Interventions such
as the administration of micro-nutrient
packets, supplemental feeding or food
transfers have proven to be moderately
successful on their own. The positive
eects of these programmes can be further
reinforced by embedding them within
TUP’s integrated framework. Lastly, from
a research perspective, it is important to
track the cognitive development10 among
the members of participant households to
ensure whether these nutritional gains
are indeed translating into higher
cognitive abilities between the short
and the long term.
Allen, L.H., and R.S. Gillespie. 2001. “What Works?
A Review of the Ecacy and Eectiveness of
Nutrition Interventions.” ACC/SCN Working Paper
Series. Manila: United Nations Administrative
Committee on Coordination Sub-Committee
on Nutrition and Asian Development Bank.
<https://www.adb.org/sites/default/les/
publication/27916/what-works-nutrition-
interventions.pdf>. Accessed 20 April 2017.
Angelucci, M. 2015. “Insurance, Investment and
Consumption: The Role of the Extended Family
in the Use of Cash Transfers.” Policy in Focus
11(1): 8–9. <http://www.ipc-undp.org/pub/eng/
PIF31_The_Impact_of_Cash_Transfers_on_Local_
Economies.pdf>. Accessed 20 April 2017.
Angelucci, M., and G. De Giorgi. 2006. “Indirect
Eects of an Aid Program: The Case of Progresa
and Consumption.” IZA Discussion Paper No. 1955.
Bonn: Institute of Labor Economics. <http://ssrn.
com/abstract=881563>. Accessed 20 April 2017.
Behrman, J., and J. Hoddinott. 2001. “An Evaluation
of the Impact of PROGRESA on Child Height.”
FCND Discussion Paper No. 104. Washington,
DC: International Food Policy Research Institute.
Benhassine, N., F. Devoto, E. Duo, P. Dupas,
and V. Pouliquen. 2015. “Turning a Shove into a
Nudge? A ‘Labeled Cash Transfer’ for Education.”
American Economic Journal: Economic Policy 7(3):
86–125. doi:10.1257/pol.20130225.
Bill & Melinda Gates Foundation. 2007. “Milking
the Formula.” The Guardian, 25 May. <https://
www.theguardian.com/world/2007/may/25/
outlook.development>. Accessed 20 April 2017.
Black, R.E., L.H. Allen, Z.A. Bhutta, L.E. Cauleld,
M. de Onis, M. Ezzati, C. Mathers, and J. Rivera.
2008. “Maternal and Child Undernutrition:
Global and Regional Exposures and Health
Consequences.” Lancet 371(9608): 243–60.
doi:10.1016/S0140-6736(07)61690-0.
Borghi, E., M. de Onis, C. Garza, J. Van den
Broeck, E.A. Frongillo, L. Grummer-Strawn,
S. Van Buuren, et al. 2006. “Construction of
the World Health Organization Child Growth
Standards: Selection of Methods for Attained
Growth Curves.” Statistics in Medicine 25(2):
247–65. doi:10.1002/sim.2227.
BRAC. 2013. “An end in sight for ultra-poverty.
Scaling up BRAC’s graduation model for the
poorest.” Brieng Note #1: Ending extreme
poverty. Dhaka: BRAC. <http://www.brac.net/
sites/default/les/BRAC Brieng - TUP.pdf>.
Accessed 20 April 2017.
Brown, L.J., and E. Pollitt. 1996. “Malnutrition,
Poverty and Intellectual Development.” Scientic
American, February. <http://www18.homepage.
villanova.edu/diego.fernandezduque/Teaching/
PhysiologicalPsychology/AllPhysio/Ll11b_
Eating/a12_Malnutrition/Malnutrition.pdf>.
Accessed 20 April 2017.
Egata, G., Y. Berhane, and A. Worku. 2014.
“Predictors of Acute Undernutrition among
Children Aged 6 to 36 Months in East Rural
Ethiopia: A Community Based Nested Case -
Control Study.” BMC Pediatrics 14 (April). BioMed
Central: 91. doi:10.1186/1471-2431-14-91.
Fiszbein, A., and N. Schady. 2009. Conditional
Cash Transfers: Reducing Present and Future
Poverty. Washington, DC: World Bank.
Gartner, L.M., J. Morton, R.A. Lawrence,
A.J. Naylor, D. O’Hare, R.J. Schanler, and
A.I. Eidelman. 2005. “Breastfeeding
and the Use of Human Milk.” Pediatrics 115(2):
496–506. doi:10.1542/peds.2004-2491.
Haseen, F., and M. Sulaiman. 2007. “How
Sustainable Is the Gain in Food Consumption
of the CFPR/TUP Beneciaries?” CFPR Working
Paper No.18. Dhaka: BRAC.
Haushofer, J., and J. Shapiro. 2016. “The Short-
Term Impact of Unconditional Cash Transfers to
the Poor: Experimental Evidence from Kenya.”
The Quarterly Journal of Economics 131(4):
1973–2042. doi:10.1093/qje/qjw025.
Hien, N.N., and S. Kam. 2008. “Nutritional Status
and the Characteristics Related to Malnutrition
in Children Under Five Years of Age in Nghean,
Vietnam.” Journal of Preventive Medicine
and Public Health 41(4): 232. doi:10.3961/
jpmph.2008.41.4.232.
Jalal, C.S.B., and Frongillo, E.A. 2009.
“Poverty Reduction Program for Extreme
Poor in Bangladesh Improves Nutritional
Status of Preschool Children.” Food
and Nutrition Bulletin 34(4): 402–11.
<http://ovidsp.ovid.com/ovidweb.
cgi?T=JS&PAGE=reference&D=emed9&NEWS=
N&AN=70055502>. Accessed 20 April 2017.
TABLE 1: Impact and spill-over eects of the TUP programme
Primary eects Spill-over eects
Weight for height (SD) 0.781*** 0.452***
Wasng (WHZ<-2SD) 0.079*** 0.119*
Height for age (SD) 0.096 0.057
Stunng (HAZ<-2SD) 0.049 0.019
Weight for age (SD) 0.520*** 0.282***
Underweight (WAZ<-2SD) 0.191*** 0.086***
point estimates are nearly twice as large
in magnitude. The proxy for food security,
the ability to secure at least two meals
a day, increases for both ultra-poor
and other-poor households. Lastly, the
availability of sanitary toilets increases
substantially, though the eects on the
prevalence of infectious disease is minimal,
since it was already low at baseline.
A pleasant surprise: What next?
Despite not being a targeted outcome,
the positive eects of the Graduation
Approach on the nutritional status of
children are remarkable. Moreover, the
detection of spill-over eects among
non-participant children makes the overall
impact more notable and interesting. The
fact that the study does not detect eects
on the likelihood of stunting, a long-term
indicator of undernutrition, does raise some
concerns and suggests that the programme
could benet from further ne-tuning to
achieve these higher-order gains.
At present, there is no evidence
disaggregating the exact pathways through
which the programme aects the nutrition
status of children. However, it is likely
that TUP’s integrated nature explains the
impact and thus presents an attractive
alternative to traditional nutrition-specic
interventions, as it addresses multifaceted
vulnerabilities that exacerbate malnutrition.
There is, however, room for improvement
from a nutritional perspective. Evidence
suggests that behavioural components
such as the explicit encouragement of
exclusive breastfeeding have had a marked
positive impact on the participants.
Therefore, appropriate messaging and
consciousness-raising eorts can have
positive outcomes. Implicit encouragement
channelled through ‘demonstration eects’
have similarly and positively aected
46
“ The positive effects
of the Graduation
Approach on the
nutritional status of
children are remarkable.
Manley, J., S. Gitter, and V. Slavchevska. 2012.
“How Eective Are Cash Transfer Programmes
at Improving Nutritional Status? A Rapid
Evidence Assessment of Programmes’ Eects
on Anthropometric Outcomes.” Social Science
Research Unit Working Paper. London: University
of London. <http://www.cashlearning.org/
downloads/q33-cash-transfers-2012manley-
rae.pdf>. Accessed 20 April 2017.
Odunayo, S.I., and A.O. Oyewole. 2006.
“Risk Factors for Malnutrition among
Rural Nigerian Children.” Asia Pacic
Journal of Clinical Nutrition 15(4):
491–95. <http://www.ncbi.nlm.nih.gov/
pubmed/17077064>. Accessed 20 April 2017.
Pelletier, D., M. Shekar, L. Du, and K. Kostermans.
2005. “The Bangladesh Integrated Nutrition
Project Eectiveness and Lessons.” Bangladesh
Development Series No. 8. Dhaka: World
Bank. <http://siteresources.worldbank.org/
NUTRITION/Resources/BNGBINP8.pdf>.
Accessed 20 April 2017.
Rahman, A., S. Chowdhury, and D. Hossain.
2009. “Acute Malnutrition in Bangladeshi
Children: Levels and Determinants.”
Asia-Pacic Journal of Public Health 21(3):
294–302. doi:10.1177/1010539509335399.
Rayhan, M.I., and M.S.H. Khan. 2006.
“Factors Causing Malnutrition among under
Five Children in Bangladesh.” Pakistan Journal
of Nutrition 5(6): 558–62. doi:10.3923/
pjn.2006.558.562.
Raza, W.A, and E. Van de Poel. 2016. “Impact and
Spill-over Eects of an Asset Transfer Program on
Child Malnutrition: Evidence from a Randomized
Control Trial in Bangladesh.” Research
Monograph No. 64. Dhaka: BRAC.
Skouas, E. 2005. “PROGRESA and Its Impacts
on the Welfare of Rural Households in Mexico.”
Research Report 139. Washington, DC:
International Food Policy Research Institute.
<http://www.eldis.org/vle/upload/1/
document/0708/DOC22746.pdf>.
Accessed 20 April 2017.
White, H. 2005. “Comment on Contributions
Regarding the Impact of the Bangladesh
Integrated Nutrition Project.” Health Policy
and Planning 20(6): 408–11; author reply 411.
doi:10.1093/heapol/czi061.
WHO. 2013. Essential Nutrition Actions:
Improving Maternal, Newborn, Infant and
Young Child Health and Nutrition. Geneva:
World Health Organization. <http://apps.who.
int/iris/bitstream/10665/84409/1/978924150
5550_eng.pdf?ua=1>. Accessed 20 April 2017.
World Bank Group and UNICEF. 2016.
“Ending Extreme Poverty: A Focus on
Children.” Briefing Note. New York: World
Bank Group and UNICEF. <https://www.unicef.
org/publications/files/Ending_Extreme_
Poverty_A_Focus_on_Children_Oct_2016.
pdf>. Accessed 20 April 2017.
World Health Organization. 2013. Essential
Nutrition Actions: Improving Maternal,
Newborn, Infant and Young Child Health
and Nutrition. Geneva: World Health
Organization. <http://apps.who.int/iris/
bitstream/10665/84409/1/9789241505550_
eng.pdf?ua=1>. Accessed 20 April 2017.
1. The results presented in this article are
based on a study by Raza and Van de Poel
(2016), published by the Research and
Evaluation Division, BRAC Bangladesh.
2. Independent Evaluation and Research
Cell (IERC), BRAC International.
3. The proportion of 0–18-year-olds living
in extreme poverty outweighs all older age
groups (50.2 per cent versus 48.8 per cent).
4. From a developmental perspective,
malnourishment represents impeded
human capital formation and reduced
future productivity and economic growth
for that society. For example, a study by the
International Food Policy Research Institute
(IFPRI) estimates that the impact of the
nutrition supplements of the Progresa cash
transfer on small children’s anthropometric
development alone could account for a
2.9 per cent increase in lifetime earnings
(Behrman and Hoddinott 2001). Another
study estimated that when children reach
adulthood, they “will have permanently
higher earnings of 8 percent as a result of
the increased years of schooling” facilitated
by Progresa (Skouas 2005).
5. BRAC (2013) considers individuals earning
USD0.60–0.70 per day to be ultra poor.
6. For instance, several participant households
were seen to have substituted lentils with meat
as a source of protein.
7. Anthropometric z-scores (height-for-weight
[WHZ], weight-for-age [WAZ] and height for
age [HAZ]) for children (under 5) are calculated
using the World Health Organization 2006
growth standards (Borghi et al. 2006). The
standard deviation (SD) indicator for an
individual serves as a reference point against
the median of a global reference population.
Children with z-scores below -2 SDs from the
median of the reference population in each
of the categories are considered wasted,
underweight and stunted (ibid.).
8. The inclusion criteria include (three of
ve have to be met): household owns less
than 10 decimals of land; the main source
of income is a female member begging or
working as domestic help; no active male
adult (female household head); school-aged
children working for pay; and no productive
or income-generating assets. The exclusion
criteria, all of which have to be met, include:
no active female member in the household;
micronance participants; and household
members receiving government benets
such as old-age pensions.
9. The training sessions on topics such as
health and nutrition are typically held in
courtyards in front of participant households,
leading to ‘demonstration eects’ of the
treatment messages.
10. This can be accomplished through simple
validated cognitive tests such as the Raven
Test for adults and Early Childhood
Development Index for children.
Photo: Asian Development Bank. Women wait with their children in the Alhaz Johurul Islam City Maternity Centre
at Mirpur, Dhaka, Bangladesh, 2012 <https://goo.gl/svxGw5>.
The International Policy Centre for Inclusive Growth | Policy in Focus 47
Challenges for addressing child poverty
in Malawi through graduation1
Edward Archibald 2
Given the growing emphasis on graduation
worldwide, it is important to consider
the relevance and implications that a
graduation approach might have for
reducing child poverty in Malawi. This
article considers the likely impact, trade-os
and potential unintended consequences
of introducing a fully edged BRAC-style
graduation programme for beneciaries of
the Malawi Social Cash Transfer Programme
(SCTP). It contends that in light of the
circumstances in Malawi, doing so is neither
feasible nor desirable. Among a range of
considerations, perhaps the most troubling
is that some of the poorest families could
be left in a highly fragile and vulnerable
situation—particularly if there are optimistic
assumptions that a BRAC-style programme
can be replicated in Malawi.
Nonetheless, graduation programmes
do pose an important question that is
pertinent to both the context of Malawi
and social policy in general: what does
it mean to ‘exit’ from a social assistance
programme? And what should it mean?
Should ‘exit’ even be part of the language
and approach of social policy? This article
will focus on this particular consideration,
with reference to Malawi’s social protection
system and the perspective of child
well-being. It contends that a graduation
approach which only considers the
poorest and most vulnerable people in
the country is not consistent with the
ideal of a coherent social protection
system. It concludes that decision-makers
and development partners should focus,
rst and foremost, on intergenerational
graduation from poverty—not ‘graduation’
or exit from a cash transfer programme.
Chronic, concentrated
and uid poverty in Malawi
Almost one quarter of Malawians are
chronically monetarily poor, and in recent
years a further one third moved either into
or out of monetary poverty (de la Fuente
et al. 2017). Malawi is one of the poorest
countries in the world, ranking 173 out of
188 on the Human Development Index,
and with an estimated 70 per cent of the
population living on less than USD1.90 per
day (World Bank 2017).3 More than 50 per
cent of Malawians live under the national
monetary poverty line (using the most
recently assessed needs basket that, in 2010,
cost MWK37,000 per person per year);4 25
per cent live in ultra poverty, dened as the
inability to satisfy food needs (de la Fuente
et al. 2017). There is also signicant variation
in levels of poverty and ultra poverty
between districts (ibid.).
Multidimensional poverty is a particularly
important consideration for children.
The assets or income held by a household
may not reect the status or decisions of
children, and indicators used to target
households may not directly aect
children’s ability to develop and thrive.
The 2016 Malawi Child Poverty Report
indicates that the level of multidimensional
child poverty is 63 per cent, whereas 43 per
cent of children live in monetary poverty
(Government of Malawi and UNICEF
2016). This suggests that programmes
which target based on income may
exclude many poor children. Child poverty
comprises several dierent dimensions,
including availability of safe drinking water,
clean sanitation facilities, age-appropriate
education, and a secure home environment
free from abuse.
How cash transfers are reducing poverty
and improving human development
The SCTP is a social protection initiative
that aims to reduce monetary poverty
and improve educational and nutritional
outcomes. The programme provides an
average of around USD10 per month
to the poorest 10 per cent of Malawian
households (targeting criteria are that
the household is both ultra poor and
labour-constrained, and the transfer is
unconditional). Targeting is undertaken
through a combination of community-
based targeting, followed by proxy means
testing (PMT): a community social support
committee creates a list of potential
beneciary households which meet the
eligibility criteria (ultra poor and labour-
constrained). The list is validated by the
community, and then ranked through
PMT before being validated again by the
community. The programme currently
operates in 18 districts of Malawi, and will
scale up to cover all 28 districts in 2017.
The SCTP is protective, and does not have
promotive or transformative objectives. It
does not include any aspects of a graduation
model and, therefore, diers signicantly
from a ‘classical’ approach such as BRAC’s
Targeting the Ultra Poor (TUP) programme.
There is no transfer of assets, provision of
agricultural or other inputs or other forms
of business support, psychosocial support
or regular case management.
The demographic prole of SCTP
beneciary households shows a large
proportion of older household heads,
children and adolescents. There are
relatively limited numbers of prime-age
adults. This is a signicantly dierent
prole from many Latin American
conditional cash transfer programmes,
which tend to comprise younger couples
with young children (Handa et al. 2017).
The SCTP has had substantial impacts
on human development indicators, even
though the transfer value represents
approximately 23 per cent of the baseline
consumption of beneciary families (ibid.,
Annex, Table 1). For example, the school
attendance rate of students aged 14–17 is
13 percentage points higher in households
that received the transfer compared
to those that did not (Abdoulayi et al.
2016). In a country with particularly low
secondary school enrolment, especially
in rural areas, this is a strong achievement.
As shown in Figure 2, the school enrolment
impact of an unconditional cash transfer in
Malawi has similar or even higher impacts
on school enrolment than conditional or
unconditional cash transfers elsewhere,
including in Latin America.
The SCTP has also produced strong results
on income poverty and food security. Over
a three-year period it reduced ultra-poverty
by 15 percentage points (ibid.). Similarly,
the level of worrying about food was seen
48
“ A graduation
approach which only
considers the poorest
and most vulnerable
people in the country
is not consistent with
the ideal of a coherent
social protection system.
“ The school
enrolment impact
of an unconditional
cash transfer in Malawi
has similar or even
higher impacts on
school enrolment
than conditional or
unconditional cash
transfers elsewhere,
including in Latin America.
Source: Handa (2016), based on Abdoulayi et al. (2016).
Source: Handa (2016), based on Abdoulayi et al. (2016).
to decrease by 18 percentage points.
On the other hand, the impacts on health
are muted, and there is no impact on
anthropometric measures of child nutrition
(ibid.). This evidence underscores the
contention that cash alone is insucient,
and there is a need to identify and
strengthen linkages to other social services
and appropriate productive activities.
Given the high level of transient poverty,
and regional variations in ultra-poverty, the
10 per cent cap on the SCTP in each district
has signicant equity implications. Up to one
quarter of all households in some districts
are both ultra poor and labour-constrained
(Juergens and Pellerano 2016a). Yet there
is likely very little variation in consumption
levels within the cohort of all who are ultra-
poor. This means that selection into the
programme can be perhaps dependent on
micro-variations within household assets
or consumption at the time of programme
targeting. Mobility above the 10th percentile
could occur by purchasing a hoe for the
household and shoes for children, rendering
a household ineligible for the SCTP.
What happens when
a beneciary exits the SCTP?
The SCTP is re-targeted every four years.
Beneciaries will exit the programme at
FIGURE 1: SCTP eligible population by age and sex
MaleFemale
(Per cent)
80+
75-79
70-74
65-69
60-64
55-59
50-54
45-49
40-44
35-39
30-34
25-29
20-24
15-19
10-14
5-9
<5
12 12
10 10
8866
44022
FIGURE 2: School enrolment impacts (secondary-school-age children)
of selected cash transfer programmes
20
18
16
14
12
10
8
6
4
2
0
87
15
66
8
99
10
3
Ghana (LEAP)
Ethiopia (PSNP)
Lesotho (CGP)
Kenya (CT-OVC
)
Malawi (SCTP)
South Africa (CSG
)
Zambia (CGP)
Zambia (MCTG)
Zimbabwe (HSCT)
Mexico
Colombia
Ecuador
The International Policy Centre for Inclusive Growth | Policy in Focus 49
“ A beneciary who
exits the SCTP does not
transition automatically
into another form of
social protection.
Source: SCTP Management Information System. Figures provided by Ayala Consulting, 27 March 2017.
this point if they are no longer labour-
constrained or are no longer ranked (by the
community and PMT) as among the poorest
10 per cent. The proportion of beneciaries
that exit the programme on recertication
is relatively small. Table 1 shows the
percentage of SCTP households that
were found to be no longer eligible upon
recertication in 2012 (in the SCTP districts
with the largest caseloads). Less than 5 per
cent were no longer labour-constrained,
and less than 2 per cent were no longer
ultra poor. These gures also highlight
that the design of the SCTP is protective
and not promotive. After four years in the
programme, SCTP households were eating
more and better food, and were sending
children to school. Yet, by and large, they
were still among the most impoverished
members of their communities.
Although recertication creates some
opportunities for a small number of new
beneciaries, available evidence suggests
that the benets of the SCTP are not
always maintained for households that
exit the programme. Recent qualitative
research of households that have exited the
programme through recertication shows
that not a single former SCTP household
surveyed was now engaged in a sustainable
livelihood (Scott, Harman, and Chinsinga
2016).5 Perhaps most tragically, many
former SCTP beneciaries reported that, on
exiting the programme, they were forced to
withdraw children from secondary school
due to an inability to aord the fees.
A beneciary who exits the SCTP
does not transition automatically into
another form of social protection. There
is no laddered system whereby SCTP
beneciaries graduate into the next
stage of the social protection system
(as is the case in Chile). Graduates of the
SCTP have instead been left to their own
devices and informal support networks.
Although there are no coherent options
for mobility between social assistance
programmes, there are some positive
signs of reform in this regard. The recently
revised Public Works Programme (PWP)
is intended to target the 15 per cent next
most vulnerable households that are not
enrolled in the SCTP. This should include
former beneciaries of the SCTP who have
exited the programme. Yet there is a clear
and currently unmet need for coherence
between the SCTP and the PWP. For
instance, an SCTP household may have
progressed closer to the ultra-poverty line
over several years of receiving transfers.
This household may have exited SCTP
during the recertication process yet still
be labour-constrained. SCTP can of course
inuence consumption levels, but there is
no evidence that it inuences indicators
related to labour constraints. As the
demographic pyramid in Figure 1 shows,
the SCTP is characterised by high numbers
of elderly people, children and adolescents.
Would a former SCTP household now be
eligible for PWP, a programme that requires
work in return for cash transfers? The
design of PWP and the broader National
Social Support Policy are silent on this issue.
Indeed, the process and design of how
beneciaries exit from the SCTP is
unclear, and it seems probable that some
SCTP ‘graduates’ may slip through the
gaps—particularly if they remain labour-
constrained and, therefore, theoretically
ineligible for public works. In most districts,
the programme cycles and beneciary
management of the SCTP and the PWP are
not yet aligned. Furthermore, beneciaries
enrolled in the revised PWP are expected
to be registered as beneciaries for a full
three years, implying that a beneciary can
neither join nor leave the PWP over that
time-frame. As with the SCTP, it is a lengthy
wait for potential PWP beneciaries in the
event of exclusion errors.
The risks of a graduation agenda
for the most vulnerable people
The attention of decision-makers in Malawi
on graduation has been primarily limited
to emphasising the need for beneciaries
to graduate from the SCTP. There has been
virtually no focus on whether beneciaries
should or could graduate from the other
major social support programmes,
including PWP (15 per cent of the
population) or the fertiliser subsidy, which
in recent years has provided benets to
about a quarter of all Malawian households
and has been shown to not target poor
people (Kilic, Whitney, and Winters 2013).
These two programmes represent about 1p er
cent and 4.5 per cent of gross domestic
product (GDP), respectively, whereas the
SCTP currently costs the equivalent of
about 0.5 per cent of GDP (Juergens and
Pellerano 2016b). Focusing on one layer of
the social protection system, particularly
the poorest and most vulnerable people in
the country, is not consistent with the ideal
of a coherent social protection system.
Graduation as a concept can be misused,
with potentially deleterious impacts.
Important nuances and complexities are
lost by the use of a word which implies
nality. The word ‘graduation’ is commonly
applied as an objective for households
beneting from a specic social protection
programme. Yet ‘graduation’ usually refers
to something fully completed or attained,
and never to be returned to. However,
the prospects of those living in poverty,
TABLE 1: Percentage of SCTP households no longer eligible for the programme
upon recertication in 2012, selected districts
District Households no longer
labour-constrained (%)
Households no longer
ultra poor (%)
Households no longer labour-
constrained and ultra poor (%)
Mchinji 2.7 0.06 0.26
Salima 1.99 0.07 0
Chipa 5.1 1.7 1.1
Machinga 4.8 0.3 0.3
Mangochi 3.7 0.26 0.26
Phalombe 4.9 1.0 0.67
50
“ The broader social
protection system and
network of social services
in Malawi are insufciently
developed to support a
graduation programme.
Photo: Tim Chesney/freeimages. Kid selling chickens in Malawi <https://goo.gl/v8T9ly>.
or close to the poverty line, are often
highly uid. Cross-country evidence shows
that poverty is dynamic and uctuating,
with many people falling in and out of
poverty (Subbarao 2002). Many people
are ‘sometimes poor’ rather than ‘always
poor’ over a certain period of time. The
movement can be in either direction:
while some people can be poor and
escape poverty, large numbers can also
be non-poor and then fall into poverty for
many dierent reasons. This highlights the
risks of assuming that SCTP beneciaries
will never return to poverty after exiting
the programme. In fact, it underscores
the importance of social protection
throughout the life cycle: a catastrophic
health event can plunge all but the
wealthiest citizens in the world into a
poverty trap which they may not
be able to overcome.
Another risk relates to the political economy
of ‘graduation’. The results of graduation
programmes in some countries have
been shown to be robust (although this
view is not universally shared). More
signicantly, expectations can be raised
and may unduly skew the discourse on
prioritisation within social protection
programmes and policies. The prominent
focus on the successes of BRAC’s graduation
programmes, for instance, can have political
economy implications elsewhere where
replicability is assumed all too easily. The
objective of sharing lessons across borders
is intrinsically good. However, the results
of programmes such as BRAC’s can lead
decision-makers to conclude that social
protection programmes in very dierent
contexts should lead to the same outcomes.
This can result in policymakers seeing
graduation programmes as a quick x to
rapidly reduce extreme poverty. In such
circumstances, incomplete information
could indirectly harm the lives of highly
vulnerable families. Access to a graduation
programme could lead to exclusion from
social protection, and may render a family
unable to mitigate their risks. The discourse
in Malawi has often focused on how many
SCTP beneciaries can or should graduate
from the programme—despite the SCTP
being designed without any promotive or
transformative components.
The broader enabling environment,
particularly the real economy, also poses
signicant challenges to implementing a
graduation approach. The macroeconomic
environment in Malawi is particularly
weak. Ination has been at 20–25 per
cent for the past two years (Record 2016).
Unemployment and underemployment
are widespread. The country has been
hit by three successive shocks in two
years: oods in January 2015, followed
by a drought, followed in turn by an
El Niño-induced drought. The cost of
humanitarian responses to these shocks
has been substantial. Economic growth
has also been weak, with rates of less than
3 per cent in 2015 and 2016 (World Bank
2017). The economy has been in crisis
for several years, and although there are
prospects of recovery, sadly Malawi is not
yet a dynamic economy on the cusp of
an exciting transformation.
The International Policy Centre for Inclusive Growth | Policy in Focus 51
The broader social protection system
and network of social services in Malawi
are insuciently developed to support
a graduation programme. Government
implementation of a graduation
programme would require increased
capacity on multiple levels, and the supply-
side constraints in Malawi are enormous.
For example, many relevant services have
highly limited coverage. There has also
been a freeze on public service recruitment
for several years due to inadequate
budget and ballooning costs, and unlled
government positions are widespread at
both district and national levels.
If government resources cannot support a
graduation programme for the SCTP, should
decision-makers look to non-governmental
organisations (NGOs)? That depends, to
some extent, on objectives and resources.
The cost of hands-on management and
implementation at village level can be
substantial, due to geographic isolation of
many communities, poor-quality transport
infrastructure, and a prevailing culture of
‘allowances’ for local travel. NGOs are capable
of implementing small-scale graduation
programmes. Nevertheless, the prospects
for the vast majority of SCTP beneciaries
would be limited; implementation at
scale—either through government or non-
government systems—is highly improbable
in the short to medium term.
Ensuring that the SCTP children
of today are not the SCTP heads
of households of tomorrow
The dividends would be greater if there
were increased investment in the SCTP
rather than a BRAC-style graduation
approach. Although a graduation
approach based on BRAC could potentially
succeed in Malawi with parallel NGO
service delivery at scale, it is unlikely to
be cost-eective. The SCTP is proven to
be delivering substantial returns in both
social and economic development. It is
the most evaluated, monitored and well-
coordinated social support programme
in Malawi. At this stage in Malawi’s
development, it is likely to be more cost-
eective to scale up the SCTP to a greater
proportion of the population and provide
a higher level of transfer, which is indexed
to ination and/or accounts for seasonal
variations. In a time of increasingly limited
resources, and with attention focused on
economy and eciency, such courageous
actions may not be popular but are likely
to deliver the best value for money.
In addition, there should be a
strengthening of the recently developed
system to create linkages and referrals
for SCTP beneciaries to social services.
This nascent system entails elements of a
classical graduation approach, including
case management, although beneciaries
will only be visited twice a year, and many
services are lacking in quantity or quality.
Nonetheless, the referral system shows
good potential to provide support to SCTP
beneciaries beyond cash transfers.
Decision-makers should focus more
attention on intergenerational graduation
from poverty in Malawi, rather than
graduation from specic social assistance
programmes. Given the economic malaise,
the concentration of ultra-poverty, the
lack of government services and the poor
coherence within the social protection
system, policymakers and donors should
instead give serious and robust attention
to how the entire social protection system
can help prevent the intergenerational
transmission of poverty.
Graduation approaches should, therefore,
view children in SCTP households as direct
beneciaries of the SCTP, and focus on their
graduation from the programme rather
than solely the household head—who may
well be elderly and/or have a disability or
a chronic illness. The graduation of SCTP
children is much more likely than that
of their caregivers. Providing too little
money to too few people, and exiting
them from a programme before they are
ready to be independent, could hinder
rather than help eorts to rid Malawi of
extreme poverty. The rst responsibility
of policymakers and development
partners is to ensure that the SCTP
children of today are not the SCTP heads
of households of tomorrow.
Abdoulayi, S. et al. 2015. Malawi Social Cash
Transfer Program Midline Evaluation Report.
Chapel Hill, NC: The Transfer Project,
University of North Carolina at Chapel Hill.
Abdoulayi, S. et al. 216. Malawi Social Cash
Transfer Program Endline Evaluation Report.
Chapel Hill, NC: The Transfer Project,
University of North Carolina at Chapel Hill.
De la Fuente, A. et al. 217, forthcoming. Poverty
Assessment, Malawi. Washington, DC: World Bank.
Government of Malawi and UNICEF. 216.
Child Poverty in Malawi. Lilongwe:
Government of Malawi and UNICEF.
Handa, S. 215. “Malawi Social Cash Transfer
Program Midline Evaluation Report.” Presentation
made at dissemination workshop of Malawi SCTP
Midline Survey, Lilongwe, May 215.
Handa, S., B. Davis, S. Daidone, A. Peterman,
T. Palermo, A. Pereira, and J. Yablonski. 217,
forthcoming. Myth busting? Confronting Six
Common Perceptions about Unconditional Cash
Transfers as a Poverty Reduction Strategy in Africa.
Chapel Hill, NC: The Transfer Project, University
of North Carolina at Chapel Hill.
Kilic, T., E. Whitney, and P. Winters. 213.
“Decentralized Beneciary Targeting in Large-Scale
Development Programs: Insights from the Malawi
Farm Input Subsidy Program.” Policy Research
Working Paper 6713. Washington, DC: World Bank.
Juergens, F., and L. Pellerano. 216a. “SCT
eligibility analysis.” Unpublished. Lilongwe:
International Labour Organization.
Juergens, F., and L. Pellerano. 216b. Social
Protection in Malawi: Summary of the Assessment
Based National Dialogue Report. Addis Ababa:
International Labour Organization. <http://
www.ilo.org/addisababa/information-resources/
publications/WCMS_493326/lang--en/index.
htm>. Accessed 24 April 217.
Record, R., P. Kandoole, E. Chilima, S.
Kalemba, and T. von Carnap. 216.Emerging
Stronger. Malawi Economic Monitor,
October 216. Lilongwe: World Bank.
<https://openknowledge.worldbank.org/
handle/1986/25339>. Accessed 24 April 217.
Scott, L., L. Harman, and B. Chinsinga. 216.
“Options for a Graduation Strategy from
Malawi’s Social Cash Transfer Programme
(SCTP).” Unpublished. London: Overseas
Development Institute.
Subbarao, K. 22. “Risk and Vulnerability
Assessments: Concepts and Methods.”
Paper presented at the ADB and InterAmerican
Development Bank Workshop on Social Protection
for the Poor, 21–25 October 22, Manila.
Thome, K., E. Taylor, M. Tsoka, P. Mvula, B. Davis,
and S. Handa. 215. Local Economy-wide Impact
Evaluation of Malawi’s Social Cash Transfer
Programme. Rome: Food and Agriculture
Organization of the United Nations. <www.fao.
org/3/a-i4458e.pdf>. Accessed 24 April 2017.
World Bank. 2017. “Poverty headcount ratio
at $1.90 a day (2011 PPP) (% of population).”
World Bank website. <http://data.worldbank.
org/indicator/SI.POV.DDAY?locations=MW>.
Accessed 24 April 2017.
1. The opinions expressed in this article are
those of the author and do notnecessarily
reect the policies or views of UNICEF.
2. Chief of Social Policy, UNICEF Malawi.
3. Figures cited are from 2010, which are the
most recently available.
4. Approximately USD50.95 as of publication date.
5. This nding is consistent with the high
levels of both chronic and ultra poverty, and
community norms relating to sharing the
benets of social services.
52
Digital inclusion for the ultra poor:
the Graduation Approach
Tatiana Rincón 1
One of the dening characteristics of
the Graduation Approach is its focus
on those living in extreme poverty,
frequently referred to as the ‘ultra poor’.
These individuals and families often
live in isolated areas where, among
other challenges, access to markets and
opportunities, nancial inclusion, secure
jobs and businesses, and education for
their children are all dicult. Moreover,
these communities are also often
excluded from the technological and
digital revolutions reshaping the world.
For instance, nearly 2 billion people
worldwide do not own a mobile phone,
and nearly 60 per cent of the world’s
population have no access to the
internet (Bauer 2016).
At Fundación Capital, our graduation
programmes have, from their inception,
worked directly with governments to
promote the development and use of
digital solutions to support both the
scaling-up of programmes and digital
inclusion. Together, these partnerships
and tools, such as e-learning platforms
for coaches2 and educational tablet
and smartphone apps for participants,
represent a key way to scale up while
keeping quality high and costs low.
Additionally, by introducing digital tools,
we are able to smartly embed monitoring
and evaluation tools in our systems,
which permits further progressive and
timely adjustments of the programme,
including channels for reporting and
settling grievances, among other
advantages. Moreover, these tools
also create the opportunity to connect
and empower vulnerable communities
and advance their economic, nancial
and digital inclusion.
Digital solutions for training
There are two sides to any training
initiative: the participant and the coach.
On the one hand, for participants, training
can mean long commutes to reach a
workshop location, high opportunity
costs such as lost wages, and money
spent on child care or transportation, costs
accumulated before even reaching the
training, where the coach may not show
up or may be poorly trained themselves.
On the other hand, coaches might receive
little support, low wages and insucient
capacitation to eectively perform their
role. For our graduation work at Fundación
Capital, we have decided to tackle these
issues from both sides, integrating digital
solutions to help alleviate some of these
limitations and eliminate bottlenecks.
E-learning: training the coaches
In most social projects, the success of
participants directly depends on the
performance of the sta working with
them. The best coaches are often the peers
of programme participants—individuals
who live in the same communities, have
experienced similar situations, speak the
same language and cultural code and are
able to ‘teach through example’.3 Peer-to-
peer learning reinforces the idea that it
is possible to succeed despite diculties
and challenges. However, peers may
lack some of the technical skills needed
tocoachother people, which is where
technology can come in.
To respond to the needs of both the
coaches and the programmes—from how
to improve and monitor performance, to
how to reduce costs without impacting
the salaries of the coaches—we have
developed an online course featuring all
the specic components of the project.
Complementing the digital learning
platform is a human component, seeing
that participants receive the support
of a virtual tutor who motivates them,
provides feedback and documents their
best contributions. Additionally, for critical
topics, coaches receive face-to-face
training in addition to the virtual training.
Coaches can study at any time, although
they must complete each module by
a specied deadline, usually one week
before their eld visit. The course can
be accessed via any digital device,
such as a computer, tablet or smartphone.
It is important to mention that all
programme sta are literate,
and many already own smartphones or
other digital devices, thereby facilitating
the use of online courses. Our courses are
designed specically with their needs
and education levels in mind.
The coaches must study the content of
and pass exams on each module to both
demonstrate and reinforce the lessons and
track their understanding, and because
the coaches’ contracts are linked to their
performance in the course. Following
the grading of the exams, if any coach
continues to have diculties, a traditional
class is arranged in the eld to address any
remaining questions or doubts. Through
the course, it is possible to verify whether
sta have the necessary skills before
theyproceed to coachthe families.
APPtitude: training the participants
We have also developed and introduced
a tablet- and smartphone-based
application for graduation programme
participants, ‘APPtitude’.4 It covers themes
such as developing a microbusiness,
nancial education and personal
development, and includes educational
games and inspirational videos featuring
the stories of successful peers. By the
end of 2016, approximately 27,500 direct
participants of our graduation projects in
Colombia, Mexico and Paraguay had been
trained using APPtitude.
By using the application, participants
train themselves in parallel with the
implementation of their productive
activity. At the beginning of each module,
participants enter their ID number, and
the system automatically conrms their
GPS coordinates, allowing project sta
to track their progress online. At the
end of each module, participants are
presented with a ‘rule of thumb’—the
most important lesson they should take
away from that module. Each module
then ends with a basic survey, which helps
determine whether participants believe
the module was useful to them and easy
to understand, and allows for continued
improvement of the application based
on real-time user feedback.
The International Policy Centre for Inclusive Growth | Policy in Focus 53
“ The use of tablet- and
smartphone-based
applications for training
both staff and
participants has
many advantages.
Photo: Screenshot of the APPtitude interface, demo video: <https://www.youtube.com/watch?v=Ocl7xdIyGiA>.
Integrated solutions
In the rst graduation projects where
we worked with governments, such as
‘Producing for my Future’ and ‘Transforming
my Future’ in Colombia, we found that
the use of tablet- and smartphone-based
applications for training both sta and
participants has many advantages. It
guarantees the quality of the content while
minimising dependence on the personal
training abilities of a given coach. It reduces
costs for the programme as well as for the
participants, who also have the exibility
to choose when to work on their training
module. It contributes to closing the ‘digital
gap’ and empowers families who have never
had access to an application before. In this
sense, having a multi-platform application
is key: families that have a smartphone can
download content to their own device, while
tablets that are circulated by coaches reach
families that do not own a device or that live
in areas without an internet connection.
The use of this type of application also
enables participants to learn at their
own pace, as modules can be repeated
as often as needed until content is
internalised. Additionally, it reaches
other household members, including
children, creating a spill-over eect.
For instance, based on the 27,500 direct
participants reached since the beginning
of this approach, we estimate that it has
indirectly reached more than 144,000
people, considering the average number
of family members that are also impacted
by the programme and the shared
learning that we have seen in our work.
With both APPtitude and the e-learning
platform, economies of scale take eect:
the more users the programme reaches,
the lower the cost per person becomes,
as xed costs and investments in
software development, coordination sta,
communication and data management fall,
and market hardware costs decrease over
time. Both applications also have a robust
monitoring backend: a metric system
that allows tracking and monitoring of
the learning activity. As participants and
coaches use the app, the information
is sent in real time and uploaded to a
database. Coordinating teams can then
access the information using a web
browser and can view reports, alarms and
key performance indicators, including
progress reports for each participant
and coach, the average time spent on
each module, and the location of each
training session. When a participant is
having diculties, coordinating teams
can instruct coaches to pay extra attention
and provide additional support to ensure
that no one is left behind. Currently, these
digital solutions, with country-specic
adaptations, are being implemented in
Colombia, Paraguay and Mexico, and will
begin in Honduras and Angola in 2017.
Digital solutions for savings
When considering the challenges facing
graduation programmes and participants,
we were aware that digital solutions for
training, made accessible on tablets and
smartphones, were just one piece of the
puzzle. By putting knowledge directly into
the hands of the poorest families we are
54
reducing the digital gap, but additional
eorts are needed to tackle other
bottlenecks such as nancial exclusion.
At Fundación Capital, we have worked
throughout Latin America, as well as
recently in Africa, to encourage the
transition from cash payments to
e-payments for the recipients of cash
transfer programmes, since e-payments
tend to be more secure, are better for rural
communities located far from banks and
represent a rst step towards nancial
inclusion. As an organisation, nancial
inclusion and education5 are at the heart
of our work. Therefore, in our graduation
programmes, we promote not only strong
nancial education but also nancial
inclusion. Indeed, nancial inclusion
allows practical nancial education
through ‘learning by doing’.
Banking has become increasingly more
accessible to clients at the lower end of
the income distribution, yet it remains far
from meeting the needs of the poorest
people. Today, an estimated 2 billion
adults worldwide still do not have a basic
bank account (World Bank 2016). Banking
locations are often far away, account
rules and norms confusing, and branches
intimidating. Additionally, misplaced
negative beliefs about accounts—such as
that the bank will take all of one’s money
through fees, or the government will kick
someone out of a given cash transfer
programme if they see they are not using
all of their transfer money—persist, and lead
to transfer recipients often withdrawing all
of their money as soon as they receive it.
Between the challenges of physical banking
and the need for more user-friendly options,
the need for products designed for the
poorest citizens is clear.
On the other hand, there is the emerging
trend of mobile phone use across all levels
of society. It is estimated that currently
almost 2 billion people have access to a
mobile phone but not to a bank (Carter
2015). With even a basic mobile phone,
mobile banking becomes possible.
“Mobile banking allows the global poor
to have access to nancial services and
transactions that most of us take for
granted. For example, employers can
transfer money to employees, allowing
them to safely store and save their income.
It also allows people to apply
for microloans or send money seamlessly
to friends and family in need” (ibid.).
Indeed, some electronic wallets have
been created by telecommunications
companies, making it possible to save
some money, but their use is still limited.
Recently we worked with Tigo, a leading
telecommunications company in Latin
America, to design the ‘Microsavings with a
purpose’ product, linked to the participant’s
mobile phone. During the rst semester
of 2017 this product will be tested with
participants of Paraguay’s graduation
project, Sembrando Oportunidades Familia
por Familia,6 itself the result of collaborative
design between Fundación Capital and the
Government of Paraguay.
In our work with the poorest families,
we have learned that people save more
when they have a clear objective in mind:
‘I want to buy a bicycle’, ‘I want to save to
improve/renovate my kitchen’, ‘I want to
save to buy a fridge to expand my business’
are the common statements of intent we
encounter daily in the eld. Using the
principles of human-centred design and
behavioural economics, we were able
to design a product that could be used
by regular participants of graduation
programmes to save towards their own
specic goals. This solution—as all of our
solutions—was created together with
the end-user and iterated based on their
feedback. The key is that the technology
is adapted to the needs of the user, and
not the other way around.
The electronic wallet is only the rst step
for participants of ‘Microsavings with
a purpose’; all of the savings are being
secured in savings accounts with a formal
bank, and the microsavings programme
with Tigo is the channel that allows for this.
We are eager to learn from this pilot,
and hopefully expand this experience to
participants of graduation programmes in
other countries around the world. We are
constantly working not only to learn from
the work within countries but between
them, which has motivated us to establish a
cross-country evaluation platform supported
by the International Development Research
Centre (IDRC) and led by Colombia’s
Universidad de Los Andes. This platform is
currently conducting process, results and
impact evaluations in the countries where
we are implementing our graduation
programmes, and through these evaluations
we will be able to learn more about the
results and impacts of our initiatives.
How to eectively use technology to
improve the well-being, and increase the
digital inclusion, of some of the poorest
and most vulnerable members of society
is among our greatest challenges. Poverty is
a complex problem that cannot be solved
with just an app or the implementation of
an electronic wallet, but technology can
play a key role in promoting high-quality
solutions, greater access to services and,
ultimately, more opportunities.
Bauer, Roxanne. 2016. “Media (R)evolutions:
Dramatic spread of internet, mobile phones
not enough to get women online.” World
Bank website. <https://blogs.worldbank.org/
publicsphere/media-revolutions-dramatic-
spread-internet-mobile-phones-not-enough-
get-women-online>. Accessed 10 May 2017.
Carter, Joe. 2015. “How cell phones are helping
to End Extreme Poverty.” Acton Institute website.
<http://blog.acton.org/archives/82494-how-cell-
phones-are-helping-to-end-extreme-poverty.
html>. Accessed 10 May 2017.
World Bank. 2016. Financial Inclusion Overview.
World Bank website. <http://www.worldbank.
org/en/topic/nancialinclusion/overview>.
Accessed 10 May 2017.
1. Director, Social and Livelihood Promotion,
Fundación Capital.
2. Coaches are programme support sta who
help participants in their personal development
throughout the programme. Normally, they
are peers of programme participants, coming
from the same region and having a similar
socio-economic background, but with higher
educational levels.
3. To learn more about our work on this
topic, please see <http://nextbillion.net/
peermentorships-helping-colombiawrite-
its-next-chapter/>.
4. For more information, see <https://www.
youtube.com/watch?v=Ocl7xdIyGiA>.
5. There is an important dierence between
nancial education (sometimes called ‘nancial
literacy’ in the global North) and nancial
inclusion. While the rst concept mainly refers
to knowledge, the second concept also entails
having nancial capabilities (knowledge put
into practice), which increases the probability
of positive behavioural change. There is also
a dierence between nancial education and
nancial literacy: while the former refers to
education for people at the base of the income
pyramid (therefore, most of them illiterate),
nancial education applies to everybody.
6. See <http://fundacioncapital.org/en/
economic-citizenship/graduation-
program/paraguay/>.
The International Policy Centre for Inclusive Growth | Policy in Focus 55
Caveat emptor: the Graduation Approach,
electronic payments and the potential
pitfalls of nancial inclusion
Paulo L. dos Santos1 and
Ingrid Harvold Kvangraven1
The Graduation Approach to poverty
reduction is inextricably bound up with
programmes promoting nancial inclusion.
Proponents for the approach see it guiding
a series of interventions that encourage
poor households to ‘graduate’ into
‘mainstream development programmes’
which are centred on the provision of
credit and other nancial services (BRAC
2014). Indeed, the approach has been
presented as a way to address the needs of
those “too poor for micronance services”
(UNHCR 2014). The presumption is that
the development and poverty reduction
needs of ‘graduates’ will be well served by
nancial inclusion initiatives.
Recent interventions by inuential
development actors have sought to
deepen the policy connection between
graduation approaches and nancial
inclusion initiatives, including through the
use of electronic payments systems (EPSs).
The Ford Foundation and the Consultative
Group to Assist the Poor (CGAP) recently
called for ‘Graduation 2.0’ approaches that
would establish and promote the “impact
of nancial services in the graduation
approach and how Graduation Approaches
are or could be synergistic with national
nancial inclusion initiatives, social safety
nets and large-scale digitization of social
transfers” (CGAP 2017).
This call parallels the eorts of the Better
than Cash Alliance (BtCA) to promote the
diusion of EPSs in developing countries.
Funded by Visa, MasterCard, Citigroup, the
Gates Foundation and USAID, the BtCA
advocates public investment in EPSs to
facilitate their use to expand the reach of
private suppliers of services in payments,
savings, insurance and loans. This includes
calls on governments to use EPSs to make
cash transfer and broader social programme
payments to encourage the nancial
inclusion of some households even before
their graduation from extreme poverty.
The promotion of nancial inclusion as
a central tool in graduation and broader
poverty reduction eorts rely on two widely
cited propositions: rst, the view that greater
access to all types of nancial services can
universally (and in itself) contribute to
poverty reduction and broader economic
development; and, second, the contention
that market-based provision of credit,
savings, insurance and payment services
by prot-driven agents is desirable over
provision by state or mandate-driven rms
such as postal or cooperative banks.
A recent study by dos Santos and
Harvold Kvangraven (2017) has drawn
on an extensive review of empirical
and theoretical work to oer a critical
assessment of these propositions and
of broader policy calls for the expansion
of EPSs to expand nancial services to
poor households. Considering the potential
impacts of dierent nancial services
separately, the study raises serious concerns
about eorts to expand volumes of lending
to poor households, communities and
regions. It also points to the important
reductions EPSs can make to the costs
of managing, safeguarding and sending
money, while underscoring the vital role
of government interventions in ensuring
quality EPSs are universally available at
low cost. Those arguments are directly
relevant to debates concerning graduation
approaches and their relationship to
nancial inclusion initiatives and EPSs.
Why credit often hurts poor people?
Inuential organisations such as the BtCA
have argued that EPSs can make a positive
contribution to development by enabling,
Photo: BRAC. Woman selling produce, 2014.
“ Recent interventions by
inuential development
actors have sought
to deepen the policy
connection between
graduation approaches
and nancial inclusion
initiatives, including
through the use
of electronic
payments systems.
56
among other things, expansions of credit
via electronic banks.2 However, recent
research has refuted the long-accepted
proposition that enabling greater lending
to poor people invariably reduces poverty
and promotes development.3 Inuential
and widely cited studies purporting to
establish the positive eects of microcredit
initiatives targeting poor borrowers in
developing countries have been shown
to suer from deep methodological
aws, and to report ndings that cannot
be replicated.4 Meanwhile, a growing
body of evidence has documented how
those initiatives primarily encourage
consumption credit—often at annual
rates of interest exceeding 100 per cent
and under predatory terms—and not
new types of productive enterprise.5
These perverse outcomes are not
accidental. They reflect the realities of
underdeveloped areas, where shortages
of skills, infrastructure and markets,
among others, create formidable
obstacles to the development of new,
higher-value-added enterprises. Without
broader policy interventions to improve
prospects for such enterprises, banks will
not generally find it profitable to finance
them. Low incomes, small markets and
other constraints ensure that even where
financial inclusion initiatives support
the development of new enterprises,
those firms typically displace previously
existing ones, yielding no appreciable
net benefits to the areas in question
(Bateman 2010).
In contrast, lenders are often able
to develop a protable business in
consumption lending in those areas,
where many people struggle to meet
basic consumption needs. Living with few
choices and many privations, the would-
be beneciaries of microcredit are often
driven into consumption loans that are
usurious or predatory. In contrast to loans
supporting new productive enterprises,
these loans do not contribute to increases
in borrowers’ income. However, they
do impose interest payment burdens,
through which lenders can appropriate
entire portions of borrowers’ meagre
incomes.6 The long-term eects of this
type of lending on development and
poverty reduction are clearly negative.
Yet in many settings where nancial
inclusion initiatives are pursued, the
comparatively higher protability of this
lending ensures that it is precisely the
type of credit lenders nd most attractive.
The geographical distance of electronic
banks from their borrowers in low-
income areas makes them even less
likely than traditional micronance
institutions to engage in lending to
new productive enterprises. This lending
requires detailed knowledge of all
aspects of the operations of the small
borrowing enterprises. Proponents
for traditional micronance initiatives
explicitly recognised this diculty and
pointed to existing social connections
among borrowers and between them
and local lenders as a basis for addressing
it. It is not clear how electronic banks
would overcome such thorny problems
to support the development of new
productive enterprises.
The likely focus of these lenders on
oering credit to poor borrowers qua
consumers is also clear in the BtCA’s
proposals, which tout the prospect that
electronic banks in low-income areas
will be able to make loans collateralised
by poor households’ income from cash
transfers or foreign remittances.7 This is
a particularly problematic idea. It would
ensure that fractions of precious money
ows aimed at improving the conditions
of some of the poorest households on the
planet are diverted into interest payments
on consumption loans. Policymakers
would be well advised to resist calls
for the expansion of such practices.
Delivering quality payments
and savings services for all
While the prospect of EPSs being used
to ramp up lending to poor people raises
serious concerns, those systems can reduce
the costs of sending and safeguarding
money. They can also promote savings.
Yet ensuring that these benets are widely
shared requires careful government
intervention and policy.
Markets for payment services are prone
to uncompetitive behaviour. Payment
systems are networks. Like languages or
computer operating systems, their value to
any given user depends on the number of
total users. Consumers and merchants may
be reluctant to switch to better or cheaper
competing suppliers simply because fewer
people use them, giving dominant suppliers
of payment services the ability to abuse
their position.8 Both Visa and MasterCard
have been accused by regulators and
merchants in the USA and Europe of
abusing their dominance of point-of-sale
(POS) electronic payment systems to
charge uncompetitive fees.9 This raises
concerns about uncompetitive behaviour
in the signicantly weaker regulatory
environments of low-income economies.
It is notable that the technologies the BtCA
promotes still include the POS systems
dominated by Visa and MasterCard,10 even
though that technology is widely seen as
having been rendered obsolete by mobile
and computer-based networks.
The uncompetitive tendencies inherent in
markets for payment services ensure that
there is an important role for governments
to play. Depending on national conditions
and capacities, there are at least three
approaches they can follow to ensure that
quality EPSs are available at low cost: rst,
governments may directly regulate the
actions and prices of private suppliers;
second, they may support the operation of
a mandate-driven supplier that eectively
sets the market price for payment services,
leaving room for competitive private
suppliers to operate protably; and, third,
governments may simply have central
banks make electronic deposits widely
available, as a new form of cash.
The resulting low-cost EPSs can also
encourage savings. In this vein, they
may function like postal banks, enabling
modest wealth accumulation and asset
acquisition by poor households and
helping boost their resilience. These
savings may also be pooled and used
to support development projects.
Development agencies can provide
valuable assistance by helping governments
develop institutional capacities to regulate
or run such systems, to identify valuable
development projects that can be nanced
with electronic savings, and to design
contracts allowing this nancing to take
place at minimal to no risk to savers.
Micro-level innovation is no substitute
for systemic development policy
Well-regulated EPSs that oer low-cost
payment services, encourage savings
and are prevented from extending
The International Policy Centre for Inclusive Growth | Policy in Focus 57
“ Living with few choices
and many privations, the
would-be beneciaries
of microcredit are often
driven into consumption
loans that are
usurious or predatory.
BRAC/Alison Wright. BRAC micronace project, Tanzania, 2014.
loans to poor people can contribute to
development and poverty reduction.
Yet such contributions will be modest.
The arguments promoting EPSs and
nancial inclusion embody a third,
widely cited proposition: that small-
scale technological or institutional
innovations can by themselves contribute
to signicant reductions in poverty and
to the broader economic development
of low-income communities, areas or
economies. This tendency to see micro-
level innovations as development or
poverty-reduction panaceas downplays
the stubborn structural and systemic
obstacles and deciencies that dene
underdevelopment, and the need for
joined-up national industrialisation
strategies and broader social policies to
overcome them.11 Without such strategies
and policies, even promising innovations
and well-intended initiatives can yield
unintended or even perverse results.
This understanding should inform
policymakers and development
practitioners to consider the likely impacts
of all ‘graduation’ interventions aiming to
make permanent, sustainable reductions
in the depth and spread of poverty in
developing economies.
Amsden, Alice H. 1992. Asia’s Next Giant: South
Korea and Late Industrialization. Oxford: Oxford
University Press.
Banerjee, Abhijit, Esther Duo, Rachel
Glennerster, and Cynthia Kinnan. 2013.
“The Miracle of Micronance? Evidence from
a Randomized Evaluation.” MIT Department of
Economics Working Paper 13-09. Cambridge,
MA: Massachusetts Institute of Technology.
Bateman, Milford. 2010, Why Doesn’t
Micronance Work? – The Destructive Rise
of Neoliberalism. London: Zed Books.
BRAC. 2014. Challenging the Frontiers of Poverty
Reduction Targeting the Ultra Poor (CFPR-TUP).
Dhaka: BRAC. <http://www.brac.net/sites/default/
les/tup_June_2014.pdf>. Accessed 25 May 2017.
BtCA, the Gates Foundation and World Bank.
2014. The Opportunities of Digitizing Payments.
Washington, DC: World Bank. <http://www.
gp.org/publications/opportunities-digitizing-
payments>. Accessed 25 May 2017.
CGAP. 2017. “Graduation into Sustainable
Livelihoods.” CGAP Micronance Gateway
website. <http://www.micronancegateway.
org/topics/graduation-sustainable-livelihoods>.
Accessed 25 May 2017.
Chang, Ha Joon. 2002. Kicking Away the Ladder
– Development Strategy in Historical Perspective.
London: Anthem Press.
Dos Santos, Paulo. 2011. “Production and
Consumption Credit in a Continuous-
Time Model of the Circuit of Capital.”
Metroeconomica 62(4): 729–758.
Dos Santos, Paulo, and Ingrid Harvold Kvangraven
(2017). “Better than Cash, but Beware the Costs:
Electronic Payments Systems and Financial
Inclusion in Developing Economies.” Development
and Change 48(2): 205–227. <http://onlinelibrary.
wiley.com/doi/10.1111/dech.12296/abstract>.
Accessed 25 May 2017.
Duvendack, M., R. Palmer-Jones, J.G. Copestake,
L. Hooper, Y. Loke, and N. Rao. 2011. What is the
Evidence of the Impact of Micronance on the
Well-being of Poor People? London: EPPI Centre,
Social Science Research Unit, Institute
of Education, University of London.
Gerschenkron, Alexander. 1962. Economic
backwardness in historical perspective. Cambridge,
MA: Belknap Press of Harvard University Press.
Gowrisankaran, G., and J. Stavins. 2002. “Network
externalities and technology adoption: lessons
from electronic payments.” NBER Working Paper
w8943. Washington, DC: National Bureau of
Economic Research.
Karlan, D., A.L. Ratan, and J. Zinman. 2014. “Savings
By and For the Poor: A Research Review and
Agenda.” Review of Income and Wealth 60(1): 36–78.
Roodman, David. 2012. Due Diligence: An
Impertinent Inquiry Into Micronance. Baltimore,
MD: Brookings Institution Press.
Roodman, David, and Jonathan Morduch.
2009. “The Impact of Microcredit on the
Poor in Bangladesh: Revisiting the Evidence.”
Working Paper 174. Washington, DC: Center
for Global Development.
Sinclair, Hugh. 2012. Confessions of a
Micronance Heretic: How Microlending Lost
Its Way and Betrayed the Poor. San Francisco,
CA: Berrett-Koehler Publishers.
Srinivasan, N. 2010. Micronance State of the
Sector in India. New Delhi: Access Development
Services, Sage Publications.
UNHCR. 2014. The Graduation Approach. Geneva:
United Nations High Commissioner for Refugees.
<http://www.unhcr.org/55005bc39.pdf>.
Accessed 25 May 2017.
1. The New School for Social Research.
2. See BtCA et al. (2014, 7–9).
3. See Karlan and Zinmann (2009), Banerjee et al.
(2013) and Duvendack et al. (2011), for instance.
4. As shown by Roodman and Morduch (2009)
and Devendack et al. (2011).
5. See Bateman (2010), Karim (2011), Roodman
(2012) and Sinclair (2012), for instance.
6. See dos Santos (2011).
7. See BtCA et al. (2014, 7).
8. See Gowrisankaran and Stavins (2002),
for instance.
9. See Mangelbaum (2012).
10. See, for instance, BtWC et al. (2014, 15).
11. See Gerschenkron (1962), Amsden (1992)
or Chang (2002), for instance.
58
Resilience and graduation
Greg Collins 1
From Bangladesh to Ethiopia and well
beyond, the contribution that graduation
approaches have made towards helping
poor households escape poverty
is well documented. Learning from
these approaches is also instructive
to the broader set of investments by
governments, donors and communities
themselves aimed at enabling these
pathways, including for the ‘ultra poor’
and those who are chronically vulnerable.
At the same time, the growing risk,
frequency and intensity of shocks and
stresses that poor households face is not
only inhibiting their ability to escape
poverty but is also causing many who have
escaped to fall back into poverty, and still
others to hover around the poverty line or
even descend into poverty for the rst time
(ODI 2014; 2016; Hallegatte et al. 2017).
How signicant is this as a (counter)
development trend? According to
the Overseas Development Institute’s
2014–2015 Chronic Poverty Report, over
60 per cent of households who escaped
poverty in rural Ethiopia between 1999
and 2009 fell back into poverty during
the same period. In rural Kenya the rate
was 44 per cent between 2004 and 2010.
As a percentage, fewer people who
escape poverty are falling back into it in
Asia. However, data from Bangladesh,
the Philippines, Vietnam and Indonesia
suggest rates of above 20 per cent, which
translates into massive numbers of people
and households (ODI 2014).
When added to the rate of near-poor
households descending into poverty or
unable to eectively distance themselves
from it, the numbers become truly
alarming. For example, a recent World
Bank report estimates that some 26 million
people are forced into poverty each year
due to natural disasters alone (Hallegate
et al. 2017). The same dynamics are also
sending those already living on less than
USD1.90 per day deeper into poverty and,
in extreme cases, resulting in recurrent
crises that manifest as repeating, large-
scale humanitarian emergencies
(ODI 2016; Hallegate et al. 2017).
In this light, sustainably reducing both
recurrent crises and poverty requires
strengthening resilience—that is, “the ability
of people, households, communities, systems
and countries to mitigate, adapt to, and
recover from shocks and stresses in a manner
that reduces chronic vulnerability2 and
facilitates inclusive growth” (USAID 2012).
The Graduation Approach’s
contribution to resilience
The real and potential contribution of
graduation approaches to strengthening
resilience and enabling sustainable
escape from poverty is signicant.
First and foremost, graduation approaches
aim to target ultra-poor households caught
at the intersection of chronic poverty,
vulnerability and exposure to shocks and
stresses—that is, those often left out of the
development enterprise.3 As with eorts
to strengthen the resilience of people
and places subject to recurrent crises
throughout Africa, Asia and beyond, this
constitutes an important shift towards
treating those not reached by market-
based, formal social protection provisions
and growth-oriented development
strategies as a development priority
rather than a perpetual humanitarian
risk. It also underscores the importance
of strengthening resilience to ensure that
people not only escape poverty and chronic
vulnerability but sustain these escapes in
the face of recurrent shocks and stresses.
Second, the tailored combination
of interventions used by graduation
approaches to help enable poverty
escapes also aligns with (and is supported
by) the growing body of evidence on
sources of resilience, which explains why
some households fall back—or deeper—
into poverty and chronic vulnerability in
the face of recurrent shocks and stresses,
while other, ‘more resilient’ households
are better able to manage shocks and
stresses, maintain their well-being and
sustainably escape from (or remain
out of) poverty and chronic vulnerability
(ODI 2016; Smith et al. 2015). This no doubt
reects an implicit understanding of the
importance of resilience to maintaining
graduation in complex risk environments.
Making this understanding more explicit
in programming and measurement would
only further enhance the contributions of
graduation approaches.
Financial services
Chief among the ‘sources of resilience’
is improving access to nancial services
among poor people—the very foundation
of BRAC and the impetus behind the
Graduation Approach. Recent research not
only arms that access to nancial services
helps explain why some households were
able to escape poverty and remain out of
it in Bangladesh, Ethiopia and Uganda, it
helps to explain why some households
were able to maintain their food security
status over multiple drought waves in
Ethiopia during 2015–2016, while other—
less resilient—households were not
(ODI 2016, Smith et al. 2015).
O-farm economic activities, employment
and diversied livelihood risk
Recent evidence also suggests that
engaging in o-farm economic activities,
including the type of sustainable
livelihoods increasingly promoted by
graduation approaches, provides both
a pathway out of poverty and a source of
resilience in the face of shocks and stresses
(ODI 2016). This is particularly true of
o-farm livelihood and income strategies
outside the agriculture sector that diversify
and reduce a household’s livelihood risk
prole in relation to climate shocks and
stresses (Nelson et al. 2016).
Graduation approaches are designed to
help ultra-poor households transition out
of precarious (last-resort) wage labour and
into more secure self-employment. Small
enterprises are making a clear contribution
on this front, albeit with a varying degree
of attentiveness to livelihood risk. However,
employment—including wage labour
accessed through migration to rural
towns and cities—constitutes an equally
important adaptive strategy, particularly
for the poorest households. As Mueller
and Chan (2015) suggest, expanding and
enabling poor households to access viable
and stable employment will become
more important—not less—as economies
develop and agricultural systems transform.
The International Policy Centre for Inclusive Growth | Policy in Focus 59
“ These systems provide
both a ‘push’ to enable
poor households to
escape poverty and
chronic vulnerability, and
a protective buffer that
allows these and other
households to remain out
of poverty in the face of
extreme events.
Transfers: assets and consumption support
Graduation approaches that combine asset
transfers and predictable, time-bound
consumption support are also making an
important contribution to resilience. This is
evidenced by the role of asset transfers and
consumption support in enabling poor
households in Ethiopia to better manage
through and begin to recover from the el
Niño-induced drought in 2015–2016, either
through transfers through the Productive
Safety Net Programme (PSNP) or the PSNP
contingency mechanism that enabled
transfers to an additional 2 million non-
PSNP households experiencing transitory
food insecurity or, in its absence, external
humanitarian assistance (Smith and
Frankenberger 2017).
This experience in Ethiopia, and the growing
body of evidence on shock-responsive social
protection, also makes clear that social
protection systems and their ability to exibly
expand in the face of shocks are critical for
eective disaster risk and drought-cycle
management (Oxford Policy Management
2016). Put simply, these systems provide
both a ‘push’ to enable poor households to
escape poverty and chronic vulnerability,
and a protective buer that allows these and
other households to remain out of poverty in
the face of extreme events that overwhelm
the ability of households and communities to
manage these events on their own.
Social capital, women’s empowerment,
aspiration and condence
The importance of graduation approaches
to resilience also extends to its very visible
contribution to building social capital
(inclusive social networks), women’s
empowerment and aspirations, as well as
the condence to mitigate, adapt to and
recover from shocks and stresses in ways
that are just now starting to be properly
understood. The existing evidence on both
sustainable poverty escapes and managing
recurrent crises makes it abundantly clear
that these social and psychological factors
are as important—and potentially even
more so—to build resilience than the other
contributions of Graduation approaches
previously outlined (Béné et al. 2016).
The contribution of resilience
to graduation approaches
The emergence of resilience as an analytic,
programmatic and organising concept
also signicantly contributes to scaling
more shock-responsive, risk-informed
approaches to graduation.
Complex and compound risk
First, resilience demands a more explicit
recognition of the risk environments in
which poor and chronically vulnerable
people live and the pernicious and
perennial threat this poses to their
ability to both escape and remain out
of poverty and chronic vulnerability.
The complex and compound nature of
these environments is also clear; it is rarely
a single event that sends households
back or deeper into poverty and chronic
vulnerability but, rather, a combination
of shocks and stresses, from large-scale,
covariate droughts and oods, to more
idiosyncratic shocks such as the loss of a
wage earner or a health crisis within the
household that unfold over time.
An explicit focus on strengthening
resilience capacities
This more explicit recognition of risk
demands a greater focus on strengthening
the set of capacities (absorptive,
adaptive and transformative) that enable
households and communities to mitigate,
adapt to and recover from shocks and
stresses and remain out of poverty and
chronic vulnerability, alongside eorts
to facilitate escapes from poverty. As the
term ‘graduation’ implies, there is likely a
threshold (distance from poverty) beyond
which households are less likely to fall
back. Yet, as much as this distance can
provide some buer, it alone is insucient
without also strengthening resilience
capacities—particularly given the extent
to which recurrent shocks and stresses can
erode this buer over time.
Graduation approaches are already clearly
contributing to strengthening these
capacities. However, this contribution would
be even greater if graduation approaches
were combined with other interventions
that not only help people escape and
distance themselves from poverty and
chronic vulnerability but also strengthen
their ability to sustain these escapes in the
face of complex risk and recurrent shocks
and stresses. The starting point for this type
of collective action is a deeper and shared
understanding of risk and resilience capacities
in the varied contexts in which graduation
approaches are being implemented.
Photo: BRAC. Woman beneciary of BRAC Poultry project, Uganda, 2012.
60
“ Strengthening
resilience requires
next-generation
investment in human
capital, including
in health, nutrition
and education, as
a complement to
investment in short-term
skills and capacities.
Photo: BRAC. Woman running small grocery from BRAC micronance project, Tanzania, 2014.
Collective action across scales,
systems and sectors
The concept of resilience provides a useful
framework in that it demands working not
only across dierent sectors but across
dierent scales—from individuals and
households to communities, systems and
countries. It does so both because resilience
capacities exist at these dierent scales and
because strengthening resilience at any
scale, including the resilience of poor and
chronically vulnerable households,
requires working across them.
Increasing income from both on- and o-
farm livelihoods as a source of resilience
and as a sustainable pathway out of
poverty illustrates the point. Graduation
approaches help empower and ‘push’
individual households through asset
transfers, consumption support, access to
nancial services and skills development.
However, the ability to do so at scale will be
severely constrained without accompanying
investment in the development of inclusive
market systems that engage the private
sector and create ‘pull’ dynamics by
expanding opportunities for employment
and enterprise (Irwin and Campbell 2015).
Similarly, investments to improve natural
resource management and the resilience of
ecological systems are critical for expanding
(and even maintaining) livelihoods that
depend on the natural resource base,
particularly during shock events
(Nelson et al. 2016; Smith et al. 2015).
Inclusive social protection systems serve
as another important example. As with
graduation programmes, social protection
provides a much-needed ‘push’ in the
form of timely and predictable transfers.
Moreover, transfers provided through
social protection systems are, in general,
not time-bound and provide critical
ongoing support to those ill-equipped
to benet from a graduation approach.
Inclusive social protection also provides a
critical protective buer when household
and community resilience capacities are
overwhelmed in the face of large-scale,
covariate shock events such as droughts
and oods, as well as idiosyncratic shocks
such as a health crisis or the loss of a wage
earner within a household. Given the
alarming rates at which households are
descending or falling back into poverty,
the existence and permanence of inclusive
social protection systems as a source of
resilience is just as critical for ‘near poor’
people and those who have escaped
poverty as they are to those who are poor
and chronically vulnerable.
Strengthening resilience also requires
working across sectors, including—but
not limited to—those in which graduation
programmes typically engage. Beyond
expanding economic opportunities,
this includes strengthening governance
through improved natural resource, conict
and disaster risk management (including
the aforementioned shock-responsive
social protection), as well as strengthening
local institutions in these and other areas.
In more fragile contexts, it also requires
the very dicult endeavour of addressing
strained State–society relations, which are
the cornerstone of transformative capacity
and shape the enabling environment within
which household- and community-level
resilience can be fully realised.
Finally, strengthening resilience requires
next-generation investment in human
capital, including in health, nutrition and
education, as a complement to investment
in short-term skills and capacities. For
example, prior-generation investment
in education, as measured by current-
generation educational attainment, is a
powerful predictor of a household’s ability
to sustain and even improve its well-being
in the face of recurrent shocks and stresses
across a wide range of contexts in which the
relationship has been examined (Smith et al.
2015; ODI, 2016). This is well recognised
by many poor and resource-constrained
households which prioritise sending their
children to school above other needs.
Conclusion
Resilience and graduation are intimately
interwoven in both purpose and approach.
Yet it is equally clear that there is much
to be gained by incorporating resilience
more explicitly into work on graduation
and ensuring that the tenets of and lessons
from work on graduation inform a broader,
collective eort to build resilience and
sustainable ends to recurrent crises and
extreme poverty. This requires moving
beyond unnecessary polemics that suggest
that graduation approaches and social
protection are somehow competing with
one another or that investment in one
sector is more important than investment
The International Policy Centre for Inclusive Growth | Policy in Focus 61
“ There is much to be
gained by incorporating
resilience more explicitly
into work on graduation
and ensuring that
the tenets of and
lessons from work on
graduation inform a
broader, collective
effort to build resilience
and sustainable ends
to recurrent crises and
extreme poverty.
in another, towards a shared, evidence-
based vision for collective action and
impact, which recognises the role of varied
investment across dierent sectors, scales
and time horizons. The emergent eld of
resilience measurement and the growing
body of evidence being generated by it
provide a fertile platform for advancing
thinking on this front.
Béné, Christophe, Tim Frankenberger, Mark
Langworthy, Monica Mueller, and Stephanie
Martin. 2016. “The Inuence of Subjective and
Psycho-social Factors on People’s Resilience:
Conceptual Framework and Empirical Evidence.”
Technical Report Series No 2: Strengthening
the Evidence Base for Resilience in the Horn of
Africa. Washington, DC: USAID Feed the Future.
Hallegatte, Stephane, Adrien Vogt-Schilb,
Mook Bangalore, and Julie Rozenberg. 2017.
“Unbreakable: Building the Resilience of the Poor in
the Face of Natural Disasters.” Climate Change and
Development Series. Washington, DC: World Bank.
Irwin, Bronwyn, and Ruth Campbell. 2015.
“Market Systems for Resilience.” Leveraging
Economic Opportunities Report No. 6.
Washington, DC: USAID. <https://www.
microlinks.org/sites/default/les/resource/les/
Market_Systems_for_Resilience__nal_508_
Compliant.pdf>. Accessed 20 April 2017.
Kidd, Stephen. 2013. “The Misuse of the Term
‘Graduation’ in Social Policy.” Pathways’ Perspectives
on Social Policy in International Development 14.
London: Development Pathways.
Mueller, Bernd, and Man-Kwun Chan. 2015.
“Wage Labor, Agriculture-Based Economies, and
Pathways out of Poverty: Taking Stock of the
Evidence.” Leveraging Economic Opportunities
Report No. 15. Washington, DC: USAID.
<http://www.ilo.org/wcmsp5/groups/public/--
-ed_emp/---emp_ent/documents/publication/
wcms_372897.pdf>. Accessed 20 April 2017.
Nelson, Suzanne, Tim Frankenberger, Mark
Langworthy, Tim Finan, and Thomas Bower.
2016. “The Eect of Livelihood Diversity on
Recovery and Shock Impact in Ethiopia, Kenya
and Uganda.” Technical Report Series No 2:
Strengthening the Evidence Base for Resilience
in the Horn of Africa. Washington, DC: USAID
Feed the Future.
ODI. 2014. The Chronic Poverty Report 2014-15:
The Road to Zero Extreme Poverty. <https://www.odi.
org/sites/odi.org.uk/les/odi-assets/publications-
opinion-les/8834.pdf>. Accessed 20 April 2017.
ODI. 2016. “Investigating the Dynamics
and Drivers of Transitory Poverty Escapes.”
Overseas Development Institute website.
<http://www.chronicpovertynetwork.org/
projects-1/2016/5/23/leveraging-economic-
opportunities-leo>. Accessed 20 April 2017.
Oxford Policy Management. 2016.
DFID Shock-Responsive Social Protection
Systems Research: Literature Review.
Oxford, UK: Oxford Policy Management.
Smith, Lisa, and Tim Frankenberger.
Forthcoming 2017. Pastoralist Areas Resilience
Improvement and Market Expansion (PRIME)
Recurrent Monitoring Survey 2014-2015 Deep
Dive: Uncovering the Pathways to Resilience.
Washington, DC: USAID.
Smith, Lisa, and Tim Frankenberger, Ben
Langworthy, Stephanie Martin, Tom Spangler,
Suzanne Nelson, and Jeanne Downen (2015).
Ethiopia Pastoralist Areas Resilience Improvement
and Market Expansion (PRIME) Project Impact
Evaluation: Baseline Survey Report Volumes 1
and 2. Washington, DC: USAID Feed the Future.
<https://agrilinks.org/sites/default/les/
resource/les/EthiopiaPRIMEVol1nal.pdf>
Accessed 20 April 2017.
1. USAID Resilience Coordinator and Director,
USAID Center for Resilience.
2. ‘Chronic vulnerability’ refers to an enduring
susceptibility to the eects of external shocks
on life and livelihoods—“a shock that is not
acute or transient, but constant and cyclical”.
See <http://odihpn.org/magazine/editors-
introduction-chronic-vulnerability/>.
3. Kidd (2013) and others have criticised the
eectiveness of graduation approaches in
reaching ultra-poor people.
Photo: BRAC/Alison Wright. Woman in maize elds producing crops independently, Tanzania, 2014.
62
Source: UNHCR (2017a).
Leaving no one behind:
graduation for refugees
Helene Kuhle,1 Alexi Taylor-Grosman 2
and Andrew Mitchell 3
Over the past three years, more people
have been forced to ee their homes
than at any time since the Second World
War, and this number continues to grow:
currently, more than 65 million people are
forcibly displaced4 globally; among them,
more than 21 million are refugees. Many
refugees nd themselves in protracted
situations of ve years or more with few
prospects for attaining a solution to
displacement because of shifting political
landscapes and unending conicts
(UNHCR 2003; 2017b).
With the number of forcibly displaced
persons caught in protracted
displacement situations on the rise, it is
increasingly evident that a humanitarian
response focusing on the short term is
inadequate. Accordingly, a growing global
consensus recognises that displacement
requires a joint humanitarian and
development response, underpinned by
long-term planning and programming
for solutions. The response must involve
a broad coalition of actors, including
governments. Planning for solutions
should include strengthening refugees’
self-reliance and resilience, as these
approaches can empower them to
live more independently of external
assistance; stabilise their means for living;
and enable them to contribute to the
local economy, while preparing them
to take advantage of whatever solution
ultimately becomes available. These
insights are reected in core high-level
policy addressing forced displacement,
including the ‘New York Declaration for
Refugees and Migrants’ (United Nations
2016a) and the ‘Comprehensive Refugee
Response Framework’ (UNHCR 2016).
Moreover, these policies link with, and
build on, the Sustainable Development
Goals and principles of the 2030 Agenda
for Sustainable Development—that no
one should be left behind, and that those
furthest behind should be reached rst
(United Nations 2016b).
The vast majority of refugees live in the
developing world, and more than 4 million
are hosted by countries where the average
income of their own citizens is far below
the extreme poverty line. Often, the socio-
economic situation of refugees matches
that of the host community. As such, it
is not surprising that in these contexts
refugees often lack access to sustainable
livelihoods and long-term support from host
governments. Moreover, they often struggle
with additional challenges pertaining to
their refugee status: they may face legal or
administrative barriers limiting opportunities
to gain lawful and decent employment
(Zetter and Ruaudel 2016), further
compounded by discrimination from the
local population, a lack of documentation
necessary for formal markets, and a lack of
language and skills that match the market
opportunities in the host country.
In response, the United Nations High
Commissioner for Refugees (UNHCR)
has been enhancing collaboration
with development actors to bridge the
current gap between humanitarian
and development eorts and ensure
that refugees—in particular, extremely
poor ones—are not left behind. UNHCR
engages in eorts to strengthen self-
reliance and resilience with the objective
of equipping and preparing its persons
of concern to respond to the protection
risks they face and to take advantage of
opportunities leading to solutions. How
a displaced person fares in the future,
whether in the country of asylum or origin
or in a third country, depends on the skills,
experiences and qualications, mental
and physical health, material assets and
attitudes maintained and developed while
displaced. These resources and qualities are
benecial not only to refugees but to their
communities, including in areas of return or
relocation, or in countries of settlement.
With support from Trickle Up, in 2013
UNHCR adopted the Graduation Approach
as a vehicle to enhance international
protection and improve refugees’ prospects
for solutions. Through the Graduation
Approach, UNHCR and its partners set
out to increase self-reliance and resilience
among refugees and host community
members living in extreme poverty.
Graduation: building
self-reliance in refugees
Eorts to build self-reliance aim to enable
refugees to meet essential needs and
to enjoy human rights in a sustainable
manner and with dignity. Participants in
graduation programmes are supported
to progress along a set of objectives,
including establishing food security,
building self-condence and agency,
increasing resilience through sustainable,
UNHCR is mandated by the United Naons General Assembly to work with others
to provide internaonal protecon and seek permanent soluons for refugees and
other persons of concern.
Soluons are secured when refugees enjoy their rights, including access to
naonal services and systems, at the same level as people they live among,
without facing any discriminaon—whether in their country of origin, country
of asylum or a third country. It entails addressing four interrelated dimensions:
economic, legal, socio-cultural and civil-polical, to ensure sustainable soluons.
Voluntary repatriaon, local integraon, reselement and complementary
pathways are all possible routes to aain soluons.
BOX 1: UNHCR’s mandate for protection and solutions
The International Policy Centre for Inclusive Growth | Policy in Focus 63
“ The vast majority
of refugees live in the
developing world, and
more than 4 million are
hosted by countries
where the average
income of their own
citizens is far below the
extreme poverty line.
“ Displacement requires
a joint humanitarian and
development response.
Source: Authors’ elaboration.
stabilised income and savings, and
establishing access to networks and
services that will continue after the end of
the project. The following components are
included in UNHCR’s Graduation Approach:
yCoaching: Intensive weekly or bi-
weekly 20–30-minute coaching
sessions throughout the programme
implementation period facilitates a
smooth transition as the participant
navigates the multi-staged
programme. The coach also helps
establish links to relevant networks
and service providers that will remain
available after the project ends.
yNetwork engagement: UNHCR
encourages its graduation participants
to interact with other programme
participants, refugees and host
community members alike, either
through savings groups, social
gatherings or group training.
This helps refugees overcome the
very common sense of isolation and
reinforces social capital, both within
the refugee community and with
the host community, which is
essential for self-reliance.
ySavings: Saving, whether in a formal
nancial institution or in an informal
group setting, is a vital tool for risk
management. Regular saving helps
participants build assets and instils the
habit of saving. In addition to helping
overcome unexpected shocks, these
assets can be used for investment and
productive risk-taking activities.
yConsumption support: Often poverty
and food insecurity inhibit households
from taking on any meaningful longer-
term livelihood strategy. UNHCR provides
food and/or cash assistance, often in
collaboration with the World Food
Programme (WFP) or other pre-existing
safety net programmes, for a xed
period. With this support, participants’
basic needs are met, allowing them to
engage in programme activities and
more proactively focus on livelihood
activities that increase their self-reliance.
yBuilding core capacities: Many refugees
arrive in their host country without
documentation or knowledge of what
services exist to support them. UNHCR
oers capacity-building on legal rights,
cultural norms and integration, and
other core skills such as household
nancial management which support
engagement in successful income-
generating activities in the country
of asylum. By building core capacities,
refugees are supported to better claim
their civil, cultural, economic, political
and social rights (UNHCR 2017c).
yLivelihoods roadmap: Coaches
work closely with their graduation
ySelf-reliance refers to the ability of an individual, household or community to meet
essenal needs and to enjoy human rights in a sustainable manner and with dignity.
yResilience refers to the ability of individuals, households, communies, naonal
instuons, systems and sociees to prevent, absorb and recover from shocks,
while posively adapng and transforming their structures and means for living
in the face of long-term stresses, change and uncertainty.
FIGURE 1: Components of UNHCR’s Graduation Approach5
Livelihoods roadmap
Self Livehood
asset
Technical
skills
training
Wage
Graduaon components
StartMidway End 5
Job support
Consumpon support
Savings
Network engagement
Coaching
Core capacity building
1
2
3
4
5
6
7
8
BOX 2: Self-reliance and resilience
BOX 3: Integrating resilience into programming means enhancing
three forms of capacities
64
“ Building the resilience
of people living in
poverty is necessary
to ensure that progress
towards self-reliance
is not eroded or reversed
in the face of shocks
or crises.
Photo: UNHCR/S.Rich. Cenelia, an internally displaced person in Colombia, makes intricate beaded bracelets.
participants to develop and implement
appropriate and decent livelihoods
strategies, both through self-
employment and wage employment
mechanisms, depending on the
context. To ensure that the roadmap
oers a sustainable path for income
growth and self-reliance, it must be
based on market opportunities, build
on participants’ nancial resources and
skills, and leverage participants’ agency
to seize these livelihood opportunities.
When refugees do not have access
to legal employment, UNHCR and
partners inform refugees about their
rights, monitor refugees and their
employers to ensure protection,
advocate on the behalf of employees
for those rights that do exist, and work
on the regulatory environment to help
build more formal opportunities.
yTechnical skills training: Following the
roadmap, technical skills training is
oered to graduation participants to
equip them to successfully engage
in the self- or wage employment
opportunities identied. Technical skills
training is often critical for refugees
to be able to meet the demand in the
local market; for example, it is not
uncommon for former city-dwellers
to arrive in rural areas or for farmers
to settle in urban areas. Graduation
participants who are engaged in
the self-employment track receive
entrepreneurship training as well
as technical training linked to their
specic livelihood activities. Examples
of technical training may include food
hygiene, good practices in animal
rearing or local agricultural best
practices. Graduation participants
engaged in the wage employment
track receive employability training
and vocational training, which is often
provided by the employers.
yLivelihoods asset transfer/employment
support: Finally, those participants
who will start their own enterprise
receive an asset transfer to start their
livelihood opportunity. This is most
frequently oered in the form of cash,
linked to the purchase of a specic
asset, though it may also be oered
in-kind. In Burkina Faso, for instance,
participants who planned to create
artisan bowls used their seed capital
to purchase raw metal, leather sheets
and tools required to fashion the
artisan products. In the UNHCR setting,
particularly in urban and peri-urban
areas where the wage employment
track is pursued, participants are also
linked to employment opportunities.
In Cairo, UNHCR and partner sta
work closely with local employers to
understand their needs, and screen
refugees, to better link candidates with
potential employers. When possible,
these programmes also leverage
UNHCR’s or governments’ existing
employment linkage programmes.
In Costa Rica, for example, UNHCR sta
use a network of private-sector actors
to oer training on refugee employment
to human resources oces; then they
map vacancies with private-sector
companies to share with refugees.
Ensuring sustainability through resilience
Building the resilience of people living
in poverty is necessary to ensure that
progress towards self-reliance is not eroded
or reversed in the face of shocks or crises
and longer-term trends (such as climate
yAn absorbance capacity protects people against household, individual and
widespread shocks.
yAn adapve capacity allows people to use new ways of managing (and adapng to)
changes in shocks and direct impacts of long-term trends.
yA transformave capacity enables people to become the agents of polical and
economic change to tackle underlying issues of power at the heart of vulnerability
to shocks and change.
The International Policy Centre for Inclusive Growth | Policy in Focus 65
change). By instilling the habit of saving,
ensuring that productive assets consider
the risk of natural hazards, and building
participants’ capacity to proactively
address uctuating markets and changing
ecosystems, the Graduation Approach
contributes to strengthen the capacity
of refugees and host communities living
in extreme poverty to cope with shocks.
While more long-term research is required
to measure how participants confront
these types of challenges, it is expected
that the increased economic and social
assets that participants acquire through the
Graduation Approach will better position
them to confront common household
and individual-specic shocks, such as
illness, loss of income, negative coping
mechanisms and protection abuses.
In UNHCR, the Graduation Approach
may furthermore address unequal
power dynamics that drive vulnerability
to risk, such as the subservient role of
women in families and communities
and the exclusion of non-nationals and
extremely poor people from community
development and local government
processes. For example, coaches are
trained to address intra-household
conicts, either by working with the
household to mitigate them or by
referring the household to necessary
and appropriate services. In addition, by
building condence and facilitating joint
planning and economic activities within
families, marginalised families are more
likely to participate in social and economic
opportunities in the wider community.
A key aspect of the Graduation Approach
in UNHCR is to ensure that participants
are linked to relevant services and social
support structures, ensuring sustainability
after graduation and empowering
people to take advantage of government
services. In Ecuador and Costa Rica,
UNHCR is exploring arrangements with
the government to incorporate graduated
participants into social services, including
social protection, training and job
placement programmes. Coaches may
also refer participants to psychosocial
counselling or legal services, if needed,
or accompany participants to health
clinics or to schools to register children.
By linking participants to existing
structures and services, the Graduation
Approach helps address the multiple
constraints of refugees living in extreme
poverty, ultimately strengthening their
transformative capacity and assuring
a life of well-being with dignity—both
while in displacement and once a solution
becomes available.
Preliminary results
While UNHCR and Trickle Up are still learning
about how best to adapt the Graduation
Approach to the specic needs of refugees
in dierent contexts, initial economic and
social results have been promising.
In Egypt, 1,275 households, including Syrian
refugees in the emergency context, Africans
living in a protracted refugee status, and
poor Egyptians, were enrolled in the pilot.
At the time of the mid-term evaluation,
UNHCR Egypt found that the average
income earned per person per month
increased by 27 per cent for participants in
Alexandria and 18 per cent for participants
working in Cairo. Six months into
implementation, 68 per cent of participants
on the wage employment track had been
placed in a job. For participants engaged
in the self-employment track, 97 per cent
of participants in Cairo and 78 per cent of
participants in Alexandria had started a
business. This is particularly noteworthy,
since legal access to work for refugees is
limited. More research on longer-term
job retention is needed (Beit Al Karma
Consulting 2016).
UNHCR Ecuador piloted the Graduation
Approach with 180 urban refugees, most
of whom hailed from neighbouring
Colombia. Just eight months after services
commenced, participants reported an
overall increase in average household and
per capita monthly incomes. Moreover, 35
per cent of participants reported generating
an income of at least the national poverty
line of USD82 per capita. At the time of
the study, 61 per cent of participants
reported access to a formal savings account,
compared to 1 per cent at the beginning of
the programme. This is especially signicant
given the challenge faced by refugees in
opening a bank account in Ecuador, and
came as a result of considerable eort by
the UNHCR Ecuador team to ensure nancial
inclusion (UNHCR and Trickle Up 2016).
In Costa Rica, UNHCR’s 200 pilot participants
included refugees from Colombia, many of
whom had been in the country for years,
and Costa Rican women who were at risk
of sexual and gender-based violence. In its
nal evaluation, UNHCR Costa Rica reported
that unemployment rates of its graduation
participants decreased from 36 per cent
to 4 per cent, while self-employment
increased from 24 per cent to 59 per cent.
The percentage of participants who had
a monthly household income equal to or
greater than the national minimum wage
increased from 15 per cent to 79 per cent.
The percentage of households that reported
saving money increased from 14 per
cent to 68 per cent. Fifty-eight percent of
participants reported increased emotional
well-being (UNHCR Costa Rica 2017).
Going forward
Since adopting the Graduation Approach
in 2013, UNHCR, with technical assistance
from Trickle Up, has piloted the
methodology with refugees in Burkina
Faso, Costa Rica, Ecuador and Egypt. In
each context, UNHCR targets at least 25
per cent of local community members
living in extreme poverty, to build buy-in
and help foster social cohesion. In addition
to blazing the trail for selecting refugees
for the Graduation Approach, these
pilots have innovated in serving urban
populations through it and incorporating
wage employment opportunities.
As part of its Global Strategy for Livelihoods
Programming 2014–2018, which
focuses specically on enhancing the
economic inclusion of refugees, UNHCR,
in collaboration with Trickle Up, is in the
process of designing and implementing
the Graduation Approach in at least
ve more countries in 2017, including
Argentina, Mozambique, Sudan, Zambia
and Zimbabwe. By the end of 2018, UNHCR
expects to expand the implementation to
more than 20 country operations.
Through these eorts, UNHCR will
continue to explore how the Graduation
Approach can contribute to the protection
of refugees. This begins by ensuring that
eorts are systematically linked with
national systems and development plans
to ensure sustainability. Of course, there
are limits to the scope of the Graduation
Approach. On the one hand, it is a time-
and resource-intensive intervention that
targets those living in extreme poverty,
rather than those who are better o; at the
same time, it requires that a participant
66
“ A key aspect of the
Graduation Approach in
UNHCR is to ensure that
participants are linked
to relevant services and
social support structures,
ensuring sustainability
after graduation and
empowering people
to take advantage of
government services.
be able to engage in livelihood activities,
excluding some of the most vulnerable.
These limitations may, however, be
mitigated by targeting participants
whose graduation from extreme poverty
will enable them to better care for
more vulnerable family members, thus
expanding protection dividends to the
wider community. In addition, other types
of livelihoods and protection programmes
may be applied to benet those not
engaged in graduation programmes.
For its part, Trickle Up, with support from
the US Department of State’s Bureau of
Population, Refugees, and Migration,
is supporting the eort to improve the
economic well-being of refugees through
continued technical assistance and stang
support to UNHCR. Trickle Up is also seeking
to engage other graduation practitioners
in the inclusion of refugees, and to engage
organisations focused on refugees to
integrate lessons from graduation into
their livelihood programming.
For UNHCR, the importance of the
linkages between protection, solutions and
development endeavours that are inclusive
of refugees, internally displaced people,
returnees and stateless persons is clear. By
committing to the Sustainable Development
Goals and the 2030 Agenda for Sustainable
Development’s aim to leave no one behind,
States and international actors promise to
proactively address the needs of persons
often excluded based on their legal status,
displacement or their recent return to homes
and communities damaged by conict.
UNHCR is hopeful that its eorts will be
one of many. With 192 signatories, the New
York Agreement from September 2016
provides a global framework for convening
development actors, governments, private-
sector actors and UN institutions in a
collective response to the global refugee
crisis. The Graduation Approach could
represent a global good practice that might
enable governments and development
actors to bridge the gap between
humanitarian and development eorts and
take over what UNHCR and partners started
by including refugees in regional, national
and local development plans.
Beit Al Karma Consulting. 2016. Final Report: Mid-
Term Evaluation of UNHCR Graduation Programme
in Egypt. Cairo: Beit Al Karma Consulting.
<https://www.micronancegateway.org/sites/
default/les/publication_les/unhcr_mid-term_
evaluation_2015_beit_al_karma_nal_report_21_
march_2016.pdf>. Accessed 25 May 2017.
UNHCR. 2003. Framework for Durable Solutions
for Refugees and Persons of Concern. Geneva:
United Nations High Commissioner for
Refugees. <http://www.unhcr.org/partners/
partners/3f1408764/framework-durable-
solutions-refugees-persons-concern.html>.
Accessed 25 May 2017.
UNHCR. 2016. Comprehensive Refugee
Response Framework: from the New York
Declaration to a global compact on refugees.
Geneva: United Nations High Commissioner for
Refugees. <http://www.unhcr.org/584687b57.
pdf>. Accessed 25 May 2017.
UNHCR. 2017a. “Solutions.” United Nations
High Commissioner for Refugees website.
<http://www.unhcr.org/en-us/solutions.html>.
Accessed 25 May 2017.
UNHCR. 2017b. “Figures at a Glance.” United
Nations High Commissioner for Refugees
website. <http://www.unhcr.org/gures-at-a-
glance.html>. Accessed 25 May 2017.
UNHCR. 2017c (forthcoming). Standing
Committee paper: Resilience and Self-reliance from
a Protection and Solutions Perspective. Geneva:
United Nations High Commissioner for Refugees.
UNHCR and Trickle Up. 2016. A Review of UNHCR’s
Graduation Program: Santo Domingo de los
Tsáchilas, Ecuador. Geneva: United Nations
High Commissioner for Refugees.
UNHCR Costa Rica. 2017 (forthcoming).
Graduation Model: Pilot project 2014–2016.
San José, Costa Rica: UNHCR Costa Rica.
United Nations. 2016a. “New York Declaration.”
United Nations website. <http://refugeesmigrants.
un.org/declaration>. Accessed 25 May 2017.
United Nations. 2016b. “The Sustainable
Development Goals Report 2016.” United Nations
website. <https://unstats.un.org/sdgs/report/2016/
leaving-no-one-behind>. Accessed 25 May 2017.
Zetter, Roger, and Héloïse Ruaudel. 2016.
Refugees’ Right to Work and Access to Labor
Markets – An Assessment. Washington, DC:
Knomad. <http://www.knomad.org/docs/
KNOMAD_study/KNOMAD%20Study%201
%20-%20Volume%201%20-%20Refugees%
20Rights%20to%20Work%20-%20An%20
Assessment.pdf>. Accessed 25 May 2017.
1. UNHCR’s Livelihoods Unit.
2. Trickle Up.
3. Operational Solutions and Transitions
Section, UNHCR.
4. ‘Forced displacement’ refers to situations
where a person has been forced to move
from their home—for example, due to armed
conicts or natural disasters. Forcibly displaced
persons include refugees, internally displaced
persons and asylum-seekers.
5. This image is used to highlight the
components and sequencing being used by
UNHCR in the design and implementation of the
Graduation Approach with persons of concern.
Photo: UNCHR Ecuador/C.Loza. Marcela, a refugee from Colombia, makes plantain chips in her new home in Ecuador.
The International Policy Centre for Inclusive Growth | Policy in Focus 67
Private-sector investment capital in
graduation: it is time to unlock sustainable
nancing at scale
Shaifali Puri 1 and Anne H. Hastings 2
Poverty is a pernicious and fundamental
problem, one that has perennially
plagued society. The United Nations
estimates that 10.2 per cent of the world’s
population lived with their families in
extreme poverty on less than USD1.90
per person per day in 2015, and the World
Bank estimates that around 767 million
people lived under this threshold in 2013.3
Moreover, children are more than twice as
likely as adults to live in extreme poverty,
and globally almost 385 million children
were living in extreme poverty. In sub-
Saharan Africa, half of all children live
below this threshold (UNICEF and World
Bank Group 2016; World Bank 2016).
Since 2000, the United Nations, together
with its Member States, has sought to align
global actors around major development
challenges, with the target to halve
extreme poverty between 1990 and 2015
as part of Goal 1 of the discontinued
Millennium Development Goals (MDGs).
Once again, in 2015, 193 countries
endorsed the Agenda for Sustainable
Development and the Sustainable
Development Goals (SDGs), rearming
their commitment to eliminate poverty in
all its forms everywhere by 2030, now with
a target of eradicating extreme poverty (as
dened by the USD1.25 a day threshold).
While a growing number of stakeholders
have joined in the pursuit of this ambitious
goal, and signicant progress has been
made in the last two decades, the poorest
subset of this group has not beneted
from this progress. Termed ‘ultra poor’, this
most underserved group is the population
that Uplift4 exclusively focuses on.5
To end extreme poverty, we must
prioritise reaching ultra-poor people rst
Within the landscape of extreme poverty,
the challenge of reaching ultra-poor
people is a particularly dicult one.
Existing at the margins, they are the
most deprived people within their
communities, with extremely limited
livelihood prospects and severe social
isolation, expending much of their
energy and time on procuring food for
their households. Often beyond the reach
of most formal social protection and
poverty alleviation programmes funded
by philanthropy and government, they
cannot meaningfully access existing
market-based support mechanisms
such as micronance. They also lack
the nancial resources to withstand
economic setbacks or to sustainably
lift themselves out of ultra poverty.
Whether due to diculty, high cost or lack
of know-how, to date, economic growth
and development interventions have
largely excluded—or failed to reach—
ultra-poor people, or address the unique
and specic circumstances that keep
them caught in poverty traps. However,
without the targeted and sustained
support they need to cultivate livelihoods
and attain social inclusion, the ultra poor
and their families are unlikely to escape
the intergenerational cycle of extreme
poverty. It has recently been noted that
they are the most dicult demographic
group to pull over the extreme poverty
line, given that most of them live in
countries with feeble or inadequate social
safety net systems and little opportunity
for formal participation in the labour
market (The Economist 2017).6
Therefore, attaining the ambitious goal
of eliminating all forms of poverty by
2030 will require both a concerted
commitment by governments and
multilateral organisations to prioritise
reaching the ultra poor, as well as a
very large commitment of resources to
craft the targeted policies, programmes
and other interventions necessary to
do so.7 According to estimates from the
United Nations Conference on Trade
and Development (UNCTAD 2014a;
2014b), achieving the SDGs will require
between USD3.3 trillion and USD4.5
trillion in investment per year, and likely
more, with an annual funding shortfall
of USD2.5 trillion. However, poverty
alleviation efforts have, to date, remained
dependent on traditional sources of
financing, including national government
programmes, overseas development aid
funding or private philanthropic capital.
These sources of funding, with numerous
constraints and political considerations,
are insufficient to address the magnitude
of the problem. For example, data
from the Organisation for Economic
Co-operation and Development
(OECD 2015) for Official Development
Assistance showed a total expenditure of
USD131.4 billion, while the Global Impact
Investing Network (2017) found that
those surveyed—who comprise merely
a snapshot of the total global impact
investing picture—currently manage
USD114 billion in impact investing assets.
It is, therefore, imperative to tap into
alternative funding resources for capital
for social protection programmes.
In other words, a significant deployment
of private-sector investment capital
will also be crucial.
Unlocking innovative nancing
mechanisms to scale up the
Graduation Approach
A number of factors account for the
dearth of scalable private investment
capital in extreme poverty eradication.
First, private-sector investors typically
require an identiable and measurable
return on capital deployed within
an agreed timeframe. Traditionally,
programmes to eradicate extreme poverty
have lacked the ability to show veried
evidence of their ecacy along specic
indicators, or a guaranteed timeline for an
expected result. Even impact investment
capital and so-called ‘patient capital’—
investments that demand both social
and nancial returns on longer time
horizons—require clearly measurable
returns, benchmarks of clear success and
a specied timeline to achieve them.
68
In recent years, however, two exciting
innovations in the elds of ultra
poverty and social nance have gained
momentum, which together have the
potential to leverage private-sector
investment capital at scale towards
eradicating ultra-poverty: the Graduation
Approach and social and development
impact bonds (SIBs and DIBs).
Pioneered by BRAC in Bangladesh,
the Graduation Approach provides
a comprehensive intervention that
specically targets the needs of ultra-
poor households. Given BRAC’s success
in Bangladesh, from 2006 to 2014
the Consultative Group to Assist the
Poor (CGAP) and the Ford Foundation
launched 10 pilot initiatives in eight
countries to test the replicability of the
BRAC approach in dierent settings.
In 2015, the ndings of six randomised
controlled trials (RCTs) assessing the
pilots were published in Science.8
The researchers concluded that the
Graduation Approach yields signicant
improvement across multiple indicators
of extreme poverty and that its outcomes
are sustainable, and continue to improve
over time (Kristo 2015). Since its
inception, it has been adapted by BRAC
and replicated and adapted by many
other implementers and governments
(Fahey and Loiseau 2016). Thus, the
Graduation Approach provides, for the
rst time, a rigorously vetted, data-driven,
evidence-based method that has
the potential to permanently eradicate
ultra poverty.
The social impact bond (SIB) and
development impact bond (DIB) models
Equally as exciting has been the
emergence of SIBs, and their counterpart
in the development context, DIBs, over
the past decade.9 Though they take
their name from the xed-income asset
class, impact bonds are not bonds in the
traditional sense of a debt-based security
that pays investors a xed interest rate
until they mature. Rather, they provide
a pay-for-performance mechanism
whereby a private investor, rather than
the government, provides the upfront
funding for the implementation of a
programme by a service provider to
deliver specic social or environmental
outcomes. A third party (e.g. government,
private donors or both) acts as the
outcome funder, returning the capital
plus interest to the upfront investor
once the specied outcomes have been
independently vetted and met. If the
outcomes are not met, the interest
plus part of the capital is lost.
Unlike traditional philanthropic or
government funding models, impact
bonds help align incentives between
service providers and funders by:
(i) providing the full cost of programme
delivery up front and incentivising
transparency and rigour around the
true costs of implementation; and
(ii) giving implementers more exibility on
implementation iteration and innovation,
and emphasising eciency and ecacy
for funders by rewarding performance and
outcomes rather than programme inputs.
Photo: BRAC/Alison Wright. BRAC community healthcare promoter, trained by BRAC to be the
village doctor, Uganda, 2014.
“ The Graduation
Approach provides, for
the rst time, a rigorously
vetted, data-driven,
evidence-based
method that has
the potential to
permanently eradicate
ultra-poverty.
The International Policy Centre for Inclusive Growth | Policy in Focus 69
As Figure 1 shows, structuring impact
bonds is a complex multi-step process.
Governments or other outcome payers,
service providers and upfront funders
must collaboratively identify the
outcomes to be prioritised within the
scope of the bond and develop detailed
metrics that will trigger payment when
outcomes are reached, and often must
rely on specialised intermediaries to
assist with this process. The parties
then need to mobilise the capital from
investors to fund the upfront costs of
delivering the services. The service
provider—or recipient of funds—then
needs to implement and execute the
project with a high degree of programme
management and transparent data
to assess milestones, enable course
correction and inform performance
measurement. Moreover, an independent
evaluator must validate and assess the
outcomes against the predetermined
metrics to determine success or failure.
Finally, governments and other outcome
payers must be able to deliver the
capital plus agreed interest for
successful completion.
Since the rst SIB was launched in the UK
in 2010, enthusiasm for impact bonds has
grown. The UK organisation Social Finance
estimates that there are currently 60 live
SIBs that have raised USD216 million in
capital, and 160 more in development,
with most activity occurring in the
USA and the UK, and the majority of the
bonds tackling areas such as workforce
development and homelessness. In
the global development context, the
application has proceeded more slowly.
To date, there have only been two active
impact bonds applied in the development
context: one focused on improving
education for girls in Rajasthan, India,
and another focused on improved
coee production in Peru.10
The Graduation Approach is well
suited to the impact bond model
Despite the slow uptake of impact
bonds in the development context, the
Graduation Approach, which provides
an evidence-based model of service
delivery along specic indicators or
metrics of progress—for example,
household consumption, hygiene,
livelihood development etc.—that
lend themselves to translation into
performance metrics, is ideally suited
to the construction of such a nancing
instrument. Moreover, as a time-bound
“ Structuring impact
bonds is a complex
multi-step process.
FIGURE 1: Social impact bonds
7. Return of principal
plus interest
1. Investment of principal
6. Pay for success
5. Evaluate impact
2. Coordinate, structure deal,
& manage performance
3. Deliver services
4. Achieve outcomes
INVESTORS
INTERMEDIARY
OUTCOME FUNDER SERVICE PROVIDER
POPULATION IN NEED
EVALUATOR
Source: Authors' elaboration.
intervention, the Graduation Approach
already has built into its methodology
clear time parameters for achievement of
the specied performance outcomes that
can be assessed to trigger payment. Thus,
with a graduation programme, investors
would be able to determine the relevant
performance metrics above baseline that
they wish to establish for each of the
indicators (e.g. household consumption,
hygiene, access to clean water, children
of school age enrolled in school, asset
growth) that the intervention addresses
in a given context, rather than having
to craft bespoke measures of impact
or target population. Moreover, they
could determine the overall success of
the programme, triggering repayment
plus interest, within a relatively short
period after investment, since graduation
interventions typically last from 18 to
24 months. This, in turn, lends itself to
reducing the complexity and transaction
costs of designing a SIB.
Additionally, because of extensive
independent RCT evaluation in a
variety of contexts, investors are able
to rely on independent evidence when
setting their risk calculations, instead
of investing in an untested method or
70
provider assertions of success.11 Finally,
thanks to BRAC’s open dissemination
of the model and technical assistance,
there is an increasing number of high-
quality implementers of the Graduation
Approach that have the capacity to
manage and execute the programmes
and work with partners eectively
to execute an impact bond.
Despite strong evidence of its efficacy,
one of the greatest barriers to scaling
the Graduation Approach has been
the programme’s relatively high cost
of implementation. Graduation has
a higher cost per participant than
many traditional poverty alleviation
programmes or simple cash transfers,
due to its ‘big push’ approach. Impact
bonds have the potential to lower this
barrier by aligning private investment
capital to take on the ‘risk ‘of funding
the programmes at greater scale and
enabling government and philanthropic
donors to pay agreed returns only on
delivery of successful outcomes. Because
service providers are incentivised to
maximise performance rather than
specific, rigid programmatic inputs,
impact bond financing could also identify
innovations that reduce costs while
maintaining— or even improving—
outcomes and better addressing the
heterogeneous needs of programme
populations. Implementers could even
receive a ‘performance bonus’ if they
identify efficiencies in delivery that
engender cost savings while maintaining
or improving outcomes.
Obstacles to the development and
implementation of impact bonds
Nonetheless, despite the promise, there
are important cautionary considerations.
Very few bonds have been attempted
in the highly complex context of global
development. The largest one—the
Educate Girls DIB in Rajasthan—deployed
USD267,000 from the UBS Optimus
Foundation to reach approximately
15,000 students in India over three years,
with the Children’s Investment Fund
Foundation (CIFF) serving as the outcome
payer. While the initial outcome metrics
are very promising, with reports that after
one year of the three-year programme
44 per cent of the targeted number of girls
had been enrolled in school, triggering
recoupment of 40 per cent of UBS’s
investment (Basu 2017), the relatively
small scope and scale of the bond, and
the paucity of other examples, may deter
investors from moving into new sectors
of development at greater scale.
Moreover, as exemplied by the
construction of the Educate Girls bond,
these nancing mechanisms are still being
underwritten and powered by private-
sector philanthropic capital, even when
aligned with investment houses such as
UBS. As a result, they remain hybridised
instruments with philanthropic funding
serving as the underwriter, and as such,
the true potential of investment capital
has not yet been demonstrated. There
is reason to believe, however, that as
the approach gains traction, and long-
term social and nancial returns are
Photo: BRAC. Samsunnahar and her children, Bangladesh, 2015.
“ Despite strong
evidence of its efcacy,
one of the greatest
barriers to scaling
the Graduation
Approach has been
the programme’s
relatively high cost of
implementation.
The International Policy Centre for Inclusive Growth | Policy in Focus 71
better quantied, there will be greater
enthusiasm for true capital investment.
Creating impact bonds involves signicant
transaction costs and a high degree of
coordination among a non-traditional
alliance of stakeholders that do not
generally work together, including: the
upfront funder; the service or intervention
provider and implementing organisation;
the outcome payer (government in
the case of a SIB, or foundation or
donor agency in the case of a DIB);
and a specialised intermediary that
can help assess feasibility and facilitate
development of the structure of the SIB,
help cross-sector stakeholders align,
and raise capital (Gustafsson-Wright,
Gardiner, and Putcha 2015). Additionally,
independent evaluators are also likely to
be necessary to verify and validate the
performance metrics triggering repayment.
As a result of this complexity, the time,
eort and resources required to develop a
bond at the outset can be steep even when
the programme implementation it funds
is relatively small in scale, which may deter
investors. Still, there is reason to believe
that as impact bonds are more widely
deployed, transaction costs will decline as
design becomes more standardised and
new models for pooling costs emerge.
The crucial role of data in
enabling innovative nance
for ultra-poverty interventions
Finally, data-driven implementation
will be essential to ensure programmes
funded by impact bonds successfully
reach ultra-poor people. Yet quality data
on programme implementations and
real-time measurements of progress are
still severely lacking due to ineciencies
in data collection by implementers
or lack of standardisation or robust
analytics performed on the data gathered.
However, we believe that technological
innovation has a crucial and exciting role
to play in helping to redress this gap
and unlock pay for performance.
To this end, Uplift has helped develop
a mobile platform called ‘Impact Atlas’
that enables accessible data collection,
real-time analysis of those data and
actionable insights and predictive analytics
that can help improve service delivery
and performance. Impact Atlas equips
eld sta to collect data anywhere, and
to deliver and track a wide range of
interventions, including productive asset
transfer, livelihood and life skills training,
nutrition, hygiene, education and savings.
Additionally, through its sophisticated
functionality, Impact Atlas provides very
thorough targeting surveys, better enabling
clean and validated data to be collected
and veried. Initially designed specically
for the graduation programme, and ‘beta
tested’ with BRAC in Bangladesh, it is now
also being piloted in the Philippines with
Opportunity International. In Bangladesh,
the use of Impact Atlas enables BRAC to
analyse aggregated programme data from
the initial baseline, something it could
not previously do. This allows BRAC to
ascertain the true relevance and impact
of its interventions in shorter time, rather
Photo: BRAC. After graduating from BRAC's programme, Samsunnahar is currently a popular tailor in her village
and grows her own vegetables, Rangpur, Bangladesh, 2015.
“ Data-driven
implementation will
be essential to ensure
programmes funded
by impact bonds
successfully reach
ultra-poor people.
72
organisations, leveraging public
funding alongside private philanthropic
and market-based capital, can make
signicant headway towards meeting the
2030 goal of eradicating extreme poverty
by ensuring that those who are the
furthest behind are no longer left behind.
Instead, they, like so many others who
have progressed out of extreme poverty
in the last few decades, will be enabled to
unlock their human potential—and that
of their children—permanently.
Banerjee, Abhijit, Esther Duo, Nathanael
Goldberg, Dean Karlan, Robert Osei, William
Parienté, Jeremy Shapiro, Bram Thuysbaert,
and Christopher Udry. 2015. “A multifaceted
programme causes lasting progress for
the very poor: Evidence from six countries.”
Science 348(6236). <http://science.
sciencemag.org/content/348/6236/1260799>.
Accessed 6 June 2017.
Basu, Ranajoy. 2017. “#BeBoldforChange:
How impact investors are helping to educate
girls in India.” Ethical Corporation website.
<http://www.ethicalcorp.com/beboldforchange-
how-impact-investors-are-helping-educate-
girls-india>. Accessed 6 June 2017.
Common Fund for Commodities. 2017. “Sustainable
Cocoa and Coee Production.” Common Fund for
Commodities website. <http://common-fund.org/
newprojects/project-overview/project-details/
news/sustatinable-cocoa-and-coee-production/?
tx_news_pi1%5Bcontroller%5D=News&tx_
news_pi1%5Baction%5D=detail&
cHash=8187b417a993f1b4b6ac6efc021ea39a>.
Accessed 6 June 2017.
De Montesquiou, Aude, and Tony Sheldon,
with Frank F. DeGiovanni and Syed M. Hashemi.
2014. From Extreme Poverty to Sustainable
Livelihoods: A Technical Guide to The Graduation
Approach. Washington, DC: Consultative
Group to Assist the Poor and New York:
Food Foundation. <http://www.cgap.org/sites/
default/les/graduation_guide_nal.pdf>.
Accessed 6 June 2017.
Fahey, Allison, and Justin Loiseau. 2016. Ending
Extreme Poverty: New Evidence on the Graduation
Approach. Washington, DC: Consultative Group
to Assist the Poor. <http://www.cgap.org/
blog/ending-extreme-poverty-new-evidence-
graduation-approach>. Accessed 6 June 2017.
Freedland, Nicholas. 2017. “Rationing, not
targeting.” Development Pathways website 11
April 2017. <http://www.developmentpathways.
co.uk/resources/rationing-not-targeting/>.
Accessed 6 June 2017.
Gustafsson-Wright, Emily, Sophie Gardiner,
and Vidya Putcha. 2015. The Potential and
Limitations of Impact Bonds. Lessons from
the First Five Years of Experience Worldwide.
Washington, DC: The Brookings Institution.
<http://www.payforsuccess.org/sites/default/
les/resource-les/impact_bondsweb_1.pdf>.
Accessed 6 June 2017.
Instiglio. 2015. The Educate Girls Development
Impact Bond: A New Finance Model for
International Development. Bogotá: Instiglio.
<http://instiglio.org/educategirlsdib/wp-
content/uploads/2015/09/Educate-Girls-DIB-
Sept-2015.pdf>. Accessed 6 June 2017.
Kristo, Nicholas. 2015. “The Power of Hope
is Real.” The New York Times 21 May 2015.
<https://www.nytimes.com/2015/05/21/
opinion/nicholas-kristof-the-power-of-hope-is-
real.html>. Accessed 6 June 2017.
OECD. 2015. “Final Ocial Development
Assistance Figures in 2015.” Organisation for
Economic Co-operation and Development
website. <http://www.oecd.org/dac/nancing-
sustainable-development/development-
nance-data/nal-oda-2015.htm>.
Accessed 6 June 2017.
Global Impact Investing Network. 2017. Annual
Impact Investor Survey, 7th edition. New York:
Global Impact Investing Network.
Sulaiman, Munshi, Nathanael Goldberg, Dean
Karlan, and Aude de Montesquiou. 2016.
“Eliminating Extreme Poverty: Comparing the
Cost-Eectiveness of Livelihood, Cash Transfer,
and Graduation Approaches.” Access to Finance
Forum No. 11. Washington, DC: Consultative
Group to Assist the Poor. <https://www.poverty-
action.org/sites/default/les/publications/
Forum-Eliminating-Extreme-Poverty-Dec-2016.
pdf>. Accessed 6 June 2017.
The Economist. 2017. “The Economist Explains:
Why the War on Poverty is About to Get
Harder.” The Economist 3 May 2017. <http://
www.economist.com/blogs/economist-
explains/2017/05/economist-explains-1>.
Accessed 6 June 2017.
UNCTAD. 2014a. “Press Release: Developing
countries face $2.5 trillion annual investment
gap in key sustainable development sectors,
UNCTAD report estimates.” United Nations
Conference on Trade and Development website.
<http://unctad.org/en/pages/PressRelease.
aspx?OriginalVersionID=194>. Accessed 6
June 2017. UNCTAD. 2014b. World Investment
Report 2014. Investing in the SDGs: An Action
Plan. Geneva: United Nations Conference on
Trade and Development. <http://unctad.org/en/
PublicationsLibrary/wir2014_en.pdf>.
Accessed 6 June 2017.
UNICEF and World Bank Group. 2016. Ending
Extreme Poverty: A Focus on Children. New
York: UNICEF and Washington, DC: World
Bank. <https://www.unicef.org/publications/
les/Ending_Extreme_Poverty_A_Focus_on_
Children_Oct_2016.pdf>. Accessed 6 June 2017.
World Bank. 2016. Poverty and Shared Prosperity
2016: Taking on Inequality. Washington, DC:
World Bank. <http://www.worldbank.org/en/
publication/poverty-and-shared-prosperity>.
Accessed 6 June 2017.
1. CEO of Uplift.
2. Uplift’s Global Advocate.
3. The World Bank denes the extreme poverty
line as USD1.90 per person per day. While
beyond the scope of this article, there is much
debate about the utility of a USD1.90 a day
benchmark for identifying extreme poverty,
given the multidimensional aspects of poverty
and the variability of relevant indicators
across geographies. Only for simplicity have
we referred to this benchmark in this portion
of our discussion.
than only relying on RCT analysis. In the
Philippines pilot, the platform will be able
to run analytics based on the targeting
surveys to help assess which factors were
more or less eective in identifying the
ultra poor. It will also facilitate an RCT
on a cohort not participating in
the programme, thus enabling more
ecient, credible evaluation.
In other words, Impact Atlas enables
implementers to easily set baselines,
establish performance indicators and
gather real-time data on participant
progress. This real-time insight enables
funders and implementers to continually
assess whether the front-line sta of
implementing organisations are meeting
performance milestones, and to change
course or adapt as needed. Scaling up
the implementation of tools such as
Impact Atlas, and growing the capacity
of implementing organisations to be data-
driven and responsive, will be crucial to
successfully deploy impact bonds at scale,
which in turn will enable the sustained
inow of private-sector investment capital.
Conclusion
We know all too well that poverty is a
systemic challenge that must be addressed
with a systems-level response. It has long
been established that a sustainable and
lasting remedy cannot depend solely on
government and philanthropy but must
be driven by an integrated, collaborative,
multi-sector approach. Though the private
sector has been notably absent to date,
and despite the complexity and obstacles
discussed in this article, we nd good reason
to believe that impact bonds present the
most exciting and promising potential to
help ll the funding gap that has slowed
the progress of ending intergenerational
poverty and to accelerate scaling of
the Graduation Approach, and perhaps
other evidence-based interventions or
adaptations that might emerge. With over 60
implementations globally and 30 ongoing
RCT studies, momentum behind the
Graduation Approach, and concomitantly
the recognition that the ultra poor must
be prioritised within the movement to end
extreme poverty, is growing (Fahey and
Loiseau 2016). What remains is mobilising
the resources to match.
We believe that, together, national
governments and implementing
The International Policy Centre for Inclusive Growth | Policy in Focus 73
4. A champion for ultra-poor people, Uplift
unites a network of visionary, cross-sector
partners and deploys advocacy, innovation
and investment to lift households out of the
most dire forms of poverty by 2030. For more
information, see <www.joinuplift.org>.
5. Denitions of ultra poverty vary, with
some describing the ultra poor as those
in the bottom half of the extreme poverty
demographic, subsisting on USD0.60 to
USD0.70 per day, or those who eat below
80 per cent of their energy requirements,
despite spending at least 80 per cent of their
income on food, with some using the term
to generally refer to those living under a
threshold of USD1.25 per day. Uplift takes as its
baseline the Multidimensional Poverty Index
(MPI) developed by the Oxford Poverty and
Human Development Initiative (OPHI), which
is an international measure of acute poverty
covering 100 countries that complements
traditional income-based measures by
capturing deprivations with respect to
education, health and living standards
<http://www.ophi.org.uk/>. See also de
Montesquiou and Sheldon (2014).
6. See also de Montesquiou and Sheldon (2014).
7. We are aware that the concept of ‘targeting’
ultra-poor people with respect to crafting
interventions and policies has been the subject
of some debate. See, for example, Freedland
(2017). Engaging in that debate is beyond the
scope of the purposes of this article. However,
based on the earlier discussion of the unique and
particularly dicult nature of ultra poverty—and
its multidimensional components that encompass
both social and nancial exclusion factors—we
believe it is critical to craft programmes that
specically identify and address the needs of
ultra-poor people. Indeed, as discussed, the failure
to do so has meant that ultra-poor people have
been left behind, despite the progress made
in eradicating extreme poverty, and are not
reached by existing social protection, livelihood
development and market-based mechanisms.
Moreover, the promotion of alternative
mechanisms to avoid exclusion errors that
have been found in targeting, such as universal
basic income schemes, is not yet supported
by evidence that universal basic income alone
would demonstrate eectiveness in addressing
the multidimensional nature of ultra poverty or
that the benets would be durable over time for
ultra-poor people. Indeed, recent analysis on cash
transfer programmes as compared to graduation
approaches suggests that there is little or no
evidence that cash alone sustains benets beyond
the duration of the intervention. See, Sulaiman,
Goldberg, Karlan, and de Montesquiou (2016).
This in no way is to diminish the importance of
improving the accuracy of targeting or protecting
against its shortcomings in programme design
and policy that supports linkages to other social
safety nets wherever possible. Rather, we believe
that the emphasis should not be on dispensing
with eorts to identify and target ultra-poor
people, but improvement.
8. See Banerjee et al. (2015).
9. Development impact bonds (DIBs) are an
adaptation of the social impact bond (SIB)
structure to the global development context.
The primary distinctions between the two
mechanisms are that SIBs have been deployed
in developed countries, and with DIBs, the
outcome payer is generally a foundation or
donor agency instead of the government.
For simplicity, we primarily use the phrase
‘impact bond’ to describe the instrument.
10. See Instiglio (2015) and Common Fund
for Commodities (2017).
11. See, for example, Banerjee et al. (2015)
and Kristo (2015). To the extent that critics
of the Graduation Approach have questioned
whether or not the claims made on its behalf
should be challenged or are potentially not as
extensive as claimed, we do not believe the
debate would or should hinder the relevance
of graduation’s strong, independently
reviewed evidence base for the purposes
of an investor’s risk assessment for a SIB.
Indeed, even with the debate, the evidence for
signicant improvement on relevant indicators
of ultra poverty is better established than
many, if not most, development interventions.
Moreover, the credibility and diversity of
the independent economists involved in the
evaluations, combined with the signicant
increase in the number of graduation pilots
around the world resulting from those
evaluations, should persuade an upfront
funder that graduation has been suciently
vetted on outcomes to merit investment.
“ Poverty is a systemic
challenge that must
be addressed with a
systems-level response.
Photo: BRAC. After being forced to beg for food, Shohiton received livestock and training from BRAC's ultra poor
programme and is now a condent entrepreneur, Rangpur, Bangladesh, 2015.
The Graduation Approach in isolation, without income generation, does not
achieve the stated goals of the programme, and income generation by
itself is also insufcient.
Nathanael Goldberg
The claims made about the success of Graduation programmes are both
misleading and exaggerated, since they give the impression that impacts
are much greater than they actually are.
Stephen Kidd and Diloá Bailey-Athias
The Graduation Approach is expected to continue to grow in scale and
inuence, with strong demand from donors and governments for
nationally scaled programmes.
Aude de Montesquiou and Syed M. Hashemi
International Policy Centre for Inclusive Growth (IPC-IG)
SBS, Quadra 1, Bloco J, Ed. BNDES, 13º andar
70076-900 Brasília, DF - Brazil
Telephone: +55 61 2105 5000
ipc@ipc-undp.org www.ipcig.org
© 2017 International Policy Centre for Inclusive Growth
ISSN: 2318-8995
