SOCIAL PROTECTION IN AFRICA:
WHERE NEXT? 1
Introduction
Has social protection in sub-Saharan Africa lost its way? There have been dramatic and
impressive achievements in the past decade, but a review of country experiences reveals that
there is still a long way to go in many countries, while in others progress appears to have stalled,
with no clear consensus among domestic and external stakeholders on the way forward. Yet
social protection remains central to correcting the inequalities created by skewed development
processes of recent decades, and it is at the heart of our main development challenge for the
early decades of the 21st century: how to spread the benefits of growth to include the poorest?
The 2002/03 humanitarian crisis in southern Africa was the trigger for a big push by the
international community on the then emerging social protection policy agenda, which has since
gone through two broad phases. In the first phase (roughly 2002-05), the strategic objective of
certain donors to prevent food aid from swamping the region was successfully achieved, social
protection was established on the policy agenda in many countries, some governments initiated
national social pension schemes or ‘productive safety net’ programmes, and donors invested
heavily in building the evidence base for social transfers, and in enhancing the capacity of
government partners to deliver social protection programmes.
The second phase (2006-10) has been more erratic. Many countries now have social protection
strategies; ‘emergency cash transfers’ have displaced or complemented food aid in food crises;
some large-scale programmes have been introduced while others have been consolidated. But in
other countries pilot social transfer projects have not scaled up and face uncertain futures, while
several African governments remain resistant to institutionalising national-level social transfers,
often revealing a strong preference for input subsidies and public works instead.
As for the development community, we cannot agree among ourselves on how to move the
social protection agenda forward – some take a ‘rights-based’ perspective while others want to
implement a needs-based ‘social floor’; some favour unconditional cash transfers while others
want to introduce conditions; some argue for social pensions while others target ‘orphans and
vulnerable children’ (OVC) or the poorest 10%; and in some countries different approaches are
being ’piloted’ in different districts simultaneously, not informed by any national vision.
Hence the question posed by this paper: where next for social protection in Africa?
1
This discussion paper, published on 1 June 2010, was co-produced by: the Centre for Social Protection
(CSP) at the Institute of Development Studies (IDS), the Social Protection Programme at the Overseas
Development Institute (ODI), the School of International Development at the University of East Anglia
(UEA-DEV), and the Regional Hunger & Vulnerability Programme (RHVP). It is the product of a meeting
hosted by RHVP in April 2010, to discuss the state of social protection in Africa and brainstorm ideas
for moving forward. Participants were: Stephen Devereux (corresponding author) and Mark Davies
(IDS/CSP), Anna McCord and Rachel Slater (ODI), Nicholas Freeland (RHVP), Frank Ellis and Philip White
(UEA-DEV/RHVP). Additional inputs were provided by Rachel Sabates-Wheeler (IDS/CSP).
Page 1 of 9
SOCIAL PROTECTION IN AFRICA: WHERE NEXT?
Current Approaches
Social protection is a relatively new policy arena within development cooperation, having been
conceived by the donor community in the late 1990s.2 During the past decade, donors and their
development partners (among whom we include ourselves) have pursued several approaches to
engage with African governments around developing national social protection programmes.
1) Financing pilots and demonstration projects: Initially – at least since GTZ first sponsored a
‘social cash transfer’ project reaching 1,027 households in Kalomo District, Zambia in 2004 –
donors assumed that a well designed and well implemented social protection intervention
that generates positive impacts would persuade governments to take over the management
and financing of these programmes and scale them up to national level. Yet, despite a
proliferation of similar small-scale externally resourced cash transfer projects across Africa
since 2004, we know of no cases where such a pilot or demonstration project has become a
government-owned social protection programme with national coverage.
2) Building the evidence base: Sizeable investments have been made in analysing the impacts
of social protection programmes – for example, RHVP’s Regional Evidence-Building Agenda
(REBA) in southern Africa, panel surveys led by IFPRI and IDS of the Productive Safety Nets
Programme (PSNP) in Ethiopia, and EPRI’s impact evaluations of social grants in South Africa.
While the evidence for multiple positive impacts of social transfers is compelling, the failure
of many governments to act on this evidence confirms that policy choices are complex
political processes, requiring much more than the presentation of research findings through
reports and stakeholder workshops.
3) Building the capacity of implementing ministries: Development partners have provided
extensive technical assistance (TA) to ministries responsible for social protection, including
seconding experts into these ministries, sending officials on training courses and study tours
to observe social protection programmes in other countries, and hiring consultants to help
draft National Social Protection Strategies. In most cases, this capacity-building has had very
limited strategic impact, because these ministries remain politically weak, with small budgets
and high staff turnover, and negligible influence over Ministries of Finance and Planning.
4) Mobilising civil society: International and local non-governmental, community-based and
faith-based organisations are often excluded by national governments from social protection
policy formulation, although these civil society actors are frequently involved in programme
implementation and delivery. Initiatives to engage civil society more strategically have
included internationally-supported consultations around the African Union’s Social Policy
Framework, the formation of the Africa Civil Society Platform on Social Protection, and the
work of the Grow Up Free From Poverty (GUFFP) coalition of international NGOs working on
social protection. Despite these efforts, most governments and development partners
continue to treat civil society as junior partners or subcontracted service providers, and
there are few success stories to report from Africa of effective mass mobilisation or popular
lobbying around social protection at national level.
2
For our purposes, social protection is understood as a specific set of policy instruments devised by the
donor community in the late 1990s, with the objectives of alleviating poverty and managing risk in
poor and vulnerable communities of the global South. Although it can be defined very broadly, social
protection in sub-Saharan Africa has in practice – as operationalised by donors and international NGOs
– been dominated by unconditional or conditional cash transfers delivered to targeted poor and
vulnerable groups, and by public works programmes for the working-age poor.
Page 2 of 9
SOCIAL PROTECTION IN AFRICA: WHERE NEXT?
5) Engaging directly with Ministries of Finance: The recognition that real power over social
protection programming lies with budget decision-makers has prompted efforts to influence
Finance and Planning Ministries directly. This is challenging, because the social protection
lobby has no entry point to these ministries and little credibility with macroeconomists, who
demand evidence not only that social transfers can generate poverty reduction, but also that
they stimulate economic growth and are the most cost-effective use of public resources for
this purpose – compared, say, with agricultural input subsidies – which is a more difficult
case to prove.
6) Lobbying parliamentarians: Efforts to influence parliamentarians (for example through
special seminars, such as the SADC Parliamentary Forum initiative) have built on evidence
from several countries that social protection can be electorally popular. While these events
usually elicit enthusiasm from MPs, this does not necessarily translate into more informed
political debate at national or regional level, nor into greater political will by governments to
implement national social protection programmes.
Despite significant investments of money, technical expertise and energy in these attempts to
convince African governments to adopt particular models of social protection, success to date
has been patchy. Several governments have – with or without donor support – introduced
significant social protection programmes: examples include the PSNP in Ethiopia, Lesotho’s old
age pension, and the OVC programme in Kenya. But many others (e.g. Zambia and Malawi) have
not yet been convinced by the array of demonstration projects, empirical evidence, study tours,
training workshops and high-level seminars which development partners have provided.
Commonly heard explanations for this reluctance include government alarm over the political
irreversibility and high recurrent costs of permanent social transfer programmes (which their
development partners insist are nonetheless affordable), fears of creating ‘dependency’ among
transfer recipients (which development partners argue are misplaced), and governments’ belief
that scarce public resources should be channelled into employment creation programmes or
directly ‘productive’ sectors such as agriculture, which will contribute to national growth.
This paper turns the focus round, onto outsiders like us who are engaged with social protection
in various roles (as donors, consultants, researchers, NGO workers), and asks – how can we do a
better job of helping to advance the social protection agenda in Africa?
Analysis
What is holding social protection back in Africa? How can we move forward? The perception of
development partners is that many African governments are reluctant to adopt a vision of social
protection that might have been externally defined but is demonstrably ‘a good thing’ – but this
might be the wrong starting point. We should look critically at ourselves – as external actors
implicated in advocating, designing and evaluating social protection interventions. So perhaps
the question should be reformulated: What’s wrong with the way outsiders are engaging in
social protection in Africa?
(1) Externally-driven social protection projects have little domestic traction
Donors and NGOs have favoured the ‘pilot project’ route for introducing social protection
(especially cash transfers) into domestic policy agendas, but with disappointingly low uptake
by African governments. Reasons for this include: (1) Pilot projects are implemented in a
way that is not representative of how they would be implemented if national in scale, with
far heavier commitments of technical and financial resources per programme participant.
Page 3 of 9
SOCIAL PROTECTION IN AFRICA: WHERE NEXT?
(2) They are nurtured and supported in order to succeed rather than to fail – so their useful
demonstration effects are limited. (3) Such projects are rarely fully integrated into local or
national policy processes, so they generate little political traction.
External actors have a tendency to act as champions for particular instruments and target
groups, depending on: (1) their institutional mandates and ideological preferences (some
donors favour conditional cash transfers, but others prefer unconditional grants); (2) what
resources they have to disburse (e.g. food aid); (3) what lines of financing they can access
(e.g. availability of resources for HIV and AIDS has skewed the provision of social transfers to
AIDS-affected groups rather than those affected by chronic poverty); or (4) which vulnerable
groups they are mandated to support (some external actors lobby for pensions for older
persons, while others sponsor child grants for OVC). The result can be confusion and mixed
messages, as donors and UN agencies push governments in different directions.
(2) Government programmes and preferences are given insufficient prominence
Often governments have their own programmes and priorities for reducing poverty and
vulnerability – and would not necessarily have chosen the programmes advocated by their
development partners, if they had had a free choice and financial support. But government
preferences are paid insufficient attention by external actors who have different visions and
favoured models of social protection. The governments of Lesotho and Swaziland decided
to implement national social pension schemes (in 2004 and 2005 respectively), against the
advice of donors and with no external financial or technical support. The governments of
Malawi and Zambia reversed donor pressure (acted on in the 1990s) to stop subsidising
farmers’ access to seeds and fertiliser, and reintroduced agricultural input subsidies in the
mid-2000s as their response to food insecurity, with mixed but generally positive results – in
Malawi, donors who were initially sceptical supported these subsidies in their second year.
Meanwhile, in each of these four countries, donors and NGOs have sponsored small-scale
cash transfer projects, which, unsurprisingly, interest the respective governments less than
their own social protection initiatives. What lessons do we (as outsiders) take from this? If
development partners are pushing models that governments don’t accept, could it be that
those models are inappropriate in terms of national policy preferences? Could we be failing
to understand and engage with the political economy of social protection policy formulation
in each country? Governments often take a production-oriented view of social protection,
informed primarily by concerns about national food security, and they must heed their
electoral cycles and political constituencies. Donors and NGOs, on the other hand, are
constrained by their own project cycles, and tend to respond to humanitarian imperatives
and MDG commitments. There is no guarantee of common ground between these agendas.
(3) Standardised imported approaches may be inappropriate
Many social protection concepts that the international community is introducing to Africa
and elsewhere derive from West European notions of social welfare, which may not be well
adapted to the specific vulnerabilities faced by poor people living in very different contexts.
For instance, the international community is starting to converge on a standardised set of
social protection instruments, to be advocated for adoption globally. This agenda is led by
the UN’s ‘social protection floor’ initiative which builds on the ILO ‘social minimum’ –
comprising social pensions, child benefits, access to health care, and unemployment
provision. Even if we accept that ‘minimums’ are enough, institutionalising this menu will
face challenges from a government perspective. (1) It is an undifferentiated blueprint – the
main concession to differences in domestic capacities is the timetable for establishing the
Page 4 of 9
SOCIAL PROTECTION IN AFRICA: WHERE NEXT?
complete set of instruments (a version of ‘progressive realisation’). (2) It is driven by generic
instruments rather than an assessment of local vulnerabilities and prioritisation of needs.
(3) Although structural unemployment is the main source of poverty in most low-income
countries, inadequate attention has been paid to the working-age poor in most social
protection toolkits.
(4) Too little attention has been paid to vulnerability
Social protection should be focused on reducing vulnerability as well as alleviating chronic
poverty directly, but the models espoused by development partners are often justified only
in terms of their poverty reduction impacts, and there are too few tools for managing risk,
reducing economic and social vulnerabilities, and building resilience. Although there is some
experience with ‘social insurance’ schemes such as weather-indexed crop insurance, futures
markets, warehouse receipts and food price stabilisation measures, these are limited and
remain marginal within the social protection discourse. This is not only a significant gap in
the policy agenda, it is a diversion of social protection away from its original intention and
mandate that needs to be rectified.
(5) Participants are not involved in social protection programming
Social protection is a pro-poor policy agenda that has been devised, designed and delivered
above the heads of the people it intends to benefit, with very little consultation on project
selection, design choices, and implementation modalities. This violates the principles of
participation that have become widely accepted and well established in most development
programming. Sometimes communities are asked to participate in community-based
targeting exercises to identify eligible individuals or households, but usually following
eligibility criteria that have been pre-defined by the programme.
Yet when ‘beneficiaries’ are asked for their feedback they often express very divergent
views, which are not always consistent with external designs – for instance, cash transfers
are not universally preferred to food aid (preferences vary by gender, distance from
markets, seasons, and so on). Moreover, it is not uncommon for direct recipients to share
their transfers within communities, reflecting a preference for wider distribution than is
stipulated by targeting criteria – even at the cost of lower benefits for themselves. This
presents a challenge to externally designed programmes that aim to distribute resources to
only targeted subsections of the poor (e.g. ‘the bottom 10%’, when the bottom 10% may
not be significantly poorer than those in the next 20-30%). Rolling out blueprint models like
the ‘social floor’ across diverse contexts risks perpetuating a ‘top-down’, non-participatory
approach to social protection.
Future Options
What future pathways can be identified for promoting social protection in Africa? Five options
are proposed here for development partners to pursue: (1) learn from government-driven
programmes and work with governments to monitor, evaluate, improve and extend them;
(2) work through appropriate institutional mechanisms, at regional and national levels; (3) learn
lessons for national implementation, rather than from pilot projects; (4) be driven less by
instruments and locate social transfers within a broader national social policy agenda, by paying
more attention to social protection objectives – including reducing vulnerability; and (5) find
new levers for supporting African governments, based on a more sophisticated understanding of
the political economy in each national context.
Page 5 of 9
SOCIAL PROTECTION IN AFRICA: WHERE NEXT?
1. Start from government programmes
Following from the above, we need to understand better why governments have particular
preferences and why there is more energy and enthusiasm for certain programmes rather
than others. ‘Lack of capacity’, ‘lack of resources’, and ‘lack of political will’ are often given as
explanations for governments’ failure to adopt recommended social protection programmes,
but there are usually good reasons for this lack of motivation: where political incentives
exist, governments can and will take action. In cases where governments have introduced
social protection interventions on their own initiative (e.g. social pensions in Lesotho and
Swaziland, or agricultural subsidies in Malawi and Zambia), observers need to learn: (1) why
they chose these instruments; (2) why they succeeded; (3) how donors can support them; or
(4) whether outsiders should leave such programmes alone and not interfere.
The principle is to understand the political drivers of policy processes – find out where the
energy is, and build on this. Countries should be encouraged to define their own domestic
social protection agendas. If governments insist on input subsidies rather than cash transfers
(for instance), an open-ended engagement is more appropriate than dismissing this as
misguided, and setting up parallel projects that are only loosely connected with government.
Since governments often depend on donors, directly or indirectly – through General Budget
Support – to finance their programmes, donors can legitimately ask governments whether
specific programmes are the most cost-effective use of resources to achieve mutually agreed
outcomes. If there are glaring gaps in the national vision for social protection, development
partners can identify these gaps and bring them to government’s attention. This suggests a
dual role for donors and NGOs: (1) assist governments to implement their own programmes,
by investing in TA and M&E to design and evaluate existing or preferred government social
protection interventions to maximise their effectiveness; and (2) work with governments to
supplement these programmes with others that reach excluded people whom donors and
NGOs feel should be reached, or that achieve different social protection objectives (that the
donors and NGOs believe are important).
2. Work through appropriate institutions
Social protection is predominantly a national, not a regional, issue – reflecting as it does
national political, economic and social imperatives. Regional institutions like SADC, IGAD and
the AU have important roles to play in defining a vision and in monitoring progress towards
agreed targets, but they do not provide strong leadership and in reality they have limited
influence over, or capacity to support, national policy-making processes. Donor-sponsored
regional initiatives, such as RHVP, can disseminate lessons across countries and host regional
events, but these agencies sit outside implementing institutions and have only indirect
influence over policy processes. Once regional institutions have produced their overarching
guidelines and ‘policy frameworks’, donors should focus their attention and efforts on design
and implementation within the framework of national policy objectives.
Donors are central actors in the broader institutional context, and have substantial influence
over the social protection policy process, due to their financial and technical resources. A key
challenge is the lack of donor harmonisation at country level, which needs to be addressed
by national initiatives to prevent donor competition around alternative approaches and
instruments, and to rationalise programming and support to governments. It is important to
note, however, that donor harmonisation is not in itself sufficient to promote effective social
protection programming. The ‘social floor’ initiative outlined above is an example of global
donor harmonisation around an agreed agenda, but is subject to many limitations – also
noted above – not least the risk of converging on a limited, pre-defined set of instruments
Page 6 of 9
SOCIAL PROTECTION IN AFRICA: WHERE NEXT?
that has not been generated through local consultations and may not be consistent with
national policies and priorities.
At the national level, donors often intervene directly in the social protection policy process,
but without providing support to existing institutions that already deliver social protection.
In Zambia, donors chose to initiate a cash transfer pilot project in a single district, using a
community-based targeting system developed by the long-standing national Public Welfare
Assistance Scheme (PWAS), which has an institutional presence in almost every district but is
cash-strapped. Given the challenges subsequently faced in scaling up the pilot project to a
countrywide nationally owned programme, donors might have been better advised to have
invested instead in strengthening the capacity of PWAS to fulfil its mandate.
3. Learn lessons for national implementation, not from pilot projects
A great deal of donor and government money has been spent on pilot social transfer projects
– advocacy, design, delivery, evaluation, training – and many of these interventions have
achieved positive impacts on transfer recipients, their households and local communities.
Lessons from these initiatives should be captured and integrated into the process of national
capacity building and policy development.
But in future pilot projects should be reserved for testing specific administrative, process or
implementation modalities (e.g. electronic delivery mechanisms) in specific contexts – rather
than for illustrating that certain approaches can have an impact on poverty. Even then, they
should be initiated only if there is already a commitment to the approach they embody, a
genuine intention to institutionalise the scheme, and a credible roll-out plan for scaling it up.
Donors should critically assess existing pilots in this light: if there is no clear government
commitment to scale them up to national programmes, then they should be phased out –
there should be no shame in a pilot failing.
What is needed now is not more evidence of impact from small experiments: it is practical
experience that will inform implementation at national scale. There is a need to stimulate
open national debate around the key constraints to adoption of large-scale social protection
programmes: specifically, government concerns about fiscal sustainability and dependency.
Many African governments are unconvinced that large-scale social grants or employment
‘guarantees’ are fiscally affordable, and they remain concerned that such programmes will
create unnecessary and damaging ‘dependence’ on the state. There is a need to resolve
these issues by promoting an informed national debate drawing on relevant evidence from
Africa and elsewhere (e.g. South Africa, Latin America, India), while taking domestic political
economy considerations into account.
Also required is learning from the experience of existing large-scale government initiatives,
and building the capacity of governments to design and deliver effective national social
transfer programmes. This includes training at national and – crucially – sub-national levels,
and the provision of technical and financial assistance in key areas of high initial expenditure
such as establishing systems for identification, delivery and M&E.
4. Encourage a broader social protection agenda
Continuing social protection programming on a ‘business as usual’ basis is not adequate: the
current dominance of social cash transfers and public works represents too limited a toolbox.
In the drive to introduce such programmes across Africa, the broader objectives and range of
instruments of social protection appear to have been overlooked. In particular, the fact that
social protection evolved out of the humanitarian relief, ‘safety nets’ and food security
Page 7 of 9
SOCIAL PROTECTION IN AFRICA: WHERE NEXT?
agendas seems to have been forgotten – social protection was a response to vulnerability,
not just to poverty. As a guiding principle, social protection programming should start from
national policy objectives, not from available instruments or favoured target groups. So if the
ultimate objective is a genuine reduction of vulnerability, rather than just poverty alleviation,
then a range of different instruments that address vulnerability needs to be introduced, to
complement cash transfers – with or without a work requirement – to the chronically poor.
Vulnerability is an elusive concept that is defined in various ways for different purposes; it
needs to be clearly conceptualised and operationalised. Chronic poverty and vulnerability
overlap and reinforce each other, but they are not the same, and they require different
responses. Vulnerability is a function of exposure and susceptibility to hazards, both of which
are intensified by poverty. In operational terms, vulnerability requires policies that reduce or
manage risk, while chronic poverty requires policies that provide a minimum acceptable level
of consumption for people who are persistently unable to achieve this, given their extreme
lack of resources and opportunities to make a living.
The policy challenge for the future is to achieve a much better understanding (located in
political and social contexts) of the impact of cash transfers of different levels provided to
vulnerable and chronically poor people. Then the costs and outcomes – in terms of both
poverty and vulnerability – can be better weighed in the balance against other risk reduction
policies, such as food price stabilisation measures. It must also be acknowledged that the
popularity of cash transfers has been somewhat undermined by the recent global food price
crisis, which has contributed to a resurgence in food aid. Appropriate roles for cash and food
transfers need to be redefined in different contexts.
This broader vision of social protection would not simply transfer cash to the chronically
poor, but also sees cash transfers as one way of tackling chronic vulnerability, in terms of
which trade-offs between approaches (cash, input subsidies, public works, price stabilisation,
asset transfers) should be properly examined. This broader vision should also incorporate
measures that combat exploitation and marginalisation (e.g. upholding the rights of workers,
women, children, people affected by HIV and AIDS, and ethnic minorities), to the extent that
these barriers constrain their access to sustainable livelihoods and/or to social protection.
5. Look to new levers of accountability
Existing approaches should continue – building government capacities, generating evidence
and disseminating lessons, engaging with Finance ministries, educating MPs about social
protection – but an improved political economy analysis and understanding is needed in
each country, to identify the ‘drivers’ of social protection policy processes (key stakeholders,
government priorities and constraints, informal as well as formal institutions, complex
patterns of allegiance and competition involving domestic and external actors). This analysis
should also uncover innovative mechanisms for supporting policy-making – e.g. recognising
ways in which greater accountability to the poor and vulnerable can emerge both from an
active and mobilised civil society and from broadly targeted social protection itself,
supported by more creative use of mainstream and electronic media for dissemination and
advocacy purposes, or involving the private sector in aspects of programme financing and
delivery where they have convergent interests, commercial synergies and comparative
advantage.
Instead of lobbying for particular social protection instruments, development partners
should first support governments to develop a national vision for social protection, and then
translate that vision into appropriate strategies, policies and programmes. This requires
Page 8 of 9
SOCIAL PROTECTION IN AFRICA: WHERE NEXT?
working closely with politicians, technocrats, implementing agencies, civil society actors,
programme participants and other stakeholders – not imposing an externally defined
agenda, but building local capacity to design and deliver nationally owned social protection
systems. The outcome should be a more organic and better grounded approach to social
protection, one that is better aligned to the political realities in each country and is driven
forward by the most effective domestic levers of influence, not by outsiders.
Principles
Following from the analysis above, ten guiding principles are proposed for future engagement of
development partners with national social protection policy processes in Africa.
1. Recognise the importance of social protection: Social protection remains a vital tool for
achieving inclusive growth in sub-Saharan Africa.
2. Support national policy priorities: First identify the national vision for social protection, then
design interventions around those objectives, starting from what is already in place.
3. Minimise policy intrusion: Ensure that externally-supported programmes do not promote
the preferences of development partners over those of domestic actors, including in terms
of selected instruments or favoured target groups.
4. Rationalise donor support: Harmonise donor activities to prevent competition, reduce
inefficiencies, and encourage the pooling of resources.
5. Encompass a diversity of approaches: Instead of importing standardised models, build
national social protection systems based on local analysis, political preferences, capacities
and prioritisation of needs.
6. Focus on vulnerability: Social protection should focus on addressing poverty and reducing
vulnerability, based on local perceptions of the nature of vulnerability.
7. Limit pilot projects: There is little justification for more social protection experiments in
Africa – the imperative now is to take lessons learnt and apply them for effective delivery at
scale.
8. Find new levers of support: Work more closely with Ministries of Finance, parliamentarians,
the private sector, domestic civil societies, and local mainstream and electronic media.
9. Involve participants: Engage social protection participants in vulnerability assessments,
programme selection, design choices, delivery, and M&E.
10. Focus on outcomes: recognise that social protection is not an end in itself (numbers of
people covered by social protection), but rather a means to an end (reduced poverty and
vulnerability).
Page 9 of 9