Tag Archives: Stephen Devereux

Dependency and Graduation

A recent research brief, written by Stephen Devereux for the Regional Hunger and Vulnerability Programme (RHVP), looks at two issues that keep cropping up in debates about social protection:  ‘dependency’ and ‘graduation’. Both issues are commonly raised by governments and donors that are sceptical about making firm, long-term commitments to social transfer programmes.

Often, skeptics talk about ‘dependency syndrome’ – a tendency for recipients of regular social transfers to become permanently reliant on these ‘handouts’, and to lose any incentive to improve their circumstances using their own initiative and resources.

‘Graduation’ is often presented as an antidote to dependency. The argument is that financial assistance to poor individuals and families in distress should be limited in scale and time in order to avoid the ‘dependency trap’, and complementary programmes should be put in place to ensure that beneficiaries are able to ‘graduate’ from ‘handouts’ and become self-reliant.

In his paper, Devereux argues that these concerns about the risks of social transfers are based on generalisations and flawed thinking. He provides evidence from several studies and real-life examples of social protection programmes to support his view.


Firstly, all societies accept that certain members must be supported – young children, older persons, the chronically ill and severely disabled. So ‘dependency’ must differentiate between those who are genuinely unable to care for themselves, and those who need temporary support but can be assisted to become self-reliant in the future.

Secondly, the concern to avoid dependency confuses two quite separate roles for social protection:

  • temporary support for economically active people who become vulnerable due to a livelihood shock or humanitarian emergency (e.g. unemployment insurance for workers who are retrenched, or famine relief for drought-affected farmers); and
  • continual support to those people who are not able to work to provide for themselves (e.g. infirm elderly widows living alone with no support).

While people in the first category clearly have the potential to ‘graduate’ once their circumstances improve (e.g. if they find a job, or when the next harvest comes in), people in the second category have much lower prospects of graduating – and should not be expected to do so.

Several social protection programmes in Africa do recognise this fundamental distinction, by combining different types of support:

  • cash or food transfers with a labour requirement for eligible people who are able to work; and
  • unconditional cash or food transfers for eligible people who are unable to work.

It is important also to emphasise that while many social protection programmes may seem to create dependency on the state in the short-term, they may actually reduce dependency on the state in the long-term.

For example, child benefits, school feeding and conditional cash transfers, all provide immediate relief to poor households, but they also enhance children’s health, nutrition and access to education. This provides big potential returns to the national economy in the next generation.

Food aid and dependency

It is widely believed that food aid damages African agriculture because it removes incentives for production. The argument is that large imports of free or subsidised food aid depress local food prices and create disincentives to farmers to produce for the market, so that production falls and dependence on further food imports increases. The distribution of food aid also stifles the development of a competitive commercial grain marketing sector.

Recently, a panel survey in Ethiopia tested this version of the dependency argument across 1,470 households. A comparison of food aid recipients and non-recipients found that recipients pursued a more diversified portfolio of livelihood activities, and were more likely to provide assistance to others.

The survey found no significant differences in social and economic behaviour between households who receive food relief and those who do not.

Even more striking, a cross-country analysis found that food aid was associated with an increase, rather than a decline in national food production.

Cash transfers and dependency

Economic theory suggests that people who receive regular free transfers will be discouraged from seeking work, especially if the value of the transfer is close to the income that the recipient could expect to earn from paid employment.

The ‘dependency trap’ refers to a situation where transfer recipients have no incentive to take steps to stop receiving these transfers (such as finding work) – or even worse, when employed people leave work and ‘choose leisure’ instead, preferring to survive on state benefits.

These fears about ‘dependency syndrome’ are common in affluent northern welfare states. But is there evidence of similar behaviour by poor recipients of social transfers in African countries?

On the contrary, Devereux says there is intriguing evidence from southern Africa that social grants have precisely the opposite effect.

In South Africa, a study of the Child Support Grant and Old Age Pension found that adults living in recipient households were more likely to seek work, and more likely to find work, than people in similarly poor households that do not receive these grants.

The authors of that study speculate that this is because social grants give beneficiaries the resources and economic security they need to invest more in job searching – by spending some of this cash on transport costs, child care, and so on.

Similar findings have been reported from large-scale conditional cash transfer programmes in Latin America.

In Namibia, sceptics greeted the introduction of a pilot Basic Income Grant in one community with assertions that this would convert the local population into idle drunkards. What actually happened was strikingly different (for details, see: Cash Transfers Reduce Dependency in Namibia).


The concept of ‘graduation’ describes a process whereby recipients of cash or food transfers ‘graduate’ from depending on external assistance, to a condition where they no longer need these transfers, and can therefore exit the programme.

Though it might seem a simple idea, graduation is extremely difficult to define and implement. Graduation is not only about attaining a certain level of food consumption or cash income. Sustainable graduation means that recipients must be able to:

  • generate enough future food and income;
  • be resilient against future shocks.

How does one then assess whether and when a household is self-reliant and resilient, and no longer needs social assistance? This requires defining graduation criteria and setting thresholds that participating households must reach.

The notion of graduation suggests steady progression up an income scale – but livelihoods in rural Africa are often unpredictable and uncertain. Farming communities and pastoralists face unpredictable weather and other threats to crop harvests and livestock herds.

Even if a household appears to have passed an income or asset threshold at a point in time, it is impossible to predict whether the household is about to suffer a major crisis (e.g. a drought or disease outbreak) that will decimate its harvest or livestock herd, leaving the household acutely vulnerable to hunger, destitution and even death.

Evidence on graduation

Two ongoing social protection programmes in Africa that have grappled with defining and implementing graduation operationally are the Productive Safety Net Programme (PSNP) in Ethiopia, and the Vision 2020 Umurenge Programme (VUP) in Rwanda.

Both case studies reveal the complexity of graduation: the difficulty of establishing realistic graduation criteria (especially incorporating indicators of resilience), the incentives that programme managers face to graduate participants prematurely (and the risks of doing so), and above all, the necessity to acknowledge that many recipients of social transfers have no prospect of graduating, and will need permanent assistance.


Evidence from many African countries challenges the notion that social transfers induce ‘dependency syndrome’.

First, the social protection that does exist in much of Africa provides very limited benefits, and is often unreliable – so that recipients cannot rely on it to predictably cover all their food and income needs.

Second, even when social transfers are more predictable, continual and generous, the overwhelming evidence is that recipients will use the grants they receive to buy more assets, invest in small businesses, or enter the job market. This in itself could be enough to lift social transfer recipients out of extreme poverty and towards self-reliance.

‘Graduation’ is a concern for social protection programming, because policy-makers strive to avoid ‘dependency’ and because financial constraints and donor funding cycles mean that social transfer programmes often have limited budgets and timeframes.

Experts have struggled to work out what graduation might mean in practice, and many eligible households have little prospect of sustainable graduation, either because of their own circumstances or because of the challenging environment in which they struggle to make a livelihood.

Devereux believes then, that social protection policymakers should perhaps worry less about minimising dependency and maximising graduation, and focus their attention more on building effective social assistance and social insurance systems.

You can download the full brief in PDF form here:

1 Comment

Filed under Uncategorized