Are social protection programmes child-sensitive?

Abstract

There is no doubt that child focus in the social protection agenda makes development and economic sense, yet child-sensitive social protection still remains elusive in some African country programmes. The case study of the Productive Safety Net Programme in Ethiopia discussed in this paper shows that the child-conditioned component in both the design and the implementation of this huge social protection programme is largely absent. Child-sensitive social programming, which discretely improves children's schooling and access to basic health care services and protects them from child labour, is recommended, with prioritisation of child labour saving assets as one of the key interventions.

Keywords:

• child labour
• child-sensitive social protection
• Ethiopia
• social protection
• Productive Safety Net Programme

1. Introduction

This research note demonstrates the negative and unintended outcomes of a well-intended social protection programme in Ethiopia, the Productive Safety Net Programme (PSNP), which was started by the government and its development partners in 2005 to graduate poor households out of poverty. A close scrutiny of the programme shows that it lacks child focus and sensitivity in its design and implementation. This is clearly observable in the larger labour-based conditional transfers and discretionary quotas on unconditional transfers embedded in the programme, which provides no guidelines on protecting children and ensuring they derive the most opportune benefit from the programme. We argue that the PSNP has had a number of indirect adverse effects on child labour and schooling. We also provide a conceptual background on child-sensitive social programming (CSSP). This is followed by a description and critique of Ethiopia's PSNP as a case study and concludes with policy recommendations.

2. Child-sensitive social protection

Social protection is increasingly being viewed as part of the response to child poverty and vulnerability (Shaffer, 2003). In policy, however, it has ‘tended to focus on vulnerable groups other than children’ (Handa et al., 2010:3). Research indicates links between poverty prevalence and a disproportionately large incidence of child poverty (Lanjouw et al., 1998). ‘Child-sensitive social programming’ is the catch-phrase used to summarise a wide range of policies and programmes (Roelen & Sabates-Wheeler, 2011) in the social protection landscape. In most of these programmes, social transfers are widely used as instruments to tackle child poverty and associated vulnerabilities discretely (Adato & Basset, 2008).

Before the CSSP phrase was coined, Crawford (2001) challenged governments to prioritise its future citizens – children and young people – in health, education and early childhood development. Later, ‘child-conditioned’ income transfers, support for maternal employment, and early childhood programmes became key apparatuses for both reducing child poverty and enhancing child development (Kamerman et al., 2003). Investment in children is linked to long-term poverty reduction, particularly in ‘critical areas of child protection such as health, education, early childhood development’ (Crawford, 2001:503). More discrete CSSP will especially prevent, manage and overcome child-related risks and vulnerabilities as they change throughout childhood (Jones & Holmes, 2010).

This discreteness has led to some UNICEF-supported diagnostic studies, for example, Bailey et al. (2011) in the Democratic Republic of Congo (DRC) and Roelen et al. (2011) in nine countries.4 Bailey et al. show that CSSP measures in the DRC are still limited mainly to formulating legislation and national child protection plans, while Roelen et al. map out lessons built around children affected by HIV/AIDS and scaling them up. These lessons include the need for government to be the leading actor and bearer of responsibility without undermining the role of non-formal modes, and reaching those children who cannot be supported through families. Roelen et al. (2011:23) observe that although this CSSP-HIV/AIDS programming in east and southern African countries is being piloted, it is still plagued by the problems of ‘translating policy and strategic rhetoric’ and putting into practice the ‘theoretical conceptualisation’ contained in the programming guidelines and policy documents and declarations by several international organisations and agencies.

There have been claims that most child-specific guidelines fail to guide practice in social protection programmes. For example, a statement issued jointly by the United Nations Children's Fund and other development partners, which spells out seven CSSP guidelines, does not provide sufficient guidance for programme planners, designers and implementers of child-friendly social programming interventions (Roelen & Sabates-Wheeler, 2011). Jones & Holmes (2010) call for the Millennium Development Goals to be revised, because they believe that in their current design they do not adequately tackle the complexity of childhood poverty and vulnerability. The literature reviewed above suggests that the lack of clear guidelines may be part of the reason why CSSP remains elusive in some country programmes. The following Ethiopian case study is used to further explore this assertion.

3. PSNP public works and direct support programmes in Ethiopia

South Africa is generally acknowledged as having the most comprehensive state-led system of cash transfer-based social protection in sub-Saharan Africa, followed by Ethiopia's PSNP (Devereux, 2001). In 2009, the Ethiopian Ministry of Agriculture and Rural Development (MoARD, 2009) reported that the PSNP annual budget stood at US$347 million (1.2% of Ethiopia's GDP). This programme covered 7 574 480 beneficiaries, with the budget for 2010–2014 exceeding US$2 billion.

The PSNP seeks to address food insecurity by smoothing consumption patterns and preventing asset depletion for food insecure households totalling 14 million people, of which more than six million are children in chronically food insecure areas (MoARD, 2006). This is to be achieved via appropriate, timely and predictable transfers of food or cash, or both, either through a public works programme (PWP) (food or cash for work) or direct transfers through a direct support programme (DSP) component. The former entails the provision of counter-cyclical employment on rural infrastructure projects such as road construction and maintenance, small-scale irrigation and reforestation, while the latter constitutes direct unconditional transfers of cash or food to vulnerable households with no able-bodied members who can participate in PWPs. In addition, the PSNP is expected to enhance child welfare by increasing household income and smoothing consumption, potentially improving education and reducing child labour. Through a third component of the PSNP, called ‘other food security programmes’, some households benefit through access to credit, agricultural extension services and technology transfers in areas such food crop production, cash cropping, livestock production, and soil and water conservation as well as irrigation and water harvesting schemes (Hoddinott et al., 2010). However, this supplementary programme is not the subject of this paper.

The PWP is the larger component of the PSNP. It constitutes approximately 85% of the PSNP budget (MoFED, 2007) and it paid eligible beneficiaries 10 birr (US$0.81) per day as at 2009/10. By design, each adult household member is to work a maximum of five days per month up to a maximum of 20 days (Gilligan et al., 2007). In cases where the labour input exceeds the maximum number of working days, the outstanding transfers have to be given to the household through a DSP. This limit on time spent on PWP activities is intended to avoid the problem of the PSNP shifting households from their usual livelihoods activities, i.e. to avoid dependence on the PSNP (an outcome to which the government is sensitive).

As this is a food security programme, PWP wages were originally calculated to equal a shopping basket containing the equivalent of 15 kg of grain, 1.5 kg of pulses and 0.5 litres of oil. Eligible households are those with a three-year continuous dependence on relief (prior to the programme) and able-bodied members who are above 18 years of age. PWP planning engages the communities to identify and prioritise work to be done. The work includes projects such as construction of roads and schools, soil and water conservation, and water development, among others.

The DSP, which is distributed in the form of unconditional cash or food transfers, is provided to labour-poor households who do not have support from their adult children or remittances from other relatives. DSP beneficiaries include, but are not limited to, orphans, pregnant and lactating mothers, the elderly, households with sick individuals and female-headed households. Though DSP transfers are unconditional, Save the Children (UK) Ethiopia piloted attendance at health and nutrition education sessions as conditions for transfer. This pilot study explored synergies between the PSNP and the health sector to produce the maximum benefits for children. The results from this study were encouraging: for instance, there were improvements in care for infants and young children and health seeking practices (Fenn, 2009).

4. Child focus of PSNP: A critique

The proportion of children who started Grade 1 at the expected official age (seven years) used to be just above 20% (World Bank, 2005). While this rate has risen to 60%, an estimated 25% drop out before finishing Grade 1, and 50% do not finish Grade 5 (Woldehanna et al., 2011). According to Ethiopia's Central Statistical Authority, 27% of school dropouts cite the need to work as the main reason for dropping out (CSA, 2005). In a survey of 3043 Ethiopian children aged between four and 15, Admassie & Singh (2001) found that 75% were engaged in different forms of work (59% paid and approximately 41% without fixed payment). The CSA's 2003 national child labour survey indicated that 62% of rural 10- to 14-year-olds were involved in economic activity (CSA & ILO, 2003). It has been observed that households under-invest in the human capital of their children during exposure to shocks (Jacoby & Skoufias, 1997). Furthermore, children who work are more likely to fail at school, and their educational attainment is reduced by almost two years (Psacharopoulos, 1997).

But how well does the PSNP address this problem? Arguably, not so well. For one thing, the PWPs ignore the circumstances of the parents or guardians and the children themselves. PSNP beneficiaries generally have a relatively lower labour capacity, so it was to be expected that DSPs, which do not require labour capacity, would occupy the greatest part of the PSNP. Yet, PWPs, which require labour, occupy the greatest part of the PSNP. Furthermore, the PSNP imposes a ceiling on the DSPs and this reduces both paid and unpaid child labour activities. This occurs despite the fact that the DSP has a strong impact on improving child education in Ethiopia (Hoddinott et al., 2010).

We also ask: what strategies do households adopt to obtain their entitlements from the PSNP? And what strategies do they use to supplement these entitlements? Empirical work conducted by Nigussa & Mberengwa (2009) in the Kuyu woreda (district) confirms general claims that PSNP entitlements are too low to support typical household consumption. Households therefore supplement these entitlements with income earned by the children in the households in both PSNP (PWP) and non-PSNP activities (during planting and weeding time).

Devereux et al.'s research (2008) could not conclude that the PSNP has resulted in fewer children being withdrawn from schools. There are other factors that may account for this outcome, such as the government's campaign urging that ‘every school-age child should go to school’ and promoting ‘expansion of schools’ initiatives carried out prior to and during the same period as its promotion of schooling. Further secondary data analysis also shows that 2% of children of households supported by the PSNP and 1% of children from non-PSNP households were withdrawn from schools during 2005/06 (Devereux et al., 2008). This effect on schooling (and the child work effect highlighted in the next paragraph) is based on a difference-in-difference regression and nearest neighbour matching method used by Devereux et al. (2008), Woldehanna (2010) and Hoddinott et al. (2010). In this method, a comparison group of non-PSNP beneficiaries with statistically average observable characteristics was selected so as to generate robust comparisons. While the difference is small, the question still remains: why were more children from PSNP-supported households (that receive additional income) withdrawn from school than children from non-PSNP households? Since primary school education is free in rural areas, it could be argued that children were withdrawn not for financial reasons but to add to the household labour pool and thus increase household income through paid work.

Woldehanna (2010) uses data from the Young Lives study of 12-year-old schoolchildren in rural and urban Ethiopia to assess the effects of the PSNP on children's work division between work and school, and how the split affects attainment of education. The analysis reveals that in rural areas PWPs increase paid child labour ‘due to direct involvement of children in public work or substitution of children for adults when adults go to public work’ (Woldehanna, 2010:183).

A further question would be: where and how did the design of the PSNP miss the mark? We argue it missed the mark because of its timetabling errors, which unintentionally supplement adult labour with child labour, and its lack of child labour-saving assets (see Section 4.3 below). We consider these situations separately below.

4.1 Timetabling errors

PWP activities are generally scheduled for the post-harvest period of April to September when food is generally in short supply. The expectation was for the PSNP to supplement income during this period. PWP activities generally come to an end once the planting season starts so that household members can return to the fields where they traditionally earn their livelihood.

The timing of the PSNP does not take into account strategies that favour childhood development. Before the programme was introduced, children generally worked only during the farming season. Once the PSNP was introduced, children started working all year round – in the fields or in PWP activities – dropping out of school if needed to contribute to household income. This would not have been the case had there been a preference for DSPs instead of PWPs.

Woldehanna's research (2010) shows that some children in rural areas missed classes for entire weeks, mainly from October to January when agricultural work (such as harvesting) is very intensive (see Figure 1). The main reasons for this were that children had to do paid work or were required for domestic and agricultural work.

Figure 1: Months when children are most absent from school

Source: Adapted from Woldehanna (2010).

4.2 Substitutability of adult labour with child labour

The PSNP adopted what is called the 80:20 principle, where 80% of the beneficiaries receive transfers in exchange for labour, while 20% are eligible to receive transfers gratuitously or unconditionally. Put simply, this principle assumes that 80% of food insecure households have the labour capacity to meet the requirements for full transfer entitlement, as it is pre-determined by the budget. But it is practically impossible that 80% of any population can be available to work since children under 18 years generally constitute approximately 50%, and nursing and pregnant women, the elderly and people living with physical and mental disabilities constitute another 25% (Adams & Kebede, 2005). This ultimately reverses the 80:20 principle to a 20:80 principle. Given that the transfer did not consider the nature or form of labour contributed, it was inevitable that child labour would be contributed in order to receive the full entitlement.

4.3 Lack of child labour-saving asset creation

Output-based PWPs can have positive external effects if the physical assets accumulated are strategically aligned to reduce child work. This strategic alignment stems from a recognition that increased household income withdraws children from work and puts them into schools, but the opposite scenario exists where there is a wider portfolio of productive assets. Greater access to productive assets in turn increases the productivity of, and demand for, child labour – and children are pulled out of school to work the assets (Cockburn & Dostie, 2007:560). What has been overlooked in linking child labour with income is that some assets can augment child labour, to the extent that households are encouraged to withdraw their children from school so as to exploit the increased returns from these assets (Cockburn & Dostie, 2007). Cockburn (2001:2) observes that

In rural Ethiopia, the principal activities of children are fetching wood or water and herding, whereas adult males are primarily involved in farming and adult females in domestic work. It therefore seems likely that the effects on child work will vary considerably depending on the types of physical assets targeted in poverty alleviation policies. In particular, targeting assets used in activities commonly performed only by adults may make it possible to avoid increased child work and reduced schooling. Furthermore, child labour-saving assets such as a nearby well or a wheelbarrow can be expected to directly reduce child work and poverty.

We argue that the PSNP did not make a ‘judicious choice of assets targeted’ (Cockburn & Dostie 2007:560) in respect of the problematic issues of child labour and school participation. The key intervention here is access to water and fuel, because the likelihood of children attending school falls significantly as the distance to the nearest water and fuel source increases. A water intervention has also shown positive spillover effects such as improvement in children's health (Fenn et al., 2012). An overview of the achievements of PWPs (Table 1) shows extensive community asset building that is said to have brought about ‘environmental transformation’ (Del Ninno et al., 2009). A relevant issue is that of increased groundwater recharge as some dry springs have started flowing again (Fenn et al., 2012). Flowing springs that are closer to home reduce the time children spend fetching water and create time for mothers to spend caring for young children, among other benefits that can lead to health improvements observed in the PSNP-Nutrition pilot project (Fenn et al., 2012).

Table 1: Sample of public works programmes supported under the PSNP, Ethiopia 2007

Activity	Result
Soil embankment construction (kilometres)	482 542
Stone embankment construction (kilometres)	443 148
Pond construction and maintenance (number)	88 936
Spring development (number)	598
Hang-dug well construction (number)	491
Land rehabilitation through area enclosure (hectares)	530
Small-scale irrigation canals (kilometres)	2 679
Tree nursery site establishment (number)	285
Seedlings produced (number)	301 778 607
Seedling planted (number)	12 883 657
Rural road construction (kilometres)	8 323
Rural road maintenance (kilometres)	20 458
School classroom construction (number)	340
Animal health post construction (number)	71
Farmer training centre construction (number)	119
Source: Food Security Coordination Bureau in Del Ninno et al. (2009:12).

As mentioned earlier, the construction of community assets did not directly reduce child labour or improve school participation, regardless of the fact that classrooms were renovated. Indirectly, a study by Nigussa & Mberengwa (2009) shows that despite a community in one of the PSNP districts saying that the shortage of potable water was its major problem, this priority was not only ignored by the authorities, the PWP sites were also located far away from participants' homes, which weakened the positive contribution of these assets to the problems of child labour and schooling.

5. Conclusion

This paper challenges a general view in the discourse of social protection – that social protection programmes are child-sensitive. The case of Ethiopia's PSNP was presented to demonstrate how concealed fundamentals in the design of social protection programmes may result in unintended outcomes in an otherwise well-intended programme. In the case of the PSNP, we identify the basic flaws as bad timetabling of its PWP component, the inevitable need to supplement adult labour with child labour, and the lack of child labour-savings assets. While these flaws may have been difficult to identify ex ante, there is now an urgent need to revise the PSNP, add provisions to enhance education and protect children from child labour, and prioritise child labour-saving assets as one of the key interventions. Lessons may be drawn from Latin American social protection programmes, where it has been noted that income effects protect children from child labour and improve their education (ILO, 2007). Child-sensitive programmes should take guidance from Article 6 of the United Nations Convention on the Rights of the Child and focus on ‘the circumstances of the child’ and ‘persons having responsibility for the maintenance of the child’. At policy and operations level the PSNP implementation manual (2006 version) prohibits children under 16 years from working on PWP activities. Child research and advocacy has made calls for upward revisions that will align the limit with the UN Convention on the Rights of the Child to 18 years.

However, even if this revision is made, everyday practice in PWP sites holds sway and supervisors perpetuate the same output-based perspective and ignore the child work issues. For instance, one supervisor quoted by Tafere & Woldehanna (2012:12) openly said ‘We do not care whether [parents] send their children or [whether they] come themselves because what we need is the job done. [Children] work better than the adults because they have the capability’.

We recommend an increase in the size of the cash transfer so that the income effect offsets the current strong substitution effect. In terms of research, there is need to find how the problem – of parents sending children to work and the PWP supervisor allowing this – can be addressed if these critical actors at community level ignore the effects of child work on schooling. One possible way is to cascade the child protection screening and oversight downwards from policy level so that PWP supervisors will enforce the age restrictions and employ only those who are eligible to provide labour.

Programme planners and policymakers need to explore of the possibility of integrating social protection programmes with other child survival interventions for complementarity and strengthening of intervention impact. The ongoing initiatives to link the PSNP with the National Nutrition Program (NNP) in Ethiopia offer an opportunity for synergy between the two programmes for better health and nutrition outcomes for children, especially when also combined with the exemption of pregnant and lactating mothers from PWP. Good performance at school is, among other things, a function of the good health and nutrition status of children (Victora et al., 2008). Baseline studies for labour based social protection programmes should consider households' labour capacity in future designs so as to eliminate the temptation to use children as sources of labour Attaching a conditional clause linking the receipt of transfers with certain child care practices and behaviours (e.g. school attendance) in a facilitated way could be a solution, given its success in some PSNP-linked projects (Fenn, 2009; Fenn et al. 2012) in Brazil, Mexico and other countries (Lindert et al., 2007). Such ‘facilitation’, i.e. including benevolent conditions, would help overcome the much criticised ‘aid conditionality’ which has adverse effects on beneficiaries.

Referring again to the diagnostic studies reviewed above that attempted to trace discrete CSSP outcomes, we recommend an intensive advocacy and oversight at implementation stage to promote appropriate child-sensitive management of PWP activities (so as to keep children at school), increased size of transfers (so as to augment the income effect) and integrating child labour-saving assets into the household and community asset creation processes. This, however, needs to rely on built-in monitoring and evaluation as part of programme implementation, with a focus on outcome and impact indicators that will reveal whether the programme or intervention is producing undesirable outcomes or is likely to do so in the long run.

Notes

⁴Angola, Botswana, Lesotho, Swaziland, South Africa, Malawi, Mozambique, Namibia and Tanzania.