Design and rural context in antipoverty transfers: Implications for programme outcomes
Armando Barrientos, Global Development Institute, University of Manchester, UK
Abstract:
The expansion of antipoverty transfers in low- and middle-income countriesfocuses on poverty reduction but it also raises important questions regarding their contribution to growth and development. The paper considers the role of design and context in programme outcomes. Distinguishing three main programmes types identified, the analysis relies on insights from theory and practice to identify rural context conditions capable of enhancing or undermining programme outcomes. It finds that context conditions are important and require careful consideration by programmers.
Introduction
The rapid expansion of large-scale antipoverty transfer programmes in low- and middle-income countries has been a feature of the last two decades. They provide direct transfers to families in poverty to facilitate their exit from poverty. An interesting fact is that a majority of antipoverty transfer programmes have startedin rural areas. Mexico’sProgresa, a pioneer initiative in Latin America, was designed and implemented as a response to the changes in rural living standards and livelihoods associated with the liberalisation of agricultureand in view of the limited effectiveness and clientelism associated with the existing antipoverty programmes(Escobar Lapati, 2002; Levy, 2006). Persistent poverty in rural areas in Mexico, highlighted by the Zapatista mobilisation, was the main focus of the programme(González de la Rocha and Escobar, 2012). The spread of antipoverty transfer programmes in the South is linked to structural transformation and globalisation processes,in which agriculture plays a large role.This paper examines the expansion of antipoverty transfer programmes focusing on the implications of the agriculture context in facilitating or limiting their effectiveness.
The rural focus in the expansion of transfer programmes should not be overstated. It is not true everywhere. Brazil’s Bolsa Escola began in urban areas at the same time as the Programme for the Eradication of Child Labour (PETI), now a component of BolsaFamília, replicated many of Bolsa Escola’s features but with a focus on hazardous child labour in agriculture.China’s Minimum Living Standards Scheme, DiBao, emerged from large cities and was extended to rural areas much later.
Interestingly, the focus of social assistance expansion in agriculture has been associated with inclusion and citizenship. The inclusion of agricultural workers and their families in the emergent pension in Sweden at the turn of the 20th century was a consequence of the opposition of agricultural political parties to contributory pension schemes proposals supported by social democratic parties (Edelbank, Per Gunnar, 2009; Kildal and Kuhlne, 2005). Tax-financed basic pensions were more likely to include agricultural workers than contribution-based schemes. The gradual extension of the old age grant in South Africa in the 1970s and 1980s to include the majority of the black population was rationalised by civil servants as a ready-made instrument to include these groups and to inject effective demand into the ‘homelands’(Barrientos, 2008). Prêvidencia Social Rural in Brazil extended earlier schemes to include workers in rural areas into formal social insurance pension funds (Lewis and Lloyd-Sherlock, 2009). In all these cases, concerns with social and economic inclusion werebehind the extension of social protection to rural areas. To an important extent, they apply to the extension of antipoverty transfers underway in developing countries today.
The materials in this paper are divided into three main sections. Section 1 argues the expansion of antipoverty transfer programmes is best understood as institution building. Section 2 introduces a classification of antipoverty transfers based on their underlying understanding of poverty. Section 3 examines the constraints and opportunities associated with the agriculture context in the design and implementation of antipoverty transfer programmes. A final section summarises the main conclusions.
1.Social Protection and Antipoverty transfers
It is important to have clarity on terminology. In international development policy discussion, social protection has been used recently to describe antipoverty transfer programmes addressing poverty and vulnerability in developing countries. This is in contrast to the usage of the term in academic discussion, especially in high-income countries. There, social protection is commonly used to include social insurance and social assistance transfers. Social insurance transfers address life course and other contingencies leading to a reduction in income generating capacity and are on paper normally financed from regular contributions by workers and their employers. Social assistance provides tax-financed transfers to households in poverty(Barrientos, 2013). This paper focuses on social assistance, understood as budget-financed programmes providing regular transfers, in cash or in kind, to households in poverty.
The World Bank, and other Washington-based agencies, use the term ‘safety nets’ to describe all programmes and interventions focused on poverty reduction in developing countries (Grosh et al., 2008; Weigand, 2008). Safety netscrucially include emergency and humanitarian assistance and are therefore broader in scope than social assistance.There is very little to be gained from collapsing emergency and public or social assistance. They have very different analytical basis, objectives, target groups, and financing. It is essential to keep firmly in mind the conceptual distinction between emergency assistance, social assistance, and social insurance.
Flagship programmeshave led the expansion of social assistance in developing countries. The reasons behind a programme approach to social protection are complex and often country specific. They include fiscal constraints, political opposition, knowledge gaps and uncertainty, the time window of international aid, and the hubris of silver bullets in international development policy. For the pioneer middle-income countries, the initial focus on flagship programmes has given way to a focus on institution building. In many countries in Latin America, Ministries of Social Development have been established with the purpose of managing and coordinating antipoverty transfer programmes (Cecchini and Madariaga, 2011). In contrast to the ‘project’ approach dominating discussion in international development, this points to the need to approach the expansion of antipoverty transfer programmes in low- and middle-income countries as institution building.
A strand of literature on the link between agriculture and social assistance emphasises the latter's potential role in providing insurance against production or consumption shocks (Dercon, 2002). A number of studies focuses in the impact of transfers in cash on asset accumulation in rural areas (Blattman, Fiala and Martinez, 2011;Macours, Premand and Vakis ,2012; Stoeffler and Mills, 2014). It is important not to lose sight of the fact that social assistance also aims to address structural factors underlying poverty. An illustration might help to pin this down. Figure 1 provides a stylised view of welfare in four different types of households. In the Figure, the solid line traces welfare over time, while the broken line denotes a poverty threshold. Household A has an upward trajectory in welfare, but one punctuated by shocks lowering welfare. Smoothing these shocks for household A would strengthen the welfare trend and lead to permanent escape from poverty. Household B’s welfare, on the other hand, drifts above and below the poverty threshold alternating poverty spells with non-poverty spells. Smoothing the shocks for household B might lead to constant welfare at or just above the poverty threshold. Household C’s welfare is not affected by shocks and shows instead a constant level of welfare below the poverty line. In this case, no smoothing of welfare would lift, by itself, this household above the poverty line. Household D’s welfare is affected by shocks, but in the context of a declining trend. In this case, smoothing welfare, whilst protecting this household from shocks, is unlikely to be sufficient to enable it to escape poverty permanently. Addressing poverty in all four households requires that the design of antipoverty transfers takes account of idiosyncratic risk and structural factors, blending welfare smoothing and redistribution.
[Figure 1 about here]
2.Programme Types
The diversity of programme design across developing countries makes it necessary to identify ‘ideal types’. Analytically, it will be useful to classify social assistance programmes into three main categories: pure income transfers; income transfers combined with asset accumulation; and integrated poverty reduction programmes.
Pure income transfers include transfers specifically targeted at households in poverty and categorical transfers targeted at individuals belonging to a demographic group considered to be particularly vulnerable. Some pure income transfers are focused on households in (extreme) poverty. For example, the Kalomo Pilot Social Transfer Scheme in Zambia provides direct transfers to the poorest ten percent of households in designated districts. Other programmes – for example, children or family allowances or social pensions – are focused on specific groups in the population thought to be especially vulnerable. South Africa’s social assistance is organised in the main through means-tested categorical grants focused on older people, people with disabilities, and children.
Income transfers combined with asset accumulation include programmes providing transfers in cash or kind, which are combined with, and facilitate, accumulation of productive assets. The term ‘asset’ is used here in its broadest sense, to include human, physical and financial assets. Linking direct transfers with interventions aimed at asset accumulation underlines the fact that programmes of this type aim to strengthen the productive capacity of households in poverty.
This type includes two families of programmes now common in developing countries. The first group includes programmes combining direct transfers with interventions facilitating household investment in human development, especially education and health. Mexico’s Oportunidades or Brazil’s BolsaFamíliaare well known examples of this family of programmes. The second group includes programmes combining direct transfers with interventions facilitating physical asset protection and/or accumulation. Examples of this type of programme include India’s Mahatma Gandhi National Employment Guarantee Scheme (infrastructure or community assets) and Ethiopia’s Productive Safety Net Programme (household and community assets). The Productive Safety Net Programme combines transfers in cash or kind with agricultural extension activities for households with work capacity. It also provides direct support to households without work capacity. These programmes require beneficiaries to supply work to create or protect household or community assets. They are also described in the literature as public works, cash/food for work, or guaranteed employment programmes.
Integrated poverty reduction programmes are an important innovation in social assistance, combining a range of interventions focused on the poorest. BRAC’s Challenging the Frontiers of Poverty Reduction – Targeting the Ultra Poorin Bangladesh provides an integrated and sequential set of interventions strengthening the nutrition and health status of the poorest households, as well as training in preparation for the transfer of productive assets. Chile’s Chile Solidario(nowIngresoEtico Familiar), together with China’s Di Bao and South Korea’s Minimum Living Standards Scheme, are examples of programmes providing an integrated set of interventions addressing a range of deficits responsible for keeping households in poverty.
The underlying understanding of poverty underpinning the programmes is the basis for this classification. Pure transfers rely on an understanding of poverty as largely to do with deficits in income or consumption. Transfers are expected to remedy these deficits and thus reduce poverty. Income transfers combined with asset accumulation share a broader understanding of poverty. They pay attention to deficits in income or consumption but, important as these are, they also aim to address deficits in productive assets. Programmes focused on human development, like Mexico’s Oportunidades, understand the persistence of poverty as arising from deficits in human capital. Ethiopia’s Productive Safety Net Programme understands poverty and its persistence as arising from the asset depletion and destruction affecting food-insecure households faced with droughts or other shocks. Programmes included in this group adopt a multidimensional understanding of poverty, but focus on a few dimensions. Integrated poverty reduction programmes also share a multidimensional understanding of poverty, but are distinguished both by a wider set of dimensions covered and by the fact that direct income transfers play only a marginal role in the overall support provided to households in poverty. Integrated poverty reduction programmes understand poverty as arising from social exclusion.
3.Programme design, context and outcomes
What constraints and opportunities arise from the agriculture context? And what are the implications for programme outcomes? The section below examines these issues for each of the ideal types identified in the last section. The analysis reviews the expectations from theory and available knowledge from practice and then considers the implications of the agriculture context programme outcomes. Table 1 summarises the main points.
[Table 1 about here]
3.1 Pure income transfers
The theoretical literature suggests that the effects generated from transfers of this type are captured by the outcomes generated by the additional income transferred, i.e. their income effects. Households receiving pure income transfers, if operating within their budget constraint, will use the transfer to increase their consumption of all normal goods and services proportionally. Consumption rises in proportion to the transfer. This expectation is in line with the findings from studies of expenditure patterns following receipt of transfers. Several studies findthat household consumption rises by around 75% – 85% of the amounts transferred on average(Angelucci and de Giorgi, 2009; Fiszbein and Schady, 2009; Hoddinott and Skoufias, 2004).
In high-income countries, the literature on pure income transfers has been dominated by concerns with incentive effects. In this literature,work disincentives are directly related to the generosity and duration of transfers(Moffitt, 2002). Policymakers aim to ensure households in poverty reach a specified basic standard of living. However,where household respond to transfers with a reduction in labour supply, they might as a consequence remain below the target standard of living. A strategy to minimise adverse incentives effects is to identify population groups with high incidence of poverty and who are less likely to show work disincentives; and ‘tag’ transfers onto them (Akerlof, 1978; Feldstein, 1987). Categorical transfers on older people, people with disabilities, and children provide appropriate examples of such groups. Because work disincentives are less of an issue with these groups, then transfers can be relatively more generous.
The impact of pure income transfers is maximised where households in poverty face constraints in the allocation of their productive resources. It is to be expected that pure transfers would be especially effective in rural areas. If demand conditions in the local economy are declining or are variable, for example as a consequence of sustained labour out-migration, income transfers help sustain effective demand. In Brazil and South Africa, old age transfers have been directed to depressed areas as a means to sustain local economies(Barrientos, 2008). This makes sense because with rapid urbanisation older people are left behind in rural areas. Liquidity or credit constraints can also shift households away from their optimal allocation of productive resources, especially in conditions where household production functions are non-linear. Martinez (2004) finds that social pensioners in rural Bolivia show more than proportionate increases in household consumption following pension receipt, when compared with social pensioners living in urban areas. He attributes this differential impact to the fact that the pension is paid once a year and therefore constitutes a significant lump sum and to the fact that Bolivian rural pensioners face liquidity constraints in purchasing seeds and fertilisers and in hiring equipment or labour. The transfers help them purchase the complementary productive inputs needed. In Brazil, rural social pensions are paid through a magnetic card. The magnetic card facilitates access to bank credit for social pensioners as it demonstrates a guaranteed income flow (Delgado and Cardoso, 2000). Income transfers can also help families in poverty overcome liquidity constraints associated with service charges, especially health care and schooling (Case and Wilson, 2000; Galiani et al., 2014; Schubert, 2008).
Several features of rural areas suggest limitations in the poverty reduction effectiveness of pure transfers. Corruption by public agencies or predatory elites could significantly reduce the value of transfers actually reaching beneficiaries. In extreme cases, the costs to beneficiaries from accessing their entitlements might absorb a large fraction of the actual transfers (Pellisery, 2008). Community or household power structures could also undermine the effectiveness of pure income transfers, through compensatory private transfers, sharing of transfers within households, gender disparities in decision making, or social stratification(Juarez, 2009; Møller and Sotshongaye, 1996; Molyneux, 2006). To the extent that households in poverty face non-linear production or consumption functions, well documented in the agriculture context, the level of the transfers could be too small to generate the desired outcome. This applies to most of the transfer programmes in sub-Saharan Africa (Garcia and Moore, 2012).
3.2 Income transfers plus asset accumulation – human development conditional transfer programmes.
Human development conditional transfer programmes combine income transfers with conditions facilitating human capital accumulation. Whereas income effects dominated in pure income transfers, conditions and other programme design features in human development conditional transfer programmes generate significant substitution effects. Substitution effects work through the structure of the transfers, conditions and co-responsibilities; and through information flows and intermediation. Normally transfers have several components: a household consumption transfer component and transfers associated with the number of children and their ages/school grade. The conventional conditions attached to continued programme participation include children schooling, utilisation of primary health care, and nutrition information and training. A common pattern is to require that children of school age enrol at school and attend regularly. Health conditions require participant households to ensure all household members attend health centres for regular checks. They also require that expectant mothers and infants to be checked up at more regular intervals and infants have a full set of immunisation. In some programmes, nutrition training for mothers is part of the conditions required for participation.