Trajectories of social protection in Africa

Abstract

Social protection in contemporary Africa is the product of several strands of social policy, from European social security systems to humanitarian relief interventions. Contributory social security mechanisms such as unemployment insurance and pensions were imported to Africa during the colonial period, but cover only a minority of formally employed workers. Food aid alleviates hunger but does not resolve problems of chronic food insecurity. Cash transfers are being promoted as an alternative to food aid, but have been criticised for being ineffective against price inflation and underwriting neoliberal economic policies. Some programmes link social transfers to public works employment and microfinance, with the aim of ‘graduating’ participants off social protection. This article argues instead for a ‘social contract’ approach – recognising the right to social protection, empowering passive beneficiaries to become entitled claimants, and introducing social audits to hold duty-bearers accountable for effective and equitable delivery of citizen-driven social protection policies.

Keywords:

• cash transfers
• food aid
• social contract
• social protection
• social security
• social welfare

1. Introduction

Social protection in Africa today is woven from several strands of complex and diverse pre-colonial, colonial and post-colonial histories. This article explores two of these strands – European social security provision and humanitarian relief interventions – to develop an argument about the limitations and potential of the evolving discourse, in terms of delivering effective and comprehensive social protection.

This paper makes its argument in three steps. First I demonstrate that the current social protection policy agenda in Africa is not indigenous but largely imported. I identify two specific trajectories of social protection in Africa (there may also be others). One trajectory started with the adaptation of European social security models, which were first applied to the formally employed workforce in African countries in the form of contributory pensions and unemployment insurance and have more recently been extended to the informal and self-employed sectors in the form of targeted cash transfer programmes such as ‘social pensions’. The other trajectory started with ad hoc humanitarian responses by the international community to food security shocks in Africa and developed towards quasi-institutionalised social assistance and social insurance mechanisms such as public works programmes for the seasonally underemployed and weather-indexed insurance for farmers.

In the second step I argue that attempts to graft these imported models onto domestic policy agendas have failed to recognise that the economic and social structures of African countries are fundamentally different from those of Europe, resulting in grossly inadequate coverage and programmatic responses that fail to meet the actual social protection needs of local populations. In the third step I propose an alternative model, based on the construction of a rights-based ‘social contract’ for social protection at national level, which would generate a third trajectory – from largely unaccountable and often externally financed social assistance projects towards accountable, nationally owned social protection systems.

2. Antecedents

Formal social security systems originated in Europe as one set of policy responses to the ‘great transformation’ (Polanyi, 1944) during the 18th and 19th centuries. The emergence of the modern nation-state and of ‘self-regulating markets’ as organising institutions of liberal economies was associated with dramatic economic transitions and social dislocation. Industrialisation triggered a mass movement of rural populations to towns and cities, to work in factories and offices. Urbanisation was accompanied by the commodification of labour, which meant that labour power previously devoted to growing food for the family was now dedicated to producing profits for employers.

‘Social protectionism’, understood here as an ‘intervention in the market economy in order to secure the cultural and social integrity of humankind and/or society (e.g. welfare-state redistribution and the regulation of the economy)’ (Drahokoupil, 2004:837), evolved to protect society against the negative consequences of market forces. Drahokoupil (2004) identifies two contradictory readings of social protectionism in Polanyi's The Great Transformation. The ‘balancing principle’ reading is that social protection insulates society against self-regulating markets – it is a necessary counter-balance to economic liberalism and can produce a sustainable equilibrium. The alternative ‘market pathology’ reading sees social protection as an interference in the functioning of the self-regulating market that will lead inevitably to its destruction. This tension can also be framed in terms of class struggle – if ‘self-regulating markets’ serve the interests of capital, ‘social protectionism’ serves the needs of labour. In the contemporary social protection discourse, the tension between economic and social concerns manifests itself in complaints (from economists, business, ministries of finance) about ‘dependency syndrome’ and the ‘unaffordability’ or ‘unsustainability’ of labour market interventions and social welfare programmes.

Urbanisation and industrialisation had two important implications for social reproduction and the ‘demand’ for and ‘supply’ of social protection. On the demand side, urbanised former farmers lost their access to subsistence through own production and became dependent on the labour market for survival. Instead of relying on rainfall and harvests for their food and income, they now relied on regular payments in the form of wages or salaries from employers, with which they purchased food and other essentials from commodity markets. On the supply side, in pre-industrial societies informal social protection, according to Polanyi, was organised according to the behavioural principles of reciprocity, redistribution and householding, but these principles were displaced by the economic rules of the market. Traditional informal providers of social protection – kin and community – were undermined by the disruption of community bonds, the weakening of affective relationships and the anomie of urban lifestyles.

Similar concerns about the dialectical relationship between ‘formal’ and ‘informal’ social protection have pervaded policy debates for decades. In South Africa, where social pensions were extended to urban Africans in 1944, rural Africans were excluded for some years because the 1943 Social Security Commission argued against providing benefits that would ‘conflict with or break down their traditional food sharing habit’ (Devereux, 2007:543).

Responsibility for guaranteeing subsistence to individuals and families who could not support themselves shifted away from relatives and neighbours and became the responsibility of the market and the state. A complex array of formal social protection instruments was devised to ensure the reproduction of the labour force. Adults with labour capacity need (contributory) social security for periods when they are not working – unemployment insurance, retirement pensions and paid maternity and sick leave. Dependants who are not working – children, older persons, people living with disabilities – need state-funded (non-contributory) social welfare if their private support systems are inadequate. The foundations of this model were devised in Bismarck's Germany in the 1880s, institutionalised in the UK following the recommendations of the 1942 ‘Beveridge Report’, which argued for a social insurance system that would guarantee a minimum standard of living ‘below which no one should be allowed to fall’, and later evolved into the range of ‘welfare state regimes’ (Esping-Andersen, 1990) now found across the European Community.

This model was also exported across much of Africa, Asia and Latin America by European administrators during the colonial era of the late 19th and early 20th centuries. In Africa, the earliest social security systems were introduced in South Africa and North Africa. Although based on employee contributions, agricultural and self-employed workers were sometimes encouraged to join (e.g. in Egypt and Tunisia), with incentives such as lower contribution rates. In francophone West Africa, voluntary retirement plans for civil servants during the colonial period became compulsory schemes based on defined benefits after independence. In former British colonies, provident funds for public sector workers were also introduced after independence, usually as compulsory savings schemes to provide insurance against unemployment or retirement. Some countries, especially those affected by armed conflict (such as Somalia) do not yet have institutionalised social security systems, while others (such as the Democratic Republic of the Congo) have seen theirs destroyed by internal instability (Bailey & Turner, 2002).

But how well do European models of social security map onto the African context? Arguably, not well at all. In most African countries, the urban and industrial transitions have not yet happened. Most Africans still live in rural areas where they depend on farming for a living and on relatives or neighbours when their livelihood fails (i.e. householding, reciprocity and redistribution). Both these features present challenges to the European social security model. First, in industrialised economies the majority of workers are employed in a ‘standard employment relationship’, with written contracts that entitle them to a regular income from which contributions to social security can be automatically deducted. A minority of people live in poverty, do not have access to regular paid work and depend on social welfare. In African countries this economic structure is reversed: a minority have contractual employment, but the majority live outside the formal sector, with no job security, irregular and unpredictable incomes and no opportunity to make regular social security contributions. Coverage of national social security schemes or contributory provident funds reaches only the few who have a ‘standard employment relationship’ in the public and private sectors – typically, less than 15% of the population. A major challenge is to ‘extend social security to all’, in contexts where most of the working population are either self-employed farmers or pursuing a meagre and often precarious living in the informal sector.

Second, rural livelihoods are very different from urban livelihoods. Seasonality brings predictable cycles of hard work and under-employment, full and empty granaries, low food prices when the harvest has been reaped and high prices when the crops are still in the field. Rural people know what it is to face the annual ‘hungry season’. Farmers face food crises, usually triggered by adverse weather events. Poor smallholders with only a little land and struggling to afford seeds and fertiliser face persistently low yields and chronic hunger. Recognising that problems in accessing inputs and food are both symptoms of low income – or ‘entitlement failure’, in Sen's phrase (1999) – many African governments have introduced cash transfers rather than food aid as a solution to chronic poverty and food insecurity. But whether cash transfers are the most effective policy response to the inherent variability and unpredictability of rural African livelihoods remains open to question.

Third, labour markets are more complex in Africa than in Europe, and are becoming more so. The notion that people are either contractually employed or unemployed does not hold where there are high levels of informal employment and self-employment. Moreover, labour markets across the world are becoming increasingly ‘informalised’ (Standing, 2008) and it is not just the unemployed who are vulnerable but also casual, seasonal and other low-paid workers. Innovative thinking is needed to extend social security coverage and ‘decent work’ principles (ILO, 2006) to vulnerable workers of all kinds, and also to defend the employment contracts and associated rights to social security that are being eroded by labour market ‘flexibilisation’.

A further impetus for social protection in Africa was dissatisfaction with decades of humanitarian relief, especially emergency food aid. Humanitarian relief is not targeted at individuals on the basis of their employment status; instead, it provides temporary assistance to families who have lost their livelihood, for instance following a drought that destroys harvests. Unlike social security, which is rules-based, humanitarian relief is entirely discretionary. Critics of emergency food aid argue that while it alleviates the symptoms of food insecurity, it does little to address the underlying causes (Clay et al., 1998). Similarly, evaluations of project food aid (food-for-work, supplementary feeding, school feeding schemes) in many countries have found little or no discernible impact on chronic food insecurity, though it should be noted that food-for-work and school feeding schemes also have broader developmental objectives. A stronger criticism is that decades of food aid in Africa have merely perpetuated the demand for food aid – by weakening farmers' incentives to grow food, inhibiting market development, distorting consumer preferences, enabling donors to institutionalise surplus disposal, and allowing aid-dependent governments to neglect investment in agriculture and rural livelihoods (Barrett & Maxwell, 2005).

On the other hand, food aid does respond directly to real food security needs. Whether cash transfers can buy food security in a self-provisioning agrarian economy depends on the functioning of local markets. If access to food is restricted at household level but not at local or national level, cash works (it restores access to food when entitlements fail), but if food availability is constrained at local or national level, then cash works less well (and food aid works better).

Cash transfers have recently been introduced as a drought response in several African countries, with varying success. In some cases where drought has caused food production to collapse, cash transfers have been effective in replacing access to food through production (growing one's own food) with access to food through markets (buying food). In drought and other emergency scenarios in Africa cash transfers can even help to strengthen weak markets, by providing an incentive for traders to bring food from surplus to deficit areas. In other countries, cash transfers have worked less well. In Malawi, Lesotho and Swaziland, emergency cash transfers were supplemented with food aid, to ensure that drought-affected households had some access to food if local markets failed. Also, the value of the cash transfer was benchmarked against the cost of a basic food basket, and in the Food and Cash Transfers project in Malawi it was adjusted every month as local food prices rose and fell. In non-emergency contexts, many PSNP (Productive Safety Net Programme) participants in Ethiopia who had initially received cash transfers chose to revert to food aid or cash plus food in 2007, when food prices started rising faster than the value of the cash transfer (Sabates-Wheeler & Devereux, 2010). In neighbouring northern Kenya, participants in the Hunger Safety Net Programme have seen the purchasing power of their cash transfers drop by two thirds since the programme started in 2008, because of rapid and uncompensated food price inflation.

Notwithstanding the persistence of food aid, especially in emergency situations, social protection in Africa has become dominated by regular cash transfers to targeted vulnerable households in rural communities, a modality that is reminiscent of an incipient social welfare system. Looked at this way, the ‘cash versus food’ debate, which has preoccupied the social protection discourse in Africa until recently, can with hindsight be reinterpreted as a struggle between the social welfare lobby and the food security lobby within development policy, with food aid being negatively characterised as backward-looking, self-serving (surplus disposal by donors) and paternalistic (denying beneficiaries choice), while cash transfers are positively characterised as forward-looking, progressive and ‘modern’. The displacement of food aid by cash transfers is not just a technical choice between alternative modalities: it signifies a broader move away from a humanitarian approach towards a welfarist approach. In a sense, the hegemony of cash transfers is a victory for the ‘modernisation’ and commodification of social protection in Africa.

3. Ideological critiques

Although they have been widely adopted and evidence of their benefits is accumulating, large-scale social transfer programmes in Africa have been subjected to ideological criticism from both the right and the left. Critiques from the right resonate with social welfare debates in Europe. Handouts, especially to able-bodied adults, are predicted to create perverse incentives for behavioural change – people who could and should be self-reliant might choose free support from the state instead (‘dependency syndrome’), or even adjust their behaviour in order to qualify for humanitarian relief or social assistance (‘moral hazard’). Examples would be if farmers in Ethiopia stopped farming in expectation of receiving food aid, or if young women in South Africa deliberately fell pregnant in order to receive the Child Support Grant (CSG).

Studies have in fact been conducted on both these scenarios, but there is little empirical evidence for dependency syndrome or moral hazard in either case (Devereux, 2010). In rural Ethiopia, food aid is not guaranteed or predictable and is usually inadequate to meet a family's food needs, so no rational farmer would stop farming in the hope of being given enough food aid, and promptly, to compensate for lost production and to smooth consumption (Little, 2008). The incentive to alter behaviour rises with the value and predictability of the transfer, so regular social welfare programmes such as the CSG are stronger candidates for ‘leakages’ to people who do not really need them than are ad hoc emergency relief interventions. In South Africa, anecdotal claims that teenage fertility has increased since the CSG was introduced in 1998 have been refuted by an empirical study that showed that teenage pregnancies had in fact stabilised by 1998, that only 20% of teenagers with children receive the CSG, and that trends in fertility among teenagers were similar across all socio-economic groups, including those who would not qualify for the CSG (Makiwane & Udjo, 2006).

Critiques from the left question the political motivation for social protection, especially in countries pursuing neoliberal economic policies, where delivering social transfers to the poorest citizens is often interpreted as a palliative measure designed to quell social unrest and deflect calls for alternative economic models or radical political reforms. Countries with high levels of inequality – such as South Africa, Namibia and Botswana, which all have well-developed social protection systems – are especially susceptible to this criticism. South Africa's post-apartheid blend of neoliberal economic policies and progressive social policies has justifiably been described as a ‘strange and often incoherent mix of policies and interventions’ (Seekings & Nattrass, forthcoming:23). Neoliberalism reproduces economic growth and prosperity for a minority alongside chronic poverty and unemployment for the majority, who receive compensation in the form of social grants – a nominal level of redistribution that contains but does not substantially reduce either headcount poverty or income inequality (Leibbrandt & Woolard, 2010) and serves to legitimise the economic growth strategy.

When South Africa's extensive social grants system is discussed as a possible model for other African countries, the phrase ‘South African exceptionalism’ is often heard – especially in the context of affordability, since South Africa (like Botswana and Namibia) is undeniably wealthier and has a higher revenue base than, say, Ethiopia or Malawi. But South Africa is also a more urbanised, less agrarian society – which explains why the European model of social security plus social grants seems more appropriate at first glance. The problem is that the South African economy is not structured like a European economy. Most significant is the fact that unemployment in South Africa does not affect just a residual minority – an extremely high proportion (30 to 40%) of adults are unemployed, and most of them have no access to social insurance (the Unemployment Insurance Fund has limited coverage) or to social assistance (the Extended Public Works Programme also has limited coverage). Social grants are designed to reach non-working vulnerable groups such as the young (the CSG), older persons (the Old Age Pension) and people with disabilities (the Disability Grant) – but many unemployed or underpaid adults end up depending on these grants: in other words, they become dependent on dependants. This is not the intention, but it is the inevitable consequence of a hybrid blend of neoliberal economic policies and progressive social policies. In other southern Africa countries where non-contributory social pensions have been introduced, including much poorer countries such as Lesotho and Swaziland, dependence of working-age adults on these transfers has also been observed.

4. Labour market linkages

A comprehensive and sustainable approach to social protection needs to address failures in the labour market as well as provide social welfare for the non-working poor. A statistical decomposition of household incomes in South Africa reveals that wealthier citizens survive mainly on wage or salary income, while the poorest survive mainly on social grants (Leibbrandt & Woolard, 2010). This raises a fundamental question. Does the rapid and dramatic expansion of social grants indicate successful social policy or a failure of labour market policy? South Africa is applauded globally for the scope and generosity of its social grants system, but a radical critique might argue that these grants effectively underwrite an unemployment rate of over 30%, rather than compelling the government to make more sustained efforts to tackle the structural causes of poverty, by intervening directly in the labour market to reduce unemployment.

That is not to say that the employed are guaranteed access to social security. In South Africa workers' rights have been extended in recent years through progressive labour legislation to ensure decent working conditions, a statutory minimum wage and access to social security, including unemployment insurance. Perversely, these positive trends have been accompanied by accelerated casualisation of vulnerable sectors of the workforce, such as domestic workers and farm workers, as employers take steps to evade the costs of meeting their obligations under the law, leaving many workers less protected than before.

Interventions such as the PSNP in Ethiopia and the VUP (Vision 2020 Umurenge Programme) in Rwanda seem to have discovered an elegant solution to the reality that social protection needs to compensate for both a failure of labour markets – with the concomitant loss of access to social security – and a failure of most African governments to deliver social welfare to their poor non-working citizens. In effect, these programmes deliver both a labour market intervention for working adults and social grants for the dependent poor. The PSNP and VUP both offer temporary employment (public works) for unemployed or underemployed adults – though without the social security and decent working conditions of a ‘standard employment relationship’ – and unconditional transfers (direct support) for poor individuals who cannot work. The unconditional transfers were always intended to be a ‘residual’ component. In Ethiopia, this residualism was codified in an informal rule of thumb – in the 1990s, a public works programme called the Employment Generation Scheme was complemented by Gratuitous Relief, but the two were supposed to be delivered in a ratio of 80:20. The assumption was that most dependants were resident in households with working adults who could be employed in public works activities. In the PSNP, ‘full family targeting’ is applied, meaning that payments are made in direct proportion to household size, allowing working adults to earn cash or food for themselves as well as their non-working dependants.

Moreover, employment in public works was never intended to be a permanent feature of the PSNP or VUP. Instead, public works participants are supposed to ‘graduate’ within three to five years, by reaching a level of food security or self-reliance where they no longer need access to subsidised employment. But the debate about graduation is confused. Firstly, when the International Food Policy Research Institute (IFPRI) designed graduation thresholds for Ethiopia's PSNP, they converted household asset values into cash and concluded that a household had graduated when the value of its assets exceeded a threshold value. This approach is evidence of a fundamental misunderstanding about the nature of agrarian livelihoods in Africa. A calculation of the poverty headcount in a panel survey in Ethiopia showed very different rates at different times of year – because of seasonality, poverty is substantially higher before the main annual harvest than after (Dercon & Krishnan, 1998). This is mirrored in findings from the nutrition literature, where adult body weights and child malnutrition rates are often reported as varying seasonally, but poverty surveys in Africa never record different poverty rates for the ‘hungry season’ and the ‘post-harvest’ season. Yet calculating the value of a household's assets at one point in time (which point in time?) says nothing about the ability of the household to withstand seasonal stress and its resilience against future shocks, which should be a key indicator of success for a social protection intervention. The preoccupation with graduation reveals a contradiction between establishing an effective and permanent safety net against cyclical seasonality and occasional crises, and a linear (but unrealistic) vision of economic growth where poor people are assisted to move smoothly up the income ladder until they are no longer poor, at which point the safety net can simply be removed.

A second limitation of graduation thinking is that it places too much pressure on cash transfers and public works to solve the structural problems that generate and perpetuate poverty and food insecurity in rural Africa. The assumption was that the vicious cycles of dependency and disincentives associated with food aid would be converted into virtuous cycles of asset accumulation and income growth through cash transfers and the infrastructure created by public works projects. But this ignores the precarious nature of livelihoods, the thinness of commodity markets and near-absence of financial intermediation, the failure by governments and donors to invest in agriculture over several decades, and the grossly inadequate asset base (in terms of land, livestock and inputs) that most farming families face. Also, despite its proponents' expectation that recipients will invest cash transfers in assets and inputs for more productive and sustainable livelihoods, in contexts of chronic and seasonal hunger, where the transfer covers only part of the family's food needs, it is inevitable that most of these transfers will be consumed. Indeed, the fact that substantial proportions of even small cash transfers are invested in agriculture, education and other productive purposes, as evidence shows, is quite remarkable.

Recognising this challenge, the PSNP added ‘other food security programmes’ and then the ‘household asset building programme’ to the menu of public works and direct support, with the aim of helping PSNP participants to generate supplementary streams of income, for instance, by keeping bees to produce and sell honey. This intervention has proved difficult to implement at scale, and the limited number of livelihood options supported means that local markets tend to be flooded quickly, so relatively few participants have successfully graduated through this route. In Rwanda, instead of livelihood packages, VUP participants are offered access to low-interest loans through the Ubudehe Credit Scheme, either as individuals or in small groups. Access to the scheme requires a business plan that demonstrates how the loan will be used to generate enough income to repay the loan within one year and generate sustainable streams of income thereafter. This component was launched in 2010 and there is no evidence yet to show whether it is facilitating large-scale graduation.

Even if working adults who participate in public works and/or take an asset building package do achieve a level of income and assets that allows them to leave the PSNP better off than when they entered it, most direct support beneficiaries by definition have no prospect of ever graduating. Although the 80:20 rule of thumb has been applied on the PSNP, the scale of the programme means that approximately one million Ethiopians are currently receiving direct support transfers. Despite its antipathy to social welfare and its fears of generating long-term dependency on social transfers, the government of Ethiopia does recognise the imperative to address the needs of these people, as well as those of public works participants who have not yet graduated when the current phase of the PSNP ends in 2014. Options for transferring the direct support caseload from the PSNP to the Ministry of Labour and Social Assistance are being seriously considered (Devereux & Teshome, 2009). Whether the same trajectory will be followed in Rwanda and other African countries where cash transfers and public works are compensating for the absence of social welfare provisioning and labour market failures remains to be seen.

5. Social contracts for social protection

The preceding discussion has identified two complementary functions of social protection: as social assistance in the absence of fully fledged social welfare systems, and as a labour market intervention in the absence of fully fledged social security or insurance mechanisms. Nonetheless, social protection does not completely compensate for the limited coverage of social welfare and social security in contemporary Africa. Two related deficits in the design and delivery of social protection need to be addressed: the absence of social contracts and the fact that most social protection is supply-led rather than demand-driven.

The first challenge is not, as is often asserted, to build political will for social protection but rather to construct a social contract for social protection in each country – which immediately problematises the dominant role of donors and international non-governmental organisations (NGOs) in designing, delivering and financing social protection in much of Africa. A rights-based approach has some merits in establishing a social contract, but rights must be realised. The terms ‘rights-holder’ and ‘duty-bearer’ are useful for identifying the actors involved and their roles and responsibilities, but a politician's signature on a document such as the 1948 Universal Declaration of Human Rights, the 1966 International Covenant of Economic, Social and Cultural Rights or the 2006 Livingstone Declaration on social protection means nothing until it produces a real programme that delivers real benefits. Central to a social contract approach is the establishment of enforceable claims to social protection. If citizens have a right to demand protection from the state, for instance, they cannot be told they have ‘graduated’ and be removed prematurely from social protection programmes.

Terminology matters. The term ‘beneficiary’ is too passive and paternalistic (beneficiaries do not necessarily get their benefits in full and on time, and they might benefit more from a different or better designed programme – beneficiaries are rarely consulted about their priorities). The term ‘client’ is too closely associated with the commodification of social protection, whereas social protection should be about social solidarity and collective responsibility, which is obscured by the language of consumers and markets. The term ‘claimant’ carries several positive connotations. First, claimants are entitled to receive certain prescribed social protection benefits (grants and/or services), rather than a vaguely defined ‘right to social protection’ that governments can easily evade through recourse to the cop-out clause of ‘progressive realisation’. Second, if social protection benefits are not delivered to people who are entitled to receive them, mechanisms can (and should) be established to enable these people to claim them. These mechanisms include social audits, rights committees, grievance procedures, or even taking the government to court. Here the constructively adversarial role of domestic civil society can be powerful, and in South Africa civil society activism has achieved notable successes in extending the access and scope of interventions such as the CSG (Proudlock, 2011). Conversely, the 2009 decision by the Ethiopian government to pass the ‘NGO law’, effectively barring NGOs from campaigning for any rights at all, is a blow to the formation of a social contract between the state and its citizens and relegates civil society participation in social protection to the role of implementing agencies or service providers.

A third argument for the term ‘claimant’ and for enforceable claims is not just that eligible individuals should receive specified social grants and services but that these should be delivered in full, on time and at minimum inconvenience (e.g. in terms of distance to pay-points, queuing time, access costs and security risks). Poor and vulnerable people are entitled to be treated with dignity and sensitivity. So the claim should cover not only the benefit but also the way it is delivered, and the claimant should be able to hold the duty-bearer to account on both aspects. This approach is more real, more enforceable, more defined and broader than an abstract right to social protection. In Africa, where most governments have signed up to the right to social protection but social protection is often delivered in the form of projects funded by external donors and implemented by international NGOs, an enforceable claim bridges the gap between unaccountable donors and unenforceable rights.

Finally, seeing social protection as an enforceable claim requires the transfer of certain powers from those who deliver benefits such as social grants to those who receive them. This goes beyond allowing beneficiaries to participate, or consulting them; it implies empowering them as claimants. Lessons can be drawn from recent innovations in India, where social audits ‘create a collaborative and constructive platform for participatory governance of social protection programs … provide the most vulnerable with a “voice” to assert their “rights” [and] hold the village and local administration accountable’ (Vij, 2011:2).

Elements of empowerment are filtering into some African programmes. The Cash and Food Transfer Pilot Project in Lesotho, which delivered cash or food to drought-affected households in 2007/08, set up help desks where people who believed they were eligible but had been excluded could lodge their complaint. The Hunger Safety Net Programme in Kenya has established a rights charter that specifies the responsibilities of administrators and beneficiaries, and rights committees to whom both included and excluded people can complain if they feel badly treated or overlooked by the programme (Devereux & White, 2010). South Africa's new National Development Plan argues for introducing social audits to its social grants programmes, ‘not only because this would enhance the effectiveness of these programmes, but also because it would empower poor and vulnerable citizens and deepen the process of democratic inclusion’ (NPC, 2011:347). These are small but significant first steps towards making social protection in Africa not just more comprehensive, but more responsive, accountable and citizen-driven.