Rights-Based Social Protection in Africa A review of approaches, challenges and possibilities Zambia Office

ii Rights-Based Social Protection in Africa A review of approaches, challenges and possibilities Jeremy Seekings University of Cape Town Report for the Friedrich Ebert Stiftung May 2018

iii The Friedrich-Ebert-Stiftung The Friedrich-Ebert-Stiftung(FES) was founded in 1925 as a political legacy of Germanys first democratically elected president, Friedrich Ebert. Ebert was a Social Democrat from a humble crafts background, who had risen to the highest political office in his country. In response to his own painful experience in political confrontation, he had proposed the establishment of a foundation to serve the following aims:  Furthering the political and social education of individuals from all walks of life, in the spirit of democracy and pluralism;  Giving gifted young people scholarhsips to universities; and  Contributing to international understanding and co-operation. The FES, which was banned by the Nazis in 1933, and re-established in 1947, continues today to pursue those aims in all its extensive activities. As a private cultural non-profit institution, it is committed to the ideas and basic values of social democracy.

iv Published and Copyright by: Friedrich Ebert Stiftung Zambia Office 5583 Great East Road P.O. Box 30554 Kalundu, Lusaka Phone:+260 211 295615/6 info@fes-zambia.org www.fes-zambia.org Commercial use of all media Published by the Friedrich-Ebert-Stiftung(FES) is not permitted without the written consent of the FES. Readers are encouraged to quote or reproduce material for their own publications, as long as they are not being sold commercially. The findings, interpretations and conclusions expressed in this volume do not necessarily reflect the views of the Friedrich-Ebert-Stiftung.

v Rights-based social protection in Africa: A review of approaches, challenges and possibilities Jeremy Seekings University of Cape Town Report for the Friedrich Ebert Stiftung May 2018

vi

Contents Contents Acronyms Abstract Acknowledgements Introduction 1. The coverage and reach of social protection in Africa 1.1 Conceptualising and measuring the‘coverage’ of social protection 1.1.1 What counts as social protection? 1.1.2 Conceptualising‘coverage’ 1.1.3 Individual and family coverage 1.2 The scope of coverage 1.3 The extent of coverage 1.3.1 Data challenges 1.3.2 Coverage of the elderly 1.3.3 Changes over time in coverage of the elderly 1.3.4 Coverage of children 1.3.5 Coverage of working-age adults 1.4 Coverage gaps 1.4.1 World Bank data on coverage gaps 1.5 Expenditure on social protection 1.6 Conclusion: Who gets what, who does not, and how? 2. Rights for Whom in Global Rights Discourses 2.1 The origins of rights-based approaches to social protection 2.2 The promise and limits to‘development’, 1950 1996 2.3 The resurgence of a rights-based approach from 1996 2.4 The rights of peasants 2.5 Gaps in the global discourse of rights to social protection 3. The Vernacularisation of a Rights-Based Approach in Africa 3.1. The African Charter and subsequent Protocols 3.2. The AU, external actors and social protection 3.3. Sub-regional organisations 3.4. Nonstate organisations and workers’ rights to social protection vii vii ix xi xiv 1 8 10 10 13 15 17 20 20 21 23 25 27 28 30 33 35 40 41 45 49 52 57 59 60 63 71 73

8 3.5. Nonstate organisations and peasants’ rights to 80 social protection 3.6. Non-state organisations and the rights to social 82 protection of‘vulnerable groups’ 3.7. National constitutions 84 3.8. Assessing the vernacularisation of the global 90 rights discourses 4. Neither necessary nor sufficient: Case-studies of 93 rights and social protection at the national level 4.1. South Africa 94 4.2. Botswana 101 4.3. Zambia 109 4.4. Senegal 117 4.5. Lessons from the country case-studies 122 5. Explicit and implicit rights in the expansion of social protection 126 5.1. Elite conceptions of rights and responsibilities 5.2. Workers’ rights and the expansion of contributory 131 programmes 5.3. Rights, responsibilities and the expansion of non- 133 contributory programmes 5.4. Beyond social assistance for the deserving poor 135 References 137

ix Acronyms ALRN African Labour Research Network APSP African Platform for Social Protection AU African Union CAADP Comprehensive Africa Agriculture Development Programme(AU) CAMSD Conference of Ministers in charge of Social Development(AU) DfID Department for International Development(UK) FAO Food and Agriculture Organisation(of the United Nations) ICESCR International Covenant for Economic, Social and Cultural Rights ILO International Labour Organisation IMF International Monetary Fund ITUC-Africa International Trade Union Congress, Africa Region NDP National Development Plan NORAD Norwegian Agency for Development Cooperation OAU Organisation of African Unity SADC Southern African Development Community SIDA Swedish International Development Agency SPIREWORK Social Protection for Informal and Rural Workers STC-SDLE Specialised Technical Committee: Social Development, Labour and Employment UNDP United Nations Development Programme UNICEF United Nations International Children’s Emergency Fund WFP World Food Programme WIEGO Women in Informal Employment: Globalizing and Organising

x

xi Abstract Any discussion of‘rights based approaches to social protection’ requires us to ask: Whose rights are being recognized and against what risks? This paper makes five main points. First, although data on the coverage or reach of social protection in Africa is very uneven, and there is weak evidence of a recent expansion of coverage or reach, there can be little doubt that very large‘coverage gaps’ remain. Good data exist for some c ountries(such as South Africa), from both administrative sources(collated by government departments) and household surveys, and for several decades(allowing analysis of trends over time). The data on most African countries are poor; administrative data are unreliable, and survey data incomplete or dated. For some countries there are almost no data. Moreover, there is no consensus among scholars or observers as to what constitutes social protection. Coverage or reach is much higher if a more inclusive definition(including programmes that provide benefits in kind, including both routine and emergency feeding schemes) is used. In terms of social groups, workers in formal employment are generally covered unevenly but reasonably generously. Almost no informal sector workers and peasants enjoy coverage outside of famines or other natural disasters. They are likely to be eligible for benefits only when they fall into the category of‘vulnerable’ individuals, i.e. the‘deserving’ categories of the elderly and dis abled, children, mothers involved in fulltime carework), and often only if they are very poor(because programmes are meanstested). The result is that there is social protection provides some coverage for people at the top end of the income distribution(through social insurance) and some for people at the bottom(through targeted social assistance), but very little for the large number of people in between. Second, gaps in the coverage of social protection reflect gaps in the ideologies underlying and discourses of social protection. The‘global’ discourse of rights to social protection that has developed over almost a century is rooted in the conditions and ideas that prevailed in Western Europe. Rival traditions focused on social citizenship(rights for citizens, especially through tax-financed social assistance programmes) and industrial citizenship(rights for workers, primarily through contributory social insurance programmes), providing protection in different ways against the risks that characterized Western European societies in the first half of the twentieth century. These traditions came together in the 1940s, briefly, and again in the 2010s around‘social protection floors’. For most of the past one hundred years, however, the discourse of workers’ rights has been dominant globally, most recently through the emphasis on‘decent work’. The kind of worker accorded rights was the worker in formal employment; this concept of rights has belatedly been applied to informal sector workers, but without momentum, and it has never paid much attention to the peasants. The neglect of peasants prompted movements rooted in different parts of the global South around rights to food or to food sovereignty. Given that most working people in Africa are either in the informal sector or are peasant farmers, the global‘workerist’ conception of rights has clear limitations in Africa and contributes to wide coverage gaps.

xii Thirdly, whilst Africa and Africans have generally been peripheral to the evolution of the global discourses of rights to social protection, the global discourse has to some extent been‘vernacularised’ in Africa, through the Charters, Protocols and other statements adopted by the Organisation of African Unity and African Union, the activities within Africa of the International Labour Organisation and other international agencies, and the efforts of regional civil society organisations. Some aspects of the global discourse have been resisted, and others adopted but with the addition of distinctively African concerns, including strong emphases on the family(as well as the individual) and on responsibilities (as well as rights). The Africanised version of the global discourse of rights to social protection remains incompletely vernacularized, however, in that it remains largely silent on the rights of peasants and rights to food. Gaps remain in the ideologies and discourses of social protection that prevail in Africa. Fourthly, case-studies of struggles around rights and social protection in selected countries show that the explicit acknowledgement of rights to social protection(in the sense used in the global rights-based approach) is neither necessary nor sufficient for an expansion of access to social protection on the part of citizens. Social and economic rights are included in the South African constitution, used by civil society, and supported by the public, but South Africa’s welfare state largely predated the formal recognition of rights. Botswana also has an extensive, if conservative welfare state, but most of the political elite is deeply ambivalent about the global approach to individual rights. Social protection in Botswana has no basis in either the national constitution nor even legislation. International agencies, NGOs and aid donors have more influence in Zambia, which is much poorer than either Botswana or South Africa, but even there they run into resistance in promoting their rights-based approach. The Francophone case of Senegal appears to be more similar to Zambia. The concept of the s ocial protection‘floor’ points to two strategies that might facilitate the expansion of social protection in Africa. On the one hand, the expansion of the‘vertical’ dimension through consolidating contributory programmes among formally-employed workers a nd expanding them to at least some informal sector workers offers a‘top down’ route to more universal social protection. On the other hand, the expansion of the ‘horizontal’ dimension – i.e. the floor – through tax-financed social assistance programmes of fers a‘bottom up’ route to more universal social protection. Whilst the former is undoubtedly important to prevent working people falling back into poverty, the latter offers much more potential for reducing poverty, especially extreme poverty. An explicitly rightsbased approach to workers’ rights is likely to strengthen the former strategy. An explicitly rightsbased approach to citizens’ rights is less likely to strengthen the latter strategy, because of the strong antipathy among political and economic elites across much of Africa to the global discourse of social and economic rights. This does not mean, however, that there is no potential for a strategy framed in a normative discourse. The predominant understanding of welfare among political elites, at least in East and Southern Africa, has an alternative normative basis, emphasising responsibilities rather than rights, and relationships rather than individual entitlements.

xiii This explains why there is deep opposition to the provision of public assistance to working-age adults, except through workfare or in times of emergency, and why there is widespread anxiety over‘dependency’ on‘handouts’. Elites show some interest in the extension of contributory social protection to informal workers, through cooperative mechanisms, but without much enthusiasm. Elite beliefs do, however, open up the possibility of expanding social protection for specific categories of people – especially the elderly and children – when kin have otherwise failed in their responsibility to provide support. Concepts of responsibility and opportunity point to the possibility of an effective, implicitly rights-based approach to expanding social protection. Even such an approach has its limits: Provision for‘workers’ and for the deserving poor leaves open the challenge of addressing poverty among peasant farmers. More attention needs to be paid to the rights of peasant farmers and to the responsibilities of society to them. The application of rights-based approaches to social protection in Africa thus entails both a challenge and an opportunity. The challenge is to vernacularise effectively the global rights discourse, taking into account the conservatism of African elites and the fact that poverty is most pronounced among peasant farmers. The opportunity is to build on the implicit recognition of rights entailed in the ideology of family, responsibility and opportunity.

xiv Acknowledgements I am grateful to the Friedrich-Ebert-Stiftung(FES) Zambia Office for the opportunity to think through the status of social protection in Africa, which resulted into a Scoping Paper to guide implementation of the Rights Based Social Protection Project. Much gratitude should be given to the Consultant and the research team who assisted with the initial research. Furthermore, I would like to extend my appreciation to the Steering Committee Members- WIEGO, ITUC-Africa, APSP and ARLN implementing the Rights Based Project in collaboration with FES, who gave valuable and critical comments to the study by providing critical additional information. Finally, special thanks go to the FES Zambia management and staff for their support to facilitate production of this study. Their commitment and patient are highly appreciated.

1 Introduction ‘[H]uman rights talk should be historically situated and socially specific … Any debate conducted on the level of moral absolutes or universal humanity is not only fruitless but ideologically subversive of the African masses.’(Shivji, 1989; quoted in England, 2006: 27) Over the past twenty-five years, there has been a dramatic proliferation of social protection programmes across much of Africa. The coverage or reach of social protection remains uneven, however, both within and between countries. In a few countries, a significant proportion of the population – and especially of the poor population – has been deemed deserve social protection. In most, social protection programmes cover or reach only a small or even very small proportion of the population. Even in countries with more extensive systems of social protection, some groups of people are not covered. South Africa, for example, makes no direct provision for unemployed working-age adults, whilst Botswana does not support all poor parents with children. Some of the models of reform turn out, on close examination, to be much less impressive than they initially appeared. The persistence of coverage gaps does not reflect an absence of proposals. In most, perhaps all African countries, governments have rejected proposals to expand social protection that their advocates – including major international organization – have argued are affordable. Overall, as we shall see further in Chapter 1,‘coverage gaps’ in Africa remain both wide and deep. One possible strategy to address these persisting coverage gaps is to emphasise rights to social protection. In this view, if national governments can be persuaded, cajoled or shamed into recognizing citizens’ rights to social protection, then citizens will be empowered in demanding and securing the implementation of social protection programmes. Internationally, enthusiasm for a rights-based approach to social protection has grown, beginning in the early 2000s. Diverse international organisations(including the International Labour Organisation[ILO], UNICEF and global NGOs such as HelpAge International) and government agencies(such as the UK’s Department for International Development[DfID]) have advocated the specification of social protection as a right, meaning that individual citizens should have constitutional or other legal rights to claim from their states sufficient benefits, whether in cash or in kind, to allow them to meet a minimum set of needs. It is argued that such rights should be universal, that there should be no discrimination in the selection of beneficiaries and the payment of benefits, except perhaps on the basis of need through targeting. Even targeting should be used sparingly, and states should commit themselves to the‘progressive realisation’ of u niversal coverage without targeting(Sepúlveda and Nyst, 2012). Moreover, as the then UN Special Rapporteur on Extreme Poverty and Human Rights, Magdalena Sepúlveda, put it in a strong statement for a rights-based approach, beneficiaries should be able‘to participate in the design, implementation, monitoring and evaluation of social protection programmes’(Sepúlveda and Nyst, 2012: 59). Such arguments have been taken up by African regional organisations, including the African Union(AU). Rights to social

2 protection are enshrined in the constitutions of several African countries, including South Africa and Kenya. This rights-based approach to social protection has been and is resisted, within some international organisations(including, notably, the World Bank) and, especially, national governments across much of Africa. Even when a rights discourse is used, it need not reflect any substantive commitment. In 2017 we attended a workshop where a government minister with partial responsibility for social protection read a speech that set out a vision of reforms – of both specific programmes and the overall policy framework – that suggested a clear a commitment to a rights-based approach to social protection. The previous day, in a private meeting, she articulated a very different, and much more conservative, view of what the government should and would do. Her public speech had, of course, been scripted by a junior official; the Minister may not have read it prior to the workshop. The speech did not reflect her private views. Nor, in all likelihood, did it provide a meaningful indication of her government’s reform programme. The‘global’ discourse of rights to social protection was reflected in‘her’ speech, but not in her private values and beliefs. In this particular country, social protection policy documents, mostly drafted by expatriate European consultants, promise much, but the policy-making national elite has shown little interest in legislating or implementing substantive reforms. Our research across much of Anglophone Africa suggests that this was not an isolated instance. Across much of the region, politicians read speeches or approve of documents that either employ a discourse of rights or imply that their governments recognize that their citizens have social and economic rights. Speeches and documents are typically written by technocratic officials, often with considerable assistance from international organisations. In private, however, politicians eschew a discourse of rights. They usually endorse programmes that extend privilege to selected groups of workers in formal employment, but they tend towards parsimony and paternalism in their treatment of the rural and urban poor. The explicitly rights-based approach favoured by many(but not all) internatio nal organisations had yet to gain much support among Africa’s political elite. The expansion of social protection in Africa has rarely been accompanied by an explicit discourse of rights, but has rather been framed through alternative discourses. As stand ard accounts of the ideological basis and discursive framing of‘social protection’ have noted, arguments about rights comprise only one of several alternative sets of justifications or frames for social protection. Munro(2008: 27), for example, identifies doctrines of rights as one of three sources of justification for the welfare state, alongside the‘analysis of uninsurable risks and other market failures’ and‘needs based doctrines’. Similarly, von Gliszczynski(2015; also von Gliszczynski and Leisering, 2016) identifies a ‘human rights discourse’ as one of four‘global discourses of social protection’, i.e. sets of ‘ideas behind cash transfers’; his other three discourses focus on poverty, development and risk. Political elites in Africa rarely refer e xplicitly to citizens’ social or economic rights. Such accounts tend to focus on the kinds of discourses of rights that are explicit in international conventions and may be replicated or reflected in domestic policy documents or debates. In terms of explicitly documented discourse, these analyses are

3 clearly correct. Rights are not mentioned in many of the documents(for example, from the World Bank) that argue for social protection. This does not mean, however, that these other discourses, that do not refer to or focus on rights, have no normative basis. Their normative basis might not be explicit, but all discourses of social protection entail implicit norms about distributive justice. Moreover, opposition to – or the constraints on –‘global’ and explicitly rights-based approaches to social protection often do not entail an opposition to rights per se, but are rather rooted in alternative and generally implicit norms and values. Contrasting approaches to social protection are often rooted in contrasting understandings of the rights and obligations that constitute the social contract between government, kin and individual citizens. This report examines rights-based approaches to social protection in Africa through not only an analysis of the norms and values that are explicitly set out in the documents produced by international agencies and organisations (see Chapter 2), but also through an analysis of the norms and values that are implicit in the arguments put forward, often verbally rather than in documents, by national and regional policy-makers. Across much of Africa, national policy-makers are wary of the ‘global’ approach to rights. They tend to hold a different understanding of distributive justice, rights and obligations. Our analysis leads us to th e conclusion that the‘global’ approach to rights is not the only understanding of rights that can(and does) justify social protection. International agencies and organisations based in the global North might engage more fully with the understandings of(or approaches to) rights and obligations that prevail among policy-making elites in Africa. This report provides a casestudy of what Levitt and Merry called the‘vernacularization’ of human rights.‘To have impact’,’ they write,‘human rights ideas must be adopted locally, must transform the consciousness of those who claim them and have some institutional teeth so that people who demand rights are at least recognized if not satisfied’ (2009: 457). ‘Rights ideas and practice ne ed to resonate with existing ideologies to be adopted, but to be legitimate as human rights they have to reflect universal principles or standards. It is the claim to universality that provides‘the magic’ of human rights for activists and establishes its transcendental character and legitimacy. In other words, the non-local dimension that points to global universals is precisely what makes human rights discourse politically powerful. This dilemma is inherent in the human rights project. More homegrown programmes emphasize localization, while more imported, cosmopolitan ones emphasize universality.’( ibid) Levitt and Merry also point to an‘advocacy dilemma’:‘ When organizations use human rights in ways that join readily with existing issues and strategies, they are more readily accepted but represent less of a challenge to the status quo ’( ibid: 457-8). This explains why‘the process of vernacularization contains more friction than flows’( ibid: 448). The‘diffusion’ into Africa of the global human rights d iscourse on social protection has been widely contested, although this contestation is hardly mentioned in the existing literatures. Based on case-studies of the vernacularisation of rights in other parts of the world, Levitt and Merry write that‘framing human rights claims in local terms and adapting

4 them to existing ideas of justice may mean abandoning explicit references to human rights language altogether and, indeed, can mean highjacking these concepts for quite different purposes’( ibid). In the African context, many fervent supporters of human rights – from both within and beyond Africa – often downplay them in their engagement with policymaking elites. In this paper we suggest that supporters of human rights might go further, in engaging with the values and beliefs of African policy-making elites in order to appropriate(or‘hijack’) them for the purpose of promoting the interests of the poor and of strengthening a rights-based culture(even if it is not precisely the same as the global human rights culture). This report begins(in Chapter 1) with an analysis of the coverage and reach of social protection in Africa. We review critically the conceptualization and measurement of social protection by the ILO, the World Bank and independent scholars. Estimates of the coverage and reach of social protection – and hence of resulting‘coverage gaps’ – vary, depending on the choice of definition of social protection, the data used to measure coverage and reach, and a series of methodological assumptions. What is clear, nonetheless, is that there are massive coverage gaps across Africa. In some respects, coverage is widespread, but parsimonious. Peasant farmers are widely protected against natural disasters(especially drought), and school children are protected against poverty through school feeding programmes. But large majorities of people across Africa face the risks of old age, illness or disability, un- or under-employment with little or no prospect of support from the state. In Chapter 2 we turn to the evolution of the‘global’ discourse of rights with regard to social protection, i.e. to the explicit concern with rights on the part of international organisations. We focus on the question of‘rights for whom?’. We consider the ways in which the rights of different categories of people – workers(and their dependents), citizens and peasants – have been addressed in the unfolding global rights discourse. Our analysis contrasts with most other accounts in tracing the concern with workers and citizens’ right s back to the period prior to 1948, and especially through the activities of the ILO, and in discussing the rights of peasants. We show how the global discourse of rights to social protection has evolved through concern with specific categories of people(first workers, and later ‘vulnerable groups’) whilst neglecting others(especially peasant farmers). The third chapter of this report examines the‘Africanisation’ of this‘global’ rights discourse through the activities of the Organisation of African Unity(OAU) and then AU, regional organisations, external actors such the ILO and HelpAge International, and regional civil society structures(especially the African Platform for Social Protection [APSP]). This chapter thus focuses on the vernacularisation of social and economic rights at the supra-national level of Africa as a whole(or of regions within Africa). The chapter notes the very limited role that African elites played in the making of global rights discourses, examines the translation of the global discourse into African regional covenants and the activities of transnational organisations. Chapter 4 turns to the vernacularisation of rights discourses at the national level within selected African countries. We examine the extent to which, and ways in which, the new

5 ‘global’ discourse of rights has been incorporated into national strategy and policy documents, and into the thinking of local elites. The chapter focuses primarily on four case-studies: South Africa, Botswana, Zambia and Senegal. These cases have been selected purposively. South Africa is the case where social and economic rights are most fully institutionalized, not only in the Constitution and legislation, but also in court cases, political rhetoric and even popular discourse. South Africa has a social assistance system that is both unusually extensive and unusually generous. Botswana also has an extensive system of social assistance, but it is parsimonious rather than generous, and the government and ruling party have long eschewed any mention of social or economic rights. Whereas South Africa has adopted a classically‘rights based’ approach, Botswana’s welfare state is based on very different norms and values. Zambia and Senegal provide Anglophone and Francophone cases in which both programmes and discourses of social protection are fluid and contested. The final chapter examines various strategies that might be employed in an effort to promote the further expansion of social protection in Africa. The chapter considers the rights-oriented strategies employed by a variety of civil society organisations. It also considers the norms and values regarding social protection among national policy-making elites, considering the case of Zanzibar in addition to the cases discussed in the previous section. The chapter concludes by drawing out some of the implications of this analysis, pointing to areas in which there is an overlap between the global rights agenda and the norms and values of political elites across most of Africa. This report argues that rights discourses, whilst neither necessary nor sufficient, can play an important role in the expansion of social protection to fill in some of the massive coverage gaps that currently exist. In order to be effective, rights discourses must be app ropriate, i.e. they must be‘vernacularised’ appropriately. The effective vernacularisation of rights discourses entails both cultural and socio-economic considerations. Culturally, rights discourses are likely to be more effective if they incorporate or make use of existing norms and values. In contemporary Africa, this might mean attacking more importance to a discourse of responsibility than to an explicit emphasis on rights. Effective vernacularisation is not simply a matter of cultural difference, however. With respect to social and economic rights, the fact that individuals, families and societies in different parts of the world face different kinds of risks requires different kinds of rights to protect them against those risks. In the advanced capitalist societies of Western Europe, social and economic rights and citizenship entail a demand for decommodification of labour, so that citizens or workers are not dependent upon their current sale of their labour. In some parts of the global South, the commodification of land and nature have generated comparable demands for decommodification. But across much of Africa the predominant risks cannot be reduced to the commodification of labour. Poverty is rooted more widely in low agricultural productivity, the weak demand for offfarm and the risks of drought or floods. In these circumstances,‘social protection’ should not be understood in narrow terms, referring only to contributory(social insurance) or noncontributory(social assistance) programmes that pay benefits in cash. In Africa, social protection encompasses also programmes that distribute food, in kind, whether through regular feeding programmes(especially, but not only, in schools) or emergency relief(in

6 times of famine). As the FAO suggests, there are also various agricultural programmes that have‘social protection functions’, including forms of crop insurance, livelihood support packages and even agricultural extension and marketing services(FAO, 2017: 30). There is a strong push towards compiling unified registers of the beneficiaries of all relevant programmes, in countries such as Malawi and Zambia. The fact that these registers will record the beneficiaries of peasant-support programmes attests to their role as a form of social protection. The effective vernacularisation of rights requires consideration of rights for whom as well as the discourse that is employed to advance a rights-based agenda.

7

8 Chapter 1: The Coverage and Reach of Social Protection in Africa It is commonplace in studies of social protection across the world to emphasise that Africa has not been left out of the global movement of expansion over the past twenty or so years. Hanlon, Barrientos and Hulme(2010) included examples from Africa in their more general analysis of the rise of social protection as a‘quiet revolution from the South’. The World Bank documented the proliferation of new social assistance programmes in Africa in the 1990s and 2000s(Garcia and Moore, 2012) into the early 2010s(World Bank, 2015: 10). In a foreword to its 2012 report, a senior World Bank staffer wrote that‘enthusiasm for conditional cash transfer programs in other regions has spilled over into the continent’, such that‘many policy makers are excited about how cash transfers can be used to meet Africa’s poverty and development goals’(Garcia and Moore, 2012: xiii). More recently, the International Labour Organisation(ILO)(2017) reports‘notable progress in the extension of social protection coverage’ in Africa (2017: 119), noting that ‘most African countries have made social protection a priority in their development strategies’ and have adopted or are developing national social protection policies and plans(2017: 131). Echoing Hanlon et al., the ILO identifies African countries with universal pension coverage as‘home grown’ schemes,‘developed by innovative African governments’( ibid: 121). 1 The Social Assistance Politics and Institutions(SAPI) project counts the number of social assistance programmes in Africa rising more than fourfold from about 20 in 2000 to about 85 in 2015, with coverage rising fivefold over this period(Hickey et al., 2018). In their foreword to a recent study of the use of evidence in policy reform, the heads of the FAO and UN ICEF emphasised that‘virtually every country in Africa has some kind of cash transfer programme’(Davis et al., 2016: v). Specific African programmes are cited as models: Ghana’s LEAP(Hanlon et al ., 2010: 45; ILO, 2014: 22), South Africa’s old age pensions, child grants(ILO, 2014: 22) and disability grants(Hanlon et al., 2010: 38-9) and Ethiopia’s Productive Safety Nets Programme(PSNP)(Hanlon et al., 2010: 47). Even in countries with poor coverage until recently, such as Tanzania and Senegal, safety net coverage is reported to have expanded rapidly in the mid-2010s(World Bank, 2017). These same studies acknowledge that, whilst not being left out altogether, Africa is being left behind. Social protection covers only a minority of either the total population or the poor. The international agencies present data that suggest that Africa has been a global laggard in terms of both the coverage of and expenditure on social protection(ILO, 2014, 2017; World Bank, 2015). The World Bank noted in 2015 that the coverage gap among the poor was most acute in Sub-Saharan Africa, with only one in ten people in the lowest income quintile covered by social safety nets(World Bank, 2015: 1). In its most recent report, the ILO reports notable progress, but it immediately qualifies this, noting that‘only 17.8 percent of the population are covered by at least one cash social protection benefit’ (ILO, 2017: 119). 1 The ILO identifies Algeria, Cabo Verde, Mauritius and South Africa. Neither the Algerian nor Cabo Verde cases have been documented, but the pension systems of Mauritius and South Africa were based on the British model and cannot be considered as home-grown ‘African’ innovations(Seekings, 2007, 2011).

9 This is not because of any lack of need. In Africa, contrary to global trends, the absolute number of people in poverty has actually been rising(although the poverty rate has fallen slowly – Beegle et al., 2016). The number of elderly people has been rising at the same time as families are less and less willing to care for them. Birth rates remain high. Across much of East and Southern Africa, AIDS resulted in rising populations of orphans. Child poverty and deprivation persist. Landlessness has fueled unemployment, especially among young adults. Chronic poverty is compounded by drought and – across much of the Sahel – war. The international agencies‘talk up’ progress in Africa in their enthusiasm for social protection. Since the early 2000s, one after another of the international and donor agencies that are active in Africa embraced social protection as an integral element in their approaches to development and the reduction of poverty(Leisering, 2009; De Haan, 2014; von Gliszczynski, 2015; von Gliszczynski and Leisering, 2016; Hickey and Seekings, 2018). Reports emphasise that the evidence from Africa as well as elsewhere shows that cash transfers can reduce poverty and, it is claimed, encourage more widereaching‘development’, and cite examples from Africa so as to persuade African policy makers to emulate these models(see also Davis et al., 2016). International organisations and their allies across Africa are right to focus on the need to persuade African policymakers of the merits of social protection. In many countries, presidents and ministers have resisted the kinds of reforms advocated by international organisations, citing concerns over either their affordability or the‘dependency’ on social grants that will undermine the willingness of poor people to work. 2 As tools in campaigns of advocacy, the international organisations’ reports should not be read as disinterested or entirely critical studies of the policy landscape. These reports do, however, provide a useful starting-point for assessing the coverage of social protection in Africa, the‘coverage gaps’ that remain, and changes over time in both coverage and gaps. This chapter surveys data on coverage by examining critically the data provided in the most recent World Social Protection Report published by the ILO (ILO, 2017), contrasting the ILO data with those presented by the World Bank and indepen dent scholars. I argue that the ILO’s estimates remain preliminary and open to improvement. First, the definition of social protection used in practice by the ILO is both ambiguous and too restricted, excluding especially programmes that provide benefits in kind(including school feeding schemes) or through workfare(even when participants receive cash benefits). Secondly, the ILO’s regional estimates seem to be based on a possibly flawed process for the imputation of missing country-level data. Thirdly, some of the country-level data that are used seem to be suspect, perhaps because data are collected through government ministries. The ILO appears to exaggerate both existing coverage(by its own definition) and the pace of expansion over recent years, whilst underestimating coverage(by a more inclusive definition). Whatever the precise extent of existing coverage, it is clear that large gaps remain: Most elderly people in Africa do not receive pensions; most poor children receive little or nothing; and peasant farming families and workers in the informal economy remain largely uncovered. States protect 2 See, for examples, Kabandula and Seekings(2016) on Zambia and Ulriksen(2016) on Tanzania.

10 their citizens against the most extreme forms of poverty(especially famine) but do little to mitigate more routine and chronic poverty and disadvantage. 1.1. C onceptualising and measuring the‘coverage’ of social protection Identifying precisely the‘coverage’ of social protection, and hence the gaps that remain, is very difficult, given the paucity of data on many African countries. The most recent detailed study of coverage across Africa – and elsewhere across the global South(and North) – is the ILO’s World Social Protection Report. Having published its first World Social Protection Report in 2014, the ILO published a second World Social Protection Report in late 2017. 3 The World Social Protection Reports use the ILO’s World Social Protection Database, which itself draws primarily on data collected from each national government 4 through a standardized‘Social Security Inquiry questionnaire’(ILO, 2016a). 5 The 2017 Report uses data collected in 2016, usually covering one or other year between 2012 and 2016. These data are supplemented with data from other sources. The Report provides much more detailed coverage of old-age pension programmes, especially social pensions(perhaps because it uses data collated by HelpAge International). Annex II to the Report provides a thorough account of the sources of data and definitions used. Annex IV comprises a series of tables setting out for most countries in the world aspects of the coverage of social protection. The data in these tables, as well as the data used to generate figures in the text, are available online from the ILO’s Social Protection Platform. 6 The ILO has put into the public domain an invaluable resource. 1.1.1. What counts as social protection? Data on the coverage or reach of social protection requires choices as to what counts as social protection, i.e. what range of schemes are taken into account and which schemes are not. The ILO’s conceptualization of‘social protection’ reflects its history as an organization concerned primarily with the regulation of employment. The ILO reports that it defines social protection as the‘set of policies and programmes designed to reduce and prevent poverty, vulnerab ility and social exclusion throughout the life cycle’(2017: 194). Vulnerability and social exclusion are neither defined nor(therefore) measured in the 2017 Report. In practice, the ILO uses a narrower definition of social protection. As the ILO itself w rites, it uses the term‘social protection’‘largely interchangeably’ with the term ‘social security’( ibid: 194-5), i.e. a term that the ILO used routinely between the 1940s and 2000s. Social protection thus encompasses both social insurance(i.e. programmes that insure people against specified risks, such as disability or illness, through 3 The ILP previously published a World Social Security Report in 2010(ILO, 2010). 4 Sections of the questionnaire on schemes are completed by Ministries of Labour or Welfare, as appropriate, whilst the sections on expenditures are completed by Ministries of Finance(ILO, 2016a). 5 The Social Security Inquiry Manual(including the questionnaire) is available online: http://www.ilo.org/wcmsp5/groups/public/---dgreports/--integration/documents/instructionalmaterial/wcms_568711.pdf. 6 http://www.social-protection.org/gimi/gess/ShowWiki.action?id=9. This was previously the Global Extension of Social Security(GESS) Platform. The ILO’s Social Security Inquiry website contains country - and programmespecific data, but only until about 2010: http://www.ilo.org/dyn/ilossi/ssimain.home.

11 contributions, almost always linked to formal employment) and social assistance(i.e. ‘non contributory’ tax or aid -financed programmes, usually targeted on the poor but sometimes‘universal’ within specified categories such as the elderly). In practice, as we shall see in Chapter 2, the ILO has for most of its history been preoccupied with social insurance, as part of its concern with the regulation of employment, with only a secondary interest in social assistance. The ILO’s definition of social security focuses on income security in the face of specific risks, together with access to health care. Social security never encompassed the full range of politics‘designed to reduce and prevent poverty’. Instead, it focused on a narrow understanding of‘work’. Social security(and hence social protection) focuses primarily on workers in formal employment(and their dependents) with some consideration given to workers in informal employment(and their dependents). The Report vaguely refers to people with‘fragile livelihoods as workers and entrepreneurs’( ibid: 25), but never refers to peasants, and refers to farmers only three times in the entire text. 7 The ILO pays no attention to farmer support programmes, including programmes that ensure income security for farmers through stabilizing the prices of agricultural produce or raise productivity through subsidized fertilizer or other inputs. The Food and Agricultural Organisation(FAO), in contrast, recognizes that many such agricultural programmes have‘social protection functions’(FAO, 2017: 30). Many of these should count as programmes‘to reduce and prevent poverty, vulnerability and social exclusion throughout the life cycle’, but do not fall under the ILO’s restrictive definition of social security. The ILO’s definition of social security also seems to exclude workfare programmes where the benefits are conditional on work. The ILO itself recognizes the difficulties in assessing social protection against the risk of‘unemployment’. In the text, the 2017 Report covers employment guarantee schemes(notably the Mahatma Gandhi National Rural Employment Guarantee Scheme[MGNREGS] in India) and other public employment programmes(such as the Productive Safety Nets Programme[PSNP] in Ethiopia or the Expanded Public Works Programme[EPWP] in South Africa)(ILO, 2017: 45). In estimating the coverage rate for the MGNREGS, the 2017 Report uses the proportion of the labour force in rural areas, and hence insured against the risk of unemployment(at least up to a limit in terms of days of work per year). This gives a coverage rate of 24 percent( ibid: 46, footnote 11), which raises considerably the overall coverage rate in South Asia(see Figure 3.16). In Table B.6 in the statistical annex, the ILO seems confused as to how to treat MGNREGS. 8 It is not clear how the ILO assesses the coverage rate of public employment programmes in Africa. Whilst(as far as I know) no African public employment programmes provide an absolute employment guarantee, as in rural India, but many African workfare programmes provide for a reasonably high probability of low-wage employment in the event of unemployment. Table B.6 seems to omit both the PSNP for Ethiopia and the EPWP for South Africa. Figure 3.16 in the ILO’s 2017 Report implausibly suggests that the‘inclusion’ of public employment programmes 7 See pages 50(in a box on Irish unemployment insurance), 82(Brazil) and 155(Kiribati); also, in tables, see pages 297, 336, 344 and 345. 8 Table B.6 itself lists the employment guarantee scheme as one of the forms of unemployment insurance in India, but a footnote reports that the participants in MGNREGS have not been included in the reported data.

12 would make no apparent difference to the coverage of social protection in Africa. This is implausible because workfare programmes in Africa benefit at least 16 million people 9 (World Bank, 2015; see further below). Excluding workfare is ironic given that the ILO has always viewed social insurance programmes at the core of social security. Social insurance schemes tie benefits to a past record of contributions, almost always linked to(formal) employment. The ILO therefore includes under coverage workers whose current employment-funded contributions render them eligible for future benefits, and current beneficiaries whose eligibility depends on past employment, but excludes workers who are eligible for a cash transfer because they are participating now in a workfare programme. The ILO focuses primarily but not exclusively on cash benefits, to insure workers and their dependents as well as retired workers against risks that might terminate or reduce their incomes. The ILO grudgingly acknowledges benefits in kind but is clearly hostile to them. This is generally not an issue for older people across the global South, who rarely receive public benefits in kind(not least because few if any states across the global South provide any significant personal care services). The ILO’s primary focus is more of an issue, however, for working-age adults and children, many of whom receive benefits in kind rather than in cash. The chapter in the 2017 Report on social protection for children explicitly focuses on cash benefit programmes whilst acknowledging that many countries provide benefits in kind, especially through school feeding programmes(2017: 14). The Report cites an estimate by the World Food Programme that, globally, at least 368 million children(i.e. one in five of the world’s children) were fed daily at school, at a total cost of about US$75 billion p.a.(see WFP, 2013). Many African countries operate school and other feeding programmes. The largest, in South Africa feeds almost 10 million children. Using WFP data, the World Bank lists school feeding programmes in African countries that reach a total of about 30 million children(World Bank, 2015). 10 Many countries also operate extra-school feeding programmes for pre-school children. The number of children benefitting from school feeding programmes in Africa is probably double the number of children registered for cash transfer programmes. The ILO data ignores almost entirely famine relief, usually delivered in kind as food aid. Drought, floods and other‘natural’ disasters are the primary risks facing rural people across much of Africa. One of the most dramatic achievements since Independence across much of Africa is that functional states guarantee a minimal level of food security. Very few people die during famines. Famine relief provided the foundations of social protection across much of Africa(and South Asia). During severe droughts, famine relief is sometimes provided to half or more of the population in the countries concerned. The World Bank(2015) lists feeding programmes, excluding school feeding programmes, that reached 20 million people across Africa. This figure varies from year to year, depending 9 It is not clear from the World Bank data whether(or perhaps when) they refer to individual participants and when they include also their dependents, i.e. all members of participants’ households. 10 A more recent study(Drake et al., 2016) includes detailed case-studies of school feeding programmes in nine African countries(Botswana, Cabo Verde, Cote d’Ivoire, Ghana, Kenya, Mali, Namibia, Nigeria and South Africa).

13 on the incidence of famine, but never drops to zero because some feeding programmes are ongoing. The ILO seems to focus on government programmes, and the ILO has a strong preference for programmes that are‘anchored in national legislation’. The ILO measures (inter alia) the legal ‘scope’ of social protection(as we shall see in the next section) but does not restrict its analysis of coverage to programmes that have a legislative basis. Moreover, it is unclear where in practice the ILO draws the line on what constitutes a government programme. Does it include, for example, the Tanzania Social Action Fund (TASAF), which has been funded by the World Bank(and, now, other donors also), but is notionally managed by a government agency? 11 The focus on government programmes might help to explain why drought relief – often provided through the World Food Programme(WFP) – seems to be largely ignored, although in some countries drought relief is managed by and through government departments. The World Bank pays no heed to legal status or government role. The ILO de facto definition of social protection – government-run, cash transfer programmes that do not entail any immediate labour obligation – contrasts with the definition of‘social safety nets’ used by the World Bank in its State of Social Safety Nets reports. The World Bank defines social safety nets as‘noncontributory measures designed to provide regular and predictable support to poor and vulnerable people’, whether in cash or in kind(World Bank, 2015: 7). The World Bank views the term‘social safety nets’ as synonymous with social assistance or social transfers. In contrast to the ILO, the World Bank includes both programmes with benefits in kind(such as feeding schemes) and workfare programmes. The World Bank excludes, however, generalized subsidies, i.e. subsidies(for example, of fuel) that are not targeted on the poor and are usually regressive in effect. The World Bank views social safety nets as a large part of ‘social protection’, along with social insurance. Because the World Bank’s brief is to reduce poverty, and social insurance is almost always regressive, the Bank focuses almost entirely on social assistance, i.e. social safety nets. 1.1.2. Conceptualising‘coverage’ The ILO’s concept of‘coverage’ also reflects its historic preoccupation with the regulation of labour.‘Coverage’ refers primarily to the risks against which workers(and their dependents) are insured. For the ILO, therefore, coverage rates refer primarily to the proportion of people who are insured directly or indirectly through contributions to social insurance funds. The ILO recognizes that people who contributed in the past and are now receiving the benefits are also‘covered’, so the coverage d ata include both contributors and beneficiaries. As the World Bank(2015: 77) notes, the concept of‘coverage’ makes 11 The ILO seems to operate on an assumption that is made explicit by the SAPI team, that there is a clear distinction between‘social assistance’ and‘emergency assistance’. For SAPI, emergency assistance – including drought relief – is charity, addressing short-term misfortune, outside of any legislative framework, and typically operated by NGOs. In contrast, for SAPI, social assistance programmes are long-term, operated by government, within a legislative framework(Armando Barrientos, personal communication, January 2018). This distinction seems to me to be very unclear, with a large number of programmes located in the murky area between these ideal types.

14 no sense with respect to social assistance;‘reach’ would be a more appropriate term (Barrientos, 2018). A universal social pension should reach 100 percent of elderly people, and thus covers the entire population against the risk of old age. It would be bizarre, however, to say that the entire population is covered by social protection. Below we shall see how the ILO contorts itself to get around this problem. Its history has also led the ILO to distinguish between three dimensions of coverage:  Scope , i.e. the number of different‘areas’ or‘branches’ of social protection, originally linked to specific risks such as unemployment or work-related injury;  extent, i.e. the proportion of the relevant population that is covered; and  level, i.e. the generosity of benefits. Historically, the ILO’s dream was that all workers in formal employment would be insured adequately against a set of specified risks. The scope of social security was thus as important an indicator of coverage as its extent and level. The identified risks reflected the ILO’s preoccupation with workers in formal employment in the more industrialised countries that were later categorized as the global North, and the specific risks that these workers faced. The ILO’s Social Security(Minimum Standards) Convention(number 102, 1952) identified eight‘contingencies’ or risks:  illness(requiring both medical care and benefits, covered in Parts II and III of the Convention),  unemployment(IV),  old age(V),  employment injury(VI),  children(i.e. family benefits, Part VII),  maternity(VIII),  invalidity(IX) and  the death of the breadwinner(i.e. benefits for‘survivors’ or dependent widows and children, Part X). These eight contingencies continue to frame the ILO’s work, including the collation of data in its 2017 Report(see ILO, 2017: Table B.2, for example). 12 As we shall see further below, there is no provision here for insurance against the risk of drought, that has been the primary risk facing most peasant farmers across much of Africa. Nor is there any provision against the risk of armed conflict, that has been a pressing concern across much of Africa and remains a major concern across much of the Sahel. The ILO was always somewhat concerned with the extent of coverage, both across occupations and national borders, but the extent of coverage became much more important when the ILO began to acknowledge that generous contributory programmes covering most or all formally employed workers might nonetheless leave massive coverage gaps because only a small proportion of the population was in formal employment. The ILO’s new f ocus in the 2000s on the extent of coverage has led to the ILO embracing taxfinanced,‘non contributory’ social assistance alongside social 12 The Social Security Inquiry Manual reports that data is collected, however, for nine risks. These are:(1) old age, disability and survivor pensions;(2) sickness;(3) maternity and paternity;(4) family and children benefits;(5) unemployment benefits and public works programs;(6) employment injury;(7) health care;(8) other social assistance; and(9) social work services(ILO, 2016a: 9).

15 insurance, and to a concern with provision for all stages of the life course. The main body of each World Social Protection Report includes chapters dedicated to each stage of the lifecycle – i.e. childhood, working-age adulthood, and old age – as well as, separately, health care. The chapter on working-age adults and the relevant statistical tables discuss, separately, mothers(i.e. in the last three months of pregnancy and the first months after birth), the disabled, the unemployed, the risk of employment injury and other‘vulnerable’ people. For each of scope, extent and level, the ILO distinguishes between legal and effective co verage. Legal coverage refers to programmes that are‘anchored in national legislation’. Effective coverage refers to programmes that are actually implemented. If programmes are not legislated, then effective coverage might be higher than legal coverage. If implementation lags behind legislation, then effective coverage is likely to be lower than legal coverage. The ILO, in contrast to the World Bank, pays no attention to precisely who gets what, i.e. the detailed distribution of benefits from social prote ction. The ILO’s approach is premised on the objective of universal coverage against a specified set of risks. Historically, this meant coverage of all workers in formal employment. In the 2000s this was finally recast as coverage of all citizens, through the provision of social protection floors. The World Bank generally prefers a targeted approach in practice. It concurs that everyone should be protected against risks but prefers market provision for the non-poor whilst focusing public provision on the poor, through safety nets . One problem with the ILO’s approach, as we shall explore further below, is that it does not distinguish between a non-universal but pro-poor system(i.e. comprising a safety net) and a non-universal but anti-poor system(for example, a system that provides expensive social insurance for formallyemployed workers subsidized through taxes paid by the poor). The ILO also reports on public expenditure on social protection, as a proportion of GDP. Public expenditure on health care is usually separated out from the forms of social protection that provide immediate or future income security. 1.1.3. Individual and family coverage Measuring the extent of coverage also requires choices about who counts as covered under recognized schemes. The ILO generally focuses on individual, direct beneficiaries, and current contributors to schemes who have or are acquiring rights to future benefits. Programmes such as South Africa’s Child Support Grant are considered as social protection for children, although the grant is paid to an adult caregiver, almost always the mother, without any restrictions on how she can spend it. The Child Support Grant might reasonably be considered social protection for a specific category of working-age adults, i.e. adults with children. In practice, recipients seem to spend most of the grant on their children, but spend part of it on general household support. Similarly, there is evidence that old-age pensioners in South Africa spend part of their pension income on supporting grandchildren and other kin. The ILO counts only the intended individual beneficiary. The World Bank, in contrast, generally estimates coverage or reach according to the number

16 of people living in households where at least one member benefits from the relevant kind of programme(see World Bank, 2015: 77-9). Whereas the ILO focuses on individual beneficiaries, the World Bank generally focuses on beneficiary households. A third dataset – the Social Assistance, Politics and Institutions(SAPI) dataset, compiled by Nino-Zarazua and colleagues – jumps between counting individuals beneficiaries and household members. If the programmes is supposedly intended to benefit a specific individual(as is the case with the Child Support Grant in South Africa), then it counts the number of individual beneficiaries. If a programme is intended to benefit the entire household(as is the case with the Tanzania Social Action Fund, for example) then all household members are counted. The definition of social protection and coverage can be very consequential. In Botswana, the only dedicated programme for children is for orphans, on the grounds that nonorphaned children in poor households are supported through the adult breadwinner, who might be employed on a workfare programme. The ILO reports that coverage of children is low, because it counts only the orphans(although it is their caregivers who actually receive the benefit). The ILO also disregards Botswana’s school feeding programme, which reaches most children and many adolescents. Moreover, because support for poor families is primarily through workfare, the ILO reports very low overall coverage: only 15 percent of Botswana’s population is reported as covered by at least one branch of social protection(2017: Table B.3). In contrast, the World Bank’s Atlas of Social Protection Indicators of Resilience and Equity(ASPIRE) database shows that 74 percent of the population of Botswana lived(in 2009) in households that benefitted from one or other social assistance programme. Most of these lived in households where there was at least one child benefiting from the school feeding scheme. Others lived in households that included at least one pensioner or the beneficiary of some other programme. The difference between 15 and 74 percent is due to both the definitions of social protection and the inclusion of indirect as well as direct beneficiaries. The contrast between the ILO and the World Bank reflects their historically-rooted conceptual differences and the limitations of their respective datasets. The ILO relies primarily on administrative data, supplied by national governments. The World Bank relies more extensively on data from household surveys, reported in its ASPIRE database. 13 The World Bank uses survey data more fully than the ILO because it is concerned to examine the distribution of benefits, and especially whether they are reaching the poor. Whereas the ILO pays no attention to precisely who gets what, the World Bank pays close attention to both‘inclusion errors’(when nonpoor people benefit) and‘exclusion errors’ (when poor people fail to benefit). 14 Using administrative data means that the ILO can report the number of children who supposedly benefit from the Child Support Grant in South Africa, but even if it wanted to do so it could not report the number of children supported indirectly through the workfare programme in Botswana. The ILO’s Social Security Inquiry Manual acknowledges that the demarcation of who is a beneficiary is sometimes complicated. Housing benefits may be for households not individuals. Some 13 http://datatopics.worldbank.org/aspire/. 14 The ILO does declare, in an endnote, the ILO‘s 2017 Report declares that it does supplement administrative data with household survey data(209), but it is unclear how far this is the case.

17 pension schemes(especially contributory ones) may provide for dependents, including as survivors after the death of the‘protected person’(ILO, 2016: 10). In practice, however, the ILO’s data seem to refer to individuals and(supposedly) direct beneficiaries, whereas the World Bank for the most part considers households including indirect beneficiaries. The overall result is that the ILO is likely to underestimate the reach of social assistance, especially in countries(including most of Africa) where policy-makers attach considerable importance to families, both ideologically and in the design of public policy. Whether one counts only the individual recipient or all co-resident household members should surely depend on more than the label attached to the programme by policymakers. Ideally, one would take into account the level of the benefit. Is the benefit sufficient to support an entire household, or only a single individual? This would bring in the third dimension of coverage – the level of the benefit – that the ILO in practice disregards entirely. The World Bank does discuss this, in terms of the adequacy of benefits in terms of poverty reduction or elimination(World Bank, 2015: 46-51). Linked to this, one might ask of a programme whether the benefit depends at all on the size of the households(or number of dependents of the individual recipient). For example, the Mozambican Programa Subsidio Social Basico(PSSB)(and its predecessor, the Programa Subsídio de Alimentos or PSA) pays benefits set at one-third of the poverty line(about US$5/month) for the individual recipient plus an additional 25 percent of this for each dependent up to a maximum of four. A recipient with four dependents would receive double the benefit(about$10/month) of a recipient with no dependents. For this programme, it clearly makes sense to count coverage including not only the 143,000 direct beneficiaries but also the 200,000 indirect beneficiaries(as of 2008)(Nino-Zarazua et al., 2012). Conversely, it makes little sense to count other household members as beneficiaries of a school feeding programme: Only the child at school eats the food. Other programmes pose a challenge, however. The Mozambican workfare programme (Programa de Acção Social Produtiva, PASP) pays a benefit equal to the poverty line, i.e. about$15/month, which is 50 percent more than PSSB pays to a recipient with four dependents. This is intended as a family income, although there is no adjustment of the benefit to take into account the number of dependents. Unfortunately, neither the World Bank nor the ILO nor SAPI even attempts to incorporate any such analysis into their estimates of‘coverage’. This remains an avenue for further research. 1.2. The scope of coverage The ILO has long measured the scope of coverage in terms of the ratification of ILO conventions and the number of areas or branches of social security that are provided for by national legislation. Both of these are reported in the ILO’s 2017 Report(in Tables B.1 and B.2 respectively). African countries have been slow to ratify the ILO’s conventions on social security. The most important ILO convention was the Social Security(Minimum Standards) Convention

18 (number 102, 1952). By 2017 this had been ratified, in whole or in part 15 , by seven African countries: Senegal(1962), Niger(1966), Mauritania(1968), DRC(1987), Togo(2013), Chad(2015) and – in Northern Africa – Libya(1975). Several other countries had ratified all or part of Convention 118: CAR(1964), Guinea(1967), Kenya(1971), Cabo Verde (1982), Egypt(1993), Rwanda(1989) and – in Northern Africa – Tunisia(1965). Mali, Benin, Burkina Faso and Sao Tome and Principe had ratified Convention 183(in 2008, 2012, 2013 and 2017 respectively). In total, only twenty African countries had ratified any of the conventions related to social protection. Even among these twenty countries, the average number of branches of social protection that were covered by ratifications was only four(out of eight, excluding medical care and provision for migrants). Table 1.1: Scope of legal coverage of social security, African countries Number Scope of Examples countries Comprehensive 5 Algeria, Morocco, Tunisia, Cabo Verde, South Africa Nearly 6 Egypt, Libya, Guinea, Namibia comprehensive Intermediate 25 Angola, Benin, Botswana, Limited 13 Sudan, Malawi, Nigeria, Zimbabwe Missing data 5 Eritrea, South Sudan Total 54 Source: ILO, 2017: Table B.2. Many countries have statutory programmes even when they have not ratified the relevant ILO Conventions. The ILO counts the number of branches of social security protection covered by statutory programmes in every country. If all eight branches are covered, then the ILO assesses that legal coverage has‘comprehensive scope’. If seven are covered, then the legal coverage is‘nearly comprehensive’. Countries that cover five or six branches are deemed to have‘intermediate scope’, whilst countries with fewer have ‘limited scope’. Table 1.1 shows the number of African countries in each category. It is not clear what this tells us. The relationships between ratifications, legal coverage by branch and effective coverage or expenditure might not be strong. More than twenty years ago, Strang and Chang(1993) examined the consequences of ratification, identifying two possible mechanisms: Conventions‘offer a script or model that actors can draw upon to design policy’(e specially when conventions define benefit categories, minimum levels and periods of coverage, and conditions for eligibility), and they‘might provide reasons for action’ through shifting policies symbolically‘out of the realm of zero -sum, partisan politi cs and into the realm of fundamental, universally recognized rights’( ibid: 242-243). 15 Countries could choose to ratify only some parts of the Convention. Of the nine substantive Parts, each coverning a different risk or branch of social security, overall ratification required ratification of a minimum of three parts: http://www.ilo.org/dyn/normlex/en/f?p=NORMLEXPUB:12100:0::NO::P12100_ILO_CODE:C102.

19 Strang and Chang found that ILO ratifications seemed to stimulate increased welfare spending in the advanced capitalist societies but not in less developed countries(LDCs): ILO standard setting does not seem to contribute similarly to the growth of welfare programs in the developing world. … LDCs tend not to ratify ILO standards on social security, and spending does not markedly increase when they do ratify them. In such cases, ratification may be best understood as a symbolic commitment to programs whose realization faces severe fiscal and organizational constraints.( ibid: 257) In the advanced capitalist countries, for which there are better data, right-wing governments were more likely to respond to ratifications than left-wing ones, presumably in large part because left-wing governments had often expanded welfare spending prior to ratification. Table 1.2: Relationships between the scope of coverage and expenditure Model A B C D E F Effective coverage All All 1 All only only 1 Africa Africa Africa only Conventions 1.4 2.4 (0.8)(1.6) ratified **** 2.0 (1.4) 1.5 (0.9)* Scope of legal coverage 15.5 (1.7) *** 3.8 (1.4)** constant 59 47-49 13(3)-2.9 (8)(5)(12) 10(9) ***(7.4) ********* Adjusted r0.04 0.17 0.55 0.16 0.11 0.32 squared n 45 71 71 6 18 18 p<0.1** p<0.05*** p<0.01* 1 Coding missing data for conventions as 0. coverage G H All All 1 0.6 (0.2) *** 0.7 (0.1) *** 8.7 (1.6) *** 0.14 82 7.3 (0.6) *** 0.26 166 and the extent of I J K Expenditure All only only 1 Africa Africa 0.12 (0.12) 0.06 (0.1) 3.0 (0.3) *** -9.5 (2.0) *** 4.1 (0.9) *** 5(0.5) *** 0.38 0 0 164 20 50 L Africa only 0.5 (0.3) 2.4 (1.6) 0.04 48 Table 1.2 reports the results of some simple regression models, regressing(1) the extent of‘effective coverage’(from ILO, 2017: B.3, discussed further in the next section below) and(2) expenditure on social protection(including health care) as a share of GDP( ibid: B.16 or B.17, for(a) all countries and(b) African countries only, on each of the number of conventions ratified( ibid: B.1) and the scope of legal coverage by branch( ibid: Table B.2, scored from 0 or 1 to 8). Results are reported for two models using data on ratifications: The second assumes that countries that are not listed in the 2017 Report had never ratified any of the social protection conventions. It is important to note that the number of African countries included in some of the models is small, because the ILO reports data on either coverage or ratification for only a minority of countries. Table 1.2 shows that, globally, there is a strong relationship between the number of conventions ratified and either the extent of effective coverage(see models A and B) or expenditure(see models G and H). In Africa, however, the relationship is weak and of little or no statistical significance(see models D, E, J and K). Globally, both the extent of

20 effective coverage and total expenditure are positively correlated with the scope of legal coverage(see models C and I), but in Africa the first is weak(relative to across the world as a whole, see model F) and the second is not statistically significant(see model L). Across the world as a whole, whether or not a country ratifies ILO conventions or has legal programmes in different branches of social security is correlated with both the extent to which citizens are covered by social protection or public expenditure on social protection. In Africa, it seems to make little or no difference whether a country has ratified conventions or has legislated social security provision across the designated risks. 1.3. The extent of coverage The ILO reports coverage for the world as a whole, for regions of the world, and for individual countries. The headline finding of the 2017 Report was that 71 percent of the world’s population have no or only partial access to comprehensive social protection(ILO, 2017: vi). More precisely, 29 percent of the global population are fully covered, 16 percent are partially covered, and 55 percent are not covered at all. Coverage rates are highest among the elderly(at 68 percent of the global population of older men and women)( ibid: ix-x). The previous(2014) World Social Protection Report put the proportion of the world’s population with no or incomplete coverage at 73 percent(ILO, 2014: xix), suggesting that coverage was rising slowly. Reported global coverage of the elderly rose most between the 2014 and 2017 Reports. The 2014 Report reported that almost half(48 percent) of the global population over pensionable age did not receive any pension(2014: xxi); the 2017 Report indicated that this had dropped to only one-third(32 percent). The ILO also reports coverage for different regions of the world. Coverage rates are lowest in Africa. The ILO reports(absurdly precisely) that 17.8 percent of the total African population are covered in the sense of receiving(or contributing to)‘at least one social protection cash benefit’(2017: xxxiii). Whilst 29.6 percent of the elderly are reportedly covered, only 15.9 percent of children and mothers with newborn babies are reportedly covered. Coverage of unemployed people is very low( ibid: x). The ILO and World Bank concur that public expenditure on social protection is also very low across most of Africa in relation to GDP(although they differ in their precise estimates, given their divergent definition of what is covered). In most countries, the proportion is less than 5 percent. Whilst there has been‘significant progress in the extension of social protection coverage’, the ILO avers, the‘development of social protection floors’ remains ‘an urgent priority in Africa’(2017: xxxiii). The ILO points to the need to extend coverage ‘to those in the informal economy’(and‘facilitating their transition to the formal economy’) through‘a mix of contributory and non contributory schemes’( ibid: xxxiv). 1.3.1. Data Challenges The precision with which the ILO reports its findings – such as‘17.8’ percent of the population of Africa are covered by at least one cash social protection benefit – is spurious. Not only are there enormous definitional challenges, as discussed above, but

21 the data on Africa are incomplete. Data are missing for some or all social protection ‘areas’ for some or many countries. In its 2017 Report, the ILO imputed values for missing data(2017: 206-8). Few countries – containing only 5 percent of the global population – failed to report on old-age pensions. Missing data for old-age pensions were therefore imputed through simply applying the regional average. Data were missing for larger proportions of the global population with respect to other‘areas’ of soci al protection. Data were missing on the coverage of children, for example, for countries with a combined total of 59 percent of the world’s children. For these other‘areas’ of social protection, data were imputed using a regression model that predicted values given the region and GDP per capita. The regional and global estimates were calculated using weighted averages of the reported or imputed national-level data. Missing data were a major problem in Africa(although not as much as in the Arab states). Data were missing on old-age pensions for countries with 26 percent of the total African population. Data were missing on unemployment for countries with 50 percent of the total African population. And data were missing on social protection for children for countries with 55 percent of the total African population. Imputing such a high proportion of the data is, at best, high risk. The ILO’s chosen methods for imputing data seem especially suspect. The basic intuition behind the imputation seems to be that a random selection of countries failed to report, so that values for these countries can be imputed from the regional average or taking into account the relationship between GDP and coverage for countries that had reported. It is likely, however, that countries with big programmes and/or that prioritized social protection were more likely to report, whilst non-reporting countries were more likely to have small or no programmes and/or have not prioritised social protection. Moreover, predictions using regressions on GDP per capita are unlikely to be reliable, given that social assistance programmes are relatively important in Africa (as the World Bank acknowledges: World Bank, 2017: 2, Figure 4) and there is a weak relationship between GDP per capita and social assistance expenditure(as originally noted by Weigand and Grosh, 2008; see further Seekings, 2016d). In addition, the reported data for some African countries are very old. In the case of data on contributory pensions, for example, the 2017 Report published data for Djibouti and Mauritania for 2002. The mean year for the data on contributory pensions in Africa was 2010. The World Bank also reports some aggregate data on the coverage or reach of social safety nets. Given that it also has incomplete data, it must either impute missing data or assume that there are no programmes in the countries that have not reported any programmes. It is not clear how the World Bank tackles this problem. 1.3.2. Coverage of the elderly The most detailed data in the ILO ’s 2017 Report concern provision for old age. Four tables in an Annex present data on the current and prospective coverage of the elderly through contributory and non-contributory programmes. Table B.9 shows coverage by country, expressed as percentage of the working age population. This appears to indicate the

22 proportion of working-age adults who are contributing to a pension(or provident) fund, either compulsorily or voluntarily, or who are covered by a guarantee of pension through a universal or near-universal, non-contributory programme. Table B.11 reports the number of active contributors to pension schemes, as proportions of both the workingage population and the labour force. Table B10 reports data on the current coverage of the elderly under each non-contributory pension programme(and the cost of these programmes). Table B.12 reports total coverage of the elderly, in some cases disaggregating between contributory and non-contributory sub-totals. Much of the data in Table B.9 – supposedly reporting prospective future coverage – are simply not credible, and are not consistent with the data shown in the following tables. Some of the data seem to be good. Botswana, which has a universal non-contributory programme, is correctly reported to provide future coverage for 100 percent of workingage adults. South Africa is also correctly reported as providing for 100 percent of workingage adults, because its means-tested programme guarantees a pension to anyone who does not have substantial, alternative income. 16 The data for a number of other countries are, however, nonsensical. For example, Kenya is reported to have 100 percent coverage, two-thirds through a mandatory contributory system, and one-third through a non-contributory system. The suggestion that one-third of the population is prospectively covered by the non-contributory scheme is broadly credible. 17 But it is certainly not the case that two out of three adults contribute to or can expect benefits from the mandatory, contributory system. Other sources reported that only 15 percent of Kenyans were covered by pensions. 18 16 Table B.9 incorrectly reports that contributory schemes in South Africa have 0 percent coverage; the error is clear from Table B.11, which records that 4 percent of the working-age population contributes to a pension fund; even this figure is an underestimate, given the statutory foundations of many private funds. 17 The Kenyan‘pilot’ non -contributory old-age pension programme was reported officially to have reached just over 200,000 elderly people aged 65 or older in 2015-16, out of a total elderly population of over 1 million(see http://www.socialprotection.or.ke/social-protection-components/social-assistance/national-safety-net-program/olderpersons-cash-transfer-opct). The ILO reported a higher figure, of just over 300,000 pensioners, The government announced in 2017 that it would introduce a universal old-age pension in 2018, although only for people aged 70 or older(http://www.helpage.org/newsroom/latest-news/kenya-to-launch-universal-pension-scheme-in-january-2018/). Recent reports suggest that more than half a million people had applied for the new pensions, to be introduced in March 2018, but at least half of these applicants were receiving pensions under the existing scheme (https://www.standardmedia.co.ke/business/article/2001267078/senior-citizens-to-get-sh4-000-from-march). 18 http://blogs.worldbank.org/psd/sustainablecities/3-hindrances-expanding-pensions-kenya.

23 Similarly, the suggestions that Mozambique 19 , Uganda 20 and Tanzania 21 all have 100 percent coverage are not credible. The reported coverage rates for Malawi 22 (28 percent), Nigeria(34 percent), Senegal(24 percent) and Zambia(48 percent) are also implausible high. Most bizarrely, much of the data in Table B.9 do not seem to be consistent with data in subsequent tables. In each of Tanzania, Uganda and Malawi, according to Table B.11, less than 4 percent of the working-age population contributes to a pension scheme. In Nigeria the proportion is less than 8 percent and in Senegal less than 2 percent. In Kenya the proportion is higher, at 11 percent, but it is impossible to reconcile even this with the 100 percent coverage reported in Table B.9. Moreover, the detailed data on noncontributory pension programmes in Table B.10 do not explain the discrepancies. The ILO reports that Kenya’s Older Persons Cash Transfer Pilot Programme supposedly reached 310,000 people aged 65 or older, corresponding to 24 percent of the elderly population, not one half. 23 Mozambique’s Programa de Subsidio Social Basico(PSSB) reportedly reached more than 341,000 elderly people, i.e. 36 percent of the elderly population of Mozambique, also well short of one half. Table B.12 reports that Mozambique has only 17 percent coverage, mostly non-contributory – i.e. less than half the rate reported in the earlier Table B.10. Uganda is reported to have total coverage of less than 7 percent, two-thirds through contributory and one-third through noncontributory programmes. In short, I am unable to make sense of the ILO’s data on the extent of old-age pensions in many African countries. Overall, the ILO calculates, 30 percent of Afr ica’s elderly population is covered by social protection. It is hard to see how this figure can be reconciled with the more credible country-level data. 1.3.3. Changes over time in coverage of the elderly The estimates in the 2017 Report for coverage or reach across Africa as a whole are also markedly higher than the estimates included in the ILO’s previous 2014 Report. The 2014 Report estimated that 17 percent of the elderly in Sub-Saharan Africa received pensions, whilst 37 percent did so in North Africa(2014: 83). Across Africa as a whole, the figure was 21.5 percent( ibid: 273). 24 The ILO seems to be suggesting that the proportion of 19 The ILO reports that half the Mozambique population is covered through the mandatory contributory system and half through a non-contributory system. HelpAge International report that, in about 2015, less than a quarter of the population aged 60 or more received the non-contributory old-age pension(http://www.pensionwatch.net/pensions/country-fact-file/mozambique). 20 Uganda is reported to have 100 percent coverage through a non-contributory system. In reality, Uganda has a meanstested system in selected districts, with no clear prospect of a national roll-out. 21 Tanzania is said to have 100 percent coverage, with over half of the working-age population covered through contributory programmes and everyone else covered by a non-contributory programme. Yet, in reality, only Zanzibar has a universal pension(and then only from the age of 70); there is no non-contributory provision on the mainland(as is recognized in Table B.10) – although more than 500,000 elderly people live in households that in 2016 benefitted from a programme targeted on poor households(TASAF, 2017: 26). Contributory programmes in Tanzania reach very few elderly people. HelpAge International reports that o nly 4 percent of the country’s elderly population receives any pension(http://www.pension-watch.net/pensions/country-fact-file/tanzania-united-republic-of). 22 The ILO itself notes, in an endnote, that the 2011 pension law‘is yet to be implemented’(2017: 343). 23 No data are provided for coverage of the elderly under the Harmonised Social Cash Transfer Programme, in the north of Kenya. 24 This estimate was m uch the same as the World Bank’s earlier estimate that‘less than 20%’ of the elderly in Africa received pensions(Forteza, Lucchetti and Pallares-Miralles, 2009: 35).

24 elderly receiving social pensions rose from just over 20 percent to about 30 percent over a few years. What is the evidence for this? Both the 2014 and 2017 Reports provide country-level data for non-contributory pension programmes, total coverage or reach of contributory and non-contributory programmes, and expenditures. Comparison of the countrylevel data from the ILO’s two reports does not support the finding that coverage or reach has risen significantly over the period covered by the reports. This supposed rise in coverage does not seem to be based on country-level data on either contributory or non-contributory programmes. The data on non-contributory programmes covers a total of sixteen African countries, including two in Northern Africa. For the nine countries for which data are provided for at least two years, growth in the total number of beneficiaries runs at about 5 percent p.a.. Given that the elderly population of Africa is rising by about 3 percent p.a., 25 the reach of social pensions has been rising very slowly. Even taking into account the two countries with new programmes, the reach of noncontributory programmes has risen by very little. 26 Nor can the supposed increase in coverage be explained by country-level data on contributory programmes. The 2014 and 2017 Reports provide coverage data for contributory programmes for 40 African countries, including five i n‘Northern’ Africa. 27 For the 11 countries for which data are provided for two years, there was little or no change in 4 countries. 28 coverage collapsed dramatically in Tunisia(from almost 70 percent in 2006 to 25 percent in 2015) and declined in Malawi(from 4 percent to 2 percent, 2010-16) and Mali(from 6 percent to 3 percent, 2010-15). Coverage rose in only four countries: in Ethiopia(from 9 percent to 15 percent, 2006-15), the Gambia(from 11 percent to 17 percent, 2006-15), Ghana(from 8 percent to 33 percent, 2011-15) and Sao Tome and Principe(from 42 percent to 52 percent, 2010-15). Taken at face value, the reported data on the coverage of contributory programmes do not suggest that overall coverage rose any faster than a snail’s pace. Given the absence of any country-level evidence that coverage rates of contributory programmes or the reach of noncontributory programmes rose significantly, the ILO’s assertion that total coverage rose from just over 20 percent to 30 percent must be based on their imputation of missing data. The 2014 Report does not discuss imputation, and it is not entirely clear how the regional estimates were derived. In its 2017 Report, the ILO 25 Calculated using WDI data for population aged 65 or older. 26 Comparison of the data in the 2014 and 2017 Reports shows the following:  For six countries, the data are for one year only although the programmes have existed for some time: Mozambique, Seychelles, Swaziland, Zambia, Egypt and Cabo Verde. For one of these(Cabo Verde), there appears to be an error, in that different data are provided for the same year.  Seven countries had existing programmes and little or no change in even the absolute number of pensioners: Algeria, Botswana, Lesotho, Mauritius, RSA, Namibia, Nigeria.  Two countries had existing programmes and rising reach: Kenya, Uganda.  Two country had new programmes, but in both cases the new programmes were sub-national in extent (Zanzibar in Tanzania, Osun state in Nigeria) and the number of new pensioners was very small. 27 Comparison of the data from 2014(Table B.9) and 2017(Table B.12) shows the following:  For 24 countries, the two reports publish the same data, for one year only.  For 3 countries, only the 2014 Report publishes data: Kenya, South Africa and Zambia.  For 2 countries, only the 2017 Report publishes data: Morocco and Nigeria.  Data for two periods in time are available for 11 countries. 28 Burkina Faso, Burundi, Cameroon and Niger.

25 explains that it imputed missing data using the regional average. I assume that this means that the average coverage rate for countries that reported data was applied to countries that did not report data. Given that countries with broad programmes – and especially those with universal programmes – appear to have been much more likely to report data, the regional average is likely to be greatly inflated by this imputation. I have no knowledge of the veracity of most of the countrylevel data. Taking the ILO’s reported contributory pension coverage rates and most recent non-contributory pension data at face value, and adding in missing data for contributory pensions in countries like South Africa and Kenya, I estimate very crudely that there are less than 12 million pensioners in Africa(of whom less than 5 million benefit from contributory programmes and less than 7 million from non-contributory programmes). This is equivalent to about 40 percent of the population aged 65 or older, and about 25 percent of the population aged 60 or older. It is impossible to know how many pensioners are younger than 60, but the figure is unlikely to be high. It is unlikely that coverage rose significantly in the 2010s. The estimate in the ILO’s 2014 Report was probably too low, and the estimate in the 2017 Report too high. 1.3.4. Coverage of Children The ILO’s 2017 Report does not provide as detailed data on child and family benefits as it does on pensions for the elderly, perhaps because there is no organization collating programme-specific data in the way that HelpAge International does for social pensions. The ILO’s data are also very incomplete with regard to the extent of coverage, and some of the data seem implausible. The ILO estimates that social protection covers 16 percent of children across Africa, at a cost of about 0.6 percent of GDP. In the relevant table(B.4) in Annex IV to the Report, the ILO provides estimates for coverage for only 14 African countries. These include Botswana, where effective coverage is put at 5.5 percent, Cabo Verde(31.5 percent) and South Africa(75 percent). It also includes Cameroon(where coverage is put at 41.5 percent) and the DRC(16 percent), which are both described as having contributory systems. 29 Zambia, which is correctly reported as having no statutory programme, has a reported coverage rate of 21 percent. The ILO’s 2014 and 2017 Reports suggest that protection of children has risen in Africa. The earlier report neither included any comparable statistical annex on the coverage of child and family benefits nor provided any summary data on coverage in the text. It did, however, report on expenditure on child and family benefits. Across Africa as a whole, only 0.2 percent of GDP was spent on child and family benefits. The proportion was higher in Northern than in Sub-Saharan Africa. Most of the data were from 2009, 2010 or 2011 (2014: Table B.13). The only country to spend more than 1 percent of GDP was South Africa. Botswana spent almost 0.6 percent of GDP, and Algeria almost 0.4 percent. Comparing the 2014 and 2017 Reports, only one country stands out: Madagascar. No data for Madagascar were reported in the 2014 Report. According to the 2017 Report, 29 The World Bank’s ASPIRE database shows very low co verage for all programmes in Cameroon(and lacks data on the DRC).

26 however, Madagascar spent 2.8 percent of GDP on child and family benefits in 2015 (2017: 406). 30 With the possible exception of Madagascar, the 2017 Report does not seem to provide any direct statistical evidence of any significant expansion of coverage of children in Africa. The Report does, however, discuss in the text two new programmes: Senegal’s PNBSF, which was piloted in 2013 and then extended countrywide, reaching(according to the Report) 3 million people(2017:124); and Namibia’s Vulnerable Child Grant, introduced in 2015, soon reaching about 130,000 children( ibid: 125). The PNBSF is a family grant, paid to 300,000 households, i.e. approximately one in five Senegalese households. Without detailed data it is impossible to gauge precisely the coverage of children, but it is unlikely that it covers much more than one in five children. The Namibian programme reaches a smaller proportion of Namibian children.(Namibia already had a programme for orphans and vulnerable children, so the combined coverage is much the same as in Senegal). Whilst these reforms only reach a minority of children, Senegal and Namibia now have the fourth and fifth highest coverage rates in Africa, according to ILO definitions. Neither reform is reflected in the statistical data published in the Report. Indeed, a graph above the text box on the PNBSF reports the effective coverage of children in Senegal at 4 percent! Africa is a young continent, in the sense of having a very high proportion of young people. More than 40 percent of the population of Sub-Saharan Africa is aged 14 or younger; almost 50 percent are aged under 18. Africa has a total child population of about 500 million(aged 0-14) or close to 600 million children(aged 0-17). If the ILO is correct in estimating that 16 percent of the continent’s children are covered by one or other social protection programme, then this means that about 90 million children(aged 0-14) or almost 110 million children(aged 0-17) are covered. It is difficult to understand how the ILO arrives at the figure of 16 percent. By far the largest programmes are in South Africa, where the Child Support Grant and two smaller programmes are paid to the caregivers of a combined total of almost 13 million children(as of the end of 2017). The World Bank has a reasonably good list of non-contributory programmes(although it misses several programmes in Botswana 31 and South Africa 32 ). In Africa as a whole, it is difficult to see how more than 25 million children benefit from cash non-contributory child or family benefits, as defined by the ILO. This corresponds to only about 5 percent of the continent’s children. For the ILO to arrive at a figure of 16 percent coverage in total, it has probably taken at face value implausibly high coverage rates for some contributory programmes(notably in Cameroon, the DRC and Madagascar), and has imputed missing data for other countries on the basis of the high(and probably inflated) data on countries in its database. The ILO might also have included some programmes that provide benefits in kind. Many more children benefit from school(and other) feeding programmes, but these are excluded from the ILO data. The World Bank(2015: 104-16) reports coverage of school 30 The ASPIRE database lacks data on Madagascar. 31 The Orphan Care Package(OCP). 32 The Foster Care Grant(FCG) and Care Dependency Grant(CDG).

27 feeding programmes by country. Across Africa as a whole, the total number of children fed at school comes to almost 40 million. This total includes more than 9 million in South Africa, 7 million in Egypt, more than 2 million in each of Burkino Faso and Sudan, almost 2 million in each of Malawi and Ghana, more than 1 million in each of Morocco and Tanzania, and close to 1 million in each of the DRC, Kenya, Senegal and Zambia. School feeding programmes have high coverage rates in countries such as Botswana and South Africa, but in most other countries they are targeted on the very poor. If the total number of children being fed at school is about 40 million, this represents less than 10 percent of the continent’s children. Much smaller numbers of pre -school children are also fed. The World Bank lists in kind non-contributory programmes, excluding school feeding programmes, with a total coverage of 20-25 million children. This includes 5 million children in the Sudan, 3 million in Senegal, 2.5 million in Ethiopia and close to 2 million in Malawi. In total, in kind programmes might cover about 15 percent of African children(at least to the age of 14). Some children also benefit indirectly from programmes that provide relief to families through adults. Children are covered as dependents. As discussed above, this is true to some extent of programmes counted by the ILO, including child grant programmes in South Africa, Senegal, Cabo Verde and Botswana, all of which pay benefits to one or other adult on behalf of children. The ILO does not count, however, workfare programmes, even when these are intended to support entire destitute families. In Botswana, for example, orphans are supported through a dedicated programme, but the state expects that non-orphaned children will be supported by their parents, if necessary through participation in the Ipelegeng workfare programme. Workfare is considered more fully in the next section. 1.3.5. Coverage of working-age adults The ILO reports on coverage of mothers, the unemployed, mothers, employment injury and disability. Reported coverage rates are low in all four categories. Only one country in Africa(Ghana) is reported to provide cash benefits to more than one in five new mothers (2017: 125).(This figure ignores the payment of the Child Support Grant to most new mothers in South Africa, and feeding programmes for new mothers in countries such as Botswana). Most countries only have contributory systems, limiting benefits to women who previously worked in formal employment. Worker’s compensation programmes are limited(2017: 126). Only two countries(South Africa and Cabo Verde) provided cash benefits to more than one in five people with severe disabilities(2017: 127). The ILO reports in the text that no African country pays cash benefits to more than one in five unemployed citizens. The tiny Seychelles supports 18 percent of its unemployed, and South Africa 10.6 percent of its unemployed(2017: 126). In the statistical annex, Botswana is also reported to support 35 percent of its unemployed(in 2015), through a severance payment programme(Table B.6). I cannot find any corroboration of this claim.

28 These data seem to exclude public employment programmes(as discussed above). 33 The World Bank(2015: 104-16) reports coverage of public employment programmes by country. Across Africa as a whole, the total number of people benefitting from workfare programmes comes to almost 16 million(excluding countries whose data are missing); this is only one-third the number in India. The largest programmes are in Ethiopia(the PSNP, with almost 7 million beneficiaries) and Morocco(with 4 million). Each of Kenya, South Sudan, Somalia and Sierra Leone have almost 1 million beneficiaries. The WFP has been experimenting with workfare programmes in place of food relief in a number of countries. 1.4. Coverage gaps Just as understandings of coverage differ, so the‘coverage gap’ means different things to different organisations. The ILO has played a leading role in work on coverage gaps. In 2002, the ILO embarked on a‘Global Campaign on Social Security and Coverage For All ’, which in turn grew into the ILO led‘social protection floors’ initiative(Deacon, 2013). The ILO understood coverage and hence coverage gaps primarily in terms of the existence or lack of statutory insurance against risk, but risk had come to be understood rather imprecisely as including vulnerability to poverty. In the ILO’s understanding, everyone should be protected against the risk of poverty, although only those people unfortunate enough to fall into poverty(whether defined in relation to a general poverty line or an individual’s prior standard of living) would actually need to receive benefits. The coverage gap could therefore be understood in terms of either the proportion of the population that is not protected against the risks that result in poverty, or the proportion of the currently poor population who is not receiving benefits. The ILO’s World Social Protection Reports are unclear as to when each of these understandings is relevant. In practice, the Reports imply that coverage gaps exist insofar as:  Children are not provided for through family or child cash benefit schemes, whether contributory or non-contributory;  Elderly people are not provided for through cash pensions, whether contributory or non-contributory;  Working-age adults are not insured against unemployment, maternity, injury, disability, sickness, or old age;  Working-age adults who are unemployed, injured, sick, disabled or are about to or have recently become mothers, receive cash benefits. The ILO Reports to not pay attention to able-bodied, working, working-age adults who are not mothers, who are living in poverty. Nor does the ILO distinguish between coverage of poor and non-poor children, poor and non-poor older people, and so on. The ILO envisages(and proposes) achieving universal coverage through a combination of a social protection floor, comprising primarily universal or(perhaps) means-tested social assistance, with more generous benefits for people who had insured themselves against specified risks through contributory programmes. If the social assistance 33 Table B.6 indicates that India spends 0 percent on non-contributory unemployment assistance, because the data do not cover MGNREGS(see endnote 8, p283).

29 programmes were not universal, they should at least provide minimal benefits for anyone who could not access contributory programmes. Progress towards universal coverage would usually entail a mix of expanding contributory and non-contributory programmes. In the ILO’s terms, there remain massive coverage gaps among the elderly, children and working-age adults across most of Africa. The total population of Africa was about 1.25 billion in mid-2015, including just over 1 billion in Sub-Saharan Africa and 230,000 in Northern Africa. In terms of age-groups, this comprised approximately 500 million children aged 0-14 and about 600 million under the age of 18. The elderly population was small: Only 3 percent(or less than 40 million) were aged 65 or older, with another 2 percent(or about 25 million people) aged 60-64. The remaining 600 million people were working-age adults. 34 The ILO reports(in its 2017 Report) coverage rates of 16 percent among children and 30 percent among the elderly. These rates correspond to about 90-110 million child beneficiaries and 10-20 million elderly beneficiaries(depending on the age limits to these categories). The ILO reports also that overall coverage across Africa is 18 percent, but the ILO seems to mean by this that 18 percent of the population(i.e. about 220 million people in total) are either benefiting from a social protection programme(whether contributory or non-contributory) or are actively contributing to at least one programme or scheme(2017: Table B.3, endnote 1). Deducting the child and elderly beneficiaries(100130 million) suggests that 90-120 million working-age adults either received benefits or contributed(with the number depending on the age limits to childhood and old age). Given that the number of workingage adults receiving benefits(by the ILO’s definition) is very small, most of the 90-120 million workingage adults deemed to be‘covered’ must be covered in the sense of being contributors. The ILO estimates for Africa-wide coverage seem to suffer from numerous flaws. First, combining contributors and beneficiaries is probably more misleading than useful. ‘Coverage’ rates would be higher in a country with a primarily contributory but very incomplete pension system than in one with a universal social pension, as long as the actual number of pensioners was small. Secondly, as we have seen, it is hard to square even the ILO’s estimates for Africa -wide coverage of children and the elderly with the country-level data published by the ILO itself or from other sources. It seems likely that the ILO’s estimate for the elderly is too high, although by only a few percentage points. The ILO’s estimate for the coverage of children seems much too high, especially given the ILO ’s focus on cash transfer programmes. If we exclude school and other permanent feeding schemes, as the ILO does, then the share of children covered by child and family benefits seems closer to 5 percent(rather than 16 percent). Feeding schemes reach about 15 percent of children, and perhaps 8 percent benefit indirectly as the dependents of adults employed on workfare programmes. There will, however, be considerable overlap between these programmes: Children in poor households may be the dependents of adults on workfare and the direct beneficiaries of school feeding programmes and even the intended beneficiaries of a cash transfer programme such as South Africa’s Child Support Grant. Overall coverage might be only 5 percent using the ILO’s definition of social protection, but might be as high as 25 percent using a more inclusive definition 34 Data from www.prb.org. See further UNECA(2016).

30 (and depending on duplication). During drought- or war-induced famine, the number of people benefitting from feeding programmes rises sharply. 1.4.1. World Bank data on coverage gaps The World Bank understands coverage gaps rather differently. Its focus is on the poor, and the absence of coverage for the poor specifically. Unfortunately, the ASPIRE database covers social protection data for fewer African countries and for fewer years than the ILO’s aggregate data. 35 Table A1 in the Appendix shows the most recent year of data in the ASPIRE database for African countries. There are no data for a small number of countries. 36 ASPIRE includes more than one round of data for sixteen countries. There is only one round of data for 23 countries. The ASPIRE data has two obvious limitations in addition to infrequent and incomplete country-level coverage, First, it relies on respondent recall, with(usually) one respondent providing information on the entire household. Secondly, it is difficult to match the programmes that exist in any particular country to the general programme categories used for presentation of data. The use and limits of the data are evident in the cases of Tanzania and South Africa. The World Bank’s ASPIRE database provides partial data on Tanzania from the three waves of the Tanzania National Panel Survey, conducted in 2008-09, 2010-11 and 201213. 37 The survey seems to have asked about: pensions(i.e. private and government combined); school and other feeding programmes; and workfare(both food- and cashfor-work). 38 Tanzania is an acknowledged laggard in terms of social insurance as well as social assistance(Ulriksen, 2016), although the small, semi-autonomous territory of Zanzibar introduced old-age pensions in 2016(Seekings, 2016). The only data in ASPIRE on social insurance concerns old-age pensions. Only 1 percent of people live in households where someone benefits from a contributory pension. The coverage rate is higher in urban areas (3 percent) than rural areas(<1 percent), but it is low everywhere. These estimates of coverage are not consistent with(but are more credible than) the data reported by the ILO(discussed above), i.e. that contributory pension programmes covered half of the working-age population. ASPIRE does not provide any data on any other social insurance programmes. With respect to social assistance, there are no data on social pensions or any other cash transfers. There are, however, data on school feeding programmes: the 35 Data – generally partial rather than complete – are available for: Benin(2003), Botswana(2009), Burkino Faso (2003, 2009, 2014), Cabo Verde(2007), Cameroon(2007, 2014), CAR(2008), Chad(2011), DRC(2004, 2012), Congo(2005), Cote d’Ivoire(2003), Djib outi(2003, 2012), Egypt(2008), Ethiopia(2005, 2012), Gabon(2006), Gambia(1999), Ghana(2006, 2013), Kenya(2006), Lesotho(2011, 2015), Liberia(2008), Madagascar(2011), Malawi(2005, 2011, 2014), Mali(2010), Mauritania(2009, 2015), Mauritius(2007, 2013), Morocco(2010), Mozambique(2009), Namibia(2004, 2010), Niger(2012), Nigeria(2011, 2013), Rwanda(2006, 2011), Senegal (2012), Sierra Leone(2012), South Africa(2006, 2011), South Sudan(2010), Sudan(2010), Swaziland(2001, 2010), Tanzania(2009, 2011, 2013), Togo(2006, 2012), Uganda(2006, 2010, 2013), Zambia(2011) and Zimbabwe(2008). See: http://databank.worldbank.org/data/download/aspire/ASPIRE_list_of_datasets_per.html. 36 Including Algeria, Angola, Burundi, Eritrea, Guinea, Guinea-Bissau, Libya and Tunisia. 37 The data are listed in the database under 2008, 2010 and 2012 respectively. 38 See ASPIRE metadata for East and Southern Africa.

31 proportion of people living in households that benefited rose from 5 percent to 7 percent between 2008 and 2012. 39 A larger proportion(rising to 12 percent of the population) benefited from one or other transfers in kind(including school feeding programmes). About 1 percent benefited from public works programmes. In total, 20 percent of the population benefited from one or other social assistance programme. The distribution of beneficiary households by quintile is shown in Table 1.3. The beneficiaries of contributory pensions are, for the most part, limited to the top two income quintiles. The coverage of social assistance is much more even across income quintiles. School feeding programmes benefit non-poor as well as poor households, although other feeding schemes are targeted more precisely on the poor. The benefits of workfare are concentrated primarily in the third and fourth income quintiles, presumably in part because the poorest quintile comprises disproportionately households without adults available to work. Table 1.3: Social protection coverage by quintile, Tanzania, 2012 Poorest quintile (%) Quintile 2 (%) Quintile 3 (%) Quintile 4 (%) Richest quintile (%) Total (%) Contributory pensions<1<1<1 2 4 1 School feeding 7 6 7 7 9 7 All in kind 14 12 14 11 9 12 Workfare<1<1 2 2<1 1 Total social assistance 20 18 20 20 18 20 Source: ASPIRE, variables per_sa_allsa.cov_q*_tot; per_si_allsi.cov_q*_tot; per_sa_ik.cov_q*_tot; per_sa_pw.cov_q*_tot; per_sa_sf.cov_q*_tot where* is the quintile number(i.e. 1 through 5). Tanzania clearly had(in 2012) and still has huge coverage gaps, whether viewed through the ILO’s or World Bank’s metrics. The coverage or reach of social protection is very far from universal. In 2012, some poor people benefitted from feeding programmes, but most had no access to any social protection. Contributory pensions had very limited coverage, and most of the beneficiaries were nonpoor. From 2014 the World Bank’s safety nets expanded rapidly in Tanzania(although there are no data to show this in the ASPIRE database). Between the beginning of 2014 and the end of 2015, cash transfers under the Productive Safety Net programme expanded from 40,000 households to more than 1.1 million households(about 10 percent of the total, with a total of just over 5 million members). These numbers remained stable through 2016, costing at total of about US$15 million for 2016 as a whole(TASAF, 2017). Assuming that there are few inclusion errors, 39 In its 2015 State of Social Safety Nets Report, the World Bank reports rather different data for coverage for 2008. As many as 77 percent of all households receive benefits from one or other safety net, almost all from school feeding scheme, at a cost of about 0.3 percent of GDP. Because only a very small proportion(4 percent) of the benefits accrue to the poorest income quintile, the effect on poverty is small(2015: Appendices D, F and G). The figures on programmatic coverage do not match these. Over 1 million children were reported as benefiting from the school feeding scheme(‘food for education’), but this is a very small proportion of the number of children in Tanzania.

32 future survey data will presumably show that the coverage gap in the poorest quintiles has been reduced but not disappeared. Tanzania is a laggard in terms of social protection, with large and enduring coverage gaps. South Africa is, in contrast, a pioneer. Table 1.4 shows coverage by income quintile in South Africa, for social pensions and all social assistance, using ASPIRE data from 2010. It is unclear precisely what is included in and what is excluded from the ASPIRE category of‘total social assistance’. The overall picture is, however, clear: Almost all poor households i nclude one or more members who enjoys some‘protection’. A large minority of poor households include at least one pensioner. Coverage is very much more progressive than in Tanzania. The distribution of benefits between quintiles is less propoor(see Table 1.5). The poorest income quintile benefits less from the pension than less poor quintiles(due to the fact that a pension will usually lift the entire household out of extreme poverty; two pensions will certainly achieve this). In South Africa, the coverage or reach of social protection is good at the household level for all except for the top (richest) income quintile and benefits are not concentrated on the rich. What ASPIRE data cannot show is whether there are persistent coverage gaps at the individual level, in that specific categories of individuals are not covered even if other members of their households are covered. Indeed, South Africa makes no provision for most unemployed, workingage adults; this category of individuals is‘covered’ only in the event that they have children and can access the Child Support Grant. Table 1.4: Social protection coverage by quintile, South Africa, 2010 Poorest Quintile Quintile Quintile Richest All quintile 2 3 4 quintile (%) (%)(%)(%)(%)(%) Social pensions 40 40 34 23 12 30 Total social assistance 86 80 68 47 22 61 Source: ASPIRE, variables per_sa_allsa.cov_q*_tot; per_sa_sp.cov_q*_tot, where* is the quintile number(i.e. 1 through 5). Table 1.5: Social protection benefits by quintile, South Africa, 2010 Poorest quintile (%) Quintile 2 (%) Quintile 3 (%) Quintile 4 (%) Richest quintile (%) Total (%) Social pensions 20 23 25 20 13 100 Total social assistance 25 25 24 17 9 100 Source: ASPIRE, variables per_si_cp.ben_q*_tot and per_sa_allsa.ben_q*_tot, where* is the quintile number(i.e. 1 through 5). The ILO’s own data underestimate the size of the coverage gaps, if defined in terms of cash benefits. Taking into account benefits in kind, and indirect forms of coverage, the gaps are perhaps less daunting. Taking into account inequity in the amount of the benefit,

33 however, reveals the extent to which social protection protects the non-poor rather than the poor. 1.5. Expenditure on social protection Coverage is reflected also in expenditure. The potential to close a coverage gap depends in part on fiscal capacity, which is likely also to be related to expenditure, with low levels of expenditure indicating‘under’’ -expenditure. The ILO, World Bank and other agencies have invested considerable effort in costing social protection packages and demonstrating thei r affordability(see Seekings, 2017a). Both the ILO’s and World Bank’s global surveys of social protection include country-level data on expenditure, usually measured as a share of GDP. As with other dimensions of coverage, it is not always clear precisely what kinds of expenditure are being measured. Do expenditure data include expenditures by subnational governments, or by international agencies? In addition, missing data must be either ignored or imputed. The ILO uses country-level data to estimate expenditure for the world as a whole, regions and, in the African region, Northern Africa and Sub-Saharan Africa separately. These data are reported in the text. In the statistical Annex, country-level expenditure data is also disaggregated into the following seven categories, all exclusive of expenditure on health(Table B.17):  Expenditure for older persons, excluding health  Expenditure for children  Expenditure for‘persons of active age’: General social assistance  Expenditure for‘persons of active age’: Social benefits, excluding general social assistance: o Unemployment o Labour market programme o Sickness, maternity, employment injury The ILO also reports in Table B.17 on total expenditure including health. Separately, the ILO reports on total expenditure over time, since 1995(Table B.16). Table 1.6: Aggregated expenditure data, as% of GDP Social protection (Excluding health) for children for working-age adults for older people World 0.4 2.2 3.3 2014 Report Africa Northern Africa 0.2 0.4 0.6 1.4 1.3 5.0 SubSaharan Africa 0.1 0.5 1.1 World 1.1 3.2 6.9 2017 Report Africa Northern Africa 0.6 0.1 SubSaharan Africa 0.7 1.1 1.2 1.0 2.3 3.7 1.6

34 Sub-total(excl health) 5.9 2.1 6.8 1.7 11.2 4.0 5.0 3.3 Health 2.8 2.6 3.2 2.6 NA NA NA NA Total(incl health) 8.8 4.3 10 4.3 NA NA NA NA Data in italics are my calculations. NA: Data not available, i.e. not published in the Report. Note that the data on expenditure by category for Africa do not add up to the reported total in the 2014 Report. Sources: 2014 Report, Table B.13, p306; 2017 Report, pp.18(children), 23(adults), 85(elderly) … The 2017 Report does not seem to include any estimate for global expenditure on social protection, but it does provide estimates for global expenditure on each of the three lifecourse stages(children, adults, elderly) as well as health, which can be used to derive a global figure. The 2017 Report also provides estimates for total expenditure at the regional level. Table 1.6 collates these data from the 2017 Report and the previous, 2014, Report. The 2017 Report indicates that expenditure in Africa is lower, as a share of GDP, than in the world as a whole, and that this is true also for every category of expenditure. The Table also suggests that expenditure is rising, in Africa and globally. Comparison of the data provided in the 2014 and 2017 Reports suggests remarkable growth in expenditure. Globally, reported expenditure on non-health social protection almost doubled from 5.9 to 11.2 percent of GDP. Expenditure rose on each of children, working-age adults and the elderly. In Africa, also, reported expenditures rose, approximately doubling: From 2.1 to 4 percent of GDP. Reported expenditure in Africa rose for each of children, working-age adults and older people. The 2017 Report puts expenditure on children in Africa at 0.6 percent of GDP, which was three times higher than the estimate of 0.2 percent reported in the 2014 Report. The increase in expenditure was even more dramatic for Sub-Saharan Africa(rising from 0.1 to 0.7 percent), offset by a decline in Northern Africa(from 0.4 to 0.1 percent). Reported expenditure in Africa on older men and women rose from 1.3 percent(and 1.1 percent in Sub-Saharan Africa) in the 2014 Report to 2.3 percent(and 1.6 percent in Sub-Saharan Africa) in the 2017 Report. The 2014 Report shows that reported expenditure on social protection(inclusive of health care) in Africa rose from less than 3 percent of GDP in the 1990s to 5.4 percent in 2009, before dropping slightly to 5.1 percent in 2010/11. As of 2010/11, expenditure was 9 percent in Northern Africa and 4.2 percent in Sub-Saharan Africa(2014: Table B.12, p297). The 2017 Report does not report regional aggregates in the statistical tables, but the figures in the text indicate that expenditure had resumed its upward path. These expenditure data are consistent with the ILO’s claim in the 2017 Report that the extent of coverage had risen across Africa as a whole. Above, I questioned whether the claimed rise in the extent of coverage across Africa was consistent with the country-level data published in the Report. The same question must be asked of the data on expenditure. Table B.16 in the 2017 Report provides data on total social protection expenditure (presumably inclusive of health) for both 2010 and 2015 for eight African countries, 40 for 40 Angola, Burkina Faso, Cote d’Ivoire, Egypt, Mozambique, Namibia, Seychelles and South Africa.

35 2010 and 2014 for five more countries, 41 and for one country 42 for 2011 and 2015. In these fourteen countries, expenditure as a share of GDP barely changed in four, 43 fell in seven 44 and rose in only three countries, and these were ones with small populations(the Gambia, Namibia and the Seychelles 45 ). On the basis of these data, it is difficult to understand how expenditure in Africa as a whole could have risen sharply, as the ILO claims. 1.6. Conclusion: Who gets what, who does not, and why? The World Bank, using a more inclusive definition of social assistance(or social safety nets) and counting indirect as well as direct beneficiaries, identifies many more poor beneficiaries than are suggested by the ILO. The ILO reports that 45 percent of the world’s population are partially or fully covered by social protection. In(Sub-Saharan) Africa, however, the proportion is only 18 percent, or 180 million. The World Bank counts a total of just over 100 million people in Africa(including North Africa) as benefitting from social safety net programs, including about 30 million children through school feeding programmes, another 20 million people through other feeding schemes, 28 million people through workfare(i.e. public employment programmes), less than 4 million through social pensions, and another 25 million people through other unconditional or conditional cash transfers(World Bank, 2015). 46 Most(and perhaps all) of these data appear to be for direct beneficiaries, notwithstanding the World Bank’s arguments(discussed above) for counting indirect beneficiaries also. The data exclude the beneficiaries of contributory (social insurance) programmes, who the ILO include in its coverage data. As the World Bank emphasizes, however, the level of protection provided by these programmes is often too modest to reduce substantially the poverty rate. Data on social assistance programmes compiled by SAPI suggest that the World Bank may have underestimated the reach of these programmes. This is in part because SAPI has collected more recent data as well as data on progammes omitted from the World Bank’s list. But it is also because SAPI reports data for in direct beneficiaries. Thus South Africa’s old age pension system has 15 million beneficiaries(of whom only 3 million are pensioners) and TASAF in Tanzania has 6 million beneficiaries(through transfers to 1.1 million households). Taking into account what programmes are and what are not included in their data, the ILO’s data on coverage seems to overestimate coverage among the elderly, children and workingage adults. The ILO’s data suggest that about 100 million children are covered by cash transfers. The World Bank and SAPI data on national programmes suggest that the more likely figure is closer to 25 million(with another 35 million benefitting from 41 Gambia, Madagascar, Mauritius, Sao Tome and Principe and Togo. 42 Uganda. 43 Cote d’Ivoire, Madagascar, Mauritius and South Africa. Expenditure in South Africa w as on a slow overall upward trend, however. 44 Angola, Egypt, Burkino Faso, Mozambique, Sao Tome and Principe, Togo and Uganda. 45 Expenditure in Seychelles rose only because expenditure in 2010 was unusually high. Expenditure in the 2010s generally was lower than in the 1990s and 2000s. 46 The World Bank also consider school and clinic fee waiver programmes as a form of social protection. I am doubtful of this, and exclude the beneficiaries form these totals.

36 feeding schemes and others as the indirect beneficiaries of other cash transfer programmes). The ILO’s da ta suggests that there between 10 and 20 million people are ‘covered’ by old -age pensions. The data on national programmes suggest that less than 5 million elderly people receive non-contributory pensions, perhaps 1 million elderly people live in households that receive a cash transfer(such as through TASAF in Tanzania), and less than 7 million people receive contributory pensions – giving a total of 13 million at most. The ILO’s data suggest that about 100 million working -age adults are covered. Using national data, I am unable to arrive at a figure even close to this. The ILO also reports on national coverage rates. The ILO sets the goal of universal coverage, including not only contributory schemes for the formally-employed but also universal social protection floors for the poor. Complete(100 percent) coverage would mean that everyone was protected against specified risks, and everyone who actually experienced that risk received cash benefits. The ILO reports that the highest overall coverage rate in Africa – i.e. what the ILO calls the‘front runner’ – is South Africa. Even South Africa‘is still only half way through the set indicator of universal coverage’, with coverage(by at least one programme) put at 48 percent(2017: 123). Effective coverage stood at 75 percent among children, 65 percent among the disabled, 93 percent among the elderly(including 74 percent under the means-tested, non-contributory pension), but only 11 percent among the unemployed. Under 4 percent of the working-age population (defined broadly as aged 15-64) was actually contributing to a contributory pension scheme. The‘scope’ of social protection is said to be‘comprehensive’ in South Africa (ILO, 2017, see above) – even though the safety net has gaping holes in it, especially through the absence of support for most unemployed people. 47 Egypt and Cabo Verde were the only other two countries with effective coverage(as measured by the ILO) of more than 30 percent of the total population(although data are not available for Mauritius). It thus seems clear that there remain massive coverage gaps in the coverage of cash transfer programmes. The ILO underestimates these coverage gaps, at the same time as it neglects the scale and importance of feeding and other in kind social assistance programmes. Moreover, as the World Bank’s ASPIRE data show, co verage in many countries is especially low among the poor, who are unable to participate in contributory systems. In general, contributory schemes are limited to formally-employed workers, mostly in the public sector or working for large private employers. Many formallyemployed workers and most informally-employed workers(and the self-employed) remain uncovered. Peasant farmers remain almost entirely outside of the social protection system, as defined by the ILO. The ILO’s neglect of distributional issue s – who precisely benefits from programmes – makes it difficult to compare either coverage or coverage gaps in the current context of less than universal coverage. A country could achieve(say) 20 percent coverage by targeting either the non-poor(for example, formal sector employees) or the very poor. The strength of the World Bank’s approach(and its ASPIRE database) is its emphasis on 47 The South African government rejected proposals for a basic income grant and has shown little real interest in an employment guarantee scheme(Seekings& Nattrass, 2005, 2015; Seekings and Matisonn, 2012).

37 distribution, i.e. on who gets what, and especially the needs of the poor. In South Africa, many but not all poor households benefit from social assistance. In Tanzania, however, many of the benefits of social assistance(in addition to almost all of the benefits from contributory pension programmes) accrue to the non-poor. Even the World Bank’s approach to distribution pr ovides only a partial picture of precisely who gets what. It focuses on programme beneficiaries in terms of the income distribution, i.e. of income quintiles. The World Bank does not report distribution in terms of categories of people, either within households or within income quintiles. The World Bank data do not allow us to identify distribution by gender. Nor do they allow us to see whether informal workers or peasants are covered. Overall, formally-employed workers are generally covered, at least for o ld age, through pension and provident funds. Poor‘incapacitated’ households – i.e. households where no one is able to work – are in more and more countries the focus of social assistance programmes, whether these are targeted on categories of individuals(pensioners, the disabled) or the households themselves(for example, through the PSSB in Mozambique or the cash transfer component of the PSNP in Ethiopia and Tanzania). Peasants are covered against the risk of drought, and may benefit from farm input subsidy programmes, which are sometimes very generous. Informally-employed workers are rarely covered at all. Non-contributory pensions and other social assistance programmes are beginning to cover peasants and informallyemployed workers in retirement, but regular provision for working-age adults is generally limited to workfare. In short, coverage gaps are very much wider than coverage. As part of their promotion of social protection, the ILO and World Bank talk up progress. The ILO reports, for example, that several African countries(including South Africa, Tanzania, Kenya and Zambia) have been‘making efforts’ to extend coverage to informal economy(ILO, 2017: 1234). Namibia has introduced a Vulnerable Children’s Grant, and several other governments have said they would introduce child grants. Senegal’s PNBSF,‘after a pilot phase in 2013, … has now been extended over the whole country, covering about 3 million persons’( ibid: 124). Maternity benefits have expanded in Ghana. Mozambique and Zambia are reported to be relaxing the targeting of their cash transfer schemes. The World Bank similarly emphasizes that‘many countries in Africa are introducing“flagship” SSN[social safety net] programs and are rapidly expanding their coverage’, citing the cases of Tanzania and Senegal(World Bank, 2017: 2). The data do not support the claim that there is a rapid expansion of social protection across Africa as a whole. Even the favoured examples are modest, with coverage reaching 10 percent of the population in Tanzania and 16 percent in Senegal in 2016 ( ibid: 3) – in countries that lack any existing cash transfer programmes for the elderly or children. The single largest contributor to the rising number of beneficiaries was South Africa, primarily as a result of the deracialisation of child grants, which was a reform with its own peculiar political circumstances and dynamics(Lund, 2008; Proudlock, 2012; Seekings, 2016f). Efforts to expand the coverage of contributory schemes among informal workers have made desultory progress.

38 It has become increasingly clear that governments(and elites more generally) across much of Africa are either ambivalent or outright hostile to the expansion of social protection. Even the ILO acknowledges, cautiously, that the expansion of family and child benefits has been‘ limited in part because of continued scepticism on the part of policymakers about providing transfers to new parents’(ILO, 2017: 124). Governments and elites prefer polices that promise to promote production, especially agricultural production, and hence reward work. Governments and elites also hold to a highly romanticised view of working adults supporting elderly, disabled and younger kin. Underlying the ambivalence(at best) of political and other elites is the character of society, the economy and politics across much of Africa. Almost everywhere, the formal sector is small. Workers in the informal sector seem hostile to the imposition on them of any obligation to contribute themselves to contributory schemes. Indee d, the ILO’s enthusiasm for the formalisation of informal work has not been matched on the part of governments across most of Africa. Peasant farmers have been left off the map of social protection almost entirely. The ILO ignores the programmes that insure peasants against the risks of drought, disease and price fluctuations, and thus underpin some degree of income and/or food security. The ILO even ignores the feeding and workfare programmes that benefit the rural as well as the urban poor(and often non-poor also). A few countries have old-age pension programmes, and others have the kind of safety net programmes targeted on poor,‘incapacitated’ households.

39 Appendix to Chapter 1 Table 1.A1: Coverage of Africa in ASPIRE database Most recent Number Countries year of survey of data countries 1999 1 Gambia 2003 2 Benin, Cote d’Ivoire 2005 1 Congo 2006 2 Gabon, Kenya 2007 1 Cabo Verde 2008 4 CAR, Egypt, Liberia, Zimbabwe 2009 2 Botswana, Mozambique 2010 5 Morocco, Namibia, South Sudan, Sudan, Swaziland 2011 5 Chad, Madagascar, Rwanda, South Africa, Zambia 2012 7 DRC, Djibouti, Ethiopia, Niger, Senegal, Sierra Leone, Togo 2013 5 Ghana, Mauritius, Nigeria, Tanzania*, Uganda 2014 4 Burkino Faso, Cameroon, Mali 2015 2 Lesotho, Mauritania Source: ASPIRE database. * The most recent data for Tanzania is 2012-13, and is included under 2012 in the database.

40 Chapter 2: Rights for Whom in Global Rights Discourses Can a rights-oriented approach help to close the wide coverage gaps that currently characterize social protection across Africa? This clearly depends on rights for whom, to what and how? The global discourse of human rights is not uncontroversial, not least because of its historical roots in western liberal theory with its foundational assumptions of the primacy of autonomous individuals(see, for example, Claeys, 2015b: 1, 83). In order to close gaps in the coverage of social protection, we must first ask how the global discourse of social and economic rights has constituted the bearers of rights(rights for whom?) and the content of those rights(rights to what?). In subsequent chapters we shall turn to the question of‘how?’. If the most important quest ion to ask of any system of social protection(or‘welfare regime’ more broadly) is‘who gets what?’, then we need to identify carefully whose rights are privileged in prevailing rights discourses.‘Human’ rights can constitute people, whether as individua ls, families or other groups, in a variety of forms.‘Welfare regimes’ across the world differ in part in whether they are oriented primarily towards workers, peasants, paupers or citizens. The global social and economic rights discourse evolved through the privileging of some of these categories over others. From the late nineteenth century, the predominant discourse of social and economic rights focused on workers, and many welfare regime s were‘workerist’, focused primarily on the incorporation of industrial and urban workers through the institutionalization of an‘industrial’ form of citizenship. An alternative discourse, rooted in European Christian religious thought, focused on the ‘social’ rights of the poor, or paupers, whether to charity or the support of the modern state. Following the Second World War, north European welfare states institutionalized the kind of‘social’ citizenship imagined by T.H.Marshall. At the same time, across most of colonial Africa and South Asia, states constructed agrarian welfare regimes, focused on peasant farmers, whose rights were imagined in terms of a‘traditional’ agrarian order. More recently, arguments that citizens have a right to a fair or rightful share in national (or global) prosperity have fueled campaigns for basic income grants and similar proposals. The relative importance of these different categories of rights-holders reflects both the diverse currents in the history of ideas and the social and economic realities of different societies. Workerist conceptions of rights(and workerist welfare regimes) developed in response to industrialization and the risks that accompanied this. Pauperist conceptions of rights were fueled by concern with the people left behind by industrialization. Agrarian rights and regimes characterized societies dominated by peasant farmers, or at least where elites imagined that the peasantry might be revived. Deepening landlessness,‘deagrarianisation’ and urbanization in the late twentieth and early twenty-first centuries undermined the relevance of agrarian welfare regimes. Each of these types of welfare regime focuses on a different category of people, protects them against a distinctive set of risks, imagines a distinct ideal society, and is justified through

41 a specific set of norms, values and beliefs. Rights discourses have evolved in relation to these diverse risks, ideals and norms. In the decade since the mid2000s, a‘global’ discourse of rights to social prot ection has become prominent. This discourse revolves around the concept of the‘social protection floor’, championed initially by the ILO but later embraced also by diverse other international agencies(including UNICEF and the World Bank), NGOs(including HelpAge International) and aid donors(including DfID). Backed by a powerful array of international organisations, this discourse has become hegemonic at the global level. It has also been resisted, by diverse actors including both peasants in Central America and political elites across much of Africa. In important respects, the global discourse represents a compromise between the two strands of thought about social and economic rights that have predominated in Western Europe for almost a century, i.e. between industrial and social forms of citizenship. The social protection floor institutionalized this compromise through its combination of‘horizontal’ and‘vertical’ elements(i.e. social assistance for citizens and social insurance for workers respectively). The global discourse is rooted in the assumptions about rights-bearing individuals that underpin both industrial and social forms of citizenship. It pays little or no attention to alternative conceptions of citizenship, including especially the ideas about collective rights that underpin agrarian forms of citizenship. In this first section of this report, we examine the history of the‘global’ discourse of rights to social protection, exploring its roots in the ideas and conditions existing in Western Europe in the early and mid-twentieth century. Examining the history of the‘global’ concern with these rights helps us to understand their application in Africa, both in the past and the present. 2.1. The Origins of Rights-Based Approaches to Social Protection The 1948 Universal Declaration of Human Rights is widely cited as the foundational statement of social and economic rights(e.g. Munro, 2008: 30; von Gliszczynski, 2015: 160). Whilst primarily concerned with civil and political rights, which had obviously been violated flagrantly during the Second World War, the Declaration included also some social and economic rights. These were based around a conception of‘livelihoods’ that extended beyond waged employment but nonetheless reflected assumptions rooted primarily in the experience of the global North. Article 22 proclaimed that‘everyone, as a member of society, has the right to social security’. The reference here to‘social security’ encompassed, at least potentially, social assistance as well as social insurance(along the lines of the Philadelphia Declaration, discussed below). Article 23 concerned employment. It stipulated that‘ Everyone who works has the right to just and favourable remuneration ensuring for himself and his family an existence worthy of human dignity, and supplemented, if necessary, by other means of social protection ’. Article 25 proclaimed that‘ Everyone has the right to a standard of living adequate for the health and well-being of himself and of his family, including food,

42 clothing, housing and medical care and necessary social services, and the right to security in the event of unemployment, sickness, disability, widowhood, old age or other lack of livelihood in circumstances beyond his control’. The reference to‘livelihoods’ potentially encompassed forms of self-employment as well as wage employment. The qualification‘beyond his control’ pointed to social risks, and made clear that rights did not extend to loafers. Article 25 added that‘Motherhood and childhood are entitled to special care and assistance’. The language expresses the ge ndered assumptions(of male breadwinners and dependent women) that were hegemonic across the global North at the time(despite the erosion of the gender division of labour during the Second World War) but did not apply across most of the global South, where women were very often also‘breadwinners’. The Universal Declaration’s references to social and economic rights were informed by earlier statements about rights, especially the series of‘conventions’,‘recommendations’ and‘declarations’ adopted by the ILO. The ILO had been established under Part XIII of the Treaty of Versailles(following the First World War) in order to promote‘social justice’ which was seen as a necessary foundation of‘universal peace’(and hence protection against the risk of another terrible war). The preamble to Part XIII of the Treaty identified explicitly‘the protection of the worker against sickness, disease and injury arising out of his employment’ and‘provision for old age and injury’ as elements in‘social justice’. The ILO first focused on these issues between 1925 and 1927, adopting two conventions and a recommendation on sickness insurance in 1927. The ILO turned again to these issues between 1932 and 1934, culminating in six conventions on compulsory insurance for oldage, invalidity, widows and orphans, for industrial and agricultural workers, in 1933, and an additional convention and recommendation covering‘benefits or allowances to the involuntarily unemployed’ in 1934. Through the 1920s and 1930s, the ILO understoo d social welfare primarily in terms of social insurance. It considered the German model not only superior to alternatives(such as the British and Scandinavian models, which emphasized strongly social assistance) but also as universally appropriate. ILO officials referred disparagingly to the alternative model of tax-financed social assistance as little more than‘social relief’, which would inevitably be replaced over time by social insurance. The ILO’s focus was on wage -earners, and their rights, not on citizens per se(Seekings, 2008b; Kott, 2010). During the Second World War, the ILO relocated from Geneva to Montreal and for a while became far more open to alternatives to the German model in its preparations for postwar‘reconstruction’. The ILO embra ced both social insurance and social assistance, representing them as equals: If‘social assistance is a progression from poor relief in the direction of social insurance’, as the advocates of the German model had argued earlier, then‘social insurance is a progression from private insurance in the direction of social assistance’. In the exemplary cases of New Zealand and Denmark,‘we can no longer say whether social assistance or social insurance predominates, but only that they possess a national system o f social security’(ILO, 1941: 82 -3; see further Seekings, 2008b). This process of rethinking was encouraged by the publication in Britain at the end of 1942 of the Beveridge Report, which set out a grand vision for the postwar welfare state. The ILO’s de liberations culminated in the 1944 Philadelphia Declaration on Income Security.

43 The Philadelphia Declaration proclaimed the imperative to extend‘social security measures to provide a basic income to all in need of such protection and comprehensive medical care’, as part of a broad strategy to raise the standard of living. More detail was provided in an accompanying Recommendation(no. 67), which emphasized that both social insurance and assistance were required‘to relieve want and prevent destitution by restoring, up to a reasonable level, income which is lost by reason of inability to work (including old age) or to obtain remunerative work or by reason of the death of a breadwinner’. Social insurance was to be preferred but social assistance was necessary also to meet the needs of people who were not covered by social insurance. The 1944 ILO also adopted a Recommendation(no. 70) on‘Minimum Standards of Social Policy in Dependent Territories’, referring to colonies(and protectorates), including almost all of Africa. This was the first international policy statement addressed directly at social protection in Africa. Here the ILO retreated from the bolder vision of the Philadelphia Declaration. The detailed report that motivated for the Recommendation emphasized the need for a‘general programme of well being and development’ because it was insufficient to regulate employment relationships. The ILO should‘contribute positively, not merely to better labour laws and administration, but also to policies that should provide for more wealth and for a better distribution of wealth’(ILO, 1944a: 26). A key factor in propelling the ILO towards a more encompassing approach was its recognition that the regulation of formal employment – including any social insurance for workers in formal employment – affected only a small proportion of even the working population in most colonies. In colonies where most people were‘engaged in independent production, possessing in their lands, crops and herds some safeguards against personal disaster, measures of social security, by which the individual usually without such resources is granted a measure of protection against the hazards of his life, must have a lower priority than policies designed to raise the productivity and remune ration of the population as a whole’ (ILO, 1944b: 43-5). The ILO advocated an appropriate balance not only between protection and development but also between different kinds of institutions. Recognising differences between European and non-European societies, the ILO suggested that the new challenges of achieving income security could be met through either‘modern’ or‘traditional’ institutions of‘social responsibility’. ‘Present conditions predicate the closest possible co -ordination between the expansion of production and the improvement of standards of living. This can best be effected, not by dictatorial means, but by full and open collaboration at all levels of production and of social life. In modern production trade unions, employers’ association s, co-operative societies and their federations are the chief units of labour collaboration. In primitive production the expanded family, the clan, the village and the tribe fill a similar role. As one form of production emerges into another, some of the characteristics of both types of organization may become welded. … In dependent territories the development of indigenous institutions of social responsibility, whether they be industrial organizations of a modern type or traditional associations capable of adjusting themselves to new conditions, is fundamental to the fulfillment by the colonial

44 powers of trusteeship and to the acquisition by the dependent community of responsibility and a capacity for partnership.’(ILO, 1944a: 9) The 1944 Recommendation specified some‘minimum standards’. Article 38 concerned the elderly, and implied that coverage should be universal, i.e. not limited to retired workers:‘Such arrangements as are practicable, having due regard to local conditions, shall be made for the maintenance and treatment of the sick and for the care of the aged, of the incapacitated and of the dependent survivors of deceased persons’. Note, again, the caveats‘as are practicable’ and with‘due regard to local conditions’. The final Recommendation discarded an article included in an earlier draft, calling for the introduction of‘compulsory insurance for the protection of employed persons and their dependents in cases of sickness and maternity, old age, death of the breadwinner and unemployment’. The draft had included a caveat(‘as soon as the necessary conditions for the operation of such insurance are present’; ibid: 59) but even this was presumably too strong for the colonial powers within the ILO. The ILO was well aware of the need to take into account local conditions. For twenty years, trade unions from the more industrialised countries had pressed for higher minimum standards, with some support from employers, wanting to curtail low-wage competition. Employers and governments from less industrialised countries, with some support from their trade unions, had generally resisted this standardization. The imperial powers were strongly opposed to any suggestion of parity between their citizens in Europe and their colonial subjects in African and else where. Paying‘due regard to local conditions’ was the formula that allowed a principle to be established whilst diluting its implementation(Seekings, 2010). The ILO’s approach to welfare in colonial territories was broadly similar to the approach formulated by the British state, presumably reflecting in large part the considerable authority of the latter as the major imperial power. In the 1930s and early 1940s both colonial officials and local political leaders in a number of colonies had expressed interest in social assistance. The Caribbean island colony of Barbados, followed by neighbours Trinidad and Tobago and British Guyana, introduced social pensions as a right. Faced with an Empire-wide flurry of excitement in response to the 1942 Beveridge Report, the British Colonial Office formulated a policy on social security that explicitly discouraged Britishstyle reforms and steered colonial governments towards‘development’ instead. 48 The underlying view had been expressed earlier by the Colonial Office’s labour advisor: ‘There seems to be some reason for doubt whether the methods of the Beveridge Report are quite suited to colonial conditions. It is designed for an advanced democratic community with a high standard of education and a long experience of social services; such a description is clearly inapplicable to the populations of the African colonies, or even the West Indies.’ 49 European-style reforms might be required for the small but significant population of urban workers, many of whom had weak links to rural society, but‘development’ was the answer for the vast majority of the population living in the countryside. 48 The key document here was a Memorandum on Social Security in the Colonial Territories, completed in mid-1943 but only published in 1944(see Seekings, 2013). 49 OrdeBrowne,‘Social Insurance in the Colonies’(21.4.43)(CO 859/78/1, document 1), p4.

45 Cooper, in his detailed study of the colonial regulation of labour, argues that the ILO played an important role in both the British and French colonial governments’ embrace of policies intended to‘stabilise’ African labour. The‘French and British governments … took the ILO agenda quite seriously in internal discussions’, and sought to use it to their advantage by claiming the moral high ground at the same time as depoliticizing labour issues(Cooper, 1996: 219-20). British and French government colonial experts met regularly to discuss(in technical terms) policy reforms. The first Inter-African Labour Conference was held in Nigeria in 1948. The policy agenda included social welfare reforms: Officials agreed that African workers should be insured against all risks, and oldage pensions should be introduced‘where tribal organization has ceased to be effective’ (quoted, ibid: 222). Whilst the small minority of the population that was formally employed could be covered by statutory social insurance, the vast majority of the population in rural areas got ‘development’ instead. In exceptional cases – such as Mauritius – colonial officials acceded to local demands for social pensions or other forms of social assistance. In some other cases, including Northern and Southern Rhodesia, the presence of destitute Europeans pushed governments into racially discriminatory provision. In general, however, development served as the substitute for the kind of‘social citizenship’ that T.H. Marshall spoke about in 1949/50. 50 This background serves to explain the specification of social or economic rights in the 1948 Universal Declaration of Human Rights(and in several constitutions adopted by newly-independent countries at this time). The imagined rights were not the rights of peasant farmers in‘developing’ colonies or countries. They were the rights – or privileges – of(mostly male) citizens in advanced capitalist economies, who were assumed to be waged workers in‘modern’ or formal employment, their dependents(wome n and children), or people who were unable to work for wages on the basis of age, infirmity or illness. Despite the description of these rights as universal, they were heavily circumscribed and reflected the conditions, ideas and concerns of the‘western’ powers. 2.2. The promise of and limits to‘development’, 1950 -1996 Despite the acknowledgement of both social insurance and social assistance in the Philadelphia Declaration, the ILO continued in practice to privilege the rights of industrial workers through the extension of social insurance through the 1950s, 1960s and 1970s. Social assistance almost disappeared from the agenda. Insofar as the rural poor – who comprised a large majority of the poor across the global South – had any right to economic sec urity, this was to be achieved entirely through‘development’. The privileging of industrial workers and social insurance was evident in the ILO’s 1952 Social Security(Minimum Standards) Convention(no. 102), which set out globally50 Prior to Marshall’s coining the phrase‘ ’ social cit izenship’, Karl Polanyi had already analysed the phenomenon in The Great Transformation: In his account, the demand for state programmes, to decommodify citizens, was in response to the inequalities generated by market commodification(Polanyi, 1944).

46 applicable minimum standards, for both coverage and benefits, for provision against poor health, disability and death, unemployment and old age, as well as family benefits. Countries that ratified the convention were obliged to provide social security in at least three of the nine listed branches. The convention stipulated that provision could be made through either social insurance or social assistance(including either targeted or universal schemes). Old-age benefits, for example, should cover either a specified proportion of workers(i.e. through social insurance) or of people with incomes below a specified minimum(i.e. through means-tested social assistance). But the convention clearly favoured contributory social insurance. This was clear in the articles that allowed for partial exemptions for any country‘whose economy and medical facilities are insufficiently developed’. Most of the exemptions were partial in that they required 50 percent coverage of the industrial workforce rather than of the total workforce or population(ILO, 1952). The Convention was a compromise, partly brokered by the Indian Government’s delegate, whose called for minima to be kept‘reasonably easy so as to make implementations possible for a large number of countries’. It would take‘a very long time’ to implement even the compromise Convention, and more onerous conditions would render implementation‘impossible’. Meanwhile, he suggested,‘Much higher standards can be laid down for advanced countries when dealing with advanced standards of social se curity.’ 51 Even the compromise Convention was ratified by very few countries. By 1970, only six countries from the global South had done so. Similar problems beset the 1966 Convention no. 128 on old age pensions, which specified minimum coverage and benefit rates. Old-age benefits should cover either ‘(a) all employees, including apprentices, or (b) prescribed classes of the economically active population, constituting not less than 75 per cent of the whole economically active population, or(c) all residents or residents whose means during the contingency do not exceed[prescribed] limits’(emphasis added). Countries whose economies were‘insufficiently developed’ might apply for a partial exemption, allowing even more limited coverage(ILO, 1966). In practice, few African countries chose to ratify most of these conventions. Many – as we shall see – did expand social insurance coverage for formally employed workers, but this class was rarely poor. Indeed, in 1968 the ILO itself published a study that showed that social insurance programmes typically redistributed in ways that worsened inequality in developing countries(Paukert, 1968). Whilst it is unclear what impact this – and other, similar research – had within the ILO, it certainly served to highlight the paradox of an organization that proclaimed its pro-poor credentials whilst championing in practice programmes that, on their own, benefited semi-privileged minorities in the global South. The UN also sought to deepen commitments to rights but did so with a limited vision of the kinds of livelihoods and risks that predominated across the global South. The International Covenant on Economic, Social and Cultural Rights(ICESCR) was adopted in 1966(although it was supposed to be binding onl y from 1976). As with the ILO’s efforts, the focus was primarily on formal employment and employment-related social insurance. Articles 6, 7 and 8 concerned employment. Article 9 recognised‘the right of everyone to social security, including social insura nce’. Social assistance was not mentioned. The rights of the selfemployed were neglected. Article 10(2) specified that‘Special protection 51 ILC, 35 th Session, Geneva, 1952, Record of Proceedings(Geneva: International Labour Office, 1953), p407.

47 should be accorded to mothers during a reasonable period before and after childbirth. During such period working mothers should be accorded paid leave or leave with adequate social security benefits.’ Article 11 specified‘the right of everyone to an adequate standard of living for himself[sic] and his[sic] family, including adequate food, clothing and housing, and t o the continuous improvement of living conditions.’ Article 10 acknowledged that some women worked, but Article 11 assumed that society generally comprised male breadwinners and their dependents. Articles 12 and 13 concerned access to health care and education(which, at primary levels, should be free; parties should work towards free education at secondary and tertiary levels also). The Covenant provided for an Economic and Social Council, which subsequently established a Committee on Economic, Social and Cultural Rights, comprising human rights experts, to monitor the implementation of the Covenant. The Covenant included in Article 2 the principle of‘progressive realisation’. Signing up to the Covenant committed states to‘ take steps... to the maximum of its available resources, with a view to achieving progressively the full realization of the rights recognized in the present Covenant by all appropriate means, including particularly the adoption of legislative measures .’ This principle of‘progressive realisation’ was more compelling than the ILO’s concern with‘taking into account local conditions’: Even if resource constraints inhibited the implementation of the listed rights, states were required to act as best they could, and to achieve more over time. Some of these articles might seem to be broadly encompassing, but they were of little import in practice. It is not clear what African countries contributed to the process of drafting the Covenant. Some countries signed the Covenant but delayed ratification. Kenya and Tanzania ratified it in 1972 and 1976 respectively. Zambia did so in 1984, and Uganda in 1987. Zimbabwe did not do so until 1991, Malawi until 1993, Namibia until 1994, Ghana until 2000 and South Africa until 2015! Neither Botswana nor Mozambique have ever signed or ratified it. 52 Both the ILO’s Conventions and the UN’s Covenant assumed that rights applied to the formal economy, and only affected the rural poor(and the growing population of urban poor) insofar as‘development’ resulted in them moving into the formal economy. This failed to confront two huge problems across much of the global South: Famine and the limits to development. Droughts(and floods) posed perhaps the most serious risk to the livelihoods of most poor people in Africa and some other parts of the global South. Colonial and later post-colonial governments across much of Africa were faced intermittently with drought and the prospect of famine. If European welfare states – especially the Bismarckian ones – were primarily concerned with protecting workers against the risks of an industrial economy, colonial states in Africa were compelled to confront the risks associated with agrarian societies, especially the risk that regional drought would cause a collapse in food production and the decimation of livestock herds, overwhelming the capacity of kin, communities and‘native authorities’ to support the destitute. In colonial India, the British 52 According to Wikipaedia.

48 authorities had long recognized the imperative of government intervention in the event of drought and famine, along the lines of‘famine codes’ that set out what should be done, when, and by whom(see Drèze, 1990). 53 There were no such famine codes in subSaharan Africa, although the settler states of South Africa and Southern Rhodesia(later Zimbabwe) did succeed in minimizing famine by the late nineteenth century and 1920s respectively(see Iliffe, 1990, on Southern Rhodesia). Elsewhere, however, colonial authorities usually intervened inadequately and too late, with the result that drought led to famine and starvation – for example, in Nyasaland(Malawi) in 1949(Vaughan, 1987; Mandala, 2005). As the‘winds of change’ swept through Africa, in the late 1950s and early 1960s, even colonial administrations were compelled to respond to droughts more effectively, and post-colonial governments faced even greater pressure. Botswana, which achieved independence during the worst drought in living memory, pioneered drought relief and recovery programmes, in association with the newly-formed World Food Programme(Seekings, 2016a). The new government of Botswana not only guaranteed in practice the right to food security of most of its citizens, but also formulated an ideology that recognized this right(as we shall see further in Chapter 4 below). The work of Amartya Sen(1981) focused attention on the ways in which famines were the consequence not simply of food shortages, but rather of a distribution of resources that left the poor unable to secure food supplies. Drought relief was thus fundamentally about decommodification. The institutionalization of drought relief and recovery programmes across most of Africa between the 1970s and 2000s represented an extraordinary expansion of social protection and social rights, albeit not in the form described in rightsoriented plans from most international organisations or organisations based in the global North. In addition, it was becoming clear that in practice development was failing to pull enough poor people into the formal,‘modern’ economy. Rural(and urban) poverty persisted. This put the ILO in a tricky position, because it began(again) to look like it was protecting the interests of a privileged minority across most of the global South, i.e. the so-called labour aristocracy. In the 1970s, the ILO deepened its efforts to promote poverty-reducing development or growth paths. An inquiry in Colombia in 1970(followed by an inquiry in Kenya in 1972) fed into new emphasis on employment creation, the satisfaction of‘basic needs’, and later the informal sector also. The ILO adopted the concept of‘basic needs’ in 1976. Applied to the case of Tanzania, this meant emphasising measures to ensure improved health, education, shelter, access to water and sanitation, and to promote smallscale agricultural production – and not simply on the need to expand formal employment. The ILO report did not mention social security, but did advocate labour-intensive public works programmes to build infrastructure(ILO, 1982). In 1974, the World Food Conference focused attention on food insecurity across much of the world. It became even more difficult to ignore distribution when structural adjustment programmes were implemented in the face of debt crises. The position of children received particular attention. In 1987, UNICEF focused on the position of children in its critical response to structural adjustment, Adjustment with a Human Face(Cornia, Jolly and Stewart, 1987). Two years later, the UN adopted the Convention on the Rights of the 53 They sometimes failed, most notably in the Great Bengal Famine of 1943.

49 Child. This recognized, inter alia, the right of‘every child … to benefit from socia l security, including social insurance’(Article 26) and‘ to a standard of living adequate for the child's physical, mental, spiritual, moral and social development’ (Article 27). The phrasing reflected the earlier ICESCR and reflected a similarly limited understanding of livelihoods and risks. Whilst families had primary responsibility for the welfare of their children, states were required,‘in accordance with national conditions and within their means’, to‘take appropriate measures to assist parents and others responsible for the child to implement this right and shall in case of need provide material assistance and support programmes, particularly with regard to nutrition, clothing and housing’. 2.3. The resurgence of a rights-based approach from 1996 Between the 1950s and 1990s, the idea of a right to social protection to eclipsed by the preoccupation with‘development’. Insofar as post -colonial Africa was concerned, social protection was more of a privilege that was or would be consequential to development, not a right to be claimed before or alongside it. This decades-long preoccupation with growth and development was challenged in 1996-97. First, in 1996, the FAO hosted in Rome a massive World Food Summit. Then, the following year, UN Secretary-General Kofi Annan called on all UN agencies to incorporate human rights into their various activities and programmes, within the framework of their respective mandates. The various UN agencies began to engage with the question of social and economic rights is an uncoordinated way. Then, in 2003, they collaborated in a workshop(in Stamford, Connecticut) to draw up an agreed‘UN Statement of Common Understanding on Human RightsBased Approach to Development Cooperation and Programming’. The ‘Stamford Common Understanding’ identified three points. First,‘all programmes of development co-operation, policies and technical assistance should further the realisation of human rights as laid down in the Universal Declaration of Human Rights and other international h uman rights instruments’. Secondly, these human rights should‘guide all development cooperation and programming in all sectors and in all phases of the programming process’. Thirdly, development cooperation should build capacities among states(and non-st ate‘duty bearers’) to meet their obligations and among rights -holders to claim their rights. 54 The ILO soon assumed a leading position in developing a new approach. The ILO began to rethink more fully its‘labourist’ approach in the late 1990s, in part in response to the apparent hegemony of the World Bank in debates over social protection. Whereas the World Bank had no interest in rights, the ILO had a long history of at least notional concern with them. The UN SecretaryGeneral’s endorsement of a rights -based approach probably fueled its reassessment of its approach. In an important early contribution, van Ginneken(1999: 3) wrote that‘the ILO had traditionally expected that all workers would sooner or later end up in secure formal sector employment’. Th is assumption had justified the ILO’s focus on the rights of workers in formal employment and on development 54 See http://hrbaportal.org/the-human-rights-based-approach-to-development-cooperation-towards-a-commonunderstanding-among-un-agencies.

50 strategies that would expand the formal sector. In recognising that there was in practice very limited scope for extending the reach of existing contributory institutions, van Ginneken exposed the inadequacy of the ILO’s existing approach. The coverage of contributory programmes might be extended from 10-25 percent of the working population (and their dependents) by an additional 5-10 percent of the working population, but not further. If the ILO was to concern itself with the rights of the poor, they must look beyond contributory programmes. Van Ginneken(1999) recommended a multi-pillar approach to social protection. Existing contributory schemes might provide for between 10 and 30 percent of the population. New(and largely unspecified) contributory schemes might provide some coverage for another 40-60 percent of the working population. But taxfinanced social assistance and other schemes would be required to provide for the poorest 30 percent of households. Van Ginneken thus no only exposed the insufficiency of contributory schemes, but also put an unusually precise figure on the scale of social assistance that would be required. In 1999 the Governing Body of the ILO initiated a major and formal re-examination of social security. The 2000 World Labour Report addressed many aspects of social security, under the title Income Security and Social Protection in a Changing World(ILO, 2000). The 2001 International Labour Conference decided to campaign on the theme of ‘Social Security for All’, and two years later formally launched a‘ Global Campaign on Social Security and Coverage for All’(although this was initially focused on community based micro-insurance). The ILO began to develop the concept of the social protection ‘floor’, i.e. a set of programmes that would guarantee a minimum level of provision across all stages of the life-course(see e.g. Cichon and Hagemejer, 2007). This marked a massive shift for the ILO, from preoccupation with what might be called industrial citizenship(for industrial workers) to social citizenship(for all citizens, as imagined by T.H.Marshall). The financial crisis accelerated the rethink and helped to mobilise support in other organisations, including the World Bank. In 2009, diverse UN agencies agreed to the ILO’s proposal to launch a Social Protection Floor Initiative(co -chaired by ILO and the WHO) as one of nine UN joint initiatives to cope with the effects of the economic crisis. The following year the ILO published two important reports: Extending Social Security to All: A guide through challenges and options (the‘Guide’)(ILO, 2010a) and a World Social Security Report(ILO, 2010b), which proved to be once-off, but was resuscitated in 2014 as the World Social Protection Report). An Advisory Group was established, chaired by former Chilean president Michelle Bachelet, to consider the proposed‘floor’. The Advisory Group reported in 2011, and in 2012 the ILO adopted Recommendation no. 202 on Social Protection Floors. Michel Cichon – who as director of the ILO’s Social Security Department from 2005 to 2012 played a central role in the social protection floor initiative(Deacon, 2013) – gushingly described Recommendation 202 as ‘ a“Magna Carta” of social protection’ (Cichon, 2013: 37) – referring the the charter of 1215 in which the then King of England recognised that his subjects had civil and political rights as citizens. The ILO Recommendation opened with a preamble that reaffirmed‘that the right to social security is a human right’. Social security was also developmental,‘ an important tool to prevent and reduce poverty, inequality, social exclusion and social insecurity, to promote equal

51 opportunity and gender and racial equality, and to support the transition from informal to formal employment’, and a useful macroeconomic tool to counter recession. For the ILO, the social protection floor initiative was linked to its broader strategy around ‘decent work’. Including the horizonal floor(comprising social assistance) meant that the ILO could sidestep criticisms that it was only concerned with non-poor workers in formal employment. T the same time, the ILO was able to move forward with the social protection floor only because the trade unions(forming one arm of its tripartite structure) were satisfied that the establishment of a horizontal floor was a complement not a substitute for the vertical strengthening of social protection through the protection and expansion of contributory schemes. Not long after, in June 2015, the ILO approved another Recommendation – the Transition from the Informal to the Formal Economy Recommendation(number 204)(ILO, 2015) – that was more directly in keeping with its long-standing workerist agenda. The Recommendation called on governments(and other actors) to‘( a) facilitate the transition of workers and economic units from the informal to the formal economy, while respecting workers’ fundam ental rights and ensuring opportunities for income security, livelihoods and entrepreneurship;(b) promote the creation, preservation and sustainability of decent jobs in the formal economy and the coherence of macroeconomic, employment, social protection and other social policies; and(c) prevent the informalization of formal economy jobs.’ For the ILO, the expansion of the rights of citizens went along with the defence and expansion of the rights of workers. Social and industrial citizenship remained closely bound together. Another important early player in the shift towards a rights-based approach to social protection was the Organisation for Economic Cooperation and Development(OECD)’s Network on Poverty Reduction(POVNET). In 2004, POVNET established a Task Team on Risk, Vulnerability and Social Protection(later renamed Social Protection and Social Policy)(Voipio, 2007). This brought together a large team of social development and social protection experts from several OECD countries(including Finland, Germany, Sweden and the UK), international agencies(including the World Bank, ILO and UNICEF), and large international civil society organisations(including HelpAge International). Another early convert to rights-based approaches was the UK’s Depart ment for International Development(DfID), which played an important role in Africa. 55 DfID had been established as a stand-alone government department in 1997. Its founding White Paper on Eliminating World Poverty pointed at a rights-based approach. Its 20 00‘Target Strategy Paper’ on Realising Human Rights for Poor People described the human rights approach to development as‘empowering people to take their own decisions rather than being the passive objects of choices made on their behalf’. DFID’s objective was‘to enable all people to be active citizens with rights, expectations and responsibilities’ and to‘claim their rights to the opportunities and services made available through pro -poor development’(DfID, 2000). In 2002 -03, DfID established in its Policy Division a Reaching the Very Poorest(RtVP) Team. The RtVP Team promoted a focus on cash transfers as a key instrument in the pursuit of poverty eradication and inclusive economic 55 This section is drawn primarily from research by Sam Hickey, included in Hickey and Seekings(2017).

52 development. The RtVP evolved into a Social Protection Team, which in turn was later incorporated into a broader Team on Equity and Rights. At much the same time, HelpAge International, which played a leading role in promoted the interests of elderly people in the absence of any dedicated UN agency equivalent to UNICEF, al so embraced social protection in the form of‘social’ - i.e. non-contributory – pensions. Later, UNICEF followed suit. By 2012, UNICEF described itself as following‘a human rights-based approach to ensuring that duty bearers are accountable to rights holde rs and that rights holders are able to claim their rights to social protection’, and an ‘equity focused approach to development’(UNICEF, 2012: 3). The heightened stature of social protection on the global agenda was reflected in its privileged position in the SDGs, agreed in 2015, in stark contrast to the complete silence on social protection in the earlier MDGs. The SDGs included a dedicated target on social protection(SDG 1.3):‘Implement nationally appropriate social protection systems and measures for all, including floors, and by 2030 achieve substantial coverage of the poor and the vulnerable’. Social protection was also mentioned explicitly in SDGs 5(Achieve gender equality, empower all women and girls) and 10(Reduce inequality within and among countries). 2.4. The rights of peasants The revival of a global rights discourse in the early 2000s focused on the rights of workers, perhaps including informally-employed workers as well as those in formal employment, and citizens. One group that was absent from the discussion was peasants. Yet, through the twentieth century, most of the world’s poor were peasant farmers, i.e. smallholders growing crops and raising livestock primarily for subsistence, sometimes for sale. Even at the very end of the century, only 20 percent of the worldwide population experiencing hunger were urban. An estimated 50 percent were peasant farmers, 22 percent were landless labourers in rural areas, and 8 percent were herders, fishers or hunters(Ziegler et al., 2011). Peasants were excluded from the rights agenda because, for most of the twentieth century, the rural poor across the global South(i.e. the‘developing’ world) were defined as a challenge for‘development’ rather than‘welfare’. ILO Conventions and UN Covenants generally assumed that it was‘development’ that created the need for ‘welfare’, in that it created the kinds of risks against which workers could be protected through social insurance(or, exceptionally, social assistance). In this view, what the rural populations of the global South needed was, first, development, and only then did states need to worry about expanding‘welfare’. This was not the whole picture. First, a formal right to food fluttered across the global South. The right to food was included(in passing) in Article 25 of the Universal Declaration of Human Rights. The 1966 International Covenant on Economic, Social and Cultural Rights – drafted with the assistance of the FAO(Claeys, 2015b: 64) – recognized the

53 ‘right to an adequate standard of living, including adequate food ’(Article 11.1, emphasis added), as well as the‘fundamental right to be free from hunger ’(Article 11.2, emphasis added). In 1975 the UN adopted the Universal Declaration of the Eradication of Hunger and Malnutrition, which proclaimed an‘inalienable right to be free from hunger and malnutrition’. This formalised what was already a de facto right to food during famines, resulting from drought, floods or war, and effected primarily through the WFP. But in practice, this implicit right to food was limited to famines in rural areas. No such right to food was extended to the chronically poor in urban areas. The underlying idea, rarely if ever made explicit, was that peasants(and the landless, rural poor) were being insured against social risks(drought, floods, war) that impaired their ability to support themselves. The discussion of a right to food was transformed in 1996, when the FAO organized a World Food Summit in Rome. The ensuing Rome Declaration on World Food Security reaffirmed‘the right of everyone to have access to safe and nutritious food, consistent with the right to adequate food and the fundamental right of everyone to be free from hunger’(quoted in Claeys, 2015b: 69). This led in turn, to the UN’s Committee on Economic, Social and Cultural Rights, which oversees the implementation of the ICESCR, issuing a‘General Comment’(no. 12) on‘The Right to Adequate Food’ in 1999. The Committee declared that: ‘The right to adequate food is realized when every man, woman and child, alone or in community with others, has physical and economic access at all times to adequate food or means for its procurement. The right to adequate food shall therefore not be interpreted in a narrow or restrictive sense which equates it with a minimum package of calories, proteins and other specific nutrients. The right to adequate food will have to be realized progressively. However, States have a core obligation to take the necessary action to mitigate and alleviate hunger even in times of natural or other disasters.’(UN, 1999: para 6) The Comment proceeded to set out various general obligations on states: crucially, to ‘respect’ the ways in which people access food, to‘protect’ people against intrusions on their ability to access food, and to‘fulfil’ the right to food in terms of strengthening people’s access to food generally and providing food directly when necessary(Claeys, 2015b: 6970). In 2009, the Committee adopted an Optional Protocol to the ICESCR making the right to food justiciable at the international level. In 2000, the UN appointed for the first time a Special Rapporteur on the right to food. The Special Rapporteur’s mandat e opened with the requirement‘to promote the full realization of the right to food and the adoption of measures at the national, regional and international levels for the realization of the right of everyone to adequate food and the fundamental right of everyone to be free from hunger so as to be able fully to develop and maintain their physical and mental capacities’. 56 ‘It soon became clear that little action was being taken with respect to the 1996 commitments’, wrote the Special Rapporteur later(Ziegl er et al., 2011: 6). The FAO hosted another World Food Summit in 2002. At the Summit, it was reported that the number of hungry people had hardly changed since 1996, and the 1996 goal of halving hunger by 2015 would not be met. An intense debate ensued, between governments, and 56 http://www.ohchr.org/EN/Issues/Food/Pages/Overview.aspx. The first UN Special Rapporteur on the Right to Food was Jean Ziegler(2000-08). He was succeeded by Olivier De Schutter(2008-14) and then Hilal Elver(2014-).

54 between them and civil society organisations, over both the conceptualization of the right to food(should it be a right to‘food security’ or even‘food sovereignty’?) and the obligations that would arise from the recognition of any right. Eventually, in 2004, the FAO drafted non-binding Right to Food Guidelines( ibid: 7-8). The Guidelines have been heralded by some( ibid) but criticized by others, primarily because the approach is seen as‘social democratic’, meaning that the state to required to mitigate inequalities whilst leaving the inequalities themselves largely untouched(Claeys, 2015b: 80). Whilst the UN and FAO were thinking about the right to food, peasants themselves in some parts of the global South were mobilising around the more radical demand for the right to food‘sovereignty’, and exerting pressure on the UN to recognise this as a peasant’s right. Pressure for the UN to adopt a declaration on the rights of peasants has been driven primarily by the transnational peas ants’ movement La Vía Campesina. La Vía Campesina was established in 1993, by movements of small farmers in various parts of Central America, North America and Europe, in response to the commodification of subsistence in the face of neoliberal capitalism. The name reflected the central importance of movements from Latin America(Desmarais, 2007; Martinez-Torres and Rosset, 2010). The nascent movement was invigorated by the 1996 Rome World Food Summit, which provided an opportunity for NGOs to confer(Claeys, 2015b: 13). As a reaction to‘neoliberalism’, the movement engaged with UN agencies with some skepticism. It asserted that organization on the ground protected peasants more than declarations by international agencies(Durnford, 2016: 66). La Vía Campesina called for ‘food sovereignty’, i.e. the right to produce food rather than a right to food, and insisted that this was a collective right, rather than an individual one, and was the right of‘peoples’ directly, rather than through the state(Claeys, 2012, 2015a, 2015b: 42-54). Durnford argues that the prospective right to food sovereignty challenges the West-centric character of most‘human rights’. In Durnford’s view, the elaboration or application of rights in the global South has generally entailed the ‘localisation’ or‘vernacularisation’ of rights originating in the global North. In contrast, the right to food sovereignty was formulated from below, in the global South(and more specifically Latin America), through La Vía Campesina. When La Vía Campesina adopted a declaration on the rights of peasants, in 2008, it soon attracted the attention of the Advisory Committee to the UN Human Rights Council. The Council noted the extreme vulnerability to poverty and hunger of‘smallholder farmers, landless workers, fisherfolk, hunters and gatherers’. At the request of the Human Rights Council, the Advisory Committee produced a report: ‘Today, 50 per cent of the world’s hungry are smallholder farmers who depend mainly or partly on agriculture for their livelihoods. Some 20 per cent of those suffering from hunger are landless families who survive as tenant farmers or poorly paid agricultural labourers who often have to migrate from one insecure, informal job to another. And 10 per cent of the world’s hungry live from traditional fishing, hunting and herding activities in rural communities. ’(UN Human Rights Council, 2012: 4) The Report identified and discussed five forms of human rights violation that contributed to this poverty:

55 (a) expropriation of land(including land-grabs), forced evictions and displacement; (b) gender discrimination; (c) the absence of agrarian reform and rural development policies: land reform dropped off the policy agenda in the 1980s, and support for small farmers (through credit, seeds, water, etc) declined with liberalisation etc (d) the lack of minimum wages and social protection; and (e) the criminalization of movements defending the rights of people working in rural areas.(ibid: 6) The Advisory Committee reported that the rights provided for by existing international conventions were poorly implemented and contained gaps, so that there was a need for a new international instrument. The Committee proposed a draft declaration. Itemised rights included: the rights to live in dignity and with an adequate standard of living; to land, seeds,‘traditional agricultural knowledge’ and other‘means of agricultural production’; to fair prices and payments; to information, freedom of association and justice; and so on ( ibid). In response, the UN Human Rights Council established in 2012 an Inter-Governmental Working Group(although there was some opposition and many members of the Council abstained). The Working Group strongly criticized La Vía Campesina’s draft Declaration. The(Bolivian) chairperson of the Working Group, in consultation with governments, civil society and others, drafted a new text in 2015. The revised text continued to focus on the right to food sovereignty, defined as the democratic and collective right of‘peoples’‘t o define their own food and agricultural system’, but some of the more radical elements in the original draft were diluted(Durnford, 2016: 68-9; see also Claeys, 2015b: 58-9). The revised text included a right to social security(Article 22): ‘Peasants a nd other people working in rural areas have the right to social security, including social insurance. … States shall recognize the rights of peasants and other people working in rural areas to social security, including social insurance, and, in accordance with national circumstances, should establish or maintain their social protection floors comprising basic social security guarantees. The guarantees should ensure at a minimum that, over the life cycle, all in need have access to essential health care and to basic income security, which together secure effective access to goods and services defined as necessary at the national level.’ The call for rights for peasants was a compromise between advocates of a right to food and champions of the right to food sovereignty(Claeys, 2015b: 108). It recognized that peasants had rights as producer-citizens, and not simply as consumer-citizens. The rise of the idea of peasants’ rights placed the ILO in an embarrassing position, caught between its progressive reformism and its historical commitment to workerism. In the face of criticisms that it had neglected peasants, the ILO pointed to a series of conventions, 57 and to its 2001 Safety and Health in Agriculture Recommendation(number 192). The 57 Specifically: The 1921 Right of Association(Agriculture) Convention(number 11), 1958 Plantations Convention (110), 1975 Rural Workers’ Organisations Convention(141), 1989 Indigenous and Tribal Peoples Convention(169) and 2001 Safety and Health in Agriculture Convention(184).

56 Recommendation included t he stipulation that,‘where economic, social and administrative conditions do not permit the inclusion of self-employed farmers and their families in a national or voluntary insurance scheme, measures should be taken by Members for their progressive coverage. This could be achieved by means of:(a) developing special insurance schemes or funds; or(b) adapting existing social security schemes’(Article 14). 58 In practice, this was largely empty rhetoric. Small farmers were sometimes important actors on the stage of welfare state-building in the global North in the early and mid-twentieth century. In the USA, as Prasad(2012) has argued, the political power of family farmers resulted in widespread state intervention in the economy, not through a European-style welfare state but through consumer protection, economic protection and progressive taxation. The reconstruction of a welfare state in the 1930s, through the New Deal, was achieved in part by mobilising agrarian discourses, images, policies and rights(Dauber, 2013). In Europe, the construction of social democratic welfare states in Scandinavia resulted from a coalition between the parties representing industrial workers and those representing small farmers. The political power of small farmers helps to explain both the initial emphasis on social assistance in the Nordic welfare states and their subsequent universal egalitarianism(Luebbert, 1984; Esping-Andersen, 1985). In Africa, peasant farmers played leading roles in many of the nationalist movements of the 1950s, but thereafter were generally marginalised politically in one-party states that sought to tax rural producers to finance development and were more anxious about the political threat posed by urban populations. The rural population often had the power to demand drought relief – and to reward governments that provided it(most notably in Botswana) – but peasants generally lacked both the political power (within the state) and economic power(in markets) of their Scandinavian or American counterparts in earlier decades. The result has been that international agencies such as the FAO have been caught up in the rising enthusiasm for social protection and for rights-based approaches generally but have not been under significant pressure from within Africa. The FAO has placed social protection at the centre of its strategic framework and endorsed a human rights-based approach to social protection. In its first Social Protection Framework(FAO, 2017), the FAO put forward a vision of social protection serving to ensure the food security of rural people and communities, protecting them before, during and after shocks and stresses, promoting resilient livelihoods and the sustainable management of eco-systems, contributing to pro-poor growth and inclusive rural development. The FAO explained its suppo rt primarily in terms of the‘economic case’ for social protection, i.e. the proven efficacy of social protection as an affordable investment in increasing production and improving food security and nutritional outcomes. But the FAO also restated its‘ commitment to promote the right to adequate food and social protection for all ’(19), citing international covenants(20). The FAO’s support for the principles of social inclusion and gender equality also led to the FAO calling for social protection to be‘transformative’ through empowering the poor and specifically women 58 See further‘Normative sources and rationale underlying the draft declaration on the rights of peasants and other people working in rural areas’, report A/HRC/WG.15/4/3 for 4 th session of the Working Group, May 2017.

57 (20-22; see also 32-3). The FAO committed itself to supporting national governments ‘ to ensure that normative frameworks and legislation promote the access to critical productive resources, services, organizations, markets, and decent employment to those living in rural areas, with a special focus on women ’(21) . Institutionally, the FAO’s Social Policies and Rural Institutions Division(ESP) includes teams working on the rights to food and social protection. 2.5. Gaps in the Global Discourse of Rights to Social Protection The global discourse of rights to social protection originated in the workerist agenda of the ILO. In the 1940s, this briefly blossomed into a broader concern with citizens ’ rights. From the late 1940s through to the 1990s, however, any talk of social citizenship was submerged under the tide of enthusiasm with development. Small groups of non-poor workers might have rights, but the urban and rural poor got development, if they were lucky. This changed dramatically from the end of the 1990s, with renewed interest in rights on the part of the ILO, DfIF UNICEF and other international organisations and aid donor agencies. Neither in the 1940s nor in the 2000s, however, did the predominant rights discourse provide for peasants. The informally-employed urban poor might get formalization, but the rural peasantry was almost entirely neglected by international organisations. It was only in the 2000s that organisations of peasants, based primarily in Latin America and Southeast Asia began to put peasants’ rights – and the right to food – on the global agenda. As we shall see in Chapter 3, little attention was paid to peasants in Africa. The global discourse of rights did not encompass explicit rights for everyone.

59 Chapter 3: The Vernacularisation of a Rights-Based Approach to Social Protection in Africa The‘global’ discourse of rights to social protection developed with, at most, a marginal contribution from African governments, civil society or individuals. As Ishay puts it, it is the‘Western concept of human rights that has prevailed’(2004: 7). Prior to the 1960s, Africans were generally represented in‘global’ fora(including the ILO and UN) by the colonial powers. The Universal Declaration of Human Rights was drafted primarily by Europeans and North Americans. Even the ICESCR, although passed in 1966, had been mostly drafted before most African colonies achieved independence and were admitted into the United Nations(see AnNa’im, 1990: 348 -51, cited in Merry, 1997: 35). This left the paternalistic task of defending colonized peoples to anthropologists. The American Anthropological Association objected to the Universal Declaration, asking how it could be ‘applicable to all human beings’ when it was‘a statement of rights conceived only in terms of the values prevalent in the countries of Western Europe and America’(Merry, 1997: 33-4). After independence, African countries participated in in the global fora where rights were discussed but generally failed to wield influence commensurate with their numbers. The global institutions continue to be dominated by players from the global North, and the global rights discourse remains one rooted in the history and culture of north-west Europe. Deliberation over social and economic rights within Africa has therefore generally entailed the‘localisation’ or‘vernacularisation’ of the‘global’ discourse, with modest modifications to take into account some of the attitudes and beliefs of African policy-making elites. Growing numbers of African scholars have criticized aspects of the global rights discourse, arguing that‘the individualist, narrow formulation of human rights is ill -suited to the African political and cultural unive rse’(Mutua, 2002: 68). But African regional fora have often been sponsored financially and‘supported’ technically by‘global’ – i.e. Northern-dominated – institutions, so that‘African’ declarations represent an uneasy combination of external priorities and Africans’ own preferences. The one dimension of the global rights discourse that was driven by activists in the global South – the call for the acknowledgement of the rights of peasants(or for‘food sovereignty’) – has received weak support from within Africa, despite the large numbers of peasants across the continent. The general criticism of the global rights discourse is that it has been overly individualistic and under-emphasizes duties. Mutua writes that: ‘… the current human rights movement mus t be understood as only a piece of the whole. Its roots in Western liberal thought and tradition necessarily deny it completeness, though not the universality of many of its ideals and norms. To paraphrase the famous metaphor, the gourd is only partially filled by the Western tradition; it falls on other traditions fully to fill it.’(Mutua, 2002: 70). For Mutua and others, the language of duty is integral to African conceptions of‘rights’. Individual rights not only coexisted with but were bound up with corporate responsibilities and obligations. Social and economic change may have encouraged individualism

60 (Howard, 1990), but communitarian ideals have been eroded and transformed rather than entirely displaced. This conception of rights and duties has major implications for social protection. This chapter first examines the vernacularisation of the global discourse of social and economic rights through the 1981 African Charter and subsequent‘Protocols’, and through the positions adopted by sub-regional organisations. The chapter then turns to the work of international organisations and NGOS in developing social and economic rights in the African context, focusing on efforts with regard to workers, peasants and specific vulnerable groups. Finally, the chapter examines the content of national constitutions across Africa. 3.1. The African Charter and subsequent Protocols The foundational document on rights in Africa was the African Charter on Human and Peoples’ Rights(or‘African Charter’), adopted by th e then Organisation of African Unity (OAU) in 1981 in Banjul(and hence sometimes called the Banjul Charter). The Charter was adopted‘partly due to external pressure on African governments to establish a regional human rights regime and partly as a response to the gross human rights violations committed by some African leaders’(Chirwa, 2005b: 323, emphasis added). Hitherto, African governments had tended to view human rights with suspicion, often justifying their ambivalence in terms of the priority or imperative of economic development (on which see Howard, 1983). With strong support from his President, Leopold Senghor, the Senegalese judge Kéba Mbaye led a team that drafted the Charter, and other African governments desisted from further obstructionism(Lindholt, 2001: 75-78; Baricako, 2008). The African Charter was‘distinctive in its attempt to append an African“fingerprint” on the human rights discourse’(Chirwa, 2005b: 323; see also Mutua, 2002). At the outset, Senghor had urged Mbaye’s team‘to dr aw inspiration from African traditions, bearing in mind the values of African civilization and the real needs of Africa’. He urged them to incorporate social, economic and cultural rights, including the right to development, and to recognize that rights must be balanced with duties(quoted in Baricako, 2008: 6). The African Charter did identify some social and economic rights, including rights to property (Article 14), work(Article 15), health(Article 16) and education(Article 17). It noted the specific importance of protecting the rights of women, children, the aged and the disabled (Article 18). Article 22 specified a general right to development: 22(1): All peoples shall have the right to their economic, social and cultural development, with due regard to their freedom and identity and in the equal enjoyment of the common heritage of mankind. 22(2): States shall have the duty, individually or collectively, to ensure the exercise of the right to development. This was a group right, of peoples, not of individuals. The Charter did not, however, mention the adequacy of the standard of living, food, social security or housing, all of which were included in the earlier International Covenant on Economic, Social and

61 Cultural Rights. The minimalist approach in the Charter reportedly reflected a concern not to overburden African states(Ssenyonjo, 2016: 92-3). There was, however, no explicit restriction of specified rights according to affordability, no reference to‘progressive realisation’ nor to any minimum set of‘core’ obligations on the part of either the state or any other group or institution. The ensuing Charter included an emphasis on values and the family and the inclusion of a section on‘Duties’. Under Article 17, the state was enjoined to promote a nd protect morals and traditional values. Article 18 identified the family as‘the natural unit and basis of society’, and enjoined the state‘to assist the family which is the custodian of morals and traditional values recognized by the community’. Article 27 recorded that‘every individual shall have duties towards his[sic] family’ as well as to society and the state. Article 29 spelt out individuals’ duty‘to preserve the harmonious development of the family and to work for the cohesion and respect of t he family’ as well as‘to respect his [sic] parents at all times[and] to maintain them in case of need’. Each individual must also‘serve his[sic] national community by placing his physical and intellectual abilities at its service’ and‘work to the best of his[sic] abilities and competence, and to pay taxes imposed by law in the interest of the society’. The individual must also‘preserve and strengthen positive African cultural values’. As Chirwa writes,‘these provisions represent an African conception of human rights, which regards rights and duties as inseparably linked’(2005b: 327). The rights of specific categories of citizens(or‘vulnerable groups’) – children, women and the elderly – were revisited later, the first through a dedicated African Children’s Charter and the second and third through‘protocols’ to the existing African Charter. The 1990 African Children’s Charter did not refer to social protection explicitly and placed the primary responsibility for children’s welfare on their parents , but it did enjoin states to provide‘material assistance and support programmes particularly with regard to nutrition, health, education, clothing and housing’(Sloth Nielsen, 2016: 159). The Children’s Charter was‘Africanised’ in one respect: It includ ed a section dedicated to the responsibilities of children that borrowed entire clauses from the African Charter.‘E very child shall have responsibilities towards his[sic] family and society, the State and other legally recognized communities and the inte rnational community’. Specific duties were listed, including‘(a) to work for the cohesion of the family, to respect his[sic] parents, superiors and elders at all times and to assist them in case of need;(b) to serve his[sic] national community by placi ng his[sic] physical and intellectual abilities at its service; … and(d) to preserve and strengthen African cultural values …’ Children were, however, explicitly spared the duties of working and paying taxes(OAU, 1990). The 2003 Protocol to the African Charter on the Rights of Women in Africa(‘African Women’s Protocol’ or‘Maputo Protocol’) entailed, first and foremost, a strong statement against discrimination against women(AU, 2003). The Protocol went beyond this with respect to, especially, reproductive rights(Article 14). The Protocol was cautious on ‘economic and social welfare rights’, however, focusing primarily on discrimination against women in employment. It did emphasise that child-raising is the responsibility of both parents(Article 13(l)). The Protocol went further than the African Charter on social and

62 economic rights in only three respects. First, it called on states to‘create conditions to promote and support the occupations and economic activities of women, in particular, within the informal sector’(Article 13(e)) and to‘establish a system of protection and social insurance for women working in the informal sector and sensitise them to adhere to it’(Article 13(f)). Secondly, it also specified a right to food security, including access to the land needed to produce food(Article 15). Thirdly, it did not include any praise for African culture or values. On the contrary, the Protocol called for‘the elimination of harmful practices’(including genital mutilation)(Article 5) and for a‘positive cultural context’(Article 17). In contras t to the African Charter, the Protocol barely mentioned the family, never mentioned duties, and mentioned responsibilities only in the context of children responsibility of both parents. The Women’s Protocol did not have the conservative features of the Af rican Charter and Children’s Charter. Later still, in 2016, the AU adopted a Protocol on the Rights of Older Persons. Article 7 (on social protection) pointed to the need for both contributory and non-contributory (universal) programmes to ensure income security. This seems to have been the first explicit reference to noncontributory programmes. It marked a shift from the Women’s Protocol, which had only mentioned contributory programmes. Article 10 stipulated that states should encourage families to support their elders, whilst also ensuring that older people have the option of residential care(Article 11). The Protocol was more respectful of existing culture than the Women’s Protocol had been. It referred to‘ the virtues of African traditions, values and practices which should inspire and characterize the provision of mutual social and communal care and support; respect for older members of society and the passing of knowledge to younger population groups ’. Article 8 did, however, call on states to‘t ake all necessary measures to eliminate harmful traditional practices including witchcraft accusations, which affect the welfare, health, life and dignity of Older Persons, particularly Older women ’ . Like the African Charter, and unlike the Women’s Protocol , it discussed the duties or responsibilities of older people, primarily to mentor younger generations(Article 20). The African Women’s Protocol and Protocol on the Rights of Older Persons were produced after the OAU had been replaced by the African Union(AU). The Constitutive Act of the African Union(signed in Lome, Togo, in 2000) neither mentioned poverty nor referred to social and economic rights. Thereafter, the AU did take social and economic rights somewhat more seriously than its predecessor, including an explicit reference to social protection in the 2016 Protocol on the Rights of Older Persons. The‘Africanisation’ of the global discourse on social and economic rights remained, however, both complex and partial. It was complex in that the veneration of the family that was evident in the African Charter and Children’s Charter could not be carried over to the African Women’s Protocol, and sat uneasily with social realities in the Protocol on the Rights of Older Persons. It was partial in that the Protocols focused on categories of people deemed vulnerable globally: children, women and the elderly. There was no discussion of the rights of peasants(or small pastoralists), or the rural and urban poor, i.e. the people most vulnerable across most of Afr ica. The Women’s Protocol denounced discrimination against women as women, but not against women as peasants. There does not appear to be any existing literature on precisely why some themes were included and others not

63 in the Charter and following Protocols, nor what role was played by personnel from(or supported by) Northern-based agencies. The Charter and Protocols have not generated much telling jurisprudence. The African Commission on Human and Peoples’ Rights(‘African Commission’) was charged with applying the African Charter, but there has been little jurisprudence over social and economic rights(or duties). In 2005, C hirwa concluded that‘it remains unclear as to when a State will be held to be in violation of its positive obligations in relation to socio-economic rights under the African Charter’(2005b: 326). Although there has been some additional jurisprudence since then(see Ssenyonjo, 2016), these have barely advanced any substantive social or economic rights. Indeed, Chirwa suggests, even the absence of any explicit qualification with regard to affordability is unlikely to be of much consequence, in that other courts are likely to follow the cautious lead of the South African courts. Many of the bolder provisions in the Charters and Protocols are unlikely to get far in the face of resistance. Brown and Oder suggest that the limitations of the African Women’s Proto col lie less‘in the substance or procedures of the Protocol’ as‘in the political, social and economic climate within which it operates and ultimately in the limitations of the law to challenge deeply embedded social, cultural and political norms’(Brown and Oder, 2016: 151). Overall, it is not clear how important the Charter and three Protocols are with respect to the claims to social protection by either individual people or entire categories of people (or vulnerable groups). The Protocol on the Rights of Older Persons and also the Children’s Charter state the need to treat older people and children with respect, recognizing their dignity. In short, the elderly and children should be considered as deserving. The African Women’s Protocol denounces discri mination. None of the documents provides a clear or strong statement of what precisely people deserve, or what non-discrimination means in practice. Moreover, the documents are silent on the social and economic rights that people – including women – have as peasants or as the urban or rural poor. In terms of both design and implementation, these documents have glaring gaps. 3.2. The AU, External Actors and Social Protection In addition to its general Charters and Protocols, the AU has also deliberated over social protection, including rights to social protection, through other regular and ad hoc structures. These deliberations did result in statements of commitment to social protection, especially the 2008 Social Policy Framework for Africa and the 2015 ‘Addis Ababa Declaration’(on Social Protection for Social Development). There was also discussion of an additional Protocol(to the African Charter) on the Rights of Citizens to Social Protection and Social Security. The meetings provided an opportunity for international organisations and local NGOs to keep questions of rights on policymakers’ agenda. Given that much of the initiative came from international organisations and allied local NGOs, it is not surprising that deliberation in these meetings has focused on the groups of primary concern to those organisations, i.e. workers and vulnerable groups.

64 At least five series of meetings can be identified:  Between 2003 and 2013 probably the most important of these were the meetings of the AU’s Labour and Social Affairs Commission. These meeting typically comprised‘ expert sessions ’ followed by sessions involving ministers of labour and social security. The Labour and Social Affairs Commission had a tripartite structure, with representatives of the‘social partners’(business and trade unions) participating in Commission meetings. It was supported(financially and logistically) by the ILO. The Commission’s first ordinary session was held in Mauritius, in 2003, and the ninth and final session was held in Addis Ababa in April 2013. 59  In 2006, HelpAge International organized, for the AU, an Intergovernmental Conference on Social Protection in Livingstone(Zambia) and a second conference later that year in Yaoundé(Cameroon), and followed these up with a ‘Livingstone 2 Process’ in 2008; the perceived need for these was presumably linked to discontent with the Labour and Social Affairs Commission.  Apparently in response to the call made in Livingstone, and presumably also reflecting discontent with the Labour and Social Affairs Commission, the AU began to hold separate‘ Conferences of the Ministers in charge of Social Development of AU Member States ’(‘CAMSDs’) every two years. The first of these, in Windhoek in late 2008, adopted the Social Policy Framework for Africa. Three subsequent CAMSDs were held in Khartoum in November 2010, Addis Ababa in November 2012, and Addis Ababa again in May 2014.  In 2011 the AU decided to replace the inter-ministerial meetings with a new system of fourteen‘Specialised Technical Committee’(STCs), most of which were to meet every second year. 60 The new STC for Social Development, Labour and Employment(STC-SDLE) replaced both the Labour and Social Affairs Commission and the parallel CAMSDs. STC-SDLE meetings involved separate meetings of experts(or senior officials) in the‘Social Development’ and‘Labour and Employment’ sectors, followed by ministerial meetings involving both ministers of labour and ministers of social development. The labour and employment stream remained tripartite. The first ordinary session of the STC-SDLE was held in Addis Ababa in April 2015 61 and a second session was held in Algiers in April 2017. 62 The next session is scheduled to be held in Guinea in 2019. The STC-SDLE discussed a Protocol(to the African Charter) on the Rights of Citizens to Social Protection and Social Security.  Finally, the AU has a Committee of Experts on the Rights and Welfare of the Child , which monitors compliance with the African Children’s Charter. Not only did this meet regularly, but it also organized‘expert consultations’ which provided an opportunity for civil society activists to promote a rights-based approach. 59 I cannot trace the venue or date of the second and fourth sessions. The third session was held in Johannesburg, in April 2005. The fifth, sixth and seventh were all held in Addis Ababa, in April 2007, April 2008 and September/October 2009 respectively. The eighth ordinary session was held in Yaounde in April 2011, and the ninth in Addis in 2013. 60 https://au.int/en/organs/stc. 61 https://au.int/en/newsevents/28072/first-session-specialised-technical-committee-social-development-labour-and. 62 https://au.int/en/newsevents/20170424/second-session-specialised-technical-committee-social-developmentlabour.

65 These processes are widely criticized, in private, by both government officials and participants from NGOs. Critics describe meetings as boondoggles, i.e. as opportunities for delegates to enjoy staying in plush hotels whilst investing little or no effort in discussions over social protection. It is not clear how often meetings do more than ratify documents that had been drafted beforehand, whether in the AU Commission or by consultants. Nonetheless, these meetings have been important for the Africanisation of the global discourse of rights to social protection in at least two respects. First, they maintained some momentum at a time when the AU itself seemed to have dragged its heels on social protection, i.e. in the early and mid-2000s. Crucially, in 2006-08, pressure from both international organisations and local NGOs pushed the AU into action. Second, the meetings pushed the AU to clarify why it was hesitant about embracing a fully rightsbased approach to social protection, and how any such approach could be combined with developmental arguments and an emphasis on the family. The history of the AU’s‘social policy framework’ seems to have begun in the first session of the AU’s Labour and Social Affairs Commission, in Mauritius in 2003. The LASC s eems to have been established to take up the concern that the poverty reduction strategy papers developed in association with the World Bank and IMF had not gone far enough. This reflected the ILO’s concern to reframe the global agenda on social protection . Having been established, the new LASC clearly needed a policy framework to frame its work, and to justify its existence. Meanwhile, AU ministers meeting in Ouagadougou adopted a Declaration on Employment and Poverty Alleviation in Africa and accompanying Plan of Action that called on governments to improve and strengthen social protection schemes. The Ouagadougou Declaration and Plan of Action emphasized the extendion of contributory schemes to any workers and their families who were currently excluded(AU, 2004), i.e. it was a workerist position clearly and explicitly linked to the ILO’s agenda at the time(see further below). A draft‘Social Policy Framework for Africa’ was prepared by a team led by the Labour and Social Affairs Commission secretariat, and discussed at the third session of the Labour and Social Affairs Commission in 2005(Wright and Noble, 2010: 112). The draft focused on social exclusion and integration, and was in parts very critical of the OAU/AU and national governments. It mentioned social protection without elaboration(AU, 2005). In 2008, a very different document, prepared by consultants at the Human Sciences Research Council in South Africa, was presented to the first CAMSD in Windhoek. The stated objective was‘ to provide an overarching policy structure to assist AU Member States in the development of their national social policies to promote human empowerment and development in their ongoing quest to address the multiple social issues facing their societies’(AU, 2008). Wright and Noble(2010) report that the draft presented in Namibia omitted any provision for social protection. After discussion in the Windhoek meeting, a section on social protection(section 2.2.3) was inserted. The Framework emphasised that social policy was necessary for inclusive development and growth but must also be consistent with development. The first of the guiding principles listed in the Framework declared that‘ Social policies must encapsulate the

66 principles of human rights, development imperatives and be embedded in the African culture of solidarity ’(AU, 2008: para 16). The section on social protection was mostly non committal. Earlier, the Framework seemed to endorse old-age pensions(in order to discourage large families). Section 2.2.3 stipulated that: ‘Member States are encouraged to choose the coverage extension strategy and combination of tools most appropriate to their circumstances. There is an emerging consensus that a minimum package of essential social protection should cover: essential health care, and benefits for children, informal workers, the unemployed, older persons and persons with disabilities. This minimum package provides the platform for broadening and extending social protection as more fiscal space is created.’( ibid: section 2.2.3) The insistence on a minimum package marked a potentially important shift in rhetoric. The overall formulation encompassed the concerns of external actors such as HelpAge International(with pensions for the elderly) and UNICEF(concerned increasingly with provision for children). It was most similar to the position of the ILO, with provision for each stage of a worker’s life, from childhood to old age. The minimum package idea retained the Eurocentric combination of workerist social insurance with pro-poor(or citizenorientist) social assistance. The scope of the‘minimum package’ would be appropriate in a European context, with a focus on protecting workers when they are unable to work because of age, unemployment or disability. It extended ou t to‘informal workers’ but did not seem to envisage provision for poor peasants or the self -employed urban poor. At the same time, the Framework made no other mention of either families or duties, suggesting a discontinuity from the African Charter and Ch ildren’s Charter. What was entailed in the‘African culture of solidarity’ was not explained. Whilst the Social Policy Framework was widely lauded outside of Africa(Wright and Noble, 2010: 112-13), it was also criticised. Wright and Noble noted that there was ‘insufficient acknowledgement of the ideological values that underpin social policy choices made, and of the potential tradeoffs that follow’ – although they did not spell out these diverse values( ibid: 113). Olivier(2009, cited in Wright and Noble, 2010) criticised the Framework on more practical grounds, for failing to think through the challenges of social security in contexts where the informal sector is large and almost entirely excluded from existing social security schemes. It is not clear who wrote – and who pushed for the inclusion of – Section 2.2.3 in the Social Policy Framework, but its insertion marked the culmination of a two-year campaign to raise the profile of social protection in Africa. In 2006, a bolder vision of social protection had been set out in the Livingstone Call for Action, agreed at a three-day Intergovernmental Conference on Social Protection that was attended by representatives of thirteen African governments. 63 In his opening speech, Zambian President Levy Mwanawasa pointed to the urgency of considering cash transfer and related programmes to combat poverty, given the weakening of the‘extended family system’. Mwanawasa referred to the‘success story that is emerging from the pilot cash transfer scheme being 63 Ethiopia, Kenya, Lesotho, Madagascar, Malawi, Mozambique, Namibia, Rwanda, South Africa, Tanzania, Uganda, Zambia and Zimbabwe. Representatives from Brazil, EU development partners, UN agencies and NGOs also attended (Schubert and Beales, 2006: 6).

67 implemented in Kalomo District.’ As Mwanawasa reminded his audience, the conference programme included a field visit to Kalomo(which was not far from Livingstone) so that participants could see for themselves. 64 The(brief) Call for Action agreed at the conference urged governments to expand cash transfers, including especially social pensions and transfers for vulnerable children and other households. Governments should cost social protection proposals and integrate these into national development plans. Donors should commit funds over the long term, and the AU should convene regular conferences(AU, 2006a). The Livingstone Call for Action reveals the difficulties in interpreting international agreements. The conference was organized in the name of the AU and Government of Zambia, but, as Leutelt writes bluntly, it‘was initiated, organized and managed by HAI [HelpAge International] and funded by DFID’(Leutelt, 2012: 17; 2016: 128). Although it remains unclear how precisely HelpAge succeeded in persuading the AU and Government of Zambia to embrace the Livingstone conference, Leutelt reports that HelpAge had previously(from 1999 to 2003) worked closely with the OAU/AU over both the Policy Framework and a Plan of Action on Aging in Africa(2016: 126). In addition, Mwanawasa’s speech was written by a Zambian technocrat, and went far beyond Zambian Government policy. Whilst very enthusiastic about social protection, even this speech downplayed the importance of rights. Five reasons were given for reforms: Social protection, Mwanawasa said, was good for economic growth, poverty-reduction, local economic development, equity and social cohesion, and(fifth and last) the fulfilment of human rights. In the conclusion to his speech, Mwanawasa emphasized not rights but responsibility:‘We have a responsibility to ensure that the weak and poor in our societies are protected.’ The official report of the conference, written for HelpAge International by Bernd Schubert(the German champion of cash transfers in Zambia and elsewhere) and Sylvia Beales(from HelpAge in London), paid more attention to rights, but also emphasized the practical benefits. Later in 2006, HelpAge organized – in the name of the AU and the Government of Cameroon – a followup‘International Workshop on Ageing’ in Yaoundé. This was attended by officials from eleven African countries(Leutelt, 2016: 130-1). 65 The Yaoundé Call for Action mentioned in passing the rights of older people, but emphasised instead their needs(and neglect). It included a call for universal social pensions(AU, 2006b). HelpAge(with further funding from DfID) followed up on this in 2008 with what it called the‘Livingstone 2 Process’ on the theme‘Investing in Social Protection in Africa’ : A series of national‘consultations’ 66 (in March 2008) and three ‘Regional Experts Group Meetings on Social Protection’. 67 The purpose was, in part, to mobilise support in civil society to exert pressure on national governments(and the AU) to raise social protection up the policy agenda. This is likely to have been influenced by the‘Drivers of Change’ 64 ‘ Speech delivered by His Excellency the President of the Republic of Zambia Mr Levy Patrick Mwanawasa, SC, on the official opening of the Inter-Governmental Conference on Social Protection held at Fairmount Hotel, Livingstone, on 21 st March 2006 ’ . 65 Uganda, Kenya, Tanzania, Ethiopia, Mozambique, South Africa, Sudan, Ghana, Lesotho, Cameroon, United Kingdom and Zimbabwe(AU, 2006b). 66 In March 2008, in Burkina Faso, Rwanda, Mozambique, Sierra Leone, Cameroon and Tunisia. 67 In June 2008, in Uganda, Egypt and Senegal.

68 programme, funded and driven by DfID, that identified civil society as a prospectively key actor in pushing conservative government departments towards policy reform(see e.g. Scott, 2007). The HelpAge initiative did lead directly to the establishment of a network of civil society organisations concerned with social protection – the African Platform for Social Protection(see below) – and to a set of recommendations presented to the AU Ministers in Charge of Social Development in Windhoek(in October 2008, see above). 68 HelpAge organised the Livingstone conference and‘Livingstone 2 process’ in order to reshape the regional agenda in Africa, through applying global undertakings in African countries(Leutelt, 2012: 18; 2016: 129). The success of this strategy remains unclear. Undertakings were included in formal declarations, but among what actors, and where, did the commitment extend beyond rhetoric? We are not aware of any public evaluation of the Livingstone processes. The Livingstone processes may appear to be successful interventions by a Northern-based NGO, with Northern experts working with allied local technocrats. But we should be cautious before we conclude that the processes generated additional enthusiasm for social pensions specifically or social protection more broadly. In Zambia itself, the enthusiasm for cash transfers – and the Kalomo pilot programme in particular – expressed in the speech read at Livingstone by Zambian President Mwanawasa was not widely shared within his government, and did not reflect government policy. Mwanawasa’s government stalled on reforms for several years(Kabandula and Seekings, 2016; see further below). More generally, there is little evidence of impact in some of the other countries that participated in the Livingstone conference, including Tanzania, Namibia, Zimbabwe and Mozambique, in addition to Zambia itself. Following its inaugural session in Windhoek, in 2008, the CAMSD seemed to avoid bold moves, and was preoccupied largely with provision for the disabled(drafting a protocol on their rights) and the elderly – both of which were groups of special concern to an array of international organisations and NGOs. A series of meetings were held with HelpAge International on establishing institutions to monitor the implementation of the Protocol on the Rights of Older Persons. It was not until the first STC-SDLE meeting, in 2015, that the AU again stepped back to consider social protection more broadly. The theme of the first STCSDLE meeting was‘Social Protection for Inclusive Development’. The preparatory documents emphasized that social protection was an investment in future development, with muted references to rights. The STCSDLE did, however, agree on an‘Addis Ababa Declaration’(on Social Protection for Social Development) that called on signatories to ‘Fully pursue the rights -based approach to social protection and social security for our citizens, aiming at inclusive development that leaves no one behind, through appropriate legal and policy frameworks supplementing the AU Charter on Human and People’s Rights’. The meeting also called on the AU to prepare a Protocol(to the African Charter) on‘the Rights of Citizens to Social Protection and Social Security’(AU, 2 015a). The proposed Protocol was discussed further at an Expert Roundtable convened by the AU in Gaborone in mid-2016. At the second STC-SDLE meeting, in 2017, it was reported that the Protocol was being considered within the higher decision-making structures of the AU. As of early 2018, no draft Protocol appears to be available online. 68 ‘Summary Recommendations from L2 process‘Investing in social protection in Africa’ to be presented to AU ministerial conference on social development, 27-31 October 2008 Windhoek Namibia ’.

69 Rights to social protection have also been discussed within the AU’s African Committee of Experts on the Rights and Welfare of the Child, which monitors compliance with the African Children’s Charter. This Committee also convenes‘expert consultation meetings’, including, in April 2014 in Cape Town, a meeting on Children and Social Protection Systems: Building the African Agenda. This activity seems to have been sponsored by UNICEF. The meeting, in Cape Town, was attended by representatives from 38 AU Member States and included senior officials from governments at Permanent Secretary and Director-General levels, directors of budget offices, line ministry representatives, national, regional and international social protection experts, and civil society organisations. The meeting recommended‘a human -rights based approach, in particular, children’s rights, to social protection’, with‘social protection goals and targets … define d and embedded in national development plans and policies, in medium and long-term expenditure frameworks, and … protected by legal instruments’; called on govts‘to fulfil their commitments and obligations’ under international and AU agreements(AU, 2014a : 1). The meeting called on the AU Commission to‘become a more active role player in the social protection arena and prioritise social protection for children’( ibid: 3). Whilst children’s rights were emphasized in meetings of experts on children in part icular, the AU in general emphasized more strongly the importance of the family. At the 2010 CAMSD2, ministers agreed a Khartoum Declaration which reaffirmed( inter alia) their commitment to‘the acceleration of implementation of relevant social protection measures to directly benefit the wellbeing of the Family in Africa’, with an emphasis on persons with disabilities, children and the elderly. In 2012, CAMSD3 discussed and endorsed a draft AU‘Common Position on the Family’. The document acknowledged tha t social, economic and demographic changes had undermined the historic role played by the family in providing‘ a sophisticated social security system, as well as support and guidance mechanisms especially in times of need and crisis such as during unemployment, sickness, bereavement, and old age ’. The AU – and CAMSD – called for renewed attention to the family through family-centered policies. The theme for the following CAMSD(in 2014) was‘Strengthening the African Family for Inclusive Development in Afri ca’. CAMSD4 was concerned primarily with child marriage, but it did also agree on a ‘Plan of Action on the Family in Africa’. The preparatory documents for the conference emphasized that the family had been neglected in discussions of poverty reduction and development.‘ Family oriented policies that target the family as a unit and its dynamics as a whole’ were said to facilitate inclusive development better than policies that focused on individuals(AU, 2014: para 6). The conference would provide an opportunity to consider how child protection and social protection policies could be used to strengthen social integration and intergenerational solidarity within families( ibid: para 9). The following year, the STCSDLE noted in the preamble to its‘Addis Ababa Declaration’ that‘the family is the fundamental unit of society and has multiple roles and functions, especially in Africa where it is central, unique and indispensable in support of all its members, hence needs all requisite support to be strengthened further’(AU, 2015a]. This emphasis on the family seems to have represented a backlash against the implication of individual rights in the Children’s Charter and Protocols.

70 Pushed by international organisations and NGOs, the AU also considered further the rights and position of informal workers. Under pressure from the ILO, the AU adopted a Programme of Upgrading the Informal Economy(in 2010) and then a Social Protection Plan for the Informal Economy and Rural Workers(SPIREWORK)(adopted at Yaoundé in 2011). SPIREWORK represented a compromise between the ILO, which pushed the idea that the the formalisation of the informal sector(as part of its broader strategy focused on‘decent work’) was an integral to solving the challenge of poverty in Africa, and the AU(and individual African countries) which pushed for employment creation( ‘job rich and inclusive growth’), especially for young people . SPIREWORK pointed vaguely to community-based insurance schemes. It acknowledged the possibility of non-contributory programmes, but without any evident enthusiasm. The AU began to refer to a target of extending coverage to 20 percent of the informal sector, although it was never clear whether the informal sector here meant informally-employed wage workers or included also self-employed peasants, urban traders, and so on. In any case, this target was not only modest but was also unrealistic without large-scale non-contributory programmes. In December 2016, the AU organized a SPIREWORK expert meeting in Lilongwe (Malawi), with funding from the European Union. The purpose for the meeting was reportedly to share findings of an AUFAO study conducted in Cameroon, Cote d’Ivoire, Kenya and Malawi on the role of cooperatives in providing social protection to rural workers. According to one online report of the meeting: ‘While the need for SPIREWORK has already been recognized in strategic AU documents and resolutions, it would seem that the study presented in Lilongwe together with the Finnish-funded SPIREWORK project designed for Zambia are today the only concrete development projects focusing on SPIREWORK in Africa to date.’ 69 SPIREWORK is discussed further below in the case-study of Zambia(Chapter 4.3). Overall, it is not clear what, in practice, SPIREWORK achieved. In the 2000s the AU was jostled into assertions of the social and economic rights of selected categories of vulnerable people, and social protection has been identified as a key mechanism in this. The 2006 Livingstone Declaration and Yaoundé Call for Action, 2008 Social Policy Framework and 2015 Addis Ababa Declaration all stressed the need for social protection to realise social and economic rights. The AU seems to have been jostled into these assertions in part by external pressure(notably from HelpAge and the ILO), and in part by pressure from advocates of reform within Africa(including NGOs) and within the AU itself. The AU’s commitment has, however, remained less than emphatic. The AU condemns clearly discrimination against women and affirms the (general) rights of children and the elderly but holds onto an understanding of social protection that assumes that populations have been proletarianized, with social protection policies filling the gaps when people cannot support themselves through formal or informal employment. There has been little consideration of the rights of either peasants or the urban self-employed. In this sense, the global approach to social and economic 69 https://www.thl.fi/en/web/thlfi-en/research-and-expertwork/projects-and-programmes/eu-social-protectionsystems-programme-eu-sps/what-s-new/the-african-union-expert-meeting-on-social-protection-for-informal-andrural-economy-workers-spirework-.

71 rights has been insufficiently Africanised, resulting in large gaps in the rhetorical reach of formal statements on social protection. In a second sense, the global approach has been Africanised in distinctive ways: Statements about rights or policy frameworks include assertions of responsibilities alongside rights, and of the importance of family over-andabove individuals. 3.3. Sub-regional organisations Within Africa, sub-regional organisations have also turned to social protection. The two primary sub-regional organisations have been the Southern African Development Community(SADC) and the Economic Community of West African States(ECOWAS). Other regional organisations include the West African Health Organisation, with a specific interest in health, and the apparently moribund Economic Community of Central African States(ECAS). ECOWAS has been largely concerned with trade and economic integration, and more recently with political stability; it does not seem to have a track record of interest in labour market and social policy. SADC, in contrast, has an interest in social policy, perhaps because of the long history of extensive cross-national migration within what has become a very regional economy. SADC’s first statement on social and economic rights was unambiguously workerist, showing the fingerprints of the ILO. The 2003 Charter of Fundamental Social Rights in SADC emphasised the importance of regional harmonisation of labour market and social security policies(Article 2) through a tripartite frame. Article 3 declared that the Charter embodied‘the recognition by governments, employers and workers in the Region of the universality and indivisibility o f basic human rights’ proclaimed in a set of international instruments including the Constitution of the ILO and the ILO’s Philadelphia Declaration. The Charter referred repeatedly to workers(as in Article 8, which called for social pensions for any retired workers who cannot meet their basic needs through other (presumably insurance) schemes). The core article on social protection(Article 10) required that SADC governments ‘create an enabling environment so that every worker in the Region shall have a right to adequate social protection and shall, regardless of status and the type of employment, enjoy adequate social security benefits’, whilst ‘ persons who have been unable to either enter or re-enter the labour market and have no means of subsistence shall be entitled to receive sufficient resources and social assistance’(SADC, 2003, emphasis added). The Charter focused on the labour market. Three years later, SADC agreed on a Code on Social Security in the SADC(SADC, 2007). Article 4 boldly affirmed th at‘Everyone in SADC has the right to social security’. It called on governments to‘establish and maintain a system of social security in accordance with the provisions of this Code and Article 10 of the Charter of Fundamental Social Rights in SADC’, to‘ maintain its social security system at a satisfactory level at least equal to that required for ratification of International Labour Organization(ILO) Convention Concerning Minimum Standards of Social Security No. 102 of 1952’ and to‘progressively raise its system of social security to a higher level, which should include achieving the meaningful coverage of everyone under the system, bearing in mind the realities and level of

72 development in the particular Member State’. Article 5.1 affirmed that:‘ Everyone in SADC who has insufficient means of subsistence to support themselves and their dependants should be entitled to social assistance, in accordance with the level of socio-economic development of the particular Member State. ’ The following Article(6.1) affirmed that: ‘ Each Member State should establish social insurance schemes and should progressively expand the coverage and impact of these schemes ’, with Article 6.4 requiring that ‘ Member States should extend social insurance coverage to the entire working population ’ and Article 6.5 referring specifically to the informal sector. Article 11 suggested that social assistance should be provided to the unemployed. In comparison with the 2003 Charter, the 2007 Code was much more inclusive, Although its focus remained strongly on social insurance, it did envisage extending coverage to people who were not formally-employed waged workers. Since 2007, however, there does not appear to have been much progress in terms of regional policy on social protection. Institutionally, the SADC Secretariat includes a Social and Human Development and Special Programmes Directorate, whose component units include an ‘Employment and Labour Sector programme’ as well as a‘special programme’ on ‘Orphans, Youth and Vulnerable Children’. The objectives of the Employment and Labour Sector programme include promoting social protection‘for workers and their families’, and the Special Programme on Orphans has completed a long set of policy documents setting out minimum levels of provision, including social protection. The rights of workers were further spelt out in a 2014 Employment and Labour Protocol. States,‘with due regard to the means avail able’, must ensure that: ‘(a) Every worker in the region and his or her dependents shall have a right to adequate social protection and shall, regardless of status and the kind of employment of the worker, enjoy adequate social security benefits; and (b) persons who are unable to enter or re-enter the labour market and have no means of subsistence shall be entitled to receive sufficient resources and social assistance.’(SADC, 2014: Article 11). The Protocol further enjoined SADC’s member states to ratify and implement the ILO’s 1952 Convention on minimum standards and to implement its 2012 Recommendation on social protection floors. It called on each member state to develop‘an integrated and comprehensive social protection system’ that provides‘meaningfu l coverage for everyone’. Whilst SADC held up the vision of universal coverage, no part of the SADC structure appears to have any clear responsibility for social protection for those people who are neither‘workers’ nor the elderly or other deserving poor. Whilst the SADC Secretariat also includes a Directorate on Agriculture and Food Security, SADC seems to view food security more as a national problem than as a problem for households or individuals. 70 It is unclear how SADC compares with ECOWAS, although it seems that the latter organization has paid more rhetorical attention to peasant farmers, perhaps because peasants’ organisations are more active in West Africa(through Réseau des 70 See http://www.sadc.int/themes/social-human-development/.

73 Organisations Paysannes et de Producteurs de l’Afrique de l’Ouest, ROPPA)(Claeys, 2015b: 35-6). Yeates(2017) points to the European Union as an example of how regional social policy can serve as a mechanism for promoting policy reform at the national level. With respect to social protection, regional organisations typ ically produce‘road maps’, or‘declarations of intent’. These can be important, symbolically and practically, acting as precursors of more substantive reform. In the case of the European Union, however, regional policy was soon backed by a degree of legal force, and was driven in part by strong and wellresourced regional institutions. SADC does not have comparable resources. It is not clear that it has significant influence on national policy-making. 2.4. Nonstate organisations and workers’ rights to s ocial protection For most of the past century the global discourse around rights to‘social protection’ was driven by the demand for employment-related social insurance. In the African context, until recently, this meant demands for social insurance for workers in the public sector and large private enterprises. Beginning in the 1950s, colonial and then post-colonial states saw social insurance as a key element in a Bismarckian strategy of co-opting or at least appeasing these workers. But the leading rol es in championing workers’ rights in Africa have been played by trade unions, the ILO and related NGOs. In the 2000s, these civil society organisations have turned their attention to the extension of social insurance to workers in small firms and especially in informal employment. This has entailed a partial vernacularisation of the global discourse of the rights of‘workers’ – partial, because the ILO and other organisations have not fully confronted the challenge of reaching the peasantry and other family-employed working people. The ILO has long been a major actor in the promotion of social and economic rights, both at the global level(see Chapter 2) and, in the 2000s, through shaping the policy agenda of both the AU and SADC(see Chapters 3.2 and 3.3). The ILO’s vision of rights has formally become an inclusive one, i.e. including not only all working people but also all citizens through social protection floors. In practice, the ILO frequently lapses into an assertion of the primacy of workers’ rights. The slippage was evident in, for example, a speech by the Deputy Regional Director of the ILO to the first AU STC-SDLE meeting, in Addis Ababa in 2015. The ILO official began by reaffirming the ILO’s commitment to helping African states‘to build effectiv e social protection systems as a means to eradicating extreme poverty in Africa’. She went onto applaud the AU’s SPIREWORK initiative(discussed above) and emphasized that the ILO saw social protection as a right. She then called specifically for a‘minimum package of social protection benefits’ to be made available‘to all workers including those in the informal economy’, without any acknowledgement that the category of‘workers’ in the informal economy in Africa poses considerable challenges to the ILO(AU, 2015a). The history of the ILO in Africa has long revolved around the tension between its primary goal of promoting the rights of workers in formal employment and the economic reality

74 that, across most of Africa, few workers in formal employment are relatively poor, and most enjoy clear privileges relative to peasants and the rural and urban poor. ILO conventions on the rights of workers have rarely seemed a priority in countries that have been, at least until recently, overwhelmingly agrarian. African states and business representatives have rarely played prominent roles in deliberations within the ILO, including through its annual International Labour Conferences. 71 The ILO has, nonetheless, remained active across much of Africa, trying to shape the rights agenda. Since the early 2000s the ILO has engaged in three closely related processes of standard-setting of relevance to social and economic rights. First, it began to promote‘decent work’. Second, it developed and championed the idea of a social protection floor, culminating in Recommendation 202 in 2012. Thirdly, it advocated the formalization of informal employment, including through Recommendation 204 in 2015. In all three respects, the ILO had been compelled to vernacularise its global discourse around workers’ rights. The only careful analysis of the ILO’s global standardsetting in these fields is Deacon’s close study of the process culminating in the ILO’s Recommendation 202 on social protection floors. Deacon does not analyse the role that African(or other) countries played during the phase of conceptual development, but he does comment on their role during deliberations within the 2011 and 2012 International Labour Conferences. In the 2011 Conference, most African countries were supportive of the general proposals. In 2012, during the decisive deliberations, they were more guarded, generally lining up with India and Europe in defending national(and perhaps even intra-national) variation and resisting the‘progressive’ amendments pushed by countries in Latin America(with support from many workers’ representatives)(Deacon, 2013: 87). A group of fifty African countries proposed an amendment that would qualify social protection floors‘in accordance with national context and considering the l evel of national economic development’. South Asian countries proposed a similar amendment. Neither amendment was accepted( ibid: 89). African countries reportedly also resisted the insertion of commitments to‘decent wages’ and‘the reduction of informality’( ibid: 94). Unsurprisingly, African states have been slow to ratify ILO conventions on social protection, as the ILO itself has acknowledged. As recently as 2010, the ILO’s flagship social security convention – the 1952 Social Security(Minimum Standards) Convention, number 102 – had been ratified by only five countries in Africa, none of which were Anglophone. 72 Neither Recommendation 202 nor Recommendation 204 were conventions, and therefore were not subject to ratification. In 2011, the AU explained the ambivalence of African countries towards Convention 102 in terms of the inappropriateness of contributory social insurance schemes in the African context, where 71 One conspicuous exception was South African trade unionist and later government minister, Ebrahim Patel, who played a major role in the ILO’s decent work agenda in the 2000s. Patel brought a distinctively South African perspective to debates. He cannot be considered as representative of the views of African representatives generally. 72 Democratic Republic of the Congo, Libya, Mauritania, Niger and Senegal(ILO, 2011: 22). All five countries accepted parts of but not the entire Convention. The Convention has subsequently been ratified(and accepted in part) by two more countries: Togo(2013) and Chad(2015).

75 most poor people were too poor to afford contributions and the state lacked the capacity to organize large-scale(or universal) social insurance(AU, 2011: para 18). 73 In practice, therefore, the engagement between the ILO and Africa has primarily taken the form of the ILO promoting its positions within Africa. The ILO has deployed considerable resources in this project, through its African Regional Meetings held every four years, 74 its offices(now in Abidjan for Africa as a whole, together with four regional and eight country offices), its sponsorship of and support for AU and SADC structures and activities, and focused research projects. The ILO’s own Regional Meetings have the standard tripartite structure, and the ILO encouraged the AU and SADC to structure some of their own institutions similarly, thereby ensuring that workers – especially workers in large firms – are prioritised. The ILO’s adoption of the decent work agenda in the early 2000s pushed it to consider actively ways of protecting people other than workers in formal employment. For the forseeable future, decent work could be a reality for only a small and non-poor section of the population. Alongside its advocacy of decent work, the ILO promoted social assistance programmes for the poor. In 2005 a team from the ILO’s Social Protection Department costed basic social protection packages for various African countries(Pal et al., 2005). The ILO’s annual African Regional Meeting in 2007 resolved that‘All African countries adopt coherent national social security strategies, including for the introduction or extension of a basic social security package that includes essential health care, maternity protection, child support for school-age children, disability protection and a minimum pension’. With funding from DfID, the ILO went on to conduct detailed country studies in Zambia and Tanzania(and Zanzibar). The ILO promoted decent work, with social protection tacked on, at different levels. The ILO initiated pilot country-level Decent Work Programmes in around 2006, before branching out morw widely from about 2009. At the regional level, the ILO organized tripartite‘Decent Work Symposia’ from 2009. At the second of these, in Yaoundé in October 2010, concrete proposals for the social protection floor were endorsed. The ‘Yaoundé Declaration’ ca lled on African governments and social partners(i.e. business and labour) to expand social protection both horizontally(through a social protection floor, providing not only for children and the elderly but also for anyone unable to support themselves through work) and vertically(through contributory schemes)(ILO, 2011: 1721). The ILO emphasized its recognition that‘ there is no single right model of social security ’: programmes must be‘ implemented and developed according to economic and social realities of each country ’( ibid: 4). The ILO also expended considerable energy on the AU’s(tripartite) LASC and subsequent STC -SDLE. It pushed the AU into adopting a Programme of Upgrading the Informal Economy in 2010 and SPIREWORK the following year. As we saw above, SPIREWORK – and the‘formalisation’ agenda more broadly – proved to be largely vacuous with respect to social protection. 73 It is telling that the 1952 ILO Convention was understood as referring only to social insurance. 74 The 9 th African Regional Meeting was held in Abidjan in 1999. The 10 th and 11 th were held in Addis Ababa in 2003 and 2007 respectively. The 12 th was held in Johannesburg in 2011, and the 13 th reverted to Addis, in late 2015. The 14 th is scheduled for 2019.

76 The ILO adopted in 2015 its Recommendation(no. 204) Concerning the Transition from the Informal to the Formal Economy(ILO, 2015; see above). The Recommendation focused, from the opening paragraph of its preamble, on the rights of workers. It included the usual caveat respecting national circumstances(para 7) 75 and pointed to the need for social protection floors(para 19), but the priority with respect to social protection was clearly the extension of‘social security’ through social insurance(para 20). In Africa, however, this was clearly not the case. Insofar as the formalization agenda focused on improving working conditions and raising productivity, then it was entirely relevant to Africa. Its relevance with respect to social protection was far less clear. This is because the character of informal employment in Africa was very different to its character in some other parts of the global South. In Africa, with the exception of South Africa, informal employment comprised mostly selfemployment(primarily on farms) and constituted about 90 percent of all employment. The ILO itself had reported data on this in a 2009 report(ILO, 2009). In Tanzania, in 2005, formal employment accounted for only 6 percent of total employment. Almost two-thirds of total employment comprised peasants working on their own land. Most of the rest comprised self-employment or unpaid family work. Non-family informal employment accounted for only about 3 percent of total employment( ibid: 10). Given that the formalization agenda focused primarily on formalizing non-family informal employment, i.e. 3 percent of all working people in the case of Tanzania, the scope for formalization to make a big difference to social protection is small in most African countries. ILO documents have generally acknowledged this implicitly rather than explicitly. For example, the 2009 ILO report on formalization in Africa did discuss briefly contributory microinsurance schemes(to cover health care costs) and Zambia’s national health insurance system, but it seemed to acknowledge that contributory schemes could not be extended very far. ‘Some countries, such as Tunisia, have engaged in reforms to extend statutory social insurance coverage. Currently, social insurance schemes, if they exist at all, usually cover only limited sections of the labour force in the formal economy. Legislators have frequently envisaged extending their cover at a later stage. In practice this extension, by bringing in successively smaller enterprises and/or including new categories of workers such as the self-employed, has not so far reached many more groups of the working population in Africa, with the exception of a few countries. Even small enterprises may present additional problems given their rudimentary accounts and arrangements for paying workers and their stronger tendency towards non-compliance. Self-employed workers are usually unwilling, and frequently unable, to pay the combined worker and employer contribution. As might be expected, schemes for self-employed workers are more successful when the government is willing to subsidize part of the premium.’(ILO, 2009: 42) 75 Although the only reference to full employment was in the context of urging states to make‘full, decent, productive and freelychosen employment a central goal in their national development and growth strategy or plan’. The only mention of job creation was in the context of‘the objective of quality job creation in the formal economy’(paragraph 14). The term‘labour intensive’ was never used.

77 The Report proceeded to discuss non-contributory cash transfers, citing ILO studies work of the affordability of a basic package of benefits. Following the adoption of the Yaoundé Declaration in 2010, the ILO drafted an Action Plan for the implementation of social protection floo rs. This Action Plan stated more boldly that‘t he experience of recent decades has shown that classical social insurance, while remaining an important tool for consolidating the formalization of a part of the workforce, has had limited effect when it comes to reaching out to groups in the informal economy’ (ILO, 2011: 7). The Action Plan pointed to microinsurance schemes, which were being considered by‘an i ncreasing number of African governments … in their strategy to extend social protection’, whilst some countries(including Ghana and Senegal) were already implementing them. But, the ILO warned,‘self -financed micro-insurance schemes have major limitations in terms of sustainability and outreach to large numbers of beneficiaries’( ibid: 8). Whilst formal social insurance and microinsurance schemes should form part of a country’s overall plan for social protection, the ILO recognised that tax-funded social assistance was likely to be far more important. The ILO’s emphasis on social assistance in Africa en tailed an important process of vernacularisation. The ILO also vernacularized its global discourse in terms of downplaying the discourse of rights in its promotion of reform. In a typical speech to the AU’s Conference of Ministers in Charge of Social Deve lopment, in 2014, the ILO’s Regional Director for Africa(Aeneas Chuma) concluded that there was a clear common vision of‘employment, social protection, rights and voice for all women and men’, but otherwise barely mentioned rights (Chuma, 2014). Funded by IrishAid, the ILO ran a project on Building National Floors of Social Protection in Southern Africa between 2014 and 2016. The project’s objective was to implement basic social protection guarantees in the three pilot countries: Zambia, Malawi and Mozambique. Output included a National Basic Social Security Strategy in Mozambique, proposals in Malawi, and policy documents in Zambia. The case of Zambia is discussed below(Chapter 4.3). The ILO was not alone in promoting the formalization of informal employment. WIEGO (Women in Informal Employment: Globalizing and Organizing) has a strong presence in Africa. Three out of five WIEGO programme directors are South African(as is the acting director of another programme). WIEGO’s board has members from both W est and East Africa. 76 Informal sector workers often not covered by some means-tested programmes because their incomes raised them out of extreme poverty. Whilst this was not a factor with respect to social pensions, it was relevant to the risks that befell working people. In addition, even public health care systems are often inaccessible to informal workers, because of their work commitments. WIEGO’s approach to social protection was initially set out in a paper by Lund and Srinivas(2000). Lund and Srinivas noted at the outset that ‘the informal economy is not 76 http://www.wiego.org/wiego/who-we-are.

78 a transitional phenomenon, but is here to stay: it will not disappear naturally as part of the development process’, therefore the task of‘addressing the dearth of social protection systems for w orkers in the informal economy cannot be postponed’(Lund and Srinivas, 2000: 1). They proceeded to pose the question:‘ Under what conditions can what kinds of workers in the informal economy(and especially poorer women) get access to what core measures of provision, which can be incrementally improved upon in the future? ’ ( ibid: 2). Lund and Srinivas emphasised that social protection needed to accommodate the heterogeneity of the informal economy, although in practice they seem to have limited their analysis to informal workers in the cash economy and did not consider subsistenceoriented peasant farmers. The core of the paper comprised a discussion of various cases of innovative social protection programmes. In terms of financing, one of these seemed to entail state-funded social insurance, one taxfinanced social assistance(South Africa’s old age pension programme), whilst the others relied on contributions from the members (sometimes with subsidies, including from donor aid). Lund and Srinivas made a point of emphasising that poor people were often willing to contribute to the costs of programmes that insured them against risks :‘ The voices from thousands of organizations are saying that very poor people can save, and do. And very poor people wish to contribute to social insurance, and do. This is a demand-driven need, around which people have successfully organized’( ibid: 29). The discussion of how the various schemes were financed made it clear, however, that programmes reaching large numbers of poor people depended heavily(or entirely) on external financing. The purpose of the initial WIEGO documents may have been to persuade the ILO to take informal sector workers more seriously. If so, they succeeded, at least in shifting the rhetoric of the ILO. Globally, the ILO embraced a formalization agenda, with the implication that contributory schemes would expand. In Africa, the ILO seemed to accept the limits to formalization, and instead looked primarily to non-contributory programmes. Perhaps in respons e to this, WIEGO tried to chart a‘middle path’ between formalization (and conventional social insurance) and the universalism of social assistance. In a later paper, Lund(2009) emphasized more strongly the need to ensure that the ‘owners of capital’ ass umed more responsibility for financing social protection for ‘workers’. Lund – who was at the time WIEGO’s Director for Social Protection – worried that both the World Bank and ILO had shifted too far away from an employment-linked approach to social protection. Lund asked whether too much emphasis was being placed on programmes funded out of general taxes. Her alternative seemed to be funding programmes out of profits earned in the value-chains in which workers worked, although she was unclear as to how this would be done, nor did she provide examples or models. Subsequent to this paper, WIEGO has thought further about the possibility of organizing groups of workers at sectoral or local levels so that they can negotiate more effectively that someone else pays at least part of the cost of protecting them against risks. For example, street vendors might negotiate better terms with wholesalers, construction workers might negotiate with employers to insure them against injury, and social protection for fishermen might be funded through a dedicated tax on fish.

79 In practice, WIEGO seems to have focused primarily on very specific groups of women workers in the informal economy, including domestic workers in countries such as South Africa, home workers in industries such as clothing manufacturing in parts of Asia and urban street vendors in West Africa. WIEGO does not seem to have engaged with the challenge of social protection for peasant women across Africa. When WIEGO officials point to models, they generally point to ones that entail massive state subsidies as well as contributions(such as Rwanda’s health insurance system). It is unclear what impact WIEGO has had within Africa. WIEGO participated in the process to review the formalization of informal employment, through SPIREWORK, that led up to the 9 th session of the AU’s Labour and Social Affairs commission(in April 2013). Although the Labour and Social Affairs Commission was a tripartite structure,(other) civil society organisations were invited to participate, although they were prevented from speaking! WIEGO later reported that they had succeeded in inserting a‘key paragraph’ into the LSAC’s report. 77 Another organization concerned primarily with the rights to social protection of workers is th e International Trade Union Confederation’s African Region(ITUC -Africa). 78 ITUC, formed in 2007 through the merger of two prior confederations, proclaims itself to be the ‘global voice of the world’s working people’, with the mission of promoting and defen ding workers’ rights. 79 Its leadership includes trade unionists from Africa. 80 ITUC-Africa claims to have 16 million declared members in 101 trade unions in 51 countries. One of its objectives is to achieve social protection for all workers. In 2009-10, the ITUC-linked African Labour Researchers Network(ALRN) compiled a study of social protection in Africa, for ITUC-Africa. The ALRN team posed the same basic question as Lund and Srinivas: How to expand social protection given that informal employment was not a transitory phenomenon? The team provided valuable summaries of social protection in a set of Anglophone African countries, but did not answer their own question beyond a general endorsement of the social protection floor concept(Kalusopa et al., 2013). In practice, ITUC-Africa seems to be proceeding down the same path as WIEGO, searching for models for the extension of contributory schemes to selected groups of urban, informal workers, such as domestic workers, market vendors and hairdressers, whilst also investing in capacity-building among its affiliated trade unions so that the individual unions can push more effectively for reforms in bi- or tripartite fora. One model that has attracted ITUC-Africa is the mutuelles de sante that workers have established in Francophone countries such as Mali and Senegal. In some cases, the state matches the contributions from workers, i.e. it pays one half. ITUC-Africa believes that African countries might also learn from the experience of European countries such as France and Belgium in building national mutual funds. The ILO has paid little attention to mutuelles, focusing primarily(in Africa) on social assistance. ITUC-Africa, however, must pay attention to the needs of the 77 http://www.wiego.org/sites/wiego.org/files/resources/files/WIEGO-e-news-Jan2013-Jun2013.pdf. 78 I am grateful to Mme Ghislaine Saizonou-Broohm(Equality and Social Protection program officer) for providing information on ITUC-Africa. 79 https://www.ituc-csi.org/about-us. 80 ITUC’s Deputy General Secretary, Wellington Chibebe, was formerly Secretary General of the Zimbabwe Congress of Trade Unions(ZCTU).

80 members of its affiliated unions, including members who have lot their formal jobs and been pushed back into informal employment. Social assistance is unlikely to offer sufficiently generous benefits for such workers, and workers might even be excluded through a means test. 3.5. Nonstate organisations and peasants’ rights to social protection Peasants have been almost entirely absent from discussions of rights to social protection in Africa. The AU has focused on selected vulnerable groups(especially children and the elderly), and its occasionally strong rhetoric has not been matched by reform on the ground. The ILO has also emphasized the needs of children and the elderly, as well as the needs of informally-employed workers, but also with limited success. HelpAge has campaigned for social pensions for the elderly, and UNICEF around the rights of children. But no major organization has championed the rights of peasants, to food sovereignty or even food security. Although Africa is, still, a continent of peasants, African actors have played only a minor role in the global struggle for recognition of the rights of peasants. La Vía Campesina was slow to organise in Africa(Desmarais, 2007: 6, 29). As of 2001, only one out of 101 of La Vía Campesina’s member organisations was in Africa. La Vía Campesina only established an African region in 2004. To promote itself in Africa, it held a major forum in Nyéléni(Mali) in 2007(Durnford, 2016: 127ff), and its fifth conference in Maputo (Mozambique) in 2008. In June 2011, representatives of peasants’ organizations affiliated to La Vía Campesina from ten African countries 81 met with activists from elsewhere in the world in a conference held in Zimbabwe. 82 In 2013, La Vía Campesina’s secretariat relocated to Zimbabwe(in line with the movement’s principle that the secretariat move every four years). In 2017, organisations from seven African countries attended La Vía Campesina’s seventh conference, in Spain. 83 La Vía Campesina’s website lists affiliated movements in seven other countries also. 84 Overall, however, neither La Vía Campesina 81 Zimbabwe, Mozambique, Democratic Republic of the Congo, Rwanda, Angola, Uganda, Tanzania, Kenya, Zambia, South Africa, Central African Republic. 82 https://cagj.org/2011/07/african-small-farmers-declaration-la-via-campesina-international-peasants-movement/. 83 These were: The Coordination Nationale des Organisations Paysannes(CNOP, Mali), the Landless Peoples Movement(LPM, South Africa), Zimbabwe Smallholder Farmer’s Forum(ZIMSOFF), the Mozambique Peasants Union(União Nacional de Camponeses, UNAC), the Plateforme Paysanne du Niger(PFPN), Conseil National de Concertation et de Cooperation des Ruraux(CNCR, Senegal), and the Ecumenical Association for Sustainable Agriculture and Rural Development(ECASARD, Ghana). See: https://viacampesina.org/en/african-peasants-highlight-struggles-via-campesina-global-conference/, and https://viacampesina.org/en/who-are-we/regions/west-and-central-africa/. 84 The Coordination Togolaise des Organisations Paysannes(CTOP), Concertation Nationale des Organisations Paysannes et des Producteurs Agricoles du Congo(CNOP-Congo, in Congo-Brazzaville), Coalition Paysanne de Madagascar(CPM), União Nacional das Associações de Camponeses Angolanos, Mtandao wa Vikundi vya Wakulima Tanzania(MVIWATA), Confédération Paysanne du Congo(COPACO/PRP, in the DRC), the Cadre National des Organisations Paysannes et de Producteurs Agricoles de la Guinée Bissau, and the National Coordinating Organisation of Farmer Associations of Gambia(NACOFAG). See: https://viacampesina.org/en/who-arewe/regions/south-and-east-africa/.

81 nor its member organisations have succeeded in putting peasants and food security on the agenda in Africa. The rights of peasants and to food sovereignty remain largely absent from African debates. The AU had not ignored agriculture or food security but approaches these from the point of view of the needs of consumers more than of the rights of producers. The AU adopted in 2003, in Maputo, a Comprehensive Africa Agriculture Development Programme (CAADP). The programme’s stated goal was to eliminate hunger and reduce poverty through agriculture, and it set targets for increased budgetary allocations to agriculture (to 10 percent of government expenditure) and increases in agricultural productivity(by a minimum of 6 percent p.a.).‘Pillar 3’ of the CAADP focused on food security. Its ‘Framework for African Food Security’ acknowledged the right to food, and mentioned the role of social protection, but these played a modest role. Critics charged that AU member states were slow even to commit themselves to the CAADP process. In 2012, the AU renewed its commitment to the CAADP, and declared 2014 to be the Year of Agriculture and Food Security. At the end of 2014 the AU adopted, in Malabo(Equatorial Guinea) a Declaration on‘Accelerated Agricultural Growth and Transformation for Shared Prosperity and Improved Livelihoods’(AU, 2014b). The Malabo Declaration included commitments to ending hunger and halving poverty by 2025, but never mentioned rights. At the same time, the AU held a series of meetings that led to the adoption of an implementation strategy for the CAADP, as well as guidelines for country-level implementation(AU, 2016). These documents did not mention rights. Nor did these documents mention social protection. The issue of rights and social protection for the rural population did arise in the context of the AU’s SPIREWORK initiative for the informal sector(see above). In 2013 the AU did initiat a process of meetings to discuss how its SPIREWORK might be integrated into the CAADP. Despite the apparent attempt to bring these initiatives together, SPIREWORK was interpreted as concerning only‘rural workers’, defined in terms of employment for wages, excluding most peasant farmers(AU, 2013, 2015b). A more fruitful initiative was driven by the FAO and UNICEF through their joint programme on‘From Protection to Production’. The programme focused on peasants(or‘small family farmers’) in Africa, and included detailed research in Ethiopia, Ghana, Kenya, Lesotho, Malawi, Zambia and Zimbabwe. In 2013-14, a series of consultative workshops was held, includi ng in Accra and Cape Town, on‘ Strengthening Coherence Between Agriculture and Social Protection ’. The FAO reported on the Ghana workshop: ‘FAO has been facilitating inter -ministerial policy dialogue at the regional level in Africa to improve coherence between agriculture and social protection. In September 2013, a workshop in Ghana brought together government officials from the ministries of agriculture, social welfare and finance from Angola, Ethiopia, Ghana, Kenya, Malawi, Niger, Rwanda and Zambia. Representatives from the African Union, World Bank, UNICEF and the World Food Programme also attended. For many countries, it was the first time that staff from different ministries discussed the links between agriculture and social protection in depth.’ 85 85 http://www.fao.org/3/a-i3920e/i3920e10.pdf.

82 Thes e informed the drafting of an impressive‘Framework for Analysis and Action’(FAO, 2016) and an accompanying‘Diagnostic Tool’. These did not invoke or discuss any discourse of rights, however. Subsequently, the FAO did endorse a rights-based approach to social protection(FAO, 2017, see above). Discourses of rights have yet to be applied substantially to the rural poor in Africa, except insofar as they are children or older people. The FAO, which is the primary international agency concerned with peasant farmers, has recently embraced a rights-based approach, but its primary institutional partners are Ministries of Agriculture which tend to apply a productivist approach to agriculture and hunger and are protective of their (expensive) farm input subsidy programmes. They may be happy to present their programmes in terms of social protection as well as production, but they are less happy about handing over responsibilities and resources to other ministries. Broader global or Africa-wide initiatives to extend social protection to informal workers – including the SPIREWORK initiative, under the AU’s auspices but shaped if not driven by the ILO – have made little progress in Africa, where self-employed peasants(and urban traders etc) constitute a very large sha re of the‘informal’ economy. 3.6. Nonstate organisations and the rights to social protection of‘vulnerable groups’ The expansion of social assistance programmes has been justified primarily in terms of the needs of‘vulnerable groups’, which generall y have comprised groups of people who are considered to be unable to work as either‘workers’ or peasants: The elderly, disabled, children, pregnant women and women whose time is taken up in care work, and in some cases the landless. The rights to social protection of these groups of people have been asserted strongly in the global discourse and(as we have seen above) have been asserted with some qualifications in the formal positions of the AU and sub-regional organisations. These groups have also been championed in Africa by international organisations and NGOs, including UNICEF and HelpAge International, as well as African organisations, including especially the African Platform for Social Protection. In their activities in Africa, these organisations have also vernacularized the global rights discourse. Non-government organisations have often held back in their advocacy of a rights-based approach, apparently assessing that they are more likely to secure national governments and policy reforms if they motivate for them on other grounds. This is very clear in the case of HelpAge International. At the global level, HelpAge has long been a strong advocate of the rights of the elderly, including(since at least 2004) to social pensions. HelpAge’s website pro vides detailed material on rights. Indeed, HelpAge summarises its function as helping older people to claim their rights. When engaging with governments in Africa, however, HelpAge downplays rights(as Leutelt noted in 2012: 11). In motivating for its 2006 conference in Livingstone, HelpAge consigned the discussion of rights to a short passage in the middle of the paper(HelpAge, 2006: 4). Its 2009 report advocating social pensions in Zanzibar did not mention rights at all(HelpAge, 2009). Its 2010 report advocating social pensions in Tanzania referred to the right of the elderly to

83 social welfare that was specified in the cou ntry’s constitution, but otherwise avoided any discussion of rights and concentrated instead on the developmental and other benefits of introducing social pensions(HelpAge, 2010). Rights were not mentioned in a paper on the Tanzanian case that was presented at a major conference in Arusha in late 2014 (HelpAge, 2014). HelpAge’s 2016 report setting out the case for social pensions in Malawi mentioned rights only in passing, in reference to Malawi’s constitution and National Social Support Policy(HelpAge, 2016: 5). Even strong supporters of a rights-based approach tend to steer around questions of rights in their engagement with national governments in Africa. UNICEF has similarly trod carefully around the question of rights in its advocacy of social protection for children. African-based organisations have also to tread carefully. The African Platform for Social Protection(APSP) comprises a network of national NGOs established to engage with the AU and national governments across Africa. 86 The APSP was established in 2008 with strong support from organisations in the global North and was subsequently funded by them. 87 International NGOs(including HelpAge) were represented on the APSP’s interim board. The APSP operates from a rights-based basis, but tailors this as necessary: ‘The APSP position is that social protection is provided for in the Universal Declaration of Human Rights(Article 22 and 23) and in ILO conventions, making provision of social protection an entitlement, not charity or a handout. However, the APSP also recognises that governments are on different levels of the social protection debate, development of policy, design and implementation of programmes. Insistence on social protection as a right, therefore, has to take into account the indivi dual country context.’(20 -21) In practice, the APSP seems to have emphasized rights more strongly than most of the international organisations. In its advocacy work – which included national consultations, regional meetings(for example, in Malawi, Kenya and Senegal in late 2012), engagement with the AU, training and study visits – the APSP has pushed strongly for rights to be recognized by ensuring that social protection programmes were enshrined in national constitutions or at least legislation. But the APSP recognizes also that it wields more influence with national governments when it can back up its arguments in favour of expanded programmes with empirical evidence of their benefits, rather than ideological assertions. The APSP was heavily involved, f or example, in the AU’s deliberations over the rights of older people, culminating in the AU’s adoption of a Protocol in 2016(see above). During 86 I am grateful to Samuel Obara for valuable insights into the work of the APSP. 87 The APSP grew out of the UK-based Grow Up Free From Poverty campaign, driven by a coalition of UK-based NGOs and faith organisations, mostly concerned with child poverty but including also HelpAge International. In 2006, the coalition embraced social protection, and also committed to supporting African-based civil society organisations to promote social protection. See: https://www.youtube.com/watch?v=KhzLKi7fosg. In 2012, in its first Review Report, the APSP reported that it had received more than$2.2m from donors(including Sida/Norad, Cordaid, Save the Children and DfID)(APSP, 2012: 6). The Review Report itself was prepared by a consultant, with funding from DfID and SIDA. In 2013-14, DfID and Save the Children ended their support, leaving SIDA and NORAD as the remaining funders(APSP, nd).

84 the AU’s Conference of Ministers in Charge of Social Development in 2012, the APSP participated in the Experts Group meeting that helped to develop the Protocol. The APSP then assisted in a workshop organised by the African AU the following year. For the APSP, this was a major achievement: The Protocol represented‘an important milestone for older persons of Africa as it will go a long way in guaranteeing their rights. As with the Policy Framework and Plan of Action on Ageing(2002) and the AU Social Policy Framework for Africa(2009), Africa is leading the way in seeking comprehensive solutions for its older citiz ens’(APSP, nd: 23). The APSP worked similarly with the AU over SPIREWORK and, most recently, over the development of a protocol on the minimum package of social protection. The APSP has also worked closely with international as well as local non-government organisations. It partnered with Save the Children in thinking through what child-sensitive social protection policy might entail. It worked with an FAO-led consortium examining the needs of peasants and informal farmworkers. One of the APSP’s objectives is to enshrine commitments to social protection in national constitutions or legislation. For the most part, it has made little progress on this. Much of the expansion of social protection in Africa has not been rooted in any constitutional or legislative commitment. 3.7. National constitutions It is widely believed that the specification of legal rights to social protection is conducive to the expansion of effective coverage. We have already considered above some aspects of the legal framework, and shown that these might not be as important as is sometimes imagined. In Chapter 1 we considered the cataloguing by the ILO, in its World Social Protection Reports , of national governments’ ratification of ILO conventions. We showed, however, that there was little or no evidence of any relationship in Africa between ratifications and the extent of effective coverage of social protection(as measured by the ILO itself). This might be because, as Wujczyk(2016) notes that in his introductory essay to t he first of the ILO volumes on the right to social security,‘ the position of social security rights still depends on whether they are guaranteed in the internal legal systems of particular countries ’(emphasis added). In at least some African countries, i nternational agreements rarely have force unless they are incorporated into national legislation. In Malawi, for example, an amendment to the constitution in 2001 meant that international agreements only become binding in Malawi if they have been incorpora ted in Malawi’s own national legislation(Chirwa, 2005b: 233). 88 In Botswana, ministers say that international conventions or charters must be incorporated into national legislation to have any standing. 89 The government Francophone countries seem to be subject to the French legal doctrine that holds that international covenants are not applicable in domestic courts (Diallo, 2016). 88 Chirwa suggests that it is unclear whether even those international agreements incorporated into law before 1994 remained binding under the new constitution(2005b: 233-4). 89 Interviews, Gaborone, 2017.

85 Several scholars have argued strongly that the recognition of legal rights to social protection in national constitutions or legislation is of crucial importance in policy reform. ‘Constitutional guarantees play a very important proactive role in introducing social rights into national legislation and in fostering their implementation’, writes the series editor of a new ILO series on The Right to Social Security in the Constitutions of the World. He continues: ‘They can control the generation of norms, provide a justification for rules, mechanisms and institutions that already exist, provide guidance in the interpretation of other rules, and influence the organization of public services. They also provide a basis for the protection of rights through institutional mechanisms, primarily through constitutional and supreme courts, thereby holding out the promise of the fulfilment of such rights in future. Where there is an individual right of action before the supreme or constitutional court, individuals are able to participate actively in the realization of their social rights. … Finally, the Constitution as a legal embodiment of human values attaches to social security rights a strong moral dimension – that of preventing the unjust denial of human dignity together with income.’(Egorov, 2016: xvi). If the constitution emphasizes social and economic rights, then legislation and actual policy reform and implementation may be more likely. If there is no constitutional framework, then legislation might serve as a somewhat weaker substitute. Barrientos emphasizes the importance of both constitutional and legislative provision for the ‘institutionalisation’ of social protection:‘Legislative or constitutional recognition is essential to establishing that access to social assistance is an entitlement, and not subject to the discretion of government officials’(2013: 208). Barrientos himself recognizes that few social protection programmes originated in clear constitutional or legislative requirements. More often,‘social assistance programmes [began] life as projects, as interventions with a limited time window, usually based on an exe cutive decision’. Programmes are often initiated, funded and regulated through executive rather than legislative processes( ibid: 207). This is especially true in Africa, where social protection has rarely made it into national constitutions or even legislation. Even when social and economic rights are recognized in constitutions or legislation, this has not always had clear and substantive effects on policy-making and implementation. One reason for this was suggested in Chapter 1.2. There is little or no evidence of any relationship in Africa between the existence of national legislation covering the different ‘branches’ of social security(i.e. what the ILO refers to as the‘scope’ of social protection; see ILO, 2017) and the extent of effective coverage. This is likely to be because the existing of legislation governing contributory pensions, for example, has no bearing on rights to, or the reach of, non-contributory pensions. This suggests a need to consider institutionalisation of social protection at the level of individual programmes as well as at the level of the country as a whole. Contributory pensions might be institutionalized, but social pensions not at all institutionalized. The SAPI database reports the degree of‘institutionalisation’ of indi vidual social assistance programmes. SAPI distinguish between‘institutionalised’,‘precariously institutionalised’

86 and‘pilot’ programmes, in part according to their legal status.‘“Institutionalised” indicates a leading role of central governments in the implementation and coordination of programmes. In such cases, programmes usually have national coverage, and are wellintegrated into legislation and the governments’ budgets, even when supported by donors’. SAPI classify as‘institutionalised’ most socia l assistance programmes in the middle-income countries of Southern Africa(Namibia, Botswana, South Africa, Lesotho) and the Indian Ocean(Mauritius and the Seychelles), along with Rwanda and Ethiopia. A number of countries have weakly institutionalised social assistance programmes. Most social assistance programmes in Francophone countries are considered to be pilot programmes, not even precariously institutionalized. In this section, we examine the specification of rights to social protection in national constitutions. In Chapter 4 we consider case-studies of selected countries, examining how the recognition of rights in either constitutions or legislation has shaped policy reform. It is not straightforward to gauge the extent of constitutional commitments to social protection across Africa. The ILO’s comprehensive review is supposed to result in ten volumes. As of September 2017, however, the ILO has published only one volume, on Europe(ILO, 2016). Other volumes dedicated to social and economic rights(e.g. Chirwa and Chenwi, 2016) do not provide a comprehensive catalogue. The fullest catalogue is provided by the FAO, which hosts a website on the‘Right to Food Around the Globe’. 90 This site collates information on the constitutional status of the right to food in most African countries, and in so doing has become a useful source on the constitutional status of social and economic rights in general(although it is incomplete on rights that have no relevance to the right to food). The website distinguishes between four levels of constitutional status: Explicit protection of the right to adequate food, implicit protection of the right to adequate food, directive principles of state policy, and international legal commitments. Overall, very few African countries have clear explicit or even implicit commitments to social and economic rights. More commonly, social and economic rights are the subject of directive principles. Governments have signed up to many international agreements, but it is often not clear what this means. In Anglophone Africa, South Africa stands out in its constitutional and legislative endorsement of social and economic rights. South Africa’s 1996 post -apartheid constitution includes a Bill of Rights that recognises explicitly rights to social protection. This case is considered in more detail below(section 3.1; also, Liebenberg, 2002). Kenya’s 2010 constitution is often held up as another case that recognizes social and economic rights(in Chapter 4, its Bill of Rights). Article 43(1) does indeed recognize rights to housing, health care, sanitation, water, and‘social security’, and‘to be free from hunger, and to have adequate food of acceptable quality’. Article 43(3) elaborates on the right to social security:‘The State shall provi de appropriate social security to persons who are unable to support themselves and their dependants’. The specific rights of children and the disabled are set out in Articles 53 and 54. These do not include explicit rights to social assistance. Article 57 specifies that‘ The State shall take measures to ensure the 90 http://www.fao.org/right-to-food-around-the-globe/.

87 rights of older persons …( d) to receive reasonable care and assistance from their family and the State.’ Article 60 provides for(inter alia)‘equitable access to land’. Zimbabwe’s 2013 Constitution stipulates(in Article 30) that‘ The State must take all practical measures, within the limits of the resources available to it, to provide social security and social care to those who are in need.’ The Constitution acknowledges as founding values or principles the rights of the disabled, women, elderly, youth and children (as well as of veterans of the liberation struggle), and‘the equitable sharing of natural resources, including land’(Article 3). The specific r ights of women, children, youth, the elderly and the disabled are protected under Articles 17, 19, 20, 21 and 22 respectively (with further elaboration under Articles 80 through 84). The elderly and veterans of the liberation struggle are accorded a wider range of rights. These include not only rights to care and food(if needed), but also to a degree of income security. The State is enjoined to‘endeavour, within the limits of the resources available … to develop programmes to give elderly persons the opportunity to engage in productive activity suited to their abilities and consistent with their vocations and desires’(Article 21(2)(c)). At the same time, ‘p eople over the age of seventy years have the right(a) to receive reasonable care and assistance from their families and the State;(b) to receive health care and medical assistance from the State; and(c) to receive financial support by way of social security and welfare; and the State must take reasonable legislative and other measures, within the limits of the resources available to it, to achieve the progressive realisation of this right’(Article 82). Veterans of the liberation struggle are provided for under Articles 23 and 84(which specifically refers to pensions). Articles 24 and 65 cover workers. Zimbabwe also recognizes explicitly rights to food and water. Article 77(b) records that ‘ Every person has the right to sufficient food; and the State must take reasonable legislative and other measures, within the limits of the resources available to it, to achieve the progressive realisation of this right ’. Article 81(1)(f) specifies that children have rights to education, health care, nutrition and shelter. Overall, and more clearly than in the Kenyan case, the Zimbabwean Constitution emphasises the importance of self-reliance and hard work. The Preamble refers to the goal of‘ a united, just and prosperous nation, founded on values of transparency, equality, freedom, fairness, honesty and the dignity of hard work’. Article 13(National Development) commits the state to the pro motion of‘private initiative and self reliance’, as well as‘balanced development’(especially between rural and urban areas). Article 15 (Food Security) requires that‘The State must(a) encourage people to grow and store adequate food;(b) secure the establishment of adequate food reserves; and(c) encourage and promote adequate and proper nutrition through mass education and other appropriate needs.’ Most other Anglophone African countries recognize social and economic rights in their constitutions weakly, at best through directive principles of state policy, or not at all. The provisions of the 1981 African Charter and subsequent Charters and Protocols have been largely ignored. Chirwa discusses in detail the case of Malawi’s 1994 constitution, which h e describes as having been written in haste, largely by‘foreign experts’(Chirwa, 2005a: 211). It seems to have taken the new constitutions of Namibia and Ghana as models in

88 including only a weak recognition of social and economic rights. The constitution does include a Bill of Rights, which recognizes rights to education and property, and for workers. Rights to health care and nutrition were recognized as(unenforceable) principles of national policy. Rights to social protection or an adequate standard of living were not recognized at all. The Bill of Rights did, however, include a right to development. Whereas the 1981 African Charter(Article 22) recognized a right to development as a collective right, the Malawi constitution implies that it is also an individual right: Article 30(2) specifies that‘t he State shall take all necessary measures for the realization of the right to development. Such measures shall include, amongst other things, equality of opportunity for all in their access to basic resources, education, health services, food, shelter, employment and infrastructure’. Section 30 also requires the state to ensure equality of opportunity – especially for women, children and the disabled – and to eradicate social injustices and inequalities. Overall, however, Chirwa assesses that the protection of socioeconomic rights in the Malawi constitution is‘inadequate’( ibid: 214) and‘half hearted’( ibid: 241). The protection of social and economic rights in the constitutions of Tanzania, Zambia and elsewhere also remains weak. Tanzania provides what the FAO optimistically concludes to be an implicit recognition of the right to adequate food. Article 11(1) requires that the state‘make appropriate provisions for the realization of a person’s rig ht to work, to selfeducation and social welfare at times of old age, sickness or disability and in other cases of incapacity’. Two directive principles are relevant also to social protection: Article 8(1)(b) directs that‘the primary objective of the Government shall be the welfare of the people’, whilst Article 9 directs the state to ensure that national wealth is used to eradicate poverty, ignorance and disease. Zambia’s constitution includes several relevant directive principles. Article 112(c) specifie s that‘ the State shall endeavour to create conditions under which all citizens shall be able to secure adequate means of livelihood and opportunity to obtain employment ’ and 112(f) directs the State to ‘ endeavour to provide to persons with disabilities, the aged and other disadvantaged persons such social benefits and amenities as are suitable to their needs and are just and equitable. ’ Francophone countries have been more influenced by French legal traditions. Chirwa writes that the wave of re-democratisation after 1989 resulted in many Francophone countries – including Benin, Burkina Faso, Gabon, Madagascar, Mali, Niger and Togo – adopting new national constitutions that recognized social and economic(as well as cultural) rights(Chirwa, 2005a: 207). The FAO website is more sanguine on the status of the right to food in these constitutions. Mali’s constitution, for example, records that education and training, work, housing, leisure, health and social protection are ‘recognised rights’(Article 17), but the import of this‘recognition’ is not clear. Niger’s constitution recognises explicitly the right to‘a healthy and sufficient diet’ alongside rights to life, physical and moral integrity, safe drinking water, education and training, but all of these are subject to the‘conditions defined by law’(Article 12). 91 91 The state is also directed to promote ‘food sovereignty, sustainable development, universal access to social services and improved quality of life’(Article 146.3).

89 Whilst food sovereignty has been incorporated into the constitutions of various Latin American countries, 92 this does not seem to have happened in any African country. Durnford writes that rights to food sovereignty in Africa exist only through legislation, including in Mali(since 2006) and prospectively in Senegal(Durnford, 2016: 63-4). The FAO reports that Senegal implicitly provides for the right to adequate food through Article 8 , which‘guarantees to all citizens fundamental individual freedoms, economic and social rights as well as collective rights’, including(specifically) the rights to work and to health, and Article 17(3), which‘guarantees to families in general and those living in rural areas in particular access to health services and wellbeing’ as well as guaranteeing‘women in general and those living in rural areas in particular the right to lighten their living conditions’. Lusophone countries seem to have embraced social and economic rights in their constitution, although sometimes rather vaguely. Angola’s constitution requires that the state adopts‘legislative initiatives and other appropriate measures to ensure the gradual and effective realisation of economic, social and cultural rights, in accordance with the available resources’( Article 28(2)) whilst Article 90(e) directs t he state to‘promote social development by ensuring that all citizens enjoy the benefits resulting from collective efforts in terms of development, specifically with regard to quantitative and qualitative improvements to standards of living’. The constitution of Cabo Verde(1992, revised 1999) is more specific. Article 67( 1) stipulates that‘Everyone shall have the right to social security, in conformity with the national development, for his protection in the unemployment, illness, disability, old age, or as an orphan and in all situations of lack or diminution of the means of subsistence or of the capacity to work’. Section(2) continue s to require that‘The State shall ensure the gradual creation of the conditions that are indispensable to the exercise of these rights, namely through the adoption of policies towards the setting up of a decentralized national system of social security and a national network of medical and hospital services’. Special provision is made for the disabled and the elderly, who‘shall have the right to special protection from their family, the society and the State’ (Article 72). Specific provision is also made for workers, including to ‘ adequate social security in case of illness, work accident, old age or involuntary unemployment, as well as to paid and periodical vacation, rest, leisure and material assistance shall, progressively, be guaranteed to all workers, in accordance with the national economic development ’(Article 58). The constitutions of some African states are distinctively African in other respects. The family features in many constitutions. For example, Malawi’s 1994 constitution recognizes the family as‘the natural and fundamental group unit of society’ and committed the state to protecting the family. The Kenyan Constitution insists that‘The family is the natural and fundamental unit of society and the necessary basis of social order, and shall enjoy the recognition and protection of the State’(Article 45 (1)). Article 25 of Zimbabwe’s Constitution similarly requires that the State protects and fosters the institution of the family. In Cabo Verde, also, the family is identified in the constit ution as‘the fundamental element and the foundation of the whole society’( Article 84(1)). Similar provisions, 92 Bolivia, Ecuador, Venezuela and Nicaragua. This has been discussed in El Salvador and the Dominican Republic also.

90 reminiscent of the provisions of the 1981 African Charter(especially Article 18(1)), are reported to be included in the constitutions of Mozambique and Namibia. Self-reliance and work also feature in several constitutions. The Tanzanian constitution also requires that the state‘shall make provisions to ensure that every person earns his livelihood’, i.e. that he or she works. Probably not coi ncidentally, Tanzanian citizens are much more likely to endorse workfare than citizens of any other African country surveyed by the World Values Survey. The constitution of Cabo Verde specifically discusses citizens’ duties.‘Everyone shall have duties tow ards his family, the society and the State and also towards other legally recognized institutions’(Article 80), including duties to ‘work within the limits of his capabilities and capacities’ and to‘pay the contributions and taxes established by law’(Ar ticle 82(b) and(c)). In sum, there is wide variation between African countries in the constitutional recognition of rights to social protection. Very few African countries enshrine clear rights to either income or food security, but some point to ill-defined rights to social security, especially for specific groups such as the elderly. Constitutional provisions are almost always vaguer and weaker than the provisions in the international agreements signed by most African states. Like the Africanised versions of international charters and protocols, national constitutions in Africa often include an emphasis on the family, and sometimes also on duties. 3.8. Assessing the Vernacularization of Global Rights Discourses This chapter has examined the‘vernacularisation’ or‘localisation’ within Africa of global discourses of rights to social protection through close analysis of the work of different organisations: The OAU and AU, which has adopted formal Charters, Protocols and other declarations; the ILO and other organisations, which have promoted the rights of workers, including in informal employment; other international and regional organisations concerned with, especially,‘vulnerable groups’ such as the elderly and children; and national governments, through their own constitutions. The overall picture seems to be a complex mosaic of strong calls, weak calls and silences, applied unevenly to different sections of the population. Supposed undertakings are strongest for workers, the elderly and children. They are weakest, if not almost entirely absent, for peasant farmers. In rights discourses within Africa, much stronger explicit attention has been paid to rights to ‘social’ security – implying some kind of right to income security – than to rights to food security(or food sovereignty). In this, the African experience contrasts with experiences in Latin America(where the struggle for peasants’ rights – and perhaps rights generally – is relatively stronger) or South Asia(where the struggle for the right to food is relatively stronger). Rights to social security have generally been associated with specific groups (workers, and perhaps the elderly), with little or no consideration of the rights of peasant farmers. The overall application of global discourses into Africa has been muted as well as selective. A range of international actors who advocate rights in general have been

91 reluctant to emphasise rights in their work in Africa and have chosen to emphasise instead the benefits of social protection in terms of poverty reduction and development. This seems to reflect a widespread perception that the global discourse of rights is not persuasive among African policy-making elites. The vernacularisation of the global discourse in African regional agreements(as well as in national constitutions) provides some evidence of this. The OAU/AU’s Charters and Protocols, as well as national constitutions, tend to emphasise the importance of family and of responsibilities or duties, to balance the emphasis on rights-bearing individuals in the rights discourse developed in or through the global North. The fact that rights to social protection have been selectively and unevenly enshrined in Africawide institutions(and debates) begs the question‘what is the import of this?’. Does it matter, and, if so, how? Wright and Noble, after enthusiastically noting various declarations of rights to social protection in the mid2000s, describe them as‘collective declarations of intent ’(2010: 116). Is this all they were? And we re they even this? The existing literature provides little in the way of systematic, empirical answers to these questions. The literature does, however, alert us to the potential risk associated with the global rights discourse. A series of scholars have pointed to the ways in which the global rights discourse assumes and endorses a vision of autonomous, rights-bearing individuals(e.g. Claeys, 2015b). Some of the risks involved with this have been explored in the Southern African context by anthropologists. Englund, in Prisoners of Freedom, points to the ways in which an emphasis on individual political and civil rights may have distracted attention from collective economic and social rights that remained, in the late 1990s, unacknowledged(Englund, 2006). His more general point is that all sorts of actors – including well-meaning NGOs – can be prisoners of their concept of rights, with the risk that their promotion of some rights serves to hide other injustices or inequities. More recently, he has contrasted an individualistic understanding of poverty – and, we might add, economic and social rights – from the understanding that prevails across much of Malawi, where he conducted his research, and elsewhere across Southern Africa. In Chichewa, Englund writes, a distinction is drawn between poverty( umphawi), which refers to the absence of food, medicine, clothing and so on, and destitution ( kusaukitsitsa), which refers to the absence of relationships as well as possessions. ‘Much of the moral debate about po verty carried out in Chichewa explores the thin line that separates the two conditions.’ Englund acknowledges that obligations have narrowed in Malawi, but they remain important, not least because the failure to honour perceived obligations might result in hostility, including through witchcraft. This was neither a solidaristic community nor a world of autonomous individuals, but a world of people linked through complex relationships(Englund, 2012). Englund contrasts the use of human rights discourses by activists speaking on behalf of the poor in Africa(specifically, in Malawi), and the ways in which‘the poor … might themselves describe their predicament’ (Englund, 2014: 199). Villagers in Malawi make claims not as individuals or even as communitarian gr oups, but as‘persons in relationships’, including especially relationships of dependency. Claims are the expression of rights that are‘particular to the relationships that enable the claims. They are relationships rights, not human rights’( ibid: 204). What

92 this means for the codification of rights is not clear. Englund himself asks:‘What difference would it make if rights were the attributes of neither individuals nor groups but of relationships?’( ibid: 199). He does not provide a clear answer. This report does not draw on ethnographic research at the village level. What we can do, however, is to examine how the global rights discourse has shaped social protection reforms at the national level, through case-studies of selected countries, and how it compares and contrasts with the informal and often implicit understandings of rights(and responsibilities) that predominate among policy-making elites at the national level. Chapter 4 of this Report examines four country case-studies.

93 Chapter 4: Neither necessary nor sufficient: Rights and social protection at the national level This Chapter examines the extent to which, and ways in which, the new‘global’ discourse of rights has been incorporated into regional and national strategy and policy documents. The section examines the legal basis of social protection programmes in selected countries, including primarily South Africa, Botswana and Zambia. Whilst these three cases are all in the same, Southern African region, they have been selected purposively in that they exemplify different patterns of the institutionalisation of both rights and social protection programmes(see Figure 4.1). South Africa is the case where social and economic rights are most fully institutionalized, not only in the Constitution and legislation, but also in court cases, political rhetoric and even popular discourse. South Africa also has a social assistance system that is both unusually extensive and unusually generous. Botswana also has an extensive system of social assistance, but it is parsimonious rather than generous, and the government and ruling party have long eschewed any mention of social or economic rights. Whereas South Africa has adopted a classically‘rights based’ approach, Botswana’s welfare state is based on very different norms and values. Zambia provides an example of a country where both programmes and discourses of social protection are fluid, and the focus of contestation. Ideally, we would have considered also a case of a country with a strong formal commitment to social and economic rights but little public provision in practice. Kenya might have fallen into this position in the early 2010s, when its programmes did not match the promises in its new constitution. Africa as a whole might be considered to occup y this space, through the AU’s formal commitments. Social protectio n Limited Extensive Formal institutionalisation of rights Weak Strong Zambia Kenya? Botswana South Africa Figure 4.1: Case-studies in terms of institutionalization of rights and extent of social protection In reality, the variation between countries might better be understood in terms of 2dimensional space rather than a 4-quadrant diagram. We currently do not have good cross-national data on either the institutionalisation of rights or effective coverage or reach

94 (taking into account the scope, extent and level of social protection. This is a topic for further research. Our intuition is that the relationship between the institutionalization of rights and effective coverage is weak. Each country case-study includes a table that summarises the social protection landscape, in terms of provision for the four categories of people identified in earlier chapters: formally-employed workers; informally-employed workers and the unemployed; peasant farmers; and‘vulnerable groups’(i.e. the elderly or disabled, children, pregnant women and women preoccupied with domestic carework), especially if they live without any working-age adults available and fit for work, in socalled‘incapacitated households’. The concept of the‘incapacitated household’ is one rooted in Northern assumptions about households having breadwinners, but given substance across much of Africa in that social, economic and demographic changes have left growing numbers of people destitute in the sense of detached from networks of obligation. The second column summarises the need for social protection, in terms of the approximate numbers of people needing social protection and the primary risks against which they need to be protected. The third and fourth columns summarise existing and prospective coverage, using the same criteria. The final two columns summarise any relevant formal rights framework and the informal rights that are discussed more extensively in Chapter 5. 4.1. South Africa South Africa is often held up as an international lodestar, in terms of both its constitutional provision for social and economic rights and the actual reach(and effects on poverty) of its social assistance programmes. South Africa clearly has both an unusually high degree of institutionalization of rights and unusually extensive social protection coverage. Table 4.1 provides an overview. 4.1.1. Rights South Africa’s post -apartheid constitution, adopted in 1996, includes a bold vision of socioeconomic rights. Section 26 declares that‘everyone has the right to have access to adequate housing’ and section 27 specified that‘everyone has the right to have access to(a) health care services, including reproductive health care;(b) sufficient food and water; and(c) social security, including, if they are unable to support themselves and their dependents, appropriate social assistance’. Both of these clauses are qualified, but the qualification uses the‘progressive realisation’ formula i ncluded in the 1966 International Covenant on Economic, Social and Cultural Rights :‘The state must take reasonable legislative and other measures, within its available resources, to achieve the progressive realization of each of these rights’. Section 28 set out a more onerous and unqualified set of rights for children up to the age of eighteen(including a right‘to basic nutrition, shelter, basic health care services and social services’). Section 29 recognised that ‘everyone has the right(a) to a bas ic education, including adult basic education, and(b)

95 to further education, which the state, through reasonable measures, must make progressively available and accessible’. Table 4.1: Formallyemployed workers Informallyemployed workers and the unemployed Social protection in South Africa Need Existing coverage Current or prospective reforms Widespread Considerable, because of the scale of formal employment contributory protection for old age, injury, unemployment, and(less so) National health insurance; Pension fund reforms poor health Workfare Considerable, primarily because of unemployment (limited); Social assistance for old age, child support Expand social insurance; Expand workfare Peasants Low: peasantry largely destroyed Minimal Limited: Land reform? ‘Vulnerable’ groups/ ‘incapacitated’ households Considerable, because of diminished kin support Social assistance: Old age pensions, disability grants, child grants Increased value of child grant benefits? Formal rights Legislated In constitution, subject to affordability and ‘progressive realisation’ clauses None In constitution, subject (except for children) to affordability and ‘p rogressive realisation’ clauses Informal rights culture (Strong expectations) Weak Limited (except that land has symbolic importance) Taken for granted (except perhaps child grants?) Not only are rights enshrined in the constitution, but most major welfare programmes are governed by legislation, i.e. they are not entirely at the discretion of state officials. Social assistance programmes are governed by the 2004 Social Assistance Act. The preamble to the Act restates the constitutional provisions regarding income security. The parastatal South African Social Security Agency(SASSA) is governed by the 2004 SASSA Act. Contributory pension and provident funds are regulated under the 1956 Pension Funds Act and 1996 Government Employees Pensions Act. Unemployment insurance is governed under the 2001 Unemployment Insurance Act. Contributory medical aid schemes are governed under the 1998 Medical Schemes Act. Further legislation governs compensation for workplace injuries. Finally, labour legislation(specifically the 1995

96 Labour Relations Act) provides for the compulsory application of industry-wide social insurance commitments once they have been agreed between employers and unions through collective bargaining. 4.1.2. Gaps The reach or coverage of South Africa’s social insurance and social assistance programmes – and the remaining gaps – have been widely analysed elsewhere(South Africa, 2002; Seekings, 2002, 2015; Seekings and Moore, 2014). South Africa has a very extensive system of social assistance programmes. As of the end of 2017, generous monthly benefits are paid to about 3.4 million old-age pensioners and 1 million disabled adults; more modest benefits are paid to the mothers(or other caregivers) of about 12.2 million children. These programmes are means-tested, but the means test serves primarily to exclude the rich, not to target the very poor(SASSA, 2018). Given South Africa’s total population of just over 50 million people, about one in three people receives a social assistance benefit; few households in the poorest half of the population include no one receiving any benefit. As of 2011/12, close to 9 million children also ate free meals at school. An Unemployment Insurance Fund paid benefits to about 0.7 million people p.a.: Most benefits are for unemployment(although they are paid to only a small fraction of all unemployed people), with some for maternity. South Africa’s semi -social insurance system pays old-age pensions to close to 1 million people, whilst about pension and provident funds have about at least 10 million contributors(although some people are counted twice or more because they are members of multiple funds)(Registrar of Pension Funds, 2017). The largest pension fund, the Government Employees’ Pension Fund, covers 1.2 million workers and about 0.4 million pensioners. 93 Just under 9 million people are‘members’ of medical aid sc hemes. The largest are Discovery Health, with 2.7 million members, and the Government Employees Medical Scheme, with 1.8 million members. 94 Finally, between 500,000 and 1 million individuals benefit directly every year from public works programmes(although the number of person-years of work created p.a. is much lower, around 200,000). 95 Whilst the system has broader reach or coverage than any other country in Africa, poverty persists: Most unemployed people do not benefit directly from any programme. 4.1.3. Rights and the expansion of social protection South Africa’s extensive welfare state largely predated the recognition of social and economic rights in the 1996 constitution, however. The post-apartheid state inherited social assistance and occasional public works programmes that redistributed from taxpayers to the poor and a‘semi social’ insurance system that provided for limited risk pooling and savings for some workers in formal employment. Means-tested social assistance programmes provided for the elderly, disabled and poor mothers, although the latter programme continued in practice to exclude the poor African majority of the 93 http://www.gepf.gov.za/. 94 https://businesstech.co.za/news/finance/205466/the-biggest-medical-aid-schemes-in-south-africa-and-how-muchtheyre-charging-in-2018/. 95 http://www.epwp.gov.za/.

97 population. State-regulated but privately-run pension funds provided for workers in formal employment. The welfare system had originally been designed to protect white(and coloured) citizens, but(with the exception of support for poor mothers and residential care for the elderly) had been formally deracialized by 1994. Social assistance programmes cost about 2 percent of GDP. The semi-social insurance system was more expensive, but nowhere near as expensive as the social insurance systems of middle-income countries in Latin America. After the transition to democracy the welfare state was reformed rather than transformed. Faced with the considerable cost of extending the existing programme for poor mothers to the African majority – which would have cost an additional 2 percent of GDP – the government considered abolishing the programme. It was persuaded to appoint a Committee of Inquiry, which recommended instead that the existing programme be replaced with a new Child Support Grant that paid much more modest benefits to all poor caregivers, although only until a child’s seventh birthday. The government did wind down state-funded residential care for the elderly rather than deracialize access. The other big change was the massive expansion of the private sector in health care, in terms of both service provision(i.e. hospitals) and funding(through contributory medical aid schemes). The state regulates this through the 1998 Medical Schemes Act. With these three exceptions, both social assistance and semi-social insurance expanded without fundamental changes. For example, access to existing disability grants and unemployment insurance was widened, and men became eligible for old-age pensions at a younger age. Whilst the new Constitution was in no way responsible for the wide scope of the social assistance system, it did encourage advocates of the expansion of social protection to employ the discourse of rights. Rights talk became general in policy debates in South Africa after 1996, as pro-poor activists sought to extend social protection to the large number of South Africans who remained poor because they were not eligible for one or other grant. In 2000, the government appointed a Committee of Inquiry into a Comprehensive Social Security System. The(Taylor) Committee was asked to propose both a long-term vision and immediate reforms, bearing in mind affordability. The Committee’s re port included one chapter dedicated to discussion of the rights enshrined in the 1996 constitution, and a large part of another on the rights of children under the constitution and international covenants(South Africa, 2002). This discussion of rights cer tainly informed the Committee’s recommendations(inter alia) that the Child Support Grant be extended to a child’s eighteenth birthday and a Basic Income Grant be introduced for adults. Civil society organisations lobbied Parliament in favour of both recommendations, stressing the rights set out in the constitution and international law. The government rejected emphatically the call for a Basic Income Grant whilst conceding the demand to raise the age limit on the Child Support Grant – although only to the age of fourteen, citing affordability. Civil society organisations continued to press for the age limit to be raised to eighteen and took the matter to the courts. Whilst the case was still before the court, the government conceded the issue, raising the age threshold to eighteen. The result was that the number of child grants paid every month rose to almost

98 12 million, and the cost of the entire social assistance system rose to more than 3 percent of GDP(Matisonn and Seekings, 2004; Proudlock, 2011; Seekings, 2016f). It might seem that the rights-based approach reflected in the Constitution and championed by civil society organisations played the key role in the expansion of the social assistance system. Civil society activists suggest that their advocacy served to fuel a‘culture of demand’ for expanded social assistance. Other factors were also important, however. Prominent among these was the need for the ruling party to demonstrate that it was doing something about poverty, which remained stubbornly high. It is very possible that the Child Support Grant would have been expanded even if the Constitution had not specified social and economic rights. Moreover, the Constitution had not(yet) resulted in wholesale expansion of the welfare system. Court cases have not led to the introduction of any new programmes that would repair the holes in the safety net, such as a Basic Income Grant. The cases that have come before the courts shed light on why the inclusion of rights in a constitution may not transform the policy landscape. A series of major court cases have tested the scope in practice of the rights included in the constitution. Most of these cases have not focused on social protection itself. The first major case( Soobramooney, in 1997) concerned access to medical treatment. The second( Grootboom, in 2000) concerned access to housing, and the third(the 2002 Treatment Action Campaign case) concerned access to medical treatment. A more recent, controversial case( Mazibuko) concerned the rights of poor households to free water. Some other cases have concerned social protection directly. In the Khosa and Mahlaule cases, the court ordered the state to pay pensions and grants to legally-resident non-citizens on the same basis as citizens. Justice Mokgoro, in her majority judgement in the 2004 Khosa case, emphasised the constitutional commitment to building‘a caring society’. Even non -citizens(at least, those who had been given permanent resident status) should not be abandoned‘to destitution if they fall up on hard times’ and – as in the case of the elderly and children – were unable to earn a living for themselves. The state should not force them into‘relationships of dependency upon families, friends and the community in which they live’(Mokgoro, 2004: 41, 48). The Mahlangu case concerned the age threshold for children, and the Roberts case the age threshold for men for the old-age pension. In Roberts, as in Mahlangu, the Government conceded the issue before the relevant lower courts handed down judgements, forestalling the legal process. In all of these cases the courts have been reluctant to trespass on the policy-making roles of the legislature and executive. They have been especially cautious in handing down judgements with major financial consequences. Most of the social protection cases that came before the courts concerned primarily discrimination and had modest financial consequences. The courts’ hostility to discrimination is rooted more in South African history than in an expansive understanding of social and economic rights. The exception – the Mahlangu case involving the age of eligibility for child support grants – was unusual, in that the Constitution elsewhere defined individuals as children to the age of eighteen. In other cases, that tested more severely the limits of social and economic rights, the courts have not required that the government assumes major new financial obligations,

99 but rather that it can provide a reasonable defence of its policy choices and demonstrate that it is committed to the progressive realization of the specified social and economic rights, even if the realization is slow. In several major cases, the courts have effectively decided against the plaintiffs: the government was not obliged to provide very expensive dialysis to a patient with multiple medical problems( Soobramooney) or housing for homeless people( Grootboom) or more free water to the poor( Mazibuko). Critics charge that‘the post -apartheid judiciary has collectively failed to act as an institutional voice for the poor’(Dugard, 2008: 215). In cases like Mazibuko, critics charged, the timid court had abdicated its responsibilities, playing an ameliorative rather than transformative role. The failure to specify a‘minimum core’ of public services that the state must provide to its citizens meant that the egalitarian aspirations of the constitution were reduced to an‘equality of the graveyard’(Bilchitz, 2010: 603). The Court was said to have failed to take adequate account of the lived experience of poverty, and of how the insufficient basic public services provided to the poor served to reproduce their disadvantage(see Wilson and Dugard, 2011; Dugard and Bohler-Muller, 2014; Liebenberg and Quinot, 2012; Langford et al ., 2013). Critics acknowledge that‘the Court cannot on its own alter the balance of forces that sustain structural advantage and disadvantage’, but they assert that‘if poverty and inequali ty are to be meaningfully tackled, the Court will need to develop a more robust approach to the adjudication of socioeconomic rights and substantive equality’(Dugard and Bohler -Muller, 2014: 248). The Constitutional Court itself disputes this interpretation of its role, arguing that the constitutional separation of powers imposes limits on the extent to which it should assume the roles of the legislature and executive. Indeed, the Court itself sometimes divided over precisely this issue of deference to the executive. Whilst reluctant to engage in policymaking, the Court required repeatedly that the executive justify its actions, demonstrate their reasonableness, and show that its policies entailed progress in the progressive realization of socio-economi c rights. This partially filled the void left by parliament’s passivity, providing a soft constraint on the executive. For critics, however, this was not enough. The Court should, they argued, formulate a new approach to the separation of powers. ‘If the popular protests which have recently engulfed townships and informal settlements across South Africa tell us anything,’ Wilson and Dugard conclude,‘ it is that the basic needs of people participating in them have not been well-served by the “democratic” process upon which the Court says they must rely’(2011: 682). Faced with the argument that the Court has to be cognisant of its legitimacy, which is ultimately its only source of power, critics suggest that‘its greatest strength’ might in fact lie‘in developing a genuinely pro-poor jurisprudence that appropriately listens and responds to the lived realities of disadvantage in South Africa’(Dugard and Bohler -Muller, 2014: 248). Albertyn and Davis(2010) went further to suggest that, because the constitutional commitment to‘liberal democracy’ was the result of a political compromise, the Court’s role should be to go beyond this, in a‘social democratic’ direction. Constitutional court judges, including judges with impeccable pro-poor credentials, have defen ded the Court’s caution. In their view, the Court should generally‘leave the details of policy formulation and implementation strictly to the executive and legislative

100 branches’. One reason for this was that experience elsewhere suggests that judicial interventionism can easily be counter-productive. Even when cases(like Mazibuko )‘fail’ in the Constitutional Court, the process may help to strengthen democratic governance through requiring that the Government account for its policy(Cameron, 2014: 249-70). Paremoer and Jung argue that the Court’s refusal to define a minimum core obligation strengthened democracy by kicking‘the scope of government obligation back into the sphere of democratic politics’(2012: 207). South Africa’s welfare state did grow af ter 1996, but it was large to begin with and its basic shape remained largely unchanged. The inclusion of social and economic rights in the Constitution may not have been very important. If it had not inherited relatively generous old-age pensions, disability grants and grants for poor, single mothers, the ANC Government might well not have introduced them, notwithstanding the 1996 Constitution. The Constitution did not prevent the ANC Government contemplating the ending of support for poor mothers; the decision not to do so was not made on legal grounds. Nor did the Constitution prevent the state winding down residential care for the elderly; had there been a legal challenge, the Court might well have found for the state on the same grounds as the Soobramooney decision. Indeed, it is possible that the inclusion of social and economic rights in the Constitution was itself in part due to the prior existence of extensive social assistance programmes. The Constitution was, in effect, recognizing and justifying what already existed. What the Constitution might prevent is any retrenchment of South Africa’s pro -poor social assistance programmes, or even any diminution in the real value of benefits, unless the state could make a very compelling case. 4.1.4. Rights and the prospects for further reforms The inclusion of social and economic rights in the Constitution does not mean that courts will expand public provision through new reforms. In the South African case, policymaking remains largely the responsibility of the executive and legislature, both dominated by a party that has generally been reluctant to expand the social assistance system. In the late 1990s the ruling party(the African National Congress, ANC) and government were preoccupied with development rather than welfare, and worried about‘dependency’. The Department of Welfare said that it accepted the need for social grants, but,‘to ensure that those receiving welfare do not become permanently dependent on state aid, social grants for certain target groups will be closely linked to job creation and other anti-poverty programmes. Successful development programmes will empower people to earn a living, move off the social security system and achieve economic independence’(South Africa, 1996: 19-20). The 1997 White Paper committed the government to the goal of ‘developmental social welfare’ and‘re -orienting[its] services towards developmental approaches’. This meant helping people to meet their own needs, through‘the development of human capacity and selfreliance’, rather than relying on the state(South Africa, 1997; see further Patel, 2005; Gray, 2006). The Minister of Welfare, Geraldine FraserMoleketsi, revealingly told parliament in 1998 that‘welfare has become associated with charity and hand-outs, with food parcels and pensions, something in which it was alleged bleeding hearts got involved’. She called for a shift in thinking about‘welfare’,

101 ‘from paternalism to self reliance’ and investment in development. The Department’s flagship programme involved training unemployed women with young children so as to reduce their‘dependence on social security’. 96 The Department of Welfare was renamed the Department of Social Development. Under a new Minister, Zola Skweyiya, development temporarily took a back seat. Skweyiya understood that grants made a real difference to the lives of their poor recipients. Nonetheless, the state contested every case brought against it in the Constitutional Court, pointing to the cost of the existing system. When the Taylor Committee put forward proposals for a more extensive welfare state, most government ministers recoiled, denouncing‘handouts’ and‘dependency’ on the state(Matisonn and Seekings, 2003; Meth, 2004; Seekings and Matisonn, 2012). The ANC and government preferred public employment programmes to grants, but failed to scale them up significantly. The ANC increasingly articulated also a conservative commitment to‘family values’. The welfare state sapped familial as well as individual responsibility, party leaders alleged. The Minister of Social Development, Bathabile Dlamini, proclaimed that it was ‘important to strengthen families as the first line of response as espoused in African values’. 97 Most ANC leaders insisted that they were opposed to building a welfare state. Even whilst Skweyiya was at the helm as Minister of Social Development, the ANC resisted calls for the introduction of a Basic Income Grant or the introduction of grants for young men and women over the age of eighteen. We are not aware of any attempts to press for such reforms on constitutional grounds – presumably because the caveats in the constitution(with respect to both progressive realization and affordability) provide the government with a strong defence. The government has been mulling over two reforms of social protection: The reform of the pension system, to encourage people to save for their retirement, and the reform of the health care system through the introduction of a socalled‘national health insurance’ system. Neither reform has proceeded very far or fast. Neither reform has been justified on constitutional grounds. 4.2. Botswana Botswana also built an extensive welfare state by the 2000s, but this welfare state had multiple conservative features and the government and ruling party were strongly opposed to the rights associated with the global rights discourse. Assessments of Botswana’s welfare state are deeply divided. The World Bank’s Garcia and Moore summarised that‘ Botswana has one of the most extensive social grant 96 Debate on Appropriation Bill, Vote no.37 – Welfare, Wed 27 th May 1998, Hansard, col 3193-5, 3201. 97 Address by the Minister of Social Development, Ms Bathabile Dlamini, on the occasion of the Gala Dinner and Opening Session of the African Union(AU) Expert Consultation Meeting on Children and Social Protection Systems in Africa, Cape Town: http://www.dsd.gov.za/index.php?option=com_content&task=view&id=592&Itemid=1. Accessed 10 May 2014.

102 systems in the region, including separate programs for elderly people, orphans, and people living with disabilities’ (2012: 227). A jointly-researched World Bank study of Botswana refers to its‘mature and complex social protection system’(BIDPA& World Bank, 2013: x). Devereux concurs:‘B y African standards, Botswana implements an unusually comprehensive set of social welfare programmes for its poor and vulnerable citizens’(2007: 554). For the UNDP:‘ Botswana provides a comprehensive regime of social safety nets with high rates of coverage. These provide a comprehensive buffer against hunger.’ 98 For other scholars, however, Botswana has been distinguished by an enduring failure to include all of its citizens in the benefits of growth. Bar-On(2001) empha sized the parsimony of the country’s welfare policies, and concluded that ‘Botswana’s social assistance scheme fails many of the very poor completely, and those who do benefit from it are receiving steadily less of the nation’s increasing prosperity’ (2001: 264). Ulriksen(2011, 2012) contrasted Botswana with Mauritius. Both were multiparty democracies with open, middle-income economies and competent state institutions. Both invested heavily in public education, as well as public health and housing. Yet in the early 2000s Mauritius spent one-fifth of public expenditure on social insurance and social assistance whereas Botswana spent only between 1 and 3 percent on cash transfers and feeding schemes(Ulriksen, 2012: 1503; see also Ulriksen, 2017). There is merit in both sets of assessments. Viewed in terms of public expenditure on social protection, Botswana lags behind many other middle-income countries in Africa (and elsewhere), although less dramatically than Ulriksen and others have suggested. 99 Viewed in terms of the overall coverage of institutionalized programmes, however, Botswana seems more of a frontrunner, with provision far exceeding that in most other African countries. Coverage is extensive but expenditure modest because primarily because most benefits are set at parsimonious levels. In 2012, the old age pension in Botswana paid P250/month compared to R1200/month in neighbouring South Africa – when both their GDP per capita and their currencies were similar in value. Table 3.2: Social protection in Botswana Formallyemployed workers Informallyemployed workers Need Growing: large formal sector Existing coverage Expanding protection for illness and oldage Growing Workfare Current or prospective reforms Unknown Unknown Formal rights Informal rights Legislated? None None Strong commitment to workfare 98 http://www.bw.undp.org/content/botswana/en/home/countryinfo.html. 99 Total public expenditure on contributory pensions and safety nets amounted to 3 percent of GDP in Botswana in 2012/13(BIDPA and World Bank, 2013: xi). In the mid-2000s, total public expenditure on social benefits in Mauritius amounted to between 3 and 4 percent of GDP, with other(unidentified) expenditures on‘welfare’ pushing this share up to between 5 and 6 percent of GDP(Mauritius, 2008).

Peasants ‘Vulnerable groups’/ ‘Incapacitated’ households Diminishing, but still important Growing Drought relief; Farmer support programmes; drought relief; old-age pensions Pensions and grants for the elderly, orphans and ‘destitutes’ Unknown Child grants 103 None Weak None Strong sense of responsibility 4.2.1. Rights The Constitution of Botswana, dating from 1965, does not refer to any social or economic rights. Most of the country’s programmes of social protection have no legislative basis. Only contributory pension and provident funds and workmen’s compensation seem to be governed and regulated legislatively. 100 The central government takes responsibility, mostly through a Ministry of Local Government rather than through a dedicated Ministry of Welfare(or Social Development), 101 and welfare programmes are integrated fully into the government’s budget processes. But citizens of Botswana have no formal, legal rights to social protection. Botswana has not ratified any of the ILO Conventions concerning SP. 102 Botswana is also one of very few countries that has not acceded to or signed the International Covenant on Economic, Social, and Cultural Rights.‘ The Government of Botswana also has a poor reporting record in regard to many of the international treaties or conventions to which it is a party.’ 103 BDP Governments have a longstanding commitment to the reduction and ultimately elimination of poverty. This is evident in BDP election manifestos as well as the Government’s successive National Development Plans(NDPs). The Government also claims to have made the MDGs‘part of the national agenda to alleviate extreme poverty and its related socioeconomic problems’(Botswana, 2016: para 2.30). Th e 10 th NDP (200916) shifted the objective from poverty reduction to poverty eradication; a‘Poverty Eradication Roadmap’ was also published. In a series of dedicated progress reports (2004, 2010, 2015) as well as the 10 th and 11 th NDPs, the Government reported on progress made. The Government also declared its commitment to achieving the SDGs (Botswana, 2016: para 2.42ff). 100 The Retirement Funds Act(no 27 of 2014) replaced the Pension and Provident Funds Act(no 18 of 1987), which in turn had replaced Act no 17 of 1965. The Workmen’s Compensation Act(no 23 of 1998) governs work -related injuries. 101 And through the Poverty Eradication Coordination Unit(PECU) in the Office of the President. 102 http://www.ditshwanelo.org.bw. 103 http://www.ilo.org/dyn/normlex/en/f?p=1000:11200:0::NO:11200:P11200_COUNTRY_ID:103303.

104 Moreover, Botswana has no social protection policy or strategy. In 1980 it issued a National Policy on Destitute Persons, and revised this in 2002, but this provided only minimal assistance to very destitute people. Botswana does have a series of policy documents that discuss poverty reduction(and, more recently, poverty eradication). These are revealingly ambivalent about social protection. The most recent 11 th NDP(for 2017-22, published in draft form in late 2016) emphasizes the expansion of opportunities for the poor‘to help themselves’ through‘sustainable livelihoods’, through expanded employment and education(Botswana, 2016: par a 2.27). Chapter 9 covers‘Social Upliftment’, which seems to be the term preferred over‘social protection’. The chapter seems to suggest that government policies build‘on notions of individual and community rights and entitlements, as well as on the sta te’s responsibilities to all its citizens’( ibid: para 9.2).‘Empowerment’ programmes that help unemployed youth to establish small businesses are preferred over cash transfers. Workfare is preferred over grants. The spectre of‘dependency’ is ever -present .‘Although appreciable strides have been made to uplift the wellbeing of citizens, increasing dependence on support programmes has been observed, and the current relief programmes tend to compete with those that have an empowerment component’. Poor harmonisation of programmes also allowed‘multiple dipping’ by beneficiaries( ibid : para 9.22). Later:‘There is a need to reduce dependence on support programmes by up scaling the provision of life skills to empower beneficiaries. A deliberate strategy for graduating people out of such dependence will need to be developed with a view to making the schemes sustainable’( ibid: para 9.49). The country’s original national‘vision’ – Vision 2016, published in 1997 – repeatedly referred to rights, but to human rights, meaning primarily civil rights(Botswana, 1997). It mentioned responsibility often. More recently, the Government carefully avoids almost any reference to rights. Botswana’s 2010 MDG progress report mentioned rights only three times: once with reference to children’s rights to education(Botswana, 2010: 29) and twice with reference to domestic and sexual violence against women( ibid: 34 and 35). The 2015 MDG progress report mentioned human rights in passing, but included only one specific reference to rights, again with respect to education(Botswana, 2015). Despite the paucity of – even aversion to – any formal declaration of rights, the political elite in Botswana has since independence held to an informal conception of responsibility for the deserving poor, which implies that the deserving poor have some informal or weak rights. Drought and famine beset Botswana at the same time as Seretse Khama’s Botswana Democratic Party(BDP) sought to build a modern state and strip the chiefs of most of their formal powers. The combination of these factors pushed Khama and his party to articulate a doctrine setting out the respective responsibilities of citizens and the new ‘modern’ state. Faced with drought, the new state adapted existing understandings of chiefly authority and responsibility, generating a doctrine that represented the state as benign, with clear responsibilities to the poor, but at the same time emphasized the responsibilities of the poor themselves to contribute to their own welfare.

105 The doctrine was conservative in its idealization of rural society, wariness of social change and distaste for urban life. It emphasised the imperative of‘self help’ and was hostile to‘dependency’ on the‘’charity of the state. At the same time, it set out the goal of social harmony(notwithstanding inequality), and this meant that the state would need to assume some of the responsibility of providing for the poor, especially during periods of drought. Khama and the BDP identified four‘fundamental principles’ that w ould guide ‘National Development’. Each of these was said to be rooted in Tswana tradition. They were united in the Tswana concept of kagisano , which was usually translated as‘unity, peace, harmony and a sense of community’(Gulbrandsen, 2012; Seekings, 2 016c). Whilst the BDP sought to appropriate tradition, its leaders were not traditionalists. They did not, for example, promote traditional clothing or rituals(Gulbrandsen, 2012: 192). They were liberal in their preference for markets over the state, were conservative on many social issues(including both the family and sex). The poor had a responsibility to work hard and take advantage of whatever opportunities were open to them; they should not rely on government‘handouts’. Improved productivity was alw ays to be preferred to welfare. But the state also had a clear responsibility to the poor, which meant providing forms of poor relief to a substantial proportion of the population. Khama was succeeded by his vice-president, Quett Masire. Close to the end of Masire’s presidency, the BDP’s ideology was rearticulated in Vision 2016. Vision 2016 started by asking what kind of society Botswana should be in 2016, i.e. fifty years after independence. The answer was: educated, prosperous and just, as well as secure, democratic, moral and‘united and proud’. Vision 2016 spelt out that Botswana would be (by 2016)‘a compassionate and caring society, offering support and opportunity to those who are poor, and including all people in the benefits of growth’(Botswana, 1997: 8, emphasis in original). Poverty would be halved by 2007, and absolute poverty would be eliminated by 2016. This would be achieved primarily through economic growth. But ‘… there will be a social safety net for those who find themselves in poverty for any reason. … There will always be those who fall into poverty as a result of misfortune. The challenge to Botswana is to provide an adequate and dignified safety net for those who are in poverty in a way that does not encourage dependency and provides a means of escape from the poverty trap and a return to productive society’( ibid: 9, 24). Vision 2016 expressed concern over the possible‘deterioration of national values’. In a section on the‘give me’ attitude, Vision 2016 insisted on the importance of avoiding‘a culture of dependency’( ibid: 25). Poverty reduction required firstly job creation, then education and basic social services. Yes,‘Botswana must establish an effective social safety net’, but this was‘to support those who are genuinely unable to obtain a minimum standard of living through their own efforts, and to assist and encourage them to become selfreliant again’. Vision 2016 invoked the Tswana concept of‘ botho ’. ‘This refers to one of the tenets of African culture – the concept of a person who has a well-rounded character, who is well-mannered, courteous and disciplined, and realises his or her potential both as an individual and as a part of the community to which he or she belongs.“ Botho ” defines a process for earning respect by first giving it, and to gain empowerment by empowering others. It encourages people to applaud rather than resent those who succeed.

106 It disapproves of antisocial, disgraceful, inhuman and criminal behaviour, and encourages social justice for all.“ Botho ” as a concept must stretch to its utmost limits the largeness of the spirit of all Batswana. It must permeate every aspect of our lives, like the air we breathe, so that no Motswana will rest easy knowing that another is in need’( ibid: 2). In the 1980s, and especially 1990s, a discourse of concern over dependency became more and more pervasive within Botswana, at least among the political and economic elite. This might have fueled the retrenchment of welfare programmes, but(as we have seen above) welfare programmes continued and indeed expanded. It is likely that conservative norms and beliefs among the elite serve to constrain the benefits associated with public programs, but not to reel in the reach of the programmes. One section of the population of Botswana which seems to have been disadvantaged repeatedly and which continues to experience discrimination is the Basarwa, historically resident in the Kalahari. 4.2.2. Gaps Botswana’s welfare state has been shaped by almost fifty years of rule by the conservative – but also modestly benign – BDP. Under the BDP, and buoyed by rapid economic growth, the social contract in Botswana was fashioned(or refashioned) along broadly conservative lines, ensuring widespread but parsimonious inclusion in an otherwise(and increasingly) unequal society. The welfare state comprises both social insurance and social assistance(including workfare and feeding programmes). Whilst most of the 120,000 public sector workers are covered for old age(through the Botswana Public Officers Pension Fund), only a small proportion of the approximately 300,000 formally-employed, private sector workers are covered(the largest scheme being the Debswana Pension Fund, for mineworkers). Social assistance comprises a universal old-age pension from the age of 65, cash grants and in kind benefits for orphans, school and other feeding schemes, workfare and small grants for the truly destitute. Programmes expand during episodes of drought. Botswana’s welfare state is conservative not only in terms of parsimony and preference for in kind benefits, but also in terms of its design, which was rooted in strong commitment to the family. Welfare policies in Botswana have long revolved around three major components: workfare for able-bodied, working-age adults(originally food-for-work, more recently cash-for-work), enabling them to support their dependents also;‘supplementary’ feeding programmes for designated‘vulnerable groups’(children, together with pregnant and lactating mothers); and additional support for selected categories of people(initially support along the lines of poor relief for‘destitutes’, and more recently universal old -age pensions and support for orphans, i.e. for categories of people who have ceased to be supported adequately by kin and cannot support themselves). In contrast to neighbouring South Africa, there is no general child support grant paying benefits to poor mothers or caregivers. As of 2015/16, about 10 percent of the population received benefits through

107 the old-age pensions, orphans and destitutes programmes. But one-third of the population benefitted directly from workfare at some point during the year, and feeding programmes reached a similar proportion. Some gaps in the system have been pointed out by trade unions, which provide a more radical and programmatic opposition to the BDP than do most of the opposition parties. In an important 2007 document, prepared with assistance from FES, the Botswana Federation of Trade Unions(BFTU) acknowledged that social protection had wide coverage, but remained narrow: ‘The BFTU acknowledges that the role o f the State in providing some form of social security in Botswana is comparatively better in the SADC region. However, the current social security system is rather reactive, indirect and unsystematic. It is fragmented and hidden in various indirect social expenditure by the State. What is clear from this analysis is that though the majority of the very poor in Botswana are provided with some basic essential services to sustain their livelihoods, these schemes are not intended to provide skills to enable the poor to get out of the poverty trap. What then remains a challenge is finding innovative strategies that would eradicate poverty in line with Vision 2016. These strategies would have to preventive and holistic in nature and therefore targeted at eliminati ng risk and vulnerability.’(BFTU, 2007: 44). The BFTU’s specific demands focused on the formalization or institutionalization of the country’s programmes. In the 2014 elections, the opposition parties proposed large increases in benefits(for example, from Pula 250/month to 750/month for the old-age pension) and expanded coverage(through, for example, lowering the age of eligibility for the old-age pension from 65 to 60 years). They did not, however, propose substantial new programmes. Indeed, oppositio n leaders generally condemned the government’s workfare programmes, viewing them as mechanisms for the BDP to buy votes rather than as effective antipoverty programmes. In their criticisms of workfare, the opposition parties sought to tap into widespread anxiety over the programme, which critics derided it as atlhama ke go jese (‘open your mouth and let me feed you’). The opposition parties thus demanded and promised the creation of good jobs on a large scale rather than workfare(Hamer, 2016). 4.2.3. Rights and the history of social protection The welfare state in Botswana has its origins in the drought relief programmes developed at the time of independence, in 1966, when the country was beset with the worst drought and famine in memory. Whilst the outgoing colonial administration had been slow to respond, the incoming BDP government headed by Seretse Khama and Quett Masire was quick to secure massive food aid from the newly-established World Food Programme. Food was distributed through school and other feeding programmes, and through workfare. Drought recurred over the following decades, but the government expanded rather than contracted its school feeding programmes, persisted with most other feeding schemes and workfare programmes(expanding them during droughts), and

108 reformed its policy towar ds‘destitutes’ to ensure that most people were protected against the risk of extreme poverty. As we have seen above, the expansion of public provision was accompanied, and to some extent driven by a normative doctrine of public responsibility for the poor. Society – through the state – had a responsibility to the able-bodied poor to ensure opportunities for work so that they themselves could support their dependents(hence workfare). Society had a responsibility to support any elderly or disabled people, or children, whose families were unable or unwilling to support them. This doctrine implied informal rights, but not explicit or formal ones. In the late 1990s and early 2000s, it pushed the BDP to the introduction of old-age pensions, reforms of the destitute grants, and the introduction of packages for orphans. BDP governments were pushed to expand public provision in part by political considerations. Welfare programmes served to buttress electoral support for the BDP, especially in the countryside. In the 1994 elections, the BDP lost a lot of its support to the opposition parties. The introduction of old-age pensions(in 1996) and subsequent reforms were in part a response to electoral competition. Similar, in 2009 and 2014 the BDP and the incumbent president, Ian Khama(son of Seretse Khama) expanded workfare programmes in part to strengthen the image of benevolence and retain support. At no point was the expansion of social protection driven or even shaped by legal rights. Botswana was a clear case of reform without any conducive rights framework. 4.2.4. Rights and the future of social protection In other countries in Africa, international and donor organisations have been a source of pressure for the expansion of social and economic rights, in partnership with local civil society organisations. In Botswana, most international and donor organisations have a weaker presence, primarily because they do not have much financial muscle. International organisations will attempt to advise the government through reports on, for example, poverty-reduction, but these reports will typically be written in a way that downplays mention of rights – with the exception of UNICEF, which has pushed for a children’s right agenda(UNICEF, 2013). The World Bank’s 2015 report on Botswana, for example, seems to have referred to rights only once, in relation to property rights(World Bank, 2015b: 68). One NGO pushing for a human rights agenda is Ditshwanelo(the Botswana Centre for Human Rights), established in 1993. Ditshwanelo explains that: ‘We work in partnership with other rights -focused organisations in Botswana, such as those focusing on gender equality or the rights of those affected by HIV/AIDS. Complementing the work of our partners, we focus on issues least supported by others, including the Basarwa/ San(Bushmen), sexual minorities and domestic workers. We have a specific Children’s Rights Programme and an Outreach Programme in Kasane, northern Botswana, and conduct campaigns on issues such as abolition of the death penalty in Botswana and the development

109 and rights issues related to the relocation of the Basarwa/ San from the Central Kalahari Game Reserve. ’ 104 Ditshwanelo’s website reveals little concern with social and economic rights. Neither it, nor the opposition parties, nor international and donor organisations, and certainly not the governing BDP seem to be pressing for the formal recognition of social and economic rights. 4.3. Zambia There is a clear need for social protection reforms in Zambia: Poverty is widespread; social and economic change compounded by AIDS have resulted in growing numbers of ‘vulnerable’ individuals who are not being supported by kin; and the coverage of existing programmes remains very uneven. The political elite, however, is for the most part not very sympathetic to the poor. Table 4.3 shows the existing contributory and non-contributory social security and social assistance schemes that exist in Zambia, that address risks faced by incapacitated households, peasants(or small-scale farmers), informal wage workers, and formal wage workers. It also shows the coverage of these programmes and describes whether the existing frameworks that support the implementation of the programmes are formal (backed by legislation) or informal(not legislated but supported by the government through the national budget). It further describes recent reforms to the programmes. 4.3.1. Rights The Zambian constitution directs the state to improving livelihoods but does not formally recognize any social or economic rights. Some government documents express a stronger commitment, but these seem to have only modest influence on actual policymaking. The inclusion of some social and economic rights, including a right to social protection, was considered in deliberations over constitutional reforms in 2016. Proposals to broaden the Bill of Rights were rejected in a referendum. The Government has incorporated some international law into national legislation. For example, the UN Convention on the Rights of Persons with Disabilities gave rise to the Persons with Disabilities Act No 6 of 2012. Social insurance is governed by legislation: The National Pension Scheme Act(Act 40 of 1996), the Workers’ Compensation Act(Act 10 of 1999), the Public Service Pensions Act (Act 35 of 1996) and the Local Authorities Superannuation Fund Act(Act 284 of 1997). A new, integrated Social Protection Bill was supposed to have been finalised in mid-2017, but seems to have been delayed. It was reported that the Bill would, for the first time, provide for social assistance. Hitherto, social assistance had been expanded through executive not legislative action. 104 http://www.ditshwanelo.org.bw/welcome.html.

110 Table 4.3: Social protection in Zambia Need Existing coverage Formallyemployed workers Small proportion of the population: old age, poor health or incapacity; etc Workmen’s Compensation and Pensions for about 800,000 people (NAPSA) Perhaps onethird of the Informally- working employed population; old None workers age, poor health or incapacity; etc Approximately one half of the FISP(with population; FSP) covers Peasants needs about 0.5m seasonal households unemployment, drought, etc ‘vulnerable’ individuals/ ‘incapacitated’ households Risks of disability, old age, poor health, unemployment SCTs: 200,000 households NAPSA: National Pension Scheme Authority FISP: Farmer Input Support Programme; FSP: Food Source: Based on compilation by H. Siachiwena Current or prospective reforms Formal rights Proposed national health insurance scheme Strong (legislation) Formalisation: See Table 4.4 Target coverage of about 0.3m workers Reduced coverage for FISP(but small expansion of FSP) Weak None Expansion of coverage Weak Security Pack; SCT: Social Cash Informal rights Strong(history of active trade unions) Weak Strong (especially during drought); because of political competition Weak Transfer Social and economic rights are alluded to in other government documents. Vision 2030 envisaged that Zambia would become‘a nation that promotes and provides sustainable security against deprivation and extreme vulnerability by 2030’. Specific targets include ‘ensuring that incapacitated and low capacity households have sufficient liveliho od security to meet basic needs, and are protected from the worst impacts of risks and shocks’, reducing the number of vulnerable and incapacitated households, and ending child labour by 2030(Zambia, 2006a). Successive NDPs have commited the Government to the reduction of poverty. The Fifth NDP(2006-10) included a chapter on social protection which declared that social protection was necessary for‘meaningful and sustained economic growth, reducing poverty, and promoting equity by reducing income inequali ty’. It also described social protection as contributing‘to the fulfilment of human rights of refugees and internally displaced persons by providing services that strengthen and secure their access to basic rights and freedom’(Zambia, 2006b). The Sixth N DP (2011-15) included a chapter on social protection that focused on promoting the livelihoods and welfare of the poorest and those most vulnerable to risks and shocks. (Zambia, 2011). The Revised Sixth NDP(2013-16) included a chapter on Social

111 Protection and Disability in which the Government committed to‘implement policies, programmes and activities that promote social assistance, social insurance/security, livelihood and empowerment … and to safeguard the livelihoods and welfare of people suffering fro m extreme poverty and/or vulnerable to risks and shocks’(Zambia, 2013). Neither the Sixth nor the Revised Sixth NDP referred to social or economic rights. The new Seventh NDP(2017-2021) refers to social protection in various chapters, and envisions Zambi a as a‘Nation with Equality and Rights, which will have firmly ensured provisions of livelihood and empowerment through SP schemes’ by 2030(Zambia, 2017: 49). This NDP also states that‘Government will in line with SDG number 1 implement nationally appropriate SP systems to achieve substantial coverage of the poor and vulnerable. This will ensure that all men and women, in particular the poor and the vulnerable, have equal rights to economic resources and access to basic services’. The fullest statement of the Government’s vision for social protection is in the National Social Protection Policy(Zambia, 2014). The Policy declares that‘all Zambian citizens have the right to social protection’(p11) but does not explain how or why this is the case. In her foreword the Minister of Community Development wrote that‘Government has a fundamental responsibility of ensuring the wellbeing of its citizens’. Successive manifestos from the Patriotic Front make brief references to rights to social protection. The PF ’s 2011 2016 manifesto stated that‘the PF recognizes the international obligations to social protection, expressed in the Universal Declaration of Human Rights and other subsequent instruments’. The 2016 2021 manifesto referred to‘the right of persons wi th disabilities to social protection’. Overall, Zambian governments have not recognized rights to social protection, and they have expanded social protection programmes very slowly. 4.3.2. Gaps Zambia has a population of about 15 million people living in about 3 million households. Of these, approximately 6 million people were employed, of whom more than 5 million were in the informal economy. Most of the small, formally-employed population(of less than 1 million people) but none of the much larger informally-employed population are covered by social security schemes. The National Pension Scheme Authority(NAPSA) is compulsory for all employed persons and‘covers all categories of employment for as long as a contract between the employer and the employ ee exists’. 105 NAPSA currently provides coverage to 677,000 workers but plans to increase coverage to 1 million workers by the end of 2017, by extending coverage to informal wage workers. 106 Similarly, it is mandatory for employees to register their employers with the Workers’ Compensation Fund Control Board(WCFCB) which administers the Workers’ Compensation Fund (WCF). The WCF provides compensation for workers disabled by occupational accidents or diseases. 107 About 112,000 workers in the public sector are covered by the public 105 http://www.napsa.co.zm/who-should-register-with-napsa/ 106 http://www.napsa.co.zm/message-from-the-director-general/ 107 http://www.workers.com.zm/index.php/about-us

112 sector pension fund. Almost the only informal sector workers in NAPSA are a small number of domestic workers who are registered by their employers: Only 3 percent of the approximately 98,000 domestic workers are registered with NAPSA; none are registered with the WCFCB, although an estimated 16 percent of domestic workers are helped by their employers to pay for health expenditure. Almost the only informal sector workers registered with the WCFCB are a small number of workers in hazardous jobs: An estimated 9 percent of casual saw mill workers are registered with the WCFCB. Small-scale farmers account for the largest number of informal workers. About 2.8 million smallscale farmers are‘employed’(i.e. self -employed) informally while only 100,000 small-scale farmers are employed formally(presumably meaning that they are large-scale or commercial farmers paying taxes and so on). None of the informal small-scale(i.e. peasant) farmers are registered with either NAPSA or the WCFCB. Furthermore, peasant farmers are responsible for meeting their own health expenditure(Phe Goursat and Pellerano, 2016). Table 4.4 shows the existing coverage of social security for three categories of informal wage workers, expressed as percentages of the total of workers employed in each category. Table 4.4. Coverage of Social Security for Informal Wage Workers, Zambia Domestic Workers Small Scale Farmers Casual Saw Mill Workers NAPSA 3% 0% N/A WCF 0% 0% 9% Payment of health expenditures by 16% 0% 35% employers NAPSA: National Pension Scheme Authority. WCF: Workers’ Compensation Fund Source: Phe Goursat and Pellerano(2016) Contributory schemes clearly cover not only a low proportion of the population, but also a low proportion of the working population. Over the past ten years, social assistance programmes have expanded, although their reach only partially closes the coverage gap. The Social Cash Transfer programme was expanded rapidly in 2014, but still only reached 145,000 households at the end of 2014 and 180,000 households at the end of 2015. The Government set a target of 500,000 households by 2016. Much smaller numbers of people benefitted from other social assistance programmes. In total, only about one in ten Zambian households benefitted from any of the cash transfer programmes. A slightly larger proportion of households – but still a minority – included children who were fed through school feeding schemes. Many more households benefitted from FISP, which provided subsidized fertilizer and other inputs for farmers. Overall, however, Zambia’s social protection system has glaring gaps, with coverage limited to two small minorities: firstly, a minority of the very poor(through cash transfers) and, secondly, the non-poor in formal employment(through social insurance).

113 4.3.3. History of Rights and Social Protection The slow expansion of social protection in Zambia is not due simply to the lack of any formal recognition of social and economic rights. The expansion of social protection in Zambia has actually been stalled, repeatedly, because of the ambivalence or even opposition of political elites. Whereas in Botswana there has been a widespread ideology of responsibility for the poor, some senior government ministers in Zambia have articulated the view that it is the poor themselves who are responsible for their poverty (and for rising out of it). This was most obvious in the mid-2000s, when the Minister of Finance and National Planning in the then Movement for Multi-party Democracy(MMD) go vernment, Ng’andu Magande, blocked the expansion of pilot cash transfers programmes. In Magande’s view, 108 poverty results from a lack of planning and effort, and the solution was simply better planning and hard work: ‘If only your children can learn from an early age when they are young, that money doesn’t fall from trees. It falls from your own work. Then these children are going to say, I have to aim to be somebody to do something to earn the money.’ He himself had understood this from an early age: ‘When I was five years old, my parents were already telling me, you will get old, now, whatever we give you, you must look after it. If we give you a chicken, make sure when you come home from school, you look after it. If we tell you to go in the bush, you are looking after cattle, when you bring them, count them. Because if … one[cow] is left in the bush, and it’s eaten[then you will remain poor]. ’ If parents are negligent in teaching their children, then their children will end up poorer than them. The state should beware supporting the poor – including even poor, orphaned children – because the poor will end up dependent on the state. Even orphans could be empowered to support themselves through their own efforts: ‘Suppose they inherited goats from the ir parents: if they are children, one of them is above eight years, …, these children can even look after a goat, or two goats, then they milk the goats, and they have that milk. They are also beginning to learn to be responsible for their life, that once you have property, you look after it.’ Even poor elderly widows should not get pensions: They had clearly failed to prepare for their old age, probably through drinking. In Magande’s view, cash transfer programmes worsen poverty by discouraging the poor both to plan and to work hard. Poor people simply become dependent on the state, and lose all initiative. This might seem like a neoliberal argument, in its scepticism about the state and its emphasis on individual investment and work. But the argument is as much conservative as neoliberal. There is little advocacy of expanding provision through financial markets. Social pensions are not denounced in favour of individual contributory pension accounts. Rather, poor people – including the elderly – are urged to provide for themselves and their kin independently of the market as well as the state, including 108 This section is based on an interview with Magande, by Abigail Kabandula and Singumbe Muyeba, Lusaka, March 2014; see Kabandula and Seekings(2016).

114 through investment in livestock. Magande’s worldview holds up a nostalgic – and surely romantic – view of a past society in which peasant farmers prospered on the basis of their hard work, discipline and planning. The opposition Patriotic Front was more sympathetic to social protection than the MMD government in the mid-2000s. A disparate group of PF leaders concurred around a broadly social democratic view of social protection. They pushed for the expansion of the Social Cash Transfer programme, first to the entire country(rather than selected districts) and then to a rising proportion of the poor in each district. Under Michael Sata(i.e. 201114), additional funding was allocated to the SCT programme, which was expanded rapidly. After Sata’s death in late 2014, however, the PF slid towards a split. By the 2015 elections, most of the‘social democratic’ wing of the PF had defected to opposition parties, leaving the PF firmly under the control of Edgar Lungu, who showed no sign of having any interest in social protection. Throughout this period, international organisations and donor agencies pushed for the expansion of social protection. The UK’s DfID and other European donors funded much of the SCT pilot programme and repeatedly sought to persuade the government to roll out the programme countrywide, with substantial Zambian funding. Social assistance was very much an international project – with some buy-in from the Ministry of Community Development. The ILO costed a package of social protection programmes and concluded that a basic package was affordable. International organisations and donors sponsored ministers and officials to go on study tours, supported local civil society organisations that were lobbying for good government and social protection, and helped(directly or indirectly) to shape the National Social Protection Policy. Studies of the politics of social protection in Zambia tend to conclude that international organisations and donor agencies did not wield much power. They were unable to push successive Zambian governments to the kinds and pace of reforms they themselves would have liked to see(Kabandula and Seekings, 2016; Siachiwena, 2016, 2017; Pruce and Hickey, 2017). 4.3.4. Future for Rights and Social Protection During 2016-17 there was some discussion of reforms to both social assistance and social insurance, to fill some of the glaring gaps in the coverage of the social protection system in Zambia. The reforms involved several ministers: Labour and Social Security for social insurance, Community Development for social assistance, and Health for national health insurance. The promised, integrated Social Protection Bill did not materialize as promised. Eventually, in December 2017, a National Health Insurance Bill was presented to Parliament. In practice, the three ministries seem to have been presiding over three separate reform initiatives that proved too difficult to integrate into a single framework. The Ministry of Health seems to have pushed most effectively, and placed health insurance reform firmly on the agenda. The Ministry of Labour and Social Security had some influence. Together with the ILO and other international organisations, the Ministry pushed to expand social protection through the partial formalisation of the informal sector.

115 The Ministry of Community Development seems to have had the least influence, and there is little prospect for an immediate and rapid expansion of social assistance. The initiatives for formalizing informal work provide perhaps the most important lesson to be learnt from the Zambian case. The data in Table 4.4 above suggests that there is some scope for the expansion of social security – covering pensions, compensation for incapacitation and health insurance – among domestic workers and saw mill workers, and perhaps also among construction workers. Extension to other categories of informal workers in urban areas(such as market traders, bus and taxi drivers) might be feasible through these workers’ own orga nisations. The lack of organization among self-employed street vendors and hawkers would seem to preclude extension to them. Peasants – who comprise by far the largest category of informal‘workers’ – do not benefit from existing formal social insurance schemes, and they do not appear to be on the radar screen of the Ministry of Labour and Social Security. It does not appear that there are immediate efforts to address social risks faced by peasants. It is not clear either how the progressive realization of rights to social protection would extend to this sector. Reforms to other interventions targeting small-scale farmers that are implemented by the Ministry of Agriculture and the Ministry of Community Development and Social Services suggest that coverage to these workers may be contracting rather than expanding. The Zambian government would reduce the number of the Farmer Input Support Programme(FISP) beneficiaries by 600,000 in September 2017 109 to improve targeting – as part of reforms to shift from giving inputs to farmers to using an electronic voucher(e-voucher) payment system. 110 At the same time, the number of beneficiaries accessing the Food Security Pack(FSP) remained at 40,000. The ILO is the lead agency in an initiative that aims to extend and promote social protection amongst Micro Small and Medium Enterprises(MSMEs) under the Zambia Green Jobs Programme(ZGJP)(Pellerano and Matandiko, 2016). The ZGJP project is a partnership between the United Nations system(including the ILO, FAO and other agencies) and the Government of Zambia with funding provided by the Finnish government. It aims to promote the potential for job creation in Zambia’s building construction industry( ibid: 2). Through the ZGJP project, the ILO organized national consultative meetings and a national policy dialogue with representatives of Government, social protection providers, sector associations, employers’, workers’ and civil society organizations. Five outcomes of the meetings and dialogue are worth emphasizing: 1. The Government committed its support to appropriate policy options that will promote the extension of social protection to the MSME sector. 2. Proposals were submitted to the draft Social Protection Bill to allow the Minister of Labour and Social Security to introduce social security schemes with appropriate benefits and incentives to enhance coverage of the informal economy. 3. Memoranda of Understanding were signed with the National Pensions Scheme Authority(NAPSA) and the Workers’ Compensation Fund Control Board(WCFC B) to jointly develop social protection and registration campaigns among employers 109 https://www.africanfarming.com/thousands-cut-fisp-new-farming-season-starts-september-1/ 110 ibid

116 and workers in the construction sector in collaboration with employers’ and workers’ organizations. 4. These Memoranda of Understanding were accompanied by campaigns to raise awareness to increase the low level of knowledge and understanding of social security services and benefits among MSME business owners and workers, including among members of the National Association for Small Scale Contractors, Copperbelt Sawmillers and Ti mber Growers’ Association, and Zambia National Association for Sawmillers. 5. Social security mobile offices were set up to make the registration process easier, while regulations to allow MSMEs in the construction sector to register with pension schemes were eased( ibid). In 2016, the ILO published a report that set out to provide a summary of the lessons learnt, the challenges and opportunities for extending the coverage of social protection to informal sector workers, particularly domestic workers, small scale farmers, construction workers, and saw mill workers. The report was explicitly framed in a discourse of rights, invoking both the Universal Declaration of Human Rights, the International Covenant on Economic, Social and Cultural Rights and the ILO’s Social Protection Floors Recommendation( ibid). In Zambia, elections in both 2015 and 2016 were closely fought, with the PF presidential candidate defeating the UPND candidate very narrowly. Would policies have developed differently had th e UPND candidate won? Some clues can be found in the UPND’s 2016 election manifesto. The manifesto was vague about its approach to both social security and social assistance. The manifesto included a very short paragraph on labour which stated that‘the ri ghts of workers will be high priority for a UPND government. We will uphold the principle of equal pay for equal work. We will also ensure that Government sets a high standard as an employe’(UPND, 2016: 9). The manifesto identified job creation and creating business opportunities as the most important intervention to Zambia’s economic challenges. However, it did not address risks associated with informal wage earners. Regarding social assistance, the UPND promised to‘put in place measurers to protect the vulnerable in our society which includes women and youth, ensuring they can fully participate in our economy and contribute to their own betterment as well as our national development’( ibid: 10). Yet the party did not specify the measures that it would implement. The UPND president, Hakainde Hichilema, is on record stating that education and agriculture are more appropriate forms of social protection than social cash transfers (see Pruce and Hickey, 2017). Even then, some UPND MPs who served in the 2011 to 2016 Parliament were advocates for the expansion of social cash transfers and collaborated with the now defunct Platform for Social Protection Zambia Chapter to establish a Parliamentary Caucus on Social Protection(see Siachiwena, 2016). It seems that the populist PF, with strong support on the Copperbelt and in Lusaka(as well as in the Northern and Eastern parts of the country), attached more importance than the opposition UPND to the extension of social protection to informal sector workers.

117 Neither the PF nor the UPND, however, attached much importance to social protection in their manifestos or campaigning for the 2015 elections. The PF has been widely regarded as a populist party, but it seems that Zambian parties are unhappy to go the voters with bold proposals on welfare reform. It is possible that Zambians are still wary of overlystatist projects, given their previous experience with Kaunda prior to 1991. In December 2017, the Zambian Government did introduce a National Health Insurance Bill, with the purported objective of improving access to and funding for health care. 4.4. Senegal In some respects, Senegal is not unlike Zambia. In both countries there is a significant minority of formally or informally-employed non-agricultural workers. Both countries has an even larger number of peasant farmers(although Senegalese farmers might be more immersed in markets, and hence more vulnerable to traders and moneylenders – see Waterbury, 1987). But the two countries have very different policy regimes, because discourses of rights and practices of social protection differ between Anglophone and Francophone countries. This section examines Senegal as an example of a Francophone case. In general, Francophone African countries are said to be‘lagging b ehind in terms of the recognition and enforcement’ of economic, social and cultural rights(Diallo, 2016: 396). This is in part because Francophone countries are supposed to recognized the major commitments of French law, but in practice fail to do so( ibid: 397). Unfortunately, they also lag behind Anglophone countries in terms of the existing literature on social protection. Within Francophone Africa, Senegal stands out in that‘basic civil and political rights’ are relatively entrenched. Economic, socia l and cultural rights might remain‘minimal and often indirect’ but are not entirely‘illusory’( ibid: 396). Francophone African countries tend to have more active peasants’ organisations, and Senegal has enacted legislation on food sovereignty. Senegal’s social protection policies share many of the characteristics of systems in other Francophone countries, but are also on the more extensive end of that range. Table 4.5: Social protection in Senegal Need Existing coverage Formallyemployed workers Minority of population Some pensions. Limited health insurance, through CSS, IPRES, FNR, mutuelles de santé Current or prospective reforms ? Formal rights Informal rights Weak?

Informallyemployed workers Peasants ‘Vulnerable’ individuals/ ‘Incapacitated’ households Minority of population Majority of population Growing None None ILO STEP programme; Plan Sénégal Émergent ? Expansion of PNBSF None None weak 118 ? ? ? As across most of Francophone Africa, social protection in Senegal has its roots in the years following the Second World War, when France’s colonial subjects enjoyed some of the rights of French citizens. Prior to 2001, Senegal’s constitutions bore the imp rint of French constitutions which provided weak, vague and qualified recognition of social and economic rights. The preamble to France’s 1946 constitution – which was supposedly recognised in the constitutions of Senegal – emphasised the right of‘the chi ldren, the mother, and the elderly workers to health protection, material security, rest and leisure’. In addition,‘each person whom, because of his age, his physical or mental condition or his economic situation, is unable to work has the right to receive adequate social assistance from the state’(quoted in ibid: 402). These provisions have been interpreted as non-binding or non-justiciable principles of state policy( ibid: 403). France itself revamped its social security system in 1946. Over the following ten years there was ongoing debate over the application of these policies in its colonies, focused on eligibility for the benefits(including especially family allowances) enjoyed by French workers. The 1952 Code du Travail extended eligibility for family allowances to all formallyemployed workers in the colonies(but not to the‘customary’ labour employed by West African farmers)(Cooper, 2012: 182). Rights for workers were institutionalised through social security funds that were generally controlled by employers and workers, with autonomy from the state: The Senegal Family Allowance Compensation Fund(1956, renamed the Social Security Fund or CSS in 1973); the Institution de Prévoyance Retraite du Sénégal(IPRES, originally established 1958, renamed IPRES in 1975) for private sector employees; the Fonds National de Retraite(FNR, National Retirement Fund) for public sector workers; and the Sickness Insurance Institutions(Institutions de Prévoyance Médicale, IPM, 1975)(Diop, 2003; Daffe and Diene, 2017). Social protection was limited to formally-employed workers and their dependents for the rest of the twentieth century. There was no provision for the rural or urban poor. The Senegalese elite articulated commitment to rights in general. The Senegalese president(Senghor) and judge president(Mbaye) played important roles in the 1981 African Charter, which did not specify any substantive social or economic rights for the poor(see above). In 2000, the Senegalese Democratic Party(under Abdoulaye Wade) defeated the long-incumbent socialist party. A new constitution was adopted in 2001. The 2001 constitution integrated the African Charter into its preamble and accorded the preamble the same legal status as the rest of the constitution(Diallo, 2016: 407-8). The constitution included a Bill of Rights that recognised rights to property, to health and

119 education, to strike and to form trade unions. Diallo describes Article 25, on wo rkers’ rights, as including a right to social protection( ibid: 408-9) – but the actual Article stipulated only that‘ Specific laws establish the conditions of assistance and of protection which the State and the enterprise accord to the workers ’, 111 which seems to be a very vague right to anything. The constitution also recognises international agreements. In 2012, Wade was in turn ousted by Maky Sall and his Alliance for the Republic. A new constitution was drafted in 2013. The draft extended somewhat the range of social and economic(as well as cultural) rights, but was not adopted(although some elements were incorporated into constitutional amendments in 2016). Although there is no legal basis for this, the general presumption is that social and economic rights are not justiciable. There have been a few cases involving rights, mostly concerning civil servants and the state(as is also the case in Benin). In the early 2000s the social security systems in Senegal(i.e. IPRES, CSS and FNR) covered(at most) about 20 percent of the population(cited in Hodges, 2008). This was a higher proportion than in most other Francophone African countries, and may have been an exaggeration. The ILO website more conservatively reports that‘coverage remains limited: security schemes apply to only 6 per cent of the active population(workers, salaried employees of the formal public and private sectors) and mutual benefit organizations to only 2 to 3 per cent.’ 112 The World Bank reported that‘formal social security coverage’ reached only 12 percent of the population, comprising 6 percent of the population covered by a formal pension, 3 percent receiving social security benefits and 3 percent having some form of health insurance(World Bank, 2013). Few poor and informal sector workers have access to the social insurance system(see further Daffe and Diene, 2017). In the 2000s, the economy grew slowly, and poverty declined slowly, from 55 percent in 2001 to 47 percent in 2011. In 2002-03, severe drought prompted policy reforms. The government committed to heavy expenditure on drought relief through interest rate subsidies and debt forgiveness. Expenditure rose from about 0.15 percent of GDP to 0.5 percent of GDP, but this was untargeted(World Bank, 2013). In the face of fuel and food price‘shocks’ in the later 2000s, the government introduced substantial subsidies on basic foodstuffs(rice, wheat, and milk), butane/natural gas and electricity, at a cost of 2.4 percent of GDP in 2008( ibid). The cost of poorly targeted and often distortionary general price subsidies reportedly rose to 3 to 4 percent of GDP in the late 2000s. The combination of economic growth, the possible reallocation of domestic resources from subsidy programmes and donor funding led international agencies to assess that there was‘fiscal space’ for the expansion of social protection. The costs of an expanded social protection system in Senegal were estimated by teams from the ILO, UNICEF and elsewhere. Pal et al.(2005), for the ILO, costed a universal old-age pension for all men and women above the age of 65 at between 0.3 and 0.7 111 https://www.constituteproject.org/constitution/Senegal_2009.pdf?lang=en. 112 http://www.socialprotection.org/gimi/gess/ShowCountryProfile.action;jsessionid=l7kQYQnJ6ypwhmVfTmnypg18RpGlWwNpbhjgt KYw8tnRJRsGw2d8!79209976?id=308.

120 percent of GDP depending on the value of the pension(varying from a modest US$0.50/day up to about$1.day, adjusted for purchasing power). A universal child benefit was costed at 2.3 percent of GDP in 2005(dropping to 1 per cent over time); a child benefit targeted on orphans would cost much less. Pal et al. also costed a targeted grant to the poorest 10 per cent of households, with benefits of just under US$14/household/month(modelled on an experimental cash transfer programme in Zambia). The cost of this would initially be about 0.6 percent of GDP, falling over time (Pal et al., 2005: 24, 30-31, 36). These cost estimates were updated and extended in 2008(ILO, 2008a: 6-7). A more generous child benefit scheme was costed by UNICEF, in conjunction with the Overseas Development Institute(ODI). Two versions of a child benefit cash transfer programmes were costed. A universal child benefit, payable for all children up to the age of 14, with benefits set at 30 per cent of the extreme(food) poverty line, would cost 6.4 percent of GDP. A targeted child benefit, limited to children in households below the poverty line, would cost only 3.7 percent of GDP(Handley, 2009). Subsequent studies costed cheaper programmes. A child benefit programme was costed at 1.7 percent of GDP by Samson and Cherrier(2009, cited in Schnitzer 2011). Schnitzer (2011: 18), who assessed that Samson and Cherrier’s proposals were‘likely not to be affordable in the current situation of recovery from the economic crisis and likely fiscal readjustment and budget austerity in the few years to come’, proposed a more modest programme, limited to fifteen very poor districts, that would cost only 0.55 percent of GDP – i.e. substantially less than the 3 percent of GDP spent on food and fuel subsidies. In summary, a series of studies costed programmes from as little as 0.3 percent of GDP to more than 6 percent of GDP. Senegalese governments nominally committed themselves to the expansion of social protection. In 2003 the Government established a National Commission on Social Dialogue to discuss social protection strategy. In 2005, Senegal adopted a National Social Protection Strategy(Stratégie Nationale de Protection Sociale, SNPS) for 2005-2015. This was incorporated into the second Poverty Reduction Strategy Paper(PRSP2, for 2006-10). The primary focus of the SNPS seems to have been the extension of health insurance, to cover one half of the population by 2015, and the establishment of some kind of system to insure rural populations from the risks of natural disasters(Hodges, 2008). Senegal’s third Poverty Reduction Strategy Paper(PRSP3), titled Stratégie Nationale de développement Economique et Sociale(SNDES), 2011-15, also included a notional commitment to social protection. The World Bank identified twelve programmes in Senegal in 2011. The largest were the Food Security Commissariat(Commissariat à la Securité Alimentaire, CSA), which provided food aid assistance to about 3.6 million people, and two school lunch programmes( the WFP’s school lunch program, Cantines scolaires, Programme alimentaire mondial[PAM] and the National School Lunch Program, Programme d’alimentation scolaire[DCaS]) which provided school lunche s to about 780,000 children in vulnerable rural and peri-urban areas. A pilot Cash Transfers for Child Nutrition program(Nutrition ciblée sur l’enfant et transferts sociaux, NETS) distributed cash grants to mothers of vulnerable children under 5 years of age. The WFP also operated a Vouchers for Food pilot program(Bons d’Achat, PAM CV) which addressed food

121 insecurity among vulnerable households, and a temporary workfare programme(Food for the Creation of Sustainable Assets). The total cost of these programmes, excluding the WFP workfare programme, came to an estimated 0.27 percent of GDP, two-third of which was funded by donors(World Bank, 2013). In addition, the Sesame programme, launched in 2006, provided free access to medical services for all elderly persons over the age of 60(Hodges, 2008). These programs were operated by several ministries and agencies, with the Ministry of Family and the Ministry of Social Action and National Solidarity responsible for most programmes focused on vulnerable groups. The government elected in 2012 emphasized social protection in its rhetoric. The new government established a dedicated office(the‘ Délégation générale à la protection sociale et à la solidarité nationale ’, DGPSN), reporting to the president, as well a s an Inter-ministerial Steering Committee on Social Protection, which included civil society and development partners, to improve co-ordination(World Bank, 2014; Daffe and Diene, 2017). It published a Plan Sénégal Émergent, which included explicit commitments to expand both contributory and non-contributory social protection, although not within any rights framework(Senegal, 2012: 80-82). In December 2012, at a three-day West Africa ‘Exchange and Learning meeting’ organized by the APSP in Dakar (co-hosted with CONGAD, i.e. the Conseil des ONG d’Appui au Développement) , the Minister of Labour ( Mansour Sy) described social protection as a‘basic human right’. 113 PRSP3 seems to have been modified, to cover 2013-17; it is not clear whether this entailed any change in the notional commitment to social protection. 114 In October 2013, the Prime Minister stated that her government's priorities included reforming social policies(including cash transfer programmes) to address vulnerability(World Bank, 2014). In April 2015, the Senegalese government co-hosted an International Seminar on Social Protection in Dakar, in partnership with the AU, UNDP and the government of Brazil. The event brought together participants from twelve African countries and Brazil, to discuss especially Brazil’s Bolsa Familia programme. 115 In 2013, the government initiated a pilot conditional cash transfer scheme, the Programme National de Bourses de Sécurité Familiale(PNBSF). This was piloted to about 50,000 households in 2013, and then rolled out – with the financial backing of the World bank – to about 300,000 households by 2017(out of a national population of 15 million people, or at least 1.5 million households). The benefit was set initially at about US$45 per quarter(i.e. 25,000 FCFA), i.e. the equivalent of US$15/month. The programme is not presented in terms of rights, however; it is a poverty-reduction programme. Senegal has a longer history of peasant organization and mobilization than Zambia, perhaps because of their long participation in the market economy. The Senegalese Conseil National de Coopération Rurale(CNCR), an affiliate of La Vía Campesina, mobilized strongly for legislation protecting peasants. Legislation in 2004 protected the 113 Reference? 114 Reference? 115 http://www.undp.org/content/undp/en/home/ourwork/development-impact/south-south-cooperation/featuredstories/featured-stories221.html .

122 food sovereignty of Senegal as a country whilst also committing the state to the rather vague goal of ensuring that peasants enjoyed a standard of living sufficient to meet their basic needs. An evaluation in 2013 reportedly found, however, that this legislation had not been implemented, and that the government had continued to neglect the peasant sector(Claeys, 2015b: 37-9). 4.5. Lessons from the country case-studies Both the need for and provision of social protection varies considerably between these four cases. South Africa is, in economic terms, a post-agrarian society: Whilst many poor households live in rural areas, very few are engaged in – or might engage in – peasant agriculture. The South African poor depend overwhelmingly on the government’s social assistance programmes and(to a much lesser and declining extent) remittances sent by kin in urban areas. Botswana is becoming a post-agrarian economy, in which the likelihood of achieving livelihoods through peasant agriculture has declined and continues to decline. Zambia and Senegal, however, remain very agrarian, with large proportions of the population engaged in peasant agriculture. Senegalese peasants are much more engaged in the market economy than their Zambian counterparts, but poverty is widespread in the rural areas of both countries. Conversely, South Africa has a much larger formal sector. These differences frame the risks against which people need to be protected and help us to understand the salience of formal and informal rights discourses. In neither South Africa(prior to 1996) nor Botswana was the expansion of social protection accompanied by any explicit commitment to social and economic rights. In South Africa, social protection was initially introduced for white(and coloured) citizens, and denied to the African majority. Some programmes were extended to the African population in the 1940s, but with deep racial discrimination in benefits. Racial discrimination in benefits was slowly abandoned in the 1980s. Whilst there were always some activists employing a discourse of rights, at no time was a discourse of rights dominant among reformers. Only in 1996, in the new democratic constitution, were social and economic rights, including to income security, formally recognized. By then, South Af rica’s welfare state was already largely built. Subsequent to 1996, the major reform – the introduction and expansion of Child Support Grants – entailed in effect the deracialisation of existing support for poor families with children. The explicit emphasis on rights in the South African constitution reflected the nature of South Africans’ struggle for democracy as a struggle against the denial of de facto rights through racial discrimination. In Botswana, also, the expansion of social protection was not associated with any explicit discourse of rights. Social protection grew out of drought relief. Public provision was bound up with an understanding of the state as a modern state, but it was rooted in an ideology of responsibility(on the part of elites, leaders and the better off) not of rights(on the part of poorer citizens themselves). In Botswana, unlike South Africa, the expansion of social protection did not generate a discourse of rights, presumably because there was no sign of any denial of access to social protection to any one section of the population

123 (with the possible and notable exception of the Basarwa). In Botswana, not even the opposition parties or civil society employ prominently an explicit discourse of rights to social protection. The cases of South Africa and Botswana show that an explicit discourse of social and economic rights is not necessary for the expansion of social protection to a large section of the population, at considerable expense, and with significant effects on poverty. The case of South Africa also suggests that the inclusion of social and economic rights in the constitution is probably not sufficient for further expansions of social protection. Since a right to income security and other social and economic rights were recognized, in the 1996 Constitution, the South African Government has not demonstrated much interest in expanding social protection. Kenya might be another example of a country where the constitutional recognition of social and economic rights does not readily generate actual policy reforms. Zambia and Senegal are cases of countries that are far more typical of Africa in that there is at most only a weak recognition of social and economic rights in the constitution, legislation or political discourse, and social protection has not expanded beyond two minority sections of the population: Workers in formal employment, many of whom (especially in Senegal) are covered by contributory programmes; and some‘vulnerable’ individuals especially in‘incapacitated’ h ouseholds, who receive very modest benefits through new social assistance schemes strongly supported by international organisations and donor agencies. Neither attempts to promote a discourse of rights not attempts to expand social protection to all poor people(rather than only a section deemed deserving) have succeeded. The facts that an explicit recognition of rights has historically been neither necessary nor sufficient for the expansion of social protection does not necessarily preclude the possibility that, in future, a rights-based approach would serve to mobilise support for and weaken opposition to social protection reforms, especially in countries(such as Zambia and Senegal) where the reach or coverage of social protection remains limited and little attention is paid to rights. Chapter 5 turns to the question of the likely politics of rights and reform in future.

124

ix Chapter 5: Explicit and implicit rights and the expansion of social protection There remain huge and glaring gaps in the reach or coverage of social protection across most of Africa. ILO data suggests that only a minority – perhaps one-third – of elderly people receive a pension, and the reach or coverage of cash transfer programmes for other poor people is even lower. Feeding schemes do reach a larger number of school children, but even they reach only a small minority of children across Africa as a whole. Coverage is uneven among workers in formal employment, growing but still limited among vulnerable groups and incapacitated households, and almost non-existent among workers in the growing informal economies and peasant farmers. Across most of Africa, people have neither formal rights to social protection nor any claim in practice on social protection programmes. Social protection has limited coverage or reach because the expansion of social protection has rarely attracted strong support among the political elite across much of Africa. In many countries, the political elite is either ambivalent or hostile to the idea that social protection might be expanded. The predominant strategy among advocates of social protection has been to point to its benefits in terms of immediate poverty-reduction and longer-term development. The World Bank is a powerful proponent of the developmental argument, in support of its proposals for targeted programmes, but many critics of the World Bank subscribe to the basic developmental argument. Despite these arguments, widespread opposition to the expansion of social protection persists across much of Africa. It is tempting, therefore, to consider a rights-based approach as an alternative strategy for promoting social protection. As we saw in Chapters 2 and 3, social and economic rights – and specifically rights to social protection – have a long his tory in‘global’ fora and have been incorporated(with modification) into the OAU and AU’s formal positions. Some national constitutions also recognize rights to social protection. Overall, however, most African governments have been reluctant to recognize rights to social protection. Social insurance programmes have generally been governed by legislation for decades, but most social assistance programmes introduced more recently have been introduced through executive fiat without any legislative basis. Bot swana’s extensive set of social assistance programmes, for example, has neither constitutional nor legislative bases. Perhaps a more energetic emphasis on rights to social protection would nudge or push governments into reform. In Chapters 2 and 3 we saw that explicit rights to social protection have most often entailed rights for workers(meaning workers in formal employment, in the‘modern’ economic sectors) and less often rights for citizens, especially the deserving poor categories of the elderly and disabled and(less often) children and their caregivers. The populations of the industrialised societies of north-west Europe comprised working people in formal employment, their dependents, and people who were unable to work because of age or infirmity. Social protection could be almost universal if it provided for workers

126 (and their dependents) and the deserving poor(who were unable to work and were not the dependents of any working people). This Eurocentric vision of social protection gave rise to the IL O’s proposals for a social protection‘floor’. The rights of workers underpin the expansion of the‘vertical’ dimension through consolidating contributory programmes among formally-employed workers. The rights of citizens underpinned the expansion of the‘horizontal’ dimension – i.e. the floor – through tax-financed social assistance programmes. An emphasis on rights in Africa might thus combine with the social protection floor agenda to encompass institutionalizing the rights of workers and the rights of citizens. In African contexts, these entail‘top down’ and‘bottom up’ routes to more universal social protection. Rights for workers would be linked to the consolidation of social insurance programmes for the formally employed and their expansion to at least some informal sector workers helping to prevent working people falling back into poverty. Rights for citizens would be linked to the expansion of social assistance programmes, especially for the very poor. An explicitly rights-based approach to workers ’ rights is likely to strengthen the former strategy. An explicitly rightsbased approach to citizens’ rights is less likely to strengthen the latter strategy, because of the strong antipathy among political and economic elites across much of Africa to the global discourse of social and economic rights. This does not mean, however, that there is no potential for a strategy framed in a normative discourse. The predominant understanding of welfare among political elites, at least in East and Southern Africa, has an alternative normative basis, emphasising responsibilities rather than rights, and relationships rather than individual entitlements. This explains why there is deep opposition to the provision of public assistance to working-age adults, except through workfare or in times of emergency, and why there is widespread anxiety over‘dependency’ on‘handouts’. Elite beliefs do, however, open up the possibility of expanding social protection for specific categories of people – especially the elderly and children – when kin have otherwise failed in their responsibility to provide support. Concepts of responsibility and opportunity point to the possibility of an effective, implicitly rights-based approach to expanding social protection. Even such an approach has its limits: Provision for‘workers’ and for the deserving poor leaves open the challenge of addressing poverty among the large proportion of the population who are, exclusively or primarily, peasant farmers. More attention needs to be paid to the rights of peasant farmers and to the responsibilities of society to them. Neither the concept of the social protection floor nor the dominant discourses of rights have much to offer people in Africa who are neither salaried or waged workers nor destitute through the lack of working kin to support them. 5.1. Elite conceptions of rights and responsibilities In interviews and discussions with politicians across a range of East and Southern African countries we find little interest in, or even acknowledgement of, the global discourse of

127 rights to social protection. Except when they are reading from speeches drafted by technocrats who have bought into the global discourse, ministers and members of parliament rarely(if ever) refer to rights, international covenants or social protection floors. Our research generally echoes the research conducted among elites in Malawi in 2008 by Kalebe-Nyamongo. Kalebe-Nyamongo interviewed about fifty people‘holding strategic positions in political, governmental, economic, communication, cultural and donor organisations, and social movements’ in order to analyse attitudes towards the poor and to cash transfers as a specific intervention by government. Most of the Malawian interviewees blamed poverty on the poor. The poor were said to be overly-fatalistic, which encouraged passivity. They used resources unwisely. And they had become dependent on handouts:‘They know that if they do not work, someone will come to help them’, said one interviewee. Despite evidence of the positive effects of cash transfers, ‘Malawi’s elites prefer policies that support those among the poor who can work to help themselves. Many respondents’ ideas about the poor – that they are passive, dependent, and fatalistic – seemed to contribute to a lack of support for social assistance schemes like cash transfers. These were seen as likely to encourage laziness or dependency.’(2014: 1) Malawian elites favoured workfare(i.e. public employment programmes) and supposedly ‘developmental’ programmes(such as m icrofinance) over cash transfers, and old-age pensions over other cash transfers. Our research suggests that political elites across most of East and Southern Africa draw on two major, and to some extent competing, sets of ideas or discourses. The one di scourse revolves around an aversion to‘ dependency ’ on‘handouts’, and a related anxiety about the effects of cash transfers on productivity and morality. This discourse has been both widespread and deep-rooted across East and Southern Africa. The competing discourse focuses on the issue of responsibility, i.e. the role of the state in the shared responsibility for the poor or disadvantaged. In this discourse, the state is often presented as having assumed some of the roles that were played formerly by kin or community. These two discourses are widely linked through a third concept, reciprocity: responsibilities are shared, and must be balanced appropriately, between individuals and the state, on the basis of a version of the reciprocity that historically has underpinned kinship and community-level social relationships. While the relative importance attached to the discourses of dependency and responsibility vary, and competing ideologies exist, there are sufficient similarities to identify a predominant, distinctive and generally conservative ideology of welfare across much of East and Southern Africa. The discourse of responsibility is central to the welfare doctrine developed in Botswana in the late 1960s and early 1970s, by the country’s first president, Seretse Khama, and his vice-president(and, in 1980, successor) Quett Masire, with the assistance of speechwriters. Botswana was born amidst terrible drought – the worst in living memory – that compelled the new Botswana Democratic Party(BDP) government to organise drought relief on a scale entirely unprecedented in the territory. The drought relief and later drought recovery programmes helped to define the new‘modern’ state that the BDP was building, and to provide the BDP with a firm electoral base. Khama and Masire sought to transfer a wide range of powers and responsibilities from traditional leaders to the new

128 modern state. These included responsibility for drought relief. The overall process of state-building was accompanied by an ideology of responsibility, as the new state assumed the responsibilities previously borne by chiefs. In this ideology, responsibility was shared between the individual and society. The state acted on behalf of society, as chiefs had previously. Society was viewed as a form of extended family, with an essential unity. Individual self-fulfilment as well as identity depended on social and communal relationships. Tradition was an important source of legitimation. Reciprocity was integral to this ideology. The state/society/family assisted an individual in need because that individual would in turn play his or her part in the collective assistance of other needy individuals. Each individual had a moral obligation to work, and to be as self-reliant as possible. People should nei ther give nor get‘something for nothing’. This was not an egalitarian ideology. On the contrary, inequality and patriarchy were generally viewed as natural. The ideology was communitarian, opposed to individualism. If any individual fails to reciprocate, then the moral imperative of collective responsibility falls away. Individuals can make claims only insofar as they are part of a network of claims and obligations. In Botswana, BDP leaders debated repeatedly denying assistance to people who failed to fulfil their social obligations. Should food and other assistance be provided to farmers who did not go out to their fields, plough the land and plant crops? Should families that did not participate in‘voluntary’ school -building programmes be allowed to send their children to the schools that were built? More broadly, should drought relief be conditional on work(i.e. workfare)? Did outsiders have any legitimate claim? In South Africa, the extensive Child Support Grant is controversial, it seems because cash is given to young(and often unmarried) mothers. Research suggests that the grant has significant benefits for children, as well as dignity to otherwise in impoverished women. But critics of the grant point to the payment of cash benefits to young unmarried women who supposedly do not fulfil their reciprocal social obligations to kin. This form of responsibility(and reciprocity) is a weak or thin version of solidarity. It is also offset by anxiety over‘dependency’ – and this anxiety explains some of the resistance to stronger versions of solidarity. Across East and Southern Africa, political and other elites share a w idespread anxiety over‘dependency’ on government programmes(‘handouts’) which are seen to sap the recipients’ will to stand on their own f eet through hard work. If an understanding of social responsibility has pushed elites tentatively towards pro-poor reforms in a range of countries, so anxieties about‘dependency’ have reeled in any nascent enthusiasm. This was not a feature of the welfare doctrine articulated in Botswana in the 1960s and 1970s. Khama and the BDP emphasized‘self reliance’ – as did, famously, Nyerere in Tanzania – meaning that they aspired both to make their countries independent of the colonial power(the UK) or other aid donors, and to ensure that their citizens became successful peasant farmers. Later, beginning around 1980, BDP leaders shifted from their praise of self-reliance to the denunciation of dependency. Dependency also meant both external dependency(for example, on the World Food Progremme or aid donors) and the

129 dependency of individual citizens(on‘handouts’ or other government programmes). The second of these was the most important in terms of the welfare doctrines that were being articulated by elites. The emergence of anxiety over dependency on handouts required two major changes in the context. First, it required the existence of public programmes that‘handed out’ benefits to citizens. At the time, anxiety about dependency entailed anxiety about existing programmes more than prospective ones. In Botswana and many other countries, anxiety focused on the possibility that peasant farmers would become dependent on drought relief. Secondly, anxiety over dependency required the privileging of economic ‘development’ in national policy -making. Dependency and development were intimately connected to each other. Whilst economic development meant rising productivity and output, dependency meant the failure to develop.‘Successful’ programmes – in developmental terms – helped the poor to raise themselves out of poverty, ideally through the acquisition of human capital or other assets, but also through ensuring survival during times of drought.‘Development’ required that beneficiaries‘graduated’ off programmes, as they prospered. Dependency marked the failure to graduate off programmes, the failure to develop, and thus the failure of development(Seekings, 2017b). The discourse of dependency often went along with a shift from a social to an economic analysis of poverty. Poverty increasingly came to be viewed not in terms of social and economic relationships but rather as a matter of individual incentives. Government programmes that provided‘handouts’ reduced the incentive to wor k, to become selfreliant, and to prosper. At the same time, however, dependency retained a fundamentally social character, because it marked the breakdown of reciprocity. Dependency entailed the failure of poor people to fulfil their moral responsibilities to the broader community, to strive to prosper and become productive members of the community, willing and able to help others in turn. In the case of Botswana, anxiety about dependency ran up against an acknowledgement of responsibility. The dilemma was set out explicitly in the 7 th National Development Plan (NDP), for 1991-1997. The Plan recorded a strong commitment to social justice and concern over the decline of extended family support, envisaging a‘social safety net’ and making specific budgetary provision for childcare(Botswana, 1991: 386). Here the government faced the dilemma:‘The major challenges Government faces is to ensure that the temporary reliance of vulnerable groups on drought relief does not become a permanent dependency’( ibid: 389). Feeding programmes for children represented an developmental investment, but support for other‘vulnerable’ groups risked encouraging dependency. The BDP Government’s solution was to keep most of its programmes tightly targeted on the very poor. The discourse of dependency in Africa has similarities to but is not identical to the discourse of dependency in the global North. In preindustrial England, dependency meant little more than subordination, without negative moral connotations(Fraser and Gordon, 1994: 313) ;‘independence’(and citizenship) were based on property -ownership. With industrialization, however, its meaning shifted to refer to people who remained outside of

130 the system of wage labour. Dependency became shameful, deviant, and stigmatised; the ‘pauper’ ceased to mean someone who was poor, and instead came to mean someone who was dependent on poor relief rather than wages; paupers were‘degraded’, their characters corrupted, perhaps inherited from their parents, perhaps due to chronic reliance on charity.‘To be a pauper was not to be subordinate within a system of productive labour; it was to be outside such a system altogether’( ibid: 317). In the USA, dependency became increasingly racialized, gendered and stigmatized. As Daniel Moynihan dec lared in 1973,‘the issue of welfare is the issue of dependency’:‘To be poor is an objective condition; to be dependent, a subjective one as well … Being poor is often associated with considerable personal qualities; being dependent rarely so’ (quoted in ibid: 309). In Africa, the concept of dependency has a different genealogy, and a slighty different meaning. Dependency had an economic element, in that dependent people were not working sufficiently to raise themselves out of poverty and economic dependency. At the same time, however, it was a fundamentally social character. Dependency marked the breakdown of reciprocity and the triumph of the individual over the social. Ferguson rightly notes that dependency entails relationships: ‘For while“dependency” is obsessively decried as a problem or a trap in social policy discourse, an ethnographic view suggests that it is really only via relations of“dependence” that most of the population survives at all. Dependence is, in this respect, not the name of the problem; it is the name of the solution.’(2015: 97) Ferguson underestimates, however, the way in which dependency entails the denial of the kinds of relationships deemed(by elites at least) to be just. The discourse of dependency in Africa had its roots in the agrarian past, and especially in the historic availability of land which meant that production was limited most immediately by labour. Dependency was rooted in the failure to work, but not in the failure to engage in wage labour as much as the failure to engage in peasant production. In Britain,‘Victorian’ morality represented dependency as a failure to conform with the moral obligation to perform wage labour, i.e. to fit into one’s appropriate position in capitalist society. In Africa, dependency meant a failure to conform with the moral obligation to be a productive peasant, i.e. to fit into one’s appropriate position in an agrarian society. In practice, a concern over dependency inclined policy-makers towards workfare and the use of farm input subsidies(to reward hard-working farmers) rather than cash transfers. Studies of public programmes across most of East and Southern Africa report denunciations of dependency. In Malawi, KalebeNyamongo’s interviewees emphasised it(Kalebe-Nyamongo and Marquette, 2015), and rival presidential candidates in the 2014 election used it to bash the briefly incumbent president, Joyce Banda(Hamer and Seekings, 2017). In Ethiopia and Rwanda, the authoritarian regimes led by Meles and Kagame denounced dependency and insisted on the imperative of work. The Kagame regime in Rwanda has held an even more dismissive view of the peasantry than the pregenocide Habyarimana regime. For Kagame, the peasantry was undynamic and needed to be modernised. Poverty was seen as the consequence of a lack of determination.‘If

131 we can utilize the resources that God has given to us to good effect, we can eradicate poverty’, Kagame told the country in 2000.‘We would like to urgently appeal to the Rwandese people to work. As the Bible say s,“he who does not work should not eat”’ (cited in Ansoms, 2009: 297; see also Lavers, 2016a, also on Rwanda, and 2016b on Ethiopia). South Africa’s Deputy -Minister for Social Development briefed an AU meeting on her country’s social protection system. She concluded with a warning that‘the main challenge … was not to create a dependency syndrome among the population to the detriment of productivity.’(AU, 2015a). International agencies recognize the depth of this anxiety about dependency. The FAO, in endorsing social protection in 2017, referred to research that it had conducted together with UNICEF that‘ contributed to strengthening the case for social protection as an investment, not just a cost, while addressing public misperceptions around dependency and labour disincentives ’(2017: 11) . In Botswana, anxiety about dependency became almost ubiquitous by the 2000s. The UNDP’s 2005 Botswana Human Development Report reported i n a section titled‘The Trap of State Provision’ that an earlier“2002 review of antipoverty initiatives in Botswana” had found that“stakeholders… cited‘welfare policies that promote a culture of dependency’ among the causes of poverty in Botswana”. Referring to drought relief and the Accelerated Rural Development Programme(ARDP), the 2002 report apparently found that“poor people were unanimous in their view of drought relief as a useful programme” but“many of the relatively well -off respondents, including civil servants, expressed concern” over the possibility that policies might exacerbate or cause poverty (UNDP, 2005: 18). Ideologies that focus on the risk of dependency inhibit the expansion of social protection to working-age, able-bodied adults and the households in which they live. The fear is that social protection will in fact disadvantage the supposed beneficiaries, as well as compromising the norms of reciprocity and the social relationships that underpin social cohesion and order. Anxiety over dependency does not, however, preclude the institutionalization of collective responsibility for the deserving poor, who include both individuals who are unable to support themselves because of age or infirmity, and people whose livelihoods have been disrupted temporarily by natural or other disasters. 5.2. Workers’ rights and the expansion of contributory programmes Norms of social responsibility have little bearing on workers’ rights. Workers are generally viewed as having rights in relation to their employers. It is the employers who, at least in the first instance, have obligations to their employees. The focus of social insurance schemes on formally-employed workers is not simply a matter of Eurocentric influence or operational convenience. It is also an ideological or discursive issue. If a worker does not clearly have an employer who is exploiting him or her, then it is unclear who has an obligation to recognise the worker’s rights. In Africa, a large proportion of informal sector workers are self-employed, whether as marginal hawkers or relatively well-off artisans. Even with manufacturing it seems that there are growing numbers of people who are formally self-employed, doing piecework, often at home. Whether any of these self-

132 employed workers have the same rights as salaried and waged workers is a difficult question. Lund(2009) began to tackle this difficult question although she provided only the preliminary outline of an answer. Her analysis is rooted in the South African case. In South Africa, as we saw in Chapter 4 most formally-employed workers enjoy a range of benefits through contributory programmes whilst the elderly, disabled and children receive social pensions or grants. Lund notes that the expansion of social assistance entailed the expansion of rights for citizens. Many workers – or at least many retired or disabled workers and many of the children of workers – are not covered as workers, but as citizens. This expansion of social citizenship has meant that responsibility has shifted from employers(and the owners of capital) to the state. Lund suggests that the‘owners of capital’ – who themselves may not be direct employers – should bear the responsibility for funding social protection for informal workers. By‘informal workers’, Lund means primarily informal wage workers not the self-employed, although she considers that some piecework‘outworkers’ are disguised wage workers. Moreover, in practice Lund seems to envisage ensuring that the costs of social protection are somehow passed onto consumers, especially when value chains are global, linking rich consumers in the global North and poor workers in the global South. This vision could encompass self-employed farmers or artisans as long as they were producing for sale rather than subsistence. Lund does not proceed to consider how this vision might be effected. It is hard to see how Lund’s vague vision of a reinvigorated workerist model of social and economic rights could be effected. A series of initiatives have put forward bold visions of the expansion of social protection to the informal sector. The AU’s 2 011 SPIREWORK programme envisaged the expansion of social protection to the 7080 percent of Africa’s working people who were in the informal economy, i.e. including peasant farmers. How this would be effected was not clear. The AU pointed hopefully at community-based health insurance schemes in a number of countries as well as National Health Insurance schemes in Ghana and Namibia. In practice, community-based schemes have generally had limited reach, and the‘national’ health insurance scheme in Ghana pro vides much less than universal coverage within Ghana(see Alfers, 2013). South Africa’s experience with contributory programmes is instructive here. South Africa’s post-apartheid governments have sought to encourage more working people to save for their retirement and to insure themselves against illness. The rich, for the most part, already self-insure, mostly through privately-run pension funds and medical aid schemes. The non-rich have been very reluctant to pay the premia required for membership of pension funds and medical aid schemes. Trade unions have resisted strongly the government’s proposals that require workers to contribute more. Government concluded that the expansion of pension funds could only be achieved through significant subsidies out of general taxation. Even the ILO recognizes that this is generally true:‘ Schemes for self-employed workers are more successful when the government is willing to subsidize part of the premium’

133 (ILO, 2009: 42). If schemes for workers were to be financed out of special taxes, the obvious question is why should scarce tax revenues be allocated to them rather than the very poor? In practice, attempts to expand social protection to informal workers have made little progress in Africa hitherto, and there is little prospect of this changing in the immediate future. The experience of Zambia suggests that there is some potential for expanding social insurance to some groups of informal wage workers, but the scale is small in comparison with the numbers of people living in poverty, especially in rural areas. It is also unclear how many countries have the administrative capacity to organize social insurance schemes that can cope with members who shift employment frequently. Consolidating social insurance within the formally-employed working population and expanding it within the informally-employed working population are undoubtedly important tasks, but these programmes will not reach the majority of the population and certainly not the poorest section of the population unless their financial basis is shifted from contributions to taxes. The workerist project seems to have very little to say about the large numbers of self-employed working people who constitute a majority of the working population across much of Africa – far more so than in most other parts of the global South(or North). Unsurprisingly, more attention has been paid to the question of rights to social protection for citizens as citizens. 5.3. Rights, responsibilities, and the expansion of non-contributory programmes The limits to an approach focused on workers’ rights to social protection funded through earnings-related contributions lie in the financial and practical challenges of organizing contributory programmes. Given these, even the ILO recognizes the need for a taxfinanced social protection floor. As the ILO, World Bank and others have argued, a basic floor – especially one that is targeted on the poor – is affordable. The constraint on the expansion of social assistance is largely political, and in significant part ideological. Hanlon, Barrientos and Hulme have argued boldly that‘just giving money to the poor’ is the right thing to do as well as being politically viable, fiscally affordable and economically developmental. Our and others’ res earch suggests that political elites across Africa do not share Hanlon et al .’s view that giving money to‘the poor’ is the right thing to do. Political elites have shown little more than perfunctory and rhetorical support for the social and economic rights advocated by international organisations, aid donors and NGOs (both foreign and local). In public, they might read speeches written by subordinates, but in private they generally express deep ambivalence if not hostility. An explicitly rightsbased approach to social assistance does not get much traction among most political elites in Africa. The absence of enthusiasm for the‘global’ discourse of rights does not mean, however, that politicians do not have their own norms and values that entail, implicitly if not explicitly, an understanding of rights. Our research suggests that most members of the political elite in East and Southern Africa tend to a conservative view of social and

134 economic rights that is at odds with the more‘progressive’ view that cha racterizes many (but not all) international agencies and NGOs. This conservative approach emphasizes responsibility more than entitlements, relationships rather than individual autonomy. Conservatives worry about the risk of dependency if the poor themselves do not fulfil their reciprocal obligations, to work in return for support, to support others in return for being supported themselves. In Africa, as elsewhere, policy-making elites view their policy options through the prism of their norms and beliefs. Advocates of specific policy models arrive in national capitals armed with statements about‘rights’ and evidence of the developmental benefits of welfare, often to be brushed off by local elites who view the basic approach with considerable unease. Their unease might reflect also their perception that they have little to gain politically from introducing reforms, but it goes beyond any such political selfinterest. If the welfare state in Europe became the vehicle for primarily a Polanyian process of decommodification, in Africa it appears as the vehicle threatening a process of ‘defamilialisation’ which worries many policy -makers. The welfare state in Africa is not viewed primarily as something to subvert the market, but as something that might substitute for the family. The welfare state is the vehicle for defamilialisation rather than decommodification, and – for conservatives – this is often not a good thing. At times it is appropriate: The state has a responsibility to substitute for the family(or community) when the family(or community) fails. But, conservatives aver, the state should beware undermining the family, and should ensure that the social relations of reciprocity that underpin the ideal family and community are replicated in the relations between people and the state. People should not be encouraged to become autonomous individuals, detached from their wider web of obligations and responsibilities. Dependency arises when they do become so detached. The themes of responsibility and reciprocity draw on the past, but they are shaped also by the power of development. Dependency entails both the breakdown of reciprocity and the failure of development. The predominant welfare ideologies in Africa are conservative, with shared concerns over the balance between responsibility and dependency. Most of the variation between these conservative welfare ideologies is based on differences in the preferred balance between responsibility and dependency. Sometimes – as in Botswana in the 1960s and 1970s, under Seretse Khama, or in South Africa in the 2000s, when Zola Skweyiya was Minister of Social Development – responsibility trumped the spectre of dependency. For many other policy-makers, however, dependency has generally trumped responsibility. In general, the two are rarely far apart. Conservatives often belief that rights and responsibilities are differentiated by gender, class or status, familial status, and often also nationality. They are instinctively hostile to universalist claims. In the predominant conservative view, men have both different and more rights and responsibilities than women, and children’s rights are far more limited than suggested in even the African Children’s Charter.

135 Many advocates of the global rights-based approach recognized these realities when they engage with government ministers and top bureaucrats. In their briefings, they downplay talk of rights, and refer more to the developmental benefits of paying grants to poor families with children or even pensions to the elderly. Some political leaders and parties in Southern Africa associate themselves with the idea of‘social democracy’, but it is not entirely clear how deep is this affinity, and cynics wonder whether it is motivated in part by a desire to access funds from social democratic parties and foundations in Europe. For the most part,‘social democracy’ tends to mean a willingness to expand the role of the state in the economy, not a statement of commitment to either solidarity or equality. Viewed comparatively, the dominant norms and values of political elites in Southern Africa resemble more the views of Christian Democrats and their equivalent than they do Social Democrats. This does not mean, however, that there are no opportunities for promoting the kinds of policy reforms associated with a rights-based approach. In Europe, as Wilensky showed long ago, there was a close correlation between the strength of Christian Democratic parties and government spending on the welfare state. EspingAndersen(1990) later suggested that this was because Christian Democrats(and similar conservatives in Britain, in the One Nation tradition of the Conservative Party) promoted a particular kind of welfare regime, one that was more stratified and more familial than its social democratic rival(see also van Kersbergen, 1995). Across much of Africa, elites’ endorsement of norms of responsibility open up spaces for reforms that do not target the poor as such, but rather focus on people who have been failed by their communities or families, including especially the destitute elderly. 5.4. Beyond Social Assistance for the Deserving Poor A strategy emphasizing explicitly the rights of workers might help to consolidate and expand slowly contributory social insurance and a strategy emphasizing responsibilities to the(deserving) poor might help to mobilise support and(crucially) defuse opposition from conservative elites. Even if these strategies were to be successful and social protection was to be expanded, it is likely that the reach or coverage of social protection in most African countries would expand from the current small minority of the population to a large minority of the population, and the benefits for many would remain very modest. South Africa’s social protection system is widel y lauded as exceptional within Africa, but even it fails to cover or reach many South Africans, including even many poor and some extremely poor people. In the South African case, a high proportion of the uncovered poor are unemployed. Elsewhere in Africa, the challenge is to cover the working poor, especially self-employed peasant farmers and others. What approaches – and what conceptions of rights – might underpin an expansion of social assistance that covers all poor people and is politically viable given the conservatism of political elites across much of Africa? It is likely to be very difficult to persuade political elites to consider unconditional cash transfers to working-age adults or to subsidise contributory schemes on a sufficiently large scale.

136 One possible approach is to consider more seriously the conservative conception of rights to equal or at least fair opportunities. In the contemporary African context, this might entail more sustained thinking about(a) the rights of small farmers in terms of opportunities to produce and perhaps sell their produce, which is bound up with the right to food or to ‘food sovereignty’, and(b) the rights of them and others to employment. The first of these requires engagement with a different set of organisations, including the FAO. The AU, ILO and FAO may have begun to consider how to integrate social protection with agricultural policies, but they do not seem to have made much progress. Some attention has been given to the second of these approaches. Conservative elites have accepted ‘workfare’ because it imposes the condition of work, which is in part a condition of reciprocity, on recipients. Workfare has been rolled out on a large scale on occasion in Africa, primarily as a mechanism for drought relief. In dia’s Mahatma Gandhi National Rural Employment Guarantee Scheme offers an important model here. Much more attention might be paid to the possibility of Employment Guarantee Schemes in Africa. The application of rights-based approaches to social protection in Africa thus entails both a challenge and an opportunity. The challenge is to vernacularise effectively the global rights discourse, taking into account the conservatism of African elites and the fact that poverty is most pronounced among peasant farmers. The opportunity is to build on the implicit recognition of rights entailed in the ideology of family, responsibility and opportunity.

137 References Acharya, Amitav(2004),‘How ideas spread: Whose norms matter? Norm localization and instituti onal change in Asian regionalism’, International Organization 58,2: 239275. Albertyn, Catherine, and Dennis Davis(2010),‘Legal realism, transformation and the legacy of Dugard’, South African Journal on Human Rights 26,2: 188-216. An-Naim, Abdullahi Ahmed(ed, 2002), Cultural Transformation and Human Rights (London: Zed). ----- and Francis Deng(eds, 1990), Human Rights in Africa(Washington: Brookings Institution). Ansoms, An(2009),‘Re -engineering Rural Society: The Visions and Ambitions of the Rwandan Elite’, African Affairs 108,431: 289-309. APSP(2012), Review Report 2010-2011(Nairobi: African Platform for Social Protection) -----(nd), Review Report 2013-2014(Nairobi: African Platform for Social Protection) AU(2003), ‘Protocol to the African Charter on the Rights of Women in Africa’(Addis Ababa: African Union). ----(2004),‘Ouagadougou Declaration on Employment and Poverty Alleviation in Africa and accompanying Plan of Action’(Addis Ababa: African Union). -----(20 05),‘Draft Social Policy Framework for Africa’, document LSC/EXP/5(III) for the 3 rd Ordinary Session of the Labour and Social Affairs Commission of the African Union, Johannesburg(April). ----(2006a),‘Livingstone Call for Action’ (Addis Ababa: African Union). ----(2006b),‘Yaounde Call for Action’(Addis Ababa: African Union). ----(2008),‘Social Policy Framework for Africa’, presented to the First Session of the AU Conference of Ministers in charge of Social Development(CAMSD), Windhoek, 27-31 October. ----(2011),‘Promoting Employment for Social Cohesion and Inclusive Growth’, 8 th Ordinary Session of the Labour and Social Affairs Commission of the AU, 11-15 April, Yaoundé. -----(2013), Concept document for Workshop for Mainstreaming Social Protection Plan for the Informal Economy and Rural Workers in the AU Comprehensive Africa Agriculture Development Programs(CAADP), Nairobi, 21-22 October. ----(2014a),‘Recommendations of the African Union Expert Consultation Meeting on Children and Social Protection Systems in Africa, Cape Town, 2830 April’. ----(2014b),‘Malabo Declaration on Accelerated Agricultural Growth and Transformation for Shared Prosperity and Improved Livelihoods’(Addis Ababa: African Union). ----(2015a),‘Social Protection for Inclusive Development: Draft Report of the Ministers’ Meeting’, First meeting of the Specialised Technical Committee on Social Development, Labour and Employment(STC-SDLE-1), Addis Ababa, 20-24 April. ----(2015b),‘Draft Document on Mainstreaming Social Protection Plan for the Informal Economy and Rural Workers in the AU Comprehensive Africa Agriculture Development Programs(CAADP)’, first session of the Specialised Technical Committee on Social Development, Labour and Employment(STC-SDLE-1), Addis Ababa, 20-24 April.

138 ----(2016),‘Country CAADP Implementation Guidelines under the Malabo Declaration’ (Addis Ababa: African Union). Baricako, Germain(2008),‘Introductory Preface: The African Charter and African Commission on Human and Peoples’ Rights’, in Malcolm Evans and Rachel Murray(eds), The African Charter on Human and Peoples’ Rights , 2 nd edition, The System in Practice, 1986-2006(Cambridge: Cambridge University Press): 1-19. BarOn, A.(2001),‘Providing So Little for So Few: Botswana’s So cial Assistance Scheme’, in Francis Wilson, Kanji, N.& E. Braathen(eds.), Poverty Reduction: What Role for the State in Today’s Globalized Economy? (London: Zed, with CROP): 246-268. Barrientos, Armando(2013), Social Assistance in Developing Countries(Cambridge: Cambridge University Press). Beegle, Kathleen, Luc Christiaensen, Andrew Dabalen and Isis Gaddis(2016), Poverty in a Rising Africa: Overview(Washington DC: World Bank). BFTU(2007), Policy position paper on social security and social protection in Botswana – 2007(Gaborone: Botswana Federation of Trade Unions). BIDPA& World Bank(2013), Botswana Social Protection Assessment(Gaborone: Botswana Institute for Development Policy Analysis and Washington DC: World Bank). Bilchitz, David(2007), Poverty and Fundamental Rights: The Justification and Enforcement of Socio-Economic Rights(Oxford: Oxford University Press). ----(2010),‘Is the constitutional court wasting away the rights of the poor? Nokotyana v Ekurhuleni Met ropolitan Municipality’, South African Law Journal 127,4: 591-605. Botswana(1991), 7 th National Development Plan(Gaborone: Ministry of Finance and Development Planning). -----(1997), Vision 2016(Gaborone). -----(2010), Botswana: Millenium Development Goals Status Report 2010(Gaborone: Ministry of Finance and Development Planning). -----(2015), Botswana: Millenium Development Goals Status Report 2015(Gaborone: Ministry of Finance and Development Planning). -----(2016), Draft National Development Plan 11(Gaborone: Ministry of Finance and Development Planning). Brown, Rebecca, and Judy Oder(2016),‘The Protection of Women’s Economic, Social and Cultural Rights in Africa’, in Danwood Chirwa and Lilian Chenwi(eds), The protection of Economic, Social and Cultural Rights in Africa(Cambridge: Cambridge University Press): 121-54. Cameron, Edwin(2014), Justice: A Personal Account(Cape Town: Tafelberg). Cichon, Michael(2013),‘ The Social Protection Floors Recommendation, 2012(No. 202): Can a sixpage document change the course of social history?’, International Social Security Review 66,3-4: 21-43. ----and Hagemejer, K.(2007),‘Changing the Development Policy Paradigm: Inves ting in a Social Security Floor for All’, International Social Security Review 60,2-3: 169 – 96. Chirwa, Danwood(2005a),‘A full loaf is better than half: The constitutional protection of economic, social and cultural rights in Malawi’, Journal of African Law 49,2: 207241.

139 ----(2005b),‘The African Human Rights Framework’, in Malcolm Langford(ed.), Social Rights Jurisprudence: Emerging Trends in International and Comparative Law (Cambridge: Cambridge University Press): 323-38. ----- and Lilian Chenwi(eds, 2016), The Protection of Economic, Social and Cultural Rights in Africa(Cambridge: Cambridge University Press). Claeys, Priscilla(2012),‘The creation of new rights by the food sovereignty movement: The challenge of institutionalizing subversion’, Sociology 46,5: 844-860. -----(2015a)‘Food sovereignty and the recognition of new rights for peasants at the UN: A critical overview of La Vía Campesina’s rights claims over the last 20 years’, Globalizations 12,4: 452-465. -----(2015b), Human Rights and the Food Sovereignty Movement: Reclaiming Control (Routledge/Earthscan). Cooper, Frederick(1996), Decolonisation and African Society(Cambridge: Cambridge University Press). Cornia, Giovanni, Richard Jolly, and Frances Stewart(1987), Adjustment with a Human Face: Protecting the Vulnerable and Promoting Growth(Oxford: Oxford University Press). Daffé, Gaye, and Mbaye Diène(2017),‘Vulnerability and Social Protection: Evidence from Senegal’, NoPoor Working Paper 3(Consortium pour la Recherche Economique et Sociale, CRES). Dauber, M. L.(2013), The Sympathetic State: Disaster Relief and the Origins of the American Welfare State(Chicago: University of Chicago Press). Davis, Benjamin, Sudhanshu Handa, Nicola Hypher, Natalia Winder Rossi, Paul Winters and Jennifer Yablonski(eds, 2016), From Evidence to Action: The Story of Cash Transfers and Impact Evaluation in Sub-Saharan Africa(Oxford: Oxford University Press, FAO and Unicef). Deacon, Bob(2013), Global Social Policy in the Making: The foundations of the social protection floor(Bristol: Policy Press). de Haan, Arjan(2014),‘The Rise of Social Protection in Development: Progress, Pitfalls and Politics’, European Journal of Development Research 26: 311-21. Desmarais, Annette Aurélie(2007), La Vía Campesina: Globalization and the Power of Peasants(Halifax: Fernwood). Devereux, S.(2007),‘Social pensions in Southern Africa in the twentieth century’, Journal of Southern African Studies 33,3: 539-560. DfID( 2000),‘Realising Human Rights for Poor People’, Target Strategy Paper(London: Department for International Development). Diallo, Fatima(2016),‘Direct constitutional protection of economic, social and cultural rights in Francophone legal systems: Senega l’, in Danwood Chirwa and Lilian Chenwi(eds), The protection of Economic, Social and Cultural Rights in Africa (Cambridge: Cambridge University Press): 396-422. Diop, Ahmadou Yéri(2003),‘Governance of Social Security Regimes: Trends in Senegal’, International Social Security Review 56,3-4: 17-23. Drake, Lesley, Alice Woolnough, Cramen Burbano and Donald Bundy(2016), Global School Feeding Sourcebook: Lessons from 14 Countries(London: Imperial College Press).

140 Drèze, Jean(1990),‘Famine prevention in India’, in Jean Drèze and Amartya Sen(eds), The Political Economy of Hunger, Volume 2: Famine Prevention(Oxford: Oxford University Press): 13-122. Dugard, Jackie(2008),‘Courts and the poor in South Africa: A c ritique of systemic judicial failures to advance transformative justice’, South African Journal on Human Rights 24,2: 214-38. ----- and Narnia BohlerMuller(2014),‘Poverty and substantive equality in the courts’, in Thenjiwe Meyiwa, Muxe Nkondo, Margaret Chitiga-Mabugu, Moses Sithole and Francis Nyamnjoh(eds), State of the Nation: South Africa, 1994-2014(Pretoria: HSRC Press): 237-52. Durnford, Robin(2016), The Politics of Transnational Peasant Struggle: Resistance, Rights and Democracy(London and New York: Rowman and Littlefield). Egorov, Alexandre(2016),‘Foreword: Making social security a constitutional right’, in ILO, The Right to Social Security in the Constitutions of the World: Broadening the moral and legal space for social justice, Volume 1: Europe(Geneva: ILO): xv-xvii. Englund, Harri(2006), Prisoners of Freedom: Human Rights and the African Poor (Berkeley: University of California Press). ----(2012),‘Poverty’, in Didier Fassin(ed.), A Companion to Moral Anthropology(Oxford: Blackwell). ----(2014),‘Cutting human rights down to size’, in Mark Goodale(ed.), Human Rights at the Crossroads(Oxford: Oxford University Press): 198-209. Esping-Andersen, Gøsta(1985), Politics Against Markets: The Social Democratic Road to Power(Princeton: Princeton University Press). -----(1990), The Three Worlds of Welfare Capitalism(Princeton: Princeton University Press). FAO(2016), Strengthening Coherence between Agriculture and Social Protection to Combat Poverty and Hunger in Africa: Framework for Analysis and Action(Rome: Food and Agriculture Organisation). -----(2017), FAO Social Protection Framework: Promoting Rural Development for All (Rome: Food and Agriculture Organisation). Ferguson, James(2015), Give a Man a Fish: Reflections on the New Politics of Distribution(Raleigh, NC: Duke University Press). Forteza, Alvaro, Leonardo Lucchetti and Mntserrat PallaresMiralles(2009),‘Measuring the Coverage Gap’, in Robert Holzmann, David Robalino and Noriyuki Takayama (eds), Closing the Coverage Gap: Role of Social Pensions and Other Retirement Income Transfers(Washington DC: World Bank). Fraser, Nancy, and Linda Gordon(1994),‘A Genealogy of Dependency: Tracing a keyword of the U.S. welfare state’, Signs 19,2 309-36. Garcia, Marito, and Charity M.T. Moore(2012), The Cash Dividend: The Rise of Cash Transfer Programs in Sub-Saharan Africa(Washington D.C.: World Bank). Gulbrandsen, Ørnulf(2012), The State and the Social: State Formation in Botswana and its Precolonial and Colonial Genealogies(New York: Berghahn). Gray, Mel(2006),‘The progress of social development in South Africa’, International Journal of Social Welfare 15.s1.

141 Greven, Thomas(2017),‘Organising the Informal Economy in Senegal’, in Virginius Xaxa, Debdulal Saha and Rajdeep Singha(eds), Work, Institutions and Sustainable Livelihoods: Issues and challenges of transformation(London: Palgrave): 303-22. Hamer, Sam(2016),‘“Our Father’s Programmes”: political branding around social protection in Botswana, 20082014’, CSSR Working Paper 320(Cape Town: Centre for Social Science Research, University of Cape Town). ----and Jeremy Seekings(2017),‘ Social protection, electoral competition, and political branding in Malawi ’, UNU-WIDER Working Paper 2017/99(Helsinki: UNUWIDER). Handley, G.(2009). Fiscal Space for Strengthened Social Protection: West and Central Africa(London: Overseas Development Institute and UNICEF). Hanlon, Joseph, Armando Barrientos and David Hulme(2010), Just Give Money to the Poor: The Development Revolution from the Global South(Sterling, VA: Kumarian). HelpAge International(2006),‘ A transformative agenda for the 21st century: Examining the case for basic social protection in Africa ’, background briefing for the Intergovernmental regional workshop, Livingstone, Zambia, 20-23 March(London: HelpAge International). -----(2009), Social Protection Policy: Responses to Older People’s Needs in Zanzibar (London: HelpAge International). -----(2010), Achieving Income Security in Old Age for all Tanzanians: A Study into the Feasibility of a Universal Social Pension(Dar es Salaam: Ministry of Labour, Employment and Youth Development with HelpAge International). ----(2014),‘Towards Universal Pension in Tanzania, paper presented at conference on ‘Social Protection: Building Effective and Sustainable Systems for Equitable Growth’, Arusha, December. -----(2016), Realising Income Security in Old Age: A study into the feasibility of a universal old age pension in Malawi(London: HelpAge International, and Lilongwe: Ministry of Gender, Children, Disability and Social Welfare). Hickey, Sam, Tom Lavers, Miguel Niño ‐ Zarazúa, and Jeremy Seekings(2018),‘Th e negotiated politics of social protection in subSaharan Africa’, UNU-WIDER Working Paper 2018/34(Helsinki: UNU-WIDER). ----and Jeremy Seekings(2017),‘The global politics of social protection’, UNU-WIDER Working Paper 2017/115(Helsinki: UNU-WIDER). Hodges, Anthony(2008),‘ Perspectives for social protection in West and Central Africa ’, unpublished paper. Howard, R.(1983),‘The full -belly thesis: Should socio-economic rights take priority over civil and political rights? Evidence from SubSaharan Africa’, Human Rights Quarterly 5: … ----(1990),‘Group versus individual identity in the African debate on human rights’, in A.A.AnNa’im and F.M. Deng(eds), Human Rights in Africa: Cross-cultural perspectives(Washington DC: Brookings Institution): 159-83. Iliffe, John(1990), Famine in Zimbabwe, 1890-1960(Gweru: Mambo). ILO(1941), Approaches to Social Security: An International Survey, Studies and Reports series M(Social Insurance) no.18(Montreal: International Labour Office).

142 -----(1944a). Minimum Standards of Social Policy in Dependent Territories. Report distributed prior to 26 th Session of the International Labour Conference(Montreal: International Labour Office). -----(1944b). Social Policy in Dependent Territories(Montreal: International Labour Office). -----(1982), Basic Needs in Danger: A Basic Needs Oriented Development Strategy for Tanzania(Addis Ababa: International Labour Office). -----(2000), Income Security and Social Protection in a Changing World(Geneva: International Labour Office). ----(2008),‘Can Low income Countries Afford Basic Social Security?’. Socia l Security Policy Briefings 3(Geneva: International Labour Organization). ----(2009),‘The Informal Economy in Africa: Promoting transition to formality: Challenges and strategies’(Geneva: ILO, 2009). -----(2010a), Extending Social Security to All: A guide through challenges and options (Geneva: International Labour Office). -----(2010b), World Social Security Report(Geneva: International Labour Office). ----(2011),‘Yaoundé Tripartite Declaration on the Implementation of the Social Protection Fl oor: The Action Plan’, draft for discussion at ILO’s 12 th African Regional Meeting, Johannesburg, October. -----(2014), World Social Protection Report 2014/15: Building Economic Recovery, Inclusive Development and Social Justice(Geneva: International Labour Office). ----(2015),‘Recommendation Concerning the Transition from the Informal to the Formal Economy’, adopted at the 104 th session of the International Labour Conference, Geneva. -----(2016a), Social Security Inquiry Manual(Geneva: International Labour Office). -----(2016b), The Right to Social Security in the Constitutions of the World: Broadening the moral and legal space for social justice, Volume 1: Europe(Geneva: International Labour Office). -----(2017), World Social Protection Report 2017-19: Universal Social Protection to Achieve the Sustainable Development Goals(Geneva: International Labour Office). ----(nd),‘Strategies and Tools against Social Exclusion and Poverty(STEP): mainstreaming gender in the ILO Global Programme, STEP’, available at http://www.social-protection.org/gimi/gess/RessourcePDF.action;jsessionid. Ishay, Micheline(2004), The History of Human Rights: From Ancient Times to the Globalization Era(Berkeley: University of California Press). Kabandula, Abigail, and Jeremy Seekings(2016),‘ Donor influence, the Minister of Finance and welfare policy reform in Zambia, 200311’, CSSR Working Paper 395 (Cape Town: Centre for Social Science Research, University of Cape Town). KalebeNyamongo, Chipiliro, and Heather Marquette(2014),‘ Elite Attitudes Towards Cash Transfers and the Poor in Malawi ’, DLP Research Paper 30(Birmingham: Developmental Leadership Programme, University of Birmingham). Kott, Sandrine(2010),‘Constructing a European Social Model: The fight for social insurance in the interwar period’, in Magaly Garcia Rodriguez, Jasmien Van Daele, Marcel van der Linden and Geert Van Goethem(eds), The ILO: Past and Present (Bern: Peter Lang): 173-95.

143 Langford, Malcom, Ben Cousins, Jackie Dugard, and Tshepo Madlingozi(eds, 2013), Socio-economic Rights in South Africa: Symbols or Substance?(Cambridge: Cambridge University Press). Lavers, Tom(2016a),‘Understanding elite commitment to social protection: Rwanda’s Vision 2020 Umurenge Programme’, UNU-WIDER Working Paper 93-2016 (Helsinki: UNU-WIDER). ----(2016b),‘Social protection in an aspiring‘developmental state’: The political drivers of Ethiopia’s PSNP’, UNU-WIDER Working Paper 130-2016(Helsinki: UNUWIDER). ----and Sam Hickey(2016),‘Conceptualising the politics of social protection expansion in low income countries: The intersection of transnational ideas and domestic politics’, International Journal of Social Welfare 25: 388-98. Leisering, Lutz(2009),‘Extending Social Security to the Excluded: Are Social Cash Transfers to the Poor an Appropriate Way of Fighting Poverty in Developing Countries?’, Global Social Policy 9,2(August): 246-72. Leutelt, Michael(2 012),‘HelpAge’s Involvement in Spreading Social Pensions in the Global South: Slow and Steady Wins the Race? ’, FLOOR Working Paper no. 16 (Bielefeld: University of Bielefeld). -----(2016), Perspectives on Policy Transfer: How International Organisations Promote Social Cash Transfer Programs in the Global South, PhD thesis, University of Bielefeld. Levitt, Peggy, and Sally Merry(2009),‘Vernacularization on the ground: local uses of global women's rights in Peru, China, India and the United States’, Global Networks 9,4: 441-461. Liebenberg, Sandra(2012),‘The Right to Social Assistance: The Implications of Grootboom for Polict Reform in South Africa’, New Agenda 3 rd quarter: 31-47. ----- and Geo Quinot(eds, 2012), Law and Poverty: Perspectives from South Africa and Beyond(Cape Town: Juta). Lindholt, Lone(2001), Questioning the Universality of Human Rights: The African Charter on Human and Peoples’ Rights in Botswana, Malawi and Mozambique( Aldershot: Ashgate). Luebbert, Gregory(1984), Liberalism, Fascism, or Social Democracy: Social classes and the political origins of regimes in interwar Europe(Oxford: Oxford University Press). Lund, Francie(2008), Changing Social Policy(Cape Town: HSRC Press). ----(2009),‘Social protection, citizenship and the employment relationship’, WIEGO Working Paper(Social Protection) no.10(Cambridge, MA: WIEGO). ----and S. Srinivas(2000),‘Learning from Experience: a gendered approach to social protection for workers in the informal economy’(Geneva: ILO). Mandala, E. C.(2005), The End of Chidyerano: A History of Food and Everyday Life in Malawi, 1860-2004(Portsmouth: Heinemann). Marshall, Thomas H.(1950),‘Citizenship and Social Class’, republished in T.H.Marshall and Tom Bottomore, Citizenship and Social Class(London: Pluto, 1992). MartinezTorres, Maria Elena, and Peter M. Rosset(2010),‘La Vía Campesina: the birth and evolution of a transnational social movement’, Journal of Peasant Studies 37,1: 149-175.

144 Matisonn, Heidi, and Jeremy Seekings(2003),‘The Politics of the Basic Income Grant in South Africa, 19962002’, in Guy Standing and Michael Samson(eds), A Basic Income Grant for South Africa(Cape Town: University of Cape Town Press): 5676. Mauritius(2008), Digest of Social Security Statistics, 2008(Port Louis: Central Statistics Office). Merry, Sally Engle(1997),‘Changing rights, changing culture’, in Jane Cowan, Marie Benedicte Dembour and Richard Wilson(eds), Culture and Rights: Anthropological Perspectives(Cambridge: Cambridge University Press): 31-55. Meth, Charles(2004),‘Ideology and social policy:“Handouts” and the spectre of “dependency”’, Transformation 56: 1-30. Mokgoro, Yvonne(2004), Judgement, in Louis Khosa and others versus Minister of Social Development and others(CCT 12/03), and Saleta Mahlaule and Altinah Hlungwana versus Minister of Social Development and others(CCT 13/03), Constitutional Court of South Africa. Munro, L.(2008).‘Risk, Needs and Rights’, in Armando Barrientos and David Hulme (eds), Social Protection for the Poor and Poorest: Risk, Needs and Rights (Basingstoke: Palgrave). Mutua, Makau(2002),‘The Banjul Charter: The Case for an African Cultural Footprint’, in A.A.AnNa’im(ed), Cultural Transformation and Human Rights in Africa(London: Zed): 68-107. NiñoZarazúa, Miguel, Armando Barrientos, Sam Hickey and David Hulme(2012).‘Social Protection in SubSaharan Africa: Getting the Politics Right’. World Development, 40(1): 163 – 76. OAU(1990), African Charter on the Rights and Welfare of the Child(Addis Ababa: African Union). Olivier, Marius(2009),‘ Reflections on the Draft AU Social Policy Framework for Africa, With Particular Reference to Social Protection Perspectives’, paper presented at the 4th Citizens Continental Conference on the 12th AU Summit, Addis Ababa, 16 – 17 January. Pal, K., C. Behrendt, F. Leger, M. Cichon, and K. Hagemejer(2005). ‘Can Low -income Countries Afford Basic Social Security? First Results of a Modelling Exercise’. Issues in Social Protection Discussion Paper 13. Geneva: International Labour Organization Paremoer, Lauren, and Courtney Jung(2012),‘The Role of Social and Economic Rights in Supporting Opposition in Postapartheid South Africa’, in Ian Shapiro and Kahreen Tebeau(eds), After Apartheid: Reinventing South Africa(Charlottesville: University of Virginia Press): 199-230. Patel, Leila(2005), Social Welfare and Development in South Africa(Cape Town: Oxford University Press). Pellerano, Luca, and P. Matandiko(2016),‘More Than Business Alone: extending and promoting social protection amongst MSMEs and informal workers in Zambia’s building construction industry’,(Lusaka: Government of the Republic of Zambia, Ministry of Foreign Affairs of Finland, and United Nations Zambia). Paukert, Felix(1968),‘Social Security and Income Distribution: A Comparative Study’, International Labour Review 98,5: 425-50.

145 Phe Goursat, M, and Luca Pellerano(2016),‘Extension of Social Protection to Workers in the Informal Economy in Zambia: lessons from field research on domestic workers, small scale farmers and construction workers’(Lusaka: ILO). Polanyi, Karl(1944), The Great Transformation(New York: Farrar and Rinehart). Prasad, Monica(2012), The Land of Too Much: American abundance and the paradox of poverty(Cambridge: Harvard University Press). Proudlock, Paula(2011),‘Lessons learned from the campaigns to expand the Child Support Grant in South Africa’, in Sudhanshu Handa, Stephen Devereux and Douglas Webb(eds), Social Protection for Africa’s Children (London: Routledge). Pruce, Kate, and Sam Hickey(2017),‘The Politics of Promoting Social Protection in Zambia’ , ESID Working Paper 75(Manchester: University of Manchester, Effective States and Inclusive Development). SADC(2003), Charter of Fundamental Social Rights in SADC(Gaborone: Southern African Development Community). -----(2007), Code on Social Security in the SADC(Gaborone: Southern African Development Community). Samson, Michael, and C. Cherrier(2009).‘Feasibility Study for a Social Cash Transfer Programme as a Lead Instrument in Childcentered Social Protection in Senegal’ (Cape Town: Economic Policy Research Institute). Schnitzer, P.(2011), Design Options, Costs and Impacts of a Cash Transfer Program in Senegal: Simulation Results(Dakar: UNICEF). Scott, Zoe(2007),‘The changing face of“Drivers of Change”’, Public Administration and Development 27,1: 85-90. Seekings, Jeremy(2002),‘The Broader Importance of Welfare Reform in South Africa’, Social Dynamics 28,2: 1-38. ----(2007),‘“Not a Single White Person Should be Allowed to Go Under”: Swartgevaar and the Origins of South Africa’s Welfare S tate, 19241929’, Journal of African History, 48,3: 375-94. -----(2008a), ‘ A Comparative Analysis of Welfare Regimes in the South: The social origins of social assistance outside of the established democracies’, in Ian Shapiro, Peter Swenson and Daniella Donno(eds), Divide and Deal: The Politics of Distribution in Democracies(New York: NYU Press). ----(2008b),‘ The ILO and Social Protection in the Global South, 19192005’, paper for the ILO Century Project(Geneva, 2008) and also as CSSR Working Paper no.238 (Cape Town: Centre for Social Science Research, University of Cape Town). ----(2010),‘ The ILO and Welfare Reform in South Africa, Latin America and the Caribbean, 191950’ , in Magaly Garcia Rodriguez, Jasmien Van Daele, Marcel van der Linden and Geert Van Goethem(eds), The ILO: Past and Present(Bern: Peter Lang): 145-72. ----(2011),‘British Colonial Policy, Local Politics and the Origins of the Mauritian Welfare State, 19361950’, Journal of African History 52,2: 157-77. ----(2013).‘The Beveridge Report, the Colonial Office and Welfare Reform in British Colonies’. Unpublished paper, University of Cape Town. ----(2014)‘Are African welfare states different? Welfare state -building in(Anglophone) Africa in comparative perspective’, paper presented at conference on Social Policy and Regimes of Social Welfare in Africa, University of Fribourg(September).

146 ----(2015),‘Income support through the welfare state’, in Jeremy Seekings and Nicoli Nattrass, Policy, Politics and Poverty in South Africa(London: Palgrave). ----(2016a)‘Drought relief and the origins of a conservative welfare state in Botswana, 1965-1980. CSSR Working Paper 378(Cape Town: Centre for Social Science Research, University of Cape Town). ----(2016b),‘Building a conservative welfare state in Botswana’, CSSR Working Paper 388(Cape Town: Centre for Social Science Research, University of Cape Town). ----(2016c),‘“A lean cow cannot climb out of the mud, but a good cattleman does not leave it to perish”: The origins of a conservative welfare doctrine in Botswana under Seretse Khama, 19661980’, CSSR Working Paper 387(Cape Town: Centre for Social Science Research, University of Cape Town). ----(2016d),‘ Welfare regimes and distribution across the global South: Theory and evidence in the construction of typologies’, CSSR Working Paper 394(Cape Town: Centre for Social Science Research, University of Cape Town). ----(2016e),‘The introduction of universal old age pensions in Zanzibar’, CSSR Working Paper 393(Cape Town: Centre for Social Science Research, University of Cape Town). ----(2016f),‘Redefining the affordability of social assistance prog rammes: The Child Support Grant in South Africa, 19982014’, CSSR Working Paper 379(Cape Town: Centre for Social Science Research, University of Cape Town). ----(2017a),‘“Affordability” and the political economy of social protection in contemporary Afr ica’, UNU-WIDER Working Paper 2017/43(Helsinki: UNU-WIDER). -----(2017b), ‘The discourse of dependency and the agrarian roots of welfare doctrines in Africa: The case of Botswana’, SozialPolitik.Ch 2. ----- and Heidi Matisonn( 2012),‘The Continuing Politics of Basic Income in South Africa’, in Carol Pateman and Matthew Murray(eds), Horizons of Reform: Basic Income Solutions around the World(London: Palgrave Macmillan). ----and Elena Moore(2014),‘Kin, Market and State in the Provision of Care in South Africa’, Soziale Welt, Sonderband 20: 435-50. ----- and Nicoli Nattrass(2005), Class, Race and Inequality in South Africa(New Haven: Yale University Press). ----- and-----(2015), Policy, Politics and Poverty in South Africa(London: Palgrave Macmillan). Sen, Amartya(1981), Poverty and Famines. An Essay on Entitlement and Deprivation (Oxford: Oxford University Press). Sepúlveda, Magdalena, and Carly Nyst(2012), The Human Rights Approach to Social Protection(Helsinki: Ministry for Foreign Affairs). Shivji, Issa(1989), The Concept of Human Rights in Africa(Dakar: CODESRIA). Siachiwena, Hangala(2016). ‘ Social Protection Policy Reform in Zambia in Zambia during the Sata Presidency, 2011 – 2014 ’, CSSR Working Paper 377(Cape Town: Centre for Social Science Research, University of Cape Town). ----(2017),‘Social policy reform in Zambia under President Lungu, 2015 2017’, CSSR Working Paper 403(Cape Town: Centre for Social Science Research, University of Cape Town). SlothNielsen, Julia(2016),‘The Protection of Children’s Economic, Social and Cultural Rights under the African Children’s Charter’, in Danwood Chirwa and Lilian Chenwi

147 (eds), The protection of Economic, Social and Cultural Rights in Africa(Cambridge: Cambridge University Press): 155-79. South Africa(1996), Annual Report of the Department of Welfare 1996(Pretoria: Department of Welfare). -----(1997), White Paper for Social Welfare(Pretoria: Department of Welfare). -----(2002), Report of the(Taylor) Committee of Inquiry into a Comprehensive System of Social Security for South Africa(Pretoria: Department of Social Development). Ssenyonjo, Manisuli(2016),‘The protection of economic, social and cultural rights under the African Charter’, in C hirwa and Chenwi(eds), The Protection of Economic, Social and Cultural Rights in Africa: 91-120. Strang, David, and Patricia Mei Yin Chang(1993),‘The International Labor Organization and the Welfare State: Institutional Effects on National Welfare Spending, 196080’, International Organization 47,2(Spring): 235-62. TASAF(2017),‘Second quarter progress implementation report, October -December 2016’(Dar es Salaam: Tanzania Social Action Fund). Ulriksen, Marianne(2011),‘Social policy development and gl obal financial crisis in the open economies of Botswana and Mauritius’, Global Social Policy 11,2-3: 194-213. ----(2012),‘Welfare policy expansion in Botswana and Mauritius: Explaining the causes of different welfare regime paths’, Comparative Political Studies 45,12: 1483-1509. ----(2016),‘Ideational and Institutional Drivers of Social Protection in Tanzania’, UNUWIDER Working Paper 2016/142(Helsinki: UNU-WIDER). ----(2017),‘ Mineral wealth and limited redistribution: social transfers and taxation in Botswana, Journal of Contemporary African Studies 35:1, 73-92. UNICEF(2012),‘Integrated Social Protection Systems: Enhancing Equity for Children’, UNICEF Social Protection Framew ork’. ----(2013),‘The Situation Analysis of Children and their Families in Botswana: Summary Report’(Gaborone: UNICEF). UN(1999),‘The Right to Adequate Food’, General Comment 12(United Nations Committee on Social, Economic and Cultural Rights). UN Human Rights Council(2012),‘ Final study of the Human Rights Council Advisory Committee on the advancement of the rights of peasants and other people working in rural areas’( United Nations Human Rights Council) UNDP(2005), Botswana Human Development Report(Gaborone: UNDP). UPND(2016), Our plan for Zambia: 2016-2021(Lusaka: United Party for National Development). van Ginneken, W.(ed.)(1999), Social Security for the Excluded Majority: Case-studies of Developing Countries(Geneva: International Labour Office). Van Kersbergen, Kees(1995), Social Capitalism: A Study of Christian Democracy and the Welfare State(London: Routledge). Vaughan, Megan(1987), The Story of an African Famine: Gender and Famine in Twentieth-century Malawi(Cambridge: Cambridge University Press). Voipio, Timo(2007),‘Social Protection for Poverty Reduction: The OECD/DAC/POVNET view’, IDS Bulletin 38,3: 45-50. Von Gliszczynski, Moritz(2015), Cash Transfers and Basic Social Protection(London: Palgrave).

148 -----and Lutz Leisering(2016),‘Constructing new global models of social security: How international organizations defined the field of social cash transfers in the 2000s’, Journal of Social Policy 45,2: 325-343. Waterbury, John(1987),‘The Seneg alese Peasant: How good is our conventional wisdom?’, in Mark Gersovitz and John Waterbury(eds), The Political Economy of Risk and Choice in Senegal(London: Frank Cass): 47-73. Weigand, Christine, and Margaret Grosh(2008),‘Levels and Patterns of Safety Net Spending in Developing and Transition Countries’, SP(Social Protection) Discussion Paper no.0817(Washington DC: World Bank). Wilson, Stuart, and Jackie Dugard(2011),‘Taki ng Poverty Seriously: The South African Constitutional Court and SocioEconomic Rights’, Stellenbosch Law Review 22: 664-82. World Bank(2012), Managing Risk, Promoting Growth: Developing Systems for Social Protection in Africa: The World Bank’s Africa Soc ial Protection Strategy 2012-2022 (…) ----(2013),‘Republic of Senegal: Social Safety Net Assessment’, Report no. ACS7005 (Washington DC: World Bank). ----(2014),‘Project Appraisal Document … for a Social Safety Net Project’(Washington DC: World Bank). -----(2015a), The State of Social Safety Nets 2015(Washington DC: World Bank). -----(2015b), Botswana Poverty Assessment, Report 88473-BW(Washington DC: World Bank). -----(2017), The State of Social Safety Nets 2017(Washington DC: World Bank). WFP(2013), State of School Feeding Worldwide(Rome: World Food Programme). Wright, Gemma, and Michael Noble(2010),‘Recent Social Policy Developments in Africa’. Global Social Policy 10,1: 111-19. Wujczyk, Marcin(2016),‘The constitutionalisation of social security rights as the way to a social justice state’, in ILO, The Right to Social Security in the Constitutions of the World: Broadening the moral and legal space for social justice, Volume 1: Europe(Geneva: ILO): 1-6. Yeates, Nicola(2017),‘Beyond the Nation State: How can regional social policy contribute to achieving the sustainable development goals?’ UNRISD Issue Brief 5(Geneva: United Nations Research Institute for Social Development). Zambia(2006a), Vision 2030(Lusaka). -----(2006b), Fifth National Development Plan, 2007-2010(Lusaka). -----(2011), Sixth National Development Plan(Lusaka). -----(2013), Revised Sixth National Development Plan(Lusaka). -----(2014), National Social Protection Policy(Lusaka). Ziegler, Jean, Christophe Golay, Claire Mahon and Sally-Anne Way(2011), The Fight for the Right to Food: Lessons learned(Basingstoke: Palgrave Macmillan).