South Africa’s hybrid care regime: The changing and contested roles of individuals,families and the state after apartheid

Introduction

South Africa has long been characterised by deep inequalities in almost every aspect of life, including access to care. Under apartheid, racialised restrictions on where people could live, what work they could do, and what they could earn framed inequality in income and hence access to care. The state provided primarily for the minority white population. The erosion and then end of apartheid resulted in major reforms of most areas of public policy, but inequality and poverty persist, as high incomes for skilled and professional people generate prosperity for some but mass unemployment and landlessness ensure poverty for many. The state now provides for many South Africans. But the post-apartheid care regime remains characterised by inequality. Moreover, pro-poor reforms have generated a conservative backlash.

South Africa’s care regime comprises a hybrid mix of market, state and kin that does not fit neatly into any of the categories identified in typologies focused on the Global North (Seekings and Moore, 2014). The regime has ‘liberal’ features: the rich buy childcare and domestic labour, insure themselves against a variety of risks, and, in their old age, purchase personal care either in their own homes or in privately run residential institutions; the state provides means-tested benefits for the ‘deserving’ poor (the elderly, disabled, and mothers or other caregivers with children) as well as public employment and school feeding programmes. Some of these social assistance programmes are, however, more generous, and their coverage is far more extensive than one would expect in a liberal regime: the elderly and disabled receive pensions worth more than US$100 per month, which is more than some people earn while working; in total, almost one in three people in South Africa receives a tax-financed monthly grant or pension, with a total cost of about 3.5% of GDP – which is more than in any other major middle- or low-income country. The state also spends increasingly heavily on childcare through ‘early childhood development’ facilities for pre-school children. At the same time, the family continues to play a very important role, with large numbers of people dependent on kin for either financial support or personal care, and extended families remain important. The welfare and care regime thus combines conservative and social democratic features alongside liberal ones.

This mix of liberal, social democratic and conservative features is rooted in South Africa’s unusual quasi-colonial and racist history. South Africa’s hybrid regime was forged in the context of colonialism, European settlement and immigration, and institutionalised racial segregation and discrimination. The economic privileges of the European (or ‘white’) population were secured primarily through the reservation of land, well-paid occupations and better business opportunities. A set of classically ‘liberal’ policies provided ‘residual’ support and care to poor white people, raising them (and especially their children) out of poverty and disciplining them into conformity with the norms of white supremacy. The indigenous ‘Native’ or African majority population was not only subject to discriminatory economic policies, which kept most in poverty, but was also largely excluded from the welfare state on the grounds that support and care were sufficiently and appropriately provided by kin within the ‘traditional’ or ‘tribal’ system. The racist dualism of the care regime was reflected in domestic labour: State regulation ensured that white men could support their families at a living standard that included employment of a low-paid domestic worker, who was usually African or ‘coloured’; ¹ the domestic worker herself was often compelled to live apart from her own children, who were cared for in a rural area by their grandmother.

In neither design nor practice was the system ever as neatly dualistic as the ideology of institutionalised racial segregation would suggest. ‘Welfare state’-builders in the early twentieth century sought to enforce the responsibilities of extended kin within white as well as African families. Social and economic change meant that the ‘radius’ of such kin responsibility shrank over time, and women and children increasingly came to be seen to have rights independent of their fathers, husbands or brothers, among many African as well as white families. Even under apartheid, public provision expanded for the African population, family law was reformed, and the perception that people had rights to grants and pensions became widespread across the whole population. Overall, however, the formal side of the care regime, providing for white and some African people, retained a distinctly ‘liberal’ character, while the enduring emphasis on the family for the African population meant that the informal side of the regime retained distinctly conservative features.

As apartheid limped towards its demise, the government moved to eliminate explicit racial discrimination in both its labour market policies and its welfare programmes. The real value of the old-age pension paid to white South Africans was allowed to decline, while the real value of the pension paid to African men and women was raised, until parity was finally achieved in 1993. Discrimination continued in other parts of the welfare state, however. Most African women were not eligible for the grants paid to single mothers. Institutional care also remained deeply discriminatory (as well as segregated). The state also continued to exempt domestic workers from regulation, ensuring that rich families could continue to enjoy the benefits of cheap domestic care. The post-apartheid state inherited deep inequality and poverty, only partially mitigated by the welfare and care regime.

This article examines South Africa’s current care regime, as a contribution to the comparative study of care regimes. It looks at how the care regime has been reformed since the end of apartheid but retains colonial features in both the design of public policies and the diversity of norms and practices in society. We focus primarily on access to care among the poor, almost all of whom are African (in post-apartheid racial terminology), contrasting this with access among the rich, who include most white South Africans along with growing numbers of African people. This article does not consider separately state provision for South Africa’s ‘coloured’ or ‘Indian’ minorities, nor for all classes.

Deracialising the provision of care

The African National Congress (ANC) under Nelson Mandela was elected into government in the first democratic elections, in 1994. The ANC moved quickly to remove the last vestiges of explicit and direct racial discrimination from the welfare state by extending the ‘privileges’ of public provision enjoyed by white citizens to the entire population. The most urgent tasks were the removal of supposed bureaucratic obstacles to access by African people to disability grants, the extension of legal access to grants for the support of poor children to African families, and the desegregation and reform of funding for residential institutions, especially for the elderly. The deracialisation of social assistance programmes entailed a massive expansion in the number of beneficiaries, which served to buttress support for the ANC (Bond, 2014; Seekings and Nattrass, 2015). This deracialisation of policy was insufficient to address the underlying inequalities and poverty, rooted in high (and rising) unemployment and landlessness. Moreover, poverty was soon exacerbated by AIDS-related illness and death. As many working-age adults became ill and died, older people were called on to support growing numbers of dependent children and, especially, grandchildren. The extension of existing social assistance programmes did provide some support. For example, many doctors signed off on applications for disability grants by unemployed patients who were physically capable of performing light work, because there was no such work available and the patients needed a cash income to maintain basic health (Kelly, 2016). Deracialising or extending existing programmes provided cash support for some categories of deserving poor but failed to address care needs across the population as a whole. In addition, such policy risked imposing an inappropriate and Anglocentric model of public welfare and care on the previously excluded African population.

There were therefore also strong calls to transform rather than simply extend the programmes that existed for white citizens. Three rather different arguments were made for transformation. A rights-based argument pushed reforms in a liberal direction: progressive technocrats and organisations in civil society pushed for the reform of public policy in line with the constitutional recognition that women and children were autonomous, rights-bearing individuals. The second argument was developmental: faced with poverty and high unemployment, what was needed was ‘development’ and job creation, not ‘handouts’, as President Mandela himself stated in his inaugural presidential address in May 1994. The third argument focused on the importance of the family, and this pushed reform in a more conservative direction.

Many existing programmes designed for white South Africans reflected the assumption that public provision was required because of the breakdown of the nuclear family. Maternal and child poverty, for example, were viewed as the result of the absence of a male breadwinner. Within the African population, however, the nuclear family was far from normal in terms of either prevalence or ideology. Despite social and economic change, extended kinship remained practically and normatively important within a significant part of the African population. Research on family norms in the late 1990s revealed marked differences between white and rural African people, with urban African people subscribing to a mix of these two groups’ views (Russell, 2003). For white South Africans, for example, marriage was a relationship between two autonomous individuals, while rural African people continued to view marriage as the incorporation of wives into the husband’s family and the connection between two families. This had clear implications for care. Placing the elderly in an old-age home was widely accepted among white people – and among urban African people – but was strongly opposed within the rural African population. Conversely, the practice of grandchildren being sent to rural areas to live with grandparents was endorsed by some African men and women (especially rural African men) but not by white people.

In practice, the government’s options were soon constrained by severe fiscal crisis. ² Deracialisation could not be just a simple expansion. Having at first encouraged a rapid expansion in the disability grant programme among African people, the government shifted to rein in and even reverse this (Kelly, 2013). Faced with estimates that the cost of deracialising the existing programme for poor mothers and children would be about 2% of GDP, the government considered abolishing the existing programme entirely. Instead, it appointed a committee which recommended replacing the existing programme with a new grant – the Child Support Grant – that would reach young children regardless of race, but would pay much lower benefits than the existing programme, with the total cost not rising above that of the existing programme (Lund, 2008). Expanding the existing residential care system for elderly white people to the entire population would require the construction of approximately 1400 new old-age homes, at massive cost. The state chose to desegregate existing institutions without increasing expenditure or the total number of places available.

These reforms were shaped by ideological factors as well as fiscal constraints. Building old-age homes for elderly African men and women was widely considered to be inappropriate (Lund, 1992: 314). The new Child Support Grants were paid to poor caregivers, not necessarily mothers, and without the former criterion of paternal absence (Lund, 2008: 51–54). As perceived fiscal constraints eased, so the government extended the Child Support Grants to older children. By March 2014, a total of 11 million child grants were paid out each month, and the cost of the programme had risen to more than 1% of GDP (Seekings, 2016). But courts invoked both customary and civil law in enforcing obligations on extended kin so as to reduce public expenditure. The Foster Care Grant, paid to people who had been awarded the care of children by the courts, was more generous than the modest Child Support Grant, and was thus attractive to extended kin looking after orphaned and even non-orphaned children. In a 2012 case, a court initially decided that grandparents and siblings had a ‘duty of support’ and were therefore ineligible for the grant, even if they were caring for children who had been abandoned or orphaned. Uncles and aunts had no such duty, however, and were therefore eligible (South Gauteng High Court, 2012). This decision was revised by the High Court, however. The High Court agreed that grandparents had a duty of support, but nonetheless could be appointed as foster parents and thus could be eligible for a Foster Care Grant (News24, 2013).

Customary law was also reformed. The 1998 Recognition of Customary Marriage Act provided for equality between men and women under customary marriage. Democratisation also led to reforms of the customary law of succession. Hitherto, the principle of male primogeniture was applied in the event of a man dying intestate. This was challenged successfully in Bhe v. Magistrate Khayelitsha, in the Constitutional Court (South Africa Constitutional Court, 2005).

In practice, both public programmes and family law were deracialised more than they were transformed. Attempts to institutionalise extended kinship obligations did not get very far. The Child Support Grant could be paid to kin other than mothers, and was supposed to ‘follow the child’, but the vast majority of recipients were mothers (even when other kin were in fact caring for the child). The state lacked the means or will to enforce the kinship obligations that were supposedly recognised under both civil and customary law; it was up to individuals to bring cases, for example claims for child maintenance or compensation for road accidents. Even customary law reforms built on the legal protections offered to white citizens under apartheid (Moore and Himonga, 2016). The core architecture of the welfare and care regime in post-apartheid South Africa in the 2010s thus remained what it had been for white citizens prior to apartheid, 75 or so years before. Reforms were modest even with respect to domestic labour. White South African families had long enjoyed the privilege of cheap domestic labour, through state policies that ensured a steady supply of female domestic labour free of any regulation of wages and conditions of employment. Democratisation led to domestic workers being covered by the same battery of rights and regulations as any other workers, and minimum wages doubled in real terms – but they remained low, lest jobs were lost (Budlender, 2013). One reform that was massively consequential was the abolition of racist restrictions on where African people could live, which meant that children could live with their parents in urban areas Domestic workers could live with their families. Despite the abolition of legal restrictions, most elderly people remained in (or retired to) the countryside, and many parents chose to leave children in rural areas with grandparents.

These reforms of social assistance and other welfare programmes, of customary law and of employment regulations, all eroded the bifurcated and dualistic system that apartheid ideologues had sought to institutionalise. In the apartheid vision, white people enjoyed support and care primarily through the market and state, while African people – who had little access to care provided through either the market or the state – were pushed into dependence on kin. The post-apartheid state built on the foundations of the system that existed for white citizens, with only modest reforms of the basic architecture. These reforms fell short of what was needed to address either poverty or inequality. The rich relied on market provision (including pensions, medical aid and personal care). For the majority of the population, however, personal care for the elderly and for children continued to be left largely to kin. Public provision for working-age adults was limited to the disabled and mothers (through child grants) and workfare (that provided for only a very small proportion of the unemployed). Kin support remained integral to the care regime.

This dominance of extension (through deracialisation) over transformation was characteristic of many other areas of public policy also, with path dependency rooted in class and racial politics (Marais, 2013; Seekings and Nattrass, 2005, 2015). In his discussion of the continuities between apartheid and post-apartheid welfare policy, Barchiesi (2011) points specifically to the enduring privileging of work or employment in elite ideology. In his account, citizenship has been reduced, ideologically, to labour. Indeed, an ideological emphasis on the dignity of work and the dangers of ‘dependency’ is widespread across much of Africa (Seekings, 2017). Nonetheless, public programmes define a large minority of South Africans as deserving of public support, and this has subverted understandings of who deserves what, when and why (Ferguson, 2015). Many South Africans navigate between the opportunities offered through markets, the state and kin, and between the norms and values accompanying these. It is the complexity of the institutional, ideological and social landscape that is distinctively South African.

Norms and practices of care in the new South Africa

While policy reforms were insufficient to address the challenges of poverty and inequality, they nonetheless outraged conservatives and prompted a political and social backlash. This meant that poor South Africans had to navigate not only through the mix of public and private provision, but also through competing social norms and values. Lower-income African men and women face acute challenges in navigating their way through the shifting institutional, social and cultural landscape. Without the resources to purchase care easily through the market, they must balance the opportunities provided by public programmes with their enduring responsibilities and obligations to kin. This navigation is complicated by two factors. First, public programmes focus for the most part on individuals, while family norms involve these same individuals in networks of kin (and, less often, community) with norms as to what it means to be a ‘good’ daughter, son, mother, father or grandmother. Second, economic change has at the same time generated opportunities (for example, for women to work) as well as pressures (most obviously, very high unemployment and hence dependency). Discourses of tradition, family and interdependence collide with discourses of rights, individuality and economic independence. The reformed public welfare system may retain some Anglocentric assumptions relating to care provision within families, but individuals must negotiate their claims on, and the demands made on them by, a wide range of extended kin. In practice, kinship no longer entails ‘inescapable moral claims and obligations’, as Fortes famously wrote (Fortes, 1969: 242). Support for most kin has become highly conditional, and this is especially true among more distant kin (Harper and Seekings, 2010).

Historically, the elderly in rural areas were the most vulnerable to poverty. Widows were especially vulnerable, if or when their sons or their late husbands’ other male kin had migrated but failed to remit any share of their earnings. The dramatic increase in the 1980s and early 1990s in the value of old-age pensions transformed the position of pensioners in kin networks. By the 1990s, pensioners were supporting entire households, including children and grandchildren. ‘Good’ parents were supposed to use their pensions to assist their financially vulnerable children and grandchildren, even if this stretched them financially; ignoring these obligations would not only be considered ‘morally outrageous but tantamount to the denial of the very kinship relationship itself’ (Sagner and Mtati, 1999: 401). In supporting their younger kin, pensioners were strengthening as well as reflecting norms of interdependence and mutual responsibility, and the subordination of individual interests to collective ones. At the same time, the elderly were strengthening their authority and their future claims to reciprocal support and care from younger kin (Bak, 2008; Button, 2017; Hoffman, 2016; Kimuna and Makiwane, 2007; Sagner and Mtati, 1999; Schatz, 2007; Schatz and Ogunmefun, 2007).

Young adults’ acceptance of their responsibilities to support their elders is, however, far from unconditional. Hoffman (2016) found that young adults expressed a willingness to care for their older kin only insofar as they had the means to do so and only when this did not interfere with the care of their children and spouses. Some younger adults fail to support their elders even when they are employed (Button, 2017). Individualist norms impede responsibilities to kin. Mathis (2011) found that employed young women tried to limit their financial obligations towards their parents by speaking of themselves as rights-bearing individuals, in contrast to the discourse of tradition used by older people. When employed adult children failed to make financial contributions towards their households, their elders considered them to be ‘uncaring’:

Today’s children, they don’t care about their parents … You find out that you raise your children but once your children get their job, they cannot support you. Instead of supporting you, your children use their money for their personal use, buying clothes, all those things. They forget that at the time they weren’t working, you were the one that was supporting them. (Unpublished quote; see Button, 2017)

This assertion of individual interests over collective well-being, and the implicit denial of both interdependence and reciprocity, is perceived to be widespread, even if there is little evidence about its actual extent.

In addition to financial support, older women provide care, especially to sick family members and to grandchildren (Ardington et al., 2010; Chazan, 2008; Fakier and Cock, 2009; Mosoetsa, 2011; Schatz, 2007; Schatz and Ogunmefun, 2007). ‘Good’ grandmothers are moral guides and teachers, passing on knowledge and values as traditions. Caring for grandchildren entails socialising them into appropriate attitudes and behaviours, including on gender roles, respect for elders, the importance of education and the undesirability of teenage pregnancy. Grandmothers generally describe the role as a joy, but many also point to the financial, emotional and physical burden involved, especially when grandchildren are unruly, disobedient and disrespectful (Blake, 2015; Bohman et al., 2009; Button, 2017; Møller and Sotshongaye, 2002).

Just as grandmothers have come to be seen as providers, so being a ‘good’ mother also increasingly entails providing financially for children (Blake, 2015: 46; Moore, 2013: 153). Rising female employment rates and declining rates of marriage and even cohabitation mean that more and more women are the primary breadwinners. At the same time, for many younger mothers, high rates of unemployment and low earnings frustrate the achievement of good motherhood. In this context, the Child Support Grant has come to play an important role, including through imparting some dignity to otherwise impoverished young mothers (Blake, 2015; Wright et al., 2015). ‘Good’ motherhood requires, however, that mothers spend their grants on their children (Blake, 2015: 47). Most mothers do use their grants to purchase food and clothing for their children, and to cover their schooling expenses (De Koker et al., 2006; Surrender et al., 2010; Wright et al., 2015). But the recipients are subjected to public scrutiny, judgement and prejudice by neighbours and kin who look out for evidence that recipients are indulging their own consumerist desires (Blake, 2015; Wright et al., 2015). Therefore grants are both empowering and potentially disempowering, as they prompt more intense surveillance.

The Child Support Grant has also prompted criticism and opposition among believers in a patriarchal order. Many men and some women, particularly in rural areas, disapproved of young, and especially unmarried, women controlling resources. Tensions intensified if the young, female grant recipients failed to contribute to their households’ expenses, even in cases where there was no similar expectation of young men. Young women who defied patriarchal norms would be accused of being selfish, irresponsible and ‘bad’ daughters. This leads to hostility to the Child Support Grant itself, because the grant is seen to be morally and socially corrosive (Dubbeld, 2013; Hickel, 2015; Mathis, 2011; Mosoetsa, 2011).

While ‘good’ fatherhood has historically been synonymous with financial provision (Mosoetsa, 2011: 63), and research shows that many absent fathers do contribute financially to their children (Clark et al., 2015: 580), social and cultural change has resulted in alternative conceptions of what it means to be a ‘good’ father (Morrell et al., 2016; Russell, 2003). Almost all the rural African participants in Russell’s research believed that it was better for fathers to be absent – but financially supportive – rather than unemployed and nurturing; yet two out of three of her urban African participants expressed a preference for an unemployed father who spent time playing with and teaching his child, instead of an absentee provider (Russell, 2003: 165). While many fathers do not reside with their biological children, physical separation does not necessarily equate to absence in terms of contact or care (Bray et al., 2010; Madhavan and Roy, 2012; see also Nkani, 2014). In a study of young men in Cape Town, 30% of fathers who did not co-reside with their children said that they saw their children every day, while a further 36% reported contact with their children several times a week (Clark et al., 2015: 579). The extent of parental contact is likely to be influenced by the status of the relationship between the biological parents. Madhavan and Roy (2012) found that unmarried fathers and paternal kin were more likely to have maintained contact with a child if, at least, ‘damages’ (i.e. compensation) had been paid from the father’s family to the family of the mother in acknowledgement of paternity. Importantly, paternal contact does not necessarily imply paternal engagement in practical or emotional care work (Nkani, 2014). Men who were less violent and who drank less alcohol, who are more communicative with women, who have more gender equitable views and more positive experiences of their own parents are more likely to be engaged in childcare. Such men were more likely to engage with children by playing with them or helping with their homework than by talking to them about personal matters or washing their clothes (Morrell et al., 2016). When fathers do not perform fathering tasks, other kin often fill the gap – in contrast to most European societies (Blake, 2015; Madhavan and Roy, 2012; Madhavan et al., 2014).

Extended kinship remains very important, and a refusal to support or care is contentious. This is in part because of unemployment, illness or death among working-age adults, so that grandparents must support or care for adult children as well as grandchildren. In return, many adults and even teenagers care for their elderly parents. It is also due to the decline of marriage and even cohabitation. Posel and Rudwick (2013: 173) report that ‘by 2010, 73 percent per cent of young African women and 28 percent of older African women had never been married and were not cohabitating with a partner, compared to 52 percent of young white women and only 8 percent of older white women’. In addition, rising numbers of women divorced, including women who had been married under customary law. While it was widely acknowledged that women married under customary law had a claim to marital property should the marriage be dissolved, in practice women rarely left a marriage with anything other than personal possessions, while the men retained marital property. Women did, however, often gain custody of the children, even when the marriage had been sealed with the transfer of bridewealth, which historically meant that children ‘belonged’ to the father’s family. Divorced fathers were believed to have a moral duty to support their children, if they had the means to do so. This was perceived to be an important way in which a father could demonstrate his commitment to his children. But in practice, divorced fathers rarely paid child maintenance; when they did, payments were irregular or insufficient (Himonga and Moore, 2015).

Both in towns and in the countryside, the decline in marriage was accompanied by the transformation of extended kinship. Maternal kin played more important roles with respect to both financial support and care. Just as the state was expanding the social grant system and recognising the rights of women and children as autonomous individuals, so these same women and children were often made dependent on extended kin networks. The expanded role of the state did not so much reduce the roles played by kin, but rather helped to transform them.

The conservative backlash against the extension of public provision

The expansion of Child Support Grants in post-apartheid South Africa has proved especially contentious. It is frequently alleged that the grants reward sexual immorality and encourage teenage pregnancy, and that recipients spend them on drink, airtime and other selfish forms of consumption. Researchers find little evidence for such assertions, but they persist, suggesting that they reflect a deeper discontent with social and cultural change. Mosoetsa (2011) points to the challenge that the grants pose to patriarchy, in that young, often unmarried, women control these resources. Hickel (2015), drawing on research in rural KwaZulu-Natal in the late 2000s, presents a more sweeping argument. In his analysis, the grant is a symbol of the liberal democracy established under South Africa’s 1996 Constitution, based on individual rights, which was rejected by many men (and women) in rural KwaZulu-Natal. ‘While they embraced the principles of racial equality and universal franchise, they questioned the underlying idea that all individuals are autonomous and ontologically equal – especially in relation to gender and kinship hierarchies – and objected to what they perceived as a systematic attack on their values by the ANC and its allies’ (Hickel, 2015: 2). Hickel’s informants contrasted the self-interested individualism of liberal democracy with hlonipha, i.e. the culture of respect, in the sense of deference to a status hierarchy based on gender and age (including veneration for the ancestors) and entailing a related system of taboos. In Hickel’s account, real and deep cultural differences persisted in post-apartheid South Africa, and these were manifested in divisions over reforms of welfare programmes.

The conservative backlash has also been articulated by sections of the political elite. The ANC leadership was wary of welfare in the mid and late 1990s on developmental grounds, whereas under President Zuma criticism has taken on a more conservative form. In 2015, in a speech to traditional leaders, Zuma branded teenage mothers as irresponsible bad mothers, claimed that they were not using the Child Support Grant for their children, and suggested that they were cheating the system. Instead of being allowed to drop out of school, they should be sent to somewhere like Robben Island – the apartheid prison for political prisoners – where they could complete their schooling, empowering them to work and support their children themselves. If they were to be given grants, then the grants should not be paid in cash, which recipients could spend as they like, but rather in vouchers that could only be used to buy designated items. ‘Should we give the money or should we have vouchers that are very specific, either to buy food or uniforms for the school or to pay for the schools – so that the money will not be used for anything except the needs of the child?’, Zuma asked. For conservatives, the problem was the immorality of young women, not the economic and other structural factors that encouraged teenagers to become mothers. In the past, Zuma continued, ‘there were no pregnancies of teenagers and people built families at the right time. Why can’t we do it?’ Now, teenage mothers were a burden on their grandmothers and on society. Their grandmothers might be deserving, Zuma said, but the teenage mothers were not (Mail and Guardian, 2015).

This conservative backlash against the expansion of public welfare idealises the extended family and blames many of the pathologies of contemporary society on its fragmentation. The government’s 2012 White Paper on Families emphasised ‘self-reliance’ – i.e. the opposite of ‘dependency’ – as well as family resilience and solidarity. It proposed to return the burden of care from the state to the family, with the state playing a more supportive role. The White Paper was silent on who cares, with what resources and under what circumstances (South Africa, 2012). This silence mirrored the ways in which much of the caring activity that happens is taken for granted, made invisible and not valued (Gouws and Van Zyl, 2014). The White Paper paid little attention to the resources required for providing adequate care and the practical support required for people to carry out their commitments in order to become self-reliant. The White Paper strongly emphasised solidarity within families and implied that there is a deficit of familial, especially intergenerational, responsibility. It paid no attention to the commitment of specific contributors (such as carers, pensioners or breadwinners), or to the values that continue to underpin such contributions. While it emphasised ‘family resilience’, which it described as the ‘inherent capacities and strengths’ that sustain families during both prosperity and adversity, the White Paper was silent on even the supportive roles that the state should play.

It seems that the post-apartheid government is at least as insistent as its apartheid predecessors on family values and shifting the balance of provision back from the state to kin. While there is probably more support for conservative values than is often acknowledged, such a reactionary approach to welfare and care reform flies in the face of the massive social, economic and cultural changes that have disrupted the roles that kin used to play, as well as the principles enshrined in the democratic constitution. The wide coverage of grants and pensions, not only directly to the recipients but also indirectly to their dependents, means that any serious programmatic retrenchment would be politically costly. However fierce the denunciations by conservatives, non-rich South Africans will continue to navigate their way across a hybrid landscape of state, market and kin, negotiating with bureaucrats, neighbours and relatives.

Conclusion

In 1994 the ANC inherited a care regime based on a combination of market provision for the rich, ‘liberal’ public provision for the ‘deserving’ poor and reliance on kin for most of the poor. Deracialisation by the apartheid state in the run-up to democratisation, and by the ANC governments after 1994, meant that a system of public provision intended to privilege the white minority came to be broadened to the entire population. Including the African majority in a system set up to privilege the white minority meant that coverage expanded massively and public welfare provision ceased to be residual, taking on social democratic features. At the same time, there was pressure for the transformation of this system to take into account either the specific ‘developmental’ needs of the disenfranchised population, or norms that had so far been excluded from consideration. There was little agreement about how far the existing system for white South Africans needed to be transformed rather than simply deracialised, or about the direction and form of that transformation. Both prospective deracialisation and proposed transformation were constrained by a perceived fiscal crisis, and a widespread agreement on the imperative of restoring economic growth after a decade of stagnation. The result was that ANC-led governments after 1994 expanded some, but retrenched other, tax-financed public programmes while denouncing the ‘welfare state’ and lauding the family in distinctly conservative ways. Enduring unemployment and poverty mean that many people remain subject to considerable pressure to support a range of dependents, and there is little consensus over which poor are deserving of direct public support. On the ground, people navigate their way between public programmes, the market and the shifting obligations and responsibilities of kinship. South Africa’s care regime remains distinctively hybrid.