Gender and child sensitive social protection in South Africa

Abstract

Drawing from a 2010 study of women receiving the Child Support Grant in an urban area of South Africa, this article discusses the link between social protection, women's empowerment and the well-being of children. It appears that the Grant enhances women's power and control over household decision-making in financial matters, general household spending and child well-being. At the same time, the data show that women continue to bear the greatest burden of care in the household and that these responsibilities significantly heighten gender inequalities. Therefore, while the Grant has benefits for child well-being and women's empowerment, it cannot on its own transform unequal and unjust social relations of power. It should be working in concert with other public programmes not only to focus on children's needs but also to strive for gender equality for poor women.

Keywords:

• care responsibilities
• child sensitive social protection
• Child Support Grant
• gender sensitive social protection
• South Africa
• women's empowerment

1. Introduction

There is growing international interest in social protection policies as tools for social transformation, particularly for gender empowerment (ILO & UNDP, 2011; Sweetman, 2011) and child well-being (Handa et al., 2011). Economic and political crises have a significant and harmful effect on women and children, who are already disproportionately represented amongst the poor. Social protection measures could be an important tool to mitigate these disparate and negative impacts. Studies have been conducted to assess the gendered impact of cash transfers in places such as Latin America (Adato et al., 2000; Molyneux, 2006; Molyneux & Thomson, 2011), but research in Africa has been less extensive (see for example Patel & Hochfeld, 2011). There is scope for research of this nature to inform gender and child sensitive social policies locally, especially in South Africa, which has a well-developed social protection environment compared with other African countries.

The feminisation of social assistance is a new trend in post-apartheid South Africa. In this article we report on research in which a gender lens was used to gain insight into the impact of the Child Support Grant (CSG) and the ways in which gender influences and mediates development outcomes for women and children. Research on Old Age Pensions in South Africa indicates that money directed to women recipients of the grant has had a positive multiplier effect on women's status and the well-being of the children in their care (Ardington & Lund, 1995; Lund, 2006; Devereux et al., 2008). On the basis of these insights we hypothesised that the CSG is likely to give women more power in the household over decisions that could improve their lives and those of their children. Increased power for women in the household may have other positive benefits for women: improving their agency, giving them access to resources and opportunities, reducing their vulnerability and the burden of care, enhancing their confidence and self-esteem, broadening their social and economic participation and increasing their autonomy. In this study we assumed that the empowerment of women due to the CSG leads to positive development outcomes for children in the form of improved health and nutritional status, better care and more regular school attendance.

The three-year study from which this article draws its data was a gendered impact analysis of the CSG with reference to intra-household relations and decision-making, access to services and support, care responsibilities and addressing the vulnerabilities of poor families in urban households (Patel et al., 2012). The focus of this article is on further understanding the link between social protection, women's empowerment and the achievement of well-being for children in an urban community.

The rest of the paper is structured as follows. Section 2 describes the CSG and the context of poverty for women in urban South Africa, Section 3 reviews the literature on the link between gender and the potential of the CSG to promote social transformation and child well-being, Section 4 sets out the method used, Section 5 discusses the findings and Section 6 concludes and offers recommendations.

2. The Child Support Grant

Social grants now reach close to 30% of South Africa's population and are one of the country's most important poverty reduction programmes, reaching 15.7 million beneficiaries in 2012 (SASSA, 2012). The decline in poverty levels, especially since 2000, irrespective of which poverty line is used, is attributed largely to the expansion of cash transfers (Van der Berg et al., 2008). Despite this progress, however, poverty, inequality and high unemployment levels remain the country's greatest challenges, with poverty continuing to be strongly associated with race, gender (Posel & Rogan, 2012), education and spatial dimensions. In particular, Africans, coloured and African women, children, youth and people in rural areas remain over-represented amongst the poor. New trends are emerging in the profile of the poor: rising urban poverty, an increasing proportion of the poor living in households with low educational levels, and an increase in the gendered nature of employment, including higher unemployment levels for women than for men (Leibbrandt et al., 2010). Two-thirds of children are living in income-poor households (Hall & Wright, 2011).

In a context of widespread poverty and inequality for both women and men, the differences in poverty rates by gender have increased over the last decade (Posel & Rogan, 2012), ensuring that poverty remains a gendered phenomenon in South Africa. Women are poorer than men (Casale & Posel, 2005; Posel & Rogan, 2012). Since women are predominantly the caregivers in this society, they bear a significant burden of mitigating the effects of poverty on their households and dependants. However, social grants (in particular the CSG and the Old Age Pension) have played an important role in redistribution and poverty reduction (Woolard, 2003; Van der Berg et al., 2008). In particular, the disproportionate level of poverty for women has meant that the ‘receipt of social grant income may have been relatively more effective in reducing particularly the depth of poverty for females and female-headed households’ (Posel & Rogan, 2012:111). There is evidence that the redistributive effect of social grants is significantly stronger when the grant goes to a female recipient (Duflo, 2003), and that households as a whole generally benefit from women's grant receipt irrespective of which household member receives the grant (Case & Menendez, 2007). Because the Old Age Pension is pooled with other household income, it produces significant benefits for children's development, especially in the areas of nutrition, health and education (Case & Menendez, 2007).

Of the three publicly funded social assistance programmes for older persons, people with disabilities and children, the CSG, introduced in 1998, is the largest and fastest growing social intervention in post-apartheid South Africa. The grant is paid to the primary caregiver of a child up to 18 years of age, and now reaches more than 11 million children (SASSA, 2012), which indicates the extensive reach of the programme. The grant is means-tested, but the criteria are relatively generous, resulting in a large number of children falling into the targeted group. The vast majority of grants go to female caregivers despite men being just as eligible to apply (Vorster & De Waal, 2008).

Although the CSG is crucial to reducing poverty, overall child poverty levels have declined only marginally since the introduction of this grant (Leibbrandt et al., 2010; Hall & Wright, 2011), possibly due to the modest amount of the grant, set at R280 per month in 2012. However, researchers have noted positive developmental outcomes in the reach of the CSG and its benefits in the form of improved school attendance and better nutritional status, and also longer-term benefits of mitigating the detrimental effects of poverty on children in the early years of life (Case et al., 2005; Agüero et al., 2006; Williams, 2007; Delany et al., 2008; Neves et al., 2009; Heinrich et al., 2012).

3. Gender, social protection and social transformation

A gender and development approach informed this study (Kabeer, 1994; Chant, 2010). This approach is based on two premises: firstly that gender is socially constructed and founded on socially acquired notions of appropriate expectations of and responsibilities for men and women, and secondly that socially patterned interactions between women and men are based on asymmetrical social positions that translate into inequality of access to resources, services and opportunities and impose constraints on women (Sabates-Wheeler & Kabeer, 2003). Finally, feminists argue that because poverty is a gendered experience researchers should take a gender perspective on poverty (Kabeer, 1994; Molyneux, 2006; Chant, 2010).

There is growing interest internationally and locally in how social protection might contribute to social transformation that tackles the multiple privations that cause gender inequality (Sabates-Wheeler & Devereux, 2008). The term ‘social protection’ is now widely used to refer to cash and in-kind support, including basic and social services that are of a public, private and informal nature to mitigate poverty, inequality, vulnerability and risks associated with social and economic changes in the society and globally (ILO & UNDP, 2011). Gender is rarely used as a lens to reveal the multifaceted social dynamics and impact of social protection, broadly incorporating labour markets and public, private, voluntary, individual, household and community support systems and networks. Anti-poverty measures to reduce vulnerability tend to be gender biased, focus on the breadwinner as the target of social policy, and take insufficient account of the gendered division of labour at household level and of the bargaining power of men in intra-household decision-making and in capturing household resources (Bertrand et al., 2003). In addition, complex intra-household relations and decisions about spending priorities and trade-offs about who seeks work, who cares for children and other dependants, and the dynamics of the care economy, are poorly understood.

We know from previous international research that there is an important relationship between a woman's empowerment and the well-being of her children (Adato et al., 2000). It is widely acknowledged that cash transfers are a relatively empowering form of social protection because they offer recipients a choice of how to use them (Ellis, 2007). Cash transfers are therefore a potentially significant pathway to achieve women's empowerment as benefits accrue from both the use of the cash itself and the opportunities for choice and control they present for women in gender unequal contexts (Adato et al., 2000; Case & Menendez, 2007; Kabeer et al., 2010).

Women's empowerment and increased access to resources have been shown to have a positive effect on children's well-being. For example, women use cash income that comes directly to them very differently from the way men do. Studies show that women's spending patterns benefit children and the family more than men's spending patterns, which tend to be focused on discretionary personal spending (Kabeer, 1994; Haddad, 1999). A local study also contends that women are ‘more likely to use resources they control to promote the needs of children’ (UNICEF, 2006:16). Other research has identified a positive correlation between higher literacy levels of mothers and improvements in children's health and education (Henshall Momsen, 2004).

While these studies point to important links between social protection, women's empowerment and children's well-being, these connections have not been substantively explored or empirically verified in South Africa, and their impact on social care in local households also remains largely unexplored (for research on this see Patel & Hochfeld, 2011; Patel et al., 2012).

To investigate the relationship between women's empowerment and children's well-being in households that receive a CSG, we used Kabeer's empowerment framework (1999) and indicators of empowerment developed by Adato et al. (2000). We drew indicators from the areas of decision-making and care responsibilities. Decision-making is a key intra-household process that is a proxy for bargaining power in relationships (Folbre, 1994; Kabeer, 1999). Specifically, decision-making about household expenditure (who spends what money on which household items) and about children (who makes financial, health, education and care decisions) are key indicators of women's status in the home (Adato et al., 2000). It would seem that women with partners or women living in homes where adult men reside are more gender empowered if they are the sole or most influential decision-maker regarding household and child related expenditure. Kabeer (1999) traces a direct link between cash resources controlled by women and their sense of agency (own goal achievement). These resources and the agency to use the money as they choose provide women with the freedom and opportunity to generate valuable outcomes that may enhance their capabilities (Sen, 1999). Other indicators of empowerment relevant to the study included women's self-belief and perceptions (individual empowerment), women's rights in the household, economic security and participation in the public domain (Adato et al., 2000).

The lack of public and private support for women's care responsibilities limits their access to employment and other opportunities that could improve life for them and their families. Care work is the unpaid labour in the home that is heavily gendered as women's work. Care involves the physical and emotional care of household members, both children and adults. Razavi (2007) argues that the care economy relies on the unpaid work of women, which remains largely invisible and unaccounted for in the calculation of gross national product. In particular, the goods and services that are crucial to satisfying human needs in the private sphere are not accounted for in economic analyses (Kabeer, 1994). Care work places a huge burden on women and this is exacerbated in South Africa by the need to care for sick household members. A household's care economy is important to the understanding of citizenship and women's disadvantage in the household in relation to income, time and freedom of movement, which is constrained if a woman is solely responsible for the supervision of children or dependent adults (Hassim, 2006; Folbre, 2008). In this study, in order to estimate the relative empowerment of women within households, we investigated who did household care tasks and how much of their time was spent on these.

Finally, child well-being was conceptualised in terms of impacts on children's health, education and access to care in line with other research of this nature locally (Agüero et al., 2006; Coetzee, 2011) and internationally (Adato et al., 2000).

4. Method

4.1 Population and research site

This study was conducted in 2010 in the urban area of Doornkop, Johannesburg. The target population was households with children aged 15 years and younger (the age of eligibility for the CSG at the time of the field work in 2010). The area of Doornkop was chosen because in a previous project, the Johannesburg Poverty and Livelihoods Study (De Wet et al., 2008), Doornkop was the area receiving the highest number of CSGs across eight of the poorest wards (municipal areas) in the City of Johannesburg. Ward 50, Doornkop, is the poorest ward of Region C and the tenth most deprived ward in Johannesburg. In this study, Doornkop had the largest number of CSG beneficiaries and the third-lowest score on the urban insecurity index measured in terms of multiple indicators of poverty, namely employment, health, food security, education, housing, access to services, overcrowding and social support (De Wet et al., 2008). Female-headed households are worse off than male-headed households across this deprived area of the city.

Despite its low urban insecurity score, the Doornkop community has access to a number of services such as free and accessible schools, primary health care clinics and free health services for pregnant women and children under six years, piped water and electrified homes (although only minimal usage is free, after which it is charged for), refuse removal and tarred streets and street lights. However, high levels of unemployment keep households very poor.

4.2 Sampling

We surveyed 343 households across Doornkop, which represents 10% of households with children in this ward. Households were sampled systematically, using a recent administrative map of Ward 50 to select municipal stands evenly across the ward. For each stand selected, we then selected one primary female caregiver of children 15 years or younger from the households on that stand. If more than one household had such a caregiver present, then the household to be interviewed was randomly selected on site using a pre-determined formula. The final sample included both CSG beneficiary and non-beneficiary households.

4.3 Data collection

To collect data we developed a questionnaire consisting of 14 sections of closed-ended questions broadly covering background household information, livelihood activities and income, food security, use of and views of the grant, partner relationships, household decision-making and care responsibilities, and dimensions of women's empowerment. The questionnaire was developed by the lead researcher in consultation with the research team and other experts and tested twice to ensure its appropriateness. The field work was conducted in July 2010 by 81 fourth-year social work students from the department of social work at the University of Johannesburg. This was integrated into their learning programme, and they were thoroughly trained to collect these data.

4.4 Limitations

Any research that includes investigating how income is used faces the problem of the fungibility of money. This means we can never be entirely certain what the CSG is used on as beneficiaries do generally have other sources of income. Since household income is pooled, separating out what the different income sources are spent on is often impossible.

The design of this study generated largely descriptive data and therefore does not permit conclusive direct causal links. However, the trends that were revealed suggest important relationships between the variables.

A further limitation is that the findings cannot be generalised to all CSG households nationally. Also, conducting the field work on weekdays excluded those with regular employment away from home, and as the questionnaire was lengthy some questions were unanswered, which left some gaps in the data. The results may, however, be generalised to other urban areas with similar social and demographic profiles. Finally, as this was a quantitative study, nuances and complexities that are best captured using a qualitative research design may have been missed.

5. Findings

5.1 Demographics of women CSG recipients

Of the 343 households surveyed, 281 were receiving one or more CSGs, covering a total of 421 children. Thus, 82% of the surveyed households with children 15 years and younger were receiving a CSG, which is a very high uptake of the grant. On average, a caregiver received 1.5 CSGs per household. Nearly half of the respondents were looking after very young children: 44% were under five years, which points to a large care burden. The demographics of the CSG recipients and their households were as follows.

Marital status: Just over a third (37%) were married or in a partner relationship and the rest (63%) had never been married, or were divorced or widowed.

Headship of households: Just over half (52%) were female headed, 38% were male headed and 10% had joint headship (self-identified).

Age of CSG recipients: Caregivers were fairly young, more than half falling between the ages of 21 and 40. The age ranges were as follows: 31% were aged 21 to 30, 26% were aged 31 to 40, 23% were aged 41 to 50 and 11% were aged 51 to 60. Caregivers who were particularly vulnerable because they were either very young or very old were a small minority: 5% were between 16 and 20 and 4% were 61 and above.

Level of education: CSG caregivers were on the whole fairly well educated. More than half (56%) had secondary school education, and 26% had a school leaving certificate. At the high end, 3% had a post-matric qualification, while at the low end 13% had only primary schooling and 2% had no schooling at all.

Source of income: Completing school does not guarantee income security: 71% of the households receiving the CSG earned R2500 or less per month, compared to only 10% of non-CSG households (p = 0.000). For 14% of the households the CSG was the only income. For the remaining 86% for whom the CSG supplemented household income, income was from multiple, diverse sources such as other grants, formal and informal employment and some support from outside their families, such as from their church or a non-governmental organisation. Regular income was most commonly from social grants, in the main the CSG and less so from other grants such as the Old Age Pension.

Figure 1 shows that only 13% of the respondents had regular wages, while 24% had occasional or irregular wages and a further 24% ran their own small businesses. More than 80% of respondents never received financial donations from outside their family. This shows that their income sources were varied and insecure with the exception of state cash transfers, which constitute regular income. The fact that 82% of all children in Doornkop were receiving a grant means the CSG contributes in an important way to reducing child poverty here, despite the small value of the grant (Patel, 2012). For all households receiving a CSG, on average the grant made up 31% (standard deviation 38) of the household's income. This means that when the income from the CSG was pooled with other household income it would almost certainly benefit all the children in the household, irrespective of whether they receive a grant or not (Case & Menendez, 2007; Delany et al., 2008).

Figure 1: Livelihood activities of CSG–receiving respondents

5.2 Women's power in the household

Questions about how decisions are made in households revealed that the women receiving the grant were the primary decision-makers. First, as Figure 2 shows, women made the financial decisions in 81% of the households: 53% were the exclusive financial decision-makers in the household and 28% shared financial decision-making with partners or other adults. Women were consistently the main decision-makers when it came to specific areas of spending, for example, food and groceries (77%), health (77%), education (81%), children's clothes (84%) and transport (74%).

Figure 2: Financial decision-making of CSG–receiving respondents

Second, the grant was being used largely to contribute to household or child needs, corroborating findings from other studies on how women use cash income in households (Kabeer, 1994; Haddad, 1999) and what CSG money is being used for (Delany et al., 2008; DSD, SASSA & UNICEF, 2012). The grant was mainly used to buy food (52% always used it for food and a further 23% often or sometimes did). It was also used to pay for school fees and uniforms (38% always used it for school costs and a further 27% often or sometimes did).

Third, decision-making related to children was even more strongly the responsibility of the woman, with 84% saying they were the sole decision-makers on children's health issues, 74% on discipline and 80% on education.

Women beneficiaries' views of the impact of the CSG on their lives were largely positive. Eighty-two per cent of recipients said that the grant had made their lives better. Sixty-five per cent said that they would not survive if it were stopped, which indicates how important the grant is for economic security for women and particularly for single women. A further 66% of respondents who were receiving the CSG believed that the grant gave them personal power and courage and 61% said it made them feel good about themselves. These are all significant indicators of personal empowerment and self-belief (Adato et al., 2000).

In summary, the women who were receiving one or more CSGs controlled the way the money was spent, were the key decision-makers in their homes, used the grant money largely for food and education, and believed the grant had benefited them and their households.

5.3 Gender and care

Eighty-six per cent of all the respondents interviewed said they spent most of their time at home on care and domestic responsibilities such as housework, shopping, care of children and care of other members of the family. Close to half of all respondents agreed that there are certain jobs that remain ‘women's work’, which indicates an acceptance of gendered roles. In addition, the CSG recipients were engaged in social care activities with their children either daily or often, such as help with homework (64%), accompanying children to school (35%) and playing with or reading to them (58%). Interestingly, CSG recipients were more likely to engage in these activities either daily or often than respondents who were not CSG recipients. The differences between CSG and non-CSG households that were statistically significant were in relation to watching television with their children (p = 0.041), providing help with school work (p = 0.009), and playing with or reading to the children (p = 0.003). We are therefore able to conclude that the caregivers who were receiving the CSG were more actively engaged in care activities with their children than non-CSG respondents, which is a positive indicator of parental involvement in promoting child well-being. There are emotional and social benefits to this, but also more clearly measurable advantages; for example, studies show that children whose caregivers are positively engaged in their education perform better at school (Desforges, 2003).

In addition, women respondents were also responsible for the care of other children who were not receiving the CSG. A total of 17% of respondents in CSG households cared for one or more non-CSG children. Seventy-four per cent of the respondents said they had not applied for grants for the other children in their care, for varying reasons, such as not having the correct documentation (46%), or not knowing how to apply for the grant (8%). The women in our sample were thus faced with additional care demands, with more dependants in households possibly in part because of the HIV/AIDS epidemic and in part because of economic and social stressors.

It is also evident from the data that the women were bearing the greatest responsibility for coping with poverty with limited resources and meeting the care needs of their families and households. Financial burdens are increased when fathers do not contribute to the maintenance of children. Sixty-one per cent of the total number of fathers (n = 102) who were not the current partners of all the women interviewed never paid private maintenance for their children. In addition, 30% of all the respondents receiving a CSG said that the fathers no longer provided support for their children now that they were receiving a CSG. This raises the question as to whether the CSG may be displacing private maintenance paid by the fathers. While no evidence prior to this exists of the CSG displacing remittances (Neves et al., 2009), remittance displacement is an unintended consequence of cash transfers in a range of other countries (Cox & Jimenez, 1995).

We were also interested in how access to basic services mediated women's domestic and care responsibilities. This picture was largely positive, with 95% of CSG households having access to electricity, 98% having running water in the house or yard and 76% having a flush toilet. About a third (34%) of the CSG recipients said that one or more of their children attended a crèche or was cared for by a child-minder.

Thus care activities were overwhelmingly the responsibility of women in these CSG households, leading to time poverty, a large domestic work burden and limited mobility (Folbre, 1994). This was somewhat eased by municipal services that alleviated the burden of domestic chores. For example, water piped directly to a household's stand relieves women from having to fetch and carry water for their homes.

5.4 Child well-being

We now turn to the question of child well-being. On the basis of CSG recipient perceptions, we considered four outcome indicators: food security, health status and immunisation, school attendance and pass rates, and family cohesion. These outcomes were not verified against administrative records or height-for-age measures of children to confirm their validity. However, the findings do provide an indication of how the caregivers perceived the well-being of the children in their care.

We used a validated Household Food Insecurity Access Scale (Coates et al., 2006) to measure household food security of respondent households. More than half (54%) of all CSG households said their households experienced severe food insecurity and a further 25% were moderately food insecure. This is despite the fact that 52% of women always spent the CSG on food. Therefore, without the CSG food insecurity would be more pronounced. The positive nutritional effects of the CSG were confirmed by Agüero et al. (2006), who found gains in child height-for-age. With respect to child health, 97% of CSG recipients said their children had been immunised, and the majority (92%) considered their children to be in good health.

All CSG children (100%) of school-going age were reported to be attending school regularly. Seventy-four per cent of school-going CSG children had not failed a grade. This was also confirmed by another study that concluded that children who received a grant were significantly more likely to be enrolled in school in the years following award of the grant (Case et al., 2005).

South Africa's history of migrant labour damaged family cohesion, resulting in large numbers of children being cared for apart from their biological parents (Lund, 2008; Hall & Wright, 2011). Acknowledging this, it was decided to make the grant payable to the ‘primary caregiver’ of the child, possibly in the extended family, but also by people outside the family. The policymakers also took account of the fact that the HIV/AIDS epidemic might worsen the situation. We were therefore interested in who cares for children and in their care arrangements. We found that all the CSG children in the sample were cared for within the family, living in households with relatives (made up of two or three generations), and 87% lived with either one or both biological parents, as Figure 3 shows. However, the dominance of women as primary caregivers and their consequent care burden is again revealed when Figure 3 is examined, as it shows that 54% of the children in the sample lived with their mother only, while a small percentage (1%) of children lived with their biological father only. Sixty-two per cent of the CSG recipients said that the grant ‘helped to keep their family together’.

Figure 3: Living arrangements of children in relation to biological parents

In addition, 87% of all grants received by the caregivers were for children who actually lived in the household with the caregiver. This is contrary to the popular belief that CSG beneficiaries receive the grant but that the children do not actually live with them. Only 9% of the total number of children who received a CSG lived away from the household, and of this 9% the majority (70%) were between one and 10 years of age. The main reasons cited for children living away from home were a lack of accommodation (21%), school attendance (16%) and because the CSG recipient needed someone to care for the children while she worked (8%). Half the children (50%) who lived away from home were in a rural area, and in these cases their main caregiver was the grandmother (49%). The children who lived away from home were thus with close relatives. Of the children living away from home, 61% had been away for one to four years, 26% for five to eight years, 11% for nine to 10 years and 3% for more than 10 years.

In summary, the children in the sample who were in receipt of a CSG were generally healthy, lived with family, mostly with one or both biological parents, and had good school attendance and reasonable performance. Nevertheless, they lived in food-insecure households and we can therefore conclude that the grant is probably used largely to pay for food.

6. Conclusions

While this research cannot confirm direct causality between the key variables, the data are suggestive of trends and relationships between variables which could be explored in further research. Two important conclusions may be drawn from the findings. First, the evidence appears to support our contention that the CSG enhances women's power and control over household decision-making in financial matters and general household spending and promotes child well-being. The fact that the women in the sample could make decisions and exercise freedom of choice about how to spend the grant suggests that the grant enabled them to generate outcomes they considered important for the quality of their lives and their children's well-being (Sen, 1999). We conclude that the CSG may give female grant beneficiaries a sense of empowerment and that it therefore has some positive transformative effects. Indeed, Adato et al. (2000:51–2) also found that Mexico's cash transfer programme for children, which is paid to women, can contribute to empowerment by ‘putting resources in women's hands’ and thus ‘directing the benefits toward expenditures that normally fall within the decision-making domain of women’. In contrast, we have limited knowledge of how the small numbers of men who receive the CSG spend the money, what the gender dynamics are in these households, and how the grant affects both gender relations and child well-being. More research is needed in these areas.

Second, the data clearly show that South African women bear greater responsibility than men for the care of children, given the entrenched gender-based patterns of care and past historical factors, such as the impact of the migrant labour system on family life. This unequal distribution of care burdens between men and women in the private domain (family and household) increases gender inequality and curtails the ability of CSG beneficiaries to engage in activities (such as earning income) outside the home. The enormous domestic and care responsibilities of CSG beneficiaries remain largely unacknowledged and invisible to policymakers. Public discourse about how grants create dependency seldom takes into account the gendered nature of care and how it affects women's quality of life. Helping to lessen the care burden on women should be viewed positively and not because we merely assume that ‘things are good for women because they are good for children’ (for a critique of this approach, see Budlender in Patel, 2009:52). Molyneux (2006) correctly cautions against positioning motherhood as the key to the successful outcome of poverty reduction programmes and viewing women merely as conduits to reach children in order to promote child well-being. Instead the empowerment of women should be supported as an end goal in itself (Kabeer et al., 2010). Thus, improving the status and position of women and overcoming unequal gender relations are important policy outcomes that need to be incorporated into social protection programmes (Jones & Holmes, 2010), even those that focus primarily on children. However, social protection programmes by themselves cannot achieve all these outcomes: they need to work in concert with other public policies and social programmes to promote both social transformation and child well-being.

Acknowledgements

The authors wish to thank Reem Mutwali for assistance with the statistical analysis for the study and the reference group, Nhlanhla Jordan and Professors Francie Lund, Shireen Hassim, Vivienne Bozalek and Trudie Knijn for their input at both the conceptualisation and analysis phase of this research. This research was funded by the South Africa Netherlands research Programme on Alternatives to Development (SANPAD), and the University Research Committee at the University of Johannesburg.