Tackling child poverty in South Africa: Implications of ubuntu for the system of social grants

Abstract

In South Africa both liberal and more communitarian and relational discourses of citizenship can be seen – the latter in the form of the southern African idea of ubuntu. Policy for assisting children, however, is dominated by the framework of liberal citizenship, most clearly through the Bill of Rights and in particular the Child Support Grant. Using analyses from a purpose-built microsimulation model we show how a neglect of children's broader relationships in the current liberal citizenship inspired policy context limits the effectiveness of the child poverty strategy. The empirical analyses demonstrate how a greater recognition by policymakers of the relational principles of ubuntu could be expected to have more effect on reducing child poverty.

Keywords:

• child poverty
• Child Support Grant
• citizenship
• microsimulation
• social grants
• ubuntu

1. Introduction

Although widely used, the term ‘citizenship’ is contested across several different strands of thinking (Jones & Gaventa, 2002), each with its own associated implications for policy approaches. This is particularly notable in South Africa with its combination of liberal (as found in the Bill of Rights) and more communitarian discourses of citizenship (in the form of the southern African idea of ubuntu). This paper uses original microsimulation modelling to question how effectively South Africa's social security package combats child poverty, given that it is based squarely on a foundation of liberal citizenship rights at the expense, in terms of policy, of ubuntu.

Within liberal approaches citizens and the state can be thought of as bound together in a ‘contract’ in which the individual is granted rights in return for fulfilling certain obligations towards the state and fellow citizens (e.g. obeying the law, participating in paid work), with the terms of that contract being keenly contested (White, 2003). The analytical starting point for discussions of liberal citizenship is often thought to be the work of Marshall (1964), yet South African discussions of liberal citizenship precede this work and continue to be central through the Bill of Rights and – in terms of social rights – the social grants. This liberal view of the detached rights-bearing individual can be counterbalanced by more communitarian citizenship thinking which sees the individual as inherently embedded in social relationships and networks which are key to facilitating identity, resource needs and belonging The concept of ubuntu, outlined below, is a specifically southern African communitarian philosophy and can be understood as a second historically rich strand of citizenship thinking in South Africa, emphasising that individuals are defined and understood primarily through their relationships with others rather than by their status as discrete individuals. These two distinct approaches to citizenship lend themselves to different policy approaches. In current policy for tackling child poverty in South Africa, liberal notions of citizenship dominate (through the Bill of Rights and social grants), yet we would argue that a greater consideration of the principles, practices and policy implications of ubuntu could lead to more effective strategies for reducing child poverty.

In this paper we first describe child poverty in South Africa and the current system of social assistance. We then discuss liberal citizenship as it relates to South Africa, and then ubuntu and its implications for the current policy regime. We present original empirical analyses from microsimulation modelling of South African household data in order to provide new empirical evidence to show the relevance of the principles and practices of ubuntu to policies for reducing child poverty. In the final section of the paper we draw together the implications for South African policies for tackling child poverty.

2. Child poverty and social assistance in South Africa

South Africa is classified as a middle-income country (UN, 2010) yet poverty rates, and especially child poverty rates, are exceptionally high. Proudlock et al. (2008) use the 2006 General Household Survey (GHS) to estimate that about 68% of children live in households with an income of less than R1200 per month. Using the 2001 Census data, Barnes et al. (2009) find that 81% of children experience income and material deprivation while Streak et al. (2009), using the 2005/06 Income and Expenditure Survey (IES), find that about two thirds of children are poor on each of three separate measures – income, consumption and expenditure.

Behind these aggregate figures lie important differences between groups. First, there remains an extreme racial disparity in the distribution of poverty and wealth (Gelb, 2003), a disparity that consistently widened throughout the early 20th century (Leibbrandt et al., 2010). It is estimated that 63% of black children live in ultra-poor3 households while this is the case for only 4% of white children; by contrast, only 1% of black children live in the most affluent households with earnings of more than R16 000 per month, compared to 29% of white children (Monson et al., 2006). In terms of geography, Leibbrandt et al. (2010) observe that poverty rates in rural areas (77%) are about twice as high as those in urban areas (39%) but that the problem of urban poverty accounts for a growing share of all poverty because of ongoing processes of urbanisation and migration. Finally, the high rate of HIV/AIDS in particular has led to a rapid increase in the number of child-headed households since the mid-1990s and although this group comprises less than 1% of all children they are highly vulnerable to acute poverty (Richter & Desmond, 2008).

For most of the apartheid era little support was provided by the government to those living in poverty and the social assistance system was highly racialised and focused on the elderly and disabled. The State Maintenance Grant (SMG) was the main source of social assistance for children but in the early 1990s less than 1% of SMG recipients were black despite the group representing more than 80% of the population. After apartheid ended the new Government of National Unity established the Lund Committee in 1995 to explore income support provision for children. The deliberations and recommendations of the Lund Committee were inevitably shaped both by political negotiation and by the economic context of the time, particularly constraints on expenditure resulting from the shift to the Growth, Employment and Redistribution (GEAR) strategy, which placed greater emphasis on fiscal constraint and a growth-driven approach to poverty alleviation than had previously been the case (Lund, 2008). It was made clear to the committee that universal expansion of the SMG to the whole population was (along with other possible reform ideas) ruled out as prohibitively expensive (Lund, 2008:18). In the end the Lund Committee proposed that a grant of R100 per month be provided to all children under the age of seven whose caregiver satisfied the means test.

Despite criticisms of the CSG, in particular because of its low value and the application of a means test (Martin & Rosa, 2002), its implementation in 1998 was an important step in embedding a social assistance transfer aimed at reducing child poverty in modern South Africa. The CSG differs from the previous SMG in many respects, one notable aspect being the shift to a child-based rather than household-based grant so that the funds can (in theory) follow the child even if the child moves to another household or as the parental situation changes (e.g. due to death of a parent as a result of HIV/AIDS) (Lund, 2008:52–3). Between 1998 and 2009 the nominal value of the grant increased to R260 per child per month (from April 2011) and the CSG has gradually been made available to older children so that those up to 18 will become eligible from January 2012. The income test was raised considerably in 2008 so that about 60% of all children in South Africa are now eligible and uptake has increased dramatically, standing at 86% in 2007 (Children Count, 2009).

The support for children through the CSG sits within the context of a broader social security regime. South Africa currently has a range of social grants besides the CSG, with the main ones in terms of claimant numbers being the Disability Grant (DG) for working age adults unable to work due to ill-health and the Old Age Grant (OAG). Both are means tested grants with a value of up to R1140 per month in April 2011. In August 2009 there were about 9.1 million child beneficiaries of the CSG, compared to about 1.3 million beneficiaries of the DG and 2.5 million beneficiaries of the OAG (SASSA, 2009). The CSG is thus received by several million more individuals than the DG or OAG but its value is much smaller.

Despite the introduction of these grants a notable gap in the country's social security regime is that there remains no provision for adults who are fit to work but unable to find work. This is a significant problem in the context of persistently high rates of unemployment – put at 24% in December 2009 (according to the more restrictive official definition4) – and there is inevitably a strong link between child poverty and parental or caregiver unemployment. Nationally 41% of children live in households where there is no adult in work, and in the province of Limpopo, which has the highest child poverty rate at 83%,5 only 28% of children live in a household where there is an employed adult (Children Count, 2008).

3. Two approaches to dealing with child poverty in South Africa

3.1 Liberal citizenship

The work of Marshall (1964) is often taken as a starting point for discussions of liberal citizenship. Marshall's account sees citizenship rights and obligations as a means to generate a sphere of equality between all citizens in order to compensate for the inequalities in outcomes inherent in capitalism (Barbalet, 1988). Marshall saw citizenship as bringing with it a triad of civil, political and social rights6 that developed during the 18th, 19th and 20th centuries. Despite being criticised for its historically Anglocentric focus (Turner, 1993), Marshall's framework remains analytically useful in thinking about citizenship in South Africa and has been applied explicitly to the South African context (Whitworth & Noble, 2008).

Indeed, liberal citizenship thinking in South Africa precedes the work of Marshall and has a long and distinguished history throughout the struggle movement of the 20th century. The idea of a package of rights for all South African citizens has its origins in the historic 1943 Africans' Claims in which the African National Congress (ANC) set out explicit claims for civil, political and social rights for all South African citizens irrespective of race or class. The response of the Smuts government to Africans' Claims was dismissive and the beginning of apartheid created a policy environment openly hostile to such debates for equality of citizenship, yet the 1954 Women's Charter and the ANC's 1955 Freedom Charter again set out such demands for universal civil, political and social rights. In many ways this historical legacy of liberal citizenship sits at the heart of current social policy reform and debate in modern South Africa. After the transition to democracy in 1994 a liberal approach to citizenship thinking became central to the vision of the post-apartheid nation through the work of the Taylor Committee and especially through the creation of the 1996 Bill of Rights in the South African Constitution, a document that is crucial in articulating a vision for the sort of society that South Africa aims to become (Motala, 2009:1), as well as giving justiciable ‘teeth’ to that vision in terms of citizenship rights.

After much political debate over the nature of social rights, Section 27 of the Bill of Rights was phrased thus: ‘Everyone has the right to have access to … social security, including, if they are unable to support themselves and their dependants, appropriate social assistance’. However, the statement contains the final caveat that ‘The state must take reasonable legislative and other measures, within its available resources, to achieve the progressive realisation of these rights’ (RSA, 1996:Section 27, 1 and 2; authors' emphasis). The Bill of Rights also contains a separate section relating specifically to the rights of children. Section 28 states that ‘Every child has the right to basic nutrition, shelter, basic health care services and social services’ (RSA, 1996:Section 28, 1.c) and here the clause relating to ‘available resources’ and ‘progressive realisation’ is, interestingly, not included. This has been interpreted by some to imply that the realisation of children's rights is not subject to limitations on available resources and therefore that children have a higher and undisputable claim on state funds (Liebenberg, 2001; Sloth-Nielsen, 2001). Others argue, however, that parents rather than the state have the primary duty and that the priority of children's rights is elevated only in instances where the parent is clearly unable to fulfil this duty (Creamer, 2002; Liebenberg, 2002), most commonly because of parental death or where the child is removed from the household for protective reasons.

The rights that the Constitution sets out specifically for children provide considerable support to policies that target children. These rights have been shown to be justiciable in court, notably in the Grootboom, TAC, Khosa and S vs M cases. Through these legal actions real progress has been made in defending children's rights and these should not be underestimated. However, the liberal approach also has drawbacks in that rights are separated and delivered as discrete packages to different citizens. Crucially, the separation of children's and adults' rights in the Bill of Rights – and the social grants which flow from them for children, but not for unemployed adult parents or caregivers – illustrates the Constitution's individualistic and somewhat atomised view of citizens. This is at odds with the more communitarian view expressed through ubuntu – and with the reality of children's lives – that relational bonds between citizens are also crucial to tackling child poverty.

3.2 Ubuntu

With links to the communitarian tradition of citizenship thinking, ubuntu is a southern African philosophy that understands the individual as embedded in and defined by broader relational ties of family, community and society (Bhengu, 1996; Coetzee & Roux, 1998; Coertze, 2001). At the centre of the concept is the Zulu saying umuntu ngumuntu ngabantu – ‘a person is a person through people’ (Shutte, 2001:23). Ubuntu sees the community as providing the relational context and support through which individuals develop and live, and it emphasises those values which forge bonds and build networks: sharing, compassion, understanding, reciprocity, kindness, solidarity and sensitivity. This philosophy is distinct from liberal conceptions of citizenship in that it is based on the emotional and relational bonds within which citizens exist and on which all citizens depend in order to fulfil their own potential rather than on the notion of a detached rights-bearing individual (Held, 1990). As citizens are conceived of as being related to and depending on each other, all are responsible for ensuring that others have everything they need.

The differences between these two notions of citizenship have practical policy implications. Of particular relevance to the effectiveness of social security policies to tackle child poverty is the way ubuntu recognises the inter-connectedness of children and adults. This is in contrast to current liberal-based policy which lacks a broader conceptualisation of children's relationships and inter-connectedness with a wider household, network and community which, in turn, have their own needs and resources. These differences suggest a need for greater recognition of the fact that children are relational beings and that households, communities and other networks may themselves act to pool resources and risks, in contrast to (and perhaps as a result of) the individualism assumed by current social security policy. It also suggests a need for more explicit recognition of the links between children's and adults' rights when seeking to tackle child poverty in South Africa. One strand of the literature on child-centred notions of citizenship has argued that children's rights are enhanced by increased autonomy (Therborn, 1993), yet the interdependence of children and adults cannot be ignored and they cannot be so easily separated that the welfare of one group can be improved independently of the other.

In many instances the principles of ubuntu have been used by the courts – either implicitly or explicitly – to interpret the principles set out in the Constitution (Keep & Midgley, 2007), yet the fragmentation of relational networks into discrete individuals is evident in the fulfilling of citizens' rights to social assistance through social grants. Thus, it could be argued that the liberal citizenship philosophy that underpins the Bill of Rights simultaneously advances and constrains the country's ability to tackle child poverty effectively. The Bill of Rights provides a valuable framework through which individuals can receive and fight for key social security transfers, but it also sees the world through the lens of liberal citizenship, which encourages the artificial separation of social security policies for children and for their network of adult caregivers.

4. Assessing welfare outcomes using microsimulation

Using original evidence from a purpose-built South African microsimulation model,7 the remainder of the paper compares evidence for the efficacy of the liberal and the ubuntu philosophies for tackling child poverty. Microsimulation is widely used as a method of analysing the impact of changes to tax and benefit policies on different groups in society (see for instance Mitton et al., 2000, for a review) and has been applied specifically to studying the impact of policy on child poverty in a number of countries (Brewer et al., 2006; Figari et al., 2009). The South African Microsimulation Model (SAMOD)8 is a static microsimulation model which tailors the EUROMOD model to the tax-benefit system and demographic context of South Africa. The model draws on the IES (2000 and 2005), the Labour Force Survey (2000, 2006 and 2007) and the Community Survey (2007) to simulate the South African population in 2007 and builds in the eligibility criteria for South Africa's key social assistance and tax policies. The model then allows detailed micro-level analyses of the impact of policy changes (both real and potential) on children and other groups.

4.1 Household composition, income pooling and child poverty

Given the differences in the availability and value of social transfers across individuals, household composition becomes a central consideration for any anti-poverty programme. Household structures for poor and non-poor households in the IES 2000 are shown in Table 1. There is no official poverty line in South Africa but a threshold of R462 per month (in 2007 prices) has been recommended by Statistics South Africa and the National Treasury (Stats SA, 2007) and is used here. The table shows that only about 4% of children in poor households do not live with a working age adult. Clearly the fact that few children live in households where there is no one of working age means that in theory most children should have access to wage income. However, in the context of high unemployment, under-employment and (for those with work) low wages, many children cannot gain much, if any, benefit from wage income generated by their working age caregivers.

Table 1: Household composition for poor and non-poor households

	Percentage living with
	Children	Working age	Oldage	Three generations	Total %	Number of persons
Poor households
Children	1.1	64.3	2.7	31.9	100	10 891 863
Working age	56.3	11.8	4.4	27.5	100	10 862 809
Old age	8.2	14.6	12.2	65.0	100	2 010 070
Non-poor households
Children	0.2	84.9	0.3	14.6	100	7 302 943
Working age	53.0	33.6	4.2	9.2	100	14 202 371
Old age	1.2	24.7	40.2	34.0	100	1 521 622

Table 1 also shows how income might itself influence household formation decisions in the policy context where social assistance is provided only to certain individuals. Previous research has found evidence to suggest that for the poor, household formation may, at least in part, be an active response to economic need in the context of high unemployment and a partial system of social grants (Klasen & Woolard, 2008). From this perspective, children and particularly old people become valuable income sources for the wider household and family. This is supported by the table, which shows that in poor households working age adults generally live with children, old people or both of these groups (88% of working age adults live with another age group). By contrast, in non-poor households only 66% of working age adults live in a household also containing another age group. While not conclusive, these data support the notion that those living below the poverty line are more likely to form household configurations in which at least one person is likely to be able to access some form of social transfer.

Previous research on income pooling within households also supports the idea that income is shared between household members rather than being ‘ring-fenced’ for the eligible recipient and may therefore encourage particular household configurations (Bertrand et al., 2003; Ardington et al., 2009). For example, while most households state that the CSG is spent directly on the child, about 20% of households say they pool the grant for wider household needs (Guthrie, 2002). Other social grants such as the OAG are also pooled by the household for a variety of reasons – sustaining the household, substituting for missing incomes of unemployed household members, and cushioning the household against unexpected economic shocks – and thus the OAG also frequently benefits children (Sagner & Mtati, 1999; Duflo, 2000). Indeed, evidence suggests a positive link between household OAG receipt and child outcomes in areas such as education, health and nutrition (Duflo, 2000; Aguero et al., 2006). Given that the value of the OAG is approximately four times that of the CSG, its potential to benefit children is clearly greater than the CSG's in the context of household income pooling.

More broadly, a household's equivalised income9 is influenced by many factors besides eligibility for social grants, such as market income (i.e. earned income), remittances and the number of people with whom any income must be shared. This point is illustrated further in Figure 1, which shows the average income derived from three main income sources (market income, social transfers and remittances) for each type of household, assuming full uptake of social grants. One issue common to poverty studies is the paucity of information about intra-household resource allocations, with equal sharing typically being assumed. A second issue in South Africa specifically is the lack of evidence about the nature and cost of adult and child needs, an issue which relates directly to the choice of equivalisation scale selected (although previous research finds that which scale is chosen makes little material difference to child poverty results in South Africa – Streak et al., 2009). The present study follows Cutler & Katz (1992) in calculating the number of adult equivalents in each household on the basis of previous South African research,10 which gives an approximate mid-point in the range of equivalence scales that have been used in South Africa.

Figure 1: Average equivalised household income from different sources for individuals in different household types (R per month)

Figure 1 shows that while households containing working age individuals typically have higher market incomes (as might be expected), the social transfer system creates even more significant differences between different types of household. In particular, only children living in households with older people have a reasonable chance of moving above the poverty line. Yet, as noted above, only about one third of children live in such households, which means that the OAG is not an effective mechanism for addressing child poverty in the aggregate.

The overall impact of each social transfer in turn on different age groups is presented in Figure 2. Assuming that every household pools all its income, the figure shows that the CSG has roughly the same impact on the proportion of children living in poverty as most other social transfers – despite the CSG being by far the largest social grant explicitly directed at children – and as intra-household transfers. For other groups the impact of the OAG on poverty rates among the elderly is particularly striking, a combination of its relatively generous level and the fact that about 40% of non-poor elderly households do not contain any children or working age adults with whom the OAG is shared (see Table 1).

Figure 2: Impact of 2008 system of social transfers on the proportion of individuals living in households with incomes below the poverty line

The above analyses illustrate the importance that income pooling and household structures have for the effectiveness of policies to tackle poverty for different age groups. The individualistic view of the right to social assistance creates some inequality, with a small proportion of children benefiting more than others by virtue of living with elderly household members.

Turning to broader strategies to tackle child poverty, in Figure 3 we evaluate the effectiveness of seven policy reforms that have been considered by the South African government: a basic income grant paid at R180 per month, an extension to the CSG to children up to the age of 18, an extension to the CSG to all children and removal of the means test, a means tested grant of R180 to the low-income working age individual, a means tested grant to the low-income working age individual of between R100 and R360 depending on existing income, an extension to the current OAG to cover men from the age of 60 (rather than 65 as at present), and an extension to the OAG and removal of the means test to make the grant universal (and taxable for those paying tax). Of these reforms only one has been implemented – the extension of the CSG to age 18.

Figure 3: Percentage of each age group living in households with post-tax and transfer equivalised incomes below the poverty line

Comparing each reform with the existing system, we find that the variable-rate low-income grant paid to the working age poor and the basic income grant have the greatest impact on poverty rates for all age groups, including children. Figure 3 suggests therefore that greater recognition of children's broader relationships provides a significantly more effective mechanism to reduce child poverty than either the current social grants regime or a range of alternative policy options. It is of course also necessary to consider the cost as well as the impact of any potential policy reform. Figure 4 shows the overall cost of each of these potential reforms as well as the cost per individual to remove them from poverty. The variable-rate grant to the working age individual is the most expensive scheme but, interestingly, is also the most cost effective in terms of the average cost of removing an individual from poverty.

Figure 4: Total annual cost of reform and average cost per individual removed from poverty

The analyses in Figure 3 and Figure 4 present a simplified view of the costs and benefits of policy reforms. The microsimulation modelling demonstrates that the current individualistic approach to tackling child poverty does not necessarily generate the most efficient or effective policies for reducing child poverty. Two points are of particular note. First, greater recognition in policy of the principles and practices of ubuntu and of children's relationships and inter-connectedness with adults offers an alternative, and arguably more effective, means of addressing child poverty. Second, there are various ways in which a greater recognition of ubuntu can be implemented, with each having positive – though differing – impacts on reducing child poverty. Particularly effective (in terms of reducing child poverty) and efficient (in terms of average cost per person removed from poverty) appear to be a basic income grant, a fixed low income grant and a variable low income grant. Moreover, while the focus here is only on potential social assistance reforms, the broader and more important message is the need to address poverty in the working age population in order to reduce child poverty more effectively, irrespective of the policy instrument used. The policy focus thus need not, and indeed should not, be restricted to social assistance reforms but should also consider employment generation policies to assist the working age poor and, as a consequence, poor children.

Naturally, further factors would need to be considered to assess the true cost of the reform and any possible unintended consequences or incentive effects. One issue is the potentially large administrative costs of any means tested scheme, and in particular of the highly effective, though administratively much more complex, variable low income grant. A second issue is the potential incentive effects or unintended consequences resulting from any such income support policies aimed at poor children or poor households. The South African debate has seen a great deal of attention paid to concerns that the CSG encourages teenage pregnancy, something which a grant targeted at households rather than children might avoid.11 Such concerns seem ill-founded, however, in that research (Makiwane et al., 2006; Steele, 2006; Moultrie & McGrath, 2007) has repeatedly found no evidence of any link between CSG and teenage pregnancy in South Africa. Thus child-based policies remain firmly on the table, though greater recognition of ubuntu highlights the point that extensions to the CSG without income support or employment generation strategies to help other household or broader network members are unlikely to alter overall child poverty levels dramatically.

5. Conclusions

In this paper we have argued that in South Africa the dominance – yet demographically partial implementation – of the liberal citizenship perspective and the relative neglect of the communitarian principles and practices of ubuntu has serious implications for the effectiveness of current government policy to tackle child poverty. While the enshrinement of rights in the Bill of Rights has significant advantages in South Africa, our analyses suggest that the system of social grants needs to take a broader, more collective and more grounded approach than is currently the case in the liberal model of citizenship and as set down in the Bill of Rights and operationalised in the current package of social grants. In particular, in seeking to tackle child poverty through the CSG the current policy formulation seems limited in that children are granted individualised rights as autonomous beings abstracted from familial and social relations and from patterns of household behaviour in terms of household formation and resource pooling. This ignores the fact that child poverty is situated in the broader context of the needs and structures of families, households and relationships. The social assistance programmes to tackle child poverty need to recognise this.

Using original microsimulation modelling, we have highlighted ways in which grants not directed at children nevertheless have important implications for tackling child poverty. Extensions to the age eligibility within the CSG are welcome, as would be increases to its real value. Despite this progress, however, the CSG continues in reality to be used to support broader household needs. Policies for reducing child poverty would be more effective if the principles of ubuntu and its application in practice were better recognised in the package of policy supports available, whether via social assistance programmes or employment generation policies or both. The fact that most children live in households that remain poor even after receiving social assistance shows that child poverty and working age poverty cannot be separated as easily as the current liberal citizenship inspired policy framework would suggest. Rather, there is a need for the principles and practices of ubuntu to be better recognised in the government's anti-poverty strategy. The government has various options available for implementing ubuntu inspired policy reforms – each with its own measurable level of cost, effectiveness and efficiency. Reform in this direction seems essential if child poverty is to be reduced in the new South Africa.

Acknowledgements

The authors are grateful to Professor Holly Sutherland and colleagues at the University of Essex for sharing the EUROMOD model.

Notes

³Ultra-poor households are households earning less than R800 per month.

⁴The South African government publishes both narrow (or official) and expanded unemployment statistics. The official figure includes those economically active individuals who were out of work in the past week, would like to work and could start within the next two weeks, and who have taken active steps to find work in the past month. The expanded version drops the active work seeking criterion.

⁵Calculated using data from the GHS 2005 based on a household poverty line of R1200 per month.

⁶In citizenship theory, examples of civil rights are the right to free speech and the right to hold property, examples of political rights are the right to vote and to form political associations, and examples of social rights are rights to cash transfers or services such as health and education. The distinctions between the three types are of course not always completely clear in practice.

⁷SAMOD was developed by Kate Wilkinson.

⁸Further details of SAMOD can be found in Wilkinson (2009).

⁹Equivalised income is used to create comparable household incomes across households of different sizes.

¹⁰In this scale, adult equivalents per household = (a + (c * b))^d using values c = 0.75 and d = 0.86, which align with the implicit equivalence scales derived by Potgieter (Woolard & Leibbrandt, 1999).

¹¹Assuming that means tests are applied on non-equivalised household incomes as is currently the case.