Social security for young people amidst high poverty and unemployment: Some policy options for South Africa

Abstract

South African youth experience extremely high levels of unemployment and poverty. Currently there is no social assistance for low-income young adults in South Africa unless they are disabled. Interventions are needed that can achieve widespread poverty alleviation, as well as help facilitate economic participation to improve lifelong earnings. In this article, six examples of social security policy options are considered, including five grants ranging from an unconditional non-means-tested grant for young people to a conditional grant for young people in training or education, plus an ‘Opportunities voucher’ that is administered through the social security system but paid out to organisations offering youth education or work opportunities. Using a tax and benefit microsimulation model to simulate the five grants, we estimate the potential numbers reached and cost, as well as the impact of these six options on poverty.

Keywords:

• youth
• social security
• microsimulation
• poverty
• unemployment

JEL codes:

• I3
• I32
• J2
• J68

1. Introduction

South Africa has a large youth population, with those aged 15 to 24 years making up almost 20% of the population (DSD, 2010). International evidence suggests that if most young people are economically active, this youth bulge can offer an opportunity for accelerated growth and poverty reduction as dependency ratios fall (DSD, 2010; Gribble & Bremner, 2012). More than one-half of those aged 18 to 24 are out of school and out of work, and also live below the poverty line (StatsSA, 2010). Much more substantial policy attention is required to promote their economic and social participation.

This article presents some of the findings of research to explore poverty impacts of selected social security options for youth. It is organised as follows: Section 2 provides the context by briefly reviewing the status of South African youth and the current social security arrangements. Section 3 introduces the social security policy options that are considered in this article. Section 4 presents findings in respect of numbers of youth that could be reached through each of these policy options, as well as potential cost and impact on poverty. Section 5 concludes.

2. Context

South Africa has very high levels of unemployment amongst young people. In 2010, 51% of 18 to 24 year olds were unemployed by the strict definition, as compared with a national average of 25.3%; while about two-thirds of all 18 to 24 year olds who were able and available to work were unemployed by the broad definition (StatsSA, 2010).

More than one-half of all 18 to 24 year olds live under a poverty line of R604 per month (2011 Rand) and more than two-thirds live under a poverty line of R1113 per month (see later Tables 2 and 3).

Table 1: Number of eligible young people and cost of provision

Option	Age group	Number of eligible youngpeople (million)	Cost of provision peryear (R million)
Status quo	15 to 17	2.071	6 461.33
	18 to 24	0	0
	15 to 24 (all)	2.071	6 461.33
A	15 to 17	2.071	6 461.33
	18 to 24	6.986	21 796.05
	15 to 24 (all)	9.057	28 257.38
B	15 to 17	2.071	6 461.33
	18 to 24	5.882	18 353.19
	15 to 24 (all)	7.953	24 814.52
C	15 to 17	2.071	6 461.33
	18 to 24	5.882	21 937.75
	15 to 24 (all)	7.953	28 399.08
D	15 to 17	2.071	6 461.33
	18 to 24	0.339	4 643.64
	15 to 24 (all)	2.410	11 104.96
E	15 to 17	2.071	6 461.33
	18 to 24	1.984	6 190.29
	15 to 24 (all)	4.055	12 651.61
F	16 to 18, used up to age 24	2.071	1 582.06

Notes: This table does not include administrative costs. Figures are based on 100% take-up.

Table 2: Poverty rates for status quo and Options A to E by age group using the lower-bound poverty line, 2011

	Poverty headcount (P0)			Poverty depth (P1)			Poverty severity (P2)
Option	15 to 17	18 to 24	15 to 24 (all)	15 to 17	18 to 24	15 to 24 (all)	15 to 17	18 to 24	15 to 24 (all)
Status quo	0.573	0.541	0.551	0.291	0.315	0.308	0.184	0.232	0.218
A	0.554	0.487	0.507	0.260	0.233	0.241	0.153	0.135	0.141
B	0.555	0.489	0.508	0.260	0.233	0.241	0.153	0.136	0.141
C	0.548	0.476	0.497	0.253	0.224	0.233	0.148	0.129	0.135
D	0.566	0.523	0.536	0.281	0.300	0.295	0.176	0.219	0.206
E	0.566	0.524	0.537	0.280	0.288	0.285	0.173	0.198	0.191

Note: Figures are based on 100% take-up.

Table 3: Poverty rates for status quo and Options A to E by age group using the upper-bound poverty line, 2011

	Poverty headcount (P0)			Poverty depth (P1)			Poverty severity (P2)
Options	15 to 17	18 to 24	15 to 24 (all)	15 to 17	18 to 24	15 to 24 (all)	15 to 17	18 to 24	15 to 24 (all)
Status quo	0.744	0.675	0.695	0.467	0.455	0.459	0.334	0.350	0.345
A	0.738	0.661	0.684	0.446	0.397	0.411	0.308	0.273	0.283
B	0.738	0.665	0.687	0.447	0.399	0.413	0.308	0.275	0.285
C	0.738	0.660	0.683	0.442	0.389	0.405	0.302	0.265	0.276
D	0.744	0.671	0.693	0.461	0.442	0.447	0.326	0.336	0.333
E	0.744	0.671	0.693	0.460	0.436	0.443	0.325	0.324	0.324

Note: Figures are based on 100% take-up.

Poverty and unemployment go hand in hand, as demonstrated by the fact that labour absorption rates of the working-age population in the poorest deciles are very low. For example, Leibbrandt et al. (2010) demonstrate that only 10% and 19% of the working-age population in the poorest income decile (decile 1) and the next poorest decile (decile 2) respectively are employed. The unemployment rates (using the narrow definition) for these two income deciles were 69% and 46% respectively, and together these two income deciles possessed a share of just 1.4% of total household income (rather than 20% if household income was spread evenly across the deciles).

Many young people do not grow up in households with adults in paid work. Hall & Wright (2010) report that in 2008 over one-third (36%) of children lived in households with no working adults and 55% lived in a household where no adults were in formal employment. The depth of poverty faced by youth acts as a barrier to accessing opportunities. Letseka et al. (2010) found that financial reasons are one of the main explanations for dropping out of school and university. In 2006, more than one-third (38%) of African youth aged 14 to 17 said they left school as they ‘did not have money for fees’ (Sheppard, 2009). Public employment programmes such as the Expanded Public Works Programme operate at a small scale relative to the unemployment challenge. The programmes to stimulate employment in small business and agriculture are still limited in their impact. The economy is growing too slowly to absorb young people, and they are generally not the first to be considered by employers. Employment projections show that even if average annual growth was to rise to 5%, unemployment amongst 15 to 24 year olds would be 44% and 31% by 2020 and 2030 respectively, in the absence of special interventions (Altman, 2013).

Although achieving high school graduation and some post-school education is an important contributor to labour market success, only 21% of African 18 to 24 year olds had completed Grade 12 or equivalent by 2007, and only 3.3% had completed post-secondary education. A first work experience is central to exiting unemployment, yet many young people struggle to access these opportunities. In 2010, only 8% of unemployed youth who lacked experience found a job, as compared with 25% of unemployed youth who had found some previous experience (National Treasury, 2011).

The ongoing refinement of policies to assist young people with attaining educational qualifications, skills and employment is essential but there is also a role for social security policy. The South African Constitution states that:

Everyone has the right to have access to [ … ] social security, including, if they are unable to support themselves and their dependants, appropriate social assistance.

(Republic of South Africa, 1996: Bill of Rights section 27)

Social security is essentially a component of the social protection floor, which consists of two main elements: universal access to essential services (such as health, education, housing, water, sanitation and other services as nationally defined); and social transfers in cash or in kind, to guarantee income security, food security, adequate nutrition, and access to essential services (ILO & WHO, 2009).

The cost and breadth of the social security programme in South Africa has grown substantially since 2005 when the South African Social Security Agency (SASSA) was established. The evidence points to the social grants. By March 2011, SASSA reached 14.6 million recipients monthly and became the main contributor to poverty reduction in South Africa in the post-apartheid era (Van der Berg et al., 2007). The payment arrangements have become increasingly sophisticated. For instance, SASSA is in the process of recording bio-informatics of beneficiaries and recipients and smart cards are being distributed to facilitate non-cash and remote transactions. This capacity offers substantial opportunity to leverage further support and information flows to enhance access to education and employment opportunities.

In spite of this, young people in South Africa have extremely limited access to social security because the system is currently aimed at vulnerable sections of the population – specifically low-income disabled people (Disability Grant and Grant in Aid), low-income older people (Old Age Grant [OAG]) and children in low-income families (Child Support Grant [CSG], and the Care Dependency Grant) as well as the Foster Child Grant (FCG) (Social Assistance Act No. 13 of 2004). The value of the CSG was R260 per month in 2011/12, reaching over 10 million children (SASSA, 2011). It has been demonstrated that the CSG, although small in value, makes a considerable difference in the lives of people on low incomes (DSD, SASSA & UNICEF, 2012; Leibbrandt et al., 2010).

Currently there is no social assistance aimed at able-bodied working-age young people (with the exception of the FCG, which can be extended to 18 to 21 year olds if the young person is completing secondary education or special education training). The only type of social assistance available to young people who are not fostered and pursuing education is the Disability Grant, if they are disabled. In terms of contributory social insurance, young people in formal employment can access the Unemployment Insurance Fund if they become unemployed. However, this will only cover small numbers of young people as most have not had the opportunity to contribute to the Unemployment Insurance Fund.

There is evidence that many youth are indirectly reached by the OAG, which is aimed at people aged 60 and over and was set at R1140 in 2011. Bhorat (2004) showed that the OAG supported 23% of unemployed youth while the Disability Grant and the CSG supported 6% and 15% of unemployed youth respectively. The likely impact is greater now, as Bhorat's study was prepared in the early days of the expansion of the social grants. These other grants are intended for other individuals in the household and so will fail to provide adequate income maintenance for the youth in that household.

There is also evidence that the current social assistance provision (for those who are too young or disabled or old to work) is used by those of working age in various ways, including assisting with job seeking activities (e.g. Schöer & Leibbrandt, 2006) and labour migration (Posel et al., 2006). This diversion of grants towards the unemployed can create household tensions, however, as it is widely understood that, for example, child grants are intended for the needs of the children (Ntshongwana et al., 2010; Surender et al., 2010).

When considering the extension of social assistance, policy-makers are challenged by concerns about affordability, dependency and perverse incentives (see Spicker, 2006; ILO, 2008; Surender et al., 2010). This needs to be weighed against the evident need to implement programmes to alleviate poverty and enhance opportunities for low-income youth.

There is strong empirical evidence that unemployed and poor people in South Africa have a very favourable disposition towards work and express a strong willingness to do so should the opportunity arise (e.g. Noble et al., 2008). The majority (83%) of poor people interviewed in the 2006 South African Social Attitudes Survey said they would take a very low paid job if they thought it would improve their position in the long run. Over three-quarters indicated that they would be prepared to move to find a job. Labour Force Surveys by StatsSA show that young people search for work for long periods of time. In 2005, one-quarter of unemployed had been searching for a job for one to three years while 35% had been looking for three years or more (Altman, 2010). This speaks to a strong motivation to work. The options below are paid at such low rates as to render negligible any potential disincentive effects, although this would have to be tested empirically.

Although some might hold a concern that society would not endorse the extension of the scope of social assistance provision, this is not necessarily warranted. Using the 2006 South African Social Attitudes Survey data, Surender et al. (2010) found that almost three-quarters of adults agreed with the statement that ‘people who can't get work deserve help in the form of social grants’, and two-thirds of those who defined themselves as not poor agreed with this statement.

3. Six social security policy options

No single policy or mechanism will resolve the challenges discussed above given the heterogeneous character of the youth population and the related reasons for inactivity. Designing social security arrangements for youth is complex. In South Africa, the scale and cost of any intervention would need to be substantial given the extent of poverty and unemployment among this age group. Interventions need to be developed that can achieve widespread poverty alleviation, as well as enable economic participation to improve lifelong earnings. This is indeed a complex question that requires an analysis of the creative intersection between social security and social and economic policies aimed at youth.

Goodin (1988) identifies six possible goals of social policy and highlights that whilst they might all be regarded as desirable, they may compete with each other in practice. His six goals are reducing poverty, promoting social equality, promoting social stability, promoting social integration, promoting economic efficiency, and promoting autonomy. All of these six goals might be regarded as desirable in relation to policies for young people and South Africa as a whole. However, in practice, it is a challenge to design programmes that meet this combination of goals. Moreover, these objectives can be interpreted in very different ways depending on which ideological discourses dominate the debate.

The commitment to an expanded provision of social security has been made in the Constitution, and in the SADC Code on Social Security (SADC, 2007). The National Youth Policy also calls for social security for young people (SADC, 2007:27). The idea of the provision of social security to young people is wholly in line with these key documents, but the type of social security could vary greatly. South Africa and the SADC region have given careful thought to the question of which types of social policies the region would seek to promote, and have defined social policies as: ‘[I]nterventions which are about promoting the well-being of all citizens and which address structural inequalities in wealth, ensure greater equity and equality for all, correct market shortcomings, reduce poverty and promote social inclusion’ (SADC, 2006). The definition of social policies that has been developed by the SADC therefore focuses particularly on Goodin's goals that relate to reducing poverty, and promoting social equality and social integration.

As well as the policy goals and cost of a programme, the reach of a programme also needs to be taken into account. Our examples focus on reaching those who exit the CSG at age 18, but – as will be seen – the options reach very different proportions of young people. Careful consideration needs to be taken as to which groups benefit and miss out from support.

Six social security policy options are considered in this paper:

1. unconditional universal grant for 18 to 24 year olds inclusive, at the CSG rate (R260);

2. unconditional means-tested grant for 18 to 24 year olds inclusive, at the CSG rate (R260);

3. unconditional means-tested grant for 18 to 24 year olds inclusive, at the CSG rate (R260), plus conditional component with top-up to the OAG rate (R1140) if in education/training;

4. conditional means-tested grant for 18 to 24 year olds inclusive, if in education/training, at the OAG rate (R1140);

5. conditional means-tested grant for 18 to 24 year olds inclusive if unemployed or in education/training at the CSG rate (R260); and

6. opportunities voucher – a one-off allocation of R3500 to recipients of the CSG or the FCG, or alternatively R6500 if they complete high school, to be applied to an activity that can raise the chances of a school-to-work transition.

In all of these policy scenarios, the current system of grants was assumed to continue as it was at April 2011, with the exception that the CSG age band was extended from 0 to 14 years to 0 to 17 years, which in practice took place in January 2012 (see SASSA, 2011).

The scenarios focus on 18 to 24 year olds inclusive as the target group (although it is recognised that older age groups might question the restriction of such support to this particular age group). These scenarios are not prescriptive but are simply examples of options that could be considered by the government. The payment amounts for the six options are also only examples and could be adjusted depending on the availability of resources and empirical evidence on the cost of meeting basic needs in the twenty-first century. A further consideration is the level of earnings that young people might achieve if in the labour market; a monthly payment at the level of the CSG would not exceed this, and the higher level of payment to those in education or training is in recognition that this group is not expected to be economically active.

As will be noted, the examples represent a range of ideological approaches to the provision of social security, from universal provision (a social democratic objective) to a means-tested and conditional grant that represents a more neo-liberal approach. Also, whilst some of the examples include conditions for ongoing receipt (Options C to F), some do not (Options A and B). In these instances eligibility is dependent on ongoing evidence of education/training/work-seeking status, in recognition that the challenges faced by young people are primarily of a structural nature (e.g. Lund et al., 2009).

3.1 Option A: Universal grant for 18 to 24 year olds

This option would entail an unconditional non-means-tested grant for all young people aged 18 to 24 inclusive. Given the extent of youth unemployment, the poor quality of education that many have received, the lack of available jobs, and the extent of poverty among young people, this would serve as a direct form of income maintenance to help to redress these social challenges. The grant could serve as an – albeit very small – contribution to meeting the material needs of young people (e.g. contributing to the costs of food, housing, clothing and transport). The rationale for this option is that categorical benefits are common internationally for groups in particular need. This option is also consistent with a social democratic or social citizenship rights-based approach to social security which, it could be argued, is consistent with the Constitutional commitment to ‘the right to have access to social security, including, if they are unable to support themselves and their dependants, appropriate social assistance’ (Republic of South Africa, 1996: section 27.1c). This option also has the advantage that it would be more straightforward to administer and thus would be more likely to reach those most in need (Mkandawire, 2005).

3.2 Option B: Means-tested grant for 18 to 24 year olds

This option comprises an unconditional means-tested grant for all young people aged 18 to 24 inclusive. As it would no longer be appropriate to means-test a ‘primary caregiver and their spouse if they have one’ (as in the case of CSG), the means-test could be applied to the young person and their spouse if they have one. In terms of mixed-age-band partners (i.e. if a young person is married to an older person), the means-test would still apply at the couple level. This grant may have some unintended consequences but this is unquantifiable and may be a negligible issue (e.g. people choosing youthful partners as an income stream, or divorcing older partners and replacing them with youthful partners). Another potential issue is that single young people living with wealthy parents would also qualify. However, if we were to apply a means-test to the parents (if present in the household), this may introduce a perverse incentive whereby people leave home sooner than they would otherwise have done in order to qualify for the grant, and put themselves at greater risk in so doing. We therefore apply the means-test to the young person and their spouse if they have one.

3.3 Option C: Means-tested grant for 18 to 24 year olds plus top-up if in education or training

This option is a variant of Option B and entails an unconditional means-tested grant made available to all young people below a certain income threshold. In addition, a substantial conditional top-up is made available to people in education and training, paid at the rate of the Disability Grant and the OAG. In consequence, unemployed youth would receive a lower amount than low-income youth in education or training. Such a grant would promote participation in education or training.

3.4 Option D: Means-tested grant for 18 to 24 year olds if in education or training

This option narrows down the focus to youth in education or training only. They are means-tested in the same way as for Options B and C. In essence, Option D is a hybrid of the CSG and FCG; like the FCG, it is extended beyond the age of 17 for those in education, but like the CSG it is means-tested. The option provides no support for unemployed young people who are not in education or training. It would penalise low-paid workers who may fall below the means-test yet not qualify because they are not in education or training. A potential (but again unquantifiable) unintended consequence might be that people may cease or defer working and enter education or training in order to qualify, but end up ultimately worse-off if the training is of a poor quality and does not enhance their employability.

3.5 Option E: Means-tested grant for 18 to 24 year olds who are unemployed or in education/training

Option E provides a modicum of income support for low-income unemployed youth, as well as for low-income youth who are in education or training. For this option, the young person (and their spouse if they have one), are means-tested. The grant is paid at the rate of the CSG. Unemployed and discouraged unemployed youth are taken into account, as are those who are in education or training. Other issues to be considered include the fact that low-paid and/or under-employed workers would be excluded from this option (although they would not be excluded from Options A, B and C).

3.6 Option F: Opportunities voucher: One-off allocation to recipients of the CSG or FCG and topped-up if completed high school

Option F is a one-off allocation to beneficiaries of the CSG or FCG. It could, in principle, be considered alongside each of the other options. For example, a voucher of R3500 could be committed to the social grant ‘wallet’ or account of a young person if they were registered for the CSG or FCG, and this amount would rise to R6500 if they completed high school. A one-off and one-time commitment would be made to beneficiaries at age 16 or at 17 if they were not registered at age 16. The recipient could access the opportunity between the ages of 16 and 24. At age 24, the voucher would expire. The resources could only be used for a suite of services that have been assessed as having a high probability of improving employability, such as a payment to an employer for a first work opportunity, a post-school training or education opportunity, investment into a small business, or the cost of books or transport. A voucher of this amount is unlikely to be enough to catalyse a full year's employment or the full cost of a post-school education opportunity. However, upon entry, some public programme funding could be channelled in the same way. For example, upon successful education or employment results, other programmes such as the National Student Financial Aid Scheme bursaries or the Expanded Public Works Programme (EPWP) non-state-sector employment incentive might follow.

The opportunities voucher would be aimed at improving access to options available in a wide variety of programmes and institutions, ranging from further education and training and higher education and training, to recruitment agencies, to private-sector recruitment, to EPWP non-state-sector employment incentives. The idea is not to develop these programmes, but rather to leverage their existence. The opportunities voucher would simply provide the financial mechanism that enables demand-side access – where education-seekers or work-seekers can access opportunity directly. The idea is to enhance the chances of getting a first opportunity, upon which further opportunities can be stimulated, and is also aimed at incentivising high school completion.

The mechanics of an opportunities voucher would differ from the social assistance grants. The programme would depend on the growing sophistication of the social grant payment services system. This will involve improved bio-informatics of beneficiaries (the child or youth) and recipients (the guardian), the introduction of smart cards, the ability to transfer funds, and the registration of vendors for non-cash transactions. To make an opportunities voucher work, the beneficiaries (the youth) would have to be registered and be given their own smart cards and accounts.

Different institutions already manage the registration of entities that could qualify. For example, the Department of Social Development manages the registration of non-governmental organisations. The South African Revenue Service has tax-compliant companies. Education institutions, whether public or private, are registered. Beneficiaries of expanded public works support are monitored by the Department of Public Works. Other providers, such as transport or educational books, could be separately registered. It is envisaged that the programme would rely on these separate registries, and no new registration process would be needed. The SASSA social grant transaction facility will already enable cash withdrawals at retail stores, transfers to bank accounts or third parties, third-party payments and the purchase of goods and services. The opportunities voucher programme would have to identify which group of entities could qualify, expanding them over time as the programme grows.

A meaningful monitoring and evaluation service would be required to ensure that funds are used for the purpose intended, and to assess the impact on youth employability. These functions could include monitoring whether the beneficiary has completed high school, monitoring the validity of registered vendors (there would be reliance on existing forms of registration such as the South African Revenue Service, the non-governmental organisation registry at the Department of Social Development, South African Qualification Authority or Skills Education Training Authorities, etc.), and generally monitoring the impact of the programme.

Such a programme would be well targeted, reaching youth in the poorest households. The payment structure is designed to encourage young people to complete high school. The grant would test whether putting resources into the work-seeker or education-seeker's hands would reduce the institutional barriers to accessing opportunity. It would test whether the government should consider partly funding education and training or the non-state-sector EPWP employment (which pays non-governmental organisations to hire at an EPWP stipend rate) through this channel.

4. Coverage, cost and impact on poverty of the options

To simulate Options A to E, a South African Microsimulation Model (SAMOD⁴) was used (Wilkinson, 2009). Microsimulation modelling is a technique whereby a set of rules (in this instance, South Africa's tax and benefit arrangements) are applied to individuals as captured in a micro-dataset such as a social survey in order to simulate, at individual level, the impact of the rules on individuals and on aggregate for the population as a whole (see for example Zaidi et al., 2009; Mitton et al., 2000). The impact of existing or hypothetical tax and benefit arrangements can then be quantified, in terms of the extent to which they reduce poverty and inequality, their cost, and their impact on different groups such as children, workless people and women. In this instance our focus is on young people. SAMOD Version 1.2 was used, which is underpinned by the National Income Dynamics Study 2008 dataset (SALDRU, 2008; Smit, 2010; Wright et al., 2011). SAMOD is a static tax benefit microsimulation model that enables the next-day financial impact to be measured. It does not take into account behavioural responses to policy changes.

In our calculations, we applied the tax and benefit rules as they were in 2011. As SAMOD V1.2 is underpinned by the National Income Dynamics Study 2008 dataset, it was necessary to recast the survey weights to a 2011 time point, inflate the incomes using the Consumer Price Index, and update the policies to a 2011 time-point.

Table 1 presents the number of young people that would have been reached in 2011 if each of the options had been in existence then, as well as the potential cost. As mentioned above, the only change to the existing tax and benefit rules in 2011 that were introduced in the simulations for the ‘status quo’ was that the CSG was extended to all children under the age of 18 with low-income caregivers: by doing so, a further two million beneficiaries were identified aged 15 to 17 inclusive, with an additional cost of R6.46 billion. As can be seen in Table 1, these additional CSG beneficiaries are included for each of Options A to E.

Option A would reach all 18 to 24 year olds, almost seven million young people, at a cost of R21.8 billion annually if paid at the same rate as the CSG. A means-tested grant for young people aged 18 to 24 (Option B) would reach 5.9 million young people aged 18 to 24 inclusive, and would cost R18.4 billion if paid at the same rate as the CSG. Option C would also reach 5.9 million young people, but the top up (to the payment level of the OAG/Disability Grant) for those in education and training would mean that the cost was virtually the same as for Option A.

Option D, offering a means-tested grant for 18 to 24 year olds in education and training, and paid at the level of the OAG/Disability Grant reaches only 339 000 people aged 18 to 24, at a cost of R4.6 billion. This is based on reported numbers of people in the National Income Dynamics Study in this age group in education or training. However, the numbers studying might expand if such a grant was introduced.

Option E, paid only at the level of the CSG, is a means tested grant for 18 to 24 year olds who are unemployed or in education/training. This option would reach just under two million young people and would cost R6.2 billion.

Option F works differently to the others and was not simulated using SAMOD. A liability is generated that the grant beneficiary could access for specified purposes and through registered vendors. This would take place upon leaving school. It is recognised that many (46%) young people leave school without completing high school. This mostly takes place from Grade 10, and especially Grades 11 and 12. Ten per cent of 16 year olds are out of school, as are 17% of 17 year olds and 31% of 18 year olds. About 59% of the cohort makes it to their final year of high school, and of these about 67% pass (Sheppard, 2009). We estimate that a commitment of R1.58 billion would be needed each year for Option F. This is based on an assumption that not all school-leavers and high school graduates will take up the opportunity, where 25% of 16-year-old school-leavers do, ramping up to 70% of high school graduates. While the grant is accessible upon leaving school, most likely taking place between the ages of 16 and 19, it is envisaged that beneficiaries would use the grant between the time of school leaving and the age of 24. This window is suggested to raise the chances of using the grant, in a context that is recognised to have many institutional barriers and a slow-growing economy. It is envisaged that up to 550 000 beneficiaries across different age groups would be eligible to access their opportunities voucher in any one year, but that an average of 270 000 would activate it annually, and a large proportion would be those in the older group who had completed Grade 11 or more.

Options A to E are focused on immediate poverty alleviation, with some of the scenarios offering top-ups to those in education. There is some evidence that, for those in deep poverty, immediate poverty alleviation can contribute to improved education and labour market participation in the longer term. The top-up for education is meant to be encouragement of human capital attainment that will affect longer term poverty reduction. Option F is focused on improving lifelong earnings, but is only a one-off payment and so can be considered in conjunction with any of the other options.

The remainder of this section considers the extent to which Options A to E raise young people above the poverty line. In order to estimate poverty, household income was calculated and then divided by the number of people in the household to give a per-capita income. The poverty lines used for the analysis are the lower-bound poverty line proposed by Hoogeveen & Özler (2006) (i.e. R322 per capita per month in 2000 prices), which is R604 per capita per month in 2011 prices, and the upper-bound poverty line (i.e. R593 per capita per month in 2000 prices), which is R1113 per capita per month in 2011 prices.

Poverty estimates are presented for these lower-bound and upper-bound poverty lines for 15 to 17 year olds, 18 to 24 year olds, and for a combined group of 15 to 24 year olds (Table 2 and Table 3 respectively) and for the total population (Table 4 and Table 5 respectively). Results are shown for the impact of the ‘status quo’ and Options A to E on poverty head counts, the depth of poverty and the severity of poverty. The poverty head count refers to the proportion of people in the age group that fall below the upper-bound or lower-bound poverty line. The poverty depth measure (P1) shows how far people are from the poverty line. The poverty severity measure (P2) places greater emphasis on people that are further away from the poverty line (Foster et al., 1984; World Bank, 2013).

Table 4: Poverty rates for total population status quo and Options A to E using the lower-bound poverty line, 2011

Option	Poverty headcount (P0)	Poverty depth (P1)	Poverty severity (P2)
Status quo	0.487	0.263	0.181
A	0.466	0.233	0.149
B	0.466	0.233	0.149
C	0.462	0.229	0.146
D	0.482	0.257	0.176
E	0.481	0.253	0.169

Note: Figures are based on 100% take-up.

Table 5: Poverty rates for total population for status quo and Options A to E using the upper-bound poverty line, 2011

Options	Poverty headcount (P0)	Poverty depth (P1)	Poverty severity (P2)
Status quo	0.629	0.403	0.298
A	0.624	0.382	0.271
B	0.625	0.383	0.272
C	0.624	0.379	0.268
D	0.628	0.399	0.293
E	0.628	0.396	0.289

Note: Figures are based on 100% take-up.

If a choice was to be made simply on the basis of impact on poverty experienced by those aged 18 to 24, Option C is the most effective; that is, the means-tested grant for young people paid at the rate of the CSG, plus a top up to the Disability Grant and OAG rate if in education/training – using the lower-bound poverty line, poverty for those aged 18 to 24 falls from 54.1% to 47.6%. Using the upper-bound poverty line, poverty for this group falls from 67.5% to 66.0%. In some respects this small percentage point change may seem disappointing, but this is because the poverty levels of this severely deprived group of the population and the households in which they live are such that it would require more than a grant at the CSG rate (for those not in education/training) to make a dramatic difference to their position in relation to these poverty lines. The cost of Option C would be substantial, at an estimated R21.9 billion per annum, reaching 5.8 million people aged between 18 and 24. It would, however, serve to reduce the financial burden placed on other (mainly poor) household members by unemployed young people.

Option C has some attractions as it provides an admittedly small amount of income support to all low-income young people, but provides additional support for those in education/training. The grant therefore financially incentivises young people to participate in education/training. People entering low-paid work or working for a small number of hours per week would continue to be eligible for the CSG-level amount but would cease to be eligible if their income, and that of their spouse if they have one, exceeds the threshold.

If the choice was to be made simply on the basis of affordability, the most affordable option would of course be to do nothing. The least expensive of the options considered is Option F, at R1.58 billion per year, although this would naturally depend on how it is designed and, as a one-off payment, would not provide a regular source of income maintenance. The most expensive option is Option C, at R21.9 billion. Notably, this is only slightly higher than Option A (R21.8 billion), which is the universal grant paid to all young people and would presumably be more straightforward and less costly to implement as there would be no means-test or variable rate of payment. Options D and E, costing R4.6 billion and R6.2 billion respectively, are much less expensive and have an almost identical impact on poverty. However, the reach of the options is quite different, with Option D targeting less than 5% of young people and Option E targeting 28%.

Options C, D and E are the options most focused on linking youth into education and employment. Options C and D would offer a monthly grant at the OAG rate to those in education and training, the component of which could reach 339 000 young people at a potential cost of R4.6 billion annually. The aim would be to encourage and support youth to obtain further education. It could have a substantial poverty reduction impact for this group, as well as improving employability. However, Option D reaches a small group relative to the whole.

Option F is a different sort of proposal – an opportunities voucher. If the vendor provided education services to the young person, there may be limited impact on immediate poverty (depending on how it is designed) and an – admittedly unquantifiable – impact on lifelong earnings. If the vendor was an employer, the impact would be immediate poverty alleviation – through earnings – as well as improving longer term employment prospects. The funds would enable a young person to get a leg through the door of the school-to-work transition, whereupon further resources would be needed. It would test the impact of a demand-side incentive that cuts out much of the red tape and institutional barriers to accessing opportunities. This option leverages the system of social grants to access labour market and education opportunities. Moreover, it takes advantage of the benefits of the system, including effective targeting of vulnerable groups, prior beneficiary identification, and an increasingly sophisticated financial system that enables payments directly to accredited providers.

5. Concluding remarks

More meaningful solutions are needed to address the extent of youth poverty and unemployment in South Africa. Significant efforts are required across a range of sectors, including the income maintenance arena as well as in relation to education and economic policy.

This article explores six options in relation to social security policy and presents findings on their possible reach, cost and impact on poverty. In practice, the options are limitless. Once one or more options are under consideration, more detailed analysis of their mechanics, costs and underpinning objectives – and intersection with other government initiatives for young people – would be required, in addition to efforts to estimate behavioural responses to such policy changes, which is not addressed in this paper. Prior to the introduction of the CSG, the Lund Committee was established in order to carefully review policy options (Lund, 2008). Given that the provision of social security for young people would represent a significant shift towards the goal of comprehensive social security, it is recommended that a process similar to the Lund Committee be established in relation to the provision of income maintenance for young people.

Acknowledgements

The authors gratefully acknowledge the support of the South African Department of Social Development. The results in this paper do not necessarily represent the views of the Department.

Notes

⁴SAMOD (The University of Oxford, The Department of Social Development of the Government of the Republic of South Africa and The University of Essex 2008) was developed using the EUROMOD framework. EUROMOD is a tax-benefit model for the European Union, supported by European Commission funding, developed by a large consortium of European researchers and coordinated by Holly Sutherland at the University of Essex.