Enabling growth: redistribution priorities for South Africa

Reads 3,058
Andrew Donaldson, 12 November 2014
If the National Development Plan is to be effectively implemented, we need clarity about the mechanisms through which growth and redistribution can be jointly advanced. Priorities include social security reform and quality improvements in social services, urban development, housing and public-transport investment. Expanding employment opportunities is the most pressing challenge, requiring policies that might include: support for labour-intensive industry and agriculture, small enterprise and informal sector development, well-targeted skills programmes, and wage or employment subsidies. Recognising the complementarity between redistributive and growth-enhancing measures is essential.


South Africa’s National Planning Commission has set the reduction of inequality as a central goal. In the Commission’s Diagnostic Report, inequality, unemployment and social fragmentation were identified as the main barriers to rapid and sustainable growth and development. The subsequent National Development Plan (NDP) has set a target of reducing the Gini coefficient from 0.69 to 0.60 by 2030 (NPC 2012:34).

The NDP is a complex narrative of inter-connected economic and social analyses, organised into twelve ‘key priority areas’. In each of these areas, there are strong redistributive elements. The underlying idea is that inclusive development and a more equitable pattern of investment in jobs, infrastructure, capabilities and institutions will reinforce growth and create a more equal distribution of income and opportunities.[1]

That redistribution can enhance growth prospects has been demonstrated, among others, in recent research published by the International Monetary Fund (see Donaldson, Econ3x3, 7 October 2014). For policy purposes, however, it is important to be clear about the channels through which redistribution is expected to support growth. It is also important to identify particular interventions or programmes that are likely to have a rapid and effective effect on growth and distribution in South Africa’s context and circumstances.

While the NDP emphasises many development priorities and redistributive interventions, it provides little guidance on their relative effectiveness, the appropriate scale and their likely impact on either growth or distribution.

Getting to more specifics: international experience

A recent IMF Fiscal Affairs Department Staff Paper (IMF 2014) provides a useful point of departure. The IMF authors note that the scale of fiscal redistribution in advanced economies is substantial: direct income taxes and transfers decrease the inequality of monetary income by about one-third, or about 15 Gini coefficient points on average. (In OECD countries, the median market-distribution Gini coefficient is about 0.45, while the median net-distribution Gini coefficient after taxes and fiscal transfers is about 0.30 – see Donaldson, Econ3x3, 2014).[2]

In these countries, inequality reduction is largely achieved through social security arrangements such as public pensions and family benefits. These serve to smooth incomes over lifetimes and pool income risks associated (inter alia) with longevity. They are often funded by earnings-related taxes or contributions, and typically the receipt of the ensuing benefits is also based on earnings. In the overall reduction of inequality in these countries, contributory social security benefits and transfers are more significant – though  non-contributory income transfers and progressive personal income tax structures do contribute to redistribution.[3]

In developing economies, however, fiscal redistribution typically plays a smaller role. This is in part because revenue and social spending are lower relative to the national income and also because social security systems are less advanced. In-kind social spending (such as spending on education, health and welfare services) may be redistributive, but the poor do not always have full access to social services, while higher education and urban services usually benefit higher income groups disproportionately.

Accordingly, the IMF’s recommendations focus on improving pro-poor social spending programmes, together with appropriate measures to enhance revenue. The authors propose that developing economies should:
(a) consolidate social assistance[4] programmes and improve their targeting,
(b) introduce and expand conditional cash transfer programmes[5] as administrative capacity improves,
(c) expand non-contributory, means-tested[6] social pensions,
(d) improve the access of low-income families to education and health services, and
(e) extend the coverage of personal income taxation.

What are the appropriate priorities for South Africa?    

In South Africa nearly everybody already has access to education and health services, though much needs to be done to improve their quality and effectiveness. In respect of social assistance, old-age pensions, disability grants and child support grants already contribute substantially to the income security of those in the bottom half of the income distribution. Social grants account for over three per cent of gross domestic product (GDP), which is high by international comparison.

Given the statutory basis for South Africa’s social grant system, the scope for the kinds of conditionality that characterise Latin American social transfers is probably limited. A greater redistribution effort could be achieved by raising the value of social grants or by extending their coverage. There definitely is scope for some improvement in the design and administration of social assistance.

However, it is the fragmented and incomplete statutory social security system which is most in need of improvement – despite several years of work on options for social security reform. South Africa’s unemployment insurance fund reaches a small fraction of the unemployed, occupational injury and disease compensation is poorly administered and the current road accident fund arrangement is financially unsustainable and excessively reliant on legal settlement procedures. Statutory social security does not yet include any provision for disability insurance, survivor benefits or earnings-related retirement funding. (These are typically financed through payroll taxes or earmarked contributions – various blends of mandatory and voluntary insurance and savings are found internationally).

As life expectancy increases in South Africa, and as urbanisation proceeds, the need for more comprehensive social insurance will become more pressing. The current non-contributory social assistance system cannot reasonably meet the expanding need for income security in retirement and for insurance that covers death, disability or involuntary unemployment.

The South African experience since 1994 also suggests that significant advances in household well-being and the distribution of opportunities can be achieved through several other areas of public spending or financial support. Finn, Leibbrandt & Woolard (Econ3x3, 2013) illustrate how progress has been made in `multidimensional’ poverty reduction through the provision of housing, electricity, water services and social infrastructure. The NDP reflects a similarly broad understanding of the development and redistribution challenge. Therefore, investment in urban development, housing and public transport remains a priority.

In respect of its tax policy, South Africa has a well-established and progressive income tax structure. The phasing out of many exemptions and allowances since 1994, the introduction of capital gains tax and other base-broadening measures have reinforced the reach, effectiveness and equity of the tax system. Improved tax administration has also contributed to its distributional fairness. A steeper tax burden on high-income earners and on estates and donations might assist in reinforcing fiscal redistribution, but the international tax literature cautions that wealth and capital are mobile and difficult to tax.

Other equity considerations also arise in tax design. Family allowances were removed from our personal income tax in the 1990s. This was part of a necessary simplification of the system. But administrative advances since then have made it possible to bring greater horizontal equity into the tax system by recognising that household living costs are higher for taxpayers with dependants than for single individuals. In keeping with international practice, a child allowance or rebate could be introduced into the tax code alongside (and aligned with) the old-age and disability allowances. That would improve the equity of the distribution of the tax burden.[7]  

A related tax-design question arises from the interaction between rising marginal tax rates (in a progressive personal income tax structure) and the deductibility of contributions to retirement funds and medical schemes. If contributions were fully deductible, for high-income taxpayers the contribution of the fiscus to retirement saving or health insurance would amount to 40% (equivalent to the maximum marginal tax rate); the fiscal contribution would be regressive in that it would benefit higher-income taxpayers more. Partly for this reason, the deductibility of medical scheme contributions has been replaced by a capped credit, which is more equitable. A similar tax treatment of retirement funding would substantially improve progressivity; however, it would need careful design and sequencing to avoid counter-productive incentive effects on work and savings.

Promoting labour market participation

While noting the effectiveness of social spending as ‘a major tool of targeting resources to South Africa’s poor,’ Van der Berg & Moses (2012) have recently argued that redistributive social spending in South Africa is approaching its fiscal limits. The implication is that the emphasis in policy should shift to the improvement of the market distribution of income (rather than relying on redistribution only).

South Africa’s unusually high level of market-based income inequality is largely a consequence of (a) high unemployment – many people do not earn any wage income – as well as (b) highly unequal earnings in the labour market – many people have jobs with low wages (Presidency 2013; World Bank 2012). Policy measures that expand opportunities in the labour market and raise the earnings of informal and low-wage workers are needed, both to reduce inequality and enhance growth. Such policies are more likely to be fiscally sustainable than income transfers. They might include support for labour-intensive industry and agriculture, the promotion of small enterprise and informal-sector development, well-targeted skills programmes, and wage or employment subsidies.

Redistributive wage or employment subsidies have an obvious appeal under South Africa’s circumstances because they would both raise incomes in the lower part of the income distribution and reduce employment costs for employers (thereby boosting the demand for labour). Currently, employment subsidies through the tax system are limited to young first-time work seekers and employees registered for formal learnerships. A general subsidy for low-wage employees would be more strongly redistributive and directly growth-enhancing, and could be structured to target direct formal employment rather than casual or ‘brokered’ employment. A general wage incentive is likely to have a stronger impact on creating employment than the present approach.

Consideration also has to be given to fiscal instruments that enhance households’ capacities to participate in work and generate income, accompanied by institutional reforms to address the barriers faced by prospective work seekers and employers. Investment in housing, public transport and access to electricity and telecommunications networks are good examples. In urban areas especially, such investments in infrastructure reduce the cost of economic participation, job search and doing business – and so play an important role in redressing the spatial imbalances of the apartheid era.

Complementarity and integration as organising principles in policy design

In government budgetary decisions, direct income support programmes compete for funding with in-kind services such as health and education. However, improved health and education outcomes flow from both public services and the own contributions of households. Therefore, policies that increase household incomes actually complement the provision of health and education services by government. Also, income policies can be targeted more easily than, for example, the free (or low-fee) provision of health, education and other public services. More broadly, policies that create employment and increase the earnings of low-wage workers will contribute to households’ capacities to invest in human development as well as to afford user charges for services.

Therefore, a framework for analysing growth and redistribution needs to recognise their macro-complementarity and the micro-linkages between a wide range of policies and programmes – income support, education, health, housing, public transport, welfare services, employment programmes, amongst others.

In practice, micro-linkages or complementarities often have to be addressed locally. If industrial growth should be supported by an expansion of further education, or if mining investment leads to housing and health-service needs, or if residential densification needs to be accompanied by new public-transport routes, these linkages have to be managed and coordinated on a local level. Thus, municipalities have critical responsibilities to coordinate and integrate development. The coordinating role of municipalities in managing the linkages between economic development, job creation, social investment and household service delivery surely deserves more attention than it currently enjoys.[8]

It must be realised, however, that high levels of inequality create special problems for this coordination challenge. In areas such as education, health care and urban transport, service provision tends to evolve in differentiated ways that reflect the spending capacity of different income groups. Often the result is a fragmented, unequal structure in which the allocation of resources and the quality of services diverge – reinforcing the division between high-quality private facilities and low-quality public services (which entrenches inequality between rich and poor). This outcome might account in part for the growth-inhibiting impact of inequality.

Therefore, in implementing policies, one should be mindful that services that are targeted to the poor can have unintended outcomes. There are collective (or network) features of education, health, transport and communication services that are better served – and more efficiently produced – through common delivery platforms. (Examples include curriculum design and the certification of qualifications, internet-based education, medical training and research, immunisation programmes, emergency rescue and recovery services, expensive hospital services, pharmaceutical logistical networks, road and rail infrastructure and commuter transport services.) It is crucial to strive for such integrated social-service and infrastructure networks – even though it is harder to achieve their benefits in an unequal society.


Developing integrated delivery systems in a context of fragmented networks in an unequal society is far from straightforward. Industry structures that accommodate both public and private service providers are more complex than those in which either markets or public provision dominates – and their dynamics are not easily analysed. This is likely to be a fruitful research arena, relating to a wide range of activities and services and many possible solutions. For example, sometimes the challenge is to find ways of keeping the rich within the public delivery system; sometimes it is to find subsidies and tariff structures that improve the access of poor people to market-based services.

This is in part a question of economic efficiency. It is about the scale economies and network characteristics of those elements of the education system, health care, transport, communications and other sectors that warrant integrated and coordinated solutions. It is also about the complementarity and linkages between public provision, regulation and market-based delivery.

A further consideration in the design of redistributive efforts is administrative simplicity. In comparison with targeted in-kind public service delivery, measures that provide universal benefits or cash transfers are institutionally less complex and less prone to leakage, corruption or bureaucratic failure.

There are also political-economy considerations. Social cohesion is served, and accountability and political oversight are likely to be more effective, when there is a shared interest in public service delivery across a broad spectrum of income groups. Where public and private services diverge, with one system serving the poor and another the rich, inequality will tend to become entrenched and the development path is likely to be unstable.


Donaldson A. 2014. Redistribution is part of the toolkit to promote growth. Econ3x3, 7 October. 

IMF (International Monetary Fund). 2014. Staff Report on Fiscal Policy and Income Inequality, IMF Fiscal Affairs Department.

Finn A, Leibbrandt M & Woolard I. 2013. The significant decline in poverty in its many dimensions since 1993. Econ3x3, 29 July.

NPC (National Planning Commission). 2012. National Development Plan. The Presidency. Pretoria.

Presidency of the Republic of South Africa. 2013. Development Indicators 2012. Pretoria.

Van der Berg S & Moses E. 2012. How better targeting of social spending affects social delivery in South Africa. Development Southern Africa 29(1), March.

World Bank. 2012. South Africa Economic Update: Focus on Inequality of Opportunity. Working Paper No. 71553. Washington DC.



[1]The redistributive impact of South Africa’s tax and expenditure programmes has been reviewed in several studies. See Servaas van der Berg (2005), Fiscal Expenditure Incidence in South Africa: 1995 and 2000, a report for the National Treasury; Ingrid Woolard, Charles Simkins, Morné Oosthuizen and Christopher Woolard (2005), Tax Incidence Analysis, the Final Report for the Fiscal Incidence Study for the National Treasury.

[2] The ‘market distribution of income’ is the distribution before payment of taxes and receipt of transfers; the ‘net distribution’ is the distribution of disposable income or expenditure after taxes have been paid and transfers (such as social grants) have been received by households or individuals.

[3] In a contributory system, benefits or entitlements are funded out of regular payments or contributions into a dedicated fund. A non-contributory system is funded out of general revenue and benefits are not linked to membership of a fund or a record of contribution to a fund.

[4] Social assistance refers to non-contributory income support or welfare benefits, such as South Africa’s old-age, disability and child support grants.

[5] Conditional cash transfers, such as Brazil’s bolsa familia programme, involve income support to households that is linked to desired behaviour of recipients, such as children’s enrolment in school or their participation in primary health services.

[6] In a ‘means-tested’ scheme, benefits are restricted to those whose income or assets fall below specified thresholds, as in the case of South Africa’s social grant programmes. ‘Universal’ schemes are not subject to means tests or income-related targeting, though other relevant criteria might apply, such as actively seeking work, disability, or retirement age.

[7] If such tax allowances are properly aligned with old age, disability and child-support grants for those below the tax threshold, it would contribute to the fairness of the combined tax and spending contributions to household income security (and simplify administration). 

[8] The idea that social problem-solving should be managed at the lowest level of government capable of managing the coordination that is required is sometimes referred to as the `subsidiarity principle’. The term has its origins in Catholic social teaching. With effect from the 1992 Maastricht Treaty, subsidiarity is one of the organising principles of the European Union. It implies local or regional management and coordination, with support from central or federal authorities.

Download article

PDF icon Donaldson 2014 Redistribution priorities FINAL.pdf
Post a commentary

This comment facility is intended for considered commentaries to stimulate substantive debate. Comments may be screened by an editor before they appear online. To comment one must be registered and logged in.

This comment facility is intended for considered commentaries to stimulate substantive debate. Comments may be screened by an editor before they appear online. Please view "Submitting a commentary" for more information.