Inequality in Post-Apartheid South Africa

In the course of globalisation and increasing liberalisation, the question of how the costs and benefits of these processes are shared is of special concern. Recent studies suggest that they are shared unequally. Although inequality between countries is declining, global inequality is on the rise owing to the fact that inequality within countries is increasing. These developments raise the question of how growth affects inequality and vice versa. Further it has to be explored what role governments have to play to ensure that all members of society benefit from growth.

Kuznets suggested that countries would initially face rising inequality as they grow richer, however, as a certain threshold of per capita income is reached the degree of inequality would decrease. From this proposition it is easy to conclude that governments do not have to be concerned about rising inequality as it eventually will decrease once the economy has reached a certain level of economic development. It may also imply that there is a trade-off between growth and equality.

More recent studies challenge Kuznets’ proposition. These studies indicate that high levels of inequality hamper economic growth. Yet, economic growth is an important instrument in the fight against poverty. It is assumed that economic growth creates employment opportunities and increases the scope for distributive policies and thus tends to reduce poverty. Therefore, any country that is determined to address the problem of poverty ought to pay attention to the nexus between inequality and growth.

In the theoretical part of this paper, arguments that support and challenge Kuznets’ hypothesis are analysed. Thereafter an interesting proposition developed by Addison and Corniais presented that suggests that there may be an efficient level of inequality. From these different arguments conclusions are drawn and policy implications presented. The theoretical part does not attempt to analyse the causes of inequality. To identify the causes of inequality, it would be necessary to undertake a case-by-case analysis. The general assumption for the theoretical part is that inequality is caused by unequal access to resources as well as inadequate economic policies which can lead to persistent unemployment and poverty. Unequal access may be caused by market failures especially in the capital and labour markets, or by historical and political developments that create uneven power constellations, exclusion, or discrimination.

The case study deals with the question of whether growth per se is responsible for inequality and persisting poverty in South Africa.

At the beginning of the section an overview of the apartheid era is presented in order to identify the roots of South Africa’s highly unequal income distribution and persistent poverty. South Africa is an interesting case as until 1994 apartheid policies discriminated against major parts of the society and created a skewed distribution of resources and income as well as unequal opportunities in the labour market among the different population groups. Furthermore, South Africa is a case in point illustrating how a government is able to influence the distributional outcome by means of its choice of economic policy. The case study also shows that the discriminatory policies of the apartheid era had a long lasting impact on the South African society, which still struggles to cope with the legacy of apartheid. Despite the strong commitment of the democratic government to fight poverty and inequality and despite the relatively good growth performance, the welfare situation in post-apartheid South Africa has hardly, if at all, improved. South Africa is still among the countries with the highest degree of inequality in the world and large parts of the society suffer from severe and persistent poverty.

Compared to other middle income countries that have similar levels of per capita GDP (e.g. Mexico, Brazil, and Malaysia), South Africa is faced with the highest poverty headcount ratios among these countries (see appendix A). However, it is difficult to draw a cross-country comparison, due to the lack of adequate data and the fact that countries apply national poverty lines in order to illustrate their welfare situation. Furthermore, it is not the scope of this paper to draw a cross-country comparison. Instead this paper attempts to analyse the evolution of growing or persistently high levels of inequality and poverty with concomitant high levels of unemployment in the post-apartheid era. Subsequent government strategies to address the problem are considered, in order to ascertain whether the government’s policies have been well targeted. The case study is primarily based on different studies that have analysed October Household Surveys (OHS), Labour Force Surveys (LFS) and Income and Expenditure Surveys (IES) as well as the data of national censuses and the All Media and Products Surveys (AMPS). Due to the limited availability of current data, the paper primarily focuses on the period from 1994 until 2005.

Chapter 2, In Response to Kuznets:

Chapter 2.1, Inequality as an Accelerator of Growth:

In growth models that follow a Keynesian notion, economic growth depends on the ability of an economy to accumulate capital.

In post-Keynesian models, owners of different production factors have different propensities to spend or to save respectively. Capitalists tend to save more than wage earners. According to some post-Keynesian models, a profit biased economy exhibits a higher savings ratio compared to a more egalitarian society and hence is better able to accumulate capital. This, in turn, leads to faster growth. Applying the Harrod-Domar growth model, that regards savings as a precondition for investment, higher savings would eventually lead to higher growth. Kaldor’s assumption that capitalists have a higher propensity to save implies that an economy in which profits are higher and real wages are lower, would tend to grow faster than an economy with lower profits and higher real wages. This model is consistent with the Kuznets hypothesis in so far as income inequality will cause an efficient outcome at the initial stage of economic development.

In contrast to these post-Keynesian models, neo-Kaleckian models see a dual function of wages: wages are costs and reduce profits; on the other hand a worker’s wage is in itself purchasing power that is important for aggregate demand. Depending on the characteristics of the economy, one can distinguish between profit-led and wage-led regimes. In a profit-led regime, the cost aspect of wages dominates and hence an increase in the wage share would have a negative impact on growth through its dampening effect on profits. On the other hand, in a wage-led regime the demand aspect prevails; any increase in the wage share would stimulate demand and that would more than compensate for the negative effect of growing costs on profits. The distinction between profit-led and wage-led regimes has an important implication for redistributive policies: being in a wage-led regime would allow for redistribution towards wage earners at the expense of capitalists without jeopardising growth. However, within a profit-led regime the redistribution in favour of wage earners would dampen growth.

Aghion suggests a further case where inequality might be more efficient than an egalitarian distribution of income. It is assumed that at the initial stage of development large investments are required to boost growth. These investment endeavours are often characterised by being indivisible. Operating in an economy that does not have a sophisticated financial sector and in which the use of shares as a way of financing large investments is difficult, a concentration of income among a small privileged group may be a pre-requisite to carry out these initial lump sum investments. However it is questionable whether the privileged group is willing to undertake investments within the underdeveloped home market. It is quite possible that the privileged group favours investment opportunities abroad and therefore channels funds to more developed markets.

In contrast to the above presented models on capital accumulation, Galor and Tsiddon take the development of differences in wages for skilled (high level of education) and unskilled (low level of education) workers into account. According to their overlapping-generations model, economic growth induced endogenously leads to a Kuznets’ style inverted-U. Galor and Tsiddon assume that technical progress is driven by human capital development. Further, the aggregate human capital of a generation has a positive influence on the level of human capital and thus the efficiency of future generations. Additionally, the better educated the parents, the higher the return on education investments of the offspring. As the level of technology increases, the yield for educated individuals of future generations rises. The higher return on human capital offers an incentive for further investments in education. Galor and Tsiddon suggest in their model that initial inequality is necessary to create sufficient incentives for human capital investments. A too egalitarian society might face the problem of being trapped at a stage where human capital investments are insufficient to create technical progress and, hence, economic growth. Once a certain level of human capital is reached, a transfer of know-how from high to low human capital sections occurs. This transfer is fuelled by the increasing technical standards used in the production process. As skills become more precious in the market, the incentives to invest in human capital increase for all, regardless of the initial human capital level. At this stage, it is crucial to note that the model assumes diminishing returns to skills, and therefore individuals that were initially located in lower income (lower skill) brackets of society tend to invest a higher proportion of their income on human capital accumulation.

In counterpoint to the above-discussed theories, Addison and Cornia argue that the Kuznets hypothesis and the abovementioned models supporting it lack empirical verification. Data relating to OECD countries only support an inverted-U shape up until the 1970s. Thereafter, most of the OECD countries experience rising inequality again. Bénabou presents evidence that highly unequal countries such as the Philippines and most of the Latin American countries grew far more slowly than most of the emerging Asian countries such as Korea, which exhibited a more egalitarian distribution pattern at their initial stage of development compared to most Latin American countries.

In the following section theories that establish a negative relationship between inequality and growth are presented.

Chapter 2.2, Accumulation and Redistribution:

The extent of inequality may influence the level of the tax rate, since the tax rate can be seen as a proxy for redistribution. High inequality may increase the pressure for more redistribution and hence for higher tax rates. Alesina and Rodrik have developed a model of redistribution and endogenous growth based on the so-called median voter theorem. Members of a society are heterogeneous in terms of their ownership of factors. In general there are two factors: labour and capital. Labour is regarded as nonaccumulative, whereas, capital is deemed accumulative. Growth depends on the society’s ability to increase its capital stock. This ability is influenced by the saving decisions of the members of the society. Furthermore, Alesina and Rodrik incorporate public services into the model that are productive and beneficial for all members of society. Taxation on capital is used to fund these services. By taxing capital, the government influences the savings decision of individuals. A high capital tax rate inhibits accumulation and thus growth; whereas, a low tax rate has the opposite effect. Due to the differences in factor endowment, each person has his or her optimal tax rate. Capitalists would find a low capital tax preferable as in turn it would maximise the economy’s growth rate; whereas, wage earners would prefer a higher tax rate which, however, would lead to lower growth. Considering the median voter theorem, the existing tax rate reflects the median voter’s preference. Therefore, an economy where the median voter is relatively poor in terms of capital compared to the average would choose a higher tax rate, and thus retard growth. The model implies that a functioning democracy exists and that government policies are aimed at redistribution. Further it is presumed that the returns to private investments are higher compared to those of public investments. Contradicting these assumptions, Easterly and Rebelo as well as Perotti provide empirical evidence that public spending, for instance on education, yields high returns and, therefore, it is instead growth enhancing.