This resource is hosted by the Nelson Mandela Centre of Memory, but was compiled and authored by Padraig O’Malley. It is the product of almost two decades of research and includes analyses, chronologies, historical documents, and interviews from the apartheid and post-apartheid eras.
The purpose of this report is to measure levels and analyse sources of inequality across regions nationally and within former development regions E (KwaZulu/Natal) and F (Eastern Transvaal).
Inequality analysis is undertaken to provide a basis for assessing differential regional consequences of national policy. The two regions chosen for more detailed examination illustrate a range of regional implications of national policy. For example, growth in the two regions has been extreme, with KwaZulu/Natal having abnormally low and Eastern Transvaal abnormally high growth. Population distributions are divergent, with Region E having a very high proportion of the population in former homelands and Region F a low proportion. The sectoral composition of the regions is also dissimilar, in that Region E is highly concentrated in manufacturing, while Region F is unique in the importance of extractive and energy sectors. Comparing these two regions hi. blights how different population and sectoral structures have affected growth and hence inequality among regions.
The study is limited by data availability to the 1980s, with some reference to the 1970s. For this period, several groups of regions have been examined. The most aggregate group is the former development regions; the next level of disaggregation distinguishes between the former homeland and provincial parts of the development regions. From this disaggregation it is possible to compare inequality both across and within provinces and homelands. In addition, comparisons can be made between former TBVC and 'self-governing' homeland areas. The final level of disaggregation is to the level of the magisterial district, which is used for the analysis of inequality within Eastern Transvaal and KwaZulu/Natal.
Conventional measures of inequality, including coefficients of variation and ratios of maxima to minima, are calculated for the three levels of regional disaggregation for several demographic and economic variables. The significance of inter- and intraregional inequality is determined by simple statistical techniques such as analysis of variance. For Regions E and F, econometric estimation of equations derived from the limited variables available allows some comparison of the structure and behaviour of each region with That of the country as a whole.
Assessment of the impact of various national policies is made on the basis of descriptive economic and demographic data for each region supplemented by regional input, import and export multipliers derived from existing input-output and social accounting matrix data. These assessments are largely qualitative rather than derived from formal, quantitative models.
Summary of results
Summary of inequality results
Several patterns emerge from the data on regional structure, performance and inequality. At the development region level there is a group of three poor regions (former development regions G, D and E) which have in common low rates of urbanisation, economic activity and participation, but high shares of females in the population, high dependency ratios and population growth rates.
At a more disaggregate level, however, the pattern is less clear. In the former provinces, for example, the most successful areas (the former PWV, Western Cape and Eastern Transvaal) all succeeded with very diverse leading sectors. The poorer provinces, on the other hand, do resemble the poor development regions in economic and demographic structure.
Former homeland areas, however, do not exhibit a common set of features associated with poverty. The poorest areas are unlike each other in demographic characteristics such as dependency ratios, female shares and urbanisation rates. What they have in common is the position of having been bantustans with impoverished resource bases.
For the limited time period considered, inequality in summary indicators such as GGP per capita generally increased over time at the level of subregions. This result is troubling, particularly because at the same time inequality declined for a number of demographic variables such as urbanisation rates, dependency ratios and participation rates.
Inequality within groups of subregions highlights other problems for distributional policy. As expected, inequality is greatest for the group including both former province and former homeland subregions. However, inequality among former homelands was also very high, indicating that race is not an exclusive determinant of income variation. Although all former homelands are poor, the economic and demographic differences among them are significant.
Within KwaZulu/Natal and Eastern Transvaal inequality is dominated by the difference between former homeland and provincial areas. Comparing inequality within each group of districts in KwaZulu/Natal, those variables for which the homeland average values differ most from the non-homeland averages are also those for which homeland inequality is lower than provincial inequality. Therefore, what makes the former homelands distinct as a group also makes them more similar to each other. This is the reverse of the result obtained at the level of development regions. In Eastern Transvaal, however, inequality among the former homeland districts is greater than across provincial areas, which supports the national result.
Summary of policy implications
Policy implications will be summarised for two sectors, manufacturing and agriculture. In both cases the regional impact of national policy depends upon the specific distribution of each region's economic activity. Those regions with disproportionately high shares of manufacturing will be particularly vulnerable to South Africa's post-liberalisation role in the international division of labour. If, for example,
national policy is designed to encourage South Africa's entry into high technology or high-end consumer goods, the richest regions will reap large benefits and the manufacturing base of the poorer regions be disadvantaged further. If, on the other hand, market opportunities are promoted either within South Africa or in neighbouring states for mass-market products, manufacturing in KwaZulu/Natal and the Border Kei areas could be the largest beneficiaries, ceteris paribus (other things being equal), with a net equalising effect.
Similarly, in agriculture the crop mix in each region determines its likely future after both land reform and liberalisation. In general, if exports are the main target of government policy, the already richer regions will gain overall because of both domestic conditions of production and international prices and demand for various crops. Policy aimed more at food production for domestic use - also because of location of crop production and the nature of domestic demand - is more likely to benefit the relatively poor regions.