Poverty can be placed in context most easily by focusing on the bipolar nature of the South African economy—a bipolarity that was maintained until recently by the legal regime of apartheid. It manifests itself in a basic split between the high living standard enjoyed overwhelmingly by the white minority and the condition of poverty in which the majority of the black population lives. Aggregate social and economic indicators clearly underline these differences. Although, based on GNP per capita and the structure of production, South Africa is considered to be an upper-middle-income country, the benefits deriving from the economy accrue disproportionately to the white minority.1 The income levels of the black population and the social indicators pertaining to this sector of the community—such as life expectancy at birth and infant mortality—are comparable to those of the poorer countries that border South Africa.
A feature of South African poverty is that it follows the traditional urban/rural split observed in many developing countries. However, absolute poverty is concentrated in regions created by administrative flat—namely, the “homelands,” which are largely the repository of women, children, and the aged, who are not engaged in the formal sector of the economy. The income per capita in these outer-peripheral areas is but a fraction of that prevailing in the metropolitan areas of the country.
In the 1970s, the black population made some progress in catching up with the living standards of the white population by moving to the economically active sectors and regions of the economy. In many respects, this progress was the result of the vigorous growth of the South African economy and its concomitant labor shortages, which enabled blacks not only to increase their employment (a trend favored by labor-starved white companies) but also to unionize and thereby to achieve real wage increases.2 However, the reduction in economic growth throughout the 1980s—associated with a worsening in the investment climate brought about by heightened political uncertainty and the imposition of sanctions against South Africa—particularly affected black employment and wage growth, thereby curbing the earlier trend toward income equalization. Thus, GDP growth slowed to about 1½ percent a year in the 1980s, well short of the 2½ percent a year growth in the population. This decline helped to increase the proportion of the economically active population that was without employment opportunities in the formal sector of the economy from 25 percent in 1974 to about 42 percent by the end of 1989.
Interracial Distribution of Income
Given that incomes in South Africa vary widely according to race, discussions of income distribution have mainly concentrated on the distribution of income shares by race,3 which provides some insight into the evolution of South African society. As Table 1 shows, the share of overall income accruing to the white minority has declined steadily over the past two decades while that accruing to the black majority has increased correspondingly.
Income Shares over Time
(In percent of total income)
Income Shares over Time
(In percent of total income)
| 1960 | 1970 | 1980 | 1988 | |
|---|---|---|---|---|
| White | 72.5 | 71.1 | 64.9 | 53.9 |
| Asian | 2.1 | 2.4 | 3.0 | 3.2 |
| Colored | 5.6 | 6.7 | 7.2 | 6.6 |
| Black | 19.9 | 19.8 | 24.9 | 36.3 |
Income Shares over Time
(In percent of total income)
| 1960 | 1970 | 1980 | 1988 | |
|---|---|---|---|---|
| White | 72.5 | 71.1 | 64.9 | 53.9 |
| Asian | 2.1 | 2.4 | 3.0 | 3.2 |
| Colored | 5.6 | 6.7 | 7.2 | 6.6 |
| Black | 19.9 | 19.8 | 24.9 | 36.3 |
If one examines the evolution of income shares by racial group alone, however, one can misconstrue the degree to which income distribution might be becoming more even, because such measures disregard the different rates of population growth among the various groups. For this reason, it is more instructive to focus on estimates of the Gini coefficient, which provide a more direct measure of income distribution between the individual members of society.4 Wilson and Ramphele (1989) (hereinafter, the Carnegie report), quoting de Lange, Roukens, and van Seventer (1986), report a Gini coefficient of 0.66 for South Africa in 1978. The Carnegie report considers this the most accurate estimate, although it also quotes Devereux (1983), who gives a 0.65 estimate for 1976 and a 0.57 estimate for 1980—a “dramatic” improvement in Devereux’s opinion. To put these numbers into perspective, the Gini coefficient for South Africa in 1978 was the highest of any of the 57 countries in the world for which data were available.5 Gini coefficients for the developed, Western economies generally range between 0.35 and 0.40, with only those economies in Latin America that suffer highly uneven income distributions having Gini coefficients close to those in South Africa.6
Using data available from the South African Bureau of Market Research (Nel and Van Wyk (1984)), one can compute a Gini coefficient time series based solely on between-race inequality (see Table 2).7
Gini Coefficients Assuming Income Equality Within Racial Groups
Gini Coefficients Assuming Income Equality Within Racial Groups
| Year | Gini Coefficient Personal Income |
Gini Coefficient Personal Disposable Income |
|---|---|---|
| 1960 | 0.55 | 0.54 |
| 1965 | 0.56 | 0.55 |
| 1970 | 0.53 | 0.51 |
| 1975 | 0.49 | 0.47 |
| 1980 | 0.50 | 0.49 |
| 1985 | 0.51 | 0.47 |
| 1987 | 0.48 | 0.45 |
Gini Coefficients Assuming Income Equality Within Racial Groups
| Year | Gini Coefficient Personal Income |
Gini Coefficient Personal Disposable Income |
|---|---|---|
| 1960 | 0.55 | 0.54 |
| 1965 | 0.56 | 0.55 |
| 1970 | 0.53 | 0.51 |
| 1975 | 0.49 | 0.47 |
| 1980 | 0.50 | 0.49 |
| 1985 | 0.51 | 0.47 |
| 1987 | 0.48 | 0.45 |
The calculations in Table 2 support the notion that income inequality in South Africa is overwhelmingly the result of income differentials between the races. In 1980, for example, the amount of inequality between the races generated a Gini coefficient of 0.50, while the total amount of ine-quality in the economy generated a Gini coefficient of 0.57. These calculations also suggest that the marked slowdown in growth during the 1980s led to a significant slowing in the pace of income equalization compared with that of the previous decade. Finally, a comparison of income distribution before and after taxes suggests a significantly more even distribution after taxes, which points to a relative degree of progressivity in the tax system.
In addition to the problems of racial inequality referred to above, the South African economy is plagued by a spatially distorted distribution of economic activity. These distortions, which have shown only a marginal tendency to narrow over the past two decades, are clearly brought out by Table 3, which distinguishes output and population between a metropolitan core area, an inner-peripheral area, and an outer-peripheral area. It is particularly striking that, as late as 1988, although almost 38 percent of the population resided in the outer-peripheral areas, it accounted for barely 7 percent of the total product.
Distribution of Income Between Regions
Distribution of Income Between Regions
| Product | Population | Product Per Capita | ||||
|---|---|---|---|---|---|---|
| 1970 | 1988 | 1970 | 1988 | 1970 | 1988 | |
| (In percent of total) | (In rand per capita) | |||||
| Outer-peripheral area | 2.7 | 7.3 | 29.7 | 37.6 | 325 | 642 |
| Inner-peripheral area | 34.8 | 32.2 | 34.1 | 25.5 | 3,591 | 4,161 |
| Metropolitan core | 62.5 | 60.5 | 36.2 | 36.9 | 6,093 | 5,396 |
| Total | 100.0 | 100.0 | 100.0 | 100.0 | 3,526 | 3,293 |
Distribution of Income Between Regions
| Product | Population | Product Per Capita | ||||
|---|---|---|---|---|---|---|
| 1970 | 1988 | 1970 | 1988 | 1970 | 1988 | |
| (In percent of total) | (In rand per capita) | |||||
| Outer-peripheral area | 2.7 | 7.3 | 29.7 | 37.6 | 325 | 642 |
| Inner-peripheral area | 34.8 | 32.2 | 34.1 | 25.5 | 3,591 | 4,161 |
| Metropolitan core | 62.5 | 60.5 | 36.2 | 36.9 | 6,093 | 5,396 |
| Total | 100.0 | 100.0 | 100.0 | 100.0 | 3,526 | 3,293 |
Absolute Measures of Poverty and the Vulnerable
Traditionally, poverty has been measured in reference to some absolute living standard. A number of such measures, differing largely in coverage, have been proposed for South Africa (see Table 4).
Alternate Measures of Poverty, 1980–851
(In rand per month)
These figures are compiled from urban centers and refer to average household sizes (5.45 persons per household). In current U.S. dollar terms, these estimates of minimum living standards ranged from $205 to $255 in 1980 and from $125 to $160 in 1985. Rural minimum living standards are estimated to be fairly similar to those in the urban centers.
Alternate Measures of Poverty, 1980–851
(In rand per month)
| 1980 | 1985 | Coverage | |
|---|---|---|---|
| Poverty datum line or housing subsistence level | 195 | 345 | Food, clothing, fuel/lighting, washing/cleansing, and transport |
| Minimum living level (MLL) | 189 | 350 | As above + tax, medical, education, household |
| Supplementary living level | 240 | 446 | MLL + recreation, personal care, pension, medical |
These figures are compiled from urban centers and refer to average household sizes (5.45 persons per household). In current U.S. dollar terms, these estimates of minimum living standards ranged from $205 to $255 in 1980 and from $125 to $160 in 1985. Rural minimum living standards are estimated to be fairly similar to those in the urban centers.
Alternate Measures of Poverty, 1980–851
(In rand per month)
| 1980 | 1985 | Coverage | |
|---|---|---|---|
| Poverty datum line or housing subsistence level | 195 | 345 | Food, clothing, fuel/lighting, washing/cleansing, and transport |
| Minimum living level (MLL) | 189 | 350 | As above + tax, medical, education, household |
| Supplementary living level | 240 | 446 | MLL + recreation, personal care, pension, medical |
These figures are compiled from urban centers and refer to average household sizes (5.45 persons per household). In current U.S. dollar terms, these estimates of minimum living standards ranged from $205 to $255 in 1980 and from $125 to $160 in 1985. Rural minimum living standards are estimated to be fairly similar to those in the urban centers.
Good time-series data on poverty are somewhat fragmented, with data covering only particular periods, regions, or professions. Among the more striking findings of the Carnegie report is that in 1980 approximately 50 percent of all South African households (including those in the TBVC states)8 and 60 percent of all blacks (80 percent of those living in the reserves) lived below the minimum living level (MLL), whereas in 1989, 40 percent of South Africa’s population (excluding the TBVC) is estimated to have lived below the MLL. A further finding is that in 1983–84 poverty ranged from 23 percent to 68 percent (measured as the percentage of households earning less than the household subsistence level) in three black townships in Natal.9
Although comprehensive studies according to racial group are not available, four groups stand out as the most vulnerable: black children, women (especially widows and wives of migrant workers whose remittances tend to be undependable), the elderly, and the disabled. As described in moving detail in the Carnegie report, these groups have few assets, their incomes are highly susceptible to outside shocks, and they do not enjoy the benefit of a dependable safety net. The vulnerable elderly generally fall into two categories: those who are entitled to pensions based on past employment but who often do not receive them for bureaucratic reasons;10 and those between the ages of 55 and 65 who have difficulty finding employment but who are not yet eligible for pensions.
An examination of the vulnerable groups reveals that they generally depend on income transfers for their subsistence since the rural areas of the reserves are simply incapable of sustaining the population densities that have been imposed upon them.11 Because the rate of excess population migration to the economically active areas was slowed down by various rules and regulations (not all directly connected to apartheid), a system of migrant labor evolved in which the inhabitants of the reserves became net importers of consumer goods, with these imports financed through remittances from the largely male workers, who were allowed to migrate to the urban areas.
Interracial Distribution of Social Indicators
Direct measures of income tend to understate the incidence of poverty because they exclude wealth, certain sources of unrealized income, and the higher costs of living that habitually face the poor. On the other hand, government action often has a stronger equalizing effect on social indicators than it does on income, so that social indicators could actually be more equal than shown by the data on the distribution of income. Table 5 summarizes a number of social indicators by racial group for which consistent time series are available.
Selected Social Indicators
1980 figures for life expectancy are computed at age 1 year.
Selected Social Indicators
| Year | White | Colored | Asian | Black |
|---|---|---|---|---|
| Life expectancy at birth (In years) |
||||
| 1970 | 68.5 | 58.0 | 62.5 | … |
| 1979 | 69.9 | 61.5 | 64.8 | … |
| 1980 | 69.5 | 58.6 | 65.5 | 58.5 |
| Literacy (Rates of adult literacy) |
||||
| 1970 | 98.0 | 69.0 | 74.0 | … |
| 1979 | 99.3 | 84.5 | 92.4 | … |
| 1980 | 99.3 | 84.5 | 92.4 | 67.0 |
| Infant mortality (Deaths per 1,000 live births) |
||||
| 1950 | 35.7 | 134.3 | 68.5 | 165.0 |
| 1960 | 29.6 | 128.6 | 59.6 | 95.0 |
| 1970 | 21.6 | 133.5 | 36.4 | 85.0 |
| 1980 | 13.1 | 60.7 | 24.4 | 70.0 |
| 1988 | 13.2 | 57.5 | 17.4 | 57.4 |
1980 figures for life expectancy are computed at age 1 year.
Selected Social Indicators
| Year | White | Colored | Asian | Black |
|---|---|---|---|---|
| Life expectancy at birth (In years) |
||||
| 1970 | 68.5 | 58.0 | 62.5 | … |
| 1979 | 69.9 | 61.5 | 64.8 | … |
| 1980 | 69.5 | 58.6 | 65.5 | 58.5 |
| Literacy (Rates of adult literacy) |
||||
| 1970 | 98.0 | 69.0 | 74.0 | … |
| 1979 | 99.3 | 84.5 | 92.4 | … |
| 1980 | 99.3 | 84.5 | 92.4 | 67.0 |
| Infant mortality (Deaths per 1,000 live births) |
||||
| 1950 | 35.7 | 134.3 | 68.5 | 165.0 |
| 1960 | 29.6 | 128.6 | 59.6 | 95.0 |
| 1970 | 21.6 | 133.5 | 36.4 | 85.0 |
| 1980 | 13.1 | 60.7 | 24.4 | 70.0 |
| 1988 | 13.2 | 57.5 | 17.4 | 57.4 |
1980 figures for life expectancy are computed at age 1 year.
The social indicators document substantial disparities between the races, which have tended to narrow significantly over the last four decades, partly owing to the large initial level of disparity evident in the figures. One should also note that these data do not include the TBVC states after 1980—which could account for a significant part of the observed improvement.
Literacy indicators show that blacks lag substantially behind all other races. Data on pupil-to-teacher ratios in 1985–87 reinforce this claim, showing ratios for blacks to be more than twice those for whites and for Asians and only slightly less than twice those for coloreds (Trotter (1990)). In 1987, only 65 percent of black children between the ages of 5 and 19 attended school, compared with over 77 percent for the other racial groupings. Moreover, 1986 data on the highest level of education achieved by race show that 37.4 percent of blacks over 20 years old had no education, while only 2.1 percent had more than 12 years of schooling, which compared with over 25 percent for whites. Black enrollment at all levels of education is growing rapidly, and the disparities in expenditures per pupil on a racial basis are narrowing. Nevertheless, in 1990 the expenditure per pupil on blacks amounted to only R 910, which compared with R 4,090 for whites, R 2,410 for coloreds, and R 3,055 for Asians.
South Africa in Comparative Perspective
This profile has made a number of claims with respect to income and other social disparities within South Africa. But how does South Africa compare with its neighbors? Table 6, drawn mainly from World Bank data, attempts to answer this question.
Comparative Economic Indicators for South Africa and Its Neighbors
As classified by the World Bank.
Comparative Economic Indicators for South Africa and Its Neighbors
| Per Capita GNP (U.S. dollars) | Per Capita Daily Calorie Supply | Life Expectancy | Infant Mortality (per 1,000) | Adult Literacy (15+, in percent) | ||
|---|---|---|---|---|---|---|
| 1989 | 1988 | 1989 | 1988 | 1985 | ||
| South Africa | 2,460 | 3,035 | 61.3 | 50.7 | 85.0 | |
| White | 6,530 | … | 73.0 | 13.2 | 99.3 | |
| Black | 670 | … | 57.4 | 57.4 | 80.0 | |
| Upper-middle-income countries1 | 3,810 | 2,980 | 67.4 | 45.0 | 77.7 | |
| Botswana | 1,600 | 2,269 | 67.0 | 41.0 | 70.0 | |
| Lesotho | 470 | 2,307 | 56.0 | 98.0 | 72.6 | |
| Mozambique | 80 | 1,632 | 49.0 | 139.0 | 27.6 | |
| Namibia | 500–1,499 | 1,889 | 57.0 | 65.0 | 73.0 | |
| Zimbabwe | 640 | 2,232 | 63.0 | 49.0 | 62.3 | |
As classified by the World Bank.
Comparative Economic Indicators for South Africa and Its Neighbors
| Per Capita GNP (U.S. dollars) | Per Capita Daily Calorie Supply | Life Expectancy | Infant Mortality (per 1,000) | Adult Literacy (15+, in percent) | ||
|---|---|---|---|---|---|---|
| 1989 | 1988 | 1989 | 1988 | 1985 | ||
| South Africa | 2,460 | 3,035 | 61.3 | 50.7 | 85.0 | |
| White | 6,530 | … | 73.0 | 13.2 | 99.3 | |
| Black | 670 | … | 57.4 | 57.4 | 80.0 | |
| Upper-middle-income countries1 | 3,810 | 2,980 | 67.4 | 45.0 | 77.7 | |
| Botswana | 1,600 | 2,269 | 67.0 | 41.0 | 70.0 | |
| Lesotho | 470 | 2,307 | 56.0 | 98.0 | 72.6 | |
| Mozambique | 80 | 1,632 | 49.0 | 139.0 | 27.6 | |
| Namibia | 500–1,499 | 1,889 | 57.0 | 65.0 | 73.0 | |
| Zimbabwe | 640 | 2,232 | 63.0 | 49.0 | 62.3 | |
As classified by the World Bank.
Compared with other upper-middle-income countries, South Africa does well on the adult literacy indicator, average on the per capita daily calorie supply indicator, and poorly on the health indicators (life expectancy and infant mortality). South Africa is richer than its neighbors, and its people eat and read better. However, some of South Africa’s poorer neighbors (Botswana and Zimbabwe) enjoy longer life expectancies and lower infant mortality, and Namibia, which is also poorer than South Africa, registers a lower infant mortality rate.
The dual nature of the South African economy becomes evident after South African indicators are disaggregated into white and black averages. These indicators clearly place the white sector of the South African economy in a category different from either upper-middle-income countries as a group or any of South Africa’s neighbors. On every indicator cited, white South Africa does much better, while black South Africa, if one allows for data uncertainties and for the fact that the TBVC states are excluded from these indicators, appears to be doing only about as well, on average, as its neighbors. Botswana and Zimbabwe outperform black South Africa on health (Botswana also does much better on income) but not on literacy. Lesotho does slightly worse, while Mozambique—an extremely poor country—does much worse than black South Africa on all counts. Although it is possible to question the accuracy of individual indicators, the overall picture is harder to dispute: black South Africa has done at best only moderately better than its neighbors despite its proximity to Africa’s most dynamic economy.
The World Bank (1990b) records a per capita GNP for South Africa of $2,290 in 1988, grouping it with the upper-middle-income Latin American countries, Hungary, and the Republic of Korea.
Hofmeyr (1990) and van der Berg (1989a) show that the real wages of blacks in mining, which were constant between 1945 and 1970, subsequently increased sharply. The Carnegie report (1989) documents a 240 percent increase in real terms of black wages in mining and quarrying between 1970 and 1985.
One issue that has in the past occupied the attention of the authorities has been the so-called poor white issue (for a survey, see Abedian and Standish (1985)). It will not be discussed here, not least because the South African Government has been attending to it since the 1920s, and white poverty appears to have been largely eradicated.
The Gini coefficient, a widely used standardized measure of income distribution, would approach zero in the case of a perfectly even distribution of income and would have an upper limit of unity at the opposite extreme.
Carnegie report (1989).
McGrath (1990), p. 94.
The data on income distribution from the Bureau of Market Research have the advantage over those of McGrath (1990) in that they include the “independent” homelands after 1980, which makes it possible to examine income distribution on a consistent time-series basis. It is important to include these homelands in any comparison because they tend to contain the poorest segments of the black population.
The Transkei, Bophuthatswana, Venda, and the Ciskei.
Carnegie report (1989).
For example, pensions are denied to people not living in the legally defined homelands.
Under apartheid, the various land acts restricted the areas where blacks could legally own land to only 13 percent of the country, and these areas were generally regarded as being below average in agricultural potential. In addition to a communal share of about 6½ hectares of grazing land, the average arable land holding per household was between 1½ and 2 hectares in 1985. According to the Urban Foundation, 56 percent of the population in the outer-peripheral areas consisted of small-scale landowners with below-subsistence production levels, while 31 percent of the population in these areas comprised resource-poornonl and holders.