Can South Africa Avoid a Malthusian Positive Check?

Citation metadata

Author: Charles Simkins
Date: Winter 2001
From: Daedalus(Vol. 130, Issue 1)
Publisher: American Academy of Arts and Sciences
Document Type: Article
Length: 7,513 words

Document controls

Main content

Full Text: 

... both historically and philosophically the doctrine of Malthus was a corrective reaction against the superficial optimism diffused by the school of Rousseau. It was the same optimism, with its easy methods of regenerating society and its fatal blindness to the real conditions that circumscribe human life, that was responsible for the wild theories of the French Revolution and many of its consequent excesses.

--"Malthus" in Encyclopaedia Britannica [1]


ALTHOUGH THE STATISTICAL EVIDENCE is not complete, it is clear that the twentieth century has seen substantial lengthening of life expectancy in South Africa. At the beginning of the century, the expectation of life at birth was about forty-eight years for white men and fifty years for white women in the Cape Colony. Colored and black life expectancies were about thirty-two years for men and forty years for women. [2] Table 1 indicates how the situation has developed during the century.

Little has happened to interrupt the general trend. War mortality in South Africa has been relatively light since 1850 and can be summarized briefly. In 1857, after nearly eighty years of border warfare in which they had lost much, the Xhosa of the Ciskei and southern Transkei succumbed to a millennial movement, which promised that their land would be magically restored if they killed their cattle and destroyed their crops. J. B. Peires estimates that about 40,000 people died as a result, mostly by starvation. [3] The Zulu War of 1879 saw 1,326 British soldiers wounded and killed. [4] The Zulu kingdom could muster about 40,000 warriors in the field; [5] of these many must have been killed, but it is unlikely that mortality could have exceeded 15,000 in total.

Worse was the Anglo-Boer War of 1899-1902, in which 7,800 British soldiers were killed or died of wounds and a further 14,000 of disease. Some 4,000 Boer men in the field died; more notoriously 27,000 Boer refugees died of disease in concentration camps. About 14,000 blacks died in similar camps. The number of Boer civilian casualties outside the camps is not known. [6] In World War I, 12,272 South Africans were killed or died in military service, and in World War II, 5,962 members of the Union Defense Force were killed in action, died of wounds, died in prisoner-of-war camps, or were accidentally killed outside South Africa. [7]

More lethal than any war--in fact, almost as lethal as all of them put together--was the influenza epidemic that reached South Africa in late 1918. Over 40 percent of the population contracted it. Within a few weeks, close to 140,000 people had died--more than 2 percent of the population.

No attempt has been made to estimate the effect of apartheid on mortality or on population size. To do so would require the construction of an elaborate counterfactual theory, which would always be controversial. And the effects would have to be interpreted in light of the fact that the population grew by more than 2.7 percent per annum between 1946 and 1970 and by more than 2.3 percent per annum between 1970 and 1994, the period when black fertility started to drop. [8] Estimates of the casualties of the political transition are available: the South African Institute of Race Relations has estimated the number of deaths from political violence between September of 1984 and December of 1994 at 21,651, the worst years being from 1990 to 1994. [9]

It must seem a cruel irony that, just as South Africa emerges from nearly a century of segregation and apartheid into a universal suffrage democracy, the downward trend in mortality has been arrested and reversed. The reason for this is the HIV/ AIDS epidemic; the United Nations believes that South Africa has 4.2 million infected people, more than any other country in the world. It estimates AIDS deaths in 1999 alone at 250,000. [10] The estimate is probably on the high side, but not of the wrong order of magnitude. Every indication is that mortality will become much worse during the next decade.

Like the influenza virus, HIV is not in itself a consequence of poverty. Unlike the influenza virus, which generally killed within a few days or weeks of infection, HIV works by weakening the immune system, rendering the carrier vulnerable to a variety of infectious diseases. The initial strength of the immune system is a function of nutritional and health status. The rate of weakening is also a function of the ability of the health-care system to make stabilizing interventions. The period from infection to death is therefore dependent on social, economic, and political variables. So is the rate of infection. One variable that will affect it is the mean age at marriage, which influences the length of premarital sexual activity. Another is the degree of general education and specific HIV/AIDS knowledge. The delineation of the social and economic variables affecting mortality is the central theme of this study. Understanding them will be crucial for an effective HIV/AIDS policy.

The first substantial modern analysis of episodes of rising mortality was Thomas Robert Malthus's Essay on the Principle of Population as it Affects the Future Improvement of Society (first edition, 1798, second and expanded edition, 1803), which exerts a continuing influence in the fields of demography and population economics. [11] A brief survey of the major tenets of Malthusianism will help to identify the weaknesses in social and economic structure rendering South Africa so vulnerable to the epidemic.


Malthus's central concern was with the tendency of a population to outstrip the resources necessary to sustain it. He identified two "checks" to population size and growth in this situation:

* the consciously chosen preventative check of late marriage, which would result in a reduction of fertility, and

* unintended positive checks that involve an increase in mortality. Positive checks were subdivided into:

* the relatively rare ultimate check of famine, and

* more common immediate checks, consisting of all customs [12] and diseases generated by a scarcity of food and all the causes independent of scarcity that weaken and destroy the human frame.

By including positive checks independent of the scarcity of foot, Malthus made them consist of more than just an "error correction mechanism." They also contain shocks independent of the population-food resource balance. The Black Death, which ravaged Europe between 1347 and 1352 and reduced the population by a third, was one such event. In the half century before the Black Death, there seems to have a slowdown in economic and population growth; the plague supervened and ver-adjusted the population downward, with the result that wages and agricultural incomes per capita rose sharply in the late fourteenth century. The pre-crisis population level was not again reached until the mid-sixteenth century.

Neo-Malthusianism applies Malthusian ideas to modern theories of economic development. It suggests the following reasons that rapid population growth undermines growth in real income per capita:

* the stock of physical capital (land, buildings, equipment, etc.) per worker declines as population rises, and with it per capita output;

* it is easier to improve educational attainment and quality in societies with low rates of population growth, since there are fewer children and young adults to be taught. High fertility means low average levels of human capital;

* pressure on natural resources increases with the rate of population growth;

* public spending on physical infrastructural investments gets crowded Out by high population growth and consequent pressure on social services;

* rapid population growth biases household allocation decisions toward consumption and away from savings because of the pressure of meeting basic needs; and

* there are no economies of scale arising from population increase. [13]

There is, of course, a debate between those optimistic or neutral about rapid population growth and Malthusians. Ester Boserup has argued that, in agricultural societies, population pressure is the cause, rather than the consequence, of agricultural innovation. Population density increase leads to greater intensity of land use; population decline favors a return to less intensive methods. Equally, one may point to theories and facts of industrially based growth that do not conform to the neo-Malthusian postulates. Three examples of counters to the neo-Malthusian view can be cited. At the theoretical level, it has been argued that high population density facilitates adoption of "green revolution technology" by making infrastructure development and technology transfer easier. It has also been suggested that under circumstances of high fertility adult labor intensity increases so that the impact of fertility on savings is neutralized. At the factual level, the population of the four leading Western nations (Bri tain, France, Germany, and the United States) grew by a factor of 5.5 between 1820 and 1987, whereas their combined gross domestic product in constant prices multiplied by 93. [14]

The prediction from neo-Malthusian theories is that there should be an inverse relationship between population growth and growth of income per capita. Cross-sectional data for recent decades do nor show such a relationship in poor countries. But this finding does not rule out the general operation of Malthusian mechanisms in the future. Nor does it rule out their operation in particular countries over a period of time.

This brief survey of Maltusian themes establishes a checklist of aspects of social and economic issues that need investigation if the vulnerability of South Africa to rising mortality is to be assessed.

It should be noted that the consequences of a Malthusian episode reach beyond rising morbidity and mortality. A recent influential study on infectious disease suggests that infant mortality correlates strongly with political instability, particularly in countries that have already achieved a measure of democracy. [15]


As background to the South African situation, the sub-Saharan African context can be sketched briefly. Of the sixteen countries of southern, central, and east Africa, excluding South Africa, only five had life expectancies at birth of more than fifty years in 1997. Five had life expectancies of less than forty-five years. Seven countries saw life expectancy drop between 1970 and 1997. [16] Ten countries had total fertility rates of 5.4 or more in 1998, indicating that the fertility transition had not yet begun. Eight countries had a GDP per capita of U.S. $1,000. Three countries had a drop in real GDP per capita in 1998 and a further four had rises of less than 0.5 percent. Real per capita incomes dropped in sub-Saharan Africa as a whole between 1970 and 1995. [17] Six countries had wars or major rebellions in progress, or have had them since 1980. [18] Nine countries were regarded as "not free" by Freedom House (a nonpartisan organization working to advance political and economic freedom worldwide) and the remaining seven as "partly free." [19]

The picture that emerges from these data is one of a general vulnerability to Malthusian positive checks. Episodes of demographic retrogression occur in response to political deterioration, adverse weather, commodity-price developments (sub-Saharan African economies have large agricultural sectors with high dependence on commodity prices), or epidemics of infectious diseases. Sub-Saharan Africa accounts for nearly half of infectious disease-caused deaths worldwide; 65 percent of all deaths in the region are caused by such diseases. [20]

Under these circumstances, the status of a country's healthcare system matters a great deal. The CIA World Fact Book ranks health systems from one (the most advanced) to five (the least developed). Category four countries concentrate epidemiological surveillance, response, and prevention in the capital, and characteristically offer health care to 40-50 percent of the population. Health systems in category five countries rely on humanitarian assistance and characteristically offer health care to only 40 percent of the population. One of the sixteen sub-Saharan African countries has a health system in category three, nine are in category four, and six are in category five. Five have improving health-care systems and three have deteriorating systems. Health systems in these countries are in a poor position to deal with infectious diseases. Only in a minority of countries are there signs of progress.

The CIA World Fact Book assesses the South African health system as being in category two (i.e., as having a national system of epidemiological surveillance, response, and prevention throughout most of the country, offering medical care to 70-90 percent of the population, having established primary, secondary, and tertiary health-care capability, and making pharmaceuticals available to the population generally). It also regards the South African system as deteriorating, a claim that would be controversial among South African health-care experts. Certainly the public system will be greatly stretched as AIDS morbidity rises. South Africa has a substantial private health-care sector, buttressed by medical aid schemes. This sector can make a contribution to reducing AIDS mortality, especially if the cost of AIDS pharmaceuticals falls.


The 1904 Population Census put South Africa's population at 5,174,000. The 1996 Census estimated the population at 40,584,000, implying a nearly eightfold increase in just under a century. The national average population density in 1996 was thirty-three persons per square kilometer, relatively low by international standards. The average rate of population growth from 1904 to 1996 was 2.26 percent per annum.

South Africa's fertility transition has been unusually stretched out because of the heterogeneity of the population. The white fertility transition had started by the beginning of the century, the Asian fertility transition started in the 1950s, the colored in the 1960s, and the black transition in the 1970s. The transitions for the three minority groups are almost over; the black total fertility rate remains above three. In the absence of AIDS mortality, the population could have been expected to peak eventually at over 60 million; J. L. Sadie's 1993 projection for 2011 was 54.1 million with an annual growth rate of 1.39 percent at that date. [21] The post-transition to pre-transition population ratio would have been twelve or more--double the post-transition to pre-transition ratios in the four leading Western countries.

HIV/AIDS is changing these expectations rapidly. The data available on HIV infection comes from tests of pregnant women attending prenatal clinics in public health facilities. The series starts in 1990, when 0.7 percent of such women were found to be HIV-positive. By 1993, the infection rate had risen to 4 percent and by 1996 to 14.2 percent. In 1998 it was 22.8 percent, and in 1999 it was 22.4 percent. Infection among these women is more prevalent in the northern than in the southern provinces, the 1999 rate in KwaZulu-Natal being 32.5 percent and in the Western Cape 7.2 percent. The unevenness in provincial prevalence rates suggests that the national rate will rise further as the epidemic spreads southward. Like the United Nations, the Department of Health believes that one in every ten South Africans is infected with HIV (among all adult men and women the estimated rate of infection is 19.9 percent). [22] Mortality will be increasingly affected by AIDS deaths, but since the data are published about four y ears in arrears, a substantial AIDS effect has not yet shown up in official statistics. Not all deaths caused by AIDS will be recorded as such; inferences about AIDS mortality will have to be made from trends in total mortality.

The Actuarial Society of South Africa has developed a model of the spread of the epidemic. It divides the population into four groups (sex workers, those infected with sexually transmitted diseases, others at risk, and those not at risk). It sets up a contagion matrix to represent cross-group risks of infection and assumes differential infectivity rates by sex and mean terms from infection to death. The model has been calibrated against the limited data available. Robert Dorrington observes that:

such an exercise is by its nature perhaps inevitably a little more art than science but briefly the aim was to set, where possible, the assumptions to be consistent with empirical studies and where this was not possible to set the assumptions by trial and error (within bounds of reasonableness) so that the output from the model reproduced the observations of the epidemic. [23]

The model projects that the population will peak at fifty million in 2011 before it starts to fall. By 2008, the model projects that the period probability of someone exactly fifteen years old dying before age sixty will be over 80 percent for both men and women. [24]

An account of national population dynamics needs to be supplemented by a description of the distribution of the population across space and by settlement type. The 1996 Census put the urban population at 53.7 percent of the whole. Formal urban areas were inhabited by 18.2 million, and another 3.5 million lived in informal urban areas. Most of the rural population (14.8 million) lived in tribal areas; a further 2.9 million lived on commercial farms, and 1.1 million lived in other rural areas (which are more like tribal areas than commercial farms). Apartheid had two effects on this distribution: it retarded urbanization, particularly from the mid-1950s to the mid-1970s, and it affected the distribution of the rural population between tribal areas and commercial farms through a number of policies, including resettlement.

Population densities, already high in some of the reserves at the time of the 1913 Land Act, have risen throughout the century. Since the 1960s there has been a rise in informal urbanization in the tribal areas. Many studies have shown that there is great dependence in tribal areas on remittances from relatives in town and on transfers from the state. What there has not been to date is a Boserupian agricultural response to rising population densities. Half of South Africa is not suited to intensive agricultural development, since it has an average annual rainfall of less than five hundred millimeters (less than twenty inches). The greater part of the tribal area population is to the east (the wetter side) of that isohyet, and there other explanations must be sought. Two candidates deserve particular mention. First, the tribal agricultural tradition is pastoral, and there has been a gender division of roles, with women assigned to arable work while men deal with animals. Second, for many, the returns on indust rial work have been much higher than those on farming activities. All this could change if nonagricultural work continues to become scarcer and land reallocation proceeds either through the land reform program or through the market.

Mortality varies between the four main settlement types. Figure 1 shows the number of children still living as a fraction of the number of children ever born by age of mother in urban formal, urban informal, commercial farm, and rural tribal areas in 1996. This fraction is an indicator of infant and child mortality. The fraction is highest in urban formal areas, then in urban informal and commercial farming areas, and is lowest in tribal rural areas. Figure 2 shows the proportion of males between ten and forty-nine whose father is still alive. Here the fraction is highest in urban formal and commercial farming areas and is lower in urban informal and rural tribal areas. This fraction is an indicator of adult male mortality. The proportion of females between ten and forty-nine whose mother is still alive is generally a good deal higher, indicating that adult mortality is significantly lower for women than for men. The differentials in mortality between settlement type are very small for adult women.


South Africa's national accounts date from 1946. Between then and 1981, GDP per capita in 1995 prices rose from R 8,290 to R 16,437, just about doubling in thirty-five years. As figure 3 shows, the trend since 1981 has been downward, with 1999 GDP per capita at R 13,847, a 16 percent drop from the peak. GDP per capita in 1999 was virtually the same as in 1994, so the first ANC government presided over a stabilization but not an improvement in average living standards.

Figure 4 shows that gross fixed capital formation as a percentage of GDP rarely dropped below 20 percent between 1946 and 1985 and was generally above 25 percent between 1964 and 1981. Since 1990, gross saving and gross fixed capital formation have been between 15 percent and 20 percent, without the capital inflows that sustained the difference between savings and investment between 1946 and 1959. Indeed, savings exceeded investment between 1985 and 1994; as a consequence of financial sanctions, South Africa was forced to reduce its foreign debt in relation to GDP. Unlike GDP per capita, savings and investment as a percentage of GDP have not stabilized; they have both dropped between 1994 and 1999.

Table 2 sets Out the components of gross savings and investment in relation to GDP at market prices.

Net household savings have dropped steadily, and net corporate saving has dropped between 1996 and 1999. These drops have been offset by a decline in government dissaving, the rate of which will not drop as fast in the future. Table 2 confirms that net foreign investment is small in relation to gross saving and gross capital formation.

Neo-Malthusian theory predicts that high fertility puts downward pressure on savings. This relationship was tested on data from the 1995 Income and Expenditure Survey as linked to the 1995 October Household Survey. The linked surveys contain a national sample of 28,585 households. Household consumption was graphed against permanent income, from which was excluded retirement lump sums, annuity lump sums, state compensation lump sums, life insurance payouts, inheritances, funeral fund payouts, insurance payouts in respect to fixed property claims, accident claims, other gratuities, and lobola [25] (bride wealth) received. Lowess smoothing (locally weighted scatterplot smoothing in which a separately weighted regression is estimated for every point in the data) was used to extract the relationship between household consumption and permanent income. This yielded a close approximation to a straight line, with an autonomous consumption of R 1,366 and a marginal propensity to consume of 0.919 for all households wit h a permanent income of up to R 1,000,000 per annum. The residuals (actual consumption less predicted consumption as a proportion of predicted consumption) were then examined for a relationship between (a) total household size and (b) the number of children under twenty years of age. The predicted relationship was found. Each child increased consumption by 0.66 percent (the coefficient was significant at the 1 percent level) and each household member increased consumption by 0.26 percent (the coefficient was significant at the 10 percent level).

Foreign investment is split into three categories in the national accounts:

* direct investment includes transactions related to the acquisition of share capital in foreign countries by establishing new businesses, or through mergers and takeovers. The equity threshold used to determine a direct investment relationship between .a direct investor and an enterprise is 10 percent. Loan capital is excluded, unless it represents permanent debt.

* portfolio investment consists of international equity and debt securities not classified as direct investment. Money market debt instruments and tradable financial derivatives are included.

* other investment comprises trade credits, loans, currency and deposits, and other assets and liabilities.

Net direct investment has been negative in six of the eight years between 1992 and 1999. It was positive in 1997 and 1999; in other years, South African direct investment abroad exceeded foreign direct investment in South Africa. Net portfolio investment has been positive throughout. It was small up until 1996. In 1997 it was R 30.6 billion, in 1998 R 20.4 billion, and in 1999 R 51.0 billion. Net other investment was negative in five of the eight years, turning positive in 1995, 1996, and 1998. Portfolio investment contains the potentially volatile "hot money" flows.

The financing of growth in 1990s South Africa appears precarious: the domestic savings rate is low, and foreign funding augmentation is relatively small and easily withdrawn. A complementary issue then arises: how productively are South Africa's investable funds and labor used? Anthony John Wright, Murray Leibbrandt, and R. T. Bell have carried out a careful study of total factor productivity growth in manufacturing, concluding that average annual TFP growth in manufacturing from 1954-1963 was 1.0 percent. From 1963-1 974 it was 0.3 percent, from 1974-1981 it was 0.3 percent, and from 1981-1990 it was -0.5 percent. They conclude:

The extent of this recent slowdown and its pervasiveness across industries outside the capital intensive category is an alarming feature of the productivity growth performance of South African manufacturing. ... Available international evidence suggests that, for manufacturing as a whole, South Africa's TFP performance has been relatively poor. [26]

There are some subsectors within manufacturing for which this is not true. But a general trend toward increasing the efficiency with which capital and labor are used has not been in evidence in the last two decades.

Figure S shows the evolution since 1980 of the capital/labor ratio in four predominantly private-sector industries: 1) mining, 2) manufacturing, 3) construction, and 4) wholesale trade, retail trade, catering, and accommodation. In constant 1995 prices, the ratio rose by over 80 percent between 1980 and 1999, with most of the rise occurring in the second decade. Employment dropped by 15 percent over the period. This development occurred during a period of high and probably rising unemployment. The neo-Malthusian prediction about the capital/labor ratio is that it will be depressed by high fertility; implicit in the prediction is the assumption of an efficient labor market. The South African labor market, instead of spreading the capital stock over all available workers, spreads it over only some of them and assigns the rest to unemployment. Unemployment is selective; it varies across space, by education, and by age and experience, as will be shown below.

Neo-Malthusian theory predicts the crowding out of public infrastructural investment by expenditure on social services. Figure 6 shows that this effect has been present in the South African economy since the mid-1980s; democratization has turned a latent tendency into a manifest one. The Medium Term Expenditure Framework, which projects national government expenditure for three years, envisages a slight downturn in the percentage of government spending accounted for by social services, but no upturn in the economic services percentage.

In summary, several neo-Malthusian factors are present in the form of the precariousness of the current level of real GDP per capita, the low level of savings and investment, the pressure of large households on savings, generally poor productivity trends, rising unemployment, and pressure on public funds available for physical infrastructure. Conditions in the labor market and South Africa's human capital are further causes for concern.


The October Household Surveys from 1994 to 1998 and the 1996 Population Census all attest to very high rates of unemployment in contemporary South Africa. This has much to do with the spatial distribution of the population in relation to centers of employment, which in turn have been greatly affected by segregation and apartheid. Table 3 shows the variation of unemployment across type of settlement.

The relatively low rate of unemployment on commercial farms reflects generally tight control over the population on those farms where residence is tied to employment, despite recent legislation designed to improve security of tenure among farm workers. For the rest, the picture is as one would expect: higher unemployment in informal urban areas than in formal, and very high unemployment in the rural tribal areas.

Increasingly it has been recognized that the accumulation of physical capital (land improvements, machinery, and the like) is not the only determinant of growth prospects. Human capital matters a great deal; indeed, societies recover more quickly from the destruction of physical capital than from the loss of human capital. The United Nations Human Development Report publishes a human development index annually. This index is constructed from life expectancy, the adult literacy rate, a combined primary, secondary, and tertiary education gross enrollment ratio, and real GDP per capita in purchasing power parity U.S. dollars. [27] In 1999, South Africa's human development index was 0.695, placing it 101st in the world. On the basis of GDP per capita in PPP$, South Africa is placed fifty-fourth, so the PPP$ rank minus the HDI rank is -47, indicating that South Africa's human capital is seriously underdeveloped in relation to its level of GDP per capita. Among ninety-four medium human development countries, only Oman, Botswana, and Gabon have greater negative discrepancies than South Africa.

What pulls South Africa down on the HDI is its low life expectancy. The combined educational enrollment index is relatively high, as is the proportion of GDP spent on education (7.6 percent in 1999). [28] The country's high expenditure on education is not in doubt. But expenditure is the input to the formation of human capital. How efficiently is it used?

South African secondary-school enrollments have risen rapidly in the last two decades. This is reflected in a rise in the number of candidates for the senior certificate, a public examination written after twelve years of schooling. In 1979, there were 85,276 candidates; by 1998, the number had risen to 552,862. [29] The average annual rate of increase of candidates over the period was 10.3 percent. But the number of passes has not risen as fast; the average annual growth rate over the period was 7.1 percent for passes as a whole and 4.1 percent for university exemption (entrance) passes. The number of university exemption passes rose from 68,700 in 1988 to 88,497 in 1994, but dropped back to 69,861 in 1998. In that year, it took an average of 189 school years to produce one senior certificate pass at the university entrance level. Figure 7 shows that the pass rate has dropped from 87 percent in 1979 to 49 percent in 1998 and that the proportion of candidates who qualify for matriculation exemption (i.e., wh o may go to university) has dropped from 38 percent to 13 percent.

The South African higher education system also has its inefficiencies:

* since the 1950s, the ratio of degrees conferred by universities to university enrollments has fluctuated between 10 and 15 percent--this in a system where the average undergraduate curriculum is between three and four years in length, and where the average postgraduate curriculum is about two years;

* the number of dropouts each year from universities and technikons (technical universities) slightly exceeds the number of graduates;

* from the mid-1980s, real expenditure per degree has shown a marked upward trend;

* the proportion of degrees awarded in the mathematical, natural, and engineering sciences has dropped from close to 25 percent in the mid-1960s to little over 15 percent in the 1990s;

* of the seventeen universities for which Department of Education publication units per lecturer could be calculated in 1989 and 1994, eleven showed a decline, five an improvement, and one remained constant;

* whereas colored, Asian, and black university enrollments have risen rapidly in universities over the last three decades, the entrance of these groups in substantial number into technikons has only come about in the 1990s; and

* apprenticeship training has been in decline since the mid-1980s. [30]

The inefficiencies of the South African secondary and higher education systems are particularly problematic in light of the severe skills shortages in the labor market. The Human Sciences Research Council reviewed trends in the occupational distribution of formal employment from the early 1970s and concluded that the higher the skill of an occupational category the faster demand for it has grown. [31] Interviews with business and professional associations confirmed that this trend would continue.

This picture is further confirmed by the fact that unemployment rates vary dramatically by highest education level, as Table 4, taken from the 1996 Population Census, indicates:

At the end of the twentieth century, the South African educational system can be described as a low-quality mass system. Over 90 percent of children aged between seven and sixteen are enrolled in school, though many of them are not spending anything like the full school year engaged in productive learning. Higher education enrollments are close to 20 percent of the twenty to twenty-four-year-old age cohort, but a great many students leave the system without completing a qualification. The consequence is a severe shortage of skilled people in the economy, especially in fields where mathematics is required. This reduces efficiency and drastically limits development options.

Skilled-labor shortages could be alleviated by liberalization of the immigration policy for people with higher education. The policy of the Department of Home Affairs is slowly and some-what reluctantly moving in this direction. Politically, it is not easy in light of high domestic unemployment.


Over two-thirds of the unemployed have not worked before. The social consequences can be traced through figure 8, which constructs a period analysis on the basis of information in the 1997 October Household Survey. At twenty, more than 50 percent of men remain in the educational system and at twenty-four, more than 20 percent. This reflects enrollment at high ages in secondary school as well as enrollment in secondary education. Young men leaving school are exposed to a lengthy period of searching for a job; at age twenty-five, about 25 percent of the cohort are unemployed, never having had a job before. At age thirty, just over 20 percent of the cohort remain in this position. By age twenty-five, about 40 percent of the cohort has become employed and by age thirty, about 60 percent. The proportion of the cohort ever married rises with employment, with a lag. It is employment that makes young men marriageable; late exit from the educational system and a lengthy period of unemployment before the first job make s marriage late. The marriage prospects of women do not depend on employment; they depend on the number of available men. The singulate mean age at marriage for South African men in 1997 was 31.5 years; that for women 28.9.

The consequences are these:

* There is considerable premarital fertility. Whites and Asians have low levels of premarital fertility--and marriage is earlier among these more affluent groups--but coloreds and blacks have premarital fertility rates that are 40-50 percent of the marital fertility rates at corresponding ages. Malthus would have approved of the late marriage age, but the key factor turning this fully into a preventative check--abstinence, or at least zero fertility--is largely absent in South Africa. Nonetheless, since premarital fertility is lower than marital fertility, the rising age at marriage has played a minor role in reducing South African fertility.

* Long sexually active premarital time stretches increase the risks of STD and HIV infection.

* A great many households have to support young adults for much longer periods than in societies that have close to full employment.

* Among Africans, the incidence of two parent nuclear households is low. The expected length of an adult's working life is short; there is late entry into employment. Mortality and morbidity rates become substantial for South Africans in their forties and beyond, with adverse consequences for incomes. Over 13 percent of women remain unmarried at age fifty, and there are many two-and three-generation female-headed households. A substantial number of children live with their grandparents off an old-age pension; AIDS mortality of younger adults will increase the number of households of this kind.


Many Malthusian and neo-Malthusian conditions can be found in contemporary South Africa. The population has increased rapidly throughout the twentieth century. Real income per capita has declined over the last twenty years; more recently it has stabilized, but the stability is precarious. The savings rate has also declined, and net savings on the part of households have almost disappeared. Recent net foreign inflows on the capital account have added only marginally to savings as a source of funds for investment. There is some pressure on savings in higher-fertility households. Rather than capital per worker being diluted uniformly by a rapidly growing labor force, it is spread over only part of it, unemployment being the fate of the remainder. South Africa's education system, despite being expensive, is inefficient and unable to supply adequate numbers of skilled and highly educated people to the labor market. Pressures for social spending have crowded out public investment in economic infrastructure. Despite rising population densities in rural tribal areas, there has been little technical change in agricultural production. Moreover, there is a candidate for a positive check in the form of an HIV epidemic, which has spread rapidly during the 1990s.

Everything we know suggests that South African mortality will rise significantly in the early twenty-first century. The incidence will vary across population groups, being lowest among whites and Asians, higher among coloreds, and highest among blacks. Better-educated South Africans will be less affected by the HIV epidemic for five reasons: they will be more informed about the means of avoiding infection, they will have stronger immune systems, they will use tests more, they will be more likely to have the income (and medical aid coverage) for the drugs to manage the infection, and they are more likely to stick to medication programs. But this is a matter of degree only, and AIDS will take a substantial toll on South Africa's already scarce human capital.

The extent of the rise in mortality will be affected by factors under the control of South Africans and others: international medical advances in HIV/AIDS management, reduction in the price of relevant pharmaceuticals, improvements in the efficiency of the South African private and public health systems, efforts to relieve the worst of South African (largely rural) poverty, improvements in the employment prospects of the young, greater efficiency in the education system.

Two related issues have been identified as crucial in recent debate: appropriate political leadership and much greater capacity and seriousness in the public health system. Helen Epstein has published an alarming chronicle of her investigations into a clinical trial of anti-retroviral drugs that raises major questions about the administration and capacity of the public health system, [32] and Edwin Cameron, in his Jonathan Mann Memorial Lecture delivered at the thirteenth International AIDS Conference in Durban, complained of a "cacophony of task groups, workshops, committees, councils, policies, drafts, proposals, statements and pledges. But all have thus far signified piteously little." [33] Similarly, Epstein concludes that South Africa has a ruler who has not adequately confronted the most deadly threat facing his country, and Cameron observes that "perhaps the most intractably puzzling episode has been our President's flirtation with those who in the face of all reason and evidence have sought to disput e the aetiology of AIDS." [34]

Without rapid clarification of goals and a special effort to improve the focus and capacity of the Department of Health, there can be little improvement in a so far poor effort to deal with a national emergency.


The idea for this study came to me while reading Professor J. L. Sadie's retirement Ph.D. thesis submitted to the Stellenbosch University in 1999. Sadie's thesis is not self-consciously Malthusian, but it amounts to the statement of a Malthusian case in relation to contemporary South Africa. While our arguments overlap in some respects, they also differ substantially in scope and emphasis. Thanks are also due to Peter Perkins and Ingrid Woolard for assisting with parts of the statistical analysis.

Charles Simkins is Helen Suzman Professor of Political Economy at the University of the Witwatersrand, Johannesburg.


(1.) "Malthus" in Encyclopaedia Britannica, 9th ed., vol. XV (1883): 344.

(2.) Charles Simkins and Elizabeth van Heyningen, "Fertility, Mortality and Migration in the Gape Colony, 1891-1904," International Journal of African Historical Studies 22 (1) (1989): Table 1.

(3.) J. B. Peires, The Dead Will Arise: Nongqawuse and the Great Xhosa Cattle-Killing Movement of 1856-7 (Johannesburg: Ravan Press, 1989), 319.

(4.) David Clammer, The Zulu War (Devon, U.K.: Newton Abbot, David and Charles, 1973), 216.

(5.) John Laband and Paul Thompson, Field Guide to the War in Zululand and the Defence of Natal, 1879, 2d ed. (Pierermaritzburg: University of Natal Press, 1983).

(6.) Brian Johnson Barker, A Concise Dictionary of the Boer War (Cape Town: Francolin Publishers, 1999), 45.

(7.) D. J. Potgieter et al., eds., Standard Encyclopaedia of Southern Africa, vol. 11 (Cape Town: NASOU, Nasionale Boekhandel, 1975), 512, 525.

(8.) Statistics South Africa, 1995.

(9.) South African Institute of Race Relations, South Africa Survey 1995/1996 (Johannesburg: South African Institute of Race Relations, 1996), 52.

(10.) UNAIDS, Report on the Global HIV/AIDS Epidemic, June 2000, 9 and Annex 2; [less than][greater than].

(11.) Thomas Robert Malthus, An Essay on the Principle of Population as it Affects the Future Improvement of Society (London: J. Johnson, 1798; 2d and exp. ed. 1803).

(12.) "In all groups except the Nguni it is not only justifiable but also customary to kill one or both of twins, or children born feet first, or cutting their upper teeth first, or presenting some other abnormality. Such children are regarded as evil omens who must be put out of the way as soon as possible lest they bring disaster on their families." Isaac Schapera, The Bantu-Speaking Tribes of South Africa (Cape Twon: Maskew Miller, 1937), 210. Custom also prohibited sexual intercourse between the birth and weaning of a child. Killing one or both twins or handicapped children and birth spacing all served to make the survival of parents and siblings more likely in a time of famine. These customs have eroded to the point of extinction in the twentieth century as a result of urbanization and the spread of formal education.

(13.) Massimo Livi-Bacci, A Concise History of World Population, 2d ed., trans. Carl Ipsen (Maiden, Mass.: Blackwell, 1997), 206-211.

(14.) Ibid., 146-147.

(15.) Ibid.

(16.) United Nations Development Program, Human Development Report, 1999 ed. (New York: United Nations Development Program, 1999), Table 8. The sixteen countries are: Lesotho, Swaziland, Botswana, Namibia, Zimbabwe, Zambia, Malawi, Mozambique, Angola, Kenya, Uganda, Tanzania, Rwanda, Burundi, and both Congos.

(17.) Central Intelligence Agency, The Global Infectious Disease Threat and its Implications for the United States (Washington, D.C.: CIA, G.P.O., January 2000), 21.

(18.) Central Intelligence Agency, The World Fact Book, 1999 ed. (Washington, D.C.: CIA, G.P.O, 1999).

(19.) Freedom House, Annual Survey of Freedom Country Ratings 1972-73 to 1999-00 (New York: Freedom House, 1999). Freedom House constructs a political liberties index ranging from one (the most free) to seven (the least free). It also constructs a civil liberties index on the same basis. "Not free" countries have a combined index of at least eleven; "partly free" countries have a combined index of between five and ten.

(20.) Central Intelligence Agency, The Global Infectious Disease Threat, 12.

(21.) J. L. Sadie, A Projection of the South African Population, 1991-2011, Bureau of Market Research Report 196, University of South Africa, Pretoria, 1993, Appendix A.

(22.) South Africa Department of Health, National HIV Sero-prevalence Survey of Women Attending Public Ante-natal Clinics in South Africa, 1999 (Pretoria: Department of Health, 2000), Figure 1.

(23.) Robert E. Dorrington, ASSA600: An AIDS Model of the Third Kind, [less than][greater than], 2000.

(24.) Ibid., Table A8.3.

(25.) Isaac Schapera, Migrant Labour and Tribal Life (New York: Oxford University Press, 1947), 143, gives lobola as the Nguni term for bride wealth and bogadi as the Sotho term.

(26.) Anthony John Wright, Murray Leibbrandr, and R. T. Bell, "Productivity Growth in the South African Manufacturing Sector: Deriving and Assessing Neoclassical TFP Measures," in K. Kruger, ed., Papers Submitted for the Competition to Promote Academic Research into Industrial Policy and Development in South Africa, vol. 3A (Sandton, South Africa: Industrial Development Corporation of South Africa, November 1993), 29-30.

(27.) Conversion of local currencies to U.S. dollar equivalents can be done in one of two ways. The easier way is to use exchange rates. Purchasing power parity conversion establishes the number of units of the local currency that would buy the same basket of goods that one dollar would buy in the United States.

(28.) UN Human Development Report, 1999, Tables 1 and 10.

(29.) South African Institute of Race Relations, South Africa Survey 1999/2000 (Johannesburg: South African Institute of Race Relations, 2000), 113.

(30.) Johannes Fedderke, Raphael de Kadt, and John Luiz, "A Capstone Tertiary Education System: Inefficiency, Duplication and Inequity in South Africa's Tertiary Education System, 1910-93," Econometric Research Southern Africa Working Paper no. 14, University of the Witwatersrand, Johannesburg, April 2000; South Africa's National Commission for Higher Education, Report of the Commission (Pretoria: National Commission for Higher Education, 1996).

(31.) Human Sciences Research Council, Skills Needs of the South African Labour Market, 1998-2003 (Pretoria: Human Sciences Research Council, 1999), 9.

(32.) Helen Epstein, "The Mystery of AIDS in South Africa," New York Review of Books 47 (12) (20 July 2000): 50-55.

(33.) Edwin Cameron, "The Deafening Silence of AIDS," first Jonathan Mann Memorial Lecture, thirteenth International AIDS Conference, Durban, 9-14 July 2000, 3.

(34.) Ibid., 4.

                  Life expectancies at birth: 1936-1941,
                           1970-1975, 1990-1995
         1936-1941        1970-1975        1990-1995
           Male    Female   Male    Female   Male    Female
Blacks      9.6     40.5    56.8     63.4    59.8     67.9
Coloreds   43.2     43.6    55.8     62.6    59.2     64.9
Asians     52.7     50.9    60.0     65.1    64.7     71.8
Whites     58.6     62.5    65.2     72.8    69.7     76.8

Source for 1936-1941 statistics: J. L. Sadie, A Reconstruction and Projection of Demographic Movements in the RSA and TBVC Countries, Bureau of Market Research Report 148, University of South Africa, Pretoria, 1988.

Source for 1970-1975 and 1990-1995 statistics: Charles Simkins, "Reconstructing South African Demographic History 1970-1995," paper presented to the 1996 Census Phase II evaluation workshop, Pretoria, December 1999.

                      Savings as a percentage of GDP
                             at market prices
                              1992  1996  1999
Saving by households          3.64  1.15  0.49
Corporate saving              5.57  7.02  4.12
Saving by government         -7.49 -5.06 -2.66
Net saving                    1.72  3.11  1.95
Consumption of fixed capital 14.90 13.04 13.21
Gross saving                 16.62 16.15 15.16
Net foreign investment       -1.52  1.34  0.36
Gross capital formation      15.10 17.49 15.52
Source: South African Reserve Bank,
Quarterly Bulletin (Pretoria) (March
2000): Tables S-126 and S-127.
                   Unemployment by settlement type, 1996
                 Male (%) Female (%)
Urban formal       20.4      30.3
Urban informal     35.8      58.5
Commercial farms    6.1      17.4
Tribal rural       54.2      67.9
South Africa       27.1      41.9
Source: 1996 Population Census;
10 percent sample.
                       Unemployment rates by sex and
                       highest education level, 1996
Highest education level               Male (%) Female (%)
Complete primary or less              34.1     51.4
Incomplete secondary                  29.1     47.0
Senior Certificate                    23.1     34.9
Less than SC plus certificate/diploma 10.9     18.4
SC plus certificate                    7.4     15.3
SC plus diploma                        5.3      7.0
SC plus degree                         2.6      3.6
South Africa                          27.1     41.9
Source: Statistics South Africa, 1996
Population Census, tabulation
from electronic data.

Source Citation

Source Citation   

Gale Document Number: GALE|A71902195