8HIV/AIDS and Older People in South Africa

Hosegood V, Timaeus IM.

Publication Details


South Africa is experiencing a rapidly growing and severe HIV/AIDS epidemic. National antenatal clinic data show a rise in seroprevalence from 1 percent in 1990 to 25 percent in 2000 (Karim and Karim, 1999; South Africa Department of Health, 2005). By 2000, 40 percent of adult deaths at ages 15-49 were due to HIV/AIDS (Dorrington, Bourne, Bradshaw, Laubscher, and Timaeus, 2001). This disastrous epidemic has enormous implications for older people. They are at risk of HIV infection and AIDS mortality themselves. In addition, many more older people face the consequences of AIDS-related illness and deaths among their own children and other relatives and of the wider social and economic changes wrought by the epidemic. The rising burden of morbidity and mortality among younger adults is likely to increase the importance of the practical contributions made by older people to their households. In South Africa, these include the contribution the monthly old age pension makes to family incomes as well as older people’s role in caring for grandchildren and other children whose parents are absent (HelpAge International, 2003).

In their international review of AIDS and older people, Knodel, Watkins, and VanLandingham (2003) extensively consider the evidence concerning the sociodemographic impact of AIDS on older people in Africa (Knodel et al., 2003). This paper complements their review by focusing on South Africa. The survival of the apartheid system in South Africa long after decolonization of the rest of the continent and the economic advantages of the country mean that older South Africans live in a very different social, political, and economic environment from older people in other African countries. These distinctions, coupled with a very severe HIV epidemic, suggest that the impact of HIV/AIDS on older people in South Africa may be very different from that in other countries with generalized HIV epidemics.

We begin by describing some of the more important social, demographic, and economic aspects of older people’s lives in South Africa and then discuss the direct and indirect consequences of the HIV epidemic for them. The last section of the paper presents data on the living arrangements of older people in rural KwaZulu-Natal in South Africa and examines their households’ experience of adult mortality, in particular AIDS mortality.


The impact of the HIV epidemic on older people is shaped by the social, political, demographic, and economic circumstances in which they live. This section highlights some of the relevant characteristics of South African society.

South Africa Has the Highest Proportion of Older People in Africa

Population aging has commenced in most African countries as a consequence of the transition to lower levels of fertility and mortality. The population age 60 or more is projected to increase sixfold by 2050 (HelpAge International, 2000). The degree of population aging has been exceptionally large in South Africa primarily because of the early onset of fertility transition. Fertility has been falling since the 1960s, when total fertility was nearly seven births per woman, with the rate of decline accelerating in the early 1980s to reach total fertility of around 3.5 in 1996. South Africa’s total fertility is currently the lowest in mainland sub-Saharan Africa (Moultrie and Timaeus, 2002, 2003).

South Africa has a higher proportion of older people in its population than any other mainland sub-Saharan African country. In 1997, 7 percent of the population were age 60 or older (Kinsella and Ferreira, 1997). The demographic profiles of the different racial groups in South Africa are markedly different.1 In the African population, 6 percent were age 60 or older in 1997 in contrast to 14 percent in the white population, whose population structure closely resembles that of European countries (Kinsella and Ferreira, 1997). By 2030 the proportion of older people in South Africa is projected to increase to 11 percent (U.S. Census Bureau, 2005).

The Sex Ratio in South Africa’s Older Population Is Exceptionally Low

In all African countries, women constitute the majority of the older population, reflecting the higher mortality of men. However, the excess in men’s mortality in South Africa has always been exceptionally large (Timaeus, 1999). In 1985, the probability of dying between ages 15 and 60 in South Africa was 24 percent for women but 37 percent for men. By 2001 it was 34 percent for women and 51 percent for men (Dorrington, Moultrie, and Timaeus, 2004). Much of the excess in men’s mortality in this age range is due to the high death rate from accidents and homicides (Dorrington et al., 2001; Hosegood, Vanneste, and Timaeus, 2004). Thus, according to the 2001 census, the sex ratio at age 60 or more in South Africa was only 62 men per 100 women, compared with 85 in sub-Saharan Africa as a whole (U.S. Census Bureau, 2005).

South Africa’s Political History Has Shaped Older People’s Lives

South Africa’s history of apartheid and labor migration profoundly influences contemporary life there. The Apartheid Group Areas Act and the labor migration system systematically divided African families by recruiting younger men and women to the centers of employment, including mining, farming, and urban areas. Restrictions on the movement and settlement of those not employed meant that most children, unemployed younger adults, and older people were required to live in rural or periurban areas (Leliveld, 1997; Mazur, 1998; Spiegel, 1987). Older people facilitated the economic migration of younger adults by caring for their grandchildren and safeguarding the family land and assets. This became essential to both maintaining the labor migration system and ensuring the long-term survival of households (Izzard, 1985; Leliveld, 1997; Mazur, 1998; Murray, 1980). Consequently, in South Africa the role of older people, particularly grandmothers, in caring for children affected by HIV, builds on a long-established set of social structures related to child care (Madhavan, 2004; Van der Waal, 1996).

In traditional African societies, older men were the principal authority figures both in their households and in the wider community. In South Africa, rapid social change and the transition to democratic government are transforming family relations. Studies have highlighted increasing intergenerational tensions and adaptations of the patriarchal structures in family and community life (Campbell, 1994; Nhongo, 2004). In most countries, fertility decline and economic development are leading to an evolution of household structures away from large, extended households toward smaller households based around the conjugal unit (Bongaarts and Zimmer, 2002). In this transition, increasing numbers of older people will live alone or with a partner. Such changes in household structure do not appear to be occurring readily in South Africa. The reasons for this include very low marriage rates, low rates of cohabitation in nonmarital relationships, the dependence of children and younger adults on the economic and material support of older people, and the limited availability of land for and cost of housing. Thus, in 1996, only 6 percent of Africans were living alone (Noumbissi, Bawah, and Zuberi, 2005). In contrast, other phenomena, such as matrifocal and other women-headed households,2 have emerged and become a common household arrangement in both rural and urban areas (Preston-Whyte, 1988).

Older Africans Have Worse Health than Older People from Other Racial Groups

The massive social and economic inequalities that the apartheid system served to maintain are mirrored in enormous inequalities in health status among the different racial groups in South Africa. In 1997, life expectancy at birth was 77 years for white women, compared with 55 years for African women (Kinsella and Ferreira, 1997). The 1990-1991 Multidimensional Survey of Elderly South Africans found that older Africans living in rural areas experienced greater health and financial problems than other older people. This can be attributed to isolation, poor housing, lack of income, poor access to health care facilities, and the political and economic marginalization that resulted from apartheid policies (Ferreira, Møller, Prinsloo, and Gillis, 1992). Other studies have found high levels of self-reported depression and ill health among Africans living in urban areas (Ferreira, 2000; Gillis, Welman, Koch, and Joyi, 1991).

The cause-specific mortality profile in South Africa also differs among racial groups. Prior to the HIV epidemic, an epidemiological transition had been occurring in all racial groups. Thus, in 2000 in an almost exclusively African population in rural KwaZulu-Natal, noncommunicable diseases accounted for 76 and 71 percent of deaths of women and men, respectively, in the age group 60 or more (Hosegood et al., 2004). The proportion of deaths from heart disease is lower among Africans than other racial groups, but they suffer higher death rates from hypertension, stroke, infectious diseases, accidents, and violence (Bradshaw, Bourne, Schneider, and Sayed, 1995).

The Majority of Older People in South Africa Receive a State Pension

In South Africa, a means-tested, noncontributory state old age pension is paid in cash to women age 60 or more and men age 65 or more. This pension is relatively large—R780 a month in 2005—and had a purchasing power equivalent to about U.S. $280. Thus, in 1998 the pension was about twice the median income per head in African households (Case, 2001; Case and Deaton, 1998; Case, Hosegood, and Lund, 2005; Mohatle and de Graft Agyarko, 1999; Møller and Ferreira, 2003). The existence of a noncontributory pension scheme is a product of South Africa’s political past and its ability and political will to currently finance large-scale, public welfare schemes. The means-tested state pension was introduced in 1928 to provide for poor whites with inadequate occupational pensions. Although African workers were generally excluded from occupation pension schemes, the state pension was extended to Africans in 1944 and the value equalized for all population groups after the democratic elections in 1994 (Sagner, 2000).

Older People Use Their Income to Support Other Members of Their Household

In South Africa, older people’s pensions do not provide only for their own needs. Many older people also use this income to support the basic needs of their family, including food, clothing, and school fees for children. For many rural African households, the state pension and, to a lesser extent, other government grants are the main source of income. For example, in the Western and Eastern Cape provinces, Møller and Ferreira (2003) found that the old age pension competed with wage earnings as the most important source of income in the households they surveyed, in terms of both access and amount. In one-third of households interviewed in a study in KwaZulu-Natal, a pension was the only source of household income (Møller and Sotshongaye, 1996).

A study examining the role of pensions in poverty alleviation in South Africa found that both older men and older women spend 30-40 percent of their income on school expenses for dependents. This was in addition to other expenditures on food and household utilities (Mohatle and de Graft Agyarko, 1999). In another study in the Western Cape province, Case (2001) demonstrated that the pension had a measurable protective effect on indicators of health and self-reported well-being. In households whose members pool their income, the presence of an older recipient of a state old age pension in the household had a significant, protective effect on the health of not only the older person themselves, but also all household members. Most studies indicate that the majority of African pensioners share their income with other members of their household. As fewer than 5 percent of pensioners live alone in South Africa (Kinsella and Ferreira, 1997), pension income is a substantial source of financial support for many younger adults and children.

There are several additional social welfare grants for which older people may apply on behalf of someone else (Hunter and Rushby, 2002; Women’s Budget Initiative, 2003). In 2001 approximately 10 percent of child support grants in a rural area of KwaZulu-Natal were held by grandmothers (Case, Hosegood, and Lund, 2003Case, Hosegood, and Lund, 2005).

The Support That Older People Provide Others Affects Their Own Welfare

The state pension policy has had a profound influence on social arrangements, particularly in rural areas. It has influenced the way in which communities perceive and relate to older people, as well as the social and economic role that older people feel obliged to fulfill as income providers (Burman, 1995). Sagner (2000) argues that, even though the African pension prior to 1994 was an inadequate amount to live on, it served to ensure that many households survived economically (Sagner, 2000). While the contribution made by pensions to poorer households may increase the self-respect and social status of old age pensioners, it also makes younger people dependent on them. Several studies and government inquiries have highlighted the problem of financial and physical abuse of older household members. Both younger and older respondents report that younger people pressuring older people for money is common (Mohatle and de Graft Agyarko, 1999; South Africa Department of Social Development, 2001). However, the old age pension may have slowed the loss of social status and increasing marginalization of older people relative to younger generations that has been described in other countries undergoing substantial social change and economic development (Du Toit, 1994; Johnson, 1989).

Poverty and Older People in South Africa

The take-up of the old age pension is high, particularly in the poorest groups and at the oldest ages. Using data from the 1998 and 1999 October Household Surveys, May (2003a, 2003b) reports that 84 percent of people age 64 or older in households that were on or below a consumption-based poverty line were receiving a government pension. However, the pension is often insufficient to support the number of household members who are dependent on it. Indeed, evidence exists that the pension may attract dependents to the household, thereby reducing the value of the pension per household member (Møller and Sotshongaye, 1996). Thus, despite the state pension, many older people live in severe poverty. In 1998-1999, 25 percent of older people were living in households whose income fell below half the poverty line, compared with 28 percent of the population as a whole (May, 2003a).


This section considers the impact of the HIV epidemic on older people in South Africa. The rapid development of a severe epidemic has resulted in substantial increases in adult mortality since the mid-1990s, which have reversed improvements in health and survival made in the 1970s and 1980s. The mortality of women ages 25-29 in 1999-2000 was 3.5 times higher than in 1985, while the mortality of men ages 30-39 doubled in the same period (Dorrington et al., 2001).

HIV Infection and AIDS Mortality in Older People

In the United States and Europe, attention has been given to both HIV prevention and the treatment and care of HIV-positive older people (Levy, Ory, and Crystal, 2003). In Africa, however, HIV/AIDS health education programs and the health services largely ignore the risk of HIV infection and AIDS in older people (Wilson and Adamchak, 2001). One of the factors contributing to this neglect is that the available data on HIV infection are restricted largely to women of reproductive age. Most of the seroprevalence data on South Africa have been collected from pregnant women seen at government antenatal clinics. However, the national HIV prevalence survey conducted by the Human Sciences Research Council collected HIV data on a small number of people age 55 or more (Human Sciences Research Council, 2002). Their report presents a graph indicating that approximately 7 percent (95 percent confidence interval, CI, approximately 4-10) of women age 55 or more and 7 percent (95 percent CI ap proximately 3-15) of men in the same age group were HIV positive. These data probably suffer from a number of biases. In particular, they probably overrepresent younger members of this age group. However, they do suggest that appreciable numbers of older South Africans are HIV positive. Moreover, as people may not develop AIDS until many years after they were infected with HIV, significant numbers of AIDS cases and deaths are likely to occur in this age group. A study in rural KwaZulu-Natal in 2000 found that 2 percent of people dying of AIDS with or without tuberculosis were age 60 or more, representing 5 percent of all deaths in this age group (Hosegood et al., 2004).

Older People as Members of Households Affected by HIV/AIDS

Although they are at risk of being infected with HIV themselves, the major impact of the HIV epidemic on older people is indirect. Knodel and colleagues (2003) identify seven pathways though which older people experience the impact of the AIDS epidemic at the family or household level: caregiving, coresidence with an ill adult child, loss of the child, providing financial or material support during the time the adult child is ill, paying for the funeral of the deceased child, fostering grandchildren, and negative community reaction. Both in South Africa and elsewhere in the African region, most studies examining the indirect impact of HIV/AIDS on older people have focused on the role of older people in caring for people with AIDS or their orphaned children, rather than on outcomes for the older people themselves, such as effects on their physical and mental health, economic status, or living arrangements.

The role of older people in the care of relatives with AIDS has been relatively well documented. Many of the empirical studies have been in the United States (Berk, Schur, Dunbar, Bozette, and Shapiro, 2003; Ellis and Muschkin, 1996) and Thailand (Knodel and VanLandingham, 2001; Knodel, Saengtienchai, Im-em, and VanLandingham, 2001a, 2001b). Qualitative studies in Tanzania, Zimbabwe, and South Africa have also shown the important role that older people play in caring for those of their children who develop AIDS (Ferreira, Keikelame, and Mosaval, 2001; Foster et al., 1995; World Health Organization, 2002). In Uganda, 48 percent of people with AIDS were cared for by a parent for at least for some time during their illness (Ntozi and Nakayiwa, 1999). Although the focus of caring has often been on women, men are also involved in supporting ill and bereaved people. In Tanzania, older men and women were equally likely to care for sick household members, although women spent twice as much time as men on caregiving activities (Dayton and Ainsworth, 2002).

In rural South Africa, many of those needing care will have been migrants, often labor migrants, prior to their illness. Their living arrange ments at their place of work often fail to provide them with the level of physical, emotional, and financial support that they can receive from their parents and others in their natal household. Therefore, they often return to their parents’ households when they become chronically sick. As many as 14 percent of the people dying in one area of rural KwaZulu-Natal in 2001-2002 had arrived in the area less than 6 months before they died (Gafos, 2003).

Widespread awareness exists in the media, research, and program arenas of the role of older people in fostering children orphaned by parental deaths due to AIDS. Studies in Zimbabwe, Uganda, and Kenya show that grandmothers, in particular, care voluntarily not only for their orphaned grandchildren but also for other closely related children (Drew, Makufa, and Foster, 1998; Guest, 2001; Nyambedha, Wandibba, and Aagaard-Hansen, 2003).

Few studies in Africa have examined the impact of HIV/AIDS on caregivers’ own health and well-being. AIDS illness and death have short-term and long-term economic consequences for households and their surviving members, including reduced economic status (Rugalema, 1999; Yamano and Jayne, 2004). Since many older people in developing countries are dependent on financial and material assistance from their children and grandchildren, increased mortality of working-age adults will weaken their support networks (Adamchak, Wilson, Nyanguru, and Hampson, 1991).

A qualitative study of grandmothers caring for a child with HIV/AIDS as well as their grandchildren in townships in the Western Cape province of South Africa found that the cost of caring for the sick person (transport, medical bills), as well as taking on more of the costs of childrearing (school fees, food), drove these older women and their households into poverty. In addition, the women reported that their health had worsened as a result of the experience due to the physical demands involved in caregiving and the emotional trauma that they had suffered (Ferreira et al., 2001).

In one of the few longitudinal studies in Africa to investigate the impact of adult deaths on older household members, Dayton and Ainsworth (2002) found that, controlling for poverty, the body mass index of older people decreased significantly in the period immediately following the death of an adult member of their household. In Zimbabwe, elevated levels of emotional and psychological stress have been reported among older caregivers (World Health Organization, 2002).

Older People Living in HIV/AIDS-Affected Communities

The macro-level economic and social impact of the HIV epidemic in Africa is also likely to have implications for older people, for example, by increasing demands on or worsening the quality of health and welfare ser vices, reducing opportunities for paid work, and adversely affecting the supply of adequate foodstuffs (Barnett and Whiteside, 1999, 2000). Water supply, sanitation, and clinic-based and hospital care are particularly important for older people and public health expenditures in South Africa fell from 8.2 percent of the gross domestic product in 1994 to 4.1 percent in 2000 (Walker, 2001).

Even in remote rural areas, older people in South Africa experience a high burden of chronic diseases due to noncommunicable diseases and disorders, principally obesity in women, hypertension, diabetes, stroke, and cancer (Bradshaw et al., 1995; Kahn and Tollman, 1999; Walker, 2001). Almost three-quarters of older people report having at least one chronic illness or ongoing health problem, and more than half of them report a physical disability (Ferreira, 2000). Government health services, particularly in rural areas, are inadequately equipped to provide long-term support for people with chronic diseases, and this is unlikely to improve in the context of the overwhelming pressure on them resulting from the tuberculosis and HIV/AIDS epidemics.


Demographic surveillance systems (DSS) with longitudinal observations on individuals and households provide opportunities to measure the sociodemographic impact of the HIV epidemic. In this section, we describe the demographic and socioeconomic characteristics, composition, and experience of adult deaths of the households of older people using data from a DSS site in rural KwaZulu-Natal. The results presented are from an analysis of a subset of longitudinal data from the Africa Centre Demographic Information System (ACDIS) (Hosegood and Timaeus, 2005).

Study Area

The study area is part of the rural district of Umkhanyakude in northern KwaZulu-Natal. It is situated about 250 km north of the provincial capital of Durban. The area includes both land under tribal authority that was designated as a Zulu “homeland” under South Africa’s former apartheid policy and a township under municipal authority. Infrastructure is poor. In 2001 only 13 percent of households had access to either their own piped water supply or a communal tap. Although this is a rural area, there is little subsistence agriculture. Most households rely on wage income and pensions. Unemployment is high: 67 percent of women and 56 percent of men ages 16-59 were unemployed in 2001 (Case and Ardington, 2004). Few local employment opportunities exist, and consequently labor migra tion is high. Approximately 40 percent of the men and 35 percent of women age 18 years or more who report that they are members of households in the study area also report that they reside outside the area most of the time (Hosegood and Timaeus, 2005).

KwaZulu-Natal is the province of South Africa with the highest prevalence of HIV infection among those attending antenatal clinics (South Africa Department of Health, 2005). An antenatal survey conducted in the study area in 1998 found that 41 percent (95 percent CI: 34.7-47.9) of pregnant women were infected with HIV (Wilkinson, Connolly, and Rotchford, 1999). Preliminary results from a population-based HIV surveillance study in the same area in 2003 found HIV seroprevalence to be 22.2 (95 percent CI: 20.4-24.1) in women ages 15-49 and 12.1 (95 percent CI: 10.4-14.1) in men ages 15-54 (Weltz and Hosegood, 2003). HIV prevalence was highest among men ages 30-34 (42.5, 95 percent CI: 31.0-54.6) and women ages 25-29 (43.2, 95 percent CI: 35.7-51.1).

Data Source

ACDIS started data collection on January 1, 2000. The demographic study area was mapped and all households were registered. The study population includes all members of households living in a dwelling in the study area. ACDIS collects data on both the resident members of a household who sleep at the dwelling most of the time and nonresident members who acknowledge the authority of the household head and visit the household at least once a year. The initial round of fieldwork registered 79,354 individuals as members of 10,612 households. Demographic and health information is collected two or three times a year from all registered households and individuals. It includes reports of all births and deaths and moves between households and in and out of the area. The causes of all notified deaths (of both resident and nonresident household members) are established by clinicians on the basis of a verbal autopsy interview conducted by nurses with the family or caregivers of the deceased (Hosegood et al., 2004). The conceptual and operational design of ACDIS has been described in more detail elsewhere (Hosegood and Timaeus, 2005).

Because of the importance of the old-age pension in South Africa for both pensioners and other members of their household, our analysis of older people’s living arrangements focuses on those of pensionable age. We present information on the individual and household characteristics of the 3,657 women age 60 or more and men age 65 or more who were resident in the study area on January 1, 2000. We also examine changes in the structure of their households by January 1, 2002, together with the structure of the households of the 528 additional people of pensionable age residing in the study area by then. The methods used to generate the data on individual and household relationships are described elsewhere and are available from the authors on request.

Characterizing Household Structure

In ACDIS, the relationship of each member of the household to its head is collected at each round, but establishing the nature of the relationship between two members when neither is the head of the household is more difficult. It is complicated by high levels of extramarital fertility and the frequent presence in households in this area of distantly related or unrelated individuals. In addition, respondents often report socially as well as biologically related individuals as kin or report kin to be more closely related than they are. This is particularly a problem with foster parents and half-siblings (Noumbissi and Zuberi, 2001; Townsend, 1997).

In contrast, ages are available for all household members and provide an unambiguous way of comparing households in the study area. Instead of classifying household structures on the basis of the relationship of other members to the older person, such as coresidence with a grandchild under age 15 (Zimmer and Dayton, 2003), we distinguish four groups of households defined on the basis of the ages of younger household members irrespective of whether the household includes any older people other than the index individual. These are: (a) households in which the older person of pensionable age is the only member or that has other older members only, (b) households with younger adult members but with no members under age 18, (c) households whose membership includes both younger adults and children under age 18, and (d) households with members under age 18 but no adult members below pensionable age. Older people are included in the analysis only as an index individual in the household in which they are resident, but we analyze the composition both of the other residents and of everyone recognized as a member of the household.

One of the primary hypotheses of both researchers and policy makers has been that the deaths of younger adults due to HIV/AIDS in Southern Africa will result in an increase in what are termed “skipped-generation households.” This concept has been defined in various ways. It is generally applied to households made up of coresident grandparents and grandchildren. However, some authors use the term for households that include younger adults, providing that they are not a child of an older household member (Knodel et al., 2003). This analysis equates skipped-generation households with our final group of households, those in which one or more older persons live in a household whose membership includes at least one child but no younger adults.

Living Arrangements of Older People in Rural KwaZulu-Natal

On January 1, 2000, 5 percent of the resident population of the study area were of pensionable age. Some 29 percent of the households in the study area had at least one resident member of pensionable age. Table 8-1 presents individual and household characteristics of these older men and women. Older men and women differ significantly (p < 0.01) in their individual characteristics. Older men are more likely than older women to be married, to be the head of the household, to live in households with more assets, and to live alone.

TABLE 8-1. Characteristics of Older People and Their Households, Persons of Pensionable Age Resident in the Demographic Surveillance Area on January 1, 2000.


Characteristics of Older People and Their Households, Persons of Pensionable Age Resident in the Demographic Surveillance Area on January 1, 2000.

The majority of people of pensionable age (87 percent) live in households with both younger adult and child members. Most of these households would be classified as “extended households” by other authors (Noumbissi and Zuberi, 2001), although our classification does not specify the kinship relationships among the members. Table 8-2 looks jointly at the composition of all members and the other residents in older people’s households. It shows that, while about 15 percent of older people live in households without any younger adult residents, less than half of those living alone or only with other older people, and less than a third of those residing only with other older people and children, live in households than have no younger adult members. Thus, few older people (2 percent) live in households consisting of only other older people and children.

TABLE 8-2. Percentage Distribution of Households with Resident Members of Pensionable Age According to the AgeComposition of Their Other Resident Members and of All Their Other Members, January 1, 2000.


Percentage Distribution of Households with Resident Members of Pensionable Age According to the AgeComposition of Their Other Resident Members and of All Their Other Members, January 1, 2000.

Noumbissi and Zuberi (2001) present estimates of the living arrangements of older people in South Africa based on census and Demographic and Health Survey data. The 1996 Census of South Africa suggests that 79 percent of Africans age 60 or more lived in extended or nuclear households, while 6 percent lived alone and 16 percent lived with both family members and nonrelatives (Noumbissi and Zuberi, 2001). While they do not specifically identify skipped-generation households, modeling by Merli and Palloni (this volume, Chapter 4) based on several sources of data suggests that by 1998 some 15 percent of older people in KwaZulu-Natal were living with a grandchild under age 15 but none of their adult children. This percentage is higher than in the data from Umkhanyakude (in the northeastern part of the Kwazulu-Natal province). Their figures also suggest that less than 2 percent of older people were living alone with one or more grandchildren, both of whose parents had died.

The differences between these three sets of estimates undoubtedly result to a considerable degree from differing definitions in the various inquiries of what “living with” someone means. The 1991 and 1996 South African censuses enumerated only the de facto population. Any household member who resided elsewhere (e.g., adult labor migrants) was excluded from the household schedule. Given the extent of labor migration in such provinces as KwaZulu-Natal, de facto data on households mask the distinction be tween households in which older people living with children do so in the absence of any adult involvement (i.e., the child’s parent has died or abandoned the child) and households in which younger adult members, although nonresident, return periodically and provide at least some social and economic support for both older person and child. One would expect to find a high proportion of skipped-generation households in South African census data simply because, as Merli and Palloni (this volume, Chapter 4) note, many children “lose a parent” to migration.

As we explain elsewhere (Hosegood and Timaeus, 2005), the limitations of the de facto approach were one reason why the ACDIS data system was designed to enumerate both resident and nonresident household members. (The other major reason was to enable ACDIS to track nonresident and occasionally resident individuals longitudinally). Given the different approach that it adopts to data collection, one would expect to find fewer skipped-generation households in ACDIS than in the census. Both perspectives on the household are valid. However, it is wrong to assume that households in which only older adults and children are usually resident only became common recently as a result of high AIDS mortality. By allowing us to look at both residents and all household members, ACDIS makes it clear that, in rural KwaZulu-Natal at any rate, most de facto skipped-generation households have younger adult members and result from high levels of circulatory labor migration rather than the death of younger adults.

Older People’s Experience of Adult Mortality Within Their Households

Over the 2-year period of follow-up, 316 older people (8 percent) resident on January 1, 2000, died. They include seven older women and seven older men who died of AIDS with or without tuberculosis. By the end of the follow-up period, an additional 528 older people were resident in the study area. These people were either below pensionable age in 2000 or only moved into a household in the study area after the start of the surveillance.

Mortality at younger ages in this area is high. Of the 3,180 older people residing in the same household after 2 years, 20 percent had experienced the death of at least one younger adult in their household and 12 percent had experienced one or more deaths of adult household members from AIDS. Household size decreased by 18 percent in households that had a younger adult death, compared with 8 percent in other households with an older person resident.

Changes in the Living Arrangements of Older People

Table 8-3 documents changes in household membership between the beginning of 2000 and the beginning of 2002 for households with older residents. Households that experienced the death of a younger adult household member are shown separately from those that did not. Older people who were already living alone with children are not represented in the first of these two groups, although in some instances younger adult members may have joined the household and then died during the 2-year period under consideration. Table 8-4 presents data on changes of residence during 2000 and 2001 for all households with a resident older member.

TABLE 8-3. Percentage Distribution According to Age of the Other Members in 2002 in Households with Residentsof Pensionable Age by Age of the Other Members in 2000.


Percentage Distribution According to Age of the Other Members in 2002 in Households with Residentsof Pensionable Age by Age of the Other Members in 2000.

TABLE 8-4. Percentage Distribution According to Age of the Other Residents in 2002 in Households with Residents of Pensionable Age by Age of the Other Residents in 2000.


Percentage Distribution According to Age of the Other Residents in 2002 in Households with Residents of Pensionable Age by Age of the Other Residents in 2000.

Comparison of the changes in household type shown in Table 8-3 suggests that older people experiencing younger adult deaths in their households are more likely to undergo substantial changes in their living arrangements than older people living in unaffected households. Younger adult deaths were more likely to change the structure of the households of older people living in households without children than of those living in households with a child as well as adult members.

The death of a young adult left 15 additional older people in households that included children but had no younger adult members. Another 15 such households were created by people leaving or joining households. But these new skipped-generation households were matched by an equal number of shifts out of this type of household. During the 2 years, 38 percent of all those living initially in households with a child but no younger adult members either moved to join another household that had a younger adult member, were joined by a younger adult, or saw the children leave the household. Moreover, the de facto data in Table 8-4 show that the number of older people who were residing with only children and other older people dropped by 11 percent during the 2 years. While older people in skipped-generation households were more likely to have had their living arrangements change than other older people, no increase occurred in the proportion of older people living alone with children.


The ACDIS data reveal the high level of younger adult mortality that older people are facing in rural South Africa, even in most their immediate social sphere, with 20 percent of them experiencing such a death in their households in the 2-year period considered. Given the relatively short period of follow-up, longer term changes in the living arrangements of older people experiencing a death in their household late in the period have not been observed. However, as one might suspect, living in a household in which a younger adult dies raises the likelihood that the older person’s household no longer contains young adults at the end of the period of follow-up.

Neither the ACDIS data nor the census data provide any evidence that the increase in adult mortality resulting from the HIV/AIDS epidemic is raising greatly the prevalence of skipped-generation households. Data from other African countries also suggest that the majority of older people continue to live in extended, multigenerational households. Recent Demographic and Health Survey data from 16 countries in sub-Saharan Africa show that fewer than 5 percent of older people were living with children and no adults (Zimmer and Dayton, 2003). Even in Uganda, a country with a relatively mature HIV epidemic, the proportion of skipped-generation households was less than 2 percent in 1995 (Ntozi and Zirimenya, 1999).

In discussing the results of their modeling of the living arrangements of older people in South Africa, Merli and Palloni (this volume, Chapter 4) also note that the household composition is not evolving in the way anticipated. They suggest that the lack of increase in older people living with grandchildren in skipped-generation households “may be because the epidemic has not worked its way through with sufficient force, because individuals and groups react in ways that conceal the trail left by HIV/AIDS, or because we may be unable to distinguish the effects associated with HIV/AIDS from those triggered by migration, which mimics the effects of HIV/AIDS on the availability of kin, or those induced by modernization, which changes preferences for coresidence.”

These suggestions echo issues we identified in the first two sections of this paper. Even prior to the HIV epidemic in South Africa, older people played an enormously important role in the care of children and the maintenance of rural households. However, it remains fairly unusual for a household with young children and older people not to include any of the children’s parents, aunts, or uncles (i.e., the older person’s adult children). In South Africa, low marriage rates, extensive labor migration, and the costs involved in establishing independent dwellings have acted to slow the trend toward nuclearization of family life usually observed in the course of development and encouraged younger adults to remain members of their parental household (Bongaarts and Zimmer, 2002). Most families would acknowledge their responsibility toward those older people who are in need of care and support themselves. Thus, it is rare to find young children caring for older people alone without the assistance of adults. Of course, there are older people in South Africa who have no surviving relatives in the next generation or who have lost touch with or quarreled with them all. Usually though, it is only in such extreme circumstances that the relatives of older people and children who find themselves coping alone after the death of an adult would not seek to alter the arrangement. This can be achieved by sending adults to help or placing the children (or the older person, or both) in another household. Thus, in Umkhanyakude, 9-10 percent of older people coreside with their grandchildren or other young children while their adult children are absent. However, less than 30 percent of these de facto skipped-generation households had no younger adult members. Moreover, only half the latter households survived for 2 years, indicating that they are vulnerable to dissolution and the dispersal of their members to other households or to change through younger adults joining the household (Hosegood, McGrath, Herbst, and Timaeus, 2004; Ford and Hosegood, 2005).

The old age pension may also encourage younger adults and children to live with, and if necessary care for, older people in South Africa. The pension may act as a “magnet” that ensures that households with older people are attractive for adult as well as child dependents. Therefore, even after the death of the older person’s own adult children, other relatives may be eager to ensure that they do not live alone. In many South African households, older people may be “burdened” by the young rather than vice versa, although in others the pension may just mean that the cost of caring for an older person is reduced.

Although we emphasize that it is rare for older people to live with children without younger adults as coresident or nonresident members of the household, we do not wish to minimize the difficulties faced by older people and children who find themselves, even temporarily, in this situation. In part the severity of these difficulties will depend on the age of the older people with whom children live. The grandmothers of most children under age 18 in South Africa will be fairly young and may be able to provide them with considerable economic, financial, and emotional care. In households containing frail older people, however, the primary direction of caregiving may be from the child or children to the older person.

This study has not examined indicators of well-being for older people and other members of households. However, we believe that for policy makers and programs to target skipped-generation households may be a very poor way of identifying the most vulnerable households. Given the importance of labor migration in South Africa, one would need to examine patterns of residence and caregiving in some detail to determine the extent to which older people and children live alone for considerable periods of time without significant assistance from younger adults and how well they are able to cope with the demands that this places on them.

The impact of HIV/AIDS on older people is receiving increasing research attention. However, there remains a serious lack of empirical data that can be used to examine a broad range of outcomes. Few surveys collect data on the morbidity, nutritional status, or mental health of older people. There are also limited data with which to assess wider impacts of the HIV epidemic on older people, such as stigmatization and isolation following AIDS deaths in their households, increased financial insecurity and increased workloads, and deterioration in the availability and quality of health and welfare services.

We have sought to highlight aspects of life in South Africa that influ ence the demographic, health, and economic characteristics of older people, as well as those that influence the impact of HIV/AIDS on them. Our population-based data from rural KwaZulu-Natal demonstrate that HIV/AIDS is very much part of many older peoples’ lives today. However, we also suggest that adaptive strategies are being adopted by many households to protect dependent individuals, including older people. The existence of a substantial noncontributory old age pension in South Africa is undoubtedly of huge benefit both to those older people whose lives have been affected by AIDS and those who have not. In addition, the fluid nature of households, the limited involvement of most of them in agricultural production, and the stretching of household groups due to migration—all of which developed in response to the political economy of apartheid—give social networks and living arrangements in South Africa a degree of flexibility that benefits many older people when adverse advents occur, such as the death of an adult child.


  1. Adamchak DJ, Wilson AO, Nyanguru A, Hampson J. Elderly support and intergenerational transfer in Zimbabwe: An analysis by gender, marital status, and place of residence. South African Journal of Gerontology. 1991;31:505–513. [PubMed: 1894155]
  2. Barnett T, Whiteside A. HIV/AIDS and development: Case studies and a conceptual framework. European Journal of Development Research. 1999;11(2):200–234.
  3. Barnett T, Whiteside A. Guidelines for studies of the social and economic impact of HIV/AIDS. Geneva, Switzerland: Joint United Nations Programme on HIV/AIDS; 2000. Best Practice Collection.
  4. Berk ML, Schur CL, Dunbar JL, Bozette S, Shapiro M. Short report: Migration among persons living with HIV. Social Science and Medicine. 2003;57:1091–1097. [PubMed: 12878108]
  5. Bongaarts J, Zimmer Z. Living arrangements of the elderly in the developing world: An analysis of DHS household surveys. Journal of Gerontology Series B: Psychological Sciences and Social Sciences. 2002;57(3):S145–S157. [PubMed: 11983741]
  6. Bradshaw D, Bourne D, Schneider M, Sayed R. Mortality patterns of chronic diseases of lifestyle in South Africa. In: Fourie J, Steyn K, editors. Chronic diseases of lifestyle in South Africa. Cape Town, South Africa: Medical Research Council; 1995. Medical Research Council Technical Report.
  7. Burman S. Child care by the elderly and the duty of support in multigenerational households. Southern African Journal of Gerontology. 1995;4(1):13–17.
  8. Campbell C. Intergenerational conflict in township families: Transforming notions of respect and changing power relations. Southern African Journal of Gerontology. 1994;3(2):37–42.
  9. Case A. Does money protect health status? Evidence from South African pensions. Cambridge, MA: National Bureau of Economic Research; 2001. NBER Working Paper No. 8495.
  10. Case A, Ardington C. ACDIS monograph: Socioeconomic factors. Mtubatuba, South Africa: Africa Centre for Health and Population Studies; 2004.
  11. Case A, Deaton A. Large cash transfers to the elderly in South Africa. Economic Journal. 1998;108(450):1330–1361.
  12. Case A, Hosegood V, Lund F. Child support grant study findings. Children First. 2003;51:45–49.
  13. Case A, Hosegood V, Lund F. The reach of the South African child support grant: Evidence from KwaZulu-Natal. Development Southern Africa. 2005;22(4):467–482.
  14. Dayton J, Ainsworth M. The elderly and AIDS: Coping strategies and health consequences in rural Tanzania. New York: The Population Council; 2002. Working Paper No. 1.
  15. Dorrington R, Bourne D, Bradshaw D, Laubscher R, Timaeus IM. The impact of HIV/AIDS on adult mortality in South Africa. Tygerberg, South Africa: Burden of Disease Unit, Medical Research Council; 2001.
  16. Dorrington R, Moultrie TA, Timaeus IM. Estimation of mortality using the South African census 2001 data. Cape Town, South Africa: Centre for Actuarial Research, University of Cape Town; 2004. [accessed February 2005]. Monograph No. 11. Available:http://www.commerce.uct.ac.za/care/Monographs/Monographs/Mono11.pdf.
  17. Drew RS, Makufa C, Foster G. Strategies for providing care and support to children orphaned by AIDS. AIDS Care. 1998;10(Suppl. 1):9–15. [PubMed: 9625890]
  18. Du Toit BM. Does the road get lonelier? Aging in a coloured community in South Africa. Journal of Aging Studies. 1994;8(4):357–374.
  19. Ellis M, Muschkin C. Migration of persons with AIDS: A search for support from elderly parents? Social Science and Medicine. 1996;43(7):1109–1118. [PubMed: 8890411]
  20. Ferreira M. Growing old in the new South Africa. Ageing International. 2000 Spring;:32–46.
  21. Ferreira M, Keikelame MP, Mosaval Y. Older women as carers to children and grandchildren affected by AIDS: A study towards supporting the carers. Cape Town, South Africa: Institute of Ageing in Africa, University of Cape Town; 2001.
  22. Ferreira M, Møller V, Prinsloo FR, Gillis LS. Multidimensional survey of elderly South Africans, 1990-1991: Key findings. Cape Town, South Africa: Human Sciences Research Council and HSRC/UCT Centre for Gerontology, University of Cape Town; 1992. Monograph No. 1.
  23. Ford K, Hosegood V. AIDS mortality and the mobility of children in KwaZulu Natal, South Africa. Demography. 2005;42(4):757–768. [PubMed: 16463920]
  24. Foster G, Shakespeare R, Chinemana F, Jackson H, Gregson S, Marange C, Mashumba S. Orphan prevalence and extended family care in a periurban community in Zimbabwe. AIDS Care. 1995;7(1):3–17. [PubMed: 7748908]
  25. Gafos M. An analysis of the associations between migration, mortality, and health care usage in the adult population of Hlabisa District, KwaZulu Natal in South Africa. London School of Hygiene and Tropical Medicine; 2003. Unpublished MSc thesis.
  26. Gillis LS, Welman M, Koch A, Joyi M. Psychological distress and depression in urbanizing elderly black persons. South African Medical Journal. 1991;79(8):490–495. [PubMed: 2020893]
  27. Guest E. Children of AIDS: Africa’s orphan crisis. Pietermaritzburg, South Africa: University of Natal Press; 2001.
  28. HelpAge International. Ageing issues in Africa: A summary. London, England: Author; 2000.
  29. HelpAge International. HIV/AIDS and ageing: A briefing paper. London, England: Author; 2003. [accessed December 2004]. Available:http://www.helpage.org/images/pdfs/briefing%20papers/HIV%20AIDS%20position%20paper.pdf.
  30. Hosegood V, Timaeus IM. Household composition and dynamics in KwaZulu Natal, South Africa: Mirroring social reality in longitudinal data collection. In: van der Walle E, editor. African households: Census data. New York: M.E. Sharpe; 2005. pp. 58–77.
  31. Hosegood V, McGrath N, Herbst K, Timaeus IM. The impact of adult mortality on household dissolution and migration in rural South Africa. AIDS. 2004;18(11):1585–1590. [PubMed: 15238777]
  32. Hosegood V, Vanneste AM, Timaeus IM. Levels and causes of adult mortality in rural South Africa. AIDS. 2004;18:1–19.
  33. Human Sciences Research Council. Nelson Mandela HSRC study of HIV/AIDS: Full report. Pretoria, South Africa: Human Sciences Research Council Press; 2002. [accessed April 2005]. Available:http://www.hsrcpublishers.co.za/user_uploads/tblPDF/2009_00_Nelson_Mandela_HIV_Full_Report.pdf.
  34. Hunter N, Rushby J. Annotated bibliography of recent research on the living conditions of the main target groups of social security grants. Durban, South Africa: School of Development Studies; 2002. Research Report No. 53.
  35. Izzard W. Migrants and mothers: Case-studies from Botswana. Journal of Southern African Studies. 1985;11(2):258–280.
  36. Johnson P. The structured dependency of the elderly: A critical note. In: Jeffreys M, editor. Growing old in the twentieth century. London, England: Routledge; 1989. pp. 62–72.
  37. Kahn K, Tollman SM. Stroke in rural South Africa: Contributing to the little known about a big problem. South African Medical Journal. 1999;89(1):63–65. [PubMed: 10070416]
  38. Karim QA, Karim SSA. Epidemiology of HIV infection in South Africa. Aids. 1999;13(6):4–7.
  39. Kinsella K, Ferreira M. Ageing trends: South Africa. Washington, DC: Department of Commerce, Bureau of the Census; 1997. [accessed December 2004]. Available:http://www.census.gov/ipc/prod/ib-9702.pdf.
  40. Knodel J, VanLandingham M. Return migration in the context of parental assistance in the AIDS epidemic: The Thai experience. Ann Arbor: University of Michigan; 2001. Population Studies Center Report No. 01-492. [PubMed: 12765712]
  41. Knodel J, Watkins SC, VanLandingham M. AIDS and older persons: An international perspective. Journal of Acquired Immune Deficiency Syndrome. 2003;33(Suppl. 2):153–165. [PubMed: 12853864]
  42. Knodel J, Saengtienchai C, Im-em W, VanLandingham M. The impact of AIDS on parents and families in Thailand. A key informant process. Research on Ageing. 2001a;23(6):633–670.
  43. Knodel J, Saengtienchai C, Im-em W, VanLandingham M. Older people and AIDS: Quantitative evidence of the impact in Thailand. Social Science and Medicine. 2001b;52:1313–1327. [PubMed: 11286358]
  44. Leliveld A. The effects of restrictive South African migrant labor policy on the survival of rural households in Southern Africa: A case study from rural Swaziland. World Development. 1997;25(11):1839–1849.
  45. Levy JA, Ory MG, Crystal S. HIV/AIDS interventions for midlife and older adults: Current status and challenges. Journal of Acquired Immune Deficiency Syndrome. 2003;33(Suppl. 2):59–67. [PubMed: 12853854]
  46. Madhavan S. Fosterage patterns in the age of AIDS: Continuity and change. Social Science and Medicine. 2004;58(7):1443–1454. [PubMed: 14759688]
  47. May J. Chronic poverty and older people in South Africa. Durban, South Africa: Chronic Poverty Research Centre; 2003a. [accessed May 2005]. Working Paper No. 25, School of Development Studies. Available:http://www.chronicpoverty.org/pdfs/JMay.pdf.
  48. May J. Extent, distribution, and nature of poverty among the elderly in South Africa. Manchester, England: Chronic Poverty Research Centre; 2003b.
  49. Mazur RE. Research in rural sociology and development, no. 7. Greenwich, CT: JAI Press; 1998. Migration dynamics and development in rural South Africa: Demographic and socioeconomic perspectives; pp. 197–225. [PubMed: 12294797]
  50. Mohatle T, de Graft Agyarko R. Contributions of older people to development: The South African study. London, England: HelpAge International; 1999.
  51. Møller V, Ferreira M. Getting by: Benefits of noncontributory pension income for older South African households. Cape Town, South Africa: Institute of Ageing in Africa, University of Cape Town; 2003. [accessed June 2005]. Available:http://idpm.man.ac.uk/ncpps.
  52. Møller V, Sotshongaye A. My family eats this money too: Pension-sharing and self-respect among Zulu grandmothers. South African Journal of Gerontology. 1996;5(2):9–19.
  53. Moultrie TA, Timaeus IM. Trends in South African fertility between 1970 and 1998: An analysis of the 1996 census and the 1998 demographic and health survey. Tygerberg, South Africa: Medical Research Council; 2002. [accessed March 2004]. Technical report. Available:http://www.mrc.ac.za/bod/trends.pdf.
  54. Moultrie TA, Timaeus IM. The South African fertility decline: Evidence from two censuses and a demographic and health survey. Population Studies. 2003;57(3):265–283. [PubMed: 14602529]
  55. Murray C. Migrant labour and changing family structure in the rural periphery of Southern Africa. Journal of Southern African Studies. 1980;6(2):139–156.
  56. Nhongo TM. Impact of HIV/AIDS on generational roles and intergenerational relationships; Paper presented at the National Research Council Workshop on HIV/AIDS and Family Well-Being; January 28-30; Namibia, South Africa. 2004.
  57. Noumbissi A, Zuberi T. Household structure and aging in South Africa: A research note; Presented at the African Census Analysis Project Virtual Conference on African Households, An Exploration of Census Data; November 21-24; 2001. [accessed April 2005]. Available:http://www.pop.upenn.edu/africahh/NoumbissiZuberi.pdf.
  58. Noumbissi A, Bawah AA, Zuberi T. Parental survival and residential patterns. In: van der Walle E, editor. African households. Census data. New York: M.E. Sharpe; 2005.
  59. Ntozi JPM, Nakayiwa S. AIDS in Uganda: How has the household coped with the epidemic? In: Orubuloye IO, Caldwell JC, Ntozi JPM, editors. The continuing HIV/AIDS epidemic in Africa: Response and coping strategies. Canberra, Australia: Health Transition Centre, Australian National University; 1999. pp. 155–180.
  60. Ntozi JPM, Zirimenya S. Changes in household composition and family structure during the AIDS epidemic in Uganda. In: Orubuloye IO, Caldwell JC, Ntozi JPM, editors. The continuing HIV/AIDS epidemic in Africa: Responses and coping strategies. Canberra, Australia: Health Transition Centre, Australian National University; 1999. pp. 193–209.
  61. Nyambedha EO, Wandibba S, Aagaard-Hansen J. Changing patterns of orphan care due to HIV epidemic in western Kenya. Social Science and Medicine. 2003;57:301–311. [PubMed: 12765710]
  62. Preston-Whyte E. Families without marriage: A Zulu case study. In: Argyle J, Preston-Whyte E, editors. Social system and tradition in Southern Africa: Essays in honour of Eileen Krige. Cape Town, South Africa: Oxford University Press; 1978. pp. 55–85.
  63. Preston-Whyte E. Women-headed households and development: The relevance of cross-cultural models for research on black women in Southern Africa. Africanus. 1988;18:58–76.
  64. Rugalema G. It is not only the loss of labor: HIV/AIDS, loss of household assets and household livelihood in Bukoba district, Tanzania. In: Mutangadura G, Jackson H, Mukurazita D, editors. AIDS and African smallholder agriculture. Harare, Zimbabwe: Southern Africa HIV and AIDS Information and Dissemination Service; 1999.
  65. Sagner A. Ageing and social policy in South Africa: Historical perspectives with particular reference to the Eastern Cape. Journal of Southern African Studies. 2000;26(3):523–553.
  66. South Africa Department of Health. National HIV and syphilis antenatal seroprevalence survey in South Africa, 2004. Pretoria: Author; 2005. [accessed December 2004]. Available:http://www.doh.gov.za/docs/reports/2004/hiv-syphilis.pdf.
  67. South Africa Department of Social Development. Mothers and fathers of the nation: The forgotten people? Pretoria: Author; 2001. [accessed March 2005]. Vol. 1, main report—Ministerial Committee on Abuse, Neglect, and Ill-treatment of Older Persons. Available:http://www.welfare.gov.za/Documents/2001/March/elder.htm.
  68. Spiegel A. Dispersing dependents: A response to the exigencies of labour migration in rural Transkei. In: Eades J, editor. Migrants, workers, and the social order. London, England: Tavistock; 1987. pp. 113–129.
  69. Timaeus IM. Mortality in sub-Saharan Africa. In: Chamie J, Cliquet RL, editors. Health and mortality: Issues of global concern. New York: Population Division, United Nations and Population and Family Study Centre; 1999. pp. 108–131.
  70. Townsend N. Reproduction in anthropology and demography. In: Kertzer D, Fricke T, editors. Anthropological demography: Toward a new synthesis. Chicago, IL: University of Chicago Press; 1997. pp. 96–114.
  71. U.S. Census Bureau. International database. 2005. [accessed August 26, 2005]. Available:http://www.census.gov/ipc/www/idbsum.html.
  72. Van der Waal CS. Rural children and residential instability in the Northern Province of South Africa. Social Dynamics. 1996;22(1):31–54.
  73. Walker AR. Changes in public health in South Africa from 1876. Journal of the Royal Society of Health. 2001;121(2):85–93. [PubMed: 11467213]
  74. Weltz T, Hosegood V. Population-based HIV surveillance in the Hlabisa DSS, South Africa; Presented at Forum 7, Global Forum for Health Research; Geneva, Switzerland. Dec, 2003.
  75. Wilkinson D, Connolly C, Rotchford K. Continued explosive rise in HIV prevalence among pregnant women in rural South Africa. Aids. 1999;13(6):740. [PubMed: 10397577]
  76. Wilson AO, Adamchak DJ. The grandmothers’ disease: The impact of AIDS on Africa’s older women. Age and Ageing. 2001;30:8–10. [PubMed: 11322679]
  77. Women's Budget Initiative. What’s available? A guide to government grants and other support available to individuals and community groups. Cape Town, South Africa: 2003. [accessed June 2005]. Available:http://www.internationalbudget.org/resources/library/govtfunds.pdf.
  78. World Health Organization. The impact of AIDS on older persons in Africa. Geneva, Switzerland: Author; 2002.
  79. Yamano T, Jayne TS. Measuring the impact of working-age adult mortality on small-scale farm households in Kenya. World Development. 2004;32(1):91–119.
  80. Zimmer Z, Dayton J. The living arrangements of older adults in sub-Saharan Africa in a time of HIV/AIDS. New York: The Population Council; 2003. Report No. 169.



We regard race as a social reality produced by racism, not a biological characteristic. This paper follows South African practice and uses the term “African” to refer to the majority racial group. In South Africa, “black” has the connotation “nonwhite.” Statistical sources on South Africa usually distinguish four population groups: Africans, whites, coloreds, and Asians.


Several types of households headed by women have been described in South Africa, including those headed by women following the death or separation of their spouse and matrifocal or female-linked households. These are households formed by younger, never married women and their children. In such households the dominant relationships are between mother and daughters and, to a lesser extent, sisters. Husbands and other male partners are not necessarily absent, but the unions may be temporary or not socially recognized (Preston-Whyte, 1978, 1988).