Aging in Sub-Saharan Africa: Recommendations for Furthering Research (2006)

This page intentionally left blank.

Economic security, health and disability, and living conditions in old age are policy concerns throughout the world, but the nature of the problems differs considerably from continent to continent and between and within countries. In industrialized countries, old age support comes to a great extent from large public or private pension and health systems. These systems are becoming increasingly strained as population aging has increased the proportion of older people. At the same time, in much of the industrialized world (Russia being a major exception), the health of the older population is, at a minimum, remaining steady and, in many places, it is improving rapidly.

By contrast, throughout most of the developing world, providing support for older people is still primarily a family responsibility. Traditionally in sub-Saharan Africa,¹ the main source of support has been the household and family, supplemented in many cases by other informal mechanisms, such as kinship networks and mutual aid societies. With the notable exceptions of Botswana, Mauritius, Namibia, and South Africa, formal pensions (whether contributory or not) or other social welfare schemes are virtually nonexistent and, when they do exist, tend to pay minimal benefits and cover only a small fraction of the elderly population (Gillian, Turner, Bailey, and Latulippe, 2000). Population aging is only beginning and, because fertility is falling, it is occurring during a temporary phase of declining dependency

burdens (see the paper by Velkoff and Kowal in this volume). Older people make up a relatively small fraction of the total population, which is expected to increase slowly, although their numbers are increasing rapidly. There are also major differences in the principal health challenges in sub-Saharan Africa compared with industrialized countries. In much of sub-Saharan Africa, gains in life expectancy that were achieved throughout the latter half of the 20th century have been eroded by the HIV/AIDS pandemic (see Chapter 2). Yet little is known about the health and disability of older people and patterns of change. In addition, traditional caring and social support mechanisms now appear to be under increasing strain (Apt, 1996; Dhemba, Gumbo, and Nyamusara, 2002; Kasente, Asingwire, Banugire, and Kyomuhenda, 2002; Mchomvu, Tungaraza, and Maghimbi, 2002; Mukuka, Kalikiti, and Musenge, 2002; Williams, 2003; Williams and Tumwekwase, 2001).

Reasons for this strain include a series of profound economic and social changes associated with development and modernization. In sub-Saharan Africa, older people have traditionally been viewed in a positive light, as repositories of information and wisdom. To date, sub-Saharan African families have shown a great deal of resilience and are generally still intact. Changes associated with development and modernization can, however, combine to weaken traditional social values and networks that stress the important role of older people in society and that reinforce traditions of intergenerational exchange and reciprocity. These changes include increasing formal education and the migration of young people from rural to urban areas, leaving older family members behind. Far more is known about the impact of these changes in other regions of the world, particularly Asia (see Hermalin, 2002), compared with sub-Saharan Africa. Yet their effects pale, in parts of sub-Saharan Africa, in comparison to the effects of the HIV/AIDS pandemic.

Sub-Saharan Africa has long carried a high burden of disease, including from malaria and tuberculosis; today it remains at the center of the HIV/AIDS pandemic. The Joint United Nations Programme on HIV/AIDS and the World Health Organization estimate that more than 60 percent of all people living with HIV are in sub-Saharan Africa—some 25.8 million in 2005 (Joint United Nations Programme on HIV/AIDS, 2006). The lives of older people may be affected by their own illness, but it is more likely, given the age structure of the pandemic, that they are affected by sickness and death of their adult children. Not only do these older people face the loss of a child or children who may well have been a vital source of support and caregiving, but many are also then faced with additional obligations and responsibilities for grandchildren and other members of their extended families.

Finally, AIDS and other social changes are occurring against a backdrop of persistent poverty and deprivation. Sub-Saharan Africa remains the least developed and least urbanized region in the world. Approximately two-thirds of the population of sub-Saharan Africa still live in rural areas and rely largely on near subsistence agriculture or traditional pastoralism for their livelihoods. In such settings, families have to be very self-reliant. Chronic poverty becomes a critical risk factor for the well-being of older people, and more than two of every five of the continent’s inhabitants remain trapped in debilitating poverty (United Nations, 2006). In addition, while the continent has witnessed a decline in the number of armed conflicts since the early 1990s, persistent violence and in some cases seemingly intractable conflicts compound the region’s problems and present critical obstacles to development in some countries (Marshall and Gurr, 2005; Porter, Robinson, Smyth, Schnabel, and Osaghae, 2005).²

Researchers in sub-Saharan Africa are only now beginning to ask how all of these factors—trends in socioeconomic conditions, changing cultural norms and values, changing levels of formal and informal social support, ongoing poor health conditions, and the AIDS crisis—are combining to affect the well-being of older people. African gerontology has expanded in recent years, much of the work based on anthropological approaches (see, for example, Makoni and Stroeken, 2002). Elsewhere—in Europe and Asia, for example—major research programs focused on older people are under way, and longitudinal studies now provide a great deal of information on economic well-being, health, and family processes (see, for example, Börsch-Supan et al., 2005; Hermalin, 2002). Comparable efforts are only beginning in sub-Saharan Africa; instead, much of what is known today comes either from censuses, which often are not particularly reliable or particularly detailed, or from small cross-sectional surveys, which often suffer from problems of nongeneralizability.³

It is against this backdrop that in 2004 the National Institute on Aging asked the National Academies’ Committee on Population to organize a workshop on advancing aging research in Africa. The workshop was to explore ways to promote U.S. research interests in aging in developing countries and to increase the capacity of sub-Saharan African governments and institutions to address the many challenges posed by the changing position of older persons in an era of AIDS. The workshop provided an opportunity for leading scientists from a variety of relevant disciplines to come together and review the evidence on economic security, health, and living conditions of older people and the ways in which critical changes are affecting their well-being.

The Committee on Population appointed an eight-member panel comprised of U.S. and African researchers. The panel had two charges: to develop the two-day workshop and to identify, subsequently, a research agenda aimed at gathering new data that would enable policy makers to better anticipate existing and changing needs of older people and to better assess the viability and potential impact of various public policy options.

In advance of the workshop, the panel commissioned 12 papers by prominent researchers to gather together recent research findings relevant to the goals of the workshop. These papers covered important domains of research on aging, including the changing demography of sub-Saharan Africa (Velkoff and Kowal), demographic impacts of the HIV epidemic on older people (Clark, Merli and Palloni, and Hosegood and Timaeus), formal and informal social security systems (Kaseke), health (Kahn et al., Kuhn et al., and Kuate-Defo), measurement (Kuhn et al.), the impact of social pensions (Lam et al. and Posel et al.), the situation of older people in urban areas (Ezeh et al.), living arrangements (Kuate-Defo and Ezeh et al.), and policy (Peachey and Nhongo). The primary focus of the workshop was to present and discuss recent studies using high-quality data from the region on the situation of older people. Selected papers from the workshop were revised and edited for inclusion in this volume. The workshop also included a focused discussion on data needs and future research directions. Thus, the principal inputs into this report are the panel’s interpretation of the literature consulted,⁴ the presentations of the commissioned papers, and the rich interactive discussion that occurred at the workshop.

Five key themes emerged from the panel’s original planning meeting, the workshop papers, and the workshop discussion. The first is the lack of basic, agreed-upon definitions crucial to the study of aging in sub-Saharan African societies. Most fundamentally, who is elderly in sub-Saharan Africa? Do the definitions used in industrialized societies have the same meaning for sub-Saharan Africa? And, in the African context, with its complex and extended family structure, what constitutes a household, the usual unit in which older people are studied (van de Walle, 2006)? Research can adequately assess the situation of older people in sub-Saharan Africa only if it is conducted in a framework that can allow for a full range of actors and impacts on the well-being of older people.

A second theme and persistent lament throughout the workshop is the lack of careful empirical research and the dearth of comprehensive data needed to rectify this situation. Third is the participants’ belief that the situation of older people in sub-Saharan Africa is changing fairly rapidly. Fourth is the need to recognize the considerable diversity across sub-Saharan Africa with respect to a wide range of indicators. The final central theme is the need to support the development of local research capacity and facilitate research. Each of these themes is amplified below.

When asked who is old, participants in recent focus group interviews in Nairobi claimed that old people can be identified in a variety of ways: by their physical attributes or appearance (e.g., gray hair, wrinkles, obvious frailty), by their life experiences (e.g., their reproductive history), or by the roles that they sometimes play in their community (see the paper by Ezeh et al. in this volume). Consequently, chronological age—which in any event may not even be known in sub-Saharan Africa—may be a poorer indicator of being elderly than social standing. Chronological age may also differ markedly from functional age, which can be the most important dimension of aging in a rural subsistence agricultural context. In general, it is important to recognize that in some sub-Saharan African settings, people who are younger than 60 may be considered old because they exhibit morbidity profiles and take on status roles more usually associated with people over the age of 60 in other settings.

Similar definitional problems surround the meaning of the term household (van de Walle, 2006). Due to the extended-family system, households are more likely to be larger, multigenerational, and less nuclear than in Western societies. For example, a recent study of household composition in Gabon found that about half of all households in Libreville and Port-Gentil

contained at least one guest who had some sort of kinship tie to the family but who did not belong to the nuclear family (Rapoport, 2004). A similar situation can be found in a large number of sub-Saharan African countries. Households can also be split across geographic locations, with families maintaining both a rural home and an urban home. This situation frequently occurs when households decide to allocate their labor resources between rural and urban areas in order to diversify risk, maximize incomes, or both (Agesa, 2004; Lucas and Stark, 1985; Stark, 1991, 1995). Sub-Saharan African family structures complicate the study of older people’s well-being, and basic definitional questions must be resolved before comparative analysis can take place easily.

As stated above, research on aging in sub-Saharan Africa is still very much in its infancy. The current situation of older people in sub-Saharan Africa is, in fact, quite poorly known, and micro-level data are available only for a limited number of countries. The World Health Surveys and the follow-up Study on Global Aging (SAGE) will improve availability of data on health for some countries. Very little information is readily available for Francophone or Lusophone (Portuguese-speaking) Africa. In addition, the range of topics addressed at the workshop illustrates the need for multidisciplinary work that cuts across traditional research domains. In Asia, by contrast, a decade of in-depth and wide-ranging research on multiple aspects of peoples’ lives has generated considerable insight into the situation of older people in that region (see, for example, Hermalin, 2002), which may well be able to inform the design and implementation of similar research in sub-Saharan Africa (Knodel, 2005).

Despite the lack of longitudinal studies, many observers believe that older people are worse off than they were in the past. There are a number of reasons why this might be the case, although there is currently very little empirical research that documents whether older people are worse or better off on most measures of welfare. Three dimensions of change that have bearing on the well-being of older persons were repeatedly raised throughout the two days of the workshop: demographic change, modernization and development, and the impact of HIV/AIDS.

Based on the demographic changes taking place, both the absolute size and the relative proportion of the population age 60 and over are projected over the next 25 years to grow faster than at all younger ages (Table 1-1,

and see Chapter 2). The increase in the population age 75 and over will be particularly noticeable.

Modernization and development have led to broad social and economic changes that put in doubt the continued viability of traditional arrangements for the care and support of older people. For example, formal education and modernization are generally associated with weakening traditional social ties and obligations and greater independence and autonomy, factors that tend to undermine traditional extended family systems. Similarly, economic development is associated with young people migrating from rural to urban areas, leaving older family members geographically isolated.⁵ Once established in urban areas, migrants tend to form new nuclear households. Although children may remit money and goods, such flows are typically irregular and may not be enough to provide much in the way of real economic security. These changes have combined to alter, probably permanently, the nature of the relationship between generations.

The HIV/AIDS epidemic has severely affected many communities across sub-Saharan Africa, with multiple impacts on older people. The vast majority of the estimated 25.8 million people living with HIV are still in their prime wage-earning years—that is, at ages at which normally they would be expected to be not only wage earners but also the principal sources of financial and material support for older people and children in their families (Joint United Nations Programme on HIV/AIDS, 2006).

There is a substantial amount of uncertainty about the future course of the pandemic. While some positive news about gradual, modest declines in HIV prevalence are emerging from East Africa and Zimbabwe, HIV prevalence has soared in Southern Africa in recent years (Asamoah-Odei, Garcia Calleia, and Boerma, 2004; Joint United Nations Programme on HIV/AIDS, 2006). In some antenatal clinics in urban areas of Southern Africa, HIV prevalence rates as high as 25 percent have been recorded, whereas they were only around 5 percent in 1990. Very high HIV prevalence—almost 40 percent of pregnant women found to be HIV positive—has been recorded recently in Botswana and Swaziland (Asamoah-Odei et al., 2004). Increasingly, because of the HIV/AIDS epidemic, older people are being asked to provide emotional and economic support both to their own children, the immediate victims of the HIV/AIDS epidemic, and to their grandchildren (Makiwane, Schneider, and Gopane, 2004; Nyambedha, Wandibba, and Aagaard-Hansen, 2003; Williams and Tumwekwase, 2001).

While modernization theory has been the most prominent theoretical

framework used to explain the ongoing changes in family support to older people in Africa, some African gerontologists have criticized the theory, arguing that it is overly deterministic and simplistic (see, for example, Ferreira, 1999). Recently, contemporary African researchers have adopted alternative theoretical frameworks to better understand how ongoing economic and social changes are affecting older people.

For example, some researchers have suggested an alternative explanation for a possible decline in material support for older persons: rising economic hardship (Aboderin, 2004; Nyambedha et al., 2003). Sub-Saharan Africa faces a greater set of development challenges than any other major region of the world, and, on average, income per capita is now lower than it was at the end of the 1960s (World Bank, 2006a, 2006b). Although the region’s per capita gross national income has grown at a rate of around 3 percent per year for the past two years, it still stands at only $600 per year. An estimated 516 million people in the region are forced to survive on less than $2 a day, and 303 million on less than $1 a day (World Bank, 2006a, 2006b). In contrast to such countries as China and India, where substantial progress has been made over the past 5 years in combating poverty, the number of extremely poor people in sub-Saharan Africa has almost doubled since 1981 (World Bank, 2005). Consequently, sub-Saharan Africa is home to a growing share of the world’s absolute poor (United Nations, 2006). Thus, rather than emphasizing weakening filial obligations, some researchers have argued that declines in support for older people may simply reflect a growing incapacity on the part of the younger generation (see, for example, Aboderin, 2004). Increasingly, it appears that sub-Saharan African societies are being asked to cope with population aging and a catastrophic health crisis with neither a comprehensive formal social security system nor a well-functioning traditional care system in place.

The second largest and the second most populous continent in the world, the diversity across sub-Saharan Africa is apparent in the region’s physical geography and climate, in its plurality of cultural heritages, official and native languages, traditions, beliefs, religions, and value systems, in its modes of production and levels of economic development, and in its diverse social and political structures. Differences across countries and cultures make generalization from small-scale studies quite problematic, and the current lack of long-term comparable data from multiple sites hinders the ability to make meaningful cross-country comparisons. This heterogeneity also implies that cross-country comparisons are never going to be possible without careful longitudinal and multidisciplinary research designs. At the same time, because many countries are now in the early stages of adapting

to their changing population age structures and because current and prospective policy responses are likely to differ among countries, it may be possible to take advantage of this diversity to design a number of natural experiments enabling countries to learn from other’s experiences (see for example, Knight and Sabot, 1990). Thus, sub-Saharan African countries that are beginning to experience population aging may learn from those whose demographic shift began earlier. Similarly, the experiences of African governments that implement some form of social protection scheme can inform others contemplating a similar program.

A current list of local hurdles to research includes lack of access to adequate funding for research and lack of highly skilled local researchers to do the work, as well as sometimes difficult administrative barriers to carrying out research. Even in industrially advanced countries, amassing the resources required to undertake multifaceted research endeavors can be extremely complicated and time-consuming, but these difficulties are multiplied many-fold in sub-Saharan Africa. Many universities in the region have been badly neglected for decades. Their limited budgets are spent predominantly on (entirely inadequate) salaries, leaving few resources for maintaining facilities or equipment, purchasing computers or other basic office supplies, or initiating and sustaining a long-term research program. Consequently, much of the best research on aging in sub-Saharan Africa has been undertaken only with technical cooperation and foreign assistance from the international community. In the long run, it is essential to help sub-Saharan African countries develop their own research capacity by strengthening their universities and by augmenting the skills of their researchers.

The current dependence on international partners is not without its own controversies. While many donor agencies emphasize the importance of capacity building, the net result of their investments in this area in the past has often been quite modest (Commission for Africa, 2005). External researchers and funding agencies may design research projects that may not align closely with the priorities of national governments. They may also design them on a fairly short (2-year) time frame, which, while appropriate for accomplishing immediate research objectives, may not be sufficient to build sustainable in-country research capacity. Furthermore, while most people would agree that collaborations involving scientists from both the north and the south are vital to making sustained scientific progress, there is still a great deal of debate and uncertainty as to what constitutes “fair scientific partnerships” (Tollman, 2004). Essentially this refers to the way that different scientists and institutions allocate the benefits and obligations

of collaborative ventures, including allocation of funds, division of research roles, extent of data sharing, and method of ascribing authorship and related credits.

Even with good collaboration and adequate funding for research, investigators frequently are confronted with unwieldy procedures for obtaining permission to carry out the research. These may take the form of institutional review boards that are unfamiliar with social science research or government agencies that are slow to respond to requests for approval or multiple agencies that must be convinced of the value of the work to be undertaken and which may have agendas that do not give high priority to aging research.

The aging of individuals and populations and the changing position and well-being of older people in sub-Saharan Africa present a set of key challenges for African nations to begin to address. Yet evidence and a strong knowledge base of information on the nature and dynamics of poverty, health, social support networks, and the changing roles and responsibilities of older people and their implications are lacking. To fill this gap and provide information vital to the development of appropriate policies and programs, there is urgent need for an enhanced research effort on aging in sub-Saharan Africa, pursuing a variety of data collection strategies and analytic approaches.

Five key areas of research—all closely interrelated—emerged as essential to the advancement of understanding of the situation of older people in sub-Saharan Africa and as necessary precursors to the development of sound aging policy in the region. These are (1) income, wealth, and expenditure; (2) health and well-being; (3) the nature of family support and social networks; (4) the changing roles and responsibilities of older people as a function of the AIDS crisis; and (5) the nature and role of various kinds of formal and informal social protection schemes. Some discussion of each of these topics is provided below. In each case, a rationale for why a particular theme was included is provided, along with a discussion of measurement problems and selected findings from recent research on each topic, including findings from the papers included in this volume.

Critical to assessing the impact of programs intended to assist older persons in poverty is the ability to measure living standards and to monitor how they change over time. Even in the poorest sub-Saharan African agrarian economies in which the great majority of the population is absolutely

poor, inequality can be considerable, so that some households are substantially better off than others (House, 1991; House and Phillips-Howard, 1990; de Savigney et al., 2005). In such economic systems, the household is typically conceived of as the decision-making unit whose principal input is the labor of household members and the principal objective is to achieve self-sufficiency in basic food production (see Chayanov, 1925).

Although several countries have participated in the World Bank’s Living Standards Measurement Study or conducted other forms of ad hoc cross-sectional surveys of income, expenditure, and consumption to provide a snapshot of household living standards at a given point in time (see, for example, National Bureau of Statistics [Tanzania], 2002; National Statistics Office [Malawi], 2005), virtually no sub-Saharan African country regularly collects nationally representative household-level survey data to monitor trends in household income and expenditure over time. More typically, the period between surveys is 10 years or more, and in some sub-Saharan African countries there has not been a large-scale, nationally representative household survey for decades.

The measurement of the economic well-being of older people in sub-Saharan Africa is complicated by a number of conceptual and practical issues. Economists have traditionally relied on income as the most important indicator of economic status at a point in time. Measuring individual or household income in sub-Saharan Africa can be very challenging, however, particularly in communities in which a significant fraction of the population is still engaged in subsistence agriculture.⁶ In most surveys of income and expenditure, rates of nonresponse to questions related to earnings tend to be high and nonrandom, and sources of income that are infrequent (such as crops that are harvested only in periods of drought or cash remittances from distant family members) can easily be missed or underreported.

Further compounding these measurement problems, older people in sub-Saharan Africa typically do not live alone; they live with other family members (see van de Walle, 2006, and the papers by Lam et al. and Ezeh et al. in this volume). Households containing only older people constitute a very small percentage of households in sub-Saharan Africa (Kakwani and Subbarao, 2005). Resources that come into households are typically shared in some fashion among various household members, and it is often extremely difficult, if not impossible, to determine how these resources are divided (Deaton and Paxson, 1992). A common supposition in the development literature is that prime-age adults who bring money into a household

receive better treatment than older people or others who make only in-kind contributions to household well-being, but there is little empirical evidence to support this conjecture (Behrman, 1997; Deaton and Paxson, 1992). In many cases, surveys simply take the household as the unit of observation in which case any analysis of differential treatment of certain members in the household becomes impossible.

A further problem with using current income as a measure of economic well-being is that it may not be closely correlated with total wealth, which may be a better indicator of economic security or future consumption. Yet obtaining accurate information about wealth can be even more challenging than measuring current income. In many sub-Saharan African settings, household wealth can be gauged only in terms of possessions of basic household items (such as tables, chairs, beds, mosquito nets, shoes), housing type (i.e., the type of construction materials used to build the house), or the number of animals or amount of landholdings. In practice, these measures have not always been good predictors of consumption per adult (Montgomery, Gragnolati, Burke, and Paredes, 2000).

In sub-Saharan Africa, attention is often focused less on levels of household income and inequality than on the proportion of households that fail to surpass a certain poverty threshold. This is not as straightforward an assessment as one might first expect because, among other things, it requires a well-accepted definition of poverty. An enormous amount of work has been undertaken over the years to conceptualize and measure poverty. Perhaps the main controversy that has arisen over how to identify the poor relates to the issue of whether poverty should be considered an absolute or a relative measure. A common approach, the origins of which can be traced back to Rowntree’s study of the population of York, England, at the turn of the 20th century, is to identify a poverty line on the basis of nutritional standards, a basic diet, and the opinions of the potentially poor themselves as to their minimum requirements for expenditure on nonfood items (Rowntree, 1901). This “basic needs” approach has been widely adopted in studies on poverty in sub-Saharan Africa (see, for example, National Bureau of Statistics [Tanzania], 2002). But for policy-building purposes, it is also important to be able to deconstruct the various dimensions of poverty and to determine the prevalence of each of these among older people.

Kakwani and Subbarao (2005) recently investigated differences in the prevalence of poverty by household composition. Drawing on available recent household survey information collected over the period 1998-2001, the authors present a profile of older people in 15 low-income sub-Saharan African countries. The sample included countries in East and West Africa, Francophone and Anglophone countries, and countries with various levels of HIV prevalence and incidence. The sample countries can be taken as broadly representative of the region. The authors found that households

containing older persons only or older persons and children only have higher income shortfalls than households with no older people, and the differences are statistically significant in most cases (Kakwani and Subbarao, 2005). Furthermore, the size of the gap among households headed by older people is much higher than those not headed by older people. There are also significant rural-urban differences, with a much higher proportion of single older people who are poor in rural areas compared with urban areas in every country.

In summary, evidence on income, wealth, and expenditure in sub-Saharan Africa that is derived from (small-scale) household survey data currently has important limitations of coverage and content when used as a gauge of the economic status of older people (Barrientos, Gorman, and Heslop, 2003). Qualitative studies, such as those sponsored by the World Bank and others in the late 1990s and early 2000s, provide a basic picture of household living standards, but they do not focus explicitly on the condition of older people. When investigations have included older people, they reveal that poverty in old age is associated with poor access to paid work, basic services, and social networks, and there is a close relationship between older people’s ability to contribute to traditional roles and responsibilities and their ability to access support (Barrientos et al., 2003).

Income and expenditure studies also typically find that women have lower incomes than men, and that female-headed households are more prone to poverty than male-headed households. Building on that finding, the paper by Kuate-Defo in this volume uses unique data from Cameroon to examine the extent and nature of gender inequalities in health in later life and the extent to which these inequalities can be explained by differences in the socioeconomic characteristics and living arrangements of older women and men. Documenting important gender differences in self-rated health and functional limitations, the author found a strong association between poor health and low socioeconomic status. But more research is needed on the health and socioeconomic status of elderly women, particularly those who are widowed and childless, who may be especially vulnerable to poverty.⁷ Parts of sub-Saharan Africa are still heavily patriarchal, and in such societies the status of women can be very low. Women may lack certain rights of ownership to property, and inheritance is through the male side of the family (Toulmin, 2006). Consequently, upon their husband’s death, widows are at risk of dispossession of their house and land by their dead husband’s kin.

In sub-Saharan Africa, good health and nutrition are often emphasized as critical components of basic needs. In 1978, African delegates joined the representatives of other nations in endorsing the Declaration of Alma Ata, which committed all governments throughout the world to the common goal of achieving health for all by the year 2000. This goal included ensuring a life that was both long and free of a heavy burden of illness. It is well known that mortality rates increase at older ages, along with functional limitations and chronic conditions. But given the extremely limited resources available for health services in sub-Saharan Africa, most public health programs on the continent are far more concerned with eradicating or at least controlling preventable childhood diseases, such as measles and diarrhea, than they are with treating chronic diseases or managing the health care of the frail elderly. This is hardly surprising given that more than one out of every six children in sub-Saharan Africa dies before their fifth birthday (Lopez et al., 2006) and that the vast majority of these deaths are preventable with very low-cost interventions such as oral rehydration salts and vaccinations that cost just a few cents each. Furthermore, the HIV/AIDS pandemic continues to take an extremely heavy toll on the young adult population. In Southern Africa, for example, life expectancy at birth has fallen from 62 to 48, and it is projected to decrease further to 43 over the next decade (Joint United Nations Programme on HIV/AIDS, 2006). For most sub-Saharan African ministries of health, the challenges prior to old age are simply overwhelming.

Despite efforts in a number of sub-Saharan African countries to decentralize their health care budgets and realign health care expenditures to emphasize prevention rather than care, most countries still spend a significant fraction of their total health care budgets treating adult illness (Feachem et al., 1992; Poullier, Hernandez, and Kawabata, 2003; Tollman, Doherty, and Mulligan, 2006). The average sub-Saharan African country spends approximately 5.5 percent of gross domestic product on health care, of which perhaps half is spent on hospital care (Poullier, Hernandez, Kawabata, and Savedoff, 2003). Given that some African countries achieve substantially better population health outcomes than others on similar budget constraints, there is reason to believe that a substantial proportion of health resources in some countries could be better programmed (Murray and Evans, 2003; Laxminarayan, Chou, and Shahid-Salles, 2006). But in order to improve the allocation of resources, it is first necessary to understand the nature of health problems that given age groups face and how they are evolving over time.

From comprehensive assessments of the changing global burden of disease, sub-Saharan African governments can anticipate that, as their popula-

tions age, the observed pattern of disease will change: noncommunicable diseases, such as depression and heart disease, and injuries will become more important, and infectious diseases and malnutrition will become proportionately less important causes of disability and premature death (Lopez et al., 2006; Omram, 1971). In 2001, communicable, maternal, perinatal, and nutritional conditions accounted for 70 percent of the burden of disease in sub-Saharan Africa, while noncommunicable diseases and injuries accounted for 21 and 8 percent, respectively (Lopez et al., 2006). On the basis of 1990 data, Murray and Lopez (1996) predicted that communicable diseases will decline to around 40 percent of the burden of disease (as measured in disability-adjusted life-years) by 2020 in Africa, while noncommunicable diseases and injuries will grow. However, life expectancy in sub-Saharan Africa was 6 years lower in 2001 than it was in 1990 (Lopez et al., 2006), reinforcing the point that a considerable degree of uncertainty surrounds the estimates of the burden of disease there (see Cooper, Osotimehin, Kaufman, and Forrester, 1998; Mathers et al., 2006) and, in any event, may not reflect conditions that prevail among the poorest deciles of the population (Gwatkin, Guillot, and Heuveline, 1999). Furthermore, scholars are still divided on the extent to which sub-Saharan Africa’s health transition will traverse a unique path. In their paper in this volume, Kahn et al. document the rising burden of chronic disease and associated risk factors affecting older persons simultaneously, with worsening infections and illness affecting younger adults. Consequently, there is growing recognition that classic epidemiological transition theory is inadequate to explain the patterns of health and disease emerging in rural sub-Saharan Africa. The study of health in later life should also include understanding the changing physical and cognitive functioning of older people and its implications on the demand for health care services.

While there is a need to describe and monitor patterns of ill health and disability among older people in sub-Saharan Africa as well as monitor the health of other household members who may be under the care of older people and to better understand the relationships between health and poverty, there are numerous challenges to measuring health status in social surveys (see Thomas and Frankenberg, 2002, and the paper by Kuate-Defo in this volume for a more complete treatment of these issues). Assessing an individual’s health in the field can be difficult and add considerable expense to any survey, particularly if it involves the use of trained medical professionals. There may also be thorny ethical issues to overcome if blood or any other biological measures are to be collected, analyzed, and stored. Because they tend to be easy and inexpensive to collect and because they have been found in numerous contexts to be quite useful predictors of future mortality, self-reported responses to questions about health status tend to be the

mainstay of health status measurement in social surveys in other parts of the world.

One problem with self-reported health, however, is that it can be sensitive to the cultural context in which it is collected, so that cross-country comparisons of the relative extent of ill health in two different communities can lead to false conclusions. For this reason, over the past 10 years, researchers have attempted to increase the comparability of self-reported health status measures by the use of anchored vignettes (Murray et al., 2002, 2003; Salomon, Tandon, and Murray, 2003). Anchored vignettes may enable researchers to create a common metric to enable more reliable analysis of cross-cultural variations in ill health, thereby strengthening the reliability and usefulness of self-reported measures of health as a measure of well-being. (See the paper by Kuhn et al. in this volume for a further discussion of the pros and cons of using self-reported measures of health in social surveys.)

A potential challenge to using self-reported measures of health in sub-Saharan Africa has to do with the fact that illness and disease can be viewed as both cultural and as biomedical constructs. Consequently, there is a wide range of ways that people perceive and treat different illnesses in sub-Saharan Africa. For example, a recent analysis of Mozambican refugees’ understanding of stroke-like symptoms found that such symptoms are considered to be both a physical and a social condition (SASPI Team, 2004). Consequently, most people with stroke-like symptoms in that community seek treatment both from Western-trained doctors and from healers or prophets (SASPI Team, 2004).

Finally, in order to understand the determinants of health and disability among older people, health care researchers in sub-Saharan Africa not only need to understand the general socioeconomic, cultural, and environment conditions, but they also need to understand how patterns of health care utilization are changing over time. Sub-Saharan African governments typically struggle to provide health care services to older people in rural areas who still tend to have far poorer access to any kind of service relative to urban dwellers. Furthermore, in many sub-Saharan African communities, traditional medicine is often the only affordable and even the only available source of health care to large sections of the population. Across sub-Saharan Africa, traditional healers outnumber allopathic medical practitioners by more than 50 to 1 (Addae-Mensah, 2005).

The above discussion highlights the need for a better understanding of the magnitude and underlying causes of ill health and morbidity among older people in sub-Saharan Africa, how these patterns are evolving over time, and the implications of those changes for older people, their families, and patterns of health care utilization over time.

The projected increase in the absolute numbers of older people in sub-Saharan Africa, particularly the projected rise in the oldest old, as well as the projected growth in the proportion of the general population over age 60 are likely to have profound implications for families and kin networks. Hence a major question for policy makers is whether the traditional family in sub-Saharan Africa can cope successfully with the demographic, health, social, and economic changes that are taking place and can continue to provide older people with the range of support that they need. At the moment, traditional support systems, based on family and kinship ties, represent a way of life for most people in sub-Saharan Africa. Economically active adults in the family or kinship network provide support to children, older people, and others who are unable to care for themselves. Social protection is the natural outcome of commonly shared principles of solidarity, reciprocity, and redistribution in an extended family.

As discussed earlier, economic development and modernization are also associated with a range of economic and social changes that combine to weaken social networks that traditionally provide care and support in later life. Migration of young people from the rural areas to the towns, for example, can leave older family members geographically and socially more isolated. Formal education also provides a major counterpoint to the family in the socialization of the young. Schools have enormous potential to influence the values, expectations, and behaviors of the young and tend to weaken traditional social ties and obligations and place greater emphasis on independent thinking and autonomy, factors that tend to undermine traditional extended family systems (National Research Council and Institute of Medicine, 2005). In addition, the AIDS epidemic has placed enormous strains on traditional institutions.

To examine changes in societal arrangements for the support of older people, one starting point, albeit imperfect, that demographers have used extensively in other settings is trends in living arrangements. Such analyses can be particularly complex in sub-Saharan Africa due to the variety of household structures, including resident and nonresident household members and multiple household memberships (see van de Walle, 2006). Zimmer and Dayton (2005) examined data from Demographic and Health Surveys conducted in 24 countries and found substantial variation in household composition. The authors found that 59 percent of older adults in sub-Saharan Africa live with a child and 46 percent with a grandchild. The authors also found that older adults are more likely to be living with orphans in countries with high AIDS-related mortality (Zimmer and Dayton, 2005).

The net implications of economic development on the welfare of older

persons in sub-Saharan Africa are extremely difficult to determine. The general view among African researchers working in this area is that the situation of older people is getting worse. Although there are a number of reasons why this might be the case, not all social changes are necessarily detrimental. It therefore is an empirical question as to whether the situation of older people has improved or worsened. Migration and urbanization, for example, are frequently charged with leading to a breakdown of traditional family structures that weaken the position of older people. But in a largely subsistence agriculture economy with surplus labor, the absence of a family member due to migration may be a net loss or a net benefit to household welfare. Conventional microeconomic models of rural-urban migration are based on the premise that migration is an individual response to a higher expected urban income (Todaro, 1969). However, in sub-Saharan Africa labor migration may also be the outcome of a family or household-level decision to diversify sources of household income, lower total household risk, or a rational response to imperfect rural credit markets (see, for example, Stark, 1991, 1995). Work on intergenerational remittances in Botswana, for example, supports the view that migrants can sometimes be considered as members of a single household that is spatially split between two locations (Lucas and Stark, 1985).

More work is needed that builds on new as well as existing analytical frameworks in order to develop testable hypotheses related to the pathways and mechanisms by which the forces of modernization and change are affecting traditional social relations.

No analysis of the situation of older people in sub-Saharan Africa would be complete without acknowledgment of the fact that the region is experiencing a very severe HIV/AIDS pandemic with important implications for older people. The region is home to more than 60 percent of all people living with HIV, around 25.8 million in 2005 (Joint United Nations Programme on HIV/AIDS, 2006). And since the start of the epidemic, HIV/ AIDS has infected roughly 50 million Africans, of whom more than 22 million have already died. Even if new transmissions were halted today, millions of Africans who are currently infected would still develop AIDS and die over the next 5 to 10 years. As noted above, in some southern African countries, life expectancy has fallen from 62 years in 1990-1995 to 48 years in 2000-2005 and is projected to decrease further to 43 years over the next decade (United Nations, 2005).

While it is important to remember that HIV/AIDS prevalence remains quite variable across the region, where it is prevalent, the epidemic can affect the lives of older people in many different ways. Knodel et al. (2003)

have developed a framework to examine a broad range of potential pathways through which AIDS can adversely affect the well-being of parents of adult children with AIDS and their possible demographic, psychological, economic, or social consequences (see Table 1-2). Although developed based on the authors’ experience with AIDS in Asia, it offers a useful starting point for formulating the issues in an African setting. Older people most directly affected by the epidemic are likely to be either infected by the virus themselves or parents of infected adult-age children. Not only do parents of adult children with AIDS have to face the pain of seeing their own children suffer and die, but they can also face serious economic hardship, both in the short term due to unforeseen medical expenses and funeral costs and in the long term due to being deprived of one of their primary sources of economic support in old age. In addition, older people, particularly elderly women, are also frequently left to care for grandchildren. Thus these older people face the double burden of having to replace lost sources of income while supporting additional family members.

Older people may also suffer indirect health consequences, such as the mental and physical fatigue associated with caregiving, additional labor force participation, exposure to TB or other opportunistic infections brought into the household by the person with HIV, and stigmatization and isolation following the death of household members due to AIDS (Dayton and Ainsworth, 2004; Knodel et al., 2003; World Health Organization, 2002). AIDS is also likely to have important indirect impacts on society in general, such as the demand and availability of health services, per capita income growth, and macroeconomic performance (National Research Council, 1996).

Finally, older people may be directly affected by the virus themselves. Approximately 6 percent of officially reported AIDS cases in Africa in 1999 affected people age 50 or older (Knodel et al., 2003). (The official number of AIDS cases worldwide is widely acknowledged to represent merely the tip of the iceberg with regard to the true extent of the epidemic, but it is the only readily available source of data for comparing caseloads by age.) But with more than 1 in every 10 adults in Africa estimated to be HIV positive, far more older people in Africa are being affected indirectly by sickness and death among the younger generation, which can have many direct and indirect consequences for their material well-being.

While most of the attention to date has focused on how the pandemic affects persons with HIV and their surviving orphans, greater recognition needs to be given to the consequences of the pandemic for older people, who in many cases are playing critical roles as caregivers for the sick and guardians for orphaned grandchildren left behind. There is still very little research on the impact of HIV/AIDS on older people in sub-Saharan Africa. The papers by Clark, Merli and Palloni, and Hosegood and Timaeus use a

variety of research strategies to help fill this gap. The chapter by Hosegood and Timaeus, for example, uses micro-level household data to establish a baseline from which they can examine the impact of HIV on older people in rural KwaZulu-Natal in South Africa, the country in the region with both the highest proportion of old people and one of the severest AIDS epidemics. The authors point out that, partly as a function of a generous social pension program and high rates of rural unemployment and underemployment, older people have maintained their traditional role as a major resource in society, caring for children and maintaining rural households even prior to the HIV epidemic. A recent study of the experiences and needs of older people in Mpumalanga, South Africa, supports this contention. The authors of that study found that almost one in three older people are either now caring for sick adults living in the household or are raising grandchildren whose own parents are either dead or away in the cities on a long-term basis (Makiwane et al., 2004). And 60 percent of all orphans in Mpumalanga are being cared for by their grandparents (Makiwane et al., 2004).

In the absence of detailed longitudinal data, simulation models can be employed to provide critical insights into how the AIDS epidemic may impact certain key demographic variables, such as residential patterns and kinship networks over a protracted period. In their chapters in this volume, Clark and Merli and Palloni use micro- and macromodel-based approaches to examine the likely number of orphans, residential patterns, and kin relations that might result from a severe epidemic.

While policy makers in sub-Saharan Africa are becoming increasingly aware of the needs of older people, there is general agreement that the types of social welfare programs in place in other parts of the world are too expensive to replicate in sub-Saharan Africa given the size of their economies (Kalasa, 2001). Thus there is a need to search for alternative approaches that might achieve a similar function but at lower cost.

The concept of social protection is one that is gaining increasing attention in development circles as a useful policy framework for addressing issues of poverty and vulnerability (Garcia and Gruat, 2003). Traditionally in sub-Saharan Africa, social protection for older people is provided by both formal and informal programs and practices that have been developed to reduce poverty and vulnerability in old age. But with per capita income

below a few hundred dollars in most sub-Saharan African countries, it is no surprise to find that formal social security systems across the region cover only a small fraction of the population. Except for Mauritius and the Seychelles and a few countries in Southern Africa, including South Africa, Botswana, Lesotho, and Namibia, all of which operate social pension schemes aimed at comprehensive coverage, most countries’ formal social security programs never reach the urban or rural poor. Except for these few countries, the extended family unit remains the main source of support for the vast majority of older people in sub-Saharan Africa when they can no longer work.

In many Western countries, formal social security is an important policy instrument for governments to redistribute wealth, combat poverty, and reduce inequalities between various segments of society. But in sub-Saharan Africa, current social security schemes are extremely marginal both in terms of percentage of the labor force that is covered and the size of pensions that are received. In most sub-Saharan African settings, national social insurance schemes cover less than 5 percent of the labor force and expend less than 1.5 percent of their gross domestic product on pensions (Fox and Palmer, 2001). Consequently, in the majority of countries in sub-Saharan Africa, social protection programs have a very modest effect on poverty alleviation. The largest social protection programs for older people in sub-Saharan Africa are occupational pension schemes, but these typically cover only people who have worked in the public sector, in state enterprises, or in large private firms in the modern sector. The self-employed, workers in the informal sector, domestic workers, and the vast majority of the population living in rural areas and engaged in subsistence agriculture or other forms of subsistence living, such as nomadic pastoralism, are still excluded from formal social security schemes and must rely on their families for support and protection when they can no longer work.

Bailey (2004) identifies several distinct patterns of social protection schemes that have developed in sub-Saharan Africa. Even though most countries did not introduce programs until after their independence, most schemes have been strongly influenced by their countries’ colonial heritage, with the types of programs in Anglophone Africa differing from those in Francophone Africa. At one end of the spectrum are countries, for example South Africa, that have introduced schemes aimed at near universal coverage. Other countries currently provide no form of social security, either because nothing has been set up yet or because previously established schemes have been dismantled or disrupted for various reasons, including, not infrequently, armed conflict. Apart from government schemes, volun-

tary private pensions can also be found in many countries, although, again, their coverage tends to be restricted to formal sector workers. Most sub-Saharan African pension schemes are financed by contributions made by both employers and employees, with the contribution rate in most cases being higher for the employer. In the case of South Africa’s social pension, the scheme is financed through general tax revenue. Given the structure of the schemes and the nature of the labor force in most sub-Saharan African countries, the vast majority of those actually covered by formal social security schemes are neither the poorest of the poor nor women.

In West Africa, several Francophone countries established a voluntary plan during the colonial period for government employees: for example, the West African Retirement Pensions Fund was modeled on a program for French civil servants that linked benefits to length of service and average earnings (Bailey and Turner, 2002). Even today, Senegal has a social security program that determines benefits through a formula that is quite similar to the system used in France (Bailey and Turner, 2002). Cote d’Ivoire, Mali, and other countries in the region have similar defined-benefit programs, with workers contributing between 4 and 9 percent of their earnings to the schemes (Bailey and Turner, 2002).

Social security programs in the countries that were former British colonies are generally more modest than those in Francophone Africa. In several former British colonies, provident funds, such as the Nigerian National Provident Fund, were established. These were seen as relatively easy to operate and amounted to compulsory interest-bearing individual savings accounts for workers that were financed from contributions from both employees and employers (Bailey, 2004). Unlike most social security programs, which typically offer survivor and disability benefits, most provident funds generally provide only a single lump-sum amount at retirement. Generally the level of the lump-sum payment is extremely modest and cannot actually support anyone in retirement. In 1993-1994, for example, retirees enrolled in the Zambian National Provident Fund each received, on average, a lumpsum payment of around US$10 (Mukuka et al., 2002). In a number of countries, for example, Ghana, Nigeria, and Zambia, these early provident funds have now been converted to defined-benefit social security systems.

In some countries, including Sierra Leone, Eritrea, and Somalia, efforts to introduce schemes have been stalled by armed conflicts. In other places, for example Liberia and the Democratic Republic of the Congo, whatever social security programs that once existed have been effectively dismantled and destroyed by armed conflicts.

Finally, a few countries have universal social security programs. Botswana, for example, has a universal flat-rate pension scheme for all residents over the age of 65. South Africa has a means-tested benefit for women age 60 and over and men age 65 and over, and Mauritius offers a basic

pension to all residents age 60 or older with supplemented earnings-related benefits (Bailey, 2004). The South African pension scheme was introduced in 1928 as a measure to provide for the poorest retired white workers. The State Pension was extended to all South Africans in 1944, and the value was equalized for all segments of society in 1993 shortly before the first democratic elections in 1994.

In addition to various types of occupational pension schemes that are contributory, some sub-Saharan African countries administer minimal public assistance or social welfare assistance programs to the needy. In Zimbabwe, for example, the government operates a Social Welfare Assistance Scheme that provides assistance to older persons, persons with disabilities, and the chronically ill (Kaseke, 2004). A similar scheme operates in Zambia, the Public Welfare Assistance Scheme, providing benefits to such vulnerable groups as older persons, widows, and the unemployed (Kaseke, 2004). These schemes are far less well documented than contributory pension schemes or provident funds: the amount of assistance provided is typically very small and the coverage of these programs is generally extremely low (Kaseke, 2004).

Formal social security schemes in sub-Saharan Africa are riddled with a number of well-known problems, including low coverage of the labor force, corruption, and inadequate benefits that are not indexed for inflation. Where occupational pension schemes or provident funds exist, they are only available to a small percentage of workers who have regular paid employment in the formal sector of the economy. In Tanzania, for example, Mchomvu et al. (2002) estimate that formal social security schemes currently cover only 6 percent of the population and about 5 percent of the active labor force, the majority of those covered being men. Schemes are also frequently characterized as suffering from poor and inefficient management and much bureaucracy, leading to high administrative costs, lack of transparency, low or even negative real rates of return on investments, and delays in receiving benefits. In some cases, the number of staff employed to administer the programs is so large that the administrative and management expenses of the fund exceed the investment income. At one point, the Zambian National Provident Fund, the main public scheme in Zambia until January 2000, had more than 1,300 employees. In 1995, administrative expenditures accounted for more than 100 percent of total revenue from contributions (Mukuka et al., 2002). In other cases, the bureaucratic machinery to administer these programs is so unwieldy that it severely hampers the effectiveness of the program. In Uganda and Zimbabwe, for example, the manual processing of claims, combined with the many stages

that a claim needs to go through before payment is dispersed, leads to long delays for legitimate recipients waiting to receive their benefits (Dhemba et al., 2002; Kasente et al., 2002). Rates of interest awarded annually to members of provident funds have invariably been negative in real terms, and the lump sums paid out generally represent no more than a few months’ earnings (Mchomvu et al., 2002). In addition to these problems, the devaluation of local currencies necessitated by structural adjustment programs has severely eroded the value of benefits in some countries (Kaseke, 2004).

No discussion of social protection of older people in sub-Saharan Africa would be complete without a description of the South African social pension program, which is quite unusual in sub-Saharan Africa with respect to its level of coverage and generosity of benefit (Case and Deaton, 1998). South African women age 60 or older and South African men age 65 and older may apply for a state pension irrespective of employment history. In 2006, the state pension was R820 per month (roughly US$115). The pension is means-tested, but the level is set at a point at which 80 percent of all age-eligible Africans may receive the pension (Lam et al., in this volume). Essentially a by-product of the dismantling of the apartheid system, the program, which was originally designed to provide protection for poor whites, is viewed in South Africa today as a way to achieve several broad development goals: providing assistance to households in rural areas, targeting women, and keeping significant numbers of households out of poverty (Ardington and Lund, 1995). It is the sole or major source of income for many poverty-stricken families (van Zyl, 2003).

A considerable body of research has been conducted on the effects of such a large and generous transfer scheme on the welfare of older people and extended family members (Alderman, 1999; Ardington and Lund, 1995; Bertrand, Mullainathan, and Miller, 2003; Case, 2001; Case and Deaton, 1998; Duflo, 2003; Ferreira, 2003; Posel, Fairburn, and Lund, 2004; Lam et al., in this volume). Case and Deaton (1998) argue that the pension program is effective in reaching the poorest households and in fact is a useful tool for reaching the poor in general, not just older people. Because so many older people in South Africa’s African population live in multigeneration households, in part because young people tend to join households that receive a pension, the state pension program transfers money into households with children. Roughly one-third of all children age 4 and under live in households in which older people receive pensions, and the percentage of children living with pensioners is even higher among the poorest income quintiles (Alderman, 1999). Case (2001) also found that in the Western Cape, in households that pool income, the state pension ap-

pears to protect the health status of all adults and children in the household. Duflo (2003) found that the impact of the program depends on the gender of the recipient: pensions received by men have little effect on children’s health status, but pensions received by women have a large impact on the physical stature of girls.

An important policy question with respect to any social welfare program has to do with the extent to which the program creates dependency and has a negative effect on labor supply. Betrand et al. (2003) found that pensions can have a negative effect on the labor supply of working-age adults residing in pension-receiving households. However, Betrand et al. (2003) investigated only the labor supply of adults resident in the household. Posel et al. (2004) argued that the social pension also affects the propensity of household members to migrate to find work, which acts as a positive supply response to the receipt of a social pension. In their chapter in this volume, Lam et al. contribute to this debate by examining how the social pension affects the decision to withdraw from the labor force by older people. By analyzing census and survey data, the authors found that, although the pension is associated with high rates of withdrawal from the labor force, the rates are somewhat less sharp than those observed for similar programs in Europe.

Another crucial policy question in the region is whether economic growth is needed in order to broaden the social safety net. Kakwani and Subbarao (2005) investigate the likely fiscal implications of providing some sort of social pension to older people in various sub-Saharan African countries and study the impacts on poverty rates. The authors have found that the fiscal cost of providing a universal noncontributory social pension to all of older people in sub-Saharan Africa would be quite high, around 2 to 3 percent of gross domestic product, a level comparable to—or even higher than—the current levels of public spending on health care in some sub-Saharan African nations. The authors argue that the case for universal social pensions also appears to be weak on welfare grounds, inasmuch as there are other groups competing for scarce safety net resources (such as families with many children) whose incidence and prevalence of poverty is much higher than that of older people.

Given that a universal social pension program appears out of reach in most countries and is difficult to defend on purely social welfare grounds, the authors then explore various options for targeted social pensions using a fixed budget constraint of 0.5 percent of gross domestic product and a fixed benefit level of 70 percent or 35 percent of the poverty threshold for older people defined as age 60 or 65 and older. Two household types were

considered: households with older people living with children and households with older people only. The authors found that the introduction of social pensions targeted to these groups would yield considerable reductions in the prevalence of household-level poverty, both for the targeted groups and for the national average. Nevertheless, as the authors point out, the operational feasibility of such a program is very weak. The administrative burden of operating such a scheme is enormous and would be likely to lead to dissolution and reformation of certain types of households in order to make them eligible to claim a pension. Bearing this is mind, the authors also investigate the fiscal implications of providing a social pension to only poor older people, regardless of the type of household in which they reside. The authors conclude that the best option is to target the pension only to the poor, keep the benefit level low, and the age of eligibility at 65 and older (Kakwani and Subbarao, 2005).

Given the problems and formidable financial and administrative hurdles to expanding formal social security schemes in sub-Saharan Africa, policy makers also need to know whether there are ways to expand and support any of the various forms of informal social protection schemes that exist around the continent as a means to provide a vital safety net for certain vulnerable populations. A wide variety of informal community-based arrangements have been developed in rural areas aimed at spreading risk among friends and extended family members, with neighbors, or with other participants. These can often involve self-help or community based-initiatives that draw on sub-Saharan African traditions of shared support and kinship networks. In parts of Zimbabwe, for example, the government has successfully reintroduced the concept of the Zunde raMambo (literally “the Chief’s Granary”), which refers to the harvest from a common field that is stored in a common granary and used at the discretion of the chief in order to ensure that the community has sufficient food in the event of a drought or a poor harvest (Dhemba et al., 2002).

There are many other examples of groups that have come together as spontaneous responses to poverty. Rotating savings and credit associations and mutual aid societies, for example, are commonly used throughout sub-Saharan Africa to compensate for failures in existing formal financial markets. In rotating credit and savings associations, participants periodically contribute fixed amounts of money and allocate the fund on a lottery or rotational basis to its members. The scheme encourages small-scale capital accumulation and savings and allows members to meet various welfare objectives, such as to pay school fees, meet medical expenses, or buy food. Funds can also be used to start or promote small businesses and acquire

assets, including livestock (Kimuyu, 1999). Burial societies are another form of rotating savings scheme. In burial societies, members pay periodically to the society, and, when the member dies, the family receives money to help offset the funeral expenses. These types of scheme are very popular in sub-Saharan Africa, particularly in urban areas. Much less is known about informal social security systems in sub-Saharan Africa than about formal social security systems, but it is generally believed that informal schemes also suffer from a number of chronic problems and in their current form fail to provide much in the way of long-term protection against various forms of risk (Mchomvu et al., 2002). Nevertheless, there is a need for more detailed country-specific analysis on the nature of both formal and informal schemes, the size and frequency of transfers, and the redistributive effects of those transfer payments on the health and well-being of older people and other household members.

Sub-Saharan African policy makers are increasingly aware of the challenges associated with population aging and with the changing needs and contributions of older people. While aging may not soon have highest priority, such actions as the development and adoption of the African Union Framework and Plan of Action on Ageing, the formulation of the African Common Position on Ageing, and the establishment, for the first time, of national policies on aging in a few sub-Saharan African countries are all indicators of a growing awareness of aging issues across the continent.

Sound understanding of the links between key social and economic trends and the economic security, health and disability, and living conditions of older people in sub-Saharan African contexts is essential if appropriate new policies to enhance their lives are to be established. The recommendations below take into account the research essential to this new understanding, the need to overcome barriers to research, and translation of research findings into programs and policies. The substantive agenda for research was laid out in the previous section on key areas for future research. There are no easy solutions to the problems discussed in this report, but unquestionably understanding of some of the key issues and causal processes we have discussed would be greatly improved if the research community had the resources to use available information and undertake new data collection efforts, particularly those with a repeated sampling or longitudinal design. The consensus of the panel is that this type of longitudinal, multidisciplinary monitoring system would be most useful if implemented in several locations using comparable design. It would not only provide a reliable benchmark on the current socioeconomic situation of older people but would doubtlessly also contribute significantly to scientific knowledge.

In turn, it could inform those charged with the development of new programs and policies for older people. The returns on such an investment may be modest initially but will accumulate over time.

The panel also wishes to emphasize the importance of facilitating research, building local research capacity, and supporting the development of a local research network in sub-Saharan Africa that can support essential studies on the nature and consequences of its population aging and the context in which it is occurring. Top priority for the immediate future should be given to building basic research infrastructure, improving access to data, removing burdensome administrative barriers to carrying out new research, and strengthening international collaboration.⁹

After consideration of the general state of knowledge about aging in sub-Saharan Africa, recent research developments and emerging opportunities, and the strength of local research capacity, the panel arrived at the following recommendations, grouped under research agenda and funding, enhancing research opportunity and implementation, and translation of research findings that they feel could help improve the future development of the field.

1. Increase Research on Aging in Sub-Saharan Africa

This report provides detail on the substantive agenda of needed research on aging in sub-Saharan Africa. There can be little doubt that, as a function of the emerging fertility transition in sub-Saharan Africa, the changing macroeconomic climate, and the impact of the devastating HIV/ AIDS epidemic, researchers are paying increasing attention to the social, economic, and demographic dimensions of aging in sub-Saharan Africa. This attention can be turned into action only with increased funding directed to this arena. Funders should consider mechanisms—existing or new—to enable the research agenda identified in this report to be carried out. These mechanisms should foster international collaborative research in ways that benefit researchers in both the developed and developing worlds.

1. Explore Ways to Leverage Existing Data Collection Efforts to Learn More About Older People in Sub-Saharan Africa

The ongoing economic and social changes taking place in sub-Saharan Africa as well as the projected changes in both the numbers and proportion of older people in sub-Saharan Africa pose a series of vital policy challenges: How are ongoing economic and demographic changes affecting the family structures, socioeconomic position, and health of older people? How is HIV/AIDS changing the roles and responsibilities of older people? Can some form of social protection scheme be designed and successfully implemented in this part of the world that will partially relieve some of the burden on sub-Saharan African families?

With such vagaries and uncertainties there is a clear need to enhance understanding of the current situation of older people in sub-Saharan Af-rica as well as to improve understanding of some of the underlying causal processes that relate social and economic change to older people’s wellbeing in the sub-Saharan African context. Given that the proportion of the population that is older is still low, at least relative to other continents, sub-Saharan African policy makers have an important window of opportunity in which to act. Furthermore, as an earlier report of the National Research Council’s Committee on Population pointed out, each sub-Saharan African country’s response to the challenges of aging is liable to be slightly different. Consequently, a number of natural experiments are either already currently under way, or shortly will be that, provided they are well recorded and documented, could be used to enable countries to learn from each other’s experiences (National Research Council, 2001). In order for this to happen, national governments will need both to invest much more in basic research and to develop mechanisms to establish common definitions that will facilitate the harmonization of data collection across countries.

Improving understanding of the situation of older people will also require a better picture of the simultaneous interplay among multiple factors, including health, economic, and social characteristics. Hence, the development and use of multidisciplinary research designs will be essential in the development and production of any new data on aging in sub-Saharan Af-rica. Furthermore, the very strong a priori assumptions that many researchers and social commentators hold concerning the deteriorating situation of older people over time imply the need for study designs that can trace the experiences of individuals over time. Experience from the United States, Europe, and Asia has shown that data collection efforts that use a multidisciplinary panel approach, involving researchers who are willing to work across traditional domains, can produce significant returns (Börsch-Supan et al., 2005; Hermalin, 2002; National Research Council, 2001). Ideally, data should be reliable, population and community based, and in-

clusive of all groups, should cover multiple domains of interest, and should be collected both prospectively and continuously.

All of these challenges suggest the need for establishing a foundation of high-quality baseline data and tracking changes in many key variables over time. But collecting high-quality longitudinal data would undoubtedly be an extremely expensive and difficult undertaking. A more feasible first step may be to take advantage of already ongoing data collection efforts. The existing network of community-based population surveillance sites (INDEPTH) offers one likely vehicle for developing such a data collection effort, rather than investing in a completely new sampling framework. Although there are some inherent limitations of site-specific studies, greater investment in a growing number of (predominantly rural) INDEPTH sites around the continent is likely to substantially enrich knowledge of the living arrangements, economic activities, and health status of older people in sub-Saharan Africa if these data are made available and analyzed in a timely fashion. In general with ongoing data collection efforts, it is important to find a balance between protecting confidentiality and increasing access by qualified researchers to these valuable data. Sub-Saharan African researchers are likely to benefit most from greater access to African censuses, surveys, and demographic surveillance site data.

1. Improve Support for Library Infrastructure and Dissemination Tools to Create a More Integrated Body of Knowledge in Sub-Saharan Africa

Researchers and policy makers in sub-Saharan Africa are often poorly informed about previous research that has taken place on their continent or elsewhere. The only sub-Saharan African journal dedicated to publishing the findings of research on various aspects of aging, the Southern African Journal of Gerontology, ceased publication in 2000 due to lack of financial support. Furthermore, even when studies are accepted and published in international journals, it is often quite difficult to obtain copies of papers locally. In fact, it is often far easier to obtain copies of research papers outside the country in which they were produced than inside it. In addition, researchers in sub-Saharan Africa are working in three main languages: English, French, and Portuguese, which slows down professional networking and the dissemination of findings. In most sub-Saharan African countries, there is no up-to-date bibliography of research or reports on aging to form a knowledge base, such as was compiled for aging research in Europe in the late 1990s (see Agree and Myers, 1998). Furthermore, there are few opportunities for national and international networking among scholars

interested in aging although the recent establishment of the African Research on Ageing Network (AFRAN) may lead to more opportunities in the future. Hence, there is a need for more support for library and information services as well as a need for greater information sharing and professional networking, perhaps through the sponsorship of more local or regional conferences. Given the rapid takeoff of electronic journal retrieval systems, such as JSTOR, increased investment in Internet access—both to high-speed Internet itself and to rights to use resources available on the Internet—may be one of the most effective means of closing the gap between continents in terms of access to existing research.

1. Improve Archiving of Past Censuses and Surveys

Generally speaking, African censuses and surveys have been greatly underutilized and much survey data collected over the past 30 years has deteriorated as a result of poor archiving. Yet even with the limited focus on aging issues in past social surveys, there may be significant potential for furthering knowledge of the social and economic conditions of older people from a more systematic analysis of previously collected data. But that cannot happen without improvements to the ways that data sets are archived and put into the public domain. Data handling and storage technology advance so rapidly that the burden of making data available in a useful format cannot rest with individual researchers. Hence there is a need for a more systematic archiving of sub-Saharan African microdata. The World Bank’s web-based African Household Survey database, the Minnesota Integrated Public Use Microdata Series (IPUMS-International), and the University of Pennsylvania’s African Census Project are good examples of initiatives designed to save previously collected data from destruction that have eased data constraints and produced new findings about older people (see, for example, Mba, 2002).

1. Improve Access to Ongoing Data Collection Efforts

If investments in new data are to be realized, better mechanisms will need to be put in place to improve storage, retrieval, and access to aging data. The experience of the Health and Retirement Survey and the Asset and Health Dynamics of the Oldest Old Study in the United States has shown that the return on research dollars is highest when the data collected are made available to the broad scientific community in a timely fashion (National Research Council, 2001; Willis, 1999). Yet in the sub-Saharan African context this may be far easier said than done. First, there are a number of ethical issues that need to be explicitly addressed when collecting any individual-level data (Cash and Rabin, 2002). But particularly in the

case of such sensitive topics as the HIV status of respondents or their other family members, it is quite easy to see how information collected by researchers could be damaging both to the individual and to others if it were disclosed to a third party. Different countries have different policies in place to protect the privacy and confidentiality of their informants. Issues of confidentiality may require developing complex informed consent procedures that may be difficult to devise and communicate when the concepts are new and foreign and the population being investigated is poorly educated.

Furthermore, much of the best research undertaken to date has been made possible only by the establishment of international research partnerships between researchers in the developed and the developing world. These partnerships are quite complex to establish and maintain, since they involve negotiating such thorny issues as fair allocation of research roles, balance in infrastructure investments, and fairness in ascribing authorship and related credits (Tollman, 2004). At the same time, high functioning North-South institutional partnerships can accomplish a great deal with regard to research training as well as research and may well offer the best prospects for the foreseeable future. Thus, the challenge will be to strengthen these partnerships in ways that both support local institutions and increase timely access to data.

1. Strengthen International Collaboration and Capacity-Building in the Short Term

There is a critical and urgent need to strengthen research capacity in sub-Saharan Africa. More and better research on various dimensions of aging in sub-Saharan Africa cannot happen without an increase in the funding for research, more well-trained local researchers, and improvements in administrative procedures that currently hinder the execution of research projects.

Many sub-Saharan African universities were badly neglected in the 1980s and 1990s. Funding for salaries, maintenance of facilities and equipment, library services, and sometimes even basic office supplies was often entirely inadequate, resulting in the demise of sub-Saharan African universities and the widespread flight of faculty into the private sector (National Research Council, 1996). At the same time, a lack of managerial and administrative capacity can often lead to inefficiencies in the way that available money is allocated.

Although there are signs that African governments are beginning to value the role that science and technology can play in national development, it is unlikely that sub-Saharan African governments are going to increase their level of research funding substantially in the near future. In the short term, foreign financial and technical assistance will remain essential

to the development of local universities and the strengthening of local research capacity. Donors should explore funding mechanisms that increase incentives for work in this area, perhaps through the establishment of special funding mechanisms, particularly those that not only advance funding for research on aging in sub-Saharan Africa but also encourage cross-national collaboration and training.

At the same time, African and other governments should do all in their power to facilitate linkages between sub-Saharan African institutions and international research centers in the United States and elsewhere by establishing agreements at the highest levels to expedite local review of projects as expeditiously as possible. Such linkages, especially if built on the basis of a strong mutual interest in collaborative research, can help local universities develop, can assist local researchers by providing funding and in-country technical assistance and training, and can help with the processing of data and the preparation of manuscripts for publication. Experience in a number of sub-Saharan African settings has demonstrated that such collaborations can lead to important scientific advances and be mutually beneficial to all institutions involved (Tollman, 2004).

1. Remove Barriers to Implementation of Research

Collaborative research with sub-Saharan African institutions requires approvals by all collaborating institutions, frequently involving more than one review board for the ethical conduct of research, as well as government offices and officials in the country in which the data collection takes place— whether representatives of the United States or the local government. The panel received informal reports of difficulties and delays in receiving the required approvals even for projects funded by the U.S. National Institutes of Health (NIH) after the usual exacting NIH reviews. One study was delayed for more than a year; another was delayed then finally dropped. If research on these important subjects is to be carried out in a timely fashion, it is essential that institutions, review boards, and government bodies at all relevant levels establish procedures and processes that make speedy review possible, without repetitive reviews of scientific merit.

1. In the Long Term, Sub-Saharan African Governments Must Give Reasonable Priority to Aging Research and Strengthen Local Research Capacity

In the long run, the importance of foreign-supported research in the region must be reduced. It will be up to sub-Saharan African national governments to prioritize aging as a focal area and to find the resources needed to be able to drive the region’s aging agenda. Numerous related processes

are already under way in Africa aimed at strengthening research institutes and building research capacity, including programs at the Council for the Development of Social Science Research, headquartered in Senegal, and the African Centre for Research and Training in Social Development, headquartered in Ethiopia. Nevertheless, sub-Saharan African governments generally need to place a greater value on the role of higher education and find funding to rebuild and strengthen local universities. In many countries, pay scales will need to be adjusted in order to attract and retain the best researchers.

1. Improve Dialogue Between Local Researchers and Policy Makers

There is an ongoing need for continued and expanded dialogue between the research and the policy communities. Researchers need to do a better job of drawing out the main policy and programmatic implications of their work, and policy makers need to better articulate what information they most need for more effective planning and program design. At the same time, there is also value in both sides engaging with older people themselves to ensure so that they are not excluded from a dialogue aimed ultimately at enhancing their future well-being. Otherwise, the danger is that local programs and policies will be only marginally based on a solid understanding of local needs and conditions, while research will continue to be undervalued by policy makers and therefore underfunded.