The workshop began with a review of fertility trends in the sub-Saharan region and their consequences. The presentations highlighted the uniqueness of the African fertility transitions, the variability of the trends between the countries in the region, and the possible consequences of these trends for the region.
John Bongaarts of the Population Council and Ann Biddlecom of the Population Division of the United Nations discussed factors that make the fertility transitions in African countries unique. Bongaarts provided context for understanding how fertility rates in Africa have followed patterns different from those observed in other countries, and Biddlecom provided additional perspective on these trends by examining differences within the sub-Saharan countries and considering possible scenarios for the future.
Bongaarts began by noting that fertility rates in the United States and many European nations declined significantly in the 19th and early 20th centuries, in some cases from as many as 8 births per woman to below 2. This trend was an important factor in their economic growth, and this set of developments is referred to as a fertility transition (Guinanne, 2011). For developing nations in Asia and Latin America, Bongaarts noted, this transition began later, in the mid-1970s. In sub-Saharan Africa, it did not begin until the 1990s and proceeded more slowly than in other develop-
ing nations. Currently, Bongaarts noted, the total fertility rate (TFR) is just above 5 births per woman for the African region, as opposed to under 3 elsewhere.
Bongaarts used four development indicators to help explain fertility trends in sub-Saharan Africa between 1970 and 2010: gross domestic product (GDP) per capita, percentage of the population attaining at least a primary education, life expectancy at birth, and percentage of the population living in urban areas. As the graphs in Figure 2-1 show, development has proceeded in all regions, but the African nations were at lower levels than the other countries for each of these variables as they began their transitions, and are at significantly lower levels today.
Bongaarts also noted that the pace of the transition in sub-Saharan nations has been slower than the pace in other developing countries, and
FIGURE 2-1 Comparison of African and other developing nations on four variables.
NOTES: PPP indicates GDP converted to show purchasing power parity rates. LDCs are least developed countries. Circles indicate timing of onset of fertility transition.
SOURCE: Bongaarts (2015).
that the pace of progress in each of the other four variables has also been slower.
The comparatively high level of fertility at a given level of development in African nations has been called the “Africa Effect,” Bongaarts noted, as researchers have sought explanations for the difference between regions in their fertility response to development.1 He noted that because the fertility transition in the region occurred later in time (though early in relation to other economic developments in the region) and at a slow pace, the overall Africa Effect for fertility is substantial. He also indicated that this is partly because of the relatively slow pace of development there compared with that of other regions at the time of their transition onset. Characteristics of this region may help explain the differences, he added, as theories of African exceptionalism have suggested.
Socioeconomic development, he explained, raises the cost and decreases the benefits of having children and it reduces mortality, especially among children. These changes in turn often lead parents to want smaller families. In many African countries, both economic and cultural traditions have strongly favored larger family sizes. Where family planning programs have been weak or nonexistent, there has been little encouragement for increased contraceptive use. Bongaarts noted that there has been variation in these factors within the region and that family planning investments in several countries have been successful. He also noted that a full accounting for the Africa Effect would require a detailed examination of the history of colonialism and other historical developments.
Biddlecom noted that trends across the African continent vary. The fertility transitions took place at varying times across the region, and both the total fertility levels at the start of the transitions and the pace at which the transitions progressed also varied. Data calculated even for sub-regions may mask significant differences from country to country, she added; Figure 2-2 shows the diversity in rates across African countries with the varying rates color coded to indicate the range of rates.
Biddlecom and her colleagues have analyzed detailed data for the sub-regions and individual countries to search for patterns. Their analysis included 130 countries and covered approximately 60 years. They identified nine distinct clusters of countries that followed similar trajectories in their fertility declines and ultimately reached a level of 3 births or fewer per woman; these trajectories are shown in graph form in Figure 2-3.
Using these nine clusters, Biddlecom explained, it was possible to develop a range of fertility rate projections for individual countries, such
1See http://news.harvard.edu/gazette/story/2007/06/researchers-analyze-%E2%80%98africa-effect%E2%80%99-the-slow-growth-of-some-economies/ [July 2015] for a general discussion of the Africa Effect.
FIGURE 2-4 Range of fertility decline projections for Ethiopia.
SOURCE: Biddlecom (2015).
as those for Ethiopia shown in Figure 2-4. The projections for individual countries, in turn, allowed Biddlecom and her colleagues to develop a range of population projections, also based on possible scenarios for regions of Africa; the range of projections is shown in Figure 2-5.
Biddlecom closed with the observation that this is a “time of uncertainty” in sub-Saharan Africa because total fertility is high across the region despite considerable diversity within it. New data expected after the workshop would make it possible to update the scenarios she presented and look for changing patterns. Biddlecom noted that the modeling does not take into consideration socioeconomic determinants that may influence changes in fertility rates and that it is important to remember the role that policy choices may play in the outcomes. Moreover, she added, data collection in many African countries is problematic, which limits the value of the data on which the models are based.
The discussion turned next to an exploration of the ways fertility and population trends may influence economic development in sub-Saharan Africa. John Cleland of the London School of Hygiene and Tropical
FIGURE 2-5 Range of population projections for sub-Saharan Africa.
SOURCE: Biddlecom (2015).
Medicine and David Lam of the Population Studies Center, University of Michigan discussed challenges that sub-Saharan countries will face in coming decades and the possibility that demographic changes will also bring opportunity. David Canning of the School of Public Health at Harvard University described an approach to modeling the potential economic effects of reducing family size.
Challenges and Opportunities
Cleland began with reference to a remark by a former chief economist at the World Bank, Francois Bourguignon, that “we really do not know what causes economic growth.” He did so to emphasize that he would offer not predictions for what will happen but a set of challenges that need to be overcome for sub-Saharan African countries to experience rapid economic growth.
His first caution was against undue optimism about the demographic dividend—that is, the economic benefit that can come after a fertility transition, when a country reduces its fertility rate and the wage-earning component of its population becomes larger in comparison to the nonwork-
ing, dependent component (often referred to as a change in age structure). When this development coincides with other favorable circumstances, such as widespread access to jobs, education, and adequate nutrition and health care, a country can achieve significant gains in productivity as a result of this change in the age structure. Many Asian countries, in particular, benefited from this set of circumstances to achieve substantial economic gains.
In the sub-Saharan countries, however, Cleland explained, demographic projections indicate that there is likely to be growth in all age bands, so the changes in the age structure in many are likely to be quite modest. Projections also indicate that growth in the urban population will be dramatically larger than growth in rural areas. Thus, not only are the demographic shifts likely to be modest, he explained, but also the boundary between workers and dependents is likely to become more blurred as rates of education and urbanization increase. The percentages of youth ages 15 to 19 who are working will likely decline, for example, diluting some of the benefit of changes in dependency ratios.
Finally, Cleland added, reproduction is much less likely to interfere with wage earning in African nations than elsewhere, because women’s employment is often more compatible with childrearing, so a decline in reproduction will have less benefit in this region. Figure 2-6 shows that the projected decline in the dependency ratio for Africa is very slow and modest compared with those for other regions that experienced the demographic dividend.2
Another challenge for sub-Saharan African countries, Cleland went on, is that the region’s school-age population is growing by nearly 9 million per year. To meet the goal of universal primary schooling, the region will need to increase its teacher workforce from 3.2 million in 2011 to 5.3 million by 2030. A further increase from 1.0 to 3.5 million will be needed in the ranks of secondary teachers. Cleland suggested that chronic teacher shortages are likely to persist for decades in countries where academic achievement is already low, and it may prove difficult to provide or sustain high-quality educational opportunities in those circumstances. Similar shortages are likely in the medical workforce, he added. Of 12 countries in the region, only 4 are on track to increase their numbers of doctors, nurses, and midwives, and even in those countries the projected numbers are not sufficient to match projected population increases.
The region’s rapid urbanization is taking place without the industrialization that accompanied it in many other developing regions, Cleland pointed out. While the urban population is doubling every 20 years,
2Broadly defined, the dependency ratio relates the number of children (0-14 years old) and older persons (65 years or over) to the working-age population (15-64 years old).
FIGURE 2-6 Projected total dependency ratio, Africa and other regions, 1950-2050.
SOURCE: United Nations (2014).
62 percent of the urban population lives in slum conditions, as compared with 35 percent in Asian countries. Few countries in the region have large-scale programs to provide low-cost housing, and many have ambiguous arrangements for urban land ownership and weak municipal governance structures. The population pressure on the housing infrastructure is likely to be relentless, in Cleland’s view, and he suggested that this circumstance might give rise to increasing slum populations, housing insecurity, violence, and threats to social cohesion.
Sub-Saharan Africa is also already the region of the world with the largest prevalence of undernourishment, Cleland explained, with 32.7 percent of the region’s population having had insufficient nutrition in the years 2011 to 2013. To meet the needs of a growing population, the region will need to double its food availability over the next 35 years, he said.
This will be difficult, Cleland explained. Agriculture accounts for 64 percent of employment in the region, but 80 percent of farms are less than 2 hectares in size and ownership rights are often insecure. Yields are not improving, and 95 percent of crops are dependent on rain, as opposed to irrigation. Seventy percent of arable soil is degraded, and the region currently imports 31 percent of its cereals, at a cost of $30 to 50 billion annually. He noted the possibility of ameliorating some of these problems, but added that many countries have reached the limits of their capacity. The ratio of the agricultural population to arable land will likely increase, he explained, which will in turn lead to overexploitation of fragile land and further soil degradation. As farms become smaller, the possibilities for innovation and the production of surplus will decline. The insecurity of many farmers’ tenure on their land is a further disincentive to invest in long-term improvements. The biggest threat, however, comes from the erratic rainfall patterns and increases in temperatures that have already begun as a result of global climate change.
Another issue Cleland pointed out is that as the region’s labor force increases by a projected 32 million annually, the pressure on employment will intensify. He said that, according to projections, over the next 10 years only one in four young people will locate a wage job; the other three-fourths will need to find employment in the informal economy. It is possible that African countries could expand light manufacturing operations and begin to reap some of the rewards of the globalizing economy. Unfortunately, however, Cleland noted, the labor force in the Asian countries with which African countries would be competing is also expected to continue growing. Because worldwide manufacturing jobs are not increasing, African countries’ expanding labor forces are not a significant advantage in this arena, in his view.
Cleland noted the significant variability in projected growth for individual countries in the region, as shown in Figure 2-7, and closed with the observation that it is likely that socioeconomic trends in these countries will continue to diverge. Some, he suggested, are likely to hit Malthusian limits, where growing populations outstrip the nation’s capacity to feed, house, and employ them. Some will remain or become failed states. These developments are likely to generate migration on a very large scale across the continent, he added, as people seek refuge or economic opportunity, and he wondered whether this could happen without causing significant civil strife.
Another View of Challenges and Opportunities
David Lam of the Population Studies Center at the University of Michigan agreed with Cleland that the expected continued population
FIGURE 2-7 Projected population percentage increases for sub-Saharan African
nations, 2015 to 2050.
SOURCE: United Nations (2014).
growth in the region, particularly among youth, will create both challenges and opportunities. Population growth in sub-Saharan Africa is projected to remain high compared with rates in other regions, Lam explained; see Figure 2-8. The TFR, currently at 4.8 births per woman, is expected to decline to 3.0 births by 2050, but will also be higher than
FIGURE 2-8 Annual population growth rates, actual and projected, 1950-2050.
SOURCE: United Nations (2014).
those in other regions: rates in south, east, and Southeast Asia and Latin America are all on track to reach or remain at a rate below 2 births per woman by that year. Thus the proportion of the population aged 0 to 14 will also remain highest in sub-Saharan Africa. The percentages of the population in older working-age groups are also likely to remain large, Lam added. Figure 2-9 shows the percentage of working-age (15 to 64) populations who were age 24 or under, for years 1975 and 2015, in 18 countries.
The dependency ratio for sub-Saharan Africa is high, at 0.85, compared with that for the world (0.52) and for Latin America (0.51) and Asia (0.47), Lam noted. However, the ratio for the world is increasing, he explained—the elderly population will nearly double by 2050, which will offset the decline in the percentages of children in populations where fertility rates are low. Sub-Saharan Africa is the only region for which the dependency ratio is projected to decline during that same period; by 2050 it will be lower than those for Europe and North America, and only slightly above that of Latin America.
FIGURE 2-9 Percentages of working age (15 to 64) populations in 18 countries who were age 24 or under, for 1975 and 2015.
SOURCE: Lam (2015).
The working-age population in sub-Saharan Africa is expected to have a net increase of 14 million in 2015, and increases of 20 and 26 million, respectively, by 2030 and 2050, Lam added. In some regions, particularly less economically developed ones, Lam explained, bulges in the youth population have meant high unemployment and social unrest. The primary challenge posed by population growth in the sub-Saharan region, in Lam’s view, is that the region will need to gain 1.1 million jobs each month in order to keep employment rates constant. If current projections are accurate, then the need will be even greater in the future: 1.6 million per month by 2030 and more than 2.0 million by 2050. However, because the growth rate in the working-age population will remain high, dependency ratios will fall even without more rapid fertility declines. Overall, Lam and his colleagues concluded that population shifts are likely to have mixed effects on economic development, and other determinants are likely to have a greater influence on outcomes.
Modeling the Effects of Demographic Change on Economic Growth
David Canning presented a simulation model of the possible effect of reduction in family size on economic growth in the region, noting that approaches to understanding the role of declining fertility in accelerating
economic development have evolved. There was little evidence, in his opinion, to support an older view that population growth slows economic growth. The newer view, that fertility decline brings a demographic dividend because of changes in the age structure, an increase in the female labor supply, and changing views of the investment in children, he suggested, can be tested empirically in several ways.
One macroeconomic approach is to use growth regressions to project per capita income based on changing age structures. This approach, he explained, can separate the effects of fertility from those of mortality or life expectancy. However, sample sizes tend to be small for this type of analysis, Canning added, so it is difficult to establish causality using it. Applying regression analysis on a micro level, using household data, can help to identify the effects of demographic changes on children and families, he explained, but may miss macro-level influences, such as the effects of public education or the ratio of labor to capital.
A third approach is to develop a macro-level simulation of the economy to investigate the effects of policy changes. This approach, Canning explained, makes it possible to see causal linkages and to include the results of micro-level studies in developing parameters, though its complexity can make results difficult to interpret and it can be computationally intensive. Canning and his colleagues used this approach to compare a baseline fertility scenario to one in which family planning policies are introduced to reduce fertility. The basic model is shown in Figure 2-10.
Canning and his colleagues adapted the model first presented by Ashraf, Weil, and Wilde (2011) by including five additional factors: endogenous savings rate, child health, effects of manufacturing and agriculture, wage distortions, and the effects of education of females on fertility. They applied the model to the case of Nigeria using data from a variety of sources covering such factors as land and natural resources, population characteristics, labor force participation and labor by sector, education and height, age-specific fertility rates, and savings. They examined three scenarios: fertility decreases resulting in projected low (1.7 children per woman), medium (2.2), and high (2.7) fertility levels by 2100. The resulting possible effects on population and per-capita income are dramatic, with the difference in the Nigerian population in 2100 between the highest and lowest scenarios amounting to some 600 million people, and the projected difference in per capita income amounting to more than US $10,000 for the same time period.
Canning and his colleagues hope to add further refinements to the model, but they conclude that it highlights the potentially large effects of fertility rates on economic outcomes: more than double the effects shown in previous analyses.
FIGURE 2-10 Simulation model for assessing fertility reduction scenarios.
NOTE: LFP = labor force participation.
SOURCE: Canning (2015).
Workshop participants discussed a variety of comments and questions about the presentations in the first session, and Jacob Adetunji of the U.S. Agency for International Development and Jean-François Kobiané of the University of Ouagadougou offered their comments.
With respect to the session on trends in fertility rates, many participants focused on the Africa Effect. One participant noted that more research is needed on the reasons why many African countries have such high fertility rates, compared with other regions. Another expressed a wish for more expert insight into the variation across the nations of the region, perhaps by means of a model that takes into account the policy choices that individual countries have made. Presenters also noted that collecting data at the national and sub-national levels is challenging, but that as improved data become available, these will be valuable avenues to pursue.
Adetunji noted that the portrait of trends in fertility decline and population growth should remind the group that the question of whether Africa has too many people is too broad because, “it depends on where you are looking.” Population growth is very different from country to country, he added, but it is nevertheless true that the implications of population projections for the region are “enormous.” Family planning programs have an effect, he added, as Rwanda, Malawi, and Ethiopia demonstrate. So, he concluded, “We have a clue about what can be done to help Africa manage its population growth.”
Regarding the session on consequences of fertility trends, presenters added a variety of points to the discussion. One noted that the sub-Saharan region has seen strong performance in the service and telecommunications sectors over the past 15 years, and suggested that it may not be necessary for the region to experience strong growth in manufacturing to prosper more generally. Cleland acknowledged that many countries have indeed experienced such growth, but noted that in most cases this is the result of changes in governance designed to improve the welfare of the population—another reason that focusing on the variation across countries would be valuable.
Another participant noted that the high rates of migration that Cleland foresees may be easier for the region to manage than Cleland fears, because there already is considerable internal migration, and structures have evolved to handle it. Several agreed that data limitations mean that forecasts, while useful, may be misleading. Participants identified some important factors that were not considered in the models, particularly political and other developments, such as changes in female labor force participation, which might significantly alter countries’ trajectories. One suggested that Latin American countries might have more similarities with the African region than Southeast Asian countries do and therefore offer more useful comparisons. Finally, a few participants highlighted the importance of equity issues: one observed, “even if we double GDP, if the wealth is concentrated in a few hands it will not change anything.”
Kobiané noted that though the three discussions of the possible consequences of fertility trends offered some differing views, they made clear that the effects of fertility decline are not systematic. They made clear, he observed, that the way fertility interacts with other development issues, such as education, governance, and leadership, in each country will influence the outcomes for population growth and economic development. He agreed with earlier comments that greater attention to the collection of comparable data across the region is needed.