Chapter 1. Output, Demographic, and Employment Transformation in Sub-Saharan Africa and Asia
- Louise Fox, Alun Thomas, and Cleary Haines
- Published Date:
- April 2017
Structural transformation has become a widely used macroeconomic measure of the quality of the economic development process.1 It is widely accepted that in order for it to happen, two processes have to be going on simultaneously: (1) a shift of GDP out of agriculture into modern industrial and service enterprises, and (2) following output, a shift in employment out of agriculture into the new nonagricultural enterprises, which necessarily involves migration and urbanization (economic densification). A demographic transition, resulting in a lower number of dependents and slower growth of the labor force, usually accompanies the output and employment transitions and facilities the shift of the labor force as the growing modern sector can absorb a larger share of new entrants to the labor force when there are not so many entering.
Since the mid-1990s, sub-Saharan Africa has had its longest continuous expansion in over 50 years. Economic growth averaged about 4 percent per annum among upper-middle-income countries and about 6 percent per annum among low-income sub-Saharan African countries. These growth rates surpass those of middle- and low-income Asia over this period (Figure 1). Moreover, during the recent financial crisis, the growth rate among low-income countries (LICs) in sub-Saharan Africa was broadly unaffected, in contrast to the sharp downturn among LICs in Asia.
Figure 1.Sub-Saharan Africa and Asia: Real GDP Growth
Source: IMF, World Economic Outlook Spring 2013.
However, even though the growth rate among sub-Saharan African economies has been very strong, the output structure of the economies has changed only moderately (Figure 2). Separating countries according to the level of per capita GDP in 2012, with threshold levels at $1,025 and below for LICs and $1,026–$4,035 for low-middle-income countries (LMICs), shows that in both types of countries, agriculture is falling as a share of GDP by about 8 percentage points combined with a corresponding rise in the share of services. At the same time the aggregate industry share remained fairly fat and manufacturing represents only about 7 percent of output in low-income sub-Saharan African countries.
Figure 2.Selected Regions: Development of Sectoral Output Shares
Source: World Bank, World Development Indicators.
The economic structure has changed more swiftly among LICs in east Asia.2 The agricultural share in these countries was higher at the beginning of the period than the share in sub-Saharan Africa, and fell by 15 percentage points of GDP. Moreover, in contrast to the experience in sub-Saharan Africa, the industrial sector made a large contribution to the output transformation. The manufacturing output share has risen by 4 percentage points in low-income east Asian countries and in the LMICs there, the manufacturing share continued to increase over the period. More important, in east Asian LMICs, by 2010, industry (including manufacturing) accounted for about one-third more output than in sub-Saharan Africa. The evolution of output shares in south Asia was closer to that of sub-Saharan Africa, especially in the LMICs (where India dominates).
The low manufacturing output share is mirrored by the low share of manufacturing exports in total exports in sub-Saharan Africa. This assessment is based on two-digit industry export data from Comtrade.3 The share of manufacturing goods in the export basket of low- and low-middle-income sub-Saharan African countries is very low, at between 10 percent and 20 percent on average over the past two decades. For east and south Asian countries, manufacturing export shares have historically been much higher, although they did take a sharp hit in the LICs of south Asia during the global financial crisis. Thus, the industrial sector among LICs and LMICs in sub-Saharan Africa is less competitive than in Asia, and, for LICs, more dominated by nontradables such as construction (see shaded yellow area in Figure 2.1, left panel).
Sub-Saharan Africa’s demographic trends are also different; fertility is higher and the demographic transition is happening much more slowly. Between 2000 and 2010 the labor force grew at a rate of only 1.2 percent per annum in east Asia and 1.7 percent per annum in south Asia, but 2.6 percent per annum in sub-Saharan Africa. The median age in sub-Saharan Africa is 18—seven years younger than the median age in south Asia, which is the next youngest region (Figure 3). These population trends suggest that the number of youth entering Africa’s working-age population will be rising for years to come. Between 2005 and 2020 the working-age population is projected to increase by over 200 million people, continuing the trend of the past decade. This trend is not expected to decline soon because a rapid, systematic reduction in fertility rates has yet to occur in sub-Saharan Africa. In the 1970s the fertility rate in Asia and Latin America was identical to the rate in Africa today, but Africa’s fertility rate is falling much more slowly than in those regions at that time (Filmer and Fox 2014).
Figure 3.Sub-Saharan Africa: Population by Age Group, 2005–20
Source: United Nations, World Population Prospects: The 2010 Revision.
How does the sub-Saharan African employment pattern compare with other regions? The previous section contrasted the output transformation in sub-Saharan Africa with the one happening in east and south Asia, and found large differences, especially with east Asia. The employment transformation shows similar gaps, especially in the lower-middle-income countries. The employment structure was estimated for the following categories (a mix of sector and employment type):
- Agricultural employment – predominantly farmers working on small holdings and consuming a significant share of their production, but including more commercialized farmers as well. Wage work in agriculture as a primary activity is included in this category as well as fishing and primary forestry (collecting wood and other forest products).
- Household enterprise (HE) employment – unincorporated, nonfarm businesses owned by households. This category includes self-employed people running unincorporated businesses (which may or may not employ family or other workers) and family members working in those businesses.
- Wage employment (industry or services) – includes all labor force participants who report working outside the agricultural sector and receiving a payment for their work from an unrelated individual. It includes the public and private sectors. This category is divided into the industry and service sectors as the relationship between output growth and employment is expected to differ. The former is more likely to be tradable, while the latter is more likely to be the public sector.
- Unemployed – labor force participants not in employment according to this paper’s broad definition.
These categories were chosen because the first two categories correspond to the Lewis-type “traditional employment” while wage employment is mostly, but not entirely, “modern” sector employment—the type of employment that should expand during a structural transition.
For the employment analysis, countries were divided up even further than in the output analysis presented in Figure 2. For comparison purposes, the upper-middle-income countries were included, and the resource-rich countries split out. The reason for the latter category is that resource-rich countries have a significantly different output–employment relationship than the non-resource-rich countries. The resource-rich countries are those whose ratio of resource exports to total exports was above 80 percent between 2008 and 2012, except for Botswana, which is categorized as an upper-middle-income country because of its high unemployment level.
The majority of Africans in low- and low-middle-income countries still worked in agriculture (either on their own family farm or as wage labor on other farms, or both) according to the regional estimated employment distribution in 2010 (Figure 4).4 For the LICs, this is not too surprising as the share of agriculture in GDP is still substantial. It is well known that the transformation in employment by sector always lags the transformation in output (more capital per worker is needed to employ people in more productive jobs).5 The high share of employment in agriculture in the LMICs is more surprising given the decline in the output share. Only in upper-middle-income countries has employment in agriculture almost disappeared.
Figure 4.Sub-Saharan Africa: Estimated Distribution of Employment by Country Type and Sector, 2010
Sources: Country household surveys; IMF, African Department database; and authors’ calculations.
The next largest category of employment is household enterprises, which are unincorporated, nonfarm businesses owned by households. They include self-employed people running businesses that may employ family members without pay but also self-employed people who run a business that employs a small number of nonfamily workers on a casual basis. The vast majority (70 percent) of nonfarm enterprises today are pure self-employment—just the owner operating the HE. About 20 percent of these enterprises include a family member in the operation, and only 10 percent have hired someone outside of the family.6 Taken together, the analysis shows that 80 percent of the labor force in 2010 was in household farms and firms—a segment commonly termed “the informal sector.”
To the extent that the employment transition consists of moving labor to the nonfarm wage employment segment, by 2010 most sub-Saharan African countries had not made much progress.7 Mirroring the output transformation, of the roughly 15 percent of the labor force in the wage sector in 2010, most were found in the services sector. Two types of wage employment — “formal” wage employment (where the employee has a contract and may be entitled to social protection), and “informal” or casual wage employment—are grouped together because most of the data sets do not allow a consistent disaggregation of wage employment into these categories. However, using a subset of countries, it was estimated that in the LICs and LMICs, about half of all nonfarm wage employment was in formal jobs.
Comparing sub-Saharan Africa to other regions using only sector of activity, shows that at similar levels of income, Asian countries have more employment in industry and less employment in agriculture (Figure 5). For this comparison, data from the Groningen sector database were used (de Vries,
Figure 5.Sectoral Shares of Employment Low- and Low-middle-Income Countries, 1990 and 2010
Sources: Groningen Growth and Development Center (GGDC) 10-Sector database; de Vries, Timmer, and de Vries (2015); and World Bank, World Development Indicators.
Timmer, and de Vries 2015); this contains a smaller set of sub-Saharan African countries than is shown in Figure 4, but a robust set of Asian countries. The agricultural sector charts reveal a reduction in the share of the labor force in agriculture in Asian economies as they advanced, but not a strong relationship in sub-Saharan Africa. Workers leaving the agricultural sector end up in both industry and services, but the share of employment in industry is much greater in Asia than in sub-Saharan Africa for a given level of GDP per capita. This picture also mirrors the output changes shown in Figure 2.
To compare regional performance on the basis of the distribution in Figure 4 (sector and employment type), the comparison sample was restricted to the few benchmark Asian countries (see Appendix Tables 2 and 4 for a GDP per capita comparison and final country list). Unfortunately, lack of employment data for the south Asian countries restricts the employment comparison to the east Asian economies plus Bangladesh.
An assessment of employment ratios across countries using the sector and employment type categories reveals that the differences between low-income sub-Saharan African and Asian countries are not stark. The share of the labor force in agriculture is not unusually high in sub-Saharan Africa, since Cambodia and Lao P.D.R. (countries with 2012 GDP per capita values ranging from $1,000 to $1,400) still have 50 percent or more of their labor force in agriculture (Table 1). Sub-Saharan African low-income countries are substantially poorer than this, on average, so the share of employment in agriculture is expected to be higher. The share of employment in HEs is actually smaller in sub-Saharan Africa than in the comparator countries, again reflecting lower income in sub-Saharan Africa. But in the LMICs, the differences become more obvious. The ratio of the employment share to the GDP share for agriculture in LICs in sub-Saharan Africa was 2.3 in 2010, while it was 2.4 in Bangladesh. But for sub-Saharan African LMICs, the same ratio was 3.4, much higher than in the comparator countries. Agricultural productivity is higher in the Asian countries as well, which has helped to reduce rural poverty to well below sub-Saharan African levels (Regional Economic Outlook Fall 2012). Labor is also more concentrated in the HE sector in LMICs in sub-Saharan Africa.
The low share of the labor force working in private industry is what makes the employment structure so different in low- and low-middle-income countries of Africa compared with the rapidly growing countries of Asia (Table 1). All the comparator countries have a larger share of employment in industrial wage jobs, because they have a high number of manufacturing jobs. Clearly the importance of agricultural commodity and mineral rents (as opposed to industrialization) in raising the per capita incomes of sub-Saharan African low-middle-income countries contributes to this discrepancy. Resource-rich countries in sub-Saharan Africa have not created much private wage employment at all. Unlike the low- and low-middle-income non-mineral exporters, the majority of the wage employment in resource-rich countries is in the public sector and high resource rents can create an economic structure unfriendly to private sector labor-intensive industry (Filmer and Fox 2014).
The lack of jobs in export-oriented manufacturing is not the only factor setting Africa apart. As discussed, a sluggish demographic transition means that the labor force is growing much faster in Africa than in Asia or Latin America. This lack of demographic transition complicates the employment transformation, because even with non-agricultural, modern private sector enterprise growth as rapid and labor intensive as occurred in the past 20 years in east Asia, a similar employment transition could not occur in sub-Saharan Africa. The enterprises would not be able to absorb the same share of the labor force because the labor force would be too big. For example, because Vietnam’s labor force grew at only two-thirds of the pace of Senegal’s over the past decade (2.1 versus 3.1 percent per annum), every dollar invested in creating labor-intensive manufacturing jobs will have a stronger effect on the structure of employment (measured as a share of the labor force) in Vietnam than in Senegal. In other words, Senegal needs 50 percent more investment in manufacturing than Vietnam needed just to bring its share of employment in industry to the level of Vietnam. In addition, the higher dependency level can be expected to result in lower private savings, reducing capital available for investment in the modern sector.
In sum, by the end of the past decade, the output transformation appeared to be underway, but moving toward services, not toward industry. The employment structure, however, reveals that looking at the output transformation only by sector in sub-Saharan Africa is deceptive. A large share of the growth in nonfarm employment was in household enterprises, not in the modern industrial and service enterprises that are expected during structural transformation. And the majority of Africa’s labor force still worked in its least productive sector—agriculture—which had yet to experience the substantial productivity growth seen in rapidly growing economies outside Africa. Meanwhile, the demographic transformation was also lagging the output transformation, which made it difficult to transform the employment structure quickly.