9.3.1 Sectoral GDP Per Capita
We use GDP per capita data from the Penn World Table (PWT) version 8.1 (Feenstra et al. 2015) and sectoral share data from the World Bank’s World Development Indicators (WDI). Combining the two datasets, Fig. 9.1 shows changes in sectoral GDP per capita (in 2011, constant USD at purchasing power parity) in log scale. We include 17 major Asian countries for analyses in this chapter: Bangladesh, China, Hong Kong, Indonesia, India, Japan, Cambodia, South Korea, Laos, Sri Lanka, Malaysia, Nepal, Pakistan, Philippines, Singapore, Thailand, and Vietnam. However, the sectoral share data for 1970 (and other years) is not available for Hong Kong, Cambodia, Laos, Singapore, and Vietnam, and thus, these five countries are excluded from Fig. 9.1 so that we can remove influences from the inclusion of sample countries, and focus on time-series changes of sectoral GDP per capita (see Appendix for data availability of our sample countries). In addition, the economic development of Japan, which started in the 1950s, was exceptional among the Asian countries, and thus, Japan’s changes in sectoral GDP are separately presented in Panel B, whereas weighted averages of other emerging Asian countries are presented in Panel A.
Similar to Japan, South Africa was exceptional compared to other sub-Saharan African economies because, for instance, its GDP in 1970 accounted for a quarter of total GDP of all 45 sub-Saharan African countries. Therefore, weighted averages of sub-Saharan countries, excluding South Africa, are presented in Panel C, whereas Panel D presents the numbers for South Africa. In Panel C, we restrict our sample to 26 sub-Saharan African countries whose sectoral share data for 1970 is available. Although we limit our analyses to the 26 countries, the average level of GDP per capita is similar in Panel C and Fig. 9.2, which includes all the 45 sub-Saharan African countries for which GDP data is available, at around 2,000–3,000 USD, and thus, these 26 countries give a representative picture of the entire sub-Saharan African region.
Figure 9.1 corresponds to the second expression in Eq. (9.1), and presents a sharp contrast in the economic development of Asian and African economies. Panels A and C illustrate that the GDP per capita was at a similar level in 1970 in Asia and Africa. In 1970, for instance, Korea’s GDP per capita was 2,088 USD, which was comparable to that of Congo (2,281 USD) or that of Tanzania (1,750 USD), and was much lower than that of Ghana (2,843 USD). The two regions’ economies have diverged, and the stagnation of African economies in the 1970s and 1980s, as discussed by Easterly and Levine (1997) among others, is evident in Panel C. The increase in Asia’s GDP per capita seems to be a result of the increase in industrial GDP per capita and service GDP per capita, whereas the industrial GDP remained stagnant in Africa.
Using the same data as Fig. 9.1, Table 9.1 presents sectoral GDP per capita and the contribution of each sector in selected years. There are three major points to note. First, the agricultural GDP per capita is almost constant in the last four decades, particularly in Africa. As Gollin et al. (2014) find, the increase in agricultural productivity is cancelled out by the decrease in the labor share in agriculture, and thus, the agricultural GDP per capita did not increase despite the productivity improvement. This supports our argument in the previous section that economic growth cannot be achieved only by developing the agricultural sector.
Secondly, similar to the agricultural sector, the industrial GDP per capita has been stagnant at a low level over the past four decades in Africa. This suggests that neither the ISI policies adopted in the post-independent era of the 1960s and 1970s nor the SAP policies in the 1980s and 1990s helped industrialization in Africa. The SAP comprised market-oriented policies without governmental intervention, which can be considered an absence of industrial policy. The absence of industrial policy and subsequent failure in industrialization in Africa in the 1980s and 1990s is reminiscent of India’s experiences as discussed by Nomura (2018) in this volume. Using data from Nigeria, Austin (2014) argues that the failure, as well as the absence, of industrial policies resulted in the failure of labor-intensive industrialization despite its potential for offering ample employment opportunities.Footnote 6
In contrast, Asian economies experienced a rapid increase in industrial GDP per capita, which led to Asian economic growth. Otsuka et al. (2017a) discuss that industrial development played a critical role in Asian economies by increasing income without expanding inequality because labor-intensive industrial sectors created ample job opportunities. Similarly, Tanimoto (2018) and Kubo (2018) in this volume illustrate that labor-intensive industrialization took place in the early stage of economic development, which was followed by the development of capital-intensive industries in Japan and China, respectively. Stiglitz and Uy (1996) and Wade (1990) argue that such economic “miracle” of Asia, led by Japan and followed by Asian Tigers (Hong Kong, South Korea, Singapore, and Taiwan), was a consequence of industrial policies not in the form of ISI policies that ignored the comparative advantage but in the form of export-oriented policies that was friendly with markets and took advantage of the comparative advantage.
Third, the service GDP per capita increased both in Asia and Africa. In particular, the modest growth of GDP per capita after 2000 in Africa has been led by service sector growth. The GDP per capita increased from 2,494 USD in 2000 to 3,240 USD in 2010, and most of the increase can be attributed to the service sector growth. This phenomenon is known as premature deindustrialization (Haraguchi et al. 2017; Rodrik 2016) and premature shift to service industries (Page 2012). In Asia, Philippines and India are examples of economic growth driven by service sector growth, and a similar pattern is observed in Rwanda and Senegal. The labor absorption capacity, however, is limited in the service sector, particularly in modern service sectors like finance and ICT without backward and forward linkages in industries. These countries experience widening inequality within each country, and thus, this pattern of economic development seems to be unsustainable.
9.3.2 Sectoral Employment
Next, we show the sectoral share of employment in Table 9.2. As the data availability is severely limited, we replace a missing observation with the nearest observation of the same country within the four-year interval.Footnote 7 We must note, however, that the number of observations is still small, particularly in the early period in our analyses. Our data shows that the agricultural share of employment has continuously decreased both in Asia and Africa, whereas the service sector share has continuously increased.
Admitting that sample countries do not overlap, Tables 9.1 and 9.2 suggest that industrial productivity has increased in Asia, but decreased in Africa, where the increase in employment share exceeds the increase in industrial GDP per capita. This stands in contrast to Duarte and Restuccia (2010) and Rodrik (2013), who find that manufacturing labor productivity unconditionally converges all over the world. Such exceptional absence of industrial productivity growth in Africa illustrates the failure of industrial development. At the same time, however, the low industrial productivity and global convergence suggests that there is potential room for increasing industrial labor productivity through industrial policies.
9.3.3 Finer Sectoral Classification
In order to analyze the contribution of specific industries to economic growth, we alternatively denote k in Eq. (9.1) as narrowly defined industries, such as mining, manufacturing, trade, and finance, rather than the broadly defined industrial sector vis-à-vis agriculture and service sectors. We use data taken from the Groningen Growth and Development Centre (GGDC) 10-Sector Database (Timmer et al. 2014), which categorizes economic activities into ten groups, which are comparable across countries and over time. Although the GGDC data has advantages over comparable and finer classifications of industries, the data coverage is limited. The sample Asian countries include China, Hong Kong, Indonesia, India, Japan, South Korea, Malaysia, Philippines, Singapore, and Thailand (nine of the original 16 sample countries) and the sample African countries include Botswana, Ethiopia, Ghana, Kenya, Mauritius, Malawi, Nigeria, Senegal, Tanzania, South Africa, and Zambia (11 of the original 45 sample countries).
Table 9.3 shows the sectoral contribution to GDP per capita and employment. The agricultural share of GDP per capita has decreased, and the share of the service sector has increased, particularly the finance sector both in Asia and Africa. In the previous sub-sections, the industrial sector included both mining and manufacturing sectors. The share of mining sector is larger in Africa than in Asia, particularly because the African sample countries incudes Botswana and Nigeria, resource-rich countries with relatively high per capita GDP in Sub-Saharan Africa. The employment share of mining, however, is small because the sector is intensive in capital-use and the employment creation is limited.
The manufacturing sector, which is represented by labor-intensive industries and absorbs large amounts of labor, presents clear contrast in the two regions. The manufacturing share of value added in Asia was already much higher than in Africa at 18.9% in 1970, and further increased to 23.3% in 2010 so that almost a quarter of Asian value-added is from the manufacturing sector today whereas the share in Africa has been stagnant at around 10%. Moreover, the value-added share of the manufacturing sector has increased more than the employment share in Asia, indicating that the labor productivity in the manufacturing sector has increased. Therefore, the manufacturing sector creates numerous high value-adding jobs, and contributes to Asian economic growth.
9.3.4 Decomposition Analysis
Table 9.4 shows the results of decomposition analyses based on Eq. (9.2). Row (A) corresponds to the first term on the right-hand side, and represents within-industry labor productivity growth. Both in Asia and Africa, this accounts for a large proportion of labor productivity growth. Row (B) represents static labor reallocation, whereas row (C) represents dynamic labor allocation. In Africa, the latter is negative and large in magnitude. Africa is experiencing urbanization and expansion of urban slums, and the urban service sector, mostly in informal economy, absorbs labor from rural areas. The productivity of such sector, however, remains low. On the other hand, the reduction of surplus agricultural labor increases agricultural labor productivity. Therefore, the contribution of dynamic labor reallocation is negative because the labor shifts away from agriculture, whose productivity increases, to other sectors—mostly the service sector, where the productivity does not increase.
We combine rows (B) and (C) to examine the total effect of labor reallocation. In Asia, the contribution is positive. Between both 1990–2000 and 2000–2010, labor reallocation contributes to about 1% of annual productivity growth. However, the contribution of labor reallocation is negative in Africa. In other words, the structural transformation does not account for economic development. Although the labor shifts from agricultural sector to other sectors, the productivity is low in these sectors.