The CORE Econ website uses essential cookies to make our website work. You may disable these using your browser settings but this may affect website functionality (such as your access to logged-in resources). We would also like to use analytics cookies to help us improve the functionality of our website and improve your user experience. These analytics cookies will be set only if you accept. We do not sell or otherwise transfer personal data or usage data to any third parties or use it for any other purpose. For more detailed information about the cookies we use, see our Privacy policy.

Informality and the capitalist economy: A perspective from the Global South

4 Informality and structural change

The Lewis model predicts that as GDP per capita (output per worker) rises in the economy, the share of the informal sector in the economy should decline over time as the capitalist sector expands. Does the data bear this out? To measure the extent of structural change, ideally we need data on the share of the informal sector and the capitalist sector, measured in terms of both GDP and employment. Unfortunately, such data is not easily available in a comparable form across countries over long periods of time. However, we do have data on the GDP and employment shares of agriculture, manufacturing, and services for a large set of countries across the world. For this reason, most studies of structural change use these shares as proxies for the size of the informal and capitalist sector.

This approximation is valid to some extent, because, in developing countries, the majority of agricultural workers are self-employed and most self-employment is concentrated in agriculture. In India, for instance, about 80% of employment in agriculture is informal, and this agricultural informal employment accounts for 60% of the informal workforce. However, the surplus labour in the Lewis model does not just consist of agricultural workers but also includes urban informal sector workers. This means that workers could move out of agriculture without finding a job in the capitalist sector, instead getting absorbed in other parts of the informal sector. Think of a farmer who leaves the village for the city, but instead of finding wage work, is forced to sell mobile phone covers on the street.

Simon Kuznets was a Russian-American economist who did pioneering empirical work building cross-country datasets that showed declining shares of agriculture in output and employment in the United States, the UK, and other high-income countries.

Let us examine some data from India that illustrates this point. Figure 9 shows the trend in the share of employment in agriculture compared to the trend in self-employment. The declining share of agricultural employment is an empirical observation known as the ‘Kuznets process’. After the year 2000, when the Indian economy started growing much faster compared to previous years, the share of employment in agriculture declined much faster than the share of self-employment.

This line chart compares India’s agricultural employment share and self-employment share from 1983 to 2023. The horizontal axis shows years, and the vertical axis shows employment share in percentages. The line representing employment share in agriculture starts at about 63% in 1983, remaining around 60% until the early 2000s, then declining to 40% in 2018, followed by a rise and ending at 42% in 2023. The line representing employment share in self-employment starts at about 57% in 1983, fluctuating slightly over time, peaking at around 58% in 2004, declining to about 51% in 2018, then rising to around 55% in 2020 before ending at about 54% in 2023.
Fullscreen
https://books.core-econ.org/insights/informality-and-the-capitalist-economy/04-informality-and-structural-change.html#figure-9

Figure 9 Agricultural share and self-employment share in India (1983–2023).

Employment-Unemployment Surveys for 1983 to 2011. Periodic Labour Force Surveys for 2017 to 2023.

With the caveat that agriculture is a proxy for what we are really interested in (self-employment), Figure 10a–f shows how data on shares in agriculture, manufacturing, and services can be used to study structural change. Using a dataset that combines information from various countries over time, structural change can be described as follows: At its early stages, the share of agriculture in GDP and employment falls and the share of manufacturing and services rises. Eventually, the share of manufacturing starts falling, while that of services keeps rising.

One pattern in the data is that as countries get richer, the share of agriculture in GDP as well as in employment decreases. Equivalently, at any one point in time, lower-income countries have a higher proportion of their workforce in agriculture. The pattern is the reverse for services. But the manufacturing or industrial sector shows a more interesting non-linear relationship. What is the reason for this inverted U-shape in the data?

deindustrialized
The process by which the share of the manufacturing sector in a country’s GDP as well as its workforce reduces over time.

Over time, as an economy grows richer, two things happen. First, faster growth of labour productivity in manufacturing as compared to other sectors allows more output to be produced with less and less labour. That is, fewer workers are needed in manufacturing as compared to services, which tend to be more labour-intensive or harder to automate. Second, as wages rise, other economies with lower labour costs can attract manufacturing industries to their shores. As a result, the rich economies have deindustrialized even as emerging economies such as China have industrialized.

The scatterplot shows multiple country trajectories over time. The horizontal axis is GDP per capita in 2011 international dollars, on a logarithmic scale from 250 to 64,000. The plot shows the share of agriculture in GDP (%). At a low GDP per capita of around 500–1,000 dollars, the share of agriculture in GDP is above 60%. As GDP per capita increases to around 4,000–8,000 dollars, the share drops to about 20%, and at around 16,000–64,000 dollars, it falls below 10%
Fullscreen
https://books.core-econ.org/insights/informality-and-the-capitalist-economy/04-informality-and-structural-change.html#figure-10a

Figure 10a Trends in the GDP share of agriculture (1819–2016).

Our World in Data. Growth and Structural Transformation.

The scatterplot shows multiple country trajectories over time. The horizontal axis is GDP per capita in 2011 international dollars, on a logarithmic scale from 250 to 64,000. The plot shows the share employed in agriculture (%). At GDP per capita of 500–1,000 dollars, agricultural employment is often 60–80%. Around 4,000 dollars, it declines to about 40%, and at 16,000–64,000 dollars, it falls below 10%.
Fullscreen
https://books.core-econ.org/insights/informality-and-the-capitalist-economy/04-informality-and-structural-change.html#figure-10b

Figure 10b Trends in the share of employment in agriculture (1800–2010).

Our World in Data. Growth and Structural Transformation.

The scatterplot shows multiple country trajectories over time. The horizontal axis is GDP per capita in 2011 international dollars, on a logarithmic scale from 250 to 64,000. The plot shows the share of manufacturing in GDP (%). Manufacturing’s share rises from about 10–20% at GDP per capita of 500 dollars to peaks of 30–50% between 4,000 and 8,000 dollars, then declines to below 20% at 16,000–64,000 dollars.
Fullscreen
https://books.core-econ.org/insights/informality-and-the-capitalist-economy/04-informality-and-structural-change.html#figure-10c

Figure 10c Trends in the GDP share of manufacturing (1800–2010).

Our World in Data. Growth and Structural Transformation.

The scatterplot shows multiple country trajectories over time. The horizontal axis is GDP per capita in 2011 international dollars, on a logarithmic scale from 250 to 64,000. The plot shows the share employed in manufacturing (%). Employment starts below 10% at GDP per capita of 500 dollars, rises to 20–30% between 4,000 and 8,000 dollars, and then remains steady or slightly declines at higher incomes.
Fullscreen
https://books.core-econ.org/insights/informality-and-the-capitalist-economy/04-informality-and-structural-change.html#figure-10d

Figure 10d Trends in the share of employment in manufacturing (1800–2010).

Our World in Data. Growth and Structural Transformation.

The scatterplot shows multiple country trajectories over time. The horizontal axis is GDP per capita in 2011 international dollars, on a logarithmic scale from 250 to 64,000. The plot shows the share of services in GDP (%). At GDP per capita of 500–1,000 dollars, services account for around 20–40% of GDP, increasing to 50–60% by 4,000 dollars and reaching 70–80% at 16,000–64,000 dollars.
Fullscreen
https://books.core-econ.org/insights/informality-and-the-capitalist-economy/04-informality-and-structural-change.html#figure-10e

Figure 10e Trends in the GDP share of services (1800–2016).

Our World in Data. Growth and Structural Transformation.

The scatterplot shows multiple country trajectories over time. The horizontal axis is GDP per capita in 2011 international dollars, on a logarithmic scale from 250 to 64,000. The plot shows the share employed in services (%). At GDP per capita of 500–1,000 dollars, service employment is often below 20%. It increases to around 40–50% at 4,000 dollars and reaches 60–80% at 16,000–64,000 dollars
Fullscreen
https://books.core-econ.org/insights/informality-and-the-capitalist-economy/04-informality-and-structural-change.html#figure-10f

Figure 10f Trends in the share of employment in services (1800–2010).

Our World in Data. Growth and Structural Transformation.

The data presented in Figure 9 and Figure 10a–f gives us a visual snapshot of structural change. But to deepen our understanding we need to delve into the experiences of particular countries. Let us use the contrasting cases of India and China to show how this type of data can be used to draw conclusions about the speed of structural change.

Figure 16.18 from The Economy 1.0 shows the shift of employment into and out of industry for seven countries. Read Section 16.11 of The Economy 1.0 to learn more about the economics of changing employment shares in manufacturing and services.

According to the World Bank’s World Development Indicators database, in 1978, when China began its economic reforms, its GDP per capita was comparable to that of India (around $400 per year in constant 2015 US dollars). By 1991, when India started its own reform journey, the two countries had diverged ($975 for China vs $530 for India) but their sectoral structures were still fairly similar. Agriculture accounted for more than half of all employment and a quarter of GDP in both countries (Figure 11). However, over the subsequent three decades the share of agriculture fell much faster in China than in India. The increasing gap is particularly evident in the case of the share of agriculture in employment. From this data, we can conclude that the pace of structural change was much faster in China.

Why might this be the case? Recall, from the Lewis model, that the pace of structural change depends on the pace of job creation in the capitalist sector. This in turn is related to the growth of output in this sector. The slower the former and/or the weaker the latter, the slower will be the pace of structural change.

There are two line charts showing trends for India and China from 1990 to 2022. The first chart shows the share of agriculture in GDP (%). The horizontal axis shows years. In 1990, India’s share is around 27% and declines steadily to about 16% in 2022, with a slight rise around 2020. China’s share starts at around 23% in 1990, and declines more steeply to about 7% in 2022, with only minor fluctuations. The second chart shows the share of agriculture in employment (%). The horizontal axis shows years. In 1990, India’s share is around 63%, decreasing gradually to about 42% in 2022, with a slight increase near 2020. China’s share starts at around 59% in 1990, and declines steadily to about 23% in 2022, with a more pronounced drop between 2005 and 2015.
Fullscreen
https://books.core-econ.org/insights/informality-and-the-capitalist-economy/04-informality-and-structural-change.html#figure-11

Figure 11 Share of agriculture in GDP and employment for India and China (1990–2022).

World Development Indicators database.

Data on GDP growth shows that from 1991 (the first year for which cross-country data is available) until the onset of the COVID-19 pandemic, the Chinese economy grew on average at 9.5% per year, while the Indian economy grew at 6.2%. Further, China’s growth was driven, at least in the initial years, by labour-intensive manufacturing industries such as textiles, apparel, toys, and other consumer goods.1 Hence it was able to accommodate a much larger share of its surplus labour in modern productive activities. In the Indian case, GDP growth has been driven by skill-intensive modern services that have failed to create jobs in adequate numbers for workers leaving agriculture.2 3 As a result, in India, the share of manufacturing in total employment has never exceeded 12 or 13%. In China, at its peak this share reached 22%.

But India’s experience is not unique. In contrast to the experience of East Asian economies of South Korea, Taiwan, and China, where growth was historically driven by labour-intensive manufacturing and thus enabled structural transformation, most low- and middle-income countries have struggled to create adequate numbers of jobs in the capitalist sector. One indicator of this struggle is the stagnant share of self-employment in the total workforce. In Kenya, for example, ILOSTAT (the ILO database) reports that the size of the self-employment sector has remained fairly constant between 1991 and 2021, at 65%.

Exercise 3 Agricultural shares in GDP in employment

Use the Our World in Data (OWID) database (links given in question 1) to answer the following questions about the changing share of agriculture in GDP and employment.

  1. In the country you live in, what share of GDP and share of employment did the agricultural sector account for in 2019 (or the latest year available)?
  2. Download the dataset for the above two indicators for all the countries in the OWID database. Choose one country each from Latin America, sub-Saharan Africa, and Asia and plot the employment share and the share of GDP share as a line chart, with years on the horizontal axis.
  3. Use your charts to compare and contrast the pace of structural change in your chosen countries.

One challenge facing countries such as India and Kenya, identified by economist Dani Rodrik, is ‘premature deindustrialization’. Deindustrialization refers to the declining importance of manufacturing in a country’s economy. The United States and other high-income countries experienced this phenomenon in the latter half of the twentieth century. However, low- and middle-income countries are already experiencing a contraction in the share of manufacturing in GDP and employment, at much lower levels of per capita income than where the United States was when it started deindustrializing—hence the adjective ‘premature’ (earlier than high-income countries did).4

Why might this be the case? One reason, suggested by Rodrik, is that low- and middle-income countries have lowered their trade barriers, opening their economies to imported manufactured goods from high-income countries. This competition from imports has put many domestic manufacturing firms in the low- and middle-income countries out of business. Another issue facing every economy today is the accelerating pace of automation and emerging technologies such as artificial intelligence, which make even traditionally labour-intensive industries far less reliant on labour. Does the resulting inability of the capitalist sector to create enough jobs to absorb surplus labour mean a perpetually dual economy for some low- and middle-income countries? This is an important question for further research.

To learn more about how institutions and policies affect labour market outcomes and economic growth, read Sections 16.8–16.10 of The Economy 1.0. Section 18.10 of The Economy 1.0 also explains the role of trade in economic growth for East Asian and other countries.

What was it about South Korea or China that made it possible for them to compete with Germany or the United States in manufacturing, while other low- and middle-income countries were unable to do so? Economists have studied some key factors that can explain such differences. These include the success or failure of land reforms which improve agricultural productivity and availability of food for the rapidly growing urban areas, early public investments in human capital (especially healthcare and primary education), the development of critical infrastructure, and cleverly designed industrial and trade policies.5 6 7 But much more needs to be done by future generations of economists to understand the diversity of experiences.

  1. Naughton, Barry J. 2018. The Chinese Economy: Adaptation and Growth. MIT Press. ↩

  2. Basole, Amit. 2022. ‘Structural Transformation and Employment Generation in India: Past Performance and the Way Forward’. The Indian Journal of Labour Economics 65(2): pp. 295–320. ↩

  3. Ghose, Ajit K. 2016. India Employment Report 2016: Challenges and the Imperative of Manufacturing Led Growth. Oxford University Press. ↩

  4. Rodrik, Dani. 2016. ‘Premature Deindustrialization’. Journal of Economic Growth 21(1): pp. 1–33. ↩

  5. Amsden, Alice H. 2001. The Rise of the Rest: Challenges to the West from Late-Industrializing Economies. Oxford University Press. ↩

  6. Kay, Cristóbal. 2002. ‘Why East Asia Overtook Latin America: Agrarian Reform, Industrialisation and Development’. Third World Quarterly 23(6): pp. 1073–1102. ↩

  7. Weber, Isabella M. 2021. How China Escaped Shock Therapy: The Market Reform Debate. Routledge. ↩