- The Lewis model (W. Arthur Lewis, 1954) explains how a poor economy grows by moving workers from a low-productivity traditional sector into a high-productivity modern (capitalist) sector.
- Because the traditional sector holds surplus labour with a marginal product close to zero, the modern sector can hire at a constant wage, so its labour supply curve is flat.
- Firms hire until \( MPL_M = W_M \). Profits are reinvested, the demand for labour shifts right, and employment grows while the wage stays constant.
- The Lewis turning point is reached when surplus labour runs out. From then on, the modern sector must pay rising wages to attract workers.
- Exam point: evaluate with Harris–Todaro (urban unemployment), capital-saving technology and the evidence on China, India, Bangladesh, Pakistan and South Korea.
What is the Lewis model?
The Lewis model is a theory of how a developing economy grows by shifting workers out of farming and into industry. It was set out by the economist W. Arthur Lewis in a 1954 article in The Manchester School called “Economic Development with Unlimited Supplies of Labour”[1]. Lewis shared the 1979 Nobel Memorial Prize in Economic Sciences with Theodore Schultz for “pioneering research into economic development research with particular consideration of the problems of developing countries”[3].
The model is also called the Lewis dual sector model, because it splits the economy into two parts that work in very different ways. Its central idea is that growth comes from moving people from work that adds almost nothing to work that adds a lot.
The two sectors
The traditional or subsistence sector. This is mostly small-scale farming, but Lewis also included casual work, petty trade and domestic service[1]. Families share the output, so each member receives roughly an equal share rather than a wage based on what they add.
The modern or capitalist sector. This covers factories, mines, plantations and modern services. Employers own capital, hire workers for a wage and aim to make a profit.
The key assumptions
- Surplus labour. The traditional sector has more workers than it needs. Lewis wrote that in such sectors “the marginal productivity of labour is negligible, zero, or even negative”[1][2].
- A constant subsistence wage. Earnings in the traditional sector are set by custom and by average output per person, not by the marginal product. This acts as a floor for wages in the whole economy.
- A wage premium. The modern sector must pay somewhat more than subsistence earnings to attract workers, to cover the costs of moving, higher living costs in towns and the loss of family support. Lewis treated this wage as fixed for as long as surplus labour lasts[2].
- Profit-maximising employers. Capitalists hire until the marginal product of labour equals the wage.
- Profits are reinvested. Capitalists save and reinvest their profits in more capital. Workers on the subsistence wage save little.
- Full employment in towns. Anyone who moves to the modern sector gets a job.
Lewis later said that the zero-marginal-product assumption is stronger than the model needs. What matters is that labour can be drawn into the modern sector at a constant wage, so that the supply of labour is “infinitely elastic”. The model still works if moving a worker does reduce farm output a little, provided the loss is smaller than what that worker adds in the modern sector[2].
The Lewis model in formulas
Use the subscript \( S \) for the subsistence sector and \( M \) for the modern sector.
1. The subsistence wage equals average product in the traditional sector. If \( Q_S \) is total traditional output and \( L_S \) is the number of people sharing it:
2. The modern-sector wage is a fixed mark-up over subsistence. If \( m \) is the premium (as a decimal):
3. Modern-sector firms hire until the marginal product of labour equals the wage.
The marginal product depends on the capital stock \( K \). More capital makes each worker more productive, so it raises \( MPL_M \) at every level of employment.
4. The capitalist surplus (profit) is output minus the wage bill.
5. Reinvested profit adds to capital. If capitalists save a fraction \( s \) of profit:
Put these together and you get a loop. Profit leads to investment. Investment raises \( K \). Higher \( K \) raises \( MPL_M \), so firms want more workers at the same wage. Because the wage is stuck at \( W_M \), all the gains from rising productivity go into profit. That is why the share of profits in national income rises during the Lewis phase, and why saving and investment rise with it[1].
The Lewis diagram explained in words
The top (modern-sector) panel carries most of the marks. Here is how to draw it.
Axes. The horizontal axis shows the quantity of labour employed in the modern sector, \( L_M \). The vertical axis shows the real wage and the marginal product of labour.
Subsistence level. Draw a dashed horizontal line at \( \bar{w}_S \), the subsistence wage.
Labour supply. Draw a horizontal line a little above it at \( W_M \). This is the modern-sector labour supply curve, flat because surplus workers will come in any number at this wage. Label the point where it stops being flat T, the Lewis turning point. To the right of T, the supply curve slopes upwards.
Labour demand. Draw a downward-sloping marginal product curve \( D_1 = MPL_M(K_1) \). It crosses the supply line at employment \( L_1 \).
Areas. The area under \( D_1 \) up to \( L_1 \) is total modern output. The rectangle below \( W_M \) up to \( L_1 \) is the wage bill. The triangle between \( D_1 \) and the wage line is the capitalist surplus (profit). Shade it.
Growth. Profit is reinvested, so capital rises to \( K_2 \) and demand shifts right to \( D_2 \). Employment rises to \( L_2 \) at the same wage, and the profit triangle grows. Repeat for \( D_3 \) and \( L_3 \).
Turning point. Once demand cuts the supply curve to the right of T, extra workers can only be hired at a higher wage.
The lower panel, if you draw it, shows the traditional sector’s total product curve. It is flat over the range of surplus workers, which shows that removing them leaves farm output unchanged. Fei and Ranis later built this two-panel version in detail and split the process into phases, depending on whether the workers leaving farming had zero, low or normal marginal products[4].
Worked example: from surplus labour to the Lewis turning point
Suppose an economy’s traditional sector has 100 workers producing 200 units of food. Of these, 40 are surplus: if they left, the other 60 could still produce 200 units. The family sharing rule pays each person the average product. The modern sector pays a 30% premium over subsistence. Its marginal product of labour with the starting capital stock is \( MPL_M = 8 – 0.2L_M \). All profit is reinvested. Each round of reinvestment raises the intercept of the \( MPL_M \) schedule by 2 units. All modern-sector workers come from the 40 surplus workers. Find employment and profit in each round, and the round in which the turning point is reached. (All numbers are illustrative.)
In words: while surplus labour lasts, the modern sector grows at a constant wage and capitalists capture the gains. Once the pool is empty, labour becomes scarce and wages start to track productivity.
The Lewis turning point
In the data, look for several signs together: faster growth in unskilled and migrant wages (including in rural areas), reported labour shortages, and a slowdown in the fall of agriculture’s share of jobs. No single sign settles it, because wages can also rise through minimum-wage laws. This is why turning-point dates are often debated.
Real-world evidence: structural change in Asia
The Lewis model predicts that, as the modern sector grows, the share of employment in agriculture should fall. The World Bank’s World Development Indicators publish this share for every country, using modelled estimates from the International Labour Organization (ILO)[5]. The table shows five Asian economies at different stages.
| Country | 1991 | 2000 | 2010 | 2020 | 2025 | Change 1991–2025 (pp) |
|---|---|---|---|---|---|---|
| China | 59.6 | 50.0 | 36.7 | 23.6 | 21.7 | −37.9 |
| India | 63.1 | 59.6 | 51.1 | 44.7 | 41.6 | −21.5 |
| Bangladesh | 69.9 | 67.6 | 47.6 | 42.5 | 44.3 | −25.6 |
| Pakistan | 40.7 | 41.7 | 42.7 | 38.3 | 36.2 | −4.5 |
| South Korea | 15.5 | 10.4 | 6.2 | 5.4 | 5.1 | −10.4 |
Employment in agriculture, % of total employment (modelled ILO estimates). pp = percentage points. Source: World Bank, World Development Indicators, indicator SL.AGR.EMPL.ZS, accessed October 2026[5]. Modelled estimates are revised from year to year, so recent values may change.
What the numbers tell you:
- China shows the textbook Lewis pattern. Agriculture’s share of jobs fell from about 60% in 1991 to about 22% in 2025,[5]. Notice that the decline has slowed sharply since 2020, which fits the idea that the surplus pool is nearly exhausted.
- South Korea had only about 15% of workers in farming by 1991, and its share has hovered around 5% since 2010[5]. It is an economy well past its turning point.
- India and Bangladesh show steady structural change, but both still have more than 40% of workers in agriculture. Both series rose around 2020, the year of pandemic lockdowns, a reminder that the process can reverse[5].
- Pakistan is the striking case. Its share barely moved between 1991 and 2016, staying in the low 40s, and has only recently drifted down to about 36%[5]. In Lewis terms, the modern sector has not grown fast enough to absorb surplus labour.
Has China passed the Lewis turning point?
China is the most debated recent test of the Lewis model. The evidence comes from wages and demography.
Rising wages. Zhang, Yang and Wang used primary surveys of wage rates and found a clear upward trend in real wages from 2003, even in slack farming seasons. They concluded that China’s era of surplus labour had ended[8].
Demography. Cai Fang linked the turning point to China’s demographic transition. He argued that the demographic dividend was shrinking and that a Lewis turning point was imminent[7].
The IMF estimate. Das and N’Diaye of the International Monetary Fund estimated China’s excess labour supply directly. Their model put excess labour at around 160 million in 2007, in line with an official estimate of 150–200 million. They projected that growth in the working-age (15–64) population would turn negative around 2020. Their central result was that, on current trends, the Lewis turning point would emerge between 2020 and 2025[6].
So economists disagree about the date. Reports of migrant labour shortages on the coast first appeared in 2004, and some economists read the wage rises that followed as the turning point[6]. Others put it around 2020–2025, based on estimates of the surplus pool. Part of the gap comes from labour market segmentation. China’s household registration (hukou) system and the limited portability of benefits held back migration, so shortages on the coast could exist alongside surplus labour inland[6]. Both views fit the sharp slowdown in the World Bank series above after 2020.
Criticisms of the Lewis model
1. Urban unemployment and the Harris–Todaro model
Lewis assumed that everyone who moves to town finds a modern-sector job. Harris and Todaro showed in 1970 why urban unemployment can persist even while migration continues[9]. In their model, a rural worker compares rural income with the expected urban wage, which is the urban wage multiplied by the chance of getting a job:
Here \( L_M \) is the number of modern-sector jobs and \( L_U \) is the urban labour force. Migration continues until the expected urban wage equals the rural wage.
Suppose the urban wage is 300, there are 60 urban jobs and the rural wage is 150. How large will the urban labour force be in equilibrium, and how many people will be unemployed?
In words: a high urban wage attracts more migrants than there are jobs. Creating 10 more urban jobs would attract more than 10 extra migrants, so job creation alone can raise urban unemployment. This is the opposite of Lewis’s smooth transfer.
2. Capital-saving and labour-saving technology
Lewis assumed that reinvested profits create jobs in proportion to the capital added. But if capitalists invest in labour-saving machinery, the \( MPL_M \) curve can become steeper rather than shifting right. Output and profits rise while modern-sector employment barely changes[10]. Profits may also be sent abroad rather than reinvested at home[10].
3. Surplus labour may be smaller than assumed
Farm labour is often fully used at planting and harvest, even if idle for part of the year. Removing workers can then cut food output and raise food prices, pushing up the subsistence wage and shrinking profits early. Fei and Ranis made this point central to their extension: unless farm productivity rises alongside industry, the transfer stalls[4].
4. Wages may rise before the turning point
Unions, minimum-wage laws, public-sector pay and multinational employers can lift urban wages even when surplus labour remains[10]. That squeezes profits and slows reinvestment. This also muddies the China debate.
Evaluation: how useful is the Lewis model today?
Strengths. The Lewis model gives a clear explanation of why growth in poor countries often comes with rising inequality at first: wages are flat while profits rise. It explains why saving and investment rates rise during industrialisation. It identifies structural change, the move of labour from low- to high-productivity work, as a main source of growth. And it gives a testable prediction, the turning point, which matches the experience of South Korea and, more recently, China[5][6].
Weaknesses. It assumes full employment in towns, job-creating investment, a fixed wage and an easy supply of food. Pakistan’s flat agricultural employment share for most of the period since 1991 shows that the transfer is not automatic[5]. Harris and Todaro showed that migration can create urban unemployment rather than modern jobs[9].
Judgement. The Lewis model is best seen as a framework rather than a forecast. It fits East Asian economies, where labour-intensive export manufacturing absorbed rural workers, better than South Asian ones, where the modern sector has grown more slowly and informal urban work is common. For an exam answer, the strongest conclusion is that surplus labour is a necessary but not sufficient condition for Lewis-style growth: the outcome depends on whether the modern sector is labour-intensive, whether profits are reinvested at home, and whether farm productivity rises to keep food cheap.
Practice questions
EasyQ1. What is meant by surplus labour in the Lewis model? Show worked answer
EasyQ2. Why is the modern-sector labour supply curve horizontal before the Lewis turning point? Show worked answer
MediumQ3. A traditional sector has 50 workers producing 150 units of output, with a sharing rule that pays the average product. The modern sector pays a 20% premium and has \( MPL_M = 9 – 0.3L_M \). Calculate the modern-sector wage, employment and profit. Show worked answer
MediumQ4. Using the Harris–Todaro model, the urban wage is 400, the rural wage is 160 and there are 80 urban jobs. Find the equilibrium urban labour force and the urban unemployment rate. Show worked answer
ExamQ5. Evaluate the usefulness of the Lewis dual-sector model in explaining economic development in Asian economies. [12 marks] Show worked answer
Knowledge and understanding (up to 4 marks). Define the dual economy, surplus labour with zero marginal product and the constant wage. Explain that firms hire where \( MPL_M = W_M \), profits are reinvested, labour demand shifts right and employment grows at a constant wage until the Lewis turning point. Credit an accurate diagram.
Application (up to 4 marks). Use data: China’s agricultural employment share fell from about 60% (1991) to about 22% (2025), and the IMF projected a turning point between 2020 and 2025. South Korea is well past its turning point at about 5%. Pakistan stayed around 40% for most of the period, showing stalled transfer.
Evaluation (up to 4 marks). Harris–Todaro: migration can create urban unemployment, not modern jobs. Capital-saving technology can raise output without raising employment. Surplus labour may be smaller than assumed, so food prices rise. Wages may be pushed up early by institutions. Judgement: fits East Asian labour-intensive industrialisation well; weaker where the modern sector is small or capital-intensive.
Frequently asked questions
What is the Lewis model in simple terms?
It is a theory of development in which a poor economy grows by moving workers from low-productivity farming into high-productivity industry. Industry hires surplus farm workers at a constant wage and reinvests its profits until the surplus runs out.
What is the Lewis turning point?
It is the point at which surplus rural labour is used up. After it, employers must raise wages to attract more workers, and real wages for unskilled labour begin to rise steadily across the economy.
Has China reached the Lewis turning point?
The evidence points that way, though the date is debated. Survey evidence shows real wages rising from 2003, and an IMF working paper projected the turning point for 2020–2025 as China’s working-age population began to shrink[6][8].
What is the main difference between the Lewis and Harris–Todaro models?
Lewis assumes everyone who moves to town gets a modern-sector job. Harris–Todaro assumes migrants respond to the expected urban wage, which includes the chance of being unemployed. So Harris–Todaro predicts persistent urban unemployment.
References
- Lewis, W. A. (1954). “Economic Development with Unlimited Supplies of Labour”. The Manchester School, 22(2), 139–191. https://doi.org/10.1111/j.1467-9957.1954.tb00021.x
- Kirkpatrick, C. and Barrientos, A. (2004). “The Lewis Model After Fifty Years”. Development Economics and Public Policy Working Paper No. 9, Institute for Development Policy and Management, University of Manchester. https://hummedia.manchester.ac.uk/institutes/gdi/publications/workingpapers/depp/depp_wp09.pdf
- Nobel Prize Outreach (1979). “The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 1979: Summary”. NobelPrize.org. https://www.nobelprize.org/prizes/economic-sciences/1979/summary/
- Fei, J. C. H. and Ranis, G. (1964). Development of the Labor Surplus Economy: Theory and Policy. Homewood, IL: R. D. Irwin. https://vufind.lboro.ac.uk/Record/12721
- World Bank (2026). “Employment in agriculture (% of total employment) (modeled ILO estimate)”, indicator SL.AGR.EMPL.ZS, World Development Indicators (last updated 8 October 2026). https://data.worldbank.org/indicator/SL.AGR.EMPL.ZS
- Das, M. and N’Diaye, P. (2013). “Chronicle of a Decline Foretold: Has China Reached the Lewis Turning Point?”. IMF Working Paper WP/13/26. International Monetary Fund. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Chronicle-of-a-Decline-Foretold-Has-China-Reached-the-Lewis-Turning-Point-40281
- Cai, F. (2010). “Demographic transition, demographic dividend, and Lewis turning point in China”. China Economic Journal, 3(2), 107–119. https://doi.org/10.1080/17538963.2010.511899
- Zhang, X., Yang, J. and Wang, S. (2010). “China Has Reached the Lewis Turning Point”. IFPRI Discussion Paper 977. International Food Policy Research Institute. https://hdl.handle.net/10568/153558
- Harris, J. R. and Todaro, M. P. (1970). “Migration, Unemployment and Development: A Two-Sector Analysis”. American Economic Review, 60(1), 126–142. https://www.jstor.org/stable/1807860
- Todaro, M. P. and Smith, S. C. (2020). Economic Development, 13th edition. Pearson. https://www.pearson.com/en-gb/subject-catalog/p/economic-development/P200000005609/9781292291192
