Dual economy and structural transformation: Lewis

Detail

1. Key words first

• Dual economy: one country holds two very different economies side by side.
• One is a poor, low-productivity traditional sector.

• The other is a richer, high-productivity modern sector.

• Structural transformation: the slow shift of workers and output from agriculture to industry and services as a country develops.

• Marginal product of labour (MPL): the extra output from adding one more worker.

• Formula: MPL = ΔOutput ÷ ΔWorkers.

• Disguised unemployment: people look employed, but their MPL is zero or close to zero. If some of them leave, total output does not fall. (This is covered in detail in employment-informal-sector.)

• Example: 8 family members work on a farm that needs only 5.

  • 3 members leave, and output stays at 50 quintals.
  • So MPL = 0 ÷ 3 = 0. The 3 are disguised unemployed.

• Surplus labour: the pool of such workers. They can be moved out of farming with no loss of farm output.

2. The Lewis model (1954; Nobel 1979)

• W. Arthur Lewis split a developing economy into two sectors:
• Traditional sector (subsistence agriculture): has surplus labour. Many workers have near-zero MPL, i.e. disguised unemployment.

• Modern capitalist sector (industry): absorbs this labour at a constant wage. The wage is set at subsistence plus a small margin. The margin is what pulls workers off the farm.

• Official bodies use the same frame:

• The IMF describes a low-productivity sector with excess labour (agriculture) and a high-productivity sector (manufacturing) [5].
• The IMF also notes that wage increases in industry are held down by farm wages while labour keeps moving from farms to industry [5].

3. How growth happens in the Lewis model

The engine: fixed wages lead to rising profits, which lead to reinvestment

• Industry can hire any number of workers at the same wage, because the rural pool looks unlimited.
• So wages stay fixed while output grows. Profits in the modern sector rise.
• Capitalists reinvest their profits, so factories expand.
• The bigger factories need more workers, so more surplus labour moves across. The cycle repeats.

Worked example (illustrative numbers)

• Subsistence income on the farm: ₹100/day. Factory wage: ₹130/day (subsistence plus a margin).
• One factory worker produces ₹200/day of output.

• Profit per worker = ₹200 − ₹130 = ₹70/day.

• 1,000 workers give ₹70,000/day of profit. Reinvesting it builds new capacity and hires more workers from the farms, still at ₹130.

• Farm output does not fall, because the workers who left had MPL = 0.

4. Fei-Ranis extension (1961)

• Fei and Ranis added agricultural productivity to Lewis's model.
• As workers leave farms, the remaining farmers must produce a food surplus to feed the workers now in cities.
• If farm productivity does not rise:
• Food becomes scarce and food prices rise.
• Factory workers then need higher wages to buy food.

• Profits fall, and the growth engine slows.

• Lesson: agriculture and industry must grow together. Industrialisation cannot simply ignore farming.

5. Harris-Todaro model (1970)

• Harris-Todaro: migrants move on expected urban wages, not actual urban wages.
• Formula: Expected urban wage = (Urban jobs ÷ Urban labour force) × Urban wage = p × W_u

• p is the chance of finding a city job.

• Migration continues while p × W_u > W_r (rural wage).

• Worked example:
• Urban wage W_u = ₹600/day. Only 60% of urban job-seekers find work, so p = 0.6.
• Expected urban wage = 0.6 × 600 = ₹360. Rural wage W_r = ₹300.

• ₹360 > ₹300, so people keep migrating even with 40% urban unemployment.

• This explains why migration continues despite urban unemployment. It also explains the growth of urban slums and the informal sector.

6. Lewis turning point

• Definition: the stage when surplus rural labour runs out.
• The IMF defines it as the point where an economy moves from abundant labour to labour shortages. Wages rise, profits are squeezed and investment falls [5].
• After this point:
• Industry must bid workers away from agriculture.

• So wages start rising rapidly.

• Exam trap: wages do NOT fall at the Lewis turning point. They rise.

• Japan: reached it in the 1960s.
• China: debated for around 2010, as coastal wages rose.
• An IMF study (Das and N'Diaye, 2013) projected that China would reach the turning point between 2020 and 2025 [5].
• The same study estimated China's excess labour at 151 million (2010) → 57 million (2015) → 33 million (2020) [5].
• Aggregate wage growth in China stayed around 15% a year over the decade before 2013 [5].
• It found that demographics (a shrinking working-age population) would be the main driver, more than policy [5].

7. India vs China vs Pakistan: sectoral shares (Class 11, Comparative Development Experiences, Table 8.3, 2022)

Sector	GVA share: India	China	Pakistan	Workforce share: India	China	Pakistan
Agriculture	18	8	24	43	23	36
Industry	28	38	21	26	32	26
Services	54	54	55	31	45	38

• GVA (Gross Value Added): the value of output minus the value of inputs used up. It measures what each sector adds to the economy.
• Relative labour productivity = GVA share ÷ Workforce share. A value below 1 means the sector's workers produce less than the average worker.
• India agriculture: 18 ÷ 43 = 0.42. A farm worker produces about 42% of the average worker's output.
• India industry: 28 ÷ 26 = 1.08. India services: 54 ÷ 31 = 1.74.

• China agriculture: 8 ÷ 23 = 0.35. The gap exists in China too, but its farm workforce is much smaller.

• China followed the classic path: agriculture → industry → services.

• India and Pakistan moved directly from agriculture to services.
• In India, 43% of workers produce just 18% of GVA in agriculture. This is surplus labour on a Lewis scale that has not yet been absorbed.
• Services share of the workforce (1980s → 2022): India 17 → 31; China 12 → 45; Pakistan 27 → 38.
• Class 11, LPG: An Appraisal: post-1991 growth was "mainly driven by growth in the service sector".

8. Latest official data on India's labour shift (PLFS and Economic Survey)

• PLFS (Periodic Labour Force Survey): MoSPI's regular survey of jobs and unemployment.

• Usual status (ps+ss) counts a person's main activity over the past year, plus any side activity.

• Share of workers in agriculture: 45.8% (2022-23) [4] → 44.8% (2024) → 43.0% (2025) [2]. (NCERT: 43%, 2022.)

• Agriculture is still the largest employer (2025) [2].

• Other sector shares, 2024 → 2025 [2]:

• Manufacturing: 11.6% → 12.1%.
• Construction: 12.3% → 12.0%.

• Other services: 12.2% → 13.1%.

• Job quality, 2024 → 2025 [2]:

• Regular wage/salaried workers: 22.4% → 23.6%.

• Self-employed: 58.2% (2023) → 57.5% (2024) → 56.2% (2025).

• Unemployment rate (share of the labour force looking for work but not finding it), age 15+, usual status:

• 6.0% (2017-18) → 3.2% (2023-24) [3].

• 3.1% (2025) [2]. Youth (15-29): 9.9% (2025) [2].

• Female LFPR (share of women aged 15+ who work or seek work): 23.3% (2017-18) → 41.7% (2023-24) [3].

• The Economic Survey 2023-24 says rural India drives this rise. It rejects the view that the rise comes mainly from distress [4].

• How many non-farm jobs India needs:

• The Economic Survey 2023-24 assumes agriculture's workforce share falls from 45.8% (2023) to one-fourth by 2047 [4].
• On that basis, India needs about 78.5 lakh non-farm jobs every year until 2030 [4].
• In other words, India's Lewis transfer must speed up sharply.

9. Premature deindustrialisation

• Deindustrialisation: manufacturing's share of jobs and output falls.
• Premature deindustrialisation (Dani Rodrik): manufacturing's share of jobs and output peaks at much lower income levels than it did for today's rich countries.

• As a result, poor countries start moving to services before they become rich through factories.

• Detailed treatment is in sectors-of-economy.

• Link to Lewis: in the Lewis model, industry is the sector that absorbs surplus labour. If industry peaks early, the absorbing sector is too small.

10. The services-led growth debate

Why services cannot easily absorb India's surplus farm labour

• Many services, such as IT and finance, need skills that most farm workers lack.
• So services raise GVA, but they employ fewer low-skill workers.
• Manufacturing absorbs low-skill labour better. Factory jobs need less formal education than software or banking jobs.
• The result in India: services have 54% of GVA but only 31% of workers (2022). Agriculture still holds 43% of workers (NCERT, 2022; also 43.0% in PLFS 2025 [2]).

11. Leapfrogging and its limits

• Leapfrogging: skipping intermediate stages of development to adopt the latest technology directly.
• India's examples:
• Mobile phones instead of landlines.

• UPI (Unified Payments Interface, a real-time mobile payment system) and digital public infrastructure, i.e. Aadhaar and the India Stack.

• Limits of leapfrogging:

• It raises productivity.
• But it creates few mass jobs for low-skill workers.
• PLFS data still show agriculture as the largest employer: 43.0% of workers (2025) [2].

Prelims Hooks

• Lewis model (1954): two sectors. Traditional agriculture has surplus labour with near-zero MPL. The modern sector hires at a constant wage (subsistence plus a small margin).
• Disguised unemployment = marginal product of labour is zero or near zero. Removing the workers does not reduce output.
• Engine of Lewis growth: fixed wage → rising profits → reinvestment by capitalists → more labour absorbed.
• Fei-Ranis (1961) added agricultural productivity and food surplus. Harris-Todaro (1970) says migration responds to expected urban wage = p × W_u.
• Lewis turning point: surplus labour runs out and wages rise rapidly. They do not fall, which is a common trap. Japan: 1960s. China: debated around 2010; the IMF projected 2020-2025 [5].
• Class 11, Table 8.3 (2022): India agriculture = 43% of workers but 18% of GVA. China's services workforce share (45%) is higher than India's (31%).
• India skipped the industry stage (agriculture → services). China followed agriculture → industry → services.
• PLFS 2025: agriculture 43.0% of workers (down from 44.8% in 2024); manufacturing 12.1% (up from 11.6%) [2].
• Economic Survey 2023-24: about 78.5 lakh non-farm jobs a year until 2030 are needed [4].
• Premature deindustrialisation is linked to Dani Rodrik: manufacturing peaks at lower income levels.

Mains Points

• India's unfinished Lewis transfer:
• 43% of workers produce 18% of GVA (NCERT, 2022). Agriculture was still 43.0% of workers in 2025 [2].
• Services-led growth raised GVA but did not absorb low-skill labour.

• So the 78.5 lakh non-farm jobs a year target [4] points to labour-intensive manufacturing (textiles, food processing, electronics assembly) and construction.

• Fei-Ranis lesson for policy:

• Moving labour out of farms needs higher farm productivity (irrigation, seeds, agri-value chains). Otherwise food inflation erodes industrial profits.

• Agriculture and industry policy must go together, not one after the other.

• Harris-Todaro and urban policy:

• Creating urban jobs alone can raise urban unemployment, because more migrants arrive.

• Balanced rural non-farm jobs, small-town growth and skilling reduce slum growth and informality.

• China's turning point as a mirror:

• China's surplus labour fell from 151 million (2010) to 33 million (2020) [5].
• As wages rise there, India has a window to attract low-skill manufacturing.
• But premature deindustrialisation and automation may shrink that window. Leapfrogging (UPI, India Stack) raises productivity, not mass employment.