·The Hindu

If AI lifts productivity, will demand sustain growth?

FlashcardsQuick reviseKey facts. One by one.Start →
In this note
  1. At a Glance
  2. Why in the News
  3. Background & Evolution
  4. Core Static Facts
  5. Multi-Dimensional Analysis
  6. Recent Developments (last 12–18 months)
  7. Prelims Hooks
  8. The 'Machines for Machines' Bet Is Being Tested in Real Time
  9. India's Wages Were Falling Behind Productivity Before AI
  10. The Best Case Against the Demand Worry, and Where It Falls Short
  11. Who Must Act So Wages Keep Pace With AI
  12. Anchors for Answers
  13. Mains Relevance
  14. Related Topics to Study Next
  15. NCERT Concepts to Revise
  16. Common Errors / Trap Areas

1. At a Glance

  • Core question: if AI raises labour productivity but real wages stay flat, the wage share falls. Workers spend most of their income, so consumption demand can slow even while potential output rises. That is a demand-side limit on AI-led growth [1].
  • Theoretical debate: Tugan-Baranovsky argued that "machines can produce machines which can produce more machines", meaning investment can sustain itself. Rosa Luxemburg argued that machines are not an end in themselves, so final consumption demand must eventually justify them [1].
  • Policy view: the IMF identifies two channels, labour displacement and complementarity [2]. The ILO warns that productivity gains may not reach workers [3].
  • UPSC relevance: links GS-III topics (growth, employment, inclusive growth, technology) with India's AI strategy, such as the Economic Survey 2025-26 and NITI Aayog's AI roadmap [8][10].

2. Why in the News

  • An op-ed by Rohit Azad and Indranil Chowdhary, "If AI lifts productivity, will demand sustain growth?", appeared in The Hindu on 20 November 2026 (p. 11). It is the second part of a two-part series on the macroeconomics of AI [1].
  • Investment in AI-related equipment has grown about 3.5 times since 2021, far faster than GDP [1].
  • At the 2026 IMF/World Bank Annual Meetings, the ILO called for productivity gains to "translate into better jobs and incomes" [3][4].

3. Background & Evolution

  • Classical debate: whether capitalism can expand without limit through investment, or is held back by limited consumption. Tugan-Baranovsky said investment can feed on itself; Luxemburg said accumulation needs final consumer demand [1].
  • Demand framework used by the article: at the global level, demand = consumption + investment + government expenditure, because exports and imports cancel out across the world [1]. The authors leave government out and analyse a "pure" market economy, i.e. only C and I [1].
  • Key mechanism: productivity ↑ with real wages flat → wage share ↓ → consumption growth ↓ [1].
  • Milestones in institutional work:
  • 2023: IMF Working Paper (Pizzinelli et al.) on AI exposure and complementarity across occupations [6][5].
  • 2024: IMF Staff Discussion Note Gen-AI: Artificial Intelligence and the Future of Work [2].
  • October 2025: NITI Aayog Roadmap on AI for Inclusive Societal Development [10].
  • 2026: Economic Survey 2025-26 chapter on AI [8]; IMF WP/26/204 on AI and aggregate labour productivity [7]; ILO statements at the IMF/WB Annual Meetings [3][4].

4. Core Static Facts

Item Fact
Labour productivity Output per worker or per hour. It rises when AI raises output per unit of labour [1]
Wage share Real wage ÷ labour productivity. If productivity rises faster than real wages, the share falls [1]
Global aggregate demand C + I + G. Net exports sum to zero globally [1]
Labour displacement channel (IMF) Tasks shift from human labour to AI capital, which reduces labour income [2]
Complementarity channel (IMF) Value added shifts towards occupations that complement AI. Labour demand rises there and falls elsewhere [2]
Who is most exposed Women and highly educated workers face higher exposure, but also high potential complementarity [6][5]
AI capex trend AI-related equipment investment is about 3.5× its 2021 level [1]
IMF WP/26/204 data Patent data for OECD countries, 2000–2017 [7]
India – AI readiness Ranked 27th in Oxford's Government AI Readiness Index 2025 (cited in a PIB-linked document, Feb 2026) [9]
India – nodal policy texts Economic Survey 2025-26 AI chapter (Ministry of Finance) [8]; NITI Aayog roadmap on AI for inclusive societal development, aimed at the informal workforce [10]

5. Multi-Dimensional Analysis

Economic

  • Under-consumption risk: if productivity gains go mostly to profits, workers' consumption slows. Investment then has to fill the demand gap [1].
  • Investment-led path: AI capex (about 3.5× since 2021) is currently growing faster than GDP. Whether it can keep doing so without consumer demand behind it is the core of the Tugan-Baranovsky vs Luxemburg debate [1].
  • Output and jobs: the article asks what happens to the global "pie" of output and jobs, and leaves aside which countries gain or lose [1].

Social / Labour

  • The ILO says adjustment happens through more than unemployment. It also shows up as weaker hiring, falling real wages, rising informality and worse working conditions [3].
  • In developing countries, the costs of adjustment may arrive before the productivity benefits [3].
  • Gains may not reach workers, especially in smaller firms [3][4].

Distributional / Equity

  • Displacement lowers labour income, while complementarity redistributes demand between occupations, so inequality can widen within the workforce [2].
  • Exposure differs by gender and education level [6][5].

Technological

  • The IMF measures AI's link to aggregate labour productivity through patent activity [7].
  • India's official framing: AI adoption should be "economically grounded and socially responsive" [8].

Governance / Policy (India)

  • The Economic Survey 2025-26 treats AI deployment as a strategic choice [8].
  • NITI Aayog's roadmap focuses on extending AI benefits to the informal workforce [10].
  • The Economic Survey's AI chapter also covers the employment questions raised by rapid adoption [8].

6. Recent Developments (last 12–18 months)

  • October 2025: NITI Aayog publishes its Roadmap on AI for Inclusive Societal Development [10].
  • 2026 (Economic Survey 2025-26): a dedicated chapter treats AI as a strategic choice for India [8].
  • February 2026: a PIB-hosted document says AI is lifting productivity and inclusion, and cites India's 27th rank in the Government AI Readiness Index 2025 [9].
  • 2026: IMF releases Working Paper WP/26/204 on AI and aggregate labour productivity [7].
  • 2026 IMF/World Bank Annual Meetings: ILO statements call for productivity gains to translate into better jobs and incomes [3][4].
  • 20 November 2026: The Hindu op-ed by Azad and Chowdhary on AI, productivity and demand [1].

7. Prelims Hooks

  • At the global level, net exports are zero, so aggregate demand = C + I + G [1].
  • If productivity rises while real wages stay constant, the wage share falls [1].
  • Tugan-Baranovsky: "machines can produce machines which can produce more machines", i.e. investment can sustain itself [1].
  • Rosa Luxemburg: "machines are not an end in themselves", i.e. growth depends on consumption [1].
  • AI-related equipment investment has grown about 3.5 times since 2021 [1].
  • IMF's two AI–labour channels: displacement and complementarity [2].
  • IMF's "displacement" channel: tasks shift from labour to AI capital, which lowers labour income [2].
  • According to IMF research, women and highly educated workers face higher AI exposure, with high complementarity [6][5].
  • ILO: AI adjustment can show up as informality and falling real wages, not just unemployment [3].
  • India ranked 27th in Oxford's Government AI Readiness Index 2025 [9].
  • NITI Aayog's AI roadmap targets the informal workforce [10].
  • IMF WP/26/204 uses OECD patent data, 2000–2017 [7].

8. The 'Machines for Machines' Bet Is Being Tested in Real Time

  • Today's AI boom looks a lot like Tugan-Baranovsky's world
  • Big tech firms that run huge cloud data centres are called hyperscalers. Since 2023 they have led 14 public deals worth about $300 billion in equity and $650 billion in debt. The money is concentrated in a handful of AI firms [17].
  • The IMF warns about circular deals. A hyperscaler invests in an AI lab, and the lab then spends that money buying computing capacity from the same hyperscaler [16][17].
  • So part of today's AI demand is investment buying other investment. That is the "machines producing machines" path the article describes [1].

  • Fast-ageing chips make Luxemburg's point

  • Hyperscalers' accounts assume their equipment lasts about seven years. GPUs (the chips that run AI) may be out of date within two years [15].
  • That gives only a short time to recover the money. Final users, meaning households and firms paying for AI services, must pay for these machines quickly.
  • Without that consumer demand, the investment cannot pay for itself. This is Luxemburg's limit [1].

  • When an investment boom is paid for with debt, a slowdown hits harder

  • Hyperscalers' share of investment-grade bond issuance has risen from under 1% to 10%. (Investment-grade bonds are loans to borrowers rated safe.) [16]
  • The IMF says a sharp cut in AI capex (capital expenditure, i.e. spending on equipment and buildings), together with tighter credit, could leave data-centre lenders unable to roll over their loans [15].
  • The IMF lists weaker or shorter-lived AI investment and profits as a trigger for a market correction [17]. If investment is all that holds up demand and it stops, nothing else is there to take its place.

9. India's Wages Were Falling Behind Productivity Before AI

  • The wage share was already falling
  • According to the ILO, India's labour income share fell from 38.5% in 1981 to 35.4% in 2013. In 2008 it dropped as low as 31.7% [11].
  • The reason was the same as in the article: average labour productivity grew faster than average real wages [11].

  • The partial recovery has not closed the gap

  • The labour share recovered somewhat between 2011 and 2019. It still stayed well below its level in the early years after the 1991 reforms [12].

  • Many workers have no wage floor that actually works

  • About 62 million workers were paid less than the indicative national minimum wage [11].
  • If AI raises productivity in a labour market like this, the gains are even less likely to reach wages. Most workers have little power to bargain for a share.

  • This is not only an Indian problem

  • In high-income countries, labour productivity rose 29% between 1999 and 2024, while real wages rose only 15% [13].
  • The ILO links the long-run fall in the global labour share to technology that automates tasks [14]. AI could speed up a trend that is already under way.

10. The Best Case Against the Demand Worry, and Where It Falls Short

  • Counter-argument 1: AI makes things cheaper, so real wages rise anyway
  • If AI cuts costs and firms lower prices, the same money wage buys more goods. Real wages rise without any pay hike.
  • What is right: this does happen in competitive markets.
  • Where it falls short: it depends on firms passing on the savings. Productivity has outrun real wages for decades, even in rich countries [13]. The AI market is also concentrated in a few firms [17], and firms with that much market power usually keep more of the gains.

  • Counter-argument 2: AI creates new work that complements people

  • The IMF's complementarity channel is real. Demand rises for jobs that work alongside AI [2].
  • Where it falls short: in developing countries the costs may arrive before the benefits [3]. India's workforce is mostly informal and less exposed to these complementary roles. A gain in a few jobs does not lift the overall wage share.

  • Counter-argument 3: the government and exports can fill the gap

  • This is fair for one country. The authors leave G out on purpose [1], and for India net exports do count towards demand.
  • But: worldwide, exports cancel out [1]. Not every country can grow by selling to the others. Government spending can fill the gap only for as long as the government can afford to keep spending.

  • Honest verdict: the demand limit is not certain. But it is the most likely outcome unless policy makes sure wages keep up with productivity.

11. Who Must Act So Wages Keep Pace With AI

  • Ministry of Labour: make the wage floor real
  • The ILO's India Wage Report says strong wage policies are key to inclusive growth. This includes a minimum wage that actually covers workers [11].
  • The Code on Wages, 2019 creates a national floor wage. The test is whether it reaches the roughly 62 million workers who are paid below the indicative minimum [11].

  • Count AI's labour effects beyond unemployment

  • The ILO says adjustment shows up as falling real wages and more informal work, not only job losses [3].
  • So MoSPI's labour surveys and the Economic Survey [8] should track the wage share and the gap between productivity and wages, not just the unemployment rate.

  • RBI and financial regulators: watch lending to AI

  • The IMF flags data-centre lenders as exposed if AI capex falls sharply [15].
  • Indian banks and NBFCs (non-bank lenders) funding data centres in India face the same risk. Supervisors should map these exposures early.

  • Use public spending when private demand is weak

  • The article's model leaves out G [1]. In practice, public spending on skilling and social protection keeps workers' purchasing power up while the economy adjusts. It also helps workers move into AI-complementary jobs [2].

12. Anchors for Answers

  • Data: India's labour income share fell from 38.5% (1981) to 35.4% (2013) [11]
  • Data: In high-income countries, labour productivity rose 29% and real wages 15% between 1999 and 2024 [13]
  • Data: Since 2023, hyperscaler-led AI deals total about $300 bn in equity and $650 bn in debt [17]
  • Data: GPUs may become obsolete in about 2 years, against a roughly 7-year assumed life for hyperscaler assets [15]
  • Report/Committee: ILO India Wage Report [11]; ILO Global Wage Report 2024–25 [13]; IMF Global Financial Stability Report, April and October 2026 [15][17]; Economic Survey 2025-26, Ch. 14 [8]
  • Law/Case: Code on Wages, 2019 (national floor wage)
  • Comparison: The gap between productivity and wages exists even in high-income countries, which shows AI gains do not reach wages on their own [13]
  • Scheme: IndiaAI Mission and NITI Aayog's AI roadmap for the informal workforce [10], the skilling route to AI complementarity

13. Mains Relevance

14. Related Topics to Study Next

15. NCERT Concepts to Revise

  • Labour productivity (Economy): The article's starting point: AI raises output per worker, but the gains may not reach wages.
  • Real wage (Economy): Flat real wages while productivity rises is what shrinks the wage share and weakens consumption.
  • Aggregate demand (Economy): The core question is whether total demand (C + I) can keep up with AI-raised output.
  • Consumption (Economy): Workers spend most of their income, so slow wage growth slows consumption demand.
  • Investment (Economy): Tugan-Baranovsky argued investment can sustain itself; AI equipment investment is up about 3.5 times since 2021.
  • Net exports (Economy): Net exports cancel out globally, which is why the authors use only C and I.
  • Paradox of thrift (Economy): It is the basic example of demand-constrained growth, which is the logic of the under-consumption argument.
  • Automation (Economy): Labour displacement, where machines take over tasks and cut labour income, is the IMF channel the note stresses.
  • Artificial intelligence (Economy): AI is the technology whose effects on productivity, jobs and demand the whole note examines.
  • Inclusive growth (Economy): The ILO and NITI Aayog both want AI gains shared widely, including with informal workers.
  • Jobless growth (Economy): India's earlier experience of GDP rising without matching jobs is a precedent for AI-led growth.
  • Potential output (Economy): AI can raise what the economy could produce, but demand may stop actual output reaching it.
  • IndiaAI Mission (Economy): It is India's main institutional push on AI, including skilling, and a likely exam link.
  • Economic Survey (Economy): The Economic Survey 2025-26 AI chapter is the note's key official Indian source.
  • NITI Aayog (Polity): Its October 2025 AI roadmap focuses on the informal workforce and is a likely exam link.

16. Common Errors / Trap Areas

  • Mixing up productivity and wages: higher productivity does not automatically mean higher real wages. The gap between them is what lowers the wage share [1].
  • Leaving out trade "for globality": trade cancels out only at the world level. For a single country such as India, net exports are still part of demand [1].
  • Swapping the two theorists: Tugan-Baranovsky held that investment can sustain itself; Luxemburg held that consumption is the limit [1].
  • Measuring AI's labour impact by unemployment alone: the ILO also counts informality, falling real wages and weaker hiring [3].
  • Attributing the Government AI Readiness Index to NITI Aayog or the IMF: it is Oxford's index, cited in a PIB document [9].

Sources

  1. 1"If AI lifts productivity, will demand sustain growth?" by Rohit Azad and Indranil Chowdhary, The Hindu, 20 Nov 2026 (from the supplied excerpt)thehindu.com · tier 4
  2. 2IMF, Gen-AI: Artificial Intelligence and the Future of Work (SDN presentation, 2024)imf.org · tier 2
  3. 3ILO, "ILO calls for productivity gains to translate into better jobs and incomes"ilo.org · tier 2
  4. 4ILO, "Making productivity growth work for people"ilo.org · tier 2
  5. 5World Bank Jobs Blog, "What we're reading about the age of AI, jobs and inequality"blogs.worldbank.org · tier 2
  6. 6IMF Working Paper WP/23/216, AI exposure and complementarityimf.org · tier 2
  7. 7IMF Working Paper WP/26/204imf.org · tier 2
  8. 8Economic Survey 2025-26, Chapter 14 (AI)indiabudget.gov.in · tier 1
  9. 9PIB document, February 2026static.pib.gov.in · tier 1
  10. 10NITI Aayog, Roadmap on AI for Inclusive Societal Development (Oct 2025)niti.gov.in · tier 1
  11. 11ILO, India Wage Report: Wage policies for decent work and inclusive growthilo.org · tier 2
  12. 12ILO, India Employment Report 2024: Youth employment, education and skillsilo.org · tier 2
  13. 13ILO, Global Wage Report 2024-25ilo.org · tier 2
  14. 14ILO, World Employment and Social Outlook: September 2024 Updateilo.org · tier 2
  15. 15IMF, Global Financial Stability Report, April 2026, Chapter 1imf.org · tier 2
  16. 16IMF Blog, Global Financial Markets Face Rising Risks From AI, Bond Yields, and Cyberattacks (Oct 2026)imf.org · tier 2
  17. 17IMF, Global Financial Stability Report, October 2026, Chapter 1: Beneath the Resilienceelibrary.imf.org · tier 2

Mains Q&A on this note

NCERT concepts in this note

Also on 20 November

All 20 November articles →