Opinion says India's real challenge is capital allocation, not jobs
India faces a capital-allocation problem rather than a jobs problem, argue Nasscom co-founder Harish Mehta and Centre for Innovation in Public Policy founder K Yatish Rajawat. They note India's capital stock grew 74% in a decade while employment rose 36%. Under the Production-Linked Incentive scheme, over ₹2.40 lakh crore across 14 sectors generated 14.15 lakh jobs, about ₹17 lakh of capital per job. They urge incentives be redirected to labour-intensive sectors.
Source
News18 — India · read the original report ↗
Desk check · some claims need care
What the desk checked (5)
- Over ₹2.40 lakh crore invested under PLI across 14 sectors generated 14.15 lakh direct and indirect jobs as of March 2026, about ₹17 lakh of capital per job. — Figures appear in the source and are attributed to PLI scheme data; the March 2026 reference date is forward-looking relative to typical reporting and needs verification.
- Food processing PLI outlay of ₹9,207 crore generated 3.35 lakh jobs, exceeding its target. — Figure appears in source, attributed to the same PLI dataset; no document citation given.
- India installed 9,100 industrial robots in 2024, overtaking Germany, as global installations reached 542,000 units. — Appears in source; an auto-generated section in the same article states 9,123 robots, an internal inconsistency an editor should resolve.
- Each additional robot per thousand workers cuts local employment by 0.2 percentage points. — Attributed in source to Acemoglu and Restrepo, Journal of Political Economy, 2020.
- US manufacturing employment fell by 2.4 million jobs (17%) between 2002 and 2022; Stanford study found 13% relative employment decline among 22-25-year-olds in AI-exposed jobs. — Attributed to US data and Stanford Digital Economy Lab respectively; not independently verified.
Analysts’ view opinion
The core argument here is an allocation argument, not a pessimism argument: capital is still arriving, but each rupee of it now buys far fewer jobs than industrial policy assumes. The PLI numbers cited make that visible in an unusually concrete way — roughly ₹17 lakh of capital per job on average across 14 sectors, against wages the authors put at not much more than ₹4 lakh a year — while food processing delivered 3.35 lakh jobs off a far smaller ₹9,207 crore outlay. The honest reading is that the capital-employment link has weakened structurally, and that the sector mix of incentives now matters more than the headline investment total.
- The decade's mismatch is the headline economics: capital stock up 74% against employment up 36%, with manufacturing's employment share stuck near 12% since 1991.
- Within the same PLI scheme, capital-per-job varies enormously by sector, which suggests automatability — not incentive design alone — is doing much of the work.
- Who gains and who pays is the real distributional question: capital-intensive, automated plants concentrate output and profits, while labour-absorbing sectors spread incomes more widely.
- The authors' MSME-plus-AI proposal is essentially a cost-of-overhead argument — cheaper compliance, sales and accounting could let more small units stay viable and hire, though the story offers no estimate of that effect.
- The measurement critique lands: GDP and investment totals can keep rising while jobs per rupee fall, so capital-per-job and district-level business survival data would be a more useful policy dashboard.
What to watch — Watch whether the next generation of incentive schemes starts publishing sector-wise capital-per-job ratios and tilts outlays toward labour-absorbing sectors such as food processing, tourism and small services — that would be the first hard sign the argument has been accepted.
This is an opinion piece by two named authors, not official policy or new reporting; it does not establish causation between automation and India's specific job numbers, the Indian IT entry-level slowdown is described as not yet conclusive, and some cited figures are forward-dated, so they should be read as the authors' claims rather than settled data.
Deep dive
Research brief · 8 facts · 8 dates · exam-readyThe brief
Context
An opinion piece by Nasscom co-founder Harish Mehta and Centre for Innovation in Public Policy founder K Yatish Rajawat argues that India's industrial and incentive policy still assumes that capital investment automatically creates jobs. They contend that robotics and now AI robotics are substitutive rather than complementary to labour, breaking that link. Using Indian manufacturing data and the Production-Linked Incentive (PLI) scheme's own disclosures, they argue India has a capital-allocation problem, not a jobs problem, and should redirect incentives towards labour-absorbing sectors and AI-enabled small firms.
Key facts
- India's capital stock grew 74% over the last decade while employment grew only 36%, per the authors.
- Manufacturing's share of employment in India has been stuck near 12% since 1991.
- Labour's share of manufacturing value added nearly halved from 22.2% to 14.3% between 2000-01 and 2011-12, even as output rose.
- Over Rs 2.40 lakh crore invested across 14 PLI sectors generated 14.15 lakh direct and indirect jobs as of March 2026 - roughly Rs 17 lakh of capital per job, for work paying on average not much more than Rs 4 lakh a year.
- PLI food processing: a much smaller outlay of Rs 9,207 crore generated 3.35 lakh jobs, beating its target, versus far more capital-intensive electronics.
- Global robot installations more than doubled over the past decade to 542,000 units in 2024; India installed 9,100 robots in 2024, overtaking Germany as the fastest-growing major market (a record 9,123 units cited in the article's quick answers).
- US manufacturing employment fell by 2.4 million jobs (17%) between 2002 and 2022 even as the economy grew and capital flowed into plants automated by design.
- Acemoglu and Restrepo (Journal of Political Economy, 2020): each additional robot per thousand workers cuts local employment by 0.2 percentage points, independent of trade or offshoring.
Timeline
- Since 1991India's manufacturing share of employment remains stuck near 12%.
- 2000-01 to 2011-12Labour's share of manufacturing value added falls from 22.2% to 14.3%.
- 2002-2022US manufacturing employment declines by 2.4 million jobs (17%).
- 2020Acemoglu and Restrepo publish robot-employment findings in the Journal of Political Economy.
- 2024Global robot installations reach 542,000 units; India installs 9,100 robots, overtaking Germany.
- Five straight months through mid-2026Challenger, Gray & Christmas data shows AI as the leading stated reason for US job cuts.
- March 2026PLI cumulative figures cited: over Rs 2.40 lakh crore investment, 14.15 lakh jobs across 14 sectors.
- September 19, 2026Opinion piece published (first published 20:33 IST).
Who has a stake
- Policymakers designing industrial and AI incentives — Asked to publish capital-per-job ratios sector by sector and use them, not headline investment totals, as a scheme design metric.
- PLI scheme beneficiary sectors (14 sectors, incl. electronics, food processing) — Capital-to-job ratios differ sharply by automatability, affecting whether incentives translate into employment.
- MSMEs and one-to-ten-person companies — Identified as India's real jobs multiplier if AI lowers overheads in sales, accounting, design, compliance and customer service.
- Entry-level workers and young jobseekers — Stanford Digital Economy Lab found a 13% relative employment decline among 22-25-year-olds in the most AI-exposed US occupations; Indian IT shows early similar signs.
- Indian IT firms — Early signs of an entry-level hiring slowdown, though the data is not yet conclusive per the authors.
- Authors: Harish Mehta (Nasscom co-founder) and K Yatish Rajawat (CIPP), both HMJM founder members — Advocating a redesign of incentives, AI infrastructure and measurement around jobs.
Why it matters
If a rupee of capital no longer reliably buys a job, India's core policy instruments - investment-linked incentives and GDP growth as a success metric - may keep expanding output while employment lags. The PLI data cited show the same incentive structure producing opposite capital-to-job ratios depending on how automatable a sector is, which has direct consequences for India's employment targets and for young entrants to the workforce.
UPSC angle
Prelims pointers
- Production-Linked Incentive (PLI) scheme: over Rs 2.40 lakh crore invested across 14 sectors, 14.15 lakh direct and indirect jobs as of March 2026.
- PLI food processing: Rs 9,207 crore outlay, 3.35 lakh jobs, target exceeded - lowest capital-per-job among cited sectors.
- India installed 9,100 industrial robots in 2024, overtaking Germany; global installations 542,000 units in 2024.
- Acemoglu and Restrepo (JPE, 2020): one more robot per 1,000 workers lowers local employment by 0.2 percentage points.
- India's manufacturing employment share has stayed near 12% since 1991; labour's share of manufacturing value added fell from 22.2% (2000-01) to 14.3% (2011-12).
- Nasscom co-founder Harish Mehta and CIPP's K Yatish Rajawat are founder members of the Hundred Million Jobs Mission (HMJM).
Mains framing
The authors argue that India's industrial policy rests on an outdated assumption - that capital formation mechanically generates employment. Earlier automation was complementary to labour; the current robotics-plus-AI wave is substitutive, with AI robots able to observe, react and multi-task, including dexterous work like packaging. The evidence cited is domestic as well as foreign: capital stock up 74% against 36% employment growth, manufacturing's employment share flat near 12% since 1991, labour's share of manufacturing value added halving, and the PLI scheme yielding roughly Rs 17 lakh of capital per job overall while food processing produced 3.35 lakh jobs from just Rs 9,207 crore. Services are not insulated - Stanford Digital Economy Lab payroll data showed a 13% relative employment decline among 22-25-year-olds in the most AI-exposed US occupations, and Nvidia's physical AI and SAP's Joule agents are being embedded into industrial robots. The proposed way forward has three parts: redirect incentives towards labour-absorbing categories such as food processing, tourism and restaurants, and publish capital-per-job ratios for every scheme; build Indian AI infrastructure (affordable compute, open models, Indian-language datasets) to enable AI-using MSMEs and micro-firms as a jobs multiplier, favouring thousands of small teams over two or three national champions; and shift measurement from GDP to district-level data on business formation, jobs, incomes and firm survival. The authors concede no incentive design can fully restore a link automation economics has structurally broken.
Key terms
- Production-Linked Incentive (PLI) scheme
- Indian incentive scheme across 14 sectors; cited with over Rs 2.40 lakh crore investment and 14.15 lakh jobs as of March 2026.
- Capital-per-job ratio
- Capital invested for each job created; about Rs 17 lakh per job under PLI overall, far lower in food processing.
- Substitutive vs complementary automation
- Complementary machines raise a worker's output; substitutive AI robots replace the worker entirely, per the authors.
- Physical AI / Joule agent platform
- Nvidia's physical AI models and SAP's Joule agents being built into industrial robots, extending AI from screens to shop floors.
- Hundred Million Jobs Mission (HMJM)
- Initiative of which both authors are founder members (www.hundredmillionjobs.org).
- Centre for Innovation in Public Policy (CIPP)
- Policy body founded by co-author K Yatish Rajawat.
Practice questions
- "India's challenge is capital allocation, not job creation." Critically examine this claim using evidence from India's PLI scheme and manufacturing employment trends.
- How does the shift from complementary to substitutive automation alter the design of industrial policy incentives in a labour-surplus economy like India? Discuss.
- Is GDP growth an adequate proxy for employment health in the AI era? Suggest alternative measurement frameworks for Indian policymakers.
Grounded only in the source report — figures and dates are the source's, not inferred.