The Broken Ladder: When Automation Arrives Before the Demographic Dividend

Bangladesh’s garment success helped lift millions; automation is now testing how long that model can last.

1. When the Ladder Breaks: Bangladesh's Disruption

Bangladesh's garment sector employs 4.22 million workers and has contributed to one of the fastest poverty reductions in history. Over the course of three decades, ready-made garment (RMG) exports transformed millions of rural workers, particularly women, into industrial operatives. This was the ladder: labor-intensive manufacturing absorbs youth bulges, generates foreign exchange, and buys time for a country to move up the value chain.

But the ladder is breaking. A 2024 study by the Bangladesh Labour Foundation and BRAC University found that automation has reduced the RMG workforce by 31 percent in a single year. The cutting phase saw 48 percent job losses; sewing, 27 percent. Factory owners plan to accelerate this: 80 percent intend to invest in automation in a span of two years. Meanwhile, job creation has collapsed from 300,000 annually to 60,000. Bangladesh faces a generational crisis not because its manufacturing failed, but because it succeeded at precisely the moment when the technology that made success possible now makes workers expendable.

This is not an outlier. It is the warning signal for a global problem: automation arrives before demographic dividends are realized. Labor-abundant economies that planned to ride youth bulges into development now face premature job destruction. The development sequence that sustained East Asian growth is breaking. And without deliberate state intervention to govern human-AI complementarity, youth bulges transform from engines of growth into reservoirs of political instability.

2. The Demographic Dividend That Assumes Absorption

The demographic dividend is deceptively simple: when the working-age population (15–64 years) grows faster than dependents, the dependency ratio falls, and savings rise. If jobs absorb this surplus labor, growth accelerates. If they don't, it doesn't.

Bangladesh's numbers reveal the opportunity and the trap. Working-age population share peaked at 67.6 percent in 2019, driven by fertility declines over the previous two decades. The demographic component contributed approximately 18 percent of per capita GDP growth over three decade-real gains. Yet this came with a structural assumption: that 1.83 million new labor force entrants annually would find productive work.

The historical template worked because manufacturing was labor-intensive and expanding. In post-war Japan and South Korea, industrial employment grew alongside output. Women entered garment factories not despite low skill requirements but because of them; the sector's labor-absorptive capacity was its competitive advantage. As workers accumulated experience, they transitioned into higher-value roles or moved into services. The ladder was continuous.

But discontinuity now prevails. Demographic dividends require not only a young population but also active labor absorption. Countries that fail this shift, where working-age youth cannot find productive employment, face what development economists term a "demographic liability": the reserve army grows, informal employment expands, and the state loses its primary channel for generating growth, legitimacy, and organization. Political scientists have documented the link: democratization depends partly on industrialization because organized labor provides the discipline and coordination that holds democratic institutions together. Bangladesh can still harness its demographic window. Yet this requires understanding why the traditional channel, labor-intensive manufacturing, is now blocked.

3. Technology Arrives Too Early: Premature Automation in Bangladesh

The automation of Bangladesh's garment sector is not primitive mechanization, the simple substitution of machines for hands. It is AI-enabled and multifaceted: automated cutting systems, algorithmic quality control, and platform-mediated supply chain pressure from global buyers who incentivize factories to reduce headcount.

Three dynamics collide. First, rising wages for low-skilled workers make automation economically attractive. Bangladesh's relative advantage, low labor costs, faces pressure as wages rise and technology costs fall. Factories that previously operated with 500 workers now operate with 200 or more workers. Second, global supply chains directly incentivize job shedding. Multinational buyers impose compliance deadlines, quality specifications, and cost targets that force suppliers to choose between raising prices or cutting labor. Automation becomes the inevitable solution. Third, AI deployment is concentrated in precisely the tasks that absorb unskilled workers: cutting, sorting, and pattern matching.

This is where Dani Rodrik's concept of premature deindustrialization becomes urgent. Deindustrialization typically occurs at high income levels, after countries have accumulated capital and moved labor up the value chain. Bangladesh is deindustrializing at lower income levels, earlier in its development sequence. Labor-saving technological progress in manufacturing is shrinking employment before the sector has absorbed its youth bulge. Rodrik's political warning is stark: premature deindustrialization makes democratization less likely and more fragile. Without organized labor, political parties lose the discipline and coordination they need to sustain democratic institutions.

Bangladesh's case is not unique. India faces identical pressures. Nigeria, facing a 77 percent youth population share by 2050, confronts automation before industrialization has even begun. Indonesia, Vietnam, and Kenya watch the same dynamics unfold. The gauntlet of global value chains tightens as automation's cost falls and its capabilities expand. What was gradual in the North would be compressed in the South.

The central paradox: Bangladesh automated because it succeeded. Its competitive advantage in labor-intensive manufacturing made it a target for automation. Success breeds its own obsolescence.

4. Political Consequences: Instability as a Rational Outcome

The political fallout is not merely an externality of economic dislocation; it is a rational outcome of unabsorbed youth cohorts.

Employment is the primary mechanism through which young people transition to stable social roles: earning income, forming households, acquiring assets, and developing stakes in institutional stability. When this mechanism fails on a scale, political economy shifts. The reserve army grows. Informal employment expands, with limited protections and chronic income insecurity. Social anxiety rises, particularly among the most educated cohort. Young Bangladeshis achieve higher educational attainment than any previous generation yet cannot find work commensurate with their credentials.

In labor-abundant economies, this creates a specific vulnerability. Where organized labor traditionally anchored political parties and provided governance capacity, unabsorbed and precarious workers lack collective structures. They are accessible to extremist recruitment, sensitive to populist narratives, and disconnected from state institutions. Migration pressures intensify. Domestic political contagion becomes more likely: grievances in one labor-abundant country cascade to others through remittance networks and diaspora communication.

Bangladesh's recent experience shows the pattern. Youth unemployment and persistent underemployment in informal sectors correlate with protest volatility. A youth cohort unable to translate education into employment becomes a vulnerability for any government.

The geopolitical dimension follows. Unabsorbed youth in strategic manufacturing economies such as Bangladesh, Vietnam, and Indonesia create pressure for migration, remittance-dependent policy, and susceptibility to external actors offering alternate employment or security frameworks. The development model that once translated demography into geopolitical stability now risks the reverse.

5. Governing Human-AI Complementarity: The Core Intervention

Rather than resist automation, an approach that fails both economically and politically, states must actively govern the substitution of machines for labor by reshaping incentives toward human-AI complementarity. This requires three coordinated instruments.

First: Labor-Augmenting Industrial Policy. Bangladesh's Board of Investment should tie automation incentives explicitly to net employment outcomes. Rather than subsidizing capital accumulation indiscriminately, industrial policy should reward technologies and production methods that increase worker productivity while preserving (or expanding) employment. This is not Luddism; it is market-shaping. The Philippines case demonstrates viability: by deliberately biasing investment incentives toward labor-intensive sub-sectors of manufacturing, policymakers influenced sectoral composition without imposing prohibitions. Bangladesh can adopt similar mechanisms: faster tax write-downs for employment-intensive automation; restricted access to automation finance for factories reducing headcount; regional dispersal incentives that direct investment toward labor-surplus areas.

The International Monetary Fund's 2025 analysis shows industrial policy's modest but consistent effects, approximately 0.5 percent value-added improvement when paired with institutional strength. The key is targeting-undiscriminating subsidies fail; employment-linked incentives work. Bangladesh should establish an automation review process—not a veto power but a transparency mechanism requiring factories to document the employment impact of major capital investments. South Korea's industrial policy in the 1970s–80s operated similarly: visible state coordination without command-and-control authoritarianism.

Second: Transitional Public Employment as Stability Infrastructure. Public sector employment has negative connotations in development circles, associated with waste, political patronage, and fiscal bloat. Yet this conflates design with concept. Framed as temporary infrastructure responding to technological disruption, public employment becomes a risk-mitigation tool.

Bangladesh should pilot a time-bound public employment guarantee in garment clusters: workers displaced by automation receive guaranteed access to public works (infrastructure, rural development, maintenance services) at prevailing wage levels, with a sunset clause tied to private sector job creation. This is not welfare; it is stability management. It maintains income, prevents reserve-army formation, and gives private employers time to create new roles. The constraint is fiscal sustainability: such schemes require either increased revenue or reallocation from unproductive spending. Bangladesh's current public employment is already substantial; recalibration, not expansion at any cost, is the lever.

Third: Human-AI Deployment Mandates. Global supply chains create asymmetric pressure: buyers incentivize automation without bearing adjustment costs. Bangladesh should adopt supplier standards explicitly tied to workforce transition. Multinational buyers, who already impose environmental, safety, and labor standards, should face contractual obligations to fund transition support proportional to workforce reduction. This formalizes what is economically true: the productivity gains from automation create surplus value; that surplus should be partially directed toward maintaining social stability in supplier regions.

This requires enforcement mechanisms. Bangladesh's nascent platform economy offers a model: digital platforms for informal workers (like Kormo) can be repurposed. Rather than simple job-matching, platforms can serve as transition infrastructure, tracking displaced workers, linking them to reskilling, channeling employer contributions, and coordinating public employment access. This digitalization of the labor market transition becomes a public good, not a private service.

6. Risks, Trade-offs, and Failure Modes

This intervention framework requires an honest assessment of constraints.

Fiscal strain is real. Bangladesh's public revenue as a percentage of GDP lags peer economies. Sustaining public employment requires either increased taxation (politically difficult) or reallocation from military, bureaucratic, or subsidy spending. The IMF's industrial policy analysis shows that untargeted sectors shrink when investment is diverted to priority industries, a real economy-wide cost. Bangladesh cannot indefinitely maintain labor-intensive manufacturing without forgoing higher-value sectors.

State capacity is limited. Industrial policy's success depends on institutional strength, effective monitoring, low corruption, and political insulation from short-term pressures. Bangladesh's civil service, while improving, remains under stress. Automation reviews, if poorly designed, become corruption vectors. Public employment schemes, if captured by political patronage, generate fiscal waste. The state must build institutional capacity alongside policy deployment.

Protectionism and trade retaliation pose risks. Employment-linked incentives border on local content requirements, instruments that trigger scrutiny and retaliatory pressure from trading partners. Bangladesh must design policies that shape incentives without formally restricting trade. This is technically possible but politically harder.

Failure modes include both ineffectiveness and over-reach. An industrial policy that succeeds only in protecting uncompetitive sectors prolongs stagnation. Conversely, state intervention that becomes too extensive crowds out private dynamism. Bangladesh's path requires calibration: protecting employment during transition, not permanently. The sunset clause matters. Public work must raise productivity, not simply employ, reskilling bundled with employment.

7. Why This Matters Beyond Bangladesh

Bangladesh's dilemma is not exceptional; it is generative. Across labor-abundant Asia, Africa, and parts of Latin America, automation is arriving before demographic dividends are absorbed. Nigeria, Pakistan, and Indonesia face identical pressures. So does Egypt, where youth unemployment exceeds 25 percent, and manufacturing capacity remains underutilized.

The development model assumed by the World Bank, IMF, and bilateral donors, labor-intensive export manufacturing as a transitional stage, is encountering technological closure. Without active state intervention to reshape human-AI complementarity, the transition becomes a trap: young cohorts enter the labor force but find neither formal manufacturing jobs nor the public services and informal sectors that once absorbed them. This generates global migration pressure, political instability, and potential geopolitical contagion.

More broadly, the premise of the current international economic order, that free markets and trade will generate development trajectories similar to the East Asian experience, requires recalibration. It is assumed that technological cycles would unfold sequentially (capital first, then automation), with time for institutional adaptation. That assumption has collapsed. Automation arrives compressed. States that do not actively govern their sequencing and distribute their costs face social breakdown.

For multilateral institutions, this signals that development policy must shift from deregulation toward active state coordination. For states, it requires honest budget trade-offs, institutional investment, and political will to prioritize youth employment as a development imperative. For global supply chains, it means recognizing that stability in production networks depends on socially sustainable transitions in supplier economies.

Bangladesh illustrates the choice starkly: states can design labor-augmenting industrial policy and transitional public employment now or manage political instability later. The ladder can be repaired, but only through deliberate governance of technology's arrival and effects. That is the lesson for leaders facing disruption at the intersection of demography, technology, and politics. It is uncomfortable. It requires state capacity and political courage. But the alternative, passive acceptance of premature deindustrialization and its consequences, is untenable for countries where billions still depend on manufacturing for their development. -Ganit

REFERENCES

1. Bloom, D. E., Canning, D., & Sevilla, J. (2003). "The Demographic Dividend: A New Perspective on the Economic Consequences of Population Change." Population and Development Review, 29(S1), 83-109.

2. UN DESA. (2023). "World Population Prospects 2022 Revision: Demographic Profile of Bangladesh." United Nations Department of Economic and Social Affairs.

3. PMC NCBI. (2023, January 12). "Harnessing Demographic Dividend Before it is Lost." PubMed Central. https://pmc.ncbi.nlm.nih.gov/articles/PMC9839200/

4. Rodrik, D. (2016). "Premature Deindustrialization." Journal of Economic Growth, 21(1), 1-33. https://www.ias.edu/sites/default/files/sss/pdfs/Rodrik/Research/premature-deindustrialization.pdf

5. Acemoglu, D., & Restrepo, P. (2018). "Artificial Intelligence, Automation and Work." NBER Working Paper w24196. https://www.nber.org/system/files/working_papers/w24196/w24196.pdf

6. Columbia University Law School, Millstein Center. (2024). "Artificial Intelligence, Automation and Work: Policy Translation." https://millstein.law.columbia.edu/

7. WEF (World Economic Forum). (2024, September 9). "Augmented Workforce: Empowering People, Transforming Manufacturing." https://www.weforum.org/publications/augmented-workforce-empowering-people-transforming-manufacturing/

8. Bangladesh Labour Foundation, BRAC University, Solidaridad Network Asia. (2024). "Automation and Its Impact on Employment of RMG Sector."

9. BRAC University Institute of Governance and Development. (2024, August). "Digitalisation in the Informal Sector: Through the Lens of Kormo." https://bigd.bracu.ac.bd/wp-content/uploads/2024/08/Conference-Paper_Kormo.pdf

10. ILO (International Labour Organization). (2024, December). "Global Employment Trends for Youth 2024." https://www.donorplatform.org/post/publications/ilo-global-employment-trends-for-youth-2024/

11. IMF (International Monetary Fund). (2025, November 24). "Industrial Policy Can Lift Productivity—but Comes With Risks and Trade-offs." https://www.imf.org/en/blogs/articles/2025/11/25/industrial-policy-can-lift-productivity-but-comes-with-risks-and-trade-offs

12. World Bank. (2012). "Public Employment and Governance in Middle East and North Africa." Policy Research Working Paper. https://documents1.worldbank.org/curated/en/339381472236811833/pdf/

13. OECD. (2025). "The Political Economy of Stabilization Programmes in Developing Countries." https://www.oecd.org/en/publications/the-political-economy-of-stabilization-programmes-in-developing-countries_802526617102.html

14. World Bank. (2012). "Transition Toward More Rapid and Labor-Intensive Growth: A Case Study Approach." https://documents1.worldbank.org/curated/en/939591468780910653/pdf/

15. Fair Work Foundation. (2023, June 13). "Bangladesh Needs a Policy Framework for Growing Platform Economy." https://fair.work/en/fw/blog/bangladesh-needs-a-policy-framework-for-a-growing-platform-economy/

16. ILO (International Labour Organization). (2024, December 31). "Global Labour Market Rebounds, but Youth Unemployment at 15-Year Low—Regional Disparities Persist." https://gcap.global/news/global-labour-market-rebounds-but-youth-unemployment-at-15-year-low-regional-disparities-persist-ilo-re/

17. WEF (World Economic Forum). (2024, August). "Global Youth Employment: This is the State of Global Youth Employment in 2024." https://www.weforum.org/stories/2024/08/global-youth-employment-future-jobs/