When a small birth cohort enters the labor market two decades later, employers may receive fewer applications. A hospital can struggle to fill a night shift while a manufacturer keeps an older technician because no apprentice knows the machine. This is a conditional pressure, not an economy-wide law: output, employment, and care demand also depend on age structure, migration, participation, and productivity. The economy has not run out of people; in some sectors it is asking fewer workers to carry more of what earlier generations built.
Scarcity can be good for workers. Employers may raise wages, loosen hiring requirements, or buy equipment that makes each employee more productive. A country with fewer people can remain rich: GDP may fall while GDP per person, wages, or some measures of material comfort rise. Productivity means more output per hour or worker; it is not the same thing as service access, fiscal capacity, or household welfare. I expect many low-fertility countries to remain comfortable for a long time.
The pressure appears in what the remaining workers have to maintain. Older people need care, but chronological age is not functional ability: healthy life, disability, and years requiring intensive support vary widely. Roads and water systems do not shrink neatly with the population; existing businesses still need someone who understands how they work. As more effort goes toward keeping inherited systems running, there may be less discretionary capacity for expansion and experimentation—the opportunity cost of having one technician also serve as the only trainer, or one hospital shift operate without a reserve.
A capable society has a substantial adaptation menu: higher participation among older adults and people previously excluded, safer and more flexible jobs, selective later work, training, capital deepening, automation and AI, migration, consolidation of facilities, prevention, healthier aging, and redesigned care. Capital deepening includes equipment, software, buildings, transport, energy systems, and organizational redesign, not only robots. Migration can enlarge a receiving country’s workforce and tax base, while redistributing workers and requiring housing, credentialing, language, and integration capacity; it does not solve global aging. Consolidation can lower unit costs in dense areas while making distance and access worse in shrinking ones. These measures can preserve GDP per person and many services for decades, and may improve wages or housing access in some places. They do not create a successor generation, and service access can still worsen when distance, geography, or missing workers is the constraint.
Work becomes scarce
Care is exposed early because demand rises with age while much of the work remains personal. A factory can install robots; a frail person still needs someone patient enough to help her stand.
Higher wages can pull more people into paid work, flexible jobs may keep some older adults employed, and better training can open work to people employers once screened out, including people with disabilities or interrupted careers. The OECD estimates that the working-age population across its members will decline by 8 percent by 2060. That is a projection of a defined age group, not a forecast that employment or output must fall by the same amount. Participation gains can offset part of a demographic decline, but they are constrained by health, care responsibilities, discrimination, and job quality.
Longer working lives can make sense when healthy life has lengthened, provided physically worn-out workers are not simply told to endure another five years. Healthy-life averages do not justify one retirement age for everyone. The relevant reform may be a safer job, fewer hours, or a different role, with stronger protection for people in strenuous work or with lower lifetime earnings.
Japan’s labor ministry documents one organizational response at a care provider in Mie Prefecture. Welfare Mie divided direct care from supporting tasks and created dedicated night-shift roles, opening some work to older and disabled employees while trying to retain experienced caregivers. The case describes a work design, not an evaluated national productivity intervention. It did not automate the relationship; it reorganized tasks around the people available to do them.
Automation buys capacity
Demographic pressure gives firms a reason to substitute machines for scarce labor. Daron Acemoglu and Pascual Restrepo found that, in their cross-country study and period, countries aging more rapidly adopted industrial robots more quickly and did not grow more slowly. That is evidence of an association in the sample, not proof that automation generally absorbs demographic shocks or that every sector can substitute machines for people.
Machines take a task, 2011 A robot installs a laminated glass roof at Tesla’s factory in Fremont, California. Aging countries have adopted industrial robots more quickly; the limits appear where judgment and responsibility cannot be reduced to a task. via Wikimedia Commons
That is encouraging. A warehouse with autonomous equipment can move more goods with fewer workers. Artificial intelligence may raise output per worker for selected tasks, contingent on complementary capital, skills, data, energy, and governance; its effects remain uncertain and uneven. Productivity is a promising way for a smaller workforce to maintain a large inherited stock of capital and services, but it does not automatically supply judgment, accountability, or embodied care.
The limits appear where responsibility and relationships cannot be reduced to a task. A robot can move a box, and software can assist a clinician, without taking full responsibility for the warehouse or the patient. Even a highly automated economy needs people who notice failures, make judgments, supervise systems, and decide what should be built next.
Productivity and the economic margin
Demographic change can split asset markets. In shrinking towns, inherited homes lose value as buyers disappear; in prosperous cities, housing can remain expensive because younger workers keep concentrating there. Consumer markets divide in the same way: national demand may soften while the places still attracting workers remain crowded. These are economic reallocations, distinct from the public cost of maintaining water lines or financing care taken up in the pages that follow.
Innovation raises a longer and less certain question. Modern discovery often requires teams, apprenticeships, and experimental slack. Research by Nicholas Bloom, Charles Jones, John Van Reenen, and Michael Webb found that the number of researchers needed to sustain Moore’s Law was more than eighteen times larger than in the early 1970s; their study concerns research effort and productivity in selected technologies, not a direct population effect. Across their cases, rising research effort offset declining research productivity.
That result does not prove that a smaller population innovates less. Education, freedom, capital, communication, and the quality of institutions determine whether potential talent ever reaches a laboratory, and artificial intelligence may let a small team attempt work that once required a large one. Population affects the pool from which teams and experiments can be assembled; it does not decide what those people will discover.
None of this implies that output per person must fall. The problem is the margin left after care and maintenance. A productive society can keep inherited systems running with fewer people for decades; under repeated contraction, it may have to repeat that achievement with fewer hands and less redundancy. That is an interpretation about discretionary and civilizational capacity, not a measured forecast of GDP or innovation. Its success should make us less fatalistic about adaptation, not indifferent to recurrence.
There is a possible feedback risk here too. Scarce younger workers may face intense career demand just as they are trying to form families, while taxes and care obligations take more of their income and time. Wage gains, housing costs, policy, and family support can push in different directions. Adaptation that protects today’s standard of living could make the next generation harder to raise unless governments, employers, and families notice the trade.
Productivity can buy enormous time. What it cannot do by itself is produce the successors who decide how to use it. That is the boundary between measured productivity and the less measurable capacity for succession, redundancy, and new projects.
Citations
- Organisation for Economic Co-operation and Development, “OECD job markets remain resilient but population ageing will cause significant labour shortages and fiscal pressures”, July 9, 2025.
- Japan Ministry of Health, Labour and Welfare, “Responding to Labor Shortages”, in Analysis of the Labour Economy 2024, 2024.
- Daron Acemoglu and Pascual Restrepo, “Secular Stagnation? The Effect of Aging on Economic Growth in the Age of Automation”, American Economic Review 107, no. 5, 2017.
- Nicholas Bloom, Charles I. Jones, John Van Reenen, and Michael Webb, “Are Ideas Getting Harder to Find?”, American Economic Review 110, no. 4, 2020.