Skip to main content

Productivity in an aging society

One of the great paradoxes of aging economies is that they grow slower, not faster, even though common sense suggests older, more experienced workers should be more productive. In fact, aging workforces and aging populations create structural headwinds to productivity growth. These headwinds include reduced investment in technology, lower entrepreneurship rates, slower adoption of innovation, higher costs for elderly support, and constrained government budgets. Understanding why aging slows productivity growth is essential to understanding why Japan, Germany, and other aging developed countries face such formidable economic headwinds, and why the global growth slowdown is partly demographic, not just cyclical.

Quick definition: Productivity growth in aging societies slows because older workers invest less in learning new technologies, firms invest less in capital and R&D (due to budget constraints and uncertain demand), and entrepreneurs are concentrated among younger age groups.

Key takeaways​

  • Aging does not increase individual worker productivity; if anything, productivity per worker tends to decline slightly with age due to reduced learning capacity and technological adaptability.
  • Firms reduce capital and R&D investment when facing an aging workforce, because expected payoff horizons shorten and skill-mismatch becomes more acute.
  • Innovation and entrepreneurship are concentrated in younger age groups, so aging workforces see lower startup rates and slower adoption of new technologies.
  • Government budget constraints in aging societies reduce public investment in infrastructure, education, and research—inputs that drive broad-based productivity growth.
  • Aging creates compositional shifts in employment, moving workers from high-productivity sectors (manufacturing, tech) to lower-productivity sectors (healthcare, elder care).
  • Aging is associated with lower aggregate productivity growth across OECD countries—Japan, Germany, and Italy all experienced productivity slowdowns as they aged.

How individual worker productivity changes with age​

The relationship between age and individual worker productivity is complex and depends heavily on the task. For routine, experience-based tasks (legal review, project management, skilled trades), older workers often have higher productivity because they have deeper expertise. For tasks requiring rapid learning or physical exertion, productivity often declines with age.

Research on productivity by age shows:

Peak productivity years: Most studies find that worker productivity peaks between ages 40–50. Before that, workers are still learning and building skills; after that, physical decline and reduced learning capacity begin to offset experience gains.

Task-dependent variation: In routine, experience-based work, older workers remain productive into their 60s. In physically demanding work (construction, nursing, agricultural labor), productivity declines more noticeably after age 55–60. In innovation-intensive work (software engineering, scientific research, design), productivity may decline earlier, as rapid technological change makes existing expertise partially obsolete.

The learning problem: As technologies change, older workers must retrain. But they often have reduced incentive to invest in retraining (fewer years to recoup the cost) and may find it harder to learn new systems. Firms similarly have reduced incentive to train older workers—an investment is worth making only if the worker will remain in the job long enough for the firm to recoup the training cost.

Tenure and wage premium: Older workers typically earn higher wages than younger workers with similar job titles, because wages reflect tenure and experience. If older workers are paid more but not proportionally more productive, firms face a wage-productivity gap. This can lead firms to prefer laying off older workers or not hiring them in the first place.

Firm-level capital investment declines with aging workforces​

When firms face an aging workforce, they often respond by reducing capital and technology investment, not increasing it. This seems counterintuitive—wouldn't firms invest more in automation to offset labor scarcity? Sometimes, but the overall pattern in OECD data shows the opposite.

Why? Several mechanisms:

Shortened payoff horizons: A piece of equipment that costs $1 million and lasts 20 years is worth investing in only if the firm expects to operate it for enough years to recoup the cost. If a firm is uncertain about future demand (as many are when facing labor shortages and potential recession), or if the equipment requires workers trained in skills that are scarce, the expected payoff horizon shrinks. The firm defers investment.

Uncertainty and demand: Aging economies are often low-growth economies (Japan, Germany), because demographic headwinds reduce consumption growth. Firms facing stagnant or slowly growing demand have less reason to invest. They maintain existing capital but do not expand. This creates an underinvestment trap.

Skill mismatch: New technologies often require skills that existing workers do not have. Training older workers in new skills is costly. Hiring younger workers with new skills is difficult in declining industries or regions. So firms operate with an aging workforce on older technology—a "low-productivity trap."

Public investment constraints: Government investment in infrastructure, research, and education declines when budgets are strained by pension and healthcare spending (as they are in aging societies). This public underinvestment reduces economy-wide productivity growth, because firms invest partly based on public infrastructure quality.

Data supports this: Japan's capital intensity (capital per worker) peaked around 2000 and has been roughly flat since, despite massive technological advancement and rising labor scarcity. Germany's capital investment as a share of GDP has been below its pre-2008 average for most of the 2010s. South Korea's capital deepening (capital per worker) has slowed notably since 2000, as its workforce began aging earlier than other countries.

Innovation and entrepreneurship in aging societies​

One of the clearest relationships in demographics is between age structure and entrepreneurship. Startups are disproportionately founded by people aged 25–45.

Research from the Kauffman Foundation and other sources shows:

  • About 35% of all startup founders are under 35 years old
  • About 50% of founders are under 45 years old
  • Only about 5% of founders are over 60

Moreover, startups founded by younger entrepreneurs tend to be in high-growth sectors (tech, biotech, energy), while startups founded by older entrepreneurs tend to be service-sector businesses (consulting, retail). This is not because older entrepreneurs are less capable, but because younger entrepreneurs have more tolerance for risk, more years to benefit from success, and greater familiarity with emerging technologies.

As societies age, the share of the population in the 25–45 age range shrinks. Fewer people are in the age cohort with the highest entrepreneurship rates. This directly reduces startup formation rates. Additionally, venture capital and angel investors—who are themselves often entrepreneurs in their 40s–50s—may become more risk-averse as they age, further reducing funding for new ventures.

Data illustrates this: Japan's startup formation rate has been among the lowest in the OECD (about 3–4% of firms are startups annually, versus 6–8% in the US). Germany's startup rate is moderate but concentrated in a few cities (Berlin) rather than broadly distributed. In both countries, aging is cited as a factor in lower entrepreneurship.

The sectors most vulnerable to reduced innovation from aging are high-tech, biotech, and advanced manufacturing—exactly the sectors that drive long-term productivity growth and wage increases.

Sectoral composition shifts: From manufacturing to healthcare​

As populations age, employment naturally shifts from goods-producing sectors (manufacturing, agriculture) to service sectors, particularly healthcare and elder care. This shift has significant productivity implications.

Manufacturing productivity: Manufacturing sectors typically have high capital intensity and high productivity per worker (e.g., an auto assembly worker supported by $100,000+ of equipment produces $150,000+ of value annually). Manufacturing also has strong productivity growth from automation and process improvement.

Healthcare and elder-care productivity: Healthcare and long-term care are inherently more labor-intensive. A nurse or home-health aide cannot be significantly more productive through automation (unlike a factory worker). Productivity growth in these sectors is slow, maybe 0.5–1% annually. Yet these sectors are expanding as shares of employment in aging societies.

When workers shift from manufacturing to healthcare, economy-wide productivity growth slows even if individual worker productivity in healthcare is unchanged. This is purely a compositional effect.

Example: Japan's sectoral shift

In 1980, about 30% of Japanese employment was in manufacturing; healthcare was about 5%. By 2024, manufacturing was about 17% of employment; healthcare and elderly care were about 13%. As this shift has occurred, aggregate productivity growth has slowed dramatically. Manufacturing productivity grew about 3–4% annually in the 1980s–1990s; healthcare productivity grew 0.5–1% annually (and much of this reflects measurement issues, not real improvement).

The composition effect: If 5% of the workforce shifts from a sector with 4% annual productivity growth to a sector with 0.5% annual productivity growth, aggregate productivity growth slows by about 0.175 percentage points. Over 20 years, this compositional drag reduces cumulative productivity growth by several percentage points.

Government investment and the fiscal squeeze​

Aging societies face budget constraints that reduce public investment in the infrastructure and research that underpin productivity growth.

Education investment: In countries with declining school-age populations, schools close and education spending per student can remain flat or increase (because pension and healthcare spending is growing faster). But education investment in higher education, vocational training, and adult retraining often falls, because public budgets are strained. This reduces the human-capital accumulation needed for productivity growth.

Infrastructure investment: Public infrastructure—roads, bridges, broadband, water systems—depreciates and must be maintained. In aging societies with high pension and healthcare spending, governments defer infrastructure maintenance and new construction. This creates underinvestment traps, where aging infrastructure reduces productivity (slow travel times, inadequate broadband, unreliable utilities).

Research and development: Government funding for basic research (universities, national labs) is often squeezed in aging societies. Private R&D spending also falls, as discussed above. This reduces the frontier of technological innovation, slowing productivity growth economy-wide.

Germany's underinvestment example: Germany spends about 20% of its government budget on pensions and healthcare. As a result, investment in education, infrastructure, and R&D is lower as a share of GDP than in the US or Scandinavian countries. Germany's infrastructure is aging; broadband coverage lags US averages; university facilities have been underfunded for decades. These factors contribute to Germany's lower productivity growth compared to the 1990s–2000s.

The mermaid: How aging affects productivity through multiple channels​

Real-world examples: Productivity slowdowns in aging economies​

Japan's Lost Decades and Productivity Stagnation

Japan's total factor productivity (TFP) growth—the increase in output not explained by more capital and labor—peaked at about 3% annually in the 1980s. By the 1990s, as Japan began aging and growth slowed, TFP growth fell to about 1% annually. By the 2010s, it was near zero. This is despite Japan's position as a global technological leader.

Several factors explain this: Japanese firms responded to aging and demographic decline by becoming cautious about investment (Keidanren, the business lobby, famously warned of impending decline in the 1990s). Younger people emigrated in search of opportunity. Startup formation virtually collapsed. Large, old firms dominated the landscape, investing mainly in process improvements rather than new products.

By the 2010s, when other countries were experiencing productivity booms from digitalization and e-commerce, Japan was lagging. TFP growth remained near zero, and labor productivity growth (output per worker hour) was among the slowest in the OECD.

Germany's Productivity Puzzle

Germany was known as a high-productivity economy in the 1990s–2000s, with strong manufacturing and engineering sectors. However, since about 2005, German productivity growth has slowed. TFP growth was near zero in the 2010s. Labor productivity growth slowed from 2%+ annually to about 0.5% by the 2020s. Productivity statistics by country are tracked by the OECD and Federal Reserve Economic Data (FRED).

Germany's workforce began aging in the early 2000s. Simultaneously, the country faced:

  • Competition from Eastern European manufacturers (lower wages)
  • Slow adoption of e-commerce and digital technologies
  • Low startup formation (especially in tech sectors)
  • Reduced public investment in infrastructure and education

As a result, German firms maintained aging capital stocks and aging workforces. Large, old firms like Volkswagen, Siemens, and SAP drove employment, but their innovation rates slowed. The lack of a vibrant startup ecosystem (unlike the US or Singapore) meant less disruptive innovation and productivity growth.

South Korea's Approaching Slowdown

South Korea remains relatively young by developed-country standards (median age 43 years) but is aging rapidly. Birth rate of 0.72 children per woman is among the lowest globally. The working-age population peaked around 2020 and is now declining.

Korean productivity growth has been robust through the 2010s but is showing signs of slowing. Startup formation, while healthy, is concentrated in a few sectors (e-commerce, gaming) and a few cities (Seoul). TFP growth has been moderate by Korean standards (1.5–2%), below the rates of the 1990s–2000s (2.5–3%).

Korean firms are investing heavily in automation and AI to offset coming labor shortages. However, this is expensive and uncertain. If successful, it could help offset demographic headwinds. If unsuccessful, Korea faces the same slow-growth trap as Japan.

Common mistakes​

Assuming older workers are automatically less productive. This is too simplistic. Older workers are more productive in experience-based roles and less productive in technology-intensive roles. Productivity varies by task, not uniformly by age.

Ignoring that firm investment decisions are forward-looking. Firms do not invest in equipment or training if they expect weak demand or uncertain tenure for workers. Aging populations create uncertainty about future demand, leading firms to underinvest even if aging also creates labor scarcity. The investment effect can outweigh the scarcity effect.

Overlooking the fiscal constraint on public investment. Government spending on pensions and healthcare is not just a transfer (money from workers to retirees). It also crowds out investment in education, infrastructure, and R&D that are critical for productivity growth. The fiscal effect of aging on productivity is substantial and often underestimated.

Confusing productivity with output. Productivity (output per worker hour) can stagnate even if total output continues to grow (if more people work more hours). In aging societies, output per worker often stagnates while total output may decline (if the working-age population shrinks). Both are problematic: stagnant productivity means no real wage growth; declining output means less aggregate consumption and wealth.

Assuming automation will fully offset labor shortage and aging effects. Automation is real and powerful, but it requires capital, technical expertise, and certainty about future demand. Aging societies often face all three constraints. Additionally, automation works best for routine, repetitive tasks, not for sectors like healthcare that are expanding most rapidly in aging societies.

FAQ​

Can productivity growth be fast enough to offset aging's effects on output?​

Theoretically, yes. If productivity per worker grew at 4% annually (higher than historical averages) and the labor force shrank 1% annually, output would still grow at about 3% annually. However, this requires sustained, exceptional productivity growth that is hard to achieve when aging is simultaneously reducing investment and innovation. No aging economy has managed this yet. Japan, Germany, and Italy have all experienced simultaneous labor-force decline and productivity slowdown.

Why doesn't aging create incentives for automation that boost productivity?​

Automation is economically viable only if the cost of the automation is less than the lifetime cost of the worker it replaces. In aging, high-wage economies, this is often true. However, firms face uncertainty about future demand, uncertain about the skill-match of aging workers with new technology, and capital-constrained by competing demands (pension and healthcare spending). The result is underinvestment in automation.

Are there sectors where aging increases productivity?​

Some sectors benefit from aging (healthcare, pharmaceuticals, hearing aids, mobility aids). However, these are sectors that expand to serve the elderly—they do not drive broad-based productivity growth. The main productivity gains in modern economies come from IT, bio-tech, advanced manufacturing, and services (finance, logistics). These tend to slow in aging societies due to reduced innovation and entrepreneurship.

Can immigration help with aging-driven productivity slowdown?​

Yes, if immigrants are younger and more entrepreneurial than the native population. Immigrants are disproportionately likely to start businesses (about 25% of US startups have at least one immigrant founder, though immigrants are 15% of the population). However, immigration requires political support and must be sustained; one-time immigration waves help temporarily but do not solve the long-term demographic problem.

Is the productivity slowdown permanent?​

Not necessarily. Technological breakthroughs (AI, biotech, fusion energy) could drive rapid productivity growth that overcomes aging headwinds. However, such breakthroughs are difficult to predict and do not always translate to broad-based growth (e.g., the internet was a huge innovation but productivity growth in the 2000s–2010s was not particularly fast). Most economists see aging as a structural headwind to productivity that will persist for 20–30 years.

Summary​

Aging populations and aging workforces create structural headwinds to productivity growth. Older workers are less adaptable to technology, firms reduce capital investment when facing uncertain demand from aging populations, entrepreneurship declines as the population ages, and government budget constraints reduce public investment. Employment shifts from high-productivity sectors (manufacturing) to lower-productivity sectors (healthcare), creating compositional drag. The result is that aging economies like Japan, Germany, and Italy have experienced simultaneous labor-force decline and productivity slowdown—exactly the opposite of what optimists hoped for. Labor-productivity trends and sectoral employment data are available from the Bureau of Labor Statistics and World Bank.

Next​

→ The demographic dividend explained