Demographic Change and Korea’s Growth Prospects: The Role of AI and Human Capital

South Korea's demographic challenge and its growth implications

South Korea's economic growth rate has been declining for some time, coinciding with one of the fastest demographic transitions in the world. The country is entering super-aged society status faster than any other nation, with a sharply rising elderly population share, a dramatically falling fertility rate now hovering at around 0.7–0.8, and a rapidly declining working-age population. The fertility rate has fallen dramatically since the 1970s and has continued to decline in recent years, driven largely by a steep rise in the share of never-married individuals. This suggests that the trend is deeply structural in nature and will be difficult to reverse quickly through policy intervention alone. These demographic shifts are already increasing fiscal pressures through higher pension, healthcare, and social spending. These pressures are set to intensify significantly in the decades to come as the elderly share of the population continues to rise steeply relative to other major economies. While Japan experienced population aging earlier, South Korea's pace of aging is even faster and will surpass that of Japan in the near future, making the challenge particularly acute.

Beyond demographics, South Korea is also faced with an increasingly uncertain external environment. Rising protectionism and geoeconomic fragmentation, as well as geopolitical conflict, are reshaping global trade and investment flows, and major economies are increasingly turning toward deploying trade, investment, and industrial policy as strategic tools. This puts additional and growing strain on South Korea's long-established export-oriented growth model, which has historically relied on open international trade. Taken together, these internal and external pressures weigh heavily on South Korea's overall growth outlook. According to OECD long-term projections, South Korea's potential growth rate could fall below 1% by the 2040s and turn negative by the late 2050s. This trajectory is not inevitable, but it is a serious warning that demands proactive and well-calibrated policy responses.

Demography is not destiny: offsetting factors

Despite these pressures, demographic change does not automatically determine the growth path, and several factors can meaningfully counteract the drag from a shrinking labor force. Female and elderly labor force participation rates have risen significantly in South Korea in recent years, and the M-shaped female participation curve, which reflected career breaks around childbearing years, has largely disappeared. The employment rate of elderly workers in South Korea is now one of the highest among the OECD countries, an important development that demonstrates the potential for labor supply adjustments to partially offset demographic headwinds. Older workers today are also considerably better educated and more experienced than previous generations, which makes existing sweeping assumptions about lower elderly productivity increasingly questionable. Indeed, with the right combination of skills, digital tools, and institutional support, older workers may in some cases achieve higher productivity than the levels that are conventionally assumed. Moreover, as the workforce ages, there is evidence that demographic pressure itself can accelerate automation, inducing technological change that is more complementary to older workers, which is a dynamic that was first highlighted by Acemoglu and Restrepo.

Applying standard growth accounting to South Korea's situation illustrates the scale of the challenge but also the scope for policy to make a difference. South Korea's working-age population is projected to decline by around 1.4 percentage points per year on average according to UN estimates, which would reduce GDP growth by approximately 0.8 percentage points if all else remained equal. However, looking at actual growth accounting data for the period 2010 to 2019, labor input contributed only around 10% of total GDP growth, while human capital quality, physical capital accumulation, and total factor productivity played far larger roles. This suggests that higher labor force participation, capital deepening, and productivity improvements could collectively offset roughly one percentage point of the growth drag from demographic decline. The key message is that the right policies and structural adjustments matter at least as much as demographic fundamentals, and that South Korea retains meaningful room to shape its own economic trajectory.

Human capital: strengths and persistent gaps

South Korea is internationally recognized for the speed and scale of its human capital accumulation over the past several decades. School enrollment rates rose rapidly during the country's industrialization period, and by most international measures the pace of increase in average years of schooling in South Korea was the fastest of any country in the world. South Korean students consistently rank among the world's top performers in international assessments of mathematics, science, and other fields, alongside other East Asian economies including Japan. In economics, human capital refers to the skills, knowledge, intelligence, and abilities acquired through education, training, and experience, and the evidence shows that it raises productivity at both the individual and economy-wide level by improving returns to capital investment, generating knowledge spillovers among skilled workers, enhancing R&D capacity, and facilitating the absorption of new technologies. The broad expansion of high-quality basic and secondary education was therefore a central driver of South Korea's economic rise, underpinning the manufacturing-led growth model that transformed the country into an advanced economy within a generation.

However, significant weaknesses remain that must be addressed if human capital is to continue supporting growth in the decades ahead. University education quality is relatively low by international standards, with only a small share of graduates coming from globally top-ranked institutions. This ratio compares unfavorably even with Japan when considered in proportion to population. A substantial mismatch between educational outcomes and actual labor market needs also persists, with many graduates entering fields where their qualifications are not optimally utilized, representing a significant loss of human capital investment. Furthermore, adult skill levels, which can be measured by literacy, numeracy, and adaptive problem-solving among working-age adults between 20 and 64, fall below the OECD average across most age cohorts and decline sharply with age. This is in part a consequence of lifetime employment systems that have historically reduced incentives for firms and individuals to invest in ongoing retraining and upskilling once initial employment is secured. In an AI-driven economy where skills become obsolete at an accelerating pace, this weakness is particularly consequential. Strengthening tertiary education quality, building robust and accessible lifelong learning mechanisms, and systematically reducing skill mismatches between education and the labor market will therefore be essential priorities for sustaining and deepening South Korea's human capital advantage going forward.

AI as a growth driver: potential and uncertainty

AI is increasingly regarded as a general-purpose technology comparable in its transformative potential to earlier breakthroughs such as steam power, electricity, and information technology. Its key distinguishing feature is that it augments cognitive tasks, such as performing prediction, analysis, optimization, and content creation, and continues to improve autonomously over time. In practical terms, AI raises productivity by automating routine tasks and freeing workers for higher-value and more creative activities, by enhancing the efficiency of R&D and innovation processes, and by driving significant investment in intangible and knowledge capital. A notable finding from micro-level research is that productivity gains from AI adoption benefit not only highly skilled workers but also less-skilled employees, pointing to the potential for broad-based gains across the workforce if adoption is well managed. At the same time, AI also poses risks, including job displacement for workers performing repetitive tasks, rising income inequality, market concentration in digital platforms, and governance and national security concerns. These risks make the labor market effects of AI particularly complex: while workers in non-routine, AI-complementary roles stand to benefit significantly, those whose tasks are largely substitutable by AI face the prospect of lower wages or unemployment. The aggregate employment outcome therefore depends heavily on skill distribution, the speed of AI adoption, and crucially the flexibility of labor market institutions in reallocating displaced workers to more productive sectors.

Estimates of the aggregate macroeconomic growth effect vary widely, reflecting genuine uncertainty about the pace of economy-wide diffusion, organizational adjustment costs, and what is sometimes described as a "productivity J-curve"—an initial period of declining or stagnant output as firms and institutions adapt before gains materialize. Conservative model-based estimates, such as those by 2024 Nobel Prize winner Acemoglu, suggest AI may raise annual growth by only around 0.1 percentage points over the next decade. More optimistic projections, including those by fellow Nobel laureate Aghion and Bunel, point to gains of up to 1.3 percentage points annually, drawing on comparisons with the historical impact of earlier general-purpose technologies such as the Internet and electricity. The actual outcome will depend heavily on institutional design, policy choices, and the degree to which AI diffuses broadly across sectors and firm sizes rather than remaining concentrated among large, technologically advanced enterprises.

For South Korea specifically, around half of all jobs are highly exposed to AI, with roughly half of those showing strong task-level complementarity and the other half facing meaningful displacement risk. An important finding for South Korea is that productivity gains from AI may be particularly large for older, high-skilled workers, while older, low-skilled workers may face greater risks, a distinction that is particularly relevant given South Korea's aging workforce. A joint study by the IMF and the Bank of Korea estimates an overall GDP gain of around 13% in the medium to long run, equivalent to roughly 1.3 percentage points of annual growth if realized within a decade. South Korea already has notable structural strengths to build on: it accounts for approximately 23% of global semiconductor exports, and an IBM survey from 2023 found that around half of large South Korean firms had already actively adopted AI. The IMF's AI Preparedness Index ranks South Korea 15th out of 165 countries, with particular strengths in digital infrastructure, innovation capacity, and regulatory frameworks. The government's Manufacturing AI Transformation (M.AX) initiative is actively promoting AI adoption in key industries including robotics, autonomous vehicles, semiconductors, shipbuilding, and drones. The principal remaining weakness lies in labor market rigidity, which constrains the reallocation of workers toward productive, AI-complementary roles and risks concentrating productivity gains in a narrow segment of the economy, particularly among large firms while small and medium-sized enterprises remain unable to adopt the technology at an adequate level. These types of firms account for the majority of employment and struggle with high adoption costs and skills shortages.

A coherent policy strategy for sustained growth

The central challenge for South Korea is to activate a virtuous cycle in which human capital, AI, and supportive institutions reinforce one another. AI increases productivity and accelerates innovation, but human capital is what allows workers and firms to effectively adopt and leverage AI. Economic growth in turn feeds back into further human capital accumulation and technological investment. Institutions and policies—spanning education, labor markets, R&D frameworks, data governance, and infrastructure—coordinate and amplify these interactions. No single reform will be sufficient on its own; the key is building synergies across a coherent and integrated policy strategy.

Four key directions for achieving this can be identified. First is extending and upgrading older workers' participation by not merely keeping them in the labor market longer, but by equipping them with the digital and cognitive skills needed to remain productive. The second direction involves reorienting education at all levels toward the skills that complement rather than compete with AI, such as critical thinking, creativity, communication, collaboration, and digital literacy. Third is deepening labor market reforms to enable genuine worker mobility and sectoral reallocation, since productivity gains from AI will remain limited and unevenly distributed if workers cannot move to where their skills are most needed. The fourth direction means integrating AI policy, human capital policy, and industrial policy into a single coherent strategy with clear goals such as achieving AI G3 status rather than pursuing isolated initiatives. South Korea is already moving in this direction, as evidenced by the Special Act on Semiconductors passed in 2026 and new economic security legislation aimed at supply chain resilience. With the right combination of these mutually reinforcing measures, a medium-term growth rate of around 2% may remain achievable for South Korea, and the lessons drawn are broadly applicable to Japan and other economies navigating similar challenges of rapid demographic aging.

Comment

IWATA Kazumasa:
South Korea and Japan face the common challenge of transitioning to a super-aged society amid intensifying technological rivalry between the U.S. and China. Assessing and analyzing the similarities and differences among both nations offers valuable opportunities for policymakers. South Korea's goal of becoming a so-called AI G3 country, one of the top three global AI powers, by accelerating AI transformation in manufacturing, with particular emphasis on the comparative advantage of physical AI, is especially noteworthy in this context.

Regarding immigration policy, the number of foreign workers in Japan currently amounts to 2.57 million, while South Korea hosts around 1.1 million foreign workers. Long-term forecasts suggest that Japan's foreign worker share could rise to approximately 14%, comparable to the average of European countries. I would like to know if there are any particular lessons related to South Korea's immigration experience that could prove useful for Japan, including how to address undocumented workers, so please share those if possible. Illegal foreign workers are a larger issue in South Korea at around 0.4 million, compared to approximately 0.08 million in Japan.

Regarding human capital, a clear positive relationship exists between human capital accumulation and total factor productivity growth. South Korea's human capital level ranks among the highest internationally and is projected to continue growing through 2040, contributing an estimated 0.5 percentage points to South Korean economic growth. Meanwhile, in Japan, the lower return on female higher education investment presents a persistent structural obstacle. Female college graduates frequently exit the workforce after childbirth and re-enter only as part-time workers, which results in a significant reduction of lifetime earnings relative to male graduates and complicating the policy for education investment. Could you please provide any insights on this topic from your research in Korea?

With respect to AI's impact on TFP, South Korea ranks fourth globally in AI-induced productivity gains, performing considerably better than Japan, which is ranked in 30th place. Japan's AI-related investment plan amounts to around 3 trillion yen across government and private sectors and appears modest when compared against the scale of commitments elsewhere. This situation raises legitimate concerns about Japan falling behind. Using South Korea as a reference and setting an explicit national AI ambition, such as targeting AI G5 status, could provide useful strategic direction. South Korea's AI G3 ambition is significant, even though Singapore and Taiwan stand as strong competitors. In terms of AI research institutions, China currently occupies all top five positions globally, with South Korea's KAIST ranked seventh, while Japan's University of Tokyo lags considerably further behind.

On labor market reform, the mandatory retirement system both in South Korea and Japan meaningfully constrains labor supply and work incentives for elderly workers, and its complete abolition, rather than merely extending the statutory retirement age, warrants serious consideration. Additionally, tax and social security systems that continue to favor traditional household arrangements also negatively influence female labor supply decisions in ways that are difficult to unwind. Early retirement trends in South Korea, with individuals around age 50 transitioning into startup activity after early retirement, could represent a promising avenue for productive use of elderly workers' accumulated human capital.

Jong-Wha LEE:
On immigration, South Korea's foreign residents stand at around 2.5 million (approximately 5% of the total population), including roughly 1.1 million foreign workers. Most workers are entering through the employment permit system and directed toward small and medium-sized firms in industries such as construction, agriculture, and services. Pathways to permanent residency are largely absent. Immigration is not viewed as a primary solution to demographic decline, as the pace of working-age population contraction is too rapid to be meaningfully offset by immigrant inflows alone. Improving productivity and restructuring industry remain the more viable long-term responses.

As regards human capital and especially for female labor force participation, career interruptions and job status penalties are recognized as central concerns, compounded by field-specific variation in returns to education. The priority, however, lies less in measuring individual returns to education and more in removing structural barriers to female labor market participation, for example through expanded parental leave and flexible working arrangements. South Korea has pursued these policies actively and achieved meaningful progress, but the broader labor market reform agenda remains more critical in addressing the problem.

Concerning AI and global competition, China's strength in large-scale engineering and physical AI is significant, while the U.S. retains leadership in frontier research and advanced chips. South Korea's AI G3 ambition isambitious, with Singapore potentially better positioned as a global finance and technology hub. Nevertheless, South Korea and Japan may hold structural advantages in economy-wide AI adoption by combining advanced digital infrastructure with strong manufacturing sectors. The aggregate GDP impact of AI for South Korea is estimated at around 13% in the medium to long run, equivalent to approximately 1.3 percentage points of annual growth if the transition is realized within a decade.

Finally, for the labor market in South Korea, rigidity and dualism remain central structural constraints, limiting worker mobility and raising transition costs as AI reshapes the employment landscape. Initial job placement carries an overly significant influence on career trajectories. As a result, workers who enter small- and medium-sized enterprises face significant difficulties moving to more productive firms later, even after accumulating valuable skills. Meaningful structural change will require time, and resistance from strong labor unions adds further complexity to the reform agenda. South Korea continues to look to Japan's experience in managing the participation of older workers as a source of policy learning in this area.