Why Labour Shortages Are Becoming Structural in the Age of AI and Decarbonisation

The OECD study shows that labour shortages in advanced economies are no longer temporary but are driven by long-term forces such as ageing, digitalisation, decarbonisation and AI, which reshape labour demand faster than skills can adjust. It finds that while digital and green transitions intensify shortages, AI and ageing change the nature of skill mismatch, calling for forward-looking education, training and labour policies.

Produced by the OECD Economics Department and the OECD Directorate for Science, Technology and Innovation under the Global Forum on Productivity, this study examines why labour shortages and skill mismatches have become persistent features of modern economies. Using new data covering 32 OECD countries between 2010 and 2019, the authors argue that shortages are no longer just temporary or cyclical problems. Instead, they are increasingly driven by long-term structural changes that are reshaping how economies produce goods and services and what skills they require.

Across most OECD countries, labour markets tightened steadily during the 2010s, well before the COVID-19 pandemic. Job vacancies grew faster than unemployment, meaning firms found it harder to recruit workers. This trend was especially strong in the United States and in knowledge-intensive sectors such as health, finance, ICT and professional services. These developments point to deep, lasting pressures rather than short-term disruptions.

Shortages Rise, Mismatch Changes Shape

While labour shortages increased almost everywhere, qualification mismatch behaved differently. The overall share of workers whose education level does not match their sector’s typical requirement stayed relatively stable. However, this stability hides an important shift. Fewer workers were underqualified, while more workers were overqualified. In simple terms, firms increasingly hired workers with more education than the job formally required, often because appropriately skilled candidates were scarce.

This pattern was strongest in knowledge-intensive sectors, where advanced skills take time to develop and cannot be quickly supplied. By 2019, more than half of all sectors across OECD countries would have been unable to fill vacancies without moving workers across sectors. In countries with very tight labour markets, such as the United States, unfilled jobs were concentrated mainly in high-skill sectors, while in countries with higher unemployment, shortages were less acute and focused on low-skill activities.

Five Structural Trends Behind the Pressure

The paper studies five major structural trends shaping labour markets. Workforce ageing reduces labour supply over time and increases skill obsolescence, especially in fast-changing sectors. Decarbonisation raises demand for green skills as economies cut emissions and transform energy-intensive industries. Digitalisation boosts productivity but also increases demand for digital skills. Deglobalisation, measured through lower import penetration, affects job creation and destruction across sectors. Artificial intelligence introduces automation and new tasks that can either replace or complement human labour.

By combining sector-level exposure to these trends with country-level progress, the study shows that digitalisation and decarbonisation are the strongest drivers of labour shortages. As countries adopt more digital and green technologies, demand for specialised workers rises faster than supply can adjust. Closing half the gap with leading countries in digitalisation or decarbonisation would significantly increase labour market tightness, highlighting how transformative these transitions are.

Why AI and Ageing Work Differently

Artificial intelligence stands out because its average effect differs from digital and green transitions. During the 2010s, AI diffusion tended to reduce labour shortages and overqualification, mainly because automation displaced certain jobs, particularly in white-collar occupations. However, the study shows that this effect depends on how AI is used. When AI substitutes for workers, it creates slack in the labour market. When it complements workers, it raises productivity and increases demand for skilled labour, worsening shortages instead.

Workforce ageing has slower but lasting effects. In the short run, ageing does not strongly affect labour shortages. Over time, however, especially in countries where ageing is already advanced, it tightens labour markets by shrinking the available workforce. Firms increasingly hire underqualified workers, leading to downskilling rather than overqualification. Ageing, therefore, contributes to shortages and changes the nature of mismatch.

What This Means for Policy

The study’s central message is that labour shortages and mismatches are structural outcomes of long-term economic change. Policies that treat them as temporary problems risk falling behind reality. Governments need to anticipate future skill needs and coordinate policies across education, labour, innovation, climate and industrial strategies.

The paper highlights “double-dividend” policies that can ease labour shortages while supporting productivity and transitions. Education and training systems can expand green, digital and AI-complementary skills. Lifelong learning can limit skill obsolescence in ageing societies. Policies that boost labour participation, support mobility across sectors and regions, manage migration effectively, and encourage innovation can all help reduce bottlenecks. Without such coordinated action, the paper warns, labour shortages could slow digital progress, delay climate goals and weaken long-term growth.