While the absolute size of the EU workforce is expected to be smaller in the coming decades, it is also likely to be more highly educated. As Figure 1 illustrates, the total size of the workforce is projected (under the CEPAM Medium scenario, SSP2¹) to change from 245.8 million to 214.1 million over the next four decades. The decline is solely attributable to fewer workers with educational levels of ‘upper secondary’ and ‘lower secondary or below’. Projected decreases in the size of the labour force come from the population with education levels of low (from 50.7 to 14.0 million) and upper secondary (from 108.2 to 74.2 million).

At the same time, the numbers of workers with a master’s degree or higher, a bachelor’s degree, or a shorter form of post-secondary education (e.g., technical training) are all projected to increase not only in their proportions of the labour force, but also in absolute terms. Significantly, these post-secondary groups are expected to grow by almost half (+45%) over the next 40 years.

National differences exist in the levels and rates of educational increases, but the general trend holds true across all EU countries. Irrespective of various changes to the age structure and labour force size in individual member states, it is anticipated that the human capital of future workers (measured by the highest level of education attainment) will be higher at every age because younger cohorts are spending more years in education than previous generations.

Looking at regional trends within the EU, we see that several eastern member states – the Czech Republic, Slovakia, Poland, Latvia, and Croatia – currently register the highest levels of post-secondary educational attainment. While CEPAM projects that all EU member states will continue moving towards higher educational attainment, Southern member states that currently have the lowest levels – i.e., Portugal, Spain, Italy, and Greece – will make the fastest gains.

In the scenario with constant participation rates, the labour force size remains stable for many western member states, due in part to the movement of working-age citizens from eastern member states. As a result of these same migration trends, the labour force size decreases in the 13 new member states (EU13). However, the labour force dependency ratios in these states would remain at intermediary levels because, on average, their populations have lower life expectancy or fewer unproductive years at older ages than their western counterparts.

Whether either population age-structures shifting towards higher ages or reductions in the working-age populations do in fact produce a smaller labour supply, depends on how age-specific labour force participation rates – particularly of women and people aged between 50 and 70 – develop in the coming decades. The extent of any economic implications also depends on labour demand, which may fundamentally shift in the coming years amid automation and other factors.

When considering future labour force size and composition, it is important not to overlook the labour force dependency ratio. This measure, together with changing productivity, in large part determines the potential economic impact of population ageing. By comparing the economically inactive population to the economically active population, the ratio captures a population’s composition both in terms of demographic structure and labour force participation.

A recurring theme in Europe’s public discussions of migration describes an unmet, ongoing demographic need that can be filled by migrants. Apprehensions revolve around the size of the future workforce and how the system might function with fewer workers. However, largely absent from such discussions are (1) considerations of existing differentials in workforce participation rates between the EU member states, specifically, important variations in female participation rates; and (2) analyses of how migration could influence the labour force dependency ratio, rather than simply the labour force size.

As can be seen in Figure 2, increasing female participation rates leads to the lowest ratio of non-workers to workers among the scenarios considered. This scenario uses observed Swedish workforce participation rates as a focal point, projecting a slow convergence of other member states towards Sweden’s age- and education-specific participation rates over the course of the next 45 years. CEPAM research shows how such a gradual development would wholly counterbalance the expected rise in the dependency ratio due to population ageing.

The next-lowest dependency ratio is produced by the Equalisation scenario, in which the female labour force participation rates reach current country-specific male levels by 2060. In this situation, the expected climb in the dependency ratio would be significantly slowed, stabilising at around 1.2 dependents per worker. This amounts to about only half the increase compared to the constant scenario, arriving at nearly 1.4 dependents.

In contrast, even in the double scenario, migration leads to only a minimal decrease in the dependency ratio. CEPAM projection results show that doubling migration, despite increasing the overall population (and labour force) size, would only slightly improve the labour force dependency ratio, without altering the broader momentum towards rising levels.

Therefore, increased migration flows – particularly of migrants with low education and low labour force participation prospects – are unlikely to alleviate the economic consequences of population ageing. Although political discourse often refers to migration as a tool for easing the consequences of population ageing, demographers agree that while immigration can affect population size, it generally has little impact on the ratio of non-working to working people. Furthermore, the human capital levels of natives and migrants matter greatly for any potential improvements in productivity, and how societies will or will not cope with the consequences of automation of work processes.

Fertility in the EU-28 varies between 1.35 children per women in Portugal and two children per woman in France and Ireland (standard deviation around 0.2), with the average total fertility rate standing at 1.61. Western members have had higher fertility levels than those in southern and eastern member states, although fertility rates appear to be slowly converging across Europe as fertility falls in the west and rebounds in the east.

According to the CEPAM Medium scenario, average period fertility in the EU will rise from the current 1.61 children per woman to 1.67 in 2030, and then to 1.71 children by 2060. In addition to higher EU-wide fertility, CEPAM Medium predicts the continuation of converging fertility behaviours across member states into the future.

Over time, EU internal mobility has the potential to strikingly change the demographic makeup of many member states depending on whether intensities of south- and east-to-west movements lessen, remain the same, or grow further. In some cases, net internal flows significantly impact the overall size of a sending country’s population.

Particularly in eastern EU member states (Bulgaria, Croatia, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, Slovenia, and Romania) smaller populations seem inevitable, primarily due to high net movement to western economies. Without out-migration, the populations of the eastern member states would naturally only shrink 11–25% by 2060, similar to the natural rates of decrease in many other European countries.

If EU internal mobility continues without stabilising at lower levels, CEPAM results indicate that the most extensive population decreases reach −38.0% for Bulgaria and −34.3% for Romania by 2060. In a double migration future, the populations of many eastern member states would fall even more dramatically, including Bulgaria (−48.6%) and Romania (−46.8%).

As for the proportion of people aged 65+ in the population, higher volumes of internal mobility generally reduce age-dependency rates in western Europe and increases them in eastern Europe. If current trends continue, a particularly dynamic process will be observed in the eastern member states where lower fertility is accompanied by high volumes of movement to other EU member states, accelerating ageing.

Aside from internal movement, western member states will continue to see expanding populations if current flows of international migration either remain at constant levels or double. Such inflows would be significant enough to increase the number of people living in the EU as a whole.

Only seven countries are unlikely to naturally decreases in population size without migration (in all presented scenarios, including those of Eurostat (2015) and national projections), mostly because they have relatively high fertility: Ireland, Denmark, France, Sweden, and the UK (TFRs around 1.8–2.0 in 2015), or an age structure in which a large share of the population is of reproductive ages: Luxembourg (median age of 39.3 years) and Cyprus (35.0 years).

All CEPAM scenarios show a decrease in the populations aged 0–14 and 15–64, and an increase in the population aged 65+. No migration leads to a reduction in the population aged 0–69, but also to a more moderate increase in the population aged 70+, shown in Figure 3. Whereas higher migration (CEPAM Double Migration) would reduce the magnitude of change for ages 0–69, it would also enlarge the populations aged 70+ by 67.8 million people in the long-run, as the migrant populations inevitably grow older.

Demographic changes are to a large extent pre-determined in the near term by population age structures that result from past fertility, mortality, and migration trends. But from the longer perspective, all components of demographic change can be influenced by policy and other initiatives.

CEPAM results demonstrate that aside from policies intending to influence fertility and mortality directly, efforts to enhance education can have important indirect consequences that lead to increases in longevity and reductions in high fertility as well. The effects of socio-economic development become clear when looking at Figure 4. The two extremes in possible global development pathways lead to wildly different worlds, over 5.5 billion people apart. By the end of the century, these divergent trajectories reach 7.23 billion (Rapid development/SSP1) and 13.6 billion (Stalled development/SSP3).

Population ageing is not confined to East Asia, Europe, and North America. Under all scenarios the world population will get significantly older as a consequence of lowering fertility rates and rising life expectancy. In the CEPAM Medium (SSP2) scenario, the proportion of the population aged 65+ increases from the current 8.3% to 17% by mid-century, and to 29% by 2100. Moreover, the faster gains in life expectancy assumed by the Rapid Social Development (SSP1) scenario result in the proportion of 65+ reaching about 20% by mid-century, and 45% in 2100.

CEPAM projections for world population have been revised upwards from the estimates in World population and human capital in the 21st century (Lutz et al. 2014), primarily due to swifter declines in African mortality and slightly higher fertility rates than were previously assumed by other international assessments. In demographic terms, higher rates of child survival have the same effect as increasing fertility. Already by 2015, the population was markedly higher than had been foreseen using the 2010 baseline data.

Spread of mass education has been very slow (with moderate progress in recent years) across Sub-Saharan Africa. Although educational attainment is expected to increase steadily over time in all of Africa’s sub-regions, it will still remain lower than other continents at the aggregate level according to the CEPAM Medium (SSP2) scenario. While the average amount of schooling for western, middle, and eastern Africa stood between 5.4 and 5.8 years in 2015, these sub-regions are projected to reach between 9.3 and 9.9 years by 2060.

If development slows or stalls (SSP3) and female education does not expand, world population already crosses the 10 billion mark around 2045 and then continues to grow over the rest of the century to 13.6 billion in 2100. For Africa, such a path would likely be associated with weak resilience to environmental changes, high mortality, and widespread poverty. Interruptions in the spread of education would in fact accelerate population growth due to differing fertility behaviours between education groups.

In this stalled development scenario, the combination of high population growth with no further expansion of schooling results in African populations becoming less educated. Schooling expansion must keep pace with the pressures from rapid population growth, otherwise CEPAM results show hardly any improvement in the proportion of people without any formal education even at the global level – from 10% in 2015 to 9% by the end of the century.

South Asia is the least educated region of Asia with the adult population averaging 6.6 years of schooling in 2015, or slightly above the average for Sub-Saharan Africa. Younger cohorts of Indian women have notably risen their levels of educational attainment. However, for the adult population of India as a whole, the low educational attainment of previous cohorts still overshadows the gains made in recent decades.

All of the projections foresee India eclipsing China to become the world’s most populous country. And although most South Asian countries send large numbers of emigrants to North America, the EU, and Gulf countries, these flows have a negligible effect on population growth of the region. Still, by 2060 the median age in South Asia is expected to rise to a point similar to that of East Asia in 2015, stabilising by mid-century.

The Japanese population has achieved global leadership in both educational attainment (ages 25+ averaged 13 years of schooling in 2015) and longevity (87.2 years for women in 2015–20, expected to climb to an impressive 95.7 by 2055–60). Japan is also further along in the ageing process, with the world’s highest share of population 65 and older. The current proportion of people aged 65+ in Japan is 7% higher than in the EU, and is expected to continue rising in the future as migration remains low and life expectancy increases.

China is rapidly catching up with the other highly educated East Asian countries after coming from far behind, and is predicted to reach 11.4 mean years of schooling in 2060. Over the same span of time, CEPAM results show China is expected to continue ageing to become 14% smaller compared to itself in 2015.

Differences between CEPAM’s projections make clear that changes in educational composition play a significant role in steering social development and population trends. The SSP3 pathway illustrates a future in which minimal investments in education and continued high fertility in Africa and parts of South Asia could reinforce each other, locking the world into a cycle of very high population growth, poverty, and high vulnerability.

In stark contrast, the SSP1 scenario shows how feedbacks between population, education, and development can alternatively result in further improvements in education and lower fertility. Many countries under this rapid development scenario would be well positioned to take advantage of the opportunities for broad economic and social advances that come from population structures shifting towards more workers and fewer dependents.

However, without sufficient education, the benefits of a maturing age structure are not guaranteed. To repeat the development successes seen in parts of East Asia and Latin America requires both prior investments in education and health infrastructure, along with positioning local enterprises to employ and retain the newfound potential of a more educated local labour force.