- In 2022, the EU had 36 people aged 65 or older for every 100 people aged 20 to 64; by 2050, this ratio will reach 55.2.
- Many European pay-as-you-go pension systems were built or expanded in the second half of the twentieth century in a different demographic context.
- Each country has several levers available: the effective retirement age, employment, contribution rates or public revenues, coverage, and the relative level of pensions.
- For two decades, European governments have undertaken reforms to the retirement age, benefits, and contribution rates, while deferring or partially cancelling certain measures in several countries.
The essentials
- The dependency ratio in Europe is rising from 36% in 2022 to 55% in 2050, according to Eurostat’s demographic projections: by 2050, fewer than two people aged 20 to 64 are projected per person aged 65 or older.
- Denmark has set the public pension age at 70 from 2040 onward; Germany is raising it to 67 by 2031; in Italy, 71 is an alternative age applicable to certain contributors in 2026, while the standard old-age pension age is 67.
- Parametric reforms are among the available levers: their success depends on building a political coalition capable of securing acceptance and their place within a long-term strategy.
- Several trajectories are possible: a path of negotiated reforms combining multiple levers, a path of imposed reforms without consensus, and a path where productivity contributes to adjustment, with very different profiles of sustainability and social cohesion.
- Adapting mechanisms inspired by the Swedish model may be more difficult in some Central and Eastern European countries where institutions for social dialogue and institutional trust are weaker.
From 36% to 55%: what the figure shifts
The dependency ratio measures the number of people aged 65 or older per hundred people aged 20 to 64. In 2022, the European Union stood at 36. The reference projection EUROPOP2023 used by the Commission is 55.2% in 2050. The projection for 2100 reaches 65.
These figures seem abstract until they are translated into financial flows. Pay-as-you-go pension systems, the dominant model in Europe, directly transfer contributions from the working population to retirees’ pensions. In any given pay-as-you-go system, changes in the ratio may require a combination of additional revenues, lower benefits, later departures from the workforce, or other sources of financing such as immigration, productivity gains, or general taxation.
Ageing is not uniform across Europe. Southern and Eastern Europe are ageing faster than the North, according to the Bruegel report on the demographic divide. Poland, Romania, and Italy show particularly degraded projections, combining low birth rates with emigration of their youngest working-age populations to Northern and Western Europe. This double movement—fewer children here, fewer workers tomorrow, and even fewer the day after—creates budget pressure that Eurostat’s projections make visible but that electoral cycles tend to render invisible.
The European Central Bank, in its report dedicated to demographic ageing, documents an additional mechanical effect: healthcare and long-term care spending rise in parallel with pensions. The total cost of ageing represents far more than the pension line alone in national budgets.
Germany, Denmark, and Italy: three laboratories for the same problem
Three countries illustrate the range of available responses and their limits.
Germany chose a progressive approach. The shift to 67 years is being phased in until 2031. The reform was negotiated with social partners in a context of full employment and strong growth, conditions that no longer exist. Today, with an economy stagnating and a population ageing faster than expected, Berlin is hesitant about what comes next: extending the legal retirement age further meets growing political resistance, particularly in the eastern Länder where life expectancy in good health remains below the national average.
Denmark adopted a different logic, more radical and more elegant. The retirement age is automatically indexed to life expectancy. When people live longer, they work longer; the rule is written into law and does not need to be renegotiated with each parliament. Denmark has set the public pension age at 70 from 2040 onward. This architecture reduces political conflict on the issue, since adjustment becomes technical rather than legislative.
It does, however, require an already-high employment rate for older workers and a labour market capable of absorbing workers aged 65 to 69—two conditions that Denmark meets better than most of its neighbours.
Italy sets an alternative age of 71 in 2026 for certain contributors with five years of contributions, while the standard age remains 67. The Italian labour market offers few jobs suited to older workers. Life expectancy in good health is shorter than in Scandinavia. Institutional trust and the reception of pension reforms must be measured separately; institutional trust constitutes a potential brake on acceptance of reforms whose benefits and costs are spread across several decades.
The Italian case shows that arithmetic alone is not enough: a technically necessary reform can remain politically blocked for years.
The limits of parametric reforms alone
Bruno Palier, a specialist in the political economy of social protection, has documented with precision what European governments took time to acknowledge: raising the legal retirement age or contribution rates are instruments for adjusting to structural demographic and financial imbalances, even if they do not by themselves change those demographics. The flow is the annual imbalance between revenues and expenditure. The structure rests on several factors: the institutions for negotiation, the legacy of previous public policies, political coalitions, and the social groups affected.
When parametric reforms are adopted without this coalition, they risk producing cycles of protest, amendment, and delay. With each new cycle, the political cost tends to rise due to progressive erosion of trust. Adjustment eventually arrives, but demography only creates pressure; the modalities of reform depend on system design and national political choices. The absence of consensus makes these adjustments more painful and less durable.
A competing reading, advanced by economists close to Philippe Aghion and the Schumpeterian tradition, emphasizes a different lever: productivity. If productivity gains are high enough, each worker produces more wealth, and the dependency ratio can be managed more easily without extending working lives. This thesis is intellectually sound. Its weak point is the time horizon: productivity gains linked to automation and artificial intelligence are real, recent work on the adoption of AI in business shows that they do not eliminate employment but reshape it, but their rate of diffusion is uncertain and their calendar does not coincide with the demographic pressure of the 2030s.
The two readings are not mutually exclusive. They are hierarchical: parametric reform is one option among others, which can also include tax revenues, changes to coverage or employment, and institutional changes. Productivity is an important variable for absorbing the effects of ageing, without solely determining the scale of pension adjustments.
The brake is institutional, not arithmetic
Demographic projections signal pressures that may require adaptations depending on the design and financial situation of each system. Some countries manage this with less conflict than others, and the reasons for this gap deserve examination.
The answer lies in three institutional variables that Bruegel documented in its analysis of Europe’s demographic divide.
The first is the quality of social dialogue. Countries with established and recognized structures for negotiation between social partners tend to produce reforms that are more durable and better accepted. Sweden is the most cited example: the 1994 reform, which transformed the Swedish points system into a notional accounts model, was negotiated over seven years between government and unions. It is today one of the most robust in Europe. Its transfer to Southern Europe encounters difficulties linked to the absence or weakness of comparable negotiation structures.
The second variable is institutional trust. A high level of trust in institutions facilitates acceptance of reforms whose benefits and costs are spread over several decades. In Romania and Poland, the institutional history of breaks in social contracts creates obstacles to accepting reforms based on long-term promises.
The third variable is the employment rate of older workers. Raising the legal age without creating employment conditions for 55- to 64-year-olds amounts to shifting the problem toward unemployment insurance and disability schemes. Italy remains among countries with low participation of older workers, but the 60% threshold has been exceeded in recent data, against more than 75% in Denmark and Sweden. Without active employment policies for older workers—job redesign, continuing education, targeted tax incentives—extending working life is, for part of the population, an unkeepable promise.
Three trajectories for the next thirty years
By 2050, three trajectories are emerging for EU member states. They are not mutually exclusive, but they produce very different results for social cohesion and fiscal sustainability.
Some Nordic countries combine multiple levers: automatic indexation of retirement age to life expectancy, targeted employment policies for older workers, high female employment rates, and policies welcoming migrants of working age. This combination helps absorb pressures linked to changes in the dependency ratio. It requires, however, coherence in public policy across multiple mandates, which is precisely the main political difficulty.
The forced parametric path is one where governments raise the legal age without building prior consensus, constrained by external forces. Reforms pass, but political fragmentation increases. The Greek case of 2010-2015 remains the reference for what this path can produce: adjustments that can significantly erode trust in institutions.
The productivity-AI path is the most uncertain. If efficiency gains from automation diffuse rapidly and widely enough, pension financing can be maintained without massive extension of working lives. Existing signals—adoption of AI in services, partial robotization of industry—are real but insufficient to establish a credible quantified scenario by 2040. This trajectory deserves to be monitored, not planned around.
Two signals will allow us to characterize the trajectory each country is pursuing: the evolution of the employment rate of 55- to 64-year-olds, which Eurostat publishes quarterly, and the evolution of political positions on pension reforms in Italy, Poland, and Romania in 2026-2027. These votes will tell whether the second path is starting, or whether these countries find a more durable compromise.
Pension rights portability, the least-used lever
One mechanism remains largely underexploited in the European debate: the portability of pension rights for intra-EU migration. Currently, a Romanian worker who has contributed ten years in France and fifteen years in Germany accumulates rights in two different national systems, with distinct settlement rules, significant administrative delays, and partial losses in transition. This complexity hinders worker mobility within the Union, precisely when ageing countries need to attract workers and countries that export them would benefit from securing their citizens’ rights abroad.
The European Commission identified this lever in its report on demographic change from July 2026. Building political agreement to move forward is more difficult than the technical mechanism, because it touches on member states’ sovereignty over their social protection systems. Facilitating portability of rights could contribute to a more fluid European labour market and to lengthening actual working lives without necessarily extending the legal retirement age.
This link between residential mobility and rights systems extends beyond housing: it touches on the capacity of institutions to accompany life trajectories rather than fix them in national categories designed for a world where people were born, worked, and grew old in the same country.
What is at stake by 2031
The immediate horizon is that of ongoing reforms. Germany completes its shift to 67 years in 2031. Italy applies adjustments to pension conditions scheduled for 2027-2028. Romania adopted a major reform in 2023, which came into force on 1 September 2024, including notably a new formula, indexation, and a mechanism linked to life expectancy. Poland restored, in 2017, the retirement age of 60 for women and 65 for men.
The question that arises is not whether these countries will reform—demographic pressure increases the likelihood of adaptations—but under what conditions a reform can be both durable and acceptable. Several countries, including Sweden, Denmark, and the Netherlands, have advanced by investing in social dialogue, employment of older workers, and transparency of mechanisms. They have treated reform as a renegotiated contract between generations, not as a constraint imposed by markets.
This architecture is transferable. It is not automatic. It assumes institutions that have the capacity and credibility to keep commitments over thirty years, something that Italy, Poland, and Romania have not yet fully rebuilt after decades of abandoned reforms. The quality of institutions and social dialogue determines the acceptability and durability of any parametric reform.
Sources
- Eurostat, demographic projections 2023–2100, https://ec.europa.eu/eurostat/web/population-demography/population-projections
- Bruegel, Demographic Divide: Inequalities in Ageing Across the European Union, https://www.bruegel.org/policy-brief/demographic-divide-inequalities-ageing-across-european-union
- European Central Bank, Occasional Paper No. 296 on demographic ageing, ECB, Occasional Papers, no guaranteed URL
- European Commission, report on demographic change, July 2026, European Commission, DG EMPL, no guaranteed URL
- Bruno Palier, Réformer les retraites, interview Nonfiction, https://www.nonfiction.fr/article-10923-reformer-a-nouveau-les-retraites-entretien-avec-bruno-palier.htm
