France Is Aging Faster Than Its Pension System Can Absorb

In 2024, France records the lowest fertility rate since the end of the First World War [1]. The Retirement Advisory Council is revising its projections downward. Without a change in demographic or employment trajectory, the deficit in the pension system will reach 2.4% of GDP by 2070 [2]. The decision awaiting 2027 is architectural: choose between three levers with radically different timelines and costs, or continue to defer.

A fertility rate at its lowest in a century

The total fertility rate stands at 1.62 children per woman in 2024 [1]. This is the lowest level since the end of the First World War. In 2010, this rate was 2.02 children per woman. The number of births in 2024 is 21.5% lower than in 2010 [1].

The decline is accelerating. Between 2014 and 2019, the total fertility rate fell by an average of 1.6% per year [1]. Between 2022 and 2023, the decline reached 6.6%, then 2.2% in 2024 [1]. This pace reflects a regime change, not a cyclical adjustment.

France remains above the European average. In 2023, the French rate stands at 1.68 according to INSEE (1.65 using Eurostat’s methodology), compared to 1.38 in the European Union [2]. But the gap is narrowing.

Finland shows 1.26 [2]. Italy reaches 1.20 [2]. Spain drops to 1.16 [2]. France lost its position in 2023 as the most fertile country in the Union, which it had held since 2012.

A first interpretive pitfall deserves to be defused. The total fertility rate captures the effects of a postponement of the age at first motherhood, which now exceeds 31 years in France [1]. Part of the decline reflects a shift in timing, not necessarily a definitive renunciation.

But the final fertility of recent generations stabilizes below replacement level. Between 1995 and 2024, the number of women aged 20 to 40 decreased by 7.9% [1]. This decline mechanically amplifies the fall in births, regardless of individual fertility levels.

An aging trend already visible in natural increase

Life expectancy at birth stands at 85.6 years for women and 80.0 years for men [1]. Baby boom generations are reaching ages of high mortality. Natural increase, births minus deaths, drops to 17,000 in 2024 [1]. This is the lowest level since the end of the Second World War.

In 1980, people aged 65 and older represented 14% of the French population [2]. This share now reaches 21%. UN median projections place it at 28% by 2050 [2]. Over seven decades, France will triple its proportion of elderly people.

Comparison with neighbors offers only meager consolation. In Italy, the median age could reach 53 years by 2050, compared to 45 today [2]. Japan combines a fertility rate around 1.3 with one of the world’s highest life expectancies [2]. But having a less degraded age pyramid than Rome or Tokyo resolves nothing if financing institutions are not scaled for the shock.

The projections from DREES, revised in March 2024 and extended in December 2025, provide concrete measure [11]. By 2050, nearly 23 million people aged 60 or older will live in France. That is 5 million more than in 2021.

Among them, 738,000 additional people will experience loss of autonomy [11]. To maintain current care levels, 365,000 nursing home beds would need to be created between 2021 and 2050, in addition to the 640,000 existing in 2021 [11]. Between 150,000 and 200,000 additional jobs will be necessary to provide care at home or in facilities [11].

The sector’s finances already reflect the tension. Between 2020 and 2023, the share of nursing homes operating at a deficit rose from 27% to 66% [11]. The autonomy branch of Social Security represented 40.64 billion euros in 2024 [3]. It remains structurally insufficient. Out-of-pocket costs continue to deepen significant territorial inequalities [11].

The causal mechanism of pay-as-you-go pensions

The demographic shock strikes at the heart of the pension financing model. In 1970, France had 3.1 active workers for each retiree [2]. In 2026, this ratio has fallen to 1.7, for 17.4 million retirees [2]. The COR projects 1.62 active workers per retiree by 2070 [2].

The annual COR report published on June 11, 2026 marks a turning point [2]. The central fertility assumption is lowered from 1.8 to 1.45 children per woman from 2028 onward. This adjustment alone adds 1 percentage point of GDP to the projected pension system deficit in 2070. The deficit thus rises from 1.4% to 2.4% of GDP [2]. This is the largest annual revision since the COR’s creation in 2000.

Pension spending represents 14.1% of GDP in 2025 [2]. This share would remain at 14.2% in 2045 according to the COR’s reference scenario, before rising to 15.3% in 2070 [2]. The current deficit reaches 5.1 billion euros in 2025, or 0.2% of GDP [2].

Increasing contributions or raising the retirement age compensates for the effect without addressing the mechanism itself. The volume of future contributors depends on the number of children born today and their actual entry into employment. National transfer accounts, developed by Hippolyte d’Albis in his work on the economics of aging, document this chain [4]: the long-term dynamics of generations determine the balances that the media moment merely manages in crisis mode.

Senior employment follows the same logic. Germany employs 75.2% of its 55-64 year-olds [7]. The Netherlands reach 75.3% [7].

France struggles at 60.4%, according to the OECD, compared to 65.2% in the EU-27 in 2024 [7]. This lag reflects institutional rules that make early labor market exit rational for both employees and employers, as Bruno Palier and Christine Erhel document [6]. A problem of incentive design, not cultural attitude toward work.

Migration balance also contributes to system equilibrium. The 2026 COR itself revised this upward in its projections, raising it from 70,000 to 150,000 people per year [2]. Roughly 80% of immigrants are of working age [5].

They arrive already trained, contribute to financing pensions, and expand the base of active workers. Over the 1994-2008 period, increased permanent migration flows correlated with GDP per capita growth and declining unemployment rates [5]. The political debate treats immigration almost exclusively through its identity and security dimensions; it sidesteps its function as a demographic and budgetary adjustment variable.

The quality of jobs matters as much as their quantity. A system financed by poorly paid jobs, involuntary part-time work, and high physical strain generates fewer resources than a system built on stable, productive employment [6]. Productivity gains convert into a contribution base only if labor market institutions direct these gains toward labor income, a point Acemoglu and Johnson document in their recent work on the distribution of productivity gains [8].

Three architectural decisions that 2027 can no longer postpone

The main choice concerns the allocation between three levers, each requiring an irreversible institutional commitment.

Rebuild the conditions for early childhood care rather than distribute birth allowances. Comparative European studies show that fertility responds to confidence in the ability to reconcile work and parenthood [9]. The actual availability of childcare for children under 3, housing conditions, and job security at the start of one’s career are the variables correlated with the decision to have a child [9].

A natalist policy that compensates for the decision through financial transfers facilitates a choice without addressing the causes of postponement or renunciation. This requires massive public investment in daycare, housing policy oriented toward young households, and securing career paths at the beginning of working life. The delay for effects on pension balance is twenty years.

Raise the employment rate of seniors to 70%, by acting on job quality rather than only the legal retirement age. The employment rate of 55-64 year-olds with chronic conditions reaches 52% [7]. Among those in good health, it rises to 69% [7].

Countries that achieve the best senior employment rates are those that have invested in work quality at mid-career [6]. Action on ergonomics, prevention of occupational diseases, and retraining reduces future dependency needs. This lever produces effects on pension accounts within five to ten years. Its counterpart is investment in working conditions that employers, at this stage, have no incentive to finance alone.

Create sustainable dependency financing before the 2030-2040 shock makes trade-offs impossible. The cost of dependency is estimated at between 41 and 45 billion euros annually in a comprehensive measure including care, accommodation, and informal assistance [11]. DREES projects 30 billion euros annually for beneficiaries of personalized autonomy assistance alone, or 1.4% of GDP in projection [11]. The current model, fragmented between the state, departments, and households, produces documented territorial inequalities and an out-of-pocket cost that middle classes cannot absorb [11]. Choosing between mandatory insurance financing, a better-funded fifth branch, or resort to targeted retirement savings for the dependency risk represents an architectural trade-off that demographics no longer permits to defer.

These three decisions are linked [4]. Financing dependency without activating senior employment amounts to increasing levies on a shrinking contributor base. Betting on immigration without improving job quality expands a base of underpaid contributors whose own future rights will be equally insufficient. Investing in natality without creating the material conditions for childcare spends without measurable effect. National transfer accounts allow measurement of actual flows between generations and orient policy toward the quality of transfers to young people, rather than replaying the debate on pension levels [4].

Sources

[1] INSEE, “2024 Demographic Review,” INSEE Première no. 2033, January 2025, https://www.insee.fr/fr/statistiques/8327319 (consulted 09/08/2026).

[2] COR, Annual Report: Developments and Prospects for Pensions in France, June 2026, https://www.cor-retraites.fr/rapports-du-cor/rapport-annuel-cor-juin-2026-evolutions-perspectives-retraites-france (consulted 09/08/2026).

[3] National Assembly, written question no. 238, 17th legislative term, PLFSS 2025: autonomy branch 40.64 billion euros in 2024, https://www.assemblee-nationale.fr/dyn/17/questions/QANR5L17QE238.pdf (consulted 09/08/2026).

[4] Hippolyte d’Albis, Economics of Life Ages, Odile Jacob, 2026, https://www.parisschoolofeconomics.eu/en/personnes/hippolyte-dalbis/ (consulted 09/08/2026).

[5] Hippolyte d’Albis, Ekrame Boubtane, Dramane Coulibaly, “Immigration Policy and Macroeconomic Performance in France,” Annals of Economics and Statistics, no. 121-122, 2016; SFDS presentation, “Immigration in France: What Effects,” 2026, https://www.sfds.asso.fr/sdoc-14815-5896cc62e08c9cdf63962130b913f9ab-2026_02_cafe_stat.pdf (consulted 09/08/2026).

[6] Bruno Palier and Christine Erhel, Working Better, PUF/Vie des idées, 2025, https://www.sciencespo.fr/centre-etudes-europeennes/fr/actualites/travailler-mieux/ (consulted 09/08/2026).

[7] OECD, 2025 Employment Outlook: Older Workers, July 2025, https://www.oecd.org/en/publications/2025/07/oecd-employment-outlook-2025_5345f034/full-report/component-7.html (consulted 09/08/2026).

[8] Daron Acemoglu and Simon Johnson, “Can A.I. Be Pro-Worker?,” The New Yorker, 2026, https://www.newyorker.com/contributors/john-cassidy (consulted 09/08/2026).

[9] Eurostat, fertility statistics by member country, series demo_find, extraction November 2024, https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Fertility_statistics (consulted 09/08/2026).

[10] INED, “Demographic Challenges and Prospects in France 2020-2050,” working paper, https://www.ined.fr/fr/publications/editions/document-travail/enjeux-et-perspectives-demographiques-en-france-2020-2050 (consulted 09/08/2026).

[11] DREES-INSEE, LIVIA model “Projections of the number of elderly people in loss of autonomy by place of residence,” December 2025; study on staffing needs, February 2026, https://drees.solidarites-sante.gouv.fr/publications-communique-de-presse/etudes-et-resultats/soutien-autonomie-personnes-agees (consulted 09/08/2026).