Levers to mitigate the impact of demographics on public finances: the case of pensions

The upward pressure of demographics on public spending will be the dominant factor which, in the absence of measures, would lead to a deterioration of the public finances of developed economies in the medium term.1 Broadly speaking, in a scenario with constant policies, AIReF projects that public spending on ageing in Spain will increase between 2022 and 2050 by more than 5 points of GDP, of which 3.4 would correspond to pension spending, compared to a 1.1-point increase in revenues from social security contributions.2It would therefore be necessary to increase revenues by 2.3 points of GDP via transfers from the government to the Social Security system in order to finance the higher pension expenditure, unless measures are taken to reduce expenditure as a percentage of GDP.3

In this article, we focus on three major levers proposed by economists to mitigate the impact of population ageing on public finances and, in particular, on pension spending: prolonging working life, boosting productivity and attracting more immigration.4 Institutions such as AIReF and the European Commission have conducted sensitivity analyses on how changes in these levers could affect pension spending as a percentage of GDP in Spain. These are illustrative results that should be taken with caution given the high uncertainty surrounding the country’s future macroeconomic path over the next 25 years. In addition to these levers, we should also consider policies aimed at bolstering private pension schemes as an essential complement to public pensions, as well as measures to incentivise child birth.5

Starting with prolonging working life, increasing the employment rate among people aged 55 to 74 would considerably ease the pressure of ageing on pension spending. The Ageing Report 2024 (AR24) estimates that successfully raising this rate in Spain to 70% by 2050, rather than the 60% predicted in its baseline scenario with constant policies, would allow pension spending to be reduced by 1.4 points of GDP in 2050 compared to the baseline scenario – almost half the projected increase in expenditure (3.4 points). Achieving an employment rate of 70% by 2050 for people aged 55 to 74 seems ambitious; in the major European economies, also with constant policies, the projections for the employment rate in this age group in 2050 fall below that figure.6 In Spain, the current rate is 54%, so reaching the level of 60% anticipated7 in the baseline scenario would already represent a substantial improvement of 6 points, which incorporates the impact of the new design for the system of incentives to delay retirement under the 2023 pension reform.8 Policies that go further in this regard, such as allowing work and the receipt of a pension to be combined, would help to further boost the employment rate among these groups.

One avenue that other countries have followed to extend working life is to delay the legal retirement age. In AR24, they analyse the impact of delaying retirement by linking it to the increase in life expectancy; specifically, applying three-quarters of the increase in life expectancy9 between 2022 and 2050. This would mean that the legal retirement age in Spain would increase from 66.2 years in 2022 to 68.2 years in 2050, as life expectancy is expected to increase by almost three years in that period. The impact of this delay in the retirement age alone would represent a reduction in pension spending of 0.5 GDP points in 2050 compared to the baseline scenario in which the retirement age rises to 67 years in 2027 and remains constant at that age from then until 2050. Denmark, the Netherlands, Sweden and Estonia have already linked retirement age to life expectancy. These countries also tend to be in more favourable positions in terms of the sustainability of their pension systems.10

Improving productivity is the second major lever analysed in this article to counteract demographic pressure on public finances. Productivity is the main source of economic growth in the medium term, so accelerating it would reduce pension spending as a percentage of GDP, although the positive impact via greater economic dynamism would be partially offset by higher pensions driven by higher wages. It is estimated that an average growth rate of total factor productivity (TFP) of 1.2% over the next 25 years, instead of the 0.8% envisaged in AIReF’s baseline scenario, could decrease pension spending in 2050 by 1 GDP point and public debt by around 20 points compared to the baseline scenario.11 If sound economic policies are implemented – in education, attracting talent, and a good institutional environment, etc. – and if the deployment of artificial intelligence has a successful impact, it is not unreasonable to aspire to achieve that pace in the medium term, even though it is an ambitious goal. To contextualise this scenario, the average annual growth rate of TFP in Spain was 0.9% between 2015 and 2019.

Finally, immigration can be a lever that also contributes in the desired direction, since it is the demographic phenomenon that has the quickest and most direct impact on the working-age population, a key element for GDP growth. It is estimated that net migration flows of 385,400 people per year on average in the period 2024-2050, as predicted by the National Statistics Institute (INE), rather than the 275,200 used in AIReF’s baseline scenario, would reduce pension spending in 2050 by 0.3 points of GDP compared to this baseline scenario, thanks to the greater growth of the economy.12 In addition, public debt in 2050 would be reduced by around 10 points of GDP compared to the baseline scenario thanks also to public revenues contributed by this group linked to their labour incomes. To contextualise the estimation of flows, in the period 2000-2023 they averaged 356,000 per year13 and reached 630,000 in 2000-2008. However, there is significant uncertainty over the impact of this lever; according to the AR24 analysis, the reduction in pension spending as a percentage of GDP thanks to immigration would be greater, albeit based on different assumptions.14