The design of pension rules must take into account, through an actuarial neutrality approach, the adjustments to pension levels that are necessary in the case of flexible retirement ages. This column argues that a change in the average retirement age has an impact not only on pension funding, but also on resources available for other public spending. All else equal, incorporating this externality would imply penalties for earlier retirement and bonuses for later retirement that are much higher than those designed to balance the pension system alone.
The design of pension rules must take into account, through an actuarial neutrality approach, the adjustments to pension levels that are necessary in the case of flexible retirement ages (Queisser and Whitehouse 2006). The pension reforms undertaken in many countries over the last few decades take this factor into account. It sometimes gives rise to public debate, which is particularly heated in countries such as France in the context of the reform decided in 2023.
Marginal actuarial neutrality takes into account the fact that early retirement has a double impact on the pension system: a downward impact on total contributions and an upward impact on pensions to be paid. Symmetrically, later retirements generate a double benefit for the system. But this approach only internalises part of the collective effects of individual retirement behaviour. It ignores the induced effects of retirement choices on the rest of the public accounts. In Blanchet and Cette (2025), we show that because it impacts the total volume of economic activity, retirement behaviour also affects the ability to finance other economic and social needs. So, the question we raise is: what adaptation of the actuarially neutral scale should be envisaged to internalise this effect as well?
The standard approach: Actuarial neutrality restricted to pension financing
We consider the stationary state of a stylised pay-as-you-go (PAYG) pension system for a representative agent living for T years with an exogenous age of entry into working life. If he retires at a certain age R, he is entitled to a replacement rate of his gross wage which will depend on his retirement age. Pensions are supposed to be financed by a uniform contribution rate on the gross wage. In the usual version of actuarial neutrality, we do not take into account the provision of public services other than the benefits provided by the PAYG pension system.
Under these assumptions, actuarial neutrality corresponds to the equality between the mass of contributions paid during working life and the cumulative flow of pensions received during retirement, whatever the age chosen for retirement. The replacement rate will depend on the uniform contribution rate on the gross wage. This contribution rate determines the size of the pension system: a smaller rate implies a higher reliance on personal savings for smoothing the profile of life-cycle consumption.
When this overall neutrality is respected for any value of retirement age, we can derive the property of marginal neutrality that tells us by how much the pension can or must change for a given shift in the retirement age. For realistic values related to France, we obtain that retiring one year earlier (later) has to be compensated by a reduction (increase) of 6.66% in the pension level, here the replacement rate. This percentage has two components. The first, which accounts for about one-third of the total, corresponds to the fact that earlier (later) retirement reduces (increases) the contribution period. The second component, which accounts for about two-thirds of the total, corresponds to the fact that earlier (later) retirement increases (decreases) the pension benefit.
Actuarial neutrality extended to all public finances
We consider now a general public finance perspective rather than one limited only to the pension system. Within the total consumption of goods and services, a part is not financed privately out of individual income, but is provided collectively, financed by a general tax applied both to wages net of pension contributions and to pensions themselves.
In that case, the problem we face is that, in their individual optimisation, agents do not take into account the impact of their retirement behaviour on the ability to finance the collective expenses, via the fact that a lower (higher) retirement age imposes, at a given level of public expenditures other than pensions, a higher (lower) general tax rate, reducing (increasing) the standard of living, or that it limits (expands) the quantity of public spending at a given level of the general tax rate.
To internalise this effect, a more comprehensive budget balance condition must include pension contributions and spending, but also the revenue from the general tax rate on, respectively, wages net of pension contributions and pensions, and, finally, the public spending beyond pensions.
Equalising this financial balance to zero gives a new formula for replacement rates compatible with the overall equilibrium of the whole public finance. If individuals retire too early, with a given global tax rate, resources are missing for the financing of public spending. If we have no other correcting instrument than the level of pension benefits, we must impose a penalty on these benefits higher than in the case of the restricted actuarial neutrality approach. In that case, a surplus should appear for the pension system, to be recycled for restoring equilibrium for public finances taken as a whole. The opposite happens if people choose to retire later: for fixed levels of global tax rate and public spending, it is outside the pension system that a surplus can appear, to be recycled in the form of a pension bonus that is larger than under standard actuarial neutrality.
The respect of this enlarged neutrality gives us by how much the pension must change for a given shift in the retirement age to keep financed both pensions and other public spending. For realistic values related to France, we obtain that retiring one year earlier (later) has to be compensated by a reduction (increase) of 11.1% per year of anticipation/postponement instead of the 6.66% of the baseline profile. Thus, this percentage would almost double when we move from an approach based on actuarial neutrality restricted to the pension sector alone to an approach extended to public finances as a whole.
Questions and prospects
To sum up, if we take into account the effects of retirement behaviour on the financing of all public spending, rather than just pensions, we come to a much stronger link between pension and retirement age than the one deriving from standard actuarial neutrality. This finding makes sense, but raises, in turn, several other questions.
A first objection is to say that, in most countries, pension systems already have enough trouble balancing themselves, without the concern to add this additional constraint of contributing to the overall balancing of public spending. But, in that case, the problem is to find another instrument to manage the external effect of retirement ages on general public finances: the pension system remains the natural place to contribute to this control.
A second set of objections deserves further scrutiny, namely the potentially anti-redistributive impact of the proposed profile. The literature offers a series of arguments in favour of profiles that are not steeper but flatter than those implied by baseline neutrality. The reason can be the necessity to cover the risk of unemployability at the end of one’s career (Diamond and Mirrlees 1978, 1986), or purely redistributive considerations (Spinnewijn et al. 2022, Kolsrud et al. 2024, Landais and Spinnewijn 2024, Giupponi and Seibold 2024), when early retirement is more likely to occur among less-favoured categories who, in addition, have lower-than-average life expectancies. The modified profile we present goes in the opposite direction. Striking the right balance between this proposal and the opposite redistributive arguments requires further scrutiny. Raising this issue of redistribution is relevant in that one generally expects the overall financing of public spending to be redistributive, hence no reason for the correction of the externality to uniformly affect people endowed with unequal earning capacities.
Along the same lines, very steep profiles would appear particularly unfair if benefiting individuals with above-average salaries and pensions who, enjoying more favourable working conditions, would spontaneously choose to retire later, even without additional bonuses. They would benefit from substantial pension increases without even changing their behaviour and therefore without contributing more to public expenditures, destabilising the pension system without any favourable counterpart in the rest of the public accounts. To avoid this inconvenience, beyond a point to be determined, a lower bonus than that provided for in the modified scale could be applied. For individuals who would still be encouraged to retire later by this reduced scale, there would be a positive return on the public finances higher than the additional pensions to be paid, a difference that could be used to finance additional public expenditure.
Source : VOXeu
