John Southall: A key concern with CDC schemes is that they may create intergenerational unfairness.
Collective defined contribution (CDC) is in the spotlight following the government’s publication of the multi-employer CDC regulations, with the first multi-employer CDC schemes potentially operating from early next year.
Most of the commentary has focused on its advantages such as an uplift in expected pension levels and longevity pooling. These are important, but they are only part of the picture. In this piece, we step back to take a broader view, exploring an angle that has yet to be fully examined in the current discourse.
There are various advantages to CDC schemes. Through pooling of longevity risk, and adopting a higher return-seeking investment strategy, CDC can offer a higher expected income for life than alternative approaches.
What is more, CDC simplifies decision-making for members, removing the need for investment decisions in the accumulation phrase, or drawdown decisions in the decumulation phase.
None of these benefits come for free. Longevity pooling compromises death benefits, higher expected returns come with greater uncertainty and simplified decisions inevitably impact individual flexibility.
Another key concern with CDC schemes is that they may create intergenerational unfairness.
In this article we identify the key sources of intergenerational unfairness and explain how accrual rates, which in a multi-employer CDC scheme can be different between members, may be adjusted to deliver a fairer balance between younger and older members.
How CDC schemes work
CDC schemes share similarities with defined benefit (DB) schemes, but don't receive deficit contributions from sponsors. Unlike Dutch CDC schemes, UK CDC schemes also can't use buffers - liabilities must always match assets. As such, benefits must be adjusted regularly to maintain balance.
How benefits are adjusted matters. For example, if scheme assets fall by 18% relative to liabilities, overall benefits must be reduced by the same proportion, but there are different ways this can be achieved. There are two main approaches:
- Scaling approach: All expected benefits payments are cut by 18%, with no change to the indexation rate of the scheme (the rate pensions are increased each year).
- Indexation approach: The rate of benefit increases (indexation) is reduced. For a scheme with a 20-year duration, reducing indexation from 3% to 2% per year results in an overall liability reduction of 18%. The impacts are much greater for longer-dated cashflows as you can see below:
Source: L&G calculations, 2026
The indexation approach stabilises short-term payment levels, typically benefitting those near or at retirement, but increases the variability of long-term payments. In effect, this allows a scheme to de-risk older members (since their payments are relatively near-term) without reducing how much of the overall scheme is invested in growth assets. But the investment risk doesn't disappear - rather it's transferred to younger members with longer time horizons[1].
The chart illustrates the greater downside risk for younger members. They also have more upside risk, but it isn't the case that "more downside is OK because there is more upside". If members are exposed to more investment risk, they ideally should be rewarded with a commensurately higher investment risk premium.
In practice CDC schemes are likely to use a hybrid approach of scaling and indexation, because a very high or low indexation rate is problematic[2].
Tackling potential unfairness
For the purposes of this article, we define fairness as follows:
Contributions equal the present value of expected benefits[3], calculated using an appropriate risk-adjusted discount rate, at the time of contribution[4].
Other objectives, such as simplicity, can be balanced against fairness but shouldn't be confused with it. Below we outline four potential sources of intergenerational unfairness and how they might be fixed by setting appropriate accrual rates which dictate how much expected pension is awarded for a given contribution. A key idea here is that there is no unfairness that - at least in principle - can't be compensated for via changing accrual terms.
(1) Time is money
Ignoring age when awarding expected pensions is "unfair" mainly because it ignores the time value of money. Providing an expected pension from retirement is much cheaper for younger members, who must wait longer to receive it. As an example, if expected real returns are 5% pa then a 25-year-old should be awarded approximately seven times[5] as much expected pension for a given contribution than a 65-year-old, purely due to the time value of money.
The solution is to allow for market-sensitive time value of money when awarding pension accrual. In other words, to achieve fairness accrual should depend on both age and market conditions. As we shall see next, however, this isn't quite the whole story.
(2) Investment risk transfer
CDC scheme regulations prescribe a best-estimate basis for the valuation of total scheme liabilities. That works for the whole scheme[6] but, for schemes that feature some indexation, is not ideal for individuals. This is because we ought to allow for both the expected level of target benefits and their risk.
A higher discount rate should be used for younger members because their promises fall further into the future where – as explained earlier - investment risk has been transferred. The discount rate should therefore include a higher risk premium. In contrast, benefits due to be paid soon are stable so should be discounted at close to risk-free rates. This isn't just saying younger people's accrual should be cheaper (our first point) - it's saying it ought to be even cheaper than that.
How much difference does this make? The answer depends on how much indexation is in the benefit-adjustment mechanism. No adjustment is needed for a pure scaling approach. For a pure indexation approach, our calculations[7] indicate that a 25-year-old could justifiably receive around double the accrual versus an age-related best-estimate approach[8]. For a sensible hybrid approach suitable adjustments are smaller and could merit around a 10-20% uplift.
(3) Perpetual motion
The way CDC schemes work is that the de-risking of members as they get older relies on there being younger members of the scheme to pass investment risk down to. However, this only works if younger members continue to join the scheme. If the music stops then a generation that had risk passed down to them doesn't get to do the same. This could be allowed for via a further boost to their accrual rates but is complex to estimate[9].
(4) Assumption uncertainty
If longevity expectations or expected returns turn out to be underestimated, the impact falls disproportionately on younger members. Younger members are essentially selling a form of insurance against assumption uncertainty to older members. But in practice this is extremely hard to quantify and explain.
Building trust for the future
CDC schemes offer attractive features, but their long-term success may depend on whether intergenerational fairness is treated as a core design consideration rather than an afterthought.
In this article we've seen that by clearly defining fairness and adjusting accrual rates to reflect differences in timing and, risk CDC schemes may be able to maintain their benefits whilst minimising potential inequities between generations.
Although perfect fairness is unattainable, balancing simplicity and equity is essential for building trust and sustainability.
John Southall is head of strategic research at L&G
[1] If a scheme adopted a cohorting approach, effectively segmenting investment experience for different generations, this would mean that benefit changes amount to largely scaling with no investment risk transfer from older to younger members. This would mean the only way to de-risk pensioners would be by reducing growth assets. In turn this would reduce the expected return on the scheme assets and expected outcomes.
[2] A negative indexation rate would result in a falling pension. A very high indexation rate e.g. 10% pa would result in very ‘back-end loaded' pension payments.
[3] Note this is calculated ex-ante (before the outcome is known), not ex-post (after the fact). Fairness means that members receive benefits of equal value to their contributions, not necessarily that every transaction is neutral in hindsight. Pooling longevity risk, for example, results in those who live shorter subsidising those who live longer, but who these people are is only known after the fact.
[4] A similar principle can apply to transfers in or out
[5] 1.05^(65-25) = 7.04
[6] Assuming the scheme is perpetual
[7] Which use risk-neutral pricing based on the same theory for pricing options
[8] For a CDC scheme with duration 20 years.
[9] assuming the scheme is perpetual one can relatively easily simulate future benefits and work out fair pricing. To allow for the chance the supply of new entrants changes requires more complicated calculations and estimations.
