Are the Young Being Done by CDC?

When the government finally authorised Collective Defined Contribution (CDC) pensions, many policy makers and journalists cheered. The Royal Mail scheme, the first CDC plan in the UK, promised a middle way between the gold‑plated certainty of defined benefits (DB) and the high risk of individual defined contribution schemes (DC). By pooling assets and sharing longevity and investment risk, CDC schemes were supposed to provide higher retirement incomes and smoother pension paths than individual pots. For the UK, facing an acute pension crisis, that promise resonated. Yet as the dust has settled, it has become clear that how we share risk matters just as much as whether we share it at all.

The hidden transfers in shared indexation

“A one-size-fits-all uplift is a bigger deal for a twenty-five-year-old because it compounds over five decades.”

The Royal Mail design uses a single lever: each year, trustees announce one pension increase or cut that applies equally to everyone. The simplicity is seductive because there are no secret buffers or ad hoc discretionary cuts. Yet the mathematics of this approach create a hidden tension. When the fund’s projected indexation moves, for example, from two per cent to three per cent, the present value of a younger member’s benefit can jump by around forty per cent. In contrast, an older retiree’s value might only move by about five per cent. The same percentage change compounds across many more future payments for a twenty-five-year-old than for a seventy-year-old. On the way down, the disparity reverses. The Pensions Policy Institute notes that flat accrual (the Royal Mail uses a uniform 1/80th of pay) also means that a pound paid five years before retirement buys the same pension as a pound contributed decades earlier. That skews value toward those near retirement and creates material cross‑subsidies.

This is not a flaw unique to the UK. Other countries that embraced CDC‑type plans have grappled with the same intergenerational trade‑off. Dutch schemes built reserves in good years to smooth downturns, but the young helped fill the buffer while those already retired were most likely to draw on it. Canada’s target benefit plans codify how surpluses and deficits are shared and require retirees to take cuts in bad times. The common thread is that if you prioritise stability above all else, you lean on younger members as the contingency fund. One observer captured it as a “heads I win, tails you lose” proposition: retirees enjoy the full benefit of the good years, while in bad years, much of the long‑term pain falls on those still accruing. This is not malice; it is a design choice that trades some fairness for simplicity. The question is whether younger savers will accept that trade‑off once they understand it.

Recent research argues that the problem runs deeper than buffers or accrual rules. There is an academic proof that sharing investment risk across generations cannot be mutually beneficial: mixing investment shocks via a uniform index inevitably creates winners and losers. Adjusting the accrual rate by age reduces, but does not eliminate, cross subsidies. Uniform indexation, therefore, delivers the illusion of fairness while hiding a heavy tilt against the young. However, longevity risk can be pooled fairly because no one knows how long they will live. A tontine‑like pool can equitably redistribute mortality credits in an actuarially fair manner. Sadly, the current UK CDC legislation prohibits a modern tontine.

Learning from abroad and from mathematics

“Across jurisdictions, the pattern repeats: cushions smooth headlines today at the cost of hidden cross-subsidies tomorrow.”

The Netherlands, long a poster child for collective pensions, has moved on from the old “doorsneesystematiek” and is now shifting to individualised accruals with a collective longevity pool under its new Future Pensions Act. Regulators learned that buffers and opaque smoothing can hide transfers and erode trust, so the new regime aims to expose and limit them. Each saver will have a notional pot that grows with actual investment returns, with a separate collective pool for longevity risk. That system preserves the benefits of long‑horizon investing while making each person’s share explicit. It recognises that fairness requires transparency about who bears what risk. That deserves credit. But it is not yet fair. Under the reform, mortality gains are allocated by age‑cohort rules and can be routed through a reserve. This keeps payments smoother, but credits are not calculated member by member, and there is no published fair‑transfer schedule. Transfers are smaller, not gone.

Canada moved early. New Brunswick legislated Shared Risk Plans in 2012, with fixed contributions, targeted benefits and mandated risk‑management rules that allow benefit cuts when funding falters. Alberta and British Columbia followed with target‑benefit legislation, and Ontario finalised a permanent Target Benefit Plan (TBP) framework in October 2024. The federal government consulted on a TBP regime for federally regulated plans, but has not implemented it. The Canadian Institute of Actuaries’ recent work frames the core challenge plainly: TBPs must balance benefit stability with intergenerational equity, because design choices on accrual and adjustment rules create transfers. In short, Canada has a clear legal chassis for collective plans, but fairness still depends on the rulebook each plan adopts.

Australia has not legislated CDC, but it has built a retirement‑phase framework that enables longevity pooling inside super. The Retirement Income Covenant requires every fund to design a retirement strategy, and Treasury’s position paper envisaged Comprehensive Income Products for Retirement that combine account‑based drawdown with a pooled lifetime income component. The Age Pension means‑test gives favourable treatment to lifetime products, with 60% of payments assessed as income and a step‑down in the assets test after a set age, which nudges designs toward pooling. Funds have responded with products like QSuper’s Lifetime Pension, which pools mortality risk and adjusts payments annually with investment results. This is a viable legal path to group self-annuitisation, but crediting rules are product-specific and not based on a published fair-transfer schedule, so “how fair” still comes down to design.

The US does not have CDC in law. The landscape is shifting. On 7 August this year, the White House issued an Executive Order that pushes the Department of Labor (DoL) and the SEC to broaden what 401(k) menus can hold, including alternative assets and lifetime income strategies. The order itself does not create pooled longevity products, but it directs agencies to revise guidance and safe harbours that have deterred plan fiduciaries. In parallel, existing rules under the SECURE Acts have already strengthened lifetime‑income safe harbours and expanded Qualified Longevity Annuity Contracts (QLACs). If the DoL explicitly recognises pooled “lifetime income investments” inside defaults, the US could leapfrog others by embedding longevity pooling in target‑date funds. For now, the door is ajar rather than open.

A better way: collective drawdown and fair transfer plans (FTPs)

A growing body of UK research, joined recently by King’s College London (KCL), proposes an alternative architecture known as collective drawdown. Instead of a shared index, every member has a notional defined contribution pot that is revalued in line with market returns. Longevity risk is pooled via an internal insurance market. This “modern tontine” style scheme shares the remaining funds of those who die earlier than expected among survivors on an almost actuarially fair basis. Investment gains and losses are not averaged across generations; each cohort shoulders the outcomes of its own investments, while sharing only the symmetric risk of living shorter or longer than expected. In simulations, this collective drawdown scheme delivers better pension outcomes than shared indexation CDC.

Separating investment and mortality pooling removes the largest intergenerational transfers; however, the KCL proposal is only approximately fair. To make it completely actuarially fair, it still needs the sort of fair transfer plan (FTP) tweaks this blog is championing. This means precisely calibrating mortality credits to age and risk. A FTP ensures that each participant’s expected present value equals their contribution, regardless of age or gender, so that no group subsidises another.

Adapting CDC legislation for fairness

If collective drawdown is such an improvement, why hasn’t it been adopted already? Part of the answer lies in regulation. UK statutes and the eighteenth‑century Life Assurance Act treat tontine‑style pools as regulated life insurance activities; there is uncertainty about whether modern longevity pooling is even permissible. The Royal Mail model has explicit enabling regulations and was politically easier to agree. Running a fair longevity pool also requires robust mortality modelling and clear communications to avoid the perception of profiting from death. Equality law could constrain how far mortality credits can vary across cohorts. None of these hurdles are insurmountable, but they do require legislative clarity and careful framing.

The Pensions Policy Institute and other researchers suggest two near‑term steps. First, allow age-graded or price-based accrual so that a pound buys a fairer slice of pension for younger workers. Second, pilot decumulation only collective drawdown schemes alongside existing CDC plans. A retiree-only tontine is arguably simpler to communicate (members already know they are in a longevity pool) and avoids the more sensitive question of varying accrual rates during the accumulation phase. However, in parallel, we should update legislation to permit full modern tontine pools and to recognise FTPs as a legitimate, regulated form of self-insurance.

Rethinking tax incentives

One reason modern tontines are often framed as drawdown products is that the current tax regime rewards saving on the way in but taxes pension income. Pension contributions in the UK attract tax relief, while the interest proportion of annuity or tontine payments is taxed as income. Each payment is split into a return of capital and interest; the capital portion is tax-free, but the interest part is subject to income tax. If we shifted the tax break from contributions to outcomes, for example, by making the interest element of annuity or tontine payouts tax-free, we could encourage people to choose lifetime income products without confining modern tontines to decumulation only. In effect, you would earn tax relief by committing to secure a retirement income rather than by simply putting money aside. This would level the playing field between accumulation and decumulation and allow modern tontines to operate in both phases of retirement planning.

Aligning tax incentives with outcomes rather than contributions would also support fairness. Younger workers would not feel penalised for staying in a longevity pool because the returns they receive in retirement would be advantaged relative to a taxed drawdown. The policy would dovetail with our goal of fair transfer plans: a transparent, actuarially fair pension that pays a tax-advantaged lifetime income.

Why fairness matters for the pension crisis

Britain’s pension crisis is real. Defined contribution pots are too small, annuities are unpopular, and most people will outlive their savings. And the Labour Party seem determined to rub salt into the wound.

What Labour’s IHT change really does to DC pensions

Collective solutions like CDC address a genuine problem: how to convert capital into a lifetime income without requiring a corporate sponsor. The Royal Mail scheme shows that unions, employers and regulators can innovate. Early modelling suggests that CDC could deliver higher average pensions than individual defined contribution. But averages are not enough. A scheme that systematically loads investment volatility onto the young or over-rewards late joiners will not command trust. The risk is that we simply replace one set of dissatisfied savers with another: defined contribution members angry about poor outcomes, swapped for young CDC members who feel short-changed.

For those of us campaigning for modern tontines [subscribers to this blog – thank you!], the lesson is clear. We need retirement products that are both sustainable and transparent. Pooling longevity risk is a powerful and fair way to provide lifetime income. Pooling investment risk across generations is not. The KCL collective drawdown proposal shows a path forward, and work on fair transfer plans demonstrates how to allocate mortality credits without hidden subsidies. By adapting CDC legislation to incorporate these ideas and by rethinking tax incentives so that lifetime income is rewarded, we can build products that appeal to both young and old. Modern tontines should be available not just at retirement but throughout the saving journey, with actuarially fair crediting and a fair tax regime.

In the coming months, the government plans to legislate for multi-employer CDC schemes and explore pooled drawdown. We should seize that opportunity to pilot collective drawdown tontines with actuarially fair crediting, and to consider shifting tax advantages to the outcomes we want rather than the contributions we happen to make. Fairness and simplicity need not be enemies. Only then will collective pensions earn the enduring confidence of those they aim to serve.

Call to arms

Modern Tontines

If you care about this, join me. I am looking for trustees, actuaries, asset managers and policymakers who want to pilot an FTP‑style crediting rule inside a collective scheme. The flexible contract in the Dutch system is the natural test bed, and the UK can run the same experiment in decumulation. Subscribe to Modern Tontines and consider becoming a founding member. You will get the research, models and policy notes first, and you will help fund the pilot work. If you are in a position to run a trial, contact me and let’s scope it now. Fairness should not be a slogan. It should be the rule.