Collective Defined Contribution pension schemes are a new way to help people save for retirement in the UK. Here we explore how CDCs of the kind Royal Mail is proposing ‘smooth’ pensions and share risk.
Since the UK passed the Pension Schemes Act 2021, there is now a new way to help people save for retirement. In Collective Defined Contribution (CDC) pension schemes, both employers and employees contribute to a collective fund. They represent a third option for pension savers alongside traditional defined benefit (DB) and defined contribution (DC) arrangements.
The IFoA has been investigating CDCs as a potential solution for UK pension savers. Part of this exploration includes the IFoA-commissioned Optimising Future Pension Plans research. The project is headed up by Professor Catherine Donnelly and her team at Heriot-Watt University. Her research:
In this blog post, Professor Donnelly analyses the Royal Mail’s CDC scheme – the first viable CDC scheme structure in the UK. She explores how schemes of this design ‘smooth’ pensions and deal with risk sharing. Register for the webinar on 23 September to find out more about this research and join in the discussion.
One of the talked about benefits of CDC schemes structured like the proposed Royal Mail one, is their ability to ‘smooth’ members’ pensions. In other words, the total pension accrued by a member steadily increases over time because:
A smooth development of a retirement pension is ideal for individuals. Research by Lusardi and Mitchell shows that individuals who plan for retirement have better retirement outcomes. Pension smoothing assists in retirement planning, by allowing individuals to better predict the ultimate amount of pension they will have in retirement.
How does the Royal Mail CDC scheme achieve pension smoothing? My team’s recent research provides an answer. In essence, the scheme transfers risk from older members of the scheme to younger members. This risk transfer is enabled by the expression of risk-sharing in the scheme: the annual pension increases. To better understand what is going on, a technical dive into the inner workings of the Royal Mail-like CDC scheme is necessary.
At the heart of these CDC schemes is the risk-sharing mechanism, whereby the total assets of the scheme is equated with the total liabilities. The free variable which permits the equality is the annual pension increase. It is assumed that pension increases are granted annually on each member’s accrued pension, with the same pension increase granted at the same time.
Moreover, when calculating the scheme’s liability at a point in time, the accrued pensions are projected into retirement using the annual pension increase calculated at that point in time. (This is not quite the nature of the proposed Royal Mail CDC scheme, which includes an inflation element in their annual pension increases, but the conclusion is the same).
The annual pension increase is chosen so that the total asset value is equal to the total liability value. For example, suppose that the annual pension increase is calculated today to be 2%. Members are told that the annual amount of their accrued pensions has grown by 2%.
This one year’s worth of pension increases – which accumulates each member’s accrued pension from one year ago to today – becomes part of the accrued benefit of each member. But the future annual pension increases of 2% per annum, used to project the accrued pensions to retirement from today, are not guaranteed.
Suppose that a new pension increase is calculated one year later. The total asset value has grown by 10% since the last pension increase calculation. The younger members’ liability values – the discounted value of their projected pension paid at retirement – will increase by far more than that of older members.
This is because the younger scheme members’ liability is much more sensitive to changes in the net discount rate than that of older members. (The net discount rate is the difference between two factors:
The younger members are further from retirement, and so the effect of compounding means that their liability value is very sensitive to the net discount rate. Suppose the new annual pension increase, which balances the 10% higher asset value with the total liability value, is now calculated to be 3%. A younger member’s liability value may increase by 40% under this new pension increase whereas an older member’s liability value may increase by only 5%. In other words, the younger member has a liability value eight times more sensitive than that of the older member.
From the members’ point of view, their accrued pensions appear to have increased all by the same proportion: 3%. While the value of their accrued pension has changed by radically different proportions, this is not obvious to them. This is not a bad thing. Nobel laureate in economics Robert C Merton makes a compelling argument for the focus to be on a member’s income in retirement and not on their underlying value. (Read his argument in the Harvard Business Review.)
Turning to an investment standpoint, suppose that the members are individually assigned the same investment strategy in the CDC scheme. The individual investment strategy serves no purpose other than to determine the collective investment strategy. But because each young member’s liability value is more sensitive to changes in the net discount rate, the scheme effectively invests them in much riskier assets than older members. There is a transfer of investment risk, from the older members to the younger ones, at each point in time. The result is that older members are less exposed to investment risk than their individual investment strategy implies, whereas younger members are much more exposed.
To see why this investment risk transfer from older members to younger members smooths members’ pensions, let us conduct a thought-experiment.
Suppose, instead, that the members had been in an individual defined contribution (DC) scheme, and they all followed the same investment strategy. If the value of their individual assets all went up by 10%, then the value of their individual, notional liability would all increase by 10%.
In this different scheme, the implied annual pension increase would be much more volatile for the older members than the younger ones. It requires a much higher pension increase to give a 10% higher liability value if they’re only a few years out from retirement (as opposed to the luxury of time younger members enjoy).
What if the asset values do the opposite and fall by 10%? The older members must endure a much lower annual pension increase to drive their liability value down by 10%.
There is no pension smoothing in this individual DC scheme. Indeed, it is likely that older members would reduce their investment in risky assets to achieve pension smoothing through a different means.
We also explored a CDC scheme that did not use pension increases to express the risk-sharing mechanism. It rather applied the same change to the value of all accrued pensions. Our exploration here did not find evidence of pension smoothing arising from other design elements of the CDC scheme. This means that using annual pension increases to enable the risk-sharing is essential for pension smoothing.
Professor Catherine Donnelly FIA is Principal Investigator of the research programme “Optimizing Future Pension Plans” at Heriot-Watt University, which is funded by the IFoA’s Actuarial Research Centre.
