This article first appeared in the FT on this link.
I know that many members of USS read this blog and some members of its executive. I hope this is interesting and fair. If you have a view please use the comments below as your correspondence is always informed.
Success in applying the rules set out in law is not reliant on loopholes (I prefer educated projections of future…
I think a lot of people are holding on for worse news! I agree with you that the market now…
[…] better social capital can come from Government but is better that it comes from Britain’s largest and most patient…
[…] ← When it’s too late to “flex” then fix (with an annuity) […]
Great to see you put the correct positive gloss on this change. Pensions were never intended to be a means…
One of the problems the USS faces is the use of Gilt yields as the basis of its valuations and targets.
It makes no sense setting a (Gilts plus 2.9%) target for liabilities stretching 70 years into the future based on the most volatile variable (20 year Gilt yields have varied from 0.6%. to 9.37% over the past 35 years and even intra-day variability frequently reaches 0.2%).
In a much smaller scheme fully open DB scheme we have set a fixed long term investment target of 5.9% p.a. This was the effective whole scheme discount rate at the 31st March 2024 Technical Provision valuation. This target is used for investment decisions and also setting contribution rates and variable benefits (our defined inflation exposure is limited to pensions in payment). In setting the target we considered multiple factors including the scheme investment performance over differing periods over the past 15 years (the data to hand and which normalised at over 7% p.a.), the yields and trends of yields of current and potential investments, and some comparative scheme performance (e.g. CAPAData default 5yr returns 30 years pre-retirement – now c6%). We will formally review the target after the 2027 valuation, but while we may notch it up a bit we are unlikely to move it substantially because the yield on Gilts and bonds is now so much higher than the dividend yield on quoted equities (c. 3% on the FT100, 1.5% on the MSCI World) meaning projected future capital gains are dominate valuations of quoted and private investments.
The Trustee now requests investment consultants and managers “decompose” projected investment performance of existing and prospective investments into three components:
1. Current cash yield (annual interest or dividend payments flowing into a cash or liquidity account under the Trustee’s control)
2. Growth (annual expected growth in the current cash yield)
3. Revaluation (the expected capital gain from the investment not arising from 1 and 2)
• A justification for the Revaluation element should be provided against some suggested measures. Timeframes relevant to the revaluation return should also be clarified. The Trustee also expects that any particular investment may have more than one element to its projected investment return and will consider its contribution to the overall portfolio.
Using that analysis we have established a three portfolio approach, matching against the “armadillo” chart of future projected benefit payments on a buy and maintain basis:
Cash Yielding Portfolio
• Assets which generate short term cash flows into the Trustee’s bank account to cover the pension benefits paid out during the year.
• A Cash Yield Target (£M p.a.) set which is designed to match cash requirements for the next N (?3) years
• To avoid need for asset sales (no liquidity risk for the Scheme)
• To permit new contributions to be invested in the Future Impact Portfolio
Future Impact Portfolio
• Assets purchased with current contributions targeting benefit payments in the 20 to 60 year time horizon.
• Sustainability Issues key
• Liquidity not an issue so allows Closed End funds and Private Capital
• Long Term Yield targeted
Buffer Assets Portfolio
• Assets invested for return which will be called upon from time to time (every 3 years?) to top up the Cash Yielding Portfolio as benefit outflow increases.
• Existing Growth Assets (66% of total assets)
• Long term Yield targeted
• Liquidity important but flexibility in timing of transfers to Cash Yielding Portfolio mitigates and minimises significance of “Value at Risk”
This approach is designed to insulate the Scheme from volatile valuation assumptions over the long term and also make meaningless some of the so called risk measures e.g. volatile gilt yields when income flows are already secured, no required asset realisations within a three year time frame mean no 1 year Value at Risk etc.
Peter Cameron Brown
I think the FT critic has fundamentally misunderstood what DB schemes are designed to do and how they differ from DC.
With DC, growth over the investment duration is everything, the aim is simply to get the member’s pots as big as possible. The scheme will be able to look at the likely investment duration, how members are likely to use the pot at the end e.g. buying an annuity or taking tax free cash and entering drawdown say, and aim the default accordingly. If members aren’t happy with the level of risk they can self-select to modify the strategy.
With DB the pot size is irrelevant to the member since they have the guaranteed income. Changes to the employer contribution may be of interest, and changes to their own contributions and benefit accrual definitely are, but they are not what the scheme is primarily aiming to do. A DB scheme is fundamentally aiming to have sufficient assets so that the right pension can be paid at the right time to the right beneficiary. That means a more cautious approach than pure DC arises.
The Trustees may choose to aim for more growth and take on more risk but that has to be with a weather eye on the strength of the covenant and future contributions that would be required if things went wrong.
Peter Cameron Brown has quite reasonably argued for DB schemes to resist closing and/or de-risking too much too soon since ongoing contributions and investment income can benefit both members and sponsor in an open scheme, but I’m pretty sure he’d also argue against going for maximum growth since that would increase the risk of problems potentially forcing the scheme to close in due course.
Where the scheme is already closed, or extremely mature and/or with questionable covenant strength over the long term (unlike universities I’d hope), then the only concern is meeting the past liabilities. It is therefore sensible to keep a large part of the investment in defensive assets so that the prime objective can be met and the risk of catastrophic losses avoided.
Oh, and I think this confusion leading to DB being viewed as a pot by members is an unintended consequence of pension freedoms, and the regulators and politicians as referring to all members, including DB members, as “savers”!
I have had many, many conversations with DB members about them wanting to move from DB because their pension is losing money (CETV’s having declined as gilt rates have gone up).
I agree the major problem with DC is that it has been designed as a savings plan not to provide an income in later life.
Investment are therefore managed to maximise value of funds (for managers’ benefit) and taking income regarded as a disinvestment to be discouraged.
“prefers to compare the returns with a liability proxy (a mixture of gilt maturities)”
Are those Index-Linked gilts, given that the pension payments are CPI-linked?
Scheme actuaries use a nominal gilt yield curve, which they assert directly matches the currency and inflation exposure of those cash flows. It embeds both the real rate and expected inflation in one number.
Using only ILGs would give a real yield, but you’d then have to add an explicit inflation assumption, introducing extra model risk and further subjectivity.
Using an ILG yield would also on occasions introduce a negative discount rate to the mix.
UK index-linked gilt real yields were last consistently negative as recently as 2022.
Thanks for your reply. I hope it does successfully “embed both the real rate and expected inflation in one number”. Our avoiding a cat-food diet depends on that.