Yesterday I published part one of Pension Oldie’s compelling argument that gilts are not the best way to discount pension liabilities for valuation purposes.
Tonight is the Mansion House Speech where we are expected to hear something about DB funding. It would be surprising, but not impossible for Pension Oldie’s ideas to be adopted by Government. We are sorely in need of innovative thinking – here it is.
Pension Oldie asks that you put your comments on the blog or send them to henry@agewage.com for transmission.
Should Pension Schemes Ditch Gilts
– At least for valuation of liabilities?
Pensions Oldie
Part 2 – The alternatives and conclusions
In part 1 of these thoughts, I set out my views on the issues associated with the use of gilt yields as the base for virtually all the current methods of estimating the assets and contributions required to pay pensions and other benefits, often very many years into the future, of a defined or targeted benefit pension scheme. In this part I set out my thoughts on possible alternatives to the use of gilt yields and my conclusions and recommendations.
The Alternatives
Alternative 1 – Inflation assumption used as discount rate:
The simplest and easiest implemented amendment that could be made would be drop the gilt yield as the basis for actuarial valuations of a pension scheme and replace it with a less volatile measure relevant to the scheme. The obvious candidate is already used in the valuation – namely the assumed future inflation rate determining the increases in the future liabilities – for pension increases and pre-retirement revaluations. This would reduce the number of the assumptions being made in the valuation to two, the mortality assumptions determining the period over which the pension is being paid and the future inflation assumption. Market determined inflation assumptions have been remarkably stable over the past 30 years and therefore this would avoid unrepresentative volatility in the present value of the liabilities. :
The main issue with using the inflation rate to discount the liabilities of the pension scheme is that you are effectively replacing one arbitrary discount rate measure with another. While the inflation rate assumption is applied to the future liabilities, discounting those liabilities back using the same assumption produces nothing more than the total real liabilities of the scheme. It will no more appropriately determine a surplus or a deficit of assets than using the historic gilt yield or any other arbitrary figure (e.g. inflation + 1%, or gilts + 0.5%, or just 5%). It would however appear to offer a more stable measure than using the highly volatile gilt yield and to that extent may be preferable as it would avoid the temptation to target an inappropriate measure in an investment policy.
Alternative 2 – Asset Based Valuations:
Asset based discount rate valuations were the norm before Technical Provision valuations were required. In an asset based valuation the discount rate is usually the weighted average of the historic investment returns of the assets held at the valuation date as a proxy for the expected return from those assets.. This has the advantage of recognising the actual assets held by the Scheme at the valuation date unlike the Gilt based valuation with its assumption that all assets will provide a return linked to the Gilt yield.
While the Pensions Regulator does recognise asset based valuations, it has stated it considers them only appropriate where the employer/sponsor provides a strong covenant towards the scheme to cover short and long term risks to investment returns or asset values.
Such risks are however often over estimated, for example the total cash dividends received from a FT100 index equity portfolio has been remarkably stable with growth highly correlated to inflation (even in the 2009 and 2020 crisis years the fall in total cash dividends was of the order of 17% and fully recovered within one year).
Similarly the worst case loss of capital value is usually measured on a one year basis in a “value at risk” analysis, however the pension scheme loss would only arise on the assets sold in that year and the forward dividend yield on the remaining assets is likely to have gone up reflecting the fall in their market value.
Asset based valuations are therefore preferable to gilt based valuations when considering a pension scheme running on into the future, either as an open scheme or with failure risk mitigated by either employer covenant or by the maintenance of a surplus within the scheme. In my view, actuaries should be required to provide a comparison under their professional standard TAS300 v2 paragraph 5 of the cost of buying-in a bulk annuity policy against the comparative position measured using an asset based valuation model.
This would help Trustees properly assess the relative risks to potential members’ benefits. Also investment strategy decisions should be based on asset rather than gilt based valuation models.
Asset return based valuation models are however not appropriate where the scheme is projected to need to sell a substantial proportion of its capital assets each year, such as where the scheme is fully closed and winding down. As indicated above, it is the maintenance of sufficient capital value that needs to be the goal in this situation.
Alternative 3 – Cash Flow Budget with Discount Rate Valuations only for the uncertain future
When a pension scheme matures so that the cash required to pay the benefits as they fall due has to be obtained from the sale of capital assets, the significance of the future yield expected from its investments diminishes to the point of disappearance in a super mature scheme. It therefore is inappropriate to seek to calculate the required market value of the scheme’s assets based on a volatile yield based discount rate.
For mature schemes where capital assets have to be sold on an annual basis, the key measure of the scheme’s capacity to pay those benefits should be based on an assessment of the cash flow required and identification of the assets to be sold in the year and the resulting effect on the following years. It is therefore suggested that the key evaluation of the scheme’s capacity to pay the benefits as they fall due should be based on a cash flow forecast with initial estimates based on the pension scheme’s accounts for the previous year. After all, well managed businesses should prepare cash flow budgets to manage the business, why should pension schemes be any different, indeed many well managed pension schemes already do?
A pension scheme forecast should project forward pension payments with inflation assumptions appropriate for each year (e.g. short term inflation above long term expectations would be reflected), lump sum payments, transfer payments, and administration costs to be paid. On the income side, interest, dividends, rent and other investment income for individual assets classes can be projected forward with appropriate individual assumptions. Redemption proceeds from fixed term investments can be mapped.
The model will then show the amount necessary to be realised from the capital assets of the scheme to meet the cash outgoings in each year with the resulting effect on income flows in subsequent years reflected and all necessary assumptions identified and assessed. Such a forecast will aid trustee decisions as to the selection of assets to be sold down to meet the cash requirements of the year and any liquidity buffers required.
It is suggested that this forecast should cover a nine year period (3 valuation cycles) with a non-volatile valuation (e.g. based on inflation discount rates or asset based on the projected future assets) used to provide an assessment of the funding requirements after the end of that period to ensure the scheme remains in surplus. It is suggested that mortality assumptions should not be reflected in the cash flow forecast, but that the forecast should be updated each year reflecting actual scheme experience.
Not only would such an approach indicate a clear yet still prudent picture of the scheme’s capacity to pay the benefits as they fall due, it would also provide a clear assessment of both the need for and the impact of any required additional (deficit or regular) contributions and their timing, It should also give clear guidance to investment managers of the scheme’s cash flow needs and its capacity to hold term investments on a buy and maintain basis and liquidity requirements and potential volatility risk. This should allow identification of the scheme’s capacity to hold private equity and infrastructure assets.
There may be other alternatives which I have not considered.
Conclusions
The UK Pensions universe has been particularly badly served by a legislative, regulatory, and advisory environment that has exclusively focused on anticipating pension scheme failure (particularly in defined benefit pension schemes).
Billions of pounds of resources have been removed from funds that would otherwise have been available to pay or enhance pensions or would have been available to sustain and grow employers’ businesses.
These funds have been lost to the Government (£3.8BN in 2022 to the Bank of England buying in the market and reselling back Gilts); the many times larger loss through the payment of inflated premiums to insurance companies subject to a different regulatory regime in so called “risk transfer” transactions justified against the volatile and at critical times over-prudent measurement of that risk; the encouragement of investment strategies which limit the potential for asset growth both for the benefit of the Members or the wider UK economy; plus the opportunity cost of employer and members’ contributions being directed into defined contribution pension pots which are much less efficient at converting those contributions into actual pensions in retirement. There is also the billions of pounds in levies collected by the PPF, essentially from employers, for which a large proportion is no longer on risk with the PPF
In their 3rd Report of the 2023/24 Session on DB funding, the House of Commons Work and Pensions Committee recommended that benefit accruing defined benefit pension arrangements should be supported. In particular they recommended that the objectives for TPR should be re-written to replace protecting the PPF with an objective to promote and enhance the future pension prospects provided by occupational pension schemes. It could also be considered whether the regulatory framework for DB pension should be aligned with that for insurance annuities and as to which alternative should be adopted or whether an innovative approach is required.
Against this background, it surely is time for a fundamental re-appraisal of the pension provision environment and that should start with a consideration of the methods of assessing and management of the resources required to pay the benefits of defined benefit pension scheme as they fall due. I hope this paper contributes to that vital debate.
PensionsOldie
13th November 2024