The USS is the largest private sector DB scheme in the UK. Its triennial actuarial valuation attracts lively debate and comment from a panoply of pension experts and laypersons, and spicy exchanges between USS, UUK and UCU on DB funding and investing.

In this blog we bring clarity to the USS valuation process. We provide a framework for USS and UUK to complete their valuation, without reliance on subjective actuarial judgement and with full transparency on risk and its potential impact on pensions and the employers' finances.

Focus On Prudent Pension Outcomes, Not Prudent Discount Rates

To guarantee DB pensions, it is necessary to hold a portfolio of gilts whose proceeds exactly match the amounts and timings of future pension payments. Investing in anything else exposes pensions to risk.  UK pensions legislation, however, does not require that DB pensions are guaranteed, only that DB schemes are funded prudently. Trustees and employer can, if they wish, embrace a degree of risk to reduce the expected cost of funding for their DB schemes.

USS trustees and Universities UK, like most trustees and employers, have opted to use this flexibility and USS do not hold gilts to match pensions but instead invest in a diversified portfolio of assets.

Given this mutual acceptance of risk, and the statutory requirement for prudence, we would propose a funding objective that explicitly considers risk and its impact on outcomes for members and employers.

For example, USS and UUK could opt to fund and invest so that the pension scheme always has enough assets to be 95% certain of paying all pensions in full without recourse to the employer. That is, they could define their technical provisions to be the assets required, so that if the scheme was fully funded on technical provisions, there would be a 95% chance of paying all benefits in full.

‍
Alternate objectives could be to fund for a 90% (or 99%) chance of paying all pensions. A higher benefit security target (of say 99%) would naturally constrain the level of investment risk and require a larger pool of assets, mostly allocated to gilts and a larger contribution rate for ongoing benefit accrual. A lower target (of say 90%) means smaller technical provisions, a less conservative investment strategy and a lower ongoing contribution rate. Funding to be 100% certain of paying all benefits without recourse to the employer would, of course, mean holding the portfolio of matching gilts described above.

Technical Provisions

We estimate USS would need reserves of about £81.8 billion to be 95% certain of paying all pensions in full without recourse to the employer. On this basis the scheme has a shortfall of £15.3 billion. If a 90% benefit security target were adopted, then reserves of £70.0 billion would be required with a shortfall of £3.5 billion.  Our estimates are based on publicly available information on USS actuarial assumptions and pension cash-flows and internal Axis Actuarial models and assumptions.

Consider The Impact Of Risk On Employers And Members

Agreeing to hold enough assets to be 95% certain (or 99% or 90%) of paying all pensions would not, on its own, be enough to complete the USS actuarial valuation. Given its large size, USS and UUK should put in place contingency plans to deal with the scenario where the 5% (or 1% or 10%) risk crystalises and demonstrate that the pension scheme is able to 'fail safely'.

A key risk is poor investment performance. That is, investment returns are lower than expected over a number of years, and as a result, the pool of assets needed to pay pensions is too small. Unless action is taken, the scheme could run out of assets before all pensions are paid. Options for dealing with this include employer recovery contributions and / or reduced discretionary benefits, and we describe below a stochastic modelling exercise that USS and UUK could carry out to demonstrate how they could safely deal with the scenario where risks crystalise on the 5% (or 1% or 10%) worst case scenario:

1. Employer Recovery Contributions. Carry out stochastic projections of the future development of the funding position. Should the projected funding level fall below a safe level in any simulation, then allow for employer recovery contributions to start and continue to be paid until funding is returned to a safe level. The modelling will hopefully demonstrate that, once the ability of the employer to make future contributions is allowed for, USS will never run out of assets and pensions will always be paid in full, even in the 5% worst case scenario. The valuation can then be completed and submitted on this basis.
2. Reduce Discretionary Benefits.  The modelling in 1 above may reveal scenarios where the employer cannot afford the contributions needed to make sure that the scheme does not run out of assets. If that's the case then repeat the stochastic analysis, but this time should the projected funding level fall below a safe level, then allow for both employer recovery contributions and a reduced level of discretionary benefits (that is no discretionary pension increases, reduced commutation factors, lower transfer values, etc) until funding is returned to a safe level. The modelling will hopefully demonstrate that, once the ability of the employer to make recovery contributions and reduce discretionary benefits are both allowed for, USS will never run out of assets and contractual pension benefits will always be paid in full. The valuation can then be completed and submitted.
3. Action Needed Now. The modelling in 1 and 2 may reveal some scenarios where USS still cannot safely pay all contractual pension benefits in full. If these are thought significant, then options include reducing future service benefits, closing for accrual or a contingent asset should these risks crystalise.

It’s worth noting that adopting a lower funding target (of say 90%) will make it more likely that employer recovery contributions and reductions to discretionary benefits may be needed in the future. A higher target (of 99%) provides more certainty on pensions and contributions - although pension costs are higher.

Benefit Security Should Be Set By The Employer And Trustees

With our proposal, the statutory requirement for prudence is not met by choosing a prudent 'gilts plus' discount rate, or by embedding a margin into a 'best estimate' discount rate or any other discount rate methodology. Instead, the requirement for prudence is met by targeting a sufficiently high level of benefit security.  For example, it would be prudent to always hold enough assets to be 99% certain that all pensions will be paid in full, but a 50% target would not be prudent.

In fact, with our proposed approach, subjective judgements on margins for prudence and the actuary's choice of discount rate methodology have no place in the valuation. This is because the technical provisions, investment strategy and future service contribution rate are all determined by the trustees and employer, via their agreement on a suitably prudent level of benefit security. The discount rate is simply a balancing item needed to give effect to the technical provisions - and it makes no difference if the actuary expresses this as a 'gilts plus' rate, an 'inflation plus' rate or if he uses a 'yield curve approach', a 'split pre and post discount rates' or any of the many other discount rate methodologies used by actuaries.  

This is exactly as it should be. The level of benefit security in a DB scheme is a decision for the employer (as paymaster) and the trustee (as scheme manager). DB benefit security should not be set by actuaries, via their judgement as to suitably prudent discount rates or their choice of discount rate methodology.  

Conclusion

Despite a plethora of published material on the USS valuation, we have found no objective mathematical modelling on risk and its potential impact in terms of contribution outcomes for employers and pension outcomes for members. It seems that the focus of hundreds of hours of meetings and discussion on the 2020 USS valuation, has not been on risk and acceptable employer and member outcomes, but highfalutin debate on the nature of prudence and the merits of various obscure discount rate methodologies.

As a result, none of the parties seem to have a good understanding on how secure pensions in USS actually are, and no way of knowing how to agree a valuation outcome that fits within the employers’ pension budget and maximises the level of benefit security for members. Given this, it’s not surprising that negotiations on the 2020 USS valuation are difficult, with mutual incomprehension on both sides and USS and UUK both hoping that exchanges of jeremiads will be enough to complete the valuation.

USS and UUK have a responsibility to pay pensions to thousands of members. Decisions they make on the 2020 valuation will impact the allocation of £billions in assets, and the level of future service benefit accrual for thousands of employees. The 2020 valuation seems an opportune time for USS and UUK to ‘up their game’, with stochastic modelling for more informed, evidence-based decisions that considers actual outcomes for employers and members.

Our framework provides full transparency on risk and would allow USS, UUK, UCU and TPR to provide direct input on a suitably prudent level of benefit security.  It makes clear the circumstances when employers might need to pay recovery contributions and the amount of those contributions. Employers and members will also have clarity on when discretionary benefits might need to be reduced and when the scheme might need to close for future accrual.

Please contact us if you would like to discuss this.

‍