This blog from Woon Wong is a very welcome addition to the  evidence that USS is architect of the predicament it finds itself in.

Despite a short balloting period, staff of 37 universities voted for strike action over changes that could see a 35% reduction in their guaranteed pensions. The Universities Superannuation Scheme (USS), the pension scheme at the centre of the dispute, explained the reforms are necessary because of the “fundamental truth” that guaranteed pensions were much more expensive today than they were in the past.

There is a widely held view that guaranteed pensions should be funded by a 100% risk-free bond portfolio and the cost is reflected in the level of risk-free interest rate. My recently completed paper shows, however, that it is possible to guarantee pensions while investing significantly in productive assets, and hence the cost of guarantee is considerably cheaper than the bond approach given the current negative interest rates.[1]

Consider a pension scheme in which the pension outgo and contributions from members remain the same from year to year. It can be shown that the cost to guarantee the promised pensions is the contribution rate required to pay for the outgo. Note this is not a Ponzi approach, as the scheme’s asset can be ensured to be larger than its liability. To put it into perspective, for the considered scheme with the same benefits as the USS, the contribution rate of guarantee is only 34.7% of payroll, whereas using real interest rate, the required rate is 58.6% of payroll. The USS is asking contribution rate as high as 57% of payroll to simply keep the current benefits.

For those who are familiar with arbitrage pricing theory in finance, the above discrepancy is unsurprising because competitive financial markets are incomplete for intergenerational products such as pensions and hence no arbitrage activities exist to ensure that interest rates can be used to price them.

It is perhaps easier to understand the cost of guarantee with simple illustrations of how increasing contribution rate can reduce a pension scheme’s insolvency risk, defined as the likelihood of asset less than liability at a distant future T years from now. For the above considered scheme, if T = 1000, the insolvency risk is less than 17% if contribution rate is 26% of payroll (which is the contribution rate of USS in 2017). If the contribution rate increases to 30.7% of payroll as is recently the case for the USS, the insolvency risk falls to a mere 2%. As a member’s lifetime is 60 years in the scheme, the probability of failure to pay for promised pensions would be less than one seventeenth of the above risk figures.

The above findings contrast sharply with Miles and Sefton’s recent analysis which shows as high as 40% chance that the USS could fail to pay for the promised pensions.[2] The reason for the difference is that the latter is based on a closed scheme (with neither new members joining nor contributions received from existing members), whereas I consider an open scheme which is in line with the recent comment of David Fairs of the Pensions Regulator (TPR) that “a truly open scheme could not mature, would not be expected to de-risk, and would be able to continue to invest in a long-term way.”[3] The USS is such an open scheme for which the above discussed cost of guarantee and insolvency risk are relevant and applicable.

The USS does not use interest rates for pricing. However, evidence suggests its past and current valuations are affected by interest rate disproportionately. My earlier paper finds a gilts-plus model based on real interest rate can explain up to 99% of valuations conducted between 2011 and 2017.[4] This implies that the USS’s recent past deficits are driven by falling interest rates rather than any true funding shortfall. The self-sufficiency liability, which forms the basis for risk management in 2020 valuation, is highly dependent on real interest rate.[5] Aon, the employers’ actuary, expressed concern “that the risk framework is more appropriate for a closed scheme.”[6]

Because the liability of USS has been repeatedly overstated by successive valuations, various economic inconsistencies surface. It can be shown that a gilts-plus model overstates the liability of a pension scheme when interest rates are falling.[7] It is thus shocking to find that the USS’s 2017 liability is larger than a gilts-plus model would produce. In the 2020 valuation, the covenant (financial strength of employers) is £20-22bn, which is insufficient to cover the difference between self-sufficiency liability (£101.5bn) and asset value (£66.5bn). However, the sum of all promised pensions in real terms is £84.4bn, which is less than the sum of asset value and covenant. The latest update indicates that the USS’s asset stands at nearly £90bn.

The USS seems to like complexity. Both TPR and Joint Expert Panel[8] find 2017 valuation highly complex. Yet strangely, as pointed out above, a single economic variable (real interest rate) can explain most if not all of it. When the high level of prudence in the 2020 valuation is challenged by stakeholders, the complexity manifests itself as “hall of mirrors” with as many as ten different lenses as answer.[9]

The following joke is shared in an informal meeting between USS’s stakeholders.

Question to actuary: “What is 2+2?”

Actuary’s answer: “What answer would you like?”

This joke does not apply to most actuaries; rather it illustrates a trust deficit faced by the UK’s largest private pension scheme.

[1] http://carbsecon.com/wp/E2018_22.pdf

[2] Avoiding a pensions disaster at UK universities | Financial Times (ft.com)

[3] DB funding code: busting a few myths | The Pensions Regulator Blog.

[4] http://carbsecon.com/wp/E2018_12.pdf

[5] https://www.thepensionsregulator.gov.uk/-/media/thepensionsregulator/files/import/pdf/understanding-db-_scheme-funding.ashx

[6] USS 31 March 2020 valuation – Aon report 9April2021.pdf

[7] Jan 2021 RES, p26-28

[8] https://ussjep.org.uk/

[9] USS 31 March 2020 valuation – Aon report 9April2021.pdf