Crucial questions on the 2017 USS valuation
This is a USSbrief, published on 16 August 2018, that belongs to the OpenUPP (Open USS Pension Panel) series. It was submitted to the UCU-UUK JEP (Joint Expert Panel) by the author on 10 August 2018.
The submission below was prepared by the Cambridge UCU Executive Committee. It is not intended to include all areas of concern surrounding the USS valuation process. We have instead focussed on questions and uncertainties that we believe are most critical to the outcome of the valuation. Without in-depth answers to these questions there can be no confidence in any scheme valuation.
Alongside these questions, the panel must be mindful of the need for our pension scheme to consider issues of intergenerational fairness and equality.
1. On the valuation process itself
We understand that the valuation considers many parameters (demographics of USS membership, salary inflation, life expectancy, etc.) and compiles these into a model that produces an estimate of the Technical Provisions. This is used to generate a probability distribution, and the 33rd centile of which is used to calculate the existence and size of any surplus or deficit, when compared with the assets held by the scheme.
Several technical aspects of the valuation are not clear to us.
How is the probability distribution derived? Where does the error in each parameter come from (which might reasonably be summed to produce the very broad distribution presented to us)? We can see no formal derivation of the error in the USS documentation. Point estimates for each parameter would arrive at a single value for deficit/surplus, not a distribution, and we wish to have clarity on how uncertainty is introduced into the model. We believe that an accurate and transparent estimate of error is just as important in establishing the appropriate level of uncertainty (width of the distribution) in the valuation as the use of the 33rd centile of the overall distribution, which is where prudence is nominally introduced.
Further, we wish to be confident that true Best Estimates have been used to derive each parameter within the model. This might mean use of historic pay award data to estimate salary inflation, rather than [CPI + a premium], for example, where [below CPI] would more accurately reflect recent history. Data supplied by USS (Methodology and Inputs for the 2017 Valuation) for the proportion of female USS members dying with dependants is patently not the best estimate of the data presented. It appears that an extra margin of prudence has been applied to these data, and we wish to see whether the current valuation process similarly overestimates the cost of the scheme in each of the parameters of the model.
It is our view that, in an unbiased valuation, the stochastic nature of the distribution is where prudence should be introduced, that is by reference to the 33rd centile, not by rounding up the cost of the constituent parameters or overestimation of error in each parameter. Public clarification of these issues is crucial for any confidence in the valuation.
2. On the September 2017 valuation and the covenant
The valuation proposed by USS in September 2017 yielded a deficit of £5.1bn. In response to concerns on the strength of the covenant (from tPR) and the September UUK survey of USS institutions, the investment strategy was revised (to adopt earlier ‘de-risking’) and more conservative technical provisions were adopted, moving towards self-sufficiency. This resulted in lower projected returns and a lower discount rate, increasing the scheme deficit to £7.5bn in the November 2017 valuation.
The views expressed by USS institutions in the September 2017 UUK survey were often provisional, being formulated at a time when many institutions were unable to convene the appropriate authorities, and were sometimes the view of a single individual within an institution. Notwithstanding the dubious authority of the views expressed, USS responded to a 42% minority of institutions (curiously, not weighted by size) who expressed a desire for ‘less risk’. As a consequence, USS revised its investment strategy to propose an immediate move from growth-seeking assets into gilts and bonds. As a consequence, foreseen investment returns were downgraded, and the size of the deficit increased.
Since then, much pressure for transparency has been exerted on the USS institutions, which revealed the highly provisional nature of their responses to the UUK survey, which were in some cases unauthorised (and hence deniable!), since they had not been scrutinised by the appropriate decision-making bodies. Several institutions, including University of Bristol, University of Oxford and University of Cambridge, then withdrew their expressed appetite for ‘less risk’, and instead stated that they were content with the level of risk implicit in the September valuation. In other words, the majority of USS institutions accepting the present level of risk has increased yet further, to include several more major stakeholder universities with large USS memberships. This is a strong re-affirmation of confidence in the existing covenant.
Why then has USS persisted in basing its present cost-sharing proposals and its forward planning on the November valuation?
3. On the relative strength of the different scheme covenants
The Pensions Regulator (tPR) expressed the view (in the letter to Sir David Eastwood of 17 September 2017) that the USS institutions’ financial condition was weaker than hitherto perceived, and the covenant therefore correspondingly weaker. The justification for this statement is not obvious, nor is it appropriate. It is for USS to put a proposal to tPR, not the other way round.
There is lack of clarity over what determined tPR’s perception of financial stability, and how this influences the valuation. tPR suggests that USS was, prior to 2017, at the highest level of covenant, CG1, but was downgraded in tPR’s estimation to CG2. In turn, the covenant has influenced USS in the valuation, leading to an initial de-risking with projected investment returns of [CPI-0.53%]. Contrast this with Teachers’ Pension Scheme (TPS) 2017 valuation, with projected investment returns of [CPI+0.24%]. The same approach applied to USS would give rise to a £9.2bn uplift in the valuation (see Table 21 in Technical Provisions Consultation document), that is, an elimination of the November valuation deficit.
What is the basis for the different treatment of TPS and USS? We understand, of course, that TPS, like some local government schemes which also enjoy a higher discount rate, is underwritten by the taxpayer, but USS, prior to September 2017, was already CG1-rated, the highest of tPR’s ratings. Why then, are their covenants and corresponding discount rates so very different, given the anticipated longevity and stability of the higher education sector?
Bill Galvin comments in ‘How decisions are made on future pensions for USS’ on the role of judgement in assessing investment returns, and hence discount rate, a theme underscored in the ‘Summary of the scheme actuary’s valuation report (Rule 76.1)’, p. 3, point 7. The applied discount rate does not begin to approach the performance of the USS investment team, and we wish to understand why, when USS’s own data shows investment returns for the past decade or more far exceed those needed to fund the scheme, such a low discount rate is applied, and why Galvin’s judgment is so pessimistic.
We believe the arithmetic underlying the valuation is at present obscure, and in some places, inaccurate. This must be remedied.
We see no justification for USS clinging to the November 2017 valuation, given that the appetite amongst the institutions for ‘less risk’ is now much lower than was presented at the time.
tPR’s approach to the covenant is self-contradictory. If CG1 (the 2017 USS grading) is the highest grade, why are other schemes such as TPS given greater latitude than USS in projected rate of return/discount rate?
This is a USSbrief, published on 16 August 2018, that belongs to the OpenUPP (Open USS Pension Panel) series. It was submitted to the UCU-UUK JEP (Joint Expert Panel) by the author on 10 August 2018. This paper represents the views of the author only. The author believes all information to be reliable and accurate; if any errors are found please contact us so that we can correct them. We welcome discussion of the points raised and suggest that discussants use Twitter with the hashtags #USSbriefs41 and #OpenUPP2018; the author will try to respond as appropriate. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.