Addendum to ‘Understanding “Test 1”: a submission to the USS Joint Expert Panel’
Sam Marsh, University of Sheffield
Click for printable PDF —combines both USSbriefs32 and USSbriefs51
This is a USSbrief, published on 1 September 2018, that belongs to the OpenUPP (Open USS Pension Panel) series, and has been submitted to the UCU-UUK JEP (Joint Expert Panel). It is an addendum to Sam Marsh’s USSbriefs32, published originally on 10 July 2018, ‘Understanding “Test 1”: a submission to the USS Joint Expert Panel’.
The addendum, based on analysis of restricted data, exposes a major flaw in the logic of USS’s Test 1, and illustrates how the selective release of information by the trustee has resulted in a skewed picture of the health of the fund, potentially misleading employers into over-estimating the risk involved in maintaining the status-quo.
6. Addendum
On 1 August 2018, in response to a request I made at a Joint Negotiating Committee (JNC) meeting, I was sent cashflow data by USS which covered two things:
(1) Projected benefit payments from April 2017 until they reach zero, under the assumption that the scheme closed to new accrual on 31 March 2017;
(2) Projected benefit payments from April 2037 until they reach zero, under the assumption that the scheme remains open to new accrual until 31 March 2037, but closes thereafter.
I also had access, from previous JNC papers, to the projected contributions and payments for the years 2017–2036 under the assumption that the scheme remained open in its current form with unchanged benefits and contributions.
The significance of this data is as follows. The data in (1) is used to calculate the liabilities for the valuation as at 31 March 2017 under different discount rate assumptions, which allows not only the replication of the surpluses and deficits under all the scenarios discussed by USS (September and November valuations, best-estimates etc), but also under scenarios that USS have never publicly stated (for example, where no de-risking occurs). The data in (2) is used to carry out the calculations behind Test 1. The data I had access to from previous JNC papers allows the comparison of asset growth with the liability calculations under various investment growth assumptions.
6.1. Findings. My main findings are as follows. All figures as at 2037 are in nominal terms (that is, not adjusted for inflation or converted to present values). All references to prudence refer to USS’s adjustments from a 50% best estimate to a 67% likelihood of success. The figures I state are those derived by my calculations: they may not precisely match any provided to the Joint Expert Panel directly by USS, but the error is likely to be small. I asked USS to correct any misinformation, and they responded to say that the liability calculations “are reasonably consistent with [their] calculations”.
6.1.1. Funding position as at 31 March 2017 with de-risking absent.
(1) The technical provisions liabilities with de-risking absent total £60.4bn: a £0.4bn deficit.
(2) The liabilities under best-estimate assumptions with de-risking absent total £48.3bn: an £11.7bn surplus.
I have not yet attempted to calculate future service costs under these assumptions, which are more complicated to derive.
6.1.2. Liabilities as at 31 March 2037.
(1) The technical provisions liabilities as at 2037 under the September and November assumptions are expected, on a prudent basis, to be £112.5bn.
(2) Note that the above figure is set by Test 1 as the self-sufficiency liabilities less the target reliance on covenant. That means the self-sufficiency liabilities as at 2037 must be £(112.5+15.8)bn = £128.3bn, the target reliance inflated by CPI as at 2037 being £15.8bn.
(3) The technical provisions liabilities as at 2037 with de-risking absent are expected, on a prudent basis, to be £93.5bn.
6.1.3. Asset growth to 2037: November assumptions.
(1) Under the assumption that the scheme remains open in its current form with unchanged benefits and contributions, the assets as at 2037 would be expected to reach £112.2bn on a prudent technical provisions basis under the November de-risking plan.
(2) This above figure is highly significant: if there is no change to the benefits and contributions, the technical provisions deficit at 2037 is expected, under the prudent November assumptions, to have reduced to £0.3bn. In other words, under the November assumptions, the current contribution rate fully funds the benefits over the next 20 years and recovers all but £0.3bn of the stated £7.5bn deficit on a prudent technical provisions basis.
6.1.4. Asset growth to 2037: September assumptions.
(1) Under the assumption that the scheme remains open in its current form with unchanged benefits and contributions, the assets as at 2037 would be expected to reach £117.9bn on a prudent technical provisions basis under the September de-risking plan.
(2) The above figure shows that if there is no change to the benefits and contributions, the technical provisions deficit at 2037 is expected, under the prudent September assumptions, to have turned into a £5.4bn surplus. In other words, under the September assumptions, the current contribution rate fully funds the benefits over the next 20 years and converts the stated £5.1bn deficit into a £5.4bn surplus on a prudent technical provisions basis.
6.1.5. Asset growth to 2037: no de-risking.
(1) Under the assumption that the scheme remains open in its current form with unchanged benefits and contributions, the assets as at 2037 would be expected to reach £123.5bn on a prudent technical provisions basis if there were no de-risking.
(2) The above figure shows that if there is no change to the benefits and contributions, the technical provisions deficit at 2037 is expected, under prudent assumptions but with no de-risking, to have turned into a £30bn surplus (£123.5bn in assets, £93.5bn in liabilities). In other words, the current contribution rate fully funds the benefits over the next 20 years and converts the £0.4bn deficit into a £30bn surplus on a prudent technical provisions basis.
7. Implications for Test 1
As mentioned in my initial submission, the scheme’s assets are ignored by Test 1. But the underlying rationale for Test 1 is to make sure that the ‘safe harbour’ of a self-sufficiency portfolio is within affordable reach. To determine whether such a portfolio is within reach, what is relevant is whether the scheme has enough assets at Year 20 to get to self-sufficiency via 7% extra contributions from Years 20 to 40. The value of the technical provisions liabilities at Year 20 is only relevant to the affordability of a self-sufficiency portfolio at year 40 if it is to act as a proxy for the level of assets in the scheme. If there is a divergence between the expected level of assets at Year 20 and the valuation of the technical provisions liabilities at Year 20, it would fit the rationale better to use the former in the definition of Test 1 than the latter.
Were the reliance on covenant metric defined to be the self-sufficiency valuation less the prudently projected value of the assets, rather than the self-sufficiency valuation less the technical provisions, applying Test 1 to the funding position as at 31 March 2017 would lead to the following conclusions:
(1) The actual extent of the reliance on covenant at 2037 would be calculated as £(128.3-123.5)bn = 4.8bn;
(2) This falls comfortably within the target maximum reliance at 2037 of £15.8bn, so no de-risking or additional contributions would be necessary to satisfy Test 1;
(3) The 2017 technical provisions deficit would be only £0.4bn;
(4) The future service costs would be significantly lower than the rates stated by USS: most likely, no increases would be necessary at all.
8. Resources
The spreadsheet used to derive the figures in this addendum is available at http://sam-marsh.staff.shef.ac.uk/uss/valuation_modeller_AUG_18_dummy_data.xlsx. Note that this spreadsheet replaces the cashflows supplied by USS, which they will not give me permission to release publicly, with dummy data. This dummy data has been reverse-engineered fairly accurately by Patrick Lee, so the spreadsheet is functional.
I have also recorded a video ‘talk-through’ of the spreadsheet, available at https://www.youtube.com/watch?v=DlMxwIytgNc.
References
[1] Universities Superannuation Scheme, 2017 Actuarial Valuation (consultation document), 1 September 2017, available from https://www.sheffield.ac.uk/hr/thedeal/pensionupdates/ussvaluation
[2] Report to the USS paper: 2014 Actuarial Valuation, First Actuarial for UCU, November 2014, available from http://ucu.group.shef.ac.uk/campaigns/pensions/disputed-2014-valuation-of-uss/
[3] Response to the Valuation Discussion Forum, USS, 22 November 2016, available from https://www.uss.co.uk/how-uss-is-run/valuation/reference-materials
This is a USSbrief, published on 1 September 2018, that belongs to the OpenUPP (Open USS Pension Panel) series, and has been submitted to the UCU-UUK JEP (Joint Expert Panel). It is an addendum to Sam Marsh’s USSbriefs32, published originally on 10 July 2018, ‘Understanding “Test 1”: a submission to the USS Joint Expert Panel’. 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 #USSbriefs32, #USSbriefs51 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.
