July 2019 newsletter: The USS deficit - it's about time someone did see it and was clear about it, too
16 Jul 2019
In our last issue, Woon Wong argued that anxieties about the state of the Universities' Superannuation Scheme were unjustified. In this article Bernard Casey1 and John Ralfe2 [respectively a retired academic and one-time senior economist at the OECD and an independent pension consultant] beg to differ. They also draw attention to some interesting, and unusual, features of the USS.
Woon Wong has recently claimed that USS has no problems.3 Most pensions economists take a very different view.4 So, too, does the Pensions Regulator. USS also seems to be having second thoughts, whilst some of those who are currently USS sponsors are having doubts about the scheme’s sustainability.
The underlying economics of defined benefit (DB) pensions — including USS — are pretty straightforward. An employer promises to pay employees a pension for life, plus a reduced pension to any surviving dependent. The pension, and lump sum, is based on the employee’s salary and number of years in the scheme and increased in line with inflation.
To give employees security for their long-term pension promises, companies must hold assets in a separate pension fund. Employers and employees make cash contributions into the fund and the money is invested — with investment returns also flowing into the fund. However, pension payments are not contingent on the value of this fund. The company is legally obliged to pay pensions — the pension scheme is a creditor of the company — whatever happens to the assets in the fund that backs it.
USS is unusual because all employers have ‘joint and several liability’ for all other employers. If a single employer goes bust the remainder pick up a share of its pension liabilities. This means the USS covenant is as strong as the strongest individual employer.
What is the present value of promises to pay pensions over several decades when calculated according to a pre-agreed formula? The answer is the same as that to the much easier question ‘what is the present value of promises to pay interest and principal on a quoted bond issued by a company and having a lifetime of several decades?’ And it is easy — the present value is simply the price at which the bond is being bought or sold. The price changes as interest rates change — it goes up if interest rates fall, it goes down if interest rates rise.5
What about a non-quoted corporate bond, such as a private placement? To calculate the present value of a company’s non-quoted private placement, we need to discount the promised interest and principal payments at the yield, or market interest rate, on a similar publicly quoted bond. The ‘Law of One Price’ means that cash flows of the same amount, the same timing, and the same credit rating, all have the same present value.
Pension promises have the same economic characteristics as such a bond with the same pattern of payments, the same timing and the same credit rating. Thus, the present value of pensions is calculated by applying a bond discount rate.
The security provided by the pension fund assets means members are much less dependent on the credit rating of their individual employer — but they still take some credit risk. The right discount rate for DB pensions is a high quality or AA corporate bond.6 The value of the pension promises goes up and down, as interest rates fall or rise, just like the bond. Nothing tricky.
The value of assets is simply their market value, which, of course, also goes up or down. The deficit or surplus is simply the value of assets minus the value of liabilities on any given day.
At March 2017, USS’s published accounts showed £60bn assets and £77.5bn IAS19 liabilities — the AA bond yardstick required for all UK DB pension schemes. This gave a whopping £17.5bn deficit. By March 2018 assets had increased to £63.6bn, and liabilities had fallen to £72bn, reducing the deficit to £8.4bn.
However Panglossian some USS members might be, the Pensions Regulator is very unhappy and has written two very tough letters to the Chair of the USS trustees.7 Meanwhile one sponsor — Trinity College, Cambridge — is planning to pull out of USS entirely. Trinity is widely recognised as being the ‘last man standing’ were other sponsors to fail — i.e.it would retain responsibility for the pensions of all USS members. Trinity would rather not be in this, ‘albeit unlikely’, situation and is prepared to pay some £30m to buy itself out. Whilst its withdrawal would not, of itself, weaken the employers’ covenant, USS is of the view that the loss of another strong sponsor might well do so.8
Trinity’s senior bursar is not alone in pointing out that ‘[e]mployers individually have an underlying incentive to game USS and free ride on the other employers’.9 Perhaps USS, too, is beginning to lift its head out of the sand. As experience shows, there are limits to the length of the road down which cans can be kicked.
1. Bernard H Casey, SOCialECONomicRESearch, London and Frankfurt (firstname.lastname@example.org).
2. John Ralfe, John Ralfe Consulting, www.johnralfe.com, JohnRalfe@johnralfe.com
3. ‘Now you see it, now you don’t’ - Is the university pension fund really in deficit?, in RES Newsletter,185 April 2019.
4. See letter by 39 economists to Mr Bill Galvin, Group Chief Executive Officer, USS Limited, 12-10-18.
5. An example of a bond fitting this description is that issued for £300m by Cambridge University in 2018. It is repayable in equal annual instalments between 10 and 50 years and pays interest at CPI plus 0.25 per cent.
6. There is an argument that it is higher for USS, since some of the employers are AAA rated.
7. Mike Birch, Director of Supervision, The Pensions Regulator, to Sir David Eastwood, Chair of Trustees, Universities Superannuation Scheme Ltd, 11-12-18.
8. See ‘Leaked: Trinity has voted to leave the UK pensions scheme. If two employers exit, staff pensions may be in uncharted territory’, in Varsity 19-5-19. https://www.varsity.co.uk/news/17530.
9. ‘Revealed: Trinity plans exit from national pension scheme, isolating college from higher education sector’, in Varsity, 8-12-2018. https://www.varsity.co.uk/news/16687.
Woon Wong replies...
Bernard Casey and John Ralfe argue that the right discount rate for defined benefit (DB) schemes is a high quality or AA corporate bond. Moreover, they say, the discount rate for the USS could be even lower since all its employers have ‘joint and several liability’ for all other employers and hence its covenant is as strong as the strongest individual employer which is AAA rated.1 In essence, it is a bond-based approach to value DB schemes in which the discount rate is determined by corporate or government bond yield. Such an argument is emblematic of the confusion in the debate on the valuation of the USS as well as other occupational DB schemes in general.
It is easy to see the problem of such an approach by considering the following. Suppose we break up the USS into two schemes, USS-A and USS-B, which are exactly identical except for their sponsors. Both schemes have the same future pensions to be paid out but the sponsors of USS-A comprise AAA-rated institutions such as Trinity College, Cambridge, whereas those of USS-B have lower credit ratings. According to the bond-based valuation, the liability of USS-A will be larger than that of USS-B because the former has a lower discount rate. This does not make sense because it is equivalent to saying that the financial strength of universities such as Oxford and Cambridge should make the scheme more expensive to fund.
The reason for the confusion, as Carne (2004) explains, is that an ‘actuarial’ valuation of a DB scheme is not quite the same as determining the present value of bonds issued by a borrower.2 For an open DB scheme such as the USS, the funds (or cash flows) to pay for the promised pensions come from two sources: investment returns on the scheme’s assets and contributions from both employers and employees. An actuarial valuation will determine the scheme to be in deficit (surplus) and a higher (lower), contribution is required if the projected cash flows are smaller (greater) than the promised pensions. The so-called discount rate is primarily determined by the rate of investment returns on scheme assets. While Bernard and John are right to acknowledge that the security provided by the scheme assets means members are much less dependent on the credit rating of their employer, their problem lies in equating investment returns on scheme assets to a bond-based discount rate. As I pointed out in RES Newsletter 185 (April 2019), there is little evidence to suggest that the investment returns on productive assets have fallen by the same amount as the decline in riskless interest rates or corporate bond yields in past decades.
Also, the use of the AA bond as a source of discount rate for all UK DB pension schemes ignores the fact that the assets of some DB schemes are predominantly gilts whereas others invest more heavily in equities. In this regard, it is worth noting that the USS is an immature scheme with its current cash flow (from scheme assets and contributions) exceeding pension outgoings. For this reason, the Joint Expert Panel (JEP) recommends that the scheme is well suited to invest in high-risk-high-return productive assets. If this and other recommendations of JEP were implemented, the USS’s own calculation shows that the scheme would be in a surplus of £0.5bn in 2017, compared with the reported deficit of £7.5bn.
The problem with the USS valuation seems to be about more than the application of financial economics. Hoping to provide a rounded view on the different approaches to valuation, Dennis Leech of Warwick University asked a question in the December 2018 USS Institutions Meeting that would allow the scheme to consider an alternate cash-flow projection method described in Carne (2004). The treatment of his question, unfortunately, demonstrates the weakness of the current management of the USS: it fails to engage with the intellectual arguments surrounding the issue of valuation and wants to manage the scheme as if there is nothing to question about the methodology.3 Indeed, the Pensions Regulator (tPR) is currently investigating the complaint by Jane Hutton, who has served on the trustee board overseeing the USS, that she was frustrated in her bid to find out if the deficit was put at too high a level in 2017.4
Finally, it is interesting to note that the regulator rebukes the USS in relation to the bond- or gilt-based valuation.5 In setting out its approach to valuing the scheme, the USS said the regulator ‘prefers measuring discount rates relative to gilts’. But such a statement was described as ‘incorrect’ as the watchdog had no preferred approach to setting discount rates. As for the other statement where the USS said the ‘discount rate adopted for the 2017 valuation of gilts + 1.20 per cent is still above the level the Regulator views as appropriate’, the regulator said this was ‘factually incorrect’ as it had ‘not commented specifically’ on the level of the discount rate. Note that First Actuarial, the actuary for the University and College Union, uses the cash-flow valuation approach and finds no funding problem with the USS.
1. See footnote 4 of the letter by Bernard Casey and John Ralfe.
2. Carne, S., 2004. Being actuarial with the truth. Staple Inn Actuarial Society. http://www.simoncarne.com/truth.pdf
3. For further details about the meeting, see https://blogs.warwick.ac.uk/dennisleech/entry/uss_institutions_meeting/.
4. See https://www.ft.com/content/fdf66d42-83ab-11e9-b592-5fe435b57a3b.
5. See https://www.ft.com/content/3b6a2f18-8e94-11e9-a1c1-51bf8f989972