We thought members would be interested in this excellent paper from the London School of Economics Pensions Advisory Group

——————————

27 November 2014

Response by the London School of Economics Pensions Advisory Group to USS Consultation on Technical Provisions and Recovery Plan

Here, as requested, are our comments on “the underlying assumptions which will be used to complete the formal valuation and more broadly the trustee’s approach as set out in the Statement of Funding Principles.”1

The views expressed are those of the LSE’s Pensions Advisory Group. The Group accepts that the scheme is facing an important challenge and that there is a need to find an equitable and stable solution. We are also clear that suggesting alternative valuation assumptions is not by itself a solution and needs to be supplemented with further measures.2 We would, however, like a solution to be founded on assumptions that reflect the genuine funding realities of the pension scheme. We find unconvincing the explanations of some of these assumptions and would welcome indications of why other options were rejected or not considered.3 We begin by noting that we share the concerns about the valuation assumptions that have been voiced by the professors of statistics, financial mathematics, and actuarial science (hereafter ‘the statisticians’) in their letter to the Trustee, which we have attached as an appendix to this letter.

We regard the following observation of the statisticians as especially telling:

“…moving to evidence-based assumptions on salary growth and RPI would show the scheme to be in healthy surplus on a neutral assumptions basis. Remove the de-risking assumptions and that surplus would be substantial. Substitute historic asset growth performance for Gilts plus and the neutral basis would show a very substantial surplus.”

In other words, the assets exceed the liabilities on a neutral or best estimate – which is to say an estimate that is neither pessimistic nor optimistic – of the value of the pension fund. It now becomes difficult to see how anything other than an “overly prudent”4 series of pessimistic departures from genuinely neutral assumptions regarding the valuation of the liabilities could transform such a surplus into the £12.3 billion deficit that is reported in the draft valuation results, although it is noted that this £12.3 billion is on a technical basis.5 Such departures also overestimate the cost of contributions for future service of pensions benefits.6 These departures seem to go too far, especially given that the Ernst and Young investigation of the strength of the covenants in a sample of universities found that they were robust.7 Given the multiple objectives of pension schemes, the degree of prudence should be optimised not maximised. We therefore ask what would an appropriately, as opposed to an overly, prudent adjustment of such a neutral best estimate be?

In answering this question, we note first that UUK’s advisor Aon Hewitt has advised UUK that “the current Statement of Funding Principles …states that, other than the discount rate, and longevity assumptions, all assumptions will be chosen on a ‘best estimate’ basis.”8 Hence, by the Trustee’s current principles, there should be no quarrel with the statisticians’ introduction of “evidence-based assumptions on salary growth and RPI”. If there is any dispute between the statisticians and the Trustee here, it will need to be narrowed down to the question of what is the best estimate of salary growth and RPI.

INFLATION. With regard to the Trustee’s estimation of the rate of RPI itself, we are reliably informed that economists and others who are expert on this matter regard 3.4% as too high rather than the best available estimate of RPI. Such a forecast for RPI would be the best estimate only if the best estimate of the gap between CPI and RPI were about 0.5% greater than the AV consultation document’s assumed gap of 0.8%-1.0% (the accuracy of which has been confirmed by the advice we have received). We also note that if the larger gap that would be necessary to justify a 3.4% RPI is assumed, then UUK’s proposed cuts to pensions will be exacerbated, given the manner in which revaluation is tied to CPI.

SALARY GROWTH. With regard to salary growth, we concur with the statisticians’ conclusion that RPI + 1% is unsupported, given the historical data to the contrary. We also note that general pay growth along these lines is unrealistic when looking forward, since there is little prospect for  RPI + 1% increases in revenue for the pre-92 higher education sector as a whole during the next Parliament. The unreality becomes especially vivid when RPI + 1% is combined with the assumption that RPI itself will be 3.4%. Our employers could nevertheless make such assumed salary growth come true by awarding increases of 4.4% per annum (or its long run equivalent) for the next several years. If, therefore, UUK does not challenge this assumption, but later refuses to award such pay increases because, as they must now foresee, they will deem them unaffordable, then employees will feel hard done by their employers twice over: first for accepting an assumption that forced cuts to their pensions, and second for failing to deliver the assumed pay increases. If, as we believe, our employers are not prepared to start awarding 4.4% pay increases, then honesty requires that UUK should request the USS trustees to revise their assumption of salary growth downward. Given a best estimate of salary growth and RPI, what would constitute a reasonably prudent adjustment of the other assumptions? In particular, what would constitute a reasonably prudent adjustment to the discount rate?

DISCOUNT RATE AND DE–RISKING. The most fundamental assumption about the discount rate is that valuation is based on the “gilts plus” method, though the Pensions Regulator is also willing to entertain a methodology based on actual asset holdings. The latter methodology is also expected to be prudent. As we understand it the “gilts plus” method is favoured by actuaries as the most prudent method. The preference for using this method needs justification in the light of a robust covenant as does the one percent prudent deduction from gilts’ returns as used in the valuation.

Bound up with this question is the extent to which USS should de-risk its investments – an investment strategy which has been justified on grounds of prudence and which also lowers the discount rate. In its 22 October 2014 submission to the JNC, UUK wrote that it “recognized the need for some investment de-risking, principally to respond to the increasing reliance which the scheme will otherwise place on the sector over time, and to help in reducing funding (and contribution) volatility.”

AV similarly states that “The trustee’s plan to reduce risk within the scheme would, over the long term, deliver increased contribution stability enabling some confidence that contributions would not become unaffordable”. We would like to make the following points regarding derisking and the discount rate:

1. De-risking is an ineffective strategy for keeping contributions affordable, since it increases the need and demand for an increase in contribution rates in order to avoid reductions in the income that retirees receive from their pensions. It is also important to realize that the level of the employer contribution rates does not necessarily capture the full effects on employers of de-risking, as academics with international mobility are likely to expect salary compensation in the form of higher pensionable salary to counteract the effects of pension reduction arising from reforms to USS.

2. Insofar as funding and contribution volatility are concerned, it would seem a more rational response, as suggested above, to set a prudent discount rate in a manner that accurately reflects the actual mixture of return-seeking and other investments – namely, ‘best estimate minus’ – rather than to de-risk investments into a mixture that more closely approximates gilts.9 We are perplexed by the intransigence of the Trustee in sticking to gilts-plus in the face of the sound arguments to the contrary that have been offered in the exchange of technical letters with UCU.10

3. USS claims that their de-risking strategy is justified on grounds that gilts provide a good match to the liabilities of the pension scheme and hence constitute a liability-hedging asset. However “there are no assets that perfectly match pension liabilities (except for purchasing annuities where the insurer takes all the risk instead). Pension liabilities move with salary inflation (no matching asset), lpi [i.e., limited price indexation] (no matching asset), have long duration (no properly matching asset) and longevity (no matching asset). Therefore, trying to use bonds to ‘match’ pension liabilities is doomed to underperform the liabilities themselves.”11 There are, of course, many other methods of de-risking and it would be helpful to know why these were rejected and why given the assumption of continuing low discount rates this therefore expensive method of derisking was chosen.

4. Funding and contribution volatility are, in significant part, a function of how volatile the deficit is. The existence and volatility of the deficit, however, is an artefact of assumptions regarding salary growth and RPI that fail to provide best estimates of these factors. (See above discussion.) Once, therefore, assumptions regarding salary growth and RPI are corrected to conform to the evidence, the volatility-based argument for de-risking substantially decreases.

Speaking more generally, it is widely accepted that one of the main advantages of a large defined benefit pension scheme such as USS is provided by the pooling of investment risks, which allows for the reaping of high returns on investment in an efficient but prudent manner over a long period of time, by smoothing over variations above and below the expected returns on return-seeking assets.12 The de-risking strategy would therefore defeat a key purpose of a defined benefit scheme.

THE RECOVERY PERIOD. We note that the AV consultation document reports that “there is good visibility regarding the robustness of the covenant over a 20 year time horizon; beyond which visibility is reduced although the expectation is that the covenant will remain robust” (emphasis added). The Pension Regulator has also indicated that longer periods than the usual ten years may be allowed where the covenant is strong. More generally, the regulator is willing to allow somewhat more optimistic estimates in recovery plans. In the light of these facts, we do not think there should be any doubt regarding the adoption of a recovery period of at least 20 years. The Trustee’s recommendation of a 15 year recovery period is, we think, another instance of “over-prudence”.

CONCLUDING REMARKS. According to the Pensions Regulator’s Code, “The trustees’ key objective is to pay promised benefits as they fall due.”13 In meeting this key objective, trustees and employers must not lose sight of the reason why it would be bad to fail to make good on promised pension benefits: because employees would be made significantly worse off on account of shortfalls in their income in retirement. We believe, for the reasons offered above, that the proposed de-risking of investments and the assumptions underlying the valuation of the liabilities of the fund have lost sight of this objective. We therefore urge “trustees and employers to use the flexibilities in the funding regime and work collaboratively” towards the achievement of a solution that will result in far less of a reduction in pension income than UUK has proposed.14

1 “2014 Actuarial Valuation: A consultation on the proposed assumptions for the scheme’s technical provisions and recovery plan”, USS, October 2014 (hereafter “AV”).These comments also take account of “USS: Consultation on Technical Provisions and Recovery Plan”, UUK cover note, 4 November 2014. We regret that our Pensions Advisory Group did not receive a further UUK cover note of 21 November 2014 in time to make use of it by your noon 28 November deadline.

2 The Group will be issuing a further statement at a later date, which addresses wider issues of pensions reform that fall outside of the remit of this technical consultation.

3 A number of other assumptions need to be reconsidered. For example, it is said that allowing for commutation would have no effect on the deficit. This argument needs to be justified.

4 http://www.thepensionsregulator.gov.uk/press/pn13-17.aspx

5 AV, table C.2.

6 AV, tables C.2 and C.5.

7 “Scheme Funding within USS: an engagement with Universities UK”, USS, December 2013.

8 “USS: Consultation on Technical Provisions and Recovery Plan”, UUK cover note, 4 November 2014.

9 “I close with an appeal to the profession to stop using the gilt yield + x% method of setting the discount rate for a valuation. To tell trustees that their scheme is 100 percent funded and then say it is 60 percent funded a short while later, using a method represented as reflecting the actual assets of the scheme, risks bringing the profession into disrepute. It is time to move away from this method which has no sound rationale and instead use methods that have a real-world interpretation, fit better with the Pensions Act 2004 and provide a firm basis for advice.” (Derek Benstead, “Pensions: The going rate”, http://www.theactuary.com/archive/old-articles/part-6/pensions-3A-thegoing- rate/ )

10 http://www.ucu.org.uk/circ/pdf/UCUHE231_att1.pdf

http://www.ucu.org.uk/circ/pdf/UCUHE231_att2.pdf

http://www.ucu.org.uk/circ/pdf/UCUHE231_att3.pdf

11 Ros Altmann, private correspondence with the Pensions Advisory Group. (Altmann is an authority on pensions as well as a member of LSE’s Court of Governors.)

12 As Ros Altmann notes: “USS is a different type of scheme from most of those in the private sector, because it is an open scheme. The private sector schemes are now almost all closed (at least to new members), which means that they are in run-off and have a shorter time horizon than an open ongoing scheme. This should allow a longerterm investment perspective for the asset allocation and assumed returns.” (ibid.)

13 “Code of practice no. 3: Funding defined benefits”, the Pensions Regulator, July 2014, para 22.

14 Quotation from the statement of purposes, ibid., p. 7.