It is well-known that small changes in the discount rate cause large variations in the present value of long-term liabilities.1 In this article, Woon Wong2 argues that the discount rate currently used to value the liabilities of the Universities’ Superannuation Scheme is too low, and that the largest higher education strike in British history followed by the forming of the Joint Expert Panel help to reveal that the claim of a deficit is fallacious.3
It has long been suspected that falling gilt yields overstate the liability of defined benefit (DB) schemes. It is thus surprising to find that, based on projected benefit payments data available from the University Superannuation Scheme (USS), ‘gilt-plus’ (gilt yields plus a fixed margin) discount rates actually reduce the reported 2017 deficit of the USS from £7.5bn to £3.4bn. This finding makes the reason behind the 2017 deficit ever more puzzling since the valuation assumed that gilt yields would revert back to their higher (2014) level in ten years’ time (thereby implying a deficit lower even than that obtained by a gilt-plus method). Regulatory guidelines require assumptions be evidence-based. If evidence-based discount rates are used, the scheme is found to be in a surplus that can be as large as £7.5bn.
Debating the discount rate
The problem with the USS’s valuations is related to the discount rate debate that was published by the Pensions Regulator (tPR) in the 2017 Annual Funding Statement for Defined Benefit Schemes.4 In that debate, proponents of gilt-plus valuations argued that low gilt yields mean low returns on other asset classes. Recent research, however, shows that gilt yields are driven by inflation whereas returns on equities (the most important asset class for pension funds) are determined by firms’ productivity in the real economy.5 The fall in gilt yields since the 1970s are the result of successful monetary policy to target inflation in order to ensure optimal economic growth, which in turn makes businesses profitable and hence healthy returns on equities.
Furthermore, Lord Paul Myners CBE (who was the UK’s financial services secretary during the 2008 financial crisis) questions the wisdom of discounting pension liabilities at the current low interest rates, which have been manipulated through quantitative easing.6 Such a view is shared by economists who regard long term interest rates as inappropriate discount rates after they have been used as monetary policy tools.7 Also, gilt yields may have been depressed relative to economic fundamentals by the Pensions Act 2004, since this requires pension managers to purchase gilts regardless of price.8 An investment bank estimates the potential demand for index-linked gilts as five times the size of the current market.9
The need for transparency
The USS has refused repeated requests by scheme members for information such as the projected benefit payments data. It is only after a member of the Joint Negotiation Committee (representing University and College Union) obtained the required information from USS that the findings reported in this article are made possible. It turns out that the reported discount rates are actually investment returns, not discount rates in the usual sense. Since the 2017 reported discount rates (investment returns) begin and remain at a very low level for ten years, the resulting actual discount rates are significantly lower; as shown in Figure1.
The above revelation raises the issue of frankness in the governance of USS. Letters have been written in the past pointing out to USS that the scheme’s deficits are the outcome of falling gilt yields rather than any true funding shortfall.10 While the USS denies the use of gilt-plus methods in the 2017 valuation, it fails to mention that the new approach actually produces a deficit that is higher than that obtained by gilt-plus discount rates. As Figure 1 illustrates, the 2017 reported discount rates being deceptively close to those of 2014 does not help convince its scheme members that USS is acting openly (in spite of the fiduciary duty of trustees to act honestly).11
The USS attributes the initial low investment returns to high asset prices buoyed by low interest rates and a possible market down rating. However, the first ten years of investment returns average only 0.93 per cent, which is significantly lower than the gilt yield of 1.727 per cent as at the valuation date. A simulation study can show that the low return assumption is equivalent to using a rate with a less than 10 per cent probability to value a DB scheme.
Also, the 2017 discount rates and inflation forecasts have varied like a rollercoaster for 50 years. It turns out that the discount rates are required to fall from year 11 to less than 2.8 per cent at year 34 in order to ensure the failure of Test 1 in the 2017 valuation.12 Needless to say, the rollercoaster variation of discount rates and inflation forecasts is inconsistent with the inflation targeting policy of the Bank of England.
Finally, both gilt yields and US interest rates were at similar levels at the time of 2014 valuation. Since Brexit, the spread of US interest rates over gilt yields has widened to around 1.3 per cent, which further confirms the assertion that gilt-plus valuation inflates pension costs. This is because (a) the assets of USS are internationally diversified, and hence (b) if the level of interestrates were to determine returns on other asset classes, it would be the US interest rate that calls the shots.
Role of academics
Earlier research has shown that managers have been known opportunistically to use downward-biased discount rates to inflate pension costs in order to obtain labour concessions.13 What happened to USS is worse: opaque valuation method based on un-evidenced assumptions provided the 2017 deficit which led to a decision to close the USS DB scheme? — an outcome that only the largest strike in British higher education history has been able to prevent. Note that as a result of continuing closure of DB schemes in the UK, the pension industry is anticipating an estimated additional £1 trillion annuities business in the next 20 years.14
The economic truth is that DB schemes enjoy intergenerational risk-sharing and time-diversification of risk that no other pension designs can match. Moreover, regulations require trustees to always act in the best interests of scheme beneficiaries, which means implementing the regulatory guidelines of evidenced-based assumptions for valuation. While the report of Joint Experts Panel (JEP) has vindicated the industry action, the recommendations of JEP remain merely advisory.15 Therefore, it is vital for economists (and other academics) to scrutinize the assumptions that underpin the valuations of USS.
1 See, for example, D Evans, ‘Climate change, the Stern Review and discounting the future’, Newsletter no. 141, April 2008, pp.8-9.
2 Reader in Finance, Cardiff Business School
3 See Letter to the Pensions Regulator.pdf for details of various calculations and discussions.
5 See W K Wong, ‘The Phantom Deficits of the Universities Superannuation Scheme.’ Cardiff Economics Working Papers E2018/12, 2018 and R.J. Caballero, E. Farhi, P. Gourinchas, ‘Rents, technical change, and risk premia. Accounting for secular trends in interest rates, returns on capital, earning yields, and factor shares.’ American Economic Review 107, 2017, pp.614-620.
7 L Booth, ‘Estimating Discount Rates.’ The School of Public Policy Publications 8, 2015.
8 R Greenword, D Vayanos, ‘Price pressure in the government bond market.’ American Economic Review 100, 2010, pp.585-590.
9 Schroders, ‘Pension funds and index-linked gilts: A supply/demand mismatch made in hell.’ 2016. Available at: 2016-06-pension-schemes-and-index-linked-gilts.pdf
10 See letter by Professors. Hutton and S Jacka of Warwick University. Available at: USSTrusteesDeficit21Nov2014.pdf
12 For a critique of Test 1 in 2017 valuation, see https://medium.com/ussbriefs/why-test-1-must-be-dropped-a-critique-of-its-design-and-implementation-1358c612a2a4
13 J. Comprix K A Muller III, ‘Pension plan accounting estimates and the freezing of defined benefit pension plans.’ Journal of Accounting and Economics 51, 2011, pp.115-133.
14 See SeminarUSS27Mar2019.pdf