Universities’ superannuation fund is accumulating surplus assets!
19 Jan 2021
Woon Wong1 has argued in these pages before that the valuation of the USS’s liabilities and the call for higher payroll contributions are incorrect.2 In this article he argues that the scheme is entirely viable and indeed accumulates surplus assets even at the pre-2017 contribution rate of 26 per cent of payroll.
1. A contrasting, positive, message
In its consultation (the ‘Consultation’) with the Universities UK (the ‘UUK’) on the proposed methodology and assumptions for the Technical Provisions in the 2020 valuation, the Universities Superannuation Scheme (the ‘USS’) reports deficits ranging from £9.8bn to £17.9bn, giving rise to contributions of 40.8 per cent to 67.9 per cent of payroll, respectively.3 Prior to the 2017 valuation, the contribution stood at 26 per cent of payroll.
In sharp contrast to the Consultation, this article provides evidence that suggests the USS is viable at 26 per cent of payroll contribution, and has been accumulating surplus assets with several benefits in waiting for the stakeholders.4 The benefits include (a) surplus assets to act as a further buffer to absorb investment risk; (b) the scheme will be self-sufficient which implies little support is required of employers; (c) future contributions lower than 26 per cent of payroll will be possible; and (d) it offers a cost-effective pension provision for the higher education sector that few other sectors and countries can rival.
The rest of this article is organised as follows. Sections 2, 3 and 4 provide evidence for the positive outlook whereas section 5 criticises the USS’s valuation methodology. Finally, summary remarks are provided in section 6.
2. The falling funding cost
There are two sources of funding to pay for liabilities, namely the contributions by stakeholders and investment returns on assets held by the scheme. Since the controversies arise mostly from the setting of the discount rate (a prudent estimate of rate of investment returns), we consider here the funding cost in terms of the required rate of investment returns, which is defined as the discount rate that equates the present value of liabilities to the asset value. Keeping contributions constant at 26 per cent of payroll, Figure 1 shows that the realised funding cost (based on realised asset value) has fallen drastically since 2011 to 1.2 per cent in real terms as of March 2020.
The first sign that the USS is sustainable at 26 per cent of payroll contribution is to note that the scheme would continue to invest significantly in growth assets, the long-term return of which is estimated by USS as 4.4 per cent, which is considerably higher than the funding cost of 1.2 per cent. Indeed, during the Valuation Methodology Discussion Forum (the ‘VMDF’) that took place earlier this year, the funding cost is projected to continue decline to negative territory in 2040.5
Falling future funding cost implies that assets would grow at a faster pace than the growth of liabilities. This prompted the USS to acknowledge that the scheme is fine in the long-term. To add to the good news, the valuation date of 31 March 2020 happens to be a low point for financial markets due to the pandemic. Since then, the USS’s assets have rebounded from £66.5bn to £75bn, giving rise to an even lower funding cost of 0.7 per cent (see Figure 1), a strong indication that the scheme is in surplus.
3. A reality check on discount rates
The positive message in the preceding section conflicts with the past and current deficits reported by the USS. Figure 2 provides an explanation.
The dotted bars in Figure 2 represent the discount rates used in the 2011, 2014 and 2017 valuations. They are significantly lower than the realised growth rates of the scheme assets (to reach the asset value in March 2020, represented by striped bars). For example, the discount rate for 2011 valuation is set at 4.1 per cent. The £32.4bn of assets in 2011 grew by 6.3 per cent per annum to reach £66.5bn in March 2020. The difference between the discount rate and the realised growth rate in 2014 has increased since, mainly due to a lowering of the discount rate.
The 2017 discount rate has fallen to 0.8 per cent. Despite March 2020 being the low point for financial markets, the realised growth rate of USS assets between 2017 and 2020 is higher than the discount rate. If the latest asset value (£75bn) is used, the realised growth rate (the rightmost bar) is considerably higher. In short, the USS’s assets have consistently grown at rates that are significantly higher than the discount rates. This not only explains the sharp fall in the funding cost over the past 10 years, but also implies that the USS’s past deficits could be due to overly pessimistic discount rates. The next section shows this is indeed the case.
4. Are ever lower discount rates justified?
This section shows that the lowering of the discount rate in both the 2017 and 2020 valuations are far more than what is justifiable by evidence.
For simplicity, suppose the USS invests only in gilts and equities. Let y and rE denote the gilt yield and expected return on equities, respectively. If the weight of gilts is w, then the expected portfolio return (rp) is given by:
rp = wy + (1-w) rE
A gilt-plus approach to the discount rate assumes perfect correlation between gilt yield (y) and expected return on equities (rE). For example, a 1 per cent fall in y is accompanied by the same fall in rE, resulting in the gilt-plus discount rate declining also by 1 per cent.
Evidence shows that the expected return on equities is broadly stable despite the falls in long-term interest rate in the past few decades.6 The implication is that as gilt yield falls by Δy in recent years, the discount rate would fall by w x Δy < Δy as w is only about 0.35 for the USS.
The USS has repeatedly claimed that it does not use a gilt-plus approach to set a discount rate. It turns out that its discount rate is lower even than that set by the gilt-plus approach. To illustrate, gilt yield fell by 3.5 per cent between 2011 and 2020. Since the discount rate in USS’s 2011 valuation is 4.1 per cent, the gilt-plus assumption would set the 2020 discount rate as 4.1 per cent minus 3.5 per cent = 0.6 per cent, which is approximately the upper end of the discount rates (0.0 per cent to 0.5 per cent) set in the Consultation. Since the proportion of low-risk assets held by the USS is less than half, the 2020 discount rate should be at least 3.5 per cent ÷ 2 = 1.75 per cent higher. A 1 per cent rise in the 2020 discount rate reduces the liability by approximately £16bn.
If readers think the pandemic may make the above example less convincing, setting the discount rate lower than the gilt-plus method is also found in the 2017 valuation. If a gilt-plus discount rate is applied to the 2017 valuation, the liability would be reduced by £4.1bn.7
5. Economically irrelevant methodology and un-evidenced assumptions
The deficits in the 2017 valuation and the current Consultation are driven primarily by the stipulation of a self-sufficiency portfolio (comprising mainly gilts and low-risk securities) in the valuation methodology to manage risks. In the former, the deficit is caused by de-risking the scheme portfolio to a hypothetical self-sufficiency portfolio over 20 years, in order to manage long-term risk. For the latter, the concern of short-term risk requires, among others, the sum of the employers’ affordable risk capacity and assets to exceed the liability of a self-sufficiency portfolio. Because of quantitative easing, the liability of a self-sufficiency portfolio becomes exorbitantly expensive, exceeding the sum of employers’ risk capacity and assets. Consequently, because of short-term risk, prudence is set at 73-85 per cent confidence level (cf. 65-67 per cent in past valuations), lowering the best estimate return by 2.1-2.6 per cent (cf. 1-1.1 per cent in past valuations) to arrive at the discount rates in the Consultation.
However, self-sufficiency is not required by pensions legislation. This is pointed out by the Association of Pension Lawyers in their recent submission to the Pensions Regulator on Defined Benefit’s funding code:
[N]othing in the legislation suggests that a move to minimise dependence on the employer’s covenant will always be appropriate or that trustees should be pushed in that direction… and of course could well be inconsistent with the sustainable growth objective [of the employer].
Moreover, successful risk management requires identifying and measuring risks that are relevant. The self-sufficiency portfolio is counterfactual and economically irrelevant to the USS because
a) It is open, immature, has positive cashflow with last-man-standing employer support.
b) There is no plan to de-risk the portfolio.
Therefore, the Joint Expert Panel (the ‘JEP’) set up in 2018 recommended more risk could be taken and criticised the hypothetical construct of a low-risk self-sufficiency portfolio in the 2017 valuation.
Consistent with the JEP’s view, in the VMDF, stakeholders disagree over the use of a self-sufficiency portfolio and suggest cash flows as a basis for managing risk in the 2020 valuation. For example, the UUK notes that
...[m]aking self-sufficiency the centrepiece of the Trustee’s risk metrics … is fraught with difficulties. We believe other methods — that are more directly linked to cash contributions — are more effective to measure risk.
The need for evidence and transparency
Regulatory guidelines require valuation assumptions to be evidence-based, and evidence suggests the impact of pandemic on valuation is considerably smaller than is implied by the low discount rates in the Consultation. A stochastic simulation finds the impact of the 1918 Spanish flu pandemic on the discount rate to be small. Intuitively, this can be understood by the long-term nature of the USS’s liabilities — they take 80 years to payoff, whereas ‘...[t]he pandemic is unlikely to have significant long-term consequences for the sector as a whole.’ (Consultation).8
Indeed, as financial markets are forward looking, the USS’s low asset value in Mar 2020 has priced in the negative shocks and increased uncertainties caused by the pandemic. The low discount rate imposed on an already depressed market is effectively double counting the price of risk. Since the USS is relatively immature, the cash from contributions alone are sufficient to pay for pension outgoes for almost 20 years. Unfortunately, the USS chooses to disregard such evidence and insists on using self-sufficiency portfolio to justify the high confidence level of prudence. Moreover, as the Consultation remarks, ‘[d]ifferent assumptions could produce lower confidence levels,’ no details are provided on what these assumptions are.9 Also, no evidence is provided by the USS to justify the assumptions that give rise to the high confidence level of prudence.
Finally, the UCU has long complained the lack of transparency in the USS’s valuation methodology, which has been described as complex by both JEP and the Pensions Regulator (the ‘tPR’). Therefore, evidence and transparency are vital for the USS to engage properly with its stakeholders, thereby resolving the disagreements in the 2020 valuation.
6. Summary remarks
Since quantitative easing to lower long-term interest rates is now an established monetary policy, gilts-based funding for pension schemes has become prohibitively expensive. The USS does not seek gilts-based funding; thus, gilts-aligned valuation methodology is inappropriate.
Most economists believe that quantitative easing benefits the economy, especially large institutions such as the USS. It is thus no surprise to find the scheme is not only viable at 26 per cent of payroll contribution, but also on a pathway to achieve self-sufficiency based on surplus assets.
What is concerning about the 2020 valuation is that the deficits and high contributions are due to issues that were neither resolved nor discussed in the VMDF. Indeed, the very low discount rates of 0.0 to 0.5 per cent in the Consultation are based on a methodology that uses un-evidenced assumptions and economically irrelevant input (a self-sufficiency portfolio).
Securing payments for accrued benefits may tempt the trustees to err on the side of excess prudence. On the other hand, an unnecessarily low discount rate using un-evidenced assumptions may render trustees breaching stakeholder trust. For tPR, securing the Pension Protection Fund and ensuring sustainable growth of employers provide similar opposing forces to the valuation. Surely, using evidenced assumptions as required by the regulatory guidelines to carry out the 2020 valuation is the best course of action for all parties concerned.
1. Woon Wong is Reader in Finance and Director of Trading Room Operations & Development, Cardiff Business School, Cardiff University.
2. No. 185 (April 2019). See aso the subseuent discussions in Nos. 186 (July 2019) and 187 (October 2019).
3. The USS is a privately funded pension scheme for pre-92 universities in the UK and the UUK is a representative organisation for the university employers.
4. Subject to agreement by the stakeholders, surplus assets may be kept in the USS via a contingent support vehicle.
5. The VMDF was formed in response to the second report of the Joint Expert Panel, and was participated by the USS, the UUK, and the University and College Union which represents the scheme members.
6. See Wong (2018) and the references therein.
7. See the letter submitted to the Pensions Regulator.
8. See page 63 of the Consultation.9. See page 26 of the Consultation.