Alarming deterioration in USS funding is based on an incoherent valuation methodology

At a 1 December Institutions Meeting*, USS scheme actuary Ali Tayebbi reported an alarming deterioration in the level of funding since the 31 March 2014 valuation. As of that 2014 date, the scheme had reported a £5.3 bn funding shortfall after factoring in agreed cuts in pensions — i.e., closure of final salary and its replacement with a hybrid of DB and DC — to bring this deficit down from about £12 bn. The scheme actuary reported that, as of 31 October 2016, the deficit had ballooned back up to £15 bn. This has arisen from a 53% increase in the valuation of the pensions liabilities since March 2014, which has outpaced a 36% increase in the value of the assets. According to the scheme actuary:

this measure of the estimate of the liabilities assumes that all of the key assumptions that were made in 2014 remain unchanged at this point. One of the key assumptions is the discount rate. And this measure assumes that the outperformance in the discount rate above the prevailing gilt yield remains at the same level [of gilts plus 1.7%] as it was in 2014. The effect of that is that the real discount rate will have fallen from around 2.5% in 2014 to around 1% as of the end of October. And that really is the driving factor for the increase in the liabilities and the estimated deficit of £15 bn as of 31st of October.

Moreover, if there are no further cuts to pensions, and assuming a deficit recovery period similar to the 17 year period agreed in 2014, the employer contribution rate would need to rise from the current 18% of salaries to an eye watering 37% in order to address such a £15 bn funding shortfall. About this, the scheme actuary commented:

I have talked about some very high numbers in terms of contribution rates. Now, to emphasise again, that at this stage, all of those assessments are based on maintaining the same approach to assumptions as was used in 2014. [my bold emphasis added]

It emerged from a later response to a question, however, that the October 2016 £15 bn shortfall is based on a fundamentally different approach to the setting of assumptions, in comparison with that used for the 2014 valuation. What emerged is that USS employs a ‘best estimate minus’ approach at its triennial valuations (i.e., the March 2014 valuation and the upcoming March 2017 valuation), but it employs a ‘gilts plus’ approach to the monitoring of the liabilities in the periods such as October 2016 that fall between valuations.

Here is the question, from Warwick’s Finance Director, which elicited this response: ‘Is USS willing to consider alternative methodologies to “Gilts +” for calculating an appropriate discount rate, such as “Best Estimate Minus”…?’

Here is the response from Guy Coughlan, USS’s chief risk officer:

…we’re going back to first principles [for the 2017 valuation]. …[CIO] Roger [Gray] explained how we come up with expected returns. We take best estimates and, for the 2017 valuation, we look at those best estimates and subtracting a margin of prudence as appropriate. …It is not a “gilts plus” process in terms of developing the valuation. It is really more of a “best estimate minus”. However, “gilts plus” is a term that we use for communication and … a process that we use in monitoring or tracking the funding level between valuation dates. So the only place that really gilts comes into the valuation process as a whole is that it has a bearing on what “self-sufficiency” is as a particular safe harbour for the valuation process. But it is definitely not a “gilts plus” process for determining the valuation inputs. [my bold emphasis added]

Coughlan’s account of the ‘best estimate minus’ method employed for determining the valuation inputs is of a piece with remarks by USS CEO Bill Galvin in his December 2014 letter in reply to the statisticians:

There has been much comment regarding the trustee’s approach to setting the initial discount rate, which has been characterised by some as a “gilts plus” approach. For clarity, although the discount rate is ultimately expressed as outperformance relative to gilts (for example, to enable comparisons to be more easily drawn for example with other schemes), the trustee’s approach is based on a “first principles” analysis, using data from different independent sources. This analysis starts with consideration of expected performance of each of the asset classes held in the fund. The expected performance of each of the various asset classes is weighted according to the trustee’s overall asset strategy to get an expected rate of return on the fund as a whole.

We learn the following, however, from Coughlan’s remarks that I have italicised above: the frightening and volatile reports of the level of the deficit between valuations are based on a fundamentally different and less reliable “gilts plus” method for estimating investment returns on the assets in the scheme. Between valuations, changes in the market yield on gilts are solely responsible for any changes in the value of the liabilities, as the discount rate and the assumed rate of inflation track these yields.

If USS’s portfolio were invested almost entirely in gilts and other assets that behave like gilts, it would make sense for their discount rate to track the gilt yield. But only 17% of USS’s portfolio consists of gilts, and a majority is invested in return-seeking assets such as equity and property whose performance do not track the gilt yield. Therefore, investment returns on such a diversified portfolio will not track the gilt yield, nor will they exhibit the volatility of changes in the gilt yield.

At the 1 December Institutions meeting, the scheme actuary offered a real time updating of the 31 October figure and reported that, as of the end of November, the funding situation had improved from a £15 bn to a £12.5 bn shortfall, given a rise in the gilt yield in November. It is not, however, credible to maintain that USS’s ability to meet pensions liabilities out of investment returns on its diverse portfolio has improved so dramatically during the last month, nor that it has deteriorated as much as the gilts-plus-driven discount rate has fallen since the 2014 valuation.

Three questions for USS:

1. Why does USS incoherently switch from a sound ‘best estimate minus’ approach to the setting of the discount rate at the valuations to an unsound (given the diverse nature of its portfolio) ‘gilts plus’ approach between valuations?

2. Why doesn’t USS instead continue to employ a ‘best estimate minus’ approach between valuations? Such an approach would reflect changes in market conditions and forecasts, not just for gilts, but across the different classes in USS’s portfolio. It would do so in a manner that gives proper weight to the extent of USS’s holdings in the different asset classes.

3. Why, moreover, does USS place any weight on funding shortfalls between valuations that are derived on a ‘gilts plus’ basis that they themselves recognise as unsuitable for their triennial valuations? Why, for example, does the scheme actuary report that USS’s Tests 1 and 2 — which involve risks associated with the need to raise employer contributions — have been violated, on the basis of a gilts-plus-generated £15 bn October shortfall?

* The linked video can be viewed by anyone who registers. Quotations of remarks by the scheme actuary and the chief risk officer are based on my transcriptions from the video. Here’s a link to a pdf of the powerpoint slides of the presentations. The relevant slides are 44–64.

Written by

Professor, Dept. of Philosophy, Logic & Scientific Method, LSE

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store