This article is by Jon Spain
Jon Spain is one of Britain’s foremost actuarial thinkers. You don’t have to be an actuary to enjoy this piece – indeed it might help for non-actuaries to peer behind the actuarial covers – as Jon allows you to! I’m grateful to Jon for sharing his thoughts with me – he’s referencing my blog
That USS have reported themselves to the Pensions Regulator was quite startling. As I understand it, under its own assumptions, they have “technically breached” one of their funding covenants due to plunging equity markets. Is this really such bad news?
So far, I have seen alternatives suggested, namely “higher funding”, “derisking to safe assets”, “discontinuing the scheme” and “Government intervention”. Below, I suggest alternative 5.
A crucial point is that, while market values at some future time will be relevant, current market values have no predictive power (Fama, 1965). When and how far the markets will recover is impossible to say. However, just because markets are low at present does not mean that they will remain low. The perception of a problem is being driven by the concept that “financial economics” has any relevance to long-term entities such as DB pension schemes. There is an abundant lack of evidence for that and it is being challenged elsewhere. Worryingly, “path dependence” is totally ignored; it really matters as opposed to just one year at a time.
The discounting principle has been known for over 2 millennia. Converting future cashflows to the present only consistently works out in reality if the discount rate is the inverse of the investment return. As the future is unknowable, there can be no uniquely correct discount rate. There will be one set of outcomes but nobody, not even an actuary, knows what that set will be. Using the evidence available is necessary and there must be room for valid differences of opinion. Equity risk premia are realistic and bonds may fail. While the USS approach may appear “prudent”, that can only be defined relative to a best estimate. So far as is public, no best estimate exercise has been undertaken so that any prudence present is impossible to define.
The real problem is that risk quantification is very poorly captured by scalars. This is especially the case when liquidity problems can’t be identified in advance. The huge concentration on risk, without reward recognition, is, in my view, unbalanced. In the real world, risks are only taken because of potential rewards (see Maurice Ewing interview, “The Actuary”, October 2018), not a new idea.
Indeed, as far back as 1952, Redington wrote that avoiding losses is the same as avoiding profits. Single numbers are not appropriate results for representing many future uncertainties, especially when we don’t even say what the result means (mean? median? mode? specified percentile?). In reality, scalars are grossly inadequate for indicating uncertainty of many possible outcomes so that discounting is inappropriate. Instead, we should be looking at multi-dimensional results with confidence intervals but we cannot do that with a deterministic approach. Instead of using discount rates alone, actuaries should show the uncertainty to the sponsors and trustees, using robustly supported stochastic projections. Let’s have more simulations rather than utterly misleading scalars – and much less dissimulation.
Indeed, there is a UK actuarial professionalism problem. This arises under the TAS regime in force since July2017, to which little attention appears to have been paid. Under paragraph [3.2] of the Framework Reliability Objective definition, transparency of assumptions is required together with communication of any inherent uncertainty. Under TAS 300 (pensions), communications shall explain comparison between discount rates used (or proposed) against expected assets return according to stated strategy. How can scalar results comply with those?
A recent technical paper (“O’Brien”) was presented on 09 March 2020, a week after TPR issued its consultation on “clearer DB funding standards”. Together with the current USS furore, a cynic might think that the paper was deliberately designed to bully UK pension actuaries even further into agreeing that the TPR guidance makes sense. Three years ago, the UK actuarial profession had the opportunity to explain why the current system is unfit for purpose. That chance was wasted and we are now faced with an even worse version.
So, let me propose a fifth alternative, namely “formal long-termism”, which has two strands. First, instead of using discount rates alone, actuaries should show the uncertainty to the sponsors and trustees, using robustly supported stochastic projections. We should recognise that the financials are more significant than the demographics. Using realistic best estimates seems likely to show that the position is nothing like as dire (see discrate.com). Secondly, this must present a superb buying opportunity for USS, improving the long-term financial position.
Well said Jon.
There is – objectively – no such thing as the liabilities. The liabilities figure has to be constructed artificially by the use of a PV formula that requires a discount rate. That is totally arbitrary because the discount rate chosen is arbitrary. That is where the comments of reporters and commentators ought to focus. Yet they ALL treat the reported pension liabilities as gospel. The standard of reporting – even in financial press – is very poor – and highly misleading.
As regards the short term reliance and covenant triggers, these have all been introduced in the past few months by the USS trustees. Short term volatility in the funding level has nothing to do with what happens in the long term. Short term reliance on covenant is really nonsense because the valuation of assets is not real because it is never realised, and the liabilities is arbitrary and does not necessarily reflect the pension benefits that need to be paid in the years ahead.
Both market values of assets and calculated liabilities can swing around dramatically without there being any changes to the long term position. We need to remember that the DPV formula for the valuation of a security is a mathematical formula. That does not mean it holds as an identity in the real world – as some people seem to believe.