Sam Marsh’s brilliant reframing of USS’s Test 1
Sam Marsh, a mathematician who is a UCU branch officer for Sheffield, has produced three short Youtube videos which have caught the attention of USS.
In what follows, I try to explain, in words which are somewhat different from his, and which draw on an interpretation of Sam’s videos by Susan Cooper (who is Oxford’s UCU pensions officer), what I take to be the most important lesson of his videos. For purposes of illustration and simplicity, I employ the assumptions of the 2014 valuation.
As I explain here and here, USS’s Test 1 mandates that the scheme must reach the point, within twenty year’s time, that it is affordable to move to a low-return “self-sufficiency” portfolio. In order to ensure that such a move is affordable, Sam shows that it would have made more sense to set the technical provisions discount rate (which determines the level of contributions for a given level of benefits) as low at it is now set (namely, gilts + 1.4%, averaged over 20 years) without also de-risking the assets in the scheme’s portfolio over 20 years. This would have amounted to a highly prudent 1.35% deduction from the 2014 gilts + 2.75% best estimate of returns on the current return-seeking portfolio. The advantage of this approach over USS’s mandated 20 year de-risking of the portfolio from equity into gilts is that it would leave us with more assets at the end of 20 years to fund a move to self-sufficiency, rather than pouring this money down the drain through de-risking.
It would have made even more sense to make use of the extra anticipated investment income and assets in order to increase the technical provisions discount rate somewhat — e.g., to gilts + 1.55% averaged over 20 years — which would have allowed for an improvement in pensions benefits per pound spent in employer and employee contributions. That would still have left a larger 1.2% margin of prudence than USS’s margin of 1.05% on the return-seeking portfolio (i.e., gilts + 1.7% compared with a best estimate of gilts + 2.75%). This would have satisfied Test 1 while avoiding needless costs to scheme members and employers by making it more expensive to provide benefits.