Recently, there has been a move to allow surplus refunds in order to help “drive UK economic growth”. In reality, the surplus can only be definitively assessed when a pension scheme is finally terminated (evidence in Courage case, around 1986).

Those with longer memories will recall that we’ve been here before when “excessive surpluses” were regulated between 1987 and 2004 (at GAD, I was part of the regulatory team). It is fair to record that we discovered that not all actuaries understood the fairly simple rules in place then and I imagine that any new rules will be somewhat more complex. If anyone is interested, a personal mindmap summary from 2002 is here (yes, I now hate the colour scheme).

In February 2025, the Association of Consulting Actuaries published a policy paper, which laid out seven key challenges that would need to be addressed. Under challenge [2], they suggested that the assets in the DB pension scheme after releasing surplus – together with any supporting covenant and security – must provide a high degree of confidence to the trustees that members will still receive their existing benefits in full.

In order to assess what is essentially a probability, one must use a stochastic process; a net present value approach simply cannot do that. Relying upon the funding basis being so strong that there should be no problem suggests that one should be asking whether the funding basis is actually too strong. If I were still a pension scheme trustee being asked for a surplus refund, I hope that my co-trustees would agree that we need the sponsor to provide robust stochastic evidence at their expense.

While it’s still too early to sense what’s really happening, refunds seem to be hardly compatible with TPR’s risk aversion.