£1 Billion Difference Between DLUHC Savings Claim and SF3 Data?
Why does SF3 data show £1bn difference from DLUHC pooling ‘savings’ claim?
Earlier in the summer I challenged DLUHC’s claim that pooling had ‘saved’ £380m. I was not suggesting they were necessarily wrong to make this claim. But the only publicly available data, the SF3 data provided to DLUHC by all LGPS funds, was in direct contrast, suggesting that investment costs paid by LGPS funds have doubled over the last 5 years. Two of the pools, LPPI and ACCESS, have also come out with big claims of savings in the past week.
The public data is undoubtedly flawed and indeed one of the few parts of the recent DLUHC consultation which I can unhesitatingly support is the proposal to standardise reporting. There are several factors which probably lead to its overstating the increase in investment costs.
One explanation is that the data historically only picked up investment costs paid directly by Scheme Managers, and not those which were subsumed into a fund’s investment performance. Over the past five years cost reporting in this area has become more honest, thereby increasing stated costs. A second is that 2021 in particular saw very strong performance by private equity firms, which generated large performance-related fees for some Scheme Managers. A third is that the allocations to expensive alternative managers has been increasing. A fourth is that DLUHC’s figure is calculated against a notional figure of what would have been spent without pooling.
Data shows a £620m investment cost increase over 5 years
All these points have some truth in them. But the most recent SF3 data published earlier this month for the 85 English and Welsh funds showed a £m increase of £620m (or 9% annualised) in investment costs over the five years to March 2023 – i.e., the approximate period since pooling started. In basis points (i.e., allowing for the increase in asset size) they have risen from 39 to 49bps or by 25%.
In blunt terms, that is a £1bn (or 31bps of total LGPS AUM) difference from the savings claim which DLUHC has used to justify the push to more pooling. Again, I need to make it clear that this does not of itself invalidate pooling, which may indeed be beneficial. However, the Government cannot rest its case on savings made so far without providing more transparency. I believe it would greatly be in its and the pools’ interest to do so: if their claim stacks up and pooling really is saving money then a lot of the resistance to it will fade away.
Investment income is becoming crucial to maturing LGPS
Another important point is the level of investment income received. This fell from 1.63% of AUM in 2019 to a low of 1.10% in 2021 as companies suspended dividends over the COVID lockdowns. It has now recovered to 1.42%. Of course most funds choose not to take income from some income-producing assets, and some managers (and pools) choose not to offer income-paying fund classes. So the potential level of income is certainly much higher.
It matters because the LGPS is maturing quite fast. In aggregate pension payments exceeded contributions in 2023 by 1.1% of AUM. 1.4% of investment income only just covers that gap. Payments of course went up 10% from April 2023 and can be expected to rise by a further 7% next year. Contributions will have been adversely affected in 2023 by the historic tendency of big employers to pre-pay at the beginning of each triennial contribution period (i.e., in 2021). In theory 2024 contributions should be increased by pre-payments, but now that cash yields 5% (as opposed to almost zero three years ago) the incentive to pre-pay has lessened and may even have disappeared.
Almost all Scheme Managers now need investment income to fill the shortfall. The few that do not are almost without exception those with relatively poor funding levels and higher levels of secondary contributions (i.e., to make up past service deficits). If funding levels improve, the secondary contributions will reduce. For example, in 2023 they constituted 12% of total contributions in aggregate, whereas in 2018 it was 25%. The sharp increase in pension payments makes it clear that in a very few years all will be needing investment income to balance their books.
25 LGPS funds already in the red even including investment income
A cause for worry is that some 25 Scheme Managers are in deficit even including investment income. Unless they wish to raise contributions even further the implication here is that they will at some point need to make one of three things happen:
Consciously raise the level of income generated and paid to them, either by shifting their asset allocation in the direction of income-generating assets, or by switching to distributing classes. Pools which do not offer the latter may like to take note.
Rely on distributions from maturing private assets
Sell investment assets to meet the shortfall
The third has the potential to be damaging if sales take place at times of financial stress, and the timing of the second is outside the control of the Scheme Manager. It is therefore clear to me that the first is the strategy of choice.
There is no shortage of assets offering secure income, starting with Government bonds and infrastructure/real estate. However, many of them exhibit long duration and their capital values will fall if bond yields rise further. There is a good argument that this does not matter, as LGPS fund liabilities also have a long duration. However, because most do not use gilts as the basis of their discount rate (or if they do, their actuarial advisers have given themselves scope to fudge the numbers), the match is not exact. From a pure investment perspective, I want to generate more income, but I also prefer to keep duration short right now. That is another challenge for asset allocators.