What is SF3 data?
Every autumn DCLG publishes its annual haul of data for the 86 English and Welsh LGPS funds (known in the trade as SF3 data). There is no certainty that it captures everything nor that each fund applies consistent principles when providing data. One known flaw, for example, is inconsistency in capturing investment expenses: some funds do not include those incurred through deductions from collective vehicle NAVs.
However, it does provide a snapshot of data fund by fund and I always find some interesting nuggets contained there. Last year I focused on investment expenses and pointed out that, in contrast to the Government’s claims of £380m of savings from pooling, in fact reported investment expenses had also risen by £620m over the five years to March 2023. This article provides an update on investment costs, and also looks at the disappointing three year investment numbers.
Investment costs remain competitive
Reported aggregate investment fees rose by 4.5% in absolute terms to £1.83bn in the year to March 2024.. That will provoke confected indignation amongst the LGPS critics, but it reflects the growth in assets. In basis point terms the fee level fell from 49bps to 47bps.
That is substantially lower than the investment costs of largely internally run Canadian pension funds, where the range is about 55bps to 70bps. The comparison reflects both under-reporting (as above) and higher allocations in Canada to higher fee private asset classes. However, it is hard to avoid the implication that, while further consolidation and pooling may deliver benefits of scale, cost savings are unlikely to be one of them.
The range of individual funds’ reported investments costs was from 4bps to 123bps. At the lower end that is almost certainly because they have only included directly incurred costs and/or have under-reported. The upper end probably reflects higher allocations to more complex strategies. The 25th and 75th percentile costs are 32bps and 69bps –as it happens both London boroughs are of a similar size.
Performance varied greatly
Performance also shows considerable variation. I looked at three-year data net of investment costs after adjusting for contributions, pension payments and net transfers.
Over three years the median performance was 11%, or 3.6% annualised. The reported range was an annualised -1.8% to 7%. This compares with the average discount rate (as at March 2022) of 4.4%, and suggests that on average LGPS funds have lost a little ground relative to their ‘target’ over the past three years. In practice, of course, the valuation of liabilities has fallen and therefore funding levels have risen, in many cases substantially.
Underperformance vs index
As context, the FTSE World equity index in GBP delivered an annualised return of 9.5% during this period. No fund would, or indeed should have all its assets in equities because of the requirement to be appropriately diversified. But even if we assume that a fund had 50% in equities and the other 50% diversified to deliver a sober 4% (i.e., the return on cash+1%), the net return should have been above 6%.
Stock selection, and in particular the choice to invest actively and specifically to tilt towards sustainable ‘climate action’ type, play a large part in the shortfall. Almost every active manager has underperformed in the last three years, because of the extraordinary outperformance of a small number of U.S. ‘tech’ stocks.
Doing good has not (yet) resulted in doing well.
Rightly or wrongly, LGPS funds tend to invest actively for stewardship and climate action reasons. They do not wish to be locked into investments in companies whose business models are not in their view sustainable. Some funds have gone further and focused on climate transition i.e., companies whose activities are designed to support or accelerate climate transition.
The thesis is that by doing good they will do well. However, in practice it has cost them money in recent years, as these funds have performed poorly, relative to the market indices. Lower exposure to oil, mining and defence was a substantial negative in 2022 following the conflicts in Ukraine and the Middle East. Under-exposure to tech stocks – or in some cases the wrong tech stocks – has also had a very significant impact in 2024. Both these factors have impacted almost all actively managed funds.
But the real culprit is the nature of the universe of companies which climate transition funds deliberately invest in. They tend to be narrow universes, with clear tilts to growth and to small caps, both of which have underperformed over the past three years.
Lower valuations of future income streams as a result of higher bond yields have not helped, but this has affected the larger growth companies too. I suspect the underperformance of the climate transition universe has more to do with the stage in their lives which many companies within this universe are at: needing harder to come by finance and, comparatively more affected by lock-downs. There was probably a degree of over-valuation prior to 2022.
This is not to suggest that investing in climate change transition is a poor investment strategy. The thesis is for the long term, and the companies may well perform much better in the future. However, it is clear that the timing has been less than optimal.
Longer term performance
The LGPS has delivered an annualised return of around 7% (4.5% real) over ten years. This compares with a range of between 3% and 10% from other public sector pension funds around the world*. These numbers will depend on the risk appetite and chosen strategic asset allocation of each fund, so cannot be directly compared. But it shows that the longer term LGPS returns, despite the experience of the last three years, are not out of kilter with other similar funds.
With the Mansion House speech and Pension Investment Review looming, I hope that the Chancellor in particular will pay attention to the fact that, whatever criticisms may be levelled at the LGPS, the outcomes have not been bad.
* Source CPP 2023 Annual Report, data from 2013 to 2022.
Comments