Chris Sier

.This article first appeared in IPE and is written by AgeWage co-founder Dr Chris Sier. Chris sent it to me to remind a friend of mine, instrumental in setting up LGPS pools of what best practice looks like. Last week at the PP awards, Border to Coast picked up several awards, they are a bright star that shines a light on how our future investment strategies, both DB and DC, can be improved.

Asset pooling (‘Pooling’) by Local Government Pension Scheme Funds, seen as a means of reducing the costs of asset management, was first formally suggested in 2015 by the then UK Chancellor of the Exchequer, George Osborne. Plans for each pool were finalised by early 2018 and eight pools were created. In the subsequent five years each of the Pools have absorbed at least some of the assets of most of their LGPS Fund clients but, despite this and in recent weeks, UK Ministers have criticised pooling for not moving fast enough. And each year Pools publish data that tries to prove claims of great success through huge fee reductions, but these are met with scepticism.

Which is a shame because my experience of the Pools with which ClearGlass has worked has been extremely positive. So, I wanted to set the record straight in this article, which relates to what I think is the first independent and quantitative study of the asset management costs of Pooling and, specifically, the Border to Coast Pension Partnership.

In 2022 Border to Coast invited ClearGlass to benchmark the asset management costs of the products it produces for its LGPS fund clients. This process involved Border to Coast giving CTI (‘Cost Transparency Initiative’) cost and performance templates to ClearGlass for each of its products. This is generally a simple process, but there were complexities involving the attribution of central Border to Coast costs to the underlying costs of external managers (that also had to submit CTI templates) to build aggregate templates. We got past this and built a bank of Border to Coast data with which we could make like-for-like comparisons to the large database ClearGlass holds on institutional fund cost and performance data.

A note on this cost and performance database: since ClearGlass launched in 2019 it has collected data on behalf of almost 1,000 UK DB pension schemes ranging in size from less than £10mn AUM to several over £50bn AUM. The total asset value of the schemes is (or was, prior to last year’s LDI crisis) well over £1 trillion AUM. Data came from over 530 different asset managers, in portfolios of all sizes, and data has been segmented into 44 pooled portfolio strategies and 12 segregated portfolio strategies. Each strategy has sufficient data to generate accurate scale curves across the full range of portfolio sizes, a total of over 30,000 portfolios. Data is granular (CTI templates have over 200 cost sub-categories) and each data point is always deal-specific. This is an important point as the CTI templates require client-specific deal data and are rejected if managers insert only share-class level data. In other words, rebates, indirect fees, implicit costs, fees paid through NAV as well as on invoice…all are captured and checked. Suffice it to say the database is huge, accurate and carefully curated to ensure like-for-like strategy matching. ClearGlass also works with non-UK clients, DC clients, and non-pension clients but these volumes are still quite low. So B2C’s data was compared to a UK dataset.

With this data we can do some interesting things. The first is to compare the data we received from Border to Coast, strategy by strategy, to our database. And the results were interesting. Every single strategy run by Border to Coast had costs that were better than the market median and therefore could be classified as Good Value-for-Money (VfM), with most also being better than the best-quartile and therefore classified as Excellent VfM. We’ve worked with over 1,000 schemes, including some LGPS funds, and haven’t encountered ubiquitous success like this before. But this is to be expected, yes? Scale counts for a lot in asset management and Border to Coast is huge, more than £40bn AUM, and the basic use-case of Pooling is to derive benefits from scale. But our benchmarks are scale adjusted, so by every measure Border to Coast is doing well.

It doesn’t stop there though. For every asset owner with which ClearGlass works, we can create a ‘unique fee benchmark’. To do this we do not look at scheme total size. Rarely are any two schemes of the same total size comparable as they have different asset allocations. In other words, any two schemes will use different strategies with different portfolio sizes, and as fees are dependent on the strategies used, the size of the portfolios, and whether the portfolio comes from a pooled fund or is a unique segregated structure, total costs will vary. To give a simple example, a scheme with £1bn AUM and 30% allocated to private markets will have much higher costs than a £1bn AUM scheme with a lower asset allocation to private markets. Instead, we look at the factors I mention above (strategies, portfolio sizes and seg vs pooled structures) and apply our benchmark fees to each portfolio and then weight each fee according to the asset allocation.

For Border to Coast the results were, frankly, incredible. Our data told us that, given the asset allocation, portfolio sizes…etc of Border to Coast, it should carry a total ongoing charge (equivalent to TER) of 42bps. We calculated instead that Border to Coast was only paying 16bps, a ‘saving’ of at least 26bps, or £121mn, per annum. That is £121 million that Border to Coast is saving its client LGPS funds each year, every year and in perpetuity. Compound that value up over 50 years or so and imagine what it does to the asset values of underlying clients. Deficit wiped out, to put it bluntly.

But what do I mean by ‘at least’? Well, the 42bps ‘unique fee benchmark’ we calculated was based upon portfolios of the size that Border to Coast can build, which are huge. How much would be saved if we created a benchmark fee level, with the same asset allocation, but based on the portfolio sizes that a typical LGPS fund would deploy? For this we scaled back portfolio sizes to roughly one twentieth, which is where a £2bn LGPS fund would operate. Under these circumstances, the benchmark fee level would have been 47bps. So in this version of events, Border to Coast has reduced fees by an even-more-incredible 31bps.

As it happens, we have data on one of Border to Coast’s clients, and we calculate that Border to Coast has actually reduced fees on the assets that have migrated from that fund to Border to Coast by a real figure of 25bps. Point made.

It doesn’t stop there though as ClearGlass has also built a Fee Savings Index, think of it as an index of scheme negotiating efficiency for the 1,000 client schemes with which it has worked, and Border to Coast is ranked at number 1. Border to Coast is the scheme in the UK that has achieved the largest reduction in fees, in relative terms, when compared to any other scheme with which ClearGlass has worked. The index also includes some schemes larger still than Border to Coast, so this is a truly remarkable achievement.

But this begs the question, how? How is Border to Coast achieving what I would describe as supra-normal levels of fee reduction? Basically, it is operating well beyond the limits of its scale and has not only forced external managers to deliver startlingly low fee levels, it has also built an internal management model that adds little to the costs of the fund-of-fund or direct fund management model it operates. To put it in the words of one external fund manager to whom I spoke recently “they really got a fantastic deal, almost unbelievable”.

I think the answer to the negotiating power exercised by Border to Coast comes in several parts.

Firstly, they started with a clean slate. They did not onboard managers and spend time rationalising lots of small portfolios. No, they started from scratch and went to market with notional AUM totals and, having won the deals with their external managers, they onboarded the AUM from client LGPS funds rather than onboarding the managers that were already being used.

Secondly, they built FOMO…Fear of Missing Out. From what I remember, managers were captivated by the potential of winning multi-billion pound portfolios to manage, and a negotiating war ensued that was related not just to the scale of the portfolios, but the idea that in winning they would exclude competitors.

Thirdly, and probably most importantly, governance and independence played a huge role. My experience of most Pools is one of consummate professionalism. Decisions are made based upon rational analysis and an independent decision-making framework, which means that managers are picked because they are good and able to demonstrate that ability.

Finally, and this applies to all institutional asset owners, you can negotiate with liquid asset classes anytime if you have good benchmarking data. But you can only negotiate with private markets managers up front. The threat of losing a mandate as a manager if you misbehave or are too expensive (again based on benchmark data) is not real in the world of illiquids.

In my opinion, therefore, asset pooling does work, but it works best if you have a clean slate, have excellent governance and are independent and rigorous. This is great news for LGPS Pooling, albeit there are questions around how to pool illiquids in any other way than starting from scratch (and there are far wider questions in my mind than just this point for illiquids, but that’s for another discussion). But it should also be great news for wider pooling initiatives in the private sector, such as the Pension Superfund and other aggregator mechanisms.

It might also mean a world where large schemes that either are, or claim to be, good at asset pooling, fee negotiating, and independent manager selection, having a role to play in the wider market. In other words, when will Border to Coast be open to clients other than LGPS because, based upon my analysis, they would do an excellent job