LGPS funding – are actuarial models surplus to requirements? 

Whilst funding levels have improved for many Local Government Pension Scheme (LGPS) Funds and their constituent employers, this is not the case for all.

The extent of the reported improvement also, however, varies by around a third of the total Funds assets. That is material by anyone’s measure. The key question on the minds of Funds and employers alike is – quite rightly - how much of the improvement is “real” and sustainable. Or putting it another way, how much is down to the underlying economic modelling now being off kilter due to a changing world? By this I’m talking about not just expectations of future risk-free returns (i.e. gilts) and whether other assets can outperform them, but as crucially – by how much do we expect those other assets to outperform them. After all growth assets comprise around 70% of the total assets for LGPS funds in England and Wales.

For LGPS Funds, the risk is two-fold:

1. A model that suggests materially higher investment returns can be delivered in the future than is actually the case, could lead to funding projections that are too optimistic and underestimate the contributions needed. This is a problem even where there is surplus, as it can lead to returning surplus assets to employers only to have to ask for them back at the next valuation, perhaps when the employer has come to rely on lower contributions. Of course, Funds apply prudence in taking decisions around contributions to help to manage potential volatility and uncertainty. And if uncertainty increases, then so does the need for prudence.

2. However, the flip side is that being overly pessimistic will lead to additional budget pressures for employers when they can least afford it. For the UK taxpayer ultimately, it puts more pressure on current generations for the benefit of future generations.

Therefore – as has always been the case, whilst model output provides a starting point, it is just that, the next stage is to overlay this with judgement to form a view on the plausibility of the model output and so the reliance that can be placed on it. This assessment needs to be taken together with the implications of varying the contribution levels on affordability now and the ability of employers to withstand step changes in contributions in the future. This is not new and Fund actuaries are experienced in guiding Officers and Committees through this decision-making process.

What is new this time, is the scale of the change in the economic environment over the short period since 2022. This has placed into question by many, the usefulness of existing models calibrated using historical information on the risk and return relationships between different assets classes. As highlighted by the Institute and Faculty of Actuaries (IFoA) paper “Climate Scorpion – the sting is in the tail” in the context of climate change analysis, but is equally applicable to LGPS valuations:

Too much complexity can lead to over confidence in the model with the model over-riding expert judgement. Even just a lack of understanding of the model, can hinder effective judgement. Transparency and understanding are crucial. The key decision makers – who are not actuaries – need to be able to understand not just what they need to pay now and the impact on budgets/local services, but what are risks and potential future financial impacts if a greater or lesser amount is paid, so they can determine their priorities for spending. After all we’re talking about billions of pounds of taxpayers’ money. Now is the time, whilst the preparations are underway for the 2025 valuations, for Funds to check-in with the actuary to ensure you are getting exactly what you need.

In addition to the model itself, there is also the calibration of the model, using past and present data - and it takes judgement to determine how much weight to put on these. In a changing environment, the weightings and judgments around these parameters can materially influence model outputs.  E.g. if there has been a paradigm shift, then over reliance on historical “evidence” to determine parameter inputs could be disastrous.

A good example of this being put into practice is in relation to the Continuous Mortality Investigation’s (CMI) use of new weighting parameters within their projections of future mortality experience. Tragically, the Covid pandemic had a material impact on death rates around the world including the UK. The CMI used judgement to exclude years where death rates from Covid were at a peak, deciding that the pandemic itself did not represent a paradigm shift and this experience was not representative for the model. Since then, they have gradually introduced more recent death data from 2022 to 2024 in the projection models.  Of course, in the current economic environment, in the context of modelling future expected returns, perhaps the reverse is true, and the risk is instead of placing too much reliance on past data to inform model inputs and too little focus on the present – but time will tell.

This phrase is often attributed to George E. P. Box, a British statistician. Or alternatively, as a great philosopher once put it: "Life is like a box of chocolates, you never know what you’re going to get"¹

So, what is the point of models?

Well firstly, and critically, without any model, we are just guessing – which is a worse state of affairs?  In addition, models provide objectivity, with so much pressure on employers for financing elsewhere, the importance of objectivity in determining LGPS funding needs cannot be understated. They also provide for a means to scenario test, therefore are an important tool for effective risk management, provided they are used with care.Through that analysis and scenario testing comes increased understanding. Although this is an increase in understanding of the model as opposed to the real world, I hear you say back to me! That is indeed true, but it gets people talking, thinking, challenging, and exploring which is ultimately what is important for delivering robust outcomes.

An analogy that most people can identify with, is that if you are going out for a long walk, you check the weather forecast – even though you know it’s not always right. If the weather forecast says it will be warm and sunny with a 70% chance it won’t rain, and you look outside and it is overcast and starting to drizzle, you take your coat (if you don’t want to get cold and wet). You overlay the weather forecast with judgment after looking out of the window at the real world – and take additional precautions, depending on what is important to you (perhaps keeping warm and dry is more important than the additional burden of carrying a coat).

So, as we head into the discussions and debate for the 2025 actuarial valuations, Funds should still check the “weather forecast”. And importantly, you should ensure it is delivered in a way you and your employers can understand so that you can apply judgement depending on your priorities. Even if the outlook appears rosier than before, you may still wish to take a coat with you on your journey through to the 2028 valuation - depending on the consequences for your finances of getting cold and wet in the interim.  If you don’t want to catch a cold, that is.

At Mercer we provide all the actuarial technical expertise and technological support you would expect in providing services to the LGPS. However it is important that this bedrock to our advice is implemented via a transparent framework that allows clear decision making and robust risk control for Funds when determining funding outcomes. The benefit to our clients is it allows LGPS Funds full control over their funding strategy whilst also being able to communicate decisions simply and clearly to stakeholders. This is even more important in the current climate of improving funding levels and budget uncertainty for employers.

Footnote

1. Quote from the film Forest Gump.