As many schemes prepare for their first valuations under the Pensions Regulator's new funding code, much of the attention has understandably been on low dependency targets, significant maturity, and reliability periods. When it comes to mortality, increasingly complex future improvement models have taken all the attention (eg the latest "CMI 2024" model). But amidst all this, the Regulator has issued a timely reminder: don't overlook the importance of base mortality.
They note that "scheme mortality is likely to be a material source of uncertainty, particularly for smaller schemes" and also "we expect many schemes will want to commission [postcode-based or experience] analysis and would generally expect all trustees to have considered doing so."
"We expect many schemes will want to commission [postcode-based or experience] analysis and would generally expect all trustees to have considered doing so."
The Pensions Regulator
Recent analysis by the actuarial profession's mortality arm, the Continuous Mortality Investigation (CMI), reminds us of the importance of tailoring mortality assumptions to a scheme's membership. Their analysis of the mortality experienced by individuals with a pension annuity over 2015-2022 calculated that, on average, male mortality rates could be twice as high in the most deprived areas compared with the least. This echoes similar findings for defined benefit pensioners and highlights the stark disparities in life expectancy across the UK.
It's an unfortunate truth that there is so much disparity in mortality in the UK and the underlying causes of this are complex. Our own analysis of the output from our Geo Focus model shows how neighbouring postcodes can have a life expectancy difference of four years, equivalent to a 15%-20% difference in the value of their pension liability. For example, in the Birmingham area, life expectancy at 65 in Solihull is 88.1 years, compared to just 84.3 years in nearby Castle Vale only ten miles away.
This shows how life expectancy can vary even within a small area. Similarly, pension amount is correlated with life expectancy, but analysis shows it only goes so far to explaining the difference in mortality outcomes between individuals, especially since many private sector scheme members won't have a full-service pension (because the scheme closed or they moved employment).
To properly understand the mortality expectations for a pension scheme, trustees and sponsors need to look at their own experience in combination with a good quality postcode-model that also takes into account lifestyle, health and wealth factors based on where people live (as our Geo Focus model does).
The Regulator is clear: mortality assumptions should be prudent, and greater margins for prudence are expected where schemes rely on standard tables without tailoring. They state, "where such a standard approach is taken, we would expect the uncertainty of experience to be reflected in a more prudent set of rates being chosen."
This is especially important for smaller schemes, where the concentration of risk in fewer members magnifies longevity risk and less credible data hampers establishing an appropriate scheme-specific assumption. Medium and large schemes on the other hand have the benefit of more data and less concentration risk but pose greater financial risk in absolute terms to their sponsors. Careful management of the risk is therefore beneficial in all circumstances. By doing so, schemes can properly prepare for the future, be it run-on or matching an insurers expectation when approach buy-in/out.
While the industry's focus has rightly been on improvements in future mortality rates through projection models, this is a timely reminder: base mortality is the bedrock of your mortality assumption. Get it wrong, and even the most sophisticated mortality projection model won't save you. Get it right, and you will have solid foundations for the next stage in the scheme journey.
