Life insurers in the UK have taken on pension liabilities worth hundreds of billions of pounds across hundreds of thousands of lives in the past decade or so, and therefore, any increase in life expectancy (i.e., a decrease in mortality rates) would, other things being equal, increase the aggregate payouts over time.
At the end of June, the Continuous Mortality Investigation (CMI) in the UK, which produces mortality data and analysis, issued its latest update, CMI_2024, which indeed showed an increase in cohort life expectancies at age 65 in England and Wales; men gained approximately three months (1.3%), and women, two weeks (0.1%).
While there are sub-trends within the new data, life insurers tend to see the CMI data as a useful benchmark, as opposed to a bible.
“The headline change in life expectancy between CMI_2023 and CMI_2024 is an increase of three months for males and two weeks for females at age 65. All else being equal, an insurer using the core CMI model could see an increase in liabilities of around 1% moving from CMI_2023 to CMI_2024, albeit this would vary depending on the maturity of their liabilities,” said Stuart McDonald, Partner at Lane Clark & Peacock.
“Importantly though, many insurers and reinsurers form their view on life expectancy independently, and the majority of insurers and reinsurers who responded to the CMI_2024 consultation indicated that their own view on life expectancy was higher than proposed in the consultation. So, in practice there may be limited impact on insurer liabilities.”
Still, the CMI data is certainly a foundational pillar in the insurance and pensions actuarial world. It is deeply embedded in Solvency II valuation frameworks for insurers, funding valuations for pension schemes, pricing models for annuities and protection products, and stress/scenario testing, for example.
This latest iteration is considered to be the result of the most significant changes to the model – announced in a consultation request in February – in many years. Specifically, the structure of the CMI Model now allows for different mortality trends at young, middle, and old ages to better reflect recent experience, and includes an explicit mortality shock in 2020 – the height of the Covid-19 pandemic – with the impact reducing in each successive year. Previous iterations of the model disregarded data from 2020 and 2021 entirely.
The consultation received plenty of feedback from members to help shape how the latest edition looks.
“The CMI had good engagement with its consultation on the proposed changes, with 27 respondents including insurers, reinsurers and actuarial consultants advising pension schemes. Consequently, the model now better reflects recent real-world data such as the impact of the Covid-19 pandemic and differences in mortality trends seen between groups. For example, mortality has been improving more rapidly for pensioners than for people of working age,” said McDonald.
As McDonald states above, CMI_2024 provides good news for pensioners and less encouraging signs for those in the age cohorts below. But the life insurer’s back book is weighted toward retirees – the last vestiges of the defined benefit pension industry, which began to decline in the late 1990s and early 2000s – of which the majority are men.
That means that, despite what seems, at just 1.3%, a small increase in life expectancy for males, the back book is more exposed to this cohort from a liability perspective.
Not quite.
“Life insurers hold significant capital buffers and risk margin against the risk of increases to life expectancies. The increase in life expectancy between CMI_2023 and CMI_2024 is very small relative to these buffers. The increase is also small relative to the falls in life expectancy seen over the last decade,” said McDonald.
Still, an increase in liabilities, even an expected one, is an increase in liabilities. And it is not only the life insurers that use the CMI model as a baseline for their own – the pensions trustees do so as well.
In recent months, there has been talk in the pensions industry about the benefits to the corporate sponsor of running on, as opposed to winding up, the scheme via a bulk purchase annuity buy-out. Lower costs (in the short term), investment flexibility and the ability to generate a funding surplus – which could benefit members (via discretionary increases) or even be refunded to the sponsor – all have their own appeal to trustees.
But the news that mortality rates are down, particularly in the older cohorts, might provide additional support for increased activity in the longevity swap market.
“Trustees and corporates have become accustomed to seeing successive CMI models reduce life expectancies. For the first time, the new model will lead to a significant increase in liabilities for many schemes if they choose to use the core model “out-of-the-box”,” said McDonald.
“This increase in life expectancies may serve as a reminder that longevity, which is often the largest unhedged risk that schemes face, is not a one-way bet. This potentially increases the appeal of longevity hedging for schemes pursuing a run-on strategy.”
The CMI Mortality Projections Model was introduced in 2009 to replace previous projections and has been updated on a broadly annual basis since then. There was a pronounced steady fall in mortality until 2011, but falls were more modest from 2011 to 2019. The increase in mortality from 2019 to 2020 was exceptional, with mortality in 2020 returning to levels previously seen in 2008 before falling again between 2020 and 2024, returning to levels similar to the previous record low mortality observed in 2019.
So, record low mortality means, in a nutshell, that pricing for DB pensions looking at a buy-in or a buy-out will go back up, right?
It is not nearly that straightforward.
“The impact of the update to the CMI model on PRT pricing is nuanced as it depends on the extent to which both the pension scheme and the insurer reflect the change in life expectancies,” said McDonald.
“PRT pricing may now appear better value to pension schemes if the CMI model change has a larger impact on their view than it does on life insurer pricing, for example.”