Industry Voice: The conundrum of predicting future longevity

clock • 4 min read
Matthew Fletcher, Associate Partner, Aon

Matthew Fletcher, Associate Partner, Aon

The UK has seen high levels of mortality since the start of the Covid-19 pandemic in March 2020.

The Continuous Mortality Investigation (CMI) estimates that, between the start of the pandemic and the end of 2021, there were over 120,000 more deaths in the UK than their baseline, which is based on the death rates that were seen in 2019. For context, a typical year would see around 600,000 deaths in the UK, so on average, mortality in 2020 and 2021 was around 10% higher in the UK than might have previously been expected.

During 2022, the number of deaths caused directly by Covid-19 has been lower than in the previous two pandemic years. Indeed, for the first few months of 2022, deaths were roughly in line with the numbers seen in 2019, leading to hopes that mortality was returning to more typical levels. However, in around April of this year, the numbers of deaths started to increase significantly. The CMI now estimates that there have been around 25,000 more deaths in 2022 to date than their 2019 baseline.

This increased number of deaths was not widely expected, especially because death rates in the summer months are usually more predictable than in the winter. This has led to considerable analysis and investigation into the likely causes. Principal candidates are:

  • Covid-19 infection having a negative long-term impact - i.e. after individuals have apparently recovered - on health, in particular circulatory diseases
  • Knock-on impacts of the pandemic on the healthcare system, with data showing that the NHS is under severe pressure
  • Disruption to diagnoses and cancellation or delay of treatments during the pandemic period
  • Additional deaths arising from very high temperatures in the summer months

The extent to which these factors are expected to continue into the long-term remains a subject of intense speculation, however it is becoming clear that the short to medium-term outlook is less positive than might have been anticipated at the start of 2022.

Projecting life expectancy

In non-pandemic years, higher-than-expected numbers of deaths in national data would flow through into pension scheme funding assumptions because they would affect the mortality trends projected by the industry-standard annually updated CMI Mortality Projections Model. All else being equal, if recent years have higher-than-expected deaths then this leads to lower-than-expected life expectancies, and hence reductions in liabilities. But in the face of the extremely high mortality resulting from the pandemic, the CMI elected to (in effect) exclude data for 2020 and 2021 from its model to avoid distorting its mortality projections.

The next version of the CMI Model (CMI_2022) will be published in mid-2023. (This is later than normal because the CMI is awaiting final adjustments to historical population estimates from the ONS following the 2021 Census).  Our estimates suggest that simply incorporating these adjustments will reduce pension schemes' liabilities by around ½%.

The CMI faces a difficult decision on how to treat 2022 mortality. On the one hand, the CMI Model must start to incorporate new mortality data at some point, otherwise it will be seen to be out of date - and variations in mortality similar in size to those seen in 2022 have been incorporated in previous versions of the model. On the other, 2022 still saw relatively large numbers of Covid-19 related deaths, as well as elevated deaths from other causes.  While it is possible that this higher level of deaths represents a 'new normal', at least in the short term, it is not clear how long this higher level may last. Finally, the impact of 2022 data on the Model is even greater because 2020 and 2021 data have been omitted - in effect the model has been coasting based on pre-2020 trends and so including new higher-than-expected deaths data has the potential to pull it abruptly in a different direction.

Looking ahead to 2023

There is no single good option. The CMI Model plays a critical role for pension schemes and life insurers, and deviating from the CMI's core model can prompt difficult questions from auditors. At Aon, our view is that the most appropriate approach is for the next version of the CMI Model to place partial weight on 2022 data. This would have the effect of:

  • reducing life expectancies to reflect lower expectations following the pandemic, but not over-reacting to the higher-than-expected numbers of deaths in 2022, but
  • at the same time, signal that, although 2022 may not be a completely 'normal' year for mortality, the CMI's intention is to revert to placing full weight on future years' data.

For further information on longevity considerations for your scheme, and potential implications to 2023 insurer pricing, please speak to one of Aon's Demographic Horizons Team.


This post is funded by Aon

More on Defined Benefit

LCP predicts further new entrant to risk reduction market this year

LCP predicts further new entrant to risk reduction market this year

Consultancy forecasts the total number of insurers will rise to 11 by the end of the year

Martin Richmond
clock 14 June 2024 • 3 min read
LGPS in a 'new era' of openness and transparency

LGPS in a 'new era' of openness and transparency

Industry experts say dialogue between employers and LGPS funds needs to be open

Holly Roach
clock 13 June 2024 • 2 min read
LGPS funds encouraged to adopt PLSA's Retirement Living Standards

LGPS funds encouraged to adopt PLSA's Retirement Living Standards

Industry experts say the standards will be ‘prominent’ through a members’ journey

Holly Roach
clock 13 June 2024 • 2 min read