Stuart Cook published the demographic study of longevity and mortality rates of UK population based on the Continuous Mortality Investigation (CMI) longevity projection model. In the article, author discusses life expectancy predictions and pension scheme liabilities.
Author: Stuart Cook
Most readers will be well aware of the marked increase in death numbers in the UK since the turn of the decade, and the resulting impact on the life expectancies of pension scheme members. This has been accompanied by the topic's rapid rise up the agendas of trustees and scheme sponsors. In March this year, the UK actuarial profession completed its analysis of deaths over 2017 and released the latest version of its mortality projection model which confirms another fall in life expectancy. In this article, we examine recent death numbers and consider what the implications might be for pension schemes.
Death numbers and mortality rates in 2017
Data compiled by the Office for National Statistics (ONS) and used by the UK actuarial profession confirms 2017 was another exceptional year for death numbers. The data, which covers England and Wales, shows that 533,000 deaths were registered in 2017. The annual count exceeds the number registered in 2016 by around 9,000 and is higher than any other year in the preceding decade.
To give the figures some colour, the following chart shows how monthly death numbers in 2017 compared to 2016 as well as to the preceding 5 year average. The chart shows how the total reached in 2017 wasn't the result of an exceptional month – monthly deaths in 2017 exceeded the 5 year average nearly every month.
Of course, death numbers alone do not tell the whole story for life expectancy, particularly since the population of England and Wales has increased and aged over the same period. The latest figures also confirm that mortality rates in 2017 were almost identical to those in 2016. This reinforces the recent trend of stalling mortality rate improvements, with virtually no improvement having occurred since 2011. The five-year average mortality improvement is now close to zero, as illustrated by the chart below.
Release of CMI_2017 - Pension scheme life expectancy falls again
The CMI model is used by most pension schemes to predict future mortality improvements. It combines an extrapolation of recently observed mortality trends with a user's chosen long-term rate of improvement. The continuation of the recent trend means that the latest version of the model predicts that improvements in life expectancy over the coming years will be slower than had previously been expected. The result is a reduction in life expectancy at age 65 of 0.2 years or nearly 1% for males and females, relative to the CMI_2016 model prediction. As the following chart shows, the CMI_2017 model marks the fifth year in a row that male life expectancy at age 65 has reduced.
Implications for trustees/sponsors
Most trustees and scheme sponsors with an upcoming valuation are likely to consider adopting the CMI_2017 projections. This alone could lead to a significant reduction in the assessed value of liabilities when compared to an earlier version of the model used at the previous valuation. It could mean that schemes find themselves further along their funding or journey plan than they expected.
Depending on specific scheme circumstances and their views on likely developments, recent mortality data in the UK may lead some trustees and sponsors to consider changes to the CMI_2017 model parameters. For example, the smoothing parameter gives users the opportunity to adjust the responsiveness of the model to the recent data and can be dialled up or down to reflect whether users think recent mortality is a prolonged blip or a new trend that will continue. The long-term trend parameter could also be under scrutiny if recent data makes users reassess their assumption for long term future improvements.
Insurance companies, who have been reluctant to immediately recognise recent experience, now seem to be adopting the latest CMI models. This should potentially result in lower pricing for buy-ins/buyouts and longevity hedging, all else being equal.
On the scheme sponsor side, accounting assumptions are also likely to be affected, meaning an improved balance sheet and profit and loss for the following year.
Heavy mortality continues into 2018
An elevated level of deaths has continued into 2018 - the number of deaths to March already exceeds the preceding five-year average by 20,000. If this level continues, the 2018 iteration of the CMI model, due next Spring, will see further reductions in life expectancy and pension scheme liabilities. It will be interesting to see how 2018 evolves and trustees/sponsors should certainly be monitoring this closely with mortality likely to remain high on the agenda for the foreseeable future.
