The Continuous Mortality Investigation has proposed to calibrate its next mortality projections model, CMI_2023, in a way that would produce cohort life expectancies at age 65 that are very similar to those in the previous model: around one week lower for men and one week higher for women.

The consultation closes on 11 March. The CMI aims to release CMI_2023 in April.

Cobus Daneel, chair of the CMI Mortality Projections Committee, said: “We have aimed to strike a balance between differing views with our proposals for CMI_2023 which would lead to similar life expectancies than in CMI_2022.”

More recent mortality has been less volatile and is more likely to be indicative of future mortality to some extent, said the organisation. 

Mortality in England and Wales in 2023 was 3% higher than in 2019, a period referred as the ‘pre-pandemic benchmark’, but it was lower than in 2020, 2021 and 2022.

Daneel said: “While mortality since 2022 has been less volatile, mortality rates are still higher than what we expected prior to the pandemic which could plausibly be interpreted in many ways: Were our pre-pandemic views too optimistic? Or are there just lingering short-term effects that will rapidly fade out as we revert to the pre-pandemic trend? Or did the pandemic have a fundamental impact on the longer-term trend?”

As a result, the CMI proposes placing 10% weight on data for 2022 and 2023 when calibrating CMI_2023 while still placing no weight on data for 2020 and 2021.

Daneel added: “Given the uncertainty [on the outlook of mortality], we encourage and expect many users to modify the model parameters or methods to reflect their own portfolios and their views of the impact of the pandemic.”

In June, the CMI placed a 25% weight on data for 2022 when calibrating CMI_2022.

Explaining the drop of weight from one model to the next, Stuart McDonald, deputy chair of the CMI, told mallowstreet the impact of this change on outputs from the models is “fairly small”, as placing 10% weight on two years is having a similar impact to placing 25% weight on one year. 

For example, he said Chart B shows that cohort life expectancy as at 1 January 2024 would be broadly the same in the proposed CMI_2023 as in CMI_2022: “Placing more weight on 2022 and 2023 would result in a further reduction in life expectancy but the majority of the CMI Mortality Projections Committee had a preference for the proposed weight, which results in little change in life expectancies since CMI_2022.”

The CMI said as weights do not affect projections linearly, placing a 10% weight on data for 2022 and 2023 results in a projection that falls nearly half-way between a projection that places no weight on these years and one that gives them full weight.

McDonald explained: “The weights have diminishing returns, so increasing weights from 0% to 10% has a much bigger impact than increasing from 10% to 20% and so on. This was also true for CMI_2022, where placing 25% weight on 2022 similarly resulted in outputs from CMI_2022 that were broadly halfway between 0% and 100% weight.”

Commenting on the CMI’s consultation, consultancy LCP explained that the committee’s decision on what to propose in the core model was split 4 votes to 3, with a narrow majority preferring to assume life expectancies remain broadly unchanged rather than reducing from last year’s model.

Chris Tavener, partner at LCP, said: “Projecting life expectancies is particularly tricky at the moment given the ramifications of the Covid-19 pandemic and the ongoing pressures on the NHS. It is, therefore, not surprising to see that the committee was split on whether life expectancies should remain unchanged or fall.”

The CMI_2023 consultation is detailed in Working Paper 183. Alongside it, the CMI has published a companion paper - Working Paper 184 - which considers the model's parameters more broadly.

The CMI intends to hold a webinar to cover this consultation and other aspects of its recent work on 26 February.