The Continuing Mortality Investigation (CMI) ignored mortality data for 2020 and 2021, when the pandemic began, and said data from these two years had no significance in predicting life expectancy and mortality. .
CMI has released a consultation paper on proposals for CMI_2023, the next annual update of the CMI mortality prediction model, scheduled for release in April 2024.
If this proposal is adopted, CMI_2023 will produce life expectancy for the cohort at age 65 that is very similar to the previous version of the CMI model, CMI_2022. This is about a week shorter for men and about a week longer for women.
The CMI model is used by UK pension schemes and insurance companies where future mortality assumptions need to be made. While the mortality experience in 2020 and 2021 will impact actuarial calculations, mortality rates in both years are exceptional and unlikely to be indicative of future mortality rates, CMI said.
“For this reason, CMI de-emphasizes the 2020 and 2021 data for the core version of the model, effectively ignoring those years,” it added.
Kobus Daniil, Chairman of the CMI Mortality Forecasting Committee, said: “Although mortality rates from 2022 onwards are less volatile, mortality rates are still higher than expected before the pandemic, and this is due to some It can be plausibly interpreted in this way.” Is our view of the pandemic too optimistic? Are there short-term effects remaining that will quickly disappear as we return to pre-pandemic trends, or has the pandemic fundamentally affected long-term trends?”
He noted that the CMI “aimed to strike a balance between different views” with its proposal for CMI_2023, which would result in similar life expectancies as CMI_2022.
“We recognize that the mortality outlook remains uncertain and there is a wide range of reasonable forecasting methods and expected mortality rates,” Daniil added.
He said the CMI was consulting on proposals to gauge the views of model users, adding: “Given the uncertainty, we encourage and expect many users to modify their own portfolios and model parameters and methodologies to reflect their views on the impact of the pandemic.” “
Recent mortality rates are less volatile and are likely to be somewhat indicative of future mortality rates. The standardized mortality rate based on registered deaths in England and Wales in 2023 was similar to the 2015-2019 average and lower than in 2020, 2021 and 2022.
In light of this, CMI proposes to give 10% weight to data from 2022 and 2023 when adjusting CMI_2023, while giving no weight to data from 2020 and 2021.
“Weights do not linearly affect forecasts, so if you put a 10% weight on data for 2022 and 2023, you will have two forecasts with no weight on those years and one with full weight on these years. The forecast is roughly in between,” CMI said. said.