Continuous Mortality Investigation

This note provides an update on our progress with the review of the CMI Mortality Projections Model, our proposals for the next version, CMI_2024, and the likely timing of next steps. We previously noted that we had identified two improvements that we intend to make to the Model: 1) Fitting to the pattern of mortality during the pandemic more accurately – by adding an “overlay” to the existing model to explicitly reflect the rise in mortality in 2020 and subsequent fall. 2) Using “multiple period terms” to more accurately reflecting changing mortality trends at different ages – for example, mortality in 2024 was lower than in 2010 at middle and older ages, but higher than in 2010 at younger ages.   Our December update noted that making those changes in isolation would have led to concerns about the shape of the fitted cohort component of mortality improvements and how this is affected by cohort constraints.   Since then, we’ve conducted significant further investigations and identified a solution which addresses these concerns.   The solution requires an amendment to the way in which cohort constraints are applied, and improvements to the algorithm used to fit the APCI model.     Together, these changes represent a significant overhaul of the Model. Given the scale of the changes, we now aim to issue the CMI_2024 consultation in the second half of February 2025. We intend to publish CMI_2024 as soon as we are able in the second quarter of 2025, with the timing depending on responses to the consultation.

Record low death rates in 2024, says CMI The Continuous Mortality Investigation (CMI) publishes frequent mortality analysis for England & Wales through its mortality monitors. Today’s updates cover the period to 3 January 2025 and in particular, the quarterly monitor sets out results for the whole of 2024. The key points are: -> In England & Wales, mortality for 2024 was nearly 4% lower than in 2023 and slightly below the previous record low seen in 2019.  -> Death rates fell in all age categories between 2023 and 2024. -> However, mortality compared to the 2014-2023 average varies significantly by age. -> Mortality in 2024 for ages 20-44 was 2% above the ten-year average, which includes the peak pandemic period, while mortality for ages 45-64 was 2% below the ten-year average and mortality for ages 65-100 was 6% below.  -> There were around 10,000 deaths involving COVID-19 registered in 2024, compared to around 17,000 in 2023 and 33,000 in 2022. -> Our analysis is based on death registrations. The impact of death certificate reform in 2024 is to increase the time taken to register deaths. This means that the numbers of deaths registered in 2024 may be around 0.25% lower than if the reforms had not been implemented. Cobus Daneel, Chair of the CMI Mortality Projections Committee, said: “2024 has had low mortality, with death rates below the previous record lows seen in 2019. However, we would typically expect mortality rates to fall over time, and mortality in 2024 has been higher than projected before the pandemic. Mortality trends have differed greatly by age. Remarkably, mortality for younger adults is still above the ten-year average. In contrast, mortality for ages 75-84 is 8% below.” All mortality monitors available from: https://lnkd.in/gepZwWn

The CMI Annuities Committee has a vacancy. This is an exciting opportunity to contribute and get involved in the work that the Committee has planned. Please get in touch if you are interested or if you would like to find out more. The deadline for applications is 12 February 2025. More details on the role itself, how to apply and who to contact if you'd like to discuss this further can be found on the IFoA Volunteer Vacancies page: https://lnkd.in/ezfPCkBS

The CMI has today published its weekly mortality monitor for week 50 of 2024. Please note that the ONS is publishing the next weekly deaths data on 15 January 2025, covering weeks 51 and 52 of 2024 and week 1 of 2025. We will therefore publish the monitors for these weeks, along with the Q4 2024 quarterly mortality monitor, on 15 January 2025. All mortality monitors available from: https://lnkd.in/gepZwWn

We have today published updated mortality monitor software and an accompanying user guide. This is a minor update to enable the software to work with an updated data format used by the ONS to publish weekly deaths data. We will publish a further update in early 2025 to work with weekly deaths data for 2025.

The CMI's Mortality Projections Committee has today published Working Paper 196: "The CMI Mortality Projections Model – Interim update". The paper contains a number of analyses that are intended to be helpful to users of the CMI Model, including: -> Analysis of mortality in 2020 to 2024 for the general population and in the CMI datasets. -> A high level summary of our plans for CMI_2024. -> Mortality improvements by Index of Multiple Deprivation (IMD). -> Mortality improvements in CMI datasets. -> An analysis of the IFoA cause of death dataset. The working paper is available here: https://lnkd.in/dTujJ2ep

The CMI has today published its weekly mortality monitor for week 49 of 2024. Please note: -> The ONS is publishing the next weekly deaths data on 8 January 2025, covering week 50. We will therefore publish the mortality monitor for week 50 on 8 January. We are then expecting data for weeks 51 and 52 of 2024 and week 1 of 2025 to be published on 15 January. -> The weekly deaths data for week 49 published today by the ONS uses an updated data format, and therefore the mortality monitor Subscriber software can no longer automatically update the data file. We are working on an update to the Subscriber software to allow it to work with the updated data format, and we are intending to publish this updated software prior to 8 January. All mortality monitors available from: https://lnkd.in/gepZwWn

The CMI Self-Administered Pension Scheme Committee has released Working Paper 195, “Mortality experience of pensioners for the period 2016-2023”. The working paper presents an analysis of the mortality experience of data covering the period 2016 to 2023. It is the latest in a series of such papers that are updated annually. The working paper is available here: https://lnkd.in/ewdjNpuc And you can find a summary of the paper in our blog, here: https://lnkd.in/eMsD2wcH

We have been reviewing the CMI Mortality Projections Model and had aimed to consult in November 2024 on proposals for the next version, CMI_2024.   This note summarises progress on the review and provides an update on timing.   We have identified two improvements that we would like to make to the Model: 1) Fitting to the pattern of mortality during the pandemic more accurately – by adding an “overlay” to the existing model to explicitly reflect the rise in mortality in 2020 and subsequent fall. 2) Using “multiple period terms” to more accurately reflecting changing mortality trends at different ages – for example, mortality to date in 2024 is lower than in 2010 at middle and older ages, but higher than in 2010 at younger ages.    However, making those changes would lead to concerns about the shape of the fitted cohort component of mortality improvements and how this is affected by cohort constraints.   We intend to consult on CMI_2024 at the end of January 2025. Our eventual proposal for CMI_2024 may or may not include the two improvements noted above, depending on whether our ongoing analysis can resolve those concerns. We intend to publish CMI_2024 as soon as we are able in the second quarter of 2025, with the timing depending on responses to the consultation.   We aim to publish the annual “interim update” working paper by 19 December 2024.

The CMI has today published its weekly mortality monitor for week 48 of 2024. All mortality monitors available from: https://lnkd.in/gepZwWn