Piero Cocevar, a member of the CMI Annuities Committee, reflects on the key findings from the Committee’s analysis of the mortality experience of pension annuities in payment in 2021.
In Working Paper 172 (available to subscribers) the CMI Annuities Committee has published its annual analysis of the mortality experience of pension annuities in payment data covering 2021 (including comparisons with 2016-2019 and 2020).
This year, for the first time, we have included the experience of enhanced annuities in our annual update; previously, we considered enhanced annuities in occasional standalone analyses. Pension buy-in data has also been included in the dataset for the first time, as part of the ‘Bulk annuities’ product type. As a result, our dataset has grown. Overall, the data volumes for 2021 increased to a sizeable 4.6 million life years of exposure, which is the largest annuity dataset we have ever analysed. It’s been very pleasing to see data volumes steadily increasing over the last few years, along with the proportion of life offices contributing to the investigation. This is testament to the collaborative efforts of everyone involved – from data contributors to the CMI Secretariat.
Our annual experience updates on the CMI’s annuities dataset usually cover a four-year period, moving on one year each time. However, with the Covid-19 pandemic we expected pension annuity experience in 2020 and 2021 to be significantly different from that of the previous years, as we have seen with population mortality. As a result, this year we have focused on the experience in calendar year 2021 on its own, with calendar years 2016-2019 and 2020 shown for comparison.
We compared actual deaths in the annuity dataset by age and gender with those expected using the ‘16’ Series pension annuity in-payment tables, adjusted for improvements in general population mortality experience using the Core version of CMI_2021, in particular, including a weight of 0% for both 2020 and 2021. We note that using different parameters for CMI_2021 would have produced different results.
The all-age, all-duration experience for males in the 2021 dataset (excluding enhanced annuities) is a little higher on both a lives and an amounts basis than the projected ‘16’ Series tables, while for females they are very close to 100% on an amounts basis and lower than 100% on a lives basis. This suggests that the impact of the Covid-19 pandemic has not resulted in higher-than-expected deaths for females in 2021, but that there is still some increase in mortality for males. It’s interesting to see that although we observed a peak in deaths in early 2021 in our previous analysis of annuitant mortality experience (Working Paper 161), the impact is fairly modest when looking at the year as a whole.
In contrast, the results for 2020 for both genders are considerably higher than the projected ‘16’ Series tables, particularly for males, with the increase in mortality predominantly likely to reflect the impact of the Covid-19 pandemic. The experience of 2016-2019 was generally very similar to the projected ‘16’ Series tables.
For both genders, the 2021 experience decreases with increasing age. We have seen similar patterns in previous years; in part, this may reflect the methodology used for the older ages in the graduated tables.
When we look at the experience in 2021, there are some differences for the various types of individual annuities. For Individual Internal annuities, the experience for males was markedly higher than the corresponding ‘16’ Series table, whereas it was similar to the table for females (or slightly lower on an amounts basis). For Individual External annuities, the experience was higher than the corresponding tables for both genders on a lives and an amounts basis.
For enhanced annuities, the overall mortality experience in 2021 is considerably higher than for Individual Internal and Individual External annuities on both a lives and amounts basis. The mortality experience is similar to the 2011–2019 enhanced annuities dataset analysed in Working Paper 155 for males on an amounts basis but heavier on a lives basis, while for females the experience is lighter than the 2011–2019 dataset on both a lives and an amounts basis.
For bulk annuity business the all-age, all-duration experience in 2021 on a lives-weighted basis is generally higher than that of Individual annuities. It is important to note that the experience is a function of the particular schemes that are in the dataset and could be influenced by particularly large schemes with heavy or light mortality.
Our analysis shows that the fall in experience between 2015 and 2019 has been greater than the improvements in CMI_2021, which broadly represent smoothed improvements within the population of England and Wales for females and (to a lesser extent) for males. The mortality experience then increases considerably in 2020 for both genders then falls again in 2021, with the male improvement lower than implied by CMI_2021 and the improvement for females higher than implied by CMI_2021.
We also carried out an analysis of standardised mortality rates, which approximately allow for changes to population structure and size over time. This analysis suggests that the average improvements over 2016–2019 in the annuities dataset are lower than those in the population for males but are higher for females. In 2020 and 2021 the mortality improvements are higher in the annuities dataset than the population for both males and females, but this finding should be considered carefully as there is some uncertainty on late reporting of deaths, particularly for 2021.
Looking further ahead, we hope to undertake an accelerated analysis of data for 2022 to investigate the extent to which excess deaths seen at population level in the latter half of 2022 were also present in the pension annuities dataset.
We’re also considering which other areas of analysis to explore over the next few years. If you have any areas you would like us to research further, please contact the committee at annuities@cmilimited.co.uk
