CMI: New “16” Series term assurance mortality and accelerated critical illness tables

Life Insurance CMI

Chris Reynolds (of PartnerRe) and Stephen Courquin (of RGA), members of the CMI Assurances Committee, set out the background to the new term assurance tables issued by the Committee.

The CMI Assurances Committee reports regularly on the claims experience of policyholders with term assurances issued by UK life insurers. Our most recent analysis – covering the years up to 2019, plus indicative analysis for 2020 – was issued to CMI Subscribers alongside Working Paper 152. The Committee has also now finalised the “16” Series term assurance mortality and accelerated critical illness tables, which we discuss here. The proposed tables were released for consultation alongside Working Paper 150 and were finalised in Working Paper 154. The tables cover mortality (including terminal illness) and accelerated critical illness under non-rated term assurance policies, based on 2015-2018 data.

How do the “16” tables differ from previous tables?

The tables are based on a much larger dataset than the predecessor “08” Series tables (and, of course, they use more up-to-date data). As a result, they reflect aspects where the experience appears to have changed from that in the previous tables; in particular:

  • The mortality tables reflect significantly higher smoker differentials at older ages, particularly for males, and changes in the select period for smokers.
  • For accelerated critical illness, the select periods are generally shorter than in the AC08 tables – most notably for female non-smokers, where it is now one year instead of five.

In addition, the new accelerated critical illness rates reflect an apparent hump in female incidence rates around age 50. This feature was evident in the crude claim rates and given a plausible “real world” explanation – we believe it arises from the commencement of routine NHS breast cancer screening – we felt it should be reflected in the rates. This means that the rates progress less smoothly from one age to the next than is customary in CMI tables.

After issuing the proposed tables, we produced an analysis of critical illness experience by cause of claim (in Working Paper 151) that provided additional support for including this unusual feature.

Why did the Committee consult on the new tables and how did you do this?

The CMI customarily consults on new tables, which is an important part of the Quality Assurance process. Although we have a number of practitioners on the Assurances Committee, from both insurers and reinsurers, the consultation gives a broader group an opportunity to comment on the tables and whether they meet the needs of users.

For these tables, we included a number of specific questions in the working paper; for example to seek users’ views on whether we were right to adjust the female critical illness rates, as noted above. These questions were also available via an online survey. In addition, the Committee hosted a webinar on the new tables in June; this was especially helpful in generating discussion and a recording of the webinar is now available to Subscribers.

To conclude the consultation, Working Paper 154 was released in August 2021. This paper summarises the feedback received and sets out the Committee’s response.  The one aspect of the proposed rates that attracted feedback warranting further consideration was in relation to the variation in mortality experience by sum assured band – additional analysis has therefore been undertaken, which is also set out in this paper. Our conclusion is that there is no justification to change the proposed tables and the final “16” Series tables are unchanged from the proposed tables issued alongside Working Paper 150.

Why do the tables use data for 2015-2018?

The CMI often uses a four-year period for the data underlying tables. This is a long enough period to avoid the fluctuations we typically see in a single year’s results but short enough that we aren’t using older data, which may be less relevant in future. With regard to the specific choice of 2015-2018, there were a number of considerations, including having a stable group of data contributors in these years. 2019 data was not yet available for these tables, but our analysis in Working Paper 152 shows that the new tables provide a good fit to the new data for 2019.

How do you expect the tables to be used?

It’s for actuaries and companies to decide whether and how they use the tables. Inevitably, there may be a delay before the tables are used in practice as each insurer will want to ensure they understand the implications of using these tables before they embark on any changes to systems, etc. However, we would hope that the tables are soon used as a benchmark for experience analysis, financial reporting, and, possibly, for pricing.

It is important that the tables are not used blindly, though. An example of where we might expect actuaries to adjust the mortality tables is to reflect variations in experience by sum assured band – this has been a recurrent feature of recent analyses, with a strong gradient to experience from higher mortality at low sums assured to lower mortality at high sums assured.

Actuaries should also be mindful that we had limited volumes of claims at younger and older ages; here we have applied considerable judgment to extend the rates so that they cover a full age range; most notably to older ages for accelerated critical illness. 

Another important area for actuaries to consider is whether and how to reflect the impact of COVID-19 in using the tables. As discussed in an earlier blog, mortality experience in 2020 was higher during the first wave of the coronavirus pandemic, whereas accelerated critical illness experience appears to have been lower in 2020 than in the preceding years, particularly for females.

What is the Committee working on next?

The Committee has now turned its attention to reporting on the mortality experience of underwritten and non-underwritten whole of life assurances in recent years and we hope to release this in a working paper in October or November.