[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries
.

Mortality: Under the microscope

The Continuous Mortality Investigation (CMI) is the longest-established unit of the profession with a research role, having been established in the 1920s (see CMIR 1 for more details). The first object of the CMI’s constitution is “to carry out research into the mortality and morbidity experience of various classes of insured lives, annuitants and pensioners”.

The first continuous investigation was of life office mortality. Subsequently, in the 1970s, an investigation into permanent health insurance (income protection) experience was launched. This was followed by critical illness in the 1990s, and self-administered pension schemes in this decade. In this article we consider the recent and forthcoming work of the CMI.

Life office mortality
The scope of the analyses that this committee has been able to undertake is still curtailed by the nature of the data received, which is largely unchanged from the days when life offices submitted paper returns. Considerable progress has been made recently with the ‘per policy’ initiative, and substantial volumes of data have now been collected that will enable more accurate and more detailed analyses to be produced once the teething problems have been overcome.

The committee also oversees an investigation into the mortality experience of impaired assured lives. Results for the period 1995-2006 have recently been published in Working Paper (WP) 36. Due to falling data volumes, the CMI will no longer collect data in the current format. WP36 includes a consultation on possible options for re-launching this investigation which has strategic importance to the insurance industry, in demonstrating the need to underwrite, to charge additional premiums for impaired lives, and to help better understand trends in non-impaired mortality.

Critical illness
The Critical Illness Committee has been grappling with a mismatch arising from collecting data on claim settlements each year that does not correspond to the exposure in that year. This has necessitated original research to tackle an issue that does not appear to have been addressed elsewhere in actuarial work.

WP33, published in July 2008, further developed the analysis methodology and presented results for accelerated critical illness experience for 1999-2002. Similar results for years to 2005 have also been sent to members. These are the first results that we have calculated that properly match claims to exposure. However, they do so in terms of settled claims, not diagnosed claims, and so still need careful interpretation. Subsequently, the committee has focused its efforts on producing realistic claim rates. It hopes to be able to present these in a WP in Spring 2009.

Self-administered pension schemes
The CMI Self-Administered Pension Schemes Mortality Committee is the youngest of the CMI committees. Most recently it published the first graduations of pension scheme mortality — the ‘S1’ Series tables — in WP35 in October 2008. The committee is currently working on further experience analysis using the latest data. It will then seek to analyse mortality improvements within the dataset, and mortality experience by industry.

Mortality projections
Perhaps the hottest mortality topic currently is that of projecting future improvements. The CMI undertook extensive research into projection methodologies during the earlier part of this decade. This is extensively documented in a number of WPs but the CMI recognised that these were inaccessible to many actuaries, and so collated sample projections in its Library of Mortality Projections in 2007.

Notwithstanding the Library, concerns have remained that the Interim Cohort Projections, or variants of them, are still in near-universal use within the profession, despite these projections being based on experience ending in 1999. Consequently, the CMI has established a working party to examine the latest data on mortality improvements and to develop a generic model, able to produce a range of different projections. The intention is to develop a relatively simple, spreadsheet-based model capable of widespread application. At the time of writing, it is hoped to issue a paper for consultation, and an illustrative version of the model by March 2009.

Gordon Sharp is chairman of the CMI’s executive committee, and a director in KPMG’s pensions team