[Skip to content]

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

A longer life

T he Continuous Mortality Investigation Bureau (CMI) has monitored pensioner mortality for many years. It produces the standard mortality tables which are widely used by UK actuaries for valuing annuity liabilities in insurance companies, and pensioner liabilities in defined benefit pension schemes. According to the CMI mortality tables, life expectancy has been improving for a 60-year-old male so that in 1955 the life expectancy was 18 years, but had risen to 22 years by 1992.
In October 2002, the CMI released three projections of the standard 1992 mortality tables into the future the short-, medium-, and long-cohort projections. These bases project recent mortality improvements, by age and year, allowing for improvements over shorter or longer periods. According to the medium-cohort projection, a 60-year-old male in 2003 should expect to live for another 27 years.

Mortality improvements and the cohort effect
What has happened? Figure 1 below sets out smoothed mortality improvements (1qx,t/qx,t+1) for the male population at age 60 and age 70 over the period 196999.
The low historical mortality improvement seen in the earlier years of the graph led to the gradually increasing life expectancy from 1955 to 1992. By the mid-1980s, mortality improvements had significantly increased, and these improvements have persisted through the 1990s. The pattern of mortality improvement is particularly associated with year of birth, leading to the projections of enhanced longevity for the cohort born between around 1920 to 1940. The rationale for the new CMI tables is set out in a recent SIAS paper. There is now evidence that these levels of mortality improvement are starting to affect people at age 80 and above in accordance with a year-of-birth cohort effect. A fuller picture of the cohort effect and the possible impacts on mortality may be found in Richard Willets’s article ‘Mortality in the Next Millennium’ (www.sias.org.uk/papers/mortality.pdf).

Where will it all end?
Why are these huge improvements to mortality happening? There is currently no complete answer to this question, though the reduction in smoking from around 1970 is likely to have played a significant role in the improvement. How far can it continue? Well, the new CMI tables are based on the assumption that cohort improvements will continue and then decline until they are reabsorbed into earlier projections. According to the long-cohort projection, the cohort effect will finally die out in 2040 with noticeable improvements to the mortality of centenarians. However, even in 2040, a 60-year-old man is not expected to live to much over 90. The CMI’s projections allow for a gentle reduction in rates of mortality improvement from 2001, so the largest impact has already been seen.
A key question for the future is whether the mortality improvements affect life expectancy but leave lifespan, the maximum age experienced, unchanged. Much further improvement in mortality rates at older ages might lead to an increase in lifespan, for which there is currently little evidence, but much theoretical discussion.
One of the difficulties is that the size of the population aged over 90 in any one country has historically been small and often poorly recorded. It has therefore been difficult to obtain adequate credible data to review mortality rates at older ages. However, by aggregating populations across national boundaries and looking at international comparisons, a picture may be formed.
Demographers Jim Oeppen and James Vaupel have reviewed data across many countries in the Science article ‘Broken bounds to life expectancy’. The article shows a remarkably consistent pattern of improvements in life expectancy. There is very little evidence that we are nearing a natural cap on life expectancy in Japan significant mortality improvements are now being recorded at ages well above 80.

What will the financial impact be?
All those providing benefits for the remainder of people’s lives will feel the financial consequences of a longer life expectancy:
– governments which provide state pensions;
– private pension schemes with a defined benefit component;
– financial services companies through annuities and equity release; and
– (of course) those people who live longer than they had expected.
The impact is particularly noticeable in a low interest rate environment since the present value of the later years’ benefit becomes much more significant. For governments, the obvious solution is to reduce the liabilities in respect of old-age benefits, for example by raising the retirement age. For a 20-year-old today who may expect to live to 90, the prospect of retiring at 60 is now distant. It is likely to become more so, as the financial implications of the population’s working for less than half its life-span become apparent.

Pension schemes
UK defined benefit pension schemes have, since 1997, valued their liabilities with respect to the statutory minimum funding requirement (MFR), which is due to be replaced. The MFR mortality basis is life table PA90 rated down two years. The PA90 table is based on data gathered more than 30 years ago. To reflect the improvements in mortality suggested by the CMI medium-cohort projection, the mortality assumptions should be replaced by PA90 rated down 12 years.
Life offices
The cost of annuities is reviewed by the CMI in its SIAS paper. Switching from the 1992 projections to the medium-cohort increases the current cost of an annuity by around 5% for ages between 60 and 75. These increased costs will be reflected in increased provisions for annuity portfolios and lower levels of income for new annuitants. Increased longevity and guaranteed returns is a particularly onerous combination. The pricing of annuity business may change too, from the current basis of rating healthy lives on age and sex, to increased segmentation.

What lessons can be learnt?
For many years, mortality has been an undervalued risk as life expectancy improved gradually. Many of the key areas where mortality risks are paramount were less visible for example, unfunded public pension provision and defined benefit pension schemes, which were not explicitly accounted for. High interest rates also meant that, in any event, the impact was limited. This is no longer the case. Mortality at older ages is changing, contributing to a shifting demographic picture with ageing populations in developed economies and its impact is much more visible in a low interest rate environment.
Better research earlier might have led to greater awareness of the likely effects of increased longevity, and earlier management of the consequences. Much of the risk is being shifted to the individual as both governments and private companies reduce their exposure to longevity, and the market in longevity risk is concentrated in a small number of annuity providers. The greater the benefits promised at older ages, the more strenuous the financial guarantees. These are particularly significant in the European Union where many countries offer generous, though unfunded, defined benefit pensions.
Changing life expectancy is a phenomenon that has an impact across the whole of society. The challenges that emerge from it are not confined to the actuarial profession, and it is not solely from within the profession that solutions must be sought.