Accounting for mortality trends

By Brian Peters

The gerontologist Aubrey de Grey believes that the first man to live for 1,000 years is alive today.  This is based on a theory that technological advances in medical science will resolve and ultimately stop the ageing process that currently leads to a life span.   On the other hand, recent actual mortality experience is showing more deaths than previous models and data predicted.  This means previous estimates about how long people will live overestimated today's projections.  Or put another way, we will live longer than our parents, in general, but not as long as we perhaps originally thought.

It is of course not known how long people will actually live and this creates uncertainty when assessing the value of obligations coming from defined benefit schemes and, at an individual level, working out how much pension you can draw from a defined contribution pot each year without running out of money before you die.  Clearly you need a much bigger pot if you are the lucky person to live for 1,000 years!

Companies rely on advice from their actuary on the future rates of mortality for many purposes including making estimates for accounting liabilities under IAS 19.  Mortality estimates are usually in two parts.  

Firstly, actuaries look at how long the population covered by the pension scheme lived in the past.  PwC have created additional analysis to support this based on the postcodes of the pension scheme membership.  This has been shown to provide a more appropriate estimate of future longevity for a particular pension scheme as it takes into account changes in lifestyle and localised factors down to a regional level.  In many cases the results show that standard approaches using national statistics or just past experience, tend to overstated how long people are expected to live.  Using the more granular data, including postcodes, means liabilities are typically reduced.  It is very simple to run and the results can be enlightening.

The second component is what future improvements in life expectancy beyond those observed, should be anticipated.  PwC has recently concluded a study for one of the largest UK pension schemes looking at what could be the drivers of increased (or reduced) life expectancy in the future, from advances in medical science to the impact of driverless vehicles or wearable technology (and the combined effect of all three).  We are using this work with other schemes to inform their thinking.

There is clearly a range of assumptions Companies can make for accounting purposes and the impact can be material, for example, one extra year of life expectancy adds may be £30-40m of liabilities for a £1 billion pension scheme.  Therefore proper understanding of what has happened in the past, and could happen in the future is essential.