New official life expectancy assumptions and what they mean for the State Pension Age

Towers Watson's picture
Printer-friendly versionPrinter-friendly versionPDF versionPDF version

LONDON, Wednesday 11 December 2013 – Towers Watson’s analysis of life expectancy assumptions published by the Office for National Statistics (ONS) today indicates that:

Impact on State Pension Age

  • Plugging the new ONS life expectancy assumptions into the Government’s recently unveiled formula for increasing the State Pension Age points to it:
    • Reaching 68 in 2036 (starting to rise in 2034)
    • Reaching 69 in 2049 (starting to rise in 2047)
    • Reaching 70 in 2063 (starting to rise in 2061).

These increases come slightly later (around one year for the increases to 68 and 69) than would have been expected based on the previous set of ONS life expectancy assumptions.

  • The expected pace of increase in the State Pension Age is very sensitive to the precise way the Government has formulated its target that people reaching State Pension Age should expect to spend ‘up to one third of their adult lives’ in receipt of State Pensions. The State Pension Age would be expected to reach 68 in 2049 (when it is now expected to reach 69) if the target was instead for this proportion to be exactly one third of adult lives at all times.

Changes to ONS life expectancy assumptions

  • The ONS estimate of life expectancy has fallen for older men – from 4.8 to 4.4 years at age 90 (there is much less change for women).  For people aged 65 in 2013, life expectancy estimates have barely changed – the estimates of 21.4 for men and 24.0 for women are within one month of the previous estimates.
  • Compared with two years ago, many people are assumed to have a better (or at least similar) chance of living to age 90 but less chance of living to 100: 
    • Men aged 30 in 2013 are assumed to have a 49.9 per cent of living to 90 (up from 49.0 per cent) but an 18.1 per cent chance of living to 100 (down from 19.9 per cent)
    • Women aged 30 in 2013 are assumed to have a 60.3 per cent of living to 90 (up slightly from 60.2 per cent) but a 25.8 per cent chance of living to 100 (down from 26.4 per cent)
    • The proportion of baby boys born in 2013 projected to reach 100 has fallen from 32.2 per cent to 30.1 per cent, but the proportion expected to live to 90 has increased from 60.2 per cent to 61.0 per cent
    • The proportion of baby girls born in 2013 expected to live to 100 has fallen from 39.3 per cent to 38.7 per cent but the proportion expected to live to 90 is unchanged at 69.8 per cent.

Explaining the changes to the ONS’s assumptions, Matthew Fletcher, a senior consultant at Towers Watson, said: “Since the ONS last crunched these numbers, we’ve learnt two things.  First, mortality improvements were faster than expected in 2011 but came nowhere close to living up to expectations in 2012 and early 2013 - just as everyone was getting used to very fast year-on-year improvements in mortality rates, these have come to a halt.  Second, the census revealed that there were fewer people in their late 80s and older living in the UK than ONS had thought.  This pointed to worse mortality rates at the oldest ages between 2002 and 2010 because the same number of registered deaths is now thought to have occurred across a smaller population.

“There is a mixture of good news and bad news here but, on its own, this data would result in a bigger downwards revision to life expectancies than we have seen today.  The reason it has not is that the ONS appears to be assuming faster improvements to mortality rates over the next 20 years than it used to, including extremely fast improvements in late 2013 and in 2014. 

“Whereas the bad experience is particularly pronounced at the oldest ages, the more rosy assumptions cover a bigger age range.  As a result, people who are already in their 90s, or not too far off, have lower life expectancies under the new assumptions. For many younger people, the changes broadly cancel each other out.

“People have become used to thinking that life expectancy can only go up.  However, this measure of life expectancy anticipates that, due to improvements in medicine and more healthy behaviour (for example, less smoking), mortality rates will fall over time.  If you are already taking the credit for that, you don’t need things to get worse for predicted life expectancy to fall – you just need them to improve less quickly than had been expected.”

Plugging the new numbers into the Government’s State Pension Age review formula

Last week, the Government announced new details of how it envisages that regular reviews of the State Pension Age will work.  Its illustrations of when the State Pension Age might go up as a result were based on the mortality assumptions underpinning the “2010-based” ONS population projections, which were published in 2011. The table below shows how the results change if the same calculation is performed using today’s data, which reflects the mortality assumptions underpinning the 2012-based population projections.   

68 69 70

Autumn Statement (using old ONS mortality assumptions)

“mid-2030s” “late-2040s” Not stated
Towers Watson estimate using old ONS mortality assumptions Starts rising in 2033, reaches 68 in 2035 Starts rising in 2046, reaches 69 in 2048 Starts rising in 2059, reaches 70 in 2061
Towers Watson estimate using new ONS mortality assumptions Starts rising in 2034, reaches 68 in 2036 Starts rising in 2047, reaches 69 in 2049

Starts rising in 2061, reaches 70 in 2063

[Published projections only stretch to 2062 but this looks the likely result]

Matthew Fletcher said: “Putting the new ONS life expectancy assumptions into the Government’s formula points to a slightly slower increase in the State Pension Age than was signalled in the Autumn Statement.  This is not the final answer, though.  All eyes should be on the next set of ONS assumptions, due in two years’ time, which are the first ones that could affect State Pension Ages.

“No one can be very confident about how long people retiring in future will live, and the Government’s wish to give people plenty of notice of their final State Pension Age puts extra weight on the assumptions that are made. The ONS will have to be very clear about what it has assumed and why – for example, why it assumes faster improvements in mortality rates at the oldest ages than most pension schemes and insurance companies do.

“Ultimately, however, future State Pension Ages will be determined as much by pressures on the public finances and electoral considerations as by views about life expectancy.  If future Governments do not like the answers that the ONS life expectancy numbers produce, a small tweak to the target retirement length will quickly solve the problem.  To illustrate how easy this is, the State Pension Age is now expected to reach 68 in 2036 under the Government’s target that people should spend ‘up to one third’ of their adult lives in receipt of State Pensions.  If the target was one-third at all times, it would not reach 68 until 2049.”

The mechanism that the Government proposes to use to revise the State Pension Age is explained in the notes to editors.

Life expectancies

The ONS has changed its assumptions about the average age at which people will die as follows:

Age in 2013 Males – old assumption Males – new assumption Females – old assumption Females – new assumption
0 90.7 90.7 94.1 94.0
45 86.0 86.2 89.4 89.3
65 86.4 86.4 89.1 89.0
80 89.8 89.5 91.2 91.1
90 94.8 94.4 95.2 95.1

Average expected age at death is sometimes higher for older groups because none of today’s 80 year-olds can possibly die before reaching 80 whereas some of today’s 45 year-olds will do. This is consistent with 2013’s 45 year-olds living longer on average than people who were 45 in 1978.  

Matthew Fletcher said: “The further into the future you go, the more uncertain life expectancy becomes. So, although the ONS has published estimates of life expectancy for children born in the 2062, no one should put too much confidence in 150 years’ worth of projections about mortality rates. 

“Most actuaries assume that historically fast improvements to mortality rates will slow down but not stop. Views on life expectancy can change within these boundaries, even without potential gamechangers such as interventions that slow the ageing process or resistance to antibiotics.”

Notes to editors

The State Pension Age adjustment mechanism

The Pensions Bill currently before Parliament would raise the State Pension Age to 67 by 2028 and provide for regular reviews of the State Pension Age.  The intention is that a review will take place every five years and that people who will be affected by any changes will be given at least 10 years’ notice.

At each review, the Government of the day would set a target percentage of adult life for people reaching State Pension Age during a specified period to expect to spend in receipt of State Pensions.  The Government Actuary’s Department (GAD) would then advise on the changes to State Pension Age needed to bring this about, with a separate body reporting on other matters that the Government considers relevant to the review (for example, inequalities in life expectancy between social groups). 

The Government has said that the life expectancy assumptions used by GAD should be those published by the ONS, which underpin its population projections.  These assumptions are updated every two years.  The 2017 deadline for completing the first formal review of the State Pension Age has been set so that the next update (the assumptions for the 2014-based projections), which are due to be published by ONS in late 2015, can be used.  It is conceivable that the increase in State Pension Age to 68 resulting from this first review could be revised at the second review without violating the “10 years’ notice” principle (which is anyway not a legislative requirement).

The Autumn Statement said that the reviews would seek to ensure that people could spend ‘up to one third’ of their adult lives (which the Government defines as beginning at 20) in receipt of State Pensions.

A paper published by the DWP alongside the Autumn Statement clarified what this means.  State Pension Age would start rising two years before the percentage of adult life in receipt of State Pensions would otherwise reach 33.3 per cent.  The calculation will weight male and female life expectancy according to the numbers of men and women who are expected to be alive at State Pension Age.

This means that people would always expect to spend slightly less than one-third of their adult lives receiving State Pensions – the new estimates of how SPA might go up imply that this percentage would vary between about 33 per cent (immediately after an increase in SPA) and a little under 32 per cent (immediately after an increase). 

All life expectancies and mortality rates referred to here are for the UK as a whole but the ONS has used provisional data on recent mortality rates in Scotland.  The UK numbers could be revised slightly when the final Scottish data is available.

About Towers Watson

Towers Watson (NYSE, NASDAQ: TW) is a leading global professional services company that helps organisations improve performance through effective people, risk and financial management. With 14,000 associates around the world, we offer solutions in the areas of benefits, talent management, rewards, and risk and capital management.

News Source : New official life expectancy assumptions and what they mean for the State Pension Age

Copy this html code to your website/blog to embed this press release.