Last week, I started to look at key issues affecting the choice of
flexible retirement income options by considering the impact of increasing
longevity on conventional annuities and their more flexible alternatives,
primarily, income drawdown and staggered vesting.
I noted that life expectancy for males in the 10 years up to 1990 had
improved by around three years for a male aged 60 (the improvement for
younger males being even higher) and around two years for females. This
increase broadly continued an improvement which was already evident over
the preceding decades.
Indications from insurance companies are that this increase has continued
since the last published mortality tables. What has all this to do with
Conventional annuities are costed on the basis of three main constituent
factors – underlying assumed interest rates (at which we will look closer
in future weeks), mortality assumptions and charges.
In very simple terms- in fact, a little oversimplified – if the insurance
company actuary assumes that an individual purchasing a lifetime annuity
with a fund of £100,000 is likely to live for five years after
purchase, then the mortality element within the annuity rate quoted will be
£20,000, returning to the annuitant the capital cost of the annuity
over his expected lifetime.
If, on the other hand, the annuitant's life expectancy is 10 years, the
mortality element within the annuity will be only £10,000. If his life
expectancy is 20 years, the mortality element will be £5,000.
These are, of course, exaggerated examples which show clearly that the
life expectancy assumed by a conventional annuity provider can and does
have a significant effect on annuity rates.
More realistic than these simplified examples, the improvement in life
expectancy over each recent decade has, for males, resulted in a reduction
of a little under 10 per cent in the annuity rate, broadly equivalent to a
1 per cent reduction in underlying interest rate assumptions.
If all other factors within the annuity calculation remained equal,
annuity rates would have fallen by a little short of 20 per cent over the
20-year period up to 2000. If this trend of increasing longevity continues,
as is widely expected, then a further reduction of up to 20 per cent might
be expected over the next couple of decades.
Perhaps the key message here is the risk warnings which should or must be
given to clients who are being recommended drawdown or staggered vesting.
Both these flexible retirement income options involve the client in
deferral of annuity purchase until typically 75.
If life expectancy continues to improve during this period of deferral,
then annuity rates can be expected to continue to fall, irrespective of
what happens to underlying interest rates. This fall could be significant.
If life expectancy continues to improve at the same rate as in recent
decades, then deferring buying an annuity for 15 years from retirement at
60 to 75 might lead to anything up to a 15 per cent reduction in annuity
But it is my experience that few IFAs have included this risk warning in
It is vital that the potential impact of increasing life expectancy is
given just as much prominence in the risk warning as the possible impact of
falling interest rates.
It should be noted that our text books advise us, or infer, that annuity
rates are based on historical mortality tables. But, increasingly, such an
assumption has been overtaken by actuarial consideration not of past life
expectancy (as shown by these mortality tables) but future life expectancy.
To use another simplified example, mortality tables compiled in respect of
experience up to 1990 might show life expectancy for a 60-year-old as
being, say, 24 years. Such expectancy does not show true expectancy as it
is based on historical data. If, at that time, continuing improvement in
life expectancy could have been predicted (which, in fact, it commonly was)
then we could envisage the following scenario.
Looking at the last published mortality tables, we might consider that a
60-year-old male can expect to live a further 24 years to age 84.
However, these tables are already out of date and more recent, although
unpublished, figures for this person indicate a life expectancy of 27 years.
Yet even this higher figure is based on past mortality experience and
takes no account of the general consensus of continuing improvement. Over
the next decade, if longevity continues to improve, then our client, at 70,
might now be advised that he is expected to live a further 19 years against
the 16 years indicated 10 years earlier for a 70-year-old.
Then, over the next decade and having reached the age of 80, he is advised
that he is expected to live a further 10 years against the eight years
indicated 10 years earlier for an 80-year-old and so on.
The above example is designed to show that an individual's predictive life
exp- ectancy is, if we assume continuing increases in longevity, much
higher than his life expectancy apparently indicated by mortality tables.
Annuity providers know this, of course, and to my certain knowledge some
of them have already based their annuity rates on the assumption of life
expectancy considerably higher than the PA90 tables, thereby factoring in
some allowance for improved longevity.
The problem remains that deferring buying an annuity is a dangerous
decision and, without doubt, one which should be taken by a client only
after having learned the impact on annuity rates of continuing improvements
to life expectancy relative to those already anticipated by annuity
This issue has already been highlighted in regulatory guidance notes over
the last few years but perhaps could or should have been brought more
strongly to our attention, perhaps including it as one of the major risk
factors in drawdown and staggered vesting recom-mendations, along with the
possibility of falling interest rates and poor investment performance.
If all this sounds more than a little negative as regards the attractions
of flexible retirement income options, then to some extent my apologies.
Readers who have attended any of my seminars will be aware that I am a keen
supporter of drawdown and staggered vesting (and, of course, combinations
of these) but we have to acknowledge and bring clients' attention to risks
in such recommendations.
If we do not, and the client suffers loss as a result of one of these
risks materialising, then we have on our hands a whole new pension transfer
review scenario as, indeed, has already been predicted in some quarters.
Now, with this key issue already discussed, what are the remaining
attractions of flexible retirement income options which can or could still
make them arecommended alternative to conventional annuity options?
Over the next few weeks, I do not want to cover issues which are already
well understood, such as flexibility of income and the level of death
benefits payable from such arrangements. Instead, I want to concentrate on
issues which are perhaps much less appreciated, starting next week with
issues relating to mortality gain, mortality drag and the nomination of
beneficiaries for death benefits.
Keith Popplewell is managing director of Professional Briefing