Last week, I looked at continuing improvements in life expectancy, the implications of which form what I perceive may be the severest risk with income drawdown recommendations – the probability that annuity rates will continue to fall over the next couple of decades, even if interest rates remain static.
I strongly believe that drawdown and staggered vesting – when properly understood and considered – should, in fact, account for a much greater share of the retirement income market than at present. Over the next few weeks, I would like to consider this view more closely.
Here, I would like to consider the impact of a concept termed by our regulators as mortality gain, though widening its meaning perhaps beyond that originally envisaged.
I am certain that every retirement income adviser has heard of and understands mortality drag, broadly describing, for a drawdown investor, the absence of crosssubsidy from annuitants who die early to annuitants who live a long time.
This cross-subsidy is a major feature of conventional annuities and particularly benefits those who live longer than the average life expectancy assumed by the insurance company (these people, therefore, being the annuitants who would lose most, in this respect, from entering a drawdown contract).
The effect of mortality drag is to require a higher underlying investment return from the drawdown portfolio than would otherwise be the case – typically, between 1.5 and 2.5 per cent over the redemption yield on long-dated gilts, the traditional underpinning investment for conventional annuities, currently standing at around 5 per cent. This would mean a required investment yield typically somewhere between 7 and 8.5 per cent when charges are taken into account.
But is mortality drag a concept applicable to every drawdown investor?
The concept was first explained and quantified – in relation to drawdown – in a 1995 regulatory guidance note which showed that, to match a conventional annuity for a male aged 60, the effect of mortality drag was to require an investment return around 1.8 per cent over the redemption yield on long-dated gilts.
The target annuity rate used in that calculation was £10,300 at a time when gilt redemption yields were around 8.15 per cent. But, although the guidance note stated the fact that the target rate was derived from a single-life annuity, it did not at that stage go on to highlight the very different conclusions which could have (and still can be) arrived at if a joint-life annuity rate had been used.
This variation on the initial mortality drag theme is a good starting point for considering what might be termed as mortality gain. This is, in fact, a term used in PIA regulatory update 55 which stipulates that “advisers should understand the concept of mortality drag and mortality gain and how it impacts on the advice given”.
In my experience, doing straw polls among conference delegates over the last few years, very few advisers are aware of the impact of mortality gain on the advice given to drawdown clients and potential clients, so here goes.
Accepting the first mortality drag chart used by Fimbra six years ago, we can now consider how the mortality drag figure would have changed if the target annuity rate had represented a joint-life pension rather than single life. The target £10,300 annuity rate would, in these circumstances, have been some 18 per cent lower at around £8,450 a year and it is easy to see that the required investment yield to match this lower pension would have been much lower than that required to match the single-life target of £10,300.
In fact, comparing the target annuity rate of £8,450 a year – which may be expressed as 8.45 per cent of the annuity purchase price – it is only 0.3 per cent higher than the redemption yield on long-dated fixed-interest gilts at the time.
We can identify a huge reduction in mortality drag where the target annuity rate is taken on a joint-life rather than a single-life basis.
Although I have simply reworked the original mortality drag figures from six years ago, here it should be noted that similar comparisons may be made at any interest rate level, with the same conclusions.
Put simply, if 18 per cent of the fund is used to buy a survivor's pension under a conventional annuity, whereas the death benefits from a drawdown contract make no such reduction, then the drawdown investor who values those death benefits will be starting off with a “gain” of 18 per cent at the outset of this contract.
If we average this gain over the 15-year term of deferment from age 60 to 75, this equates to an annual gain in simplistic terms of 1.5 per cent, thereby reducing the critical yield by this amount although the compounding effect makes the actual numbers a little less simple than this.
Mortality gain, then, is particularly important to clients who would have selected a joint-life option had they bought a conventional annuity, whereas the severest impact of mortality drag is suffered by clients who would have purchased a single-life annuity.
I strongly suggest, therefore, that advisers ensure in their illustrations to potential drawdown clients that they use a target annuity rate appropriate to the shape of conventional annuity the client might have purchased. This practice will lead to much lower critical yields for many clients, indicating that a greater proportion of retirees are likely to be attracted to the drawdown concept.
This is the major part of what the PIA called mortality gain in guidance note 55. On this point alone, however, critics of the importance of mortality gain would point out that there has been no gain at all, simply a reduction in drag. This observation relates to the fact that, even using joint-life annuity rates as a target, the required investment rate of return, though lower than for a single-life annuity, is still higher than the gilt redemption yield and, therefore, some element of drag still exists.
While such criticism is fundamentally well-founded, it ignores the fact that a joint-life conventional annuity – even one which provides a 100 per cent surviving spouse's pension – does not provide equivalent death benefits to the drawdown contract. I would suggest that, when this is taken into account, mortality gain does, in fact, exist as the regulator's term suggests, and not only as a reduction in mortality gain.
Why? Consider this – if a pension scheme member buys a 100 per cent joint-life conventional annuity, what type of annuity is the survivor left with in the event of his death? What shape of annuity is left behind and at what rate has that annuity been calculated?
You should be able to identify that the survivor would be left with a single-life annuity but calculated at joint-life rates applicable to the survivor's and deceased's ages at the time the annuity was first purchased.
Expanding this question, what would the annuitant be left with in the event of the spouse's earlier death and what type of annuity would be left in the event of both dying?
Keith Popplewell is managing director of Professional Briefing