The process of converting retirement savings assets into an income, which has come to be known as decumulation, needs first to be placed in context. The decumulation phase of retirement savings is that period when an income is drawn from the investment fund as contrasted to the savings period when savings are being contributed.
It is perhaps not well understood that the total pension income achievable is heavily dependent upon investment performance in this post retirement period. In fact, under an investment strategy delivering uniform returns over the entire term of the scheme, the investment accrual post-retirement is greater than the pre-retirement result for most. The importance of post-retirement investment returns increases in relative terms if contribution savings are made later in the working lifetime, and obviously if the period in retirement, and funds permit, is longer.
The relative importance of contributions, investment returns and taxes is also merits bearing in mind. Contributions typically account for less than 10% of the total pensions finally paid with investment returns accounting for the 90% balance. The tax components are also important – under current UK tax concessions and deferrals, they account for 47.5% of the pensions finally payable.
This immediately raises two issues. First, George Osborne’s often repeated line at the time of introduction of “Freedom & Choice” to the effect: “It’s your money” when almost half of it is in fact our (tax) money. Even if all pensioners paid income tax at the highest current marginal tax rate of 45% (and they don’t), it is evident that there is a tax subsidy here. Second, if we move from the current EET arrangement to TEE, and savers maintain their present disposable income, then contributions will be smaller with the consequence that the size of pension pots will decline. The exemption from income taxes of the retirement income will not offset this decline. One incidental effect of all retirement income being tax exempt would be the effective end of the 25% tax-free lump sum. Indeed, the move to tax-exempt retirement income would remove one of the few barriers to the immediate consumption of all retirement savings.
Though more could be said about tax incentives, the subject of this article is decumulation. The effects of uncertainty and variability of investment returns appear not to be well understood. The first point is that the order in which investment returns are achieved matters. Low returns early in the decumulation process will restrict the amount that can be paid if total depletion is to be avoided. We illustrate this below (figure 1) with a sequence of random returns; the same returns are then reordered so that the losses are experienced first. The pension income drawn is 5% of the value of the pot in both cases. The outcomes are starkly different, as is shown in table 1.
The minimum pension is £2.00 in the random case and just £0.10 in the ordered. Though these pensions are in each case 5% of the value of the pot, the average pension received with random ordering is £3.69, and just £0.47 when ordered, and highly uncertain as is indicated by the volatility. The ordered sequence generates a pension income that is just 13% of the random; it is clear that luck plays a big part in individual pension outcomes.
Figure 1: Evolution of value of pension pot under ordering and randomness.
Considering, as is illustrated in figure 2, a fixed annual drawdown (of £5) brings out some further points. First, the ordered sequence exhausts its funding after just eight years, far less that average retirement period expectations.
Figure 2: Evolution of the residual pension pot under fixed amount drawdown.
The second point to note here is that random pot also ultimately exhausts its value, after 76 periods. This is not predominantly an effect of the return sample properties. The arithmetic mean of the sample was 4.85% over the period prior to exhaustion, and that small difference from the £5 annual withdrawal would take far longer than is evident to fully deplete the pot. This is an effect of the volatility of the process, which generates a drag on compound realised returns. With 10% volatility, as is used in this illustration, this is expected to be 0.5% p.a. If the volatility of the investment process were 20%, this would be expected to be 2% p.a. and have the effect of depleting the pot far faster – then it is expected to last just 30 years.
Costs, management charges and fees, are obviously important to the performance and viability of a drawdown pension, but the major issues there were covered in my Professional Pensions article “Why costs and fees matter.”
These calculations and illustrations draw out some of the principle technical issues faced by drawdown strategies. First, that investment returns are extremely important; holding cash with little or no investment yield is deficient in this regard and particularly so when this is held in the early years of retirement. Second, the variability of investment returns is important in terms of the evolution of the value of the investment pot even if no assets are liquidated in markets to produce income. The challenge is, self evidently, to select an investment strategy which generates sufficient returns but is not excessively risky.
The second part of this article will cover approaches to this problem.
 These figures are derived from a forty year term with annual contributions of 15% followed by 25 years in retirement drawing a two thirds final salary pension and a fixed investment return of 5%.
 These are random drawings from a normal distribution of 5% mean and 10% volatility. The sample stats are 4.72% mean and 9.1% volatility.
Con Keating is a genius – (Pension Plowman)
All very good, but how is the man/women in the street supposed to understand the above, even my eyes started to glaze over by the 2nd diagram!
There are plenty of blogs on here talking about this from an ordinary persons viewpoint. Con is an expert writing for experts
Not even sure it’s all very good, Henry.
I used a spreadsheet of my own to model pay over 40 years with 15% contributions and 5% investment returns, but I also included an arbitrary 3% for annual inflation.
I found that contributions made up more like 40% (than Con’s 10%) of the accumulated pensions pot after 40 years. I also had a final pot which wasn’t sufficient to buy a two-thirds final pay annuity using an arbitrary 20-for-1 valuation; it bought more like two thirds of two thirds, which would be less than half my final inflated pay. I agree the resulting pot could be used instead for retirement income drawdown, but I’d personally rather target a larger pot so I can also enjoy a tax-free lump sum and build in some margins for uncertainty over the expected number of years of retirement, market yields at particular points in time, etc.
So, what needs to change in the “genius” model?
Most actuaries of my acquaintance would say the 15% annual contribution was way too low, but on that particular assumption I’m quite prepared to side with Con. It should be sufficient.
His 5% fixed investment return, on the other hand, strikes me as insufficient, unrealistic and unnecessarily pessimistic. I don’t see how you’re able to invest in a series of bonds of longer and then shorter durations, which redeem from 40 years onwards and which are somehow not subject to fluctuations in market interest rates. Since investment grade credit can currently be issued in the UK for quite a lot less than 5%, this also suggests higher risk bonds than I’d trust for pensions.
Higher returns than 5% can be obtained by using prudent yield and growth assumptions and investment managers who can deliver/compound such yield and growth over the long term without permanent losses of capital. I believe such managers do exist, but others may not.
I don’t think Cons is a bond based investment model.
I think the 10% calculation allows for tax relief on the contributions, and also for investment returns on the funds after retirement commences. The key is “Contributions typically account for less than 10% of the total pensions finally paid with investment returns accounting for the 90% balance. ”
Regarding current annuity rates, a cost of closer to 30:1 is probably more appropriate (allowing for inflation and a partner’s pension). Thus a million pound pot may only provide an income in retirement of £30-£35 K a year (all dependent on gilt rates, mortatlity, expenses, reserving requirements etc)
Regarding the 5%, this is the pre-retirement investment assumption. A 25 year old should be able to invest in equities early on (with the potential for higher than 5% a year returns). The model just makes an assumption that overall the return can be modelled as 5% per annum. Each of us is free to determine our own return assumption – but the point is once you allow for tax relief and reasonable expense assumption, the contributions paid are a small proportion of the final sum – we all rely on the power of compound interest … and tax relief.
A number of providers have started to tackle the issue of sequence order risk. Schroder Life has a Flexible Retirement Fund which aims for a maximum drawdown of no more than 8% (ie a fall from the highest point to a low point); an annual return of inflation plus 2%; and the ability to take a 5% per annum income. Their literature indicates that the member has a 80% chance of not running out of funds after 30 years.
Jonathan, you must be an actuary; you certainly write like one.
Why only invest in equities “early on”?
Why not use appropriate equities (as opposed to equity markets or equity indices) as part of your post-retirement portfolio as well?
(your last post does not seem to have a reply button)
Kindest words anyone has said to me re writing like an Actuary 🙂
I used the expression invest in equities “early on” as shorthand to calm the nerves of all those people who say you should lifestyle to “less risky” gilts closer to retirement. I think it is appropriate to consider equities at all stages but for the purposes of the article – which was about sequencing risk etc – I don’t think the discussion of types of investment is important. The model used a 5% return and the comment was bonds would not get that kind of return. We are free to make our own calculations and use whatever rate we think is appropriate. The “shock” point was that the actual contributions represent a very small proportion of the final fund: the bulk is made up of tax relief and investment returns (and dare I say it employer contributions?)
Have a good day.
If you want to hear more about the Wise Men – listen to me and Katie Evans on Paul Lewis’ Moneybox at 18 minutes 25 seconds via this link http://www.bbc.co.uk/programmes/b06nhpxq