Five strategies to combat sequence risk

It’s well known that one of the uncertainties retirees face in decumulation is sequence risk. This is where the same returns happening in a different chronological order can lead to different outcomes. Last year, we explained how sequence risk might be smaller than you think, busting the common myth that it’s the dominant decumulation risk.

Nevertheless, it remains a substantial source of uncertainty, explaining at least a quarter of the variance in investment returns achieved in decumulation. It is also an unrewarded risk.

A natural question arises, therefore: how can investors combat sequence risk? And do they need to be careful when doing so? We outline five strategies below, giving their pros and cons.

(1) Buy an annuity

One way to eliminate sequence risk (and many other risks) is simply to buy an annuity (annuitise). However, clearly this isn’t ideal for those willing to take more risk in the hope of higher retirement income, or for those looking to leave capital to future generations. Partial annuitisation is a potential compromise. Overall, though, we believe annuitisation is best thought of as a potential solution to rising longevity risk, rather than sequence risk.

(2) Take less investment risk

Another approach is to take less investment risk with non-annuitised assets, given that sequence risk is proportional to the volatility of asset returns. The drawback is that with lower investment risk there is also less scope to generate excess returns and meet the types of objectives tied to exceeding the income provided by annuities.

(3) Lower withdrawals

The less you withdraw, the lower sequence risk is. At one extreme you could avoid withdrawing anything. This would eliminate sequence risk and would be great for any inheritance objectives. The problem is that such a strategy provides no retirement income, which is usually the key objective, so clearly is not a sensible solution in most cases!

(4) Adaptive (or flexible) spending

Adapting spending to how the pot value fares through time (rather than spending predetermined amounts each year regardless of performance) tends to help improve outcomes and mitigate some sequence risk.

Depending on exactly how money is spent (in practice there are likely to be limits to how flexible the pensioner can be), adaptive spending can help to reduce or even eliminate the effects of sequence risk on bequest amounts. However, it doesn’t eliminate sequence risk in respect of retirement income, even though it reduces it.

As an example, someone spending a certain pre-determined percentage of their remaining pot each year (where the percentage could increase with age) will leave a sum behind that doesn’t depend on the order that investment returns arrived in. But they still will have enjoyed a better retirement, in terms of the income received over that period, if the bad returns arrived later than the good returns. This is illustrated for a heavily simplified example in the chart below:

(5) Rising glidepaths

The final strategy we consider, and seriously suggested by some, involves increasing the percentage in growth assets over time. One way to think of it is as a result of splitting assets into two pots: a relatively low-risk pot, which is spent from for the first 15 years, say, and a risky pot that isn’t touched during that period.

At the end of the period, perhaps aged 80, you might use the risky pot to buy an annuity (given longevity risk becomes higher in old age). The idea is that by not touching the risky assets, the order their returns arrive in doesn’t matter i.e. you’ll have combatted sequence risk.

There are a few problems with such an approach:

• It involves a higher proportion of the pot being invested in risky assets as time goes by (a ‘rising glidepath’). This is usually the opposite of what pensioners want![1]
• It can involve an abrupt change in asset allocation in the future e.g. switching from 100% growth to an annuity
• Last, but not least, our own research shows that a rising glidepath increases overall risk despite reducing sequence risk, if pensioners withdraw adaptively. This is explained in this blog

Evaluate risk holistically

As the last point underlines, sequence risk is only a part of overall risk. In general, it can be dangerous to fixate only on one dimension of uncertainty. When choosing between strategies we prefer to look at many scenarios that allow for investment, sequencing and longevity risk combined. We can then study the resulting range of outcomes, such as whether they run out of money.

After all, it’s outcomes that investors ultimately care about.

[1] Pensioners normally prefer to take less risk as their future time horizon / future life expectancy reduces. Rules of thumb, such as investing 100 minus their age in equities, reflect this preference (even if this rule can be questioned in other respects)