Life expectancy is an average and you are not one

Both approaches to planning for an unknown lifespan work. What differs is what they cost you, and the cost is what this sets out.

The average is the wrong number

Life expectancy figures describe the average of a distribution, and roughly half of any group lives beyond that figure. Planning a portfolio to run out at exactly the average is therefore planning deliberately for something close to a coin-flip outcome. The distribution has a long tail, and a meaningful proportion of people live a great deal longer than the average.

Conditional life expectancy also rises with age, so somebody who has already reached retirement has a longer expectation than at birth. Using the figure quoted at birth for somebody in their sixties therefore understates the horizon substantially.

Why longevity risk is different

Most financial risks can be diversified or hedged, whereas the risk of living a long time cannot be diversified by an individual. It can be pooled across many people, which is what guaranteed income products do, and that pooling is the actual service being purchased. The price of that pooling reflects mortality assumptions, prevailing interest rates and the provider's own margin.

People frequently compare such products against investment returns, which measures something else entirely. What is being bought is protection against a specific risk rather than an investment return at all.

Planning to a percentile instead

A more robust approach plans to an age well beyond the average rather than to the average figure itself. That produces a lower sustainable withdrawal and a considerably lower chance of the money running out.

Put simply, the cost is spending less throughout, in exchange for protection against an outcome that may never occur. How far beyond the average to plan is a personal judgement informed by health and by family history. Actuarial tables are published in most countries and are more useful for this than any general rule of thumb.

The couple version

For a couple the relevant horizon runs until the second death, which is materially longer than for either individual. Plans built on one person's life expectancy underestimate the period over which the money has to last. Survivor arrangements on pensions and guaranteed income differ enormously and are chosen at purchase, frequently irreversibly.

On an ordinary week, the income of a surviving partner commonly falls by considerably more than their spending does.

This is among the most consequential decisions in the area and among the least discussed before it has to be made.

What flexibility buys

A plan able to reduce discretionary spending in poor conditions can start from a higher withdrawal than one that cannot. Flexibility is therefore an alternative to caution rather than something added on top of it, and the two are frequently confused. Deciding in advance which spending counts as discretionary makes that flexibility real rather than merely theoretical.

For most people, guaranteed income covering essentials converts the longevity problem into a much smaller one for whatever remains. Combining a floor of guaranteed income with a flexible portfolio is a common structure for precisely this reason.

Some of this will suit you and some will not, and that is the point.

Revisiting as you go

The horizon shortens with each year that you live, and a plan can be recalculated periodically rather than fixed once and for all at the outset. Annual recalculation naturally increases the sustainable withdrawal as the remaining period gets shorter.

It also responds to how the portfolio has actually performed rather than to an assumption made years earlier. This approach produces a variable income, which suits some households and emphatically does not suit others. Which structure fits your own circumstances is exactly the question regulated advice exists to answer.

Side by side

The takeaway

Half of everyone outlives the average. Plan for the tail, not for the middle of the distribution.

Pick the one that costs you least, and let the rest wait.