Average life expectancy in wealthy countries is around 79 for men and 83 for women, and it is a population average rather than a prediction about anyone. Lifestyle shifts it substantially and unevenly: heavy smoking removes about a decade, which is larger than every other commonly cited modifiable factor put together.
The number is useful for planning and useless as a forecast, and the distinction matters because it is routinely used for the second.
Why does the figure rise as you age?
Because it is conditional. Life expectancy at birth averages in everyone who dies young; life expectancy at 65 does not, because you have already survived those years.
That is why a 65-year-old’s remaining expectancy plus 65 exceeds the at-birth figure — often by several years. Any retirement or annuity calculation should use the conditional figure at the relevant age, and using the at-birth number instead understates how long the money has to last.
It is the same effect that makes historical life expectancies misleading. A figure of 35 in a pre-industrial population reflects enormous infant mortality rather than a world where nobody reached 50.
What actually moves it?
A short list, in a very uneven order.
| Factor | Rough effect |
|---|---|
| Heavy smoking | −10 years |
| Sustained regular exercise | +3 to 5 years |
| Obesity, severe | −3 to 5 years |
| Heavy alcohol use | −2 to 5 years |
| Diet quality | +1 to 3 years |
The smoking figure is consistent across large cohort studies and is the reason it dominates any list like this. Stopping recovers much of it, and the recovery is strongly age-dependent — quitting before 40 avoids most of the excess risk.
Everything below the first row is real, smaller, and much harder to isolate, because the behaviours cluster. People who exercise regularly also tend to smoke less and eat differently, and disentangling those in observational data is the central difficulty of the field.
What does the figure not include?
The two largest determinants, both of which are outside individual behaviour. Genetics accounts for a meaningful share and is not modifiable. Socioeconomic position accounts for a larger one than most people expect — the gap between the most and least deprived areas within a single wealthy country routinely exceeds ten years, which is the same order as heavy smoking.
Any calculator working from lifestyle inputs alone is therefore describing one slice of the variance. It is the slice you can act on, which is a reasonable thing to model and a poor thing to mistake for the whole.
Why is the median more useful than the mean?
Because mortality is not symmetric. The mean is pulled down by deaths at every age below it, while the median is the age by which half the cohort has died — and for planning purposes the median and the upper percentiles are the figures that describe the risk you are managing.
A cohort with a mean of 83 typically has a meaningful share still alive at 95. Planning to the mean therefore fails for close to half of the people it is applied to, which is an unusual definition of a plan.
How should a planning figure be chosen?
Longer than the average, deliberately. Half the population outlives the mean by definition, and running out of money is a much worse failure than leaving some behind.
Financial planning conventionally uses a figure well into the nineties for that reason, or models the distribution rather than a point. The 4% rule article covers what a thirty-year horizon assumes and where it stops holding.
The same asymmetry applies to insurance and to any decision where being wrong in one direction costs far more than the other. The average is the wrong input for an asymmetric problem.
What is healthy life expectancy?
The years lived in good health rather than the years lived, and it is the figure that changed most in the last few decades. Total life expectancy rose faster than healthy life expectancy in many countries, which means the extra years came with more of them spent in poor health.
For retirement planning that gap is the one that matters. Years of active retirement and years of care have very different costs, and a single lifespan figure treats them as identical.
Questions people ask
Why do women live longer? Consistently, across essentially every country and period, by three to five years. The causes are partly biological and partly behavioural, and the gap has narrowed somewhat as smoking rates converged.
Does the figure account for future medical advances? No. It is built from current mortality rates, so it implicitly assumes medicine stands still — which for a forty-year projection is a substantial and unstated assumption.
Is a calculator like this any use at all? As a way of seeing the relative size of the factors, yes. The ranking is more robust than the numbers, and the ranking is the actionable part.
How much does stopping smoking recover? Most of it if stopped before 40, and a meaningful share at any age. It is the single largest effect available in either direction.
Does the gap between countries close? Slowly and unevenly. Within wealthy countries the differences are now small compared with the differences between regions inside them, which is why the socioeconomic gradient is the more informative comparison.
Use the conditional figure, plan past it, and read the ranking rather than the number. The life expectancy calculator shows what each factor contributes, and the retirement age calculator is where the figure actually gets used.