Could curing ageing lead to a population crisis? This is an attempt at some (very) simple calculations to get some idea, based on population projections by the United Nations (UN) and Institute for Health Metrics and Evaluation (IHME).
These calculations support a video exploring this by Andrew Steele, author of Ageless: The new science of getting older without getting old.
The simplest and most extreme case we’ll consider is a complete eradication of death in 2025. This is obviously absurd, for a number of reasons: we can’t possibly develop treatments that fast; even if we could, we couldn’t roll them out universally that fast; and, even if we somehow did cure ageing and roll out those treatments in just a few years, there would still be other ways to die, from traffic accidents to infectious diseases.
Nonetheless, this provides the most extreme imaginable scenario with which to benchmark our level of concern.
Without embarking on a complicated demographic modelling exercise, I’ll just take the two best-known population projections and assume no-one dies after 2025, ignoring any effect this might have on birth rates.
The UN has three headline scenarios, the high, medium and low variants, which differ depending on assumptions made about birth rates. I’ve plotted these below.
By 2050, the population would be 8.9bn, 9.7bn or 10.6bn under the low, medium and high variants, respectively. By 2100, the population would be 7.3bn, 10.9bn or 15.6bn.
What if no-one died? Under the medium variant, the population by 2050 would be 11.7bn. That’s 20% larger than if we’d not eradicated death.
By 2100, it would be 11.7bn: 69% larger.
That compares to a difference of 19% between the normal, death-included low and high variants in 2050, or 110% between them in 2100.
So, similar or even larger differences exist with plausible future variation in birth rates than if we were to literally cure death which, to reirerate, is a rather extreme assumption.
Let’s try a similar exercise with the IHME projections. They publish a number of different scenarios, with the ‘reference’ scenario as a baseline. We’ll look at this, and the two most extreme in population terms: ‘slower’ (which assumes slower development of poorer countries, and thus higher birth rates), and ‘SDG’ (which assumes rapid development in accordance with the Sustainable Development Goals, plus increased education for women, slowing birth rates more rapidly.)
This next graph is a bit confusing but I’m hoping if you’ve read this far you’l be willing to try to make sense of it! It includes all three UN variants (Low, Medium and High), and the deathless UN Medium variant (which is the higher line in the same colour); and it also includes the three IHME scenarios (reference, sdg and slower), plus a deathless SDG scenario to see if literally curing death ‘saves us’ from the risk of underpopulation by the end of the century in this case.
As you can see, the IHME scenarios tend to end up with lower populations than the UN ones, but they’re all in the same broad range, and literally curing death has an effect that’s broadly in line with the size of effect that adjustments in birth rate is expected to produce.
I wanted to try something just slightly more realistic than literally curing death, so I decided to try not of eradicating all death, but just the age-related fraction of it.
To perform these calculations, I have assumed a very simple change to death rates resulting from a ‘cure’ for ageing: that everyone aged 35 or over will have the same death rate as people aged between 30 and 34. This is quite an extreme assumption, and I’ve made it to make the most extreme possible projections again.
Regardless of the exact age you choose, annual death rates are very low for most people in most countries under the age of 60. The low death rate of young people means that the number of deaths caused by ageing is very large. When making that calculation, I made an opposite ‘extreme’ assumption that death rates would freeze at the end the our 30s, rather than the start. This was to ensure that my results would be conservative in the opposite direction—if anything, underestimating deaths due to ageing—to make my case as robust as possible.
This does mean that the ageless scenario here and the number of deaths caused by ageing in those stats don’t quite tally up—but all the code for these calculations is available on GitHub, so you can play around with these scenarios, check my working, or develop better ones if you’re interested!
The other major group projecting population is the IHME. Their forecasts are rather different to the UN's, predicting that global population will peak mid-century and decline after that.
This graph shows the UN’s medium variant (bottom line), the ageless version of that (ageing completely cured in 2025, the next line up), and the deathless version from before (the line above that).
After curing ageing, the population would be 11.3bn in 2050 (1.6, or 16% larger), and 17.3bn (6.4, or 59% larger) in 2100.
As a reminder, that compares to a difference of 1.7 (19%) between the normal, death-included low and high variants in 2050, or 110% (8.3) between them in 2100.
Thus (obviously!), this ‘worst-case’ scenario is made slightly less extreme by only curing age-related death rather than all death. The two scenarios are compared to the regular medium variant in the next graph which, from top to bottom, is deathless, ageless and regular medium variant.
Finally, this graph shows the IHME’s reference scenario with (blue) and without (dark blue) deaths due to ageing, compared to the UN medium variant with (green) and without (dark green). Again, the differences here show that existing variation between model variants or scenarios is smaller but not massively so than totally curing ageing in a few years’ time. Given a more realistic timeframe, it’s fair to assume that treatments for ageing will only modestly increase uncertainty on future population projections.
Life expectancy assumptions in population projections
Nonetheless, more work is most definitely needed here: though the changes are only likely to be relatively modest compared to existing scenarios, the lack of interest from demographers about future trends in life expectancy is striking.
Taking the IHME projections as an example, this is what they anticipate happening in (from top to bottom) Japan, all high-income countries averaged, and the US until 2100. None of the scenarios offer any real variation, and they all project that life expectancy will increase more slowly now than in the past: a projection which has repeatedly been proven wrong in previous forecasts.
In some less wealthy countries (India and Nigeria here), there is a bit more variation, but presumably mainly driven by mortality which is reduced through development, since it’s the ‘slower met need and education’ scenario at the bottom. It’s also quite surprising to me that they think these countries will plateau at lower levels than the high-income countries—why wouldn’t people in Nigeria and India eventually live as long as rich people do today?
It would be straightforward to add the option to make this slightly more realistic by gradually reducing age-related deaths over a few decades, rather than entirely removing them instantaneously or, even better, to gradually reduce them starting with countries with the highest GDP per capita. If anyone would like to do this please make a pull request!
Obviously best of all would be some more thorough modelling of these consequences, but that would require some proper demographic expertise…
About this report
This report was generated from an R script which performs the underlying calculations. The code and console output have been removed to make it easier to read. If you want to see the code, you can check out the latest version at GitHub, and if you want to view the console output, see population.txt.