I recently heard Simone and Malcolm Collins talk about pronatalism on Spencer Greenberg's podcast. They make a pretty clear case that the coming demographic collapse is inevitable and, and further claim this will be catastrophic for humanity and for the economy. If so, you might think that pronatalism was a key pressing problem worthy of a considerable investment in resources. Do the claims hold up? And does pronatalism--increasing birthrates across the world--rank anywhere amongst other cause areas EAs deem important, and if so, where in the priority ranking could it lie?

I looked at some of the data points they've draw attention to and demographic collapse at least throughout Western Europe, the Americas, and East Asia--just about everywhere except Africa, the Middle East, and South Asia--is a pretty undeniable fact. The picture for global fertility rates seems less clear, and I wonder whether Metaculus and professional estimates of future populations have properly taken into account recent declines of fertility rates as well as the continuing declining trajectory of fertility rates in all parts of the world (including throughout sub-Saharan Africa). 

In the graphs, ~2.1 is the "replacement rate" of fertility, and numbers below that indicate shrinking populations. The Americas, Europe, and East Asia are already below replacement, and Middle East and North Africa and South Asia may get there within 20 years.

The trend in Sub-Saharan Africa is in the same direction, although at present, fertility in that region still indicates an expected doubling in population from the current generation of child-bearing-aged women to the next one.

Will it really be so catastrophic? AI is expected to automate a large number of roles. If it does, then the decline in the numbers of people available to work will not threaten economic decline as much as it would as if the economy continued to be as labor-intensive as it is today. Metaculus predicts the labour force participation in the US amongst working-age people will decline from 63% now to 58% in 2035 and 46% for the 2040s

If AI is likely to decrease the demand for work--seems plausible or even likely--why exactly is demographic collapse bad or dangerous or harmful? In that case it doesn't seem like it will cause economic collapse, because we'll no longer need the labor from younger workers to sustain the economy.

One reason: on total utilitarian grounds. For contemporary humans, life is worth living, and Malthusian ideas that more humans will mean a worse planet generally seem unconvincing (though reasonable people may disagree!). So, it seems that in general, every extra birth is a net positive for the world from a total utility standpoint, and that's sufficient enough to be a pronatalist.

But perhaps it not sufficient enough to prioritize pronatalism. In an era where we are facing down truly existential risks from AI and other human technologies, it seems like those issues are much more relevant for maximizing total utility in the very long term compared to whether humanity enters 2100 with a population of 9, 10, or 11 billion seems less important for maximizing total utility over the very long term. In fact, though I am mostly unconvinced by e.g., climate-change-based Malthusian stories, considering the pace of decarbonization, it seems possible that other forms of overcrowding, given the social and political structures we have, might precipitate military conflict of the sort that raises the risk of catastrophic warfare. 

I'm curious what you think about the idea

  • that AI is likely to mostly alleviate the risk of economic collapse derived from the coming demographic collapse, or whether you think there is still a risk of economic disruption.
  • the total-utilitarian long-termist critique that, although raising birthrates in the short-term is net positive, it seems relatively less important for maximizing total utility over the long term, because reducing catastrophic and existential risk form most of our leverage in the 21st century over total utility in the long-term future.

I can sort of anticipate a pushback that perhaps more births now will result in a larger human population right out into the long-term future; I suppose how plausible that idea is depends on whether we think there are likely to be future periods of population constraints, say, in the 22nd or 23rd century, after we have hopefully dealt with pressing AI risk but before humanity goes multi-planetary and multi-stellar.

So considering all the above, how important is pronatalism? Catastrophic risk arising from demographic collapse seems unlikely, so the case has to rest on the intrinsic value of additional lives, relative to other means of bringing about additional full human lives (i.e., avoiding deaths of infants and children). We can try about a quick BOTEC.

The average person worldwide can expect to live to the age of 73. For a committed total utilitarian who doesn't accept the person-affecting view, to a first approximation, bringing a new person into the world is a gain of 73 life-years, worth the same as preventing the death of one infant who would have lived the same number of years. Of course this is not quite right as the death of a person almost always has negative effects on those around them who will mourn their loss, although these are likely to be small relative to the value of the person's life itself. And most of us are probably willing to give at least a little bit of weight to the person-affecting view, where we care more about improving welfare of those who already exist than people we might choose to cause to exist. But with those caveats, a total utilitarian might value bringing a person into the world at roughly the same order of magnitude as they would preventing one death.

Famously, Givewell estimates they can save one life in Nigeria through donations to AMF for the value of around $3100. For the sake of simplicity, let's assume the value of one life anywhere is roughly equal. Then, pronatalism might be competitive as a cause area if we can cause on extra birth for the value of $3,100.

The cost of raising a child in the United States is estimated to be around $300,000. Coincidentally, Zvi Mowshowitz estimates around the same figure to generate one extra birth when used as a direct childraising subsidy. So, the actual raising of a child doesn't seem particularly competitive, as a cause area. However, there may be a number of interesting public policy interventions in which one extra birth could be gained for much lower than the cost of raising a child. For instance, it's been estimated that car seat regulations that tell parents what sort of car seats they must use for their children prevent around 8,000 births a year in the United States. Changing such a policy, if we counterfactually value one extra life at $3,100, might return $25m for every year the carseat policy remains in place.

That seems competitive with global poverty interventions if we accept a thoroughgoing total utilitarian, non-person-affecting view of life.

Does that make a great cause area? I think most of us don't really intuitively fully accept a non-person-affecting total utilitarian view. In other words, even if you think that all else being equal, a world with 100 people enjoying life is better than a world with only 99 enjoying lives equally as good, you still think it is more valuable to save the life of an existing person than to counterfactually bring one extra person into the world.

It seems to me the most interesting discussion to have on pro-natalism stands apart from the pure direct effect of bringing people into the world. Primarily, there is the impact on possible catastrophic risk (and thereby, possible knock-on effects for existential risk). But having ruled out economic collapse as a possible effect of collapsing birthrates, do we have anything else to worry about?





More posts like this

Curated and popular this week
Relevant opportunities