This was written in my role as researcher at CEARCH, but any opinions expressed are my own. I will check comments on the EA Forum version more often.
This report uses population dynamics to explore the effects of a near-existential catastrophe on long-term value.
Summary
Global population would probably not recover to current levels after a major catastrophe. Low-fertility values would largely endure. If we reindustrialize quickly, population will stabilize far lower.
Population “peaking lower” after a catastrophe would make it harder to avoid terminal population decline. Tech solutions would be harder to reach, and there would be less time to find a solution.
Post-catastrophe worlds that avoid terminal population decline are likely to emerge with values very different to our own. Population could stabilize because of authoritarian governments, prescriptive gender roles or civil strife, or alternatively from increased collective concern for the future.
Conclusion: Near-existential catastrophes are likely to decrease the value of the future through decreased resilience and the lock-in of bad values. Avoiding these catastrophes should rank alongside avoiding existential catastrophes.
Introduction
In this report I use population dynamics to explore the question “What are the long-term existential consequences of a non-existential catastrophe?”. I do not claim that population dynamics are the only, or even the most important, consideration.
Others have written about the short-term existential effects of a global catastrophe. Luisa Rodriguez argues that even in cases where >90% of the global population is killed, it is unlikely that all viable groups of survivors will fail to make it through the ensuing decades (Rodriguez, 2020). The Global Catastrophic Risk Institute has begun to explore the long-term consequences of catastrophe, although they consider this “rather grim and difficult-to-study topic” to be neglected (GCRI).
What comes after the aftermath of a catastrophe is very difficult to predict, as life will be driven by unknown political and cultural forces. However, I argue that many of the familiar features of population dynamics will continue to apply.
Even without a catastrophe, we face a possible population problem. As countries develop, their populations peak and begin to decline. If these trends continue, global population will shrink until either we “master” the problem of population, or we can no longer maintain industrialized civilization (multiple working papers, Population Wellbeing Initiative, 2023). It could be argued that this is not a pressing problem. It will be centuries before global population drops below 1 billion, so we have time to overcome demographic decline or to make it irrelevant by relying on artificial people. But in the aftermath of a global catastrophe there may be less time and fewer people available to solve the problem.
Longtermists may argue that most future value is in the scenarios where we overcome reproductive constraints and expand to the stars (Siegmann & Mota Freitas, 2022). My findings do not contradict this. But such scenarios appear to be significantly less likely in a post-catastrophe world. And the worlds in which we do bounce back seem likely to have values very different from our own.
Population recovery after a catastrophe
In this section I examine three models for determining population growth. I find that full population recovery after a major global catastrophe is unlikely, and that the worlds which do recover are likely to emerge with values very different from those of the pre-catastrophe world.
It’s worth noting that a catastrophe need not inflict its damage at one point in time. The effects of some historical famines and pandemics have unfurled over many years, and may have triggered larger civilizational collapses. I assume in this report that a catastrophe ends when population reaches its lowest point, although I recognize that this is an artificial distinction.
1 Natural Limits Approach
Resource-constraint, or Malthusian, models stem from the premise that human populations will increase while there are sufficient resources available. When the population consumes all available resources, mechanisms like famine kick in to keep the population in check.
This approach might predict that post-catastrophe population will bounce back to the present-day trajectory—or even exceed it, if ways are found of increasing or rationing the available resources.
But demographic data simply does not match this model. Countries with abundant resources like Canada and Australia have fertility rates well below replacement levels, while poverty-stricken Nigeria has a fertility rate of 5 children per woman. Population growth appears to be positively correlated with food scarcity (Our World in Data, 2018).
Global population is projected to peak at ~10 billion later this century largely because people are choosing to have fewer children, not because of mass mortality triggered by sparse resources.
The Malthusian model may apply to populations of digital people, where growth could be capped by the amount of energy or computational power available. But the key biological reproductive parameters, fertility and mortality, appear to change in parallel with social and economic development.
2 Fertility Approach
Fertility rates were high throughout all of recorded history, right up until France became the first country to experience a decline in the 1700s (Cummins, 2008, Blanc, 2023). Since then, the overall trend has become clear: as living conditions improve, fertility falls. Identifying the specific drivers of low fertility is more difficult, but may help us determine what post-catastrophe population change will look like.
Our World in Data have an excellent report on fertility in which they identify likely fertility indicators (OWID, 2017), which I have divided into the two broad categories “Cultural” and “Economic”.
Cultural
Economic
Women’s labor participation
Child mortality
Attitudes to contraception
Access to contraception
Women’s education
Pension replacing children as old-age safety net
Children’s status
Societal norms
Media
Family planning
Religion
Consider a country with low fertility (eg. Canada: fertility rate 1.5). Which of the key fertility indicators would change in the aftermath of a near-existential catastrophe?
Clearly, all of them could change. But some would be more resilient than others.
Child mortality, contraceptive access and pensions would almost certainly be transformed amid disruption to the global supply chain. We can expect, at least in the short term, an increase in fertility for these reasons. These indices should gradually “recover” to their pre-catastrophe states as the post-catastrophe world redevelops.
The effect of a catastrophe on attitudes to contraceptives & family planning, the roles of women & children and the influences of media & religion are far less clear. Some cultural factors may be reversed: for example, reverting to an agricultural economy may decrease our emphasis on educating children, both reducing the cost of raising them and increasing their utility as productive workers. But even in such scenarios, key cultural barriers to high fertility are likely to remain in place.
It seems almost obvious that fertility culture is “sticky” over time. We see it in rapidly developing countries, when the fertility rate begins to drop many years after improvements to quality of life begin to reduce mortality rates. Similarly, we can expect low-fertility culture to be “sticky” after a catastrophe: suppose you are a typical young woman in a wealthy country, who plans to have only 1-2 children. If a terrible catastrophe thrusts society into a pre-industrial state, you won’t suddenly want to have 5 children. Any increases in fertility are likely to be moderate and, if contraceptive access is restored quickly, short-lived.
In short, catastrophe is likely to affect economic indicators more than it will affect cultural ones. The global economy may be destroyed, but the culture of the “new world” that emerges would be partly shaped by survivors and their pre-catastrophe values.
So fertility may rise in the aftermath of a catastrophe, but not to pre-industrial levels of 5+ children per woman. This can facilitate rapid population growth, to be sure. If women each bear an average of 4 children who survive into adulthood, population will double once a generation. Within three generations, a world of 1 billion could recover to 8 billion.
However such sustained, explosive growth has never been seen in the past. Periods of high growth occur during a “demographic transition” and follow a predictable pattern, which we explore in the next section.
3 Demographic Transition Approach
Broadly, there have been two types of population growth: the slow, faltering growth of pre-industrial states and the population explosion that accompanies demographic transition. Demographic transition is the change from a high-mortality, high-fertility society to a low-mortality, low-fertility society. Because mortality rates begin to drop first, there is a period when mortality rates are lower than fertility rates, leading to population growth.
For populations to recover after a catastrophe there would need to be a long period of slow growth, or a period of demographic transition, or both.
If slow growth is to lead to full population recovery, it would need to last for a very long time (~2000 years to recover from 1 to 8 billion at 1000-1500AD growth rates). This would be a very risky scenario, as it would imply that either (a) society has reverted to a pre-industrial state of high mortality and high fertility and runs the risk of never re-industrializing or (b) society is “stuck” at an intermediate level of development with low growth rates—indicative of cultural stagnation. These scenarios seem unusually likely to feature the lock-in of bad values (MacAskill, 2022).
If society is to reindustrialize after the catastrophe it seems likely that the process will either begin long after the event, when enough time has passed for key things to be rediscovered (Rodriguez, 2022[1]), or very soon after the event, while survivors retain pre-catastrophe knowledge and culture. The former scenario would involve slow growth and a demographic transition, but seems very risky as there would be plenty of time for things to go wrong. The latter scenario, in which we immediately begin to rebuild, would involve a demographic transition as reindustrialization led to falling mortality and fertility rates. But how big would it be?
Rather than get too deep into analyzing a situation with so many variables, I will present a simple thesis: the stickiness of fertility culture means the boom won’t be big enough for full population recovery.
Put it this way: if a catastrophe reduces global population to 1 billion, a level last seen in 1800:
Our fertility behavior will not fully revert to that of 1800
Fertility rates will decline to 2023 levels much faster than they did after 1800
Point 1 results from the argument made in the previous section that fertility behavior is “sticky”, and that fertility rates after the catastrophe will be strongly influenced by those before the catastrophe. It is possible that fertility rates will rise over time, as new generations are raised in the undeveloped post-catastrophe world and adopt larger family sizes. But this initial lag is enough to ensure that population won’t fully recover to 2023 trajectories.
If the survivors begin to rebuild civilization immediately, it seems likely that society will develop more rapidly than it did in 1800, which supports point 2. We won’t have to wait for key discoveries about electricity, medicine etc. to be remade from scratch: books and artifacts will speed up the process.
Still, it won’t be easy. The loss of tacit knowledge will be a blow—having access to information about hydroelectric power plants does not mean you have the expertise to run one. Another hitch appears to be fossil fuel depletion, which will make it harder for a nascent industrial society to secure cheap energy. But even if these bottlenecks slow down economic progress, it’s possible that some of the low-fertility cultural factors will endure.
All things considered, it seems highly likely that development, and thus fertility, would recover to 2023 levels in far less than 200 years. This shorter, less potent boom would result in population peaking at a lower level than it is currently projected to (~10 billion).
In summary, for a post-catastrophe world to recover through slow growth would entail a long, risky period of recovery. Immediate reindustrialization would be safer, although the boom may well be muted. Only a very large population boom could take a world of 1 billion back to 8 billion. Such booms are possible, but they would require an unprecedented reversal of fertility trends.
What kinds of worlds bounce back?
Population recovery after a mass-mortality catastrophe is far from assured. In the most optimistic scenarios, where humanity is able to quickly raise living standards, we can expect fertility rates to rapidly return to their current trends (2.3 and falling). It seems likely that such scenarios will lead to population peaking far lower than their current projected maximum of around 10 billion (UN, 2022[2]). In this section we explore whether the worlds in which population does recover are “less worth saving”.
Drawing from the previous section, the kinds of worlds that bounce back are more likely to have one or both of the following properties:
A long period of slow growth
A prolonged population boom
Property (1) is fraught with risk. Persistent slow growth has only been observed in pre-transition countries where fertility rates are high and life expectancy is low (see OWID, 2019). Long, slow growth would very likely mean that the catastrophe was so severe that it was not possible to immediately begin rebuilding society, and therefore that subsequent generations were unable to preserve basic scientific knowledge (such as germ theory) or pre-catastrophe culture (women’s rights, investment in children’s education). There is no guarantee that such a preindustrial society would be able to rekindle advanced civilization (Mokyr, 2016), especially with disadvantages such as an altered climate and depleted fossil fuel deposits (Dartnell, 2015) and so eventual extinction seems to be a likely outcome.
It is possible that (1) emerges from a culture that is “in control” of population growth. Perhaps the survivors feel a renewed sense of common responsibility to repopulate the earth in a gradual, measured way. This would be unprecedented—the nearest modern-world analogue is Israel, where fertility norms have been unusually resilient to cultural changes.
Property (2), on the other hand, describes situations where the post-catastrophe demographic boom is enough for full population recovery. This requires a considerable period of high fertility and low mortality. Low mortality is likely, as the biggest contributors to reduced mortality have been relatively cheap, low-tech innovations. Sustained high fertility rates are likely to be the product of civil strife and faltering economic recovery—like Nigeria, whose civil war, economic collapse and periodic ethnic conflict may have contributed to a stubbornly high fertility rate.
In short, the scenarios that lead to full post-catastrophe population recovery are those in which human fertility behaviors are radically reversed, either due to persistent low living conditions or to a shift in values. These scenarios are likely to be more disruptive than scenarios in which the population does not fully recover, and we can be less certain about the types of societies that will emerge from them.
The second stage of this report explores the consequences of population not fully recovering after a catastrophe.
What are the consequences of population peaking lower?
With the exception of Israel, every wealthy country has below-replacement fertility levels. If current trends continue, global population will decline precipitously in the 22nd century. Eventually there will be too few of us to maintain a modern society.
Fortunately we have time to respond. Culture could adapt in a way that enables sustainable reproduction. Technological advances could make procreation much easier.
But after a major catastrophe in which population never recovers to its current levels, there will be far less time. We can be less confident that we will overcome demographic decline. And the worlds that do overcome it are likely to be culturally very different from the world today.
Pathways to security
I model the interaction between states according to the diagram below. A “mature” society is one in which mortality and fertility rates have decreased far enough to cease population growth. I assume that after this, global population will begin to decline until (a) industrial civilisation can no longer be maintained or (b) we find some way to stop the population decline.
The key question in this section is to what extent will lower peak population make it harder to avoid demographic decline? I will also be asking what kinds of worlds manage to avoid demographic decline after peaking lower?
I distinguish between two possible end states: “Existential security”, which I take to be equivalent to attaining multiple independent interstellar colonies (Arepo, 2022), and “Extinction”.
This section will focus on near-existential catastrophes in which the survivors are immediately able to begin rebuilding society.
The risk of terminal population decline
Even without a near-existential catastrophe, terminal population decline is a significant threat. Not only will it be harder to reach for the stars with a falling population; establishing interstellar colonies will be futile if fertility rates remain below replacement levels. In order to survive in the long term, our civilization will have to somehow reverse an established demographic trend: as living standards improve, fertility rates fall.
This problem is not seen as a top priority within longtermism—perhaps partly due to the assumption that AI or other technology will one day make controlling population growth much easier. But there are reasons to be pessimistic about this, especially in post-catastrophe scenarios where populations peak lower.
Let’s take a look at some of the ways we might overcome terminal population decline. Much of what follows is inspired by a working paper from the Population Wellbeing Initiative* (Kuruc, K and Weston, G. How large is the value of humanity’s future? 2023) which explores long-run population trends in the absence of technology that can create digital people. To plug the gap I add in some of my own analysis on digital people.
Ways we could overcome terminal population decline—and some reasons to be pessimistic about them:
Space colonies. Fertility is low in wealthy countries with large unsettled territories (Canada, Australia), even though they are far more hospitable than other planets. There is no reason to think that space colonies alone will reverse the fertility decline.
Artificial wombs. Pregnancy is only a small part of the cost of having children, so artificial wombs may not significantly shift procreative decision-making. Past advances in assisted reproductive technologies seem not to have done much to increase fertility (Lazzari et al, 2023).
Government interventions. Hungary has shown the most success in increasing fertility, going from 1.2 to 1.6 in the decade after 2011 (World Bank, 2023). But Hungary is not on track to eventually reach replacement rates, despite the costly incentives. Even if all of the fertility increase is attributed to policy, which it probably shouldn’t be (IFS, 2018), it seems unlikely that government incentives alone will ever turn the tide.
Heritable Fertility. Inevitably, high-fertility subgroups grow faster than the general population. It has been suggested that this ensures the demise of secularism (Kauffman, 2010), and it seems to imply that eventually most of the global population will belong to a high-fertility subgroup. Unfortunately, the theory fails to account for the fact that fertility is not fully heritable, and that subgroups tend to become less extreme as they grow and assimilate with the wider population.
Change in values. If childrearing becomes more prestigious or more desirable relative to seeking career success, more people will choose to have children. This could happen in a future where most “work” is automated, or where attitudes towards women change significantly. This would be a significant reversal of the trends we have seen in the past, where both women’s rights and the cultural capital attached to careers tend to increase as living standards improve.
Life extension. Longer lives would give people more time to pursue multiple life goals like raising children and building a career. However, this would likely be ineffective without parallel breakthroughs in reproductive technology.
Digital people. Non-biological reproduction would be much easier to control and may only be limited by constraints on resources like energy or computing power. If digital people are possible, and are developed in time, we could end our reliance on biological reproduction.
Decline to a pre-industrial state. Population levels might fall so far that global trade, industrial production etc. become infeasible. The resulting societies would likely have much higher fertility and would begin to grow again. This process seems likely to act as a buffer from complete extinction.
The above paints a concerning picture, and it looks even more concerning in the aftermath of a near-existential catastrophe. We have relied on population growth to maintain levels of technological progress and economic growth in the past century, and cresting population brings the threat of stagnation in the present one (MacAskill, 2022). In futures where the population peaks in the low billions, there will be less time and fewer people available to find a fix.
In light of this, the main scenarios in which we overcome fertility decline either (a) involve a major sift in values, or (b) are harder to achieve in a world with much lower population:
Governments impose effective pro-fertility measures. Given that governments are not currently able to bring low fertility back to replacement levels, this scenario likely means that governments either have much more relative power than they do today, or that they have much stronger incentives to boost fertility. Increased government power is closely linked with authoritarianism. Stronger incentives to boost fertility could come from a more far-sighted society seeking to avert long-term decline, or from competition & conflict with other governments.
Post-catastrophe society develops pro-fertility norms & values. Modern-day high-fertility subgroups are predominantly religious (the Hutterites and the Mormons in the US, the Haredi in Israel) and have prescriptive views on women’s roles. If this was the source of pro-fertility norms, the consequences for civilization’s long-run values could be bad. It is possible, however, that pro-fertility norms arise from an increased sense of collective responsibility for restoring global civilization, or from “the end of work” in an age of automation.
Life extension & reproductive breakthroughs. The very possibility of radical life-extending treatments is questionable, and breakthroughs seem less likely in a lower-population world with fewer researchers.
Digital people. Creating digital people may not be possible. If it is possible, it will be more difficult to attain in a future world with lower population.
Decline to a pre-industrial state. All known pre-industrial societies have had high fertility, so we should expect this to be the case in a near-empty world that cannot sustain industrialization. It’s possible that this society could “grow back” and one day attain existential security—but there will be plenty of opportunities for things to go wrong.
To summarize, the worlds in which a low-population post-catastrophe society overcomes terminal population decline are often associated with authoritarianism, conflict or limited roles for women. Some of these worlds arise from cultures of long-term thinking and collaboration, and others from technological innovations like life extension or digital people. But overall, a global catastrophe would make it less likely that we would find a technological fix.
Implications
This report aims to use population dynamics to increase our understanding of the consequences of near-existential catastrophe. It does not claim that population dynamics are the only, or even the most important, consideration.
Nevertheless, this analysis leads to some concrete implications for cause prioritization:
We should update the value of preventing near-existential catastrophes, since they appear to both increase the risk of extinction and to increase the risk of future worlds with values very different from our own.
Interventions should be partially judged by the effect they have on our chances of eventually overcoming terminal population decline—either by creating/preserving life[3]*, or by increasing the chances of finding a cultural/technological “solution”. Optimism about technological fixes for demographic decline seems unjustifiably high, and optimists should do more to argue their case.
Conclusion
This report aims to use population dynamics to give a fresh perspective on how to weigh near-existential catastrophes against full-existential catastrophes.
We found that it is likely that population would not recover to its previous trajectory in the aftermath of a mass mortality event. The scenarios in which population does recover are often associated with a radical change in fertility culture, or a slow and faltering recovery. These worlds may be worth saving, but appear to be riskier than the world which we currently inhabit.
Terminal population decline may become a serious problem in the coming centuries, and the pathways to averting it are less reliable than many assume. The problem of terminal decline becomes even more acute in worlds with a lower population, as there are fewer people and less time to find a solution. Again, the worlds that do make it through give cause for concern: most known high-fertility cultures are associated with some combination of low living standards, limited women’s rights and poor education.
Population dynamics suggest that a near-existential catastrophe would introduce existential risks and threats to long-term value that were previously not considered. Hence we should update the value of preventing near-existential catastrophes.
Note that Rodriguez published this work in 2022 with the disclaimer “This is a rough draft I wrote between October 2019 and April 2020. It’s incomplete, and doesn’t reflect updates in my views in the 2+ years since I worked on it.”
The UN population projections after 2100 are out-of sync with the trends that we have seen so far in the wealthiest countries. The projections assume that global fertility levels will stabilize around the rate of replacement. However, while there is evidence for a slight uptick in fertility among the wealthiest countries, none of these “uptick” countries have seen their fertility rates reach replacement levels. Hence the UN’s projections of eventual stability seem difficult to justify.
Population After a Catastrophe
This was written in my role as researcher at CEARCH, but any opinions expressed are my own. I will check comments on the EA Forum version more often.
This report uses population dynamics to explore the effects of a near-existential catastrophe on long-term value.
Summary
Global population would probably not recover to current levels after a major catastrophe. Low-fertility values would largely endure. If we reindustrialize quickly, population will stabilize far lower.
Population “peaking lower” after a catastrophe would make it harder to avoid terminal population decline. Tech solutions would be harder to reach, and there would be less time to find a solution.
Post-catastrophe worlds that avoid terminal population decline are likely to emerge with values very different to our own. Population could stabilize because of authoritarian governments, prescriptive gender roles or civil strife, or alternatively from increased collective concern for the future.
Conclusion: Near-existential catastrophes are likely to decrease the value of the future through decreased resilience and the lock-in of bad values. Avoiding these catastrophes should rank alongside avoiding existential catastrophes.
Introduction
In this report I use population dynamics to explore the question “What are the long-term existential consequences of a non-existential catastrophe?”. I do not claim that population dynamics are the only, or even the most important, consideration.
Others have written about the short-term existential effects of a global catastrophe. Luisa Rodriguez argues that even in cases where >90% of the global population is killed, it is unlikely that all viable groups of survivors will fail to make it through the ensuing decades (Rodriguez, 2020). The Global Catastrophic Risk Institute has begun to explore the long-term consequences of catastrophe, although they consider this “rather grim and difficult-to-study topic” to be neglected (GCRI).
What comes after the aftermath of a catastrophe is very difficult to predict, as life will be driven by unknown political and cultural forces. However, I argue that many of the familiar features of population dynamics will continue to apply.
Even without a catastrophe, we face a possible population problem. As countries develop, their populations peak and begin to decline. If these trends continue, global population will shrink until either we “master” the problem of population, or we can no longer maintain industrialized civilization (multiple working papers, Population Wellbeing Initiative, 2023). It could be argued that this is not a pressing problem. It will be centuries before global population drops below 1 billion, so we have time to overcome demographic decline or to make it irrelevant by relying on artificial people. But in the aftermath of a global catastrophe there may be less time and fewer people available to solve the problem.
Longtermists may argue that most future value is in the scenarios where we overcome reproductive constraints and expand to the stars (Siegmann & Mota Freitas, 2022). My findings do not contradict this. But such scenarios appear to be significantly less likely in a post-catastrophe world. And the worlds in which we do bounce back seem likely to have values very different from our own.
Population recovery after a catastrophe
In this section I examine three models for determining population growth. I find that full population recovery after a major global catastrophe is unlikely, and that the worlds which do recover are likely to emerge with values very different from those of the pre-catastrophe world.
It’s worth noting that a catastrophe need not inflict its damage at one point in time. The effects of some historical famines and pandemics have unfurled over many years, and may have triggered larger civilizational collapses. I assume in this report that a catastrophe ends when population reaches its lowest point, although I recognize that this is an artificial distinction.
1 Natural Limits Approach
Resource-constraint, or Malthusian, models stem from the premise that human populations will increase while there are sufficient resources available. When the population consumes all available resources, mechanisms like famine kick in to keep the population in check.
This approach might predict that post-catastrophe population will bounce back to the present-day trajectory—or even exceed it, if ways are found of increasing or rationing the available resources.
But demographic data simply does not match this model. Countries with abundant resources like Canada and Australia have fertility rates well below replacement levels, while poverty-stricken Nigeria has a fertility rate of 5 children per woman. Population growth appears to be positively correlated with food scarcity (Our World in Data, 2018).
Global population is projected to peak at ~10 billion later this century largely because people are choosing to have fewer children, not because of mass mortality triggered by sparse resources.
The Malthusian model may apply to populations of digital people, where growth could be capped by the amount of energy or computational power available. But the key biological reproductive parameters, fertility and mortality, appear to change in parallel with social and economic development.
2 Fertility Approach
Fertility rates were high throughout all of recorded history, right up until France became the first country to experience a decline in the 1700s (Cummins, 2008, Blanc, 2023). Since then, the overall trend has become clear: as living conditions improve, fertility falls. Identifying the specific drivers of low fertility is more difficult, but may help us determine what post-catastrophe population change will look like.
Our World in Data have an excellent report on fertility in which they identify likely fertility indicators (OWID, 2017), which I have divided into the two broad categories “Cultural” and “Economic”.
Cultural
Economic
Consider a country with low fertility (eg. Canada: fertility rate 1.5). Which of the key fertility indicators would change in the aftermath of a near-existential catastrophe?
Clearly, all of them could change. But some would be more resilient than others.
Child mortality, contraceptive access and pensions would almost certainly be transformed amid disruption to the global supply chain. We can expect, at least in the short term, an increase in fertility for these reasons. These indices should gradually “recover” to their pre-catastrophe states as the post-catastrophe world redevelops.
The effect of a catastrophe on attitudes to contraceptives & family planning, the roles of women & children and the influences of media & religion are far less clear. Some cultural factors may be reversed: for example, reverting to an agricultural economy may decrease our emphasis on educating children, both reducing the cost of raising them and increasing their utility as productive workers. But even in such scenarios, key cultural barriers to high fertility are likely to remain in place.
It seems almost obvious that fertility culture is “sticky” over time. We see it in rapidly developing countries, when the fertility rate begins to drop many years after improvements to quality of life begin to reduce mortality rates. Similarly, we can expect low-fertility culture to be “sticky” after a catastrophe: suppose you are a typical young woman in a wealthy country, who plans to have only 1-2 children. If a terrible catastrophe thrusts society into a pre-industrial state, you won’t suddenly want to have 5 children. Any increases in fertility are likely to be moderate and, if contraceptive access is restored quickly, short-lived.
In short, catastrophe is likely to affect economic indicators more than it will affect cultural ones. The global economy may be destroyed, but the culture of the “new world” that emerges would be partly shaped by survivors and their pre-catastrophe values.
So fertility may rise in the aftermath of a catastrophe, but not to pre-industrial levels of 5+ children per woman. This can facilitate rapid population growth, to be sure. If women each bear an average of 4 children who survive into adulthood, population will double once a generation. Within three generations, a world of 1 billion could recover to 8 billion.
However such sustained, explosive growth has never been seen in the past. Periods of high growth occur during a “demographic transition” and follow a predictable pattern, which we explore in the next section.
3 Demographic Transition Approach
Broadly, there have been two types of population growth: the slow, faltering growth of pre-industrial states and the population explosion that accompanies demographic transition. Demographic transition is the change from a high-mortality, high-fertility society to a low-mortality, low-fertility society. Because mortality rates begin to drop first, there is a period when mortality rates are lower than fertility rates, leading to population growth.
For populations to recover after a catastrophe there would need to be a long period of slow growth, or a period of demographic transition, or both.
If slow growth is to lead to full population recovery, it would need to last for a very long time (~2000 years to recover from 1 to 8 billion at 1000-1500AD growth rates). This would be a very risky scenario, as it would imply that either (a) society has reverted to a pre-industrial state of high mortality and high fertility and runs the risk of never re-industrializing or (b) society is “stuck” at an intermediate level of development with low growth rates—indicative of cultural stagnation. These scenarios seem unusually likely to feature the lock-in of bad values (MacAskill, 2022).
If society is to reindustrialize after the catastrophe it seems likely that the process will either begin long after the event, when enough time has passed for key things to be rediscovered (Rodriguez, 2022[1]), or very soon after the event, while survivors retain pre-catastrophe knowledge and culture. The former scenario would involve slow growth and a demographic transition, but seems very risky as there would be plenty of time for things to go wrong. The latter scenario, in which we immediately begin to rebuild, would involve a demographic transition as reindustrialization led to falling mortality and fertility rates. But how big would it be?
Rather than get too deep into analyzing a situation with so many variables, I will present a simple thesis: the stickiness of fertility culture means the boom won’t be big enough for full population recovery.
Put it this way: if a catastrophe reduces global population to 1 billion, a level last seen in 1800:
Our fertility behavior will not fully revert to that of 1800
Fertility rates will decline to 2023 levels much faster than they did after 1800
Point 1 results from the argument made in the previous section that fertility behavior is “sticky”, and that fertility rates after the catastrophe will be strongly influenced by those before the catastrophe. It is possible that fertility rates will rise over time, as new generations are raised in the undeveloped post-catastrophe world and adopt larger family sizes. But this initial lag is enough to ensure that population won’t fully recover to 2023 trajectories.
If the survivors begin to rebuild civilization immediately, it seems likely that society will develop more rapidly than it did in 1800, which supports point 2. We won’t have to wait for key discoveries about electricity, medicine etc. to be remade from scratch: books and artifacts will speed up the process.
Still, it won’t be easy. The loss of tacit knowledge will be a blow—having access to information about hydroelectric power plants does not mean you have the expertise to run one. Another hitch appears to be fossil fuel depletion, which will make it harder for a nascent industrial society to secure cheap energy. But even if these bottlenecks slow down economic progress, it’s possible that some of the low-fertility cultural factors will endure.
All things considered, it seems highly likely that development, and thus fertility, would recover to 2023 levels in far less than 200 years. This shorter, less potent boom would result in population peaking at a lower level than it is currently projected to (~10 billion).
In summary, for a post-catastrophe world to recover through slow growth would entail a long, risky period of recovery. Immediate reindustrialization would be safer, although the boom may well be muted. Only a very large population boom could take a world of 1 billion back to 8 billion. Such booms are possible, but they would require an unprecedented reversal of fertility trends.
What kinds of worlds bounce back?
Population recovery after a mass-mortality catastrophe is far from assured. In the most optimistic scenarios, where humanity is able to quickly raise living standards, we can expect fertility rates to rapidly return to their current trends (2.3 and falling). It seems likely that such scenarios will lead to population peaking far lower than their current projected maximum of around 10 billion (UN, 2022[2]). In this section we explore whether the worlds in which population does recover are “less worth saving”.
Drawing from the previous section, the kinds of worlds that bounce back are more likely to have one or both of the following properties:
A long period of slow growth
A prolonged population boom
Property (1) is fraught with risk. Persistent slow growth has only been observed in pre-transition countries where fertility rates are high and life expectancy is low (see OWID, 2019). Long, slow growth would very likely mean that the catastrophe was so severe that it was not possible to immediately begin rebuilding society, and therefore that subsequent generations were unable to preserve basic scientific knowledge (such as germ theory) or pre-catastrophe culture (women’s rights, investment in children’s education). There is no guarantee that such a preindustrial society would be able to rekindle advanced civilization (Mokyr, 2016), especially with disadvantages such as an altered climate and depleted fossil fuel deposits (Dartnell, 2015) and so eventual extinction seems to be a likely outcome.
It is possible that (1) emerges from a culture that is “in control” of population growth. Perhaps the survivors feel a renewed sense of common responsibility to repopulate the earth in a gradual, measured way. This would be unprecedented—the nearest modern-world analogue is Israel, where fertility norms have been unusually resilient to cultural changes.
Property (2), on the other hand, describes situations where the post-catastrophe demographic boom is enough for full population recovery. This requires a considerable period of high fertility and low mortality. Low mortality is likely, as the biggest contributors to reduced mortality have been relatively cheap, low-tech innovations. Sustained high fertility rates are likely to be the product of civil strife and faltering economic recovery—like Nigeria, whose civil war, economic collapse and periodic ethnic conflict may have contributed to a stubbornly high fertility rate.
In short, the scenarios that lead to full post-catastrophe population recovery are those in which human fertility behaviors are radically reversed, either due to persistent low living conditions or to a shift in values. These scenarios are likely to be more disruptive than scenarios in which the population does not fully recover, and we can be less certain about the types of societies that will emerge from them.
The second stage of this report explores the consequences of population not fully recovering after a catastrophe.
What are the consequences of population peaking lower?
With the exception of Israel, every wealthy country has below-replacement fertility levels. If current trends continue, global population will decline precipitously in the 22nd century. Eventually there will be too few of us to maintain a modern society.
Fortunately we have time to respond. Culture could adapt in a way that enables sustainable reproduction. Technological advances could make procreation much easier.
But after a major catastrophe in which population never recovers to its current levels, there will be far less time. We can be less confident that we will overcome demographic decline. And the worlds that do overcome it are likely to be culturally very different from the world today.
Pathways to security
I model the interaction between states according to the diagram below. A “mature” society is one in which mortality and fertility rates have decreased far enough to cease population growth. I assume that after this, global population will begin to decline until (a) industrial civilisation can no longer be maintained or (b) we find some way to stop the population decline.
The key question in this section is to what extent will lower peak population make it harder to avoid demographic decline? I will also be asking what kinds of worlds manage to avoid demographic decline after peaking lower?
I distinguish between two possible end states: “Existential security”, which I take to be equivalent to attaining multiple independent interstellar colonies (Arepo, 2022), and “Extinction”.
This section will focus on near-existential catastrophes in which the survivors are immediately able to begin rebuilding society.
The risk of terminal population decline
Even without a near-existential catastrophe, terminal population decline is a significant threat. Not only will it be harder to reach for the stars with a falling population; establishing interstellar colonies will be futile if fertility rates remain below replacement levels. In order to survive in the long term, our civilization will have to somehow reverse an established demographic trend: as living standards improve, fertility rates fall.
This problem is not seen as a top priority within longtermism—perhaps partly due to the assumption that AI or other technology will one day make controlling population growth much easier. But there are reasons to be pessimistic about this, especially in post-catastrophe scenarios where populations peak lower.
Let’s take a look at some of the ways we might overcome terminal population decline. Much of what follows is inspired by a working paper from the Population Wellbeing Initiative* (Kuruc, K and Weston, G. How large is the value of humanity’s future? 2023) which explores long-run population trends in the absence of technology that can create digital people. To plug the gap I add in some of my own analysis on digital people.
Ways we could overcome terminal population decline—and some reasons to be pessimistic about them:
Space colonies. Fertility is low in wealthy countries with large unsettled territories (Canada, Australia), even though they are far more hospitable than other planets. There is no reason to think that space colonies alone will reverse the fertility decline.
Artificial wombs. Pregnancy is only a small part of the cost of having children, so artificial wombs may not significantly shift procreative decision-making. Past advances in assisted reproductive technologies seem not to have done much to increase fertility (Lazzari et al, 2023).
Government interventions. Hungary has shown the most success in increasing fertility, going from 1.2 to 1.6 in the decade after 2011 (World Bank, 2023). But Hungary is not on track to eventually reach replacement rates, despite the costly incentives. Even if all of the fertility increase is attributed to policy, which it probably shouldn’t be (IFS, 2018), it seems unlikely that government incentives alone will ever turn the tide.
Heritable Fertility. Inevitably, high-fertility subgroups grow faster than the general population. It has been suggested that this ensures the demise of secularism (Kauffman, 2010), and it seems to imply that eventually most of the global population will belong to a high-fertility subgroup. Unfortunately, the theory fails to account for the fact that fertility is not fully heritable, and that subgroups tend to become less extreme as they grow and assimilate with the wider population.
Change in values. If childrearing becomes more prestigious or more desirable relative to seeking career success, more people will choose to have children. This could happen in a future where most “work” is automated, or where attitudes towards women change significantly. This would be a significant reversal of the trends we have seen in the past, where both women’s rights and the cultural capital attached to careers tend to increase as living standards improve.
Life extension. Longer lives would give people more time to pursue multiple life goals like raising children and building a career. However, this would likely be ineffective without parallel breakthroughs in reproductive technology.
Digital people. Non-biological reproduction would be much easier to control and may only be limited by constraints on resources like energy or computing power. If digital people are possible, and are developed in time, we could end our reliance on biological reproduction.
Decline to a pre-industrial state. Population levels might fall so far that global trade, industrial production etc. become infeasible. The resulting societies would likely have much higher fertility and would begin to grow again. This process seems likely to act as a buffer from complete extinction.
The above paints a concerning picture, and it looks even more concerning in the aftermath of a near-existential catastrophe. We have relied on population growth to maintain levels of technological progress and economic growth in the past century, and cresting population brings the threat of stagnation in the present one (MacAskill, 2022). In futures where the population peaks in the low billions, there will be less time and fewer people available to find a fix.
In light of this, the main scenarios in which we overcome fertility decline either (a) involve a major sift in values, or (b) are harder to achieve in a world with much lower population:
Governments impose effective pro-fertility measures. Given that governments are not currently able to bring low fertility back to replacement levels, this scenario likely means that governments either have much more relative power than they do today, or that they have much stronger incentives to boost fertility. Increased government power is closely linked with authoritarianism. Stronger incentives to boost fertility could come from a more far-sighted society seeking to avert long-term decline, or from competition & conflict with other governments.
Post-catastrophe society develops pro-fertility norms & values. Modern-day high-fertility subgroups are predominantly religious (the Hutterites and the Mormons in the US, the Haredi in Israel) and have prescriptive views on women’s roles. If this was the source of pro-fertility norms, the consequences for civilization’s long-run values could be bad. It is possible, however, that pro-fertility norms arise from an increased sense of collective responsibility for restoring global civilization, or from “the end of work” in an age of automation.
Life extension & reproductive breakthroughs. The very possibility of radical life-extending treatments is questionable, and breakthroughs seem less likely in a lower-population world with fewer researchers.
Digital people. Creating digital people may not be possible. If it is possible, it will be more difficult to attain in a future world with lower population.
Decline to a pre-industrial state. All known pre-industrial societies have had high fertility, so we should expect this to be the case in a near-empty world that cannot sustain industrialization. It’s possible that this society could “grow back” and one day attain existential security—but there will be plenty of opportunities for things to go wrong.
To summarize, the worlds in which a low-population post-catastrophe society overcomes terminal population decline are often associated with authoritarianism, conflict or limited roles for women. Some of these worlds arise from cultures of long-term thinking and collaboration, and others from technological innovations like life extension or digital people. But overall, a global catastrophe would make it less likely that we would find a technological fix.
Implications
This report aims to use population dynamics to increase our understanding of the consequences of near-existential catastrophe. It does not claim that population dynamics are the only, or even the most important, consideration.
Nevertheless, this analysis leads to some concrete implications for cause prioritization:
We should update the value of preventing near-existential catastrophes, since they appear to both increase the risk of extinction and to increase the risk of future worlds with values very different from our own.
Interventions should be partially judged by the effect they have on our chances of eventually overcoming terminal population decline—either by creating/preserving life[3]*, or by increasing the chances of finding a cultural/technological “solution”. Optimism about technological fixes for demographic decline seems unjustifiably high, and optimists should do more to argue their case.
Conclusion
This report aims to use population dynamics to give a fresh perspective on how to weigh near-existential catastrophes against full-existential catastrophes.
We found that it is likely that population would not recover to its previous trajectory in the aftermath of a mass mortality event. The scenarios in which population does recover are often associated with a radical change in fertility culture, or a slow and faltering recovery. These worlds may be worth saving, but appear to be riskier than the world which we currently inhabit.
Terminal population decline may become a serious problem in the coming centuries, and the pathways to averting it are less reliable than many assume. The problem of terminal decline becomes even more acute in worlds with a lower population, as there are fewer people and less time to find a solution. Again, the worlds that do make it through give cause for concern: most known high-fertility cultures are associated with some combination of low living standards, limited women’s rights and poor education.
Population dynamics suggest that a near-existential catastrophe would introduce existential risks and threats to long-term value that were previously not considered. Hence we should update the value of preventing near-existential catastrophes.
Note that Rodriguez published this work in 2022 with the disclaimer “This is a rough draft I wrote between October 2019 and April 2020. It’s incomplete, and doesn’t reflect updates in my views in the 2+ years since I worked on it.”
The UN population projections after 2100 are out-of sync with the trends that we have seen so far in the wealthiest countries. The projections assume that global fertility levels will stabilize around the rate of replacement. However, while there is evidence for a slight uptick in fertility among the wealthiest countries, none of these “uptick” countries have seen their fertility rates reach replacement levels. Hence the UN’s projections of eventual stability seem difficult to justify.
Although pursuing population increase is probably not competitive with other longtermist interventions (Siegmann & Mota Freitas, 2022)