It is a bit tricky to balance writing new posts and updating old ones. As this library of societal collapse research grows, there is an ever larger backlog of posts that would like to see some updates. Even though the update posts usually have a readership equal or even slightly higher than the other ones here, I do think that having regular completely new posts is important to keep things interesting. So, I will continue to try to find a balance here, but if any of my readers have a strong preference here, I’d be curious to hear it. Also, I recently was on a podcast, where I summarized many insights from this blog here. Check it out, if you would like to see a bit more conversational version of this living literature review. But now, the actual updates.

Subscribe now

What could go wrong?

This post is meant to give an overview of what all kinds of academic and political institutions think constitutes a global risk. Basically, it is a collection of collections of catastrophes. To this collection I have added two new reports. The first one is by the UK Ministry of Defence, (2024).

Such kinds of catastrophes are also highly relevant to defense. Therefore, defense ministries also try to map out the different risks their countries might face in the future. An example of this would be a report by the UK Ministry of Defence (2024). They tried to map out what the strategic trends will be until 2055. A strategic trend is not the same thing as a catastrophe, but still the main drivers of trends they identify look eerily familiar to the kinds of events we have covered in this post so far. More specifically, they identify six main drivers of global change:

• Global power competition

• Demographic pressure

• Climate change and pressure on the environment in general

• Technological advances and connectivity

• Economic transformation and energy transition

• Inequality and pressure on governance

They assess that we are in for a wild ride, because many of these trends push the world in different directions. For example, economic transformation and technology will connect the world at an increasing rate, while global power competition will lead to fragmentation. Or similarly, that more and more people globally feel empowered to speak up and mobilize, while at the same time autocratic governments seem to be on the rise.

The risks and trends are also assessed on their impact and uncertainty. Here AI comes out on top for causing the most uncertainty, while global power competition is assigned the position of the most impactful one for the state of the world. As they assess global power competition as the most impactful one, much of the rest of the report is trying to map this out in more detail. They think that the United States will remain the most powerful nation, but will face increasing competition from China, Russia’s future depends on the outcome of the war in Ukraine, India will play a major role, but struggle with internal unrest and many of the middle power will band together, to have a chance in a world of weakening international agreements.

A second report I have added is by Graham et al. (2025), which translates the methodology of the Swiss national risk assessment to Australia:

What has been encouraging here is that good national risk assessments can work as an inspiration for other actors. For example, in a recent report by Graham et al. (2025), a group of researchers used the methodology of the Swiss national risk assessment and transferred it to Australia and showing that even under traditional risk assessment (accounting for both likelihood and magnitude of the consequences), the annual expected impacts of some global catastrophes are far higher than many of the more frequent threats (like floods or fires) and thus Australia should invest more to prevent damages from global catastrophes.

End time economics

What are the economic consequences of global catastrophes? This is generally quite hard to assess, as we are often lacking the data to even calibrate our models on. However, there is still some historical data that can be used. One paper that has accumulated such a dataset is by Blouin et al. (2024):

A paper that looks more systematically into the impacts of industrial destruction is by Blouin et al. (2024) (Disclaimer: I am a co-author on this one). The general idea is that, while we cannot easily assess the damages a nuclear war might do to industrial output, we can look into history, to find case studies of how destruction of industry leads to declines in industrial outputs and extrapolate from this. These case studies reach from major hurricanes to the Second World War (Figure 1).

Figure 1: Relationship of incapacitated industrial infrastructure and subsequent decline in industrial output. Based on historical case studies and modelling studies.

In addition to establishing this relationship between industrial destruction and declines in output, the study also assesses how much of industrial capacity would be destroyed in US/Russia and India/Pakistan conflicts. This can be estimated from plausible target lists and the destruction caused by the nuclear weapons detonating above those targets. In the case of the US/Russia conflict the estimate is that this could destroy roughly 3 % of global industrial capacity, which would translate to a 24 % loss in global industrial output from the detonation damage alone. This strongly nonlinear relationship between destroyed capacity and declining output highlights the vulnerability of economies to large scale disruptions.

Systemic risk and the polycrisis

The ongoing polycrisis and systemic risk in general are hard to pin down. This is not also true in general, but even more so the case when it comes to policy making. Still, there are some attempts to think about how even such complex topics might be governed. One of those attempts is by Studzinski et al. (2025):

Coordinating around such complex problems is quite difficult. Where even to start? One thing that is clear, that global coordination is needed one way or the other. An attempt to get this started was the United Nations’ Summit of the Future. The summit was meant to start the reform of the United Nations 75 years after their founding and resulted in the Pact for the Future, which spelled out 56 goals which should guide the way to tackle the existing global challenges. How well this worked out and what we might do better in the future is discussed in Studzinski et al. (2025). They start with the criticism that the Summit of the Future did have a too narrow view on individual hazards and not really the bigger picture that would be needed to address the challenges we face. But the world as it is, is faced by threats that are larger than the sum of their parts. They are interconnected and reinforce each other.

The inability of global governance to effectively address the current global problems is caused by a variety of factors. One of them is that the situation is just tricky in general, and so it is to be expected that it cannot be solved easily and quickly. But also many of the existing institutions for global governance have ambiguous aims and missions, are often working against each other or are trying to find the least disruptive policy, in an attempt to please everyone. All in a world where international agreements are taken less seriously and the funding for global governance is decreasing.

But this does not mean that such summits are a waste of time. Ultimately, events like this are one of the few avenues of global governance that currently exist. Future summits need to get rid of their single hazard focus and instead focus on the interconnected problems we actually face. In addition to that, regional organisations like the European Union or the Shanghai Cooperation Organization could step up their game and implement measures to tackle systemic risks in their own borders, to showcase how this might also work on a global level. None of this is easy and likely needs a paradigm shift on how systemic problems are viewed, but getting this right is essential for the coming decades.

Economic inequality and societal collapse

One big factor that has come up in many of the papers I have been reading for this blog is inequality. It is just generally corrosive to societies if it gets too high. A recent paper by Alfani et al. (2025) does a comparative analysis of inequality in the Han, Roman and Aztec Empires:

Another large comparison of how inequality plays out in comparable polities is by Alfani et al. (2025). They created a dataset to compare inequality in the Han Empire and the Roman Empire. In addition, they compared both to a previous dataset from the Aztec Empire. Thankfully, for us the Roman and Han Empire were quite good at record keeping. While comparable on many axes, the Roman Empire tended to be less centralized and more self rule in the provinces.

Alfani and colleagues calculated the wealth of the different regions by the rate of their urbanization and population density, based on the assumption that more and larger cities and a higher population meant a province was richer. This showed that for the Han Empire the central province was quite rich, while the other provinces were much less so. In comparison, the Roman Empire had also a very rich central region, but also several other regions (e.g. in North Africa) which had considerable wealth. This is likely shaped by how both states used their military and bureaucracy. The military is usually deployed in frontier regions, meaning that wealth gets transferred there from the center, while the bureaucracy tends to accumulate in the central region, syphoning away wealth from other regions. The Han Empire had small military and large bureaucracy, while it was the other way around for the Romans.

In addition, to this inequality between regions, the paper also calculates the inequality in the population in general. They do so by tracking the size of the different groups of people in the empires (like peasants, aristocrats, merchants, etc.). This allows them to calculate a Gini index for the Han, Aztec and Roman Empire and compare it to the present day United States (Figure 1).

Figure 1: Comparison of income shares in the Roman, Han and Aztec Empire and the United States.

This highlights stark differences between the different empires, with the Aztec Empire being the most unequal one by far, followed by the Han and then the Roman Empire. The United States has not yet reached the levels of inequality of those past empires. However, what is notable, that relatively speaking the poorest 10 % in the United States are poorer than their counterparts in all the three other empires.

These differences in inequality are also mirrored in differences in stability. From those empires considered, the one with the highest inequality (the Aztecs) was the most unstable one, while the one with the lowest inequality (the Romans) was most stable. Obviously, these are only three data points and thus not super reliable, but it fits in with the general trend of the instability-inducing power of inequality.

Trade collapse

This post is about the importance of trade for the modern world and what might happen if it gets disrupted. I added a short paragraph to discuss a paper that looks into the importance of ports:

In addition to these chokepoints in general, we can also see that ports in general differ a lot in their importance for global trade. A paper by Verschuur et al. (2022) analyzed the world’s 1300 most important ports for global supply chains. Based on a large supply chain database, they tracked how many goods flow through all these ports. They find that in value terms, around 50 % of global trade flows through ports. Also, the ports differ wildly in their importance with some major ports like Antwerp, Los Angeles or Singapore having a highly important role in global trade flows (Figure 2). This means if any of those ports would be disrupted, it would likely have wide reaching ripple effects.

Figure 2: Globally important ports for imports and exports.

Share your favorite

Thank you for your interest in this academic experiment. Here’s a great song as an appreciation.

If this post (or others here) got you thinking, consider leaving a like. It helps more people stumble onto this blog.

As you made it this deep, I assume that you like the stuff that is written here. If you want to do me a little favor, I would really appreciate it if you could share your favorite post with someone who might enjoy it.

Go ahead, you can do it right now. Nothing more to read in this post anyway.

Much appreciated!

Until next time

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me an email to existential_crunch at posteo.de and we can schedule something.

Subscribe now

References

• Alfani, G., Bolla, M., & Scheidel, W. (2025). A comparison of income inequality in the Roman and Chinese Han empires. Nature Communications, 16(1), 3248. https://doi.org/10.1038/s41467-025-58581-0

• Blouin, S., Jehn, F. U., & Denkenberger, D. (2024). Global industrial disruption following nuclear war. EarthArXiv. https://doi.org/https://doi.org/10.31223/X58H9G

• Graham, J., Boyd, M., Sadler, G., & Noetel, M. (2025). Mapping Australia’s Risk Landscape: A Comparative Analysis of Global Catastrophic Risks and Traditional Hazards (SSRN Scholarly Paper No. 5253625). Social Science Research Network. https://doi.org/10.2139/ssrn.5253625

• Studzinski, N. G., Kent, R., & Korowicz, D. (2025). Towards the Governance of Global Systemic Risk: Reforming the Summit of the Future. Global Governance: A Review of Multilateralism and International Organizations, 31(2), 113–136. https://doi.org/10.1163/19426720-03102002

• UK Ministry of Defence. (2024). Global Strategic Trends: Out to 2055.

• Verschuur, J., Koks, E. E., & Hall, J. W. (2022). Ports’ criticality in international trade and global supply-chains. Nature Communications, 13(1), 4351. https://doi.org/10.1038/s41467-022-32070-0

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

Existential risk is not something static which lives outside of humanity. It is something humans shape. This can be directly, such as things that are created or deployed by humans like nuclear weapons. Or it can be indirectly, like whether or not humans have sufficiently prepared our infrastructure for major volcanic eruptions. This means that human actors can and do increase the hazards we face or worsen our vulnerability to existing hazards. But not everybody has the same influence on this. My neighbours do not have any influence on the question if there will be a nuclear war between Pakistan and India. The power and resources to influence such events is usually concentrated in very few groups (1). But who might want to destroy the world? Who might have the capacity to destroy the world, but no motivation to do it? How are we vulnerable to these risks? And what happens if catastrophic capabilities and motivations of an actor align?

Subscribe now

Individual bad actors and the “Doomsday Button”

Two major early works on agential x-risk were Torres (2018a) and Torres (2018b). They try to catalogue all the kinds of agents (individuals or groups) that could have the potential to destroy the world, either on purpose or by accident. They highlight that there should be increasing global concern about those who express omnicidal beliefs, especially when faced with increasing accessibility of dual-use emerging technologies, meaning technologies with potentially high destructive potential like AI or biotech. Also you never know what future scientific breakthroughs have in store, there is always the chance that one day a technology might pop up which is super easily accessible and super destructive (2). The only chance we might have in such a case is to try to understand the actors who might want to use such weapons.

If they had the chance, who would willingly commit “omnicide” - literally, the act of killing everyone? Torres (2018a) develops a framework to assess these kinds of actors. The idea here is that the methods by which bad actors try to destroy the world is less important than their motivation. If their motivation is understood, there is a higher chance to stop them. The first insight here is that this means that we can exclude most conventional terrorists. For example, rebels fighting against an authoritarian regime in their home country do not have an incentive to use weapons of total destruction. If your goal is to free the people in your lands, it does not really make sense to kill them. This is a bit abstract, but Torres (2018a) suggests a simple thought-experiment to determine who we should consider apocalyptic agents: Imagine there is a button you could press, which would instantly kill all humans. According to the paper, every person who would push this button is an apocalyptic agent.

This means we are left with people who think that the current state of Earth (or even the universe) is so irredeemably bad that the only reasonable thing to do is to kill everybody. While this is certainly a very extreme position, we know from history that such people have existed. Torres outlines six examples of ideologies whose adherents might press the “Doomsday Button”. While Torres (2018a) gives the framework, Torres (2018b) provides the examples to fill the framework, so let’s combine the two here:

Apocalyptic terrorists

These are people who believe in some kind of coming apocalypse, would like to see it sooner and contribute to this end if they can.

An example cited here is James Ellison. He was the leader of “The Covenant, the Sword, and the Arm of the Lord”, a far right, anti-government militia in the United States. While they started as a “normal” Christian fundamentalist cult, Ellison formed them into an active terrorist group. He had the idea the Second Coming of Christ would only happen after an apocalyptic race war between white Europeans and everybody else. This means that if he would be able to start a world destroying race war, Christ would return sooner. Ultimately, the group was not really successful in attaining their goals, but one could easily imagine them pushing the “end of the world” button (3).

Misguided moral actors

“Misguided” in the context of the Torres papers refers to actors who do not subscribe to the long-term transhumanist consequences of classical utilitarianism. This means especially believers in negative utilitarianism, who subscribe to the idea that only the prevention of negative experience matters. Therefore, any world which includes suffering should be avoided. As all worlds contain suffering, none should exist. Or said more simply, dead people cannot feel pain.

Fortunately, this quite extreme view is only held by some philosophers tightly locked away in academia. However, nihilism cults like ‘efilism’, the anti-natalist ideology behind the 2025 bombing of a Palm Springs fertility clinic, have surfaced more in recent years, but they at least seem to be sticking with conventional weapons for the moment.

Radical ecoterrorists

These are people who think that Earth (or Gaia) would be better off without humans. We had our chance to live in peace with this planet, we spoiled it, so we should just be extinct and let this planet find its way back to equilibrium again.

Examples of such ideas in the wild are the fringes of the deep ecology movement. One of the more well known groups affiliated with these ideas is the Gaia Liberation Front. They state that their mission is to extinguish all humans, as we cannot be trusted with Earth. To bring this about they think that a pandemic that can only target humans would be most well suited, because other methods are too impractical (like mass suicide) or too destructive (like nuclear war).

Idiosyncratic actors

This category includes ‘lone wolves’ or fringe cells. They can be part of loose communities online or evolve organically alone, and often pull from a personal or eclectic range of ideologies. We can find many such examples in the US school shootings epidemic, and a particularly famous one is Eric Harris, who was the shooter at the 1999 Columbine High School massacre. In his diaries he explained that he wanted to see the world burn and kill mankind. Fortunately Eric Harris did not have the means to see the world burn, and the ‘lone actor’ concept has received a lot of counter-terrorism attention over the last few decades when it comes to dangerous weapons.

Non-human actors

Besides these actors where we can give historical examples, Torres also argues that there are two further, more hypothetical groups: value-misaligned machine superintelligence and belligerent extraterrestrials. But for obvious reasons we can only speculate about the ideas and incentives for such actors.

What might bring people to press the Doomsday Button?

Besides Torres, there are other people who have looked into such feelings in the general population. In other words, what might cause ordinary people to feel motivated to press a doomsday button if one existed? There will always be insane people, malignant narcissists and frothing zealots to account for, but wider systemic problems can raise the net level of extreme beliefs in the general population who otherwise wouldn’t hold them, which might increase levels of terrorism without proper social support.

Based on data from British elections, how people change their beliefs when they feel threatened is explored by Stevens & Banducci (2022). They show that growth spikes in extremism might be correlated with feelings of threat, both in normative and personal spheres. Personal threats include things like feelings of existential insecurity. Normative threats can include climate change (mentioned by Torres in the papers described above), feelings of cultural erosion, or increasing authoritarianism. Apocalypticism and doomsday sentiments have historically seen greater proliferation during times of crisis, such as the Cold War. Feelings of existential insecurity might mean that a greater-than-average number of people feel increasingly drawn to beliefs that the world is better off destroyed or fundamentally altered than it is continuing along its current trajectory. However, issues arise if we rely too heavily on the idea of “citizen terror”, which we shall explore now.

Problems with the “individual actors” framing

So far we have explored in great detail what has and what might motivate individual actors or small groups to believe in - or even to try to cause - omnicide. However, this does not mean that this is the most likely way that omnicide or global catastrophe might be caused by humans.

There is a recurring theme in existential risk literature (for example the work by Torres explained above) that strongly emphasizes the risk of terrorists or single actors causing a global catastrophic event. However, this argument lacks somewhat, as there might be actors who theoretically might want to destroy the world, but they clearly lack the resources to do so and it seems plausible that they won’t get them anytime soon. In addition, in the last decades the world has seen massive investments into international CBRN counter-terrorism work, further decreasing the chance that a splinter group will have the chance to end the world. Finally, the question is: would most of the people holding the beliefs outlined above actually be able to extinguish all of humanity?

In general, it seems pretty hard to find people who would want to destroy the world for real and also have the capability to do so. Kallenborn & Ackerman (2023) thoroughly went through the literature in an attempt to find such groups, but weren’t able to find any that tick all the boxes. Groups usually have the motivation to destroy the world, but not the capabilities for a direct attack. Though they think it is plausible that such groups might exist in the future, focusing on building resilience and addressing societal vulnerabilities is the best way to lower the risk. They also highlight that while there currently is no terrorist group which could destroy the world, there have been groups that had the will and capability to spoil risk mitigation measures, but got their timing wrong. This means, currently there is no ongoing risk mitigation measure for global risk, but for example, if there would be an asteroid on the way to Earth, terrorists would likely be able to sabotage a mission to deflect it from hitting Earth.

The problem of making apocalyptic terrorism the center of agential risk focus is explored in Meggitt (2020). The category of “apocalyptic terrorism” is deceptive and should not be attributed lightly as it leads to false positives. Terrorism analysts have shown that ‘apocalypticism’ is often flexible and debated within terrorist groups like ISIS, used as a strategy to gain bargaining power, or to boost recruitment or membership retention in hard times; its emphasis waxing and waning relative to the group’s fortunes.

An interesting case study is presented by Juergensmeyer (2003). He explored the motivations of members of Aum Shinrikyo, a group often used in x-risk discussions as an example of genuine omnicidal intent and follow-through. After conducting interviews with multiple cult members after the Sarin gas attack, both free and in jail, Juergensmeyer concluded that despite their apocalyptic rhetoric, their goal was not to annihilate humanity or cause a GCR. Instead they believed that their organisation would help them survive an apocalypse caused by external nuclear forces. Their actual motivation was being a part of those survivors that would rebuild the world after the apocalypse and that they would hold power in Japan. Their apocalyptic framing served them for meaning making, but was not their actual goal. There is no great evidence that they really wanted to cause human extinction.

Since many of Torres’ papers were written, the field of existential risk has benefited from the inclusion of researchers with counter-terrorism, biosecurity, politics and CBRN expertise, as well as simply having had more time to consider the multiple hazard and high vulnerability areas of this risk. While recognising the importance of monitoring dangerous ideologies in numerous but incapable actors, existential terrorism has moved away from this narrow focus and is starting to focus more on systemic vulnerability risk factors and their exploitation by malicious actors. Kallenborn and Ackerman (2023) advocate for resilience-strengthening solutions as well as monitoring on the agents side. Meanwhile, examinations have also grown to include other agents, capabilities and systemic vulnerabilities too.

Institutional bad actors

It seems far-fetched that an isolated individual could destroy the world today, even if they had ideological motivations to do so. However, institutions which endanger the world already exist. Not only should we ask “who might press the doomsday button?”, but also “who might build it, and should they?”.

Institutions are made up of networks and groups of people, have access to greater resources and legal protection to build capabilities, and often use systemic vulnerabilities to their advantage. Understanding these institutions, their capabilities and the wider societal vulnerabilities at play is integral to building a fuller and more accurate picture of agential x-risk.

A comparison of threats from malicious actors at different scales is done by Sandberg & Nelson (2020). They developed a simple model for this in biorisk (Figure 1). The idea is that if you want to be a successful bioterrorist, there are many steps you have to do right. This starts from having the insights of how you want to kill everyone, implementing and testing the idea, evading capture, and scaling the enterprise to global proportions. Each of these steps has a specific set of difficulties, each of which has to be overcome in order for an actor to progress to the next step. Overcoming these difficulties often requires having resources and progress is harder the fewer resources you have. Due to these constraints, their model comes to the conclusion that, even though less powerful malicious actors are more numerous, it is more likely that less numerous but more powerful actors would be able to overcome more difficulties and therefore progress through the entire risk chain. Though Sandberg and Nelson also acknowledge that this might change in the future (if these steps get considerably easier as technology progresses), the results of their model indicate that most current risk from malicious actors is enabled by their place inside resource-centralising institutions.

Figure 1: The model of which steps you have to master to be able to cause global biorisk.

This scary possibility is one that CBRN, de-radicalization and biosecurity experts have devoted huge amounts of time and resources to minimizing: a high-powered actor becoming sufficiently motivated to deploy the catastrophic capabilities they are already in possession of, and no-one around them being educated to notice or report it until it is too late.

But what happens when institutions’ very incentives are to build these capabilities and keep them secret? A highly destructive, highly accessible weapon which might be discovered in the future is, most likely, not going to come from a DIY amateur biolab, or a single genius researcher. It is much more likely to come from a group of people working in an AI development lab, biosecurity facility, or defence-tech company. Their investors, another group of people with highly centralised resources, may have a strong incentive to market, use or sell this new invention. Because of this level of resource and power centralisation, it’s no exaggeration to say that, in today’s world, the agential threat posed by terrorist groups is dwarfed by that posed by Peter Thiel or Sam Altman, enabled by the institutions that they operate within and the systems vulnerabilities those institutions exploit.

Specifically focussed on the institutional actors is a short overview article by Kemp (2021) (4). In it Kemp argues that existential risk is mainly driven by big, institutional actors, as everybody else has too few resources to meaningfully move the needle. The pathways that have been identified so far to end the world are things like artificial general intelligence, catastrophic biological threats, climate change, lethal autonomous weapons systems and nuclear weapons. You cannot build nuclear weapons in your backyard, neither can the lifestyle of an individual meaningfully influence how many ppm carbon dioxide are in the atmosphere.

To do these things you have to have lots of power and lots of resources, and lots of time to work on them without being stopped.They are something only very large actors like the military-industrial complexes of several countries, the fossil fuel industry, or Big Tech can enable or accomplish. In many (or maybe all?) of these institutions we can see that their profits are privatized, while the potential downsides are public.

Let’s look at some examples:

• Artificial general intelligence: There are very few AI companies who actually build powerful new models, and the resources needed to build such models are highly concentrated in a few groups of people. Presently the biggest ones are OpenAI, xAI, Anthropic and Deepseek. It is unclear if their models even could lead to general AI, but what is clear is that it is extremely resource-intensive and therefore difficult to work on them, because they need large amounts of electricity and compute to be built and run.

• Climate change: Globally around 100 producers are responsible for the vast amount of carbon dioxide emissions. Similarly, more than half of the global cumulative emissions fall to just three main actors: the US, EU-27 and China. Finally, only six countries hold 80 % of the global fossil fuel reserves.

• Nuclear weapons: Almost 90 % of all nuclear weapons are held by two countries: the United States and Russia. And for the remaining 10 % you just have to add 7 other countries. The maintenance and production of American weapons is bundled up into just 28 companies. The command chain to use those weapons are quite short. In the United States the president has sole authority to order a nuclear strike. There are four key advisors, but their consent is not required.

All this means that most human-made catastrophic hazards are highly concentrated in small networks of powerful actors. The United States military could easily be considered as the largest contributor to global risk. They emit gigantic amounts of carbon dioxide, they control the most advanced nuclear arsenal, are at the forefront of developing autonomous weapons and surveillance technologies and have by far the largest military in the world.

What makes such actors so tricky to reign in is that they often use their power and resources to distribute misinformation, so public opinion is swayed and they can continue their mission. Clear examples here would be NATO trying to discredit nuclear winter research (5) or the fossil fuel industry financing large misinformation campaigns for decades. Similarly, these powerful organizations often lack efficient oversight, which allows them to keep risks secret. Like the fossil fuel industry knowing about the fact of global warming for decades.

In the future, some of those pathways to global catastrophes might get into the reach of individual actors, but for now it seems pretty obvious who creates global risk and we should focus our efforts on those actors. The motivation of these actors is likely not to ruin the world, but we cannot get around the fact that they do endanger us all. They are driven to accept these risks because they have a profit motive and face competitive dynamics.

There are also attempts to combine these two strands of ideas focussed on individual and institutional actors. A report by Althaus & Baumann (2020) argues that we should also look at malevolent individuals in positions of power. They argue that some of the worst things ever done to humans were caused by dictators, who had large institutional power behind them. Clear examples of this would be Hitler or Stalin. These individuals likely did those horrible things because they had power, but also because they scored high in Machiavellianism, narcissism, psychopathy and sadism. These are often properties that individuals often have who are power and status seeking and also harbour extreme beliefs. Therefore, existential risk mitigation should also be done via exploring these traits, how they can be detected and how people with these traits can be excluded from ever gaining power. Something like a similar screening process is already done in counter-terrorism, for example in high security biolabs. If more scrutiny can be directed towards actors already operating within powerful institutions, agential x-risk might be significantly lowered via this type of intervention.

Policy around global risk

The concentration of global catastrophic risk in the hands of relatively few powerful institutions raises a critical question: how do those responsible for governing these risks actually perceive this challenge? Understanding how policymakers navigate this landscape of concentrated power is essential, as they represent one of the few potential checks on these institutional actors. However, policymakers themselves face significant constraints. Research by Nathan & Hyams (2022) provides direct insight into how those closest to global catastrophic risk governance perceive their own capacity to address these challenges, revealing both the opportunities and structural barriers that shape policy responses to our most dangerous institutional actors. They interviewed 16 civil servants, civil society group representatives and individuals from the private sector in the UK, US or Europe and tried to distill these interviews. This brought up four themes on how these people think about global risk:

• Scepticism: They realize that current solutions are often lacking and that the government is too often just in a reactive role. New systems are needed to plan before disaster happens. This skepticism directly reflects the challenge of regulating the powerful institutional actors identified earlier. As one UN policy lead put it regarding pandemics: “if a crisis hits tomorrow, it’s too late.”

• Realism: Global catastrophic risks are just a very complex problem, which often requires long times to implement solutions. This is just really hard in a policy world which is focussed on the short term. As one health security expert noted, there’s “realpolitik because you can have as many agreements as you want, but if you’ve got the UK and the United States producing something, politically, their first obligation is clearly their own citizens.” This reflects the challenge of governing institutional actors whose influence spans national boundaries while policymakers remain constrained by domestic political pressures.

• Influence: Individual policymakers feel like they can be a force for good when they do good work on global catastrophic risk. For one to implement potential solutions, but also just simply by making the topic something that can get more easily discussed. One interviewee emphasized the importance of formative conversations with “people a lot brighter than me who are maybe going to end up in a sovereign wealth fund, maybe advising the government of Saudi Arabia or Abu Dhabi… If their framing of the world is just really, really slightly different and then they’re making a decision that relates to tens of billions of dollars being allocated in a particular area rather than on another.”

• Governance outside government: The idea here is that governance also happens a lot outside of government. This means we should not only rely on the policy world, but also work on creating more accountable governance structures in other places like academia or companies. As one expert on emerging biotechnology policy stated: “where I find the most optimism is where I dream of creating supplementary spaces that create bridges between regulation and scientists that sit just under regulatory decision-making that augment the process.” This reflects recognition that traditional state-based governance may be insufficient to control the institutional actors driving global catastrophic risk.

The research reveals a fundamental tension: policymakers understand the risks posed by concentrated institutional power, but feel constrained by the very power asymmetries this concentration creates. They see opportunities for influence through language, agenda-setting, and non-governmental governance mechanisms, yet remain skeptical about their ability to fundamentally alter the incentive structures that drive institutional bad actors.

Conclusion

The takeaway from all this is that large risks from individual actors is something we should be concerned with and prepare for, but the ultimate problem is not them, but some of the institutions which have been built to shape and control parts of society. Their incentives are often misaligned with society and we need more accountability and control over them. Policy makers are thinking about this, but it is just hard. The research shown here highlights that the actors posing the most risk globally are often held unaccountable and even the policy makers who think most about them, are unsure how they can make a difference. This means we need to restructure our system, so accountability exists again. And I might just sound like a broken record, but I think this is again an area where we can see that we have to make our society more democratic, that we have to experiment with new ways to distribute power. It might not fix all of our problems, but more decentralized decision making does lend itself less to unaccountable power centers by definition (6).

This post was reviewed by Mel Cowans. If you enjoy it, or are interested in any of the concepts discussed, feel free to reach out to them on Linkedin.

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

Subscribe now

How to cite

Jehn, F. U. (2026, March 11). Who shapes global risk?. Existential Crunch. https://doi.org/10.59350/zfdhw-d9e42

Endnotes

(1) This is especially the case for nuclear weapons, as the short time span to decide if you want to launch the weapons or not, makes it impractical to include a lot of shareholders. Kurzgesagt has a nice video which highlights this urgency.

(2) A more formal explanation of how this might come about can be found in Nick Bostrom’s vulnerable world hypothesis.

(3) As long as the person pushing the button is white, of course.

(4) Yes, I know that this is a newspaper article and not a paper, but I think it is a pretty good source and also includes lots of references to underline its arguments. You can also find a more detailed account of this in chapter 20 of Luke Kemp’s Goliath’s Curse book, which I summarized in another post.

(5) I have written about this here.

(6) See here for a more detailed argument why I think democracies are essential.

References

• Althaus, D., & Baumann, T. (2020). Reducing long-term risks from malevolent actors. Center on Long-Term Risk. https://longtermrisk.org/files/Reducing_long_term_risks_from_malevolent_actors.pdf

• Juergensmeyer, M. (2003). CHAPTER 6 Armageddon in a Tokyo Subway. In Terror in the Mind of God: The Global Rise of Religious Violence (3rd ed.). University of California Press. https://www.jstor.org/stable/10.1525/j.ctt4cgfbx

• Kallenborn, Z., & Ackerman, G. (2023). Existential Terrorism: Can Terrorists Destroy Humanity? European Journal of Risk Regulation, 14(4), 760–778. https://doi.org/10.1017/err.2023.48

• Kemp, L. (2021). Agents of Doom: Who is creating the apocalypse and why. https://www.bbc.com/future/article/20211014-agents-of-doom-who-is-hastening-the-apocalypse-and-why

• Meggitt, J. J. (2020). The Problem of Apocalyptic Terrorism. Journal of Religion and Violence, 8(1), 58–104. https://doi.org/10.5840/jrv202061173

• Nathan, C., & Hyams, K. (2022). Global Catastrophic Risk and the Drivers of Scientist Attitudes Towards Policy. Science and Engineering Ethics, 28(6), 50. https://doi.org/10.1007/s11948-022-00411-3

• Sandberg, A., & Nelson, C. (2020). Who Should We Fear More: Biohackers, Disgruntled Postdocs, or Bad Governments? A Simple Risk Chain Model of Biorisk. Health Security, 18(3), 155–163. https://doi.org/10.1089/hs.2019.0115

• Stevens, D., & Banducci, S. (2022). What Are You Afraid of? Authoritarianism, Terrorism, and Threat. Political Psychology, 43(6), 1081–1100. https://doi.org/10.1111/pops.12804

• Torres, P. (2018a). Agential risks and information hazards: An unavoidable but dangerous topic? Futures, 95, 86–97. https://doi.org/10.1016/j.futures.2017.10.004

• Torres, P. (2018b). Who would destroy the world? Omnicidal agents and related phenomena. Aggression and Violent Behavior, 39, 129–138. https://doi.org/10.1016/j.avb.2018.02.002

This post was originally an invited contribution to the United Nations Independent Scientific Panel on the Effects of Nuclear War. I’m also sharing it here as an overview of the current state of nuclear winter research (+ related topics) and the open questions the field is grappling with. It was written by me with extensive input from the team at the Alliance to Feed the Earth in Disasters (ALLFED), whose comments and suggestions greatly improved it.

The research around the impacts of nuclear war began with immediate effects such as explosions and fallout. A vast literature exists and has been building since the mid-20th century. The literature around the climatic effects is smaller, partly because the climatic effects were only scientifically recognized in the early 1980s (Turco et al., 1983), but also due to nuclear winter research being partly defunded in the late 1980s and 1990s (Turchetti, 2021) and because general interest declined after the dissolution of the Soviet Union in 1991. There have been some studies in the intervening decades, but scientific attention increased again after several philanthropically funded studies used modern climate models and provided the data for other researchers to build on (Robock et al., 2023). After the Russian invasion of Ukraine, the topic also gained more attention in public and policy discussions (Helfand et al., 2022). The body of research has been growing since then, albeit slowly due to funding constraints, yet many questions around nuclear war and subsequent nuclear winter remain unanswered. Below, I summarize the current state of knowledge, focusing on climate, the food system, and where gaps exist.

The climate aspect of nuclear war has been explored extensively in comparison to other aspects. Beginning with the original nuclear winter study (Turco et al., 1983), the literature progressed through intermediate studies (e.g. Robock et al. (2007) to more recent studies that provided the data for the reemergence of the field (Coupe et al., 2019; Toon et al., 2019). These were used for a set of papers that explored facets of post-nuclear war climate, such as UV radiation changes (Bardeen et al., 2021), the state of the ocean (Harrison et al., 2022), and the emergence of a so-called Nuclear Niño (Coupe et al., 2021). More recently, there has been a push to make this data more accessible to enable further study (Harrison et al., 2025).

What has become clear from nuclear winter climate research is that the climate impacts of a nuclear war would be significant, and would lead to abrupt cooling (beginning within weeks and reaching its lowest values 2–3 years after the war), and global effects (impacting both Hemispheres) (Coupe et al., 2019). Some research disputes the severity of these effects from US national laboratories like Los Alamos (Reisner et al., 2018). However their data and code are not easily available, which prevents independent researchers from reproducing their calculations (Robock et al., 2019).

Constraints on the scientific discussion around climate and nuclear winter research include that the number of studies remains limited and concentrated among a small number of researchers, and many rely on the same underlying modelling data. Moreover, the group around Alan Robock that conducted much of this research has been criticized for overly pessimistic assumptions (Reisner et al., 2019). The debate around this is summarized in Hess (2021). Given these dynamics, a Coupled Model Intercomparison Project (CMIP)—but for nuclear winter instead of global warming—or, at minimum, several attempts to reproduce the climate simulations using other models, could provide a more robust foundation for future research.

Building on the data from nuclear winter climate research, a significant number of studies explore the food system and how it might react to the sudden shift in climate following a nuclear war. For example, Jägermeyr et al., (2020) estimate that even a limited nuclear war would have major consequences for agriculture, primarily due to a reduction in temperatures, precipitation and sunlight. There is also research that studies the impacts of increased UV radiation on agriculture, though this impact seems to be constrained to the most severe scenarios (Shi et al., 2025). Importantly, the food system impacts would not be limited to the combatant nations, nor even to the hemisphere in which the war occurred. In addition to climate effects on agriculture, countries would likely also face a cascading disruption of trade systems (Jehn, Gajewski, et al., 2025) and the countries which are hit by nuclear weapons would see their industrial production plummet (Blouin, Jehn, et al., 2024). Besides many other negative impacts, this would decrease availability of fertilizer and pesticides, which would impact yields beyond the climate impacts (Moersdorf et al., 2024).

Xia et al. (2022) used a comprehensive model of the food system to explore how many people globally would be affected by famine after nuclear war, arguing that the death toll from famine could be higher than direct fatalities from the nuclear war. This result assumes that global food trade would cease completely and that adaptation would be limited. Other research that models continued trade alongside large-scale and rapid adaptation concludes that many— potentially even all of these famine deaths—could be preventable given significant international cooperation, technological responses, and food aid (Rivers et al., 2024). However, further studies into trade show that there would be severe disruptions in the trade system, even in smaller nuclear exchanges (Jehn, Gajewski, et al., 2025). Together, these findings show that the food system’s response to a nuclear war is not fully understood.

We now know that inaction following a nuclear war would lead to widespread famine, while rapid and large-scale interventions could prevent it. However, research that explores the factors that make these outcomes—or the paths between them—more or less likely is missing.

Another major gap is an assessment of how plants would grow under nuclear winter conditions. Crop models are not calibrated to the estimated nuclear winter light conditions. Growth chamber trials are currently in progress, led by researchers at the Pennsylvania State University and Alliance to Feed the Earth in Disasters (ALLFED), but they are not yet complete and will represent only a first data point. Pilot-scale field trials that simulate sets of possible nuclear winter conditions would be a valuable next step towards high-quality evidence and would reduce uncertainty.

The longer term effects of nuclear war will crucially depend on responses and adaptations of the food system. ALLFED has contributed to research on the effects of nuclear war on food systems, including how different adaptation pathways could shape societal consequences in the years and decades following a nuclear war. Much of this work focuses on resilient food solutions, defined as food sources that could compensate for severe global shortfalls in traditional food production, such as those caused by a nuclear winter (see García Martínez et al. 2025, for a comprehensive review). Examples include crop relocation (Blouin et al., 2025), seaweed (Jehn et al., 2024), greenhouses (Alvarado et al., 2020), and single-cell protein grown in bioreactors (García Martínez et al., 2022). Without minimizing the profound and catastrophic consequences of nuclear war, this research is intended to support preparedness and informed decision-making, and is also relevant to strengthening resilience to other global food shocks.

In the past, research on resilient foods has focused primarily on modelling and the initial exploration of the research landscape. This work suggests that there is nothing in physics or current technology that would prevent humanity from producing enough food in a nuclear winter; rather, the main challenges lie in distribution, trade, and coordination—that is, in the systems required to rapidly scale and distribute resilient foods.

However, this part of the research landscape is almost completely unexplored. There are no clear ideas for how trade and cooperation could be maintained following a nuclear war, yet understanding this would be essential to increasing resilience for non-combatant countries. Without this understanding, there is a high chance that cascading trade bans could shut down global trade and lead to famine. On a much smaller scale than would be expected following a nuclear war, there were comparable dynamics during the 2007–2008 rice crisis. This highlights a range of questions that are primarily political. How would remaining food be distributed? What measures could food-importing countries take to address international supply shortages? How would emergency decisions be made when every delay leaves more people hungry? These questions require answers before a nuclear war, not after.

What also becomes clear from the literature is that many of the negative consequences of a nuclear war have not been examined yet. For example, a recent study explored the consequences of increased frost depths after a nuclear war and how this would impact drinking water infrastructure (Lamilla Cuellar et al., 2026). The study found that the projected increases in frost depth would likely disrupt drinking water supply across large parts of the Northern Hemisphere. While drinking water problems have been explored before in nuclear war research, this specific pathway has not. There are likely many research questions that have not been asked yet.

Besides these unknown unknowns, there is also a wide range of known unknowns, which have been highlighted in the literature (e.g. in Robock et al. (2023)) and need to be understood better:

• There is little modern research on the ecosystem impacts of nuclear war. One study suggests that it would likely lead to a mass extinction (Kaiho, 2023).

• Long-term Earth system impacts remain underexplored. Running large climate models is expensive, and most nuclear winter simulations therefore end after 15–20 years. However, the climate could experience long-term consequences after that.

• Nuclear war would not occur in isolation, but would interact with other already stressed components of the Earth system, like planetary boundaries. To date, only limited exploratory research exists (Jehn, 2023).

• A major unresolved question concerns how burnable modern cities are. Nuclear winter only happens if firestorms loft soot into the stratosphere. How much soot can be lofted depends on the amount of burnable material in the burning cities. Existing estimates span a wide range, from “nuclear winter is impossible” to “nuclear winter is guaranteed” (Hess, 2021; Wagman et al., 2020).

• Research on the economic consequences of nuclear winter is sparse, even for basic assessments such as how much food prices might rise (e.g. Hochman et al. (2022)). As a result, how nuclear war would affect the global economy remains poorly understood.

• There is some indication that nuclear winter research has been actively discouraged (Turchetti, 2021) and that many of the think tanks studying nuclear war have potential conflicts of interest (Egeland & Pelopidas, 2025). Understanding this in more detail would help in assessing the robustness of different research studies.

Most of the abovementioned studies are also relevant to other global catastrophes. This suggests that framing this work in an all-hazards approach (Sepasspour, 2023) would be appropriate, focusing on research which provides insights across a range of global catastrophic risks and prioritizing approaches that are effective across multiple scenarios. For example, improving models of the effects of nuclear war on climate and agriculture could also have benefits for research on other catastrophes with potential climate and food system impacts, such as major volcanic eruptions.

Key arguments from this submission for potential inclusion in the report:

• Nuclear winter is understudied relative to its importance, largely due to funding constraints, leaving only a small number of researchers active in this field.

• Climate impacts would be severe and rapid, but current results rely heavily on the same underlying data, making replication particularly important.

• Food system impacts would be severe and would extend far beyond combatant countries. Large scale famine is likely if no quick adaptation happens.

• If trade is maintained and resilient foods are employed at scale, a large fraction of famine deaths could likely be prevented.

• The critical bottlenecks for societal response likely are cooperation, trade, inequality and coordination, not physical constraints.

• Potential conflicts of interest in nuclear policy think tanks warrant caution in interpreting research reports.

Key research gaps that need to be addressed:

• No model intercomparison project exists for nuclear winter. Replication using different models would strengthen the evidence base.

• Crop models are not calibrated for nuclear winter light conditions.

• Trade and cooperation mechanisms post-war are almost completely unexplored.

• City flammability estimates range from “nuclear winter impossible” to “guaranteed” and need to be constrained further.

• Resilient foods appear promising for preventing global famine, but require more research, piloting, and policy support.

• Ecosystem impacts, long-term Earth system effects (beyond 15–20 years), economic consequences, and interactions with planetary boundaries are poorly understood.

  Subscribe now

How to cite

Jehn, F. U. (2026, February 18). Nuclear war, nuclear winter, and the food system. Existential Crunch. https://doi.org/10.59350/hkngk-5fq39

Recommended references and resources

Climate

• Primary climate assessment of a major nuclear war, which underpins many subsequent studies (Coupe et al., 2019).

• Climate assessment of smaller-scale nuclear wars (Toon et al., 2019).

• Potential strong El Niño-like response following nuclear war would disrupt oceans and climate (Coupe et al., 2021).

• Dependence of simulated climate effects on soot production and assumptions about burnable material, with substantial uncertainty in key inputs and the climate impacts (Wagman et al., 2020).

• First major updated study of nuclear-winter climate since research conducted in the 1980s (Robock et al., 2007).

• Recent review of the state of the nuclear war literature, with a focus on climate effects (National Academies of Sciences, Engineering, and Medicine, 2025).

• Role of latent heating in enabling soot to reach the stratosphere (Tarshish & Romps, 2022).

Environmental

• Interactions between nuclear winter and overstepped planetary boundaries (Jehn, 2023).

• Assessment forest resources following nuclear war (Winstead & Jacobson, 2022).

Radiation

• Disruption of public health systems in a nuclear winter (Vilhelmsson & Baum, 2023).

• Increased ultraviolet radiation following major nuclear wars due to ozone loss (Bardeen et al., 2021).

• Overview of radiation injuries and early fallout effects (Smith & Smith, 1981).

• Estimation of direct nuclear war casualties (Habbick, 1983).

Global socioeconomic systems

• Vulnerability to fuel dependency under declining trade conditions, illustrated using island nations (Boyd et al., 2023).

• Estimated decreases in industrial output if a certain percentage of industrial production is disrupted (Blouin, Jehn, et al., 2024).

• Disruption of food supply chains by the loss of electricity e.g. after a nuclear electromagnetic pulse (EMP) (Blouin, Herwix, et al., 2024).

• Projected reductions in renewable energy production in a nuclear winter (Varne et al., 2024).

• Long-term economic impacts of nuclear war on involved countries for decades (War et al., 1986).

• Unprecedented disruption of global food trade following nuclear war (Jehn, Gajewski, et al., 2025), alongside increasing fragility in the food system over time (Puma et al., 2015).

• Large increases in global food prices following even a small nuclear war (Hochman et al., 2022).

• Non-combatant countries would face massive disruptions due to climate effects and disruption of the global economy (Green, 2024).

• Several globally important ports during a nuclear winter, inferred from the identification of critical ports (Verschuur et al., 2022) would be frozen in a nuclear winter (Harrison et al., 2022) (these results are only implied, but Verschuur identifies important ports, and Harrison shows where sea ice is projected to extend to).

• Cascading impacts in the financial system following nuclear war, with losses potentially reaching hundreds of billions of dollars (Gajewski et al., 2025).

• Loss of access to pharmaceuticals if trade collapses following a nuclear war, illustrated using New Zealand as a case study (Wilson et al., 2025).

Agriculture

• Overview of resilient foods and the scenarios in which they are most useful (García Martínez et al., 2025), including a research agenda for food system adaptations to nuclear winter with over 100 proposals across agriculture, technology, infrastructure, planning, and policy.

• High likelihood of famine after nuclear war, due to dropping yields (Xia et al., 2022).

• Potential of seaweed as an especially effective resilient food in nuclear winter (Hinge et al., 2025; Jehn et al., 2024).

• Reduced availability of agriculture inputs (e.g. fertilizer, pesticides) following nuclear war that could reduce production by up to 70%, partly in addition to climate-caused reductions (Moersdorf et al., 2024).

• Crops are adapted to their local climate conditions, and most would fall outside their typical climate range during nuclear winter (McLaughlin et al., 2025), with potential for crop relocation to maintain production (Blouin et al., 2025).

• Additional yield reductions from increased ultraviolet radiation in the most severe nuclear wars (Shi et al., 2025).

• Severe global consequences for agriculture from even regional nuclear wars (Jägermeyr et al., 2020; Özdoğan et al., 2013).

• Potential to prevent global famine through food system adaptation and maintained trade even under severe nuclear winter conditions (Rivers et al., 2024), including through the use of non-agriculture-based foods (García Martínez et al., 2024). Evidence suggests these foods could be nutritionally adequate to prevent macro- and micronutrient deficiencies (Pham et al., 2022). However low-income countries may lack sufficient access and endure malnutrition without continued trade and significant aid (Asal et al., 2025).

• High concentration of the food system makes it vulnerable to large-scale disruptions such as nuclear war (Clapp, 2023).

• Estimated food supply capacity in New Zealand in a nuclear winter, using frost resistant crops (Wilson, Payne, et al., 2023).

Ecosystems

• Potential transition of ocean systems to a new, potentially stable state following nuclear war, with unclear long-term consequences (Harrison et al., 2022).

• Substantial increases of species extinction risk following nuclear war and the resulting nuclear winter (Kaiho, 2023).

• Negative impacts on fisheries due to the climate changes following nuclear war, particularly in already overfished systems (Scherrer et al., 2020).

Other

• Historical assessments of responses to nuclear winter, informed by volcanic cooling (Peregrine, 2018, 2021; Wilson, Valler, et al., 2023).

• Potential resilience of islands to many of the effects of a nuclear winter (Boyd & Wilson, 2023).

• Disruption of drinking water infrastructure across much of the Northern Hemisphere (Lamilla Cuellar et al., 2026), but could potentially be protected if sufficient measures are taken (Kamana-Williams et al., 2025).

• Global Catastrophic Food Failure as a new category to better assess food system catastrophes like nuclear war (Wescombe et al., 2025).

• Climate change will likely increase the risk of nuclear war (Egeland, 2025).

• Prevalence of financial conflicts of interest among think tanks involved in nuclear war policy (Egeland & Pelopidas, 2025).

• High likelihood of large-scale, long-term blackouts following nuclear war, with detailed analysis in Petermann et al. (2011); an English summary is available (original source in German): https://existentialcrunch.substack.com/p/the-consequences-of-blackouts

• Limited global food storage of approximately three-quarters of a year of a year (Laio et al., 2016).

• Overview of the global catastrophic risk space and the role of nuclear war and nuclear winter within it (Jehn, Engler, et al., 2025).

• Overview of divergent viewpoints on soot emissions during a nuclear war (Hess, 2021).

• A summary of the primary effects of nuclear war (Baum & Barrett, 2018).

• Potential resilience options for disruption of fossil fuels (Nelson et al., 2024) and electricity (Williams et al., 2025).

• A probabilistic model of uncertainty in climate and crop impacts of nuclear war and the cost-effectiveness of interventions (Denkenberger & Pearce, 2018).

References

• Alvarado, K. A., Mill, A., Pearce, J. M., Vocaet, A., & Denkenberger, D. (2020). Scaling of greenhouse crop production in low sunlight scenarios. Science of The Total Environment, 707, 136012. https://doi.org/10.1016/j.scitotenv.2019.136012

• Asal, Z., Martínez, J. B. G., Hinge, M., & Denkenberger, D. (2025). Nutrition in Abrupt Sunlight Reduction Scenarios: Analysis and prevention of malnutrition in low-income regions. EarthArXiv. https://eartharxiv.org/repository/view/10499/

• Bardeen, C. G., Kinnison, D. E., Toon, O. B., Mills, M. J., Vitt, F., Xia, L., Jägermeyr, J., Lovenduski, N. S., Scherrer, K. J. N., Clyne, M., & Robock, A. (2021). Extreme Ozone Loss Following Nuclear War Results in Enhanced Surface Ultraviolet Radiation. Journal of Geophysical Research: Atmospheres, 126(18), e2021JD035079. https://doi.org/10.1029/2021JD035079

• Baum, S., & Barrett, A. (2018). A Model For The Impacts Of Nuclear War. Global Catastrophic Risk Institute.

• Blouin, S., Herwix, A., Rivers, M., Tieman, R. J., & Denkenberger, D. C. (2024). Assessing the Impact of Catastrophic Electricity Loss on the Food Supply Chain. International Journal of Disaster Risk Science. https://doi.org/10.1007/s13753-024-00574-6

• Blouin, S., Jehn, F. U., & Denkenberger, D. (2024). Global industrial disruption following nuclear war. https://eartharxiv.org/repository/view/8166/

• Blouin, S., Rivers, M., Hinge, M., Antonietta, M., Jimenez, I., Jehn, F. U., & Denkenberger, D. (2025). Strategic crop relocation could substantially mitigate nuclear winter yield losses. EarthArXiv. https://eartharxiv.org/repository/view/10178/

• Boyd, M., Ragnarsson, S., Terry, S., Payne, B., & Wilson, N. (2023). Mitigating imported fuel dependency in agricultural production: Case study of an island nation’s vulnerability to global catastrophic risks. Risk Analysis, n/a(n/a). https://doi.org/10.1111/risa.14297

• Boyd, M., & Wilson, N. (2023). Island refuges for surviving nuclear winter and other abrupt sunlight-reducing catastrophes. Risk Analysis, 43(9), 1824–1842. https://doi.org/10.1111/risa.14072

• Clapp, J. (2023). Concentration and crises: Exploring the deep roots of vulnerability in the global industrial food system. The Journal of Peasant Studies, 50(1), 1–25. https://doi.org/10.1080/03066150.2022.2129013

• Coupe, J., Bardeen, C. G., Robock, A., & Toon, O. B. (2019). Nuclear Winter Responses to Nuclear War Between the United States and Russia in the Whole Atmosphere Community Climate Model Version 4 and the Goddard Institute for Space Studies ModelE. Journal of Geophysical Research: Atmospheres, 124(15), 8522–8543. https://doi.org/10.1029/2019JD030509

• Coupe, J., Stevenson, S., Lovenduski, N. S., Rohr, T., Harrison, C. S., Robock, A., Olivarez, H., Bardeen, C. G., & Toon, O. B. (2021). Nuclear Niño response observed in simulations of nuclear war scenarios. Communications Earth & Environment, 2(1), Article 1. https://doi.org/10.1038/s43247-020-00088-1

• Denkenberger, D. C., & Pearce, J. M. (2018). Cost-effectiveness of interventions for alternate food in the United States to address agricultural catastrophes. International Journal of Disaster Risk Reduction, 27, 278–289. https://doi.org/10.1016/j.ijdrr.2017.10.014

• Egeland, K. (2025). Disentangling the Nexus of Nuclear Weapons and Climate Change—A Research Agenda. International Studies Review, 27(1), viaf003. https://doi.org/10.1093/isr/viaf003

• Egeland, K., & Pelopidas, B. (2025). No such thing as a free donation? Research funding and conflicts of interest in nuclear weapons policy analysis. International Relations, 39(1), 125–147. https://doi.org/10.1177/00471178221140000

• Gajewski, Ł. G., Hinge, M., & Denkenberger, D. (2025). Quantitative, Data-driven Network Model for Global Cascading Financial Failure (No. arXiv:2502.12980). arXiv. https://doi.org/10.48550/arXiv.2502.12980 García Martínez, J. B., Behr, J., &

• Denkenberger, D. (2024). Food without agriculture: Food from CO2, biomass and hydrocarbons to secure humanity’s food supply against global catastrophe. Trends in Food Science & Technology, 150. https://doi.org/10.1016/j.tifs.2024.104609

• García Martínez, J. B., Behr, J., Pearce, J., & Denkenberger, D. (2025). Resilient foods for preventing global famine: A review of food supply interventions for global catastrophic food shocks including nuclear winter and infrastructure collapse. Critical Reviews in Food Science and Nutrition, 0(0), 1–27. https://doi.org/10.1080/10408398.2024.2431207

• García Martínez, J. B., Pearce, J. M., Throup, J., Cates, J., Lackner, M., & Denkenberger, D. C. (2022). Methane Single Cell Protein: Potential to Secure a Global Protein Supply Against Catastrophic Food Shocks. Frontiers in Bioengineering and Biotechnology, 10. https://www.frontiersin.org/articles/10.3389/fbioe.2022.906704

• Green, W. (2024). Nuclear War Impacts on Distant, Non-Combatant Countries [Policy Brief]. Toda Peace Institute.

• Habbick, B. (1983). Casualties in a nuclear war. Canadian Journal of Public Health = Revue Canadienne de Sante Publique, 74(1). https://pubmed.ncbi.nlm.nih.gov/6850478/

• Harrison, C., Faulkner, W., Coupe, J., Asante, E. K., Bardeen, C., Garza, V., Jägermeyr, J., Lovenduski, N. S., Robock, A., Rojas, K., Scherrer, K., Toon, O. B., & Xia, L. (2025). Accessible Climate and Impact Model Output for Studying the Human and Environmental Impacts of Nuclear Conflict.

• Harrison, C. S., Rohr, T., DuVivier, A., Maroon, E. A., Bachman, S., Bardeen, C. G., Coupe, J., Garza, V., Heneghan, R., Lovenduski, N. S., Neubauer, P., Rangel, V., Robock, A., Scherrer, K., Stevenson, S., & Toon, O. B. (2022). A New Ocean State After Nuclear War. AGU Advances, 3(4). https://doi.org/10.1029/2021AV000610

• Helfand, I., Lewis, P., & Haines, A. (2022). Reducing the risks of nuclear war to humanity. The Lancet, 399(10330), 1097–1098. https://doi.org/10.1016/S0140-6736(22)00422-6

• Hess, G. D. (2021). The Impact of a Regional Nuclear Conflict between India and Pakistan: Two Views. Journal for Peace and Nuclear Disarmament, 4(sup1), 163–175. https://doi.org/10.1080/25751654.2021.1882772

• Hinge, M., Grilo, V. A., Jehn, F. U., Martinez, J. B. G., Dingal, F. J., Roleda, M. Y., & Denkenberger, D. (2025). Seaweed cultivation: A cost-effective strategy for food production in a global catastrophe. Aquaculture International, 33(5). https://doi.org/10.1007/s10499-025-01978-x

• Hochman, G., Zhang, H., Xia, L., Robock, A., Saketh, A., Mensbrugghe, D. Y. van der, & Jägermeyr, J. (2022). Economic incentives modify agricultural impacts of nuclear war. Environmental Research Letters, 17(5), 054003. https://doi.org/10.1088/1748-9326/ac61c7

• Jägermeyr, J., Robock, A., Elliott, J., Müller, C., Xia, L., Khabarov, N., Folberth, C., Schmid, E., Liu, W., Zabel, F., Rabin, S. S., Puma, M. J., Heslin, A., Franke, J., Foster, I., Asseng, S., Bardeen, C. G., Toon, O. B., & Rosenzweig, C. (2020). A regional nuclear conflict would compromise global food security. Proceedings of the National Academy of Sciences, 117(13), 7071–7081. https://doi.org/10.1073/pnas.1919049117

• Jehn, F. U. (2023). Anthropocene Under Dark Skies: The Compounding Effects of Nuclear Winter and Overstepped Planetary Boundaries. Intersections, Reinforcements, Cascades: Proceedings of the 2023 Stanford Existential Risks Conference, 119–132. https://doi.org/10.25740/zb109mz2513

• Jehn, F. U., Dingal, F. J., Mill, A., Harrison, C., Ilin, E., Roleda, M. Y., James, S. C., & Denkenberger, D. (2024). Seaweed as a Resilient Food Solution After a Nuclear War. Earth’s Future, 12(1), e2023EF003710. https://doi.org/10.1029/2023EF003710

• Jehn, F. U., Engler, J.-O., Arnscheidt, C. W., Wache, M., Ilin, E., Cook, L., Sundaram, L. S., Hanusch, F., & Kemp, L. (2025). The state of global catastrophic risk research: A bibliometric review. Earth System Dynamics, 16(4), 1053–1084. https://doi.org/10.5194/esd-16-1053-2025

• Jehn, F. U., Gajewski, Ł. G., Hedlund, J., Arnscheidt, C. W., Xia, L., Wunderling, N., & Denkenberger, D. (2025). Food trade disruption after global catastrophes. Earth System Dynamics, 16(5), 1585–1603. https://doi.org/10.5194/esd-16-1585-2025

• Kaiho, K. (2023). An animal crisis caused by pollution, deforestation, and warming in the late 21st century and exacerbation by nuclear war. Heliyon, 9(4). https://doi.org/10.1016/j.heliyon.2023.e15221

• Kamana-Williams, B., Feng, X., Lamilla Cuellar, J. E., Peterson, R., & Denkenberger, D. (2025). Protection of subterranean water infrastructure in an abrupt sunlight reduction scenario. EarthArXiv. https://eartharxiv.org/repository/view/9098/

• Laio, F., Ridolfi, L., & D’Odorico, P. (2016). The past and future of food stocks. Environmental Research Letters, 11(3), 035010. https://doi.org/10.1088/1748-9326/11/3/035010

• Lamilla Cuellar, J. E., Palm, R., Denkenberger, D. C., & Jehn, F. U. (2026). Frost depth increase under a nuclear winter scenario projected to sever piped-water access in the Northern Hemisphere. Water Security, 27, 100193. https://doi.org/10.1016/j.wasec.2025.100193

• McLaughlin, C. M., Shi, Y., Viswanathan, V., Sawers, R. J. H., Kemanian, A. R., & Lasky, J. R. (2025). Maladaptation in cereal crop landraces following a soot-producing climate catastrophe. Nature Communications, 16(1). https://doi.org/10.1038/s41467-025-59488-6

• Moersdorf, J., Rivers, M., Denkenberger, D., Breuer, L., & Jehn, F. U. (2024). The Fragile State of Industrial Agriculture: Estimating Crop Yield Reductions in a Global Catastrophic Infrastructure Loss Scenario. Global Challenges, 8(1), 2300206. https://doi.org/10.1002/gch2.202300206

• National Academies of Sciences, Engineering, and Medicine. (with Board on Atmospheric Sciences and Climate, Nuclear and Radiation Studies Board, Committee on Independent Study on Potential Environmental Effects of Nuclear War, Division on Earth and Life Studies, & National Academies of Sciences, Engineering, and Medicine). (2025). Potential Environmental Effects of Nuclear War. National Academies Press. https://doi.org/10.17226/27515

• Nelson, D., Turchin, A., & Denkenberger, D. (2024). Wood Gasification: A Promising Strategy to Extend Fuel Reserves after Global Catastrophic Electricity Loss. Biomass, 4(2), Article 2. https://doi.org/10.3390/biomass4020033

• Özdoğan, M., Robock, A., & Kucharik, C. J. (2013). Impacts of a nuclear war in South Asia on soybean and maize production in the Midwest United States. Climatic Change, 116(2), 373–387. https://doi.org/10.1007/s10584-012-0518-1

• Peregrine, P. N. (2018). Social Resilience to Climate-Related Disasters in Ancient Societies: A Test of Two Hypotheses. Weather, Climate, and Society, 10(1), 145–161. https://doi.org/10.1175/WCAS-D-17-0052.1

• Peregrine, P. N. (2021). Social resilience to nuclear winter: Lessons from the Late Antique Little Ice Age. Global Security: Health, Science and Policy, 6(1), 57–67. https://doi.org/10.1080/23779497.2021.1963808

• Petermann, Th., Bradke, H., Lüllmann, A., Poetzsch, M., & Riehm, U. (2011). Was bei einem Blackout geschieht: Folgen eines langandauernden und großräumigen Stromausfalls. edition sigma. https://doi.org/10.5445/IR/140085927

• Pham, A., García Martínez, J. B., Brynych, V., Stormbjorne, R., Pearce, J. M., & Denkenberger, D. C. (2022). Nutrition in Abrupt Sunlight Reduction Scenarios: Envisioning Feasible Balanced Diets on Resilient Foods. Nutrients, 14(3), 492. https://doi.org/10.3390/nu14030492

• Puma, M. J., Bose, S., Chon, S. Y., & Cook, B. I. (2015). Assessing the evolving fragility of the global food system. Environmental Research Letters, 10(2), 024007. https://doi.org/10.1088/1748-9326/10/2/024007

• Reisner, J., D’Angelo, G., Koo, E., Even, W., Hecht, M., Hunke, E., Comeau, D., Bos, R., & Cooley, J. (2018). Climate Impact of a Regional Nuclear Weapons Exchange: An Improved Assessment Based On Detailed Source Calculations. Journal of Geophysical Research: Atmospheres, 123(5), 2752–2772. https://doi.org/10.1002/2017JD027331

• Reisner, J., Koo, E., Hunke, E., & Dubey, M. (2019). Reply to Comment by Robock et al. on “Climate Impact of a Regional Nuclear Weapon Exchange: An Improved Assessment Based on Detailed Source Calculations.” Journal of Geophysical Research: Atmospheres, 124(23), 12959–12962. https://doi.org/10.1029/2019JD031281

• Rivers, M., Hinge, M., Rassool, K., Blouin, S., Jehn, F. U., Martínez, J. B. G., Grilo, V. A., Jaeck, V., Tieman, R. J., Mulhall, J., Butt, T. E., & Denkenberger, D. C. (2024). Food system adaptation and maintaining trade could mitigate global famine in abrupt sunlight reduction scenarios. Global Food Security, 43, 100807. https://doi.org/10.1016/j.gfs.2024.100807

• Robock, A., Oman, L., & Stenchikov, G. L. (2007). Nuclear winter revisited with a modern climate model and current nuclear arsenals: Still catastrophic consequences: NUCLEAR WINTER REVISITED. Journal of Geophysical Research: Atmospheres, 112(D13), n/a-n/a. https://doi.org/10.1029/2006JD008235

• Robock, A., Toon, O. B., & Bardeen, C. G. (2019). Comment on “Climate Impact of a Regional Nuclear Weapon Exchange: An Improved Assessment Based on Detailed Source Calculations” by Reisner et al. Journal of Geophysical Research: Atmospheres, 124(23), 12953–12958. https://doi.org/10.1029/2019JD030777

• Robock, A., Xia, L., Harrison, C. S., Coupe, J., Toon, O. B., & Bardeen, C. G. (2023). Opinion: How fear of nuclear winter has helped save the world, so far. Atmospheric Chemistry and Physics, 23(12), 6691–6701. https://doi.org/10.5194/acp-23-6691-2023

• Scherrer, K. J. N., Harrison, C. S., Heneghan, R. F., Galbraith, E., Bardeen, C. G., Coupe, J., Jägermeyr, J., Lovenduski, N. S., Luna, A., Robock, A., Stevens, J., Stevenson, S., Toon, O. B., & Xia, L. (2020). Marine wild-capture fisheries after nuclear war. Proceedings of the National Academy of Sciences, 117(47), 29748–29758. https://doi.org/10.1073/pnas.2008256117

• Sepasspour, R. (2023). All-Hazards Policy for Global Catastrophic Risk (Technical Report Nos. 23–1; p. 37). Global Catastrophic Risk Institute. https://gcrinstitute.org/papers/068_all-hazards.pdf

• Shi, Y., Montes, F., Di Gioia, F., Xia, L., Bardeen, C. G., Anderson, C., Gil, Y., Khider, D., Ratnakar, V., & Kemanian, A. (2025). Adapting agriculture to climate catastrophes: The nuclear winter case. Environmental Research Letters. https://doi.org/10.1088/1748-9326/adcfb5

• Smith, J., & Smith, T. (1981). Radiation injury and effects of early fallout. Br Med J (Clin Res Ed), 283(6295), 844–846. https://doi.org/10.1136/bmj.283.6295.844

• Tarshish, N., & Romps, D. M. (2022). Latent Heating Is Required for Firestorm Plumes to Reach the Stratosphere. Journal of Geophysical Research: Atmospheres, 127(18), e2022JD036667. https://doi.org/10.1029/2022JD036667

• Toon, O. B., Bardeen, C. G., Robock, A., Xia, L., Kristensen, H., McKinzie, M., Peterson, R. J., Harrison, C. S., Lovenduski, N. S., & Turco, R. P. (2019). Rapidly expanding nuclear arsenals in Pakistan and India portend regional and global catastrophe. Science Advances, 5(10), eaay5478. https://doi.org/10.1126/sciadv.aay5478

• Turchetti, S. (2021). Trading Global Catastrophes: NATO’s Science Diplomacy and Nuclear Winter. Journal of Contemporary History, 56(3), 543–562. https://doi.org/10.1177/0022009421993915

• Turco, R. P., Toon, O. B., Ackerman, T. P., Pollack, J. B., & Sagan, C. (1983). Nuclear Winter: Global Consequences of Multiple Nuclear Explosions. Science, 222(4630), 1283–1292. https://doi.org/10.1126/science.222.4630.1283

• Varne, A. R., Blouin, S., Williams, B. L. M., & Denkenberger, D. (2024). The Impact of Abrupt Sunlight Reduction Scenarios on Renewable Energy Production. Energies, 17(20), 5147. https://doi.org/10.3390/en17205147

• Verschuur, J., Koks, E. E., & Hall, J. W. (2022). Ports’ criticality in international trade and global supply-chains. Nature Communications, 13(1), 4351. https://doi.org/10.1038/s41467-022-32070-0

• Vilhelmsson, A., & Baum, S. D. (2023). Public health and nuclear winter: Addressing a catastrophic threat. Journal of Public Health Policy. https://doi.org/10.1057/s41271-023-00416-7

• Wagman, B. M., Lundquist, K. A., Tang, Q., Glascoe, L. G., & Bader, D. C. (2020). Examining the Climate Effects of a Regional Nuclear Weapons Exchange Using a Multiscale Atmospheric Modeling Approach. Journal of Geophysical Research: Atmospheres, 125(24), e2020JD033056. https://doi.org/10.1029/2020JD033056

• War, I. of M. (US) S. C. for the S. on the M. I. of N., Solomon, F., & Marston, R. Q. (1986). The Consequences of Nuclear War: An Economic and Social Perspective. The Medical Implications of Nuclear War. https://www.ncbi.nlm.nih.gov/books/NBK219185/

• Wescombe, N. J., Martínez, J. G., Jehn, F. U., Wunderling, N., Tzachor, A., Sandström, V., Cassidy, M., Ainsworth, R., & Denkenberger, D. (2025). It’s time to consider global catastrophic food failures. Global Food Security, 46, 100880. https://doi.org/10.1016/j.gfs.2025.100880

• Williams, B., Croft, H., Hunt, J., Viloria, J., Sherman, N., Oliver, J., Green, B., Turchin, A., B, J., Pearce, J., & Denkenberger, D. (2025). Wood Gasification in Catastrophes: Electricity Production from Light-Duty Vehicles. Energy Engineering, 122(4), 1265–1285. https://doi.org/10.32604/ee.2025.063276

• Wilson, N., Payne, B., & Boyd, M. (2023). Mathematical optimization of frost resistant crop production to ensure food supply during a nuclear winter catastrophe. Scientific Reports, 13(1), 8254. https://doi.org/10.1038/s41598-023-35354-7

• Wilson, N., Valler, V., Cassidy, M., Boyd, M., Mani, L., & Brönnimann, S. (2023). Impact of the Tambora volcanic eruption of 1815 on islands and relevance to future sunlight-blocking catastrophes. Scientific Reports, 13(1), Article 1. https://doi.org/10.1038/s41598-023-30729-2

• Wilson, N., Wood, P., & Boyd, M. (2025). Capacity to manufacture key pharmaceuticals in New Zealand after a global catastrophe. New Zealand Medical Journal, 138(1625), 44–58. https://doi.org/10.26635/6965.7053

• Winstead, D. J., & Jacobson, M. G. (2022). Forest Resource Availability After Nuclear War or Other Sun‐Blocking Catastrophes. Earth’s Future, 10(7). https://doi.org/10.1029/2021EF002509

• Xia, L., Robock, A., Scherrer, K., Harrison, C. S., Bodirsky, B. L., Weindl, I., Jägermeyr, J., Bardeen, C. G., Toon, O. B., & Heneghan, R. (2022). Global food insecurity and famine from reduced crop, marine fishery and livestock production due to climate disruption from nuclear war soot injection. Nature Food, 1–11. https://doi.org/10.1038/s43016-022-00573-0

This post here was written in collaboration with the Geopolitical Insight and Education Foundation. It was originally published here. They are think tank focused on democratic stability. This post was improved with lots of ideas, comments and suggestions by Bennett Iorio and Sandro Sousa. Check out their work, they will post several additional pieces about democracy by other authors in their commentary program ‘Drivers of Instability & Effects on our Democracies’. I am cross-posting it here to this living literature review, as it summarizes nicely many of the arguments I have made about the advantages of democracy.

Subscribe now

The proportion of the global population living under the spectrum of political systems classed as a democracy is near its all-time high, and stands (very approximately) around 4 billion, or around half of the global population. From a historical perspective, democracies have had quite the run over the last two centuries: Before the revolutions of 1848, 0% of the global population lived in what we would today recognize as a democracy, as none of the existing states satisfied various essential democratic properties like truly accountable leadership, as well as free and fair multiparty elections (Lührmann et al., 2018; Our World in Data, 2025), though the 100 years preceding the 1848 revolutions saw significant progress in this direction in Europe and the new United States. Since 1848, democracies have seen a meteoric rise, reaching their high water mark around 2010. However, progress has stalled and since 2010, the proportion of population living under democratic rule has decreased (Figure 1). Some of this may relate to demographic trends, with aging populations and decreasing overall percent of population in “traditional” democratic heartland territory like Europe, North America and Japan as compared to significant population growth in the global south, often in countries with democratic deficits. As a pointed example, the world’s fastest growing country, Niger, is also ranked amongst the lowest in terms of democracy and freedom.

Figure 1: Proportion of global population living democracies or autocracies (1789 to 2025). This figure should be read with the caveat that democracy is not a binary category but a spectrum, and regime classifications inevitably simplify complex institutional realities.

But the main and more dire driver of this change is the erosion of democratic standards and ‘backsliding’ in many democratic polities. ‘Democratic’ status in countries that in 2010 were still flawed or even full democracies, like Hungary, Slovakia, India and Russia, is now under threat, with many now being classed as ‘electoral autocracies’. Put simply, democracy as a system has proven highly resilient in the preceding two hundred years, growing its population share enormously, but not fundamentally invulnerable to reversal.

This is quite concerning on many levels; first and foremost because democracies are the only form of government where leaders can be held truly accountable and the population is a stakeholder in government, not only a subset of powerful citizens. There is ample evidence too that democracy as a ‘self-correcting mechanism’, that is, a system with robust checks and balances has been a significant factor in the staying power of democratic systems (Aspen Institute, 2017). But there is a second order reason that democracies should be protected and nourished, not only for their own sake, but on a systemically important level: They are more effective than other governance systems at both preventing and managing disasters and catastrophes. So, protecting democracies is not only a moral imperative for our societal wellbeing, it’s also necessary to avoid destruction, collapse and human harm after large catastrophes.

Why democracies are better in crisis management and prevention

Historical evidence

As stated above, democracies in their modern definition have only existed since 1848, so issues of correlation (with other major historical trends, such as industrialization, colonialism and decolonization, etc.) present a statistical problem for how we understand and disentangle the cause and effect of disaster mitigation, and we cannot easily look into deep history to understand how democracies face threats. However, democracies have certain features that define them and which can also be found in pre-democratic societies. This means when we look into history, we can sample societies that score high on these democratic values and use them as a proxy for how democracies might have fared under those circumstances, and draw conclusions on the impact of these proto-democratic mechanisms and their impact on societal resilience.

An approach like this was implemented in two studies by Peter Peregrine (Peregrine, 2018, 2021). In the first study Peregrine collected data from 33 societies which faced a total of 22 catastrophic climate disasters (mostly drought). The idea of the study was to test whether greater local participation and more community coordination/governance, or high enforcement of societal norms more effectively increase resilience of societies against catastrophe. To track this, Peregrine coded variables, including how steep hierarchies were, and how much social norms were enforced (e.g. by looking how much rituals and dwellings were standardized). This resulted in an overall score of how much social norms were enforced and how steep hierarchies were. These scores were then compared against societal outcomes (change in population, health, etc.). The study’s result demonstrated that societies with stricter enforcement of norms (a proxy for less democracy) had poorer outcomes than less hierarchical societies (a proxy for democracies), which tended to have better outcomes. A similar argument was made in the recently published book Goliath’s Curse (Kemp, 2025). In it Kemp, describes how lootable resources allow wealth accumulation, which leads to monopolized violence and enforced hierarchies. On a wide range of historical case studies, he shows that these enforced hierarchies are inherently unstable and that humans tend to prefer to live in more egalitarian societies.

Building on this, the second Peregrine study used a similar approach, but focussed on the Late Antique Little Ice Age (the coldest volcanic winter in the last 3000 years, and used by societal researchers as a rough proxy of a nuclear winter). Just as with the first study, Peregrine compared societal outcomes with markers of hierarchy and participation. And just as with the first study, more flexible, participatory societies fared better. A useful modern analogue can be found in contrasting democratic and authoritarian responses to the COVID-19 pandemic, which constituted a large-scale, exogenous stress test for state capacity, social compliance, and adaptive governance. China’s response relied on extreme centralization and coercive enforcement, including prolonged mass lockdowns, mobility controls and information suppression. While these measures initially reduced transmission, they generated significant second-order costs: Delayed detection of outbreaks, brittle compliance, economic dislocation, and ultimately a loss of global public trust, as well as destruction of trust in the state’s own policies within the country (Zhang et al., 2024). By contrast, countries such as Denmark adopted a markedly different approach, emphasizing transparency, decentralized implementation, and voluntary compliance grounded in high institutional trust. Denmark imposed relatively limited and time-bound restrictions, avoided prolonged school closures, and relied heavily on public communication and local discretion - yet achieved comparatively low excess mortality and faster social recovery (Mathieu et al., 2020). Importantly, this contrast should not be read as a simple function of regime type alone; demographic scale, urban density, and health-system structure are relevant too. Nevertheless, the comparison reinforces the historical pattern identified by Peregrine of a highly centralized, norm-enforcing systems appearing effective in the short term but tending towards brittleness under prolonged stress, whereas participatory systems with distributed authority and trust-based compliance exhibit greater adaptive resilience.

A much larger dataset, which captures all major polities of the last 5000 years shows similar results (Hoyer et al., 2025). This dataset measured crisis outcomes (e.g, French Revolution, Fragmentation of the Mughal Empire or Meiji Restoration in Japan) and how they relate to the properties of the involved polities. They find that one important factor if a state manages to introduce reforms to end a crisis is the hierarchical complexity. More concretely, they find that there needs to be a sufficiently complex state with a power base outside of the administrative apparatus of the state. Which is exactly what democracy is for; stable democracies trend away from highly centralized power structures, and instead rely on balancing power between institutions and interest groups.

There is also a companion paper to Hoyer et al. (2025). In this other paper, the aim was to understand what countries can do to prevent a crisis before it occurs (Hoyer et al., 2024). To understand they searched for cases of societies, which came right before the brink of crisis, but then successfully led to peace again. The four clearest cases for such a behavior they find are the conflict of the orders in the early Roman Republic (494-287 BCE), the Chartist movement in England (1819-1867), the reform period in the Russian Empire (1855-1881) and the Progressive Era in the USA (1914-1939). The critical element they shared to allow reform before the crisis became unmanageable was that they managed to generate enough elite buy-in and a broad societal alliance. This provided enough power to implement reforms against the special interests of powerful elites. Once again, this indicates that democracies are better at managing crises. Although it is also challenging to form broad alliances in a democracy and oppose vested interests, because of the central idea of decentralised power, it is still likely to be more effective at managing such issues than other forms of government.

Modern evidence

The Peregrine studies provide compelling evidence that more participatory, less hierarchical societies performed more effectively in the management of extreme scenarios in the past. However, there are significant systemic differences between pre-modern societies and modern, global democracies; the next section provides evidence that modern democracies are more stable and handle crises better, too.

Planning for and managing disasters and future outcomes

Management and outcomes of disasters are improved by first understanding, and second preparing for them. One proxy used to assess this is the presence and magnitude of national risk assessments; national risk assessments are the attempt by nations to map out the space of possible futures and what dangers might await them there. Countries with strong national risk assessments have, in this sense, made the first step in preparing well for disasters. But not all countries tend to undertake national risk assessments and those which do often differ in their focus. Generally, more democratic governments tend to create better risk assessments than autocratic ones. To compare their farsightedness we can look at how different countries treat global catastrophic risk in their national risk assessments, as these are usually the worst tail risks that can happen (Figure 2).

Figure 2: Overview of what kind of global catastrophes different countries consider in their national risk assessments. “Tangentially” refers to the hazard being mentioned, but not discussed further.

There is evidence that at least some states treat global catastrophic risk as a serious object of national planning, with Switzerland standing out for the breadth and depth of its assessment. The United States is not included in this comparison, as it does not produce a single, consolidated national risk assessment; instead, relevant analyses are distributed across multiple agencies, making them difficult to evaluate as a coherent whole. That said, the United States is distinctive in having produced a risk assessment focused explicitly on global catastrophic risks, with plans for periodic updates on roughly decennial timescales.

All of the countries included in this comparison are democracies, and there appears to be no publicly available national risk assessment of comparable scope produced by an autocratic government. This does not imply that such assessments do not exist in classified form, but this in and of itself is also a problem. National-scale risk assessment is inherently complex, and important pathways to catastrophe are easily overlooked. Public availability allows external scrutiny and iterative improvement. New Zealand provides a useful illustration: earlier versions of its national risk assessment were classified, but the decision to publish a subsequent iteration exposed the document to substantial criticism, which in turn informed revisions and strengthened overall resilience (Boyd & Wilson, 2021).

Beyond ex ante planning, it is also instructive to examine state performance once disasters actually occur. Lin (2015) provides a systematic analysis of this question, examining 150 countries between 1995 and 2009, including their state capacity, how democratic they were and how they responded to disasters. The latter was measured by using publicly available data on how many people were affected by a disaster (death, injury or homelessness). Government expenditure divided by GDP was used as a proxy for state capacity and democratic status was assessed by using the Polity IV dataset. The central finding is that disaster impacts are substantially lower in countries that combine democratic governance with high state capacity. Neither democracy nor capacity alone is sufficient: autocratic systems, as well as democracies with weak state capacity, experience significantly worse outcomes when disasters strike. Lin (2015) attributes this effect in part to democratic accountability, which increases political costs for leaders who fail to manage disaster response effectively, while emphasizing that accountability must be matched by administrative and fiscal capacity to translate into effective action.

When it comes to prevention, there is also evidence that more democratic processes tend to create more forward-looking outcomes. Citizens’s Assemblies aim to bring together a large number of citizens, who are a representative sample of their society. These are provided information and time to create a broad perspective and suitable solutions for societal problems. Lage et al. (2023) examine the outcomes of such assemblies in the European Union in the context of climate policy, finding that their recommendations were consistently more ambitious than those advanced by national governments and more closely aligned with what climate science suggests is required. Because these recommendations emerge from structured deliberation by ordinary citizens rather than political elites, they often command higher levels of public legitimacy. The citizens’ assemblies held in Ireland on abortion provide a well-documented illustration: their recommendations extended well beyond what had previously been considered electorally feasible, yet ultimately received broad public support once the deliberative process underpinning them was widely understood. In short, democracies tend towards high disaster-preparedness, and both their governments and civil societies tend to take steps to understand and reinforce against disasters.

Democracies tend to avoid very bad outcomes

These studies suggest that democracies perform better both in preparing for disasters and in managing them once they occur, provided they possess sufficient state capacity. In addition, there is evidence that democratic systems are also more effective at avoiding severe negative outcomes in the first place.

Democracies rarely go to war against each other: There are very few cases in history where democratic states fought each other. When the analysis is limited to full liberal democracies under contemporary definitions, this number effectively falls to zero. While there is no single, definitive theory explaining this pattern, a prominent explanation (Democratic Peace Theory) runs that it is politically difficult to justify aggression against another fully democratic state to a domestic electorate (Reiter, 2017; Tomz & Weeks, 2013).

Stable democracies very rarely have famines: Historically, famines have been among the most destructive events societies can face, often resulting in mass mortality and political instability by undermining the most basic conditions for survival. Democracies, however, display a markedly stronger record in this regard. Outcomes depend in part on how both famine and democracy are defined, but according to (Hasell & Roser, 2013), only a single famine has occurred in a democratic state since 1870, whereas autocratic regimes experienced dozens over the same period (Figure 3).

Figure 3: Death from large famines compared by political regime.

Democracies avoid economic growth for its own sake: There is a persistent argument that autocratic systems are better at long-term planning than democracies, on the grounds that they are not constrained by electoral cycles of four to five years. This claim is examined by Millemaci et al. (2024). They assess the relationship between regime type, economic growth, and longer-term societal outcomes - such as education, health, public transport provision, and responsiveness to public preferences - comparing democracies and autocracies across these dimensions. Their findings indicate that democracies perform at least as well as autocracies in terms of economic growth on average, while the most extreme outcomes at both ends of the distribution - very high growth and severe contraction - are observed in autocratic systems. When societal outcomes are considered, the contrast is even clearer. Democratic countries deliver systematically better outcomes for their populations on these measures (Figure 4).

Figure 4: Comparison of positive societal outcomes (Long-term Background - LTB - index consisting of factors like health, public transport etc.) between democracies and autocracies. A higher index indicates better outcomes for general society. The y axis is a kernel density estimate, this means it is a smoothed approximation of the data points present. In simpler terms this just means higher values on the y-axis indicate more countries are concentrated around that LTB index value.

Democracies often reverse autocratization: It has been established that democracies tend to avoid negative outcomes better and more reliably than autocracies. Accordingly, democratic backsliding into autocracy constitutes a negative outcome in its own right. At the same time, democratic systems also exhibit a degree of self-stabilization in response to autocratic erosion. This dynamic is examined by Nord et al. (2025), who analyzed the evolution of democratic regimes between 1900 and 2023, tracing changes in democratic quality over time. Within this time series, the authors identify episodes in which countries experienced abrupt shifts toward either greater democratization or increased authoritarianism. They find that approximately half of all episodes of autocratization are reversed within a relatively short period, defined in this study as no more than five years. This pattern has strengthened over time, with contemporary democracies reversing autocratizing trends in approximately 73% of cases.

What threatens democracies today?

The reasons for the contemporary decline in democracy are likely numerous and complex (see Figure 1). A review by Loughlin (2019) characterizes this decline as a slow and often opaque process. Rather than occurring through overt coups, democratic erosion today typically proceeds through the gradual degradation of democratic institutions and practices. Hungary provides a frequently cited example; although formally classified as a democracy, the government under Viktor Orbán has, over time, dismantled institutional checks and balances and reshaped the electoral system in ways that substantially reduce the likelihood of electoral defeat. Comparable dynamics can be observed in other established democracies: Examples often cited include the Alternative für Deutschland in Germany and the MAGA movement associated with Donald Trump in the United States. While the mechanisms of democratic erosion are increasingly well documented, the reasons for their apparent acceleration in the present period remain less clear.

Loughlin identifies several contributing factors:

• Globalization: Rising globalization has reduced states’ capacity to regulate their domestic economies, as they are increasingly exposed to external forces beyond their direct control. This constrains important avenues for national self-determination.

• Inequality: Economic inequality has increased in many countries over recent decades, worsening material conditions for significant segments of the population in democratic societies. This has eroded trust in traditional democratic processes, which had promised broad-based improvements in living standards.

• Focus on the individual: Neoliberal capitalism, the dominant economic model across most democracies, places a strong emphasis on individual autonomy. In combination with rising migration and inequality, this emphasis can weaken social cohesion and collective identification. Such collective identification, however, is often important for sustaining democratic legitimacy, particularly when democratic outcomes do not align perfectly with individual preferences.

• Economic power of companies: Large multinational corporations, such as Amazon or Apple, exert growing influence over public policy through lobbying and related mechanisms. This can generate voter frustration when political decisions appear to reflect corporate interests rather than popular preferences.

While these are significant claims, they are frequently echoed in other papers examining the relationship between societal conditions, inequality, and democracy. For instance, Centeno & Cohen (2012) discuss neoliberalism’s history and emphasize similar issues as Loughlin. In the late 20th century, neoliberalism emerged as an attractive ideology promising both democracy and prosperity through free markets. As Centeno & Cohen (2012) note, this was especially compelling in the 1990s when the success of the “Asian Tigers” seemed to demonstrate how global market mechanisms could lift countries out of poverty. The apparent triumph of capitalism over the Soviet model further cemented the appeal of neoliberal ideas.

The core tenets were straightforward: Minimize government intervention, privatize public services, and deregulate markets. The theory was that this would maximize economic freedom which would, in turn, guarantee political freedom. This “indivisibility thesis” - the idea that economic and political freedom were inseparable - became deeply influential in policy circles.

However, the implementation of neoliberal policies had several consequences that increased societal vulnerability:

• Increased Inequality: The focus on deregulation and tax reduction, particularly for the wealthy, has led to growing inequality.

• Weakened State Capacity: Through extensive privatization and deregulation, states reduced their ability to respond to crises and provide public goods.

• Broken Feedback Loops: As wealth and power became more concentrated, decision-makers became increasingly insulated from the consequences of their choices, as they mainly hear what needs to be done from lobbyists. This has made it harder for societies to recognize and respond to emerging threats.

A more concrete example of how economic power enables large companies to influence the political system is provided by Supran et al. (2023). They examined ExxonMobil’s company records, showing the company knew by the early 1980s how much fossil fuels would heat the planet. These models accurately predicted recent decades’ warming. Despite this, ExxonMobil used its influence to spread misinformation that contradicted their findings and to persuade policymakers to avoid legislation that would impede their business.

Taken together, these factors share a common underlying feature. They can all be understood as consequences of neoliberal capitalism, which prioritizes market mechanisms as the primary mode of societal coordination while seeking to limit the role of the state.

Actionable insights to make democracies stronger again

Taken together, the preceding examples indicate that strong democratic processes are a central component of societal resilience to catastrophic risks. In addition to enhancing crisis preparedness and response, such processes are also associated with higher levels of citizen wellbeing in the present. This raises the question of how democratic systems might be strengthened. Addressing this challenge requires confronting a set of structural weaknesses, notably high inequality, diminished state capacity, and disrupted democratic feedback loops.

Fixing inequality: Giving citizens the resources and time to participate in democracy

Democratic participation is time-intensive. Broadening participation in decision-making can improve the quality of outcomes, but it also requires sufficient space for deliberation. If citizens are to engage more fully in democratic processes, such engagement must be feasible alongside their working lives. In contemporary economies, however, many individuals work long hours. While some do so by choice, for others this is a necessity driven by basic material needs such as housing and subsistence. For the latter group, increasing discretionary time and security is a prerequisite for meaningful democratic participation. One possible approach is a universal basic income (while the interpretation of trials remain contested, they generally point towards better outcomes on an individual level with regard to outcomes on health and time), which could provide individuals with a financial buffer that enables greater civic engagement. More generally, a wide range of policies that reduce economic pressure - such as extended parental leave, improved childcare provision or higher taxes on multinational corporations and inheritance - could serve similar functions.

Fixing weak state capacity: Making people more excited about democracy

Beyond having the time to participate, citizens also require sufficient motivation to engage in democratic processes. This dimension is among the most difficult to influence through policy. Nonetheless, existing research identifies factors that can increase citizens’ engagement with, and confidence in, the political community they inhabit. One such factor is whether the state is perceived as accessible, competent, and reliable. When public services deteriorate - for example, when local facilities close, public transport is unreliable, or libraries lack adequate funding - it becomes difficult for citizens to view the democratic system as effective. Strengthening public provision therefore plays an important role in restoring confidence in democratic governance. Increased state capacity is also directly relevant to catastrophe preparedness, as it underpins the infrastructure and administrative processes required to mitigate and respond to large-scale risks.

Fixing broken feedback loops: Creating the structures and processes which allow more direct participation in democracy

Once citizens have both the time and motivation to participate, appropriate institutional processes and structures are also required to channel this engagement effectively. A wide range of potential institutional responses could contribute to this goal. At the most ambitious end of the spectrum, this would involve a substantial reallocation of political authority toward lower levels of government. Such arrangements have historical precedent in the United States, most notably in the tradition of town halls. During the formative period of the United States, local councils exercised significant authority and provided the democratic authority and foundation for constitutional development, as locally elected representatives participated directly in drafting governing frameworks. Although much of this authority was later consolidated at the state level, this evolution does not imply that decentralization is inherently irreversible. Contemporary examples also exist from which such models could be drawn. Switzerland, for instance, allocates substantial decision-making authority to subnational levels of government. Similarly, the autonomous region of Rojava in northern Syria represents an ongoing experiment in localized and participatory democratic governance.

A more readily implementable solution could be citizens’ assemblies. Such assemblies have demonstrated a capacity to generate forward-looking policy recommendations that also command relatively high levels of public approval. They can be convened across a wide range of policy domains. One timely application would be the governance and regulation of artificial intelligence. It is also conceivable that citizens’ assemblies could be organized on a transnational basis to contribute to forms of global risk assessment, integrating democratic deliberation with efforts to strengthen global resilience. Another potentially effective approach is the use of participatory budgeting, which grants citizens a degree of authority over how local governments allocate public funds. This approach is grounded in the assumption that local populations are often best placed to identify the challenges they face and to prioritize responses accordingly. These approaches could be further strengthened through the use of digital technologies. A growing number of platforms enable accessible and remote forms of democratic deliberation, lowering barriers to participation and coordination (for example Decidim).

Conclusion

Democracies face significant pressure, yet they remain essential both for sustaining high quality of life and for building societies that are resilient to future risks. The policy options outlined here are not intended as an exhaustive set of solutions, but rather as illustrative examples of how democratic systems might be strengthened. Many additional interventions may also prove effective, provided they expand citizens’ time and resources for participation, foster motivation to engage in democratic self-governance, and establish institutional processes that shift meaningful authority toward lower levels of decision-making.

References

• Aspen Institute. (2017). Democracy as Self-Correction. Aspen Institute Centre for Enterprise. https://www.aspeninstitutece.org/article/2017/democracy-as-self-correction/

• Boyd, M., & Wilson, N. (2021). Anticipatory Governance for Preventing and Mitigating Catastrophic and Existential Risks. Policy Quarterly, 17(4), Article 4. https://doi.org/10.26686/pq.v17i4.7313

• Centeno, M. A., & Cohen, J. N. (2012). The Arc of Neoliberalism. Annual Review of Sociology, 38(Volume 38, 2012), 317–340. https://doi.org/10.1146/annurev-soc-081309-150235

• Hasell, J., & Roser, M. (2013). Famines. Our World in Data. https://ourworldindata.org/famines

• Hoyer, D., Bennett, J. S., Whitehouse, H., François, P., Feeney, K., Levine, J., Reddish, J., Davis, D., & Turchin, P. (2024). CRISES AVERTED How A Few Past Societies Found Adaptive Reforms in the Face of Structural- Demographic Crises. OSF. https://doi.org/10.31235/osf.io/hyj48

• Hoyer, D., Holder, S., Bennett, J. S., Francois, P., Whitehouse, H., Covey, R. A., Feinman, G., Korotayev, A., Ustyuzhanin, V., Preiser-Kapeller, J., Bard, K., Levine, J., Reddish, J., Orlandi, G., Ainsworth, R., & Turchin, P. (2025). All Crises are Unhappy in Their Own Way: The Role of Societal Instability in Shaping the Past. Social Science History, 1–33. https://doi.org/10.1017/ssh.2025.10113

• Kemp, L. (2025). Goliath’s Curse: The History and Future of Societal Collapse. Viking.

• Lage, J., Thema, J., Zell-Ziegler, C., Best, B., Cordroch, L., & Wiese, F. (2023). Citizens call for sufficiency and regulation—A comparison of European citizen assemblies and National Energy and Climate Plans. Energy Research & Social Science, 104, 103254. https://doi.org/10.1016/j.erss.2023.103254

• Lin, T.-H. (2015). Governing Natural Disasters: State Capacity, Democracy, and Human Vulnerability. Social Forces, 93(3), 1267–1300.

• Loughlin, M. (2019). The Contemporary Crisis of Constitutional Democracy†. Oxford Journal of Legal Studies, 39(2), 435–454. https://doi.org/10.1093/ojls/gqz005

• Lührmann, A., Tannenberg, M., & Lindberg, S. I. (2018). Regimes of the World (RoW): Opening New Avenues for the Comparative Study of Political Regimes. Politics and Governance, 6(1), 60–77.

• Mathieu, E., Ritchie, H., Rodés-Guirao, L., Appel, C., Gavrilov, D., Giattino, C., Hasell, J., Macdonald, B., Dattani, S., Beltekian, D., Ortiz-Ospina, E., & Roser, M. (2020). Excess mortality during the Coronavirus pandemic (COVID-19). Our World in Data. https://ourworldindata.org/excess-mortality-covid

• Millemaci, E., Monteforte, F., & Temple, J. R. W. (2024). Have Autocrats Governed for the Long Term? Kyklos. https://doi.org/10.1111/kykl.12425

• Nord, M., Angiolillo, F., Lundstedt, M., Wiebrecht, F., & Lindberg, S. I. (2025). When autocratization is reversed: Episodes of U-Turns since 1900. Democratization, 0(0), 1–24. https://doi.org/10.1080/13510347.2024.2448742

• Our World in Data. (2025). Countries that are democracies and autocracies. https://ourworldindata.org/grapher/countries-democracies-autocracies-row

• Peregrine, P. N. (2018). Social Resilience to Climate-Related Disasters in Ancient Societies: A Test of Two Hypotheses. Weather, Climate, and Society, 10(1), 145–161. https://doi.org/10.1175/WCAS-D-17-0052.1

• Peregrine, P. N. (2021). Social resilience to nuclear winter: Lessons from the Late Antique Little Ice Age. Global Security: Health, Science and Policy, 6(1), 57–67. https://doi.org/10.1080/23779497.2021.1963808

• Reiter, D. (2017). Is Democracy a Cause of Peace? In Oxford Research Encyclopedia of Politics. https://doi.org/10.1093/acrefore/9780190228637.013.287

• Supran, G., Rahmstorf, S., & Oreskes, N. (2023). Assessing ExxonMobil’s global warming projections. Science, 379(6628), eabk0063. https://doi.org/10.1126/science.abk0063

• Tomz, M. R., & Weeks, J. L. P. (2013). Public Opinion and the Democratic Peace. American Political Science Review, 107(4), 849–865. https://doi.org/10.1017/S0003055413000488

• Zhang, M., Liu, B., Xiang, G., Yan, X., Ling, Y., & Zuo, C. (2024). Navigating the shift: Understanding public trust in authorities amidst policy shifts in China’s COVID-19 response. Humanities and Social Sciences Communications, 11(1), 1716. https://doi.org/10.1057/s41599-024-04224-6

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

To better understand societal collapse we not only must examine the factors that could lead to collapse, but also what might happen during and after collapse. One crucial thing here is economics. What are the direct effects of collapse on an economy? How would prices change? Would trade continue? All important questions, but very difficult to answer due to the complexities of how our economies work. Due to this difficulty only very little research exists for this and those that do exist usually only consider a single aspect or simplified scenarios.

Subscribe now

Just a single bomb

One example here is a briefing paper by the UK non-profit Article 36 (Article 36, 2015). They analyze the economic consequences of the explosion of a single nuclear weapon. They argue that the effects of such an explosion, be it intentional or not, will have global consequences, but also note that they can only provide some rough sketches, due to the inherent complexity of the matter.

Their key points are:

• Massive loss of life: Even a single, low yield (100 kT) nuclear weapon could plausibly kill and injure around 300,000 people. The effects would be way worse if we would use the larger bombs we have today.

• Destruction of industries: Often industries are concentrated in a region due better use of economies of scale. Therefore, a single explosion could completely disrupt a key sector of a national economy.

• Destruction of infrastructure: Cities tend to be the main hub of infrastructure, be it education, public transport or administration. If those are disabled, the whole region suffers from it.

• Heavy burden on public finances: It would be extremely costly to rebuild the destroyed city. Also, public health costs in the region would be quite elevated for decades to come.

• Unclear recovery: Due to the high cost and unclear systemic effects, the effects of even a single bomb could make it quite hard for the affected country to recover, even with outside help.

Besides those points the paper repeatedly mentions that these are only vague predictions, as our current economic models have a hard time predicting such unprecedented events. We do have some data from the Second World War where Japan and Germany bounced back quite strongly, even though they endured much stronger damage than a single bomb could do. For example, Hiroshima and Nagasaki both took around 10 years to get back to the population levels they had before the bombs were dropped. However, the biggest bombs we have in use today have one to several megatons of explosive yield, which is several magnitudes larger than the only nuclear weapons that were ever dropped in anger. Therefore, it remains quite uncertain if we could use the reconstruction of Japan and Germany as a comparison class.

Industrial disruption

A paper that looks more systematically into the impacts of industrial destruction is by Blouin et al. (2024) (Disclaimer: I am a co-author on this one). The general idea is that, while we cannot easily assess the damages a nuclear war might do to industrial output, we can look into history, to find case studies of how destruction of industry leads to declines in industrial outputs and extrapolate from this. These case studies reach from major hurricanes to the Second World War (Figure 1).

Figure 1: Relationship of incapacitated industrial infrastructure and subsequent decline in industrial output. Based on historical case studies and modelling studies.

In addition to establishing this relationship between industrial destruction and declines in output, the study also assesses how much of industrial capacity would be destroyed in US/Russia and India/Pakistan conflicts. This can be estimated from plausible target lists and the destruction caused by the nuclear weapons detonating above those targets. In the case of the US/Russia conflict the estimate is that this could destroy roughly 3 % of global industrial capacity, which would translate to a 24 % loss in global industrial output from the detonation damage alone. This strongly nonlinear relationship between destroyed capacity and declining output highlights the vulnerability of economies to large scale disruptions.

Modeling agricultural prices after a nuclear war

The studies above relied more on case studies and historical comparisons, but can we still learn something from them by applying economic models to global catastrophes?Due to the difficulties of using economic models so far outside of their training data, very little research on this topic exists. For example, to my knowledge there is just a single study that even attempts to model how a nuclear war (or similar events) might impact the global economy. In this case the study looks at food prices after a nuclear war (Hochman et al., 2022).

If you know of important studies I’ve missed, please email me and I’ll update this article!

The study by Hochman and co-authors uses a general equilibrium model originally developed to understand the economic effects of climate change (later we will discuss why this might be a problem). The model simulates most countries and a total of 15 sectors, of which 9 represent the food system (mainly different crops like rice, wheat etc.), while the rest is for mining, manufacturing, construction and transport. They force the model by providing the changes in agricultural productivity due to worsened climate after a nuclear war. Their scenario is a smaller exchange between India and Pakistan, resulting in roughly 2-4°C decrease of the global mean temperature.

The model makes one quite big assumption: it does not consider the destruction caused by the bombs. It basically models a world where everything is as today, with the one exception that agriculture is suddenly much less productive. This means it also does not consider big knock on effects. India is a large food exporter and having it effectively removed from the market by the destruction from a nuclear war would send additional shockwaves rippling through the global economy. In addition, it is the most populous country in the world and many of the people living there would likely want to migrate from India in such a scenario.

But even this incomplete picture of the situation already shows that this would massively impact our food system. They find that global calories would likely decrease by 10-12 % and also see short time price peaks. However, their model also finds that prices for staple foods quickly go down again, as farmers switch away from livestock and vegetables to produce more staple crops. All this comes with one additional caveat: the results are only valid under the assumption that global trade continues as is, which might not be the case, after such a massive shock.

Do we even have the right tools to model global catastrophes?

The papers we have discussed hint that any global catastrophe would likely be very bad for our overall economic outlook. But it also shows that we do have large difficulties in even applying our current tools to such tasks. Modeling the economics after nuclear war was only possible by excluding the direct impacts of the war and modeling trade is done on such an abstract level that it is often difficult to draw direct lines to the way trade is actually occurring. This leads to the question: Do we even have the right tools to model and analyze our economy after global catastrophe?

One global risk where this question is discussed more thoroughly is climate change. A well known paper here is by Weitzman (2009). The main argument is that the way cost-benefit analysis is done for climate change is fundamentally flawed, as it does not properly account for the uncertainty around catastrophic outcomes. The core problem is structural uncertainty: we don’t just face unknown outcomes, we don’t even know the true probability distribution of those outcomes. We have to estimate it from limited data. When you account for this estimation uncertainty mathematically, you get a distribution with fatter tails than climate economic models typically assume. Extreme events become more probable than thin-tailed distributions would suggest.

Weitzman builds a toy model which showcases why this matters. In his model, the probability of catastrophe declines polynomially as the catastrophe gets worse, but the disutility (how much the catastrophe hurts) grows exponentially (the more you lose, the more each additional loss hurts). Exponential growth beats polynomial decline. This means that even very low-probability catastrophic outcomes can dominate the entire expected damage calculation. The problem is that climate economic models usually use thin-tailed distributions and arbitrary cutoffs for extreme damages. This means they are effectively assuming we can ignore catastrophic outcomes because their probability is negligible. But if the tails are actually fat, this assumption breaks down. If a model ignores the fat tails of the distribution, it has to make much more positive predictions by default, as it ignores the worst parts.

Weitzman’s critique applies to every model that does not consider tail risks properly, no matter the underlying structure. But in complex, global questions like trade policy analysis, climate economics, and large-scale policy simulation one kind of model is used especially often: general equilibrium models. These kinds of models try to simulate an entire economy and how prices react to changes (e.g. production shocks). Their problems have been summarized in a review by Stiglitz (2018). Generally, one of the things you want from a model is to predict catastrophic changes, because those are the ones that hurt most. If GDP growth is 2.8 % or 2.9 % does not matter that much, but if you have a market crash, that hurts. The problem is that general equilibrium models fail at exactly that. They have quite a hard time predicting sudden downturns.

This often gets attempted to be fixed by introducing additional parameters. However, additional parameters always make it easier to just produce a nice looking overfit. The problem with these kinds of models is that they work on quite a high level of abstraction. It doesn’t actually model how households, companies or countries behave or react. It just calculates an equilibrium price. And thus it cannot really tell you anything about changing expectations or values. The models solve this by simply assuming that humans are completely rational actors. But they are not and their behavior is what we actually want to predict, not the future state of some idealized price. This means these models assume the thing we are most interested in as static. As they fail this, they also fail to predict catastrophic outcomes, as these are based on changing expectations, behavior and values. The simple structure of the models also does not account for amplification of shocks in the system. It assumes that shocks can only ever come from the outside. However, from systemic risk research we know that amplification of risks in a system and a risk arising from the system itself regularly happen.

When we think about global catastrophes, another problem creeps in. Population is usually assumed to be an external factor, meaning that the model just assumes a stable population a priori. However, after global catastrophes the loss of life is likely one of the most important factors for economic consequences (2).

Finally, the economic models only look at a single catastrophe at a time. What would happen if we have massive damages from high warming plus a pandemic (3)? We simply do not know.

Are we going to be able to model economics after a global catastrophe anytime soon?

Can you calculate a price after a catastrophe? Maybe, but it will always be a guess, as most of the available tools that just are not made for the task. Economic models are ill equipped to model the effects of global catastrophes, but even the first attempts made highlight the potential for massive consequences. To improve predictions here economic models have to focus on outcomes outside of GDP, like lives lost and we also have to work on improving the missing feedbacks. Empirical work could improve, but this likely only works for climate change, as there simply do not exist any reliable training data sets for catastrophes like nuclear war. Overall, it seems that it will be quite some time until we can make any usable predictions here and in the meantime we should be quite careful of any economic prediction that involves global catastrophes. This all harks back to modeling of a whole economy being really difficult. We cannot even accurately forecast recessions today, but at least we have data to calibrate the models on. Simulating global catastrophes thankfully cannot rely on empirical data, but this makes it much more difficult as well.

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

Subscribe now

Endnotes

(1) Agricultural optimization aligns with local climates. Short-term climate shifts therefore incur higher costs than long-term ones. As climate shifts, initial productivity decline is substantial, but over time, agriculture adapts and starts to resemble places with similar climates before the shift happened. Comparing productivity across current climate zones may forecast long-term outcomes, yet may underestimate short-term damages, given the time required for agricultural adjustment.

(2) In an earlier post we have discussed how a shrinking population is one of the parameters that can bring a society back into equilibrium with the carrying capacity of its surroundings. However, this equilibrium can be on a much lower population and technological level.

(3) This example is actually something that seems to be getting more likely over time. Increased human population and animals migrating to cope with higher warming produce much more opportunities for disease to be transmitted from animals to humans.

How to cite

Jehn, F. U. (2026, January 14). End time economics. Existential Crunch. https://doi.org/10.59350/10f7p-hc578

References

• Article 36. (2015). Economic impacts of a nuclear weapon detonation.

• Blouin, S., Jehn, F. U., & Denkenberger, D. (2024). Global industrial disruption following nuclear war. EarthArXiv. https://doi.org/https://doi.org/10.31223/X58H9G

• Hochman, G., Zhang, H., Xia, L., Robock, A., Saketh, A., Mensbrugghe, D. Y. van der, & Jägermeyr, J. (2022). Economic incentives modify agricultural impacts of nuclear war. Environmental Research Letters, 17(5), 054003. https://doi.org/10.1088/1748-9326/ac61c7

• Stiglitz, J. E. (2018). Where modern macroeconomics went wrong. Oxford Review of Economic Policy, 34(1–2), 70–106. https://doi.org/10.1093/oxrep/grx057

• Weitzman, M. L. (2009). On Modeling and Interpreting the Economics of Catastrophic Climate Change. The Review of Economics and Statistics, 91(1), 1–19. https://doi.org/10.1162/rest.91.1.1

It’s been half a year since the last time I updated old posts, so here we go again. It’s always a bit odd to go back to old posts. It feels like those were written by someone else and in a way they were. They are past me’s work, not present me’s. But I hope that the additions I am making here are just as interesting as the original posts were to you. Also, I think these updated posts are a nice reminder to my readers that for this blog here it makes sense to go through the backlog. They are still relevant, get updated and I hope that you can learn something from them. This round of updates even saw some updates to the first post I have officially written for this living literature. So, if you are a recent subscriber and like what I produce here, make sure to check out those older posts (or maybe recommend them to somebody else). I am hopeful you will get something out of it.

And here are the updates.

Subscribe now

Mapping out collapse research

In this post I gave an overview of the main schools of collapse. The post started with a simple figure that showed how these schools related to each other. I now updated this plot to better show how I think these schools differ.

Figure 1: Schools of collapse research.

Also, since I originally wrote this post, the great book Goliath’s curse by Luke Kemp was published. This introduces a new theory of collapse, so I had to make sure that the post reflects this. So, I added to paragraphs that discuss Kemp’s book in more detail.

Goliath’s curse

“Dominance hierarchies doom societies”

Goliath’s curse is the most recent addition to major collapse theories (introduced by Luke Kemp in 2025). It draws inspiration from all of them. Similarly to Structural demographic theory, it is more focussed on the social aspects, but also discusses major shocks, especially in the form of global catastrophic risk today.

The general idea of Goliath’s curse is storable resources enable dominance hierarchies that create inequality, weakening societies until shocks trigger collapse. This is all based on a large body of evidence, which starts from the belief that humans favor egalitarian societies, but that societies easily end up in a spiral of dominance. This spiral is started by people gaining access to storable resources (like grain). Once those are available, they can be amassed and inherited. Quickly creating a power differential, which allows some to gain power. With this power often comes violence and coercion. People can only bear so much of it and they rebel if it gets too bad. This cycle has repeated countless times by now and if we are not careful, the global society of today might end up in the same way. So, better to reduce inequality now, before a global collapse makes everyone equally poor (2).

There is also the idea that collapse does not necessarily have to be bad. This was especially championed by Kemp (2025). Collapse ceases to be a problem if you are not really connected to the state anyway and if the state you lived in was coercive in the first place. Historically, this was quite often the case, with many people’s lives improving after the state around them collapsed. They did not lose much, except freedom from taxes and the main people who had a bad time were the elites of the collapsing society. This does not mean that is always good for the average Joe. Especially, in today’s highly connected world, it would likely be devastating and deadly for many, but we should keep in mind that collapse also can have upsides if it ends an authoritarian regime.

Finally, I wanted to update the conclusion of the post to better reflect my learnings from writing about collapse for almost three years now.

What also has become more clear to me while reading about all these different ways to think about collapse is that they all revolve how civilizations increase or decrease resilience and the likelihood and severity of external shocks they have to face. Obviously, all these theories focus on different answers to what causes these resilience changes and shocks, but also they are not really mutually exclusive. It seems more to me that they are focusing on facets of the same picture. And so if I had to lump them all together, I would say that collapse theories are about the forces that change resilience and how this resilience interacts with shocks. I tried to visualize this in Figure 6.

Figure 6: Example resilience trajectory for an idealized civilization.

This is meant to show resilience shifts over time due to the decision a civilization makes. It can go up and down and every so often it gets hit by a shock of a certain size. If the shock is smaller than the resilience, the civilization moves on. If the shock is handled well, the resilience might even increase afterwards. But it can also lead to further deterioration. Finally, at some point the civilization is hit by a shock so large, it cannot be handled anymore and it collapses. I like this framing, as it is largely agnostic to what changes resilience and thus allows all kinds of theories being lumped together here. It also explains why very low resilience civilizations can survive for a long time. They just got lucky when it comes to the magnitude of shocks. Similarly, high resilience civilizations can still get toppled by a very high magnitude shock when they get unlucky.

These are just my two cents of how this all might fit together after reading a few hundred papers and a bunch of books around the topic. It is far from a finished theory, but maybe it inspires some of my readers to dig deeper.

Famine’s role in societal collapse

A famine is one of the most devastating things that can happen to a society. They usually start by some shock to the food system. This can be all kinds of things, ranging from climate to conflicts or even simply mismanagement. To explore this further, I added an additional paper by Jehn et al. (2025).

If we want to have an even deeper look into crop production shocks specifically, we can look at Jehn et al. (2025) (Disclaimer: I am the first author of this paper). The idea of this paper was to look at the total crop production and not separate them into individual time series for each individual crop. To do this, the individual crops were aggregated using their caloric values, resulting in one calorie production time series for each country. The shocks were detected by comparing the actual caloric values in each year with the amount that would have been expected in that year, given the long term trend. This allowed us to detect those years that experienced the largest crop production shocks in each country (Figure 2).

Figure 2: The largest crop production shocks per country. Darker colors indicate larger shocks.

As you can see the shocks can be quite substantial. Especially in Africa several countries have experienced crop production shocks of 60 % or more. But even the world’s largest producers have faced significant shocks. The United States in 1983 for example had a crop production shock of almost 30 %.

By looking through historical documents and things like FAO reports, almost every shock could be linked to an historical event which plausibly caused it. This allowed us to also compare the different crop shock reasons to each other (Figure 3).

Figure 3: Swarm plots showing the magnitude of crop production shocks across different cause categories. The black line indicates the mean. Single points show all individual country-level shocks. For each category the largest shock is labelled with year and country it occurred in.

This revealed that climate induced shocks both produce the largest shocks in most countries, but also cause the most severe shocks in general. The vast majority of those climate shocks were droughts. But most of the other categories identified also contain major shocks, indicating that a sufficiently large crisis of any kind, can be devastating to agriculture.

Climate anomalies and societal collapse

When we look at which environmental factors are usually related to societal collapse, climate clearly comes out on top. Droughts are especially devastating to any food system and bring societies to their breaking point. To explore this in more detail, I added another case study in the form of Norman et al. (2025).

Another example of this very clear link between climate and crisis is a paper by (Norman et al., (2025). They also want to explore how climate influences societies. To do this they look at a large tree ring dataset, which gives them an overview of the climate across the Roman Empire for its entire existence. In this climate dataset they look for all the years where a major battle happened in the Roman Empire (for a total of 106 battles). For the years of all the battles, as well as the ten years before and after each battle, they look in the timeseries of the regions where the battle happened, if there are any unusual climate events. They find that these major battles are significantly more likely when there were climate extremes beforehand. In the majority of cases the relevant climate extreme happened three years before the battle took place. They explain this by saying that the climate shock needed time to percolate through society: A climate shock happens, food gets scarce, tension starts to rise and is released in a major battle. The climate events that cause this are usually droughts or extremely hot weather.

Economic inequality and societal collapse

Across many theories of collapse and risk theories, inequality often pops up as a strong cause for decreased societal resilience. It leads to less social cohesion, state capacity and general wellbeing. However, around this topic there is often the argument that this has to be tolerated, as inequality is a byproduct of capitalism and capitalism is the economic form which led billions out of poverty. To challenge this, I added a discussion of Sullivan & Hickel (2023) to my post about inequality:

But to track down the roots of present day inequalities it makes sense to dig even deeper. There is a great paper by Sullivan & Hickel (2023), which explores how poverty developed from the 16th century until today. The motivation behind the paper is to show that deep poverty is not something that every society experienced and which was only fixed by the wonders of economic growth in capitalism, but instead that in many countries people were actually fine and only started to suffer once capitalism started. They think that the former view was able to take hold, as for these arguments per capita GDP is used as a proxy of how poor people were. However, this proxy has three big problems:

• There are many examples in history where GDP per person improved, while people clearly experienced worse life. They specifically highlight the Philippines between 1820 and 1902. In this time period GDP per person rose by 15 %, but pretty much every record at the time only talks about how living standards are just completely deteriorating.

• Past GDP per person is based on calculating purchase power parity. This means you calculate how much stuff people were able to buy with the money they had at the time (4). However, these calculations are highly error prone, especially for food.

• Capitalism existed for a long time before GDP per person started to improve. The basis of capitalism started to get established in the 16th century. However, from 1600-1800 GDP per person in many countries actually declined.

To avoid those problems, they want to look at measures of how well people actually lived, not some abstract measure such as GDP. For this they selected real wages (specifically how long someone had to work to get a day’s worth of food for a four person family), height and mortality rates. They track all these markers from the 16th century till today in Europe, Latin America, Sub-Saharan Africa, South Asia and China.

Their clearest finding is that in all of those places there was little to none extreme poverty. Most workers were able to provide for a family of four. The time periods where they were able to find extreme poverty were usually during major wars or in some exceptional countries like Peru who were suffering from especially harsh colonialism.

This changed once the given region started to transition to capitalism. In all cases height and real wages decreased, while mortality went up. Some of these trends were reversed when strong social democratic or socialist movements came to power in the first half of the 20th century. However, many of these advances in human well being were rolled back again after neoliberalism’s rise in the later 20th century.

This does not mean that becoming a socialist state is always the best way to avoid poverty, as there were also examples like China where this ended in the population getting poorer. However, it clearly shows that capitalism cannot be singled out as the main reason wellbeing improved. Instead it looks more like it is actually holding many societies back.

Share your favorite

Thanks for reading so far in the post. Here’s a great song as an appreciation.

As you made it this deep, I assume that you like the stuff that is written here. If you want to do me a little favor, I would really appreciate it if you could share your favorite post with someone who might enjoy it.

Go ahead, you can do it right now. Nothing more to read in this post anyway.

Much appreciated!

Subscribe now

Until next time

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me an email to existential_crunch at posteo.de and we can schedule something.

References

• Jehn, F. U., Mulhall, J., Blouin, S., Gajewski, Ł. G., & Wunderling, N. (2025). The Largest Crop Production Shocks: Magnitude, Causes and Frequency. EarthArXiv. https://eartharxiv.org/repository/view/10133/

• Kemp, L. (2025). Goliath’s Curse: The History and Future of Societal Collapse. Viking.

• Norman, C., Schwinden, L., Krusic, P., Rzepecki, A., Bebchuk, T., & Büntgen, U. (2025). Droughts and conflicts during the late Roman period. Climatic Change, 178(5), 87. https://doi.org/10.1007/s10584-025-03925-4

• Sullivan, D., & Hickel, J. (2023). Capitalism and extreme poverty: A global analysis of real wages, human height, and mortality since the long 16th century. World Development, 161, 106026. https://doi.org/10.1016/j.worlddev.2022.106026

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

As long as we are not in the year 2100, we will not know exactly what path humanity will have taken when it comes to climate change in this century. But we can at least make educated guesses about the trajectory we are on. Given this, it seems that a good amount of climate science is aimed at hopeful temperature ranges below 2°C, which we will likely miss. The focus on below 2°C stems from the aspirational climate targets that humanity has set itself. We have less research on the currently most probable looking temperatures far above 2°C and even less on low probability, but high impact tail risk warming. By tail risk, I mean both the chance of unexpectedly high emissions and the chance of unexpectedly high temperatures from a given level of emissions. This post aims to give an overview of why I think that climate science is at least partly looking in the wrong direction to understand such risks.

Subscribe now

Reasons for pathways above 2°C

First, let’s explore why it seems to have gotten more likely again that we will end up in a more than 2°C world. While there has been some hope before 2020 that climate action was gathering the force it needs to limit warming below 2°C, the last few years have had more of a feeling of disappointment in them. Both from the way the climate system is behaving, as well as how society is doing.

Green backlash

Before COVID hit, there were widespread climate protests in many countries. However, only a few years later it seems like the main news stories are that climate action is just too costly and we should abandon it and funnel all the money into the economy instead. This is not only my own impression, but also the topic of a recent review paper by Bosetti et al. (2025). They tried to understand the political consequences of climate policies. Based on their findings they argue that dissatisfaction around climate science mainly focuses on two issues:

• Economics: The distributional consequences. Climate policies, as all policies, produce winners and losers. If too many people think that climate policies will harm them, they will try to elect someone else.

• Culture: A broader pattern of distrust towards the scientific and political elite (1).

They summarize the overall process in a very straightforward conceptual model (Figure 1). The basic idea is politicians implement climate policies, these impact citizens, who elect politicians based on this impact. Easy.

Figure 1: Conceptual model of how climate policies, politicians and citizen preferences interact

They go on to explain all three steps of this in more detail:

Impact of climate policies

Climate policy typically involves two main components: promoting renewable energy and reducing fossil fuel use. Renewable energy projects often face local opposition, like people opposing turbines near their homes. Policies targeting fossil fuels include traffic restrictions, bans on combustion engine vehicles, increased fuel taxes, and coal mine closures.

Bosetti et al. (2025) find a consistent pattern across their case studies: these policies increase support for anti-climate politicians, but only when affected citizens receive no compensation. When compensation is provided, like improved public transport alongside traffic restrictions, support for climate action holds steady. People appear willing to accept climate policies when they’re compensated for losses and even more when they are included in decision-making. Unfortunately, these important steps are often skipped, as they cost additional money, which governments seem unwilling to spend. This compensation and representation gap typically benefits right-wing parties advocating against climate action.

Advocacy against climate action from populist right wing parties

Populist right wing parties particularly emphasize that there could be a trade-off between climate action and economic development, which is too high to pay. This is grounded in their belief that global warming is not actually that bad and a general opposition to higher taxes, often based on a neoliberal belief system. These messages allow them to frame climate action as a competition about resources between pure, working class people and corrupt elites, who only push climate action for their own gain. This messaging is often mixed with other themes like “climate hysteria”, which aims to discredit everybody who speaks up in support of climate action, as an overzealous activist.

There are some populist right wing parties (e.g. in Finland) who generally accept the claim that global warming is a reality, but are still against doing anything, as they frame it as unfair that they should pay, while others stand by idly. Similarly, it is often argued that one’s country is too small to make a difference, so only the major polluters are required to do anything.

Finally, climate action means change and conservative parties are all about maintaining the status quo. This argument is then pushed further by arguing that the country is in a pure and beautiful state (or more likely was) and that climate action will push it further away from this ideal.

Political consequences of the green backlash

We now know that climate action without compensation can trigger a significant backlash and that populist right wing parties are most likely to benefit from them. This has a significant impact on the likelihood of climate action decreasing or even being rolled back. The most prominent example here is Donald Trump. Besides many other things, he withdrew the US from the Paris Agreement (twice), defunded the Environmental Protection Agency and liberalized the drilling for oil and gas.

But we can see the same happening in Europe. Generally, the greater the influence of populist right-wing parties, the weaker the climate action. This impact increases if they manage to control a relevant part of the government.

Summing it up

Overall, this shows that we are currently in a time of slowing climate commitment, as exactly those parties who advocate against it, are able to gather momentum. However, as many of the grievances seem to be driven by a lack of compensation, this also opens the door for more successful climate policies in the future. But unfortunately, we do not yet live in this world.

Path dependence

This rise of parties that are hellbent on delaying or stopping climate action has also shown up in the progress on climate action. If you have been involved in the climate movement or are interested in climate change, you probably have seen the kind of graph that shows past emissions and how much these would have to be reduced in the next decades to limit warming to 1.5°C or 2°C. You might have noticed that the slope of the emission reduction trajectory gets steeper every year, as countries fail to step up their climate game.

This pattern has become so obvious that there is even a whole UN report just about that topic (UN Environment Programme, 2023). In the report they review the state of emissions reductions and how it has developed over the last few years. Every year (with the exception of 2020 due to COVID) we set a new record in emissions, while the pathway we should be on points down, not up. Though, at least it seems that the emissions are growing more slowly every year, which implies that we should see them pointing down in a few years. But even if we see such a change in the direction of the trend in a few years, it will be too late to easily set us at a 1.5°C path again. The UN report also created an updated version of the graph (Figure 2) that shows the discrepancy between where we should be and where we actually are. This one clearly shows that we are very much on a pathway to above 2°C.

Though to be fair to the right wing populist parties who are currently working on dismantling climate action, progress wasn’t stellar before they showed up either. It is just that they slowed down the crawl even more.

Figure 2: Climate pathways for the next decades for current policies and what we would actually need to get to 1.5°C or 2°C degrees. Y-axis is the amount of gigatons of carbon dioxide we can emit.

Hot 2020s

While we seem to be living in a world where climate action decreases, at the same time the climate system itself may be responding more strongly than anticipated. Even if emissions trajectories stabilize, higher climate sensitivity or stronger feedbacks could push us toward higher temperatures than current models predict.

Here I want to highlight a recent study by Minobe et al. (2025). They argue that we are currently in a time where almost every new year is a new record in temperature. There no longer is a “normal” year. However, even in this more extreme world the years 2023 and 2024 have been extraordinary. Especially, oceans are severely affected by the rapid warming. To study this, they came up with a new statistical test which tracks abnormal-record breaking years and use this to analyze ocean and atmosphere data from 2023 and 2024. They find that these two years were warmer than we should have expected, even after factoring in El Niño events. This implies that we might be on a quicker climate pathway than standard models predict, might be caused by either climate sensitivity being higher than the median estimate or that feedbacks are stronger. In both cases higher temperature scenarios are becoming more likely. Though we need more data to be sure.

While this is just a single study, it is not the only one that has recently made the point that climate might be warming more quickly, as the last few years have been quite exceptional.

This implies we face dual risks: both insufficient emissions reductions and potentially stronger climate responses to whatever emissions we do produce. Both pathways lead to higher temperature becoming more likely.

Do we actually know what might happen in more extreme warming?

Temperature focus

The three arguments I made above, a backlash to the green transition, slowing climate action and accelerating warming all point that we have a very good chance to end up well above 2°C warming. But how good are we actually prepared for that? Because theoretically, we all agreed that warming well under 2°C (or even below 1.5°C) is our collective aim as a species. As this was the agreed goal, there was also an incentive to focus just on those temperature ranges and scenarios that explore those. How much did this incentive shape research? This is the question that Jehn et al. (2022) tackle (Disclaimer: I am the first author on this paper).

The approach of the paper is simple: Read in all documents produced by the IPCC and track how often they mention different temperatures. The result is Figure 3.

Figure 3: Temperature mentions across the different IPCC assessment reports.

It shows that the range of temperatures explored always focused most on temperature below 2°C, but really emphasized them in AR6, which was the first report after the Paris Agreement. This highlights that up until recently, most research energy was spent on temperatures we are quite unlikely to remain below.

Scenario focus

I am not aware of any study which updated those numbers using the IPCC documents published after 2022, but there haven’t been that many new documents anyway. But what exists is an interesting review and perspective paper by Meinshausen et al. (2024). In it they discuss how climate scenarios have been used in the past and how they should be used in the future. I am not going to recap the whole thing, because it is quite long, but what I want to highlight is that they find that we miss a suitable scenario that can be used, which is in line with the current promised policies. These will likely lead us to a world with far above 2°C warming, albeit with a high uncertainty. Past scenarios often explored temperature ranges higher and lower than this, but not really this pathway in particular. Meinshausen and co-authors therefore argue that in the future we should create such a specific scenario that starts from the place we are actually in right now. This would equip us with the knowledge needed for the pathway we will probably end up in.

This does not mean that we should all suddenly focus on RCP8.5 again, because this worst case scenario is practically impossible by now (2), but a more sensible approach to higher temperature and especially considering tail risk. This means the chance that we end up in a high temperature scenario, due to unforeseen feedback or higher than expected climate sensitivity.

Also, scenarios can only be a piece of the puzzle. What I think would be especially helpful is research around compounding and cascading events. The IPCC is quite good at pointing to things like the likelihood of extreme weather events increasing. However, what the research generally lacks is an understanding of complex cascades happening above 3°C. In such a warming scenario we can expect events like multiple-breadbasket failures, overstepping tipping points or large scale ecosystem collapse. What are the long term consequences of this? How might they interact? There is very little research about this.

Exploring tail risk is a good idea actually

The two last papers show that there is a clear gap in the research, both in the most probable bath, but also in more extreme warming. This means we need more research here urgently, to know what we are up against and what might happen if we get really unlucky. However, often when such catastrophic outcomes are highlighted as being worthwhile for research, they are often dismissed as only contributing to pessimism and climate inaction.

Thankfully, there is also a perspective paper by Davidson & Kemp (2024), which lays out why it is instead quite valuable to explore more extreme warming and that it is an essential part of good climate research and policy. They argue that by now there is plenty of evidence that exploring extreme warming does not misrepresent science. Also, there is no evidence that they slow decarbonisation or lead to apathy. There are even multiple studies that highlight that fear-based messaging is effective at altering intentions and behaviors, when they are framed as actionable.

They distance themselves from climate doomerism. They think of themselves and this approach rather as climate realism. Not seeing the more extreme futures as a certainty, but as a probability that has to be explored, to be safe from unpleasant surprises. Their motivation is to understand these futures better, so we can avoid them, not preparing for near term collapse (2).

They think there are three approaches which are most helpful to study these climate endgames:

• The first is foresight, which involves scientists, economists, and philosophers creating systematic assessments of catastrophic futures through tools like climate models, expert surveys, and forecasting exercises. While this approach aims to provide evidence-based probability estimates, it often falls short by using overly simplistic methods that miss important cascading risks and societal responses.

• The second approach is agitation, where political actors and activists use images of catastrophic futures to mobilize climate action. This includes warnings from groups like Extinction Rebellion showing that disaster is possible but avoidable. While agitation can effectively communicate stakes and counter false optimism, it risks depoliticizing the crisis by framing it as a universal threat that obscures different interests and positions.

• The third approach is fiction, where writers and artists create speculative narratives of climate-changed worlds. Climate fiction serves unique functions beyond entertainment—it challenges taken-for-granted assumptions through estrangement, offers reflexive commentary on how different communities already experience disaster differently, and reveals the limits of our ability to imagine unprecedented futures. While fiction may seem too speculative, it actually helps highlight important but overlooked futures and encourages epistemic humility (3).

The authors argue these three approaches are complementary and should be better integrated, while always maintaining awareness that extreme outcomes are uncertain possibilities to be avoided, not inevitable certainties to accept.

Conclusion

All this taken together it seems like we are on a challenging climate path of far above 2°C warming and we do not yet know enough about what might await us in such a future. However, it is great to see that people are working on this and it also was a pleasant surprise while reading the first paper (Bosetti et al., 2025), that apparently more climate policy does not automatically lead to more backlash. It just needs to be compensated and done more democratically for the outcomes to be much better.

Finally, I want to highlight especially the climate tipping points community. After many discussions and reading their papers, it seems to me that this is the part of the climate change community which takes tail risks most seriously. Also, they now publish their most recent knowledge every year in a global tipping points report. It’s great to see that they are gaining traction in their research around more extreme futures.

Now we just have to put in the work to avoid those.

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

Subscribe now

How to cite

Jehn, F. U. (2025, November 26). Does climate science focus on the right temperature range?. Existential Crunch. https://doi.org/10.59350/g64bp-jjf43

Endnotes

(1) If you want to learn more about where this distrust might come from, take a look here.

(2) Climate Brink wrote a nice explainer about why RCP8.5 is not likely.

(3) People mean a lot of different things when they talk about collapse. If you really mean collapse of industrial society, I don’t really see what preparation now will give you. Chances are really high you just work hard for the privilege to starve last.

(4) The best example of this is arguably “Ministry for the Future”. I can recommend asking climate scientists what they thought about the book. High chance they have read it and an opinion.

References

• Bosetti, V., Colantone, I., De Vries, C. E., & Musto, G. (2025). Green backlash and right-wing populism. Nature Climate Change, 15(8), 822–828. https://doi.org/10.1038/s41558-025-02384-0

• Davidson, J. P. L., & Kemp, L. (2024). Climate catastrophe: The value of envisioning the worst-case scenarios of climate change. WIREs Climate Change, 15(2), e871. https://doi.org/10.1002/wcc.871

• Jehn, F. U., Kemp, L., Ilin, E., Funk, C., Wang, J. R., & Breuer, L. (2022). Focus of the IPCC Assessment Reports Has Shifted to Lower Temperatures. Earth’s Future, 10(5), e2022EF002876. https://doi.org/10.1029/2022EF002876

• Meinshausen, M., Schleussner, C.-F., Beyer, K., Bodeker, G., Boucher, O., Canadell, J. G., Daniel, J. S., Diongue-Niang, A., Driouech, F., Fischer, E., Forster, P., Grose, M., Hansen, G., Hausfather, Z., Ilyina, T., Kikstra, J. S., Kimutai, J., King, A. D., Lee, J.-Y., … Nicholls, Z. (2024). A perspective on the next generation of Earth system model scenarios: Towards representative emission pathways (REPs). Geoscientific Model Development, 17(11), 4533–4559. https://doi.org/10.5194/gmd-17-4533-2024

• Minobe, S., Behrens, E., Findell, K. L., Loeb, N. G., Meyssignac, B., & Sutton, R. (2025). Global and regional drivers for exceptional climate extremes in 2023-2024: Beyond the new normal. Npj Climate and Atmospheric Science, 8(1), 138. https://doi.org/10.1038/s41612-025-00996-z

• UN Environment Programme. (2023). Emissions Gap Report 2023: Broken Record. https://www.unep.org/interactives/emissions-gap-report/2023/

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

Space weather often comes up as a topic when discussing what environmental hazards might lead to societal collapse or at least a massive societal disruption. The main idea here is that geomagnetic storms coming from the Sun create failures in the electrical grid so large that they cannot be repaired in a reasonable timeframe, resulting in a permanent blackout. And as we have learned from other posts in this series, having a blackout that lasts longer than a few days would be really, really bad.

The odd thing is that the damage assessments that I have come across differ wildly, ranging from “a large geomagnetic storm would totally overwhelm global society” to “this is a non-problem, which could just lead to minor disruptions and a short, local blackout at worst”. So, let’s explore this a bit deeper which of those two views is closer to reality.

Subscribe now

What is a geomagnetic storm?

Before we dive into the details of damages and mitigation, first let’s define a few key terms around space weather. I am relying for this section and the next one on nice reviews by Baker et al. (2014), Johnson et al. (2016) and Eastwood et al. (2017):

• Space weather: All the radiation and particles the Sun throws at us and how they interact with Earth’s magnetic field.

• Solar flares: Sudden bursts of electromagnetic radiation from the Sun’s surface, ranging from radio waves to X-rays and gamma rays, caused by the release of magnetic energy.

• Solar wind: The continuous stream of charged particles (mainly electrons and protons) that flows outward from the Sun in all directions.

• Coronal mass ejections (CME): This is when the Sun shoots especially large amounts of charged particles into space.

• Geomagnetic storm: The large magnetic disturbances in Earth’s magnetic field which happen when strong solar wind (often enhanced by CMEs) hits Earth.

How the damage happens

When a geomagnetic storm happens, the geomagnetic field around Earth twists and bends under the onslaught of the Sun’s emissions. This movement creates geomagnetically induced strong currents in everything on Earth which is large, and able to conduct electricity. The things thus most affected are the electrical grid and pipelines. Both represent long metal lines, able to capture the currents. For pipelines this is not that big of a problem, as it just increases corrosion. This is bad for long term maintenance, but not something catastrophic. Train lines sometimes have short-term disruption or there can be fires at relay stations. Things are different for the electrical grid. Long distance transmission lines need low electric resistance to be able to work well. But this low resistance also means that the currents from the geomagnetic storm are more easily absorbed by them. Essentially, these long distance transmission lines work as gigantic antennas which capture the currents induced by the storm. If current in the grid fluctuates wildly due to the geomagnetic storm this can cause damage and trigger safety shut downs.

Some experts argue that the existing protections put into place against low frequency currents like from the geomagnetic storms are insufficient to prevent large-scale damage to the electrical grid. This manifests especially in the electrical transformers. The additional current can damage them, let them age more quickly due to heating and lead to potentially complete loss of function. Generally, this damage builds up over hours of exposure and is less an instant failure of the system. Still, this damage is a big problem, because these transformers are fairly big and if you order one today, it can be months to years until it gets delivered. Delivery times would obviously be longer if this happens at a lot of places at once. Therefore, theoretically, you could end up in a situation where blackouts last for weeks or even longer. This would be highly destructive, and we explore how destructive further down in this post (1).

• Collapse of the electrical grid is clearly the worst outcome we can get from a geomagnetic storm. However, it is not the only one we should be concerned about:

• Communication disruptions: Geomagnetic storms inject energy into the ionosphere, this can disrupt high frequency radio.

• Disruption of global synchronization and navigation: Global navigation satellite systems (things like GPS) could be disrupted for several days. This would be annoying for navigation in general, but more importantly, we also use it for global synchronization for all kinds of things, like our financial system. This would not be possible during the storm.

• Aviation problems: Airplanes might be exposed to higher radiation doses than normal. This would not have any immediate effects, but would condemn a random collection of people on the airplanes to cancer later in their life. Also communications with airplanes might be lost and their navigation systems could be affected.

• Cascading damage: The storms can cause voltage fluctuations on the power lines which can damage other components and lead to automatic cascading grid shutdown routines

Historical examples

From history we can point to several examples of massive geomagnetic storms hitting Earth (also described in Eastwood et al. (2017)):

• 1859: The Carrington Event is the largest geomagnetic storm which hit Earth since we can accurately record them. It disrupted telegraph lines in many places, but ultimately did not cause much damage, due to the lack of very vulnerable parts of society at that time.

• 1921: This storm was likely on a similar magnitude as the Carrington Event, but we can’t be sure due to uncertainty in assessing long past events. It caused fires in several telegraph stations in Sweden.

• 1967: This storm was large enough to strongly disrupt communication in the United States. In particular, the military got quite concerned about it and thus started to look into space weather effects more seriously.

• 1989: Roughly 1.5-3 times smaller than Carrington, this storm led to a large blackout in Canada and damaged two larger transformers in the United Kingdom.

• 2003: Large, but not massive storm led to a short blackout in Sweden.

• 2012: A Carrington class storm was ejected from the Sun, but missed Earth.

• 2024: This was the most recent large geomagnetic storm which hit Earth. While it led to auroras to be visible all throughout Europe, its geomagnetic impacts were somewhat smaller than the 1989 storm, meaning not much damage occurred (a comparison of this storm with earlier ones is done in Lawrence et al. (2025)).

Where on Earth does this happen?

The effects of the geomagnetic storms would not be felt globally all at once. However, the storms can last up to a week and the Earth rotates every 24 hours, meaning that every place on Earth could potentially be reached. But in the exposed areas the distribution is not random. Due to the shape of the magnetic field there are preferential areas that could be hit. Thankfully, Maffei et al. (2023) used some physics based models to predict where these areas are today and also where they will be in five decades from now. They slowly change, as the Earth’s position in space is not completely stable, but wobbles around. Figure 1 shows the areas with the highest probability of extreme space weather hitting that area. This shows that Northern Europe, Southern Canada and the Northern United States have the highest chance of being hit. However, this does not mean that everywhere else is safe. It just means that the likelihood is lower. During the Carrington Event auroras could be seen as low as the equator.

Figure 1: Bands of highest probability of large geomagnetic field fluctuations due to geomagnetic storms.

A second global effect is disruption to satellites orbiting the Earth. As the coronal mass ejections push more particles towards Earth, the density of the atmosphere increases. This in turn means that satellites experience more drag, lose speed and thus altitude. Many satellites have thrusters on board to counter this, but in a very severe storm those might not be enough, leading to a loss of an unknown, but potentially high number of satellites.

How large could these storms get?

The impacts of the storms depend a lot on how large they are. We clearly know that Carrington Event sized storms can happen and we should expect them roughly every 150. However, this is not the absolute maximum. The size of the storms follows a very fat tailed distribution, meaning much larger, but rarer events are possible. Unfortunately, we don’t really have a long time series from the Sun’s activity. What we do have is a sky full of stars, many of which are similar to the Sun. Vasilyev et al. (2024) used a large sample of Sun-like stars to check how often they produce large solar flares. They found that in their sample the stars produced flares 1-1.5 magnitudes larger than Carrington roughly every 100 years. This means the Sun might have some nasty surprises for us in store, but we do not know for sure if the Sun could even produce storms of this magnitude.

There is also some research that looks into paleo records of how large the Sun’s flares can become. A recent example of this is Bard et al. (2023). When solar flares hit Earth’s atmosphere they create radioactive carbon-14. Trees absorb this and fix it in their growth rings. This means when you have very old remains of trees you can use the carbon-14 concentration as a proxy for the size of solar flares that hit Earth while the tree was still growing. Bard and his co-authors were able to find 14,300 year old tree remains. In those trees they found carbon-14 concentrations which indicate that a solar flare around a magnitude larger than the Carrington Event must have hit Earth.

These two lines of argument imply that the Sun is probably capable of creating geomagnetic storms of a magnitude larger than the Carrington Event. However, the maximum size of a geomagnetic storm is limited by the strength of the geomagnetic field as well. Vasyliūnas (2011) created a theoretical physical model and uses it to argue that the largest possible storm change in magnetic field is roughly 1.5 times the size of the Carrington event in terms of depression of the magnetic field at the tropics, although this is not exactly the same as the peak rate of change of magnetic field which is responsible for high voltage transformer overheating and other electricity grid issues. So, storms above 1.5 Carrington size could still produce more damages, for example by more quickly ramping up to the maximum value or changing more rapidly, even though their maximum value of magnetic field strength depression is topped out.

The dangers might also be getting more imminent. Jasinski & Velli (2025) looked at activity data from the Sun of the last few decades and found that the Sun is currently on a trend of increased activity in general, meaning that for the next few decades the occurrence of a larger geomagnetic storm might be higher. The magnetic field of the earth is also on a weakening trend, decreasing the natural geomagnetic shield from CME charged particle radiation.

Estimates of damage

As we thankfully have not yet faced a Carrington-sized event with our modern, highly digital civilization, we have to rely on modeling to understand potential damages. Estimates vary widely. The most pessimistic, commonly-cited paper is by Schulte in den Bäumen et al. (2014). They modeled geomagnetic storms similar to the 1989 Quebec event, determining which geographic areas would likely be affected (Figure 1). They assumed affected regions would lose 10% of their electricity production capacity for a year, which is the estimated time needed to replace damaged transformers and restore grid infrastructure. Using a global supply chain model covering 187 countries and ~16,000 economic sectors, they traced how this electricity shortage would cascade: production losses in electricity-dependent sectors, disrupted supply chains, and reduced international trade as countries cannot import goods that haven’t been produced. This yielded estimated global losses of 2.4-3.4 trillion dollars (3.9-5.6% of global GDP), with roughly half the damage occurring in countries outside the directly affected zones. The difference in their estimate to what actually happened in 1989 is that they assume regions more important to the global economy are hit. But their research seems like a worst case under pessimistic assumptions to me. They also model Carrington sized events, but never state how much those would cost. If you naively scale this on the magnitude of the storm, you end up with something like 5-10 trillion dollars in damages.

On the other side of the scientific spectrum we have Oughton et al. (2019). They first simulated the electrical grid of the United Kingdom and how geomagnetic storms of different sizes would create damages in the grid. Based on this they then check how much the economy gets disrupted and how this cascades into supply chains. Their estimate is that a Carrington sized event would leave up to 60 million people in the United Kingdom without power, but assume that the grid would be restored in days to weeks. The damage of 20 billion dollars they come up with is then based on lost economic transactions during the period. Though this does not consider global effects in general and also does not model that other countries might face the same problem. Finally, it also does not consider the damages that would occur during a week-long blackout, which would likely be much more severe than just lost economic transactions. If we naively scale this up to global numbers based on the United Kingdom’s contribution to global GDP, we end up with roughly 660 billion $ damages globally.

The difference between those numbers is the roughly difference between the amount of money the United States spends on military and a third of the total United States GDP. This is a massive difference. And even the larger number does not really grapple with what happens if a region experiences a longer blackout. Also, disruption of satellites, especially megaconstellations and global navigation satellite systems are not considered here at all.

Finally, there are also the grid operators themselves who maintain and supervise the grid. They are acutely aware that these things can happen and over the last decades the countries most likely affected by geomagnetic storms have hardened their grid to withstand them. However, it is quite difficult to get a picture of how far these things have actually come. The most recent big government report I could find was the U.S. Department of Energy (2019). They gave an overview of the state of preparations in the United States. It states:

• The industry is required to have plans for what they do when geomagnetic storms hit.

• They have to do a damage assessment every 60 months.

• The industry is in discussion with the government about implementing physical grid hardening.

• Major gaps remain in understanding how large geomagnetic storms would interact with the grid.

Overall, it sounds like the risk is taken seriously, preparations are being worked on, but are far from finished. This report is only for the United States, but I would assume that countries that are even more exposed like Canada or Sweden are prepared better, while countries like China, which aren’t affected during smaller geomagnetic storms are likely much worse and would have big troubles in severe storms. Due to this and many other uncertainties, the range of damages possible seems quite large to me.

How can we prepare?

Thankfully, there are at least some things that can be done to prevent or at least decrease damage when a geomagnetic storm hits. Especially important here is accurate forecasting of geomagnetic storms and their impacts. Oughton et al. (2019) also look at scenarios with better forecasting than we have today and this reduces their damage estimate by two thirds, as the most vulnerable parts of the electrical grid can just be disconnected. This means they are not damaged and can be added to the network again immediately after the storm ends. Generally, it seems to me that forecasting is the one area where the most progress was made in the last decade or so, with satellites being added for monitoring and the forecasts now usually good enough to predict geomagnetic storms by around ten hours. This means due to better forecasting a large amount of damages could likely be prevented, even in larger storms.

Besides forecasting, there are also several other options on how you could change the physical infrastructure to be more resilient (as explained in Johnson et al. (2016) and implemented in Canada, the US, and Sweden):

• Capacitor blocking: This means installing low-ohmic resistors in the grid. These slow down the flow of the incoming currents, providing protection by reducing the maximal loads this way.

• Transformer redundancy: Just have some spare transformers that you can use to replace damaged ones quickly.

• Grid topology changes: Include grounding points every 50 to 75 km, to limit how much harmful current can accumulate in transmission lines. Also, having alternative routes in your grid, so disrupted lines don’t bring the whole thing down.

Conclusion

The existing damage estimates appear somewhat inaccurate and omit important pathways, such as damages from extended blackouts. Moreover, the Sun can likely produce geomagnetic storms considerably larger than the Carrington Event, yet nearly all papers treat Carrington as the worst-case scenario. Yet, there appears to be an upper limit to storm severity based on how Earth’s magnetic field functions.

On the mitigation side, forecasting has improved substantially, and the past decade has seen many grid hardening measures implemented. Yet government reports indicate much work remains and large uncertainties persist. Given the current state of knowledge, the potential severity of geomagnetic storms remains unclear.

That said, there are reasons for cautious optimism. This is an active research field with relatively strong government interest compared to other global risks, suggesting the dangers may decrease over time. Additionally, geomagnetic storms appear to be more regional than global in nature when it comes to the worst impacts. While effects could potentially be felt across most of the globe, transformer damage would likely affect only certain regions. This would certainly be expensive to repair and could cost many lives if managed poorly, but there doesn’t seem to be a realistic pathway to a global blackout.

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

Subscribe now

How to cite

Jehn, F. U. (2025, October 29). Auroras, space weather and the threat to critical infrastructure. Existential Crunch. https://doi.org/10.59350/g95ww-xym22

Endnotes

(1) One caveat: many high voltage transformers have backups at the station already, which would take days or weeks, rather than months, to be put into position and powered on.

References

• Baker, D. N., Jackson, J. M., & Thompson, L. K. (2014). Predicting and mitigating socio-economic impacts of extreme space weather: Benefits of improved forecasts. In A. Ismail-Zadeh, A. Kijko, I. Zaliapin, J. Urrutia Fucugauchi, & K. Takeuchi (Eds.), Extreme Natural Hazards, Disaster Risks and Societal Implications (pp. 113–125). Cambridge University Press. https://doi.org/10.1017/CBO9781139523905.012

• Bard, E., Miramont, C., Capano, M., Guibal, F., Marschal, C., Rostek, F., Tuna, T., Fagault, Y., & Heaton, T. J. (2023). A radiocarbon spike at 14 300 cal yr BP in subfossil trees provides the impulse response function of the global carbon cycle during the Late Glacial. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 381(2261), 20220206. https://doi.org/10.1098/rsta.2022.0206

• Eastwood, J. P., Biffis, E., Hapgood, M. A., Green, L., Bisi, M. M., Bentley, R. D., Wicks, R., McKinnell, L.-A., Gibbs, M., & Burnett, C. (2017). The Economic Impact of Space Weather: Where Do We Stand? Risk Analysis, 37(2), 206–218. https://doi.org/10.1111/risa.12765

• Jasinski, J. M., & Velli, M. (2025). The Sun Reversed Its Decades-long Weakening Trend in 2008. The Astrophysical Journal Letters, 990(2), L55. https://doi.org/10.3847/2041-8213/adf3a6

• Johnson, M., Gorospe, G., Landry, J., & Schuster, A. (2016). Review of mitigation technologies for terrestrial power grids against space weather effects. International Journal of Electrical Power & Energy Systems, 82, 382–391. https://doi.org/10.1016/j.ijepes.2016.02.049

• Lawrence, E., Beggan, C. D., Richardson, G. S., Reay, S., Thompson, V., Clarke, E., Orr, L., Hübert, J., & Smedley, A. R. D. (2025). The geomagnetic and geoelectric response to the May 2024 geomagnetic storm in the United Kingdom. Frontiers in Astronomy and Space Sciences, 12. https://doi.org/10.3389/fspas.2025.1550923

• Maffei, S., Eggington, J. W. B., Livermore, P. W., Mound, J. E., Sanchez, S., Eastwood, J. P., & Freeman, M. P. (2023). Climatological predictions of the auroral zone locations driven by moderate and severe space weather events. Scientific Reports, 13(1), 779. https://doi.org/10.1038/s41598-022-25704-2

• Oughton, E. J., Hapgood, M., Richardson, G. S., Beggan, C. D., Thomson, A. W. P., Gibbs, M., Burnett, C., Gaunt, C. T., Trichas, M., Dada, R., & Horne, R. B. (2019). A Risk Assessment Framework for the Socioeconomic Impacts of Electricity Transmission Infrastructure Failure Due to Space Weather: An Application to the United Kingdom. Risk Analysis, 39(5), 1022–1043. https://doi.org/10.1111/risa.13229

• Schulte in den Bäumen, H., Moran, D., Lenzen, M., Cairns, I., & Steenge, A. (2014). How severe space weather can disrupt global supply chains. Natural Hazards and Earth System Sciences, 14(10), 2749–2759. https://doi.org/10.5194/nhess-14-2749-2014

• U.S. Departement of Energy. (2019). Geomagnetic Disturbance Monitoring Approach and Implementation Strategies. U.S. Departement of Energy. https://www.energy.gov/sites/prod/files/2019/06/f64/DOE_GMD_Monitoring_January2019_508v2.pdf

• Vasilyev, V., Reinhold, T., Shapiro, A. I., Usoskin, I., Krivova, N. A., Maehara, H., Notsu, Y., Brun, A. S., Solanki, S. K., & Gizon, L. (2024). Sun-like stars produce superflares roughly once per century. Science, 386(6727), 1301–1305. https://doi.org/10.1126/science.adl5441

• Vasyliūnas, V. M. (2011). The largest imaginable magnetic storm. Journal of Atmospheric and Solar-Terrestrial Physics, 73(11), 1444–1446. https://doi.org/10.1016/j.jastp.2010.05.012

Life feels more rough, because it is. At least this is what I took as one of the main messages from Peter Turchin’s book “End Times: Elites, Counter-Elites, and the Path of Political Disintegration”. In it Turchin takes us on a grand tour of his structural demographic theory and how its insights can be applied to modern day America.

Subscribe now

The mechanisms of structural demographic theory

The main idea behind structural demographic theory is that societies tend to regularly experience disintegrative phases and that the processes behind this pattern can be quantified and analyzed mathematically. How often you end up in trouble mostly is determined by how your society is set up and especially how many elites you have in comparison with how many positions you have for those elites. If an elite does not have a position that they think they deserve, they will use their powers to shape society in such a way that they get a position after all. This can quickly spiral into much disruption or even things like revolt or civil war, as those disgruntled elites often are open to quite drastic measures to get into power again. The number of elites is mostly determined by how many kids the elites had and by the inequality of wealth in a given country. When they have more kids and the money to educate them, elite numbers can rise quickly. Additionally, wealth means power and if you have much more wealth than almost anybody else, you can block others from positions of power, further increasing competition.

But before we dig into all this and how it currently plays out in the US, we first have to understand a few terms and concepts.

What are elites even?

Elite can be a vague term and people’s definition of what an elite is can differ. In the context of structural demographic theory an elite simply is a person who holds power in society. Especially in capitalist societies, the amount of power you have very strongly correlates with the amount of money you have. However, in addition to elites who wield power through money, you can also have elites who wield power by coercion (an army general), administration (the head of the central bank) or persuasion (your favorite celebrity). You can also have several kinds of power unified in one person. For example, economic power can be used to also gain political power.

The main mechanisms

The wealth pump

One thing that elites tend to do is to extract wealth from the rest of society. While this is manageable at some level, if it increases over time you have a problem. Turchin use the term “wealth pump” for this, meaning that an ever increasing share of the wealth your society has is ever more concentrated in a small group of people. This increasing wealth concentration has several consequences.

Elite overproduction and counter-elites

The first thing is that you have an increasing number of elites. Because the easiest way to become an elite is to simply have rich parents. If the elites grab an ever increasing share of wealth from society, this means elites can distribute more wealth to more of their children. A second way to become an elite is upward social mobility. However, your society only has so many positions for those wannabe elites. Ultimately, only one person can be the president, no matter how many people would like to have this position. Therefore, the more elites you have, the more of those elites feel like they should be on top of society, but there is just no room for them anymore. These failed elites then become counter-elites, which means people who want to upend the current system, so they can get into an elite position after all.

To get back to a stable society again, you have to somehow reduce elite numbers again or expand the number of elite positions. Historically, this has mostly happened by overproduced elites simply being killed off in wars, purges, epidemics and the like. But there are also a few instances, like the New Deal, where this elite reduction happened by non-violent means and through re-distribution of wealth.

Popular immiseration

In parallel to elites getting richer, the general population gets poorer, because the additional wealth for the rich has to come from somewhere. When you get poorer, you usually also lose access to things that are valuable to you, think health care, education or freedom to do the things you want to do. This sucks, so people also tend to get less happy when they get poorer. Over time this builds up resentment in the population, which destabilizes society. This gets resolved once more of the societal wealth is distributed again to the poorer parts of society.

Falling fiscal health and weakened legitimacy

As wealth gets siphoned away towards the rich, this also means that the state gets less of it. However, to maintain public services, the state needs a certain amount of money. So, as the state gets poorer, it also loses its ability to provide for its citizens. You can actually see this in many present-day countries where public services have gotten worse and worse over the last few decades, as neoliberal politics reduced taxes and allowed more wealth accumulation in the hands of the few.

These deteriorating public services are a problem for your society, because it makes life worse for everyone and it also reduces the legitimacy of the current regime. If everything around you crumbles, the bus only comes once a day and your public swimming pool had to to close down, it gets difficult to feel like your society is making progress or moving in the right direction. This failure gets associated with the current regime and also partly with the way your society is organized in general. Like if you are in a democracy, and everything crumbles, people get the idea that democracy is not really working for them.

Outside pressure (geopolitics, climate, pandemics)

All of the three things I mentioned above (elite overproduction, popular immiseration, failing fiscal health) contribute to reducing the resilience of your society. The problem is that your society gets somewhat regularly confronted with outside pressure. This can be climate (a drought), a pandemic (black death) or a neighbouring state deciding that they want to wage a war against you. Your society does not really have much control if these things happen and on what magnitude. This means we can imagine your society as having a certain resilience value, based on how well it manages elite overproduction, popular immiseration and fiscal health. This resilience value is then tested by outside pressure on random times and in a random strength. If the pressure is larger than your resilience, you collapse or at least have a major societal crisis. This means even very resilient states can fail when they are challenged by a large outside pressure, but also that very fragile states can last for a long time if they get lucky. If you collapse or have a big crisis, this usually resets the wealth distribution in your society and the cycle can start anew.

Figure 1: Flow chart of how the basic mechanism of structural demographic theory works.

Additional mechanisms

Besides this main loop of structural demographic theory, Turchin also describes a bunch of other processes which can happen and which reinforce the loop or explain parts of it better.

Goldstone’s theory of ideological change

In addition to the main structural demographic cycle explained above, there are also other mechanisms at work. One of them is how the ideological landscape fractures once there are too many elites. All of these failed elites try to push their narrative of why the country is struggling and why it was not able to provide them the elite position they deserved - typically with themselves placed as leaders of these splinter ideological camps. The phases of how this fracturing developed were described by Jack Goldstone. He thinks the fracturing cycle happens in three main steps: The state struggles to maintain the power over the societal narratives, as more and more counter-elites try to push their narratives. Once the old regime has lost legitimacy completely (in the past often accompanied by collapse), the different factions in the counter-elites fight for the authority over the remaining society. One of those fractions wins the authority, which allows them to suppress the other narratives by reconstructing new political, religious and social institutions.

The ruling class

Who controls society? A question often asked, and often answered very differently. In many early states the question was easy to answer. Those who controlled the military, controlled the state. But over time persuasion became a much more important factor. It is much better to convince others that helping you is in their self interest, then forcing them at gunpoint. In the past this was often done through religion. This is the reason you find so many god kings in history. They merge the power of coercion and persuasion. Religion is also still used today, like in Iran where the regime clearly builds its power on religion, but also in the United States where many things are still justified using Christianity.

In many countries today, the military is still an important factor, but in most mature democracies the military is not really a power factor anymore. Especially in the United States if you belong in the ruling class or not is based on your wealth. Running for office is so expensive now that only independently wealthy or highly sponsored politicians even have a chance in the race to get elected.

The motives to get to power are quite obvious here. Most people want their wealth to grow rather than diminish. This seems to be especially the case for the rich in society. But while they might try to have policies implemented that profits them in some way, there is no big overarching plan. Just many small interactions of the wealthy interacting with the policy makers. Over time aligns the interest of society with their interest. This also means there is no power center. Just 1000s of rich individuals follow their own incentives, but as a whole class they push society towards their collective needs. And none of this is a secret. There is no hidden kabal. Just people following their incentives, which unfortunately leads to bad societal outcomes.

How structural demographic theory plays out in the United States

The two narratives of American society

After we now have the basics of structural demographic theory down, we can look at how all those processes play out in the United States. In the United States there are currently two main narratives of how the country explains itself to itself, think slogans like “The system abandoned the white working class” versus “The data proves life is better, the bigots are just mad about losing privilege”. Both groups who tell these narratives think that the other group is one that is really destroying the country they love.

The first narrative is those of the poor, white working class. For them their life just got worse and worse over the last few decades. Everything in their life is on the cusp of breakdown, their health, the infrastructure around them, everything crumbles. This leaves them frustrated and especially when other people call them privileged for being white. They don’t feel privileged. Don’t the others see that they are barely hanging on?

The other narrative is that the last 100 years of the United States have been a clear success story. The life of everyone is getting better (just look at the “Our World in Data” graphs!). If you work hard, you can get rich and live the life you always dreamed of. If you are poor or your life just does not come together, it is just because you did not work hard enough. If you would really be trying, you would be thriving. If people claim that things are getting worse, they are simply rejecting scientific data. Why can’t they see that they are so obviously wrong?

What are the trends in the United States when it comes to societal outcomes?

The tricky thing with both of these narratives is that they are partly true and partly wrong. Both are anchored in reality, but only in a part of it, only for a certain group. Life in the United States really has gotten significantly better over the last decades, but only for the richest 10% or so. The vast majority of people however, are worse off than their parents. GDP might be rising, but if the real wages decline each year, this just means that the majority of people get poorer even as the GDP is rising. Two thirds of the American population earn less in real wages than their parents did, while at the same time costs for the basic necessities of life like health, education and housing have all gone through the roof. This makes getting it even harder to get the education you would need to even have a chance of making it. All of these trends have been especially strong for men and Black Americans.

We can find the effects of this even in the physical and mental health of Americans. Since the 1960s Americans have not been getting any taller, while other rich countries continued to grow in height. Even more stark is the trend reversal in life expectancy. It has been on a downward trend in the United States since around 2013, while also still climbing for other rich countries.

This tough situation for many Americans is also mirrored in how they cope mentally. Over the last decades the deaths of despair have been rising. These are deaths caused by suicide and drug abuse. Here too the burden is especially high on middle aged men with little education.

So, we see life in the United States can be great, but only if you are rich. If you are in the majority of the population, life in the United States is hard and has been getting harder for decades. But how did the United States even end up like this?

Historical causes for the mess the United States are in right now

The plutocratic roots of the United States

Around 1500 plutocracies were quite common in Europe. However, pretty much all of them were conquered by militocracies in the few hundred years afterwards. This constant pressure of warfare weeded out the states that could not fend for themselves and it turned out that states run by the wealthiest members of society just weren’t very good at that. However, we had one exception in Europe, where plutocratic rule could be maintained: England. England was able to maintain only a relatively small army by being an easily defendable island. This was enough to deter most attempts to conquer the island. And this rare plutocratic state was one major contributor to the colonization of the American continent. This gave the United States strong plutocratic roots. And as the United States also ended up in defendable territory, with neighbors militarily weaker than itself, it never really had to give up on plutocratic rule.

The American Civil War

Before the American Civil War there was discontent brewing in the United States. The center of power were the aristocratic Southern slaveholders in combination with Northeastern merchants, bankers and layers. The South produced the agricultural goods and the North made sure to sell it all around the world. In this arrangement, the South was generally more important, as the structure of the United States at the time led them to control half of the Senate, even though they had a lower population. Also, most of the richest Americans all lived in the South.

In the decades before the civil war you saw several signs of things getting worse. The relative wage declined, there were more urban riots and the number of millionaires skyrocketed. As the country industrialized the Northern elites got richer than the Southern ones and increasingly disliked that they had more money, but less power. They wanted to reform the state in a way that was better to their business interest. However, these reforms would have been to the detriment of the South, which therefore blocked it. As the conflict lines deepened, you also had more elite overproduction in parallel. It was caused by industrialization creating sustained economic growth while labor oversupply (from rapid population growth and immigration) kept worker wages stagnant, directing all the fruits of that economic growth to the wealthy class instead. We can also see more and more violence being done (like more riots and people having fist fights in the Senate), the closer we get to the outbreak of the civil war.

While the general story of the American Civil War is that it was fought to free the slaves, Turchin thinks this is only a fig leaf for the real reason: Slavery was the economic foundation of the Southern economy. If you remove slavery, the South would decline in importance and thus the Northern elites could do what they want.

Racism kills solidarity

One major hurdle to end plutocratic rule in the United States was racism. In many European countries you had large socialist movements, which united all of the working class and gave them strong leverage to push for change. Even though much of the gains from these days have been scaled back again in Europe, they pushed harder than the United States and managed to slow down scale back more efficiently. The working class in the United States often failed at this, because they were split by racism. The plutocratic elite could use racism to drive a wedge between black and white workers. This decreased their bargaining power and left them in a worse position than their European counterparts. Due to this, the American workers never managed to force so much redistribution as in Europe and the plutocratic elites could maintain their position.

The reforms of the Progressive Era

Still, we can see that the labour movement in America was able to achieve some wins during the Progressive Era (roughly 1890s to 1920s). What changed during the Progressive Era? Several factors converged to force even American elites to make concessions. The two decades around 1920 saw unprecedented crisis pressure: nearly four million workers struck in 1919 alone, the Battle of Blair Mountain became the largest armed insurrection since the Civil War, and race riots killed over 1,000 people during the “Red Summer.” This wasn’t periodic unrest, it was a sustained threat to the entire system. External events like the Russian Revolution provided additional pressure, showing that entire systems could collapse if they lost legitimacy.

Crucially, elites realized they could make strategic concessions without abandoning their fundamental position. The immigration restrictions of 1921-1924 reduced labor oversupply and boosted wages, but business elites accepted this constraint because it helped stabilize the system. Progressive reforms often worked around racial divisions rather than challenging them directly, like focusing on white-dominated industries while reinforcing certain racial boundaries. These kinds of compromises brought some peace to society and laid the groundwork for the New Deal.

The New Deal almost turned the country around for good

The New Deal reshaped American society towards more redistribution of wealth. It had become possible because in the decades before, discontent had been rising sharply again. More and more labor conflicts of which an ever higher percentage also turned violent. Sometimes even resulting in federal troops being sent in to kill striking workers. The New Deal didn’t overcome racism - it temporarily sidestepped it by explicitly excluding Black Americans from the social contract, just like in the reforms of the Progressive Era. This morally compromised arrangement enabled white working-class solidarity by removing the racial economic competition that plutocrats had previously exploited. Additionally, the severity of the Depression crisis made some elites willing to accept significant redistributive policies to prevent complete system collapse.

At the same time, you had strong ideologies pushing for change all over the world. Anarchism, communism and facism all represented themselves as an alternative to capitalism. This danger coming both from within and without forced the American elites to acknowledge that something had to change. This change was the New Deal. It introduced that the rich had to pay considerably more in taxes and that this income would be redistributed to the poor and built up the infrastructure of the United States. These policies decreased wealth inequality without any bloodshed, thus decreased elite overproduction and popular immiseration and led to several decades of social peace.

One of the most important outcomes of the New Deal was an unwritten contract between business, workers and the state. It gave workers the right to organize and bargain collectively and that the gains from economic growth would be shared with everyone in society. A social cooperation between commoners, elites and the state.

Breakdown of the consensus on how society should look like and the rise of Reagen

While the New Deal brought peace to American society for a few decades, this all started to come apart in the 1970s and 1980s, but the roots of this can already be found in the 1950s. If you track how far away from each other the Democrats and Republicans are politically, you see that around the 1950s they started to drift apart, with the left wing people sorting themselves into the Democrats and the right wing people into the Republicans. This sorting and thus the political polarization have only been going up since then. By now both parties’ more extreme wings argue for a complete restructuring of society. This means we are in the middle of the ideological change cycle by Jack Goldstone which I described above.

A clear break from the New Deal consensus can be seen especially in the 1980s. The people coming to power at that time, exemplified by Ronald Reagan (1), started to dismantle the pillars of postwar society, guided by a new economic theory, neoliberalism, which argued that rules and redistribution are bad and that capitalism should be unchained. This disruption is when we see the start of the decline of real wages. The ideas of neoliberalism were supported by neighbouring anti-collective ideas like objectivism and meritocracy. These argued for egoism actually being good, collective action being bad and how rich you are only reflects how hard you worked and nothing else (2). All of these ideas are excellent if you really want to make the wealth pump go brrrr (3), but not that great if you want to have a stable society for the long term. It might be able to pump up your GDP, but what is a high GDP worth if people die earlier each year?

Everybody goes to university now

As all of these processes jacked up the wealth pump, it predictably led to an overproduction of elites. In the 1950s around 15 % of the population went to college. In the United States of today, it is around two thirds. This means, even if you are really good, your chances of making it to the top are slim. This leaves a lot of people frustrated. Wasn’t the idea that education would make sure you get your share of the pie? In the American system this gets worsened by the high amount of student debt that people accumulate while getting an education. Not being successful after going to university hurts even more if you are crushed by debt afterwards.

This higher competition between elites also leads to more cheating and other anti-social behaviour. This happens because people get more desperate, when they realize the competition is so high that they probably will not make it, no matter how hard they try. It was further emphasized by philosophical ideas like objectivism, which tell you that it is fine to only look out for yourself.

No political influence for the poor

All these forces have led to ever diminishing influence of the poor in American society. If you have money, you can easily influence the public opinion, by pushing your narratives in the media. Also, you can invest in lobbying and buy your access to politicians directly. There is an extensive piece of research about this by the political scientist Martin Gillens. He gathered thousands of policy issues between 1981 and 2002. For each of those issues he checked what was ultimately decided and how public opinion was thinking about the issue. He found the political outcomes in the United States only aligned with the interests and wishes of the richest 10 % and their interest groups. Everybody else’s preferences were simply not reflected in national policy decisions. This is a consequence of decades of shifts in wealth in power in American society.

So, what can we expect for the future?

Given the trends we have seen above, what might the future hold for the United States? Turchin is relatively pessimistic here. He thinks that the 2020s are bound to be a rough time, no matter what we do now. The problems are just too deeply entrenched to quickly resolve them. The first chance he sees of conditions getting better are the 2030s, but only if we manage to lay the groundwork for this better trajectory now (4). To get to this better future, the United States has to find a way to break free again from their plutocratic rule, just as they did during the New Deal. If they don’t manage this Turchin predicts that things will only get worse, with ever more violence and disruption. This could get as bad as a new civil war.

But who might bring about this change? The problem in the American society of today is that pretty much all of the elites are still deeply entrenched in neoliberal economic ideas. Up until recently the main political fights in the United States were all centered on culture war issues (and it is still a major focus). There was little disagreement about most of the economic policy between Democrats and Republicans. But on the edges of both parties, there are dissidents who think about how things might be done differently in the future.

The dissident left is centered around people like Bernie Sanders or Alexandria Ocasio-Cortez. They think about more socialist ways to organize the American economy and are the reaction to the abandonment of the working class by the Democrats under Bill Clinton. The dissident left had their chance to power when Bernie Sanders almost became the presidential candidate, but they were overpowered by the more established, neoliberal Democrats. Since then, the dissident left did not have much to celebrate. Their cultural ideas are much discussed in the form of culture war, but their criticism of existing economic and military structures rarely finds its way into mainstream media.

The dissident right on the other hand has been much more successful over the last decade. They found their champion in Donald Trump. While Trump sees himself as a classic elite, many of his followers and allies (like Steve Bannon) see themselves as a revolutionary force (counter-elites). Their first attempt to dismantle the system after the election of 2016 failed. However, instead of assimilating into the ruling class, they became more radicalized by this experience. This radicalization has spread to the Republicans in general and by now a large chunk of them have taken up the ideas of the dissident right. Some of the ideas are even aimed against the interest of the plutocratic elite, like tariffs for example. But mostly they just seem out for disruption and domination. The big question is: How successful will they be in their second attempt to break the current system?

Obviously, Turchin cannot answer this question and the history of future America remains to be written. However, he points to history again to give us an idea of how things might play out. When we look at historical crises, it becomes clear that most societies follow a very similar path into crisis. Specifically, the one described in structural demographic theory. The wealth pump activates and things go downhill from there. However, once the actual crisis starts, things get interesting. Because here, history shows that the path of societies diverge again and every society seems to map out their own unique path on how they resolve their crisis. Unfortunately, history also shows that in the vast majority of examples, the crisis is not resolved peacefully, but by the horsemen of the apocalypse. But still, there are at least some examples in history where societies were able to solve their problems peacefully. Turchin thinks the best examples here are England in the Chartist period, (1819-1876) and Russia in its Reform Period (1855-1881). Both countries used peaceful reform and a redistribution of wealth from the rich to the poor to bring back societal stability. Or even in the United States we can point to the New Deal as an almost successful example of making society more equal again for the long term.

This means there is hope that we can resolve the crisis of the future peacefully as well. Europe will be interesting to watch here, as we have many nations with a relatively similar structure and environment, but wildly differing levels of inequality. It will be a good test case for the idea that low inequality makes societies more resilient.

Conclusions

At first glance, this book does not offer you much to hope for. It mainly seems to say that times are tough, and will remain tough for the foreseeable future. If we move things in the right direction, they might get a bit better again in a decade or so. If you look a bit more closely, it offers at least some glimmers of hope. It tells us that reform is necessary, yet often fails, but it also highlights that there are some examples, like the New Deal, where successful redistribution was archived (for some time at least). And in contrast to past societies, we now have at least some understanding why societal crises happen and what might be done about them. But what I found perplexing about this book is how timid it is in its recommendations about what might be done. It hammers home again and again and again the point that wealth accumulation of the rich is the main process that drives instability in society. Yet, it does not call for a complete restructuring of society or taxing billionaires out of existence. It merely suggests we should work towards something like the New Deal again and hope for the best. But if we take the book seriously, shouldn’t it argue for a whole new way to structure society? Come up with ways that never again allow the wealth pump to run its course?

I would say yes, this new structure of society is exactly what we should be looking for. But why are you even listening to me? I am just a scientist who did not manage to get a permanent position at university and now I write a blog for a living. According to this book, this makes me a counter-elite who just tries to push their narrative to get power in society again. So, maybe take it with a grain of salt. But I suspect many are feeling like something fundamental needs to change, even if we can’t quite put our finger on what.

Subscribe now

Appendix

I wrote those sections, but had to cut them, as they were in the way of the flow of the post, but they were just too interesting to not include them here.

A brief history of quantitative history/Seshat/Structural Demographic Theory

The idea to quantify history is actually quite old. The trouble is that doing it is very hard, which left most previous attempts going nowhere. Turchin traces back the roots of the theory to Jack Goldstone. While Goldstone had some initial success in understanding history using large datasets, he did not get much recognition and struggled with a lack of structured data, which could be used to make the big historical comparisons more accurate. This is the place where Peter Turchin comes into the story. He got to know Jack Goldstone and the two started collaborating. This collaboration was an inspiration for the creation of the Seshat database (which I have discussed on this blog on several occasions). This database contains a vast treasure trove of quantified historical data. All in the same formats, so it can actually be compared and analyzed statistically. This allowed the ideas of Goldstone to really shine, because now the data was available to make more solid analysis and predictions. To create such datasets is actually quite involved. It took them 7 years from starting to collect data to the first published paper. The Seshat team consists of people with a wide range of expertise from anthropologists, historians, archeologists and data scientists. The data they find is coded by research assistants, this codification is then checked by other research assistants and more senior researchers, as well as by experts in the relevant fields when the coding is difficult/ambiguous. This means every data point in Seshat had several people look at it and verify it and the more complicated ones always had an expert involved. In my opinion, this is the best historical data source out there and if you reject Seshat, you also kinda have to reject all other historical empirical claims, as Seshat is the biggest one. If you don’t trust this one, it would be weird to trust smaller ones.

Plutocracy versus democracy versus autocracy versus militocracy versus theocracy

You can structure your society in a lot of ways, but one of the most important factors is the question of how power is structured. A plutocracy is a system where the wealthy elite hold political power, with governance effectively controlled by the richest individuals or corporations. Democracy involves rule by the people, typically through elected representatives who are accountable to citizens via regular elections and civil liberties. An autocracy concentrates power in a single ruler or small group with no meaningful checks on their authority, often suppressing opposition and dissent. Militocracy places military leaders in control of government, where armed forces directly govern rather than serving civilian leadership. A theocracy is ruled by religious authorities or leaders who claim divine mandate, with laws and policies based on religious doctrine rather than secular principles. Each system represents fundamentally different approaches to organizing political power and determining who has the authority to make decisions for society. There is often also no clear cut between them and you can have things like an autocracy which still conducts elections, which are at least somewhat competitive, but strongly titled towards the rulers.

Historically, militocracies and theocracies were quite dominant and often intertwined, but nowadays it is mostly plutocracies, democracies and autocracies we end up in. To highlight how even similar countries can end up in different spectrums of this, Turchin compares the fates of Belarus, Russia and Ukraine after the fall of the Soviet Union.

Russia

After the Soviet Union crumbled, Russia implemented privatization on a breakneck speed. This served the plutocratic elite well, because they could amass tons of money. This money in turn allowed the rich to influence politics on a massive scale, quickly turning the country into a plutocracy. However, this got so bad, so quickly that the remaining elites in the military and administration realized that they had to do something, or lose all of their wealth and power. They used their remaining power to place Putin in the role of president. While he and his allies were corrupt. They were less corrupt than the plutocratic elites they ousted. This stabilized the country and life became better again for many Russians. This in turn meant when large protests broke out against an ever more autocratic regime, they did not gather enough momentum. Too many people felt that things were going in the right direction and did not want to rock the boat. This allowed the autocrats to consolidate power even more and turn Russia into a full fledged autocracy.

Belarus

When Belarus became independent, the first election was won by Alexander Lukashenko. He saw how the total privatization in Russia had wrecked the economy, so he opposed the same moves in Belarus and kept the key industries under state control. As he controlled the state and the state controlled all key industries, he could use this to gather favors. He focussed on the military. This served him well when large protests broke out against him in 2020, because he could use his contacts in the military to crack down hard on protesters. He was also strongly supported by Russia. This now leaves Belarus somewhere on the border of an autocracy and militocracy.

Ukraine

In contrast to Belarus and Russia, in Ukraine the rich could maintain power for much longer. The other elites in society were not able to hold up against them, which allowed the start of a massive wealth pump. This was partly due to Ukraine being much more split between east and west. Roughly half of the country wanted to join the EU, while the other half wanted closer ties with Russia. This meant that there was no unified power base against the control of the rich and several revolutions just led to another plutocratic faction gaining power. They were able to siphon away much of the wealth of the country, leaving Ukraine with half of the GDP per capita than Belarus, but a lot of billionaires. This streak of plutocratic rule ended when Viktor Yanukovych took power. He was massively corrupt and tried to steal both from the public and the other oligarchs. This was such a blatant power grab, that it allowed enough forces in Ukraine’s society to unite to remove him from power and elect a new president. As Yanukovych was very pro-Russian, his removal also meant that Russia feared it would lose influence in Ukraine. Therefore, the revolution was also the trigger of Russia getting more involved in Ukraine’s politics, which first led to the civil war in the Donbass and ultimately to the Russian invasion of Ukraine. How the country will go into the future very much depends on how the war ends. If it does not end in a total disaster for Ukraine, there is a good chance that it will lead to a more democratic Ukraine, as the war has reduced inequality by destroying much of the wealth. Also, the Ukrainian people are really fed up with having large political influence by both Russians and plutocrats.

How to cite

Jehn, F. U. (2025, October 1). Inequality all the way down. Existential Crunch. https://doi.org/10.59350/3zhee-h1x35

Endnotes

(1) It’s always Ronald Reagan isn’t it? I feel like every time I read about this guy it’s “Ronald Reagan got elected and X started to get worse”.

(2) This is your regular reminder that the later Ayn Rand relied on the very Social Security and Medicare benefits she spent her career arguing against.

(3) Are people still making this joke?

(4) Or more like yesterday, as the book was published in 2023.

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

Life on Earth depends on a stable climate. Both the biosphere and our global society are calibrated to a narrow band of habitability. The global temperature is governed by a lot of interacting systems, but it mostly boils down to how much energy from the Sun reaches the planet’s surface and how much of incoming energy stays in the Earth’s system. Humans mess with this system by trapping more energy here, by releasing more carbon dioxide than the Earth’s system can handle. However, there are also two other, more hypothetical ways humans might impact the climate, which instead mess with the amount of incoming radiation: 1) intentional stratospheric aerosol injection to combat climate change and 2) the release of massive amounts of soot as the consequence of a nuclear war. All three are linked by their influence on the global climate, and in this post I want to discuss their potential interactions and why we should keep the others in mind when we talk about any of them.

Subscribe now

Three dials in the same system

Climate change

If you don’t manipulate the Earth system, the amount of incoming radiation is pretty much fixed, as the sun has a fairly stable output of energy (at least on human timescales). The main variable that changes the temperature on Earth is how much energy remains on it, which in turn is strongly balanced by the amount of carbon dioxide in the atmosphere. Other gases and processes also matter a lot for the temperature (e.g. water vapor, albedo, or the tilt of the Earth). But here I want to focus mostly on carbon dioxide, because it is the one most easily manipulated by humanity. Carbon dioxide exists in a delicate balance when Earth is left undisturbed. New carbon dioxide is emitted from volcanoes and existing carbon dioxide is removed when rock weathers. The rate of rock weathering in turn is influenced by how warm and moist it is, which is also influenced by carbon dioxide concentrations. This means you have a self-regulating thermostat for the global temperature. However, since industrialization humanity has been messing with this thermostat. We have been adding much, much more carbon dioxide than can be removed quickly by weathering. This means we put the system out of equilibrium, we trap more energy on the planet and thus warm it (1). The timescale we are talking about here is decades to centuries. Thankfully, the Earth system is large enough to compensate for at least some of our messing around.

One great paper that tries to explore the most extreme consequences of our messing around is by Kemp et al. (2022). In it Kemp and co-authors try to shine a light on what research has already been done on the most extreme facets of climate change and how it might interact with other catastrophes. They highlight that our knowledge about the most extreme scenarios is scarce, especially when we want to discuss how a world which is out of balance due to climate change might react if this interacts with other anthropogenic impacts like a nuclear war or stratospheric aerosol injection.

Nuclear winter

Nuclear winter is a hypothesized consequence of a nuclear war, where soot particles block out sunlight and cool the planet. This process is explained in detail in Coupe et al. (2019), who conducted a big nuclear winter modeling study. The idea is that the explosion of nuclear weapons ignites whole cities all at once. These fiercely burning cities get so hot that they create a massive firestorm. A firestorm creates so much heat that air rises very quickly above it, like in a chimney, and new, oxygen-rich air is sucked in from the surroundings, creating an even hotter fire. This can create such a strong effect that the soot from the fire is sucked up into the upper atmosphere. Once there the black soot particles heat up and create warmer air around themselves, which carries them upwards, a process called self-lofting. Once they are high up in the atmosphere, they are distributed globally and block out the sunlight. We know this is possible in principle because big forest fires are able to loft smaller amounts of soot up into the stratosphere regularly. The main tricky question around this is: how burnable are cities actually? Depending on this the likelihood of a firestorm changes and how much soot is produced. As we haven’t been torching down whole cities all at once since the middle of the last century, there is little data to base this on, and the modelling studies disagree widely. If you want to learn more about these debates, see here for a detailed discussion of the state of nuclear winter research and here for a history of nuclear winter research and why we know so little about it.

The climate impact would be quite severe. The global land surface temperature could drop by up to almost 10°C after only two years in the most extreme simulations (2) (for land surfaces) and even for much smaller nuclear wars as discussed in Xia et al. (2022) temperatures in cropland areas could drop 4-6°C Thankfully, it only takes about 10 years to get back to normal. The tricky thing is to survive that long, as it would strongly impact food systems.

Stratospheric aerosol injection

Stratospheric aerosol injection involves deliberately releasing aerosols into the stratosphere to reduce incoming sunlight. Unlike nuclear winter scenarios where soot particles absorb and block light, this approach typically uses sulfate particles that scatter sunlight. The process would require aircraft to continuously disperse these particles at high altitude, since they are naturally removed from the atmosphere over time. This would have to be done on both hemispheres at once.

This geoengineering approach remains deeply controversial. It represents a massive intervention in Earth’s climate system with potentially unpredictable consequences. However, it also offers a potential emergency response if conventional climate mitigation efforts prove insufficient and might help avoid triggering tipping points. So far, there have only been a few smaller experiments, but it seems likely that the resources needed to employ it on a larger scale are within the power of many larger nation states. As it is not clear how much sulfate we would need exactly, a slow ramp up period would be sensible to make sure that we do not overdo it.

One paper that dives deeper into the consequences is by Tang and Kemp (2021). They argue that these consequences can happen directly from the way the radiation management is deployed, but also by interacting with other hazards. They frame this as it being a latent risk, meaning a risk that lies dormant, but can later be triggered. This is what is often called the “termination shock”, a sudden warming after the dispersion of sulfate is stopped, especially if carbon dioxide emissions kept on rising while the sulfate was in the atmosphere. Additionally, the sulfate emissions could have negative consequences on ecology mainly by shifting weather patterns in hard to predict ways.

Stratospheric aerosol injection also requires near-perfect international coordination for decades, or possibly even centuries, to balance out who profits and who loses from their deployment. This would likely be extremely difficult to pull off, due to differing regional impacts and deciding who will pay for the program. Additionally, the infrastructure and planes to implement the program would be an easy target for hostile actors.How these processes potentially interact

We see that each of the three processes has a lot of destructive potential on its own, but here I want to highlight the dangers and consequences we might face if they interact. They all influence the global temperature and weather patterns, but partly push it into different directions.

How these processes potentially interact

We see that each of the three processes has a lot of destructive potential on its own, but here I want to highlight the dangers and consequences we might face if they interact. They all influence the global temperature and weather patterns, but partly push it into different directions.

Stratospheric aerosol injections + nuclear winter

Tang and Kemp explicitly discuss this potential interaction. Both stratospheric aerosol injection and nuclear winter push the climate to a cooler state. This means if a nuclear war happens, while sulfate aerosols are emitted, this could add together and lead to an additional cooling. Also, there does not exist any research that looks at how soot and sulfate might interact in the atmosphere.

The nuclear war, which would have to happen before the nuclear winter, would also lead to additional disruption, especially if high-altitude electromagnetic pulses were used, which would destroy electrical equipment on continental scale. If this would affect the nation(s) which are dispersing the sulfates in the atmosphere, this would lead to an instant stop of dispersion and thus to a termination shock, right after the climate had recovered from the nuclear winter. It seems likely that once the program is disrupted, it would be difficult to restart it again, due to the general chaos and disruption. This means stratospheric aerosol injection would make nuclear war more dangerous, and vice versa. Though if you manage to maintain it, it could help to create a smooth transition in and out of nuclear winter.

This scenario of a global catastrophe like nuclear war disabling the ability to continue with solar radiation management is also mapped out in detail in Baum, Maher, and Haqq-Misra (2013). They even go so far to suggest that such a scenario might be a potential way for human extinction, because such a massive climate shift would be hard to muddle through and even if you make it, you would live in a world which is ravaged by high global warming.

Theoretically, this also implies that you could gradually phase out the sulfate emissions to counter global cooling. Though this would likely be very difficult to time it just right, especially in a world that has just gone through a nuclear war, which would make global cooperation hard to come by. However, the research I could find that looks into the costs of stratospheric aerosol injections (e.g. Smith (2020)) usually comes up with estimates around 10-20 billion US dollars per year and per degree Celsius of warming avoided. As this is relatively cheap in comparison to many other things that states do, it might be possible to pull this off by a single country, avoiding the whole coordination problem after nuclear war.

All this means that stratospheric aerosol injection seems like a very high risk, high reward strategy if you also consider the different hazards we phase.

Climate change + nuclear winter

Climate change influences the state of the Earth system and thus the baseline from which a nuclear winter would happen. This is explored in Jehn (2023) (Disclaimer: This is one of my own papers). The paper reviews how crossing planetary boundaries might affect nuclear winter scenarios. The analysis is mostly pessimistic, though climate change could potentially offer one benefit: if nuclear winter occurs in an already warmed world, the temperature drop might be easier to manage. This is because our climate adaptation measures—both in infrastructure and ecosystems—typically lag behind actual warming, creating a buffer. However, this remains highly speculative since no climate models have examined this combined scenario. It is also plausible that a nuclear winter would hit even harder in a world that has prepared for heat, not cold.

Another paper I want to highlight here is by Egeland (2025). This paper tries to entangle how climate change and nuclear threats intersect. It highlights a number of interactions:

• Many nuclear weapon sites are in areas that will be highly affected by climate change and will have to be moved, which is costly and often politically difficult.

• Melting Arctic sea ice will make it harder for Russian submarines to hide, decreasing deterrence and thus making the nuclear world less stable.

• Climate change increases conflict in general and thus also the danger of nuclear weapon usage.

• Globally, we are seeing increased defense spending. This will likely increase carbon emissions, as the military is already responsible for around 5% of global carbon emissions.

• Defense spending is often prioritized over spending for avoiding climate change.

• Nuclear weapons are a cheaper way to project force than traditional military assets. This means if climate change decreases economic growth, there will be an incentive to focus more on nuclear weapons to save money (3).

• Climate change will likely lead to lots of migration, as people flee uninhabitable places. This will lead to more destabilization.

All of this points to a self-reinforcing spiral: climate change makes the world less stable, a less stable world focuses more on defense, more defense spending means less focus on avoiding climate change. This in turn implies more climate change leads to a higher chance of a nuclear war and thus a nuclear winter.

Climate change + stratospheric aerosol injection

Stratospheric aerosol injection is only a partial fix for climate change. As discussed above and in Tang and Kemp (2021), it does not solve the underlying problem of elevated carbon dioxide emissions, it just mitigates some of the effects. This means other negative effects of elevated carbon emissions, like ocean acidification, will just continue and destroy ecosystems. Besides this and the already discussed problem of the termination shock, there are also ethical problems when using stratospheric aerosol injection:

• Stratospheric aerosol injection can be used as an excuse to delay ending carbon emissions. Much of the current climate change action is driven by the fact that there is no alternative to solving climate change. Stratospheric aerosol injection isn’t one, but it can be perceived as such and therefore potentially decrease climate action.

• Also, stratospheric aerosol injection leads to thorny moral issues. It seems possible that the sulfate in the atmosphere might change precipitation patterns. This would often hit countries which haven’t been large greenhouse gas emitters historically, like India. While India will also be hit by climate change, they would be understandably pissed if someone turned off their rain season by injecting sulfate in the upper atmosphere.

• The nation(s) that control the program might have an incentive to shape the climate in a way that is preferable to them, such as trying to preferentially cool some regions compared to others.

All three at once

I think it is safe to say that a scenario of high emission climate change, nuclear winter and a termination shock, due to interrupted stratospheric aerosol emissions would be catastrophic. As far as I know, there is no piece of research that looks into the question of what these three things together might lead to. But let’s imagine a scenario:

It is 2075. Humanity has been engaged in stratospheric aerosol injection for some decades now. This is mostly done by China, because they have the resources to do so, and lots to lose by more warming. Unfortunately, this has led to a slow down of climate action and emissions have continued. While the sulfate has managed to limit warming to around 1°C, it has also led to an ever weaker rain season in many parts of Asia. Relationships between China and India (as well as many other states that now don’t have a rain season anymore) have gotten more and more tense over this. Finally, India decides that enough is enough and starts shooting down the Chinese planes that bring the sulfate into the stratosphere. This is seen as an act of war by China and they start attacking Indian air defenses. The conflict quickly spirals to a nuclear conflict, as both states see it as a struggle for their future. As predicted by many wargaming scenarios, the nuclear conflict cannot be contained, and all other nuclear weapon states are drawn into the exchange, resulting in an all-out nuclear war, which devastates the Northern Hemisphere. Temperatures quickly drop globally. As much of the world economy is destroyed and the rest is massively disrupted, there is little coordination on how to react to the coming nuclear winter, resulting in even more deaths due to famine. As the skies slowly clear again, the termination shock kicks in and the world is catapulted from freezing temperatures to +3.5°C warming in just 10 years.

I leave it as an exercise to the reader to estimate how much of our global society would remain after this chain of events.

Conclusion

I think all this highlights that we know very little about potential hazard interaction on global scales, but what we know all points in the direction that we should put massive efforts into making sure none of these things come true. Climate action is the most important thing, because it allows us to avoid the others. If climate change does not go much above 2°C (or preferably even less) we don’t need stratospheric aerosol injection, as these temperatures are likely still manageable. Also, if we rein in climate impacts, we won’t have as many global conflicts, thus reducing the likelihood of nuclear war. The safest path is emissions reduction, there are no easy fixes. This would also help avoid potential catastrophic interactions we haven’t even foreseen yet.

We should also keep in mind here the different timescales. Climate change is pretty much baked in for centuries. This means for this whole time period, we have to make sure that we don’t trigger the shorter time frame catastrophes of termination shock or nuclear winter.

Pretty much all of the papers I cited here also emphasize that we need more system-level analysis, not just individual hazards. Our world is very complex, especially if you want to consider natural and anthropogenic systems at once. This means considerable research is needed, but preferably on a more holistic level, trying to assess the whole system, not just facets. This is hard, but if we want to really understand the risks that we are facing, I see no way around it.

Subscribe now

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

How to cite

Jehn, F. U. (2025, September 3). Manipulating the global thermostat. Existential Crunch. https://doi.org/10.59350/existentialcrunch.167348184

Endnotes

(1) If you want to understand this in more detail, Ruddiman (2013) is a good explainer and quite comprehensive.

(2) A recent report by National Academies of Sciences, Engineering, and Medicine (2025) questions if such a stark temperature drop could be reached with present day arsenals, but keeping it in here to illustrate the potentially very large effects.

(3) Though it does not necessarily have to be true that nuclear weapons really are the cheapest option. If all of the other points of Egeland (like flooded military bases or decreased usefulness of submarines) become true, nuclear weapons might get more expensive as well. Similarly, larger reliance on nuclear weapons incentives also spending more on missile defense, which is extremely expensive. Finally, it might only be cheap(ish) if you already have the technology and don't have to start from scratch.

References

• Baum, Seth D., Timothy M. Maher, and Jacob Haqq-Misra. 2013. “Double Catastrophe: Intermittent Stratospheric Geoengineering Induced by Societal Collapse.” Environment Systems & Decisions 33(1):168–80. doi:10.1007/s10669-012-9429-y.

• Coupe, Joshua, Charles G. Bardeen, Alan Robock, and Owen B. Toon. 2019. “Nuclear Winter Responses to Nuclear War Between the United States and Russia in the Whole Atmosphere Community Climate Model Version 4 and the Goddard Institute for Space Studies ModelE.” Journal of Geophysical Research: Atmospheres 124(15):8522–43. doi:10.1029/2019JD030509.

• Egeland, Kjølv. 2025. “Disentangling the Nexus of Nuclear Weapons and Climate Change—A Research Agenda.” International Studies Review 27(1):viaf003. doi:10.1093/isr/viaf003.

• Jehn, Florian Ulrich. 2023. “Anthropocene Under Dark Skies: The Compounding Effects of Nuclear Winter and Overstepped Planetary Boundaries.” Pp. 119–32 in Intersections, Reinforcements, Cascades: Proceedings of the 2023 Stanford Existential Risks Conference. The Stanford Existential Risks Initiative.

• Kemp, Luke, Chi Xu, Joanna Depledge, Kristie L. Ebi, Goodwin Gibbins, Timothy A. Kohler, Johan Rockström, Marten Scheffer, Hans Joachim Schellnhuber, Will Steffen, and Timothy M. Lenton. 2022. “Climate Endgame: Exploring Catastrophic Climate Change Scenarios.” Proceedings of the National Academy of Sciences 119(34):e2108146119. doi:10.1073/pnas.2108146119.

• National Academies of Sciences, Engineering, and Medicine. 2025. Potential Environmental Effects of Nuclear War. Washington, D.C.: National Academies Press.

• Ruddiman, William. 2013. Earth’s Climate: Past and Future. New York: WH Freeman.

• Smith, Wake. 2020. “The Cost of Stratospheric Aerosol Injection through 2100.” Environmental Research Letters 15(11):114004. doi:10.1088/1748-9326/aba7e7.

• Tang, Aaron, and Luke Kemp. 2021. “A Fate Worse Than Warming? Stratospheric Aerosol Injection and Global Catastrophic Risk.” Frontiers in Climate 3.

• Xia, Lili, Alan Robock, Kim Scherrer, Cheryl S. Harrison, Benjamin Leon Bodirsky, Isabelle Weindl, Jonas Jägermeyr, Charles G. Bardeen, Owen B. Toon, and Ryan Heneghan. 2022. “Global Food Insecurity and Famine from Reduced Crop, Marine Fishery and Livestock Production Due to Climate Disruption from Nuclear War Soot Injection.” Nature Food 1–11. doi:10.1038/s43016-022-00573-0.

When we study societal collapse, we typically examine it through the lens of elites. Using the records and remains left by emperors, nobles, and the literate classes who had the most to lose when their world ended. But what if we’ve been asking the wrong question all along? After years of reviewing collapse literature ranging from quantitative databases like Seshat to case studies spanning millennia, certain patterns emerge again and again. When I encountered Luke Kemp’s “Goliath’s Curse: The History and Future of Societal Collapse” his analysis aligned with many of the conclusions this blog has been exploring: that inequality drives fragility, that democratic participation enhances resilience, and that our current global system may be more vulnerable than it appears. But Kemp adds an important perspective that most collapse research overlooks. He challenges elite-centered views of history. Rather than simply asking why societies collapse, he asks whether collapse is necessarily the disaster we assume it to be. Perhaps the real question isn’t why collapse is so catastrophic, but catastrophic for whom?

Subscribe now

How bad is collapse and for whom?

Societal collapse is usually seen as a bad thing. It implies chaos, violence and misery. And this is certainly true for some part of the population, but not for everyone. Historically, most states were quite horrible to live in. For example, both the Roman Republic and the Roman Empire were at war for the majority of their existence, while also keeping a large chunk of its population as slaves. Even for a normal farmer, life often wasn’t great. You had to pay a lot of taxes, while often not getting much in return. So, for the average person in the Roman Empire, your life had a high chance to end in war, poverty, or slavery. But this isn’t the Rome we often remember. When we think of Rome we think of all the monuments they build, the art they produced and their politics. But these aspects are representative only for a small sliver of the inhabitants of the empire. We talk about the cultural accomplishments instead of the war, poverty and slavery, because history isn’t written by a poor peasant in the alps or a slave in the mines in Spain, but by the rich, the landowners, the priests, the politicians. This is the case for pretty much every empire.

We see societal collapse as this uniquely bad thing, because we take the accounts of the richest people of the past as representative for the whole population. But what does collapse look like if we instead focus on the living conditions of the average person? What if we have a people’s history of collapse? This is the question that Kemp’s book is trying to answer.

This does not mean that Kemp thinks that collapse has no downsides. It can surely have horrible consequences. For example, if you have a region that specialized in a certain crop, to trade it with other parts of its empire for everything it needs, this region will have massive problems when long distance trade ends because an empire has collapsed. But still, for many of the poorer people, this collapse might mean the end of slavery and a general betterment of their living conditions. The effects of collapse strongly depend on where you are and also when you are, as the consequences of collapse have changed over time as well.

The state of nature?

When we think about what human life was like before we started to record our history, there is one quite dominant idea: the state of nature. The state of nature was an idea made iconic by the philosopher Thomas Hobbes. It assumes that human life before states were established was nasty, brutish, and short, due to humanity being locked in a war of all against all. We therefore need a social contract between the state and its citizens. The state provides order, while the citizens submit to the state for their own protection from each other. While this view on humans is quite popular (especially on the right side of the political spectrum), it is not actually grounded in any data. Hobbes just extrapolated from his personal experience of living through a civil war and concluded that this has to be true for all of humanity. The great thing is, we now have data to verify these claims and this data paints a very different picture to what Hobbes had in mind.

Before humanity started to live in permanent settlements, egalitarian bands of up to 200 hunter-gatherers roamed the lands. We know they had regular exchanges with other bands, because the remains of these groups have quite a high genetic diversity. But genes weren’t the only thing they exchanged. We can also find evidence that they conducted long distance trade 200,000 years ago, transporting goods like obsidian or shells for around 200 km. They also had communal projects, which brought together thousands of people. The most famous example here is Göbleki Tepe. A massive temple-like structure which was built by cooperating hunter-gatherer tribes.

But not only did humans cooperate and trade over long distances, their life also wasn’t nasty, brutish and short. Instead the evidence we find for violence is generally pretty low, albeit highly variable. Looking at anthropological evidence from hunter-gatherers we find values between 0 and 55 %. However, these studies are unreliable with the violence coming from settlers, farmers, and the effects of alcohol. In the archaeological record, the largest systematic studies have found signs of only around a 1. % lethal violence rate throughout the paleolithic. Almost all the deaths are attributed to one site: Jebel Sahaba, which saw a surge of violence during the environment tumult as the world exited the ice age and entered the stable warm period we now know as the ‘Holocene’. Lethal violencerates today are 0.9 %, or 2.2 % if you include suicide. What is also striking about these findings is that we basically see no warfare in the evidence from hunter-gatherers. If there was violence, it was likely between individuals, not between groups.

Our egalitarian origins

This relatively peaceful existence was likely possible, because humans are quite egalitarian in their origin. Only around 11,000 years ago do we see the first consistent markers of hierarchy, like bigger graves or larger houses. Present day nomadic hunter-gatherers still maintain an egalitarian lifestyle. They make sure that everyone who tries to dominate the hierarchy is brought down again. First with social pressure, but they even go so far to kill members of their tribe, if that member continuously tries to raise themselves about others. This dominance aversion is made also easier by a hunter-gatherer lifestyle. If somebody tries to dominate you, you can simply move somewhere else. We can also see this in our biology. Human males generally lack the attributes that species usually have with steep dominance hierarchies, like males being much larger than females and having natural weapons like big and sharp teeth. Additionally, our eyes are much more white, than gorillas or chimps. This is likely an evolutionary adaptation, as it allows others to gauge what we are focussing on, therefore making our intentions more obvious to others, which builds trust.

But in addition to this egalitarian streak, we are also obsessed with status. This also makes sense from an evolutionary perspective. If your status in your group gets too low, the group might expel you, which would mean certain death in many environments. But also, the higher your status is, the more likely it is that people will give you what you want.

When we think about status there are generally two ways you can get it: dominance or prestige. These have quite different implications. Prestige based status is the more prosocial and helpful version. It helps the whole group if people who can do something particularly well get more status. Also, status based on prestige cannot be really inherited and you therefore do not have accumulation of status over generations. Dominance on the other hand is seizing status by sheer force. It is based on violence, intimidation and controlling resources. If possible humans usually try to avoid individuals who seek status by dominance, while seeking those who build status by prestige. Not all individuals have an equal desire for status: a desire for status through dominance is higher in men and those who rank higher in the dark triad.

Our origins seem to be shaped by people seeking prestige and cooperation in an egalitarian setting. Unfortunately, this all started to change once humans started to gain access to new resources.

The birth of Goliath

The idea of societal collapse is only really applicable to humans once we settled. While hunter-gatherers fluctuate widely in their population numbers, these were driven by other processes than what we usually consider to be a societal collapse. Their number waxed and waned in direct correlation with the general productivity and climate of the region they stayed in. When the climate got cooler and productivity fell, so did human population numbers. While these are factors that also influence settled humans, they can maintain higher population numbers for longer, even in more adverse settings, because they can create a storage for emergencies. Hunter-gatherers had on average a much better diet than agriculturalists. However, they also had more periods of famine and starvation. This means when you settle down and focus on a single ressources, you are worse off on average, but shocks don’t hit you as hard.

Creating a settlement only makes sense when you have some abundant resources nearby. These can be fertile farming lands, but any reliable resource works. Many of the first settlements were founded near naturally abundant places, like rivers with plenty of fish. While this made humans more resilient to the fluctuation of the planet, it also allowed dominance based hierarchy to creep in. Though the surplus in and of itself is not the problem. The tricky thing is if the resources could be easily seen, stolen and stored. If you combine with caged land and monopolised weapons Goliath can rise. Once you live for generations at the same place and you focus all your lifestyle around a few key resources, accumulating wealth becomes possible. And once you have more resources than others you can trade them for power.

Imagine a fishing village. One family finds an especially plentiful place to fish and is able to make it storable (e.g. via smoking). This means they can accumulate more fish than the others. This gives them opportunities like hosting a feast or helping out others in times of need. These opportunities create a social debt which is owed to that family. They can leverage this social power to shape the decisions made in the village. To make this work, they not only have to gather these extra resources, they also have to make sure that nobody else takes them. To defend your resources, you need weapons and people to wield them. If somebody objects to this, there is not much they can do. They cannot easily leave, as their life is centered around the local resource as well and they cannot challenge the leading family, as they control the weapons and the social power in the village.

We started with a family that gathered a bit more fish than the others and ended up with this family now having a small militia, while everyone in the village owes them a favor. This showcases how dominance based status hierarchy easily develops once you have a storable resource, people cannot easily leave and monopolized weapons make it hard to resist. Finally, these resources and dominance structures can be easily inherited, making it easier to build and refine them across generations.

Figure 1: The recipe for dominance based status hierarchies

This does not mean that we always have to end up in this situation, just that it is a strong natural attractor in the space of ways to organize your society. This is what Kemp calls Goliath: a collection of aligned dominance based hierarchies across different areas of life which control labor and energy.

Once a Goliath is established, it aims to accumulate more and more resources. This tends to increase violence. Both in and between societies. It increases in the societies, as the dominance has to be enforced to keep people in line and it increases between societies, as other societies represent a threat, which could be solved best preemptively. This clearly shows in our archeological evidence. Violence increases across the board and we also see mass killings and human sacrifice, which basically did not exist before. This means Goliath has shifted the focus from between individuals to between groups and therefore is the origin of war. We see this empirically in Japan, Europe, Mesoamerica, China, North America, and the Near East.

Collapse in the first settlements

While history shows that human societies tend to gravitate to this dominance hierarchy trap, it also shows that especially at the beginning these hierarchies were quite fragile. It took thousands of years from the first settlements until the majority of humans stopped being hunter-gatherers. This means for quite a long time farmers and hunter-gatherers lived in parallel. If they had the choice people tended to remain hunter-gatherers as this gave them agency over their life and when the times were good, a much more comfortable life. Also, this co-existence of lifestyle opened up a way to avoid dominance based hierarchies. If you lived in a settlement that became too hierarchical for your taste, you simply left.

In contrast to this peaceful rejection of hierarchy, many early settlements also show a violent rejection of hierarchy. However, others also seemed to empty in a slow, peaceful trickle of people leaving. Kemp relies here on a lot of examples, but the pattern is always the same. A lootable resource leads to the founding of a settlement. This settlement is egalitarian at first, but over time we see more and more evidence of hierarchy (e.g. temples, bigger homes for some), but only up to a point. Once the inequality becomes too steep, people rise up in rebellion and kill their oppressors, or ‘vote with their feet’ and abandon the settlement. This usually is triggered by an external shock like a drought. Sometimes this does not lead to an abandonment of the settlement, but instead to a time of prosperous, yet egalitarian growth. However, in many cases after some time hierarchy creeps back in and the cycle starts anew. What is striking is the universality of the pattern. We can find it all around the globe, be it the Americas, Europe, or Asia. What we also see is that more egalitarian societies tended to survive for longer, implying an inherent instability of dominance hierarchies.

Figure 2: The boom and bust cycle of dominance hierarchies in early human societies.

We can see both in the egalitarian and the dominance based society that they developed tools and myths that entrench their current style of being. One of the main things here is religion, which is especially helpful for dominance hierarchies. You can use religion to defer to a higher power and develop stories like a monarch being essential for a good harvest. Once this hierarchy is toppled, people tend to emphasize decentralization in all facets of their life. The more decentralized you are, the more difficult it is to be dominated.

The structure of early states

Over time agricultural societies came to outcompete hunter-gatherers due to their ability to face food crises better, maintain higher population numbers, coordinate on bigger scales and actively focussed on expansion and military expansion. This meant that Goliath could grow bigger, because now wherever they extended to, more lootable resources were present, as people had already shifted to agriculture. But there were additional factors that allowed Goliath to grow bigger. Bronze enabled the creation of better weapons, domesticated animals and crops made looting resources easier, and higher population densities made dominating a territory more lucrative. All of this can clearly be seen in the data from the Seshat database (1). The best predictors for more hierarchy in a region are its agricultural productivity and the adoption of military innovations.

These shifts enabled dominance hierarchies on a larger scale. While up until then the most you could dominate was a settlement or two, now you could dominate whole regions. These new and large extractive violent dominance based hierarchies are what we now call the first states. Especially in Egypt, all these factors came together just right and allowed it to grow into the first major state.

To get back to Hobbes, we see that the opposite of his social contract happened. States did not bring order to barbarians, but instead early states created systems that extracted resources from farmers, which got little in return and could have easily lived without these states. If we compare this to today, the most similar organizations are actually those of organized crime like the Mafia. Once you have established this crime-like structure of a state, slavery is the next logical step, as it allows you the ultimate exploitation of another human (though you obviously can also have slavery without states).

Another factor that allowed these early states is linked to how people react to perceived threats. If you live in a dominance based state, you are aware that there are other states out there, which might attack you. This makes people more open to authoritarianism, as it promises more safety and people tend to become more authoritarian if they are threatened. The early states were all shaped by similar forces. Each of these states had some form of authoritative hierarchy (political power), enforced by force (violent power), supported by rituals and narratives that legitimized their authority (informational power), and extracted resources (economic power) from a defined population (demographic power).

An argument often made at that point is that hierarchy is needed to create great things. But this is not actually true. In many regions we can find big projects like temples or irrigations before we can find signs of hierarchy. It often is the other way around. Egalitarian groups build something worthwhile, which leads others to take interest in it and build a dominance hierarchy to control it.

Another important point to note here, we find in the archeological evidence that the most bloody and deadly phase of these early states was not their collapse, but their founding. Their rise coincided with conquest and war, while their collapse mostly just meant that people were free of their oppressors, at least until the next Goliath rose. The overall process of how the earliest states collapsed was quite similar to the ones in settlements, albeit on a larger scale and with more variance introduced by climate and conquest.

Collapse in early states and empires

The next logical development in human organization were empires. Essentially, empires employ the same extractive mechanism we explored for settlements and the early states, albeit at a much larger scale. You usually have a center of power that vacuums in the resources from all corners of the empire, like Rome in the Roman Empire. They are large scale dominance hierarchies. For the majority of the population living in them, they usually decreased their living standards and led to higher inequality.

While their overall dominance based structure is quite similar to early settlements and states, the way collapsed started to shift here. While early settlements and some of the early states faced rebellions which reverted them to a more egalitarian structure, we see less and less of these occurring the larger and more powerful the states and then empires became. They still tended to crumble regularly, but it was less due to egalitarian rebellions, but more due to external factors.

If we look at the first batch of empires that all arose at a similar time during the early Bronze Age (think Akkadian Empire or the Old Egypt Kingdom), they likely were toppled by a megadrought (called the 4.2ka event), which was a too large shock for their highly unequal societies. The drought disrupted their food production and the hierarchy collapsed, as the lack of food meant more complex states could not be supported (though regional variation on how this played out was high). It was different for the empires of the late Bronze Age (think Baylonians, Hittites, etc.). They had a more complex breakdown which we now usually call the Late Bronze Age collapse (2). This collapse included raiding sea people, a climate shift, problems with food production and other factors. But one that is very important here is that these empires relied on each other for resources like food or tin (3). Once these interconnections were disturbed, most of the empires crumbled and fell, they could not weather the impacts without this additional support.

In both cases, the late and early Bronze Age collapses, a dominance hierarchy was established which was so strong that it was more difficult to topple from within. However, it could be brought to fall by outside pressures and like with our earlier examples of collapse on smaller scales, our archeological evidence points to the fact that for most people this collapse was good actually, as their height and health improved after the collapse. They suddenly were freed from the extractive Goliath, which left them more resources for themselves and to cope with the more difficult climatic conditions. Though we can also see bloodshed and conflict here.

Figure 3: Rise and fall of empires due to external pressures.

Besides this end of the empires due to external circumstances, there is also the pathway via internal pressures. However, these internal pressures weren’t the wish of the people to revert to egalitarian societies, but rather the subversive power of corruption and a system dependent on never-ending growth.

The clearest example of this way to end an empire is the Roman Empire. Rome was built on conquest. It had a strong military which was at war almost constantly. This ongoing conquest brought in lots of resources, which allowed low taxes, while also being able to tolerate high corruption. But these resources flowed predominantly to the elites. This was stable as long as the conquests kept going and everybody was better off than before. However, over time new conquests became more difficult. Many of the regions that Rome bordered to later on were poor (e.g. the Germanic tribes) or so powerful they could not be conquered (e.g. Sasanian Empire). This meant the constant inflow of resources by conquest stopped. While the elites could still acquire more and more resources by corruption and dominance, the general population got poorer. Also this was over accompanied by overworking the resources of the land. The decrease in resources led to factionalized elites. All this left the empire in such a weakened state that it could not resist outside pressure anymore and collapsed. However, this collapse was quite bad for a vast part of the population, especially the ones that had been part of the empire for hundreds of years, as they had become dependent on a large interconnected market of specialized production. When this vanished, many regions had to completely re-order how and what they produced.

Figure 4: Collapse of conquest dependent empires.

This pattern was not unique to the Roman Empire, but could be found in many of the empires of antiquity. An especially interesting case here is China. It suffered from similar problems as Rome, but it lasted for longer. This longer time frame also gives some more interesting data to look at when it comes to how extractive the Chinese empire was. We can find a clear correlation between how extractive the state was and how long the individual emperors lasted. The less extractive their policies were, the longer they remained on the throne. However, this also leads to the problem that a less extractive empire is often a weaker empire, as it cannot maintain larger armies. This in turn makes it more vulnerable to conquest. This is the reason why so many of the states throughout human history have relied on dominance hierarchies. Dominance hierarchies make it easier to defend yourself, as you can more easily mobilize manpower and violence. While an egalitarian society makes life better for those living in it, it also makes them potentially more vulnerable against the dominance hierarchies around them.

What we can also see when we look at collapsing empires is that different parts of them collapsed on different timeframes and especially their general culture never really collapsed, but just transitioned to a new system. Political collapse was often rapid, occurring within years or decades. Economic breakdowns typically followed within a century. However, shifts in information systems and population dynamics unfolded over several centuries. Some aspects remained unaffected entirely.

We can also see that there is a clear underlying pattern. Empires collapsed when their internal inequality became so high that it decreased their resilience. Once their resilience was low, they were toppled by the next catastrophe that came along and it did not really matter if it was climate, war or something else.

Collapse, colonialism and capitalism

As we have seen so far, the reach of Goliath increases in step with the capabilities of humans. When we were early farmers Goliath could be as big as a settlement, it then grew to early states, then to continent spanning empires. The next step were the colonial empires which spanned the whole globe. The colonial conquest of empires like the Spanish or British relied strongly on their improved technology, but also on weakening other empires via disease and fueling internal disagreements. This global conquest was arguably one of the most brutal things humanity has ever done. It is estimated that in the Americas, Australia and Hawaii up to 90 % of the population died and this population stayed low for centuries because they were so violently oppressed. The disruption was so strong that it even touched upon regions that were never colonized directly, but that were still wrecked by imported diseases and political instability. Interestingly, this birth of the global Goliath is seldom counted as collapse, because from a European perspective it led to growth and urbanization, even though the cost was the extinction of whole cultures. Maybe we should be much more cautious about the beginning of empires and not their end. If we look at history, the rise is pretty much always much more brutal than the end.

At the same time we can see the beginnings of capitalism. Capitalism and colonialism were a good match for each other. Capitalism relied on growth and extracting resources, while colonialism allowed access to a large amount of resources on the cheap. It also ushered in a new form of inequality that had not existed before. Early capitalism was advantageous for very few people. Even in the heartlands of empires the majority suffered under horrible living conditions. For example, even though England was clearly the single country which profited most from colonialism and early capitalism, life expectancy fell by almost 20 years and the height of adult men by almost 10 cm.

Thankfully, the majority of horrors of colonialism only lasted for around 100-200 years. Since then the world has largely (at least formally) decolonized and living conditions have improved again worldwide. However, this time brought with it a lasting change. Arguably, since then no state has collapsed in a lasting fashion. Sure, countries have gained and lost territories and from time to time some nations become failed states, but none collapsed like the states and empires in earlier history. We have clear transitions of powers, albeit sometimes violent ones. Even failed states still have some access to global markets and they usually stay “failed” only for a few years, before a new government forms.

So, is our modern system the end of collapse? Have we slain Goliath? Not really. We still have empires (e.g. Russia, China or the United States), we just have stopped calling them that way. Many of our modern institutions are still dominance based hierarchies, created and maintained to extract resources. What actually happened is that the product of colonialism and the conflicts of the 20th century created a global Goliath, which contains every nation on this planet. This Goliath is self-stabilizing. If states struggle, other states stabilize it again, as this is often an opportunity to make money. We have seen in many failed and struggling states that a big wave of privatization occurred, after which they were part of the global economy again. But there are other reasons as well that make the collapse of dominance based extractive systems much less likely now. Modern weapons make it much more difficult to resist, but what is likely the strongest reason: We take dominance based hierarchies for granted, because humanity has lived under their control for thousands of years now. This does not mean collapse is impossible now. It just means it has to be a shock that hits the whole global system at once.

A general theory of collapse

Before we dive into what these global shocks might be, let’s recap a bit the main patterns we can see from all these historical ideas and examples. Much of this is related to inequality. More equal and inclusive societies tend to weather shocks much better. Also, more unequal societies tend to suffer more from a variety of problems which make them less well equipped to react to crises. Wealth gives you power. The more wealth you have, the more you can shape society to be more like your preferences and the personal preferences of a single person seldom overlaps with which would be best for a society as whole. Wealth inequality is self-reinforcing. In our modern societies the wealth you gain from having capital is much bigger than the wealth you gain from labor. Therefore, the rich tend to get richer, which in turn increases their undue influence on society. This further gets entrenched via debt. If I owe someone money, it gets much more difficult to resist them, as they have an additional power over me. But also via rich people buying things like media companies or lobbying teams to influence politics. Power begets power.

Besides these direct negative consequences, inequality also leads to a variety of follow-up problems. Inequality leads to unequal power which leads to corruption. The greater the discrepancy in power, the more opportunities arise to exploit this power. Corruption decreases trust in the state and in other people, removing the glue that holds society together. Inequality also makes oligarchy easier. If you have more wealth and power, it gets easier to carve out a part of society that you tightly control. As wealth inequality grows, more elites compete for high-status positions. Aristocrats, lawyers, nobles, and generals—those with resources—either organize rebellions or take over existing ones. This rise in elite competition is central to structural-demographic theory, a model first developed by Jack Goldstone and refined by Peter Turchin (4). It just seems that inequality is inherently corrosive to every society.

In addition, to these factors which are very directly related to inequality, there are also two other factors that come into play:

Imperial Overstretch: Empires like Rome, the Qin, and the Mongols believed their rule was divinely ordained and should be universal. The British and other colonizers saw themselves as civilizing savages, while Islamic rulers expanded the umma. Conquest offered legitimacy and practical benefits: neutralizing threats, securing resources, and rewarding elites with land and plunder. However, expanding borders often created new enemies and increased costs, leading to “imperial overstretch.”

Declining Energy Ratio: Societies depend on energy, and the ratio of energy expended to energy gained, called “Energy Return on Investment” (EROI), is crucial. Many civilizations, like the Western Roman Empire, faced declining energy ratios due to dwindling resources and environmental degradation. However, a poor energy ratio alone rarely causes collapse—it contributes, but other factors are typically involved.

These problems are interconnected. Resource extraction, whether from people or nature, faces diminishing returns. The severity depends on a society’s resilience and its ability to adapt. Larger empires with advanced technology and control systems can expand further and are less prone to rapid collapse. However, extractive institutions and inequality breed oligarchy, corruption, and elite competition. As these factors interact, they worsen each other—corruption fuels inequality, and oligarchy leads to poor environmental management. The state becomes fragmented, with elites draining resources, gradually hollowing it out.

Taking all this together, Kemp formulates a general four step process of collapse (Figure 5). Extractive institutions lead to inequality. This in turn leads to weakening processes like corruption. Once the society is weakened enough, it essentially just waits for a trigger to start the process of collapse.

Figure 5: The four steps to collapse.

What makes societies resilient against collapse

There are some universal principles, such as knowledge storage and creation, which large states and empires excel at. For instance, historical empires often stored and passed on technological advances, contributing to long-term resilience. Similarly, strong religious or ideological unity, and the ability to reform and decentralize (as seen in the Byzantine Empire) also help reduce intra-elite competition, which makes societies more resilient.

However, there’s a strong argument that more inclusive and democratic systems are an even more important factor during crises. Studies by Peter Peregrine (5) show that societies with greater political inclusivity and collaboration tend to be more adaptable to catastrophes like climate change or disease outbreaks. Modern democracies are often more resilient in part because they allow diverse viewpoints to shape decisions, leading to more comprehensive solutions during disasters (6). Citizens’ assemblies, which have proven successful in addressing polarized issues, exemplify this. Also, there are many other examples where the collective wisdom of crowds works pretty well: forecasting, foresight, collective intelligence. Distributed information processing seems to result in better decisions. Also, this decreases the influence of people high in the dark triad and the corrupting influence of power.

Yet, many modern democracies are still indirect in how they channel citizen input, limiting their potential. A more direct application of participatory decision-making could improve resilience even further.

But resilience is a double-edged sword: a system can be resilient even when it’s a bad one. This makes democracy crucial because it not only fosters resilience but also has the best chance of creating a good system that’s worth sustaining. Interestingly, societal collapse often was a path to enable more inclusive government. In many historical examples, collapse led to less dominance hierarchy and more egalitarian societies afterwards. So, in some way societal collapse could be seen as a cultural adaptation to too high inequality and too little inclusivity. Also, as we discussed earlier, debt is an important mechanism to enforce dominance hierarchies, but if your society collapses, this also usually erases all debt and provides the following societies a clean slate on which they can start again.

Modern day collapse

Historically, it looks like societal collapse is less of a problem than you might think. When we look at history, the most brutal and horrible events almost exclusively happened when a new Goliath rises (e.g. the rise of the Mongol Empire or the conquest of the Americas). When groups of humans try to rise to the stop in their own society they often use violence. When they succeed and start conquering their neighbors with their newfound power, it usually gets even more brutal. However, on the other side of the arc of history, in societal collapse we seldom see brutality at that scale. Instead, we see societies that erase their debts, level hierarchies and decentralize, so they can live more egalitarian again.

Does this mean that even today a societal collapse would possibly be an event that makes things better instead of worse? Unfortunately, the answer is probably no. On the contrary, it seems like we have built ourselves a trap. While the global Goliath continues to lead to extractive dominance hierarchies, it has also led to a rise of new technology. Technology that would make any potential conflict, might it be caused by collapse or otherwise, much more deadly. This includes things like nuclear weapons, but also gain of function research, AI or industry loss via an electromagnetic pulse.

Besides these tools and processes which make the breakdown of international order a collapse would likely imply more dangerous, we also have made the planet itself much more vulnerable and inhospitable due to climate change and overstepped planetary boundaries.

The final problem is that we created a global system which is highly interdependent. There is no easy going back to lower levels of complexity. Even ignoring the problem that everything in our system is aimed at evermore growth, economic complexity is not something you can easily dial back, if it gets too difficult to maintain. Instead, we have a network of networks. If one part of these networks fail, it seems likely that it will bring the whole system crashing down. This system is pretty resilient against smaller shocks, but as the global Goliath works towards more inequality and more inequality makes us less resilient, it looks like we are steering towards a world that will get less and less resilient until we get unlucky and a hazard happens that is larger than our diminished resilience. Modern supply chains optimize for efficiency over resilience. Just-in-time delivery means factories often have less than a month’s supply of critical components. For example, during the 2011 Thailand floods, these factors caused factory closures across multiple continents, as stocks of things like hard drives quickly depleted.

Besides the obvious reason that this collapse would be bad, we should also avoid it because chances are that over time it could lead to the rise of a new global Goliath. As we have seen this rise was probably the most violent thing ever done in human history. We should make sure to not repeat this.

What to do?

We live in paradoxical times. While the last section might sound all doomy, it is only half of the story of our times. Yes, we have built ourselves a trap and when it would snap, we would all have a pretty bad time. However, modern humanity is also better equipped to handle danger than ever before. We have the science to understand what is going on, there are ideas out there on how we could solve it. We “just” have to do it. This might even be easier than we generally think, because if you boil it down, there are very few key players that have to change. These are exactly the ones you would guess. Two nations control almost all nuclear weapons, the vast majority of climate damage is done by very few companies, only very few companies have the capability to even have the chance to create general AI or mass surveillance which is also implemented by a handful of key companies. These agents of doom are usually dominance hierarchies themselves, which push them towards making decisions that only favor their growth and continuation and nothing else.

When we summarize the arguments made in Kemp’s book, one thing that becomes pretty clear is that humans like to live in egalitarian and free societies. Paradoxically, societal collapse in the past has been a way to get back to more egalitarian and free societies. But this came at a cost. We lost technological progress, many people lost their lives. So, if this is the case: why don’t we cut out the middleman? If more egalitarian and free societies are the upside that collapse brought us, we could also get there without the downsides.

When wealth inequality and centralized power is the root of all evil, we can redistribute wealth and we can break up power monopolies. Even today the idea of universal basic income and other ideas along those lines get a bit more popular every day. If democracies lead to more resilient societies, we can strengthen them today. Citizens’ assemblies have a good track record of doing exactly that and they get used more and more often. But we need this democratic push not only in governments. We need it on all levels of society. Because if we don’t democratize the few institutions and governments that have brought us into this situation, it will be difficult to get out.

Or to say it differently, let’s build a future where we have an egalitarian and equal society that decides in consent what brings the most health, happiness and general welfare for everyone.

Let’s bring an end to dominance hierarchies!

Subscribe now

How to cite

Jehn, F. U. (2025, August 6). The people's history of collapse. Existential Crunch. https://doi.org/10.59350/3qg73-sxp76

Endnotes

(1) A vast historical dataset, which we have already mentioned a bunch of times on this blog: 1, 2, 3, 4

(2) We discussed this collapse more in an other post.

(3) We discussed the consequences of such trade collapse in an other post.

(4) I explain this theory here.

(5) We discussed those studies here.

(6) For more explanations on why democracies increase resilience, see here.

The following text presents a condensed summary of "The state of global catastrophic risk research: a bibliometric review" by Jehn et al. (2025), published in Earth System Dynamics. The paper represents the first systematic bibliometric analysis of global catastrophic risk (GCR) and existential risk (ER) literature, examining a large number of documents to map the field's development, identify research clusters, and assess current challenges. The full paper includes extensive supplementary materials, detailed methodological descriptions, and over 280 references and can be found here (open access): https://esd.copernicus.org/articles/16/1053/2025/

This is a bit of a different post to what I usually cover here, as it is a summary of a single paper I authored. I am making an exception here, as I think that this paper will be interesting for many of my readers, as it covers pretty much every facet of research I have discussed on this blog so far.

Subscribe now

Visual Abstract

Introduction

The field of global catastrophic risk (GCR) and existential risk (ER) research has grown considerably over the past two decades, evolving from a nascent area of study into an established research domain. This first systematic bibliometric analysis of the GCR/ER literature examines all large number documents to understand how the field has developed, what topics it covers, and what challenges it faces.

GCR refers to risks that could cause the death of a significant fraction of humanity or major loss of global well-being, while ER encompasses risks of human extinction or similarly severe permanent curtailments of humanity's potential. These concepts emerged distinctly in the 2000s, with key foundational texts introducing and refining the terminology. The field's intellectual roots trace back to realizations about nuclear weapons' dangers and climate change threats, which brought awareness of humanity-scale risks into mainstream consciousness.

Growth and Development of the Field

The analysis reveals remarkable growth in GCR/ER research, expanding from approximately 10 documents annually in 2010 to over 150 by 2023. This growth trajectory shows the field's increasing academic legitimacy and societal relevance, though a notable dip occurred in 2020, likely due to researchers pivoting to pandemic-related work during COVID-19.

GCR/ER documents per year by cluster.

Through bibliographic coupling analysis, ten distinct research clusters emerged, each representing different aspects of GCR/ER research: Foundations, Artificial Intelligence, Climate Change, Governance, Pandemics, Transhumanism, Global Resilience and Food Security, Risk Management and Mitigation, Reasoning and Risk, and Emerging Biotechnologies. These clusters demonstrate the field's interdisciplinary nature, touching on disciplines ranging from philosophy and ethics to climate science, agriculture, and technology studies.

Research Clusters and Their Interconnections

The Foundations cluster contains seminal works defining GCR and ER concepts, establishing methodological approaches, and making philosophical arguments for why preventing these risks should be a global priority. This cluster serves as the intellectual backbone of the field, with other clusters frequently citing its key texts.

The Artificial Intelligence cluster, representing about a quarter of annual publications, examines risks from advanced AI systems. Research here spans from technical analyses of AI alignment problems to philosophical discussions about digital consciousness and the ethics of artificial entities. The dominance of AI-related research reflects both the field's origins and contemporary concerns about rapid technological advancement.

Chord diagram of how the clusters relate to each other. The thicker the line is between two clusters, the more their references overlap.

Climate change research within the GCR/ER framework explores catastrophic warming scenarios, tipping points, and governance challenges. Unlike traditional climate research, this cluster often focuses on worst-case outcomes and their systemic impacts on human civilization. The Pandemics grew rapidly during COVID-19, examining not just biological risks but also governance failures and societal vulnerabilities exposed by global health crises.

The Global Resilience and Food Security cluster addresses practical interventions for maintaining food systems during catastrophes. This research area emerged around 2015 and focuses on "resilient foods" that could sustain populations following nuclear war, volcanic eruptions, or other sunlight-blocking events.

Geographic and Demographic Patterns

The field exhibits significant geographic concentration, with approximately 60% of publications originating from researchers in the United States and United Kingdom. This concentration reflects where key GCR/ER research institutions are located but also highlights a problematic lack of global diversity in perspectives on truly global risks.

Gender representation shows even starker imbalances, with roughly 75% of authors being male. This disparity is particularly strong in certain clusters like Global Resilience and Food Security (85% male authors), while only the Pandemics and Biotechnology clusters approach gender parity. Among the most prolific researchers in the field, gender imbalance becomes even stronger, with none of them being female.

Gender balance in GCR/ER research based on all clusters.

Gaps and Future Directions

Several notable gaps emerge from this analysis. Despite nuclear war often being the first GCR people think of and its clear potential for global catastrophe, it remains underrepresented in GCR/ER literature. This possibly reflects the existence of established nuclear security research communities that predate GCR/ER terminology. Similarly, ecosystem collapse receives surprisingly little attention given its potential scale and likelihood.

The field lacks dedicated publication venues, with research scattered across diverse journals. No single journal serves as a focal point for GCR/ER research, potentially hampering community building and knowledge synthesis. Conference opportunities remain geographically limited, primarily concentrated in the US and UK.

Policy uptake of GCR/ER research also remains limited despite growing academic output. While some progress appears in initiatives like the UN's Pact for the Future and the US Global Catastrophic Risk Management Act, translation from research to policy implementation proceeds slowly.

Implications and Recommendations

This analysis suggests several priorities for advancing GCR/ER research. First, creating dedicated publication venues and geographically distributed conferences would help consolidate the field while improving accessibility. Second, actively addressing gender and geographic imbalances through targeted support and collaboration could enrich perspectives on these global challenges.

The field would also benefit from stronger connections to adjacent research areas like systemic risk, disaster risk reduction, and international relations. These connections could provide methodological tools and policy pathways while avoiding redundant efforts. Greater emphasis on understudied risks like ecosystem collapse and great power conflict could provide more comprehensive risk coverage.

Most critically, improving research-to-policy pipelines requires developing concrete scenarios, standardized assessment frameworks, and regional engagement strategies. The successful integration of climate science into policy frameworks provides a model for how GCR/ER research might achieve similar influence.

While challenges remain in diversity, coordination, and implementation, the field's rapid development and increasing sophistication offer hope for better understanding and mitigating humanity's greatest threats.

Subscribe now

Time to update posts again. I know it is not that long ago since the last update post, but there are just so many interesting things I want to share with you. The first thing observant readers probably have already noticed: Every Substack post now has its own historical image. You can watch all of them by browsing through the archive. I tried to find ones that fit the general vibe of the post and I am quite happy with the result. If you are curious about those pictures, they are the work of Albert Kahn and the photographers he hired. Kahn was a rich banker, who liked the arts and one day thought: “Wouldn’t it be cool if we photograph everything on the planet?”. And as he had the means to do so, the project became a reality. He hired several photographers, who took a total of around 70,000 pictures all around the world. I find them quite fascinating. They often have a kind of silent beauty and rich texture. If you want to learn more about this, here is an article and here is an archive of the majority of the pictures.

Subscribe now

Another important piece of news here is that every entry in this blog can now be cited. You can find all entries on Rogue Scholar. Every post gets a DOI with rich metadata (it even includes the references). More info about this below.

And now the actual updates.

Trade collapse

Food trade generally is a good idea, as it gives you access to more resources than you would have otherwise available. However, this also exposes you to the agricultural production shocks of other countries. I added a paper by Keys et al. (2025) to the post, which explores how trading food can make us both more and less vulnerable to climate shocks, depending on who we are trading with.

Another paper which digs deeper into the danger of trade amplifying shocks throughout the system is by Keys et al. (2025). They wanted to understand how climate hazards are transmitted through global food trade. More specifically, they focussed on how both hot and dry and hot and wet climate shocks reduced agricultural production and how this reduction impacted both the producing country and their trading partners. To model this, they ran a climate model (CESM 2) 100 times for 10 years in each run. Every run has slightly different starting conditions and thus develops differently over time. In total these 100 runs represent the plausible variability of the climate over 10 years. The variability they found thus allowed them to assess which parts of the Earth have a chance of being exposed to climate shocks.

With this knowledge they could assess how much of our global agricultural production might be affected by those climate shocks. To check how these production shocks might then impact trading partners, they used trading data to understand how much food countries are importing and from where. This allowed them to calculate how much of the food supply in a given country might be affected by local as well as climate shocks in other countries (Figure 5). This showed them that there are many countries that can expect that 30 % or more of their crop supply might be disrupted due to climate shocks in the next ten years (with up to 93 % in the case of Serbia).

Figure 5: Exposure of crop supply for different countries to climate shocks. Black lines indicate single model runs. Violin plot overall indicates the distribution of the exposure throughout all runs.

Generally, they found that the best shield against climate shocks is having as many countries you import food from as possible, which should also be distributed throughout different climate zones. If you only rely on your local crops and those of your neighbours, it is likely you will have a bad time due to climate change at some point.

Another paper I wanted to add here is by Jain (2024). It explores how food trade happens on a sub-country level. It is the most detailed analysis of this kind I have read so far and reveals quite interesting patterns.

To dig even deeper into the concentration, we can look at how trade happens on a sub-country level. This was recently explored in quite some detail by Jain (2024). The main idea behind the paper was that we would probably be able to understand trade better, if we could look at it in more detail, meaning admin levels below the country level. Problem is that this data mostly does not exist. This means we have to create it.

To do this, Jain trained machine learning models on national scale-trade relations of food trade and predictive variables like crop area, livestock count, population and transportation networks. The resulting model allowed Jain to predict the out and ingoing flow of trade for lower admin levels for all countries. These numbers were then adjusted until the sums lined up with the actually traded amounts on a country level. Due to the high uncertainties involved, the resulting numbers are likely still somewhat off, but they reveal quite interesting patterns of trade (Figure 2). It turns out production is not only concentrated on a few countries, but even in those countries most of the production comes from a small number of regions. This means even a relatively local shock can quickly have global consequences.

Figure 2: Sub-national global cereal trade flows.

Simulating the end of the world

One way to understand the past better is to try to model it. This can often lead to surprising conclusions and strengthen or discard past hypotheses. In my original post, I highlighted several models. In this update, I added a paper by Kondor et al. (2024), which uses an agent based model to explore the question if holocene population dynamics can be explained by people fleeing aggressive neighbours.

Landscapes of fear

To give an example of agent based modelling, I want to discuss a model by Kondor et al. (2024). They build this to understand how inter-group conflict has influenced population dynamics in prehistory. We know from archeological data that there have been boom-and-bust patterns in non-state societies in Mid-Holocene Europe, but it is still debated what exactly caused them. One theory is that this is due to conflict. Partly by just killing people, but potentially more strongly by people avoiding areas where aggressive neighbours might roam.

To test this Kondor and co-authors created an agent based model which goes beyond direct casualties to capture how the mere threat of violence reshapes where people live. They call this the "landscape of fear," borrowing from ecology where prey animals avoid areas with predators even without actual predation.

Their models include two key mechanisms. When conflicts break out, people abandon vulnerable settlements and crowd into defensible locations like hilltops. The presence of aggressive groups also prevents normal expansion into new territories, keeping populations concentrated in "safe" zones.

Testing against radiocarbon data from Mid-Holocene Europe, they found that a scenario with zero direct conflict casualties could still reproduce the observed boom-and-bust population patterns. The models show that these indirect effects alone can cause large-scale regional population declines lasting centuries, cyclical patterns matching archaeological evidence, and population concentration leading to resource stress and elevated mortality.

This research reveals how conflict's psychological and social impacts can be far more devastating than direct violence. Fear itself becomes a powerful demographic force, reshaping entire landscapes and potentially explaining major population patterns in human prehistory.

How to write a living literature review

I extended the post by a section that explains how you can more easily find specific papers you are looking for.

Finding specific papers

Sometimes you already have a hunch of what kind of papers you are looking for, like when there is a gap in argument that would profit from a fitting citation. For this case, there are also a bunch of very helpful tools out there:

• Elicit: This is a search engine optimized for finding the papers you need and it can also automate information extraction. Just ask whatever interests you and it will try to find papers which answer your questions. It is kind of optimized for medical papers, but in my experience it can also be quite helpful for many other fields.

• Connected Papers: The main usage of this tool is to find the web of papers that surrounds a specific paper. You put in the name of a paper that you find interesting and you’ll get a network of papers relevant to it. These papers are selected based on similarity, which is calculated by checking if papers are overlapping in their citations and references.

• Research Rabbit: This is a bit of a mix between the other two tools. Allowing you to search in the network of papers you have specified.

I also added a new section that explains how I use Rogue Scholar.

Making everything citable

As this blog is essentially just another form of doing research, I wanted to make sure that my blog is both findable via academic channels and also that the people I cite can become aware of what I am doing here. This seemed quite difficult to accomplish, until I came across Rogue Scholar, which aims to be a home for scientific blogs. It is a platform which automatically creates DOIs for my blog posts, as well as extracting the references I have, so other people can see that I cited them.

To also set this up for your living literature review you have to go through a few steps:

• Create an account at Rogue Scholar (optional).

• Fill out the blog questionnaire. This gives Rogue Scholar the information it needs to figure out if your work is actually a scientific blog.

• Wait a day or so.

• Once Rogue Scholar has decided that your work is a scientific blog, they will go through all your existing posts to archive them and provide them with DOIs. The end result will look something like this.

Every time you add a new post on your blog, this will also be added to the archive. The DOI forwards you to the original blogpost, unless this is not available anymore, in which case it falls back to an archived version in the Internet Archive.

When I have updated posts, I have to inform Rogue Scholar manually, as Substack does not automatically provide this information (Wordpress for example does). Rogue Scholar will then update the version on their end as well.

You can also add additional meta data like your ORCID or funding information, if you want to.

Overall, it is a pretty cool solution! And now every post of mine also has a “How to cite” section.

Counterfactual catastrophes

In this post I explored how we can use counterfactual reasoning to better understand collapse and catastrophes. I especially highlighted the idea of storylines. As this term tends to be a bit vague and people understand it differently, I added a review paper that analyzed the main ways storylines are used in academia.

Using storylines can sometimes be a bit confusing, as people tend to understand quite different things under the term. Thankfully, a recent review by Baulenas et al. (2023) tries to clear up this confusion by highlighting the main ways people tend to use the term storylines. They identified three main ways:

• Scenario-based storylines: These are qualitative descriptions of how the world might evolve. A typical example would be the IPCC’s Shared Socioeconomic Pathways. They describe the general idea of a potential future, which then can be integrated and modeled with climate models.

• Physical climate storylines: This is the kind of storyline which is meant by Shepherd et al. as explained above.

• Discourse-analytical storylines: This kind of analysis aims at capturing what is discussed in climate policy to understand the main narratives and metaphors.

As you see, these approaches are quite different. I think for societal collapse research, all of them could be helpful. I would love to read it if someone did the work to capture what the main narratives and metaphors around societal collapse are.

Baulenas et al. also recommend that we could combine these three approaches to create more useful storylines. For example, doing a physical climate storyline and then doing a workshop about it to try to find out what kind of narratives this sparks in the participants.

Democratic Resilience

While many of the things I highlight on this blog are more on the negative side of things, there is also good news to be found. One of those good news is a paper by Nord et al. (2025) which shows that democracies are apparently becoming more resilient against the call of autocracy.

But there is also some evidence that democracies seem to be a strong attractor to get back to, once you have been a stable democracy for a while. This is explored in Nord et al. (2025). They wanted to understand how democracies fare in the present “wave of autocratization”. To do so, they looked into history, as this is not the first time democracies face headwinds. They used a dataset of democracies from 1900 to 2023, which tracked how democratic these countries were over this period.

This provided them with a time series for every country. In this timeseries they looked for breaks when countries suddenly are becoming less democratic or less autocratic. Using this approach they found a process which they simply call “U-Turns”. This means that a given country turned toward autocracy for a while, but then quickly reversed course and became more democratic again.

They identify 102 of these U-Turn episodes. Which is significant, because this represents 52 % of turns towards autocracy. This means roughly half of all democratic backsliding gets quickly resolved again (meaning in less than five years). And it gets even better, because if we only look at the last 30 years 73 % of all turns towards autocracy got quickly reversed. So, even in the present day we apparently have a good chance to save our democracies. Additionally, in recent decades these U-Turns mostly start before a country becomes an autocracy, hinting that democracies got more resilient against the call of autocracy. Though, it remains a bit unclear why exactly.

Share your favorite

As a little thank you for reading my stuff, here’s a song I recently enjoyed (hope you like folk punk).

If you made it this far in the post, I assume that you find my writing interesting. I am still working on growing my audience, so if you want to do me a favor, please send your favorite article to someone who might find it interesting.

Go ahead, you can do it right now. Nothing more to read in this post anyway.

Much appreciated!

Subscribe now

Until next time

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me an email to existential_crunch at posteo.de and we can schedule something.

How to cite

Jehn, F. U. (2025, July 16). July 2025 Updates. Existential Crunch. https://doi.org/10.59350/rnf1r-jh869

References

• Baulenas, E., Versteeg, G., Terrado, M., Mindlin, J., & Bojovic, D. (2023). Assembling the climate story: Use of storyline approaches in climate-related science. Global Challenges, 7(7), 2200183. https://doi.org/10.1002/gch2.202200183

• Jain, S. (2024). Mapping Global Cereal Flow at Subnational Scales Unveils Key Insights for Food Systems Resilience. Research Square. https://doi.org/10.21203/rs.3.rs-5204730/v1

• Keys, P. W., Barnes, E. A., Diffenbaugh, N. S., Hertel, T. W., Baldos, U. L. C., & Hedlund, J. (2025). Exposure to compound climate hazards transmitted via global agricultural trade networks. Environmental Research Letters, 20(4), 044039. https://doi.org/10.1088/1748-9326/adb86a

• Kondor, D., Bennett, J. S., Gronenborn, D., & Turchin, P. (2024). Landscape of fear: Indirect effects of conflict can account for large-scale population declines in non-state societies. Journal of The Royal Society Interface, 21(217), 20240210. https://doi.org/10.1098/rsif.2024.0210

• Nord, M., Angiolillo, F., Lundstedt, M., Wiebrecht, F., & Lindberg, S. I. (2025). When autocratization is reversed: Episodes of U-Turns since 1900. Democratization, 0(0), 1–24. https://doi.org/10.1080/13510347.2024.2448742

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

For this post I aim to essentially create a little repository of catastrophic risk lists, so other people can find those more easily and get an overview of how groups of experts with different backgrounds categorize and evaluate risks. The motivation behind both this post and the articles I will be citing, is the idea that to be safe, you have to be prepared and to be prepared you have to know what lurks behind the horizon. Also, I want to show how different groups of researchers go about this and how they focus on sometimes quite different kinds of events. Finally, I want to highlight those catastrophes that most people seem to agree on and how well these risk assessments are taken up by policy makers.

Subscribe now

Gotta catch ‘em all

The most extensive of these catastrophe lists which I am aware of is by Tähtinen et al. (2024). They tried to list essentially every possible future threat that societies could face. To bring it all together they looked at 1714 documents and interviewed 54 experts, which left them with 153 potential future crises. They emphasize that this is the most extensive list of things that go wrong on a societal level that there is and I tend to share this assessment. They categorize these crises in 6 distinct categories:

• Political Crises (34 identified): These emerge through societal conflicts, power struggles, and dysfunctional dynamics of existing institutions. The researchers divided political crises into three subcategories:

  • Shifts in Governance, Regimes & Power Relations (11 crises): Including phenomena like government overthrow and established surveillance society

  • Socio-political Polarisation (8 crises): Covering crises like rising terrorism and disinformation crisis

  • Interstate Conflicts (15 crises): Encompassing various forms of warfare and international tensions

• Economic Crises (23 identified): These emerge from market failures and dysfunctionalities in economic systems. They’re categorized into three subcategories:

  • Collapsing Economic Institutions (9 crises): Including systemic banking crises and stock market crashes

  • Market Frictions & Disparities (9 crises): Covering issues like hyperinflation and mass unemployment

  • Adversities in the Public Economy (4 crises): Including fiscal crises and widespread corruption

• Social-Cultural-Health Crises (25 identified): These emerge from socio-political relations, demographic dynamics, and exogenous disruptions. The researchers divided them into three subcategories:

  • Demographic Turn (5 crises): Including migrant crises and aging population issues

  • Social Erosion (9 crises): Covering phenomena like extreme social stratification

  • Blights to Public Health (11 crises): Including pandemics and healthcare service crises

• Technological Crises (31 identified): These escalate in socio-technological systems through technological development and dependency. They fall into three subcategories:

  • Socio-Technical Dissonance (10 crises): Including uncontrolled AI and militarization of space

  • Technological Hazards (7 crises): Covering incidents like major nuclear plant accidents

  • Inaccessibility of Technologies (14 crises): Including energy crises and broadband inaccessibility

• Legal Crises (8 identified): These concern failures in legal systems affecting equality, efficiency, or trustworthiness. They’re divided into two subcategories:

  • Inefficient Legal System (5 crises): Including widespread distrust in legal systems

  • Discriminative and Oppressive Legal System (3 crises): Covering issues like loss of individual liberty

• Environmental Crises (32 identified): These consist of physical phenomena emerging through natural systems or human disruptions. They’re categorized into two subcategories:

  • Natural Hazards & Extremes (23 crises): Including extreme weather events and biodiversity loss

  • Human-Made Environmental Emergencies (9 crises): Covering issues like pollution crises and geoengineering failures

Escalating global crises

There are also more directed ways to gather lists of potential crises beyond broad cataloging efforts. The United Nations Office for Disaster Risk Reduction (UNDRR) report “Hazards with Escalation Potential: Governing the Drivers of Global and Existential Catastrophes” (Stauffer et al., 2023) specifically focuses on identifying hazards that could escalate into global catastrophes or existential threats to humanity.

The report defines hazards with escalation potential as those that, when combined with corresponding vulnerabilities and exposures, could trigger cascades leading to global catastrophic or existential outcomes. Through a comprehensive methodology involving literature review, expert surveys examining 302 hazards from the Hazard Information Profiles (HIPs), and expert consultations, they identified a subset of ten hazards with particular potential for catastrophic escalation. These ten primary hazards are then sorted into four categories:

• Geohazards: Volcanic aerosols that could trigger global cooling and food system collapse

• Biological hazards: Deadly pandemics, antimicrobial resistance, and harmful algal blooms

• Technological hazards: Nuclear weapons and nuclear winter, radiation agents, infrastructure disruption, and hazards related to the Internet of Things

• Social hazards: International armed conflicts and environmental degradation from conflict

Beyond these specific hazards, the report highlighted two crucial cross-cutting risk drivers: climate change (as an amplifier of multiple hazards) and artificial intelligence (as a transformative process with potentially far-reaching implications).

The researchers were able to identify shared characteristics that make these escalating hazards especially dangerous. These include exponential growth and self-propagation, global geographical scope, severe cascading impacts across multiple systems, ability to trigger irreversible systemic shifts, capacity to bypass established response mechanisms, erosion of trust and cooperation, high uncertainty, shared ownership between separate, national governments, technological origins, and development-driven emergence.

The report emphasizes that human choices drive these risks. How societies invest in critical infrastructure, political systems, military capacity, and technological development creates opportunities but also introduces risks. Many of the most dangerous escalation scenarios are directly linked to human decision-making and technological development, highlighting that human agency is a critical factor in both creating and potentially mitigating these catastrophic threats.

Similar in scope and approach is the Global Risks Report by the World Economic Forum (WEF) (World Economic Forum, 2025). Like the UNDRR report this surveyed actors from academia, government, the private sector and civil society, but an even larger sample (900 experts). The goal in the WEF report was to identify which risks we face in the short term (2 years) and long term (10 years). Generally, the vibes are pessimistic and the majority of survey respondents think that we are in for a rough ride in the short term and an even rougher ride in the long term. This is also a more negative assessment than in the WEF risk reports of the last few years. The survey respondents trace their more negative assessments back to deepening geopolitical and geoeconomic tensions. Especially, in the short term they rank state-based armed conflicts as the number one risk that will cause trouble in 2025. Though this does not mean that they think that armed conflict will be the most severe risk. In the short term the survey respondents think this will be misinformation and in the long term extreme weather events. In addition to ranking the risks, the report also tries to explore how they interact with each other and what kind of risks tend to have a large influence on others (Figure 1). I found this one especially interesting, because it frames inequality as the key issue of our time, as it influences a lot of other potential risks in a very negative way. This is very much in line with inequality also coming up as one of the main culprits for past crises, which we discussed here.

Figure 1: An interconnected map of the global risk landscape

Such kinds of catastrophes are also highly relevant to defense. Therefore, defense ministries also try to map out the different risks their countries might face in the future. An example of this would be a report by the UK Ministry of Defence (2024). They tried to map out what the strategic trends will be until 2055. A strategic trend is not the same thing as a catastrophe, but still the main drivers of trends they identify look eerily familiar to the kinds of events we have covered in this post so far. More specifically, they identify six main drivers of global change:

• Global power competition

• Demographic pressure

• Climate change and pressure on the environment in general

• Technological advances and connectivity

• Economic transformation and energy transition

• Inequality and pressure on governance

They assess that we are in for a wild ride, because many of these trends push the world in different directions. For example, economic transformation and technology will connect the world at an increasing rate, while global power competition will lead to fragmentation. Or similarly, that more and more people globally feel empowered to speak up and mobilize, while at the same time autocratic governments seem to be on the rise.

The risks and trends are also assessed on their impact and uncertainty. Here AI comes out on top for causing the most uncertainty, while global power competition is assigned the position of the most impactful one for the state of the world. As they assess global power competition as the most impactful one, much of the rest of the report is trying to map this out in more detail. They think that the United States will remain the most powerful nation, but will face increasing competition from China, Russia’s future depends on the outcome of the war in Ukraine, India will play a major role, but struggle with internal unrest and many of the middle power will band together, to have a chance in a world of weakening international agreements.

Global catastrophic risk and existential risk

Another community that is fond of making lists of catastrophes is the one focussed on global catastrophic risk (1). Global catastrophic risk involves threats that could cause widespread human mortality or severely reduce well-being worldwide. The research here revolves around how such risks might occur and how they could be prevented or managed. However, it is still debated which catastrophes should count here. But I want to quickly highlight the most extensive summary studies here and what kinds of catastrophes they include.

Ord (2020)

This book is one of the foundational texts of the field of global catastrophic risk. In it, Toby Ord wrestles with the question of what catastrophes are a global risk to humanity and what the chances are that they might lead to human extinction. The main catastrophes he comes up with are: asteroids, volcanoes, stellar explosions, nuclear weapons, climate change, environmental damage, pandemics, AI and global authoritarianism. Overall, he thinks there is a one in six chance that one of these leading to an existential catastrophe this century, with his best bet being AI.

ÓhÉigeartaigh (2025)

Very similar in its aim to Ord, but more recent (and considerably shorter), is this review by Sean ÓhÉigeartaigh. It is specifically about what experts have said in the past about which catastrophes might lead to human extinction and how likely they might be. The list is similar to Ord, but a bit longer. In addition to those already mentioned by Ord, ÓhÉigeartaigh also considers solar luminosity increase (the Sun getting brighter in its life cycle), physics experiments going wrong (e.g. accidentally creating a black hole inside of Earth) and unknown future technologies (e.g. atomically precise manufacturing) (2). Overall, ÓhÉigeartaigh assesses that getting to human extinction is just pretty difficult. For most of the catastrophes, you always have some kind of holdout. For example, if you have a supervolcano cooling the Earth, there are still places like Australia that would be pretty habitable. However, he thinks what might get us all are more the cascading failures of critical systems like power and food supply chains and the resulting chaos and reduced resilience against further catastrophes. Finally, he highlights that the largest danger comes from ourselves. We still have nuclear weapons and might build artificial general intelligence in the future, with very uncertain consequences for all of us. He ends with a sentence that summarizes this succinctly: “Humans may be hard to wipe out, but that will not stop us giving it our best shot.”

Avin et al. (2018)

This study has a bit of a different approach from Ord. Instead of trying to come up with the main existential catastrophes that are out there, they try to think about how we can classify them best. The way they go about this is by thinking about what critical systems are affected (physical, biogeochemical, cellular, anatomical, whole organism, ecological, sociotechnological), how the catastrophe spreads globally (natural global scale, anthropogenic networks, replicators) and on what scale prevention and mitigation failures happen (individual, interpersonal, institutional, beyond institutional). They then try to sort a list of example scenarios into those categories. More specifically they decide to look at asteroid impact, volcanic super-eruption, natural pandemic, ecosystem collapse, nuclear war, bioengineered pathogen, weaponized artificial intelligence, geoengineering termination shock. This is not meant as an exhaustive list, but more to showcase their classification system. While their classification system is interesting in and of itself, it also highlights that there is one system that is especially vulnerable to many of these global catastrophes: the food system (3).

Sepasspour (2023)

There are many further studies who discuss what might count as a global catastrophic risk, but I have highlighted Ord, ÓhÉigeartaigh and Avin et al., because I think they are the best ones. The final one I want to highlight here is Sepasspour (2023). While the general direction of this article is more focused on the importance of an all-hazards approach to global catastrophic risk (4), it starts with a literature review I want to highlight. In this, Sepasspour looked at 31 books and 7 reports to try to find out which are the most often mentioned global catastrophic risks and comes up with the following list: artificial intelligence, biotechnology, climate change, ecological collapse, near-Earth objects, nuclear weapons, pandemics, supervolcanic eruption. Given my own experience in the field, this seems spot on to me and should probably be considered the list of “main” global catastrophic risks.

What future catastrophes might await us?

Besides these catastrophes, about which most already agree that they are big and bad and we should do something about them, there are also potentially less known catastrophes out there, which could become a big problem in the future. Dal Prá et al. (2024) tried to map out such risks in a horizon scan. To do so, they gathered a team of 32 scientists (disclaimer: I was one of them) and asked them first to submit global issues that they thought were flying under the radar. This resulted in 96 potential future risks, which were brought down to those 15 that were deemed as most important by the majority. Here’s a short description:

• Integration of AI in Nuclear Weapons Systems: The increasing integration of AI into nuclear weapons systems introduces new risks by potentially destabilizing the strategic balance, enabling faster decision-making with less human oversight, and creating vulnerabilities to cyberattacks that could lead to accidental launches.

• State Capacity Deficits: Growing government debt, declining trust in institutions, and reduced political stability are hindering governments’ abilities to provide essential services and effectively respond to catastrophic risks, creating a dangerous decrease in state capacity precisely when we need it most.

• Cascading Failures in Global Food Systems: The complexity of our global food system makes it vulnerable to disruptions that can trigger domino effects across food, economic, and health systems, potentially leaving hundreds of millions more people hungry when crop failures, conflicts, or supply chain disruptions occur.

• Climate Change-Induced Displacement: As climate change pushes nearly one-third of humanity outside the historical human climate niche, mass displacement will interact with other crises like conflicts and food insecurity, creating a mutually reinforcing web of challenges that overwhelms societies’ abilities to respond.

• Unclear Future of the Ocean Carbon Sink: The ocean’s crucial role in regulating Earth’s climate by absorbing CO₂ and excess heat may be weakening, potentially accelerating climate change and disrupting marine food webs that billions of people depend on.

• Declining Epistemic Robustness: Our collective ability to make decisions based on accurate information is being eroded by digital media, ideological biases, and the lowering costs of producing disinformation, making it increasingly difficult to coordinate responses to global catastrophic risks.

• Supercharged Surveillance States: Advanced surveillance technologies combining facial recognition, biometrics, and social media monitoring are enabling unprecedented social control, potentially locking societies into surveillance regimes that reduce citizens’ agency and ability to respond to extraordinary crises.

• AI in Bioengineering Arms Races: AI is removing barriers to the development of biological weapons by enabling more people to manipulate biological data for creating proteins and viral vectors, with risks of accidental releases or deliberate use in conflicts.

• Collapse of the Truth in the Age of AI: Generative AI is accelerating our inability to distinguish fact from fiction, creating a societal environment where truth is increasingly devalued and our ability to respond to real threats is severely limited by confusion and mistrust.

• Instability and Collision of Objects in Earth’s Orbit: The proliferation of satellites raises the risk of triggering the Kessler Syndrome—a cascade of collisions generating space debris faster than it decays—which could render low Earth orbit unusable and threaten modern communications, GPS systems, and critical Earth observation capabilities.

• Collapse of Trade Networks Due to Key Hub Destruction: Our just-in-time global supply chains are vulnerable to disruptions at critical trade hubs that could trigger cascading effects, causing export bans, conflicts over resource access, and severe impacts on vital sectors.

• Political Radicalization Driven by Ecological Destabilization: Climate change and environmental threats can push people toward more authoritarian views, creating self-reinforcing cycles where rising authoritarianism makes environmental problems harder to address, further accelerating societal polarization.

• Large-Scale Heat Stress: At just 2-3°C of warming, extreme heat stress will be exceeded during large parts of the year across much of the tropics where half of humanity lives, causing neurological damage, economic collapse, and potentially starting a global socio-economic tipping cascade.

• Accelerated Development of Autonomous Weapons Systems: The rapid proliferation of autonomous weapons systems without adequate oversight is undermining international norms about when nations are at war, blurring conflict boundaries, and increasing the risk of inadvertent escalation.

• Termination Shock from Solar Radiation Management: If solar geoengineering efforts using stratospheric aerosol injection were disabled by disaster or international conflict after implementation, the rapid warming that would follow could be catastrophic, potentially causing multiple degrees of temperature rise in a short period.

This is a pretty wild mix from a very wide range of fields, but I think it gives an interesting overview of what experts concerned with global risks are thinking about. And just like in the review by ÓhÉigeartaigh one thing becomes very clear: all of these risks are of our own doing. The largest future risk for humanity is humanity.

What kind of catastrophes do policy makers consider?

Now we know what kinds of catastrophes experts of global risk see as most important right now and potentially important in the near future. But this is only the first step. It is also important how these risks are considered in policy. Because if none of these insights make it into law or agreements, they are pretty much toothless. Therefore, let’s look a bit at the uptake of these catastrophes by policy makers.

United Nations

One very interesting study in this context of global catastrophic risk and policy uptake is Boyd & Wilson (2020). They examined the UN Digital Library to assess which existential threats receive attention from global policymakers. They conducted a systematic keyword search for terms related to existential risks (like “human survival” and “global catastrophe”) as well as specific risk categories such as nuclear war, artificial intelligence, and synthetic biology. Their analysis revealed that nuclear war dominated the discussion, accounting for 69% of explicit existential risk mentions, while other potentially catastrophic threats like artificial intelligence, synthetic biology, and supervolcanic eruptions received minimal attention or were completely absent from UN documents.

They concluded that there’s a significant imbalance in how international governance bodies address different existential threats, with some major risks being severely neglected despite their potential severity. They argue that this gap should be addressed through three main approaches: creating dedicated UN bodies for each major existential threat (similar to existing ones for nuclear disarmament and near-Earth objects), establishing an overarching body to coordinate across all existential risks, and potentially considering the rights of future generations through the UN Human Rights Council. Their findings suggest that global cooperation on existential risk mitigation needs significant strengthening, especially for emerging technological threats that could develop rapidly in the coming years.

What kind of catastrophes do national risk assessments consider?

I have already discussed national risk assessments in another post, so I won’t go into much detail here, but to quickly recap it: Democratic countries often conduct comprehensive national risk assessments that include global catastrophic threats. A review of assessments from Switzerland, the United Kingdom, the Netherlands, New Zealand, and Canada shows varying coverage of major risks. Switzerland has the most thorough approach, addressing artificial intelligence, climate change, near-Earth objects, geomagnetic storms, pandemics, and supervolcanic eruptions. The UK similarly covers most risks except near-Earth objects. Other democracies like the Netherlands, New Zealand, and Canada have more limited coverage, though climate change and pandemics appear consistently across all assessments.

Overall, it seems here that the inclusion of global and/or a bit more unusual risks strongly depends on the country. Some of them do very thorough analysis, while others mostly focus on small, local risks like floods and immediate global risks like climate change or pandemics. It is not really clear to me if there’s a pattern here.

If you have studies that might shine a light on why some countries focus more on global risks than others, please let me know.

What has been encouraging here is that good national risk assessments can work as an inspiration for other actors. For example, in a recent report by Graham et al. (2025), a group of researchers used the methodology of the Swiss national risk assessment and transferred it to Australia and showing that even under traditional risk assessment (accounting for both likelihood and magnitude of the consequences), the annual expected impacts of some global catastrophes are far higher than many of the more frequent threats (like floods or fires) and thus Australia should invest more to prevent damages from global catastrophes.

Takeaway

When looking at all these different risk complications, it seems to me that global risk experts are most concerned with nuclear threats, artificial intelligence, pandemics and climate change. Also, it becomes pretty clear that generally all risks are either directly caused by humanity or made worse by humanity. While it is obviously bad that we are our own nightmare, it also means that all of these risks are within our own agency. Things caused by humanity can probably also be solved by humanity. A good place to start seems to be to make sure that the risks we face are more consistently included in things like national risk assessments or global policy making, so that we can better explore what might be done about them. Also, inequality pops up in all kinds of assessments as being pretty bad. So, it seems like a good idea to also take care of that.

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

Subscribe now

Endnotes

(1) If you want to know about this field, I wrote a whole systematic review about it.

(2) The famous grey goo.

(3) The implications of this are discussed in another post of this living literature review.

(4) Which essentially means that we should primarily focus on interventions which have a positive effect on all or most global catastrophic risks, instead of getting bogged down in trying to find the best solution for each global catastrophe.

How to cite

Jehn, F. U. (2025, June 18). What could go wrong?. Existential Crunch. https://doi.org/10.59350/k1dc2-9x428

References

• Avin, S., Wintle, B. C., Weitzdörfer, J., Ó hÉigeartaigh, S. S., Sutherland, W. J., & Rees, M. J. (2018). Classifying global catastrophic risks. Futures, 102, 20–26. https://doi.org/10.1016/j.futures.2018.02.001

• Boyd, M., & Wilson, N. (2020). Existential Risks to Humanity Should Concern International Policymakers and More Could Be Done in Considering Them at the International Governance Level. Risk Analysis, 40(11), 2303–2312. https://doi.org/10.1111/risa.13566

• Dal Prá, G., Chan, C., Burkhanov, T., Arnscheidt, C. W., Cremades, R., Cremer, C. Z., Galaz, V., Gambhir, A., Heikkinen, K., Hinge, M., Hoyer, D., Jehn, F. U., Larcey, P., Kemp, L., Keys, P. W., Kiyaei, E., Lade, S. J., Manheim, D., McKay, D. A., … Sutherland, W. (2024). A Horizon Scan of Global Catastrophic Risks (SSRN Scholarly Paper No. 5005075). Social Science Research Network. https://doi.org/10.2139/ssrn.5005075

• Graham, J., Boyd, M., Sadler, G., & Noetel, M. (2025). Mapping Australia’s Risk Landscape: A Comparative Analysis of Global Catastrophic Risks and Traditional Hazards (SSRN Scholarly Paper No. 5253625). Social Science Research Network. https://doi.org/10.2139/ssrn.5253625

• ÓhÉigeartaigh, S. (2025). Extinction of the human species: What could cause it and how likely is it to occur? Cambridge Prisms: Extinction, 3, e4. https://doi.org/10.1017/ext.2025.4

• Ord, T. (2020). The Precipice: Existential Risk and the Future of Humanity. Hachette Books, 480 pp.

• Sepasspour, R. (2023). All-Hazards Policy for Global Catastrophic Risk (Technical Report Nos. 23–1; p. 37). Global Catastrophic Risk Institute. https://gcrinstitute.org/papers/068_all-hazards.pdf

• Stauffer, M., Kirsch-Wood, J., Stevance, A.-S., Mani, L., Sundaram, L., Dryhurst, S., & Seifert, K. (2023). Hazards with Escalation Potential Governing the Drivers of Global and Existential Catastrophes. United Nations Office for Disaster Risk Reduction.

• Tähtinen, L., Toivonen, S., & Rashidfarokhi, A. (2024). Landscape and domains of possible future threats from a societal point of view. Journal of Contingencies and Crisis Management, 32(1), e12529. https://doi.org/10.1111/1468-5973.12529

• UK Ministry of Defence. (2024). Global Strategic Trends: Out to 2055.

• World Economic Forum. (2025). The Global Risks Report 2025.

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

When we talk about societal collapse, we usually talk about the factors that led to the collapse of a given civilization. However, you could also turn this around and ask what factors allow civilizations to avoid societal collapse and major crises. Knowing more about this would be quite valuable, because generally, you would rather avoid major crises, instead of having to try to solve them once they started.

Subscribe now

Under what conditions are long-term risks tackled before they become a crisis?

Exactly this kind of question is tackled in a review by Shwom and Kopp (2019). They are motivated by climate change, as they see this as the main long-term risk we face right now, but the results they find are generalizable. This one is especially interesting, as they try to tackle this question from an as wide angle as possible and include insights from a lot of different fields. They structure their review along several questions:

What risk perceptions would make action on long-term risk more likely?

They start with a focus on psychology. From their findings they argue that the main reason from a risk perception standpoint is that people have more psychological distance the more long term risks are. This mostly means if something feels more or less directly relevant for your life. The psychological distance increases if an event is farther away in the future and if its outcomes seem more uncertain and ambiguous. While we cannot change how far something is in the future, we can change how uncertain or ambiguous something feels. Shwom and Kopp highlight the value of fiction here, especially if it is concerned with the suffering of future people and invites the people interacting with it to take the perspective of the future harmed people. But also other things like sharing experiences related to the risk in group discussions seem to make it more concrete for people (1). When the risk becomes more concrete, it tends to trigger a reaction of fear and anxiety and these negative emotions are then often the driver to move people to act in precaution of the risk.

When will societies weigh the costs and benefits of long-term risk as action worthy?

In the answer to this question Shwom and Kopp focus on expected utility theory. With this they mean assessing if the cost of acting against the risk now is higher or lower than the future damages of the risk, discounted by how far they are in the future and how likely they are. Calculations like this are done by many institutions in the form of cost benefit analysis. In analysis of many past cost benefit analysis, it has become apparent that most institutions heavily discount the future and thus are biased against action against long-term risk. Interestingly, this seems to be different for individual humans. Shwom and Kopp cite here several studies that gave people the option to assign value to environmental crises across different time horizons and found that around 30-50 % of their respondents don’t do time discounting at all (2). This means if an environmental risk is 5,10 or a 100 years away does not matter to them, they want us to avoid it all. Unfortunately, it seems that these studies don’t offer any good answers on what makes a large chunk of the population not have future discount rates for environmental crises. But a good way forward here might be to assess the discrepancy between what individual people want and what institutions (which consist of these people) want.

What social, cultural and institutional factors would predict that society will act to address long-term risks?

The first two questions showed that people differ in their assessment of long term risk and can be moved to care more about it. However, societal reaction is still lacking, so there must be other factors at play. Based on their review Shwom and Kopp think that the missing piece here is how society processes information. Obviously, what kind of and how you get information is shaped by the way your society is structured. But what the research here seems to be showing is that to get action going to tackle a long-term risk, there is some kind of focussing event needed. These are often disasters that relate to the long-term risk. Such events allow the society to clearly identify the problem and the suffering inflicted provides the motivation to change things. However, for this to be really successful, the crisis needs to be socially amplified. This means you have to have a critical mass of experts and media people interested in the topic before it happens and once the focussing event happens they are able to amplify the topic into the general discourse of the society. This process can be distorted by bad actors which withhold information or produce disinformation. So, you have to keep those in check if you want your social amplification to work. One thing that helps is to emphasize the existing scientific consensus, as there is research showing that people tend to be more open to act on a risk in general, if they know about it and have the impression that the scientists are all pointing in the same direction.

Summing this up

My main takeaways from Shwom and Kopp are that you have to inform people about the long-term risk, work on making it more concrete for them by working with fiction and highlighting the scientific consensus. In parallel you have to work on getting experts and people in the media interested in the topic in general. Then you have to wait for a focussing event and use this to amplify your message around the risk and force society into action.

However, Shwom and Kopp also emphasize in their review that these are very preliminary results and there is surprisingly little research around long-term risk governance in general. In addition, this research is spread throughout many disciplines, which often do not know what the other fields are doing. So, we need a lot more research here and especially empirical one.

Quantitative history to the rescue

This is the part where everyone’s favorite quantitative historical database Seshat comes into play. A recent article by Hoyer et al. (2024) tries to make a first exploration of how averted crises play out using this data. However, this is quite a tricky thing to look at. By definition, averted crises cannot easily leave a big mark in history, because they did not happen. To be able to detect an averted crises they lean on structural demographic theory. We discussed this briefly in the first post in this series, but here’s a quick refresher.

In structural demographic theory the idea is that your society goes through a cycle of increasing pressure, which gets released at some point and then the cycle starts anew. This goes through several steps:

• Your society is growing, there are enough resources for everyone, everything is quite stable.

• Your growth starts to slow down, because you hit some limits (e.g. there is no more territory to conquer). Elites start to amass more and more of the remaining wealth and the general populace starts to get poorer.

• As the elite’s number keeps growing at the cost of everybody else, they too run into limits, as there are only so many prestigious positions to go around. This leads to competition within the elites.

• The state starts to run out of resources, as the high number of competing elites takes an ever greater share of the available wealth for themselves.

• The pressure of a poor and unhappy general populace, infighting elites, and fiscally stressed states results in a disruption which resolves the pressure. This can be anything from major reform to a civil war.

This means that there is a similar process in the build up of stress for all societies, but the important question is how and at what stage this pressure is relieved? Hoyer and co-authors here use the analogy of a Ginkgo leaf. There is one fairly common path to crises, but many ways to exit (3).

Figure 1: How a crisis plays out similar to a Gingko leaf.

The key indicators to understand where in the cycle you are is popular immiseration, elite overproduction, intra-elite competition and fiscal distress. These are all things you can measure (or at least proxies of them). Therefore, to find cases of averted crises you have to measure these indicators and find historical events where they were very high, but the pressure was relieved in a productive way, which did not result in destructive events like civil wars.

This is exactly what Hoyer et al. did, and they found four events where you have a big pressure build up, but no immediate crisis, but instead a release of pressure (4).

The case studies

Conflict of the Orders in the early Roman Republic (494-287 BCE)

The early Roman Republic was quite an achievement, because it managed to move away from monarchy to a republic. While this led to some redistribution of wealth and political power, it still meant that most of the power was held by a small number of aristocrats and the general populace (the plebs) did not have much say. These differences between the aristocrats and the plebs increased over time, as the aristocrats amassed ever more land and small holder farmers were pushed to the margins. Also, the early Roman Republic was almost permanently at war. This led to a massive influx of wealth, which was mainly funneled to the aristocrats. Therefore, Rome had popular immiseration, elite competition around the relatively small amount of public offices, but little fiscal stress. These highly unequal circumstances led the plebs to essentially go on strike and demand both political power and a more equally shared wealth. Such strikes happened several times and allowed the plebs a larger share in the societal wealth, while also giving them institutionalized legislative authority. Once this was in place, the pressure was relieved and the crisis was averted.

Rome was probably successful here, because the elites were willing to do a reform. This willingness was likely due to the constant warfare. The elites were reliant on the plebs for their armies. If no agreement was reached, Rome risked conquest by neighboring states. Also, the conquests of other territories made it easier to reform, as the additional wealth made redistribution of wealth less of a hot topic, as there were enough spoils of war, that the elites were open to share some of that without really losing most of their wealth and privileges

Chartist movement in England (1819-1867)

The United Kingdom grew a lot in the 17th to 19th century and acquired a large number of colonies. The exploitation of those colonies meant a steady influx of wealth to England, which contributed to early industrialization, which in turn accelerated economic growth as well. However, the spoils of this quickly growing empire were distributed very unequal. Most people had next to nothing and many of the jobs in industry were quite hazardous as almost no regulation existed. Also, there was no universal suffrage, so most people did not even have political influence to try to improve their lot. This resulted in growing unrest and societal stress in the 19th century. From the structural demographic theory indicators the popular immiseration shows up most clearly. Real wages had been declining since 1750. There was potentially an elite overproduction, because university admissions rose sharply. Fiscal stress however was not really present, as the empire was flush with wealth from its colonies. As in the example of the Roman Republic, the pressure was relieved via reforms. This happened in several steps, as the first attempts at resolution were not far reaching enough or did not get enough buy-in from power-holders. These reforms increased welfare considerably (5), introduced more laws around labor and created universal suffrage. These laws could only be passed after there had been massive strikes and rebellions, as before that the elites were not open enough for a compromise that would cut into their profits and status. But still it is a remarkable outcome, as in most of such cases labour unrest is simply suppressed.

As with the Roman Republic, England could reform more easily because it imported wealth and resources from its colonies (though this clearly harmed people living in these colonies). England was also able to relieve population pressure through emigration, making domestic reforms easier to implement.

Reform period in the Russian Empire (1855-1881)

Over the course of the 18th century the Russian Empire had expanded a lot. This had led to an influx of wealth to the central parts of the empire at the cost of the periphery. Overall, the economy had been growing quickly, but the population grew even quicker, meaning the real wages were declining. Also, the societal structure in Russia still allowed serfdom, which meant that most farmers were tied to the land they worked on and not that far from being slaves. This was in sharp contrast to the liberal ideas from Enlightenment thinkers, which were making the rounds in Europe at the time. This made the general population very dissatisfied with their situation. The elites on the other hand had profited considerably, as they had been able to carve out a lot of privileges. They also did not have that much intra-elite competition, as they were united against the states that aimed at curtailing these privileges. All of this was combined with empty state coffers, due the many wars that were fought by Russia in the 19th century. When Alexander II came to power, he aimed for wide ranging reforms, to stabilize the situation. He led reforms to free the serfs, modernize education and judiciary. However, these reforms went strongly against the interests of the elites and as there had been assassinations and coups in the past. This led to Alexander II becoming cautious, which in turn made the reforms less ambitious.

These toned down reforms resulted in only partial success. The reforms brought some peace because peasants had some demands met, but wealth inequality remained largely unaddressed. For many freed serfs, “freedom” mainly meant paying the same amounts to the state that they previously paid to aristocrats. This left many grievances unresolved, which temporarily pacified the country for some decades but ultimately led to revolution in 1917. Unlike the fully successful cases, Russia failed to achieve sufficient elite buy-in for sufficient redistribution of wealth and power.

Progressive Era in the USA (1914-1939)

The American Civil War was fought around the issues of slavery, states’ rights and economic and cultural differences. While it resulted in slavery being abolished, it was not able to resolve many of the systemic issues in the United States. There was still systemic racism, the rapid industrialization disrupted the society and the central government tried to reassert its central authority. These factors lingered or grew throughout the 19th century. At the same time the country grew massively, both in territory and economic size. As in England, this often came together with hazardous working conditions for many and few rights to protect you. Also, much of the wealth was funneled to a small group of elites, who grew over time and started to fight over political influence. Fiscally, the state was in a somewhat neutral position, it did not have much debt, but the overall state structure was rather small and inflexible, with little reliable income streams. All of this meant that you had a massive amount of pressure building up.

This was partially relieved by the First World War, which focussed much of the public attention away from social issues. However, there had also been a positive trend even before the war. In the first decade of the 20th century, labour laws were strengthened and unions started to grow considerably. Even some of the major industrialists started to advocate for reform and introduced better working conditions in their factories. These trends were reinforced by the Great Depression. The economic collapse demonstrated that further major systemic changes were necessary to restore peace and prosperity. This crisis ultimately prompted Roosevelt’s New Deal reforms, which raised general welfare and reduced instability across society by redistributing wealth and creating stronger safeguards against economic hardship.

Main factors

The case studies reveal several patterns in successful crisis aversion. First, expanding territory and economic growth help maintain social peace by allowing greater wealth distribution without requiring elites to surrender existing assets. This often coincides with increased national unity when facing external threats.

Second, even after significant social pressure builds, resolution through reform remains possible. The challenge is pushing reforms far enough against elite resistance and managing to forge a broad societal alliance that has enough power to maintain the momentum of the reform projects. Without sufficient reform depth or a too small coalition, as in the Russian Empire case, you only get temporary relief rather than lasting stability.

A consistent theme across these cases is that when a small elite group amasses a large portion of society’s wealth, it leads to hardship for the majority. Peace was only reestablished when wealth became more equally distributed and democratic participation expanded to include more of the population.

The critical element in all successful cases was securing enough elite buy-in and a broad enough societal alliance to implement meaningful reforms. This societal alliance can be made of state power holders, elite factions or other interest groups. They have to be strong enough to assert authority over the special interests of powerful elites while having the legitimacy to implement changes.

Also, it seems that the conditions have to get “sufficiently bad” to allow reform. In many of these situations, reform would have been possible much earlier than it was actually implemented, but there was not enough pressure in the system, to highlight to the elites that they had to tolerate that their privileges would be partially dismantled.

Takeaway

Obviously, we have to take this study with a grain of salt, as it is only four case studies and these always have the potential to be cherry picked. Hopefully, the Seshat team will publish a follow up with a bigger dataset, so we can be more sure of the results here.

However, all these insights align quite well with many other topics that we have discussed on this blog. For example, that inequality seems to be an important factor for societal collapse in general, that forging a broad societal alliance and having elite buy-in is important if you want to do reform or that making your society more democratic seems to be a promising way to make it also more resilient against crises.

Therefore, I am hopeful that follow-up studies using larger datasets will confirm these conclusions, though we must wait for their publication before drawing definitive lessons for contemporary crisis prevention.

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

Subscribe now

Endnotes

(1) Another example here are the storylines we explored in an earlier post.

(2) Problem with these kind of studies is that for the participants nothing is really on the line, so we cannot be sure if they really behave like this in their everyday life.

(3) If you want to learn more about this, the CrisisDB team also has another big study about the database, which I have discussed here.

(4) This does not mean that there are only four cases in history, these are just some of the most striking ones they singled out as case studies.

(5) Though only if you were a white male living in England. Everybody else still had a bad time.

How to cite

Jehn, F. U. (2025, May 21). Reform or ruin. Existential Crunch. https://doi.org/10.59350/phqtd-z9p60

References

• Hoyer, D., Bennett, J. S., Whitehouse, H., François, P., Feeney, K., Levine, J., Reddish, J., Davis, D., & Turchin, P. (2024). CRISES AVERTED How A Few Past Societies Found Adaptive Reforms in the Face of Structural- Demographic Crises. OSF. https://doi.org/10.31235/osf.io/hyj48

• Shwom, R., & and Kopp, R. E. (2019). Long-term risk governance: When do societies act before crisis? Journal of Risk Research, 22(11), 1374–1390. https://doi.org/10.1080/13669877.2018.1476900

So, how do you actually write a living literature review? In the more than two years that I have been writing this (time really flies), sometimes I get asked how I go about writing the posts here. With this post, I want to answer this question. While this is focused on a living literature review, I think these notes are probably helpful for everybody who has to do at least some scientific writing or wants to do more of that in the future. Also, at least the first two thirds of the post is exactly the same approach I use to write scientific papers.

Subscribe now

Finding relevant literature

I think the most essential part of the process is finding relevant papers to read and write about. This is more difficult than you might think. If you are working in a highly specialized discipline, you can often just subscribe to your favorite topic on ArXiv and you are golden. However, once you start to do interdisciplinary work (which this living lit review clearly is) you run into the problem that there is no obvious place to look for papers. Over time, I have found some places that yield reasonable results, but if you are on your interdisciplinary journey, you just have to accept that your false positive rate will always be pretty high in the papers you look at. In the following, I will give a quick description on how you can find relevant literature.

Using good keywords

Before we dive into the different services you can use here, one meta point. When you are using many of these services, you have to define what kind of research you are looking for. Here you have to tread the fine line between too general and too specific. You have to find a keyword which is used in the research around the topic you are interested in, but it should not be so obscure that you have to set thousands of alerts to get a paper a week. On the other hand, if you use something like “climate change” as your keyword, you will be absolutely flooded with emails. An example of a keyword on the level of specificity that regularly provides results, but does not overwhelm me with alerts is “Global Food Trade”. Too broad would be “Global Trade” and too specific would be “Global Wheat Trade”.

So far I have found two good ways to generate these goldilocks keywords:

• Feed a few papers to the LLM of your choice and ask it what good keywords would be if you were looking for similar research.

• Search for papers you find interesting on OpenAlex (more about OpenAlex below). Once you get the results of your query, you can add a keywords tab there. These are automatically generated keywords by OpenAlex, based on their own bibliometric analysis. These often capture the ideas you are looking for quite well and can be used to search for more.

Google Scholar alerts

Google Scholar tries to index a large chunk of scientific literature. While it has some blind spots (like that it does not index Zenodo), it usually does a good job in finding papers. When you have a profile at Google Scholar, you can set alerts. This means you get an email every time a new paper is found by Google Scholar which mentions the keyword you specified.

Research Gate

Research Gate is a website for researchers to share their work and connect with other researchers. In the home tab of their website you are presented with papers that might be interesting to you. I have found that the algorithm often works reasonably well, it is just cumbersome to train. It is based on the papers you have uploaded, which people you follow and what papers and topics you have interacted with in the past. This means it takes some work to put in all this info and find interesting researchers to follow, but once this is done, it sometimes surfaces papers I do not find anywhere else.

OpenAlex

OpenAlex is a bibliographic catalogue of scientific papers. It is similar to other bibliographic catalogues like Web of Science, but it has the big advantage that it is completely free and available to everyone. You can search directly on their website, but they also have an API that lets you request papers. To make the usage easier, I wrote some code (thanks Claude) to query the API. This allows you to request papers about specific topics in a defined date range and you will get a nice csv file which you can browse through to check if you found interesting papers.

Newsletters of research orgs

Well, this one is pretty self-explanatory. Research orgs like to boast about the stuff they publish. So, find the newsletters of the research orgs you like and subscribe to it.

TERRA (unfortunately deprecated)

CSER did have a tool that tried to automatically find papers that are relevant to global catastrophic risk. Unfortunately, they recently discontinued it and you can only download the past results. Though, this still contains a lot of gems.

Going to conferences

Find big, interdisciplinary conferences and try to look at as many scientific posters as you possibly can. Obviously, the error rate is pretty high here, but this also gives you the most recent research and papers to look out for in the not too far future and you can easily look at several hundred posters in a few hours.

Looking at the reference list

Scientific papers reference other papers for their claims. If you found a claim interesting, just dig up the referenced paper.

Finding specific papers

Sometimes you already have a hunch of what kind of papers you are looking for, like when there is a gap in argument that would profit from a fitting citation. For this case, there are also a bunch of very helpful tools out there:

• Elicit: This is a search engine optimized for finding the papers you need and it can also automate information extraction. Just ask whatever interests you and it will try to find papers which answer your questions. It is kind of optimized for medical papers, but in my experience it can also be quite helpful for many other fields.

• Connected Papers: The main usage of this tool is to find the web of papers that surrounds a specific paper. You put in the name of a paper that you find interesting and you’ll get a network of papers relevant to it. These papers are selected based on similarity, which is calculated by checking if papers are overlapping in their citations and references.

• Research Rabbit: This is a bit of a mix between the other two tools. Allowing you to search in the network of papers you have specified.

Reading and taking notes

Alright, now you have a big stack of papers you might be interested in, but you still have to read them. It can be tricky to find a set-up that allows you to comfortably read a lot of papers (at least it was for me). But by now I have a workflow that allows me to read papers easily and still be able to extract the information without too much overhead:

• Add the papers to Zotero: Zotero is a great, free tool to organize the things you want to read. It automatically gets you all the metadata for the paper and you can even add them directly from your browser.

• Get the papers on your reading device: I am using a Boox Note Air 2. I like these because they have an E-Ink display, but under the hood this is just a normal Android tablet, so all your normal apps work. But you can use any device you enjoy reading on, as long as it allows making notes. To get the papers on the tablet, I just have a folder on my computer that synchronizes with the tablet. To get the papers in the folder I use the Attanger extension for Zotero.

• Read and annotate the papers: On the Boox I can just underline everything in the paper I find interesting and scribble handwritten notes in the margins. Once I have finished a paper, I sync it back to my computer. I select the papers to read by the date they were added, so first in, first out.

• Get the notes out of the paper: Zotero allows you to automatically extract everything you have underlined in a pdf, so I do just that.

• Making your notes permanent: To make use of your notes, you have to find a permanent place for them, where they can be easily connected with other things you write. For this I use Logseq. In Logseq you can import papers directly from Zotero. Once you have done that, this creates a new page for your paper with all the underlined text being directly added. On this page I then order the notes and write down all my handwritten notes. I also add tags to find things more easily in the future. I also specifically tag all the papers that I think would make a good post and link them to other papers of that topic. Logseq allows you to query your notes. I use this feature to automatically create a list of all the post ideas I have written down anywhere in my notes.

Coming up with ideas

Once you have done all these steps to find, read and digest papers, I think you have already done the most important things you can do to come up with ideas to write about. Because in my experience the amount of ideas you have is very directly correlated with the amount of papers you read. This might seem obvious, but it wasn’t to me when I started to become a scientist (and nobody bothered to tell me). I think this is because in my experience many scientists just don’t read that much, so they never realize what they are missing. I started to move in the direction of reading more after I had finished my PhD, because I noticed that I did not have many good research ideas. On my journey to find out how to have better ideas, I came across the book “Constructing Research Questions” by Mats Alvesson and Jörgen Sandberg. I won’t go into much detail here, because I have already written a summary here, but the main gist of the book is “You have to read more! If you think you are reading enough, read more! If you think you read too much, you don’t, read more instead.” And I think this is spot on. Because if you want to have good ideas, you have to know what others have already done, in what direction your field is developing, what techniques and methods are out there and what the big questions are that people are grappling with. The only way to get this information reliably is to read a shitload of papers. Pretty much all the good ideas I had I can attribute to just reading more and I think simply reading more is a skill that is underutilized by many scientists. All the posts in my living literature review started their life as notes scribbled on pdfs. Same goes for all the papers I have written since my PhD.

Writing a post

This is actually the easiest part, because you have already done all the hard work in the previous steps. The way I go about it is just to look at the automatically generated list of post ideas in Logseq and pick one that catches my interest on that day. Then I take a quick look at all the notes of the relevant papers and draft an outline of my argument. In many cases I can just use my previously made notes almost one to one. Once I have a first draft going, I usually send it to one of the people from whom I have summarized papers and ask for feedback. This results in a positive answer most of the time, because people like it when you are interested in their work and are happy to tell you where you might have misunderstood their argument or technique. As soon as I have incorporated the feedback, I put the post on a schedule to be published.

Technical setup

To distribute my posts, I have two separate systems. Github Pages for my permanent archive of posts and Substack for an easy way for people to subscribe to my work and also to get some more engagement via their algorithm.

Github Pages

I use Github Pages for making sure that I have an easy way to make my posts available for the long term and be able to track changes. Also the actual posts are just stored as markdown files, so if Github tries to do enshittification, I can easily move somewhere else.

Get a Github account

Don’t think I have to explain that one.

Set up your Github Pages with Beautiful Jekyll

Beautiful Jekyll is a template for easy to modify static websites. They have a good get started guide, but I’ll give an overview + some extras:

• Fork the Beautiful Jekyll repository and make sure to rename it to YOURUSERNAME.github.io This allows Github to understand that it is supposed to make a Github Page out of this. This will create a static website at https://YOURUSERNAME.github.io

• Set-up Github Desktop on your computer. This program allows you to download from and upload to the repository you just created. Use it to clone your repository to a local folder of your choosing. Once you change anything in that folder Github Desktop will see this and allow you to upload the changes again to your online repository. Once they are uploaded, Github will build the website again with your changes.

• In your local folder of your repository, you should find a file called _config.yml. This file lets you change pretty much everything on your website. Play around and find out.

If you like how I have set up things, you can also just do these steps, but fork my repository instead of Beautiful Jekyll.

Publishing posts

Your repository should contain a folder called posts. In this folder you can put markdown files with the following naming system YYYY-MM-DD-name_of_your_file.md. When you upload this file to your repository using Github Desktop, Github will create a new post on the date you have specified in the name. If the date is in the past, it will create the post immediately. See here for an example of how these files are formatted in markdown.

Substack

Setting up a Substack is pretty self-explanatory and there are many tutorials out there, so I won’t go into detail here. But what I found to be most helpful to get a wider readership was recommendations from other substacks. To get them, just talk to people that write about similar things as you do and convince them that your work is relevant to their readers as well.

Making everything citable

As this blog is essentially just another form of doing research, I wanted to make sure that my blog is both findable via academic channels and also that the people I cite can become aware of what I am doing here. This seemed quite difficult to accomplish, until I came across Rogue Scholar, which aims to be a home for scientific blogs. It is a platform which automatically creates DOIs for my blog posts, as well as extracting the references I have, so other people can see that I cited them.

To also set this up for your living literature review you have to go through a few steps:

• Create an account at Rogue Scholar (optional).

• Fill out the blog questionnaire. This gives Rogue Scholar the information it needs to figure out if your work is actually a scientific blog.

• Wait a day or so.

• Once Rogue Scholar has decided that your work is a scientific blog, they will go through all your existing posts to archive them and provide them with DOIs. The end result will look something like this.

Every time you add a new post on your blog, this will also be added to the archive. The DOI forwards you to the original blogpost, unless this is not available anymore, in which case it falls back to an archived version in the Internet Archive.

When I have updated posts, I have to inform Rogue Scholar manually, as Substack does not automatically provide this information (Wordpress for example does). Rogue Scholar will then update the version on their end as well.

You can also add additional meta data like your ORCID or funding information, if you want to.

Overall, it is a pretty cool solution! And now every post of mine also has a “How to cite” section.

Promoting your work

This is my least favorite part of the whole endeavour, so I might be doing this suboptimally. But by now I have several channels where I notify people when I have written something new. I will write a short summary of my new post and why I think it might interest people. I then publish this summary, together with a picture from the post and a link to it on LinkedIn, Substack Notes and Mastodon (with a bridge to Bluesky). I also send out an email to all the subscribers of my newsletter. The vast majority of readers come via the email newsletter, but there is enough readership from all the other channels to justify doing the extra work there.

How to start your own living literature review?

Congratulations, you now know how to do your own living literature review. If this got you interested in writing one, you can. Open Phil is still looking for people who want to try out this way of scientific writing. You can find me info here.

Subscribe now

How do you go about finding papers and ideas? How do you process them? Feel free to share in the comments or write to existential_crunch at posteo.de if you want to discuss this further.

How to cite

Jehn, F. U. (2025, April 23). How to write a living literature review. Existential Crunch. https://doi.org/10.59350/2vd97-whc20

Once again it is time to update the existing posts with new papers I have found. This time the majority of them revolve around how trade and democracy keep our society stable.

I also have a little request to my readers: Do you have any paper recommendations for me to read? I have a lot of channels to go hunting for new, interesting papers, but I figured I would be interested in what my readers find relevant. If you have a suggestion, feel free to leave it as a comment here or email me to existential_crunch at posteo.de

Subscribe now

And now the actual updates.

Trade collapse

Food distribution, not just production, plays a critical role in food security. This principle extends to other essential goods as well. I previously examined the consequences of trade disruption or collapse in an earlier post. Since then I have come across two additional papers which explore new facets of this topic, especially in relation to domestic production.

How important is domestic production in comparison with trading food?

Another way to look at food trade is to explore when it is helpful to import a lot of food and when this is more of a liability and you should rather have grown your domestic food production. This is explored in an article by Verschuur et al. (2024). The researchers developed a bilateral trade model for 177 countries and four major crops (maize, wheat, rice, soybean) to simulate how different types of shocks affect food availability and prices. They modeled several scenarios: the Ukraine war, energy price shocks, trade restrictions, and a "polycrisis" combining all three—against 54 years of weather-driven production variability.

Key findings about trade as a buffer:

• Trade generally helps mitigate localized shocks by allowing countries to source food from alternative suppliers. For example, when Ukraine's exports were disrupted, importing countries shifted to other suppliers.

• In typical years, countries with high import dependency experienced smaller reductions in consumer surplus, suggesting trade networks provide resilience against moderate disruptions.

• During the Ukraine war scenario, international trade flows increased as countries sought alternative suppliers, demonstrating trade's adaptive capacity.

However, trade has limitations as a buffer:

• During extreme "tail risk" events (the worst 10% of scenarios), higher import dependency became a vulnerability rather than an asset—especially for maize and rice. Countries with greater domestic production were better buffered against severe shocks.

• Trade can transmit and amplify shocks across the global system. The compound "polycrisis" scenario resulted in consumer price increases of 23-52% across all crops, affecting virtually all countries simultaneously.

• Trade restrictions (like India's rice export ban) severely impacted specific regions, particularly Sub-Saharan Africa and Central Asia.

The energy price shock proved most damaging overall, increasing input costs globally, while different regions showed varying vulnerabilities to specific shock types. The research suggests that both trade connectivity and domestic production capacity are needed in a "risk layering" approach to food system resilience.

In essence, trade provides crucial flexibility for the food system to adapt to shocks, but becomes less effective as buffers during compound, global-scale disruptions. This is a similar conclusion as Yıldırım & Önen (2024), but Verschuur et al. additionally back this up with modelling.

In the post about trade, I have also rewritten the concluding paragraph to account for both the present political situation and to include a report on geopolitics and trade.

This also highlights the importance of geopolitics for trade. We have seen that political ramifications had been a major factor when the trade network collapsed during the Bronze Age and also the shifts after Russia’s invasion of Ukraine are obviously linked to politics. But we can look at even broader trends here. These decades-long geopolitical changes and their influence on trade are explored in Piñeiro & Piñeiro (2024). Specifically, they looked at how much trade was worth in comparison with GDP and found that it peaked before World War I, crashed down in the period of 1914 - 1945 and has consistently grown since then. However, it also seems that either the growth has slowed down or might even have reversed after ~2020. They argue that this recent shift has been happening as countries are stepping away from international treaties around trade. In parallel, many countries are starting to sort themselves into power blocks, especially around China and the United States. They decrease their trade with everyone outside their power block and increase it in their power block. Given the developments of the United States isolating itself more and more since Trump has become president again, it seems likely that in the coming years trade as a share of GDP will decrease and that the power block around the United States might shatter into smaller groups of countries, which would further decrease trade.

In summary, the global trade system, particularly in the food sector, is inherently vulnerable and it is very difficult to predict how disruptions might turn out. This highlights that it should be a priority to increase resilience, so we don’t follow the same path as the empires of the Bronze Age. It seems that domestic food production is an important factor here if you want to be resilient against major shocks. Food trade only seems to help well for smaller shocks. Additionally, the current geopolitical world situation seems to imply that we will see decreased trade in the coming years. While this is likely bad in general, it might increase resilience for some countries, as they have to become more self-reliant. However, globally this is a worse outcome, both now, as everything becomes more expensive when trade decreases, but also after major crises, as these tend to be managed better if countries trust each other and work together.

Democratic resilience

When the value of democracy is discussed, an argument that often comes up is that democracies are just too slow and cannot really look beyond their election periods. While looking into this, I found two very interesting papers that present quite solid counter-arguments against this view. The slowness of democracies is actually needed to make good decisions and over the long term, this leads to much better outcomes for the citizens of democratic countries. I used these two papers to extend my argument about the value of democracy for societal resilience.

Comparison to authoritarian regimes

But democracies are not only better at maintaining peace, they also tend to fare better when a geopolitical conflict breaks out. Brands (2018) used a review to explore how democracies and authoritarian regimes tend to approach great-power conflict and their conflicting worldviews in general. Economically, democracies consistently outperform autocratic regimes in building the wealth essential for strategic success. This stems from the free exchange of information, stable legal frameworks, and protection of individual and property rights that foster investment and innovation. Democracy's decentralized power structure helps avoid rash economic decisions with potentially disastrous consequences, like China's Great Leap Forward or going to war unprovoked. Additionally, democratic militaries tend to perform at higher levels than their authoritarian counterparts, as they typically employ professional forces comfortable with delegated authority and operational flexibility, while authoritarian regimes try to centralize power to be safe from coups.

Authoritarian regimes may boast advantages in speed and decisiveness, democracies excel in sustainable decision-making processes over the long term. The very features that can make democratic governance appear slow and inefficient—checks and balances, power transitions, and robust public debate—are precisely what enable reasoned deliberation and necessary course corrections. Conversely, authoritarian systems, which concentrate power in a small elite while stifling debate, are more prone to significant errors when leaders make emotional or illogical decisions without adequate feedback mechanisms. This long-term decision quality may prove especially critical when addressing complex global catastrophic risks that require sustained attention and adaptive management rather than merely swift action.

A similar point is often made around the ability of autocrats to govern for the long term. Here the argument is that democracies fail to foresee problems over long time horizons, because they can only think along their election periods (so mostly 4-5 years), while autocratic regimes are not bound by this and can therefore plan for decades ahead. The validity of this argument is explored in Millemaci et al. (2024). To study this, they looked at economic growth and long term positive societal outcomes (think education, health, public transport, taking public opinion into account) and compared it between democratic and autocratic countries. For economic growth they find results similar to previous literature, meaning that generally democratic countries outperform autocratic regimes. However, autocratic regimes have heavier tailed distributions. This means that while autocratic regimes are generally worse when it comes to economic outcomes, they also represent the largest outliers for both very slow (or even negative economic growth), but also very fast growth. But economic growth in and of itself is not something we really care about, because a higher GDP does not automatically mean that people are better off. This makes the study of Millemaci and co-authors especially interesting, as they also take into account positive societal outcomes in addition to economic growth and here the results are very clear. Democratic countries across the board deliver better societal outcomes to their citizens in things like health or education. This finding even stays robust when you control for how rich the country is in general. Therefore, Millemaci et al. conclude that there are no benevolent dictators who govern for the long term. They might be able to boost economic growth, but this only means they have more riches for themselves and their cronies. Their population would have a much better quality of life, if their country were democratic instead.

Lessons from the past for our global civilization

In this post I explored what lessons we can get for our present world by studying history. These lessons are especially relevant in a time of complex crisis, as we have now, as they give additional points of reference and orientation. I extended the discussion in the post by another paper which also argues for the relevance of premodern history for today.

The relevance of premodern history for today

After the above-mentioned papers have explored from what kind of societies we can learn, I also want to discuss a review paper by Haldon et al. (2024), which explored the ways history is relevant for policy makers and especially, how we can make sure to take reliable lessons from the past.

They argue that we rely on history all the time to make political decisions, because if you face a crisis, it is only natural to try to figure out how past decision makers reacted in similar situations. You need something to ground yourself in. However, this runs into several problems. History is complex and policy makers are often stripped for time. This means they tend to cherry pick easily, as they default to the most well known history facts from their own country. However, this leaves out the vast majority of lessons we could take from history. For example, this leaves out lessons from large sample and modelling studies, both of which try to capture the more general trends in history.

Haldon and co-authors also make several suggestions on how we could improve history uptake in politics in the future. They think that especially expert elicitations are underused here. With this they mean to ask a large group of experts about a specific topic, which is relevant to policy and then aggregate all those voices into one report, which highlights the main lessons to be drawn and also in what areas the experts' opinions diverge from each other. In addition to this, they suggest we should aim for qualitative-quantitative data integration. With this they mean combining the strength of large datasets, with specific case studies. Use the large dataset to identify a recurring trend and then give an example of how this trend played out in a particular historical crisis, which is relevant to what the policy maker is looking for (1). Their hope is that such an approach will make it easier to rely on history for policy makers, as in current debates, the insights of scientists from the natural sciences are often ranked higher in their trustworthiness. It is important for history to bridge this gap, as otherwise many important insights from history will remain unheard in our policy debates.

How long until recovery after collapse?

In this post I had explored how long it might take a society to go from hunter-gatherers to modern day civilization. I added another short paragraph to discuss a bit more the factors that decide if a state is founded or not.

Likely, the factor relevant here is you have to have an environment which is conducive to agriculture. This not only includes climate, but also things like your local soils or if you have plants and animals available which can be domesticated. This argument is underpinned by a study from Borcan et al. (2021). They collected a dataset which consists of the majority of societies who created states. What they wanted to find out was the question if usually you had agriculture first and then the formation of a state or the other way around. Their results show that this is the case. In pretty much all cases you had first agriculture and state formation only afterwards. Besides this they found that geography, climate, distance to other states and time since the first settlement founded in the area all play a role in the time it takes to go from hunter-gatherers to early states.

Further reading

If you come this far in the post and are still searching for more interesting content, here are some suggested readings from the archives:

• Science denial and nuclear winter: Nuclear winter is the idea that after a nuclear war, soot will partly block out the sun and cause temperatures to fall globally. This hypothesis is contested, partly for scientific reasons, but also for political ones. In this post I explore the latter.

• How long until recovery after collapse? In this post I try to estimate how long it might take us to recover from collapse of different levels of severity.

• Is societal collapse just a random event? One of the possible causes that has been suggested for collapse is that sometimes societies just get unlucky and are destroyed by pure chance of several bad events happening at the same time. This post explores if this idea is a helpful way to frame things.

Subscribe now

Until next time

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me an email to existential_crunch at posteo.de and we can schedule something.

How to cite

Jehn, F. U. (2025, April 16). April 2025 Updates. Existential Crunch. https://doi.org/10.59350/mm5xg-mpb70

References

• Borcan, O., Olsson, O., & Putterman, L. (2021). Transition to agriculture and first state presence: A global analysis. Explorations in Economic History, 82, 101404. https://doi.org/10.1016/j.eeh.2021.101404

• Brands, H. (2018). Democracy vs Authoritarianism: How Ideology Shapes Great-Power Conflict. Survival, 60(5), 61–114. https://doi.org/10.1080/00396338.2018.1518371

• Haldon, J., Mordechai, L., Dugmore, A., Eisenberg, M., Endfield, G., Izdebski, A., Jackson, R., Kemp, L., Labuhn, I., McGovern, T., Metcalfe, S., Morrison, K., Newfield, T., & Trump, B. (2024). Past Answers to Present Concerns. The Relevance of the Premodern Past for 21st Century Policy Planners: Comments on the State of the Field. Wiley Interdisciplinary Reviews: Climate Change. https://doi.org/10.1002/wcc.923

• Millemaci, E., Monteforte, F., & Temple, J. R. W. (2024). Have Autocrats Governed for the Long Term? Kyklos. https://doi.org/10.1111/kykl.12425

• Piñeiro, M., & Piñeiro, V. (2024). Geopolitical changes and their implications for agricultural trade negotiations. https://hdl.handle.net/10568/151905

• Verschuur, J., Murgatroyd, A., Vittis, Y., Mosnier, A., Obersteiner, M., Godfray, C., & Hall, J. (2024). The impacts of polycrises on global grain availability and prices. https://doi.org/10.21203/rs.3.rs-3969801/v1

• Yıldırım, C., & Önen, H. G. (2024). Vulnerabilities of the neoliberal global food system: The Russia–Ukraine War and COVID-19. Journal of Agrarian Change, n/a(n/a), e12601. https://doi.org/10.1111/joac.12601

What do you think about when you hear the term revolution? Likely, you think of the French Revolution and things like the Storm of the Bastille, the Guillotine or Louis XVI’s execution. It is on the forefront of our minds, because the French Revolution escalated so quickly and so violently that it left a permanent mark on the world. It left us with the idea that once you start a revolution, it is a torrent that takes everything with it. This shaped what we expect of a revolution and provided us with a script on how to conduct one. Revolutionaries ever since have taken up this script again and again. Especially, the October Revolution took strong inspirations from the French one. This includes: the revolution having to devour its own children, the unmasking of a hidden enemy of the people, and the split into two extreme factions. The fact that everybody only thinks about the French Revolution annoyed Hannah Arendt to no end. She argued that there is a revolution which was actually (mostly) successful, which we could get inspired by instead: the American Revolution. To set everybody straight, she took it upon herself to lay out in detail, how these two revolutions happened, the dynamics inherent in them, the failure of the French Revolution, the success of the American Revolution and what the latter should have done to make democratic government worthy of that term.

Subscribe now

What even is a revolution?

The ultimate aim of revolutions is freedom—specifically, public freedom, which means the ability to shape one's political environment. To achieve this, liberation must come first, freeing people from economic or political constraints. Yet liberation alone is insufficient. To have public freedom you need a republic, meaning a state where the citizens decide the trajectory of their nation. This is in contrast to private freedom, meaning being left alone from state intervention in your private life. This kind of freedom is also possible for many in authoritarian regimes.

Revolutions are a force that has shaped our history, especially in the 20th century. They are different from the rest of politics, because they have to include at least some violence. If you want to change the overall makeup of a state, there is no chance that everybody in your society will be alright with this. So, from the beginning if a revolution happens, it really shakes things up. This does not have to be done via physical violence, but it can also involve violence in other spheres. For example, a revolution in norms can mean violence to existing social structures, traditions, and hierarchies. The key insight is that revolutions fundamentally reorganize power - whether through physical force, institutional change, or cultural transformation.

Revolutions are a relatively recent phenomenon, they did not exist before the 18th century. Before this, revolution was inconceivable. To even think about a revolution, you have to understand that time is not cyclical and especially that who is rich and powerful in a society and who is poor and oppressed are not the only state a society could take. This shift in perspective was powered by the Enlightenment and secularization. If it becomes clear that the shape of the society you live in is not ordained by God, this means you actually have the chance to mold your society into something different from what it currently is.

The Enlightenment thinkers were not the first to think more deeply about public freedom. The idea of public freedom is quite old and was already discussed in ancient Athens. The main idea was and is that to be free you have to create a political framework which allows people to meet on equal terms. This means in such a framework everybody starts with the same, equal rank. The assumption was that you can only be free if you interact with people who are neither above nor below you. This also means that you are not free if you rule above others, because you have no one equal to you.

We also have to distinguish a revolution from a rebellion. A rebellion is aimed to replace the person in power, like a king being replaced by another, more successful king. However, this is quite different to a revolution. A revolution aims at creating a completely new societal system, which allows people to be free. While we still have rebellions today, revolutions could only start once you realized that structural change is possible.

Revolutions have the problem that they want to destroy the old order, while also creating and stabilizing a new one. This means they need to use violence, but at the same time try to constrain the violence, so it does not destroy themselves. Both the French and the American revolution did not start out as revolutions. Both simply wanted to reconstruct rights they thought the monarchy had deprived them of. The upcoming revolutionaries were quite convinced that they would just restore an already agreed upon contract that the monarch failed to comply with. However, they ended up creating something new. They saw it as self-evident that every human has inherent political rights. However, this was a completely new idea, which had not existed before. Once they took power, they realized that being able to shape your society and what you can do within it is something very powerful and alluring. This made them push further and introduced the self-reinforcing nature of the revolutions.

Arendt terms these experiences public freedom and public happiness. With this she means the opportunity to partake in politics and the joy gained from shaping your own environment. In both the French and American Revolution, the experience of public happiness and public freedom was something completely new for everyone involved. Up until then, you simply had no opportunity to shape your environment politically. This gave the revolutionaries an additional motivation to push for more self-governance. They wanted to maintain this opportunity to govern their own fate.

Why did the French and American Revolution take such different paths?

Liberation versus freedom (aka the social question)

Very poor people are often pro revolution, because their current state sucks and every change is better than staying as it is. However, if your life is so harsh that every change is good, this also means that there is a lot of desperation going on, which pushes the revolution to just focus on liberation, to bring as many people out of poverty as quickly as possible. However, if you only focus on liberation, freedom is often sacrificed, because it can come into conflict with liberation, e.g. by complicating political processes. This conflict between balancing liberation and freedom is what Arendt calls “the social question”.

The overall focus on liberation during revolutions can be traced back to Karl Marx. He assumed that as long as people are poor, they cannot be free. Therefore, to make people free, you have to liberate them. His idea was that a small group of elites owns the means of production and keeps the surplus that the workers produce for themselves. In a feudal society, it was hard to get out of this trap, but after the industrial revolution a new working class formed which had a stronger bargaining power, as they were more concentrated spatially, which allowed them more effective resistance.

Liberation was a larger topic in France than in the American colonies. In France, you had a very unequal society, with the vast majority of people being desperately poor. This was less the case in America, where misery was much less common. However, this is only partly true, because the Americans still maintained slavery at that time. This meant that there was maybe a similar amount of misery in America as in France, but it was more hidden from the overall society. Nonetheless, this more equal situation between the white population in the American colonies allowed them to focus less on liberation and more on freedom.

The Americans saw freedom as the higher good, because they saw it as essential for having a good life. On the one hand, living in a free society allows you to more easily follow your private happiness (e.g. starting a family), because when you are in a free society, if the state does something that inhibits your private happiness, you can get involved and change it. A free state also allows you to get public happiness, the joy from shaping your own political environment in the free deliberation of equals. This is what the founding fathers in America saw as the ultimate source of meaning and the basis for a life well lived.

The general will

As established above, liberty and freedom can come into conflict, especially if a majority of the population lives in misery. This was especially strong in the French Revolution, as in pre-revolution France the gap between the rich and the poor was so vast. Generally, the rich were seen as morally corrupt, as they did not change the misery of the masses, even though they would have had the means to do so. From this moral failure, the French revolutionaries concluded that wealth corrupts humans and as the poor have no wealth, they are therefore uncorrupted and morally superior to the rich. Building on this the French revolutionary thinkers came up with the concept of the general will. The idea is that the general will is the sum of all individual wills of the population. As this is dominated by the poor, it should give superior direction on where to aim the revolution. The problem with this theoretical idea is that you cannot really know the general will, as this would require a hive mind. This is obviously impossible, especially in a feudal society before the industrial revolution. Therefore, the general will needed an interpreter in the form of the revolutionary. As they could not really know what the general will was, they just intuited it and their intuition was that they had to relieve the poor of their misery right now and everybody who stood against this aim was a traitor to the revolution. The problem was that you cannot simply abolish misery in a whole country in a month. Large scale changes like these need careful planning and years or even decades to implement. The revolutionaries did not grasp this and so they thought their failure to end misery in France was only possible due to treason. These traitors to the revolution had to be purged, which led to the reign of terror and our view of the French revolution as brutal and failed.

So, why did the same not happen in the USA? As mentioned above, the white population in the USA did not face as much misery as the poor in France, as they had outsourced this misery to the slaves. This made wealth redistribution a less urgent topic to the white population in America. Additionally, as most white citizens had their basic necessities taken care of, they had the time to think about higher goals, like public freedom. Finally, as the distance between the rich and the poor was much smaller, they had more opportunities to engage in discussion and so there was a wider knowledge of what people actually thought. This led to a structure in the USA that was much more based on deliberate discussion on different levels of government and less to the centralized idea of the general will.

Compassion versus solidarity

This different approach to the general will versus the deliberate discussions can also be found in which values were upheld in the different revolutions. During the French Revolution the value which was seen as most important was compassion. This was based on the idea that compassion is the opposite of reason, which was seen as being too much inward focussed and leading to egoism. Therefore, you should foster compassion, as it is focussed on the suffering of your fellow humans and motivates you to take actions that will make the lives of others better. The problem with compassion is that it can only ever be focussed on a single person, you cannot have compassion for a whole nation. Also, as it is focussed on the suffering right here, right now, it aims to alleviate this suffering without delay. This meant that the focus on compassion, also led to a focus on liberation.

Arendt thinks that a much better concept to focus on is solidarity, as it is the more abstract version of compassion. It allows you to employ reason better and also allows an easier focus on a large class of people, instead of focusing on the individual. Also, it allows you to treat others more as equals to yourself, as in contrast to compassion, which always implies a power difference. This kind of solidarity was the focus in the American Revolution, as their colonial experience had taught them that they have to be able to rely on each other to survive.

Self-governance

The colonial experience of the Americans also led to another factor that shaped the revolution. They had much more experience in self-governance. The American settlers traveled across a whole ocean to start their project in the Americas. This meant that they were far away from the British government and had to make most decisions on their own. Additionally, in the colonies, you had to rely much more on others than in the United Kingdom. Creating new towns in an uncharted continent meant that you had to trust others to keep up their end of the bargain. If everybody is on their own, you will not make it. Another factor was that people cannot be forced to participate in your endeavors. If they don’t like what you do, they can easily leave. All these factors coming together meant that many decisions in the American colonies were done via negotiation on a local level. Due to these discussions many people were already familiar with democratic procedures before the revolution even happened. Therefore, the Americans had a much easier time than the French to adapt to this new form of government.

The French in contrast lived in a steeply hierarchical feudal society where the king had massive influence in every domain that caught his attention. This strict hierarchy meant that almost no one in France had the opportunity to practice open deliberation and consent in decision-making before the revolution broke out. As it turned out, leading a revolution, while also learning how to do democracy was too much at the same time and contributed to the failure of the French Revolution.

Absolutism versus constitutional monarchy

Revolutionary ideas don’t exist in a vacuum. They are shaped by the society they are developed in. This can be seen in the way both the Americans and the French developed their revolutionary ideas. Both looked at the kind of governments they were accustomed to: absolutism and constitutional monarchy. The French oriented a lot around absolutism, because this had been the kind of government in France for a long time. This meant that when they thought of power, they thought of strict hierarchies and centralized power. In this image, they created many of their revolutionary institutions. Essentially, they removed the king and tried to replace him with the people, but kept everything else in a similar power structure. This meant that they also did not think much about separation of powers, because such a separation had never been present before in France. Their removal of the king left a king-shaped hole in their state structure and so they tried to create structures that centralized power, because this is what had been necessary for absolutism.

While the Americans also focussed at some ideas of absolutism (mostly because much of the theories around at that time had been developed under absolutism), they also took up many ideas from the constitutional monarchy of the United Kingdom. This meant that they thought much more about separating powers and introducing checks and balances, because this is what the constitution in the United Kingdom had done.

Arendt argues that we got really lucky that a revolution happened in America. Otherwise, we would have only seen revolutions in previously absolutist states like Russia and France and their failure to always end up in a centralized dictatorship, because they oriented around absolutism too much.

Founding a new nation

The role and structure of a constitution

A major step of a revolution is writing a constitution. You start a revolution because you want to change the status quo. However, once you have overthrown the old regime and started to implement new rules, you somehow have to keep those new rules in place, to make sure that you don’t get overthrown yourself. Most successful constitutions have a deep distrust of centralized power and focus on checks and balances. They have to strike the balance between constraining the government from having too much influence on the life of its citizens, while also being strong enough to stop one part of society from dominating other parts.

Like in many other details of the revolution, the French and American Revolution took different paths here. The French laid their emphasis on abstract rights that they saw as inherent in humans by birth. The Americans thought that rights are not something you inherit by birth, but that instead rights are something that humans give each other. In this spirit, they structured their constitutions based on separations of power and checks and balances. Their idea was that only power can keep power in check, and therefore you need separate power bases in the government which keep an eye on each other. The French constitutions, however, never really came off the ground. They were too abstract to be really implemented, as they focussed more on theoretical rights of humans and less on constraining power.

Power and authority

The difficulties of implementing a stable constitution can be traced back to power and authority. Both are necessary for a stable society. Power in this case means giving someone else the opportunity to influence your life. For power to exist, you need at least two people and a relationship between them. Power can be freely given or taken by force. If you want to have a democratic society, you have to find some mechanism that brings people to give away power over them freely. This is where authority comes into play. You have to have some convincing justification on why people should allow you power over them. Before the Enlightenment, this authority was often derived from religion: “I should have power over you because God said so”. However, once you start secularization and people believe less in God, this argument becomes much less convincing.

Another way to get authority is to refer to a more glorious past, which you presumably orient around. This was for example done in the Roman Republic. Rome saw itself as a reincarnation of Troy and thus Greek culture. From this they derived their authority: “I should have power over you, because I adhere to the ideas of a great culture of the past, much better than all of us in the here and now”.

However, if you start a revolution, this also does not really work well, because the whole idea of a revolution is to replace the current system with something new. But if you have something completely new, you cannot link back its authority to what came before. This means after a revolution, you need a new way to derive authority. Arendt argues that in this case, the authority comes from the act of foundation: “I should have power over you, because we both together created the rules that we follow now”. During the American Revolution this was especially powerful, because you had a stepwise process of delegation of power under mutual consent. On the lowest levels you have town halls, which gave everybody the chance to contribute and elect a representative for the writing of the constitution of the state people were living in. These representatives in turn selected those who should write the constitution of the Union of all states. This allowed every citizen in the American colonies to contribute at least somewhat to the constitution and thus experience public freedom and public happiness.

The trade-off between public freedom and stability

This puts you in a bit of a difficult situation. Because as we have established above, one thing the revolutionaries realized is that the big upside of starting a revolution is being able to shape your own political surroundings. Writing a constitution is the ultimate expression of public freedom. The Americans even wrote this striving to shape your own environment in their declaration of independence, where they put the pursuit of happiness as one of the main goals. While this is today often framed as striving for your own private happiness, it seems more likely that at least some of the founding fathers meant public happiness here. This is what their whole revolution had been about. However, if writing a constitution is the best thing ever, you should also make it available to everyone after you. If you also think that the constitution you just wrote is pretty good, you want to prevent it from being changed. This means you either have to be open to your constitution being changed completely, or deny everybody else the public happiness of writing a constitution. If you chose the latter, this also means that you have no authority over everyone who did not participate in the foundation of your state.

The Americans tried several ways to get out of this problem of authority:

• Having a supreme court: The supreme court checks if the rules of the constitution are still upheld and strikes everything down which they think does not follow those rules. While having such an institution grants some authority by making sure the agreed upon rules are upheld by everyone, it does not solve the problem of not being able to change the foundational rules and thus denies newcomers real public freedom.

• Allowing constitutional amendments: This way you can adapt the constitution to a changing environment. However, it still denies you the option to change the underlying rules, you can just add additional rules.

• Recreating your constitution every generation: The founding fathers discussed including the rule that the constitution would have to be rewritten every generation, to make sure that everybody is still on board. However, this idea was never implemented.

This means that the American Revolution was not able to find a solution to the problem of authority beyond the first generation of founders. Thus, it ultimately failed to bring true public freedom and public happiness, but it was still better at this than everything that came before.

How to truly bring public freedom and public happiness?

The problems of direct and representative democracy

We now know that the American Revolution ultimately failed to archive its aims of public freedom and public happiness. And in the eyes of Arendt, no state or revolution since then has really achieved this, because she thinks that both direct and representative democracies come with flaws which ultimately deny public freedom. They all default to focus on liberty, and private happiness at best. This creates environments where the best you can hope for is being left alone by the state or maybe being somewhat supported by it, but not having the opportunity to shape your overall kind of government. In addition, many of them even see public happiness and public freedom as obstacles in the way of liberation and private happiness.

This is plausibly rooted in the way many of our modern democracies are organized. You often have either opportunities for direct democracy (mostly implemented as referendums), but this suffers from being strongly coupled with public opinion, and the unorganized public opinion is highly volatile and thus not conducive to good government. It needs to be refined and organized in some way to be productive and produce long term good government. Every human has potentially an opinion on every topic you could ask them, but only in deliberate debate, these opinions can become reasonable policy.

One way to have these discussions is representative democracy. Here, citizens elect those representatives (often around every 4 years) which they think represent their opinions and values best. These representatives can then debate with each other to find the policies which best overall represent what the population wants. While this is less volatile than direct democracy, but it comes with other problems:

• Citizens are essentially powerless except on election day.

• The representatives are also fallible humans and as they are mostly without checks for the four years they are in office, they can easily be captured by special interests.

• It cements the rule of the few. Almost nobody can experience public freedom except the few people who become career politicians.

This means that neither direct nor representative democracy are a valid path if you want to have both stability and public freedom.

Societies, town halls and the dangers of party politics

So, if the common ways to structure a democratic government are ruled out, what is left? We can find some inspiration and cautionary tales in both the American and the French revolution. In the case of the American Revolution, the hints for how to implement continuous public freedom can be found in the role of town halls in the original writing of the constitution. As laid out above, they allowed deliberation on the lowest level of government and a step-wise escalation of power in ever more potent assemblies on state and then Union level. They were not considered in the constitution, but this does not mean that such a structure can never again be implemented again.

Similarly, during the French Revolution all across France so called “societies” formed. These were local clubs which were spontaneously funded all across the republic. They started because people felt both the need to discuss politics and also govern their local administration. They were true places to experience public freedom and public happiness, as they were open to everyone who wanted to discuss and govern on a local level.

However, during both revolutions, these decentralized organs of power came into conflict with the more centralized power of representative government. These fights revolved around three main points:

• How much public freedom can be allowed if it comes into conflict with liberation?

• Should the government represent the general will of the population or the democratic diversity of opinions?

• Should power be monopolized or distributed?

All of these fights were won by the proponents of more centralized power, meaning more focus on liberation, representation of the general will and monopolized power. But why did these two structures even come into such a stark conflict? Here, Arendt thinks the culprit is party politics. If you have political parties, these tend to radicalize over time and start to see things more in an “everybody who is not with us is against us” frame. This process is especially strong in countries with a history of absolutism, as we can see with the French and the Russian Revolution. In both, we had a party system which radicalized itself and fought each other until only one party was left standing, which then assumed centralized and monopolized power.

Another problem of the party system is that it forces private citizens to find alternative ways to get involved, if they want to have more influence than casting a vote every four years. One of the classical pathways to do so is using lobbying to influence the political representatives. However, this process is highly undemocratic and favors those with large amounts of wealth, and ultimately leads to the capture of the parties by special interests.

Council republics as a way to unite stability and public freedom

These dangers inherent in party politics were also very present to the founding fathers, especially Thomas Jefferson. One suggestion he made was to change the United States to a council republic. This would mean keeping a system similar to the foundation of the United States, which would have town hall councils open to everyone, these would then elect representatives for the next higher level, e.g. a county. This process of electing representatives would repeat until it would reach the highest level of the country. Using this approach would produce both a structure which is stable, while also having a very low bar of entry for everyone, thus allowing public freedom and public happiness for everyone who wants it.

Such council republics seem to regularly arise spontaneously during revolutions. Another example are the societies mentioned earlier. But there are many other examples of council republics:

• Paris Commune in the revolution of 1871.

• Soviets after the February Revolution in Russia in 1917.

• Räte in Germany in 1918/19, especially in the Bavarian Council Republic.

• Soviets in the Kronstadt rebellion of 1921 against the party dictatorship in Revolutionary Russia.

• Councils during the Hungarian Revolution of 1956.

In all these cases, the councils were crushed by the power of the party system and military force. Councils just did not fit their view of how a revolution should work. Neither Marx nor the theoreticians of the French Revolution included councils in their thinking, and so they also weren’t part of the revolutionary lore. However, this recurrence of councils again and again in so many revolutions in places not really connected with each other, show that they are something that people naturally flock to and which they enjoy participating in.

A council republic also has the clearest separation of power possible. Essentially, every citizen keeps the power of every other citizen in check. This happens at every level of the council systems, as on each level only individuals of equal power and rank confront each other, always allowing to constrain overly power hungry individuals.

Lesson for today?

So, why haven’t we implemented such a system on a large scale today, if it is so great? One big problem is that all the incentives of an already established party system are against allowing councils. It would require them to both cede power and see the political process less as a “them against us” and more as a collaborative process. Similarly, council republics are seen as a challenge to the status quo, and every such challenge will face headwinds from those profiting from the current systems or seeing it as optimal. For example, one criticism of council republics in this regard is that people don’t actually want public freedom and public happiness, but instead only private happiness and liberation, while not really caring about who is in charge, as long as they face no interference in their private lives. Finally, many of the council republics which have existed in the past have had problems with administration, as they sometimes elected people to administrative roles on political grounds. As none of the council republics mentioned so far existed for a longer period of time, they also did not have the time to sort out these kinks in the systems.

And actually, we kinda have implemented them on a variety of scales and configurations, we usually just don’t call them council republics. A large-scale example would be Switzerland. While not being a council republic, Switzerland is organized in a very federal way, delegating much of its decision power to lower and local levels of government. Another large-scale example is Rojava. This autonomous region in northern Syria has implemented a system of governance that shares similarities with the council republic concept. It employs a bottom-up approach to governance, with local councils playing a significant role in decision-making. An example on a smaller scale would be citizens’ assemblies. Such assemblies are meant to resolve difficult political questions in open discussion and give recommendation to the political entity which instituted them. These have been done on local and regional level, but there are also some nationwide examples or even some that cover the whole European Union.

Besides that, the current times also seem much better soil for the implementation of council republics than before. Arendt argues that some of the main problems included that most people had to focus on liberation instead of public freedom, because otherwise they would just starve. Also, many of the people that participated in the failed revolution lived under absolutist governments, this meant they had no chance to practice public participation. Both of these things are not true anymore for many countries. Especially in Europe, very few people are so poor that they are on the verge of starvation, and there exist many opportunities today to learn how to engage in discussion with others. This means people today are much better prepared for something like a council republic. However, my own main criticism of the book is that it does think too little about the economic circumstances needed to be able to participate in politics. Even today many more people would be interested in doing politics, but are barred from it, because they have to work so hard that at the end of day, they have no energy to still get involved in politics. They might not be starving, but if you have to decide between being able to pay your rent and being active politically, the choice is a difficult one. Therefore, if you had something like a universal basic income, this would be a much more level playing field and I see it as an important component of future democratic systems.

Though following Arendt’s argument means that the establishment of such council republics is still called for. The American Revolution failed in bringing true public freedom, but this does not mean that it has to stay this way forever. Our current democracy is not enough. In addition, in the current world we have additional opportunities due to technology to implement decentralized, deliberative systems. For example, this has been done in Taiwan, where every citizen can bring forward topics into the political process, if they can convince enough others that it is worthwhile.

The question is no longer whether such systems are possible, but rather how we might adapt and implement them to strengthen democratic participation in our own contexts. The tools and knowledge exist—what remains is the political will to experiment with and refine these approaches to build more inclusive and deliberative democratic institutions.

Subscribe now

How to cite

Jehn, F. U. (2025, March 19). Revolution and democracy. Existential Crunch. https://doi.org/10.59350/w4pr7-e0e43

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

Imagine a row of dominoes where knocking over one piece triggers a cascade that topples them all. This is systemic risk - when a single failure can bring down an entire system. Now imagine several such cascade failures happening simultaneously at a global scale and making each other worse. This occurrence of heightened global systemic risk and its actualization is the global polycrisis we face today. Let’s explore these concepts that help us understand the complexity of modern global threats.

Subscribe now

What is systemic risk?

I am relying here a lot on a recent systemic risk governance paper by Schweizer & Juhola (2024) which is about systemic risk governance, but also explains nicely what systemic risk is about. To understand what systemic risk might be, we first have to understand what a system is. The definitions here tend to be fuzzy, but generally everything is a system when you have a bunch of elements which interact with each other and which have a boundary to the rest of the world (1). The interaction between these elements is the most important part, as they create the complex behavior of the system which is difficult to impossible to derive from inspecting the elements in isolation. The scale is not important however. My body can be seen as a system, as can a company, the electrical grid or the banking sector. The important thing is that within the boundaries of the system, you have elements interacting with each other.

To understand the factors involved here, we also have to look at an additional paper here by Mark et al. (2023) who tried to lay out in a clear way the definition of polycrisis, but also touch upon many of the concepts which are important to understand systemic risk. More specifically, they talk about what a risk consists of. They see risk as the product of the potential of something to disrupt your system (the hazard), the vulnerability of your system to this hazard and if your system is exposed to them. The hazard that could disrupt your system can either build up gradually (e.g. the infrastructure of a country slowly getting worse, because needed investments aren’t being made) (2), which we call systemic stress or they can increase rapidly (e.g. when an earthquake disrupts many systems at once and so drastically and quickly heightens the risk of something going wrong).

Following from these explanations of what a system and what a risk is, a systemic risk is the potential for disruption of a single element in my overall system, that brings the whole system crashing down. This usually happens via feedback loops or risk cascades. This means you have a disruption that reinforces itself or damages more other elements in every time step. To make some examples:

• Imagine a severe blockage in a major artery, such as the coronary artery, which supplies blood to the heart. This blockage could lead to a heart attack, causing the heart to pump less effectively. The reduced pumping efficiency impacts the delivery of oxygen and nutrients throughout the body, affecting vital organs like the brain, kidneys, and liver. These organs may begin to fail due to insufficient blood supply, leading to cascading failures across multiple systems in the body.

• A systemic risk involving a company is the collapse of Lehman Brothers in 2008. As a major international financial services company, Lehman Brothers’ bankruptcy triggered a domino effect that led to a wider banking collapse and contributed significantly to the 2008 financial crisis.

• Another example is the interaction of the electrical grid with telecommunication. Imagine you have a severe blackout due to a geomagnetic storm. After a few hours to days the telecommunication system breaks down. However, to restart the electrical grid you need telecommunication to coordinate and so the disruption gets much larger as the systems rely on each other.

Schweizer and Juhola identify three main properties which systemic risks usually have and which make them so difficult to deal with:

• Complexity: When you have systems that consist of many elements and connections, like the electrical grid, it becomes very hard to determine what is going on, because there are so many feedback loops and unexpected interactions, that it is just too much to keep track of easily. Also, your system can have elements which change in their behaviour over time (in this case we call them complex adaptive systems), like the behavior of companies changing after the 2008 crisis after new regulations were made. Finally, systemic risk can spread from one system to another, like in the example of the electrical grid and telecommunication above. This is called a ripple effect.

• Uncertainty: As systemic risks are so complex, it is often quite difficult to determine what is cause and what is effect. Also, as the elements in the system change, over time the system can react differently to the same input (this is called a regime shift). Taking all these difficulties together, we get deep uncertainty, which just means that things are so complicated that you cannot give meaningful forecasts anymore, as everybody disagrees what is even going on and what might be done about it.

• Ambiguity: This complexity and uncertainty lead you to the situation that the data and knowledge you have about your systemic risk allow several, conflicting interpretations.

Schweizer and Juhola also explain what you should consider, if you want to understand a specific systemic risk:

• Understand the risk emitting part of your system. Not all elements in your system are equally dangerous to the stability of the system overall. Often you can identify key elements. These are often those where one system interacts with another system.

• Figure out which institutions and organisations are involved in your systemic risk and which regulatory rules they follow. If you want to reduce systemic risk, you have to know who is even involved and what rules they adhere to.

• Understand the debates around your specific systemic risk. As systemic risks are so complicated, people tend to disagree about what to do about them. Only if you roughly understand all the sides, you should proceed on tackling it.

Finally, Schweizer and Juhola talk about what you have to do to govern systemic risks well. This part of the paper feels a bit vague to me, so I won’t go into too much detail. But the main gist is that they argue that systemic risks are super complex, uncertain and ambiguous. Therefore, if you try to solve systemic risks alone or in a small (homogenous) group, you are extremely likely to fail, as you will overlook some important connections and facts. This means that systemic risk governance has to make sure to include a large number of stakeholders, so you even have a chance of understanding the dynamics you want to govern and what might be done to constrain them.

Recently, a discussion also started around global systemic risk (Centeno et al., 2015). Here the idea is that due to globalisation, we have created one single system, which therefore also has its own inherent systemic risk. Centeno and co-authors highlight this system is especially vulnerable, as it is a system of systems, extremely complex and governed by the profit incentives of capitalism. In a capitalist system you are paid for efficiency, not for resilience against seldomly occurring risks. While this might be less of a problem if you have a regional trade network of non essential goods, it can easily become a major crisis if you have a global trade network of essential goods. A classic example here is food. While in any given year many countries are better off in buying their food on the world market, as this allows you to have cheaper food, this puts you in a perilous position if the trade stops (e.g. due to export bans or extreme climate disruptions). In such a case, you suddenly cannot buy food anymore, but as you also stopped producing food, you now have a famine on your hands.

But Centeno et al. argue that this is not the only global systemic risk. They also highlight:

• Finance: The global finance system means that economic disruption in one sector of a given country can lead to disruption in a completely different sector in another country. For example, the abovementioned collapse of Lehman brothers, contributed to the financial crisis of 2008, which in turn decreased food availability due to disruption of the global market.

• Critical infrastructure: One recent example here is the blockage of the Suez Canal by the Ever Given transport ship. The ship blocked the canal and led to supply chain issues for months all around the world. Modern supply chains optimize for efficiency over resilience. Just-in-time manufacturing reduces costs but leaves little buffer against disruptions. When semiconductor production was disrupted during COVID-19, cascading effects hit industries worldwide.

• Climate change: Climate change touches on pretty much every other topic globally and increases the risk there. Both abruptly due to more extreme weather events, but also chronically due more strain on all systems in a warmer world.

What is a polycrisis?

We now know that global systemic risk is the potential for disruption on a global scale, which is then realized because a single element in the system fails. The polycrisis is essentially the perfect storm we are experiencing right now of multiple global systemic risks being triggered at the same time, making each other worse and leading to a much more difficult response, as you have to put out so many fires at once.

Currently, there are two overlapping, but somewhat conflicting definitions in use. There is the definition coming more from the climate science/planetary boundary camp, which is laid out in detail in Lawrence et al. (2024) and the definition by the history camp, explained in Holder et al. (2024), as well as Mark et al. (2023) (which we also discussed above). Both camps are in agreement that we currently are in a polycrisis, that this is bad and that we should do something about it. They also agree that a polycrisis is about interacting systems on a large scale, which are under stress and are hit by shocks. The main thing where they disagree is how we should think about the term crisis. The climate science camp argues that we should use crisis in the sense that a potential risk has materialized, while the history camp would rather define crisis as a prolonged period of heightened risk where crucial decisions have to be made, independent from the fact if the potential risk has materialized or not. While there are some differences in how these scholars conceptualize crisis, both approaches emphasize the interaction of multiple systems under stress, with risks manifesting through both acute shocks and chronic pressures.

Lawrence et al. (2024) define polycrisis as “the causal entanglement of crises in multiple global systems in ways that significantly degrade humanity’s prospects”, but let’s unpack that step by step. With crisis they mean a sudden event which quickly causes a large amount of harm on a lot of people. They argue that what is so distinct about the current polycrisis is that we have so many crises at the time, which are also all interacting with each other.

They think we need such a new term, because in classical risk assessment you usually only focus on a single chain of events, like how a storm of a certain size would impact your country. And while we surely need this kind of risk assessment, it just doesn’t work for the global situation we find us in, because if you try to solve our global crises one by one, you will fail, as they feed into and interact with each other. If you don’t look at that interaction, you won’t be able to understand and thus solve them.

Lawrence et al. also try to establish a framework on what properties a polycrisis has:

• Multiple causes: As many causes of the polycrisis happening in parallel it is often difficult to establish what caused what. They also interact with each other, which leads to emergent problems.

• Non-linearity: Global systems don’t react in a linear fashion. This means that even a small disturbance can bring the whole thing down.

• Hysteresis: Once a system in your society is pushed out of its stable state, it is often not easily reversible and you have to push much harder than the impact that removed the system from its stable state.

• Boundary permeability: There is no clear boundary to the interacting crises, neither spatially (going from local to global) nor in the kind of system (natural, social, technological).

• Black swan outcomes: The nature of such complex risk is that you have a very wide range of outcomes, some with extraordinarily large damage and it is quite hard to predict where you will end up.

Lawrence et al. argue that we are currently in a polycrisis and that things are getting worse. They think that the main crises of which the polycrisis consists are the aftereffects of the COVID-19 pandemic, economic problems, volatility in food and energy markets, geopolitical conflict, political instability and civil unrest due to economic insecurity, political polarization, declining institutional legitimacy, and the effects of climate change. All these factors interact with each other and make each other worse.

They also think that these crises have some main underlying causes. The Earth is leaving its holocene climate equilibrium as we keep emitting greenhouse gases, the global energy system is starting to shift away from fossil fuels, the geopolitical world is changing to a world not centered around American hegemony, the global economic systems is shifting away from the neoliberal regime to something new and we don’t know yet what and artificial intelligence disrupts the world with unknown consequences.

From my reading of the history camp, I would think that they agree with pretty much everything that Lawrence et al. lay out here, with the exception of how they use crisis of course, but Holder et al. (2024) also have something else to contribute to this discussion, which I found quite helpful to wrap my head around the idea of the polycrisis.

They argue that much of our confusion around the term polycrisis happens because we think that something either is or isn’t a polycrisis. Instead we should think of a polycrisis as a spectrum. A crisis is more a polycrisis if the spatial scale is larger, there are more systems involved and there is a longer and larger build up of stress in the systems. This means on the one side you would have a very localized crisis, in a single system with no systemic stress involved. An example of this could be a bridge that collapses due to an unexpected hazard, like an earthquake. The other side of this spectrum is the world we live in today. Due to globalization we have essentially connected every system with all other systems, the scale is global and as we have been building up systemic stresses in the form of overstepped planetary boundaries and rising inequality.

Is this our first polycrisis?

Both camps also agree that polycrises have happened in the past and will happen again after we are past this one. They can and have been solved.

One paper that explores these past polycrises more is Hoyer et al. (2023). They use the Seshat database and look at a variety of past polycrises to better understand how and why they happened. Hoyer et al. show how multiple factors affect how well societies handle polycrises. While institutional structure matters, they also emphasize how ecological conditions and social bonds between groups shape responses. For example, they highlight how societies often fail when elites block needed changes.

Crises happen all the time and usually societies are able to solve them. Only when you increase the vulnerability of your society or when you are unable to address the vulnerabilities of your society you start to have a real problem. Hoyer et al. argue that one of the main processes here is when the state gets captured by elites. Once power, wealth and social privilege are entrenched, collective actions become much more difficult, as those people who have the power to change things, are very much interested in keeping the status quo. To highlight this, they looked at a bunch of case studies:

• Qing dynasty in China: In their early period, strong institutions - especially their granary system - helped them weather climate challenges from the Little Ice Age. However, similar environmental challenges in the 19th century contributed to the dynasty’s collapse, not because the threats were worse, but because social pressures had weakened state capacity. Overpopulation, a rigid bureaucracy creating frustrated elites, and corruption had eroded the institutions that previously provided resilience. This shows how environmental threats don’t exist in isolation - their impact depends heavily on social and political context.

• Ottoman Empire: The Ottoman Empire in the 16th century demonstrates how societies can survive severe environmental and social pressures. During this period, the empire faced multiple challenges: the Little Ice Age brought cold temperatures and drought (including the worst drought in Ottoman history in the 1590s), while the Celâlî Rebellions created internal unrest as local elites competed for declining resources and positions. Combined with external wars, these pressures led to a polycrisis that killed half the population in some regions. Yet unlike many societies facing similar challenges, the Ottoman Empire survived. Their resilience stemmed from maintaining effective infrastructure and resource distribution systems, particularly water management. While the Qing dynasty’s similar systems later collapsed due to corruption, the Ottomans sustained theirs until conditions improved. This allowed them to maintain stability despite decades of stress, though at significant human cost in some regions. Their experience shows how societies can survive multiple overlapping crises if they maintain strong institutions and resource distribution systems.

• Monte Albán in Mexico: The hilltop city of Monte Albán in Mexico shows how rising inequality, rather than environmental stress alone, can drive societal decline. For nearly 1000 years from 500 BCE, Monte Albán thrived despite its arid location by maintaining relatively equal social structures. This included distributed public spaces, similar-sized houses for elites and non-elites, and collective infrastructure projects. When Monte Albán declined around 800 CE, many blamed drought and warfare. However, recent evidence suggests a different story. While the region did face drought, nearby settlements actually grew as Monte Albán shrank. The timing coincided with growing inequality - elite houses became larger, ruler imagery more prominent, and public spaces less accessible. People “voted with their feet,” abandoning the city as social disparities widened. This pattern appears across pre-contact Mesoamerica - settlements with more equal, collective governance tended to better survive environmental challenges. The case demonstrates how environmental threats often become dangerous only when combined with social inequality.

Interaction of systemic risk with global catastrophic risk

Another topic which interacts with systemic risk and the polycrisis is global catastrophic risk. Arnscheidt et al. (2024) explored this in much detail, aiming to map out how events that would constitute a global catastrophe are made more likely by systemic contributions. The outcomes considered a global catastrophe are much larger than those usually considered in the polycrisis literature (think large asteroid impact instead of global financial crisis and geopolitical tension) Global catastrophic risk mitigation is just harder in a more unstable world. If countries are constantly trying to put out the fires of the polycrisis, they just have fewer resources left to do something about global catastrophic risk, which leaves our society more vulnerable. Arnscheidt et al., similarly to Hoyer et al., also highlight that one key thing here is the vulnerability of our society. A hazard is only a problem for you as long as you are vulnerable to it. Unfortunately, it seems that the global system we have built enables worse outcomes than in the past. We have more new and dangerous technologies, a lot of built-up vulnerabilities (like loss of social cohesion) and have a higher global connectedness which leaves us more vulnerable, even to crises which start far away. This can also be seen already today that “small” events like the blockage of the Suez Canal by the Ever Given, was an event that was felt globally, even though it was just a single ship, blocking just one of the many trade routes we have globally.

Another important factor that Arnscheidt et al. highlight is so-called latent risk. The idea is a risk that only becomes a problem once the conditions around it change. A classic example of this is stratospheric aerosol injections to reduce global warming. This introduces the risk of termination shock, which means the sudden warming if the aerosol injections would halt. This halt could also be involuntary due to global disruptions like war.

Arnscheidt et al. also argue that our global risks have increased due to systemic changes we have introduced, like climate change, nuclear weapons or AI. These increase our vulnerability to many other risks as well. Highlighting these systemic contributions to the vulnerability is important, as often in global catastrophic risk research, hazards are framed as something that you cannot really do something about, like “oh, I guess somebody is just going to build AGI and there’s nothing we can do about that, we should just focus on alignment”. However, the increased risk from such hazards is not inevitable. We could have addressed arising hazards in the past politically, and we can still do so in many cases. For example, if we can find agreements that reduce global warming, climate change mitigation becomes much less important.

Polycrises can erode the social and institutional capacity needed to handle both global catastrophic risk in general, but especially the preventive measures needed to stop hazards from growing. Political polarization and declining trust in institutions make coordinated responses to global threats more difficult.

How might we govern this mess?

Coordinating around such complex problems is quite difficult. Where even to start? One thing that is clear, that global coordination is needed one way or the other. An attempt to get this started was the United Nations’ Summit of the Future. The summit was meant to start the reform of the United Nations 75 years after their founding and resulted in the Pact for the Future, which spelled out 56 goals which should guide the way to tackle the existing global challenges. How well this worked out and what we might do better in the future is discussed in Studzinski et al. (2025). They start with the criticism that the Summit of the Future did have a too narrow view on individual hazards and not really the bigger picture that would be needed to address the challenges we face. But the world as it is, is faced by threats that are larger than the sum of their parts. They are interconnected and reinforce each other.

The inability of global governance to effectively address the current global problems is caused by a variety of factors. One of them is that the situation is just tricky in general, and so it is to be expected that it cannot be solved easily and quickly. But also many of the existing institutions for global governance have ambiguous aims and missions, are often working against each other or are trying to find the least disruptive policy, in an attempt to please everyone. All in a world where international agreements are taken less seriously and the funding for global governance is decreasing.

But this does not mean that such summits are a waste of time. Ultimately, events like this are one of the few avenues of global governance that currently exist. Future summits need to get rid of their single hazard focus and instead focus on the interconnected problems we actually face. In addition to that, regional organisations like the European Union or the Shanghai Cooperation Organization could step up their game and implement measures to tackle systemic risks in their own borders, to showcase how this might also work on a global level. None of this is easy and likely needs a paradigm shift on how systemic problems are viewed, but getting this right is essential for the coming decades.

Conclusion

To summarize this: systemic risk is when a single element in your system is enough to disrupt the whole thing and a polycrisis is when several systemic risks turn into actual crises and interact with each other on a global scale.

Understanding systemic risk and polycrisis reveals both the fragility and resilience of our global systems. While a single failure can trigger cascading collapses, history shows that societies can build institutions and practices that help weather multiple crises happening at once. The challenge we face today is maintaining and strengthening these protective systems even as our world grows more interconnected and complex.

Subscribe now

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

Endnotes

(1) System boundaries are analytical tools rather than natural divisions. Climate impacts don’t stop at national borders, and financial crises spread through complex global networks.

(2) Looking at you Germany.

How to cite

Jehn, F. U. (2025, February 19). Systemic risk and the polycrisis. Existential Crunch. https://doi.org/10.59350/bhpka-t7g93

References

• Arnscheidt, C. W., Beard, S. J., Hobson, T., Ingram, P., Kemp, L., Mani, L., Marcoci, A., Mbeva, K., hÉigeartaigh, S. Ó., Sandberg, A., Sundaram, L., & Wunderling, N. (2024). Systemic contributions to global catastrophic risk. OSF. https://doi.org/10.31235/osf.io/wcj9e

• Centeno, M. A., Nag, M., Patterson, T. S., Shaver, A., & Windawi, A. J. (2015). The Emergence of Global Systemic Risk. Annual Review of Sociology, 41(1), 65–85. https://doi.org/10.1146/annurev-soc-073014-112317

• Holder, S. L., Ainsworth, R., Aldrich, D., Bennett, J. S., Feinman, G., Mark, S., Orlandi, G., Preiser-Kapeller, J., Reddish, J., Schoonover, R., Turchin, P., & Hoyer, D. (2024). The Spectrum of (Poly)Crisis: Exploring polycrises of the past to better understand our current and future risks. OSF. https://doi.org/10.31235/osf.io/3bspg

• Hoyer, D., Bennett, J. S., Reddish, J., Holder, S., Howard, R., Benam, M., Levine, J., Ludlow, F., Feinman, G., & Turchin, P. (2023). Navigating polycrisis: Long-run socio-cultural factors shape response to changing climate. Philosophical Transactions of the Royal Society B: Biological Sciences, 378(1889), 20220402. https://doi.org/10.1098/rstb.2022.0402

• Lawrence, M., Homer-Dixon, T., Janzwood, S., Rockstöm, J., Renn, O., & Donges, J. F. (2024). Global polycrisis: The causal mechanisms of crisis entanglement. Global Sustainability, 7, e6. https://doi.org/10.1017/sus.2024.1

• Mark, S., Holder, S., Hoyer, D., Schoonover, R., & Aldrich, D. P. (2023). Understanding Polycrisis: Definitions, Applications, and Responses (SSRN Scholarly Paper No. 4593383). Social Science Research Network. https://doi.org/10.2139/ssrn.4593383

• Schweizer, P.-J., & Juhola, S. (2024). Navigating systemic risks: Governance of and for systemic risks. Global Sustainability, 7, e38. https://doi.org/10.1017/sus.2024.30

• Studzinski, N. G., Kent, R., & Korowicz, D. (2025). Towards the Governance of Global Systemic Risk: Reforming the Summit of the Future. Global Governance: A Review of Multilateralism and International Organizations, 31(2), 113–136. https://doi.org/10.1163/19426720-03102002

This post is part of a living literature review on societal collapse. You can find an indexed archive here.

While writing this living literature review on societal collapse, one topic that keeps popping up again and again is economic inequality. It seems that this inequality makes societies inherently less stable and less able to internally coordinate and thus more vulnerable to collapse. Some examples of us exploring how inequality makes society more vulnerable:

• Democracies are likely the best form of government if you want to be more resilient against collapse and democracies work less well if your society is highly unequal (Link).

• Many of societally harmful processes seem to increase with higher inequality (Link).

Therefore, this post is meant to explore the topic of economic inequality in more detail. We also touch upon how economic inequality contributes to inequalities of power.

Subscribe now

Historical examples of the influence of inequality

There are many examples of how inequality has lead to worse outcomes for the societies involved, but I would like to highlight two examples here:

Comparing Egypt’s and England’s reaction to the Black Death

This case study is described in the CrisisDB paper, which we discussed extensively in a previous post. Hoyer et al. (2024) discuss how both Egypt and England reacted to the Black Death in the 14th century. They choose this comparison because at that point in time both countries were quite similar when it came to infrastructure, population, agricultural commerce and overall complexity. However, the plague affected them quite differently. While both countries faced a high death toll (up to one third of the population), they chose different strategies to cope with this crisis. Egypt’s rulers decided the best way to address their problems was to leave the struggling peasants on their own and instead funnel their resources in great and expensive construction projects. This resulted in a rise in grain prices, higher land rents and an overall drop in production, leaving Egypt much worse off than it was before the plague. England’s ruling class on the other hand did not manage to suppress their peasants as effectively, which allowed the peasants to push for higher wages. This meant that the elites, rather than the poor population, had to cover much of the costs of the pandemic. This resulted in rising living standards in England and likely contributed to England being the source of the industrial revolution.

In another recent paper by Cohn (2023) the impacts of the Black Death on Europe are explored in more detail. He highlights similar points as Hoyer and explains that the results we found in England, can also be found in much of Central Europe as well. In most of these countries, the much decimated peasant population managed to get into a better position by labor becoming the scarcest factor in the whole supply chain. It also diversified the agriculture away from only farming labor intensive wheat and instead creating a more diversified agricultural landscape with other cultures like vineyards or olive groves. This was a bottom up adaptation. Peasants just would not accept anymore the grueling conditions they had to endure before. During this time wealth inequality dropped considerably due to the better bargaining positions of the peasants, leading to overall much increased health and productivity. However, there were also places like Russia which managed to suppress their peasants brutally. This then resulted in similar problems like in Egypt, lowered productivity and a much slower development of the country in general.

The comparative effects of inequality in the Roman, Han and Aztec Empire

Another large comparison of how inequality plays out in comparable polities is by Alfani et al. (2025). They created a dataset to compare inequality in the Han Empire and the Roman Empire. In addition, they compared both to a previous dataset from the Aztec Empire. Thankfully, for us the Roman and Han Empire were quite good at record keeping. While comparable on many axes, the Roman Empire tended to be less centralized and more self rule in the provinces.

Alfani and colleagues calculated the wealth of the different regions by the rate of their urbanization and population density, based on the assumption that more and larger cities and a higher population meant a province was richer. This showed that for the Han Empire the central province was quite rich, while the other provinces were much less so. In comparison, the Roman Empire had also a very rich central region, but also several other regions (e.g. in North Africa) which had considerable wealth. This is likely shaped by how both states used their military and bureaucracy. The military is usually deployed in frontier regions, meaning that wealth gets transferred there from the center, while the bureaucracy tends to accumulate in the central region, syphoning away wealth from other regions. The Han Empire had small military and large bureaucracy, while it was the other way around for the Romans.

In addition, to this inequality between regions, the paper also calculates the inequality in the population in general. They do so by tracking the size of the different groups of people in the empires (like peasants, aristocrats, merchants, etc.). This allows them to calculate a Gini index for the Han, Aztec and Roman Empire and compare it to the present day United States (Figure 1).

Figure 1: Comparison of income shares in the Roman, Han and Aztec Empire and the United States.

This highlights stark differences between the different empires, with the Aztec Empire being the most unequal one by far, followed by the Han and then the Roman Empire. The United States has not yet reached the levels of inequality of those past empires. However, what is notable, that relatively speaking the poorest 10 % in the United States are poorer than their counterparts in all the three other empires.

These differences in inequality are also mirrored in differences in stability. From those empires considered, the one with the highest inequality (the Aztecs) was the most unstable one, while the one with the lowest inequality (the Romans) was most stable. Obviously, these are only three data points and thus not super reliable, but it fits in with the general trend of the instability-inducing power of inequality.

Moral collapse and inequality

Another perspective on how inequality undermines societal stability comes from Blanton et al. (2020) (1). They studied 30 premodern states to understand how governmental fairness affected their resilience. Their key finding was that states with more equitable practices - like fair taxation, limits on leadership power, and impartial courts - generally achieved higher citizen welfare. However, these more equitable states faced an interesting paradox: while they didn’t collapse more frequently than less equitable states, when they did collapse the impact tended to be more severe.

The authors suggest this happens because more equitable states enable citizens to build complex, interconnected systems of mutual benefit. When elites begin abandoning their societal obligations through corruption or tax avoidance, it triggers a cascade of defection from these cooperative arrangements. Citizens who see leadership enriching themselves at society’s expense become less willing to contribute their fair share, creating a self-reinforcing cycle of declining social cohesion. The authors demonstrate this pattern through case studies of major civilizations like the Roman Empire and Ming Dynasty, where increasing elite capture of resources often preceded collapse.

This research reinforces how economic inequality, particularly when driven by elite exploitation, can erode the social foundations that make complex societies possible. When those at the top stop contributing proportionally while still extracting benefits, it undermines the moral consensus that encourages broader social cooperation.

The tricky thing is to understand why this change occurs in the first place. If the elites of your society have contributed so far, why do they suddenly defect? I haven’t yet come across a super convincing argument here, but one possibility is the idea of elite overproduction. This concept is a part of the structural-demographic idea of collapse, which we have discussed in the first post of this living literature review. To quickly recap the idea here:

Secular cycles describe a recurring pattern in history. It starts with a growing population that also has room to grow. The growth leads to more resources for everyone, which leads to an overall cooperative atmosphere. However, at some point, the room to grow shrinks, so that people can only get more resources if they take it from others. Population increases also depress real wages whilst also leading to an overproduction of elites relative to elite positions.This decreases trust and peace until it possibly triggers a redistributive event, which could be something like a civil war, which in turn decreases the population and the cycle starts anew.

This is a good starting point, but I think there is much more research needed here to get a better understanding on why this happens. For example, this could also be caused by long run cultural shifts or a change in climate.

The importance of reducing inequality across different time horizons

Let’s also consider how economic inequality manifests across different time horizons. Here Schmidt & Juijn (2024) provide a great summary. The aim of their paper is to make the argument that economic inequality is a problem across short, mid and long term time horizons and should therefore be tackled no matter how you discount the future.

Short term here means around 50 years. In such a time span we have lots of empirical research to study how inequality plays out in different societies. Schmidt and Jujin make both an individual and a societal argument. The individual argument is that life satisfaction seems to increase with wealth, but it does so in a sublinear fashion. This means that if you give the same amount of money to a rich and a poor person, the poorer person will get more life satisfaction out of the same amount of money. Therefore, if you want to have more life satisfaction in your society overall you should distribute wealth more equally. The societal argument is that if we look at societies there seems to be a clear statistical relationship of more equality and better societal outcomes. More equal societies have better mental and physical health, higher levels of trust, better educational outcomes and less crime. All things which likely make your society more stable (like having more trust and effective institutions). This means the short term case for reducing inequality is pretty strong.

In the medium term (500-1000 years) it gets a bit more tricky to understand how inequality plays out, because we have much less data. Schmidt and Jujin think that we can still extrapolate from the short term outcomes. If your society has more trust, better health outcomes and better education, it seems highly unlikely that this will be worse for society than the other way around. In addition, there is the danger of path dependency, meaning if you have inequality now, it might reinforce itself and you are stuck with it for a long time. Schmidt and Jujin also make the argument that there is some research that tries to ground this argument empirically. Specifically they highlight research that shows that how you distribute wealth in your society also influences how you distribute political and legal power in your society. Meaning, if your society has a high economic inequality for a long time, this will cement power structures in the hands of those that have the most wealth. It could also lead to less buy-in of the general population, as they feel left out (similarly to the moral collapse idea explained above, Schmidt and Jujin refer to the work of Acemoglu and Robinson a lot). As we know that more democratic governments are less vulnerable to catastrophes, this is a concerning outlook and it seems likely that reducing economic inequality is also a good thing for your society in the medium term.

To further strengthen this argument for the medium term Schmidt and Jujin also look at potential counterarguments. The strongest one they find is the idea by Tyler Cowen that over longer time periods economic growth is the most important factor that increases welfare and we should therefore be very careful around redistributing wealth. If reducing inequality would decrease economic growth, this would therefore be a strong argument against reducing inequality. There are two competing worldviews at play here. One being that more inequality creates stronger incentives to work hard and take risks, therefore increasing economic growth, the other being that more inequality reduces how well your society can function, due to decreased trust and underdeveloped public goods, therefore decreasing economic growth. The ultimate answer to this discussion seems still to be discussed a lot, but Schmidt and Jujin think that the best empirical datasets from the OECD we have right now seem to point in the direction the less inequality leads to more economic growth. So, if you consider economic growth the most important factor for welfare, you should be in favor of reducing inequality.

However, if reducing inequality really creates more economic growth, we might run into another problem: climate change. Generally, economic growth is coupled with carbon dioxide emissions. But this relationship of growth and emissions differs between countries and Schmidt and Jujin highlight that recent research has found that in many countries we can find a positive correlation of income inequality and carbon emissions, meaning the per capita emissions are lower if there is less inequality. This is potentially due to better societal coordination and higher trust in more equal societies, which enables better environmental protection. So, it seems that over the short and medium term, we can be reasonably sure that reducing inequality is a good idea. But what about even longer time horizons?

Longer term here means a few thousand years and beyond. This part of the argument gets more hypothetical and abstract, as we don’t have anything empirical we could ground this in. However, the general idea is quite straightforward. Over long time horizons the largest danger for humanity likely comes via things like global catastrophic risks, think nuclear war or runaway climate change. If we can reduce the chances of such events occurring, this is likely the most important thing we can do for the long term future. As we know from the arguments around short and medium term inequality reduction, reducing inequality leads to a more stable society and better institutions, which in turn are better able to react to crises. If people are more trusting due to lower inequality, then it becomes easier to work on international problems like climate change or arms control.

Because we do not know how the future will play out, a good opportunity to influence it positively will be to give the people living in that future a stable society and good institutions, because this will allow them to be more flexible. Therefore, reducing inequality is likely good for long term time horizons as well.

The Current State of Inequality: A Tale of Two Trends

The current state of global inequality presents a complex and seemingly paradoxical picture. While overall global inequality has actually decreased over recent decades, this decline masks significant variation both between and within countries. Inequality between nations has fallen as countries like China and India have grown faster than rich countries. However, this has been accompanied by rising inequality within many nations. This trend has been highlighted nicely in two recent papers Kanbur (2019) and Chrisendo et al. (2024).

This divergence can be seen in several clear patterns:

• Rich Countries: Countries like the US and many European nations have seen increasing internal inequality despite their overall wealth. As Chrisendo et al. (2024) show, around 68% of the global population now lives in areas where inequality has increased.

• Asian Economies: Large Asian economies like China have experienced dramatic growth alongside sharp increases in inequality. China’s Gini coefficient (2) rose from 0.35 in 1995 to 0.53 in 2010, though it has shown some signs of plateauing since then.

• Latin America: Countries like Brazil have managed to reduce inequality, though from very high initial levels. Brazil’s Gini coefficient fell from 0.55 to 0.47 between 1990 and 2021, achieved through targeted policies like cash transfers to poor families and minimum wage increases.

• Africa: The picture in Sub-Saharan Africa shows perhaps the most variation, with inequality decreasing in Western Africa but increasing in Eastern and Southern Africa. This region also faces the double burden of high inequality often combined with very low incomes.

What makes this picture particularly important is its implications for societal stability. Historical evidence, like the contrasting responses to the Black Death in Egypt and England, suggests that how societies manage inequality can profoundly affect their resilience and capacity to handle crises. The fact that inequality is rising for such a large proportion of the global population - while we simultaneously face increasing global challenges - raises important questions about our collective capacity to address future crises.

However, the varied outcomes across regions demonstrate that rising inequality is not inevitable - policy choices matter. Countries that have successfully reduced inequality have typically done so through deliberate interventions, whether through social programs, minimum wage policies, or other redistributive measures. As Kanbur emphasizes, while technological change and globalization create pressures toward rising inequality, national policy can mediate these forces - though increasingly this requires international coordination to be fully effective.

These patterns suggest we are living in an age of rising inequality not because inequality is universally increasing, but because the economic forces of technological change and globalization are aligned to exert upward pressure on inequality. How societies respond to these pressures through policy choices appears to be the crucial factor in determining outcomes.

How did we end up with the current state of inequality?

This current situation of rising in counties and lowering inequality between countries can plausibly be attributed to neoliberalism. This arc of neoliberalism over the last decades is described in a lot of detail in Centeno & Cohen (2012). In the late 20th century, neoliberalism emerged as an attractive ideology promising both democracy and prosperity through free markets. As Centeno & Cohen (2012) note, this was especially compelling in the 1990s when the success of the “Asian Tigers” seemed to demonstrate how global market mechanisms could lift countries out of poverty. The apparent triumph of capitalism over the Soviet model further cemented the appeal of neoliberal ideas.

The core tenets were straightforward: minimize government intervention, privatize public services, and deregulate markets. The theory was that this would maximize economic freedom which would, in turn, guarantee political freedom. This “indivisibility thesis” - the idea that economic and political freedom were inseparable - became deeply influential in policy circles.

However, the implementation of neoliberal policies had several consequences that increased societal vulnerability:

1. Weakened State Capacity: Through extensive privatization and deregulation, states reduced their ability to respond to crises and provide public goods.

2. Increased Inequality: The focus on deregulation and tax reduction, particularly for the wealthy, has led to growing inequality.

3. Broken Feedback Loops: As wealth and power became more concentrated, decision-makers became increasingly insulated from the consequences of their choices, as they mainly hear what needs to be done from lobbyists. This has made it harder for societies to recognize and respond to emerging threats.

All of these things tend to make a society more vulnerable to societal collapse as shown in this post and many others in this series (3). This highlights that we have to shift away from many of the neoliberal ideas of the last decades to have a better chance to face the challenges that are likely ahead of us in the next decades.

But to track down the roots of present day inequalities it makes sense to dig even deeper. There is a great paper by Sullivan & Hickel (2023), which explores how poverty developed from the 16th century until today. The motivation behind the paper is to show that deep poverty is not something that every society experienced and which was only fixed by the wonders of economic growth in capitalism, but instead that in many countries people were actually fine and only started to suffer once capitalism started. They think that the former view was able to take hold, as for these arguments per capita GDP is used as a proxy of how poor people were. However, this proxy has three big problems:

• There are many examples in history where GDP per person improved, while people clearly experienced worse life. They specifically highlight the Philippines between 1820 and 1902. In this time period GDP per person rose by 15 %, but pretty much every record at the time only talks about how living standards are just completely deteriorating.

• Past GDP per person is based on calculating purchase power parity. This means you calculate how much stuff people were able to buy with the money they had at the time (4). However, these calculations are highly error prone, especially for food.

• Capitalism existed for a long time before GDP per person started to improve. The basis of capitalism started to get established in the 16th century. However, from 1600-1800 GDP per person in many countries actually declined.

To avoid those problems, they want to look at measures of how well people actually lived, not some abstract measure such as GDP. For this they selected real wages (specifically how long someone had to work to get a day’s worth of food for a four person family), height and mortality rates. They track all these markers from the 16th century till today in Europe, Latin America, Sub-Saharan Africa, South Asia and China.

Their clearest finding is that in all of those places there was little to none extreme poverty. Most workers were able to provide for a family of four. The time periods where they were able to find extreme poverty were usually during major wars or in some exceptional countries like Peru who were suffering from especially harsh colonialism.

This changed once the given region started to transition to capitalism. In all cases height and real wages decreased, while mortality went up. Some of these trends were reversed when strong social democratic or socialist movements came to power in the first half of the 20th century. However, many of these advances in human well being were rolled back again after neoliberalism’s rise in the later 20th century.

This does not mean that becoming a socialist state is always the best way to avoid poverty, as there were also examples like China where this ended in the population getting poorer. However, it clearly shows that capitalism cannot be singled out as the main reason wellbeing improved. Instead it looks more like it is actually holding many societies back.

What can we do about it?

So, our current state of inequality is far away from the optimal state and it is important for both the short, medium and long term future to bring it down again. But how might we go about that?

As we have seen above, a pandemic that kills a third of the population is one option, but this is not the preferred way of reducing inequality and as we have seen above also not even guaranteed to reduce inequality. One example of reduced inequality without a brutal shock like a pandemic is described by Marienbach (2024). He explored the city of Teotihuacan in Central America. This city is quite interesting because it showed a very low Gini coefficient of 0.12 (5). This is an incredibly low value and essentially unheard of in modern states. The closest country today is Slovakia with 0.24. However, in contrast to Slovakia, Teotihuacan was the dominant power in its region and outperformed the less equal cities around it.

Marienbach goes on to look at the most common explanation of why states or cities are able to reduce their inequality to a lower level. The main things he looks at are pandemics, mass warfare, revolutions and state failure. None of these apply to Teotihuacan. Before the arrival of the Spanish, there were no large-scale epidemics in America, the city was much stronger than its surrounding cities and therefore did not need to result in large-scale warfare and it also experienced neither revolution nor state failure. So, what kept the city on a low inequality for several hundred years? Marienbach suggests it was their system of government. Due to a volcanic eruption in the vicinity in the region, many neighboring cities have to be abandoned. This led to a large influx of people into Teotihuacan. This resulted in around 20 groups of similarly powerful social units inhabiting the city. They kept each other in check and prohibited that a single faction could take power and amass wealth.

This also ties in with the findings above that inequality is not a given, but depends on how we govern and how we decide to distribute the wealth in society. Vodovnik (2024) makes some suggestions here on how a new politics that is able to withstand societal collapse might be oriented. He ties into the critique of neoliberalism and that this political ideology has left behind one very important connecting idea in society: care. In a society focused on only individuals which are always in competition with each other it becomes difficult to build shared projects and ideas and make sure that everybody is taken care of. To accomplish this we need to enable people to have both the ressources and time to do the work that is necessary to have a democracy. This means for example to be able to even afford to spend time on work in politics. As I discussed in another post, an important stepping stone here might be a universal basic income (6). This would allow people to do the care work which is ultimately necessary for a stable democracy and thus the ability to withstand large scale catastrophes.

Building a more equal society requires what Vodovnik calls “caring democracy” - a system where care for others is not subordinated to market logic but is understood as the foundation of political and economic life. This means a democratic system that prioritizes social welfare and equitable resources distribution, so its citizens gain the freedom of not having to constantly worry about their future and present day needs (7). While we may not achieve the remarkably low inequality of ancient Teotihuacan, history shows that alternatives to our current system are possible when we reimagine the basic relationships that structure our society. The question before us is not whether change is possible, but whether we can make the transition thoughtfully and deliberately (e.g. by implementing gradual policy reforms) rather than waiting for a crisis to force our hand.

Or to make things more concrete we can also look at a paper by Easterlin (2012) who found the gap in life satisfaction between rich and poor in both capitalist and socialist countries. He found that generally the gap was considerably smaller in societies that were either socialist or had a strongly redistributive capitalist economy (think Norway). This means if we want to maximize overall life satisfaction and economic equality in society we have to implement redistributive measures at least as strong as in Norway. This would address two problems brought by neoliberalism that we identified above. Namely, that we decrease inequality with all its bad consequences for societal resilience and fix the broken feedback loops, as in societies where economic power is distributed more equally, no single person can easily hijack the system just by being rich.

Thanks for reading! If you want to talk about this post or societal collapse in general, I’d be happy to have a chat. Just send me a mail to existential_crunch at posteo.de and we can schedule something.

Subscribe now

Endnotes

(1) We also discussed this paper in the participation post.

(2) The Gini coefficient is a measure of inequality. A value of 1 means one person has all the wealth, 0 means all people have the same amount of wealth.

(3) For example, how basic democracy and state capacity are quite important for societal resilience.

(4) Similar calculations are done today for example when comparing military budgets. Nominally, Russia has a much smaller military budget than the United States. However, the same amount of money is able to buy much more tanks in Russia, as labor and resources are much cheaper there.

(5) This was calculated by looking at the size distribution of houses in the city to serve as a proxy of overall wealth.

(6) Universal basic income also has some potential problems and is not the ultimate solution to all our problems, but it could be a good first step and I did not want to make this post even longer with a discussion of the upsides and downsides of it.

(7) Present day examples of this would be the “Wellbeing economics” in countries like New Zealand

How to cite

Jehn, F. U. (2025, January 15). Economic inequality and societal collapse. Existential Crunch. https://doi.org/10.59350/yrvk7-w7589

References

• Alfani, G., Bolla, M., & Scheidel, W. (2025). A comparison of income inequality in the Roman and Chinese Han empires. Nature Communications, 16(1), 3248. https://doi.org/10.1038/s41467-025-58581-0

• Blanton, R. E., Feinman, G. M., Kowalewski, S. A., & Fargher, L. F. (2020). Moral Collapse and State Failure: A View From the Past. Frontiers in Political Science, 2. https://doi.org/10.3389/fpos.2020.568704

• Centeno, M. A., & Cohen, J. N. (2012). The Arc of Neoliberalism. Annual Review of Sociology, 38(Volume 38, 2012), 317–340. https://doi.org/10.1146/annurev-soc-081309-150235

• Chrisendo, D., Niva, V., Hoffmann, R., Sayyar, S. M., Rocha, J., Sandström, V., Solt, F., & Kummu, M. (2024). Income inequality has increased for over two-thirds of the global population. Research Square. https://doi.org/10.21203/rs.3.rs-5548291/v1

• Cohn, S. (2023). The Black Death: Collapse, resilience, transformation (M. Centeno, P. Callahan, P. Larcey, & T. Patterson, Eds.; pp. 192–206). Routledge. https://eprints.gla.ac.uk/292026/

• Easterlin, R. (2012). Life satisfaction of rich and poor under socialism and capitalism. International Journal of Happiness and Development, 1(1), 112–126.

• Hoyer, D., Holder, S., Bennett, J. S., François, P., Whitehouse, H., Covey, A., Feinman, G., Korotayev, A., Vustiuzhanin, V., Preiser-Kapeller, J., Bard, K., Levine, J., Reddish, J., Orlandi, G., Ainsworth, R., & Turchin, P. (2024). All Crises are Unhappy in their Own Way: The role of societal instability in shaping the past. OSF. https://doi.org/10.31235/osf.io/rk4gd

• Kanbur, R. (2019). Inequality in a global perspective. Oxford Review of Economic Policy, 35(3), 431–444. https://doi.org/10.1093/oxrep/grz010

• Marienbach, M. (2024). Egalitarianism Without Catastrophe? Teotihuacan’s Challenge to Walter Scheidel’s Four Horsemen Theory. The Maastricht Journal of Liberal Arts, 15. https://doi.org/10.26481/mjla.2024.v15.1017

• Schmidt, A. T., & Juijn, D. (2024). Economic inequality and the long-term future. Politics, Philosophy & Economics, 23(1), 67–99. https://doi.org/10.1177/1470594X231178502

• Sullivan, D., & Hickel, J. (2023). Capitalism and extreme poverty: A global analysis of real wages, human height, and mortality since the long 16th century. World Development, 161, 106026. https://doi.org/10.1016/j.worlddev.2022.106026

• Vodovnik, Ž. (2024). Possibilities at the End of the World: (Re)Imagining Care as Politics. Teorija in praksa, 387–405. https://doi.org/10.51936/tip.61.2.387