Hide table of contents


This post summarizes some thoughts (mostly possible research questions) around the possible collapse, balkanization, or radical transformation of the United States (collectively, "Collapse").

To be clear, I do not think that Collapse is either imminent or likely in the foreseeable future. Nor do I even think that Collapse is probably competitive with other top causes. However, I do think that:

  1. p(Collapse within 50 years) is non-trivial (>0.5%).
  2. The disutility from Collapse could be extreme in certain scenarios.

As EAs, we care about expected utility, so Collapse may well be worth worrying about despite being unlikely. Thus, I think it may be worth some EAs exploring Collapse as a possible cause area.

The thoughts here are extremely rough. I normally like to publish stuff that is more polished than this, but I wanted to get this out there and don't want to invest more in polishing it. I hope others will critique or build on these thoughts.

Note that I had been thinking about this subject from an EA perspective the past few weeks, though Wednesday's riot at the Capitol certainly made me think about this more seriously.

Meta: Estimating the Probability of Collapse and Specific Scenarios

Here are some rough thoughts on sources of information we could use to estimate the probability of different Collapse scenarios and effects. Overall, studying the expected value of Collapse more seems quite valuable to me.

Evidence from Markets

US equities are at all-time highs, and the US's bond rating is good. The US is currently able to sell bonds at negative real interest rates. Thus, global financial markets do not seem worried about Collapse, or at least not a Collapse that would imperil the US financial system. (One could imagine, however, an oligarchic government in which US business functioned well but the public was seriously harmed.)


Are there any superforecasting questions about Collapse directly? Which questions would indirectly inform p(Collapse)?

Public Sentiment

A number of polls ([1], [2]) have found that a significant chunk of the US believes we are headed towards civil war. Even if one is skeptical of the public's ability to accurately predict such events (as I am), the perceptionof an impending civil war could itself be escalatory and therefore self-fulfilling. However, most people seem to be living quite normal lives (and not, e.g., preparing to flee the country), so people's revealed beliefs may be quite different and less dire.

Academic Disciplines

Disciplines like history and political science could presumably tell us a lot about the probability, indicia, leading indicators, and severity of Collapse scenarios.

Collapse Scenarios

How might a Collapse play out? What might its morally relevant effects be?

Civil War Scenario

The most dramatic, but probably least likely, outcome is a civil war.

The US has a population of 328 million. Presumably a civil war in the US would lead to significant loss of life in the US, as well as significant decline in the standards of living and economic activity for its residents.

Other Tumultuous Scenarios

Other tumultuous Collapse scenarios might include:

  1. Bloodless or limited-violence coup d'état
  2. Invasion by a foreign power (including following periods of domestic unrest or after long-term military and economic decline)
  3. Escalation of domestic terrorism and civic unrest
  4. Secession attempts with ambiguous outcomes
  5. Unlawful expansion of federal government powers
  6. Cessation of democratic rule in the US

Peaceful Scenarios

Collapse need not be violent or tumultuous. For example, there could be a legal agreement to split the country into different independent countries. Although difficult to imagine, the US could also enter into a treaty with an independent country that would integrate the two, fundamentally altering each.

Indirect Effects

Besides the direct effects of tumultuous collapse scenarios (such as loss of life and economic harms from armed conflict), what effects of Collapse might worry EAs?

Nuclear Security

A glaring problem during Collapse would be nuclear security. Use of nuclear weapons during a Collapse conflict could be possible. In addition, security of the country's nuclear arsenal during any transitional periods would become extremely important.

How well prepared is our nuclear arsenal for domestic unrest? What can we do now to ensure the US nuclear arsenal is secure during Collapse?

Science Research Funding

The US leads the world in R&D spending. Disruption from collapse could seriously diminish this spending and therefore key advances in biomedicine and other beneficial technologies.


A number of top AI labs are owned or operated in the US. Collapse could affect these labs by destroying them or their value or destabilizing the environment in which they operate. Collapse could lead to a post-Collapse legal regime in which AI labs were more susceptible to coercion or expropriation. Collapse could make development of AGI or TAI in non-US countries more likely. Seizure of key AI assets or facilities during conflict also seems possible.

In the more distant future, Collapse could provoke militaries into using unsafe AI in conflict situations.

Liberal Democracy and the International Balance of Power

Collapse could mean that the US ceases to be a great power. This may enhance the relative power of the EU, China, and Russia. The implications of this for global security and the continued expansion of liberal democracy are unclear but seem negative. Even more so if the post-Collapse balance of power influences which countries lead in AI and space exploration/settlement.

Humanitarian Aid

The US government the largest contributor to international humanitarian aid. Collapse could plausibly result in tens of billions of dollars in reduced international AID.

Flow-Through Economic Effects

In addition to the direct economic effects of Collapse (i.e., on Americans), Collapse could reduce demand for foreign labor and thereby reduce income among the global poor.

Critical Global Infrastructure

What effect could highly disruptive Collapse scenarios have on global infrastructure, such as the Internet or global financial markets? Could this be significant globally?

Possible EA Interventions

If, after much more investigation, we conclude that Collapse is worth taking seriously, what actions might EAs consider taking?

Movement Infrastructure

One key sub-cause would be minimizing the degree to which EA suffers as a result of Collapse.

Geographic Diversification

One way to hedge against Collapse could be to invest more in building the EA movement outside the US, to ensure that it survives Collapse. However, the size of the movement in the UK alone is probably sufficient to ensure this already.

Contingency Planning

Large EA organizations based in the US may want to plan to switch operations to non-US countries if needed, though immigration issues could make this difficult. Where feasible, setting up foreign branches could be wise.

Financial Planning

Similarly, large organizations could consider hedging against Collapse by (partially?) switching to non-US financial institutions, investing in non-US assets, and keeping reserves in multiple currencies.

Preventing Collapse

Prevention is often more effective than amelioration. How might EAs reduce the probability of Collapse?

Voting Reform

The Center for Election Science is working to implement Approval Voting. Could this help elect unifying candidates? If so, would that reduce p(Collapse)?

Could other possible reforms—such as the National Popular Vote Compact, HR 1, or switching to a parliamentary system—help?

Constitutional Amendments

Which plausible Amendments to the US Constitution could stabilize the United States?


Could fighting misinformation help reduce domestic sources of violence and instability?

Domestic Law Enforcement

Is it feasible and desirable to expand funding for law enforcement specifically targeted at domestic sources of instability?

Gun Control

Is gun control (to reduce the probability and severity of domestic violence) feasible and desirable?

Ameliorating Collapse

How can we make Collapse less bad if it does happen? How can we prevent or mitigate the possible bad effects described above?


How can we make a post-Collapse order in the former US as good as possible? What institutions should we build post-Collapse?


Sorted by Click to highlight new comments since:

I certainly agree that this is worth thinking about, but I also think it's worth suggesting that the analysis here is be a bit myopic. Of course, it seems particularly relevant because many EAs are in the US. And it seems unconceivable that the world will change drastically in this one particular way, but far larger plausible changes are on the horizon.  (Though as I've noted n various conversations for a while, Americans might want to personally consider their options for where else they might want to live if the US decline continues.) 

And if the worst case happens, we're still likely looking at a decades-long process, during which most of the worst effects are mitigated by other countries taking up the slack, and pushing for the US's decline to be minimally disruptive to the world. Nations and empires have collapsed before, and in many cases it was was bad, even very bad. (Though in other cases, like the dissolution of the British empire, there were compensating changes, like the rise of the US and the far more egalitarian and peaceful post-WWII order.) So preventing a bad collapse is plausibly as important a cause as preventing another pandemic like COVID-19 - albeit far less certain to occur, and far less certain to be bad for the world. And it's not of the same order of magnitude of many other longtermist causes, since it's highly likely that conditional on the unlikely case of severe collapse in the US, humanity will be fine.

All that said, again, I don't disagree with the analysis overall - this is worth taking seriously.

Thanks David. Great analysis as usual :-)

I'm not actually sure we disagree on anything. I agree that

if the worst case happens, we're still likely looking at a decades-long process, during which most of the worst effects are mitigated by other countries taking up the slack, and pushing for the US's decline to be minimally disruptive to the world.

I definitely also agree that it behooves EAs to try to avoid myopia. I have tried to do so here but may very well have failed!

In terms of expected disvalue, I would guess that severe and rapid collapses (more like the USSR than France or Spain) are the most important, due to the nuclear insecurity and possible triggering of great-power conflict.

As for cost-competitiveness with other longtermist interventions, it seems that increasing nuclear security from domestic instability is actually pretty tractable and may be neglected. If so, that suggests to me that it may be approximately as cost-effective as most marginal nuclear security work generally. The only other things that seem plausibly cost-effective to me now are contingency planning for key longtermist institutions so that their operations are minimally disrupted by a turbulent decline.

I agree that we agree ;)

I particularly endorse the claim about tractability and effectiveness of technical changes to internal nuclear weapon security  and contingency planning, both with moderate confidence.

Re: probabilities, I've been working on a search engine for probabilities as a small side project, and some of the ones I could find are:

Will the USA's Labor Force Participation Rate be lower in 2023 than in 2018?: 70%

Second US civil war before July 2021?: 1% (Metaculus doesn't allow lower probabilities)

Before 1 January 2022, will the U.S. Senate expand the scope of matters for which a filibuster cannot be used?: 3%

SCOTUS impeachment before 2030: 7%

Will another 9/11 on U.S. soil be prevented at least through 2030?: 75%

Longbets series: By 2040 will the percentage of college-aged U.S. citizens who are attending postsecondary educational institutions in the United States drop at least 50% from the level in 2011?: 75%

Date USA Metaculites face emigration crisis: 15% before Dec 25, 2030

Will US life expectancy at birth for both sexes fall below 75 years before 2040? : 13%

Will at least one US state secede from the Union before 31 December, 2030?: 5%

Coup-cast: Probability of a coup in the US in 2021: 0.08%. (at 0.08% per year, the probability in 50 years would be ~4%.)


  • Most of these are from Metaculus, which isn't surprising given the sheer volume of questions it has and its willingness to include long term and somewhat weird questions.
  • A >0.5% per 50 years sounds like a reasonable probability to me.
  • I imagine my probabilities would depend on the specific definition of "collapse". If something like the fall of the Soviet Union would also count as a "collapse" then I could imagine going up to a couple of percentage points.
  • having a question on Metaculus about this seems like a cheap win.
  • Added the cause candidate tag to this.

Very helpful. Thanks!

I was surprised you did not mention nuclear war as a cause of the decline of the US. If you take Luisa Rodriguez's average estimate of US-Russia nuclear war, 0.38% per year, that's about 20% chance in 50 years. And that does not take into account possible US-China nuclear war. I think even if nuclear winter did not happen, just the war would cause a significant decline in the US. So would that meet your definition?

Thanks David! I guess I was implicitly thinking of scenarios where the decline of the US was not caused by a GCR, since such cases would already qualify for EA prioritization. But agree that decline of the US due to a GCR would meet my stated definition of Collapse.

[Epistemic status: Uncertain, and also not American, so this is a 3rd party perspective]

As for the likelihood of some form of collapse, to me the current trajectory of polarization in the US seems unsustainable. Nowadays, members of both parties are split about whether they consider members of the other party "a threat to their way of life"(!)  and feelings towards the other party are rapidly declining.  

I do not think that this is just a fluke, as many political scientists argue that this is driven by an ideological sorting and a creation of a "mega-identity", where race, education and political leanings now all align with each other. Political debate seems overwhelmingly likely to get more acrimonious when disagreement is not just about facts, but about your whole identity, and when you consider the other side to be your enemy.  

It is only a slight overstatement to say that members of both parties live in two very different realities. There is almost no overlap  in the trusted news organizations  and the unprecedentedly constant approval rating of Donald Trump indicates that neither side changed their mind much in response to new information coming in.   

On the up side, "67% comprise 'the Exhausted Majority', whose members share a sense of fatigue with our polarized national conversation, a willingness to be flexible in their political viewpoints, and a lack of voice in the national conversation.” My worry is that this majority is increasingly drowned out by the radical voices in traditional and social media. 

It is also pertinent that political collapse can happen very fast and without much warning, like the Arab Spring and the collapse of the Soviet Union showed, which came unexpected to observers. Decline can also take the form of persistent riots/unrest where no one party has the political capital/strength to reach an agreement with the rioters or to stop it. Consequently, if decline of the US seems likely and bad, I would worry about it possibly happening quickly (<10 years).

Figuring out how to move politics towards the exhausted majority seems interesting. They probably care about stability a lot more than hyper-partisans do.

Peaceful Scenarios

Collapse need not be violent or tumultuous. For example, there could be a legal agreement to split the country into different independent countries. Although difficult to imagine, the US could also enter into a treaty with an independent country that would integrate the two, fundamentally altering each.

These seem like quite different scenarios to the others discussed. If the US agreed to let California become independent, or annexed Canada, I would not expect any threat to nuclear security, or AI lab integrity, or drastic loss of life in the process. Annexing Canada could even potentially help continue US international hegemony through increased population and GDP, though it might be bad in other ways.

Yeah, they are definitely quite different, and probably less important from an EA perspective. I just included them for completeness because of the definition of "Collapse" I gave.

I am aware of two (short-term) questions related to civil war scenarios on Metaculus:

The issue with these interventions suggested for preventing collapse is that they generally have much more pressing impacts besides this. For instance, of course approval voting is great, but its impacts on other political issues (both ordinary political problems, and other tail scenarios like dictatorship) are much more significant. More generally, stuff that makes America politically healthier reduces the probability that it will collapse, and the converse is almost always true. So not only is the collapse possibility relatively unimportant, it's mostly unnecessary baggage to carry in your cognitive model.

As for movement infrastructure, a similar logic probably applies as EA organizations have many other priorities with these things.

not only is the collapse possibility relatively unimportant,

I'm not sure this is true, though, if I'm right that certain Collapse scenarios are GCRs, because they could cause or entail great-power war or nuclear war.

Sorry for a disorganised comment, but here's a recent summary I've wrote about why most people, including academics, are wrong about polarisation (ping me for references and details):
There’s been a lot of high quality research on polarisation since at least the 80s, spearheaded by Keith Poole and Howard Rosenthal. Around the time of Trump’s election, the subject became extremely popular and a lot of people (including me) decided to research this issue. Unfortunately, many of those scholars (including political scientists unfamiliar with polarisation research) believe things that seem to make sense intuitively but that contradict the large body of pre-existing research on polarisation. 
Findings in the area have established that many seemingly likely causes of polarisation are not causes of polarisation; including polarisation of the electorate, party leadership, polarisation over social issues, and gerrymandering. It seems most people (including academics) can get away with claiming these were the main causes, it seems to be a very common belief (I thought it myself until I fact-checked this); alas, it’s well-established it’s just not so. The only mildly well-established likely cause is economic inequality. It is the only trend that follows the pattern of raising polarisation; but it’s an open question whether there’s an actual causal link from inequality to polarisation. Another plausible contender is polarisation of partisans (people heavily involved in politics), but experts seem split over this issue. It is my impression that the latest research favours the view that this was not a relevant cause and that party sorting better explains the data used in favour of partisan polarisation.
The NOMINATE-type measures are the most frequently cited basic data supporting the view that Republicans are now more extremists than they ever wore in the last century. But it does not really directly support this view on its own. NOMINATE-type measures are strictly relational, that is, they measure how far things are from each other, not how far they are from an absolute point such as the centre. Therefore, the data is just as compatible with Republicans moving away from the centre as it is with the whole thing (Republicans and Democrats) moving towards conservatism as part of a long-term macro-trend that changed the US Party System. Nolan McCarty (Poole’s most frequent recent co-author) agreed this is the case by email, but that he believes Republicans drove the trend due to contextual reasons (e.g. Gingrichism, I assume).

Fascinating read and potential concept, especially given what we have witnessed in the US over the past few months. I am interested in the framing of this piece around 'collapse' and more specifically how 'collapse' may be differentiated from a more general reduction in relative power? Is there something specific about a 'collapse' that differentiates it from the standard tectonic shifts of power that we have observed over the course of all human history that makes prioritization more important? 

In my opinion, the framing of a cause prioritization around 'collapse' creates risks of a false dichotomy, whereas more general prioritization on something like say institutional decision-making, policy-making or just more generally politics could provide a vast majority of the potential benefits that a cause priority focus on 'collapse' could bring. Simultaneously, these more general areas would likely provide more relevant insights with a higher likelihood of implementable positive social impact. 

I think the evidence from the financial markets is a bit weaker.

First, let's imagine predicting that the forecasting platform will stop operating and assume that forecasting is only incentivized by points on this platform. The reasonable prediction is that platform will continue to operate because otherwise, points will become meaningless. Same about predicting existential risk (because if it occurs, one won't be able to claim a prize).

The US collapse will be devastating for the financial markets (plausible to me unless the USA will gradually lose power and importance, in which case interventions are less crucial). The incentives assumption seems plausible to me as well. So the market might not be a reliable predictor of it.

I am pretty confident that's wrong. The disanalogy is that with financial markets, you can presently withdraw money and move it to safer assets or spend it on present consumption.

More from Cullen
Curated and popular this week
Relevant opportunities