The Metaculus democracy threat index

Astral Codex Ten
Subscribe Sign in
The Metaculus Democracy Threat Index
...

Jun 25, 2026
117
138
2
Share
In recent posts on Trumpand dictatorship, people have asked me - how do you know you’re not suffering from Trump Derangement Syndrome?
I take this seriously; we’ve all lost loved ones to this condition. The best check on my reasoning would be an objective measure of the health of American democracy. There are several “democracy indices” that purport to do this, but they have a mixed reputation. My impression is that most _current_ accusations of bias are relatively weak - I agree with Claude’s analysis here - but they rely enough on “expert” opinion that I don’t expect them to convince a skeptic.
The newest entrant in this space - **Metaculus Democracy Threat Index** - works differently, and deserves a closer look.
Metaculus is a prediction site - like a prediction market, except that no money changes hands. People can record their guesses for how future events will turn out, which get aggregated by an algorithm (currently just a recency-weighted median, although they’ve done fanicer things in the past).
Their Democracy Threat Index is a collection of 153 questions relevant to US democracy. For example:

This question says there’s a 3.5% chance that a political party will keep an opponent off the ballot in a state election in 2026. You can see that 86 people have made forecasts about this. When probabilities like this go up on the 153 questions, the index gets higher.
This has some advantages over traditional democracy index design. It’s transparent: you can see all the questions and how they fit together. It’s crowdsourced, so there’s limited opportunity for ideologues or biased experts to put their fingers on the scale. It does a good job limiting itself to things which naturally seem democracy-related, resisting the pressure to add “and they support my preferred policy” to the definition of democracy.
But the biggest change is that Metaculus leverages modern forecasting science (popularly called “superforecasting” - although I think the people with that trademark would be unhappy to hear me use it this way). Does this help?
We care most about questions like “Did they cancel elections?” or “Did they murder protesters?” But these are coarse binary outcomes - when someone wants to know if democracy is “under threat”, they want to know when there’s _increased chance_ of these things happening, even though they haven’t happened yet. As a forecasting platform, Metaculus can distinguish not just between “keeps elections” vs. “cancels elections”, but between 10% chance of cancelling the next elections vs. 50% chance. This both lets them focus on the big questions (rather than the small questions that are most likely to differ between one mildly-concerning regime and another) and give finer-grained estimates (a 10% vs. 50% chance of elections getting cancelled, rather than just ‘not cancelled yet’).
What are the remaining risks/biases?
- **Susceptibility to crowd attack:** If a hundred Democrats join Metaculus and give maximally pessimistic answers to all democracy-related questions, would that make Trump look worse? Currently Metaculus has near-perfect security through obscurity - nobody cares about this index enough to attack it. I asked them what happened if that changed, and they said they had good security, and that medians are naturally more secure than means to this kind of threat. I’m still disappointed to have to rely on the security of one centralized site.
These problems could be solved by transitioning from Metaculus to a true prediction market. Prediction markets’ main advantage over Metaculus-style forecasting engines is resilience to attack (because new bettors are incentivized to come take the attacker’s money). Unfortunately right now Metaculus is focusing on being a responsible academic institution, and prediction markets are focusing on attracting sports-gambling-obsessed degens, so we’ll have to settle for the former.
That still leaves one possible source of risk:
- **Susceptibility to question selector bias:** Who decides which questions get added to the Index? Right now, it’s a “nonpartisan group” called Bright Line Watch. Are they _really_ nonpartisan? You can see some discussion here, but even if we trust them, it’s disappointing that the question of trust has to come up at all, given the otherwise-trustless design.
I thought about this most when seeing the question about whether government departments will change policy in response to bribes, which mentions as a possible bribe vector an “expenditure to a cryptocurrency” - for example, some high official creates a crypto token, and the would-be-bribe-giver buys into it. Democrats and Republicans are corrupt in different ways, and privileging memecoin buyers is so far a uniquely Republican form of corruption. If they’d missed that vector, the bribery question would have been biased toward finding corruption in Democrats (who so far haven’t use memecoins as much). But how chould we prove that they didn’t _include_ that vector specifically because they searched very hard for Republican-coded forms of bribery and made sure to include them, while thinking less hard about Democrat-coded forms? Are they allowed to use their judgment and say that there seem to be more things like this on one side of the aisle than the other?
Maybe it would be safer to stick to really obvious questions like “will they cancel elections?” But even this might be subjective - maybe Republicans are more likely to cancel elections, but Democrats are more likely to censor speech. Is censoring speech as much of a “threat to democracy” as canceling elections? Hard to say.
As far as I can tell, questions like these make it hard to conclude anything from the index’s “historical backfilling”:

That is: the index was started in 2025. The 2021-2022 and 2023-2024 data points come from retrospectively checking how many of the questions they invented in 2025 had already happened by 2021-2022 etc. There are many reasons it could be fewer besides declining democracy. For example, it might have been boring to start with questions that had already happened. Conversely, it could be especially exciting to consider new (as of 2025) threats to democracy like memecoin corruption.
On the other hand, as far as I can tell the third and fourth data points - the squares - are pretty believable. These represent forecasters’ opinions when the index was first formed, and their forecasts about what will be true two years later. They seem to expect that democracy as measured by these questions won’t get any worse over the remainder of Trump’s term - which I find reassuring.
Here’s another interesting graph:

The expected index level for 2027-2028 over time. We see that concerns about democracy peaked in November/December 2025, when the expected value of the index was 47%. Then they started declining in mid-December, and are currently stable at 39%. What changed? Maybe the results of various November special elections (both the elections going without a hitch, and Trump’s preferred candidates losing), or the failure of a grand jury to indict Trump opponent Letitia James in a trial widely perceived as politicized. I respect this graph for telling me a non-obvious thing (the level of predicted threat to democracy decreased in December 2025) that makes some sense in retrospect.
As time goes on, the Index will become more valuable insofar as it gets further from any possibility of biasing the questions. That is: any biases in the question set are the biases of 2025, when (for example) Republicans are more into memecoin-based corruption than Democrats. But as time goes on, the biases will drift - the biases of 2035 will be correlated <1 with the biases of 2025 (for example, Democrats might take up memecoin-based corruption, or Republicans might drop it) and as long as the question set is stable, biased question-selectors can’t optimize it for the most up-to-date biases.
Suppose that the 2028 election is Newsom vs. Vance, and that each party claims the other’s nominee will destroy democracy. We could have a conditional prediction market - what will the Democracy Index be in 2032 conditional on Vance/Newsom getting elected? Or we could just watch how the Democracy Index responds to shocks in election outcome percentages: for example, if Newsom has a bad debate and his chances drop 5% on Polymarket, does that drive the Democracy Index up or down?
What would we need to make this really work? My wish list looks like:
- Duplicates of some of these questions on Polymarket and Kalshi, as a backup against attacks on Metaculus.
- More thought put into the questions, including a policy on how new questions will (or won’t) be added.
- More forecasters, and existing forecasters adjusting their guesses more often. There’s now $10,000 in prize money on the line, so consider participating!
- * *
Subscribe to Astral Codex Ten
By Scott Alexander
P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary.
Subscribe
By subscribing, you agree Substack's Terms of Use, and acknowledge its Information Collection Notice and Privacy Policy.

117
138
2
Share
Previous
138 Comments

[14h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282842043 "Jun 26, 2026, 12:10 AM")
Managed Democracy calls for aid. Only you can answer, Helldiver.
IYKYK

[14h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282846403 "Jun 26, 2026, 12:21 AM")
I can only hope this eventually adds a "Will a politician attempt to change a prediction market threat to democracy index" question, both because that would be the best possible signal of its success and because it would be hilariously hard to interpret the resulting number.

[14h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282848026 "Jun 26, 2026, 12:25 AM")Edited
Formally removing another party's candidates from the ballot. Interesting criterion.
What if instead a party were to, I don't know, allow large numbers of noncitizens into the country and let them vote? Does Metaculus have a question on that? Or would that just not count as a "threat to democracy" because "voting good?" Just what is "democracy" anyway?

[13h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282851600 "Jun 26, 2026, 12:34 AM")
Also they missed "attempting to remove a candidate from the ballot", a milestone which has already been hit.

[13h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282867297 "Jun 26, 2026, 1:15 AM")
This gets weird - because of course someone who stands to gain from it will make the *attempt* if it is legitimate to do so.
IE: If your opponent sexually assaults a child on 5th avenue, and you try to indict them for it I think that's on your opponent, not you.
However, a big part of the game of that thing we call "politics" is "reframing your actions to make it look like you're not doing that obviously underhanded thing" , so what would be a framing that could detect this ?
I don't think people are concerned with "prosecuting the political opposition" per se they are concerned with, for example:
- Handling Hillary Clinton with kid gloves on the security breach
- The FBI committing crimes to wiretap Trump
- As already mentioned, Letitia James
"Not conceding the election" is another pretty wild one, because we have:
- Ran a scheme to send alternative electors to the certification to vote for him (Trump)
- Engaged in a 2 year long criminal process on the basis of fabricated information to impeach the President on the basis of a fraudulent election (House Democrats)
What "counts" is most of the question, and there's tons of room for "whataboutism".

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282900725 "Jun 26, 2026, 2:46 AM")
That assumes prosecution automatically removes a candidate from the ballot, which doesn't necessarily follow. I'd think that the removal of the candidate is a more straightforward measure of threat to democracy than the prosecution (which is obviously still impactful).

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282909132 "Jun 26, 2026, 3:11 AM")
Eugene Debs was on the ballot while in prison. I had mused that would be the most entertaining outcome for Trump in the most recent election, particularly if SCOTUS ruled he couldn't pardon himself and he couldn't trust his VP to pardon him if he resigned, so he spent his whole term governing from behind bars.

[9h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282938087 "Jun 26, 2026, 4:50 AM")
But the way the questions are phrased, though, it's not about who is right or wrong, it's about whether there's a threat to democracy.
If someone refuses to concede an election then either (a) they're incorrectly claiming an election was rigged, which is bad for democracy, or (b) the election really was rigged, which is also bad for democracy.
In a healthy democracy you could be reasonably confident that neither of these things would happen, because the voting system would be secure and everyone would have high confidence that significant fraud isn't possible.

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283056255 "Jun 26, 2026, 11:52 AM")
"If someone refuses to concede an election then either (a) they're incorrectly claiming an election was rigged, which is bad for democracy, or (b) the election really was rigged, which is also bad for democracy."
Then we've already hit that milestone, because of course everyone knows about Trump and the 2020 election, but there were also those claiming the 2024 election was rigged to let Trump win.
So we've got partisans of both parties claiming illegal/rigged elections, and even back in 2016 the whole Not My President crowd where some nutjobs on the Internet liked to claim Hillary was the President Elect and that was it, Trump was not president.
Didn't Stacy Abrams also indulge in some "my opponent only won because the election was dodgy" claims?
So if democracy is imperilled where there are claims of election rigging, then that has already happened and democracy is indeed imperilled.

[31m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283101342 "Jun 26, 2026, 1:30 PM")
See for example Turkey. Turkish Constitution requires the President to have a university diploma. An opponent of Erdogan, Imamoglu, was stripped of his diploma by his alma mater, thus rendering him ineligible to stand in the election.
Now this wasn't technically done by the government, but by a university. But it still counts IMHO, as the university was obviously pushed to do this via Erdogan's people.
This also illustrates quite nicely how hard is it to protect democracy via laws alone.

[13h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282864623 "Jun 26, 2026, 1:08 AM")
That there is no “Will an opposing party authority convict a presidential candidate of a crime that isn’t normally prosecuted?” makes me worried this is too sensitive to question choice to be a good early warning system.

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283056929 "Jun 26, 2026, 11:54 AM")
Convict or accuse? We've danced this dance already, I think, if we're going for accusations of naughty-naughty.

[12h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282883025 "Jun 26, 2026, 1:56 AM")
Noncitizens voting in federal elections is illegal, so that question reduces down to "will there be widespread voter fraud in a future election?"
Metaculus has this question at 0.9% chance, and most of that probability is the nitpick that the Trump administration will probably *claim* that this happened regardless of how much fraud there actually is.
https://www.metaculus.com/questions/41883/widespread-non-citizen-voting-occur-in-2026-us-midterms/

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282898011 "Jun 26, 2026, 2:38 AM")Edited
There are locales trying to expand voting rights to noncitizens (e.g. https://www.latimes.com/california/story/2026-06-17/measure-to-give-non-citizens-right-to-vote-in-la-elections-headed-for-nov-3-ballot), so it's not just fraud. Whether this is a threat or improvement to democracy is up for debate.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282920717 "Jun 26, 2026, 3:48 AM")
To make clear, the measure would expand voting rights to non-citizens in citywide elections. Non-citizens living in LA would be able to vote for the mayor of LA, not president. The parent comment mentioned voting in federal elections, which is illegal for non-citizens and will remain illegal for the foreseeable future.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282926955 "Jun 26, 2026, 4:11 AM")
Right, hence the mention of specific locales. But it does show that the left is politically amenable to letting noncitizens vote, so I disagree that it's definitely out of the question for the foreseeable future, especially when the point is forecasting the future.
And appeals to absurdity or legal unlikeliness are out the window when we have leftist political candidates (soon-to-be congresswoman, as she just won the Dem primary in New York) not only arguing that illegal immigrants who have committed murder and other serious crimes should be protected from deportation, but also that prisons should be abolished entirely (https://nyeditorialboard.substack.com/p/darializa-avila-chevalier-on-housing).

[6h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282980929 "Jun 26, 2026, 7:44 AM")
Right, hence the mention of specific locales
Sure, just the original comment could be misconstrued & wanted to avoid any confusion.
Re: fringe congressmen, it's not fair to define an entire party based on the most extreme people sent to congress. Republicans in Georgia sent Marjorie Taylor Greene to Congress 3 times, and she believes that California wildfires were started by a Rothschild-owned space laser. I imagine most republican-leaning districts aren't on board with that, but in a non-competitive district you can easily get wackos.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283014439 "Jun 26, 2026, 9:52 AM")
Oh I would certainly not define the entire party based on the DSA types, but I also give them more a chance to get their wacko policy agendas into law because they're more of an organized bloc than right wing fringe loonies like MTG, who as far as I know is an isolated case. They've been pretty effective at coopting the Dem voter base in deep blue areas to get more DSA types into power, so I could definitely see a trajectory where they form a significant enough caucus within the Dem party to be able to inject something like noncitizen voting into the Dem platform, and eventually make it into actual legislation when Dems are in power again.

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283057457 "Jun 26, 2026, 11:55 AM")
I could imagine someone being honestly confused why they can vote in this election but not that election, or assuming that if they can now vote in these elections then that means they can vote in that election.
Needn't be deliberate fraud.

[30m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283101920 "Jun 26, 2026, 1:31 PM")Edited
Two observations:
- I believe this is not a "US federal elections democracy index", it's a "US democracy index". City elections becoming corrupted when federal elections aren't is still a degraded democracy
- various people might disagree about the answer to "will illegals be allowed to vote in federal elections in the foreseeable future?". Considering we are talking about prediction markets here it seems entirely fair to have a question like this in a prediction market about democracy. If predictors all agree that it's certainly unlikely and continues to be unlikely under a new democratic president, that's a useful signal.
(Which gets back to the fact that the central question here is whether noncitizen voting is a threat or improvement to democracy, not whether it's likely or the scale of it. Whether it's federal or local only matters for the scope of the index, whether it's likely or unlikely would be the literal result of the market itself so "it's unlikely" is not a counterargument, just a single prediction by a single individual)

[5h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283002845 "Jun 26, 2026, 9:10 AM")Edited
Worth mentioning that noncitizen voting rights already exist, or existed in the past, in many jurisdictions commonly regarded as "democratic".
https://en.wikipedia.org/wiki/Non-citizen_suffrage

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283033897 "Jun 26, 2026, 10:54 AM")
I find it questionable to argue that making more people vote would be something that should be classified as a threat to *democracy*. You can think it's a short sighted ploy by a specific party to buff its numbers, but the same tends to be true of many changes in electoral law, districting, etc. Sometimes a change is good for democracy in the long run but still will only get enacted when, situationally, it's also good in the short term for the prospects of a specific party. That's not amazing but it doesn't destroy democracy because in the long run the system stays while things like "this demographic votes for me" change and dissolve.
[](https://substack.com/profile/6053602-stephen-pimentel?utm_source=comment)
[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282916653 "Jun 26, 2026, 3:35 AM")
_User was indefinitely suspended for this comment. [Show](javascript:void(0))_

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282922730 "Jun 26, 2026, 3:55 AM")
Author
Banned for this comment - obvious attempts to pick fights without any useful contribution.

[9h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282944336 "Jun 26, 2026, 5:14 AM")
Can anyone explain what actually prevents non-US-citizens from registering to vote?
You'd think it would be pretty easy to force people to provide proof of citizenship when registering; either a birth certificate or a certificate of naturalisation would work.
But according to this https://www.migrationpolicy.org/content/noncitizen-voting-us-elections the main security is "the form requires applicants to affirm, under penalty of perjury, that they are citizens" ... is there really no actual check?

[6h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282984368 "Jun 26, 2026, 7:58 AM")Edited
Here's an example, for California: https://elections.cdn.sos.ca.gov/pdfs/voter-id-and-reg-requirements.pdf
If you have a social security number or California state ID number you have to include it and they will check against the database. You can also register affirming under penalty of perjury that you are a citizen but have neither ID number, as you have said. (EDIT: also note that if you register online or by mail without an ID number, you'll need to show an ID the first time you vote, but it doesn't have to be a photo ID and doesn't apply if you register in person.)
I would disagree that penalty of perjury is "no actual check". It seems like it would be risky to commit fraud on a government form if you're already in the country illegally. Also if millions of people were suddenly registering without ID numbers someone in the bureaucracy would probably notice and whistleblow.
But I think the goal is that a tiny chance of fraud is an ok tradeoff for making it a lot easier for actual citizens to vote (in the vein of "let 10 guilty men go free rather than convict one innocent man", but with a better ratio). Personally, I'd support sidestepping the issue by having a real national ID system that's free and automatic, though you'd still need to worry a bit about fraud/accessibility in the ID system itself.

[5h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282996643 "Jun 26, 2026, 8:47 AM")
You know you can live in the country legally without being a citizen right?

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283060987 "Jun 26, 2026, 12:03 PM")
I do, in fact. What part of the comment is this relevant to?

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283061554 "Jun 26, 2026, 12:05 PM")
I would disagree that penalty of perjury is "no actual check". It seems like it would be risky to commit fraud on a government form if you're already in the country illegally.

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283066591 "Jun 26, 2026, 12:16 PM")
Got it, thanks. I also think it would be risky to commit fraud on a government form if you are in the a government database that can be easily double-checked by an administration that cares to.

[3m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283115856 "Jun 26, 2026, 1:57 PM")
Yeah, isn't this basically just the reason there isn't stricter ID checks on this? One side seems to think there's widespread fraud among non-citizens, the other side seems to think that's unthinkable for the reason you give; but I guess that's just what living in an extremely polarized world looks like.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283018458 "Jun 26, 2026, 10:06 AM")
It seems like it would be risky to commit fraud on a government form if you're already in the country illegally.
On the other hand, people who are in the country illegal are already selected for being willing to take the risk of breaking the law. I don't think there's really a large number of people doing it since the benefit is not as great as being able to live in the US, but I also think it's silly that the CA government is so against checking that they legislated *disallowing* requiring voter IDs https://legislation.ballotpedia.org/elections/bill/10007

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283070027 "Jun 26, 2026, 12:24 PM")
Overall I think it's good to keep voting rules consistent across municipalities, and I prefer lighter ID laws *in the absence of a universal ID program*, so I can't say I disagree with CA's decision. But I don't think your perspective is unreasonable.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283022288 "Jun 26, 2026, 10:18 AM")
"widespread" does a lot of work there.

[30m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283101736 "Jun 26, 2026, 1:30 PM")
You also have to be able to *detect* voter fraud, and afaict there is no actual way to detect voter fraud in California.
Also their verification of citizenship for voting registration is: do you say you're a citizen? good, they don't verify anything else. they don't even verify if you live at the address on your form.
Please note that I am not claiming there is widespread voter fraud. I am claiming *the system provides no way of knowing if there is or isn't*.

[12h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282883330 "Jun 26, 2026, 1:57 AM")
If a party did this, I'm not sure "threat to democracy" would be the right term, but it would definitely disrupt the established democratic system in a bad way. Fortunately, no party has even attempted to do this; given the number of readily available threats to democracy, it seems mostly unhelpful to come up with unlikely hypotheticals.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282898691 "Jun 26, 2026, 2:40 AM")Edited
Who knows if it'll ever be attempted at the state or federal level, but the left seems amenable to letting noncitizens vote.

[14h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282848951 "Jun 26, 2026, 12:28 AM")
I’m finding the idea that Democrats and Republicans corrupt in different ways a little ridiculous. Oh if you exclude the meme coins and oh if you exclude the staggering corruption from trump and his cronies, all the carveouts to business interests, not to mention the freaking Supreme Court (Harlan crow anyone?), then yeah democrats could be more corrupt. And the idea that democrats would restrict speech more than republicans is hilarious after Jimmy Kimmel.

[13h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282858195 "Jun 26, 2026, 12:51 AM")Edited
A stereotype of Democrats, but not Republicans, is that their corruption is generally funneled through their NGO-industrial complex and via rules that target preferred groups for special handouts (government contracts go to women/minorities/etc…), so there potentially being different modes of corruption is a fair worry.

[25m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283104275 "Jun 26, 2026, 1:35 PM")
Also broadly turning a blind eye to fraud.
Republican corruption seems more personal enrichment; Democratic corruption is more often greasing the wheels of machine politics.

[13h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282867501 "Jun 26, 2026, 1:16 AM")
The people who were banned personally from major social media platforms during the Great Terror of 2015-2023 have a different perspective on how bad sending the third-rated late night talk-show host to time-out for a month is.
There were a few incidents during the campus Gaza protest crackdown and the Charlie Kirk assassaination aftermath that were a little taste of how bad things were back then, but speech is much freer now than it was 5-8 years ago.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282904272 "Jun 26, 2026, 2:57 AM")
It's very difficult to read this in any way other than, "I got banned from Twitter, that's basically the same as somebody going to jail for a Facebook post."

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282912608 "Jun 26, 2026, 3:22 AM")
I wasn't banned, but a lot of people I like to read were. The breadth and quality of public discourse was severely hampered by both the absense of these specific voices and the chilling effect from the fear of the banhammer.

[9h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282936061 "Jun 26, 2026, 4:42 AM")
But isn't that just the right of a private corporation to serve who they want to serve?

[8h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282952955 "Jun 26, 2026, 5:50 AM")
That's ridiculous. It's firstly documented that the administration intervened to put political pressure on these sites, and it doesn't change anything that Democrats used this sort of 'private' thuggery to push their political viewpoints.

[8h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282953220 "Jun 26, 2026, 5:51 AM")
A functioning media is a core part of a functioning democracy. If enough of the population’s media exposure is mostly limited to the platforms in question or sources downstream of them, a unanimous decision by that set of platforms not to cover certain viewpoints would prevent the democracy from addressing those viewpoints effectively.

[5m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283115243 "Jun 26, 2026, 1:56 PM")
Even if you think it was overreach and censorship, you have to agree that there was actually some legitimate public interest in censoring COVID misinformation.
There was no legitimate interest in the Kimmel stuff though. This was strictly because the president had his feeling hurt, wanted to hurt his opposition, and send a message to limit criticism against him on public airways. This speaks deeply to the differences about fundamental views about Freedom of Speech between the two parties, EVEN IF more censorship happened during the COVID era.
(none of this is to mention the fact that it was primarily private platforms doing the censorship during the COVID era and not the President himself)

[13h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282870093 "Jun 26, 2026, 1:22 AM")
Again, I repeat. NGO-industrial complex is no worse than traditional GOP carveouts for business interests and their beloved “public-private partnerships” that they outsource rather than giving governments any extra funding. How about when republicans turn a blind eye to white collar crime and basically “turn off” any financial regulatory barriers and encourage cronyism.
Oh also I forgot when Obama and Biden were calling to prosecute Fox News for their speech.

[12h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282889970 "Jun 26, 2026, 2:15 AM")
You seem to think these replies are attempting to establish which is a worse kind of corruption, rather than that the typical kinds of corruption differ enough by party to make the Metaculus early warning system more about partisan question selection than real “democracy is at risk” proxies. Your additional examples strengthen, rather than weaken, that worry.

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283039390 "Jun 26, 2026, 11:10 AM")
If the corruption differs not just in quality but in seriousness, obviously the fact that only the worse kind gets weighted is not in itself a form of bias.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282896201 "Jun 26, 2026, 2:33 AM")
They didn't threaten to prosecute Fox News, but they had an absolute ton of "geez if you don't block stories about Hunter Biden/possible lab origins of covid/etc., we might have to regulate you good and hard" conversations with the social media companies.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282898006 "Jun 26, 2026, 2:38 AM")
Hysterical considering trump threatened to arrest Zuckerberg.
And guys. Scoreboard: name a wealthy Westernized country/society that is actually super right wing and very pro-democracy.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282917039 "Jun 26, 2026, 3:36 AM")
Japan?

[6h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282990236 "Jun 26, 2026, 8:22 AM")
Japan is socially mixed. They haven't legalised gay marriage, but they just passed a plan to deliver nationwide LGBT awareness and education to prevent discrimination. Hardly "super right wing".

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283021733 "Jun 26, 2026, 10:17 AM")
Please, Japan's conservative party just won the election by a landslide, and currently their main platform is tightening immigration restrictions. You can't possibly think some passing lip service to "LGBT awareness" makes the political landscape "socially mixed", the social attitude towards LGBT in Japan is comparable to America in the 90s.

[27m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283103476 "Jun 26, 2026, 1:34 PM")
This doesn't work anymore because "tightening immigration" has been a liberal policy for a long time now. Most liberal politicians agree on tightening immigration all across the West. Even still, having some uncomfortable opinions on gays does not make you "super right wing". It's kind of impossible to be super right wing and pro-democracy.

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283041046 "Jun 26, 2026, 11:14 AM")
I think Japan is pretty right wing, it's just also been very non-fascist/militarist about it in its last ~80 years due to the consequences of losing the war. So it's mostly just conservative in the most literal sense of the word - resistant to change, in almost all forms (they still use fax machines, lots of tech adoption is slow, etc).

[26m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283104086 "Jun 26, 2026, 1:35 PM")
Sure then that's just not "super right wing" by any reasonable and good faith definition of the question. It's just garden variety right wing.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282929303 "Jun 26, 2026, 4:19 AM")
Poland?

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283062238 "Jun 26, 2026, 12:06 PM")
What do you mean by "super right-wing"? Because that covers a lot of ground, and if you're only smuggling in the assumption "right wing of any stripe is the same as Nazi Fascist Horribleness", then you're going to go on about "that country isn't wealthy", "that country isn't Westernised", "that country isn't right-wing" etc. if people give you examples.

[6h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282987790 "Jun 26, 2026, 8:12 AM")
There's nothing wrong with this btw. They tried to direct the country towards the set of values that they wanted, that was never the problem with the right either. The problem (to put it mildly) is that the right does not care about laws, procedures, or ethics, when they try to implement their values.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283019697 "Jun 26, 2026, 10:10 AM")
You think hiding news stories that may reduce their electoral chances is simply "directing the country towards the set of values they wanted"? Isn't that also "not caring about laws, procedures, or ethics when trying to implement their values"?

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283043640 "Jun 26, 2026, 11:21 AM")
I think a lot of the COVID stuff was mainly done because "we're in an emergency and spreading false information that fosters panic or makes people act suboptimally is a danger to all". You can argue that it wasn't that serious or that you actually think that information was correct but it's not necessarily a selfish motive. Generally speaking, while truth-finding is hard, having widespread fake information, especially at critical times, is also a danger. It's honestly hard to navigate if one tries to tackle that kind of problem in good faith - the market effectively *doesn't* fix this and has even lots of incentives to be bad actors.

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283043829 "Jun 26, 2026, 11:22 AM")
Trump et al. tried to downplay the severity of covid when it first happened (doesn’t that fit the definition of hiding information). That is literally their approach to any form of bad news, deny it, lash out at fake news etc. Texas, Florida and some other red states are currently outdoing themselves trying to emulate Trumpism at the state level, so the “bias” cannot be isolated to one individual but clearly the whole party has this problem.
Also, how old are you? I’m sure you are also outraged by the Iraq War and how the Bush administration lied to the American public about WMDs, completely distorting the “news”.

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283060675 "Jun 26, 2026, 12:03 PM")Edited
"the right does not care about laws, procedures, or ethics, when they try to implement their values"
What a biased comment. I think there is plenty of evidence to make this exact same statement about the left.
Cancelling people and making them lose their livelihood for espousing conservative values (as was the norm just a few years back) was legal, procedural, and ethical?
And there are less abstract examples:
I would invite you to do some research on Pedro Sanchez, current president of Spain and shining star of global Socialism. He is one of the most machiavellian politicians to ever grace this earth and has eroded democracy in this country probably to a greater extent than Trump has in the US.
He got to power in his party's primaries by stuffing ballet boxes, got his wife a position as a tenured professor in one of the most prestigious universities in Spain (despite her not even holding a college degree), his campaign was funded by his wife's family's brothel business, his top minister just got convicted to 24 years prison for corruption and the top minister that replaced this guy is also in prison awaiting trial, he set up a sort of under-the-radar mafia squad with orders to intimidate police investigators, judges, media outlets who investigated his family and ministers' corruption, he controls the constitutional courts, the current Attorney General is his lackey (former one, also his lackey, was convicted of misuse of office), state-funded media is completely sycophantic to him, the national statistics institute (which publishes his popularity polls) is controlled by him (and always favorable to him), his spiritual guide and mentor, Jose Luis Zapatero (forrmer Socialist president of Spain), is under investigation for influence trafficking and corruption (among other things, they found 1.5 million dollars worth of undeclared jewels in a safe in his office, apparently gifts for helping Venezuela move oil and gold illegally out of the country under Maduro), he hasn't had enough congressional support to even pass the federal budget for the past 4 years, has routinely attacked judges (in complete disregard for the separation of powers) who have his wife, brother and ministers on trial, and the list goes on...
As a result of all this, polls are massively unfavourable to him. Despite that, he refuses to step down. Why? Because - he alleges - the danger of the Right coming to power is too great for the Spanish people. (ie. rather than letting them decide this in an election, he has decided for them). Basically taking a giant dump on laws, procedures and ethics in order to forcibly implement left-wing values...
You will rarely see a greater erosion of Western democracy than you will see in Spain at the moment, and this from one of the current heros of global Socialism...

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283063286 "Jun 26, 2026, 12:09 PM")
"his campaign was funded by his wife's family's brothel business"
Once upon a time that would have been called "living off immoral earnings" but of course we're all sex-positive sex-work-is-real-work now 😁

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283064454 "Jun 26, 2026, 12:11 PM")
Some of the brothels were gay brothels, so this was seen as "not so bad"...

[31m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283101308 "Jun 26, 2026, 1:29 PM")
Yes, "Cancelling" (as far as it exists) is perfectly legal. After Charlie Kirk died the right got many people fired for so much as an innocent comment. While I obviously disagree with their values, I am not going to have any legal case against someone asking someone else's boss to fire them.

[27m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283103420 "Jun 26, 2026, 1:34 PM")
not to mention the freaking Supreme Court (Harlan crow anyone?)
Free concert tickets and getting to act in a play, anyone?

[26m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283103982 "Jun 26, 2026, 1:35 PM")
And the idea that democrats would restrict speech more than republicans is hilarious after Jimmy Kimmel.
Yes, one shitty late-night comedian versus all the COVID-era speech suppression. though tbf some of that started under Trump! He never tried to set up a disinformation board, though. thank god biden wasn't stupid enough to tough out that fight.
[](https://substack.com/profile/6538309-vaughn-svendsen?utm_source=comment)
[13h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282872504 "Jun 26, 2026, 1:29 AM")
_User was indefinitely suspended for this comment. [Show](javascript:void(0))_

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282905837 "Jun 26, 2026, 3:01 AM")
Speak for yourself! This American will be fighting for democracy for longer than the current political fad of fashionable illiberalism

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282923285 "Jun 26, 2026, 3:57 AM")
Author
Banned for annoying low-effort comment that ignores the point of this post.

[12h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282886105 "Jun 26, 2026, 2:05 AM")
My input on potential question-selection bias:
"Will senior DOJ leaders override normal procedures to open or accelerate a case against a high-profile presidential opponent or critic in the following years?"
Note the phrase "override normal procedures". Looks suspiciously like this was inserted specifically to exclude the Biden-era prosecutions of Trump.
"Will senior DOJ leaders override normal procedures to protect the President, his family, or his allies from a legal case in the following years?"
This was resolved "yes" for Trump's EPA quashing the investigation into Senator Jim Jones's coal holding, but not for president Biden's pardon of his son Hunter.
"Will a SCOTUS decision weakening donor disclosure rules for political spending or electioneering communications be good law in the following years?"
Why is this assumed to be bad? Is there no privacy interest? Would overruling NAACP v. Alabama be considered good then?
"Will the U.S. federal government open formal investigations of at least 5 top-50 universities in the following years?"
Why are top-50 universities prima-facia assumed to be innocent? I can almost guarentee that they are all in violation of federal civil rights laws.
"Will U.S. authorities transfer any current or former U.S. citizen to a prison abroad in the following years?"
Wouldn't normal extraditions count? Am I misunderstanding something here?
It's also notable what isn't on the list:
Nothing about a state or city allowing noncitizens to vote. A lot of people are worried about that.
Nothing about demographic change. Seems like changes to the electorate itself should be considered as a possible threat mode.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282895730 "Jun 26, 2026, 2:31 AM")
Why is this assumed to be bad?
Even if we grant it is bad, why grant that it is undemocratic? Private donations are undemocratic, but private voting is democratic? Not obviously.

[8h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282954317 "Jun 26, 2026, 5:56 AM")
Former is one vote per dollar, latter is one vote per person?

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282900212 "Jun 26, 2026, 2:45 AM")
> Why are top-50 universities prima-facia assumed to be innocent?
Because everyone is prima-facia assumed to be innocent?

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282906520 "Jun 26, 2026, 3:03 AM")
Uh, this doesn't appear to reply to his point that a legitimate government could prosecute >5/50

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282905364 "Jun 26, 2026, 3:00 AM")
Note the phrase "override normal procedures". Looks suspiciously like this was inserted specifically to exclude the Biden-era prosecutions of Trump.
Trump attempted to prevent the legitimately elected President from taking office via creating slates of fake electors and starting a riot to try to pressure various people in government into throwing out the legitimate results of the election, I don't really see why exactly it would be bad for him to have gone to jail for this.
Frankly the Biden era prosecutions of Trump were too slow.
This was resolved "yes" for Trump's EPA quashing the investigation into Senator Jim Jones's coal holding, but not for president Biden's pardon of his son Hunter.
Pardons are, unfortunately, in the Constitution, otherwise the Jan 6 thugs would still be in prison rather than out and about molesting children &c.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282908764 "Jun 26, 2026, 3:10 AM")
You're right that whenever someone says, "prosecuting political enemies is bad," there is almost always an implied, "unless they did something really bad and deserve it". One might think that career prosecutors at the Justice Department would be competent to judge this, but the one time they pulled the trigger they were directly overruled by the American electorate, so apparently not.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282931231 "Jun 26, 2026, 4:25 AM")
Preemptive pardons seem like a ridiculous concept, though. I'm not familiar with the part of the Constitution that regulates pardons, but are there really no limits against preemptively pardoning someone before charges have even been laid or trials conducted?

[1h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283088126 "Jun 26, 2026, 1:03 PM")
It's a bad precedent. I've seen defences of it based on "well Biden was afraid that when Trump got into power, the MAGAs would go after his family with fake accusations and lawfare in revenge", but that equally now applies to any other president in office, including Trump.
He gave all his family, friends, and associates pre-emptive pardons? Well you see, he was afraid they'd be persecuted once he was out of power.

[6h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282977924 "Jun 26, 2026, 7:32 AM")
Do we know if the Jan 6 convictees committed more child sex offences since their release than we’d expect from their statistical profiles? I agree that jail is an appropriate response to the Jan 6 incident, but I don’t think anyone should be using political prosecutions to target non-political crimes.

[1h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283086281 "Jun 26, 2026, 12:59 PM")
"the Jan 6 thugs would still be in prison rather than out and about molesting children &c."
Chazza, things can get hot and heavy in the comments here but there are still some standards of evidence. Provide evidence that this allegation is more than just "oh let's call all the people we don't like bad names" or don't do it.
I don't want to see flame wars on here, I can get that elsewhere.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282907830 "Jun 26, 2026, 3:07 AM")
"Normal procedures" seems like a reasonable caveat because prosecuting major political figures isn't in itself undemocratic, if your concept of democracy includes the idea that political leaders should not have blanket immunity or otherwise be above the law. Prosecuting *plausibly or likely innocent* political figures to manipulate their eligibility for office is undemocratic, but the publicly available evidence suggests that Trump is not plausibly or likely innocent.
Pardons have a separate question, which was appropriately resolved Yes for 2023-24.
Considering demographic change as a threat to democracy is... well, I'll just say an extremely controversial and partisan-coded position, and not something one would expect a non-ethnonationalist organization to include in such an index.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282923590 "Jun 26, 2026, 3:58 AM")Edited
"Leftists have not subverted most institutions to be biased in their favor" is also an extremely controversial and partisan-coded position, yet it's the background assumption of most "democracy indices" (this one is no exception), which is the real reason why it won't convince any skeptical conservative.

[1h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283090346 "Jun 26, 2026, 1:08 PM")Edited
Yeah, my view on this is that Metaculus is not rigged, but the natural constituency for people who have heard of it, contribute to it, and are consulted by it is the liberal college-educated Democratic-voting bloc.
So there may indeed be biases there just simply because the mindset of everyone is "this thing good, the other thing bad, and the other lot always do the other thing".
If there are conservative to right-wing superforecasters, yippee! But are there?

[7h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282973548 "Jun 26, 2026, 7:13 AM")
Thanks for compiling this list, I agree some of these examples are egregiously targeted.
I did think one question you listed was actually fine. Weakening donor disclosure laws seems likely to be anti-democratic because corruption is a constant concern. Money is easy for an individual to provide in quantities that matter to an individual politician. Membership of political organisations is different, because “membership” on its own provides a much smaller tangible benefit to an individual politician. There are trade-offs here, but statistically, I suspect most times donor disclosure law changes have been about corruption, and most membership disclosure changes have been about political association. I’d be interested if you have examples where the opposite seems true.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282903234 "Jun 26, 2026, 2:54 AM")
I don't think these sorts of predictors are necessarily very good at actually predicting what we might intuitively think of as "a threat."
For example, imagine you had a prediction market in 1941, immediately after Pearl Harbor. You might reasonably sit down and calculate the balances of forces between Hitler, Stalin, Roosevelt, Churchill, and Hirohito, and conclude that there was virtually no way for the Axis powers to win. The industrial power of the USA and USSR were just too great, the failure of political collapse during Barbarossa combined with Nazi genocidal actions in the East mean that the Soviet population will almost certainly fight to the bitter end, Japan was too poor and too reliant on oil that it doesn't have good sources of, the Chinese theater bogged down the IJA while providing minimal strategic gains, Germany lacks the naval capacity to invade Britain so it's forced to rely on bombing which has never worked, democracies are extremely bloodthirsty once attacked so the US and UK are unlikely to give up even as the war grinds out, etc, etc.
But the only reason that is true is that everyone on the Allied side take Hirohito and Hitler very seriously as threats, and are extremely unlikely to change their minds! It would be crazy to conclude, "See? Only a 0.3% chance of them winning. You have TDS (Tojo Derangement Syndrome)."

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282911339 "Jun 26, 2026, 3:18 AM")
I don't think anyone took seriously the idea that Al Qaeda would take over the US, but they were still pretty concerned about terrorist attacks. War isn't all or nothing.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282916191 "Jun 26, 2026, 3:33 AM")Edited
Of course, but because liberals are so wonderfully civilized, they refuse to deal with threats before they become a problem. Nothing they'd do would change just because Trump is more or less of a threat. So all you can do is calculate probabilities and pray to God.

[8h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282961551 "Jun 26, 2026, 6:25 AM")Edited
What point are you making? Are you worried that if the Allied leaders believed in prediction markets, they would’ve taken Germany/Japan less seriously and the odds of an Axis victory would’ve increased? I think if the leaders were that likely to lessen their response, observers at the time would learn of that, and the prediction markets would take that into account. We already have some politicians who are more or less commonly attacked for being “too swayed by opinion polls”, people would have a sense whether a given politician listened to them much.
Or do you mean, democracy might be “under threat” but the threat might have a low chance of winning? I agree that’s likely, but less of an issue. Winning a war is awful, and in many ways worse than life after losing a war (though it was clearly right for the US to enter WWII). I don’t think winning a fight against anti-democratic forces is nearly as bad as living in an undemocratic country.
(If politicians cared that much about markets, there’d probably be interference from people wanting to shift the market, due to other investors not being *that* confident in their positions. The market outcome would fall somewhere between the “true” aggregate prediction and the desired outcome of the most-committed manipulator. But I don’t think that’s what you’re getting at.)

[5h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283000863 "Jun 26, 2026, 9:03 AM")
There could be a weird feedback loop where the probability of success determines how much resources you want to spend to secure that success. Could work in different ways, like:
"There is already a 100% chance of victory. Maybe we are sending too much resources, let's try a smaller amount and see what happens on the prediction markets."
"There is a 0% of victory. Clearly all resources we send are wasted, stop sending anything."
Both arguments seem reasonable in isolation (and the enemy propaganda would definitely make them). Of course it would be ironic if the latter situation happened as a consequence of reacting to the former.
A probability 50% is probably the one most motivating to send more resources. But I guess it depends on how the media would frame it.
Or a better example, imagine 99% predicted chance of victory. Some people would argue "hey, send a little more resources and make it 100%, we are so close to guaranteed victory". Other people would argue "hey, 98% is almost as good as 99%, plus we can still change our minds later it things start going wrong, and the extra resources could be spent on our education or healthcare instead". By buying thousands of bots on social networks, you could push the consensus in either direction.

[7h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282967698 "Jun 26, 2026, 6:50 AM")
Comparing today's prediction markets with what they might have led to in 1941 doesn't make a whole lot of sense. The information landscape was very different back then, the exchange of ideas limited by speed of delivery (literally!), much larger trust in institutions than today. Probably the average person had a lot less leasure time to spend on deep diving into political and military matters.
Even within this different context I'm not sure the predictor would be as bad as you think it might be. We can crudely think of the outcome of a "prediction site" as the aggregation of all available data, recency-biased and people's-belief-weighted. And more specifically for prediction markets, belief-as-currency-weighted. There are sources for the beliefs, and the stronger the source, moving up and into an actual information advantage (a big no-no in regulated financial markets, but almost a de-facto feature of prediction markets), the more chance someone will take advantage. In 1941, you'd probably have the few people in power taking the Axis forces very seriously use their information advantage to reap the benefits. Were prediction markets taken seriously by the media class, maybe they would have helped shape public consciousness, too.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282905475 "Jun 26, 2026, 3:00 AM")
Some observations on potential bias from a strongly anti-Trump, pro-proportional-representation perspective:
-- There is a question on the crude magnitude of gerrymandering (efficiency gap >= 20 seats in the House), but that doesn't fully capture the antidemocratic impact of widespread tit-for-tat gerrymandering. "Our safe seats cancel out your safe seats" is not actually a win for representativeness of the legislature or responsiveness to popular sentiment.
-- Reconciliation bill with no CBO input on Byrd Rule compliance seems weirdly orthogonal to democracy. It's not clear what a better measure of Senate "skew" would be, since given the baked-in skew of the Senate's constitutional design, it's not obvious whether ending the filibuster would be a net win for representativeness.
[](https://substack.com/profile/50720562-jon-smith?utm_source=comment)
[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282911189 "Jun 26, 2026, 3:18 AM")
_User was indefinitely suspended for this comment. [Show](javascript:void(0))_

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282923403 "Jun 26, 2026, 3:58 AM")
Author
Banned for this comment.

[7h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282966451 "Jun 26, 2026, 6:45 AM")
Before Trump’s second presidency, I heard someone argue that sane Republicans, especially the Supreme Court (I forget the language they used) would use a second Trump term to push through normal Republican objectives with unusually little resistance, but that those same sane Republicans would stop Trump from having a third term because normal Republican values include two-term presidencies. Trump’s bluster would never become more than bluster. I feel like that prediction has gone surprisingly well - the Supreme Court has blocked a lot of Trump’s craziest and most unconstitutional, least conventionally-Republican ideas.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283023711 "Jun 26, 2026, 10:23 AM")
That kinda makes sense, I could definitely see Trump's crazy ideas actually having an anchoring effect to make conventionally Republican agendas normal by comparison and thus pass more easily.

[11h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282913062 "Jun 26, 2026, 3:23 AM")
This post is a bit confused about the way predictions are now aggregated on Metaculus. You essentially described the Metaculus Prediction, which was deprecated in 2024:
https://www.metaculus.com/faq/#metaculus-prediction
The Community Prediction, which is used today, is just the recency-weighted median:
https://www.metaculus.com/faq/#community-prediction
It turns out there wasn't enough difference between the two to keep the old version of the Metaculus Prediction around. That means Metaculus manages to perform this well even with a pretty simple metric! That said, pro forecasters do still routinely outperform the community, so the site might find a better way to take advantage of that in the future.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282923712 "Jun 26, 2026, 3:59 AM")
Author
Thanks, fixed. I'm still confused by this, though. Is this basically an admission that superforecasting is bunk and we can get the same effect by polling random people?

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282929172 "Jun 26, 2026, 4:18 AM")
I don't think so, but I am also confused by this. The FutureEval hub shows a pretty stark gap between pros and the community at large:
https://www.metaculus.com/futureeval/
Maybe they just weren't weighting it well? 🤷 I've never asked.
The community forecast isn't random people, though, since the sample is significantly narrowed to "does forecasting" and then narrowed further based on particular interest (or commitment, in the case of tournament questions).

[7h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282968764 "Jun 26, 2026, 6:54 AM")
I wonder how much of the difference between pros and the community comes from excluding low-volume markets and low-effort answers? They say “The pro and community performance baselines are calculated using all questions where both humans and bots made forecasts — from the first forecast of our first AI model to today.”

[9h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282940959 "Jun 26, 2026, 5:01 AM")
This sentence confused me until I read the comments: "You can see that 86 people have made forecasts about this, although it’s the people with a long history of making better-than-average forecasts whose opinions drive the result."

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283011239 "Jun 26, 2026, 9:41 AM")
Author
You're right, I edited that out, thanks.

[35m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283099138 "Jun 26, 2026, 1:25 PM")
no. the flattering-to-metaculus explanation is that superforecasters and regular people both have a similar error spread, so most of the useful signal comes from aggregation itself. the less flattering explanation is that everything in metaculus is motivated by trying to impress various bigwigs, so when the simple aggregation and the superforecaster aggregation were different, it was opaque to explain to bigwigs why the superforecasters disagreed, so it wasn't useful.

[26m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283103999 "Jun 26, 2026, 1:35 PM")
Metaculus employee here. The answer is no.
- Pro forecasters still completely crush random people (even random Metaculus forecasters!) in basically all tournaments.
- Pro aggregates are better than random user aggregates, as shown by the FutureEval leaderboard posted by Nathan below, however part of this is that Pros probably spend slightly more time on the questions and are incentivised to cooperate, while the broader community does that less.
The reasons the Metaculus Prediction didn't significantly outperform the Community Prediction imo is that
- Good forecasters are overrepresented in most questions, as skill and the number of forecasts one makes are heavily correlated.
- The old way of measuring skill simply wasn't all that good. The new system is much better and we are working on a new version of reputation weighting at some point.

[10h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282919753 "Jun 26, 2026, 3:45 AM")
Has anyone looked into which questions caused the index to decline early in the year? This seems like it should be traceable to a specific cause.

[9h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282941417 "Jun 26, 2026, 5:03 AM")
While this post tries to achieve objectivity, it occurs to me that you could go in the other direction and make multiple subjective indexes representing different political positions with different weights.
Maybe one group believes gun control is tyranny and another believes the real tyranny is denying gender affirming care to adolescents. Why not track both separately? Then we can have an objective measure of the real, observed level of subjective tyranny!
(Obviously there are some issues to work through. You would need to track the measures across a decade or two, to separate the underlying ideological definition of tyranny from the day to day partisan wrangling. If candidate A wants to require ID checks for SIM cards and candidate A's supporters are suddenly in favour of this, we wouldn't want the metric to shift suddenly.)

[7h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282969637 "Jun 26, 2026, 6:58 AM")
You could. The post claims that everyone who’s done this before Metaculus has done that. You could make those indices in Metaculus as well, but it’d be harder to claim they’re “democracy indices” instead of “tyranny indices”, and commenters like Daniel above could find longer lists with more egregious examples of political leanings.

[8h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282954979 "Jun 26, 2026, 5:58 AM")
As time goes on, the Index will become more valuable insofar as it gets further from any possibility of biasing the questions. That is: any biases in the question set are the biases of 2025, when (for example) Republicans are more into memecoin-based corruption than Democrats. But as time goes on, the biases will drift - the biases of 2035 will be correlated <1 with the biases of 2025 (for example, Democrats might take up memecoin-based corruption, or Republicans might drop it) and as long as the question set is stable, biased question-selectors can’t optimize it for the most up-to-date biases.
This can also go the other way.
Suppose right now the authors make a competent and honest attempt at keeping the index unbiased. Then we freeze the question set, and in a few years things have drifted to make the index biased.
A priori, today I have no reason to believe the index is more likely to drift in one direction than the other; but that doesn't mean I should expect the index to drift towards less bias. (Ie variance can be high, even if the mean is 0.)

[8h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282955762 "Jun 26, 2026, 6:01 AM")
The problem with "Democracy Threat Index" as a concept is that even granting all of the methodology problems (and pretty much every single one of those is an attempt to make Trump look bad and the Dems look good) the only thing that such an Index can measure is how imperiled Democracy is, but not the source of the problem.
The underlying assumption "this got worse under Trump's rule, so Trump is worse for democracy" doesn't even survive the most basic scrutiny. If Trump files lawsuits against Universities, yes possibly he is attacking these institutions, but also obviously possibly they really are corrupt and the actual threat was the Biden administration welcoming their illegal racism. Trump White House revoked security clearances while Biden didn't was an example given on that page. Well yeah, are you surprised? All those officials supported Biden and didn't support Trump. If they were MAGA you can bet Biden would have revoked too.
Any number of things can happen where you will have a debate about whether the problem is that this admin is acting or that the previous one didn't. The real issue is it's already clear American democracy is under threat because both sides are willing to ignore the rule of law, justice and basic humanity to get their priorities in. This sort of index just doesn't help you determine who is causing the problem.
If you ask me it's absolutely obvious that the Democrats started this and the Republicans are waking up and fighting back in kind, which they absolutely should do.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283011400 "Jun 26, 2026, 9:41 AM")
Author
I think the conditional prediction market structure that I mention in the last paragraph should get around that objection.

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283026319 "Jun 26, 2026, 10:32 AM")
Curious to hear what you mean by the Democrats started this, what would you point to as the starting point? I've heard it argued that it started with Republican backlash against Obama (e.g. birtherism) due in some part to racism, and from the other side that the root cause was Obama being unusually divisive in pushing idpol.

[7h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282973929 "Jun 26, 2026, 7:15 AM")Edited
I am unclear how Metaculus can handle actions that are both legal and constitutional, or at least not in direct conflict with the law or the constitution, but the practice of which can range very widely from the relatively normal and innocent to the grossly and seemingly corrupt (though not in legal terms!) overuse. Obvious examples are gerrymandering, presidential pardons and executive interference in the employment and careers of civil servants.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283011675 "Jun 26, 2026, 9:42 AM")
Author
I think this is a judgment call (and unfortunately adds to the nontrustlessness of this), but I think an 80-20 would be to stick to monitoring the obviously unconstitutional things and admit that this is only a measure of a strict conception of "democracy threat" rather than general civic unhealth.

[6h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282984453 "Jun 26, 2026, 7:59 AM")Edited
I'd be interested in seeing this expanded to other countries using the same questions... but a lot of the questions really are very US-coded.
The advantage of having this index expanded is that a value of 39.3 could be placed in some context. Right now it's a bit hard to parse what exactly a 39.3 threat to democracy means. Does that mean there's almost a 40% chance that the US becomes a monarchy or an outright undemocratic oligarchy? Or is that a 40% chance that important democratic institutions are destroyed and the will of the people ignored but with a chance to come back given enough public support?
Essentially, does this mean 40% to become Hungary (before Orban was voted out) or 40% to become North Korea. Or is it neither?
Placing other countries on the index would make it a lot more useful.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283011740 "Jun 26, 2026, 9:43 AM")
Author
Great point.
[](https://substack.com/profile/105678704-marcus-a?utm_source=comment)
[6h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/282991901 "Jun 26, 2026, 8:29 AM")
_User was indefinitely suspended for this comment. [Show](javascript:void(0))_

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283011848 "Jun 26, 2026, 9:43 AM")
Author
Banned for AI written comment.

[5h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283000135 "Jun 26, 2026, 9:00 AM")
Metaculus, per a quick Wikipedia check, is based in the United States.
I seriously doubt that US-based institutions are capable of saying "this country is a dictatorship now", or even "this country is a hybrid regime now". Way, way too vulnerable to government pressure, both formal and informal, especially as the question isn't some abstract one but one which gets at the heart of the US government's legitimacy. Even foreign institutions are vulnerable to this, to a degree, because of the size of the US market.
This doesn't have to look like taking down the whole site for dissent, it can look like getting favorable regulatory treatment/contracts/help for the board's other businesses if you make sure the website displays the "right" answer.

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283029650 "Jun 26, 2026, 10:42 AM")
I dunno, I feel like people on CNN and MSNBC etc. (not to mention even more nakedly left-biased institutions) say this kinda stuff all the time without consequence. As others have pointed out, the questions in this case seem pretty obviously designed to make Trump look bad.

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283057803 "Jun 26, 2026, 11:56 AM")
I don't watch either of those networks, but I get the sense they keep their criticism firmly in acceptable bounds and don't touch any red lines, and assure their audiences that it'll all be okay if the Democrats win next time around (or if not, it's only because fixing Trump's damage takes time, incremental change, etc.). And I know they're owned by member's of Trump's class.
I highly doubt that they (or Metaculus, for that matter) are capable of or even interested in resisting authoritarian consolidation, especially if Trump gets leading DNC figures on board with it first.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283008748 "Jun 26, 2026, 9:32 AM")Edited
Where American domestic policy is concerned, I think there is a wild card potentially explosive enough in its effect that it can upset even the most careful predictions: the United States are unique amongst western democracies in possessing a constitution which theoretically allows nearly unlimited power to its chief executive. It was of course very much the intention of the framers of the constitution to make the executive the most powerful branch of government, but I doubt they could have foreseen the advantage Trump has taken of it. The US, for instance, is the only democracy which nominates its elected head of state effective (not simply symbolic or ritual) commander of all its armed forces.
(The President of France also has the title of Chef des armées (Head of the Armies), but it is the Ministry of Defence which administers the armed forces - Macron could not start firing senior officers for disagreeing with him).

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283041385 "Jun 26, 2026, 11:15 AM")
I kind of disagree. The system does give pretty extraordinary power to the chief executive, but the limits are also pretty clearly articulated. He has unlimited power only in the sense that "what if the army just did what he wanted and congress refused to impeach him etc etc" which would be true of any executive.

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283043329 "Jun 26, 2026, 11:21 AM")
It would of course be true of any chief executive, but I am not actually concentrating specifically on Trump, but on a system that can allow Trumps to exercise virtually unlimited power. Commander of the Armed Forces is one tool would-be Trumps can wield; perhaps even more dangerous is the apparently limitless ability to govern by executive decree, thus bypassing Congress entirely - as we have seen!

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283045103 "Jun 26, 2026, 11:25 AM")
That's pretty far from "unlimited" though, as we have *also* seen. There are fairly clear and well-understood limits on what can be done with an executive order, especially if congress is motivated to counter it.

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283050687 "Jun 26, 2026, 11:39 AM")
"Motivated" - aye, there's the rub! But I agree, the present US constitution does not give the chief executive an absolute carte blanche to do whatever he wants; but it allows pretty awesome latitude to anyone prepared to make use of it. Does any other western democracy?

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283071389 "Jun 26, 2026, 12:27 PM")
Any system is going to struggle if the other branches "let" the president or prime minister do a thing. I consider "motivated" to be an extremely powerful check on the executive. It seems we both agree that congress *could* constrain the executive in a number of ways, it simply chooses not to in many ways at present...but that's a decision. "The president can get away with anything as long as it's sufficiently popular" is a *pretty big* check on the power of an executive in a democracy, because the whole point of a democracy is that the government does popular stuff and doesn't do sufficiently unpopular stuff.

[31m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283101279 "Jun 26, 2026, 1:29 PM")
To elaborate, there is not some abstract factor of Congressional Willingness To Constrain The Executive. If Trump started using executive orders to ban puppies and kittens, one imagines that Congress would be rather motivated to stop that. If the Democrats win a huge Congressional majority in the next election, then Congress would suddenly start restraining the executive much more than it has been.
The degree to which Congress attempts to constrain the president is a function of the degree to which Congress disagrees with the president's actions. You could say that the executive has awesome latitude to do anything Congress is fine with, and that is perhaps a nontrivial feature of the US system, but when you put it that way there is an obvious check on his power.

[13m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283110396 "Jun 26, 2026, 1:47 PM")
It seems we both agree that congress *could* constrain the executive in a number of ways, it simply chooses not to in many ways at present...but that's a decision.
Ah, but how free is the free will of the majority of the Republican party in today's Congress; when a simple Trumpian snarl will bring most of them to heel? I may tell myself that I have 'chosen' to approach a chained wild beast no closer than its chain will stretch, but my choice was surely constrained, if by nothing more than self-interest!

[8m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283113389 "Jun 26, 2026, 1:53 PM")
Why is Congress afraid of a simple Trumpian snarl? It's not as if he's going to push them out a window, their fear is that if they go against Trump, they will lose the next election to a candidate who likes Trump more than they do, because the voters like Trump and wish to elect candidates who like Trump.
If you have a problem with that outcome you have a problem with democracy itself.

[4h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283011195 "Jun 26, 2026, 9:41 AM")
I understand the need to implement questions pertinent to the time (technological change means there are many more alternative ways to undermine democracy).
But why isn't the first port of call for this question to take questions similar/ the same as other Democratic Backsliding measures as opposed to generating questions on the fly?
It seems like we're replacing one intractable issue (slow updating of the metrics, "experts" being untrusted and possibly partisan) with its own (incommensurability over time, the measure's tendentiousness and partisanship to the challenges of the time, threats of participants being untrusted and possibly partisan...).
Perhaps we should start with the same set of questions, but in the Prediction market format, to fully isolate those benefits and avoid bringing on others?
True, updating the measures to the time is important, but like opening Pandora's Box, it's difficult to close. Without constraint on updating of the metrics, we'll just end up with a hodge podge of different measures that reflect the specifics of the most recent case of fraud.
Would a Meta-metric be more fruitful?
At the beginning of each year, questions are chosen based on current affairs. Over time, deterioration in those metrics are measured.
They are then aggregated (e.g. 2022, 2023, 2024, 2025 metrics are all rolled in to one). We can view democratic deterioration from the perspective of discrete points in time, but aggregating as such would "wash out" the effects of specific contemporary concerns (each having a lower impact).
This approach would age like fine wine (i.e. better over time with more measures) and it could be bolstered with a "core" set of democratic backsliding questions at the beginning that reduces the weight of specific questions.
We would perhaps still be incommensurate on aggregate, but on an individual year basis we wouldn't be and we could directly measure performance over time for these yearly measures (as opposed to total set of questions changing over time based on current affairs).
Partisanship is possibly diluted, particularly by the "core" set of questions (but isn't entirely solved).

[3h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283039662 "Jun 26, 2026, 11:11 AM")
I am kind of confused what the question "Will senior DOJ leaders override normal procedures to open or accelerate a case against a high-profile presidential opponent or critic in the following years?" I feel the problems with this question should be sufficiently obvious to a non-lawyer. What are normal procedures? Grand juries? That happened already. Arguably that happened going to back to forever. Charging people for stuff that "doesn't normally get charged?" Every white collar defendant thinks that's what's happening, and they're not always wrong.
I could say "this has already happened with Comey/SPLC" and someone would potentially comment "The SPLC committed pretty obvious financial crimes as detailed here" and then I could say "that's not a crime" and then we'd be...having a criminal case and a trial. That's what the process is FOR, to see if normal procedures have been followed.

[2h](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283053303 "Jun 26, 2026, 11:45 AM")
"Their Democracy Threat Index is a collection of 153 questions relevant to US democracy."
Huh? From the Metaculus link you provided:
"Our new U.S. Democracy Threat Index combines 39 concrete forecast questions into a single metric tracking institutional resilience."
Where is this 153 coming from?

[43m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283095164 "Jun 26, 2026, 1:17 PM")Edited
The fundamental problem I see is that these questions amount to predicting instability/polarization rather than authoritarianism under one central leader.
On the broader topic, I think the "democracy protector" crowd must address things like the annulment of the 2024 Romanian election under laughable claims of "Russian social media bots", the unprecedented level of speech arrests and police brutality in the UK, and the broader problem of leftie parties importing millions of people who are going to vote for them no matter what. All three of those are existential threats to any reasonable definition of democracy, and all three have already happened.
Instead, these people talk about the prospect of Trump suddenly declaring a military coup in 2027 after more than a decade of a decidedly and unnecessarily deferential attitude towards judges and congressmen.

[27m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283103444 "Jun 26, 2026, 1:34 PM")
The first two are genuine abuses, but I really take issue with
the broader problem of leftie parties importing millions of people who are going to vote for them no matter what
this is empirically false, and in any case undermines not democracy (how is it democratic to deny anyone the franchise?) but the nation-state system.

[22m](https://www.astralcodexten.com/p/the-metaculus-threat-to-democracy/comment/283105918 "Jun 26, 2026, 1:38 PM")
i really feel like the worst challenge to a "democracy threat index" isn't various accusations of bias, it's the concept of an index itself. which is more persuasive: "the chance of riots / violence at democracy-related protests killing more than 100 people moved from 20% to 30%" or "The Metaculus Democracy Index moved from 20 to 30"? things like the DJIA or S&P500 are naturally grown from an already existing set of data, all of that data already obsessively being monitored by investors. the trading data can be backtested. metaculus recieves a millionth of the attention for the stock market, so scott being reassured that "the index isn't moving much" can be explained by "no one is looking at that particular wing of this hyper-niche site". i would also be willing to bet if something newsworthy happened that would move the democracy index, metaculus would shoot themselves in the foot by launching 30 more questions about it, diverting attention away from the existing questions.
Top Latest Discussions
Sixty-eight years of highly defective people
Jan 16•Scott Alexander
2,074
876
306
You just keep on trying till you run out of cake
Jan 21, 2021•Scott Alexander
1,590
483
38
Ivermectin: Much More Than You Wanted To Know
Nov 17, 2021•Scott Alexander
442
2,113
5
See all
Ready for more?
Subscribe
© 2026 Scott Alexander · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture