I think a 95% probability is too high. There are only four elections between now and 2038, so even if we assume the odds are even and independent for all four, that still gives a 1/16 chance that Republicans lose all of them. But the odds should actually be slightly higher than that because the probability of Republicans winning each election isn't independent. Most presidents win reelection, so if a Democrat wins in 2028, they'll probably win in 2032 as well.

Plus, we should factor in the probability that certain political trends make it harder for Republicans to win the presidency, giving Democrats a winning streak. This type of thing has happened in the past, and there's no reason to think that presidential politics will always stay as evenly divided as it currently is. So this introduces yet another way that the probabilities are non-independent that makes it more likely that Republicans lose all four than it would be if they were independent.

Of course, there's also thermostatic public opinion, which weighs in the other direction (it makes the probabilities non-independent in a way that makes a winning streak less likely). But I'm not sure it's enough to outweigh the other two factors.

@JosephNoonan Thermostatic opinion weighs way harder than anything else here. The last time one party won 4 presidential elections in a row was FDR, and that was when it was legal for him to have 4 terms and he had literal World War 2 propping him up. The last two elections have been some of the closest ever. The incumbency advantage is less than ever (and may even be negative, if you look at approvals and margins).

@SemioticRivalry Hm, I don't know if it says that much that it hasn't happened since FDR since the chances of it happening in that time frame would only be a little more than 50% if what I'm saying is true. But if the incumbency advantage is negative, that definitely makes it much less likely for the same party to win four times in a row. I know the incumbency advantage is shrinking, but I've never heard of it being negative.

@JosephNoonan It being negative is little more than the purest form of Trump-voter cope (they desperately want Biden's current polling to reflect his upcoming popularity in the 2024 elections)

@JosephNoonan incumbent margins have been worse than their first election on average in recent history, and by a decent margin. you can also observe what happened to Obama, trump, and Bidens approvals after they took office.

@SemioticRivalry I don't think we can count Biden since his reelection bid hasn't even happened yet. Trump and Obama definitely count as examples, but before that we get Bush and Clinton, who did better in their reelection bids than their initial elections.

@dgga I don't see how that's "Trump voter cope." Looking at Biden's approval rating as a negative sign for his reelection seems pretty reasonable to me, and insisting that it means nothing seems like wishcasting. It only gets into "Trump voter cope" territory if they look at it as the sole sign of reelection chances, ignoring the fact that Trump is also extremely unpopular, and all the factors that could give Biden an advantage.

Carter did 12 points worse in his re-election, Reagan did 8 points better, HW did 13 points worse, Clinton did 2 points better, Bush did 3 points better, Obama did 3 points worse, Trump did 2 points worse. Averaging it out has incumbents doing 2.6 points worse. Of course you could go back further and find different results but I'd say the more recent info says more about 2024. In general, I wouldn't care that much about an N=6 sample, but I'm skeptical of the pro-incumbency evidence. I think most of incumbency is "hey I know that guy, I'm gonna vote for him!" but in a presidential election that is like a black hole for media attention and sucks up all the oxygen in the room, this really doesn't apply. I think that model hasn't worked in several decades, though it still works on down ballot races.

We can also look at favorables. When Biden took office, he had +8 favorability. Now he has -14. When Trump took office, he had -8 favorability, now he has -16.

This from YouGov is particularly striking: Obama was very well liked when he entered office, very quickly became controversial, and then was very well liked the second he was leaving office. Being the figurehead for all of the problems in the world is difficult!

@PlasmaBallin 😄

This is of course very unlikely to resolve no, but I'll note that DuN did say

I would bet a significant amount that we won't see a (democratically elected) republican president before 2040.

Emphasis mine. I think it'd be interesting to see a version of the question where the republican is required to win the popular vote, not just the electoral college. Would still be high... but not this high.

@Joshua That clarification was actually edited in later. I was thinking of asking whether a Republican would be elected by 2040 but changed it to match the original prediction better. But yeah, I think the probability gets a lot lower if the Republican has to win the popular vote as well.

@JosephNoonan @Joshua the clarification was indeed edited in but it is meant as a caveat to rule out blatantly undemocratic shenanigans like a successful insurrection or some sort of over-the-top ridiculous gerrymandering effort slapped on top of the status quo to blatantly underwhelm electoral integrity. I did not mean popular vote and I consider the presidents so far to be democratically elected regardless of the popular vote.

I'll be sitting on it and might bet on the trends in this if I deem it to be worthy in shorter term (this market just might dive in the next elections when republicans get absolutely crushed). I have little faith in the current ~50-50 balance of power staying as it is until 2040.

I don't bet on politics.