I’m very surprised that in this article you didn’t address the elephant in the room that could affect the polls in the near future. Trump doubling and tripling down on Harris’s race/ethnicity. Millions of people in the US are biracial and yet Trump can’t grasp the concept of identifying as more than one ethnicity when you are indeed BOTH. Add that to dufus Vance and his weird ideas about childless people and they look like complete idiots to people who don’t float in the MAGA orbit.
"I’m very surprised that in this article you didn’t address the elephant in the room that could affect the polls in the near future."
There are already plenty of pundits who try to predict the polls (and typically fail). Nate is valuable precisely because he *doesn't* try to predict them.
Fair point, but I think he does give hot takes from time to time. He said democrats shouldn’t get “over their skis” about the recent positive polling for Harris, so that’s what I was responding to. Isn’t that predicting the polls in a way? Don’t got too excited because this probably isn’t a trend? And Trump has punched himself squarely in the groin even since those poll numbers were added, but he didn’t mention it. Thats fine, but major campaign blunders matter.
I'm completely anti-Trump, but I'm not convinced that his statements were the disaster that is being attributed to them by the MSM and others. First, let's get it out of the way that he was both lying and making false statements. Were they any worse than Biden's address at the black college commencement? Reasonable people can disagree, but I don't think they can disagree that it's a close call. Now on to the point. People may not want to admit it, but many blacks may not feel that Harris is naturally one of them. Trump doesn't have to get the majority of the black vote. He just wants to get 20 or 30%. This is possible and playing on the black, particularly male black vote, on his remarks it way move his ball forward.
Couple tips for your next attempt at concern trolling Norm. Real people who are anti-Trump tend not to use the term "MSM" and assume without evidence that it's being unfair to him. They also rarely engage in whataboutism like claiming Biden's ho-hum speech at Morehouse was somehow on a similar level to Trump's racist screed of an interview at NABJ.
I don’t know how black voters will feel about it. I don’t know if it will hurt him with any particular demographic. I just think it hurts him across the board for him and his surrogates to keep posting family photos of her in traditional Indian clothes like that should embarrass her in some way. The one he posted on Truth Social had the caption, “She’s always been Indian.” The more I think about it maybe that is the real strategy. Back to the Obama birther rhetoric…”she’s foreign.” Again, that only resonates with a portion of his base like the ones here in rural East Texas. It cannot help him with new votes.
What do the people who wouldn’t vote for her because “she’s foreign” think about Vance’s wife? It seems like she would give them pause. (But maybe not, considering that never seemed to affect Mitch McConnell.)
I think ppl are going to hear about these Trump comments and not even know what the Biden comments are that you’re referring to. Seeing this as a strategy to shore up a portion of the black vote seems pretty out there to me.
Obviously. He’s had a solid base since day 1. There’s nothing he could do to lose their votes. Some of those people are my family members, friends, and work friends. I know all about it. I live in rural Northeast Texas.
"Extreme" by whose standards? There remains a convincing majority of Americans who support Israel over Hamas. That probably also includes a majority of the subset of Democrats who will actually show up at the polls in November (i.e. not the much-hyped but ever-elusive "youth vote").
Would choosing Shapiro as her running mate lose her Michigan? Perhaps, but I think that would be offset by other purple states (and not just Pennsylvania) being brought into her column that she otherwise would have lost.
I think it is much ado about nothing. Its a pretty mainstream viewpoint on protestors and Israel. I think Harris has shown she's threading the needle on it and I expect, given his comments, they will continue to draw that distinction between Israel's rights to defend themselves and what they are currently doing in Gaza.
The evidence there is really thin tbh, unless a mainstream news outlet substantiates this more I don’t think this should be a concern with picking Shapiro
If someone gave me better than 1/100 odds, I would bet on a tie. I think she will take PA, WI, and MI, but then I think she will lose in Nebraska 2nd, which would be a tie if there were no other surprises.
The model is currently giving her better chance in NE2 than any of PA, WI, and MI. Let's stick to methodology and data, rather than any feelings we might have :)
Nate just recently wrote about this, the model will probably give Harris like a .2% overall win chance based on Shapiro. But of course it’s possible she could see a more measurable polling bump in PA because of him which would then factor in. (Fingers crossed?)
The model would wait for polling, but as to the question of what would happen to polling, it does subjectively feel like a case where VP pick will matter more than usual.
Why? I guess because Kamala *is* a VP, and because her VP pick will almost certainly be of a different gender and race and those things are not meaningless. In '08 Biden helped Obama on the margins, surely.
Can you explain a bit the tail probability that leads to the model showing a 5.3% (as of the 8/1 snapshot) chance of Kennedy getting at least one EV? Other than maybe ME-2, I can't see how he wins any of the split EVs, so otherwise we are talking about him pulling a plurality win in a state.
Or is the model also allow for "rogue electors"?
[edit to add]
For the 5.3% of simulation runs where RFKjr wins at least one EV, what is the distribution of his nationwide PV %?
Third party candidates often have one state where they do unusually well. The model knows that. While at the 5.3% probability level, it only has Kennedy getting 4.5% of the national popular vote, if that is heavily concentrated in one small state it could be 51% of the vote there. It's a crazy scenario, but the model knows that crazy scenarios do sometimes happen.
I should note that Kennedy could only need to win 33.4% actually, going with the not-too-crazy assumption that in this unlikely hypothetical scenario that the remaining balance of votes in that state were evenly split between Harris and Trump.
I decided not to dive into that because of the complexities of which candidates Kennedy would be drawing votes from to get to his plurality. I should have just said "plurality".
Yes, and RFKjr only needs a plurality, not 51% (or he could win via RCV-IRV in ME-02 or ME-statewide via being a second choice of both Harris and Trump voters)
I agree that crazy things can and do happen; 5% seems *enormously high, hence my asking about the underlying distribution.
Note that Gary Johnson in 2016 (with a party apparatus behind him and 50 state ballot eligibility, neither of which RFKjr has so far) wound up getting 3.28% nationwide, not that much behind where the model thinks RFKjr is "polling" now. Other than New Mexico, where Johnson got 9.34%, his best state was NoDak at slightly less than 2x his national final vote result.
Assume that fourth and fifth parties wind up getting a few percent in the state/CD that RFKjr winds up winning (yes, I know the 5.3% is the probability over all states plus WDC and the NE and ME CDs), and that Harris and Trump are basically dead even. Other than RCV-IRV in ME-02, that means RFKjr has to get a plurality. Let's call the plurality a mere 30% (see fourth and fifth parties - I have no idea for example where Jill Stein and the Greens are for 2024). With RFKjr coming in with a tailwind somehow and getting 6% nationwide, he still has to score 5x his national vote percentage in the relevant state/CD.
AFAICT that kind of concentration hasn't happened in the 20th C - the closest I can find is 1968 when Wallace got very close to 5x his nationwide 13.5% in AL and MS. LaFollette got about a bit over 3x his nationwide 16.6% to win Wisconsin in 1924.
He still needs to get in the ~30% range to win the IRV races, because he needs to beat one of Harris or Trump into second place, and he also needs the one he doesn't beat to fall decently far short of 50% so he can overtake them in the runoff.
Agreed. Best case for an EV is about 30% with a lot of minor party votes and a reasonably even split between Trump and Harris (I can easily see Trump voters in an IRV race picking RFKjr for second preference; I'm not sure what Harris voters would do - they have a bigger coordination problem).
Which state is his support heavily concentrated in? Seems like there would likely be some sign already if there was a strong state in particular for him.
If there was a state where he was regularly polling at 15% while the major candidates are both in the upper 30s, I don't think that would be enough to be giving him much higher than 5% chance, would it?
I think that would give him a much higher than 5% chance - 3rd party candidates are hard to poll (a lot of polls don't even include them), so one doubling their expected votes doesn't sound that unlikely to me (especially when you allow for how much can happen between how and election day).
If you take his current probabilities for each state (etc) and assume they are all independent, you get a probability of him winning at least one seat of 10.1%. They obviously aren't independent, which is why the model is actually giving him a lower probability. His highest probability in any state is Alaska where he has 1.4%. There don't seem to be any polls for Alaska since Harris entered the race, so I think the model is guessing quite a lot there...
Agreed that third party vote polling and vote projection should be considered very very noisy (such thin historical data, all but 2016 effectively pre-social medio too). Makes complete sense that AK would be the best single state according to the model (historically high elasticity, right up there with NH, in particular)
Also, and I don't think the model takes this into account (nor _should_ the model it get that far into the weeds), according to a claimed capture from the RFKjr campaign website about a week ago (no URL given and I could not find the specific data when I just looked), RFKjr does not have ballot access in ME, which (ME-02) is his best chance at getting an EV via RCV-IRV rather than having to have a plurality.
It's the probability to win at least 1 EV not exactly 1 EV, eg it's the probability that he wins at least one of the states or districts. 5% seems about right
I'd assume that if the election were tomorrow, the odds would be a lot smaller of Kennedy capturing an EV. Some decent portion of the chance, in other words, represents scenarios where the race shifts in his favor between now and then...
oh, very definitely. That's why I am curious what the distribution of his nationwide PV is for the 5.3% of total scenarios wherein he wins at least one EV.
Looking at strong third party finishes post WW2, so mostly Wallace 68, Perot 92, and Johnson 16, it looks like the plausible best case regional concentration, ignoring special factors applying to Wallace 68, is 3x nationwide vote, and even that is stretching it for RFKjr. So yeah, with RFKjr winning say 12% nationally he *might* get enough concentration to squeeze out a plurality win, but looking at e.g. Johnson 2016 I suspect RFKjr needs to hit at least 15% nationally, getting pretty close to Perot 1992 territory (19% nationwide). Perot came pretty close to winning ME-02, losing to Clinton 38-33, so yeah, conditional on RFKjr getting well above 15% nationally one or more EVs is pretty plausible.
Well, based on today's model update, the actual answer might have been more like "the model just isn't that confident that really weird outcomes aren't more likely than we think due to small sample size of the historical polling data." Chance of RFK winning an electoral vote is way down after the model tweak:
"This model run also introduces a small technical change that somewhat reduces the instance of outlier outcomes in individual states, but you really have to squint to see any effect. The most notable difference is that the model now gives RFK Jr. roughly a 2 percent chance of winning at least one electoral vote, down from 5 percent before"
I addressed both those, and I agree that _conditional on_ RFKjr getting into the 15% range nationwide (getting close-ish to Perot 1992, who got 18.9%) it would not be particularly surprising (I don't see how RFKjr achieves the level of regional concentration that Wallace had).
Hence if the model thinks RFKjr has a 5% chance of getting more than 15% nationwide, I agree re EVs and I change to being curious about the distributions that have RFKjr pulling a Perot 92 level nationwide vote.
I still don't see it, although in the model's defense :) there are so few examples of serious** third party runs that the underlying variance has to be considered high.
** I'm not sure whether RFKjr counts as a serious run in the sense that Perot was.
But RFK is no Perot or Wallace! He’s not riding a nationwide wave of people who think the two main parties are the same, and he’s not someone with a strong regional base. What’s his upside story that makes him more likely to get an elector than Jill Stein or Ralph Nader or Ron Paul?
Hmm, I _did_ look before posting, but somehow I missed what you describe. I was thinking of emailing Eli; now I will do so and plead that he point me to the "already addressed here" comment.
If you haven't seen it, the model was updated today to reduce the probability of extreme events in individual states, so the probability of Kennedy getting at least one EV is now only 2.0%. It seems they had unrealistically fat tails in the distribution for each state - there was a bit of discussion about that in yesterday's article.
The US Constitution (which has not been amended recently) has nothing directly to say about rogue, aka faithless, electors. I *think* the state level faithless elector "laws" are state statutes, not state Constitutional provisions, but I could easily be wrong about that.
Chiafalo v. Washington (where it was a Washington state statute in question) says that a state *may* punish a faithless elector, not that the states *must* punish them. Nor AFAIK does Chaifalo directly speak to the recognition of a such an electoral vote sneakily, or cheekily :), offered in the face of punishment.
AFAIK Ray v Blair (1952) is still good law, explicitly allowing states (or in the specific case a political party) to require a pledge, but not addressing enforcement of the pledge. See also Justice Jackson's dissent in that case.
Kamala doesn’t even have to campaign anymore, just let Trump and Vance keep saying stupid things (especially Vance). If Trump loses he only has himself to blame at this point.
Black men. Trump does pretty well among men. Black male unemployment remains higher than black female unemployment, so Trump's economic message resonates better. And the gap between black men and black women is growing - DEI in the spotlight. If some black men feel less racial affinity for Harris, they may vote on sex before race.
Oh, so you did understand what I was saying about the MC poll (which was good for Harris in more than just MI, which I also explained) being important. You were just bad faith playing dumb to rile me up. It’s like you are a character written by Sartre himself to demonstrate his point about people with evil views. And now, I assume, you will begin doing the same thing to anyone who bites here.
The MC poll had an absolutely crazy result in MI. I have to wonder if the magnitude of the change there is having an outsize effect on the model's output.
There is a Glengariff Group poll with a very slight Harris lead. And a Public Opinion Strategies poll and a Civiqs poll with them equal. There are only 9 polls total for MI, though, so the model will be heavily relying on national polls and polls from states with similar demographics.
Dow Jones drops 700 points as PMI drops below 50 and initial jobless claims reach levels not seen in a year. When does the "fundamentals" portion of the model get updated?
The question is whether Harris gets the blame for current conditions in the country. She is after all part of the current administration.
EDIT: I'm not talking about the drop in the market being significant, I'm talking about WHY the market dropped. Namely PMI missed big.
You're right—the market had a one day decline. It had an all-time high—checks notes—two weeks ago. On news this bad, Harris is definitely toast! (Over in the land of Copium)
You are as wrong about this as you are about politics:
-The strong productivity growth (2.3% vs 1.7% expected) coupled with lower unit labor costs (0.9% vs 1.8% expected) suggests that economic growth in Q2 2024 was organic and not driven by inflation. This indicates a healthier economic environment, not an economy on the verge of recession.
-the combination of strong productivity, lower labor costs, and modest manufacturing slowdown suggests the economy may be heading towards a "soft landing", which continues to be the majority view
-Both the PMI (49.6) and ISM (46.8) manufacturing indices show contraction in the manufacturing sector, but nothing indicative of an imminent recession
-we're still on a glide path for rate cuts
-The mix of strong productivity growth with signs of economic slowing suggests a more balanced and potentially sustainable growth pattern, which could help avoid boom-bust cycles that often lead to recessions.
This is an exact quote from a financial website: "Meanwhile, manufacturing data from the PMI (S&P Global) showed a slightly better read than expected, at 49.6 vs. 49.5. This is the first reading under 50 since January and may be indicating the economy is slowing down from the brisk pace in Q4 2023.
The ISM manufacturing number for July was also released, and showed a 46.8 reading, which was less than expected (48.8)."
Don’t bother. He’s not trying to have an actual discussion. He’s willfully misinterpreting the data to get smart people to tie themselves in knots. There is a famous Sartre you’re about people like him.
There has been open discussion in the economic/business media that the economy might slip into recession by the election. You have missed this? It isn't a conspiracy though doesn't look likely.
If you took the time to backtest, then you would know the PMI 50 threshold is a poor leading indicator of recession - not much better than chance. We certainly could still have a recession, but the PMI movement has little-to-nothing to do with it.
That is not my argument. I suggest looking at the totality of indicators (such as the Sahm rule) for agreement across measurements to try and gauge the direction of the economy.
So let me get this straight: according to you, that disaster show Trump put out yesterday is irrelevant, but daily fluctuation of stock market of 1.5% is super important?
The stock market dropped a little over 1% based on slightly poor results in the manufacturing sector for one month, which represents about 20% of the US economy. There is nothing in that that suggests an imminent recession.
I have done. There is no headline saying "Recession on the way". It just reports the numbers - a small fall in the stock market after slightly disappointing manufacturing numbers. You keep telling people to read CNBC, but it doesn't support what you are claiming. You are forming your own interpretation based on what it says and that interpretation is fundamentally flawed.
That's sensationalist headline writing. You have to actually read the article. It's just people adjusting their predictions very slightly in a downward direction.
It’s a shame trolls like this can’t be blocked here. It diminishes the ability for everyone else to have reasonable discussions. I wonder if Nate can do anything about it? The deluge of repetitive, agenda-pushing posts is what ruins social media.
Contrary opinions are fine, but can you just posts whatever nonsense you want once per article and be done with it?
Here's a question: what do you think is the majority opinion of economists about whether an imminent recession is likely? Hint: it is different from yours.
“Infidelity does not consist in believing, or in disbelieving; it consists in professing to believe what he does not believe. It is impossible to calculate the moral mischief, if I may so express it, that mental lying has produced in society. When a man has so far corrupted and prostituted the chastity of his mind, as to subscribe his professional belief to things he does not believe, he has prepared himself for the commission of every other crime.”
Agreed on this point. If I was Jerome, I would have begun moving rates down two meetings ago. They need to be looking at the 2nd derivative of the economy and not just the 1st.
You are in essence arguing "The trend is your friend" when you make statements like that. And if the economy is not on the brink of recession there's no need for premature cuts that run the risk of reigniting inflation.
Going off of the explicit and implicit goals of the Fed, which are to lower inflation to (a completely arbitrary) 2% in a way that leads to a “soft landing” and does not put the economy into a tailspin. I am suggesting that acceleration of the economy would be beneficial to consider in combination with the velocity of the economy instead of the velocity by itself. It would be a mistake to assume that the only way to justify cuts would be if we were on the brink of recession. That is simply not how monetary policy works.
Au contracted, that is precisely how the Fed works. The historical record is clear: the Fed always cuts late. I suspect that the primary consideration is that taking the foot off the brake too soon runs the risk of reigniting inflation.
Again, a less risk averse approach that prioritized a "soft landing" over taming inflation with no potential for resurgence would have led to cuts at the start of the year.
Two hundred and seventy is exactly what Harris needs. With 269, she will lose according to the complicated procedure of elections in the House of Representatives in case of a tie. It’s not one elector more than needed, as you wrote.
Odds are more states have a Republican majority than a Democratic majority in terms of House members. Remember, the threshold is 26 states, not a majority of House votes.
It is highly unlikely that 26 states will vote for her in the House. However, Trump might not secure them either. In this case if a Democratic miracle happens in the Senate, Shapiro or Kelly will get the support of 51 Senators and become the acting president.
I’m a bit confused. The article says it’s a toss up but when I look at the model Kamala is in the 30s. Is it possible I need to force some kind of a refresh?
I’m very surprised that in this article you didn’t address the elephant in the room that could affect the polls in the near future. Trump doubling and tripling down on Harris’s race/ethnicity. Millions of people in the US are biracial and yet Trump can’t grasp the concept of identifying as more than one ethnicity when you are indeed BOTH. Add that to dufus Vance and his weird ideas about childless people and they look like complete idiots to people who don’t float in the MAGA orbit.
"I’m very surprised that in this article you didn’t address the elephant in the room that could affect the polls in the near future."
There are already plenty of pundits who try to predict the polls (and typically fail). Nate is valuable precisely because he *doesn't* try to predict them.
Fair point, but I think he does give hot takes from time to time. He said democrats shouldn’t get “over their skis” about the recent positive polling for Harris, so that’s what I was responding to. Isn’t that predicting the polls in a way? Don’t got too excited because this probably isn’t a trend? And Trump has punched himself squarely in the groin even since those poll numbers were added, but he didn’t mention it. Thats fine, but major campaign blunders matter.
"Isn’t that predicting the polls in a way? Don’t got too excited because this probably isn’t a trend?"
I read it more as "there is insufficient evidence yet to adduce a trend", not "there probably isn't a trend".
I'm completely anti-Trump, but I'm not convinced that his statements were the disaster that is being attributed to them by the MSM and others. First, let's get it out of the way that he was both lying and making false statements. Were they any worse than Biden's address at the black college commencement? Reasonable people can disagree, but I don't think they can disagree that it's a close call. Now on to the point. People may not want to admit it, but many blacks may not feel that Harris is naturally one of them. Trump doesn't have to get the majority of the black vote. He just wants to get 20 or 30%. This is possible and playing on the black, particularly male black vote, on his remarks it way move his ball forward.
Couple tips for your next attempt at concern trolling Norm. Real people who are anti-Trump tend not to use the term "MSM" and assume without evidence that it's being unfair to him. They also rarely engage in whataboutism like claiming Biden's ho-hum speech at Morehouse was somehow on a similar level to Trump's racist screed of an interview at NABJ.
"[M]any blacks may not feel that Harris is naturally one of them."
Tales pulled from your ass.
I don’t know how black voters will feel about it. I don’t know if it will hurt him with any particular demographic. I just think it hurts him across the board for him and his surrogates to keep posting family photos of her in traditional Indian clothes like that should embarrass her in some way. The one he posted on Truth Social had the caption, “She’s always been Indian.” The more I think about it maybe that is the real strategy. Back to the Obama birther rhetoric…”she’s foreign.” Again, that only resonates with a portion of his base like the ones here in rural East Texas. It cannot help him with new votes.
What do the people who wouldn’t vote for her because “she’s foreign” think about Vance’s wife? It seems like she would give them pause. (But maybe not, considering that never seemed to affect Mitch McConnell.)
I think ppl are going to hear about these Trump comments and not even know what the Biden comments are that you’re referring to. Seeing this as a strategy to shore up a portion of the black vote seems pretty out there to me.
Naturally - Which one of them?:
https://www.science.org/content/article/genetic-study-reveals-surprising-ancestry-many-americans
Yes. She will have a problem with the black male. They are still not ready for a female president and they do question her ethnicity.
What do you mean about questioning her ethnicity? I’ve never heard that before and we know who her parents are so not really any mystery.
There’s plenty of Republicans who find Harris as undesirable as you find Trump. They’ll vote for him because he’s not Harris.
Obviously. He’s had a solid base since day 1. There’s nothing he could do to lose their votes. Some of those people are my family members, friends, and work friends. I know all about it. I live in rural Northeast Texas.
Would love to see how race projections change if she chooses Josh Shapiro.
That was exactly my question. The probability map as presented makes Shapiro an appealing choice.
I think that if she goes with Shapiro that PA is out of play. I know that VP’s generally do not matter, but I think he would be the exception.
If she goes with Shapiro she probably loses Michigan though. Shapiro's extreme views on Israel would not go well there
Are his views extreme, or is it that he’s Jewish? Be honest.
"Extreme" by whose standards? There remains a convincing majority of Americans who support Israel over Hamas. That probably also includes a majority of the subset of Democrats who will actually show up at the polls in November (i.e. not the much-hyped but ever-elusive "youth vote").
Would choosing Shapiro as her running mate lose her Michigan? Perhaps, but I think that would be offset by other purple states (and not just Pennsylvania) being brought into her column that she otherwise would have lost.
Ya i dont buy that either. Shapiro condemned Bibi and will make clear hes better than alt.
I think it is much ado about nothing. Its a pretty mainstream viewpoint on protestors and Israel. I think Harris has shown she's threading the needle on it and I expect, given his comments, they will continue to draw that distinction between Israel's rights to defend themselves and what they are currently doing in Gaza.
Are Shapiro's views on Israel materially different from Kelly, Beshear, or Walz?
The vast majority of Americans are fine with helping the good guys pummel the shit out of the bad guys.
Has anyone done a meaningful analysis showing his comments on Israel are to the right of Kelly or Beshear?
https://penncapital-star.com/government-politics/womens-group-says-kamala-harris-should-consider-how-pennsylvania-josh-shapiro-handled-2023-sexual-harassment-complaint-against-aide/
The evidence there is really thin tbh, unless a mainstream news outlet substantiates this more I don’t think this should be a concern with picking Shapiro
Any major news outlets?
Media is reporting sexual harassment in his office. Someone was paid off to keep quiet. Sorry Shapiro is out.
Man if that was disqualifying, I'd hate to show you the Republican ticket...
If someone gave me better than 1/100 odds, I would bet on a tie. I think she will take PA, WI, and MI, but then I think she will lose in Nebraska 2nd, which would be a tie if there were no other surprises.
The model is currently giving her better chance in NE2 than any of PA, WI, and MI. Let's stick to methodology and data, rather than any feelings we might have :)
That would be a compelling bet at 200:1 but no one is gonna give you that.
Nate just recently wrote about this, the model will probably give Harris like a .2% overall win chance based on Shapiro. But of course it’s possible she could see a more measurable polling bump in PA because of him which would then factor in. (Fingers crossed?)
It'll be interesting to see how it changes regardless of who she chooses. Shapiro may be the best choice, though, or maybe Andy Beshear.
The model would wait for polling, but as to the question of what would happen to polling, it does subjectively feel like a case where VP pick will matter more than usual.
Why? I guess because Kamala *is* a VP, and because her VP pick will almost certainly be of a different gender and race and those things are not meaningless. In '08 Biden helped Obama on the margins, surely.
polymarket folks seem to put PA further in Harris' direction; I was wondering if they were thinking the same thing, about Shapiro.
Yep. Polymarket is already pircing in Shapiro.
Can you explain a bit the tail probability that leads to the model showing a 5.3% (as of the 8/1 snapshot) chance of Kennedy getting at least one EV? Other than maybe ME-2, I can't see how he wins any of the split EVs, so otherwise we are talking about him pulling a plurality win in a state.
Or is the model also allow for "rogue electors"?
[edit to add]
For the 5.3% of simulation runs where RFKjr wins at least one EV, what is the distribution of his nationwide PV %?
Third party candidates often have one state where they do unusually well. The model knows that. While at the 5.3% probability level, it only has Kennedy getting 4.5% of the national popular vote, if that is heavily concentrated in one small state it could be 51% of the vote there. It's a crazy scenario, but the model knows that crazy scenarios do sometimes happen.
I should note that Kennedy could only need to win 33.4% actually, going with the not-too-crazy assumption that in this unlikely hypothetical scenario that the remaining balance of votes in that state were evenly split between Harris and Trump.
yep, as noted in my reply to Thomas Dalton.
possibly actually a bit less than 33.4% allowing a couple of percent for 4th and 5th parties (libertarian, green, etc.)
I decided not to dive into that because of the complexities of which candidates Kennedy would be drawing votes from to get to his plurality. I should have just said "plurality".
Yes, and RFKjr only needs a plurality, not 51% (or he could win via RCV-IRV in ME-02 or ME-statewide via being a second choice of both Harris and Trump voters)
I agree that crazy things can and do happen; 5% seems *enormously high, hence my asking about the underlying distribution.
Note that Gary Johnson in 2016 (with a party apparatus behind him and 50 state ballot eligibility, neither of which RFKjr has so far) wound up getting 3.28% nationwide, not that much behind where the model thinks RFKjr is "polling" now. Other than New Mexico, where Johnson got 9.34%, his best state was NoDak at slightly less than 2x his national final vote result.
Assume that fourth and fifth parties wind up getting a few percent in the state/CD that RFKjr winds up winning (yes, I know the 5.3% is the probability over all states plus WDC and the NE and ME CDs), and that Harris and Trump are basically dead even. Other than RCV-IRV in ME-02, that means RFKjr has to get a plurality. Let's call the plurality a mere 30% (see fourth and fifth parties - I have no idea for example where Jill Stein and the Greens are for 2024). With RFKjr coming in with a tailwind somehow and getting 6% nationwide, he still has to score 5x his national vote percentage in the relevant state/CD.
AFAICT that kind of concentration hasn't happened in the 20th C - the closest I can find is 1968 when Wallace got very close to 5x his nationwide 13.5% in AL and MS. LaFollette got about a bit over 3x his nationwide 16.6% to win Wisconsin in 1924.
He still needs to get in the ~30% range to win the IRV races, because he needs to beat one of Harris or Trump into second place, and he also needs the one he doesn't beat to fall decently far short of 50% so he can overtake them in the runoff.
Agreed. Best case for an EV is about 30% with a lot of minor party votes and a reasonably even split between Trump and Harris (I can easily see Trump voters in an IRV race picking RFKjr for second preference; I'm not sure what Harris voters would do - they have a bigger coordination problem).
Which state is his support heavily concentrated in? Seems like there would likely be some sign already if there was a strong state in particular for him.
If there were already signs of him having strong support somewhere, the probability would be much higher than 5%.
If there was a state where he was regularly polling at 15% while the major candidates are both in the upper 30s, I don't think that would be enough to be giving him much higher than 5% chance, would it?
I think that would give him a much higher than 5% chance - 3rd party candidates are hard to poll (a lot of polls don't even include them), so one doubling their expected votes doesn't sound that unlikely to me (especially when you allow for how much can happen between how and election day).
If you take his current probabilities for each state (etc) and assume they are all independent, you get a probability of him winning at least one seat of 10.1%. They obviously aren't independent, which is why the model is actually giving him a lower probability. His highest probability in any state is Alaska where he has 1.4%. There don't seem to be any polls for Alaska since Harris entered the race, so I think the model is guessing quite a lot there...
Agreed that third party vote polling and vote projection should be considered very very noisy (such thin historical data, all but 2016 effectively pre-social medio too). Makes complete sense that AK would be the best single state according to the model (historically high elasticity, right up there with NH, in particular)
Also, and I don't think the model takes this into account (nor _should_ the model it get that far into the weeds), according to a claimed capture from the RFKjr campaign website about a week ago (no URL given and I could not find the specific data when I just looked), RFKjr does not have ballot access in ME, which (ME-02) is his best chance at getting an EV via RCV-IRV rather than having to have a plurality.
It's the probability to win at least 1 EV not exactly 1 EV, eg it's the probability that he wins at least one of the states or districts. 5% seems about right
yes, I know that :)
5% seems very high to me (e.g. extrapolating from Perot or 2016), hence wondering what the prior distribution the model uses was.
I'd assume that if the election were tomorrow, the odds would be a lot smaller of Kennedy capturing an EV. Some decent portion of the chance, in other words, represents scenarios where the race shifts in his favor between now and then...
oh, very definitely. That's why I am curious what the distribution of his nationwide PV is for the 5.3% of total scenarios wherein he wins at least one EV.
Looking at strong third party finishes post WW2, so mostly Wallace 68, Perot 92, and Johnson 16, it looks like the plausible best case regional concentration, ignoring special factors applying to Wallace 68, is 3x nationwide vote, and even that is stretching it for RFKjr. So yeah, with RFKjr winning say 12% nationally he *might* get enough concentration to squeeze out a plurality win, but looking at e.g. Johnson 2016 I suspect RFKjr needs to hit at least 15% nationally, getting pretty close to Perot 1992 territory (19% nationwide). Perot came pretty close to winning ME-02, losing to Clinton 38-33, so yeah, conditional on RFKjr getting well above 15% nationally one or more EVs is pretty plausible.
Well, based on today's model update, the actual answer might have been more like "the model just isn't that confident that really weird outcomes aren't more likely than we think due to small sample size of the historical polling data." Chance of RFK winning an electoral vote is way down after the model tweak:
"This model run also introduces a small technical change that somewhat reduces the instance of outlier outcomes in individual states, but you really have to squint to see any effect. The most notable difference is that the model now gives RFK Jr. roughly a 2 percent chance of winning at least one electoral vote, down from 5 percent before"
Remember that 5% is a common one-tailed test demarcation of statistically significant.
Don't forget George Wallace! Between him and Perot coming quite close in 1992, it actually happens a bit more than you might think.
I addressed both those, and I agree that _conditional on_ RFKjr getting into the 15% range nationwide (getting close-ish to Perot 1992, who got 18.9%) it would not be particularly surprising (I don't see how RFKjr achieves the level of regional concentration that Wallace had).
Hence if the model thinks RFKjr has a 5% chance of getting more than 15% nationwide, I agree re EVs and I change to being curious about the distributions that have RFKjr pulling a Perot 92 level nationwide vote.
I still don't see it, although in the model's defense :) there are so few examples of serious** third party runs that the underlying variance has to be considered high.
** I'm not sure whether RFKjr counts as a serious run in the sense that Perot was.
But RFK is no Perot or Wallace! He’s not riding a nationwide wave of people who think the two main parties are the same, and he’s not someone with a strong regional base. What’s his upside story that makes him more likely to get an elector than Jill Stein or Ralph Nader or Ron Paul?
Yes, you and I know the difference! But I imagine the model does not.
The model has seen enough polling data to know the difference!
One of their pieces does explain this. I wish I could remember which, but maybe a comment from Eli on the main model post
Hmm, I _did_ look before posting, but somehow I missed what you describe. I was thinking of emailing Eli; now I will do so and plead that he point me to the "already addressed here" comment.
If you haven't seen it, the model was updated today to reduce the probability of extreme events in individual states, so the probability of Kennedy getting at least one EV is now only 2.0%. It seems they had unrealistically fat tails in the distribution for each state - there was a bit of discussion about that in yesterday's article.
Model (and constitution!) does not allow for rogue electors. Heaven forbid that ever happens in a way that contradicts the outcome of the vote.
The US Constitution (which has not been amended recently) has nothing directly to say about rogue, aka faithless, electors. I *think* the state level faithless elector "laws" are state statutes, not state Constitutional provisions, but I could easily be wrong about that.
Chiafalo v. Washington (where it was a Washington state statute in question) says that a state *may* punish a faithless elector, not that the states *must* punish them. Nor AFAIK does Chaifalo directly speak to the recognition of a such an electoral vote sneakily, or cheekily :), offered in the face of punishment.
AFAIK Ray v Blair (1952) is still good law, explicitly allowing states (or in the specific case a political party) to require a pledge, but not addressing enforcement of the pledge. See also Justice Jackson's dissent in that case.
Kamala doesn’t even have to campaign anymore, just let Trump and Vance keep saying stupid things (especially Vance). If Trump loses he only has himself to blame at this point.
It will be funny to watch him blame his loss on everyone and everything other than himself tho lol
I don't think he'll go to jail. But maybe house arrest at Mar a Loco?
270 electoral votes is not “one more than she needs to win”, it’s exactly what she needs to win.
Just signed up. Would be nice if you took Paypal so I don't have to sprinkle my credit card numbers all over the web
Try privacy dot com (I’m not sponsored by them or anything)
Pen15
Pen island.
I can do it too.
So is Polymarket just going to track Nate's model from here in? Predicit went and flipped yesterday.
PolyMarket normally prices slightly more in favor of the Republican candidate than 538 models - there’s a slightly baked in pro-GOP bias there.
Trump emphasising Kamala's Indian heritage seems really odd, but it looks like a deliberate strategy. Can anyone explain it?
Is it just a thowaway comment that got out of hand or is there some reason why they think this will help reduce her lead among Black voters?
I just think he runs his mouth without a plan a lot.
That is certainly true, but this seems organised his supporters have lines to take and a coordinated message.
My take is it can’t hurt - anyone who is Trump-curious to begin with is not going to be offended, and maybe some buy “she is not who she says she is.”
Black men. Trump does pretty well among men. Black male unemployment remains higher than black female unemployment, so Trump's economic message resonates better. And the gap between black men and black women is growing - DEI in the spotlight. If some black men feel less racial affinity for Harris, they may vote on sex before race.
Also, isn't this all based on essentially one poll (Morning Consult) that had crazy results for MI?
Oh, so you did understand what I was saying about the MC poll (which was good for Harris in more than just MI, which I also explained) being important. You were just bad faith playing dumb to rile me up. It’s like you are a character written by Sartre himself to demonstrate his point about people with evil views. And now, I assume, you will begin doing the same thing to anyone who bites here.
Gotta love intellectual dishonesty.
No single poll is going to move the model significantly.
The MC poll had an absolutely crazy result in MI. I have to wonder if the magnitude of the change there is having an outsize effect on the model's output.
You don't have to wonder. Nate knows how to construct an election model. It does not overreact to single polls.
https://www.realclearpolling.com/polls/president/general/2024/michigan/trump-vs-harris
What's the other polling in MI that shows Harris with a lead there?
There is a Glengariff Group poll with a very slight Harris lead. And a Public Opinion Strategies poll and a Civiqs poll with them equal. There are only 9 polls total for MI, though, so the model will be heavily relying on national polls and polls from states with similar demographics.
In other words a bunch of polls showing a tie and one massive outlier. I would be very careful if that Morning Consult poll is thrown into the mix.
Dow Jones drops 700 points as PMI drops below 50 and initial jobless claims reach levels not seen in a year. When does the "fundamentals" portion of the model get updated?
The question is whether Harris gets the blame for current conditions in the country. She is after all part of the current administration.
EDIT: I'm not talking about the drop in the market being significant, I'm talking about WHY the market dropped. Namely PMI missed big.
You're right—the market had a one day decline. It had an all-time high—checks notes—two weeks ago. On news this bad, Harris is definitely toast! (Over in the land of Copium)
Did you miss the part about PMI? I'm not talking about the stock market going up and down, I'm talking about a recession.
You are as wrong about this as you are about politics:
-The strong productivity growth (2.3% vs 1.7% expected) coupled with lower unit labor costs (0.9% vs 1.8% expected) suggests that economic growth in Q2 2024 was organic and not driven by inflation. This indicates a healthier economic environment, not an economy on the verge of recession.
-the combination of strong productivity, lower labor costs, and modest manufacturing slowdown suggests the economy may be heading towards a "soft landing", which continues to be the majority view
-Both the PMI (49.6) and ISM (46.8) manufacturing indices show contraction in the manufacturing sector, but nothing indicative of an imminent recession
-we're still on a glide path for rate cuts
-The mix of strong productivity growth with signs of economic slowing suggests a more balanced and potentially sustainable growth pattern, which could help avoid boom-bust cycles that often lead to recessions.
"ISM" stands for "Institute for Supply Management". It's the organization that calculates PMI. It's not a measure itself. PMI for July was 46.8.
I think we're done here.
This is an exact quote from a financial website: "Meanwhile, manufacturing data from the PMI (S&P Global) showed a slightly better read than expected, at 49.6 vs. 49.5. This is the first reading under 50 since January and may be indicating the economy is slowing down from the brisk pace in Q4 2023.
The ISM manufacturing number for July was also released, and showed a 46.8 reading, which was less than expected (48.8)."
Don’t bother. He’s not trying to have an actual discussion. He’s willfully misinterpreting the data to get smart people to tie themselves in knots. There is a famous Sartre you’re about people like him.
Quoting something is meaningless if you don't understand what is actually being said.
Yeah, we're done here.
What recession? If you want to be taken seriously and not as a troll, you are doing it wrong.
Open a browser and go to cnbc or something.
No, you open a browser and go to cnbc or something. There is no recession. End of story.
If you're so certain of that why don't you go to cnbc?
There has been open discussion in the economic/business media that the economy might slip into recession by the election. You have missed this? It isn't a conspiracy though doesn't look likely.
If you took the time to backtest, then you would know the PMI 50 threshold is a poor leading indicator of recession - not much better than chance. We certainly could still have a recession, but the PMI movement has little-to-nothing to do with it.
That is not my argument. I suggest looking at the totality of indicators (such as the Sahm rule) for agreement across measurements to try and gauge the direction of the economy.
So let me get this straight: according to you, that disaster show Trump put out yesterday is irrelevant, but daily fluctuation of stock market of 1.5% is super important?
You can tell who doesn't watch the financial news...
Yes I can: the vast majority of relevant electorate.
That's not going to make any difference. Do you think recessions don't happen if nobody watches the news?
You think there's going to be a recession in the next 3 months? If so, you don't even know what a recession is.
Open up a browser and go to cnbc.
The stock market dropped a little over 1% based on slightly poor results in the manufacturing sector for one month, which represents about 20% of the US economy. There is nothing in that that suggests an imminent recession.
Open up a browser and go to cnbc.com
I have done. There is no headline saying "Recession on the way". It just reports the numbers - a small fall in the stock market after slightly disappointing manufacturing numbers. You keep telling people to read CNBC, but it doesn't support what you are claiming. You are forming your own interpretation based on what it says and that interpretation is fundamentally flawed.
The banner on cnbc right now reads something like "PMI reawakens investors recession fears". Their top article goes into more detail.
EDIT: It just got bumped because Amazon missed sales targets for Q2.
That's sensationalist headline writing. You have to actually read the article. It's just people adjusting their predictions very slightly in a downward direction.
What is ambiguous about "reawakens recession fears"?
Oh my god are you going to do this the whole fucking election?
Trolls got to troll. MAGAts got to be MAGAts, unfortunately. :(
It’s a shame trolls like this can’t be blocked here. It diminishes the ability for everyone else to have reasonable discussions. I wonder if Nate can do anything about it? The deluge of repetitive, agenda-pushing posts is what ruins social media.
Contrary opinions are fine, but can you just posts whatever nonsense you want once per article and be done with it?
I did block him. He stills show up. It’s very weird.
Oh, I guess I didn’t look very hard. Thanks for the tip that it is possible!
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
Here's a question: what do you think is the majority opinion of economists about whether an imminent recession is likely? Hint: it is different from yours.
Are you asking them today or yesterday?
Please go ahead and show me the list of prominent economists who _today_ are saying that an imminent recession is likely
My point is that I am pretty sure reasonable economists change their mind based on data. Like the kind we had today.
So no specifics economists, just your own speculations
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
“Infidelity does not consist in believing, or in disbelieving; it consists in professing to believe what he does not believe. It is impossible to calculate the moral mischief, if I may so express it, that mental lying has produced in society. When a man has so far corrupted and prostituted the chastity of his mind, as to subscribe his professional belief to things he does not believe, he has prepared himself for the commission of every other crime.”
- Thomas Paine, The Age of Reason
Go to cnbc.com and read a few articles. Then come back here and post.
I did go to cnbc.com and searched
Keyword recession and amazingly at least twice a month, every month there have been articles stating that we are heading into a recession.
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
I could read everything in their backlog. It won't change the fact that you are obviously an intellectually dishonest troll.
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
Is the Trump/Vance campaign paying you by the post or by the word? Just curious . . .
https://www.cnbc.com/2024/08/01/stock-market-today-live-updates.html
I think this is a positive as it makes it more likely the Fed cuts interest rates.
The Fed cuts based on lagging indicators. If they cut it will be too late to stave off a recession.
Bill Ackman, for example, wanted cuts at the start of the year.
Agreed on this point. If I was Jerome, I would have begun moving rates down two meetings ago. They need to be looking at the 2nd derivative of the economy and not just the 1st.
You are in essence arguing "The trend is your friend" when you make statements like that. And if the economy is not on the brink of recession there's no need for premature cuts that run the risk of reigniting inflation.
Going off of the explicit and implicit goals of the Fed, which are to lower inflation to (a completely arbitrary) 2% in a way that leads to a “soft landing” and does not put the economy into a tailspin. I am suggesting that acceleration of the economy would be beneficial to consider in combination with the velocity of the economy instead of the velocity by itself. It would be a mistake to assume that the only way to justify cuts would be if we were on the brink of recession. That is simply not how monetary policy works.
Au contracted, that is precisely how the Fed works. The historical record is clear: the Fed always cuts late. I suspect that the primary consideration is that taking the foot off the brake too soon runs the risk of reigniting inflation.
Again, a less risk averse approach that prioritized a "soft landing" over taming inflation with no potential for resurgence would have led to cuts at the start of the year.
Two hundred and seventy is exactly what Harris needs. With 269, she will lose according to the complicated procedure of elections in the House of Representatives in case of a tie. It’s not one elector more than needed, as you wrote.
Actually the next Congress would decide, who havent been elected yet.
Odds are more states have a Republican majority than a Democratic majority in terms of House members. Remember, the threshold is 26 states, not a majority of House votes.
It is highly unlikely that 26 states will vote for her in the House. However, Trump might not secure them either. In this case if a Democratic miracle happens in the Senate, Shapiro or Kelly will get the support of 51 Senators and become the acting president.
Who knows what the next House delegation will look like?
https://centerforpolitics.org/crystalball/trumps-contingent-cy-plan/
I’m a bit confused. The article says it’s a toss up but when I look at the model Kamala is in the 30s. Is it possible I need to force some kind of a refresh?
I see 44.6% for Kamala right now. It was 42% yesterday, and 38% the day before.
I’m having this issue too. It updates when I’m in a browser, but is stuck on the numbers from a few days ago in the app.
It's a known thing that the app doesn't update the numbers. You need to visit the browser page.
And that’s after the Reep convention and before the Dem convention bounce!
The model knows that, though, so that's taken into account. It is projecting a fall in Trump's polling numbers after his convention bounce wears off.
How is harris getting so little cred in georgia? She has a better shot at georgia than Biden did in 2020
Probably a paucity of quality polling.
Interesting. I think Vivek would have helped Trump a lot compared to his pick.