Those elections last week may nothing one thing they do mean is that that socialist anti semite who became the mayor in New York will be an albatross around every Democrat in the midterm elections, and the Republicans will control the house and the Senate and of course the White House for another two years
One year. Two years from now there's a decent (but not overwhelming) probability of a Democratic majority in the House and a razor-thin Senate majority for the GOP. And holding 53 seats in the Senate is very different from holding 51 or 50. Especially if Trump continues to shoot himself in the foot.
Can't disagree with much of this. The only structural points I would add---and they are very significant---is that as deportations continue (already about 2m total), a Ds will lose a significant number of voters (ahem) from those. Second, as voter roll purges continue, that also will continue to shift things. To my knowledge, required voter roll purges have NOT yet taken place in two big CA counties, Orange and Los Angeles. Meanwhile, there has not been any sign yet of a voter registration shift slowdown working for Rs. Last week NC saw the D lead of 175,000 in 2020 fall to 6,000; PA fell again; NM showed a nearly 6,000 R gain/D loss since September. Seth Keshel ("Capt'n K's Corner") had a nice rundown last week of all these changes.
Note that, per your calculations, R's gained 'net' registrations in NJ and VA since the last set of comparable elections. Yet they performed worse. Ciattarelli's vote share was 4.8 points lower than it was in the last gubernatorial election in 2021, even with the shifts in registration. Similar trends emerged in the VA elections. According to your methodology we should have seen a directional shift against Democrats, even given off-year overperformance--yet we saw the opposite. Clearly you are overstating the predictive power of your 'net registration' metric when it comes to electoral outcomes, given what we've seen since 2024.
The Republicans were out of power in 2021. Off-year elections favor high-propensity voters (Democrats) and the party out of power (Republicans in 2021, Democrats in 2025).
Anything else is just wishcasting. We could each add a half dozen factors to the pile that seem plausible, but all it goes to is Dan Kahan’s point that intelligent people are *more* prone to motivated innumeracy than less intelligent people. See also: the left’s belief in 2023 special elections and the Selzer poll.
Even were I to grant all of that, (I don’t) it would really only reinforce my point, since Larry has been consistent in treating his net registration gain calculations not only as electorally deterministic, but deterministic in a linear way. (Thence his weekly apocalyptic prognostications of Democratic Party doomsday) Clearly that is incorrect—else we should be seeing a deterioration in relative Democratic performance here—and suggests it may indeed best be read as the motivated enumeration I pegged it as from the get-go.
Voter registration is also objective data, unlike polls which are essentially worthless. But a slow long-term trend towards Republican registrations won't mean much from year to year. OTOH, if the Democrats can't figure out a way to reverse the trend, it will result in the party eventually cracking.
Even assuming those are indeed the only species of data worth consuming, the problem here is the assumption of some strong linear relationship between the two. That’s the baseline assumption of Larry’s whole line of argumentation. Thence all that talk of R gains in New Jersey being the ‘canary in the coal mine’ for Democrats, and his dismissive attitude towards the huge net gains of independent voters versus both major parties, because ‘it doesn’t matter, Rs are gaining more’.
If the argument is ‘both of those variables are related, but in a non-linear, weakly correlated way, wherein you can’t really effectively predict voting trends—and thence electoral outcomes—from changes in net registration, even on a ‘directional’ basis’, then these election results don’t pose a problem for you. If, like Larry, it’s instead ‘you can infer the Democrats are electorally doomed from net voter registration trends for the past five years, since changes in votes cast and electoral outcomes are a function of marginal changes in registration’, then they do.
The changes in voter trends can be swamped by the economy and other factors. So in that sense, no it's not a strong linear relationship. (I don't think Larry ever claimed it was?) Rather the baseline for what Republicans need given the economy (and off-year elections) is slowly but surely moving in their favor.
Also note that the rise of independents is not a good sign for the left. Some of them are like me and don't want to get fired from their jobs, so they register as independents even though they're really Republicans. Others remain loyal to the Democrats, but it is a sign that the Democratic brand is increasingly toxic.
Man, honestly, I am convinced that you do not have a job. You post the same half-baked gibberish to every single article, no matter the topic.
There is absolutely no evidence that deportations will cost Dems votes. First off, your premise assumes (without any actual evidence) that those being deported even voted at all. If they did vote, then that suggests that they are citizens, which means deporting them is illegal. I either case, it’s an ignorant statement with no basis in fact. Plus, Trump won in part because he received a large percentage of Latino votes. You aren’t seriously suggesting that those Latinos who voted for Trump are illegal immigrants, are you?
The election results in VA were very predictable. VA federal and NGO employment are the equivalent of oil in Texas. Many people effected personally, and many more downstream. NJ appears to be a combination of Trump loathing and the inability of Reps to discuss mass migration.
The election does not, however, resolve what should happen to the roughly 9 million migrants who entered the US without a valid asylum claim. Statistically speaking, 90% of Biden's 10 million will not qualify for asylum. They were not fleeing war or a natural disaster. Neither poverty, domestic abuse or crime has ever been the basis for a valid US asylum. Under current law the vast majority of Biden's migrants will eventually lose asylum claims, years after they have become entrenched in the US, given birth to US citizens and 1/2 have enjoyed perpetual tax payer subsidizes.
Dems have turned their Achilles Heel into a sword, but Reps will likely figure out a better way to message the subject. The most effective would be to remind Americans nothing prevents Dems from following the last 10 million unexpected Dem new arrivals, with 10, 20, 30 or 40 million more, when Dems retake the WH.
It certainly would not be difficult. The WH controls tourist visas. Every year the US issues between 10-11 million. And every year 300K-400K visa holders never leave the US. The "overstays" often purchase new identities, illegally "emigrating" in a quick and cheap manner.
The duration of tourist visas are generally capped at 6 months, but the annual number issued is not limited. That means if Dems decided to inform the world, 30 million US tourist visas will now be issued annually, fist come first serve, Reps could not block them. Nor could Reps do anything when Dems ignore a 33% overstay rate, rather than the current 3% or 4% rate.
Forget Biden's 10 million migrants in 4 years. Imagine 10 million a year, for 4 years, arriving in stealth. The removal of that many people would be logistically impossible, even if the majority of Americans would support deportations.
A lot has been made in the media about this recent election. Other than the news reaction, I don't think it means much, other than possibly the NYC mayor.
We already know, those that don't like Trump, really really don't like Trump. That explains the higher than normal turnout and media reaction for such an insignificant event.
The Governors elections were in states that would normally be expected to go Democrat. No real surprise except for the media speculation the NJ or Virginia might be a surprise and go Republican. They were simply wrong.
In most other places there were no significant offices being decided this year. Where I live, the only offices on the ballot were for Public Service Commission where everyone was a complete unknown.
Next year should be a better indication but this year will be soon forgotten, with one exception, NYC Mayor.
The truth is, the voters had such poor choices on the ballot. Sliwa, simply not a realistic candidate but the Republicans have zero chance there anyway.
Cuomo? He was the, hold your nose and vote candidate, since he was the only other option.
He should have been a defendant, not a real candidate, but this is New York and Trumps last minute, reluctant endorsement was really a reverse psychology move that likely hurt Cuomo more than helped.
A total wild card, but should be interesting to see if anyone without any real management experience (even forgetting the controversial aspects) can manage such a large and complex operation better than the corrupt and often inept, previous management.
Then, add in the controversial stuff.................
I guess even in 26 the midterms will only give us an indication of what the midterm vote is like, it seems presidential years are a species unto themselves.
I wish I were privy into the kinds of statistics both parties are probably getting about the Hispanic vote in both places. I'd like to know what percentage of the NJ and VA Hispanic vote is of Puerto Rican vs Mexican origin for instance. I personally know people who have been deported now, I'm sure Hispanic voters also know deportees. The issue has been getting lots of negative coverage and it must be a factor in turnout and voting.
I suspect deporting illegal immigrants is only having a positive impact on the Hispanic vote because legal immigrants want affordable housing and good-paying jobs. Keep in mind the lesson of Starr county, which is right on the Texas border with Mexico. It is 97% Hispanic.
I don't have the numbers either, but if you look at Puerto Rican history it has always leaned a bit further left than Mexico. My family's also Puerto Rican and from personal experience (which I'll grant is not a scientific measure of anything) those of them that were mildly pro-Trump (or rather just anti-incumbent) in 2024, are definitely anti-Trump now, although not necessarily enthused by Democrats.
Great article, as expected from Mr. Bahareen and whenever the TLP drills down into the subtleties of the numbers.
I said the turnout margin, if substantial, would bode well for Democratic gains in 2026, and I'll stand by that prediction. Ceteris Paribus, 2026 will probably be decent-to-very-good for the Dems. 2028 will be determined by the economy and is anyone's game. The Supreme Court may well do Trump a big favor there by preventing him from continuing his chaotic tariff policies, and giving the nation a massive tax cut if they are lifted.
But here's the rub: if the opposition's turnout is already high, a rational politician would try to avoid doing things that would incentivize it to increase even more. To the disbenefit of Republicans, Trump is a singularly irrational politician, and will probably continue to engage in siege warfare tactics that will provoke even *higher* Democratic turnout. That irrationality will also prevent him from making effective adjustments that would inspire the whole of his 2024 coalition to come back home. That's the other trend at work here that isn't talked about very much: Yes, it's true the Democrats may be in danger of continuing to narrow their coalition. But Republicans aren't doing anything to widen theirs, either. That cuts against their coalitional structural advantage over the Democrats. Two stagnant coalitions + economic problems + high Dem turnout = 2028 Democratic blowout.
Reagan built the long-running post-1970s Republican majority by widening his coalition and converting Democrats. Remember the term 'Reagan Democrat'? I know quite a few. We don't see anything like that going on here.
Remember: Trump won't be here in 2028. The Republican Party has no identity without him, and the electoral results for the GOP when Trump isn't on the ballot aren't the kind of results that scream 'permanent GOP majority'.
One thing to remember about turnout. The better the turnout the better the chances for Republicans. The current elections last week were, as Ruy said, a low turnout situation. Studies have shown that a majority of voters who stayed home in the Presidential election were Trump supporters.
I was talking about the *opposition's* turnout, i.e. Democratic turnout.
The key variable actually isn't total turnout, it's turnout *differential*. I.e., what is the difference between the degree to which Democrats turn out vs. the degree to which Republicans turn out to vote. High differentials in opposition voting can undercut a party's advantage in the overall distribution of registered voters.
To use an extreme example just to illustrate the point, consider an election that occurs in a state or nation that polls at (or has a registration advantage of) 48% Democrat, 50% Republican. Now let's say A.) Overall turnout is high, but B.) in this high turnout election, the number of Democrats that turn out is four times higher than the number of Republicans that go to the polls.
In that scenario, Democrats will without a doubt win said election. That's why it would be better for Republicans to govern in ways that will incentivize *low* turnout amongst Democrats, or--God forbid--in ways that will win some Democrats over. They're currently not doing that.
Lest we forget, the 2024 election was the 12th-closest in history. Democratic permabears like Ruy don't talk about that--but given that fact I wouldn't take anything for granted if I were a Republican, despite some of the trends in the party's favor.
What we have found is that Ruy speaks to us. His analyses are sharp and are data based, which turns us on. He has also shown a great ability to have predictions, which is unusual.
So, you and we might have some disagreements, but that OK!
I agree that he has his place. Permabears--like permabulls--are often wrong, but they are also often right, and will make points that the permabulls will miss. (And vice versa) Both are worth listening to, agree or disagree, to broaden your perspective and sharpen your thinking. Goes for investing *and* for politics!
Excellent analysis! I think that the Democrats will misinterpret the 2025 election results and overplay their hand going forward. The current infighting over ending the government shutdown shows how deeply divided the Democrats are, and those divisions will not be easily resolved or papered over in 2026 and 2028.
Nice analysis by Baharaeen. Surprised he made no assessment of the significance for the midterms when Trump will not be on the ballot and where is endorsement will be resisted in key swing districts. Will the VA and NJ phenomenon continue? Will MAGA candidates win Republican primaries? Trump's lame duck status will be solidified if House, and possibly the Senate, turn Democratic.
Those elections last week may nothing one thing they do mean is that that socialist anti semite who became the mayor in New York will be an albatross around every Democrat in the midterm elections, and the Republicans will control the house and the Senate and of course the White House for another two years
One year. Two years from now there's a decent (but not overwhelming) probability of a Democratic majority in the House and a razor-thin Senate majority for the GOP. And holding 53 seats in the Senate is very different from holding 51 or 50. Especially if Trump continues to shoot himself in the foot.
Don’t drink and post Bob, not a good look.
Can't disagree with much of this. The only structural points I would add---and they are very significant---is that as deportations continue (already about 2m total), a Ds will lose a significant number of voters (ahem) from those. Second, as voter roll purges continue, that also will continue to shift things. To my knowledge, required voter roll purges have NOT yet taken place in two big CA counties, Orange and Los Angeles. Meanwhile, there has not been any sign yet of a voter registration shift slowdown working for Rs. Last week NC saw the D lead of 175,000 in 2020 fall to 6,000; PA fell again; NM showed a nearly 6,000 R gain/D loss since September. Seth Keshel ("Capt'n K's Corner") had a nice rundown last week of all these changes.
Note that, per your calculations, R's gained 'net' registrations in NJ and VA since the last set of comparable elections. Yet they performed worse. Ciattarelli's vote share was 4.8 points lower than it was in the last gubernatorial election in 2021, even with the shifts in registration. Similar trends emerged in the VA elections. According to your methodology we should have seen a directional shift against Democrats, even given off-year overperformance--yet we saw the opposite. Clearly you are overstating the predictive power of your 'net registration' metric when it comes to electoral outcomes, given what we've seen since 2024.
The Republicans were out of power in 2021. Off-year elections favor high-propensity voters (Democrats) and the party out of power (Republicans in 2021, Democrats in 2025).
Anything else is just wishcasting. We could each add a half dozen factors to the pile that seem plausible, but all it goes to is Dan Kahan’s point that intelligent people are *more* prone to motivated innumeracy than less intelligent people. See also: the left’s belief in 2023 special elections and the Selzer poll.
Even were I to grant all of that, (I don’t) it would really only reinforce my point, since Larry has been consistent in treating his net registration gain calculations not only as electorally deterministic, but deterministic in a linear way. (Thence his weekly apocalyptic prognostications of Democratic Party doomsday) Clearly that is incorrect—else we should be seeing a deterioration in relative Democratic performance here—and suggests it may indeed best be read as the motivated enumeration I pegged it as from the get-go.
Voter registration is also objective data, unlike polls which are essentially worthless. But a slow long-term trend towards Republican registrations won't mean much from year to year. OTOH, if the Democrats can't figure out a way to reverse the trend, it will result in the party eventually cracking.
The only worthwhile data on political trends is:
- how people vote
- how people are registered
Even assuming those are indeed the only species of data worth consuming, the problem here is the assumption of some strong linear relationship between the two. That’s the baseline assumption of Larry’s whole line of argumentation. Thence all that talk of R gains in New Jersey being the ‘canary in the coal mine’ for Democrats, and his dismissive attitude towards the huge net gains of independent voters versus both major parties, because ‘it doesn’t matter, Rs are gaining more’.
If the argument is ‘both of those variables are related, but in a non-linear, weakly correlated way, wherein you can’t really effectively predict voting trends—and thence electoral outcomes—from changes in net registration, even on a ‘directional’ basis’, then these election results don’t pose a problem for you. If, like Larry, it’s instead ‘you can infer the Democrats are electorally doomed from net voter registration trends for the past five years, since changes in votes cast and electoral outcomes are a function of marginal changes in registration’, then they do.
The changes in voter trends can be swamped by the economy and other factors. So in that sense, no it's not a strong linear relationship. (I don't think Larry ever claimed it was?) Rather the baseline for what Republicans need given the economy (and off-year elections) is slowly but surely moving in their favor.
Also note that the rise of independents is not a good sign for the left. Some of them are like me and don't want to get fired from their jobs, so they register as independents even though they're really Republicans. Others remain loyal to the Democrats, but it is a sign that the Democratic brand is increasingly toxic.
Man, honestly, I am convinced that you do not have a job. You post the same half-baked gibberish to every single article, no matter the topic.
There is absolutely no evidence that deportations will cost Dems votes. First off, your premise assumes (without any actual evidence) that those being deported even voted at all. If they did vote, then that suggests that they are citizens, which means deporting them is illegal. I either case, it’s an ignorant statement with no basis in fact. Plus, Trump won in part because he received a large percentage of Latino votes. You aren’t seriously suggesting that those Latinos who voted for Trump are illegal immigrants, are you?
The election results in VA were very predictable. VA federal and NGO employment are the equivalent of oil in Texas. Many people effected personally, and many more downstream. NJ appears to be a combination of Trump loathing and the inability of Reps to discuss mass migration.
The election does not, however, resolve what should happen to the roughly 9 million migrants who entered the US without a valid asylum claim. Statistically speaking, 90% of Biden's 10 million will not qualify for asylum. They were not fleeing war or a natural disaster. Neither poverty, domestic abuse or crime has ever been the basis for a valid US asylum. Under current law the vast majority of Biden's migrants will eventually lose asylum claims, years after they have become entrenched in the US, given birth to US citizens and 1/2 have enjoyed perpetual tax payer subsidizes.
Dems have turned their Achilles Heel into a sword, but Reps will likely figure out a better way to message the subject. The most effective would be to remind Americans nothing prevents Dems from following the last 10 million unexpected Dem new arrivals, with 10, 20, 30 or 40 million more, when Dems retake the WH.
It certainly would not be difficult. The WH controls tourist visas. Every year the US issues between 10-11 million. And every year 300K-400K visa holders never leave the US. The "overstays" often purchase new identities, illegally "emigrating" in a quick and cheap manner.
The duration of tourist visas are generally capped at 6 months, but the annual number issued is not limited. That means if Dems decided to inform the world, 30 million US tourist visas will now be issued annually, fist come first serve, Reps could not block them. Nor could Reps do anything when Dems ignore a 33% overstay rate, rather than the current 3% or 4% rate.
Forget Biden's 10 million migrants in 4 years. Imagine 10 million a year, for 4 years, arriving in stealth. The removal of that many people would be logistically impossible, even if the majority of Americans would support deportations.
Must have been a slow news day.
A lot has been made in the media about this recent election. Other than the news reaction, I don't think it means much, other than possibly the NYC mayor.
We already know, those that don't like Trump, really really don't like Trump. That explains the higher than normal turnout and media reaction for such an insignificant event.
The Governors elections were in states that would normally be expected to go Democrat. No real surprise except for the media speculation the NJ or Virginia might be a surprise and go Republican. They were simply wrong.
In most other places there were no significant offices being decided this year. Where I live, the only offices on the ballot were for Public Service Commission where everyone was a complete unknown.
Next year should be a better indication but this year will be soon forgotten, with one exception, NYC Mayor.
The truth is, the voters had such poor choices on the ballot. Sliwa, simply not a realistic candidate but the Republicans have zero chance there anyway.
Cuomo? He was the, hold your nose and vote candidate, since he was the only other option.
He should have been a defendant, not a real candidate, but this is New York and Trumps last minute, reluctant endorsement was really a reverse psychology move that likely hurt Cuomo more than helped.
A total wild card, but should be interesting to see if anyone without any real management experience (even forgetting the controversial aspects) can manage such a large and complex operation better than the corrupt and often inept, previous management.
Then, add in the controversial stuff.................
Thanks, I love all that wonky numbers stuff.
I guess even in 26 the midterms will only give us an indication of what the midterm vote is like, it seems presidential years are a species unto themselves.
I wish I were privy into the kinds of statistics both parties are probably getting about the Hispanic vote in both places. I'd like to know what percentage of the NJ and VA Hispanic vote is of Puerto Rican vs Mexican origin for instance. I personally know people who have been deported now, I'm sure Hispanic voters also know deportees. The issue has been getting lots of negative coverage and it must be a factor in turnout and voting.
I suspect deporting illegal immigrants is only having a positive impact on the Hispanic vote because legal immigrants want affordable housing and good-paying jobs. Keep in mind the lesson of Starr county, which is right on the Texas border with Mexico. It is 97% Hispanic.
2012: 86% Obama, 13% Romney
2016: 79% Clinton, 19% Trump
2020: 52% Clinton, 47% Trump
2024: 42% Clinton, 57% Trump
https://en.wikipedia.org/wiki/Starr_County,_Texas
I don't have the numbers either, but if you look at Puerto Rican history it has always leaned a bit further left than Mexico. My family's also Puerto Rican and from personal experience (which I'll grant is not a scientific measure of anything) those of them that were mildly pro-Trump (or rather just anti-incumbent) in 2024, are definitely anti-Trump now, although not necessarily enthused by Democrats.
Great article, as expected from Mr. Bahareen and whenever the TLP drills down into the subtleties of the numbers.
I said the turnout margin, if substantial, would bode well for Democratic gains in 2026, and I'll stand by that prediction. Ceteris Paribus, 2026 will probably be decent-to-very-good for the Dems. 2028 will be determined by the economy and is anyone's game. The Supreme Court may well do Trump a big favor there by preventing him from continuing his chaotic tariff policies, and giving the nation a massive tax cut if they are lifted.
But here's the rub: if the opposition's turnout is already high, a rational politician would try to avoid doing things that would incentivize it to increase even more. To the disbenefit of Republicans, Trump is a singularly irrational politician, and will probably continue to engage in siege warfare tactics that will provoke even *higher* Democratic turnout. That irrationality will also prevent him from making effective adjustments that would inspire the whole of his 2024 coalition to come back home. That's the other trend at work here that isn't talked about very much: Yes, it's true the Democrats may be in danger of continuing to narrow their coalition. But Republicans aren't doing anything to widen theirs, either. That cuts against their coalitional structural advantage over the Democrats. Two stagnant coalitions + economic problems + high Dem turnout = 2028 Democratic blowout.
Reagan built the long-running post-1970s Republican majority by widening his coalition and converting Democrats. Remember the term 'Reagan Democrat'? I know quite a few. We don't see anything like that going on here.
Remember: Trump won't be here in 2028. The Republican Party has no identity without him, and the electoral results for the GOP when Trump isn't on the ballot aren't the kind of results that scream 'permanent GOP majority'.
One thing to remember about turnout. The better the turnout the better the chances for Republicans. The current elections last week were, as Ruy said, a low turnout situation. Studies have shown that a majority of voters who stayed home in the Presidential election were Trump supporters.
I was talking about the *opposition's* turnout, i.e. Democratic turnout.
The key variable actually isn't total turnout, it's turnout *differential*. I.e., what is the difference between the degree to which Democrats turn out vs. the degree to which Republicans turn out to vote. High differentials in opposition voting can undercut a party's advantage in the overall distribution of registered voters.
To use an extreme example just to illustrate the point, consider an election that occurs in a state or nation that polls at (or has a registration advantage of) 48% Democrat, 50% Republican. Now let's say A.) Overall turnout is high, but B.) in this high turnout election, the number of Democrats that turn out is four times higher than the number of Republicans that go to the polls.
In that scenario, Democrats will without a doubt win said election. That's why it would be better for Republicans to govern in ways that will incentivize *low* turnout amongst Democrats, or--God forbid--in ways that will win some Democrats over. They're currently not doing that.
Lest we forget, the 2024 election was the 12th-closest in history. Democratic permabears like Ruy don't talk about that--but given that fact I wouldn't take anything for granted if I were a Republican, despite some of the trends in the party's favor.
I had to look up the term "permabears." :)
What we have found is that Ruy speaks to us. His analyses are sharp and are data based, which turns us on. He has also shown a great ability to have predictions, which is unusual.
So, you and we might have some disagreements, but that OK!
I agree that he has his place. Permabears--like permabulls--are often wrong, but they are also often right, and will make points that the permabulls will miss. (And vice versa) Both are worth listening to, agree or disagree, to broaden your perspective and sharpen your thinking. Goes for investing *and* for politics!
p.s. Thanks for a nice, respectful interchange.
Excellent analysis! I think that the Democrats will misinterpret the 2025 election results and overplay their hand going forward. The current infighting over ending the government shutdown shows how deeply divided the Democrats are, and those divisions will not be easily resolved or papered over in 2026 and 2028.
Nice analysis by Baharaeen. Surprised he made no assessment of the significance for the midterms when Trump will not be on the ballot and where is endorsement will be resisted in key swing districts. Will the VA and NJ phenomenon continue? Will MAGA candidates win Republican primaries? Trump's lame duck status will be solidified if House, and possibly the Senate, turn Democratic.