On the menu today: I put it off as long as I could, to account for any last-minute changes, but the story of the 2026 midterms is largely a static one. As I wrote back in late September, “When the president’s job approval is around 40 percent or below, and the national average price of gas is nearly $4.50 a gallon, and the national average price of diesel is $6.50 a gallon, the president’s party should expect to get crushed.” This morning, the president’s job approval rating is at 39 percent in the RealClearPolitics average, the national average price of gas is $4.15 a gallon, and the national average price of diesel is $6.29 a gallon — a slight improvement, but unlikely to be large enough for most consumers to notice. What kind of election results do you think Republicans will get in an economic environment like that? What kind of election results do they deserve in an economic environment like that?
I think by Wednesday morning, we will be saying, “That was a pretty darn rough midterm election for the Republicans.” Read on.
Brace Yourselves, Republicans
Back in the 2020 cycle, all the polls showed Maine Senator Susan Collins behind, often well behind, her Democratic rival Sara Gideon. Collins ended up winning by 8.6 percentage points.
Collins is the exception that proves the rule; if you have never or almost never led a poll in your race, you are unlikely to win when all the votes are counted. You might be surprised how many people find that to be a controversial statement.
I understand that a large segment of the public, and lots of Republicans, doubt public polling, in part because they remember the Ann Selzer poll showing Kamala Harris winning Iowa, or Francesca Hong leading the Wisconsin Democratic gubernatorial primary polling and losing, or Abdul El-Sayed enjoying a huge lead in Michigan senatorial primary polling and barely winning. But few remember all the primary or general election polling that turned out to be reasonably accurate, because the results were not surprising. (In 2024, InsiderAdvantage and Emerson showed Trump winning Iowa comfortably. No one talked about those polls much, because Iowa had been trending red, and almost everyone expected Trump to win Iowa.)
Or people remember the fact that pollsters Median Strategies and Wedgewood Polling were made up and completely faked their results. Those are darkly funny stories of fraud, demonstrating how the hunger for updates will get certain people to believe any numbers they hear, assuming they must be legit. But Marist, Siena, YouGov, Public Opinion Strategies, SurveyUSA, Ipsos, Emerson — whatever you think of their methodology, they actually exist.
We should recognize the existence of outliers, and we should recognize that outliers might be ahead of the curve of everyone else on some factor. But we should not think that the outliers represent the most likely scenario.
If nothing else, watch where the parties and their affiliated groups choose to spend their money.
If you’re a candidate who has consistently trailed all autumn, I am skeptical that you are about to mount an amazing come-from-behind victory. Almost 55 million votes have already been cast in this election cycle; we will probably see turnout a little bit higher than the 112 million ballots cast in 2022. So probably 40 to 50 percent of the ballots in this year’s election have already been cast.
THE SENATE
Seats where the party in power will not change . . .
Maine: I predict Susan Collins will win against Democrat Troy Jackson, in part because the 2020 results suggest there’s a portion of the Maine electorate that likes Collins just the way she is and doesn’t feel like telling any nosy pollster who calls on the phone or texts them.
New Hampshire: I predict Democrat Chris Pappas will beat Republican John Sununu.
South Carolina: Darline Graham may be a nice person, but she entered a race that is usually one of the easiest wins for any Republican in any state and turned it into a nail-biter. I think she hangs on over Democrat Annie Andrews, but it’s strikingly close for this state, and a painful demonstration of the risks of letting the president pick and endorse a candidate, based upon personal relationships, instead of letting the state’s primary voters go with whatever better-known and trusted representative or state legislator they like best.
Georgia: Incumbent Democratic Senator Jon Ossoff is going to cruise to reelection. While Ossoff says he isn’t interested in running for president, Republicans will regret not being able to scuff him up a bit with a hard-fought race. If Ossoff doesn’t run, he will be a top-tier running mate option for the 2028 Democratic presidential nominee. Departing GOP Governor Brian Kemp — you know, the guy Trump called a “fool” and a “clown” — would have, at minimum, given Ossoff a real race.
Florida: Republican Ashley Moody, appointed as Marco Rubio’s temporary replacement, will dispatch Angie Nixon, a Democratic Socialist, fairly easily.
Michigan: This will be an extremely unpopular prediction, but I think that Abdul El-Sayed will defeat Mike Rogers. I’ve been bearish on this race for a while; El-Sayed has led the polls pretty consistently, and sometimes by a surprisingly wide margin.
Alas, I think the traits, statements, and stances that make El-Sayed utterly unelectable and unacceptable to conservatives are unnoticed or shrugged off by the average Michigander. A WDIV/Detroit News poll of likely voters conducted from October 19 to 22 found that 42.2 percent of respondents felt positively about Abdul El-Sayed, and 44.8 percent felt negatively. That might sound not-so-great, but the numbers for Rogers were 32.2 percent favorable and 49.2 percent unfavorable. As of mid-October, Republicans had spent more than $92 million in advertising against El-Sayed; if that advertising has had an effect, it is well hidden.
Democratic Representative Haley Stevens’s decision to join Rogers at church in Detroit Sunday morning and to endorse Rogers spurred some last-second hesitation on this prediction, but a) it’s not clear that a lot of voters will change their mind because of Stevens’s endorsement and b) more than 2.1 million votes have already been cast in Michigan. Four years ago, 4.5 million Michiganders cast their ballot. (If you think that polls oversample Democrats — and many likely do — and Haley Stevens represents a significant group of Democrats who find El-Sayed too antisemitic and extreme to support, shouldn’t he be polling much worse than he is?)
As of 2021, Michigan is home to about 105,000 Jewish adults, comprising 1.4 percent of the state’s electorate; as of early 2025, Michigan’s Muslim population is estimated at about 240,000 or 2.4 percent of the state. El-Sayed believes that his wink-and-a-nod approach to Hasan Piker and Cenk Uygur will add more votes to his total than he will lose. Unless significant numbers of non-Jewish voters find El-Sayed’s approach odious, the math appears to be on his side.
El-Sayed will run well behind the Democratic gubernatorial nominee, Jocelyn Benson, but in the end, it won’t be enough in a blue state in a Democratic year. Michigan automakers have spent the past two years screaming their heads off about the tariffs and trade war with Canada. That’s a huge headwind for Mike Rogers to overcome.
Minnesota: As much as I think Michele Tafoya has run just about a near-perfect campaign, the combination of this being a Democratic-leaning state and a Democratic-leaning year means Peggy Flanagan is the one more likely to end up in the Senate. The last time a Republican won an election statewide in Minnesota was Tim Pawlenty in 2006. If Tafoya is on the verge of pulling off the mother of all Senate race upsets, don’t you think she would have polled above 46 percent at least once?
Kansas: Yes, incumbent Roger Marshall is showing strikingly weak numbers in a state that is usually nearly automatic for Republicans, but as I wrote at the beginning of October, this is a state that has a lot of Republicans and not nearly as many independents and Democrats. Voters in Kansas can register by party, and as of Monday morning, 44 percent of mail-in ballots returned and early in-person votes were cast by registered Republicans, and 41 percent by registered Democrats. Marshall hangs on, but it’s much closer than usual in a traditional red state.
Nebraska: I didn’t buy into the “independent” (wink, wink) Dan Osborn hype two years ago, and I don’t see any reason to start buying into it now. In a year where Democrats are leading in a whole bunch of red states, Osborn hasn’t led incumbent Republican Pete Ricketts in any survey that wasn’t sponsored by his own campaign or a Democratic group.
Alaska: I had been bearish on this one, too, but I do think the New York Times article about how Democrats recruited and set up the other Dan Sullivan is going to damage the Democratic nominee, Mary Peltola.
Please note that way back in June, our Kamden Mulder reported:
The impostor candidate was historically registered to vote under the name “Daniel J. Sullivan, Jr.” yet his campaign filings did not match those records. Both the “J” and the “Jr.” were notably dropped off the candidate’s name.
The fake Sullivan’s initial approach was even more brazen: He requested to file under the name “Dan S. Sullivan” — a name identical to the incumbent’s — despite the fact his actual middle name begins with a “J.”
As we know, news doesn’t become real to Democrats until it is reported in the New York Times.
Note that in Alaska, of the more than 86,000 early votes cast so far, 28 percent are from registered Republicans, and just 19 percent are from registered Democrats.
U.S. Senate Seats that flip:
North Carolina: Democrat Roy Cooper will likely be declared the winner not too long after the polls close. A previously Republican-held seat flipping to the Democrats is something of an ominous omen for Republicans the 2028 presidential race. Like Ossoff’s reelection in Georgia, the disappointment for Republicans is not the loss, but the failure to make it competitive.
Ohio: Democratic former Senator Sherrod Brown has enjoyed a very steady narrow lead, and he returns to the chamber, defeating appointed Republican Jon Husted. At least Husted can turn to male modeling as a backup career.
Iowa: A lot of Republicans think Ashley Hinson will hang on in Iowa over Democrat Joshua Turek. Polling in this race is very tight, while the polling in Iowa’s governor’s race shows Democrat Rob Sand well ahead of Republican Zach Lahn, and two of the state’s four all-Republican House seats are rated “lean Democratic” by Cook Political Report, and a third is rated “toss-up.” We’re in an era where ticket-splitting is getting rarer, so I think the Democrats eke out a win in the U.S. Senate race, too. Hinson is currently pledging to be “a thorn in [Trump’s] side if he’s wrong for our state,” which I think is an indicator of how her campaign sees the mood of the state’s electorate.
Texas: A few weeks ago, I wrote that Democrat James Talarico needed about five things to break his way in this race to be competitive, and he’s gotten about seven of them.
I know polling isn’t the end-all and be-all of assessing who’s on course to win, but . . . if Paxton were on course to win, don’t you think he would be leading more than two of the last 39 polls, as of Thursday? (There are a handful of ties in there, too.)
Our friends in the Democratic Party often wail about the sinister influence of “dark money” and demand that free speech, in the form of political advertising, be regulated by the government. Well, outside groups have spent $221 million trying to convince Texans not to vote for Talarico. So far, it appears to have been about as effective as the $1 billion Mike Bloomberg spent running for president, winning only American Samoa. You can spend as much as you want on advertising a brand of dog food, but if the dog won’t eat it, the owners won’t buy the brand.
That adds up to four seats flipping from Republican to Democrat, and none in the other direction, leading to 51 Democrats and 49 Republicans . . . which would lead to us spending November and December wondering if Pennsylvania Democratic Senator Jon Fetterman will remain in his party or caucus with the Republicans.
THE GUBERNATORIAL RACES
Ohio: I predict Democrat Amy Acton beats Vivek Ramaswamy, with a lot of recriminations about the decision to nominate Ramaswamy and his late commercial featuring Acton’s estranged mother.
Georgia: Republican Rick Jackson narrowly wins over Keisha Lance Bottoms. If this goes the other way, it’s a catastrophic night for Republicans.
Florida: Byron Donalds narrowly wins over Democrat David Jolly, raising questions about whether the state’s Republicans have gotten a little overconfident.
Iowa: As alluded to above, Democrat Rob Sand is on course to defeat Republican Zach Lahn.
Kansas: In one of the few Republican pickups of the evening, I predict Republican Ty Masterson will defeat Cindy Holscher.
Texas: Republican Greg Abbott narrowly wins over Democrat Gina Hinojosa.
Nevada: Incumbent Republican Joe Lombardo narrowly wins.
Arizona: Democrat incumbent Katie Hobbs will win a second term over Republican Andy Biggs. Outside of Trump’s victory here in 2024, it is fascinating to see how badly the Arizona GOP does in statewide races.
Is it even a prediction to say Democrats win the governor’s races in the blue-wall states of Pennsylvania, Wisconsin, and Michigan? I might as well predict Xavier Becerra will win the California governor’s race, too. Becerra recently bragged, “I don’t know if we’ve ever had a governor who comes from where I come from.” While that statement makes him sound like an immigrant, he was born and raised in Sacramento.
THE HOUSE
The U.S. House of Representatives currently has 219 Republicans or Republican-caucusing independents and one vacancy in a GOP-leaning district, and 214 Democrats and one vacancy in a Democratic-leaning district, so effectively a 220–215 split in favor of Republicans.
I think when all the votes are counted — which will take weeks out in California — Democrats pick up twelve seats, for a 227–208 advantage. That may not sound like much, considering the Senate and gubernatorial races, but there just aren’t enough competitive seats to see a 31-seat swing like in 2006 or a 63-seat swing like in 2010.
If the House Leadership Fund has stopped running ads to defend incumbents Tom Kean Jr. in New Jersey, incumbent Representative Mariannette Miller-Meeks in Iowa, and the open seat in Nebraska . . . they’re not going to keep those seats. Max Miller is not going to win reelection in Ohio.
ADDENDUM: If you’re wondering about my track record, in 2024, I got all my U.S. Senate predictions right; I predicted Donald Trump would win the presidential election but that Harris would win Michigan and Wisconsin; I also correctly predicted that Republicans would keep control of the House.
In 2022, I wrote, “I think we’re on the higher end of a red-wave year, and approaching ‘red tsunami’ territory.” The night was nowhere near a red tsunami, but Republicans did gain nine seats and win a majority in the U.S. House of Representatives.
At the end of the year, Politico put my prediction on their list of “Worst Predictions of 2022,” right alongside a Financial Times editor insisting Russia would not invade Ukraine; Douglas Macgregor insisting “This war is really over for the Ukrainians” on March 17, 2022; and a lawyer’s confident insistence, “Ron DeSantis will lose his race for governor of Florida.”
Unsurprisingly, I feel like those other guys’ predictions were worse.