This is an unusually vexing election. Political history and President Trump’s unpopularity argue for one set of outcomes, but geography and the gravitational pull of partisanship make the opposite case.
We know that, for better or worse (usually worse), midterm elections are referenda on the sitting president. The president’s party has suffered net losses in House seats in 18 of the 20 midterms since the end of World War II. President Trump’s approval ratings currently sit at 37 percent (disapproval 59 percent) in the New York Times polling average and 38 percent (58 percent disapproval) in Nate Silver’s Silver Bulletin, basically the lowest midterm job approval for any president since Jimmy Carter. To those who lamely argue that the election is still 11 weeks away, I would note that we have never seen a major directional change during a midterm election year in modern times.
Simply put, those who belong to or lean toward a president’s party are rarely highly motivated, and that lack of enthusiasm translates into a drop in turnout. Conversely, those in or leaning toward the opposition party tend to still be upset about the loss two years earlier, are highly motivated, and turn out in larger numbers. True, turnout drops across the board from a presidential election to the midterm two years later. But the drop-off in midterm turnout within the president’s party is significantly greater than the drop for the opposition party.
In key states where we have seen reputable polling, Trump’s approval is running between 6 and 14 points below his vote share, a big drop in two years. The gap between Trump’s share of the 2024 vote and his current approval ratings is the lowest in Alaska and Maine, and the highest in Iowa, North Carolina, Ohio, and Texas.
Where Democratic Senate candidates are doing best, the recipe is pretty much the same. First, they are winning a higher percentage of votes from Democrats than the GOP candidates are from Republicans. Second, Democratic candidates are getting a higher share of independent votes than Republican candidates. Third, Democrats are getting a larger share of the women’s vote than Republicans are of the men’s vote.
That Democrats are overperforming in a lot of red states is unmistakable, but they are also encountering a bumpier road in some blue, Democrat-tilting states like Maine, where Republican Sen. Susan Collins is seeking a sixth term. Some disagree, but I think it will be an uphill battle for Democrats to hold onto their open seat in Michigan, after Motor State Democrats made an unforced error in selecting Abdul El-Sayed as their nominee. In the Minnesota open seat, they nominated potentially problematic Lt. Gov. Peggy Flanagan.
To capture a Senate majority, and to match the midterm average of a 26-seat House gain, Democrats have to do something that neither party has succeeded in doing in a long time: They need defections, specifically convincing Republicans to vote for a Democrat. For Democrats, doing well among fellow party members and independents is not enough in many red states. This becomes critical particularly if they have to win five GOP seats to make up for losing one of their own.
According to the Pew Research Center’s invaluable Validated Voter surveys after the last three presidential and last two midterm elections, the share of Democrats voting for a Republican or Republicans voting for a Democrat has been astonishingly small. In fact, in both 2020 and 2024, 95 percent of Democrats voted Democrat for president; the same percentage of Republicans voted for Trump. In each case, defections were either 4 or 5 percent. In 2016, 94 percent of Democrats voted for Hillary Clinton (5 percent for Trump); 92 percent of Republicans voted for Trump (4 percent for Clinton). In midterm elections, the parties vote even more monolithically. In the 2018 midterms, Democrats voted 97 to 4 percent for their party’s House candidate, Republicans 98 to 3 percent. In 2022, Democrats voted for their party by 98 to 2 percent; Republicans voted 97 to 3 percent for theirs.
In top Senate races over the last two elections, the voting has been almost as lopsided. Using the Fox News exit polls, which have larger state-level samples, in the 2024 Arizona Senate race Democrat Ruben Gallego carried Democrats 95 to 3 percent; Republicans voted for Kari Lake 85 to 13 percent. In Michigan, Elissa Slotkin won Democrats 95 to 3 percent; Mike Rogers won among Republicans 90 to 7 percent. Jon Tester won his fellow Democrats in his unsuccessful Montana reelection bid, 98 to 1 percent; Republican Tim Sheehy carried his fellow Republicans 85 to 13 percent. In Nebraska, independent Dan Osborn, the de facto Democrat, won the vote of Democrats 97 to 3 percent; Sen. Deb Fischer carried her Republicans 82 to 18 percent. In Ohio, Democratic incumbent Sherrod Brown carried his fellow Democrats 95 to 3 percent; Republican Bernie Moreno won his party members 86 to 11 percent.
Things were much the same in 2022. In Arizona, Democrats voted to reelect Sen. Mark Kelly by 96 to 2 percent; Republicans backed Blake Masters by 85 to 13 percent. Democratic Sen. Raphael Warnock won among his fellow party members, 99 to 0 percent; Republican Herschel Walker won 91 percent of Republicans to just 6 percent who voted for Warnock. In North Carolina, Democrats voted for Cheri Beasley 96 to 2 percent; Republicans voted to reelect Sen. Ted Budd, 93 to 5 percent. Ohio Democrats voted 98 to 1 percent for Tim Ryan; Republicans voted 92 to 7 percent for J.D. Vance. John Fetterman carried Pennsylvania Democrats 96 to 3 percent; Mehmet Oz won 87 percent of GOP voters to 9 percent who voted Fetterman. In Wisconsin, Mandela Barnes pulled 99 percent of the Democratic vote; Sen. Ron Johnson, 1 percent, but Johnson carried Republicans 97 to 3 percent.
Trump is a huge liability for Republicans; there is no question that the GOP is going to have huge turnout problems, meaning that in a lot of races, Democrats are going to greatly exceed expectations. But converting Republican voters to actually back a Democrat—even for a day—is so much harder now. Democrats have to challenge the electoral physics that have been increasingly in place for about 20 years, especially since 2016. Poll numbers show they will have some success, but winning a majority will still be gravity-defying.