Midterm victories in Pennsylvania, Michigan and Wisconsin gave Democrats hope of retaking the Rust Belt battleground states that handed the presidency to Donald J. Trump in 2016.
Yet success in the midterms might not mean as much for Democratic presidential candidates as the party might think. Nearly two-thirds of voters in six battleground states who voted for President Trump in 2016 — but for Democratic congressional candidates in 2018 — say they intend to back the president against each of his top rivals, according to recent polling by The New York Times Upshot/Siena College.
The results suggest that the party’s winning formula in last year’s midterms may not be so easy to replicate in a presidential election. The Democrats’ relatively moderate House candidates succeeded in large part by flipping a crucial segment of voters who backed the president in 2016. If these voters remain open-minded again in 2020, Democrats will have a ready-made blueprint for winning back the crucial Rust Belt battlegrounds.
This group is only a sliver of the electorate — 2 percent of registered voters — and is not representative of all voters. They are overwhelmingly white, 60 percent are male, and two-thirds have no college degree. But the president’s strength among them helps explain why he is highly competitive in states that Democrats carried just one year ago.
Many of the voters who said they voted Democratic but now intended to vote for Mr. Trump offered explanations that reflect longstanding theories about why the party out of power tends to excel in midterms.
Michelle Bassaro, 61, is a Trump supporter, but in the midterm election, she voted for the Democrat in her district to balance the administration’s power. She said she had voted for Republicans when Democrats were in the White House for the same reason, consistent with research that shows that some people intentionally vote for divided government.
Another reason was local: The Democrat promised to bring more jobs to her area, Nanty Glo, Pa. (The name comes from a Welsh phrase that means “streams of coal,” but its coal jobs have disappeared.)
Voters often think differently about state and national issues. Some said they had voted for their local Democrat in the midterms because the person had served well for a long time, or because the candidate’s policies would directly help their community. But presidential politics were another story, they said. Many of the white working-class voters in the Rust Belt who supported the president in 2016 were traditionally Democratic voters who backed President Obama in 2012 and even continued to vote Democratic down-ballot in 2016. Democrats generally held on to these voters in 2018, but the reasons many of them voted for Mr. Trump, like his promises on immigration or the economy, could still be relevant.
Michael Townsend, 38, a high school-educated construction worker in Dunmore, Pa., was a lifelong Democrat — until he voted for Mr. Trump.
“In the last couple years, the Democrats had kind of been losing the work, and I thought Trump might get us that work,” he said. “And to be honest, I’ve been in construction 21 years and the last two years were the best years I’ve ever had.”
He voted for the Democrat in the midterms because he liked his ideas on less polarizing local issues, like veterans affairs and opioids, while he said the Republicans were too focused on Washington politics. He has also been intrigued by Bernie Sanders. But he’ll probably back Mr. Trump again, he said.
Mr. Townsend, who lives just outside Scranton, is in a district that swung from a 12-point victory for Barack Obama to a 10-point win for Mr. Trump in 2016. On the same day in 2016, the district voted to re-elect its Democratic congressman, Matt Cartwright, who won again in 2018.
The district’s continued Democratic tilt down-ballot, even after it flipped at the presidential level, bears out the tendency of congressional races to lag geographic shifts in presidential elections, particularly if the district is controlled by the party out of power.
Nowhere was that more true than in the South, which remained Democratic in the House for decades after Republicans started carrying it in presidential elections.
Danny Destival, 56, who runs a greenhouse supply business in Panama City, Fla., said he’s “been a Southern Democrat all my life.” But in 2016, he cast his first Republican vote because he liked that Mr. Trump was a businessman, not a politician — and he disliked Hillary Clinton.
His main priority is voting for “the person who’s going to get more done” — that’s why he stuck with the Democrats in the midterms — but at the national level, he said, the Democrats have disappointed him on that front.
“If you’re going to Washington, you need to do something,” he said. “If the only thing you’re going to do the whole time you’re there is try to get rid of the president, that’s a problem. I mean, Trump is not a great person, but you’ve got to get some work done.”
Other voters say they are preparing to take an even greater leap: vote for Mr. Trump after supporting Democratic congressional candidates in 2018 and Mrs. Clinton in 2016.
In the survey, 7 percent of those who supported Mrs. Clinton in 2016 said they now approved of the president’s performance — despite his personality and his Twitter account, many said.
“In 2016, I hated both” candidates, said Juli Anna California, 57, a nurse from Coral Springs, Fla. “I went with Hillary because Trump had no history as a politician.”
Mr. Trump has convinced her, though — not with his character, but with his policies.
“He’s not exactly the person I’d have as my best friend,” said Ms. California, who currently lives in Los Angeles as a traveling nurse. “But he’s a great president. Most politicians just talk about doing things, but Trump does them.”
Scott Will, 51, an equipment operator in Ligonier, Pa., also voted for Mrs. Clinton in 2016, and will vote for Mr. Trump next year. So will much of his family, union workers who had been “die-hard Democrats.” Mr. Will, who started college but left to get married before graduating, credits Mr. Trump’s trade deals and pledge to bring jobs back to the United States.
“In years past, it seemed like Democrats were supposedly supposed to be for the working man and for unions,” he said. “But I can say this: I will not be voting Democratic this election.”
Many of the voters cited economic strength as a major reason to support Mr. Trump in 2020, even if they didn’t support him last time. Also, certain voters who support Trump said they had soured on Democrats because of partisan fighting, culminating in impeachment hearings.
Matthew Headley, 41, is a general contractor and owns a pizza business in Grand Blanc, Mich. He has mostly voted for Democrats, including for Mrs. Clinton, whose experience impressed him, but plans to vote for Mr. Trump.
Mr. Headley, who did not finish high school, likes what the president has done for the economy: “The wheels are turning in the right motion for a lot of people who it wasn’t for the longest time.”
“The Democratic Party fell apart on the heels of Trump winning,” he said. “The harder they’re going after Trump, the more they’re just alienating people and pushing them away.”
The appeal of moderate Democratic candidates in a year of strong Democratic recruiting may have also been a factor in 2018.
Margaret Foster, 84, a retired real estate agent in Prescott, Ariz., said the Democrats had become “the socialist party.” Still, she has supported Kyrsten Sinema, her Democratic senator, whom she called honest and willing to work with Republicans.
Mr. Trump is “an egotistical, overbearing man,” she said — but said that doesn’t change what he’s achieved.
“You’re all going to be very surprised because all these quiet little Christian women aren’t saying anything right now, but they are going to vote for Trump again,” she said.
Of course, a smaller but significant share of those who recently switched from Mr. Trump to a Democrat said they intended to vote for one of the leading Democratic presidential candidates next year. And these voters could be a key to victory.
But a broader look at past midterm results suggests they offer no predictive power — and can even be false signals — for the presidential races that follow.