“Are we rewriting America’s Electoral College map tonight?”

In the predawn hours of Nov. 9, 2016, CNN’s chief national correspondent, John King, posed the question that would come to define the mechanics of Donald Trump’s victory. It had been 28 years since a Republican had won Wisconsin, Michigan or Pennsylvania—states belonging to a much-hyped “blue wall” that narrowed the GOP’s path to 270 electoral votes. As it became clear, however, that Trump would carry all three states, and secure blowouts in the traditional battlegrounds of Ohio and Iowa, a new reality took hold. Whereas Republicans once believed they could win the presidency only by using the George W. Bush playbook—appealing to affluent, suburban populations in diverse states such as Virginia, Colorado and Nevada—Trump proved the truer route to the White House went through working-class whites in the industrial Midwest.

Story Continued Below

The narrative of a sweeping realignment was irresistible. Analysts branded the Rust Belt as “Trump Country.” Streaming into greasy spoons, shot-and-a-beer bars and Walmarts across the Midwest, reporters pumped out anecdotal insights into the forgotten masses in middle America. And Republican officials, almost instantaneously, shifted their outlook. It didn’t matter that Trump had won Wisconsin, Michigan and Pennsylvania by fewer than 78,000 votes combined—not enough people to fill a popular Big Ten football stadium—or that his path through those states was necessitated by his defeat in once-purple areas of the country. Trump had redrawn the Electoral College map, they insisted, and the implications were far-reaching: Rust Belt Democrats could soon become an endangered species.

Now, on the eve of the 2018 midterm elections, Republicans are bracing for a massacre in the Midwest.

“When we woke up after the 2016 election, there was a real possibility that we were seeing a realignment among white working-class voters in the Midwest—and that they could go the way that white working-class voters have gone in the South over the past generation,” said Matt Grossman, a political scientist at Michigan State University. “But two years later, there’s no sign that those gains are holding or being extended. Instead, there are a lot of campaigns in that region where Republicans are struggling to be competitive.”

Trump bears much of the responsibility. His approval rating has plunged, by double digits, in most Midwestern states. His presidency has energized the Democratic base in ways Hillary Clinton’s candidacy never could. His party’s rewrite of the tax code was disproportionately beneficial to wealthy people and corporations; to the extent the law is popular with voters, he barely tries to promote it. And his trade warring has been burdensome to farmers and blue-collar workers in manufacturing-heavy swatches of middle America, with numerous Republicans from affected states, including Speaker Paul Ryan of Wisconsin and National Republican Congressional Committee chairman Steve Stivers of Ohio, privately pleading with Trump and his aides to find a resolution.

And yet, blaming Trump ignores some fundamental realities on the ground. For one thing, Republicans failed to recruit—or cultivate—top-tier candidates in many high-profile races. Democratic challengers outraised GOP incumbents across the region and the country. Complacency has beset portions of the conservative base, with a combination of tax cuts, judicial nominees and regulatory reforms leaving some Republicans fat and happy. Above all, not everyone voting against the Republican Party on Tuesday is voting against Trump; the president persuaded lots of Democratic voters in 2016 without durably changing their party affiliations.

Still, Trump says his name is on the ballot, and the electorate appears to be behaving accordingly. In visits this fall to Michigan, Wisconsin, Pennsylvania, Ohio, Iowa and Minnesota, and in conversations with Republican voters and officials, I found signs of alarm relating to 2018—and a sense of foreboding related to the repercussions beyond.

In Michigan, which Trump carried by 10,704 votes, Democrats are expected to take over numerous state legislative seats, win back the governorship and hold onto Debbie Stabenow’s Senate seat despite a spirited challenge from blue-chip Republican nominee John James. Meanwhile, Democrats are threatening to win a pair of congressional districts, the 8th and 11th, that are anchored in traditionally Republican suburbs. Democrats are also poised to win the race for attorney general—and perhaps most significant of all, the secretary of state’s office, giving the party considerable power to liberalize voting laws before the 2020 election.

“Democrats could make it structurally harder for Trump,” said Rob Steele, the Republican National Committeeman from Michigan. “If the secretary of state is a Democrat, then you’ll see a significant change in how people are looking at voting laws and ballot integrity. My message is, if you want to help Trump, you have to elect these other Republicans.” Steele knows that demographics are working against the GOP; even in predominantly white Michigan, the state is growing increasingly urban and college-educated, which explains the competitiveness of the 8th and 11th districts. “The more money in those areas,” he said, “the more Democratic they get.”

In Wisconsin, which Trump carried by 22,748 votes, Democrats have already begun flipping state legislative seats—prompting warnings of a “Blue Wave” from Governor Scott Walker, which irritated Trump and the White House political operation. Walker himself is in trouble: After two terms in office (and three elections, including the failed recall election in 2012), many Republicans believe the governor will lose to Democrat Tony Evers. In the other prominent statewide race, incumbent Democratic Sen. Tammy Baldwin—who, like Stabenow in Michigan, is steady but unspectacular—is cruising to a comfortable victory over Republican challenger Leah Vukmir. Wisconsin Republicans also worry they could lose the attorney general’s office, though one thing the party isn’t sweating is congressional seats: The only competitive race is in the 1st District to replace Ryan, and Democrats likely made a fatal mistake in nominating Randy Bryce, a union member with personality but tons of political baggage, instead of Cathy Myers, who, Republicans feared, would have flipped the seat.

“This political climate is one that I’ve never seen in my lifetime,” says Chris Goebel, chairman of the GOP in Walworth County, one of Wisconsin’s conservative strongholds. “In the past, you voted with your pocketbooks. But people get thrown out of office when voters are this angry.” At a party unity event in late September, the longtime stars of the state GOP—Walker, Ryan and Reince Priebus—celebrated past victories in tones befitting a memorial. Walker told me this political environment was harsher than in cycles past and argued that Trump owed his success in the Midwest to strong state parties. Without them, he implied, it’s tough to imagine the president running the table in the Rust Belt again. “I would argue the reason [Trump] was able to do that was because Republicans had made the case in those states and showed we could actually get things done,” Walker said.

In Pennsylvania, which Trump carried by 44,292 votes, Republicans are staring into the abyss. Due in large measure to a court-executed redistricting of the state’s gerrymandered congressional districts, Democrats could flip a half-dozen seats in the state; at worst, they should win four. Meanwhile, the state’s top Democratic incumbents, Gov. Tom Wolf and Sen. Bob Casey, are both coasting to reelection. And, as in Michigan and Pennsylvania, Democrats are expected to win back a healthy number of Republican-held state legislative seats.

One reason is that wayward Democrats are simply coming home. “Pennsylvania does have many sons and daughters of what we could call Reagan Democrats, particularly in the central and western part of the state. They went for Donald Trump in big numbers—they put him over the top,” said John Brabender, the pre-eminent GOP strategist in the state. “But there was this wild assumption that because they voted for Trump in 2016, that they would vote for Republicans in 2018. But these people are still Democrats. They’re still going to vote for Democrats. And what we forget is that it wasn’t a vote on Donald Trump up-or-down in 2016; it was Donald Trump versus Hillary Clinton. If Hillary Clinton was on the ballot against some of these Republicans, we’d be doing a lot better.”

Even in those Midwestern states won handily by Trump, there are red flags galore.

In Ohio, which Trump won by 447,000 votes—a relatively whopping 8 percentage points—Republicans are expected to lose seats in the Legislature for the first time in a decade. They are in danger of losing the deep-red 12th Congressional District , which they just spent millions to defend in a summertime special election. Republicans are also likely to lose the majority of statewide elections. In the U.S. Senate race, incumbent Democrat Sherrod Brown will likely win by double digits over the challenger, Republican congressman Jim Renacci. And in the governor’s race, despite incumbent Republican John Kasich leaving office with approval ratings in the 50s, the race is a toss-up between Democrat Richard Cordray and Republican Mike DeWine.

In Iowa, which Trump won by 147,000 votes—or 9.5 percentage points—Democrats are poised for gains across the state. Republicans control three of Iowa’s four congressional seats, and all of them are being contested. Rod Blum, the incumbent in the 1st District, is likely a goner, and David Young in the 3rd District is hanging on for dear life. Even Steve King, the immigration hard-liner who represents the ultraconservative 4th District, faces a legitimate challenge from Democrat J.D. Scholten. Democrats are likely to pick up numerous legislative seats and could even retake control of the state House. And statewide, among other races Republicans are worried about, Governor Kim Reynolds is in deep trouble against Democratic challenger Fred Hubbell.

And in Minnesota, which Trump lost by 45,000 votes, the only good news for Republicans is a likely pickup in the vacated 8th Congressional District. Otherwise, the state looks bleak: Democrats appear certain to flip two Republican-held congressional seats, in the 2nd and 3rd districts, while holding onto the governorship and both U.S. Senate seats. Even the attorney general’s race—in which Democrat Keith Ellison, the former congressman, has been dogged by allegations of sexual assault—is thought to be slipping away from Republicans.

Strategists in both parties have stressed the volatility of this election cycle, and Trump’s victory in 2016 demonstrated that polling can be unreliable. Still, in the final weeks before the midterm elections, Republicans in the Rust Belt were in agreement that Nov. 6 would be a rough night at best—and a disaster at worst.

The three states of greatest concern for the GOP continue to be Michigan, Wisconsin and Pennsylvania. This isn’t simply because Democrats are poised to pitch a shutout against Republicans in the highest-profile races there, but because without their combined 46 electoral votes, Trump would have lost the presidency. His 306 electoral votes (304 in the final tally, because of the defections of two electors) looks comfortable on paper—until you consider the threadbare margins in those three Midwestern states.

It’s true that Trump was close to winning elsewhere: He lost New Hampshire by less than 1 point, Minnesota by less than 2 points and Nevada by less than 3 points. But those states offer relatively few electoral votes; remember that Trump carried Florida and its 29 electoral votes by barely more than 1 point, and won the rapidly diversifying states of Arizona and North Carolina by less than 4 points each.

All of which ensures that Trump’s path to reelection will once again run through the Midwest; that if he cedes any other parts of his winning 2016 map, most essentially Florida, the president will likely have to run the table in Iowa, Michigan, Ohio, Pennsylvania and Wisconsin.

“I think that he pulled an inside straight in 2016, and he hasn’t done that much to change his cards, so he may have to pull another inside straight,” David Axelrod, chief strategist to the Obama campaigns, said in a recent interview for POLITICO’s Off Message podcast. “If the results are what they look to be, I think it’s an important harbinger for him that Republicans just didn’t do well in those states that delivered the presidency to him.”

Tuesday’s midterm elections mark the beginning of the 2020 campaign. And for all the focus on suburbanites fleeing the GOP, a blue wave in the Midwest would be indicative of Trump’s base turning against his Republican Party.

“There’s a lot of talk about these swing House districts that are disproportionately suburban, but the polls are showing similar movement among the white working class,” said Grossman, the MSU political scientist. “When you have states like Ohio and Iowa that swung very strongly toward Trump, and this year they are seeing Democratic voters returning to the fold, those Democratic gains would be disproportionately among the white working class.”

One constant of Donald Trump’s ascent to the presidency and his two years in power is how behavior that would be not just risky but downright stupid for any normal politician ends up working smartly for him.

This is the essence of the Trump Mystique—a three-year record in which he regularly demonstrated that many of the normal precedents, patterns and truisms of American politics simply do not apply to him. This mystique—Is it real or illusion? Is his patented sorcery still working?—is among the big questions being tested in Tuesday’s elections.

Story Continued Below

Trump’s own decisions over the past month have put the issue—whether Trump has defied political gravity or merely delayed its impact—in even sharper relief than it would have been anyway.

It would be smart, viewed through a conventional prism, for a president who has never commanded majority support to try to float above the midterms and allow politicians of his own party to keep their elections locally focused. It seems stupid to unite and energize the opposition in their loathing by insisting that congressional elections are a national referendum on himself.

It would be smart, if playing by normal rules, for a leader presiding over the best employment numbers in decades to make an economic argument his main push against the headwind that the incumbent president’s party historically faces in midterm elections. It seems stupid to reduce this to secondary status in favor of picking scabs over immigration and societal violence in the days before voting.

In the disoriented state of contemporary politics, however, it seems stupid for anyone to pretend to be smart in predicting the results of Trump’s decision to turn the volume up to 11 on Trumpism.

As Trump himself cast the implications for Tuesday in a weekend stop in Georgia: “I wouldn’t say it’s as important as ’16, but it’s right up there.”

Some dynamics seem inescapably true. One is that at nearly every important turn when traditional political logic would have pointed toward softening the tone and broadening support—from his 2016 acceptance speech to the 2017 inaugural address and countless other occasions since—Trump took the opposite path and along the way tightened his connection to his most devoted supporters.

A vivid recent example was over the sexual assault allegations against his Supreme Court pick, Brett Kavanaugh. For a few days Trump deferred to prevailing wisdom that he needed to treat accuser Christine Blasey Ford respectfully and project an open mind on the merits. Before long he returned to his customary instincts and attacked Ford, Democrats and the media, while cheering on Kavanaugh’s own attacks on Democrats.

For every Republican operative who thinks Trump’s midterm strategy is nuts—one senior GOP strategist running competitive statewide races said the president’s image took a 15-point hit in internal campaign polling over the past 10 days—there is a Democratic operative who worries that Trump’s polarizing approach just might allow him to beat the odds as he did in 2016.

But that same approach raises the cost of GOP setbacks for Trump, who has often made clear his own view that power is partly a matter of perception, and preserving an aura of strength and success. A narrow House loss, for instance, would surely be explained as the result of normal historical patterns. In the case of a national blowout, no matter if Trump blamed others, the result would be like a baby with a paunch and comb-over: No way to deny paternity.

“I do think it’s a little unfair to put it all on him, because you start behind the eight-ball,” said a senior GOP Senate strategist, pointing to the usual historical pattern with a president’s first midterm election. “What I think is different [in 2018] is that while the president always has the ability to define the agenda, he takes all of the oxygen out of the air. The reality is, these races are completely national. And while there’s always a national bent to congressional races, there’s really no escaping it this time.”

A senior White House official said political advisers applied a three-prong test this fall in deciding where to send Trump. One was whether they could find good rally venues. Two was data suggesting which districts were especially promising if Trump could manage to ignite GOP-leaning voters who might normally vote in presidential elections but not midterms. Third was protectiveness, trying to avoid races where Trump would risk being blamed for a race that was a likely loser anyway.

The president’s vituperative attacks on Democrats and race-baiting immigration rhetoric broke new ground on divisiveness, but in one sense he was making a calculation—can a president influence the midterms to advantage?—familiar to three of his recent predecessors.

In 1994, Bill Clinton’s advisers urged him to take it easy and mostly stay off the campaign trail in favor of the White House and overseas trips. He didn’t buy it—convinced he could persuade voters to back him and Democrats if he could just get in front of enough of them. Polling suggested otherwise, and political aides later concluded that an unseasoned president’s own efforts helped fuel the GOP’s historic congressional takeover that year.

In 2002, the backdrop of 9/11 one year earlier changed the landscape for George W. Bush. Stressing national security themes, he helped Republicans make historically unusual congressional gains.

In 2010, Barack Obama saw a conservative backlash over spending to combat recession and the financial crash, as well as the Affordable Care Act. He campaigned in some districts where he was welcome, but he knew it wasn’t doing much good. “There’s no doubt this is a difficult election,” he said at a Cleveland rally. He was right: November brought a “shellacking,” as he called it, that lost the House and reached deep into statehouses around the country.

Similar results in the opposite direction against Republicans on Tuesday will not only put subpoena power in the hands of the president’s political foes—it could lead the handful of prominent Trump dissenters in the national GOP to urge others to join their cause.

“Yeah, he’s going to lose the House,” said Bill Kristol, editor at large of the Weekly Standard and a leading Trump critic. “They’re gonna lose eight to 10 governorships probably. So, where is the brilliance? Where is the political magic? … He got 46 percent of the vote in 2016. It looks like Republicans are going to get, if they’re lucky, 46 percent of the vote [or lower]. … So what has Trump done for the party?”

Not that Trump will admit as much. Terry Sullivan, who managed Marco Rubio’s 2016 campaign, suggested that one key aspect of Trump’s mystique is that he will argue that his mystique is undimmed no matter the result. “Don’t take my word for it. Ask him tomorrow,” Sullivan said Monday. “Don’t take my word for it, ask his supporters. He will say that candidates that he campaigned with won and the ones who didn’t want to campaign with him lost. And the ones that lost that he campaigned with did better than they would have if they hadn’t campaigned with him—he made the race closer, so much closer.”

For all Sullivan’s evident sarcasm, Michael Strain, director of economic policy at the American Enterprise Institute, effectively agreed that Trump’s activities in the closing days of the campaign might help in some districts but won’t be the decisive factor if the evening ends in a big GOP defeat. “I think that the cake on the president is kind of baked—that people have a view of the Republican Party under Donald Trump” that won’t swing widely based on any day’s headlines, he said. “That suggests to me that if the president were talking about the economy and not talking about the caravan, that wouldn’t necessarily be a better strategy to get Republicans to win.”