Donald Trump’s braggery about sexual assault sparked immediate outcry and condemnation from Republican Party leaders, with many pulling their support. Although House Speaker Paul Ryan didn’t reverse his endorsement, he said he’d stop defending Trump and told members of his caucus they were free to support or oppose the GOP’s nominee as they pleased. Many wondered if Trump’s meltdown would lead to a true landslide loss, but so far the polls don’t quite show it—yet. He trails by 6.7 points nationally in the Huffington Post Pollster average, virtually unchanged from his 6.4 percent deficit when the Access Hollywood tapes surfaced on Oct. 7.
Still, with a little under three weeks to go, we can’t yet say whether things will get worse for Trump, since we simply have no modern precedent for a major party abandoning their nominee. However, there is one telling downballot precedent that offers many parallels: Missouri’s extraordinary 2012 Senate election.
In that race, Democratic Sen. Claire McCaskill was unpopular with voters and faced a hostile electorate in a red state. It’s easy to forget now, but she narrowly trailed Republican Rep. Todd Akin at the time he won the GOP primary (with McCaskill’s now-legendary assistance). But soon after earning his party’s nomination, Akin’s campaign totally fell to pieces. On Aug. 19, he infamously defended a ban on abortions that allowed for no exceptions by preposterously claiming pregnancy could not result from rape. In lines that will live forever, Akin declared, “If it’s legitimate rape, the female body has ways to try to shut the whole thing down.”
Amazingly, Akin’s shocking ignorance now manages to pale in comparison to all the countless offensive things Trump has said. Akin’s statement was more than outrageous enough, though, and it led his own party’s leaders to call on him to drop out of the race—not out of any moral conviction, mind you, but because they quickly concluded he’d destroyed his own candidacy. But the same kind of bull-headed stubbornness that McCaskill knew would benefit her if Akin became the GOP nominee also ensured that he wouldn’t step aside at the establishment’s behest, and his outside financial support dried up.
As a result, many pundits immediately presumed Akin had rendered himself dead man walking, but—and this is the crucial point—the polls simply didn’t show it. Two weeks after “legitimate rape,” Akin only trailed McCaskill by 1 percent, according to the Pollster average. Even a full month later, he was behind just 3.3 percent, and on Election Day—two and a half months after his meltdown—polls “only” showed her 8.7 points ahead. You can see it all unfold in the chart below from Huffington Post.
Yet McCaskill ultimately prevailed in a 15.7 point landslide, which few polls came close to predicting. Akin’s example could matter greatly for Trump, since polls taken even two weeks after his implosion similarly have not shown a sharp swing toward Clinton yet. But as Akin shows, they soon could.
Presidential elections are more polarized than downballot ones, and many of the so-called “game changers” of past races don’t actually swing many votes. However, unlike in practically every modern election, Trump’s party leaders and elites will no longer be there to defend him from every attack and prime Republican-leaning voters to circle the wagons around him. This unprecedented situation makes it plausible that Trump’s numbers could decline further under sustained scrutiny. Even more critically, droves of Republicans might simply not vote because they are demoralized by the prospect of an unwinnable race.
And lower Republican turnout could cause Democrats to overperform their current polling numbers both up and down the ballot. While many Republican-leaning voters did ultimately abandon Akin, they still showed up to vote for Mitt Romney, who easily won Missouri by 9 points. However, it’s almost always easier to motivate voters to vote in races at the top of the ticket than downballot, where the stakes are less obvious and partisan allegiances can be weaker.
In one recent example, Democrats reaped the devastating consequences of putting forth a non-credible candidate in Nevada’s 2014 gubernatorial race, which was the top contest in the state that year. That led to abysmal turnout among Democrats, which in turn led to shockingly heavy downballot losses for Team Blue. Now just imagine something like that on a national scale.
It’s not just Trump’s temperament that’s wreaking serious havoc for Republicans—his lack of a traditional campaign could as well, particularly since he relies more on voters with lower education levels than past Republican presidential nominees have. That demographic, though, is less likely to vote than more-educated voters. At the same time, Trump has leaned heavily on free media coverage to power his candidacy, while Clinton and the Democrats have pumped hundreds of millions of dollars into TV ads and a get-out-the-vote operation. It’s unclear what precise impact this disparity could have, but one political scientist estimates it could cost Trumps 6 net points in swing states.
One other way Trump could depress Republican turnout is with his now-central claim that the election is “rigged.” There isn’t a shred of evidence to back him up, of course, and this line of argument dangerously undermines our democracy itself because Trump holds sway with millions of partisans as the face of his party. However, one political science experiment found that such claims could backfire by lowering turnout among voters on the side making such accusations because their partisans figure there’s no point in voting if the outcome is fixed. Trump, in other words, is telling his own supporters that they shouldn’t bother casting ballots.
With all of these ways Trump could depress turnout, pollsters could have an even more difficult time estimating who is a likely voter, leading them to overestimate Republicans. Another problem pollsters face is the higher level of support for third-party candidates and the unusually large number of undecided voters, compared to prior elections. Importantly, Clinton usually performs better in polls that exclude Libertarian Gary Johnson and Green Party candidate Jill Stein. What these polls reveal is that third-party supporters and undecideds utterly despise Trump, but aren’t yet sold on Clinton either. If they come home to a major party—as often happens later in the campaign—that could boost Clinton further.
Taken together, these issues with polling and turnout mean that just like with Todd Akin, Trump could indeed lose by a larger margin than the polls say now, and even than what they say on Election Day. And this finding is key because, as the chart below shows, generic congressional ballot polling has mostly tracked the two-way presidential election margin in polls over the last several weeks:
A rising Clinton tide could very well lift Democratic boats downballot, and even if there are still many split-ticket Clinton voters who favor congressional Republicans, a wave could nonetheless wash away vulnerable Republican incumbents.
At present, Democrats lead by 5.5 percent in the generic congressional ballot polling average. Thanks largely to gerrymandering, Democrats might need something closer to an 8 percent lead in the nationwide popular vote to win a majority, and many entrenched Republican incumbents likely poll better when named than a more abstract “generic Republican” would. That means Democrats still have a ways to go in the generic ballot polls before they’re positioned to win the House.
And it’s still possible Trump will fare better than Todd Akin, since he is after all running in a more polarized political environment where most Republican elites haven’t yet denounced him despite many high-profile defections. However, it’s plausible that Trump continues to trend downward—particularly as allegations of actual sexual assault continue to mount—and underperforms his polling thanks in part to lower turnout. If Trump does decline, downballot Democrats’ 5.5 percent congressional polling lead could plausibly grow wide enough that they could take the House. We’ll know soon enough.
