Mike Segar / Reuters

Do you smell that in the air? No, it’s not the first breezes of fall. It’s the aroma of Democratic panic.

The latest trigger is a poll that CNN released late Wednesday afternoon. The poll shows Donald Trump beating Hillary Clinton 46-41 in Ohio, with Gary Johnson taking 8 percent, and Trump winning 47-44 in Florida, with Gary Johnson at 6 percent. Both figures are based on screening for likely voters, not just registered voters. A national CNN poll last week showed Trump winning by a point 49-48.

“This shouldn’t be close, but it’s close,” President Obama said at a fundraiser Tuesday night.

CNN’s poll of those two crucial battleground states is the latest in a series of polls that have shown good news for Trump. The gap between the two candidates has shrunk to one of its smallest points so far. Not since Trump’s post-Republican National Convention bump has the GOP nominee seemed to be in such a strong position. Certainly, mid-September has been decisive in presidential elections before. John McCain had built a lead over Barack Obama before the economy collapsed on September 15, 2008; McCain never recovered.

Of course, Trump also had farther to go to close that gap: With the exception of that bump, Trump has trailed in the race all along, and what’s changing here is momentum. Polling averages still show Clinton leading, and the forecasts from both FiveThirtyEight (64 percent change of Clinton win) and The Upshot (78 percent chance of Clinton win) show her with the edge. But those numbers have shrunk, and if the current trend continued, she could easily lose. As a Citibank analyst noted, in a comment that seems like it ought to be unnecessary but is not, “35% probability events happen frequently in real life.”  

It’s that sort of idea that has some liberals flipping out. After all, this was supposed to be an easy contest! How can it be that a candidate who has lied repeatedly and blatantly, insulted minority groups, not bothered to learn anything about policy, and proposed banning Muslims from the country be even close to winning? Just a few months ago, even straight reporters were speculating that the race might be over, and that Trump had practically no way to win. Now, it looks like a nailbiter.

Or maybe all of this is wishful thinking from Trump fans and a classic example of what Obama adviser David Plouffe described as bedwetting, the tendency among Democrats to freak out and panic at the slightest sign of trouble. But privately, many Republicans who are horrified by Trump but have taken solace in the expectation that he would lose by a wide margin are starting to freak out too.

One interesting way to think about this is to look at Huffington Post Pollster’s polling average:


Clinton vs. Trump


What sticks out here is the Trump side of the graph. Clinton’s standing bounces up and down, but Trump follows a regular pattern. First, he has a ceiling right around 42.5 percent, and he’s never exceeded that. Second, at more or less regular intervals, he climbs up to that ceiling, and then he commits a grievous gaffe. In mid-May, he was flying high, but then the Trump University scandal and his attacks on Judge Gonzalo Curiel knocked the wind out of his sails. (Remember that? When Trump endorser Paul Ryan called his comments textbook racism? Strange times.) He then regained his standing at the RNC, just in time to pick a fight with Khizr and Ghazala Khan and tumble again.

One way to think through the current Trump boom is to watch and see whether he’s able to break through the magical 42.5 percent ceiling. Based on past experience, the key to that is whether he’s able to avoid committing another catastrophic error.

After months of failed “pivots” and “resets,” the newest Trump team does seem to have done a better job of keeping their candidate on message and, most importantly, on script. A campaign synonymous with bumbling has started to look nearly professional.

In part, that’s a triumph of bar-lowering, and it doesn’t mean Trump hasn’t said some bizarre things; there’s the burgeoning Trump Foundation scandal, too. It may be that there’s some fatigue setting in with Trump’s many outlandish comments, or it may be, as many progressives insist, that this is all the media’s fault for covering Clinton too harshly. But whatever the reason, he’s mostly managed to let the focus stay on Hillary Clinton, from her failure to hold press conferences to her continued email troubles to her recent illness. It turns out that staying out of the public eye helps Trump, just like it helped her during the weeks when she sat back and let him self-destruct.

In the meantime, will Clinton pick up on the panic among Democrats and panic herself? After a volatile 2008 campaign, she has run a low-key, stable operation this time around, other than rumors of a shakeup, never executed, after her poor showing in the New Hampshire Democratic primary.

Despite the run of strong polls, Trump faces some serious structural deficits. He’s struggling even among white voters, trailing behind Mitt Romney’s pace. He also has to contend with an electoral map weighted against a Republican, which means that Clinton can afford to lose a few swing states, while Trump has to be nearly perfect. It’s also common for Republicans to gain some ground when pollsters switch from measuring registered voters to polling likely voters, but it’s tough to know just what a likely voter is, especially because Trump has made little effort to build a large-scale get-out-the-vote operation.

Just as it was far too early to declare Clinton a prohibitive favorite two weeks ago, it’s far too early to declare Trump is on a path to victory now. But recent polling provides a useful corrective. Trump is still the underdog, but yes, he can win.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.