Is bipolar disorder a requirement for covering an American Presidential campaign? Possibly. It seems that every gyration in the polls results in a mania of stories hyperventilating about the end of this candidate or that candidate. Over the past week, Donald Trump’s has been all but counted out of the Presidential race due to his "cratering" poll numbers. The numbers for Trump are down, but whether he is really plummeting is still a question mark.
Trump’s real problem is twofold: 1) he cannot break out of the low 40s; and 2) his polling is remarkably durable. If Trump cannot break through his polling ceiling, he simply cannot beat Hillary Clinton. And every day that goes by he becomes more and more reliant on a major blunder by Clinton and her team.
The post-convention polls actually show an increase in support for Trump, but a larger increase for Clinton. From June 7th, when Clinton clinched the Democratic nod, to July 18th, that start of the GOP convention, the average ballot test was 44.5/39.7 favoring Clinton. Since the end of the Democratic convention (July 29th), the ballot test has averaged 46.9/40.6 for Clinton – and that includes a suspicious McClatchy/Marist poll showing Clinton with a 15-point lead, larger than any poll since March 22nd. The most recent three polls (released August 5-7) are only slightly worse with a 47-40 Clinton advantage.
It’s the same story for both candidates’ favorable/unfavorable numbers. Trump’s numbers have hardly moved from 34.1/59.8 pre-convention to 34.0/60.2 post-convention. Clinton remains under water in her favorable/unfavorable, but has improved from 39.1/55.3 pre-convention to 42.4/53.0 post-convention.
It’s bad news that Clinton is growing her lead (by just 1%) and shrinking her favorability deficit, but the real problem for Trump is that he just cannot break out of the low 40s in his polling. Of the 75 polls conducted by B+ or higher graded pollsters by FiveThirtyEight, Trump has only topped 45% in six of those polls. And three of those were clustered in the mid-July run-up to the Republican convention (and on the heels of the release of the FBI report on the Clinton e-mails). Furthermore, Trump has only led Clinton in nine polls, five of which were in the same mid-July period. Trump’s most common number has been 41%. On the other hand, Clinton has not only consistently led Trump, she has hit 50% or higher 18 times.
It is impossible to say if Clinton’s improving numbers are a result of her actions – choosing Kaine, getting a strong Sanders endorsement, etc. – or as a result of Trump’s inability to take a punch and his determination to recapitulate every feud. But the fact is that Trump is headed, if slowly, in the wrong direction and has shown little ability to bounce back on his own. His best poll bump was in the aftermath of the FBI Clinton e-mail report.
If there is one saving grace for Trump, it is that he performs much better with likely voters. Most of the polling is conducted among registered voters. However, USC/LA Times, Rasmussen, Monmouth and Reuters/Ipsos conduct a likely voters screen. That screen gives Trump a major boost. Since the conventions there have been eight polls of likely voters, giving Clinton a 45.4/41.3 advantage. Trump has even managed to be within one point of Clinton on three of those polls.
Of course, even these polls put Trump behind. No matter how you slice the numbers – and it’s easy to cherry pick polls you like – Trump simply has not shown that he can break through his polling ceiling. With 90 days until Election Day, Trump is running out of time to show he can really move his numbers.