The month of August, with its back-to-back political conventions, has passed without much fundamental change in the presidential race. Just as it was in July, post-Republican convention polling pegs Joe Biden’s lead over President Trump at somewhere between 8 and 10 points.

Two national, live-telephone-interview polls conducted after the conclusion of the GOP convocation, by CNN and Selzer & Company/Grinnell College, showed Biden leading by 8 points. The Quinnipiac University survey put Biden up by 9 points, while Monmouth University’s poll showed a 10-point spread. These are all polls worth taking seriously. There should be more coming in shortly from the three broadcast networks and Fox News.

In each of the four polls, the former vice president held between 49 and 51 percent of the vote; each had the incumbent between 41 and 43 percent. That 41-43 percent support matches precisely the range of Trump’s job-approval numbers—approval ratings being the most reliable predictor of how an incumbent president will perform.

Worth keeping in mind is how these numbers fit into a broader context. During Trump’s first year in office, the Gallup Organization gave him an average 38 percent job-approval rating, the lowest of any first-term, elected president since the end of World War II and 11 points worse than the second-lowest, Bill Clinton. For Trump’s second year, he averaged 2 points higher at 40 percent, the lowest second year of any postwar elected president. In his third year, he ticked up another 2 points to 42 percent, the second-worst third year in office behind Jimmy Carter’s 37 percent.

That an incumbent president can have a six-month run of 50-year-low unemployment levels and still never hit 50 percent in any major national poll suggests that he’s got not just a firm ceiling, but an unprecedentedly low one.

At the same time, he has an uncanny ability to absorb the most negative of stories, like news that his lawyer paid $130,000 to a porn star to keep her quiet, without his numbers taking a hit. That is why those who think the contents of Bob Woodward’s book will be damning are so wrong; there is nothing that could be said or written, nothing that Trump can say or do, that would crack his base. Those expecting Trump to drop at the next negative story forget that it is hard to take a drop if you’ve never risen in the first place.

But this is also why those who are convinced that he will make a spectacular comeback are likely to be wrong as well. His standing is impervious to events. The vast majority of voters decided a long time ago whether they liked him or not, whether they would approve of his actions, and whether they would support his candidacy. For many voters, this cake is baked.

As for those who keep coming back to the nostrum that he won four years ago, proved everyone wrong, and can do it again: They ignore that this time, he doesn’t have Hillary Clinton as an opponent. While Clinton is an incredibly bright and accomplished person, she accumulated an enormous amount of political baggage during her quarter century on the national stage. Until the #MeToo movement arrived, her husband went through his career as if he had a Teflon coating—nothing stuck to him. But Hillary was Velcro; seemingly everything stuck to her.

Comparing how Hillary Clinton was perceived in the 2016 campaign with Biden now defies logic. The CNN poll taken at the beginning of September 2016 showed Clinton with 41 percent of registered voters giving her a favorable rating and 57 percent giving her an unfavorable rating, for a net favorability of -16 percent. Trump showed a 42-percent-favorable, 56-percent-unfavorable split, for a -14 net rating. The new CNN poll gives Biden a 48-percent-favorable, 43-percent-unfavorable rating for a net of plus-5 points. Trump’s current numbers are 40 percent favorable, 56 percent unfavorable—a net -16 points.

Obviously, we are talking about anticipating human behavior, which is always problematic. But given that voting has already begun in some states, Trump does not even have 54 days to make a comeback. Similarly, an October Surprise may need to happen in mid-September for it to have any impact. It is a good bet that a lot of votes will be cast before that first debate on Sept. 29.

I understand the caution that many in my business have after the surprising outcome in 2016, but the only way this year resembles 2020 is that they both are presidential years, Trump is the Republican nominee, and both years begin with a 2. That’s it. Alan Greenspan’s irrational exuberance has given way, in my mind, to irrational caution. This is not 2016.

This story was originally published on on September 11, 2020

More from the Cook Political Report