A headline from CNBC today, as the president’s approval rating sinks to dangerous levels for his party two months out from a midterm:

If you’d showed me a headline like that a year ago, I’d have guessed it was effectively impossible for an incumbent to be sitting at 40 percent approval at the same time. Certainly it’d be impossible for that incumbent to see his approval *dropping* in that sort of economic environment.

And yet:

The most amazing thing about this, notes Enten in his piece for CNN, is that Trump’s approval rating this year has been famously steady. I noticed that myself a few weeks ago. Until recently, the guy simply didn’t budge from 43 percent or so. Rain or shine, weird tweets or no tweets, Mueller news or no news: 43 percent. People’s minds about him were made up. Love him or hate him, you had your view and nothing was going to change it.

But then something did change. Not a lot, but meaningfully. When you’re stuck in the low 40s on approval, all it takes to turn a bad midterm into a wipeout is a further decline of a few points. And Trump is, as Enten notes, in decline. In fact, his average approval at RCP today is the lowest it’s been in six months.

“[S]omething the President did had a real impact on public perception,” Enten surmises. Right — but what? What irresistible force managed to budge the immovable object of his 43 percent approval? Or is it less a case of a single incident jolting his numbers than a steady accumulation of pressure finally dislodging them? If I had to endorse a silver-bullet explanation, I’d guess there’s a backlash to how Trump reacted to McCain’s death. That’s the only thing I can think of that might have turned a chunk of Republicans from pro-Trump to moderately anti. Some veterans might have been rubbed the wrong way by POTUS rushing to raise the flag over the White House from half-staff, signaling that he was doing the absolute minimum to honor a war hero. It wasn’t just media types and bloggers who noticed it either, remember; the American Legion publicly called Trump on it, urging him to lower the flag again out of respect (which he finally did). Maybe, for some Republicans, that was one show of pettiness too many.

If you prefer the “accumulation of pressure” theory, though, you have various things to point to. There’s the McCain fiasco; there are the Cohen and Manafort convictions; there’s the Woodward book and the anonymous NYT op-ed, both of which accuse Trump of being dangerously unfit for office. Just because the public thinks it was unfair to Trump to knife him in the back anonymously in the Times doesn’t mean they’re not taking the accusations against him seriously. If just 5-10 percent of Republicans and indies who are otherwise well-disposed to him found the Woodward material and the NYT op-ed alarming, that alone might be enough to knock him off 43 percent approval. Note the independent numbers in particular:

The good news is, every trend ends eventually. The bad news is, the GOP needs this one to end soon. And not just to end but to reverse itself.

There’s a third theory. Neither a silver bullet nor an accumulation of pressure explains the dip in Trump’s ratings. It’s a seasonal thing, a matter of voters starting to pay attention to politics again as the midterms approach. He wasn’t at 43 percent for months because opinions about him had turned to granite, he was at 43 percent because most Americans had checked out for the summer and were giving pollsters their general impression of Trump when asked. Now that they’re watching the news again, those impressions are being refined. If that’s what’s happening, even with economic fireworks bursting overhead, that’s probably not a trend that’s going to turn around much before November. And given the spectrum of other ominous numbers for the party right now, that ain’t good.