“I’m just gonna tell you point-blank: Donald Trump can win this if he hangs on,” said Rush Limbaugh to his audience yesterday. “I actually, folks, really believe that if he hangs on and continues down this road, at some point there’s gonna be a shift in public opinion and the vast majority of the American people are gonna end up with him on this — if he hangs in there.”

I don’t believe Rush “really believes” that.

That’s the sort of thing you say when your audience is pure populist, the cardinal rule of populism being to show your mettle by standing and fighting even when you can’t win, and even when doing so will actually cost you on balance. Limbaugh can’t betray ambivalence here, especially when Trump himself is still standing firm for now. It would damage his populist cred. But he also can’t say “this isn’t worth it but let’s fight on anyway” because, after all, who wants to be told that the fight is pointless? So he’s stuck claiming that victory is still possible and that the public will see the light — if only this goes on a bit longer.

That would be a hard sell even if the polls were frozen right now, with voters still withholding judgment on POTUS as this plays out. They’re not frozen. For Rush’s prediction to come true and the public to rally to Trump, we’d need an outright reversal in the trend. What’s supposedly going to drive that reversal?

The key point from this new survey isn’t that Trump’s job approval is down, it’s who’s driving the decline. For the past year his approval rating has barely moved; seemingly nothing he does can win over the majority of voters who disapprove of him and nothing he does can alienate the minority who approves. But the shutdown has upset that dynamic. For the first time in a long time he’s seeing some Trump-friendly groups frown.

The biggest one-month drop came from white women without a college degree, who went from +20 on Trump a month ago to -4 now. If you want to chuck that poll because it comes from NPR and PBS and thus can only be “fake news!”, remember that Trump’s favorite pollster has also seen him drop from 49/48 in mid-December to 44/55 today. The share of people in the RCP poll of polls that disapproves of him is tied today for its highest mark since January 23 of last year. And the NPR poll isn’t the only one to find him suddenly losing ground with groups that are usually pro-Trump. CNN saw the same thing happening a few days ago.

There are many ways to respond to all that. “Eh, it’s a momentary blip. Once the shutdown’s over he’ll be back to normal.” That’s likely true, although not definitely. “If Trump’s numbers are suffering, Democrats’ approval must be suffering too.” Also likely true, although presidents facing reelection have more to worry about here than the opposing party writ large does. “The polls were wrong in 2016, they’re wrong now, they’ll be wrong in 2020.” That’s the blind-faith approach to bad news, but okay. “I really believe that if he continues to go down this road the polls will magically turn around” is weaker than any of those replies. It’s blind faith minus even a recent historical example to point to for evidence.

Here’s the truly scary number from the NPR/PBS poll:

The president also faces some significant headwinds for re-election in 2020. Just 30 percent of registered voters said they will definitely vote for Trump in 2020, while 57 percent said they will definitely vote against him.

Just 76 percent of Trump supporters, 69 percent of Republicans and 58 percent of white evangelicals say they will definitely vote for him. Many, if not most, of them will likely vote for the president, but their softness in supporting him for re-election is a sign of vulnerability.

I think Trump really believes that a “silent majority” supports him, never mind that he lost the popular vote three years ago, never mind that the *highest* approval he’s ever had in the poll of polls was 46 percent (two weeks after his inauguration). If he’s right then he can continue the way he’s going: Feed the base, get them excited to vote, trust that the silent majority will deliver. But he’s betting his entire presidency on that proposition. If he’s wrong, if in fact 50+ percent have now been so alienated that they’re resolved to voting against him, then he’s cooked. That is, for all the hype about Trump’s base slipping during the shutdown, that’s not really his chief worry. They’ll come back in due time. The worry is that some small but momentous chunk of independents and centrist Democrats have been so deeply estranged from him by now that he just can’t get to 270 electoral votes anymore. Does the trend in his job approval lately suggest that he’s easing that estrangement or making it worse?

Nate Silver’s thinking back to the midterms and wondering if Trump had the same takeaway he did:

The lesson of the midterms, in my view, was fairly clear: Trump’s base isn’t enough. The 2018 midterms weren’t unique in the scale of Republican losses: losing 40 or 41 House seats is bad, but the president’s party usually does poorly at the midterms. Rather, it’s that these losses came on exceptionally high turnout of about 119 million voters, which is considerably closer to 2016’s presidential year turnout (139 million) than to the previous midterm in 2014 (83 million). Republicans did turn out in huge numbers for the midterms, but the Democratic base — which is larger than the Republican one — turned out also, and independent voters strongly backed Democratic candidates for the House.

Right. MAGA Nation didn’t stay home for the midterms. Turnout was historic. They just got out-voted by Democrats and independents, both of whom also turned out massively. “Given that,” wrote Silver, “perhaps 2018 is a better model for 2020 than 2016.” Perhaps so. And if it is, it’ll almost certainly involve a Democratic nominee unsaddled by the scandal baggage or unpopularity that helped sink Hillary. The shutdown will come and go but the feeling among reluctant 2016 Trump voters that there are too many bad surprises from him to offset the good ones might not. Which is why Rush said what he said: To believe that this process isn’t damaging his reelection chances instead of helping them, that this hasn’t been a backfire, you need to believe that the public will have an almost magical awakening about the virtues of a shutdown standoff over the wall, and soon.

I think Trump will hang in there until the SOTU now, just to give one more big speech about all of this and hope that it sparks the awakening.