Trump, Unpopular

Enough to start convincing Republicans that there's a problem.

Wow that’s one unliked president.

Donald Trump is at or near new approval lows in the various survey aggregators. He’s down as I write this to 39.1% at FiftyPlusOne. Nate Silver estimates him at 41.3% (not quite a low there; he’s been at 41.0, including early yesterday). The New York Times says 41%. And Cook Political comes in at 41.1%. In each case, that’s down from the beginning of the month, although not sharply so (basically between half a percentage point and a full percentage point).

To be sure: he’s nowhere near overall record lows. The historical floor for presidential approval during the polling era (so no Herbert Hoover) is about 25%: Harry Truman, Richard Nixon, Jimmy Carter, George H.W. Bush, and George W. Bush all spent time at around that mark.

But Trump is making history nonetheless. Using FiftyPlusOne and the data at G. Elliott Morris’s site, and if we’re considering him a first-term president, then he’s recently passed his own record and is now dead last in net approval at this point into a presidency. Indeed from Dwight Eisenhower through Trump II only Joe Biden and Trump I ever dipped below Trump’s current net approval of -18 at any point in their first two years. He’s basically one bad poll away from beating the record for lowest net approval at any point in the first two years; he’s already within a percentage point of Biden, and two of his own first term.

If we want to follow Seth Masket and consider Trump a second-term president, he’s doing a bit better — sixth place out of seven through the first 53 weeks, with only Richard Nixon scoring lower. Nixon’s decline was truly impressive – after all, he won 49 states in November 1972 only to have his support collapse by January 1973. As Morris points out, Trump had the second-biggest decline one year into his second term (and that was a week ago, when he was at only -16.2 net approval.

Not only that, but among second-term presidents (beginning with Ike, but excluding Lyndon Johnson) only Nixon and George W. Bush hit a lower point in net approval than Trump is at right now. At any point in their second terms.¹

By the way, on Nixon: It’s absolutely true that his initial collapse was almost certainly all about Watergate, with the scandal breaking open early in his second term and yielding constant headlines in March and April. Watergate certainly still mattered in 1974 – the Saturday Night Massacre was in October 1973, and after that impeachment was always looming. But Watergate was joined by the Arab oil embargo that also happened in that busy October, and not only created gas lines and shortages but also plunged the nation into the “stagflation” recession that continued well into Gerald Ford’s presidency. Yes, people don’t like the economy right now, but for the most part Trump seems to have earned his unpopularity without a recession, without a big inflation spike, and without a casualty-producing war.

That’s all just basic facts. Now I’m going to reprise my most speculative and unscientific theory that I’ve ever blogged about: The 40% rule. Basically, the (speculative! unscientific!) idea is that falling below 40% is the point where the president’s party starts really understanding how unpopular the president really is, and starts acting with that in mind.

Why 40%? Here’s the logic. First of all, partisans always interpret polls optimistically. And it’s easy to do. After all, we never know exactly what the truth really is, and it’s easy for partisans to just choose the polling average – or the individual survey – that matches their hopes. Thanks to house effects (even high-quality polls might systematically lean a bit to one party or the other) and random variation, there’s always a fairly wide spread in what different polls say. Over the week so far, I see polls ranging from 37 to 45% approval.² Which means optimistic Republicans are going to think he’s basically at 45%.

That’s step one. Step two is that partisan optimists can easily conclude that a 45% approval president will probably bounce back a bit – say, 3 to 5 percentage points – to wind up back in break-even territory. It just doesn’t seem like much. And it sort of makes sense. If the president, like Trump now, has recently slumped? He’s likely to regress to mean and regain that ground. If he’s improving? That’s obvious: Just project the trend to continue.

All that has always been true. For both parties. Not just Republicans. It’s presumably even more the case now that more people are getting their news from party-aligned media.

Granted, not every partisan is an optimist, but most tend to be about such things. After all, if you like Trump, you probably don’t think that you’re a freakish loser who likes an unpopular president; you’re a lot more likely to think that you’re normal, and others either agree with you now or will after they understand things better – or after the president’s achievements have had time to sink in.

Once the approval numbers fall under 40%, however, it gets increasingly difficult to believe the president is actually popular. And that’s when party politicians beyond those in the toughest seats start defecting, and it gets harder and harder for the president to persuade those politicians on anything.³ That doesn’t mean (for example) that House Republicans will go from enthusiastic Trump supporters to filing impeachment resolutions overnight, but it is a real dividing line. It’s the point, for example, where co-partisans of a first-term president start to entertain the possibility of a nomination challenge.

Or at least that’s what I see.

1. Johnson is excluded from the charts I’ve linked to presumably because “same point” of his presidency doesn’t align with his re-election. Looking around, it appears to me that LBJ might or might not have gone below -18 net approval. It’s worth adding a caveat here: Over time fewer and fewer people are neutral on presidential approval, especially at the early stages of the presidency, with more and more people who don’t approve of the president’s performance moving right away to disapprove. Still, Trump’s numbers are awful, even adjusting for that.

2. And that’s just looking at quality polls. There are always junk surveys out there, although I suspect they don’t fool many political professionals. I could be wrong about that, however!

3. Remember as always that congressional Republicans vote for plenty of things that Democrats hate because they support them, not because Trump tells them to.