At this point, there's not much doubt that Trump's approval ratings are falling, not precipitously, but steadily. On trend, he will shortly be below 40 percent in the 538 composite poll tracker, where he hasn't been in quite some time.
Of course, this doesn't mean he's doomed in 2020 or anything like that. Still a long way to go. Bur it's hardly a point in his favor either. Nate Silver's rundown of the data on 538 does a good job of summarizing the various ways in which this trend might play out.
I think Silver has the right of it that the biggest potential bad implication of this trend for Trump isn't so much the rating itself but rather what the current political dynamic may indicate about Trump's strategy for 2020.
"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.
Plenty of presidents, including Obama, Clinton and Reagan, recovered from poor midterms to get re-elected. But those presidents typically sought to pivot or “triangulate” toward the center; we don’t know if the political rebound occurs if the pivot doesn’t. Instead, Trump has moved in the opposite direction. Despite some initial attempts at reaching out to the center, such as in passing a criminal justice bill in December and issuing trial balloons about an infrastructure package, Trump’s strategy of shutting down the government to insist on a border wall was aimed at placating his critics on the right, such as Rush Limbaugh and Ann Coulter, and members of the House Freedom Caucus.
Maybe Trump took some of the wrong lessons from 2016. Trump may mythologize 2016 as an election in which he was brought into the White House on the strength of his base, but that isn’t necessarily why he won. And even if it was, trying to duplicate the strategy might not work again...
[P]erhaps 2018 is a better model for 2020 than 2016. In the midterms, voting closely tracked Trump’s approval ratings, and he paid the price for his unpopularity. According to the exit poll, midterm voters disapproved of Trump’s performance by a net of 9 percentage points. Not coincidentally, Republicans also lost the popular vote for the House by 9 percentage points.
There’s plenty of time for Trump’s numbers to improve, but for now, they’re getting worse. So while the shutdown’s consequences may not last into 2020, it has been another step in the wrong direction at a moment when presidents have usually pivoted to the center."
Of course, I guess it's possible that the current dynamic may change and Trump will start getting credit for standing firm on the wall. But the latest data suggest this is very unlikely indeed. Check out the data below from the latest Pew poll and how few people both support the wall and would not want to see the government re-opened with getting one.