If Trump's popularity ratings remain low, the question is whether he can repeat his 2016 feat.
New polling suggests that Trump will have difficulty doing that because, for now, the 2020 election looks like it will be a referendum on him.
An NBC News/Wall Street Journal poll out this weekend found that Trump trailed a generic Democratic candidate by 7 points in a hypothetical 2020 matchup. The important statistic here isn't so much that Trump was losing (it's still early 2019, after all). It's why Trump was losing. Trump trailed because the same voters gave Trump a -6 point net approval rating (approval rating - disapproval rating).
This was not the first NBC News/Wall Street Journal poll to show Trump losing to a generic Democrat. Back in December, Trump trailed a generic Democrat by 14 points. His net approval rating in that poll was -11 points. (The average poll shows Trump is about this popular currently.) Again, the key statistic here is Trump's margin is directly related to his own popularity.
In limited polling, the well known Democratic candidates seem to doing as well against Trump by what you'd expect given his popularity. In a January Glengariff Group poll of Michigan voters, Trump trailed Joe Biden by 13 points and Bernie Sanders by 11 points. The same poll put Trump's net approval rating among Michiganders at -9 points. A poll from Quinnipiac University of Texas voters out last week showed something similar: Trump's position against the most well-known Democrats in Texas (Biden, Beto O'Rourke and Bernie Sanders) matched his approval rating.
This 2020 data does not look at all like what happened in 2016. In the election, Trump had a -22 net favorability rating (favorable - unfavorable). Yet, he only lost the popular vote by 2 points and won the electoral college.
Trump was able to win in 2016 because that election came down to a choice between him and Hillary Clinton. Clinton was the second-least-liked major party nominee of all time, and there was 18% of the electorate that liked neither Clinton nor Trump. Trump won this 18% of the electorate by 17 points and with it the election. In other words, Trump was seen as the lesser of two evils.
The 2020 polling looks a lot more like what occurred in 2018 than 2016. Last year, Democrats won the national House vote by 9 percentage points. Not surprisingly, Trump's net approval in the exit poll was -9 points. That is, if you liked Trump, you voted Republican. If you didn't like Trump, you voted Democratic.
Democrats were likely aided in 2018 by the fact that a lot more people liked the Democratic Party in 2018 than liked Clinton in 2016. In the 2018 exit polls, just 10% of the electorate viewed the Democratic Party and Republican Party unfavorably. That's only a little more than half who viewed Clinton and Trump unfavorably in 2016. Additionally, Democrats won this 10% of the electorate by a point, which is far better than Clinton did among those who disliked Clinton and Trump.
This points to the difficulty of Trump's road ahead if his own numbers don't improve: Clinton is not going to be his opponent in 2020. The default position voters have right now is to make the 2020 election about the incumbent. To win, Trump is likely going to have to make the 2020 a choice between two sides. This won't necessarily be easy, as 2018 showed. And even if 2020 is a choice between two sides, there's no guarantee that Democrats will nominate someone who is anywhere close to as disliked as Clinton was in 2016.
Of course, this leads to the question of who Democrats will nominate. The ball is in the court of Democratic primary voters. If they pick a candidate who allows the election to be about Trump, said Democratic candidate have a good chance of winning. If, however, the candidate is unpopular, it gives Trump an opportunity do what he did in 2016. He could win a choice election without a majority of people actually liking him.