I've spent part of this week reviewing the black vote in 2016, and anticipating the 'slant' in 2020. So I'll lay out my observations:
1. If you use the Pew research done....2016 was an odd year for black votes. Roughly 7-points lower than in 2012 or 2008. One might get the impression that neither Trump or Clinton thrilled the general black population.
2. Trump basically got 8-percent of the black vote that did show up in 2016. Right now, if you use the various data collections....they suggest that he's in the 20-percent range on black votes.
3. If you use CNN-data, then it suggests that black men are three times more likely (in 2016) to vote for Trump, than black women. No one can really explain this particular detail.
4. Best estimate, four million blacks didn't vote in 2016. In a couple of red-state wins for Trump....it probably mattered.
5. If you use the theory that roughly 25-percent of blacks by 2020 will vote for Trump (three times the 2016 number), what states really matter? The higher populated states with blacks? Mississippi, Florida, Maryland, Louisiana, They all have 30-percent or more of their population as black. Trump won three of these in 2016....leaving Maryland as the state with 'flip-potential'.
6. The one city that matters? That's an interesting topic. Detroit has a statistical number of 82-percent black. It may not matter because Trump won Michigan, but it was a fairly slim win in 2016.
So I would suggest that between the black and Latino 'plus-up'....there's probably well over two-million votes that might be flipped in 2020....if Trump runs again.