I wouldn't be surprised by solid D victories in these states. Trump's approvals have been in the toilet in WI, MI and PA since like four months into his Presidency and these states all gave Democrats resounding victories in 2018.
Add on top of this that Trump won these states by less than a percent and it increasingly looks difficult for Trump to win them again. Of course a 2016 redux where both candidates end up covered in mud and enough people say "to hell with it" and vote for Trump again is possible but it's probably much less likely than we all think.
One does not win a state in which one's disapproval is in the fifties. The only ways in which President Trump wins either of Michigan, Pennsylvania, or Wisconsin again are
(1) tampering with the electoral process
(2) splintering of the center-to-left vote among multiple nominees, or
(3) the President solving his problems of credibility.
Demographics alone disfavor this President in these three states (as practically all others) over time, and the President can do nothing to offset the tendency of strongly-D Millennial voters supplanting older, somewhat-R voters over 55 as many of the latter die off or go senile.
President Trump obviously cannot undo his gaffes, his clear support of economic elites over everyone else from Day 1, and the questions of personal integrity.
It is more likely, should Presidential disapproval nationwide be in the fifties, that some conservatives will prefer an alternative who better espouses free-market solutions and supports traditional 'family values' in deeds as well as words. Such people will not vote for the Democrat under any circumstances.
On (1), the "Deep State" is already in action. Democratic officials will be much more cooperative with the FBI this time should there be monkey business with Democrats' proprietary databases for getting out the vote.
On (2), it is up to Democrats to nominate someone presentable.
On (3), the President needs miracles.
Our reliance on the polls last time proved to be a massive disappointment and presented inaccuracies. Michigan definitely won't be voting for another TPP globalist.
People who do this sh**t are so annoying. "Polls were wrong once, therefore we can never trust polls again," without any understanding of why polling was wrong and whether and how that applies to current races. It's such lazy analysis.
I do not go to horse races, but long-shots occasionally win. That is part of the excitement of some bettors who think that they can handicap better than the supposed experts. In 2012 I did my own early handicapping of the Presidential race, and some of my cautious models made more for ease of calculation had certain key races (those in Colorado, Florida, Missouri, Ohio, and Virginia) as roughly 50-50 propositions for both Romney and Obama. These were 'independent events' in that nothing that went on in another one of those states was likely to have an effect in another. At some point, as other states became irrelevant as sure wins for the President, one of those states would be enough to win a second term for Obama. They are different enough and distant enough from each other that winning or losing could be an 'independent event'. I ruled out states such as Arizona, Georgia, and North Carolina because if Obama was winning any one of those states he was surely winning another of them (Colorado if Obama was winning Arizona, Florida if he was winning Georgia, and Virginia if he was winning North Carolina).
If the chance of Romney winning one of those states was one half, then the chance of him winning overall was one half to the fifth power, or one in 32. That allows a simple calculation that the probabilistic chance of Obama winning was 31 chances in 32, or about 93%.
OK, so Missouri spiraled out of contention, but Obama's chances got much higher (maybe 75% or so in Colorado and Virginia, maybe 60% in Ohio) while Florida remained a 50-50 proposition. At such a point, the chance of Romney winning was about one in 80. Toward October, Colorado and Virginia both became nearly sure things, little advantages for Obama in Ohio compounded, and Florida was still shaky, and giving a 10% chance for Romney in both Colorado and Virginia, about 30% in Ohio, and about 50% in Florida, my trusty calculator gave me about one chance in 740 for Romney. With independent events, the chances of everything going right is multiplicative, and multiplying small numbers together gets even smaller numbers... and bigger reciprocals.
Obviously we are nowhere near the 'time-running-out' phase of the 2020 Presidential election. We do not have a lame horse in the race, but we have one that now has inadequate speed for winning a race if it were being held today. Maybe the political culture will change, and Trump's message will become more popular. But that is a big assumption.
A long-shot won in 2016.
