Time for a good poli-sci wonk post on a slightly chilly October day.
We're 20 days out from the election and things are edging in Clinton's direction. Things would have to make a significant turn to give Trump a win...but not that much of a term.
If we use Real Clear Politics' framework, they have Clinton at 256 electoral votes before toss-ups. Let's give her Minnesota's ten EV and Maine's statewide 2 (they do one each for their two congressional districts and Trump is on pace to win the more-rural Second district) for sake of argument and get her to 268, two short of a 270-vote win.
Trump would then have to win, going east to west, New Hampshire (down 3.6% per RCP) North Carolina (down 2.6), Florida(down 3.6) Ohio (up 0.7) Indiana (up 4.5), Iowa(up 3.7) Arizona (down 0.2) and Nevada(down 4.2). If Trump can run that table, he gets to 270 on the schnoz.
If he gets a 4% swing from various factors (bad poling, some extra-ugly Wikileaks stuff worse than the confirming-Democratic-stereotyping stuff we saw in the recent batch, a new terror attack, bad health from Clinton) and avoid having another moral hasmat spill like the one that tipped me into the Never Trump camp two weeks ago, it could easily be done.
If we get that 270-268 squeaker, all kinds of wonk's-dream (or nightmare) scenarios could unfold
(1) Hanging Chad-Part Deux. One of those above states will likely be recount close, where it's at least too close to call until the wee hours of Wednesday morning or maybe later. If it is really in recount range (under 0.2% or so), it will be lawyer-time, as every friggin' absentee and provisional ballot will be double and triple checked.
(2) The McMullin Factor. Utah is listed as a Leaning Trump state by RCP, but it might be better placed in the Toss-Up catagory, as he is in a three-way race there with Clinton and former CIA guy and BYU grad Evan McMullin running as a straight-arrow conservative alternative to Trump. McMullin only managed to get on the ballot in 11 states, but his birth state of Utah is one of them.
McMullin has been doing well in Utah polling, enough so that 538 did a piece noting that he could well win Utah. If you envision the 270-268 Trump scenario above and then slide Utah's six EV to McMullin, it becomes 268 Clinton, 264 Trump, 6 McMullin. Then, no one gets a majority and the House (voting by state delegation rather than individual) gets to pick from the top three. The Senate would get to select from the top two, leaving then without the option of slipping us a Mindy Finn (McMullin's running mate) and having to settle on either Pence or Kaine.
The 538 came before the most recent polling out of Utah which had things 30 Trump, 29 McMullin, 28 Clinton. If that isn't an outlier and more folks become aware of a non-protest-vote option beyond the two standard evils, we could have things thrown to the House.
(3) Unfaithfully Yours- Just because the Republicans get 270, there is no guarantee that Trump will get 270. Most states require electors (generally party loyalists unlikely to go rogue) to vote for the party's nominee, but it is an open constitutional question whether those state-law requirements are enforceable. There's usually a rogue (or "unfaithful") elector every decade or so, but none that's changed an outcome of a presidential election.
If one Republican elector would go rogue, that would knock Trump out of a majority and throw things to the House as in the McMullin scenario. An interesting scenario might emerge if multiple electors go rogue and vote for different candidates. What if things are 267 Trump, 267 Clinton, 1 McMullin, 1 Rubio 1 Cruz and 1 Sanders (as a Democratic elector joins the going-rogue party)? How do you break the three-way tie for third?
Some interesting and real possibilities. Long-shots, to be sure, but very feasible.