To update then
In the immediate aftermath of locking up the Republican nomination we saw the GOP establishment senior figures unleash their devastating broadside. As predicted though, it didn't amount to much. Basically (with the possible exception of Paul Ryan) it was led by yesterdays men and only seems to have driven wedge between the party and the base.
Democrats momentarily celebrated for about 10 days as a succession of polls emerged showing Hillary extending her lead and winning states like Arizona and tied in Georgia. This was the blue landslide they'd all told was inevitable then? Well as I said, it didn't last long. What we hear from them now is you can't take polls in isolation and its too early, as a series of new polls show Trump performing well in the key swing states. Within a week Trump had restored the lead both
Ohio +4
Pennsylvania -2
Florida -1
New Hampshire -2
In some other national polls he enjoys a lead, in others he's losing. Nate Silver is now hypothesising that Hillary's inability to put away Bernie Sanders is a drag on her, whilst Trump continues to do the hard man posturing having put away a more complete field before her. Trump for his part seems to be putting out a different policy daily (often in contradiction to what he said only a week earlier). It doesn't seem to hurt him. John Kerry was crucified for this flip flopping. Trump is much worse. However his supporters only need to listen for the bit they like the sound of. They seem to latch onto that. They then ignore the contradiction that casts it into doubt. It's pic n mix stuff.
This looks like a highly charged, emotionally unstable, low information, and increasingly paranoid electorate. Either that or American polling organisations are the world's worst and are simply flinging darts at a dartboard. It's not unheard of the see differences of 15% taken in the same period. Even from day to day you can see similar confusion. Even Italy doesn't generate this kind of volatility where you routinely see swings of 10% every week
At the time of writing, the three most recent national polls have the following (these were all reported within 48 hours of each other)
CBS/ NYT = Clinton +6
Fox News = Trump +3
Rassmussen = Trump +5 (Rass have a house bias and I'm prepared to downgrade them)
What's going on?
https://www.washingtonpost.com/opin...e32c58-1c47-11e6-8c7b-6931e66333e7_story.html
I think the battle is becoming more complex than purely geography as it happens, and Hillary faces a very real threat that she's going to get caught out fighting a campaign of explanation. Once you find yourself having to explain to the voters why your policy is better than the opponents, you're very often in trouble
The other issue of course is that of the much vaunted Hispanic vote. Trump has a narrow flight path that allows him to continue to alienate this bloc so long as doing so brings him corresponding new low-information white voters. It's all about where these voters are concentrated.
One decidely ropey looking poll gave Trump a highly improbable 55/45 win amongst Hispanics. If you reproduce that result on the Romney map, Trump flips Florida, New Mexico, Nevada, and Colorado. Guess what though. It's not enough to win. In other words his path to the white house doesn't go through the Hispanic/ Mexican community. He can afford to continue to marginalise them just so long as he doesn't over-do-it to the point where he loses Arizona, or to the point where Latino Cubans in Florida come out in sympathy for the Mexicans
Now if you increase the Hispanic turnout by 6%, and allocate it exclusively to the Democrats, and do the same for the non college educated white vote, Trump wins. If you reduce the turnout for the black vote (the consequence of the other two groups increasing) Trump wins a landslide. This is because of where these populations are located. He can afford to lose places like NM, NV, CO and CA by bigger margins, as they're already in Democrat in hands, so long as he's gaining OH, WI, PA, and IA which on this crude extrapolation, he does
Now there is an obvious flaw in the logic. Just because Trump calls Mexican Hispanics names, doesn't mean that non college educated white voters are drawn into him like a moth to a light. There is likely to be a much stronger elasticity amongst the offended group. If you operate a ratio of 2:1 then, Clinton wins, (narrowly) albeit it doesn't take much for the college educated white vote to break for Trump to make this a knife edge win for him.
After I played about with Nate's state flipping tool, trial and error, I kind of hit on a ratio of 3:2. That is to say he can alienate 3 Hispanics into voting Democrat, so long as doing so gains him 2 non college educated white voters. If he achieves this, he wins, provided he doesn't overdo it and lose Arizona in the process. The key group (not for the first time) is the Cuban vote in Florida who can't be guaranteed to vote in line of sympathy with the Mexicans given historical emnity between the two, and that the Cuban has traditionally gone for the Republican party with its older members in particular disapproving of Obama reproachment with Raul Castro