We had a bunch of new polls of Pennsylvania earlier this week but the data nerds are focused on New Hampshire, Colorado, and Wisconsin right now, seemingly convinced that no matter how tight the numbers get, the Great White Whale of the mid-Atlantic will elude the GOP again. Trump’s trailed in every poll of the state (except one) this year, hasn’t he? The state’s gone blue in every election since 1988, hasn’t it? At least one poll of PA showed Romney tied with Obama there in the final days of the 2012 election, didn’t it? How’d that work out? You can find only so much electoral fool’s gold in Pennsylvania before concluding that there’s no real gold in the ground at all.
But. These same analysts will tell you, quite rightly, that once a candidate’s lead shrinks to three points or so, the risk that the polls will miss predicting the winner starts to get real. Polls of likely voters are based on turnout estimates across various key demographics; when the lead is down to three, even small mistakes in those estimates can overlook an upset brewing. Hillary’s current lead in Pennsylvania: Three points exactly as I write this. Hmmm.
Three different polls in the last five days have her lead down to one or two points, and yet Pennsylvania continues to be treated mostly as a gimme. That may be due in part to the fact that her own numbers there appear to be holding up (check the graph here) even as Trump’s have begun to climb. That’s not what’s happening in New Hampshire: Clinton is down sharply in the last few days while Trump is up sharply, which has produced the current toss-up. The thinking, I guess, is that Trump will simply run out of gas in PA after having trailed there for so long — never mind that we could have said the same thing about NH a few days ago, as he hadn’t led a poll there all year. FiveThirtyEight still gives her a better than 75 percent chance of holding the state. And yet, guess where Hillary is holding her final mega-rally on Monday night, with special guests Barack and Michelle Obama and Bill and Chelsea Clinton. Right. That’s a lot of political firepower for a state that’s supposedly in the bag.
Endless smart analyses of why Pennsylvania held so much potential for the GOP this year have been written over the last few months. The state’s been getting redder over time; the hope was that the party would nominate someone who could take advantage of that this year. The distaste for Trump felt by the sort of white college grads who populate the suburbs around Philly had seemed to extinguish his chances, though. Now we’re back to a three-point game with reports nationally of wayward Republicans — like those suburbanites! — finally coming home to the GOP amid other reports that black turnout, which is what Hillary’s counting on in Philly, is lagging. What if Trump pulled the upset to end all upsets on Tuesday night by winning PA? Let’s look at a map:
That’s Clinton 237, Trump 229 with Pennsylvania and Ohio red. It’s a testament to how difficult his path is this year that, even with those states in his column, he still needs to win Florida and probably North Carolina as well to win the election. He can get to 270 without NC, but he’d have to win multiple other battlegrounds to do it. And needless to say, there’s no realistic path without FL. If you can believe it, with nearly 5.3 million early votes having already been cast there, the GOP had a registration advantage this morning … of 1,833 votes. Follow that last link and you’ll see that Democratic strategists are so confident of victory, given the strong early turnout among women and Latinos, that one of them allegedly predicted a win of one or two points for Clinton when he met privately with Obama there yesterday. We’ll see, as Republicans in Florida are pleased with early turnout themselves. One quirk of this election, purely from the standpoint of election-night suspense, is that because Florida and North Carolina are so important, are located on the east coast, and have robust early voting, we should know pretty quickly on Tuesday how competitive the election will be. If both states are called for Trump, the rest of the evening will involve high drama and some cliffhanger vote-counting in Pennsylvania. If both (or just Florida, really) break for Hillary, you can skip the end of the movie, safe in the assumption that you know how it ends.
By the way, a new poll out today in Michigan has Clinton’s lead there down to just four points at 42/38, although you wouldn’t want to bet heavily on a guy who’s below 40 in a poll four days out from the big vote. On the other hand, three other polls of Michigan this past week have also been tight, showing Clinton up no more than three points and by as little as one. FiveThirtyEight still has her at a nearly 80 percent favorite to win — which, given the margins and notwithstanding its tradition as a blue state, seems awfully high. If Trump’s Rust Belt constituency is finally showing up at the last minute, Tuesday night is going to be pure panic for Democrats.