After a recent bout of volatility, the polls in the presidential contest have mostly settled down into a consensus that Hillary Clinton is likely to defeat Donald Trump by a popular-vote margin very similar to that by which Barack Obama defeated Mitt Romney four years ago.
I say “consensus” because there are two tracking polls that have been very friendly to Trump that still show him ahead: the USC/L.A. Times survey (Trump up by 3 in a two-way matchup) and the IDB/TIPP poll (Trump up by 2 in a four-way matchup).
But that’s it: Every other creditable public poll released in the last week has Clinton ahead, and as Nate Silver noted late yesterday, they are at the last minute beginning to converge in terms of Clinton’s margin:
[T]he number I have on my mind today is “4.” That’s because it kept coming up over and over as national polls were released today: It seemed like every pollster had Clinton leading by 4 percentage points.
Silver actually measures the polling average as giving Clinton a 2.9 percent lead, which he would bump up to 3.8 percent for the “highest rated pollsters.” RealClearPolitics’ polling average shows Clinton up by 3.3 percent. At the Huffington Post, it is Clinton by 4.6 percent. Obama won in 2012 by 3.9 percent. And it is perhaps fitting that Obama’s approval rating as measured by Gallup just reached 56 percent, his highest number since October 2012, when he was at 57 percent. For all the craziness we have gone through this year, Clinton/Trump is looking a lot like Obama/Romney. Maybe there is even more to the “partisan polarization” explanation of this election than most people grasp.
Looking at the final state polling, there is a similar picture of a slight drift toward Clinton, but more than with the national popular vote, uncertainty prevails. In the three states most likely to destroy Trump’s hopes of becoming president — Nevada, North Carolina, and Florida — he still has tiny, tiny leads in the RealClearPolitics polling averages. Indeed, in Nevada in particular, there’s a fascinating divergence between the polls and the sizable Democratic advantage shown in early voting. Late trends along with early-voting statistics help explain why all of the major forecasters, including FiveThirtyEight and the Upshot, now project all three of these states as winding up in Clinton’s column, albeit narrowly. There’s no huge Trump surge in one of the “Clinton firewall” states like Pennsylvania, Michigan, or Wisconsin, either, to offset the potential loss of the “true battleground” states, though Pennsylvania has tightened significantly.
All in all, a Trump win would require a pretty significant polling error, particularly if you believe the polls have not fully captured Clinton’s advantage in Election Day turnout operations. We should know a lot shortly after eight o’clock EST, when polls will have closed in Florida, New Hampshire (where late polls have been all over the place), North Carolina, Ohio (where Trump has maintained a solid lead in the final polls), and Pennsylvania. If any of those states are called quickly for Clinton, it could be curtains for Trump.