You’ve probably already OD’d on reading about Friday’s cryptic Comey memo, which re-opened the email can of worms (if only to reveal the can was empty), from every legal and messaging angle. I don't have anything to add about it from those perspectives, so I’ll just stay in my lane here and answer the questions about “what does this do to the polls?!?” It’s a pretty simple answer, though: nothing so far. The Daily Kos Elections model is still holding steady in the mid-90s, currently giving Hillary Clinton 96 percent odds of victory according to Monday’s numbers.
It shouldn't be hard to see why: a) many people simply haven’t heard the story (believe it or not, most people don’t follow the news as obsessively as your average Daily Kos reader), focusing instead on football or on earning enough money to survive or simply having a life on the weekend; b) many people who’ve heard the story just walked away from it with more question marks than they had before; and most importantly c) thanks to the intense levels of political polarization in today’s society, most people who fully digested the story already made up their minds long ago, and what few truly undecided swing voters still remain are more likely to be making their last-minute decisions based on their own personal economic situations or the ol’ “have a beer with” test, instead of the baffling sequel to a story that was highly technical and boring the first time around.
(There’s also: d) 21 million people have already voted, which we’ll discuss in more detail shortly. And while the early voters are probably the most hardcore partisans of the electorate, the ones least likely to be swayed by a story like this, that’s still possibly 1/5th of the electorate (there were about 125 million votes cast in 2012) that is already locked in.)
Part of the problem, though, is that we haven’t seen a lot of polls with field dates extending through Friday. At the state level (which is what our model is interested in), the latest round of CBS/YouGov polls extended through Friday (and also had a “brief recontact survey” on Saturday, though it’s unclear whether that affected the polls’ toplines). Those polls’ results looked entirely normal compared with previous averages: Clinton up 8 in Pennsylvania, up 3 in Colorado and North Carolina, and down 2 in Arizona. (You might be concerned about the one poll that came out over the weekend that really falls in the “bad news” category, a Florida poll from Siena that had Donald Trump leading. It, however, was completed on Thursday.)
Similarly, at the national level (which our model doesn't ask about), all we’ve seen that covers the post-Friday period were the tracking polls from the Washington Post and IBD. Those saw a drop of either one or two points from before; that’s small enough that we’ll need to wait several more days to see if there are further drops or if that’s just standard day-to-day fluctuation. The Huffington Post Pollster rolling national averages are still at a 7-point lead for Clinton in the 2-way race and 6 points in the 4-way race (though it’s closer if you switch to the “less smoothing” setting, which is to be expected since, again, all we’ve seen in the last few days is those trackers). The first fully post-Friday poll will be released on Monday morning, from Politico/Morning Consult, though, so by the time you read this, depending on the result, people may have resumed running around with their hair on fire.
You might be wondering, “Yeah, but that Florida poll from Siena had Trump up by 4! How can the numbers have stayed the same?” Well, we’re so jammed with polls by this point in the cycle that one poll that stands out like a sore thumb tends to get quickly overwhelmed by all the rest of the data. Saturday and Sunday saw two other Florida polls released: a Marist poll with Clinton up 1, and a Gravis poll (for Breitbart) with Clinton up 1, but that gets turned into a +4 thanks to the correction for partisan pollsters. Taken together, those three polls basically cancel one another out, having little impact on the overall Florida average (which is currently a 46-43 average lead for Clinton, and a 91 percent chance of Clinton winning the state).
Which brings me to that other poll from this weekend that stood out, in the other direction: the one from Craciun Research giving Clinton a 47-43 lead in Alaska. Although there aren’t anywhere near as many Alaska polls as there are Florida polls, there are still enough recent Alaska polls that this new one only moved the average a small amount; the Last Frontier is still seeing an average lead of 42-36 for Trump, with Clinton still having only a 1 percent chance of winning there.
Finally, there’s one other state that you may be wondering about, even though no polls have come out there in more than a week: New Mexico, which is in the news because Donald Trump made a left turn in Albu-koi-kee while on his way to the Coachella Valley Carrot Festival, and held a rally while he was there. That’s apparently part of a leaked plan to start putting some blue states (Michigan, New Mexico, and Wisconsin) into play. That, however, is not based on any sort of polling reality: our New Mexico average gives Clinton a 43-35 lead on average, and a 99 percent chance of winning.
In fact, what’s really going on is that paths are getting further and further closed off for Trump, and he has to start looking elsewhere for the kind of triple-bank-shot he’d need to win. You’re probably already familiar with the proposition that Clinton only has to win Colorado, New Hampshire, Pennsylvania, and Virginia to win (and also Michigan and Wisconsin, if you think of them as swing states); she can still lose Florida, Iowa, Nevada, North Carolina, and Ohio so long as she holds that firewall.
But now it’s looking extremely likely she’ll win Nevada as well. That’s not based so much on the polling, which gives her a 45-43 lead on average and an 80 percent chance of winning, as what’s happening with early voting. The early voting numbers are, while encouraging, kind of ambiguous in, say, Florida and North Carolina. However, they’re very unambiguous in Nevada, according to the mighty Jon Ralston. He reports that Democratic voters are almost on pace with where they were in 2012, with Democrats with a statewide lead of more than 30,000 votes; they’re getting to the point where Trump would need to win by at least 20 percent among independent voters in order to have a shot at winning in Nevada.
That’s important because, as few electoral votes as it has (6), it’s a state Trump absolutely can’t afford to lose. A scenario where Clinton wins Nevada but Trump somehow wins New Hampshire or even Colorado still wouldn’t get Trump to 270 votes. He gets 263 if he wins New Hampshire but loses Colorado, 268 if he wins Colorado but loses Nevada. If he loses Nevada, he’d need to win New Hampshire plus Colorado, or else a bigger state like Michigan, Pennsylvania, or Virginia. (Substituting Wisconsin for Nevada would get him to 269, where he’d have to hope for a victory in the House.)
The other parlay would be to fit New Mexico somewhere in there. New Mexico is about the same size as Nevada (with 5 electoral votes). So, conceivably, Trump could win by losing Nevada but picking up Colorado and New Mexico instead, which would take him to 273. Of course, that takes you back to the more fundamental problem, which is that he has a 1 percent chance of winning in Colorado, and, separately, a less than 1 percent chance of winning in New Mexico.
Finally, Democratic odds of having 50 seats or more in the Senate are holding steady at 65 percent. Here too it’s the same old story: there are three core races—Nevada, New Hampshire, and Pennsylvania—and the Democrats need to win two of the three (assuming they’ve already won the easier races in Illinois, Indiana, and Wisconsin). The rank order of those three tends to fluctuate from day to day; on Sunday, we had 61 percent odds in Pennsylvania, 56 percent odds in New Hampshire, and 50 percent odds in Nevada, but on Monday, we have 63 percent odds in Pennsylvania, 58 percent odds in Nevada, and 48 percent odds in New Hampshire (with Maggie Hassan and Kelly Ayotte literally tied at 45.3-45.3 in our polling average).
There are also three more states where the Democratic candidate is still in contention: Missouri, where Democratic odds are at 40 percent, North Carolina, at 30 percent, and Florida, at 13 percent. Remarkably, we got two polls in North Carolina’s Senate race on Sunday, and both of them had the race exactly tied. If that holds, with five different races with a margin of two or less in our polling average, it could well be (in echoes of the 2006 election) days after the election before we know who holds control of the Senate.