With one day left before long-awaited November 8, there is a consensus of opinion outside official GOP circles that Hillary Clinton is probably going to win a plurality of the national popular vote and a majority of the Electoral College. Among those media organizations that project win probabilities, all of them have Clinton as the favorite, with percentages ranging from FiveThirtyEight’s 65 percent (a figure that has occasioned some controversy with other forecasters and with Democrats) to the Huffington Post’s 98 percent and Princeton Election Consortium’s 99 percent plus.
And yet, a host of factors, including widely ranging poll results, a significant undecided and minor-party vote, a large and unusual map of battleground states, and a general sense that this is an irrational election year, are all keeping a lot of observers — not to mention members of Congress, potential executive and judicial-branch nominees, and plain voters — nervous. So here are some questions and concerns to keep in mind before placing any large bets.
Are the polls showing some Trump surge? Not really. Though there is something to be said for the idea that Trump has closed the polling gap he has experienced against Clinton for most of the general-election campaign. The two most-cited polling average systems vary in how much they rate relatively old surveys: RealClearPolitics drops them entirely while HuffPost phases them out slowly. Methodological justifications aside, this means RCP tends to magnify very recent trends, which at the moment benefit Donald Trump. So it’s not surprising RCP gives Clinton a 2.2 percent lead in a four-way race, while HuffPost gives her 4.6 percent advantage.
But averages aside, there are some unusual ranges of opinion shown by the polls this year. The week after the final presidential debate there was one national poll showing Clinton ahead nationally by 14 points, and another showing her down one point. There are also variations by types of polls, with Clinton typically doing best among traditional live-interview polls (of which there have been fewer than usual this year), less well among online polls, and least well among automated phone polls. And there have been two major tracking polls (regular surveys of a fixed sample of voters) in the mix this year, one of which (USC/LATimes) usually has shown Trump leading when no one else did (mostly because of how the poll measures and weighs likelihood to vote).
Looking at averages avoids the problems associated with the natural partisan tendency to look only at favorable polls, but could blur methodological differences that will eventually be shown to matter a lot.
All these complications are, well, more complicated at the state level where the mix and frequency of polls, along with different sets of pollsters, varies even more.
Suffice it to say that if Donald Trump wins the national popular vote, it will be a shock, though not a terminal shock. One estimate is that the average pollster missing the final result by three points would not be all that out of the ordinary. After all, the polls on average underestimated Obama’s percentage of the vote in 2012 by 2.7 percent. The same kind of error could hide the advent of a President Trump — or the emergence of a big Clinton win.
Okay, Trump is likely trailing in the popular vote. But based on what we know about the key battleground states, does he have a realistic route to 270 electoral votes? The short answer is “maybe yes.” That is because he has a decent chance of breaking through what we have thought to be a “Clinton Firewall” group of states he was — even as of very recently — not supposed to have a chance to win.
The general-election “battleground map” (largely based on the results of the relatively close 2012 election) going into this election consisted of Colorado, Florida, Iowa, Michigan, Nevada, New Hampshire, North Carolina, Ohio, Pennsylvania, and Virginia. (Some might have added usually Republican Arizona and usually Democratic Wisconsin.) During the course of the campaign, two battleground states, Iowa and Ohio, have pretty regularly leaned to Trump. Meanwhile, Colorado, Michigan, New Hampshire, Pennsylvania, and Virginia have pretty regularly leaned to Clinton; they are sometimes called her “firewall.” Florida, Nevada, and North Carolina, along with the second congressional district of Maine (which assigns its electoral vote independently), have emerged as true battlegrounds. The basic math of getting to 270 electoral votes requires Trump to take Florida, Nevada, and North Carolina. But even if he wins them all plus the second congressional district of Maine, he’d still need one “Clinton Firewall” state to get to 270. Late polls showing Trump pulling even or near-even in New Hampshire have given him his first really plausible “path to 270.”
But bad news for Trump from early voting in Nevada (more about that below) may mean he’ll have to win another “Clinton Firewall” state, and none of them are within easy reach. Meanwhile, if Clinton wins North Carolina and Nevada, it may take Trump a couple of “firewall states” — say, Michigan and Pennsylvania — for Trump to reach 270. And if Clinton wins Florida, that is probably the ball game no matter what.
This just brings us back to what might be the key fact in this election: Clinton has many paths to victory, while Trump has few and each of those is a bit of a stretch. However, if he pulls off the sort of triple bank shot it could take to win 270 electoral votes, there is a decent chance he could do so without a popular vote plurality. For that to happen to Democrats twice within 16 years would be pretty maddening, eh?
Clinton’s get-out-the-vote operation in battleground states is generally considered much stronger than Trump’s. Is that likely to be a difference maker? Or does Trump have his own ways of countering that? One of the many unusual things about this presidential general election is that one of the two major-party candidates has a fine, conventional get-out-the-vote operation, complete with rich voter data for targeting purposes and a lot of field offices, while the other has close to zilch. This doesn’t mean Trump won’t benefit from the RNC’s investments in voter data and message targeting, and state GOP turnout operations. But it’s not much of a fair fight. Estimates of the value of Clinton’s ground game in the states where it is operating range from 1 to 3 percent of the vote — which is, needless to say, a lot in a close state.
But it is important to remember that Trump has his own, unconventional get-out-the-vote, base-mobilization technique: It’s those big monster rallies he holds. A week or so ago, some observers were amused by Trump’s decision to personally campaign all over the “Clinton Firewall” states; not only was he spending time in difficult turf when he could have been focused on the “true battleground” states, but he seemed to be spreading himself too thin. But that’s the perspective of people used to candidate-centered rallies being the cherry on top of the whipped cream of a campaign’s final effort. For Trump, it’s the whole sundae. How much good it will do him no one knows.
The odds are Clinton’s ground game will make a close win more comfortable.
What have we learned from early-voting results? There’s a fair amount of talk in the media that early-voting Latinos might win it all for Clinton by helping her carry Nevada and Florida. Does the evidence support that? The implications of early voting are probably not as clear as suggested in some of the reports claiming that early voting shows a clear Clinton victory. For one thing, no one knows infallibly how early voters have voted. For most early-voting states, we know the party registration of early voters. But there are always crossover voters, and it’s at best a guess exactly how independent voters will split. A few southern states (e.g., North Carolina) also report voting by race in order to comply with the Voting Rights Act. That enriches the ease of interpreting the data to a significant extent.
The other imponderable is whether one party or the other is disproportionately driving its constituents to vote early, which produces big early-voting margins but not necessarily victory. For example: The final survey of Iowa by the much-respected pollster Ann Selzer shows Clinton winning early voters (roughly a third of the electorate) in that state by 22 points. But Trump leads among non-early voters by 21 points, leaving him ahead seven points overall.
Another way to put it is that Trump’s very lack of organization means a higher percentage of his vote probably won’t show up early. But they may well show up late, and in the end a vote’s a vote.
On the other hand, if a very high percentage of the vote is cast early, a sizable lead could well be decisive. With more than two-thirds of the vote in Nevada usually cast early, the big Democratic advantage posted in early voting will be very difficult — difficult, but not impossible — to overcome on Tuesday.
It is also helpful to understand the varying rules for early voting. In Florida, for the first time, those who voted absentee ballot in the last election (who usually lean Republican) were automatically mailed ballots instead of having to request them. That is increasing the overall number of early voters, but with unclear implications for the parties and candidates.
Will late news developments from the FBI or WikiLeaks change anything? Nine days ago FBI director James Comey freaked out a lot Americans by blandly informing Congress his people were looking into a few hundred thousand emails found on Anthony Weiner’s laptop. On Sunday he blandly informed Congress his agents had found nothing to change his July recommendation against any prosecution of Clinton over the emails. It is probably too late to determine if this second action by Comey will have any impact on swing voters, such as they are, or on the enthusiasm levels for either candidate.
WikiLeaks has also kicked out some more illegally procured Podesta emails, and has constantly hinted more are on the way before Tuesday.
It’s always possible that last-minute news can have some influence on a small number of voters. But if you add up the nearly one-third of voters who have already cast ballots, and the very high percentage of the rest who are highly partisan, we’re not talking about a lot of impressionable voters. And in most cases, “news” isn’t as mixed as it sometimes seems. For example: Comey’s letter to Congress today would seem to exonerate Clinton, but on the other hand, it is one more very crucial day in which the subject of “Clinton emails” will dominate the news. Don’t expect a big effect.
Will “enthusiasm” among committed supporters of Clinton or Trump be what determines the election outcome? Or will the undecided voters decide for the rest of us? Every election we rehash ancient arguments over the value of mobilization over persuasion, of “base voters” versus “swing voters,” and of “enthusiasm” versus “late-deciding undecideds.” This election probably won’t resolve those arguments once and for all, in part because voters on both sides seem motivated by antipathy to an opposing candidate as much as by “enthusiasm” for the favored candidate. In the end, it does not matter if people motivated to vote do so with a smile or a snarl.
As for “swing” or “undecided” voters, there are two conflicting indicators. On the one hand, the number of truly “independent” voters has been steadily declining for years as the parties and major ideologies polarized their supporters. On the other hand, an unusually high percentage of voters have been expressing an unwillingness to decide between Trump or Clinton, sometimes by expressing a preference for minor-party candidates Gary Johnson and Jill Stein (or in Utah, Evan McMullin). The undecided vote has been steadily declining in the last month, and there are no particularly strong indications what is left of it will “break” to either candidate. So the odds are turnout — and how shrewdly the campaigns have estimated and targeted their votes — will be the name of the game. As noted above, there is more than one way to skin a cat — or to get a “base” voter to the polls. If it really is all about “enthusiasm,” Trump may have an advantage. If it is about organization, Clinton is in command.
When will we know who has won? If Florida gets called early on Tuesday night, particularly if it is called for Hillary Clinton, we may know a lot pretty early. But do stay tuned to make sure that if it is called Florida is not subsequently uncalled, or left to the U.S. Supreme Court to decide. With that court now deadlocked 4-4 between liberals and conservatives, they probably will not get to decide another presidential election anytime soon.