Sabato's Crystal Ball

The Trump Surge

Or is it a Clinton Fallback?

Larry J. Sabato, Kyle Kondik, and Geoffrey Skelley, Sabato's Crystal Ball September 19th, 2016

facebooktwittergoogle_plusredditpinterestmail

And then, everything changed.

Well, not everything, but enough to generate the first major revision in our electoral map, and all of it is in Donald Trump’s direction for now.

Let us make our view perfectly clear: We still believe that Hillary Clinton is more likely than Trump to win the election, and she still has the advantage in the Electoral College. Yet it is equally apparent that she has stumbled badly in recent weeks, fueling Trump’s polling advance. And the Republican nominee has more pathways to 270 electoral votes than he did before.

Clinton disappeared for long stretches before Labor Day to do fundraising and failed to define any overriding positive message about what a Hillary Clinton presidency would mean. Then came the “basket of deplorables” gaffe (which some believe might actually help Clinton with her party base) and the lie by omission about her pneumonia, a serious condition that deserves scrutiny. Instead of transparency, there was a bold gambit by Clinton to avoid any health disclosure that might give credence to longtime, far-right theories that she was at death’s door. This ill-advised error reinforced the public’s view that Clinton is secretive and untruthful.

At the same time, two other developments helped Trump. His new (third) campaign team, especially Kellyanne Conway, seems to have been able to force some discipline on the unpredictable billionaire. The number of middle-of-the-night tweets has diminished, the use of scripts and the teleprompter is up, and Trump is doing events other than giant rallies filled with true believers who encourage his worst instincts — though he continues to make outrageous statements with some frequency, such as Friday’s suggestion that Clinton’s Secret Service team should be disarmed. (Almost any other candidate would be disqualified by the kinds of remarks Trump has offered. In this sense, there truly is a double standard in media coverage that benefits Trump.)

Second, and as important as anything else we’ve mentioned, the pollsters have switched from highlighting the trial heats of all registered voters to a more select group they themselves define, called likely voters. The composition of this group is influenced by the response rates of partisans and their enthusiasm, or lack of it, to vote in this election. It will surprise no one following this contest to learn that Trump supporters, and Republicans generally, are more committed just now. Clinton has not lit a fire under her supporters, and to the contrary, her missteps have disheartened them a bit. It is also true that Democrats frequently get engaged later in the campaign season, and as a result, we might expect the enthusiasm gap to decline somewhat in the seven weeks remaining until Nov. 8. Clinton’s best surrogates (Barack Obama, Michelle Obama, Bill Clinton, Joe Biden, Bernie Sanders, Elizabeth Warren, etc.) can help, but only Hillary Clinton herself can provide enough incentive for many voters to show up.

The third-party candidates aren’t helping Clinton either. Put together, Libertarian Gary Johnson and Green Jill Stein seem to be hurting Clinton more than Trump, particularly among the youngest voters — voters who have little use for Trump but also are hardly inspired by Clinton. Politico’s Steve Shepard found that one survey, from Fox News, showed  about a quarter of voters under 35 backing Johnson or Stein. Another, from Quinnipiac, pegged third-party support amongst millennials at an eye-popping 44%.

Clinton tends to do a little better in most surveys with the one-on-one matchup with Trump than in the four-way polls, and it’s easy to see why.

No doubt the nominee’s better polls have helped nudge some reluctant Republican officeholders and donors more toward Trump, and that’s a plus for him. At the same time, these close surveys may provide fuel to get more Democrats fully motivated. Most have thought Trump couldn’t win; pollsters now say otherwise, and it’s a shock to the Democratic system. Whatever Clinton’s positive message turns out to be, it’s the FOT Factor (Fear of Trump) that is her best weapon. Fear of Hillary (FOH) assists Trump, of course, but Trump himself lights the fire under most of his zealous backers.

Each of these candidates has given the other plenty of ammunition. The latest is a Trump present to Clinton: his somewhat reluctant revival of the odious birther issue. To his great discredit, Trump spent years cheerleading the crackpots, conspiracy theorists, and bigots who pushed the phony claim that Barack Obama wasn’t born in the United States, and thus was ineligible for the presidency. Trump had never renounced that position — until a few days ago, his hand forced by renewed press questioning. For the first time, Trump admitted Obama was born in the U.S., but with breathtaking cheek, he also claimed Hillary Clinton started the controversy in 2008 (false, though some Clinton backers outside the campaign hierarchy mentioned it). Then Trump laughably took credit for ending the controversy by “forcing” Obama to release his birth certificate in 2011 — after which Trump shamelessly continued to talk darkly about the matter. The outrage that flowed from Trump’s rewriting of history galvanized not just African Americans but Democrats of all stripes. Will this have an impact on the polls? We shall see. As we noted last week, the candidate who is the focus of more, usually unfavorable, news coverage this year tends to lose ground. It may be Trump’s turn.

In addition, the very fact that Trump is now seen as tied with Clinton, with some in the media saying he has a fair to good chance to win, may wake up somnolent, unstimulated Democrats and give Clinton a poll and turnout boost by means of the FOT factor.

The debates are just up the road, beginning a week from today. Few voters switch sides, and most tune in to cheer for their side. Still, this election has become mainly about the contrasting engagement of the two party bases. The performance turned in by each candidate will either energize or depress the base, and thus affect the polls somewhat. But that subject is one we’ll address in the next issue.

One more thing before we get to the new map: It’s reasonable to ask why we are making substantial changes to our ratings in a special Monday issue just four days after an update where we made only minor alterations.

In the midst of writing that earlier update, we were stunned by the flood of bad numbers for Clinton, which compounded erosion that had been taking place over the course of the previous few weeks. Some of this was natural — Clinton’s post-convention lead was probably artificially high — but late last week her numbers started to curdle, both before and after our regular Crystal Ball release on Thursday. Credible pollsters like CNN/ORC, Bloomberg/Selzer, CBS News/New York Times and others produced polls showing Trump up three to five points in the typically critical state of Ohio, up in Florida, up in Iowa, and virtually tied nationally.

It is possible, perhaps even likely, that we’re just in a low point of Clinton support: She was absent from the trail and suffered from a series of bad headlines, coming to a climax last weekend with her pneumonia-induced fall captured on camera. Democrats across the country might just not have been participating as much in polls, much like what happened to Barack Obama in 2012 after the first debate. Many Democrats believe — hope? — this is the case. But let’s not over-apply the lessons of 2012: That was a more stable race between an incumbent and a traditional challenger — and, crucially, the Democrats were seeking just a second straight term in the White House, not a third, when public fatigue with the ruling party can grow. There was no measurable third-party support in 2012, and the candidates did not have the bad favorability ratings registered by both major nominees this time. Yes, much of the fluctuation in polls can be artificial, but are we so sure that nothing has changed? Some are, but consider us sufficiently moved by the new numbers. Clinton of course can rebound, but it is at least possible that the recent changes are two campaign ships passing in the night, with Clinton dropping and Trump moving into an actual lead.

Many have been considering the Brexit vote in the United Kingdom a few months ago, when in a shock and disappointment to elites across the pond, British voters narrowly opted to leave the European Union. Brexit’s surprising victory is sometimes portrayed as a massive failure of polls, but in aggregate the polls presented a mixed picture. Though “Remain” was leading in the HuffPost Pollster average, it was only by a hair: the final average pegged the race as 45.8% Remain to 45.3% Leave. So it was reasonable to look at that race as a Toss-up, but many assumed the undecideds would move to Remain. That assumption was wrong.

Brexit isn’t necessarily a great parallel for our election, both because it was a ballot issue while ours is a choice among candidates, and because the British Brexit electorate (6% nonwhite) and the U.S. presidential electorate (approaching 30% nonwhite) are dramatically different. Clinton, even now, retains a lead in national polling averages, though it’s dwindling. She’s up a little over three points in HuffPost, and only about a point in RealClearPolitics, which is more sensitive to the most recent surveys. That’s a little bit better for Clinton than the final Brexit polls for Remain but not dramatically so. The danger for analysts, and voters, would be to assume that a basically tied race would break for Clinton, because so few analysts can imagine Trump as president. We cannot make that mistake come November.

Now, to our state-by-state ratings. Our practice here at the Crystal Ball is governed by one word: patience. Many of you have written us for months, wanting to know why we haven’t changed our March 31 map more significantly. The short answer is: We didn’t need to do so. We don’t flip states every time a poll suggests one candidate or the other has moved ahead somewhere. We wait until we see a credible pattern emerge. Our early bet was that Clinton might run the table in the swing states. This has been replaced, at least for the moment, by a recognition that Trump has surged (or Clinton has fallen, or Republicans are more excited than Democrats) in several states.

We are also re-introducing the Toss-up in a few places. Although we tried to avoid the rating this cycle, we assumed circumstances would eventually force our hand and push us to look at a few states as true coin flips. As always, though, the Crystal Ball will call every state and every race for Congress before Election Day.

Simultaneously, and related, the Democratic position in the fierce battle to control the U.S. Senate has deteriorated. Naturally, in an age with relatively little ticket-splitting, when Trump moves up, so do GOP chances to hold the upper chamber. Democrats had counted on a big Clinton victory to lift them to at least 50 Senate seats, if not an outright majority.

That isn’t currently happening, as we explain below. Obviously, if Clinton regains her mojo and revs up the Democratic base to a greater degree, Democrats might still be able to capture the Senate — which will be critical to Clinton’s ability to get Supreme Court and other nominees confirmed, for instance, should she be elected.

Electoral College: Trump’s standing improves across the map

Table 1: Crystal Ball Electoral College ratings changes

Map 1: Crystal Ball Electoral College ratings

Our Electoral College changes this week are considerable.

First of all, Florida, Nevada, and North Carolina, three diverse, growing states, move from Leans Democratic to Toss-up. Polls have been close in all three for months. There are reasons to think Clinton will still win each: She should out-organize and out-advertise Trump in the trio, which may especially be vital in the Sunshine (Mega)State, as Politico’s Marc Caputo recently noted. But Florida barely voted for Obama in 2012, and Clinton has had persistent weakness in Nevada. You could argue that North Carolina may be the likeliest of the three to vote for Clinton, strange-sounding since the state didn’t even back Obama in 2012 (but did in 2008). But beyond Trump, Republicans have local problems there — Gov. Pat McCrory (R) is unpopular and is a slight underdog for reelection. Be that as it may, all three states look like coin flips to us, as does the single electoral vote in Nebraska’s Second Congressional District.

Trump’s strength with white voters who do not have a college degree is especially helpful to him in Iowa and Ohio, two places where Trump’s polling has been strong lately. Dueling Bloomberg and CNN Ohio polls on Wednesday, both with Trump up five, were reinforced with a Suffolk survey on Thursday finding Trump up three. Ohio, a traditional bellwether state — one of us wrote the book on it recently — typically does vote slightly more Republican than the nation, and it seems like a safe bet that Trump will do better in Ohio than he does nationally. In a very close national race, it’s not hard to imagine Ohio voting for Trump while Clinton wins the Electoral College. If so, it would be just the third time in 31 elections that Ohio would have voted for the loser. What’s essentially impossible to imagine is Ohio voting for Clinton while Trump wins nationally, given history (no Republican has ever won without it) and the state’s demographics (a little more favorable to Trump than much of the rest of the Midwest). Iowa is a state where Democrats have maintained a tenuous advantage based on their over-performance with white voters (compared to the national average). But that edge may be eroding, and Trump has led five of the last eight polls there (and was tied in a sixth). Both states move from Leans Democratic to Leans Republican for now. Iowa doesn’t appear likely to be decisive this year — though Trump may do better there than in any state that Obama won in 2012 — and Ohio probably won’t provide the crucial 270th electoral vote this year either, but a Clinton recovery in the Buckeye State would foreclose Trump’s chances. Maine’s Second Congressional District also moves toward Trump for similar demographic reasons.

Clinton retains an edge in Colorado, Michigan, New Hampshire, Pennsylvania, Virginia, and Wisconsin, although as the race got tighter nationally the numbers seemed to close in at least some of these states, too. An NBC News/Marist poll last weekend had Clinton only up two in New Hampshire. The University of Mary Washington had Clinton up just three in Virginia (though the Democratic pollster Public Policy Polling had Clinton up six), and the Detroit Free Press had Clinton up three in Michigan. Clinton got a better poll on Saturday night — Muhlenberg College/The Morning Call found her up eight in Pennsylvania. Still, there’s no Clinton “firewall” if Trump moves into a national tie or even starts taking the lead — the states all move to at least some degree when the national polls swing significantly in one direction. These half-dozen states, which represent the inner walls of Clinton’s Fortress Obama, move from Likely Democratic to Leans Democratic.

Finally, Missouri — a one-time bellwether itself that for a century almost always voted for the winner — has become reliably Republican over the last decade. There’s just no path for Clinton there given the current state of the race, so it moves from Likely Republican to Safe Republican.

Despite all of these ratings changes toward Trump, an Electoral College majority still at least leans toward Clinton — but just 272 electoral votes, a bare majority.

Clinton’s slide boosts Republican Senate chances

Table 2: Crystal Ball Senate ratings changes

Map 2: Crystal Ball Senate ratings

Right before Labor Day — when Clinton still appeared to be on the path to a comfortable, Obama-style national win — we shifted some Senate races toward the Democrats, thinking that Trump would end up sinking Sens. Pat Toomey (R-PA) and Kelly Ayotte (R-NH). But with national polls very tight, Ayotte and Toomey find themselves in a better position because they’ve both demonstrated the ability to run ahead of Trump in the polls. If Trump wins either New Hampshire or Pennsylvania, it probably means he will be president — and that would almost certainly mean that Ayotte and/or Toomey would be winning as well. If Trump falls off — certainly possible — then Ayotte and Toomey could go back to being underdogs. But in a close national race that is also close in New Hampshire and Pennsylvania, Ayotte and Toomey could definitely still hang on. So we’re moving both from Leans Democratic back to Toss-up.

Ever since the general election began, the political world has been debating whether Republican Senate candidates should repudiate Trump. At the very least, most have kept their distance. Yet Trump’s improvement in the polls could allow them to survive, even if Trump doesn’t win their states.

Indiana is a different story. There, it’s ex-Sen. Evan Bayh (D) who needs to run ahead of Hillary Clinton, because she almost certainly will lose the state by a wide margin. A recent WTHR/Howey Politics poll showed Bayh up just four points, 44%-40%, a troublesome sign for Bayh: Because the former senator already has strong name ID, his opponent, Rep. Todd Young (R, IN-9), may ultimately have the clearer path to a plurality because he’s the Republican running in a Republican state, although Bayh does mainly have a positive statewide profile. Ultimately, Trump’s margin matters here, too. Mitt Romney won the state by 10 points in 2012. If Trump does as well or better than that, Bayh’s path becomes a lot rockier. So too would the path of Democrat John Gregg, who is seeking the open governorship against Republican Eric Holcomb. Gregg, like Bayh, has a lead and more name ID in the Howey poll but probably has less room to grow. Trump’s vice presidential nominee, incumbent Gov. Mike Pence, has united the GOP factions behind the ticket in a normally Republican state, and it benefits the whole Hoosier GOP ticket.

One other change is shifting Ohio’s Senate race from Leans Republican to Likely Republican. Sen. Rob Portman (R) has been pulling away for the last couple of months, and Democratic outside groups have cut bait on former Gov. Ted Strickland (D). Portman is running a strong campaign and he and fellow Republicans continue to hammer Strickland over his time in office, which coincided with the national 2008 economic collapse. This was entirely predictable, and it has worked quite well. The only dangers to Portman were his lack of statewide name ID – which he has effectively corrected — and a drag from the top of the ticket. Now that Trump could win or come close to winning Ohio, the threat he posed to Portman appears to be eliminated.

What’s worse for Ohio and national Democrats is that while there are some young potential statewide prospects waiting in the weeds for an opportunity — including Cincinnati City Councilman P.G. Sittenfeld (D), who Strickland handily defeated in the Senate primary — they used this prominent statewide slot on Strickland, who is almost certainly running his last race. Compare that to Missouri, where Secretary of State Jason Kander (D), a fresh face, is pushing Sen. Roy Blunt (R). Kander, who released one of the best ads of the whole cycle last week, is an underdog, but he has received strong, bipartisan reviews for his performance so far, and even if he loses, he is getting exposure that might help him in a future run. We have Missouri’s Senate race (Leans Republican) now listed as more competitive than Ohio’s (Likely Republican).

Because of the increasing party polarization of the electorate and the erosion of ticket-splitting in recent Senate elections, we’ve thought the winning presidential party would have a leg up in the race for the Senate. That would almost assuredly be the case if Trump won the presidency — it would be very surprising if Democrats won the Senate in such a scenario (and of course Republicans would hold the House, too — a likely outcome in any event). But the now-tight presidential race suggests that perhaps Clinton could pull off a narrow victory that still allows the GOP to hold the Senate. If so, and given the very favorable Senate map coming up for Republicans in 2018, it’s possible that a President Hillary Clinton would face a hostile House and Senate for as long as she was in office (whether one term or two).

Of course, that assumes there is a President Hillary Clinton — an election result that seems to us to be more in question now than at any point since Donald Trump wrapped up the Republican presidential nomination.