Is There A “Dire Risk” of Trump Losing Georgia?
A front page story in today’s New York Times dives into the difficulties Republicans are facing following Donald Trump’s declaration that “the shackles have been taken off me,” and his willingness to go against the Republican Party. The story lists lists states and GOP candidates who are at risk of losing, but it also includes this:
Democrats are moving swiftly to exploit Mr. Trump’s crumbling position in the presidential race, aiming to run up a big margin of victory for Mrs. Clinton and extend their political advantage into the congressional elections next month.
Mrs. Clinton’s campaign has concluded that at least two traditionally Republican states, Georgia and Arizona, are realistic targets for her campaign to win over. And Republican polling has found that Mr. Trump is at dire risk of losing Georgia, according to people briefed on the polls, who spoke on the condition of anonymity.
Current public polling still seems to keep the Peach State in Trump’s camp. The Real Clear Politics average has Trump up by 4.7 points. FiveThirtyEight gives Trump an 80.2% chance of winningn in its polls-plus model, although all three of FiveThirtyEight’s models show some deterioration in Trump’s chances. And the Times’s Upshot model gives Trump a 78% chance of winning here.
Of course, the conventional wisdom says that private polling done by the parties is more accurate than public polls. You can look back to 2014, when the polling done by the Georgia GOP showed wins by Nathan Deal and David Perdue, according to GOP officials, despite some public polling that showed Jason Carter or Michelle Nunn ahead. And those party polls are by definition not public, although if someone wants to share, we’re right here, you know.
:: Update ::
With respect to public polling vs. private polling, Mark Rountree of Landmark Communications called to remind me that my memory of the late public polling in the 2014 elections compared to the internal GOP polls was unclear. While it’s true that many polls showed Deal or Perdue behind or tied in their races, Landmark’s polling was in general more accurate than national polling firms, and, in the end, got the races right. And our own Will Kremer had the play by play at the old place, as GAGOP Chair John Padgett castigated the polling in an interview with Tim Bryant, and later Rountree had a chance to respond.
We’re glad to clarify what happened, as usual we regret the error and blame others, and are eagerly awaiting updated presidential polling from Landmark.
About Author
Jon Richards
Add a Comment
You must be logged in to post a comment.
This is supposedly ‘Republican’ polling, yet Clinton pulled her advertising from Georgia a couple of weeks ago. My BS detector pinged when I read “…according to people briefed on the polls, who spoke on the condition of anonymity.”
yeah, i got the sense that this article was at least a few weeks into the past.
No.
I will be the contrarian here. I’m predicting a Trump loss here in Georgia. If not a Trump loss, then he wins with less than 50% and Hillary is within a point from winning. I’ve seen some internal polling that has him losing the Northern suburbs of Atlanta by 10 points. If white educated suburban women stay at home in mass numbers then Hillary can win Georgia. Polling shows that they are the single group that is most likely to stay home becasue they have an extremely high unfavorable rating of both candidates. You heard it here first folks. Clinton wins Georgia.
And I say ‘so what’. Georgia is unlikely to swing the electoral count one way or the other.
And I doubt he is hurting Isakson enough for Johnny to lose either.
The true “dire risk” for the Republican party would be for Trump to win nationally. He’s toxic and should have been recognized as such from jump. Follow Ryan’s lead and cut your losses by paying more attention to the down ballot races.
Yes, I think the “dire” part is overrated—like Benevolus says, Georgia is unlikely to swing the Electoral count. If Trump were to lose Georgia, then certainly he would be losing Florida, North Carolina, Ohio and Pennsylvania—and thus the election. If Clinton were to win, it would be icing on the cake for her, like running up the score in a football game. But if Clinton were to win Georgia, or even come within a few points (like under 5) that would certainly install confidence in the state’s Democratic party looking toward the 2018 election, which unofficially gets underway four weeks from today (the day after the presidential election).
Whether she can win the state depends in large part on the 29-county metro Atlanta area. Obama narrowly (by barely half a point) won metro Atlanta in 2012, but lost by nearly a 3-2 margin in the rest of the state. Obviously Clinton would need to win the Atlanta area by a lot more than Obama to have a chance, and probably prevail overall in the state’s other metro areas. Back in 2002, the Perdue-Barnes race for governor was about even in what was then metro Georgia (the boundaries of metro areas as defined then), but Perdue trounced Barnes by nearly 3-2 in the non-metro (rural counties), which ended the Barnes era at the Gold Dome. You figure Trump will win a lot of rural, exurban Georgia, places like the Gainesville-centered 9th District (Doug Collins) and the Rome-Dalton based 14th District (Tom Graves). but GOP margins may be more in question in say the north Fulton-East Cobb 6th District (Tom Price) and Forsyth-Gwinnett 7th District (Rob Woodall). Price’s district backed Rubio over Trump in the March 1 presidential primary while the Gwinnett portion of Woodall’s district is trending Democratic. As for Augusta (Columbia and Richmond Counties), well we know Trump will win Columbia and Clinton Richmond, but will there be some GOP erosion even in Columbia County, which hasn’t backed a Democrat for president in 40 years?
Doubtless both major parties are polling presidential numbers in several State House districts in metro Atlanta, but neither has released public results. For the Democrats, that may be to lull Republicans into a sense of complacency, create the perception that none of the State House districts that backed Romney last time are in play; while for the Republicans, the polling results may be so bad or mediocre in some districts, they worry about a depressed turnout November 8 if early polls show Trump behind Clinton in normally GOP areas—and thus put at risk some legislative seats. Of course the question November 8 is not continued GOP majorities at the Gold Dome, but rather (especially in the House) the size of the margins.
Of course, if you look at all three of those models, you see the last data they’re relying on are polls from September 22 or earlier. (The one exception being the Reuters/UPI ‘polls’ that are pulling out national subsamples on a state by state basis, and are pretty much useless). What was happening then? Pneumonia and Deplorables, with Trump and Clinton being pretty close nationally. In a world where they’re close nationally, you’d expect a small to solid GOP lead in Georgia.
But since then: Debate I, the Sexual Assault Tape, Debate II, and him going nuclear on Ryan, the results of which have seen Trump swing back down to close to double digits behind nationally. If he’s losing badly nationwide, that’s enough to swing Georgia real close. Public polls haven’t been in the field to see any change in Georgia from all of that, and so haven’t reflected that. But you can bet the private pollsters, beyond being generally higher quality than the public ones, haven’t also gone dark for three weeks. If they’re seeing something, more likely than not it’s there, and out of date public polls aren’t much of a refutation.
Georgia’s going to be won or lost in the Atlanta suburbs. If he gets wiped out in the Atlanta metro, which will make up more than 60% of the state’s raw vote, that could be enough to overcome the 15-20 point GOP lean of the rest of the state. He loses the 32 counties of the Atlanta metro by 10 points or more, he loses Georgia. If you were designing something to make him absolutely toxic to the college educated women from Marietta to Sandy Springs to Lawrenceville, that’d be the last week or so.
30 counties. Stupid typo. Apparently Hall County was recently added to the Atlanta MSA. (hence 30 instead of 29) With his 33,400 vote margin there, Romney narrowly won the Atlanta Metro by 28,800 votes out of 2.2 million. Obama won it by 48,800 in 2008. She’s going to need to win it by at least 200,000 to have the possibility to carry the state, probably closer to 250,000, which was McCain’s margin outside of the ATL proper/suburbs/exurbs. Obama won Fulton/Cobb/Dekalb/Gwinnett 60-39 and 58-41. If she can get that margin up to 65-35 or better, that’d be enough, everything else being equal, to make Georgia real, real close.
I still think Georgia is not in play though some fresh polling would be great given the changes to the race since the last round. Great comments above. 538 ran an article after clintons big bounce after the conventions that showed what national polling lead puts what states in play. I wish I could find a link but I believe they put Georgia in play with Clinton having an 8-10 point national lead. I’m also still fascinated by how trumps lack of ground game will affect him. When the republican primaries were still being contested he would consistently underperform his polling. Add on to that, Obama significantly overperformed his polling in 2012. Obama was only polling up about 1% on the eve of the election and actually won by 4%. A lot of that was the democratic turn our machine much of which Clinton has inherited. If you get in a scenario with trump underperforming per his pattern and democrats drumming up a big turnout things could get very lopsided very quickly.
One thing about the R vs D lean within the state. I’d expect it to be much less than R + 12 this year because of the way Trump, while not doing substantially worse than Romney among white voters overall, makes that happen by doing better with white voters without a college degree and worse than white voters without one. You can see this in how Virginia, Colorado, and North Carolina played out this cycle, as opposed to the last two:
2008:
USA: D+7
Colorado: D+8.95 (So D+1.95 lean)
Virginia: D+6.3 (So R+0.7, was the state closest to national margin)
North Carolina: D +0.3 (So R+6.7)
2012:
USA: D+3.9
Colorado: D+5.37 (D+1.5)
Virginia: D+3.87 (R+0.03, again the state closest to the national margin)
North Carolina: R+2.04 (R+5.94)
In 2016 in all those states Clinton is running ahead of her national margin (or, in North Carolina, at it, which makes a big shift in a cycle) to the point that the Republicans didn’t bother with Colorado and Virginia after the conventions, even as national polls shifted. (In the same vein Iowa and Ohio, which have a lot of non-college educated whites, are running better for him vs. his national polls than they did for Romney or McCain). Those are states, like Georgia, where college educated whites have made up a big chunk of the Republican coalition, and he’s been toxic for them in the Denver, Washington, Charlotte, and Raleigh/Durham suburbs. 538 had an interesting article today on these two shifts (http://projects.fivethirtyeight.com/clinton-trump-vote-maps-2016/) showing that even in a world where Trump loses by just 0.9 more points than Romney nationwide, that internal churn could make Georgia one of the narrower Republican wins. Why? A quarter to a third of the Romney coalition in the Atlanta metro are the same educated white voters he’s losing in VA/CO/NC. And the white working class vote in GA mostly left the Dems already, so there’s not much for them to lose. Throw them together, and you see why 538’s county by county map is all upside for the Dems.
David C. thanks for the update on 30 counties now in metro Atlanta—wow, quite a change from the old 1960s definition of Clayton, Cobb, DeKalb, Fulton and Gwinnett. Think it was 28 counties for the 2012 cycle, then 29 with Morgan in 2013 and now Hall. The metro area equation starts to get complicated when you start getting into adjoining metro areas, like Atlanta and Athens and then consolidated metro areas.
On Virginia, Trump’s chances were never good…even in the bad 2014 cycle for the Democrats, Mark Warner was re-elected to the Senate (though just barely—he apparently took his race for granted after his easy 2008 win). Virginia Republicans dread presidential years because heavily Democratic Northern Virginia likes to show up on those occasions, but not so much in the off-years such as for governor (understandably so, NOVA focused more on federal matters than happenings in Richmond). So the story is that if Republicans could not win the state in an off-year, they certainly would not do so in a presidential year, of course especially with Kaine on the ticket.
Back to Georgia—you probably would have to see Clinton win Cobb and Gwinnett to have a chance in this state—and neither county has voted Democratic for president since 1976, for the peanut farmer from Sumter County.. But Gwinnett’s demographics are such that at the very least, if Clinton loses Gwinnett, it won’t be by much. In 2012, Romney got 55% in Cobb and 54% in Gwinnett; a generation earlier (Reagan landslide of 1984), Cobb voted 77% for Reagan, Gwinnett 79% (of course both counties, but especially Gwinnett, were far less populated than they area today—easier to run up high GOP percentages in small or relatively small counties).
What will be especially interesting to see is how the “country-club” areas of metro Atlanta vote—places like East Cobb, Buckhead, Dunwoody and Vinings.
Thanks Jon. We tracked the 2014 election regularly with polls, like we’ve done in every election since 1995. It was released here, as seen on the polling aggregator site RealClearPolitics.com:
Governor ’14 – Our poll numbers: Deal 51% to Carter 45%, election result was 53% to 45%.
http://www.realclearpolitics.com/epolls/2014/governor/ga/georgia_governor_deal_vs_carter_vs_hunt-5254.html
Senate ’14 – Our poll numbers: Perdue 50% to Nunn 46%, election result was 53% to 46%.
http://www.realclearpolitics.com/epolls/2014/senate/ga/georgia_senate_perdue_vs_nunn_vs_swafford-5253.html
We released our latest Landmark Communications (Landmark Opinion) poll of Georgia likely voters today. Trump leads Clinton 48-42%, with 4% for Johnson.
Sen. Isakson leads 50%-37% against Jim Barksdale, with 5% for Libertarian Allen Buckley.
1400 interviews, 2.7 margin of error, conducted October 11-12th.