Reëlection Watch: October 13, 2012
This week has been another time of major change. We have much more post-debate data.
What’s the latest news? Let’s dive in and see.
In the national popular vote matchup of President Barack Obama versus Republican nominee Mitt Romney, the Republican has taken the lead for the first time ever. After the Presidential debate, Obama’s numbers have fallen and Romney’s have risen. The President’s position is now 2.5 points behind President George W. Bush’s on this date eight years ago, and 8.5 points below his position four years ago.
Going into the debate, Obama’s favorability polls all exceeded 50 percent. Since then, two of four dipped below that magic number. Romney’s favorability polls have continued to improve, and he finally broke the 50 percent ceiling on a single FOX News poll (one for which Obama also was above 50). Obama has consistently held a three to five point edge over Romney all year, and was at six points going into the debate. After the debate, his lead dropped back to three points.
The national polls, for the first time, suggest that Obama might lose the upcoming election.
As of yesterday, Intrade had Obama at 62, down seven points from last week, and 17 from two weeks ago. The markets still believe that Obama will be reëlected, but they show far less confidence than they did a couple of weeks ago.
Overall, thinks on the national level look sketchy for the President.
Last time I mentioned states with early in-person and absentee voting open. This time, I’m separating them out so you can see which are which. The green states are accepting absentee ballots, but not early in-person ballots, while the gold states are accepting both.
All battleground states other than Colorado and Nevada are at least taking absentee ballots; those last two open this coming week. Ohio and Iowa are the only two battleground states with open early in-person voting. North Carolina’s in-person voting opens next week.
The Electoral College
Here’s what the Electoral College looks like, based on current polling data:
With the recent exposure of Gravis Marketing, I’m dropping them from the model. It turns out that dropping them didn’t impact the results, as I was already compensating for their five-point Republican lean relative to the consensus. The timing of polls this week is especially important in determining the contour of the Romney debate bump. I outline the details in each state below.
Here are the 15 states with new data since last time, covering only those discussed around the Internet as “leans” or “tossups”, from reddest to bluest:
- Arizona hasn’t been polled recently, but I’m now starting to include Intrade markets in these descriptions (though not as input to my model). And Intraders are confident that the Grand Canyon state will go for Romney; his security shows a 90 percent chance of him winning. Arizona remains “Likely Romney”, though not by enough to drop off the list entirely.
- North Carolina was polled twice after the Presidential debate. Rasmussen saw Romney ahead, by three points. Accounting for Rasmussen’s, that puts North Carolina in Romney territory by about two points, which is consistent with the overall national post-debate shift. Of those requesting absentee ballots, 54 percent are registered Republicans. North Carolina remains a “Leans Romney” state. Intraders give Romney a 73 percent chance of carrying the Tarheel State.
- Florida got polled five times after the Presidential debate. I mentioned the Rasmussen and WeAskAmerica polls last week. In addition, the University of North Florida conducted a poll that straddled the debate, and found Obama up by four points. It’s a ten-day poll from a group without any history this cycle, so I’m not lending it any credence here. More recently, NBC/Wall Street Journal/Marist saw Obama up by one, while Tampa Bay Times/Miami Herald/Mason-Dixon, American Research Group, and Rasmussen found Romney up by seven, three, and four, respectively. All in all, it appears that Marist is the outlier. Florida still “Leans Romney”. Floridians have been able to vote for nearly two weeks. Intraders give Romney a 61 percent chance of picking up Florida’s 29 electoral votes.
- Virginia was moved to “Tossup” last week, because the two post-debate polls published then by Rasmussen and WeAskAmerica saw Romney leads. The four subsequent polls tell a conflicting story. Public Policy Polling and CBS/New York Times/Quinnipiac’s polls conducted late last week both found Obama ahead, by three and five points, respectively. NBC/Wall Street Journal/Marist’s Sunday-to-Tuesday poll came away with Romney ahead by one. And Rasmussen’s Thursday poll had Romney up by two. The conflicts leave Virginia a “Tossup”, which strongly suggests that Suffolk University is mistaken in calling the Old Dominion a certain Romney state. Virginians have been able to vote by absentee ballot for three weeks now, and Obama was ahead for about half of that time. The Intrade market has Romney slightly ahead, at 53 percent.
- Colorado was polled before the Presidential debate by Republican-funded McLaughlin/American Conservative Union, who saw a four point Romney lead. Given the provenance, timing, and lack of historical data, I don’t count this in the mix. After the debate, six firms conducted polls. American Research Group, The Denver Post/SurveyUSA, and CBS/New York Times/Quinnipiac found Romney ahead by four, one, and one point, respectively. On the other hand, the University of Denver and Rasmussen came away with Obama leads by four and one point, respectively. Other than Rasmussen, all polls were conducted over several days after the debate (SurveyUSA on the 9th and 10th). This makes it exceptionally difficult to shape the bump contour. If we lend a lot of credence to Rasmussen, the bump fully dissipated by last Sunday, but SurveyUSA suggests otherwise. Last time I moved Colorado to “Leans Obama”. This week, I’m moving Colorado back to “Tossup”. Intraders disagree with me (and Rasmussen), assigning a 54 percent chance to Romney.
- Nevada got polled after the Presidential debate by the Las Vegas Review Journal/SurveyUSA, Rasmussen, Public Policy Polling, and Suffolk/KSNV. All except SurveyUSA were conducted entirely after the debate (SurveyUSA straddled it), and all except Rasmussen saw small Obama leads. Rasmussen found the race tied. Obama’s house-bias adjusted lead from these four polls is one point, small enough to move the Silver State to “Tossup”, particularly since most of the polls were late enough for the debate bounce to dissipate. Intraders still have faith in the President here; he is trading at 70 percent.
- New Hampshire was polled Tuesday by Rasmussen, who saw the race as tied. American Research Group polled Tuesday through Thursday and found Romney up by four. While both have Republican biases, accounting for the biases still moves the Granite State into “Tossup”, and much higher on this list. On Intrade, Obama has the lead at 58 percent.
- Ohio was polled eight times after the Presidential debate. I mentioned Rasmussen and WeAskAmerica last week. The six new post-debate polls were conducted by American Research Group, CNN/Opinion Research, SurveyUSA, NBC/Wall Street Journal/Marist, and Rasmussen (again). All found Obama ahead except for ARG, who has been to the right of the consensus. ARG saw Romney ahead by one point. On balance, Ohio barely remains “Leans Obama”. Buckeyes have been able to vote for two weeks now, and absentee ballots appear to be coming in from the urban zones at a higher rate than in 2008, though thus far they still represent only three percent of the 2008 vote total. The Intrade market gives the edge to Obama, at 59 percent.
- Iowa was polled by Rasmussen, who saw a two point Obama lead last Sunday. Iowa remains “Leans Obama”, which is important since Iowans can currently vote by absentee ballot, and in person…and have been doing so; over 200,000 votes have already been cast, about 15 percent of the total 2008 vote total in Iowa. On Intrade, Obama has the lead here, with a 60 percent chance.
- Wisconsin got polled three times after the Presidential debate by Public Policy Polling, CBS/New York Times/Quinnipiac, and Rasmussen. They saw Obama ahead by two, three, and two points, respectively. It’s clear that Obama’s lead here diminished appreciably from its pre-debate highs, but by how much really? Both PPP and Quinnipiac are smeared across the post-debate week, while Rasmussen’s was conducted this past Tuesday. This suggests that Obama remains ahead by about four points among likely Badger voters. This late, a steady four point lead would be “Likely Obama”, but a four point lead that is closing is “Leans Obama” instead. For now, Wisconsin moves from “Likely Obama” to “Leans Obama”. Intraders remain confident in Obama here, trading him at 69 percent.
- Michigan was polled five times since the Presidential debate. Baydoun/Foster, EPIC-MRA, the Detroit News, and Rasmussen all saw Obama ahead, by three, three, seven, and seven points, respectively. Despite being funded by the Democrats, Baydoun/Foster has consistently been far to the right of the consensus, by an average of five points. On the other hand, EPIC-MRA has been pretty close to the consensus. Rasmussen’s poll was the most recent, on Wednesday. All of this implies that Obama’s lead in Michigan is probably closer to the Detroit News number. Michigan stays “Likely Obama”, but no longer by enough to be off the list entirely. At Intrade, Obama is trading at 81 percent here.
- Pennsylvania is back on the list this week. Polls were conducted after the Presidential debate by Siena, Susquehanna, the Philadelphia Inquirer, and Rasmussen. All saw Obama ahead, by three, two, eight, and five points, respectively. Siena was the only pollster to straddle the debate, and Rasmussen was the only one to exclude the early post-debate bump days. Susquehanna was unchanged from its results of September 18–20, when it was about seven points to the right of the consensus. Overall, then, it appears that Obama lost about two points in the Keystone State, which isn’t enough to yet move it from “Likely Obama”, though it’s clearly not as safe as it was in late September. Unsurprisingly, the markets at Intrade agree with this one, putting the President at 83 percent.
Movement this week has been pretty consistent, and it points in Romney’s favor. Colorado, Nevada, and Wisconsin shifted one column to the right. Based on the model, Obama has a probable 275 electoral votes, down 28 from last week, but still enough to win. It’s been more than three consecutive months in which Obama could lose all tossups and still stay in the White House.
Are we seeing a momentum shift, or a new plateau? The numbers I’m seeing look like the latter, since more recent numbers show a slight reversion in the pre-debate direction (though most of the boost has been sticky). For now, then, it appears that Romney’s high-water mark was October 4th and 5th. On Tuesday, both candidates have an opportunity to create a similar shift. Historically, however, the first Presidential debate has had a greater impact than subsequent debates.
The national polls show Romney ahead by close to two points, while the state tallys indicate a comfortable Obama lead. They can’t both be right, since there can realistically be a maximum of a point or so by which one can lose the national popular vote and still win the electoral vote. Historically, the state-by-state tallies have been more accurate than the national ones, which leads me to believe that Obama still has the edge.
If I had to predict an Electoral College result based on the model, I’d keep things close to where they have been for three months. The Tossups would go to Obama, only because they are currently a hair to his side of the middle in the polls. That prediction would give Obama 303, and Romney 235. That’s the same as it has been for a couple of months.
How do you feel about these predictions? Do you differ on them? If so, how, where, and why?
- Obama Walloped On Intrade Early In Debate
- Reëlection Watch: October 6, 2012
- Poll: Romney Leads Obama By 7 Points In FL
- NATE SILVER: Obama Is Collapsing
- Oct. 10: Is Romney Leading Right Now?
- Rasmussen Poll: Romney ahead 49–47 percent in first, full post-debate survey