Reëlection Watch: October 20, 2012
This week has been yet another time of major change as we close in on Election Day. For the first time in a long time, my Electoral College prediction has shifted.
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 maintained his RealClearPolitics lead from a week ago, though it has shrunk a hair. Gallup’s tracking poll is the only thing holding Romney above Obama, and Gallup is sufficiently far from the consensus as to be considered an outlier. Historically, when Gallup is far from the consensus, Gallup is the errant pollster. Investor’s Business Daily and Rasmussen both indicate a tie at the national level. The President’s position is now two points behind President George W. Bush’s on this date eight years ago, and eight points below his position four years ago.
After the first debate, half of Obama’s favorability polls had him below 50 percent. In the past week (and all before the second debate), they rose again above that magic threshold. Romney’s favorability polls have continued to improve, and he now matches Obama. Obama held a three to five point edge over Romney all year until this week. The two candidates are now tied in this metric.
Obama’s job approval polls show him about 1.5 points worse than Bush was in 2004.
Like last week, the national polls suggest that Obama might lose the upcoming election.
As of yesterday, Intrade had Obama at 62, unchanged from last week.
Overall, things on the national level continue look sketchy for the President.
Early voting has spread to 49 states this week. The green states are accepting absentee ballots, but not early in-person ballots, while the gold states are accepting both.
All states other than Pennsylvania are at least taking absentee ballots. Sixteen states allow in-person voting today; of the battleground states, this includes Ohio, Iowa, Nevada, and North Carolina.
The Electoral College
Here’s what the Electoral College looks like, based on current polling data*:
Here are the key states with new data since last time:
- North Carolina was polled once in the past week. Public Policy Polling saw Romney ahead by two points, but their left bias translates to something more akin to four points. Of those requesting absentee ballots, 54 percent are registered Republicans. North Carolina moves to “Likely Romney” as the clock winds down. The Romney campaign feels the same way; they pulled their advertising dollars out of the state earlier this week, in favor of spending more money elsewhere. Intraders give Romney an 80 percent chance of carrying the Tarheel State, up seven points from last week.
- Florida got polled four times in the past week. Public Policy Polling, Rasmussen, CNN/Opinion Research, and FOX News saw Romney up by one, five, one, and three points, respectively. The overall weighted and corrected average translates to a three point lead for Romney, but they’re trending in Romney’s direction. Absentee ballots have been coming in; about four percent of the 2008 total have already been cast, with Republicans outvoting Democrats by a four point margin. Florida moves to “Likely Romney”, based on the consistency of pro-Romney polls in the Sunshine State, the trend in Romney’s direction, and the systemic advantage Romney already enjoys. Floridians have been able to vote for nearly three weeks. Intraders give Romney a 65 percent chance of picking up Florida’s 29 electoral votes, up four from last week.
- Virginia was polled this past week by American Research Group and Rasmussen, who saw a one and three point Romney lead, respectively. Rasmussen’s adjusts down to two. It’s not that Obama has been trending down so much as the undecideds in the Old Dominion are breaking heavily for Romney, according to recent polls. All in all, it’s close enough to keep Virginia a “Tossup”, in part because Virginians have been able to vote by absentee ballot for a month now, and Obama was ahead for much of that time. The Intrade market has Romney barely ahead, at 51 percent, down two from a week ago.
- Colorado had two polls published this week. WeAskAmerica found Romney ahead by a point, while Public Policy Polling, the only one to poll after the second debate, saw Obama up by three, which adjusts to something more akin to a tie. There’s still no evidence that Romney’s debate bump has faded here. Colorado stays a “Tossup”. Intraders agree with me; the Colorado market is tied between the two candidates. This represents a four point shift in Obama’s direction this past week.
- New Hampshire was polled this week by Suffolk/7News, who saw the race as tied, Rasmussen, who saw Obama up by one, and Public Policy Polling, who saw Romney up by one. Rasmussen’s poll adjusts to a two point lead, and PPP’s adjusts to three, but on balance the three polls aren’t enough to move the Granite State a column left; it stays “Tossup”. On Intrade, Obama has the lead at 62 percent, up four from last week.
- Ohio was polled twice this past week. SurveyUSA, Rasmussen, and FOX News saw leads for Obama by three, one, and three points, respectively. Including the house biases, that keeps Obama narrowly in the lead. Ohio still sits just on the “Leans Obama” side of the border with “Tossup”. Buckeyes have been able to vote for three weeks now. The Intrade market gives the edge to Obama, at 62 percent, up three from a week ago. But Ohio’s market is one of the few competitive states where the sum of the two candidates don’t add up to 100 percent; this market adds up to 101.7.
- Iowa was polled four times this week. American Research Group saw the race tied, while WeAskAmerica found Obama ahead by three, and NBC/Wall Street Journal/Marist saw an Obama lead of eight points. On the other hand, Public Policy Polling saw a one point Romney lead, which would typically adjust to three. This is a real mess to decipher. Marist and PPP are the only two to cover days after the second Presidential debate, and they differ by nine unadjusted, and 11 adjusted, points. Both diverge from the historical trend and each other, which suggests that they’re both outliers, though Marist by a bit more than PPP. I think this converges on Obama up by two or three points, so Iowa remains “Leans Obama”, a position it has held during the entire time ballots have been able to be cast. On Intrade, Obama has the lead here, with a 63 percent chance, up three from last week.
- Nevada got polled again this week, this time by the Las Vegas Review Journal/SurveyUSA and Rasmussen. This time, both saw small Obama leads. Rasmussen found the race tied. Obama’s lead crept up this week. This is a tough call for the model, which sees the following: growing lead for Obama, little time before Election Day, and yet a relatively small lead. This is what has led to the whipsaw of “Leans Obama” two weeks ago, “Tossup” last week, and now “Leans Obama” again, and moves several states down this list. Intraders still have faith in the President here; he is trading at 69 percent, down a point from a week ago.
- Wisconsin got polled this week by Marquette University, who saw Obama ahead by one point, and NBC/Wall Street Journal/Marist, who came away with Obama up by six. This is a tough call. Every poll after the first debate, except for this latest Marist poll, showed Obama up by one to three points. But, like their Iowa poll, Marist’s is the only one to include people surveyed after the second debate. I’m leaving Wisconsin as “Leans Obama”, but we should look for a trend here and in Iowa. Intraders remain confident in Obama here, trading him at 68 percent, down one from a week ago.
- Pennsylvania was polled three times this week. Public Policy Polling, the Morning Call, and Quinnipiac saw Obama leads of seven, four, and four points, respectively. That puts Obama in the lead by somewhere between four and five points; that this lead has remained consistent as the debate bump faded in the Keystone State means we keep it at “Likely Obama”. Pennsylvania does move one step up this list this week, though. The markets at Intrade agree with this, putting the President at 79 percent, down four from a week ago.
- Michigan was polled once this week, the day after the second debate, by EPIC-MRA, who saw a six point lead for Obama. This is significant, because last week they found Obama up by only three. On the other hand, the consensus of the four pollsters last week was more like five or six. Regardless, Michigan stays “Likely Obama”, but not by enough to be off the list entirely. At Intrade, Obama is trading at 85 percent here, up four from last week.
- Connecticut makes this list this week, not because the outcome is particularly in question, but rather because the outcome has been suggested at various times to be in question. The two polls that came out this week should put this to rest. Siena and Hartford Courant/UConn both show commanding leads for Obama. At no time this year has a poll shown Obama ahead by fewer than seven points, and these latest two show him ahead by twice that much. Connecticut is “Likely Obama”, and barring a huge shift between now and the end of the month, we won’t see Connecticut on the list again. This heavy Obama lean is reflected at Intrade, where Obama is given a 90 percent chance of winning the state.
Three states shifted this week. Only Nevada moved one column to the left, putting six more electoral votes in the “Leans Obama” zone. Florida and North Carolina moved from “Leans Romney” to “Likely Romney”. This is the first time states moved into “Likely Romney” since we started this series. Based on the model, Obama has a probable 281 electoral votes, down 22 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.
The Vice Presidential debate didn’t have a visible impact in the polls, and it’s still too early to discern the broader impacts, if any, from the second Presidential debate.
The national polls show the race effectively tied, while the state tallies continue to indicate a comfortable Obama lead. 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.
What if I’m wrong, and the national numbers are right? If so, then we should shift the states about two points to the right. Romney would pick up Virginia, Colorado, and New Hampshire, giving him 261 electoral votes. Romney would then need ten more electoral votes. In this scenario, Ohio would be an up-the-middle tossup, and would therefore be to the 2012 election what Florida was to the 2000 election. While the systemic forces in Ohio aren’t as strong as they are in Florida, they are probably enough to push Ohio to Romney if the national numbers are right, a further reflection of 2000.
If I had to predict an Electoral College result based on the model, which puts more emphasis on the state polls, I’d shift things a hair from last week. Both tossups would move into Romney’s column. This would give Obama 281, and Romney 257, the first change in over two months.
How do you feel about these predictions? Do you differ on them? If so, how, where, and why?
Author’s note: The map erroneously had New Hampshire as “Leans Obama” instead of “Tossup”. I regret the error.
- Reëlection Watch: October 13, 2012
- About that Gallup poll: Is Romney really up by 7? And will Obama win the election anyway?
- Romney has edge in key national poll, but swing states remain close
- Romney campaign, feeling confident, pulls some staff out of N.C.
- With 284,000 Florida votes already in, absentee vote data more critical than polls
- Intrade Odds Show Romney Gaining After Libya, First Presidential Debate
- How to Make Money on the Presidential Prediction Markets
- Even with Gallup, This Race is Tight
- Absentee-ballot war: 500,000 Floridians have voted; Democrats slightly trail Republicans
- Polls: Gallup still has Romney up, Marist shows Obama leads in two swing states