Some Final Election Wildcards

The recent national poll news is overall bad for Romney, suggesting a reversal in momentum in favor of Obama. The one possible saving grace is that this momentum may primarily be an artifact of increased Obama support in Northeastern states where Romney was not competitive to begin with. The Pew poll, which shows Obama up by three–his best current result–states: “Another notable gain for Obama, perhaps reflecting Hurricane Sandy’s effect on the race, comes in a region he was already secure in: the Northeast. He has increased his lead over Romney from nine points (52%-43%) to 21 points (56%-35%) there over just the past week.” A very rough calculation suggests that this surge of support in the Northeast could account for approximately 1.5% of the 3% point lead. If other polls showing an even tighter race are getting similar results, it leaves open the possibility of a repeat of 2000–an Obama popular vote victory resulting from blowout victories in the Northeast, and a narrow Romney electoral college majority, with Romney squeaking by in either Ohio or Pennsylvania. Another wildcard is whether any hurricane-related gains by Obama will dissipate as the hurricane fades in the news, except for stories about those who are still suffering. A final wildcard is whether state polls are properly sampling early voters–from what I’ve read of pollster methodology, anyone who says that they voted early is automatically put down as a “likely voter.” I can easily imagine people who don’t really want to be polled but are too polite to hang up telling pollsters they already voted in the hopes that this will signal an early end to the interview. (UPDATE: And, though I haven’t checked the figures myself, others who claim to have say that the polls are substantially overestimating early voters compared to the known early voting figures–which would benefit Obama in the polls).

We’ll find out tomorrow. Meanwhile, I agree with Ilya (and Intrade) that Obama is about a 65-70% favorite.

UPDATE: It’s worth noting that in 2000, before the election everyone was talking about the possibility of Bush winning the popular vote and Gore winning the electoral college. Of course, it turned out to go precisely the other way.

FURTHER UPDATE: The new Politico poll, conducted Sunday and Monday, strongly favors Romney. The headline number is 48-48, but the internals tell an even better story: (1) Romney leads independents by 14%; (2) Romney’s 47% is 44% definite, 4% probable (I assume this is something like 43.7 plus 3.7&, rounded down to 47%); Obama’s is 44% definite, 2% probable, 1% lean (rounded); (3) Obama’s disapproval rating is 53%; (4) Number 3 suggests that the undecideds will break against Obama.

IDB/TIPP also has a new poll out, showing a 1+% Obama margin. It’s a very odd poll because it shows Romney and Obama almost even in the West, and also almost even in the South. (By contrast, the last, pre-hurricane iteration of the poll, much more plausibly, showed an Obama landslide in the West, and a Romney landslide in the South. The poll also shows a 17% Romney lead among Catholics, compared to only 1% pre-hurricane, which seems like an implausibly large shift. Maybe these anomalies balance out, maybe it’s garbage in, garbage out. In any event, the ultimate result is very similar to the final pre-hurricane poll.

The final Rasmussen poll, conducted Saturday-Monday, still has Romney +1.

FINAL UPDATE: Nine “final” major national polls show the popular vote within 1% in either direction. Two show Obama with a 3% lead. Nate Silver, heavily weighting state polls, gives Obama a 2.5% popular vote margin. If the final popular vote comes closer to the 2.5% Obama margin than to RCP’s .7% Obama margin, kudos to Silver. My “money” is on the national polls.