November 5, 2012 My not-very-confident election predictions
First, the popular vote.
The RCP (realclearpolitics.com) average is 47.8 Obama, 47.3 Romney. In both 2004 and 2008, the RCP poll of polls was pretty accurate (see here and here) , so I don’t see why this year should be different. What happens to the undecideds? I would guess that they break about 55-45 for the challenger, since that’s more or less what happened in 2004. So (assuming that, is in 2008, 1.5 percent of voters go with minor parties) the 3.4 percent remaining undecided break 1.9 for Romney and 1.5 for Obama, the final result is: 49.3 Obama, 49.2 Romney.
What about the electoral college? Just to make this easier, let’s look at the RCP Electoral Map. (You can click on links for each state to see the polls referenced below). It lists 201 electoral votes as pretty sold for Obama, 191 for Romney. To save time, let’s assume that’s the case. It lists 11 states as tossups, so let’s go through them.
Colorado: a true dead heat. Five recent polls according to RCP, three lean Obama two lean Romney. However, they are all within each other’s margin of error: that is, the differences between them are statistically insignificant. In each of them Obama is between 47 and 50, and Romney is between 45 and 50 (which means if he is “really” at 47.5 the differences between them are all within a 3 percent margin of error). In a case this close, I assume that the tie goes to the Republican: that is, between late deciders and GOP enthuasiasm Romney pulls it out. So now its 201-200 Obama.
Florida: Romney is pretty clearly ahead if you average the last few polls together. Of the last four polls (ending Oct. 30), three have Romney at 50 percent or 51 percent. (The outlier has him at 47- again, not a statistically significant difference). It seems to me that even if undecideds break evenly Romney carries Florida. 229-201 Romney.
Iowa- Every recent poll but one shows Obama ahead. Unless there is a nationwide shift to Romney (which certainly could happen!) Obama wins. To put it another way, if Romney wins the national popular vote by a significant margin he carries Iowa and some other states that now seem to be leaning Democratic. But if (as I suspect) it will be basically a 50-50 election, Iowa is in the President’s corner. 229-207 Romney.
Michigan- Another state where every poll but one shows Obama ahead. On the other hand, the poll that shows Romney winning has a really large sample size compared to the others. On the OTHER other hand, Rasmussen (which generally tends to favors Republicans) has Obama 5 points ahead. With somewhat less confidence I’m calling Michigan for Obama. 229-223 Romney.
Neveda- Every single poll says Obama wins Neveda, though all are close. 229-229 Romney.
New Hampshire- the five most recent polls all show either a tie or a slight Obama lead. For the reasons stated in my discussion of Iowa, I think this is enough to say Obama wins NH. 233-229 Obama.
North Carolina- Every single recent poll shows either a tie or a Romney lead. So just as that combination justifies calling Iowa and NH for Obama, it justifies calling NC for Romney. 244-233 Romney.
Ohio- Of the last twelve polls, one shows a tie. The rest show Obama leads of up to 6 points but mostly in the 2-4 point range. Unless there’s a nationwide Romney surge, I think that’s enough to justify calling Ohio for Obama. 251-244 Obama.
Pennsylvania- As with Ohio, one poll shows a tie, the rest a small Obama lead. I think that justifies calling Pa. for Obam (271 votes!)
And just to finish up…
Virginia- Some polls go one way, some the other. In that situation I’m guessing tie goes to the Republican, as in Colorado. 271-257 Obama.
Finally, Wisconsin- Where again every poll but one shows Obama ahead, and that a tie. So that means to me that in a 50-50 election Obama wins Wisconsin for a 281-257 victory.
Note that if Romney had picked Rob Portman of Ohio and Portman had delivered Ohio’s 18 votes, that would have given Romney 275 votes and the election. Hmmm….
First of all, it seems to be a given that the Republicans will pick Ben Nelson’s Nebraska seat, and that Olympia Snowe’s Maine seat will go to Independent Angus King (which means there will be 2 Independents in the Senate, both of whom I suspect will caucus with Dems). So going into the “tossup bloc” the Republicans have a net pickup of zero.
But there are a dozen races that RCP calls tossups. What happens to them?
Arizona- The last two polls show Republican Jeff Flake ahead, and the last one that rules otherwise was in early October. So the Republican will hold this open seat (previously held by Jon Kyl).
Connecticut- In the summer it looked like Linda McMahon might pick up this open seat for the Republicans. But every poll taken in the last few weeks has Rep. Chris Murphy ahead, so I think he holds the seat for the Dems.
Indiana- Right-winger Richard Mourdock beat longtime incumbent Dick Lugar in the primary, and was apparently leading until mid-October. But some ill-conceived remarks caused Mourdock to slip; even Rasmussen has him behind moderate Democrat Joe Donnelly. DEM PICKUP.
Mass.- Scott Brown was running ahead of Harvard prof Elizabeth Warren in the summer but has fallen behind, probably due to Democrats coming home in this Democratic state. DEM PICKUP.
Mo.- Another Republican pickup blown by a weak-candidate. Republican Todd Akin was at first ahead of scandal-ridden Democrat Claire McCaskill, but has also made various ill-concieved remarks. No poll shows Akin ahead. I note, however, most polls show this to be pretty close so I wouldn’t be amazingly shocked if Romney coattails carried either Akin or Mourdock of Indiana to victory. Nevertheless I am guessing McCaskill holds the seat.
Montana- Incumbent Democrat Jon Tester has run a tight race unless at-large Congressman Denny Rehberg. Some polls lean one way, some the other. It seems to me that when in doubt, assume that straight-ticket voting happens in large enough quantities that whichever party gets a state’s Presidential votes will pull their party’s Senate candidate to victory in a tight Senate election. So I am guessing Rehberg noses out Tester. R PICKUP
Neveda- Every recent poll has shown incumbent Republican Dean Heller leading (albeit by modest margins) or tied with his opponent. R hold.
North Dakota- This open seat, formerly held by Kent Conrad, is likely to be picked up by Republican Rep. Rick Berg, who has lead in the last few polls. R PICKUP.
Ohio- Incumbent Sen. Sherrod Brown has had a ugly race against Republican Josh Mandel. Every poll shows a small Brown lead or a tie, just as at the Presidential level every poll shows a small Obama lead or a tie in Ohio. So IF I am right in guessing that Obama does win Ohio, Brown does as well. (On the other hand if Romney surges all bets are off).
Pennsylvania- Bob Casey has had a tougher than expected race against coal tycoon Tom Smith. But no poll shows him trailing. So I think he holds (barely).
Virginia- This Democratic open seat features two former governors, Republican George Allen against Democrat Tim Kaine. As in Virginia’s Presidential race,some polls lean one way, some the other. So I think that this race goes the same way as the Presidential race, which in Virginia (for reasons stated above) means R PICKUP.
Wisconsin- A tight race between lesbian Rep. Tammy Baldwin and former Gov. Tommy Thompson. Every poll in the last week or so shows either a tie or a Baldwin lead. On the other hand, I can’t help suspecting that there are people who won’t vote for a gay candidate who won’t admit it. And if one in 100 people (or more precisely, one in 50 Baldwin voters) fall in this category that’s enough to put Thompson over the top. So I am guessing an upset here- R PICKUP.
So there are a grand total of 3 seats lost by Republicans (Maine, Mass., Indiana) and 5 gains (Nebraska, Montana, North Dakota, Virginia, Wisconsin) which means a Senate with 49 Republicans, 49 Democrats, and 2 Independents. I suspect King of Maine will caucus with the Dems, since Republicans have campaigned much more aggressively against him than Dems. So total 51 DEMS 49 REPS.
Note that the Republicans will have lost 2 races due to nominating crazed right-wingers (Indiana and Missouri).
Based on what I’ve read it doesn’t seem like the Democrats will gain much in the House. I am not familiar enough with each House race to guess who wins. But RCP lists 224 R seats, 178 D seats and 33 tossups. If the tossups break 17-16 R, the Republicans get 241 seats – exactly what they have now.