Month: November 2010

As we reach the end of yet another long, grueling circus of a campaign, let us reflect on a few of the larger issues in play on Tuesday before charging into the predictions. (N.B.: I got exactly one Senate race wrong in 2008 and two in 2006, and in both years I was inadvisedly blinded by my hatred of Mitch McConnell and Bob Corker. Past results do not necessarily indicate future performance, but to continue this trend I should get 0 wrong in 2010 and then -2 wrong in 2012.)

The GOP has done a poor job of managing expectations, setting themselves up for defeat even in victory. Do not misunderstand me. I don't mean literal defeat, as in failing to gain seats in both chambers. They are taking the House with 99.5% certainty and they will pick up at least four Senate seats, possibly more. The problem is that their year-long rhetoric and the overwhelming sense of cockiness radiating from the party elite really puts them in a bind. They and their media surrogates have been predicting such an overwhelming, crushing defeat for the Democrats on Tuesday that even a slightly better-than-expected performance by the blue party will read as a GOP defeat. The problem with predicting a 60-seat pickup is that when you "only" pick up 40 it looks like you underachieved. Gaining 40 House seats is really good for one election. But the Democrats would be able to look at a 40-seat loss and say "Is that it? What happened to the tidal wave of Teabaggers we've been hearing about?" In fact, I will be very surprised if the Democratic talking point after Tuesday is anything other than "This hardly looks like the revolution Glenn Beck promised."

Now. On to the races.

I hate House predictions. There are just too many races in play for one person to meaningfully track, analyze, and comment on them. The vast majority of predictions are the GOP gaining between 50 and 55 seats. My poorly-informed guess is that they will underperform that slightly based on the strength of some of the recent generic ballots. Generic ballots are a terrible tool overall, but Alan Abramowitz has done some pretty neat analysis of Gallup's generics over the years suggesting that a 4-6 point advantage for the GOP in generics corresponds to a 44-50 seat pickup in the House. I'll go with the 45-50 range and give Alan the credit if I'm right.

And now the Senate.

First, I've moved four additional races out of the Competitive category since the last update:

Lincoln is toast in Arkansas, and after some initial indications that the races might be somewhat competitive Portman (OH) and Ayotte (NH) have really pulled away from their Democratic rivals. All three of those seats are likely safe R, and of course the surprise nomination of Christine O'Donnell has taken the Delaware race out of play.

That means that of our 37 (!!!) races this year, more than 2/3rds of them – 27 in all – are slam-dunks:

These races represent a 3-seat pickup for the GOP, with the North Dakota, Arkansas, and Indiana races switching parties.

Finally, let's look at (few) the competitive races:

Isn't that something? Only 10 competitive races, and I'm being generous to include two of them (MO and KY). First, let's talk briefly about the six races with solid predictions:

Illinois: Giannoulias's lead in recent non-Rasmussen polls, combined with the sheer power of the state Democratic Party in even the worst of times, suggest he will take this one by a hair. Ironically, it may be Libertarian Mike Labno (currently polling about 6%) who sinks Republican Mark Kirk in the final tally.

Kentucky: I'd bet money on Rand Paul with a high level of confidence, but his recent surge of bad press leaves just a sliver of hope for Jack Conway. It's been a while since the Bluegrass State sent a Democrat to the Senate, and this hardly seems like the year to do it. Paul wins.

Missouri: Roy Blunt appears to have this one in the bag, and I'm calling it competitive only because of my longstanding policy of never betting against a Carnahan in Missouri.

West Virginia: Joe Manchin has run a great campaign, including a dumb but stunningly effective TV spot, and has the lead in the late stages of a very tight race. This would be a big moral victory for the Democrats.

California: Bless their little hearts for trying, but this race has simply never been that close despite the best efforts of the right to say so. Obama won this state by over a million votes. Boxer should be able to hang on by a few thousand.

Washington: Murray has a slim lead in the polls heading into the final turn. That combines with two factors – the liberal tendencies of the state and the meatheadedness of GOP opponent Dino Rossi – to favor Murray. Her lead has never been large but it has been consistent.

And finally, the four Coin Flip races. I would not bet money on any of these. I would not even bet someone else's money on any of these. They are, with one exception, Too Close to Call in every sense of the phrase. I hate making predictions on these because so much will depend on turnout and the small number of late deciders. These races are close enough for just about any small change to matter.

Wisconsin: The numbers on this one aren't that close. Ron Johnson leads in the polls and you should probably bet money on him. But Feingold is good, and he got a very late boost by winning the endorsement of essentially every newspaper in the state over the past week. Ron Johnson has done everything possible to hurt his own chances, mostly by opening his mouth near cameras and reporters. The numbers say Johnson, but I'll take a risk and say that this race ends up breaking the GOP's heart. Feingold by a hair (Hold D).

Colorado: No clue on this one. None. It is as close to a statistical and qualitative tie as any race can be. Bennet is a bland, forgettable non-entity and Ken Buck is a raging asshole. Who wins when those two personality types face off? Based solely on Colorado's recent trend toward the left and Bennet's late break in the polls, I'll go with the incumbent with 0% confidence. This race comes down to turnout. Who is more organized, Colorado Springs or Boulder? Bennet by half a hair (Hold D).

Pennsylvania: Toomey has led throughout and PA is odd politically, with the old joke noting that it is Philly and Pittsburgh with Alabama in between. My brain says I have to stick with Toomey on this one, but as close as this race has gotten (and given Obama's 10-point win in 2008) we can't look entirely shocked if Sestak pulls an upset. Toomey wins it (Pickup R).

Nevada: Good lord. Harry Reid is just awful. Sharron Angle is categorically insane. How is a voter to choose? Note the curious fact that nearly every poll shows a statistical tie and the distance between Reid and Angle is attributable entirely to a handful of Rasmussen polls that show her with 5-point leads. HMM. Angle could very well win this one given the number of A) lunatics and B) Midwestern retirees living in Nevada, but Reid may very well pull it out. Without Rasmussen's data, this race is a tie. FWIW. Angle by like 7 votes (Pickup R).

So if Ed is right, that leaves the GOP with a pickup of five seats. I am probably wrong about Wisconsin, if the poll numbers are to be trusted, but odds that I am wrong about WI are as good as the odds of Reid prevailing in NV so it could be a wash. But my theory is, what fun would it be to make the same predictions as everyone else, just blindly following the polls? I'll go out on a limb and call Feingold an upset winner along with tight Democratic holds in Colorado, Illinois, and Washington.

Final prediction: +5 GOP: 46 R, 52 D (+ 2 Independent Democrats)