The state of the Senate playing field -- as portrayed by the election models of the Washington Post, the New York Times and FiveThirtyEight -- has been enough to give even the most veteran political observers some whiplash of late. A few weeks back, I wrote that the models seemed to be coalescing around a Republican majority in 2015. Then came movement toward Democrats -- including WaPo's Election Lab showing that the party had a slight edge to retain their majority.
That left plenty of people understandably confused. So, in an attempt to smooth out what seems like the daily peaks and valleys for in the fight for Senate control, I decided it made the most sense to look at where the trio of models showed the race on the six Mondays between now and Nov. 4. (Why the models move around -- and why you should be wary about putting too much faith in any prediction made by any of the models on a single day -- is well explained in this Monkey Cage blog post.)
As of today then, all three models agree that Republicans are slight favorites to win the Senate majority. The Post's Election Lab is the most bullish (65 percent probability of a Republican majority) while the Times LEO model pegs the chance at 55 percent and Nate Silver's FiveThirtyEight puts it at a shade under 55 percent.
In all three models, Republicans are favored in five Democratic-held seats -- the largely non-competitive open seats in Montana, South Dakota and West Virginia as well as the contests in Louisiana and Arkansas where incumbents Mary Landrieu and Mark Pryor, respectively, are struggling to overcome the deep unpopularity of President Obama in their states. (Worth noting: In all three models, Pryor is given better odds than Landrieu at winning.)
If the models are correct, that would put Republicans at a five-seat gain and the GOP would need just one more pickup to claim the majority. It's at this point where the models disagree a bit -- and things start getting complicated. And, it's also in these next four races where the majority likely will be decided. Let's go through them.
* Alaska: FiveThirtyEight (56 percent Republican), LEO (62 percent Democratic), Election Lab (81 percent Democratic)
* Colorado: FiveThirtyEight (50 percent chance for both sides), LEO (55 percent Democratic), Election Lab (67 percent Democratic)
* Iowa: FiveThirtyEight (52 percent Democratic), LEO (53 percent Republican), Election Lab (70 percent Republican)
* Kansas: FiveThirtyEight (64 percent Independent), LEO (54 percent Independent), Election Lab (73 percent Republican)
The first three states listed above are all potential majority makers for Republicans. Relying solely on the wisdom of the models, Iowa's open seat where Joni Ernst (R) is facing off against Bruce Braley (D) looks like the best chance of the three for the GOP. But then there is Kansas where Sen. Pat Roberts (R) finds himself in an increasingly difficult race against independent Greg Orman who, apparently, will not need to worry about a Democrat on the ballot splitting the anti-Roberts vote. Orman's unwillingness to disclose which party he would caucus with makes the attempt to handicap what it would mean if he beat Roberts all the more difficult.
(For those wondering about Kentucky and Georgia -- Democrats' other two pickup opportunities -- both seem to be fading from sight. All three models show Republicans with a strong chance of holding both seats. In Kentucky, the probabilities range from 85 percent to 99 percent; in Georgia it ranges from a 77 percent chance of a Republican victory to a 95 percent one.)
For Republicans hoping to expand the playing field to counter Roberts' emerging problems, the best chance to do so looks like North Carolina. All three models show Sen. Kay Hagan (D) as the likely winner but two of the three -- FiveThirtyEight and LEO -- put her odds around 75 percent. (Election Lab says Hagan has a 99 percent chance of winning.) Both Michigan, an open seat, and New Hampshire, where Sen. Jeanne Shaheen (D) is running for reelection, are longer shots, according to all three models -- although Election Lab is, again, far more bullish about Democrats' chances than FiveThirtyEight and LEO.
So, six weeks and one day until the election, Republicans have the edge, according to the election models. It's not conclusive. But it's there.