Are things getting better for Senate Democrats? Certainly many of the better (more reliable) statistical models seem to suggest they are. Nate Silver’s FiveThirtyEight moved from a 64 percent chance of the GOP gaining a majority, predicted on Sept. 3, to a 54.7 percent chance on Sept. 15. As of Sept. 16, The New York Times’ Upshot model, nicknamed Leo, put GOP chances at 51 percent; they were at 67 percent on Aug. 26. The conventional wisdom also appears to have shifted over the past week. What, if anything, has happened to cause this shift?
Polls show Democratic Sen. Kay Hagan of North Carolina leading GOP challenger Thom Tillis. (Chip Somodevilla/Getty Images)A little bit of the change can be attributed to methodological shifts among forecasters; as statistical modelers add new elements to their computations, the new data affect the output of their models. But that does not explain all of the shift. The most significant reason seems to be that in this year’s competitive Senate races in purple states—those where either Barack Obama or Mitt Romney won by narrow margins—Democrats are, for the most part, holding their own or even improving their odds.
In North Carolina, for example, Democratic incumbent Kay Hagan has now built a lead over GOP state House Speaker Thom Tillis in both private and public polling. Hagan led by 4 points in the Sept. 5-9 Elon University poll, 45 percent to 41 percent among likely voters. It is still a competitive race, and Tillis could very well win the seat, but for now the momentum seems to be with Hagan.
In the fight for the open seat in Michigan, Rep. Gary Peters seems to be getting a firmer grip on the contest, and his chances of winning have increased. Colorado’s Democratic incumbent, Mark Udall, has built up a small but measurable lead over GOP Rep. Cory Gardner. Again, this is still a very close race, but Udall looks a bit better now than he did earlier in the cycle. In Iowa, Rep. Bruce Braley seems to have a tiny lead over GOP state Sen. Joni Ernst. The Democrat is still under-performing compared with how he should be doing, but he now looks a little better.
Models also show Republican Pat Roberts in Kansas shifting from safe to endangered, although it is far from certain which party independent Greg Orman would sit with if he upset Roberts (or, for that matter, how reliable a vote he would be for Democrats if he were to win and join their conference). Rather than an obvious shift in Democrats’ favor, I still see uncertainty in Kansas.
At the same time, in Georgia—not quite a swing state but not deeply red, either—things are looking up for Republicans, with David Perdue starting to pull away from Democrat Michelle Nunn, although the race is still very competitive. Polls are producing conflicting results in New Hampshire; some show former Sen. Scott Brown closing the gap with incumbent Democrat Jeanne Shaheen, but I remain skeptical that the race has moved that much. Brown is a transplant who hasn’t worked out as well as Republicans had hoped.
The key for Democrats is to either pray and work to help one of the three scarlet-state, Democratic-held seats survive, or to pick off one Republican seat.
What isn’t moving for Democrats: the group of six seats they’re contesting in deeply red states that Romney carried by big margins. Things still look hopeless for Democrats trying to keep their open seats in Montana, South Dakota, and West Virginia. In Alaska, there is no reason to believe the situation has improved for Mark Begich. Romney won this state by 14 points, and in the aftermath of a Begich TV ad that has been widely seen as grossly deceptive, things have probably gotten a bit worse. Arkansas’s Mark Pryor certainly doesn’t look any better than he did a month or two ago; Romney won Arkansas by a whopping 24-point margin in 2012. There is little reason to believe that Mary Landrieu will hit the 50 percent-plus-one-vote mark on Nov. 4 to escape a Dec. 6 runoff in Louisiana, and nothing to suggest that a runoff with GOP Rep. Bill Cassidy would turn out well for the Democrat. Her share of the vote does not appear to be elastic—that is, likely to expand—in a runoff; instead, the conservative and Republican vote will likely consolidate.
The key for Democrats is to either pray and work to keep one of the three scarlet-state, Democratic-held seats, or to pick up one Republican seat in a state such as Georgia, Kentucky (which is looking increasingly difficult for Democrats), or Kansas (who knows what will happen there?). If Begich, Pryor, or Landrieu survive, or if Democrats pick off Georgia, Kansas, or perhaps Mitch McConnell in Kentucky, then Republicans would have to win a swing or light-blue state contest somewhere, which would mean beating Udall, Hagan, or Shaheen, or winning Iowa or Michigan.
But if Republicans can just ensure that reliably Republican voters stay Republican in 2014, then the GOP doesn’t need to win a single state that is purple, or blue of any hue, to win the Senate. The bottom line is that the odds of Republicans scoring a net gain of seven, eight, or more seats have gone down, but in my judgment a net gain of six—the minimum number they need for a Senate majority—remains pretty likely.