There are dozens of different permutations of ways for the Democrats to hold onto control of the
Senate in this year's election (well, thousands of permutations, if you're willing to entertain bizarre-but-not-mathematically-impossible possibilities like the Democratic candidate winning the Senate race in Wyoming). But the most common pattern comes down to this: the Democrats protect two of their four red-state incumbents (Mark Pryor in Arkansas, Mary Landrieu in Louisiana, Mark Begich in Alaska, and Kay Hagan in North Carolina).
Several weeks ago, it was looking like the Democrats were on track to protect three from that list, with only Pryor trailing in the averages. Landrieu, however, ran into some trouble last week with a Fox News poll that put her down by 13 points against her opponent, Bill Cassidy. Begich and Hagan, by contrast, looked fairly healthy. That's not so surprising with Hagan; she's in a state that Barack Obama only barely lost in 2012, and one where a lot of the attention right now is on the very unpopular GOP-controlled state legislature, where her opponent, Thom Tillis, is state House speaker.
Begich is in a much redder state, at least at the presidential level, and moreover, he won in 2008 by only a paper-thin margin, against the backdrop of presidential turnout instead of a midterm, and only after his opponent, Ted Stevens, was convicted of corruption. Based on that, you might assume he'd be DOA ... but he seems to have a unique appeal tailored to his state's idiosyncrasies (unlike Hagan, who seems to come across more as Generic Moderate Democrat).
Unlike many red states, Alaska has a large federal workforce and a large union presence, and its successful politicians tend not to be across-the-boards conservatives but rather pro-gun, pro-oil, and pro-getting-more-stuff-for-Alaska. In addition, most outside observers agree that Begich has been running a strong campaign, with good advertising and a ground game oriented toward getting as many votes as possible out of Alaska's Native population in its far-flung rural areas.
Unfortunately, the polling has taken a turn against Begich, making it that much harder for the Democrats to get that second hold. Most importantly, a Public Policy Polling poll from Tuesday found Begich down 43-41 against his Republican opponent, Dan Sullivan. This, by itself, was enough to push Begich's odds of winning from 56 percent on Monday down to 36 percent on Thursday. Begich has sported 5-point leads in several other recent polls, but those were both Democratic internals, which get weighted down in our model.
In fact, aside from Democratic internals, the only two other polls taken since Sullivan clinched the nomination on August 19 have also had Sullivan in the lead: by 2 in a Rasmussen Reports poll, and by 6 in a YouGov poll. It's possible we're seeing a similar phenomenon to what we saw in Kentucky, another race with a hotly contested establishment/outsider GOP primary: Alison Lundergan Grimes consistently led in the polls until the primary was over, and then consistently trailed after that. Probably what happened there was the Matt Bevin supporters, before the primary, swearing they'd never vote for Mitch McConnell, and then after the primary grudgingly admitting that, yes, they'd vote for McConnell after all. We may be seeing a similar situation in Alaska, with extreme tea partier Joe Miller's supporters finally admitting that they'll vote for the generically conservative Sullivan after all.
We'll discuss the other Senate races, and the gubernatorial races, over the fold:
Pryor, in Arkansas, did get a bit of good news on Wednesday, in the form of a poll from Suffolk/USA Today giving him a 2-point lead over Tom Cotton. You'd think that would push his odds up in the Arkansas Senate race, but the Suffolk poll is more than compensated for by the Public Policy Polling poll that came out on Tuesday giving Cotton a 5-point lead. That pushed Pryor's odds down from 37 percent last Monday to an even worse 29 percent now.
If you're looking for some good news, Jeanne Shaheen's chances in New Hampshire were bolstered by a new poll from New England College giving her a 7-point lead over Scott Brown, which pushed her odds up from 77 percent to 82 percent. New Hampshire's not really a pivotal race that's central to determining Senate control, though, so between Alaska and Arkansas, Democratic odds overall in the Senate dropped to their lowest point since we've started running the model a month ago: from 41 percent on Monday down to 37 percent now. The median and modal number of Democratic-controlled seats remains steady at 48, though.
In fact, that's the beginning of something unusual, if you take a brief look at Vox's aggregate-the-aggregators page: For what I think is the first time, the Daily Kos Election Outlook, which is one of the all-polls models, is more pessimistic about the Senate than most of the models that incorporate "fundamentals." For instance, the New York Times' Leo model puts Democratic odds at 40 percent, and 538 puts Dem odds at 41 percent. (Huffington Post's model moved in line with everybody else, too; that leaves only Sam Wang's Princeton Election Consortium model standing on its own in Dem-friendly territory.)
One thing that I've been repeating ad nauseam for several weeks is that our model was more optimistic than the fundamentals models because it seemed like the Democrats were overperforming what the fundamentals (i.e. presidential approval, generic ballot, how the in-power party tends to fare in midterms) would predict. Well, we may have gotten to a point where the Democrats are now underperforming the fundamentals in the polls ... though that seems to be heavily driven by a handful of recent polls in key races that have kind of an outlier-ish smell about them (like the Fox News poll in Louisiana, and the Quinnipiac polls in Colorado and Iowa). So that's one potential advantage to the fundamentals approach: It can insulate you somewhat against outlier-caused volatility.
Finally, let's turn to the gubernatorial side. There wasn't a lot of change in the big picture, with 22 still the median number of gubernatorial seats for the Democrats (an improvement of one over now), and a 54 percent chance of gaining seats, down from 57 percent on Monday. Democratic odds in a couple sparsely-polled races fell a bit with new polls, though; in Alaska, that same PPP poll found quasi-Dem independent Bill Walker leading Republican incumbent Sean Parnell only 42-41; that small lead was enough to push Walker's odds from 66 percent down to a totally even 50/50.
Two dueling internals in Arizona also pushed Democrat Fred DuVal's odds down, from 41 percent to 26 percent. One of those was a Dem internal showing DuVal down 2; considering that he's been down 1 in a YouGov poll and the race had been tied in a Rasmussen poll, I'm not really sure what was to be gained by leaking that poll.
Instead, the best polling news on the gubernatorial front came in Hawaii, where a Merriman River Group poll put Democrat David Ige ahead of Republican Duke Aiona and independent Mufi Hannemann 43-39-8. That's the largest lead Ige has seen, after trailing Aiona in several pre-primary polls, and it pushed his odds from 52 percent to 67 percent. But it's playing out as we expected, that Ige would consolidate support and get on better financial footing after dispatching Neil Abercrombie in the primary, and it seems likely that Ige's position will only improve from here in that dark-blue state.