In 2012, my race ratings were stellar but not perfect: I blew one state in the presidential race (Florida), one Senate race (Montana), and a dozen House races, although I had the Democrats gaining 8 seats, which turned out to be exactly correct. If anything, my 2012 pattern would suggest I’m slightly pessimistic. However, in 2010 I underestimated the size of the Republican wave, although I did expect the GOP to pick up the House. With all that said, here is a brief explanation for my race ratings.
For the Senate and Governorships, where I predict the exact margin, a race within three points is a Tilt race, a race with a margin of four to six points is a Lean race, and a race with a margin of seven to twelve is a Likely race. I have also attached confidence intervals; if my projections are accurate, 60% of Tilt races, 75% of Lean races, and 95%+ of Likely races should be called correctly. I’m substantially less cautious than professional prognosticators like Charlie Cook, who I believe contribute little to no information because of their excessive caution in moving races away from Tossup status. I’m going to begin with the Senate.
Safe Republican:
Alabama (Sessions 100-0), Idaho (Risch 68-32), Wyoming (Enzi 64-32), Mississippi (Cochran 64-35), Tennessee (Alexander 63-34), Oklahoma (Inhofe 61-35), Maine (Collins 61-35), Texas (Cornyn 60-38), South Carolina (Graham 60-38, Scott 57-43), Kansas (Roberts 58-39), Nebraska (Sasse 57-41), Montana (Daines 56-40)
Nothing is too controversial here; I believe that with everything going on, Montana is unwinnable for Team Blue and they will not invest a penny.
Safe Democratic:
Delaware (Coons 100-0), Rhode Island (Reed 100-0), Hawaii (Schatz 62-36), Virginia (Warner 58-41), New Jersey (Booker 56-42)
These are all self-explanatory as well. Not many Safe D seats up this cycle.
Likely Republican:
South Dakota (Rounds 48-37-11), West Virginia (Capito 56-44)
Rounds hasn’t quite been able to put the race away but it’s very tough to see a path to victory for Weiland; South Dakota’s liberal and center-left votes max out at 40%.
Likely Democratic:
New Mexico (Udall 56-44), New Hampshire (Shaheen 54-43), Oregon (Merkley 53-42), Minnesota (Franken 51-42), Michigan (Peters 52-45)
I see no path to victory outside of scandal for any of the five Republican challengers listed here. That includes Terri Lynn Land in Michigan, who had an outside shot if she turned out to be a great candidate or 2014 turned out to be a great Republican year, but neither are occurring.
Lean Republican:
Kentucky (McConnell 51-47), Louisiana (Landrieu at 47% in jungle primary, Cassidy 54-46 in runoff), Georgia (Perdue 49-48, 53-47 in runoff)
At this point, it’s worth talking about each race. I don’t see a favorable enough climate for Democrats where a Republican incumbent in a 60%+ Romney state could lose, even if that Republican is as disliked as Mitch McConnell. Grimes is a great candidate, and getting to 47% in 2014 as a Democrat running for Senate in Kentucky will be extremely impressive, but it won’t send her to Washington. In Georgia, I don’t see how Nunn can get to 50% in November, and her support will drop at least slightly in a runoff, no matter how hard it is contested. If anything, I might be projecting a closer runoff than would actually occur. And if there’s any race I might be too high on for Democratics due to rose-colored glasses, it’s this one. Finally, there’s Louisiana, where I know many are projecting a closer race. Let’s think about this. In 2008, the best Democratic year in Mary Landrieu’s adult life, she got 52%. People can quibble about what exactly November’s political climate will be, but it surely will be at least 4 points worse for Democrats than 2008 was. So how does Landrieu lose almost nothing from her 2008 totals? Weak opponent? Cassidy’s essentially Generic R. Better Black turnout? Certainly not, although Katrina returnees will help some in New Orleans. Performing better with Whites? No way. So I don’t see how she wins. Louisiana is a Lean R race, and Mary Landrieu is the most vulnerable incumbent in the Senate.
Lean Democratic:
Iowa (Braley 52-46), Alaska (Begich 50-45), Colorado (Udall 51-46)
All of these races are considered Tossups by overcautious prognosticators. Let’s examine why they aren’t. In Iowa, Republican Joni Ernst is more conservative than anybody Republicans have won Iowa with since essentially forever. Bruce Braley is trying his damndest to make sure she sets the record for most conservative senator from Iowa, but I can’t see how a state that voted Democratic in six of the last seven presidential elections votes for Joni Ernst. At some point, people will realize she is much more conservative than them, and she’ll lose, although she’ll keep it respectable. In Alaska, polling tends to underestimate the incumbent. In addition, voters are quite elastic. Begich starts with a higher floor than any red state Democratic except Hagan and has a swingier electorate than she does. I’ve always been high on his chances, and that isn’t changing now. In Colorado, Mark Udall probably couldn’t lose to anybody in this climate, but Cory Gardner has a poor geographic base and is at least as conservative as Udall is liberal. In a tossup state, that’s not a recipe for success against an incumbent.
Tilt Democratic:
North Carolina (Hagan 50-47)
Kay Hagan has the good fortune of running against a Republican people hate. That can make a mediocre candidate win in a mediocre climate, and it likely will here.
Tilt Republican:
None
Pure Tossup:
Arkansas (Cotton 49-49)
I try to keep Tossups to races where I really just can’t figure out what is going on, and this election is one of them. Both sides still have a good shot at victory, and their paths rely on persuading rural Blue Dogs to either finally make the leap to voting R, something they have only done for Senate in 2010 (Cotton) or give the guy they’ve always liked one more term, despite national Democratic policies (Pryor). I also personally find Cotton to be a poor candidate, in contrast to Steve Daines, Shelley Moore Capito, Dan Sullivan, or Bill Cassidy. I just don’t know what will happen, but I’m giving the smallest of edges to Cotton, especially after the latest poll.
Projection: GOP gains in Montana, South Dakota, West Virginia, Louisiana, and Arkansas, leading to a tied 50-50 Senate with Democratic control
As for the statehouses (note that I didn’t rate Arizona or Rhode Island because of their competitive primaries)…
Safe Republican:
Idaho (Otter 66-28), Tennessee (Haslam 67-31), Wyoming (Mead 65-32), South Dakota (Daugaard 64-33), Nevada (Sandoval 59-36), Alaska (Parnell 53-32-15), Oklahoma (Fallin 59-39), Alabama (Bentley 58-40), Texas (Abbott 56-41), New Mexico (Martinez 57-43), Nebraska (Ricketts 55-42)
None of these are considered competitive by analysts, although some lump a few of the ones with closer margins into the Likely R category. But none of these Republicans will lose.
Safe Democratic:
New York (Cuomo 63-32), Hawaii (Ige 62-37), Pennsylvania (Wolf 60-37), California (Brown 61-39), New Hampshire (Hassan 58-40), Oregon (Kitzhaber 57-40), Vermont (Shumlin 57-40), Maryland (Brown 56-42)
See Safe Republican.
Likely Republican:
Iowa (Branstad 54-42), South Carolina (Haley 54-44), Ohio (Kasich 53-43)
All these would require everything to go right for the challengers, probably including the R brand becoming toxic due to impeachment or a major government shutdown, and even then it’s probably a narrow R win.
Likely Democratic:
Massachusetts (Coakley 54-42), Minnesota (Dayton 52-43), Maine (Michaud 46-38-16)
Cutler will fade as almost every 3rd party candidate does, giving Michaud a comfortable win.
Lean Republican:
Georgia (Deal 51-45), Illinois (Rauner 50-45), Wisconsin (Walker 51-47)
In Georgia, Deal should outrun Perdue by a bit, but scandals could move this race to Tilt R. In Illinois, Pat Quinn is looking like a goner. He looked like one last time, too, though. But I still give Rauner a decent edge. In Wisconsin, Scott Walker has always looked like he’d win by two to six points, and I’m sticking with the middle of that range right now. But remember, Lean means 75% confidence…
Lean Democratic:
Connecticut (Malloy 51-46), Florida (Crist 52-47)
I’m less worried about these two Democrats than polling shows. Malloy is in a blue state and is running against a weak candidate, and Crist is just getting hammered on air right now and will likely recover once he’s at close to parity with Rick Scott, the same way Kay Hagan has against Thom Tillis. Plus Rick Scott barely won in the 2010 wave against a weaker candidate.
Tilt Republican:
Michigan (Snyder 50-48), Arkansas (Hutchinson 50-47)
Both of these races are in flux, with Schauer seemingly improving in his challenge to the Governerd in Michigan, and Hutchinson pulling away in red Arkansas. But right now, they both stay here, just to the Republican side.
Tilt Democratic:
Colorado (Hickenlooper 49-47)
The embattled Hickenlooper should win, but he may well get held below a majority, certainly not bolstering his VP hopes.
Pure Tossup:
Kansas (Davis 47-47)
I decided that to balance out my other tossup going Republican, this one would go Democratic. And thus, I have Paul Davis narrowly beating incumbent Sam Brownback in a recount.
Projections: Republicans defeat an incumbent in Illinois and pick up Arkansas, while Democrats defeat incumbents in Pennsylvania, Maine, Florida, and Kansas. In the end, Democrats net 2 governorships.
House of Representatives:
I’m only listing Lean races (75%), Tilts (60%), and two Tossups (50%). I’m not listing AZ-1 (Kirkpatrick) or LA-5 (McAllister), as the primaries have yet to occur.
Lean Republican:
CA-52 (Carl DeMaio vs. Scott Peters(i) ), MT-AL (Ryan Zinke favored to take the open seat), NV-3 (Joe Heck), MN-2 (John Kline), WI-6 (Glenn Grothman favored to win primary and general), NY-21 (Elise Stefanik favored to take the open seat), NY-23 (Tom Reed), MI-1 (Dan Benishek), MI-7 (Tim Walberg), IA-4 (Steve King), WV-2 (Alex Mooney favored to take the open seat), MI-8 (Mike Bishop favored to take the open seat)
I’m expecting 75%, or 9 out of 12 of these, to go to the GOP, and in fact that may be lowballing it a bit. The seats are a mix of open seats in purple or light red territory, Republican incumbents who won relatively narrow victories in 2012, and polarizing archconservatives (King, Walberg, and Grothman, although he’s also in an open light red seat). The final seat is CA-52, where I’ve always been pessimistic about Scott Peters’s chances; this was confirmed by his poor June showing in the Top Two primary and I consider him the most vulnerable House Democrat.
Lean Democratic:
CA-7 (Bera), CA-26 (Brownley), NY-11 (Dominic Recchia favored over incumbent Michael Grimm), NH-2 (Kuster), ME-2 (Emily Cain favored to take the open seat), MN-8 (Nolan), GA-12 (Barrow), IL-10 (Schneider), NY-18 (Maloney)
Here we have six vulnerable freshmen, one open seat, one Blue Dog in a red seat without an A-list challenger, and one indicted incumbent. I doubt many of these ratings should be too surprising.
Tilt Republican:
AZ-2 (Martha McSally favored over incumbent Ron Barber), NE-2 (Terry), FL-2 (Southerland), IL-13 (Davis), VA-10 (Barbara Comstock favored to take the open seat), NY-1 (Lee Zeldin favored over Tim Bishop), NJ-3 (Tom MacArthur favored to take the open seat)
The two incumbent losses I project here are due to the fact that each only barely won in 2012; a slightly worse showing will lead to defeat. Lee Terry is someone who is always overestimated, so I put him at only Tilt R. VA-10 and NJ-3 are essentially even seats where I believe the year’s slight Republican tilt will push Republicans to victory along with ancestral Republican leanings. IL-13 is an inverse of my projections for Barber and Bishop. Finally, Southerland has a stronger opponent than 2012, but the seat is still a tough one for Democrats to win against an incumbent.
Tilt Democratic:
CO-6 (Andrew Romanoff favored over incumbent Mike Coffman), NH-1 (Shea-Porter), FL-25 (Garcia), MA-6 (Tierney)
Mike Coffman is a far-right congressman in a swing district; he’s got an opponent running to the center-left and should lose based on the district’s lean and his own ideology. Carol Shea-Porter doesn’t have great opponents, so I’m giving her a slight edge. Joe Garcia has a strong opponent but his incumbency gives him a narrow edge. Finally, if Tierney didn’t lose in 2012 I don’t think he will in 2014, but I’m putting him at Tilt rather than Lean because second-time candidates like Tisei can often learn from their mistakes the first time around.
Pure Tossup:
WV-3 (incumbent Democrat Nick Rahall vs. Evan Jenkins) and IA-3 (Republican David Young vs. Democrat Staci Appel)
In the first, I have the same problem as Pryor vs. Cotton: we don’t know what way the Blue Dogs will swing yet. In the latter, both candidates are underwhelming (similar to the Iowa Senate race), but as neither is as ideologically out of sync with the district as Ernst is with Iowa, it’s a Tossup.
Projection: Democrats pick CA-31 (open seat, not listed) and defeat Mike Coffman in CO-6 and Michael Grimm in NY-11. Republicans pick up UT-4 and NC-7 (open seats, not listed), NY-21 (open seat, listed) and defeat Scott Peters in CA-52, Ron Barber in AZ-2, and Tim Bishop in NY-1. Two seats are pure tossups. Therefore, Republicans gain between 2 and 4 seats.
Finally, the R vs. R races. In CA-4, I project a Tossup between Tom McClintock and Art Moore, who is running to his center. It all depends on how many Democrats pick the lesser of two evils instead of skipping the race; if 100% of voters vote in this race McClintock loses. In CA-25, I project a Tilt to Steve Knight due to geographic advantage, but this could be negated by Tony Strickland holding his base better. In WA-4, I project a Lean to Don Newhouse. It’s CA-4 but with no incumbency edge for Clint Didier.
Agreements? Disagreements? Questions? Let me know in the comments.