Monday, November 3, 2014

Final Election Predictions + 2014 Election Night Guide

One more dawn, one more day, one day more!

Here are my final election predictions. After an extremely rocky road to the Election day with polls showing 10 tossups according to RCP, this is going to be an exciting night! The Democrats are looking to lose some House seats (I predict eight) and gain three Governorships. The Senate is too close to call and we may not know the final outcome until January.

Reps will automatically pick up Montana, South Dakota and West Virginia. Arkansas looks increasingly likely to go red too. The other seats however will be explained in the post.

Also, I will be including county baselines with the two party vote. The baselines were calculated using the Obama 2012 numbers as well as numbers from a recent statewide race that one of the candidates on the ballot ran in before 2014. This statewide race is factored in to look at regional trends while Obama 2012 looks at recent trends. Also, click on the maps for the full picture.

Here is a Google doc with a detailed baseline breakdown, keep this handy during election night:
https://docs.google.com/a/usc.edu/spreadsheets/d/1qnT5zDXyZwGhJKYL3Dl6oomua5aJ_XI1QubxxwOTjMw/edit#gid=36061957

Alaska:
Alaska does not report by county. My only note though is that the first votes in Alaska usually lean Republican because the rural areas report late. The rural areas have a large Native American population and Begich has been aggressively courting their vote. While I am not as optimistic here, they can help narrow the race.

Prediction: Sullivan +1

Colorado:
Jefferson Jefferson Jefferson, if you want to know the final result in Colorado, look at Jefferson County (also known as JeffCo). It is a suburban Denver County with a leaning Democratic north and leaning Republican south. The baselines have Udall at 49.9% there in the two party vote. Other important counties to watch are Arapahoe (Udall needs to win it by 3.6% or more) and Larimer County (Udall is at 49.4% in the two party vote). These three counties will decide the election, especially Jefferson. Watch these three counties only, period.

I predict that Udall squeaks by here, this may go to a recount. The reason is that polling always underestimates Dems in Colorado and with the race tightening a bit; Udall has a good chance to pull it out. Udall by 0.5-1.

(Baselines are Obama 2012 and Udall 2008)



Map legend:
Dark Red = Republican by 10 or more points
Light Red = Republican by 0%-9.9%
Light Blue = Democratic by 0%-9.9%
Dark Blue = Democratic by 10 or more points

Georgia:
Polls have shown an extremely tight race expected to go into runoff. Watch Gwinnett and Cobb Counties. If Nunn gets 47% of the vote in those counties, she has a great chance at entering the runoff. These counties were once Republican strongholds but Hispanic and African American growth have made them less Republican like the rest of Georgia. Also, Nunn should perform well in south Georgia due to her Father’s popularity with voters there. Look at Decatur County as a barometer for the runoff, if she breaks 50% there, she will probably break 50% statewide, if not, then she will be in a runoff.

Prediction is that this race goes to runoff. Nunn leads by 1 on November 4th.
In the Gubernatorial race, Deal leads by 2, also goes to runoff.

The baselines include the 2012 Presidential election and the 2010 Gubernatorial election because the Democratic candidate had rural strength.



Iowa:
There are no baselines for this race because neither candidate has run statewide. Due to historical trends though, Braley needs a strong margin out of Polk County (10 or more points) and needs his margin in Johnston County to be 30 points or more. Overall prediction is that Braley barely pulls it out thanks to the ground game and the history of Dems getting underestimated in Iowa polls (the RCP average had Obama up by 2 and he won by 6).

Prediction: Braley by 1.

Kansas:
To win, Orman needs to run up the score in the eastern Kansas urban areas. He needs to win Leavenworth County by more than 30 points. Counties Orman must win include Sedgwick (Wichita) and Johnson (Kansas City suburbs). If Orman loses both of them, he is done. If he wins one, he will probably lose and if he wins both, he is in good shape.

Prediction: Roberts by 1, I see Kansas coming home to the Republicans.

Kentucky:
For Grimes to win, she must run up numbers in Jefferson County (Louisville) to offset rural Republican margins. Democrats have collapsed in rural Kentucky so she needs urban margins. A winning Grimes electorate will have her winning Jefferson County by 61.7% or more of the two party vote. Important bellwethers to watch include Pike County (Grimes 50.3%) and Lyon (Grimes 50.1%) Pike County is in coal country and will show how well Grimes is performing among the coal country voters.

Grimes has been sinking in recent polls. A Marist poll shows her down by nine and while I do not expect her to lose by that much, I do not see her winning this and this is hard for me to say because I am a major supporter. I predict McConnell wins by 5.

(Baselines are Obama 2012, Grimes 2011 and McConnell 2008)



Louisiana:
More detailed explanation on Louisiana baselines + close counties here: http://racesandredistricting.blogspot.com/2014/10/louisiana-county-benchmarks.html
The final prediction is a runoff. I will release a final prediction for the runoff a few days before it after seeing polling and the November 4th election results.

New Hampshire:
Shaheen has led in most polls here (and New Hampshire underestimated Dems in 2012,) plus Scott Brown (R) has reinforced his narrative as a carpetbagger by thinking Sullivan County was in northern New Hampshire. The county to watch is Rockingham, if Shaheen can win it or lose it by less than five, Brown should drive his truck back to Massachusetts.

Prediction: Shaheen by 3.

North Carolina:
Many Democrats are optimistic about North Carolina. The early votes show Democrats have much higher turnout than 2010 plus polling shows Hagan with a narrow lead which is important because unlike other states, Democrats are not severely underestimated in polling. For Hagan to win, she needs to turnout the urban Democratic counties and hit these marks in them (Mecklenburg 61%, Orange 70%, Guilford 59%) The two major bellwether counties are New Hanover and Nash so if Hagan wins those, expect a good night for her.


I predict Hagan wins by 3.
(Hagan 2008 and Obama 2012)



Overall, the Senate range could be anywhere from Republicans gaining four seats to eight seats. The playing field is wide open and the Senate appears to be a coin flip, especially with new polls on Sunday and Monday confirming that.

The three major Gubernatorial races I will be watching are the Florida and Wisconsin Gubernatorial races, both because I dislike Scott and Walker and will be extremely happy if they lose.

Florida:
Florida always has extremely close elections and this one will be no exception. For Charlie Crist (D) to win, he must perform well in the I-4 Corridor, a formerly swingy area that is trending Democratic. He must win Hillsborough County by at least five points (Alex Sink (D) who ran for Governor in 2010 lost it by four). Another I-4 county to watch include Orange County which is Democratic but has bad midterm turnout. Crist must win by at least 15 points there. The most populous and important county to watch in Florida is Miami Dade County. Crist needs to win by at least 19 points there and he can do that by winning that important Hispanic vote.

Prediction: Crist by 2.

(Note: Crist may overperform the baselines in counties around Tallahassee thanks to residual Democratic strength there but the change should not be enough to impact any of the other major county baselines. The baselines btw are from the 2010 gubernatorial election and the 2012 Presidential election).




Michigan:
Rick Snyder (R) has seen his lead evaporate and is in a close reelection contest with Mark Schauer (D). In order for Schauer to win, he needs to win Wayne County (Detroit) by at least 40 points and win Oakland County which is a big bellwether. He also needs to win or come within 3 points in Macomb County.

Prediction: Snyder by 1, the Peters coattails may not be enough.

Wisconsin:
Most polls show this race within one point. Burke has run a solid campaign attacking Walker for job losses instead of the union debacle. The major counties to watch here are Dane and Milwaukee which have a large Democratic base. The anti Walker vote should be high there (Burke needs to win Milwaukee by 30 points to offset Walker’s overperformance in rural Wisconsin as predicted by the baselines,) and Burke needs to win Dane County by 41 points. A good bellwether county is Racine County, it leans slightly Republican but if Burke can keep the margin within one point or win it, she will be the next Governor.

Prediction: recount, Burke +0.3
Baselines are from the 2012 recall and the 2012 Presidential election.





No comments: