Nate Silver's Blog, page 83

Are Democrats Courting Chaos In 2020 By Limiting The Power Of Superdelegates?

Welcome to FiveThirtyEight’s weekly politics chat. The transcript below has been lightly edited.

micah (Micah Cohen, politics editor): We’re here to talk about superdelegates!!!!!!

Everyone’s favorite subject, right?

clare.malone (Clare Malone, senior political writer): Extremely 2016 up in here.

micah: (This is my least favorite topic.)

clare.malone: I can’t imagine why. It’s so sexy, and the debate is totally based in facts about what happened during 2016.

micah:

Politics Podcast: The Democrats And The Trump Impeachment Question

By Nate Silver, Clare Malone, Micah Cohen and Jody Avirgan, Nate Silver, Clare Malone, Micah Cohen and Jody Avirgan, Nate Silver, Clare Malone, Micah Cohen and Jody Avirgan and Nate Silver, Clare Malone, Micah Cohen and Jody Avirgan












 












More: Apple Podcasts |
ESPN App |
RSS
| Embed


Embed Code

Michael Cohen’s suggestion last week that President Trump directed him to break the law has renewed questions about whether Democrats should aim to impeach the president. In this episode of the FiveThirtyEight Politics podcast, the crew lays out four arguments for how Democrats should approach the issue of impeachment. The team also discusses John McCain’s legacy and previews Tuesday’s primary elections in Arizona and Florida.

You can listen to the episode by clicking the “play” button above or by downloading it in iTunes , the ESPN App or your favorite podcast platform. If you are new to podcasts, learn how to listen .

The FiveThirtyEight Politics podcast publishes Monday evenings, with occasional special episodes throughout the week. Help new listeners discover the show by leaving us a rating and review on iTunes . Have a comment, question or suggestion for “good polling vs. bad polling”? Get in touch by email, on Twitter or in the comments.

The 5 Big Takeaways From Our House Forecast

Democrats are favored to gain control of the House of Representatives in this year’s midterm elections, according to the FiveThirtyEight forecast model. But — a very FiveThirtyEight-ish sentence follows — the range of possible outcomes is wide and Democrats’ prospects are far from certain. Relatively small shifts could allow Republicans to keep control of the House, or could turn a blue wave into a tsunami.

What’s behind all of this? Our methodology post goes into a lot more detail about how our forecasts are calculated. But that explanation is rather abstract, so in this article, I’m going to focus on how these factors are playing out given what we know about the political environment this year.

Theme No. 1: A broad consensus of indicators point toward Democrats performing well

In contrast to our presidential forecasts, which are heavily dependent on polling, our House model uses a broad mix of polling and non-polling indicators, including factors such as fundraising totals and historical trends in midterms. Those indicators look both pretty good for Democrats and remarkably consistent with one another:

The Lite version of our forecast, which focuses as much as possible on district-level and generic ballot polls, projects Democrats to win the popular vote for the House by 7 or 8 percentage points.
The Classic version of the model, which incorporates a lot of non-polling metrics such as fundraising and past voting in each district, also shows Democrats winning the popular vote by 7 or 8 points.
The generic ballot, which influences all three versions of our forecast, has generally shown Democrats with a lead of … 7 to 8 percentage points.
And finally, our model calculates a starting assumption about the race based on historical trends in midterms since 1946 and presidential approval ratings. It also implies that Democrats “should” win the House popular vote by about 8 percentage points — just what the other metrics show.

So you’d expect Democrats to do pretty well based on the historical propensity of opposition parties to gain ground in midterm elections, especially under unpopular presidents. And Democrats are doing roughly as well as you’d expect them to according to most indicators of the national environment.

There are a couple of exceptions — indicators that are a little out of the consensus — but both of them fall on the better-for-Democrats side of the consensus. First, Democrats have done really impressively in fundraising. Their candidates have raised more in individual contributions than Republicans in 71 of the 101 districts rated as competitive by the Cook Political Report, despite the fact that about two-thirds of these districts feature Republican incumbents. That’s unusual. Most challengers significantly trail in fundraising at this point in the cycle. Meanwhile, the results of special elections have been very good for Democrats. Our model doesn’t actually use special election results in its forecasts, but they’re part of a coherent alternative narrative in which there’s upside for Democrats relative to what our forecast shows. Donating money and voting in special elections are tangible indicators of voter engagement, and it’s possible that they point toward a Democratic enthusiasm advantage that could become clearer later on in the cycle.

Theme No. 2: However, there’s some feast-or-famine risk for Democrats

It’s much to Democrats’ credit that there are so many districts in play in all corners of the country. (Based on our accounting, Democrats have fielded a nominee in all but three of the 435 congressional districts nationwide.) But if you had to pinpoint the exact districts that Democrats should hope to win to gain 23 seats and take the House majority, you’d have a pretty hard time. We have only 215 seats rated as favoring Democrats — “lean Democrat” or stronger — which is fewer than the 218 they need to take the House.

Nonetheless, Democrats are favored to win the majority if current conditions hold because they’ll have a bunch of opportunities, even as underdogs, to win those extra seats: 14 toss-up races, 19 “lean Republican” races and 53 “likely Republican” contests. Those are a lot of lottery tickets to punch, even if Democrats aren’t necessarily favored in any individual race.

But Democrats would have a problem if there’s a shift in the national climate toward Republicans, or even if there’s a relatively modest systematic polling error in the GOP’s favor on Election Day. All of the sudden, they’d lose most of the toss-up races along with some of the “lean Democrat” races — and the “lean Republican” and “likely Republican” seats would become an uphill climb at best.

The flip side to this is that if the political environment gets better for Democrats, their seat gains could pile up at an accelerating rate. There are a plethora of districts that are 10 to 20 points more Republican than the country as a whole, a lot of which were gerrymandered to be “safe” for Republican candidates — but where the gerrymanders could fail in the event of a large enough wave.

Theme No. 3: Incumbents — especially Republican incumbents — are really vulnerable

The first line of defense for a party hoping to maintain its majority is incumbency. Even if the national political climate is unfavorable, its incumbents may be popular enough in their districts to withstand the wave.

The issue for Republicans is that the incumbency advantage has been weakening over time — and it appears to be especially flimsy this year. In the 1990s, incumbents overperformed the partisan baseline of their districts by somewhere on the order of 20 percentage points. (So, for example, a district that might favor Republicans by 2 points in an open-seat race would go to the GOP by 22 points if there were a Republican incumbent running.) In more recent elections, as Congress has become less and less popular, the incumbency advantage has eroded to more like 10 to 12 percentage points. And between Republicans’ anemic fundraising, GOP incumbents’ voting records — which are highly aligned with President Trump’s positions, even in purple districts — and reasonably good district-by-district polling for Democratic challengers, our model is projecting only about a 6-point advantage for GOP incumbents this year. Plus, a lot of Republican incumbents have retired.

Our forecast also shows a relatively narrow advantage for Democratic incumbents. But Democratic incumbents have little exposure in the House: Any Democratic representative who was strong enough to survive the GOP waves in both 2010 and 2014 probably won’t have any problems this year. (It’s a entirely different story in the Senate, where there are lots of vulnerable Democratic incumbents who were last re-elected in the strong Democratic year of 2012.)

Theme No. 4: Potential Democratic gains are broad-based, across all regions of the country

One factor helping Trump in 2016 was that he really needed to beat his polls in only one part of the country, the Midwest, to defeat Hillary Clinton in the Electoral College. (Outside of the Midwest, the polls were reasonably accurate and even underestimated Clinton in some states.) By contrast, Republicans are facing a multi-front assault in the House this year:

In the Northeast, they have a lot of exposure in New York and New Jersey, which were once bastions of moderate Republicanism but which have become increasingly inhospitable to this brand of politics — and in Pennsylvania, where court-ordered redistricting resulted in a bad map for Republicans and where a lot of GOP incumbents have retired.
In the South, they face pressure because of demographic change in states such as Georgia and Virginia — and increasingly in Texas.
In the Midwest, there’s the risk of reversion to the mean with Trump off the ballot, especially as the GOP coalition in these states has come to rely on voters without a college degree who don’t always participate in midterm elections.
And in the West, there are 14 Republican-controlled seats in California and another four in Washington that look increasingly out of place as the Pacific Coast becomes a somewhat literal “blue wall.”

As it happens, projected Democratic gains are almost evenly distributed between the four Census Bureau regions: The Classic version of our model projects them to gain nine seats in the Midwest, nine in the South, nine in the Northeast and eight in the West. Note that Democrats could completely flop in any one of these regions and yet still (just barely) win enough seats to take the House.

Our forecast shows Democrats gaining House seats all over the country





Democratic-Held Seats


Census Region
Total Seats
Current
FORECASTED*
net gain




Northeast
78
51
60
9


Midwest
94
33
42
9


South
161
48
57
9


West
102
63
71
8

* Forecasts are derived from the Classic version of FiveThirtyEight’s House model as of Aug. 16.

Theme No. 5: Democrats need to win the popular vote by a fairly wide margin

The Classic version of our model gives Democrats a near certainty (about a 98 percent chance) of winning more votes than the GOP in the race for the House — but “only” a 3 in 4 chance of winning the majority of seats. This discrepancy between votes and seats reflects a combination of gerrymandering, voter self-sorting1 and incumbency, all of which favor Republicans to some degree. Thus, in the Classic version of our forecast, Democrats would need to win the popular vote by about 5 percentage points in order to become favorites to win the majority of seats in the House. And in the Lite and Deluxe versions, the break-even point is closer to a 6-point popular-vote win.

Nonetheless, these margins aren’t as bad for Democrats as they might be. At earlier points in the cycle, Democrats had appeared to need more like a 7- to 8-point advantage in the national popular vote to be favored to claim the majority of seats. Since then, the Republican edge has been eroded by retirements, by Pennsylvania’s redistricting and by the relatively weak GOP incumbency advantage (see Theme No. 3). All of this might seem like splitting hairs, but because so many indicators (see Theme No. 1) point toward Democrats winning the popular vote by a margin of something like 7 or 8 percentage points, these subtle differences are important.

I’ll be on vacation next week — excuse me, I’ll be investigating ground-level conditions in Maine’s 2nd Congressional District — but we’re going to be returning to these themes again and again between now and Nov. 6, so let’s call it a day. As a bonus, though, here’s a table put together by my colleague Julia Wolfe showing what our Classic forecast thinks of the race in every district in the country.

The odds for all 435 house races

According to the Classic version of FiveThirtyEight’s House model, as of 6 p.m. on Aug. 16, 2018

probabilities that……


district
Incumbent
A Democrat Wins
A Republican Wins




AK-1
Don Young
24.31%
75.69%


AL-1
Bradley Byrne
0.05
99.95


AL-2
Martha Roby
2.39
97.61


AL-3
Mike Rogers
0.18
99.82


AL-4
Robert B. Aderholt
0.00
>99%


AL-5
Mo Brooks
0.18
99.82


AL-6
Gary Palmer
0.02
99.98


AL-7
Terri A. Sewell
100.00
0.00


AR-1
Rick Crawford
0.16
99.83


AR-2
French Hill
22.65
77.34


AR-3
Steve Womack
0.08
99.92


AR-4
Bruce Westerman
0.13
99.87


AZ-1
Tom O’Halleran
94.16
5.83


AZ-2
Open seat
90.01
9.99


AZ-3
Raul Grijalva
99.97
0.03


AZ-4
Paul A. Gosar
0.06
99.94


AZ-5
Andy Biggs
0.20
99.80


AZ-6
David Schweikert
13.39
86.61


AZ-7
Ruben Gallego
99.98
0.00


AZ-8
Debbie Lesko
21.54
78.46


AZ-9
Open seat
98.82
1.18


CA-1
Doug LaMalfa
13.94
86.06


CA-2
Jared Huffman
>99%
0.00


CA-3
John Garamendi
99.87
0.13


CA-4
Tom McClintock
12.50
87.50


CA-5
Mike Thompson
99.95
0.00


CA-6
Doris O. Matsui
100.00
0.00


CA-7
Ami Bera
98.13
1.87


CA-8
Paul Cook
0.00
100.00


CA-9
Jerry McNerney
99.86
0.14


CA-10
Jeff Denham
70.77
29.23


CA-11
Mark DeSaulnier
>99%
0.00


CA-12
Nancy Pelosi
>99%
0.00


CA-13
Barbara Lee
99.99
0.00


CA-14
Jackie Speier
>99%
0.00


CA-15
Eric Swalwell
>99%
0.00


CA-16
Jim Costa
99.77
0.23


CA-17
Ro Khanna
>99%
0.00


CA-18
Anna G. Eshoo
>99%
0.00


CA-19
Zoe Lofgren
>99%
0.00


CA-20
Jimmy Panetta
99.88
0.00


CA-21
David Valadao
64.34
35.66


CA-22
Devin Nunes
2.23
97.77


CA-23
Kevin McCarthy
0.11
99.89


CA-24
Salud Carbajal
97.27
2.73


CA-25
Steve Knight
77.44
22.56


CA-26
Julia Brownley
99.82
0.18


CA-27
Judy Chu
100.00
0.00


CA-28
Adam Schiff
>99%
0.00


CA-29
Tony Cárdenas
>99%
0.00


CA-30
Brad Sherman
>99%
0.00


CA-31
Pete Aguilar
99.82
0.18


CA-32
Grace Napolitano
>99%
0.00


CA-33
Ted Lieu
>99%
0.00


CA-34
Jimmy Gomez
99.98
0.00


CA-35
Norma Torres
>99%
0.00


CA-36
Raul Ruiz
99.68
0.32


CA-37
Karen Bass
>99%
0.00


CA-38
Linda Sánchez
>99%
0.00


CA-39
Open seat
34.57
65.43


CA-40
Lucille Roybal-Allard
99.89
0.00


CA-41
Mark Takano
99.99
0.01


CA-42
Ken Calvert
2.25
97.75


CA-43
Maxine Waters
>99%
0.00


CA-44
Nanette Diaz Barragán
100.00
0.00


CA-45
Mimi Walters
58.04
41.96


CA-46
J. Luis Correa
>99%
0.00


CA-47
Alan Lowenthal
99.97
0.03


CA-48
Dana Rohrabacher
66.27
33.73


CA-49
Open seat
74.91
25.09


CA-50
Duncan D. Hunter
8.17
91.83


CA-51
Juan Vargas
>99%
0.00


CA-52
Scott Peters
99.80
0.20


CA-53
Susan Davis
99.97
0.03


CO-1
Diana DeGette
>99%
0.00


CO-2
Open seat
99.78
0.22


CO-3
Scott Tipton
40.54
59.46


CO-4
Ken Buck
3.63
96.37


CO-5
Doug Lamborn
2.79
97.21


CO-6
Mike Coffman
64.57
35.43


CO-7
Ed Perlmutter
99.69
0.31


CT-1
John B. Larson
99.97
0.02


CT-2
Joe Courtney
99.93
0.07


CT-3
Rosa L. DeLauro
99.98
0.02


CT-4
Jim Himes
99.84
0.16


CT-5
Open seat
96.33
3.66


DE-1
Lisa Blunt Rochester
98.77
1.23


FL-1
Matt Gaetz
0.01
99.99


FL-2
Neal Dunn
0.02
99.98


FL-3
Ted Yoho
1.18
98.82


FL-4
John Rutherford
0.03
99.97


FL-5
Al Lawson
99.99
0.01


FL-6
Open seat
28.17
71.83


FL-7
Stephanie Murphy
97.43
2.57


FL-8
Bill Posey
1.95
98.05


FL-9
Darren Soto
99.86
0.14


FL-10
Val Demings
100.00
0.00


FL-11
Daniel Webster
0.07
99.93


FL-12
Gus M. Bilirakis
1.56
98.43


FL-13
Charlie Crist
99.60
0.40


FL-14
Kathy Castor
100.00
0.00


FL-15
Open seat
27.57
72.43


FL-16
Vern Buchanan
11.59
88.41


FL-17
Open seat
0.37
99.63


FL-18
Brian Mast
7.58
92.42


FL-19
Francis Rooney
0.64
99.36


FL-20
Alcee L. Hastings
100.00
0.00


FL-21
Lois Frankel
100.00
0.00


FL-22
Ted Deutch
99.88
0.12


FL-23
Debbie Wasserman Schultz
99.69
0.31


FL-24
Frederica Wilson
100.00
0.00


FL-25
Mario Diaz-Balart
28.18
71.82


FL-26
Carlos Curbelo
37.89
62.11


FL-27
Open seat
97.11
2.89


GA-1
Buddy Carter
0.21
99.79


GA-2
Sanford D. Bishop Jr.
99.89
0.11


GA-3
A. Drew Ferguson
0.01
99.99


GA-4
Hank Johnson
>99%
0.00


GA-5
John Lewis
100.00
0.00


GA-6
Karen C. Handel
4.62
95.38


GA-7
Rob Woodall
29.40
70.60


GA-8
Austin Scott
0.00
100.00


GA-9
Doug Collins
0.00
>99%


GA-10
Jody Hice
0.01
99.99


GA-11
Barry Loudermilk
0.06
99.94


GA-12
Rick Allen
1.76
98.24


GA-13
David Scott
>99%
0.00


GA-14
Tom Graves
0.00
>99%


HI-1
Open seat
>99%
0.00


HI-2
Tulsi Gabbard
>99%
0.00


IA-1
Rod Blum
72.87
27.13


IA-2
David Loebsack
96.99
3.01


IA-3
David Young
66.97
33.03


IA-4
Steve King
17.55
82.44


ID-1
Open seat
0.04
99.96


ID-2
Mike Simpson
0.31
99.69


IL-1
Bobby L. Rush
>99%
0.00


IL-2
Robin Kelly
>99%
0.00


IL-3
Daniel Lipinski
99.98
0.02


IL-4
Open seat
>99%
0.00


IL-5
Mike Quigley
>99%
0.00


IL-6
Peter J. Roskam
25.34
74.66


IL-7
Danny K. Davis
>99%
0.00


IL-8
Raja Krishnamoorthi
99.89
0.11


IL-9
Jan Schakowsky
>99%
0.00


IL-10
Bradley Schneider
99.91
0.09


IL-11
Bill Foster
99.94
0.06


IL-12
Mike Bost
38.37
61.63


IL-13
Rodney Davis
32.50
67.50


IL-14
Randy Hultgren
35.98
64.02


IL-15
John Shimkus
0.03
99.97


IL-16
Adam Kinzinger
2.23
97.77


IL-17
Cheri Bustos
99.86
0.14


IL-18
Darin LaHood
0.02
99.98


IN-1
Peter Visclosky
99.98
0.02


IN-2
Jackie Walorski
7.46
92.54


IN-3
Jim Banks
0.26
99.74


IN-4
Open seat
0.72
99.28


IN-5
Susan W. Brooks
0.93
99.07


IN-6
Open seat
0.11
99.89


IN-7
André Carson
99.99
0.01


IN-8
Larry Bucshon
0.77
99.23


IN-9
Trey Hollingsworth
23.53
76.47


KS-1
Roger Marshall
0.00
>99%


KS-2
Open seat
39.00
60.99


KS-3
Kevin Yoder
21.21
78.78


KS-4
Ron Estes
12.73
87.27


KY-1
James Comer
0.01
99.99


KY-2
Brett S. Guthrie
0.12
99.88


KY-3
John A. Yarmuth
99.35
0.64


KY-4
Thomas Massie
0.06
99.94


KY-5
Harold Rogers
0.00
>99%


KY-6
Andy Barr
47.30
52.70


LA-1
Steve Scalise
0.00
>99%


LA-2
Cedric Richmond
100.00
0.00


LA-3
Clay Higgins
0.09
99.91


LA-4
Mike Johnson
0.20
99.80


LA-5
Ralph Abraham
0.07
99.93


LA-6
Garrett Graves
0.00
>99%


MA-1
Richard E. Neal
100.00
0.00


MA-2
James McGovern
99.98
0.02


MA-3
Open seat
99.90
0.10


MA-4
Joseph P. Kennedy III
100.00
0.00


MA-5
Katherine Clark
>99%
0.00


MA-6
Seth Moulton
99.97
0.03


MA-7
Michael E. Capuano
100.00
0.00


MA-8
Stephen F. Lynch
100.00
0.00


MA-9
William Keating
98.78
1.22


MD-1
Andy Harris
1.10
98.90


MD-2
C. A. Dutch Ruppersberger
99.98
0.02


MD-3
John P. Sarbanes
99.99
0.01


MD-4
Anthony Brown
>99%
0.00


MD-5
Steny H. Hoyer
99.99
0.01


MD-6
Open seat
98.83
1.17


MD-7
Elijah Cummings
>99%
0.00


MD-8
Jamie Raskin
>99%
0.00


ME-1
Chellie Pingree
99.44
0.56


ME-2
Bruce Poliquin
40.79
59.21


MI-1
Jack Bergman
18.99
81.01


MI-2
Bill Huizenga
5.88
94.12


MI-3
Justin Amash
1.92
98.08


MI-4
John Moolenaar
0.44
99.56


MI-5
Daniel Kildee
99.79
0.21


MI-6
Fred Upton
14.96
85.04


MI-7
Tim Walberg
38.06
61.94


MI-8
Mike Bishop
42.62
57.38


MI-9
Open seat
98.89
1.11


MI-10
Paul Mitchell
0.24
99.76


MI-11
Open seat
65.19
34.81


MI-12
Debbie Dingell
99.99
0.01


MI-13
Open seat
100.00
0.00


MI-14
Brenda Lawrence
>99%
0.00


MN-1
Open seat
45.34
54.66


MN-2
Jason Lewis
76.13
23.87


MN-3
Erik Paulsen
65.79
34.21


MN-4
Betty McCollum
99.98
0.02


MN-5
Open seat
>99%
0.00


MN-6
Tom Emmer
0.19
99.81


MN-7
Collin C. Peterson
85.48
14.52


MN-8
Open seat
35.44
64.56


MO-1
William “Lacy” Clay Jr.
>99%
0.00


MO-2
Ann Wagner
10.17
89.83


MO-3
Blaine Luetkemeyer
0.02
99.98


MO-4
Vicky Hartzler
0.08
99.92


MO-5
Emanuel Cleaver
99.78
0.22


MO-6
Sam Graves
0.06
99.94


MO-7
Billy Long
0.01
99.99


MO-8
Jason Smith
0.01
99.99


MS-1
Trent Kelly
0.07
99.93


MS-2
Bennie G. Thompson
99.53
0.00


MS-3
Open seat
0.27
99.73


MS-4
Steven Palazzo
0.45
99.55


MT-1
Greg Gianforte
12.18
87.81


NC-1
G.K. Butterfield
>99%
0.00


NC-2
George Holding
10.74
89.25


NC-3
Walter B. Jones
0.00
100.00


NC-4
David Price
>99%
0.00


NC-5
Virginia Foxx
5.19
94.81


NC-6
Mark Walker
21.51
78.49


NC-7
David Rouzer
7.18
92.82


NC-8
Richard Hudson
14.12
85.88


NC-9
Open seat
50.35
49.64


NC-10
Patrick T. McHenry
0.30
99.70


NC-11
Mark Meadows
0.34
99.66


NC-12
Alma Adams
>99%
0.00


NC-13
Ted Budd
37.21
62.79


ND-1
Open seat
1.34
98.66


NE-1
Jeff Fortenberry
0.15
99.85


NE-2
Don Bacon
58.20
41.80


NE-3
Adrian Smith
0.00
>99%


NH-1
Open seat
75.26
24.74


NH-2
Ann Kuster
98.55
1.45


NJ-1
Donald Norcross
99.98
0.02


NJ-2
Open seat
87.08
12.92


NJ-3
Tom MacArthur
43.91
56.09


NJ-4
Chris Smith
6.66
93.34


NJ-5
Josh Gottheimer
99.02
0.98


NJ-6
Frank Pallone Jr.
99.98
0.02


NJ-7
Leonard Lance
63.22
36.78


NJ-8
Albio Sires
>99%
0.00


NJ-9
Bill Pascrell Jr.
>99%
0.00


NJ-10
Donald Payne Jr.
>99%
0.00


NJ-11
Open seat
72.34
27.66


NJ-12
Bonnie Watson Coleman
>99%
0.00


NM-1
Open seat
97.92
2.07


NM-2
Open seat
24.28
75.72


NM-3
Ben R. Lujan
99.98
0.01


NV-1
Dina Titus
99.98
0.02


NV-2
Mark Amodei
1.31
98.69


NV-3
Open seat
66.86
33.14


NV-4
Open seat
78.55
21.45


NY-1
Lee Zeldin
10.95
89.05


NY-2
Pete King
18.79
81.21


NY-3
Thomas Suozzi
99.14
0.86


NY-4
Kathleen Rice
99.87
0.13


NY-5
Gregory W. Meeks
100.00
0.00


NY-6
Grace Meng
99.97
0.00


NY-7
Nydia M. Velázquez
99.99
0.00


NY-8
Hakeem Jeffries
100.00
0.00


NY-9
Yvette D. Clarke
>99%
0.00


NY-10
Jerrold Nadler
>99%
0.00


NY-11
Daniel Donovan
24.32
75.68


NY-12
Carolyn Maloney
>99%
0.00


NY-13
Adriano Espaillat
>99%
0.00


NY-14
Joseph Crowley
>99%
0.00


NY-15
José E. Serrano
>99%
0.00


NY-16
Eliot Engel
99.98
0.00


NY-17
Nita Lowey
99.95
0.00


NY-18
Sean Patrick Maloney
98.26
1.74


NY-19
John Faso
52.13
47.87


NY-20
Paul D. Tonko
99.98
0.02


NY-21
Elise Stefanik
7.38
92.62


NY-22
Claudia Tenney
71.17
28.83


NY-23
Tom Reed
4.86
95.14


NY-24
John Katko
36.66
63.34


NY-25
Open seat
99.72
0.28


NY-26
Brian Higgins
>99%
0.00


NY-27
Open seat
24.50
75.49


OH-1
Steve Chabot
58.23
41.77


OH-2
Brad Wenstrup
5.50
94.50


OH-3
Joyce Beatty
>99%
0.00


OH-4
Jim Jordan
4.48
95.52


OH-5
Robert E. Latta
0.31
99.68


OH-6
Bill Johnson
0.04
99.96


OH-7
Bob Gibbs
11.38
88.62


OH-8
Warren Davidson
0.06
99.94


OH-9
Marcy Kaptur
99.97
0.02


OH-10
Michael Turner
6.42
93.58


OH-11
Marcia L. Fudge
>99%
0.00


OH-12
Troy Balderson
48.95
51.05


OH-13
Tim Ryan
99.95
0.05


OH-14
David Joyce
18.93
81.07


OH-15
Steve Stivers
2.63
97.37


OH-16
Open seat
4.99
95.01


OK-1
Open seat
0.75
99.25


OK-2
Markwayne Mullin
0.01
99.99


OK-3
Frank Lucas
0.00
>99%


OK-4
Tom Cole
0.02
99.98


OK-5
Steve Russell
23.67
76.33


OR-1
Suzanne Bonamici
99.96
0.04


OR-2
Greg Walden
0.26
99.73


OR-3
Earl Blumenauer
100.00
0.00


OR-4
Peter DeFazio
99.17
0.83


OR-5
Kurt Schrader
99.61
0.39


PA-1
Brian Fitzpatrick
21.52
78.48


PA-2
Brendan Boyle
>99%
0.00


PA-3
Dwight Evans
>99%
0.00


PA-4
Open seat
99.63
0.37


PA-5
Open seat
99.98
0.02


PA-6
Open seat
97.60
2.40


PA-7
Open seat
86.77
13.22


PA-8
Matt Cartwright
93.52
6.48


PA-9
Open seat
0.90
99.10


PA-10
Scott Perry
14.85
85.15


PA-11
Lloyd Smucker
27.28
72.72


PA-12
Tom Marino
0.18
99.82


PA-13
Open seat
0.01
99.99


PA-14
Open seat
1.23
98.77


PA-15
Glenn W. Thompson
0.08
99.92


PA-16
Mike Kelly
4.76
95.24


PA-17
Conor Lamb
92.96
7.04


PA-18
Mike Doyle
100.00
0.00


RI-1
David Cicilline
99.98
0.02


RI-2
Jim Langevin
99.97
0.03


SC-1
Open seat
15.00
85.00


SC-2
Joe Wilson
0.85
99.15


SC-3
Jeff Duncan
0.15
99.85


SC-4
Open seat
0.05
99.95


SC-5
Ralph Norman
12.36
87.64


SC-6
James E. Clyburn
>99%
0.00


SC-7
Tom Rice
1.06
98.94


SD-1
Open seat
1.44
98.56


TN-1
Phil Roe
0.00
>99%


TN-2
Open seat
0.03
99.97


TN-3
Chuck Fleischmann
0.13
99.87


TN-4
Scott DesJarlais
0.76
99.24


TN-5
Jim Cooper
99.97
0.03


TN-6
Open seat
0.00
>99%


TN-7
Open seat
0.03
99.97


TN-8
David Kustoff
0.02
99.98


TN-9
Steve Cohen
>99%
0.00


TX-1
Louie Gohmert
0.00
>99%


TX-2
Open seat
7.84
92.16


TX-3
Open seat
0.77
99.22


TX-4
John Ratcliffe
0.00
>99%


TX-5
Open seat
0.08
99.92


TX-6
Open seat
6.64
93.36


TX-7
John Culberson
49.32
50.68


TX-8
Kevin Brady
0.00
>99%


TX-9
Al Green
100.00
0.00


TX-10
Michael T. McCaul
2.62
97.38


TX-11
K. Michael Conaway
0.00
>99%


TX-12
Kay Granger
0.01
99.99


TX-13
Mac Thornberry
0.00
>99%


TX-14
Randy Weber
1.75
98.25


TX-15
Vicente Gonzalez
99.90
0.10


TX-16
Open seat
99.98
0.02


TX-17
Bill Flores
0.52
99.48


TX-18
Sheila Jackson Lee
>99%
0.00


TX-19
Jodey Arrington
0.00
>99%


TX-20
Joaquin Castro
99.94
0.00


TX-21
Open seat
17.81
82.17


TX-22
Pete Olson
14.17
85.83


TX-23
Will Hurd
72.36
27.64


TX-24
Kenny Marchant
4.37
95.62


TX-25
Roger Williams
7.09
92.91


TX-26
Michael Burgess
0.12
99.88


TX-27
Michael Cloud
0.48
99.52


TX-28
Henry Cuellar
99.95
0.00


TX-29
Open seat
99.99
0.01


TX-30
Eddie Bernice Johnson
99.99
0.00


TX-31
John Carter
20.11
79.88


TX-32
Pete Sessions
11.74
88.25


TX-33
Marc Veasey
99.99
0.01


TX-34
Filemon Vela
99.95
0.05


TX-35
Lloyd Doggett
99.99
0.01


TX-36
Brian Babin
0.02
99.98


UT-1
Rob Bishop
0.03
99.97


UT-2
Chris Stewart
2.53
97.43


UT-3
John R. Curtis
0.03
99.97


UT-4
Mia Love
19.01
80.99


VA-1
Robert J. Wittman
1.50
98.50


VA-2
Scott Taylor
8.14
91.86


VA-3
Robert C. Scott
100.00
0.00


VA-4
A. Donald McEachin
99.86
0.14


VA-5
Open seat
48.78
51.22


VA-6
Open seat
0.34
99.66


VA-7
Dave Brat
32.08
67.92


VA-8
Don Beyer
>99%
0.00


VA-9
Morgan Griffith
0.16
99.84


VA-10
Barbara Comstock
74.71
25.29


VA-11
Gerald E. Connolly
99.99
0.01


VT-1
Peter Welch
99.98
0.02


WA-1
Suzan DelBene
99.87
0.13


WA-2
Rick Larsen
99.88
0.00


WA-3
Jaime Herrera Beutler
10.00
90.00


WA-4
Dan Newhouse
0.80
99.20


WA-5
Cathy McMorris Rodgers
28.14
71.86


WA-6
Derek Kilmer
99.93
0.07


WA-7
Pramila Jayapal
>99%
0.00


WA-8
Open seat
44.36
55.64


WA-9
Adam Smith
100.00
0.00


WA-10
Denny Heck
99.90
0.10


WI-1
Open seat
20.76
79.24


WI-2
Mark Pocan
100.00
0.00


WI-3
Ron Kind
99.27
0.73


WI-4
Gwen Moore
>99%
0.00


WI-5
F. James Sensenbrenner
2.01
97.99


WI-6
Glenn Grothman
29.23
70.77


WI-7
Sean P. Duffy
0.89
99.11


WI-8
Mike Gallagher
0.76
99.24


WV-1
David McKinley
0.30
99.70


WV-2
Alex Mooney
3.65
96.35


WV-3
Open seat
6.03
93.97


WY-1
Liz Cheney
0.02
99.98

Odds may not sum to 100 percent due to rounding and third-party candidates.

CORRECTION (Aug. 17, 2018, 5:45 p.m.): The first name of Rep. Mark Pocan of Wisconsin was misspelled in an earlier version of the table in this article.

Model Talk: Let’s Talk About The 2018 House Forecast

By Nate Silver and Galen Druke and Nate Silver and Galen Druke












 












More: Apple Podcasts |
ESPN App |
RSS
| Embed


Embed Code

Our 2018 House forecast is now live! That also means that the FiveThirtyEight Politics podcast is resurrecting “Model Talk,” in which Nate Silver answers questions about what goes into the forecast and how it’s reacting to new developments. This is the inaugural episode of the 2018 midterm season.

You can listen to the episode by clicking the “play” button above or by downloading it in iTunes , the ESPN App or your favorite podcast platform. If you are new to podcasts, learn how to listen .

The FiveThirtyEight Politics podcast publishes Monday evenings, with occasional special episodes throughout the week. Help new listeners discover the show by leaving us a rating and review on iTunes . Have a comment, question or suggestion for “good polling vs. bad polling”? Get in touch by email, on Twitter or in the comments.

Politics Podcast: Let’s Talk About The 2018 House Forecast

By Nate Silver and Galen Druke and Nate Silver and Galen Druke












 












More: Apple Podcasts |
ESPN App |
RSS
| Embed


Embed Code

Our 2018 House forecast is now live! That also means that the FiveThirtyEight Politics podcast is resurrecting “Model Talk,” in which Nate Silver answers questions about what goes into the forecast and how it’s reacting to new developments. This is the inaugural episode of the 2018 midterm season.

You can listen to the episode by clicking the “play” button above or by downloading it in iTunes , the ESPN App or your favorite podcast platform. If you are new to podcasts, learn how to listen .

The FiveThirtyEight Politics podcast publishes Monday evenings, with occasional special episodes throughout the week. Help new listeners discover the show by leaving us a rating and review on iTunes . Have a comment, question or suggestion for “good polling vs. bad polling”? Get in touch by email, on Twitter or in the comments.

How FiveThirtyEight’s House Model Works

We’ve been publishing election models for more than 10 years now, and FiveThirtyEight’s 2018 House model is probably the most work we’ve ever put into one of them. That’s mostly because it just uses a lot of data. We collected data for all 435 congressional districts in every House race since 1998, and we’ve left few stones unturned, researching everything from how changes in district boundary lines could affect incumbents in Pennsylvania to how ranked-choice voting could change outcomes in Maine.

Not all of that detail is apparent upon launch. You can see the topline national numbers, as well as a forecast of the outcome in each district. But we’ll be adding a lot more features within the next few weeks, including detailed pages for each district. You may want to clip and save this methodology guide for then. In the meantime, here’s a fairly detailed glimpse at how the model works.

Overview

The principles behind the House forecast should be familiar to FiveThirtyEight readers. It takes lots of polls, performs various types of adjustments to them, and then blends them with other kinds of empirically useful indicators (what we sometimes call “the fundamentals”) to forecast each race. Then it accounts for the uncertainty in the forecast and simulates the election thousands of times. Our models are probabilistic in nature; we do a lot of thinking about these probabilities,and the goal is to develop probabilistic estimates that hold up well under real-world conditions. For instance, Democrats’ chances of winning the House are between 7 in 10 and 3 in 4 in the various versions of the model upon launch — right about what Hillary Clinton’s chances were on election night two years ago! — so ignore those probabilities at your peril.

Nonetheless, if you’re used to the taste of our presidential forecasts, the House model will have a different flavor to it in two important respects:

As compared with the presidential model, the House model is less polling-centric. Instead, it uses a broader consensus of indicators. That’s partly out of necessity: House districts don’t get much polling, and the polling they do get often isn’t much good. It’s also partly out of opportunity: With 435 separate races every other year, it’s possible to make fairly robust empirical assessments of which factors really predict House races well and which don’t.
House races are far more localized than presidential races, and this is reflected in the design of the model. In presidential elections, outcomes are extremely correlated from state to state. It wasn’t a surprise that President Trump won Michigan given that he also won demographically similar states such as Wisconsin and Pennsylvania, for instance. Sometimes that sort of thing happens in congressional elections too; although Democrats are favored in our initial forecast, even a relatively minor polling error could tilt the race back toward Republicans. Nonetheless, about three-quarters of the uncertainty in the House forecasts comes from local, district-by-district factors. If the presidential model is laser-focused on how the polls are changing from day to day and what they say about the Electoral College, the House model’s approach is more diffuse, with the goal being to shine some light into the darker corners of the electoral landscape.

Three versions of the model: Lite, Classic, Deluxe

In 2016, we published what we described as two different election models: “polls-only” and “polls-plus.”1 This year, we’re running what we think of as three different versions of the same model, which we call Lite, Classic and Deluxe. I realize that’s a subtle distinction — different models versus different versions of the same model.

But the Lite, Classic and Deluxe versions of the House model somewhat literally build on top of one another, like different layers of toppings on an increasingly fancy burger. I’ll describe these methods in more detail in the sections below. First, a high-level overview of what the different versions account for.

The layers in FiveThirtyEight’s House forecast





Which versions use it?



Layer
Description
Lite
Classic
Deluxe




1a
Polling
District-by-district polling, adjusted for house effects and other factors.
✓
✓
✓


1b
CANTOR
A system which infers results for districts with little or no polling from comparable districts that do have polling.
✓
✓
✓


2
Fundamentals
Non-polling factors such as fundraising and past election results that historically help in predicting congressional races.

✓
✓


3
Expert forecasts
Ratings of each race published by the Cook Political Report, Inside Elections and Sabato’s Crystal Ball


✓

Lite is as close as you get to a “polls-only” version of the forecast — except, the problem is that a lot of congressional districts have little or no polling. So it uses a system we created called CANTOR2 to fill in the blanks. It uses polls from districts that have polling, as well as national generic ballot polls, to infer what the polls would say in districts that don’t have polling.

The Classic version also uses local polls3 but layers a bunch of non-polling factors on top of it, the most important of which are incumbency, past voting history in the district, fundraising and the generic ballot. These are the “fundamentals.” The more polling in a district, the more heavily Classic relies on the polls as opposed to the “fundamentals.” Although Lite isn’t quite as simple as it sounds, the Classic model is definitely toward the complex side of the spectrum. With that said, it should theoretically increase accuracy. In the training data,4 Classic miscalled 3.3 percent of races, compared with 3.8 percent for Lite.5 You should think of Classic as the preferred or default version of FiveThirtyEight’s forecast unless we otherwise specify.

Finally, there’s the Deluxe flavor of the model, which takes everything in Classic and sprinkles in one more key ingredient: expert ratings. Specifically, Deluxe uses the race ratings from the Cook Political Report, Nathan Gonzales’s Inside Elections and Sabato’s Crystal Ball, all of which have published forecasts for many years and have an impressive track record of accuracy.6

Within-sample accuracy of forecasting methods

Share of races called correctly based on House elections from 1998 to 2016

Forecast
100 Days Before Election
Election Day




Lite model (poll-driven)
94.2%
96.2%


Fundamentals alone
95.4
95.7


Classic model (Lite model fundamentals)
95.4
96.7


Expert ratings alone*
94.8
96.6


Deluxe model (Classic model expert ratings)
95.7
96.9

* Based on the average ratings from Cook Political Report, Inside Elections/The Rothenberg Political Report, Sabato’s Crystal Ball and CQ Politics. Where the expert rating averages out to an exact toss-up, the experts are given credit for half a win.

So if we expect the Deluxe forecast to be (slightly) more accurate, why do we consider Classic to be our preferred version, as I described above? Basically, because we think it’s kind of cheating to borrow other people’s forecasts and make them part of our own. Some of the fun of doing this is in seeing how our rigid but rigorous algorithm stacks up against more open-ended but subjective ways of forecasting the races. If our lives depended on calling the maximum number of races correctly, however, we’d go with Deluxe.

Collecting, weighting and adjusting polls

Our House forecasts use almost all the polls we can find, including partisan polls put out by campaigns or other interested parties. (We have not traditionally used partisan polls in our Senate or presidential forecasts, but they are a necessary evil for the House.) However, as polling has gotten more complex, including attempts to create fake polls, there are an increasing number of exceptions:

We don’t use polls if we have significant concerns about their veracity or if the pollster is known to have faked polls before.
We don’t use DIY polls commissioned by nonprofessional hobbyists on online platforms such as Google Surveys. (This is a change in policy since 2016. Professional or campaign polls using these platforms are still fine.)
We don’t treat subsamples of multistate polls as individual “polls” unless certain conditions are met.7
We don’t use “polls” that blend or smooth their data using methods such as MRP. These can be perfectly fine techniques — but if you implement them, you’re really running a model rather than a poll. We want to do the blending and smoothing ourselves rather than inputting other people’s models into ours.

These cases are rare — so if you don’t see a poll on our “latest polls” page, there’s a good chance that we’ve simply missed it. (House polls can be a lot harder to track down than presidential ones.) Please drop us a line if there’s a poll you think we’ve missed.

Polls are weighted based on their sample size, their recency and their pollster rating (which in turn is based on the past accuracy of the pollster, as well as its methodology). These weights are determined by algorithm; we aren’t sticking our fingers in the wind and rating polls on a case-by-case basis. In a slight change this year, the algorithm emphasizes the diversity of polls more than it has in the past; in any particular race, it will insist on constructing an average of polls from at least two or three distinct polling firms even if some of the polls are less recent.

There are also three types of adjustments to each poll:

First, a likely voter adjustment takes the results of polls of registered voters or all adults and attempts to translate them to a likely-voter basis. Traditionally, Republican candidates gain ground in likely voter polls relative to registered voter ones, but the gains are smaller in midterms with a Republican president. Furthermore, some polls this year actually show Democrats gaining in likely voter models. The likely voter adjustment is dynamic; it starts with a prior that likely voter polls slightly help Republicans, but this prior is updated as pollsters publish polls that directly compare likely and registered voter results. (If you’re a pollster, please follow Monmouth University’s lead and do this!) In mid-August, for example, the model treats likely-voter and registered-voter polls as roughly equivalent to each other, but this could change as we collect more data.
Second, a timeline adjustment adjusts for the timing of the poll, based on changes in the generic congressional ballot. For instance, if Democrats have gained a net of 5 percentage points on the generic ballot since a certain district was polled, the model will adjust the poll upward toward the Democratic candidate (but not by the full 5 points; instead, by roughly half that amount — 2.5 points — depending on the elasticity score8 of the district). As compared with the timeline adjustment in our presidential model, which can be notoriously aggressive, the one in the House model is pretty conservative.9
A house effects adjustment corrects for persistent statistical bias from a pollster. For instance, if a polling firm consistently shows results that are 2 points more favorable for Democrats than other polls of the same district, the adjustment will shift the poll part of the way back toward Republicans.10

The House model does use partisan and campaign polls, which typically make up something like half of the overall sample of congressional district polling. Partisanship is determined by who sponsors the poll, rather than who conducts it. Polls are considered partisan if they’re conducted on behalf of a candidate, party, campaign committee, or PAC, super PAC, 501(c)(4), 501(c)(5) or 501(c)(6) organization that conducts a large majority of its political activity on behalf of one political party.

Partisan polls are subject to somewhat different treatment than nonpartisan polls in the model. They receive a lower weight, as partisan-sponsored polls are historically less accurate. And the house effects adjustment starts out with a prior that assumes these polls are biased by about 4 percentage points toward their preferred candidate or party. If a pollster publishing ostensibly partisan polls consistently has results that are similar to nonpartisan polls of the same districts, the prior will eventually be overridden.

CANTOR: Analysis of polls in similar districts

CANTOR is essentially PECOTA or CARMELO (the baseball and basketball player forecasting systems we designed) for congressional districts. It uses a k-nearest neighbors algorithm to identify similar congressional districts based on a variety of demographic,11 geographic12 and political13 factors. For instance, the district where I was born, Michigan 8, is most comparable to other upper-middle-income Midwestern districts such as Ohio 12, Indiana 5 and Minnesota 2 that similarly contain a sprawling mix of suburbs, exurbs and small towns.

The goal of CANTOR is to impute what polling would say in unpolled or lightly polled districts, given what it says in similar districts. It attempts to accomplish this goal in two stages. First, it comes up with an initial guesstimate of what the polls would say based solely on FiveThirtyEight’s partisan lean metric (FiveThirtyEight’s version of a partisan voting index, which is compiled based on voting for president and state legislature) and incumbency. For instance, if Republican incumbents are polling poorly in the districts where we have polling, it will assume that Republican incumbents in unpolled districts are vulnerable as well. Then, it adjusts the initial estimate based on the district-by-district similarity scores. For instance, that Republican incumbent Carlos Curbelo is polling surprisingly well in Florida’s 26th Congressional District will also help Republicans in similar congressional districts.

All of this sounds pretty cool, but there’s one big drawback. Namely, there’s a lot of selection bias in which districts are polled. A House district usually gets surveyed only if one of the campaigns or a media organization has reason to think the race is close — so unpolled districts are less competitive than you’d infer from demographically similar districts that do have polls. CANTOR projections are adjusted to account for this.

Overall, CANTOR is an interesting method that heavily leans into district polling and gets as close as possible to a “polls-only” view of the race. However, in terms of accuracy, it is generally inferior to using …

The fundamentals

The data-rich environment in House elections — 435 individual races every other year, compared with just one race every four years for the presidency — is most beneficial when it comes to identifying reliable non-polling factors for forecasting races. There’s enough data, in fact, that rather than using all districts to determine which factors were most predictive, I instead focused the analysis on competitive races (using a fairly broad definition of “competitive”). In competitive districts with incumbents, the following factors have historically best predicted election results, in roughly declining order of importance:

The incumbent’s margin of victory in his or her previous election, adjusted for the national political environment and whom the candidate was running against in the prior election.
The generic congressional ballot.
Fundraising, based on the share of individual contributions for the incumbent and the challenger as of the most recent filing period.14
FiveThirtyEight partisan lean, which is based on how a district voted in the past two presidential elections and (in a new twist this year) state legislative elections. In our partisan lean formula, 50 percent of the weight is given to the 2016 presidential elections, 25 percent to the 2012 presidential election and 25 percent to state legislative elections.
Congressional approval ratings, which are a measure of the overall attitude toward incumbents.15
Whether either the incumbent or the challenger was involved in a scandal.
The incumbent’s roll call voting record — specifically, how often the incumbent voted with his or her party in the past three Congresses. “Maverick-y” incumbents who break party ranks more often outperform those who don’t.
Finally, the political experience level of the challenger, based on whether the challenger has held elected office before.

In addition, in Pennsylvania, which underwent redistricting in 2018, the model accounts for the degree of population overlap between the incumbent’s old and new district. And in California and Washington state, it accounts for the results of those states’ top-two primaries.

In open-seat races, the model uses the factors from the list above that aren’t dependant on incumbency, namely the generic ballot, fundraising, FiveThirtyEight partisan lean, scandals, experience and (where applicable) top-two primary results. It also uses the results of the previous congressional election in the district, but this is a much less reliable indicator than when an incumbent is running for re-election.

But wait — there’s more! In addition to combining polls and fundamentals, the Classic model compares its current estimate of the national political climate to a prior based on the results of congressional elections since 1946, accounting for historic swings in midterms years and presidential approval ratings. This prior has little effect on the projections this year, however, as it implies that Democrats should be ahead by about 8 points in the popular vote — similar to what the generic ballot and other indicators show.16 To put it another way, the results we’re seeing in the data so far are consistent with what’s usually happened in midterms under unpopular presidents.

Incorporating expert ratings

Compared with the other steps, incorporating expert ratings and creating the Deluxe version of the model is fairly straightforward. We have a comprehensive database of ratings from Cook and other groups since 1998, so we can look up how a given rating corresponded, on average, with a margin of victory. For instance, candidates who were deemed to be “likely” winners in their races won by an average of about 12 points:

What do ratings like “lean Republican” really mean?




Expert Rating
Average margin of victory




Toss-up
0 points


“Tilts” toward candidate
4 points


“Leans” toward candidate
7 points


“Likely” for candidate
12 points


“Solid” or “safe” for candidate
34 points

Based on House races since 1998.

But, of course, there are complications. One is that there’s an awful lot of territory covered by the “solid” and “safe” categories: everything from races that could almost be considered competitive to others where the incumbent wins every year by a 70-point margin. Therefore, the Deluxe forecast doesn’t adjust its projections much when it encounters “solid” or “safe” ratings from the experts, except in cases where the rating comes as a surprise because other factors indicate that the race should be competitive.

Also, although the expert raters are really quite outstanding at identifying “micro” conditions on the ground, including factors that might otherwise be hard to measure, they tend to be lagging indicators of the macro political environment. Several of the expert raters shifted their projections sharply toward the Democrats throughout 2018, for instance, even though the generic ballot has been fairly steady over that period.17 Thus, the Deluxe forecast tries to blend the relative order of races implied by the expert ratings with the Classic model’s data-driven estimate of national political conditions. Deluxe and Classic will usually produce relatively similar forecasts of the overall number of seats gained or lost by a party, therefore, even though they may have sharp disagreements on individual races.

Simulating the election and accounting for uncertainty

Sometimes what seem like incredibly pedantic questions can turn out to be important. For years, we’ve tried to design models that account for the complicated, correlated structure of error and uncertainty in election forecasting. Specifically, that if a candidate or a party overperforms the polls in one swing state, they’re also likely to do so in other states, especially if they’re demographically similar. Understanding this principle was key to understanding why Clinton’s lead wasn’t nearly as safe as it seemed in 2016.

Fortunately, this is less of a problem in constructing a House forecast; there are different candidates on the ballot in every district, instead of just one presidential race, and the model relies on a variety of inputs, instead of depending so heavily on polls. Nonetheless, the model accounts for four potential types of error in an attempt to self-diagnose the various ways in which it could go off the rails:

First, there’s local error — that is, error pertaining to individual districts. Forecasts are more error-prone in districts where there’s less polling or in districts where the various indicators disagree with one another. (In West Virginia 3, for example, the fundamentals regression thinks Democrat Richard Ojeda should be a huge underdog — but the only poll of the race has him ahead!) Some districts are also swingier (or more elastic) than others; conditions tend to change fairly quickly in New Hampshire, for instance, but more slowly in the South, where electorates are often bifurcated between very liberal and very conservative voters.
Second, there’s error based on regional or demographic characteristics. For instance, it’s possible that Democrats will systematically underperform expectations in districts with large numbers of Hispanic voters or overperform them in the rural Midwest. The model uses CANTOR similarity scores to simulate these possibilities.
Third, there can be error driven by incumbency status. In some past elections, polls have systematically underestimated Republican incumbents, for example, even if they were fairly accurate in open-seat races. The model accounts for this possibility as well.
Fourth and finally, the model accounts for the possibility of a uniform national swing — i.e., when the polls are systematically off in one party’s direction in almost every race.

Error becomes smaller as Election Day approaches. In particular, there’s less possibility of a sharp national swing as you get nearer to the election because there’s less time for news events to intervene.

Nonetheless, you shouldn’t expect pinpoint precision in a House forecast, and models that purport to provide it are either fooling you or fooling themselves. Even if you knew exactly what national conditions were, there would still be a lot of uncertainty based on how individual races play out.

Related:












Odds and ends

OK, that’s almost everything. Just a few final notes:

We’ve made educated guesses about the identity of the nominees in states that haven’t yet held their primaries. We’ll definitely be wrong about a few of these, and we’ll change them once the primaries are held. For the time being, we’re also assuming that incumbent Republican Chris Collins will successfully be able to withdraw from the race in New York’s 27th District, but we’ll reinsert him if it looks like he won’t be able to.
Once we publish the race-by-race pages, you’ll notice that we also project turnout in each district, based on factors such as the citizen voting-age population and turnout in past midterms and presidential races. This is important in understanding the relationship between the national popular vote and the number of seats that each party might gain or lose. As of forecast launch, our model implies that Democrats would need to win the House popular vote by 5 to 6 percentage points to have a break-even chance of winning a majority of seats.
Finally, I should emphasize that we do not make ad-hoc adjustments to the forecasts in individual races. They’re all done strictly by algorithm. Nor do we implement major changes in the program once the model has been released. With that said, we will correct bugs, especially in the first week or two after the model is out.18 So if you see something that looks awry, please just let us know.

2018 House Forecast

How do you like your House forecast?

Lite

Keep it simple, please — give me the best forecast you can based on what local and national polls say

Classic

I’ll take the polls, plus all the “fundamentals”: fundraising, past voting in the district, historical trends and more

Deluxe

Gimme the works — the Classic forecasts plus experts’ ratings

Forecasting the race for the House

Published Aug. 16, 2018 at 11:00 AM

7 in 10

Chance Democrats win control (70.4%)

3 in 10

Chance Republicans keep control (29.6%)

AVERAGE

MEDIAN

CURRENTBREAKDOWN

CURRENTBREAKDOWN

Breakdown of seats byparty

↑

Higher

probability

267 D168 R

247 D188 R

227 D208 R

227 R208 D

247 R188 D

80% chance Democrats gain 13 to 55 seats

80% chance Democrats gain 13 to 55 seats

10% chance Democrats gain fewer than 13 seats

10% chance Democrats gain fewer than 13 seats

10% chance Democrats gain more than 55 seats

10% chance Democrats gain more than 55 seats

+55

+32 Democratic seats

AVG. GAIN

+13

Our forecast for every district

The chance of each candidate winning and projected vote share in all 435 House districts

Cartogram

Map

Solid D

≥95% D

Likely D

≥75% D

Lean D

≥60% D

Toss-up

50% nonincumbent party

= one district

District totals by category

189

8

19

10

19

53

137

MAJORITY

Chance of controlling the House

1 in 10

1 in 4

1 in 2

3 in 4

9 in 10

70.4%

70.4%

29.6%

29.6%

NOV. 6

ELECTION DAY

TODAY

AUG. 14, 2018

Seats controlled by each party

157-278

187-248

EVEN

247-188

277-158

227-208

227-208

Popular vote margin

Sept.

Oct.

Nov.

R+10

R+5

0

D+5

D+10

D+7.2

D+7.2

KEY

AVERAGE

80% CHANCE OF FALLING IN RANGE

How the House has swung historically

The projected swing of our current forecast along with the swing of every House election since 1924

Net advantage

Swing

0

100

200

300 seats

0

100

200

300

1926

1930

1934

1938

1942

1946

1950

1954

1958

1962

1966

1970

1974

1978

1982

1986

1990

1994

1998

2002

2006

2010

2014

2018

◄ More Democratic | More Republican ►

COOLIDGE

COOLIDGE

HOOVER

HOOVER

ROOSEVELT

ROOSEVELT

TRUMAN

TRUMAN

EISENHOWER

EISENHOWER

KENNEDY

KENNEDY

JOHNSON

JOHNSON

NIXON

NIXON

FORD

FORD

CARTER

CARTER

REAGAN

REAGAN

H.W. BUSH

H.W. BUSH

CLINTON

CLINTON

W. BUSH

W. BUSH

OBAMA

OBAMA

TRUMP

TRUMP

Most likely outcome

Most likely outcome

How this forecast works

Nate Silver explains the methodology behind our 2018 midterms forecast. Read more …

* For races in which the general election candidates haven’t yet been determined, we’re showing a leading primary candidate until a nominee is picked. Vacant seats are assigned to the party that previously held them for the purposes of seat totals and flips.

Forecast models by Nate Silver. Design and development by Aaron Bycoffe, Rachael Dottle, Ritchie King, Ella Koeze, Andrei Scheinkman, Gus Wezerek and Julia Wolfe. Research by Andrea Jones-Rooy, Dhrumil Mehta, Mai Nguyen and Nathaniel Rakich. Notice any bugs? Send us an email.

How Competitive Would New York 27 Be With Chris Collins On The Ballot?

UPDATE (Aug. 16 6:39 a.m.): GOP Rep. Chris Collins is trying to get his name removed from the ballot in New York’s 27th Congressional District — an effort that will likely be decided by a legal fight. The district is much more likely to stay red if Republicans can manage to replace Collins: A new survey of likely voters in NY-27 shows several potential GOP candidates with double-digit leads. With Collins on the ballot, however, the district could become competitive, as Nate details below.

In the era of President Trump, it’s become fashionable to presume that politicians can do whatever they like and get away with it. But if recent elections to Congress are any guide, scandals do have large and measurable effects. So when U.S. Rep. Chris Collins, the Republican from New York’s 27th Congressional District, was arrested on insider trading charges on Wednesday morning, it took a seat that had looked to be fairly safe for Republicans and put it into the competitive category.

I’m going to be fairly circumspect in this article because I’m knee-deep in finalizing our House model, and I don’t want to scoop our own forecast. But one of the things we evaluated in designing that model is the electoral effects of scandals, based on the data set of scandals put together by my colleague Nathaniel Rakich.19

Below is a list of scandal-plagued incumbents since 1998 who made it to the general election and faced an opponent from the opposite party.20 I’ve compared each incumbent’s actual margin of victory or defeat against a projected margin based on a “fundamentals” model that accounts for: (i) the incumbent’s previous victory margin,21 (ii) the partisan lean of the district,22 (iii) the generic ballot at the time of the election, (iv) congressional approval ratings at the time of the election (which are a good proxy for the overall mood toward incumbents), and (v) the incumbent’s congressional voting record (representatives who break with their party more often overperform on Election Day). This is a slightly pared-down version of what our House model will look at, but it should be a fairly robust and reliable model.23

How much do scandals hurt incumbents?

It depends on how competitive the district is

Year
District
Incumbent
Projected Margin Of Victory
Actual Margin of Victory or Defeat
Net Effect Of Scandal




1998
GA-6
Newt Gingrich
31.2
41.4
10.2


1998
ID-1
Helen Chenoweth
23.4
10.5
-12.9


1998
IL-6
Henry J. Hyde
32.6
37.2
4.7


1998
IN-6
Dan Burton
51.4
55.3
3.9


2000
GA-7
Bob Barr
21.0
10.5
-10.5


2004
OH-14
Steven C. LaTourette
34.0
25.5
-8.5


2006
MI-14
John Conyers, Jr.
78.1
70.6
-7.5


2006
PA-10
Donald Sherwood
25.1
-5.9
-31.0


2008
FL-16
Tim Mahoney
7.6
-20.2
-27.8


2008
NY-15
Charles B. Rangel
82.4
81.3
-1.2


2010
MA-6
John F. Tierney
22.3
13.9
-8.4


2012
FL-26
David Rivera
12.5
-10.6
-23.2


2012
NY-11
Michael G. Grimm
10.3
5.4
-4.9


2012
TN-4
Scott DesJarlais
24.8
11.5
-13.3


2016
NC-9
Robert Pittenger
27.2
16.4
-10.9


2016
NH-1
Frank C. Guinta
5.4
-1.3
-6.8


2016
TX-27
Blake Farenthold
28.5
23.4
-5.1




Overall average
30.5
21.5
-9.0




Districts less competitive than NY-27
45.7
42.0
-3.7




Districts more competitive than NY-27
19.8
7.1
-12.7

Shaded districts were more competitive than NY-27 based on their partisan lean.

On average, the scandal-ridden incumbents … won re-election by 21.5 percentage points! But that’s quite a bit worse than their projected margin of victory, which was 30.5 percentage points. The net effect of a scandal is about 9 points, therefore. (This finding is reasonably consistent with previous research on the topic.) Fourteen of the 17 incumbents underperformed their projection by at least some amount, and the three exceptions came a relatively long time ago, in 1998. (There’s no evidence of the effect of scandals decreasing in recent elections; if anything, it’s increased slightly over the course of the data.)

Moreover, the effect of scandals is potentially greater in competitive districts, where the other party has an opportunity to mobilize a real alternative. Let’s use New York’s 27th Congressional District as a dividing line, for instance. It has a FiveThirtyEight partisan lean of R +22, meaning that it’s 22 points more Republican than the country as a whole based on its voting in recent presidential and state legislative elections.24 That type of district is ordinarily quite safe, but is just on the fringe of what could become competitive if everything breaks right for the opposing party — for example, in an election in a wave year against a candidate who just got arrested by the FBI. In districts less competitive than NY-27, scandals cost the incumbents only 4 percentage points, on average. But in districts that were as competitive or more competitive than NY-27, candidates with scandal issues underperformed their fundamentals by an average of almost 13 points.

So does that make Collins’s race a toss-up? You could do a little mental math: If the scandal costs him 13 percentage points, and the national environment favors Democrats by 6 points, that could produce a 19-point swing toward Collins’s Democratic opponent, Nate McMurray — almost enough to offset the strong Republican lean of the district. But you’d be leaving one thing out: Collins is still an incumbent, and incumbents usually outperform the partisan lean of their districts.

In fact, the incumbency bonus in recent elections has been in the very low double digits — on the order of 12 percentage points.25 (It used to be quite a bit higher.) That’s just about the same as the magnitude of the scandal penalty. The typical scandal, therefore, essentially wipes out a candidate’s incumbency advantage and makes the district perform similarly to an open-seat race. But it doesn’t necessarily reverse the advantage. Republicans would be favored to win an open-seat race in NY-27, even amid a very blue national political environment, so they’re probably still favored with Collins on the ballot too.

There’s one more complication, however, which is that this data suffers from survivorship bias. The candidates with really bad scandals will often retire rather than seek re-election, or they may lose in their primary. If all scandal-plagued incumbents were forced to be renominated, we’d probably observe a scandal penalty even larger than the 10 or 12 points we’re showing here.

But in some ways, Collins and the New York GOP are in a position where their hand has been forced. New York has already held its primary and Collins is the nominee; the general election is in only three months. He seems disinclined to bow out. And it isn’t entirely clear whether it would be possible to replace Collins on the ballot even if Republicans wanted to.26 This is the type of scandal that might have induced a retirement if it had occurred a year ago, but the GOP may not have that choice.

The Cook Political Report moved NY-27 from noncompetitive to its “Likely Republican” category after the news on Wednesday morning. I might go one step further and put it in the “Lean Republican” category instead, even though it’s a really red district. (It went for Trump by 25 points in 2016.) Soon, we’ll be able to tell you what the FiveThirtyEight House model thinks too, so it’s back to work on that.

But in general, Republicans face a very long list of potentially competitive districts — places where Democrats aren’t necessarily favored at even odds, but have a fighting chance when they have no real business doing so. That list got one seat longer after Collins’s arrest. Cashing in a few of those lottery tickets is what might turn a near-miss for the Democrats into a narrow majority — or a narrow majority into a wave.

Check out all the polls we’ve been collecting ahead of the 2018 midterms.

Is Chris Collins Toast?

In the era of President Trump, it’s become fashionable to presume that politicians can do whatever they like and get away with it. But if recent elections to Congress are any guide, scandals do have large and measurable effects. So when U.S. Rep. Chris Collins, the Republican from New York’s 27th Congressional District, was arrested on insider trading charges on Wednesday morning, it took a seat that had looked to be fairly safe for Republicans and put it into the competitive category.

I’m going to be fairly circumspect in this article because I’m knee-deep in finalizing our House model, and I don’t want to scoop our own forecast. But one of the things we evaluated in designing that model is the electoral effects of scandals, based on the data set of scandals put together by my colleague Nathaniel Rakich.1

Below is a list of scandal-plagued incumbents since 1998 who made it to the general election and faced an opponent from the opposite party.2 I’ve compared each incumbent’s actual margin of victory or defeat against a projected margin based on a “fundamentals” model that accounts for: (i) the incumbent’s previous victory margin,3 (ii) the partisan lean of the district,4 (iii) the generic ballot at the time of the election, (iv) congressional approval ratings at the time of the election (which are a good proxy for the overall mood toward incumbents), and (v) the incumbent’s congressional voting record (representatives who break with their party more often overperform on Election Day). This is a slightly pared-down version of what our House model will look at, but it should be a fairly robust and reliable model.5

How much do scandals hurt incumbents?

It depends on how competitive the district is

Year
District
Incumbent
Projected Margin Of Victory
Actual Margin of Victory or Defeat
Net Effect Of Scandal




1998
GA-6
Newt Gingrich
31.2
41.4
10.2


1998
ID-1
Helen Chenoweth
23.4
10.5
-12.9


1998
IL-6
Henry J. Hyde
32.6
37.2
4.7


1998
IN-6
Dan Burton
51.4
55.3
3.9


2000
GA-7
Bob Barr
21.0
10.5
-10.5


2004
OH-14
Steven C. LaTourette
34.0
25.5
-8.5


2006
MI-14
John Conyers, Jr.
78.1
70.6
-7.5


2006
PA-10
Donald Sherwood
25.1
-5.9
-31.0


2008
FL-16
Tim Mahoney
7.6
-20.2
-27.8


2008
NY-15
Charles B. Rangel
82.4
81.3
-1.2


2010
MA-6
John F. Tierney
22.3
13.9
-8.4


2012
FL-26
David Rivera
12.5
-10.6
-23.2


2012
NY-11
Michael G. Grimm
10.3
5.4
-4.9


2012
TN-4
Scott DesJarlais
24.8
11.5
-13.3


2016
NC-9
Robert Pittenger
27.2
16.4
-10.9


2016
NH-1
Frank C. Guinta
5.4
-1.3
-6.8


2016
TX-27
Blake Farenthold
28.5
23.4
-5.1




Overall average
30.5
21.5
-9.0




Districts less competitive than NY-27
45.7
42.0
-3.7




Districts more competitive than NY-27
19.8
7.1
-12.7

Shaded districts were more competitive than NY-27 based on their partisan lean.

On average, the scandal-ridden incumbents … won re-election by 21.5 percentage points! But that’s quite a bit worse than their projected margin of victory, which was 30.5 percentage points. The net effect of a scandal is about 9 points, therefore. (This finding is reasonably consistent with previous research on the topic.) Fourteen of the 17 incumbents underperformed their projection by at least some amount, and the three exceptions came a relatively long time ago, in 1998. (There’s no evidence of the effect of scandals decreasing in recent elections; if anything, it’s increased slightly over the course of the data.)

Moreover, the effect of scandals is potentially greater in competitive districts, where the other party has an opportunity to mobilize a real alternative. Let’s use New York’s 27th Congressional District as a dividing line, for instance. It has a FiveThirtyEight partisan lean of R +22, meaning that it’s 22 points more Republican than the country as a whole based on its voting in recent presidential and state legislative elections.6 That type of district is ordinarily quite safe, but is just on the fringe of what could become competitive if everything breaks right for the opposing party — for example, in an election in a wave year against a candidate who just got arrested by the FBI. In districts less competitive than NY-27, scandals cost the incumbents only 4 percentage points, on average. But in districts that were as competitive or more competitive than NY-27, candidates with scandal issues underperformed their fundamentals by an average of almost 13 points.

So does that make Collins’s race a toss-up? You could do a little mental math: If the scandal costs him 13 percentage points, and the national environment favors Democrats by 6 points, that could produce a 19-point swing toward Collins’s Democratic opponent, Nate McMurray — almost enough to offset the strong Republican lean of the district. But you’d be leaving one thing out: Collins is still an incumbent, and incumbents usually outperform the partisan lean of their districts.

In fact, the incumbency bonus in recent elections has been in the very low double digits — on the order of 12 percentage points.7 (It used to be quite a bit higher.) That’s just about the same as the magnitude of the scandal penalty. The typical scandal, therefore, essentially wipes out a candidate’s incumbency advantage and makes the district perform similarly to an open-seat race. But it doesn’t necessarily reverse the advantage. Republicans would be favored to win an open-seat race in NY-27, even amid a very blue national political environment, so they’re probably still favored with Collins on the ballot too.

There’s one more complication, however, which is that this data suffers from survivorship bias. The candidates with really bad scandals will often retire rather than seek re-election, or they may lose in their primary. If all scandal-plagued incumbents were forced to be renominated, we’d probably observe a scandal penalty even larger than the 10 or 12 points we’re showing here.

But in some ways, Collins and the New York GOP are in a position where their hand has been forced. New York has already held its primary and Collins is the nominee; the general election is in only three months. He seems disinclined to bow out. And it isn’t entirely clear whether it would be possible to replace Collins on the ballot even if Republicans wanted to.8 This is the type of scandal that might have induced a retirement if it had occurred a year ago, but the GOP may not have that choice.

The Cook Political Report moved NY-27 from noncompetitive to its “Likely Republican” category after the news on Wednesday morning. I might go one step further and put it in the “Lean Republican” category instead, even though it’s a really red district. (It went for Trump by 25 points in 2016.) Soon, we’ll be able to tell you what the FiveThirtyEight House model thinks too, so it’s back to work on that.

But in general, Republicans face a very long list of potentially competitive districts — places where Democrats aren’t necessarily favored at even odds, but have a fighting chance when they have no real business doing so. That list got one seat longer after Collins’s arrest. Cashing in a few of those lottery tickets is what might turn a near-miss for the Democrats into a narrow majority — or a narrow majority into a wave.

Check out all the polls we’ve been collecting ahead of the 2018 midterms.

Politics Podcast: Should The Press Respond To Trump’s Attacks?

FiveThirtyEight












 












More: Apple Podcasts |
ESPN App |
RSS
| Embed


Embed Code

President Trump’s attacks on the press have reached a new level in recent weeks. On Sunday, he called the press, “very dangerous & sick” and wrote that the media can “cause War.” The FiveThirtyEight Politics podcast team talks about what the goal of the president’s rhetoric is and how the press should respond.

The crew also previews the last special congressional election to happen before the midterms — in Ohio’s 12th District on Tuesday — and reviews the key primaries to watch that night.

You can listen to the episode by clicking the “play” button above or by downloading it in iTunes , the ESPN App or your favorite podcast platform. If you are new to podcasts, learn how to listen .

The FiveThirtyEight Politics podcast publishes Monday evenings, with occasional special episodes throughout the week. Help new listeners discover the show by leaving us a rating and review on iTunes . Have a comment, question or suggestion for “good polling vs. bad polling”? Get in touch by email, on Twitter or in the comments.