A wicked early look at 2020 Senate and gubernatorial races

In recent posts, I began to take a wicked early look at the 2020 U.S. elections. First, I assessed the field of Democrats seeking to challenge Republican President Donald Trump in 2020. Then I turned to the 2020 presidential election itself, pondering how Democrats would potentially fare against Trump.

Now I turn my attention to

1. The 34 elections for United States Senate (“Senate”) to be held in 2020.
2. The three gubernatorial elections to be held 2019 (Kentucky, Louisiana, Mississippi) and the 11 gubernatorial elections to be held in 2020.

My goal is primarily to provide the view from 30,000 feet: what the “fundamentals” in each race reveal about the overall partisan landscape—and what the likelihood is Democrats will have the Senate majority in January 2021 (and cut into the Republican advantage in governor’s mansions) As such, I only briefly discuss actual or potential candidates in these races, other than incumbents seeking reelection.

“Fundamentals” are simply the sum of three values:

1. The state’s partisan lean, measured by my 3W-RDM (weighted[1] three-election average of the difference between a state’s Democratic [minus Republican] margin in a presidential election and the Democratic [minus Republican] margin in the total national vote in that election).
2. The estimated effect of incumbency (incumbent office-holders tend to receive a higher percentage of the vote than an open-seat candidate of the same party).
3. The national partisan lean, as measure by the “generic ballot” question (variations on “If the election for were held today, would you vote for the Democratic candidate, the Republican candidate, or some other candidate?”)

**********

Just bear with me as I explain how I estimated the effect of incumbency for Senate and gubernatorial elections. As usual, unless otherwise noted, election data come from Dave Liep’s indispensable Atlas of U.S. Presidential Elections.

Senate. I first calculated an “expected margin of victory”[2] for each Democratic Senate nominee in the 35 Senate elections in 2018[3], the 34 Senate elections in 2016[4] and the 35 Senate elections in 2014[5]: a state’s 3W-RDM plus the national Democratic margin (minus Republican percentage of all votes cast) in that year’s elections. Using three elections years guarantees a minimum of two Senate elections from each state. The margins for the three previous Senate election years are:

2014 = D-5.8%

2016 = D+0.9%

2018 = D+9.9%

Next, I subtracted each actual margin (Democratic minus Republican) from the “expected” margin. I then calculated three averages of these differences within each election year:

1. Races with Democratic incumbents
2. Races with Republican incumbents
3. Open-seat races (where expected margin is for party currently holding the office)

Within each election year, then, the effect of incumbency for Democrats is simply the first average minus the third average[6], while the Republican advantage is the second average minus the third average[7]. And the estimated effect of incumbency for each party is the weighted average (2018=3, 2016=2, 2014=1) of the election-year averages.

For Democratic Senate incumbents, the effect is +4.4 percentage points (“points”), and for Republican Senate incumbents the effect is +2.6 points. Somewhat arbitrarily, I divide these values by 1.5 for incumbents who have won a special election, but not yet served a full six-year term and by 2.0 for incumbents who were appointed to the seat and have yet to face the voters.

Governor. Complicating these calculations is that five states hold their gubernatorial elections in odd-numbered years; thus, in November 2019, Kentucky, Louisiana and Mississippi will elect a governor, as will New Jersey and Virginia in November 2021.

As a result, I analyzed data from two-year cycles: 38 gubernatorial elections in each of 2017-18[8] and 2013-14[9], and 15 gubernatorial elections in each of 2015-16[10] and 2011-12[11]; going back to 2011 guarantees at least two gubernatorial elections from each state (with New Hampshire and Vermont, which hold gubernatorial elections every two years, included four times[12]). The calculations were otherwise the same, except for calculating a four-cycle weighted average (4,3,2,1)[13]: for Democratic gubernatorial incumbents, the effect is +5.7 points, and for Republican gubernatorial incumbents the effect is +8.5 points.

That the effect of incumbency is stronger for governors than for Senators reflects how partisan Senate elections have become.

**********

Let us now turn to the elections themselves. I base the “national lean” of D+6.0 on generic ballot polls listed on FiveThirtyEight.com, which have varied between D+2 and D+9—and mostly between D+5 and D+7—over the last few weeks. While this value is broadly in line with the last four Senate election years (weighted average=D+4.3 points; unweighted average from last two presidential election years= D+6.5), it is much higher than the last four gubernatorial election cycles (weighted average=D-0.6 points; unweighted average from last two presidential election years= D-3.2).

2020 Senate elections. Republicans currently hold 53 Senate seats, with 47 held by Democrats (including Independent Senators Angus King of Maine and Bernie Sanders of Vermont, who caucus with Democrats), meaning that to win back the majority in 2020, Democrats need either to win a net four seats, or win a net three seats and win the presidential election (Democratic Vice President would break 50-50 tie).

Table 1. 2020 Senate election overview

Name	State	Run 2020	3W-RDM	INC	Nat Lean	Total	Last margin	First elected
DEMOCRATS
Edward Markey	MA	Yes	22.1	4.4	6.0	32.5	23.9%	2013
Jack Reed	RI	Yes	18.0	4.4	6.0	28.4	41.3%	1996
Richard Durbin	IL	Yes	14.7	4.4	6.0	25.1	14.6%	1996
Chris Coons	DE	Yes	12.5	4.4	6.0	22.9	13.6%	2010
Cory Booker	NJ	Yes	12.0	4.4	6.0	22.4	13.5%	2012
Jeff Merkley	OR	Yes	8.7	4.4	6.0	19.1	18.9%	2008
Tom Udall	NM	No	6.5	0.0	6.0	12.5	—	N/A
Gary Peters	MI	Yes	2.2	4.4	6.0	12.6	13.3%	2014
Mark Warner	VA	Yes	1.5	4.4	6.0	11.9	0.8%	2008
Tina Smith	MN	Yes	1.5	2.9	6.0	10.4	10.6%	2018
Jeanne Shaheen	NH	Yes	0.1	4.2	6.0	10.3	3.3%	2008
Doug Jones	AL	Yes	-28.4	2.2	6.0	-20.2	1.7%	2017
REPUBLICANS
Susan Collins	ME	Yes	5.9	-2.4	6.0	9.5	37.0%	1996
Cory Gardner	CO	Yes	2.2	-2.4	6.0	5.8	1.9%	2014
Joni Ernst	IA	Yes	-4.7	-2.4	6.0	-1.1	8.3%	2014
Thom Tillis	NC	Yes	-6.0	-2.4	6.0	-2.4	1.6%	2014
David Perdue	GA	Yes	-9.6	-2.4	6.0	-6.0	7.7%	2014
Martha McSally	AZ	Yes	-9.7	-1.2	6.0	-4.9	—	Apptd 2019
John Cornyn	TX	Yes	-15.3	-2.4	6.0	-11.7	27.2%	2002
Lindsey Graham	SC	Yes	-15.7	-2.4	6.0	-12.1	15.5%	2002
Cindy Hyde-Smith	MS	Yes	-18.5	-1.6	6.0	-14.1	7.3%	2018
Steve Daines	MT	Yes	-18.6	-2.4	6.0	-15.0	17.7%	2014
Dan Sullivan	AK	Yes	-19.2	-2.4	6.0	-15.6	2.1%	2014
Bill Cassidy	LA	Yes	-22.2	-2.4	6.0	-18.6	11.9%	2014
Pat Roberts	KS	No	-23.4	0.0	6.0	-17.4	—	N/A
Lamar Alexander	TN	No	-25.8	0.0	6.0	-19.8	—	N/A
Ben Sasse	NE	Yes	-25.8	-2.4	6.0	-22.2	32.8%	2014
Mike Rounds	SD	Yes	-25.8	-2.4	6.0	-22.2	20.9%	2014
Tom Cotton	AR	Yes	-28.2	-2.4	6.0	-24.6	17.1%	2014
Mitch McConnell	KY	Yes	-28.7	-2.4	6.0	-25.1	15.5%	1984
James Risch	ID	Yes	-34.2	-2.4	6.0	-30.6	30.7%	2008
Shelley Moore Capito	WV	Yes	-35.5	-2.4	6.0	-31.9	27.7%	2014
James Inhofe	OK	Yes	-38.1	-2.4	6.0	-34.5	39.5%	1994
Mike Enzi	WY	No	-45.7	0	6.0	-39.7	—	N/A

At first glance, Democrats appear to have a significant advantage in the 2020 Senate elections (Table 1): of 34 Senate elections scheduled for November 2020, fully two-thirds (22) are currently Republican-held. And of those 22 seats, fully 73% (16) are potentially more vulnerable because they include…

• 11 incumbents who first won in 2014, when Republicans won the overall Senate vote by nearly 6 points, comprising
  • Five who defeated a Democratic incumbent (Cory Gardner, Thom Tillis, Dan Sullivan, Bill Cassidy, Tom Cotton)
  • Four who won a Democratic-held open seat (Joni Ernst, Steve Daines, Mike Rounds, Shelly Moore Capito)
  • Two who won a Republican-held open seat (David Perdue, Ben Sasse)
• Two incumbents yet to serve a full term:
  • Martha McSally—who had just narrowly lost to Democrat Kyrsten Sinema a few weeks earlier—was appointed to fill the Arizona Senate seat left vacant when Republican incumbent John McCain died,
  • Cindy Hyde-Smith was appointed to the Mississippi Senate seat in April 2018 when incumbent Republican Thad Cochran retired for health reasons, narrowly winning election in her own right that November over former Agriculture Secretary Mike Espy.
• Three open seats (Kansas, Tennessee, Wyoming).

Moreover, only one currently-Democratic seat appears particularly vulnerable as of now: Jones’ seat in deep-red Alabama (D-28.4); a reasonable estimate is that Jones would lose to a generic Republican by around 20 points. Even with the full effect of incumbency (+4.2), a repeat of Democrats’ strong overall performance in 2018 (D+9.9) and a pro-Democratic error of 5.4 points in 3W-RDM (the average miss over time), Jones would still be down about nine points to a generic Republican. Yes, Jones overcame similar odds in December 2017, but that was against a severely compromised Republican opponent.

And while first-term Democratic Senators Gary Peters of Michigan and Tina Smith of Minnesota (who won by double-digits in November 2018 after being appointed to replace Democrat Al Franken in December 2017) could be vulnerable—along with Jeanne Shaheen of New Hampshire and Mark Warner of Virginia, who narrowly won reelection in 2014—if Democrats only break even in 2020, as of now, they appear quite likely to prevail. They would join six safe Democratic incumbents (Ed Markey, Jack Reed, Richard Durbin, Chris Coons, Cory Booker[14] and Jeff Merkley) and a likely-safe open seat in New Mexico (with Democratic United States House of Representatives member [“Representative”] Ben Ray Luján a strong candidate to win the seat).

2020 New Mexico Democratic Senate candidate Ben Ray Luján,

However, Democrats should not be banking on New York Senator Chuck Schumer switching from Minority to Majority Leader in January 2021 just yet. While as many as 16 Republican-held seats are arguably vulnerable, only two are in states that even lean Democratic: Maine (D+5.9) and Colorado (D+2.2). And while Gardner is clearly vulnerable (he underperformed by about four points in 2014, when he beat incumbent Democrat Mark Udall), even a slight improvement by Republicans in the total national Senate vote puts that seat at toss-up status, at best. And Collins has been winning statewide in Maine since 1996, including winning her fourth term by an eye-popping 37.0 points!

Plus, the next four most vulnerable Republican incumbents (all finishing their first term)—Ernst, Tillis, Perdue and McSally—represent states averaging 7.5 points more Republican than the nation as a whole; this is why even in a very good Democratic year the fundamentals have these races “toss-up” at best. Moreover, while it is true that Ernst, Tillis and Perdue won in 2014 by an average of just 5.9 points (with McSally losing by 2.3 points in 2018)—a hair over the overall Senate Republican that year—all four now have the modest added advantage of running as incumbents in lean-Republican states. And where Democrats have a strong candidate to run against McSally—former astronaut Mark Kelly, husband of former Representative Gabby Giffords (D-AZ)—other strong candidates such as former Iowa Governor (and Secretary of Agriculture) Tom Vilsack and former Georgia House Speaker Stacey Abrams have ruled out running for the Senate in 2020.

Left to right: former Representative Gabby Giffords and 2020 Democratic Senate candidate Mark Kelly

Abrams and Vilsack are not the only high-profile Democrats choosing not to challenge vulnerable incumbent Republican Senators. Former National Security Advisor Susan Rice will not challenge Collins, while former Colorado Governor John Hickenlooper is running for president in 2020 instead.

Beyond those six races, Democratic chances to flip seats only get slimmer. Former South Carolina Democratic Party chair Jaime Harrison is formally exploring a bid against Senator Lindsey Graham. And while former Representative Beto O’Rourke (D-TX; running for president) and Representative Joaquin Castro (D-TX) passed on a run, Air Force veteran Mary Jennings “MJ” Hegar, who came within 3 points of defeating incumbent Representative John Carter (D-TX) in 2018, plans to run against Senator John Cornyn. Even with Democrats winning nationally by six points, however, the fundamentals suggest both Harrison and Hegar begin their races down around 12 points.

2020 South Carolina Senate Democratic candidate Jaime Harrison

2020 Texas Senate Democratic candidate MJ Hegar

Mississippi, meanwhile, will see a rematch between Espy and Hyde-Smith as she seeks a first full term. But while he came within about seven points of unseating her in 2018, this will be a tough Senate race for Democrats to win, as the fundamentals have him down by 15.1 points—similar to the Democratic position against first-term Senators Daines in Montana (where outgoing Democratic Governor Steve Bullock is apparently running for president instead) and Sullivan (who only defeated Democratic incumbent Mark Begich by 2.1 points in 2014) in Alaska.

2020 Mississippi Democratic Senate candidate Mike Espy

That leaves 11 Republican-held Senate seats which average 30.3 points more Republican than the nation. Even with three open seats it is very difficult to see how Democrats flip any of them. One intriguing exception, however, could be in Kentucky, where Senate Majority Leader Mitch McConnell (quite unpopular at home) is seeking a seventh term; if Air Force veteran Amy McGrath (who, like Hegar, came within three points of defeating an incumbent Republican Representative in 2018—in this case Andy Barr) were to run, she may be able to overcome the fundamentals showing a generic Democrat down 25.1 points to McConnell.

The bottom line?

While there are several plausible paths for Democrats to win back a Senate majority in 2020…

1. Win presidency; Jones win in Alabama; win Maine and Colorado and one of Iowa, North Carolina, Georgia, Arizona
2. Win presidency; Jones lose in Alabama; win Maine and Colorado and two of Iowa, North Carolina, Georgia, Arizona
3. Lose presidency; Jones win in Alabama; win Maine and Colorado and two of Iowa, North Carolina, Georgia, Arizona
4. Lose presidency; Jones lose in Alabama; win Maine and Colorado and three of Iowa, North Carolina, Georgia, Arizona
5. Any of 1-4 above but substituting wins in even more Republican states such as Texas, South Carolina, Mississippi, Alaska and Kentucky.

…a great deal would have to go just right for Democrats in each scenario. In fact, it is easy to foresee anything from Democrats net losing a handful of seats (Alabama and some combination of Michigan, Minnesota, New Hampshire and Virginia) to winning a clear majority (holding Alabama, sweeping the six most vulnerable states and maybe even picking off South Carolina and/or Texas and/or Mississippi and/or Kentucky) is possible.

The silver lining for Democrats, though, is that forcing Republicans to invest money, time and resources in states like Alabama, Alaska, Arizona, Georgia, Iowa, Kentucky, Mississippi, North Carolina, South Carolina and Texas makes it that much harder for them to beat Democratic incumbents in Michigan, Minnesota, New Hampshire and Virginia.

2019-20 Gubernatorial elections. Republicans currently occupy governor’s mansions in 27 states, with Democrats occupying the remaining 23.

Three gubernatorial elections will be held in 2019, all in southern states averaging 23.1 points more Republican than the nation (Table 2). The lone Democrat is John Bel Edwards of Louisiana, and while the fundamentals have him down to a generic Republican by 10.5 points, he is generally popular with voters in his state and thus more likely than not to win reelection. By contrast, the only Republican governor to seek reelection this year—Matt Bevin of Kentucky—is the least popular governor in the country; still, the fundamentals have him beating a generic Democrat by a whopping 31.2 points. As for the open seat in Mississippi, the fundamentals have a generic Republican defeating a generic Democrat by 12.5 points.

This means that the likeliest outcome is no net change in partisan control of governor’s mansions in 2019—though that could mean the parties switch control in Louisiana and Kentucky!

Table 2. 2019-20 Gubernatorial election overview

Name	State	Run 2019/ 2020	3W-RDM	INC	Nat Lean	Total	Last margin	First elected
2019 DEMOCRATS
John Bel Edwards	LA	Yes	-22.2	5.7	6.0	-10.5	12.2%	2015
2019 REPUBLICANS
Phil Bryant	MS	No	-18.5	0.0	6.0	-12.5	—	N/A
Matt Bevin	KY	Yes	-28.7	-8.5	6.0	-31.2	8.7%	2015
2020 DEMOCRATS
John Carney	DE	Yes	12.5	5.7	6.0	24.2	19.2%	2016
Jay Inslee	WA	No	12.1	0.0	6.0	18.1	—	N/A
Roy Cooper	NC	Yes	-6.0	5.7	6.0	5.7	0.2%	2016
Steve Bullock	MT	No	-18.6	0.0	6.0	-12.6	—	N/A
2020 REPUBLICANS
Phil Scott	VT	Yes	27.7	-8.5	6.0	25.2	14.9%	2016
Chris Sununu	NH	Yes	0.1	-8.5	6.0	-2.4	7.0%	2016
Eric Holcomb	IN	Yes	-16.3	-8.5	6.0	-18.8	6.0%	2016
Mike Parson	MO	Yes	-15.9	-4.3	6.0	-14.2	—	Succ 2018
James Justice	WV	Yes	-35.5	-8.5	6.0	-38.0	6.8%	2016
Doug Burgum	ND	Yes	-29.4	-8.5	6.0	-31.9	57.1%	2016
Gary Herbert	UT	No	-33.1	0.0	6.0	-27.1	—	N/A

Looking ahead to 2020, two states currently governed by Democrats, Delaware and Washington, are all-but-certain to remain in Democratic hands, with Governor John Carney poised to reprise his nearly-20-point win in 2016 and a Democrat (state Attorney General Bob Ferguson?) heavily favored to succeed Governor Jay Inslee (running for president instead).

Equally certain to remain in Republican hands are West Virginia and North Dakota (where James Justice—who switched parties after winning as a Democrat in 2016—and Doug Burgum will seek reelection), as well as Utah, where Governor Gary Herbert is term-limited from seeking reelection.  The fundamentals in these states have Republicans ahead by 32.3 points over a generic Democrat.

That leaves six races which could be competitive—although Governors Eric Holcomb of Indiana and Mike Parson (who became governor in June 2018, following the resignation of Eric Greitens, just elected in 2016) of Missouri—are ahead in the fundamentals by 14-19 points.

North Carolina Governor Roy Cooper barely defeated Republican incumbent Pat McCrory in 2016, and while the fundamentals have him beating a generic Republican by 5.7 points, this race would be a pure toss-up in a neutral partisan environment. Montana is another story, though, with Bullock retiring after two terms (and 16 consecutive years of Democratic governors); the fundamentals suggest a generic Republican would win back the governor’s mansion in Helena by 12.6 points (and that is with Democrats winning by six points nationally).

That only leaves two New England Republican governors who just won reelection last year, but who the fundamentals see as highly vulnerable: Phil Scott, who won by nearly 15 points in deep-blue Vermont (D+27.7), and Chris Sununu, who “only” won by 7.0 points in swing-state New Hampshire. If they did not lose in 2018, though, it is unlikely (though not impossible) they will lose in 2020.

The bottom line?

As of May 2019, the 14 gubernatorial elections in 2019 and 2020 will most likely result in a net gain of 1 (with Republicans winning the open governor’s seat in Montana) governor’s mansion, expanding their overall lead to 28-22—but races this year in Kentucky and Louisiana, and next year in Indiana, Missouri, New Hampshire, North Carolina, and Vermont could yet surprise.

Until next time…

[1] The most recent election is weighted “3,” the 2^nd-most recent election is weighted “2” and the 3^rd-most recent election is weighted “1.”

[2] That is, relative to the Republican candidate. I excluded data from special elections such as the December 2017 Senate election in Alabama.

[3] For the California Senate election, I used the total votes for Democratic, Republican and all-other-party candidates in the June 5, 2018 “jungle primary.” For the Mississippi special Senate election, I used the results from the runoff election on November 27, 2018. For the Maine and Vermont Senate races, I counted as “Democratic” votes those cast for Independent Senators Angus King and Bernie Sanders, respectively, since each man caucuses with the Democrats (and there was no Democratic Senate nominee in Vermont); in Maine, I counted the Democratic votes as “other.” Notably, counting votes for King and Sanders as “other” (and Democratic votes in Maine as “Democratic”) only changes the national Democratic margin from +9.9 percentage points to +9.4.

[4] For the California Senate election, I used the total votes for Democratic, Republican and all-other-party candidates in the June 7, 2016 “jungle primary.” For the Louisiana Senate election, I used the results from the runoff election on December 10, 2016.

[5] I excluded the Alabama Senate race in which Republican incumbent Jeff Sessions ran unopposed.

[6] These values were +0.9% in 2018, +6.5% in 2016 and +10.6% in 2014.

[7] These values were +2.6% in 2018, +3.6% in 2016 and -0.7% in 2014.

[8] I counted the 2018 Alaska gubernatorial election as a Democratic open seat after Independent Governor Bill Walker suspended his reelection campaign on October 19, 2018, throwing his support to Democratic nominee Mark Begich.

[9] I counted Walker as a Democrat in 2014 Alaska gubernatorial election (though counting him as “Other” would have made little material difference). I counted the Rhode Island gubernatorial election as a Democratic open seat although outgoing Governor Lincoln Chafee was an Independent (who briefly sought the 2016 Democratic presidential nomination).

[10] For the 2015 Louisiana gubernatorial election, I used data from the runoff election held November 21, 2015.

[11] Because incumbent Republican governor Bobby Jindal easily cleared the 50% threshold on election day 2011, for the 2011 Louisiana gubernatorial election, I used the sum of all votes cast for the candidate of each political party (Republican, Democrat, Other) that day.

[12] West Virginia is counted three times because it also held a special gubernatorial election in 2011.

[13] Democratic incumbency “advantage” was +2.0% in 2017-18, +6.3% in 2015-16, +5.7% in 2013-14 and +18.9% in 2011-12; the corresponding Republican values were +17.3%, -3.4%, +10.3% and +5.1%.

[14] Or whoever replaces him, should he become the next president or vice president of the United States.