It is time to take a wicked early look at the 35 elections for the United States Senate (“Senate”) scheduled to end on November 3, 2026. Republicans are defending a 53-45-2 majority, although two Independents – Angus King of Maine and Bernie Sanders of Vermont – vote with Democrats to organize the Senate. If Democrats net four seats in 2026, they regain control of the Senate after losing it in 2024. At this early stage, based almost entirely on what I call fundamentals, this does not appear especially likely – though there is a minimally-plausible path or two.
By “fundamentals,” I mean the sum of…
- State partisan lean,
- Estimated incumbency advantage and
- National partisan lean
…which is a baseline estimate for how a generic Democrat would fare against a generic Republican, all else being equal.
State partisan lean is how much more or less Democratic than the nation a state voted for president in the last three elections, weighted by recency (2024=5, 2020=3, 2016=1). I call this measure 3W-RDM. It is how you would expect a generic Democrat to fare against a generic Republican in a state if the two parties split the national vote evenly, all else being equal.
Incumbents have historically had an advantage when seeking reelection, albeit a shrinking one. I thus calculate estimated incumbency advantage (“EIA”) for every Senate election. I first calculate an expected margin of victory for every Senate election held from 2014 to 2024 (n=208), when all 100 Senate seats came up for election at least twice. Expected margin of victory is state partisan lean plus the difference between the Democratic and Republican percentages of all Senate votes cast in an election cycle. The two most recent cycles featuring the 2026 set of Senate seats are 2020 and 2014. Unless otherwise specified, election data come from Dave Leip’s invaluable website.
2024 = D+2.0 percentage points (“points”)[1]
2022 = D+0.1 points
2020 = D-1.9 points
2018 = D+9.9 points[2]
2016 = D+1.7 points
2014 = D-6.1 points
For example, Democratic New Mexico Senator Martin Heinrich ran for reelection in 2024, when New Mexico’s 3W-RDM was D+6.3. Counting every vote cast in that year’s 35 Senate elections, Democratic Senate candidates received 50.1% and Republicans received 48.1%, a Democratic margin of 2.0 points. Thus, a generic Democrat would be expected to win a 2024 Senate race in New Mexico by 6.3 + 2.0 = 8.3 points. Heinrich actually won by 10.1 points, for EIA = 10.1 – 8.3 = +1.8.
Over the last six elections, 90 Democrats have run for reelection, and they average EIA of +4.4, adjusting for number seeking reelection and recency.[3] The 81 Republicans running for reelection during this same period averaged EIA of +2.8. In the 37 open seats – no incumbent sought reelection – the Democratic candidate averaged EIA of +2.9.
As of late September 2025, eight Senators – four Democrats and four Republicans – have announced they will not seek reelection in 2026 (Tables 2 and 3). I assign EIA = D+2.9 to these open seats. Similarly, to the three Democrats and seven Republicans running for reelection for the first time in 2026, I assign EIA of D+4.4 and R+2.8, adjusted for percentage of term served. Republican Senator Markwayne Mullin of Oklahoma won a special election in 2022 to serve out the term of Republican Senator James Inhofe, who retired in February 2022. Mullin thus gets “incumbency” credit for two-thirds of his term. Republicans Ashley Moody and Jon Husted get “incumbency” credit for one-third of their term because they were appointed in January 2025 to fill the vacancies left by Marco Rubio of Florida and J.D. Vance of Ohio, who became Secretary of State and Vice President, respectively.
That leaves seven Democrats and 10 Republicans running for reelection for at least a second time. If they first ran for reelection in 2020, I use EIA from that year. Otherwise, I used a weighted average of 2014 and 2020, with 2020 weighted twice as much.
This brings us to national partisan lean, which I estimated in recent elections using the generic ballot[4] polls curated by FiveThirtyEight.com. However, with the disquieting demise of this pioneering data-driven journalism website, there is no longer a reliable repository of newly-released polls. For now, then, I average the generic ballot poll averages calculated by RealClearPolitics (strong Republican bias), Ballotpedia (unclear bias), DecisionDeskHQ (unclear bias) and RacefortheWH (strong Democratic bias). As of late September 2025, these values are D+3.6, D+5.5, D+4.6 and D+3.3, respectively, for an average of D+4.2.
However, because the generic ballot reflects partisan lean nationally, and not solely in the 35 states holding Senate elections in 2026, I adjusted my metric by the difference between the Democratic margins in the total vote cast for United States House of Representatives (“House”) and for the Senate in the last six elections:
Table 1: Democratic margin in all votes case for House and Senate, 2014-24
| Year | House Margin | Senate Margin | Difference |
| 2024 | D-2.7 | D+2.0 | +4.7 |
| 2022 | D-2.8 | D+0.1 | +2.9 |
| 2020 | D+3.1 | D-1.9 | -5.0 |
| 2018 | D+8.6 | D+9.9 | +1.3 |
| 2016 | D-1.0 | D-0.7 | +0.3 |
| 2014 | D-5.8 | D-6.1 | -0.3 |
On average, the Senate Democratic margin was only 0.05 points lower than the House Democratic margin over the last six elections. In the last two elections featuring this set of seats, though, Senate Democratic candidates fared a mean 2.7 points worse – while in 2018, the last time a Republican president faced a midterm election, Senate Democratic candidates fared 1.3 points better. Wanting to capture information from every recent Senate election, I thus calculated a weighted-average[5] difference of -1.0.[6] At this point in the cycle then, I assume there will be a 3.2-point Democratic margin in all votes cast for Senate in 2026.
**********
We start with the Democrats. Table 2 below summarizes the fundamentals for the 13 seats they will defend in 2026.
[Eds. note: Table updated November 9, 2025 with recalcuated Democratic win probabilities: normal distribution with mean difference between fundamentals and final margin of 0.6 and standard deviation of 11.2, based on the last 241 Senate elections.]
Table 2: 2026 Senate elections – 13 Democratic held-seats:
| Name | State | Run 2026? | 3W-RDM | INC | Nat Lean | Expected | Last margin | First elected/apptd | P (Dem win) |
| Ed Markey | MA | Yes | 27.2 | 11.8 | 2.4 | 41.4 | 33.1% | 2013 | 100.0% |
| Jack Reed | RI | Yes | 15.3 | 19.6 | 2.4 | 37.4 | 33.1% | 1996 | 99.9% |
| Chris Coons | DE | Yes | 14.5 | 8.8 | 2.4 | 25.8 | 21.5% | Appt 2009 | 98.8% |
| Jeff Merkley | OR | Yes | 13.2 | 12.7 | 2.4 | 28.3 | 17.6% | 2008 | 99.3% |
| Dick Durbin | IL | No | 12.8 | 2.9 | 2.4 | 18.2 | — | N/A | 94.1% |
| John Hickenlooper | CO | Yes | 9.7 | 4.4 | 2.4 | 16.5 | 9.3% | 2020 | 92.2% |
| Cory Booker | NJ | Yes | 9.5 | 6.9 | 2.4 | 15.8 | 24.2% | 2012 | 91.2% |
| Ben Ray Lujan | NM | Yes | 6.9 | 4.4 | 2.4 | 13.7 | 6.1% | 2020 | 87.8% |
| Mark Warner | VA | Yes | 6.0 | 11.0 | 2.4 | 19.4 | 12.1% | 2008 | 95.3% |
| Tina Smith | MN | No | 3.6 | 2.9 | 2.4 | 9.0 | — | N/A | 77.1% |
| Jeanne Shaheen | NH | No | 2.8 | 2.9 | 2.4 | 8.2 | — | N/A | 74.8% |
| Gary Peters | MI | No | -0.9 | 2.9 | 2.4 | 4.8 | — | N/A | 63.3% |
| Jon Ossoff | GA | Yes | -3.0 | 4.4 | 2.4 | 3.8 | 1.2% | 2020 | 61.1% |
Democrats start the cycle at least 7:1 favorites to win nine of these 13 seats, including the open seat in Illinois, where Democratic House Member Raja Krishnamoorthi has an early edge over two black women: Democratic House Member Robin Kelly and Lieutenant Governor Juliana Stratton.
Two Democratic women are retiring in states that lean Democratic nationally: Tina Smith in Minnesota and Jeanne Shaheen in New Hampshire. In the former, Democratic House Member Angie Craig and Lieutenant Governor Peggy Flanagan are the early favorites to be the Democratic nominee against token Republican opposition. In the latter, Democratic House Member Chris Pappas will face one of two former Republican Senators, Scott Brown or, far more likley, John E. Sununu. Both races start Lean Democratic.
That leaves two Senate seats held by Democrats that are currently competitive.
In Michigan, a toss-up state at the national level, Democratic Senator Gary Peters is retiring after two terms. Two young House Members – Mallory McMorrow and Haley Stevens – are vying with former Wayne County Health Director Abdul El-Sayed for the Democratic nomination, with state Attorney General Dana Nessel still deciding whether to run. Thus, Democrats could be hurt by a highly competitive primary. The likely Republican nominee is former House Member Mike Rogers, who lost the 2024 Senate race by 0.3 points. Still, Democratic overperformance in open Senate elections and a Democratic environment nationally make the Democrats initial 5-3 favorites, or Tilts Democratic.
In Georgia, Jon Ossoff seeks a second term after narrowly flipping this Senate seat in 2020, securing the first Democratic Senate majority in six years. Georgia has moved 3.5 points more Democratic nationally since then – and it was one of only six states where 2024 Democratic presidential nominee Vice President Kamala Harris won more votes than 2020 Democratic presidential nominee Joe Biden. Her 74,384-vote increase topped all states. Ossoff’s Republican opponent will be one of two House Members, Buddy Carter and Mike Collins, or former Tennessee Volunteers head football coach Derek Dooley, endorsed by outgoing Republican Governor Brian Kemp. A changing electorate, incumbency and a Democratic environment nationally makes Ossoff an early 35-3 favorite, or Tilts Democratic.
Bottom line: Democrats start as 3-1 favorites to win 11 of their 13 Senate seats up for election in 2026. Jon Ossoff is only modestly favored to win reelection in Georgia, while Democrats are only slight favorites to retain the open seat in Michigan.
**********
Turning to the Republicans, Table 3 below summarizes the fundamentals for the 22 seats they will defend in 2026.
[Eds. note: Table updated November 9, 2025 with recalcuated Democratic win probabilities: normal distribution with mean difference between fundamentals and final margin of 0.6 and standard deviation of 11.2, based on the last 241 Senate elections.]
Table 3: 2026 Senate elections – 22 Republican-held seats:
| Name | State | Run 2026? | 3W-RDM | INC | Nat Lean | Expected | Last margin | First elected/apptd | P (Dem win) |
| Susan Collins | ME | Yes? | 5.9 | -9.7 | 2.4 | -1.4 | 8.6% | 1996 | 42.6% |
| Thom Tillis | NC | No | -3.8 | 2.9 | 2.4 | 1.6 | — | N/A | 53.2% |
| Ashley Moody | FL | Yes | -8.9 | -0.9 | 2.4 | -7.4 | — | Appt 2025 | 23.5% |
| Jon Husted | OH | Yes | -10.7 | -0.9 | 2.4 | -9.2 | — | Appt 2025 | 18.9% |
| John Cornyn | TX | Yes | -11.3 | 4.5 | 2.4 | -4.4 | 9.6% | 2002 | 32.6% |
| Joni Ernst | IA | No | -12.0 | 2.9 | 2.4 | -6.6 | — | N/A | 25.7% |
| Dan Sullivan | AK | Yes | -13.5 | 8.4 | 2.4 | -2.7 | 12.7% | 2014 | 38.3% |
| Lindsey Graham | SC | Yes | -16.3 | 6.4 | 2.4 | -7.4 | 10.3% | 2002 | 23.4% |
| Roger Marshall | KS | Yes | -17.4 | -2.8 | 2.4 | -17.8 | 11.4% | 2020 | 4.9% |
| Steve Daines | MT | Yes | -19.9 | 10.4 | 2.4 | -7.1 | 10.0% | 2014 | 24.5% |
| Cindy Hyde-Smith | MS | Yes | -21.0 | -2.8 | 2.4 | -21.4 | 10.0% | 2020 | 2.4% |
| Bill Cassidy | LA | Yes | -21.6 | -1.8 | 2.4 | -21.0 | 25.8% | 2014 | 2.7% |
| Pete Ricketts | NE | Yes | -21.9 | 0.7 | 2.4 | -18.8 | — | Appt 2023 | 4.1% |
| Bill Hagerty | TN | Yes | -28.0 | -2.8 | 2.4 | -28.4 | 27.0% | 2020 | 0.5% |
| Tommy Tuberville | AL | No | -29.4 | 2.9 | 2.4 | -24.0 | — | N/A | 1.4% |
| Steve Rounds | SD | Yes | -29.4 | -3.8 | 2.4 | -30.8 | 31.5% | 2014 | 0.2% |
| Mitch McConnell | KY | No | -30.0 | 2.9 | 2.4 | -24.6 | — | N/A | 1.2% |
| Tom Cotton | AR | Yes | -30.1 | -36.5 | 2.4 | -64.2 | 66.5% | 2014 | 0.0% |
| Jim Risch | ID | Yes | -34.8 | 7.8 | 2.4 | -24.6 | 29.4% | 2008 | 1.2% |
| Markwayne Mullin | OK | Yes | -35.3 | -1.9 | 2.4 | -34.5 | 26.5% | 2022 | 0.1% |
| Shelley Moore Capito | WV | Yes | -41.5 | -5.9 | 2.4 | -45.0 | 43.3% | 2014 | 0.0% |
| Cynthis Lummis | WY | Yes | -46.2 | -2.8 | 2.4 | -46.6 | 46.1% | 2020 | 0.0% |
Republicans start the cycle as at least 19-1 favorites in 13 of these 22 elections, including open seats in Alabama and Kentucky. In the former, first-term Senator Tommy Tuberville is running for governor, making Republican state Attorney General Steve Marshall the likely next Senator. In the latter, long-time Senate Republican leader Mitch McConnell is retiring after seven terms. Had he run, Democratic Governor Andy Beshear might have made this election competitive, but otherwise one of two Republicans – House Member Andy Barr and former state Attorney General Daniel Cameron – will succeed McConnell.
That leaves nine Republican-held seats that, at this stage, are at least minimally plausible for Democrats to win.
Working from least to most plausible, Republican incumbents Steve Daines in Montana and Lindsey Graham in South Carolina are surprisingly weak incumbents. In 2020, Daines underperformed expectations by 10.4 points against Democratic Governor Steve Bullock, while Graham underperformed by an average 6.4 points in 2014 and 2020. Daines and Graham continue to win solely because their states lean more than 15 points more Republican than the nation, meaning it would take either a historic Democratic wave and/or especially strong Democratic opponents to unseat them. Neither are apparent as of late September 2025.
In Iowa, Republican Senator Joni Ernst is retiring after two terms. Over the last decade, Iowa has shifted from 2.0 points more Democratic than the nation to 12.0 points more Republican. Thus, even in an open seat in a good Democratic year, either former State Senator Jim Carlin and House Member Ashley Hinson starts as roughly a 3-1 favorite to beat any Democrat, with former state senator Zach Wahls the early frontrunner over Des Moines School Board chair Jackie Norris and state representative Josh Turek.
That brings us to a key swing state – in 2000. Florida has become 6.5 points more Republican at the national level since then, which helps to explain why Democrats were demolished there in 2022 and 2024 after barely losing two statewide elections in 2018. And while the appointed Moody could be vulnerable next year, especially if Democrats replay 2018, only one Democrat – Brevard County school board member Jennifer Jenkins – has formally declared a bid.
As of late September 2025, then, these four elections are probably out of reach for Democrats, though Democrats still need to invest resources in Florida to right their ship in a state with 30 Electoral Votes, while Iowa has suffered from President Donald Trump’s tariff regime.
Which brings us to Ohio, another state that lost its swing. It is now 10.7 points more Republican than the nation at the national level, which is why Husted has an estimated 81.1% chance to win against a generic Democrat. However, former Democratic Senator Sherrod Brown is a brand-name, his career focus on affordability perfectly fits the moment, and he is running in 2026. Brown lost a bid for a fourth term in 2024 by only 3.6 points (while President Donald Trump won Ohio by 11.1 points). Summing Brown’s weighted-average EIA of +3.7 and Husted’s +0.9 gives Democrats a 29% chance to win. This feels too low, especially as Democratic turnout in Ohio dropped by 145,462 from 2020 to 2024.[7] That is 70% of the 206,434 by which Brown lost in 2024. Thus, Ohio starts the cycle as Tilts Republican.
On paper, two Republican incumbents seeking reelection should not be vulnerable, as their states average 12.4 points more Republican than the nation. Except that Dan Sullivan and John Cornyn are weak incumbents, with EIA of -8.4 and -4.5, respectively. And both states are trending Democratic, gaining a mean three points over the last three cycles. Going in reverse order – Texas’ Cornyn faces a bruising primary against state Attorney General Ken Paxton as he seeks a fifth term, potentially fracturing state Republicans along establishment/MAGA lines. Even if Cornyn survives, he will likely be severely weakened – whereas the once-indicted Paxton would be extremely vulnerable. If 2024 Democratic Senate nominee Colin Allred or state representative James Talarico – or, if she runs, House Member Jasmine Crockett – can then somehow overcome Texas’ traditionally low turnout,[8] Democrats could finally win a Senate seat there for the first time since Lloyd Bentsen won reelection in 1988 (while also Democratic nominee for vice president). Allred overperformed expectations by 1.5 points in 2024, losing by “only” 8.5 points. Texas thus starts the cycle as Tilts Republican.
Despite its Republican history, Alaska is winnable for Democrats in 2026, especially if the state moves another 1.5 points less Republican. A Democrat such as Ann Diener, sales manager for the Fairbanks Daily News-Miner, starts less than three points behind Sullivan, while former House Member Mary Peltola – who has already won statewide twice – could easily overcome that. Until and unless Peltola announces a Senate bid, this election starts Tilts Republican.
Maine’s Republican Senator Susan Collins has defied political gravity for three decades, winning five times in a state averaging 6.6 points more Democratic than the nation. In the strong Democratic year of 2020, she won by 8.6 points, for an EIA of +12.6. Add her 37-point blowout in 2014, and Collin’s EIA jumps to +22.4. If Collins runs for reelection, she probably beats a generic Democrat – but the fact she has yet to launch a formal campaign is why her EIA is -9.7, the average of -22.4 and and the open seat EIA of 2.9. However, both Sullivan Harbor Master Graham Platner and Democratic Governor Janet Mills, who is stepping down after two terms, are showing early polling strength. Each has disadvantages, though. Platner has a controversial history, while Mills is 77 years old. That said, Mills overperformed by 8.5 points in her successful 2022 reelection bid, while Platner fits the left-wing populism of the Democratic base. This election starts as a Toss-Up.
Which leaves the one Senate seat Democrats are currently more likely than not to flip: North Carolina. Republican Senator Thom Tillis is retiring, giving Democrats an opening in a state that has shifted 3.0 points more Democratic relative to the nation during the Trump years. Harris received 31,083 more votes here than Biden did in 2020.[10] Tillis won his Senate seat in 2014 by narrowly defeating incumbent Democrat Kay Hagan. That is the last time North Carolina had a Democrat in the Senate. Given this is an open seat in a close state in a good Democratic year, there is a 53% chance a generic Democrat beats a generic Republican, essentially a toss-up. Except, like Brown in Ohio, former Democratic governor Roy Cooper is a brand name. In 2016 and 2020, Cooper was elected governor by an average 2.4 points, while Trump won North Carolina by an average 2.5 points. The last time Cooper ran for an open seat (2016), he overperformed expectations by 9.7 points. Cooper’s likely Republican opponent is Republican National Committee chair Michael Whatley, who is seeking his first elective office. Taking everything into account, this election starts Leans Democratic.
Bottom line: At this early stage, the only Republican-held seat likely to flip is the one in North Carolina, though Maine could shift quickly from its toss-up status. And while Republicans will field weak incumbents in Alaska, Florida, Montana, Ohio, South Carolina and Texas, while looking at an open seat in Iowa, each of these states is at least 8.9 points more Republican than the nation (average R+12.3). Thus, a lot would have to go right to flip most of these seats.
Overall outlook. There are three ways the 35 Senate elections in 2026 can go for Democrats. The “right down the middle” scenario, based on Democrats winning the total Senate vote by about three points, is that they hold all 13 of their seats while flipping North Carolina and possibly Maine. The challenge then becomes finding two more seats to flip to regain the majority. The path of least resistance is Brown winning in Ohio, a Democrat like Peltola winning Alaska competitive, and a fractured Texas Republican Party handing Democrats a Senate seat. That is a lot of “ifs,” though it is just barely plausible. In this scenario, then, Democrats gain 0-5 seats.
The “long dark night of the soul” scenario is that 2026 resembles 2020 and 2014: Democratic Senate candidates lose nationwide by 3.3 points. In this scenario, Democrats likely lose the seats in Georgia and Michigan, while struggling to hang on in Minnesota and New Hampshire, where Sununu decides to run. Cooper could buck the trend and flip North Carolina, but more likely in this scenario is a Democratic loss of 0-4 seats.
The “a party can dream” scenario, finally, is that 2026 resembles 2018: Democratic Senate candidates win nationwide by nearly 10 points. In this scenario, Democrats easily win all 13 seats they currently hold and flip Maine (Collins retires) and North Carolina. They also win Alaska (possibly Peltola runs), Ohio and Texas by low single-digits. That already gives Democrats a 52-48 majority. Meanwhile, the races in Florida, Iowa, Montana and South Carolina all head to a recount. Democrats could thus flip as many as nine seats, giving them a 56-44 majority. In this scenario, Democrats gain 5-9 seats.
Splitting the difference gives Democrats a range of 0 to 5 seats gained, with a gain of 1-2 the most plausible. Thus, a bit more than a year from the 2026 elections, Republicans are favored to retain control of the Senate, but with fewer seats.
Until next time…and if you like what you read here, please consider making a donation. Thank you!
[1] Counting Independents King, Sanders and Dan Osborn of Kansas as Democrats. Otherwise, Democrats won all Senate votes cast by 1.0 points.
[2] Counting King and Sanders as Democrats. Otherwise, Democrats won all Senate votes cast by 9.4 points.
[3] 2024=6, 2022=5, 2020=4, 2018=3, 2016=2, 2014=1
[4] Some varaiation of “If the election in your U.S. House district were held today, would you vote for the Democratic candidate, the Republican candidate, a different candidate or are you not sure?”
[5] 2024=1, 2022=1, 2020=6, 2018=4, 2016=1, 2014=5
[6] Using only data from 2020, 2014 and 2018, the difference drops to -2.4, using a 3-2-1 weighting scheme.
[7] Turnout for Trump increased by just 25,288.
[8] In 2020 and 2024, 65% of eligible voters cast a ballot, on average, compared to 51% in Texas.
[9] For what it’s worth, a September poll by the Democratic polling firm Public Policy Polling gave Kleban an 8-point lead over Collins.
[10] The other states where Harris received more votes that Biden are Wisconsin (37, 363), Utah (2,284), Nevada (1,711) and Maine (580). In the other 44 states and DC, Harris averaged 142,537 fewer votes than Biden, led by California (-1,834,460), New York (-625,343), Florida (-614,007), Texas (-423,695) and Illinois (409,022).
Just Bear With Me... 
2 thoughts on “A Wicked Early Look at the 2026 US Senate Elections”