The road to a Senate majority
runs through rural America.
By now, everyone knows how much is riding on this election. Families are paying more for groceries, health care and electricity while Washington lurches from one round of cost chaos to the next, and the corruption has become hard to keep up with.
Donald Trump heads into Election Day with historically low approval ratings, but that alone won’t hand Democrats a majority. This is still going to be a hard year for Democratic candidates, especially in rural America.
That’s what the national numbers miss. A generic-ballot poll or a Trump approval rating samples the country as a whole, where about 14 percent of people live in non-metro counties (RUCC 4–9). But the Senate counts states, not people, and this year’s states are far more rural: in the average Class II state, more than a quarter of residents (26.5 percent) live in non-metro counties. To win the majority, Democrats need to win 17 of the 35 races on the ballot, the 33 regular Class II races plus special elections in Ohio and Florida. Line those states up by the 2024 presidential vote, and the 17th-best state for Democrats is Alaska, which Trump carried by about 13 points while winning the national popular vote by 1.5. In other words, the seat that decides the majority sits about 11 points to the right of the country as a whole. A national environment that looks good for Democrats can still leave them short in the Senate.
Every two years a third of the Senate faces voters, and the third on the ballot Tuesday, Class II, is the most rural of the three by every measure we ran. That means the road to a Senate majority runs straight through small towns and farm country. The outcome will turn on rural voters, and on the local leaders who have spent the last two years organizing their neighbors and making sure Democratic voters in those communities turn out to vote.
Our findings
- This year’s Senate seats are the most rural on the map. In the 33 Class II states, 18.2% of people live in non-metro counties (RUCC 4–9), against 13.8% in the United States as a whole and 11.3% and 13.4% in the other two classes. Sixteen of those states are at least one-quarter non-metro. Only 19 of the 50 states are, so 16 of the 19 have a Senate race this year.
- Republicans hold 14 of those 16 rural states. New Hampshire and New Mexico are the only two held by Democrats.
- Democrats can win the House in metro America. If every metro county (RUCC 1–3) returned to its 2008 margin, Democrats would win about 230 House seats, even if non-metro counties did not move at all.
- They cannot win the Senate that way. Metro gains alone leave Democrats at 48 seats. Obama’s 2008 margins in every county produce a 50–50 Senate, with Vice President JD Vance breaking ties.
- The 51st seat takes a rural result no Democrat has managed since at least 2008, Obama included. Non-metro counties have to come in near R+11, against Harris’s R+36 and Obama’s R+13. Even then, North Carolina is a dead heat.
What we’re watching
on election night.
On election night, we’ll watch how each county swings from Harris’s 2024 margin, not who carries it. Remember, Democrats will lose almost every non-metro county even on a great night. What matters is by how much.
- Non-metro (RUCC 4–9) counties need to swing about 25 points toward Democrats on average. Harris lost them by 36. The Senate math needs about R+11. A county that swings 15 points is a good story, and it still means we lose the Senate.
- Metro (RUCC 1–3) counties need 4 to 5 points. Harris won them by 4. Obama won them by 11. Below that range, the House gets nervous. Above it, the House is done early.
- North Carolina and Florida count fast. If their non-metro counties come in near the 25-point swing, the Senate is live. If they come in near 10, stop watching the Senate and start defending the House.
Rural voters have more influence in 2026
than in any midterm of the last ten years.
The Senate seats on the ballot this year, Class II, are the most rural of the three Senate classes, and more rural than the country as a whole.
Senators serve six-year terms, but the whole Senate is never up at once. The Constitution split the first senators into three groups, or classes, so that one-third of the seats would be elected every two years. In 1789 the senators drew lots to decide which class each seat joined, and every state admitted since then has had its two seats placed in different classes. Today Class I has 33 seats, Class II 33 and Class III 34.
Each class comes up every six years: Class I was last elected in 2024 and is next up in 2030, Class III was last elected in 2022 and is next up in 2028, and Class II, last elected in 2020, is up now. Because the classes rotate on a six-year cycle and midterms come every four years, the same class lands in a midterm only once every 12 years. The 2018 midterm elected Class I, and the 2022 midterm elected Class III. Class II last faced a midterm electorate in 2014 and won’t again until 2038.
Pick a class to see the states with a Senate seat in it. Every state has two senators in two different classes, so each state lights up in two of the three views. Light blue in the Class II view: Ohio and Florida, which also hold special elections this year for their Class III seats. Hover over (or tap) a state for both of its classes.
Senate classes: U.S. Senate. State boundaries: U.S. Census Bureau, via us-atlas.
To determine which Senate class is most impacted by rural voters, we took every county in the country, labeled it metro (RUCC 1–3) or non-metro (RUCC 4–9) using the USDA’s official 2023 Rural-Urban Continuum Codes, and grouped the counties by which Senate seats their states elect. The group on the ballot this year, Class II, came out as the most rural by all five measures below.
The first three columns are the three groups of Senate seats, by the year they are next on the ballot; the last is all 50 states together. Each row measures how rural the states in that column are. Class II, up this year, is the most rural of the three classes on every row, and more rural than the country as a whole on every share. Non-metro counties are RUCC 4–9; the most remote category is RUCC 9.
Population: 2020 Census. County types: USDA Economic Research Service Rural-Urban Continuum Codes, 2023. A state appears in two class columns because it elects two senators in different classes. United States: the 50 states, not including the District of Columbia. “Average state” is the unweighted mean of each state’s non-metro share.
In the Class II states voting for Senate this year, 18 percent of people live in non-metro counties, almost one in five. In the country as a whole, it is about 14 percent, and in the other two groups 11 to 13 percent, closer to one in eight. That means rural voters make up a much bigger share of the people choosing senators this time than they do of the country, or than they usually do in a Senate election.
The state count matters most, because every state gets two senators no matter its size. A Senate win in Wyoming counts the same as one in California. In the average state on this year’s ballot, about a quarter of the people live in non-metro counties. In the average state in the other two groups, it is closer to a fifth. Sixteen states on this year’s ballot are at least one-quarter non-metro, compared with 10 and 12 in the other cycles. Only 19 of the 50 states clear that bar, so 16 of the 19 have a Senate race this year.
The road to a Senate majority
runs through rural states.
Republicans control the Senate 53 to 47. To take it back, Democrats need to gain four seats overall, and most of the seats they would have to win are in states where rural voters carry real weight.
Two special elections are also on the ballot this year, in Ohio and Florida. Both are Class III seats held by Republican appointees, Jon Husted and Ashley Moody, so they are not counted among the 33 Class II seats above. That makes 35 Senate races in all.
Most of the seats on this year’s ballot are already in Republican hands, 20 of 33 regular seats plus both specials. The rural ones are even more lopsided. Of the 16 states on the ballot where at least one in four people live in non-metro counties, Republicans hold 14. New Hampshire and New Mexico are the only two held by Democrats.
Each bar is a state with a Senate race this year, sized by the share of its people who live in non-metro counties (RUCC 4–9) and colored by the party that holds the seat now. The dashed line marks one in four; the dotted line marks the national share, about 14%. Red bars dominate the top of the chart.
Population: 2020 Census. County types: USDA Rural-Urban Continuum Codes, 2023. Ohio and Florida are special elections for Class III seats. New Jersey and Rhode Island have no non-metro counties.
The races Democrats themselves are focused on show the same thing. The Democratic Senate campaign committee’s $40 million program targets seven seats to pick up: Alaska, Florida, Iowa, Maine, North Carolina, Ohio and Texas. The seats it most needs to hold are in Michigan, Georgia and New Hampshire. Eight of those ten states are more rural than the country as a whole, where about 14 percent of people live in non-metro counties. Only Florida and Texas are not.
Two more rural states could decide the majority, and in both the leading challenger is an independent, not a Democrat. In Montana, Steve Daines dropped his reelection bid at the filing deadline, and Republican Kurt Alme now faces independent Seth Bodnar and Democrat Alani Bankhead. In Nebraska, independent Dan Osborn, backed by the state Democratic Party, is in a close race with Republican Pete Ricketts: a YouGov poll in early October had Osborn ahead 48 to 47, and a New York Times/Siena poll released Oct. 31 had Ricketts ahead 49 to 43.
The ten seats in the Democratic Senate campaign committee’s program, seven to pick up and three to hold, sorted by the share of each state’s people who live in non-metro counties (RUCC 4–9). Nationally, the share is about 14%.
| State | Who holds the seat now | Share of people in non-metro counties |
|---|---|---|
| Maine | Republican incumbent | 40.2% |
| Iowa | Republican, open seat | 38.6% |
| New Hampshire | Democratic, open seat | 37.0% |
| Alaska | Republican incumbent | 32.6% |
| North Carolina | Republican, open seat | 21.0% |
| Ohio (special) | Republican appointee | 18.1% |
| Georgia | Democratic incumbent | 16.7% |
| Michigan | Democratic, open seat | 16.5% |
| Texas | Republican, incumbent lost primary | 9.9% |
| Florida (special) | Republican appointee | 3.1% |
Population: 2020 Census. County types: USDA Rural-Urban Continuum Codes, 2023.
Democrats need a historic performance
in rural America to take back the Senate.
Democrats can take the House in metro America. They will win the Senate only by doing better in rural America than any Democrat has since 2008.
For reference, Obama lost non-metro counties (RUCC 4–9) by 13 points in 2008. Harris lost them by 36 in 2024. Over the same years, the metro (RUCC 1–3) margin slipped 7 points and the non-metro margin fell 23. Most of the damage is rural.
Everything changed in 2016. Democrats went from R+19 in non-metro counties in 2012 to R+34 in 2016, and no Democrat has climbed back since. The Senate number we need is better than any Democrat has managed since at least 2008, Obama included. Getting there takes a swing of about 25 points from Harris’s margin in non-metro counties. That would be more historic than Trump’s rural wave in 2016, which moved those counties 15 points.
The Democratic presidential candidate’s two-party margin in metro counties (RUCC 1–3, navy) and non-metro counties (RUCC 4–9, red), each year on that year’s own turnout. Above zero, Democrats won the group; below zero, Republicans did. The dashed line is the non-metro margin a Senate majority takes this year.
Two-party margin, (Democratic − Republican) ÷ (Democratic + Republican), summed over each group’s counties. Sources: county presidential results 2008–2024; 2012 and 2016 certified returns from the MIT Election Data and Science Lab.
The House runs through the suburbs. The Senate runs through rural America.
To see what each chamber takes, we started every 2026 race at the 2024 presidential vote and moved counties back toward their 2008 results. If every 2026 race tracked the 2024 presidential vote, Republicans would keep both chambers: 46 Democratic senators and 205 Democratic House seats. Democrats start 13 House seats short.
If every metro county (RUCC 1–3) returned to its 2008 margin, Democrats would win about 230 House seats. On today’s turnout, that adds up to D+8 across metro counties, about 4 points better than Harris did. The same move gets the Senate to only 48. When we put every county in America back at its 2008 margin, we reach a 50–50 tie in the Senate, and JD Vance breaks the tie. The 51st seat requires non-metro (RUCC 4–9) margins better than Obama’s.
Each row moves counties from their 2024 presidential margins back toward 2008. The margin columns show where that leaves metro (RUCC 1–3) and non-metro (RUCC 4–9) counties as a whole on 2024 turnout. The seat columns count Democratic wins; Democrats need 51 senators and 218 House seats. Only the last row wins both.
Senate: each state’s 2024 presidential margin plus the average shift of its counties, weighted by 2024 votes, plus the 34 Democratic seats not on the ballot. House: each district’s 2024 presidential margin on the 2026 lines plus the shift of the counties its voters live in. Try other scenarios with the predictor on The Rural Desk’s home page.
The seats that decide the majority
The table below follows every Senate race that changes hands in at least one of those scenarios, plus Montana, which comes closest without flipping. Michigan and Maine are not close once counties move toward 2008. Iowa, Ohio and Florida flip as soon as every county is back at its 2008 margin.
North Carolina is the 51st seat, and it is the last to move. It stays Republican with every county at its 2008 margin and turns only when non-metro counties do 2.5 points better than that.
Each row is a Senate race that changes hands in at least one scenario, plus Montana. The three margin columns show the 2024 presidential result in the state and our predicted margin in the last two scenarios above. Blue: Democrat ahead. Red: Republican ahead.
Predicted margins come from the presidential vote alone, so they say nothing about candidates. In Montana and Nebraska, the leading challenger is an independent; the Democratic presidential margin stands in for them.
Midterms punish the president’s party, not Democrats
Since 2008, Democrats have done better in a midterm than in the presidential year before it only once: 2018, the one midterm with a Republican in the White House. In 2010, 2014 and 2022, with a Democrat as president, their national House margin fell 6 to 17 points from two years earlier.
That is the midterm penalty, and it falls on the president’s party. In the 20 midterms since 1946, the president’s party lost House seats in all but two, 1998 and 2002. This year a Republican is president, so history favors Democrats the way it did in 2018, when their national House margin improved by almost 10 points.
Each bar is a midterm, measured against the House vote two years earlier, the last race for the same office. Blue: Democrats gained ground. Red: Democrats lost ground. The president at the time is in parentheses. The only gain came under a Republican president. Hover over a bar for the margins before and after.
Points of national margin, Democratic minus Republican share of all U.S. House votes, midterm year minus the presidential year before it. Source: Brookings Institution, Vital Statistics on Congress, Table 2-2.
Even a 2018-sized swing doesn’t reach the Senate number. Our scenarios need non-metro (RUCC 4–9) counties to swing about 25 points from Harris’s margin. A national swing like 2018’s is well under half of that, so Democrats would need rural counties to swing far more than the country does.
The map is hard.
Local Democrats are ready for it.
There’s no doubt about it. The math is hard for Democrats this year. Taking back the Senate is a massive challenge, and it means doing better in rural America than any Democrat has since 2008.
But all across the country we’ve met hundreds of vocal locals working day and night to elect candidates who will put an end to the extraordinary costs, chaos and corruption we’re seeing from MAGA Republicans.
The map is hard, but local Democrats are ready for it. They have spent the last ten years building back, and they are better organized now than at any point since the early 2000s. That strength comes from county parties that meet all year, from neighbors who run for school board and county commission, and from people who knock doors in towns where everyone knows their name.
Just looking at the math, winning back the House while falling short in the Senate would still make this a historic year. And when the results come in, we need to celebrate cutting margins in rural America just as much as we celebrate the counties that flip.
How we did this
Definitions
- County types. USDA Rural-Urban Continuum Codes (RUCC) 2023: metro (RUCC 1–3) and non-metro (RUCC 4–9). The most remote category is RUCC 9: non-metro counties with an urban population under 5,000, not adjacent to a metro area. “Rural” in this report means non-metro.
- Senate classes. Class I is next elected in 2030, Class II in 2026 and Class III in 2028. Ohio and Florida hold special elections this year for Class III seats; they count as 2026 races in the scenarios but not in the Class II rurality figures.
- Margin. The two-party margin, (Democratic − Republican) ÷ (Democratic + Republican). R+11 means the Republican led by 11 points.
Senate-class rurality
Each county’s 2020 Census population is assigned to its state, and each state to its two Senate classes. The share of counties counts every county in a class’s states; the share of people sums their population; the average state is the unweighted mean of each state’s non-metro share.
The scenarios
The scenarios use the same model as the predictor on The Rural Desk’s home page, at its default settings. Every county moves in a straight line from its 2024 presidential margin to its 2008 margin; metro (RUCC 1–3) and non-metro (RUCC 4–9) counties move separately. The last scenario carries non-metro counties 2.5 points past their 2008 margins, the far end of the predictor’s non-metro slider. A state’s Senate result is its 2024 presidential margin plus the average shift of its counties, weighted by 2024 votes. A House district’s result is its 2024 presidential margin on the 2026 lines plus the shift of the counties its voters live in, weighted by each county’s votes and the share of its residents in the district. Of the Senate seats not on the ballot, Democrats hold 34 and Republicans 31; Democrats need 51 because the vice president breaks ties, and 218 House seats.
Data sources
- County presidential results, 2008–2024. The Rural Desk’s county results files; 2012 and 2016 replaced with certified county returns from the MIT Election Data and Science Lab.
- Midterm turnout check. 2022 U.S. House votes by county, MIT Election Data and Science Lab precinct returns; where a county’s total is missing or implausible, as in uncontested races, estimated from ballots cast in the U.S. Election Assistance Commission’s Election Administration and Voting Survey.
- House districts. 2024 presidential results by 2026 district, The Downballot; 2020 Census tract populations placed in Census TIGER/Line 2026 district boundaries.
- Rurality and population. USDA ERS Rural-Urban Continuum Codes, 2023; U.S. Census Bureau, 2020 Census.
- National House vote, 2008–2024, and midterm seat changes since 1946. Brookings Institution, Vital Statistics on Congress, Table 2-2. Each midterm is compared with the House vote two years earlier.
Limitations
- The presidential vote is a stand-in. The scenarios move partisanship, not candidates. Senate and House candidates run ahead of or behind the top of the ticket, and this year several of the closest races feature independents or open seats.
- Counties within a group move together. Every metro county moves the same share of the way back to its own 2008 margin, and so does every non-metro county. Real elections are uneven.
- Missing history. Alaska, Connecticut and Oglala Lakota County, South Dakota, lack comparable older county results and stay at their 2024 margins.
- Not a forecast. The scenarios use no polling. They show what each chamber takes, not what will happen.