Who will win the 2026 midterm elections? Polls and predictions

Democrats have not led a generic congressional ballot by this much in nearly two decades. The Silver Bulletin polling average broke D+6.1 in early May — the widest margin recorded at this stage of any midterm cycle since 2008. Five months out from November, the raw numbers look devastating for Republicans.

But polls do not win seats. Maps do.

Every forecast model this cycle faces the same question: not whether Democrats are ahead — that data is unambiguous — but whether a six-point national lead translates into the 218 seats needed to flip the House. Nearly every quantitative model says yes. The margin is narrower than the topline suggests. The variables are more unsettled than in any midterm since the Supreme Court dismantled the preclearance framework a decade ago.

Nate Silver’s model puts the Democratic probability of a House majority at roughly 85 percent. The Economist’s forecast assigns Democrats a 19-in-20 chance. Prediction markets on Polymarket are pricing Democratic House control at 78 percent. These are not marginal probabilities. They describe an environment where the baseline expectation is a flip. The gap between the 78 percent the market assigns and the 95 percent the statistical models project is itself telling: traders are pricing in something the models cannot fully capture.

The structural variables that could cap Democratic gains fall into three buckets. None are small.

The first is redistricting. Republican-controlled state legislatures have redrawn maps since 2022 that net the party an estimated 5 to 14 additional safe seats, depending on how courts rule on outstanding challenges in Louisiana, Georgia, and Alabama before November. The Supreme Court’s decision in Louisiana v. Callais last term gutted the remaining teeth of Section 2 of the Voting Rights Act, allowing states to eliminate majority-minority districts that had protected Democratic incumbents across the South. Legal experts tracking the fallout at the Brookings Institution estimate the ruling alone could shift the median-seat threshold by two to three points toward Republicans — meaning a D+6.1 environment behaves more like D+3 or D+4 when mapped onto actual districts. Democrats need something closer to a seven- or eight-point national lead to feel confident about a governing majority.

Redistricting narrows the lane. Turnout narrows it further.

Donald Trump’s job approval sat at 37 percent in the NPR/PBS News/Marist survey conducted April 27-30. The lowest of his presidency. Lower than any first-term midyear reading for a sitting president since Jimmy Carter. Gas prices — the single most politically combustible consumer metric — are straining household budgets for 81 percent of Americans in the same poll. The Iran-Hormuz disruption that drove crude past $95 a barrel in April has not yet abated. Pump prices have eased slightly in May, but the administration has no short-term lever to bring them down further. Historically, when the president’s party loses the economic-approval question by double digits, the midterm is a wave.

Yet the Republican base is not behaving like a party bracing for a wave. Internal GOP polling shared with CNN earlier this month shows Trump’s approval among self-identified MAGA voters holding between 70 and 95 percent, depending on the survey. These voters are not defecting. The risk, Republican strategists have told reporters at NPR, is that they stay home. Midterm turnout among the party out of power is historically driven by anger at the president. When the president is from your own party, the intensity flips. Without Trump’s name on the ballot, GOP operatives are privately forecasting an enthusiasm gap of 8 to 15 points relative to 2018 and 2020. A depressed Republican base that still votes is better than a fired-up one. A depressed one that does not vote at all is a disaster that no gerrymander can fix.

Democrats face a mirror-image problem. The party’s grassroots is energized — the NPR/Marist poll found Democratic enthusiasm for the midterms running at its highest level since 2006 — but the coalition that delivered the popular-vote margin in 2024 was fragile. Suburban women, younger voters, and Black turnout in urban districts all underperformed relative to registration targets in that cycle. None of those groups can be taken for granted in a midterm. Democratic strategists who spoke to UnHerd and The Bulwark in recent weeks have warned against reading the D+6.1 number as a guarantee. “Complacency is the only thing that can lose this,” one senior House Democratic aide told The Bulwark on background — a sentiment echoed across the caucus.

Elaine Kamarck, a senior fellow at Brookings who has modeled midterm outcomes across six decades, put the Republican position in plain terms: “Barring unforeseeable game-changing events during the next 14 months, the probability that Republicans will lose control of the House is very high.” The conditional — “barring unforeseeable events” — is worth pausing on. Fourteen months is a geological age in American politics. In the equivalent window before the 2022 midterms, the Supreme Court overturned Roe v. Wade, inflation hit 9.1 percent, and gas prices spiked past $5 a gallon. The models did not price any of those events in May 2022. They cannot price the 2026 equivalents now.

Then there is the Senate, where the arithmetic cuts in the opposite direction. Democrats are defending seats in Georgia, Michigan, and New Hampshire — all states Trump carried or nearly carried in 2024 — while the Republican map is almost entirely red-state terrain. Even in a wave environment, Democrats face a plausible ceiling of 50 or 51 seats. Control of the upper chamber is a narrower, more contingent question than the House. The models reflect it: The Economist gives Democrats roughly a 60 percent chance of holding the Senate majority; Silver Bulletin puts it slightly lower. Neither number is comfortable for the party in power. If Republicans hold the Senate while losing the House, the legislative consequence is paralysis. The confirmation consequence — judges, cabinet officials, independent agency heads — is a Republican Senate’s to control. That alone makes the upper-chamber fight strategically consequential even in what looks like a Democratic year.

What the polling cannot measure is whether the 2026 electorate will look more like 2018 — when Democratic turnout surged and the party gained 41 House seats — or 2022, when the expected red wave collapsed because independent voters broke against the GOP’s most Trump-aligned candidates in swing districts. Candidate quality, more than any polling aggregate, may determine whether D+6.1 translates to 225 Democratic seats or 215. That variable no model can capture five months before the ballot is set.

The direction of travel is clear. The magnitude is not. A Democratic House majority in the 116th Congress was built on a D+8.6 popular-vote margin that produced 235 seats. The current polling points to a smaller national margin, applied to a more gerrymandered map, yielding a narrower majority. Domenico Montanaro, summing up the NPR/PBS News/Marist findings, captured the dynamic: “Democrats lead by 10 points on the congressional ballot test” — a number that, if it holds, rewrites the historical midterm playbook. History says the president’s party loses ground. The models says it loses the House. The only open question, six months out, is whether it is a rout or a squeaker.