OUR MIDTERMS FORECAST PREDICTS PAIN FOR DONALD TRUMP / THE ECONOMIST

Department of prognostication

Our midterms forecast predicts pain for Donald Trump

Just how bad could things get for the Republicans?

Photograph: Getty Images

The sun rises in the east; the other queue always moves faster; and the president’s party loses the House of Representatives in the midterms. 

Starting with the Democrats’ “thumping” blue wave in George W. Bush’s second term, this pattern has held firm every four years. 

Even Republicans’ weak performance in 2022, derided as a “red ripple”, ended Joe Biden’s trifecta.

And although Donald Trump prides himself on breaking rules, in this respect he looks embarrassingly conventional. 

The Economist’s new statistical forecast of the 2026 Congressional elections gives Democrats a whopping 95% chance of gaining at least the three seats needed to flip the 435-seat lower house. 

More surprisingly, despite a Senate map that looks nearly impregnable on paper, the model estimates that the party has a 46% probability of taking over the upper chamber as well.

Given the unpredictability of elections—as shown by Mr Trump’s own victory in 2016—how can we be so certain? 

Although individual House races are thinly polled and frequently deliver upsets, district-specific surprises usually cancel out across hundreds of concurrent elections.

If a party wins far more seats than expected, that is usually because it wins more votes than expected nationwide. 

And in modern times the national popular vote for the House has generally landed close to estimates based on “generic-ballot” surveys, which ask respondents which party they plan to support in Congress. 

Among decided voters, Democrats now lead these polls by around 53% to 47%.

Polling errors do still occur: in 2020 House Republicans’ share of the major-party vote exceeded the generic-ballot average by 2.8 percentage points, which was worth nearly 20 seats. 

However, adding other types of data can partly reduce this risk. 

The results of special elections for vacant legislative seats, in which Democrats have excelled, have also been broadly accurate, particularly in midterms. 

And the president’s approval rating, now 17 points underwater, provides another glimpse of a blue wave building offshore.

There is plenty of time for these variables to change. 

But generic-ballot polling in midterms tends to be fairly stable, edging towards the party out of power over time as less engaged voters start to pay attention. 

On top of this, the more respondents who have not made up their minds, the worse the president’s party tends to do on election day, implying that “undecideds break for the challenger” is the rare political adage that data actually support.

Control of the House is ultimately determined in individual races, not by a nationwide vote. 

However, unlike in past years when district lines strongly favoured one party, current maps give Republicans only a modest edge. 

Thanks above all to a potent gerrymander of California’s congressional districts in retaliation for Republicans’ gerrymander of Texas, Democrats have so far fought an unprecedented mid-cycle redistricting war to a rough draw. 

And if Virginia voters approve a referendum today that would draw 3-4 more Republicans out of their seats, even this slight bias would vanish (unless the state Supreme Court overturns the new map).

Our forecast of the national popular vote for the House among the two major parties, trained on every Congressional election since 1942, does leave room for surprises. 

The range containing 95% of its simulations runs all the way from 50.7% for the Democrats—a near tie, similar to Republicans’ milquetoast showing in 2022—to 55.6%, which would be the biggest House landslide since 1976. 

But given relatively fair maps, even the bottom of this range would give the party a 50/50 chance of flipping the lower chamber.

Our House forecast is much more confident in a Democratic victory than the 85% probability implied by prediction markets. 

We would not recommend betting on the Democrats solely based on our model, given that punters use many types of information that our forecast lacks. 

Above all, it assumes that current district lines are final. In fact Republican-controlled states, particularly Florida, may enact last-minute gerrymanders of their own. 

And if the Supreme Court overturns the Voting Rights Act’s protection of House districts with large non-white populations, these re-draws could be unusually brutal. 

We also assume that House incumbents awaiting primaries will be re-nominated, and do not incorporate race-specific polls or information about non-incumbent candidates until both major-party nominees are chosen.

A corollary of the House looking like a shoo-in for Democrats, rather than a dogfight, is that a dogfight is unfolding instead in the Senate. 

There, Democrats need to flip four of the 100 seats to gain control—a tall order given that all but two of their potential targets are in states that Mr Trump won by double digits in 2024. 

The party must defend a vulnerable open seat in Michigan, and are only clearly on track for one gain: Roy Cooper, the former governor of North Carolina, leads polls by around six points. 

Although Maine is a light blue state, Democratic primary voters appear poised to forsake Janet Mills, their sitting 78-year-old governor, in favour of Graham Platner, an oyster farmer with a record of impolitic comments about sexual assault and a tattoo containing a Nazi military symbol. 

Such liabilities would provide ample fodder for Susan Collins, the state’s moderate incumbent, who is a proven electoral outperformer.

At the same time, Democrats have scored recruiting coups in two light-red states. 

Sherrod Brown, a longtime incumbent in Ohio, lost in 2024 but ran far ahead of Kamala Harris. 

He is now mounting a comeback in a far friendlier political environment, and polls show him nearly tied. 

Similarly, Mary Peltola, who won a statewide race for Alaska’s only House seat in 2022 but was narrowly defeated two years later, has led Dan Sullivan, the incumbent Republican, in all public polls this year. 

Moreover, Republican primary voters in Texas could well oust John Cornyn, a mainstream incumbent, in favour of Ken Paxton, the state’s attorney general, who faced securities-fraud charges for nearly a decade and was impeached by the state legislature. 

That would open the door to making Democrats’ long-held dream of “Blexas” a reality.

Our model sees Democrats as underdogs in all of these states, and in an open seat in light-red Iowa as well. 

However, any scenario in which they are doing well enough nationally to win even one of these states is probably also one in which they sweep the easier contests as well, putting them on the cusp of the decisive 51st seat. 

If anything, our forecast may be underestimating the Democrats’ Senate chances. 

Prediction markets already make them a narrow favourite to flip both chambers.