Democrats on track to outperform 2026 House forecasts

In this era of rampant online gambling, a fitting way to frame this year’s U.S. House contests is around the odds of Democrats winning a majority. To put it bluntly, those odds are about as good as anyone can ever expect to find. But just how many seats will they flip? The smart money sets the number at around 17. I’d take the over.

The Democrats start five seats down (219-214), with two vacancies in play—Florida’s 20th District and Texas’ 23rd District. One of the two is a deep blue constituency, which effectively trims the deficit to four (Texas’ 23rd District is majority Hispanic and supported Trump by 15 points but looks more competitive now). To regain the gavel, then, the Democrats need just three net wins.

That’s not a big spread to overcome, but the betting card is short, and the GOP does have a couple more cards to play. To elaborate, district-by-district analysts see 381-395 outcomes as “safely” or “likely” squared away, yielding an effective Republican cushion of seven or eight seats (largely because of its redistricting efforts this year, a point to which I will return). That leaves only about three or maybe four dozen districts that are truly competitive. Here’s why I’m wagering that the Democrats will win nearly all of them.

Races and polls

The race-by-race prognosticators are characteristically cautious, anchoring their estimates on district-level factors like partisan balance, incumbency, and candidate quality and discounting national-level dynamics like economic trends and presidential approval. The theory that politics turns on local issues hasn’t been true for some time; the national trends carry outsized importance now.

As such, some forecasters rely on generic ballot polls, which tally vote intentions for each party without mentioning specific candidates. They currently favor Democrats by about 6 percentage points, which may not sound seismic, but midterm voters tend to migrate toward the out-party as campaigns progress. In fact, this century, when an out-party has enjoyed a late-summer lead like this, it has netted between 31 and 63 seats. That big of a landslide is probably out of reach nowadays (because the number of districts with any semblance of partisan parity has cratered). Still, those precedents speak to what a 6-point lead really represents.

Presidential patterns

We can reach the same conclusion, though, without even looking at election polls. It is something close to an ironclad law that the president’s party loses seats in midterm elections—typically more than two dozen of them. In fact, in only three such elections since the Civil War has the president’s party not lost seats: 1934 (reflecting New Deal optimism), 2002 (rallying post-9/11 patriotism), and 1998 (responding to perceived impeachment overreach).

What did those years have in common? A remarkably popular president. Clinton’s approval stood at 66%. Bush’s stood at 63%. And though FDR’s is less clear because scientific polling had yet to be invented, the robust economic rebound of that year connotes broad public support.

Indeed, by contrast, when a president’s approval rating is underwater, his seat loss spikes substantially. Because Trump’s approval (mid/high 30’s) is among the lowest of any president at this point in his term since 1946—when Truman’s Democrats lost 55 seats—history portends an uphill battle. And with gas prices and mortgage rates both up, job growth and consumer confidence both down, and Iran tensions and ICE shootings both unabated, Trump’s numbers are unlikely to rebound anytime soon—certainly not before voting begins in two-to-six weeks.

Truth be told, Trump is probably even less popular than previous presidents were with similar ratings. While dissatisfied Democrats are willing to voice that discontent on surveys, modern-day Republicans tend to “cheerlead” and loathe to give pollsters the supposed satisfaction of hearing their disapproval. But their enthusiasm may still be waning, and there’s no guarantee they’ll cast ballots for down-ballot Republicans. Thus, the projected seat loss from Trump’s unpopularity may be lowballing it.

A big part of the presidential party’s midterm seat loss is about turnout. Unpopular presidents tend to dampen enthusiasm among supporters and boost it among opponents. In Trump’s case, it applies even when he is more popular; the “cult of personality” nature of his appeal draws supporters out of the woodwork when he’s on the ballot but not when he isn’t.

Equally important to the turnout story—and often overlooked among pundits—is the “diploma divide” that has realigned the two parties’ class identities. That upheaval has taken Democrats, as Paul Begala said, “from the factory floor to the faculty lounge,” but it has also taken them to the ballot box. In other words, because (a) education is such a powerful predictor of voting, and (b) Democrats have become the ones who tend to have more of it, they (c) tend to have a turnout advantage in low-visibility elections (non-presidential years) now.

What about …

There are several objections to consider, of course.

First, what about the 2022 midterms, when Democrats performed well despite Biden’s unpopularity? Well, while the Democrats did gain a seat in the Senate, the party still lost nine seats in the House—three times what Republicans can afford this year. And as I mentioned earlier, Biden’s number then did not mean the same thing that Trump’s does now. A meaningful share of Biden’s disapprovers were frustrated Democrats who were never going to vote red. By contrast, relatively few of Trump’s disapprovers are Republican voters, making his disapproval rating more consequential.

But what about the systematic survey bias in favor of Democrats? Swing-state polls overestimated Democratic support in 2016, 2020, and 2024, so why shouldn’t we take the Democrats’ current polling lead with a grain of salt? Because the pattern did not hold in either of the last two midterms: 2018 and 2022 were both relatively accurate polling years; if anything, the 2022 polls underestimated the Democrats’ showing.

So, what about all the radicals the Democrats have been nominating who stand stiff odds of winning over swing voters? That narrative doesn’t hold up; leftist victories have received outsized media attention, but just as many centrists have been winning overall, and in the real House battlegrounds, Democratic primary voters have mostly stuck to mainstream candidates. Even the so-called radicals are not necessarily unelectable. Their leftist rhetoric is usually concentrated on economic issues—reflecting a rising tide of populism among not just Democrats but independents and Republicans as well.

Well, what about the fundraising advantage that Republicans enjoy? Republican Party committees and allied groups currently hold a half-billion-dollar cash edge over their Democratic counterparts. But that top-line gap doesn’t show up where it matters most: In the second quarter of 2026, Democratic challengers outraised Republican incumbents in about two-thirds of the toss-up districts. And even where a real gap does exist, money has diminishing returns; candidates need enough cash to get their message out, but beyond that, those additional millions buy less and less.

Finally, what about the Republicans’ new gerrymanders? Indeed, those mid-cycle redistricting gambits have genuinely hurt Democrats; the median district is about two points more Republican than it was before they redrew the maps. But that damage is already baked into the competitive-seat math that I laid out at the beginning. It’s the reason that Democrats will not gain 50 seats. They are in a pretty good position to gain around 25-30, though, which would amount to a decisive rebuke of the Trump sequel and his many foot soldiers on Capitol Hill.