Democrats are still likely to flip the House, but it's close

GOP-led redistrictings are shrinking Dems' error margins

Predicting midterm election outcomes is a tricky business. They don’t follow the usual models or polling forecasts as closely as presidential elections tend to, polling tends to be spottier, and it can be hard to pick up on an important trend among voters until it’s too late. But my look at some predictions suggests that Democrats are still likely to take control of the House this November, even while Republicans shrink their margin for error through aggressive redistricting.

G. Elliott Morris estimates that, in the wake of last week’s ruling by the Virginia Supreme Court invalidating that state’s recent redistricting, Republicans now have a nationwide 6-ish net seat boost thanks to their redistricting campaign over the past year. As he notes, there’s some uncertainty with this estimate and it could be better or worse for Democrats in the end. How does that figure into our expectations for the election?

Race To The White House has a helpful forecast model based on polls, previous election history, fundraising, and more. Their latest expects Democrats to pick up 19 seats. Now, that forecast was done after the April Virginia redistricting, which gave Democrats four new favorable districts, but before last week’s ruling overturning it. So let’s say it’s a forecast of 15 pickups for Democrats. That’s still well in excess of the 3-4 seats Democrats need for a majority, but it’s closer.

The Cook Political Report doesn’t quite do forecasts this way. But by their numbers (after last week’s Virginia ruling), 208 districts are rated as “solid Democrat,” “likely Democrat,” or “lean Democrat,” meaning that Democrats would need to pick up 10 of the 18 toss-up districts to have a majority. If we split the toss-ups down the middle (which is not necessarily how things will go), that gives Democrats an expected 217 seats, or a gain of roughly two or so based on how several House vacancies go. This puts the House on the knife’s edge.

The generic ballot polling question — just asking people whether they want Democrats or Republicans to win control of the House — currently has a five-point Democratic margin. (Nate Silver has it about a point higher.) That’s consistent with a Democratic gain of one to two dozen seats. (Although there’s some debate as to why the Democrats’ margin in this polling question is just five percentage points given how unpopular President Trump is.)

Now, for what it’s worth, I did a “fundamentals” forecast model of midterm elections that relies upon economic growth (real disposable personal income, to be specific), presidential approval, and the number of seats the president’s party currently holds, drawing on data going back in 1950. That is, it ignores stuff like current House polling, fundraising, redistricting, polarization, war, etc. (Fundamentals-based forecasts tend to do better than ones based on early polling, although that wasn’t true in 2022.) I won’t get into the weeds on that model here (although please see my earlier Tusk post on the subject if you’re interested), but the relevant income growth measure came in pretty weak, growing just 0.4% over the past year. All that translates to the Democrats picking up 46 House seats.

This is a much larger gain for Democrats than most people expect and that polls predict. Historically, it’s not that unusual — it’s on par with the losses by the president’s party in the midterm elections of 1958, 1966, and 1974, and considerably less than Democrats lost under Bill Clinton in 1994 and Barack Obama in 2010. But, yes, if my forecast is right, quite a few people, including me, will be surprised. It’s wildly out of line with other predictions, but I’m reporting it here anyway because science or something.

Okay, that’s a lot of forecasting for you. Let’s try to put it all together:

Importantly, this is a wide range for the forecasts, running from Democrats remaining in a narrow minority to them having a substantial majority. With the exception of the Cook Report one, all expect Democratic seat gains well in excess of the 3 or 4 seats they need to obtain a majority. So yes, it’s closer, but the smart money is still on Democrats taking the House — the question is the size of the majority they’ll have to work with.¹

Now, there’s undoubtedly more tumult to come, and there’s plenty of reason to think that the way we’ve predicted elections in the past just won’t work in 2026 given everything the parties are doing to change the state of play and given Trump’s stated plans to reduce turnout among poorer voters and people of color. So feel free to toss out anything I’ve written above. But it’s what I’m going with.

1. I don’t place too much stock in betting markets, but for what it’s worth, the folks at Kalshi still give the Democrats a 75% chance of taking the House