Republicans are acting like there’s a Blue Wave coming

Could Democrats retake the House ... *before* the midterms?

The Republican party thought they’d won a slim majority in the U.S. House of Representatives (218 to 216)¹ in the 1930 midterm elections. The 72nd Congress opened thirteen months later — this was before the 20th Amendment moved Congress’s start date to January. But by then, the breakdown of the chamber was 219 to 213 in favor of Democrats. How come? Fourteen Representatives-elect died between November 1930 and December 1931. And Democrats flipped three seats in the subsequent special elections — enough to take control of the House before the 72nd Congress even started.

We don’t have 13-month periods before Congress opens anymore. But it’s still technically possible for special elections to flip control of the House between general elections. Is it likely? No. The majority party usually has enough seats to withstand a few special election losses. But narrow congressional majorities are becoming more common, and Republicans’ current House majority is the smallest in modern history. So could Democrats retake the House before the 2026 midterms?

Right now, Republicans hold 218 House seats — the minimum needed for a fully seated majority, though votes for the speakership aren’t affected by absences — and Democrats have 213. There are four vacancies: two from Democratic representatives who passed away in March and two in Florida from Mike Waltz and Matt Gaetz. That’s a 50/50 split between open Democratic and Republican seats, but Republicans have consistently underperformed in special elections post-2024. Last week, for example, Republicans lost a Pennsylvania State Senate seat that Trump won by 15 points in 2024. So the Democrats should have no trouble holding the two seats they’re defending — those districts voted for Kamala Harris by 40 and 22 points in the 2024 presidential election.

What about Florida? Well, neither seat should be competitive or even close. Trump won Florida’s 1st congressional district by 37 points in 2024 and its 6th district by 30. But so far, things are much tighter than those numbers would suggest. The early vote in FL-1 is R +19 compared with R +40 in 2024; in FL-6, it’s R +9 versus R +25 in 2024. Fundraising is also lopsided in favor of the Democrats for both races. Josh Weil — the Democratic candidate in FL-6— has raised more than 10 times as much as Randy Fine, the Republican. (Although Elon Musk’s super PAC is trying to even the odds.)

Now, Florida’s 1st is still almost certainly a safe Republican hold — although the Republican margin could shrink dramatically relative to 2024. But the 6th district could just perhaps shift enough to give Democrats a fighting chance. Beyond the spending difference, lagging Republican early turnout, and a Libertarian candidate in the race, Fine is not a good candidate. If you don’t want to take our word for it, listen to Florida Governor Ron DeSantis:

Regardless of the outcome in that, it’s going to be a way underperformance from what I won that district by in 2022 (as a candidate for re-election as Governor) and what the President won it by in November. They’re going to try to lay that at the feet of President Donald Trump. That is not a reflection of President Trump. It’s a reflection of the specific candidate running in that race.

Special elections are hard to poll accurately, but the polls that have been released also look bad for Fine. A survey from St. Pete Polls had Fine leading by only 4 points (48 percent vs. 44 percent). And in an internal Republican poll from Fabrizio Ward, Fine trailed Weil by 3 points (41 percent vs. 44 percent).

Now if we assume Democrats win the two special elections they are defending and Republicans hold Florida’s 1st, that would bring the composition of the House to 219 R vs. 215 D. That razor-thin margin would get even thinner if Fine loses in the 6th district — which we can’t stress enough, is still unlikely. But even if that doesn’t happen, how likely is it that three more Republican seats open up and are won by Democrats before the 2026 midterms?

How common are special elections — and why do they happen?

The first question is: how many special elections typically happen in between general elections? Well, in interregnums between general elections from 1973 through 2024, an average of nine special U.S. House elections took place. Here, I’m counting all special elections, including those that ended up falling on Election Day in a general election year or featured two candidates from the same party. It makes sense to exclude these races when analyzing the possibility of party flips. But for now, let’s just focus on how frequently seats open up.

So far, we only have four special elections scheduled for the 2025-2026 cycle — well below the historical average. In fact, only one cycle in our dataset had fewer than four special elections (1999-2000). Of course, there’s still basically a year-and-a-half to go. So we might expect more vacancies before November 3rd, 2026. But there’s a lot of variance cycle-to-cycle. We could see a deluge of special House elections this cycle like we did between 2016 and 2018 (there were 14). Or it could look more like 2002-2004, when there were only 5.

Next question: What can we say about why special elections occur? To figure that out, we coded the cause of each vacancy associated with each special election shown above. Since 1973, about 35 percent of special House elections have been triggered by the death or illness of a member of Congress. And another 34 percent happened because House members took new jobs — they joined the executive branch, were appointed to a Senate seat, took a job in the private sector, etc.

Scandals are the next-most frequent cause of vacancies, representing 16 percent of special elections. The rest were either due to legal or procedural issues (overturned results, mid-cycle redistricting) or other reasons (personal reasons, a sick family member, etc.).

Now some of this does vary by cycle — cycles after a new president takes office tend to have lots more appointments, for example — but a good 85 percent of special House elections are caused by deaths, new jobs, or scandals. So what does that tell us about how many more special elections we’ll have in this cycle?

Pulling Elise Stefanik’s nomination is what we call a “tell”

Well, we’re unlikely to see many more Republican vacancies due to members taking new jobs, in part because Republicans recognize their predicament. Last week, Donald Trump withdrew Elise Stefanik’s nomination to be UN ambassador, ostensibly so she could stay in the House and bolster the slim Republican majority. Now, Stefanik is from an increasingly red district in Upstate New York: Trump won it by 21 points in 2024. If the White House is worried about a special election there, it’s safe to say that they won’t be nominating too many House members to the executive branch before 2026. And that same logic holds for Republican members taking other jobs, resigning because of scandals, or for other personal reasons. Trump is lenient about personal indiscretions, and House Republicans will heavily discourage members from leaving for any reason, including scandal.

So that leaves death or illness as the major potential cause of special elections before the 2026 midterms. It’s a bit morbid, but it's something that’s fair to consider from an actuarial standpoint. The median age in the U.S. House is 57.5 years old.

Now it’s not easy to predict when people will die or fall ill. But Republican members of the House are slightly younger than Democrats (59.2 years old versus 59.7, on average). And older members tend not to be in competitive seats, though as we’ve seen, the definition of “competitive” can expand when there’s a big backlash against the incumbent party.

Since 1973, the average number of House special elections caused by deaths per cycle is about three. But a member passing away doesn’t necessarily equal a flip. First, the open seat would have to be held by a Republican. Second, it would have to be flippable. Only 9 of the 29 Republican seats the Cook Political Report rates as competitive are held by members aged 60 or older.

So there’s a big element of fate here. If you exclude vacancies caused by new jobs and scandals, the average number of special House elections per cycle drops to 4.7. So unless Republicans start feeling more confident about their chances and open up special elections for other reasons, a switch in chamber control before the midterms is unlikely but not impossible.²

What to watch for on Tuesday

But Tuesday’s special elections in Florida can still provide useful signals. First, watch for the degree to which the Democratic candidates overperform. If they significantly close the gap with Republicans, the chance of the House flipping — conditional on enough Republican seats opening up — is presumably higher. And second, if Fine loses — again, this probably won’t happen — then Democrats will be one seat closer to retaking the majority.

Plus, every lost seat will still make legislating harder for Trump. And, of course, the results might provide a hint about what might happen next November. Relative to other analysts, Nate and I err on the side of being skeptical about the forecasting power of special election results, mostly because they often feature lopsided turnout that won’t be repeated in general elections (in recent years, Democrats have particularly benefited from these lower-turnout contests).

But pulling Stefanik’s nomination is an acknowledgment of political gravity for a White House that often seeks to defy it. Midterms usually go badly for the president’s party; Trump’s approval rating has already flipped negative, and Democrats will probably have an enthusiasm advantage.

The lack of competitive seats because of gerrymandering significantly curbs Democrats’ upside. Still, let’s just throw this as a bone out there for our Democratic readers: there are currently 98 Republican House members holding seats in districts that Trump won by 21 points or fewer — his margin in Stefanik’s NY-21.³ Most of these seats are unlikely to flip, or might require the sort of “extra element” like a scandal that would ordinarily trigger a resignation (and special election) if the GOP majority weren’t so thin. But they represent the outer boundary of potentially competitive seats in the event of a massive backlash against Trump. We’re not about to start forecasting next year’s midterms yet: let’s wait (at least) a year for that. But Republicans are acting like there’s a blue wave out on the horizon.

1. There was also one Farmer-Labor Party member who planned to caucus with the Democrats.

2. Another caveat here is that about 20 percent of House special elections between 1973 and 2024 were held on Election Day in general election years. So those technically couldn’t give Democrats control of the House before the 2026 election.

3. This includes 3 districts that Harris won in 2024.