The 2026 Midterms, 100 Days Out: What the Polling and the Q2 Money Filings Actually Show

The Bank Account That Doesn't Match the Candidates

On June 30, 2026, the Republican National Committee reported $128.5 million in the bank. The Democratic National Committee reported $16.3 million — and it owed $18.5 million on top of that, meaning its true position was negative^[4]. By that measure alone, Republicans hold almost every money advantage there is.

Now look one level down, at the same set of filings. Democratic House candidates outraised 14 of the 17 Republican incumbents sitting in the most competitive districts in the country^[5]. Five of those Republicans — Reps. Miller-Meeks, Nunn, Edwards, Bresnahan and Perry — now have less cash in their own account than the Democrat trying to beat them^[5]. Both facts are true. Both come from filings covering the same three months.

That contradiction is the honest starting point for where the 2026 midterms stand, about 100 days out. Voters are telling pollsters they lean Democratic, and have been for months^[1]. Republicans are sitting on a much bigger overall campaign war chest^[4][9]. Neither fact cancels the other. Understanding why requires knowing exactly what each number is actually counting.

Which Pot of Money Are We Talking About?

"Money" in a congressional election isn't one number. It's at least five separate pots, and each tells a different story. There's the national party committee (RNC vs. DNC), the congressional campaign committee that funds House and Senate races directly (NRCC vs. DCCC, NRSC vs. DSCC), the candidate's own account, outside super PACs, and small-dollar platforms like ActBlue.

On the congressional committees, the gap that looked huge at the national level nearly disappears. The NRCC reported $92.7 million to the DCCC's $79 million — a difference of under $14 million^[4]. The NRCC calls this a record quarter and a "massive" advantage in its own press release^[8]. Fifteen percent is a real edge. It's not the eightfold gap the national committee numbers suggest.

Outside money does tilt Republican, and by a lot. Senate Leadership Fund, a Republican super PAC, held $238.6 million — the largest war chest on record for any super PAC^[9]. Its Democratic counterpart, Senate Majority PAC, held $126.5 million^[9]. Congressional Leadership Fund on the House side reported $141.8 million, roughly double House Majority PAC's total^[9]. Democrats, meanwhile, lead on small donors: ActBlue processed $586 million in the quarter, a record for a midterm, from 15 million gifts averaging $39 each, with 573,000 new donors^[12]. That money flows straight to candidates — which is exactly the account where Democrats currently lead.

A widely shared conservative talking point turned this into a single ratio: Republicans have $677 million to Democrats' $136 million, a roughly 5-to-1 advantage capable of "erasing" the Democratic polling lead^[13][32]. That number is built by adding MAGA Inc., a pro-Trump super PAC worth roughly $400 million, to the RNC, NRCC and NRSC — while counting only the DNC, DCCC and DSCC on the other side, with no Democratic super PAC included at all^[33][4]. Senate Majority PAC alone holds $126.5 million that the comparison leaves out^[9]. It also leaves out candidate committees entirely, which is where Democrats are ahead^[5][10]. None of that makes the underlying Republican committee advantage fake. It just means the viral version of it is comparing different things on each side of the ledger.

There's also a finding that rarely makes it into money coverage at all. Political scientists Joshua Kalla and David Broockman pooled 49 field experiments on campaign ads and contact in general elections. Their best estimate of the persuasive effect on how people actually vote is close to zero^[14]. Money still matters — it pays for staff, organizing and turnout operations. It just means a cash ratio isn't the same thing as a vote ratio, in either direction.

What Voters Are Actually Saying, and Why the Numbers Don't Agree

The "generic ballot" is a simple poll question: which party's candidate would you vote for in your district? It's one of the oldest tools for reading a midterm's mood. Right now it favors Democrats — but by how much depends entirely on who's asking.

Nate Silver's polling average had Democrats up 6.1 points on July 24^[1]. Pew Research, surveying registered voters July 6–12, found a similar 43-37 Democratic edge, with 20% still undecided^[2]. Fox News, polling July 17–20, found Democrats up 53-46^[16]. Emerson College, polling July 19-20, found them up 53-42, an 11-point spread^[15]. That's a range of 6 to 11 points, in the same month, on the same question. No single poll is "the" state of the race.

President Trump's approval rating is the more stable number. Mid-July averages put it near 39% approve, 57% disapprove^[3]. Pew found 34% approve, 64% disapprove — essentially unchanged since April^[2]. Approval among independents has fallen to 34%, down from 48% at his inauguration^[3]. This matters because approval is the single measure most closely tied to how many seats a president's party loses in a midterm^[20].

History gives that a hard edge. The president's party has lost House seats in 38 of 41 midterms since the Civil War, and in 20 of the last 22 since 1938^[6][20]. The two modern exceptions both had unusual causes: Bill Clinton's 66% approval during impeachment in 1998, and the post-9/11 surge behind George W. Bush in 2002^[20]. Neither condition exists today. Gallup finds an average loss of 37 seats when a president's approval sits under 50%, but the range is enormous — anywhere from 8 to 63 seats since 2000, depending on how bad the environment actually gets^[6][20].

One indicator has looked unusually strong for Democrats: low-turnout special elections. Across 39 state legislative special elections since 2025, Democratic candidates ran a median 10.4 points ahead of the 2024 presidential result in the same seat, outperforming that baseline in 34 of 39 races, and Democrats have flipped 30 seats to zero for Republicans^[18]. That's one of the better-regarded midterm predictors. But it's measured in races where only a few thousand people vote — a small, highly engaged slice of the electorate — and similar overperformance has historically compressed as Election Day gets closer^[18][19]. Whether that edge holds when turnout is ten times higher in November is simply not knowable yet.

There's a deeper version of that same uncertainty. Over the past two cycles, the two parties have partly swapped voters: Republicans have gained lower-propensity voters who tend to skip midterms, while Democrats have gained higher-propensity voters who show up for everything^[2][19]. If that holds, it could mean Democratic midterm strength is now structurally larger than old history suggests. Or it could mean special elections and other low-turnout data are systematically overstating Democratic strength for November. Nobody has a 2026 turnout file yet, so neither read can be confirmed or ruled out.

The Map Nobody Voted For

Even a clear national mood doesn't translate directly into seats, because the district lines themselves have moved. Since 2025, mid-decade redistricting — redrawing maps between census counts, once rare — has become a routine partisan weapon. Texas, Missouri and North Carolina redrew their maps for Republicans, worth up to seven seats combined; California countered with a voter-approved map worth up to five seats for Democrats^[7][26]. NBC News reports analysts see a realistic net Republican gain of five to seven seats once both sides are weighed^[7].

A Supreme Court ruling helped open that door further. On April 29, 2026, the Court's decision in Louisiana v. Callais narrowed when a district can be challenged as illegal racial vote dilution under the Voting Rights Act's Section 2, and it separately found that Louisiana's own second majority-Black district was itself an unconstitutional gerrymander^[25]. That combination makes it easier for states to eliminate majority-minority districts going forward, with more lawsuits and possible map changes still working through the courts^[25][26]. Not every state has moved the same way — Indiana Republicans declined to redraw at all^[7].

Republicans currently hold the House by only about three seats, so the districting fight matters enormously on its own^[7]. On the Senate side, Democrats need a net gain of four seats to take control. Of the 35 seats up in 2026, 23 are held by Republicans, but most of them sit in states Trump carried, which is why race raters like Cook Political Report and Sabato's Crystal Ball still gave Republicans an edge for Senate control in July, even after moving Alaska and Ohio to Toss-up and North Carolina to Lean Democratic^[21][22][23].

This is the strongest version of the case that structure will blunt whatever the polls are showing. But it has a real weak spot. Maps drawn for maximum efficiency in a normal year can backfire badly in a genuinely bad one — spreading a party's voters thinner across more districts wins more seats when conditions are neutral, and loses more of them when conditions turn against you. NBC News notes several of the newly drawn Republican seats remain competitive despite the redraw^[7]. And with litigation still unresolved in several states, the final net number for 2026 isn't fixed yet^[26].

Two Honest Scorecards

Strip away the spin, and each party's position looks less like a single story and more like a genuine split ledger.

Republicans hold real structural and financial advantages: the national committee cash gap, the outside-group money edge, and a redistricting map built to their benefit^[4][7][9]. They're also facing real headwinds: an approval rating stuck in the high 30s, a generic ballot deficit that has held for months, and a base measurably less engaged than the opposition — Pew found 39% of Democrats had thought "a lot" about the election, against 22% of Republicans^[2][1][3].

Democrats hold their own real advantages: a consistent generic ballot lead, a record small-dollar quarter, a candidate-committee fundraising edge in the most competitive races, and a wave of special-election overperformance^{[1][12][5][18]}. They also carry real weaknesses. The DNC's own balance sheet is negative once its debt is counted^[4]. Congress as an institution is deeply unpopular — Pew found just 25% viewed it favorably, versus 72% unfavorable — and that unpopularity doesn't spare the party currently out of power^[2]. Contested primaries are producing insurgent nominees who test well with the base but are unproven in swing districts; Brookings' own analysis finds those insurgents winning mostly on economic arguments rather than cultural ones, which cuts differently depending on the district^[29][28].

One more piece complicates the Democratic side of that ledger. Some analysts argue that the party's strong polling numbers reflect Trump's weakness more than any actual repair to the Democratic brand, since Democratic favorability hasn't caught up to the generic ballot lead^[28]. Whether that distinction matters in November is exactly the kind of thing that can't be settled from a July snapshot.

The Rule That Changed on the Last Day of the Count

There's one more piece of this that most of the money coverage published so far has missed entirely, because it happened at the exact moment the numbers above were being finalized. On June 30, 2026 — the very day the second-quarter fundraising period closed — the Supreme Court ruled 6-3 in National Republican Senatorial Committee v. FEC that limits on coordinated spending between party committees and candidates are unconstitutional^[34].

Here's why that matters. Until that ruling, party committees like the NRCC or DCCC could either give a candidate a small, capped direct contribution, or spend money "independently" on that candidate's behalf — meaning no coordination at all, and much higher ad rates than a campaign pays for its own airtime. Coordinated spending removes that wall. Committees can now work directly with candidates at candidate-rate pricing, often several times cheaper than the independent-spending route^[34].

In practice, that gives whichever side holds more committee cash a new, faster way to convert it into direct help for individual candidates. Right now, that's Republicans, at both the national and congressional committee level^[4]. It's a real reason for caution about reading the Democratic candidate-money edge in isolation — that edge existed under the old rules, and the rules just changed^[34]. But how much of the Republican committee advantage actually gets converted into ad buys for specific outraised candidates won't show up until the next round of filings. No Q3 numbers exist yet to say^[34].

That's the shape of what's still open. Nobody has a 2026 turnout file to say whether Democratic overperformance in low-turnout specials will hold at full midterm scale^[18]. The net redistricting effect is still moving through litigation, with estimates ranging from a realistic five to seven Republican seats up to a theoretical thirteen^[7][25][26]. Outside investor research, like Morgan Stanley's client notes, treats a split Congress as the likely outcome and points out that midterms have historically moved markets very little — a useful check on any framing that treats a single month's polling as decisive^[31]. And forecasts made 100 or more days out have a documented poor track record: the polling average has already drifted a full point since June 1, and roughly a fifth of registered voters told Pew in July they weren't committed to either party yet^[1][2]. What happens between now and November — on the economy, in the courts, or with candidates not yet nominated — hasn't happened yet. That's not a gap in this snapshot. It's the honest state of the race.