Six weeks ago, a prominent television pundit declared Democrats were heavy favorites to retake the House, with the only remaining question being the size of their majority. The RealClearPolitics average on July 2 showed Democrats up 7.4 points on the generic ballot.
Today that lead sits at 5.3 points and is still falling. The question now is whether it drops far enough, fast enough before November. And the data suggests it might.
Here is the number every Republican strategist should have pinned to their wall: in 2018, the RCP generic ballot average improved by more than four points for Democrats between June 1 and July 30. This year, over that same window, it moved three points in the opposite direction, toward Republicans. The 2018 wave that handed Democrats 40 House seats was built on accelerating Democratic momentum. This cycle is showing the opposite. Anyone predicting a 2018 replay needs to explain why the trajectory looks nothing like 2018.
As originally reported, the individual polls tell the same story. WaPo/Ipsos has Democrats up three points, down from five in April. Morning Consult shows Democrats at plus three. Harvard-Harris has them at plus two. Reuters/Ipsos lands at plus two. RMG Research also shows Democrats at plus two, representing a seven-point swing toward Republicans since May alone. One Economist/YouGov poll showing Democrats at plus four was built on a D+7.5 sample. When you need to oversample Democrats by seven and a half points to produce a four-point lead, the underlying electorate is not where you need it to be.
The structural math remains brutal for Democrats. Due to redistricting, they need to win the national popular vote by three to four points just to flip the House. Most recent polls put them at D+2 to D+3, which is already below that threshold. That is before factoring in the well-documented historical pattern of generic ballot leads shrinking further between now and Election Day. It is also before accounting for pollsters undercounting Republican support by nearly three points, as happened in 2024.
The financial picture makes everything worse. The NRCC has $92.7 million in cash on hand. The DNC is carrying $18.5 million in debt and has reportedly asked vendors to hold their invoices until after the election. Total committee infrastructure tells an equally stark story: Republicans have $677 million versus Democrats’ $136 million.
The party that is supposed to be riding a wave cannot afford to run the ground game a wave requires.
None of this means Republicans will hold the House. But the people confidently predicting a Democratic takeover are increasingly relying on trendlines that are moving in the wrong direction, a financial disadvantage that is real, and polling samples that require generous assumptions to produce the numbers they want to see.