Key takeawayPre-election polls and prediction markets suggest Democrats could gain congressional seats, but a Senate majority still depends on several closely contested states, while the order in which votes are initially counted could exaggerate their apparent advantage.
What has changed in the pre-election picture?
- As of 2 November, prediction markets implied a probability above 70% of Democrats winning a Senate majority, compared with about 60% two weeks earlier; the probability of winning a House majority exceeded 90%.
- Democrats lead by an average of 8.9 percentage points in national congressional voting-intention polls. Compared with Republicans' 2.6-percentage-point lead in the national House popular vote in 2024, this represents a swing of about 11 to 12 percentage points. The Democratic lead has changed little over the past few weeks.
- Goldman Sachs estimates that if Democrats win all the House seats in which prediction markets favor them, they would gain a net 25 seats or so, bringing their total to 240 in the next House. The bank believes a Democratic House victory would matter more for the policy outlook than the Senate outcome.
Why is the Senate outcome still uncertain?
- Democratic candidates lead in polls for their party's existing seats and 7 Republican-held seats, but the closest races remain within the margin of error. Goldman Sachs notes that in recent midterm elections, Republicans have outperformed final pre-election polls in Republican-leaning states such as Alaska, Ohio and Texas.
- Goldman Sachs believes that if Democrats lose the Ohio Senate seat and also underperform expectations in Texas, their chances of winning a Senate majority would diminish. In that case, Democrats would need to win Alaska, Maine and Iowa in addition to North Carolina; Iowa is considered harder to win than Texas or Ohio.
How much does the early-voting advantage reveal?
- Early-voting data in key Senate battleground states show that the Democratic share of early votes has risen in each compared with 2024, including in Texas, where most ballots are cast before Election Day. Some states do not publish party registration data, so model estimates are used; Georgia uses past primary participation as a proxy.
- When people who say in polls that they are “almost certain” to vote are also included, the gap between the parties in voter participation narrows. Goldman Sachs believes this may reflect a greater tendency among Democratic voters to vote early, while Republican voters are more concentrated on Election Day.
- In Texas, for example, Goldman Sachs expects early vote counts could show a Democratic lean several percentage points greater than the final result. If the race is as close as polls suggest, media outlets may not call it until late at night or the next morning.
Which results should be watched first on election night?
- Kentucky's sixth congressional district could provide an early signal for the House: Trump won the district by 15 percentage points in 2024. If the Democratic candidate loses by only a few percentage points, Goldman Sachs believes that would show Democrats performing as strongly as polls suggest; however, counting is relatively slow in the district, and early results will also include early votes that lean Democratic.
- Results from Florida's fourteenth congressional district are usually available relatively quickly. A recent poll in the district showed the Democratic candidate leading by 5 percentage points; Goldman Sachs believes a Democratic loss could mean the poll overstated the party's support.
- For the Senate, Goldman Sachs expects Ohio could be among the close races called earlier; Michigan and Maine are unlikely to have confirmed results before Wednesday, while Alaska could take several days if the race is close.