Midterm polls have good news, great news, and bad news for Democrats. Wait, what?
The slightly complicated history of the generic ballot.
If you’re a Silver Bulletin reader, you know that the party out of power (this time around, Democrats) tends to gain seats during the midterm elections. You probably also know that there’s a difference between a tsunami like 1994, a wave year like 2006, and a ripple like 2022. So, how is the 2026 race for Congress shaping up?
It’s several months too soon to direct you to our election forecast, especially when redistricting fights are still ongoing in a few states. However, we’ve launched a page for our 2026 generic congressional ballot polling average, which we’ll be updating nearly every day from now through November. And it’s never too early for a Model Talk.
Let’s start with the basics. What are generic ballot polls? Actually, there are two types of surveys that qualify under this banner. One asks voters whether they plan to vote for a Democrat or a Republican for the U.S. House in the upcoming election. Occasionally, an enterprising pollster will look up the names of the exact candidates in each voter’s district and include them (i.e., Mariannette Miller-Meeks or Christina Bohannan). But usually they just go by the party labels, which is why pollsters call this the generic ballot. The second type of poll asks voters whether they’d prefer for Democrats or Republicans to control Congress. We include these too since, empirically, they produce nearly identical results to the other type of survey and increase our sample size.
Even with nearly 300 days to go before the midterms, the generic congressional ballot is a reasonable predictor of the House popular vote. Let’s compare three things. First, the generic ballot as it appeared on equivalent days in midterm cycles since 1994 — that is, 285 days before Election Day — based on running our algorithm retroactively on past election cycles. Next, the generic ballot average as of Election Day in each midterm. And third, the House popular vote: that is, the total number of votes cast for Democrats and Republicans.
Except, instead of looking at Democrats and Republicans, let’s remove the party labels and consider the president’s party versus the opposition party instead:
As you can see, the opposition party usually prevails. Despite the presence of a couple of unusually strong midterm years for the president’s party — 1998, generally attributed to Republican overreach on Monica Lewinsky, and 2002, generally attributed to the Sept. 11 attacks — the opposition party won the House popular vote by 4.4 points on average. That would likely be enough for Democrats to gain control of the House, given that redistricting has been something of a wash so far, and that the House map (as opposed to the Senate map) doesn’t currently have much overall bias.
True, there’s a lot of uncertainty in the calculation. The correlation between the generic ballot at this point in the cycle (early January) and the eventual House popular vote is 0.53, which is decent but not great. And the empirical margin of error on the January polling average, compared to the popular vote, is 5.7 points.
Democrats currently lead by 5.3 points in our generic ballot average. That implies that everything from an 11.0-point Democratic win to a 0.4-point Republican edge is within the margin of error, which covers 95 percent of possible outcomes.
The former would qualify as a blue tsunami. Yes, the number of Democratic seat gains would be somewhat curbed in the House because of extensive partisan gerrymandering. But that’s the point at which even supposedly safe seats could break, potentially turning some gerrymanders into dummymanders. With such a large tailwind at their backs, Democrats would also be highly competitive in a half-dozen or more Republican-held Senate seats. While there would be few guarantees in the upper chamber, they’d be in a decent position probabilistically.
On the flip side, a 0.4-point edge for the GOP in the national popular vote would be a disappointment for Democrats. Still, the Republican margin would actually be smaller than in 2022 or 2024, which they won by 2.7 and 2.6 points, respectively. So even here, considering the narrowness of the current Republican majority, it’s not out of the question that Democrats could win the House. The Senate would almost certainly be out of play, however. Democrats would stand plausible chances in Maine and North Carolina, but probably not elsewhere, and they need four pickups. States like Georgia and Michigan, currently held by Democrats, could also be tough to defend.
Overall, that’s good news for Democrats. But where does that leave us? And what’s the “great news” and the “bad news” that we referred to in the headline?