The Bilateral Disgust Contest

The standard midterm narrative runs like this: the president is unpopular, therefore his party will lose seats, therefore the opposition will gain them. The first two steps have modest historical support, as discussed in the previous article in this series. The third step is treated as automatic, a simple consequence of the first two. If voters are unhappy with the party in power, they vote for the other party.

But what if they are also unhappy with the other party?

That question is not hypothetical. It describes the current political landscape with precision, and it exposes a gap in every forecasting model being used to predict the 2026 midterms.

Both Sides of the Ledger

Trump’s approval rating sits at roughly 35 percent. The commentary class treats this as a flashing red light for Republicans. But the same CNN/SSRS poll that produced that number also found that only 28 percent of Americans hold a favorable view of the Democratic Party. The Republican Party is at 32 percent favorable, which is underwater by any standard but still four points higher than the opposition.

Read those numbers side by side. The president’s approval is 35 percent. The party that is supposed to benefit from his unpopularity has a favorability of 28 percent. The entity the electorate is expected to turn to as an alternative is viewed less favorably than the entity they are supposedly fleeing.

This is not a normal midterm environment. In previous cycles where a president’s party lost significant seats, the opposition was viewed as a credible, at least passably acceptable alternative. When Republicans lost 40 seats in 2018, the Democratic Party’s net favorability was roughly even. Today it has shifted to net negative by nearly 30 points. The opposition party has gone from acceptable alternative to broadly disliked institution in the span of one midterm cycle.

The Priority Gap

Favorability is abstract. The CNN/SSRS poll asked a more specific question that cuts deeper: “Do Congressional Democrats have the right priorities?” Seventy-four percent said no. Only 25 percent said yes.

That 74 percent includes a majority of Democrats themselves. Fifty-five percent of Democratic voters said their own party’s congressional leadership has the wrong priorities. CNN’s own analysis noted that Democrats are more likely to say their own party has the wrong focus than Republicans are to say the same about Trump. Among independents, both Trump and congressional Democrats are viewed as equally off-topic, with about three-quarters saying each has the wrong priorities.

A party whose own voters say by a majority that it is focused on the wrong things faces a problem that no amount of presidential unpopularity can solve. You cannot convert dissatisfaction with the incumbent into votes for your candidates if your candidates are seen as working on things voters do not care about.

The Cohesion Asymmetry

The internal health of each party’s coalition tells a starker story than the topline numbers.

Only 19 percent of Republican-aligned adults say their party is mostly divided, compared to about one-third of Democratic-aligned adults, a 14-point cohesion gap in the GOP’s favor. Seventy-six percent of Republicans view their own party favorably, compared to only 63 percent of Democrats viewing theirs. The enthusiasm gap runs 13 points; Democrats cannot get two-thirds of their own coalition to say they like their own party.

On internal fracture lines, the asymmetry holds across every dimension the poll measured. On the Democratic side, 72 percent say the Israel divide is causing problems within the party. About two-thirds see problematic divides over priorities and ideology, and 58 percent see division on whether to cooperate with Trump. On the Republican side, every comparable number is 10 to 20 points lower: 54 percent on focus, 52 percent on ideology, 52 percent on opposing Trump, 47 percent on Israel. The GOP has divisions, but its own coalition experiences them as less severe.

Sixty-seven percent of Democrats say the fight over whether to lean left or move toward the center is causing problems. That is not a party that disagrees about tactics. That is a party without consensus on its own identity. When your voters cannot agree on what you stand for, winning elections requires the other side to be so toxic that your incoherence becomes irrelevant. At 35 percent approval, Trump is unpopular, but the question is whether he is unpopular enough to overcome the opposition’s self-inflicted dysfunction.

The Double-Hater Trap

About one quarter of the public holds a negative view of both parties. These “double haters” currently prefer Democrats in the upcoming midterms by 31 points. On its face, this looks like a Democratic advantage. Look closer at the reasons and the advantage becomes fragile.

The most common reasons double haters dislike Democrats are viewing them as do-nothing (22 percent), not standing up enough to Trump and the GOP (11 percent), or too liberal (10 percent). Their most common reasons for disliking Republicans are failure to stand up to Trump (14 percent), not caring about people (10 percent), and Trump himself (8 percent). The double haters are preferring Democrats not because they like Democrats, but because their reasons for disliking Republicans are more visceral and personality-driven while their reasons for disliking Democrats are about passivity and incompetence.

“Vote for us, we’re useless but at least we’re not them” is not a message that generates turnout enthusiasm. These are the voters most likely to stay home, vote third party, or change their minds between now and November. Building a midterm strategy on the assumption that double haters will show up and break your way is building on sand, particularly when the 2016 and 2024 elections demonstrated that double-hater voters can break sharply toward Republicans when the political environment shifts.

Why the Models Break

The forecasting models used to predict midterm outcomes, including the Tien/Lewis-Beck model, the Abramowitz model, and various generic-ballot models, share a common structural assumption: low presidential approval translates into votes for the opposition because the opposition is there to catch the disaffected voters. None of these models contain a variable for opposition-party favorability, opposition-party priority alignment, or opposition-party internal cohesion. They treat the out-party as a passive receptacle for anti-incumbent sentiment.

When both parties are at historically low favorability simultaneously, this assumption fails. The mechanism the models depend on, voter migration from the president’s party to the alternative, is impaired because the alternative is also generating its own disaffection. Disaffected voters have three options: vote for the opposition despite disliking it, stay home, or vote on non-partisan issues like pocketbook concerns. The models assume the first option dominates. The current data suggests the second and third options may be far more significant than historical patterns would predict.

The generic ballot currently shows Democrats ahead by about 6 points among registered voters. In a normal midterm with a president at 35 percent, historical patterns would suggest something more like 8 to 12 points. The compressed margin may itself be evidence that Democratic unpopularity is suppressing what would otherwise be a larger structural advantage. The models will not detect this because they have no mechanism for it.

What This Means for November

The midterms will not be decided by the president’s approval rating in isolation. They will be decided by the relative capacity of each party to mobilize its supporters while attracting the narrow slice of persuadable voters in competitive districts. On the mobilization side, the data favors Republicans: higher internal cohesion, stronger intra-party favorability, and a base that views itself as united. On the persuasion side, the picture is muddier, because the voters being persuaded are choosing between two options they dislike.

The commentariat treats Trump’s 35 percent as a referendum verdict. The data says it is half of a bilateral disgust contest, and the other half, the opposition’s own unpopularity, is receiving a fraction of the analytical attention. Until the forecasting models account for what happens when both sides are underwater simultaneously, their predictions will carry a systematic bias toward overestimating opposition gains, because they are modeling a world where dissatisfied voters have somewhere appealing to go. In 2026, they may not.

Polling data from CNN/SSRS (March 26-30, 2026), CNN economic approval (March 2026), Newsweek/RealClearPolitics favorability aggregates, and The Post Millennial analysis of Democratic priority ratings. Model references from the American Presidency Project (UCSB) and Algara et al. 2025.