US: Did Dobbs Give Democrats A Path Back?

Spoiler: Probably Not

One of the more annoying things about covering the US this year, as someone with a spotty at best record of predicting it, is the fact that there’s a lot of different ways to slice through the morass of (not very good) data out there, and depending on your lean, you can make a (somewhat) coherent case for Democrats losing 10 House seats and gaining Pennsylvania’s Senate seat or you can make a case for 30 House losses and 3 or 4 Senate losses from the same pool of data. This all comes down to what you think of the same pool of data, and whether or not there’s a risk that the polls will systemically miss again in the same way as 2016 and 2020.

This is much different than in 2020, because even then, the vast, vast majority of the data pointed to a blue wave, and all of the metrics did. In 2016, the district polls pointed to a much worse showing for Hillary with White Working Class voters than the state polls and the national polls did, but in 2020 all the metrics – state polls, national polls, district polls, approval ratings for the President – pointed to a wipeout for the GOP and a landslide for Biden.

This time, there’s a bunch of conflicting signals – the state polls point to a status quo, D+4 environment, in aggregate, Biden’s approval is some god awful number suggesting a GOP landslide, and the generic ballot polls are somewhere in between – which means that unlike in 2020, which was a simple game of “do you trust the polls yes/no?”, this is a much more complicated game of discerning which metrics are “more” likely to be right or wrong.

I’ve made my stance on the state polls pretty well known at this point (“Stop believing dogshit state polls”, to coin a phrase), but in the nearly two weeks since Roe, we have had 6 national polls which released crosstabs for both their Generic Ballot poll and their Biden Approval. Democrats, on average, lose the white vote across these 6 polls by 10% - 39%-49%. Across those same 6 polls? Biden’s net approval is -24%, 35%-59%. And this is what will decide things in November.

Given that nobody can accurately poll Hispanics and Black voters, my Generic Ballot modelling has priced in a 2020-style result with Black voters and a worse for Dems than 2020 result with Hispanics and Asians. They’re constants in my model, because I needed a way to analyze change in Democratic fortunes that wasn’t “20% of Black people changed their minds on Biden in a week, so the Generic Ballot swung heavily”. Conceding that I have already corrected for the likelihood that Biden Approval polls are missing a lot of non-white, Biden-disapprove, Dem voters, there’s still an ocean between the Generic Ballot and what an implied GB from Presidential approval would show. And here’s why it’s the main question of the midterm.

Across those six polls, if you just take the Biden white approval numbers as your benchmark, Republicans will win the House popular vote by 4.8%. If you take those polls’ white Generic Ballot support levels, Democrats would win the House by 5.5%. Sound insane? Biden lost whites by 12% in 2020, so another couple of points of swing towards Dems with whites is worth a point and a bit nationally. That gap is the difference between a GOP landslide and Democrats winning North Carolina and Wisconsin in the Senate. It’s also the litmus test for whether you’ve learned your lesson from 2020 or not.

It's always been the case that US pollsters can’t poll Black and Hispanic voters well, and the fact that Democrats beat their polls in Nevada seemingly every cycle before 2020 was the proof for that. What would habitually happen with Hispanics is why Beto O’Rourke beat his 2018 polls – the pollsters were mostly bang on the money about his white support, and then undershot his Hispanic support levels by 20%. In 2020, that flipped – they overshot Democratic support with whites, and got the Hispanic numbers pretty close to right. The 2020 polling miss is often charactizered as being about working class Trumpian voters, but that was much more the 2016 miss – in 2020, they thought college educated whites would swing 30 points left, and it was much closer to 10 points. They got the direction of travel right, but the magnitude completely wrong. Here? I’d bet real money that the same thing is happening.

Like, if you think Congressional Democrats are going to do better with white voters than Joe Biden did, with the current shit economy, I have oceanfront property in Kalamazoo for you, but that’s what these overly fulsome white voter splits point to. Will there be some Biden-disapprove, Dem voting whites? Sure, but not this many. Whether it’s intentional or not, many people who thought Democrats were going to have a very good 2020 and a good 2021 are going back to the well now, and they’re wrong to.

“Ah, but don’t you have Democrats as Senate favourites?” I hear someone ask. Yes, because the GOP nominated four dogshit candidates in the four majority deciding states. Blake Masters is supportive of a federal personhood amendment, Adam Laxalt is mediocre at best, Herschel Walker has undiagnosed CTE and Dr. Oz is apparently as popular as drinking your own piss in Pennsylvania, so yes, I have Democrats as narrow Senate favourites – as does the 538 model when they take out the “expert ratings”, which has a GOP +6 projected national environment. If the GOP fail to win the Senate, it will be entirely because of horrible candidate recruitment, not the environment.

If you want to be charitable to Democrats, there’s a chance they could scrape to a neutral environment, which could be enough to get them a very slender House majority, if incumbency really helps and the GOP’s swing to them is particularly unfriendly (aka, they rack up huge swings in right-trending seats and lose a bunch of suburban tossups), but their chances of winning the House are probably sub-5% this year. There’s a lot of hope attached to this group of white voters who disapprove of the President but say they’ll vote for Democrats. Most likely, a lot of those expressing no opinion on the Generic Ballot question are Biden-disapprovers who will vote GOP.

You want hope, Democrats? It’s the Senate, where abortion helps Democrats hugely by shifting the political terrain from unfriendly issues around economics to healthy terrain for them, especially in 68% pro-choice Nevada and 63% pro-choice Arizona. But the fanciful notion that Democrats could find some path forward in the House or some fiction about North Carolina or Wisconsin Senate is just that. There’s been some good news for Democrats’ electoral prospects since Dobbs, but there’s a risk a lot of Democrats are setting themselves up for heartbreak – again.