Pre-registering some takes on the midterms
Why I think Democrats will undershoot their polling and what it means
New Bad Takes episode about Ye, free speech, antisemitism, and all the rest is here for your listening pleasure.
Takes will fly fast and furious once the results of the midterm elections are in, but for the sake of intellectual hygiene and discipline, I think it’s both useful and important to pre-register some takes on the outcomes so I can’t just make up post hoc explanations when the numbers are in.
My big picture expectation is that polls and poll-based forecasts are overestimating Democrats’ odds, so a result that is actually pretty good by the normal standards of midterms is going to play as a crushing disappointment.
According to 538, Republicans have a 72 percent chance of taking the House, while Democrats have a 64 percent chance of holding the Senate. Those forecasts seem D-skewed to me. I will be genuinely shocked if Democrats hold the House. The only precedents for that happening in remotely recent history are the 9/11 election and the 1962 midterms held right after the resolution of the Cuban Missile Crisis. I’m not saying it’s impossible — we do have those two examples — but it’s difficult to understand why that would happen this fall. By contrast, Republicans picking up a net of one Senate seat would be a completely banal outcome. The fact that polling is giving it only a 1-in-3 chance of happening is, I believe, a consequence of the polling being skewed.
And I think that this continued skewed polling is a problem.
The political parties tend to react to public polling with a fandom mentality, so Democrats like polls that say they are doing well and get annoyed by more skeptical polls or by people who are skeptical of the public media polling. But this is a big mistake. One reason that I think Democrats end up falling short in certain races is that unreliable polling tells them that they are winning. We’ve lived through years of takes on why Hillary Clinton lost in 2016, but the most boring explanation is a technical one: her team’s polling told her she was winning. If it had told her something else, she would have used a different message. Similarly, Biden’s 2020 polls (like the public polls) oversampled very Covid-cautious people, which led him to overemphasize a Covid-cautious message. If his campaign had better polling, they would have made smarter choices and done better.
The official Yglesias forecast
I really do think 538 is the best modeling team in the business, so I don’t want to diverge my predictions from theirs too much.
But I note that while their 2016, 2018, and 2020 Senate forecasts were all pretty good, the GOP did overperform in all three cases. I’m writing this on Monday evening, and right now 538 says that both Marco Rubio and Michael Bennet are heavy favorites to win, but that Bennet is slightly safer than Rubio. But that sounds absurd to me. I really like Val Demings and congratulate her for taking on this race, but it would be wild for an incumbent Republican to lose in a Trump state with a Democrat in the White House. By contrast, while it would be surprising for Bennet to lose, he is up against an opponent who has made meaningful efforts to distinguish himself from the national GOP, which is exactly the kind of situation in which Bennet might conceivably lose.
And the whole 538 forecast is shot through with this kind of “believe it or not” weirdness. They say Mark Kelly is safer than Ron Johnson. That Tim Ryan is more likely to upset JD Vance in Ohio than Don Bolduc is to upset Maggie Hassan in New Hampshire.
I agree, broadly speaking, with the 538 forecast that all of these are unlikely events. But in all these cases, the Democrat-friendly outcomes would be history-defying events that cry out for special explanation, while the GOP-friendly outcomes would be explained pretty easily as “the incumbent president was unpopular because of inflation, so these things happen.”
And this is important not because I think Bennet and Hassan are likely to lose, but because the history-defying outlier forecasts are a giveaway that the polling in general is skewed. Another tell is that according to polling, Democrats’ worst state among the real battlegrounds is Nevada, which happens to be the one battleground state where polling has been reliable in recent cycles.
I’m going to go with:
10 percent chance that Democrats gain one or more Senate seats
20 percent chance that there is net zero change of seats.
30 percent chance that Republicans gain net one seat.
20 percent chance that Republicans gain net two seats.
20 percent chance that Republicans gain net three or more seats.
So that’s my forecast — we’ll see who’s right.
If I’m right, here’s why
One thing to note about my forecast is that while it is pessimistic for Democrats relative to the public polling as of 10/17/22, it is wildly optimistic relative to the map and the underlying fundamentals. If you just assume “Republican candidates will generally run one point ahead of Trump,” that implies they pick up two seats.
And that’s a very conservative assumption based on the history of midterm elections.
So in a broad sense, everyone is agreeing that the GOP is set to underperform expectations for the out party in the midterms. And the big reason why everyone agrees is their awful candidate selection.
Without compromising at all on policy or even on nutty election denialism, Republicans could have put themselves in a much better position to win in Georgia, Arizona, and Pennsylvania just by nominating real professional politicians. Adam Laxalt in Nevada is doing so much better than the clowns that Trump recruited in those other states — running them was a tremendous fumble. Democrats are especially lucky that in New Hampshire, the GOP has a great potential candidate (the incumbent governor) who declined to run. Instead, their nominee is an extremist bozo. And a true nightmare scenario would have also seen Larry Hogan challenging Chris Van Hollen in Maryland.
But despite it all, I think Republicans are going to take the Senate. In part, this is because I think Democrats made two critical missteps that squandered their mid-summer surge.
Fumbling the advantage
Democrats’ summertime peak featured the conjunction of two things:
The Dobbs decision overturning Roe v. Wade upended the normal logic of thermostatic backlash by delivering an unpopular, high-salience policy change in the direction of the GOP.
The passage of the Inflation Reduction Act (IRA) not only managed to totally avoid any kind of grassroots mass opposition, but it also finally gave Democratic Party elected officials something to say in response to questions about inflation.
The IRA didn’t solve Democrats’ inflation problem — only actually bringing inflation way down would accomplish that. But it did mitigate the problem. A candidate who got a question about inflation could say something like “Inflation is really bad, that’s why I voted for the Inflation Reduction Act that will reduce the deficit [popular], let Medicare negotiate prescription drug prices [very popular], cap the cost of insulin [popular], and take an all-of-the-above approach to increasing domestic energy production [popular].” In other words, something that is objectively bad for Democrats (inflation) turns into a discussion of some popular Democratic ideas.
The problem is that Joe Biden stepped on this with a sweeping student debt forgiveness program.
It’s hard to say that deficit reduction is your answer to inflation when you’re doing a massive deficit increase. It also makes it harder to portray yourself as someone who cares about inflation when you’re doing something inflationary because of a campaign promise you made several years ago under completely different macroeconomic circumstances. Back in 2020 when inflation wasn’t a problem, debt forgiveness advocates routinely (and correctly) described it as stimulus.
The right thing for Biden to do would have been to publicly and frankly apologize to his supporters for disappointing them and for breaking a promise, but to say it was the only responsible way to react to changed circumstances.
Dobbs, meanwhile, delivered a big jolt to Democrats and continues to be a millstone around the GOP’s neck. But individual Democratic Party candidates keep squandering it by staking out a very odd “no restrictions on abortion under any circumstances” position rather than promising to return to the pre-Dobbs status quo or even copying the actually existing policy in liberal states. I don’t really understand the communications breakdown between the pro-choice groups and candidates that has led to them being so poorly equipped to talk about this. But it’s a huge failure.
I’d also note that while progressives routinely tell me it’s impossible to expect activists to exert discipline, pro-life groups have done exactly that since Dobbs. You do not see safe-seat House Republicans tweeting in favor of a national abortion ban, Mitch McConnell is not committing to hold a floor vote on one, and anti-abortion activists are not staging a sit-in in his office. You know and I know that the GOP base is committed to a very extreme position on abortion. But they are giving their members latitude to try to win the election, and then basing their post-election strategy on the seat count. It’s smart politics.
Learning from Tim Ryan
Speaking of smart politics, I think Tim Ryan has run an excellent campaign, drawing clear contrasts between himself and the national Democratic Party while also going after his opponent viciously. And when the votes are counted, I think he’ll overperform the fundamentals by a considerable margin. He is, I believe, the only Democratic statewide candidate who eagerly courted Joe Manchin’s support — recognizing that there are probably plenty of Obama-Trump voters in Ohio who think of Manchin as an example of the good kind of Democrat that they don’t make so much of anymore.
I also expect Ryan to lose. Because running as a Democrat with a Democrat in the White House in a state that Trump won by 8 points is really hard. But I do think he’ll overperform. And in doing so he offers a roadmap for the kind of campaign that I think Democrats should have run in Pennsylvania and Wisconsin and with some modification in North Carolina.
Despite their struggles in Senate races, I do fundamentally think that Democrats know what they need to do to win. They actually repeatedly show the ability to run strong overperformances like the one Steve Bullock put up in 2020. The problem is that Trump won Montana by 16. You need to bring that kind of energy to Ohio where he won by 8, and bring Tim Ryan’s energy to knife’s-edge states like North Carolina and Wisconsin.
Which gets us back to bad polling.
Political campaigning is an inexact science and reasonable people differ around the margins about tactics. But fundamentally, I think the professionals know what it means to appeal to a right-of-center electorate. The problem is that Democrats keep miscalibrating their actual appeals because they are repeatedly over-optimistic about the nature of the actual electorates that they are facing. People sometimes construe polling bias as a question of biased pollsters loading the dice in favor of Democrats. But that’s not at all what’s happening. Instead, the propensity to respond to polls has become correlated with certain other psychological dispositions that influence political behavior.
It means that the answers-polls electorate is to the left of the votes-in-elections electorate, so Democrats keep getting a skewed read of the landscape and miscalibrating their own races. The actual country is simply less-educated, lower in social trust and openness to experience, and more right-wing than the country that shows up in surveys. Democrats can win those voters; they just need to realize the practical necessity of doing it.
I could be wrong
Even in my relatively pessimistic forecast, Democrats have decent odds to hold the Senate. That would be an extraordinary achievement in historic terms. But I think elation at that prospect should be tempered by the reality that it all comes down to Republicans nominating bozos. In the House, where candidate quality matters much less, everyone thinks the bozos will probably win.
And the stakes here shouldn’t be underplayed: a Democratic Senate means that Joe Biden will be able to shuffle his cabinet and continue to fill judicial vacancies.
But a GOP House can — and says that it will — provoke a debt ceiling crisis. And it could refuse to certify the 2024 presidential election result, setting the stage for a constitutional crisis or even a coup. And I continue to think there is a remarkable mismatch between how progressives portray the stakes and what progressives are actually willing to do. The MSNBC host Joy-Ann Reid says “literal fascism” and “female serfdom” among other things are on the table, which seems pretty bad.
So to tempt voters away from literal fascism, have they been given candidates in the purple districts (D+4/R+4) who disagree with progressives about gun control? Who support banning late-term abortions? Who have qualms about trans women competing against cis women in college sports? Who favor changing asylum law to try to cut off the flow of migrants arriving at the southern border? Who think it’s a problem that college admissions offices discriminate against Asian applicants and low-income whites? I’m not saying every candidate in every swing district should dissent from party leaders on all those subjects, but how many dissent on any of them? For that matter, given that everyone agrees gasoline prices are politically significant, how much effort did Nancy Pelosi, Chuck Schumer, and Joe Biden put into thinking about how how to spur an oil production recovery back when they were planning the transition?
You don’t need to like or approve of the fact that some voters are more fearful of the mainstream Democratic Party policy agenda than they are of GOP kooks. But factually, it is true. And the mainstream party leaders have repositioned themselves to the left of where they were 10 years ago based on the incorrect idea that persuasion no longer mattered and you could win elections purely by amping up the base.
Taking those kinds of ideological risks is a plan that makes sense if you believe the downside to losing is low. And, indeed, I think Reid is in fact overstating the risks for effect. But I still think losing is pretty bad. And if you agree with that, you ought to try hard to win — including trying ideological flexibility.
If Democrats shock the world and retain both the House and the Senate, I will eat crow and say all my complaining and malingering has been overblown and the top leaders have positioned things correctly. But I think the odds are overwhelmingly high that Democrats will lose at least the House and that they will regret that loss badly. I can only hope they will also learn to regret the ideological rigidity that made the loss almost inevitable.
I've come to believe that in the post 2016 world we are watching less a contest of who is trying to win, but who is trying harder to lose. The funny thing is there probably is a pretty hard to displace Democratic majority out there. It's just that it's composed mostly of people who think everyone should have social security when they get old, healthcare should be affordable (however we get there), needy children should have free lunch at school, and that sort of boring stuff. They're live and let live and support the social changes of the 60s and 70s but are not animated at all by the identitarian and other more out there theories coming out of universities and grad schools, that are... well self evidently stupid.
These people are of course available to the GOP as well, which is doing its own disservice by putting itself in thrall to a megolomaniac who wherever possible gets crazy, clearly incompetent people nominated. However as long as the GOP's core commitment as a party is to gutting the welfare state I think it would be harder to sustain.
What's baffling to me is that no one seems ready to pivot despite how obvious this is.
It baffles me that progressives aren’t being more cautious given the US’ status as a democracy is literally under threat (and in states like Wisconsin is already significantly eroded). There’s a massive risk impatience over social progress irrecoverably sets us back. By contrast a more incremental approach seems to have been working pretty well in recent decades.