The midterms should be a stake through the heart of the mobilization myth
Turnout was great for Republicans, but they lost most of the key races anyway
In recent years, I’ve become obsessed with making an argument that would have struck people as too obvious to even write about 20 years ago: Swing voters are real (as I wrote for Vox in 2018), and you can’t win elections by mobilizing a secret non-voting majority of leftists.
Now that all the races are resolved, the 2022 midterms were pretty clearly the most catastrophic defeat yet for mobilization theory. As Nate Cohn explained in detail in the New York Times last Thursday, Republicans decisively won the turnout battle in key states, even while losing the preponderance of important races. Most years require an in-the-weeds effort to parse whether mobilization or persuasion mattered more because they both point in the same direction (a subject we’ll return to), but this year there was no such dilemma. Differential turnout explains less than 0 percent of Mark Kelly, Raphael Warnock, or Catherine Cortez Masto’s wins — it was an R-leaning turnout year, but they all won.
We don’t yet have the same level of visibility into exactly who voted in Pennsylvania, where Democrats had an even better performance in statewide races. But we do know that comparing 2020 to 2022, the PA county with the largest turnout decline was the city of Philadelphia. That’s what you would expect to see if PA had the same turnout dynamics as those other six states, and it would be very unusual for a trend to hit all six states that we have full data for and not represent a national trend.
Democrats won key races by persuading a small but nonzero number of Republicans to vote for them.
The triumph of Kemp/Warnock crossover voters
On its face, this should be good news for progressives.
Democrats did pretty well despite suffering the normal midterm demobilization problems. “Normal presidential election turnout in the presidential election year” is a pretty easy bar to hit, and this year’s results indicate that House Democrats in particular could do better in 2024 without any huge revolution in party ideology or message. But a lot of people on the left are very dug-in on the bigger picture conceptual argument about mobilization because they believe mobilization theory gives them license to practice irresponsible, undisciplined politics. That’s how we ended up with this slightly bizarre urge to give credit to Stacey Abrams (who lost her Georgia statewide race rather badly) for the fact that Raphael Warnock managed to get more votes than his opponent.
I don’t think Abrams’ loss reflects poorly on her — I don’t see how any Democrat using any set of tactics could have beaten Brian Kemp in this midterm. It looked early in the cycle like a MAGA primary challenger might beat Kemp, in which case Abrams would have had a fighting chance and we could critique her choices.
But the fact that she lost badly is relevant context for Warnock’s win: he scored the plurality in the original election and won the runoff because a lot of Kemp voters voted for him. And if you look at how the down-ballot races went in Georgia, it’s clear that these aren’t just Brian Kemp superfans or people who have some huge problem with Stacy Abrams. The whole GOP ticket won, and the party controls both houses of the state legislature. As in all elections, a majority of Warnock’s voters were hardcore Democrats and a majority of Republicans voted for Herschel Walker, but the reason Warnock won is that a sliver of people who are otherwise party-line GOP voters backed him.
Perry Bacon, Jr. looked at these results and concluded that “Stacey Abrams didn't win. But her ideas did.”
He then lists a bunch of Abrams-associated ideas that really were vindicated by the race, and it’s absolutely true that she was right to say in the wake of 2016 that Democrats had better odds of growing in Georgia than of winning back Ohio and Iowa. But I think it’s also indisputably true that along with the ideas Bacon attributes to her, Abrams is associated with the idea that Democrats could organize their way to victory. And that’s just not what happened.
There’s also not some huge mystery as to how Warnock’s campaign persuaded crossover voters. Especially in the runoff, his allies made direct appeals to Kemp voters urging them to vote for Warnock.
And obviously this worked in large part because Walker had a lot of serious flaws as a candidate. But Warnock’s campaign manager, Quentin Fulks, describes their deliberate efforts to court these voters.
“There could have been other campaign operatives or another campaign that could have said, ‘OK, Herschel Walker has all this baggage, so we’re just going to run to the left and just try to turn out as many of our voters and just let Republicans eat their own,” he told Bloomberg’s Sahil Kapur. “We didn’t do that.”
This idea of deliberately courting crossover voters is so banal that it hardly seems worth analyzing. But it really did go out of style in the wake of liberals’ shock and horror at the idea that anyone would vote for Donald Trump. Normally, when you lose an election, the first order of business is to figure out how to convince some of the people who voted for the other guy to change their minds next time. But lots of progressives found Trump so appalling that the idea of trying to do outreach to his voters was beyond the pale. Even though Hillary Clinton’s infamous analysis put only half of them in the basket of deplorables, there was very little interest in even trying to reach the other half. But there just isn’t some other way of doing politics.
Democrats need Trump voters’ votes
The states that Joe Biden won in 2020 are all represented by two senators, but two of those senators — Ron Johnson and Susan Collins — are Republicans.
Democrats’ current majority rests on the backs of Sherrod Brown, John Tester, and Joe Manchin, all of whom represent Trump states and face very steep re-election battles in 2024. It’s also notable that there aren’t a lot of targets that are clearly better than Ohio. In 2022, North Carolina, Florida, and Texas were all closer than Ohio. But Florida’s politics seem to have moved further to the right since then, Democrats have come up decidedly short in the last two North Carolina Senate races, and Texas remains a pretty firmly red state.
That’s not to say Democrats should give up on NC, FL, and TX. But those three states plus Ohio and Iowa are where you have to look if you want to obtain governing majorities down the road. And these are states where it’s hard for Democrats to win. You need to either find Obama voters who flipped to Trump and convince them to flip back, or else find long-time Republicans who never liked Trump all that much but still preferred him to a Democrat and find a way to court them.
Realistically, this probably involves making some unpalatable concessions on policy and ideology. Not unpalatable in the sense that “leftists will be mad but Matt thinks these are good ideas on the merits,” but, like, actually unpalatable. You need to find people who have enough bad opinions that they’ve been voted solidly Republican for the past several cycles, and get them to vote for you — probably by agreeing with them about some stuff.
But the very least a candidate can do is the kind of “hey, I respect you and I want your vote” stuff that we see from Warnock. There’s just no other way to do things. If you want to get a paid parental leave program passed, you need to elect some more people who support paid parental leave. Ideally, they would also oppose right-wing policy change in most areas. But there might be very, very few topics on which they actually endorse progressive change. The reality, though, is that a senator who favored a paid parental leave program and otherwise opposed sweeping new legislation from the left or right would constitute a huge improvement over the status of NC/FL/TX/IA senators.
It’s a shitty reality, and I totally understand why folks resist facing up to it. And that’s what makes the mobilization myth so perennially tempting — it would be nice to think there’s some magical alternative to dealing with an unpleasant reality. But there isn’t.
There’s no tradeoff here
Once you give up on the magical idea of mobilizing the base instead of finding ways to make swing voters like you, it’s easier to see that there actually isn’t a tradeoff here anyway. In other words, you should absolutely try to maximize the turnout of sporadic voters who are likely to vote for you. But there’s no reason to believe there’s a tension between that goal and trying to appeal to swing voters, because the boring truth is that sporadic voters are less politically engaged and less ideological than non-voters. Successful but boring messages (like Catherine Cortez Masto and John Fetterman talking about how they think it’s good when cops arrested criminals) are a perfectly good mobilization strategy. They let Cortez Masto and Fetterman seem like sane, sensible human beings to the kind of people who are not that interested in politics and only sometimes vote.
Indeed, one of the things that’s so striking about the 2022 crossover vote data is that it’s extremely rare to have a situation like the one we saw in Georgia, Nevada, Arizona, and Pennsylvania where Democrats won swing voters while doing badly on turnout.
In a normal year, you either get great turnout and do well with independents (like in 2018) or you get crappy turnout and tank with independents (2014). But in 2022, Democrats did badly on turnout — admittedly not nearly as bad as in 2014 — while nevertheless winning a bunch of key races thanks to crossover voters.
The key is that while Democrats won the preponderance of the most important races, their overall 2022 performance wasn’t very good. They got about 48 percent of the two-party vote, which should have been consistent with losing the Georgia, Nevada, Arizona, and Pennsylvania Senate races. And recall our Georgia case: Democrats really did do badly in most Georgia races in November. That’s how we know so precisely that Warnock won thanks to crossover voters. Taken on the whole, Democrats flopped on both turnout and vote choice. Warnock himself just did really well. Here’s a chart David Shor put together showing that Democratic incumbents in tossup races ran much stronger than Democrats in races that tilted clearly D or R. That’s smart politics; it’s good to run your best races in the most important spots instead of coughing up the likes of Herschel Walker and Blake Masters.
But I think it also confirms that if the national environment had been better — if the gas price drop was a little faster, if crime started falling in 2021 after the 2020 surge rather than plateauing, if the Dobbs ruling came out closer to the election — Democrats would have done better nationally on both vote-choice and turnout. That could have meant re-election for Steve Sisolak in Nevada or Mandela Barnes winning a Senate seat in Wisconsin.
Relearning the obvious
I think the boring truth about all this is that there’s a fairly obvious division of labor in politics. Politicians and their teams need to develop high-level messages that they think will be popular with their voters. In office, they need to try to implement ideas that are consistent with that message (which can include doing something quietly without drawing attention), and while campaigning, they need to do targeted mobilization outreach to sporadic voters. But that mobilization is largely a technical question that doesn’t alter the basic mandate to try to craft a popular message.
And advocacy groups need to try to convince people they are right about stuff.
This is, again, an almost absurdly obvious thing to write, but persuading people to change their minds is a very powerful tool. Marriage equality was won largely because people changed their minds. Marijuana laws keep liberalizing because people have changed their minds. Transgender and gender-fluid identities are winning more acceptance because people have changed their minds. I’ve been trying for years to convince people that we should increase immigration levels to the United States. The needle is moving on this, but my side hasn’t yet won the argument, and we’re currently seeing a mini-backlash because the asylum situation isn’t being handled well.
Because this is a big, high-salience national political issue, it’s hard for incremental change in public opinion to alter policy. We need to keep trying to convince people.
On land use, things are different and there are thousands of relevant local political jurisdictions. The issue doesn’t sharply track partisan polarization, so every small increase in the number of people persuaded by the case for reform opens up the scope for more policy change.
This kind of thing is the hard work of political advocacy, and while it’s a good idea to make your case to candidates and elected officials, there’s also no percentage in getting angry that people don’t want to embrace politically risky ideas. Our job as advocates is to be effective at advocating, try to change minds, and create new possibilities. Too many advocates got tied in with mobilization theory, started insisting there were cheat codes to ignore public opinion, and were writing checks about turnout they couldn’t possibly cash. Politics is very important, but at the end of the day it’s also kind of boring, and there isn’t some exciting hidden way to win — you have to appeal to mass opinion.
In this debate, you're never going to persuade the pro-turnout people that they're wrong. To win it, you're just going to have to organize the pro-persuasion people and get them to the polls.
After all, there's a huge untapped reservoir of pro-persuasion people who already agree with you. You just need to mobilize them, by emphasizing the most extreme and unpersuasive pro-persuasion lines, and demonizing everyone on the pro-turnout side. That should do the trick.
I would posit that Democrats didn't do very much to persuade Republicans to vote for them. Instead, the Trump-selected candidates were demonstrably awful. Dr. Oz is a charlatan and a resident of NJ. Herschel Walker can't string three sentences together and, oh by the way, had affairs, multiple kids out of marriage and funded abortions. Kari Lake came closest to a normal candidate, but she was all-in on "the election was stolen" silliness, as we can see from her current, meritless lawsuits.