The Sweep: Swing States and Voter Registration Trends

Campaign Quick Hits

If I were a rich man: In July, the Trump campaign raised $165 million, which was $25 million more than the Biden campaign. But last month, the Trump campaign raised $210 million—$150 million less than the Biden campaign. We’ll get the full FEC reports this week to tell us all the nitty gritty and the most important number: cash on hand. But if the reports of the Trump campaign’s spendthriftiness are accurate (the campaign manager had a car and driver?!), that number may be even more disheartening for Team GOP.

Look who’s voting: In 2016, about 40 percent of Americans voted before Election Day. In 2020, that number may be closer to 60 percent. North Carolina already mailed out more than 500,000 ballots. Georgia, Wisconsin, Minnesota, and Michigan will all begin voting this week. And not surprisingly, absentee ballot requests are still lopsided. Nearly 500,000 more Democrats have requested absentee ballots in Florida and 250,000 more in North Carolina. As I’ve said before, though, the vote-by-mail numbers this year will tell us far more about how people are voting by party and far less about the eventual outcome of the race in any given state. What does all this mean for campaigns? Most important, perhaps, is that money that comes in the door from this point forward is about to become far less valuable because it will be harder to spend it in meaningful ways as more and more votes are already cast.

Contact (not starring Jodie Foster): The Trump campaign announced it had contacted its 100 millionth voter through door knocking and phone banking this week. The Trump Victory army includes “2,000 paid field staffers in 17 states and more than 2 million volunteers.” According to the New York Post, the “RNC says it has knocked on the doors of 12 million potential voters in battleground states since mid-June — that’s around 1 million a week — while the Biden campaign has knocked on zero.” Mirroring the Obama 2012 re-election bid, if Trump wins, you can bet there will be books dedicated to nuggets like this: “the GOP’s new ground-game strategy uses data to drive even the most mundane decisions, such as where to locate a local field office.”

Checking in on how debate prep is going: “President Donald Trump has not held a single mock debate session, and has no plans to stage a formal practice round,” according to NBC. Why? “Trump has repeatedly told aides that he’s not worried about debating Biden because the former vice president is likely to have a gaffe moment or stutter.” But then his campaign manager tried this blatant, if logically perplexing, attempt to raise expectations: “Joe Biden is not formidable anywhere else but he is formidable on the debate stage.”

Hola, voy a votar por Donald Trump: In what is only the latest example of how political allegiances continue to shift as a result of Trump, three polls this past week in Florida found Trump winning Hispanic voters over Joe Biden. Even in Miami-Dade County, which Trump lost by 30 points in 2016, he is now leading Biden 47 percent to 46 percent among Hispanic voters. If these numbers hold, they will have a huge impact on the politics of the immigration debates to come and could be the biggest long-term story for the future of the Republican Party coming out of 2020. And if Biden loses the presidency, will Democrats look back on their convention and the party’s focus specifically on black identity politics, as the biggest mistake of the race?

Interesting polls of the week:

• In the wake of The Atlantic story that said the president referred to people who died in wars as “losers” and “suckers,” Morning Consult found little changed in military households. Trump is still leading Biden by 10 points and 53 percent approve of his job performance—a number largely unchanged since he took office. Even so, Biden is up 7 points from where Clinton was in 2016. 

• Is Trump flagging with religious voters? A recent academic survey “predicts an 11 percentage point swing toward Biden among evangelicals and Catholics who backed Trump in 2016, based on input from both demographics across five major 2020 battleground states: Florida, Michigan, North Carolina, Pennsylvania and Wisconsin.” Last month, a Fox News poll showed Biden winning 28 percent of white evangelicals, which would be a 12 point uptick for Democrats from 2016.

• Yikes. A majority of voters in Arizona, Florida, Michigan, North Carolina, Pennsylvania and Wisconsin told pollsters they believe both Biden and Trump are mentally unfit to be president. 

Andrew Egger saw two headlines last week coming out of Pennsylvania that seemed to say exactly the opposite of one another. And so he dug in. I love when that happens … 

A Tale of Two Party Registration Numbers

If you’ve spent much time following swing state polls over the last few months, you’d be forgiven if you’d started to suspect the 2020 cake is baked.

Let’s unpack some of the numbers out of Pennsylvania to see how things may still be a little messier than they seem at first glance—and for a good illustration of how the parties massage their messaging to put the best possible construction on a number of distinct and sometimes contradictory data metrics.

Start, again, with the polls: By this metric, Biden is significantly up in Pennsylvania. This spring, while the Democratic primary was still in gear, the smattering of statewide polls we were seeing were breaking both directions roughly evenly. That’s changed dramatically over the summer: Out of the 46 Pennsylvania polls recorded by FiveThirtyEight’s polls index since July 1, 44 have shown Biden in the lead; in 29 of those, that lead exceeds 5 points. One September poll by Rasmussen Reports showed a dead heat, and one July poll commissioned by the conservative think tank American Principles Project gave President Trump a 1-point edge.

But over the last few weeks, Republican and Democratic party officials and strategists in the state have been squabbling over a different metric entirely: Voter registrations. Like most states, Pennsylvania keeps publicly available lists of registered voters, including information on party affiliation. This information can help campaign-watchers get a sense of partisan sea changes at a more fundamental level than polling data can always provide. The snapshot of voter sentiment polls record is passive—the pollster calls you—while registering to vote is an active statement of intent that tends to correlate strongly with actually getting off your butt and going down to the polling precinct come November.

So what do the Pennsylvania voter registration numbers tell us? Well, it’s a bit of a snarl. In fact, you can find headlines that seem to tell exactly contradictory stories about Pennsylvania: Here’s NBC telling us that “Democrats have voter registration advantage in four battleground states,” and here’s Politico informing us that “Since 2016, Republicans have netted nearly seven times as many registered voters in Pennsylvania as Democrats.”

What accounts for this disparity? Well, it turns out that “new registered voters” can actually mean two different things. If you’re talking about “new registered voters” in the sense of “people who were not previously registered to vote, but are now,” it’s Democrats who have the edge: Data analysis from Democratic data company TargetSmart shows that 922,000 new voters have been added to the rolls over the last four years, with about 132,000 more Democrats than Republicans among those numbers.

But Republicans have an advantage of their own to flaunt in the voter numbers: Despite new registrations tilting away from them, they’ve actually gained significant ground when it comes to the overall registered-voter picture: They’ve got nearly 200,000 more voters on the books than they did this time in 2016, while Democrats sit only about 30,000 above their 2016 numbers. While several factors contribute to this phenomenon, the most important seems to be the significant number of Rust Belt former Democrats who are now part of the Trump coalition. They’re not newly registered voters, but they’re new to the GOP.

It will perhaps not shock you to learn that state political operatives tend to think the metric that favors their own party is the one that will prove decisive (or at least tend to say as much to inquiring reporters). Republicans argue that the shift shows that the white working-class voters who broke for Trump over Hillary Clinton have found in the president’s first term ample reason to justify that choice, and that their change in registration signals that what was once a trial run is now looking more like a permanent realignment.

“You switch your registration for a reason. There’s an intensity to that,” GOP consultant Charlie Gerow told The Dispatch. “You don’t just wake up one morning and say, ‘I think I’ll be a Republican now.’ So if it was just folks that were inclined or trending to vote for Trump, they clearly are cemented into his column now, because folks that register for the first time or change their registration tend to vote in the next election, almost uniformly.”

State Democrats, for their part, argue that this overstates the case, and that the shift is at least partially accounted for by nonpolitical quirks in the data like inactive voters falling off the rolls. But when it comes to those areas where historic Democrats are switching to the red team, they concede they’re likely not getting those voters back.

But will it matter in 2020? In an interview with The Dispatch, TargetSmart CEO Tom Bonier argued that, when it comes to assessing this year’s campaign relative to 2016’s, the net registration numbers that favor Republicans are a “lagging indicator”: These former Democrats who are now registering Republican are voters who helped push Trump over the top last time, not new pickups for the Trump campaign.

“All that tells us is that Pennsylvania wasn’t as Democratic as we perhaps thought it was four years ago,” Bonier said. “It doesn’t tell us that it’s getting less Democratic now. That’s the key distinction.”

New registrations, Bonier argued, are definitionally the opposite: They represent a change in affairs between the last election and this one. And by that metric, it’s the Democrats who hold a modest but significant edge. 

This sort of thing might sound like the hair-splitting of operatives with too many charts in their heads and too much time on their hands. But the strange reality of this cycle is that, if it’s anything like 2016, it will be a cataclysmic choice between visions of America that comes down to a bare handful of voters in states with razor-thin margins like Pennsylvania, Michigan, and Wisconsin. And in that case, little electoral changes like this might make all the difference in the world.

You Better Work, B****

The race is on and so is a global pandemic. Assuming you feel strongly enough about who you are voting for that you’d like to help your chosen candidate, what’s the most effective thing you can do between now and Election Day, with or without a mask?

And remember that persuasion is not the name of the game. Turnout is. There’s very little evidence that there’s much of anything campaigns can do to persuade voters to vote for their candidate. But they can get voters to vote. So the real question is what’s the most effective thing you can do to get registered voters who already like your guy off their butts and to the polls (or the mailbox as it were)? 

This is what we call the field operation. Let’s go through your options. 

Yard sign

What is it: If you’ve never had the pleasure of building hundreds of yard signs, you may have never pondered the existential question of ‘what is a yard sign.’ At its most basic, it is a three-sided frame made of wire onto which a very tight plastic sleeve shuffles its way over. But a yard sign is also a symbol of the perennial struggle between darkness and light, hope and reality.

For all his hundreds of insightful, thought-provoking pieces, Philip Bump at the Washington Post has never touched my heart the way this 2015 headline did: “Sorry campaign managers: Lawn signs are only 98.3 percent useless.” 

As he correctly identified, “[t]he problem with lawn signs, as any campaign manager would probably tell you, is that they are expensive, annoying, logistically tricky to distribute and — most importantly — don’t seem to do much of anything.” But candidates are obsessed with them. For a campaign manager to persuade her candidate not to invest in yard signs, a study “would essentially have had to demonstrate that candidates who used them lost before a candidate would have second thoughts about the efficacy of lawn signs. And even then, he’d probably still buy them.”

Does it work: Yard signs are everywhere. Bump, in fact, cites one study that found their use had quadrupled between 1984 and 2012. But the most recognized study in this area found that the yard sign that you, as a supporter, place in your yard has zero effect. That’s right. Your yard sign is entirely pointless. Although randomly tossing them along public streets had a modest effect in down ballot races, it’s still hard to justify the cost and all my hours sleeve-hustling over those wire frames.

Verdict: If you must, you shall. But don’t fool yourself. You’re not actually helping your candidate so much as virtue-signaling to passersby. 

Phone banking

What is it: In the olden days, a phone bank was a plastic picnic table surrounded by plastic chairs and a dozen hardwire phones (hey kids, we call those landlines; pronounced lændlaɪn) manned by a dozen old ladies (aka bluehairs).

But some campaigns have more money than time. In 2004, for example, both presidential campaigns made 1.6 million phone contacts in the four days before the election. But while Republicans had a million volunteers to make those calls, the Democrats had only 250,000 and made up the difference with paid workers, which some have cited as their reason for losing a tight race.

Fast forward to 2020, it’s never been easier to phone bank. Campaigns can send you a script and automatically connect you to the numbers on your list anywhere in the country as you sit in your bathrobe. No mask needed!

Does it work: Political scientists Alan Gerber and Donald Green found that “a volunteer phone bank that reaches 1,000 people will produce about 28 new voters” and costs around $36 per vote. Not bad. (And especially compared with robocalls, which not surprisingly “never raise voter turnout. They have no mobilization effect, and no persuasion effect either.”)

But reaching 1,000 people isn’t as easy as it used to be. As more and more people ditch their landlines, it is increasingly hard to get people to pick up unknown numbers on their cell phones and match cell phones to voter files. The DNC, however, announced earlier this year that it had “purchased high-quality cell phone numbers for every possible voter in all 50 states and the District of Columbia which will be integrated into the DNC’s national voter file and add tens of millions of new numbers.” 

Verdict: If you want to boost your candidate, this is a no brainer. Use your extra hour after dinner each night to knock out 10 phone calls and feel good knowing you’ve done your part.

Door knocking

What is it: Door knocking is the introvert’s nightmare. At its most basic, a group of volunteers will head out with a folder containing a map, a list of addresses, door hangers to leave on the “no answers,” and pamphlets for the ones that do. When they get to the neighborhood, they will break up into couples and try to speak to voters in their own homes. 

Does it work: That Gerber Green study found that 5,000 doors—if we assume a high response rate in which 1,000 people actually answer—will generate around 40 new voters for about $33 a vote. But remember, those “new voters” aren’t switching their vote. They are just showing up to vote when they would not have done so otherwise.

In a look at Obama’s 2012 ground game, it looks like it probably made less than a point’s difference. And there was a potential downside as well. Campaigns have less control over canvassers. And in 2012, these “individuals who were interacting with swing voters on the campaign’s behalf were demographically unrepresentative, ideologically extreme, cared about atypical issues, and misunderstood the voters’ priorities.” In other words, they often turned voters off while thinking they were persuading them. 

There’s also a relatively new thing called deep canvassing. The idea is that “instead of short, scripted conversations where canvassers essentially tell voters what they should believe, deep canvassing involves asking questions and listening empathetically to answers.” Initial research has shown it can be effective—even at the ever-elusive persuasion—but it would also be nearly impossible for a campaign to utilize because of the training required for those thousands of volunteers across the country.

Verdict: It’s good exercise in good weather. If you’re a Trump fan, this is the way to go. But bear in mind, you may want to tone it down because it turns out voters “can be less receptive to those who don’t share similar characteristics like age, gender or accent” and are not nearly as ideological as you. 

It will be extra fascinating this year because we have the control experiment we’ve always wanted. Biden’s campaign is foregoing door knocking entirely despite the conventional wisdom that Obama’s superior ground game won the 2008 and 2012 elections. (But see, Hillary Clinton’s superior ground game.) Trump’s campaign, on the other hand, is boasting 100 million doors. Note: This is the number of doors knocked and not the number of people who answered their doors—although in this work-from-home era, the response rate may actually be higher than previous years.

Friends and family

What is it: I hate to even suggest it, but this one is pretty simple: Post your ‘I voted’ sticker on Facebook and shoot a text to 20 friends reminding them to vote. It’s literally the least you can do, right? Sure, but it also might be the most effective. 

Does it work: We know through any number of experiments that social pressure is the most effective turnout tool. In fact, there’s nothing that political scientists appear to agree on more. As Dylan Matthews over at Vox noted last week, “there’s widespread agreement and optimism across the political scientist/data consultant spectrum about the possibilities of ‘relational voter turnout’ that exploits people’s friendships and social attachments. As just one example, the effect can be as high as 13 percentage points in local elections. 

It turns out that if you instituted purely evidence-based electioneering, “you might have fewer campaign ads clamoring for your attention in every form of media you encounter, but you’d be hearing a lot more campaign ads from your friends and family.”

Verdict: With great power, comes great responsibility. Don’t make me regret this … but … post away!

Photograph by Rosenfeld/Getty Images.