Early Voting Looks Good for Harris But Could Be an Illusion

Early in-person voting in North Carolina. Photo: Allison Joyce/Bloomberg/Getty Images

In a national election as close as this one — and this one is really close — a lot of attention will obviously be paid to any real voting data we can find. And in an era when early voting in one or multiple forms is now available nearly everywhere, it’s tempting to analyze early voting data as it is compiled. According to the University of Florida Election Lab tracker, as of October 21, 14.9 million votes had already been cast, with just over 10 million mail ballots having been returned and 4.7 million cast in person. Total mail ballots requested by voters (in states where that is a requirement) amount to 56 million.

By comparison, total early voting in 2020 reached 101 million ballots, about two-thirds of the total vote, with 65 million involving mail ballots and 36 million cast in person. No one expects this year’s early vote to match 2020’s, in part because a number of states temporarily relaxed voting-by-mail requirements to reflect pandemic conditions, and in part because voters themselves were more likely to avoid Election Day lines.

Reports about which party (or presidential candidate) is doing well in early voting should come with a number of huge asterisks. Only 23 states report early voting numbers according to voters’ party registration status. Among those (roughly half of early votes cast so far), according to Election Lab, registered Democrats represent 46.5 percent of early voters and registered Republicans represent 32.5 percent with the rest (20.9 percent) being unaffiliated.

Should the Kamala Harris camp be enthused about this (heavy asterisks!) “lead”? Maybe. Votes already cast are “banked” votes, which allows the campaign to focus on undecided or unmobilized voters. But almost everywhere, early voters are very likely to be people who would have voted in any circumstances, so a given party’s heavy early vote share doesn’t necessarily mean a heavy overall vote. It’s also worth remembering that early-voting breakdowns in non-battleground states may affect the national popular vote but won’t determine the election winner. And the patterns in battleground states are more complicated than the national numbers indicate.

First of all, there is no voter registration by party, and thus no early vote partisan breakdowns, in three battleground states: Georgia, Michigan, and Wisconsin. Georgia does, however, report voting by racial demographics as part of its compliance with the Voting Rights Act, and so far those participating in Georgia’s very heavy early voting look demographically much like the electorate as a whole (59 percent non-Hispanic white, 28 percent non-Hispanic Black). In North Carolina, which has both partisan and demographic reporting on early voting, Democrats have a very slight advantage over Republicans in early voting (with nearly a third of early voters being unaffiliated) and Black early voting is a bit below the Black percentage of the electorate. In a third early state, Arizona, Republicans actually lead Democrats by a 43 to 36 percent margin in early voting. And in Nevada, which now has universal mail-ballot availability, Democrats have a very narrow advantage (38 to 37 percent) in early voting.

So the national picture of strongly pro-Democratic early voting is really reflected (as best as we can tell from the available data) only in one battleground state, albeit a very important one: Pennsylvania, where registered Democrats have a 63 to 27 percent advantage in early voting so far. But the Keystone State is unusually polarized in preferred voting methods; in 2020, only 16 percent of those casting mail ballots (the only form of early voting available there) were Republicans, even though the state wound up being dead even.

There’s one other thing about early voting that distinguishes it sharply from voting in general: It skews very old. In the nine states (including battlegrounds Georgia and North Carolina) that report the age of early voters, 49 percent were over 65 and only 16 percent were under 40, according to Election Lab.

Nationally, polls are still showing a big partisan division in willingness to vote early or in in person on Election Day, as reflected in a new USA Today/Suffolk national survey:

A new USA TODAY/Suffolk University Poll shows the vice president leading the former president by 63%-34%, close to 2-1, among those who have already voted.

That preference turns around among those who plan to wait until Election Day to vote, with Trump ahead 52%-35%.

As some states have begun early mail-in and in-person voting, one in seven respondents said they had already voted. A third said they plan to vote early; that group supported Harris by 52%-39%. And nearly half said they’ll wait until Election Day.

Overall, Harris was favored by 45%, Trump by 44% — a coin-toss contest.

How this will play out on Election Night this time around is more than a little unclear, depending on how rapidly votes are counted and which votes are counted first, which varies state to state. In Pennsylvania and Wisconsin, a legislative ban on pre-processing mail ballots could slow down the overall count and make the early count skew Republican, as was the case in 2020, when Donald Trump declared victory based on initial returns. But Georgia is promising to count all mail ballots on Election Night this year. Arizona has a particularly heavy reliance on mail ballots dropped off on Election Day, which could slow down the count. North Carolina has made it less likely early in-person ballots will be counted and reported on Election Night. Nevada is the one battleground state where mail ballots postmarked on Election Day will be counted if received later, though Republicans are still fighting that permissive approach in the courts. So the landscape is complicated and constantly shifting. The results will obviously sort themselves out in the days following November 5 (the director of Fox News’ “decision desk” predicts the race will be “called” the Saturday after Election Day, as it was in 2020), but that does provide a large window for paralyzing uncertainty, conspiracy theorizing, and (in Trump’s case) premature victory claims.