It’s all happening.
Completed early voting ballots are pouring into county clerk offices from Nevada to Illinois. Candidates are flooding the airwaves with last-minute calls for support. In Georgia, Cobb County residents have stood in snaking three-hour lines, waiting for the chance to cast their votes.
It’s far too early to read the tea leaves from the early voting returns. Different states and localities have different early voting practices, and there are still 18 days to go before Nov. 6, the first national election day since Donald Trump stunned the country by winning the presidency.
But given the polls of voter enthusiasm, the astronomical fundraising numbers, and the remarkable number of ballots cast in this year’s special elections and primaries, there’s plenty of reason to believe this is going to be a uniquely high turnout midterm election.
University of Florida professor Michael McDonald, who runs the United States Election Project blog, expects that some 45-50 percent of eligible U.S. voters will participate in the midterms—a figure not seen in a midterm election since 1970. In the 2014 midterms, slightly over a third of eligible voters, or 37 percent, cast ballots.
McDonald is basing that prediction in part on the high rates of early voting in states like Georgia, where turnout is three times higher than it was at this time in 2014.
“The initial early voting data we’re seeing is very unusual,” McDonald told TPM. “I’ve never seen anything like this.”
In Georgia, a state that maintains individual-level voting data, black voters are turning out in huge numbers compared to the previous midterm cycle.
“You can see that African-Americans voting right now are less likely to have voted in 2014 than the whites who have voted so far,” McDonald said.
As McDonald—and some of his critics—have pointed out, Democratic gubernatorial nominee Stacey Abrams has carried out an aggressive effort to coax her supporters to vote early. Those high rates of early returns in the Peach State and elsewhere could just signal that the most motivated, enthusiastic voters are rushing to the polls, and that the numbers could level off by Election Day.
For McDonald, the early voting data is useful “as a sort of check on the polling data. If we’re seeing something very consistent with the polls, it just gives us higher confidence that the polling data is giving us a good signal.”
Public polls have fairly consistently shown Democrats with a strong shot at retaking the House and a number of swing state governor’s mansions, with the GOP likely to maintain control of the Senate. Democrats have expressed greater enthusiasm for casting ballots in this cycle, but polls show Republican voters are feeling above-average enthusiasm, too.
The talk of a “blue wave” hinges in large part on historical patterns. The president’s party tends to take a hit in midterm elections. Having Trump, specifically, in the White House is another boon for Democrats, since the President is a polarizing figure whose popularity is underwater even in many red states.
Democrats have also won a string of local and congressional special elections in tricky places, like a state Senate seat in rural Wisconsin and U.S. House seat in Pennsylvania. They’ve pulled in remarkable fundraising hauls, significantly out-raising Republicans in the fight for the House and in races like the Texas Senate battle between Rep. Beto O’Rourke and Sen. Ted Cruz.
But these indicators of a strong year for Democrats don’t tell the whole story. Republicans have a major geographic advantage. The enthusiasm gap benefiting the Democratic Party has narrowed in recent weeks, as the vicious confirmation battle for Supreme Court Justice Brett Kavanaugh helped polarize angry voters on either side of the aisle.
And while Democrats appear to be breaking from their pattern of sitting out midterm cycles, Republican voters who were highly engaged in 2014—an exceptionally low turnout election—are just as engaged this fall.
George Khalaf, president of Arizona political analytics firm Data Orbital, told TPM that he was “shocked” to compare same-day numbers from the 2016 presidential election and 2018 midterms in Republican Maricopa County, where a roughly similar number of raw ballots had come in. The numbers were far lower in liberal Pima County, where Tucson is located, compared to 2016.
With all of these crossing signals and intersecting factors at play, some elections experts refrain from reading anything into early voting numbers at all.
“My advice is to literally just ignore everything,” wrote 538 editor in chief Nate Silver.
“It’s extremely difficult to infer much from early-early voting numbers,” Paul Gronke, founder of Reed College’s Early Voting Information Center, told TPM in an email.
“We know from the past that early voters tend to be ‘decided’ voters, and I would not be at all surprised to discover that Democrats in particular are fired up this election cycle,” Gronke said. “To the degree that ‘fired up voters’ overlaps with states with a substantial amount of early voting, we should see higher numbers. We *may* be able to use the early vote figures as a measure of overall enthusiasm and interest, but past attempts to do so have not performed well.”
It may become easier to read into early voting numbers as more states adapt mechanisms for it, and more voters become accustomed to casting ballots this way. A 2016 Pew Research Center survey documented the steadily growing share of the electorate that relies on “nontraditional voting.”
As for the 2018 cycle, McDonald joked, “All will be revealed in just a few weeks’ time. The fog in that crystal ball is going to start fading away.”