The Early Vote: A Handy Guide

With millions of swing state voters having already cast ballots, what can the early vote trends tell us about November 5th?

People line up for early voting at a polling station at the Black Mountain Public Library in Black Mountain, North Carolina on October 21, 2024. Photo by Jim WATSON / AFP.

There’s something happening in the early vote across the battleground states: More Republicans, and fewer Democrats, are voting early.

But wait, is that bad for Kamala Harris? Looking at the early vote in states that track party affiliation, you might start to conclude that Republicans are going to have a good year, especially in places like Nevada where they currently have an early vote lead. And even in places like Pennsylvania, the percentage of the early vote held by Republicans has increased considerably since 2020.

For example, here’s what the firm TargetSmart shows about the increasing share of the Republican early vote in Pennsylvania, which is all mail-in ballots:

The Republicans had just 19.4 percent of the vote in 2020 and 2022, but now they have 28.6 percent! And the Dems have fallen from 71.5 percent in 2020 to 62.3 percent in 2024. Is this a warning sign?

We need to zoom out and take some deep breaths. As I’ll discuss in this piece, the main thing that the early vote numbers tell us is that fewer Democrats are voting early, largely because there isn’t a pandemic anymore. In fact, if you look at the raw numbers of mail-in votes, comparing 2024 to 2020, Republicans have only added a little less than 30K by this time twelve days out from the election, while Democrats have dropped off by around 320K. Here’s the important thing to keep in mind: Those Democrats are still going to VOTE, but they will do so in person on Election Day. And that has ramifications for the so-called “red mirage” we are used to seeing.

It is true that in some states, such as North Carolina and Nevada, Republicans are using in-person early voting more, at the urging of their leaders and despite mixed messaging from Trump. But that also means that the impact of their Election Day surge will be relatively smaller.

In today’s piece, I’ll walk through several things to understand about the early vote and survey where we are twelve days out from Election Day. I’ll look at how things are going in the Blue Wall Great Lake states of Pennsylvania, Michigan, and Wisconsin and discuss the concept of the Democratic firewall of mail-in ballots in the first two. I’ll also explain how this will impact the so-called “red mirage” we have seen in previous elections.

Then I’ll look at what we can see in the Sunbelt states, particularly in North Carolina, Georgia, and Nevada and discuss how the early vote appears to confirm how close things are in those states. Finally, I’ll touch upon a true wild card in the election—the unaffiliated voters—and provide some data about how at least some experts believe they will break.

Things look pretty “normal” in the Blue Wall States

The early votes in Michigan and Pennsylvania have been by mail-in ballot. And as we’ve seen in recent elections, Democrats have requested and turned in many more ballots than Republicans.

Here is how things stand in the critical battleground of Pennsylvania as of October 23rd, according to data from the Election Lab project run by Prof. Michael McDonald of the University of Florida.

Democrats have returned around twice the number of ballots as Republicans so far, and importantly their rate of return is about six percentage points higher. That tends to show a bit more enthusiasm to vote among Democrats than Republicans. The percentage difference among returned ballots is, however, lower than in 2020, when Democrats outpaced Republicans by around 64 to 24 percent.

And here’s something to watch. While the percentage of mail-in ballots returned by Republicans has increased considerably from 2020 (largely because the number of Democratic mail-in ballots requested and returned has dropped off sharply in favor of Election Day voting), the demographic make-up of those votes has remained remarkably stable. This is particularly true when looking at the gender composition of the early vote.

2024 is a post-Dobbs election, and women voters have been making the key difference in special elections ever since the summer of 2022. Statistician and data modeler for the group Force Multiplier, Marsha Kessler, has been exchanging post-Dobbs election data with me, and together we looked at the average swing toward the Democrats in special elections since Dobbs. What she found was fascinating.

Zeroing in on the swing states (at my request), Kessler confirmed that the “Dobbs effect” was evident in the swing states where special elections had occurred (16 in PA, 8 in MI, 5 in WI, and 3 in GA). On average, these states swung 7.5 points toward the Democratic candidates post-Dobbs.

If women are turning out enthusiastically to vote early in the key swing states, that would be a good sign for Harris. So what do the numbers show?

Let’s look again at Pennsylvania, which has rich data that breaks down early voting by party registration and gender. Even while the number of Republican early voters was rising and the number of Democratic voters was falling, the percentage of women early voters has remained the same across recent elections, including this one.

Recalling the chart from the top of this piece, we can see that the total number of registered Democratic early voters in Pennsylvania is down sharply while the number of registered Republican voters is up slightly from prior elections, creating a 9-point move in the percentages for both.

But now let’s look at the gender of these voters across these same years. As you can see, the gender make-up of the electorate in Pennsylvania has hardly changed at all, despite the sharp drop in Democratic early voters and the rise in Republican:

This is curious because typically around 60 percent of Democratic voters are women, but only around 52 percent of Republican voters are. You would think that this would drag down the gender disparity in the total vote, but it hasn’t by very much at all.

That’s what leads me to conclude that the early vote in Pennsylvania, impacted as it is by a post-Dobbs world, indicates that many of the same conditions that led to Democrats romping in 2022 across the statewide races in Pennsylvania are at play. Of course, we won’t know for sure until all the votes are counted. But nothing in the early vote demographics leads me to worry that women in Pennsylvania are losing interest in the election. In fact, the data points the other way.

The same can be said for Michigan, although it doesn’t track party affiliation so the data is a bit harder to sift through. Women voters, who will break heavily for Harris, still carry slightly more of the early mail-in vote than they did in 2020 and 2022, even though far fewer Democrats are choosing to vote early by mail this year. 

The firewalls are rising in Pennsylvania and Michigan. Will they hold?

In past elections, where mail-in and early votes figured more heavily into the final tallies, there was often talk of a “blue” or “Dem” firewall. And to some extent, that talk continues today because mail-in balloting still figures highly into the total in Pennsylvania and Michigan.

But what is the Dem firewall? Simply put, if Democrats vote in large enough numbers before the election, then Republicans will have a harder time overcoming that lead and jumping the “wall” to burn through to victory. 

In 2020, for example, Democrats in Pennsylvania built a formidable firewall of around 1.1 million more mail-in votes than Republicans had cast. That was out of around 2.6 million mail-in votes cast that year. (“No party affiliation” voters turned in 283K ballots). Here are the statistics from the 2020 election for Pennsylvania:

This year, there are so far around 1.1 million ballots cast out of 1.9 million requested. There are three more days left to request mail-in ballots, so we may get up to 1.95 million requested.

I’ll spare you the back-of-napkin math here, but based upon past voting trends (including how the no-party folks broke), and how many more Dems will vote on Election Day this year, there’s a general sense that a “Dem firewall” of around half a million mail-in votes before Election Day will be sufficient to hold off the GOP Election Day surge. But this shouldn’t be relied upon as gospel by any means! We simply don’t know come Election Day how many voters from either party will show up. Still, the higher that firewall grows, the better Democrats will feel. 

So where does the Dem firewall stand now in Pennsylvania? As of yesterday, the “ballot edge” (or Dem firewall) stood at around 356K. (Thanks to election watcher Joshua Smithley who is dutifully tracking this and knows Pennsylvania election data very well!)

The firewall has grown between 6K and 30K a day, depending on which county is processing batches and how large they are. I would not be surprised to see it just barely hit that half-million mark by Election Day.

In Michigan, the lead is more difficult to pin down because that state doesn’t break mail-in ballots down by party affiliation. But the folks at TargetSmart have developed a “modeled party” dataset that estimates the party breakdown based on a host of factors, and here is how they believe the parties are faring in the early vote out of the 1.3 million votes cast so far:

Looking at their own datasets, Patrick Schuh of America Votes estimates that the Dem firewall in Michigan will grow to 489K:

And once again, it is Democratic women who may be driving this good early turnout in Michigan to help build this Dem firewall. As Tom Bonier of TargetSmart noted, using their “modeled party” dataset,

I've been saying for several days now that the early vote looks strong for Dems in MI. A huge driver for that is turnout among modeled Dem women. At this point in ’20 GOP women were turning out at a higher rate than Dem women (3.6 pts); right now Dem women have a 5.7 pt advantage: 

What about Wisconsin?

The third piece of the Blue Wall is the swing state of Wisconsin. The state sadly doesn’t maintain any party registration data from the early vote. And like the other two Blue Wall states, Wisconsin has seen a sharp decline in early mail-in voting because we’re no longer in a pandemic. That makes prognostication a bit tricky.

Early in-person voting has only just begun in that state, too, but it smashed records on Day One, with 97,436 votes cast on that single day alone. That’s nearly triple the number on the first day of early voting in 2022 and 20K higher than in 2020. But is it Democratic enthusiasm, Republican enthusiasm, or a bit of both?

Early data in Wisconsin indicates something similar to the other states, especially among women voters. Despite the decline in mail-in voting, largely from Democrats where women comprise a much bigger share of voters than in the GOP, the gender make-up of the Wisconsin early vote hasn’t changed from prior elections. Here is the Wisconsin tally as of yesterday with 13 days to go:

The state went narrowly for Trump in 2016, then swung back narrowly to Biden in 2020, and now is balanced on a knife’s edge according to many polls. Wisconsin Democrats had a great year in 2023 with a judicial election that went heavily their way, and they’re looking to keep that momentum going in the general election. All indications are that it will be a close contest, as usual.

The “red mirage” may prove fleeting on Election Day in key swing states

One key takeaway from the drop in early mail-in voting, especially in states like Pennsylvania that have only early mail-in balloting and not “early in-person voting” like other states, is that the phenomenon of the “red mirage” will be far less pronounced.

The red mirage happened because of the peculiar way mail-in ballots are processed and counted. In a state like Pennsylvania, by law they’re not even permitted to start processing and counting ballots until Election Day (blame that on GOP-led efforts to prevent early counting). In the past, that has meant that same-day votes on Election Day, which skewed heavily GOP, would tend to be counted first, with big counties slow to provide final dumps of ballot counts.

As the news networks carried the count live, it appeared in certain states that Republicans were far ahead on Election Day, even though millions of Democratic ballots were yet to be counted. Donald Trump used that to his advantage by claiming early victory and demanding that counties “stop the count”—which of course would have left Democrats disenfranchised by the millions. 

This year, the red mirage is likely to be far less apparent. First, a lot more Democrats will vote in person on Election Day, based on the number of mail-in ballots requested by the parties. That will tend to blunt the surge of GOP votes that day. Second, some counties, such as Allegheny County (home to Pittsburgh), have pledged to process and count their mail-in ballots shortly after the polls close. This could result in a reverse mirage, in fact, where it seems like Harris is far ahead because the Allegheny mail-in ballots get added in well before the totals from rural counties arrive.

In short, we should keep all of that in mind and not jump to any early conclusions about where things stand until mail-in ballots are processed and big counties have reported in on Election Night.

Nevada flashes warning signs for Democrats

Jon Ralston reports on elections for the Nevada Independent, and he usually has a good sense of how things are going in the #WeMatter state. (And in this close election, Nevada may in fact matter, but less so than the other larger swing states.)

The glum news for Democrats is that early voting in Nevada is going well for Republicans. Many had expected Democrats to take an early lead from mail-in and early in-person voting, but instead Republicans have pulled slightly ahead, with Democratic voter-heavy Clark County, home to Las Vegas, looking sluggish.

Here is how things look in Nevada with 12 days to go:

It is far too early to begin assuming the state will flip red, however. As I’ll discuss below, in other Sunbelt states where the early vote looks tied, a Marist poll of voters who have already voted shows Harris leading by double digits. So the race may look tied at this point, but it may not actually be. And Clark County will show up; it’s just a question of by how much. Enthusiasm will matter in these final days.

Still, at this rate we may be talking about a rural red wall instead of a blue firewall when it comes to the early vote in Nevada. Democrats will need to step it up and soon to keep Nevada’s six electoral votes in the Harris column.

The race is tight and enthusiasm is high in the bigger Sunbelt swing states

There is one thing election analysts all agree upon: Early voting in the Southern swing states is setting records. As the New York Times noted in a chart out earlier today, Georgia and North Carolina top the list for the percentage of the electorate that has already voted, at 31 and 30 percent respectively.

The early voting surge in Georgia is particularly sweet to Democrats, given how hard the state government has tried to make it for citizens to vote. As the Times reported,

Among the states that made changes to early voting rules since 2020 were Georgia and North Carolina, both important swing states. In Georgia, the Republican legislature and governor passed a sweeping law that decreased the time to request absentee ballots, imposed strict new ID requirements for those ballots and significantly limited the availability of absentee ballot drop boxes.

For its part, North Carolina is defying the odds by showing big turnout despite large parts of the state being ravaged by hurricanes and a two-week-delay ordered by the Supreme Court so that Robert F. Kennedy, Jr.’s name could be taken off the ballots.

So, who is voting early in the Sunbelt? In North Carolina, there is an even split among Democrats, Republicans, and independents, with each around a third of the electorate. This is a big departure from past years when Republicans voted early in lower numbers.

Georgia doesn’t track party registration among early voters, but here are TargetSmart’s “modeled party” figures, showing Republicans with a slight edge over Democrats in the early vote.

These numbers don’t tell the whole story, however. Marist was out with a poll of the Sunbelt swing states this morning, and it found that in North Carolina, Harris was leading Trump among those who had already cast votes by 55 to 43, and in Georgia by 54 to 45. Conversely, among voters who haven’t yet cast ballots, Trump was leading Harris 53 to 45 in North Carolina and in Georgia by 52 to 46.

In Arizona, Republicans are out ahead of Democrats in the early vote by about six percentage points. But about one in five voters has listed no party, and these voters could determine the entire election there. The Marist poll similarly found that, among early voters in Arizona, Harris was leading Trump 55 to 44, with the numbers pretty much reversed for those who haven’t voted yet. 

I caution not to read too much into any single poll, but the Marist Sunbelt polls do suggest, generally speaking, that independents who voted early are breaking for Harris, and that there may be defections among Republicans as well. Indeed, this pretty much has to be the case, or the tied counts in the early votes can’t be squared with these poll results favoring Harris. 

I’d say that’s good news for Harris, but if the polls are correct, then any advantage now is likely to be offset by votes still to come from Trump supporters.

Unaffiliated voters could tip the election toward Harris

One final observation. There’s a bit of good news for Democrats to be found in the high numbers of unaffiliated voters casting ballots in states like Pennsylvania and North Carolina. A sizable number of these are newly registered, younger voters, especially in Pennsylvania which has automatic voter registration in place for those obtaining state IDs or driver’s licenses.

Citing reporting by CBS, Forbes noted that among the newly registered voters since 2022 (the year of the Dobbs decision) “49% of the 17.82 million are Democrats, 34% are Republicans and 17% are unaffiliated with either party.” That surge in new voters was given another jolt when Harris entered the race in July, moving enthusiasm among Democrats from 55 percent in March to 78 percent by August, according to Gallup. During the same period, enthusiasm among Republicans increased only slightly from 59 percent to 64 percent.

These new voters could make the real difference come Election Day when most young voters cast ballots. Here is the age breakdown among early voters so far in the battleground states, according to data from TargetSmart. As you can see, most young people typically vote last minute, and they will be critical to offsetting the older, Trumpier vote among people over 50.

While the early vote can’t provide crystal clear answers to how the election will go, it has provided enough information to indicate that we currently remain on track to perform as well as we did in past elections in Pennsylvania and Michigan, meaning Harris has a slight advantage in those two states. As I like to remind nervous Democrats, if she wins those two states, and as expected she wins the single electoral college vote from Nebraska’s second Congressional District around Omaha, she need only win one more of the larger battleground states—Wisconsin, North Carolina, Georgia or Arizona—to clinch the presidency. 

Under these conditions, I would rather be the Harris campaign looking to win one of those states than the Trump team trying to win all four.