Comparing early voting apples to apples

The Chron’s writeup of the first five days of early voting is an incomplete picture.

Is turnout up or down in Texas in the 2024 election? It depends on how you look at it.

Since early voting began Monday, more than 3 million of the 18.6 million Texans registered to vote have already cast their ballots, with many counties across the state reporting long lines at the polls.

When compared to the last presidential election in 2020, the percentage of voters showing up to the polls in person is higher this year. But the overall voter participation rate this cycle is lagging when factoring in mail-in ballots, which soared in popularity amid the coronavirus pandemic.

[…]

In Harris County, in-person participation is slightly down from 14% to 13% of registered voters. Polling places in some Houston-area counties, especially Fort Bend and Montgomery, are surpassing or meeting 2020 turnout numbers.

Despite the high volume of voters, Texas’ overall participation rate is lagging slightly behind 2020. In the first three days of early voting, 15.1% of registered Texas voters cast ballots, slightly down from 15.7% in 2020, according to the Secretary of State’s office.

That decrease can be in part attributed to a quirk of the 2020 election, when many voters used mail-in ballots amid the coronavirus pandemic.

In this year’s election, only 221,417 mail ballots had been cast as of Wednesday, less than half of the mail ballots that were cast in the first three days of early voting in 2020.

[…]

Brandon Rottinghaus, a political science professor at the University of Houston, said it’s too early to tell whether statewide turnout offers good or bad news for Democrats’ hopes in the hot race for U.S. Senate between Ted Cruz and Colin Allred. More than half of all voters are expected to choose early voting.

Suburban areas like Fort Bend and Denton County, north of Fort Worth, have trended towards Democrats in recent cycles and are currently leading in turnout. And while overall turnout is similar to 2020 levels so far, urban areas could steadily catch up over the weekend, when most rural counties opt to close early voting sites.

There are potential trouble spots for Democrats. The number of votes cast in Dallas and El Paso Counties as of Wednesday decreased from 2020, even though both have registered thousands of new voters.

“Democrats are hoping for blockbuster turnout, and this isn’t necessarily it so far,” said Rottinghaus.

Rottinghaus also noted that the gender gap in voters so far appears to be on par with in 2020.

Some election officials have noted that higher turnout has been sustained into the third and fourth days of early voting, rather than tapering off after the first few days as in past elections.

There are three things that I would add to this:

1. The drop in mail ballots means little to nothing in terms of total output. This is because 2020 was such a singular year and there was so much more of a focus on (and a fight about) voting by mail thanks to COVID. I would bet substantial money that the vast majority of still-living and still-in-Harris-County people who voted by mail in 2020 but are not voting by mail this year will vote in person. Someone with access to the voter rosters in both years could do a spot check on this now, looking to see what the mail voters from 2020 are doing this year.

I’m old enough to remember that the unprecedented surge in early voting in 2008 led to some wildly optimistic projections about the final vote tally, based on past early-versus-Election-Day comparisons and the misplaced belief that the surge in early voting meant a commensurate surge in overall turnout. That just wasn’t the case. A lot of people had changed their behavior, casting an early ballot where before they had voted on Election Day. That was indeed a result of the enthusiasm a lot of Democrats had in 2008 – I lost track of the times I heard a first-time early voter say some variation of “well, if I get hit by a bus tomorrow, at least I voted” that year – but these were reliable voters speaking. It was a shift in when people voted that caused the early voting numbers that we had not seen before.

2. That same shift in behavior, caused in large part by anxiety over COVID and the fear that the longer one waited the more one risked getting sick and not being able to vote, that drove not just voting by mail but also the huge first (four-day) week turnout. I’ve said this before and I’ll say it again, this time with mail ballots included: In 2020, the first four days of early voting saw an average of 128,021 ballots cast per day. For the remaining 14 days of early voting, that daily average was 67,205. If you want to be generous and exclude the four weekend days, where there would be fewer mail ballots, the daily average was 73,627. That’s a huge dropoff.

The general pattern in most years is a big first day (which includes all of the mail ballots that were received before in person voting started) and a mostly consistent in person showing up to the last day or two, with the last day usually being the biggest in person day of the whole cycle. That wasn’t true in 2020, though the last day was the biggest in person day after the first week. In 2008 and 2012, more total votes were cast in the first five days than the last five days (this is comparing weekdays to weekdays). In 2016 there were more total votes cast in the first five days than the last five, but that was entirely attributable to the 61K mail ballots that had arrived by Day One. The in person total was slightly higher in the last five days.

The point I’m making here is that the comparison from here on out will be very different because the pace of voting slowed down quite a bit in 2020 after that first week. The coverage of early voting so far doesn’t make that clear – it leaves one with the feeling that we will still be trying to catch up to 2020 all next week when in fact we’ll be running well ahead of it. That may or may not get us ahead of the “turnout as a percentage of registered voters” number, but we’ll be right around it. That’s a question better answered later anyway.

3. And that brings me to the third point, which is that – again, driven by the COVID conditions and fears – many more people voted early in 2020 than in other years as a share of the overall turnout. In 2008 and 2012, 63% and 65% of the total vote (mail and in person) was cast before Election Day. In 2016, it was 74%. In 2020, it was a whopping 88% of the vote. Basically, almost no one voted on Election Day in 2020 – about 204K total people, not much more than the 169K votes cast on the first day of Early Voting.

The trend in voting behavior has been more people voting early over time. This isn’t just the case for Presidential elections but for all elections, including the city elections (I very much underestimated the early vote share in both the November election and the December runoff in 2023) and even those dumb HCAD elections this year. So I do expect there to be a higher early vote share than 2016’s 74%. I just don’t expect it to be as big as 2020’s 88% was. Some people do still like to vote on Election Day. My completely out of thin air guess for this year is that about 80% of the total vote will be cast before November 5. I’m saying all this to say that while Early Voting is a big part of the story, it’s not the whole story. Whatever we get in early voting this year, I would be cautious about projecting what it means for the final number based on what happened in 2020. Look back a little farther than that and see if it affects your opinion. It is also the case that someone with all of the relevant voting rosters could look and see how many people who had voted reliably in past elections had not yet done so this year. That would be a better tell.

Hope this helps. I’ll be back on the daily EV tracking tomorrow. Have you voted yet?

Related Posts:

This entry was posted in Election 2024 and tagged , , , , , , , , , , , , . Bookmark the permalink.