November 6th, 2022:

First, some slightly outdated numbers from the Chron.

Fewer Harris County voters cast ballots during this year’s early voting period than in the 2020 and 2018 elections, according to unofficial voter counts released after polls closed on Friday night.

From Oct. 24 to Nov. 4, about 736,000 people had voted at Harris County’s 99 early voting locations. They accounted for about 28 percent of Harris County’s 2.6 million registered voters.

Local voters are taking to the polls to elect dozens of local offices, including Harris County judge, as well as to vote on $1.2 billion in bond proposals and on statewide races for governor, lieutenant governor, attorney general and others.

But so far, the number of voters has lagged behind the turnout in recent November elections.

In 2020, which was a presidential election year, more than 1.4 million people, about 57 percent of registered voters in Harris County, voted early. In 2018, early voting turnout was 855,711 people or 36.6 percent of registered voters.

There was an uptick in recent days of voter turnout. On Friday, more than 95,000 people voted in person, the highest daily total during the two-week early voting period. The second-highest voting day was Thursday, when more than 75,000 people voted. Long lines and waits of more than 2 hours were reported at some locations on Friday. After polls closed at 7 p.m., there were still 200 people in line to vote at some locations, including NRG Stadium, according to the Harris County Election Administrator’s Office.

I’ll get to the numbers in a minutes, but this story has a publication time of 9:14 PM and refers to a tweet posted at 7 PM by the Elections office. The final in person vote count for Friday was actually 105K, so waiting till later to publish (the voters file with the final count hit my mailbox at 11:24 PM) might have been advisable. Be that as it may, this is what we got when all was said and done. Final EV totals from 2018 are here and from 2014 are here. The final totals for 2022 are here.

Year     Mail    Early    Total
===============================
2014   67,967  307,280  375,247
2018   89,098  766,613  855,711
2022   57,871  692,478  750,349

In the end, early turnout for 2022 was 87.7% of what it was in 2018, while in person turnout was 91.6% of the 2018 number. That’s down, but as noted on Friday, the gap has narrowed. All three final days of early voting had higher in person totals this year than in 2022. Comparisons to 2020 are not particularly interesting – this is a non-Presidential year, and in case you forgot there were three weeks of early voting in 2020, thanks to the pandemic. We may be hard pressed to match 2020 EV totals in 2024.

The question is, what might final turnout look like? The trend in Presidential years is that fewer votes are cast on Election Day. In 2012, 35% of votes were cast on Election Day, in 2016 it was 26% of the vote, and in 2020 – again, in a year with three weeks of early voting, and also a record number of votes by mail – just 12% of the vote was on Tuesday. That trend is much less pronounced in off years. In 2010, 44% of all votes were cast on Election Day, in 2014 it was 45%, and in 2018 it was 29%.

What that means is that if Election Day in 2022 is like it was in 2018, we’ll get about 307K votes cast (*) for a final total of about 1.057 million. If it’s more like 2014, we’ll see 614K votes cast, for 1.364 million total. Going by the estimate of about 2.53 million total voters in Harris County, that would be 41.8% turnout of registered voters on the low end, and 53.9% on the high end. That compares to 41.7% in 2010, 33.7% in 2014, and 52.9% in 2018.

My high end scenario, in other words, would mean that 2022 exceeds 2018 both in absolute numbers as well as in turnout percentage. That feels a bit exuberant to me, but not out of the question. I think the 29% turnout on Election Day is probably too low – I’ll get into that more in a minute – so let’s split the difference and say 37% of the vote happens on Tuesday. If that’s the case, then 1.191 million votes will be cast, or 47.1% turnout. That’s down from 2018 but not by much. It feels plausible to me, with the proviso that we’re all just flat-out guessing here.

One argument for why we might get a larger portion of the vote cast on Tuesday than we had in 2018 is that more voters came out in the final days of early voting this year. The last three days of early voting in 2018 accounted for about 27% of the final in person early vote total. This year, about 35% of the in person vote came out on the last three days. That at least suggests the possibility that more people are taking their time to get to the polls. Does that necessarily carry over to voting on E-Day instead of voting early? Maybe, I don’t know. We’re dealing with the tiniest possible sample sizes here, so you can read anything you want into this stuff. That’s why I try to talk in terms of ranges of possibility. Different years are, well, different. It’s plausible to me that the Election Day share of the vote this year could be higher than it was in 2018, but that doesn’t mean it will be, nor does it mean that if it is it will be as much as it was in previous years. Pick your adventure here. Have you voted yet?

(*) – Just a reminder that some number of mail ballots come in between Friday and Tuesday, which means that the mail totals you see on the official Election Night returns don’t match what I’ve got here from the daily EV tallies. In 2018, there wound up being just over 100K mail ballots, which means there were another 11K that came in after the “final” totals posted above. My guess is we’ll get between 65-67K total mail ballots this year. All this means that my calculations for the Election Day vote share are slightly off, but it’s not worth worrying about. The basic contour is still whether we get an Election Day more like 2018 or more like 2014/2010.