Welcome to the Daily Kos Elections early voting roundup, which appears every weekday until Election Day. Click here to find out if and when early voting is available in your state.
With one week to go until Election Day, the statistics from Monday’s early vote were, on balance, a disappointment for Democrats. It consisted of the best day of the in-person early vote period thus far for the Republicans in Nevada and also showed a (narrow) bounceback for Republicans in Florida as they continued to offset narrow in-person gains for the Democrats with larger leads among returned mail-in ballots.
Much of the online discussion of early voting this week, however, seems to be revolving around the demographic makeup of those that have participated in the process thus far. A couple of studies released over the past couple of days indicate a potentially worrisome (for Democrats, at least) dive in the participation of African-American voters in early voting. So, the questions now appear to be twofold: 1) Are the studies indicating flagging black turnout accurate? and 2) Does this necessarily portend trouble for the Democratic ticket?
FLORIDA
One of the places where the turnout performance (or lack thereof) is a big issue is Florida, where an explosive report was published on Monday by University of Florida Prof. Daniel Smith. Smith argued that African-American proportion of the early in-person vote (with eight days left in the election) has plummeted from 25 percent of the total early vote down to 15 percent of the early vote. A rising proportion of Latino votes relative to 2012 has offset that somewhat, but the white share of the in-person early vote has risen from 60 to 64 percent, according to Smith.
On the surface, this seems potentially disastrous, given that African-Americans are, far and away, the most loyal Democratic voting bloc in terms of their voting preferences. Also, the rising white share of the electorate would be cause for alarm for Democrats as well, given that Republicans routinely win the white vote, both nationally and in Florida.
However, there are some caveats here. Smith noted that, because of the different election schedule in 2016, the 2016 numbers reflected eight days of early voting, whereas 2012 reflected just two days of early voting. Given that we are talking about proportions of the early vote here, and not raw numbers, this is a critical distinction. As Daniel Nichanian noted, correctly, this meant half of the data that Smith was utilizing from 2012 came on a “Souls to the Polls” Sunday, whereas this year those weekend days are a considerably smaller fraction of the sample.
If there is a reason for apprehension (ranging from mild-to-severe, depending on your general tendency to fret), it is in the partisan statistics for in-person early voting. With more than half of the early voting period exhausted, Democrats only lead Republicans by 56,000 votes. That is a considerably smaller gap than in 2012. However, local Democratic guru Steve Schale offers a reason for optimism on that end:
Another interesting point that a very smart observer mentioned to me yesterday: More traditional Election Day Republicans are voting early than Election Day Democrats. According to the file I use, about 620,000 people who have voted early already are voters who in 2012 voted on Election Day. In other words, in the absurd terms of today’s politics, these are early voters that the parties are “cannibalizing” from their own Election Day turnout.
Republicans have been making the case that Democrats have closed the [vote-by-mail] gaps by cannibalizing their own vote, but here is an interesting finding: Republicans are actually doing it more to themselves. Right now, about 34,000 more 2012 Election Day Republicans have early than Democrats.
If that is true, it means that a Republican edge among actual Election Day voters (which statistically was the likely outcome in 2012, given the Democratic early vote edge) might be muted this year. Which makes the current near 50/50 breakdown between the two parties less of a bad sign for Democrats.
For a deeper dive into the Florida early numbers, check out this post from our own community member Northern Pol.
NORTH carolina
On the other hand, the racial voting patterns in North Carolina are growing a little more difficult to explain. Last week, it was clear that white voters were overperforming their 2012 numbers in early voting, and African-American voters were underperforming their 2012 numbers. But the early lag in African-American in-person early voting had at least one simple potential explanation: the huge decrease in availability of early voting centers, particularly in areas with a high proportion of African-American voters.
However, with nearly 1.9 million votes tallied (according to the VoteTracker project of the conservative Civitas Insitute), there are some very troubling signs. Despite a number of “Souls to the Polls” events over the past weekend, the African-American share of the early vote in North Carolina has barely moved. As of this morning, black voters have made up 22.5 percent of the early vote. By contrast, in 2012, black voters comprised just over 27 percent of the early vote.
There’s always the chance that these numbers will recover as we head into the final days of early voting in North Carolina (they have slowly increased over the course of the past five days), or that black voters will simply make up a larger share of the Election Day electorate. But this is surely something that the Clinton campaign is keeping an eye on, and almost certainly explains why President Obama is heading to the Tar Heel State twice later this week.
OHIO
Here the evidence is a little more mixed, but still disconcerting. Earlier today, the Cuyahoga County Board of Elections estimated that their turnout for this election (an estimate based, in part, on early voter participation) will amount to 62 percent of registered voters. That is considerably lower than 2012, when turnout was 70 percent of registered voters. Given that (a) this is a 30 percent African-American county in a state that is only 12 percent African American, and (b) this is a county that went Democratic by a 69-30 margin in 2012, this would seem to be very bad news.
The offsetting good news, as we relayed in yesterday’s update, is that the two large Democratic vote sinks in the state (Cuyahoga and Columbus’ Franklin County) had both seen a distinct uptick in early voting performance over the weekend, putting them closer to their 2012 early vote shares, though still lagging. In addition, in Cuyahoga, Democrats are returning their mail-in ballots at a considerably higher rate than Republicans (74 percent to 64 percent), even though overall mail-in balloting has only amounted to about 20-25 percent of the expected ballots cast in the county.
FINAL THOUGHTS
Some of these numbers definitely should raise a healthy level of concern, but there is an important data point that might mitigate some potential trouble spots we’re seeing. A weekend review of polling data by Ipsos/Reuters suggests that Hillary Clinton is doing quite well among early voters in Texas, Arizona, and Ohio, and leading the overall national early vote by 15 points. We can readily explain Clinton doing well in Texas and Arizona (hint: Latino vote!), but Ohio’s story, demographically, has been that the early vote seemed to play right into Trump’s hands.
Now, this is but one poll, but it might indicate that, as with so many phenomena in 2016, what we think will happen and what happens might be two very separate things. To wit: Polls out of both Florida and North Carolina (the two states we flagged above) seem to hint that Clinton is leading the early vote. If she is doing this well in the early vote in states that, demographically, seem less friendly toward a Democratic nominee than in 2012, it might tell us that the white vote is not as Republican-leaning as we would guess, based on recent voter history.
We look for tea leaves in these early vote numbers (especially given the paucity of quality polling that is available this cycle), but it’s worth noting that we base our assumptions on what is good or bad for the two candidates based on what has been historically good or bad for the two parties they represent. But there is no guarantee, of course, that past is prologue. That’s why, while we continue to pore over early vote data, we must caution ourselves to set in concrete any conclusions we draw from the numbers we see.