We have nearly final figures on the early voting totals in each state, and the ballot return rates continue to force us to reconsider many of our previous assumptions about who votes early and where early voting is most popular.
Early voting as a percent of the 2012 early vote, as shown in the first map, is blowing old totals out of the water. Charles Stewart has reported on the pace of early voting in North Carolina and Daniel Smith has been doing the same for Florida. Other states with extraordinarily high numbers compared to 2012 include Arizona (106%), Georgia (124%), Maryland (167%), and Nevada (109%). As I noted in an early posting, the Massachusetts and Minnesota numbers are misleading because they have relaxed their early voting laws.
These numbers are more impressive when viewed as a proportion of the total 2012 vote. This indicates higher turnout overall, or an electorate that is switching wholesale to early voting. In either case, it means that in the states that rank high on this list (Arizona, Colorado, Florida, Georgia, Nevada, North Carolina, Oregon, Tennessee, and Washington), it’s not going to be enough any more to have an election day GOTV operation.
Election day is rapidly becoming election week(s) in many parts of the country.
Just wow, that’s all I can say. Off to class!
A couple quick posts before I head off to class. Let’s start with the most up to date EV rates compared to 2012 rates. These can be heavily influenced by changes in state laws, as in Massachusetts, so interpret with caution.
(This post co-written by Paul Gronke and Brian Hamel, Department of Political Science, UCLA)
Even the prospects of ghosts, goblins, Russian hackers, and Wikileaked emails don’t seem to be able to stem the tide of early voters in 2016. It’s truly a historically early early vote total nationwide. (Take that sentence, Doug!)
Five days ago, we reported that in a number of key battleground states, close to 50% of all early votes cast in 2012 had already been cast in 2016. With less than a week till Election Day, these numbers continue to climb
The highest rate is in Louisiana, where the total early votes cast this cycle have already exceeded the total cast in 2012 over six more days of balloting. Sec’y Tom Schedler talked about this last week at the Bipartisan Policy Center. This may be a result of good election planning. After all, if the Tigers beat ‘Bama this Saturday, there could be a few folks who won’t wake up until after Election Day! Something unique is happening in the Bayou State.
Florida, Georgia, and Tennessee are also showing very high ballot return numbers, at or exceeding 80% of the 2012 totals.
Overall, twenty-nine states have racked up more than half as many early votes as were cast over the full early voting period. For many states last week, these totals only included no-excuse, vote by mail ballots, but now millions of more voters are heading to early in-person voting locations in most major battleground states.
Out of Politico’s eleven battleground states, in fact, only New Hampshire and Pennsylvania don’t report more than half of the early vote has already been cast (if we use 2012 as a baseline). This is going to create a huge problem for Donald Trump if, as some sources report, his early vote targeting operation lags Hillary Clinton’s.
The numbers look even more forbidding for Trump if you look at the early vote totals as a percentage of the total 2012 vote. These numbers have to be viewed with a bit of caution; state registration totals will have grown and shrunk during this period, and 2012 voter turnout is obviously an imperfect predictor of 2016 total. However, these numbers aren’t going to have moved that much, and more importantly, the relative position of states will have changed even less.
In Florida, nearly a requirement for a Trump victory, early votes exceed 44% of all votes cast in 2012. In Nevada, not a key to a Trump victory, but a state that Harry Reid wants to deliver for Clinton, 45% of the 2012 total has already been cast. In Tennessee, things are looking brighter for Trump–over half of the 2012 totals are in the books, although we have no data on the partisan breakdown of these ballots.
Michael McDonald notes that only in Nevada do the party ballot return totals indicate a clear Democratic advantage, and the election electorate will break for Trump (it has to if more Democratic votes are cast early). Still, this may mean that Clinton can redeploy her targeting resources in Florida and other states in a way that Trump may not be able to.
These stunning figures could be the result of a number of factors, including voters simply responding to the convenience of early voting, the intensity of the campaign, the distinctiveness of the candidates, increased attention from the campaigns and parties on early voting mobilization, or any combination of the three. What is clear, however, is that early voting could be a key indicator of who is winning the horse race.
Data courtesy of Michael McDonald and the United States Election project.
(This is a guest posting from Brian Hamel, PhD student in the Department of Political Science at UCLA)
Early ballots are flowing in rapidly to local election officials throughout the country. As of October 26, according to Michael McDonald of the United States Election Project, over 12 million ballots have been processed.
This is a stunning figure, given that we are 13 days away from Election Day and 10 days before a number of key battleground states end early in-person voting. (You can find the whole early voting “schedule” on our Early Voting Calendar)
Even more surprising is how many early votes have already been cast as a percentage of all early votes cast in 2012. To illustrate this, we used McDonald’s data and display early voting rates (darker red indicate a larger percentage of early ballots cast in 2016 as a percentage of all early ballots cast four years ago).
As shown, in a number of states, these figures approach or even exceed 50%. This is particularly true of a number of key battleground states. For example, in Florida, Iowa, and Virginia, 42.5%, 49.9%, and 49.2%, respectively, of 2012 early ballots have already been cast in 2016. We also see one of the largest shares in Arizona (50.7%), a traditionally red state where Clinton and Trump are in a statistical dead heat and where Clinton recently placed a $2 million ad buy .
These figures suggest that early voting is becoming an increasingly important aspect of American elections. Observers have long noted the sheer convenience of early voting, and the trend over just four years suggests that voters are responding to the convenience of these options.
Indeed, these trends are likely to structure campaign strategy over the final few weeks, as campaigns look to early voting returns—and the partisan breakdown of these returns—before deploying their resources. Stay tuned for more data and commentary on early voting in 2016!
This one in Green Bay, WI. At least it doesn’t concern Packers fans!
The local clerk cites “partisan advantage,” among other reasons, why she refused to establish a satellite early voting location at University of Wisconsin, Green Bay. It doesn’t help her out when the person requesting the location is a local Green Bay Democrat.
I’ve written in the past about ongoing conflict over an early voting location in Boone, NC, located at the student center at Appalachian State University, where I taught for two years. The ASU location also leaned heavily Democratic relative to the surrounding county, but in the Boone case, the student center also was the hub of a county-wide bus system, had ample parking, and was within easy walking distance of downtown. The county offices, in contrast, had little parking and were served by fewer bus lines.
The ASU location, in summary, was actually an excellent location!
This is a hard one! Not sure what the right call is on this location. Anyone with more information on the ground?
Journalists are hungry at this point for any information that might give them some special insight into the eventual outcome in November, and early voting seems particularly valuable because, unlike a poll, early voting reflects, obviously, actual voting. What’s much less clear is what we can conclude from early turnout at this juncture. As Michael McDonald warns routinely in his blog posts, these are early indicators, and may reflect changes in campaign mobilization, changes in state and local practices, along with voter interest and enthusiasm.
We have, for example, no research that I know of that correlates absentee ballot requests, returns, and eventual vote totals in a state. Comparisons between 2016, 2012, and 2008 seldom take into account in a systematic way spending rates and mobilization efforts by campaigns. Speculation about these things may make for good copy, but I don’t spend a lot of time poring over these numbers.
I am quite willing to speak to reporters, however, on the topic of “voter regret,” and whether early voting is “bad” because of the dynamic nature of “this year’s campaign.” These questions come in predictably just about this time, when the early vote totals really start to pile up, and reporters are on full alert, filing multiple stories about the campaign.
Here’s the basic math: only a relatively small percentage of total votes have been cast at this point (two weeks out from Election Day). And a relatively small percentage of registered voters (and an even smaller number of likely voters) are truly undecided at this late stage. This translates into a tiny theoretical percentage of voters who will cast an early vote and who are likely to change their vote, ignoring in that calculation lots of research indicating that most early voters are decided voters who are not going to be swayed by information provided over the next two weeks.
The best estimate we have of the total numbers of ballots cast at this point is 6.5 million. This is take from McDonald’s daily update on twitter. 
This sounds like a big number, but we expect over 130 million ballots to be cast in 2016. The early vote totals to date are about 5% of the total expected. Three days ago (October 21), it was less than 4%. Five days ago, the total was under 2%. That bears repeating: 20 days before the election, less than 2% of ballots were cast early.
Now, let’s see how many voters are really likely to be undecided at this stage of the campaign. According to Lynn Vavreck, a YouGov conducted three weeks ago found just 8% of registered voters were undecided. These undecided “… are less interested in politics and the news, less partisan, and less likely to hold opinions on issues dominating campaign discussions. Essentially, they think less about politics.” A HuffPollster study conducted a week ago found a somewhat higher percentage of undecided, 14%, yet Natalie Jackson found little space for movement within this segment.
Now we can put everything together. Who are these “early” early voters? Are they voters who will change their minds? All the evidence we have accumulated from past elections says “no.” Few early voters report that they regretted their choice, and most early voters report that they made up their minds relatively early. They tended to be more ideological and partisan than the average voter. Now, we have never had a study that samples early voters on the fly, after they cast their ballot, and follow up later on. Instead, we have to rely on post-election reports, and there may be some rationalization going on.
Nonetheless, as much as it may disappoint reporters, for the vast majority of citizens, the campaign is essentially over. It’s possible that there may be a late breaking scandal, or late breaking information, or a late breaking event that could move millions of citizens to reevaluate these two candidates who have been in the klieg lights for nine months. Every reporter is dying to break this year’s October surprise. But those surprises almost never occur, and given how strongly voters feel about these two candidates, there is very little that will move voters at this point.
Election “day” has begun! Enjoy the ride.
In fact, millions of absentee ballots will be arriving today at county elections offices. These ballots may have been delivered right on time by the postal service, or dropped off at a drop box, or hand carried into the local elections office.
In so called “postmark” states (Alaska, California, Illinois, Maryland, New York, North Dakota, Utah, Washington, and Utah), ballots need only be postmarked by November 8th (oddly enough, November 7th in Utah) and can arrive a few days to two weeks later.
Finally, in a few states, such as Arizona, Montana and California, you can even drop off your absentee ballot at the local polling place. (Editor’s Note: If anyone knows a comprehensive list of these states, I’d appreciate a link.)
These late arriving ballots can easily make the difference in close races.
In Multnomah County, OR, elections director Tim Scott estimates that anywhere from 110,000 to 150,000 ballots will arrive today–that’s more than 20% of the total registered voters in the county, so likely more than 30% of the final tally! Statewide, assuming total turnout of 80%, nearly 500,000 ballots that will be making their way into county offices today.
In Maricopa County, AZ, the second largest election jurisdiction in the country, over 100,000 ballots are typically dropped off on election day, according to Tammy Patrick of the Bipartisan Policy Center who worked in Maricopa for a number of years.
What happens to these absentee ballots? Are they counted immediately or are they counted at the close of polls? What about the absentee ballots that arrived prior to Election Day? I’ve received a flurry of questions about this today.
This means that in most states, the first results will include absentee ballots that arrived prior to Election Day, but will almost certainly not include ballots that arrived on Election Day.
The late arriving ballots will generally not even begin to be processed until Wednesday. While this may frustrate politicians, their supporters, and Americans who want to see an announcement of the final results, the delay is necessary to assure the security and integrity of our elections system.
First, in many states, absentee ballots cannot be processed before the close of the polls because election officials have to check to make sure that no one has voted twice, once by absentee ballot and a second time at a polling place. This is the kind of security measure that exists in the American election system and is often ignored by those making unsubstantiated charges of “rig
ging.”
Next, the absentee ballots need to be processed. This involves a multistep
process:
In most states, none of this will happen on election night, when election officials have already worked a very long and arduous day. Election officials need their beauty sleep, just like the rest of us!
And they’ll be working hard to provide full and accurate results for days and weeks after November 8th.