It's no secret that the way for Barack Obama to win North Carolina is for him to expand the electorate and turn out "unlikely" voters.
Previously in my NC Early Voting Diaries, we have looked at early voting primarily in aggregated numbers, without really knowing whether the people who are voting are likely or unlikely voters. But today I have the data ready to allow us to peer under the hood and see whether the people who are voting are likely voters or not. As it turns out, there are quite a few Sporadic, Unlikely and New Registrant voters - and it is clear that Obama is cleaning Romney's clock among this category of voters.
Specifically, as I define them, 56.7% of the people who have voted so far are likely voters. But those 56.7% of voters contribute only 20% of Obama's estimated vote margin. The remaining 80% of Obama's estimated margin comes from where it matters - from the 43.3% of voters who fall into the categories of Sporadic, Unlikely, and New Registrant Voters.
We will look in more detail at the numbers on this below, but first let's look at some indicators that OFA is seeing something good in the NC early vote numbers:
Michelle Obama is Coming to Charlotte on Monday:
First lady Michelle Obama is coming to campaign in Charlotte, NC on monday, the day before election day. Oh yeah, and Bill Clinton is coming to Raleigh on Sunday.
Now, why do you think that OFA would send Michelle Obama to Charlotte? Well, for one thing, she was definitely a hit the last time she was in Charlotte:
But why would OFA send Michelle Obama, one of the most effective advocates for Barack Obama, to North Carolina on the very day before election day? If we believe all the electoral college maps that show North Carolina colored in dark red, wouldn't that be a terrible misallocation of resources?
If North Carolina were a lost cause and unwinnable, do you think they would send her to Charlotte?
No way! If that were the case, she'd be spending all her time in Florida, Ohio, Virginia, or Colorado, especially one day before election day.
Some have argued that the only reason Michelle Obama is going to Charlotte is to reward people who volunteered at the Democratic National Convention but didn't get to see President Obama's DNC speech because of the rain. But if that were the case and this were really just a token gesture to make disappointed volunteers feel better, why would Michelle Obama be coming to Charlotte on the day before election day, rather than much earlier? If it were just a make-up event, then you would think that they would have the makeup event at a less critical time. Moreover, OFA continues to pour millions of dollars into television advertising in North Carolina, and the Obama ground game continues its all out effort to turn out ever more new voters.
So what are the numbers that are persuading OFA that NC is worth the continued investment of time, resources, and money?
First of all, the polls show a very close race in North Carolina (unless you believe Rasmussen and Gravis Marketing).
Over roughly the past two weeks, PPP has found North Carolina tied not once but twice, and Elon University has also found a tied race. Civitas and High Point University have both found Romney up by only one point.
As we all remember, Obama was down in just about every poll in NC in 2008, but he won anyway thanks to OFA's superior ground game, which converted "unlikely voters" into actual voters.
Obama is Turning Out Sporadic Voters, Unlikely Voters, and New Registrants:
But perhaps more important than polls are the hard numbers of actual voters who have voted. President Obama has built up a large lead in early voting, but more importantly, he has built up that lead where it counts. OFA has not concentrated on turning out likely voters that it knew would vote regardless, whether on election day or early. Instead, the Obama campaign has focused its attention on expanding the electorate to include Sporadic Voters, Unlikely Voters, and New Registrants.
In a recent summary of early voting trends nationwide, OFA Field Director Jeremy Bird wrote:
Our campaign is about inclusiveness, and we’ve always been focused on bringing new people into the process. As our supporters vote early in huge numbers around the country, we have the perfect example: We’re turning out voters who have been traditionally less likely to participate, sometimes called “sporadic” voters.
A common misconception about early vote is that both parties have a set number of voters, and all early vote does is let some of them cast their ballots before Election Day. That’s simply not true. What early vote does is help us mobilize sporadic voters by giving them more time and more convenient ways to make their voices heard. It also broadens the universe of voters and frees up more of our get-out-the-vote resources later, especially on Election Day. When you look inside the numbers so far, among sporadic voters it’s not even close.
More sporadic Obama voters are voting than sporadic Republicans in the battleground states.
Non-midtermvoters: Across nine battleground states, Democrats have a 19.7 point advantage in ballots cast among non-midterm voters. More than half (51.5 percent) of non-midterm voters who have voted already are Democrats, while fewer than a third (just 31.8 percent) are Republicans.
For example, in North Carolina, 51.5 percent of those who have already voted are Democrats, compared with just 25.1 percent who are Republicans. That’s a major advantage. And among these non-midterm voters who have voted in North Carolina so far, 87 percent of them are youth (under 35), African-American, Latino, or new registrants (registered after the 2008 election).
This is not something unique to North Carolina. The same thing is happening in Florida, where OFA Florida Director Ashley Walker says
"This isn't 2008. We don't have 15 days of early vote. We have 8 days, and so it's a different race," Walker said. "When you really dig down and start looking at at these numbers in who is turning out with these vote-by-mail numbers and early vote numbers, more of our sporadic, irregular voters than theirs by a three-to-one margin. And that means we have more old faithfuls to come out on election day. I'm not going to try to bullshit you - it's a tight race, it's a really close race, but any spin they're trying to feed you that we're behind where we were in 2008 is just spin....It's a totally different race. The opportunities and the rules of the game are totally different."
- And, in a new OFA memo
(warning, PDF file), we learn that:
This cycle, our teams registered 1,792,261 voters in key battleground states – nearly double the number of voters the Obama campaign registered in 2008. These new voters are already voting in early vote states. In fact, 28 percent of them – 345,233 – have already voted. In North Carolina, for example, 137,808 of the voters OFA-NC registered have already voted in a state the President won by just 14,000 in 2008, and that will come down to the wire again on Tuesday.
In total, 335,687 new registered voters have voted so far in NC, so that means that 41.0% of all registered voters who have voted so far were registered by OFA. That is a very striking number. (Thanks to MBishop1 from DKE's Saturday Open Thread for the link!)
Breakdown of Voters by Vote Likelihood:
For this analysis, I am defining Likely Voters, Sporadic Voters, Unlikely Voters, and New Registrants as follows:
Likely Voter - Someone who voted in both the 2008 and 2010 General Elections.
Sporadic Voter - Someone who voted in either the 2008 or 2010 General Election, but not in both.
New Registrant - Someone who registered to vote after the 2010 General Election
Unlikely Voter - Someone who did not vote in either the 2008 or 2010 General Election, and who is not a New Registrant.
And I am assigning votes from each combination of Party Registration and Race as follows:
So for example, I am counting each Unaffiliated white voter as .28 votes for Obama and .72 votes for Romney. These numbers are based off of exit polls, as explained in the Day 1 North Carolina Early Voting diary.
So, the table below shows how the 2,524,217 North Carolinians who have already voted break down by party registration, race, and voter category (likely voter, sporadic voter, unlikely voter, or new registrant). There is also another category called "Reg After 10/27." These are people who do not show up in the voter data that I downloaded from the North Carolina board of elections. The voter data was updated on 10/27/2012, so I assume that these are either entirely or largely new registrants who registered and voted in one stop during One Stop Early Voting after 10/27. Below is a chart with the breakdown:
Overall, President Obama is beating Mitt Romney by an estimated 1,333,353 votes to 1,190,731 votes, a margin of 142,623 votes. This is lower than Obama's 2008 Early Voting margin, but a small margin from unlikely voters is arguably better than a large margin from likely voters, since most likely voters will vote on election day.
And indeed, as it turns out, Obama's estimated vote margin is disproportionately coming from "sporadic voters," "unlikely voters," and "new registrant voters." This (much simpler) chart illustrates that point:
46.6% of Obama's estimated margin is coming from "sporadic" voters who voted in either 2008 or 2010 but not in both, even though they only make up 24.7% of early voters.
Another 6.7% of Obama's estimated margin is coming from "unlikely" voters (people who voted in NEITHER 2008 nor 2010 even though they were registered before election day of 2010). These voters make up only 5.3% of early voters so far.
Another 26.7% of Obama's estimated margin is coming from the two categories of New Registrant voters who registered to vote after the 2010 election. These New Registrants only make up 13.3% of early voters.
But what if Obama's margin is really bigger?:
Several commenters have pointed out that Obama's margin may be bigger than my estimates. Indeed, this could be the case. Every single poll of NC Early Voters so far has found that early voters are voting for President Obama by a greater percentage than my projections would indicate.
So, what happens to Obama's margin if we change our assumptions a bit? What if we assume the following?:
1) White Democrats who are Sporadic Voters, Unlikely Voters, and New Registrants are less likely to be Dixiecrats than "Likely Voters." So instead of voting for Obama at a rate of 83%, White Democrats in this category are voting for Obama at a 90% rate.
2) White Unaffiliated Voters who are Sporadic Voters, Unlikely Voters, and New Registrants are more likely to support Obama So instead of voting for Obama at a rate of 28%, White Democrats in this category are voting for Obama at a 50% rate. Because the number of registered "unaffiliated" has gone up by so much since 2008, and because so many students, people with advanced degrees, and people from the northeast are registering as unaffiliated, unaffiliated white voters are more Democratic than I have been assuming.
Although I can't say that I completely buy into those two assumptions, they are clearly not are not entirely implausible. So what happens if we change those two assumptions?:
As a result of these changes, our estimated Obama margin goes up to 256,611 - almost as large as Obama's 2008 vote margin from early voting. Instead of winning only 52.8% of early voters, this means that Obama would win 55.1% of early voters, which is closer to what polls have been saying.
Who has voted - and who has NOT voted?:
Now let's look at who has voted - and perhaps more importantly, let's look at who hasn't voted. In a previous diary, I noted that if total turnout increases from 2008 at the same 5.9% rate as the Voting Age Population has increased since 2008, then we should expect total turnout in 2012 to be about 4,563,876 voters. That means that we should expect about 2 million people to vote on election day.
Whether Obama is ahead by 142,623 votes, 256,611 votes, or by some other number of votes, we know that more White Republicans have voted early in 2012 than in 2008. But that means that there are fewer White Republicans left over in the pool of 3,427,143 active registered voters, from which our roughly 2 million election day voters will be drawn. Are there enough White Republicans left over for Romney to make up Obama's early vote lead and win?:
There are 989,839 Likely voters left who have not yet voted. We know that some very large percentage of these 989,839 people will vote on election day. 381,106 White registered Republicans make up 38.5% of these 989,839 people, and only 12.6% of those 989,839 "Likely Voters" who have not yet voted are African American. However, if in addition to African Americans, you throw in White Democrats and other Minority voters, then 41.7% of the 989,839 people are either African American, other minorities, or White registered Democrats.
The problem for Mitt Romney is that even if ALL 381,106 Likely Voting White Registered Republicans vote, and even if ALL 381,106 of them vote for Romney, that only gets Romney 381,106 votes, which is not nearly enough. In order to win, he will need, as a rough minimum, about 1.1 million votes. So that by itself does not get Romney where he needs to be.
The total number of White Republicans who have not voted - of ALL vote history categories - is only 995,165. Thus, even if every single White Republican who has not yet voted votes, and even if all of them vote for Romney, that would not get Romney enough votes to win. Of course, that will obviously not happen. And likewise, it is equally obvious Romney will get many votes from Unaffiliated voters and from White Democrats. But how many? Who will the 2 million election day voters be?
And that is the central problem and the central question that will be answered in North Carolina on election day. In order to make up Obama's lead, Romney needs not only extremely strong election day turnout from White Republicans, but he also needs to do very well among White unaffiliateds, and he needs to peel off a large number of White Democrats. If Obama can stop that from happening, and if Obama can turn out enough Sporadic, Unlikely, and New Registrant Democratic voters, then Obama will win North Carolina.
Obama has a large pool of potential voters to draw from. As we noted, 38.5% of Likely voters who have not yet voted are White Republicans (WR) and 41.7% of Likely Voters who have not voted are either African American, Other Minorities, or White Dems (AA/OM/WD).
But once we move past Likely voters, the share of White Republicans plummets and the Democratic advantage balloons.
Among Sporadics who have not yet voted, 43.6% are AA/OM/WD, and only 27.7% are WR.
Among Unlikelies who have not yet voted, 43.3% are AA/OM/WD, and only 24.5% are WR.
And among New Registrants who have not yet voted, 38.4% are AA/OM/WD, and only 21.1% are WR.
African American Voter Turnout is Ridiculously High:
One other thing that immediately pops out from this data is just how astronomical African American turnout in North Carolina has been. Across every voter category - whether likely, sporadic, unlikely, or new registrant, African American turnout has soared far above turnout of all other voter groups:
There are two ways to interpret this fact:
1) Because African American turnout has been so high in early voting, there are fewer African Americans left to vote on election day.
2) Because African American turnout has been so high in early voting, this signals that African American turnout among those who have not yet voted is likely to be proportionally very high on election day (even if it is smaller in raw numbers than White Republican turnout).
I would guess that both of the above interpretations are partly correct.
Voting on Friday:
At least 242,476 people voted on Friday, with more undoubtedly to be added as counties process their voter records.
Overall in the last week of early voting - not including Saturday -, turnout was clearly lower than in 2008. In part, this may be because of Hurricane Sandy, especially in the early part of the week. But another part of it may have been that so many people voted in the first few days of early voting that there were not as many people left to vote in the last week. Nonetheless, cumulatively 129,062 more people have voted than at the same time in 2012.
This represents a 5.3% increase in early vote turnout over 2008 - which is pretty close to the overall 5.9% increase in turnout that would be required to get to 4,563,876 total votes.
Click on the picture below for a full sized chart.
Since Monday, the estimated vote split between Obama and Romney has been extremely close - although Obama has continued to add to his total vote margin. In fact, on Friday, Obama added to 2,939 votes to his vote margin, which is not that far behind the estimated 5,018 votes by which Obama won the same Friday 4 years ago. And again, that number will likely go up as more data is processed by counties.
Early Vote Turnout Charts:
Minority and White Democratic turnout continues to increase, and continues to be high, but after Sandy has fallen back to 2008 levels.
White Republican early vote turnout continues to be high, but of course as we noted above, this reduces the pool of White Republicans who may vote on election day.
Obama's cumulative margin has leveled off at a bit under 150,000 votes, which is clearly less than in 2008. However, as discussed above, if you tweak the assumptions a bit in ways that are not entirely unreasonable, you can get a much larger Obama margin that is more consistent with the polls.
Obama's daily margin approached his 2008 daily margin more closely than it has in recent days.
Obama's estimated cumulative vote percentage is converging somewhat with 2008, but it still a bit below.
And after Sandy, Obama's estimated daily vote percentage has very clearly converged with 2008.
Previous NC Early Voting Diaries:
Day 1 & Methodology