Is Georgia in Play?

There is a new poll out today from Democracy Corps that shows Obama down only 2 points in Georgia. That's a substantially closer margin than most other recent polls, but it does beg the question, "Could Obama win Georgia?"

After the Mississippi primary back in the spring, I wrote a post asking that same question about Mississippi and the answer was basically, "no." In order for Obama to have a realistic shot in Mississippi, African American turnout would have to go up by about 10 percentage points over 2004 levels, or Obama would have to win substantially more white voters than Kerry did.

Georgia, it turns out, is a somewhat easier task, but still a long shot. Take a look at this graph:



This grpah shows the percentage of white voters that Obama would have to win in order to carry the state, given a variety of other assumptions (I assume 5% of the electoral will be both not White and not Black and that Obama will win 60% of these voters). The three blue lines correspond to differing levels of Black support for Obama, and on the X axis you have increasing degrees of Black turnout as a percentage of total voters. You can see pretty easily that as Black turnout increases, the share of White voters Obama needs to win goes down. That makes perfect sense. Also, the stronger the Black support is for Obama, the less White support he needs. Again, no surprise there.

Now, for the moment let's look at the middle blue line, which assumes 90% of Black voters will cast a ballot for Obama. This is a reasonable, though perhaps conservative, assumption given that 88% of African American voters supported Kerry four years ago. With 90% of the black vote, and 2004 turnout levels (as indicated by the red arrow on the left), you can see that Obama would need just under 35% of the White vote to win. Kerry got 23%, and even in today's very optimistic Democracy Corps survey, Obama's only winning 24% of Whites.

So, basically, Obama needs three ingrediants to win Georgia (or any state in the South for that matter):

1) Strong support among Black voters approaching unanimity
2) Black turnout at record levels
3) Slightly stronger support from White voters than Kerry got

To be more specific, for a squeaker win, Obama needs better than 90% Black support, an electorate with about 35% Black voters, and 25% of the White vote. Can he do it? Let's take them one at a time.

1) Better than 90% Black support? The Democracy Corps survey pegs his Black support at 90% even. The second most recent survey, from Research 2000, has it at 91%. Interestingly, both surveys show McCain with anemic levels of Black support, far lower than even Bush got four years ago (2% and 5% respectively, compare to 12% for Bush). Both polls, therefor, suggest that there are between 5 and 7% of Black voters who are still "undecided." If those folks split evenly, or even largely to McCain, that means Obama's support among African Americans will be closer to 92 or 93%. So, first ingredient: check.

2) African Americans make up about 30% of Georgia's population, but were only 25% of Georgia's electorate four years ago. Is a ten percent jump possible? The early data says that maybe it is. Early voting has already begun in Georgia, and already close to 900,000 ballots have been cast. That's 27% of the total ballots cast four years ago. So far, African Americans have made up about 35.5% of all early voters. Can that rate hold througout the election? I don't know, but for now, second ingredient: check.

3) Democracy corps has 24% of White voters supporting Obama, Research 2000 has it at 21%. Of course, these are polls, so we can't really know exactly where Obama's White support currently is. At the very least, it's not outside the realm of possibility that he gets to 25%.

The bottom line is that Obama has a real, but slim chance to win Georgia. This might be one state to watch on election night (call it an upset special, if you will).