Southern Whites didn't vote for Obama (and by association, won't vote for Democrats). So argues the New York Times' Nate Cohn in a provocative piece entitled "Southern Whites' Loyalty to G.O.P. Nearing That of Blacks to Democrats." Mr. Cohn goes on to warn, "The collapse of Democratic support among Southern whites threatens the party's ability to control government and enact its agenda."
Of course, this flies the face of states like Virginia and Florida which Obama successively won in 2008 and 2012. Mr. Cohn dismisses these victories, noting that an "...influx of Northerners has transformed metropolitan enclaves of Virginia, North Carolina and Florida into liberal bastions of Yankee expats, much of the South remains largely untouched."
Fortunately for Democrats, Mr. Cohn's analysis of the changing nature of Southern electorates falls short. Starting with the Democrats' embrace of Civil Rights in the 1960s the South drifted from the old Solid South of monolithic Democratic control towards Republican control, reaching a peak in the last decade. Since then states like North Carolina and Virginia have drifted back towards the Democrats. Obama even lost Georgia, the next Southern state that may be turning purple, by only 5.2 percentage points in 2008 and 7.8 points in 2012. Georgia's 2.6 point swing from 2008 to 2012 was less than the national 3.4 change, suggesting a continuing trend in Georgia towards the Democrats even as Obama's margin of victory narrowed in 2012.
So, are Southern Whites' loyalty to G.O.P. nearing that of Blacks to Democrats, as Mr. Cohn's headline boldly proclaims? To assess this claim, to understand how Virginia turned purple, and to assess how Georgia may soon follow suit, I plot the 2008 % Obama vote (computed for Obama and McCain vote shares only, a.k.a. the "two party vote") against the 2010 Census % Black voting-age population in all precincts in Georgia, South Carolina, and Virginia. (I analyze 2008 data that I created for these states for redistricting purposes.)
These graphs demonstrate in all three states that the most heavily African-American precincts voted nearly unanimously for Obama. No surprise here. However, while some of the most heavily White precincts did vote at similarly high rates for McCain, there is large variation among these precincts. Indeed, even in South Carolina Obama was nearly tied with McCain in some of these White precincts. Based on these graphs we can easily rate as "false" Mr. Cohn's assertion that Southern Whites' loyalty to the G.O.P. is nearing that of Blacks to Democrats, as political scientist Dr. Bartels took Mr. Cohn to task for. This may be true for some Southern Whites, but not all of them.
Can Georgia become a battleground state? Three distinct features about the Virginia graph that are different from the South Carolina and Georgia graphs help answer this question.
First, there are clearly more heavily Black precincts in Georgia than the other two states. According to the 2010 Census, 30 percent of Georgia's voting-age population is Black, compared to 27 percent in South Carolina and 19 percent in Virginia. Given African-American's high support for Democratic candidates, this makes it easier for Democrats to turn the state purple.
Two other factors are working against the Georgia Democrats, however. In Virginia, Obama support in the heavily White precincts ranged between about 20 percent to 70 percent. In Georgia (and similarly in South Carolina), Obama support in the heavily White precincts ranged between about 0 to 40 percent. These numbers appear in line with survey data reported by Dr. Bartels, showing Southern White support for Obama in 2012 at about 30 to 35 percent. Obama's vote share is thus not only shifted downward in Georgia compared to Virginia, but the range is narrower, too. As a consequence of this narrower and lower range in the Whitest precincts, the slope of Obama's support from the Whitest to the Blackest precincts is more steep in Georgia and South Carolina than in Virginia.
Virginia turned into a battleground state because Whites are willing to join in coalitions with African-Americans to elect Democratic candidates. In comparison, Georgia Democrats appear to have their work cut out for them. However, there is another important distinction between Virginia and Georgia that may give Democrats hope. The Obama campaign expended considerable resources in Virginia in 2008, but after briefly flirting early with the idea of playing in Georgia, the campaign turned its attention elsewhere. The greater willingness of Virginia voters to support Obama may thus be a campaign effect. A Democrat with a well-funded campaign and an attractive candidate to Georgians could very well win a statewide election, as may happen this November.
UPDATE (4/25/14 at 10am)
In response to comments that the graphs may be affected by Hispanics and Asians, I've re-computed the graphs to display the % White non-Hispanic voting-age population on the x-axis. I've then taken the inverse so that the graphs align with the ones above, such that a greater value represents higher minority concentrations. The substantive conclusions remain largely the same: some Southern Whites are nearing a level of support for McCain similar to Blacks for Obama, but not all, or even most.