Election Review, Part 2: The South Atlantic

Election Review, Part 2: The South Atlantic

By Jay Cost and Sean Trende - January 12, 2009

Today we continue our series on the 2008 election with a look at the South Atlantic Division.

This was an eventful part of the country on Election Day. Barack Obama swung three states here, as many as any other division in the country. What's more, he swung Virginia and North Carolina - neither of which Bill Clinton ever won.

As we shall see, this division was characterized by some important interstate differences. Accordingly, we'll be best served by dividing it into sections: West Virginia; Maryland, Delaware, and the District of Columbia; Virginia and North Carolina; South Carolina and Georgia; and Florida. Before breaking this division down by section, it is worth noting one consistent effect: the African American vote shifted toward Obama.

African American Vote in South Atlantic.jpg

The first thing we notice is that, with the exception of Georgia, the proportion of African American voters does not appear to have increased by an appreciable amount. In fact, in several states it appears to have declined. Nevertheless, Obama apparently received a boost in his share of the African American vote overall. In 2004, George W. Bush was able to win at least 10% of the African American vote in these states. In 2008, John McCain appears to have been lucky to win 5% anywhere in this division. As we shall see, this made a critical difference.

West Virginia

While West Virginia is a part of this division as far as the Census Bureau is concerned, its political behavior this cycle was more like the states in the South Central. Obama's performance here was one of the worst for a Democrat in the state's history. Let's begin by looking at how different types of areas have voted in recent cycles, again relying on our Census Bureau Statistical Area (CBSA) categories.1

West Virginia By Region, 1996-2008.jpg

As with the states in the South Central, McCain actually improved upon Bush in the rural and small town areas. This has been a crucial factor in the state's swing in the last decade or so. Clinton's statewide victory depended on big wins in these parts - and so did Bush and McCain's.2

West Virginia, 1996-2008.jpg

We note that Bill Clinton carried most of the state with only a few exceptions. Putnam county (suburban Huntington/Charleston), the areas in the Mid-Ohio Valley around Parkersburg, and the eastern panhandle went Republican, as they generally have since at least 1960.

In the last several election cycles the map has changed drastically. Both Kerry and Obama carried only a handful of counties in the state, for the worst Democratic showing since George McGovern. But unlike McGovern, Kerry and Obama did not lose national landslides.

Although the map from 2004 to 2008 shows only modest changes, the "yes/no" nature of its coding masks deep shifts. Although the eastern panhandle continued to vote for the Republican ticket, it shifted toward the Democrats in terms of voting percentage. This likely represents these counties' close proximity to Washington, D.C. and the northern Virginia media market, which makes the two easternmost counties (Jefferson and Berkeley) essentially extensions of the northern Virginia suburbs (on which more later). Obama carried Marion County (Fairmont, West Virginia) and added Monogalia County, perhaps reflecting his strength among college students at West Virginia University.

But if the eastern panhandle moved against its historic Republican roots, the southwestern portion of the state moved even more dramatically against its Democratic roots. This area - the coal producing area of the state - has voted Democratic since open class warfare erupted in the 1920s with the UMW's organizing of the mines. Even the hapless Kerry campaign carried McDowell and Boone counties by twenty points. Obama carried those counties, but by margins half as large. He also became the first Democrat in recent memory to fail to carry Logan County, site of a massive labor uprising in the 1920s, and the only county in the state George McGovern carried in 1972. Obama lost it by double digits.

In the last ten years, West Virginia has gone from being a reliably Democratic state at the presidential level to being reliably Republican. The reasons why are beyond the scope of this essay, though our intuition is that these reasons are similar to the reasons why the South Central region shifted away from Democrats, and that a Bill Clinton-type of Democrat could still carry the state with relative ease.

Maryland, Delaware, and the District of Columbia

Maryland, Delaware, and D.C. voted for Kerry in 2004 - and each gave Obama a larger share of the vote. We saw part of the reason above. While African American turnout does not appear to have been substantially higher in any of these places, Obama appears to have done better among African Americans in Maryland and Delaware than Kerry did in 2004.

The other explanation has to do with the white vote, as can be seen in Maryland.

Maryland Exit Polls, 1996-2008.jpg

Obama appears to have done better among white voters than John Kerry or Bill Clinton - and his improvement appears to have been across every category covered here. The same was true in Delaware and D.C. Unsurprisingly, given this data, Obama's improvement was geographically broad. He gained ground relative to Kerry in every county in Delaware and Maryland. Another notable feature: these states were not targeted by the Obama GOTV effort, and yet turnout here was up by more than it was in targeted Florida.

Virginia and North Carolina

Bill Clinton came close to winning both states in 1996, but fell just short. This cycle, Obama flipped both. Let's see how he did it.

We noted above that neither Virginia nor North Carolina saw a substantial increase in the share of the African American vote. However, we did see that African Americans apparently moved in Obama's direction. The two states have something else in common - white voters apparently moved Obama's way, too. So, Obama won both because of gains among whites and African Americans.

Let's start with Virginia.

Virginia Exit Polls, 1996-2008.jpg

We see that Obama seems to have improved among most subgroups listed here. The only exception is with white Democrats, who might have ticked a bit to McCain and the GOP. These voters might be located in the southwest corner of the state, in the ninth congressional district, which has some typically Democratic counties on the West Virginia border that voted Republican this cycle (some of which have not voted for a Republican since Nixon's re-election bid).

What is interesting here is that while McCain seems to have done worse than Bush among whites, he did just as well (if not a little better) than Bob Dole, who won the state in 1996. The big difference between Dole and McCain is that the former won 20% of the African American vote in the Old Dominion, while the latter won just 8%. The other relevant factor is that while the share of African Americans in the electorate between 2004 and 2008 was roughly constant, it seems to have been lower in 1996. So, McCain did worse than Dole among African Americans, who made up a greater share this cycle than they did in 1996.

Of course, while McCain did as well as Dole among white voters as a whole - it does not appear that he won the same white voters as Dole did. Unfortunately, we do not have exit poll data by sub-regions of the state, but we can examine candidate performance by CBSA.

Virginia By Region, 1996-2008.jpg

We might infer that the movement among whites in these CBSA categories is even greater than these graphs indicate; Virginia has several rural counties that are majority black, which would obscure any movement toward Republicans among rural whites. Obama did better in the large city category, namely the Hampton Roads and Richmond City areas, and he posted his biggest gain in the mega city category, namely metro Washington. Obama won nearly 3 out of 5 voters in greater DC - and this success was likely not solely due to his improvement among African Americans, since black voters make up less than 10% of the population in most of the counties and cities of northern Virginia. Given everything we have seen, it is probable that a shift among white voters played a role, along with shifts among the mix of Hispanic, Middle Eastern, and Asian minorities that inhabit Fairfax (which are difficult to measure by exit polling data).

So, we can tentatively conclude that white Obama voters were more "metropolitan" than white Clinton voters. A look at a map of the state over time will amplify this.

Virginia 1996-2008.jpg

We see that Clinton carried a good number of the mountainous counties on the West Virginia border, the heavily African American counties on the eastern North Carolina border, Nelson, Buckingham, and Prince Edward counties in the middle of the state (the latter of which was the site of the Virginia desegregation case that was incorporated into and argued with Brown v. Board of Education, and which closed its public schools for five years in the wake of the decision), and a few counties in the tidewater area.

Obama shed voters, even from Kerry's losing coalition, in the western portion of the state, carrying only Montgomery County (Virginia Tech). He and Kerry added Albemarle County outside fast-growing Charlottesville (UVa), and he performed well in the African American rural counties. He also added suburban Henrico county near Richmond, and carried some counties in the Hampton Roads area that Kerry and Clinton failed to carry. But the biggest gains are obvious, coming in northern Virginia. Obama became the first Democrat since LBJ to carry Loudoun and Prince William counties, and the second to carry Fairfax (Kerry was the first).

The same process seems to have occurred in North Carolina as well, as the following chart makes clear.

North Carolina Exit Polls, 1996-2008.jpg

Again, we see McCain did at least as well among whites as Dole, though he apparently did worse than Bush across all subgroups listed here. That was one source of Obama's success. The other was the African American vote. Again, it was not a greater share of the vote compared to 2004, but it seems to have been up compared to 1996; and McCain apparently did worse than Dole did among African Americans.

So, like Virginia, North Carolina flipped because of a shift among whites and African Americans. In both states, the movement of whites was possibly due in part to movement by whites aged 18-29. In Virginia, Obama improved upon Kerry by 5 points among white youths, from 38% to 43% of the two-party vote. Given the small sample size, this is not statistically significant - but the movement in North Carolina, where Obama improved upon Kerry by 22 points (going from 34% to 56%) was.

Also like Virginia, while whites as a whole voted this year as they did in 1996, there appears to have been intrastate variation. We cannot be sure, unfortunately, but we can get a good sense of things by looking at the vote by CBSA category. North Carolina lacks a metropolitan area on the scale of Northern Virginia, but we still can make some useful comparisons.

North Carolina By Region, 1996-2008.jpg

Once again, Obama did worse than Clinton in the rural and small town regions. He made his greatest improvements in the larger metropolitan areas - Charlotte, Raleigh, and Durham. Again, we cannot be certain, but it is reasonable to guess that Obama improved upon recent Democratic candidates among "metropolitan" white voters.

Let's examine the maps.

North Carolina, 1996-2008.jpg

Obama lost Clinton's strength in the largely white western panhandle, carrying only Buncombe (Asheville), Watauga (Appalachian State) and Jackson County (Cherokee reservation). In the Piedmont area he added Forsyth County (Wake Forest), Wake County (Raleigh), and some outlying counties with moderately high African American percentages (roughly 33%). In other words, judging by the maps, most of his improvement comes from adding a few counties with college towns, middling African American populations, or urban centers. This is consistent with our observations above.

South Carolina and Georgia

While whites as a whole shifted noticeably in the upper South Atlantic - there was little movement in the lower South Atlantic.

Starting in South Carolina, we see whites might have ticked a point or two toward Obama, but McCain still did about as well as Bush, and apparently better than Dole.3

South Carolina Exit Polls, 1996-2008.jpg

South Carolina is also interesting because it saw a higher increase in voter turnout from 2004 than any other state in the region except for North Carolina - even though the state was not targeted by either campaign's GOTV effort.

Georgia saw a fairly tight race this cycle, but this was not due to white voters, who supported McCain just as strongly as they supported Bush, and more strongly than they supported Dole.

Georgia Exit Polls, 1996-2008.jpg

Obama's improvement upon Kerry here was due to winning 98% of the African American vote, which seems to have been up by a good bit this cycle.4


Finally, we look at Florida. Obama enjoyed a 3-point victory here, but he did not improve substantially upon Kerry among white voters.

Florida Exit Polls, 1996-2008.jpg

There was no significant change between 2004 and 2008, as we can see here. Of course, this is not to say that all white voters voted the same way in 2008 and 2004. Instead, as with some of the other states we have reviewed, it might have been the case that some whites moved toward Obama while others moved toward McCain. Indeed, a look at the vote by CBSA category indicates that this is a real possibility.

Florida Results By Region, 1996-2008.jpg

We see Obama at his weakest in the rural regions. In Florida, he actually did worse than Kerry - thanks in part to a decline in Democratic fortunes in the Florida panhandle. Meanwhile, Obama did better than Kerry in metro-Miami. Additionally, he won the cities in the "I-4 Corridor" - metro Orlando and metro Tampa - neither Clinton nor Kerry did that.

Again, whites as a whole did not move. How then did the Sunshine State tip to the Democrats? We saw part of the answer above. Obama seems to have improved upon Kerry with African Americans. The other piece of the puzzle was Hispanic voters. Obama won 58% of the two-party Hispanic vote, compared to Kerry's 44% and Clinton's 48%. This was a major triumph for the Democrats - as the Hispanic vote in Florida has typically been reliably Republican, thanks to strong support in the Cuban-American community. We saw the same movement in Texas last week, and when we cover the West we will again see movement among Hispanics toward the Democrats. For Republicans, this might have been the most worrisome trend in November.

Here is the change in time in the Sunshine State. Note the shift in the pandhandle as well as through the center of the state, the "I-4 Corridor."

Florida, 1996-2008.jpg

What is especially interesting about Florida is that, compared to the country as a whole, it did not particularly move against the Republicans. In 2004 the state was 1-2 points more Republican than the country as a whole. In 2008, it was 2-3 points more Republican. This is somewhat surprising, given that the state was particularly hard hit by the subprime mortgage debacle, and has a large constituency that is particularly concerned about their 401k's. Given that the state is dominated by Republicans at the state and federal level, we might have expected movement more like we saw in Nevada.

In conclusion, two points are salient. First, like the South Central division we find Obama typically doing little better than Kerry, and worse than Clinton, in the rural and small town counties. Meanwhile, he typically improved upon Kerry and sometimes even Clinton in the large and mega cities. Thus, it's fair to conclude that, at least in the South, Obama's voting coalition was more urban and less rural than Clinton's.

Second, unlike the South Central division, racial polarization seems not to have increased in this division. In fact, in most states with large minority populations, it actually seems to have declined, as white voters in Maryland, Delaware, Virginia, and North Carolina apparently shifted toward Obama. Whites in South Carolina, Georgia, and Florida did not move appreciably - but it is important to note that Democrats already tended to do reasonably well among white voters in Florida. Kerry, for instance, won 42% of the white two-party vote in 2004. Thus, Obama's success in this division seems to have been due to a multi-racial voting coalition - one that included increased margins among whites, African Americans, and Hispanics, especially in the northern reaches of the area. This was not the case in the South Central regions, where white voters either did not move or ticked a bit toward the GOP.



[1]To review, here are our metropolitan codes, as developed via the CBSA's.

Labels for CBSAs.jpg

[2] The "mega city" we see for West Virginia is actually Jefferson County, on the tip of the eastern panhandle of West Virginia. It's classified as an outlying county in the Washington-Arlington-Alexandria Metropolitan area.

[3] If there was real improvement for Obama among white voters in South Carolina, it is possible that it occurred in metropolitan Charleston, where he improved nearly 6 points upon John Kerry. This would be consistent with what we found in Virginia and North Carolina - where Obama apparently did better with white voters as the metropolitan area grew larger. On the other hand, Charleston County itself is about 1/3 African American, which could account for much of this shift. Ultimately, this is the limitation when dealing with vote data in racially heterogeneous counties.

South Carolina By Region, 1996-2008.jpg

You'll note here that Obama did better in the rural regions than he did in any other place we have examined. The reason for this is that African Americans make up a greater proportion of the rural counties in South Carolina.

[4] Perhaps unsurprisingly, Obama saw his greatest improvement upon Kerry in metropolitan Atlanta.

Georgia By Region, 1996-2008.jpg

It is possible that he improved among white voters here - but did worse with whites in other parts of the state. That would be consistent with what we seem to have found in Virginia and North Carolina - and it would help explain how he did better than Kerry statewide, but worse in the rural counties. On the other hand, metropolitan Atlanta has a massive African American population; the movement toward Obama there is hardly unsurprising given that fact. Ultimately, the data we have does not allow us to confirm this intuition.

Sean Trende can be contacted at

Jay Cost and Sean Trende

Author Archive

Follow Real Clear Politics

Latest On Twitter