I’ve been pretty interested in gubernatorial races recently and I was struck by how politics in the south often moves in one direction only, so if you do about as well twice, the differences probably reflect long term changes, rather than short term quirks. I decided to look at our medium term trends. I compared our most recent results to the last time we got a reasonable similar result. Then I mapped out county level shifts. The colors represent a shift towards (blue) us or away (red) from us. The shades are increments of 10%, although the darkest shade covers anything >40%.
Georgia
- 2014 result: Deal 52.8- Carter 44.8, 8.0 pt loss
- Campaign comparison is 2002: Perdue 51.3- Barnes 46.2, 5.1 pt loss
- Reddest swing: Franklin County, which has about 22,000 people, 87% white, and lies in the north, on the SC border. We went from a 23 pt loss to a 63 point loss
- Bluest swing: Rockdale County, pop 89,000, 41% white, and in the Atlanta metro. Between the two elections, it grew by about 20%. We went from an 18 pt loss to an 18 pt win.
There’s a pretty obvious urban-rural divide. You can see on this map where the Atlanta metro is, as well as Athens, Columbus, Albany, Savannah, and the Warner-Robbins metro. In rural areas, the north trended away more than the south did, but we have a higher minority base in the south. We’re improving our relative performance in the growing areas. The bad news is that for several quickly growing red counties, a blue shift doesn’t overcome growth rates, and Carter’s raw vote margin either wasn’t much better or was actually worse than Barnes’s. The good new is that if we keep improving our margins, suddenly we’ll see massive returns as we approach parity. However if we don’t improve our margins in the Atlanta exburbs we’ll see our gains eaten up by them. Regardless, Atlanta’s direction is good for us. That, combined with better performances in mid-sized cities is a path to 50%+1 in the near future.
South Carolina
- 2014 result: Haley 55.9- Shaheen 41.4, 14.8 pt loss
- Campaign comparison is 2006: Sanford 55.1- Moore 44.8, 11.1 pt loss
- Reddest swing: Edgefield County, pop 27,000, 59% white, and lies on the middle of the Savannah River. We went from a 20 pt win to a 21 pt loss. Strom Thurmond was from Edgefield.
- Bluest swing: Richland County (containing Columbia), pop 407,000 47% white and is dead center of the state. We went from winning by 19 pts to 36 pts.
There’s less of an urban/rural shift at play here than a general movement of our coalition from the Upstate to the Lowcountry. That’s not earth-shattering, but it does speak to a more complex dynamic than just urban/rural. Some of it of it is that Upstate is attracting a lot of new people...white people, and not the liberal kind. Another part is the Lowcountry was always more cosmopolitan and educated than Upstate. Relatively, I mean. This is South Carolina, not South Chelsea. There’s no obvious positive trend here, but neither is there a negative one. I can’t help but wonder what elections would look like if we had more organization strength in the Tricities.
Kentucky
- 2015 result: Bevin 52.5- Conway 43.8, 8.7 pt loss
- Campaign comparison is 2003: Fletcher 55.0- Chandler 45.0, 10.1 pt loss
- Reddest swing: Knott County, pop 16,000, 98% white, and lies in the middle of the Appalachia. We went from a 25 pt win to a 17 pt loss. Knott County is entirely dry, and at 35%, the Appalachian county with the 7th highest poverty rate.
- Bluest swing: Nicholas County, pop 7,000, 98% white and is in the northeast of the state. We went from a 4 pt loss to a 21 pt win.
This one’s pretty simple. Urban areas swung towards us, the more culturally southern rural areas swung somewhat right, and Appalachia swung hard right. This map is so red it almost makes you forget we did better in 2015 than in 2003. On one hand losing Appalachia is bad for us, because we have no natural base left. On the other, it’s shrinking both absolutely and relative to the state, and many of our voters aren’t switching, just dropping off the map. If we can find a way to win back a portion of our coal country voters, while turning out better in the cities, that’s a winning coalition. But we should still worry about the west and east of the state, because we have a long way to go until we hit rock bottom.
Arkansas
- 2014 result: Hutchinson 55.4- Ross 41.5, 13.9 pt loss
- Campaign comparison is 2002: Huckabee 53.0- Fisher 47.0, 6.1 pt loss
- Reddest swing: Greene County, pop 44,000, 95% white, and is in the concave northeast corner. We went from a 16 pt win to a 18 pt loss
- Bluest swing: Philips County, pop 19,000, 35% white and lies on the Mississippi. Our winning margin went from 1 pt to 33 pts. Philips County has suffered heavy population losses, about 500 people a year for about 2 decades now. It also holds the unfortunate distinction of having the most recorded lynchings of any US county.
Democrats seem to have the opposite trend here. While we made some progress in the heavily yellow dog areas near the Mississippi, we barely gained in the cities and lost ground in the suburbs. The rural northern portion of the state especially turned on us. I tried, but I have nothing optimistic to say about this. We’re improving in places with a declining population, along with stagnant Pulaski County. It’s not just that red areas are growing faster than blue areas, it’s that we’re also doing worse in them as they grow. For our long term health in the state, we probably need to figure out how to do better in the northwest corner, but that’s a tall order. This part of the state started moving away from us under FDR and became stubbornly Republican in 70s.
Oklahoma
- 2014 result: Fallin 55.8- Dorman 41.0, 14.8 pt loss
- Campaign comparison is 1998: Keating 57.9- Boyd 40.9, 17.0 pt loss
- Reddest swing: Craig County, pop 15,000, 67% white, and is on the Kansas border, near Arkansas. We went from a 27 pt win to a 14 pt loss. About ¼ of this county is Native or part Native. We still maintain a 2:1 registration advantage in Craig County.
- Bluest swing: Oklahoma County, pop 783,000, 65% white and is in the center of the state, mostly know for OKC. We went from losing by 33 pt loss to a 6 pts. Our margin improved here by about 40,000 votes, and in 2016 Clinton did the best of any democratic presidential nominee since Humphrey or LBJ, depending on your metric.
This one’s fairly clear. We made headway in OKC/Norman, stayed about even in the rural western parts of the state, and took heavy losses in Little Dixie. I feel a lot better about this state than I thought I would. You could even say I feel OK about it. OKC is responsible for a lot of the state’s population growth, and the suburbs are moving our way too. Riding that shift, working with the substantial native population, and figuring out where our missing Tulsa voters went (raw votes dropped by about 10%, while population’s up by about 15%) is probably our key to victory here. We also have four real candidates for governor in 2018, which suggests I’m not the only one who sees a path to victory here.
Virginia
- 2013 result: McAuliffe 47.8 - Cuccinelli 45.2, 2.6 pt win
- Campaign comparison is 2001: Warner 52.1- Earley 47.0, 5.1 pt win
- Reddest swing: Buchanan County, pop 23,000, 96% white, on the point where the west of the state stops curving up. We went from a solid 32 pt win to a painful 38 pt loss. The population’s collapsing with the coal industry, and it recently earned a small national profile for giving Trump his biggest win in a competitive primary.
- Bluest swing: Norfolk City, pop 246,000 47% white and on the southern coast. Our win went from 14 to 46 points.
A couple states will split large cities from their counties. My native Maryland does it for Baltimore. Virginia does it for 38 goddamn cities. Why? I think it’s because they don’t want people to be able to understand the state visually. So keep that in mind. There’s a high chance this shows exactly what you expect. We did better in NOVA, Richmond, and Tidewater, and worse in the rural parts of the state, especially coal country. It is interesting that Charlottesville moved our way while Roanoke didn’t. The good news is we don’t have to get too caught up in the specifics here, as Tim Kaine, T-Mac, Obama, and Clinton have all showed that as long as you get the margins you need in NOVA/Richmond/Tidewater, you win. And not only that, but those areas are growing quickly, while the rest of the state is stagnating. Long term, Virginia looks great for us. I hesitated putting it on here because it’s not really southern anymore, but I decided to keep it because it was in 2001.
Texas
- 2014 result: Abbot 59.3- Davis 38.9, 21.4 pt loss
- Campaign comparison is 2002: Perry 57.8- Sanchez 40.0, 17.8 pt loss
- Reddest swing: Lamb County, pop 13,000, 74% white, stranded somewhere between Lubbock and Amarillo. We went from a decent 10 pt loss to a mind blowing 76 pt loss.
- Bluest swing: Travis County (containing Austin), pop 1,200,000 68% white and in south-central Texas. We went from a stalemate to a 30 pt win.
This is stark. DFW, Houston, San Antonio/Austin, College Station, and Victoria. We improved in these areas and collapsed everywhere else. This was the state that convinced me to add a fifth red color because there were much more severe shifts. What’s notable is that we did not do better in Hispanic border areas. This may be due to the races of our candidates, or maybe just our dismal turnout in 2014. There’s another weird feature in Texas. In all the other states, there was a weak positive correlation between an increase in votes cast in a county and increase in our share of them. In Arkansas the correlation’s negative, another reason for my pessimism. In Texas though? The correlation’s so strong that 63% of a county’s swing can be explained by turnout changes. This means 1 of 2 things. Either our voters stopped voting and we just need to bring them back, or Republicans made all their gains in (relatively) shrinking counties, and for the most part they’re maxed out there, winning 90+% of the vote in some places. Both options are promising.
Alabama
- 2010 result: Bentley 57.9- Sparks 42.1, 15.8 pt loss
- Campaign comparison is 1986: Hunt 56.4- Baxley 43.6, 12.8 pt loss
- Reddest swing: Cherokee County, pop 26,000, 86% white, in the northeast of the state, near Gadsen. Cherokee County is perhaps most famous for an F4 tornado on Palm Sunday 1994, which killed 20 worshipers while destroying their church. We went from a 20 point win to a 34 point loss.
- Bluest swing: Chambers County, pop 34,000, 59% white, in the middle of the Georgia border. Mississippi Burning was filmed there. We went from a 24 pt loss to a 2 pt win.
Start by taking note of the years here. The 2014 result was the biggest Republican winning margin in the state since 1868, when the military had essentially outlawed the Democratic party and William Hughes Smith won the governor’s race unopposed. Even 2010 only had two precedents: 2006, which I didn’t use, and 1986, when the liberal (for Alabama) democrat won a bitterly contested primary to succeed George Wallace and the conservative flank balked. This map actually doesn’t look dissimilar from an actual election map in Alabama. The white, more Appalachian north moved right, while enfranchisement in the black belt improved our margins in the south. We also did better in the cities. Jefferson, Montgomery, Madison, and Mobile counties all moved our way. Unlike in most states though, the suburbs declined to follow them. I think of Alabama as the most Republican state in the country, and this shows why. The north is done with us, the black belt is close to maxed out, and we’re only making progress in the cities, not the whole metros. There may yet be a way to win Alabama, but it would likely require getting those suburban voters who haven’t seemed to budge yet. Well that and a lot of luck.
North Carolina
- 2016 result: Cooper 49.0- McCrory 48.8, .2 pt win
- Campaign comparison is 2008: Perdue 50.3- McCrory 46.9, 3.4 pt win
- Reddest swing: Tyrell County, pop 4,000, 55% white, It’s coastal and near the outer banks. We went from a decent 40 pt win to a 17 pt loss. Tyrell county was the least populous county in the state, even before it lost 8% of its population in the last 5 years.
- Bluest swing: Meckelenburg County (containing Charlotte), pop 1,050,000, 55% white and in the center of the SC border. We went from winning by a couple hundred votes to to a 30 pt win. Mecklenburg, on the other hand, is the most populous county in the state.
North Carolina, unlike a lot of the south, does not have one predominating urban area. From west to east, the leftward swings are from Asheville, ASU, Charlotte, the research triangle, and Wilmington. In the rural areas, the coastal areas stand out for their rightward swing. This is probably from Bev Perdue’s appeal in the east being traded for the more business-friendly metropolitan Cooper, while Pat McCrory’s model transitioned from Charles Mathias to Mike Huckabee, along with McCrory’s good press from storm cleanup. These are not short term trends, however. There have been 3 close gubernatorial elections in North Carolina. The third occurred in 1972, the same year Nixon’s crushing performance in the state also inflicted upon us Jesse Helms. Comparing 1972 to 2016 gives you nearly the same map, just with higher intensity.
Louisiana
I did something different here, because every competitive gubernatorial election in Louisiana is very different. The 1991, 2003, and 2015 elections share some similarities. They were all held 12 years apart from each other. The previous governor was a Republican, but the runoff was between two non-incumbents. The Republicans messed up by nominating someone ill-suited to the Louisiana electorate, but for different reasons. A Klansman in 1991, a racial minority in 2003, and an adulterous diaper fetishist in 2015. We nominated an old-fashioned southern moderate and won by 22, 4, and 12 points, respectively.
I realize the key is hard to read. This is a map of which Democrat’s performance was the best. Green is John Bel Edwards, purple is Kathleen Blanco, and orange is Edwin Edwards.
JBE’s base was the Baton Rouge metro. Edwin Edwards did best on the bayou and in the urban areas around New Orleans and up north. What’s striking about this map is that Edwin Edwards and JBE both did much better overall than Blanco, but she beat them in all of the rural areas outside of the bayou. The most extreme example of this is La Salle Parish, in the middle of the state. La Salle Parish is ¾ rural, has about 15,000 people, 86% of them white, and Duke won it 67-33. Vitter took it by a punishing 71-29. But it backed Blanco 61-39.
With that in mind, where were the candidates’ weak areas?
Same colors as before, but for the worst performances. Blanco’s pattern is the most obvious. She couldn’t come close to the Edwardses in urban and suburban areas. She had the worst relative performance in the counties containing New Orleans, Baton Rouge, Shreveport, Monroe, and Lake Charles. Outside of the metros, JBE was stronger in the blacker areas. While this would seem weird considering who their opponents were, 1991 wasn’t too long after the civil rights era and black registration, organization, and turnout took a long time to build. You might suspect Blanco had a different coalition than the Edwardses. A simple parish level correlation analysis suggests that’s the case. I felt 24 years of population change and multiple natural disasters made a precinct analysis too sensitive. Edwards and Edwards managed a decent .69. However, Blanco’s r-value with JBE is .42 and only .14 with Edwin. That’s crazy-it means how a parish voted in 1991 had almost no bearing on how it voted in 2003. No matter what, Louisiana’s going to be tough for us, at least for the next couple decades. We should be at least encouraged by how JBE’s coalition looked. Blanco’s was very much a throwback to another time, and was entirely unsustainable. And JBE did well in the suburbs, which are major growth areas.
Odds and Ends
- I know that Kentucky was a Union state, Oklahoma didn’t exist then, and Texas is Texas. I still consider them to be predominately southern.
- Conversely I don’t consider Florida or Missouri southern enough politically. Also this diary is already too long.
- You may be noticing a distinct lack of Mississippi and Tennessee. Our fortunes have declined precipitously in Mississippi. Our last 3 gubernatorial elections all set the record for our worst we had done since Reconstruction. Despite that, we have a good chance there in 2019. Tennessee’s a fucking disaster zone though. We lost in 2014 by 48 points and 2010 by 32 points. There’s no point in comparing how badly Some Dudes got demolished.