US House Redistricting: Georgia

[Bacon King]

I could easily see the Republicans doing something like this for the Congressional map:



1st: 56.1% R, 64.1% white
2nd: 60.5% D, 50.9% black
3rd: 64.6% R, 69.8% white
4th: 80.7% D, 51.7% black
5th: 76.3% D, 52.6% black
6th: 59.9% R, 66.7% white
7th: 57.9% R, 51.2% white
8th: 61.5% R, 65.1% white
9th: 71.0% R, 84.7% white
10th: 61.7% R, 73.7% white
11th: 65.1% R, 72.5% white
12th: 54.6% R, 58.8% white
13th: 71.0% D, 57.8% black
14th: 74.3% R, 81.7% white

Barrow's district loses a quarter of his black voters and they're replaced with the Augusta suburbs. Barrow's also cut out of his own district, so he'd have to move to Augusta or run in the 1st District. A Kingston-Barrow race would be slightly risky for the GOP, considering the district is made much more marginal with the inclusion of the black parts of Savannah, but it's still a R+10 district so Kingston should be fine (the new 12th district is R+10 as well).

Woodall's district might become more marginal later in the decade but he's apparently pushing hard to have a Gwinnett-only district; in this map the 7th is R + 17 so he's secure for a while.

Ultimately, a 10-4 map, with the four Democratic districts being VRA seats and all Republican districts being R +20 or more except Kingston's and Barrow's current districts, both R+10, but that's required to eliminate Barrow.

[Southern Senator North Carolina Yankee]

Shouldn't that last line say "both R+10, but" instead of "both R+20, but"?  

[Verily]

I think that's mostly realistic, but that GA-07 is risky. Not in 2012, but by 2018 or 2020... Gwinnett County is changing fast, and the Georgia Republicans know it. They'll split it up across districts.

[Miles]

Would your GA-02 be considered Racial Gerrymandering?

[Vazdul (Formerly Chairman of the Communist Party of Ontario)]

Would your GA-02 be considered Racial Gerrymandering?

I don't see how. It's more compact than most black-majority districts, and for the most part respects county boundaries. I don't see how they could uphold something like NC-12 and toss out this version of GA-2.

[krazen1211]

Would your GA-02 be considered Racial Gerrymandering?

I don't see how. It's more compact than most black-majority districts, and for the most part respects county boundaries. I don't see how they could uphold something like NC-12 and toss out this version of GA-2.

A GA-02 that splits Macon along racial lines like that (along with other counties) has already been rejected in Georgia district court.

[Bacon King]

Shouldn't that last line say "both R+10, but" instead of "both R+20, but"?  

Corrected

I think that's mostly realistic, but that GA-07 is risky. Not in 2012, but by 2018 or 2020... Gwinnett County is changing fast, and the Georgia Republicans know it. They'll split it up across districts.

Strategically speaking they should do it, of course, but I'm not sure at this point if they will. Georgia's redistricting session meets next week so we'll have a better idea at what the GOP wants to do then.

Would your GA-02 be considered Racial Gerrymandering?

Well, not nearly bad enough to be an unlawful gerrymander.

Would your GA-02 be considered Racial Gerrymandering?

I don't see how. It's more compact than most black-majority districts, and for the most part respects county boundaries. I don't see how they could uphold something like NC-12 and toss out this version of GA-2.

A GA-02 that splits Macon along racial lines like that (along with other counties) has already been rejected in Georgia district court.

Curious here. What case?

[Bacon King]

Oh, wait, you're talking about Miller v. Johnson, doh.

I think this is much worse than my 2nd district:



Also the map-makers could just claim it's incumbency protection, which is perfectly permissible.

[Verily]

More importantly, Miller v. Johnson was decided at a time when the preferred black candidate could genuinely win a 40% black district in that area, so drawing a 55% black seat and a 30% black seat denied blacks the ability to elect their preferred candidates in two different seats. Not any more.

[BigSkyBob]

More importantly, Miller v. Johnson was decided at a time when the preferred black candidate could genuinely win a 40% black district in that area, so drawing a 55% black seat and a 30% black seat denied blacks the ability to elect their preferred candidates in two different seats. Not any more.

Which 40% Black seats in the deep South have a Black representative again?

When the VRA was passed, it was assumed the Democrat nominee would win just about every seat in the deep South. The VRA was passed to stop the "cracking" of Blacks so that a Black nominee could win.

[Bacon King]

Which 40% Black seats in the deep South have a Black representative again?

Sanford Bishop won the 2nd District after the post- Miller v. Johnson 1995 redistricting; at the time it was 60% white.

[Verily]

Which 40% Black seats in the deep South have a Black representative again?

Sanford Bishop won the 2nd District after the post- Miller v. Johnson 1995 redistricting; at the time it was 60% white.

Also, preferred black candidate =/= black person.

[krazen1211]

Oh, wait, you're talking about Miller v. Johnson, doh.

I think this is much worse than my 2nd district:



Also the map-makers could just claim it's incumbency protection, which is perfectly permissible.

That it certainly is. Whether it is in the GOP's best interest to create such a district for the perpetually 'barely winning' Bishop is another question.

[Bacon King]

A safe district for Bishop is definitely in the best interest of the GOP; they need a Democratic vote sink down there because without one Austin Scott and/or Westmoreland would be at risk.

[Bacon King]

Here's my best effort for a 11-3 plan...



1st: R+9, 64% white
2nd: R+8, 58% white
3rd: R+13, 64% white
4th: D+20, 50% black
5th: D+26, 59% black
6th: R+13, 65% white
7th: R+16, 61% white
8th: R+8, 59% white
9th: R+22, 80% white
10th: R+13, 71% white
11th: R+16, 69% white
12th: R+7, 61% white
13th: D+16, 55% black
14th: R+26, 83% white
(edit: PVI's are approximations because I dun goof'd; see next post)
Dubious legality, I think, under the VRA.

Bishop and Barrow both get cut out of their districts; the new 2nd and 12th favor the GOP but could still go Democratic in a bad year. The 8th district would see a Bishop vs. Austin Scott fight where the latter would be favored, but still under significant risk. The Kingston vs. Barrow matchup happens like in my other map.

The 3rd, 10th, and possibly even the 6th could all be at risk in a bad year later in the decade: Westmoreland would hate Macon being placed into his district, and would also have to deal with Coweta and Fayette trending to the left later in the decade; Broun would hate getting a second college town put into his district (GCSU in Milledgeville) and would face unfriendly trends in Henry, Newton, and Spaulding. Tom Price is a bit more insulated from demographic change thanks to the higher property values in north Fulton (Alpharetta/Roswell/etc.), but I've always considered it inevitable that Buckhead will eventually start voting Democratic like similar neighborhoods on the East Side of Atlanta already do.

So, yeah. A 11R-3D map that would possibly be a stable but could just as easily have 10D-4R delegation later in the decade.

[Verily]

You're overstating the PVIs in your calculation. PVI isn't Democratic minus Republican (or the reverse); it's the difference between one party's vote and the most recent result. The absolute maximum Democratic PVI, using 2008 as a base, is D+47; the absolute maximum Republican PVI is R+53.

On the assumption that 2008 was 53D-47R, a 60D-40R seat is D+7, not D+20 or D+14. A 60R-40D seat is R+13, not R+20 or R+26.

[Bacon King]

You're overstating the PVIs in your calculation. PVI isn't Democratic minus Republican (or the reverse); it's the difference between one party's vote and the most recent result. The absolute maximum Democratic PVI, using 2008 as a base, is D+47; the absolute maximum Republican PVI is R+53.

On the assumption that 2008 was 53D-47R, a 60D-40R seat is D+7, not D+20 or D+14. A 60R-40D seat is R+13, not R+20 or R+26.

Haha, wow, thanks Verily. All these years I've been entirely misunderstood what PVI meant. All I can think to say is, "derp."

[muon2]

You're overstating the PVIs in your calculation. PVI isn't Democratic minus Republican (or the reverse); it's the difference between one party's vote and the most recent result. The absolute maximum Democratic PVI, using 2008 as a base, is D+47; the absolute maximum Republican PVI is R+53.

On the assumption that 2008 was 53D-47R, a 60D-40R seat is D+7, not D+20 or D+14. A 60R-40D seat is R+13, not R+20 or R+26.

Haha, wow, thanks Verily. All these years I've been entirely misunderstood what PVI meant. All I can think to say is, "derp."

More accurately the PVI (or Cook PVI after its originator) is found by taking the average of the last two presidential elections considering only the two party vote. The PVI then measures the difference between the district and the nation as a whole.

Obama took 53.7% of the two party vote nationally, and Bush took 51.2% in 2004. The net is 51.3% for the Dems for the two cycles. A district's current PVI comes by comparing the district average to a baseline of 51.3% Dem.

[krazen1211]

Here's my best effort for a 11-3 plan...



1st: R+9, 64% white
2nd: R+8, 58% white
3rd: R+13, 64% white
4th: D+20, 50% black
5th: D+26, 59% black
6th: R+13, 65% white
7th: R+16, 61% white
8th: R+8, 59% white
9th: R+22, 80% white
10th: R+13, 71% white
11th: R+16, 69% white
12th: R+7, 61% white
13th: D+16, 55% black
14th: R+26, 83% white
(edit: PVI's are approximations because I dun goof'd; see next post)
Dubious legality, I think, under the VRA.

Bishop and Barrow both get cut out of their districts; the new 2nd and 12th favor the GOP but could still go Democratic in a bad year. The 8th district would see a Bishop vs. Austin Scott fight where the latter would be favored, but still under significant risk. The Kingston vs. Barrow matchup happens like in my other map.

The 3rd, 10th, and possibly even the 6th could all be at risk in a bad year later in the decade: Westmoreland would hate Macon being placed into his district, and would also have to deal with Coweta and Fayette trending to the left later in the decade; Broun would hate getting a second college town put into his district (GCSU in Milledgeville) and would face unfriendly trends in Henry, Newton, and Spaulding. Tom Price is a bit more insulated from demographic change thanks to the higher property values in north Fulton (Alpharetta/Roswell/etc.), but I've always considered it inevitable that Buckhead will eventually start voting Democratic like similar neighborhoods on the East Side of Atlanta already do.

So, yeah. A 11R-3D map that would possibly be a stable but could just as easily have 10D-4R delegation later in the decade.

I would consider sacrificing compactness, in favor of, say, swapping Hall and Clarke Counties.

[Bacon King]

I would consider sacrificing compactness, in favor of, say, swapping Hall and Clarke Counties.

Not realistic. The state leadership won't do anything that could marginalize its Hall County CD pet project, especially not for Broun of all people. Besides, such a swap would be unforgivably ugly unless you cut into the northern part of that 12th district, and doing so would essentially be granting it to Barrow.

[krazen1211]

http://savannahnow.com/news/2011-08-10/kingston-okay-district-including-all-chatham-if-it-keeps-its-military-bases

Kingston is willing to take Chatham and Effingham.

[Brittain33]

http://savannahnow.com/news/2011-08-10/kingston-okay-district-including-all-chatham-if-it-keeps-its-military-bases

Kingston is willing to take Chatham and Effingham.

Next challenge is for someone to draw a map that does that and keeps all four bases in. How hard can that be, if they're willing to make the lines wonky? It's not like Austin Scott's district is going to become that much more competitive.

[dpmapper]

http://savannahnow.com/news/2011-08-10/kingston-okay-district-including-all-chatham-if-it-keeps-its-military-bases

Kingston is willing to take Chatham and Effingham.

Next challenge is for someone to draw a map that does that and keeps all four bases in. How hard can that be, if they're willing to make the lines wonky? It's not like Austin Scott's district is going to become that much more competitive.

It's really quite easy, and looks reasonable too.  You just have to avoid going too far inland until you reach the Florida border.  Statesboro, Waycross, Jesup - all of them should be in other districts.  Result is about a 56% McCain district. 

[Bacon King]

Done very easily. I kept Waycross in the district to minimize county splits and because it looked uglier without it. The district's 54.9% McCain, 44.4% Obama. Cook PVI: R+7.

[lowtech redneck]

Done very easily. I kept Waycross in the district to minimize county splits and because it looked uglier without it. The district's 54.9% McCain, 44.4% Obama. Cook PVI: R+7.

What's the population and Republican voting percentage of the portion of Lowndes County you included?  Kingston would probably prefer a district with a higher probability of remaing safe in 10 years in the event that Savannah takes off economically and attracts more out-of-state migrants, so I was wondering whether the rural counties further inland would be a better fit (assuming the Republicans can make up the difference elsewhere to take down Barrow).

Thanks, btw; I just got into amateur redistricting (I still have a lot more to learn, not to mention precint-level data to acquire), and your maps have been very helpful.