Race-Based House Districts

[Antonio the Sixth]

Hi again! So since I felt a little bad about my 50,000th post project being kind of a redux of an old hit, I decided I needed something a little different to complement it. I will try to keep this project updated regularly in conjunction with the 100 States one. It's possible that either one falls by the wayside at some point if I get busy, but hopefully the hiatus won't be too long either way.

However, because I suck at coming up with original ideas, this time instead of building on an earlier project of mine, I will instead be taking over someone else's. Specifically, I'm talking about jimrtex's "New Zealand Style Apportionment" concept. I'll assume he doesn't mind, given that that thread has been dormant for 8 years. Besides, I intend to take this project in a different direction by focusing it on what election outcomes would look like under such a scenario (hence the placement on this board rather than the demographics one).

So, without further ado, let's get started.


Apportionment

First, I want to redo the apportionment process. While the big picture is the same as in jimrtex's project, I followed a slightly different procedure than he did and so got slightly different results. First, I included DC in the House (but not Puerto Rico) and thus raised its size to 436. Like jimrtex, I count Hispanics as a single category regardless of race. But unlike him, I don't take care of splitting up the "multiracial" category, and instead just consider it as its own block. It is a little awkward, since a person who's Black and Asian presumably won't find much in common with a White and Native American, but I still feel that it's the only solution that doesn't devolve into splitting hairs about who gets lumped into what ethnic box.

Given these criteria, this is the apportionment we get. I use the Harmonic Mean rule to apportion the extra seats (because it's one of the apportionment rules most favorable to smaller groups, and ethnic minorities might rightfully see it as a guarantee).

Group	Population	Seats (Th)	Seats (Min)	Quota	Seats (Tot)
White	196,817,552	277.94	277	709,255	278
Black	37,685,848	53.22	53	704,470	53
Hispanic	50,477,594	71.28	71	706,015	71
Asian	14,465,124	20.43	20	706,036	21
AIAN	2,247,098	3.17	3	655,404	3
NHPI	481,576	0.68	0	Inf	1
Other Races	604,265	0.85	0	Inf	1
Multiracial	5,966,481	8.43	8	704,376	8

The next step is apportioning each group's seats across states, of course. I already have most of the calculations ready, but I'm going to cut it short for today since it's pretty late. Like jimrtex, I will be combining states that fall significantly below the quota in order to avoid excessive population disparities across districts. For White people, the apportionment turns out the same as in jimrtex's versions, but for other races I made different choices in terms of combining states (for example, prioritizing contiguity over exact population equality). I'll show my work in the coming days, and then of course I'll get into drawing the actual districts.

Again, hope you enjoy, and comments/questions/criticism always welcome.

[Antonio the Sixth]

All right, I know my OP had very little stuff in it, but I swear I'm doing my best to keep both of my new projects afloat. I hope there's some degree of interest in this. I will get to the cool maps soon enough, I swear. In the meantime, here's a little grunt work I have to get through.


Apportionment (continued)

Let's next apportion each group's seats across states. Obviously, due to the reduced number of seats to apportion, keeping the minimum 1-seat-per-state requirement is unreasonable. For Whites, given their numbers, I still largely abide by it, and only make exception for Hawaii (merged with Alaska) and DC (merged with Maryland), which are less than half the quota. That leaves the least populous seat, Wyoming, at around 2/3rd of the quota. For minority groups, I was a lot more cavalier about merging states, since there's really no point in preserving state lines when there's about as many (or less) seats as there are states. I tried to keep all seats within 10% of the quota, and put a hard cap at 20%. Here's what it looks like.

White: 278
California: 21
Texas: 16
New York: 16
Florida: 15
Pennsylvania: 14
Ohio: 13
Illinois: 12
Michigan: 11
North Carolina: 9
Georgia: 8
Indiana: 7
New Jersey: 7
Virginia: 7
Massachusetts: 7
Washington: 7
Missouri: 7
Tennessee: 7
Wisconsin: 7
Minnesota: 6
Kentucky: 5
Arizona: 5
Colorado: 5
Alabama: 5
Oregon: 4
South Carolina: 4
Louisiana: 4
Iowa: 4
Oklahoma: 4
Connecticut: 4
Kansas: 3
Utah: 3
Arkansas: 3
West Virginia: 2
Mississippi: 2
Nebraska: 2
Nevada: 2
Idaho: 2
Maine: 2
New Hampshire: 2
Montana: 1
New Mexico: 1
Rhode Island: 1
South Dakota: 1
North Dakota: 1
Vermont: 1
Delaware: 1
Wyoming: 1
--------------------
MD+DC: 5
AK+HI: 1


Black: 53
Georgia: 4
Texas: 4
Florida: 4
New York: 4
California: 3
North Carolina: 3
Virginia: 2
Louisiana: 2
Ohio: 2
Michigan: 2
Pennsylvania: 2
South Carolina: 2
Missouri: 1
--------------------
WV+KY+IN+IL: 4
AL+MS: 3
MD+DC: 3
AR+TN: 2
NJ+DE: 2
ME+NH+VT+MA+RI+CT: 1
WA+OR+NV+AZ+AK+HI: 1
OK+KS+NE+CO+NM+UT: 1
WI+MN+IA+ND+SD+MT+WY+ID: 1


Hispanic: 71
California: 20
Texas: 13
Florida: 6
New York: 5
Illinois: 3
New Jersey: 2
North Carolina: 1
Washington: 1
Pennsylvania: 1
Nevada: 1
--------------------
NM+AZ: 4
SC+GA+AL+TN: 2
CO+UT: 2
DE+MD+VA+WV+KY: 2
ME+NH+VT+MA+CT+RI: 2
ID+OR+HI+AK: 1
MI+WI: 1
OH+IN: 1
IA+MN+ND+SD+NE+MT+WY: 1
MS+LA+AR+MO:1
OK+KS: 1


Asian: 21
California: 7
New York: 2
New Jersey: 1
--------------------
TN+AL+MS+LA+AR+TX+OK+KS: 2
GA+FL: 1
VA+WV+NC+SC: 1
PA+MD+DE+DC: 1
MN+IA+ND+SD+NE+MT+WY+CO+UT+NM+AZ: 1
OH+IN+MI+WI+KY: 1
OR+AK+HI: 1
WA+ID+NV: 1
IL+MO: 1
ME+NH+VT+MA+RI+CT: 1


AIAN: 3
AK+HI+OR+CA+NV+AZ+NM: 1
MN+ND+SD+NE+KS+OK+MT+WY+ID+CO+UT+OR+WA: 1
Everything else: 1


NHPI: 1


Other Races: 1


Multiracial: 8
HI+CA+NV+AZ+NM+UT: 2
AK+WA+OR+ID+MT+WY+CO+ND+SD+NE+KS: 1
ME+VT+NH+MA+RI+CT+NY+PA: 1
OH+IN+IL+MI+WI: 1
GA+FL+AL+MS+TN+KY+LA: 1
TX+OK+AR+MO+KS: 1
NJ+DE+MD+DC+VA+WV+NC+SC: 1


Next update, I'll start drawing the White districts and discussing how they might vote. Thanks for bearing with me.

[P. Clodius Pulcher did nothing wrong]

Pretty cool

[Antonio the Sixth]

All right, time to get into the heart of things.


White Districts

Here, I'll be going over White redistricting state by state, posting maps of the 2012-2020 districts as I think they would have been drawn by a mostly nonpartisan actor (given that a lot of the drive toward gerrymandering has historically been based on racial motivations or pretexts, it's not implausible to imagine that separating seats a priori based on race would lead to fairer maps).

I'll do my best to guesstimate how these districts would have voted throughout the decade, in both House and Presidential elections. I'll be basing these guesses largely on Adam Griffin's excellent project on White vote for Obama in 2012. I can only wish that something of the same kind existed for 2016, but I'm not aware of it, so I will have to draw up rough estimates of the trends. It goes without saying that the margin of error of all this will be enormous, so I highly welcome any feedback and alternative takes on the politics of these districts.

To start simple, today I'll do all the one-seat States, and then move up the list until I get to California (oh boy...).


Alaska and Hawaii

According to Adam's map, Hawaii Whites are strongly Democratic (though not as much as nonwhites), while Alaska Whites are strongly Republican. AK makes up about 57.5% of the district's White CVAP, so this works out slightly to Republicans' advantage. This is a district that Republicans are favored in, but that can flip for Democrats in the right year. That being said, using 2012 as a baseline might overstate Democratic strength due to Obama's strong favorite son effect in Hawaii. It is possible that Clinton lost the district by more than twice Obama's margin. In addition, Alaska Whites probably turn out at a higher rate than Hawaii Whites. Still, in light of Young's underperformance, I'm going to go out on a limb and say this seat flipped in 2018.

2008 Presidential: 51% McCain, 46% Obama
2012 Presidential: 50% Romney, 46% Obama
2016 Presidential: 49% T***p, 41% Clinton

2013 Representative: Don Young
2015 Representative: Don Young
2017 Representative: Don Young
2019 Representative: Ed Case


Wyoming

Nothing to see here, of course. A Titanium-R seat will become slightly more Titanium. If Democrats are winning this one, the GOP might just have collapsed as a party.

2008 Presidential: 67% McCain, 30% Obama
2012 Presidential: 71% Romney, 25% Obama
2016 Presidential: 71% T***p, 19% Clinton

2013 Representative: Cynthia Lummis
2015 Representative: Cynthia Lummis
2017 Representative: Liz Cheney
2019 Representative: Liz Cheney


Delaware

This one is a bit trickier. Adam's map suggests that Romney narrowly carried the state, suggesting that the district would be a GOP-leaning battleground. On the other hand, the fact that the ticket including the ultimate favorite son Joe Biden still lost it might mean that it's really more Republican than it looks. Still, Obama did almost certainly carry it in 2008, and it's certainly not unwinnable for the party. 2018 would be close, but ultimately, the Republican trend over the past 8 years makes me doubtful that Democrats could take this seats even in such a good year for them. I have no idea who would actually represent the seat though, since who knows who the Republicans would run.

2008 Presidential: 52% Obama, 47% McCain
2012 Presidential: 50% Romney, 48% Obama
2016 Presidential: 52% T***p, 42% Clinton

2013 Representative: <R>
2015 Representative: <R>
2017 Representative: <R>
2019 Representative: <R>


Vermont

There's really not much to say here. Vermont's electorate is almost entirely White to begin with, so making it only White will change very little to its politics. Still Titanium-D State, still safely in Pete Welch's hands.

2008 Presidential: 67% Obama, 31% McCain
2012 Presidential: 66% Obama, 31% Romney
2016 Presidential: 56% Clinton, 31% T***p

2013 Representative: Peter Welch
2015 Representative: Peter Welch
2017 Representative: Peter Welch
2019 Representative: Peter Welch


North Dakota

Not much would change here. Taking away the Native American vote (primarily) makes the State about 7 points more Republicans. This might have made it harder for Democratic candidates to hold the seat in the mid and late 2000s, back when they actually won it IRL, but no Democrats have come close to winning it this decade. All this does is make a solidly Republican seat slightly more solid.

2008 Presidential: 57% McCain, 41% Obama
2012 Presidential: 62% Romney, 35% Obama
2016 Presidential: 66% T***p, 24% Clinton

2013 Representative: Kevin Cramer
2015 Representative: Kevin Cramer
2017 Representative: Kevin Cramer
2019 Representative: Kelly Armstrong


South Dakota

See above. Actually, Democrats lose slightly more in SD than in ND (probably because the former has slightly more Native Americans). As a result, the two districts are about equally Republicans, rather than SD being somewhat more Dem-friendly. Still, nothing would change here.

2008 Presidential: 58% McCain, 40% Obama
2012 Presidential: 62% Romney, 35% Obama
2016 Presidential: 66% T***p, 27% Clinton

2013 Representative: Kristi Noem
2015 Representative: Kristi Noem
2017 Representative: Kristi Noem
2019 Representative: Dusty Johnson


Rhode Island

Rhode Island has only one seat in the White-only apportionment, meaning that only one of the two White Democrats who currently represent it would be able to hang on. Since Langevin is far more senior than Cicilline, my best guess is that he would get the spot. In terms of partisanship, RI would be a tad more Republican, but not enough to seriously worry Democrats there. Even Clinton comfortably won the White vote there.

2008 Presidential: 60% Obama, 37% McCain
2012 Presidential: 60% Obama, 38% Romney
2016 Presidential: 52% Clinton, 41% T***p

2013 Representative: James Langevin
2015 Representative: James Langevin
2017 Representative: James Langevin
2019 Representative: James Langevin


New Mexico

NM Whites get only one seat, compared to the three seats the State as a whole (which is still plurality-White in its CVAP) gets IRL. This is a significant reduction in Whites' voting power. In partisan terms however, it ultimately works out in the GOP's benefit. If we trust Adam's calculations, New Mexico Whites are more Republican than any of the State's current CDs, even the 2nd. Even in his 2008 landslide in the State, Obama still lost the White vote by 10 points. So, we can guess that Pearce would easily have been succeeded by Herrell with little fuss here.

2008 Presidential: 54% McCain, 44% Obama
2012 Presidential: 56% Romney, 39% Obama
2016 Presidential: 53% T***p, 34% Clinton

2013 Representative: Steve Pearce
2015 Representative: Steve Pearce
2017 Representative: Steve Pearce
2019 Representative: Yvette Herrell


Montana

This is largely the same story as the Dakotas, with Democrats bleeding a few points as a result of losing the Native American vote. Here however, this potentially makes more of a difference given Montana's occasional flirtations with Democrats. For example, Obama came within 2.5 points of beating McCain, but among Whites only, he comes 10 points short. Similarly, Greg Gianforte would probably never have to seriously worry about his prospects to either win the seat in 2017 or hold it in 2018.

2008 Presidential: 53% McCain, 43% Obama
2012 Presidential: 59% Romney, 38% Obama
2016 Presidential: 59% T***p, 32% Clinton

2013 Representative: Steve Daines
2015 Representative: Ryan Zinke
2017 Representative: Ryan Zinke
2019 Representative: Greg Gianforte

[America Needs a 13-6 Progressive SCOTUS]

Simply changing the numbers in the URL gives you the page you are probably looking for

[Antonio the Sixth]

Simply changing the numbers in the URL gives you the page you are probably looking for

Oh! I didn't know about this, fascinating!

I note that it was made by reagente, whom I vaguely recall was working on this project a while ago, but didn't know he'd finished. I don't see him around much anymore. I'd have liked to ask him about his methodology and whether/how it differs from Adam's. I'm going to take his estimates into account but also compare them to my extrapolations from Adam's analysis and overall trends. Hopefully those estimates more or less fit together.

[Antonio the Sixth]

OK, so I calculated the "swing" in the White vote by state, according to Griffin and Reagente's estimate (this requires a bit of guesswork on the Griffin end, since he only has Obama's vote share, but I just assumed that the Other vote share was the same as IRL). This is the map we get:

Compare with the overall swing map:


The broad, general pattern is similar (of course), but something just isn't right. First of all, too many states swung to T***p overall despite Whites swinging to Clinton. I know that there was some degree of demobilization among Blacks and Hispanics compared to 2012, but that can't explain the full extent of this discrepancy. Overall, the map above would imply that T***p's margin of victory among Whites was identical or near-identical to Romney's. And it's hard to see enough mobilization for Whites in NC and NM to swing Democratic by over 5 points, while general electorate swings Republican. Conversely, if we accept the mobilization thesis, it's hard to understand then why California Whites barely swung while the State as a whole clearly did?

It seems clear to me that the two estimates have significant methodological differences. Obviously I don't know enough to be able to tell which methodology makes more sense, though I'd love to see Griff and Reagente debate it. So I'm going to be drawing from both in my guesstimates here, but not necessarily commit to either. We'll see what comes out.

[Antonio the Sixth]

New Hampshire



NH-1: McCain+1, Romney+6, T***p+7. A surprisingly Republican district, and would be so even if nonwhites could vote in it. Really goes to show that the current NH map is really one of those dumb "competitiveness" gerrymanders. Since it's very suburban, however, this is the kind of area where Democrats can expect to overperform. I'm guessing the GOP held the seat until 2018, when it narrowly flipped Democratic. I guess Pappas represents it now, but he faces a tough reelection.

NH-2: Obama+16, Obama+10, Clinton+2. Conversely, this used to be a very Democratic district, although a lot less so recently, because the more rural Western counties have swung massively to T***p in 2016. Still, its PVI is comparable to the real NH-2, which hasn't elected a Republican in this round of redistricting, so I'm assuming the seat stayed Democratic all through, presumably with Ann Kuster.

[bagelman]

Have you considered using DRA 2020? Might save you a headache or two in the larger states. Great project and I'm excited to see more.

[Antonio the Sixth]

Have you considered using DRA 2020? Might save you a headache or two in the larger states. Great project and I'm excited to see more.

I've tried to use DRA 2020, but it wouldn't let me make an account (kept saying that the email I was using was "already in use", even when I tried different emails). If you have any tips for how to sign up, I'm all ears.

[Antonio the Sixth]

Maine



ME-1: Obama+21, Obama+19, Clinton+13. This is basically the RL ME-1 with borders slightly adjusted to better respect county lines and compactness. Its PVI is pretty much the same if you include the full population. Excluding nonwhites can only shave one or two points points off the Democratic margin. Chellie Pingree has been holding this seat for a while and isn't about to lose it

ME-2: Obama+10, Obama+7, T***p+11. Conversely, this is basically the real ME-2. There's even less difference here, since this is one of the Whitest districts in the country. Nevertheless, the 2018 election was close enough that even a minor shift could have cost Golden his seat. I can say for sure that Michaud occupied this seat until 2015, and Poliquin from 2015 to 2019, but I can't say for sure who holds it now. My guess is Golden wins by a recount-worthy margin and Poliquin throws an even bigger fit and possibly succeeds in delegitimizing RCV. Yeesh.

[Antonio the Sixth]

Idaho



ID-1: McCain+30, Romney+34, T***p+34. The more "competitive" of the two ID seats since it's centered around Boise, not that that's saying much. This would be the seat held by Fulcher and then Labrador, interestingly a pretty different one from the one they occupy IRL.

ID-2: McCain+35, Romney+44, T***p+42. Obviously titanium-R, although this seat probably displays some interesting cultural diversity, with the heavily Mormon Southeast and the more culturally populist Northern panhandle. Either way, Mike Simpson can probably have this seat if he wants it.

[Antonio the Sixth]

Nevada


Used a Clark Co. insert here, for obvious reasons.

NV-1: McCain+1, Romney+6, T***p+10. Even this very urban seat, concentrated around Las Vegas, would vote Republican in most elections. While the heavy Hispanic (and to a lesser extent Black and Asian) presence makes the area solidly Democratic, its White electorate is solidly Republican, though not overwhelmingly so. Obama came very close to winning the district (and maybe did so, since obviously there's a margin of error to my estimates). Still, that's a steep hill for Democrats to climb, and I doubt they have or will at any point this decade. Sadly for humanity, Danny Tarkanian can probably have this seat. He'd face a close reelection in 2018 but my guess is he'd narrowly survive.

NV-2: McCain+20, Romney+26, T***p+29. Democrats don't even come close here, of course. I guess Mark Amodei has held it ever since he first ran.

[Kingpoleon]

Just something you might want to consider: there are 3-4 million Arab Americans who probably don’t like those numbers so much.

[Antonio the Sixth]

Just something you might want to consider: there are 3-4 million Arab Americans who probably don’t like those numbers so much.

Well, yeah, the Census' definition of "White" is pretty ridiculous at this point, and if I was in charge of it I'd definitely add a "Middle Eastern" category to make things clearer. Unfortunately, I have no choice but to work with the census categories I have, so Arabs get to vote in the White seats. They'd probably still be in a decent position to elect one of them, since Wayne County has its own Congressional District and Dearborn probably weighs heavily within it.

[Antonio the Sixth]

Nebraksa



NV-1: McCain+18, Romney+26, T***p+28. The most "competitive" of the two districts, but don't expect much of it. Even Douglas County Whites voted for Romney and T***p by comfortable margins, and this district extends much farther into the Nebraskan hinterland than its RL counterpart. Terry would never have lost reelection here, and so would probably still be occupying his seat.

NV-2: McCain+30, Romney+36, T***p+39. Titanium R, of course. Smith and Fortenberry both live in this districts (albeit on opposite ends of it), so they would have to battle it out in a primary. My understanding is that Fortenberry is a bit more of a heavyweight, and so would have the advantage, but Smith might have the advantage of hailing from a more rural part of the district, which would probably play well in this electorate. I'd be happy to get input from anyone with knowledge of NE politics.

[Sol]

I know you don't go over there, but Lewis Trondheim did a similar (but not the same, so no worries) project over on AAD that you might find interesting.

[Antonio the Sixth]

I know you don't go over there, but Lewis Trondheim did a similar (but not the same, so no worries) project over on AAD that you might find interesting.

Oh! I'm definitely interested. Can you give me a link (by PM if the mods don't want direct links posted)?

[Antonio the Sixth]

Mississippi



MS-1: McCain+75, Romney+77, T***p+77. You get the picture: this is Adamantium-R territory. Republicans have achieved such a one-party status among MS Whites that Democrats wouldn't even bother running. There would be quite a bit of competition for this seat, since the recent representatives for IRL districts 3 and 4 would both be in this district. I'm not sure who would prevail, but I guess Steven Palazzo would have an edge from representing more of his original constituency

MS-2: McCain+74, Romney+75, T***p+77. Basically everything I said of MS-1 also applies here. Except that Nunnelee and then Kelly would have an easier time holding this district with little competition.

[America Needs a 13-6 Progressive SCOTUS]

Tbh, I think I would find this project more interesting if you also upped the size of the US House in it by a substantial level, like you are doing in your 100 states project.

[Kamala's side hoe]

Just something you might want to consider: there are 3-4 million Arab Americans who probably don’t like those numbers so much.

I don't think Ralph Nader and Justin Amash would care that much. Brad Avakian and Cyrus Habib might wonder why they're being left out of a broader "MENA" category.

[Antonio the Sixth]

Tbh, I think I would find this project more interesting if you also upped the size of the US House in it by a substantial level, like you are doing in your 100 states project.

Yeah, sorry but I don't have it in me right now. It actually took me a year or so to draw just the 278 White districts, and I'm still not done with districts for other races, so as much as I like the idea of expanding the House, I haven't and probably won't be able to find time for it. Note that I'm not actually drawing the districts for my 100-states project, all I'm doing is a rough guess.

[Antonio the Sixth]

Well, overall size of the house aside, let's proceed to our final two-seat state before moving to those with increasingly more complex district maps. And this one is a #populist classic!


West Virginia



WV-1: McCain+17, Romney+33, T***p+47. After our little tour through Adamantium-R Mississippi, this almost looks like a competitive district by comparison. Still, that massive WV trend that we've all been talking to death about for a decade is very real, and we can see it here. Downballot Democrats would obviously do a lot better than Hillary did, but "a lot better than Hillary did" is still not enough to win anymore. In 2012, though, it would have been enough for Rahall to hand on (the district he represented IRL has a similar partisanship. But even if he didn't retire to run for Senate, I could still see him going down in the very unfavorable year (perhaps even to Shelley Moore Capito, who wouldn't have had a seat in the House until then since she represented the Charleston area IRL).If not in 2014, then in 2016 this seat would safely be a Republican pickup.

WV-2: McCain+20, Romney+31, T***p+46. This is the more historically Republican part of the state, although this difference has all but disappeared with the uniformizing Republican trend of the past decades. Still, this means that the seat has probably been Republican since 2010 if not earlier. David McKinley is the natural choice to represent it.

[Sol]

I know you don't go over there, but Lewis Trondheim did a similar (but not the same, so no worries) project over on AAD that you might find interesting.

Oh! I'm definitely interested. Can you give me a link (by PM if the mods don't want direct links posted)?

Here you go! link

[Antonio the Sixth]

I know you don't go over there, but Lewis Trondheim did a similar (but not the same, so no worries) project over on AAD that you might find interesting.

Oh! I'm definitely interested. Can you give me a link (by PM if the mods don't want direct links posted)?

Here you go! link

Oh, I see! So he apportioned 435 House seats to each racial group regardless of size. That is excellent to reveal more fine-grained demographic pattern, and hopefully it fulfills Solid's desire for a larger House. On the other hand, actually applying this principle to our government would result in gross malapportionment (unless you introduce things like weighted voting). So yeah, thankfully Lewis and I had very different goals in mind with these project, so I'm not unwittingly plagiarizing him.

But yeah, fascinating stuff! Thanks for the link.