Redistricting the United States

Republicans have a built-in advantage in the Senate. In the last congress, when the Senate was split 51-49, Republican Senators represented only 367 congressional districts between them, whilst the 51 Senators caucusing with the Democrats represented 503 congressional districts between them.

This is, to a large extent, a function of the way America’s states were formed. They began as a hodge-podge of different territories of varying histories and sizes and more were added rather haphazardly according to matters of population, prestige and transient electoral politics.

Meanwhile the Senate, a legacy of the time when states’ rights was more than a convenient label for segregationists to hide behind, gave two Senators to each state whether they had fifty representatives in the House or five. Ten states possess a majority of America’s electoral votes, yet they are outnumbered in the Senate by representatives of the eleven states with only three or four electoral votes.

And since a low population in a state tends to correlate with rurality and since the rural vote is generally extremely conservative, it is no surprise that there is a Republican skew to the Senate as a general rule.

For all the importance of redistricting, Democrats could gain much more if they could redistrict across state lines. They could give themselves a massive advantage in presidential races and a good chance of veto-proof Senate majorities. They could create a Permanent Democratic Majority that would last much longer than Karl Rove’s sorry attempt.

So let’s give this a go. A few suggestions have been made in the DE-AL comments thread, but I want to take this much further. Let’s dream of what we could do in an ideal world, and while we’re at it get some practice in before the next census is upon us

Before I begin, I’ll give the ground rules under which I’m operating:

1) All new states should be contiguous unless they include Alaska or Hawaii. Alaska and Hawaii must be attached to states with Pacific borders. Yes, I know Sarah Palin would enjoy being Governor of Oklahoma, but let’s be realistic with this exercise that bears no relation to reality.

2) All new states should have a roughly equal population. I’m lazy, so I’ve just made 15 states of 8 congressional districts each (in fast-growing areas of America) and 35 of 9 congressional districts each.

3) All new states should have some rationale to their existence beyond gerrymandering. In other words, don’t get too ambitious. Yes, you could make a new state stretching from Idaho to New Orleans, but there would be nothing real holding the state together. Stick to areas of common culture, religion, or geographical location.

With that aside, here’s a map of 50 new American states:

And here are their descriptions:

Olympia: This state is essentially Washington west of the Cascades, plus Alaska and northern Oregon. As Washington is likely to pick up a district in 2012, I’ve given it only eight CDs.

It contains AK-AL, OR-1, WA-1, WA-2, WA-3, WA-6, WA-7, WA-8 and WA-9. By my calculation, it’s 2004 PVI would have been D+5.5. That may not be entirely accurate, as throughout I’ve just added up the PVIs of the various CDs and averaged them.

Cascadia: One of four vote sinks in sagebrush country, Cascadia is made up of Idaho, eastern Washington, eastern Idaho and north-eastern California. I’ve given this eight CDs as well, which may have been a mistake given that most of this state is uninhabitable desert, but I had the idea that the Mormon population and global warming might keep this region growing at an above-average rate.

It contains CA-2, CA-4, ID-1, ID-2, OR-2, WA-4, WA-5 and WA-8. Estimated PVI is R+11.4.



Mormonistan: I don’t think this is one of my better states. Originally it included WY-AL, but due to me accidentally making San Francisco a 10-CD state, I had to give that up and take in CA-19 instead. This makes this district a little too Californian (it goes right up to the edge of L.A., because CA-25 is a huge district) and not really Mormon enough (although it is still centred on Utah and Cow Country). Also, I should probably have put Las Vegas elsewhere, but perhaps the rapid growth there will make this a winnable state in a couple of decades.

The state contains CA-19, CA-25, NV-1, NV-2, NV-3, UT-1, UT-2 and UT-3. Estimated PVI R+10.

San Jose:  This state takes in the east coast of California between L.A. and S.F, plus the Republican districts immediately inland of there. It was by far the easiest of the California replacements to make.

It contains CA-17, CA-20, CA-21, CA-22, CA-23, CA-24, CA-27, CA-28 and CA-30. Estimated PVI D+6.1.

Arizona: Nothing to see here. It’s a state with eight congressional districts and fast growth, so I decided just to leave it as is and call it a day. PVI should be around R+1.25.

Red Mexico: This is a combination of Colorado and New Mexico minus CO-04. I picked this name because I didn’t like the sound of New Colorado and I figured that this would be the place where Tom Tancredo would least want to live.

Estimate PVI is R+1.1, but I think that would be significantly changed if we were using 2008 PVIs.

Rio Grande: The Texas districts were a bastard to deal with, because almost all of them have either insanely Republican PVIs or very swingish numbers, despite being in 60% Hispanic districts. This is partly because of low Hispanic turnout, partly because of the Bush effect and his high numbers with Hispanics in 2004, and partly because the Delaymander was very clever. All four of the states I made from Texas have Republican PVIs, but two of them are winnable and swinging our way and it’s probably worth leaving the other two as vote sinks to help us speed that.

Rio Grande is made up south Texas districts and contains just about every Democratic district not held by Chet Edwards or a black representative. The districts in question are: TX-15, TX-16, TX-20, TX-21, TX-23, TX-25, TX-27 and TX-28. Estimated PVI is R+0.1.

North Texas: This area is a disaster zone for Democrats. Given Obama’s appalling performance here, it seems likely that one of TX-11, TX-13 and TX-19 will have the worst PVI in the country once the 2008 numbers come out. Bush is probably still popular here.

The district contains the the three districts already mentioned, plus TX-3, TX-12, TX-24, TX-26 and TX-32. It’s PVI is a stunning R+18.

East Texas: This area isn’t quite as bad as North Texas, but it’s a close-run thing. As a vote sink, this is more or less ideal. My only regret is that I couldn’t put Eddie Bernice Johnson’s Dallas district to good use.

The new state contains TX-1, TX-4, TX-5, TX-6, TX-8, TX-17, TX-30 and TX-31. Estimated PVI is R+11.6.

Louissippi: This is a state made up of Louisiana and the two southern districts of Mississippi, MS-3 and MS-4. I figure since Louisiana’s going to wrong way and we’ve been having no luck there, it’s best just to cut out the bits of territory we can use elsewhere and leave the rest well alone. Thanks to LA-02, it’s only R+6.2, but I wouldn’t put money on Mary Landrieu winning re-election here.

Kanslahoma: The third of the sagebrush vote sinks, this new state is made from all of Oklahoma, Kansas minus Dennis Moore’s district and CO-04. Sadly, I think it rather destroy any hope of Betsey Markey or Kathleen Sebelius moving up to the Senate, since it’s an imposing R+12.7.

Northern Plains: This puts a whole lot of empty space on the map where lines used to be. This vast state comprises Montana, Wyoming, Nebraska, the Dakotas, MN-7 and IA-5, just because I don’t think Iowans deserve Steve King.

The PVI is R+11.9, although Obama’s performance here may have improved that somewhat, and we hold enough Senators in the area already that we might still stand a chance here.

Minnisconsin: The last of the states which I had the room to label on the map, this is just an amalgam of Minnesota and northern Wisconsin. In an earlier form it also included the Upper Peninsula, because I figure it fits in better with Wisconsin that Minnesota, but my aforementioned mistake with San Francisco torpedoed that idea.

The districts present here are MN-1, MN-2, MN-3, MN-4, MN-5, MN-6, MN-8, WI-7 and WI-8. PVI is D+2.9.

Pacifica (1): This state stretches from Portland, down the Oregon coast to the northern outskirts of San Francisco, before turning inland to head towards the Nevada border and allow us to redistrict out Dan Lungren in 2012.

The districts here are CA-1, CA-3, CA-5, CA-6, CA-7, CA-10, OR-3, OR-4 and OR-5. The PVI is a very healthy D+9.5.

San Francisco (2): I tried not to allow this district to become too Democratic, but I failed (and, as previously mentioned, also tried to give this new state ten CDs). There’s definitely room for an attempt to be made to absorb more of wingnut country, who will be horrified by having two Senators with San Francisco values. Made up of CA-8, CA-9, CA-11, CA-12, CA-13, CA-14, CA-15, CA-16 and CA-18, it works out at D+18.8.

West L.A. (3): L.A. is large enough that it quite obviously has to be two states. The only question was how to accomplish this. If one were willing to cheat horribly, you could make some appalling spaghetti slices to match inner-city districts with Republican strongholds in the exurbs.

But personally, I couldn’t be bothered. I tried to absorb the less Democratic districts here, but pretty comprehensively failed, leaving West L.A. with a PVI of D+21. The districts used were CA-29, CA-31, CA-32, CA-33, CA-34, CA-35, CA-36, CA-37 and CA-46.

East L.A (4): The name is perhaps misleading here, since this district extends right out to the southern border of Orange County, encompassing CA-26, CA-38, CA-39, CA-40, CA-42, CA-43, CA-44, CA-47 and CA-48. Still, it works out as a comfortable D+1.7 and Hispanic growth and Obama’s strong Californian performance suggest that that Democrats will actually have more of an advantage here.

San Diego (5): Making a Democratic stronghold out of San Diego is almost impossible. So I cheated a little, rather than extending the district out to Phoenix, I merely joined it to Hawaii. This gives you a PVI of exactly 0, but Obama’s performance in Hawaii will have moved that in the right direction. The exact San Diego area districts used were CA-41, CA-45, CA-49, CA-50, CA-51, CA-52 and CA-53.

Houston (6): Made up of the Texas districts not used thus far (2, 7, 9, 10, 14, 18, 22 and 29,) this state gives us a PVI of R+2.3. Such a margin is not overwhelming, however, and I’d suspect that it would begin to lean strongly Democratic by 2016.

Arkasouri (7): A combination of the four Arkansas districts, the St. Louis districts of MO-1, MO-2 and MO-3 and the south Missouri districts of MO-7 and MO-8, Arkasouri clocks in at R+1.1. Disappointingly, this is also one of the few relatively marginal states that is probably trending in the wrong direction. I didn’t manage to create a strong progressive state based around Memphis and St. Louis. I’m open to suggestions as to how this could be done.

Missouwa (8): Made up of KS-03, the Missouri districts not in Arkasouri and all the Iowa districts bar IA-05, Missouwa is R+0.2. I can only argue that the Iowa caucuses will probably keep up turnout enough there that they’ll be able to balance out the rural Missouri parts of the state.

Illisconsin (9): I have to admit that this is a state that looks strange on a map, and would probably look strange even if I hadn’t drawn it. It might in fact have made more sense for down-state Illinois to go with Missouri and Iowa to go with the Quad Cities and western Wisconsin. Then again, I like the portmanteau.

The districts used were IL-12, IL-14, IL-15, IL-16, IL-17, IL-18, IL-19, WI-2 and WI-3. The PVI is R+0.2.

Wiscago (10): Beginning in the Madison area, this state takes a narrow path along the coast to draw in the North Side of Chicago and the northern suburbs. In between, it takes in enough of exurbia to wind up as a comfortable but not super-safe D+3.9. The constituent districts are IL-5, IL-6, IL-8, IL-9, IL-10, WI-1, WI-4, WI-5 and WI-6.

Undetroit (11): Originally this state was made up of all Michigan except for the north and the UP, Detroit and the Oakland county congressional districts. But with MI-1 back in the district, I had to reduce the south-east border by giving up on MI-8.

Other districts in this form of it are MI-2, MI-3, MI-4, MI-5, MI-6, MI-8, MI-10 and MI-12. At R+0.6, it has a very mild Republican lean.

Greater Chicago (12): The centre and South Side of Chicago, combined with some suburbs to the south and south-west and the Gary area make this new state. Joe Donnelly’s district was thrown in to make up the numbers and lower the PVI, but it’s an urban enough state that the PVI is still D+16. The consistuent districts are IL-1, IL-2, IL-3, IL-4, IL-7, IL-11, IL-13, IN-1 and IN-2.

Toledo Wedge (13): Harking back to an 1830s border dispute, this state is made up of the bluest bits of Michigan and the most insanely red bits of north-east Indiana and north-west Ohio. In an earlier and rather better iteration, it included Marcy Kaptur’s district, but that’s now been replaced with MI-8.

The other constituent parts are IN-3, MI-7, MI-11, MI-13, MI-14, MI-15, OH-4 and OH-5. Detroit counts for enough to give this a PVI of D+3.7.

Ohio River (14): Mostly filling gaps in my map, this state leaves relatively slim pickings for Democrats. It’s centred on Indiana, but replaces the Gary area with Louisville and the north-east with western Ohio and some Cincinnati suburbs.

It’s R+7.3. Obama’s performance in Indiana plus Democratic successes in the non-Indiana parts of the state might suggest better times ahead, but this will always lean Republican. The districts enclosed within are IN-4, IN-5, IN-6, IN-7, IN-8, IN-9, KY-3, OH-1 and OH-8.

Mississippi River (15): An attempt to make a Democratic district extending along the Mississippi, this state comes in at R+1.9 because Bennie Thompson’s district and the Memphis area can’t quite balance out western Kentucky, central Tennessee and north-eastern Mississippi. To be marked down as Democratic, it would probably have to extend up towards St. Louis instead.

The constituent parts are KY-1, KY-2, MS-1, MS-2, TN-5, TN-6, TN-7, TN-8 and TN-9.

Greater Alabama (16): This is quite a simple district, made up of all Alabama plus FL-01 and FL-02. It’s also a bit of a nightmare for Democrats, but considering that only 9% of Alabama whites voted for Obama, I think we knew that anyway. The PVI is R+9.

North Florida (17): Taking in Georgia’s coastline before heading down to Orlando and  west to the Gulf of Mexico, this is basically a cracker district, but with a black majority district unfortunately trapped within. The exact constituent parts are FL-3, FL-4, FL-5, FL-6, FL-7, FL-9, FL-12 and GA-1, which combine for a PVI of R+4.9. Like the other Florida-based states, it has eight congressional districts.

Central Florida (18): If this state appears messy, I blame the Florida 2002 redistricting plan. For the most part it’s a relatively simple formation, made up of the south-east portions of the Orlando metropolitan area and the land heading roughly south-west of there to Tampa, but for gerrymandering purposes it also includes FL-23, which has an absurd shape, being comprised of a chunk of land inside and a couple of squiggles into downtown Miami. Some sort of landswap with the state of Miami would certainly seem sensible here. Nevertheless, Alcee Hasting’s district is Democratic enough to turn this from a swing state to a Democratic bastion at D+3.1. Other districts in the state are FL-8, FL-10, FL-11, FL-13, FL-15, FL-16, FL-23 and FL-24.

Miami (19): Made up of those bits of Florida not already used, this state is mostly a narrow coastal strip on the Atlantic, with the numbers being made up by the Keys, the very southern tip of the Mainland and Fort Myers. It clocks in at a fairly comfortable D+6.8.

Atlanta (20): This is my attempt to produce a Democratic state from Georgia based on 2004 results. It’s centred on metro Atlanta and the Black Belt, but for population reasons it also has to include the districts of Lynn Westmoreland and Tom Price.

Right now it’s D+2.5, but I fully expect that Obama’s performance added a couple of points here. The districts within it are GA-2, GA-3, GA-4, GA-5, GA-6, GA-8, GA-12 and GA-13.

Sumter (21): Yes, I picked that name purely to annoy. This fast-growing new state is made up of all South Carolina, plus a western eextension made up of GA-7 and GA-10 that stops just outside Atlanta. It’s not going to be easy picking for Democrats in the short-term, as it’s R+9.25.

Charlotte (22): Containing pretty much all of North Carolina bar the Triangle and the Appalachian regions, Charlotte clocks in at R+4.1. 2008 results notwithstanding, I suspect you could get a better map if you were prepared to make some ugly looking states. Districts used are NC-1, NC-2, NC-3, NC-6, NC-7, NC-8, NC-9 and NC-12.

Southern Appalachia (23): Another vote sink, this new state is made up of northern Georgia, western North Carolina and eastern and middle Tennessee. I’d expect this to be a state with little internal cohesion but a lot of cultural cohesion. I don’t think you could expect to elect a single progressive here.

The PVI is R+12.5 and the districts are GA-9, GA-11, NC-5, NC-10, NC-11, TN-1, TN-2, TN-3 and TN-4.

Northern Appalachia (24): The slightly less red companion to its southern neighbour, this is actually a state where Democrats could still compete at lower levels. Containing KY-4, KY-5, KY-6, OH-6, VA-6, VA-9, WV-1, WV-2 and WV-3, it isn’t going to produce great progressives, but you could get some decent economic populists and union supporters from here. PVI is R+6.2.

Real Virginia (25): It’s not really what Nancy Pfotenhauer meant, but I like the name. This version takes in North Carolina’s Triangle, all of southern Virginia bar the very western portions and goes up into Communist Country at its northern extnet. And all for a PVI of R+1.3, which is probably now reversed. The specific districts here are NC-4, NC-13, VA-1, VA-2, VA-3, VA-4, VA-5, VA-7 and VA-11.

Baltimore (26): Despite the name, this district is a sprawl. It begins in the DC suburbs (but does not include DC, as I’m leaving that out of this restructuring), curves round the Chesapeake Bay, moves on to the Eastern Shore and continues up through Delaware into southern New Jersey. I’ve just noticed that it’s not actually contiguous, since Elijah Cummings’ district extends to the shoreline, but since I’m too lazy to fix that we’ll just pretend that that district has given away a two foot wide stretch of shoreline to this new state.

So this State of Baltimore (without the city of Baltimore, much like Kansas) has a PVI of D+5.3. It takes in DE-AL, MD-1, MD-2, MD-3, MD-5, NJ-1, NJ-2, VA-8 and VA-10.

Allegany (27): This is a quite blatant gerrymander, stretching from the Pittsburgh metro area down through rural Maryland into the DC suburbs and Baltimore. I imagine there’d be a degree of cultural clash between the more liberal Maryland portions of the district and the culturally conservative Appalachian regions, but there’s a decent chance you could elect a black Senator here and the D+7.6 PVI is not to be sniffed at. The districts used are MD-4, MD-6, MD-7, MD-8, PA-4, PA-9, PA-12, PA-14 and PA-18.

Columbus (28): This state is essentially made up of the bits of Ohio I couldn’t use elsewhere. Still, for a rump it’s a lot better than it could have been and its PVI is only R+1.2. Extending from Cincinnati through the central portion of the state, it once reached the Cleveland outskirts but I had to change it to take in Marcy Kaptur’s Erie-side district and it now gets no closer than the Mahoning valley.

The component parts are OH-2, OH-3, OH-7, OH-9, OH-12, OH-15, OH-16, OH-17 and OH-18.

Lakeside (29): Based around the coastlines of Lakes Erie and Ontario, this state stretches from Cleveland to Rochester. It also extends out wildly to enclose PA-5. In my defence, it looked a lot less weird when I also had NY-29 in here, but that had to change to help clear up my mistake with San Francisco.

The PVI is D+5.8 and other unmentioned districts here are NY-26, NY-27, NY-28, OH-10, OH-11, OH-13, OH-14 and PA-3.

Delaware River (30): This district is based around Upstate but also extends down through the Poconos into Philadelphia’s outer suburbs and northern New Jersey. It’s mildly Democratic, with a PVI of D+2.1, although that’s only because I cheated a little and extended it down into the north-western edge of New York City.

The current seats are NJ-5, NY-17, NY-22, NY-24, NY-25, NY-29, PA-10, PA-11 and PA-15.

Philadelphia (31): This was a comparatively simple district to draw. Luckily, Philadelphia’s centre and blueing suburbs easily swamp the portions of the T here and I extended the district into New Jersey just because I needed somewhere to put Adler’s district. If you wanted to respect state lines where possible, you could swap it for Patrick Murphy’s district for no difference whatsoever.

The PVI is D+6.9 and the districts are NJ-3, PA-1, PA-2, PA-6, PA-7, PA-13, PA-16, PA-17 and PA-19.

North Jersey (32): At this point, I’d like to apologise for the map. If I could draw, I would have made a better one. It’s not very clear here, so let me just explain that this is basically just central and northern New Jersey, plus Patrick Murphy’s Pennsylvania district for geographical compactness.

The PVI is D+8.2 and the districts used are NJ-4, NJ-6, NJ-7, NJ-8, NJ-9, NJ-10, NJ-11, NJ-12 and PA-8.

Manhattan (33): One of two NYC districts, this one has a whopping D+28.3 PVI. Unfortunately, urban areas generally tend to clump together and produce unnecessarily large Democratic PVIs and this trend is exacerbated because of NYC’s geographic location. On the plus side, I reckon that a minority candidate would be the favourite to win a Senate seat in a state like this.

This state is made up of Manhattan, Staten Island, most of the Bronx, parts of Queens and Brooklyn and Albio Sires’ New Jersey district. That means NJ-13, NY-7, NY-8, NJ-12, NJ-13, NJ-14, NJ-15, NJ-16.

Long Island (34): This area isn’t quite as Democratic as Manhattan. Still, it does have a suitable impressive D+19.2 PVI, so I wouldn’t worry about losing it. It’s basically made up of areas east of Manhattan State, curving around Jamaica Bay, by which I mean NY-1, NY-2, NY-3, NY-4, NY-5, NY-6, NY-9, NY-10 and NY-11.

Greater Connecticut (35): Not much to say here. This is just Connecticut plus New York down to Yonkers and the congressional districts in Massachusetts based around Worcester. PVI is D+8.5 and constituent parts are CT-1, CT-2, CT-3, CT-4, CT-5, MA-2, MA-3, NY-18 and NY-19.

Cape Cod (36): Again, there’s little to see here. This is just Rhode Island stuck together with eastern Massachusetts. PVI of D+16.2, districts are MA-4, MA-5, MA-6, MA-7, MA-8, MA-9, MA-10, RI-1 and RI-2.

Algonquia (37): I’ll close this out with a state that’s essentially charged with watching the border for a Canadian invasion. Made up of all Maine and New Hampshire, Vermont and New York down to Albany (NY-20, NY-21 and NY-23) it’s D+4.8.

If we count anything below R or D+2 as a swing state and anything above R or D+5 as a safe state, then we end up with the following statistics:

Democratic safe: 176 EV (17 states + DC)

Democratic lean: 75 EV (7 states)

Swing (Democratic edge): 11 EV (1 state)

Absolutely balanced: 11 EV (1 state)

Swing (Republican edge): 108 EV (10 states)

Republican lean: 30 EV (3 states)

Republican safe: 127 EV (12 states)

I’ll be the first to admit that Republicans could win on this map. But it wouldn’t be easy. Democrats have more than half of their total locked up and since it’s hard to flip swing states, they essentially start at 251 EV to Republicans’ 157.

Then add in the fact that Republicans have a lot more marginal states to defend and that some of those states are trending blue pretty fast, and we begin to see a map where a 50:50 race leans distinctly Democratic.

Meanwhile, in the Senate there are at least 30 seats where there is no reason for Democrats to nominate a minority and a good half a dozen states where minority candidates could easily win election.

I think this would be a better map than the one we have now. But if you can improve it further, or suggest a new map entirely, post it in the comments.