Draw the Congressional Districts of the Alternate States!

[Fuzzybigfoot]

Made the fourth 52.3% black, to cover historical standards. 



New districts:

CD-4: 84%-15.1% Obama (Safe D)

CD-6: 49.6-49% McCain (Lean/likely R)


Those were the only districts I had to change, btw. 

[Fuzzybigfoot]

Election shades:

[muon2]

Election shades:

But wouldn't the Dems gerrymander Erie so that there was just one safe R seat, and all others lean or safe D?

[Fuzzybigfoot]

Election shades:

But wouldn't the Dems gerrymander Erie so that there was just one safe R seat, and all others lean or safe D?

Perhaps in this alternate reality, the districts are drawn by a bi-partisan panel instead of by the legislature.     

[Antonio the Sixth]

Election shades:

But wouldn't the Dems gerrymander Erie so that there was just one safe R seat, and all others lean or safe D?

Perhaps in this alternate reality, the districts are drawn by a bi-partisan panel instead of by the legislature.     

Actually, my goal for this project was to assume that everything was exactly like real life except for the States. So yes, a Dem gerrymander is probably more likely.

[Fuzzybigfoot]

Election shades:

But wouldn't the Dems gerrymander Erie so that there was just one safe R seat, and all others lean or safe D?

Perhaps in this alternate reality, the districts are drawn by a bi-partisan panel instead of by the legislature.     

Actually, my goal for this project was to assume that everything was exactly like real life except for the States. So yes, a Dem gerrymander is probably more likely.

Darn!  Back to the app! 

[muon2]

Election shades:

But wouldn't the Dems gerrymander Erie so that there was just one safe R seat, and all others lean or safe D?

Perhaps in this alternate reality, the districts are drawn by a bi-partisan panel instead of by the legislature.     

Actually, my goal for this project was to assume that everything was exactly like real life except for the States. So yes, a Dem gerrymander is probably more likely.

Darn!  Back to the app! 

Here's your challenge. Keep Cleveland intact and get 48.5% BVAP. Split no townships or cities and get 6 D districts all of which voted at least 56.0% for Obama in 2008. It can be done.

[Antonio the Sixth]

Sorry 'bout that Fuzzy, but I'm sure you don't actually mind working on DRA a little more, do you?

[Fuzzybigfoot]

I hope a 49.2% BVAP will do!



CD-1: 60.5%-37.9% Obama (Safe D)

CD-2: 53.5%-44.9% McCain (Safe R)

CD-3: 56.5%-42% Obama (Safe D)

CD-4: 83.9%-15.2% Obama (Safe D)

CD-5: 56.8%-41.7% Obama (Safe D)

CD-6: 56.5%-41.9% Obama (Safe D)

CD-7: 56.4%-41.6% Obama (Safe D)

[muon2]

I hope a 49.2% BVAP will do!



CD-1: 60.5%-37.9% Obama (Safe D)

CD-2: 53.5%-44.9% McCain (Safe R)

CD-3: 56.5%-42% Obama (Safe likely D)

CD-4: 83.9%-15.2% Obama (Safe D)

CD-5: 56.8%-41.7% Obama (Safe likely D)

CD-6: 56.5%-41.9% Obama (Safe likely D)

CD-7: 56.4%-41.6% Obama (Safe likely D)

Now you're thinking like an evil gerrymanderer!

Only CDs 1 and 4 are safe D. CDs 3,5,6, and 7 are all D+4 based on the 2008 election. Those would be normally classified as likely D, as I have edited above.

For the record, here was my gerrymander. I chose to make the Parma-Mentor CD slightly weaker, to a D+3, but I boosted the Akron and Youngstown CDs to D+5 and D+7.



ER-1: 59.2%-38.2% Pres'08; D+7; Uncompetitive D
ER-2: 56.9%-41.5% Pres'08; D+4; Competitive D
ER-3: (48.5% BVAP) 83.1%-16.0% Pres'08; D+30; Uncompetitive D
ER-4: 56.0%-42.6% Pres'08; D+3; Competitive D
ER-5: 43.5%-55.0% Pres'08; R+10; Uncompetitive R
ER-6: 57.4%-41.0% Pres'08; D+5; Competitive D
ER-7: 59.8%-38.0% Pres'08; D+7; Uncompetitive D

[Fuzzybigfoot]

I hope a 49.2% BVAP will do!



CD-1: 60.5%-37.9% Obama (Safe D)

CD-2: 53.5%-44.9% McCain (Safe R)

CD-3: 56.5%-42% Obama (Safe likely D)

CD-4: 83.9%-15.2% Obama (Safe D)

CD-5: 56.8%-41.7% Obama (Safe likely D)

CD-6: 56.5%-41.9% Obama (Safe likely D)

CD-7: 56.4%-41.6% Obama (Safe likely D)

Now you're thinking like an evil gerrymanderer!

Only CDs 1 and 4 are safe D. CDs 3,5,6, and 7 are all D+4 based on the 2008 election. Those would be normally classified as likely D, as I have edited above.

For the record, here was my gerrymander. I chose to make the Parma-Mentor CD slightly weaker, to a D+3, but I boosted the Akron and Youngstown CDs to D+5 and D+7.



ER-1: 59.2%-38.2% Pres'08; D+7; Uncompetitive D
ER-2: 56.9%-41.5% Pres'08; D+4; Competitive D
ER-3: (48.5% BVAP) 83.1%-16.0% Pres'08; D+30; Uncompetitive D
ER-4: 56.0%-42.6% Pres'08; D+3; Competitive D
ER-5: 43.5%-55.0% Pres'08; R+10; Uncompetitive R
ER-6: 57.4%-41.0% Pres'08; D+5; Competitive D
ER-7: 59.8%-38.0% Pres'08; D+7; Uncompetitive D

Very nice!  

[Antonio the Sixth]

Now that's a gerrymander! Great job to both!

Is anyone interested in making a Republican gerrymander in Adirondack? I'm really curious to see how many seats they could grab (my guess was 6 out of 9).

[jimrtex]

The legislature redistricted the House after the 1980 census, but like in 1971, the Texas Supreme Court said it did not comply with the Texas Constitution due to excessive cutting of counties.  There were 10 small counties with population less than one quota that were split.  Some, such as San Augustine, were quite small.  In addition, most counties with even small surpluses had quasi-floterial districts.  For example, El Paso was entitled to 5.059 representatives, and had 5 districts, plus a small portion in a district extending east to Odessa.  There were 8 such larger counties that could have had whole districts contained within their boundaries and be withing constitutional limits of deviation.

In addition, the court noted that 3 counties had their surplus divided among two districts.  Of particular concern was Nueces County, which was entitled to 2.827 districts, but with only one wholly in the county.  The voters in Nueces County probably were not too concerned since they would have totally dominated all three districts, but it was a clear violation of the Texas Constitution, that large counties be apportioned the whole number of representatives they were entitled to.

Rio Grande 1981 House Districts



County splits are symbolic, since I only had text of legislation available.  Note that Nueces county had one district entirely in the county, as well as one extending into Aransas, and another extending into Kleberg.

Clements v Valles was the Texas Supreme Court decision that overturned the legislature plan.  Bill Clements was the first Republican governor since reconstruction, and had the honor of having the lawsuit styled after him, even though the legislature was firmly in Democratic hands.  

The plaintiffs in the case had demonstrated that the deviation could be within the 10% limit, and reduce the number of counties split.  The court also rejected the State's argument that retaining surpluses within larger counties would discriminate against minorities, since it would tend to make districts within the county larger.  This of course only true of some counties.  A county with entitled to 4.8 representative would have a nominal surplus of 0.800.  But if apportioned 5 representatives, would average 96% of the ideal population.

After the Supreme Court decision, the Legislative Redistricting Board took over.  Comprised of the Attorney General, Comptroller,  Land Commissioner, House Speaker, and Lieutenant Governor, in 1981, it was firmly in Democratic hands.  Given direction by the Supreme Court, they fashioned a new map.  One advantage they may have had, is that they had less self-interest in preserving existing districts.  Splitting small counties is one way to minimize changes to districts.  Otherwise the requirement to compose districts of whole counties, usually requires wholesale rearrangement each decade.

The districts drawn by the LRB were challenged on a number of grounds, including being racially discriminatory, politically discriminatory, too much population deviation, and dividing communities, including Baytown and Montgomery County.

The Attorney General made his Section 5 submission to the USDOJ, and then the Secretary of State, appointed by the governor, made a separate submission, in which he suggested several problems with the LRB's plan.  The USDOJ later objected to the districts in most of the larger metropolitan areas, and to a Val Verde district.  This stalled the court proceedings in Terrazas v Clements, since the USDOJ objections meant that the LRB districts could not be used.  The federal court was left with considering whether to use the 1970s districts - by 1980, the largest had 4.5 times the population of the smallest, or not holding elections.   The USDOJ eventually, withdrew its objections, except with respect to Bexar, Dallas and El Paso counties.  The federal court then fashioned an interim plan meshing the LRB plan, with plans submitted by MALDEF for Bexar and El Paso counties.

This was an interim plan, to be used for the 1982 elections.  The court warned that if the State did not adopt a plan that was precleared by the USDOJ, that any plan they drew would likely require greater population equality than one drawn by the legislature.   This presumably would have meant wholesale splitting of counties, since the largest deviations are typically found where combinations of counties are forced as the only way to get within 5% (or 10% total deviation).  This might have run afoul of a modern SCOTUS which has directed federal courts to respect state laws and legislative priorities in redistricting cases to the extent possible.

The 1983 legislature adopted the LRB plan with the MALDEF changes for Bexar and El Paso.  The committee report noted that with the election of Mark White as governor, the Secretary of State had changed, and that the USDOJ had informally approved the Bexar and El Paso maps.

1982 Rio Grande House Districts



Several changes are notable.   The splits of El Paso (5.059 entitlement), Webb (1.040), and Bexar (10.424) were eliminated.  While the surpluses in El Paso and Webb were minuscule, that in Bexar was substantial.  Further, Bexar previously had 11 districts, so that the LRB change meant a loss of a district.  So even though not splitting Bexar resulted in an average 4.2% deviation, it was within the presumptive 5% (or 10%) OMOV limits.

Dallas County had a similar surplus (16.407).  In Dallas, 17 districts were apportioned.  This was primarily done so Dallas would only lose one representative from 18 to 17.  Increasing to the next whole number is legal if the population "substantially" warrants it.  In this case, the average Dallas district had a -3.5% deviation.

Even though Dallas had a nominally smaller surplus than Bexar, (0.407 vs 0.424), rounding Dallas up can be justified on apportionment grounds.  Dallas would have the greater right to the additional seat under a list method, regardless whether the arithmetic mean: (n+(n+1))/2, geometric mean: sqrt(n*(n+1)), or harmonic mean: 2*n*(n+1)/(n+(n+1)), is used.  The harmonic mean might be preferred for this application since the goal is to minimize population deviation.

Overall, the LRB eliminated county splits of 8 large counties where the surplus could be absorbed within the county, and all 10 small county splits.

There were three large counties that had their surplus distributed between two districts, the same as in the legislative plan.  But in all three instances under the LRB, the county had the correct number of districts wholly within the county.

Legislative Plan: Brazoria, Nueces, and Denton.
LRB Plan: Brazoria, Jefferson, and Collin.

Brazoria is adjacent to three large counties, and one moderate sized county.   With a fairly large surplus (1.788) it is difficult to find partners,   Most of its matched up with Fort Bend (1.379), but together there population of (3.167) was slightly too much for 3 districts, so the excess was combined with Matagorda.  Conceivably, Fort Bend could have been the double split county, but one or the other would be necessary.

Jefferson in the southeast corner had a large surplus (2.645) and only three neighbors.  In the legislature's plan, the use of split surpluses was avoided by dividing nearby counties of Liberty and San Augustine.

Nueces was double split under the legislature's plan and had only one district wholly in the county despite a population of 2.827.  Clearly the double split was not needed, since the LRB plan eliminated.

Denton (1.509) and Collin (1.529) could have been paired for a total population of 3.038.  In fact, the populations were similar enough that a conventional floterial district could have been used.  But neighboring Grayson had a population of 0.947, just outside the 5% limit (the overall deviation was 9.95% so it may not have been possible to use an asymmetric range that included Grayson.   Instead under the legislative plan, a small portion of Denton was added to Grayson; and under the LRB plan a small portion of Collin was added.  So one of the counties was double split so as to avoid a OMOV was split.  It also appears that this was done to avoid splitting a smaller county.

So it appear the rule is:

(1) Don't split larger county unless necessary to avoid a OMOV violation.  And always provide the whole number of representatives.
(2) Don't double split larger county unless necessary to avoid a OMOV violation, and (1) is insufficient.   And always provide whole number of representatives.
(3) Don't split smaller counties unless necessary to avoid a OMOV violation, and (1) and (2) are insuffiicient.

[Sol]

Would Adirondack really have a GOP leg? I could see the state senate, but I doubt they'd have the trifecta.

[Antonio the Sixth]

Would Adirondack really have a GOP leg? I could see the state senate, but I doubt they'd have the trifecta.

It would have been a bit close, but I'm pretty sure that, in a landslide year like 2010, republicans would be favored to win the State Assembly as well. IRL, this happened in States with a more Democratic PVI...

[muon2]

Would Adirondack really have a GOP leg? I could see the state senate, but I doubt they'd have the trifecta.

It would have been a bit close, but I'm pretty sure that, in a landslide year like 2010, republicans would be favored to win the State Assembly as well. IRL, this happened in States with a more Democratic PVI...

I also think they would have taken it in 2010. Perhaps I'll model their GA this weekend, like I did RG last weekend.

[Sol]

Would Adirondack really have a GOP leg? I could see the state senate, but I doubt they'd have the trifecta.

It would have been a bit close, but I'm pretty sure that, in a landslide year like 2010, republicans would be favored to win the State Assembly as well. IRL, this happened in States with a more Democratic PVI...

What about the governorship? I doubt Paladino would have won...

[Antonio the Sixth]

Would Adirondack really have a GOP leg? I could see the state senate, but I doubt they'd have the trifecta.

It would have been a bit close, but I'm pretty sure that, in a landslide year like 2010, republicans would be favored to win the State Assembly as well. IRL, this happened in States with a more Democratic PVI...

What about the governorship? I doubt Paladino would have won...

I've been making conjectures and discussing about AD local politics in the main thread on the What If board. To be honest, I have very little idea what the State's politics would look like and there are very few politicians I can name who could credibly occupy the State's major offices. That said, I think it's fairly reasonable to assume that a Republican would be swept in office in the 2010 GOP wave.

[jimrtex]

Here is the apportionment for a weighted 150-member Rio Grande legislature.



Counties with an entitlement of more than one were guaranteed their own representative.  The collective entitlement of these 17 counties was 131.279, which was rounded to 131.  Their base entitlement (the sum of the integer part of their entitlement) was 124.  The difference between 131 and 124, which is 7, was apportioned on the basis of a list method using the the harmonic mean.  The harmonic mean is favorable to smaller counties, and this case resulted in Bastrop (1.436) gaining its 2nd member, while Travis (19.826) was left with 19.  This is desirable, since a single member from Bastrop would have a weight 43.6% above parity, while two members are each only 21.8% below.  Meanwhile the 19 Travis representatives will only average 4.3% above parity.  And of course, each county's total weight will be precisely proportional to its population.

The remaining 19 representatives (150-121) were distributed among the 52 smaller counties.   I tried to apportion a single district to the largest of these counties (those closest to 1).  These counties may have in the past been above 1.0 and had their own district, but lost population share; or in the future may gain population share, and be required to have their own districts.  This provides stability.  There were 6 large-small counties (Burnet, Kerr, Val Verde, Medina, Atascosa, and Wilson).

The remaining 13 representatives were distributed among the 46 smallest counties, generally trying to apportion representation to larger regions (eg an area entitled to around 3 representatives would have 3 districts).

An alternative would be to maintain districts over time, letting their weight grow or shrink in response to a change in population share, and only making adjustments when necessary.  This would provide stability over longer periods.  In Texas, a district in a zero-growth area would decline by half in population share over 38 years.  So many districts could be retained for multiple redistricting cycles, since even slow growth would delay the need for adjustment.

In rural areas, it may be possible to simply combined districts every few decades, while in high growth suburban areas a district might be split, or perhaps 2 districts rearranged into three groups.

There is no structural reason to have 150 reprsentatives, even though the total weight is 150.  In such a system, Travis and El Paso could have been apportioned an additional representative, based on their share of the total population, rather than their share relative to other counties.  And the same is true for the districts in the smaller counties.  Districts could simply be permitted to drift across a range of values (2/3 to 4/3 of parity), with redistricting only occurring when a district is out of range.  Redistricting may be simpler at that time, when it may be possible to combine two districts or split a district, rather than trying to fashion 3 districts from parts of four, or 4 districts from parts of 3.

[Sol]

Would Adirondack really have a GOP leg? I could see the state senate, but I doubt they'd have the trifecta.

It would have been a bit close, but I'm pretty sure that, in a landslide year like 2010, republicans would be favored to win the State Assembly as well. IRL, this happened in States with a more Democratic PVI...

What about the governorship? I doubt Paladino would have won...

I've been making conjectures and discussing about AD local politics in the main thread on the What If board. To be honest, I have very little idea what the State's politics would look like and there are very few politicians I can name who could credibly occupy the State's major offices. That said, I think it's fairly reasonable to assume that a Republican would be swept in office in the 2010 GOP wave.

I think Carl Paladino would be the obvious choice for the ADGOP, and he'd have done as badly in AD as in NY.

[jimrtex]

Would Adirondack really have a GOP leg? I could see the state senate, but I doubt they'd have the trifecta.

Local politics would have developed differently.  If the state were Republican, the party would have an easier time recruiting town councilmen and county officials for the legislature.  A potential Democrat candidate would have to (1) pay the filing fee; (2) try to raise campaign funds, or paint the signs himself; (3) get elected; and (4) if elected, be assigned to the pencil-sharpening committee.  As a result, the Democratic candidate might be someone who has the money for the filing fee; and was planning to get a tattoo anyhow, and "Vote for Smith" with a couple of eagles looks cool, and knows a graffiti artist who will paint his 1983 Votemobile.   They may be willing to try again and again in two years.

[muon2]

To understand what AD might be like for the 2010 remap, I constructed a hypothetical Senate for the 2000s. I set the size to a typical state average 31 Senators. I assumed that the court ordered a neutral map in 2001 that minimized county and municipal splits. This plan used 2000 data and maintains a 10% range and only splits Buffalo and Amherst among munis.



Using the 2008 presidential results to estimate PVIs I find that 8 SDs have D+6 or higher, 6 SDs have D+0 to 2, 6 SDs have R+0 to 2, 6 SDs have R+3 to 5, and 5 SDs have R+6 or higher. A neutral year would by 17R-14D, but it's a very competitive map that has about a third of the seats able to swing from election to election.

NY elects its senators to two-year terms, so if AD kept the same arrangement, all would have been up for election in 2010. If the Pubs kept their seats and gained the competitive Dem seats, they would have had a commanding 23-8 margin for the 2011 redistricting session.

The state assembly would probably not have been as severe, but I can reasonably assume that there would be a comfortable margin for the Pubs there as well. It's safe to assume that they would hold the trifecta for the congressional remap.

[Antonio the Sixth]

Very nice!

Are you interested in drawing the R-gerrymandered 2010 map as well?

[muon2]

Very nice!

Are you interested in drawing the R-gerrymandered 2010 map as well?

Indeed. The trifecta certainly helps. A 6R-3D map is pretty easy, but I wanted to see if it could be done with no chops to cities or towns. Here's a gerrymander one could do with that restriction.



AD-1: D+10; O 62.6%, M 35.6%
AD-2: R+4; O 49.2%, M 49.1%
AD-3: R+3; O 50.0%, M 49.1%
AD-4: R+4; O 49.1%, M 49.2%
AD-5: R+2; O 50.7%, M 47.4%
AD-6: D+16; O 62.6%, M 30.3%
AD-7: R+7; O 46.1%, M 52.2%
AD-8: R+6; O 46.8%, M 51.8%
AD-9: D+11; O 63.7%, M 34.8%

Of course this shows that keeping municipal integrity alone is not much aid to stop gerrymandering.

[JerryArkansas]

Very nice!

Are you interested in drawing the R-gerrymandered 2010 map as well?

Indeed. The trifecta certainly helps. A 6R-3D map is pretty easy, but I wanted to see if it could be done with no chops to cities or towns. Here's a gerrymander one could do with that restriction.



AD-1: D+10; O 62.6%, M 35.6%
AD-2: R+4; O 49.2%, M 49.1%
AD-3: R+3; O 50.0%, M 49.1%
AD-4: R+4; O 49.1%, M 49.2%
AD-5: R+2; O 50.7%, M 47.4%
AD-6: D+16; O 62.6%, M 30.3%
AD-7: R+7; O 46.1%, M 52.2%
AD-8: R+6; O 46.8%, M 51.8%
AD-9: D+11; O 63.7%, M 34.8%

Of course this shows that keeping municipal integrity alone is not much aid to stop gerrymandering.

Muon, I like that and it works very well.  Every time I tried, I always tried to create a version of the NY 28, and it would always lead to a 5-4 map no matter what.