Texas House redistricting.

[jimrtex]

This is a first look at Texas House districting.



The numbers for each county represent 1/1000 of the quota of 194,303. Ideally, the population of a multi-county district will add to 1000, with a range of 950 to 1050 permitted. The Texas Constitution requires the apportionment among the counties be "as near as may be" which suggests that the target is 1000, and not 1050 to 950. Any variation must be justified.

The color coded counties are larger counties, that must be apportioned one or more representatives (an exception is Johnson which is near to one). My lower threshold was 0.900.

The red counties may have one or more whole districts within the county and it is possible for each district to have a population equivalent to 0.950 to 1.050. An exception is Travis, which may have 7 districts with a population of 0.9486 (94.86% of the ideal). The actual test promulgated by the SCOTUS is a range of 10%. If we can keep the largest district below 1.0486 (104.86% of the ideal) then the 7 Travis districts will be accepted - though they will have to have quite identical populations.

The counties in yellow will have to have one or more districts wholly within them and another district with a portion of the county combined with other whole counties or similar parts of other counties in a single member district.

The Texas Constitution provides for apportionment to combinations of whole counties. If County A was entitled to 2.4 representatives and County B was entitled to 1.6 representatives, then County A would be apportioned 2 representatives, County B would be entitled to 1 representative, and County A and County B be collectively would be entitled to 1 representative.

This seems fair, until you realize that County A would have 60% of the votes for the floterial district, while only providing 40% of the surplus population. This violates equal protection.

The Texas Supreme Court has harmonized the Texas Constitution with permitting two districts being drawn within County A, one district being drawn wholly within County B, and one district combining areas within the two counties representing the surplus population of the two counties, so 60% from County B combined with 40% from County A.

Equal protection (Voting Rights) require drawing of single member districts. The Texas Supreme Court has ruled that this is just a manner regulation, and is consistent with the Texas Constitution.

So ideally the state would be divided into multiple-county districts made up of smaller counties electing a single representative OR single counties divided into one or more single-member districts. But this is not possible and have the range of deviations be less than 10%.

So instead we create what might might be considered quasi-floterial districts. These are drawn in multiple counties with two or more representatives, with exactly one district crossing county lines.

In our earlier example, County A and B would be apportioned 4 representatives, with two districts entirely in County A, one entirely in County B, and one spanning the border between the counties.

Because this type of district is not entirely consistent with the Texas Constitution we should minimize their use and combine counties with surpluses where feasible.

Only as a last resort should a smaller county be divided (there has been one exception every decade since the 1970s).

Large Counties (in Red):

Harris will have 24 districts, the same as now with an average deviation of +1.5%.

Dallas will have 13 districts, a loss of one, with an average deviation of +3.5%.

Tarrant will have 11 districts, the same as now with an average deviation of -1.2%.

Bexar will have 10 districts, the same as now with an average deviation of +3.4%.

Travis will have 7 districts, one more than now with an average deviation of -5.1%. The slightly excessive deviation may be justified by the fact that most of the larger counties have a positive deviation. Collectively the 9 counties would be entitled to 73.591 representatives and will be apportioned 73 for an average deviation of 0.8%.

Were Travis removed from the group (its 7th district would be partially in another county), the average deviation for the other 8 counties would grow to 1.4% which is a bit largish when applied to almost 1/2 the state population.

Williamson will have 3 districts. Currently its 3rd district extends into other counties. The average deviation will be +4.5%.

Brazoria will have 2 districts. Currently its 2nd district includes Matagorda. The average deviation will be -4.3%.

Bell will have 2 districts. Currently its 2nd district includes Lampasas. The average deviation will be -4.6%. This should permit a Killeen based district. Currently Killeen is split. The alternative would be for the Temple district to loop into Lampasas.

Ellis will be one district, with a deviation of -1.0%. Currently, there is no county it could be paired with, and Henderson is divided with part attached to Ellis. Ellis having a district of its own will permit elimination of the one 2010 violation of the Texas Constitution.

We now start looking for quasi-floterial districts and forced groupings.

Collin and Denton have shared a district the past several decades, and it would be possible again. The two would have 10 districts, one more than they have now. But Rockwall must be paired with Collin, since no other county fits with Rockwall. Kaufman and Van Zandt must also be paired.

There is no combination of Grayson and Fannin that works, therefore it must be Grayson-Cooke-Montague. This forces Wise with Denton.

Johnson can be paired with Somervel or Bosque. The Johnson-Bosque combination is a bit ugly, but Parker and Hood can't be placed in the same district. Parker and Hood must both go west, so Somervel should go with Hood, resulting in Johnson with Bosque.

El Paso for the past few decades has had five districts, increasingly underpopulated. This will no longer be possible. It will have 4 districts with the fifth exploding east to include all of the Trans-Pecos.

Reducing the number of quasi-floterial districts will result in Hidalgo and Cameron being paired as they are now, with Starr added in. This will result in the intercounty district being Starr-western and northern Hidalgo-and a bit of northeastern Cameron. A bit ugly bu no worse than currentlly exists. Keeping Hidalgo and Cameron separate would require 8 districts and stretching far to north a la congressional fajitas.

It would be possible to pair Hidalgo and Webb, linked by Zapata and Starr, but the districts would all have to have large deviation of +4.4%.

Nueces which currently has two districts will have to go outside the county. Aransas, Jim Wells, and Kleberg are all possibilities. It may come down to fitting other districts.

Comal is somewhat trapped. Comal-Blanco-Llano is possible but ugly. Pairing Comal and Hays is simpler. Guadalupe must be paired with Gonzales.

Galveston must be paired with Chambers as it is now.

We would like to link two pairs among Fort Bend-Brazos, Fort Bend-Montgomery, Montgomery-Jefferson, Smith-Montgomery, Smith-Brazos, and Brazos-McLennan. Some of these combinations may block other districts from being created.

In West Texas, none of Potter, Randall, Lubbock, Ector, Midland, Tom Green, Taylor, Wichita, and Parker may be combined. This means that they will each have to be the core of 10 districts in west Texas (Lubbock will have one of the 10 in the county). This may strongly influence the shape of the districts.

[ProgressiveModerate]

Thanks you for doing all the work of compiling this.

Texas redistricting rules combined with the relatively small size of the state House districts and Democrat's massive geography edge means that an R gerrymander can only go so far. Think about it this way, if Republicans pulled all the stops, Dems would still get:


Houston:
~12-13 seats in Harris county
2 seats in Ft Bend

(~15 total D sinks)

Dallas:

-11 seats in Dallas county; it's possible to draw 2 R-leaning seats on 2020 numbers in university park and the NE corner of the county but they are both really marginal and would quickly fall if current trends continue
-4 seats in Tarrant County
-3 seats between Collin and Denton (this would probably be a dummymander by the end of the decade due to Collin County geography but do it's job for now)

(~18 total D sinks)

San Antonio:

-8 seats, some more marginal

Austin:

- 7 seats in Travis (duh)
- a seat in Williamson (dummymander)
- a seat in Hays

(~9 seats)

El Paso:

4 seats + a swing seat

RGV:

7 seats + a swing seat in rural north

Others:

-Loredo based seat
-Corpus based seat
-Jefferson Co based seat
-Killeen based seat (or 2 R-leaning swing seats)

If you add this all up even in an R gerrymander, Democrats would win a minimum of 65 seats assuming nothing insane happens, (they currently hold 67 seats), and would have plenty of potential openings. This is prolly one of the states where the fact that Rs have complete redistricting control shouldn't discourage Dems too much on the legistlative level.

[jimrtex]

This is the final apportionment. The number represents the number of members for the district. For the multi-county districts, all but one will be contained in the main county(-ies).



This shows deviation per district in percent.



Dallas gained a seat back. The surpluses for the 9 single-county districts were mostly greater than 0.5 with a net surplus for all nine of 0.611. Dallas was entitled to 13.451 districts or just shy of 13.5. Putting the seat back changed a surplus pf +3.47% to -3.97%.

I had miscalculated the deviation for the Kaufman-Van Zandt district. Instead of +4.4% it is +5.4%. If challenged in court, I will accept that deviation greater than 5% requires justification by the state. In this case, I will argue that the strong preference for respect of county boundaries in the constitution warrants singular exceptions. There is only one other multi-county single member district with an absolute deviation greater than 4.0% and no others above 3.0%. The same argument will be made with respect to the 7th Travis seat. Overall the standard deviation is 2.51% (assuming all districts in a single county had the same population.

An alternative will chop off a nose-shaped area of Van Zandt with under 1000 persons in a so-called Brennanmander.

Districts will be numbered based on county population in inverse order.

(127-150) Harris: 24 districts the same as now.

(113-126) Dallas: 14 districts the same as now.

(102-112) Tarrant: 11 districts the same as now.

(92-101) Bexar: 10 districts the same as now.

(85-100) Travis: 7 districts, one more than now. An attempt will be made to make the western district a Republican district.

(3,80-84) Collin-Rockwall: 6 districts, one more than now. The current outermost district is a Frisco-Rockwall wraparound that will be divided giving Rockwall a majority (55%) district.

(75-79) Denton-Wise: 5 districts, one more than currently in Denton alone. The new district will be 2/3 in Denton - likely in the western and northern areas.

(55-56,70-74) Hidalgo-Cameron-Starr: 7 districts, the same as Hidalgo-Cameron currently have. 4 districts will be entirely in Hidalgo, 2 districts entirely in Cameron, and one across all three counties (Hidalgo 49%, Starr 34%, Cameron 17%). The current Hidalgo-Cameron district was a last minute change and links La Joya with Cameron County. The new district will shift westward into Starr, with a link through the (virtually) unpopulated northern portion of Hidalgo, to a tidbit of Cameron.

(?,66-69) El Paso+9 counties of the Trans Pecos. 5 districts, the same as currently in El Paso. Previously, El Paso precariously held on to 5 low-population districts. This was only possible while El Paso had the equivalent of 4.75 or more representatives. Once it fell below that level the deficit had to be made up, plus the loss of population share for El Paso. This will shake up the districts in the county, but should still yield 5 Democrats (the new area would be competitive but it will be placed in a district with southeastern El Paso. It could conceivably provide a comeback for Pete Gallego.

(?,62-65) Fort Bend plus 6 counties to the west. 5 Districts. Currently, Fort Bend has 3 districts and dominates a fourth which includes two other counties. Four districts will be in Fort Bend. The fifth will be largely outside the county, and will likely represent someone from one of the other counties.

(?,?,59-61) Montgomery-Brazos-Waller-Washington-Grimes. 5 districts. 3 districts will be entirely in Montgomery. Currently, the 3rd extends into Waller. One district will be entirely in Brazos. The other three counties along with small bits of Montgomery and Brazos will form the fifth district.

(56-58) Williamson county will have three districts. Currently the 3rd district incudes Burnet and Milam.

(55,56) Cameron (see above).

The remainder of districts will be assigned in sequence.

[jimrtex]

This version redoes the South Texas districts.





The district west of San Antonio might have drawn VRA scrutiny under the original map, despite the fact that Trump carried seven of the nine counties. From a partisan outlook it featured a pairing of Democratic representative in a Republican district.

The green district despite its shape has a good COI based on ring counties around San Antonio. The blue district will be winnable for a Republican, particularly J.M.Lozano.

The Victoria district is also a more natural district.

I'm going to try to move Coryell into a West Texas district. This will involve splitting the San Antonio ring district in half, and migrating the two halves around the I-35 corridor up to Waco. It could get messy. involving about a dozen districts.

[jimrtex]

This version splits Coryell from the McLennan district, and divides the San Antonio ring district in half. In essence we have four half-districts, two near Waco and two near San Antonio and we have to rejoin the two eastern and the two western districts by sliding the two together.





This makes the second McLennan district more sensible. Before it would have combined Coryell, Hill, and Limestone with a small part of McLennan in a sort of super doughnut.

It also creates an additional district in West Texas and reduces the deviation in the area to about 0.5% per district.

The Wilson district now wanders a bit, but it is better than the old Walker district which was ugly.

Overall, this is an improvement.

The next map will add Burnet to Travis. This will reduce the range to 10.05%.

[EastAnglianLefty]

Presumably if you wanted to keep within 5% you'd need to pair Kaufman with Dallas? Or does the redistricting language state that counties that can stand alone must do so?

[jimrtex]

Presumably if you wanted to keep within 5% you'd need to pair Kaufman with Dallas? Or does the redistricting language state that counties that can stand alone must do so?

I'll call such districts quasi-floterial.

Districts under the Texas Constitution are made up of whole counties, but these districts may overlay. As apportionment devices they are mathematically and legally sound.

In this case Dallas is entitled to 13.451 representatives and and Kaufman 0.748. Dallas would be apportioned 13 representatives, and Dallas+Kaufman would be apportioned one. Historically, representatives would be elected at-large by position. So on Dallas County ballots, there would be 14 races, and Kaufman would have one.

Dallas voters would dominate the election since they constitute 94.7%.

After the OMOV decisions of the 1960s, the legislative gerrymanderers said OH! Goody we can carve up the map without regard to the Texas Constitution.

The Texas Supreme Court in Craddick v Smith ruled that the Texas Constitution can be interpreted in a way that harmonizes it with equal protection. So in that case an area of Dallas that has 13/14 of the total population (13.184) would elect 13 representatives from single member districts, while the remainder (surplus) of 0.266 would be placed in a district with Kaufman. Instead of a surplus being treated as belonging to the whole county, it was treated as residing in a specific area.

The plaintiff in the lawsuit was Tom Craddick who was in his 20s at the time, and one of a handful of Republicans in the House. The Democrats drew a line down his street. Craddick is now serving in his 27th term.

Because these sorts of districts are not fully consistent with the Texas Constitution they should be avoided. They are bad, but not creepy abhorrent bad.

The Texas redistricting scheme was partially upheld by the SCOTUS in White v Weiser. The main thing you will read is the finding that at-large election of representatives was ruled unconstitutional. The decision upheld the district court decision that applied to Dallas and Bexar county. Subsequent litigation applied that finding to all large counties but Hidalgo. The legislature has since just used single member districts universally. The Texas Supreme Court has ruled that single member districts is a manner regulation (the constitution only deals with apportionment not how the apportioned representatives are elected).

But the SCOTUS overturned the district court decision with regard to the amount of deviation, which ranged from +5.8% to -4.1% (this case is where the 10% range limit comes from. A range of 9.9% was accepted).

The SCOTUS has subsequently determined that the 10% range is a burden-shifting limit. Below 10%, the challenger is required to demonstrate that the deviation is unconstitutional. In the Arizona legislative redistricting case, the SCOTUS determined that Republicans had not met that burden - that deliberately under-populating Democratic districts was not unconstitutional if done for VRA reasons. Above 10%, the state must justify the excess.

I now have the range down to 10.05%. The excess of 0.05% is equivalent to 97 persons in  districts with an average population of 194,303.

But back to your original question. It has become the received wisdom that quasi-floterial districts should only be employed if necessary to fit a large county into the permitted range. Counties entitled to more than 10 representative (i.e. Harris, Dallas, Tarrant, and Bexar) can always satisfy that limit.

That guidance may be wrong. For example in 2011, Ellis could have been paired with Dallas. Ellis was just below 0.950 and could not be paired with any neighbor. But it paired with Dallas it would have avoided the egregious violation of the Texas Constitution which was the splitting of Henderson County.

The 2020 case is a little bit different because I am fully complying with the Texas Constitution at least with respect to the Kaufman-Van Zandt district, and almost complying with the US Constitution - where the 10.0% range is arbitrary.

Summary: IMO, large counties may be placed in quasi-floterial districts, but it is not necessary in this particular case.

[jimrtex]

This version adds Burnet to Travis, reducing the deviation for Travis from -5.1%, to -1.5%. Since the -5.1% would be the average for 7 districts it would likely take heroic efforts to keep the worst case above -6.0%. In cases like, Williamson (4.5%) and Brazoria (-4.3%) there is at least a tiny bit of wiggle room before exceeding 5% deviation and we have fewer districts.

I originally tried to distribute the loss of Burnet throughout West Texas, but this left the districts collectively underpopulated.

Adding Atascosa to a western district placed east and west into better balance. The cumulative western excess is around 6% or around 12,000 people. It also reduced the deviation in the eastern areas. It reduces the standard deviation to 2.200%.

While looking at this map, I realized that Blanco almost perfectly matches the excess population of west Texas. It will be added to the Travis+Burnet district.

[jimrtex]

Presumably if you wanted to keep within 5% you'd need to pair Kaufman with Dallas? Or does the redistricting language state that counties that can stand alone must do so?

I'll call such districts quasi-floterial.

Districts under the Texas Constitution are made up of whole counties, but these districts may overlay. As apportionment devices they are mathematically and legally sound.

In this case Dallas is entitled to 13.451 representatives and and Kaufman 0.748. Dallas would be apportioned 13 representatives, and Dallas+Kaufman would be apportioned one. Historically, representatives would be elected at-large by position. So on Dallas County ballots, there would be 14 races, and Kaufman would have one.

Dallas voters would dominate the election since they constitute 94.7%.

After the OMOV decisions of the 1960s, the legislative gerrymanderers said OH! Goody we can carve up the map without regard to the Texas Constitution.

The Texas Supreme Court in Craddick v Smith ruled that the Texas Constitution can be interpreted in a way that harmonizes it with equal protection. So in that case an area of Dallas that has 13/14 of the total population (13.184) would elect 13 representatives from single member districts, while the remainder (surplus) of 0.266 would be placed in a district with Kaufman. Instead of a surplus being treated as belonging to the whole county, it was treated as residing in a specific area.

The plaintiff in the lawsuit was Tom Craddick who was in his 20s at the time, and one of a handful of Republicans in the House. The Democrats drew a line down his street. Craddick is now serving in his 27th term.

Because these sorts of districts are not fully consistent with the Texas Constitution they should be avoided. They are bad, but not creepy abhorrent bad.

The Texas redistricting scheme was partially upheld by the SCOTUS in White v Weiser. The main thing you will read is the finding that at-large election of representatives was ruled unconstitutional. The decision upheld the district court decision that applied to Dallas and Bexar county. Subsequent litigation applied that finding to all large counties but Hidalgo. The legislature has since just used single member districts universally. The Texas Supreme Court has ruled that single member districts is a manner regulation (the constitution only deals with apportionment not how the apportioned representatives are elected).

But the SCOTUS overturned the district court decision with regard to the amount of deviation, which ranged from +5.8% to -4.1% (this case is where the 10% range limit comes from. A range of 9.9% was accepted).

The SCOTUS has subsequently determined that the 10% range is a burden-shifting limit. Below 10%, the challenger is required to demonstrate that the deviation is unconstitutional. In the Arizona legislative redistricting case, the SCOTUS determined that Republicans had not met that burden - that deliberately under-populating Democratic districts was not unconstitutional if done for VRA reasons. Above 10%, the state must justify the excess.

I now have the range down to 10.05%. The excess of 0.05% is equivalent to 97 persons in  districts with an average population of 194,303.

But back to your original question. It has become the received wisdom that quasi-floterial districts should only be employed if necessary to fit a large county into the permitted range. Counties entitled to more than 10 representative (i.e. Harris, Dallas, Tarrant, and Bexar) can always satisfy that limit.

That guidance may be wrong. For example in 2011, Ellis could have been paired with Dallas. Ellis was just below 0.950 and could not be paired with any neighbor. But it paired with Dallas it would have avoided the egregious violation of the Texas Constitution which was the splitting of Henderson County.

The 2020 case is a little bit different because I am fully complying with the Texas Constitution at least with respect to the Kaufman-Van Zandt district, and almost complying with the US Constitution - where the 10.0% range is arbitrary.

Summary: IMO, large counties may be placed in quasi-floterial districts, but it is not necessary in this particular case.

An addendum.

I looked at what would happen if we did create a Dallas-Kaufman pairing. It would have 14  representatives. But since we had already apportioned 14 representatives to Dallas we, have the population of Van Zandt (0.306) to distribute among other districts. If we were to split this among 10 districts, that would be 3.1% per district. Even if we chose 30 other districts it would be 1.0% per district.

In a way it is misleading to show Dallas as a surplus of 3.4% for 13 districts, or a deficit 3.9% for 14 districts. If we had a surplus (excess) of 3.4% for 13 districts in West Texas, we would say that map shows bias against West Texas.

An alternative measure would be 5% x sqrt(magnitude) as a limit.

If Dallas had between 12.820 and 13.180 we could apportion 13 representatives wholly within Dallas County.

Of our original 9 single-county districts, only two Tarrant and Ellis qualify.

Tarrant 10.863 (11 representatives for range 10.834 to 11.166)
Ellis 0.990 (1 representative for range 0.950 to 1.050)

Limit in red is exceeded for those below.

Dallas 13.451 (13 representatives for range 12.820 to 13.180)
Bell 1.908 (2 representatives for range 1.929 to 2.071)
Williamson 3.134 (3 representatives for range 2.913 to 3.087)
Travis 6.640 (7 representatives for range 6.868 to 7.132)
Bexar 10.341 (10 representatives for range 9.842 to 10.158)
Harris 24.349 (24 representatives for range 23.755 to 24.245)
Brazoria 1.915 (2 representatives for range 1.929 to 2.071)

This would not necessarily result in a surfeit of border crossings.

Dallas could be paired with Rockwall, which would permit Collin and Denton to be paired. No net change in border crossings.

Bell with Falls or Milam. One additional crossing.

Travis with Williamson and Burnet. No net change.

Bexar with Kendall, Bandera, and Medina (there are many variations). One additional border crossing.

Hays plus Comal is barely outside the limit for two districts.

Harris plus Fort Bend, Waller, and Austin. No net change.

Brazoria, Galveston, Matagorda. One additional crossing. Brazoria would have a small attachment to Galveston, one whole district, and most of a district with Matagorda.

This would require either a change in the Texas Constitution, or a ruling that the concentration of small deviations in an area is a violation of equal protection. Both are unlikely.

[jimrtex]

This places Blanco in the Travis floterial, with the ripple effects to west Texas.





If Iowa had 13 congressional districts and 10 Polk counties the map would look like west Texas. Districts comprised of grid-like counties look better and there are many variations. With this many districts it would be impossible to get the small deviations found in Iowa.

Staring at the map, the 1.1% deviation for the Hood district sticks out like a sore thumb. The next map eliminates that.

[jimrtex]

Presumably if you wanted to keep within 5% you'd need to pair Kaufman with Dallas? Or does the redistricting language state that counties that can stand alone must do so?

I'll call such districts quasi-floterial.

Districts under the Texas Constitution are made up of whole counties, but these districts may overlay. As apportionment devices they are mathematically and legally sound.

In this case Dallas is entitled to 13.451 representatives and and Kaufman 0.748. Dallas would be apportioned 13 representatives, and Dallas+Kaufman would be apportioned one. Historically, representatives would be elected at-large by position. So on Dallas County ballots, there would be 14 races, and Kaufman would have one.

Dallas voters would dominate the election since they constitute 94.7%.

After the OMOV decisions of the 1960s, the legislative gerrymanderers said OH! Goody we can carve up the map without regard to the Texas Constitution.

The Texas Supreme Court in Craddick v Smith ruled that the Texas Constitution can be interpreted in a way that harmonizes it with equal protection. So in that case an area of Dallas that has 13/14 of the total population (13.184) would elect 13 representatives from single member districts, while the remainder (surplus) of 0.266 would be placed in a district with Kaufman. Instead of a surplus being treated as belonging to the whole county, it was treated as residing in a specific area.

The plaintiff in the lawsuit was Tom Craddick who was in his 20s at the time, and one of a handful of Republicans in the House. The Democrats drew a line down his street. Craddick is now serving in his 27th term.

Because these sorts of districts are not fully consistent with the Texas Constitution they should be avoided. They are bad, but not creepy abhorrent bad.

The Texas redistricting scheme was partially upheld by the SCOTUS in White v Weiser. The main thing you will read is the finding that at-large election of representatives was ruled unconstitutional. The decision upheld the district court decision that applied to Dallas and Bexar county. Subsequent litigation applied that finding to all large counties but Hidalgo. The legislature has since just used single member districts universally. The Texas Supreme Court has ruled that single member districts is a manner regulation (the constitution only deals with apportionment not how the apportioned representatives are elected).

But the SCOTUS overturned the district court decision with regard to the amount of deviation, which ranged from +5.8% to -4.1% (this case is where the 10% range limit comes from. A range of 9.9% was accepted).

The SCOTUS has subsequently determined that the 10% range is a burden-shifting limit. Below 10%, the challenger is required to demonstrate that the deviation is unconstitutional. In the Arizona legislative redistricting case, the SCOTUS determined that Republicans had not met that burden - that deliberately under-populating Democratic districts was not unconstitutional if done for VRA reasons. Above 10%, the state must justify the excess.

I now have the range down to 10.05%. The excess of 0.05% is equivalent to 97 persons in  districts with an average population of 194,303.

But back to your original question. It has become the received wisdom that quasi-floterial districts should only be employed if necessary to fit a large county into the permitted range. Counties entitled to more than 10 representative (i.e. Harris, Dallas, Tarrant, and Bexar) can always satisfy that limit.

That guidance may be wrong. For example in 2011, Ellis could have been paired with Dallas. Ellis was just below 0.950 and could not be paired with any neighbor. But it paired with Dallas it would have avoided the egregious violation of the Texas Constitution which was the splitting of Henderson County.

The 2020 case is a little bit different because I am fully complying with the Texas Constitution at least with respect to the Kaufman-Van Zandt district, and almost complying with the US Constitution - where the 10.0% range is arbitrary.

Summary: IMO, large counties may be placed in quasi-floterial districts, but it is not necessary in this particular case.

An addendum.

I looked at what would happen if we did create a Dallas-Kaufman pairing. It would have 14  representatives. But since we had already apportioned 14 representatives to Dallas we, have the population of Van Zandt (0.306) to distribute among other districts. If we were to split this among 10 districts, that would be 3.1% per district. Even if we chose 30 other districts it would be 1.0% per district.

In a way it is misleading to show Dallas as a surplus of 3.4% for 13 districts, or a deficit 3.9% for 14 districts. If we had a surplus (excess) of 3.4% for 13 districts in West Texas, we would say that map shows bias against West Texas.

An alternative measure would be 5% x sqrt(magnitude) as a limit.

If Dallas had between 12.820 and 13.180 we could apportion 13 representatives wholly within Dallas County.

Of our original 9 single-county districts, only two Tarrant and Ellis qualify.

Tarrant 10.863 (11 representatives for range 10.834 to 11.166)
Ellis 0.990 (1 representative for range 0.950 to 1.050)

Limit in red is exceeded for those below.

Dallas 13.451 (13 representatives for range 12.820 to 13.180)
Bell 1.908 (2 representatives for range 1.929 to 2.071)
Williamson 3.134 (3 representatives for range 2.913 to 3.087)
Travis 6.640 (7 representatives for range 6.868 to 7.132)
Bexar 10.341 (10 representatives for range 9.842 to 10.158)
Harris 24.349 (24 representatives for range 23.755 to 24.245)
Brazoria 1.915 (2 representatives for range 1.929 to 2.071)

This would not necessarily result in a surfeit of border crossings.

Dallas could be paired with Rockwall, which would permit Collin and Denton to be paired. No net change in border crossings.

Bell with Falls or Milam. One additional crossing.

Travis with Williamson and Burnet. No net change.

Bexar with Kendall, Bandera, and Medina (there are many variations). One additional border crossing.

Hays plus Comal is barely outside the limit for two districts.

Harris plus Fort Bend, Waller, and Austin. No net change.

Brazoria, Galveston, Matagorda. One additional crossing. Brazoria would have a small attachment to Galveston, one whole district, and most of a district with Matagorda.

This would require either a change in the Texas Constitution, or a ruling that the concentration of small deviations in an area is a violation of equal protection. Both are unlikely.

A second addendum:

This is the distribution of districts. The red XXXXX are the districts in the four large counties of Dallas, Tarrant, Harris, and Bexar. Of the 91 other districts, 62 (68%) are within 1% of ideal (less than 2000 difference from 192,000). None of the other districts are within that range.

-5% to -4%	XXXXX
-4% to -3%	XXXXXXXXXXXXXX
-3% to -2%	XXX
-2% to -1%	XXXXXXXXXXXXXXXXX
-1% to 0%	XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
0% to 1%	XXXXXXXXXXXXXXXXXXXXXXXX
1% to 2%	XXXXXXXXXXXXXXXXXXXXXXXXXXXX
2% to 3%	XXXX
3% to 4%	XXXXXXXXXXXX
4% to 5%	XXXX
5% to 0%	X

We can with diligence get most districts close to that narrow range, with a few isolated outliers.

While this probably shows a flaw in the Texas method, it is unlikely to have a court rule that way. They will see that Bexar gets screwed, and Dallas pulls out a plum, but everything is within 5%.

Oddly, the plan approved in White v Regester had quasi-floterials in Harris, Bexar, and Tarrant counties.

[jimrtex]

This reduces the deviation for the Hood district, such that the deviation for all West Texas districts is less than 1%.

Overall, the standard deviation is 2.174%.

The districts are getting uglier.

Theory: Reducing deviation decreases compactness when districts are formed from political subdivisions.





The next map will center the Lubbock district around Lubbock, and place the Potter and Randall districts in the Panhandle. The current map was drawn so as to shove these districts over to the New Mexico border so as to make the unanchored Howard district more compact.

[President Punxsutawney Phil]

From one Texan to another - cool stuff! Thank you for your hard work! Keep it up!

[viscountviktor]

How can I actually see the maps you're talking about?

Also are you saying Dallas is going to get 14 seats?

I'm very interested in what the Texas GOP do in Tarrant and Williamson counties.

Williamson they could go 2R-1D but those R seats would be around Trump +10 and I could see them both going Dem by mid decade. Alternatively they can give Dems two seats now and create 1 safe GOP seat.

Tarrant, they need to give Dems at least 4 seats (1 more than now) otherwise their current map will fall apart. But even then, the GOP seats will only be Trump +10 to 15. Would they give Dems another seat to shore themselves up?

[ChiefFireWaterMike]

How can I actually see the maps you're talking about?

Also are you saying Dallas is going to get 14 seats?

I'm very interested in what the Texas GOP do in Tarrant and Williamson counties.

Williamson they could go 2R-1D but those R seats would be around Trump +10 and I could see them both going Dem by mid decade. Alternatively they can give Dems two seats now and create 1 safe GOP seat.

Tarrant, they need to give Dems at least 4 seats (1 more than now) otherwise their current map will fall apart. But even then, the GOP seats will only be Trump +10 to 15. Would they give Dems another seat to shore themselves up?

You need to get 10 posts on the site

[viscountviktor]

Thanks

[President Punxsutawney Phil]

Thanks

Just head to Forum Community and make 7 more posts. Doesn't matter where.

[jimrtex]

This map concentrates the Amarillo districts in the Panhandle, and the Lubbock districts around Lubbock.





Amarillo sprawls across the Potter-Randall line and I-30 (US 66) is also near the line. It is as easy to get to Vega (Oldham) and Pampa (Grey) from either of the two big counties.

The second Lubbock district comes into the city of Lubbock (the current version touches the Texas Tech campus) so it reasonable to keep the districts tight. I particularly like having Plainview (Hale) in the district.

The red district sprawls but what do expect for 27 counties. Though Howard (Big Spring) is the most populous county, it only has 14% of the population.

I just realized I can convert it into a doughnut AND improved equality.

Standard deviation for the latest map is 2.167%.

[jimrtex]

This map completes the doughnut around Lubbock.





I think this map illustrates why Iowa districts are considered exemplary: they are constructed from a square grid.

In the rural district it requires 5 counties: Howard, Gaines, Scurry, Nolan, and Terry to contain a majority of the population. 13 of the 27 counties have fewer than 5000 persons.

[jimrtex]

I've been thinking about the how the Texas Constitution has been interpreted.

Sec. 26.  APPORTIONMENT OF MEMBERS OF HOUSE OF REPRESENTATIVES.  The members of the House of Representatives shall be apportioned among the several counties, according to the number of population in each, as nearly as may be, on a ratio obtained by dividing the population of the State, as ascertained by the most recent United States census, by the number of members of which the House is composed; provided, that whenever a single county has sufficient population to be entitled to a Representative, such county shall be formed into a separate Representative District, and when two or more counties are required to make up the ratio of representation, such counties shall be contiguous to each other; and when any one county has more than sufficient population to be entitled to one or more Representatives, such Representative or Representatives shall be apportioned to such county, and for any surplus of population it may be joined in a Representative District with any other contiguous county or counties.

The text in red is the goal of the apportionment scheme.

The text in green is the provision for floterial districts. As part of the overall apportionment scheme it is valid.

It simply says that for a county entitled to 1.5 representatives, that it form a district with one representative, and that for the surplus of 0.5 that it be placed in an additional district. This is clearly preferable to granting the county two representative or only one representative.

The problem is not apportionment, but rather what happens electorally. The county is placed in the floterial district based on its surplus population, but all voters in the county, including those apportioned the single representative may vote in the floterial district. This violates equal protection under the US Constitution.

The Texas Supreme Court has said that the spirit of the Texas Constitution is followed by placing an area containing the surplus with whole counties or areas containing the surplus from other counties, but this should only be done to comply with equal protection.

But what about the Texas Constitution itself.

Clearly Dallas (13.451), Harris (24.349), and Bexar (10.341) have surpluses. If we do not recognize these surpluses we are not following the fundamental principle in the Texas Constitution that apportionment of representatives among the counties be as "nearly as may be".

[jimrtex]

This map treats Dallas (13.451), Harris (24.349), and Bexar (10.341) as having a surplus.





Dallas was paired with Rockwall (Collin was paired with Hunt in place of Rockwall), Harris was paired with Fort Bend, Waller, and Austin, and Bexar was paired with Kendall, Bandera, Medina, and Uvalde.

These changes released Atascosa, Washington, Burleson, Hopkins, Wood, and Rains. These had to be slid together to replace the extra Dallas representative.

Overall this reduces the standard deviation to 1.473%. This is a substantial improvement from where it was before I started refining the map.

This is the updated distribution. 110 of 150 districts would be within +/- 1%. 138 would be within +/- 3%.

-5% to -4%	XXXX
-4% to -3%	XX
-3% to -2%	XXXX
-2% to -1%	XXXXXXXXXXXXXX
-1% to 0%	XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
0% to 1%	XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
1% to 2%	XXX
2% to 3%	XXXXXXX
3% to 4%	XX
4% to 5%	XXX
5% to 0%	X

[jimrtex]

Harris, Fort Bend, Waller, and Austin counties collectively will have 29 representatives.



24 districts will be wholly in Harris, 4 districts will be wholly in Fort Bend, and one will combine the surplus of Harris, the surplus of Fort Bend and the two smaller counties.

Approximate population in this quasi-floterial district will be:

Harris 63K
Waller 56K
Fort Bend 45K
Austin 30K

In general, I intend to use the school districts as guides to drawing districts in Fort Bend and Harris counties. They represent a recognized community of interest as well as having more regular and stable boundaries than cities. In addition, public education is the largest component of the state budget.

School district boundaries do not necessarily conform to county boundaries. Districts were originally created by county school boards, and provided elementary education to largely rural students prior to motorized transportation. Consolidation enabled the independent school districts to offer high school education, and in some cases considered ease of transportation, even when it meant crossing county boundaries. For this exercise I will largely treat parts of districts in different counties as separate.

Districts will conform to election precincts which do not always match up with school district boundaries. If the house district boundaries were made to match election precinct  boundaries then the election precincts would have to be adjusted.

[jimrtex]

This shows the relative population of the ISD's in Harris. You will notice that the election precincts do not closely align with the ISD boundaries. Historically, school board elections were not held on the same date as general elections (it was actually illegal at one point to have city or school board elections in November). School districts held their own elections, often in the the spring, and might use elementary schools as polling places and attendance zones as precincts). It is required that general election precincts match up with city limits of cities greater than 10,000. The intent is to permit the cities to use the same polling places as counties, and simplify voting rolls. If I were doing this for real, I might conform the house districts to the school district boundaries.



0.323 is needed for the quasi-floterial district. We start with Waller ISD, and continue into Katy ISD. This has the advantage that both districts extend into Waller County. The cities of Waller and Katy are both on the county line. We can create another district entirely in Katy ISD, and a small amount detached to permit a Spring Branch ISD district. This has a total population of 2.372. If we use this it will put the five districts that are totally or partially in Fort Bend, plus the Katy and Spring Branch districts at roughly 0.7% over the ideal, which is reasonable.

Cy-Fair ISD at 3.025 easily can be divided into the 3 districts.

Houston ISD plus Alief ISD equals 9.038 which would produce one district in Alief ISD, seven in Houston ISD, and one cross-ISD district.

The northwestern ISDs of Tomball ISD, Klein ISD, Spring ISD, and Aldine ISD don't have obvious alignment into districts, but we can create districts that can be identified as Tomball-Klein, Klein, Spring, and Aldine. Collectively we have a total of 4.473. So we have a remnants of about half a district. We will get back to this later.

The northeastern ISDs of Humble ISD, Sheldon ISD, Galena Park ISD, Huffman ISD, Crosby ISD, and Goose Creek (Baytown) ISD have a population of 2.890 which corresponds to 3 districts. Moreover, we can group these as Humble; Sheldon, Galena Park, and Channelview; and Goose Creek, Crosby, and Goose Creek which each have roughly the population of a district. The latter three would form a long skinny area that is east of the San Jacinto River which isolates them somewhat from the rest of the county (the river prevents residential development).

The southeastern ISDs of Pasadena ISD, Deer Park ISD La Porte ISD, and Clear Creek ISD has a population of 2.551. Like the northwestern district there is an excess of about one half district.

We cross-over from Aldine ISD and Pasadena ISD into Houston ISD to create a 5th northwestern and a third southeastern district, reducing the number of districts wholly in Houston ISD to 6 districts. The combined area which stretches from Alief to Tomball to Clear Creek is 16.062. Large areas of the city of Houston are in Aldine ISD and Pasadena ISD.

[I’m not Stu]

Will the TXGOP try to protect as many incumbent Republicans as possible, or will they go with weird shapes that draw out incumbents?

[jimrtex]

Will the TXGOP try to protect as many incumbent Republicans as possible, or will they go with weird shapes that draw out incumbents?

In 2010 they paired Republicans, either getting one of the pair to retire or trying to make it a "fair" battle. They were able to create new districts in suburban counties to compensate. In 2010, the population of Republican-held districts was equivalent to four extra districts. So basically they could keep the same number of districts but with increased margins.