Texas House 1970 LRB plan.
Representatives Per Apportionment Area
Deviation per representative.
This is of interest because it eventually established the 10% maximum deviation rule for legislative districts. In 1971, the legislature reapportioned for the first time since
Reynolds v Sims. The legislature drew a map that was invalidated by the Texas Supreme Court in
Smith v Craddick because it only vaguely followed the Texas Constitution.
Reapportionment was then turned over to the Legislative Redistricting Board. Texas still used multi-member districts in larger counties with the exception of Harris where they drew 23 single member districts. They did eliminate floterial districts, so districts were established in Potter(1), Lubbock(2), Ector(1), Wichita(1), Taylor(1), El Paso(4), Hidalgo(3), Cameron(1), Nueces(3), Bexar(11), Bell(1), Tarrant(9), Grayson(1), Smith(1), Jefferson(3), Harris(23*), Galveston(2), and Brazoria(1) and areas including a surplus population were attached to adjacent counties.
Dallas(18), McLennan(2), Travis(4), Denton(1), Gregg(1) were close to a whole number of districts and were treated as such.
El Paso(4), Hidalgo(3), Nueces(3), Bexar(11), Tarrant(9), Jefferson(3), Galveston(2), Dallas(18), McLennan(2), and Travis(4) were all constituted as multi-member districts elected by position (e.g. there were 18 countywide races in Dallas County). Only Harris(23*) was configured in single-member districts. These multi-member districts were overturned as a result of litigation, but have no importance to apportionment (except by 2021 the legislature has forgotten they are actually charged with apportionment).
There were four instances where the surplus of a county was divided:
Hidalgo: One surplus area was attached to the surplus area of Cameron; the other was attached to Willacy, Kenedy, and Kleberg.
Brazoria: One surplus area was attached to Fort Bend; the other was attached to Matagorda and Wharton.
Smith: One surplus area was attached to Upshur, Marion, Titus, and Cass; the other was attached to Henderson, Anderson, and Freestone.
Jefferson: One surplus area was attached to Hardin and Jasper; the other was attached to Orange.
These were of dubious constitutionality. If they corresponded to floterial districts, then Smith would have elected one representative, and voted in two floterial districts.
In addition, Red River County was divided. Bowie County was entitled to 0.908 representatives. In 1960, it had been given its own district based on an entitlement of 0.939. But by evolving standards a -9.2% deviation might have been regarded as too much.
Bowie's neighbors of Red River, Morris, and Cass were entitled to 0.192, 0.165, and 0.323 representatives, respectively. A combination of Bowie and Morris would have an entitlement of 1.073, or a +7.3%. This might have been accepted. However, when I attempted to draw a House map starting from such a district, I was not able to. Drawing House districts in East Texas is hard because the population of the individual counties is fairly large, so coming up with combinations of say three counties that have a population equal to 95% to 105% is difficult. And once you find one such combination, it may conflict with another combination. Or those two may interfere with the next and so on.
The district of Cherokee, Houston, Leon, and Limestone does not reflect a community of interest. It represents contiguous counties with a population approaching the quota.
Because many East Texas counties have odd shapes following rivers, the resulting maps look like gerrymanders.
So what the LRB did was to draw a district entitled to 2.051 representatives, and then divide Red River, attaching a portion to Bowie, so that one district had a deviation of +2.2% and the other of +2.9%.
Since that time, it has been observed that for a map to comply with equal protection, that a violation of the Texas Constitution might have to be made. I think perhaps that this has morphed into a belief that there must be one violation.
In the maps I have shown what I have referred to as apportionment areas, whole counties entitled to close to a whole number of representatives. It is an approach that I tend to follow in all redistricting efforts. It decomposes a complex problem into smaller problems. One could imagine that the drawing of districts within an apportionment area could be delegated to a local commission. If you followed along as I drew my 2020 map, you will see it was the method I used.
This map shows the actual district boundaries.
1971 LRB plan modified as a result of Graves v Barnes (PDF)There are some other oddities. Bexar was entitled to 11.125, and under modern standards would have been apportioned 11 representatives with 1.1% deviation (about 800 persons in districts with 75,000 persons).
But instead six-counties from Hays to Frio were attached. By themselves, the six counties were entitled to 0.996 representatives, only 300 under the quota.
The legislature added an overly large portion of Bexar, such that the district ended up +4.6% deviation, while the Bexar deviation was reduced to 0.7%.
Harris was entitled to 23.336 representatives, Galveston to 2.275 representatives, and Chambers to 0.163. They were combined into an apportionment area entitled to 25.774. The deficit for the entire area of -0.226 was barely an improvement over the initial surplus of +0.336 for Harris alone.
As soon as the LRB issued their map, it was challenged (repeatedly) in federal court. After consolidation, a federal district court in
Graves v Barnes ruled that single-member districts must be used in Dallas and Bexar counties. The legislature had drawn single-member districts in Harris County. In addition, the district court ruled that the state plan as a whole violated equal protection because of excessive deviation without justification.
Supreme Court Justice Powell stayed that portion of the district court decision, while permitting the division of Dallas and Bexar into single-member districts to proceed.
For the 1972 election, multi-member districts continued to be employed in El Paso(4), Lubbock(2), Tarrant(9), McLennan(2), Travis(4), Nueces(3), Hidalgo(2), Galveston(2), and Jefferson(3). They would be gone by 1976.
The SCOTUS in
White v Regester reversed the district court's decision with respect to overall deviation, while affirming the portion of the decision with respect to single member districts in Dallas and Bexar counties. On remand, the court ordered all but two counties to be divided into single-member districts. The legislature in 1975 went ahead and divided all counties into single-member districts and it has been that way ever since.
But here we are more interested in the SCOTUS decision on overall deviation.
A careful observer will note that the deviation ranged from +5.8% (Harrison-Rusk) to -4.1% (Victoria-Calhoun) for an overall range of 9.9%. It is because of this occurrence that the overall range is used, rather than a +/- 5% deviation. The SCOTUS failed to apply any tests such as mean absolute deviation or standard deviation to determine whether a good faith effort for population equality was made. As a result, legislators have tended to push as many districts as possible to the extremes.
In the LRB plan, 75 districts were within 1% of the quota (assuming perfect division of counties), 125 were within 2%, and only 6 were more than 4%.
Justice Brennan in his dissent bitterly/presciently observed:
... one can reasonably surmise that a line has been drawn at 10% -- deviations in excess of that amount are apparently acceptable only on a showing of justification by the State; deviations less than that amount require no justification whatsoever.
The following map could have been drawn (if the JRB rather than the LRB had drawn the map). Deviations range from -4.66% to 4.45%, with a total range of 9.11%. Moreover, the map is more compliant with the Texas Constitution, by eliminating the split surpluses in Smith, Jefferson, Brazoria. Like the LRB plan there would have been a violation related to Bowie County. What I have shown as a 2-seat district would have had to be split. If the Iowa rule had been used, then the division would occur in the most populous county which is Bowie.
The split might have been: Bowie 0.672, Cass 0.323; and Bowie 0.236, Red River 0.192, Franklin 0.071, Titus 0.224, Morris 0.165, and Camp 0.107. This division would permit Texarkana to fit in the first district. Also the population in Cass is somewhat concentrated towards the east (Atlanta and Queen City).
JRB plan. Magnitude. In districts entitled to more than one representative, an area entitled to N-1 representative would be drawn in the largest county, where the representatives would have been elected at large by position, with the exception of Harris, where 23 single-member districts would be drawn. After 1975, all such large counties would be drawn as single-member districts.
An area including the surplus population would be attached to the other counties in the apportionment area.
Deviation is calculated assuming perfect division of each apportionment area. This is for measuring quality of the apportionment.
I used the same method for recognizing surpluses as I did for the 2020 map:
(1) A surplus will be recognized if the absolute deviation for the apportionment area will be greater than 5%.
(2) A surplus between 0.250 and 0.750 will always be recognized;
Rule (2) only applies to larger counties (5 or more representative). For smaller counties, Rule (1) is triggered when they have surpluses in that mid-range.
Surpluses between 0.250 and 0.750 are closer to one-half (0.5) than they are to zero or one, and better fit the classic example of counties with surpluses around 0.5 (0.4 and 0.6) sharing a floterial district. Smaller or larger surpluses might not have been recognized prior to 1965 (Reynolds v. Sims). A county entitled to 1.2 or 1.8 representatives might have been apportioned 1 or 2 representatives with little concern about the deviation.
Based on these rules, surpluses were recognized for:
Bell 1.668 (2) -16.6%
Brazoria 1.451 (1) +45.1%
Cameron 1.880 (2) -6.0%
Ector 1.230 (1) +23.0%
Galveston 2.275 (2) +13.8%
Grayson 1.115 (1) +11.5%
Hidalgo 2.432 (2) +21.6%
Jefferson 3.279 (3) +9.3%
Lubbock 2.402 (2) +20.1%
Nueces 3.182 (3) +6.1%
Potter 1.213 (1) +21.3%
Smith 1.301 (1) +30.1%
Taylor 1.311 (1) +31.1%
Wichita 1.633 (2) -18.4%
Not only would these counties have localized deviations, collectively they would have lost 2.372 districts. This is likely a systemic problem. We would normally expect more counties entitled to around 1 representative than around 2 representatives, more around 2 than 3, and so on. Similarly there would be more counties entitled to one representative that would have a small surplus, than those that were almost entitled to two representatives and have a deficit. Bell, Cameron, and Wichita would be happy to simply be rounded up. All the others would feel cheated if rounded down.
Two large counties had their surplus recognized. Under modern standards, their error in representation would be distributed among all districts in the county. But this is a system error. There is no difference between denying representation on a consistent basis to all persons in Houston or Fort Worth, than if you overpopulated districts represented by Democrats, or racial minorities (or Anglos) or Republicans, or farmers. All are a systemic bias in violation of equal protection.
Harris 23.336 (23) +1.5%
Tarrant 9.596 (10) -4.0%
These larger counties were kept whole in my map. Bexar and El Paso did have surpluses recognized in the LRB plan. The collective surplus of 0.344 is relatively small.
Bexar 11.125 (11) +1.1%
Dallas 17.782 (18) -1.2%
El Paso 4.813 (5) -3.7%
McLennan 1.977 (2) -1.2%
Travis 3.959 (4) -1.0%
These counties were entitled to a single representative. Both the LRB and JRB added to small counties to Tom Green which brought the overall entitlement closer to the quota. This is entirely consistent with the Texas Constitution. No counties were divided.
The LRB added Zapata to Webb. The resulting surplus of +3.4% was slightly larger than the initial deficit. It might have been justified by the isolation of Zapata, but my map found a reasonable alternative.
The LRB plan did not give a district to Orange, but added a tiny area (3300) of Jefferson to Orange, thus splitting Jefferson's surplus. This may have been done to keep the overall deviation range within 10.0%. But as my map demonstrates the two large districts with deviations of 5.8% and 5.7% were not required.
Denton 1.013 (1) +1.3%
Gregg 1.017 (1) +1.7%
Orange 0.953 (1) -4.7%
Tom Green 0.952 (1) -4.8%
Webb 0.976 (1) -2.4%
More to come ...