Coalition for Non-Partisan Redistricting
for adopting redistricting plans in Texas
The Legislature would not adopt a specific district map, but would adopt an act specifying the parameters according to which a computer program would draw the district map.
The computer program to be used would be Target, a program already written under the aegis of the Texas Legislative Council, consisting of a random map-generating engine written in Scheme, a dialect of LISP, with a user interface written in Visual Basic.
The parameters would be as follows:
1. District boundaries would snap to county boundaries, except for counties that are split.
2. Only counties with a population of 60,000 or greater would be split, and within those counties, district boundaries would snap to township boundaries if possible.
3. Each district would have equal population, to within a error of 0.01 percent
4. Each district would be compact, defined by minimizing the value of p²/4A for each district, where p = perimeter, A = area.
5. Districts would be contiguous, so that any two points are connected by a line that lies entirely within the district.
6. Districts would be simply connected, so that any closed loop of points within the district may be shrunk to a point, so that there are no "holes".
7. There would be no consideration of any other demographic factors than number, such as ethnicity or voting history.
8. The program would be run once for a period not to exceed 24 hours, and whatever map it produced would be the district map for the next election.
9. Redistricting would be done using the above procedure every two years, in January of a regular election year, to draw the districts for the next election.
10. The source code of the Target program would be published on the web site of the Texas Legislative Council, and made freely available to all persons for examination and use.
11. There would be continuing development of the Target program to improve its performance and usability.
12. The operation of the program to draw districts would be supervised by the grand jury of the county in which the State Capital is located, and outside observers would be allowed to verify that the program had not been compromised.
1. The State Constitution requires that the Legislature adopt redistricting plans, but it does not require that it adopt a bill with a specific map. Adopting a process such as the one described above to produce a map using a computer arguably satisfies the requirement without the need for a constitutional amendment. It is not an unconstitutional delegation of legislative authority to assign such a task to the impersonal operation of a computer.
2. The congressional district plan approved by the federal court for the 2002 election is valid only for that election, and does not remain the plan for subsequent elections in the absence of further action by the State Legislature.
3. The federal Voting Rights Act contemplated district plans drawn manually taking voting patterns into account, and sought to avoid ethnic or political biases to which that method is susceptible. However, if the parameters for the program are chosen as described above, with no human intervention, the plan produced should meet the requirements of the Act and of the court precedents on these issues.
The main alternative proposal would create a "bi-partisan" Redistricting Commission composed of equal numbers of Republicans and Democrats, with perhaps one tie-breaker member, who would draw a map and present it to the Legislature for adoption. There are several difficulties with this proposal that would be avoided in the main proposal for computer-generation of district maps:
1. Having the map drawn by such a Redistricting Commission does not avoid the problem that whatever it draws will be subject to the same kind of contention in the Legislature that has resulted on maps drawn in other ways. The only way to avoid such divisiveness is to remove the consideration of a particular map from the Legislature, and leave to them only the adoption of specifications.
2. A Redistricting Commission would be an unnecessary expense, because it can't do anything that can't be better done by the computer program, supervised for the short period it is run by the local grand jury, which should be selected by sortition, that is, at random, from the community, and whose only job would be to make sure the operation of the program was not compromised.
Contents | RTF Version | Text Version