Is GIS Still An Effective Political Redistricting Tool?

By Joe Francica

In a recently article published in the Proceedings of the National Academy of Sciences (PNAS) by Dartmouth College Professor Ben Forest (pictured at right), the author questions the effectiveness of GIS in the political redistricting process.In a press release issued by the college, Dr.Forest presents three arguments that affect the political representation process:_
  1. The U.S.population has become more racially and ethnically diverse in the last 40 years.
  2. There has been a growing legal commitment to fair political representation.
  3. GIS allows states to create districts with very precise political and demographic characteristics.
I asked Dr.Forest to clarify some of the points of his thesis in the following interview:

Joe Francica (JF): You state that "even where states use GIS to expand representation, other factors like bureaucratic and institutional practices end up limiting" the technology.If there is a basic lack of understanding of how the technology provides the basis for redistricting, are legislatures and redistricting technologies headed for a collision where perhaps the courts will end up mandating the use of certain criteria, geographic data or other digital information to further restrict the use of GIS?

Ben Forest (BJ):
In that particular quote, I am referring to limitations on political representation, rather than to limitations on the technology.What I tried to show with the Texas example [in the article] is that institutional practices ultimately excluded "communities of interest" that were not defined by boundaries.

As for the second part of the question, courts (and the Constitution) already mandate the use of certain data (Census population data based on actual counts, for example, rather than population estimates), and regulate the use of some data more generally (particularly involving race and ethnicity).On the other hand, courts have signaled that they will not regulate the use of political data used for partisan purposes.

My sense is that courts will probably not impose the use of specific data on their own, for example, by requiring that legislatures use a particular measure of compactness.The Texas case is complicated because the original Vera decision was a reaction against nonwhite-majority districts.The court argued, in essence, that the state put too much emphasis on race; the court saw the failure to consider "communities of interest" as evidence of this excessive consideration of race.

Legislatures or voters themselves (through ballot initiatives) are, of course, free to specify the kind of data to be considered in redistricting.

JF: Can you foresee some means to make "communities of interest" an absolute digital representation so that they can be factored into the redistricting process?

BJ: I think there are two different issues here.First, both Texas and Arizona did use GIS (or related technologies) to create representations of communities of interest.My general point is that such representations - based primarily on boundaries - are not the only way to think about communities of interest.

One can imagine other ways to represent communities in GIS.For example, a geographer at Ohio State, Mei-Po Kwan, has pioneered the use of GIS to trace the "time geographies" of individuals.At its most basic, a "time geography" is just the path that an individual follows during his or her daily routine.If you combined the "time geography" of 1,000 people, for example, you might find that 150 of them commuted along the same stretch of freeway at the same time of day (morning and afternoon rush hours).This convergence of paths arguably creates a "community of interest" that is not defined by boundaries.I am not arguing that this is a superior way to think about communities of interest, or that states should consider such "commuting communities" in redistricting.Rather, I simply argue that these sorts of representations have not found their way into redistricting GIS.

My second point (and this might be less clear in the PNAS article than elsewhere) is that "community of interest" is a relatively imprecise term that can refer to many different things.The inability to create an "absolute digital representation" of such communities rests from the fuzziness of the definition, not on any particular technical limitation.

JF: You state that "GIS technology has not generally fulfilled" certain promises to resolve certain political conflicts.Aren't you blaming the tool for the inequities for which politics are more to blame?

BF: Actually, I think this is my point.GIS, as a technical tool, cannot overcome the very basic contradictions in American political representation.I am not "blaming" GIS for this failure, but rather trying to correct what I saw as overly optimistic expectations about what GIS would do for democratic representation, and to clearly separate technical and political issues.In fact, I thought of the article in part as exonerating GIS from this "failure"!

JF: It is not clear from your thesis exactly how the redistricting process was carried out in your examples.Would you favor a weighting system that categorizes each boundary file? And would that not lead to more political problems?

BJ: I am not advocating any particular formula for redistricting.The Constitution and certain Federal laws (the Voting Rights Act) place certain restrictions on redistricting, but beyond that, I think states can and should determine the basic values they want to see reflected in redistricting.For example, a large state might want to emphasize geographic compactness, while a small state might want to emphasize the representation of distinct political interests. Once the legislature has set these goals, however, I think a properly designed redistricting commission can apply these rules with less political bias than an elected body.

JF: You suggest that limiting the redistricting process to GIS professionals who are non-partisan would be a better solution.However, 1) isn't that negating the essential element of the redistricting process for which political parties have fought for years; and 2) how does any single entity remain apolitical? That seems to be wishful thinking.

BF: I am not sure I understand the first part of the question, so my answer may be off base.The most important "essential element" that political parties have sought is partisan advantage (bias).I think eliminating (or limiting) partisan bias is a good idea, even if parties have sought it for years.Or did I miss the point of the question?

The second part of the question has to do with the design of redistricting commissions.Badly designed commissions would simply replicate or even exaggerate partisan interests or corrupting influences, while properly designed ones would minimize such problems. (I think several of the opponents of the current redistricting initiative in California may favor redistricting commissions in principle, but argue that a badly designed commission is worse than none at all.) I am not an expert on this particular issue, and it has been the subject of considerable discussion lately.Consequently, I have tried to always add the qualifier "properly designed" to "commissions" without getting into the nitty-gritty how to design a commission properly.I agree that it is naive to think that redistricting commissions will always or automatically rise above the political fray, but I think it is reasonable to discuss ways to minimize harmful political influences.

Courts are perhaps a good analogy here.Few people think that courts are totally apolitical, but we recognize them as legitimate because they use principled decision-making rather than explicit political considerations in rendering judgments.(I realize the irony of this comparison in the middle of a Supreme Court nomination process!)

There was a good article in The New York Times on October 23, "Who Should Redistrict?" (Dean E.Murphy) that lays out the case against redistricting commissions.It might be of interest.

JF: You state that GIS has "rationalized" the process but not "produced better democracy." Would you favor a non-factored, aspatial process whereby you simply use Theissen polygons to create districts based on an equally weighted demographic distribution?

BF: Again, I am not advocating any particular set of redistricting criteria.Commentators have suggested various redistricting procedures, including variations of the process you describe.The short answer is that such processes typically produce many different possible solutions, so picking one is still a political question.

On one level, questions about the design of districts are part of a much broader discussion about electoral systems.Different systems favor different principles (political responsiveness vs.stability, for example).It is not that one system is necessarily better than another, but different systems produce different effects.I think that it is extremely important that we (citizens, representatives) discuss the principles that we value most, and select an election system (and districting principles) that reflects those values.In my view, once this happens, a (relatively) neutral, (relatively) nonpolitical body is in the best position to then apply these principles in a fair and unbiased manner.

JF: Where does GIS go from here? Seems likely that the problems will only increase given wider use of the technology by all states and also consultants called into the process.

BF: Hmmm...I have been told to never make predictions, but I can see several different possibilities.These are all speculation (some more than others).

First, it is possible that courts, state legislatures, Congress and Federal agencies will together muddle out a compromise.I think there is an historical analogy for this.In the late 1980s and early 1990s, the Department of Justice and states started emphasizing the creation of black- and Hispanic-majority districts, producing some well-known gerrymanders.At the time, some of the more dire predictions suggested that this would lead to the total "balkanization" of the electorate along racial lines.In a series of decisions, the Supreme Court reigned in this practice, forbidding states from making race the "predominant" consideration in redistricting.It is not a legally elegant (or even clear) standard, but it serves to limit the extreme cases.

Second, there is the "Armageddon" scenario, where the dominant political party in each state creates new districts for each election that enable them to perpetuate their power indefinitely (or for a very long time).I think this is fairly unlikely (and some states like Colorado have constitutions forbidding such practices), but this is the possibility that really keeps me up at night.Again, history provides a good analogy: Between the 1920s and 1960s, rural legislators managed to block redistricting to balance population in most states.By the early 1960s, many states had huge problems with malaportionment, and rural interests wielded far more power than their population warranted.As populations grew more imbalanced, the anti-redistricting forces became more entrenched.The Supreme Court stepped in and ruled that malaportionment was unconstitutional, but if they had not done so, it is hard to see how the normal electoral process could have broken this logjam.

Third, I think that states will continue to integrate redistricting GIS into their general information technology infrastructure.In 2011 or perhaps 2021, states may not maintain a separate, dedicated redistricting GIS, but may use systems that they have for general state use.This is not necessarily good or bad, but it may require careful consideration of privacy issues (as does all technology with data-matching capability).Consultants and parties, however, will certainly want to maintain their own, separate systems.

Fourth, I think the big story for GIS will be (and has been) its use in campaigns and elections, rather than in just redistricting.GIS are fantastic tools for tracking and analyzing campaign activities like fund-raising, canvassing, voter registration, "get out the vote" activities, etc.Parties already make use of GIS for such activities, so I think there may be a lot of demand for more "user friendly," portable systems.Like redistricting, I think the new technology favor groups that can make the best use of it, (i.e., political parties), but to the extent that such activities increase democratic participation, I see them as a very positive development.Moreover, my sense is that grassroots and community organizations can make more effective use of GIS for these activities compared to redistricting, so again I think this could be a very positive force for democratic participation.

Published Tuesday, November 15th, 2005

Written by Joe Francica

If you liked this article subscribe to our bimonthly newsletter...stay informed on the latest geospatial technology

Sign up

© 2017 Directions Media. All Rights Reserved.