Wharton professor Steven O. Kimbrough discusses the Supreme Court's recent decision on gerrymandering along with Mimi McKenzie of the Public Interest Law Center, and Thomas Wolf of the Democracy Program at the Brennan Center for Justice at New York University.

The Supreme Court’s decision last week that it would not make a ruling on what constitutes excessive partisan gerrymandering in defining the boundaries for various voting districts has left that responsibility with Congress and, more immediately, with state legislatures, courts and related offices. The court had been considering two cases of alleged excessive partisanship — one from North Carolina and another from Maryland — in reaching its conclusions.

Steven O. Kimbrough, Wharton professor of operations and information management, recently appeared on the Knowledge at Wharton radio show on SiriusXM to discuss how the Supreme Court’s decision not to become involved in gerrymandering cases will likely play out. He was joined by Mimi McKenzie, legal director at the Public Interest Law Center, and Thomas Wolf, counsel with the Democracy Program at the Brennan Center for Justice at New York University. (Listen to the full podcast above.)

Kimbrough is also a co-author of the opinion piece below, in which he outlines his ideas for the best ways to arrive at the boundaries that define voting districts, including the use of computer algorithms to remove some of the human bias that often creeps into such decisions.

On the radio show, he compared the recent decision to the Supreme Court’s action on pornography in 1964, pointing out that the court had said, “We can’t define pornography, but we can sure recognize it when we see it.” From there, noted Kimbrough, the court said it could “make lots of rulings about it even though we can’t define it,” by studying the evidence carefully and then deciding what to do.

By contrast, the court rejected the definitions of gerrymandering offered by the plaintiffs in the case, according to Kimbrough. He went on to characterize the Supreme Court’s decisions as saying in effect: “We don’t know how to define unfairness. We recognize it — it’s unjust and unfair — but instead of looking at the evidence, we’ll …. basically forbid the courts from looking at the evidence.”

Noted McKenzie: “The court has certainly abdicated its responsibility to protect the fundamental right to vote…. There should be protections under federal constitution for sure. It’s a sad day.”

Wolf called the 5-4 decision a “stunning shoulder shrug from the Supreme Court. The court had a perfect opportunity in these cases to finally lay down a clear rule that took out the worst abuses in our redistricting process.” In Wolf’s estimation, there were no legitimate barriers preventing the court from making a decision on what excessive partisanship looks like, but nevertheless “they walked away. It is pretty jaw dropping when you think about the ultimate consequences of this….”

“When you create extreme partisan gerrymandering in particular, you are letting legislators pick voters rather than the other way around.”

Wolf added: “Legislatures need to reflect us. When you create extreme partisan gerrymandering in particular, you are letting legislators pick voters rather than the other way around.”

In a look at some more fair ways of drawing up voting district boundaries — or redistricting — below, Kimbrough and his co-authors recommend methods that employ more public input, among other things. The co-authors are Lee Hachadoorian, assistant director, professional science master’s in geographic information systems, Temple University; Peter Miller, researcher in the democracy program at the Brennan Center for Justice at New York University School of Law; and Frederic H. Murphy, professor emeritus, Fox School of Business, Temple University. 

Redistricting plans are usually created by state legislatures and signed into law by governors. Redistricting as it is practiced is not about finding good plans or serving the public’s interest. It is about serving the interests of the politicians who draw up the plans. Redistricting is of the political parties, by the political parties and for the political parties. This description is of course an exaggeration, but not by much.

The current redistricting regime is increasingly seen as unacceptable, because it is prey to abuse of power in redistricting, known as gerrymandering. There is now much agitation for change in how redistricting is done. This agitation may or may not produce real changes anytime soon, although we do note successful redistricting reform ballot initiatives in 2018 in Colorado, Michigan, Missouri and Utah. Whether or not activism is successful, however, it is certain that redistricting plans will be subject to intense scrutiny from many different interest groups, bringing the latest in computerized “big data” and analytics technology to bear. By consequence, redistricting will be very much in the public eye. News organizations, good government groups and political parties will see to it that this happens. They have been among the strongest proponents of redistricting reform.

The situation prompts an interesting question: What constitutes a good redistricting plan? While we have plenty of examples of bad plans, there has been very little discussion or investigation of what good plans look like. The public’s views are not seriously consulted during legislative redistricting and it is not known what they are. Why don’t we ask? The authors of this article founded the Redistricting Values Discovery Project for this purpose of discovering what the public’s values are with regard to redistricting, and what it wants in redistricting plans.

“A key prerequisite to designing good redistricting plans is to realize and face up to the fact that tradeoffs have to be made.”

A key prerequisite to designing good redistricting plans is to realize and face up to the fact that tradeoffs have to be made. For example, minimizing partisan advantage is often touted as a good thing, as is minimizing the splitting of counties and municipalities into multiple districts. In general you cannot minimize both in a single plan. It’s not possible. So how much partisan advantage (regardless of party) should we give up in order to reduce place splitting? This is the kind of question, regarding tradeoffs and maintaining party neutrality, that we are asking the participants in our values elicitation sessions. Our questions do not in fact assume that either minimizing partisan advantage or minimizing splitting of places is a good thing. Instead, we draw attention to apparently significant values and let people speak for themselves. If they say they want to increase partisan advantage or splitting of places, for example, we record what they tell us.

While it is early days in our project, fascinating patterns are emerging from the data we have collected so far. Most striking to date is that participants do seem to be divided, without rancor and only very roughly, into two groups. One group seems to be oriented towards favoring what we call “partisan fairness.” The other group seems to be oriented towards favoring what we call “maintaining the integrity of place” (aka placemaking). Both orientations can be measured in multiple ways.

Both groups see positive value in having partisan fairness and in maintaining integrity of place, but they differ in which they think is more important. For example, participants may be asked to choose between two redistricting plans, A and B. Plan A is strong on integrity of place and weak on partisan fairness, while plan B is strong on partisan fairness and weak on integrity of place. Which would you prefer, if you had to choose between just these two plans? We are finding substantial numbers of people on both sides.

Provisionally and very roughly, respondents who are settled, older, and rural or suburban have tended to prefer plan A (no place splitting), while respondents who are mobile, younger and urban have tended to prefer plan B (partisan fairness). The appeal of justice and fairness in plan B is universal and easy to understand. Respondents who favored type A plans have told us of the importance in their lives of counties, school districts, municipalities, neighborhoods and other markers of place. They want their representatives to be responsive to voters concerned with these important places.

There are dozens of tradeoffs like those between plan A and plan B. They are in fact made when a redistricting plan is selected, whether we realize it or not. Yet, how these tradeoffs should be made, and even what they are, is mostly unknown and greatly under studied.

* * *

The overarching purpose of our Redistricting Values Discovery Project is to contribute to public deliberation on redistricting. As we noted above, this discussion is happening and will intensify, whether or not legislatures continue to try to control the process. By talking to people and eliciting their values pertaining to redistricting, we can inject into the discussion nuance, perspective and points of view, which might otherwise go unnoticed. Public discussion of redistricting has heretofore focused on comment in the media and on discussion of maps associated with proposed plans. In asking people about what they value in redistricting plans, we add a new dimension to deliberation on redistricting.

Values elicitation — asking people about their values regarding redistricting — is central to our project and will continue throughout. We are also actively engaged in complementary work in support of public deliberation, including:

  • Developing and administering educational exercises with real data (Miller et al., 2018). Participants find the exercises engaging and are usually able to produce redistricting plans that are superior to the plans developed by state legislatures, across a range of widely-held values. Doing the exercises is itself a means of discovering redistricting values, both for observers and the participants themselves.
  • Creating algorithms and computerized software that can discover large numbers of valid redistricting plans (Kimbrough and Miller, 2018).
  • Building and applying analytics tools that can process and assess large numbers of redistricting plans in order to inform public deliberation and select good plans (Knowledge at Wharton, 2011; Gopalan et al., 2013)
  • Publishing scholarly research, position papers and messaging statements that report our results and contribute to public deliberation on redistricting (Kimbrough et al., 2018).

In closing, we offer an idealized vision of how public deliberation on redistricting could work: Redistricting plans continue to be developed and chosen during a 1–2 year period following release of the decennial census data, as they are today. During this period, all interested parties are invited to propose any number of plans for consideration. Proposed plans are made public in standard formats so that all interested parties can examine the plans and comment on them. Vigorous public deliberation ensues through hearings, town hall meetings, public debates, editorials, blog posts, and other forums. At the end of the redistricting period, a carefully chosen and well-staffed non-partisan commission chooses a plan for implementation. The commission is tasked with thorough evaluation of the public deliberation and with choosing as best it can in the interests of the public. Legal challenges to the commission’s choice would be burdened with demonstrating that a clearly and substantially better plan had been presented to the commission and rejected. In deciding the case, the courts would judge in the same manner mandated to the commission: an honest grappling with the public’s values, with the plans on offer, and with the evidence-based comments provided by the public.

If this vision appears unrealistically idealistic, we counter that much of the description can be implemented in the public square by interested individuals and organizations. It can happen without the legislature acceding to a non-partisan commission or even a fair process, for it is a way to make public opinion matter and be a force in the outcome. We have designed the Redistricting Values Discovery project and its components, described above, for the sake of furthering this vision of deliberation. We welcome collaborators and partners.

Further Reading

Gopalan, R., Kimbrough, S. O., Murphy, F. H., and Quintus, N. (2013). The Philadelphia districting contest: Designing territories for City Council based upon the 2010 census. Interfaces, 43(5):477–489.

Kimbrough, S., Miller, P., and Murphy, F. (2018). A new approach to redistricting in Pennsylvania and beyond. Philadelphia Inquirer in print and online at http://www.philly.com/philly/opinion/commentary/gerrymandering-pennsylvania-redistricting-court-20180105.html.

Kimbrough, S. O. and Miller, P. (2018). Evolutionary heuristics for redistricting and zone design. Joint EURO/ALIO International Conference 2018 on Applied Combinatorial Optimization. https://events.unibo.it/euroalio2018.

Knowledge at Wharton (2011). A New Approach to Decision Making: When 116 Solutions Are Better Than One.

Miller, P., Kimbrough, S., and Schacht, J. (2018). Simulating redistricting in the classroom: A binding arbitration decision game using Louisiana census data. PS: Political Science & Politics, 51(3):664–668. doi:10.1017/S1049096518000410.