Scholars advocate groundbreaking use of technology in Wisconsin redistricting cases

This blog post continues the focus of the Law School’s Lubar Center on redistricting.

Wisconsin could be one of the first states in the nation to draw legislative and congressional maps with powerful computer technology originally developed to expose gerrymandering, under an approach being advocated in state and federal courts.

Two overlapping groups of number-crunchers, mostly professors from the University of Wisconsin-Madison and Marquette University, are urging judges to let them use the technology to create districts that would improve representation for people of color.

One group, calling itself the Citizen Mathematicians and Scientists, has been admitted as intervenors in the redistricting case before the Wisconsin Supreme Court. The other group, dubbed the Citizen Data Scientists, has been denied intervenor status in related litigation before a three-judge federal panel, but will be allowed to file friend-of-the-court briefs in that case.

In both cases, the groups’ members represent themselves as nonpartisan researchers. But two of them, and several of their attorneys, have previous ties to Democratic and progressive interests in high-profile litigation around redistricting and other political disputes.

The technology that they are advocating uses high-speed computers to create an “ensemble” of hundreds or even thousands of possible maps for comparative analysis. Its evolutionary path starts with former U.S. Supreme Court Justice Anthony Kennedy and winds through Wisconsin.

In a 2004 Pennsylvania case, Vieth v. Jubelirer, the high court refused to find that partisan gerrymandering violated federal law. But Kennedy, a frequent swing vote, suggested in a concurring opinion that the practice might be unconstitutional if someone could come up with “clear, manageable, and politically neutral standards” for measuring how it disadvantaged particular voters.

Wisconsin’s 2011 legislative redistricting became a test case for meeting the goal that Kennedy had laid out. Leveraging their control of the governor’s office and both houses of the Legislature, Republicans pushed through maps that cemented their majorities in the state Senate and Assembly for a decade to come.

Democrats’ first challenge to the maps, in Baldus v. Wisconsin Government Accountability Board, succeeded only in altering the boundaries of two Assembly districts on Milwaukee’s south side. But with Kennedy’s words in mind, Democrats tried again, using a new technique to measure the “efficiency gap”—the number of votes cast for each candidate beyond the bare majority needed to elect them—in districts overwhelmingly tilted toward one party or the other.

A three-judge federal panel ruled in the Democrats’ favor in the 2016 case, Gill v. Whitford, the first time a federal court had struck down a partisan gerrymander. But before the Supreme Court could decide on the merits of that case, Kennedy retired. In a 2019 North Carolina case, Rucho v. Common Cause, Kennedy’s successor, Justice Brett Kavanaugh, joined the majority in a 5-4 ruling that dashed hopes that the high court would overturn any partisan gerrymander. Gill was soon dismissed.

Meanwhile, Jowei Chen, associate professor of political science at the University of Michigan, had combined the efficiency gap and ensemble approaches to demonstrate the degree of partisan gerrymandering in Wisconsin.

For a 2017 paper in the Election Law Journal, Chen used a computer to generate 200 possible legislative maps. He found the GOP-drawn Assembly map’s efficiency gap in favor of Republicans was “over twice as large as even the most biased of the 200 plans produced by the non-partisan computer simulation process,” demonstrating an “extreme electoral bias.”

Until now, however, the ensemble approach hasn’t been used to propose new maps, partly because many of the randomly generated maps do not comply with the standards of the Voting Rights Act. That landmark federal legislation requires districts to be drawn in a way that ensures members of racial and linguistic minorities can elect representatives of their choice.

In a forthcoming paper for the Election Law Journal, Tufts University mathematics professor Moon Duchin joined attorney Sam Hirsch and two data scientists from Duchin’s Metric Geometry and Gerrymandering Group Redistricting Lab in proposing a way to use the ensemble approach to draw maps that would comply with the act.

Duchin, a nationally known redistricting researcher, helped Wisconsin Gov. Tony Evers’ People’s Maps Commission draw its proposed legislative and congressional maps. Her lab also works with civil rights groups to strengthen voters’ rights.

Hirsch is a partner at the Chicago-based law firm of Jenner & Block, which is part of the legal team representing the scholars in the Wisconsin cases. The firm also represents a similar group in Minnesota, where that state’s supreme court has appointed a panel of judges to consider redistricting plans if Minnesota’s legislature and governor cannot agree on new maps.

One of the most common ways to comply with the Voting Rights Act has been to draw Black or Hispanic “supermajority” districts, increasing the likelihood that lawmakers of color would be elected from districts where their group constitutes the overwhelming majority of the population.

But some U.S. Supreme Court decisions have raised questions about whether this approach violates the 14th Amendment’s equal protection clause, as Duchin and her colleagues note in their paper. Instead, these individuals have devised an algorithm that uses detailed election returns and demographic data to calculate the likelihood that minority-preferred candidates will win both primaries and general elections in a proposed district.

“The computational redistricting methods available today didn’t exist a decade ago,” Hirsch says. “Mathematicians, computer scientists, and data scientists have worked hard in recent years to develop evenhanded mapmaking techniques. Their expertise—and their maps—could be very helpful to the [Wisconsin] Supreme Court, especially given how quickly the court may have to act.”

John D. Johnson, research fellow in the Lubar Center for Public Policy Research and Civic Education at Marquette Law School (and not a participant in the paper), says the approach advocated by Jenner & Block’s clients would be “a departure from the way these ensembles of algorithmically generated maps have historically been used.”

“Typically, advocates have used these massive ensembles of random maps to demonstrate how severely gerrymandered Republican-generated maps are, with the aim of convincing courts to overturn them,” Johnson says. “The Duchin paper, by contrast, is trying to help liberals in Democratic-controlled states draw maps that maximize minority electoral power without falling afoul of the law.”

Neither Duchin’s team nor the Wisconsin groups characterize themselves that way.

Three of them—UW-Madison computer science professors Somesh Jha and Stephen Joseph Wright and applied mathematics professor Jean-Luc Thiffeault—belong to both the state and federal groups. Two others—Marquette math professor Gary Krenz and associate professor Sarah Hamilton—belong only to the state group. The rest—Michael Switzenbaum, Marquette professor emeritus of environmental engineering; Madison health actuary Leah Dudley; and Joanne Kane, a research assistant for the National Conference of Bar Examiners—belong only to the federal group.

“The Citizen Mathematicians and Scientists are professors who have a nonpartisan interest in seeing the redistricting process proceed fairly and transparently for all Wisconsin voters,” their attorneys say in a document filed in the state case, echoing a similar representation in the federal case.

However, Switzenbaum and Dudley were also among the plaintiffs in Gill. In that case, they were described as “supporters of Democratic candidates and policies,” whose “ability to affiliate with like-minded Democrats and to pursue Democratic associational goals has been impaired by the current [GOP-drawn] plan,” in a court document submitted by their attorney, Doug Poland.

Poland is now litigation director of the progressive Law Forward, which is representing the voting rights groups who were among the original federal plaintiffs in the current Wisconsin redistricting litigation. He also represented the Democratic plaintiffs in Baldus, while Hirsch represented the Democratic plaintiffs in Vieth.

The Wisconsin professors’ local lawyers include former Madison City Attorney Michael May and Sarah Zylstra, both of the Madison-based firm of Boardman & Clark. They represented Democratic state senators in the 2002 litigation that led a three-judge federal panel to draw a new Wisconsin legislative map. Zylstra also represented the state Democratic Party in its challenge to the 2010 “lame-duck” laws that limited the powers of Evers and Attorney General Josh Kaul.

Hirsch and Poland say those Democratic ties don’t mean they are advocating for a map that favors Democrats.

“Our clients aren’t in this case to advance the political interests of Democrats or Republicans or any other piece of the electorate,” Hirsch says. “They want maps that will be fair and effective for all Wisconsinites. And fairness and neutrality are far more likely if a computer draws a million maps and systematically evaluates them than if you entrust one person to draw one map by hand and just hope for the best.”

Poland agrees that people who favor Democratic policies can put those views aside and support fair maps. He says his current clients—Black Leaders Organizing for Communities, Voces de la Frontera, and the League of Women Voters—also are seeking fair maps, not Democratic gerrymandering.

Anthony LoCoco, deputy counsel of the conservative Wisconsin Institute for Law & Liberty, which is representing the original state plaintiffs in the current litigation, says he doesn’t know much about the scholars’ groups or their technology, but WILL didn’t oppose their intervention. He says he will wait to see their maps before judging whether they are partisan.

Besides the scholars and some of the original plaintiffs on both sides, most of those who were involved in the redistricting litigation as of Oct. 24 are major political figures. Also:

  • Thiensville attorney Daniel Suhr filed a friend-of-the-court brief with the Wisconsin Supreme Court in support of WILL’s position that the redistricting issue belongs in the state court. He described himself as “a public-minded attorney who comes as a true friend-of-the-court with no agenda besides his personal view on the correct outcome.” Suhr is a managing attorney for the Chicago-based Liberty Justice Center.

In Wisconsin, Liberty Justice Center is representing the conservative MacIver Institute in its federal lawsuit demanding that Evers grant the think tank’s employees access to news media briefings. WILL is representing former Gov. Scott Walker as a friend of the court in support of MacIver’s position.

  • Not bothering with the formalities of intervenor or friend-of-the-court status, Middleton resident William R. Taylor simply wrote a letter with his ideas about the case to U.S. District Judge James Peterson.

Taylor urged the federal court to draw a map dividing each of the state’s eight congressional districts into four state Senate and 12 Assembly districts. That is likely beyond the judges’ power, because it would require a change in state law to reduce the Assembly from 99 to 96 seats and the Senate from 33 to 32 seats. The Senate rejected a similar concept in 1971, when the state had nine congressional districts.

Peterson, as chief of the three-judge federal panel, thanked Taylor for his thoughts and promised to share them with his colleagues.

This Post Has One Comment

  1. Tom Kamenick

    Thank you for your thorough look at what’s going on with all of this. The ensemble method can be a really useful tool. But drawing districts to benefit one race is just as bad if not worse as drawing districts to benefit one party.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.