Opinion | Why computers alone can’t fix Michigan gerrymandering

Algorithms don’t remove all human biases in redistricting, because humans create the algorithms and, in the end, still must pick among multiple computer-drawn maps.

Eric Lupher is president of the nonpartisan Citizens Research Council of Michigan.

Proposal 2 on the November statewide ballot seeks to move responsibility for drawing the lines that define congressional and legislative districts away from the Michigan Legislature and give it to an independent commission made up of randomly appointed citizens.  

Some have asked, why stop there? Why not create an algorithm, a formula, some sort of digital tool, free of human meddling, to guide how the commission should draw maps? People, even if they do not see themselves either as Democrats or Republicans, have innate biases that can arise when asked to perform a task such as redistricting. Couldn’t we just leave it to the computers to do it for us?

Before turning this critically important function of our democracy over to HAL 9000, the self-aware computer from the movie “2001 A Space Odyssey,” let me offer a few observations. 

Let’s start with an understanding of the term algorithm. Dictionary.com defines it as “a set of rules for solving a problem in a finite number of steps.” In the case of redistricting, an algorithm would be a mathematical formula that would sort people – specifically the communities people live in – into congressional and state legislative districts.

Therein lies the first issue: Redistricting algorithms do not exist on a store shelf for the state to purchase and plug in the requisite number of districts into which the state should be sorted. Someone has to create the algorithm. And that person will bring some level of bias to the task. What’s more, the state official(s) responsible for hiring that person will also have a partisan interest.

While an algorithm certainly can meet all criteria given, it is only as good as the data it uses. Anything not inputted (and even some of the criteria that are ranked lower) may get ignored in ways that have unpredictable, undesirable outcomes for other reasons and should not be the only authority.

That leads us to a second issue: There is no single solution or “best” district plan. In fact, we are able to judge the extent to which Michigan’s political boundaries are gerrymandered (unfairly drawn) because academics have used the same data and criteria used by the state to create hundreds of alternative maps. Someone has to decide among the options produced by the computers.

Also, some of the redistricting criteria required under Proposal 2 are rather subjective. Over time the meaning of “communities of interest” or the requirement that the districts should not provide one political party with a disproportionate advantage may become established law. Until that time, humans will have to decide their meaning. An algorithm has no way of doing this on its own.

How is it that a computer formula is to going to reconcile instances where redistricting criteria are inconsistent with one another? Prioritizing helps, but humans will ultimately have to decide. Consider the first and last guidelines established under Proposal 2 – districts must comply with the federal Voting Rights Act and should be reasonably compact.

If mapmakers create a compact minority-majority district, it would be highly “packed” (meaning, for example, that it would place as many African Americans that typically vote for Democratic Party candidates in a single district to minimize the number of elections they can influence). Humans will have to decide which suburban communities will be included with districts that lie predominantly in majority black Detroit.

Finally, we should make it clear that algorithms have been used for past redistricting exercises.

Republicans have controlled the Michigan Legislature for the past several election cycles. The political power brokers hired consultants who used their own algorithm for drawing maps. That algorithm contained most of the criteria included in the proposal. The algorithms created several iterations of gerrymandered maps from which the party leaders choose their preferred maps.

Redistricting has been permanently changed by technology; computers can create several options in minutes rather than days. The question that Michigan voters must answer is who should be responsible for choosing the criteria used in these algorithms and selecting the final output that will sort us as voters.

Proposal 2 asks voters if they wish to create a redistricting commission as an independent decision-making body to make those decisions. Because it would be comprised of equal numbers of Republicans and Democrats and include a number of independent voters, the intent is that some neutrality, or at least compromise, would be introduced to the design of the algorithm.

With an understanding that even a neutrally designed algorithm is capable of producing several iterations of district maps, the proposal creates processes through which people favoring both major parties and people favoring neither must come to a consensus on which map is the “best” for congressional, state Senate and state House districts.

With any luck, we can avoid a Skynet-type (warning: Terminator reference) conflagration along the way.

You can access the Citizen Research Council’s analysis of this and the other Michigan ballot questions at https://crcmich.org/ballot-issues/.  

Bridge welcomes guest columns from a diverse range of people on issues relating to Michigan and its future. The views and assertions of these writers do not necessarily reflect those of Bridge or The Center for Michigan. Bridge does not endorse any individual guest commentary submission.

If you are interested in submitting a guest commentary, please contact Ron FrenchClick here for details and submission guidelines.

Facts matter. Trust matters. Journalism matters.

If you learned something from the story you're reading please consider supporting our work. Your donation allows us to keep our Michigan-focused reporting and analysis free and accessible to all. All donations are voluntary, but for as little as $1 you can become a member of Bridge Club and support freedom of the press in Michigan during a crucial election year.

Pay with VISA Pay with MasterCard Pay with American Express Donate now

Comment Form

Add new comment

Dear Reader: We value your thoughts and criticism on the articles, but insist on civility. Criticizing comments or ideas is welcome, but Bridge won’t tolerate comments that are false or defamatory or that demean, personally attack, spread hate or harmful stereotypes. Violating these standards could result in a ban.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.


John Chamberlin
Wed, 09/12/2018 - 9:31am

A very useful and lucid explanation. The CRC's new report on Proposition 12-2 (independent redistricting commission) is must reading for citizens who want to understand the details of the proposal.

Yi-Li Wu
Wed, 09/12/2018 - 4:22pm

Agree! (Just a minor correction---the proposal number on the ballot is 18-2....meaning proposal 2 for the year 2018).

Kevin Grand
Wed, 09/12/2018 - 2:23pm


We cannot let people draw political districts because they have inherent biases.

We cannot allow a computer to draw political districts because they will also have inherent biases because of those who programmed them.

And exactly what was the whole point behind this post again?

Yi-Li Wu
Wed, 09/12/2018 - 4:20pm

The point is that replacing humans with machines doesn't eliminate bias, because ultimately humans control the process. So you have to find some way to compensate for human nature. And this is the purpose of the CHECKS AND BALANCES that are built into Proposal 2 and the citizen-run redistricting commission it would create. ( BTW, the details of this commission were created based on best practices from existing commissions in other states, synthesized with the ideas expressed by Michigan voters during 33 town halls around the state). First, the commission's business is TRANSPARENT, open to public scrutiny from start to finish, so that everyone has the ability to know WHO is doing it, HOW they are doing it, and WHY they are doing it, and ALL citizens have the ability to express their views at many points in the process. Before any maps are drawn, the commission must hold 10 public hearings around the state. During the map drawing and evaluation process, all meetings are public, and any materials, computer programs, etc. that are employed must be placed on the public record. Then, when the maps are proposed, they must have another 5 public hearings. Furthermore, the commission is intentionally multi-partisan, consisting of 4 members from each of the two major parties (currently Dem and GOP) plus 5 voters who do not affiliate with the major parties. For a map to be adopted, 2 commissioners from each of these groups (Dem, GOP, non-affiliated) must agree. This compels the commission to reach decisions by compromise, resulting in maps that are as FAIR and IMPARTIAL as is humanly possible. Yes, humans are gonna human. And Proposal Two is designed to eliminate the flaws of the current process, a current process that encourages and enables the worst of human nature. The current process is controlled by majority party politicians in Lansing and their special interest donors, whose aim is to mold our state in ways that benefit them, regardless of whether they are serving the voters' best interests. All done behind closed doors, in secret, with NO accountability. And you cannot "vote them out" because they have manipulated the district lines to guarantee their re-election and continued majority. Both parties do it, and the next redistricting is in 2020. The choice is clear: Status quo? or reform? We have potholes in our roads, pollutants in our water, and our school system is sliding fast. That is why Michiganders of all political persuasions are supporting Proposal 2 for redistricting reform.

Kevin Grand
Thu, 09/13/2018 - 10:12pm

That sounds nice on the surface.

Now tell me EXACTLY how you intend to find out who is really independent and how do you propose addressing party sycophants who will try and slip through the system to get on this commission?

Agnosticrat 2.0
Sat, 09/15/2018 - 2:40pm

Everyone slow down. We are going too fast for Kevin and I want him to have a chance to understand.

Kevin Grand
Sun, 09/16/2018 - 11:19am

Oh, I've understood what the democrats were up too months ago when they announced their little scheme.

And like I've brought up elsewhere; how does replacing gerrymandering, with another form of gerrymandering, fix the original problem of gerrymandering?

The democrats backing this proposal have been dancing around that troublesome question for all this time without adequately answering it.

Sat, 09/15/2018 - 2:38pm

Thank you Yi.
That is one of the best explanations I’ve read.

Wed, 09/12/2018 - 8:09pm

Damned if you do, damned if you don't.
Still, the proposal includes members of both worthless parties and in numbers to overrule any "independents".
Sounded like such a great idea to start with and the usual suspects mucked it all up.

Ben W Washburn
Wed, 09/12/2018 - 9:07pm

4 times in my life, I have personally done the prevailing analysis and determination of elective districts for the Detroit School Board and the Wayne County Commissioners. These are some of the lessons that I learned. The first aim is to make every district as close as possible to the same number of constituents as indicated by the most recent census. State law has allowed up to a 16% variation between districts. I was usually able to get that down to less than a 1% variation. The Wayne County Charter wisely requires that commission districts include whole cities so far as possible. That again is a wise criterion, because this is the way that most folks are already best organized, politically. The Charter also required that districts be as square as possible. That prevents most of the more egregious kinds of gerrymandering. The Charter also provided the opportunity for the minority party (Republicans, in the case of Wayne County) to provide an option which better fit these required criteria. They made an honest effort, but were unable to come up with a better plan. But, it is important that that opportunity be provided. School Board seats were and are non-partisan. The School Board was not required to consider any other options, but we nevertheless did. Here, my main criterion after equalizing the number of constituents, was to recognize the existing clear-cut boundaries between Detroit's many organized communities. Very few of these groups span major freeways and traffic arteries. It is very hard for any candidate for office to describe his or her district to their included school communities without reference to these kinds of boundaries. And it is equally hard for their constituents to envision the district in which they reside without reference to these kind of boundaries. Some folks at Wayne State came-up with an alternate plan which was computer generated, but it totally sliced and diced all of these boundaries, and made no common sense to anyone involved. So far as is possible, all political districts should recognize and accommodate the ways in which their constituents are already organized to express their political will.

Who Bought You
Wed, 10/24/2018 - 10:53am

I don't think you understand how algorithms work. A computer algorithm is unbiased and it can easily be tested to prove that it is unbiased. It would be simpler cheaper and faster. People have already made algorithms that can do this and can be statistically verified to be unbiased.