Many states configure their electoral districts in ridiculous ways to benefit a certain party or candidate over another. This drawing of a potential district in Chicago, Illinois is an example of the ways voting districts are manipulated to work in favor of a party.Many states configure their electoral districts in ridiculous ways to benefit a certain party or candidate over another. This drawing of a potential district in Chicago, Illinois is an example of the ways voting districts are manipulated to work in favor of a party.

Every ten years the U.S. Census Bureau launches hundreds of thousands of data-collecting workers into the country. Their basic task is to tally up all the people living in the United States and note changing demographics, and although what comes back is an immense sweep of numbers, this work is, unfailingly, humanizing. It’s an image of who we are: awkward, crude, confusing, like a class picture in middle school. When I think census I get all nostalgic for the imagined American ideal, fresh scrubbed census workers hitting the road, knocking on doors, talking to their regional cohort over a tall glass of milk.  In the age of Google Maps, there is something so delightfully old fashioned about the whole thing I begin to believe again in the genius of democracy. 

There is a mean average of 711,00 people jammed into the each of the 435 congressional districts across the United States.  Each state, in turn, has its own legislature consisting of smaller districts allowing for more local representation. After every new census, new district maps need to be drawn. On the national level, states with greater population growth pull congressional districts from states with population loss, but even if a state maintains their current level of districts, they still need to redraw their maps due to shifts within their borders.  This practice, known as redistricting, was intended to keep democracy sharp and the electoral scales balanced. However, according to The Atlantic’s Robert Draper, “Redistricting today has become the most insidious practice in American politics—a way, as the opportunistic machinations following [the] census make evident, for our elected leaders to entrench themselves in 435 impregnable garrisons from which they can maintain political power while avoiding demographic realities.”  Redistricting permits politicians to appeal to a narrow slice of the electorate, which encourages a regressive, horse-and-buggy kind of thinking on the part of people running for office. 

“Redistricting today has become the most insidious practice in American politics—a way, as the opportunistic machinations following [the] census make evident, for our elected leaders to entrench themselves in 435 impregnable garrisons from which they can maintain political power while avoiding demographic realities.” 

What partisan redistricting does is eliminate the need for the give and take of politics. If you stack voting blocs in your favor, you don’t need to answer to your constituents and you never need to compromise. When representatives are not concerned about being voted out in a general election and are forced to answer only to their base, the result is an unbalancing, a tipping of scales toward the more extreme views of a particular party. Party polarization is one concern, another is that partisan redistricting causes representatives to be allotted unproportionally. If a state assembly has 100 individual districts, basic rules of symmetry call for both parties to translate popular support into representation. Accordingly, if the state voted 55% Democrat, it is reasonable to expect to see 55 Dems in the state legislature. However, if a political party gets 48% of the state-wide vote and captures 60 assembly seats, there might be a problem with that state’s political symmetry. I’m looking at you, Wisconsin!

According to Dr. Jamie McKown, James Russell Wiggins Chair in Government and Polity at COA, there are a number of factors to consider when it comes to redistricting. Chief among them when trying to create more equitable political representation is defining what norms we, the electorate, value. What do we want? McKown points out the paradox in trying to create fair district maps: does fair necessarily equal good? And what exactly does fair mean when it comes to electoral politics? Questions like these beg for more than simplified sound bite responses. McKown addresses these topics in class and plans to further scrutinize them in a new COA course tentatively titled The Seven Deadly Sins of American Politics. When McKown and his students explore some of the more vexing political issues (like redistricting, campaign finance reform, or the electoral college), he likes to use the analogy of having a holey winter coat. You don’t need your winter coat until you need it and then you realize (again) there are holes in it but you wear it anyway vowing (again) to fix or replace it before next season comes along.

The Gerry in Gerrymander

The Framers left redistricting up to the states; subsequently, if one political party is in charge of the state legislature, it will seek to draw a map warped toward favoring others in the party. This method of friendly map making has become known as “gerrymandering.” Gerrymandering gets its name from an 1812 political cartoon satirizing a redistricting map drawn under the administration of Massachusetts Governor Elbridge Gerry.  All of the other districts in the map were relatively compact but one, and it, according to the cartoon, was shaped like a reptile. (Gerry + Salamander = Gerrymander). 

Gerrymandering gets its name from a 1812 political cartoon satirizing a redistricting map of Massachusetts under Governor Elbridge Gerry. One district shaped like a reptile earning it the name Salamander (Gerry + Salamander = Gerrymander).Gerrymandering gets its name from a 1812 political cartoon satirizing a redistricting map of Massachusetts under Governor Elbridge Gerry. One district shaped like a reptile earning it the name Salamander (Gerry + Salamander = Gerrymander).

In the two hundred years since Gerry made his mark, the practice of gerrymandering has become so ingrained in the US political system that teams of specialized mapmakers advise states involved in redistricting how best to maximize their party’s returns on election night.  The two principal tactics used in gerrymandering are “cracking” and “packing.” To “crack” a district is to dilute the power of like-minded voters by splitting them across multiple districts and to “pack” a district is to concentrate like-minded voters together in one district to reduce their voting power in other districts. For examples of how egregious the practice has become, one need not look any further than recent court cases involving Wisconsin, Maryland, North Carolina, and Texas. 

Over time, many protections against blatant partisan redistricting have been built into the legal system: the First Amendment’s prohibition on viewpoint discrimination, the Equal Protection clause of the Fourteenth Amendment, the Voting Rights Act of 1965, and a whole lot of court interpretations regarding “compactness” and “contiguity” of districts.  While the Supreme Court has weighed in on the issue, saying redistricting to create a partisan advantage can be unconstitutional, it has also indicated that a certain level of gerrymandering is to be expected.  With both parties looking to bend the rules in their favor, the question becomes, literally and metaphorically, what is over the line?  If the Supreme Court gives a definitive ruling on the constitutionality of extreme party gerrymandering in June, what guidelines will states be given in determining what fair district shapes look like? 

Shapes: Fair and Otherwise

If you have a problem with your car you get yourself a mechanic; if you have a problem creating shapes you get yourself a geometer. (Yes, Virginia, there is a geometer.) A collective calling itself the Metric Geometry and Gerrymandering Group (MGGG) formed to help experts and the public understand the math of redistricting. The MGGG was organized by Moon Duchin, an associate professor of mathematics at Tufts University, in order to “bring mathematicians together with experts in law, politics, and voting rights as we head into the 2020 census.” The MGGG hopes to see its work used as a tool for holding politicians accountable when they create district maps.  In order to create good districts one first has to hypothesize a good district shape, and this is where the math comes in. 

When attempting to create equitable district maps, the ideal factors often considered are proportionality, competitiveness, partisan fairness, and minority representation. Unfortunately, these are easier said than done, and often at least one of them is sacrificed.When attempting to create equitable district maps, the ideal factors often considered are proportionality, competitiveness, partisan fairness, and minority representation. Unfortunately, these are easier said than done, and often at least one of them is sacrificed.

Dr. David Feldman, a math and physics professor at COA, applied and was chosen to attend an MGGG workshop aimed at raising awareness of the gerrymandering problem. This past fall, Feldman gave a Human Ecology Forum at COA where he outlined several desirable features often considered when talk turns to creating equitable district maps: proportionality, competitiveness, partisan fairness, and minority representation. All of this seems sensible until you start doing the work of drawing lines around populations—what, if anything, will be sacrificed?  

There is a whole host of minutia to consider when tackling a problem as intrinsic to the American electoral system as gerrymandering. The plaintiffs in Gill v. Whitford, Wisconsin’s gerrymandering case before the Supreme Court, have offered two standards to gage political redistricting: political symmetry and “the efficiency gap.” To illustrate the difficulty of creating better district maps, Feldman suggested looking at the pros and cons of just one of these tests, the efficiency gap. The efficiency gap measures “wasted” votes, all votes cast for losing candidate and all the extra votes cast for the winning candidate are considered “wasted” (note: every time Feldman used the adjective “wasted” to describe votes, he visibly cringed). One then adds up both numbers, finds the difference between the two sides, and divides that by the total number of votes in a state. This will yield a single percentage: the efficiency gap. The lower the percentage the better the map. Feldman used the graphic below to illustrate how redistricting might take shape when considering the efficiency gap:

A Formula Goes to Court: Partisan Gerrymandering and the Efficiency Gap. Bernstein and Duchin authors.A Formula Goes to Court: Partisan Gerrymandering and the Efficiency Gap. Bernstein and Duchin authors.

Plan I shows a pretty equitable map (six districts split evenly between plain and starred voting blocks), each district is a 4-3/3-4 split, which will produce highly competitive races due to candidates needing to appeal to a closely divided electorate.  Plan II shows that same district only now in a highly gerrymandered state.  A large proportion of plain voters have been packed into one district and all the other districts have a 4-3 majority for the star party. This map would be flagged as having an efficiency gap of -1/3. Plan III shows the same district outcomes as Plan I, and even though both have an efficiency gap of zero, the result in Plan III is noncompetitive elections. Plan III would not be flagged by the efficiency gap. This is where the problem lies.

Enter the Algorithm 

In the 1986 case Davis v. Bandemer, the Supreme Court agreed it had the power to intervene in cases of partisan gerrymandering but declined to do so because, according to Justice Kennedy, the court lacked a “manageable standard” to indicate when it had occurred. The efficiency gap has been getting a lot of press because it gives courts a single, judicially discernable and manageable standard. But, as Duchin points out, the major flaw in the efficiency gap is that “partisan gerrymandering is a multi-dimensional problem” impossible to reduce to a single number. It’s like someone giving you only the area code of their phone number, you get a slice of information but it’s woefully incomplete. 

Feldman supports a multidimensional approach to the problem of gerrymandering. The MGGG and other mathematicians, political scientists, and computational experts working on gerrymandering advocate building algorithms able to explore the enormous universe of possible districting maps. If a programmer is given specific rules for fair districting (i.e. contiguous districts, respect county borders, comply with Voting Rights Act, etc.), computers can generate thousands upon thousands of maps giving the public the ability to flag outliers in a legislature’s proposed redistricting plans. One such team at Duke compared the current Wisconsin legistlative district map to 19,184 alternatives. The graph below shows how much of an outlier Wisconsin actually is. 

Evaluating Partisan Gerrymandering in Wisconsin. Herschlaga Raviera and Mattinglya authors.Evaluating Partisan Gerrymandering in Wisconsin. Herschlaga Raviera and Mattinglya authors.

Computational mathematicians at the University of Illinois Urbana-Champaign using completely different algorithms found similar results to those of the Duke group.  That’s the sort of check and balance that makes Feldman excited about this technology.

People Vote, Land Does Not

It would be ideal if all districts could be carved up into neat packable shapes, but they can’t.  There are a number of factors that thwart the dream of hexagonal districts, chief among them is the reality of self-gerrymandering. Feldman noted that larger urban areas across the country tend to be more Democratic and rural areas lean more toward Republicans…this is a truism in the U.S., except in the south where sometimes it’s not. Geographical borders also come into play when maps are drawn, rivers, mountain ranges, and inlets bisect districts. All of these different types of borders must be taken into account. In addition, Feldman added, each state has a unique political geography that varies from region to region.  Some states, due to their particular demographics, might require strangely shaped districts, districts that are far from compact. This must be part of the equation especially when we consider the age old adage: people vote, land does not. 

Not all maps can be drawn into neat packable shapes due to state borders, urban centers, and natural features such as lakes, rivers, and mountain ranges. Computers can help with this work, but a political process and judicial ruling is needed to solve the country-wide problem of gerrymandering.Not all maps can be drawn into neat packable shapes due to state borders, urban centers, and natural features such as lakes, rivers, and mountain ranges. Computers can help with this work, but a political process and judicial ruling is needed to solve the country-wide problem of gerrymandering.

To arrive at a meaningful critique of any redistricting plan, Feldman stated the need to simulate many different alternatives. It does little good to compare Wisconsin’s district map to that of Rhode Island. We want to “respect each state’s unique political and geographical topography,” ergo we want to compare Wisconsin to other possible Wisconsins. Computers can help do this work but computers alone will not solve the gerrymandering problem. With all the possible variables you cannot program a computer to spit out a single “best map” of a state, but computers can show when district maps are disenfranchising voters. The Court can help by issuing a ruling that declares once and for all that partisan gerrymandering undermines basic rights guaranteed by the constitution. And, instead of offering up a silver bullet, the Court could acknowledge the value of a multidimensional approach when it comes to drawing good maps. Committed scholars and researchers can help by giving people the knowledge and tools they need to hold map-drawers accountable. And you can help, dear reader, by educating yourself and using the wealth of tools at your disposal.

When it comes to gerrymandering, both Feldman and McKown agree that the solution will be more political than judicial. A lot of the nuts and bolts will be left up to voters. If enough of the electorate concludes partisan gerrymandering is a disease eating the host body from the inside out, the maps will become more equitable.  Collectives like the MGGG are doing their best to keep the conversation pointed toward political action and nonpartisan solutions. 

“Democracy becomes a government of bullies tempered by editors. The editors standing in the privilege of being last devoured.”

Writing in a similar time of unbalancing, Ralph Waldo Emerson remarked, “Democracy becomes a government of bullies tempered by editors. The editors standing in the privilege of being last devoured.” When it comes to gerrymandering, the temporary gains one party can make are not worth the erosion of confidence in our representative system. 2020 is coming over the horizon, slouching toward Bethlehem, PA, to alert us, to offer us a precious gift: an awkward, crude, confusing picture of who we are.  It is up to us, the electorate, to move beyond political payback and middle school myopia because what we think we already know often obscures the amazing thing we’ve never encountered before. 

Sources for this article: The Atlantic, The Washington Post, The New York Times, and the amazingly patient Dave Feldman.

To find out more about the Metric Geometry And Gerrymandering Group go to their website and read up. Many great links, articles, ways to get involved: https://sites.tufts.edu/gerrymandr/