Time is ripe for computational redistricting to expose gerrymandering

The U.S. Census Bureau recently released data required by states to draw their new congressional district maps. This will begin the decennial effort of congressional mapping when state legislative or independent redistricting committees draw maps in anticipation of the midterm elections in 2022. The high stakes involved in drawing such maps typically lead to partisanship divides, making the entire process highly contentious and controversial.    

Gerrymandering will be the topic du jour in the news over the next several months, as each party in each state maneuvers to design congressional districts in a manner that will give their candidates the greatest opportunity to win. With Maryland, North Carolina and Pennsylvania ranked among the most gerrymandered states in the country, eyes will be on them and many others as they work through their mapping process. 

This begs the question, what is gerrymandering?  

Gerrymander is the process of drawing political maps to give one party an advantage in district elections. It effectively allows a political party to assign voters to districts in a manner that sways the outcomes of elections in their favor.  

Gerrymandering has been in existence even longer than the name. Elbridge Gerry, a signer of the Declaration of Independence and an early 19th-century governor of Massachusetts, designed a state Senate district in Essex County to give his party an advantage in the ensuing election. The Boston Gazette recognized his shenanigans, labeling it “gerrymandering,” a combination of Gerry’s name and salamander, the shape of the district that he drew. The name stuck. 

Packing and cracking are the tools of gerrymandering. Packing concentrates a large number of voters with similar political leanings into a small number of districts so that they win overwhelmingly in them but have no influence in other districts. Cracking breaks voters of similar political leanings into several different districts, effectively diluting their influence on election outcomes.   

The outcome of gerrymandering is that politicians pick their voters rather than voters picking their representatives, a process that is precisely backward in a democracy. Without elections that require candidates to work for their votes, elected representatives are likely less inclined to be responsive to the wishes and needs of their constituents. 

What is the solution? Transparency.   

The practical problem is that it is impossible to simply look at a map and assess whether it is gerrymandered, and to what degree.   

How then can one achieve transparency with map properties shrouded in darkness? 

Computational redistricting, which I work on developing, [AO1] is the path forward. Computer models and algorithms are available to both draw maps and evaluate them. Measures exist for assessing gerrymandering, such as efficiency gaps (a measure of “wasted” votes) and partisan asymmetry (a measure of how changes in vote share impact changes in districts won).  No one measure provides a silver bullet for exposing gerrymandering, but collectively, they represent a set of tools to shine a bright light on gerrymandered maps and the associated destructive mapping partisanship.   

The abundance of voting data from past voting records makes computational redistricting a powerful weapon to both combat and, perhaps surprisingly, facilitate gerrymandering.  Constraints such as population balance, compactness (districts are tightly drawn) and contiguity (districts are connected) can all be incorporated into the models to create districts that are visually appealing, but may still be gerrymandered.  In other words, the “eye test” may be blind to gerrymandered maps. Thus, the belief that computational redistricting alone can make the mapping process fair is an illusion.  

What computational redistricting provides is a 21st-century tool for mapping, making the process faster and easier.   

Analogously, the efficiency and effectiveness of engineering and science improved with the transition from slide rules to calculators to computers. However, the fundamental rules of computation did not change. The same applies to computational redistricting. It can facilitate the design of maps, but the resulting maps created are only as fair as the inputs and constraints incorporated into the models and algorithms.   

What computational redistricting does is create an environment for transparency, so that when politics taints the mapping process, its impact can be quantified and exposed. Such transparency will inform voters and politicians what they can expect from any proposed maps, and how gerrymandered they may be.   

Computational redistricting is the sword that can slay the gerrymander dragon. It can shine a bright light on the egregious intent of politicians who draw maps, and in many cases, expose how little choice voters are being given. That level of transparency may even spur some voters to rethink their votes, and provide some unexpected outcomes on Election Day that may even surprise the gerrymanderers. 

Sheldon H. Jacobson, Ph.D., is a professor of computer science at the University of Illinois at Urbana-Champaign. His research group on computational redistricting is committed to bringing transparency to the redistricting process using optimization algorithms and artificial intelligence.