During my sophomore year of high school, I had the opportunity to engage in a three week intensive course which examined the effects of Gerrymandering in United States government. In this course, we learned that throughout American history, Congressional district map makers have strategically used voting information to create congressional district maps that favor a particular party, or in some cases, even promote hateful ideology. One of these heinous gerrymanders actually occurred in my hometown of Chagrin Falls Ohio, where mapmakers during the 1950’s strategically drew map lines to exclude a predominantly black community from the Chagrin Falls school/ voting districts (this may also be classified as an example of “redlining”, but because school districts also have a direct effect on what is on a voting ballot, it also may be considered an example of “gerrymandering”). I realized during this course that gerrymandering poses a grave threat to the integrity of United States elections. Unfortunately, the United States constitution gives little insight into what constitutes a “fair” district, and many state legislators struggle to create legislation that properly addresses how to identify a “gerrymander”, and how to avoid them when creating maps.
How can I employ Game Theory to better understand and identify the tactics used by mapmakers that create “Gerrymandered” Maps? Additionally, how can these game theory models inform those working to create legislation that can assist in preventing Partisan Gerrymandering?
A Case Study:
In order to better understand how Gerrymandering actually works, I have provided two examples that illustrate both a “Republican” and a “Democratic” gerrymander.
According to the 2020 CNN Election Results, Maryland voters elected 7 Democratic candidates to the House of Representatives, and 1 Republican candidate to the House of Representatives (2020 House Election Results). Although Maryland is a strong Democratic state (in the 2020 presidential election 65% of voters voted for the Democratic candidate), Maryland elected 87.5% Democratic candidates to the House of Representatives (2020 House Election Results). Similarly, in Ohio, voters elected 12 Republican candidates to the House of Representatives, and 4 Democrat candidates to the House of Representatives (2020 House Election Results). Although Ohio is considered my many to either be a swing state or lean Republican state (in the 2020 presidential election 53% of voters voted for the Republican candidate), Ohio nonetheless elected 75% Republican candidates to the House of Representatives (2020 House Election Results).
Despite both Maryland and Ohio having clear support for both parties within their states, their Congressional elections make it seem as if Ohio is a 75% Republican state, and Maryland an 87.5% Democratic state (2020 House Election Results). However, presidential election data reveals that this is clearly not the case. So why does this happen?
The answer to this dilemma- Gerrymandering.
Gerrymandering occurs when map makers strategically draw Congressional district maps to favor a particular political party. Here are some examples of what these Gerrymanders look like in both Maryland and Ohio.
Ohio- 16 Seats, 12 Republican, 4 Democrat:
Maryland- 8 Seats, 7 Democrat, 1 Republican:
- One person One Vote- Making sure that a voter in Ohio has the same say in Congressional Elections as a voter in California, Maryland, Kentucky, or any of the other 49 US states
- Many people are unaware that local and state legislators are typically the ones behind these Gerrymanders, not federal legislators
- Packing and Cracking:
- Map makers will pack the opposing voters into as few districts as possible and give their party a slight lead in the remaining districts
- Contiguous Districts to the Extreme:
- For many states, the only requirement for a Congressional district is that they be “contiguous”. There cannot be districts where one part of the district is not connected to another. Some map makers will create districts that are not compact, but are connected to each other to maximize their chances of winning
- Nonvoting Residents:
- Most states require that each Congressional district have the exact same population. In order to achieve a partisan Gerrymander without adding opposing voters, a map maker might add voters into their district that do not typically vote, or who are not allowed to vote. For example, if a Democratic map maker in Maryland did not want to add any new Republicans to a Democratic district, they could add a prison to increase the population of a particular district, but because prisoners cannot vote, they would not have any impact on the election
- The Federal Government should not step in and create new laws unless they absolutely have to
- Partisan Gerrymanders violate the 14th and 15th Amendments, and any constitutional violation should be addressed by the federal government if states fail to uphold those amendments
- Any legislation should be reactionary, and not make drastic changes to the constitution
A map maker wants to know if they should “pack and crack” the opposing party, or create fair districts in a state with 16 congressional representatives. Although the state is split almost 50/50 in the presidential election the map maker believes that if they employ packing/ cracking, they can win up to 12 districts. However, the map maker also recognizes that if there is a higher voter turnout for the opposing party than expected, it could potentially hurt them (as the 12 “lean” districts all of a sudden turn into extremely competitive districts).
Player 1- The Map Maker
Strategy A- The Map Maker decides to pack/ crack
Strategy B- The Map Maker does not pack/ crack, and creates 16 fair districts (8 that are lean one way, and 8 that lean another way)
Player 2- Opposing Voters
Strategy C- Higher than expected voter turnout
Strategy D- Normal voter turnout
A,C (The map maker uses packing/ cracking, the voter turnout is higher than expected)
This is the most dangerous scenario for the map maker. If the opposition party has a higher turnout than expected, the “lean” districts could flip for the opposition party.
A,D (The map maker uses packing/ cracking, the voter turnout is normal)
This is the best scenario for the map maker, as their strategy of “packing” works perfectly. They just barely win in 12 districts (as the other 4 districts have mostly voters from the opposite parties “packed” into them).
B,C (The map maker does not pack/ crack, the voter turnout is higher than expected)
If the mapmaker decides not to pack and crack and the voter turnout is higher than expected, a map maker can expect evenly distributed districts (8 and 8).
B,D (The map maker does not pack/ crack, the voter turnout is lower than expected)
If the mapmaker decides not to pack/ crack, and the voter turnout is normal, this will likely help the mapmaker (as they are in a state which geographically favors their party, and a normal turnout will likely help them go over the top).
SOLVING THE GAME:
After using movement diagrams to solve this non-zero sum game, it is revealed that there is no initial movement diagram solution to the game.
Instead, in order to find a solution, I first employed the Nash Equilibrium method:
Map Maker plays A 2/7 of the time, and B 5/7 of the time
Voters play C 2/3 of the time, and D 1/3 of the time
Using Nash Equilibrium, it can be concluded that the most common outcome of the “packing and cracking model” is for the map maker to create fair districts a majority of the time, and for the voters to have a higher turnout than expected a majority of the time. If this game is played multiple times, the most common outcome will likely be 16 competitive districts (8 each) with the map maker opting not to “pack and crack”.
If the mathematically “safest” result is for the mapmaker to avoid packing and cracking, then why is it an all too common practice among map makers? For starters, lets examine some things the above model might have left out:
- Player 2 (the voters) actually do not have much of a direct say over whether turnout is “high”, “low”, or “normal”. Because Player 2 is reflective of thousands of voters, they cannot simply decide to have a “high” turnout whenever they would like.
- As Ms. Scott eluded to, other factors (such as voter suppression or failing to uphold the 14th and 15th amendments) can impact voter turnout.
- Although there are associated risks with Packing and Cracking, there is much less of a risk in employing this strategy today compared to twenty years ago. Whereas predicting voter turnout might have been extremely difficult twenty years ago, advanced data analysis technology makes predicting voter turnout, along with how people will likely vote much easier for map makers. There is much less of a “risk” today than this model likely reflects.
Ultimately, although this model might have some flaws, it does indicate one thing- if there is no external factors that make “packing and cracking” a “safe” option for a map maker (such as voter suppression), than the map maker will likely opt to create “safer” districts in order to avoid the major risks associated with “packing and cracking”.
Model 2- Map Makers Under Pressure
To better understand why map makers successfully pass “Gerrymandered maps”, I used a backward induction Game Theory model to analyze the process a map maker may go through when deciding how to approach creating their maps.
A map maker is tasked by Political Party A with creating a map that will win them 12 congressional district seats out of the 16 available seats (even though the state is much closer to 50/50). Here are some of the things the map maker must consider when creating the map:
- If the map maker fails to deliver 12 seats, Party A (which controls the state legislator) may never hire them again.
- If the map maker creates districts that are not compact, the map maker/ Party A may be sued outside organizations (example- they look like Maryland District 4)
- If the map maker fails to convince Party B that the map is fair, the map maker/ Party A might be sued by outside organization
- If there is clear evidence of “packing and cracking”, the map maker might be sued by outside organizations
A map maker must create a map that abides as much of the above criteria as possible, while also delivering 12 seats to Party A, and convincing Party B that the map is competitive. The map maker wants the map to pass in the state legislator, which requires a majority of Party A (and some members of Party B) to support it.
There are three players in this game:
Player 1- The Map Maker
Player 2- Party A
Player 3- Party B
Typically, the map maker is the first person to “make a move” (as they are the ones proposing the map to Party A and Party B). Then, Party A (the majority party) will then “approve” or “disapprove” the map (if Party A does not believe the map is going to benefit them, then they will likely tell the mapmaker to redraw the map, or fire the mapmaker). Finally, Party B is usually the final party to have a say, as they are usually the deciding factor in if the map is approved in the state legislator.
A map maker wants to create a map that will pass in the state legislator and please Party A so they can keep their job as the map maker. Party A hopes to maximize the amount of seats they are awarded (they believe they can earn up to 12 of the available 16). Finally, Party B wants to pass the map and is willing to settle for 6.
If the map maker proposes a map that earns Party A 12 seats, they earn 6 points. If the map maker proposes a map that earns Party A 10 seats, they earn 10 points. However, if Party A vetoes their map and decides to go with another map, the map maker loses 5 points.
If Party A approves a map that earns them 12 seats, they earn 10 points. If Party A approves a map that earns them 10 seats, they earn 5 points.
If Party B approves a map that earns them 4 seats, they earn 0 points. If Party B approves a map that earns them 6 seats, they earn 6 points. If party B votes for a map that Party A did not recommend, they lose 10 points (as the map will likely go to a revote which will not be as favorable for Party B).
THE ACTUAL GAME:
(Payoffs as follows- Map Maker/ Party A/ Party B)
Outcome- 6 Points Map Maker, 10 Points Party A, 0 Points Party B
Ultimately, the result of the above Backward Induction chart reveals that the map maker(s) and Party A (the majority party) have the most say over the final Congressional District Map. Because the map makers are for the most part, employed by the party in power (in this case Party A), they are most likely to create a map that favors benefiting the party in power, as opposed to compromising. While compromising might be the most beneficial outcome for the map maker, the risk of losing favorability in the party (or the in-game equivalent of losing points) far outweighs the prospect of giving Party A what they want. Then, when Party A inevitably decides to support the map that gains them the most congressional seats, the vote is then passed on to Party B. Party B must decide whether they vote for what they believe (a 10 seat map), and start the map making process all over again, or compromising and by giving Party A 12 seats. Unfortunately for Party B, if the process starts over again, they are likely to have less time to compromise, and risk having to vote for a map that might be even more “Gerrymandered” than before.
The Backward Induction Model reveals that the party in power, along with the map maker(s) have an extraordinary amount of power in drawing Congressional District Maps. Because these rules apply to both federal and state congressional maps, the maps that are approved are often favorable to the state legislators that are members of the party in power, thus allowing them to continue to remain in power, and draw/ pass gerrymandered maps.
After using Game Theory to analyze the process behind Gerrymandering, I drew two major conclusions:
- If there is no “voter suppression”/ “external factors” that influence the results of an election in a state, then there is less of a chance that a map maker will “gerrymander” a Congressional District Map.
- Map makers and the current party in power within a state have an institutional advantage that allows them to draw and pass gerrymandered maps.
When reflecting upon Ms. Scott’s recommendation to pass “preventative” legation as opposed to creating new legislation, it becomes abundantly clear that the easiest way to prevent Gerrymandering has little to do with actually creating new Gerrymandering legislation. Instead, making sure that the 14th and 15th amendments are upheld, and ensuring that the principal of “one person, one vote” is upheld is the best way to prevent Gerrymandering. Although map makers/ the party in power have an institutional advantage, that institutional advantage weakens when the integrity of the election is upheld.
Following the 2020 Presidential Election, a large nationwide debate ensued between Republicans and Democrats regarding the integrity of United States’ Elections. In response to this debate, some states (such as Georgia) have either considered, or actually passed legislation that aims to revise the procedures behind statewide elections (Cox). Unfortunately, many have criticized these laws for creating new voting inequities that will hinder the ability of certain groups to have their voice properly represented in the Untied States Government. If we are to properly address the issue of Gerrymandering in the United States, we must first carefully reexamine current and past state legislation that impacts voting equity. To prevent Gerrymandering, we must ask ourselves if our laws truly uphold “one person, one vote”.
“2020 House Election Results.” CNN, Cable News Network, www.cnn.com/election/2020/results/house.
Black, Shana. “Chagrin Falls Park, Black Neighborhood Next to Chagrin Falls, at Heart of Last Week’s Protest Controversy.” Cleveland Scene, Cleveland Scene, 19 Jan. 2021, www.clevescene.com/scene-and-heard/archives/2020/06/09/chagrin-falls-park-black-neighborhood-next-to-chagrin-falls-at-heart-of-last-weeks-protest-controversy.
Cox, Chelsey. “Georgia Voting Law Explained: Here’s What to Know about the State’s New Election Rules.” USA Today, Gannett Satellite Information Network, 12 Apr. 2021, www.usatoday.com/story/news/politics/2021/04/10/georgia-new-voting-law-explained/7133587002/.
Ingles, Jo. “Federal Court Says Ohio’s Congressional Map Is Unconstitutional.” The Statehouse News Bureau, www.statenews.org/post/federal-court-says-ohios-congressional-map-unconstitutional.
Ingraham, Christopher. “Analysis | How Maryland Democrats Pulled off Their Aggressive Gerrymander.” The Washington Post, WP Company, 27 Apr. 2019, www.washingtonpost.com/news/wonk/wp/2018/03/28/how-maryland-democrats-pulled-off-their-aggressive-gerrymander/.
Ricci, David, and Katie Scott. “Gerrymandering Empathy Interview .”
“State-by-State Redistricting Procedures.” Ballotpedia, ballotpedia.org/State-by-state_redistricting_procedures.
“Who Draws the Maps? Legislative and Congressional Redistricting.” Brennan Center for Justice, www.brennancenter.org/our-work/research-reports/who-draws-maps-legislative-and-congressional-redistricting.