Background on Gerrymandering
Under the American political system model the legislative, executive and judicial branches form the three corners of the government triad. They are the key institutions that provide for a balance of power. The federal and each state constitution provide for this system of checks and balances so as to prevent any branch from abusing its authority. The judiciary is an independent non-partisan branch. The executive branch is headed by the chief executive of the government who is elected statewide in the case of the governors, or nationally by the people via the Electoral College in the case of the President and Vice President of the United States.
The law-making authority is vested in a legislature which is composed of elected representatives chosen by districts at the state level. At the federal level, senators are elected statewide, each state being represented by two senators. Like the state legislatures, representatives are elected to Congress by districts of roughly equal population on a state-by-state basis. Through this model legislators are elected at a local level from districts of roughly equal population. This gives each citizen an equal voice and vote by way of their elected representatives.
In order to ensure that the votes and voices of the people being expressed through their elected representatives are equal, districts must be redrawn periodically as the population within a state and between the states changes. This is achieved through the process of reapportionment and redistricting. At the federal level congressional representation in the House is reapportioned by Congress every ten years following each federal decennial census. However the process of redistricting, the act of drawing new district boundaries is left entirely to the state legislatures. The state legislatures are also responsible for redistricting their own constituent districts as well.
The process of redistricting is therefore one of high political stakes. In each state the majority party in control of the individual houses of the legislature at the time of redistricting controls the process. The majority party therefore will enact legislative redistricting that favors its own party. This is commonly referred to as Gerrymandering. The term is derived from a redistricting law that was enacted by the Massachusetts legislature in 1812 and signed into law by then Governor Elbridge Gerry. The word is a portmanteau of the root words ‘Gerry’ for Governor Gerry, and ‘salamander’, an amphibian that a political cartoonist reckoned a senatorial district resembled. The term was meant as a derogatory critique of the party in power (Democratic-Republicans) that drew the districts. It has been used ever since then to describe in negative terms the process of redistricting that is egregious in nature from one’s political perspective.
Many factors play into the process of redistricting. It is a complicated process. Geography naturally plays a big role. So too do demographics as well as political patterns of voting. The process has evolved over time as census practices have become more sophisticated and technology has advanced. Initially legislators relied on simple geographic units such as counties and townships as district building blocks. By the mid 19th Century railroads began to have a huge impact on the growth of cities and settlement of the United States. The once small cities began to mushroom into large metropolises. It became necessary to draw geographically smaller districts in urban areas in order to maintain equality of district populations. Legislators began utilizing city wards as building blocks of districts in these situations. By the turn of the 20th Century detailed census enumeration techniques allowed districts to be drawn down to the city block level.
As the process of redistricting became more refined and sophisticated, so too did the practice of gerrymandering. Increasingly districts became much more complex and less compact in their shapes. By the 1960’s computers were being used extensively in census tabulation. These new tools were in turn utilized by legislators to aid in drawing new districts. By the 1990’s computer usage had become highly sophisticated, resulting in almost incomprehensible looking districts. Gerrymandering had risen to a level never before seen in American politics.
Some districts were so drawn in order for states to be in compliance with the 1965 Voting Rights Act. This law mandated that no substantial ethnic, racial or linguistic minority member be denied the right to vote for the candidate of his or her choice. This meant that districts had to be drawn in ways that did not overtly dilute the voting power of these minorities. The process was further complicated by state laws requiring that districts be of equal population and conform to compactness and contiguity standards. Often these legal objectives conflicted with one another.
These conflicting mandates resulted in litigation as redistricting plans routinely ended up in legal challenges. Since 1965 the courts, particularly federal courts have ruled consistently in favor of plans that result in the maximum number of minority districts possible, regardless of the resulting shape of districts. Advancements in computer technology have enabled legislators to sculpt districts with high precision in very short turnaround times. This capability has been applied to partisan districting as well.
The result of all these advancements has been highly gerrymandered legislative districts all across the nation. Legislators have become quite adept at designing districts that serve the legislators first and the electorate secondarily. Consequently the rate of retention of incumbents has hit historic highs. It is now very difficult to unseat an incumbent or to reverse party control of the legislatures. Since incumbents have such an advantage they attract far more political campaign contributions than do challengers. This in turn helps the incumbency become even further entrenched. Thus special interests with deep pockets have come to dominate legislative prerogatives. Consequently the legislative branch, the one that was originally designed to be the one closest to the people has become grossly insulated from the electorate.
Politicians, political scientists, attorneys, jurists and the electorate are well aware of the problems resulting from excessive and extreme partisan gerrymandering. Various solutions have been proposed such as requiring once again that districts conform to strict standards of compactness and contiguity, and preservation of the territorial integrity of existing municipalities and or counties where possible. The prohibition against using election data and residency of incumbents or potential challengers in drawing of district boundaries are other proposals. Perhaps the most important proposals have been in removing the legislators from the process entirely as they constitute a conflict of interest.
Regardless of the solutions proposed or even utilized, there is a genuine need for all interested parties to have a firm understanding of the process past and present. There is an acute need for people to be able to study what was done in the past in order to be able to comprehensively understand the process and how it can be improved. The several states have recorded the results of previous redistricting cycles dating back to the beginning of the process. However these earlier legislative district plans are not easily accessible in any comprehensive manner. There is no single repository of this information that is easily accessible. Furthermore very little of this information is in a format that can be utilized by modern analytical systems.
The US Census Bureau has been collecting and organizing digital geospatial versions of statewide legislative districts dating back to the early 1990’s. These have been compiled to aid state legislators and the public in the redistricting process. They are helpful in this regard but are limited in their scope. Students of political science would benefit greatly if a comprehensive database of such geographic features were readily available.
In 2012 a nationwide database of Historical Congressional Districts were released to the general public. This database was compiled by Jeffrey B. Lewis, Brandon DeVine, and Lincoln Pritcher with Kenneth C. Martis of the University of California Los Angeles (UCLA) working under a grant from the National Science Foundation. The features in the spatial files are in ESRI Shapefile format which allows them to be utilized in several off-the-shelf GIS (Geographic Information System) software applications. The districts mapped encompasses all congressional districts of the several states from 1789 to 2012. The boundaries used were derived from an earlier database of a nationwide collection of Historical County Boundaries developed by the National Historical Geographic Information System (NHGIS) and the Atlas of Historical County Boundaries developed and maintained by the Newberry Library. Lewis, DeVine and Pritcher were the principal researchers. Martis is the author of The Historical Atlas of United States Congressional Districts: 1789-1983. (New York: The Free Press, 1982) and served as a consultant on the project. His atlas was heavily used a source for the database.
The MapStory Gerrymandering Initiative: National Efforts
A group of MapStorytellers, including myself, became interested in this topic in 2016 when an important redistricting case was accepted by the Supreme Court of the United States. The case titled Gill v. Whitford, involves a challenge to the current Wisconsin Assembly District Plan currently in place and adopted by the legislature in 2011. It is a potential landmark case because the plaintiff is challenging the plan on the basis of partisan gerrymandering. If the high court rules in favor of the plaintiff it could result in the redrawing of most if not all state congressional and legislative districts across the nation. At a minimum it may very well have an effect on all future legislative plans in the United States.
In 2017 I started compiling a comprehensive StoryLayer of all US historical congressional districts. This would allow users to view an animation of all congressional districts throughout US History. It was also decided that this StoryLayer would include attribute information about partisan control of each district. Such attribution will allow the user to see not only the changing shapes of districts, but also the shifting control of districts by party affiliation. The data could also be used to analyze the compactness and contiguity ratings of each district.
The UCLA Historical Congressional Districts database was cleaned of errors and enhanced. Territorial features were assigned standard codes that were used to uniquely identify each state congressional district. START_DATE and END_DATE fields were added to coincide with the beginning and ending of each Congress. These fields were added so as to allow for the animation of each feature. The ID codes were then updated to conform to the start and end of each congress as well as when special events occurred such as the admission of a new state or territory.
The individual files were initially organized by each congress which span a period of two years. I began compiling them into a single master file. The simplest way of doing this was to append each file into a master. By the time the first century of congresses were compiled the master file had reached two gigabytes of data. ESRI’s ArcMap software was being used to do this work and the file reached the 2 Gigabyte limitation of the application. I had to come up with a way to reduce the file size. A solution was found by eliminating duplication and simplifying the district features.
I knew that states redistrict every ten years following each decennial census. Most districts were therefore unchanged for a period of ten years rather than just two. File size was reduced by using each feature for as long as it was valid rather than with each congress. The number of features were reduced by a factor of four. I also simplified many district polygons by extending coastal districts out to the three-mile maritime boundary of each state. By eliminating complex shorelines the overall size of the master file was reduced by a factor of ten. Using the maritime boundary made sense because it is the true legal boundary of the state and therefore each coastal district. Doing so had the added benefit of compiling spatial data that would provide improved metrics when calculating compactness and contiguity measures.
Nitin Gadia and Laurence Cramer, two other MapStorytellers, helped by sorting lists of former congress members in order to obtain party affiliation for each district. I am currently about halfway through the process of compiling the master file and expect to have that phase completed in a few weeks. Once that has been achieved the partisan data will be incorporated into the master file. At that point the master file will be uploaded as a StoryLayer. Then each state will be separated from the master file and uploaded as an individual StoryLayer. Future plans call for enhancing the data to include the names of congress members and compactness and contiguity scores.
The MapStory Gerrymandering Initiative: Wisconsin Efforts
Simultaneously and in parallel with the congressional database, I am compiling a similar dataset of the Wisconsin legislative districts. These files have the same file structure as the congressional district data. Currently I have successfully mapped the Wisconsin Assembly and Senatorial districts between 1983 to 2019. These represent the districts in existence for the 86th through 103rd Legislatures. The Assembly and Senatorial Districts StoryLayers are up and running and there is now a test MapStory for the Assembly Districts layer. Check it out:
I plan to extend this dataset to include all historical legislative districts for Wisconsin dating back to 1848 when it became a state, and if possible even further back to 1836 when Wisconsin Territory was organized. I am currently working in conjunction with Jonathan Marino on this project. Marino is the Director of the MapStory Foundation and currently resides in Madison, Wisconsin. He has already acquired valuable source information from the state library in Madison. We believe that the Wisconsin dataset will be useful in helping citizens understand the issues associated with the current case before the Supreme Court.
Help out! Get Involved
In the long term we would like to continue expanding this project to cover all fifty states at the state legislative district level. Once Wisconsin is completed I have an eye on New York State. Starting in 2010 I began mapping all the historical municipal boundaries of New York State. These data will be useful in mapping the historical state legislative districts since most counties were divided along town and city boundaries.
The MapStory Foundation has been a great help in supporting and investing in this project. We are looking for volunteers interested in taking on other states. Please message me if you’d like to get involved!
Historical Congressional Districts, UCLA dataset
Historical County Boundaries developed by the National Historical Geographic Information System (NHGIS)
Atlas of Historical County Boundaries developed and maintained by the Newberry Library.