By Roy A. Minet. Op-ed originally published by lancasteronline.com 4 April 2018.
Fixing the deplorable gerrymandering rampant for decades has garnered much attention recently.
A congressional map drawn by the state Supreme Court is in place for the 2018 midterm elections. But the map will be redrawn after the 2020 census. If we're to end, or at least diminish, gerrymandering, impartiality will be required.
Impartially defining electoral districts need not be difficult, overly complicated, time-consuming or contentious. As always, the first step to identifying the best solution is to correctly identify the requirements and clearly write them down in order of importance:
1. One person, one vote: It is a hard requirement that each of multiple districts must contain, as nearly as is reasonably possible, the same number of eligible electors. This is an obvious good thing, and the U.S. Supreme Court has decreed it.
2. Impartiality: The process by which districts are determined must not give any systematic advantage or disadvantage to any group or faction.
3. Understandability: The process by which districts are drawn should be understandable by a reasonably bright high school student. (In fact, anything having to do with voting and elections should meet this requirement.)
4. Verifiability: It should be possible for citizens or organizations to independently verify that districts are correctly drawn.
5. Process must be well-defined and stable: It should be clearly and publicly spelled out. It should not be changeable on a whim or when different people are implementing it. Enshrining the process in the state or even U.S. Constitution would be a good idea.
6. Precinct atomicity must be preserved: Precincts are very small areas of roughly 1,000 voters that are determined locally based on available polling places and their proximity to voters. It is unnecessarily disruptive if redistricting requires redrawing precinct boundaries. Therefore, each precinct should be entirely contained within a single district. (If precincts straddling a district boundary should need to be merged, the merged precinct lands in the district from which most of its voters came until the next redistricting is done.)
7. Contiguity: It is usually required that voting districts be geographically contiguous, and that no district be completely contained within any other district. This supports requirements 3 and 4.
8. Compactness: This can have several definitions. Fortunately, it is not critically important. Compactness does make it easier for candidates who must repeatedly traverse the district for campaign purposes and easier for elected representatives to commune with constituents. Primarily, it supports requirements 3 and 4.
One thing not in the requirements is "fairness." People sling the word "fair" around all the time, but the criteria by which they judge fairness can vary radically. Without understanding the specific criteria, the word is meaningless.
Seriously considered solutions seem to revolve around establishing an unbiased commission that figures out how to draw boundaries. No individual is unbiased, so what that means is a commission on which it is hoped opposing factions keep each other in check. A commission doesn't guarantee requirement 2 and definitely doesn't satisfy 3, 4 and 5.
A superior approach is to define a procedure that satisfies all requirements. It doesn't matter who (or what) executes the procedure, the same impartial boundaries are the result. A procedure that well satisfies all requirements (except 6) has already been defined. It is called "splitline."
The splitline procedure very simply divides a state into two sections having the desired populations using the shortest possible line. If more than two districts are needed, the process is repeated (as many times as necessary) on one or both of the two sections until the desired number of equal population districts has been drawn. Maps are viewable online that show the splitline congressional districts for each state.
In order to meet requirement 6, I recommend changing "the shortest possible line" of the splitline method to "the shortest distance along precinct boundaries." [EDITOR'S NOTE: this change may introduce enough problems that it is not worth it,; question discussed on the splitlining variants page.] Because of precinct granularity, this will introduce small errors in population (inconsequential for large districts, perhaps 1 percent for very small districts containing only 25 or 30 precincts).
Splitline districts are always contiguous and maximally compact (geometrically). They are based only on the boundaries and populations of precincts; no voting history or registration data are used. The procedure is easy to understand. If you're familiar with the state's population distribution, you can see that the lines have to be pretty much correct by just looking at them on a map. Lots of individuals and organizations are capable of independently verifying the boundaries. Also, it should be obvious that splitline can be done in minutes by a computer at near-zero cost.
There will be two main objections. First, splitline is necessarily going to ignore geographic features, and it will divide cities and counties. That's not an actual problem. As proof, we've lived just fine for decades with many such divisions caused by gerrymandering. Some such divisions are unavoidable with any method.
The second complaint will be that some faction or another doesn't receive "fair" representation. Whether a real or imagined issue, it is not something that can be solved by adjusting district boundaries – wrong mechanism. Remedies that might be considered are multiple-representative districts, ranked-choice voting (other than instant-runoff voting) and proportional representation. But that's a whole other column.
Roy Minet, a Mount Joy resident, was the Libertarian Party's nominee for Pennsylvania auditor general in 2016.