Elected officials are regularly tempted to exercise their power in ways that benefit themselves and their friends at the public expense. A good example is gerrymandering, the practice of drawing district lines to help ensure a desired result in future elections. Both parties do it, and the practice dates far back in history.
Gerrymandering often results in strangely shaped political districts in which it is very difficult for voters to unseat incumbent politicians.
In a classic single-party gerrymander, the party in power packs opposition voters densely into as few districts as possible, thus enabling its own voters to lead by a comfortable margin in a maximum of districts. When a legislature is under split party control, the theme is often bipartisan connivance: you protect your incumbents and we’ll protect ours. Third-party and independent voters, as is so common in our system, have no one looking out for their interests.
Some other strategic political purposes can be served by gerrymandering: Weak incumbents can be spared scrutiny of their performance by assigning
them tracts that fall short of being coherent political communities,
perhaps combined slivers of multiple metropolitan areas with little in common. It’s expensive and time-consuming for a challenger to campaign or advertise against an incumbent in such a district. Party bosses can also punish their own party’s lawmakers for being too independent-minded by drawing them unfavorable districts.
The process feeds apathy. Residents who have not even figured out which district they are in are less likely to keep track of how well their representative is serving their interests.

The Constitutional Background

Our Constitution puts states in charge of apportioning their own legislatures, while dividing the corresponding power as to congressional districts between them and Congress.[1] The Supreme Court’s one-person-one-vote rulings require equal or nearly so population in districts within a state. The Voting Rights Act of 1965, following the Equal Protection Clause, bans districting done for a racially discriminatory purpose, which adds a sometimes-complex overlay of requirements.
Although the Supreme Court has been urged to ban politically motivated gerrymandering, it has thus far declined to do so. Its rationale: it could identify no principled and objective standard to apply that would not draw it into a multitude of complicated local disputes.
Fortunately, ideas for reforming gerrymandering are many. They fall into two main categories:

  1. Rules on who is responsible for drawing district lines
  2. Rules directing the shape or extent of districts

Who should draw the lines?

One of the ideas that recurs most frequently is to make the process bipartisan, or at least avoid empaneling a majority of loyalists from a single party. The second largest party thus winds up in a negotiating position, perhaps with one or more neutrals or tie-breaking votes in between.
A newer trend, which has caught on especially in Western states in recent years, is to entrust redistricting to a more fully independent commission of citizens not holding office. Elected officials themselves, their families, and political pros are frequently excluded.
In a category of its own is the system used in Iowa (as well as many countries outside the U.S.). It assigns redistricting to the same nonpartisan civil service staff that provides legislative services such as bill analysis at the capitol. Although Iowa’s system is often praised for its fair results, it may owe some of that success to features of the local political scene not replicated everywhere. For example, Iowa has a fairly even party balance and a legislative staff whose nonpartisan bona fides are accepted by lawmakers of both parties.
Under any of these systems, the law can go further by prescribing the powerful step of “blinding” the line-drawers to politics – that is, directing them not to consider such factors as current party registration, past voting records, or the residence of any individual, such as an incumbent.

What should districts look like?

The most essential task in reform is to provide clear and objective rules for governing how districts are drawn. The three most widely accepted standards are as follows:
All parts of a district should touch. Although this seems obvious, careful language helps prevent such tricks as circuitous connections over water.
Intuition tells us the difference between a district shaped like a turtle and one shaped like a tapeworm. But trusting to intuition is not necessary: at least two mathematical measures of compactness are widely employed. Colorado’s constitution prescribes the “total-perimeter test”: “Each district shall be as compact in area as possible and the aggregate linear distance of all district boundaries shall be as short as possible.” Other states use a “radius” or “length/width” test.

ntuition tells us the difference between a district shaped like a turtle and one shaped like a tapeworm.”]
Where possible, districts should respect the boundaries of smaller political subdivisions, such as counties and towns. One convenient measure of congruence is the number of county or town splits in a plan, with lower numbers ordinarily better.
Other criteria are sometimes prescribed, but if too many are introduced, and if the commission is given latitude to balance among them, then a dangerous degree of discretion is reintroduced into the process.

The Role of Technology

Technologically, gerrymandering is a bit of an arms race. Politicians with access to so-called big data can now efficiently sort voters down to precincts, city blocks, and even buildings. That is why the problem will get worse absent correction. Yet quantitative methods hold out hope for the reform side as well, and not only by providing objective, replicable measures of goals like compactness.
Geographic information systems (GIS) methods now allow members of the public using inexpensive software to analyze the full data set behind a map. In several states, that has meant members of the public could offer maps of their own or make well-informed critiques of legislators’ proposed maps. In one triumph for citizen data use, the Pennsylvania Supreme Court invalidated a map drawn by lawmakers as clearly inferior to a map that had been submitted independently by an Allentown piano teacher.
Redistricting reform makes sense as a safeguard against the entrenchment and insulation of a permanent political class. Voters should choose legislators, not the other way around.

[1] (Article I, Section 4 of the Constitution: “The Times, Places and Manner of holding Elections for Senators and Representatives shall be prescribed in each State by the Legislature thereof; but the Congress may at any time by Law make or alter such Regulations.”)