Gerrymandering Is a Useless Concept. Redistricting Is a Choice Among Values.

Originally published in the Mischiefs of Faction blog.

By Jonathan M. Ladd

Redistricting is a perennial topic of political debate in the United States, and for good reason. House districts have important consequences. But the debate often bogs down over which maps are "gerrymandered" or "nonpartisan." I would argue that, while it has historical roots, gerrymandering is a useless concept in modern American politics. It confuses much more than in clarifies because there is no common agreement on a definition of gerrymandering. As a result, people end up talking past each other.

When thinking about districts, I always start with the six possible goals that David Butler and Bruce Cain list in their classic book, Congressional Redistricting:

1. Equal Population
2. Matching Natural and Political Boundaries
3. Compactness and Contiguity
4. Party Fairness (an unbiased seats-to-votes curve between the parties)
5. Ethnic Fairness (substantial numbers of minority ethnic group members elected)
6. Party Competition (close elections and party alternation in a substantial number of districts)

A central point Butler and Cain make is that it is generally impossible to achieve all these goals at once. In their words:
"...the problems involved in securing equitable solutions to reapportionment puzzles are complex because, at almost every point, principles or goals that everyone agrees to be desirable are in conflict with each other" (p. 65)

The original 1812 gerrymander
Gerrymandering was first used to describe a Massachusetts state senate district drawn in 1812 under Governor (and founding father) Elbridge Gerry to aid Gerry's Democratic-Republican Party. Because the district resembled a salamander, the Boston Gazette labeled it a gerrymander. Since that point, it has been all downhill in maintaining a widely accepted definition of gerrymandering.

The problem is that most contemporary people define a gerrymander as a deviation from the values that they wish districts would promote. But people disagree about which values those are. In Baker v. Carr (1962), the Supreme Court ruled that districts must have equal populations. So all plans now start by fulfilling goal #1, but people disagree about which of the five other goals maps should try to achieve. Whenever a map doesn't achieve their preferred goal, they call it gerrymandered. If two political writers are arguing about whether a map is gerrymandered, and one's goal is compactness and the other's goal is party fairness, they will talk past each other. Since people disagree about redistricting's goals, and thus what gerrymandering means, it is better to set aside the "gerrymandering" concept and just specifically discuss the values different maps emphasize or neglect.



For instance, currently Democratic voters are more prevalent in densely populated urban areas, while Republicans are more spread out in suburban and rural regions. Maps that maximize compactness are very advantageous to the Republican Party and thus lack party fairness. With all the Democrats packed into urban districts, many of their votes are wasted to produce excessive victory margins. Republicans are more spread out and thus win by smaller margins in a larger number of suburban and rural districts. On the other hand, this could be easily remedied by drawing long districts that combined urban, suburban, and even rural areas, so that fewer Democratic votes are wasted. But by improving party fairness, this would greatly reduce compactness.

By at least some plausibly important criteria, any plan will fall short. No statistical or computer technique, nor any nonpartisan commission, can avoid making a choice among goals. Relying on these methods rather than the legislative process to draw lines only gives the nonpartisan commissioners or those designing the computer algorithm the power to make this trade-off, whether they realize it or not. How we trade-off among goals is a moral and political choice, not a technical challenge.


Finally, this makes clear how redistricting can have important consequences. Political science has shown fairly conclusively is that redistricting is not a major cause of the growth in congressional polarization in the past several decades. The simplest way to see this is consider that polarization has increased almost as much in the Senate, with obviously fixed districts, as in the House. But just because it didn't cause our polarization, doesn't mean redistricting doesn't have consequences.

The tradeoffs involved in redistricting can still have big political effects. As discussed above, you could draw districts that favored Republicans by making them geographically compact and packing lots of Democrats into urban districts. This sacrifices party fairness, but achieves compactness and even allows for ethnic fairness, because you get some districts with high percentages of ethnic minorities. After the 2010 census, Republican state legislatures drew districts in just this way, producing a Republican bias in the national seats-to-votes curve. It is also theoretically possible to draw lines that pack Republicans into a small number of districts where many of their votes are wasted and spread out the Democratic vote. You would just have to sacrifice compactness and ethnic fairness.

Competitiveness versus incumbent protection is also a choice. You could draw lines so that most districts have a close to even partisan division with only a few safe seats. Or you could make almost all seats safe for long-term incumbents with just a handful of closely divided swing districts. But you couldn't have both, and to get either extreme, you would probably have to sacrifice compactness. Finally, if you wanted, you could even maximize party fairness by drawing lines so whichever party had the majority of votes was highly likely to win a majority of seats. But you would likely have to sacrifice compactness and ethnic fairness to do that.

But none of these possibilities is indisputably "gerrymandered" or "not gerrymandered." Nor is any indisputably "nonpartisan." These are all just different tradeoffs among many reasonable goals. You can achieve a lot through redistricting, just not everything at once.