The furor over the Supreme Court case considering limits on gerrymandering has reminded me of a conference I attended years ago. This was a meeting mostly composed of legal scholars focused on progressive reforms to the American judicial system. Gerrymandering, which is a hot topic today, wasn’t a major focus of the conversation. But we did get one presentation from a distinguished scholar of redistricting matters who explained that redistricting is a particularly hard problem to solve because there are a number of different goals that are mostly incompatible:
- It seems like a system should offer partisan fairness such that control of a legislature is typically in line with the population’s overall preference.
- Districts should align communities of interest and correspond in some sense to real places that we can characterize, like “the South Side of Chicago,” rather than just be arbitrary zones, like “some of the suburbs of San Antonio and some of the suburbs of Austin plus a big disconnected patch of rural Texas.”
- Racial minority groups should get fair representation. A state like Georgia that’s 30 percent black shouldn’t have an all-white congressional delegation.
- There should be fair-fight districts with real electoral competition, not just everyone segregated into safe seats that protect incumbents.
- Districts should be compact and look like a nice checkerboard of squares and triangles, rather than a bunch of crazy squiggles.
Sometimes you can make this all work together, but oftentimes you can’t. Creating majority-minority districts to ensure racial representation can look a lot like “packing” Democratic voters into lopsided seats. Aiming at fair fights sounds nice but will end up violating communities of interest. Aiming for partisan fairness will necessarily involve some odd squiggles, since neighborhood-level partisanship can be very disparate.
So I asked this scholar: “What about proportional representation?”
She said that when she teaches redistricting law, she does proportional representation last because it solves all the problems and the point of the class is for the students to work through the different complexities and legal doctrines governing the American system. That seems smart as a pedagogical approach, but as an agenda for political reform, solving all the problems is a good idea.
The many flavors of proportional representation
The basic idea of proportional representation is that instead of each place having a single representative selected by either plurality or a runoff system, you aggregate a bunch of people’s votes and then assign seats to parties in proportion to their popularity.
Proportional representation can work in a bunch of different ways, but there are two broadly popular schemes:
- In a party list system, voters simply mark a ballot for a party and then seats are allocated based on the share of the votes that party got. If the Cranky Oldsters party gets 25 percent of the vote, they get 25 percent of the seats, and if that works out to 40 seats, that means the first 40 people on their list get seats.
- In an alternative vote system (endorsed by conservative magazine National Review’s editorial, Reihan Salam), individual candidates still run for office, but instead of single-member districts, each state might be a big district with multiple members. In a state like Maryland with eight House members, voters would rank a bunch of candidates in order of preference, and then a formula would determine which eight people get the seat.
There are also innovative systems out there, like Jameson Quinn’s proposal for PLACE voting. The main downside of an AV system is that AV ballots are very complicated. A party list system, by contrast, is very simple, but some people don’t like the way it breaks the direct connection between voters and politicians. Germany and New Zealand use a system called mixed-member proportional representation that combines features of both to offer a simple ballot and personalized representation, with the main downside being that the under-the-hood details of how it works are complicated.
But while these different systems all have some pros and some cons, they all fundamentally solve the redistricting dilemma. The main way they do that is by simply making decisions about where the boundaries go much less relevant. Most states wouldn’t have any district boundaries at all. A big state like Texas or California might need to be sliced into three or four chunks, but because the outcomes are guaranteed to be proportional, the exact details of the chunking don’t matter very much.
Proportional representation solves many problems
The big upside to all flavors of proportionality — and the reason places hardly ever backslide from proportional to nonproportional systems — is that it basically solves all the map-drawing problems in one fell swoop.
Questions about representation for ethnic minority groups or communities of interest are taken care of exactly the way they should be — from the bottom up by voters and parties. If Latinos want to vote for fellow Latinos, then they will end up being represented by Latinos in proportion to their numbers. Or if Latino identity loses salience relative to other factors (ideology, geography, socioeconomic class, whatever) they won’t be.
Proportional representation also ensures that almost every district is in some sense a “fair fight.” Even if a state like Georgia has a pronounced lean toward the GOP or a state like Massachusetts has a strong lean in the other direction, the margin of victory ends up mattering a lot. And since the margin of victory matters, parties have incentives to try to communicate with, appeal to, and mobilize voters in every corner of the country. That helps boost participation and engagement, but also ensures that no incumbent can feel so “safe” in his seat that he doesn’t need to try to work hard, avoid scandal, and otherwise do his job.
Since most states wouldn’t need to be subdivided at all under PR, the “squiggly district” problem would be entirely eliminated except insofar as some states (looking at you, Maryland) are themselves squiggly. States that do need to be subdivided could be sliced in compact ways without sacrificing any other values. People who live in cities wouldn’t have their votes devalued due to some arbitrary notion of “clustering.” And nobody’s vote would be “wasted” — the views of Republicans in Los Angeles and liberals in West Virginia would count just as much as the views of anyone else.
Courts should think boldly about solutions
Until the 1962 Supreme Court case Baker v. Carr, it was commonplace for state legislature districts in the United States to contain wildly different numbers of people.
There’s no black-letter constitutional text requiring that state legislative districts be drawn with equal populations, and obviously the US Senate operates along exactly the opposite principle. So for the first 150-plus years of the republic, state legislatures apportioned themselves however they wanted to. This often meant leaving maps based on decades-old population data in place, deliberately overweighting rural areas, and other violations of democratic principles. But as the Warren Court began handing down civil rights rulings in the 1950s, pressure grew to confront states’ untrammeled boundary-drawing power.
Fifty-five years later, the ruling feels like common sense, but it was extraordinarily controversial at the time with Justice Felix Frankfurter and others arguing that for the Court to insert itself into the redistricting process violated the separation of powers. But having decided otherwise, the Court’s majority articulated a principle of “one person, one vote,” drawing on legal and political principles that could just as easily have been articulated as “all citizens’ votes should carry equal weight” — i.e., electoral systems should be proportional.
Not only does a proportionality mandate solve the problems of redistricting more elegantly, it solves the problem of judicial meddling with electoral mechanics much better. The Supreme Court fairly clearly doesn’t want to set itself up as a national district-drawing body that is perpetually being asked to decide whether a given setup counts as “gerrymandering” or natural “clustering” or whatever. Mathematical tools like the “efficiency gap” are a useful first step in quantifying partisan gerrymandering, but they still leave tons of other line-drawing goals in play. The judicial system’s experience with the effort to ban racial gerrymandering only to find itself sucked into a million debates over whether such-and-such a gerrymander is really about disenfranchising black people (which is bad) or just about screwing over Democrats (which is okay) is an unhappy precedent.
Requiring states to adopt proportional systems would be more disruptive in the short term, just as the requirement that districts be equal in population was disruptive. But the distinction between a proportional and a nonproportional system is fairly clear and doesn’t leave a ton left to argue over. It would solve the substantive problem of excessive partisan gerrymandering and the procedural problem of how to avoid excessive policing of partisan gerrymandering and solve the other dilemmas of redistricting at the same time.
Fixing everything in one fell swoop would make for a boring law school class, but as a policy outcome, it’s a pretty great idea.
