Democracy Blog Essay

Redistricting Reform and the 2018 Elections 

This November, voters in four statesMichigan, Missouri, Colorado, and Utah – will vote on ballot measures that would create independent commissions or other non-partisan processes to draw congressional or state legislative districts, as Josh Douglas and Rick Hasen have noted in previous posts in this series.  At the same time, litigation challenging partisan gerrymandering as unconstitutional continues in several states, most prominently in a case from North Carolina, where the three-judge federal court struck down nearly all the state’s congressional districts (full disclosure:  I am part of the legal team currently defending that judgment before the United States Supreme Court).

Underlying the entire issue of redistricting is the question of what constitutes a fair map.  The issue is not just whether voters will endorse independent redistricting commissions, but what substantive criteria those commissions will be instructed to use.  This is not as straightforward a question as many people intuitively think.  As a conceptual matter, there are two fundamentally different approaches to answering that question, and it is worth noticing the distinct approaches these various ballot measures embrace.

The first approach is what I have called a process-oriented one.  Redistricters would be instructed to take a set of criteria into account that reflect appropriate democratic values for designing districts, but that do not include partisan political considerations.  Thus, in “partisan-blind” redistricting, districts would be designed to meet standards like equal population; compliance with the Voting Rights Act; keeping pre-existing political units (like towns, cities, and counties) together, to the extent possible; respecting communities of interest; and keeping districts reasonable compact and contiguous.  Under this approach, a “fair” map would reflect these criteria, and whatever partisan political consequences resulted from such a process would not affect the fairness of the plan.

The second approach is instead focused on exactly those consequences:  it is an outcome-oriented approach that would seek to ensure a map is “fair” in the sense that the likely outcome of elections under the plan would be that each political party would end up with roughly the same percentage of seats as the percentage of votes it received.  Under this approach, many of the values listed above would be of secondary importance to the primary goal of seeking to design a map that would generate partisan outcomes that match each party’s share of the overall statewide vote.   The “process-oriented” and “outcome oriented” approaches define two poles of the spectrum; one can imagine approaches that try to merge these approaches, though doing so risks becoming a muddle that leaves redistricters with a great deal of discretion unless the proper tradeoffs between these different objectives are identified with precision.  But having these two distinct conceptual frameworks in mind is helpful in evaluating the different ballot initiatives.

The Missouri measure, for example, which affects only state (not congressional) legislative districts, endorses the most extreme outcome-oriented approach of any statute or ballot measure I’ve seen.  This measure provides standards that a non-partisan demographer, who designs the maps in the first instance, is required to use.  After specifying that maps should be based on equal population, comply with federal law, and meet certain racial fairness requirements, the initiative then states that the demographer  is to do the following:

Districts shall be designed in a manner that achieves both partisan fairness and, secondarily, competitiveness. Partisan fairness means that parties shall be able to translate their popular support into legislative representation with approximately equal efficiency. Competitiveness means that parties’ legislative representation shall be substantially and similarly responsive to shifts in the electorate’s preferences. . . .

In any plan of apportionment and map of the proposed districts submitted to the respective apportionment commission, the non-partisan state demographer shall ensure the difference between the two parties’ total wasted votes, divided by the total votes cast for the two parties, is as close to zero as practicable.

The initiative, which proceeds to enact a type of efficiency gap test, requires that these outcome-oriented objectives take priority over making sure that districts are contiguous, compact, or keep towns and other units together (I don’t actually believe a demographer would in fact ignore contiguity for the purposes of ensuring fair partisan outcomes, but that’s how the initiative reads.)

Meanwhile, the Utah initiative appears to take a process-oriented approach, or perhaps to try to merge process and outcome approaches, but in a way likely to foster confusion.  Thus, the commission is told that it must meet, to the greatest extent practical and in a specific order of priority, seven process-oriented criteria, which include the usual factors, along with ones such as “following natural and geographic features, boundaries, and barriers.”  But then, after this specific list of rank-ordered priorities that must be followed to the greatest extent practical, the next provision tells the commission not to “divide districts in a manner that purposefully or unduly favors or disfavors … any political party.”  Utah’s commission would be only advisory. The legislature can reject the commission’s map and enact its own plan. However, it would be bound by the same criteria as the commission and would have to issue a report explaining why its plan did better at satisfying these criteria.

To tell the commission to follow process-based criteria and not engage in purposeful partisan gerrymandering is one thing.  Both of those objectives can be accomplished together.  But what does it mean to tell it both to follow these criteria and also not to “unduly favor or disfavor” a political party?   The “unduly favor” standard sounds like an effects-based or outcomes-based measure that applies even if the commission acts without any partisan intent.  So if the commission first follows all the process-based criteria, but the map that results would then favor one party or the other – in terms of how many seats that party gets compared to its statewide vote – what is the commission supposed to do?  Sacrifice some of the process criteria?  Or not pay attention to the outcomes after all?  If the process-oriented criteria are supposed to have priority, how much room does that leave to adjust the map to avoid “unduly” favoring or disfavoring a political party?

Michigan’s initiative, which applies to both Congress and state districts, is similar to Utah’s in trying to blend process and outcome criteria, but with a different set of priorities.   Thus, Michigan’s proposal first requires compliance with federal law, contiguity, and then respect for communities of interest.  But the next requirement, in order of priority, is that the “districts shall not provide a disproportionate advantage to any political party,” using “accepted measures of partisan fairness.”  Only after that requirement is met is the commission then required to reflect pre-existing boundaries of towns, cities, and counties and to be reasonably compact.

Because the commission’s plans cannot provide a disproportionate advantage to a party, this requirement goes beyond ensuring that the commission not act with partisan intent.  The effects of its plan must still not advantage any party.  In addition, since this requirement is given priority over drawing compact districts or keeping towns and the like together, the Michigan initiative might be read to require that districters use bizarrely shaped districts and break up towns, cities, and counties whenever necessary to ensure that the map produces fair partisan outcomes, in the sense that a party’s proportion of seats corresponds to its proportion of statewide votes.

Finally, Colorado‘s dual initiatives (one for Congress and one for state districts), take another approach  altogether.  They also try to blend process and outcome approaches, but the outcomes that the commission is required to pursue are competitive elections, not partisan fairness.  By their terms, these initiatives start with the standard requirements of population equality and compliance with federal law, then require preservation of communities of interest and compact districts.  After that, the initiatives require that map drawers, “to the extent possible, maximize the number of politically competitive districts.”

There can be a significant difference between maximizing competitiveness and maximizing “fair” partisan outcomes.  A plan in which all districts are designed to be 53% Party A voters and 47% Party B voters (based on past voting patterns) would be one in which all districts are competitive.  But it might be that Party A would then win all the seats.  Depending on precisely how “competitiveness” is interpreted, then, the Colorado commission might operate with a very different sense of “fair districts” than that in Missouri or Michigan or Utah.

My general aim here is to highlight that there are different understandings of what makes a districting plan fair or neutral, with the process-oriented and outcome-oriented approaches defining two fundamentally different, conflicting understandings.  By looking closely at the substantive standards in these pending voter initiatives, we can see how these different conceptions play out in practice among commission proponents in different states.  And given the ambiguities in some of these proposals, we can start anticipating where the conflicts will emerge, and potentially litigation will arise, should any of these initiatives succeed in November.

No other country with single-member election districts like ours leaves the power to draw these districts in the hands of the most politically self-interested actors, the politicians whose power and seats will be affected. I have long argued this is an inherently pathological situation for a democracy and that we should move this power into the hands of independent commissions. But shifting to commissions cannot avoid the fact that substantive choices must still be made about how we ought to define fair maps and what criteria commissions, or any other redistricting body, ought to follow in order to design fair maps.

The proposals on the ballot in these four states agree that redistricting should be taken out of the hands of self-interested state legislatures. But they show that on the deeper question – what makes a map fair – there remain differences of view.